Automation in Analytical Chemistry - Analytical Chemistry (ACS

Automation in Analytical Chemistry - Analytical Chemistry (ACS...

0 downloads 211 Views 7MB Size



Automation in Analytical Chemistry Automation is one of the most spectacular current achievements in the field of analytical chemistry. Whether it be related to selection, preparation, or measurement of a sample, or to separation and measurement of desired constituents, instruments are being devised to make the operation automatic. A recent symposium, held at the ACS meeting in San Francisco, illustrated the progress made in such areas as titrimetry, calorimetry, and spectrometry. The author, who was active in initiating this program, has prepared this summary of progress in which he notes that, spectacular as the advances have been, the d a y of the completely automatic control laboratory is not likely in the near future for many materials. OME 15 years ago an outstanding young organic chemist was telling graduate students t h a t little advance had been made in analytical chemistry in the preceding five decades. Consequently, he advised them not to major in an area which obviously offered so little opportunity. How justified was such an opinion? In reply, the writer suggests considering what lias been happening in one area of analytical chemistry—namely, aiitomation. One might, with much justification, select other new developments, such as reagents, materials, techniques, or procedures. To the writer, however, developments in automation are the most spectacular current achievements in the field. Under the title, "Analytical Automatons," he has presumed for a decade to speak on the nature and significance of this trend. Analysts have not agreed upon a concise definition for automation. The usage here is merely the writer's. For over two decades he has considered analytical methods in terms of the iinit operations involved. Thus, in quantitative work with a heterogeneous polycomponent sample there may be selection, preparation, measurement, and preliminary treatment (s). Then for the desired constituents in it there may be separation (s), and there are always measurements. Some methods, such as the densimetric determination of sulfuric acid in a lead storage battery, involve only the final operation of measurement. Others, such as the complete analysis of a granite rock for some 20 components, involve many operations. Considered from this viewpoint,


Figure I. Prototype printing analytical balance, consisting of a single-pan balance, an electronic translating unit, and an eight-place printing adding machine. {Courtesy of W m . Ainsworth and Sons)

automation is the demanning of one or more of these unit operations. If completely automatized, there is nothing manual to do but maintain the measuring device in satisfactory operating condition. If only partially automatized, one or more operations remain manual. The past half century has brought tremendous changes in both quali-

tative and quantitative analysis. Periodicals, patents, government bulletins, books (especially monographs) and manufacturers' technical publications reflect much of the change. Perhaps most convincing of the new day is a visit to a modern, analytical research laboratory of a diversified chemical organization,

M . G . M e l l o n , professor of analytical chemistry, Purdue University, was born at Conneaut Lake, Pa., in 1893. H e received his B.S. degree at A l l e g h e n y C o l l e g e (1915), M.S. (1917) and Ph.D. (1919) at O h i o State. H e has devoted his life to teaching, having been an assistant instructor at O h i o State ( 1 9 1 5 - 1 8 ) , and instuctor ( 1 9 1 8 19). H e became an assistant professor of quantitative analysis at Purdue in 1919, associate professor in 1925 and professor in 1931. He served on the Manhattan District Project in 1944. H e is a member of the American Chemical Society, American Association for the Advancement of Science, Coblentz Society, Society for A p p l i e d Spectroscopy, Society of Technical W r i t e r s and Editors, Association of A n a l y t i c a l Chemists, American Society f o r Testing Materials, O p t i c a l Society of America, Indiana A c a d e m y of Science, and the Indiana Chemical Society. H e received the Fisher A w a r d in A n a l y t i c a l Chemistry 1952, the Anachem A w a r d in 1953, and the Austin M . Patterson A w a r d in Chemical Documentation, 1957. His major areas of interest are colorimetry, absorption spectroscopy, and chemical literature. VOL. 3 0 , N O . 12, DECEMBER 1958






Figure 2. Automatic "Recordomatic" titrator. (Courtesy of Precision Sci­ entific Co.) and to a plant with automatic con­ trol of process streams. A glance backward aids in appre­ ciating more fully what has been accomplished in this direction and in recognizing unsolved problems of this kind. I t is 19 centuries since Pliny re­ ported t h e colorimetric detection, and presumably estimation, of iron in aqueous solution. B u t only within the year has a so-called com­ pletely automatic instrument—the Technicon Autoanalyser—been ad­ vertised for such a determination. Even so, it is not apparent how this device could be applied to many problems without several prior op­ erations, as in the determination of titanium in silicate rock. The origin of the principle of the equal-arm balance is uncertain, b u t such a device is shown in inscrip­ tions of ancient Egypt. For gravi­ metric analytical measurements it has been applied certainly since the advent of t h e quantitative period of chemistry (ca. 1775). Automa­ tion of balances is beginning (Fig­ ure 1) ; but, to the writer's knowl­ edge, the completely automatic bal­ ance is still a dream. The succeeding century saw the first flowering of analytical chem­ istry. Included were the develop­ ment of gravimetry, the establish­ ment of titrimetry, a n d t h e dis­ covery of other -imetries. Then followed a marked decline, com­ pared to other areas. Physical in­ terpreters arrived t o delve into the how and the why of things chemi­ cal. Hybrid biochemists emerged, and soon the further-crossed physi­

cal biochemists aspired to unravel the mechanism of life itself. M a n y others followed. Little wonder t h a t analysts—the weighers, the precipi­ tators, the titrators—were all b u t submerged by the successive waves of these new interests. Modern developments represent once more the realization of the indispensability of analytical chem­ istry. The old perennial is flower­ ing again. Certainly industrial chemists know t h a t we buy, make, use, sell, and investigate chemical systems largely on the basis of tests and analyses. T o m a k e an ever greater number of more kinds of determinations in less time and with fewer determinators is a daily prob­ lem in many laboratories. Auto­ mation of methods is p a r t of t h e effort to meet the demands. As yet no device has appeared for prepar­ ing and handling precipitates, b u t we do have a number of automatic titrators (Figure 2) for solutions ready for titration.


GUARANTEED* by your Distributor*


The recent symposium a t t h e American Chemical Society meet­ ing a t San Francisco on the litera­ ture of automation as applied t o analytical chemistry was the first such program of which the writer is aware. I t had t o be chiefly in­ troductory. Following a general survey, specific papers dealt briefly with titrimetric, calorimetric, a n d spectrometric methods. For several years many current developments have been summa­ rized in the reviews of G. D . P a t t e r ­




son, J r . , in ANALYTICAL C H E M I S T R Y .

Hundreds of references have been cited each time. Although much has happened and is happening, this subject is still largely neglected in textbooks on analytical chemistry. In considering briefly the nature of what has been done, and in sug­ gesting things undone, it is helpful to view analytical methods in terms of the unit operations already men­ tioned. T h e sequence inevitably begins with sampling and ends with measurement. W h a t intervenes de­ pends upon circumstances. A completely automated method includes all t h e operations which are necessary in a given case. T h e measuring instrument indicates a n d / o r records some value, such as density, absorbance, p H , or per­ centage composition. An excellent

* 75 Liters Per Minute * Vacuum to 0.1 Micron (Or Better) * Starts Quickly * Operates Quietly * Cleans Easily * Ideal for Diffusion Pump Operations

*See opposite page for list of P R E C I S I O N Distributors. Pump guaranteed to pro­ duce 0.1 micron vacuum on trapped M c L o u d guage. For complete information on the new Vacuum Pump request bulletin # 6 0 4 from your " P R E C I S I O N " Distributor.

Since 1920 The Finest i n Q u a l i t y Laboratory



: · ;

C O .

• · "

3725 WEST CORTLAND STREET CHICAGO 4 7 , ILLINOIS Circle No. 27 A on Readers' Service Card, page 97 λ VOL. 3 0 , N O . 12, DECEMBER 1958


2 7 A


Figure 3. Carbon dioxide recorder. (Courtesy of Cambridge Instrument Co.)

With a few simple movements, this analyst has made 10 determinations of molybdenum in steel. It took him less t h a n three minutes. This is why the Coleman Universal Spectrophotometer is called the

"Workhorse of the laboratory" The Universal performs precise routine analyses faster and easier t h a n any other spectrophotometer. Its use of a diffraction grating monochromator avoids electronic and mechanical adjustments, allows valid use of calibration curves; eliminates need for continual preparation and use of concurrent standards; insures reproducibility, simplifies and speeds analytical operations . . . Yet the Universal has the flexibility to meet the special problems t h a t arise in every laboratory and is quickly adapted for trace determinations, fluorimetry, nephelometry and other unusual analyses. Only Coleman offers this brilliant balance of speed, precision, reliability and versatility. Write

for complete

informationBulletin B-241.


Universal Spectrophotometer C O L E M A N I N S T R U M E N T S I N C . , DEPT. A , M A Y W O O D , I L L I N O I S For further information, circle number 28 A on Readers' Service Card, page 97 A 28 A



example is the recorder familiar in steam power plants for determining carbon dioxide in the flue gas. In one design (Figure 3) a pointer indicates continuously the percentage of carbon dioxide, and a pen records the values on strip chart paper. The only manual operation is changing the record paper. Examination of this application reveals several important points. The method is applied to a process stream, the dominant variable component of which is the desired constituent, carbon dioxide. The variation in the thermal conductivity of the passing gas stream is proportional to the content of carbon dioxide, and the magnitude of this variation is sufficient to make the measurement satisfactory. The built-in sampling device is assumed to provide for heterogeneity of the gas stream, or heterogeneity is neglected. Another more familiar example is the electrometric determination of p H by means of a suitable electrode dipping into an aqueous process stream. In this case the final step in automation may be included —that is, the control of the p H of the stream at some predetermined value. This last step, of course, is not analytical. The success of these and other kinds of methods has led to some excessive optimism. Several years ago, for example, a famous physicist upbraided the writer for taking


students' time to discuss the principles of different kinds of methods of separation and measurement, when, according to him, it was obvious t h a t soon the application of radioactive isotopes would m a k e other methods obsolete. I t was too much to resist the temptation to inquire j u s t how soon one might expect to drive his automobile somewhere in the vicinity of a Geiger counter in order to determine the adequacy of the ethylene glycol in the r a d i ator. I n t h e same vein, in the September 1957 issue of Chemical Engineering appeared the s t a t e ment, " I t ' s not h a r d to foresee the time when completely a u t o m a t i c control laboratories—sampling, analysis, recording—will be comm o n . " Actually, nothing of the kind seems likely in the near future for m a n y materials, as noted later. In all q u a n t i t a t i v e methods some physical property of a system is measured. I n the two examples cited it was thermal conductivity of t h e gas and the potential of an electrode in the solution. Such a single measurement can be related t o the a m o u n t of a given desired constituent in a polycomponent material only if variation in the magnitude of the property measured is a function of this desired constituent. Ordinarily this means t h a t (1) t h e concentrations of other constituents are reasonably constant, (2) these constituents do not affect the p r o p erty measured, or (3) they are m u t u a l l y compensating in any such effects. Usually these restrictive situations are not found. Thus, how can one determine 10% of v a n a d i u m in steels of variable composition by measuring the density of the alloy? The completely automatic method, then, seems t h u s far t o be restricted largely to the determination of a single constituent in the polycomponent system of a process stream. T h u s , we m a y so measure the p H of an industrial water, b u t not by the same measurement the dozen or more constituents present. Another possibility, of course, is such a system which is s t a t i o n a r y w i t h respect t o the measuring device. An example is the continuous determination of ozone in the air a t a given location to show the effect of time, temperature, illu-

Are You Interested in Vitamins

/ .




If you are interested in vitamins, you'll like the Coleman Photofluorometer. I t offers the fastest and easiest way to find and measure Thiamine and Riboflavin in foods, beverages, cereals, pharmaceuticals, animal feeds, plant and animal tissues as well as for metal complexes, and a number of other fluorophors of special interest to medical science. The Coleman Electronic Photofluorometer (Model 12C) is direct-reading, AC line operated, stable and reproducible with a sensitivity range capable of encompassing both trace quantities and concentrates of fluorescing constituents. I t is a typical COLEMAN instrument—sound in science—eminently practical in use and low in cost.

Ask for New Bulletin B-245.







For further information, circle number 29 A on Readers' Service Card, page 97 A VOL. 3 0 , N O . 1 2 , DECEMBER 1 9 5 8





Courtesy Esso Research & Engineering


THE IDEAL PRESS FOR MAKING KB* PELLETS FOR INFRARED SPECTROSCOPIC ANALYSIS Alto for forming pellets for x-ray and other types of spectroscopic analysis * 20 Ton Capacity Hydraulic Press * Accurate alignment in op­ eration * Guided moving platen * Sturdy 3 column construction * Even pressure distribution for uniformly dense pellets * Self-contained * Bench mounted * Hand operated * 9 " χ 9 " platen area * 2 2 " max. vertical opening, adjustable

A multi-purpose unit with quickly adjusted vertical opening, by handwheel, per­ mits many different set-ups in minimum time. Used around the world for RESEARCH, CONTROL & TESTING. Available accessories include: Heat­ ing and Cooling Platens; Temp. Con­ trols, Auxiliary low pressure gauges; Fast air closing; Extruding units; Testing units, etc. Available in 30 and 50 ton sizes. Write for Bulletin.

LOOMIS ENGINEERING & MANUFACTURING CO. Dept. A , Route 4 6 , Caldwell, N . J. Circle No. 30 A on Readers' Service Card, page 97 A 30 A



Figure 4. "Atomcounter" direct reeding emission spectrometer. Jarrell-Ash Co.)

mination, and other variables. Any such measurement of a proc­ ess stream is applicable, of course, to a similar batch sample. If, when the sample is suitably placed, the measuring instrument indicates or records the magnitude of an appro­ priate property, the method is auto­ matic to this extent. Preferably, the meter should be graduated to give values directly in the desired terms, such as per­ centage or milligrams per liter. Steady progress is being made in this direction. For example, over many years the familiar I and I0 values were read on spectropho­ tometers. Then came instruments indicating directly the ratio I/In, or some value related to it, such as absorbance or log A. Finally, came recording instruments, which have become almost indispensable to a b ­ sorption speotroscopists for rapid scanning of large numbers of samples. Even these expensive instruments do not indicate directly the amount of a given constituent, such as the percentage of manganese in steel. This "desired constituent step" has been taken in certain precalibrated filter photometers designed for determinators in routine laboratories. There are interchangeable gradua­ tion scales, one for each constituent to be measured. One simply turns into position the scale applying to the desired constituent.

(Courtesy of

Some progress, in a few instances phenomenal, has been made to have the measuring instrument indicate all, or most, of the constituents of a polycomponent batch sample. A very striking example is the "direct reader" emission spectrometer (Fig­ ure 4 ) . W i t h the sample in position and the excitation switch closed, in much less than a minute a dozen or more minor constituents in metal, for example, will be indicated, each on a separate meter. As important as this development is, the instru­ ment does not (1) take the sample from the molten metal, (2) cast it in a form for handling, or (3) place the sample in position for measure­ ment. This instrument, like x-ray and fluorescence spectrometers and radioactimeters, measures atomic properties. Each element measured has its own detector. In spite of remarkable progress in measuring operations, it seems fair to state t h a t still it is impossible to obtain a complete analysis of a batch sample of m a n y common polycomponent materials. Ex­ amples are highly alloyed steels, water, 'wood, milk, body fluids, ceramic materials, soil, and many naturally occurring organic sub­ stances. N o doubt F. D . Rossini could argue t h a t Ponca City petro­ leum is one of the most difficult. In general, this is attributable t o lack of detectors for specific radi-


Τ") ο you know t h a t the conventional batchtype rotor of the Servall SS-1 Superspeed Centrifuge* may be adapted for the unique Servall 8- to 1-Tube Continuous Flow System?

You can centrifuge gallon quantities of material in one continuing operation at flow rates of up to 400 ml per minute. We have been delivering Servall Continuous Flow Systems for more than a year. This work-proven instrument is available now. Write us for infor­ mation and literature: Bulletin


* For models with rotors, bear­ ing Serial Nos. 1800 and up. Continuous Flow System available with many other Servall Superspeed Centrifuges

' « §

An independent company: not connected with any other centrifuge manufacturer. Established 1934


Circle No. 34 A on Readers' Service Card, page 97 A 3 4 A




Figure 5. Podbielniak "Thermocon" Series 8700, an automatic recording apparatus for low-temperature sepa­ ration, plotting dual curves for boil­ ing point, and thermal conductivity. (Courtesy of Podbielniak, Inc.)

cals, such as sulfate, or compounds, such as pyridine. Although efforts toward automa­ tion have centered on measurement of desired constituents, very signifi­ cant progress has been made in various other unit operations. Only brief mention can be made of these. Sampling may be a very formi­ dable problem in heterogeneous materials, especially solids. For example, how can one automatically sample the Rocky Mountains of Utah for potassium? There are automatic selectors for process streams, of course, as there are also automatic grinders and mixers. Relatively little automation has been achieved in measuring batch samples, whether solid, liquid, or gas. Also little has been done to elimi­ nate the manual preliminary opera­ tions so often necessary prior to separation a n d / o r measurement. Fusion, dissolution, complexation, and adjustment of conditions, such as p H and oxidation state, are com­ mon examples. Unless the constituents of a polycomponent system can be measured in the presence of each other, there must be separation (s). Volatiliza­ tion, precipitation, electrodeposition, and partition methods of sev­ eral kinds involve well known operations. Very great difficulties may center here, as in a petroleum

Figure 6. "Kromo-Tog" Model K-5 separator and recorder for gas chro­ matography. (Courtesy of Burrell Corp.)

sample. This subject is extensive, but mention may be made of some automation. Three noteworthy ex­ amples are Podbielniak stills, mass spectrometers, and gas chromatographs (Figure 5). I n the last two, separation and measurement may not appear to the analyst as discrete operations. Analytically inexperi­ enced teachers who disparage quali­ tative analysis should note t h a t in all three of these methods one has to determine what is being measured. Final note may be made of at least partially automated aids for handling analytical data, such as many kinds of calculators and I B M sorting machines. A complicated example is the automatic data proc­ essing equipment in use at the N a ­ tional Bureau of Standards for emission spectrometric analyses. Simultaneous determinations of up to 18 elements are recorded on punched cards for high-speed statis­ tical analysis for homogeneity. Conclusion

Much has been done in the past 50 years toward automating ana­ lytical methods. Because of their high cost, and often their limited application, some of the present in­ struments can hardly be afforded or justified in many small laboratories. Even if they could be, Ave still seem a long way from having the instru­ ment which will determine all de­ sired constituents, in any relative proportions, in any kind of inor­ ganic or organic substance.