FISHER AWARD ADDRESS
A CENTURY OF COLORIMETRY M. G . MELLON Department of Chemistry, Purdue University, Lafayette, Znd.
T
HE speaker’s first reaction to the announcement that he would be given the Fisher Award was,of course, one of satisfaction. Selection for this high honor, by the procedure used, could not fail to move one. Most of u g no matter how certain we may feel about the essential soundness and value of our work, welcome ita acceptance and approval by fellow workers. This particular manifestation of others’ confidence and esteem is indeed a most generous recognition. Then came less comforting reactions. NO small responsibility rests upon one who followvs the succession of distinguished prior awardea-N. H. Furman, G. E. F. Lundell, I. M. Kolthoff, and H. H. Willard. More important, the present scheme provides no way for the speaker to share the honor with a score of former graduate students who contributed so much to whatever he has been able to do. Subject. Since nothing strikingly new and of ~ ~ f f i c i egeneral nt interest is at hand to present on this occasion, the alternative chosen is to sketch a perspective of the developmellt and aPPlication of colorimetry. Two reasons may be offered for this Palticular selection. First, this area of analytical chemistry has been the speaker’s dominant research interest for over two decades. Secondly, the year 1952 marks the centennial of the publication of Beer’s often-cited paper (10). Practically Contemporary with his paper was the announcement of several chemical and instrumental developments which were to have far-reaching interest and use. These included, among others, Duboscq’s balancing instrument and several of the c~asuicalCOlOr-fOrming reactions. Obviously, just a century has passed since these important events. A backward look over this period should give a viewpoint for better appreciation of modern colorimetry. The essential nature of the developments involved illustrates what has been.going on very generally throughout analytical chemistry during the Past hundred years. The basic outline followed haa been developed by the speaker over the past two decades for all quantitative methods, including colorimetry. In general, the analysis of many PolycomPonent materials employs (1) chemical transformations, if necessary, to render desired constituents measurable; and (2) physical methods to make the measurement. The latter operation include the determination of a property simply as a property. Definitions. Perhaps a t this point we should clarify the usage of the words “colorimetry,” “colorimeter,” and “colorimetric.” When these terms first appeared in the analytical literature is uncertain, but by 1900 they seem to have acquired a fairly definite analyticd meaning for most chemists. Colorimetry was a method of, and colorimeter an instrument for, measurement or analysis by which a colored solution of unknown concentration was compared visually with a standard of known concentration value. Colorimetric is the adjective. With the invention and wide application of other types of instruments, especially photoelectric photometers, visual comparison came to be relatively less used. Although photoelectric instruments bear little resemblance to the early devices, many people still refer to all of them as colorimeters. In all of this chemists are concerned with the nature of the colored component and with its amount-that is, with chemical analysis. To the physicist, a t least in modern times, the terms colorimetry and colorimeter mean something very different. His primary concern is neither the nature of the colorant nor its amount. Rather he measures or specifies the color. TO him, 924
then, colorimetry is a branch of optics deallng with the meaeurc ment of color as color, and colorimeters are the instrument. used. In recent years the speaker has tried to follow the usage ot physicists by limiting the term colorimeter to devices, such atristimulus matching instruments, which measure color in ternia of equivalent stimuli. He has even ventured to call them “stimulimeters.” Then devices uscd lor chemical analysis are absorptiometers. For the purpose0 of this paper, however, the terms colorimetry and colorimeter cover both the early chemical and the modern physical usages. Subsequent to 1900 absorptiometr) 17 used for analytical designation. CHEMISTRY
The term chemistry is used to include the principal preparative steps required to convert a debired constituent, or something equivalent t o it, into a form suitable for measuremerit. In colorimetric methods this covers especially the transformations lvhich yield the colored product to be measured. The general desiderata of such colored forms, arld of any color-forllnng reactions used to produce the eSstenis, have been sumlnari/e,~ elsewhere (89). If the objective of measurenient is color specification, usual11 there is no chemistry. One simply measures the color as is. bfuch chemistry may have been employed, of course, in manufacturing the material. Some desired constituents are self-colored, and to such an extent that their absorptiometric determination is possible. Especially is this true of certain clneses of organic compounds. Con]parable inorganic systemsare relatively uncommon, but a fine example is the permanganate ion. B~~~~~~the great bulk of constituents now considered susceptible t o absorptiometric determination must either be rendered measurable, or their color properties improved, these possibilities are our chief concern at this The advantage of using some reagent to develop or enhance a color of a desired constituent has long been recognled. According to Nierenstein (lo.@,Pliny reported the fist color-forming reagent about 60 A.D. 1% w& a solution of gallnuts. Used to es-
Table I. Ch,.,,nologicai Listing of Selected ~~~l~ &lorForming Reactions Date 1838
Author hmpadius
1852 1841 1856 1862
Herapath Neasler Eggertz
1863 1864 1867 1868 1869
Kemting Sprengel Braun Holland Struve
1872 1870 1876
fronstadt fhhonn Muter
1883
Johnson Fischer Christel
;::: grz& 1884
;:td
1892
1898
$Edale Vernon Richards Jolles and Neurath
Constituent
co
--- -
Fe Fe “8
C
NO,
Reactant “‘OH Fe (CN)( NCS -
-
NO$-
Mo NO1 I; Mo, Ti Au Salicylic acid K Aldehydes Albumin H?S Picric acid Creatinine Tannin AI NCSSiOi--
Nessler’s reagent Self-colored oompound Brucine Phenoldisulfonic acid NCSIH+ Self-colored. in CSS HaOt SnClr-FeCl: PtIs-Schiff’sreagent Picric acid Methylene blue KCN Picric acid L FeCh o 5 y y d extract
+
Moor--
+H
+
V O L U M E 2 4 , NO. 6, J U N E 1 9 5 2
*
timate iron in vinegar, it is said still to serve the same purpose. Today, 19 centuries later, color-forming reactants and their reactions are verr active areas of analytical research. Let us now examine the status of color-forming chemistry a century ago. Time could not he taken for critical searching on what is a difficult library problem; but with the aid of Toe’s extensive bibliography ( 1 2 7 ) arid some incidenhal checnking. Table I was compiled. There is much of possible historical interest, ill this tabulation, but attention is directed t o onlj- a few points. It seems apparent, first of all, that there is little prior to 1852. Ilrrapath‘s proposal of thiocyanate in 1852 ( 4 7 ) as a reagent for iron is a noteworthy coincidence for our purpose. Its continued use in 1052 indicates the importance of the discovery. The fact. too, that we are still arguing 3i)out the nature of the coniples is something to mntemplate. In historical importance probal)ly Xessler’s reagent for ammonia (101) is barelg second to Herapath’s reagent for iron. In onc sense it may stand even first. As the years passed, a t least a score of other reagrnts lvere found for iron, some of which possess certain properties superior to those of thiocyanate. For ammonia, however, Xtwler’s reagent as yet seems to have no real competitor, in spit(, of its recognized defects. So \vel1 known is the i,eac*tiorithat the words nesslerize and nesslerization are not uiiconimon. These ta-o reagent*, then, ivere among the earliest, and they remain today among the most familiar. That neither possesses t,he desiderata of an ideal reagent, is shown, a t least in part, by the extensive literature lvhich has accumulated on each one. h rc>viex of the work on either method \vould constitute a paper of consideratile size. The c*omplete story for these two examples would illustrate a s e q u e n c ~of steps more or less typical of the history of many similar rwgents. The order of the events may well be these: (1) Someone proposes the reagent, usually for a specific situation: ( 2 ) it.s use is e\;tenc!eci, if possible, to other constituents, for fenrealistic anal>-sts expect to find the hypothetical specific reagent for anything: (33 in the course of its application to polycomponent systems, t h e question of what constituent(s) may interfere must be settled, together with the extent of, and the conditions controlling, the interference; and ( 4 ) the nature of the colorforming process, including itieritification of factors affecting t,he rtwtion and the proposal of means for their control, should be settled, Of courw, the ator>-of Herapath’s and of Sessler’s reagents do not apply t o all systems. h fe\v achieve an important place quiclily. Pome never rate highly, and many are of little importance. One cannot avoid \vondering what \vi11 be the fate bj- 2052 oi the numerous reagents being proposed each year. Referring to Table I again, it is evident that rather early such reagents included i1 ) inorganic systems for organic and inorganic constituents, anti ( 2 ) organic systems for inorganic and organic coristituents. This practice has continued through the years. Its occurrence is sufficient’lj- general to raise the quest,ion of whether a-e should not large]>-disregard the distinction between. organic and inorganic chemistry in analytical teaching. \Ve might turn our attention to other early, \vell-kno\vn colorforming reacations, such as phenoldisulfonic acid for nitrate ion, hydrogen peroxide for molybdenum and titanium, Schiff’s reagent for altieh>-des,ant1 molybdate for soluble silica. For t,he sake of brevity, however, let us skip the intervening decades t o sonie\vhere near the present. [That now is concerns us most, at least from a practiral standpoint. The present decade of our survey finds us with a vast array of color-forming reagents and reactions. I n the case of iron and more recentl>- of uranium, for example, the novice in the field must be hewildered by the wealth of information available for a variety of reactants. .4s another example, which one of a dozen or more possible reductants would the newcomer select t o form a hetero!iols blue from the corresponding acid’? Altogether, the
925 possibilities are so varied for many constituents that the speaker tiied several years ago (89) t o classify the more important kinds of reactions, as they apply t o elements, ions, radicals, and compounds. Thousands of papers deal, d t least in part, with different aspects of this subject. To take a specific esample, there must be half a thousand of them relating, directly or indirectly, to the role of heteropoly compounds in absorptiometry (90). Another extensive group concerns metal chelate complexes. The record continues for other kinds of reactants. I n many cases, of course, authors emphasize applications rather than the nature of the chemistry involved. Organization of this v idely scattered material into usable secondary sources has progreseed rapidly in the past two decades, but the work is incomplete and often disconcertingly uncritical. Several aorks of importance may be mentioned. Diehl’s paper ( 2 5 ) marked a milestone in meticulous reviewing of the chelatd t \ pe of complex. This subject hss been brought up to date and extended in a new monograph on chelate compounds by Martell and Calvin (80). llellan’s monograph (84) Mas followed closely ti\ that of Yoe and Sarver (la??),both being on organic reagent?. The amazing amount of such information is revealed still more hv \\.elcher’s treatise (123) which alrsady has reached four volumes, v i t h more t o follow, a t least as supplements. I n general, thebe m orks do not deal exclusively with color-forming reactants. It must suffice here t o state that the chemistry of color-forming reactions is verr much alive as lv3 entar 1952. Recent revim the speaker (85) have emphasized this point. Usually, the investigative activity has the following principal objectives: (1) to find new reagents possessing more nearly the properties of an ideal reactant; (2) t o discover any factors contributing t o variability of results with either old or new reactants, and, as far aq possible, t o determine optimal operating conditions; and ( 3 ) to elucidate the nature of the color-forming process and the product. PHYSICS
1i.e come now to the measurement of the colored system. The operation has two possihle objectives: ( 1 ) the determination of the nature and/or the amount of the colorant present; and ( 2 ) the determination of the color, as such. This usually means measuring the reflecting power of opaque bodies and the transmitting or absorbing power of nonopaque systems. For convenience, separate consideration is given t o methods of analysis of materials, and to methods of designating their color. Currently the speaker prefcrs to call the former ahsorptiometry and the latter colorimetry. CHEMICAL ANALYSIS
On]?- the quantitative aspect of chemical analysis can be considered here. General topics encompass the theory or basis of t,he measurement, and the nature of the means employed. I n the present case this includes the laws of absorption and the principles of absorptiomet,ers. Laws of Absorption. In the course of transmission through a homogeneous, isotropic, nonmetallic medium, homogeneous radiant energy is reduced in intensity by absorption (disregarding surface reflection). If the medium is a substance such as railway signal glass, the magnitude of the reduction depends upon the thickness of the glass. If the medium is a solution, the reduction depends upon both t h e thickness of the solution and the concentration of the absorbing solute in the solvent. Experimental study of such systems led to the formulation of two laws, neither of which was proposed recently. Their present stat,us is of concern, however, hecause absorptiometric chemical analysis rests upon their validity. A publication by Bouguer in 1729 ( I f ) concerns the effect of thickness of the medium. His generalization, currently ksiown as Bouguer’s law, thus far antedates the start of the period of our
926
ANALYTICAL CHEMISTRY
survey. It is a matter of interest that this law has the distinction of having no known exception to its application. For decades this law was attributed to Lambert. Since his publication (69) did not appear until 1760, there seems no question of priority. He did redefine the law in more precise mathematical terms than Bouguer. It has been generally accepted that Beer, in his paper published just a century ago ( f O ) , first related the effect of concentration of solute to the reduction in intersity of a beam of radiant energy traversing a solution. The generalization is known as Beer’s law. Only a year ago Pfeiffer and Liebhafsky (106), following a reexamination of Beer’s paper, decided that he did not treat as symmetrical variables the concentration of the solution and the thickness of the layer measured. Instead, he seems to have been concerned with the combined effect. Although these men found no specific mention of “Beer’s” law until 1889 ( l d f ) , they do recommend contjnuing to call the law by his name. There is not the high degree of conformity for this law as for Bouguer’s. Extensive study has been devoted to chemical and instrumental causes of divergence. It should be kept in mind that Beer’s law presupposes the use of monochromatic radiant energy and the maintenance of the same kind of absorbing entities in solutions of different concentrations. A general statement of the case has been made elsewhere (92). Suffice it to state here, that careful analysts test specific systems for conformity. The combined Bouguer-Beer law, as applied to a solution, may be stated in the Familiar logarithmic form as loglo 1 / T
abc
=
A
in which T is the transmittance, a the absorptivity, b the thickness, c the concentration, and d the absorbance. Absorptance, or unity minus transmittance, has not yet been assigned a symbol in this usage. The absorptivity, a, is a measure of the characteristic absorbing power of the absorber. I n a paper shortly after that of Beer, Roscoe and Bunsen (108) emphasized the importance of the absorptivity, a, long called the extinction coefficient, Many who used Saperian logarithms called it an absorption coefficient. I n this connection, a word on symbols may not be amiss. Those used in the foregoing logarithmic statement of the combined law represent the recent recommendations (63)of a joint committee of the Society for Applied Spectroscopy and the American Society for Testing Materials. The variations in symbols used, even during the past half century, reflect little credit on the ability of scientists to agree upon such usages. Rugged individualism has been rampant. Kayser (69) struggled with the problem and we still have it. Something of the state of confusion is shown by the tabulations of Drabkin ( 2 7 ) and of the speaker (91 ), Only passage of time will reveal the reception accorded the latest proposal. The century we are considering opened, then, with the laws already proposed which are basic for absorptiometric analytical determinations. Subsequent work has concerned the conformity of specific systems to Beer’s law, and with interpretations of, and corrections for, divergence therefrom. Instruments. Measurement of the colored constituent is a physical operation, as ale all other similar measurements in quantitative chemical analysis (88, 103). I n any case, one determines the number of times the standard goes into the unknown. Some writers refer to all instruments used in absorptiometry as photometers. Fundamentally, no matter what the means employed, one determines the light-absorptive power of the colored system. I n a relative way this is accomplished by comparison of the solution of unknown concentration with a standard of known color value. I n an absolute way it is accomplished by determining the transmittance (occasionally the absorptance) or absorbance of the solution. Developments in instrumentation since 1852 are probably the
Table 11. Chronological Listing of Some Early Comparimeters Date
1827 1852 1853 1864 1870 1873
Author Houton-Labilladih Herepath Milller Siemens and Halske Morton Harvey
1876 1878 1879 1891 1895
Hehner Leeds Wolff Gallencamp Kriiss
Davis
Principle Dilution Standard series “Complementiir-colorimeter” Glass standards Duboscq-balancing Round-bottomed “Nessler” tubes Flat-bottomed “Nersler” tubes Cylinders, with taps Row-rack “Kessler” tubes Hehner tubes, Duboscq “head” Wedge cell Hehner tubes, polarization prism
Reference (6.9) (48)
(97) (113) (96) (42) (28)
(46) (71) (186) (56) (67)
most striking aspect of the changes in absorptionietry. One hundred years ago daylight must have been the chief illuminant, test tubes and similar vessels were the most likely containers for the unknown and standard solutions, and the human eye was the receptor for the light transmitted by the solutions. These simple means represent the essential elements of a measuring instrument. Superficially, the changes to date appear very great, and in some ways they are. However, the measuring instrument, no matter how complicated, incorporates a source, a container, and a receptor. To evaluate the significance of the main changes, therefore, one must examine these fundamentals more closely. The items concerned are primarily improvements in the basic elements, or changes in the direction of objectifying and automatizing the instruments. A critical survey of the many developments in details, even for particular instruments, would be an extensive project. .4s such coverage is beyond the intent and scope of this paper, mention is made only of what appear to be especially significant items. RIore about instruments, particularly those of the past two decades, may be found in a number of books (76, 91 110, 116, 1.27). COMPARIJIETERS. Color-matching a solution of u n k n o m concentration with a standard is comparison, and the devices used, when they are named, are “comparinieters” (comparators). Considering the three instrumental elements, it is obvious that early developments probably would concern the absorption cells and the manner and means of their use. I n these cells one may vary relatively either the concentration or the thickness in order to color-mdtch two colored solutions of different concentrations. For a few decades after 1852 the instrumental developments were in this direction. Ultimately, four variations in method were evolved for coniparing the unknown and standard(s): standard series, dilution, duplication, and balancing. For measurements with standard series a notable, although simple, early development was the so-called Seasler tube for vertical observation of a long column of liquid. This container .was found so advantageous that it is still used, especially in water analysis. In fact, we have “nesslerimeters.” Interestingly, the optically plane bottom appeared only a decade or so ago, and the ground-glass top not until 1951. Permanent seriea qtandards for instable systems were an early problem and they still are. Certain inorganic solutions and colored glasses were recommended before 1900, and many modern devices use one or the other. Developments in balancing devices took two general directions: (1)change in total thickness of the colunin of the standard and/or the unknown solutions, by adding or removing qolution from the tube(s); and (2) change in the effective thickness of the standard and/or the unknown by changing the height of a tube or prism dipping into the column(s) of solution. Control of total thickness was accomplished by simple Hehner cylinders, and variations of the idea, as found in instruments described by Kruss and Kruss (68). Control of effective thickness is illustrated in devices based ~
V O L U M E 24, NO. 6, J U N E 1 9 5 2
.
o n Duboscq’s balancing comparimeter. According to Childs (16), this instrument, which achieved such wide distribution and use, came on the scene just after Beer’s paper (1854). It seems to have been described first by Morton (96). Two special variations in the Duboscq-type instrument are of interest for certain applications. The first, for bicolored systems, employs a double absorption cell, one within the other, the effective depth of each being separately variable. The second, Thiel’s “8bsolutekolorimeter” (118), has multiple absorption cells, and uses a standard solution having an absorbance of 0 . 5 throughout t h e spectrum. Other variations n ere made in comparimeters as the decades passed. Included were wedge-type cells, improved optics, direct reading scales, occasionally restriction of the range of the source, and, in one or two cases, photocells as receptors in a balancing instrument. Table I1 lists some early developments. PHOTOMETERS. Leaving comparimeters with this brief statement, we turn t o photometers, including both the filter- and the spectro- types. These instruments do not require constant use of a standard, as do most comparimeters. Consequently, some manufacturers state that this is the distinctive feature, as well as one of the chief advantages, of photometers. Actually, unless one cares to employ unchecked scales, a standard of some kind is required for such checking and/or lor determining a calibration curve, often under the specific conditions likely t o prevail in R given situation. Perhaps the first distinguishing feature of photometers is their means for measuring the brightness (intensity) of the transmitted light, relative to that of the incident light. If I represents the the latter, two main approaches to the problem may former, and IO be recognized among the different ways proposed. In the first kind essentially one determines IO and I . IO may represent the blackness of a line on a photographic plate or the swing of a galvanometer produced by the incident beam, and I the same thing for the emergent beam. In the second kind the incident beam may be divided, one part traversing the solution and the other not (or the solvent only). The measurement problem, then, is to reduce the brightness of the latter beam to that of the former. Since various mechanical and optical photometric devices serve, the name, photometer] may thus have become attached to the instruments. The second distinguishing feature of these instruments is the means used to restrict the light passed to a reasonably narrow spectral band. The origin of all the ideas incorporated is often obscure. One may trace some of them certainly as far back as the “Polarisationkolorimeter” of Kruss (68). I n this connection various works are of reference value (13, 38, 60, 65, 74, 76, ’76, 87, 91, 111). Filter Photometers. I n filter photometers the light-restrictive means usually employed is selectively transmitting glass, an idea used by Leonard0 da Vinci as early as 1519 ( 8 8 )to vary the kind of light passed for studying paints. From the Kruss book (68) the speaker could not select with certainty the date of the first filter photometer. Perhaps Pulfrich’s note, published in 1894 (106),will do. His instrument, as subsequently developed by Zeiss, has had wide use, especially io Europe. Two later visual instruments of this general type are those of the American Instrument Co. ( 3 ) and of E. Leitz, Inc. ( ? 3 ) . The Zeiss, -4minc0, and Leitz instruments illustrate, respectively, the application of mechanical, neutral wedge, and polarization photometric devices. I n the meantime, both photovoltaic and photoemissive types of photocells were commercialized in this country. As early as 1926 Rejmann ( 1 0 7 ) described absorptiometric methods using photocells. In the next two decades a general scramble ensued to substitute this kind of receptor for the human eye. Shortly after 1930 instrument followed instrument (see 119, Vol. I, pp. 82-104). Some designers used photovoltaic cells to avoid amplifying the current. Others argued t h a t photoemissive cells have more de-
921 sirable characteristics. Both one-cell and two-cell instruments were involved, with many variations in electrical circuits. Things have settled down somewhat, but the end probably is not here. An enticing sales argument for photoelectric instruments was their so-called objectivity. Undoubtedly they have displaced many balancing and less expensive visual instruments. However, informed workers in the field know that one, or several, photocells are not in themselves a guarantee of reliability. The first photoelectric instrument known to the speaker had no filter. Then came a variety of filters, glass being the most common. Generally a single glass was used, although this nearly always meant the passage of a fairly wide spectral band of light. The next step was composite filters, of two or more different glasses, in order to narrow the band. The most recent advance is the interfzrence glass filter, some having effective band widths of approximately 15 mM. I n this respect they are thus much better than some so-called spectrophotometers. Spectrophotometers. T o define a spectrophotometer, as distinguished from a filter photometer, is not easy. I n general, the essential difference lies in the means used to isolate the spectral band of light passed, and nearly always in the width of the band. This means is a monochromator in spectrophotometers. Significant instrumental developments probably go back little before 1870. In 1860 Govi (399)had used what seems to be the first simple spectrophotometer, but it was 1873 when Vierordt (120) proposed spectrophotometry as a method of quantitative analysis. The story since Vierordt’s publication is too varied and extensive for more than superficial attention. Thus, 72 possible variations in spectrophotometric arrangements have been suggested ( 5 5 ) . As a whole, the developments concern many details for light sources, monochromators, photometers, absorption cells, receptors, and accessories for handling each of these in a working assembly. The earliest contributions have been summarized by Gibson (38), Kayser (69), and Kruss and Kruss (68). The photographic plate had early use (3X) as a receiving medium, especially for qualitative purposes. But even with a rotating sector or Spekker photometer, introduced by the Hilger company about 1910 a,nd 1931, respectively, such technique is tedious. Things might have been different had the Razek-Mulder curve-recording instrument (98) not been short-lived. As it was, much in the \my of quantitative analysis awaited other developments. Improved visual equipment marked the first of these. Although preceded somewhat by other instruments, such as those of Crova, Glan, Hufner, Lummer and Brodhun, and Wild (38, 68), the Konig-Martens spectrophotometer (64, 81) probably deserves top consideration. It appeared early, and the quality is such that it remains one of the standardizing instruments a t the rational Bureau of Standards. I n this country the introduction of the relatively cheap Keuffel and Esser color analyzer ( G I ) , in spite of design deficiencies, was important. With its direct-reading scales for rotating sector photometer and wave length, transmittance or reflectance values could be quickly obtained. This !+as the speaker’s first spectrophotometer, and it is still used for teaching principles. Other contemporary instruments of similar design were more expensive. The most important lyere the Gaertner (34) and the Bausch and Lomb ( 9 ) about 1922 in this country, and the HilgerNutting (26) a decade later in England. They were equipped with slit width control and polarization photometers. The Bausch and Lomb instrument was unique in its Duboscq-type, variable-depth cell attachment. All visual photometers suffer, of course, from the low sensitivity of the human eye toward the ends of the spectrum, and from the fatigue incident to long-continued matching of the optical field, especially at low light levels. The general status of visual instruments in 1925 is summarized in Gibson’s report (Y8), with its extensive bibliography. As a
ANALYTICAL CHEMISTRY
928
‘
group, they served t o advance quantitative spectrophotometry during the general period of 1900 to 1935. About 1920 we find the advent of the photocell as a receptor in this kind of instrument. At the National Bureau of Standards an assembly was built by Gibson (T?). For various reasons commercialization of the idea progressed slowly. Shortly after 1930 Hardy invented his photoelectric, double-beam recording instrument, with a Van Cittert double monochromator. The photocell is merely a null-point detector. The patent was not issued until 1935 (41). I n the next decade, especially, this instrument had a profound influence, long being referred t o as the recording spectrophotometer. The appellation remains appropriate if the instrument is equipped with an automatic integrator for tristimulus colorimetry ( 2 1 ) The speaker was fortunate to acquire, in Sovember 1934, the first commercial instrument of this design, as manufactured by the General Electric Co. It remains in use today. His new one, of latest design, differs chiefly in details which facilitate operation and maintenance. Although it has four different, easily shiftable ordinate cams, this does not include all the improvements developed in one laboratory (62). Many laboratories could not afford the General Electric instrument, and most routine quantitative work did not require a recorder. Also the spectral band width is too wide (10 mp) for some work with highly selective absorbers, and the range is the visual region of the spectrum. To meet this situation several instruments were introduced in this country. First came those incorporating relatively cheap replica gratings. Their linear dispersion made it easy t o get constant spectral band width throughout the spectrum. The Spectrophotelometer was made by the Central Scientific Co. (112). Then followed the various Coleman instruments ( I ? ) , the best of which was the short-lived Model lOS, which has a compact double monochromator. On it and the Cenco instrument one has a choice of four band widths, the narrowest being 2.5 mp. Shortly afterward (15) came Beckman’s single-beam, quartzprism instrument (Model D U ) , which operates with a spectral band width close t o 1 mp. This and t h e later glass-prism instrument ( 9 3 ) (Model B ) have achieved unusual popularity and thus greatly extended spectrophotometric work. The Hilger Uvispek instrument (49), appearing about 1947, has the same general design. Then, in 1946, Cary announced a new recording instrument to cover both the ultraviolet and visible regions (14). I n addition to having a double beam design, it has high resolving poFver. The speaksr’s late model is built t o record either absorbance or transmittance, a flexibility not previously available. Probably the most spectacular new development is the indicating instrument of the American Optical Co. ( 4 ) . On pressing a w i t c h , the epectrophotometric curve of a sample appears instantaneously on a cathode-ray tube. There is provision for making a photographic record within a minute. The place of this interesting instrument, when further refine& remains to be established. Through the years a number of different ordinate and abscissa scales have been advocated for plotting spectrophotometric curves. As several seem to have advantages for particular purposes, the speaker has never been able to decide that any one of them is the method t o use. Analytical Applications. The selection of a color-forming reaction, and of an instrument for the subsequent measurement, are incidental, albeit important, items in considering an analytical method as a method. But in terms of the “analysis of things as they are,’’ in the words of the late G. E. F . Lundell, our concern is the applicability of such a method. We do not develop an absorptiometric, or any other, procedure t o entertain ourselves. Writing a definitive history of the application of absorptiometric methods during the nineteenth century seems to await the devotion and perseverance of some analytical historian. -4t least
such a survey is unknown to the speaker. Consequently, his statements concerning the period prior to 1900 are based upon a superficial examination of only a few of the most likely sources of information, Because what we now know and are able t o do seem most important, emphasis is placed on the current status of the subject. Brief historical comparison of absorptiometry (colorimetry) with gravimetry and titrimetry seems relevant. The essential principle of the equal-arm balance, the measuring inst’rument of gravimetry, has been known for many centuries. Only recent decades have brought refinements and accessories. Gravimetric methods of analysis probably gained their impetus in the tinip of Lavoisier, and they were extended and largely perfected during the nineteenth century. Sames of many famous chemists of the period might, be cited as contributors, possibly the greatest of all of them being Berzelius. Tit,rimetry was started much later, the first such method being attributed t o Descroizilles (24) in 1806. Developments followed rapidly so that these methods Fere well established by 1900, Mohr’s famous “Titrierniethode” having reached its sixt’hedition of 887 pagesin 1886 (94). Flasks, pipet’s,and burets had become the, means of measurement. Significant inst,rumental changes in the twentieth century comprise chiefly a number of different means for determining end points, including some automatization of the operation. I n this later period, of course, arose work on a micro scale, both in titrimetry and gravimetry. Table 111. Percentage of Total Articles Devoted to Specific Methods Kind of Method Absorptiometric Gravimetric Titrimetric
1929-30 11.1
1980-51 24.1
18.7
7.3 27 0
30.8
Turning now to absorptiometry, the application of methods as we know them today seems t o have been a half century behind those for titrimetry. There are undoubtedly many incidental references scattered through the journals prior to 1900. Examples are Herapath’s and Sessler’s methods for iron and ammonia, respectively, in water analysis, and Eggertz’s method for carbon in steel. The authors of famous early texts apparently considered coloiimetry much less important then than gravimetry and titrimetrj . Thus, works such as those of Fresenius (SS’), Rose (SO, I O S ) , and Wohler (124) do not mention colorimetric methods. As late as 1900 the fourth edition of Lunge’s famous treatise on applietl analysis ( 7 8 ) contains only a small number of such methods. Development8 occurred a t a n increasingly accelerated rate as the decades passed, both in the chemistry of preparing measurable systems and in the instruments used. Comparable progress marked the application of such procedures to materials. The earliest book known to the speaker which is devoted to absorptiometric methods is by Snell (IId), published just three decades ago. Whether this Emall book served to catalyze the production of those that followed is probably undeterminable, but things did begin to happen. KO doubt the contemporary instrumental developments played an important part. For some years the subject expanded in a general direction t o include the determination of both inorganic and organic constituents in a variety of materials. Yoe’s comprehensive work in 1928 (Id?’) did much t o stimulate interest. The second edition of Snell’s book (with Mrs. Snell) had grown by 1936 t o two sizable volumes. Less comprehensive but competing European books are represented by those of Allport ( Z ) , Delory ( d S ) , and Lange ( 7 0 ) . Now we await the completion of the third volume in the third edition of the treatise of the Snells (116). During this period more and more absorptiome&ric methods were being included in each succeeding edition of official methods, such as those of the hmerican Public Health Association (6) and the American Society for Testing Materials (6).
V O L U M E 2 4 , NO. 6, J U N E 1 9 5 2 Several current works are devoted entirely t,o such methods, as applied to specific kinds of materials. Haywood and Wood (49) limited theirs t,o the use of the Hilger filter photometer for the determination of certain metals in alloys. Sandell’s critically evaluatcd methods cover most of the metals (110). Ext,ensive collections of methods for clinically important, coristit,uents appeared first in Hoffman’s compilation (61) and much later in Fiater’s collection (31). The Baul;ch and Lonib (a), the Leitz (Ti?), and the Pulfrich ( 6 6 ) collections contain directions for using t h w r instruments for various clinical determinations. The hooks by JIorton ( 9 6 ) , Heilnieyer ( 1 , .@), and Hunter (64)are still morr sprcialized in being devoted t o Rpectrophot,oiiietric nirtthotis applied t o biochemical prohienis. .in occasional title m:iy he ~iiisleatlirigin the implication that such methods require a 1):irticular instrument. This will rarely be true esccpt in the siniultantwus det,ermination of polycomponent constituents in 11.3 where narroL\-hand instruments are needed. r has seen all these secondary sources on applications appear. .lcwmpanying this changing scene, he has observed another, diff(~rentbut just as striking. I t is in orginal-source plriodical?. Thus, Table 111 s h o w what happened in 20 years in .~S.\LYTIC.II, CHEMISTRY. S o doubt other analytical journals she\\- the PXIIIP trend. I n this . ~ u l v e yof applications t,wo specific items arc’ of particuh i . c.uri~i~nt interest. The first conccsrns the range of concentration ot’ dt4red constituent considered measurable by absorptiomc~tricnicans. Perhaps from the earliest times such methods n-ere thought of as being limited to low percentage ranges, often of t,he ordt.1- of 1 nig. per liter, as in xvater analj . This idea persist,s, at loast implied, in the title of Sandell’s hook, “The Colorimctric I)cLtc>rrnination of Traces of ;\Ietals.” .In interesting developnient during the past decade is the extension of this range to include at 1c.ast crrtain constituents ivhich are major components in niat,(arial%.Several years ago Mehlig (82, 83),by careful att,ention to details, was able to handle ores containing u p to 58TC of i i ~ ~ rthrx i . rcsults being thz equal of titrimetric values. More rercBiitl>. IIiskey (bo),in approaching the problem differently, has dealt with the theory involved and n-ith certain instrumental posssihilities. Bastian’s papers ( 7 ) report result,s obtained in this manner on certain industrial matvrials. Thus, it would seem that such methods, like t,hose ol gravinictry and titrimet,ry, are being atl:rptetl t o the determination of major constitucnts. The secori.1 item concerns polycomponent systems. Interfering coiistitucnts have long been the bane of analysts, and separations of sonic kind have equally long heen the answer. The important placc ol t hcsc separative processes has hecn summarized elsev-hcre (86 J . T o obviate thew gcnerail!. objcctioiiable steps thcre havc heeii conic indirect methods available, as in the final stage of the gravimetric determination of the alkalies in rock analysis. .itlaptation of absorptiometric means t o thc simultaneous deterinination of two or more absorbing constituents is probably the outstanding single contribution of spectrophotometry to quantitative anal!.&. Such applications arc most important in the infrared rrgion of the spectrum, but they are also very valuable for certain polycomponent colored systems. A fiiic example is tlic ti-trrniination of the two chlorophylls in tlirl presence of each other (18). I l u c h earlier, of courst’, \\-eigcrt (122) ~ r o r k e don a four-component system of dyes. THE %IEASURE.MENT OF COLOR
The mctallurgist making stcel is conceriied iritli knowing \vliether thtxre is carbon in his iron, hon- niuch there is, and what propei,tics, such as hardness and tensile strength, heat treatment gives the finished product. Similarly, a nioderri testing laborator). may be concerned not onlj- with determining the nat.ure and/ or the amount of colorant in a material, but also with measuring t h i : color itself. The colors observed on viewing objects depend upon several vxriahles, probably the most important being (1) the spectral dis-
929 tribution of the illuminant, ( 2 ) the conditions of illumination and viewing, (3) the nature of the object, and (4)the “normalit,y” of the observer. The general subject of color, including its description and measurement, is so extensive that here only a f e n items can be included. Several recent books (12, 29,40,58,91,chap. IX; 100; I % ) , summarize contemporary information significaiit for our purpose. __ _______ ~~
Table IT. Date 1861 1872 1875 1881
1886 1892
Chronological Listing of Early Instruments for 3feasuring Color
Autlior Principle Staniiner Ciiroi~ioscop~ Stanimer Glass standards King Caraine! solution Crooked et al. Soliitions Lovibond Glass standards Hazen Pt and Co cliloride solutions
Refercncc 1116)
iff;) 10.7)
(2;)
(7,)
‘44) -_.-
Coiisicltration of possible variability in combinations of the factors mentioned indicates a t once the difficulty of objective csperimrntal work. I n brief, invcstigators have tried t o agree on reporting in terms of a standard illuniinant and a normal observer. Control of the illuniinant for color-matching purposes took two gent.r.al directions: (1) the additive combination of light sources, each of iiarroiv spectral range; and (2) the subtractive combination of elements, each absorbing selectively, hut differeritl!-, from a continuous source. Long before such instrunieiitation alqwared, \re had specification and nicasureni~ntin terms of color-matching the o1)ject ivith so-called standard samples. Such means are still important today, as evidenced by atlases surh as those of 5Iaerz and Paul ( 7 9 ) , llunsrll (99),and Ostn-:tld (56). The evolution of color names, and their relation t o surh standards, can only be mentioned. Practical instrumentation for the analyst turned toivard subtractive assemblies. Table IT lists a f e n early instruments. Such means remain in use in instruments such as tho Stammer, Union, and Donaldson colorimeters, the Lovihoritl tintometer, and the Saybolt chromonieter. Let us skip many years of work on attempted dcvelopmeiit of additive assemblies to come near present tristimulus practice. This technique amounts to det,ermining, or calculating, the amounts of three arbitrarily fielect,edprimaries required to colormatch a sample. The report of Troland’s colorimetry committee of the Optical Society of America (119) was an important, event in this country. I t marked a step toward calculating tristimulus values, from a spectrophotometric r u r v ~ :of the sample, in t.erms of a given illuniinant and a “normal” observer. Both trichroniatic and nionochroma tic spwifications of color were included. Even more important was the report of t h e International Comniission on Illuminat,ion (I.C.I.) in 1931 ( 1 9 ) , including a description of illuniinants .I,B, and C. Hdrdy’s book ( 4 0 ) greatly facilitated application of the methods. Finally, there may be nientioncd the report of the new colorimetry Committee of the Optical Society of America by Jones et al. (57). This book, just published, marks the culmination of years of study by a distinguished committee. Estensivc applicat,ion of the I.C.I. method depended upon easy acquisition of the necessarl- spectrophotometric curves. Of course, the General Electric recording instrument was the answer. Finally, in 1919, came the announcement of two automatic tristimulus integrators (21, 104) which yield trichromatic values directly as the curves are drawn. Such equipment is now commercially available. Thtxse developments cnable us t o measure a color and to rcport the numerical values in percentages of the three primaries, red (z),green (y), and violet ( z ) , and in terms of dominant wave lengt,h ( V I P ) , colorimetric purity (yo),and luminance (yo). hccompanying these developments came the introduction of a number of so-called tristimulus photoelectric colorimet,ere. I n general, these instruments, usually much simpler arid less ex-
930
ANALYTICAL CHEMISTRY
pensive than good spectrophotometers, yield coIor readings with their three filters often approximating closely the comparable trichromatic data yielded by calculation from a spectrophotometric curve. The Hunter instrument is an example (36). CONCLUSION
This resume has touched some of the significant developments in colorimetry and absorptiometry during the century beginning in 1852. The last half of the nineteenth century brought many of the fundamental developments, both in chemistry and physics, on which later progress was based. The twentieth century has seen their growth and flowering. Kow we are reaping the harvest. The number of relevant books cited in the bibliography reflects the present situation. There were no such works in 1852. The period opened with very feF color-forming reagents or reactions in use. We begin 1952 with a formidable amount of such information. Instrumentally, we have extended our facilities for analysis from visual observation of a test tube in daylight to complicated devices which will indicate or record numerical values, or even logarithmic functions of these data. An increasing number of simultaneous polycomponent determinations are practical. In fact, we seem now to be in a better position physically to make measurements than we are chemically to produce systems satisfactory for measurement. Water analysis appears to have been the chief example of applications for some years after 1852. Xow we not only have a polyvolume general treatise, but also a number of single volumes, devoted to specific materials. Every issue of our journals brings more methods. Finally, in color measurement we have come from rough colormatching, through such matching in terms of carefully standardized systems, to numerical terms. The latter include tristimulus colorimetric readings and I.C.I. values based on spectrophotometric curves. In terma of Lord Kelvin’s basis of evaluating u subject, attainment of this numerical stage should have advanced colorimetry to the level of science. LITERATURE CITED
(1) Abderhalden:,E., editor, “Handbuch der biologischen Arbeitsmethoden, section on color measurement of body fluids, Abt. 11, Teil 2, Halfte 2, pp. 2337-66, Berlin, Urban & Schwarrenberg, 1931. (2) Allport, N. L., “Colorimetric Analysis,” London, Chapman and Hall, 1945. (3) -4merican Instrument Co., Bull. 2080 (1940). (4) American Optical Co., “Rapid Scanning Spectrophotometer,” Preliminary Announcement, C9-1051 (1951). (5) American Public Health Association, “Standard Methods for the Examination of Water and Sewage,” New York, American Public Health Association, 1946. (6) American Society for Testing Materials, “ASTM Methods for Chemical Analysis of Metals,” Philadelphia, American Society for Testing Materials, 1950. (7) Bastian, R., et al., .4~.4~.CHEM.,21, 972 (1949); 22, 160 (1950); 23, 580 (1951). (8) Bausch and Lomb Optical Co., “Bausch and Lomb RIonochromatic Colorimeter,” Rochester, N. Y., Rausch and Lomb Optical Co., 1951. (9) Bausch and Lomb Optical Co., Catalog D-111,p. 235. (lo) Beer, A,, Ann. Physik u. Chem. [2],86, 78 (1852). (11) Bouguer, P., “Essai d’optique sur la gradation de la lumihre,” 1729. (12) Bouma, P. J., “Physical Aspects of Color,” tr. by R. DeGroot, New York, Elsevier Book Co., 1947. (13) Brode, W. R., “Chemical Spectroscopy,” New York, John Wiley & Sons, 1943. (14) Cary, H. H., Rew. Sci. Instruments, 17, 558 (1946). (15) Cary, H. H., and Beckman, A. O., J . Optical SOC.Am., 31, 682 (1941). (16) Childs, E., “The Tools of the Chemist,” p. 162, New York, Reinhold Publishing Corp., 1940. (17) Coleman Electric Co., Bull. 104. (18) Comar, C. L., and Zsoheile, F. P., Plant Physiot., 17, 198 (1942).
(19) Commission Internationale de I’Eclairage, Proceedings Eighth Session, pp. 19-29, Cambridge, England, 1931. (20) Crookes, W., et al., Chem. News, 43, 174 (1881). (21) Davidson, H. R., and Imm, L. W., J . Opfical SOC.Am., 39, 942 (1949). (22) Davis, G. E., Chem. S e w s , 27, 299 (1873). (23) Delory, G. E., “Photoelectric Methods in Clinical Bioohemistry,” London, Hilger and Watts, 1949. (24) Descroizilles, F. A. H., “Xotices sur l’alcalim8tre,” 1806; see J . Chem. Education, 28, 508 (1951). (25) Diehl, II., Chem. Revs., 21, 39 (1937). (26) Dowell, J. H., J . Sci. Instruments, 8 , 382 (1931); IO, 153 (1933). (27) Drabkin, D. L., section on “Spectrophotometry” in “Medical Physics,” by 0. Glasser, Vol. 2, p. 1070, Chicago, Year Book Publishers, 1950. (28) du Fresne, R., “Trattato della Pittura di Lionardo da Vinci,” Paris, G. Langlois, 1651. (29) Evans, R. M., “Introduction to Color,” Kew York, John Wiley 8: Sons, 1948. (30) Finkener, R., “Handbuch der analytischen chemie,” 6th ed., Vol. 2, Leipsig, J. A. Barth, 1871. (31) Fister, H. J., “Manual of Standardized Procedures for Spectrophotometric Chemistry,” New York, Standard Scientific Supply Co., 1950. (32) Frerichs, R., “Handbuch der Physik,” section on photographic spectrophotometry, T’ol. XIX, pp. 688-704, Berlin, J. Springer, 1928. (33) Fresenius, C. R., “System of Instruction in Quantitative Chemical Analysis,” edited by S. W.Johnson, New York, John Wiley 8: Sons, 1875. (34) Gaertner Scientific Corp., Bull. 142-74. (35) Gallencamp, T,,Z . angelc. Chem., 5, 549 (1891). Laboratory, Pamphlets, 1950. (36) Gardner, Henry -I., (37) Gibson, K. S.,J . Optical SOC.Am., 2-3, 23 (1919); J . Optical SOC.Am.-Rei. Sci. Instruments, 7, 693 (1923). (38) Gibson, K. S.,et u l , , J . Optical SOC.Am., 10, 169 (1925). (39) Govi, G., Compt. rend., 50, 156 (1860). (40) Hardy, A. C., “Handbook of Colorimetry,” Boston, Technology Press, 1936. (41) Hardy, A. C., U. S. Patent 1,987,441 (1935); J . Optical Soc. Am., 25, 305 (1935). (42) Harvey, S.,Chem. S e w s , 27, 262 (1873). (43) Haywood, F. W., and Wood, A. 8. R., “Metallurgical Analysis,” London, -4.Hilger, 1944. (44) Haren, 9.,Ani. Chem. J., 14, 300 (1892). (45) Hehner, O., Chem. S e w s , 33, 185 (1876). (46) Heilnieyer, L., “Spectrophotometry in Medicine,” tr. by A. Jordan and T. L. Tippell, London, -4.Hilger, 1943. (47) Herapath, T. J., J . Chem. Soc., 5, 27 (1852). (48) Herapath, T. J., J . prakt. Chem., 56, 255 (1852). (49) Hilger and Watts, Ltd., Pamphlet S.B.318/6 (1949). (50) Hiskey, C. F., et al., ANAL.CHEM.,21, 1440 (1949); 22, 1464 (1950); 23, 506, 1196 (1951). (51) Hoffman, W ,S., “Photelometric Clinical Chemistry,” New York, W.Morrow and Co., 1941. (52) Houton-Labillardiitre, M., J . Pharin.! 1827, 610. (53) Hughes, H. K., et al., Ax.4~.CHEY.,in press. (54) Hunter, F. T., “Quantitation of Mixtures of Hemodobin Derivatives by Photoelectric Spectrophotometry,” Springfield. Ill.. C. C Thomas. 1951. (55) Institute of Paper Chemistry, Paper Trade J., 105, S o . 18, 135; No. 19, 27 (1937). (56) Jacobson, E., Granville, IT. C., and Foss, C. E., “Color Harmony Manual,” Chicago, Container Corp. of America, 1949. (57) Jones, L. A,, et al., “Science of Color,” New York, T. Y. Crowell Co., 1952. (58) Judd, D. B., “Color in Business, Science, and Industry,” New York, John Wley & Sons, 1952. (59) Kayser, H., “Handbuch der Spektroscopie,” Vol. 1, pp. 1-128, “History of Spectroscopy,” Leiprig, S. Hirael, 1900-12. (60) Kessler, H., section on colorimetry in “Handbuch der biologischen Arbeitsmethoden,” E. Abderhalden, ed., Abt. 11, Tiel 1, pp, 689-736, Berlin, Urban & Schwarrenberg, 1925. (61) Keuffel, C. W,, J . Optical Soc. Am.-Rea. Sci. Instruments, 11, 403 (1925). (62) Kienle, R. H., and Stearns, E. I., Inslrumenfs, 20, 1057 (1947). (63) King, J. F.,J . Chem. Soc., 28, 1052 (1875). (64) Konig, A., Ann. Phusik u. Chem. [3],53, 785 (1894). (65) Kortum, G., “Kolorimetrie und Spektralphotometrie,” Berlin, Springer-Verlag, 1948. (66) Krebs, IT., “Clinical Colorimetry with the Pulfrich Photometer,” Jena. C. Zeiss, 1936. (67) Kriiss, H., 2.Instrumentenk., 14, 283 (1895). (68) Kriiss, G., and Kriiss, H., “Kolorimetrie und Spektralanalyse,” Leipsig, L. Yoss, 1909.
931
V O L U M E 2 4 , NO. 6, J U N E 1 9 5 2 Lambert, J. H., “Photometria sive de mensura et gradibus luminis, colorum, et umbrae,” 1760. Lange, B., “Kolorimetrische rlnalyse,” Berlin, Verlag Chemie, 1941.
Leeds, A. R., Ch&. News, 37, 229 (1878). Leitz, E., Inc., “Handbook for Rouy Photometcr,” New Tork, E. Leitz, Inc., 1948. h i t s , E., Inc., Pamphlets A to F, “Leifo Photometer,” 1937. Ley, H., “Handbuch der Physik,” section on spertrophotometry, 1’01. XIX, pp. 613-66, Berlin, J. Springer, 1928. Lowe, F., “Handbuch der Physik,” section on colorimetry, Vol. XIX, pp. 667-87, Berlin, J. Springer, 1928. Lothian, G. F., “Absorption Spectrophotometry,” London, Hilger and Watts. 1949. Lovibond, J. W., J . SOC.Chem. Ind., 6, 522 (1886). Lunge, G., “Chemisch-technische Untersunchungsmethoden,” Berlin, J. Springer, 1900. Maern, A., and Paul, hl. R., “Dictionary of Color,” Sew Tork, McGraw-Hill Book Co., 1950. hlartell, A., and Calvin, M., “Chemistry of the Metal Chelate Compounds,” Xew York, Prentice-Hall, Inc., 1951. Martens, F. F., Verh. deut. p h y s i k . Ges., 1, 280 (1899). Mehlig, J. P., IND.ENG.CAEM.,ANAL.ED., 9, 162 (1937). Mehlig, J. P., and Hulett, H. R., Ibid., 14, 869 (1942). Mellan, I., “Organic Reagents in Inorganic Analysis,”’ Philadelphia, Blakiston and Co., 1941. MellOn, M.G.,A S A I . . CHEM., 21, 3 (1949); 22, 2 (1950); 23, I
2 (1951): 24, 2 (1952).
Ibid.,22, 1342 (1950). hlellon, M. G., “Colorimetry for Chemists,” Columbus, Ohio, G. F. Smith Chemical Co., 1945. Mellon, RI. G.. “Methods of Quantitative Chemical Analysis,” New York, hincmillnn Co., 1937. Mellon, M. G., Proc. Am. SOC.TeatingiVfuterials, 44, 733 (1944). Mellon, M.G., Record Chem. Progress, 11, 177 (1950). Mellon, AI. G.. et al., “Analytical Absorption Spectroscopy,” New York. .John Xiley 8: Sons, 1950. Ibid., pp. 94-101. Miller, W.C., et al., J . Optical SOC.Am., 39, 377 (1949:. Mohr, F., “Lehrbuch der chemisch-analytischen Titriermethode,” 6th ed.. by A. Classen, Braunschweig, Vieweg und Sohn, 1886. Morton, H., Chem. IVews, 21, 31 (1870). Morton, R. 9., “Application of Absorption Spectrophotom-
etry to the Study of Vitamins, Hormones. and Coenzymes,” London, A. Hilger, 1942. Muller, A., J . prakt. Chem., 60, 474 (1853). Mulder, P. J., and Ranek, J.. J . Optical SOC.A m . , 20, 155 (1930). Munsell Color Co., “Munsell Book of Color,” Baltimore, Munsell Color Co., 1929.
(100) (101) (102) (103)
Sational Bureau of Standards, Civc. 478 (1950). Nessler, J., Chem. Gat., 14, 446 (1856). Nierenstein, M., Isis, 16, 439 (1931); Analyst, 68, 212 (1943). Patterson, G. D., Jr., and Mellon, M.G., J . Chem. Education,
26, 468 (1949). (104) Peterson, E. W.,et al., J . Optical SOC.Am., 39, 1055 (1949). (105) Pfeiffer, H. G., and Liebhafsky, H. A., J . Chem. Education, 28, 123 (1951). (106) Pulfrich, C., 2.Instrumentenk, 14, 210 (1894). (107) Reimann. S.P., Proc. SOC.Ezptl. Biol. Med., 23, 520 (1926). (108) Roscoe, H., and Bunsen, R., Ann. Physik u. Chem., [2], 101, 235 (1857). (109) Rose, H., “Ausfuhrliches Handbuch der analytischen Chemie,!’ Vol. 2, Braunschweig, F. Vieweg und Sohn, 1851. (110) Sandell, E. B., “Colorimetric Determination of Traces of Metals,” 2nd ed., New York, Interscience Publishers, 1950. (111) Scheibe, G., section on spectrophotometry in “Handbuch der
biologischen Arbeitsmethoden,” E. -ibderhalden, ed., Berlin, Urban & Schwarnenberg, 1931. (112) Sheard, C., and States, Jl. S . , J . Optical SOC.Am., 31, 64 (1 941). (113) Siemens, E. W., and Halske, J. G., 2. anal. Chem., 3,218 (1864). (114) Snell, F. D., “Colorimetric Analysis,” New York, D. Van Nostrand Co., 1921. (115) Snell, F. D., and Snell, C. T., “Colorimetric Methods o f Analysis,” New York, D. Van Sostrand Co., 1948-53. (116) Stammer, C., Dinglers polytech. J., 159, 341 (1861). (117) Ibid., 203, 137 (1872). (118) Thiel, A., “ilbsolutkolorimetrie,” Berlin, de Gruyter and Co., 1939. (119) Troland, L. T., et al., J . Opticnl SOC.A m . , 6, 527 (1922). (120) T’ierordt, K., “Die Anwendung des Spektralapparatus zur
(121) (122) (123) (124) (125) (126)
Photometrie der Absorptionsspektren und zur quantitativen chemischen A4nalyse,”Tuobingen, H. Laupp, 1873. Walter, B., Ann. Physik u. Chem. [3], 36, 502, 518 (1889). Weigert, F., Ber., 49, 1496 (1916). Kelcher, F. ,J., “Organic Analytical Reagents,” Kew Tork, D. Van Yostrand Co.. 1947-48. M’ohler, F., “Handbuch of Inorganic Analysis,” ed. by A. W. Hoffmann, London, Walton and Maberly, 1854. Wolff, C . H . , Pharm. Ztg., 24, 587 (1879). Wright, W.D., “Measurement of Color,” London, A. Hilger,
1944. (127) Yoe, J. H., “Photometric Chemical Analysis,” Vol. I, “Colorimetry,” Sew Tork, John \Gley & Sons, 1928. (128) Toe, J. H., and Sarver, L “Organic .Inslytical Reagents,” New York, John Tiley & Sons, 1941. ~ ~ E C & I I ’ Efor D review .January 29, 1952. Accepted .4pril 14, 19.52. Adrlresa rlrlirercrl o n the presentation of the Fisher .4aard in Bnalytical Chemistry, a t the l 2 l s t llceting of the AMERICASCHEMICAL SOCIETY, Buffalo, N. Y.
Hot-wire Manometers for Chemical Applications HANS VOX UBISCH, Department of Physics, A . B . Atornenergi, Stockholni, Sweden An investigation has been made with the aim of iniproving hot-wire manometry, with emphasis on the micron and lower millimeter range. The present paper is a summary of results, which are published elsewhere in full detail. Procedures and formulas for precision manometrj- on pure gases and on binary mixtures, together with the manometric constants for 19 gases, are given. However, a hot-wire gage is not an “absolute manometer,” and a certain amount of empirical calibration cannot be avoided. Manometers using an electrical circuit including a feed-back action operate in the entire range between and 50 mm., if air is considered. They can also be used as conductivity meters, and arc especially suitable as a head of control or monitoring devices. The results presented suggest how to design gages having optimum or special performance. They enable users to predict the response of their gages to gases not previously tested.
H
OT-wire or Pirani manometers have attracted great attention for a long time because they cover a range of pressure where suitable direct-reading liquid manometers are not availablc, they are able to measure vapors of easily condensed compounds, they show very little interference with the gases or vapors dealt with, and they can be constructed in a neat and convenient form. T h e useful pressure range of hitherto available types of hot-Fire gages (6) is, for chemical applications, believed t o be from t o 10-1 mm. of mercury, or perhaps only from 10-8 mm. of mercurv if t h e occasionally large zero drift is taken into account. The lower limit should be sufficient for very many applications, but t h e upper one a t 10-1 mm. of mercury, or a t about 1 n m . of mercury for a thermocouple gage, is too low to provide an overlap 4 ith a simple liquid-level manometer. Recent investigations (11, 12) have shoxn, however, t h a t t h e range can easily be extended upward t o tens of millimeters of mercury in t h e case of air, resulting in a gage t h a t has some essentially new features. During these investigations it h a s also become necessary t o deal a little more in detail with t h e