Page:Journal of the Optical Society of America, volume 30, number 12.pdf/86

18. Calibration Data on General Electric Recording Spectrophotometer. J. L. Michaelson and W. R. Fanter, General Electric Company.

Due to the general interest shown by members of the Optical Society of America in spectrophotometry, it has occurred to the authors that it would be interesting to present calibration data obtained from various General Electric spectrophotometers. Therefore, during the past two years we have recorded rather complete data on the performance of these instruments, which will be presented.

Bibliography: A. C. Hardy, J. Opt. Soc. Am. and Rev. Sci. Inst. 18, 96 (1929). J. L. Michaelson, J. Opt. Soc. Am. 28, 365 (1938). Orrin W. Pineo, J. Opt. Soc. Am. 30, 276 (1940).

19. The Importance of Optically Clean Absorption Cells in the Determination of the Concentration of Dye Solutions. S. Q. Duntley, Massachusetts Institute of Technology.

Spectrophotometric transmission data are often used in determining the strength of dye solutions. Customarily, effects due to the presence of the absorption cell and to the solvent are compensated by placing a duplicate cell filled with undyed solvent in the comparison beam of the photometer. It is the purpose of this paper to investigate the errors in dye strength determinations which may result when this compensation is imperfect. Such imperfect compensation may occur during a series of measurements due to contamination of the outer cell surfaces by fingerprints, spilled dye solution, etc. By the application of differential calculus to Beer’s law, it can be shown that the greatest change in transmission for a given percentage change in dye concentration occurs when the transmission is 37 percent. It can further be shown that the errors arising from imperfect cell compensation can be slightly reduced by using a lower value of transmission. However, it is never possible to reduce the error to negligible proportions by this device. For example, with a pair of cells which, when filled with solvent, yield an apparent transmission factor of 99 percent, the resulting error in ∆C/C is greater than 50 percent even when the sample transmits no more than 15 percent. This leads to the conclusion that the use of concentrated solutions does not excuse the operator from checking the condition of his cells before each measurement. The treatment is extended to show the effect of imperfect cell compensation on data taken with the “five times,” cam available for the Hardy recording spectrophotometer.

20. The Viewing Angle of Reflectometry. Elliot Q. Adams, Lamp Development Laboratory, General Electric Company, Nela Park.

Many instruments for the rapid measurement of the reflectance of matte surfaces provide for illumination, or viewing, at an angle of 45°, although it has been known for some time that the angle of equivalence with diffuse illumination is, for most matte surfaces, appreciably greater than 45°. Thus A. H. Taylor and C. H. Sharp, in discussing^[1] the Taylor absolute reflectometer, speak of viewing at 50°, while McNicholas^[2] reports; “Thus for matt samples of the kind herein studied (of not extremely low reflectance), one would be quite safe in choosing an angle of observation of 55° and assuming an accuracy of 1 or 2 percent.” It appears not to have been pointed out that there is a priori reason for illuminating at an angle of approximately this value: If diffuse illumination is to be replaced by illumination from a finite number of points, the nearest equivalence will be secured by locating the sources at the corners of a regular polyhedron, i.e., of a regular tetrahedron, octahedron or cube, the number of sources being, respectively, 4, 6 and 8. For an opaque plane surface, the sources behind the plane will not contribute to the illumination, hence may be omitted. If the remaining sources are so located as to make equal angles with the normal to the surface, and if the reflectance of the surface does not vary with azimuth, the normal brightness will, by symmetry, be unchanged if the 2, 3 or 4 sources are replaced by a single source of the total luminous intensity, at the location of one of them. The angle from the normal will be, in each case, that between the threefold and fourfold axes in the cubic crystallographic system. This angle is tan^-1${\sqrt {2}}=$54°44′8+″. To the degree of accuracy of the reciprocity principle, the apparent reflectance will be the same for normal illumination and viewing at an angle of 54'44’8",

21. An Extreme Case of the Performance of the Eye versus that of the Spectrophotometer.^[3] I. H. Godlove, E. I. duPont de Nemours and Company.

The addition of small amounts of Crocein Scarlet to Tartrazine was supposed by Draves^[4] to be a case where the eye can detect smaller additions than a spectrophotometer; but it was shown by Nutting,^[5] using the Hardy spectrophotometer,^[6] that the reverse is true if the measurements are not confined to a single wave-length, as was done by Draves. This case is one in which the two spectral absorption curves involved are very different in shape. The opposite extreme, where the curves have similar shapes, includes cases where, at least under industrial working conditions, the spectrophotometer cannot detect as small additions as the eye. This extreme is illustrated by mixtures of Pontacyl Carmines 2B and 6B Conc., which are not chemically identical but have very similar spectral absorption curves. For equal weights of standardized dyes, the long wave portions of the curves for the solutions can be practically superimposed, the short wave portions being parallel; somewhat different relations hold for the reflection curves of dyeings. Two very experienced dye testers find that 2.5 percent of the latter dye mixed with the former can be seen to change the color of the skeins, but the mixture would usually be passed as not “off-shade,” while a 5 percent admixture would cause positive rejection. Using the curves of the buffered solutions of the two dyes normally employed by us for spectrophotometric standardization, the hue change due to a 25 percent admixture of the latter dye probably can be detected by inspection of the curves; but the curve due to a 10 percent admixture (with 90

1. A. H. Taylor and C. H. Sharp, Trans. Ill. Eng. Soc. 15, 811 (1920).
2. H. J. McNicholas, Bur. Stand. J. Research I, 29 (1928).
3. The present paper will appear in full in J. Opt. Soc. Am.
4. C. Z. Draves, J.Opt. Soc. Am. 21, 336-46 (1931).
5. R. D. Nutting, J.Opt. Soc. Am. 24, 135-8 1929).
6. A. C. Hardy,J. Opt. Soc. Am. 18, 96-117 (1929).