SEARCH

SEARCH BY CITATION

Keywords:

  • inductively coupled plasma–mass spectrometry;
  • element/calcium ratios;
  • foraminifera;
  • calcite

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Instrumentation, Standards, and Sample Preparation
  5. 3. Development of the Instrumental Method
  6. 4. Results and Discussions
  7. 5. Conclusion
  8. Acknowledgments
  9. References
  10. Supporting Information

[1] A method has been developed for rapid and precise simultaneous determination of nine element/Ca ratios in foraminiferal tests directly from intensity ratios using external, matrix-matched standards on a quadrupole inductively coupled plasma–mass spectrometer (ICP-MS). All quantification isotopes are determined in pulse mode to avoid cross-calibration. Small argide (40Ar26Mg) interferences on 66Zn are corrected by using two additional Mg and Zn standards. A stable signal, conducive for high-precision measurements, is obtained by cone conditioning. Variable calcium concentration has negligible effect on Li, Al, Mn, and Sr, but Ca concentrations for standards and samples need to be constrained at a similar level for precise measurements of Zn, Cd, and U. Aliquots of samples are first analyzed for Ca concentrations on an inductively coupled plasma–atomic emission spectrometer (ICP-AES), and the remaining solutions are diluted to Ca concentration of 100 ppm for ratio measurements to assure data quality. The long-term reproducibility of the method yielded precisions of Li/Ca = 2.4%, B/Ca = 4.2%, Mg/Ca = 1.4%, Al/Ca = 14%, Mn/Ca = 0.9%, Zn/Ca = 2.8% (1.2∼7.8 μmol/mol) and 5.1% (0.5∼1.2 μmol/mol), Sr/Ca = 0.9%, Cd/Ca = 2.4% (0.07∼0.24 μmol/mol) and 4.8% (0.01∼0.07 μmol/mol), and U/Ca = 2.5% for foraminiferal samples as small as 60 μg.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Instrumentation, Standards, and Sample Preparation
  5. 3. Development of the Instrumental Method
  6. 4. Results and Discussions
  7. 5. Conclusion
  8. Acknowledgments
  9. References
  10. Supporting Information

[2] The chemical composition of foraminiferal calcite accumulating in deep sea sediments has been widely used to reconstruct chemical or physical properties in ancient oceans [e.g., Boyle, 1988; Elderfield and Ganssen, 2000; Marchitto et al., 2000; Hall and Chan, 2004]. Large differences in the abundance of chemical constituents of foraminiferal calcite have required element ratios in foraminiferal shells to be determined by a combination of different methodologies and instrumentation [e.g., Boyle, 1981; Boyle and Keigwin, 1985; Marchitto et al., 2000]. Because elements are analyzed sequentially in solutions of different dilutions by different methods, the approaches are usually time-consuming and not always conducive to high-precision ratio determinations. Simultaneous generation of element ratios has the advantage of reducing overall sample size requirements and analysis time, and improves the precision of correlation between proxies. ICP-MS techniques provide rapid and precise measurements of several elemental ratios by full use of the rapid-scanning capabilities of ICP-MS and its linear response over 8–9 orders of magnitude of dynamic range using dual-mode acquisition. Successful ICP-MS methods include utilization of a combination of internal 45Sc standardization and isotope dilution (ID) techniques [e.g., Lea and Martin, 1996] and sector field double-focusing ICP-MS [Rosenthal et al., 1999]. ID involves complex calibrations and calculations and increases the potential for spectral interferences due to the requirement to measure 2 isotopes per element. Magnet jumps are required on sector field ICP-MS because of large mass intervals during element ratio measurements, and therefore a significant fraction of the overall analysis time is spent on the magnet settling, which reduces the duty cycle and increases the influence of signal fluctuations and degrades final precision. In this paper, we present an alternative approach for precise simultaneous determination of nine element/Ca ratios on a quadrupole ICP-MS by exploiting the rapid scan ability. All samples are analyzed at similar Ca concentrations ([Ca] = 100 ppm) to that of external standards based on initial [Ca] measurements of aliquots on ICP-AES to overcome matrix-dependent effects. Despite limitations of lower sensitivity relative to sector field ICP-MS, the method presented here allows rapid and precise determination of multiple element ratios in purified foraminiferal shells using sample weights as small as 60 μg. The long-term reproducibility over a three-month period is comparable to or better than other methods. We describe the method and then show element/Ca ratios of core top Cibicidoides wuellerstorfi from the north Atlantic Ocean to demonstrate its potential.

2. Instrumentation, Standards, and Sample Preparation

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Instrumentation, Standards, and Sample Preparation
  5. 3. Development of the Instrumental Method
  6. 4. Results and Discussions
  7. 5. Conclusion
  8. Acknowledgments
  9. References
  10. Supporting Information

[3] The method described here was developed on a Perkin-Elmer Elan DRC II instrument at the University of Cambridge. Operation conditions are outlined in Table 1. The instrument sensitivity was optimized daily using a 10 ppb Mg-In-U standard. The sensitivity of the instrument is ∼40 kHz for 1 ppb 115In at a sample uptake rate of 60 μl/min. Plasma robustness was monitored by constraining CeO/Ce ratio within 3% to refrain formation of polyatomic oxides. A microautosampler (Cetac ASX100) was used for sample introduction. Because boron can be leached from glassware under acidic conditions, a quartz Cyclonic spray chamber was used to minimize B blanks. Given the limited foraminiferal sample size, small diameter peristaltic pump tubing (ID: 0.03 mm) and a glass micro-concentric nebulizer Micromist FM005 (Glass Expansion, Australia) with a narrow orifice were employed to minimize solution consumption, producing an uptake rate of ∼60 μl/min at a pump rate of 12 rpm. One measurement for the determination of nine element ratios over ∼4 min requires ∼250 μl solution at a calcium concentration of 100 ppm, equivalent to ∼60 μg purified calcite.

Table 1. Instrument, Hardware, and Acquisition Parameters
Instrument and Introduction SettingsValue/Description
  • a

    CPS per ppb of each element.

  • b

    RSD, relative standard deviation (precision) of typical counting rate.

  • c

    Potential interferences are 7Li18O on 25Mg, 46Ti on 46Ca, 40Ar26Mg, and 1H214N18O16O2+ on 66Zn, and 87Rb on 87Sr.

RF power, W1300
Detection modepulse
Scan modepeak hopping
Auto lenson
Gas flow rate, L/min 
Plasma15
Auxiliary1.2
Nebulizer0.99–1.02
AutosamplerCetac ASX-100
NebulizerGlass expansion, micromist, FM005
Spray chambercyclonic quartz
Tube0.03 mm ID (red-orange)
Uptake rate, μl/min60
Washout time, s60
Uptake time, s65
Signal acquisition parameters 
IsotopeDwell Time, ms/amuTypical Count Rate (CPS/ppb)aRSD,b %
Li-71025002.0
B-111010003.0
Mg-25c0.58002.0
Ca-46c20.5<1.0
Al-270.5150001.5
Mn-550.512000<1.0
Zn-66c525002.0
Sr-87c0.51500<1.0
Cd-1116015002.0
U-238209000∼4.0
Sweeps/reading250  
Readings/replicate1  
Replicates6  

[4] Trace metal clean procedures were adopted throughout to minimize laboratory contamination. Multielement stock standard mixtures were prepared gravimetrically by spiking a 10,000 μg/ml Ca standard with appropriate amounts of Li, B, Al, Mn, Zn, Sr, Cd and U mono-elemental 1,000 μg/ml certified ICP-MS grade stock solutions. Concentrations of minor elements in the primary calcium standard were measured by ICP-AES and included in the calculations of the element ratios of stock standards. Cross-contamination from other high-purity primary standards was negligible due to the small volumes used. Each stock solution has all minor and trace elements of interest and the final ratios of solutions were spaced linearly to contain the natural ratio ranges expected in both planktonic and benthic shells. Working standards were made by diluting the stock standard solutions with 0.075M HNO3 to give calcium concentrations of 100 ppm.

[5] The approach of using isotopes to calibrate elemental ratios or concentrations assumes standards and samples have identical or similar (natural) isotopic abundances. The lithium standard purchased for this study was artificially depleted of 7Li (92.32% vs. 92.48% for natural). The natural abundance of 11B (80.17%) is also different from values of foraminiferal samples which are expected to be 80.40–80.43% if assumed to have δ11B ratios of 25∼27‰ [Palmer and Pearson, 2003]. However, it is safe to ignore these errors because the correction factors (0.9983 for Li and 0.9968–0.9971 for B) are very close to unity.

[6] Ten to twenty individual foraminifera tests were handpicked from disaggregated sediments and cleaned by a combination of the methods of Barker et al. [2003] and Rosenthal et al. [1997] modified from the procedure described by Boyle and Keigwin [1985]. Briefly, ∼300 μg shells were crushed gently to just allow any chamber fill to escape during the following cleaning stages. Successive deionized water and methanol washes were performed to remove clays followed by coarse grained silicates removal under a binocular microscope. More rigorous reductive and oxidative steps were carried out in the clean laboratory to remove ferromanganese oxides and organic matters, respectively. Before dissolution, samples were rinsed twice with dilute acid (0.001M HNO3) to remove any adsorbed contaminants. Finally, cleaned samples were dissolved in 200 μl 0.075M HNO3 for measurements. To overcome matrix effects (see section 3.5), 20 μl was first diluted to 250 μl for Ca content analysis by ICP-AES [de Villiers et al., 2002] and the remainder was diluted to a Ca concentration of 100 ppm for element ratio determinations by ICP-MS. When ICP-AES is not available, ICP-MS can be employed for initial [Ca] measurements.

3. Development of the Instrumental Method

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Instrumentation, Standards, and Sample Preparation
  5. 3. Development of the Instrumental Method
  6. 4. Results and Discussions
  7. 5. Conclusion
  8. Acknowledgments
  9. References
  10. Supporting Information

[7] The method developed here took advantage of the fast scan capability of quadrupole ICP-MS to determine intensity ratios. Element/calcium ratios were determined directly from drift-corrected intensity ratios using external, matrix-matched standards, a procedure similar to that described previously by Rosenthal et al. [1999] and de Villiers et al. [2002]. However, several parameters had to be carefully considered in order to analyze additional elements.

3.1. Detection Modes and Choice of Isotopes

[8] The large abundance differences between major (Ca, Mg and Sr) and trace (e.g., Zn, Cd and U) elements have previously required dual (analog and pulse) modes to be used [Lea and Martin, 1996; Rosenthal et al., 1999]. We determined all isotopes using pulse mode to avoid the need for cross calibration, or the determination of an extra calcium isotope in analog mode. To achieve this, extremely short integration times were allocated to low abundance, clean (no isobaric interference) isotopes of major elements (Table 1). 46Ca was selected because of its very low abundance (0.004%) and only 0.5 ms/amu dwell times were set for 25Mg and 87Sr. The most abundant, clean isotopes were chosen for trace elements to obtain reliable and high signals (Table 1). Because of the relatively high Li and B contents in foraminifera tests and the high abundances of quantification isotopes (7Li: 92.5% and 11B: 80.1%), 10ms/amu was allocated for both 7Li and 11B. A short dwell time of 5ms/amu was set for 66Zn due to its high ionization efficiency. Cd and U are among the least abundant elements in foraminifera tests. Considering the low sensitivity caused by low isotopic abundance (12.8%) of 111Cd and the low ionization efficiency (<70%) of Cd [Jarvis et al., 1992], a large portion of the analysis time was allocated for this isotope to improve counting statistics. In contrast, only 20ms/amu was required for 238U due to its high abundance (99.3%) and high mass-dependent sensitivity.

3.2. Blank and Memory Effect

[9] Blanks were measured using the same acid as used for sample dissolution and dilution. When compared with typical ratios of foraminifera tests, blanks were <1% for Ca, Mg, Sr and Li; higher blanks were observed for Cd (<2%), and U (<5%) due to their extremely low abundance and for Zn (<4%) because of the difficulty of complete purification of the dissolution acid. The B blank was substantially decreased to ∼5% by the employment of a quartz spray chamber, compared with ∼30% when using a glass spray chamber. Determination of B at trace levels by ICP-MS is plagued by a memory effect which results from the tendency of boron to volatilize as boric acid from the sample solution layer covering the inside surfaces of the spray chamber [Al-Ammar et al., 1999, 2000]. Aqueous and gaseous ammonia have been used to overcome this memory effect [Al-Ammar et al., 1999, 2000] but are unsuitable when additional elements are measured. Because only ultratrace levels of boron (1∼10 ppb B) were analyzed, we were able to minimize the memory effect by allocating longer washout and uptake times. On the basis of our experiments, blanks for all isotopes were stable during a typical run (∼5 hr). For each sample or standard, the net intensities were calculated by subtracting the average blanks from the average raw intensities of 6 replicate scans.

3.3. Spectral Interferences

[10] Abundance sensitivity may cause inaccuracies or imprecisions in the determination of 11B by ICP-MS. Signal scanning over masses 8–13.5 using a 0.075M HNO3 indicates that 11B does not suffer from the tail of the adjacent 12C peak. Because of the high concentrations of Ca, Mg and Sr and the absence of polyatomic oxides under well-controlled plasma conditions, the potential 46Ti, 7Li18O and 87Rb interferences on 46Ca, 25Mg, and 87Sr are insignificant. This was confirmed by observation of identical Mg/Ca and Sr/Ca ratios for samples measured both by ICP-MS (this method) and ICP-AES [de Villiers et al., 2002]. The third most abundant Cd isotope 111Cd (12.8%) was selected in preference to the more abundant 112Cd (24.1%) or 114Cd (28.7%) because of the absence of an isobaric interference from Sn, which is present in foraminiferal tests.

[11] Although free of isobaric interferences from monoisotopic species, it has been documented that 66Zn can suffer overlap by some plasma-related (e.g., 40Ar26Mg) and/or acid matrix-related (e.g., 1H214N18O16O16O+) polyatomic species [Mason et al., 2004]. To test the potential interference from 1H214N18O16O16O+, we compared the 66Zn intensities for quartz distilled H2O and nitric acid of two different concentrations (0.075M and 1.0M). We found similar intensities for the three solutions, suggesting negligible 66Zn interference from HNO3 related polyatomics. Due to the prevalence of zinc in nature, it was not possible to obtain a Zn-free Mg standard. Therefore possible interferences from 40Ar26Mg were evaluated by comparing the slopes of calibration plots for Mg and Zn serial standards (Figure 1). Interference from 40Ar26Mg was indicated by a steeper Zn slope for Mg serial standards than that for Zn solutions (Figure 1a). The formation of the argide (40Ar26Mg) interference was proportional to Mg content in solution. A correction was applied off-line as follows:

  • equation image

where CF is the correction factor and K1, K2, K3 and K4 are slopes indicated in Figures 1a and 1b, and

  • equation image

Generally, CF values are around ∼0.002, implying small, and sometimes negligible, interferences. However, the correction becomes important when foraminiferal samples have high Mg and low Zn contents, such as planktonic foraminifera from tropical areas. Measurements of CFs at the beginning and end of analyses give similar values, indicating constant formation rate of the argide interference over the course of a day. In practice, CF was calculated daily from one Mg and one Zn standard and the same CF was applied for all standards and samples analyzed.

image

Figure 1. (a and b) Mg interferences on Zn illustrated by steeper slopes for Mg serial standards (open squares) than for Zn (solid squares). A correction factor (CF) can be calculated by K1, K2, K3, and K4 assuming formation of 40Ar26Mg argide is proportional to Mg content, which is supported here. In this case, CF is 0.0022. Zn concentrations for Mg and Zn series were calculated from those of the primary standards. CF was calculated daily by Ks defined by 0.075M HNO3 and two (one each of Mg and Zn) diluted standards.

Download figure to PowerPoint

3.4. Drift Correction and Ratio Calibrations

[12] When measuring high-calcium samples such as foraminifera, data quality can be seriously affected by decreasing sensitivity caused by Ca deposition on sampling cones. We found it possible to attain stable sensitivities for foraminiferal solutions by a “cone conditioning” procedure: simply injecting pure Ca solution (100 ppm) to the instrument for 0.5–1 hours before optimization. Conditioning proved extremely helpful because more samples could be analyzed in a run and precise measurements were assured by stable signals. The mechanism behind conditioning is unknown but we suspect that calcium deposition on cones (mainly the skimmer) reaches a maximum or saturation state by calcium injection during the conditioning stage. When cones were well conditioned, signal drift was usually gradual (10∼20%) and corrected off-line using linear interpolation between two consecutive drift monitors with intermediate concentrations among standards, similar to sample-standard bracketing (SSB) methods used for isotope analysis on multiple-collector ICP-MS. Drift correction standards were analyzed every 3 samples in order to closely monitor the instrument performance and achieve optimal correction quality in case of sudden changes in drift. We acknowledge that a drift monitor every 3 samples is very frequent and in the future we will consider increasing this to every 5 samples. The intensity ratios for standards and samples were obtained by dividing the blank, interference and drift corrected intensities off-line in a spreadsheet. Intensity ratio results for a suite of standards were used to construct calibration curves following the method employed for precise Mg/Ca and Sr/Ca analysis by ICP-AES [de Villiers et al., 2002]. The calibration curves determined from multiple standards are linear and R2 are usually greater than 0.999. The slopes and intercepts of these calibrations were applied to calculate element/calcium ratios of samples from their intensity ratios. Typically, 6–10 hours are required to analyze 8 external calibration standards to give nine 8-point calibration lines and 50–80 samples interspersed with drift correction standards every 3 samples.

3.5. Matrix Effects

[13] The implications of a Ca matrix effect on trace elements are critical for accurate ratio determinations. Figure 2 presents element/calcium ratios analyzed for seven serial dilutions with [Ca] ranging from 60 to ∼240 ppm. All solutions were diluted from a single concentrated mixed standard and therefore contained identical ratios. On the basis of these results, the calcium concentration has a negligible effect on accuracy for Li/Ca, Mg/Ca, Al/Ca, Mn/Ca and Sr/Ca ratios over the studied [Ca] range. In contrast, increasing Ca concentration lowers measured Zn/Ca ratios, raises U/Ca ratios and increases scatter in Cd/Ca. These observations contrast with those of sector field ICP-MS [Rosenthal et al., 1999] where Mg/Ca and Cd/Ca ratios decrease with increasing calcium concentrations and the calcium effect on U/Ca is insignificant, suggesting that differences in instrument design or plasma condition influence matrix-induced mass discrimination. To obtain reliable ratios for all elements, all sample solutions were analyzed at 100 ppm [Ca]. If only the minor elements (Li, Mg, Al, Mn and Sr) are targeted, the pre-screening by ICP-AES could be omitted when sample sizes are limited. Zn/Ca, Cd/Ca and U/Ca ratios could also be considered reliable without pre-screening, provided [Ca] can be held within the range of 80–120 ppm using raw 46Ca intensities of samples and correction standards. However, off-line corrections are not recommended for precise ratio measurements due to the instability of the matrix effect changing with time and working conditions (Figure 2b) when [Ca] of samples are far from 100 ppm.

image

Figure 2. Calcium-induced matrix effects on (a) Li/Ca, Mg/Ca, Al/Ca, Mn/Ca, and Sr/Ca and (b) Zn/Ca, Cd/Ca, and U/Ca ratios over [Ca] of 60∼240 ppm. Horizontal dashed lines indicate ±2.5% envelopes and vertical shaded areas represent [Ca] ranges for reliable element/calcium determination. The true ratios for the standards are Li/Ca = 10.25 μmol/mol, Mg/Ca = 2.55 mmol/mol, Al/Ca = 0.20 mmol/mol, Mn/Ca = 51.5 μmol/mol, Zn/Ca = 4.13 μmol/mol, Sr/Ca = 1.59 mmol/mol, Cd/Ca = 0.099 μmol/mol, and U/Ca = 9.95 nmol/mol. Accuracy (%) = [(measured ratio-true ratio)/true ratio]*100%.

Download figure to PowerPoint

3.6. Long-Term Precision and Detection Limit

[14] Analytical robustness was examined over a three-month period by analyzing several solutions with ratios ranging from extremely low to those typical of the compositions of foraminifera. Over the three-month period, diluted (100 ppm [Ca]) standards were prepared for several times from stock solutions. Cross-checking with a new set of standards indicates that the stock solutions are stable at least over the timescale of half a year. Long-term precision and accuracy for standards compared with analytical precisions of other methods (Table 2) shows that precisions are comparable to or improve on other methods developed on different ICP-MS instruments. For example, the long-term precision of 2.4% for Li/Ca compares with a previous reported value of 3.5% [Hall and Chan, 2004]. Without using ammonia, B/Ca ratios could be analyzed with an RSD of 3–5% although precisions degraded at low ratios. The data quality of Zn/Ca, Cd/Ca and U/Ca were comparable to other methods. Precisions of Mg/Ca and Sr/Ca are lower than those obtained using other methods [de Villiers et al., 2002; Schrag, 1999; Rosenthal et al., 1999], but are usually sufficient for foraminiferal analysis due to their wide ranges in ratios. Because Al/Ca is only used as an indicator of contamination from silicates and almost all samples have Al/Ca ratios below detection limit, the precision for Al/Ca is sufficient for its purpose.

Table 2. Detection Limit and Long-Term Precision and Accuracy for Elemental Ratios Determined by the Current Method for External Standards of Various Concentrations; and Short-Term Reproducibility of Alternative Methods
RatioRangeaD.L.bRSD,b %Acc.,b %nbRSD of Other Methods, %
  • a

    Mg/Ca, Sr/Ca, and Al/Ca: mmol/mol; U/Ca: nmol/mol; others: μmol/mol.

  • b

    D.L. (detection limit) = 3 * SD/m, where SD is the standard deviation of several measurements of a sample whose ratio is close to detection limit; m is the slope of the calibration curve; Unit is that for ratio range. It is assumed here that the SD of the signal from samples near the D.L. is similar to the SD from blanks [Harris, 2002]. RSD% (relative standard deviation) = [SD of measurements/average ratio]*100%; Acc.% (accuracy) = [(average of measured ratios-true ratio)/true ratio]*100%; n, number of replicate analyses.

  • c

    Quadrupole ICP-MS [Hall and Chan, 2004].

  • d

    Sector field ICP-MS [Wara et al., 2003].

  • e
  • f
  • g

    Sector field ICP-MS [Rosenthal et al., 1999].

  • h

    Quadrupole ICP-MS, ID method [Lea and Martin, 1996].

  • i
  • j

    ICP-AES [Schrag, 1999].

  • k

    Quadrupole ICP-MS [Russell et al., 1994].

Li/Ca5.3–29.40.52.420.39120∼3.5c
B/Ca41–221154.172.57120<10d
Mg/Ca0.38–5.000.031.390.61120<0.3e; 0.21f; 0.45g
Al/Ca0.02–0.460.0514.069.30120 
Mn/Ca13–1940.30.930.23120<1g; 5h
Zn/Ca1.2–7.80.052.830.8288<2∼3i
Zn/Ca0.5–1.2 5.051.6932 
Sr/Ca0.45–2.010.020.920.34120<0.3e; 0.21f; 0.45g; 0.1∼0.2j; 1.5∼2h
Cd/Ca0.07–0.240.0052.370.93501.70g; 2.5∼3.5h; 3∼6i
Cd/Ca0.01–0.07 4.800.8070 
U/Ca5.5–290.52.541.091201.4g; <7k

[15] At low B/Ca (<∼60 μmol/mol), Zn/Ca (<∼2.0 μmol/mol) and Cd/Ca (<∼0.05 μmol/mol) ratios (ratios mainly found in planktonic foraminifera; benthic foraminifera usually have much higher ratios), analytical reproducibility was affected by the blank and low instrumental sensitivity. As a result, relatively poor long-term precisions were obtained for B/Ca (3–7%), Zn/Ca (3–6%) and Cd/Ca (3–8%) over these extremely low ranges (Table 3). These ratios are very close to their respective detection limits (Table 2). The maximum absolute errors associated with B/Ca, Zn/Ca and Cd/Ca are 3, 0.06 and 0.001 μmol/mol, which are much smaller than natural variation ranges for B/Ca (50–100 μmol/mol), Zn/Ca (0–2.5 μmol/mol) and Cd/Ca (0.005–0.050 μmol/mol) in planktonic foraminifera [Rickaby and Elderfield, 1999]. Therefore this poorer precision does not complicate paleoceanographic interpretations for such samples.

Table 3. Long-Term Analytical Reproducibility for B/Ca, Zn/Ca, and Cd/Ca at Low Ratiosa
B/Ca, μmol/molRSD,b %Zn/Ca, μmol/molRSD,b %Cd/Ca, μmol/molRSD,b %nb
  • a

    Long-term here is 3 months.

  • b

    RSD% (relative standard deviation) = [standard deviation of measurements/average ratio]*100%; n, number of replicate analyses.

416.70.386.00.0118.216
483.60.744.10.0205.116
623.91.234.30.0313.222
782.71.763.50.0472.816

4. Results and Discussions

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Instrumentation, Standards, and Sample Preparation
  5. 3. Development of the Instrumental Method
  6. 4. Results and Discussions
  7. 5. Conclusion
  8. Acknowledgments
  9. References
  10. Supporting Information

[16] Element/Ca ratios of core top C. wuellerstorfi from the north Atlantic Ocean were analyzed to illustrate application of the method. The foraminifera shells were cleaned according to the procedure described in section 2 and the results are presented in Table 4 together with Cd/Ca and Zn/Ca data of some nearby core tops from other studies. The samples have Mn/Ca < 50 μmol/mol (Table 4), indicating the absence of contamination from Fe-Mn oxides and/or MnCO3 overgrowths [Boyle, 1983]. Mg/Ca ratios decrease with decreasing bottom water temperature, reflecting the temperature control of Mg incorporation into benthic foraminiferal shells [Lear et al., 2002]. Sr/Ca ratios show a decreasing trend with increasing water depth, which is consistent with previous studies [Elderfield et al., 1996; Lear et al., 2003]. Considering differences associated with bioturbation, chemical compositional heterogeneity between foraminifera from different samples, local hydrographic, and cleaning effects [Boyle, 1984, 1995], the Cd/Ca and Zn/Ca data show good agreement with previous studies which involved several separate methodologies (see Table 4). Cd/Ca of BOFS 11K obtained from this study is similar, within natural variation, to that given by Elderfield et al. [1996]. The higher Zn/Ca ratio relative to the nearby core V29-193 is expected [Marchitto et al., 2002] because V29-193 has a shallower water depth and lower nutrient level. The core tops have a Li/Ca range of 12.4–14.0 μmol/mol, consistent with the results of C. wuellerstorfi from a similar water depth in the Caribbean basin [Hall and Chan, 2004]. U and B in benthic foraminifera have not been studied previously. U/Ca ratios of analyzed samples lie within the range of planktonic foraminifera [Russell et al., 2004], while B/Ca of C. wuellerstorfi are much higher than those of planktonic foraminifera analyzed by Wara et al. [2003]. Systematic studies are required to investigate the paleoceanographic implications of both proxies.

Table 4. A Comparison of Results From This Method to Published Ratios for Core Top C. wuellerstorfi From the North Atlantic Oceana
Sample IDLatitude, °NLongitude, °WWD, MBWT,bMg/Ca, mmol/molZn/Ca, μmol/molSr/Ca, mmol/molCd/Ca, μmol/molLi/Ca, μmol/molB/Ca, μmol/molMn/Ca, μmol/molU/Ca, nmol/molReference
  • a

    Al/Ca ratios of all samples are below detection limit.

  • b

    Bottom water temperature was estimated by World Ocean Circulation Experiment (WOCE) data set (http://cdiac.esd.ornl.gov/).

BOFS 11K55.2020.4020042.91.152.261.360.09012.4202248.5this study
BOFS 11K55.2020.402004    0.080    Elderfield et al. [1996]
V29-19355.4018.731326  2.08 0.068  50 Marchitto et al. [2002]
NEAP 8K59.4422.1226492.60.980.761.350.04113.4218487.9this study
NEAP 10B59.9023.302221    0.046    Elderfield et al. [1996]
V29-20461.1923.001849  1.22    67 Marchitto et al. [2000]
NEAP 20B42.2928.2428783.41.401.641.290.04914.0203135.8this study
CHN82 21PG43.3029.832103    0.056    Boyle [1988]
CHN82 15PC43.2328.132155    0.051    Boyle [1988]
CHN82-20PC43.5029.873020  1.60    18 Marchitto et al. [2002]

5. Conclusion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Instrumentation, Standards, and Sample Preparation
  5. 3. Development of the Instrumental Method
  6. 4. Results and Discussions
  7. 5. Conclusion
  8. Acknowledgments
  9. References
  10. Supporting Information

[17] The method developed allows rapid and precise simultaneous determination of 9 element/calcium ratios in foraminifera tests using a quadrupole ICP-MS. Stable signals over a measurement session are maintained by cone conditioning. A small 40Ar26Mg interference on 66Zn is corrected by daily measurement of two additional Mg and Zn standards. Matrix effects on Zn, Cd and U are minimized by controlling the Ca concentration. The technique is efficient in generating elemental ratios of paleoceanographic interest simultaneously from a single purified foraminiferal carbonate, thereby reducing overall sample size and analytical time and improving the correlation between proxies.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Instrumentation, Standards, and Sample Preparation
  5. 3. Development of the Instrumental Method
  6. 4. Results and Discussions
  7. 5. Conclusion
  8. Acknowledgments
  9. References
  10. Supporting Information

[18] We thank Linda Booth for picking the foraminifera analyzed in this study. The manuscript was improved following constructive reviews by Tom Marchitto and Daneil Sinclair and helpful comments from the Associate Editor, Pamela Matin. Research funding from the Gates Cambridge Trust (to J.Y.) and EU grant EVK2-CT-2002-00135-6C (to H.E.) is appreciated.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Instrumentation, Standards, and Sample Preparation
  5. 3. Development of the Instrumental Method
  6. 4. Results and Discussions
  7. 5. Conclusion
  8. Acknowledgments
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Instrumentation, Standards, and Sample Preparation
  5. 3. Development of the Instrumental Method
  6. 4. Results and Discussions
  7. 5. Conclusion
  8. Acknowledgments
  9. References
  10. Supporting Information
FilenameFormatSizeDescription
ggge703-sup-0001tab01.txtplain text document1KTab-delimited Table 1.
ggge703-sup-0002tab02.txtplain text document2KTab-delimited Table 2.
ggge703-sup-0003tab03.txtplain text document0KTab-delimited Table 3.
ggge703-sup-0004tab04.txtplain text document1KTab-delimited Table 4.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.