i r ; I i ■■ 4 • i ■ ■.r? ^E«^^^*^B X m • .,. . . - . i ; i . d^^'^T'X o o o MBL/WHOl Library mfs 3 o ^L == n 5 X -i 00 LD s TOE( ^^^^ i_j CSL Th ENCYCLOPEDIA .f MICROSCOPY Edited by |— GEORGE L CLARK Research Professor of Anahjtical Chemistry/, Emeritus, Universihj of Illinois, Urbana, Illinois REINHOLD PUBLISHING CORPORATION, NEW YORK CHAPMAN & HALL, LTD., LONDON Copyright © 1961 by REINHOLD PUBLISHING CORPORATION All rights reserved Library of Congress Catalog Card Number 61-9698 Printed in the United States of America by The Waverly Press, Inc., Baltimore, Md. CONTRIBUTING AUTHORS Kenneth W. Andrews The United Steel Companies Ltd., England B Albert V. Baez Harvey Mudd College Thomas F. Bates Pennsylvania State University John H. Bender Los Alamos Scientific Laboratory James R. Benford Baiisch & Lomb, Inc. C. G. Bergeron University of Illinois A. Bergstrand Sabhatsbergs Hospital, Sweden F. P. BOHATIRCHUK University of Ottawa, Canada W. BOLLMANN Battelle Memorial Institute, Switzerland Willi AM. A. Bonner Southwestern Medical School of the Uni- versity of Texas R. BOEASKY University of Illinois D. E. Bradley Associated Electrical Industries, England Floyd Dunn University of Illinois J. Dyson Associated Electrical Industries, England N. A. Dyson University of Birmingham, England Wesley B. Estill Sandia Corporation M. L. Feeney University of California Medical Center F. Gordon Foster Bell Telephone Laboratories James A. Freeman U. S. Public Health Service William J. Fry University of Illinois Charles C. Fulton Department of Health, Education and Wel- fare L. K. Garron University of California Medical Center R. L. Gregory University of Cambridge, England Richard D, Cadle Stanford Research Institute George L. Clark University of Illinois Germain Crossmon Bausch & Lomb, Inc. D Norman L. Dockum General Electric Company H M. E. Haine Associated Electrical Industries, England Olle Hallen University of Goteborg, Sweden F. A. Hamm Minnesota Mining and Manufacturing Company Burton L. Henke Pomona College lU Contributing Authors L. L. Hundley Soiithwestern Medical School of the Uni- versity of Texas I Shixya Inoue W Dartmouth Medical School J. ISINGS Central Laboratory T.N.O., Holland Kazuo Ito Ja'pan Electron Optics Laboratory Co.. Ltd., Japan William Johnson The United Steel Companies Ltd., England Raymond Jonnard The Prudential Insurance Company of America B. E. Juniper Oxford University, England L. Marton National Bureau of Standards Ludwig J. Mayer General Mills, Inc. C. W. Melton Battelle Memorial Institute P. O'B. Montgomery Southwestern Medical School of the Uni- versity of Texas Erwin Muller Pennsylvania State University Jean M. Mutchler Linde Company N Joseph D. Nicol Michigan State University W. C. Nixon University of Cambridge, England J. Nutting University of Cambridge, England K Ernest H. Kalmus University of California P. M. Kelly University of Cambridge, England Charles J. Koester American Optical Company MoToi Kumai University of Chicago William R. Lasko United Aircraft Corporation Donald E. Lasko wski Armour Research Foundation of Illinois Institute of Technology K. Little Nuffield Department of Orthopedic Surgery, England M W. K. McEwEN University of California Medical Center Ronald H. Ottewill University of Cambridge, England D. H. Page British Paper and Board Industry Re- search Association, England H. H. Pattee Stanford University Ong Sing Poen Pomona College R Johannes A. G. Rhodin Bellevue Medical Center T. G. Rochow American Cyanamid Company George L. Royer American Cyanamid Company J. Salmon Laboratoire de Biologic Vegetate I, France IV Contributing Authors R. L. deC. H. Saunders Dalhousie University, Canada C. M. Schwartz Battelle Memorial Institute ' D. Scott National Engineering Laboratory, Scotland Michael Seal University of Cambridge, England K. C. A. Smith Pulp and Paper Research Institute of Canada, Canada John D. Steely Los Alamos Scientific Laboratory J. H. Talbot Transvaal and Orange Free State Chamber of Commerce, Africa V. J. Tenner Y Motorola Corporation w Ernest E. Wahlstrom University of Colorado Masaru Watanabe Japan Electron Optics Laboratory Co., Ltd., Japan Erwin K. Weise University of Illinois Alvar p. Wilska University of Arizona R. E. Wright Shell Chemical Company PREFACE With genuine pleasure and pride this "Encyclopedia of Microscopy" is presented as the fourth in a series of Reinhold contemporary^ integrated compilations of rapidly developing areas of science, following the "Encyclopedia of Chemistry" (1957), the "Encyclopedia of Chemistry Supplement" (1958), and the "Enc^'^clopedia of Spectroscopy" (December, 1960). Actually, "Spectroscopy" and "^Microscopy" were planned, projected, assembled and edited concurrently as twin volumes, but of necessity there was an interval of a few months in publication of the two, or a slightly delaj'ed birth, as it were, of the second twin. These two great instrumental techniques, valuable in so many disciplines, are so inti- mately related and interwoven that the simultaneous development of encyclopedias was the most logical procedure. Even though microscopy as a science is about three centuries old and spectroscopy only one, the common background and origins are well exemplified in the respective historical articles by Professor E. K. Weise. It may not be surprising, then, that one Preface was written originally for both Encyclopedias. This has appeared already in "Spectroscopy," and it is hoped that users of this volume will have the opportunity to read this more extended introduction to the pair of volumes, as well as to browse in a kindred science. The "Encyclopedia of Microscopy" is the fruit of the joint efforts of a trulj' international team of dedicated microscopists — English, Scotch, Canadian, South African, French, Ger- man, Swiss, Dutch, Swedish, Japanese and American — and it is this fact that gives such unique flavor, value and good will to the able and devoted coverage of a science which is as wide and boundless as the world itself. This Encyclopedia, of course, is a mosaic of 26 kinds of microscopj^ alphabetically arranged, and in most cases wdth numerous alphabetical subtopics under each. The numerous illus- trations in this picture book — diagrams of all kinds, photographs of microscopes and related instruments, and the micrographs made with them — speak eloquently' for themselves as superb art as well as science, heretofore unpublished in most cases. Carefully chosen lists of general and cross references, it is hoped, will add greatly to the value and usefulness of this volume to inquisitive students and laymen seeking information and illumination in this area which extends the powers of men's vision a millionfold or more, and to the experts who continue to build an ever-new science. The Editor is indebted personally to all the eminent scientists throughout the world who have contributed vitally important advice and encouragement, as well as one or more arti- cles based on experience and devotion in a specialized field. The entire list of authors pre- sented in the front pages is indeed a Roll of Honor. Special mention is due to colleagues at the University of Illinois — Professors E. K. Weise of the Department of Mining and Metal- lurgical Engineering, and R. Borasky, Director of the Electron Microscope Laboratory; and to the two loyal and able secretaries serving consecutively, Mrs. Ruth Tuite (1958-9) and Mrs. Claretta ]\Ietzger (1959-60), with whose help the entire task of planning, organiz- ing and editing has somehow been accomplished. The patience, enduring faith, guidance and technical aid by the publishing staff, especi- ally G. G. Hawley, Executive Editor, and Alberta Gordon, were indispensable factors in bringing both Encyclopedias to material fruition, and in looking ahead to new worlds of vu Preface science to bring between the covers of future members of this series of encyclopedias. The deep personal challenge and satisfaction to the Editor upon becoming a Professor Emeritus may somehow be reflected in the last paragraph of the Preface to the "Encyclopedia of Spectroscopy." George L. Clark Urbana, Illinois January, 1961 Vlll CONTENTS AUTORADIOGRAPHY 1 Autoradiography of Tissue, Norman L. Dockum 1 Shadow Autoradiography, E. Borasky 11 CHEMICAL MICROSCOPY 13 Alkaloids and Alkaloidal-Type Precipitatiox, Charles C. Fulton 13 Chemical Microcrystal Identifications, Charles C. Fulton 20 Forms of Microcrystals, Charles C. Fulton 37 History, Charles C. Fulton 38 Mixed Fusion Analysis, Donald E. Laskowski 42 Nitrogen-Bonded Radicals: Identification, Charles C. Fulton 46 Observing Microcrystals, Charles C. Fulton 51 Opium, Origin of, Charles C. Fulton 55 Purpose, Charles C. Fulton 56 Quinoline as a Reagent, /. M. Mutchler 57 Reagents for Microcrystal Identifications, Charles C. Fulton 58 Sympathomimetics and Central Stimulants, Charles C. Fulton 65 ELECTRON MICROSCOPY 72 Aerosols Containing Radioactive Particles, R. Borasky 72 Blood, James A . Freeman 77 Botanical Applications, D. E. Bradley 80 Cell Ultrastructure in Mammals, Johannes A. G. Rhodin 91 Ciliated Epithelia Ultrastructure, Johannes A. G. Rhodin 116 Colloids, Lyophobic, R. H. Ottewill 123 Crystal Lattice Resolution, G. L. Clark 145 Dislocations in Metals. See Transmission Electron Microscopy of Metals — Dislocations and Precipitation, p. 291 Electron Optics: Electron Gun and Electroalignetic and Electrostatic Lenses, M. E. Haine 147 Fibers (Textiles). See General Microscopy, p. 343 History of Electron Optics, L. Marlon - 155 Im.\ge Form,\tion Mechanism, L. Marlon 159 Ividney Ultrastructure, Johannes A. G. Rhodin 163 Leaf Surfaces, B. E. Juniper 177 Metals by Transmission, P. M. Kelly and J. Nutting 181 Microtomy. See General Microscopy, p. 385 Minerals, Thomas F. Bates 187 Paint Surface Replica Techniques, W. R. Lasko 200 ix Contents Pathology: Kidney, Anders Bergstrand 206 Plastics. See General Microscopy, p. 390 Pulp and Paper. See General Microscopy, p. 394 Reflection I, D. H. Page 220 Reflection II, Michael Seal 223 Replica and Shadowing Techniques, D. E. Bradley 229 Replicas, Removal from Surfaces, V. J. Tennery, C. G. Bergeron, and R. Borasky . 239 Resinography. See p. 525 Scanning, K. C. A. Smith 241 Selected Diffraction, J. H. Talbot 251 Snow Crystal Nuclei, Motoi Kumai 254 Special Methods, Masaru Watanahe and Kazuo I to 259 Specimen Preparation — Special Techniques at Los Alamos 270 A Modified Aluminum Pressing Replica Technique, J. H. Bender and E, H. Kalmiis 270 Preparation of Aerosols, E. H. Kalmus 272 Dispersion of Aerosols, J. H. Bender and J. D. Steely 272 Sorting and Fractography of Particles, J. H. Bender and E. H. Kalmus 273 Uncurling Carbon Replicas, W. B. Estill and E. H. Kalmus 274 Staining, Electron, R. Borasky 274 Tissues (Connective), Bones and Teeth, K. Little 276 Transmission Electron Microscopy of Metals — Dislocations and Precip- itation, W. Bollmann 291 Unsolved Problems, Franklin A. Hamm 307 Wear and Lubrication, D. Scott 308 Wilska Low- Voltage Microscopy, Alvar P. Wilska 314 ELECTRON MIRROR MICROSCOPY, Ludwig Mayer 316 FIELD EMISSION MICROSCOPY, Erwin W. Mailer 325 FLUORESCENCE MICROSCOPY, G. L. Clark 332 FLYING SPOT MICROSCOPY, P. O'B. Montgomery, Wm. A. Bonner, and L. L. Hundley 334 FORENSIC MICROSCOPY, Joseph D. Nicol 338 GENERAL MICROSCOPY 343 Fibers (Textile), J. I sings 343 Industrial Research, C. W. Melton and C. M. Schwartz 363 MiCROScopiSTS AND RESEARCH MANAGEMENT, Gcorge L. Royer 381 Microtomy, Olle Hallen 385 Plastics, J. I sings 390 Pulp and Paper, /. I sings 394 HISTORADIOGRAPH. See X-Ray Microscopy Contents INDUSTRIAL HYGIENE MICROSCOPY, Germain Grossman 400 INFRARED MICROSCOPY, G. L. Clark 411 INTERFERENCE MICROSCOPY 412 Fibers (Textile). See General Microscopy, p. 343 Industrial Research, Application to. See General Microscopy, p. 363 Instrument Classification and Applications, J. Dyson 412 Plastics. See General Microscopy, p. 390 Pulp and Paper. See General Microscopy, p. 394 Theory and Techniques, Charles J. Koester 420 LIGHT (OPTICAL) MICROSCOPY 434 Comparison Microscopes. See Engineering Microscopes, p. 438 Design and Construction of the Light Microscope, James R. Benford 434 Engineering Microscopes, G. L. Clark 437 Fibers (Textile). See General Microscopy, p. 343 Hardness Tests. See Engineering Microscopes, p. 438 Heating Microscopes. See Engineering Microscopes, p. 438 Industrial Research, Application to. See General Microscopy, p. 363 Introscope. See Engineering Microscopes, p. 438 Magnetography: The Microscopy of Magnetism, F. Gordon Foster 440 Measuring Microscopes. See Engineering Microscopes, p. 439 Optical Theory of the Light Microscope, James R. Benford 445 Origin and History, E. K. Weise 454 Particle Size and Shape Measurements and Statistics, Richard D. Cadlc 464 Plastics. See General Microscopy, p. 343 Projection Microscopes. See Engineering Microscopes, p. 438 Pulp and Paper. See General Microscopy, p. 343 Schmaltz Profile Microscope. See Engineering Microscopes, p. 438 Stereoscopic Microscope. See Engineering Microscopes, p. 439 METALLOGRAPHY 468 Industrial Research, Applications To. See General Microscopy, p. 343 Transmission Electron Microscopy of Metals — Dislocations and Precipi- tations. See p. 291 Wear and Lubrication. See Electron Microscopy, p. 308 MICROMETRON automatic microscope, G. L. Clark 468 MICRORADIOGRAPHY. See X-Ray Microscopy, p. 561 OPTICAL MINERALOGY, Ernest E. Wahlstrom 470 Petrographic Thin Sections, R. E. Wright 473 PHASE MICROSCOPY 476 Anoptral Microscope, A. Wilska 476 Fibers. See General Microscopy, p. 343 xi Contents Industrial Research, Applications to. See General Microscopy, p. 363 Plastics. See General Microscopy, p. 390 Pulp and Paper. See General Microscopy, p. 394 Theory and Microscope Construction. See Optical Theory of Light Mi- croscope, p. 445 POLARIZING MICROSCOPE 480 Basic Design and Operation. See Optical Theory of Light Microscope, p. 445 Design for Maximum Sensitivity, Shinya Inoue 480 Fibers (Textiles). See General Microscopy, p. 343 Industrial Research, Application to. See General Microscopy, p. 363 Plastics, See General Microscopy, p. 390 Pulp and Paper. See General Microscopy, p. 394 REFRACTION OF LIGHT, REFRACTOMETRY AND INTERFEROMETRY 485 Angle Refractometry, R. Jonnard 485 History of Light Refraction, R. Jonnard 494 Interferometric Methods, R. Jonnard 502 Refractometric Applications, R. Jonnard 515 RESINOGRAPHY, T. G. Rochow 525 STEREOSCOPIC MICROSCOPY 538 Basic Design, Operation and Use, J. R. Benford 538 Engineering Microscope. See Light (Optical) Microscopy, p. 434 "Solid-Image" Microscope, Richard L. Gregory 540 TELEVISION MICROSCOPE, G. L. Clark 542 ULTRAMICROSCOPY 544 Design and Operation. See Optical Theory of the Light Microscope, p. 445 ULTRASONIC ABSORPTION MICROSCOPE, Floyd Dunn and Wm. J. Fry 544 ULTRAVIOLET MICROSCOPY 548 Basic Principles and Design, J. R. Benford 548 Color Translating TV Ultraviolet Microscope, G. L. Clark 550 Image Formation by a Fresnel Zone Plate, Albert V. Baez 552 X-RAY MICROSCOPY 561 Bone Structure and Aging by Contact Microradiography 591 See Medico-Biologic Research by Microradiography, p. 591 Contact Microradiography, H. H. Pattee 561 Diffraction Microscopy, G. L. Clark 569 Eye Research Applications, W. K. McEwen, M. L. Feeney, and L. K. Garron ... 571 Xll Contents Geological, Mineralogical and Ceramic Applications of Microradiography, W. Johnson 574 Grainless Media for Image Registration, G. L. Clark 581 Histology by the Projection Microscope, R. L. deC. H. Saunders 582 Inter-Relation of Techniques for the Investigation of Materials, K. W. Andrews 586 Iron and Steel Applications of Microradiography, K. W. Andrews and W. Johnson 587 Medico-Biologic Research by Microradiography, F. Bohatirchuk 591 Microangiography with the Projection Microscope, R. L. deC. H. Saunders. . 627 MicROFLUOROScoPY. See Contact Microradiography, p. 561 Plant Microradiography, ./. Salmon 636 Point Projection X-Ray Microscopy, W. C. Nixon 647 Production of Continuous and Characteristic X-Radiation for Contact and Projection Microradiography, N. A. Dyson 653 Projection Microscopy, Ong Sing Poen 661 Reflection Microscopy (Kirkpatrick), G. L. Clark 672 Two-Wave (Buerger) Microscope, G. L. Clark 674 Ultrasoft X-Ray Microscopy, Burton L. Henke 675 Vascular and Dental Applications of Projection Microscopy. See Micro- angiography, p. 627 Xlll Autoradiography AUTORADIOGRAPHY OF TISSUE* This article discusses techniques used for autoradiography of human and animal tis- sues and, to a limited extent, plant tissues. Methods are outlined which will enable those relatively inexperienced in autoradi- ography to plan an experiment in\'olving alpha, beta and gamma emitters and to pro- ceed with the experiment confident of ob- taining meaningful autoradiograms. A cur- rent bibliography for the period 1954 to 1959 is that of Johnston (1). Boyd (2) has cited numerous literature references. Foreign journals often cany detailed papers relating to autoradiography which should not be overlooked as a source of information. The technique which produces an image on a photographic plate or film when radio- active material is opposed to it is called autoradiography. The result of the exposure of an emulsion to a radioactive specimen is called an autoradiogram. The autoradiogram supplies a graphic record of the sites of depo- sition of radioactive isotopes within or on a tissue and may be macroscopic, as in the case of plant leaves, or in some cases micro- scopic, at or below the cellular level. Au- toradiography of tissues containing alpha, beta and gamma emitters may be performed. Suggested Autoradiographic Tech- niques for Alpha Emitters; Pluto- nium In industries where contamination of per- sonnel with radioactive substances is pos- * Work performed under Contract No. AT(45- 1)-1350 between the Atomic Energy Commission and the General Electric Company. sible, low level detection procedures are necessary. Plutonium, an alpha (5.14 Mev) emitting radionuclide with some gamma ra- diation and a 24,300 year half-life, may be used as an example to illustrate one auto- radiographic technique by which either par- ticulate or soluble material may be graphi- cally localized. A 24-hour sputum sample (3) from a per- son known to have inhaled plutonium was taken and fixed in 10% formalin, as was a human biopsy of skin removed from a punc- ture wound in the hand. The samples were dehydrated in "Cellosolve" (glycol mono- ethyl ether), cleared in xylene, and em- bedded in paraffin. Sections were cut and floated on a water bath, transferred by a clean slide to a crystalhzing dish containing distilled water. Working under light filtered by a Wratten OA filter, a 5 /x NTA emulsion coated on a 1 x 3 inch slide with a thin gela- tin protective "T" coat was shpped under the section. The excess water was drained on to filter paper; the slides were then placed in a Hght-tight plastic box made for this pur- pose. A small vial of a desiccant (CaS04) lightly closed with a cotton plug was added. The appearance of the tracks from pluto- nium which was in solution is characterized in Figure 1 (a human sputum specimen), where individual alpha tracks proceed in a straight line. The appearance of the tracks in sputum when plutonium is in the par- ticulate form is illustrated in Figure 2, in which alpha tracks arise from a central point, with some random single tracks in the field. Thirty-three sections from the plutonium- AUTORADIOGRAPHY *# 50/i %^ 4 i i I "^yft Fig. 1. Autoradiogram of stained section of human sputum illustrating diffuse alpha-track pattern of plutonium in solution. {Courtesy Stain Technology^) . contaminated skin biopsy (4) were auto- radiographed by the above processing method. All of the two-hundred fifty indi- vidual skin sections were scanned in a thin window alpha detector. Radioanalysis showed that the entire skin biopsy specimen contained 0.0046 microcurie and the indi- vidual sections varied from to 362 disinte- grations per minute. The appearance of one autoradiogram of a human skin section is illustrated in Figure 3, in which alpha tracks may be seen arising from particles. In addi- tion, random alpha tracks may be seen. Suggested Autoradiographic Tech- niques for Beta Emitters; I^^S Ru^**^, Sr«o, and P^^ I^^i, with an eight day half -life, is predomi- nantly a beta emitter with some gamma emission. It was administered to sheep in an acute and chronic feeding experiment (5) to determine the effect of radioiodine on the thyroid gland of grazing animals. The auto- radiographic response of the thyroids of some of these animals is of interest. Thyroid samples were routinely fixed in Bouin's solution for from 8 to 12 hours, fol- lowed by dehydration in "Cellosolve" and nor- mal parafl&n embedding. Adjacent sections were selected from the ribbons. One section was stained routinely with hematoxylin and eosin; the other was floated into a crystalliz- ing dish containing distilled water, from which it was floated onto a slide bearing a 5 fj. NTB emulsion plus a protective "T" coat. The slides, for animals which had received 100 nc P'^^ I.V. and which were sacrificed after 1 hour were then exposed for 5 days; the slides from animals receiving 100 nc I.V. in 0.9% saline and sacrificed after 4 hours were exposed for 53-^ days ; and slides for ani- FiG. 2. Autoradiogram of stained section of human sputum illustrating star formation of alpha tracks originating from plutonium particle in cen- tral position. {Courtesy Stain Technology^). ALTOKVDIOGRAPHY OF TISSUE ^.^•^ 100 /x Fig. 3. Autoradiogram of human skin showing distribution of plutonium and alpha tracks ap- proximately 800 M beneath stratum corneiun. Hema- toxylin and eosin preparation. (Courtesy Acta Radiological) . mals which had received 5 mc per day for 129 days were exposed for 9 days. The last animal was sacrificed during gestation and had previously suckled its dam for 4 months. The dam had also been receiving 5 mc a day, so that some additional I^^^ was received by the lamb through the milk. The autoradiograms were exposed in light- tight boxes containing "Drierite." The slides were warmed to room temperature and de- veloped in D-19 at 18°C for 5 minutes, water rinsed and fixed in x-ray fixer for 30 minutes. They were water- washed and stained with hematoxylin and eosin. Sections that adhere to the emulsion can usually be stained satis- factorily, if the staining procedure is not pro- longed to the point where overstaining of the emulsion is apparent. The appearance of the autoradiograms of sheep thyroid tissue from animals adminis- tered 100 /xc intravenously and sacrificed in 1 hour, illustrated in Figure 4, shows the grain response to be relatively uniform over both the colloidal areas and the epithelial cells. The colloidal area in animals admin- istered 100 jiQ. and sacrificed in 4 hours is shown in Figure 5. The grain distribution is more pronoimced over many of the col- loidal areas, indicating a variable uptake of P^^ by the colloid in some of the follicles. For animals fed 5 /xc a day for 129 days (Figure 6) there is an even distribution of grains over the colloidal areas in thyroid follicles which are greatly decreased in size from the normal. Areas showing no grain density indicate the edema surrounding the few remnant follicles that are present. The microfollicles register a limited grain den- sity. Little or no interstitial tissue is present. Si,— ■ 3P^ - ***.- 7^' ■* ■»^ •1* ■» r .i ■ 1 : • r^ \ * #.■ 4 « * 1 1. ; - » ^ ^jHKh ,v ■'•^ ■* ^ ^ , ^ ^ ■^ J r^ •-. „__ ^ w ' , # r^ B- *« «. ■> Pi i i -^ h- -50/i-^, 4," Fig. 4. Autoradiogram of hematoxylin and eosin stained sheep thyroid. Animal was adminis- tered 100 /iC of I'^i I.V. and sacrificed in one hour. Limited grain density increase over colloidal area and epithelial cells. AUTORADIOGRAPHY case the ruthenium was administered in an insokible particulate form. Ru^°^ particles were administered to mice by intravenous and intratracheal injection in 0.1 % aqueous dispersions of the wetting agent, Tween 80 (6). The animals were sacri- ficed after 100 to 420 days exposure to beta radiation. One experiment in autoradio- graphic quantitation was conducted on a single mouse into which 47 /xc of Ru^"® par- ticles, ranging in size from 0.5 to 0.8 micron, were injected into the tail vein of the mouse. A portion of these particles lodged in the lungs of the animal. The lungs were excised after 100 days exposure and fixed in 70% ethyl alcohol. Autoradiographic processing was standard, as outlined above. The re- sult was a series of sections through the entire lung; one section was stained with orcein, one unstained autoradiogram fol- » t-IOOft-) Fig. 5. Autoradiogram of sheep thyroid four hours following administration of 100 nc of I'^^ Hematoxylin and eosin stain. Specific increase in grain density over various colloidal areas indicates variable uptake by the follicles. It is important that the investigator be aware of the type of response to be expected after processing certain types of tissue and report the appearance of representative au- toradiograms that are taken to illustrate the uptake of the radioisotope at selected times following administration. Specifically, the autoradiographic response in the thyroids of the experimental animals mentioned is that of single grains diffusely covering specified areas of the thyroid. Other types of emul- sions will register the location of I^^^ as tracks rather than diffuse grains. Changes in the development procedures are necessary to register the I^^^ as tracks rather than grains. It is interesting to compare at this point the autoradiographic response of Ru^"®, an- other beta emitter having some gamma emission, and a half-life of one year. In this h*\ '4*' «• • V ^^ % **' ^ * i • • " " " ■ * ■• J * •» % 50/X- .-^^ :# ^ t Fig. 6. Autoradiogram of sheep thyroid 129 days following administration of 5 /xc of I"' per day. Negative grain density is apparent over areas of edema, with a limited uptake in the colloidal areas of the microfollicles. 4 AUTOKADIOGRAPIIY OF TISSUE lowed it, and this in turn was followed by a section stained by the Mallory method, to register connective tissue locations in rela- tion to the "spot diameter" regponse of the unstained autoradiograms. A "spot diam- eter" is the grain response of the emulsion to a single radioactive particle. The staining sequence was followed through the entire lung. All of the sections were measured for Ru^"^ content on a mica window beta counter. The insoluble ruthenium particles elicited a specific "spot diameter" response in the NTB emulsion. This was a well defined cir- cular area of developed emulsion grains over the precise location of the deposited par- ticles, as illustrated in Figure 7. Knowing the total activity density for a specific slide, the activity density of a single particle was taken as proportional to the diameter of the darkened area of the emul- sion. Thus, if a number of particles were present on a single slide, the total counting rate was compared with the sum of all of the darkened areas on that slide. Therefore, each particle could be assigned an activity density evaluation directly proportional to the size of the individual "spot diameter" measurement of the emulsion. The "spot diameter" darkenings of the grains of the emulsion varied in size and ranged from 8 fj, to 170 fjL. A quantitative straight -line func- tion exists between the sum of the spot diameters or a single-spot diameter and the total activity of each autoradiogram. This particular technique provides a means of measuring the radioactivity of a single radioactive particle or many particles within tissue when the exposure, processing and development are standardized. The par- ticles occur in a pattern of distribution such that the accumulated dose from nearby par- ticles is sufficient to initiate fibrosis. By means of the serial sections stained specifi- cally for connective tissue, and the autora- diograms, it is possible to reconstruct the tissue in depth and study the relationship ^ 4 GRAIN DENSITY Fig. 7. Stained lung-autoradiogram prepara- tion illustrating deposition of Rui^^Qa particles with resultant spot diameter darkenings. (Courtesy J. Biol. Phot. Assoc.^^). of the fibrosis to the specific location of the deposited radioactive particles. The autoradiographic technique for Sr^oS04 in lung tissue of mice, administered 1 /iC per animal by inhalation, is mentioned here for comparison with the foregoing meth- ods. Sections from lung tissue fixed in 80% ethyl alcohol were processed in a manner similar to lung tissue containing Ru*"^ par- ticles (7). The film used in this case, how- ever, was 25 n No-Screen x-ray, single coated on 1 X 3 inch slides. These slides had a pro- tective coating. As counting techniques indi- cated a low activity densitj^ for the sections, they were exposed for 8 weeks. The use of X-ray emulsion made it possible to register activity densities that were approaching the lower limits of detection by normal counting methods. The autoradiogram of the lung section containing Sr^", illustrated in Figure 8, shows both a diffuse and particulate response to the radioactive material. The adjacent his- autok\i)I()<;h\phy ■7 ;' ■ v»-lf % ,^ PARTtCfe€- -, REAC; h-6Q0/i."i Fig. 8. Autoradiogram of lung section from chronic Sr'''S04 inhalation experiment. Particulate and diffuse distribution with heavy background grain density of x-ray emulsion visible. (Courtesy J. Biol. Phot. Assoc.^^). tological section is included, as Figure 9, for comparison purposes. A noticeable back- ground of grains, larger than in the NTA and NTB emulsions, is observable. This is due in part to the 25 ^ thickness of the x-ray emulsion and in part to the greater sensitiv- ity. Even greater sensitivity could have been obtained had the film used been of the double coated type. However, the image rendered would have been more diffuse because the emulsion layers are coated on either side and separated by an acetate base. In this case detail has been sacrificed to record low activity density deposition of diffuse and particulate Sr^". The next example is that of a beta emitter, F'\ which has a 14.5 day half -life (8). A gross, fast survey method for the detection of P^- administered to rats by intraperitoneal injection of 2.5 /xc per gram of body weight is outlined. The rats were sacrificed 24 hours following administration of the P^^ ^s Nao- HP^^o^ The emulsion of choice in this case was 25 n, single layer, No-Screen x-ray emulsion on 1 x 3 inch slides. Detail was sacrificed for rough localization of P^^ by using an emulsion with a grain size varing from 3 to 5 /i. The increase in the sensitivity of the emulsion and the thickness insured a positive record- ing of the radioactive sites, with a minimum exposure period. Noticeable disadvantages with this techniciue are the large grain size which makes cellular autoradiography im- possible. The extreme thickness of the emul- sion makes illustration of the recorded data on the film difficult, except by macrographic methods. Figure 10 is an autoradiogram of a rat ovary section from an animal sacrificed 24 hours following intraperitoneal administra- '^W ,»*-e<,y.4-" I'i't^?'. %. I— 500/i-li,, Fig. 9. Hematoxylin and eosin stained histo- logical section adjacent to Fig. 8, for orientation purposes. {Courtesy J . Biol. Phot. Assoc.^^). ALTORADIOGRAPHY OF TISSUE .-" >~^jjgg «»>- -*'*rie'i-'«Trs '."^s^K-? •-»»• * # Fig. 10. Autoradiogram of rat ovary 24 hours after administration of 2.5 juc of P'^ pgj. gram of rat. Concentrations of P^^ appear over cellular ele- ments lining the follicular cavity, surrounding the ova and in some corpora lutea. {Courtesy J . Biol. Phot. Assoc.^^). tion of P^2_ ^n increased grain density is noticeable in the area of three developing ova. The autoradiogram may be compared with the adjacent hematoxylin and eosin stained section in Figure 11. In direct contrast with this gross survey method for the localization of P^- is the ex- treme detail and precise localization ob- tained by Guidotti (9). Guidotti mounted paraffin sections, 2 to 4 ;u thick, which were stained, dehydrated, and allowed to air dry. These were given a thin coating of 1 % "Plexi- glas" solution in chloroform. The chloroform was allowed to evaporate completely in a dry atmosphere. An emulsion having a dried thickness of 100 to 150 microns, was pre- pared from Ilford G-5 type emulsion in gel form and glued to the section by means of a 15 % solution of shellac in absolute alcohol. The surface of the emulsion was then cleaned with absolute ethyl alcohol, to remove the impermeable layer. This method allows the fixing and staining of the tissues by means of all of the reagents commonly employed in histology, without any damage to the emulsion. The exposure of the tissue was for 24 hours at 2°C. The emulsions were proc- essed by the Temperature Development Method. Good adhesion and minimum sep- aration between specimen and emulsion are obtained, thus permitting reliable extrapo- lation of the electron tracks from P^-. The Temperature Development Method of Dilworth, Occhialini and Payne, 1948, and Dilworth, Occhialini and Vermaesen, 1950, as modified by Guidotti, is worthy of specific mention. Essentially it consists of soaking the slides in distilled water for 30 to 40 min- utes, beginning at room temperature, and '^ Fig. 11. Stained histological section adjacent to Fig. 10 of rat ovary 24 hours after administra- tion of 2.5 Aic of P'2 per gram of rat. Three develop- ing ova are seen in central portion. (Courtesy J. Biol. Phot. Assoc.'^). AUTOH ADIOGR A1»IIY then lowering the temperature gradually to about 4°C. The cold phase of development is for about 30 minutes. The developer used in this case was Amidol. The cold developer is poured off and the surface of the slides dried with filter paper. Warm slides to 22° to 24°C in a thermostatically controlled tank. The temperature of the warm stage must be chosen by trials. The slides are then placed in a stop bath of 0.2% acetic acid solution for about 30 to 40 minutes at 4°C. Fixation is at 4°C in a 40 % sodium thiosul- phate solution until the emulsion is clear. The emulsion is thoroughly washed at the low temperature and the emulsion allowed to dry slowly. The preparation is then capped with a thin coverslip. One of the reasons for Guidotti's success with P^- autoradiography is the varied tem- perature development allowing all layers of the emulsion to be penetrated by the devel- oper; another is the thinness of the section, which is considerably less than most routine preparations. This allows the site of origin of the beta tracks to be more accurately located. Fig. 12. Autoradiogram of rat ovary and adja- cent tissue illustrating even diffuse distribution of Cs"'. Cs^^'^, a beta and gamma emitter, with a half -life of 30 years, is an example of a water- soluble isotope that presents difficulties re- garding methods for autoradiographic regis- tration. The usual fixing fluids leach Cs''^ from tissues as will any contact with water; there- fore, a mixture of 80% acetone and 20% benzene was used as a fixative (10). This combination was decided upon after ex- perimentation in which the leaching loss was determined by radiochemical analysis. There was a minimum cesium loss when this fixa- tive was used. Tissues were then processed through "Cellosolve" and benzene and blocked in paraffin. Since the sections could not be floated on water for transfer to the emulsion, sections were attached to the No-Screen x-ray emul- sion by warming the slide until the emulsion became tacky, then the sections were at- tached by finger pressure. The slides were then exposed, the paraffin dissolved by im- mersion in xylene. The slides were run down to water from "Cellosolve" and the autora- diograms stained and the slides capped. Figure 12 is an illustration of Cs'^^ deposi- tion in a rat ovary. It will be noted that there is a homogeneous distribution in the ovary and the surrounding tissue. There is no evi- dence of leaching into the emulsion area not covered by the tissue. Figure 13 is the adja- cent section for histological comparison. Greater definition could have been obtained by using an NTB emulsion. Limited Autoradiographic Techniques for Plant Tissue An interesting experiment performed by Hungate et at. (11) involved exposure of plants to fallout from the experimental burn- ing of an irradiated fuel element. The fission products resulting from the burning of the fuel element were carried by the air to plants located downwind from the burning site. Later the plant leaves were exposed to double-coated Type KK indus- 8 alt()hai)io(;raphy of tissue trial x-ray film. The gross autoradiographic result of this exposure of two plant leaves is illustrated in Figure 14. It will be noticed that there is a general darkening over the entire leaf with an increased density at its periphery. In addition, there are small black circular areas of increased grain density that are relatively indistinct. These circular areas of increased grain density suggest that some of the fission products were in the particulate form. In order to substantiate the observation regarding specific deposition of particles, additional 1x3 inch plates coated with a 100 M NTB^ emulsion and a "T" coat were exposed to portions of the plant leaves for 4 days and developed in D-19 developer. These also demonstrated a particle response that was more detailed than that observed with the double-coated x-ray emulsion. The detailed microscopic particle response of fission material on plant leaves is illustrated in Figure 15. In this case the autoradio- 7 ; \$* Fig. 14. Glu^^ auloradiograms of plant leaves following e.xposure by air of the leaves to fission products. An over-all darkening with increased density at the peripher}^ is noticed. Some small circular spot densities can be seen. graphic technique demonstrated the presence of particles on lea^'es that was not detectable by routine counting procedures. Fig. 13. Stained tissue section of ovary and re- lated tissue adjacent to section used for autoradi- ography of Cs"' in Fig. 12. General Discussion Up to this point practical problems involv- ing autoradiography have been discussed. ]\Iany specific applications have been AUTORADIOGRAPHY >^>.-^^ ^'^:;mt PARTICLE REACTION ^-lOO^^ Fig. 15. Microscopic autoradiographic re- sponse of leaf exposed to fission products. Specific spot grain responses indicate that some of the ma- terial was in particulate form. omitted. Among them is the classical work involving plutonium deposition in the bones of dogs administered plutonium. This is the excellent work of Jee (12) in which the results of autoradiographic methods involv- ing hard bone are discussed in detail. Autoradiographic techniques have been applied to electron microscopy. O'Brien and George (13) have examined sectioned yeast cells, previously suspended in a Po^^*^ solu- tion. In this case the specimen grids were coated with the emulsion by touching them to a drop of diluted NTA emulsion. Borasky and Dockum (14) examined Pu^^^ particles contained in an aerosol sample collected on "Formvar" coated grids. The grids bearing the particles were chrome-shadowed and placed on "Lucite" plugs held upright in wells drilled in a metal block. A small wire loop was placed in diluted NTA gel emulsion and retracted, thus forming a very thin layer of emulsion. The loop bearing the emulsion was then lowered over the specimen grid and the loop retrieved by lifting the plug and withdrawing the loop from beneath the plug. The grids were exposed for 4 hours in a light- tight box when high activity specimens were examined. The emulsion-coated grids on the "Lucite" plugs were immersed in D-19 de- veloper for 5 minutes, then washed in dis- tilled water for one minute, after which they were fixed in liquid x-ray fixer for 3 minutes, washed and dried. The individual grains of the alpha tracks could be plainly seen when examined by an RCA EMU-2C electron microscope. The point of origin of the tracks was located by the shadow cast by the Pu^^^ particle. George and Vogt (15), using unshadowed grids, examined plutonium particles collected on millipore filters, and prepared electron micrographs of selected areas of particles before and after autoradiography, thus dif- ferentiating radioactive from non-radioac- tive particles. It is probable that small cubes of tissue of approximately 150 microns could be infil- trated in diluted gel emulsion long enough to penetrate the tissue. These small cubes could be exposed for the desired amount of time, developed by the Temperature Devel- opment Method, after which the cubes could be embedded in methacrylate and sec- tioned by ultra-thin sectioning methods. Selected areas of cells could be studied in relation to the developed grain deposition either by electron microscopy or by oil im- mersion phase microscopy, if mounted on a glass slide (16). This is an avenue that will lead the microscopist into cellular and sub- cellular autoradiography. REFERENCES 1. Johnston, M. E., "A bibliography of bio- logical applications of autoradiography, 1954 through 1957", UCRL-8400, 1958. Johnston, M. E., "A bibliography of biologi- cal applications of autoradiography, 1958 through 1959", UCRL-8901, 1959. 10 SHADOW AUTOR ADIOG K A PHY 2. Boyd, G. A., "Autoradiography in biology and medicine", Academic Press, New York, 1955. 3. DocKUM, N. L., Coleman, E. J., and Vogt, G. S., "Detection of plutonium contami- nation in humans by the autoradiographic method". Stain Technology, 33(3), 137-142 (1958). 4. DocKUM, N. L., AND Case, A. C, "Autoradio- graphic analysis of plutonium deposition in human skin", Acta Radiologica, 50, 559-64 (1958). 5. Marks, S., Dockum, N. L., and Bustad, L. K., "Histopathology of th3'roid gland of sheep in prolonged administration of 1"^", Am. J. Path., 33, 219-250 (1957). 6. Dockum, N. L., and Healy, J. W., "Spot diameter method of quantitative auto- radiography of ruthenium^o^ particles in lung tissue". Stain Technology, 32(5), 209- 213 (1957). 7. Unpublished data, W. J. Bair. 8. Vogt, G. and Kawin, B., "Localization of radioelements in rat ovary". Document HW-53500, p. 120-123 (Unclassified), 1957. 9. GuiDOTTi, G. AND Levi Setti, R., "Auto- radiography of tracks from beta particle emitters in tissues". Stain Technology, 31, 57-65 (1956). 10. Unpublished data, N. L. Dockum. 11. Hungate, F. p., Uhler, R. L., Cline, J. F. AND Stewart, J. D., "Decontamination of plants exposed to a simulated reactor burn", Document HW-63173 (Unclassified), 1959. 12. Jee, Webster S. S., Arnold, J. S., Mical, R., Lowe, M., Bird, B. and Twente, J. A. "The sequence of histopathologic bone changes in bones containing plutonium", p. 148-189. Univ. of Utah Radiobiology Lab- oratory Annual Progress Report, COO-218, 1959. 13. O'Brien, R. T., and George, L. A. II, "Prep- aration of autoradiograms for electron mi- croscopy". Nature, 183, 1461-1462 (1959). 14. Unpublished data, Borasky, R. and Dockum, N. L. 15. George, L. A. II, and Vogt, G. S., "Electron microscopy of autoradiographed radioactive particles", Nature, 184, 1474-1475 (1959). 16. Dockum, N. L., Vogt, G. S., and Coleman, E. J., "Applications of autoradiography in biological research", J. Biol. Phot. A,s.soc.,27, 1-18 (1959). Norman L. Dockum SHADOW AUTORADIOGRAPHY Shadow autoradiography is a technique that enables one to differentiate between radioactive and non-radioactive particles. The source of the particles may be suspen- sions or aerosols. The technique is described below. Droplets of particle suspensions are placed on clean glass microscope slides, spread and allowed to dry. Aerosol particles may be collected directly on the glass substrate by gravity or by impaction methods. If the par- ticles from aerosols are collected on mem- brane filters, the membrane filter is dis- solved in a suitable volume of acetone and (a) (b) Fig. 1. Alpha emitters. Photomicrographs of chrome-shadowed particles before (a) and after (b) autoradiography. Alpha-emitting radioactive particles are those surrounded by star clusters of alpha tracks. (Courtesy N. L. Dockum). 11 AUTOR ADIOGKAPIIY (a) [b) Fig. 2. Beta emitters. Photomicrographs of chrome-shadowed particles before (a) and after (b) autoradiography. Beta-emitting particles are covered by spots of dense granules. (Courtesy of N. L. Dockum and R. Borasky, Nucleonics, 15, 110 1957). W ■'^ % ^ ££^ '♦•- i "* K Wl^ ■■* • ' ^ ^ . .^^•%' >5* (b) Fig. 3. Photomicrographs of superimposed negatives of areas before and after autoradiography. (a) Alpha-emitting particles, (b) Beta-emitting particles. 12 SHADOW ALTORADIOGRAPIIY aliquot s of the acetone suspension treated in the same manner as described above. The glass slide containing the particles is shadowed with chromium at an acute angle (e.g. 30°) (1). The shadowed slide is next examined in the optical microscope and areas of interest are recorded by stage mi- crometer readings and by photomicrography. Following the microscopic examination, a section of 5 micron nuclear track emulsion stripping film (e.g. NTA for alpha-emitting particles and NTB for beta-emitting par- ticles) is floated over the sample and allowed to dry. The latter step is performed in a dark room using a series I Safelight filter. The slide bearing the particles and emulsion is placed in a light-tight box with a desiccant for a suitable standardized time for exposure at 3°C. The exposed slides are developed in D-19 developer for 5 minutes, washed, fixed, dehydrated and cleared, and a coverslip mounted over the emulsion. The slide is re- examined in the microscope and previously selected areas are photographed. Alpha activity is manifested by star clus- ters of alpha tracks (cf. Fig. la and lb). Beta activity is manifested by spots of dense granules (cf. Fig. 2a and 2b). Beta activity may be determined quantitatively by spot- diameter autoradiography (2). The method for differentiating radioactive from non-radioactive particles is as follows. A sheet of tracing paper is placed over the photomicrographs or negatives of selected areas of the shadowed slide before auto- radiography. Particle positions are noted with small circles. The tracing is next placed over the photomicrograph or negatives of the slide after autoradiography and activity centers noted with a check mark or cross. The tracing now has small circles represent- ing non-radioactive particles and crosses or check marks denoting radioactive particles. Siriking results are obtained by superimpos- ing the negatives of the areas before and after autoradiography (cf. Figure 3). The physical characteristics of the particles are determined from the photomicrographs of selected areas on the shadowed slide before autoradiography. REFERENCES 1. Dempster, W. T., and Williams, R. C, Anat. Record, 96, 27 (1946). 2. DocKTJM, N. L. AND Healy, T. W., Stain Tech- nologij, 32, 209 (1957). 3. DocKUM, N. L. and Borasky, R., Nucleonics, 15, 110 (1957). R. Borasky Chemical microscopy ALKALOIDS AND ALKALOIDAL-TYPE PRECIPI- TATION An attempt is made in the article on Chemical Microcrystal Identifications to keep the discussion of practical work on a sufficiently general basis so that any analyti- cal chemist concerned with identification may be able to see some application to his own work. If a consistent effort is made to use and develop this kind of chemical iden- tification, it is usually best to work from tests that are already known and satisfac- tory in a restricted field, and to extend the coverage first to chemically related com- pounds, having the same or modified re- agents, and by stages, systematically try to cover entirely different groups of compounds with other reagents and new crystal-produc- ing reactions. 13 CHEMICAL MICROSCOPY The traditional tests on the well-known alkaloids are made on aqueous solutions. About 80 substances, mostly natural alka- loids, or of natural origin although syntheti- cally modified, have been studied repeatedly by different investigators for their reactions in such tests. Table 1 summarizes the data for 16 of these with the 11 reagents which give most of their major aqueous micro- crystal tests. These reagents are equally applicable to numerous other alkaloids. It is luifortunate that publication of the isolation of a new plant alkaloid is almost never accompanied by some individual or highly characteristic analytical test by which subsequent investi- Table 1. Some Important Microcrystal Tests FOR ^ ^LKA LOIDS i) ■3 o to K- 1 s 1— i u u i-i PQ d < c o o K u < c 5 en u r2 < u u 8 Cli U W ■a o fa other Caffeine (c) C c C — — — HAuBr4 Ephedrine a C — a (ac) a — C — — — HaPtIs rgt Nicotine c c ac c C c C c c — — HoPtBre Aconitine ac ac a a a a a — a C ■ — HMn04 ; HCIO4 Scopolamine c C a C C c (a) (ac) ac — C I-KI, 1:35 Hyoscyamine C C a c C c c a c (ac) c I-KI, 1:35; Br-HBr Atropine C c a c ac C (c) (a) c ac c I-KI, 1:50; Br-HBr Morphine C(l) C C(2) ac a a (c) (a) a(c) c — HCl-HAuBr4 (HCl soln) Codeine C c(a) C a a a c a a(c) (e) — Hgl2-NaCN & Nal (3) Diacetyl-mor- a ac ac ac a(c) ac C C(4) c C a HAuBr4 in (2 + 3) phine (Her- H2SO4 oin) Meperidine a ac a ac ac c ac C C a c Pbl2 in K acetate soln Procaine ac a(c) a C a C c(a) c c(a) (a) (c) H.PtBre ; Br-HBr Cocaine a a a C C C a C(5) c c c HaPtBre ; Pblo in K acetate soln Quinine a ac a ac a(c) c ac a a a(c) C Herapathite test; Na2HP04 ; HgCU & HCl (6) Strychnine c a(c) C C C C C c C c C CrOs ; numerous others Narcotine a a a a a a a a a C a K2Cr04 ; K acetate soln a, amorphous precipitate, not crystallizing c, crystals form, either directly or from an amorphous precipitate a(c), amorphous precipitate, crystallization poor or uncertain c(a), normally crj^stals but precipitate may remain amorphous ac, discrepant reports, usually because of variations in reagents or because crystallization, while fairly certain, is quite slow (a), (c), or (ac), amorphous precipitate or crystals may not form in most tests because of lack of sensitivity — , no result C, major crystal tests 1-6, Figures. 14 ALKALOIDS AND ALKALOIDAL-TYPE PRECIPITATION gators might distinguish it. When, in some study, alkaloids are isolated from plant ma- terial on a small scale, and are not among those commercially available, one usually finds that, even though the plarit from which they came is known, and the alkaloids found are distinguished by analytical reactions, it is still impossible to tell which one if any corresponds to some previous^ studied al- kaloid of that plant, without going through most of the original procedure. This necessi- tates a large quantity of material, obtaining Fig. 1. A well-known crystal test: morphine with iodine-KI (aqueous solutions). ""% ''^. "t fh ^* % ^^' Fig. .3. A less-known alkaloidal crystal test: codeine with Hglo-NaCN and Nal (aqueous solu- tions). Fig. 4. A well-known crystal test: diacetylmor- phine (heroin) with H2PtCl6 (aqueous solutions). : ■ >ti I Fig. 2. A well-known crystal test: morphine with K2Cdl4 (aqueous solutions) . melting-points, and analyzing for elementary composition, except in the rare event that a sample of the original isolate can be obtained for comparison. This kind of gap between research and application is very common. Other Compounds with Aqueous Re- agents. Alkaloids do not constitute a dis- tinct chemical group. Their precipitation with the "general alkaloidal reagents" de- pends upon their being compounds of basic 15 CHEMICAL MICROSCOPY Fig. 5. A well-known crystal test: cocaine with H2PtCl6 (aqueous solutions). Y^^--M^ "W Fig. 6. A less-known alkaloidal crystal test: (2) quinine with mercuric chloride and HCl (aque- ous solutions). nitrogen, usually of a fairly high degree of complexity. There are now numerous drugs that are not considered alkaloids, but which have the same characteristics: they are nitro- gen-bases, soluble in dilute acids, precipi- tated by exactly the same general reagents, and give characteristic crystals with certain of them (different ones, depending on the compound), just as the alkaloids do. Among these are the synthetic narcotics, local anes- thetics, antihistamines, antimalarials, atro- pine-like drugs and some kinds of tranqui- lizers. In some cases, whole classes of these drugs are recent introductions, and for all of them, the number in use has expanded greatly in recent years. E. G. C. Clarke has given microcrystal and color tests for 101 alkaloids (including those synthetically modified), and for 15G of these other drugs (to the date of writing), with the same reagents. Other classes of drugs which are amine bases are the sympathomimetics and central stimulants. Some of these also give good tests with the traditional aqueous reagents. Others, relatively simple in structure and often having hydroxyl groups, may be too water-soluble in their compounds for good tests in this way, and precipitation in phos- phoric-acid solution is used. Many different acidic and anionic re- agents are available for basic compounds and it would seem that it should be equally possible to use basic reagents or cations to provide microcrystal tests for acidic com- pounds, at least those of some complexity. As a matter of fact, tests have been pub- lished for various acids using salts of silver, lead, copper, nickel, mercury, zinc, and pal- ladium, usually in a solution made slightly basic with ammonia, pyridine, triethanol- amine, or the like. Such tests are useful but as yet most of them hardly seem in the same class with the better alkaloidal tests, either for proving identity or in sensitivity, and much study to develop good tests is needed. Aqueous Precipitation of the Free Substance. Many alkaloids are precipi- tated from acid solution by making the solution basic, or even, in the case of weak bases, by merely reducing the acid strength, for example with potassium acetate. Charac- teristic crystals are often produced and the tests also give chemical information merely by the fact of precipitation. The strength or weakness of the base is shown to some degree by the strength of the basic reagent required for sensitive precipitation. K2Cr04 may pre- 16 ALKALOIDS AND ALKALOIDAL-TYPE PRF.CIPITA TION cipitate the chromate of a strong base but its really sensitive tests are those in which it acts as a basic group reagent for weak com- plex alkaloids. Substances soluble in both strong aqueous acids and bases may be insoluble at some intermediate point. The method is obviously general and acidic substances may be pre- cipitated from alkaline solution by ordinary acids. This is one way of obtaining crystal tests for barbitm-ates. While using various ways of separating the free substance for microcrystal tests, the chemical meaning of the reactions should always be noted. Tests are also made on an acidic substance (which may be in fine particles, or amor- phous, to start with) by dissolving it in NaOH solution, adding HCl in slight excess, and allowing the drop to evaporate to dry- ness. The reagents leave merely grains of NaCl, which are isotropic, and therefore in- visible with crossed nicols; the acidic sub- stance set free may crystallize, and may often be birefringent and easily seen, and then its refractive indices and other optical properties can be determined. Halogen Reagents for Nitrogen Com- pounds in Aqueous Solutions. Bromine in water, or in HBr or NaBr solution, and iodine in KI or HI solution, are the reagents commonly used; iodine may also be used in HBr or NaBr solution. These are very gen- eral precipitants not only for the alkaloids but in general for amine compounds of any complexity. In a somewhat different type of reaction their use extends even to compounds in which the nitrogen no longer has appre- ciable basic properties. The alkaloid-type reaction takes place in acid solution, or at least the alkaloid is com- bined with acid in a salt and itself acts as a cation. When an amine-derivative is pre- dominantly acidic and is precipitated as an iodine compound in acid solution, it may not be clear whether the reaction should be attributed to residual basic qualities, or not. The related but anionic reaction is most clearly seen as a distinct type in neutral or slightly basic solution; occurs best with a high concentration of iodine. This type of reaction, which has not been explored nearly enough occurs with some barbiturates and with other compounds without appreciable basic properties, and also with some com- pounds that can react either as bases or acids, including the alkaloids caffeine, theo- bromine, theophylline, and colchicine, which give both the alkaloidal-type and the ani- onic-type of iodine precipitation. They all have acidic C=0 groups, / Oxygen Acids. These may be divided into two kinds, complex and simple. The complex oxygen acids, phosphoritungstic, phosphorimolybdic, arsenimolybdic, sihco- tungstic, etc., are very general, precipitating from aqueous solution all the alkaloids and related compounds, and, generally speaking, potassium and the heavier alkali metals, ammonium and the simple amines, as w^ell. As a rule they give crystals only with the simpler bases; most of their alkaloidal pre- cipitates are amorphous. Phosphorimolybdic acid is, however, very useful for comparative precipitation. Instead of the absolute con- centration of reactive substance, sensitivi- ties may be reported in terms of the relative sensitivity compared with phosphorimolyb- dic acid. This is also important in classifying alkaloids by precipitation. The simpler oxygen acids used as alkaloi- dal precipitants are perchloric, chromic, permanganic, etc., used either as free acid or alkali salt. Permanganate especially fre- quently reacts as an oxidizing reagent and is itself reduced, but it gives some excellent crystals with compounds not readily oxi- dized. Unlike the complex oxygen acids these compounds are not very general or sensitive as reagents, and so are chiefly use- ful with fairly complex bases which are easily precipitated. Chlorochromic acid, HCrOsCl, is more sensitive and general than CrOs , 17 CHEMICAL MICROSCOPY but still gives similar precipitation and crj^s- tals. Anionic Complexes of Central INIetals. These reagents are commonly called "gold chloride", "platinum chloride", etc., and their precipitates are called "double salts", but actually the metals are in anionic com- plexes and the precipitating agents are really chlorauric acid, HAuCU , chloroplatinic acid, H2PtCl6 , etc. Thus the reagents are given by those metals a7id anions which will form anionic complexes of this particular type, and the group might be subdivided either according to the metals or the simple anions concerned. The metals are central to the long periods of the periodic table; the anions are halogens (Cl~, Br~, I~) and "pseudohalogens", such as CN~, SCN~, NO2-, N3-, etc. The iodide reagents of bismuth and plati- num are colored and so sensitive and general in aqueous solution that they are commonly used for spraying to bring out the spots of alkaloids or related compounds in paper chromatography . The following table gives the reagents of greatest value for microcrystals, designated as R; other definite reagents within the scope of this table are indicated by a small r. Chloride Bromide Iodide Cyanide Gold (+3) R R — R Platinum (+4) R R R R Palladium (+2) R r Mercury (+2) R r R r Cadmium (+2) r r R Bismuth (+3) r r R A "gold iodide" reagent is in use, made from HI and chlorauric acid, but the HAUI4 immediately decomposes, and the chief effects seem to be due to iodine-HI, rather than any gold compound, with a few results due to an aurous iodide complex, a gold (-f-1) reagent. The following might also be mentioned: chlorides (chloro-acids) of Fe, Zn, Sn (stan- nous and stannic), UO2 (uranyl), and VO (vanadyl), which are most effective with a high content of HCl to form the chloro-acid ; iodides of Pb, Sn (+2), Zn, and Ag, with alkali iodide; the complex cyanides of Fe, both ferro- and ferri-, and also nitroprusside; and complex thiocyanates of Ft, Hg, Sn (-f2), Cd, Zn, Co (4-2), Ni (+2), Mn (+2), and Cr (-F3). Most precipitates with thiocyanate reagents do not crystallize very readily. A remarkable exception is reinecke salt, NH4Cr(NH3)2(SCN)4 , which is very general and yet gives many crystals. Double complexes are also possible; e.g., mercuric cyani-iodide or mercuric chloro- iodide, made by dissolving the insoluble Hgl2 in cyanid-? solution or in strong HCl, respectively. On the whole this group is the most useful of all fo^^ niicrocr^^stal tests, and by using other strong acid; besides HCl the use of these reagents may be extended over all compounds of basic nitrogen. Simple Halides. These are useful with relatively complex bases, particularly iodide and thiocyanate, used as alkali salts. Organic Reagents. The best reagent of this group is picric acid, and several others are highly nitrated compounds, e.g., styph- nic and trinitrobenzoic acids. A few other kinds are known, e.g., sodium alizarin sul- fonate. Sodium tetraphenyl-boron is a new type of reagent (in some ways alhed to the complex oxygen acids), recently introduced, which is very useful for crystalline precipi- tates with free amines in volatility tests (without acidification), and quite sensitive to the lower amines in this way. Reagents in Strong Acids. The study of strong acids as media for the precipitation of nitrogen-bases began with hydrochloric acid, which, because many of the precipitat- ing agents are halides, often has special effects. Some reagents were already known which require a high content of HCl either to prevent precipitation of basic chloride (e.g., SbCls) or to form the chloro-acid 18 ALKALOIDS AND ALKALOIDAL-TYPE PRECIPITATION (e.g., FeCls). These were more or less assimi- soluble in H3PO4 tliaii in water, and far lated to the ordinary aqueous reagents; but more soluble in acetic acid than in water, some 20 or 25 % of concentrated HCl in the The reagents with acetic acid are therefore reagent will profoundly affect the precipi- less general than others, and find their espe- tation and crystal-forming 'properties of cial application with highly complex com- HAuCls , for example. Next, reagents made pounds, very easy to precipitate, particularly with concentrated HCl were tried; then in- those that yield only amorphous precipitates stead of using only aqueous solutions of the in the conventional tests. Instead of having substance tested, concentrated HCl was also to study new precipitating agents which are used for this purpose The value of other less and less general in effect, we can simply acids was later discovered. It became evi- try using acetic acid in the test-drop, as a dent that instead of dissolving the substance medium for new, useful tests. tested in quite a number of different acids, The effect of phosphoric acid in increasing it would be simpler to apply the reagents, the range of the tests is even more impor- in the various acids, directly to a Httle tant. There are in fact three effects. One is of the solid substance. This is now the the effect of a nonvolatile liquid medium in method used, but precipitati^on from solution allowing a test to stand as long as may be could be used. The precipitating agents are desired for crystallization to occur. H3PO4 the same ones that have long been used in has no very strong tendency to absorb water, aqueous solution to precipitate alkaloids. nor, when it is used already a little diluted, The acids proved to have very different to dry out. It has no side effects correspond- effects. The ones chiefly used are diluted or ing to sulfonation, or to the withdrawal of syrupy (85-88 %) H3PO4 , diluted H2SO4 , water from the molecule of a dissolved sub- concentrated HCl, and (2 + 1) acetic acid, stance, both often given by H2SO4 even when Concentrated H2SO4 would react with too it is somewhat diluted. Another effect is many of the substances to be tested, and that of any strong mineral acid in suppress- would decompose nearly all precipitating ing acidic qualities of amphoteric substances agents, but may be used up to (1 + 1) and enhancing their basic qualities Third strength with chloride reagents, as strong as and especially important is the particular (2 + 3) with bromides, and up to (1 + 3) effect of phosphoric acid in increasing the with iodides. Acetic acid may be used up to insolubility of precipitates in it, as already the glacial strength except that it then tends mentioned. to creep and spread all over the slide, and The most useful of all the crystal-produc- for this reason (2 + 1) is usually the maxi- i"g compounds are the chloro- and bromo- mum strength employed. ^^^^^ «^ ^'^^^ ^^^.^ platmum, the iodide re- ^-r -J. ■ r J ^T_ i XT- • -^ i- agents of platinum and bismuth, and Now it IS found that the precipitation- . ^,. ^,^ ^ . ,. ^^^ , ,' , „» , . , ,,!•«. .- 1-, . 1 TT o/-i lodme-KI or lodme-Hl. Among other ad- eft ect is not greatly different in diluted H 2^04 , ,, 11 , 1 J ^1 , ^^_, ^ .... vantages these are all colored, and crvstals or m concentrated HCl from that in plain ^ ji.i u j-ij-^- "^-uj ^ formed by them can be readily distinguished water, only a little greater m diluted H2SO4 ^^.^^^ ^^.^.^^^^ crystalline material, or from and a little less m concentrated HCl; and ^lystals, e.g., phosphates, formed simply by these acids are chiefly used, as media, to ob- ^^le acid used. They are compatible, with tain certain crystallization effects not ob- proper formulas, with all the acids men- tainable in plain water. The most surprising tioned, and extend the use of microcrystal effects are obtained with phosphoric and tests for identification to all compounds of acetic acids. The compounds formed by the basic nitrogen. Phosphoric acid reagents in precipitants with bases are immensely less particular are used for relatively simple, and 19 CHEMICAL MICROSCOPY feebl}^ basic and partly acidic compounds, and those of high solubiHty in water. Inorganic Precipitation with Re- agents for Basic Nitrogen. Potassium and ammonium have similar solubilities for their salts, and so give similar precipitation reac- tions; and, in a more general view, the al- kaloids and other compounds of basic ni- trogen react much like the heavier alkali metals. Cesium (ion) is precipitated by most of the "alkaloidal" reagents and rubidium > \ N^sr-^ n Fig. 7. Sodium bromaurate crystals. Na2S0 (solid) with HAuBr4 in H3PO4 . Fig. 8. Magnesium bromauraurate crystals MgCl.-eHsO (solid) with HAuBr4 in H3PO4. (Pho- tographed with red-sensitive plate.) by perhaps a third of them, from aqueous solution. The most general of such reagents often precipitate ammonium and potassium; and conversely, a reagent used to precipitate potassium is worth trying as a general pre- cipitant of nitrogenous bases. Sodium ion is not precipitated from aque- ous solution by the "alkaloidal" reagents, but in a medium of syrupy phosphoric acid the relationship of the lighter elements be- comes apparent. Not only sodium, (Fig. 7) but even lithium, magnesium (Fig. 8), zinc, and cadmium, become more or less subject to such precipitation. Even the hydrogen compound verges on insolubility, and for example bromauric acid itself, HAuBr4 , will precipitate from H3PO4 solution with drying. These inorganic reactions have been studied very little, except the formation of bromau- rates. Charles C. Fulton CHEMICAL MICROCRYSTAL IDENTIFICATIONS The microscope should be used throughout chemistry; it is surely an obvious instrument (and not a new one) simply for taking a closer and better look at small things. Chem- ists should turn to it as naturally as to test- tubes or the bunsen burner, whenever it will help. It has applications requiring knowledge and study, too, particularly the polarizing microscope, which is by far the most valu- able. Its especial use, discussed here, is in a basic branch of analytical chemistry— the making of identifications. This is a subsid- iary science in its own right, not well ex- plained by the vague general term "qualita- tive analysis". It is, of course, a part of qualitative analysis, as distinguished from quantitative, but it is a special part, in which we ought to use properties, tests, and reactions especially characteristic, or even specific for individual substances. Physical properties, although obtained by some gen- eral method, are often used when they can be accurately measured in such a way as to 20 CHEMICAL MICROCRYSTAL IDENTIFICATIONS distinguish an individual compound even The science of making chemical identifica- from those others that are closely related, tions is important in forensic chemistry (cf. Likewise, a chemical reaction may be general Forensic Microscopy, p. 338), especially in law for a whole class of compounds, and yet we enforcement, because many legal cases in- may be able to distinguish particular results, volve questions of identity rather than For example, a color reaction of phenols may quantity, as regards narcotics, poisons, po- give different colors with different phenols, tent drugs, adulterants, contaminants, sub- The value of the microscope is to see dif- stitutes, and so on, and such cases require ferent results given by different substances exacting proof of the identity of the sub- even in the same general precipitation reac- stances concerned. The uses are more gen- tion. eral, however, for questions of chemical Chemical tests are mainly divided into identity which can be answered in similar two kinds — color tests and precipitation ways may arise in any field involving chemi- tests. Now the ordinary spectrophotometer cal substances. Usually too little attention (cf. Absorption Spectroscopy — ^Visible and is paid to the possibilities of microscopic Ultraviolet, in "Encyclopedia of Spectros- chemistry, which may provide tests for very copy") may be regarded as giving not only a minute quantities, or such firm assurance of means of studying what might be called the identity as to be the preferred method, when "colorimetric" properties of a substance it- it can be used by anyone acquainted with self, but also, when a reagent is used, an it, regardless of the amount of substance enormous extension of the color tests of available for tests. chemistry. Not only are comparisons and Microcrystal tests are exceptionally good colorimetric readings made better, in the for court purposes, because they are as sim- visible range, but also they are extended pie and direct as tests can be. In many cases, deep into the ultraviolet, where nothing just looking at the crystals under the micro- at all could be seen with the unaided eye. scope enables the analyst to make the identi- In a somewhat analogous way, the micro- fication, and the things looked for are the scope enormously extends not merely the characteristics of an actual, visible com- study of the form and crystal properties of pound of the substance to be identified. It a substance itself, but also the ordinary seems to be pretty generally realized that precipitation tests of chemistry, without simple, direct tests are the best for forensic changing their essential character as chemi- purposes. What sometimes seems to be over- cal reactions. Not only are details of the looked or not realized as it should be, is that precipitate, when it is crystalline, seen bet- they are also best for chemical purposes, ter, and down to a smaller size, but with the In these microcrystal tests the polarizing polarizing microscope the observation of microscope should by all means be used, and crystal characteristics is extended to things the relation to optical crystallography is, of of an entirely different nature from those course, close. Now optical crystallography, that can be seen with the unaided eye by for which, in the field of mineralogy, the ordinary light. Thus microscopic identifica- polarizing microscope was developed, is of- tions by crystal forms and properties are at ten an ideal means of identifjdng the crystal least potentially capable of extension from species, for any substance that is already substances which are already crystalline to present in microscopically sizable crystals all substances that give precipitation reac- or fragments of crystals, and in fair pro- tions in ordinary chemistry, or indeed to portion in the material examined; and de- those that give any kind of crystal-forming termination of refractive indices plays a reaction. large part. On the other hand, microscopic 21 CHEMICAL iMICROSCOPY chemical identifications, which are the sub- ject here, usually depend upon the forma- tion of crystals by the action of a chemical reagent, which may even pick out the sub- stance in a complex mixture for the analyst. Refractive indices are very seldom used, but optical-cry stallographic properties which can be observed directly, without separating the crystals from the solution in which they are formed, are used, and therefore the polarizing microscope is needed. The two procedures go well together, but they are distinct. Figure 1 illustrates crystalline de- posits examined between crossed nicols; the compound is aminotriazole. Figure lb repre- sents only 0.4 microgram in the deposit. Other types of crystallization of a sub- stance without a reagent can be seen in the photographs of r/-, (//-, and (d + dl)-airi- phetamine hydrochloride and of NH4CI (see pages ()7-69). It should be noted that optical crystallog- raphy for substances already crystalline focuses upon the exact crystal species pres- ent, determined by such things as the acid with which a basic substance is combined, and even the proportion of molecules of water of crystallization, and sometimes the particular one of two or several polymorphic forms of the same substance, for these things often make major differences in the refrac- tive indices and all the other crystal proper- FiG. 1. Aminotriazole as seen with the polarizing microscope, between crossed nicols: (a) deposit from evaporated aqueous solution; (b) deposit of 0.4 microgram. 22 CHEMICAL MICKOCRYSTAL IDENTIFICATIOxNS ties. In the microscopic chemical tests, on the other hand, one can test directly for a basic drug (e.g., morphine, or amphetamine) without necessarily being concerned with whether it is originally present as the sulfate or hydrochloride or in some other combina- tion, and similarly for other kinds of sub- stances. Indeed, when a substance must be sepa- rated from a complex mixture, whether by extraction, chromatography, or some other means, its original state of combination is lost, and to use optical crystallography, or for that matter, to get a melting-point or x-ray diffraction pattern, it must be obtained in some kind of crystalline form. The great- est sensitivity, simplicity, and directness for any test requiring crystals exist if the crys- tals are obtained as some very insoluble com- pound which still crj^stallizes quite readily and in forms that are recognizable by direct inspection and observation. This is precisely the idea of the chemical microcrystal tests. This subject is, therefore, a branch of chemistry. It involves the use of the micro- scope in making identifications by means of chemical reactions, especially precipitation reactions, yielding crystals. The optical crystallography used is limited and usually need not be of a specialized kind. In fact the chemistry used is not very unusual either, but naturally the reagents and reactions are selected as those especially useful in con- junction with the microscope. The science is to some extent a blend of chemistry and microscopy, but analysts should particularly keep in mind the chemical meaning of the tests, and development of the field will pro- ceed along chemical lines. It is strange indeed that it has been so generally neglected by chemists, over a period of a hundred years. Advantages and Disadvantages. The outstanding advantages of this kind of iden- tification tests are: (1) The high assurance with which micro- crystal identifications can be made. (2) The direct character of most of these tests. (3) Their simplicity, convenience, and speed. (4) Selectivity, in the sense of noninter- ference, or no great interference, by most impurities. (5) High sensitivity, not necessarily in degree of dilution, but especially, with the aid of the microscope, in respect to amount of the substance for which the test is made. In addition to various objections that are often made but have no very substantial ba^is, the main disadvantages are: (1) The results are not readily classifiable. (2) Applicability of coverage limited in organic chemistry. (3) The general neglect. When an identification is wanted for a chemical substance the identification should be specific. However, it is seldom reached by a single specific test, but usually as the only conclusion possible from two or more — com- monly several — group tests or characteristic tests and properties. Often most of the chemical tests used are rather general, and are then supplemented with some physical measurement, often made on a derivative, to show differences between closely related compounds within the ascertained group, or provide final confirmation. The combined characteristics then become specific. Much misunderstanding of the microcrys- tal tests seems to result from a failure to appreciate their chemical character. In a chemical microcrystal test we know, fij'st of all, what reagent we used; and we know, or ought to know, what its chemical effects may be. We take a close look through the micro- scope at the crystals formed. Often we can recognize them immediately by their visible characteristics, assuming we have seen them before, and always remembering that we know what reagent was used. The results are far more definite for indi- vidual compounds than in most chemical tests, l)ut the procedure of identification is 23 CHEMICAL MICROSCOPY essentially the same, and these tests can well be used along with others, especially color tests for the spot-plate or designed for mi- nute residues, or "spot tests" as developed by Feigl. However, the microcrystal tests are usually so highly characteristic, that two or three of them, even just using the ordinary microscope, and making comparisons, as necessary, with a known sample, will often make an identification certain, without nec- essarily having or obtaining any other type of information. With the polarizing microscope the results are especially definite since a number of fur- ther observations can be made, even without removing the crystals from the solution in which they form. They do not have to be separated, washed, dried, recrystallized, and so forth, as for melting-point determinations and most other types of further tests, includ- ing refractive indices. All of the following are simply a matter of direct observation: not only the shape of the crystals, their grouping, color, size, and so on, but also bire- fringence, whether high, medium, low, or absent; whether extinction is inclined, and if so, the angle of extinction; in the case of colored crystals whether dichroism is pres- ent, and if so, the two different colors shown, and if the crystals are regularly elongated, the direction of dichroism (sign of absorp- tion), and in the case of colorless elongate crystals, the sign of elongation. All these things can be observed without having to treat or prepare the crystals in any way, once they form in a solution. Microcrystal tests are especially useful for distinguishing among closely related reactive compounds. Usually suitable tests can be found even to distinguish an isomer from the corresponding racemate, by completely different crystals. In fact, such tests have been used to identify (/-amphetamine even in the presence of d/-amphet amine, and vice versa. E. G. C. Clarke has shown how micro- crystal tests can sometimes be used to dis- tinguish microgram amounts of a d-isomer from the Z-isomer, and without using an op- tically active reagent, which is another possi- bility. The question of identifying a d- or Z-isomer is important in some cases now be- cause the ^isomer of certain new synthetics is restricted as a narcotic while the d-isomer is also offered commercially as a different drug not subject to narcotic restrictions. The distinction is based on the fact that the race- mate will give crystals in certain cases where a separate isomer will not. Besides knowing and having a suitable reagent, the analyst needs one known isomer, let us say the d-. If he mixes some of this with the suspected substance and it is 1-, then he has the race- mate and can get the appropriate crystals, but if the suspected substance is d- and he mixes d- with it, then of course he still can- not get crystals due to the racemate com- pound. If he has both known isomers for a cross-check, this kind of test can be com- pletely certain. The question about microcrystal tests so often asked, "what will they do on mix- tures?" — reveals fundamental misunder- standing. Most identification tests require that a substance to be identified be fairly pure. This is perhaps especially true of most physical and physicochemical tests, because they are so very general — melting-point tests, for example. A strictly chemical test or reaction, on the other hand, is bound to be selective in some degree, and will pick out a substance in the presence of all kinds of impurities ex- cept those that react with the same reagent, and occasionally even in the presence of compounds that react, but in a different way. A microcrystal test, where not merely the fact of a reaction, but crystals of a particular kind are looked for, cannot in general be expected to work on a complex of closely related, reactive compounds, although some tests that are remarkably searching, in this sense, are known. The application is often 24 CHEMICAL MICKOCRYSTAL IDENTIFICATIONS to a residue that has been separated as satis- may be put at about .000-4 microgram of factorily as possible from a mixture. atropine. "Selective" is a term used in two different In general, however, the writer agrees with senses: a reagent may demonstrate that a Chamot and Mason: "It is perhaps unfor- compound belongs to a particular group, or tunate that so much stress has been laid it may actually pick out a reactive com- upon the sensitivity of microscopical crystal pound among others that are unreactive tests, both because their advantages of ease or that do not react as strongly or in the same and directness are thereby obscured, and way. Thus one kind of identity test may because the impression is given that they require an absolutely pure substance and be are appropriate only in case extremely mi- specific, or characteristic, or merely selective nute amounts of material or low concentra- (in the first sense); while another type of test tions are to be dealt with." Sensitivities are may not require great pmity, and therefore important — -sometimes very important, even when specific or characteristic for a com- absolutely essential in such a science as pound may also be quite selective in the sec- toxicology — -but it does seem that when a ond sense, as well. The latter is very true of microcrystal test is described in any detail, the chemical microciystal tests. sensitivity is likely to be the one point These tests will work on mixtures with seized on, in abstracts for example, to the nonreactive material, with little or no inter- exclusion of everything else. This degrades ference, and can generally be applied directly the microscopical crystal tests to the level to medicinal tablets, for example, in which of mere detection tests, which are often sen- a drug is mixed or diluted with material that sitive but very general, and worthless for is inert in the chemical reaction concerned, identification whereas the purpose of the The reagent itself seeks out the particles or microcrystal tests is identification, whatever molecules of the substance that interests us, the amount of substance available. Their ad- among particles and molecules of all sorts vantages are not only in ease and directness, of other (nonreactive) substances. In the as mentioned, but in providing highly char- same way, these tests may be just the kind acteristic and even specific identification needed for extracts or isolates when it is not tests, not mere detection. To the extent that possible to get extremely high purity without sensitivity statements obscure this, they risk of losing the little bit of material that should be played down; at the same time it is all that is available. should be realized that the better micro- When procedures on a very small scale are crystal tests are exceedingly sensitive, and combined with the use of the microscope, as thus quite adequate for identification on al- is done by Clarke, a quite common sensitiv- most any minute scale, ity for a highly characteristic crystal test The statement is rather frequently en- will be 2 or 3 hundredths of a microgram, countered, that in general color tests are Atropine gives tiny characteristic crystals more sensitive than precipitation (or micro- with the right variety of iodine-KI reagent crystal) tests. This is both incorrect and down to an aqueous dilution of 1 to 1,000,- fallacious. Some compounds give excellent 000. Using a fair-sized ordinary drop of 0.04 color tests and few and poor precipitation ml the crystals can therefore be obtained tests, or none; others are just the reverse, with 0.04 microgram; by using a microdrop Some compounds, e.g., morphine, are re- of 0.1 microliter (.0001 ml), the extreme markable for both kinds, and in such cases sensitivity is .0001 microgram; or, as the they are hkely to be of the same order of sen- crystals are hard to find at the extreme limit, sitii-ity, as a consultation of published state- the working sensitivity, using microdrops, ments in regard to morphine tests will show. 25 CHEMICAL MICROSCOPY The sensitivities can be great for both, but oped. They are destructive aUke to the speed, they are usually due to different reactive simplicity, ease, and directness of the micro- radicals or atoms, sometimes both kinds crystal tests, in most cases without any con- present in the same molecule, as with mor- current gain, since the observation of char- phine, and sometimes only the one or the act eristic crystals was already adequate to other present for any known and useful tests, the pmpose of identification. Much of the It would be almost as sensible for a chemist literature on the use of picric, styphnic, and to cut off one hand, as to refuse to use one or picrolonic acids, reinecke salt, etc., is of this the other kind, according to their merits in type, particular cases. A related erroneous idea sometimes en- The preeminent radical for color tests is countered is that special measures should be probably the phenolic hydroxyl. Even more taken to get the crystals as well-formed crys- preeminent for precipitation and micro- tallographically as possible — whereas in fact, crystal tests is the basic or partly basic ni- crystals ordinarily forming in characteristic trogen atom, no matter whether it is "pri- distortions have far greater value. In most mary", "secondary", or "tertiary", or cases even procedures such as warming or whether it belongs to a straight chain of adding alcohol are of value only when a atoms, or is attached to a benzene ring, or is particular compound is suspected that needs itself part of a ring. The chief count eractant special measures to induce crystallization to the precipitating effect of basic nitrogen with a much-used reagent, and then a special is acidic oxygen, which shows why color reagent is usually better if one can be found tests may occasionally "oppose" crystal that will readily give crystals withovit such tests, i.e., the change of a molecule to in- measures. elude the preeminent radical for color tests Crystals given by a particular organic may weaken and lessen the sensitivity of compound may be specific, that is, of a kind precipitation tests, especially for relatively given by no other substance. Sometime ob- simple compounds. Paradoxically, however, jection is made to calling such tests specific, oxygen itself may be basic, in certain cases, but on grounds that would prevent any tests and then forms oxonium salts, which are sub- ever being called specific: that not every ject to precipitation tests. possible substance has been tried, or even While an identification may well be based that some new compound might be discov- on several results obtained separately by ered or synthesized sometime, that might be different disciplines, it is unfortunate that confused with the old one. To the writer it numerous researchers consider the form and seems more reasonable to call a test specific obvious characteristics of crystals only as if it is specific, so far as we know or can rea- dressing for procedures resulting in micro- sonably foresee, for our knowledge never can melting points, refractive indices, or other be infinite. numerical determinations. That is, they On the other hand, we should exercise make the tests over in the usual pattern of considerable caution in designating a test as organic confirmatory tests: formation of a specific. The crystals are, in general, direct derivative, isolation of the derivative and compounds of the substances for which the often its further purification (e.g., by recrys- tests are made, and therefore in a particular tallization), and physical measurement of case they are more or less different from some characteristic of the isolated, purified those given by any other substance, just as derivative. Such procedures have been pro- each substance is, in itself, finally unique, if posed even in the alkaloidal field, where the tests could be carried far enough to take in microcrystal tests are already best devel- all its physical and chemical properties. 26 CHEMICAL MICROCRYSTAL IDENTIFICATIONS However, the question is whether the crys- tals in a particular case are sufficiently differ- ent from most others, or from all others, to be distinguished from them by informed direct observation. The writer is satisfied to call most of these crystal tests, even most of the very good ones, characteristic rather than specific; not only when we know of some definite resemblance but also whenever the crystals are some general type that might occur with some other compound, even if we do not yet know what it would be. However, as an example of a specific test the intensely pleochroic crystals of hera- pathite, as a test for quinine, may be cited. These crystals of quinine iodosulfate (Fig. 2), produced by a suitable reagent, are rec- ognizable at once and the test is so sensitive to cjuinine, and so insensitive to interference, that the writer has used it as a direct test for quinine in the diluted, adulterated, impure heroin mixtures of the illicit drug traffic. Of course, one almost never depends wholly upon any single test by itself, even if it is specific and the intense fluorescence of qui- nine sulfate solution in ultraviolet light, which is characteristic but not specific, is also easily observed. Other cinchona alkaloids give highly char- iA *^% ./«■■ ^|i' ^'•i*'^ . ^-^v^ ■y^-^^ y Fig. 2. Crystals of quinine iodosulfate (hera- pathite) (with polarized light). lodine-KI reagent C-3 or a similar reagent. Fig. 3. Crystals produced by quinidine with lodine-KI reagent C-3 or a similar reagent. wn Fig. 4. Crystals produced b.y Cinchonidine with lodine-KI reagent C-3 or a similar reagent. acteristic, wholly different, iodosulfate crys- tals, with the same reagent used for quinine (Figs 3, 4, 5). A strange objection is often made to crys- tal tests as, supposedly, a reason for not us- ing them: that one substance with one re- agent may give two or more different kinds of crystals. This undoubtedly happens, but, in the first place, it also happens with all other tests that depend on crystal form and properties, including melting points. As a difficulty it is very unlikely to be too serious 27 CHEMICAL MICROSCOPY Fig. 5. Crystals produced by Cinchonine with lodine-KI reagent C-3 or a similar reagent. with microcrystal tests, because the crystals are not some that come to the analyst al- ready formed mider unknown conditions, but are formed under test conditions, and comparisons should be made with a known sample under as nearly as possible the same conditions. Secondly, even if one crystal type was un- expectedly changed into another which was unknown to the analyst, this would pre- sumably result in a failure to make the identification, but not in an erroneous identi- fication. There is probably no way of testing known that may not sometimes fail to give a recognizable result because of some kind of interference with the usual test conditions. On the other hand, if both kinds of crystals are known, even if the reason for one chang- ing into the other is not, the test is still good. Finally, the objection is especially strange because in most cases, having two or more types of test-crystals is a positive advantage. If we can learn the reason for the change — and this is usually not very hard to do, and is known in many cases — then we have two tests instead of one, which together often provide enough evidence for identification in themselves. Or, the two kinds of crystals may occur in the same test, in which case both types together characterize the sub- stance, and the single test is all the more likely to be specific or very highly character- istic. One crystal type is occasionally known to be changed to another by temperature or by stirring. However, the most common cause is a different concentration of the sub- stance, especially relative to the precipitat- ing agent in the test -drop. Particular advan- tage is taken of this in the new tests made by addition of a strongly acid reagent to a very little of the dry substance tested. A cover-glass is applied and this prevents the tested substance from diffusing equally through the test-drop. Thus the precipita- tion and crystallization take place at differ- ent concentrations in each single test even with only a minute amount of the substance. Therefore when a substance gives two or more completely different types of crystals at different concentrations, they are nearly always obtained in each single test, so that such tests are generally highly characteristic and may comparatively often be of specific rank. For example, in one such test with an iodine-KI reagent in HCI-H3PO4 , morphine gives four kinds of crystals (Fig. 6a), de- pending upon concentration: black needles, Fig. 6a. Crystals given by extracted Morphine with Iodine-KI reagent M-2. 28 CHEMICAL MICROCRYSTAL IDENTIFICATIONS brown threads in rosettes, tiny rectangular orange-brown blades, and comparatively large brown to red square-cut and jagged birefringent plates. All four -kinds can be obtained with only a few micrograms of morphine (in a small spot) ; the black needles are sensitive, with no special effort to reduce the scale, to 0.1 microgram. The test is ex- traordinarily resistant to interference, that is, the morphine need not have a high degree of purity to yield good crystals; in fact, the black needles and red plates have been ob- tained on a little dried poppy-juice (Papaver somniferum) without first separating the morphine from the other opium alkaloids or even purifying the alkaloids as such (Fig. 6b). That specific tests are available for some substances is only one factor in deciding what tests to use in given circumstances. Even a very general reaction may be useful to prove the absence of a compound. For general identification work, reagents will be preferred which give many different charac- teristic results, rather than "tailor-made" reagents for a few specific results, particu- larly in using them before one has much idea what is present. Other important factors are sensitivity, and ease of obtaining crystals even when the substance is not quite pure. Some specific tests rank high in these ways also, but they do not necessarily go to- gether. Regardless of other factors, there are always cases where what we want, when we can have it, is a test that, when successful, will prove the identity of the substance then and there, beyond doubt. This is possible with some microcrystal tests. The very diverse nature of microcrystals is at the same time the strength and weakness of the method. The results are highly char- acteristic or specific, but they are hard to classify. Largely, this is due to the fact that the tests are not measurements; the results are not even directly numerical at all, and so they lack the automatic classification that numbers give. It is not that the tests are Fig. 6b. Crystals given by Morphine in dried juice from Popaver somniferum — ^no purification of the morphine, not even any separation of the al- kaloids — with lodine-KI reagent M-2. good for only a few particular results: plenty of good results are obtainable. The difficulty is to have a way to look them up if they are not already familiar. Possible solutions are being sought, now that the problem is becoming acute. Even if anyone wanted still to limit the scope of the tests to some 50 familiar alkaloids, it is not possible even in the drug field, for there is now a host of new drugs that react in the same way (by precipitation and crystal- forming reactions) with the same traditional reagents. Several different solutions of the classification problem are possible in prin- ciple, at least such that punched-card sorting would quickly lead one from results on an unknown back to the proper "known" if it had been previously studied and classified. However, at present, and probably for some time, the analyst has to rely largely on chemical classification, involving method of isolation, color reactions, precipitation re- actions regarded chemically, and any other clues. Chemical evidence may reduce the number of compounds that have to be con- sidered to something manageable, and then microcrystal tests will show exactly what the substance is, if "knowns" for compari- 29 CHEMICAL MICROSCOPY sons are available. This goes well beyond about this state of affairs is that it should using the tests as merely "confirmatory" of be a challenge and a stimulus to an analyst something already known or suspected, but to promote application of the microscope to even that is one of their useful functions. any problems of identification that occur in The second real disadvantage listed above his own work, for these tests is the limited field of applica- It may be pointed out that these disad- bilitj^ or lack of coverage outside a limited vantages, serious as they are in some cases, field. It has already been pointed out that in no way interfere with the use of micro- the applicability is in fact not nearly so lim- crystal tests by an analyst acquainted with ited as generally supposed, but the disad- them, in: (a) confirmatory tests, where ten- vantage might be stated more accurately tative identification has been made, and that tests have as yet been well worked out comparison with a known sample is possible; only for inorganic ions and the familiar al- (b) in proving the presence of a particular kaloids. The same reagents and procedures compovmd (here, if the compound sought as for alkaloids can be used for a great many turns out not to be present, the test may do other drugs, e.g., antihistamines, and with more than most to show what is present); some changes they can be extended over all (c) in deciding, among a few alternatives, to derivatives of basic nitrogen, and even fur- which one an unknown corresponds: this is ther, but certainly other types of organic often possible by microscopic test-compari- precipitation have not been studied nearly sons with the known alternatives even if enough, and in fact with the many new drugs nothing has been previously published or the coverage is inadequate even for those studied on these particular compounds; (d) giving tests with the traditional reagents. in distinguishing between closely related Rather than any danger of making erro- compounds (including distinction of an neous identifications, as seems to be feared isomer from the racemate, or even of the by many (perhaps rightly in the case of the d- from the ^form, as previously explained); totally inexperienced), the bane of the ex- (e) in a larger and larger field of tests as the perienced analyst is the finding of an occa- experience of the analyst grows, sional compound, in the line of work, which yields beautiful crystals, sometimes with practical norK several different reagents, but which, never- As textbooks are nearly non-existent, theless, he still cannot identify. This is due some advice on practical work will be at- much more to lack of study of the com- tempted. In the limited space it still must pounds than to the difficulty of looking up be rather general. those already studied, because of lack of First, accustom oneself to the micro- classification. The individual chemist or a scope by using it. Much analytical work will single laboratory cannot possibly keep abreast be helped enormously by preliminary micro- of developments. scopic observation of the material to be Both of the disadvantages just discussed examined, either by reflected light (if it is are related to general neglect, which also opaque) or by transmitted light. This may, has other featm'es. The analyst must often for example, show at the outset whether the train himself; he usually cannot obtain material is a mixture or appears to be all one much instruction, or even textbooks, in this substance. field. At the present time the writer does Then, bring the chemical work down to not even know of a good textbook for micro- the level of convenience for microscopic ob- crystal tests for alkaloids that is now in servation. A precipitation test can be ob- print. The only good thing that can be said served just as well with a drop of solution 30 CHEMICAL MICKOCKYSTAL IDENTIFICATIONS on a slide as with several ml in a test-tube, before using the microscope at all. Color tests on the spot-plate are a useful adjunct. Accustom oneself to working on some such scale, which will probably be found more convenient than the "usual" one, anyway. Keep a polarizing microscope at hand, and use it. If a polarizing microscope cannot be used, polarizing attachments for the ordi- nary microscope are helpful, but are at best a poor makeshift. Observation of the micro- crystal tests is almost always by transmitted light, although it would be possible to use reflected light in many cases. "Low Power" of 80-100 X is usually sufficient for the tests, at least for all the initial observations. If the crystals are very small, a higher power is then used. Make a practice of looking at the result of a precipitation test microscopically, whether the textbook mentions crystals or not. For example, the iodoform test for ethyl alcohol is made quite definite for the product iodoform (although in any case it is not specific for ethanol) by microscopic observa- tion, although textbooks often fail to men- tion this and may only refer to the odor. Pictures that can be consulted are helpful in dealing with a test not yet familiar, and when time permits or the case is important enough the analyst should take them of his own results, as a reminder to himself and information to others. There is a real art in taking photomicro- graphs of known crystals so that they will most plainly show the truly characteristic forms that should be looked for. Pictures for reference, however, valuable as they are, are no substitute for at least a little actual ex- perience, and no substitute, either, for actual comparison of the results given by an un- known with those given by a known com- pound. Most of the familiar tests, for organic as well as inorganic substances, are made with aqueous solutions. Usually the ordinary flat microscope slide is quite satisfactory. Com- monly, a cover-glass is not used, unless or until examination by high power is wanted. The test is observed while evaporation takes place. This time can be prolonged by setting a petri-dish over the slide when it is not under immediate examination. Cavity slides are less convenient for focusing, but the evaporation is more gradual, and can be stopped by simply laying a slide over the one on which the test is made. Cavity slides are also more convenient for mixing three or four ordinary drops, in more complex tests. The reagent drop may be added directly to the drop tested, or placed beside it on a plain slide and the two drops allowed to flow together. Reagent is usually added in about the same size drop as the solution tested. The actual concentration of substance tested is therefore reduced by about half in the test-drop, but it is customary to report the sensitivity or optimum concentration, etc., in terms of the solution tested, before it is mixed with reagent. A full drop of water (let fall from a fairly wide opening) is about 0.05 ml. Use smaller rather than larger "drops", by touching the pipet to the slide and letting just a little flow out, by dropping from a narrow orifice, or by using a rod to transfer a drop. In most tests with no especial need to conserve ma- terial ordinary 1-ml pipet s may be used for handling both the solutions to be tested and the reagents. Not only crystals, but also amorphous precipitation, or absence of precipitation, are observed. Amorphous precipitates may be finely divided or curdy, or in drops, the latter often a prelude to the formation of large crystals. Crystals may form directly, or from an amorphous precipitate. As evaporation is usually not controlled, the time during which a test is observed is then only that required for the drop to dry up. Test-crystals often form around the edge, especially with dilute solutions, as the drop dries. The reagent it- self may crystallize out when evaporation has gone far enough, and a reagent-blank 31 CHEMICAL MICROSCOPY should always be run in comparison with any mm diameter) are used to make these micro- test, until the analyst is thoroughly familiar drops and add reagent (see Clarke), with crystals or any other effect that may To prevent too rapid evaporation a micro- arise from the reagent alone. test is made in a hanging drop. It may be Test solutions can be made up of weighed put on an ordinary slide and inverted over quantities in a few ml or less of water, or a cavity slide. This procedure is also used made right on the microscope slide. The for an ordinary small test-drop (say 0.01 ml), analyst soon learns to judge the amount of simply as a means of preventing or greatly substance needed, say about 0.2 mg if there slowing evaporation and the slide can be is no need to conserve the substance, giving reinverted whenever rapid evaporation is de- with an ordinary drop of about 0.04 ml a sired. As an alternative to sealing-in, to be concentration of about 1:200, which is very quite sure of preventing evaporation while good for many microcrystal tests, usually the test stands, a drop of water, or better, better than more concentrated. Most of the a large "blank" test-drop, which will give tests are best at concentrations between virtually the same humidity as the hanging 1 : 100 and 1 : 5000 (0.4 mg to 8 7 in 0.04 ml) ; drop, can be put in the depression of the the very good ones still succeed at greater cavity slide. The enclosure technique also dilutions, requiring only a few micrograms slows the evaporation of other substances even in an ordinary drop, and the tendency besides water, e.g., iodine from solutions is always to study and use more and more which have only a low content of iodide to sensitive tests. hold it in solution. The average analyst needs micro tests far With a standard "thin" slide (thickness more than "micro" procedures and equip- not over 1 mm) the usual second power ment, aside from the microscope itself. It is magnification (objective 20 or 21 X, or 8 often more convenient to carry out an ex- mm) can be used through the slide. If ob- traction, for example, with ordinary equip- servation under still higher power is wanted, ment, and volumes of, say, 10 ml of solution of course it is necessary to mix the test-drop and 5 or 10 ml of solvent at a time, for a few on a cover-glass. In Clarke's technique the shake-outs, even if the amount of material cover-glass is sealed in with 25 % gum-arabic is quite small and the final residue micro- solution to prevent its slipping and ensure scopic. Often, in fact, the analyst has to only extremely slow evaporation. The seal process quite large amounts of material to may be made not quite complete at fii'st if get an extremely minute residue. Chemistry it is desired to allow a little evaporation has reached a stage where more and more (until precipitation occurs) before finally sensitivity is wanted in tests, and they are sealing. Square 25 mm cover glasses may be almost never too sensitive if satisfactory in most convenient, other respects. In new tests for basic substances in a me- A further long step downward can be dium of strong acid, a very small amount of taken when necessary, by using small drop- solid material is put on a slide (finely pow- lets. If all the substance available will make dered or in a thin deposit in one spot), a only one small "macro" droplet of solution drop of reagent placed near it, and then a of, say, 0.01 ml, it will still be possible to cover-glass, ordinarily of 18-mm size (round try up to about 100 microcrystal tests, using or square), is added so that the reagent microdrops of as little as 0.1 (mm)^ or 0.1 flows over the solid. Or, the drop of reagent microliter, each containing only 1 to 0.02 is put on the cover-glass, which is then in- microgram of substance, to be within the verted over the substance. The reagent con- best range for most tests. Thin solid rods (1 sists of some precipitating compound in solu- 32 CHEMICAL MICROCRYSTAL IDENTIFICATIONS tion in a strong acid, most often syrupy H3PO4 , (2 + 3) H2SO4 , concentrated HCl, or (2 + 1) acetic acid — acids which give re- agents of very different effects even with the same precipitating compound. With a medium of strong phosphoric or sulfuric acid the drop cannot dry up, but its moisture content may gradually change on standing (generally increasing, sometimes considerably), depending on the humidity or dryness of the air. These tests are usually observed during only about two hours, so that ordinarily no special precautions are needed. Small cover-glasses of 12-mm diameter are obtainable. A test made under this size can be inverted over a cavity slide, a procedm'e which will very nearly prevent any evapora- tion of a volatile acid or water, or changes due to humidity or dryness of the air. This is generally sufficient control for at least 24 hours without sealing-in. A further refine- ment is to put a drop of diluted sulfuric acid of a certain strength in the cavity of the lower slide, to regulate the humidity. H2SO4 (45 -j- 55) will correspond pretty closely to a reagent of syrupy H3PO4 ; H2SO4 (1 -f 5) will provide controlled humidity; H2SO4 (55 + 45) provides gentle drying for H3PO4 reagents. Reagents in strong acids may also be added to aqueous solutions, or applied to solids in a form somewhat diluted with wa- ter, without a cover-glass. Then formation of a precipitate and crystallization may occur promptly, or only as the drop evaporates to a higher acid strength, particularly in us- ing phosphoric acid. Sometimes drying in a desiccator may be useful, e.g., with HAuCU in H3PO4 , or a controlled humidity to allow drying just to a certain degree, or to prevent complete drying when the air is very dry, e.g., with the reagent HsBile in (1 -f 7) H2SO4 . A hanging drop of reagent is also used in tests for volatile substances. This is more or less familiar in chemical and microcrystal tests for ammonia from ammonium salts, and as a test for urea when it is decomposed to ammonia by urease. Microcrystals tests in the hanging drop for the readily volatile d- and c?/-amphet amine were perhaps first used by Griebel, with aqueous reagents. The ma- terial (e.g., a small amount of scrapings from a tablet containing an amphetamine salt), or its solution, is treated with dilute alkali on a cavity slide, and a hanging drop is in- verted over it. There are now at least a dozen such sympathomimetic drugs regu- larly on the market, that are readily volatile, to be tested in this way, and others that are less readily volatile. It is certainly not yet realized how far such tests can be extended, how many sub- stances usually considered nonvolatile are in fact slightly volatile with water vapor even at room temperature and can be detected in a hanging drop with an exposure of, say, two hours, for example, I- and c?/-ephedrine, and phenmetrazine. Gentle heat from above, e.g., from a desk lamp, may be used to shorten the time required, and may improve results, e.g., in the vaporization of pentylenetetrazol, but a longer exposure at room temperature is as good or better for most compounds. On the acid side benzoic and salicylic acids may be mentioned as sufficiently vola- tile with water vapor for vaporization tests. A hanging drop of an aqueous solution of reagent, above an aqueous solution on a cavity slide, does not dry up or expand un- less considerable hygroscopic material is present in one solution, and a long exposiu'e may be used. For basic vapors, a reagent in diluted H3PO4 , (1 + 4) to (1 + 2), may be exposed for some two hours or more, with some increase in the size of the hanging drop due to absorption of water. The slide with the hanging drop is then reinverted to allow evaporation, even put in the desiccator if necessary, and tests can be obtained for a great many basic substances even if only slightly volatile. A convenient purification as well as a test 33 CHEMICAL MICROSCOPY for a volatile base is to catch the vapor in a substances, many such tests undoubledly hang;ing drop of dilute HCl (say 1 % by will be found. However, the most familiar volume of the concentrated acid), over an microcrystal tests at present are those for alkaline drop on a cavity slide. Then the alkaloids and related substances, and related slide with the hanging drop is reinverted and tests for other derivatives of basic nitrogen, the drop allowed to dry up, leaving the hy- drochloride of the volatile base. This solid Criteria for Selecting the Best Identi- hydrochloride may then be examined micro- iicatioii lests scopically, if it crystallizes, or it may be (1) Highly characteristic crystals (which tested with HAuBr4 in H3PO4 , or some other may be so by reason of two or more types reagent of high acid content; or with this occurring in the same test -drop, or for other purification it may be simply dissolved in reasons besides form); (2) Easily recogniz- water for some test. able crystals; (3) Crystals forming very Some substances decompose in alkaline readily; (4) Crystals of the same type or solution and give off a volatile base. For types over a wide range of concentration or example, the drugs levarterenol, epinephrine, with widely different amounts of material and isoproterenol, all diphenols and of simi- (when direct addition is used); (5) Crystals lar uses, give off ammonia, methylamine, reasonably stable; (6) Crystals not essen- and isoproterenol give off ammonia, methyl- tially changed by the presence of nonreactive amine, and isopropylamine, respectively, and material (e.g., tablet excipients); (7) Crys- can be readily distinguished in this way, with tallization not prevented nor badly altered a hanging drop of sodium tetraphenylboron when impurities of other nitrogenous bases solution. The assm'ance of the identity of are present in quite minor amounts (ex- the substance given off that is afforded by perience with actual uses of the test affords microcrystals raises these tests much above the best information on this point); (8) High the usual chemical decomposition tests. sensitivity; (9) Colored crystals are prefer- Still, a decomposition test is usually not able (i.e., use of a colored precipitating com- the best, if a direct test is available. In show- pound); (10) The reagent by itself (blank ing contamination by rodent urine, instead test) should not give any possibly confusing of the indirect test for urea by decomposition crystals; (11) A permanent reagent is prefer- with urease and detection of the ammonia able; (12) A slow-evaporating or non-drying evolved (a ubiquitous substance, and usu- reagent is preferable; (13) A reagent of gen- ally detected merely by its alkalinity, al- eral usefulness for microcrystals is preferable, though microcrystals can be used here, too) or alternatively (for some purposes) one that a direct microcrystal test with xanthydrol does not precipitate at all with most other can be used. In (2 -f 1) acetic acid this is alkaloids and amines; (14) Comparisons, amazingly prompt and sensitive for an or- both in general, and with the most closely ganic reaction, and results in crystals of a related compounds available, should estab- urea compound, dixanthylurea, which are lish the meaning and valvie of the test for observed directly, with obvious forensic and the particular substance. other advantages for the test. This is an example of the scattered microcrystal tests Future Lses which are possible and depend on varied It seems likel^M hat for the near future the organic reactions. usefulness of microcrystal tests will remain If analysts, generally, would try to find chiefly in the drug field, as regards the sub- microcrystal tests whenever this type would stances tested, and in general in regulatory be the most advantageous way of identifying and law-enforcement work. Here, the highly 34 CHEMICAL MICKOCRYSTAL IDENTIFICATIONS individual character of the tests is much more an advantage than a defect. Here, highly characteristic and specific tests, the results of which can be photographed for possible introduction in court or in other proceedings, are welcomed. Here, it is a great advantage that the tests are simple and di- rect, and distinctions made on the basis of obvious, visible characteristics. The identi- fications clearly distinguish between closely related compounds, even between an isomer and the racemate, and by certain tests, be- tween the d- and Z-isomers; and this is just what is needed in the drug field and in forensic chemistry and related work. In these applications the microcrystal tests are already used, and so too are color tests, of a kind with which the crystal tests form a natural partnership. The classification of the microcrystal tests, to enhance and extend their usefulness, re- mains a problem. Compounds are onty too likely to be met with occasionally, to which the tests are applicable but which still can- not be identified. This of course is due not merely to lack of classification, but is, at present, primarily because the tests for so many new drugs and other compounds have not been studied at all. However, this is a related aspect of the problem, for if a satis- factory system of classification is once worked out, the necessary experiments to find where all the compounds likely to be met with fit into the scheme are more likely to be made. WTiether a classification of crystal results should be primarily chemical, or morphologi- cal with the different reagents, or based simply on the fact of crystallization of each sub- stance with certain reagents out of a stand- ard list, there can be no reasonable doubt but that the science of microcrystal tests should be integrated with analytical chem- istry, not pushed aside as a "specialty" nor even regarded as being more microscopy than chemistry. The idea of using the microscope to make tests better for identification by simply look- ing at the different crystals formed in the usual precipitation reactions and other crys- tal-forming reactions of chemistry seems so sensible and also so obvious that it is hard to explain how it can be overlooked, as it usually seems to be. The idea develops nor- mally in three stages: (1) Microscopic observation of chemical tests: The microscope is used to distinguish the crystals occurring in chemical precipita- tion tests and other reactions, and thus to particularize them further; and the tests which are found especially characteristic are noted. (2) Microscopic study of microcrystal tests: The tests thus found of especial value be- come in themselves the basis for identifica- tion, and the best reagents thus found are tried systematically on other, related com- pounds. In this stage, appreciation of the chemical nature of the tests sometimes al- most disappears. (3) Chemical extension of microcrystal tests: Chemical reagents and reactions are reconsidered and studied with the objective of improving microcrystal tests and extend- ing them to new groups of compounds. As the tests come to cover a wider area in or- ganic chemistry than simply "alkaloids", their chemical meaning in various cases be- comes important. The chemist needs the microscope, not only for primarily observational use, as in looking at the material he is dealing with, and as a convenient aid in all kinds of chemi- cal procedures and research, for example, whenever he has a deposit or residue and merely wants to know whether it is crystal- line or not. . . . He needs the polarizing mi- croscope, not only as a means of identifying crystals of chemical substances by refractive indices and other properties shown by opti- cal crystallography. . . . Most of all, in his own capacity of chemist, he needs it, as this article has tried to show, in analytical iden- tification chemistry, as the prime means of 35 CHEMICAL MICKOSCOPY making chemical tests more definite. The science of microcrystal tests is ah-eady useful now, especially in the drug field as well as in inorganic chemistry, but it needs to be fully integrated with analytical chemistry, and extended over all organic compounds that can give characteristic, identifying crystals in chemical reactions. REFERENCES 1. Stephenson, Charles H., "Some Micro- chemical Tests for Alkaloids; including Chemical Tests of the Alkaloids Used," by C. E. Parker, Philadelphia and London, J. B. Lippincott Co., 1921. 2. Kley, p. D. C. (a) Behrens-Kley, "Mikro- chemische Analyse," 1st part, (as 4th ed. of "Anleitung zur Mikrochemischen Analyse," by H. Behrens), Leipzig, Leopold Voss, 1921. (Inorganic), (b) Behrens-Kley, "Or- ganische Mikrochemische Analyse" (as 2nd ed. of "Anleitung zur Mikrochemischen Analyse der Wichtigsten Organischen Ver- bindungen," by H. Behrens), Leipzig, Leo- pold Voss, 1922. 3. Amelink, F., "Schema zur mikrochemischen Identifikation von Alkaloiden; ubersetzt von Marga Laur." (In German). Amster- dam, N.V.D.B. Centen's Uitgevers Maat- schapij, 1934. 4. Geilman, W., "Bilder zur qualitativen Mikro- analyse anorganischer Stoffe," Leipzig, Leopold Voss, 1934. Republished by Author- ity of the Alien Property Custodian, by J. W. Edwards, lithoprinted by Edwards Brothers, Ann Arbor, Michigan, 1944. 5. Rosenthaler, L., "To.xikologische Mikro- chemie," Berlin, Gebriider Borntraeger, 1935. Republished by Authority of the Alien Property Custodian, by J. W. Edwards, lithoprinted by Edwards Brothers, Ann Arbor, Michigan, 1946. 6. Wagenaar, M., a series of articles on micro- chemical tests for particular alkaloids and related substances, (in Dutch), Pharma- ceutisch Weekblad voor Nederland, 64-67 (1927-30); 70 (1933); 72-73 (1935-36). 7. Herzog, Alois, "Mikroskopische Bilder fiir den Chemiker," Zeiss Nachrichten, 2 Folge, Heft 5, 149-81 (1938), and Heft 6, 183-215. (English summary at end of the volume.) 8. Whitmore, W. p., and Wood, C. A., "Chemi- cal Microscopy of Some Toxicologically Im- portant Alkaloids," Mikrochemie , 27, 249- 334 (1939); "Scheme for the Microchemical Separation of Some Toxicologically Impor- tant Alkaloids," Mikrochemie, 28, 1-13 (1940). (Both in English.) 9. Chamot, Emile Monnin, and Mason, Clyde Walter, "Handbook of Chemical Micros- copy, Volume II. Chemical Methods and Inorganic Qualitative Analysis," New York, John Wiley & Sons, 1940. 10. DucLOux, Enrique Herrero, "Notas Micro- quimicas sobre 'Doping'," Buenos Aires, Peuser Lda., 1943. 11. Fulton, Charles C. (1) Reagents: "The Pre- cipitating Agents for Alkaloids," Ain. J. Pharmacy (April, 1931); "New Precipitating Agents for Alkaloids and Amines," Am. J. Pharm., (Feb. & Apr., 1940); "The Relation of Alkaloidal Chemistry to Inorganic, and the Use of Bromauric Acid as a Reagent for Inorganic Microcrystal Tests," J. Am. Pharm. Assn., Scientific Edition, 31, 177 (1942). (2) Classification by aqueous pre- cipitation: "Alkaloids and Their Reagents," Am. J. Pharm. (May, 1939); for earlier, see /. Association Official Agricultural Chemists, 13, 491 (1930) . (3) Particular substances and development of new-type reagents: Atro- pine, ihid., 12, 312 (1929); Cocaine and Pro- caine, Ajn. J. Pharm. (July & Aug., 1933); Heroin (with G. D. Williams), A7n. J. Pharm. (Sept., 1933); Morphine, /. Lab. Clinical Medicine (March, 1938) ; Cinchona alkaloids, Ind. Eng. Chem., Anal. Edition, 13, 848 (1941) ; Morphine, Heroin, Dilaudid, Cocaine, in American Journal of Police Science, incorporated in Journal of Criminal Law and Criminology, 32, 358-65 (1941) ; Methadone, Narceine, Thebaine, in United Nations document (mimeograph) E/CN.7/ 117, 14 April 1948, (also includes photomi- crographs of the natural narcotine crystals of opium— see "Origin of Opium"), and Codeine in Add. 1 to this document, 22 Sept. 1948; Colchicine, Jour. AOAC, 41, 756 (1958). 12. Hartshorne, N. H., and Stuart, A., "Crys- tals and the Polarizing Microscope," London, Edward A. Arnold & Co., (1943), 2nd ed. 1950. (Optical Crystallography.) 13. Farmilo, Charles G., Levi, Leo; Oestrei- CHER, P. M. L.; AND Ross, R. G. "Micro- crystal and Colour Tests for the New Syn- thetic Narcotics," Bulletin on Narcotics, 4, No. 4, 16-42 (1952). 14. Clarke, E. G. C, and Williams, Margaret, "Microidentification of the Opium Alka- loids," Bulletin on Narcotics, 7, No. 3-4, 36 FORMS OF MICROCRYSTALS 33-42 (1955); "Microchemical Tests for the Identification of Alkaloids," /. Pharmacy and Pharmacology, 7, 255-62 (1955). 15. Clarke, E. G. C. (a) "Microchemical identi- fication of Sugars as Osazones", J. Phys- iology, 135, 28-9 P, from "Proceedings of the Physiological Society," December 1956. (b) In /. Pharmacy and Pharmacology, "Microchemical Identification": Local Ana- esthetics, 8, 202-6 (1956); Some Less com- mon Alkaloids, 9, 187-92 (1957); Antihista- mines, 9, 752-8 (1957); Antimalarials, 10, 194-6 (1958). Atropine-Like Drugs, 11, 629-36 (1959). "Microchemical Differentia- tion between Optical Isomers of N-Methyl- morphinan Analgesics," 10, 642-4 (1958). (c) "Microchemical identification of some modern analgesics," Bulletin on Narcotics, 11, No. 1, 27-44 (1959). Charles C. Fulton FORMS OF MICROCRYSTALS Descriptions and classifications of micro- chemical crystal forms have scarcely any relation to orthodox crystallography. What is needed is not a deduction of the crystal system — even if that were generally possible, which it is not — but simply a straightforward way of describing and classifying the forms just as they are seen as a result of the chemi- cal test. In the comprehensive study and classifi- cation of microcrystals, along with /orm goes size, and the latter is measurable. In the fol- lowing classification of forms, however, the concern is not with the actual measured size, but only with considering, in relation to form, how many different measurements, of value to a description or classification, might be made on a particular type of crystal. Crystals exist, of com'se, in three dimen- sions, but may extend significantly only in one direction, length, if they are fine needles, —or only in two, length and breadth, if they are plates or blades of negligible thick- ness. In the latter case, moreover, they may lie flat, especially if under a coverslip; so that such crystals, certainly as we see them, extend simply in two dimensions. The number of dimen.sions subject to use- ful measurement will not (in general) be greater than the number of dimensions of ex- tension, but may be smaller. In the case of a regular hexagonal plate, for example, only one measurement is needed to complete the description (so far as form and size go) — no matter whether we make it as the length of a side, or as a radius, or as the diameter from one vertex to the opposite one. The writer uses the term "grain" for crys- tals which resemble the familiar grains of sand, salt, sugar, etc., as they appear under moderate magnification. The three dimen- sions of "extension" are essentially equal, so that only one measurement is needed (the diameter) to complete a description in a particular case. These dimensional concepts give six classes; but there are two distinct kinds of crystals extending in three dimensions, each requiring two different measurements. They may best be taken as two separate classes (see table). It is also convenient to separate elongate or directional plates from blades. Together with a zero class for crystals of no significant dimensions (appearing as mere specks with the magnification usually used), this arrangement therefore provides nine classes in all (0-8) for the simple crystal forms. Distortions, contortions, and skele- tons, as well as aggregates, should be referred to the corresponding sunple forms. Skeletonized crystals are here classed with the forms from which they are derived (which may occur along with them), rather than with the simple forms they may finally resemble in their parts. For example, some crystals vary from square plates to crosses. The flat, thin cross should still be assigned to the same class as regular plates although it might be described as having arms of blades. Skeletonized forms can often be distin- guished from aggregates by the birefring- ence. If a complex form extinguishes and brightens as a whole, it usually is basically 37 CHEMICAL MICROSCOPY Table of Crystal Forms 0-0 Specks; precipitute often aniorphous-lookiiiji with l)irefriiigent specks seen Avith crossed nicols. 1-1 Needles (fine) distortions — hairs aggregates — fans, tufts, sheaves, bundles, rosettes, dendrites of needles; burs; wigs, spirals, balls of hairs 2-1 Plates essentially regular (triangles, squares, hexagons, octagons, disks) unformed or vague — precrystalline disks, smudge rosettes distortions — irregular but not definitely elongate plates, flakes skeletons— crosses, stars (formation all one crystal) 2-2 Elongate or Directional Plates (diamonds, rhomboids, elongate hexagons, etc.) skeletons — X's, flat nail shapes, flat combs, etc. twins — hourglass shapes, books, etc. aggregates — rosettes of plates; stepped or segmented plates; spears (built of dia- monds) 2-2 Blades (thin flat forms with length considerably greater than breadth) distortions — irregular blades; splinters, ribbons skeletons — serrate and feathered blades aggregates — fans, sheaves, rosettes, dendrites of blades 3-1 Grains (cubes, polyhedra, pyramids, gems, kernels, sharply angular grains) unformed or shapeless — globes, globulites, spherulites; nuggets aggregates — granular clumps; sphero-rosettes of kernels; chains of grains 3-2 Rods, prisms; spindles, thick coarse needles,— (unimmeasurable cross-section) distortions — sticks, wires aggregates — fans, sheaves, rosettes, clusters, dendrites of rods, etc. 3-2 Regular Tablets (regular plates with thickness) distortions — tablets irregular but not definitely elongate skeletons — regular forms with thickness (e.g., crosses); regularly branching iso- tropic dendritic skeletons 3-3 Elongate Tablets and Bars with edges — chisels; thick ribbed blades skeletons — thick combs, ladders, etc.; dendritic anisotropic skeletons aggregates — clusters of bars; irregular multiformed dendrites a single crystal, but if different parts brighten and extinguish independently, the whole must be composed of different simpler crys- tals growing from each other or from the same point. The above table summarizes the sug- gested classification, including numerous kinds of distortions, skeletons, and aggre- gates. This is not a final answer to the prob- lem, but shows that the many extremely di- verse kinds of crystals can be grouped in a small number of classes, using primarily the two simple concepts of extension and meas- urement. Charles C. Fulton HISTORY The microscope was invented between 1590 and 1609, and thus was available dur- ing the very beginning of scientific chem- istry. Scientists of the 1700's used it to look at all sorts of small things, including natural crystals and such things of chemical nature. Some trvie chemists, as soon as there were any who deserved this name, used it, in an observational way, in their chemical re- searches. True "chemical microscopy", although not then known by this name, began at least as early as 1742. In that year "The Micro- scope Made Easy", by Henry Baker, Fellow 38 HISTORY of the Royal Society and Member of the Society of Antiquaries of London, was pub- lished. Baker stated the following, under the chapter heading, '"Of Salts in. Mineral Wa- ters": "The Microscope may be of great Service to determine by ocular Examination, what kind of Salts our medicinal Springs are charged with, whence to form a Judgment in what Cases their Waters may be drank to Advantage." Five kinds of "fossile salts" were then enumerated. Marggraf, in 1747, first published his re- search, which proved the presence of a true sugar in the beet, and thus laid the founda- tion of the German sugar industry. Despite his early date, he called himself and may truly be called a chemist (then "chymist"), and he used the microscope as a matter of course, as an aid in this research, remarking that little white crystals, looking like those of sugar, could be seen sprinkled over dried slices of the root, as a preliminary indication of the presence of sugar. Another book by Baker, "Employment for the Microscope", was published in 1753. This work had two parts, of wiiich Part I, con- tainmg 232 pages out of 442 for the book, had the title: "An Examination of Salts and Saline Substances, their Amazing Configura- tions and Crystals, as formed undex* the Eye of the Observer." His method was to pre- pare a saturated solution of a salt and then observe the formation of the crystals as a drop, spread out on a slide, began to dry up. There were 55 short chapters on dif- ferent kinds of salts, and 9 plates out of 17 for the book illustrated their crystals. Not much was yet knowai about chemical com- position, and Baker did not use reagents in his procedure. He was primarily a micros- copist and such chemistry as he had was about as crude as it could be ; but it may be pointed out that he preceded Raspail, often cited as having been "the first" or "earliest" in microchemistry, microscopical chemistry, or chemical microscopy, by 80 years; also he preceded Emich, who still is often consid- ered a pioneer, by over 150 years. Chemistry became fully scientific in the late 1700's and in the early 1800's micro- scopic chemistry was still basically an obser- vational science. Chemists examined rocks, sections of plants, any and all kinds of things under the microscope, as a means of learning something about their chemical constitution. By this time, they tried reagents on the things seen, to observe w^hich parts or par- ticles reacted and how. The reagents might be of any kind and in fact color reactions were perhaps most used. Crystals, certainly, were observed, as they had been by Baker and Marggraf, but they were still usually natural crystals, or crystals of common chemicals, or of substances isolated by some chemical process; as yet they w'ere very sel- dom crystals formed in tests intended to produce them for diagnostic purposes. This microscopical chemistry, which was the original "microchemistry", was the chemis- try of substances observed through the mi- croscope. Eminent chemists of this transitional time who used microscopic methods in some of their chemical researches included WoUaston in England and Dobereiner in Germany. Other chemists also used the microscope on occasion, chiefly to look at their material to be worked on; and they included some mi- croscopically observed data in their writings. Thus Hehvig notes that Hiinefeld had in- cluded mention of the crystal forms of the alkaloids then known in a book of 1823. Most emphatic of the early chemists in recommending the microscope for the uses just described, and indeed throughout chem- istry, was F. V. Raspail, whose "Xouveau Systeme de Chimie Organique, Fonde sur des Methodes Nouvelles d 'Observation", w^as first published in Paris in 1833. Organic chemistry then and for some time still was the chemistry of living things, their constit- uents and products; Raspail may be called a biochemist in modern terms. 39 CHEMICAL MICROSCOPY At about this time a new means of ob- not the first, as stated by the "Dictionary serving crystals microscopically was devel- of National Biography", for Talbot had al- oped, that is, the polarizing microscope, ready taken the first photomicrographs of Double refraction had been observed by history as early as 1835, and reported on his Bartholinus as long ago as 1069, and about work in January 1839. Both men used the 1678 was partially explained by Huygens, solar microscope, on which Reade made im- who later in the century also made further provements. Talbot specifically called atten- observations bearing on polarization. The tion to the value of the photography for definitive discovery of polarized light is recording the appearance of microscopic, credited to Mains, over a hundred years chemical crystallizations, later, in 1808. By 1811 it was completely At least as early as 1838 toxicologists were explained in terms of light theory by Fres- concerned with how they might identify al- nel and by Arago. The optical crystallog- kaloids, for obviously no drastic treatment raphy of minerals was developed rapidly by could be used as in the separation of a min- Malus, Biot, and others, especially Sir David eral poison, and then there was the problem Brewster from 1811 to 1819 and later. of distinguishing these poisons from harm- Brewster seldom referred to magnifying less, possibly prevalent, natural bases. Al- aids but used them as convenient; e.g., by ready some suggestion had been made that mounting a plano-convex lens on a plate of microscopic distinctions might be possible, tourmaline or agate as an analyzer, and even, but in 1838 there was as yet little or no idea on occasion, directing polarized light through of using a reagent to produce crystals which a compound microscope. Sources of the could be seen microscopically to be different polarized light were reflection (Mains' origi- for different substances, nal observation), or oblique transmission This idea, however, gradually appeared through a set of plates of glass, as well as during the 1840's and 1850's. It was present, light transmitted through tourmaline, which somewhat rudimentally, in some work of has extreme dichroism. A black mirror could Thomas Anderson of Edinburgh, who de- also be used to "analyze" by reflection; and scribed test-forms of some free alkaloids and also, as early as 1819, Brewster had discov- their thiocyanates in 1848. A specific crystal ered how to extinguish one of the images of test appeared in 1852-53, with Herapath's calcareous spar, nearly perfecting the ana- discovery of the remarkable quinine iodo- lyzer, and anticipating, in a way, Nicol's sulfate, and his own recommendation that invention. the crystals could be formed in a drop of Nicol invented his remarkable prism in reagent as a test for quinine. Both of these 1828. In 1834 (W.) H. F. Talbot made the were included in "The Micrographic Die- polarizing microscope a definite instrument, tionary", by J. W. Griffith and Arthur Hen- and immediately applied it to chemical sub- frey, first edition 1856. In 1859 Taylor, in jects. This w^as the same Talbot who became the second edition of his toxicology, "On one of the founders of photography, and of Poisons", gave seven different microcrystal modern archeology. Apparently chemists in tests for strychnine, and even made refer- general were not alert to utilize the new in- ence to the "polarizing properties" as help- strument, and only in recent years have ing to characterize the crystals, as well as they begun to borrow it back from the min- giving some other microcrystal tests, includ- eralogists. ing one for cyanide using crystals given by J. B. Reade, another chemist, microsco- HCN vapor in a hanging drop of AgNOs ; pist,and photographic discoverer, took photo- but the microcrystal tests were not system- micrographs in 1839. However, they were atically developed. 40 HISTORY The new science of microcrystal identifica- tion tests came of age in the 1860's. Perhaps it may be said that the idea had crystalhzed with the issuance of the prospectus of Worm- ley's book in the United States in 1861. This pubhcation was delayed by the Civil War, and actually the honor of the first book of this science, at least the first findable by the writer, goes to Helwig, whose "Das Mikros- kop in der Toxikologie" was published in Germany in 1864 and 1865. Chamot and Mason have also mentioned a similar book by Erhard, "Die giftige Alkaloide u. d. Aus- mittelung auf Mikroskopischen Wege", 1866. Wormley's great book, "Microchemistry of Poisons", was fuially published in 1867 (1869). He gave attention to sensitivities and established the science on a firm basis. His book is a landmark; the tests are still good but, of course, the number of com- pounds covered has now become quite inade- quate. A second edition was issued, without very much change, in 1885. In those days, most poisons were inorganic or else natural plant alkaloids, and crystal tests were developed for both kinds. Later, such tests were extended over the whole field of inorganic chemistry, but their ex- tension to other organic compounds besides the alkaloids has been \'ery slow and meager. Inorganic tests were further developed by Haushofer in "Mikroskopische Reaktionen" (1885), and by Klement and Renard in "Reactions Microchimiques" (Brussels, 1886). Behrens' outstanding work, with some publication as early as 1882, culminated near the end of the century in "Anleitung zur Mikrochemischen Analyse", 1895-97. He brought the inorganic part to a high level, and made a strong effort to develop the science throughout the organic field, with many examples of reactions and descriptions of crystals suitable for microscopic identifi- cation tests. In spite of this, the majority of chemists even today think of such tests as suitable only for alkaloids, if they use them at all. A new edition of Behrens' work, by Kley, in 1921-22, represented the fourth edition for the inorganic part, the second for the organic part. The inorganic field has been developed further, and admirably, by Chamot and Mason, especially in volume II of the second edition of their "Handbook of Chemical Mi- croscopy" (1940). The work on alkaloids was carried on in excellent fashion by Stephenson, "Some Microchemical Tests for Alkaloids", 1921, and Amelink, "Schema zur mikrochemischen Identifikation von Alkaloiden", 1934. Ame- link gave attention to just a few compounds besides alkaloids, reacting with the same reagents. "Toxikologische Mikroanalyse", by Ro- senthaler, 1935, included tests for various inorganic and other organic compounds as well as alkaloids and their modern analogs, and is very valuable although it often seems not too well based on the best preceding work. (It was reissued in the United States in 1946, and is still "in print".) While Behrens was especially influential, "microchemistry" usually meant the use of the microscope in making chemical identifi- cations, especially by means of reactions pro- ducing characteristic crystals. However, in the last 40 years, to most chemists the term "microchemistry" has come to mean merely small scale chemistry. Meanwhile the oldest type of microchemistry, i.e., the chemistry of things observed through the microscope, also survived and was further developed by botanist-chemists, Tunmann and Molisch in particular; it now includes considerable use of crystal-forming reagents, but is not lim- ited to them. Only the beginning of optical crystallog- raphy has been noted above, and no attempt has been made to trace the development of micro-sublimation, or micro-melting-points and the modern fusion microscopj'. 41 CHEMICAL MICROSCOPY Numerous studies giving microcrystal tests for various alkaloids and occasionally other compounds can be found in the periodi- cal literature of the last century and this one. In 1932 and 1940 the present writer reviewed the field of "alkaloidal" reagents used for such tests, and began the work of extending them to cover all compounds of basic ni- trogen. In spite of the solid body of work on mi- crocrystal tests cited above, and a consider- able total of scattered work, largely on al- kaloids, the general impression of past history, and certainly the present situation, is one of neglect. In most fields, the authors who have contributed to the development of microscopic chemistry are surprisingly out- numbered by those others who do not recog- nize any application of the microscope to that field at all. This is most astonishing in toxicology, for here the science of crystal tests began. Hel- wig thought it strange that many toxicolo- gists seemed unaware of the obvious advan- tages of the use of the microscope in their science, but this is still true nearly 100 years later, as published works show. The neglect is at least ec^ually great elsewhere. In most colleges and universities, courses in ciuali- tative analysis or analytical chemistry, whether inorganic or organic, usually do not even mention use of the microscope; if a query about it is made, it is passed off as a "specialty". How has this neglect come about? Un- doubtedly one factor is simply that the field underwent some development very early. It has often happened that a subject developed earliest is not developed best, for later re- searchers hesitate to work in a field already partly tilled. A most important factor, however, has been the strangely long period through which chemistry has passed, during which only quantitative applications were considered to have any value at all. The microscope is not primarily a quantitative tool; this is one limitation on the tests. It has even happened that reagents already well-known and useful as alkaloidal precipitants for microcrystal tests were renamed "Wagner's reagent" and "Mayer's reagent", for example, for alleged quantitative uses almost devoid of value. For three-quarters of a century or more it was the use of the microscope in chemistry, in one way or another, either purely obser- vational, or a necessary aid in tests with crystal-producing reagents, that was com- monly known as "microchemistry". Then, this name was appropriated for what is merely chemistry on a small scale, often not even on a "micro" scale at all; this new "mi- crochemistry" had quantitative aspects, and by about the 1920's it was enthusiasically received. Emich himself spoke of the micro- scope as indispensable, and described nu- merous microcrystal tests, largely derived from Behrens, in his "Lehrbuch der Mikro- chemie" (1911), endorsing also the use of optical crystallography; but since then the microscope seems to have been lost in the shuffle. The use of microcrystal tests in the or- ganic field never has disappeared among narcotic chemists, law-enforcement chemists in general, drug analysts, and others. At the present time E. G. C. Clarke in England is doing very effective work in extending the coverage of microcrystal tests, once again in toxicology. The flood of new drugs in recent years has vastly increased the demands and difficulties for reliable identification tests. The need of extended development of micro- crystal tests was never greater than it is to- day. Charles C. Fulton MIXED FUSION ANALYSIS Microscopic mixed fusion analysis consists of the microscopical observation of the melt- ing and solidification behavior of mixtures of two or more fusible substances. Two main methods of mixture preparation are em- 42 MIXKD FUSION ANALYSIS ployed: the components may be thoroughly mixed physically or a contact preparation may be made. Melting and solidification phenomena are usually observed on prepara- tions contained between a microscope slide and cover-glass; however, it is sometimes desirable to contain the mixture in a sealed micro-capillary tube. Equipment. The equipment necessary for microscopic mixed fusion analysis is usu- ally quite simple. Any microscope with magnification up to about 200 X and capable of accommodating a hotstage is satisfactory. Some method for achieving polarized light is desirable; although this may be as simple as two pieces of "Polaroid" mounted to serve as a polarizer and an analyzer. For observa- tions at various temperatures, a good hot- stage is mandatory. Hot stages covering the range 30-350°C are commercially available as is a coldstage capable of functioning above — 100°C. In addition, there are published in the literature designs for both hot and cold stages applicable at more extremes of tem- perature. Temperatures are measured either with a thermometer or a thermocouple. The commercial stages employ thermometers. Temperatures and heating rates are usually controlled with a variable resistor or a variable transformer. It is imperative that the temperature measuring devices be accurately calibrated with known melting point standards at a heating rate nearly identical with that at which an unknown is to be measured. Usually, a standard heating rate (such as 3°C/min) is selected and all calibration and melting point determinations are made at this heating rate. It is very helpful to have a good voltmeter in parallel with the hotstage heating element so that heating rates may be approximately ad- justed to the desired value at any given temperature from predetermined stage char- acteristics. A hot bar is a very useful adjunct to the hotstage microscope, especially for contact preparations. These devices are commer- cially available or may be constructed in the laboratory. However, it is possible to employ an alcohol lamp, a micro burner, or even a soldering iron to achieve the same ends. A micro sul)limator for purification of com- pounds is also a useful item for microscopic mixed fusion analysis. Contact Preparation Methods. Con- tact preparation methods rapidh^ yield in- formation concerning the nature of the phase diagram of a two-component system. A more restricted definition of microscopic mixed fusion analysis would include only contact preparation methods. A contact preparation is made in the following fashion. A small amount of the higher melting component (A) is melted between a microscope slide and cover-glass so that, on solidification, approxi- mately one half of the area of the cover-glass contains crystalline material. A small amount of the lower melting component (B) is then placed adjacent to the crystals of component (A) and in contact with the cover-glass. When the slide is warmed so that component (B) melts, the melt flows under the cover-glass and dissolves a portion of component (A). The entire preparation is then allowed to solidify. It is freciuently de- sirable, in order to insure adequate mixing, to melt back the preparation so that most of the higher melting component is melted and then to allow the entire preparation to resolidify. A concentration gradient, ranging from pure A on one side to pure B on the other side, exists in such a preparation. All inter- mediate compositions exist in some area of the zone of mixing. ^Microscopical observa- tion of the mixing zone dm'ing the cooling process yields information concerning the nature of the phase diagram between A and B. If the system is simple eutectic, both components will crystallize rather rapidly vmtil the crystal fronts reach the mixing zone. Crystal growth will then proceed at a greatly reduced rate. When the eutectic tem- perature is reached, fine grained crystals of 43 CHEMICAL MICROSCOPY eutectic composition will crystallize rapidly without appreciable change in velocity as throughout the mixing zone. The appearance the preparation is cooled. The resultant solid of a eutectic zone is quite characteristic and crystals will appear homogeneous through easily recognized once the observer is famil- the preparation and the polarization colors iar with the phenomenon. will be uniform. On heating, the entire prep- Molecular addition compound formation aration will melt at the same temperature as is readily observed in that a third solid the two starting components. Small amounts phase may be seen to crystallize in the mix- of impurities modify the behavior only ing zone of the two-component system. One slightly. If the two components are not or two eutectic zones are also observed, de- identical, there will normally be a marked pending on whether the addition compound depression of the melting point in the mixing melts incongruently or congruently. The zone. various types of solid solution are also read- Frequently, if a contact preparation be- ily detected on solidification of the melt, tween an unknown and an easily supercooled With solid solution, both component A material such as thymol is allowed to digest and component B crystals are seen to grow on the hotbar, the unknown will develop well through the mixing zone with a change in defined crystal faces in the mixing zone. It growth velocity but without the appearance is then possible to measure accurately such of an area of eutectic composition or a mo- quantities as profile angles, extinction an- lecular addition compound. If there is partial gles, dichroisom, and refractive index rela- miscibility of liquid A and liquid B, the two tive to the melt for identification purposes, liquid phases may be seen in the mixing It is also sometimes possible, with the proper zone. choice of second component, to nucleate un- Observations made during heating of a stable polymorphic forms of a given com- contact preparation serve to confij-m the pound. This enables the investigation of more rapidly obtained conclusions drawn polymorphic forms difficultly available or from observations made on cooling. In addi- unattainable by other means, tion, it is possible to measure the various An identification scheme based upon eu- melting points such as the eutectic tempera- tectic melting points has been published by tures, molecular compound melting points, the Koflers. Unknown compounds are sub- maxima or minima in solid solution systems, divided into various restricted temperature and the melting points of the two starting ranges and two reagents are used for each components. The accuracy of such measure- temperature range. The eutectic temperature ments is dependent on the accuracy of between the unknown and the two reagents, calibration of the hotstage and the care with together with the refractive index of the melt which the proper heating rate is maintained, determined by the Koflers glass powder Aside from the general determination of method serves to identify the unknown com- the type of phase diagram between two com- pound. Over 1200 compounds have been so ponents and the determination of the signifi- catalogued and either the contact prepara- cant temperatures, there are other applica- tion method or the method of mixtures may tions of the contact preparation method, be used. In principle, this method may be Identity or non-identity of an unknown com- applied to any fusible compound provided pound and a suspect compound may be rap- suitable reagents are chosen, idly established both by observation on A different method of identification based cooling and by observation on heating. If upon molecular addition compound forma- the two are identical, the growing crystal tion and applicable to aromatic compounds front will pass through the mixing zone has been published by Laskowski, Grabar, 44 MIXED FUSION ANALYSIS and McCrone. A contact preparation be- tween the reagent (2,4,7-trinitrofluorenone) and the unknown estabhshes if the unknown is of the class which forms addition com- pounds with the reagent. If an addition com- pound forms, the preparation is observed microscopically on heating. Identification is achieved on the basis of melting point of the unknown, the molecular addition compound, the eutectic between the reagent and the addition compound, and the eutectic be- tween the addition compound and the un- known. Besides these four melting points the color of the addition compound is also ob- served. In several systems the addition com- pound was found to exist in two polymorphic forms, hence additional melting points are available for characterization. This method is rapid and is applicable to liquids as well as solids. It is also applicable if the addition compound melts incongruently. The above identification scheme has been applied to alcohols by Laskowski and Adams. Although the alcohols do not gener- ally form addition compounds, they may be converted to 2,4,6-trinitrobenzoates. The 2,4,6-trinitrobenzene grouping leads to ad- dition compound formation with a variety of aromatic substances. Contact prepara- tions w^ere made between various trinitro- benzoate esters and both naphthalene and phenanthrene as reagents. Four significant temperatures (three if the addition com- pound melts incongruently) were obtained with each ester and each reagent. Satisfac- tory discrimination was achieved among all all of the alcohols investigated. The proce- dure of reacting a functional group with a suitable reagent so that the resultant deriva- tive forms addition compounds with other reagents offers promise for wide applicability of this method of identification. Method of Mixtures. Mixtures of known composition may be prepared by weighing the components together and grinding until a uniform composition is achieved. Since only small amounts of material are required for determination of a melting point with a hotstage microscope, it is possible to deter- mine temperature-composition diagrams rapidly with a minimum expenditure of ma- terials. The points of initial and final melt- ing are easily determined microscopically. In addition it is frequently possible to ob- serve polymorphic transitions in one or both of the starting components. In such mixtures, it is also possible to determine microscopically the effect of com- position and temperature on crystal growth velocity, nucleation rate, rate of nucleation and growth of unstable polymorphic forms, and the effect of added components on crys- tal habit. The method of mixtures is appli- cable to any number of components. SELECTED REFERENCES The literature on microscopic mixed fusion analysis as well as the various experimental tech- niques involved is covered extensively in the books by Kofler (1) and McCrone (2) and the re- view by Cecchini (3). Specific applications of mixed fusion analysis include the identification of aromatic compounds (4), the investigation of molecular compound formation (5), (6), isomor- phic relations between organic compounds of sulfur and selenium (7), identification of alcohols (8), identification of fibers (9), effect of composi- tion on crystal habit (10), and studies on crystal growth velocity (11). Although this list of references is not intended to represent an exhaustive survey of the literature on microscopic mixed fusion analysis, it does serve to orient the reader in the general area. 1. Kofler, L., and Kofler, A., "Thermo- Mikro-Methoden zur Kennzeichnung Or- ganischer Stoffe und Stoffgemisch," Wag- ner, Innsbruck, 1954. 2. McCrone, W. C. Jr., "Fusion Methods in Chemical Microscopy," Interscience Pub- lishers, New York, 1957. 3. Cecchini, M. A., Selecta Chimico, 16, 95 (1957). 4. Laskowski, D. E., Grabar, D. G., and McCrone, W. C. Jr., Anal. Chem.. 25, 1400 (1953). 5. FtJRST, H., AND Praeger, K., Chem. Tech., 11, 653 (1958). 6. Laskowski, D. E., and McCrone, W. C. Jr., Anal. Chem.. 26, 1497 (1954). 45 CHEMICAL MICROSCOPY 7. Cecchini, M. a., and Giesbrecht, E., J . Org. Chem., 21, 1217 (1956). 8. Laskowski, D. E., and Adams, O. W., Anal. Chem., 31, 148 (1959). 9. Grabar, D. G., and Haessly, R., Anal. Chem.., 28, 1580 (1956). 10. Arceneax, C. J., Anal. Chem., 27, 970 (1955). 11. Gilpin, V., J. Am. Chem. Soc, 70,208 (1948). Donald E. Laskowski NITROGEN-BONDED RADICALS: IDENTIFICATION Most of the really excellent microcrystal tests known for organic compounds are due to basic nitrogen, but the crystals are so affected and modified by other elements and radicals, even quite separated from the ni- trogen atom, and by the whole structure, that they identify the molecule of a specific substance as a whole. That they can do so is, in fact, their great value. However, some- times it may be as well, and more conven- ient, to distinguish closely related com- pounds by precisely the points in which they differ. The difference may be in the radicals on a basic nitrogen atom, which volatilize while still attached to it in a deamination re- action. For example, the three USP drugs levarterenol, epinephrine, and isoproterenol, all diphenols and of similar uses, decompose spontaneously in alkaline solution to give ofT respectively, ammonia, methylamine, and isopropylamine, which can be distinguished readily by microcrystal tests in a hanging drop. Still more important are cases where it is desired to learn the identity of radicals on the nitrogen atom as a step in analysis, either in identifying an unknown as some- thing already known, or in formulating the structure of a compound whose precise con- stitution is not known. In either case there are of course chemical tests, such as those given by Feigl, for distinguishing primary, secondary, and tertiary amines; but the microcrystal tests are more definite and show exactly what radicals are attached to the nitrogen atom, provided the deamination reaction can be obtained. With many com- pounds, stable to alkali alone, alkaline oxida- tion with permanganate will cause deamina- tion, the nitrogen carrying with it any simple carbon-hydrogen (non-oxygenated) radical attached to it, as it comes off. For example, hydroxyamphet amine and methoxamine yield ammonia in this way, while phenylephrine and ephedrine yield methylamine (ephedrine volatilizes slowly unchanged, if not oxidized). Hordenine in the same way yields dimethylamine, cho- line yields trimethylamine, and N-ethyl- ephedrine yields methylethylamine. A test of an antibiotic, the exact structure of which was not known, showed that the amine com- ing off, either spontaneously from alkaline solution or more rapidly with alkaline oxida- tion, was unmistakably dimethylamine, showing that the nitrogen atom bore two methyl groups in the original compound. Hanging-drop tests above an alkaline so- lution apply also, of course, to basic com- pounds that are themselves volatile. Also there are cases where an ester is soon hy- drolyzed by alkali, and if basic nitrogen is present in the part that supplies the alcohol rather than the acid of the ester, the simpler basic compound resulting will very probably be volatile, as occurs with procaine and other synthetic anesthetics. Such decomposition compounds can be detected and identified by microcrystals in the hanging drop. However, the tests given below are suggested specifi- cally for the simplest compounds resulting from deamination. The same tests are of course directly applicable to any minute amounts of the lower amines, whether formed by decomposition of a larger mole- cule or already individual compounds. The tests may be useful in several fields, e.g., botanical chemistry, as well as food and drug work. Procedure. Stir a very little of the sub- stance into a drop of 5% NaOH in the de- pression of a cavity slide. Set a plain slide 46 NITROGEN-BONDED RADICALS: IDENTIFICATION on the cavity slide and put on it, over the fringencc, dichroism if it occurs (as with center of the cavity, a little droplet of sodium diethylamine hioniaurate), etc., as well as tetraphenylboron solution (1:20 in water); form. To the case of the red and brownish- then invert this slide. Tetraphenylboron is yellow hexafi;onal crystals with dimethyl- extremely sensitive to ammonia and the amine (as well as the similar crystals, lower amines, and if one of them is evolved, brownish -yellow only, with methylethyl- characteristic crystals form rapidly in the amine) it is surprisingly easy to get a good hanging drop. Examine them with a polariz- interference figure (a uniaxial cross), and ing microscope with the slide in place, or determine the sign of the crystal (positive), transferred over an empty cavity to prevent Working sensitivities, except for reagent further formation. In case of a mere trace of 3, are of the order of hundredths of a micro- ammonia, compare with a blank test using gram of the evolved amine captured in the the same reagents. Ammonia is so common hanging drop. Reagent 3, although less sen- from various causes, including contamina- sitive than was desired, can be vised for tion, and the tests so sensitive, that often highly characteristic crystals with as little its crystals are not very distinctive, as those as 4 or 5 micrograms of ethylamine. The of the lower amines are. alternative given for ethylamine, reagent 4, Usually a result appears within a few min- has the desired sensitivity, but has disadvan- utes, if at all. If there is no result in an hour tages due to the reagent itself, and because or so, add a drop of 1 % KMn04 solution and the crystals of the particular test are less eas- invert over the cavity a fresh droplet of ily distinguishable from others than is the tetraphenylboron solution. Or better, if case with the other recommended tests. The there is no shortage of material, run the test tests with reagent 5 can be obtained in the with oxidation at the same time as the one presence with NH4CI. (See table), with alkali alone. The oxidation test may Compare results closely with the crystals give a different result even when there is given by knowns, until ciuite familiar with some evolution of a volatile base. Examine them. the crystals wdth the slide still inverted, over Selection of a "best test" for ethylamine an empty cavity. gave the most difficulty. Its test with If either technique is effective for crystals IIAuBr4 in 2H-iP04- IHBr was passed over at apparently due to a lower amine, prepare the time the table was drawn up, because another test. Use a hanging drop of simply good crystals form only at the periphery of 1 % of concentrated HCl in water. After an precipitation. However, they are very char- exposure which may be judged from the acteristic, divided hexagons, quite different previous reaction, or up to about an hour, from the crystals of methylamine (or any reinvert the slide and allow the drop to dry others so far seen) with this reagent. HAuBr4 up (it may be put in a desiccator if hygro- in 2H3P04-1 (Acetic acid) also gives good re- scopic). Examine the residue with a polariz- suits. These tests may be compared with the ing microscope and then test with one of the tw^o that had been selected for the table, reagents suggested below, according to the Various examples of microcrystals of com- identification or indication of the tetraphen- pounds are illustrated in Fig. 1 for NH? , and ylboron test. Fig. 2 for methyl-, ethyl-, dimethyl- and tri- Put a droplet of reagent on a small cover- methylamines. glass, then invert it upon the residue of hy- Test for Aninioiiia with Formalde- drochloride. Examine the crystals, using a hyde. There should be no difficulty in iden- polarizing microscope, with magnifications tifying ammonia with certainty in the fore- of about 100 X and 200 X, observing bire- going procedure, noting the small crystals 47 CHEMICAL .MICROSCOPY Recommended Tests Structure Base given off Recommended reagent Reagent No. H / — N \ H Ammonia HAuBr4 in H3PO4 , (20) 1 H / — N \ CH3 Methylamine HAuBr4 in 2H3P04-lHBr, (24) 2 H / — N \ C2H5 Ethyl amine HAuBr4 in 9H3P04-2H20-15HBr, (26) or HaPtle in diluted H2SO4 , UOO) 3 4 CH3 / — N \ CH3 Dimethylamine lodine-KI reagent B-1 5 CH3 / — N \ C2H5 Methylethylamine lodine-KI reagent B-1 5 CHs +/ — N— CH3 \ CH3 Trimethylamine lodine-KI reagent B-1 5 H / — N CH3 \ / CH \ CH3 Isopropylamine HAuBr4 in H3PO4 , (20) 1 C2H5 / — N \ C2H5 Diethylamine HAuBr4 in H3PO4 , (20) 1 48 NITROGEN-BO.NDKU KADICALS: IDENTIFICATION with tetraphenylboron, the characteristic isotropic crystalUzation of NH4CI, and the characteristic crystals with HAiiBr4 in H3PO1 . However, a test especially for am- % 3:^ '••'is ^•%:#^ 4 Fig. la. Ammonium chloride isotropic crystal- lization (from water or dilute HCl). 66X. monia is desirable, as it is the commonest product. Ammonia condenses very readily with form- aldehyde to form methenamine (hexa- ^,-:^ > X-, Fig. Ic. NH4CI with HAuBr4 in H3PO4 , (20) crystals at periphery of precipitation, with a slightly moist reagent. lOOX. Fig. 2a. Methylamine (from epinei)liriue in Fig. lb. NH4CI deposit with HAuBr4 in volatility test with Na tetraphenylboron (1:20 in H3PO4 , (20). 135X. water), in hanging drop. lOOX. 49 CHEMICAL MICROSCOPY k>'' I -»» * , . \ \ ' i %* . A I '^ Fig. 2b. Ethylamine HCl with 1.3 HAuBr4 in 9H3P04-2H20 15HBr, (24), 135X. ^ ;>.^,- / Fig. 2c. Ethylamine HCl with 1.8 HoPtle in diluted H2SO4 , (100). lOOX. methylenetetramine), (CH2)6N4 . The low organic amines cannot give such a highly- condensed product, and under the conditions specified here will hardly give more than trace reactions nevertheless they will inter- FiG. 2d. Dimethylamine HCl with iodine-KI reagent B-1. Yellow-brown diamonds and hexa- gons, fairly large at periphery of crystallization. Similar crystals are given by methylethylamine HCl, but only dimethylamine gives red hexagons forming among and from the little crystals (2e) . « A Fig. 2e. Dimethylamine HCl with iodine-KI (5:80) in H3PO4 (2:1) (iodine-KI reagent B-1). (direct addition) lOOX. Crystal red plates, form- ing from and among little brownish yellow plates or flakes. 50 OBSERVING IMICKOCRYSTALS ^**t 1. ^ t^^^ i"^ ^.^* ^" ^ ^ V, ■+4» ^ If V'V^ ## # Fig. 2f. Trimethylaniine HCl with iodine-KI (5:80) in H3PO4 (2:1) (iodine-KI reagent B-1). (direct addition), crystals black, opaque. lOOX. fere more or less with the ammonia reaction. The recommendation of the test, therefore, is still for cases where ammonia is the only product coming off, or at least the chief one. Proceed as previously described, but in- stead of fixing the evolved ammonia with dilute HCl, use a hanging drop of neutral, 1% formaldehyde solution (0.1 ml of the usual concentrated formaldehyde, 37 % by weight, diluted to 4 ml with water). After exposure above the alkaline solution, rein- vert the hanging drop and allow it to dry. Methenamine gives branching isotropic crystallization, frequently with tripartite forms. Examine for a trace soon after drying, as it has a slight volatility. In a blank test there is at most a very slight deposit of tri- oxymethylene left from the formaldehyde solution itself, which usually does not show anything definite microscopically, and does not react in the following crystal tests. To confirm methenamine, redissolve the deposit in a little droplet of water, add a droplet of iodine-KI solution (1:1 g in 100 ml), and preferably rein vert o\'er a cavit}^ containing a little crystal of iodine — ^this prevents evaporation of iodine from the test- drop, and the crystals can be examined at relative leisure. Characteristic birofringent crystals form at once. Methenamine can also be confirmed with reagent 3 above, HAuBrj in 9H.3P04-2H.>0- 15HBr. Put a droplet of the reagent on a small cover-glass and invert on the dry resi- due. Allow a short time for the formation of characteristic crystals with a very small amount . Of the lower amines, methylamine gives the most noticeable results, including minute microcrystals, but it could not possibly be confused with ammonia if more than a trace of the latter is concerned, or if the character- istic crystals are obtained and observed. The formaldehyde test for ammonia, al- though confirmed by the usual sort of basic- nitrogen microcrystal tests, in the formation of methenamine illustrates the use of ciuite a different reaction for microscopical chem- istry. Charles C. Fulton OBSERVING MICROCRYSTALS To make full use of microcrystals, the chemist-analyst must learn to see and under- stand all the characteristics that may possi- bly distinguish them. In studjdng a particu- lar test, not every detail descriptive of the crystals needs to be permanently recorded, but everything that can readily be observed should be observed, in deciding what is worth writing down for a formal description, and what should be the points of compari- son to establish identity between known and unknown. Too often those who try such tests content themselves with a very superficial observation of the crystal forms alone. It is surprising how much can be done with such observations, often using mereh^ the ordi- nary microscope; it is not nearly as bad as failing to use the microscope at all, but still 51 CHEMICAL MICROSCOPY a gross neglect of potentialities. The polariz- interference colors mainly of a higher order, ing microscope is necessary, and this means Even with crystals having a deep color of a good instrument, not just "polarizing at- their own the birefringence may be expressed tachments", which are no more than a make- as dim, moderate, or bright. With crystals shift. that are colorless or only lightly colored in A magnification of about 100 X is ordi- themselves the place of an interference color narily used, and about 200 X when a higher in the first order can be specified as precisely power seems advisable or necessary. as the uniformity of the crystals warrants; Looking at crystals through the micro- the order of higher colors can be determined scope, the chief characteristics of form and with the quartz wedge, aggregation immediately catch the eye. (See Note, on more than one crystal, whether "Forms of Microcrystals", p. 37.) Size, not extinction is parallel to a principal side (or measured but rather loosely appreciated rela- to the general direction of an irregular crys- tive to other crystals commonly seen, is also tal), or bisects a principal angle, or is oblique, obvious. Further things to observe will now The possibility of measuring an important be suggested. angle, at least approximately, by using the Note how many dimensions of the crystals revolving stage, should not be forgotten, may be worth measuring. Note whether the This applies both to angles of form and to crystals are fairly uniform, or show diversity the angle of extinction, within one type, or whether there are two Unless the angle of extinction is close to or more distinct kinds. Look over the whole 45°, if the crystals are elongate, or direc- precipitate, if there are many crystals, to tional at all, the sign of elongation is impor- note how some forms develop from others, tant, and finding it usually requires no more Thus disks or shapeless plates when more than pushing in the red plate and observing perfectly formed may be hexagons, or per- the result. With high-order interference col- haps octagons. Squares and hexagons often ors it can usually be determined with the skeletonize into 4- and 6-armed stars; ob- quartz wedge. All crystals can be put into longs and rhomboids into X-shapes. That four groups by the sign of elongation: -1-, the diamond shape is related to the hexagon — , =t, and indeterminate, is frequently evident; also diamonds may Birefringence often shows whether a com- extend into daggers and spears. Often very plex form is all one crystal or an aggregate irregular forms can be related to a fairly of several. X-shaped crystals are usually simple form. Regarding details of descrip- skeletons of a rhomboid or an oblong, along tion, note particularly the ends of blades, the diagonals. The direction of the acute an- rods, prisms, bars: whether square, slanting, gle shows the elongation of the parent form, pointed, incised, ragged. A number of points regardless of distortions of the arms. (Fig L) in regard to form have to be noted in connec- Colored crystals are nearly always due to tion with birefringence. a colored reagent, and to this extent the Usually the next step is to look at the color does not distinguish the substance crystals with crossed nicols. First, note the tested, but the color is more significant the degree of birefringence. It will vary, of course, more it differs from the usual color produced with the thickness of individual crystals, but in crystals by the particular reagent. The very often the crystals of a precipitate will color between crossed nicols is often signifi- appear uniformly of much the same color, a cant. Some crystals are even known which do weak gray, or gray-white, or bright white to not extinguish completely but turn a differ- yellow, in the first order, or, on the other ent color at "extinction" positions, hand, nearly all may show different brilliant Dichroism is common in crystals with 52 OBSERVING MICROCRYSTALS some colored reagents, giving not merely the fact of dichroism to distinguish certain sub- stances, but two different colors to be ob- served and specified. Some crystals (mostly with iodine reagents) are pleochroic, with three extreme colors, and as seen on the slide usually shoAV quite a variety of colors with polarized light, changing with rotation of the stage. Pleochroism is the general term (in- cluding dichroism), but as there are usually only two quite different colors, and moreover an individual crystal as it lies on the slide can only show two extreme colors with rota- tion of the stage, the writer prefers to use the term dichroism when it is applicable. Dichroism can usually be noted merely by rotating the stage, but when it is feeble it may be necessary to test the extinction posi- tions, which show the extreme colors, to ob- serve it. A crystal is turned to extinction position between crossed nicols, then ob- served using only the polarizer or only the analyzer, then observed again in the next extinction position, at right angles. The change in color may be slight, or great; di- chroic microcrystals are known with iodine reagents which change from slight yellowish to black, with bromauric acid which change from pale yellowish (sometimes appearing colorless) to deep bright red, with iodopla- tinic acid which change from pink to dark blue, or from green to purplish red. Note not only the two different colors, but also their orientation. The crystal is said to have a positive sign of absorption when the darker color is "lengthwise", negative when it is "crosswise". The sign of absorp- tion, which also depends on elongation, nearly always agrees with the usual sign of elongation, when both can be observed. Dichroic crystals often show a peculiar quality of Color in ordinary light, and also show the deep dichroic color where they overlap at right angles. Pleochroic (tri- chroic) crystals may be recognizable even in ordinary light. Observe the relation of various forms to Fig. 1. dZ-Amphetamine with HAuCli in (1 + 2) H3PO4 , applied directly to tablet material. Crossed nicols. 66X . The X-crystals show negative elongation. birefringence. Hexagons may or may not show birefringence. When they do not show it, lying flat, there are often interspersed rods that do. Crystals appearing square or four- parted may be isotropic, or only some of them may show birefringence, in this case usually not very strong even when present (interference figures possible). All may show definite birefringence — even quite high — and may extinguish parallel to the crosshairs in some cases, or diagonally with crystals of a different kind. Thus crystals of the same form may be of quite different types when birefringence is also considered. Examine for interference figures — this requires at least a 20 or 21 X objective — when sizable, transparent cr.ystals show only low or no birefringence, especially if in the latter case the crystals show birefringence when tilted, or some of them are more or less birefringent (tilt of the axes in the crystal), or are accompanied by other crys- tals (possibly of the same crystal system, but a different elongation) which are quite 53 CHEMICAL MICROSCOPY birefriiigent. Interference figures, while im- portant in optical crystallography, have been very Httle used in niicroerystal tests. In most cases it would be a waste of time to look for them, as they could not be found. On the other hand as soon as the analyst learns to recognize the kinds of crystals on which it may be possible to find them, they become an added diagnostic characteristic of value, and can often be obtained, even using the 20 or 21 X objective, quite clearly on surprisingly small crystals (e.g., down to about 25 M diameter, in the case mentioned in the previous article). Moreover, the sign of the crystal can then usually be obtained, as distinguished from the sign of elongation, which may or may not be the same, or the crystals may not be elongate. Some kinds of crystals will give only indistinct figures, and for test purposes it is not worth trying to see them. A good example, but one seldom followed, was set by the Behrens-Kley text in stating actual sizes of microcrystals. No doubt merely saying small, medium, or large is often sufficient, and of course in a compari- son one can instantly see whether the con- trol crystals are of the same order of size as the crystals obtained with the sample. Moreover, one must often allow for change of size with various factors; dilution or stir- ring, for example, may cause diminution in size. However, if an ocular micrometer (kept in an extra ocular) has once been cali- brated it certainly does not take long to use it, and the size of fairly uniform crystals under controlled conditions (i.e., in the test as usually made) can be a valuable and measured characteristic. Also the ratio of length to breadth of oblongs, for example, is then measured rather than estimated, and similar proportions for other shapes. These refinements may be reserved for important tests, but ought not to be completely neg- lected. Refractive phenomena have been seldom noted in these tests. Measurement of refrac- tive indices can be used in a few cases with- out invoking special procedures of filtration, purification, recrystallization, etc.; e.g., when a bircfringent acidic substance is pre- cipitated from dilute NaOH solution with HCl, and the drop then allowed to dry up. However, even this involves more than sim- ply observing the crystals in the test-drop in which they form, the topic here. Observation with an ordinary microscope which supplements the polarizing micro- scope is often useful, chiefly as a matter of convenience when it is troublesome to change objectives on the polarizing micro- scope because they are the type sliding on, while those on the ordinary microscope are on a turntable. Sometimes one wants to see what the crystals look like in ordinary light; the appearance of pleochroic or highly di- chroic crystals is sometimes noteworthy. Darkfield observation may also be used. Many kinds of crystals show up remarkably well, but there is not the distinction between crystals and non-crystals usually obtained with polarized light and crossed nicols. The role of darkfield therefore cannot be more than secondary. In fact, no vital distinctions (not seen otherwise) have been observed in this way. At present it is not particularly recommended, since to have a darkfield ready for immediate use still another micro- scope might be reciuired. With a phase micro- scope both ordinary light and darkfield effects may be observed, but phase micros- copy has not proved of value in these crystal examinations. There is one other mode of observation of great advantage in a few cases, namely, use of incident light. This cannot supplant transmitted light, which is far more useful, but can supplement it. A light should be set up beside the microscope, which will throw a beam down on the stage; it is then no trouble at all to shut off the transmitted light and turn on the incident light. Usually the results are negative; that is, the crystals can 54 OPIUM, ORIGIN OF Fig. 2. Dihydromorphinone with HoPtBre in strong H2SO4 , applied directly. These crystals are opaque; photographed by reflected light. be seen much better by transmitted light and show no special phenomenon with inci- dent light. However, there is added diagnos- tic value in crystals that show up well (espe- cially on a dark background) or in an interesting way with incident hght. (Fig. 2.) Certainly not all these means of obser\'ing microcrystals will be used in routine tests, because then the analyst will be told or know from experience what to look for. However, he should know them all, and would do well to try them on new tests and in examining unknown crystals. Charles C. Fulton OPIUM, ORIGIN OF One of the best tests for determining the origin of seized opium is simply examination of a smear between crossed nicols of the polarizing microscope, using a magnification of about 80-lOOX (and higher when de- sired). Certain kinds of opium, notably Indian (Fig. la) and Iranian, are full of well- formed, highly birefringent, rod crystals, while other kinds, notably Turkish (Fig. lb) and Yugoslav, contain much less crystalline material, and that mostly in the form of small shapeless or roundish particles. Often not e\'en a single well-formed rod can be found in a smear of Turkish opium, while Indian opium contains a multitude of rod crystals. 8till other kinds of opium, as Afghan (Fig. Ic), are intermediate between these types. A small amount of the opium is treated with a drop of water and spread out on a slide. The water is primarily just a dispersing agent and the crystals can be seen floating in it. However, the examination is best made after the smear has dried up. With a strong light, the brown amorphous material is fairly transparent in a thin layer when dry. Alkali solution can be used to dissolve most of this other material, leaving the crystals, but generally this is not necessary. The usual sort of examination of a crude drug with the ordinary microscope will dis- close some of the large crystals in Indian or Iranian opium; in fact this distinction from Turkish opium was mentioned by The National Standard Dispensatory (Hare and others) in 1905, and Levine, in examining samples of seized opium for the U. S. Nar- cotics Bureau in 1945, used the crystal rods as one origin test, along with some others, for distinguishing Indian opium. However, the crystals are not at all easy to see in any number with the ordinary mi- croscope, or using only a polarizer, but spring brilliantly into view when the nicols are crossed. The present writer began using the polarizing test in 1947, also for the U. S. Narcotics- Bureau, and show^ed the crystal- line material to be narcotine, in work on methods for determining the origin of seized opium, which was later continued at United Nations, and which finally resulted in the U. N. Narcotics Laboratory now at Geneva. The number and form of the crystals de- pend partly on the content of narcotine and partly on the physicochemical reaction of the other constituents. This one test, of ao CHEMICAL MICROSCOPY (a) (b) (c) Fig. 1. Opium, (a) Indian; (b) Turkish; (c) Afghanistan. course, is not sufficient by itself to prove a particular origin, but in conjunction with various chemical assays and ash analysis it is very useful. After years of study, it is still the easiest origin test to make, and at the same time one of the best. Charles C. Fulton PURPOSE A fundamental error, which seems to be prevalent in modern "microchemistry", is the idea that the microscope has value in chemistry only as an adjunct to procedures on a small scale. Actually this is the least of its uses; and the early chemists, say from 56 QUINOLINE AS A REAGENT QUINOLINE AS A REAGENT 200 to 100 years ago, who are cited for the istry were chiefly observational: noting the beginnings of "microchemistry", were not, constituents in material to be analyzed, and primarily, groping for small-scale procedures making identifications by microscopic ap- when they turned to the microscope. They pearances existing in the material. The last had a much better appreciation of its real use was later extended to identification of value. mineral crystals by their properties in polar- The uses of the microscope in chemistry ized light, and now is gradually being applied comprise at least the following: in chemical science. About a hundred years (1) Taking a closer and better look at ago, identification by means of chemical things, and observing their minute charac- microcrystal tests was developed in some teristics, regardless of the amount of mate- aspects, chiefly inorganic and alkaloidal, but rial available. the possibilities here were grossly neglected (2) In particular, observing characteris- w^hile chemistry "went quantitative", and tics that cannot be seen at all with the un- still await anything like adequate develop- aided eye: for a century and a quarter now ment. this has meant not merely characteristics too small to be seen by the unaided eye, but Charles C. Fulton also those revealed by polarized light and an analyzer, plus compensators and the other fittings of the polarizing microscope. This use also does not depend on whether Quinoline is a useful reagent in chemical much or little material is available. microscopy for detection of a number of cat- (3) Making identifications by microscopic ions and as a group reagent for the elements observations and microcrystal tests, for named below. Pure quinoline produces char- which the microscope is essential, again re- acteristic crystals with the solid chloride of gardless of the amount of material available, any single one of the following: divalent (4) Using the microscope as an adjunct to cobalt, copper, iron, manganese, mercury, procedures on a small scale. nickel, cadmium, calcium, and zinc, and These uses are not mutually exclusive or monovalent copper. When more than one of even distinct; they overlap greatly. There is these chlorides are present, mixed crystal no intention of saying here that the early formation may occur, so additional tests are chemists never thought of the fourth of the needed for confirmation, but the mixed crys- above uses: of course they did, but they had tals formed are still a good indication of primarily in mind the other uses, which which cations are most probably present, modern chemical science seems to have for- When a cation which normally forms a col- gotten, and which most modern chemists ored product with quinoline is involved in disregard or overlook. When the early chem- mixed crystal formation with a cation which ists turned to small-scale procedures it was, normally forms a colorless product, the re- at least as often as not, to adapt them to sultant crystal generally shows the shape of microscopic observation, rather than the the colorless specie and the color of the col- reverse, to adapt the microscope to small- ored specie. WTien two cations which nor- scale chemistry. mally form colored products are involved, All the uses apply both to the materials they may form off -colored crystals; heating to be analyzed or studied, and to the results will usually develop characteristics which re- of chemical reactions, particularly the crys- semble one of the components, tals resulting from chemical precipitations. A drop of sample solution which has been The earliest uses of the microscope in chem- converted to chlorides is evaporated on a 57 CHEMICAL MICROSCOPY slide ; as the slide is removed from the source of heat, a coverglass from which a drop of quinoline is han<>;ing is placed on the solids and when the slide has cooled the prepara- tion is examined under a microscope. If larger, more euhedral crystals are desired, the slide may be gently warmed and then allowed to cool, but the cupric chloride- quinoline compound, if present, should be identified prior to this because its color is destroyed by heating. If the sample is a dry solid, it may be mounted in quinoline to establish the presence of the above-named cations as chlorides. The characteristic crystals formed by quinoline with the metal chlorides are de- scribed below, grouped by color. Yellow crystals indicate cuprous copper or ferrous iron. The copper compound appears as needles or rhomb-shaped plates and tablets; the iron compound appears as pleo- chroic (pale yellow to yellow) rectangular plates. Blue crystals indicate cobalt, nickel, or cupric copper. The cobalt compound forms pleochroic (light blue to blue) rhomb-shaped or rectangular plates and tablets; the nickel compound forms pleochroic (violet to blue) blue-violet plates and tablets; the copper compound forms pleochroic (green to blue to blue-violet) crystals shaped like elongated hexagons or footballs. Colorless, needle-like crystals indicate cal- cium or cadmium; these two are not readily differentiated from each other. Colorless, well-defined crystals indicate manganese, mercury, or zinc. The manganese compound forms small elongated rectangu- lar plates; the mercury compound forms pseudo-hexagonal plates; the zinc compound forms rhomb-shaped plates and tablets. The chlorides of the less commonly en- countered elements indium and thallium (thallous) also react with quinoline to yield colorless crystals. The indium compound forms diamond -shaped and rectangular plates; the thallium compound forms hexag- onal, rhomb-shaped, and rectangular plates. Primarily, (juinoline serves as a group re- agent; a positive test indicates that one (or more) of the above-named metal-chlorides is present and a negative test indicates ab- sence of an apprecial)le ([uantity of any of them; under favorable conditions qviinoline offers a specific test for these cations. When the sample is a solid, ciuinoline may distin- guish these metal chlorides from their oxides, or sulfates, or free metals, and it simultane- ously distinguishes the valence state in the cases of copper chlorides or iron chlorides. J. M. MUTCHLER REAGENTS FOR MICROCRYSTAL IDENTIFICATIONS The reagents given here are primarily precipitant s of organic compounds contain- ing basic nitrogen. They include, but go far beyond, traditional alkaloidal reagents. A few give good inorganic tests for certain ions; a few extend to nitrogenous compounds that are almost completely acidic, such as the barbiturates. Each of the outstanding precipitating compounds, HAuBr4 for example, makes a number of quite different reagents, by the use of different solvent media, particularly: (1) Syrupy H3PO4 ; diluted H3PO4 , (2) Diluted H2SO4 (up to (1 + 1) for chlorides, (2 + 3) for bromides, (1 -f 3) for iodides), (3) Water; and aqueous solutions only slightly acid, or made acid only by the pre- cipitating compound itself; sometimes even neutral, slightly basic, or alkaline, (4) Concentrated HBr (40%); diluted HBr, (5) Concentrated HCl (38%); diluted HCl, (6) Acetic acid (usually diluted at least (2 -|- 1) simply to prevent its spreading all over the slide). These are given above in the order of de- 58 REAGENTS FOR MICROCRYSTAL IDENTIFICATIONS dining effectiveness for precipitation. Crys- tallizing effectiveness varies with the kind of substances tested, and is greatest when a resultant compound is neither too soluble nor excessively insoluble. When insolu- bility is very high, the precipitate is likely to come down amorphous, and crystalliza- tion may not be obtainable. However, the very best tests combine ease of obtaining crystals with a high sensitivity which usu- ally means great insolubility of the product formed. All these media have specific effects dis- tinct from the general solubility effect. Combinations of media are used to get the best possible results in some cases. The reagents may be added to aqueous solutions of the substances tested (the tra- ditional way), usually without a cover glass; or, particularly those with high concentra- tion of acid, direct to a very little of the solid substance, usually with a cover glass added. In the list below, the major reagents of the most widely applicable precipitating com- pounds are given first, then a selection of others, more or less in the order of declining precipitating power, so far as this can be reconciled with a certain listing of related reagents together and related compounds near each other. The list as a whole will be found to bear but little resemblance to lists of traditional ''alkaloidal reagents", al- though the best of the traditional reagents are included. Iodine-Iodide Reagents The precipitating compound, iodine-HI or iodine-KI, or presumably the anion Is", has the widest range of any. However, the con- ditions for precipitation and suitable crys- tals vary greatly depending on the type of substance, and more types of reagents are used than for any other precipitating com- pound. Only those of greatest established value are given here. Inorganic applications certainly exist, at least with reagents made with H3PO4 , but have not been studied by anyone, to the writer's knowledge. With Phosphoric Acid. (1) lodine-HI- H,PO, reagciU. Iodine 0.08 g, HI (57%) 0.5 ml, syrupy H3PO4 (85-88%) 3.5 ml. May be kept in a small rubber-bulb dropping bottle. Remake when it has lost considerable iodine strength. Very wide range of pre- cipitation, but precipitates of many sub- stances remain amorphous or in drops. Used for crystals with various sympathomimetics, aminoacetic acid, etc. (direct addition). (2) Iodine-KI reagent B-1. Mix 2 ml iodine-KI solution (5 g iodine and 80 g KI in water to make 100 ml) with 4 ml syrupy H3PO4. Pour off from any KI that crystal- lizes out. The mixed reagent keeps quite well in an 8-ml rubber-bulb dropping bottle. Used for barbiturates and similar compounds (added to the slightly alkaline aqueous solu- tion); also (added directly to the hydro- chloride) for some of the simplest amines, etc. (3) Iodine-KI reagent M-2. Mix 2 ml I-KI solution (5:30 g in 100 ml) with 3 ml coned HCl and 3 ml syrupy H3PO4 . Used espe- cially for morphine (direct addition). With Acetic Acid. (4) Iodine-KI reagent C-3. Mix 0.4 ml iodine-KI solution (10:10, dissolved in about 15 ml water, then diluted to 100 ml), 2.0 ml glacial acetic acid, 3.2 ml water, 0.4 ml (1 -f 3) H2SO4 . Remake when it loses strength. Used especially for ciuinine, and for other cinchona alkaloids, which form iodosulfate crystals; also for codeine, etc. (direct addition). Wide range of useful crystallization with complex compounds. (5) Iodine-KI reagent 0-1. Mix 4 ml aqueous I-KI reagent No. 2 (1 : 1.75 g in 100 ml) with 2 ml glacial acetic acid. Originally labeled "0-1" because of crystals with a number of the opium alkaloids and their synthetic modifications. Aqueous. (6) Concentrated aqueous iodine reagents. Three of these have already been mentioned when used as stock solutions. They may also be used as reagents added to 59 CHEMICAL MICROSCOPY aqueous solutions, when a high concentra- tion of iodine is wanted, (a) I-KI, 10:10 (g in 100 ml). This is saturated with iodine; often the most useful ratio for crystals, although iodine evaporates from it comparatively readily. Also the more dilute solution (1:1) is chiefly used for complex compounds (such as alkaloids). (b) I-KI, 5 : 14. Used for atropine in trace ; also, added to bicarbonate solution and with added KCl (or other similar cation), for caffeine, theobromine, etc. (c) I-KI, 10:50. Used for colchicine (neu- tral or bicarbonate solution) ; etc. (7) A series of aqueous ^'alkaloidal" iodine reagents. The precise ratio of iodine to iodide has usually been ignored as an important factor, and is sometimes not even stated. However, it is vital. Using 1 g iodine to make 100 ml solution, the following amounts of KI give distinctly different reagents, of declining sensitivity (of precipitation) in some cases, and different crystals in many cases: 1 g; 1.75 g; 2.75 g; 5 g; 10 g; 20 g; 35 g; 50 g. Dissolve the iodine and KI in no more water than necessary and dilute to 100 ml after complete solution. Used for in- numerable precipitations and microcrystals of alkaloids and many related compounds (generally by addition to aqueous neutral or slightly acid solutions). Bromauric Acid Reagents The bromauric acid reagents as a group, and HAuBr4 in H3PO4 , and in concen- trated HCl, particularly, are probably the most useful of all know^n reagents for micro- crystal tests with compounds of basic ni- trogen. 1 g of the commercial "gold chloride" (HAuCU -31120) converts to about 1.3 g HAuBr4 ; dilutions of acid with water are given in terms of volume of concentrated acid -1- volume of water, e.g., (2 + 3)H2S04 ; final volume (from 1 g of the starting ma- terial) in parentheses at the end: these features may be used in specifying a par- ticular reagent precisely, especially in places where the full formula is not immediately in view. (8) HAuBri in H.POa . HAuCl4-3H20 crystals 1 g, HBr (40%) 1.5 ml, H2O 1 ml, syrupy H3PO4 (85-88%) 17.5 ml. Virtually shows the limits of the basic quality of the N atom. Used for all sorts of simple N bases, feeble ones, and those partly acidic; for certain inorganic cations; for oxonium com- pounds; also for sympathomimetic drugs, etc. (direct addition). (9) HAuBri in {3 -\- l)HzPOi , {1 -j- 3)- HzPOa , {2 + 3)H2SOi , water, cone. HCl, (1 + 3)HCl, or {2 -\- 1) acetic acid, etc. HAuCl4-3H20 1 g, HBr (40%) 1.5 ml; one of the media, e.g., (2 -f- 3)H2S04, to make 20 to 30 ml solution. In this concentration especially used for addition to aqueous solu- tions; for direct addition to residues of an alkaloid or similar compound a greater dilu- tion, to (45), or (60), may often be prefer- able. Considering both aqueous and direct uses, HAuBr4 in cone. HCl is perhaps the best single reagent known for alkaloid-type compounds. 1.3 HAuBr4 in (1 -{■ 3)HC1, (45), is used especially for caffeine, theo- bromine, etc. (direct addition). (10) 1.3 HAuBvi in 2HzP0vl{2 -f 3) HiSOi, {90). Dilute 1 part of 1.3 HAuBr4 in (2 + 3)H2S04 , (30)— preceding formula— with 2 parts by volume of syrupy H3PO4 . Used especially for the fully basic relatives of amphetamine, etc. (direct addition to crushed tablet material containing a salt of the base). (11) HAuBrA in H^PO, and HBr. (a) 1.3 HAuBr4 in 2H3P04- lHBr,(24). HAuCl4-3H20 1 g, HBr,(40%) 8 ml, H3PO4 16 ml. Used for methylamine hydrochloride especially; also for ammonium salts, iso- propylamine and diethylamine hydrochlo- rides; etc. (b) 1.3 HAuBr4 in 9H3P04-2H20- 15HBr, (26). HAuCl4-3H20 1 g, HBr(40%) 15 ml, H2O 2 ml, H3PO4 9 ml. Used for ethylamine and methylamine hydrochlorides, etc. 60 REAGENTS FOR MICROCRYSTAL IDENTIFICATIONS Chloraiiric Acid (Gold Chloride) Rea- gents (12) HAuCh in H^PO, , (1 + 2)HzP0i , (1 + l)H2S0i , water, cone. HCl, (1 + 3)HCl, or {2 -{- 1) acetic acid, etc. 1 g HAuCU -31120 in 20 ml (usually) of the appropriate sol- vent; further dilution (60) with H3PO4 , (1 + 1)H2S04 , coned HCl, or (2 + 1) acetic acid may be used for tests of direct addi- tion with easily precipitated compounds. In the traditional procedure, that is, simply in water and applied to aqueous solu- tions, HAuCU is probably the best single reagent known for alkaloid-type compounds. HAuCU in (1 -^ 2)H3P04 is especially used in hanging drop tests for volatile bases and in direct addition to solids (salts of simple bases, etc.) — as well as for addition to aque- ous solutions. No cover-glass is used and water is then allowed to evaporate from the test-drop if necessary for precipitation and crystallization. Bismuth and Platinum Iodide Reagents HsBile and H2Ptl6 are very general precipi- tants, so much so that they are the two pre- cipitating agents most com.monly used — though only in aqueous solution — for devel- oping spots of alkaloid-type compounds in paper chromatography. (At this time the writer does not yet know of any paper-chro- matographer who has used phosphoric acid, or even diluted sulfuric, to increase the gen- erality and sensitivity of these reagents.) Bismuth Iodide Reagents An aqueous (strongly acid) HsBile reagent is a traditional one from the last century; but as commonly made, it soon decomposes in part; then some effects are still due to the HsBile , but others to iodine-KI. Some users age their reagent for a time to obtain consistent results. Amelink even added io- dine crystals to have a mixed reagent from the start. The reagents given here, however, depend upon HsBile alone. For crystal purposes, phosphoric acid has been difficult to use because colored crystals are likely to form due to the reagent itself. The use of Bils and HI to make up the rea- gent has not as yet solved any important problems either, so the formulas given here simply employ a concentrated bismuth nitrate solution. The reagents have a bright orange color and are to be remade when they darken appreciably. This occurs very soon with (1 -f 7)H2S04 , and too rapidly for convenience even with aqueous reagent, if no preservative is used. Sodium hypophos- phite is here introduced as a preservative, which makes the reagents comparatively long-lasting. Cone. Bi(N03)3 soln: Dissolve 50 g bis- muth subnitrate in 70 ml (1 -f 1)HX03 and dilute to 100 ml with w^ater. (13) H^Bile in {1 + 7)H2SOi or in water, etc. (a) HgBile in (1 -f 7)H2S04 . KI 1.25 g, H2O 2.0 ml, (1 + 3)H2S04 2.5 ml, cone. Bi(N03)3 soln 0.5 ml, Na hypophosphite 0.05 g. Mix. Used especially for hanging- drop tests and direct-addition tests for sym- pathomimetics, etc. (b) HsBile (aqueous). Omit the diluted H2SO4 , using simply 4.5 ml water. Used direct for theophylline and related com- pounds; also for addition to aqueous solu- tions (traditional). (0.04 g hypophosphite is enough.) (c) Double strength HsBile (aqueous) may sometimes be preferred. KI 2.25 g, Na hypophosphite 0.05 g, HoO 4.0 ml, cone. Bi(N03)3 soln 1 ml. Mix. Platinic Iodide Reagents 1 g H2PtCl6-6H20, with iodide, makes about 1.8 g H2Ptl6 . (14) HiPth with H^POa , (A and B). Formulas for small dropping bottles. (A) 1.8 IloPtle in (2H + 1)H3P04 , (250), with minimum Nal. Dissolve 0.04 g Nal in 0.5 ml H2O; mix with 1.8 ml H3PO4 , and add 0.2 ml of 1:20 aqueous platinic chloride 61 CHEMICAL MICROSCOPY solution (1 g H2PtCl6-()H.,0 in 20 nil water). Keeps well for only about a week to a month at most; can be made up on half the above small scale. (B) 1.8 HsPtle in (4 + l)H;iP().. , (250), high Nal. Dissolve 0.5 g Nal in 0.:] ml water, add 2.0 ml H3PO4 and 0.2 ml of 1 :20 aqueous platinic chloride solution. Let stand about a day to develop good differentiation from (A). Keeps much better than the pre- ceding, (A); in a few cases a rather old reagent even gives the best results for cer- tain crystals. Both of the above are among indispensable reagents for sympathomimetics, central stimulants, and other compounds which are of simple structure or partly acidic, not too easily precipitated. (15) HiPth in diluted H2SO4 , (100). H.PtCle soln (1:20 in water) 0.8 ml, (1 -f 3) H2SO4 3.2 ml, Nal 0.46 g. Preferably let stand at least 2 or 3 hours (or overnight) before use. Thereafter it slowly deteriorates, but can be used for a long time. The crystals it gives change to some extent with the age of the reagent. Also, the reagent itself de- posits colored crystals as it partially dries, first around the edge of the cover-glass, where the solution is not covered. Used for ethylamine hydrochloride; uses not much explored. (16) H-iPth {acid aqueous). HoPtCle soln (1:20 in water) 4 ml, HCl 1 ml, Nal 1.25 g. If the HCl is omitted, the alkalinity of com- mercial Nal may cause precipitation. Only a little acid would be needed to overcome this, but Amelink indicates generally better re- sults in an acid than in a "neutral" test- drop, anyway. Platinic Bromide Reagents (17) H.PtBr^ in H.POi , (/ + 3)H,P0i , (^ -f 3)HoS0a , water, HBr, {1 -+- 3)HCl, or (2 + 1) acetic acid, etc. H2PtCl6-6H20 1 g [makes about 1.3 g H^PtBrg]; HBr(40%) 2.5 ml, appropriate solvent to make 20 ml (usually). The aqueous reagent (which can be made with NaBr instead of HBr) is excellent although strangely neglected in the past. New uses (hanging drop and direct, for sympathomimetics, etc.) especially concern H.PtBrg in (1 + 3)H:iP()4 . The reagent with syrupy H3PO4 develops some precipitation in the bottle, but the clear supernatant solu- tion can be used (rather than adding enough water to prevent precipitation). (18) HoPtBrCl^ in cone. HCl. If the above formula is used with concentrated HCl for the solvent, even with twice as much HBr (not exceeding the molecular ratio of 1 HBr to 5 HCl), there is evidence — from micro- crystals — that the precipitating compound is HoPtBrCU . The four other possible Br-Cl combinations form in proportioned mixtures of the strong acids. They can be used by direct addition, for the particular effects; as originally worked out on morphine they were diluted to (60) with the strong acids and some water (up to one-fifth) for best results. Much dilution with water tends to give HoPtBre , regardless of more HCl than HBr being present. Platinum and Palladium Chlorides (19) HoPtCh in H,POi , (/ + 3)H,P0, , (1 -f 1)H2S0a, water, (1 + 3)HCl, etc. 1 g H.PtCle-eH.O in 20 ml of the solvent. The aqueous reagent is of course the tra- ditional one and very valuable. The reagent with (1 + 3)H3P04 is particularly used in the hanging drop for d- and rfZ-amphetamine. (20) HoPdCU in H,POi , or water, etc. PdCl2-2H20 1 g; cone. HCl 0.9 ml (in H3PO4) or 0.8 ml (hi water, etc.); H3PO4 , or water, etc. to make 20 ml. Aqueous or partly aqueous reagent may be made wdth NaCl 0.75 g, instead of the HCl (forming NasPdCb). Tetraphenvlboron and Reinecke Salt (21) Na tctraphcnijlboron, 1 g in 20 ml water (not acidified). Used especially as a hanging drop; remarkable for its sensitivity 62 REAGENTS FOR MICROCRYSTAL IDENTIFICATIONS and crystals with ammonia and the lower amines. (22) Reinecke salt, NHiCr{NHs)2{SCN), . A fresh, approximately saturated, aqueous solution is used. Stir a little of the com- pound into about 0.5 ml water at room tem- perature, to provide a few drops for use. Properties of the reagent begin to change within a few hours. Rosenthaler says that it is useless when the ferric salt test for thio- cyanate ion can be obtained (red color). Some very interesting crystals with lower amines have been obtained with a still- effective reagent aged for a number of hours — e.g., overnight — but the writer does not know how to control these changes or sta- bilize any such intermediate stage. Gradu- ally, in two or three days, effectiveness is completely lost. Bromine-Bromide Reagents Although various forms of bromine- bromide reagents should be possible and valuable (as with iodine-iodide), the diffi- culty of keeping any reagent without the bromine evaporating, and its disagreeable character, have so far prevented anything but a limited aqueous use. (23) Br in HBr solution; Br in NaBr solution. HBr(40%) 10 ml, water 90 ml; or NaBr 5 g, water 100 ml. Saturate with bromine. Used for barbiturates, etc., as well as for basic compounds. Other Reagents for Aqueous Tests The following are added to aqueous solu- tions (usually of a salt of the base tested), unless otherwise stated. Complex Oxygen Acids. (24) Phosphori- timgstic acid. Obtainable commercially. 10 g in 100 ml water. Sihcotungstic acid is used similarly. (25) Phosphorimolyhdic acid with HNO?, . 10 g of the commercial phosphomolybdic acid in 90 ml water and 10 ml coned HNO3 . Phosphorimolyhdic acid has especial value as a standard for sensitivity determinations. Formerly the reagent for this purpose was made with only a few drops HNO3 ; but there is probably no sufficient reason to maintain a separate formula for it; the form with 10% HNO3 may be better for crystals as well as being used foi- the follow- ing reagent. (26) Phosphorimolyhdic acid with H^POi . A fresh solution of the commercial phospho- molybdic acid is decolorized immediately by H3PO4 , but the preceding solution with HNO3 , after standing at least 6 weeks, is more stable. To 4 ml of the stabilized yellow phosphorimolyhdic acid add 0.6 ml syrupy H3PO4 and mix. The solution should remain yellow for use (it will keep for a few days). Less sensitive but gives crystals (e.g., with narceine and atropine) which are unobtain- able with the usual phosphorimolyhdic r.cid solution. Platinum and ^lercury Thiocyanates. (27) H^PtiSCN), . H2PtCl6'-6H20 I'g, water 20 ml, NaSCN 0.95 g. This hardly has out- standing importance for microcrystals so far, but probably is the best of the "normal" thiocyanates (i.e., aside from Reinecke salt); and the writer takes this opportunity of correcting a former statement that onh- about a third as much NaSCN need be used : the NaSCN must be sufficient to replace all six CI atoms, or the reagent will precipitate on standing. (28) K2Hg{SCN)i. Dissolve 3 g KSCN in 100 ml water and saturate with Hg(SCN)2 (5 or 6 g recjuired). This is an outstanding reagent for inorganic microcrystals of a num- ber of cations, especially Zn, Cd, Cu, Co, and Au. It precipitates alkaloidal-type com- pounds but its value for this is minor. Mercuric Iodide Reagents. (29) KoHgli. Dissolve 2 g KI in 100 ml water and saturate with Hgl2 (nearly 3 g Hgl2 required). Only occasional value for crystals but much used for general tests of alkaloid-type precipita- tion. (30) Hglo and HCl. Dilute 27 ml cone. HCl to 100 ml with water (makes an actual 6a CHEMICAL MICROSCOPY 10% HCl), then saturate with Hgl2 (does not take much). Used both for aqueous solu- tions and direct addition to the solid, for heroin, etc. (31) Hgh'NaCN and Nal. Dissolve 0.5 g good NaCN in 100 ml water and saturate with Hgl2 (use 4J^^ g). (This is a reagent in itself, and more sensitive, but the reagent with excess Nal is recommended more highly for crystals.) Filter the saturated solution and add 8 g Nal per 100 ml. Crys- tals with codeine, etc. Cadmium and Lead Iodides. (32) K^Cdh . Cdl2 5 g, KI 4.5 g, water 100 ml. (33) Phi 2 in K acetate solution. Dissolve 4 g lead acetate (3H2O) and 30 g potassium acetate in water to make 100 ml, and add glacial acetic acid drop wise just to faint acidity to methyl red (reacts brown instead of yellow); then add 4.5 g KI. Many crystals with both the above. Mercuric Chloride Reagents. (34a) Simple HgCl2 (5%) is commonly used but usually added to solutions containing dilute HCl, or the hydrochloride of the base, so that actually more or less of the following chloro-acid is present. (b) HHgCls . HgCla 5 g, coned HCl 1 ml, water 99 ml. (c) NaHgCls . HgCl2 5 g, NaCl 0.75 g, water 100 ml. (d) HgCl2 & HCl. HgCl2 5 g, cone. HCl 15 ml, H2O 85 ml. This is less sensitive; useful especially for quinine. Ferric Chloride Reagents. (35a) FeCls in HCl; HsFeCle . FeCU-GHzO 10 g, cone. HCl 100 ml. For addition to aqueous solu- tions (cocaine, methadone, etc.). (b) HsFeCle for direct addition to solids: FeCl3-6H20 10 g, coned HCl 17.5 ml, water to make 100 ml. Organic Reagents. (36) Picric acid is outstanding. (a) A saturated aqueous solution (about 1K%). Many crystals, (b) A 0.2% aqueous solution. Picric acid crystals (10%) water) 0.2 g, water 100 ml. Cinchonine is an example of a base yielding crystals far more readily with this dilute reagent than with the satu- rated. This weak solution may also be used for direct additions. (c) Half-saturated Sodium Picrate. Pre- cipitate sodium picrate from saturated picric acid solution with concentrated sodium acetate. Filter, then prepare a saturated solution of sodium picrate. Filter this from excess crystals and dilute with an equal vol- ume of water. Crystals with bases are often obtained more readily than with the satu- rated acid. (37) Other nitro-organic reagents. Styphnic acid (trinitroresorcin) and Trinitrobenzoic acid are used in saturated solutions. Simple Oxygen Acids. (38) CrOz and HCrOzCl. (a) 5 % CrOs . The chloroacid or its salt (following formulas) is more sensitive and better. (b) HCrOsCl. Stock solution of 20 g CrOa in water to make 100 ml (this may be used as a concentrated CrOs reagent in itself). Reagent: 1 ml of foregoing stock solution plus 2 ml water and 1 ml coned HCl. Will keep for some time; slowly darkens. (c) NaCrOsCl. Add 1 g NaCl to 4 ml of the 5 % CrOs . Keeps well. (39) Perchloric acid. 5 % solution of HCIO4 or NaC104 . (40) Permanganic acid oxidizes most alka- loids and other bases, but the stable crystal- line permanganates are quite distinctive. Unfortunately reducing impurities can easily spoil a test. (a) KMn04 2 g in 100 ml water, with a few drops of syrupy H3PO4 . (b) HMn04 in dilute H2SO4 , for direct addition. Use a stock solution of 1 % KMn04 . Reagent: 2 ml stock solution, 1 ml (1 -f 3) H2SO4 . Make up fresh for use. Used for cocaine, meperidine, methadone, etc. Basic Reagents Precipitation of the free base by addition of the reagent to a neutral or slightly acid solution of the salt of an alkaloid, etc. 64 SYMPATHOAll.METICS AND CENTRAL STIMULANTS (41a) 5% NaOH. Sometimes a concen- trated solution is used. (b) 5 % Na3P04 . The alkalinity is about the same as for 5% Na2C03 , -which is more often used, but effervesces when added to an acid solution. (c) 5% K2Cr04 . Precipitation of a chromate of a strong base is possible, but the principal use and value is as a basic group reagent for the weak alkaloids that are quite insoluble in water. It must, of course, be added to solutions that are only slightly acid, or, if strongly acid, the effect is that of K2Cr207 (similar to CrOs). (d) Concentrated K acetate, 30 g in water to make 100 ml. For precipitation and crystals with very weak insoluble bases. Cyanides (42) Gold Cyanide Reagent. Dissolve 1 g HAuCU-SHaO crystals in 20 ml water. Add 0.5 g NaCN a little at a time; if there is im- mediate precipitation add just enough NaCN to redissolve; otherwise 0.5 g should be the right amount. Makes a colorless solution, which sould not turn red litmus blue (if it does, acidif}^ with acetic acid). (43) Platinum Cyanide Reagent. Dissolve 1 g H2PtCl6-6H20 m 18 ml water and add 1.5 g NaCN. Solution may warm up and be- come rather brown; if not, warm a little on the water bath until it just begins to darken, then cool. Cautiously acidify with 2 ml (1 -t- 3) H2SO4 . There is usually a small amount of brown precipitate, apparently due to the reaction going a little too far, in part ; this is removed by filtering either before or after acidifying. The solution is brown but not a dark brown. (44) HiFeiCN)^ with H^POi . A stock so- lution is kept of 10 g K4Fe(CN)6-3H20 in water to make 100 ml. Reagent: Mix 3.5 ml of the stock solution with 0.25 ml syrupy H3PO4 . Has to be made fresh as it will not keep. Simple Halides and Pseudohalides. (45) 5 % solutions of KI, NaSCN, NaNOz . They are also used in concentrated solutions. Charles C. Fulton SYMPATHOMIMETICS AND CENTRAL STIMULANTS Most of these are relatively simple com- pounds, compared with alkaloids (q.v.), and are not so readily precipitated. Some which have alcoholic and phenolic hydroxyl groups are not precipitated at all by the usual aqueous reagents, which are described in the article "Alkaloids and Alkaloidal-type Pre- cipitation". The following 14 tests are se lected for tabulation in Table 1. Five rea- gents are applied directly to a very little of the solid, with addition of a cover-glass : (1) 1.3 HAuBr4 in H3PO4 , (20) (2) 1.3 HAuBr4 in 2H3P04-1(2 + 3)- H2SO4 , (90) (3) 1.8 H2Ptl6 in (2H + 1)H3P04 , (250), with minimum Nal (4) 1.8 H2Ptl6 in (4 -f 1)H3P04 , (250), high Nal (5) Iodine-HI-H3P04 reagent. Three reagents, which contain more water than the five above, are used in two tests each. They are applied to the solid without a cover-glass, and let stand for partial evap- oration; they are also used as the reagent of a hanging drop, in volatility tests above an alkaline solution on a cavity slide. In the latter case the test-slide is in general rein- verted after exposure to the vapor (of up to two hours), and evaporation to a higher content of the non- volatile acid then occurs. (6, 7) HsBile reagent in (1 -f 7)HC2S04; 6 direct application, (7), volatility test (8, 9) HAuCU in (1 + 2)H3P04 , (20) (10, 11) 1.3 HoPtBre in (1 + 3)H3P04 , (20) The two following reagents are also used for volatility tests (the first permitting rein- version, but the second only as a hanging drop) : (12) HaPtCle in (1 + 3)H3P04 , (20) 65 CHEMICAL MICROSCOPY to H Z < H CD a.S < S H to (D PS o ^ -u CO H OJ trl W o H C J "^ «i m « fJ C) ^ IZ T-H H J m ■< H « "5 + >yil 0-3 a m TO . _ ■(-3 to -» I— I CJ o CO + .2 o o to ^^ . d. K%"0 .^ -w -^ ^ 2.2 g ^ 'D (u '•+J '-t^ '-i^ '•+J ' c3 c3 Cw c3 'o'o'o'o > > > > -< ^ t^O O o 0^0^^ O O OOU O-^oO 0J2 0000 OS-O 0-T3 Q o 0) c a, ID tp c3 a S >> X O a) rt O O =3 -►^ a c c tj G o "^ 2-" G"^ =; _o a^H "t: c cS O o o rt • O o N o3 (-• +i fi ^^ (—1 o 0^ r- "T"* • rH c - >> £3 c a OJ G G tami etam meth ?^:J G tp a c s s <» -G a S Methamphe -Methamph he nyl propyl -►J a ;_ T3 -i-i .. s up IV Amph 5 Mephen Propylh c 3 ^ "^o '^ Ph 2^ -TS -e-^P-. H 66 SYMPATHOMIMETICS AND CKNTKAL STIMULANTS (13) Sodium letraphenylboron in water decomposition) in a lianging drop of dilute (1-20) HCl (about 1% by volume of the eoncen- (14) Volatility tests are also made by trated acid) above a drop of alkaline solution catching the volatile amine -(or amine of on a cavity slide, then evaporating the HCl \ m V Ik f / N '-I •> r- * \ >N^ ••. \i ^ t '*«fe » \ 1 4 \ Fig. la. Hydroxvamphetamine with HAuBrj Vt^ o i t? u i • -^u i o tta n • u u^-. /on^" J ■ • .1 /^ w l^ I ^^^- 2a. 1-Ephednne with 1.3 HAiiBrj in in H3r(J4 (20), red spindles (hrst formed), brown ott pr^ i /o i oa tt cfi mew CENTRAL STIMULANTS Fig. 5c. Nylidrin, 1.3 HAuBr4 in 2H3P04-1 (2+ 3) H,804 (90), on standing, with humiditv. lOOX. amphoteric compounds (even when mainly acidic and only feebly basic) in which the strong H3PO4 brings out the basic character; this also occurs with bromauric acid in H3PO4 . Iodine reagents are indispensable but not more so than bromauric acid; the iodine precipitates more often fail to crystal- lize, and different reagent-formulas are needed for them. In general, the precipitates are not neces- sarily crystalline. With complex and defi- nitely basic compounds, such as the alkaloids and antihistamines, the precipitates with HAuBr4 in II3PO4 , even when it is added to an aqueous solution, are far too insoluble for the best results. Such precipitates are usually amorphous or too minutely crystalline to have any value for microscopic crystal tests. On the other hand, compounds that are rela- tively simple but too water-soluble or too feebly basic, or both, to yield precipitates from an aqueous solution with reagents dis- solved in water, will generally yield beautiful crystals with bromauric acid in a medium of phosphoric acid, although occasionally only drops are formed. The crystals show great differences from one compound to another, not only in their forms, but also in colors, birefringence, and dichroism. The reagent is added directly to the dry substance to be tested. Crystals of the bromaurate com- pound are easily distinguished from undis- solved material, or any other crystals that may form, by their color. The reagent also has a general use in de- termining whether any compound capable of this precipitation is present. A little powder from a tablet, for example, may be scattered thinly on a slide, a drop of the reagent and a cover-glass applied, and bromaurate crystals or precipitation looked for under the micro- scope immediately and after standing. In this direct addition there is some danger that a compound with too great bromaurate insolubility may not show up, because the insoluble precipitate may completely cover the surface of the material and prevent further solution. Therefore, some less sensi- tive reagents should also be tried, or the test tried on an ac^ueous or dilute acid solution of the substance, before concluding that no compound of basic nitrogen is present. Precipitates may be due to other kinds of basic substances. These include: (a) Inorganic: all the alkali metals, and magnesium and zinc, in particular, in the form of thcMr salts, as well as ammonium and hvdroxvlnniine. 71 ELECTRON MICROSCOPY (b) Compounds of basic oxygen: i.e., com- pounds of a certain complexity and capable of forming oxonium salts also react (for example coumarin). These are rare by com- parison Avith the compounds of basic nitro- gen, but in a general test it should be remem- bered that some organic bases do exist which do not contain nitrogen. The reagent: Gold chloride crystals (HAuCU-SHaO) 1 g (makes about 1.3 g HAuBr4); HBr (40%) 1.5 ml; H2O 1.0 ml; syrupy (85-88%) H3PO4 to make 20 ml. This may be named in short, HAuBr4 in H3PO4 ; or in more detail as 1.3 HAuBr in syrupy H3PO4, (20). Excellent crystals for identification tests may be obtained with aminoacetic acid, betaine, glutamic acid, urea, acetamide, etc., as well as with many sympathomimetic drugs and other substances. Charles C. Fulton Electron microscopy AEROSOLS CONTAINING RADIOACTIVE PARTICLES The electron microscope is an ideal tool for the analysis of aerosol particles. This per- tains especially to particles with submicronic dimensions down to 0.002 micron (2 X 10"^ cm). Electronoscopic observations provide data on size, shape, aggregation tendencies and population density of the particles. In addition it may be possible to obtain electron diffraction data as a means for chemical identification (1). Correlation of electrono- scopic observations with data obtained by other analytical methods makes possible a complete description of a particular aerosol. The purpose of this article is to describe the electron microscopic appearance of particles from aerosols containing Sr^'^S04 , Ru^'^'Oa , or Pu23902 . Methods Particles from the various aerosols were collected directly on "Formvar" coated elec- tron microscope supporting grids or on mem- brane filters. In the latter case the particles had to be transferred to coated grids before observation in the electron microscope was possible. This was accomplished by a modi- fied Kalmus technique (2). Each aerosol was represented by an appropriate number of specimen grids (minimum of six grids per aerosol) to provide a statistically valid sam- ple of particles. All specimen grids used were preinspected for cleanliness, or pre-shadowed to delineate contaminations. Screens representing the various aerosols were surveyed in the electron microscope and appropriate fields were photographed at magnifications of 2,000X, 6,500X, and/or 10,000 X. The electron micrographs illus- trating this report are photographic enlarge- ments. Size distribution data were obtained by measuring several hundred particles chosen at random on the prints representing the different samples. The particles were all measured in the same direction. Observations and Discussion Description of the Particles. Strontium Sulfate. The particles obtained from aerosols containing Sr^''S04 are characteristically in the form of needles. Tj^pical examples are illustrated in Fig. 1. For the most part the needles occur in clusters. The dark cuboidal or spherical material noted in these micro- graphs may be undissolved membrane filter 72 AEROSOLS CONTAINING RADIOACTIVE PARTICLES Fig. 1. Particles from an aerosol containing Sr*'S04 . 1. Particles transferred from membrane filter to a "Formvar" -coated screen. 80OOX. The broad gray band in the lower left corner is a common contaminant in this type of preparation. 2, 3 and 4. Selected fields illustrating particles collected directlj^ on "Formvar"-coated screens. 25,000X. from which the particles were transferred, or they may be small particles of Pluronics, a dispersing agent present in the aerosol gener- ation suspension. The individual needles are rarely longer than 1 micron, or wider than 0.05 micron. Ruthenium Dioxide. Particles obtained from aerosols containing Ru^''^02 are illus- trated in Fig. 2. The particles are character- istically three-dimensional chain aggregates, each aggregate consisting of many roughly spheroid unit particles. Ruthenium dioxide 73 ELECTHON MICKOSCOl'Y Fig. 2. Particles from an aerosol containing Ru'''^02 . 1. Particles transferred from membrane filter to a Formvar coated screen. 6,375X. The aggregate particles are slightly fused because of exposure to the electron beam in the electron microscope. 2. Three-dimensional chain aggregate particle. 14,450X. This preparation was shadowed with palladium. Negative print. Note the spheroid nature of the particles making up the aggregate. 3. and 4. The same field before and after pro- longed exposure to the electron beam. 6,375X. The arrow in figure 3 points to a large ag- gregate particle. The arrow in figure 4 points to the same particle after it was exposed to the electron beam for 60 seconds. Such behavior is common for ruthenium dioxide par- ticles. The aggregate particle is too large to dissipate the heat evolved liy the impact of the electron beam on the specimen, and the melting point is low enough to cause fusion of the small particles into the resultant sphere. 74 AEKOSOLS CONTAINING RADIOACTIVE PARTICLES Fig. 3 The large particle (a) in Fig. 3 is about 0.65 /x across and about 0.65 ^ high d/m = 0.38 if particle is Pu()2 with a S.G. 11.44. Particle (b) is appro.xi- mately 0.05 m- The aggregate particle (c) consists of a countless number of grains 0.05 fj. and less. (About 4500 X.) Fk,. 1 In Figure 4, the large particle (a) appears to be made up of several cubes and a brick shaped particle. The volume of this particle can be ap- proximated very roughh'. Particles like these are frequently observed. This particle illustrates the discrepancy between actual or real volume and calculated volume based on a single linear meas- urement. The actual volume is roughly guessed to be about 7 >x^. The calculated volume using the dimension shown is about 46 m'- If this particle is sensitive to the electron beam, that is, when exposed to high beam intensity the ag- aggregates meh and fuse to form a single sphere. Large aggregates are more sensitive to the electron V)cam than small aggregates. Plutonium Dioxide. Particles from Pu-'^02 containing aerosols are characteristically- cubic or brick shaped. These are illustrated in Figs. 3, 4, 5, 6, and 7. Unhke Sr^^SO^ and Ru^oeQo which occur predominantly as ag- gregate particles, Pu-^^Oo occurs predomi- nantly as individual non-aggregated parti- FiG. 5 Figure 5 shows a particle that appears to be almost a perfect cube approximately 0.4 /x on a side. Volume is therefore about 0.064 ju'. Particles of this shape are commonly found in these sam- ples. If this particle is PuOs (S.G. 11.44) the dis- integration rate is about 0.09 d/m. The halo sur- rounding the particle is probably the remains of moisture that were associated with the particle. The round globule just outside the halo is an arti- fact — namel}' a bubble of carbon produced in shadow casting. (About 15,000X.) Fig. 4. — Continued represents a PuOo particle (S.G. = 11.44) the actual disintegration rate is somewhere near 10 d/m, whereas on a calculated basis the activity density would lu' aljout 62 d/m. This micrograph also illustrates the variation in jiarticle size en- countered in these samples. Compare particle (a) 3.55 M across with particles labelled (b) 0.05 ix across. (About 4500 X.) 75 ELECTRON MICROSCOPY luu. 6 Figure 6 shows another characteristic shape (brick-shaped) of particles found in this sample. (About 15,000 X.) Fig. 7 In Figure 7 are shown particles that are ob- served occasionally. This particle is surrounded by a halo indicating moisture was associated with this particular particle. The particle is an aggregate of different sized particles standing one on an- other. The variation in width of the particle is manifested by the difference in width of the shadow. The widest portion of this particle is about 0.4 n and its height is about 1.75 y.. (About 15,000X.) cles. Aggregate particles of Pu-^^O-) rarely consist of more than 5 or 6 cubes. The particles shown in Figures 3, 4, 5, 6, and 7 were collected directly on tungsten ox- ide and carbon preshadowed specimen screens. The specimens were shadowed again with chromium at a 30° angle before exam- ination in the electron microscope. Double- shadowed particles therefore would indicate contaminants on the specimen grid prior to exposure to the aerosol. Physical Data. Pertinent information on the physical characteristics of the particles from each of the aerosols is presented in Table 1 and Figure 8. The data and observations presented above are very general in that they were based on measurements of 100 particles rep- resenting a single aerosol. However, analysis of other aerosols containing Sr^''S04 , Ru^"®- O2 , or Pu-^^02 gave similar results. Al- Table 1. Physical Characteristics of Particles from Aerosols Containing Sr^oSO* , RU106O2 , OR Pu^^Oa Aerosol Contains Size Range (m) Mean Size ±S.D. Per Cent 0.5 m or less Remarks Sr'<'S04 0.05 0.38 74 Particles - 1.30 ± 0.22 are nee- dles or needle clusters. RuiosQa <0.05 0.36 75 Particles - 1.30 ± 0.29 are three dimen- sional chain ag- gregates of small spheriod type par- ticles. PU"902 0.05 0.20 99 Particles - 0.60 ± 0.09 are cubic or brick- shaped 76 i I BLOOD [Z~\ PARTICLES FROM AEROSOL WITH Sr^SO^ ^^ PARTICLES FROM AEROSOL WITH Ru'^^Oj {■ PARTICLES FROM AEROSOL WITH Pu^'^O, 1 050 060 070 SIZE IN MICRONS Fig. 8. Particle size distribution though the size range and mean size for the particles from the different aerosols are similar, there are striking differences in shape and aggregation tendencies. REFERENCES 1. KuMAi.MoTOi, "Encyclopedia of Microscopj^," 1961. 2. BORASKY, R., AND Mastel, B., AEC R + D Report, No. H.W. 46722 General Electric Co., Richland, Washington, 1956. 3. Fitzgerald, J. J. and Detwiler, C. G., KAPL-1088, General Electric Co., Schenec- tady, New York, 1954. R. BORASKY BLOOD* This method is designed to involve the least possible technical manipulation of the blood sample both before and after fixation. The sample was centrifuged to concentrate the buffy coat, thereby obtaining minimal contamination by erythrocytes. No antico- agulant or any other foreign substance was added to the blood prior to fixation. About 6-7cc of blood was obtained by venipuncture either (a) by withdrawal with a lOcc syringe fitted with a 20-gauge needle * (Excerpt of fixation, preparation, obtaining of, and microscopy of specimens taken from "Elec- tron Microscopic Atlas of Normal and Leukemic Human Blood") and transference to a lOcc Lusteroid centri- fuge tube (International), precooled to 5-10° C; or (b) by needle drip directly into the tube. The syringe had been previously silicon-coated with Dow-Corning 200, 2 per- cent in ecu , by immersion and baking for 3^-1 hour at 450-550° C. The needle had been coated with 10 percent aqueous Armour Monocote [tris-(2-hydroxyethyldodecyl)- NH4CI] by immersion, draining, and air drying. The ice-cooled sample was then centrifuged at 1500 rpm for 15 minutes at 0° C (relative centrifugal force — 265; Inter- national model PR-2, refrigerated, angle head centrifuge). The buffj- coat was aspi- rated with a silicon-coated pipette and trans- ferred to a glass tube containing 5cc of 1 percent Veronal-buffered (pH — 7.4) OSO4 at 5-10° C. It was fixed for }^-l hour, usually the former. Between the successive steps (3-^-1 horn-) of fixation, dehydration, and methacrylate infiltration, the specimen was centrifuged for 1-lH minutes at 1500 rpm (relative centrifugal force — 385; Claj^-Adams Safeguard centrifuge) in glass tubes (alcohol dissolves Lusteroid!). After each centrifuga- tion the supernatant fluid was decanted, the next fluid added, and the tube manually agitated to produce a suspension. The last methacrylate suspension (6 parts n-butyl, 1 part methyl) was permitted to settle by gravity in 00 gelatin capsules for 3'^-l hour 77 Eosinophil Fig. 1. (a) Cell of normal V)lood It Lymphocyte Fig. 1. (h) Cell of normal blood 78 BLOOD ^:^«r Neutrophil Fig. 1. (c) Cell of normal blood Monocjte I Fig. 2. Cell of normal blood >.- %> 79 ELECTRON MICROSCOPY V " *-vt fit -■• ■Hi .-i^il^^ M- «« J5 Neutrophilic Promyelocyte Fig. 3. (a) Cell of granulocytic leukemia V ^ E:!^^ aitj* Neutrophilic metamyelocyte Fig. 3. (b) Cell of granulocytic leukemia Neutrophilic myelocyte Fig. 3. (c) Cell of granulocytic leukemia to avoid close packing. This also eliminated bubble formation during polymerization, which was performed overnight at 47° C with dry heat in an oven. Sections were cut on a Porter-Blum ultramicrotome using a glass knife and were mounted on copper grids covered by Formvar or carbon mem- branes. Three RCA electron microscopes were used for viewing and photography — an EMU-2, an EML-IB, and an EMU-3. The micrographs were taken on 2 by 10 inch or 3M by 4 inch Kodak lantern-slide medium plates and were printed by projection en- largement. Neither the negatives nor prints were retouched. James A. Freeman BOTANICAL APPLICATIONS Introduction. The high resolution ob- tainable with the electron microscope per- mits it to be used to great advantage for a mmiber of different problems in botanical 80 BOTANICAL APPLICATIONS Fig. 1. Section through part of a pollen wall (acetolyzed and chlorinated) of Rhododendron ponticum, XHflOO. (Afzelius, by courtesy of Grana Palynologica) research. The two main techniques employed have been thin sections and surface repUcas, the former being particularly valuable in plant cytology where internal detail is of interest, and the latter being mainly applica- ble to taxonomic and morphological studies. In addition it is possible to obtain informa- tion in a number of particular cases using direct examination. The various results ob- tained in different fields of botany will be briefly described here. Paly no logy. Pollen morphology has been fairly widely studied in the electron micro- scope. The results may be of interest in fields other than botany, for example, in studying the history of post-glacial flora by means of pollen analysis. In addition it may be of in- terest to those working on such problems as hay fever and asthma. Two specimen preparation techniques, namely sectioning and replicas, provide dif- ferent pictures of the sporoderm. Thin sec- tions provide a great deal of information concerning its stratification, but in order to obtain a complete picture of the sporoderm of a given pollen grain it is desirable to have a knowledge of the surface topography in addition to sub-surface stratification. Earlier work on pollen grains was confined to the study of thin sections in the electron micro- scope (1, 2, 3). It is difficult to prepare sections of pollen cell walls because they are extremely hard. However, advances in the technique of ultra- microtomy have permitted a considerable amount of information to be obtained in this way (4). The stratification of the sporoderm Fig. 2. Shadowed carbon replica of the surface of a fresh pollen grain of Rhododendron ponticum, X7000. is an extremely complex subject and cannot be discussed in detail here. An interesting study of both sections and replicas of pollen grains has been carried out by Miihlethaler (5), who interprets elec- tron micrographs of sections and carbon replicas (6) in terms of existing terminology on pollen morphology. An interesting com- parison between the electron micrographs obtained by different authors of the same type of pollen grain is shown in Figures 1 and 2. Figure 1 shows a section through part of an acetolyzed and chlorinated pollen wall of Rhododendron ponticum (4), taken by Afzelius. This can be compared directly with the carbon replica of Rhododendron ponticum taken by the author and shown in Figure 2. It can be seen that the section shows indica- tions of a surface structure which is similar in character to that revealed clearly in the replica. The potentialities of the electron micro- 81 ELECTRON MICROSCOPY scope compared with the Ught microscope in the study of pollen grains have been discussed by Bradley (7). For example, when studying pores, their outline can generally just be distinguished in the light microscope. With the electron microscope, however, the entire morphology is clearly resolved. A comparison between the pores of Plantago media and Plantago lanceolata is shown here. The pore of P. media (Figure 3) has a ragged outline and contains a number of irregularly scat- tered large protrusions; that of I\ lanceolata is circular and completely different in form. This difference can just be detected in the light microscope, but the true structures cannot be resolved. Surface replicas have indicated that the effect of the acetolyzation process on the sub-microscopic structure of the pollen grain surface is negligible. It might be expected that the use of powerful reagents such as those employed in acetolyzation would pro- duce artefacts in the sporoderm. This is not the case with pollen grains studied in the light microscope and there appears to be no Fig. 3. Shadowed carbon replica of a fresh pol- len grain of Plantago media, X9000. (Courtesy of the New Phytologist) Fig. 4. Shadowed carbon replica of a fresh pol- len grain of Plantago lanceolata, X9000. (Courtesy of the New Phytologist) noticeable effect at electron microscope levels of resolution. An important problem in the study of pol- len grains is the distinction between ap- parently identical grains of different species. If a separation of these species could be ob- tained it would be of considerable value in quaternary research. Preliminary electron microscope studies of Cannabis and Humu- lus, Coryhis and Myrica have not produced the distinction hoped for. In the former case Cannabis and Humulus appeared identical in the electron microscope with regard to their surface structures. However some slight but definite structural distinction was found between Coi'ylvs and Myrica grains. Rowley (8) has used both sectioning and surface replicas in an exhaustive study of the pollen wall in eleven species in the Com- melinaceae. No basic differences were found in the structural elements making up the mature pollen wall; morphological ^•aria- tion at light microscope level was due to variations in the arrangement of these elements. Rowley also studied the develop- 82 BOTANICAL APPLICATIONS ment of the pollen grain of Tradescantia pahidosa. It is of considerable interest that he found that the basic form of exine sculp- turing orginated very early in development. The intine was not recognizable until much later. The entire structure of the sporoderm, both internal and external, can be full}^ elucidated by the judicious employment of replicas and thin sections. Moss Spores. The spores of mosses and similar plants present a similar problem in replication and sectioning to pollen grains. Afzelius, Erdtman and Sjostrand (3) have studied the fine structure of the outer part of the spore wall of Lycopodium davatum using thin sections, the results indicating that the spore wall is divided into two layers, the outer being laminated and the inner granulated. Moss spores have not been studied ex- tensively, the only example being by Bradley (9), who shows electron micrographs of the surface structure of spores of Atrichum undu- latum and Dicranella heteromalla. It seems that the value of studying such specimens in the electron microscope is somewhat limited. Fungi. The direct examination of fungus spores of a number of different species was carried out by Gregory and Nixon (10). The spore structure is of interest in studies of asthma as is the case with pollen grains. Direct electron micrographs only provide a silhouette of the spores and little can be seen of their surface structure; the use of replicas, however, shows the surface structure clearly as in Figure 5. An interesting application of the electron microscope using both surface replicas and sections has been carried out independently by two authors on the division of Saccharo- myces cerevisiae (11, 12). The morphology of the different types of yeast bud scars and the mechanism of the division process was studied by replicas in the case of Bradley (12) and sections in the case of Agar and Douglas (11), both authors independently reaching similar conclusions. Algae. A group of algae which has been studied extensively in the electron micro- scope is that comprising the diatoms. So much work has been carried out that it is impossible to include more than a brief ref- erence. Much of the work was done in the earh^ days of electron microscopj^ and sev- eral important contributions were made by MuUer and Pasewaldt (13), Kolbe and Golz (14), Hustedt (15), and Hendey, Cushing and Ripley (16). The electron microscope is continually being used, generally as a tax- onomic aid in studies of new species or popu- lations. The light microscope is fully adequate for distinguishing the diatom genera, and also for separating the great majority of species. However, the electron microscope permits a much fuller examination of the submicro- scopic details of the silica valve and thus enables a far better understanding of the meaning of its finer visible features to be achie^'ed. In addition, the hope is that it will provide useful information about the de- velopment of this fine structure. The electron microscope has also been used in the study of one family and three genera of algae belonging to the Chryso- FiG. 5. Shadowed carbon replica of a spore of the fungus Russula mairei, X9000. 83 ELECTRON AllCHOSCOI'Y Fig. 6. Shadowed carbon replica of the calcite scales of a species of the family Coccolithophorida- ceae, X 10,000. phyccae. The first of these to be described here, CoccoHthophoridaceae, is of wide interest since the unicellular organisms form minute calcite scales. These scales form sedimentary chalk deposits. They are deposited on the ocean bed after the cells have died and the protoplasm has disintegrated. There is a very large number of species in the family and classification is a difficult matter when the light microscope is used because of the lim- ited resolution available. However, the elec- tron microscope provides the increased reso- lution recjuired for a full taxonomic study and as a result much useful information has been obtained which is of particular value to both botanists and palaeontologists. As in many other cases, electron micros- copy of the scales (coccoliths) requires a detailed terminology which is provided by Halldal and Markali (17). These authors give a comprehensive survey of a large num- ber of species of the genus using direct examination in the electron microscope. Though much information can be gained by studying coccoliths directly, the results are much clearer if carbon replicas are prepared (18). Figure fi shows a carbon replica of the calcite scales and the full surface topography is clearly shown. The remaining two genera studied in de- tail in ihe electron microscope are Synura and MaUomonas. These are closely related to the CoccoHthophoridaceae , bvit the latter are marine organisms whereas the former are fresh-water organisms. Synura and Mal- lomonas are also unicellular and covered with scales, but these arc generally much smaller and are composed of silica. The genus Synura contains only a few species. These have been studied in detail in the electron microscope by Manton (19), Fott (20), Petersen and Hansen (21) and Harris and Bradley (22, 24). The latter two authors concentrated on their taxonomy us- ing the electron microscope, but Manton studied internal morphology and employed thin sections. The genus MaUomonas contains a much larger number of species, many of which have only been discovered recently. The elec- tron microscope has been of great value as a taxonomic aid, since the classification of the genus depends almost entirely on the struc- ture of the minute silica scales covering the organisms. Harris and Bradley (23, 24), also Harris (25), have studied the scales directly and used carbon replicas to show up their fine structure. Asmund used direct examina- tion and has studied the occurrence of Mal- lomonas species in Danish ponds (26). Although the cells of many MaUomonas species disintegrate when dried, most of them become sufficiently rigid after careful fixa- tion to be studied complete in the electron microscope. Figure 7 shows a direct electron micrograph of a scale of a species of Synura and Figure 8 shows a carbon replica of a complete MaUomonas. A small genus, ChrysosphaereUa, about which relatively little is known, has also been studied in the electron microscope. It is rather similar to Synura and only consists of two or three species. The cells are also coated with silica scales and have long spines at- 84 BO r A\ IC AL APPLICATIONS Fig. 7. Direct electron micrograph of scales of Synura eichinulata; the structure at the base of the spine is internal thickening, X 13,500. t ached to them. ChrysosphaereUa is rela- tively uncommon compared with Mallo- monas and Synura. Petersen and Hansen (27) have studied some organisms associated with the surface of the water as opposed to those types of phyto-plankton which are free-swimming. By means of a special technique they were able to study single cells as they were situ- ated on the water surface before drying. Marine algae have been studied in detail by Parke, Manton and Clarke (28, 29) who used direct examination and sections. These authors are concerned mainly with the micro-anatomy and taxonomy of Chryso- chromulina. Their descriptions are extremely detailed and informative. Manton and Clarke (30) have also given a detailed and interesting description of the spermatozoid of Fucus serratus. This inter- esting organism is shown in Figure 9. Man- ton and Clarke ascertained by comparing UV micrographs and electron micrographs that the body shrinks but does not alter its shape in the electron microscope. The func- tion of the fine hairs on either side of the front flagellum is not known. The proboscis, which is highly mobile in the living state, is a fimnel-shaped membrane surrounding the front flagellum and attached to the body at the base. When dry, the organ is flattened and thirteen characteristic concentric thick- enings can be seen. The study of the flagella is particularly interesting since the morphol- ogy can be compared with other flagella and cilia. Electron micrographs of the disinte- grated front flagellum show it to be com- posed of eleven strands; in the rear flagel- lum there are only nine. It is interesting to note that fern cilia bear a close numerical relationship (31) yet there is no phyletic or Fig. 8. Shadowed carbon replica of a complete cell of Mallomonas coronata, X9000. (Courtesy of Research) 85 ELECTRON :\IICKOSCOPY Fig. 9. Direct electron micrograph of a sha- dowed spermatozoid of Fucus serratus; the struc- ture of the proboscis is particularly well shown, X 18,750. {After Manton and Clarke, courtesy of the Annals of Botany) structural relationship with broAvn algae. Eleven strands can also be found in ani- mals, for example, Paramoecium (32), the sperms of domestic fowl and fish. The prevalence of the number eleven suggests some fundamental property in the geometric relations of fibres. Bacteriology. The electron microscope has been used extensively in the study of bacteria and related organisms. The de- velopment of the thin sectioning technique has permitted bacteria to be sectioned and internal structiu'es to be examined in great detail. In general the study of the surfaces of bacteria using replica techniques is not particularly rewarding, but in a few cases it has been possible to obtain useful informa- tion in this way. The study of bacterial cytology is a complicated and controversial subject which cannot be described in any detail here. Much of the controversy arises from interpretations of electron micrographs of thin sections of bacteria. The appearance of sections of bacteria using different types of embedding and staining techniques is always at variance. The recent development of the use of epoxy resins for embedding specimens prior to sectioning (33) has indi- cated that some of the previous work using methacrylate embedding materials is sus- pect. There is no doubt that the wealth of information now available on this subject is becoming much more co-ordinated. The use of surface replicas in bacteriology has generally been connected with taxonomic studies such as in the case of the genus Bacillus (34). Here it was shown that the surface sculpturing of spores was different in different species, so that once again the electron microscope proved to be a valuable taxonomic aid. Plant Cytology. Plant cytology now covers a wide field. The electron micro- scope has been used in the study of cell walls, mitochondria, chromosomes and other cyto- plasmic inclusions. The specimen techniques required are variable, much of the work being carried out using thin sections, but direct examination and replicas of cell walls have provided much information on their structure. Detailed general reviews of the electron microscopy of the plant cell have been given by Miihlethaler (35) and Buvat (36) and some of the more important findings are described here. The Cytoplasm. The structure of the cytoplasm varies according to the fixing agent and may appear as granular or in the form of a fine network. It seems probable that the reticulate structure does not corre- spond to the living state. Studies of the distribution of the various albumens and nucleic acids in the cytoplasm have been attempted by forming heavy metal complexes to act as specific electron stains (37), Strugger (38) combined OSO4 fixation with uranyl-acetate treatment and 86 BOTANICAL APPLICATIONS was able to resolve sub-microscopic filament- on their development (39, 40). The various like elements. stages in the development of the barley Plastids. These have received a great deal chloroplast are shown in Figure 10. Firstly of attention, much work being carried out (1) a proplastid develops in the leaf meri- FiG. 10. A diagrammatic representation of the barley chloroplast (see text). (After Diter von Wettstein, courtesy of Hereditas) 87 ELECTRON MK.HOSCOPY Fig. 11. Section through a proplastid from Be- gonia, X25,000. (By K. Muhlethaler) Fig. 12. Section through a further developed proplastid from Begonia, X 25,000. {By K. Muhle- thaler) Fig. 13. Section through a chloroplast of Elo- dea canadensis, X 16,000. (By K. Muhlethaler) Stem. This shows more or less the same struc- lurc as the mitochondria. Next (2) the size iiicroasos and the plastid center, containing a tul)ular structure, is formed. Starch grains now appear (3), and then the material for the formation of the layer structure pro- trudes radially from the center (4). The starch breaks down (5) and the lamellae l)egin to form. They multiply by thickening and splitting until the continuous lamellar structure of the chloroplast (G) is formed by the fusion of short lengths. The resulting plastid is traversed by continuous double lamellae (7). Finally further splitting forms the grana of the fully differentiated chloro- plast (8). It is interesting to compare Figure 10 with the electron micrographs of Miih- lethaler showing plastid development in Begonia and the chloroplast of Elodea canadensis (Figures 11-13). The proplastid from Begonia (Figure 11) is generally similar to the drawing of the barley proplastid (Figure 10) and the further development of the Begonia plastid (Figure 12) is like stage (4) of the description. The fully-differ- entiated chloroplast of Elodea canadensis (Figure 13) is also generally similar to that of barley. From this it may be inferred that the general pattern of chloroplast develop- ment is similar in different species. It is not possible to study the structure of the chloroplast in detail here. The structure is generally similar in all plants, but there is much variation in the dimensions and spacings of the lamellae. The chloroplast is the site of photosynthesis, the chlorophyll being concentrated in the grana. The grana are distributed in the stroma which forms the body of the chloroplast. The Nucleus. Sections through the nu- cleus show chromosomes in the despiralized condition and with good resolution a fine granular structure can be detected in the chromosomes and nuclear cytoplasm. How- ever, the electron microscope has added little to our knowledge of chromosomes. I 88 BOTANICAL APPLICATIONS Mitochondria. The electron microscope shows that the mitochondria are funda- mentally different morphologically from other cell particles. As in the case of animal mitochondria, those of plant cells show a double-membrane system. Within the mito- chondrion are complex fine structures, the cristae mitochondriales, which consist of in- FiG. 14. Shadowed portion of primary cell wall Valonia, X5500. (After Steward and Miihlethaler, courtesy of the Annals of Botany) vaginations of the mitochondrial membrane reaching from one wall to the other. These structures are highly variable from cell to cell and between different species of plants. It is now known with reasonable certainty that the mitochondrion is the site of respira- tory activity of the plant. The Cell Wall. Many studies of the pri- mary cell wall have been obtained by macer- ating the cells and reducing them to a very thin film. These thin films are in the form of sub-microscopic strands of cellulose which in turn can be split into even finer threads by means of such techniques as ultrasonic irra- diation. The micro-fibrils obtained in this way are elementary in nature because they correspond to the crystalline micellar strands which can be detected by means of x-ray diffraction. It can be seen in Figure 14 that the micro-fibrils of the primary cell wall are interwoven to form a very dense network. In the secondary cell wall Avhich consists entirely of micro-fibrils of cellulose, the strands tend to lie parallel instead of in the form of a network as shown in Figure 15. In successive lamellae the orientation of the parallel fibrils is shifted. Fig. 15. Shadowed portion of secondary cell wall of Valonia, X11,0()0. {After Steward and Miihlethaler, courtesy of the Annals of Botany) 89 ELECTRON MICROSCOPY Fig. 16. Pit membrane in a simple pit in the radicle of maize, X 15,000. (After Milhlethaler, courtesy of Die Naturwissenschaften) It is possible to study changes in the pri- mary cell wall structure during cell division. The manner in which the cellulose micro- fibrils are deposited can be examined. The secondary cell wall contains characteristic perforations known as pits. These pits have been studied extensively in the electron microscope and whereas optical evidence suggested that the pit membrane, which stretches across the perforation, acts as an impassable barrier, the electron microscope indicates that it is porous in nature (Figure 16). Cell wall growth has been studied in detail in the onion root tip by Scott et al. (40). Conclusion. The electron microscope is clearly very valuable in botanical research. It is extremely useful as a taxonomic aid, and much morphological information has been obtained in the field of plant cytology. It remains for these results to be correlated with biochemical investigations. Acknoivledgments. The author would like to thank the following for material and advice: Professor and Mrs. T. M. Harris, University of Reading; Dr. K. Miihlethaler, Eidgenossische Technische Hochschule, Zurich; and Dr. B. E. Juniper, University of Oxford; also Dr. T. E. Allibone, F.R.S., Director of the Research Lab- oratory, for permission to publish this article. REFERENCES 1. Fernandez-Moran, H. and Dahl, A. O., Science (Lancaster, Pa.) 116, 465 (1952). 2. MiJHLETHALER, K., Mikroskopie, 8, 103 (1953). 3. Afzelius, B. M., Erdtman, G., and Sjos- TRAND, F. S., Sv. Bat. Tidskr., 48, 155 (1954). 4. Afzelius, B. M., Grana Palynologica, 1, 22 (1956). 5. MtJHLETHALER, K., Platita, 46, 1 (1955). 6. Bradley, D. E., Brit. J. Appl. Phys., 5, 96 (1954). 7. Bradley, D. E., New Phytol., 57, 226 (1958). 8. Rowley, J. R., Grana Palynologica, 2, 3 (1959). 9. Bradley, D. E., Mikroskopie, 13, 180 (1958). 10. Gregory, P. H. and Nixon, H. L., Trans. Brit. Mycol. Soc. 33, 359 (1950). 11. Agar, H. D. and Douglas, H. C., J. Bac- teriol., 70, 427 (1955). 12. Bradley, D. E., J. Roy. Micros. Soc, 75, 254 (1956). 13. MtJLLER, H. O. AND Pasewaldt, C. W. a., Natunviss., 30 (1942). 14. KoLBE, R. W. AND GoLZ, E., Ber. dtsch. hot. Ges., 61, 91 (1943). 15. HusTEDT, F., Arch, hydrobiol. Plankt., 41, 315 (1945). 16. Hendey, N. I., Gushing, D. H. and Ripley, G. W., /. Roy. Micros. Soc, 74, 22 (1954). 17. Halldal, p. and Markali, J., Avhandlinger Utgitt av Det Norske Videnskeps-Akademi, Oslo. Mat-Natruv. Klasse, No. 1 (1955). 18. Bradley, D. E., J. Appl. Phys., 27, 1399 (1956). 90 CELL ULTHVSniUCTLRE I\ MAMMALS 19. Manton, I., Proc. Leeds Phil. Soc, 6, 30G (1955). 20. FoTT, B. AND LuDviK, J., Prcslia, 29, 5 (1957). 21. Petersen, J. B. and Hansen, J. B., Biol. Medd. Dan. Vid. Selsk., 23, 1 (1956). 22. Harris, K. and Bradley, D. E., Discovery, 17, 329 (1956). 23. Harris, K. and Bradley, D. E., J. Roy. Micros. Soc, 76, 37 (1957). 24. Harris, K. and Bradley, D. E., ./. gen. Microbiol., 18, 71 (1958). 25. Harris, K., J. gen. Microbiol., 19, 55 (1958). 26. Asmund, B., Dansk. Bat. Arkiv., 18, 7 (1959). 27. Petersen, J. B. and Hansen, J. B., Saertyk. Af. Bot. Tid., 54,93 (1958). 28. Parke, M., Manton, I. and Clarke, B., /. Mar. Biol. Assn. U.K., 35, 387 (1956). 29. Parke, M., Manton, I. and Clarke, B., /. Mar. Biol. Assn. U.K., 37, 209 (1958). 30. Manton, I. and Clarke, B., Ann. Bot., 15, 461 (1951). 31. Manton, I. and Clarke, B., /. Exp. Bot., ii, 125 (1951). 32. Jackus, M. a. and Hall, C. E., Biol. Bull., 91, 141 (1946). 33. Glauert, a. M., Nature, 178, 803 (1956). 34. Bradley, D. E. and Franklin, J. G., J. Bac- terial., 76, 618 (1958). 35. Muhlethaler, K., Naturwiss., 44, 204 (1957). 36. Buvat, R., Ann. des Sci. Nat. Bot., 11th Series, 19, 121 (1958). 37. Lamb, W. G. P., Stuart-Webb, J., Bell, J. L. G., BovEY, R., and Danielli, J. F., Exp. Cell Res., 4, 159 (1953). 38. Strugger, S., Naturwiss., 43, 357 (1956). 39. Muhlethaler, K., Private communication. 40. DiTER von Wettstein, Hereditas, 43, 303 (1957). 41. Scott, F. M., Hanmer, K. C, Baker, E. and Bowler, E., Amer. J. Bot., 43, 313 (1956). D. E. Bradley CELL ULTRASTRUCTURE IN MAMMALS The term "ultrastructure" refers to the fact that the cell structures accounted for here have been analyzed with the aid of the electron microscope. The finer details of cell structures cannot be resolved with the light or phase contrast microscopes and, there- fore, have been considered as being beyond or ultra this level of resolution. Methods The techniques applied in modern electron microscopy are found elsewhere (electron microscopj^: specimen preparations). It should be emphasized, however, that thin sectioning (section thickness about lOOA) is required to permit the best resolution. Several specially designed microtomes are commercially available such as the Porter- Blum (U. S. A.), the Sjostrand (LKB, Sweden), the Sitte (Reichert, Austria), the Moran (Leitz, Germany), the Hanstra (Philips, Holland). The microtome preferred by the author is the LKB Ultrotome (Swe- den) which is the most recent and most superior in design. The preservation of the tissue has been thoroughly worked out by Dr. Palade (U. S. A.) and Dr. Sjostrand (Sweden). It is believed that any analysis of the cell ultra- structure today can be made without the risk of describing artifacts because so much is known about how various factors may influ- ence the cell structures: the tonicity and acidity of the fixative, the temperature, post mortem changes, etc. The contrast of the electron microscopic picture (electron micrograph) can be en- hanced by applying special staining tech- niques, either during the fixation or dehy- dration periods or after sectioning. The fixative itself gives, however, cjuite sufficient contrast for most studies. The most com- monly used is osmium tetroxide (often called osmic acid, although it is not an acid). After fixation, the specimen is dehydrated in graded alcohols and subsequently embedded in liquid plastics (usually a mixture of butyl and methyl methacrylates). The monomers are polymerized and the embedded tissue block is then ready to be sectioned. Cell Shape The general shape of cells varies from tis- sue to tissue (Fig. 1). This has been ac- curately analyzed with the light microscope and very little has been added to previous 91 ELECTROIV MICROSCOPY Fig. 1. Plasma cell in loose connective tissue of the human epididymis with most of the features present which usually are found in a mammalian cell: nucleus (N), nucleohis (Ne), Golgi zone (G), mitochondria (M), ergastoplasm (E). The cell is freely suspended in between the connective fibers and displays a number of surface extensions (S). Magnification 10,500X descriptions by applying electron micros- ruled out, as for example in the epidermis copy. However, the relationship between and in the heart muscle. Only in rapidly cells has been elucidated clearly. The old dividing cells, as for instance in the testis, conception of cells being connected by inter- can one find two or more cells interconnected cellular bridges to a syncytium has been at a certain stage of their differentiation. 92 CELL ULTKASTRUCTURE L\ MAMMALS Nucleus ranged in a layer outside the proteins. Also As a rule, there is only one nucleus in each ^ mosaic arrangement of tlic lipid and pro- cell. The striated muscle cell is an exception ^^"^ molecules has been suggested. The ap- and may contain several nuclei. A double- P'^'"^'"^ uncertainty is explained by the fact contoured membrane surrounds the nucleo- ^^'^^ ^'^^y ^^^^^^ ^^ known about what struc- plasm with a total thickness of about 250A ^^"'^ ^^ stained most intensely with osmium (Fig. 3). Discontinuities have been demon- tetroxido the proteins or the lipids, strated in the nuclear membrane, reminiscent ^*'*^^ Surface. The plasma membrane on of pores. There does not seem to be a free ^^^ ^^'^® surface of cells shows four types of communication between the nucleoplasm differentiation— microvillus, brush border and the cell cytoplasm, however, because extension, stereocilium, and cilimu. the "pores" appear to be plugged by a dense Microvilli. The microvillus is essentially a substance of unknown nature. The structure ^^^/'^ ^"^^ ^'^"^ projection of the cytoplasm of the nucleoplasm or chromatin is finely '^^'hich is covered by the plasma membrane, granulated. The granules have a diameter of ^^^ micro\'illus does not contain any pecu- about 250A and are clustered in a zone ^^^^' structures but a slightly dense ground adjacent to the nuclear membrane but can, substance. The free surface of most epithelial in addition, be seen distributed evenly cells does display a varying number of micro- throughout the nucleoplasm. The nucleolus ^^^ ^^'^^^ ^ g^eat variation in length and represents a heavy aggregation of these thickness (Fig. 4). Supposedly, the microvilli granules (Fig. 1). Very little has been done ^^^^ resorptive functions and certainly so far on the ultrastructure of the chromo- contribute to the increase of the cell surface, somes. Brush border extensions. The brush border extensions are longer than the microvilli, Centriole mostly thicker and occur in greater abun- The cell center or the centriole (centro- dance. They all are of the same length and .some) is located in the neighborhood of the are found on the surface of the intestinal nucleus, mostly within the Golgi zone of the epithelial cells (Fig. 5) and on the proximal cell. The centriole is a round or slightly convoluted tubule cells of the nephron (cross elongated body with size and structure es- reference: kidney ultrastructure). Similar sentially similar to the basal body of the structures (Fig. 8) are also seen in the ef- cilium (cross reference: ciliated epithelia) ferent ducts of the testis (cross reference: which represents a dense cortex and a lighter ciliated epithelia ultrastructure). Although core. The cortex is composed of nine paired essentially representing extensions of the filaments and a matrix which embeds the apical cell cytoplasm, the brush border ex- filaments. The core is structureless (Fig. 2). tensions do contain some fine and dense _, ^w , striations oriented longitudinally, sometimes Flasnia Membrane t. a- i ^^i "^i ^i extendnig down nito the apical cytoplasm The plasma membrane is the outermost below the level of their bases. Histochemical limit of the cell. It has an average thickness tests seem to prove that the enzyme alkaline of 70-100 A and stands out as a single dense phosphatase is associated with the brush line in the electron micrograph. It is com- border extensions. posed of lipid and protein molecules, their Stereocilia. Stereocilia are longer and mutual arrangement remaining unknown. It narrower than the brush border extensions, has been assumed that the dense line seen in but the lumibcr of stereocilia on each cell is the electron micrograph may represent the about identical with that of the brush border protein layer. The lipids would then be ar- extensions. Each stereocilium contains three 93 ELECTRON MiCHOSCOl'V W^' lA ms M 1 '» 4P ^r^*^ Gr Go \ •; .» CS' ur II * »n^.. Fig. 2. Detail of a plasma cell (human epididymis). The nucleus (N) is enveloped bj' a triple-layered membrane. The Golgi Zone (Go) has both lamellar and vesicular components. In addition, dense large granules (Gr) are seen, each surromided by a single membrane. In the center of the Golgi area is the centriole (CS) cut at an angle through its fibrillar components. Above the mitochondrion (M) is the ergastoplasm (granular endoplasmic reticulum) E, with its membrane bound flat cisternae and at- tached RNA granules. At ms the cisternae are more spherical. This shape is predom- inant after cell fractionation and centrifugation. These rounded structures then cor- respond to the microsome -fraction of the cell. The cell border is seen at P with some collagenous fibers in the interstitial space. Magnification 33,000X 94 CELL ULTKASTKUCTIRE L\ MAMMALS Fig. 3. Detail of a proximal tubular cell of the mouse kidney. A number of struc- tures are seen which can be encountered in most mammalian cells. The nucleus (N) with its triple-layered envelop. Two mitochondria are present, one sectioned longi- tudinally (MI), the other cross cut (M2). In addition to several microbodies (m), Golgi membranes and vesicles (Go) a dense large body (D), may be seen. Magnifi- cation 67, OOOX to five longitudinal fine filaments. In man, mediate stage between the ordinary brush the stereocilia are found in the duct of the border extensions and the ciha. epididymis (Fig. 6). Their function is not Cilia. The ciha are extremely long ex- clear because they do not have any con- tensions of the apical cell cytoplasm. They tractibility. They seem to form an inter- are coarser than the stereocilia and display 95 ELECTRON IVIICROSCOPY \ ..J^ l.u w I Fig. 4. Typical arrangement of microvilli (V) on the surface of an epithelial cell in the distal tubule of the mouse kidney. The microvilli are slender, short processes with seemingly poor rigidity, usually widely spaced. Mitochondria (M) are seen to- gether with a few small vesicles in the apical cytoplasm. Magnification 15,000X L 1,0 M Fig. 5. Brush border extensions of the mouse intestinal mucosa. Brush borders are closely packed, rather rigid surface processes. They display a higher density than the microvilli. Some mitochondria (M) and several absorbed lipid droplets (L) are seen. Magnification 28,000X 96 CELL ULTKASriU C: Tl KE L\ >L\MMALS 1% "^^ Fig. 6. Stereucilia on the surhice of epithelial ceils in the human ductus epididy- mis. The stereocilia are long and slender and lack basal bodies. They do not seem to have any mobility, but a few basal rootlets can be resolved. Magnification 10,800X a peculiar inner structure of fine longitudinal are all joined in the tip of the cilium and in filaments which are arranged in nine pairs the basal body below the cell surface (cross peripherally and two single in the center reference: ciliated epithelia ultrastructure). (Fig. 7). The filaments are contractile and Tubular invaginations. The ireeceWsuriace Fig. 7. Cilia on the surface of the cells of the rat trachea. Cilia are shorter and coarser than stereocilia and display distinct fine inner structures which terminate in the basal body (B). Magnification 31,500X 97 ELECTRON MICROSCOPY displays small tubular invaginations which membrane close to the free surface of all penetrate about half a micron into the cell epithelial cells (Fig. 8). In a sense, they are (Fig. 8). They are found mostly in cells with reminiscent of the coopering bands of a a high rate of fluid uptake. It has been as- barrel, although located inside the barrel, sumed that they represent another means by The terminal bar of one cell is always located which fluid and substances can be taken in opposite a similar strvicture in a neighbor- by a cell without first penetrating the plasma ing cell. Furthermore, the intervening space membrane. Their activity has been com- between the cells is filled with a denser struc- pared with the uptake of water by an ameba ture and is smaller than usually recorded in (pinocytosis). The word "micropinocytosis" other places, undoubtedly indicating the (or membrane flow) has been suggested, firmness by which the cells are attached, indicating that once the fluid or substance Desmosomes. The desmosomes are in a has been taken in by the microtubules, it section having almost the same appearance can be entirely surrounded by the tubular as the terminal bars. However, they repre- membrane, forming a vesicle which can be sent only local points of attachment and are transported elsewhere in the cell (cross actuafly paired button-like structures with reference: kidney ultrastructure, ciliated one ''button" attached to the intracellular epithelia ultrastructure). aspect of the plasma membrane of either Cell Border. The plasma membrane cell (Fig. 8). In addition, the intercellular which faces a neighboring cell is called cell space between the two "buttons" is larger border. than in other places and is occupied by a Lateral inter digitations. This portion of the substance of high density which frequently plasma membrane frequently displays small contains even denser structures of lamellated lateral projections of the cell cytoplasm nature. The desmosomes are most abundant which penetrate into the cell in juxtaposi- in the cells of the epidermis (Fig. 14) but are tion (Fig. 14). In some instances, the pro- also found in other epithelial cells. Struc- jections become quite numerous, as is the tures of identical appearance are also demon- case between the cells of the ciliary epithe- strated in striated muscle and in the specific hum of the eye or in the intestine. It was first tissue of the heart (the impulse conducting believed that the lateral interdigitations system). Besides their adhesive function in helped in maintaining cell cohesion; how- these tissues, they probably also serve as ever, it has been suggested recently that they points of less resistance across which the rather are the result of a certain compression impulse of contraction can travel with much or expansion during different functional higher speed than elsewhere, stages of the cells, much like what happens Basal Surface. The basal plasma mem- to the bellows of an accordion. The varia- brane which faces the basement membrane tion of volume does not occur in the cell is from time to time elaborately infolded, itself to any large extent, but rather to the Epithelial cells. This is particularly the intercellular space, which is expanded by case in the cells of the nephron, of the ciliary fluid from time to time. Cell cohesion is, on epithelium of the eye, and of the choroid the other hand, definitely accomplished by plexus of the ventricles of the brain. The two peculiar structures which are closely infoldings sometimes reach to the depth of related to the cell border — the terminal bars half the cell and may also be seen inter- and the desmosomes. digitating laterally with each other. Un- Terminal bars. The terminal bars are ring- doubtedly, they serve to increase the basal like reinforcements of the cell border at- surface of the plasma membrane upon which tached to the inner aspect of the plasma enzymatic activity can more readily occur. 98 CELL LLTKASTRUCTURE L\ MAMMALS Jk. .Xj Fig. 8. Detail of epithelial cells of the human ductus efferens (connection between the testis and the epididymis). The surface of the cell's is provided with brush border extensions (B) extending into the lumen (L) of the duct. Tubular invaginations (T) of the surface membrane descend into the upper part of the cells. The cell boundaries (CB) are held together by terminal bars (Tb) close to the surface, and by desmosomes (De) at lower levels. Except for mitochondria (M), these cells display large dense granules (D) as well as extremely osmiophilic bodies (P) , some of which may represent pigments, others lipid granules. A large vacuole (Va) is seen in the center. Magnifi- cation 26,000X 99 ELECTRON MICROSCOPY Nerve cells. The extremely elongated cyto- plasmic extension of a nerve cell is called the axon. The axon is covered by a number of Schwann cells along its course. It is the cytoplasm of the Schwann cell which wraps itself around the axon to form the myelin sheath. The axon and the myelin sheath to- gether represent the nerve fiber. The main component of the myelin sheath is the plasma membrane of the Schwann cell which builds up the myelin sheath in a varying number of layers. It is, therefore, justifiable to consider the myelin sheath as being a specialized infolding of the plasma membrane of the Schwann cell. Basement IVIembrane The basement membrane forms the struc- ture upon wliich most cells rest. It is a structureless layer with a thickness varying between 400 and 1000 A (Fig. 13). Accord- ing to histochemical tests, it does contam mucopolysaccharides but so far no peculiar ultrastructure has been detected in its homogeneous layer. The old conception of the basement membrane being a layer with a thickness of several microns in some tissues has been ruled out with the aid of the elec- tron microscope. The thick basement mem- branes seen in the light microscope contain, in addition to the just described homogene- ous layer, a number of fibrillar structures most of which are reticular fibers (Fig. 10) with some additional collagenous ones. It is not quite clear what kind of cell is responsible for the formation of the homogeneous base- ment membrane seen in the electron micro- scope. In some instances, it is surely laid down by fibroblasts and it should, therefore, be looked upon as being part of the connec- tive tissue. However, sometimes no fibro- blasts can be detected in adult tissue in connection with the basement membrane and it is, therefore, beheved that other cells may have the ability of laying down this struc- ture during their differentiation. Cell Organelles In the cell is recorded a number of or- ganelles some of which are large and have a definite form— rmioc/iondna, microhodies, large granules, and pigments. Other or- ganelles have more flexible form— Go/gi apparatus, vesicles, and ergastoplasm (rough surfaced endoplasmic reticulum). Within the homogeneous ground substance of the cytoplasm, as it was looked upon by means of light microscopy, structures have been detected with the electron microscope which are of rather small diameters, but which definitely should be listed among the cell organelles— /?A^A-6franw?es and glycogen gran- ^Jl PS Mitochondria. The mitochondria are dis- crete bodies within the cell. They may vary in number, size, and shape, but their ultra- structure is remarkably unchanged from cell to cell and from tissue to tissue. In the fight microscope, they can be selectively stained by Janus Green B, but even so, it is sometimes difficult to distinguish them from other granular structures in the cell. The mitochondria are surrounded by a double-contoured membrane, the thickness of which is in the neighborhood of 180 A. The matrix of each mitochondrion has a higher density than the surrounding cyto- plasm. The matrix itself is homogeneous or slightly granulated. It is traversed by a vary- ing number of double-contoured membranes (or cristae) which mostly are arranged parallel to each other. The inner mito- chondrial membranes (or plates) with a thickness of roughly 150A are always con- nected with the mitochondrial capsule for some distance, but there is no open connec- tion between the cytoplasm of the cell and the mitochondrial matrix (Fig. 3). Extremely electron-dense spherical bodies of different sizes are sometimes seen embedded in the mitochondrial matrix between the mem- branes. The mitochondria are the carriers of 100 CELL ULTKASnaCTLRE IN MAMMALS most cellular enzymes and one believes that the membranous struct m'es of their interior represent either the enzymes proper or the surfaces upon which the enzymes and the metabolites interact. Microbodies. The microbodies are spheri- cal and usually smaller than the mitochon- dria. They are surrounded by a single mem- brane, about 50A thick. The matrix has the same appearance as that of the mitochondria but lacks the inner membranes of the latter (Fig. 3). They have been demonstrated so far only in kidney, liver, cortical adrenal cells, and in the delta cells of the endocrine portion of the pancreas. They may represent the precursors of mitochondria although this has not been convincingly proved. It is still a mystery how mitochondria develop. Some people believe they can arise de novo from the ground substance of the cytoplasm, where possibly the microbodies would con- stitute an intermediate stage. Others con- sider a splitting or budding of already exist- ing mitochondria to be more likely, as is known to occur during cell mitosis. Large Granules. The large granules en- countered in cells of various tissues are usually very dense; this has led most in- vestigators to believe that they contain lipids. They have, therefore, often been called cell lipid granules. However, the den- sity varies from time to time and it is diffi- cult to predict if this is due to a certain func- tional variation or if it mainly reflects a difference in structures. Most granules in cells represent a secretory product and as such will be dealt with below (Secretion). The re- maining large granules are of two types - — spherical granules and granules with ir- regular outlines. The large spherical granules are surrounded by a distinct single mem- brane and display a medium dense structure- less matrix (Fig. 3). It is most likely that they represent end products of substances taken in by the cells by means of a micro- pinocytotic activity through the tubular invaginations of the plasma membrane. By a certain metabolic process, the engulfed fluid, and therein dissolved substances, become concentrated and now appear as dense granules. This is surely the case with macrophages and has also been dem- onstrated in connection with uptake of proteins by the proximal convoluted cells of the nephron and of the intestinal cells. The granules with irregular outlines most likely represent lipid granules which the cell handles as part of its metabolism (Fig. 8). Structural evidence is at hand for a certain interaction between the lipid granules and the mitochondria, as for instance in liver cells. The intense blacken- ing of lipid granules by osmium tetroxide indicates that these granules contain un- saturated sulfhydryl groups. Saturated fats do not take stain with osmic acid; this can be clearly demonstrated in fat cells where the fat globules are convincingly stained with Sudan III for light microscopy but in the electron microscope show up as un- stained vacuoles with a bordering thin membrane. Pigments. The pigments represent an- other type of spherical granule which can be encountered in a number of cells. Similar to lipid granules, they stain intensely with osmium tetroxide (Fig. 8). The pigments of the retina are surrounded by a single mem- brane. Their matrix is homogeneous. The pigments encountered in the cells of the epidermis (often called melanin granules) do not have a limiting membrane but unveil an abundance of small pigment micelles each of which has a diameter of about 75A. The pig- ments of the epidermis are supposedly formed within special cells, the melanocytes, and migrate into the basal cells of the epidermis. Golgi ApiJaratus. The Golgi apparatus is located near the nucleus mostly surrounding its one pole like a halo. It consists of a sys- tem of paired membranes, small vesicles and 101 ELECTRON MICKOSCOI'Y granules. The membranes have a smooth surface and enclose clear spaces. The num- ber and length of the membranes vary, presumablj'^ because of different functional stages of the Golgi apparatus. There are mostly several pairs of membranes arranged in parallel form with each other and one can occasionally see that the clear space which each pair of membranes encloses is distended to a vacuole of varying size (Fig. 3). The small vesicles are quite numerous and bordered by a smooth membrane. Fre- quently, the clear centers of the small vesicles become condensed, thus obtaining the appearance of small granules (Fig. 2). The diameter of the small vesicles and the granules ranges between 200A and lOOOA. The appearance of the Golgi apparatus as a whole varies greatly from cell to cell and from tissue to tissue. It is best developed in secretory cells where its function presimiably is involved in the secretory process. Evi- dently the secretion products are prefab- ricated elsewhere in the cell, but the end products appear as small secretory granules within the Golgi zone. Here they enlarge and migrate eventually to the upper part of the cell. In non-secretory cells, the Golgi ap- paratus presumably plays an important role in the metabolism of the cell, either by offering its membranes as surfaces for en- zymatic activity or by being utilized as a system of channels for the intracellular flow of metabolites and fluid. Small Vesicles. Vesicles of the same order of magnitude as those found within the Golgi zone may be traced elsewhere in the cell. Their origin is unknown but they could possibly be derived from the Golgi ap- paratus. In some instances, as in non-ciliated cells of the ciliated epithelia of the bronchi and bronchioles (cross reference: ciliated epithelia ultrastructure), in the distal con- voluted tubule cells of the kidney, and in Fig. 9. Surface area of a dark, intercalated cell of the collecting tubvile of the mouse kidney. The cytoplasm is pervaded by abundant microvesicles, one in connec- tion with the surface (at 9). Tiny microvilli (Vi) extend into the tubular lumen (Lu). In between the vesicles, which all are bound by a smooth membrane, are abundant granules (arrows) with the size of about 150A, probably corresponding to granules which in other cells have been demonstrated to contain RNA. Magnification 57,000X 102 CELL ULTRASTRLCTURE L\ MAMMALS the parietal cells of the gastric mucosa, they appear in great abundance in the luminal portion of the cell. They seem to migrate towards the cell surface and fuse with the surface plasma membrane (Fig. 9). It may be assumed that they participate in a certain secretory process, involving the transport of large amounts of fluid to the cell surface. Similar vesicles are a prominent feature of the capillary wall, where many such struc- tures are found thi'oughout the endothe- lial cytoplasm as well as in connection with both the surface and the basal plasma membrane (Fig. 10). The theory has been forwarded that they represent the structural evidence of fluid being transported across the capillary wall. In nerve endings and in connection with synapses, similar vesicles have been demonstrated to contain acetyl- choline. The presence of synaptic vesicles has suggested that these might be involved in either the formation of the chemical mediator or in its transmission across the synapse. Ergastoplasm (rough-surfaced endo- plasmic reticulum). In the light micro- scope, areas with basophilic structures have been called ergastoplasm. The electron microscopical analysis of such areas reveals highly complicated systems of paired mem- branes, each membrane having a thickness of about 40A (Fig. 1). Each pair of membranes surrounds a light space called cisterna. It has been demonstrated in nerve cells that the cisternaeof the Nissl body communicate with each other. The cytoplasmic aspect of each membrane is studded with numerous small granules with a thickness of 150A (Fig. 2). These granules have been separated from the membranes by cell fractionation and subse- quent differential centrifugation. Enzymatic tests prove that they contain ribonucleic acids Fig. 10. Part of a lymph capillary in the human epididymis. The cytoplasm of the capillary endothelium (L) contains a large number of microvesicles. One mitochon- drion (M) is seen. At the top is the capillary lumen (Lu) and at the bottom part of a fibroblast (F) as well as several cross sectioned collagen fibrils (C). The lymph capil- laries lack the usual type of basement membrane. Instead, a network of fine fibrils (Re) probably of reticular nature, establish the immediate base upon which the endo- thelial cells rest. Magnification 34,oOOX 103 ELECTRON MICROSCOPY and the granules or particles have, therefore, been called RNA-particles. The number of membranes varies from tissue to tissue; In cells engaged in heavy protein synthesis, such as the exocrine cells of the pancreas, the ergastoplasm dominates the cell with its complicated pattern of membranes and cisternae throughout the entire cell ex- cept in the Golgi area. Evidently the mem- branes and cisternae produce the precursors of the zymogen granules which, at a later stage appear within the Golgi apparatus. In nerve cells, the Nissl bodies have the same ultrastructure, participating in the produc- tion of proteins needed within the cell itself. Again in other cells the ergastoplasm may be seen as scattered short pairs of membranes with the RNA-particles attached. The terminology used in connection with the ergastoplasm is somewhat confusing. As mentioned, the term "ergastoplasm" refers to basophilic areas which can be seen in the light microscope. The term "endoplasmic reticulum," introduced by Porter and Kail- man (1952), originally referred to a structure presumably vesicular or tubular w^hich was observed in whole cells in tissue cultures. Extended studies demonstrated that the endoplasmic reticulum corresponds to the ergastoplasm. When later structures like infoldings of the plasma membrane, pino- cytosis vesicles, and the membranes of the Golgi apparatus were sufficiently analyzed, it was found that some of these structures could be in direct continuity with the er- gastoplasm. It was, therefore, suggested that all membranes wdthin the cell, whether smooth or rough surfaced, may represent diverse differentiations of one single mem- branous system. Therefore, the ergasto- plasm wdth its RNA-dotted membranes is usually referred to as the rough-surfaced endoplasmic reticulum, whereas the Golgi membranes, cytoplasmic vesicles and in- folded or invaginated portions of the plasma membrane are called smooth-surfaced endo- plasmic reticulum. For a more detailed dis- cussion of this problem, consult Haguenau (1958) International Review of Cytology, VII. Another example of a purely smooth- surfaced endoplasmic reticulum is foimd in striated muscle cells, here called sarcoplasmic reticulum.Ii is an elaborate network of smooth tubules around the mj^ofibrils with expan- sions of the system specifically localized in relation to the Z-band. It has been suggested that this system might function in the inward spread of the excitation impulse to contract. Microsonies. When using differential cen- trifugation the membranes and cisternae of the ergastoplasm break up to form small spheres which can be isolated at a certain cent rifugat ion speed. They then represent the microsome fraction (Fig. 2). RNA-particles. In most cells, the cyto- plasm displays an abundance of small granules or particles with a diameter of about 150A (Fig. 9). They appear single or in clusters of 3-5 and are freely dispersed throughout the cytoplasm. They are iden- tical in size and shape to the particles which are attached to the membranes of the ergastoplasm. It has been clearly demon- strated that either type of granules contains ribonucleoproteins and is, therefore, called either RNA- or RNP-particles. Glycogen Granules. Particles have been demonstrated in the cytoplasm of the stri- ated muscle and in the heart muscle which have a diameter ranging between 150A and 300A. Thus, they are somewhat larger than the ribonucleoprotein particles with which they can easily be confused. The larger particles are more variable in their size and shape and have less sharply defined margins. Histochemical tests seem to prove that they contain glycogen and it is, therefore, likely that they represent a particulate form of glycogen. One has not been able to demon- strate similar granules in mammalian liver cells. The glycogen-rich areas of the liver cell cytoplasm usually show a diffuse, cot- ton-wool texture of low density when using solely osmium tetroxide as a tissue stain. 104 CELL ULTKASTRUCTLRE IN MAMMALS However, in appl^dng an additional stain of heavy metal to the thin sections, as for in- stance phosphomolybdic acid, the large glj^cogen particles also become visible in liver cells. Fibrillar Structures Fibrillar structures can easily be resolved with the electron microscope. The}^ may be located within the cell cytoplasm as is the case with myofilaments, tonofilaments, and the neurofibrils, or they are found at the outer surface of the cells or in the interstitial space, here recognized as collagen, reticular fibers and elastin. Intracellular. Striated muscle filaments. The most evident myofilaments are found in the striated skeleton and heart muscle cell (Fig. 11). Here they occupy the main portion of the cell oriented longitudinally, and they represent the contractile elements of the muscle cell. The myofilaments extend along the whole length of a sarcomere which is the structural, repeated unit of the muscle fiber. The thickness of the individual myo- filament varies within different areas of the myofibril, with its smallest diameter related to the area of the I-band and the H-band, and its largest diameter within the Z-band, S-band, and M-band. In a stretched muscle fiber, presumably corresponding to the re- laxed position, the mean diameter of the myofilament is about lOOA in the H-band, whereas the same value for the M-band is in the neighborhood of 150A. During muscle contraction, the diameter of the individual myofilament increases about three times as compared to its stretched diameter. It has been proposed that each myofilament in turn is composed of three subunits. Each subunit consists of rows of rodlets measuring 20A in length. The subunits are assumed to represent cables of supercoiled a-helices of protein molecules. The main constituents of the myofilaments are the proteins actin and myosin. As yet, it has not been convincingly proved what part of the myofilament represents the actin and what represents the myosin. It has been sug- gested that actin is represented by one set of filaments and myosin by another set. There is also a num})er of theories about how the contraction occurs from a structural point of view. Further extensive investiga- tions are needed until a definite solution is arrived at regarding the striated muscle. Smooth muscle filaments. In the smooth muscle cell, the contractile elements are not as easily demonstrated as in the striated one (Fig. 12). This is particularly true in the smooth muscle cells of the blood vessels. Although it is well known that these cells do contract, the cytoplasm is remarkably devoid of any fibrillar structures. However, in the smooth muscle cells of the small bronchi and bronchioles of the lung as well as in those of the intestinal wall, a longitud- inal striation can more readily be seen. The diameter of the filaments here average 150A. The length of each filament is more difficult to determine and it seems that it does not extend for a length of more than half a micron. The difference in ultrastructure be- tween the striated and the smooth muscle filaments seems to reflect the difference in their function, since it is well known that the smooth muscle cell has a much slower rate of contraction than the striated one. Tonofilaments. In the cells of the epidermis an elaborate system of tonofibrils crisscrosses the cytoplasm. The tonofibrils are well dis- tinguished in the light microscope. Their ultrastructure is characterized by an abun- dance of .small tonofilaments, oriented longi- tudinally to the axis of the tonofil^ril (Fig. 13). Each tonofilament has a thickness of about 190A and an approximative length of 0.5 micron with seemingly tapered ends. The filaments have a light core and an electron dense wall, the latter with a diameter of about 70A. It has not as yet been possible to determine whether the tonofilaments are spindleshaped or slightly twisted around 105 Fig. 11. Myofibril in the specific tissue (im- pulse conducting system) of the steer heart. The longitudinallj' arranged myofilaments are seen with their various thickenings related to particu- larly the Z and M bands. Mitochondria (M) and microvesicles (Ve) are abundant throughout the sarcoplasm. Magnification X 49 ,000 Fig. 12. Longitudinal section through the con- tractile region of the sarcoplasm of a smooth mus- cle cell of the mouse intestine. Fibrillar units may be distinquished (arrows) and a certain longitud- inal arrangement is vaguelj' indicated. In the con- tractile region is always present a certain number of dense, ovoid structures (0), typical for the smooth muscle sarcoplasm. At the cell boundary (CB) are aggregations of microvesicles (Ve). The mitochondria (M) are clustered in the central por- tion of the cell. Magnification X47,000 CELL ULTUASTKL'CTURE IN MAMMALS 'aSESii.. Fig. 13. Detail of a basal cell of the human epi- dermis. The cytoplasm is run through by abundant tonofilaments (Tf) most of which are sectioned longitudinally. They display a hollow structure (arrows) and originate from dense areas (X) of the plasma membrane which faces the basement mem- brane (BM). Magnification 83, 000 X each other m the formation of the tonofibrils. The tonofibrils originate from and terminate at the desmosomes which are buttonhke structures associated with the plasma mem- Cl ^H^mSi^ ^i^-*' a2ju Fig. 14. Cell l)()iiiul:uu'.-> ui luu adjacent cells (Cl and C2) of the basal part of the human epider- mis. The cells are highly interdigitated and the mutual attachment established through desmo- somes (De) in association with which the in- tercellular space is wider than elsewhere. The tonofilaments (Tf) terminate (or originate) at the desmosomes. Magnification 90,000X brane (Fig. 14). Histochemical and bio- chemical tests prove that the main com- ponent of the tonofilaments is keratin, a tough noncontractile, in young tissue quite 107 ELECTRON MICROSCOPY elastic, structure, which gives the cells of the organ and is probably dependent on the age epidermis a certain elasticity and firmness, and the function of the fibril. Each collagen Neurofibrils. In the axoplasm of nerve cells fibril is in turn composed of small protofibrils still another fibrillar cytoplasmic structure or tropocollagcn units with a diameter of loA can be found. The neurotibrils of classical and a length of 2G00A. The tropocollagcn histology represent aggregates of axon fila- units are tied up in a staggered fashion to ments large enough to be resolved in the form the collagen fil)ril. The bands of the light microscope. Deposition of heavy met- collagen fibril represent discontinuities in the als, as with the histological silver and gold staggered arrangement of the protofibrils technics, favors the detection of the very and stand out in the electron micrograph thin fibrous structures. In the early days of because heavy metals have a greater affinity electron microscopy, these structures were for this irregularity. The formation of the mistaken for tubules, hence the first name to immature collagen seems to occur at the sur- be coined was neurotubules. Presently, there face of the fibroblasts, the main cell type of seem to be two kinds of fibrils in the axo- all connective tissues. Small fibrils without plasm — the protoneurofibrils and the neuro- periodicity appear at the surface of the filaments. The proto7ieuro fibrils appear as fibroblast. These unit fibrils organize out of smooth threads with a thickness of about 60 or polymerize from material present at the to 80A. Their length ranges between 0.5 to cell surface, and from here the fibrils, in 1 micron. They have an irregular course and many cases already in bundles, are shed into are seen to branch and interconnect. The the intercellular space. The fibrils will first neurofilaments have a thickness of about appear with an axial periodicity of 210A, but 150A with indefinite length. They are some- as they increase in size they also change into times of a double-edged appearance. Their a 640A periodicity. The subsequent fibril surface is smooth and they do not seem to growth apparently occurs by accretion of constitute any part of the endoplasmic materials from the general environment of reticulum. The function of either type of the intercellular spaces and not by fusion of fibrillar structure is unknown. smaller fibrils as believed earlier. Subsequent Extracellular. Collagen. The main fine layers of collagenous material are deposited component of the connective tissue is the upon the core, represented by the tropocoi- collagen fibers. The width of the collagen lagen unit fibril. The origin of the tropocol- fiber is about one micron and its length is lagen fibril is not settled, but considering the indefinite. Each fiber is in turn built up of presence of abundant rough-surfaced endo- numerous collagen fibrils which also may be plasmic reticulum in the cytoplasm of the seen single or in groups of two or more fibrils fibroblast which presumably is involved in scattered in the interstitial tissue (Fig. 15). the protein synthesis, it seems justifiable to Within each group of collagen fibrils the suggest the following mechanism as being fibrils are usually parallel with each other, the most likely regarding the role of the The collagen fibril has a thickness which fibroblast in the formation of the collagen varies between 400 and 2400A depending on fibrils. From the cisternae of the rough-sur- the age (Fig. 10) ; its length is indefinite. The faced endoplasmic reticulum of the fibroblast fibril has an axial periodicity of light and the monomeric form of the tropocoUagen dense bands with a length of each period of unit fibril is discharged to the environment of approximately 640A. In the light and dense the cell and is here ciuickly induced to segments can be seen several smaller bands polymerize by enzymes resident in tem- of varying density and thickness. The length plates at the cell surface or in the unit fibrils and number of bands varies from organ to themselves. 108 CELL ULTRASTRUCTURE IN MAMMALS Fig. 15. Connective tissue of the tracheal submucosa of the rat showing elastic fibers (E), collagen fibrils (C), fibroblasts (F) and reticular fibrils (Re). Magnification 24, 000 X Reticular fibers. In the light microscope fine fibers are found in loose connective tissue and in the interstitial substance of cartilage which can be elect ively impregnated with silver and they are, therefore, often called argyrophil fibers (Fig. 15). Their submicro- scopic structure is like that of collagenous fibers. They have a thickness of about 200 A and have the characteristic cross striations of collagen. In growing tissue it has been demonstrated that the bundles of fibers increase in thickness and finally lose the ability to be impregnated with silver. The reticular fibers found in the interstitial sub- 109 ELECTRON ^IICROSCOPY stance of cartilage do not always show the cross striatioiis (Fig. 10), and it may, there- fore, be assnmed that not all reticular fibers can be transformed into collagenous ones. Elastin. Elastic fibers are found in loose connective tissue (Fig. 15). They have a thickness of about one micron and seem to have an indefinite length. They are not fibrillar but have a homogeneous appearance in the light microscope. Elastic fibers at most show only a weak positive birefrin- gence, but become strongl}^ birefringent on stretching. This is caused by an orientation of the submicroscopic components in the direction of the fiber axis. These ultrastruc- tural components are difficult to demonstrate in osmium-fixed specimens, but it has been possible to distinguish thin filaments with a thickness of about 70A at the periphery of the elastic fiber. Hence, it seems that the elastic fiber has two main components. The dominating structure is a non-fibrous dense cement substance, presumably an albumi- noid which embeds the less apparent elastic filaments. Crystals There are only a few examples of where crystals may be found in normal mammalian tissues. Intracellular. Intracellularly located are the so called crystalloids of the interstitial (Sertoli) cells of the human testis. These crystals are probably of protein natvn-e. The crystals can be seen in the light microscope. They are usually elongated structures with rounded or pointed ends. Each cr3^stalloid body is made up of numerous dense granules with a diameter of about 150A. They are spaced about 190A apart along two axes which are approximately at right angles to each other. This pattern is thought to represent the arrangement of macromole- cules in the lattice of a protein crystal. The function of the crj^stals and the reason for their presence in the Sertoh cells of the testis is unknown. It has been suggested that these cells have an endocrine glandular function and, therefore, the crystals may be involved in the production of the hormone. Extracellular. Extracellularly located crystals are encountered in the bone tissue where they make up the strong and resistant component of the skeletal system. The crys- tals have a width of about 35A. They are scat- tered in the extracellular matrix of the bone tissue but they are also lined up within the collagen fibrils with their long axes oriented parallel to the long axis of the collagen fibril. Selected-area electron-diffraction of these structures has revealed that they are crys- tals of apatite, more specifically hydroxy- apatite [Caio(P04)6(OH)2]. Similar crystals appear under pathologic conditions in areas undergoing calcification like in aging car- tilage, and they also form the main com- ponent of concretions precipitated through- out the urinary system (Fig. 16). Secretion As already mentioned, most of the large granules encountered in secretory cells are associated with secretory processes and have, therefore, been called secretory granules. Although formed within the cell and from the beginning being a part of the cytoplasm, they become discharged and perform their action outside the cell territory. There are two types of secretory cells, namely the exocrine and the endocrine cells. Exocrine Secretion. The exocrine secre- tory granules appear within the Golgi zone, but are evidently preformed in association with the ergastoplasmic sacs (or the cis- ternae of the rough-surfaced endoplasmic reticulum). The fii'st indication of secretory granules within the Golgi zone is a condensa- tion of the Golgi vacuoles or a swelling of the Golgi vesicles and small granules. Sei'ous secretion. In secretory cells with a serous production, the indi^-idual secretory granules are usually surrounded by a single membrane and there is no indication of a coalescence of granules. The granules mi- 110 CELL I LTR ASTRUCTl RE IN >IA>L>IALS Fig. 16. Hydroxyapatite crystals, located in a concretion, experimentally pro- duced in the tubular lumen of the proximal convolution of the rat kidney by injecting subcutaneoush" large doses of parathyroid hormone. (From Engfeldt, et al., 1958). Magnification 16o,000X grate toward the luminal part of the cell Mucous secretion. In secretory cells with a where, in some instances, it can be seen that production of mucin, the secretory granules the limiting membrane of the granule fuses are not bound by a membrane. Frequenth', with the surface plasma membrane and the fusion of several mucous granules occurs, granule empties its content into the lumen, and the discharge of mucin does not occur 111 ELECTRON MICROSCOPY until the whole luminal part of the cell (the goblet) is filled by a muhittide of mucous granules which at all limes are discharged into the lumen. It is believed that either type of exocrine cell has the ability to re- generate new secretory granules. Accord- ing to earlier theories, the cell would disin- tegrate after the secretory cycle is completed. Endocrine Secretion. The endocrine granules are usually smaller than their exocrine relatives. They are also more elec- tron-dense and are mostly surrounded by a single membrane. Few high resolution stud- ies have so far been performed on endocrine glands, but judging from the data available, the formation of the endocrine granules is similar to what is known about the exocrine. A varying amount of rough-surfaced endo- plasmic reticulum (ergastoplasm) as well as an abundance of RNA-particles seem to contribute the necessary prerequisites for the production of the early stages of endo- crine granules. In the beta-cells of the endo- crine portion of the pancreas, it has been quite convincingly demonstrated that the granules first appear in the Golgi zone. The matter of bringing these products in contact with the capillaries of the gland is still not fully explained but evidence is at hand for a migration of the endocrine granules toward that part of the cell which faces the capillary. In some instances, it has also been possible to demonstrate that the membrane of the granule fuses with the plasma membrane, thus giving the dense granule the oppor- tunity to be discharged into the small extra- cellular space existing between the plasma membrane and the basement membrane which surrounds the endocrine cells. From here, the content of the granule may quite easily diffuse into the interstitial space and from there into the capillary. Further studies are, however, needed to prove that this occurs in relation to all endocrine organs. Ground Substance of the Cytoplasm In light microscopy, the term "ground substance" referred to that part of the cyto- plasm which was not organized as mito- chondria, Golgi apparatus, specialized cell inclusions like zymogen granules, pigments, or structures such as myofibrils. With the introduction of the ultrastructural era, it was demonstrated that the ground substance does contain particular well-defined struc- tures like cytoplasmic membranes of various kinds, vesicles, RNA-particles, osmiophilic large granules, and various fibrillar struc- tures. These ultrastructures may, of course, still be regarded as being part of the "ground substance" of the cytoplasm. However, it may also be justifiable today to call that W'hich is as yet not defined as discrete struc- tural entities as representing the ground sub- stance awaiting new techniciues to be de- veloped before this portion of the cytoplasm can be described. In most electron micro- graphs, a certain homogeneous background characterized by a certain electron density can always be recorded. This "background" represents the ground substance in our present concept of the cytoplasm. It is con- ceivable that this background contains a great variety of salts and ions as well as carbohydrates, proteins and fats w^hich available staining techniques and resolving power of the electron microscope fail to bring out. Until these are further developed, let us consider the homogeneous background of our electron micrographs as being the true ground substance. In doing so, we may create the necessary challenge to explore the un- known structural world of the atoms of the cytoplasm. REFERENCES General Reviews Selby, C. C, "Microscopy. II. Electron micro- scopy: a review," Cancer Research, 13, 753 (1953). Sjostrand, F. S., "Electron microscopy of cells and tissues," in "Physical Techniques in Biological Research," 3, 241 (1956). Eds. G. Oster and A. W. Pollister, Academic Press, Inc., New York. Sjostrand, F. S., "The ultrastrvicture of cells as revealed by the electron microscope", Int. Rev. Cytol., 5, 455 (1956). 112 CELL ULTRASTRrCTURE IN MAMMALS Oberling, C, "The structure of the cytoplasm," Int. Rev. Cytol., 8, 1 (1959). Miller, F., "Orthologie und Pathologie der Zelle im elektronenmikroskopischen Bild," "Vehr. Deutsch. Ges. Pathologie," p. 261. Gustaf Fischer Verlag, Stuttgart, 1959. Selby, C. C, "Electron microscopy: techniques and applications in cytology," in "Analytical Cytology," p. 273, Ed. R. C. Mellors, Mc- Graw-Hill Book Company, Inc., New York, 1959. "The cell," Vol. 2: Cell Constituents, Eds. J. Brachet and A. E. Mirsky, Academic Press, Inc., New York and London, 1960. Basement Membrane Weiss, P. and Ferris, W., "The basement la- mella of amphibian skin", /. Biophys. Bio- chem. Cytol., 2, 275 (1956) Suppl. VAN BrEEMEN, v. L., ReGER, J. F., AND CoOPER, W. G., "Observations on the basement mem- branes in rat kidney," /. Biophys. Biochem.. Cytol., 2, 283 (1956)' Suppl. Cent Hole Yamada, E., "The fine structure of centriole in some animal cells", Proc. 1st. Reg. Conf. in Asia and Oceanic, p. 247, Tokyo, 1956. Bernhard, W. AND DE Harven, E., "Sur la presence dans certaines cellules de Mammi- feres d'un organite de nature probablement centriolaire", C. r. Acad. Sci. Paris, 242, 288 (1956). de Harven, E. and Bernhard, W., "Etude au microscope ^lectronique de I'ultrastructure du centriole chez les vertebras" Z. Zellf., 45, 378 (1956). Cell Surface Fawcett, D. W., "Structural specializations of the cell surface," in "Frontiers in Cytology," p. 19, Ed. S. L. Palay, Yale University Press, New Haven, 1958. Rhodin, J. AND Dalhamn, T., "Electron micro- scopy of the tracheal ciliated mucosa in rat," Z. Zellf., 44, 345 (1956). Rhodin, J., "Anatomy of kidney tubules" Int. Rev. Cytol., 7, 485 "(1958). Weiss, P., "Cell contact," Int. Rev. Cytol., 7, 391 (1958). Cilia Fawcett, D. W. and Porter, K. R., "A study of the fine structure of ciliated epithelia," J. Morph., 94, 221 (1954). Rhodin, J. and Dalhamn, T., "Electron micros- copy of the tracheal ciliated mucosa in rat," Z. Zellf., 44, 345 (1956). Rhodin, J., "Ciliated epithelia". Int. Rev. Cytol., 10 (1962). Collagen Gross, J., "The behavior of collagen units as a model in morphogenesis," J. Biophys. Bio- chem. Cytol., 2, 261 (1956) Suppl. Crystals (extracellular) Robinson, R. A. and Watson, M. L., "Crystal- collagen relationships in bone as observed in the electron microscope," Ann. N. Y. Acad. Sci., 60, 596 (1955). Knese, K.-H. and Knoop, A.M., "Elektronenop- tische Untersuchungen iiber die periostale Osteogenese," Z. Zellf., 48, 455 (1958). Glimcher, M. J., "Molecular biology of mineral- ized tissues with particular reference to bone," Rev. Modern Physics, 31, 359 (1959). Crystals (intracellular) Fawcett, D.W. and Burgos, M. H., "Observations on the cytomorphosis of the germinal and in- terstitial cells of the human testis," in Vol. 2, "Ageing in Transient Tissues, Ciba Foun- dation CoUoquia on Ageing," p. 86, Eds. G. E. W. Wolstenholme and E. C. P. Millar, J. & A. Churchill, Ltd., London, 1956. Elastin Hall, D. A., "The fibrous components of con- nective tissue with special reference to the elastic fiber," Int. Rev. Cytol., 8, 212 (1959). Endocrine Secretion Ferreira, D., "L'ultrastructure des cellules du pancreas endocrine chez I'embryon et le rat nouveau-ne," /. Ultrastructure Research, 1, 14 (1957). MuNGER, B. L., "A light and electron microscopic study of cellular differentiation in the pan- creatic islets of the mouse," Am. J. Anat., 103, 275 (1958). Lacy, P. E., "Electron microscopic and fluorescent antibody studies on islets of Langerhans," Exp. Cell Research, 7, 296 (1959) Suppl. Ekholm, R. and Sjostrand, F. S., "The ultrastruc- tural organization of the mouse thyroid gland," /. Ultrastructure Research, 1, 178 (1957). Herman, L., "An electron microscope study of the salamander thyroid during normal stimu- lation," J. Biophys. Biochem. Cytol., 7, 143 (1960). 113 ELECTRON MICROSCOPY Endoplasmic Reticuhim Palade, G. E., "Tlie piidoplasmic reticulum" J. Biophys. Biochem. Cytol., 2, 85 (1956) Suppl. Haguenau, F., "The ergastoplasm: its history, ultrastructiire, and biochemistry," Int. Rev. Cytol., 7, 425 (1958). Exocrine Secretion Sjostrand, F. S. and Hanzon, V., "Membrane structures of cytoplasm and mitochondria in exocrine cells of mouse pancreas as revealed by high resolution electron microscopy," Exp. Cell Research, 7, 393 (1954). Rhodin, J. and Dalhamn, T., "Electron micro- scopy of the tracheal ciliated mucosa in rat," Z. Zellf., 44, 345 (1956). Palay, S. L., "The morphology of secretion," in "Frontiers in Cytology," p. 305, Ed. S. L. Palay, Yale University Press, New Haven, 1958. Ekholm, R. and Edlund, Y., "Ultrastructure of the human exocrine pancreas, J. Ultrastruc- ture Research, 2, 453 (1959). Glycogen Bernhard, W. and Rouiller, C, "Close topo- graphical relationship between mitochondria and ergastoplasm of liver cells in a definite phase of cellular activity," /. Biophys. Bio- chem. Cytol., 2, 73 (1956), Suppl. Fawcett, D. W. and Selby, C. C, "Observations on the fine structure of the turtle atrium," J. Biophys. Biochem. Cytol., 4, 63 (1958). Watson, M. L., "Staining of tissue sections for electron microscopy with heavy metals," /. Biophys. Biochem. Cytol., 4, 475 (1958). Golgi Apparatus Dalton, a. J., "A study of the Golgi material of hepatic and intestinal epithelial cells with the electron microscope, Z. Zellf., 36, 522 (1952). Sjostrand, F. S. and Hanzon, V., "Ultrastruc- ture of Golgi apparatus of exocrine cells of mouse pancreas," Exp. Cell Research, 7, 415 (1954). Gatenby, J. Bronte, "The Golgi apparatus," R. Micr. Soc, 74, 134 (1955). Haguenau, F. and Bernhard, W., "L'appareil de Golgi dans les cellules normales et canc^reuses de vertebras," Arch, d'anatomie micr. et morph. exp., 44, 27 (1955). Dalton, A. J. and Felix, M., "A comparative study of the Golgi complex," J. Biophys. Biochem. Cytol., 2, 79 (1956) Suppl. Large Granules Rhodin, J. and Dalhamn, T., "Electron micros- copy of the tracheal ciliated mucosa in rat," Z. Zellf., 44, 345 (1956). NiLSSON, O., "Ultrastructure of mouse uterine surface epithelium under different estrogenic influences," /. Ultrastructure Research, 1, 375 (1958). Zelander, T., "Ultrastructure of mouse adrenal cortex," ./. Ultrastructure Research, 2, (1959) Suppl . Ladman, a. J. and Young, W. C, "An electron microscopic study of the ductuli efferentes and rete testis of the guinea pig," J. Biophys. Biochem. Cytol., 4, 219 (1958). Microbodies Rhodin, J., "Correlation of ultrastructural organi- zation and function in normal and experi- mentallj' changed proximal convoluted tubule cells of the mouse kidney," Thesis, Karolinska Institutet, Stockholm, 1954. Rouiller, C. and Bernhard, W., "Microbodies and the problem of mitochondrial regenera- tion in liver cells," /. Biophys. Biochem. Cytol., 2, 355 (1956) Suppl. Engfeldt, B., Gardell, S., Hellstrom, J., iVEMARK, B., Rhodin, J. and Strand, J. "Effect of experimental hyperparathyroidism on renal function and structure," Acta Endo- crinologica, 29, 15 (1958). Microsomes Palade, G. E. and Siekevitz, P., "Pancreatic microsomes," /. Biophys. Biochem. Cytol., 2, 671 (1956). Siekevitz, P. and Palade, G. E., "A cytochem- ical study of the pancreas of the guinea pig," /. Biophijs. Biochem. Cytol., 4, 309 (1958). Mitochondria Palade, G. E., "The fine structure of mitochon- dria," Anat. Rec, 114, 427 (1952). Sjostrand, F. S., "Electron microscopy of mito- chondria and cytoplasmic double mem- branes," Nature, 171, 30 (1953). Steffen, K., "Chondriosomen und Mikrosomen (Spharosomen)," "Encj^clopedia of Plant Physiology," p. 574, Ed. W. Ruhland, Springer-Verlag, Berlin, 1955. Siekevitz, P. and Watson, M., "Cytochemical studies of mitochondria," J. Biophys. Bio- chem. Cytol., 2, 639 (1956). Palade, G. E., "Electron microscopy of mito- chondria and other cytoplasmic structures," in "Enzymes: Units of Biological Structure 114 CELL I LTR AS TRUCTl RE L\ MAMMALS and Function," p. 185, Ed. O. H. Gabler, Academic Press, Inc., New York, 1956. Nerve Cells Palay, S. L. and Palade, G. E., "The fine struc- ture of neurons," /. Biophys. "Biochem. Cytol., 1, 69 (1955). Fernandez-Mo ran, H. and Brown, R., "The submicroscopic organization and function of nerve cells," Exp. Cell Research, 5, (1958) Suppl. Neurofibrils EsTABLE, C, Acosta-Ferreira, W., and Sotelo, J. R., "An electron microscope study of the regenerating nerve fibers," Z. Zellf., 46, 387 (1957). ScHMiTT, F. O., "Molecular organization of the nerve fiber," Rev. Modern Physics, 31, 455 (1959). Nucleus Afzelius, B. a., "The ultrastructure of the nuclear membrane of the sea urchin oocyte as studied with the electron microscope," Exp. Cell Research, 8, 147 (1955). Watson, M. L., "The nuclear envelope. Its struc- ture and relation to cj'toplasmic membranes," J. Biophys. Biochem. Cytol., 1, 257 (1955). Haguenau, F. and Bernhard, W., "Particulari- tes structurales de la membrane nucleaire," Bulletin du Cancer, 42, 537 (1955). Watson, M. L., "Further observations on the nuclear envelope of the animal cell," /. Bio- phys. Biochem. Cytol., 6, 147 (1959). Nucleolus Bernhard, W., Bauer, A., Gropp, A., Hague- nau, F., and Oberling, C, "L'ultrastructure du nucleole de cellules normales et cancereuses," Exp. Cell Research, 9, 88 (1955). Pigments Falk, S. and Rhodin, J., "Mechanism of pig- ment migration," "Proc. Stockholm Confer- ence on electron microscopy," p. 213, 1956, Eds., F. S. Sjostrand and J. Rhodin, Almquist and Wiksell, Stockholm, 1957. Wellings, S. R. and Siegel, B. V., "Role of Golgi apparatus in the formation of melanin granules in human malignant melanoma," /. Ultrastructure Research, 3, 147 (1959). Wellings, S. R. and Siegel, B. V., "Electron microscopy of human malignant melanoma," J. Nat. Cancer Inst., 24, 437 (1960). Charles, A. and Ingram, J. T., "Electron mi- croscope observations of the melanocyte of the human epidermis," J . Biophys. Biochem. Cytol., 6, 41 (1959). Plasma Membrane Robertson, J. D., "The molecular biology of cell membranes," in "Molecular Biology," p. 87, Ed. D. Nachmansohn, Academic Press, Inc., New York and London, 1960. Reticular Fibers Jackson, S. F., "The morphogenesis of avian ten- don," Proc. Royal Soc, B, 144, 556 (1956). Wassermann, F. and Kubota, L., "Observations on fibrillogenesis on the connective tissue of the chick embryo with the aid of silver im- pregnation," J. Biophys. Biochem. Cytol., 2, 67 (1956). Porter, K. R. and Pappas, G. D., "Collagen for- mation by fibroblasts of the chick embryo dermis," /. Biophys. Biochem. Cytol., 5, 153 (1959). Zelander, T., "Ultrastructure of articular carti- lage," Z. Zellf., 49, 720 (1959). RN A -Particles Palade, G. E., "A small particulate component of the cytoplasm," J. Biophys. Biochem. Cy- tol., 1,59 (1955). Palade, G. E., "Microsomes and ribonucleopro- tein particles," in "Microsomal Particles and Protein Synthesis," p. 36, Ed. R. B. Roberts, Pergamon Press, New York, 1958. Small Vesicles Rhodin, J. "Anatomy of kidney tubules," Int. Rev. Cytol, 7, 485 (1958). Hally, a. D., "The fine structure of the gastric parietal cell in the mouse," J. Anat., 93, 217 (1959). DE Robertis, E., "Submicroscopic morphology of the synapse," Int. Rev. Cytol., 8, 61 (1959). Fawcett, D. W., "The fine structure of capil- laries, arterioles and small arteries," in "The Microcirculation," p. 1, Eds. S. R. M. Rey- nolds and B. W. Zweifach, The University of Illinois Press, Urbana, 1959. Vial, J. D. and Orrego, H., "Electron micro- scope observations on the fine structure of parietal cells," J. Biophys. Biochem. Cytol., 7, 367 (1960). Smooth Muscle Caesar, R., Edwards, G. A., and Ruska, H., "Architecture and nerve supply of mamma- lian smooth muscle tissue," J. Biophys. Bio- chem. Cytol., 3, 867 (1957). Thaemert, J. C, "Intercellular bridges as proto- 115 ELECTRON MICROSCOPY plasmic anastomoses between smooth muscle cells," J. Biophys. Biochem. Cytol., 6, 67 (1959). Stereocilia Rhodin, J., "Ciliated epithelia," Int. Rev. Cytol., 10 (1962). Striated Muscle Hodge, A. J., "Fibrous proteins of muscle," Rev. Modern Physics, 31, 409 (1959). "Structure and function of muscle," Vol. 1, Struc- ture, Ed. G. H. Bourne, Academic Press, Inc., New York, 1960. Tonofilaments Selby, C. C, "An electron microscope study of the epidermis of mammalian skin in thin sec- tions," J. Biophys. Biochem. Cytol., 1, 429 (1955). Odland, G. F., "The fine structure of the inter- relationship of cells in the human epidermis," /. Biophys. Biochem. Cytol., 4, 529 (1958). Setala, K., Merenmies, L., Stjernvall, L., and Nyholm, M., "Mechanism of experimental tumorigenesis. IV. Ultrastructure of inter- follicular epidermis of normal adult mouse," /. Nat. Cancer Inst., 24, 329 (1960). Johannes A. G. Rhodin CILIATED EPITHELIA ULTRASTRUCTURE Ciliated epithelia are so called because many of the cells which form the epithelial layer are provided with a great number of small motile structures on their surface — the cilia. The ciliated epithelia are found in con- nection with organs and tissues where a sur- face has to be kept clean and moist, as in the respiratory tract (nose, trachea, bronchi), or in small ducts where a certain propulsion of the content of the duct is facilitated and aided by the beating of the cilia as in the male reproductive tract (efferent ducts of testis) and in the female reproductive tract (oviduct). Function The ciliated epithelium in mammals is characterized by several cell types (Fig. 8) the most prominent of which is the cili- ated cell. The other cells have been classi- fied as tion-ciliated cells which comprise secretory cells of two types {serous and mucous or goblet cells), the brush cells and the hasal cells. The functions of the different cells de- pend on the location of each cell. The goblet cell is predominant in all ciliated epithelia and keeps the surface moist by discharging continuously a more or less viscous mucin. The secretion product of the serous cells either dilutes the mucin and/or adds en- zymes (activators) to the content of the ducts. The brush cells, presumably young cells which have migrated from the basal portions of the epitheliimi, are primarily the precursors of the ciliated cells. They may probably also be transformed into mucous cells as well as serous cells. The appear- ance of the brush cells of the efferent ducts of the testis indicates that their function here is mainly a secretory one. However, certain structural features speak for the fact that they also have absorptive functions. The ciliated cells with their abimdant sur- face extensions, the cilia, participate in keeping the mucus blanket moving. The ciliary beat is rather complicated. It is com- posed of a rapid forward stroke and a slow backward stroke. The cilium is rigid and erected during the forward stroke, whereas it folds itself beneath the mucus in a limb and yielding movement during the backward stroke. The activity of the cilia is syn- chronized in local areas and moveinents of the cilia over larger areas have been com- pared with the appearance of a rye field when the wind blows across it. The basal cells are presumably the precursors of all the cells of the ciliated epitheUa. The mitotic activity among these cells is high and they migrate to the upper portions of the epi- thelial layer in order to replace ciliated or serous cells when they are sloughed off and lost in the moving mucus blanket. 116 CILIATED EPITIIELIA ULTRASTRUCTURE T. ..: ■ MU vt f ^/:^^r-; ".\vf^:.- ^'f^^_._ • -A/i ^ 'i'../'.-.- BM 'iff^T!11fffi Fig. 1. Longitudinal section of the pseudostratified columnar epithelium of the human trachea composed of ciliated cells (C), mucous cells (G), brush cells (BC), and basal cells (B). At the top is the tracheal lumen with the mucous blanket (MU); at the bottom, the basement membrane (BM) with its abundant reticular and collag- enous fibrils. Only the basal cells rest on the basement membrane. Intercellular spaces (I) are evident, and a wandering cell, presumably a lymphocj-te (L), is located in the space between the basal cells. Magnification 1,800X 117 ELECTRON iMICROSCOPY Structure are attached lo the nienihraiiesof the rough- Goblet cell. The goblet cell cytoplasm is surfaced endoi)lasinic reticulum (ergasto- characterized by a high content of ribonu- plasm). Precursors of the mucin are evi- cleic acid (RNA) particles, some of which dently formed within the cisternae which are Fig. 2. Detail ol I'ig. 1 shuwing twu muc-ous cells (G) and one ciliated cell (C). The mitochondria (m) are abundant in the upper part of the ciliated cell in which the Golgi apparatus (go) is also seen. The cilia (cl) emerge from basal bodies at the top of the cell into the tracheal lumen together with a nmnber of microvilli (mv) beneath which a dense structure, the terminal web (tw), is resolved. The cells are attracted by the terminal bar (tb). The mucous cells (G) are filled with mucin granules which are about to be discharged at the cell surface. Magnification 6,000X 118 CILIATED EPITHELIA ULTRASTRUCTURE bound ]iy these membranes. The mucous state when the upper part of the mucous cell granules appear, howe\-er, within the smooth is transformed into a goblet filled by numer- membranes of the Golgi apparatus where a ous large mucous granules. A certain fusion heavy accumulation of various sized mucous of mucous granules occiu's intracellularly granules can be identified previous to the before they are discharged at the surface of 1 ^'U 0.2J1 #' •TmiM ifUfr .*' Lff. ^ Fig. 3. Each cilium emerges into the lumen from a basal body. Rootlets (r) and a lateral fibrillar projection (arrow) secure the ciliary filaments in the luminal part of the cell cytoplasm. The lateral ciliarj' filaments are continuous from the ciliarj^ body to the tip of the cilium. The central filaments cannot be demonstrated within the basal body. Magnification 39,000X Fig. 4. Cross sectioned cilia display a ring of nine paired peripheral filaments and two central single filaments. The plasma membrane which shows up as a .single dense line is well preserved in all but two cilia. Magnification 100,000X 119 ELECTRON MICROSCOPY the epithelium. The nucleus of the goblet cell is pushed to the base of the cell because of the heavy accumulation of mucous gran- ules. After the discharge of the mucin, the cell resumes its resting columnar shape and a new production cycle of mucin starts again (Figs. 1,2). Serous cell. The structure of the serous cells is quite variable, depending on what organ the cells are located in. In the trachea only few serous cells have been recognized, but in the fine bronchi and in the hronchioles they become quite abundant. Their cyto- plasm has an abundance of RNA-particles, presumably related to the ability of the cells to secrete enzymes. In addition, a large number of cytoplasmic vesicles indicate that the cells are turning out fluid or substances dissolved and transported within the mem- brane-bound vesicles. It is believed that these cells play an important role in the mechanism involved in pulmonary edema. In the oviduct of some mammals (man not included), the serous cells display a large Fig. 5. A schematic representation of the fine structure of the mammalian cilium. (After Rhodin and Dalhamn, 1956). mmiber of secretion granules. They appear first within the Golgi zone and migrate from here to the cell surface where they are dis- charged without previous fusion with one another. The granules seem to have some nutritional or enzymatic relation with the ovum when it passes along the oviduct to the uterus. The surface of the serous cells is characterized by a number of microvilli, the size of which is smaller than either the cilia of neighboring cells or the brush border ex- tensions seen on the cells of the intestine or the proximal convolution of the nephron in the kidney. The nucleus is located in the center of the cell and is not dislocated during the secretory cycle, the latter being struc- turally less obvious than what is noticed in the mucous cells. Basal cell. The basal cells are always lo- cated near the basement membrane upon which most of the cells of the ciliated epi- thelia rest. The basal cell is round or slightly elongated and does not reach the surface of the epithelium. Its cytoplasm contains small fibrils of unknown function. Eventually, the cell migrates to the upper part of the epi- thelium where it gains contact with the sur- face and starts to differentiate into a cell type which is ready to replace any of the two kinds of cells that dominates the epi- thelium, the ciliated and the mucous cell (Fig. 6). Brush cell. It is probable that the brush cell represents a basal cell which has recently moved to the surface and started to differen- tiate into a ciliated cell. This is quite obvious in the trachea, where several features of the brush cell are identical with the basal cell. The surface of the brush cell is covered with a large number of brush border-like exten- sions. In man, a dense plate is found at a distance of about half a micron beneath the surface of the brush cell, presumably the site of differentiation of the ciliary basal bodies. Another typical feature of the brush cell is the clustering of small mitochondria beneath this plate in the brush cell of the 120 CILIATED EPITHELIA L LTRASTRUCTURE n f'r # n% ^ :i. ■" **«^, nil ^ nu « \ rm ^^ I ■* % * A*^ • IN •^. B ;\ % K ■\. V ,♦*. I . #««(» - (C « BM Fig. 6. Detail of Fig. 1 showing four basal cells (B). The nuclei (nu) are large and prominent features of these cells with several darker bodies, the nucleoli, the latter a possible indication of mitotic activity. The mitochondria (m) are mostly aggregated beneath the nucleus. The basal cells rest on the basement membrane (BM). The in- tercellular spaces (I) are sometimes the site of lymphocytes (L). Magnification 5,600X human trachea. In the efferent ducts of the that they are obviously secretory cells. The human testis, the brush cells display an siu'face structures are predominantly of elaborate rough-surfaced endoplasmic reticu- brush border type, but resemble more lum, a multitude of freely dispersed RNA- those found in the kidney than those particles, and a large Golgi zone, indicating of the intestine. IVIoreover, a highly de- 121 ELECTRON MICROSCOPY Iji mv Fig. 7. Detail of Fig. 1 showing the top part of a brush cell with its numerous and slender microvilli (mv). Three cilia (ci) are seen, possibly indicating that the brush cell is being transformed into a ciliated cell. Mitochondria (m) are more abundant than in ordinary ciliated cells, and particularly evident is the large number of quite small mitochondria and small vesicular structures in between. The arrow points to a dense structure with a light center which is reminiscent of a cross sectioned basal body. A developing cilium? The dense line beneath the microvilli, the terminal web (tw), is present only in relation to the microvilli. Magnification 17,500X 122 COLLOIDS, LYOPHOBIC veloped system of surface invaginations, reminiscent of tubules, speaks for the fact that structural evidence for micropinocytosis is present. It is, therefore, concluded that the brush cells of the respiratory tract and those of the male reproductive tract are func- tionally different, although the structure of their surfaces is almost identical (Figs. 2, 7). Ciliated cell. The ciliated cells have a cytoplasm which contains only a small amount of RNA-particles and endoplasmic reticulum. These cell organelles are probably used only for maintaining the restricted amount of protein that is synthesized for metabolic processes within the cell. The number of cilia per cell is large; in the rat trachea, it amounts to between 250 and 300. The cilium is covered by the plasma mem- brane and its interior represents an exten- sion of the cytoplasm of the cell. Within this cytoplasm is a number of distinct fibrils oriented longitudinally. They originate from the basal body which is located intracellu- lar ly below the level of the cell surface. There are two central single filaments and nine pe- ripheral double ones. They all join in the tip of the cilium. The central ones divide and split within the basal body and wrap around its central core. The peripheral filaments ex- tend below the basal body and terminate at various levels in the upper part of the ciliated cell as ciliary rootlets. Beneath the basal bodies are clusters of mitochondria, the carriers of enzymes in all cells. It is believed that the filaments of the cilium are contrac- tile. The motion center is represented by the basal body, and the energy recjuired for the contraction is derived from the nearby mitochondria. It has been demonstrated that the ciliary beat can occur as long as the con- tact between the basal body and the cilium proper is maintained. Physical damage, which involves a break between the cilium and the basal body will, therefore, stop the ciliary beat. However, it has also been shown that toxic gases as well as cigarette smoke at a certain concentration stop the Fig. 8. A schematic representation of the col- umnar ciliated epithelium of the rat trachea. Con- trary to that of man, this epithelium lacks an evident layer of basal cells (BC) although some may be seen in between the bases of the ciliated (C) and the non-ciliated cells. Among the non-cili- ated cells are found mucous cells (C) in various states of mucous secretion, and brush cells (BRC). The basement membrane is thinner in rat than in man. (After Rhodin and Dalhamn, 1956). ciliary activity. Furthermore, infections like influenza damage the ciliated cells so dras- tically that eventually these cells die and become sloughed off. They are then replaced by basal cells which develop into brush cells and from there into cihated cells (Figs. 3, 4, 5). REFERENCES Fawcett, D. W., and Porter, K. R., "A study of the fine structure of ciliated epithelia," /. Morph., 94, 221 (1954). Rhodin, J. and Dalhamn, T., "Electron micros- copj- of the tracheal ciliated mucosa in rat," Z. Zellf., 44, 345 (1956). Rhodin, J., "Ciliated epithelia," Int. Rev. CytoL, 10 (1962). Johannes A. G. Rhodtn COLLOIDS, LYOPHOBIC The Colloidal State The colloidal state is essentially that state in which matter exists with at least one dimension in the size range 10~^ to 10~^ cm. In this state can be included large molecules 123 ELECTRON >IICROSCOPY such as proteins, enzymes, viruses and high 5G00 A) the maximum resohition obtainable polymers, which ha^•e molecular weights in is of the order of 2000 A, while with ultra- the region of 1,000 to several million and violet light resolution of the order of 800 A exist in molecular solution, and smokes, may be obtained; such resolution is totally mists, gels, etc.. In this size range, many inadequate for the examination of most col- inorganic materials can be prepared as two- loidal dispersions. However, the wavelength phase systems of small particles in liquid of an electron beam produced at an accelerat- media, and are spoken of as colloidal dis- ing potential of 80 kV is 0.043 A and there- persions or sols. The latter groups are usually fore theoretically resolution of the order of termed hjophobic colloids, and it is with this one A unit should be possible. Final resolu- class of colloidal material that this article is tion, however, is limited by the difhculty of mainly concerned. correcting the lens aberrations and with The most important properties of a lyo- earlier microscopes the resolving power was phobic colloidal dispersion are: only of the order of 30 to 50 A. With many (a) the size and shape of the particles, modern instruments the resoh^ing power is and whether the system is dispersed or floe- of the order of 5 A and hence it is possible to culated, resolve objects of the order of atomic di- (b) the chemical structure of the particles, mensions. (c) the nature and structure of the sur- £„_g Experimental Techniques (d) the mode of nucleation and growth of Preparation of Specimen Supports. the particles, and the possible production of The essential criteria for a supporting mem- monodisperse sols, brane are that it should be rigid enough to (e) the electrical charge on the surface withstand manipulation, remain stable in (electrical double layer), and its relation to the electron beam during examination and stability. have a thickness of the order of 100 A. Many Moreover, in phenomena such as nucleation, types of materials have been suggested as growth, coagulation and aging of sols the supporting membranes (1) and probably the dynamic aspects of the system have to be membranes most commonly used are pre- considered and a knowledge of changes in pared from "Formvar" or nitrocellulose. The the system with time is required. The elec- "Formvar" or nitrocellulose membrane is tron microscope may be employed to obtain formed on a dish of distilled water and then answers, or some of the answers, to all these picked up on the surface of a copper mesh questions with the exception of (e). It must grid (2). These films, however, are not suffi- be remembered, however, that as specimens ciently stable for high -resolution work; un- are subjected to a high vacuum (ca. 10~^ mm less they are carefully prepared, they have a mercury) in the electron microscope they tendency to drift under the influence of the cannot be investigated in their natural liciuid electron beam. Greater stability can be environment. The exposure of the specimen achieved by evaporating a thin layer of car- to a beam of high energj^ electrons also means bon on to the plastic film, but care must be that suitable precautions must be observed taken not to increase the support thickness to prevent heating, with subsequent sublima- beyond the limits required for high resolu- tion or decomposition of the specimen; tion. examination of large specimens may be pre- One of the most stable supporting mem- cluded by this factor. branes can be obtained by evaporating ear- In the case of an optical microscope, when bon onto carefully cleaned glass slides or viewing by white light (average wavelength freshly cleaved mica. On immersing the slide 124 COIJ.OIDS, LYOPIIOBIC in water, the carbon film floats on the sur- rectly, concentration can be achieved by- face and can be transferred to the grid. If means of electrodecantation (3). the colloidal dispersions are spread on the Deposition of Sols on Supporting glass slide and allowed to dry before applying Membranes. The simplest method of trans- the carbon, the particles remain embedded in ferring a sol sample to the supporting mem- the film on removal and can be used for direct brane is by the use of a fine loop of platinum viewing. Although carbon forms a very wire. The wire can readily be cleaned by stable film it is not suitable for the examina- flaming before transference of the specimen, tion of all materials. Electrostatic effects are An alternative method, often useful for often encountered which make it difficult to quantitative measurements, is to spray the obtain drop adhesion and solutions contain- sol on to the supporting membrane by means ing surface-active agents usually disrupt the of a nebulizer. Freeze-drying of the sample film; in these cases nitrocellulose-carbon films is often useful and this can be carried out are the most useful with the liquid applied simply by placing the grids on a copper block to the nitrocellulose side. immersed in a freezing mixture; the aqueous Silicon monoxide has also been used to drops then freeze immediately on making obtain a stable supporting membrane, with contact with the supporting membrane, little structure (2). Specimen Contrast. In order to obtain Preparation of Colloidal Dispersions, an image of the particle in the electron mi- Where colloidal dispersions are produced by croscope, electrons must be scattered out of the interaction of two ionic reagents, the the field so that they do not reach the pho- sol produced usually contains considerable tographic plate. The amount of scatter gen- quantities of extraneous electrolyte. This, if erally depends upon the atomic number and left in the sol, tends to crystallize on the sup- the density of the specimen, and it is for this porting membrane and the resultant crystals reason that materials composed mainly of may be confused with the colloidal particles carbon, hydrogen, nitrogen and oxygen, i.e., during examination. Even if actual crystal- of the same composition as a nitrocellulose lization does not occur poor backgrounds supporting membrane, are difficult to ob- may result, or flocculation of the sol may serve. In the latter case the specimen con- occur due to the high concentration of elec- trast can usually be increased by staining or trolyte reached during evaporation . Removal by shadowing with a heavy metal vapor such of any extraneous electrolytes is therefore ^s that of chromium, gold or uranium in a advisable either by dialysis or electrodialysis; ^^Sh vacuum. care must be taken, however, not to remove ^^^'"^^ ^^ ^^^ low penetrating power of stabilizing potential-determining ions. Small electrons it is necessary to use very thin samples may be rapidly dialysed against specimens if mtenor details are to be ob- distilled water in cellophane dialysis sacs; served (see later). With crystalline specimens f i.i-i- 1 r-i- considerations other than random scatter lor electrodialysis a number of simple pieces , ... c , , , , ., 1 1 . V have to be taken into account (2). ot apparatus have been described which can „ , . /• r- i, . , , « . , , ,., , /..N T., , Kesolution ot Colloidal Particles. In be readily constructed (3). Electrolyte may j- ^ • .• r n -j i ^- i . , , , ^ ^ -^ -^ direct examination of colloidal particles, also be removed by passage of the sol ^^^jy ^^e two-dimensional aspects of the through a suitable ion-exchange resin pro- panicle are seen. In this connection it was vided that precautions are taken to avoid realized by von Borries and Kausche (4) that contamination of the sol by particles or crystalline colloidal particles, which should complex ions from the resin. be bounded by plane faces intersecting in WTiere sols are too dilute to be used di- geometrical lines, should be revealed as well 125 ELECTRON MICROSCOPY ..^..H. Fig. 1. (top) Diagram illustrating resolution of the shape of a colloidal particle. The shape of the particle can only be recognized if 6 > 5. (bottom) Comparison of the size of particles of different shape required for resolution of that shape. (After Borries and Kausche). defined shapes in microscopes of infinite re- solving power. Such a condition cannot be reahzed in practice, however, and the efi^ect of finite resolving power has to be considered, von Borries and Kausche (4) supposed that the geometrical boinidaries of the object appeared in the image as boimdaries of finite width, within which the intensity fell con- tinuously from that of the particle to that of the background. The effective width of the boundary was considered to be twice the resolving power, 5, of the microscope and the physical boundary of the particle at a point midway between the particle and back- ground intensities. As a consequence of the finite resolving power the image boundary at the intersection of two straight lines is rounded, and it was assumed that the curva- ture was such that the arc of a circle was tangential to the intersecting edges at a dis- tance 5 from the point of intersection ob- tained by geometrical construction (see Fig. 1). Thus recognition of shape is only possible if the length of the straight portions of the edges, b, is greater or equal to 8. li h < 8 the particle would appear circular and in fact many early workers found that small colloidal particles appeared circular, an effect often due to lack of resolving power. ( )n this basis it is clear that the exact shape of a particle having less than six corners should be easier to recognize and therefore ti'iangu- lar particles should be recognizable as such at much smaller dimensions than the hex- agonal type. Conversely, octagonal plates must be seven to ten times larger than the triangular particles in order not to appear circular. The sizes of particles, relative to a triangle, required for the resolution of a defi- nite shape are illustrated schematically in Fig. 1. In order to test the resolution of an elec- tron microscope it is useful to have a suitable test object, and it has been found (5) that silver and silver iodide sols can be prepared which contain particles having sizes ap- proaching the limits of present-day micro- scopes. Particles having dimensions of the order of 5 A can be clearly resolved from the background (Fig. 2) ; the fact that these par- ticles could be reproduced on separate photo- graphic plates clearly established their identity as colloidal particles. The limita- tions in resolving power appear to be mainly governed by chromatic errors (i.e., stabiliza- tion of high tension) and lens aberrations. The precise measurement of particle sizes be- low 10 A becomes difficult due to phase con- trast effects at this level of resolution. A suitable test object for high resolution work is also found in the case of metal phthalo- cyanines; for example, the metal-bearing «> « ''.•'«,'*« ** * * ♦*'* » ♦-1 * . • '. ' '/,♦.•.■",', •• " *-'♦ • .'• • . ,' ; . .*. *• " • . •;■.•■'*...•**. ,*...-. -,•«•♦■.« « . ■.'•,•?.' "", "* *. ♦"••■• *• * * ' ■■ ■ '■' .*•. • * •'• . ' l1<1^i^ *■■ • • *;' • '♦ Fig. 2. Electron micrograph of silver iodide sol of small particle size. 126 COLLOIDS, LYOPHOBIC planes of platinum phlhalocj-aninc can be resolved in the electron microscope and are found to be 11.97 A apart (6). It is also pos- sible to check resolution by means of Fresnel fringes which are visible when' the objective lens is slightly off -focus (7, 8). The use of image intensifiers with the electron micro- scope may well prove useful in the future in the field of high-resolution work. Shadow Casting. Normally, only a two- dimensional aspect of the colloidal particle on the grid can be obtained. This is insuf- ficient for many purposes particularly if the 3-dimensional shape of the particle is re- quired. In order to enhance contrast and obtain an approximate idea of the vertical height and shape of the particle, shadowing with a heavy-metal vapor such as that of gold, platinum, chromium or uranium may be employed. The metal is evaporated onto the sample in a high vacuum at a suitable angle; usually a special evaporator unit is needed to meet the strongest requirements of high resolution work (1). From the known angle of shadowing and the length of the shadows, a three-dimensional picture of the particle shape can be built up. These shapes can be checked by constructing models and shadowing with a beam of light. IMore reli- able information can be obtained if the ma- terial is shadowed in two directions at right Ofn Fig. 3. Colloidal silver iodide particles sha- dowed with uranium at 50°, a) and c) in one direc- tion, b) and d) in two directions at right angles. Fig. 4. Diagram of particle models based upon shadowed micrographs of the type illustrated in Fig. 3. angles, but it is essential that the shadowing be very light; overdeposition of metal com- pletely obliterates the first shadow. In Fig. 3 micrographs of specimens shadowed with uraniimi at 50° in one direction only and with uranium at 50° in two directions at right angles are shown. From these micro- graphs particles were found to have shapes such as those shown diagrammatic ally in Fig. 4. Replication. One of the most useful tech- niques for determining the exact shape of particles and also their surface structure is replication; this technique is, moreover, in- valuable for the study of specimens which under normal conditions are decomposed by the action of the electron beam. The tech- nique is carried out by depositing samples of the sols either on clean glass slides or freshly cleaved mica and allowing them to dry. A thin film of carbon is then evaporated on to the slide; normally it is advantageous at this stage to shadow very slightly with a heavy metal vapor such as chromium. The carbon film is then removed from the slide using as a hquid substrate a solvent for the embedded particles. When this solvent is a concentrated salt solution it is advisable to follow by wash- ing with more dilute solutions of the salt and then to give a final rinse in distilled^ wa- ter. For the best results from the replication technique it is essential that the films em- ployed should be extremely thin {ca. 100-300 A). 127 ELECTRON MICROSCOPY Fig. 5. Carbon replicas of silver iodide parti- cles shadowed with chromium at 50°. Results obtained by the replica technique with subsequent shadowing are illustrated in Fig. 5. Distribution Curves. It is possible from electron micrographs of a field containing a large number of particles to determine not only their shapes but also the frequency distribution of diameters, or appropriate di- mension. Once these factors are established the surface area of a sample may be obtained, and from a knowledge of the physical density, the weight of the particles. Thus a distribu- tion curve can be made by plotting the fre- quency of appearance of a certain diameter against that diameter. A typical example is given in Fig. 6. Strictly, a number of photo- graphic plates should be taken of different fields and several thousand particles meas- ured in order to obtain a truly representative curve. In practice, however, counts are usu- ally made on 300 to 400 particles; for reason- able representation it is essential to take several micrographs of different parts of the field. Such a curve enables information to be obtained on the degree of polydispersity of the system with respect to size and shape and can be of great assistance in following rate processes such as nucleation, particle growth and coagulation. The shape of the size distribution curve depends on the rates of the nucleation and growth processes (see later). In general there is good agreement be- Iween the size of particles determined by electron microscopy and those determined by other methods. Determination of Absolute Particle Number per unit Volume. Two procedures can be employed for this determination. In the first the sol is sprayed on to the grid in the form of fine droplets using a high-pres- sure nebuUzer (9, 10). Under favorable con- ditions the drops are clearly visible and assuming the diameter of the dried drop to be the same as that in the spray, the drop volume can be estimated. A typical drop formed by spraying a polystyrene latex sus- pension is shown in Fig. 7. Thence from a count of the number of particles contained in the drop the number of particles per unit volume of the original sol can be calculated. In the second method it is essential for accurate results that a monodisperse sol should be used. Electron microscopy can be used to determine the diameter, or in the case of non-spherical particles the appropri- ate dimension, and the volume of a particle 40 30 % PARTICLES 20 lO - ! j— - ! I 1 i ! r^ ■ ! 1 1 — ■ i i i ! j 1 II; ill! ! — ! — 1 ' i '— ! ! !—l ! 1 1 1 . 1 300 600 900 I200 I5CXD PARTICLE DIAMETER IN A° Fig. 6. Particle size distribution curve for a sol of silver iodide. 128 COLLOIDS, LYOPIIOBIC calculated. Then if the particle density is known the weight per particle can be calcu- lated and if a subsidiary determination is made of the weight concentration of the sol used, the number of particles per unit volume can be calculated. Both methods have been used extensixely, but in some cases care must be exercised in applying these methods since it is not easy to determine drop diameters accurately, and the second method may yield spurious re- sults, either because of particle shrinkage (beam intensity too high), or because the particles do not possess a well-defined geo- metrical shape. A comparison has been made (11) between values for particle numbers per unit volume determined by electron microscopy, and those determined by other methods, such as, direct ultramicroscopic counting using a flow method, turbidity measurements and counts using a haemocytometer cell. The results of the comparison which was carried out using polystyrene latex particles are given in Table 1. In general it was found that the dry- weight method yielded results very close to those determined by other methods, whereas the spray method tended to give rather high results in terms of absolute numbers. A Table 1. A Comparison Between Particle Numbers per ml Determined by Electron Microscopy and by other Methods Polystyrene Latex Particles Method 0.216 m Diameter 1.029/1 Diameter Electron Microscopy, spray method Electron Microscopy, dr}^ weight Particle counter (flow method) Turbidity measure- ments Haemocytometer 5.60 X 10'^ 1.93 X 10" 1.90 X 10'3 2.53 X W 1.76 X lO'i 2.01 X W 1.26 X W 3.40 X 1011 method for the direct application of micro- drops of reproducible known volume is clearly desirable. Processes Involved and Destruction in Sol Formation Fig. 7. "Droplet" of polj'styrene latex parti- cles obtained by spraying with a Vaponefrin nebu- lizer. A variety of methods exist for the prepara- tion of colloidal dispersions of different tj^pes (3) and of these perhaps the most commonly used, and the most studied, is the mixing of two ionic solutions. A typical example is the formation of a sol of silver bromide by mixing silver nitrate and potassium bromide at concentrations sufficiently high to exceed the solubility product ; for a stable sol the stabi- lizing ions Ag+ or Br~ have to be present in certain proportions. In most cases of low- .solubility inorganic materials stable sols can be formed provided that a stabilizing ion is present . The nuclei originally formed in the solu- tion, which are usually very small crystals, grow to form the larger sol particles which are usually known as primary particles. This phenomenon is known as ageing, and is usu- ally explained as the growth of extremely small particles to form larger ones either by regular addition to the lattice, i.e., smaller particles going into solution so that the larger ones can grow at their expense, or by a process of ordering of the disordered lattice 129 ELECTRON M l( .IU)S( .( )I»Y of the particles originally formed. Sols of primary particles are often stable for long periods of time, but addition of electrolyte beyond a certain concentration, or the addi- tion of strongly adsorbed organic ions, causes the particles to clump together (coagula- tion); this process may, or may not, be ac- companied by recrystallization to form even larger particles, depending on the system. The process of sol formation and destruc- tion of the sol either by coagulation or recrystallization can be represented sche- matically in the following manner: Ions Nuclei growth (ageing) Large crystals (ageing)/^ J Primary particles \ Coagula Most of the stages represented in this scheme can be investigated by electron microscopy and will therefore be considered separately. Nucleation. Nucleation can be defined as the formation of a discrete particle of a new phase in a previously homogeneous solu- tion. Nuclear or amicronic sols contain par- ticles which cannot be resolved as discrete entities in the ultramicroscope. They can be prepared from many materials, e.g., gold, silver, silver iodide etc.. Electron micro- scopic examination of these sols shows par- ticles down to 10 A or less (see Fig. 2) which can be clearly resolved. The resolution of these particles constitutes one of the highest resolutions so far achieved with the electron microscope. From the point of view of col- loid chemistry this illustrates the small size range in which colloidal particles can exist and it is of considerable interest that the regular shapes of many of the particles ap- pear to be maintained down to the limits of resolution. Thermodynamically, the smaller particles would be expected to have a larger solubility than the larger ones and thus would be expected to go into solution as ions and deposit on the larger particles to increase their size. Charge would be expected to in- fluence this process (12), but in view of the large value of the free energy of most solid- liquid interfaces it is doubtful whether this does in fact play a significant role. Several theories have been proposed for the mechanism of nuclei formation in dilute solu- tion, of which the impurity, organizer and fluctuation mechanisms appear to have re- ceived the most attention. The impurity theory is based on the idea that nuclei are introduced into the system as foreign bodies, e.g., dust particles; it has been found, for example, that in the preparation of colloidal gold different sols are obtained according to the state of the glass vessel used. However, it was concluded by Turkevich, Stevenson and Hillier (13), who prepared gold sols un- der many different conditions, that impuri- ties were not a variable in their investigation. These authors proposed the organizer mecha- nism to account for the formation of nuclei in gold sols. Their suggestion was that the nucleating agent, e.g., hydroxylamine, grad- ually built up a complex between the gold ions, chemically binding a large number of gold ions and reducing agent molecules into large macromolecules. It was suggested that the latter underwent a molecular rearrange-i ment to give metallic gold and oxidation products of the reducing agent. Some sup- port was lent to this hypothesis by the na- ture of the reducing agents, but there are clearly many conditions under which such a mechanism cannot apply. The fluctuation theory of nucleation is probably that most widely accepted. It is based on the hypothesis that the formation of a nucleus occurs only when a statistical fluctuation of the ionic (atomic or molecu- lar) concentration brings a sufficiently large number of ions together to form a particle of thermod3niamically stable size. This theory has been shown to apply to the forma- tion of colloidal sulfur (14). Studies of Nucleation bv Electron 130 COLLOIDS, LYOPHOBIC Particles N No.of Portlcles Per Unit Volume. Particle Fig. 8. a) Particle size distribution curve for sol and b) nucleation curve obtained there- from. (After Turkevich, Stevenson and Hillier). Microscopy. Two methods can be used to examine nucleation by electron microscopy (13) firstly, quenching the nucleation process at different times and directly examining the samples and secondly examination of the particle size distribution curves of the formed sol. In the first method the reaction can be quenched at suitable time intervals either by addition of a suitable reagent to stop the reaction, or by a large dilution to slow the reaction down by several orders of magni- tude. Thus from a direct examination of the samples obtained at different times, particle size distribution curves can be obtained for each sample, and a nucleation curve con- structed. ,The second method, which is less tediotis than the first, was suggested by Turkevich, Stevenson and Hillier (13) and consists of determining the nucleation curve from the particle size distribution curve of the com- pleted sol. The method is based upon the assumption that the principal cause of spread in particle size of the sol is the spread in time in nuclei formation. Thus particles formed in the early stages of nucleation commence to grow immediately, while nuclei formed later have smaller sizes corresponding to a shorter growing time. In the final sol therefore the particles first formed have the larger size and the size distribution curve can be con- sidered as a distorted image of the nticleation curve. Thus if a particle size distribution curve of the type shown in Fig. 8a is con- sidered, particles of diameter Dx may be chosen, which correspond to the diameter at- tained by these particles at a time tx on the nticleation curve (Fig. 8b). Thus at a time tx there are Nt^ particles per unit volume, a number which can be expressed as the frac- tion of the number of particles eventually formed at infinite time, i.e., N{t). The total number of particles formed up to time tx is represented by the area shaded in Fig. 8a or the number of particles with diameter greater than D{tx) is given by. NiO = niD) dD 0(tx) Tiu'kevich, Stevenson and Hillier (13) found that the growth of gold nuclei was given by an eciuation of the form / = (a — log D)/b, where a was a function of the rate constant and of the time at which an arbi- trarily selected reference particle formed; h was proportional to the rate constant. From independent observations of a and h, nuclea- tion curves w^ere constructed from particle size distribution curves. The Growth Process. The formation of a sol involves two processes, formation of nuclei and growth of nuclei. If the formation of nuclei is slow and growth rapid the sol consists of a small number of large particles; if nuclei formation is rapid and growth slow the result is a large number of small parti- cles. When both processes are slow, a broad distribtition of sizes is obtained. In nuclea- 131 ELECTRON MICROSCOPY tioii the number of particles formed as a f miction of time is studied, whereas in the case of growth the rate of increase in the size of the particle with time is the important factor; strictly, in order to study growth the number of nuclei should be kept constant. One method of maintaining this condition in practice is to add nuclei to a slightly super- saturated solution of the growing species. The growth process which appears to have been investigated in most detail by electron microscopy is that of colloidal gold. Turke- vich, Stevenson and Hillier (13) took ad- vantage of the fact that in a slightly acid solution of chlorauric acid and hydroxyla- mine hydrochloride, in a very clean closed vessel, colloidal gold was not produced until a sufficient number of nuclei were intro- duced. Thus when the growth medium was inoculated with nuclei, the chlorauric acid was reduced by the hydroxylamine and the metallic gold was deposited only on the nuclei; hence the nuclei increased in size but not in number. The mean diameter of the resulting particles, Dg , was shown to be given by D, = Dn , 3/M„ -f Mc, M„ where Dn was equal to the mean diameter of the nuclei and Mn and Mce were the respec- tive masses of the metallic gold in the nuclei and ionic gold in the growth medium. From an examination of the particle size distribution curves obtained from the ki- netics of citrate reduction determined chem- ically, it was found that the growth law was of the form dt = kD where k was a constant dependent on tem- perature and reagent concentration but not on particle size. The growth of gold particles in monodis- perse gold sols produced by the action of sodium citrate on chlorauric acid has also been studied by Takiyama (15) using elec- tron microscopy. The size of a particle at a time t (minutes) was expressed by the mole number of one particle, x (mole), as calcu- lated from the mean particle diameter D by the relation, X = 4/37r(D/2)'p/M where p and M are the density and molecular weight of gold, respectively. It was found that the rate of growth was expressed by the equation, — = A;x2/3(x„ — x) , at where x and Xoa are the mole numbers of the particles at a time t and after completion of growth; k was a rate constant. It was found that the growth process was autocatalytic with respect to the surface of the gold parti- cles. The Ageing Process. A lyophobic sol is never stable in the thermodynamic sense and is always proceeding in the direction which involves a decrease in the surface free energy of the solid-liquid interface. Thus there is always present a tendency for the total surface area of the sol to decrease, until a pseudo-stable equilibrium is reached; at this stage the sol consists of dispersed pri- mary particles. The process of change from the initial sol, which may consist of a large number of small particles, to the final "sta- ble" sol which may consist of a small number of large particles is termed ageing; the dis- solving of smaller particles accompanied by growth of larger ones is sometimes termed Ostwald ripening (Fig. 9). The rate of ageing may be slow or fast, according to the condi- tions and the material employed. In the case of barium sulfate, for example, the ageing process even at room temperature is rapid and large crystals are formed ; it is difficult to prepare a very stable finely dispersed sol. On the other hand, in the case of silver iodide, sols of finely dispersed particles can be pre- pared which are stable for several years. The 132 COLLOIDS, LYOPHOBIC "^^:^^ -o ^^r-^ ""c^.. -. ' "-c"-"" . r* f*. X> ^ " ' ■ > V<' ;> '. "^^ VHI ' • r ■ ( , ^ ^. V ' f""' ' _^- , ' '">o ■G> • , ^ ■; • i -; Lf^u ■ »^ ^. ^ a i ^ J ^ ^^8 ABC Fig. 9. Sequence of electron micrographs of carbon replicas showing Ostwald ripening of silver iodobromide emulsion crystals in a solution containing gelatin and ammonium bromide at 50°C. a) immediately b) 5 minutes, and c) 20 minutes after mixing. (By courtesy Messrs Ilford Ltd.) ageing process is very important industrially, principally in the production of photographic emulsions, where the nuclear sols produced in the presence of gelatin are allowed to "Ost- wald ripen" before coating onto plates or film base. The ageing process in poh^disperse systems is usually considered to involve the smaller particles going into solution and the larger particles growing at their expense. An alter- native explanation is that coagulation of the small particles occurs followed by recrystal- lization of the coagula to form regular particles. Most evidence would appear to favor the former mechanism but in some cases mosaic crystals have been found (see for example Fig. 17a) which would tend to favor the latter. It is possible that in prac- tice both mechanisms occiu' with the former usually being the predominant one. Electron microscopy forms a suitable method for the examination of the ageing process since both the average size of the particles and the number present per unit volume of sol may be evaluated at a given time. Moreover, from the shape of the par- ticles formed during the ageing process, it is possible to tell whether growth of the parti- cles occurs preferentially in certain direc- tions. A detailed study of the ageing of silver bromide sols has been carried out by Kolthoff and his collaborators (16). An interesting example of the ageing proc- ess is found in the case of vanadium pentox- ide sols. The sol particles formed in the initial sol, e. g., in a Biltz sol (17) have been found to be small needles several hundred ang- stroms long and 140 A thick. On ageing these are transformed into fibrous crystals several microns in length. In detailed studies on the ageing process (18, 19) it was found that the large numbers of small needle-like particles were redistributed to give smaller nmnbers of large filaments; growth was attributed to recrystallization of the fibrils. Formation of Monodisperse Sols. Inti- mately connected with the study of nu- cleation and growth is the problem of pro- ducing monodisperse sols. The latter may be defined as sols in w^hich all the particles con- tained therein have exactly the same size and shape. The conditions for the preparation of monodisperse sols, which are verj^ important from the viewpoint of colloid chemistry, have been investigated in detail by LaMer and his collaborators (20), and may be il- lustrated by consideration of Fig. 10. Thus if a slow chemical reaction occurs which con- tinuously generates molecules of a disperse phase, the concentration of these molecules increases steadily, passes the point of satura- 133 ELECTRON MICKOSCOPY Concentration o( Otspcrtc PhQtc in Solution. Fig. 10. Schematic diagram illustrating the mode of production of monodisperse sols. (After La Mer). tion A and exceeds at B the level at which the rate of niicleation becomes appreciable. When the rate of production of molecules is slow, however, the sudden appearance of nuclei relieves the supersaturation so rapidly and effectively that the region of nucleation (II) is restricted in time and no nuclei are formed after the initial outburst. Hence the nuclei produced grow uniformly by a dif- fusion controlled process (region III) and a sol of monodisperse particles is obtained. If the initial solutions used are not very dilute, then the rate of production of mole- cules becomes so rapid that their concentra- tion in solution continually exceeds the saturation concentration (Co) and continuous creation of nuclei in addition to growth oc- curs. Thus a polydisperse sol is formed since the size of any particle depends upon the stage at which it was formed. A luimber of monodisperse sols have been prepared and investigated by electron mi- croscopy and other methods. The formation and properties of monodisperse sulfur sols were investigated by LaMer and collabora- tors (20) apparently without detailed exami- nation by electron microscopy. The mono- disperse sols most widely investigated by electron microscopy are undoubtedly those of polystyrene latex (21, 22) and sized samples of these particles are now widely used for the magnification calibration of electron micro- scopes. Watillon, Grunderbeeck and Hautecer (23) by reducing selenium oxide with hydrazine in the presence of amicronic gold particles produced monodisperse sols of selenium. Ex- amination by electron microscopy showed the particles to be almost perfectly spherical with a deviation from perfect sphericity of about ±2%; the spherical shape was con- firmed by shadowing experiments (see Fig. 11). Selected area micro-diffraction showed the particles to be essentially amorphous. ^Monodisperse barium sulfate sols have been prepared by Takiyama (24) by decom- posing the barium-EDTA complex with hy- drogen peroxide in the presence of am- monium sulfate. The particles were shown by electron microscopy to be spindle 6 Fig. 11. Monodisperse sols (a) electron micrograph of selenium sol particles {bij courtesy of Dr. A. Watillon), (b) carbon replica of silver bromide sol particles shadowed with chromium at 60°, (c) carbon replica of silver iodide sol particles shadowed with chromium at 60°. 134 COLLOIDS, LYOPIIOBIC shaped, and the mole number x per particle was calculated from the equation X = ^7r(a/2)(6/2)2p/M where a and h were the mean length of the long and short axes, respectively, and p and M were the density and molecular weight of barium sulfate. The size of the particles was found to increase with the concentration of the reagents used and the relation between the mole number of the particle and the concentration of the reagents (C) was given by the expression xC" = K where a and K were constants. ]\Ionodisperse gold sols have also been prepared by Takiyama (15) using the reduc- tion of chlorauric acid by sodium citrate. The sol particles thus prepared were found to be almost spherical, the average diameter being 172 A with a standard deviation of 13 A. Silver bromide and silver iodide sols have been prepared in a monodisperse form by Ottewill and Woodbridge (25). The silver bromide sols were prepared by slow cooling of a hot solution of silver bromide, which was slightly supersaturated at room temperature, and the silver iodide sols by dilution of the potassium iodide complex with water. The silver bromide particles were found to be cubes and the silver iodide particles rhom- boids; typical micrographs of both types of sol particles are given in Fig. 11. Effect of Additives on Sol Formation. It is well known that molecules of dyes or surface-active agents often show a preference for adsorption onto particular crystal faces (26) and thus can exert profound influences on crystal shape (27). Support for the idea of preferential adsorption on certain crystal faces has been found in electron microscope studies on the coagulation of sols. For exam- ple, in the coagulation of hexagonal plates of silver iodide by dodecylpyridinium ions (28) it was found that the plates were joined by , .^. a . ^a^u. ;:C ,»Jb Fig. 12. Electron micrographs illustrating the injfluence of additives on the formation of silver iodide sol particles, a) sol formed in the presence of 2.2 X 10"'' M mercaptotriazole, b) sol formed in the presence of 4 X 10" M dodecylpyridinium io- dide, c) sol formed in the presence of 7.74 X 10~^ M dodecylpyridinium iodide. edge to edge adhesion suggesting a preference for adsorption of the ion on the 0110, 1010, 1100, Olio, lOlO and IlOO faces. The influence of different media on the shape of sol particles can conveniently be investigated by electron microscopy. In Fig. 12a are shown particles of silver iodide formed in the presence of 2.2 X 10~^ M mercaptotriazole; comparison with Fig. 12b shows that the crystal form has been altered from predominantly flat plates or tetrahedral particles to rod-like particles. It is advisable when such changes are noticed to carry out micro-diffraction experiments on the parti- cles to determine whether any of the additive has been incorporated into the crystal. High concentrations of surface-active agents, particularly those above the critical micelle concentration, often have a consider- able influence on the sol formation process (28). In the case of silver iodide particles, it appears that a twofold adsorbed layer of sur- face-active agent ions is formed on the particles at the nucleus stage; the particles formed are finely dispersed with many of diameter less than 25 A (Fig. 12c). Owing to the strongly adsorbed layer the ageing proc- ess appears to be retarded and a protected 135 ELECTRON MICROSCOPY nuclear sol is obtained. Further growth is coagula obtained can be determined from slow and the sol shows a fairly narrow parti- micrographs. These are found to vary con- cle size distribution. siderably according to the system and coagu- An important additive to the silver halide lating agent employed, typical examples sols formed for coating photographic plates being edge to edge adhesion of flat plates, is gelatin. An electron microscope examina- irregular clumps, chains and massive "lace- tion of the growth of silver bromide particles like" aggregates of particles embedded in in gelatin has been carried out by Ammann- additive (see Fig. 13). Brass (29), and the effect of a number of other polymers of high molecular weight by ^he Structure of Sol Particles Perry (30). In the latter case the nature of Micro -diffraction Examination. With the polymer was found to have a profound an electron beam, as with an x-ray beam, a effect on the growth process and on the regular crystal lattice acts as a diffraction morphology of the crystals obtained. grating and gives rise to a diffraction pat- Studies on Flocculation. The floccula- tern. Thus the electron microscope may be tion of sols is usually considered to include used as a diffraction camera. In earlier both the process of coagulation, i.e., the ac- machines a different specimen holder was tual adhesion of the particles, and the subse- often used for diffraction but in many mod- quent processes of recrystallization etc. ern machines the specimen is left in the same Coagulation can often occur during the dry- position and hence diffraction and micros- ing down of specimens for electron micro- copy can be carried out consecutively on scopic examination and as such is of purely the same specimen. Moreover, with the elec- nuisance value. Direct electron microscopic tron optical arrangements available it is studies, however, do provide a useful method possible to select a specimen area of diameter of obtaining information on coagulation down to 2,000 A and to obtain a diffraction processes provided precautions are taken to pattern from this region alone. It is also pos- eliminate artefacts occurring during the sible to modify instruments so that diffrac- drying down process. tion patterns are produced under the same Electron microscope studies on the coagu- illuminating conditions as used for micros- lation of silver iodide sols have been carried copy (32). Hence by this means, single col- out by Mirnik, Strohal, Wrischer and Tezak loidal particles can be isolated and diffraction (31) and by Home, Matijevic, Ottewill and patterns obtained directly from them; in the Home (28). The former authors studied the case of thin plates with a cross-sectional dis- formation and coagulation of the sol as a tance greater than 2000 A it is possible to function of time and the latter authors the isolate particular regions for examination, effect of dodecylpyridinium iodide, at vari- Thus with colloidal particles which are single ous concentrations, on the sol formation crystals a symmetrical spot pattern is ob- process. tained (see Fig. 14). If the inclination of the So far electron microscopy does not appear single-crystal specimen to the electron beam to have been employed to carry out quanti- is changed, the diffraction pattern alters ac- tative studies on the kinetics of coagulation, cording to the amount and direction of the but it does form a useful method of confirm- inclination. This effect has been studied in ing whether effects observed in quantitative detail bySuitoandUyeda (33) using lamellar measurements by other methods, e.g., spec- gold crystals. For the detailed interpretation trophotometry, are really to be attributed to and analysis of micro-diffraction patterns coagulation or to other effects such as re- the relationship between the reciprocal lat- crystallization. Moreover, the form of the tice of the crystal and the Ewald sphere must 136 COLLOIDS, LYOPHOBIC K* i Fig. 13. Electron micrographs of varioiLS types of flocculation, a) side to side ad- hesion of hexagonal plates, b) formation of chains, c) formation of clumps, d) "lace- like" aggregates of particles embedded in added surface active agent. be considered, since the section of the former by the latter can be regarded, ap- proximately, as the electron diffraction pat- tern obtained. Apart from the direct use of diffraction patterns to obtain the structm-e of the par- ticles, it is very useful in the colloid chemical field to use the "x-ray" structure of the bulk material, if known, to identify the colloidal particle, or confirm its identity, and to de- termine the orientation of the particle; moreover, the indices of the crystal faces ex- posed can be obtained from the disposition of the spots. Diffraction analysis can also be used to give an indication of the imperfec- tions present in the particle. The main limitation to the use of this technique on crystalline colloidal particles is the thick- ness of the particles since for more than a certain thickness of the particle, which varies according to the nature of the material, too much of the beam is scattered, or absorbed, for a distinct pattern to be obtained. In Fig. lob is shown a selected area micro- diffraction pattern from a thin colloidal par- ticle of silver iodide (thickness ca. 100-200 A) and in Fig. 15a a selected area micrograph of the portion of the particle from which the diffraction pattern was obtained. The pat- tern is that expected for a hexagonal crystal of silver iodide of a = 4.59 A and c = 7.49 A resting on the OOOi plane. The clarity of the diffraction pattern demonstrates clearly the almost perfect crystalline nature of colloidal particles of this type. If instead of a single particle, a field is taken containing a number of small particles then a ring pattern is obtained (see Fig. 14b). Only those planes which satisfy the Bragg equation contribute to the pattern and thus a series of discrete rings are obtained; if only a small number of particles are present the rings are broken up into spots. A typical ring pattern, obtained from a group of silver iodide particles of particle size 300-400 A, is shown in Fig. IG. The angular breadth of the diffraction line depends upon the diame- ter of the particle D and the wavelength of the incident radiation X, the quantity \/D usually being termed the Scherrer breadth. Thus theoretically an estimate of particle size can be obtained from the breadth of the diffraction lines (2, 34). However, other facts 137 ELECTRON MICKOSCOl'Y Electron Beam Electron /O o -3020 Beam \/ B ^-^ 20fO 21 10 1 1 So OOOJ loTo (a) (b) Fig. 14. Schematic representation of micro-diffraction, a) diffraction pattern ob- tained from a single crystal particle, b) ring pattern obtained from a collection of colloidal particles. Fig. 15. Micro-diffraction pattern from a col- loidal particle of silver iodide, a) selected area of particle, b) micro-diffraction pattern from this se- lected area. such as finite aperture of illumination, etc. play a part in line broadening. Dark Field Image Analysis. A useful complement to micro-diffraction experiments is dark-field image analysis. This method was developed by Mollenstedt and others (35, 36? 37) and has principally been applied to lamellar crystals. The technique consists of isolating a single spot on the diffraction pattern by means of an aperture in the plane of the objective diaphragm, from which the part of the electron beam, focused as the spot, passes through the small aperture; the latter is then moved aside from the normal position on the axis into the position for dark field. Thus a dark-field image can be obtained which corresponds to the portion of the crys- tal containing the lattice planes from which diffraction was actually taking place. In this manner each spot of a single crystal diffrac- tion pattern can be studied individually and 138 COLLOIDS, LYOPHOBIC a set of dark-field images obtained which correspond to the sections of the crystal con- taining the diffracting lattice planes. Com- parison of the dark-field images with the bright -field images reveals that each black line in the latter becomes a bright line in the former, and thus the Miller indices of the lattice planes giving rise to the bright lines in dark field can be identified. An in- teresting study using this technique has been carried out by Suito and Uyeda (33) on lamellar colloidal gold crystals, in which they were able to identify the crystal planes giv- ing rise to the striped patterns often observed on thin gold crystals (see Fig. 18). Studies of Internal Structure in Thin Colloidal Particles. Many colloidal parti- cles, when studied by direct electron micros- copy, appear completely opaque, and there- fore any interior details which might be expected to be visible, such as grain bound- aries from mosaic growth, dislocation lines, etc., are completely obscured. Dislocation lines usually appear dark on micrographs, an increase of contrast which is thought to be due to the increased Bragg reflection from the strained region around the dislocation line. Moreover, studies on thin metal foils of aluminum, stainless steel, etc., have shown that such imperfections are readily visible in the electron microscope (38, 39). Very few direct observations have been recorded on the presence of these defects in colloidal par- ticles, although such defects may play an Fig. 16. Micro-diffraction pattern from a group of colloidal silver iodide particles, a) selected area of particles, b) micro-diffraction pattern. important role in the stability of colloidal systems. In fact, observation of these defects in colloidal particles does not appear to be an easy problem and may well mean that the particles are very close to perfect crystals. For observations of interior structure it would appear necessary for the thickness of the particles to be of the order of onlj^ a few hundred angstrom units. Thin crystals may often be prepared by the controlled ageing of nuclear sols; this method is particularly successful in the case of silver iodide and some detailed studies on thin hexagonal plates of this material have been carried out by Home and Ottewill (40, 41). Some crystals were found to exhibit a mosaic appearance, sections of different con- trast, which would correspond to different crystal orientations, being clearly visible (Fig. 17a); extinction contours were also clearly visible. In other crystals dark bands were observed (Fig. 17b) which appeared to be due to the presence of dislocations or stacking faults; these were often seen to migrate across the crystal under the influ- ence of the beam. In the case of thin lamellar particles, such as those of gold, striped or spotted patterns are often obser\-ed; a typical example is shown in Fig. 18. These patterns are thought to arise from diffraction effects due to local curvature of the crystal, that is the sub- strate with which the crystals are in close contact undergoes local curvature in the form of a valley; the crystals follow this curvature and have a common axis of bend- ing along the \'alley. The thin strips which accompany the central one can be considered to arise from reflections which correspond to the subsidiary maxima which surround the main diffraction spots (33). The crystal planes giving rise to these diffraction effects can be identified by dark-field image analy- sis. The displacement of the subsidiary maxima from the main spot, which precisely satisfies the Bragg condition, is closely re- lated to the thickness of the crystalline 139 ELECTKO\ MICROSCOPY Ol M. I ^-H Fig. 18. Patterns formed on lamellar gold crys- tals due to diffraction effects. Fig. 17. Electron micrographs of thin colloidal particles of silver iodide showing the presence of imperfections, a) mosaic crystal, b) dislocation lines and hexagonal dislocation loop. ness of such crj\stals (33). Reasonable agree- ment was obtained between the thickness of particles obtained by this method and that suggested by shadowing experiments. Examination of the Surface Structure of Colloidal Particles by Replication. Closely related to the problem of internal structure of particles is the surface structure, since the history of the mode of growth of a particle is often recorded in its surface, in the form of growth spirals, kink sites, twin plane grooves, etc. Such information can usually best be obtained by an examination of very thin replicas of the particles shad- owed very lightly with a heavy metal ^'apor (42). However, in the case of very thin crys- tals, composed mainly of elements of low atomic weight, it is often possible to detect details of surface structure by direct micros- copy or by lightly shadowing the specimen with a heavy metal vapor. Typical exam- ples obtained with a metal detergent crystal, barium tetradecyl sulfate, are shown in Fig. 19. The spiral terraces of the plate-like crys- tals are clearly visible, the lengths of the shadows indicating the depth of the steps to be of the order of 50 A, i.e. bimolecular, and indicate that growth probably occurs by a screw dislocation mechanism. Similar effects have been observed with single crystals of the paraffin n-nonatriacontane and stearic acid (43). In the latter case the crystals were I ^ t Fig. 19. Electron micrographs of barium tetra- decyl sulphate crystals showing surface structure, a) shadowed with chromium at 50°, b) direct micrograph. particle. This fact has been utilized as a method for the determination of the thick- 140 COLLOIDS, LYOPHOBIC replicated with silicon monoxide and the rep- licas shadowed with palladium vapor. Dis- location centers were clearly visible, indicat- ing that growth had occurred by a screw dis- location mechanism ; the spiral step heights were found to be of the order of 45 A (see Fig. 20). An interesting example of the use of elec- tron microscopy in elucidating the growth mechanism of crystals occurs in the case of silver bromide. It was suggested by Berri- man and Herz (44) that the reason for tabu- lar growth in silver bromide (sodium chlo- ride-type structure) was the occurrence of twinning on the 111 plane; some confirma- tion was found for this hypothesis in that Laue photographs taken by them exhibited six-fold symmetry. Additional support for the mechanism was obtained by Hamilton and Brady (45) in an electron microscope ex- amination of shadowed carbon replicas of tabular silver bromide crystals. The replicas were examined at an angle of 45° to the in- cident beam when the convex and concave intersections at the twin planes were visible on the edges of the crystal. Examples of crystal replicas showing the presence of twin planes are given in Fig. 21. The fact that in the immediate vicinity of a crystal imperfection the lattice has a higher chemical potential than in the more perfect parts means that when the crystal is placed in a solvent preferential attack occurs at this point. Thus, provided the reaction is stopped before extensive solution of the crystal oc- curs, etch pits are formed at the sites of preferential attack. In the case of colloidal particles this technique appears to have been employed primarily on silver halide crystals in an attempt to detect dislocation sites. For example, Hamilton, Brady and Hamm (46) carried out chemical etching-studies on large grains of silver bromide and sih^er bromoio- dide emulsions. Etching was carried out by immersion of the particles, for a hmited time, in a silver halide solvent. The particles were then replicated with carbon, and the replicas Fig. 20. Silicon monoxide replica of a stearic acid crystal, shadowed with palladium, showing growth from a single dislocation and bimolecular steps. {By courtesy of Dr. I. M. Dawson) Fig. 21. Carbon replicas of colloidal particles showing evidence of twinning, a) silver iodide, b) silver bromide. Arrows indicate the position of twin planes. examined after shadowing with platinum- palladium at a 5 to 1 angle. The concentra- tion of chemical etch pits and the geometry of the pits were found to be dependent on the solvent used; potassium bromide gave octa- hedral pits and sodium sulfite and potassium cyanide dodecahedral pits. A general increase in etching on certain faces was found after intentionally straining the grains, but the effect was-not sufficiently strong to establish a one-to-one relationship between etch pits and dislocations. The etching experiments did not provide any conclusive e\'idence that normal grains of silver bromide were poly- crystalline in nature. A micrograph of a carbon replica of a sil- ver bromide crystal etched with potassium cyanide is shown in Fig. 22. 141 ELECTRON MICROSCOPY Fig. 22. Carbon replica of a silver bromide par- ticle etched with potassium cyanide. (By courtesy of Dr. J. F. Hamilton) Stability of Colloidal Particles in the Electron Beam. One of the difficulties en- countered in electron microscopy is that the amount of energy which is transferred from the beam to electrons in the specimen can often be in excess of the chemical binding energies. Hence, precautions must be taken against possible decomposition of the speci- men in the beam. Stability depends on whether, after excitation of electronic energy levels, the substance reverts to its original structure or to a new configuration. Thermal stability is also important, since consider- able temperature changes of the specimen can occur; temperatures of the order of 100°C or more can easily be obtained. In many cases decomposition of the specimen in the beam is rapid, making direct examina- tion impossible; For example, both silver chloride and silver bromide rapidly decom- pose to yield a mass of metallic silver (47, 48); thus replica techniques are normally used for examination of these crystals (49). The interaction of the specimen with the beam, however, may frequently be helpful in studying the decomposition of crystals. It was shown by Sawkill (50) that single crystals of silver azide could be decomposed in the beam of the electron microscope and that the decomposition could be followed in (l(;1ail by examining micro-diffraction pat- terns and electron micrographs taken at various stages. In a similar type of study Goodman (51) has examined the dehydration of single crystals of magnesium hydroxide to magnesium oxide imder the influence of the electron beam. Electron microscope observa- tions have also proved useful in studies on the motion of electrons and holes in photo- graphic emulsion grains (52) and in studies on latent image formation (53). A useful technique for examining dynamic changes in crystals of colloidal dimensions is to employ a cinecamera to record the image obtained on the fluorescent screen of the microscope. The first observations of this type were carried out by von Ardenne (54) using a camera fitted into the microscope, and later by Preuss and Watson (50) using an external cinecamera. In recent studies on the movement of dis- locations in thin metal films (38, 39) and dynamic changes in silver iodide particles (40, 41) cine recordings were made of the phenomena taking place. In this work the fluorescent viewing screen was tilted at a suitable angle and recordings were made by photographing directly through one of the observation windows at microscope magnifi- cations of 40,000 X or 80,000 X. A Kodak cine special camera was used and modified to take a 1 in. f/0.95 Angenieux lens at a work- ing distance of 15-20 cm; a speed of 16 frames per second was generally used. The studies on silver iodide were made directly on colloidal particles. Two types of effects were noticed under the influence of the beam — mobile changes of contrast within the particles and filament growth from the particles. The first effect was obtained with hexagonal plates of silver iodide. Changes of contrast were observed under the in- fluence of the electron beam which were highly mobile and migrated within the par- ticle boundaries at rates dependent on the beam intensity. The nature of these changes 142 COLLOIDS, LYOPIIOBIC and the speed with which they occurred are illustrated in Fig. 23. Particles are shown which have undergone considerable changes within the period of two frames of cine film (He sec). The electron-transparent regions moved very rapidly inside the crystal with- out effecting any change in the external shape, and continued to do so indefinitely under constant beam conditions. With certain types of silver iodide parti- cles, mainly the tetrahedral variety, fila- r i I- lOOO A' H Fig. 23. Sequence from a cine film showing a silver iodide particle undergoing changes of con- trast. ^_OJj^ Fig. 24. Filament growth from a single tetra- hedral particle of silver iodide. ments appeared to be pushed outwards from the interior as the intensity of the electron beam was increased. Fig. 24 illustrates typi- cal filament growth from the interior of a particle. In many cases the filaments ap- peared to be ribbons with widths as low as 30 A; these remained, however, quite rigid and were able to push holes in the support- ing membrane. Some filaments also show^ed well-defined contrasting bands (see Fig. 24). Filament growth was also observed in some rather coagulated regions which received strong electron irradiation. A sequence from a cine film of filament growth is given in Fig, 25. Many other dynamic processes should be amenable to investigation by electron mi- croscopy using this type of technique. General Morphology of Colloidal Par- ticles. The number of substances which can be prepared in the colloidal state is very 143 ELECTROiN MICROSCOPY Fig. 25. Sequence from a cine film showing the growth of filaments from irradiated silver iodide particles. large. Furthermore, the size, shape and struc- ture of particles vary considerably from material to material. No attempt has been made in this article to cover all the work carried out on colloidal particles nor to con- sider in detail the question of general mor- phology. This subject has been reviewed elsewhere (56) and it was concluded that the morphological forms in which colloidal par- ticles are formed could be subdivided into amorphous small particles, small regular forms (spheres, cubes, hexagons, octahedra, etc.), fibers and plates. Most of these forms have been described in this article but the main emphasis has been laid upon the de- scription of techniques which enable any colloidal particle to be examined in the elec- tron microscope. Acknowledgments. It is a pleasure to record my thanks to Mr. R. W. Home for his continuous en- thusiastic collaboration in much of the work described in this article. I should also like to express my thanks to Drs. I. M. Dawson, G. F. Hamilton and A. Watillon for generously supply- ing the micrographs acknowledged in the text. REFERENCES 1. CossLETT, V. E. AND HoRNE, R. W., Vacuum, 5, 109 (1955). 2. Hall, C. E., "Introduction to Electron Mi- croscopy," McGraw Hill Book Co., New York, 1953. 3. Kruyt, H. R., "Colloid Science," Vol. I, Elsevier, Amsterdam, 1952. 4. VON BORRIES, B. AND Kausche, G. A., Kol- loid-Z., 90, 132 (1940). 5. Ottewill, R. H., and Horne, R. W., Kol- loid-Z., 149, 122 (1956). 6. Menter, J. W., Proc. Roy. Soc, A236, 119 (1956). 7. Hillier, J. and Ramberg, E. G., /. Appl. Phys., 18, 48 (1947). 8. Haine, M. E. and Mulvey, T., J . Sci. Instr., 31, 326 (1954). 9. Williams, R. C, and Backus, R. C, /. Am. Chem. Soc, 71, 4052 (1949); J. Appl. Phys., 21, 11 (1950). 10. G^ROVLD, C.B.., J. Appl.PhTjs., 21, 183 (1950). 11. Ottewill, R. H. and Wilkins, D. J., J . Colloid Sci., 15, — (1960); in press. 12. Knapp, L. F., Trans. Faraday Soc, 17, 457 (1922). 13. Turkevich, J., Stevenson, P. C, and Hil- lier, J., Disc. Faraday Soc, 11, 55 (1951). 14. LaMer, V. K. AND Kenton, A. S., /. Colloid Sci., 2,257 (1947). 15. Takiyama, K., Brdl. Chem. Soc. Japan, 31, 944 (1958). 16. Kolthoff, I. M. AND Bowers, R. C, J. Am. Chem. Soc, 76, 1503 (1954). 17. Biltz, W., Ber., 37, 1095 (1904). 18. Takiyama, K., Bull. Chem. Soc. Japan, 31, 369, 555 (1958). 19. Kerker, M., Jones, G. L., Reed, J. B., Yang, N. P., and Schoenberg, M. D., J. Phys. Chem., 58, 1147 (1954). 20. LaMer, V. K., Ind. Eng. Chem., 44, 1270 (1952). 21. Harkins, W. D., /. Am. Chem. Soc, 69, 1436 (1947). 22. Bradford, E. B., Vanderhoff, J. W., and Alfrey, T., /. Colloid Sci., 11, 135 (1956). 23. Watillon, A., van Grunderbeeck, F., and Hautecler, M., Bull. Soc. Chim. Belg., 67, 5 (1958). 24. Takiyama, K., B^lll. Chem. Soc. Japan, 31, 950 (1958). 25. Ottew^ill, R. H., and Woodbridge, R. F., in press. 26. Buckley, H. E., "Crystal Growth," 458, Wiley, New York (1952). 144 CRYSTAL LATTICE RESOLUTION 27. Rehbinder, p., Disc. Faraday Soc, 18, 151 (1954). 28. HoRNE, R. W., Matuevic, E., Ottewill, R. H., AND Weymouth, J. W., Kolloid-Z., 161, 50 (1958). 29. Ammann-Brass, H., Chimia, 10, 173 (1956). 30. Perry, E. J., J. Colloid Sci., 14, 27 (1959). 31. MiRNiK, M., Stromal, P., Wri.scher, M., and Tezak, B., Kolloid-Z., 160, 146 (1958). 32. Reicke, W. D., "Proc. First Regional Euro- pean Conference on Electron Microscopy," 98, Stockholm, 1956. 33. SuiTO, E. and Uyeda, N., "Proc. Interna- tional Conference on Electron Microscopy," 223, London, 1954. 34. Heidenreich, R. D., Phys. Rev., 62, 291 (1942). 35. Mollenstedt, G., Optik, 10, 72 (1953). 36. Rang, O., Z. Phys., 136, 465, 547 (1953). 37. Ito, K. and Ito, T., J. Electronmicroscopy (Japan), 1, 18 (1953). 38. HiRscH, P. B., HoRNE, R. W., and Whelan, M. J. Phil. Mag., 1, 677 (1956). 39. Whelan, M. J., Hirsch, P. B., Hornb, R. W. AND BoLLMANN, W., Proc. Roy. Soc, A240, 524 (1957). 40. Horne, R. W. and Ottewill, R. H., /. Phot. Sci., 6, 39 (1958). 41. Horne, R. W. Ottewill, R. H., "Fourth International Conference on Electron Mi- croscopy," Berlin, Springer Verlag, 1, 140 (1960). 42. Bradley, D. E., Brit. J. Appl. Phys., 10, 198 (1959). 43. Anderson, N. G. and Dawson, I. M., Proc. Roy. Soc, A218, 255 (1953). 44. Berriman, R. W. and Herz, R. H., Nature, 180, 293 (1957). 45. Hamilton, J. F., and Brady, L. E., /. Appl. Phys., 29, 994 (1958). 46. Hamilton, J. F., Brady, L. E. and Hamm, F. A., /. Appl. Phys., 29, 800 (1958). 47. Levenson, G. I. P. AND Tabor, J. H., Sci. and Indust. Phot., 23, 295 (1952). 48. Klein, E., Mitt. Forsch. Agfa, 10 (1955). 49. Hamm, F. A., and Comer, J. J., /. Appl. Phys., 24, 1495 (1953). 50. Sawkill, J., Proc Roy. Soc, A229, 135 (1955). 51. Goodman, J. F., Proc Roy. Soc, A247, 346 (1958). 52. Hamilton, J. F., Hamm, F. A., and Brady, L. E., J. Appl. Phijs., 27, 874 (1956). 53. Hoerlin, H. and Hamm, F. A., J. Appl. Phys., 24, 1514 (1953). 54. von Ardenne, M., Z. Phys., 120, 397 (1943); Kolloid-Z., 108, 195 (1944). 55. Preuss, L. E. and Watson, J. H. L., /. Appl. Phys., 21, 902 (1950). 56. TuRKEvicH, J. AND HiLLiER, J., Atiol. Chcm., 21, 475 (1949). R. H. Otteavill CRYSTAL LATTICE RESOLUTION The ultimate goal of electron micro.scopy is, of course, the resolution of atoms in any structure, and this possibility has been ana- lyzed theoretically by several of the eminent workers. If atoms or molecules are regularly arranged in a crystal lattice there is a much stronger chance of resolved image formation than in the case of two isolated atoms be- cause there are definite phase relationships between electrons scattered from neighbor- ing noncoherent atoms. The resolving power of the best electron microscopes produced in the world is limited first by diffraction error and spherical error to about 2.8 A, and chro- matic error and astigmatism increase this to at least 7 A. At this value it should be possi- ble to observe crystal lattices in crystals with fairly large lattice parameters of the order of 10 A or greater. Great success prior to 1956 in the investi- gation of macromolecular crystals of viruses and proteins by the replica technique had been achieved by Wyckoff and associates at the National Institutes of Health (3). The surface of a needle-shaped crystal of the jack bean protein concanavallin, with molecule weight 42,000, among the smallest thus far replicated, reveals a rectangular net of about 62 X 87 A of particles (30-40 A in diameter) which are not in contact. The most likely next step downward in dimensions would be for crystals of organic molecules of inter- mediate molecular weights, sufficiently thin and properly oriented for direct transmis- sion micrographs. Menter (1) was the first to produce elec- tron micrographs of the lattice planes in crystals. He was particularly fortunate in his choice of metal phthalocyanins, especially 145 ELECTRON MICROSCOPY Fig. 1. A packing drawing of the nickel phtha- locyanine structure viewed along the bo-axis. The metallic atom is black, the nitrogen atoms are line-shaded. (R.W. G.Wyckoff,'' Crystal Structures'^) the copper and platinum derivations of phthalocyanin, a blue dye with a flat ring structine (Fig. 1) and the metal atom in the center. The crystal structure of platinum phthalocyanin may be considered to com- prise fairly widely spaced planes of heavy metal atoms (c^2oi = 11.94 A) embedded in a matrix of the light elements nitrogen, car- bon and hydrogen. The thin ribbon habit of the crystal is such that when supported on a specimen grid the 20l planes are almost parallel to the electron beam. The electron micrograph at a magnification of 1,500,000 (Fig. 2) shows the clearly resolved planes spaced 11.94 A apart. Similarly the copper phthalocyanin shows a spacing of 9.8 A. These parallel lines are the image of the projection of the 20l planes seen edge on. Imperfections are seen sometimes in the form of edge dislocations (Fig. 2) caused by in- complete planes. It has been possible also to resolve the 111 planes of the inorganic crys- tal sodium faujasite, a silicate with the com- position 2 Al2O3-CaO-Na2O10SiO2-20H.,O, with a spacing of 14.37 A. The mechanism of image formation de- pends upon the fact that the ihiii crystals form a cross grating diffraction spectrum. With a 50-mi(*ron objective aperture the spectra contributing to the image from the {)hthalocyanin crystals are (20l), (402), (201) and (402). These spectra recombine with the zero order beam in the image plane and form an image of the crystal grating in accordance with the simple Abbe theory of image formation by a lens. Calculations by Menter suggest that it should be possible to resolve planes with spacings considerably smaller than 10 A providing the divergence of the illuminating beam is made sufficiently small. Figs. 2. (a). Electron micrograph of part of crystal of platinum phthalocyanine showing image of lattice planes 12 A apart. Magnification 1,500,- OOOX. (b). Part of crystal of platinum phthalocy- anine showing edge dislocation. The dislocation line is perpendicular to the plane of the paper. Magnification as before, (c). Sketch copied from (b) showing exact position of extra plane of mole- cules. (Menter) 146 ELECTRON OPTICS If this is accomplished in practice there an extra dark band caused by osmium at- will be new possibilities for studying crystal tached to the double bonds, structures and detecting imperfections. The The RIAS workers have compared sepa- images are poor Fourier projections and fail rate micrographs of multilayers made from to reveal detail obtained by x-ray analysis; Cis, C22; and C36 acids salts. Periodicities of but they do reveal the exact location of im- the bands show ratios of 16 to 22 to 36. With perfections, thus permitting the direct study longer chain barium soaps, the lighter bands of the behavior of dislocations in solids under increase in width while the dark bands (me- a variety of physical conditions. tal ions) remain constant. The width of the Very recently, as reported in the July 4, dark bands is greater than the actual thick- 1960 Chem. Erig. News, it has been demon- ness (2 to 3 A) of the metal ion double strated that electron microscopy can show sheath, but resolution of the electron micro- the actual structure of soap multilayers at scope goes down only to 8 to 15 A. the molecular level. Two workers at the The band image of the micrographs repre- Research Institute for Advanced Stud\^ in sents actual ''multilayer architecture" and Baltimore, Md., have made micrographs that not merely interference fringes. For one picture highly regular sequences of Hght and thing, structures of 40, 80, or 120 double dark bands. And the spacing of these bands layers of soap molecules (made of saturated conforms with the length of fatty acid chains fatty acids) show exactly 40, 80, or 120 that make up the soaps, they find. bands, as the case may be. Moreover, such This combination of electron microscopy details as line dislocations in the micrographs and film techniques is the first observational are noted as in the case of the metal phthalo- method to confirm the double-layer spacing cyanins. of multilayers. Dr. Hans J. Trurnit told the references 34th National Colloid Symposium sponsored , ,^ ^ „. ,,„ , ,;r. t^ 1 1 4 ^01 T^- • • r ^11 • 1 ^1 • 1- Menter, J. W., "Electron Microscopy, Pro- by the ACS Division of Colloid Chemistry, ceedings of Stockholm Conference, Sept. at Lehigh University. The double spacing 1956," Academic Press, N. Y., 1957, p. 88. (previously derived from optical and x-ray 2. Neider, R., Ibid., p. 93. methods) results from a head-to-head, tail- ^- Wyckoff, R. W. G., Koninkl. Nederl. Akad. to-tail arrangement when the layers are laid W^tenschappen, Amsterdam, Proc, B59, 449 down on the surface. Dr. Trurnit and his associate, George George L. Clark Schidlovsky, build the multilayer soap films on methacrylic ester slides. Then they expose DISLOCATIONS IN METALS. See TRANS- sample strips of the slides to osmmm tetrox- MISSION ELECTRON MICROSCOPY OF ide (contrast inducer and fixing agent), im- METALS-DISLOCATIONS AND PRECIPITA- bed them in the polymerizing plastic, and tION d 291 slice them into thin sections (about 500 A.). They could not use free fatty acids because of ELECTRON OPTICS: ELECTRON GUN AND their solubiUty in the plastic bed. ELECTROMAGNETIC AND ELECTROSTATIC Dark bands in the electron micrographs lcinoco correspond to sheaths of metal ions in the The electron optical system of the electron soap layers, when the metal used has a high microscope comprises (a) an electron gun in enough atomic number to give electron scat- which electrons emitted from the tip of a hot tering. Magnesium ions do not register, but tungsten hairpin-shaped filament are ac- calcium, barium, and other such ions do celerated to produce a narrow conical di- show up. Also, unsaturated fatty acids show vergent beam of electron illumination to ir- 147 ELECTRON MICROSCOPY radiate the object; (b) a condenser lens to concentrate the beam onto the object and in- corporate an aperture system to control the angular aperture of the irradiating beam; (c) an objective lens and objective aperture followed by one or two projector lenses to form an image of the object on a fluorescent screen or photographic plate magnified usu- ally between 1,000 and 200,000 times. The condenser lens may be a double one to allow the reduction of the area of illumination to minimize heat dissipation at the object. The objective lens must have the minimum spher- ical and chromatic aberration and astigma- tism consistent with meeting the geometric requirement of allowing adequate space for the object holder and objective aperture, and their manipulation. The projector lenses must be designed to give the widest range of magnification possible without image dis- tortion. The Electron Gun The electron gun used in the electron microscope utilizes a tungsten wire hairpin cathode, an apertured grid and an anode. An example of a typical electrode system is shown in Figure 1. The anode is at ground potential and the cathode at the full 50-100 kv accelerating potential. The grid is oper- ated with a negative bias of a few hundred volts with respect to cathode. To obtain even barely adequate final im- age intensity at magnifications of 50 to 100,- 000 required for high resolution working, the requirements for gun performance are strin- gent. Since the illumination beam angle is limited to give optimum resolving power, the main gun requirement is to give the maxi- mum possible current density per unit solid angle of emergent beam. Langmuir (1937) has shown that a theoretical limit to the cur- rent density per unit solid angle (or "bright- ness") is imposed by the spread in the emis- sion velocities of the electrons from the cathode. The limiting value of brightness (iS) is given by: (3 = pco/TrkT a) where pc is the emission current densitj^, cpo the accelerating voltage, k = Boltzman's constant (8.6 X 10"^ e.V./°K) and T the absolute temperature. It can readily be shown that the current density obtainable in a focused image of the virtual gun source is independent of the mag- nification of the focusing system where the beam angle is fixed. Haine and Einstein (1952) have shown that the electron micro- scope gun will give this theoretical bright- ness, for a wide range of geometrical con- figurations of the electrodes, provided op- timum bias conditions are maintained. A certain choice of geometrical configurations will give a narrower angle of divergence of the beam and hence conserve total current. Thus, increase of the cathode shield diame- ter and reduction of the height of the cathode behind the shield both reduce beam angle and therefore total current. Under such con- ^^ ATHODE SHIELD 7^ / / NODE (a) FLAT SHIELD Cb") RE-ENTRANT SHELD Fig. 1. The typical geometry of an electron gun. 148 ELECTRON OPTICS ditions the optimum bias potential is in- creased but the brightness and hence the image intensity is unchanged. Haine, Einstein and Borcherds (1958) have discussed the use of automatic bias, i.e., the generation of bias potential by the po- tential drop across a resistance connected in series with the high voltage supply. Apart from the simplicity of this arrangement, the negative feedback action of the gun mutual conductance and series resistor gives a high degree of stabilizing action to the beam cur- rent. Further restrictions are imposed on the choice of geometrical configuration to en- sure that the bias potential is maintained at the optimum value. To obtain adequate image intensity for very high resolution working (3 to 10 A), the electron gun cathode must be operated at an emission density of 1 to 3 amp/cm-. This emission current density requires an operat- ing temperature at which the tungsten cath- ode life, as limited by evaporation, is only a few tens of hours (Bloomer 1957). Electron Lenses The requirements of the electron micro- scope are met with magnetic or electrostatic lenses which provide a "bell-shaped" magnetic field or electrostatic potential distributions along the axis. The magnetic lens comprises a solenoidal excitation coil wound on an axi- ally symmetric kon circuit with a gap in the core and an axial hole bored to allow the passage of the electrons. Basically the lenses comprise parallel pole pieces spaced S cm apart and an axial hole diameter D cm with an excitation NI ampere-turns applied be- tween the pole pieces (Figure 2). Little or no advantage is gained by changing the shape of the pole pieces from this simple geometry. The electrostatic lens comprises basically three parallel plate electrodes with axial holes. The outer plates are at ground poten- tial and the central one at the negative po- tential of the electron gun cathode. The elec- trodes are usually shaped to minimize the surface electric field strength to avoid flash- overs, but the lens theory applied to the O CWDI 002 0-03 0-04 005 0-0 6 Vr/CNI)' Fig. 2. Curves showing the focal properties of magnetic lenses as functions of the excitation param- eter. Vr/{NI)\ 149 ELECTRON MICROSCOPY simple sliape gives results of adequate ac- curacy for practical purposes (Archard 1954). The relations between the geometric fac- tors, applied excitation and the focal proper- ties and aberrations have been calculated and measured by many different workers (e.g., Glaser, 1941, Ramberg, 1942, Lenz 1950, Liebmann andGrad, 1951). The results are now well established and it is more rele- vant to describe these results than the meth- ods for deriving them. The results are more complete for the magnetic than for the elec- trostatic lens, but those for the latter are adequate to show its inferiority. All the rele- vant properties are given in a series of uni- versal curves derived from data calculated by Liebmann and Grad (1951). The first order focal properties of impor- tance are the focal length (Jo , for objective and/i , for projector) and focal distance (zq). These parameters can be expressed in the simple universal curves of Figure 2 in terms of the ratios /o/(>S -f D), fi/(S + D), Zo/ (S + D) and the excitation parameter Vr/ (Niy where eVr is the relativistically cor- rected electron energy, and NI the effective ampere-turn excitation of the lens. By effec- tive is meant the ampere-turn excitation drop across the lens gap. The dotted line in the figure represents the thin lens approxima- tion given by / = 25 VriS + D)/{Niy (2) It is seen that the projector focal length passes through a minimum. This is of im- portance in that it determines the maximum magnification which can be obtained with a given stage length. Of paramount importance is the spherical aberration of the objective lens which sets the ultimate limit of resolving power. The spherical aberration is defined by the con- stant Cs where Cs a^ gives the radial error in position of a ray leaving the object at angle a with the axis, the distance being measured in object space. The value of Cs/f is given within an error of ± 10 % for values of the ratio S/D between 0.2 and 2 by the full curve of Figure 3. The 16 Cs/f 14 12 lO ■' " r T~ / \ / 1 y ^ tA / ^^^^0'^'^'^ ^ THIN L Cs/f = ENS APP 5f/Cs■^ ROX 0-2 0-4 0-6 08 lO 1-4 1-6 1-8 f/(Si-D) Fig. 3. Variation of the ratio CJj with// (5 -h Dy. 20 150 ELECTRON OPTICS 4000 (nO . - •■-^*~*^^- 1 .^^ 6000(ni) \ ^^- - r'"~""..---' — 8ooo(ni) 8000(ni) Fig. 4. The resolving power as a function of the pole piece spacing, magnetic field and excitation for 100 KeV electron energies. dotted curve shows an approximation which becomes accurate for weak lenses and is given by: CJf = 5[//(S + D)Y (S) The resolving power of the microscope is theoretically limited by spherical aberration (requiring a minimum aperture angle) and diffraction (requiring maximum aperture angle) to a minimum value (d) at an opti- mum aperture angle a, given by: d = 0.43 C'/^XS/" a = 1.4(X/C.)i/* (4) (5) Figure 4 shows the variation of resolving power under optimum aperture conditions as a function of pole piece spacing (S), mag- netic field strength (Hp) and excitation (A^7) for 100 Kev electron energy. It is seen that for a given field strength a flat optimum oc- curs around a particular value of spacing. A further parameter of importance is the chromatic constant (C<.) which gives the variation of focal length with small changes in the high voltage or lens current Sf = CciSV/V - 257/7) (6) 8f must clearly be kept within the depth 151 ELECTRON MICROSCOPY lo Cc/f 0-9 0-8 0-7 6 ..^l I i j ^ OOI 002 003 004 005 006 O07 008 O 09 OIO \2 Vr/(NI)^ Fig. 5. The variation of the ratio of the chromatic constant (Cc) to the focal length with the excita- tion parameter Vr/(Niy. of focus of the instrument. The value of Cc/f is given by the curve of Figure 5 as a function of the excitation parameter. To ensui'e no limitation in resolving power or contrast, the variation in focal length (5/) due to ripple on the electron accelerating voltage or lens current supplies must be kept below one quarter the depth of focus of the instrument: 5/ < 0.7 dVX This requires a voltage and current sta- bility meeting the requirement: dV/V - 25/// < 0.7 dyxCc For a lens near the minimum focal length (Cc '^ 0.4 cm) and for a resolving power near the optimum ('~2-4 A), the required stabilities are in the region of 2 or 3 parts in a million (Haine, 1960). The magnetic lens has the peculiar property of rotating the image with respect to the object. Astigmatism arises in objective lenses as a result of very small departures from axial symmetry of the pole piece bores or faces. Only the elliptical component of asymmetry is of significance. To an adequate approxima- tion the astigmatic distance between the tangential and sagittal foci (za) is given by the expression: Za = 1005(2 + 3 S/D)Vr/iNiy (7) where 5 is the departure from symmetry. For the astigmatism not to limit the re- solving power, Za should be small compared with the depth of focus. To achieve the nec- essary symmetry tolerance, which may be a few millionths of an inch for optimum re- solving power, represents an almost impossi- ble mechanical engineering task. Fortun- ately, it is possible to correct a small degree of residual astigmatism by the inclusion of a weak cylindrical lens of variable power and orientation. The progress of correction has been discussed in detail by Haine and IMul- vey (1954). Design Considerations The number of stages of magnification in- cluded in the microscope is dependent on the maximum magnification and the range of magnification required. A fairly definite op- timum of three stages can be deduced, giv- ing two stages following the objective. The various factors affecting the choice of stage length and the focal length of the various lenses include the maximum and range of magnification, the minimum being limited by image distortion. The practical design of a lens for given pole piece geometry must ensure an adequate magnetic circuit, ade- quate heat dissipation from the excitation coil and a design which can be manufactured within very close symmetry tolerances. The magnetic design is discussed by Mulvey 152 40 004 OOI ELECTRON OPTICS 0025 Vr /(Nl)^ Niy/vv Fig. 6. The image rotation angle in radius plotted as a function of NI/\/V\- and Vr/iNI)^ (on top scale). (1953). Permissible thermal loading of the coil varies from about 800 ampere-turns per square centimeter of cross-sectional area of coil for a lens without water coohng, to 1200 ampere-turns per sq cm for a carefully de- signed water-cooled coil with no interleaving paper. The values are fairly independent of the wire size used, which may be chosen to give a coil impedance most suitable for the current supplies. It was at one time thought that the great precision of symmetry required in the objec- tive lens required the pole pieces to be manu- factured separately from the main iron shroud and fitted precisely within its bore. That this is not so, and in fact leads to un- necessary complication, has been shown by Haine (1954). Although some manufacturers still utilize separate pole pieces, the tendency is toward the simpler lens made from two pieces of iron. Electrostatic Lenses The data for electrostatic lenses cannot be expressed by such simple means as for mag- netic lenses. The variation of focal length and spherical aberration with the dimensions for three aperture unipotential lenses of spac- ing (*S), central electrode aperture diameter D and thickness T are given in Figures 7 and 8. The objective cannot be immersed in the electrostatic field and it will be seen that the spherical aberration is about an order of ten greater than for the magnetic lens. 153 ELECTRON MICKOSCOPY 4-5 4-0 \ V , s ^ ^ — 1 — '/// \ \ y /// v/ 3b 3-0 2-5 2-0 \\ \\\ ri y/ // \ \ /i-o/ / /''V \\\ /// l'5 l-O ■s. -5 O 1 T/D O 0'2 0-4 0-6 0-8 I-O 1-2 1-4 1-6 IB Fig. 7. The focal properties of an electrostatic lens as a function of its geometry. 90 80 70 Cs/S 60! SO 40- 30 20- lO V s o S / i D \\\ ' U- / T '/ ' // We 1 y-o \ I / \\\ ^!;s ^0\^ 1 — • 1 1 T/O 1-2 1-4 O 02 0-4 0-6 0-8 I-O Fig. 8. The spherical aberration of an electrostatic lens as a function of its geometry. 154 HISTORY OF ELECTRON OPTICS REFERENCES Archard, G. a., "Proc. Internat. Conference on Electron Microscopy," London (1954). Pub- lished by the Royal Microscopical Society (1954). Bloomer, R. N., Brit. J. Appl. Phys., 8, 83 (1957). Glaser, W., Z. Physik, 117, 285 (1941). Haine, M. E., "Proc. Internat. Conference on Electron Microscopy," London (1954). Pub- lished by the Royal Microscopical Society (1954). Haine, M. E., "The Electron Microscope" (E. and F. N. Spon Ltd., London). Haine, M. E. and Einstein, P. A., Brit. J. Appl. Phys., 3, 40 (1952). Haine, M. E., Einstein, P. A., and Borcherds, P. H., Brit. J. Appl. Phijs., 9, 482 (1958). Haine, M. E. and Mulvey, T., /. Sci. Instr., 31, 326 (1954). Langmuir, D. B., Proc. I.R.E., 25, 977 (1937). Levt, F., Z. Angew. Phys., 2, 448 (1950). Liebmann, G. and Grad, M. E., Proc. Phys. Soc (London), 64B, 956 (1951). MuLVET, T., Proc. Phys. Soc (London), 66B, 441 (1953). Ramberg, E. G., J. Appl. Phys., 13, 582 (1942). M. E. Haine FIBERS (TEXTILES). See GENERAL MICRO- SCOPY, p. 343. HISTORY OF ELECTRON OPTICS The history of electron optics starts more or less with the discovery of cathode rays. The earliest experiment of Pliicker in 1859 already showed a rectilinear propagation of these rays. Ten years later, Hittorff dis- covered that cathode rays can be deflected by a magnetic field, and that an axially sym- metrical magnetic field will concentrate such rays. Another ten years elapsed before Crookes demonstrated a better proof of rec- tilinear propagation. The first attempt to calculate the trajectory of charged particles in fields of forces originated with Riecke in 1881. These were rather incomplete, and the first quantitative theoretical and experi- mental studies on the deflection of an elec- tron beam had to wait until 1897 when J. J. Thomson determined the ratio of the charge to mass of the elementary particles carrying the elementary quantity of electricity and confirmed, in a manner of speech, the beauti- ful calculations performed earlier the same year by H. A. Lorentz. The first intentional use of a solenoid for concentration of cathode rays was done by Wiechert in 1899. He used a relatively long solenoid; a short coil for concentration was not used until 1903 by H. A. Ryan. In the same year, the first electrostatic concentration of an electron beam was performed by Weh- nelt. Very extensive calculations of trajec- tories of charged particles in field.s of forces were carried out in 1907 by C. Stoermer. The basic equations of electron optics are im- plicitly contained in the work of Stoermer, although this had not been recognized for many years. The haphazard use of electron optical ele- ments and of calculations was followed by a more deliberate one in the middle twenties of this century. De Broglie's thesis in 1924 on the wave nature of the electron became the foundation of physical electron optics. The formal foundation of geometrical electron optics was set down in 1926 by Busch. Busch actually considered a magnetic coil as an optical lens, and derived the lens equation for it, although he failed to use the coil as an imaging element. The combination of these two discoveries stimulated thinking in two directions. His- torically, the first was the development of ap- plied physical electron optics in the form of electron diffraction. The momentous dis- covery of diffracted electron beams by Davis- son and Germer m 1926-27 was soon fol- lowed by the similar experiment of G. P. Thomson. On the other hand, geometrical electron optical methods were first applied around 1929, when cathode ray oscillographs were built on the basis of electron optical elements. These attempts were pursued simultaneously by Briiche iri Germany and by Zworykin in the United States. The next step in the development of geo- metrical electron optics was the study of ax- ially symmetrical fields as lens elements. 155 ELECTUON MICKOSCOPY These studies started probably around 1929. ber with two-way motion of the specimen At least, this is the date of a patent applica- and an air lock, as well as a photographic tion on electrostatic lenses by I^oll. The chamber for internal photography of the fii-st publications on the study of electro- electron micrographs. This photographic static and magnetic lens elements date from chamber was also provided with an air lock. 1931 when Davisson and Calbick published The first electron micrographs of biological a note on the focal length of an electrostatic objects in 1934 were made on material im- lens and Knoll and Ruska studied the be- pregnated with osmium salts because of the havior of magnetic lenses. Further studies of then prevalent notion that the electron electrostatic lenses were published by Briiche beam would destroy any biological material, and collaborators. Within a year, it was recognized that this Experimental electron microscopy had its impregnation was not absolutely necessary beginning in 1931. The formal beginnings are and that by reducing the total exposure time marked by the patent application of Riiden- and beam intensity one can achieve biologi- berg and the publication of a paper by Knoll cal pictures without destroying the material, and Ruska; however, the subject goes back This was possible because of the introduction at least three or four years earlier to discus- of the internal photography described above, sions m physics colloquia in the Berlin area During 1934, Ruska had also demon- as seen from a publication by Gabor. Gabor's strated that the electron microscope had a name should be mentioned here in another resolution somewhat greater than that of the respect too. The studies of Knoll and Ruska light microscope. These observations, com- on magnetic electron lenses were greatly bined with the theoretical prediction that the facilitated by the fact that an ironclad mag- electron microscope is able to reach a re- netic lens had been developed in 1926-27 by solving power exceeding that of the light Gabor. During 1931-32, the first instru- microscope by a factor of several hundred, ments which merit the name of electron led to increasing efforts toward an improve- microscopes had been built. The fii'st of these ment of the resolving powder of the electron probably is one built by Knoll and Ruska for microscope. the observation of emitting objects, which From the beginning, it was understood utilized magnetic lenses. A second was the that a good part of the mechanism of the instrument built by Briiche and Johannsen, image formation is due to scattering of the with electrostatic lenses, to study emission electrons within the transmission-type ob- phenomena. Before the end of 1932, Mar- jects. Both Ruska and Marton referred to a ton had a simple instrument using a magnetic scattering mechanism in their 1934 papers, lens to investigate emission phenomena, and In 1936, Marton made the first quantitative later transmission phenomena. During 1933, attempt to explain the image formation on two instruments were built. One was Knoll the assumption of multiple scattering by the and Ruska's revised and improved transmis- object. The same approach was used a couple sion instrument, using two magnetic stages of years later by von Ardenne. Very soon, and having a limiting magnification of however, it became obvious that the average 12,000. IVIarton's second instrument also specimen of electron microscopy is much too used magnetic lenses for transmission ob- thin for multiple scattering. Consequently, servation, but it was simpler and its limiting ]\Iarton and Schiff in 1941 studied image for- magnification was of the order of 2,000. The mation assuming single scattering. In the first biological observations were carried out postwar years, single scattering calculations with Marton's two-stage instrument in 1934. were further improved by von Borries, Lenz Marton built in 1934 an improved two-stage and others, instrument, incorporating a specimen cham- In the meantime, efforts were under way 1.56 HISTORY OF ELECTRON OPTICS to improve the performance of the electron two-stage magnification. Around 1941, Hil- microscopes. These were twofold — ^one was lier attempted to use a compound projection building new microscopes of which a few lens. Three-stage magnification was intro- merit brief mention: The- instrument by duced in 1942-43 by Marton. This idea was Scott and iVIcMillen (1936) for the emission- taken up, probably independently, late in type observation of biological subjects. This 1944 by Ruska and von Borries in a patent was the first compound emission microscope application, and in 1944-46 by Le Poole. Le in America. Second, the instrument of Mar- Poole also introduced selected area diffrac- tin, Whelpton, and Barnum in England tion as an adjunct to electron microscopy, (1937) which sparked the British work in this using a three-stage arrangement, field; and the third, the new instrument of The postwar years have seen the rapid Ruska and von Borries (1938). This last in- development of commercially available strument paved the way to the fii'st commer- models in different countries. An important cial model, completed in 1939. There were contribution to the present-day development also advances in applications. In 1937, Mar- was recognition of the role of lens astigma- ton had taken the first bacteriological pic- tism by Hillier and Ramberg in 1946. Recog- tures. These were followed in 1938-39 by nition of this defect led, in the ensuing years, improved bacteriological pictures by Ruska to the development of different types of and collaborators. The first virus pictures electrostatic or magnetic stigmators and were obtained by Helmut Ruska, brother of produced the present high resolving power Ernst Ruska, in 1940. of modern instruments. The existence of the In 1938-39, some papers also appeared high resolution transmission type micro- by von Ardenne analyzing some of the de- scope stimulated the development of several fects of the microscope. The most important related instruments. First of these was the so- single item probably was an analysis of the called scanning microscope proposed by von alignment defects on the basis of which von Ardenne in 1938, followed by further im- Ardenne built a very complicated and highly provements by Zworykin, Hillier, and Sny- successful instrument establishing a tem- der in 1942. In 1939, the electron shadow porary world record of 30 A for resolving microscope was proposed by Boersch. In the power. Coincident with that development same year, 1939, von Ardenne proposed the was development also of the first electro- X-ray shadow microscope. The practical de- static microscope by Mahl and Boersch, as velopment of this instrument had to wait well as the development of the fu'st Canadian until the 1950's when Cosslett and Nixon microscope by Burton, Prebus and Hillier. built the first working instrument. Another This was followed by the development of the related instrument is the microprobe ana- first American commercial model — -the RCA lyzer of Castaing first described in 1951. Ion Type A was completed by Marton in 1940, microscopes are another group of derived followed by RCA Type B in 1941 by Zwory- instruments. The first attempt was made kin, Hillier, and Vance. These two last- with lithium ions in 1947 by Boersch, In named instruments were the first ones to use 1949, Magnan and Chanson published the highly regulated electronic power supplies, first description of a proton microscope. The German work relied on the stability of The 1932 attempts of Knoll and Ruska, as storage batteries to achieve the required well as Briiche and Johannsen, in emission constancy of the magnetic lenses. microscopy has been mentioned earlier. The Japanese work on electron microscopy resolving power of these emission micro- started in 1939 partly by Higashi and Tani, scopes was quite poor, and no progress was also by Tadano. made until Recknagel in 1941 developed a All the above-described instruments used theory of immersion objectives. In 1942, 157 ELECTKO.N MlCKOSCOl'Y Mecklenburg and Malil produced much bet- ter emission micrographs with thermionic electrons. Secondary electrons for producing emis- sion microscopy had been hrst used in the middle thirties. A crude image had been shown in 1933 by Zworykin, and a year or two later Knoll had shown better results. The best results to date are those of Mollcn- stedtin 1953, who replaced the primary elec- trons with ions for the excitation of second- ary electrons. Photo emission as a source of emission microscopy had been first used by Briiche and by Pohl in 1933 and 1934. The method has been further developed in the hands of Grivet and Septier in 1956. The most spectacular emission microscope is not a true microscope in the optical sense. Field emission microscopy started with a two-dimensional model of Johnson and Shockley in 1930. Very soon thereafter, in 1935-37, Mueller developed the point emis- sion microscope. This was followed in 1951 In' the invention of a field ion microscope with a best achieved resolving power of about 2.7 A. A method derived from dark field micros- copy is the so-called schlieren observation of electrostatic and magnetic fields and other perturbations of the optical medium. One of the first observations of this kind was pub- lished by Boersch m 1937, when he demon- strated the dark field image of a vapor stream. In 1943-44, von Ardenne deliber- atel\^ produced schlieren conditions. Due to the war, this paper was not known in the United States, and it was independently dis- covered by Marton in 1946-47 who with Simpson and Lachenbruch applied this method for quantitative evaluation of mag- netic fields. As this review is hmited to the electron microscopical aspects of electron optics, a very brief enumeration of other de^•elop- ments of electron optics may be sufficient. The oscilloscope tube development was started in 1894 by A. Hess and in 1897 by F. Braun. These early tubes used gaseous discharge with cold cathodes. The hot cathode in the cathode ray tube was introduced by Wehnelt in 1903. The first proposals for the use of cathode ray tubes as image transmitting ele- ments were made in 1900-07 by Dieckmann and Glage, as well as by Rosing. Storage of the image now used in television pickup tubes was first announced in 1908 by Camp- bell-Swinton. Modern development is linked to the names of Zworykin between 1925-33, Round in 1926, Farnsworth in 1927, and Henroteau in 1929. Mass spectrographs are another interest- ing chapter of electron optics. They were first conceived by J. J. Thomson in 1897; 180° magnetic deflection was fu'st used by Classen in 1908. Ten years later, Dempster used magnetic focusing; and in 1919, Aston invented ^-elocity focusing. Modern mass spectroscopy (developed in 1932-34) is linked to the names of Herzog and Mattauch w^ho developed the electron optics of mass spec- troscopy devices, as well as Bainbridge, Bar- ber, and Stephens. The development of electron optical theory is linked to three names essentially: Stoermer we mentioned earlier, but the for- mal theory of axially symmetrical devices was not developed until the years 1933-36. Foremost are the names of Glaser and Scher- zer: the first used an essentially Fermat- Hamiltonian approach, whereas the latter preferred the trajectory method. The first linking of geometrical electron optical theory to wave mechanics was done by Glaser in 1943. We now present a short description of the development of physical electron optics. Electron difTraction instrumentation for the observation of crystallographic structures has been greatly improved by the addition of a magnetic lens to the diffraction instru- mentation by Lebedeff, in 1931. Modern diffractographs took advantage of the high resolving power of combined electron optical 158 IIVIAGE FORMATION MECHANISM and diffraction elements as manifested by the instrument developed by Cowley and Rees in 1952. Boersch has contributed greatly to the understanding of the physical optics of elec- tron microscopical phenomena. First he had shown, in 1936, the correlation between dif- fraction diagram and the electron optical image. Recognition of this fact was essen- tially responsible for the development of all modern combined electron microscopical and diffraction instrumentation. In 1941, he showed the correlation between image prop- erties of crystalline objects and Bragg reflec- tion. Then, in 1943, he discovered the exist- ence of Fresnel diffraction in electron microscopical images. This discovery led in 1946 to the observation of phase contrasts by Hillier and Ramberg and to the method called "microscopy by reconstructed wave- fronts" invented by Gabor in 1948. Electron interferometry started with a proposal by Marton in 1952. Following this proposal, Marton, Simpson, and Suddeth built a three-crystal, wide beam interferome- ter in 1953. About the same time, accidental interferences had been observed in electron microscope images by Rang. The use of Fres- nel biprisms for electron interferometry was introduced in 1954 by Mollenstedt and Diiker. L. Maeton IMAGE FORMATION MECHANISM Two criteria are commonl}- used for judg- ing an image formed by an optical system: one is resolution, and the second is contrast. Some of the factors governing both are com- mon, and for this reason the ensuing discus- sion will consider both. The essential mecha- nism, responsible for both qualities, is scattering of electrons in the broadest phys- ical sense. It is customary, however, to desig- nate the coherent effects of scattering from an aggregate of atoms by the term "diffraction," reserving thus the expression "scattering" to the manifestations of localized electron- specimen interaction at the atomic or quasi- atomic level. These two effects constitute the major contributions to resolution and con- trast in the image. Two minor contributions are called "refraction" and "absorption." Refraction is that process in which the major change brought about by the scattering is limited to a change in the phase of the wave function, while in absorption there is a large change in the amplitude of the wave func- tion due to the scattering. In this discussion of image formation, we will consider only those effects which arise from the interaction of the electrons with the object and those modifications of these effects which are produced by the proper- ties and defects of the optical system. How- ever, the general effect of the optical proper- ties of the system on image formation will not be treated. Diffraction is due to the wave nature of the electron which produces deviations from the simple geometrical model of energy propagation. These deviations are manifested by the appearance of dark and bright bands, the Fresnel diffraction fringes. The least re- solved distance of an optical instrument (most commonly called resolving power) can be taken to be the distance at which the first diffraction maximum from a point-like object coincides with the zeroth diffraction maximum of the next object. Using this so- called Ra3^1eigh criterion in the Abbe rela- tion, we obtain 5 = 0.61X n sin a (1) where 8 is the least resolved distance, X is the wave length of the electron (X = h/p, where h is the Planck constant and p is the momentum of the electron), and a is the aperture angle of the optical system, n is the refractive index of the electron optical me- dium wliich for many applications can be as- sumed to be unity. The numerical constant 0.61 changes if another criterion is substi- 159 ELECTRON MICROSCOPY tutod for the Rayloish criterion in the defini- makes an angle (satisfied by equation (S)) tion of the least resolved distance. Equation with a lattice plane of the crystal, essentially (1) can be derived either from the theory of specular reflection may occur from this lat- Fraunhofer diffraction or from the quantum tice plane. If the optical system of the elec- mechanical uncertainty relation. Both rela- tron microscope has a wide enough aperture tions give qualitatively the same results to collect both the primary beam and the within a small numerical factor. reflected (diffracted) beam, and the objective Contrast is also governed to some extent is properly focused so that the two beams by the diffraction phenomena at the edge of are brought to the same focus, the intensity an object. Using the theory of diffraction, distribution in the image of the crystal will the fringe distribution at the edges of an appear to be w^hat may be called "normal." opaque or semi-opaque object can be calcu- If the objective aperture, however, is small lated ; and such calculations have been found with respect to the Bragg angle, such that to be in reasonably close agreement with the reflected beam will be intercepted (most experiment, provided that in the case of of the intensity being in the reflected beam), semi-opaque objects a refractive index of the the directly transmitted part of the image material, corresponding to an internal po- will appear unusually dark. A further mani- tential of 10-15 ev, is assumed. The experi- festation for this phenomenon is when a wide ments show that fringes appear in out-of- aperture system is slightly defocused and focus images and that the fringe spacing both the dark and bright images appear side depends upon the degree of defocusing of the by side. This variation in intensity, due to image. The apparent contrast at the edge of Bragg reflection, contributes also to appear- an image is optimum in the slightly defocused ances of the so-called extinction contours, image. For perfect focusing, the contrast is These extinction contours appear in slightly found to be lower than for the out-of-focus bent crystals where, accidentally, part of the condition. crystal happens to be at the proper angle to The intensity distribution in Fresnel dif- show Bragg reflection. Under action of the fraction fringes can be interpreted as inter- electron beam, thin crystals may warp, show- ference phenomena. Accidentally occurring ing a displacement of the extinction contours interference fringes can be observed also in while under observation, selected specimen areas illustrating again the Related also to Bragg reflection is the ob- role played by interference phenomena in the servation of crystal defects, such as disloca- contrast in the image plane. tions, stacking faults, etc. Due to the exist- In the case of crystalline specimens, dif- ence of these crystal defects, certain lattice fraction from the lattice planes can produce planes show different orientations from the an important modification of the intensity surrounding lattice planes and produce in distribution in the image plane. Electrons this manner a marked contrast in the final which are scattered from a crystal lattice image. will show diffraction maxima according to While the above considerations are re- Bragg's equation stricted to Fresnel diffraction and to crystal- „ , . line diffraction, Gabor has demonstrated that dmraction phenomena may be impor- where d is the distance between the lattice tant in noncrystalline specimens too. He pro- planes, is the angle between the direction posed the use of a coherent electron beam for of the incident beam and the diffracting forming a diffraction image, called hologram, planes, and n is an integer. of the specimen. In principle, such a diffrac- If an electron beam is directed so that it tion image should contain all information 160 IMAGE FORMATION MECHANISM about the specimen except the phase. Through the use of proper optical equipment, the image can be reconstructed from such a hologram in a process called "microscopy by- reconstructed wave fronts." Although prac- tical difficulties until now made it impossible to reach high resolutions by this method, its existence amply shows the importance of dif- fraction in the image forming process. At extremely high resolution, crystal lat- tices and other periodic structures can be ob- served provided that the aperture of the sys- tem allows the zero order spectrum to pass together with at least one of the first-order diffraction spectra. This is in agreement with the Abbe theory of image formation in light optics. Moire patterns are due to the com- bination of a doubly diffracted beam, orig- inating on overlapping crystals, with the in- cident beam. The intensity at any point of the image is governed by the number of electrons which have been scattered within the solid angle formed by the aperture of the objective lens. This number can be WTitten as for the elastically scattered electrons is given by where '^'" - ^ (z - fy- (5) '^j t^^^c ^'l^f *^ >'»^^ L^--" ,d^ ^mI *-T-^iii«* Fig. 3. Cross section of a mouse kidney proximal convolution (top) and a distal convolution (bot- tom) with surrounding capillaries (C). In both tubules the nuclei (N) and mitochondria (M) are evi- dent. The lumen is usually closed in the proximal convolution but open (Lu) in the distal. Numerous brush border extensions (B) line the cells in the proximal convolution but only few microvilli (Vi) in the distal convolution. Vacuoles (V) containing autofluorescent material are prominent. Occasionally, a cilium (Ci) is encountered in the distal convolution. Magnification 2,900X. cupying the space that normally would be mal convolution and the abundance of brush called the lumen of the tubule. Evidence is at border extensions seems to facilitate this ac- hand showing that about 85 per cent of the tivity of the cells by largely increasing their urine reabsorption takes place in the proxi- absorptive surfaces. Therefore, the w'ay the 168 KIDNEY ULTRASTRUCTURE urine seems to be passed along the proximal convolution is probably more like the slow filtration through a finely porous sieve than like the rapid drainage of a sink with a wide open plumbing system. The "pores" of this particular sieve are then represented by the narrow slits between the millions of brush border extensions. It can be i^ecorded in high resolution micrographs that the plasma membrane which covers the extensions has a thickness of about 50A. However closely packed, the extensions do not get in closer contact than lOOA apart. The intervening space is occupied by a substance with less density than the plasma membrane. It has been suggested that this substance represents an additional layer of possibly lipid mole- cules with a depth of 50 A, identical with the intervening layers between adjacent cell bor- ders. When the urine passes along the proxi- mal convolution, it expands slightly the two lipid layers of adjacent brush border exten- sions, thus creating the narrow slits men- tioned above. By this mechanism is estab- lished a much closer contact between the urine and the surface membrane of the brush border extensions than would be the case with a wide open lumen. Some substances, as for instance proteins, which cannot be absorbed through the sur- face membrane, are taken up by the tubular invaginations. These structm'es are located between the bases of the brush border ex- tensions and represent minute tubules which are in open connection with the tubular lumen (Fig. 8). The tubular invaginations are ex- panded to vacuolar profiles when fluid and substances are taken in. A certain condensa- tion of the vacuolar content is noted con- comitantly with a breaking-off of the con- nection between the expanded tubular invagination and the lumen of the nephron. The condensed vacuole can then be found anywhere in the cell and as the condensation of its contents proceeds, the vacuole is trans- formed into what has been called a large granule in electron microscopy. This whole mechanism of taking in fluid-dissolved sub- stances has been called micropinocytosis or membrane flow. The mechanism by which substances other than protein are taken up by the cells of the proximal convolution is structurally not clear. It involves substances and ions such as glucose, sodium, potassium, phosphate, sulfate, amino acids, urea, and creatinine to mention some of them. The reabsorption of water is a passive process, but most of the substances listed involve an active process which requires energy. The enzymes recjuired for these active processes are supplied by the multitude of mitochondria present in the cells of the proximal convolution (Fig. 4). The ultrastructure of the mitochondria has been thoroughly investigated (cross reference: cell ultrastructure) and it is believed that the efficiency of the mitochondrial work is in- creased by the number of mitochondria and the presence of structurally intact internal membranes. Organelles have been recorded in these cells which presumably constitute mitochondrial precursors. They have been called microbodies, and represent small spher- ical bodies without internal membranes and with a single membranous capsule as com- pared with the double-contoured mitochon- drial outer membrane. Structures have been found in the basal portion of the cells of the proximal convolu- tion which probably facilitate the flow of re- absorbed fluid and substances through the cell body. They represent infolding s of the basal plasma membrane which extend to a varying degree into the cell (Fig. 4). The mitochondria have a close relationship with these infoldings and it is tempting to as- sume that this facilitates a certain interac- tion between the enzymes carried by the mitochondria and the infolded plasma mem- brane (Fig. 8). Once the reabsorbed fluid and substances have obtained contact with the basal plasma membrane, they may penetrate it by means of an enzymatic activation (Fig. 9). And when the extracelhilar space is 169 Fig. 4. Basal part of a proximal convoluted tubule cell of the mouse kidnej'. The plasma membrane which faces the basement membrane (BM) is highlj^ infolded, but at the point where it turns (arrow) another portion of the cytoplasm is always interposed. The basal cell cytoplasm is thus divided into coarse lamellae which contain mitochondria (M) and abundant RXA particles (R). Magnification 32,000X. Fig. 5. Basal part of a collecting tubule cell of the mouse kidney. As in the proximal convolution, the basal plasma membrane is infolded, but as a rule, the turning point (arrow) can be seen without anj' interposed cytoplasmic lamella. The cytoplasmic lamellae are here thinner than in the proximal con- volution and not broad enough to house mitochondria (M). Some RNA-granules (R) are located in the lamellae. The basement membrane (BM) is quite thin. Magnification X46,000. 170 BM Fig. 6. Two cells of the thick, a.- ivstiiig on the basement membrane (BM) and joined by a terminal bar (TB). The cell boundary is indicated by the dotted line. The nucleus (N) is located in the apical part of the cell with the Golgi zone (Go) close to it. Extending into the tubular lumen (Lu) are numerous microvilli (Vi). The cytoplasm is finely granulated because of the presence of abundant RNA particles. The autofluorescent vacuoles (V) are less numerous than in the proximal convolution. The mitochondria (Ml) are long and densely packed. Sometimes, their shape is more like an ice cream bar than that of a sausage, which can be seen when they are sectioned at an angle (M2). The basal cell membrane is deeply infolded and the turning points clearly seen at the arrows show that cytoplasmic lamellae, also containing mitochondria, are always interposed in an interdigitat- ing fashion. Magnification 11,000X. 171 Fig. 7. Survey of the papilla of the rat kidney showing mostly cross sectioned collecting tubules (D), thin segments of Henle's loop (T) and capillaries (C). The cells of the collecting tubule are cuboidal with few mitochondria, easilj^ recognized cell boundaries, and tinj- microvilli on the surface. The cells of the thin segment are fairly squamous in sections showing scalloped appearance due to the frequently inter- digitating, narrow cell extensions. The endothelium of the capillaries is extremely thin. In the electron micrograph, there seems to be no chance of mistaking a thin segment for a capillary because of the dif- ference in thickness of the lining cells, but also because of the obvious staining of the blood plasma in the capillaries. The interstitial cell (X) is unidentified. Magnification 2,700X. 172 KIDNEY ULTRASTRUCTURE reached, the access to the surrouncUng capil- laries is easily achieved. The Golgi zone of these cells is quite re- stricted and it is believed that it is mainly- involved in secretory processes, particularly in the synthesis of protein used either by the cell itself or for extracellular purposes. This hypothesis is supported by the fact that rough-surfaced endoplasmic reticulum is ab- sent in these cells, but the cytoplasm is filled by numerous RNA-particles which presuma- bly are the structural evidence of cytoplasmic proteins. The Loop of Henle. When the nephron leaves its convoluted portion in the cortex of the kidney, it assumes a straight course down into the medulla and papilla. The first part of this portion is called the straight de- scending loop of Henle. Once down in the papilla, it bends back again and the turning part is called the thin segment of HenWs loop. PROXIMAL CONVOLUTED TUBULE Fig. 8. Schematic representation of the basic structures of the proximal tubule cells of the mouse kidney: nucleus (N), mitochondria (M), microbodies (m), Golgi apparatus (Go), auto- fluorescent vacuoles (V), large dense granules (D), ribonucleoprotein particles (R), brush border extensions (B), tubular invaginations (Ti), plasma membrane (PM) with infoldings, terminal bars (TB), and basement membrane (BM). (After Rhodin, 1954). cell borders bcsame^f rnembrane secfioma " cytoplasmic laimllat cytoplasmic lamtliaz frcm eel Is pulled auaq terminal bars rnifochondria nucleus cell borderi ■ THIN SEGMENT OF HENLE'S LOOP Fig. 9. Three-dimensional reconstruction of cells of the proximal convolution and thin seg- ment of the mouse kidney. The reconstruction is not based on serial sections. (After Rhodin, 1958). From there on, it approaches again the cortex and the neighborhood of the glomerulus. This portion is called the ascending (thick) limh of Henle's loop. It should be stressed that in most nephrons, the thin segment comprises a considerable part of the descending loop, the hairpin turn itself, and for some distance, the ascending loop. Straight Descending Loop. The cells of the straight descending loop of Henle have a light cytoplasm due to a certain scarcity in mitochondria and RNA-particles. The sur- face of the cells shows scattered brush border extensions, but these are shorter and coarser than those found in the proximal convoluted portion of the nephron (Fig. 10). The lumen of the tubule is frequently open, presumably because of the relatively few brush border extensions. Tubular invaginations of the surface plasma membrane are few and the basal infoldings of the plasma membrane, so numerous in the proximal convoluted part, are shallow and few, often completely absent. 173 ELECTRON :\lICKOSCOPY ^^^M^^ 6 Fig. 10. Diagram showing the essential fea- tures of the colls lining different parts of a tj'^pical cortical nephron in mammalian kidney as seen with the electron microscope. 1) Collecting tubule (arched or proximal part) : dark intercalated cell; 2) Collecting tubule (arched or proximal part): light cell; 3) Proximal convo- luted tubule: proximal part; 4) Distal convoluted tubule: intercalated part (Schaltstiiek); 5) Proxi- mal convoluted tubule: terminal portion; 6) Distal convoluted tubule: thick (ascending) limb; 7) Thin segment of Henle's loop. (After Rhodin, 1958) These phenomena seem to indicate that less resorptive activity is performed by the straight descending loop as compared with the convoluted proximal part of the nephron when judging from a structural point of view. Thin Segment. The epithelium of the thin segment of Henle's loop is of a squamous type with extremely flattened cytoplasm (Fig. 7). The cell surface shows only scattered short microvilli and tubular invaginations are not recorded. The mitochondria are scarce and exceedingly small. The cytoplasm is light due to a small number of RNA-par- ticles. Basal infoldings are present only be- neath the nucleus where the cytoplasm is of greater thickness than elsewhere. The at- tenuated part of the cell shows some quite characteristic features of this portion of the nephron. It displays a large number of cyto- plasmic extensions which rest on the base- ment membrane. These extensions resume the shape of the arms of a starfish and they interdigitate frequently with similar exten- sions of neighboring cells (Fig. 9). This pat- tern is reminiscent of the way the epithelial cells of the glomerular capillaries are ar- ranged. However, in the case of the thin seg- ment, the interdigitated cell processes al- ways show a terminal bar close to the surface which presumably secures the firm attach- ment of individual cell processes to one an- other. It can, therefore, be assumed that there do not exist any "slit-pores" between the cells of the thin segment as was indi- cated to be the circumstances regarding the glomerular capillary epithelial cells. In the papilla and deeper portions of the medulla, the descending and ascending parts of the thin segment are closely opposed to each other as well as to the capillaries (vasa recta) and the collecting ducts. It has been suggested that the close juxtaposition would facilitate the mutual exchange of fluid, sub- stances and energy because of the close re- semblance of this system to the basic princi- ple of a counter current system with streams moving in opposite directions. The ultra- structure of the thin segment seems to sup- port this hypothesis. The thin walls of its loop are evidently ideal to serve this ex- change of fluid and substances. Ascending (Thick) Limh. The ascending limb of Henle's loop is slightly wider than the other portions of the nephron and has, therefore, been called the thick limb. The cells have a cuboidal shape and a free open lumen is always to be found. The extreme multitude of mitochondria makes this sec- tion of the nephron stand out more clearly than the others, both in light and electron microscopy. Not only are the mitochondria numerous, but they also have a coarse and elongated shape as compared to the spherical form of the other parts of the nephron (Fig. 6). The mitochondria are densely packed and arranged with their long axes perpendicular 174 KIDNEY ULTR VSTRUCTURE to the basement membrane of the tubule, convokition is less coiled than the proximal They extend from the basement membrane one. The cells become small and a wide lumen to a level of about one micron from the cell opens up (Fig. 3). The surface of the cells surface. The mitochondrial fine structure is usually has longer microvilli than is the case of the same appearance as elsewhere in the in the thick limb, but the length varies from nephron with the exception that the inner cell to cell. The mitochondria decrease no- membranes (or plates) are more frequently ticeably in number and size, as does the seen arranged parallel with the long axis of number of RXA-particles, features which all the mitochondrion. contribute to a light appearance of the cell The free surface of the cells displays a cytoplasm. A few shallow and narrow in- number of small and scattered microvilli. The foldings of the basal plasma membrane can luminal portion of the cells, more or less de- be recorded, but they are too narrow to ad- void of mitochondria, is pervaded by abun- mit any mitochondria to be enclosed in the dant microvesicles bounded by a smooth small cytoplasmic strands they give rise to single membrane. Some of the vesicles are (Fig. 5). A fair number of small vesicles is connected with the surface plasma mem- still to be found in the luminal part of most brane. The Golgi apparatus of the cells oc- of the cells, but cells can be recorded which cupies a restricted area around the upper part are completely devoid of these structures, of the nucleus. Its fine structure is identical When the distal convolution approaches with the one recorded in the proximal con- the neighborhood of the vascular pole of the volution. The basal plasma membrane is glomerulus, it becomes attached to the angle deeply infolded between the elongated mito- between the afferent and efferent arterioles, chondria leaving but a narrow strip of cyto- This part of the distal convolution has been plasm in between. The RNA-particles are called the macula densa because it can be numerous and scattered among the small seen that the cells of the distal convolution vesicles in the apical cytoplasm. here assume a more columnar shape than It is obvious that the work performed by elsewhere, thus causing the nuclei to stand the thick limb requires a big supply of en- out very clearly in close juxtaposition (Fig. zymes judging from the great number of 1). Apart from their extreme columnar shape, mitochondria present in the cells. Sodium is the cells of the macula densa do not show any taken up by these cells by an active process peculiarities as far as their fine structure is which also leads to a simultaneous retention concerned which would discriminate them of water. Furthermore, formation of ammonia from other cells of the distal convolution, and the acidification of the urine seem to take Cortical Collecting Duct. This part of place in this portion of the nephron. It may the nephron immediately follows the distal be reasonable to assume that the numerous convolution and serves only as a short con- microvesicles are structural evidence for one nection between this part and the collecting or both of these processes. Similar micro- tubule. It is characterized by two cell types, vesicles are a prominent feature of the pari- the first' identical with the one occurring in etal cells of the gastric mucosa. These cells the distal convolution, the second, with the supposedly are responsible for the production appearance of the cells found in the collect- of the hydrochloric acid of the stomach. ing tubule. As the cortical collecting duct is Distal Convolution. The thick ascend- gradually transformed into the collecting ing limb of Henle's loop is gradually trans- tubule, the cell type reminiscent of that of formed into a convoluted position, the dis- the distal tubule disappears. Because these tal convolution, when the nephron again cells seem to be interposed between the main reaches the cortex of the kidney. The distal cell type, they have been called intercalated 175 ELECTRON ^IKKOSCOPY cells. Tlie intor('alat(Hl coll has a large luunbcr of sphorical initDchoiulria wliich arc clustered mostly above the nucleus, thus giving the cell a dark appearance which has lead some investigators to call it a dark cell (Fig. 10). This seems justifiable when considering that the main cell type has few mitochondria and a light cyt(»i)lasni. This cell is therefore called a light cell. There are other structurally important differences between the two cell types. The luminal surface of the dark cell usuall}^ has quite abundant microvilli, whereas the light cell has few or none. The vesicles of the luminal part of the dark cell cytoplasm and the RNA-granules are numer- ous as against a scarcity of these structures in the light cell. The similarity between the dark intercalated cells of the cortical collect- ing duct and the cells of the distal convolu- tion strongly suggests that the former ac- tually represents a certain variety of cells of the distal convolution which are distributed along the cortical collecting duct. Collecting Tubule. The collecting tubule begins in the outer cortex and runs in a straight course toward the medulla, each nephron being connected to a collecting tu- bule by the cortical collecting duct. The con- vergence of successive orders of collecting tubules in progressively deeper layers gives rise to vessels of mcreasing caliber until fi- nally the papillary ducts are reached. These ducts discharge their contents into the renal pelvis. It is believed that very few processes re- lated to the composition of the urine occur in the collecting tubule. The final concentra- tion of the urine seems, however, to be estab- lished here, possibly mediated by a certain reabsorption of sodium chloride. It has also been suggested that the permeability of the cells of the collecting tubules may be in- fluenced by the action of the antidiuretic hormone (ADH) of the pituitary. Structurally, the cells of the collecting tubule are rather poor (Fig. 7). The cells are of a cuboidal type with a very light cyto- plasm containing a small amount of IINA- particles and a few small and widely scat- tered spherical mitochondria. Large granules of the lipid type occur frequently, together with a small Golgi complex. The luminal sur- face displays a varying number of very short microvilli and the basal plasma membrane shows remnants of extremely shallow infold- ings. The cell borders are straight and the cells are held together by distinct terminal bars close to the surface. The nuclei are fairly large, occupying a good portion of the cell body. REFERENCES Sjostrand, F. S. and Rhodin, J., "The ultra- structure of the proximal convoluted tubules of the mouse kidney as revealed by high reso- lution electron microscopy," Exper. Cell Re- search, 4, 426 (1953). Hall, B. V., "Studies of the normal glomerular structure by electron microscopy," Proc. A7inual Con}. Nephrotic Syndrome, 5th Conf., p. 1, 1953. Rhodin, J., "Correlation of ultrastructural or- ganization and function in normal and experi- mentally changed proximal convoluted tubule cells of the mouse kidney," Thesis, Stockholm 1954, Karolinska Institutet. Pease, D. C, "Fine structures of the kidney seen by electron microscopy," /. Histochem. Cyto- chem., 3, 295 (1955). HalL; B. v., "Further studies of the normal struc- ture of the renal glomerulus," Proc. Annual Conf. Nephrotic Syndrome, 6th Conf., 1955. Rhodin, J., "Electron microscopy of the glomeru- lar capillary wall," Exper. Cell Research, 8, 572 (1955). Pease ; D. C, "Electron microscopy of the vascu- lar bed of the kidney cortex," Anat. Rec, 121, 701 (1955). Pease, D. C, "Electron microscopy of the tubular cells of the kidney cortex," Anat. Rec, 121, 723 (1955). RusKA, H., Moore, D. H., and Weinstock, J., "The base of the proximal convoluted tubule cells of rat kidney," /. Biophys. Biochem. Cytol., 3, 249 (1957). Rhodin, J., "Anatomy of kidney tubules," Int. Rev. Cytol., 7, i85 (1958). Rhodin, J., "Ergebnisse der elektronenmikros- kopische Erforschung von Struktur und Funktion der Zelle," Verhandl. deutsch., Gesellsch. Path. 41st. Tagung, 18, 274 (1958). 176 LEAF SURFACES Rhodix, J., "Electron microscop}' of the kidney," Am. J. Med., 24, 661 (1958). Johannes A. G. Rhodin LEAF SURFACES The electron microscope and its associated techniques have now reached a stage in de- velopment where they can be used by biolo- gists as tools of research, not simply as instruments for making interesting new mor- phological discoveries or for confirming in- formation gained from other sources. Our investigations, using the new techniques, have demonstrated the existence of a fine structure on the surfaces of many plants. Al- though some such structure beyond the reso- lution of the light microscope had been in- ferred, because plant surfaces vary widely in their ability to repel water droplets, its morphology, diversity, and the problems of its development had never been suspected. Prior to this work several attempts had been made to examine the surfaces of animal and plant cuticle with the electron micro- scope. Holdgate, Menter and Seal (1954) used reflection electron microscopy to study insect cuticle. This technique, although suc- cessful in demonstrating changes in the insect cuticle, suffered from the disadvantages as- sociated with reflection electron microscopy; the difficulty of interpretation is due to dis- tortion and a restricted magnification only a little above that of the light microscope. Most of the work on plant cuticle has used some form of replica technique. The pioneer work by Mueller, Carr and Loomis (1954) and Schieferstein and Loomis (1956) used liquid polyvinyl alcohol as the f