Multi-element analysis of emeralds and associated rocks by k0 neutron activation analysis

Multi-element analysis of emeralds and associated rocks by k0 neutron activation analysis R.N. Acharyaa,*, R.K. Mondalb, P.P. Burtea, A.G.C. Naira, N.B.Y. Reddyc, L.K. Reddyd, A.V.R. Reddya, S.B. Manohara aRadiochemistry Division, Bhabha Atomic Research Centre, Trombay, Mumbai 400 085, India bAtomic Mineral Division, Begumpet, Hyderabad, India cDepartment of Applied Geology, S.V.U.P.G. Centre, Cuddapah AP, India dDepartment of Chemistry, S.V.U.P.G. Centre, Cuddapah AP, India Received 19 February 1999; received in revised form 21 April 1999; accepted 18 November 1999 Abstract Multi-element analysis was carried out in natural emeralds, their associated rocks and one sample of beryl obtained from Rajasthan, India. The concentrations of 21 elements were assayed by Instrumental Neutron Activation Analysis using the k0 method …k0 INAA method) and high-resolution gamma ray spectrometry. The data reveal the segregation of some elements from associated (trapped and host) rocks to the mineral beryl forming the gemstones. A reference rock standard of the US Geological Survey (USGS BCR-1) was also analysed as a control of the method. 7 2000 Elsevier Science Ltd. All rights reserved. Keywords: Emerald; Beryl; Associated rocks; k0-NAA; Gamma-ray spectrometry; Elemental concentration profile 1. Introduction Emeralds are a class of gemstones formed when specific additional metallic and earth elements, often at trace level, are present in beryl, beryllium aluminium silicate [Be3Al2(SiO3)6]. The colours of various types of beryls are due to the optical absorption spectrum in the range of 400–700 nm of visible region facilitated by the presence of chromophoric transition metal ions (Wood and Nassau, 1968). The colours of emeralds are dark green, pale green or blue and red depending on the presence of metals like Cr and V, Fe and Mn respectively (Encyclopedia of Chemical Technology, 1994). Emeralds containing more than 0.1% of Cr pos- sess dark green colours. Emeralds are the result of unusual geochemical processes, in which Be occurs in granite and granite pegmatites and Cr occurs in basic and ultrabasic rocks (Ma et al., 1993). Instrumental Neutron Activation Analysis (INAA) is being extensively used as an analytical tool in the environmental, the biological, the geological and the cosmological fields. It provides precise and accurate results with high sensitivity and selectivity for a large number of elements. Although useful for some purposes routine instrumental analyses by AAS, ICPMS/AES and electrochemical methods are cum- bersome and tedious, especially for multi-element analysis in complex matrices. INAA is more e€ec- tive for trace element analysis in the presence of other elements in varying matrices. The advent of high-resolution gamma ray spectrometry using Applied Radiation and Isotopes 53 (2000) 981–986 0969-8043/00/$ - see front matter 7 2000 Elsevier Science Ltd. All rights reserved. PII: S0969-8043(99 )00272-9 www.elsevier.com/locate/apradiso * Corresponding author. HPGe detectors has increased the potential of this technique. Compared to the conventional relative method, the k0 NAA method (Simonits et al., 1975; De Corte and Simonits, 1989; De Corte, 1992) is versatile for multi-element analysis. In the k0 NAA method, only one standard (gold) along with the unknown test sample is irradiated. The k0 NAA method has recently been applied by McOrist et al. (1994) for trace element analysis of gemstones like Australian opal and by Fardy and Farrar (1992) for Argyle diamonds. Yu (1995) characterised di€erent ruby samples by Energy Dis- persive X-ray Fluorescence (EDXRF) spectrometry. Particle Induced X-ray Emission (PIXE) technique has also been used by Tang et al. (1988, 1989) to characterise di€erent natural and synthetic ruby samples. Multi-element analysis of a natural ruby using the k0 NAA method was carried out by us (Acharya et al., 1997). In the present work, samples of green emerald and associated rocks (trapped and host rocks) together with a sample of beryl (pale green in colour) have been analysed. The samples were collected from the Tikki area of Rajasthan, India. The emeralds were crystal- line, short prismatic with distinct longer prismatic striations. Their size varied from 2 to 6 mm in length and 1 to 2 mm in width. They were transparent, with yellowish green colour and vitreous luster. The emerald formations were located in a host rock. There were two emerald crystals of the above specified sizes and these crystals were separated by a rock of 2–3 mm thickness. The rock in between the emeralds has been named trapped rock. The rock that surrounds the emeralds and the trapped rock has been named host rock. Both the rocks (trapped and host rocks) together are termed associated rocks. The thickness of the host rock was around 10 mm and it was carefully scraped and collected for analysis. Samples of emerald, trapped and host rocks were ground in an agate mortar to pre- pare samples of 100–200 mesh particle size for analy- sis. Multi-element analysis of host rock, trapped rock and emerald was carried out by the k0 NAA method using gold as single comparator. The beryl mineral was also analysed to compare its composition with that of the emerald. The beryl sample was collected from the same geological formation as the emerald. The beryl was opaque whereas the emerald was trans- parent. The emerald crystals were clean and free from fracture. Prismatic crystals of beryl are emeralds. Both the emerald and the beryl were collected from pegma- tite–granite rock in general and the emerald, in par- ticular, was from the ultramafic rock. Elemental profiles of certain elements such as Cr, Cs, Rb, Sc, Fe and V of host and trapped rocks and emerald are used to understand the process of segregation and are dis- cussed in this paper. 2. Experimental 2.1. Irradiations and radioactive assay The samples of emerald, beryl, host and trapped rock were powdered separately in an agate mortar and samples weighing about 15–50 mg each were packed in polypropylene tubes (2 mm ID). Four sub-samples from each group with an accurate concentration of gold in the range of 5–15 mg held in separate polypro- pylene tubes were sealed and irradiated in the E8 pos- ition of the swimming pool APSARA reactor, Trombay, BARC. The irradiation time was varied between 30 min and 7 h depending on the half life of the activation products. The neutron fluence rate in this position was of the order of 1012 n cmÿ2 sÿ1. The pneumatic irradiation facility at CIRUS reactor at Trombay, BARC was used for short irradiations where the duration of irradiations was 30–60 seconds and the neutron fluence rate was of the order of 01013 n cmÿ2 sÿ1. The ratios of subcadmium to epicadmium neutron flux (f) were 52.222.7 and 80.025.3 for the E8 pos- ition of APSARA reactor and the pneumatic ir- radiation position at CIRUS reactor, respectively (Acharya et al., 1997). Samples were assayed for gamma-ray activity of the …n, g† activation products using an 80 cc HPGe detector coupled to a PC-based 4K channel analyser in an eciency calibrated position with reproducible sample-to-detector geometry. To avoid true coincidence e€ects the sample-to-detector distance was maintained at 12–15 cm; depending upon the level of activity. The detector system had a gamma-ray energy resolution of 2.3 keV at 1332 keV. The activity of each radionuclide was followed as a function of time to ensure purity and identity. Gamma-ray standard 152Eu was used for eciency cali- bration of the detector, at di€erent distances between the sample and detector in a stable source-to-detector geometry. 2.2. The k0 NAA method and calculation The disintegration rate (dps) of the radionuclide, formed when an element is subjected to neutron acti- vation, is given by, dps ˆ � NAyw M � s � f � S �D …1† where, NA is Avogadro number, y is the isotopic abun- dance, w is the weight of the element, M is atomic weight, s is the capture cross section, f is the neutron fluence rate, S ˆ 1ÿ eÿlti is the saturation factor, D ˆ eÿltc is cooling correction factor, l is the decay con- stant, ti is the duration of irradiation and tc is the cooling time, R.N. Acharya et al. / Applied Radiation and Isotopes 53 (2000) 981–986982 The activation product when assayed by gamma-ray spectrometry, the peak area …PA† corresponding to the photo peak, was calculated by summing the counts under the peak and subtracting the linear Compton background or by the program ‘‘SAMPO’’. The ex- pression for PA is given by, PA ˆ � NAyw M � s � f � S �D � C � e � g …2† where, C ˆ …1ÿ eÿl�CL†=l is the correction factor for decay during counting, CL is the clock time, e is gamma ray detection eciency and g is gamma ray abundance. The PA area has been corrected for dead time by multiplying with CL/LT factor (Acharya et al., 1997), where, LT = live time. Specific count rate …Asp† for the activation product of the element of interest is given by, Asp ˆ PA S �D � C � w ˆ � NAy M � s � f � eg …3† where the quantity ‘w’ is in microgram. Similar, the specific count rate for the comparator (gold) can be written as, A�sp ˆ P �A S � �D� � C � � w� ˆ � NAy � M� � s� � f � e� � g� …4† The specific count rate ratio of the activation products of element of interest to the comparator is given by the term kanal. kanal ˆ Asp A�sp ˆ …NAy=M†s � f � e � gÿ NAy �=M� � s� � f � e� � g� ˆ …M � � y � g � s � f � e†ÿ M � y� � g� � s� � f � e�� …5† The term ‘s � f‘ can be bifurcated as …sthfth ‡ I0fe), where sth = thermal neutron cross section, fth = ther- mal/subcadmium neutron fluence rate, I0 = resonance integral = „1 ECd …s…E †=E † dE, where ECd ˆ 0:55 eV = cadmium cut-o€ energy and fe = epicadmium neu- tron fluence rate. The above Eq. (5) is modified as, kanal ˆ M� � y � g � ÿsthfth ‡ I0fe� � e� M � y� � g� � ÿs�thfth ‡ I �0fe� � e� …6† Taking fth=fe ˆ f and I0=sth ˆ Q0, Eq. (6) can be writ- ten as kanal ˆ …M � � y � g � sth † M � y� � g� � s�th …f‡Q0 † � e�ÿ f‡Q�0 � � e� ˆ k0, th " …f‡Q0 † � eÿ f‡Q�0 � � e� # …7† where, k0, th ˆ …M � � y � sth � g†ÿ M � y� � s�th � g� � …8† After correcting the Q0 value for the non ideal epither- mal neutron flux distribution by the parameter a (De Corte et al., 1975) and inputting k0, exp (De Corte and Simonits, 1989) instead of k0, th the final expression for kanal can be written as, kanal ˆ k0, exp "ÿ f‡Q0…a† �ÿ f‡Q�0…a† � # e e� …9† Since kanal ˆ Asp=A�sp or kanal ˆ PA=S �D � C � w � A�sp as per Eq. (3), the concentration of the element of interest (w in microgram) is given by, w …mg† ˆ PA S �D � C � A�sp � kanal …10† Finally, the concentration of the ith element …Ci in mg gÿ1 or mg kgÿ1) is calculated using the relation, Ci ÿ mg kgÿ1 � ˆ " Ap, i A�sp � kanal # …11† where, Ap, i ˆ PA=S �D � C � w, where w = weight of the sample in g, A�sp = the specific count rate of the comparator and the symbol � refers to the parameters of the comparator, gold. The k0, exp values are taken from the compilations of experimental k0 values by De Corte and Simonits (1989). As the factor k0 is used for the calculation of the concentration of the element the method is referred as k0 NAA method. The kanal is cal- culated by inputting the corresponding k0, exp, f, Q0…a† and e as per Eq. (9). This expression is the simplified form of kanal assuming negligible contribution of neu- tron self shielding (i.e., the self shielding correction fac- tor for thermal (Gth) and epi-thermal …Ge†11). The final concentration in mg kgÿ1 is determined by substi- tuting the corresponding peak areas corrected for sat- uration and decay using Eq. 11. Relevant nuclear data are taken from the compilations of Browne and Fire- stone (1986) and IAEA (1987). Further details of the calculations and input parameters like f and a are given elsewhere (Acharya, 1997). The precision and ac- curacy of the method is confirmed by analysing the US Geological Survey rock standard reference material, USGS BCR-1 (Gladney et al., 1983). R.N. Acharya et al. / Applied Radiation and Isotopes 53 (2000) 981–986 983 3. Results and discussion Results obtained on elemental concentrations of beryl (E1), emerald (E2), trapped rock (E3) and host rock (E4) are given in Table 1. The uncertainties on measured concentrations of elements in Table 1 are2 1s and they are the unweighted standard deviations of four independent measurements. Elemental concen- trations measured in USGS BCR-1 are the mean values from triplicate measurements with their stan- dard deviations (2 1 s) and are given in Table 2 and the values are in overall agreement with consensus values (Gladney et al., 1983). The measured concen- trations of V and Yb are compared with the reported average values by Flanagan (1973). The quoted errors represent precision in the measured elemental concen- trations based on the triplicate measurements and are in the range of 2–11%. The total number of elements that are measured in E1, E2, E3 and E4 are 15, 16, 11 and 15 respectively. The concentration profiles of the major elements: Na, Mg, Al, K, Ca and Fe are plotted in Fig. 1. The concentration profiles of some of the minor and trace elements: Sc, V, Cr, Mn, Co, Cl, Rb and Cs are plotted in Fig. 2. The seven key elements of significance of a gemstone are Mg, Al, Sc, V, Cr, Mn and Fe (Fardy and Farrar, 1992) and are present in the emeralds analysed; on the other hand V is not detected in beryl. Igneous rocks derived from magma contain Na and K minerals in varying concentrations. In the present samples of emerald and its associated rocks, K is absent and Na is present in varying con- centrations. This indicates that the rocks analysed con- Table 1 Elemental concentration of beryl, emerald and associated rocks (in mg kgÿ1 unless % is indicated)a S.N. Element Beryl E1 Emerald E2 Trapped Rock E3 Host Rock E4 1 Na% 0.5820.01 0.7620.04 5.3020.28 7.2720.39 2 Mg% 4.8520.21 1.1220.10 N.D. N.D. 3 Al% 10.0320.41 10.0220.34 8.0120.48 8.9320.47 4 K% 0.1220.01 N.D. N.D. N.D. 5 Ca% N.D. N.D. 0.7820.04 1.2720.08 6 Sc 3.1220.13 28.8820.37 5.4820.23 0.0420.002 7 V N.D. 39.0421.23 N.D. N.D. 8 Cr 452.58213.11 615.24234.39 71.2123.12 17.0421.10 9 Mn 281.90211.70 46.4522.48 82.4222.03 58.1522.01 10 Fe% 1.5420.12 0.2320.01 0.0220.001 0.5620.03 11 Co 17.5221.12 1.3820.05 N.D. 0.5320.03 12 Zn 374.49217.23 N.D. N.D. N.D. 13 Cl N.D. 268.25210.75 478.51225.01 1048245 14 Br N.D. N.D. N.D. 5.8720.25 15 Rb 33.3422.03 70.6423.45 N.D. N.D. 16 Cs 115.4623.83 135.5925.28 29.9721.25 1.1020.04 17 Ba N.D. 260.29215.12 N.D. N.D. 18 La N.D. 3.6220.08 N.D. 1.9020.05 19 Sm 0.2020.02 N.D. N.D. 0.2020.01 20 Eu 0.2920.02 0.8320.05 N.D. 2.4020.14 21 Th 0.3120.01 0.7520.04 3.4420.15 1.4820.08 a (1) N.D. — not detected. (2) Uncertainties:21s from four independent measurements. Table 2 Elemental concentration of USGS BCR-1 (in mg kgÿ1 unless % is indicated) S. No. Element Measured Consensus valuesa 1 Na% 2.6020.12 2.4320.08 2 Mg% 2.2620.18 2.0820.01 3 Al% 7.0020.20 7.2120.13 4 K% 1.4820.10 1.4020.07 5 Ca% 4.9320.15 4.9720.11 6 Sc 31.0821.51 32.821.7 7 Ti % 1.4620.06 1.332.06 8 V 412.0829.04 399b 9 Cr 17.3521.25 1624 10 Mn 1564227 1410240 11 Fe% 9.6520.17 9.3820.22 12 Co 35.8620.81 36.321.6 13 La 24.4021.37 2520.08 14 Ce 55.1221.21 53.720.8 15 Eu 2.1420.24 1.9620.05 16 Yb 3.3120.15 3.36b 17 Hf 5.1120.34 4.920.3 18 Th 6.1720.37 6.0420.6 a Gladney et al. (1983). b Flanagan (1973). R.N. Acharya et al. / Applied Radiation and Isotopes 53 (2000) 981–986984 tain only sodic feldspar minerals. Since the gemstone, emerald, is embedded in granite rocks, some of the sodium might have migrated during formation of the gemstones. The concentration of sodium is increasing in the order of emerald, trapped rock and host rock. In the case of beryl the chemical analysis reveals that in addition to sodium, potassium is present. The con- centrations of Al in both emerald and beryl are found to be about 10 % (Table 1) and is in agreement with the formula [Be3Al2(SiO3)6] for beryl. Magnesium is a macro component both in emerald and in beryl. Mag- nesium ion may coexist alongwith Be ion in the void space available in the silicate structure. Cordirerite, Al3Mg2(Si5Al)O18, which has a structure similar to beryl, contains Mg ions only instead of Be ions. Emer- ald contains higher concentrations of V, Cr, Sc, Cs and Rb (Fig. 2) compared to associated rocks. From Fig. 2, it is observed that the chromium content of emerald is the highest and decreases in the following order, emerald (620 mg kgÿ1) trapped4 rock (71.2 mg kgÿ1) 4 host rock (17 mg kgÿ1), indicating depletion of Cr from associated rock to emerald. The enrichment of some trace and other elements in emerald could be the result of migration of these elements from associ- ated rocks. This could be due to two reasons: (i) in the formations of gemstones, these elements might have segregated and (ii) these elements might have been weathered away from rocks over time after formation of the gemstones. The beryl structure consists of a series of SiO4 and BeO4 tetrahedra connected with AlO6 octahedra in the ratio 6:3:2 to give the composition Be3Al2( SiO3)6 (Wood and Nassau, 1968). Based on ionic radii, except perhaps for Li+, alkali ions would be expected to enter only the hexagonal channels (Wood and Nassau, 1968). Thus, the presence of Cs+ is expected with Na+, K+ and Rb+ as observed in the present studies. Impurity ions such as Fe3+ and Cr3+ are expected to substitute at the aluminium site on the basis of their ionic radii. Concentrations of Si and Be could not be measured by the present method as their measurements are not easily amenable to INAA. Cl and Br are absent in beryl whereas Br is present only in the host rock. Ba is present in emerald. In general, most of the elements in question concentrate in the emerald. Host rock con- tains the rare earth elements in higher concentration level and they are not detected in trapped rock. Thor- ium is present in all the samples at trace level. The mineral beryl, however does not contain the chromo- phoric element V, having a composition similar to that of emerald. 4. Conclusion INAA is used for measuring the concentration of el- ements in di€erent rock samples. Data on elemental concentrations suggest that the segregation of trace el- ements takes place in the formation of an emerald. Concentration profiles of elements are used to charac- terise the emeralds. Our results show the presence of Mg as a major and Co as a trace constituent in emer- ald and the absence of expected elements such as K, Ga and Ca reported in the PIXE/PIGE results of Ma Fig. 1. Concentration profiles of major elements of E1, E2, E3 and E4. Fig. 2. Concentration profiles of some of the minor and trace elements of E1, E2, E3 and E4. R.N. Acharya et al. / Applied Radiation and Isotopes 53 (2000) 981–986 985 et al. (1993). These results di€er from the findings of Ma et al. (1993). Acknowledgements We are grateful to the personnel of APSARA and CIRUS reactors for their co-operation in irradiating our samples. The authors express their sincere thanks to Dr. P.S. Rao, National Institute of Oceanography, Goa, India for his helpful suggestions during prep- aration of the manuscript. References Acharya, R.N., Burte, P.P., Nair, A.G.C., Reddy, A.V.R., Manohar, S.B., 1997. Multi-element analysis of natural ruby samples by neutron activation using the single comparator method. J. Radioanal. Nucl. Chem. 220 (2), 223. Browne, E., Firestone, R.B., 1986. In: Shirley, V.S. (Ed.), Table of Radioactive Isotopes. Wiley, New York. De Corte, F., 1992. Problems and solutions in the standardis- ation of Reactor Neutron Activation Analysis. J. Radioanal. Nucl. Chem. 160 (1), 63. De Corte, F., Simonits, A., 1989. k0 -Measurements and re- lated nuclear data compilation for (n, g† Reactor Neutron Activation Analysis. IIIb Tabulation. J. Radioanal. Nucl. Chem. 133 (1), 43. De Corte, F., Moens, L., Sordo-El Hammami, K., Simonits, A., Hoste, J., 1975. Modification and generalization of some methods to improve the accuracy of a determination in the 1=E 1‡a epithermal neutron spectrum. J. Radioanal. Chem. 52 (2), 305. Anon., 1994. Gemstones. In Kroschwitz, Jacqueline I., Howe- Grant, Mary. Encyclopedia of Chemical Technology, 4th ed. vol. 12. Wiley, New York. p. 417–479. Fardy, J., Farrar, Y.J., 1992. Trace element analysis of Argyle diamonds using Instrumental Neutron Activation Analysis. J. Radioanal. Nucl. Chem. Letters 164 (5), 337. Flanagan, F.J., 1973. 1972 values for international geochem- ical reference samples. Geochim. Cosmochim. Acta 37, 1189. Gladney, E.S., Burns, C.E., Roelandts, I., 1983. 1982 Compilation of elemental concentration in eleven USGS rock standards. Geostand. Newsl. 7 (1), 3. IAEA, 1987. Handbook of Nuclear Activation Data. Technical Report Series No.273. International Atomic Energy Agency, Vienna. Ma, X.P., MacArthur, J.D., Roeder, P.L., Mariano, A.N., 1993. Trace element fingerprinting of emeralds by PIXE/ PIGE. Nucl. Instrm. Meth. B75, 423. McOrist, G.D., Smallwood, A., Fardy, J.J., 1994. Trace el- ements in Australian opals using neutron activation analy- sis. J. Radioanal. Nucl. Chem. 185 (2), 293. Simonits, A., de Corte, F., Hoste, J., 1975. Single-comparator methods in Reactor Neutron Activation Analysis. J. Radioanal. Chem. 24, 31. Tang, S.M., Tang, S.H., Tay, T.S., Retty, A.T., 1988. Analysis of Burmese and Thai rubies by PIXE. Appl. Spectrosc. 42 (1), 44. Tang, S.M., Tang, S.H., Mok, K.F., Retty, A.T., Tay, T.S., 1989. Study of natural and synthetic rubies by PIXE. Appl. Spectrosc. 43 (2), 219. Wood, D.L., Nassau, K., 1968. The characterisation of beryl and emerald by visible and infrared absorption spec- troscopy. Amer. Mineral 53, 777. Yu, K.N., 1995. Identification of rubies by energy-dispersive X-ray fluorescence spectrometry. Appl. Spectrosc. 48 (5), 641. R.N. Acharya et al. / Applied Radiation and Isotopes 53 (2000) 981–986986