THE PARTICLE SIZE ANALYSIS OF PIGMENTS X¾ITH THE Q UANTIJ•IET 689 REFERENCES (1) Analyst 80 156 (1963). (2) Allen, T., Particle Size 2kleasurement (1968) (Chapman & Hall). (3) Irani, R. R. and Callis, C. F. Particle Size: Measurement Interpretation and Application (1963) (Wiley, New York). (4) Brit. J. Appl. Phys. õ (Suppl. no. 3) (1954). (5) British Standard 3406: Part 4 (1963). (6) A•talysis by microscopical methods for particle size distribution ASTM: E20-51T (1951). (7) Metals Research Ltd., I•Ielbourn, Roysion, Hefts. (8) Quantitative Television Microscopy, Micro-67, Microscope lO (April 1968). (9) Cole, M. Microscope and Crystal Front, •õ 148 (1966). (10) Fisher, C. Particle Size Analysis 77 (1967). (11) Kaye, B. H. Paint Oil Colour J. 1õ4 372 (1968). (12) Collins, G. F. Particle Size Analysis 65 (1967). (13) Smith, M. J. Microscope 10 123 (1968). (14) Herden, G. Small Particle Statistics (1960) (Butterworths, London). (15) Eckot•, R. K. The Microscope and Crystal Front 14 490 (1965). (16) Harwood, 5I. G. Brit. J. Appl. Phys. suppl. $, • S193 (1954). (17) Strang, A. Microscope 10 181 (1968). (18) Mendelsohn, M. L. et al. Advances in Optical and Electron 3Iicroscopy 9, (1968). (Academic Press, New York). (19) Franklin, A. G., Iron Steel Inst. 9,0 181 (1959). Introduction by the lecturer As one might expect, there are both advantages and disadvantages to be en- countered when using the Quantimet system. The former include rapid and relatively painless sample evaluation, while the principal disadvantage is usually presented by the nature of the specimen. Not all types of sample are suitable for evaluation by means of this technique although, provided a suitable optical image can be produced, useful results can be obtained. This is shown by the wide range of materials which are currently being examined with the instrument. In terms of the image presented to the instrument, the count is effectively absolute, i.e. if there is a particle seen on the screen, then the instrument counts it. The limitations to this result from the size, and the contrast and focus of the features while they are being examined. If the particle is partly out of focus in the microscope system, then it may or may not be counted, depending on how dense it is. In prac- tical terms, however, the results must be regarded' as relative, due to the variations which arise during sampling. Thus, emphasis should be placed upon establishing reproducable differences between two specimens, rather than upon a constant value for any single specimen. With the assistance of the instrument, sampling as a problem becomes much more obvious and much more easily solved. On the one hand we have more time to examine variations arising from different sampling techniques (unclouded by observer errors), while, on the other hand, we are able to examine greater numbers of particles in order to reduce the level of the error. It is interesting to note at this stage that we have found increased sampling levels to be useful in two distinct ways. Not only does it reduce the uncertainty of an evaluation (and hence the difference between two

690 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS consecutive evaluations of the same material), but it increases the observed level of difference between different samples. If increased counting loads are expected above those normally handled by the system, they may be accommodated by means of extensions to it. These include a data multiplexer which may be used in combination with a variety of display and output systems, such as typewriter output, punched tape, and accumulating digital displays. In combination with this there is an automatic stage device which moves the stage in a predetermined manner so as to present up to 500 fields of view for examination in 10 min. This means one immediately has a problem of handling data, rather than evaluating the specimen. Just as with the output, a variety of input systems may be used, enabling us to cover a wide range of sizes, from nanometres to centimetres. These include an electron microscope, visual microscopy (with incident or transmitted light, and dark field or phase contrast) and, at the macro end of the scale, there is an epidiascope attachment. This flexibility allows for a wide variety of applications and the Quantimet is beginning to enter a variety of fields, including experimental pathology and neurology bac~ teriology, for measuring the size of bacteria colonies and quantifying measurements taken from microbiological assaying haematology for counting cells odentology for measuring the size and number of cracks in the teeth in X-rays pharmacology, botany, and a very large field involving metallurgy, which is the original application of the instrument. It will size anything provided a suitable visual image is provided for it, including grain size of inclusions, and various phases in the metal. It is being used in solid state physics, mineralogy, nuclear physics, photography, radiobiology, air pollution, fibre technology, powder technology, pigment technology, and to study sieves, and aerosols. DISCUSSION MR. C. PUGH: How much does this rather versatile instrument cost? Ta•. L•.CTUR•R: The basic instrument as described above costs just under f5 000. It varies a little, depending on the optical system. This stage is really a research instrument. The time involved is largely taken up by winding from one field of view in the specimen to another, and recording the results manually. The next stage of the instrument, the Multiplexer Unit and a teletype output, add another •3 000 to the basic price. This considerably reduces the work load in- volved and releases a maj or part of the operators' time for other things. DR. R. PUGH: The Quaniim•t converts a volume into a projected area. Hence for solid suspensions, the volume to area conversion is only 100% true for perfect spheres. Do you have any data which will assess the amount of error incurred with irregular particles? T•I• L•.CTUR•R: Quite a lot of information has been published on this problem, which is a steriological one. The variations in particle shape can introduce many errors when the surfaces are randomly orientated. I myself am not particularly qualified to answer your question because I have no direct personal experience. The