Isotope ratios of molybdenum determined by thermal ionization mass

Molybdenum is an essential nutrient for human beings,* 1·2 but data on which to ... samples and correcting forfractionation with the 92/98 ratio.9. T...
0 downloads 0 Views 708KB Size
1717

Anal. Chem. 1009, 85, 1777-1722

Isotope Ratios of Molybdenum Determined by Thermal Ionization Mass Spectrometry for Stable Isotope Studies of Molybdenum Metabolism in Humans Judith R. Turnlund,’ William R. Keyes, and Gary L. Peiffer USDAIARS, Western Human Nutrition Research Center, P.O.Box 29997, Presidio of San Francisco, California 94129

Methods were developed to separate and purify Mo from biological samplesand to measure isotopic ratios in 1 fig of Mo. A magnetic sector, thermal ionization mass spectrometer was used with simultaneous collection of five isotopes. Isotopic ratios were corrected for mass fractionation by iterative normalization using the 96/98 ratio. Ion beam intensity was enhanced by using a doublefilament configuration, loading samples onto evaporation filaments with silica gel and boric acid. A triple-isotope-dilutionapproach was used, so the method could be applied to two-tracer studies of Mo metabolismin human subjects. 94Mowas added to samples prior to purification to quantify the total Mo content of samples and to determine the amounts of enriched “Mo and lWMoappearing in urine and fecal samples of study participants.The three ratios, 94/98,97/98, and 100/98, were determined with within-run precision of from 0.06 to 0.10% (RSD). Precision of the ratios between replicates was from 0.05 to 0.08%.

INTRODUCTION Molybdenum is an essential nutrient for human beings,Q but data on which to base dietary intake recommendations are very limited. The only tracer study of Mo metabolism in humans was reported in 1964.8 For that study, the radioisotope WMo was used in four cancer patients to follow ita disappearance from blood and excretion in the urine and feces. Stable isotopes provide a means of studying Mo metabolism in humans with no risk from exposure to radioisotopes.4 The concentration of Mo in the dieta and tissues of humans is very low, in the nanogram to microgram per gram range. In order to analyze stable isotopes of Mo for studies of ita metabolism in humans, a method for precise analysis of isotopic ratios using 1pg of Mo was needed. We planned to collect samples from humans for at least 3 weeks following the administration of stable isotopes and expected isotopic enrichments to fall to less than 1% Therefore, our goal was to measure natural Mo ratios with external precision and accuracy of within 0.1 % Methods have been reported to determine isotopic ratios of Mo in 5-40-pg samples from meteorites, from rocks, and from high-purity Mo metal using magnetic sector, thermal ionization mass spectrometry (TIMS). External (betweenrun) precision of 0.24.7% (RSD) was achieved by several

.

.

* Author to whom correspondenceshould be addreseed.

(1) National Research Council, Recommended Dietary Allowances, 10th ed.; National Academy Press: Washington,DC, 1989. (2) Rajagopalan, K.V. Nutr. Rev. 1987,45, 321-328. ( 3 ) Roeoff, B.; Spencer, H. Nature 1964, 202, 410-411. (4) Turnlund, J. R. J. Nutr. 1989,119, 7-14.

investigators using 15-50 pg of Mo, a single filament of either tantalum or rhenium, and a secondary electron multiplier (SEM).”7 One of these investigators corrected for fractionation, normalizing with the 95/98 ratio.7 Rees reported precisionof 0.06% (within-run RSD) for the 92/98 ratio using a 5-pg sample and 5 h of preheating.8 Moore achieved an internal precision of 0.1% (95% confidence interval) with only 30 min of preheating by using triple filaments and 40-pg samples and correctingfor fractionation with the 92/98ratio.9 Tamura used a carbonizingagent for ionization enhancement to measure isotope ratios in 5-25-pg samples by Faraday collector with external precision of 0.08-0.19% after 50-70 min of preheating, normalizing to 92/100.10 Qi-Lu loaded 20 pg of Mo in an ammonium molybdate solution onto triple filaments, added nitric and boric acid, preheated samples for 3 h, and used a Faraday collector to achieve internal and external precision of 0.003% after normalizing with 94198.11 A recent review of isotope dilution mass spectrometry (IDMS)12discusses the principles and procedures of this highprecision analytical method. We report herein methods to separate Mo from biological samples and to measure isotopic ratios by TIMS in samples containing only 1pg of Mo with sufficient accuracy,precision, and speed to study Mo metabolism in humans. The TIMS methods described include a new ionization enhancement technique and automated, simultaneous collection of five Mo isotopes. Triple-isotope-dilution procedures, equations for calculating masses of total Mo and Mo isotopes from three isotopic ratios, and procedures for correctingfor fractionation using a fourth ratio are described. We also report ranges of isotopic ratios and enrichments measured in fecal and urine samples collected during the first 6 d following oral and intravenous administration of Mo isotopes.

EXPERIMENTAL SECTION Human Study. Human volunteers confined to a metabolic research unit were fed a diet with a Mo content of 120pgld. After adaptationto the level of dietary Mo intake for 6 d, 33 pg of g?Mo was infused into an arm vein of study participants. Six d later, 95 pg of *WMo was substituted for the usual Mo in the diet. Complete urine and fecal collections were taken to measure isotopic enrichment and to quantify the amounts of Mo and the oral and intravenous isotopes that were eliminated. The experimental protocol and consent forms were reviewed and approved by the Institutional Review Committee of Letterman Army Medical Center and the US.Department of Agriculture Human Studies Review Committee. Isotope Solution Preparation and Administration. Mo metal powders enriched in wMo (94.25 atom %), 97Mo(94.25 (6) Murthy, V. R. Geochim. Coschim. Acta 1963,27, 1171-1178.

(6) Crouch, A. C.; Tuplin, T. A. Nature 1964,202, 1282-1284. (7) Wetherill, G. W. J. Geophys. Res 1964,69,4403-4408. (8)Rees, C. E. Znt. J. Mass Spectrom. Zon Phys. 1969,3,71-80. (9) Moore, L. J.; Machlan, L. A.; Shields, W. R.; Gamer, E. L. A n d . Chem. 1974,46, 1082-1089. (10) Tamura, 5.Mass Spectrosc. (Tokyo) 1975,23, 49-59. (11) Qi-Lu;Masuda, A. J. Am. SOC. Mas8 Spectrom. 1992, 3, 10-17. (12) Fasaett, J. D.; Paulaen, P. J. A d . Chem. 1989,61,643A-649A.

Thk, artlcle not 8ub)ect to U.S. Copyright. Published 1003 by the Amerlcan Chemical Society

1718

ANALYTICAL CHEMISTRY, VOL. 65, NO. 13, JULY 1, 1993

Fecal collections

Urine collections

Homogenize (blender)

Thaw

Lyophilize I I Crush to fine powder

Heat on shaking water bath I I Weigh 94Mo solution into Pyrex beaker

I

I

I

I

IS'column

2Nco1umn

- 1. Load sample dissolved in 6 N HCI - 2. Rinse 6 N HCI

- 3. Elute Cu 2.5 N HCI - 4. Rinse 2.5 N HCI - 5. Elute Fe, Zn 0.005 N HCI

1. Load Mo

in 3 N HCI

2. Rinse deionized H 2 0

3. Elute Mo 1 N HCI

- 6.Rinse 0.005 N HCI

- 7. Elute Mo 1 N HCI Weigh 94Mo solution into crucible,add deionized H 2 0 , equilibrate overnight

I

I

Dry on hot plate

hY

Anion exchange resin chloride form

Cu F e + Z n

Mo

hY

Anion exchange resin chloride form

Mo

Fbwe 2. Diagram of the lon-sxchange procedures used to separate and purlfy Mo from blologlcal samples.

Ash in mufile furnace at 500°C

L

I I Sample completely

Add 1:l HNO 3 , heat, dry

ashed ?

1

yes Dissolve sample in 6 N ultrapure HCI Flgure 1. Outline of sample preparatlon procedures. atom %), and lWMo (97.42 atom %) (Oak Ridge National Laboratory, Oak Ridge, TN) were weighed into acid-washed Teflon beakers, a 3/1 mixture of ultrapure HCl and HNOs was added, and the solutions were heated until the Mo was dissolved. The enriched solutions were transferred into acid-washed polypropylene bottles and diluted with deionized water to the desired concentrations. The exact concentrations of the isotope solutions were determined by isotope dilution using a lo00 ppm natural Mo standard solution (Ricca Chemical Co., Arlington, TX). The isotope solutions were diluted for use as feeding ("JOMo),infusion (@'Mo),or isotopic diluent PMo) solutions. For each isotope, a set of mixtures of several amounts of the isotope solution was combined with the Mo standard solution. The enrichments were calculated to span the range of enrichments expected in our study. They were analyzedto comparemeasured vs expected isotopic ratios. Sample Collection. Samples were collected in new plastic containers. Fecal samples were frozen immediately after collection and combined into 6-d pools before homogenizing. Urine was collected in 24-h periods, except on the days when isotopes were infused, when urine was collected in 8-h periods. Twentyfour-hour urine collectionswere combined into 6-d pools, except after isotope infusion, when the 8-h collectionswere not included. Care was taken to avoid trace element contamination during all phases of sample collection and preparation. The precautions included handling all samples with powder-free plastic gloves, weighing, measuring, heating and transferring samples under plexiglass enclosures, either in a clean room or in a laminar flow bench, using ultrapure acids and acid-washed containers, and keeping samples covered. Sample Preparation. An outline of sample preparation procedures is given in Figure 1. Details of the procedures have been described previously.13 Fecal collections were combined into 6-d pools prior to sample preparation. Five-gram subsamples (13) Turnlund, J. R.;Keyes, W.R.J. Micronutr. Anal. 1990, 7,117146.

of lyophilizedfecalsamplesand 250-g subsamplesof urine samples were used. A stable isotope solution containing 3 pg of MMo was added as the isotopic diluent. Both fecal and urine samples were ashed in a muffle furnace. Certified reference materials were prepared using the same procedure as for the fecal samples. The crucibles, breakers, and watch glasses used in ashing, the columns, the Teflon beakers used to concentrate the samples after the columns, and all plastic test tubes were acid-washedin 6 N HC1 and rinsed with deionized water. In addition to acid washing, the crucibles and beakers were then refluxed with 8 N HNOa and rinsed with deionized water before use. Mo Separation and Purification. All procedures related to mineral separation and purification were performed in a clean room. Anion-exchangeresin (AG l-X8,200-400 mesh, chloride form, Bio Rad Laboratories, Richmond, CA) was used. Glass chromatography columnsof 10-mm inner diameter (i.d.) packed to a length of 9.5 cm were used for samples with dry weights over 3 g (fecal and some diet samples); the column volume was 3.0 mL. For other samples, a 7-mrn4.d. column packed to a length of 7.5 cm was used; the column volume was 0.8 mL. All Mo fractions eluted from these columns were subsequently purified using 3-mm4.d. columnspacked to a length of 11cm; the column volume was 0.4 mL. A previously reported procedure for the separation of Zn, Cu, and Fela was modified to includethe separation of Mo. A diagram of the modified procedure is given in Figure 2. After samples were loaded onto columns, they were rinsed with 6 N HCl and Cu was eluted with 2.5 N HC1. Fe and Zn were eluted together with 0.005 N HC1. The Fe and Zn were separated later. Columns were rinsed with 0.005 N HCl and then Mo was eluted with 1.0 N HC1. The first 3 column volumes of 1.0 N HC1 were allowed to pass through the columns, then 7-8 column volumes were collected as the Mo fraction. This fraction was acidified to 3 N with HC1 and applied to a 3-mrn4.d. anion-exchange column. Columns were rinsed with deionized water and Mo was eluted with 1.0 N HC1 in 8 column volumes. The procedure for urine sampleswas modified to collectonly Mo, using a proceduresimilar to second columns, except samples were loaded in 6 N HCl and rinsed with 2.6 N HC1before rinsing with deionizedwater. After separation, the Mo fractions were concentrated to about 10 pL in Teflon beakers. Recovery of Mo from the columns was determined first by using a Mo standard and then from the biological samples. A known quantity of Mo in an atomic absorption standard was run through two anion-exchangecolumns. The 7-8 column volume Mo fraction was diluted and analyzed by furnace atomic absorption spectrophotometry. Recovery of Mo from fecal samples was also determined with atomic absorption spectrophotometry. A more precise determination of the Mo recovery from urine samples was done by isotope dilution using TIMS,

ANALYTICAL CHEMISTRY, VOL. 65, NO. 13, JULY 1, 1993

Table I. Sample Heating Parameters for Mo Ratio Measurements. filament control control param value detector minlstep heated 0 10 5 5 5 5

E I I I I I, E

FC FC IC IC IC IC

0.1 A 3.0 A 0.2

v

1.0 v 10 mV

200 mV

SEM SEM SEM SEM FAR

FAR

1719

Table 11. Mass Balance Equation and Equations for Triple-Isotope-Dilution Calculation of Masses of Natural Mo, and Isotopic Tracers. mass balance equation for Rta:

Equations for isotope dilution calculations:

a Abbreviations: E, evaporation filament; I, ionization filament; FC, filament current; IC, ion current (amplifier voltage); SEM, secondary electron multiplier; FAR, Faraday cup.

adding isotopic diluent to different subsamples before and after column procedures. Isotope Ratio Determinations. Isotope ratios were determined with a computer-controlled 90° magnetic sector, thermal ionization mass spectrometer (Finnigan MAT Model 261,Bremen, Germany).13The collector assembly contains five Faraday collectors,one fixed and four with manually adjustable spacing. The fixed central collector is electronically exchangeable with a secondary electron multiplier (SEM). The five detectors were used for simultaneous collection of the Mo isotopes 94,96,97, 98,and 100. The operating software was extensively modified by us to include specializedfilament heating, ion beam focusing, and data reduction procedures.'* Sample magazines holding 13 pairs of double filaments were inserted into a filament bakeout device which automatically heated the filaments under vacuum at a filament current of 4.5 A for 15 min to remove dissolved gases and organic impurities. Mo samples were loaded onto the cleaned 0.025 mm X 0.76 mm zone-refined rhenium filaments in a laminar flow bench, using an automatic sample loading device (Finnigan MAT). A 250-W heat lamp was placed 10 cm above the magazine to aid in sample evaporation. Silica gel suspension (10 pL) was placed on the filament, which was then heated at 1 A for 3 min to evaporate the liquid. A 5-pL drop containing between 1 and 2 pg of Mo (calculated from estimated excretion and recovery data) in 1-10 p L of dilute HCl solution was placed on the filament and dried at 1 A for 3 min. A 5-pL drop of 0.25 M ultrapure H 8 0 3solution was added, and the current was increased to 1.5 A for 1 min and then automatically increased at 0.5 Almin to 2.0 A and turned off. After 13sampleswere loaded onto the evaporation filaments, the ionization filaments and extraction plates were added and the sample magazine was inserted into the mass spectrometer. Automatic analysis began after the ion sourcepressure reached 3 X 1V Pa. The sample heating procedure is described in Table I. At the end of each heating step, the ion beam at the pilot mass (98for natural samples or 94 for wMo-enriched samples) was optimized by adjusting the magazine rotation and lens and accelerating voltages. The SEM was used to center and focus the ion beam initially, and the central Faraday collector was used in the later steps. When the ion beam intensity at the pilot mass reached 10 mV, control passed from the filament heating module to the data collection module. The ion beam was refocused,amplifier baseline measurements were made, and the ion beam intensity was increased to 200 mV. This was done by increasing the ionization filament current to a maximum of 4.1 Aand then, if needed, increasingthe evaporationfilament current. The ion intensities were integrated for 8 s, with 5-5 idle time between measurements. Measurement of the intensities of the five isotopes was simultaneous. After 10 sequential measurements of the five ion beam intensities, the relative amplifier gains were measured and a data reduction module calculated the averageof the 10 measurements for each isotopic ratio, after correcting for amplifier baselines and relative gains. A second data reduction module then carried out isotopedilution and iterative normalizationcalculations,using and 100/98ratios for the 96/98ratio to correct the 94/98,97/98, isotopic fractionation. The individual ratio measurements were subjectedto the Dixon outlier test for extreme meanz4and outliers (p < 0.1) were eliminated from the averages. ____________

(14) Dixon,

W.J. Ann. Math. Stat.

1480, 22, 68-78.

+ a,# + a,@ a2,W + a 8 + a&P a&P + a,# + a# a,,W

= b, = b,

(3)

= b,

a R is the isotopic ratio, A the isotopic abundance,M the maw of total Mo, W the atomic mass, f the fraction of sample weighed for analysis, t the total Mo in sample, n the natural Mo (unenriched), s the 'OOMo-enriched Mo fed to subjecta, e the "Mo-enriched Mo infused in subjects,d the wMo-enrichedMo addedas isotopicdiluent, i the isotopeenrichedin feedingsolution (100),j the isotope enriched

in isotopicdiluent(94),k the reference isotope (981,and mthe isotope enriched in infusion solution (97). Each set of 13 samples included an unenriched sample of the same matrix. These unenriched samples were used to monitor column recovery, detect sample oontamination, and establish the natural isotopicratios used in enrichment and isotopedilution calculations. Isotope Dilution Calculations. The total Mo and the enriched isotope contents of fecal or urine pool samples were determined by triplaisotope dilution, using the measured isotopic ratios of the subsamples, the unenriched samples of the same matrix, and the solutions enriched in 1°0Mo,WMo, and MMo; the amount of MMoadded; and the fraction of the total pool weighed for analysis. The subsamples contained natural Mo,the WMo and 1wMo tracers, and the MMo isotopic diluent added after sample collection. The mass balance equation for the isotopic ratios of these samples is shown in Table I1 (eq 1). The equation for Rt&is given; equations for Rt* and R& are the same, with j or m substituted fori. The terms of the mass balance equations for each ratio were rearranged and then simplifiedby substitution (Le., Rnnfor AnJAq) to form the three linear equations (eq 2)in Table 11. Substituting matrix coefficients produced a set of three equations in three unknowns,eq 3 in Table 11. These were solved,

1720

ANALYTICAL CHEMISTRY, VOL. 65, NO. 13, JULY 1, 1993

using the determinantsof the coefficients (eq 4), for natural Mo and amounts of the enriched isotopes appearing in a sample. The total Mo content of the fecal or urine sample was calculated using eq 5. Fractionation corrections were used to improve the precision and accuracy of the isotopic ratios.13 W ,Ma, and Me were calculated from eq 4, using the measured 94198,97198, and 1001 98 ratios. These masses were then used in a variation of eq 1 to calculatean interim 96/98 ratio, beginning an iteration procedure. The interim 96/98 value was used to obtain normalized values of the 94198,97198, and 100198 ratios, which were then used in Ma, and Me, completing eq 4 to obtain corrected values for Mn, one iteration. This procedure was repeated four times. The change in the ratios in the fifth iteration was less than 0.01 % . The enrichment (E)of a sample in isotope i or j was calculated from the ratios using

E ( % ) = (Rt - R")/R" X 100

RESULTS AND DISCUSSION Mo Separation. Our first efforts to separate Mo from the other minerals in our samples focused on separation of Fe and Mo, since they elute at HC1 concentrations in close proximity. We found that a reasonable separation could be obtained by eluting Fe with 1.5 N HC1 and then Mo with 0.5 N HC1. When establishingoptimal conditionsfor Mo elution, we observed that Mo was retained on columns when HC1 concentrationsabove 1.5 N were used. It began to elute slowly after 7-10 column volumes of 1.5 N HC1 or after 4-6 column volumes of 0.5-1 N HCl, but was retained at HCl concentrations of 0.005 N or less. This is relatively consistent with previous observations in studies of the adsorption of Mo onto the type of resin we used at varying concentrations of HCl.I5 Based on this observation,we adopted a procedure of eluting copper with 2.5 N HCl, eluting iron and zinc with 0.005 N HC1, and then eluting Mo with 1 N HC1. The advantage of this approach is that Fe elutes very quickly at 0.005 N HC1, while Mo remains on the columns. Thus, the procedure provided a better separation of Fe and Mo than when Fe is eluted with 1.5 N HCl and Mo with 0.5 N HCl. A broadly tailing Mo elution pattern is characteristic of Mol6 and required elution of Mo in a large volume of 1 N HCl for adequate recovery. The recovery of Mo after two successive anion-exchange columns averaged approximately 50 % Higher recoveries could be achieved by collecting up to 12 column volumes, but the additional yield was small compared to the additional volume collected, and by using smaller diameter first columns (3-mm i.d. rather than 7 mm), but small columns at this stage slowed the elution procedure greatly. Recovery from urine samples, determined by isotope dilution of samples before ashing and again after column separation and purification, was 50 f 6 % and average recovery from fecal samples was similar. Isotopic Ratio Determination. A minimum ion beam intensity of 10 mV (1x A) at each isotope was required in order for ratio uncertainty due to counting statistics to be below 0.1 % . To obtain this minimum level, the intensity at the 94 pilot mass was raised to at least 200 mV (2 X A) for analysis of the 94Mo-enrichedsamples. A 1-pg sample was sufficient for ratio determinations of natural samples. A 2-pg sample of Mo was used for analysis of samples highly enriched with 94Moin order to provide sufficient ion beam intensity of the other isotopes.

.

(15) Scadden,E.M.;Ballou, N. E. TheRadiochemistryofMolybdelulm; National Academy of Sciences,National Research Council: Washington, DC, 1960. (16) Hicks, H. G.; Stevenson, P. C.; Schweiger, J. S. J. Chromatogr. Sci. 1978, 16, 527-533.

Table 111. Abundances of Naturally Occurring Mo Ieotopes isotope mass

reference range" (atom % )

measuredb(atom %)

92 94 95 96 97 98 100

15.05-14.74 9.35-9.11 15.93-15.78 16.71-16.56 9.6-9.48 24.42-24.00 9.63-9.60

14.852 & 0.006 9.266 0.005 15.921 0.003 16.663 & 0.004 9.565 0.007 24.115 & 0.005 9.618 0.005

" From De Bihvre and Barnes.l* Mean analyses of natural Mo standard.

** *

SD based on set of 10

Mo was eluted from columns, concentrated, and loaded onto filaments in a chloride solution. External precision of analysis was similar to that achieved by loading samples in a nitric acid solution, so the chloride form was used for convenience. Mo has a high ionization potential, 7.1 eV,'I which results in low ionization efficiency. Therefore, an ionization enhancement technique was used to enhance the ion beam intensity. Neither a single-filament silica gel technique nor the use of a double-filament configuration produced sufficient beam intensity, so a procedure was developed using silica gel with a double-filament configuration. During method development,we compared use of boric acid to phosphoric acid for the final loading step. Internal precision was 0.1-0.2% for boric acid and 0 . 2 4 3 % for phosphoric acid, so boric acid was used. Since 1-2-pg samples often resulted in a short-lived Mo beam, preliminary focusing and heating was done by using the SEM to monitor a low-intensity beam, which preserved the sample for analysis. The final heating step was carried out in the first data collection module. Final focusing and amplifier baseline measurement were followed immediately by simultaneous ratio measurements on the declining signal, and the time-consuming measurement of relative amplifier gains was carried out after data collection. The rapid sample heating conserved sample for increased signal life during analysis and reduced analysis time to 35 min per analysis. Given the large number of fecal, urine, and plasma samples generated by a human nutrition study, sample throughput was an important consideration in the development of methods. Emission of Mo ions from the rhenium of the ionization filament was not observed when the ionization filament current was limited to 4.1 A, well below the 4.5-A current used in the filament bakeout device to clean the filaments. Therefore, after the ionization filament current reached 4.1 A, the evaporation filament current was increased as needed, up to the same limit of 4.1 A, to provide the desired ion beam intensity. Interference from an unknown compound was observed occassionally at mass 100. This resulted in variability of the 100/98 ratio, and analysis was repeated. Unexpected enrichment of the 9 4 M ~9, 7 M and ~ , ImMo isotopes was observed in some natural samples, apparently due to contaminationby enriched Mo. %Meenrichedsamplea contained more 94Mothan the other enriched isotopes and enrichment of natural samples was most often observed in 9 4 M ~It. was most common in natural urine samples prepared along with enriched samples, which were ashed in pyrex beakers. Some of the natural Mo standard samples, which were not subjectedtoashing before column separation, showed similar enrichment. In most cases, enrichment of natural samples was small. When it was observed, analysis was repeated when enough purified sample remained. In some, (17) Heumann, K. G. In Inorganic Mass Spectrometry; Adame, F.; Gijbels, R., Van Grieken, R., Eds.; John Wiley & Sons: New York, 1988, pp 301-376.

ANALYTICAL CHEMISTRY, VOL. 65, NO. 13, JULY 1, 1993

Table IV. Natural Mo Isotopic Ratios and Precision uncorrected ratios measd internal external isotopes ref ratio6 ratio precisionc ( % ) precisiond (% ) 94/98 96/98 97/98 100/98

0.3832 0.6910 0.3959 0.3992

0.3801 0.6874 0.3956 0.4010

0.17 0.10 0.08 0.08

normalized ratios0 internal external precisionc ( % ) precisiond (% )

measd ratio

0.56 0.27 0.12 0.30

1721

0.3842 0.6910 0.3966 0.3988

0.10

0.05

0.06

0.08 0.06

0.08

a Normalized with 96/98= 0.6910. From De Bihvre and Barnes.l* Average relative standard deviation for 10ratio measurements. d Relative standard deviation for 10 replicate analyses of natural Mo standard.

Table V. Isotopic Ratios of Enriched Mo Solutions 94/98 ratioa expected measuredd expected 0.3909 0.4148 0.4837 0.7251 1.4132

0.3907 f 0.0002 0.4145 i 0.0003 0.4851 f 0.0005 0.7250 f 0.0009 1.4129 f 0.0033

97/98 ratiob measuredd

0.3985 0.4054 0.4308 0.5034 0.7573

expected

0.3986 f 0.0003 0.4054 f 0.0002 0.4304 f 0.0002 0.5032 f 0.0002 0.7569 f 0.0005

100/98ratioC measuredd

0.4052 0.4124 0.4377 0.5099 0.7628

0.4052 f 0.0008 0.4123 f 0.0003 0.4379 f 0.0010 0.5107 f 0.0002 0.7632 f 0.0003

a MMoadded to achieve approximate enrichments of 3,9,30,90,and 300%. "Mo added to achieve approximate enrichments of 1,3,9, 30, and 90%. 'WMo added to achieve approximate enrichments of 1,3,9, 30, and 90%. Mean f SD of 10 ratio measurements.

but not all, cases the enrichment repeated. This suggests that the contamination may have occurred during sample preparation and occasionally at the stage of sample loading and analysis, but there was not a single source of contamination. The precautions taken to clean glassware used in sample preparation failed to eliminate the enrichment of natural samples. With a few exceptions, the degree of enrichment observed would contribute an error of well under 1% to the calculation of total Mo, 91M~, or lWMoin fecal and urine samples. Table I11 shows the measured and reference natural abundances of the seven stable Mo isotopes. The measured values are all within the range of reference values.18 The 94/98, 96/98, 97/98, and 100198 ratios before and after normalization corrections and their precision are shown in Table IV. Mass fractionation of about 0.2 % per mass number was observed in the uncorrected isotopic ratios. Internal precision was 0 . 1 4 2 % (RSD of 10 ratio measurements) and external precision was 0.1-0.6 5% (between-sample RSD) for the uncorrected ratios. Normalization improved internal precision slightly. Since all isotopes were measured simultaneously, the effects of most ion beam intensity variations on the isotopic ratios were eliminated. Fractionation changes during data collection were minimal, due to the short duration of the data collection (about 3 min). Table IV shows that normalization improved the external precision of all ratios. The 94/98 ratio improved from 0.56 % before normalization to 0.05 % after normalization. After normalization, internal and external precision of all ratios was 0.1 5% or less. Expected and measured isotopic ratios in three sets of mixtures of a certified Mo reference solution with solutions enriched in ~ M o97M0, , and 1WMo are shown in Table V. The close agreement between expected and measured ratios at widely differing enrichments reflects the accuracy and precision of the method. Table VI compares our measurements of the concentrations of Mo in National Institute of Standards and Technology (NIST) standard reference materials with the certified values. All of our measurements were within the certified ranges. Isotopic Ratios a n d Enrichments of Experimental Samples. Table VI1 shows the ranges of the 100/98 and (18) De BiBvre, P.; Barnee, I. L. Znt.J . Mass Spectrorn. Ion Processes 1986,65,211-230.

Table VI. Concentrations of Mo in Certified Reference Materia 1s concentration (pg/g) reference material certified measured0 SRM 1571orchard leaves SRM 1577a bovine liver RM 8431 mixed diet a

0.3 f 0.1 3.5 f 0.5 0.288 i 0.0296

0.2304 f O.OOO6 3.731 f 0.005 0.2932 f 0.0023

Based on duplicate analysis using "Mo as the isotopic diluent.

* Recommended concentration.

Table VII. Urinary and Fecal Mo Excretion in 6-d Collections following Oral and Intravenous Administration of Stable Isotopes to Four Young Men Consuming 120 pg of Mo/d 6-durine 6-d fecal collection collection total Mo content (pg/d) oral dose of 95 pg of lWMo 100/98ratio enrichment (% ) 6-d excretion (pg) intravenous dose of 33 pg of "Mo 97/98 ratio enrichment ( 5% ) 6-d excretion (pg)

91-104

12-23

0.71-0.74 79-86 44-51

0.67-0.72 67-80 7.0-11

0.54-0.55 36-38 21-22

0.42-0.45 4.9-14 0.48-0.52

97/98 ratios, their enrichments, and the Mo concentrations in fecal and urine samples for 6-d periods following oral administration of lWMo and infusion of WMo. Urinary excretion of 91M~was highest during the 8-h collection immediately following the infusion, with mean enrichment of 185% In the collection which included days 2-5 following the infusion, mean enrichment fell to 14% and continued to decline to less than 1% 25 d after the infusion. 97M0 enrichment was higher in the urine than in feces, and very little of the infused 97Mowas eliminated via the feces. Mean enrichment of 9% was observed in feces during the 6 d following the infusion, declining to 4% during the next 6 d and to 1%after that. More of the oral dose of 'WMo was eliminated in the urine than in the feces (49 vs 9%),but the difference early after the feeding was not as great as it was after the infused dose, when 64% of the infused dose of WMo was excreted in the urine and 2% in the feces. The much higher enrichment of feces following an oral isotope dose was

.

1722

ANALYTICAL CHEMISTRY, VOL. 65, NO. 13, JULY 1, 1993

due to unabsorbed dietary lwMo. Some unabsorbed lwMo from the oral dose was eliminated in feces in the next 6 d. In addition, fecal excretion of 97M0 following infusion suggests that some absorbed 1wMo would be excreted into the gastrointestinal tract. The combination of these resulted in fecal 1wMo enrichment of 9 7% , which fell to 1%after 18 d. The dietary level of Mo used for this work, 120 pg/d, is within the recommended dietary intake range of 75-250 pg/ d.1 If dietary intake were higher, enrichment would be lower, and if dietary intake were lower, enrichment would be higher than found in this study. Higher doses of stable isotopes would result in higher enrichment of fecal and urine collections and lower doses would result in lower enrichment. Results of a human study of Mo metabolism, which includes the dietary level reported here, as well as higher and lower levels, will be reported in another paper.

CONCLUSIONS The method described here provided relatively rapid, precise analysis of Mo in l-pg samples. The use of simultaneous collection, double-filament configuration, and ionization enhancement with silica gel and boric acid enabled us to achieve high precision with smaller samples and generally shorter analysis time than other reported methods. Our precision of 0.05-0.08% was accomplished with samples of 1 p g of Mo and an analysis time of 35 min. The method is suitable for analysis of samples from studies of Mo absorption, excretion, and kinetics. These studies will provide data on Mo which have not been possible to obtain previously.

RECEIVEDfor review November 19, 1992. Accepted March 9, 1993.