High-precision measurement of calcium isotopes in carbonates and related materials by multiple collector inductively coupled plasma mass spectrometry (MC-ICP-MS) Ludwik Halicz,a,b Albert Galy,a Nick S. Belshawa and R. Keith O'Nionsa aDepartment of Earth Sciences, University of Oxford, UK OX1 3PR bGeological Survey of Israel, 30 Malkhey Israel St., 95501 Jerusalem, Israel. E-mail: Ludwik@mail.gsi.gov.il Received 6th August 1999, Accepted 21st September 1999 Multi-collector ICP-MS has been used for the precise measurement of natural variations in the isotopic composition of Ca.The interference of Ar in the Ca mass region is assessed and the repeatability of the 44Ca/42Ca ratio of a sample calcium solution relative to the National Institute of Standards and Technology (NIST) Standard Reference Material (SRM) 915a Calcium Carbonate (Clinical) standard is better than 0.1ù at 95% conÆdence. Variations in sample 44Ca/42Ca ratio are expressed as d44Ca units, which are deviations in parts per 103 from the same ratio in the NIST SRM 915a Ca standard.Measurements of d44Ca are presented for terrestrial and marine carbonates, which show a variation of up to 0.7ù, in agreement with previous studies by thermal ionisation mass spectrometry (TIMS). Introduction Calcium is the Æfth most abundant element in the silicate earth, with an abundance of y3%. It has six stable isotopes, 40Ca, 42Ca, 43Ca, 44Ca, 46Ca and 48Ca and is an essential element in both animal and plant tissues. Isotopic studies of Ca are relatively few and include investigation of environmental samples as well as the dietary inØuence on calcium intake by the human body.1 A study of Ca isotopes in natural samples2±5 has revealed a variation up to y4ù in the 44Ca/40Ca ratio,4 suggesting that biological fractionation of Ca isotopes takes place in the food chain.4,5 Additionally, Ca in oceanic carbonate sediments has been observed to be lighter than the seawater from which it precipitated.5 Recent studies have concluded that the Ca isotope composition of the oceans are maintained at their present value through biological fractionation.4,5 Conventional ICP-MS is a relatively insensitive technique for the measurement of calcium isotopes with a typical precision of y5ù for isotope ratios.This is a consequence of the relatively poor peak shapes achievable with quadrupole mass spectrometers and the requirement for sequential rather than simultaneous measurement of the isotopes of interest.More precise measurement of Ca isotope ratios (y0.1ù) have been made using thermal ionisation mass spectrometry (TIMS) using a double-spike to allow corrections to be made for instrumental mass fractionation.2±6 This technique, although accurate, has not been widely exploited, in part, because of the necessity to accurately calibrate spikes and to chemically purify Ca prior to isotopic analysis. The technique reported here employs a multiple collector (MC)-ICP-MS to produce highprecision measurements of Ca isotopes even in the presence of 40Arz.This technique is relatively rapid and has the additional advantage of requiring minimal sample puriÆcation. Experimental Calcium isotopes The measurement of Ca isotopes using Ar-ICP source mass spectrometers is potentially hindered by the presence of 40Arz, preventing accurate measurement of 40Caz. However, Ca has Æve other isotopes, of which three are both sufÆciently abundant and convenient for measurements to be made at high precision.These are 42Ca, 43Ca and 44Ca with abundances of y0.65%, 0.13% and 2.1%, respectively. Interference at these masses may include molecular species and doubly charged ions.7,8 In addition, 46Ca has a very low abundance, of only 0.003%, and 48Ca, although of adequate abundance (0.18%), is too far removed from 42Ca on the mass focal plane of the spectrometer for simultaneous collection of all Ca isotopes to be possible.Furthermore, possible interference from 46Ti and 48Ti also serve to make 46Ca and 48Ca less suitable than the lighter isotopes of Ca. For 42Ca, 43Ca and 44Ca to be useful any interference at these masses, which may include molecular species and doubly charged ions,7,8 must be either absent or sufÆciently small that an accurate correction can be made. In addition to the spectral interference mentioned above, there are possible scattered background contributions from the large ( y5 nA) 40Arz and 40Caz ion beams, which enter the collector array along with the masses of interest.This study describes a technique, which both minimises and corrects for, potential isobaric interference on 42Ca, 43Ca and 44Ca and other non-spectral contributions to the background. Sample preparation All samples and standards were prepared in dilute 0.1 M nitric acid solution and introduced into the Ar-plasma source through a modiÆed Cetac (Omaha, NE, USA) MCN 6000 desolvating nebuliser.This device minimises the introduction of H2O, CO2, O2 and N2 into the plasma thus reducing the abundance of interfering molecular species. Natural samples of carbonate mineral were dissolved in 3 M HCl at room temperature. The residue after dissolution was separated by centrifugation and the supernate evaporated to dryness and then re-dissolved in 2 M HNO3. Samples and standards were diluted with 0.1 M HNO3 to 20±30 ppm Ca to obtain the best counting statistics for mass spectrometric analysis.J. Anal. At. Spectrom., 1999, 14, 1835±1838 1835 This journal is # The Royal Society of Chemistry 1999Mass spectrometry Ca isotope ratios were measured using a Nu Instruments MCICP- MS (Nu Instrument Ltd, Wrexham, Wales). This instrument is a double focusing magnetic sector instrument with variable dispersion ion optics and a Æxed array of 12 Faraday collectors.9 The normal operating conditions adopted for the mass spectrometer are summarised in Table 1.The isotopes of interest, 42Ca, 43Ca and 44Ca, were positioned as indicated in Fig. 1. Since the Faraday collectors are Æxed and the instrumental mass dispersion can be varied by a factor of y2, these isotopes were collected by Faradays 4, 8 and 10, respectively, within the overall array. In this conÆguration the 40Arz and 40Caz ion beams also enter the collector and are a potential source of scattered ions and secondary electrons.It is important, in this case, to assess the contribution of scattered ions to the measured signals. Faraday collectors corresponding to m/z 41.5, 42.5, 43.25, 43.75 and 44.25 are used for this purpose, allowing corrections to be applied for elevated background levels. Mass discrimination for Ca isotopes in the MC-ICP-MS is y5% u21 and is monitored with reference to an external Ca isotope standard, using the standard±sample±standard bracketing technique. Samples and standards of similar Ca concentration (20±30 ppm) are introduced into the instrument in 0.1 M HNO3.Results and discussion The measured 44Ca/42Ca isotope ratios are expressed relative to the same ratio in the National Institute of Standards and Technology (NIST) Standard Reference Material (SRM) 915a Calcium Carbonate (Clinical) Ca isotope reference standard as follows: d44CaÖùÜ~fÖ44Ca=42CaÜsample=Ö44Ca=42CaÜSRM 915a{1g |1000 Previous studies2,4,5 reported variations in 44Ca/40Ca or 40Ca/44Ca isotope ratios as d44Ca.However, the variation in 40Ca abundance by decay of 40K implies that 44Ca/40Ca isotope ratios are not only dependent on mass-dependant fractionation. For instance, the 44Ca/40Ca isotope ratios differ by up to 0.3ù between crustal rocks and mantle-derived material only because of differences in age and K/Ca ratios.10 By analogy with the notation used for all other stable isotopes we prefer to normalise to 42Ca, the lighter isotope not affected by radiogenic decay of other elements.Interferences The mass spectrometric measurement of Ca isotopes is potentially complicated by the presence of either spectral or non-spectral interference. Accurate analysis depends upon an understanding of, and allowance for, these as follows. Non-spectral interferences. The optimum situation in magnetic sector mass spectrometry is where the background signal levels are indistinguishable from the intrinsic background noise of the detector system.The region of the mass spectrum around m/z 40 in MC-ICP-MS instruments may present both spectral problems from the 40Ar isobar on 40Ca as well as non-spectral interference arising from scattering of the 40Arz and 40Caz ion beams entering the collector housing. As a result elevated levels of background may occur on all detectors, which in this case is approximately 1 mV (Fig. 2). Correction for this non-spectral contribution was made by the simultaneous monitoring of signal levels in the relevant Faraday collectors at positions either 0.25 or 0.5 u removed from the Ca masses.Thus for 44Caz the scattered background was monitored at 44.5 u, whilst for 43Caz and 42Caz the backgrounds at 43.25 and 42.5 were monitored. Interferences arising from multiply charged species at these monitor positions are not present, and similarly low signal levels are observed at 41.5, 43.25, 43.5, 43.75, 44.25 and 45 u (Fig. 1). Doubly charged ion interferences.The most likely source of doubly charged ion interference is from 88Sr2z, 86Sr2z and Table 1 Instrumental operating conditions and signal measurement parameters RF power 1400W Plasma gas Øow rate 12 l min21 Interface cones Nickel Acceleration voltage 4 kV Ion lens setting Optimised for max. intensity Instrument resolution y300 Mass analyser pressure 361029 mbar Detector 12 Faraday collectors Nebuliser Microconcentric Spray chamber temp. 75 �C Desolvator temp. 160 �C Sweep gas (argon) 3.65 l min21 (optimised daily) Sample uptake rate 70 ml min21 Typical Ca44 sensitive 0.3 V (ppm)21 Sampling time 3 repetitions of 20610 s Fig. 1 Typical background levels from spectral interferences and scattered contribution from 40Arz observed during nebulisation of 0.1M HNO3. The collector conÆguration is that used for static measurement of 42Ca, 43Ca and 44Ca described in the text. The precision of background measurements is 3±4% RSD. Fig. 2 Spider diagram of background signal in the mass range 41.5± 45.5 u during nebulisation of 0.1 M HNO3 (solid line) and a 30 ppm solution of calcium in 0.1M HNO3 (dashed line).The scattering of 40Arz and 40Caz are observed to be very similar. 1836 J. Anal. At. Spectrom., 1999, 14, 1835±183884Sr2z on 44Caz, 43Caz and 42Caz, respectively. These interferences cannot be mass resolved whilst maintaining high ion-transmission using the current generation of small geometry MC-ICP-MS instruments. 87Sr2z, however, may be monitored at 43.5 u. After background correction at 43.5 u, based upon interpolation of the backgrounds at 43.75 and 43.25 u, a correction for Sr2z is applied to the measured 44Caz, 43Caz and 42Caz, if necessary. Under the conditions adopted for Ca isotope analysis, Sr2z/Srz varies between 0.02 and 0.05 and for natural carbonates the 87Sr/86Sr ratio will be in the range 0.706±0.720.11 For a maximum Sr2z/Caz ratio of 0.05, this range of 87Sr/86Sr ratios would introduce an uncertainty in d44Ca of less than 0.5ù.In natural carbonates with less than 2000 ppm of Sr, the uncertainty arising from the Sr correction is below the precision of the d44Ca measurement. If required, separation of Ca from Sr is easily achieved by liquid chromatography. In some samples, particularly silicates, 87Rb2z may also be present at 43.5 u, but this is not usually a problem with carbonates. Molecular interferences. Potential molecular interferences in the Ca mass region include 14N2 16Oz and 12C16O2 z on 44Caz and 40ArH2 z on 42Caz.7,8 The use of a desolvating nebuliser acts to reduce O, N, C and H interference to an insigniÆcant level (Fig. 1). The measured interferences on Ca isotopes, after correcting for non-spectral contributions are less than 0.2 mV at 42, 43 and 44 u. Repeatability of standard Ca isotope ratios The Nu Instruments MC-ICP-MS produces Ca peaks with Øat tops at a working mass resolution of y300, as required for high-precision Ca isotope ratio measurement. The Ca isotopes of interest are positioned in the multiple collector for simultaneous measurement as shown in Fig. 3. The optimum Ca concentration for the standard solution is 15±30 ppm, given the sensitivity for 44Ca of y0.3 V (ppm)21. Higher concentrations of Ca were found to cause a reduction in instrument sensitivity due to material deposition on the sample and skimmer cones. Ca isotopes of samples or standards analysed by MC-ICPMS are not corrected for instrumental mass discrimination, either with internal or external standards.The standard± sample±standard bracketing technique has been adopted here to examine repeatability of measurement, as this permits a correction to be made for instrumental drift. Individual measurements of sample and standard isotope ratios comprised 20 measurements of 10 s integration with simultaneous measurement of 42Ca, 43Ca and 44Ca and 87Sr2z at 43.5 u as described above.The stability of uncorrected of 44Ca/42Ca and Fig. 3 Peak shapes and coincidences of calcium isotopes 44Ca, 43Ca, 42Ca and strontium double charge 87Sr (m/z~43.5) with the variable mass dispersion optics arranged for measurement of isotopes in the mass region of the mass spectrum. Fig. 4 Evolution of the measured, uncorrected calcium isotope ratio (44Ca/42Ca) of standard and samples through time. $, NIST SRM 915a; , CaCO3 from Aldrich; %, CaCO3 from Alfa; , specpure CaCO3 from Merck; © speleotherm 2-8-E3.Table 2 Ca isotopic composition for selected carbonate and commercial reagents (N~number of replicates during a 2 month period) Sample d44Ca °1s N Commercial Ca reagent– SRM 915aa CaCO3 0.01 0.08 9 Specpureb CaCO3 0.34 0.10 3 Alfac CaCO3 0.54 0.05 4 Aldrichd Ca solution 0.54 0.11 2 Merckb CaCO3 0.61 0.14 4 Continental environment– 2-8-E3 Speleotherm Israel 0.25 0.16 3 2-8-G Speleotherm Israel 0.17 0.19 4 2-8-J Speleotherm Israel 0.04 0.08 3 SA 310 Calcrete Israel 0.44 0.11 4 SA 495 Calcrete Israel 0.49 0.20 5 Marine environment– Chalk Jurassic France 0.65 0.14 2 Pocillopora Coral Red Sea 0.60 0.02 3 Acropora Coral Red Sea 0.58 0.08 2 Metamorphic rocks– Carrara Marble Italy 0.75 0.05 3 Spar Calcite Israel 0.58 0.03 3 aNIST, Gaithersburg, MD, USA.bMerck, Darmstadt, Germany. cJohnson Matthey, Karlsruhe, Germany. dMilwaukee, WI, USA. J. Anal. At. Spectrom., 1999, 14, 1835±1838 183743Ca/42Ca ratios during extended runs of up to 5 h was 0.1 and 0.15ù h21, respectively (Fig. 4). The corresponding typical 2s precision after correction for background and interference was 0.1ù. The precision of a measured sample using the sample± standard bracketing technique was observed to be 0.1ù, assessed from the external repeatability measured over a period of 2 months (Table 2). The measured isotope ratios were found to be unaffected by the presence of magnesium in concentrations up to twice that of calcium.Several carbonate samples from continental, marine and metamorphic sources were analysed (Table 2). These show a range of 0.7ù on d44Ca. Conclusion The performance of the MC-ICP-MS for Ca isotope ratio measurement is evaluated and shown to equal the precision and accuracy of the best results obtained by TIMS. Elevated background levels resulting from ion scattering of 40Arz and 40Caz are corrected for during the analysis and 87Sr2z is monitored to allow corrections for Sr interference.The throughput of this proposed method is about 12±15 samples per day compared to 2±3 for TIMS. The range in isotopic composition of pure calcium obtained commercially (y0.7ù) conÆrms previous studies2 and suggests the use of a homogenised and well-distributed standard, such as NIST SRM 915a, for normalisation of stable isotopic composition of Ca. The preliminary results for several carbonate samples indicate a different isotopic composition of Ca in marine and terrestrial environments.It is necessary to investigate the systematics of the biological and geochemical control of calcium isotope rs in the global calcium cycle. Acknowledgements This research has been supported by a grant from the Natural Environment Research Council. AG has been supported by the EC through the TMR ``Marine Record of Continental Tectonics and Erosion'' No. ERBFMXCT 960046. References 1 R. I. Price, G. N. Kent, K. J. B. Rosman, D. H. Gutteridge, J. Reeve, J. P. Allen, B. G. A. Stuckey, M. Smith, G. GuelÆ, C. J. Hickling and S. L. Blakeman, Biomed. Environ. Mass Spectrom., 1990, 19, 353. 2 W. A. Russell, D. A. Papanastassiou and T. A. Tombrello, Geochim. Cosmochim. Acta, 1978, 42, 1075. 3 I. Platzner and N. Degani, Biomed. Environ. Mass Spectrom., 1990, 19, 822. 4 J. Skulan, D. J. DePaolo and T. L. Owens, Geochim. Cosmochim. Acta, 1997, 61, 2505. 5 P. Zhu and J. D. Macdougall, Geochim. Cosmochim. Acta, 1998, 62, 1691. 6 M. L. Coleman, Earth Planet. Sci. Lett., 1971, 12, 399. 7 S. Sturup, M. Hansen and C. Molgaard, J. Anal. At. Spectrom., 1997, 12, 919. 8 N. M. Reed, R. O. Cairns, R. C. Hutton and Y. Takaku, J. Anal. At. Spectrom., 1994, 9, 881. 9 N. S. Belshaw, P. A. Freedman, R. K. O'Nions, M. Frank and Y. Guo, Int. J. Mass Spectrom., 1998, 181, 51. 10 B. D. Marshall and D. J. DePaulo, Geochim. Cosmochim. Acta, 1989, 53, 917. 11 W. H. Burke, R. E. Denison, E. A. Hetherington, R. B. Keopnick, H. F. Nelson and J. B. Otto, Geology, 1982, 10, 516. Paper 9/06422B 1838 J. Anal. At. Spectrom., 1999, 14, 1835±1838