NEW INSTRUMENTS AND METHODS OF MEASUREMENT

LIGHT FILTERS FOR THE NEAR-INFRARED REGION OF THE SPECTRUM

A. A. Il’ina

The region of the infrared spectrum adjacent to the visible region and extending from 0.75 to 2 μ has recently attracted the attention of many investigators. The appearance of photographic plates sensitized up to 1.2 μ, and of photoelements (oxygen–silver–cesium ones among photoelements of the emission type, thallium sulfide and silver sulfide ones among photoelements of the photovoltaic type) sensitive in the same region, has considerably simplified experimental technique and has given rise to a number of interesting applications. Such are applications connected with infrared photography, spectroscopy, spectrophotometry, etc.*) These applications have created a need for suitable light filters: in some cases it is necessary to filter out the visible part of the spectrum and transmit only the infrared rays; in others, on the contrary, to block completely the whole infrared part, including the region transparent to glass optics, i.e. precisely 0.75—2 μ; finally, in purely practical applications it is important to be able to transmit narrow portions of the spectrum without using monochromators, i.e. to have at one’s disposal so-called monochromatic light filters.

Unfortunately, in the near-infrared region the formulation of light filters has not been developed in such detail as in other spectral intervals, and therefore every new indication of the possibility of using one or another substance as a light filter is met with great interest.

As in most questions of modern physical optics, the pioneer here was R. W. Wood⁵, who as early as 1909 made his well-known photographs of landscapes in infrared rays and for this purpose used a filter consisting of a cuvette of cobalt glass,

*) On infrared photography there is Clark’s book¹. On biological applications see Ettger’s article¹ᵃ in Uspekhi fizicheskikh nauk. On calorimetric and spectrophotometric applications see the work of Stierns². Some information on this question may be found in the recently published collection Optics in Military Affairs³. See also the new edition of Strong’s book on the technique of the physical experiment⁴.

filled with a solution of potassium dichromate. Subsequently, he also constructed a solid filter consisting of two cobalt glasses with a gelatin film glued between them, dyed with methyl orange (transmits from \(0.68\mu\), maximum about \(0.8\mu\)).

Later, in connection with the needs of aerial photography, on the one hand, and motion-picture photography*, on the other, black filters transparent to infrared rays were issued by several firms. Such are, in particular, the filters of Corning and Agfa (their optical characteristics can be found in the catalogs). Since these filters are not widely dis—

Fig. 1. Transmission spectra of some organic dyes.
\(1\) — direct dark green, \(2\) — diamine fast violet BBN, \(3\) — methylene blue BB, \(4\) — benzofast black 4CL, \(5\) — cube azo green, \(6\) — naphthol green B, \(7\) — gallaz-green (in the visible part the curve coincides with 5), \(8\) — indanthrene black BB. Curves \(1, 2, 3\) are for dyed acetate films, \(6, 7\) for gelatin films, \(4, 5, 8\) for cellophane films.

* The use of film sensitive to the near-infrared region in combination with the corresponding filter makes it possible to replace very expensive shots of night scenes with shots taken in full sunlight.

widespread in our physics laboratories, we shall not dwell on them and shall consider, on the basis of the literature data, those filters that can easily be made by any researcher in a laboratory setting.

Let us first of all discuss filters that limit transmission on the side of short wavelengths. As in the ultraviolet, this problem is solved more simply than the reverse one: a large number of substances are known that sharply increase their transmission on going to longer wavelengths. For example, very many organic dyes (having green, blue, violet, or black coloration) give an extremely sharp increase in transparency in the far red

Fig. 2. Transmission spectra of black glasses;
1 — KS-12 (2 mm); 2 — the same, 5 mm; 3 — KS-13 (3 mm), 4 — the same, 8 mm.

and near infrared regions of the spectrum. As early as 1913 Pfund^8 pointed out the good transmission in the near infrared of certain dyes, such as, for example: cyanin (rise of the transmission curve from 0.66 μ), methyl-grün (0.68 μ), Janus-grün (from 0.7 μ), naphthol-grün (from 0.76 μ), etc. As an illustration, in Fig. 1 we present a series of transmission spectra of organic dyes*). Generally speaking, examination of the available data on the transmission spectra of organic dyes shows that only in rare cases does their transmission boundary lie beyond 0.8 μ, and it is very difficult to find a dye with a transmission boundary at \(\lambda > 0.85\,\mu\) and with a sufficiently steep rise of the curve.

In Fig. 2 are given the transmission spectra of the black glasses KS-12 and KS-13. From comparison of these curves with the transmission curves of organi—

*) With the kind permission of the authors, we borrow these curves from the work, prepared for press, of E. Shpol’skii, S. Khuludov, and S. Baranov^9.

...of organic dyes it is seen that black glasses have a much lower transmittance at the maximum and a more extended rise of the curve. It is much more difficult to find light filters that transmit the visible part of the spectrum and absorb the entire infrared part, or that have a steep drop of the curve toward the long-wave side. Something similar is given by various salts of copper and nickel. In Fig. 3 are given the absorption spectra of glasses colored

Fig. 3. Absorption of colored glasses:
1—glass with ferrous oxide, 2—glass colored with copper (green), 3—glass with chromium salts, 4—cobalt glass, 5—nickel glass (violet).

with salts of various metals, and in Fig. 4—the absorption spectra of aqueous solutions of the same salts (according to Dreisch’s work[^6]). Heat-protective glasses (greenish glasses containing ferrous oxide) cut off long-wave radiation, but have a gently sloping curve in the near infrared. It is very difficult, however, to select a filter transmitting, for example, \(0.7\,\mu\) and not transmitting wavelengths greater than \(0.8\,\mu\). Wavelengths greater than \(1.2\,\mu\) are usually removed by means of \(5\)–\(6\) cm of water.

Uranium and didymium glasses possess narrow absorption bands in the near infrared. The latter are often used in modern spectrophotometry for calibrating and checking spectropho-

...photometric setups. In a short paper by A. Pfund,^7 published in 1939, there is a description of several new light filters for the near infrared region, whose spectral characteristics are quite interesting. In this work, above all, several substances are indicated that absorb red and infrared rays up to \(0.85\,\mu\). Fig. 5 gives the transmission spectrum of vanadyl sulfate \((\mathrm{VO})(\mathrm{SO}_4)\). In a saturated solution this substance has a blue-black color. Such a solution was measured in a layer of \(1\ \mathrm{cm}\) relative to water (curve \(A\)). The same figure gives the transmission curve of a light filter composed of two glass plates and a 1.5-millimeter layer of a saturated solution of vanadyl sulfate. In both cases, the weak absorption of the red and near-infrared region of the spectrum up to \(0.9\), the steep rise of the curve on passing to longer waves, and the sufficient transmission at the maximum make this substance very valuable.* Pfund indicates that, through a light filter composed of vanadyl sulfate and a red signal Corning filter, the solar disk was not completely obscured.

In the same article—

Fig. 4. Absorption of solutions of inorganic salts:

\(1\)—\(\mathrm{FeSO}_4\), \(5\ \mathrm{mm}\) \((1\ \mathrm{M})\);
\(2\)—\(\mathrm{CuSO}_4\), \(0.72\ \mathrm{mm}\) \((0.5\ \mathrm{M})\);
\(3\)—\(\mathrm{Cr}_2(\mathrm{SO}_4)_3\), \(10\ \mathrm{mm}\);
\(4\)—\(\mathrm{CoCl}_2\) in \(\mathrm{HCl}\), \(0.9\ \mathrm{mm}\) \((0.5\ \mathrm{M})\);
\(5\)—\(\mathrm{NiCl}_2\), \(2\ \mathrm{mm}\) \((1\ \mathrm{M})\).

Fig. 5.

\(A\)—transmission of a solution of \((\mathrm{VO})(\mathrm{SO}_4)\) (relative to water),
\(B\)—the same relative to air.

* The transmission of vanadyl sulfate in the green and blue parts of the spectrum can easily be removed with the aid of almost any dark-red filter.

transmission spectra are given of several other substances, interesting because the limits of their transparency are shifted still farther toward long wavelengths. These spectra are shown in Fig. 6. Curve \(A\) represents the transmission of celluloid (\(0.58\,\mu\)), colored milori blue; \(B\) is the transmission of copper butyl phthalate (of unknown concentration), dissolved in \(\mathrm{CCl}_4\). Curve \(C\) refers to green bottle glass (this glass contains a large percentage of copper) in a layer of \(1.2\,\mathrm{mm}\). Curve \(D\) is a filter made of basic lead carbonate. This substance was ground in water containing a small quantity of gum arabic, and then applied as a thin layer to glass. After drying and separation from the glass (for details see Pfund’s article), the resulting film was mounted on a cardboard frame and measured. Curves \(C\) and \(D\) are of interest only because they begin their rise at wavelengths greater than \(1\,\mu\); however, the increase in transparency proceeds so slowly that these filters, of course, leave much to be desired.

Fig. 6.
\(A\)—milori-blue dye, \(B\)—copper butyl phthalate,
\(C\)—green bottle glass, \(D\)—white-lead film.

The same author had earlier\(^{10}\) published data on light filters made from powders of \(\mathrm{ZnO}\) and \(\mathrm{MgO}\) (cutting off rays \(\lambda < 1\,\mu\)) and from metallic films\(^{11}\). Especially interesting is a selenium film, which gives a rather sharp rise of transmission from \(1\,\mu\).

In the same work by Pfund there are several interesting remarks also on filters of the inverse type. First of all Pfund draws attention to the fact that light filters made from aqueous solutions of copper salts, which noticeably absorb red rays beginning already at \(0.6\,\mu\), can nevertheless transmit something in the near infrared. For example, the transmission curve of \(\mathrm{CuSO}_4\) (Fig. 7), after passing through a minimum at \(0.9\,\mu\), already rises appreciably by \(1.05\,\mu\), where water has a transmission band. If one takes into account the great intensity of the radiation of an incandescent lamp

Fig. 7.
\(A\)—transmission of a \(5\,\mathrm{cm}\) layer of water,
\(B\)—\(\mathrm{CuSO}_4\) solution with respect to water,
\(C\)—\(\mathrm{Fe}(\mathrm{NH}_4)_2(\mathrm{SO}_4)_2\) in solution with respect to water.

about 1 μ, it follows from this that, contrary to the opinion very widespread among experimenters, a solution of CuSO₄ does not guarantee absorption of the near infrared part of the spectrum. Pfund further gives the transmission spectrum of the compound Fe(NH₄)₂(SO₄)₂ (Mohr’s salt), whose absorption band, as can be seen from Fig. 8, is shifted

Fig. 8.

Transmission: A—iodine in CCl₄. B—solution of Fe(NH₄)₂(SO₄)₂ (18 mm) in water.

toward longer wavelengths in comparison with CuSO₄. The best heat filter recommended by Pfund consists of a saturated solution of Fe(NH₄)₂(SO₄)₂*), to which a little CuSO₄ has been added. Almost complete absorption at 1.05 μ and high transmission in the visible part make this filter unique. The transmission of an 18-millimeter layer of this solution can be seen in Fig. 7, where the transmission curve of a solution of iodine in CCl₄ (5 mm layer) is also given. For the test, both solutions were placed successively before an element illuminated by a 100 W “Mazda” lamp. In the case when the iodine solution in CCl₄ was placed before the thermoelement, the deflection of the galvanometer reached 215 mm; application of the second light filter (having 89% transmission in the infrared region) reduced the deflection of the galvanometer to 1.1 mm. Such a pair of additional or crossed light filters can be very useful, for example, in the study of infrared fluorescence.

*) Chemically pure salt was dissolved in distilled water that had first been boiled and then cooled; this solution was kept in a large glass tube with a tap at the lower end. A layer of mineral oil was poured on top, protecting this solution from oxidation.

References

1. Clark, Photography by Infrared.
   1a. Eggert, W. F. H., 1935.
2. Stearns, J. O. S. A., 33, 97, 1943.
3. Optics in Military Affairs, vol. I, 1946.
4. J. Strong, Procedures in Experimental Physics. New York, 1945.
5. R. Wood, Physical Optics, p. 455 ff.
6. Th. Dreisch, Z. Physik., 40, 714, 1927.
7. A. H. Pfund, J. O. S. A., 29, 56, 1939.
8. A. H. Pfund, Z. Wiss. Phot., 12, 341, 1913.
9. E. Shpolsky, S. Khludov, and S. Baranov. Transmission spectra of transparent colored films (in press).
10. A. H. Pfund, Phys. Rev., 36, 71, 1930.
11. A. H. Pfund, J. O. S. A., 23, 374, 1933.