Journal of Radioanalytical and Nuclear Chemistry, Vol. 216, No. 2 (1997) 179-181

Multielemental NAA of geological micro-grain samples 'for origin identification Bangfa Ni,* Zhanqin Yu,* Pingsheng Wang,* Gaokui He,* Weizhi T'mn,* Erkang Wang** *China Institute of Atomic Energy, P.O. Box 275-50, Beijing 102413, P. R. China **Department of Geoscience, Nanjing University, Nanjing 210008, P. R. China (Received June 20, 1996)

Thirty two elements were determined in five pieces of presumed cosmic dust samples with the weights of 5 to 25 lag by neutron activation analysis (NAA). All the interferences from fission, threshold reactions and T-spectra were corrected for and different counting geometries normalized. Enrichment factors with reference, to Cl-choudrite are about)200 and 100 for lithophile refractory elements (e.g., V, Th, Hf and W) and ram earth elements fREE), respectively. Deficiencies were observed for Co, Cr, Mg and Na, etc. The Cl-chondrite normalized REE patterns are close to those of extraterrestrial materials, with no anomaly of Eu, indicating an extraterrestrial origin of the grain samples analyzed.

Introduction Cosmic dust carries abundant information on cosmochemistry and cosmophysics and is another important material besides meteorites in exploring the origin and evolution of the solar system and the universe. The identification of cosmic dust has long been studied by geologists. Although the unique mechanic features and the existence of meso-stable minerals in cosmic dust can be used for identification. The multielement composition is a far more accurate judgement of the origin of a presumed cosmic dust9 NAA is one of the most statable techmques in tackling this microanalysis problem for its inherent advantages of high sensitivities and accuracies for multielements and non-destructive nature. The identification of cosmic dust has been investigated using NAA by several authors in recent years, 1-3 however, the number of elements determined is not large enough, especially lack of data for REE. In the present work, 32 elements including 10 REE were determined in five pieces of presumed cosmic dust samples. The comprehensive results provide a solid foundation in origin identification. 9

o

~

9

9

o

Reactor irradiations and measurements

Grain samples were irradiated together with standards, comparator (Au) and Quality Control (QC) sample NBS SRM 1633a for 5 minutes at a horizontal channel of the 15 MW HWRR in our Institute. The irradiated grain samples were tranferred into a precleaned AI foil bag (for long irradiation after counting), and counted with a HPGe T-ray spectrometer (1.9 keV, 26%). After two weeks of decay, the grain samples were packed together with chemical standards, comparator, QC sample (NBS SRM 1633a) for long irradiation. The irradiation and measurement schedule is given in Table 1.

Results and discussion A commercial software SPECTRAN-F was used for peak analysis. A home-made hybrid K0-relative program ADVNAA 7 was used for qualitative and quantitative analysis. Analytical results of the five grain samples for 32 elements are listed in Table 2. In Table 2, the figures in parentheses are relative uncertainties in percentage, which

Experimental Sample preparation

Later type granite samples were taken from Yanshan Mountain, Western Suzhou city. The rock samples were first dry-cleaned with compressed air, and then crashed with sample crashing machine. Five grain samples (presumed cosmic dusts) were picked out under binocular microscope after eleclromagnetic w The grains were weighed with a microbalance ( + 1 ~tg, Perkin-Elmer, AD-7, USA) and wrapped into clean weighing papers for short irradiations. 0236-5731/97/USD 1ZOO 9 1997 Akad~miai Kiadt, Budapest All rights reserved

Table 1. INAA scheme of grain samples Irradiation time flux

Decay time

Counting time, s

5m I 9 1013 n. cm-2. s-1 l17h 3.5 9 1013 n. cm-2. s-1

6m 40 m 7h 15 h

400 1 500 2000 8 000

15d

40000

Elements determined AI, Ca, Cu, V, Ti, Mg Dy, K, Mn, Na Cu, Dy, K, Mn, Na Sm, Ho, Na, K, Ga, La, Yb, As, Sb, W, U Ce, Cr, Co, Cs, Eu, Fe, Hf, Lu, Nd, Sc, Tb, Th, Zn

Elsevier Science S. A., Lausanne Akad~miai Kiadt, Budapest

BANGFA NI et

al.: MULTIELEMENTALNAA

OF GEOLOGICAL MIICRO-GRAIN SAMPLES

Table 2. INAA results of grain samples, in lag/g

Element AI* As Ca* Ce Co Cr Cs Cu Dy Eu Fe* Ga Hf Ho K* La Lu Mg* Mn Na Nd Sb Sc Sm Tb Th Ti* O V W Yb Zn

No-8-2 (23 lag) 85.6 4.7 33 130 102 170 1.5 360 11.1 2.97 548 43.5 11.7 2.23 5.2 55.4 1.79 430 1770 74 3.9 33.6 8.1 1.9 22.8 8.2 25.2 767 10.6 13.8 240

(3.3) (8.5) (13) (3.5) (3.2) (5.4) (9.6) (8.5) (5.6) (5.5) (3.1) (4.3) (4.6) (4.1) (3.7) (3.2) (3.5) (3.6) (3.2) (29) (8.5) (3.2) (1.1) (11) (4.3) (15) (6.5) (3.6) (5.8) (3.6) (8.9)

No-3 (9 I.tg) 167 4.4 67 205 71.0 235 3.2 498 25.2 5.06 195 42.2 10.7 4.85 10.2 80.9 3.27 ' 14.6 800 4230 75 1.7 46.3 13.7 6.2 29.0 6.2 39.8 382 17.7 29.1 208

(3.2) (22) (13) (4.1) (3.5) (4.6) (9.2) (6.3) (6.9) (8.6) (3.2) (5,3) (9.9) (4.9) (3.5) (3.3) (5.4) (14) (3.2) (3.1) (11) (23) (3.5) (3.5) (8.5) (4.9) (16) (8.3) (5.5) (7.9) (3.7) (9.1)

No-8-4 (5 lag) 206 124 810 116 854 1.8 246 41.9 11.6 254 30.8 18.3 16.9 15.9 206 5.35 25.3 861 3940 320 4.1 78.1 16.9 7.4 55.8 19.5 1460 2100 24.9 33.3 312

(3.4) (15) (3.6) (3,6) (4.5) (13) (8.6) (6.8) (8.8) (3.4) (19) (5.3) (4.5) (4.2) (3.2) (6.1) (14) (3.3) (3.2) (14) (23) (3.1) (3.5) (12) (4.6) (23) (3.2) (3.9) (12) (5.0) (12)

No-5 (18 lag) 81.4 2.3 28 112 108 344 2.7 640 14.5 2.65 520 22.7 11.8 3.1 4.0 2.1 3.23 5.6 505 766 29 1.4 107 6.4

(3.2) (18) (15) (4.3) (3.6) (4.3) (15) (9.6) (6.9) (9.2) (3.2), (5.1) (12) (5.2) (3.5) (3.3) (4.5) (18) (3.2) (3.2) (18) (18) (3.1) (4.2)

23.7 6.1 58.4 514 13.9 21.8

(5.6) (23) (5.1) (3.9) (5.6) (3.5)

No-10 (25 lag) 111 (3.2) 6.5 (14) 70 (12) 280 (3.3) 72.3 (3.4) 659 (3.3) 2.1 (12) 219 (12) 19.8 (9.3) 4.87 (4.9) 598 (3.1) 15.8 (12) 15.5 (4.3) 4.5 (5.5) 7.6 (3.7) 101 (3.1) I~.33 (3.6) 6.9 (21) 442 (3.2) 1790 (3.1) 179 (13) 40.7 13.3 4.4 34.6 9.4 217 1500~ 8.2 14.7 116

(3.1) (3.6) (7.7) (3.5) (21) (3.6) ~3.4) (11) (4.2) (14)

i

*The content unit is in mg/g.

consist of counting statistics and estimated total systematic errors for which a conservative 3% are given to each data. Interferences from 235U fission were corrected for by both relative and I K o methods newly developed by our group.4 The interference factors (IFs) were determined to be 0.245, 0.247, 0.280 and 10.4 for Nd, Mo, Ce and Zr, respectively. I F s for La and Ba are dependent of decay time. Spectral interferences from 239Np (indicator nuclide of U) to 153Sm and 177Lu and possible threshold reaction interferences were also corrected for using both relative and I K o methods.5-7 In most cases, 64Cu was used as indicator of Cu. 511 keV peak areas contributed by 24Na were corrected for by a measured S(511)/S(1368) ratio of 24Na, which is 0.0301 at the experiment condition. The use of the hybrid extended K0-relative NAA technique developed by our lab has made it possible to provide an internal validation for the elements with corresponding chemical standards irradiated (a cross check between results from K0 and relative standardizations), and 180

to determine elements without standards irradiated by K 0. As a result, the accuracies are improved and all the elements detected quantified. Among the difficulties encountered in NAA of I.tg samples are calibration (or standardization) and quality control. It is hard to have chemical standards or quality control samples that can be counted at the same geometry as I.tg grain samples because of the large differences in gross activities. For the former, very few (if any) backing materials can be found to have impurity concentrations low enough to support 10-1~ -12 g elements (that is about 100-1 lag/g of elements in ~tg sample). For the latter, all the existing Certified Reference Materials (CRMs) are certified at 100 mg or larger sample size. The parametric technique for normalization of different counting positions based on EID (Effective Interaction Depth) law developed by o u r lab 6 completely solved this problem. The data in Table 2 show that most elements are enriched at some extent compared to Cl-chondrite. The

BANGFANI et al.: MULTIELEMENTALNAA oF GEOLOGICALMIICRO-GRAINSAMPLES

1000

No-3 No - 5 No-8-2 No-8-4 No - 1

0

~

100,

10

I

La

Ce

1

I

1

1

t

I

1

|

1

I

1

I

Pr

Nd

Sm

Eu

Gd

Tb

Dy

He

Er

Tm

Yb

Lu

Fig. 1. Cl-chondrite normalized REE patterns of micro-grain samples

lithophile refractory elements, such as V, Th, Hf and W have the largest enrichment factors about 200, EFs of REE are about 100, of AI, Ca, Sc, Ti, V etc. are about 50. Deficiencies were found for Co, Cr, Mg and Na. As can be seen from Fig. 1, the Cl-chondrite normalized REE patterns show a uniform enrichment with a slight higher EF for light REE. This flat mode indicates that little fractionation among REE has occurred, implying an exla'aterrestfial origin.

References 1. Ot~d C:-gV~G et al., Chinese Science (B), 10 (1986) 1089 (in Chinese). 2. R. G ~ A ~ t ' ~ et al., Science, 201 (1978) 1119. 3. MA Sm.L~ et al., J. Radioanal. Nucl. Chem., 114 (1987) 329. 4. Tlma WEI2~, J. Radioanal. Nucl. Chem., 111 (1987) 449. 5. NI BANGFA,Atomic Energy Science and Technology, 22(1) (1988) 109 (in Chinese). 6. TIAra WEtzm et al., J. Radioanal. Nuel. Chem., 170 (1993) 27. 7. T~N W ~ n et al., J. Radioanal. Nuel. Chem., 179 (1994) 119.

181