THE OF , AND .

BY C. S. VENKA~'ESWARAN. (F~/om the Department of Physics, Indian Institute of Science, BangaIore.)

Received October 12, 1935. (Communicated by Sit" C. V. Raman, Kt., F.R.S., ~r

1. Introduction. THE remarkably brilliant crimson luminescence of ruby was at first observed by Edmond Becquerel 1 by irradiation with sunlight and was later on, the subject of detailed investigations at the hands of several investigators notably Lecoq de Boisbaudran, 2 Du Bois and Elias, 3 Schmidt, 4 and Niendenhall and Wood. 5 As has been observed by these authors, ruby furnishes an interest- ing example of a yielding sharp lines of luminescence and its spectrum consists of two intense lines at 6926 and 6942* accompanied by other lines or bands on either sides. The cathodo-luminescence of ruby was investigated by Sir William Crookes 6 and his results agree well with those obtained by visible ]ight. The origin of this luminescence was the subject of prolonged controversy beween Becquerel and Crookes on the one baud and Lecoq on the other. The former authors maintained that the agent causing lumines- cence was alumina while the latter insisted that the small trace of present was responsible "for the same. The detailed investigations of Wied- mann and Schmidt 7 and later by Sehmidt s succeeded in showing that care- fully frac'tionated pure alumina is non-luminescent and a small trace of chromium (1:10,000) is sufficient to make the lines appear intensely. Recently, Tanaka 9 showed that almost all the lines of the ruby coincide with the chromium series discovered byhim and from that he was led to conclude that the small trace of chromic oxide present in ruby is the cause of its luminescence.

I Becquerel, E., Am~ de Chim. et Phys., 1859, 57, 125; 1861, 62, 90. 2 Lecoq de Boisbaudran, Compt. Rend., 1887, 103, 1107, 1224; 1824, 104, 330, 478, 554. 3 Du Bols, H., and Elias, J.. Ann. d. Phys., 1908, 27, 233; 1911, 35, 635. 4 Schmidt, G. C., Ann. d. Phys. U. Chem., 1904, 15, 622 5 Mendenhall, E. C, and Wood, R. W., Phil. Mag., 1915, 30, 316. o * The numbers in this and the subsequent pages denote the in Angstrom units. 6 Crookes, W., Chem. News, 1887, 55. 25 and 56, 59. 7 Wledmann and Schmidt, ,4nn. d. Phys. U. Chem, 1895, 56 201. s Schmidt, G. C., Loc. cir. o Tanaka, four. Opt. Soc. Amer., 1924, 8, 287. 459 460 C.S. Venkateswaran

Alumina in solid solutions with the oxides of other metals like , manganese, , samarium, etc., is also known to exhibit fluorescence, 1~ the colour of which varies with the active metal used. The luminescence of alumina in the form of aluminates or silicates ill the natural state, namely, topas, and smaragd, etc., has also been investigated qualitatively but their complete spectral character and its bearing on the origin of lumines- cence has not been fully studied. In continuation of some unpublished work by Mr. Bhagavantam, the author has investigated the fluorescence spectra of the emerald, sapphire and ruby with a Fuess glass spectrograph of large aperture and having a of 100.~ per millimetre in the 6900 region. A pyrex mercury arc as well as a carbon arc has been used as sources of illumination. A deep blue solution of cupro-ammonium sulphate surrounding the solid cut off the radiations of tile source beyond 5500, where tile fluorescence of these appear. Ilford hypersensitive panchromatic plates, H and D 2500, are used for photographing the spectrum. The fluorescence of sapphire shows close similarity to the well-known spectrum of ruby ; bnt the emerald, which however belongs to the family, possesses striking dissimilarities in its spectrum though some general features characteristic of the ruby are also noticeable in this case. 2. Results. Emerald.--The most interesting of the substances investigated is tile emerald. It was available in the form of round beads of about 1.5 cm. to 2.5 cm. in diameter and in different tones of colour ranging from greyish to pale green. These beads possess a fairly constant density of 2-66. A transparent green in the form of a cut jewel with polished faces, which gives ,almost the same spectrum as the beads, has a of 1.5751 for the D-lines of sodium. A chemical analysis t of one of the beads gives the percentage composition of the as 3 BeO. AI~Oa. 6 SiO~ with 0.3 % of and an easily detectable quantity (about 0.25 ~ of chromium as well as small traces of rare earth elements. The fluorescence is very weak in these beads, an exposure of about twenty-four hours being required to obtain a fairly intense spectrum under the most favourable circumstances while the ruby has yielded the main lines in a minute. Its spectrum con- sists of two intense and sharp lines at 6806 and 6837, the latter being stronger and broader than the former and a system of diffuse bands on either sides. In order to understand which part of the spectrum of the source is responsible

lo Handbuch der Exptl. Phys., 1928, 23, Part 1, p. 425. t The chemical analysis was carried out by Mr. N. Jayaraman of the Department of General Chemistry of this Institute to whom the author's best thanks are due, The Fluorescence of Ru@, Sapphire and Emerald 46i

for this fluorescence, the intense lines of the mercury arc have been isolated by suitable filters and used separately for excitation and tile fluorescence appears equally well and unchanged in character for the 4046, 4358 and 5461 radiations of the arc. The wave-lengths of the lines are measured by comparison with the iron arc spectrum and the mercury arc lines and are classified in Table I together with their relative intensities. The wave-lengths given for the two intense lines at 6806 and 6837 are correct to -I- 1 ~x. An enlarged picture of the spectrum is reproduced in Fig. 1 of the accompany- ing Plate. In clew of the characteristic dissimilarity of the spectrum of emerald from that of ruby, the author bus also examined fairly transparent pieces of a bluish green and a deep green (smaragd) beryl. According to chemical analysis, also carried out by Mr. Jayaraman, the bluish green crystal has a structure 3 BeO. AlzOa. 6 SiO~ but contains about 3 % iron and only a very much smaller percentage of chromium than the emerald, barely detect- able during the analysis. It has yielded no fluorescence even on prolonged exposure of more than forty-eight hours though the Raman line n 647 cm.-1 appears in four hours. The deep green crystal possesses an intense fluores- cence between 4950 and 5450 in agreement with the observations of some earlier workers for beryl? 2 The significance of these results are discussed in a later section. Sapphir~.--This was available in the form of light blue beads having a density of 3-95, thus evidently belonging to the family. Blue crystals possessing deeper shades of colour are also examined in the form of cut stones and all of them have given an identical spectrum. The fluorescence, however, being extraordinarily feeble, very long exposures of about forty- eight hours are required to record the spectrum and this, perhaps, accounts for the absence of any investigation on the luminescence of this mineral. The spectrum consists of two intense and sharp lines at 6927 and 6942 and a series of bands, resembling the spectrum of ruby and is illustrated in Figs. 4 and 6. The results are given in Table I along with those of emerald. Ruby.--In order to furnish a comparison for the spectra of the two former substances, the fluorescence of a few beads of natural ruby of density four have also been examined. An exposure of about four hours has yielded a very intense spectrum showing a number of interesting features which have not been observed previously. It be mentioned that in the previous investigations of ruby instruments of large dispersion of about 3 ~_ per milli- metre but only poor light-gathering power have been used by Mendenhall

n Nisi, H., proc. Phys. Math. Soc. Jap., B., 1932, 14, 214. 12 Gmetiffs Handb~tch der /lnorg. Chem., 1930, 26, Be. 21. 462 C.S. Venkateswaran

TABLE I. (a) Emerald at 35 ~ C. (b) Sapphire at 35 ~ C.

Wavelength I I Intensity i in A.U. I ntensi ty in A.U. / I 7130 weak broad band 6946 very weak and diffuse band 7060 very weak band 6992 weak band 690S weak, diffuse band 6942 very strong line 6837 very strong line 6927 strong sharp line 6806 strong sharp line 6802 broad diffuse ban, I 6736 strong band weak band 6633 medium band 6753 6578 weak band 6690 medium band 6592 strongnarrow ban, and Wood and Du Bois and Elias. While the latter authors recorded a large number of lines on either sides of the intense doublet at 18 ~ C., Wood and co-worker ~a were unable to reproduce their results at 23 ~ C. With an exposure of one minute at a temperature of about 35 ~ C. the author has been able to record the two lines reported by Wood and his co-worker with good intensity. At longer exposures many more lines and bands appear, some of which are observed for the first time. That Wood and 3,Iendenhall did not obtain any but the most intense lines, is perhaps due only to the low intensity of the spectrograph and short time of exposure. The results obtained by the author are given in Table II together with those of the above authors. Besides the bands recorded by Du Bois and Flias, four new bands with eentres approxi- mately at 7089, 7164, 7222 and 7266 in the red end and three bands at 6650, (~495 and 6430 and a fairly sharp line at 6814 superposed by a weak band have been observed. These bands are also followed by a continuous spectrum between 6100 and 6900 which is especially noticeable in the over-exposed picture (Fig. 3). Because of the low dispersion of the spectrograph used, the sharp lines between 6946 and 7016 reported by the earlier workers are not resolved in the author's spectrum and have appeared only as a blacken- ing in the Plate. Fig. 5 in the accompanying Plate shows that the lines 6927 and 6945 are of unequal intensity as they are in sapphire. Fig. 2 gives an intense picture showing all the lines and bands clearly, particularly the

13 Mendenhall and Wood, Lee. cit. The Fluoresceuce of Ruby, Sapphire and Emerald 463

TABI,F, II. Ruby.

MendenhaU William Author--35~ Du Bols and Elias--18~ and Wood-- 23 ~C. Crookes

Wavelength Wavelength Intensity Wavelength Wavelength in A.U. Intensity in A.U. in A.U. in A.U.

7266 very weak band

7222 very weak band ~176176

7164 weak band ~176176176

7125 medium narrow band 7130 diffuse darkening

7089 medium narro~ band t~176 dark space

7059 medium narrow band 7060 diffuse darkening

7027 medium sharp line 7046 narrow band

6992 weak sharp line 7016 narrow band

Unresolved blackening group oflines (diffuse)

6945 very strong sharp line 6941 very strong band 6946 6942

6927 strong sharp line 6926 strong band 6932 6937

6814 weak sharp line (very weak continuous spectrum superposed)

6791 weak sharp llne 6790 narrow band

6753 medium diffuse band 6760 band

6690 strong broad band 6690 diffuse band 6707

6650 weak diffuse band ....

6592 strong narrow band 6590 weak band 6598

6495 weak band **,.

6430 very weak band ~176176 6514 {approximate

Continuous spectrum ,.o. centre of between 6100 and 6900 continuous spectrum) line at 6814 which almost coincides with one of the emerald lines. Fig. 3 which is an over-exposed spectrum shows the new bands observed in the red end as well as the continuous spectrum. 464 C.S. Venkateswaran

3. Discussion of Results. Just as the ruby and the sapphire differ very little in their physical properties such as density, refraction and , a dose similarity exists also in their fluorescence spectra as can be seen from Figs. 2 and 4. According to Stillwell, ~4 the blue colour of the natural sapphire is due to chromous oxide while the red col~ur of ruby is caused by a small percentage of chromic oxide. On the other hand there are some 15 who presume that the blue colour is entirely dne to ferric or oxide and that chromium is completely absent in it. Artificial sapphire resembling the natural stone has also been successfully prepared by fusing a mixture having a composition ~* of 98% A1208, 1,5~ FesO, and 0.5% TiO,. The only differences observed by the author in the fluorescence spectra of ruby and sapphire are : (1) the fluorescence in sapphire is extremely feeble as compared with that of ruby, (2) the doublets in ruby at 6791 and 6814 as well as, at 7059 and 7089 are replaced by bands in sapphire. While it is not justifiable to draw any con- clusion from these regarding the state and nature of impurities causing colouration in them, it is to be inferred that in both the materials the same agent is responsible for fluorescence and as is pointed out in the beginning, chromium is present in some form or other in both the stones. In the case of the emerald in which alumina is largely diluted by the presence of the oxides of beryllium and silicon, the fluorescence spectrum is very different from that of ruby as can be seen from Figs. 1 and 2. There are also some striking similarities in their spectrum. The prominent lines of the emerald spectrum at 6806 and 6835 are shar~p and it is remarkable that these lines almost coincide with the weak doublet at 6792 and 6814 in ruby. The occurrence of diffuse bands on either sides of these lines is also a feature in common with the spectrum of the latter. These broad bands bear a resemblance to the diffuse bands of ruby at temperatures higher than 200 ~ C. and may sharpen at lower temperatures. It may be concluded from these that the luminescence in emerald is also, probably, caused by chromium. The fact that beryl containing very little or no chromium does not give this characteristic line spectrum, lends support to this view. But it is difficult to understand why the strong doublet at 6926 and 6945 in the ruby should be suppressed almost completely and the doublet at 6792 and 6824 should appear a little shifted as the most prominent lines in emerald. The line at 6792 appears also intensely in the cathodo-luminescence of spineW (

14 Stillwell, C. W., J. Phys. Chem., 1926, 30, 1441. ~5 Krans and Holden, Gem Materials, 1925, 100. 16 Verneuil, A., Compt. Rend., 1910, 150, 186. 1~ Crookes, W., Loc. cir. C S. /enkateswaran. Proc. In& Acag. Sci., A, va/. IZ, P/. XXIV.

FIG. 1, Emer

Ruby FIG. 2. (1 hou

Rub? FIG. 3. (4 horn

Sapp FIG. 4. ( pale

S~pp (deep

FIo. 5. Ruby (1 min.) Fla. 6. The F/goresceme of Ruby, Sapp/dre end EmeraM 465 aluminate). A proper explanation of this cannot be given at this stage; but it is possible that the presence of the other oxides in the crystal is re- sponsible for this behaviour. A study of the luminescence of alumina fused with chromic oxide and a large percentage of other oxides may throw light on this problem. It would also be of interest to examine the fluorescence of emerald at various temperatures, particularly at the liquid air temperature and compare the behaviour with that of ruby under identical circumstances. The author hopes to undertake work on these lines soon and the results will be communicated in due course. 111 conclusion, the author desires to express his grateful thanks to Prof. Sir C. V. Raman, for his keen interest and constant encouragement in the course of the work as well as for kindly placing his valuable collection of gems at the disposal of the author. Surnma~Lv. The photo-luminescence of natural crystals of ruby, sapphire and emerald (beryl) has been studied at room temperature with a Fuess spectrograph of high light-gathering power. The ruby has yielded a number of new bands which have not been obtained hitherto. The spectrum of sapp~Jire corres- ponds closely to that of ruby but with slight differences. The spect,rum of emerald consists of t~o sharp !ines at 6806 and 6835 accompanied by other diffuse bands and is compared with that of ruby. The fluorescence in sapphire and emerald is also discussed with respect to the origin of luminescence in them. Photographs of the spectra illustrate the paper.