Multi Element

Applied Radiation and Isotopes 53 (2000) 981±986

www.elsevier.com/locate/apradiso

Multi-element analysis of emeralds and associated rocks by k0 neutron activation analysis R.N. Acharya a,*, R.K. Mondal b, P.P. Burte a, A.G.C. Nair a, N.B.Y. Reddy c, L.K. Reddy d, A.V.R. Reddy a, S.B. Manohar a a

Radiochemistry Division, Bhabha Atomic Research Centre, Trombay, Mumbai 400 085, India b Atomic Mineral Division, Begumpet, Hyderabad, India c Department of Applied Geology, S.V.U.P.G. Centre, Cuddapah AP, India d Department 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 pro®le

1. Introduction Emeralds are a class of gemstones formed when speci®c 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-

* Corresponding author.

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 ®elds. 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 cumbersome and tedious, especially for multi-element analysis in complex matrices. INAA is more e€ective for trace element analysis in the presence of other elements in varying matrices. The advent of high-resolution gamma ray spectrometry using

0969-8043/00/$ - see front matter 7 2000 Elsevier Science Ltd. All rights reserved. PII: S 0 9 6 9 - 8 0 4 3 ( 9 9 ) 0 0 2 7 2 - 9

982

R.N. Acharya et al. / Applied Radiation and Isotopes 53 (2000) 981±986

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 Dispersive 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 crystalline, 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 speci®ed 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 prepare samples of 100±200 mesh particle size for analysis. 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 transparent. The emerald crystals were clean and free from fracture. Prismatic crystals of beryl are emeralds. Both the emerald and the beryl were collected from pegmatite±granite rock in general and the emerald, in particular, was from the ultrama®c rock. Elemental pro®les 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 discussed 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 polypropylene tubes were sealed and irradiated in the E8 position 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 ¯uence 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 ¯uence rate was of the order of 01013 n cmÿ2 sÿ1. The ratios of subcadmium to epicadmium neutron ¯ux (f) were 52.2 2 2.7 and 80.0 2 5.3 for the E8 position of APSARA reactor and the pneumatic irradiation 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 152 Eu was used for eciency calibration 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 activation, is given by,   NA yw dps ˆ s
f
S
D …1† M where, NA is Avogadro number, y is the isotopic abundance, w is the weight of the element, M is atomic weight, s is the capture cross section, f is the neutron ¯uence rate, S ˆ 1 ÿ e ÿlti is the saturation factor, D ˆ e ÿltc is cooling correction factor, l is the decay constant, ti is the duration of irradiation and tc is the cooling time,

R.N. Acharya et al. / Applied Radiation and Isotopes 53 (2000) 981±986

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 expression for PA is given by,  PA ˆ



NA yw s
f
S
D
C
e
g M

…2†

ÿl
CL

where, C ˆ …1 ÿ e †=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. Speci®c count rate …Asp † for the activation product of the element of interest is given by, PA Asp ˆ ˆ S
D
C
w



 NA y s
f
eg M

…3†

where the quantity `w' is in microgram. Similar, the speci®c count rate for the comparator (gold) can be written as, Asp ˆ

P A ˆ   S
D
C 
w



 NA y  s
f
e
g M

…4†

The speci®c count rate ratio of the activation products of element of interest to the comparator is given by the term kanal. kanal

Asp …NA y=M†s
f
e
g
ˆ  ˆÿ Asp NA y =M s
f
e
g ˆÿ

…M
y
g
s
f
e†
M
y
g
s
f
e

…5†

The term `s
f` can be bifurcated as …sth fth ‡ I0 fe ), where sth = thermal neutron cross section, fth = thermal/subcadmium „ 1 neutron ¯uence rate, I0 = resonance integral = ECd …s…E †=E † dE, where ECd ˆ 0:55 eV = cadmium cut-o€ energy and fe = epicadmium neutron ¯uence rate. The above Eq. (5) is modi®ed as, kanal


ÿ M
y
g
sth fth ‡ I0 fe
e
ÿ
ˆ M
y
g
sth fth ‡ I 0 fe
e

…6†

Taking fth =fe ˆ f and I0 =sth ˆ Q0 , Eq. (6) can be written as kanal ˆ

…M
y
g
sth † …f ‡ Q0 †
e
ÿ
M
y
g
sth f ‡ Q0
e

"

…f ‡ Q0 †
e
ˆ k0, th ÿ f ‡ Q0
e

983

# …7†

where, …M
y
sth
g†
k0, th ˆ ÿ M
y
sth
g

…8†

After correcting the Q0 value for the non ideal epithermal neutron ¯ux distribution by the parameter a (De Corte et al., 1975) and inputting k0, exp (De Corte and Simonits, 1989) instead of k0, th the ®nal expression for kanal can be written as, kanal

"ÿ
# f ‡ Q0 …a† e
ˆ k0, exp ÿ f ‡ Q0 …a† e

…9†

Since kanal ˆ Asp =Asp or kanal ˆ PA=S
D
C
w
Asp as per Eq. (3), the concentration of the element of interest (w in microgram) is given by, w …mg † ˆ

PA S
D
C
Asp
kanal

…10†

Finally, the concentration of the ith element …Ci in mg gÿ1 or mg kgÿ1) is calculated using the relation, ÿ
Ci mg kg ÿ1 ˆ

"

Ap, i Asp
kanal

# …11†

where, Ap, i ˆ PA =S
D
C
w, where w = weight of the sample in g, Asp = the speci®c 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 calculated by inputting the corresponding k0, exp, f, Q0 …a† and e as per Eq. (9). This expression is the simpli®ed form of kanal assuming negligible contribution of neutron self shielding (i.e., the self shielding correction factor for thermal (Gth) and epi-thermal …Ge †11). The ®nal concentration in mg kgÿ1 is determined by substituting the corresponding peak areas corrected for saturation and decay using Eq. 11. Relevant nuclear data are taken from the compilations of Browne and Firestone (1986) and IAEA (1987). Further details of the calculations and input parameters like f and a are given elsewhere (Acharya, 1997). The precision and accuracy of the method is con®rmed by analysing the US Geological Survey rock standard reference material, USGS BCR-1 (Gladney et al., 1983).

984

R.N. Acharya et al. / Applied Radiation and Isotopes 53 (2000) 981±986

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 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Na% Mg% Al% K% Ca% Sc V Cr Mn Fe% Co Zn Cl Br Rb Cs Ba La Sm Eu Th

0.5820.01 4.8520.21 10.0320.41 0.1220.01 N.D. 3.1220.13 N.D. 452.58213.11 281.90211.70 1.5420.12 17.5221.12 374.49217.23 N.D. N.D. 33.3422.03 115.4623.83 N.D. N.D. 0.2020.02 0.2920.02 0.3120.01

0.7620.04 1.1220.10 10.0220.34 N.D. N.D. 28.8820.37 39.0421.23 615.24234.39 46.4522.48 0.2320.01 1.3820.05 N.D. 268.25210.75 N.D. 70.6423.45 135.5925.28 260.29215.12 3.6220.08 N.D. 0.8320.05 0.7520.04

5.3020.28 N.D. 8.0120.48 N.D. 0.7820.04 5.4820.23 N.D. 71.2123.12 82.4222.03 0.0220.001 N.D. N.D. 478.51225.01 N.D. N.D. 29.9721.25 N.D. N.D. N.D. N.D. 3.4420.15

7.2720.39 N.D. 8.9320.47 N.D. 1.2720.08 0.0420.002 N.D. 17.0421.10 58.1522.01 0.5620.03 0.5320.03 N.D. 1048245 5.8720.25 N.D. 1.1020.04 N.D. 1.9020.05 0.2020.01 2.4020.14 1.4820.08

a

(1) N.D. Ð not detected. (2) Uncertainties:21s from four independent measurements.

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 are 2 1s and they are the unweighted standard deviations of four independent measurements. Elemental concentrations measured in USGS BCR-1 are the mean values from triplicate measurements with their standard 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 concentrations of V and Yb are compared with the reported average values by Flanagan (1973). The quoted errors represent precision in the measured elemental concentrations 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 pro®les of the major elements: Na, Mg, Al, K, Ca and Fe are plotted in Fig. 1. The concentration pro®les 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 signi®cance 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 concentrations. This indicates that the rocks analysed conTable 2 Elemental concentration of USGS BCR-1 (in mg kgÿ1 unless % is indicated) S. No.

Element

Measured

Consensus valuesa

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Na% Mg% Al% K% Ca% Sc Ti % V Cr Mn Fe% Co La Ce Eu Yb Hf Th

2.6020.12 2.2620.18 7.0020.20 1.4820.10 4.9320.15 31.0821.51 1.4620.06 412.0829.04 17.3521.25 1564227 9.6520.17 35.8620.81 24.4021.37 55.1221.21 2.1420.24 3.3120.15 5.1120.34 6.1720.37

2.4320.08 2.0820.01 7.2120.13 1.4020.07 4.9720.11 32.821.7 1.332.06 399b 1624 1410240 9.3820.22 36.321.6 2520.08 53.720.8 1.9620.05 3.36b 4.920.3 6.0420.6

a b

Gladney et al. (1983). Flanagan (1973).

R.N. Acharya et al. / Applied Radiation and Isotopes 53 (2000) 981±986

Fig. 1. Concentration pro®les of major elements of E1, E2, E3 and E4.

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 concentrations of Al in both emerald and beryl are found to be about 10 % (Table 1) and is in agreement with

985

the formula [Be3Al2(SiO3)6] for beryl. Magnesium is a macro component both in emerald and in beryl. Magnesium 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. Emerald 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) trapped 4 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 associated 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 contains the rare earth elements in higher concentration level and they are not detected in trapped rock. Thorium is present in all the samples at trace level. The mineral beryl, however does not contain the chromophoric element V, having a composition similar to that of emerald. 4. Conclusion

Fig. 2. Concentration pro®les of some of the minor and trace elements of E1, E2, E3 and E4.

INAA is used for measuring the concentration of elements in di€erent rock samples. Data on elemental concentrations suggest that the segregation of trace elements takes place in the formation of an emerald. Concentration pro®les of elements are used to characterise the emeralds. Our results show the presence of Mg as a major and Co as a trace constituent in emerald and the absence of expected elements such as K, Ga and Ca reported in the PIXE/PIGE results of Ma

986

R.N. Acharya et al. / Applied Radiation and Isotopes 53 (2000) 981±986

et al. (1993). These results di€er from the ®ndings 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 preparation 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 standardisation of Reactor Neutron Activation Analysis. J. Radioanal. Nucl. Chem. 160 (1), 63. De Corte, F., Simonits, A., 1989. k0 -Measurements and related 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. Modi®cation 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., HoweGrant, 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 geochemical 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 ®ngerprinting of emeralds by PIXE/ PIGE. Nucl. Instrm. Meth. B75, 423. McOrist, G.D., Smallwood, A., Fardy, J.J., 1994. Trace elements in Australian opals using neutron activation analysis. 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 spectroscopy. Amer. Mineral 53, 777. Yu, K.N., 1995. Identi®cation of rubies by energy-dispersive X-ray ¯uorescence spectrometry. Appl. Spectrosc. 48 (5), 641.