ART. XXXV.-On the Absorption Spectra of certain Blue Solutions. (From Spectro-photometric Determinations by

Mr. F. B. PITCHER.)

[Contributions from the Physical Laboratory of Cornell University. Communicated by Edward L. Nichols.]

II.

BLUES and violets obtained by absorption in pigments and solutions, differ in several respects from those colors which approximate in hue to the longer wave-lengths of the spectrum. They are much less completely saturated, as a rule, and they show irregularities of composition, not commonly met with in absorption reds and yellows.

It is an easy matter, for instance, to find reds the absorption spectra of which show no rays lying beyond the D line, but blue solutions transmit all the wave-lengths of the visible spectrum in considerable quantity; so that an attempt to isolate a pure absorption blue, by increasing the density of the solution or the thickness of the absorbing layer, or by diminishing the intensity of the light traversing the latter, results in the extinction of the shorter wave-lengths along with those lying in the green, yellow and red. It is moreover a matter of common knowledge among spectroscopists that the absorption spectra of very many blue substances, possess a more or less prominent red band, with indications frequently of selective absorption in other regions.

In view of the fact that precise knowledge of the distribution of energy in the visible spectra of this class was almost entirely lacking, our acquaintance with them being for the most part confined to what may be learned by inspection, it seemed of interest to submit the solutions of certain characteristic substances of the class in question, to a spectro-photometric analysis. This has been done, under the writer's direction, by Mr. F. B. Pitcher, from whose Thesis upon this subject the measurements to be described in this paper have been taken. Mr. Pitcher's method differed in few important particulars. from that employed by the writer in his "Spectro-photometric Study of Pigments," already described in these pages. The spectro-photometer was one of Glan's pattern, in which the very ingenious, but far from sensitive, polarizing apparatus had been removed; its place being supplied by a pair of Nicol's prisms before the slit. To the right hand of the spectroscope slit and distant from the same about 60cm, was placed the comparison flame, an Argand gas burner. Its rays, having been rendered parallel by passage through a lens of short focal length, passed through the two Nicol's prisms already referred

* This Journal, vol. xxviii, p. 342.

*

1.

to and entered the lower half of the slit, after total reflection in a pair of small right-angled prisms. At a somewhat greater distance to the left of the slit, was placed a similar Argand burner without a condensing lens; its rays entering the upper half of the slit through another pair of reflection prisms. Between this second flame and the slit was a cell with faces of plane-parallel glass, within which were placed the solutions to be studied. The layer of liquid thus interposed was 1cm in thickness. Figure 1 shows the arrangement of these four reflection prisms. In this excellent device, which is due in part to Crova, the edges of the inner prisms are in contact with each other and with the slit. The spectra of the two sets of rays, thus introduced into the collimator tube, appear therefore with sharply defined adjacent boundaries, separated only by a black "line. Corresponding wave-lengths are in the same vertical plane and it only remains to isolate in turn the various regions to be compared, by means of an adjustable diaphragm in the eye-piece, and to bring the two spectra to equal intensity, in each region successively, by means of the Nicol's prisms.

In its complete form, this type of spectrophotometer should have two sets of Nicol's prisms. One pair in the path of each ray, the

inner Nicol of each pair being fixed with its plane of polarization parallel to the slit. This arrangement is advisable whenever the brilliancy of the ray under investigation is considerable; and it becomes necessary whenever in any region the intensity exceeds that of the corresponding portion of the comparison spectrum.

The instrument in such a case possesses the advantage of complete symmetry. The two rays are reflected the same number of times and at the same angles. They enter the collimator tube polarized in the same plane, and throughout their entire passage to the eye they suffer precisely similar treatment.

There are, however, a great many cases, notably in the study of the light reflected by pigments, and of absorption spectra, where the transmitted ray suffers marked diminution, in which the loss by polarization is such as to render advisable the removal of the second pair of Nicol's prisms. This may be done, whenever it becomes necessary, in order to obtain the requisite brightness of all parts of the spectrum, as in the case of the measurements now under consideration, without sensibly vitiating the accuracy of the determinations.

The instrument having been arranged in the manner already described; the glass cell was placed in the path of the rays

from the second Argand, and the intensity ratio between the spectrum thus transmitted and that of the comparison flame was obtained for nine regions, lying between the A line and the G line of Fraunhofer.

The wave-lengths of these regions, were as follows:

These readings afforded a correction factor for each region of the spectrum, by means of which the selective absorption due to the glass cell and the condensing lens could be eliminated.

Solutions were then made of a number of blue pigments. Three of these were of known composition, namely, Prussian blue, artificial ultramarine and indigo. Five others were commercial preparations of "bluing" of unknown composition. These were included in the list, at the writer's suggestion, for the purpose chiefly of testing the applicability of the spectrophotometric method to the detection of such coloring matters as are not amenable to the ordinary spectroscopic analysis.

The strength of the solutions was so chosen in each case as to show decided color, and at the same time to transmit easilymeasurable quantities of light throughout the spectrum. The results were indicated by means of curves, wave-lengths being taken as abscissæ, percentage of light as ordinates; that transmitted by the empty cell in each region being taken as unity for that particular region.

Figure 2 shows the curves thus obtained for ultramarine (curve a), indigo (curve b) and Prussian blue (curve c). They exhibit the nature of the selective absorption excited by each solution upon the rays traversing it. The horizontal line at the top of each diagram may be considered as representing the brightness of the spectrum of the rays passing through the empty cell, as compared with that of the light transmitted by the solution under investigation. Vertical distances from that line to the curve measure the degree of absorption in each portion of the spectrum. Each curve is characteristic of the solution to which it pertains and, in so far as the number and position of the maxima and minima are concerned, it is independent of the strength of the solution.

These results afford an excellent illustration of the use of the spectro-photometer in the location of absorption bands not

sufficiently marked to be identified by ordinary inspection, and in the detection of regions of maximum brightness in spectra which to the unaided eye do not exhibit any such peculiarity.

The measurement of the five commercial blues gave curves, which approached so nearly to those obtained in the case of the three pigments already mentioned, as to leave no question as to the nature of the substance to which each of them owed its color. Indeed the ease and certainty with which the pigment could be identified was such as to warrant the belief that by the use of this instrument a large number of absorption spectra, which are quite beyond the reach of the spectroscopic methods now in vogue, may be identified with the same degree of certainty which now attends the recognition of blood or chlorophyl. To this end the first step must be the careful determination of a large number of characteristic curves pertaining to the various solutions to which the method is to be extended. This having been properly done, the identification of solutions. in which selective absorption takes place to a measurable extent, will be in most cases a comparatively simple matter.

2.

Mr. Pitcher's experiments were also extended to two interesting solutions in which the addition of an alkali produces a change of color. These, which represent quite distinct classes, were litmus solution and the sulphate of copper. The results are shown in figures 3 and 4. The litmus solution was taken in its three charac- 6 teristic conditions, acid, alkaline and neutral. The acid solution (figure 3, curve d), is marked by almost complete transparency to rays of the regions between B and C, and to slight minima in the neighborhood of the D line, and just beyond F. The addition of ammonia, until the well known color- 2 change indicative of neutrality took place, brought about a modification of the curve to the form shown in curve e. The bright region in the 6 red disappeared; the absorption in the yellow was increased, and the brightness of the remainder of the spectrum was diminished, by

A

C

D

E

F

G

about one half. The addition of ammonia, to the extent necessary to produce the color-change denoting alkalinity, further widened and deepened the absorption band in the yellow and increased the absorption throughout the entire spectrum. The diminution in intensity was greater in the red, however, than in the regions lying beyond the great absorption band.

For the determination of the influence of ammonia upon copper sulphate, a concentrated solution of the salt was prepared and its curve obtained by the method already described.

This solution when viewed by transmitted light, in a layer only one centimeter in thickness, shows but feeble color, and the intensity curve (figure 4, curve g), indicates but little absorption excepting in the extreme red. Noteworthy is the well-marked maximum between the E and F lines and the relatively large amount of absorption in the violet.

The addition of ammonia was found to produce, strong absorption throughout the orange and yellow, the minimum corresponding in position with that observed in the case of alkaline litmus (curve f). Beyond the point of maximum, between E and F, in the curve of the neutral salt, the ammonio-sulphate had become almost completely transparent, with just a trace remaining of the small absorption band lying beyond the F line. (See curve h.)

The almost untouched domain of the spectro-photometric investigation of absorption spectra is an extensive one and by no means devoid of interest. The writer hopes ere long to add to the present contribution, studies of some of the aniline colors and of the influence of acids and alkalies upon the spectra of the so-called "color-test" solutions.

August 1, 1888.