Distribution of some major and trace elements in ...

Journal of Radioanalytical and Nuclear Chemistry, Vol. 262, No. 2 (2004) 345–354

Distribution of some major and trace elements in Danube Delta lacustrine sediments and soil L. C. Dinescu,1 E. Steinnes,2 O. G. Duliu,3* C. Ciortea,1 T. E. Sjøbakk,2 D. E. Dumitriu,1 M. M. Gugiu,1 M. Haralambie1 1 Horia

Hulubei National Institute of Physics and Nuclear Engineering, Magurele, P.O. Box MG-6, RO-077125, Bucharest, Romania 2 Norwegian University of Science and Technology, Department of Chemistry, NO-7491 Trondheim, Norway 3 University of Bucharest, Department of Atomic and NuclearPhysics, Magurele, P.O. Box MG-11, RO-077125, Bucharest, Romania (Received March 25, 2004)

Sediment cores collected from lakes Mesteru and Furtuna (eastern part), Sontea channel and soil samples collected from Caraorman bar, all located in the Danube Delta, were analyzed for 42 elements (Ag, Al, As, Be, Na, Mg, P, S, K, Ca, Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Ga, As, Se, Br, Rb, Sr, Y, Mo, Ag, Cd, In, Sn, Sb, Cs, Ce, Hf, Hg, Tl. Pb, Bi, Th. U) by instrumental neutron activation analysis (INAA), thick target proton induced X-ray emission (TT-PIXE) and inductively coupled plasma-mass spectrometry (ICP-MS). The INAA and TTPIXE yielded total concentrations whereas the ICP-MS data reflected the fractions soluble in 14M HNO3. The ICP-MS data exhibited surface enrichment relative to the lower part of the sediment core of Cu, Zn, As, Ag, Cd, In, Sn, Sb, Hg, Tl, Pb, and Bi, most prominently by Cd and Hg. Their vertical distribution in the investigated cores generally reflected the pollution history of recent sediments in Danube delta, showing a steady increase until the end of the 1980s followed by a slow decrease after 1990. The vertical profiles of most remaining elements were characterized by a relatively uniform distribution along the cores. In some cases, the concentrations of As, Cd, Cu, Cr, Mn, Ni and Pb exceeded minimum thresholds of safety, as defined by the Romanian regulations. The elemental composition of the sediment below 20 cm depth (total concentrations) was similar to that of the upper continental crust (UCC) for most elements. Values distinctly higher than UCC were observed for As, Sb (factor ~5) and Cr, Ni, Cu (factor 2 to 3). The nitric acid soluble element concentrations in the soil samples in some cases showed increased values at the surface as compared to 30 cm depth, either due to air pollution or to the action of plants. In no case a large contribution to the topsoil from atmospheric deposition was evident, indicating that the surface contamination of the sediments was mainly by riverine transport.

Introduction The rapid industrial development in European countries during recent decades has a side effect by the introduction of toxic metals, fertilizers, or pesticides in many ecosystems.1–8 These pollutants released into the environment enter the atmospheric and hydrological circulations and are finally deposited on riverbeds, in reservoirs or river deltas. In this way, the lacustrine sediments, continuously enriched by all kinds of pollutants, become a long term record for the past history of contamination processes.6,8 For that reason, the investigation of the vertical distribution of various pollutants in sedimentary cores can furnish useful information concerning these processes. The Danube Delta represents a recent land formation, its history beginning by the end of the last glaciation.9 Geographically it is a flat region with a surface of 5,640 km2 (Fig. 1). Both morphologically and historically the Danube Delta can be divided into two regions:10 Fluvial (western part) and Fluvial-marine (eastern part including the Razelm-Sinoe lacustrine complex). Scattered over the entire Delta are more than 150 lakes and swamps where the bottom sediments are intensively bioturbated by various invertebrates which spend their life cycle buried in the mud.11,12 This fact also contributes to the redistribution of pollutants within the sediments.

One criterion for establishing whether an element is present as a pollutant or not is its vertical distribution in soil or sediment. In the absence of any noticeable mixing or re-working processes, the presumed pollutants usually exhibit increased concentrations in the upper, more recent layers of the sediments.3–6,8 Each country has issued its own regulations stating the limits of concentrations of heavy metals or other pollutants that can be considered acceptable or dangerous. In Romania, these limits have been established by the Minister of the Environmental Protection and published in the Official Monitor of Romania,13 and they refer to soil and sediments for industrial areas on one hand and for domestic allotments, playing fields, open spaces as well as natural parks and reservations on the other. They include both soil and sediments, but between these categories there are no differences. In this way, to be considered as a pollutant, any potentially harmful compound must satisfy simultaneously two criteria: (1) to present an increased concentration near the sediment surface and (2) to exceed the legal limits. Since the Danube Delta sediments are derived from a large part of the European continent a comparison of the obtained data with the mean concentrations of the same elements in the Upper Continental Crust (UCC)14 could be very useful in interpreting experimental data with respect to possible pollution processes.

* E-mail: [email protected] 0236–5731/2004/USD 20.00 © 2004 Akadémiai Kiadó, Budapest

Akadémiai Kiadó, Budapest Kluwer Academic Publishers, Dordrecht

L. C. DINESCU et al.: DISTRIBUTION OF SOME MAJOR AND TRACE ELEMENTS IN DANUBE DELTA

Fig. 1. Schematic map of the Danube Delta showing, in small rectangles, the locations of sampling places (Mesteru and Furtuna lakes and Caraorman bar)

High precision nuclear and atomic techniques of quantitative analysis such as instrumental neutron activation analysis (INAA), thick target proton induced X-ray emission (TTPIXE) and induced coupled plasmamass spectrometry (ICP-MS), are currently used in the investigation of environmental samples such as sediments or soils due to the very high sensitivity.15,16 In the present work, these three analytical techniques were used to study the vertical distribution of major and trace elements in three sedimentary cores (Lake Mesteru, Lake Furtuna-eastern part and Channel Sontea) as well as some soil samples collected in the southeastern part of Danube Delta (Caraorman bar) (Fig. 1). The main purpose of this work was to uncover the existence of any pollution of the Danube Delta sediments with the elements in question. 346

Experimental Lake hydrology Lake Mesteru is located in the western part of the Fluvial Delta (Fig. 1). It has a relatively small surface (2.9 km2), and is permanently fed by the channel Mila 23. During the last 15 years the water circulation of this channel has diminished due to a progressive accumulation of sediments. At the same time there is a continuous communication between Lake Mesteru and the neighbor, Lake Lung by means of a small natural channel. Lake Mesteru is characterized by soft, lacustrine sediments having a thickness up to 50 cm and a low sedimentation of 0.20±0.07 g.cm–2.y–1 as determined from 137Cs vertical profiles.17


The Furtuna lake located in the central part of the same region (Fig. 1) is a relatively big lake (surface 11.6 km2) by respect to previous ones. It communicates with the channel Sulina by means of at least five channels. The most important of these is Sontea, which transports a substantial quantity of sediments yielding a high sedimentation rate in the lake (0.6–0.8 g.cm–2.y– 1).17 The sediments are of fluvial type, especially in the south-western part of the lake. The other four channels maintain an active exchange of detritus with the Sulina and Chilia branches [18]. Assuming an average annual sedimentation ratio of 0.7 g.cm–2.y–1 the sediment at 50 cm depth would have an age of greater than 100 years. This age corresponds to the end of 19th century, when the countries within Danube hydrographic basin were at the beginning of industrialization, and consequently the amounts of heavy metals discharged into the Danube River from industrial sources may have been very low. Samples Sediment: In November 1996 three sediment cores containing mainly recent deposits were collected in lake Furtuna (eastern part), lake Mesteru, and channel Sontea (near lake Furtuna) by using a piston corer. The resulting cores with a diameter of 7 cm, were stored in vertical position. Prior to analysis the cores were cut in half longitudinally, carefully examined for any sediment disturbance and then sectioned into pieces 1–2 cm thick. These samples were dried at 105 °C, homogenized by grinding and sieved (1.6 mm). From each sample two aliquots were taken for analysis. One sample of about 0.3 g was subjected to INAA without further pretreatment. For TTPIXE analysis about 1 g of each sample as well as of reference sediment IAEA SL119 were pressed at 8 MPa into pellets of 8–12 mm diameter and 1–3 mm thickness. More details concerning sediment characteristics and composition are presented in Reference 20. Soil: In September 2001, soil samples were collected at three sites within the region of Caraorman bar by means of a hand corer. Samples were taken at the surface (5 cm thick) and at approximately 30 cm depth. The samples were dried at 105 °C, ground and homogenized. The same procedure as above was used for sample preparation. INAA measurements For INAA samples and standards (IAEA reference sediments SL1 and SL3)19 were wrapped in plastic bags and irradiated for 4 hours in a wet vertical channel of the Institute of Physics and Nuclear Engineering VVR-S Reactor, at a thermal neutron flux of 1.2.1013 n.cm–2.s– 1. The irradiated samples were measured after a cooling

time of 6 to 8 days (counting time: 1000 seconds) and 20 to 25 days (counting time: 5000 seconds), by using a Ge(HP) detector with a FWHM resolution of 1.9 keV at 1332 keV (60Co) and a relative efficiency of 30%.21 The detector was connected to a Canberra MCA (8192 channels) through an AccuSpec A board. OS2/Genie-PC software was used for the gamma-spectrum processing. TTPIXE measurements Pressed pellets were analyzed by TTPIXE using 3 MeV protons delivered by the 9 MV Van de Graaff tandem accelerator of NIPNE. The analyses were carried out in vacuum. A collimated beam (3 mm diameter) bombarded the target oriented at an angle of 45° with respect to the beam direction. The beam current entering the scattering chamber, measured by using an Ortec Model 439 current digitizer, was maintained at 1–10 nA, and the samples were typically irradiated for a collected charge of 10 µC. The emitted X-rays were measured by a Canberra Ge(HP) detector, having an energy resolution of 180 eV at 5.9 keV and placed at an angle of 90° to the incident beam direction. For the present measurements, the Xrays passed through the Be windows of the scattering chamber (0.25 mm thick) and detector (76 µm thick), and a 3 cm air gap. The signals from the detector preamplifier were processed with a Tennelec spectroscopic amplifier Model TC244, enabling pile-up rejection, and then fed into a Model 1520 Canberra ADC and mixer-router. Data manipulation and storage were performed with a Canberra S100 counting system, based on an IBM personal computer. The X-ray spectra21,22 were analyzed using the code LEONE, which models the X-ray peaks with Gaussian functions and subtracts a polynomial background (1 to 3 degree). The peak areas derived from the fitting routine, corrected for X-ray attenuation and detector efficiency, were then used to determine the elemental concentrations. Quantitative analysis was based on normalization to the beam charge and measuring a sample of the standard reference material IAEA SL-119 at identical conditions. ICP-MS measurements Samples of about 0.4 g were weighed into Teflon bombs and decomposed with 4 ml 14M HNO3 in a microwave oven. After dilution to 0.5M HNO3 the samples were subjected to analysis by a sector field ICPMS using a Thermo (Finnigan) Element instrument (Bremen, Germany). The RF power was 1150 W. The sample was introduced using a CETAC ASX 500 autosampler (Omaha, USA) with a peristaltic pump (pump speed 1 ml/min). The instrument was equipped with a concentric Meinhard nebulizer connected to a 347


Scott spray chamber, and a quartz burner with a guard electrode. The nebulizer argon gas flow rate was adjusted daily to give a stable signal with maximum intensity for the nuclide 115In. The instrument was calibrated using 0.5M HNO3 solutions of multielement standards at appropriate concentrations. After each sample 0.1M HNO3 (Suprapur) was flushed through the sample introduction system to reduce memory effects. To check for possible drift in the instrument, a standard solution with known elemental concentrations was analyzed for every 10 samples. In addition, blank samples (0.5M HNO3, Suprapur) were analyzed for approximately every 10 samples. The samples were analyzed in random order. Results and discussion Data from different analytical techniques The data presented in this paper include totally 42 elements. The results by different analytical techniques, however, are not directly comparable in all cases. The values from INAA and TTPIXE represent the total contents of the elements in the samples. The concentrations obtained by these techniques are, therefore, directly comparable. The elements determined by INAA were K, Sc, Cr, Fe, Co, Zn, As, Br, Rb, Sb, Hf, Th, and U. INAA is also favorable for the determination of several rare earth elements (REE), but since the REE geochemistry of these elements in Danube Delta sediments had been discussed in details in a previous paper23 they were not included here. The elements determined by TTPIXE were K, Ca, Ti, Cr, Mn, Fe, Ni, Cu, Zn, Ga, Br, Rb, Sr, Zr, and Pb. In six cases there is an overlap between the elements determined by both techniques. As shown in Table 1, results by the two techniques, excepting K and Rb are in good agreement. Data from INAA and TTPIXE are suited, e.g., for

comparison with standard geochemical listings of elements such as the estimated composition of the upper continental crust.14 The ICP-MS data on the other hand are based on the fraction of each element in the sample dissolved by the treatment with concentrated nitric acid. Any form of the elements bound to the surface of the soil or sediment particle, including additions by water or air pollution, will be soluble in the acid. When it comes to the fraction of an element bound in the lattice of mineral particles, however, it depends on the character of the mineral whether that fraction is soluble or not. In the case of typical lithophilic elements contained in silicate minerals a considerable fraction is generally insoluble in nitric acid. This means that the ICP-MS data in many cases do not represent the total content in the sample and are thus not directly comparable with the INAA and TTPIXE data. On the other hand, the ICP-MS data are likely to be better suited for disclosing any contribution from pollution or other processes concentrated in the surface sediment or soil, since they are less likely to reflect the part of the element contained in the natural mineral material. The elements determined by ICP-MS were Be, Na, Mg, Al, P, S, Ca, Sc, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, As, Se, Br, Rb, Sr, Y, Ag, Cd, In, Sn, Sb, Cs, Ba, Ce, Pr, Sm, Hg, Tl, Pb, Bi, and U, 36 in all. The quality of the determinations (excepting Be, Na, Sc, Se, Br, Ag, Sn, Ba, Pr, Sm, Hg, Bi, and U) was checked by simultaneously analyzing three humic soil reference samples for which published recommended values existed for the HNO3-soluble fractions of most of the above elements.2 A comparison of present data for the three reference samples with the literature values is shown in Table 2. Excepting Zn and As, whose concentration are higher than the recommended values, in the other cases the agreement can be considered satisfactory.

Table 1. Concentrations of six elements in a sediment core from Lake Futura Est as determined by INAA and TTPIXE Depth, cm 0.5 2.5 4.5 6.5 8.5 11 15 19

348

K, mg.kg–1 INAA PIXE 18.8 11.7 19.2 11.3 21.1 13.5 18.1 11.9 15.6 9.2 20.0 10.7 16.3 12.6 15.1 6.5

Cr, mg.kg–1 INAA PIXE 95.5 89.0 93.8 91.9 121.0 85.0 103.0 99.7 111.0 98.3 107.0 79.7 89.4 97.3 81.2 89.6

Fe, g.kg–1 INAA PIXE 36.5 44.2 37.9 42.5 44.2 40.9 42.0 42.2 42.2 42.4 44.6 40.6 36.5 43.4 33.8 40.5

Zn, mg.kg–1 INAA PIXE 169 185 159 168 194 173 171 183 170 158 146 130 118 128 107 118

Br, mg.kg–1 INAA PIXE 13.6 11.9 12.4 17.1 14.2 11.2 12.7 12.1 10.5 16.2 13.3 16.4 11.9 16.8 13.1 10.5

Rb, mg.kg–1 INAA PIXE 44.7 70.1 102.0 63.0 117.0 57.1 117.0 59.6 80.8 64.7 104.0 54.4 97.0 53.0 78.3 49.7

L. C. DINESCU et al.: DISTRIBUTION OF SOME MAJOR AND TRACE ELEMENTS IN DANUBE DELTA Table 2. Analysis of humic soil reference samples 28 by ICP-MS (in mg.kg–1). Excepting Zn and As whose values are systematically higher, all other elements reported in the present work are in good agreement with the recommended values Element Mg Al P S K Ca V Cr Mn Fe Co Ni Cu Zn As Rb Sr Y Cd In Sb Cs Ce Tl Pb

Present work 761 1600 1285 1900 1100 2730 2.43 1.16 151 951 1.60 5.06 9.7 94 1.12 10.6 24.7 0.76 0.58 0.004 0.33 0.061 4.14 0.091 21.7

H-1 Recommended28 778 ± 17 1850 ± 100 1240 ± 90 1540 ± 140 1270 ± 70 2710 ± 40 2.40 ± 0.24 1.37 ± 0.15 144 ± 8 941 ± 97 1.51 ± 0.08 4.69 ± 0.28 7.18 ± 0.47 53.9 ± 2.6 0.85 ± 0.04 10.5 ± 0.3 25.8 ± 1.4 0.78 ± 0.06 0.43 ± 0.02 0.008 ± 0.002 0.37 ± 0.08 0.102 ± 0.004 4.8 ± 0.5 0.085 ± 0.005 19.2 ± 1.0

Present work 477 880 812 1500 700 2210 3.14 1.40 138 811 0.79 3.01 8.3 70 2.08 4.6 19.5 0.32 0.83 0.010 0.43 0.134 1.26 0.093 31.9

Vertical distribution of metals in sediments As indicated above, the ICP-MS results are probably most suitable to indicate any surface enrichment of elements caused by pollution or other factors. Results from the upper 50 cm of a sediment core from lake Mesteru seem to distinguish the elements that are increased near the sediment surface due to pollution of the Danube water. As seen in Fig. 2, some of these elements, i.e., Cu, Zn, As, Cd, Hg, Tl, and Pb are enriched in the upper sediment layer. However, the maximum concentrations are evident a few cm below the sediment surface, probably indicating a significant reduction of industrial activity during the last decade in the former communist countries within the Danube river drainage area.6,24 On the other hand, elements also generally known to be associated with pollution, such as V and Ni, show no such surface enrichment. A closer look at the data reveals that also As and Bi, and probably even Ag, In, and Sn, are enriched near the sediment surface. The near-surface enrichment is most prominent for Cd and Hg, by about a factor of 5. All the above elements are well known as air pollutants.25 From all the investigated sediments, the ones in Lake Mesteru seem to be the most polluted, followed by Lake Furtuna (East) and Sontea Channel. This is in good

H-2 Recommended28 550 ± 14 1140 ± 140 770 ± 80 1410 ± 100 820 ± 140 2510 ± 210 3.33 ± 0.72 2.00 ± 0.26 148 ± 0.05 934 ± 74 0.92 ± 0.04 3.00 ± 0.25 7.18 ± 0.47 43.7 ± 2.5 1.76 ± 0.04 6.18 ± 0.28 24.8 ± 1.6 0.38 ± 0.08 0.69 ± 0.03 0.011 ± 0.001 0.74 ± 0.06 0.196 ± 0.017 1.61 ± 0.27 0.099 ± 0.011 31.9 ± 3.0

Present work 515 2200 1010 2150 860 1550 7.73 4.15 53 2600 8.07 113 518 284 10.4 6.7 25.6 1.14 2.40 0.012 0.90 0.180 4.72 0.147 84.3

H-3 Recommended28 564 ± 26 2480 ± 250 850 ± 80 1710 ± 100 930 ± 80 1620 ± 70 6.82 ± 0.67 4.46 ± 0.85 51.5 ± 2.3 2620 ± 240 7.69 ± 0.25 90 ± 6 401 ± 20 149 ± 6 6.6 ± 0.6 7.37 ± 0.34 28.6 ± 2.0 1.09 ± 0.13 1.66 ± 0.05 0.107 ± 0.024 1.60 ± 0.10 0.197 ± 0.003 4.9 ± 0.7 0.137 ± 0.011 71.2 ± 3.7

agreement with differences in lake hydrology, since Lake Mesteru has a less open water circulation than Lake Furtuna and Sontea Channel, and thus a more significant fraction of pollutants may be deposited there. Most of the other elements show no appreciable variation in their HNO3-soluble concentrations within the upper 50 cm of the sediment, which is to be expected for typically lithophilic elements during a time period with small variations in sedimentation rate. In addition to V and Ni this type of distribution is evident for Na, Mg, Al, Sc, Cr, Fe, Co, Rb, Cs, Ba, and Th. Some elements show a variation with depth in the sediment that may be due to natural factors rather than pollution. In the case of S, Mn, and U there is a maximum at about 20 cm depth which is probably related to variations in the redox level with depth in the sediment. The concentrations of Ca and Sr are markedly higher in the uppermost 15 cm of the sediment than at greater depth. This difference is also evident in the TTPIXE data for total concentrations of the same elements. Also for P the HNO3-soluble concentration is higher near the surface. This peculiarity, also found in the Matita Lake,23 seems to correlate with the presence of a great amount of mollusk shell debris in the uppermost part of sediments.

349


Fig. 2. Vertical profiles of Ni, Cu, Zn, As, Cd, Hg, Tl, and Pb in lake Mesteru lacustrine sediments as determined by ICP-MS

350


Vertical distribution of metals in soils In a similar way as for the sediments the ICP-MS results for soils from the surface layer and at 30 cm depth may indicate contribution from pollution due to atmospheric deposition or occasional flooding. However, the action of plants to transport elements from deeper layers to the soil surface by root uptake and transfer to the green parts works in the same direction. After the death of the green parts of the plant the elements are enriched in decaying organic material at the soil surface. Surface enrichment of some elements in the soil relative to the level at 30 cm depth was observed (Table 3). Among the elements most enriched in the surface layer are Zn, Ag, Cd, Sb, Hg, and Pb, which are all typical components in long-range transported pollution aerosols in the atmosphere.25 On the other hand, both Cu, Zn, Cd, and Ba are known to be appreciably enriched in topsoils due to the upward transport by plants.26 The observed surface soil levels, however, are low compared to areas where air pollution is a major source of these metals to the topsoil27 and indicate that atmospheric deposition is of much less importance than riverine transport in contributing to the surface contamination of sediments in the Danube Delta. Comparison with Romanian legal limits for heavy metal pollution In Fig. 3 sediment data for seven metals (Cr, Mn, Co, Ni, Cu, As, and Pb) obtained by INAA and TTPIXE as well as Cd obtained by ICP-MS (no data for this element were available from the other techniques) are presented in whisker plots, and compared to the regulatory limits for heavy metals specified by Romanian environmental authorities.13 It must be pointed out that, according to Romanian legislation, sediments are considered as “sensitive environment”. For that reason the “minimum threshold of safety”, as we have used, has the lowest values. As seen the observed concentrations of Cr, Mn, Ni, Cu, As, Cd, and Pb in some cases exceed the “minimum threshold of safety”. The exceedance is particularly marked in the case of As in Mesteru Lake and Cd and Pb in Sontea channel. Danube Delta sediments in relation to composition of upper continental crust

Co, Ni, Cu, Zn, Ga, As, Rb, Sr, Zr, Sb, Hf, Pb, Th, U) below 20 cm depth in the investigated sediments, in comparison with the corresponding UCC concentrations as reported by TAYLOR and MCLENNAN.14 It is assumed that the sediment below 20 cm depth is not significantly influenced by pollution as discussed above, and thus representing the original natural composition. Among the major elements Ti, Mn, and Fe are close to the UCC, whereas Ca is more variable and exceeds the UCC value by a factor of 3 in Furtuna lake and Sontea channel. In the case of trace elements Sc, Co, Zn, Co, Hf, Pb, and U levels are close to The UCC values, whereas Cr, Ni, and Cu (factor 2–3) and As and Sb (around a factor of 5) are distinctly higher than the corresponding UCC values. The reason for this is not known. The higher Cr concentration, also reported previously in other papers, seems to be characteristic of Danube Delta sediments.7,11,23,24 In general, the results seem to support the idea that the Danube catchment has a general element composition quite similar to that of the UCC. Comparison of ICP-MS values with total concentrations As stated above, a direct comparison of results from the “total content” techniques INAA and TTPIXE with those from ICP-MS is not possible because the latter reflects only what is soluble in concentrated nitric acid. On the other hand the present data may illustrate to what extent the extensively used HNO3 decomposition method is capable of dissolving lithophilic elements from the mineral matter present in riverine sediments. Comparison of ICP-MS values with either INAA or PIXE for sediment samples below 20 cm depth indicate the extent of dissolution by nitric acid as it follows: Ni, As, Br and Pb more than 95%, Ca, Fe and Zn between 80 and 90%, Co, Cu and Mn between 70 and 90% and Sc, Cr and Rb between 50 and 60%.

Table 3. Surface enrichment in soil relative to 30 cm depth (mean value for 3 sites) Element S Cu, Zn, Br, Ag, Cd, Sb, Ba, Hg, Pb Be, P, Rb, In, Tl V, Mn, As, Sn, Cs Mg, Al, Ca, Cr, Fe, Co, Ni. Sr, Y, Ce

Relative enrichment 8.3 2.5–3.1 2.0–2.5 1.5–2.0