Elemental ratio determination via ICP-MS and DCP-AES : methodology to extract climate records from coral aragonite

Citation
Elemental ratio determination via ICP-MS and DCP-AES : methodology to extract climate records from coral aragonite

Material Information

Title:
Elemental ratio determination via ICP-MS and DCP-AES : methodology to extract climate records from coral aragonite
Creator:
Hilderbrand, Douglas C.
Place of Publication:
Tampa, Florida
Publisher:
University of South Florida
Publication Date:
Language:
English
Physical Description:
vii, 73 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Aragonite ( lcsh )
Coral reef ecology ( lcsh )
Paleoclimatology ( lcsh )
Dissertations, Academic -- Geology -- Masters -- USF ( FTS )

Notes

General Note:
Thesis (M.S.)--University of South Florida, 1998. Includes bibliographical references (leaves 59-61).

Record Information

Source Institution:
University of South Florida
Holding Location:
Universtity of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
025151268 ( ALEPH )
40178504 ( OCLC )
F51-00138 ( USFLDC DOI )
f51.138 ( USFLDC Handle )

Postcard Information

Format:
Book

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

ELEMENTAL RATIO DETERMINATION VIA ICP-MS AND DCP-AES: METHODOLOGY TO EXTRACT CLIMATE RECORDS FROM CORAL ARAGONITE by DOUGLAS C. HILDERBRAND A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science Department of Geology University of South Florida May 1998 Major Professor: Terrence M. Quinn, Ph.D.

PAGE 2

Graduate School University of South Florida Tampa, Florida CERTIFICATE OF APPROVAL Master's Thesis This is to certify that the Master's Thesis of DOUGLAS C. HILDERBRAND with a major in Geology has been approved by the Examining Committee on April 9, 1998 as satisfactory for the thesis requirement for the Master of Science degree Examining Committee: Major TerreR"Ce)M. Duinn, Ph.D. Membe' Peter Harries, Ph. D.

PAGE 3

ACKNOWLEDGMENTS I would like to express my dearest thanks to Terry Quinn for giving me the opportunity to explore my unique interests and providing encouragement, guidance, and expertise throughout my graduate school experience. I also would like to thank the members of my thesis committee, Dr. Jeff Ryan Dr. Johan Schijf and Dr. Peter Harries for their endless efforts and instruction in geochemistry It was never easy, but definitely worth it. I also would like to extend a special thanks to Thierry Correge for allowing us to attempt to reproduce an outstanding analytical technique. A special thanks to my family, friends, loved one, and canine companion for being especially patient. It has been a long time coming. A sincere thank you is given to the faculty and staff of the University of South Florida Geology Department for a Grade A education. A final thank you to the 1998 El Nino for keeping me indoors so I could finish writing this thesis. This study was funded in part by the National Science Foundation and in part by a grant from the Creative and Research Scholarship program at USF.

PAGE 4

TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES ABSTRACT I. II. INTRODUCTION Corals as a Climate Proxy Statement of Problem Objectives CORALLINE AND GEOCHEMICAL BACKGROUND Coral Setting Geochemical Background III SAMPLING PROCEDURE IV. EXPERIMENTAL PROCEDURE ICP-MS Instrumentation DCP-AES Instrumentation Reagents Sample Preparation Choice of Isotopes Difficulties with ICP-MS Measurement Optimization of ICP-MS Optimization of DCP i iii iv v 1 1 2 3 5 5 5 9 11 11 11 14 17 19 20 21 24

PAGE 5

v. ICP-MS MEASUREMENTS 26 SrI Ca Measurements 26 U/Ca and Ba/Ca Measurements 30 VI. DCP MEASUREMENTS: MG/CA AND SR/CA 33 VII. RESULTS AND DISCUSSION 39 Precision of Measurements 39 Isotope Ratio Curves 41 Instrumental Comparison 53 Sample Consumption and Output 53 Comparison of Elemental Ratio Utility 56 VIII. CONCLUSIONS 57 IX. REFERENCES 59 X. APPENDICES 62 Appendix A 63 Appendix B 67 Appendix C 72 Appendix D 73 ii

PAGE 6

LIST OF TABLES Table 1 ICP-MS Operating Conditions for Ca, Sr, U, and Ba 12 Table 2 DCP Operating Conditions for Mg, Ca, Sr, Sc 13 Table 3 Coral Standard Composition 15 Table 4 Calibration Standards Composition 16 Table 5 ICP-MS Analysis Protocol for Ca, Sr, U, Ba 31 Table 6 DCP Analysis Protocol for Ca Mg, Sr 38 Table 7 Reproducibility of New Caledonia Coral Standard on the ICP-MS 40 Table 8 Comparison of Instrumental Results 42 Tabl e 9 Reproducibility of New Caledonia Coral Standard on 43 the DCP iii

PAGE 7

LIST OF FIGURES Figure 1 Location Map of New Caledonia 6 Figure 2 Sample Size Experiment 18 Figure 3 Dead Time Experiment 23 Figure 4 Blank Variability During DCP Run 25 Figure 5 Drift Variability During DCP Run (Blank Corrected) 25 Figure 6 43ca Standard ICP-MS Calibration Curves 27 Figure 7 48ca Standard ICP-MS Calibration Curves 28 Figure 8 86sr Standard ICP-MS Calibration Curves 29 Figure 9 Ca Standard DCP Calibration Curves 34 Figure 10 Mg Standard DCP Calibration Curves 35 Figure 11 Sr Standard DCP Calibration Curves 36 Figure 12 U/Ca Ratios Measured by ICP-MS 44 Figure 13 Sea Surface Salinity Instrumental Record 45 Figure 14 Sea Surface Temperature Instrumental Record 46 Figure 15 U/Ca Ratios vs ORSTOM SSS (Smoothed Data) 47 Figure 16 U/Ca Ratios Sampling Curve 49 Figure 17 Sr /Ca Sampling Curve 50 Figure 18 Ba/Ca Sampling Curve 51 Figure 19 DCP 88sr/40Ca and Mg/Ca Ratios 52 Figure 20 Sr /Ca Ratio Comparison 54 iv

PAGE 8

ELEMENTAL RATIO DETERMINATION VIA ICP-MS AND DCP-AES: METHODOLOGY TO EXTRACT CLIMATE RECORDS FROM CORAL ARAGONITE by DOUGLAS C. HILDERBRAND An Abstract Of a thesis submitted in partial fulfillment of the requirements for the degree of Master of Science Department of Geology University of South Florida May 1998 Major Professor: Terrence M. Quinn, Ph.D. v

PAGE 9

A major goal of coral-based paleoclimate research is to extract multicentury proxy climate records at monthly resolution Achieving this goal requires the ability to measure elemental ratios precisely and rapidly which is important because variations of these ratios in coral skeletons have been documented to reflect changes in environmental conditions in the shallow tropical oceans. U I Ca, Ba/ Ca, and Sr /Ca isotopic ratios were determined on powdered samples from a New Caledonia coral via inductively coupled plasma mass-spectrometry (ICP-MS). For Ca and Sr, calibration curves with correlation coefficients averaging 0.99975 were made using 45Sc and 89y as internal standards. The U and Ba were determined by isotope dilution. Mg/Ca and Sr /Ca elemental ratios from the same samples were determined using calibration curves via direct current plasma atomic emission spectrometry (DCP). Precision of the ICP-MS determinations was evaluated using a solution of dissolved coral powder. Relative standard deviation (RSD) values for Ca, U, Sr, and Ba were 0.25 % 3.0%, 1.3 %, and 2.1 %, respectively. With the exception of Ca, RSD values were not precise enough to define a robust environmental signal; coral paleoclimate studies require RSD values to be < 1.0%. ICP-MS productivity averaged 12 solutions per hour, of which 6 were coral samples. DCP preliminary determinations of Mg/Ca and Sr /Ca ratios yielded RSD values of 1.0% and 0.80%, respectively. Productivity averaged 24 solutions per hour, of which 6 were coral samples. Despit e the lack of precision, a record from 1967 to 1981 of 238Uj40Ca ratios displays trends that are remarkably similar to an instrumental record of sea surface salinity (SSS) determined from ocean waters near the coral collection site. In contrast, the measured 238U j40Ca record is not similar to a companion r ecord of sea s urface temperature (SST). The striking similarity vi

PAGE 10

between SSS variations and 238Uj40Ca variations is evidence that U/Ca ratios cannot be considered to be a pure paleothermometer. Ba/Ca variations displayed no discernible trend, which is consistent with the notion that seasonal upwelling of cold, nutrient-rich and Ba-rich deep water can likely be discounted as an environmental factor influencing the coral geochemistry at New Caledonia. Analytical procedures are instrument-specific; precision and productivity differences between ICP-MS and DCP instrumentation are documented. The failure to reproduce results using an existing published procedure demonstrates the need for an interlaboratory coral standard for elemental and isotopic ratio work. Ab strac t Approved: "w'VV>')\ >=< \ .""""",... Major Professor: Terrence M Quinn, Ph.D. Associate Professor, Department of Geology Date Approved: Vll

PAGE 11

I. INTRODUCTION Corals as a Climate Proxy Coral skeletons are composed of the calcium carbonate mineral aragonite, but other elements can substitute for Ca in the aragonite structure. Geochemical signals in corals reflect the environmental conditions in which the coral grew, analogous to other climate proxies such as tree rings and ice cores. Specifically, strontium and magnesium are incorporated into the coral skeleton as a function of sea surface temperature (SST) (Becket al., 1992; Mitsugushi et al., 1996). Uranium is controlled by SST, but may have competing factors including salinity and alkalinity (Min et al., 1995; Shen and Dunbar, 1995). Barium is associated with nutrient supply which increases significantly during upwelling events (Lea and Boyle, 1993; Shen et al., 1992; Lea et al., 1989) Therefore, the determination of Sr/Ca, Mg/Ca, U/Ca, and Ba/Ca ratios in corals can provide information on environmental conditions throughout the life span of the corals. As paleoclimate indicators, corals are unique in the acquisition of crucial information not currently availab l e for tropical regions They have the potential to produce climate records many centuries beyond conventional instrumental data which only range back 20 to 100 years. Corals can produce century-scale SST records with up to weekly temporal resolution (Cole 1994) Climate modelers and climatologists need tropical environmental data, especially SST records, to better understand climate variation and to validate 1

PAGE 12

their climate models for future change (Cole, 1994; Dunbar and Cole, 1993). SST records are very important because climate change is regulated by oceanatmosphere coupling (Broecker and Denton, 1989). Currently, greenhouse warming and El Nino are two highly debated topics in terms of forcing function and response of the climate system. Coral records of SST since the Industrial Revolution may permit the evaluation of the role that anthropogenic forcing plays in climate change. Statement of Problem The rapid and precise measurement of elemental ratios in coral-based climate studies has been difficult. Previous studies (Min et al., 1995; Beck et al., 1992) have measured elemental ratios by thermal ionization mass spectrometry (TIMS) with precisions of %o (2cr). TIMS instrumentation, however, is expensive and slow (Le Cornec and Correge, 1997) with a productivity of only one coral sample determination per hour. Hart and Cohen (1996) were able to analyze Mg Sr, and Ba using ion microprobe techniques with precisions of 0.3-3%. Shen and Dunbar (1995) have developed more rapid uranium measurements by inductively coupled plasma-mass spectrometry (ICP-MS) but required calcium to be measured separately by flame atomic absorption spectrometry (FAAS). A recent external isotope dilution method by Lea and Martin (1996) was discounted as a viable option to measure calcium and strontium with good precision by Le Cornec and Correge (1997). Le Cornec and Correge (1997) have developed a method that can measure U/Ca and Sr/Ca together on the same aliquot and same instrument precis ely and rapidly. This study tried to reproduce their U, Sr, and Ca results 2

PAGE 13

along with Ba. Furthermore, Mitsuguchi et al. (1996) proposed that Mg/Ca ratios closely track SST and can be measured rapidly by inductively coupled plasma-atomic emission spectrometry (ICP-AES). The present study also tried to determine Mg/Ca ratios by direct current plasma-atomic emissions spectrometry (DCP-AES), commonly abbreviated DCP. Can these recent geochemical techniques be reproduced on different instruments? Does the DCP, which has never before performed these measurements, have the potential to reproduce the results achieved by more expensive ICP-MS instruments? Do the elemental records from the coral display trends consistent with the instrumental records taken from a nearby meteorological station ? These are some of the questions this study has tried to answer. Objectives The purpose of this study was to determine U/Ca, Sr/Ca, Mg/Ca, and Ba/Ca ratios precisely and rapidly using the same aliquot from a New Caledonia coral with ICP-MS and DCP instrumentation. From these ratios, elemental records were compared to instrumental records from New Caledonia to specify the environmental controls Precision was based on long-and short-term reproducibility of a gravimetrically made coral standard. High sample output was a very important objective because the real utility of using corals as paleoclimate indicators is the production of century-scale SST records at monthly resolution. The procedure must be cost-and time efficient to be able to measure thousands of coral samples By trying to reproduce an established technique, this study attempted to make improvements and assess any machine-specific problems. Because both 3

PAGE 14

machines measured Sr /Ca ratios from the same coral samples, a direct comparison between the ICP-MS and DCP results was performed. 4

PAGE 15

II. CORALLINE AND GEOCHEMICAL BACKGROUND Coral Setting The coral sampled in this study was a Porites lutea coral collected offshore of Amedee, New Caledonia (22S, 167E) in the South Pacific Ocean (Figure 1) in August, 1992. This coral was sampled previously as part of a stable isotopic study by Quinn et al. (1996). The coral core contained approximately 355 years of growth The New Caledonia coral was chosen for the following reasons: 1) the waters which it inhabited were representative of open-ocean marine waters (with little influence from terrestrial freshwater), 2) daily SST and SSS (sea surface salinity) data have been collected for over 20 years by the French research group ORSTOM (Institut Francais de Recherche Scientifique et Technique pour le Developpement en Cooperation), 3) the instrumental SST record exhibits high-amplitude annual variations, and 4) previous studies measuring 8180 allow for geochemical comparisons (Quinn et al., 1996). Geochemical Background Uranium, barium, strontium, magnesium, and calcium all have very different concentrations in corals. The aragonitic structure of corals is 40 .04 weight percent calcium. Uranium concentrations average 3 ppm (IJ.g/ g), strontium concentrations average 8000 ppm, and magnesium concentrations 5

PAGE 16

1-40'E 1ro'E 100'E 17o'E 18()E 170'W .. ... Coral Sea Figure 1 Location Map of New Caledonia The coral sample was drilled just offshore of Amedee Island, located 20 km due south of Noumea, New Caledonia. 6

PAGE 17

average approximately 1100 ppm. Each element has its own set of geochemical characteristics. Developing a procedure that measures these different elements together has been very challenging SrI Ca ratios were first determined to reflect SST in the 1970s (Weber, 1973; Houck et al., 1977; Smith et al., 1979) However, only recently has the SST precision improved from 3.0C to .5 C by Beck et al. (1992) when they measured Sr/Ca ratios by TIMS. Sr/Ca ratios are anticorrelated to SST with a rate of change of about 0 .9% per 0C. Although Sr /Ca ratios display such low sensitivity, they remain extensively measured. U/Ca ratios were first related to SST by Min et al. (1995) and Shen and Dunbar (1995) They measured U/Ca ratios by TIMS which produced annual variations mimicking Sr /Ca ratios The process by which uranium is incorporated into the coralline structure is more ambiguous. Shen and Dunbar (1995) suggest that uranium is incorporated as but it is unknown if it even substitutes for calcium in the crystal lattice or complexes and substitutes for the carbonate anion U /Ca ratios anticorrelate strongly with SST except that the fractional change is six times greater (5.0% fOC) than for Sr/Ca. Competing factors have been proposed to influence the U/Ca ratio and include salinity, coral species extension rate, and alkalinity (Min et al., 1995; Shen and Dunbar, 1995). This study focuses on competing factors with SST influencing U /Ca ratios. Mg/Ca ratios have been the latest paleoclimatic tracer to be correlated to SST as reported by Mitsuguchi et al. (1996). Mg/Ca ratios are directly correlated to SST and show variations in Mg/Ca ratios four times greater than that of Sr /Ca ratios Mg substitutes for Ca because they have the same charge and similar ionic radii Possible competing factors against SST include 7

PAGE 18

biological variations in growth rate and interspecies differences (Mitsuguchi et al., 1996) Barium content in the coral skeleton displays a negative correlation with SST because it is associated with upwelling of colder, nutrient-rich, deeper waters (Shen et al., 1987; Lea and Boyle, 1993; Lea et al., 1989 ; Tudhope, 1994). Upwelling is the vertical displacement of warm nutrient-poor surface waters by cold, nutrient-rich source waters When determining SST records, it is important to determine any influence from upwelling activity. Upwelling events are recorded in the skeleton as increased Ba/Ca (J.lmol/ mol) elemental ratios. Upwelling has not been observed in New Caledonia and Ba/Ca ratios from this coral should support these observations. Typical Ba/Ca values in warm nutrient-poor waters range between 2 0-4.0 J.lmol/mol, while upwelling events increase the Ba/Ca ratio above 7.0 J.lmol/mol (Tudhope,1994) Although none of these paleoclimate tracers yield a complete climate record, together they have great potential for uncovering paleoclimate data that cannot b e reco v ered otherwise The development of an analytical technique to measure these ratios rapidly and precisely using the same sample aliquot has enormous utility in the field of paleoclimatology 8

PAGE 19

III. SAMPLING PROCEDURE The New Caledonia coral Porites lutea was analyzed previously in a study by Quinn et al. (1996), who gave a description of coral sampling in detail. The total growth of the coral core for this study was 492.5 mm over 40 years, of which 12 years were analyzed. The average annual growth was calculated to be 12.3 mm/year Growth rate varied from 1.08 to 1 32 em/year during the 40-year interval. Drilling was performed with monthly temporal resolution (12 samples per annual density-band couplet). Samples represented 1.03 mm of coral growth (drill bit moved 1.03 mm along the growth axis). Sampl e depth was 0.6 mm. The sampling device used in this study was a computer-aided tria x ial sampler (CATS) A computer program was designed to accommodate the specifications of thi s study. The computer program used to perform the drilling in thi s study was HILDY.ACL. The CATS drilled a rectangle with th e dimensions 1.03 mm x 1.25 mm for each coral sample. After every 12th sa mple drilled a deeper "hole" sample was drilled to serve as a yearly marker. The holes were about 1 mm deeper than the coral samples. Quinn et al. (1996) describes the instrumental setup in detail as well as options that can be performed for unrelated studies (e.g., mollusk shells) As the coral was drilled a metal prong collected the coral powder. The powder was transferr e d to a 1 5 ml microcentrifuge tube using a small finepoint paintbrush. Any remaining coral powder was removed with a 9

PAGE 20

compressed gas spray can. Pretreatment of the microcentrifuge tubes included a 1 % nitric acid bath (>24 hours), a deionized water rinse, and an oven drying. Once the powder was secured in its labeled tube, the tube was sealed and ready for analysis The target weight of the powder samples was 2 mg. However, due to inconsistencies within the coral slab (density fluctuations), collected powder weights ranged from 0.9 to 3.2 mg. 10

PAGE 21

IV. EXPERIMENTAL PROCEDURE ICP-MS Instrumentation The inductively coupled plasma-mass spectrometer (ICP-MS) used in this study to determine Sr / Ca U /Ca, and Ba/Ca isotopic ratios was a Fisons Plasmaquad PQS located in St. Petersburg, Florida in the Dept. of Marine Sciences at the University of South Florida. Samples were introduced via a Meinhard-type/ concentric glass nebulizer and a water-cooled quartz spray chamber. Ions were detected by a Galileo Channeltron type 4870V continuous dynode electron multiplier. All parameters, including xyz positioning of the plasma relative to the sampler cone, ion lens voltages and all plasma gas flows were manually controlled ICP-MS operating conditions are given in Table 1. DCP-AES Instrumentation The direct current plasma-atomic emission spectrometer (DCP) used to determine Mg/Ca and Sr/Ca elemental ratios is an ARL Spectraspan 7located in Tampa, Florida in the Dept. of Geology at the main campus of the University of South Florida The machine was set up for routine multielement analysis, with simultaneous measurement of Mg, Ca, and Sr. Cassette parameters, acquisition parameters and wavelengths used for each element are summarized in Table 2. 11

PAGE 22

Table 1 ICP-MS Operating Conditions for Ca, Sr, U, and Ba Method Acquisition parameterAcquisition tim e (sec) Scan Mode Calculation Type Isotopes Analyzed Spacing Points per peak Dwell time (Jls) Scan s per replicate Number of replicates Detector voltage (V) P l as m a parameterCool gas flow (L/min) Auxiliary gas flo w (L/min) Nebulizer gas flow (L/min) Powe r (kW) Sample uptake r ate (ml/min) Aliquot Volume Requirement (ml) Ca,Sr U,Ba 30 Peak Jumping Isotope Ratio Ca-43, Ca-48, Sc-45 U 238 U -235 Sr-86, Sr-87, Y -89 Ba-138, Ba135 10 240 250 5 DAC steps 3 3 plateau 13.5 0-2 0.7-0.9 1.35 =1 4 10 240 187.5 N

PAGE 23

Table 2 DCP Operating Conditions for Mg, Ca, Sr, Sc Physical Parameters : Nebulizer gas flow setting Entrance slit position Acquisition parameters: Integration Time (sec) Replicates Wavelength Parameters-Multi-element cassette: AMU Magnesium 24.31 Calcium 40.08 Strontium 87.62 Scandium 44 .96 Cassette Parameters : multi-element Peaking slit (micron) 25 Analysis slit (micron) 200 Exit slit (micron) 100 19 8-20.0 optimized for Mg lines 10 5 Wavelengths 280.27 317.93 407.77 424.68 Order 28 28 28 28 ...... (J..)

PAGE 24

Reagents A coral standard, four Ca-Sr calibration standards, U-Ba spike, U-500 standard, and a ScY standard solution were all used in the measurement of U/Ca and Sr / Ca ratios. All solutions made at USF's St. Petersburg campus were gravimetrically made using high purity deionized water (18 Megaohms em) from a Milli-Q/RG unit, SPEX plasma mono-elemental 1,000 ppm standards, and pre-cleaned Teflon bottles Relative weighing precision was below 0 1 %. The U-Ba spike was made at the University of Michigan and the U-500 standard was made at the University of New Mexico. All solutions were refrigerated and sealed with parafilm when not in use. Two coral standard solutions were prepared in the course of this study. The first coral standard had a total volume of 1.0 Land a calcium concentration of approximately 50 ppm while the second coral standard had a total volume of 500 ml and a calcium concentration of approximately 44 ppm. The target calcium concentration for both standards was between 40 and 50 ppm. These concentrations were in the middle of the calibration line. The same amount of U-Ba spike was added to the second coral standard (thereby doubling the U-Ba spike concentration) to correct for underspiking in the first coral standard. Table 3 lists the composition of both coral standards. The coral powder used in the coral standards was drilled from the same coral slab as the actual coral samples Using a hand drill, 10 g of coral powder were collected and shaken vigorously to homogenize the powder. Knowing that 40.04 % of aragonite is calcium, the necessary amount of powder was calculat e d to make a standard with a 40-50 ppm calcium concentration. Four calibration standards were made along with a blank solution consisting of 1 % nitric acid. The standards contained increasing 14

PAGE 25

Table 3 Coral Standard Composition Nominal Dilution Factor=8,000 Coral Standard 11/12/96 5/1/97 Total Weight: 1005 4 g 502.8 g Coral Weight: 126.2 mg 59.5 mg Nitric Acid: 14.45 ml 7.22 ml Scandium: 103 9 mg 49.7mg Yttrium: 103 1 mg 49 .4mg U-Ba spike: 59.78 g 60.94 g Remainder: 1% nitric acid Nominal Concentrations 11/12/96 5/1/97 Calcium 50 ppm 45ppm Strontium l.Oppm n/a Scandium 100 ppb 100 ppb Yttrium 100ppb 100 ppb Calculated Concentrations 11/12/96 5/1/97 Calcium 49ppm 46ppm Strontium 0.95ppm n/a Scandium 103.3 ppb 98.8 ppb Yttrium 102.5 ppb 98 2 ppb concentrations of calcium 40, 60, 80 ppm) and strontium 1.0, 1.5, 2.0 ppm) along with constant concentrations of scandium ppb) and yttrium ppb). The precise compositions of the calibration standards are summarized in Table 4. The U-Ba spike made at the University of Michigan was added to the coral samples to measure U and Ba by isotope dilution. The nominal 15

PAGE 26

Table 4 Calibration Standard s Composition Composition: Standard #1 Standard #2 Calcium (g) 1 .0337 (:::20ppm) 2.0599 Strontium (g) 0 .359 (:::0.75ppm) 0.4715 (:::1.0pprn) Scandium (g) 0 .0494 (:::100ppb) 0 .0491 (:::100ppb) Yttrium (g) 0.0493 (;:::100ppb) 0 .0471 (:::100ppb) Total Volume (ml) 500 500 Standard #3 Standard#4 Calcium (g) 3 .0901 (:::60ppm) 4.1174 (:::80pprn) Strontium (g) 0 .7769 (""1.5ppm) 0 9974 (:::2.0pprn) Scandium (g) 0.0489 (:::100ppb) 0 0485 (:::100ppb) Yttrium (g) 0 .4880 (:::100ppb) 0 0488 (=100ppb) Total Volume (ml) 500 500 concentration of the spike was given to be 0 662 ppb uranium and 3 83 ppb barium. The isotopic ratio of 238U j235U was also measured to be 0 0239 and the 138Baj135Ba ratio was measured to be 0.0380. To separately confirm the nominal concentration of the University of Michigan spike, a U-Ba standard was gravimetrically made. Barium and uranium (SPEX plasma monoelemental @1,000 ppm) stock solutions (0.4885 g and 0.4907 g, respectively) were added to 500 ml of 1 % nitric acid The results of the spike measurement corrected the uranium concentration to 0 848 ppb (28% increase) and the Ba concentration to 4 .18 ppb (9% increase) As a second check, a U 500 standard was prepared at the University of New Mexico using certified reference materials from New Brunswick Laboratory The U-500 standard had a measured U concentration of 105.2 .5 % ppb. This second spike calibration yielded aU concentration of 0 897 ppb, only 6 % greater than the previous correction (Appendix D). 16

PAGE 27

Also added to the coral samples was a solution consisting of approximately 100 ppb scandium and yttrium. Actual concentrations, determined gravimetrically, were normalized to 100 ppb. This solution was added to each coral sample as an internal standard. Three separate Sc-Y solution standards were made throughout this study using the same stock bottles of Y and Sc Spex standards each time Each solution was prepared identically. The Sc andY concentrations of each solution were used in the isotopic ratio determinations Yttrium and scandium SPEX solutions were added to one lite r of 1% nitric acid. To achieve 100 ppb concentr ations, approximately 0.1 g of SPEX standard @1,000 ppm was added. For th e DCP measure ments the same coral standard was analyzed to determine precision and to be compared to ICP-MS coral standard values. In addition, four calibration standards along with a blank solution were made separately from th e ICP-MS calibration standards. The standards contained increasing concentrations of calcium (::::5, 10, 20, 30 ppm), strontium (::::0.25, 0.5, 0.75, 1.0 ppm), and magnesium (::::0.02, 0 .05, 0.1, 0 2 ppm). Ge and Sc were added to the standards at 10 ppm and 100 ppb, respectively, with the intention of using them as internal standards, but DCP results demonstrated that internal standard corrections were unnecessa ry. Sample Preparation Samples were prepared in a clean lab at least 24 hours before analysis to allow for The coral powder in the 1.5 ml tubes was transferred into 15 ml polystyrene centrifuge tubes with plug seal caps The tubes were sterile and required no further cleaning. A few of the initial coral samples analyzed showed evidence that a substantial amount of powder 17

PAGE 28

was left behind. A simple experiment indicated that the powder was sticking to the s ides of the microcentrifuge tube due to humidity and static attraction Samples with increasing coral powder weights were analysed. The results from the experiment (Figure 2) illustrate that for weights below 1.0 mg, U/Ca, Sr/Ca, and Ba/Ca ratios were very different from the consistent values achieved from the samples with 1 0 mg or greater coral powder weight. Sample weights must be at least 1.0 mg for accurate meas urements to be performed on the ICP-MS 8.00......------------------------, .-----------------------------------c: 6.00 .!2 -;;; c: 4.00 Cl .Sl -;;; r::.: 2. 00 .8 c: cv E CJ Ui 0 .00 0 0: ("i()"6 nwlar) ------88Sr/40Ca (10-6 molar) EB--Ba/Ca (IJ.mol/mol) \ \ ffi---ffiEB EB --------ffi ).. . '''ffi : .... .Ell -2.00 0 .00 1 .00 2.00 3.00 4.00 5.00 Powd e r W e i g ht (mg } Figure 2 Sample Size Experiment Samples with increasing coral powder weights from 0 .25 g to 4.2 g were analyzed for U, Sr Ba, and Ca. Consistent values were achieved on samples with 1.0 g or greater of coral powder. Below this threshold, precision falls profoundly. 18

PAGE 29

To transfer all the powder from the microcentrifuge tubes to the 15 ml polystyrene centrifuge tubes, the Sc-Y solution which was previously added to the coral samples was instead first flushed into the 1.5 ml microcentrifuge tubes and then poured into the larger centrifuge tube. Samples 3, 7, 14, 15 had very high analytical uncertainties within the sample analysis were not used in data analysis. Approximately 0.5 g of the U-Ba spike and 10 g of the Sc-Y solution were added to the coral powder. It was very important to record the weights of both the added U-Ba spike and ScY solutions, as these weights were necessary to make accurate isotopic ratio calculations. The spike weight is inserted into the isotope dilution formula, while the Sc,Y solution weight is needed to determine the Sc,Y concentrations in the samples. Total weight of the coral solution aliquots averaged 11 g. Once the samples were prepared, the caps were placed tightly on the tubes and shaken vigorously. The samples were then refrigerated to minimize evaporation. On the day of analysis, the samples were brought to room temperature and again shaken vigorously. On average, the samples were l eft to equilibrate for 48 hours. After ICP-MS analysis, the sample tubes were stored in plastic bags until DCP analysis. Choice of Isotopes Calcium, strontium, uranium, and barium all have multiple isotopes. Abundance and interference factors were considered in determining which isotopes were best to measure. The most abundant isotope of calcium, 40Ca, suffers an isobaric interference with 40Ar and therefore cannot be used with an argon plasma. Also, 40Ca, 44Ca, and 42Ca were too abundant given theCa content of the coral and the ICP-MS detector limitations. Isotopes 43Ca and 19

PAGE 30

4 8Ca were measured because of their low natural abundances (0. 135% and 0.187%, respectively). Isotopes 4 3Ca and 48Ca were both sufficiently low in abundance to make quantitative ICP-MS measurements. The small interference of doubly charged 86Sr on 43Ca was negligible as tested by Le Cornec and Correge (1997). For strontium, 88Sr concentrations were too high (abundance of 82.58%) surpassing the saturation threshold (Le Cornec and Correge, 1997). Therefore, 86Sr was chosen because of its low abundance (9.86%) The most abundant natural U and Ba isotopes, 238u (99 27% abundance) and 138Ba (71.70% abundance), were measured to maximize count rates. Difficulties with ICP-MS Measurement There are many difficulties inherent in measuring U I Ca, SrI Ca, and Ba/Ca ratios with the same instrument (ICP-MS) and on the same aliquots First, there is an extreme variation in mass to charge ratios (m/z) between uranium (238 amu), barium (138 amu), strontium (88 amu), and calcium (43 amu). This large difference compromises the optimization of the instrumental parameters, especially ion lenses (Le Cornec and Correge, 1997) The second difficulty is the extreme variation in elemental concentrations of the coral, especially between 48ca and 238u (748 ppm and 3 ppm, respectively). The calculated nominal diluted concentrations of 43Ca and 48Ca in the coral standard are 65 ppb and 90 ppb, respectively. The ICP-MS can get inundated with calcium and not even detect uranium if this problem is not addressed. Thirdly Sr /Ca ratios must be measured very precisely because the strontium variation with respect to temperature is only 0.9% per C. 20

PAGE 31

Optimization of ICP-MS ICP-MS operating conditions (Table 1) were set to obtain the best sensitivity for U and the best precision for Ca and Sr Two separate methods were used to measure the four isotopic ratios in order to obtain the best results. Ca and Sr were measured by internal standardization and the construction of a standard calibration line from the four standards. U and Ba were measured by classic isotope dilution after the addition of a U-Ba spike. The ICP-MS was allowed to warm up and stabilize for 45 minutes before each day of sample analysis Also the machine was tuned and masscalibrated (additional 15 minutes) using a mass calibration standard consisting of multiple elements of different masses. Blank corrections were made during every analysis. To maximize precision, three corrections were performed: mass bias dead time, and krypton interference Mass bias is a fractionation phenomenon inherent in mass spectrometric data (Russell et al. 1978). Certain masses are preferentially detected by the ICP-MS over other masses. The observations occur systematically. However, possibly due to instrumental drift, the mass preference can reverse. The effect is relatively smaller for heavier elements due to relative changes in fractionation compared to mass. Nevertheless, Ca, Sr, and U had to be corrected for mass bias. The mass bias on Ca was the most severe. In addition to having a light mass, calcium isotopes have a relatively large mass difference between isotopes. For Ca, the 43Caj48Ca ratio was measured to correct for the effect of mass bias on both 43Caj45Sc and 48Caj45Sc ratios. If there were no mass bias, the 43Caj48Ca measured ratio would be equivalent to the 43CaJ 4 8Ca natural abundance ratio Usin g equations derived from Russell et al. (1978), the mass 21

PAGE 32

bias correction for each ratio was determined. The formula below was used for 48Caj45Sc: (48Caj45Sc)MBC = (48Caj45Sc)meas*(mass 48Ca/mass 45Sc)exp (1) where MBC = mass bias corrected and exp = log[(43Caj48Ca)nat/(43Caf48Ca)meas]/log[M43Ca/M(48Ca)] (2) where exp is the exponential factor, (43Caj48Ca) nat is the natural abundance ratio and M is the isotopic mass. The exponential factor normalizes the measured 43Caj48Ca ratio to the known natural 43Caj48Ca abundance ratio The 86Sr j89y ratio was corrected for mass bias using the natural 86Sr j87Sr abundance ratio Mass bias varied greatly from sample to sample with as much as a 15 % correction in the ratios Dead time is the interval after each incoming ion where the pulse counter does not count incoming ions because the previous pulse is still decaying away. This recovery time causes concentrations to be underestimated only when the count rates are too high. Dead time was determined using the formula (Correge, pers. comm.): TCR=MCR / 1-DT*MCR (3) where MCR is the measured count rate (counts per second) TCR is the total count rate (actual counts per second), and DT is the dead time (seconds) For seven samples of increasing Sr concentrations, 86Sr and 88Sr were measured. Because 86Sr is approx imately 8 times more abundant than 88Sr, it 22

PAGE 33

was assumed that the measured 86Sr counts per second (cps) equaled the actual 86Sr cps. The measured count rate of 88Sr was calculated from: MCR(88Sr)=[88Sr j86Sr*MCR(86Sr)] /1 +DT*88Sr j86Sr*MCR(86Sr) (4) where 88Sr j86Sr = 88Sr(abundance)*88Sr(mass) j86Sr(abundance)*86Sr(mass). (5) 88Sr and 86Sr MCR values were plotted against each other (Figure 3). The curve shows the effects of dead time The 88Sr vs. 86Sr (measured) curve flattens out as the 88Sr count rates increase Dead time was determined from equation (4) after plugging in the 88Sr and 86Sr MCR values. Dead time for the detector was calculated to be 5.33 nanoseconds. For the count rates measured in this experiment, a dead time below 20 ns causes no signal deviation. Therefore, dead time was determined not to be a factor. 20 00 --,----------------------r --o-SSSr (10"6) Calculated 15 00 ?n theoretical /-:c < 10.00 0 .. actual <0 5 00 0 5 10 15 20 86Sr 00"5) MCR Figure 3 D ead Time Experiment Dead time is illustrated by the difference between 88sr measured and 88sr theoretical calculation. Notice the difference between the theoretical and actual measurement increases as the 88sr concentration increases 23

PAGE 34

Krypton interference can be a problem due to impurities in the argon gas 86Kr, present in the argon gas, interferes with 86Sr. Unfortunately, it is very difficult to measure krypton because its signal is both unstable and variable. One way to determine if krypton may be a problem is to measure 86Sr background counts in blank solutions. The blanks that were run had very small background counts of 86Sr. Therefore, krypton interference was assumed to be negligible. Optimization of DCP No special adjustments were made to optimize for DCP analysis. The procedures used were those routinely used for multi-element major and trace element determinations Mg Ca, and Sr lines used were those used in a standard 20-channel multi-element cassette and were selected for an optimal combination of sensitivity and minimal spectral interferences at 0.1-100 ppm concentration ranges Peaking signal max imization, was performed using a 25 micron slit and the machine was tuned each day before analysis. For each run, 6 blank solutions and 17 drift monitors were analysed. The drift monitor was the highest concentration calibration standard. A blank correction and drift correction were performed to account for background noise and machine-related signal fluctuation Sources of short-term variation include physical variations in sample injection crosstalk between the analytes and rin se solution, and electrical power supply fluctuations. A long-term variation source was the decay of torch consumables (anodes and cathodes). Blank count variability for Ca and Mg was small with RSD values of 1.6% and 0 7%, respectively; Sr was greater with a RSD value of 2 .9% (Figure 4). The variability of the drift monitor within a run followed the same trends as the 24

PAGE 35

25 blank. RSD values for Ca and Mg were 1.6% and 1.7%, respectively. For Sr, the RSD value was 2.7% but the intensity oscillated along the mean (Figure 5). 14. 00 12.00 ..!a 10.00 ::3 p., 8.00 .9 g 6.00 .=: 4.00 2.00 0.00 0 -----.... ... --.. .. .--" __._ Mg Blank ........ Ca Blank ......... Sr Blank ........................................................ .......... ...... .................... .......... ............. 1 2 3 4 5 6 Blank# 7 Figure 4 Blank Variability During DCP Run Six blank solutions were analysed for an entire 2.5 hour run. The lines are near parallel indicating that the background count rates fluctuated very little. 85.00 80.00 75.00 (I) ... ..!a ::3 70.00 p., 0 :-::2 65.00 c 60.00 .=: 55.00 50.00 45.00 0 .+. ..... --.. : ..... ... . .. ......... ---MgDrift Ca Drift .... .... Sr Drift ---.....,__ A-t--4Y" ...... r ....... -:t-:-.-. .. .... . -........ -.. -:r ... 5 10 15 Drift Monitor # 20 Figure 5 Drift Variability During DCP Run (Blank Corrected) Seventeen drift monitor solutions were analysed for an entire 2.5 hour run. Mg and Ca fluctuation is significantly less than for Sr. However, for every element, the trends match perfectly.

PAGE 36

V. ICP-MS MEASUREMENTS Sr /Ca Measurement Sr and Ca were measured by an internal standardization technique developed by Le Cornec and Correge (1997) The technique involved the addition of 100 ppb Sc andY to all calibration standards and coral samples. This value was chosen because it yields target ratios for 43Caj45Sc, 48Caj45Sc, and 86Sr j89y close to unity (Le Cornec and Correge, 1997). These ratios were corrected for mass bias by measuring 43Caj48Ca and 86Sr j87Sr and normalized to the nominal internal standard concen trations of 100 ppb Sc andY. These elements were chosen as inte rnal standards because : 1) they are monoisotopic elements, 2) Sc andY are not found in significant concentrations in corals, 3) the Ca-Se and Sr-Y pairs are very similar in mass, first ionization energy, and show a similar behavior with time in the plasma (Le Cornec and Correge, 1997). Determination of Ca and Sr concentrations in the coral samples was performed by the calculation of calibra tion lines relating concentrations to 43Caj45Sc and 48Caj45Sc ratios for Ca and 86Sr j89y ratios for Sr. Day to day reproducibility of the calibration lines was very good (Figures 6, 7, 8). R values for Ca averaged 0 9998 and never fell below 0.9997. R values for Sr averaged 0.9997 and never fell below 0.9996. The Sr and Ca concentrations were derived from the equation of the regression lines (Figures 6 7, 8). 26

PAGE 37

u (f) ,_, ..,. .......... co u ...., .... 1.40 --.-------------------------, 1.20 1.00 0.80 0.60 0.40 0.20 -:-.J--4 3Ca Day 1 -43Ca Day 2 -o43Ca Day 3 +Ca Day4 --y = -0.00045062 + 0.016083x R= 0.99998 --y = -0.00046666 + 0.01624x R= 0.99994 -y = 0.0080957 + 0.01668x R= 0.99999 --y = -0.0029673 + 0.016868x R= 0.99977 0. 00 -t---.-----.----.---.---,-r---.---,--,---,------.----,---,----,-r---.---.--.----,-----1 0 20 40 60 80 Ca Cone. (ppm) Figure 6 43ca Standard Calibration Curves Four calibration curves are plotted representing day to day reproducibility. Variability among the four calibration lines is minimal. The equation and r values of each calibration line are given. 100 N 'l

PAGE 38

u (f) l/) ... ......... ell u co .... 2.50 -,--------------------------, D48Ca Day 1 -----48Ca Day 2 2.00 -1 I -o48Ca Day 3 ----+-48Ca Day 4 1.50 1.00 --y = -4.1709e-05 + 0.024854x R= 0.99998 0 50 --y = -0.0010185 + 0.025114x R= 0.99994 --y = 0.012423 + 0.0258x R= 0.99999 y = -0.0046962 + 0.026088x R= 0.99976 0.00 -+---y--.,---,---,---,-.--r--.----.---,----.---,---,----y-,-,----.---,--,.---i 0 20 40 60 Ca Cone. (ppm) Figure 7 48ca Standard Calibration Curves Four calibration curves are plotted r e pres enting day to day reproducibility. Variability among the four calibration curves is minimal. The equation and r values of each calibration curve are given. 80 100 N CX>

PAGE 39

2.20 2.00 1.80 1.6 0 >-< 0. 1 .40 .Jj) CQ 1.20 1.00 0.80 0.60 0.6 ... .... o ....... 86Sr Dav 1 86S r Day 2 -o-86Sr Day 3 -+-86Sr Day 4 0.8 1 -y = 0.00049192 + 1.0604x R= 0 99996 --y = -0.0066568 + 1.07 17x R= 0.99989 --y = 0.012479 + 1.0336x R= 0.99994 --y = 0.028239 + 1.03x R= 0.99976 1.2 1.4 1.6 1.8 Sr Cone. (ppm) F i gure 8 86sr Standard Calibration Curves Four calibr atio n lin es r epresenting day to day reproducibil i ty show very littl e va ri ability The equatio n and r values o f eac h cal ibration curve are give n 2 N \0

PAGE 40

88Srj40Ca molar ratios were then calculated using the known isotopic abundances of 88Sr and 40Ca. The analysis protocol along with the aliquot contents is described in Table 5. First, the four calibration standards and the blank were analyzed. Coral samples were analyzed in groups of five, with the coral standard, U-500 standard, and blank interspersed. After every tenth coral sample, a new calibration line was made. New calibration curves were made because many parameters, such as configuration and residue build-up on the cones, properties of the lenses and plasma conditions, can alter the sensitivity and the effect of interfering species. The maximum ouput of this method was eight (8) Sr /Ca coral sample determinations per hour. U/Ca and Ba/Ca Measurements Uranium and barium were measured using classic isotope dilution. Isotope dilution is a method based on the determination of the isotopic composition in a mixture of a known quantity of spike with an unknown quantity of the actual sample. A 235U_135Ba spike was added to the coral samples during preparation. The ICP-MS measured 238U j235U and 138Baj135Ba ratios. The isotopic ratios were then multiplied by the isotopic masses to get an abundance ratio. The abundance ratio was then applied to the isotope dilution formula (Faure, 1986): Nw = Sw*Wn/Ws(Abas-Rrn*Abbs/Rrn*Abbn-Aban) (5) where Nw =weight of the unknown element, Ws = atomic weight of spike, Sw = weight of the spike in the aliquot, Wn=atornic weight of element, 30

PAGE 41

Table 5 ICP-MS Analysis Protocol for Ca, Sr, U, Ba A. Calcium, Strontium Run (20 Samples) L abel Blank Ca 20 standard Ca 40 standard Ca 60 standard Ca 80 standard Coral Samples rinse Coral Standard Sequence of Analyses Blank Ca 20 standard Ca 40 standard Ca 60 standard Ca 80 standard rin se Coral Standard Coral Samples #1-5 Coral Standard Coral Samples #6-10 Aliquot Contents 1.0% Nitric Blank 20ppmCa, 0.75ppmSr 100ppbSc, 100ppbY 40ppmCa, l.OppmSr 100ppbSc, 100ppbY 60ppmCa, 1.5ppmSr, 100ppbSc 100ppbY 80ppmCa, 2.0ppmSr, 100ppbSc, 100ppbY coral powder (12mg) 14ml Sc-Y spikeppb, O.Sg U-Ba spike 1.0 % Nitric Blank coral powder, 100 ppbSc andY, U-Ba spike Blank Ca 20 standard Ca 40 s tandard Ca 60 standard Ca 80 sta ndard rinse Coral Standard Cora l Samp l es #11-15 Coral Standard Coral Samples #16-20 B. Uranium, Barium Run (20 samples) Label Blank U standard Coral S t andard Coral Samples Sequence of Analyses Blank U 500 standard Coral Standard Coral Samples #15 U 500 s tandard Coml Samples #61 0 Aliquot Contents 1.0 % Nitric Blank U-500 cora l powder, 100 ppbSc andY, U-Ba spike coral powder (1-2mg), 14ml Y,Sc spike@100ppb, O.Sg U-Ba s pike U -500 sta ndard Coral S t andard Cora l Samp l es #11-15 U -500 standard Coral Samples #16-20 31

PAGE 42

32 Abas =abundance of isotope A (235U and 135Ba, respectively) in the spike, Aban =abundance of isotope A in nature, Abbs =abundance of isotope B (238U and 138Ba, respectively) in the spike, and Abbn = abundance of isotope B in nature. For uranium, the 238U j40Ca (molar) ratio was calculated because this is the form used in the U/Ca-SST equation (Appendix B). The isotopic ratio 238Uj40Ca was multiplied by the 238Uj40Ca abundances ratio and the elemental mass U /Ca ratio. For barium, the final form was Ba/Ca (molar) (see Tudhope,1994; and Lea and Martin, 1996).

PAGE 43

VI. DCP MEASUREMENTS: MG/CA AND SR/CA Mg/Ca and Sr /Ca elemental ratios were determined using a five point calibration line for each element. Four gravimetric calibration solutions of increasing Mg, Sr, and Ca concentrations were made plus an acid blank. Mg, Sr, and Ca concentrations of the calibration standards were determined from the linear dynamic range (LDR) and detection limit for each element. For Mg, the LDR is 0.002-60 ppm, with a detection limit of 0.2 ppb. The lowest calibration standard has a concentration of 0.02 ppm, within the LDR by only a factor of ten. Sr concentrations were set for the middle of the LDR (0.003-10 ppm) at a concentration of 0.25 ppm. The highest concentrated calibration standard had a Ca concentration of 30 ppm, within the LDR (0. 009-100 ppm). Calibration curves for Ca and Mg indicate no matrix interference effects The determinations of Ca, Mg, and Sr concentrations in the coral samples were calculated from calibration lines linking concentrations to blank and drift corrected intensity (Appendix C) Figures 9, 10, and 11 are a test of the calibration standard solutions run in duplicate without corrections. An equation for each regression line and r values of each calibration line are shown next to each line. The r-values for Ca and Sr were 0.9995 and 0 9997, respectively The Mg calibration line had the lowest r-value with 0.9953. The Mg calibration was performed a second time and a r-value of 0.9990 was achieved which is comparable to both Ca and Sr calibration lines. 33

PAGE 44

,.-.. (J) Q) .$ ::l 0.. ......., .c ..... (J) Q) ...... ........ 120 e 100 80 60 y = 7.3028 + 11.172x R = 0.99953 0 I I I I I I I I I I I I I I I I I I I I I I I I I 0 2 4 6 8 Ca Cone. (ppm) Figure 9 Ca Standard Calibration Curves The ca lcium standard calibration curve is shown. Intensity (count rate) was plotted against the four standard concentrations of calci urn from 1 ppm to 10 ppm. The equation is located in the upper left of the graph and the R value was 0.99953 10 12

PAGE 45

--.. Vl Q) 0... ..__. b 'Vi s::: Q) ..... s::: ....... 35 30 25 20 15 10 5 y = 18.852 + 496.32x R = 0.99551 0 0 0.005 0.01 0.015 0.02 Mg Cone. (ppm) Figure 10 Mg Standard Calibration Curves The magnesium standard calibration is shown. Intensity (count rate) was plotte d against four standard concentrations of magnesium from 0.002 ppm to 0.025 ppm. 0.025 0.03 UJ 01

PAGE 46

140 120 100
PAGE 47

Typically, to obtain the highest precision on DCP measurements, an internal standard correction (similar to the ICP MS procedure) is performed to remove the effects of physical sample input variations While Sc and Ge were added to the coral samples to serve as internal standards, the only necessary corrections were for the blank and to account for drift (Figures 4, 5). The analysis protocol and the aliquot contents of the DCP runs are described in Table 6. Coral samples were diluted 2 : 1 with a Sc-Y spike solution One reason for this dilution was to avoid saturating the Ca signal. In a typical analytical run on the USF DCP machine, 12 coral unknown samples and four calibration standards were analysed twice along with 6 blank analyses, at 7-10 cup intervals and 17 drift monitor analyses at 4 cup intervals All data are collected as raw intensities with blank corrections drift corrections, and calibration to standards performed offline as a spreadsheet operation. One complete analytical run time is two hours. The analy s i s protocol was designed to conduct high-precision multielement measurements on silicate samples Given the behavioral similarities of the Ca Mg, Sr lines and the simple solution matrices of coral samples, some streamlining of the analysis protocol to increase sample throughput is possible Specifically a reduction of the number of drift monitor analyses will increase productivity Drift variability is minimal and fluctuates along the mean value The number of drift monitor solutions to b e analysed for e ach run can be reduced from 17 to 4 This will reduce the time it takes to analyse a complete run from 2 hours to 1.5 hours. 37

PAGE 48

Table 6 DCP Analysis Protocol for Ca, Mg, Sr Analysis consists of 12 coral samples per run. Label blank drift monitor Standard 1 (low) Standard 2 (medium) Standard 3 (high) Standard 4 (very high) Samples Coral Standard Sequence of Analyses Blank Drift Monitor Sample 1 Sample 2 Drift Monitor Sample 3 Standard 1 (low) Sample 4 Drift Monitor Blank Drift Monitor Standard 2 (medium) Samp l e 5 Samp l e 6 Drift Monitor Standard 3 (high) Sample 7 Sample 8 Drift Monitor Standard 4 (very high) Sample 9 Sample 10 Drift Monitor Blank Drift Monitor Sample 11 Sample 12 Drift Monitor Aliquot Contents 1 Oppm Ge, 1 OOppbSc 1 Oppm Ge, 1 OOppb Sc, 0.2ppm Mg, 30ppm Ca, 1.0ppm Sr 1 Oppm Ge, 1 OOppb Sc, 0.02ppm Mg, 5ppm Ca, 0.25ppm Sr 1 Oppm Ge, 1 OOppb Sc, 0.05ppm Mg, 1 Oppm Ca, 0.5ppm Sr 10ppm Ge, 100ppb Sc, 0 1ppm Mg, 20ppm Ca, 0.75ppm Sr 1 Oppm Ge, 1 OOppb Sc, 0.2ppm Mg, 30ppm Ca, 1.0ppm Sr 2.5ml coral sample, 2.5ml Sc,Ge spike Coral powder, 100ppb Sc andY, U-Sa spike Sample 1 Standard 1 (low) Sample 2 Drift Monitor Blank Drift Monitor Standard 2 (medium) Samp l e 3 Samp l e 4 Drift Monitor Standard 3 (high) Sample 5 Sample 6 Drift Monitor Standard 4 (very high) Sample 7 Sample 8 Drift Monitor Blank Drift Monitor Sample 9 Sample 1 0 Drift m onitor Sample 11 Sample 12 Coral Standard Drift Monitor Blank 38

PAGE 49

VII. RESULTS AND DISCUSSION Precision of Measureme nts The repeatibility (short-term) and reproducibility (long-term) of the ICP-MS values were determined by analyzing coral standards throughout every run. Both short-term and long term precision are necessary because : 1) Sr /Ca variation in samples is particularly small 2) ICP-MS runs often take a few hours to be completed, and 3) century-scale coral records represent thousands of coral samples requiring many days of analysis Table 7 shows the re s ults of the coral standard analyses Three coral standards were analyzed during each run. Table 7 data represents the coral standard (labeled NC-5/1/97) tested after the introduction of the U-500 The previous coral standard before the i ntroduction of the U-500 produced higher RSD values. Improvement in the precision was achieved because 1) the second coral standard con s isted of twice the U-Ba spike concentration to eliminate underspiking and 2) the analytical procedure was changed to measure Ca and U separately instead of measuring Ca and U simultaneously. This allowed for th e ICP-MS to be tune d to each elemental mass range instead of a mass betwee n Ca and U These changes to the second coral standard increased th e Ca precision (43 Ca and 48Ca RSD values of 0.23 % and 0.26%, respectively) to levels comparable to those reached by Le Cornec and Correge (1997) (0.21% and 0 .25% respectively). 238U j40Ca RSD value was still too high at 3.07%. This was not comparable toLe Corn ec and Correge (1997) 238Uj40Ca 39

PAGE 50

Table 7 Reproducibility of New Caledonia Coral Standard on the ICP-MS Coral Standard NC-5/1/97 43Ca (ppm) 48Ca (ppm) Ca ave. (ppm) U (ppm) 238U I 40Ca (*1 0"-6) 8/5/97 (run#1) 46.000 46.013 46.006 2.49 1.072 46 .009 46.018 46.013 2 .51 1.081 46.060 46.079 46.069 2.42 1.043 8/8/97 (run#2) 46 118 46.134 46.126 2.46 1.057 45.863 45.873 45.868 2.46 1.058 45.853 45.843 45.848 2.54 1.095 8/8/97 (run#3) 45.802 45.768 45.785 2.54 1.093 45.880 45.877 45.879 2.38 1.024 45 .903 45.892 45.897 2.32 0.997 MEAN 45.943 45.944 45.944 2.46 1.058 SD 0.107 0.121 0.114 0.075 0.032 RSD (%) 0.233 0.264 0.249 3.040 3.067 w:.. 0

PAGE 51

RSD value of 0.43%. 88Sr j40Ca RSD value (1.31 %) was lower than U/Ca but still higher than LeCornec and Correge (1997) values (0.2 8%) A summary of the precisions ac hieved by this study, Le Cornec and Correge (1997), and Min et al. (1995) are listed in Table 8. The precision of the second coral standard by DCP was determined from two runs during the same day (Table 9). For more significant results, further coral standard analysis must be performed However, the results appear very promising. Sr /Ca and Mg/Ca elemental ratios were determined and the respective RSD values were 0 .8% and 1.0%. These values were lower than the ICP-MS results and more comparable to other results (summarized in Table 8). Isotope Ratio Curves Samples B-1 to B-174 (120 samples), representing years 1967 to 1981, were plotted against their measured 238U j40 Ca ratio to produce a 14-year continuous curve (Figure 12). The 238U j40Ca ICP-MS values averaged 1 .17 x1Q6 molar. A comparison of the 14yea r U/Ca curve with instrumental sea surface salinity (SSS) (Figure 13) and sea surface temperature (SST) (Figure 14) curves from ORSTOM illustrate that a corresponding shift occurred in both the SSS record and the U /Ca record When a five-point smoothing function is applied to both U/Ca and SSS, th e curves are strikingly similar (Figure 15). U / Ca ratios thought to be strictly SST dependent, appear to b e also infl u e nc e d by SSS. The shift was the result of a slight cooling trend in the region as illustrat e d in th e SST record A d e crease in SST caused a decreased in r a infall. This directly led to an increase in SSS. The climatic shift occurred around 1976. The U/Ca curve s hifts upward in 1976 and 41

PAGE 52

Table 8 Comparison of Instrumental Results Average Elemental Ratios* Ave 238U/40Ca (lOA-6 molar)* Ave. 88Sr/40Ca (lOA-3 molar)* Ave. 138Ba/135Ba (10A-6 molar)* Mg/Ca (lOA-3 molar)* ICP-MS DCP ICP-MS Current Study Current Study Le Cornec and Correge (1997) 1.17 7.76 2.50 7.8 5.80 *Average elemental ratios were calculated from actual cora l samples (not standards) Relative Standard Deviation Values from Coral Standards 43CaRSD 0 .23% 0.21% 48CaRSD 0.26% 0 25 % U/CaRSD 3 .07% 0.43% Sr/Ca RSD 1.31% 0.80% 0.28% Mg/CaRSD 1.00% Ba/Ca RSD 2.10% **Lea and Boyle determined Ca concentrations by flame atomic absorption ICP-MS TIMS Lea and Min et al. (1995) Boyle (1992) 2 .0-3.0% 1.24 7 77 0 .20% 0.10% N

PAGE 53

Table 9 Reproducibility of New Caledonia Coral Standard on the DCP Coral Standard: NC-5 / 1 / 97 Date 88Sr I 40Ca*10"-3(mol) Mg/Ca*10"-3 (mol) 10/11/97 (run#1) 7.646 4.682 7.660 4.706 7 813 4 673 7 703 4.646 7.656 4.712 7.688 4.791 10/11/97 (run#2) 7.634 4.670 MEAN 7.686 4.697 SD 0.061 0.047 RSD 0.795 1.001 VJ

PAGE 54

C1l u 0 ..,. ......... :::::> co "' N 1.50 f .1. p;, j 1,! t. Jli l"i' T A 1 1 1 I j'\ i I T f I lr h /I .n r d i o I t r ,. I I I tl t'\ 4 ,j ! '! l j I t j'il'oi. f I I l r Ji.J I t 'I' T i u t ll ,!l fl11; \ J. 111 r 1 V A ;, '! l. ; :, I ll 1.40 1 .30-1.201.101.00-0 9 0 -l-r--r-,.--.-r---.-......-r--,---,--,r-T"'--r--r-T'-,.-,.--,--,---,---r---r---r---r-r---r-1980 1978 1976 1974 1972 1970 1968 Year Figure 12 U/Ca Ratios Measured on ICP-MS The U I Ca results are plotted against time of coral growth as determined by o18o data (Quinn et al., 1996) A dramatic shift to increasing values occurred during 1976 and peaked in 1979 t

PAGE 55

tf) tf) tf) :2: 0 p:: 0 36.50--.---------------------, 1 +I 36.00 -j .,. 11 z -''l 'I ".l; I le -1 35.50I 'i \ 'jl ,: A r + , j l -.: I A }J! lr I i i 1 I \I li I 35.00 l I I I I I I I I I I I I I I I I I I I I I I I I I I I I 1980 1978 1976 1974 1972 1970 1968 Year Figure 13 Sea Surface Salinity Instrumental Record The instrumental SSS record from New Caledonia is plotted against the time The measurements were taken by the French research group ORSTOM. The SSS record shows oscillations with a minimum during the summer of 1976.

PAGE 56

0 '-" lJ) {j) lJ) 0 30.00----.-----------------------, 28.00-r $\ i' j l 26.oo I t fl r t f oe f1 1 \ JJ. \ I I '! '!"" \ i i. l I I \!1 I I 24 oo j \ 1! 1 1 r 1 I!',. f 1 I . 1 I I I I I I : I I ,; I I \ I j 11/ 22 00-1 \J t ll \it -1 1 1.! A i 20.00 - ;\ H I! t I i 1 rt fl I i : l : ,1. I fi.. I 1 I i 1 I \J \ : I 1 f \ . f \ . \ i 'II ll i I I fl. I ... I 11 0 t i \1" t \J. ! y. I l 18.00 1980 1978 1976 1974 1972 1970 1968 Year Figure 14 Sea Surfa ce Temperature Instrumental Record The instrumental SST record from New Caledonia is plotted against time ORSTOM performed the measurements. The SST record exhibits a noticeable cooling trend starting in 1976. Historically, this was a time of marked cooling globally. ,p. 0'\

PAGE 57

::0 ..c 0 0 8 (J) (J) (J) ::2 0 0 36.4 0 1. 50 238U I 40Ca (smoothed) 36.2 0 -4. ORS TO:M SSS (smoothed) 1.40 36.0 0 I 1.30 I I Ja 35.80 I I I I I I I I I II Ill' u .... \': 1.20 35.60 I I I l1 J1 1 I I ,I I 1 I 11 1 I ltf I I I I I I I I I II '-I II I 35.40 I I 11 I I I r 1. 1 0 II I I I 35 .20 -j .. II 1.00 1980 1978 1976 1974 1972 1970 1968 1966 Year Figure 15 U / Ca vs. ORSTOM SSS (Smoothed Data) 5 -point smoothing curves were made from the U /Ca and ORSTOM SSS data The curves both shift to increased values from 1976 to 1980 and level off at1981. 00 c::: ........ 0 ....... "' 9 0 0 .... ::r-'J

PAGE 58

correlates very well with the SSS curve. Absolute dates were determined using o18o data (Quinn et al., 1996) from the same coral. RSD values from the coral standard were used for the error in the U I Ca 14-year curve (Figure 16) The 88Sr j40Ca curve (Figure 17) represents only seven years of coral growth from B-1 to B-90. Annual cycles reflecting SST were faintly produced and the measurement was stopped at B-90. At the time of the study, it was decided that 88Sr j40Ca ratios could not be determined precisely enough to warrant further study. The sensitivity of the Sr /Ca-SST relation (0.9%/C) apparently is too low to be resolved by the Fisons ICP-MS. However, the average 88Sr j40Ca values are very close to average values achieved by Min et al. (1995) from the same coral species and location using TIMS (Table 8). Similar to the 88Sr j40Ca curve, the Ba/Ca curve (Figure 18) seven years of coral growth from B-1 to B-90. Fluctuations in the Ba/Ca record appear to have been driven by SST variation in the surface waters. The record does not contain any evidence of upwelling, unlike Ba/Ca curves for corals in areas of known upwelling. For example, a coral off the coast of Oman displayed annual cycles of Ba/Ca as colder deeper water upwelled during the summer seasons (Tudhope, 1994) Ba/Ca (molar) ratios never reached values above 4.0 g Upwelling events recorded in Oman (Tudhope, 1994) and the Galapagos Islands (Shen et al., 1992) measured Ba/Ca (molar) ratios above 5 0 11g/ g and as great as 10. 0 g. DCP measurements of Sr /Ca and Mg/Ca were plotted against each other along with their repre senta tiv e errors (Figure 19). Although th e curves only represent one annual cycle (12 samples), the curves closely mimicked each other. The 88Sr j40Ca results from both the ICP-MS and DCP were very 48

PAGE 59

0 r;: .... .... r;: 0 E < 0 c r;: u 0 """ ....._ :::> 00 <') N 1.60 1.50 1.40 1.30 1.20 1.10 1.00 0 90 0 co 0\ ...... I 0\ ...... 0\ ....... 0\ ...... i l T T T' . 0\ ...... 0 ('.. 0\ ...... co \{) 0\ ...... ,. 111Wr\ --R-N' i i '-'-illl t\ -I ..t .. I ':" ._.,. .,. f T .. 1 l Ji ii . JLj Ji, l Lt. ,.,,_ 1 r-.. .Jf ... l l .l t l -1-"1 . 1 20 40 60 80 100 120 140 160 Sample# Figure 16 U/Ca Sampling Curve The 14-year U/Ca ratio curve was produced by measuring 174 samples via ICP-MS Error was taken from the RSD value of the coral standard U/Ca ratios The error value was +/-3.07%. .,t:.. \0

PAGE 60

.... 0 E ...... C<) I < 0 c c;: u 0 "
PAGE 61

.... C1l 0 E (0 I < 0 .... ......, C1l u --...... C1l j:Q 0 co 0\ ........ 0\ ........ 0\ ........ 0\ ........ 5.00 4.00 3 .00 2.00 I -I 0.00 I I I I I I I I I I I I I I I IiI I I I Iii I I I I i I I I I I I I I I I I I I I 1 10 20 30 40 50 60 70 80 90 Sample# Figure 18 Ba/Ca Sampling Curve The Ba/Ca sampling curve represents seven years of coral growth (Samples B 1 to B 90). Ba/Ca ratios never exceed values above 4 0 (lo-6) molar unlike Ba/Ca curves for corals in areas of upwelling. Error .1 taken from the RSD value of the coral standard Ba/ Ca ratios. Dramatic increases at samples 40 and 70 are most likely explained by mass bias, machine drift and/ or deterioration of sampling cones 01 .-4

PAGE 62

6.20 6.10 -------Sr /Ca (x10A-3 molar) m-:tvfg/Ca (xJOA3 molar) '!) 6.00 1'0 0 E 7 5.90 . . . 0 ,...... >< ......... u 5.80 . -0.0 5.70 5.60 5.50 0 2 .1 4 6 Sample# Figure 19 DCP 88Sr/40Ca and Mg/Ca Ratios . . . . . . . 8 10 Sr/Ca and Mg/Ca curves from 12 samples (B-41 to B-52) Both curves represent a one-year annual cycle and have representative error bars taken from the RSD values of the coral standard. Mg/Ca error was +/-1.0%. Sr/Ca error was +I -0.8%. 12 7.60 7.65 7.70 00 00 (./) '"1 7 75 n Ill 7 80 E 0 > w 7.85 3 7.90 7.95 8.00 0 Si .::!, (Jl N

PAGE 63

comparable (Table 8). Future work will focus on extending these curves as more samples are analyzed. Instrumental Comparison Sr /Ca ratios for samples B-41 to B-52 were determined on both the DCP and ICP-MS machines so that a direct comparison of the curves could be made along with the representative errors (Figure 20). The DCP measurement produced a more defined annual cycle and preliminary results favored the DCP over the ICP-MS for the measurement of Sr /Ca ratios. Caution should be noted that the comparison was for only 12 samples and for Sr /Ca ratios only. However, the results demonstrated at the least the possibility that Sr /Ca could be measured by less expensive DCP instrumentation. This study documents the need for an interlaboratory coral standard. An interlaboratory coral standard will allow for direct instrumental comparisons and elemental ratio calibrations. DCP and ICP-MS comparisons and multiple ICP-MS instrument comparisons are difficult without a well characterized coral standard. Such a standard would have its elemental concentrations independently measured by muliple laboratories and multiple instruments in an analagous fashion to other well charcterized geochemical standards (e. g PDB for oxygen isotopes and SRM-987 for strontium isotopes). Sample Consumption and Output Sample consumption sample preparation, and sample output have been very important in this study due to the nature of coral samples and the 53

PAGE 64

r.J) :2 I p.. u --1-o m 0 E ('I") I < 0 ...... m u ........ 1-o r.J) 8.00 7 95 r T 1. .. I /l'l 7 90 1 1 7.85 .. .. I I 7 80 !../ ...... I 7 75 Sr /Ca (xl0/\-3 molar) DCP 7 70 Sr /Ca (x10"-3 molar) ICP-fv1S I 7.65 2 4 6 8 10 Sample# Figure 20 Sr/Ca Ratio Comparison Sampl es B-41 to B-52 were analysed for Sr / Ca on both the ICP MS and the DCP. The DCP curve is much more representative of an annual cycle as the ICP-MS cur ve The error bars for each S r/Ca curve were added. The DCP Sr/Ca error was +/-0.8% whil e the ICP-MS error was greater at+ I -1.3%. Caution should be noted as ICP-MS precision is based on long term analysis, while DCP precision is based on the results of one complete run. 8.00 _j 7.95 T I 7.90 (.f) ""t ........ I () 7.85 :l> >< ...... 0 > 7.80 I (;..) 3 0 7.75 iii ""t ....... Cl Q 7.70 T --\1 7.65 7 .60 12

PAGE 65

use of corals as a climate proxy Sample consumption was set at a 1 mg minimum from the sample size e xperiment (see Figure 2) This compared favorably w ith TIMS sample consumption of at least 2 mg for U determinations. Sample throughput must be on the order of 12 samples per hour to warrant the cost and time to produce century-scale climate records. Two factors have determined sample output: sample analysis and machine time Samples must be prepared quickly and the analysis time must be well under 5 minutes per aliquot. Le Cornec and Correge (1997) were able to analyze 50-80 coral s a mples daily for Sr Ca, and U In this study, the productivity was 30 coral s a mples dail y for Sr Ca, and U There were many reasons why this study was unable to reproduce the results achieved by Le Cornec and Correge (1997). First although both studies measured coral samples by ICP-MS, the actual machine s were very different The ICP-MS in this study did not have a computer-controlled peristaltic pump to minimize rinse times Further, the tubing length was much longer in the USF lab than the ORSTOM lab requiring more rinse time. Secondly machine tune-up time was longer in this study. Correge (pers comm ) averaged only 20 minutes of warm-up time where as this study averaged 45 minutes to an hour. Thirdly, calibration standards at USF were run after every 10 coral samples while Le Cornec and Correge (1997) only ran calibration standards once at the beginning of each day. The remaining lost time was caused by different acquistion parameters (see Table 1). All these factors have added up to make productivity in this study 50% of that achieved by L e Cornec and Correge (1997). Some of the factors can be adjusted (analysis t i me and analysis protocol) while the others are specific to the machine (tune-up time and pump configuration). 55

PAGE 66

Comparison of Elemental Ratio Utility From this study, it appears that Sr /Ca ratios are too difficult to determine via a Fisons PQS ICP-MS with the precision required. Coral paleoclimate studies require RSD values to be <1.0%. Sr /Ca RSD values from this study were 1.3%. Sr /Ca ratios have a sensitivity of 0.9% per C RSD values above 1.0% for Sr/Ca result in poor signal to noise ratios. Although U/Ca ratio RSD values were .0%, the high sensitivity of U/Ca (5.0% per C) result in acceptable signal to noise ratios. This study has illustrated that U /Ca ratios are not pure paleothermometers and SSS is a significant competing factor Ba/Ca ratios displayed no discernible trend consistent with upwelling events RSD values were comparable to values (.0%) achieved by Lea and Boyle (1992). Initial DCP results have shown that both Sr /Ca and Mg/Ca measure ments can be determined precisely and rapidly. This study illustrates the utility of measuring multiple geochemical tracers in discerning competing environmental factors. Sr /Ca, U /Ca, and Mg/Ca all have been shown to be correlated to SST. However, competing factors (such as SSS for U/Ca) decouple these relationships. By measuring multiple elemental ratios, competing factors can be isolated and a more accurate climate record can be extracted from corals. 56

PAGE 67

VIII. CONCLUSIONS 1) This study attempted to reproduce a technique developed at ORSTOM de Bondy on a Varian Ultramass ICP-MS and published by Le Cornec and Correge (1997) using a Fisons PQS ICP-MS instrument at USF-Marine Science Calcium data were produced with comparatable precision (RSD=0.23%) at the two facilities. Uranium and strontium data generated at USF, however, did not yield the precision (RSD values of 3.07% and 1.31 %, respectively) achieved in previous studies. 2) SrI Ca and Mg I Ca ratios were also determined by DCP Initial results proved that a rapid, precise, cost-efficient method for determining elemental ratios can be developed with the addition of an autosampler and a reduction in the number of drift monitor solutions analysed from 17 to 4 in the DCP protocol 3) MgiCa ratios have the potential to be an excellent SST proxy from the small amount of data collected using DCP instrumentation Further study is required, specifically an analysis of long-term reproducibility of the results achieved in this study. 4) U I Ca ratios appear to be correlated with SSS, not solely with SST as has been prev iously thought. A profound shift in both U ICa ratios and SSS instrumental values in 1976 illustrate the strong correlation between these two parameters. 5) The sensitivity of th e Sr I Ca-SST rel a tion (0.9%IC) was too low to be r esolved by the Fisons PQS ICP-MS 57

PAGE 68

6) Reproducing the results of Le Cornec and Correge (1997) on the USF ICP-MS was difficult due to differences in the specific attributes of the instrumentation (e g para meters at USF were manually controlled while parameters at ORSTOM were computer controlled, different operating conditions, and different hardware, specifically the electron multiplier, spray chamber, and nebulizer) 7) ICP-MS techniques are machine dependent and differences will occur in the reproduction of results. This limitation must be recognized and alterations of the analytical procedure must be made according to the specifications of each machine. 8) This study documents the differences between experimental results achieved in this study compared to other studies. There is a need for an interlaboratory coral standard that will allow for direct instrumental comparisons and elemental ratio calibrations 58

PAGE 69

REFERENCES Beck, J.W., Edwards, R.L., Ito, E., Taylor, F.W Recy, J., Rougerie, F., Joannot, P and Henin, C. (1992) Sea-surface temperature from coral skeletal strontium/ calcium ratios. Science 257: p. 644-647 Broecker, W.S, and Denton, G.H. (1989) The role of ocean-atmosphere reorganizations in glacial cycles. Geochim. Cosmochim. Acta 53: p. 2465-2501. Cole, J.E. (1994) R econstructing and predicting climate variability, Geotimes. 39: p. 12-15 Dunbar, R.B., and Cole, J E (1993) Coral R ecords of Ocean-Atmosphere Variability: Special Report No. 10, NOAA Climate and Global Change Program, UCAR, Boulder, CO, 38pp. Faure, G., (1986) Principles of Isotope Geology, 2nd ed. Wiley, New York, 589 pp. Hart, S.R., and Cohen, A.L. (1996) An ion probe study of annual cycles of Sr /Ca and other trace elements in corals Geochim. Cosmochim. Acta 60: p. 3075-3084. Houck, J.E., Smith, S.V., and Jokiel, P.L. (1977) The response of coral growth rate and skeletal strontium content to light intensity and water temperature Proc. 3rd Intl Coral Reef Symp. 2: p. 425. Lea, D W., and Boyle E .A. (1993) Determination of carbonate-bound barium in corals and foraminifera by isotope dilution plasma mass spectrometry. Chern. Geol. 103: p. 73-84. 59

PAGE 70

Lea D.W and Martin, P.A (1996) A rapid mass spectrometric method for the simultaneous analysis of barium, cadmium, and strontium in foraminifera shells Geochim. Cosmochim Acta 60 : p 3143-3149 Lea D W ., Shen, G T and Boyle, E.A (1989) Cor a lline barium records tempora l variability in equatorial Pacific upwelling Nature, 340: p. 373-376. Le Cornec, F and Correge T. (1997) Determination of Uranium to Calcium and Strontium Calcium Ratios in Corals by Inductively Coupled Plasma Mass Sp e ctrometry Journal of Analytical Atomic Spectrometry 12 : p 969-973 Min, G.R Edwards, R.L. Taylor F. W., Recy J., Gallup C.D and Beck, J W. (1995) Annual cycles of U / CA i n corals and U/Ca thermometry. G e ochim. Cosmochim. Acta 59: p 2025-2042 Mitsuguchi T. Matsumoto E., Abe 0., Uchida T and Isdale, P.J (1996) Mg/Ca Thermometry in Coral Skeletons Science 274 : p. 961-963. Quinn, T M., Taylor F.W Crowley T J ., Link S .M. (1996) Evaluation of sampling resolution in coral stable isotope records : A case study using records from New Caledonia and Tarawa. Paleoceanography 11: p 529-542. Rus sell, W.A., Papanastassiou, D.A and Tombrello, T .A. (1978) Ca isotopic fractionation on the Earth and the other solar system mate rials. Geochim Cosmochim Acta 42: p 1075-1090 Shen G T., Boyle E A., and L ea D .W. (1987) Cadmium in corals as a tracer of historical upwelling and industrial fallout. Nature 328: p 794-796. 60

PAGE 71

Shen, G T Cole, J.E., Lea, D .W., Linn, L J., McConnaughey, T.A., and Fairbanks, R.G. (1992) Surface ocean variability at Galapagos from 1936-1982: Calibration of geochemical trac e r s in corals Pa leoceanography 7: p 563-588. Shen, G.T., and Dunbar, R.B. (1995) Environmental controls on uranium in reef corals Geochim. Cosmochim. Acta 59 : p 2009-2024. Smith, S.V., Buddemeier, R.W., Reda l je, R.C. and Houck, J .E. (1979) Strontium-calcium t hermometry in coral s keletons. Science 204 : p 404-407. Tudhope, S. (1994) Extracting high-resolution climatic records from coral skeletons Geoscientist 4: p. 1 7-20. Weber, J N. (1973) Incorporation of strontium into reef coral skeletal carbonate. Geochim. Cosmochim. Acta 37: p. 283-299 61

PAGE 72

62 X. APPENDICES

PAGE 73

63 Appendix A Sample # Distan ce ( e m) 238U /40Ca ( 1 0 -6) 88S r /40Ca (10"-3 ) 1388a/40Ca (1 0"-6) 8 1 13.60 1 .269 7 .845 2.160 B-2 13.70 1 .317 8 3 2 0 1.917 8-3 13. 81 1 .358 9.173 0 .271 8-4 1 3.91 1.408 8 .961 1.555 8-5 14. 0 1 1.436 8 7 01 1.946 B-6 14.12 1.410 9 .120 0.512 8 7 14.22 1.442 8 .504 2 .666 8-8 14.32 1.405 8 .979 1.000 8-9 14.42 1.454 8 .880 1 .775 8-10 14.53 1.2 5 6 8.774 1.413 8 -11 14.63 1.242 7 .400 1 .244 8-12 14.73 1.343 7 .064 0 .844 8 1 3 14.84 1.43 9 6 9 7 5 0 .724 8-14 14.94 1.365 7.869 1.598 B-15 15.04 1.463 7.530 1.221 8 -16 15.15 1.481 7.455 1 .187 8-17 15.25 1 .414 8 .107 1 .626 8-18 15.35 1.387 7.908 1.482 8-19 15.45 1.572 6.722 -0.421 8 -20 15.56 1.258 8.044 1.651 B-21 15.66 1.378 7 670 1.180 8 -2 2 15.76 1.366 7.730 1.406 B-23 15.87 1 411 7 5 5 9 -0.507 8-2 4 15.97 1.377 7 677 0.512 8-2 5 16.07 1.407 7 7 3 9 0.424 B-26 16.18 1.498 7 .915 -2.626 8-27 16.28 1.268 7.640 1.238 B 2 8 16.38 1.289 7.5 7 9 1.112 8-29 1 6 .48 1.291 7.5 6 1 1 .048 8-30 16.59 1.345 7.569 -0.390 B-3 1 16.69 1 2 6 4 7 .523 1.567 B-32 16.79 1.1 9 6 7.566 1.642 B-33 1 6 9 0 1.29 4 7 .668 0 .107 B-34 17.0 0 1 .380 7.749 1 .057 B-3 5 17.10 1 .304 7 .893 1 .130 B-36 17.21 1.295 7 .903 1.011 B-37 17. 31 1 .233 7.891 0 .265 B-38 17. 41 1.1 9 4 7.679 0.957 B3 9 17.51 1.188 7 .653 0 .886 B-4 0 17.62 1.255 7.95 3 1 .542 B-41 17.7 2 1.190 7 .813 2 .091 8-42 17.82 1.232 7 .839 2 .113 8-43 17.93 1.248 7 786 2 .193 8 -44 1 8.03 1 .362 8 .112 2 .167 B-45 1 8 1 3 1.388 7 791 2 .449 B4 6 18. 2 4 1.363 7 .971 2 .229 8-47 1 8 .34 1.304 7 .912 2.182 B-48 18.44 1.259 7.876 2 .186 B-4 9 1 8.54 1 1 9 6 7.790 2 .045 8-50 18.65 1.191 7.859 2 .065 8 51 1 8 .75 1.030 7 .666 2 .022 B-52 18.85 1.116 7.825 1 .954 8-5 3 18 .96 1 .097 7.808 1.966 8-54 19.06 1.153 7 .814 1 .854

PAGE 74

64 Appendix A (continued) 8-55 19.16 1 0 1 71 7.823 1.811 8-56 19.27 1 .175 7.921 2 .067 8-57 19.37 1.271 7 .965 2.036 8 -58 19.47 1.227 7.910 2.105 8-59 19.57 1.130 7.883 1.901 B-60 19.68 1 .164 7.748 2.000 8-61 19.78 1.119 7.864 1.842 8-62 19.88 1.133 7.781 1.586 8-63 19.99 1.103 7.743 1 .850 B-64 20.09 1.096 7.721 1.458 8-65 20.19 1.138 7 .843 1 712 8-66 20.30 1 .135 7.854 1.825 8-67 20.40 1.199 7.981 1.965 8-68 20.50 1.235 8.034 2 .051 8-69 20.60 1.234 7.909 1 .826 8-70 20.71 1.180 7 .893 1 .806 8-71 20.81 1.202 7.454 2.770 8-72 20.91 1.204 7 .686 2.642 8-73 21.02 1 .246 7.541 2.678 8-74 21.12 1. 241 7.358 2.702 8-75 21.22 1.193 7.488 2.624 8-76 21.33 1.179 7.476 2.606 8-77 21.43 1.170 7.421 2.625 8-78 21 .53 1.295 7.748 2.701 8-79 21.63 1.278 7.690 2 .701 B-80 21.74 1.327 7.800 2.789 8-81 21.84 1.314 7.956 2 .958 8-82 21.94 1.278 7.782 2.768 8 -83 22.05 1.283 7.798 2.724 8-84 22.15 1.200 7.623 2.794 8-85 22.25 1.306 7.070 2.985 8-86 22.36 1.204 7.766 2.839 B-87 22.46 1.153 7.635 3.025 8-88 22.56 1.147 7.578 2.858 8-89 22.66 1.246 7 .821 2.790 8-90 22.77 1.315 7.568 2 .817

PAGE 75

65 Appendix A (continued) Sample# Distance* (em} 238U/40Ca (1 0"-6) 8-91 22.87 1.223 8-92 22.97 1 .275 8-93 23 .08 1 .322 B -94 23 .18 1.236 B-95 23.28 1.237 B-96 23.39 1 .160 B-97 23.49 1 1 70 B -98 23.59 1 .123 8 99 23.69 1 082 8-100 23.80 1 225 B-1 0 1 23.90 1 .183 B-1 02 24.00 1 .114 B-103 24.11 1 .180 B-104 24.21 1 .165 B-105 24.31 1 .236 B-1 06 24.42 1.171 B 107 24.52 1 219 B-108 24.62 1 .245 8-109 24.72 1.202 B-11 0 24.83 1 .198 B-111 24.93 1 .154 B-112 25.03 1.164 B-113 25.14 1 .1 01 B-114 25.24 1 .112 B-115 25 .34 1.1 01 8-116 25 45 1 .147 8-117 25 55 1 .203 B-118 25.65 1 .198 B-119 25 .75 1 .223 B-1 20 25.86 1 .208 B-121 25 .96 1 .083 8-122 26.06 1.197 B-123 26.17 1 096 B-124 26.27 1 .182 B-1 25 26.37 1 .133 B-126 26.48 1 .128 B-127 26.58 1 .058 B-128 26.68 1.085 B-129 26.78 1.043 B-130 26.89 1 .109 B-131 26. 99 1 .149 8-132 27. 09 1 .160 8 133 27.20 1 .221 B-134 27.30 1 .194 B-135 27.40 1 .183 B-136 27.51 1 .127 B-137 27.61 1 .276 8 138 27.71 1.177 B-139 27.81 1 .125 B-140 27. 92 1 121

PAGE 76

66 Appendix A (continued) B-141 28.02 1 .034 B 142 28. 12 1 .077 B-143 28. 23 1 .187 B-144 28.33 1 .205 B-145 28.43 1.252 B-146 28.54 1.230 B-147 28.64 1 .128 B-148 28.74 1 1 05 B-149 28.84 1 .123 B-150 28.95 1.070 B-151 29.05 N/A B 152 29.15 0.964 B-153 29 .26 1 .012 B -154 29.36 1 .023 B-155 29.46 1.002 B -156 29.57 0 .979 B-157 29.67 1 1 01 B-158 29.77 1 141 B -159 29.87 1 .167 B 160 29. 98 1 .156 B-161 30.08 1 .163 B-162 30.18 1 .132 B-163 30.29 1 .091 B-164 30.39 1 .189 B-165 30.49 1 .124 B-166 30.60 0.991 B -167 30. 70 1 091 B-168 30. 80 1 .124 B -169 30. 90 1 071 B -170 31.01 1 .076 B-171 31.11 1 .135 B -172 31.21 1 1 07 B-173 31.32 0 951 B-174 31.42 1 .393 Ind icates the distance down axis from the surface of cora l to th e end of the subsample.

PAGE 77

Appendix B Master Scxucbhtet lor Strorrtill'n and bY ICP MS Calibration Standards Ca (g) Cam Sc (g) Sc ml total (g) total ml Ca 20 1. 0420 1.0159 0 .0492 0.0498 500.00 Ca40 2.0860 2 .0337 0.0488 0.0484 500.00 Ca 60 3 .1290 3 .0506 0 .0492 0.0488 500.00 Ca 80 4 .2000 4 .0948 0 .0491 0.0487 500.00 Reortss ion 431-45 At()'Cission 48145 &.llndb 0 .01580089 -0.00232812 a andb 0 .02444332 .00420504 9.932EOS 0.0055675 0 .00016303 0 .00913876 r-squaro 0 .99992099 0 .00456762 r-s.quare 0 .99991104 0 .00749752 25309.77 56 2 22479.7519 2 I 0 .52804248 4.1726E 1 .26364902 0 .00011243 Sr (g) Srm v (g) v m total !o: tota l ml Sr 0 75 0 .3767 0.3733 0 .0493 0 .0489 500.00 Sri 0 .4850 0 .<806 0 .0491 0.0487 500.00 Srl.5 0 .7484 0 741 7 0 .0491 0 .0487 500.00 Sr2 0 .9969 0 !1879 0.0487 0.0483 500.00 Ca(ppm Sc(ppb 43Ca/49Ca 20.38 97. 61 0.6123 40.79 96.82 0.5822 61.19 97. 61 0.5695 82.13 97. 41 O.SH7 Regres sion 86189 aandb 1.03415434 0 .00827297 r-squaro 0 .99987202 15625.9837 0 .96951296 Sr (ppm v (l>pi>J 86Sr/87Sr 0 7 4 4 97.42 1.3880 0 .958 97.02 1.3950 1.479 97.02 1.3950 1 .970 96.23 1.3960 --exponent 43Ca/45Sc 43145 MBC 43 1 45 SeC 49Ca/45Sc 49145 MBC .49760437 0. 3237 0.3311 0.3232 0.5286 0 511 g .95598315 0 .6307 0.6587 0 .6377 1 .0830 1.0182 1.15653846 0.9360 0.9865 0 .9629 1.6430 1 .5248 1 .07388585 1.2690 1.3325 1.2980 2.2090 2 061 1 Mas s bias lonnula lor Calcium. L OG 0 646741J2 t LOG 42. 958766 1 47 .952533 MBCmass bi as cotTection 0 .01169424 0 .01135843 0 .0078728 2 0 .00012396 exponent 86Sri89V 86189 MBC 86189 VC .2717253 0. 7900 0. 7974 0.7768 0 .163 12185 1.0410 1.0352 1 .0044 0 .16312185 1.6060 1.5970 1.5495 0 .22506468 2.1400 2 .1235 2 .0435 Mass 48 / 45 SeC 0 .4997 0 .9858 1 .4884 2.0078 --0\ '-l

PAGE 78

Appendix B (continued) Coral S1andards Name Sc ( Q) Sc (ml) y (Q) Y(ml) tota l (Ill cstd 1 0 .1039 0 .10296304 0 .1031 0 10217025 1005.4 cstd 2 0 .1039 0 .10296304 0.1031 0 10217025 1 005. 4 cs t d 3 0 1039 0 .10296304 0.1031 0.10217025 1005.4 Sc Y sol #2 0 .1038 0 .10286394 0 .1022 0. 10127837 48Ca/45Sc 48/45 MBC 48/45 SeC 86Sr/87Sr exponent 86Sr/89Y 1 .278 1.17751757 1 2 1156818 1 .387 -0.33402528 0 .9679 1 .279 1 .18077302 1.2 1491777 1 .393 0 .03910286 0 .9833 1.272 1 .18489876 1 2 1916281 1.401 0 .53411477 0 .9928 total (ml) Sc ( ppb) y (ppb ) 43Ca/48Ca 1001 .69373 102.891728 101 .691502 0 .5625 1001 .69373 102.891728 101 .691502 0 .5644 10 0 1 .69373 102.891728 101.691502 0 5731 1000 102.966802 100.974532 r eference formula 86 /8 9 MBC 86/89 YC Ca (ppm) 43 Ca (p pm) 48 0 .97904856 0 .99560918 49.7254968 49.7384552 0.98198258 0 .99859284 49.8837903 49.8754902 0 .97478453 0 .991273 0 3 50.0427266 50.0491591 expo n ent 43Ca/45Sc -1 .26900139 0.7187 -1.23833795 0.722 1 .0992375 0. 7289 Ca (ppm) avg Sr (ppm) 49.731976 0 .95141983 49 .8796402 0 .95430495 50.0459428 0 9 4722688 43/45 MBC 0 .76136232 0 76379321 0 .76623396 Sr (g)/Ca (g) 0 .01913095 0 .01913215 0.01892715 43145 SeC 0.78337885 0 .78588003 0 78839137 88Sr/40Ca 0 .00745647 0 .00745694 0 .00737704 (]'\ 00

PAGE 79

Appendix B (continued) Coral Samples s pike (g) total (g) Sc (ppb) y (ppb) 43Ca/48Ca 12_aa1718 0 .4885 14.4462 99.48 97.56 0.5742 lpaa1 -728 0 .4878 13.9248 99.36 97.44 0 .5796 12_aa1-738 0.4888 13.6819 99.29 97.37 0 .5742 lpaa1-748 0.4897 13.1713 99.14 97.22 0.5771 paa1-758 0.4915 12.5407 98.93 97.02 0.584 2 paa1-768 0.4913 12.6388 98.96 97.05 0 .5845 paa1-778 0 .4918 12.7233 98.99 97.07 0 .5837 paa1-788 0 .4819 12.1100 98.87 96.96 0.5838 paa1-798 0.4903 12.2960 98.86 96.95 0.5829 paa1-808 0.4888 12.3028 98.88 96.96 0 .5875 48/45 SeC 86Sr/87Sr exponent 86Sr/89Y 86/89 M8C 86/89 YC 0.3757 1 .3820 -0.64620048 0 .3197 0.3269 0.3189 0 .5683 1.3740 -1.1480377 0 .4827 0.5021 0.4892 0 .4345 1.3780 -0.89675491 0.3692 0.3807 0.3707 0 .3692 1.3850 -0.45876011 0.3137 0.3187 0 .3098 0.4424 1.3800 -0.77138692 0 .3760 0.3861 0.3746 0.4338 1.3810 -0.70877104 0.3691 0.3782 0.3670 0.3973 1.3810 -0.70877104 0.3368 0.3451 0 .3350 0.7354 1.3820 -0.64620048 0.6393 0.6536 0.6337 0.6097 1.3850 -0.4587601 1 0 .5321 0.5405 0 .5240 0 .5446 1.3780 -0.89675491 0 .4762 0 .4911 0.4 762 exponent 43Ca/45Sc 43/45 M8C -1.08180065 0.2325 0.2442 -0.9966832 0.3535 0.3699 -1.08180065 0.2695 0.2831 .0359904 0.2298 0 .2409 -0.92479891 0.2773 0.2892 -0.92013049 0.2719 0.2835 -0.93258495 0.2488 0.2596 -0.93102721 0 .4612 0.4811 -0.9450565 0.3821 0.3989 -0.87357758 0.3423 0 .3562 Ca {ppm) 43 Ca (ppm) 48 Ca (ppm) avg 15.5235146 15.5404003 15.5319574 23.4061812 23.4233903 23.4147858 17.9352251 17.9495343 17.9423797 15.2605416 15.2780582 15.2692999 18.25461 18.2718976 18.2632538 17.9041447 17.9209256 17.9125352 16.4084678 16.4273985 16.4179332 30.2526771 30.2592383 30.2559577 25.1031332 25.115254 25.1091936 22.4345507 22.4509025 22.4427266 43/45 SeC I48Ca/45Se 0.2430 0.4049 0 .3675 0 .6100 0 .2811 0.4693 0 .2388 0.3982 0.2861 0.4747 0.2806 0.4652 0.2569 0.4263 0.4757 0.7899 0.3943 0.6555 0.3522 0 .5827 Sr (ppm) Sr ( g )/Ca (g) 0 .29704734 0 .01912491 0 .46174715 0.01972032 0.34715272 0.0193482 0.28827521 0.0188794 0.35088241 0 .01921248 0.34359152 0.01918162 0 .31260795 0.01904064 0.60149083 0 .01988008 0.49542444 0.0197308 0.44912011 0 .02001183 48/45 M8C 0.3776 0 .5720 0.4377 0.3724 0 .4472 0.4384 0 .4014 0.7438 0.6167 0.5508 88Sr/40Ca 7.454E-03 7.686E 7 .541E-03 7 .358E-03 7.488E-03 7.476E 7.421 E-03 7.748E-03 7.690E-03 7.800E-03 0'\ \0

PAGE 80

Appendix B (continue d ) Coral standards U WnfWs = 1.01239762 Na m e tota l wt (g) Ca meas (pp m Ca meas (mg} coral meas (mg) spike wl (g ) 236U/235U abund. ratio u (J.Ig) U co r al ( ppm) 238U/40Ca temp ( "C) U cstd 1 1005.4 49.731976 49.6162067 124.40 59.76 6 .332 6 .25201943 0.329 2 .65 1 .140E-06 23.49 csld 2 1005. 4 49.6796402 49.9641231 124.77 59.76 6.436 6 .35470579 0.335 2 .69 1 .156E-06 23.14 csld 3 1005. 4 50.0459426 50.1307073 125.19 59.78 6.366 6 .26556997 0.331 2 .65 1 .139E-06 23.50 average= 124.79 average= 23.36 Ba WnfWs = 1 .01650629 Name total wt (g) Ca meas (ppm Ca meas (mg) coral meas (mg ) spike WI ( g) 1368a/135Ba abund. ratio Ba (J.Ig) Ba coral ( ppm ) 138Ba/40Ca Ba/Ca molar cstd 1 1005. 4 49.731976 4 9 .6162067 124. 40 59.76 1 .3670 1 .33726646 0.491 3 .94 2.126E-06 2.674E-06 cstd 2 1005. 4 49.6796402 49.9641231 124.77 59.76 1.3390 1 .30967551 0.479 3 .64 2.069E-06 2 .797E-0 6 cs l d 3 1005.4 50.0459426 50.1307073 125. 19 59.76 1.3430 1 .3137865 0.460 3.84 2.069E-06 2.797E-06' average= 124.79 average 2 .623E-06 _,

PAGE 81

Appendix B (continued) Samples total wt (g) Ca meas (ppm Ca meas (mg) coral meas (mg) spike wt (g) 238U/235U abund. ratio U (ng) U coral (ppm) 238U/40Ca temp ( C) U paa1-71B 14.4462 15.5319574 0.22355063 0 .558 0 .4885 3.747 3 .699671 1 .558 2.79 1 .202E-06 2 2 .16 [paa1 72B 13.9248 23.4147858 0.32484429 0 .811 0.4878. 5 .386 5 .31796852 2 .269 2.80 1 .204E-06 22. 11 [paa1 -73B 13.6819 17.9423797 0 .2445809 0.611 0 .4888 4 .228 4.17459541 1.767 2 .89 1 .246E 21 .22 [paa1 74B 13.1713 15.2692999 0.20037514 0.500 0 .4897 3 .467 3 .42320773 1.442 2 .88 1 241 E-06 21 .33 paa175B 12.5407 18.2632538 0 .22818969 0 .570 0.4915 3 .771 3 .72336786 1 .578 2.77 1 .193E06 22.36 paa1-76B 12. 6 388 17.9125352 0.22555838 0 .563 0.4913 3.688 3 .6414162 4 1 .542 2 .74 1 .179E06 22.66 paa177B 12. 7233 16.4179332 0 .20812024 0 .520 0 .4918 3 .383 3 .34026875 1 .412 2 .72 1 .170E-06 22.85 12.1100 30.2559577 0 .36504897 0 .912 0 .4819 6 .526 6 .44356898 2 .741 3 .01 1 .295E-0 6 20.16 lpaa1-79B 12.2960 25.1091936 0.30760451 0 768 0 .4903 5 .387 5 .31895589 2 .281 2 97 1 .278E-06 20. 51 [paa1 80B 12.3028 22.4427266 0 .27509054 0 .687 0 .4888 5 .032 4.96843995 2.118 3 .08 1 .327E-06 19.46 Ba WnNis = 1.01650629 Name total wt _(g)_ Ca meas (ppm Ca meas (mg) coral meas {mg) spike wt (g) 138Ba/135Ba abund. r atio Ba ( ng) Sa coral ( ppm ) 138Bat40Ca Ba/Ca molar paa171B 14.4462 15.5319574 0 .22355063 0 .558 0 .4885 1 .0230 1 .0007488 2 .869 5 .14 2 .770E-06 3 .746E-03 [paa1-72B 13.9248 23.4147858 0 .32484429 0 .811 0 .4878 1 .3590 1 .32944049 3 .975 4 .90 2 .642E-06 3 .572 E -03 [paa1-73B 13.6819 17.9423797 0.2445809 0 .611 0.4888 1.0740 1.0506395 3 .035 4 97 2 .678E06 3 621 E-03 [paa1-74B 13.1713 15.2692999 0 .20037514 0 .500 0.4897 0 .9070 0 .88727191 2 .508 5 01 2 .702E06 3 .653E-03 [paa1 75B 12.5407 18.2632538 0 .22818969 0 .570 0.4915 0.9880 0 .96651008 2 .774 4 .87 2 .624E06 3 .548E03 [paa1 76B 1 2 .6388 17.9125352 0 .22555838 0 .563 0 .4913 0. 9724 0 9512494 2 723 4 .84 2 .606E-06 3 .524E-03 [paa1 77B 12.7233 16.4179332 0 .20812024 0 .520 0 .4918 0.9109 0 .89108708 2 .531 4.87 2 .625E-06 3 .550E-03 [paa1-78B 12.1100 30.2559577 0.36504897 0 .912 0 .4819 1 .5450 1 .51139481 4 .568 5 01 2 .701E-06 3 .652E-03 [paa1-79B 12.2960 25.1091936 0 .30760451 0 .768 0 .4903 1.3160 1 .2873757 8 3 .849 5 01 2 701 E-06 3 .652E03 paa180B 1 0 .27509054 _ 0 .687 0.4888 1.2320 3 .554 5.17 2 .789E-06 3 771 E -03 '-l

PAGE 82

Appendix C Spreadsheet l or DC P Analys i s o f Mg. Sr and Ca I Sampl e san'fllel san'flle2 l san'flle3 s an'fl le4 San'fl leS san'flle 6 san'flle7 san'flle8 San'fllo9 sample t O SarJll l e11 sample 12 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 sample San'flie 2 samplo3 sample4 sampleS sampleS sample7 sampleS s a mple9 s ample tO sample 1 sample12 MQ (WI01a) 0 .027210999 0 .02743931 0.032328418 0 .02166361 1 0.036456121 0 .025521387 0 .02751744 0 .025683942 0.020613021 0 .0223005 0 .02496613 0 .019419137 Ca ( wt%) 7 .663927172 7 .333703732 9 .202013479 6 .201764493 11. 34114814 7 .613749763 8.237474048 7 .260184988 5 .608678419 6 2 12049207 6.803748976 5 .483488775 13.75 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 S ST(MQ/Ca) 36.4 7269036 3 8 9 14370431 35.9953658 35.73902267 32.17713196 33.93499669 33.78807035 3 6 3 0 795959 38.06640338 36.97569706 37.99309955 36.355776631 Mg/Ca( m ol ) 0 .005854977 0 .0061 69954 0 005793402 0 .005760334 0.00530085 0 005527 615 0 .005508661 0 .005833727 0 .006060566 0 .005919865 0 .00605 1 11 0 .0058398951 Sr (ppm) 0. 149535932 0.1443 5974 7 1 0 .184989985 0.12592807 0 .226526153 0 .155522275 0 .168757883 0 14 7878097 0 111548443 0 123059833 0 133597076 0 109474609 88Sr/ 40Ca{m ol) 0 .00760486 0 .0076721971 0 .007835422 0.00791415 0.007784995 0 .007961429 0 007984855 0.007938769 __ ()_,_007751754 0 .007721087 0 .007653246 0.0077813231 i:j

PAGE 83

Appendix D Firs t Sp i k e Cal i bration I U Wn/Ws = 1 .01239782 U spike (ppb ) 0 .848 Name stand wt (g) spike wt (g) 238U/235U abund ratio U (ng) U s tand (ppb) meas/a c t u al U (ng) correct. U stand ( ppb ) cal 1 1 .9696 0 .9842 11.10 10.959794 7 .709 3.93 0 .816 9.875 5 03 cal2 0 .4886 0 .9823 2 83 2 .79425379 1 83 1 3.76 0 781 2.346 4 .82 cal3 0.9806 0 .9792 5.70 5 .62800233 3 .772 3 86 0 .802 4 832 4 .95 cal4 3 .9075 0 .9763 21.60 21.3271667 16.222 4 .17 0 8 6 5 20 780 5 .34 cal5 3 8643 1 .9633 11.00 10. 8610571 15. 2 2 8 3.96 0 821 19 506 5 .07 ideal= 1 .816 actual = 4 .815 0 .817 average 5. 04 Ba Wn/Ws = 1.01650629 Ba sp ike (ppb) 4 18 Name total wt (g) spike wt (g) 138Ba/135Ba abund ratio Ba (ng) S a stand (ppb) meas/ actual Sa (ng) correct S a s tand ( ppb ) cal 1 1 .9696 0 .9842 1 .470 1 .43802613 8 .059 4 .11 0.857 8.795 4 48 ca l 2 0 .4886 0 .9823 0 .459 0 .44901632 2 137 4.39 0 .916 2 333 4 .79 cal 3 0 .9806 0 .9792 1.050 1 .02716152 5 .429 5 56 1 .159 5 925 6 06 cal4 3 .9075 0 .9763 2.630 2 .57279506 16.452 4 23 0 .882 17 955 4 .61 cal 5 3 8643 1 .9633 1.430 1 .39889617 15. 562 4 .04 0 .843 16. 984 4 .41 ideal = 0 .643 actual= 4 .794 0 931 a v erage 4 87 Spike concentrations u sp ike (ppb)_ 0 .848 Ba s p ike (ppb) I 4 18 Second Spike Calibration w / U500 U Wn/Ws = 1 .00604659 U spike (ppb ) 0 897 Name stand wt ( g) spike wt ( g) U 500 238U/ 2 3 5U 238U/ 235U abun d. rat i o U (n g ) U s tand ( ppb ) me a s/ actua l u (ng) correct U s tand (ppb) cal 1 0 .0522 4 .g7o3 0 .950 0 .375 0 .400 0 .39485581 3 .985 76. 62 0 .728 5 .400 103 .83 cal2 0 .1008 4 .9723 0 .950 0 531 0 566 0 .55911583 7 .895 78.61 0 .747 10.697 106.51 ac tual= 105 2 0 0 0 .738 a v erage = 105.17 Spike concentrations U spike (ppb) I 0.897 -'----------2;:1


printinsert_linkshareget_appmore_horiz

Download Options

close
No images are available for this item.
Cite this item close

APA

Cras ut cursus ante, a fringilla nunc. Mauris lorem nunc, cursus sit amet enim ac, vehicula vestibulum mi. Mauris viverra nisl vel enim faucibus porta. Praesent sit amet ornare diam, non finibus nulla.

MLA

Cras efficitur magna et sapien varius, luctus ullamcorper dolor convallis. Orci varius natoque penatibus et magnis dis parturient montes, nascetur ridiculus mus. Fusce sit amet justo ut erat laoreet congue sed a ante.

CHICAGO

Phasellus ornare in augue eu imperdiet. Donec malesuada sapien ante, at vehicula orci tempor molestie. Proin vitae urna elit. Pellentesque vitae nisi et diam euismod malesuada aliquet non erat.

WIKIPEDIA

Nunc fringilla dolor ut dictum placerat. Proin ac neque rutrum, consectetur ligula id, laoreet ligula. Nulla lorem massa, consectetur vitae consequat in, lobortis at dolor. Nunc sed leo odio.