mercury in blood _02.docx

8/11/2019 mercury in blood _02.docx

http://slidepdf.com/reader/full/mercury-in-blood-02docx 1/22



Keywords: Hg; Inductively coupled plasma mass spectrometry; Cold vapor atomic absorptionspectrometry; Urine;Whole blood

1. Introduction The standard method for the determination of

mercury in biomedical samples is cold vaporŽ . w xatomic absorption spectrometry CVAAS 19 .UCorresponding author. Fax: q1-507-266-4176.Ž .

E-mail address: [email protected] D.E. NixonAfter sample digestion or pretreatment to decom-

pose organic mercury compounds, total mercury

can be easily determined by stannous chloridewxreduction 4,5,7 . This method has been applied inwxa batch mode without digestion 3,4,8,9 and byw xffow-injection 57. For effective analysis thismethod requires at least 500 l of whole blood orw x1000 l of urine 7 .The method does have some limitations. The0584-8547r99r$ - see front matter1999 Elsevier Science B.V. All rights reserved.Ž .PII: S0584-8547 99 00057-9

(

) D.E. Nixon et al.r Spectrochimica Acta Part B: Atomic Spectroscopy 54 1999 1141 1153 1142method used in our laboratory uses large amountsof strong acids and strong oxidants to assure thew xcomplete decomposition of organic mercury 7 .The sample preparation step is labor-intensiveand time consuming. In addition, cysteine in urinesamples or oxidizing agents such as iodine pre-vent the effective reduction of mercury by stan-

nous chloride and the formation of nascent ele-mental mercury. Iodine can be present in urine



after use as a radiological contract agent.The potential advantages of inductively coupledŽ .

plasma mass spectrometry ICP-MS for routine

heavy metal screening are as follows: only 300 lof urine and 150 l of whole blood are requiredw xto perform the analysis 10 ; essentially no sample

preparation is required, only dilution with a dilu-ent containing acid and internal standards. Sig-nificant savings in sample volumes, reagents, tech-nician time and analysis time are realized.Easy reduction of inorganic mercury to ele-mental mercury and the volatility of elementalmercury, prerequisites for effective CVAAS anal-

ysis, present major problems for the determina-tion of mercury by ICP-MS. It is well known thatafter a sample containing mercury is nebulizedinto a conventional spray chamber, significantresidual mercury signal can be detected manyminutes afterwards. The mercury signal fails toreturn to baseline and all subsequent samplesappear to contain mercury, whether they do ornot. Retention of mercury in spray chambers andlong washout times with nitric acid have beenwxdocumented 1113 . This phenomenon has beenduplicated in our method development for ICP-MS heavy metal screening using 1% nitric acid asthe diluent. Mercury either adsorbs onto the spraychamber walls or is retained as vapor in the deadwxvolume of the spray chamber 14,15 . The sameadsorptionvaporization phenomena are observed

w xwith mercury solutions in nitric acid 16 . Withnitric acid alone in solution, mercury adsorptiononto the container walls is prevented but notw xmercury volatility 16 . The addition of gold ordichromate was required to prevent volatilizationw x16 .Several solutions to this problem of mercuryspray chamber memory have been published. Flow

injection and mercury reduction followed by goldamalgamation has been used to eliminate the



w xaerosol spray chamber memory 17 . Mercury issubsequently released from the gold by heatingand detected by ICPMS. Another solution is theŽ

. w xdirect injection nebulizer DIN 12 . This devicealso eliminated the spray chamber. Conventionalspray chamber-pneumatic nebulization has beenused with diluents such as tetramethyl ammoniumŽ . w xhydroxide TMAH14 , ethylenediaminete-Ž .

w xtraacetic acid EDTA in TMAH 18 , 25% HClw xwxwith cysteine 19, or gold 13,15,20 to modifysamples to reduce the memory effect. RecentlyHBr has been used to reduce mercury memory inw xspray chambers 11 . All of these solutions appearto work but, for routine clinical applications, wefavor the simple dilution approach coupled withconventional spray chamber-pneumatic nebuliza-tion. While the DIN appears to give excellentsensitivity with very small samples, clogging might

present problems with whole blood samples. Cer-tainly the ffow injection-gold trap-ICP-MS methodeliminates matrix effects and memory problems,

but the method is as slow as CVAAS. The mostviable and compatible methods with routine clini-cal analysis involve only simple dilution.

Our approach to the determination of mercury by simple dilution-pneumatic sample introductionis based on the knowledge that dichromate hasw x

been used as a modifier-oxidant for CVAAS 21w xand graphite furnace mercury determination 22and dichromate has also been effectively used asw xa preservative for mercury standard solutions 16 .When dichromate is present in solution, mercury

w xtiter does not change with time 16 . The addition



of dichromate to acidified mercury standards pre-vents both the volatilization and adsorption lossesw xof mercury 16. Since dichromate is a strongoxidant, compounds such as mercurous sulfide

Ž .cinnabar are completely solubilized and remainw xin solution 21 . With dichromate present, reduc-tion of mercury to insoluble compounds such asqŽ .0Hg

calomel or volatile Hg is prevented. Inthis paper we describe the use of dichromate in2% HCl as a diluent for mercury analysis ofwhole blood and urine. HCl was chosen becauseinitial trials with 1% HNO showed that mercury3

Page 3

( )

D.E. Nixon et al.r Spectrochimica Acta Part B: Atomic Spectroscopy 54 1999 1141 1153 1143memory effects are severe. In addition HCl isw xeffective for organic mercury solubilization 23w xand total mercury determination by ICP-MS 19 .In addition to the combination of dichromatewith HCl, gold was evaluated to determine if it istruly effective with biomedical samples. Gold wastried with and without added dichromate.

A rapid clearing spray chamber using a highvolume wash solution combined with high volumegas ffush has been designed to suppress matrixw xeffects and increase sample throughput 24 . Theliquid and gas ffush combination reduced analytesignals for iron and copper over four orders ofw xmagnitude in less than 20 s 24 .The use of a high volume liquid ffush alone wastried to help drive mercury from the spray cham-

ber between samples. In this ‘fast pump’ mode a microprocessor was used to switch the sampling



pump from normal 1 mlrmin. to the ‘rabbit’ mode between samples. The spray chamber was

purged with approximately 2.5 mlrmin of diluent between samples.Results for the ICPMS analysis of whole blood

and urine were compared with our ffow injectionw xCVAAS method 9 . Commercially available qual-ity control samples, interlaboratory comparisonsamples from the Centre de Toxicologie du Que-

bec, and patient samples were compared by bothtechniques.

2. Experimental 2.1. ICPMS facility All ICP-MS experiments were performed on aPerkin-Elmer Sciex Elan 5000A instrument

equipped with an AS-90 autosampler. Nickel sam- pler and skimmer cones were used for all experi-ments. Typical operating parameters are listed inTable 1. The mass ffow controller for nebulizergas ffow, aerosol injector orifice position relativeto the sampling cone, and mass spectrometer lenssettings were adjusted on a weekly basis. Counts

per second were maximized for Rh, Mg, and Pb by sequentially adjusting the above parameterswhile a solution containing 10 grl of each ana-lyte in 1% HNO was nebulized into the plasma.3Typical counts for Rh after optimization wereq6Ž . Ž .5=10counts=l r s=mg . Peak widths werealso adjusted at the same time to maintain a0.7"0.1 amu peak width for 103Rh and a maxi-

Table 1ICPMS Conditions for the analysis of urine and whole blood for mercuryInstrumentPerkin-Elmer Elan 5000APower1000 WPlasma gas ffowrate15 lrminIntermediate gas ffowrate0.8 lrmina

Nebulizer gas ffowrate0.8 lrmin



NebulizerPerkin-Elmer Gem Tip Cross FlowTMSpray chamberRyton

Double PassCones

NickelResolution0.7"0.1 amuIsotopesHg sum of 198, 199, 200, 201, 202; Bi 209

bPb sum of 204 , 206, 207, 208Dwell time100 ms

Sweeps25 per readingMode

Normal resolutionSample preparationUrine: 1:10; whole blood: 1:20 dilutionDiluent2% hydrochloric acid; 57 mg K Cr O rl227Internal standard200 g Birla Nebulizer gas ffow changes with optimization.

b Pb 204 corrected for Hg 204.

Page 4

( ) D.E. Nixon et al.r Spectrochimica Acta Part B: Atomic Spectroscopy 54 1999 1141 1153 1144

mum difference in peak width between 24 Mg and208 Pb of 0.02 amu.2.2. Isotopes The sum of masses 198, 199, 200, 201, and 202were used for mercury. Lead 204 corrected formercury, 206, 207, and 208 were summed andused for memory effect comparisons. Bi 209 wasused as the internal standard.2.3. Reagents Stock solutions of Hg, Pb, Au, and Bi were

purchased from J.T. Baker Chemical Company

Ž .



Phillipsburg, NJ as Baker Instr-Analyzed AtomicSpectral Standards. Concentrations were nomi-nally 1000grml except gold which was 500grml. GR Volumetric Grade K Cr O and GR

227Grade hydrochloric acid was purchased from EMŽ .Science 480 S. Democrat Rd., Gibbstown, NJ .Reagent quality water was produced by process-ing laboratory-distilled water through a BarnstedŽ

NANOpure treatment system Barnsted-Thermo-

.lyne; Dubuque, IA .2.4. Laboratory ware The following disposable laboratory items wereŽ used: Oxford pipette tips, no. 885-091341 Mono-

ject Scientific, Div. of Sherwood Medical, St..Louis, MO ; polystyrene 75=12-mm 5-ml tubesŽ .

Ž no. 55.476 and 6.0-ml screwcap vials no..Ž .61.542 Sarstedt, Inc., Princeton, NJ . All itemswere used with no additional cleaning.2.5. Certified controls Specimens with pre-analyzed mercury concen-trations were purchased from the National Insti-tute of Standards and Technology; Gaithersburg,

Ž MD, USA NIST-SRM 2670 toxic metals in.freeze-dried urine; normal and elevated levelsand Bio-Rad, ECS Division, Anaheim, CAŽLyphochek urine metals control; level 2; two .lots . Survey whole blood and urine samples wereobtained by participation in Interlaboratory qual-ity control program from the Centre de Toxi-cologie du Quebec.

2.6. Patient samples Patient samples sequestered for method com-



Urine samples were diluted 1:10 with the dichro-mateHCl diluent and whole blood was diluted1:20. Whole blood samples were vortexed tothoroughly mix diluent with the blood but notcentrifuged. Minimal cellular debris and no pro-

tein denaturing was produced by 1% HCl.2.9. Memory effect experiments The effectiveness of dichromate, gold, and the‘fast pump’ purge mode was determined by taking readings at 10-s intervals with a dwell time of 100ms and 25 sweeps for the sampling sequencesdescribed as follows. To test equilibrium andwashout, three blanks of the trial diluent weremeasured first. The sample probe was then in-serted into the test solution containing matrix andŽ

.mercury or lead plus the test diluent. The mea-surement sequence was started as the liquid sam-

ple entered the nebulizer. Measurements at 10-sintervals were accumulated until a steady stateŽ .signal was observed 1020 readings . The probewas then inserted back into the test diluent con-Ž .taining no matrix or mercury or lead and themeasurements resumed. This sequence was fol-lowed rigidly for all test solutions so that differ-ences in equilibrium and washout could beobserved and calculated. For the trial of the ‘fast

pump’ purge mode the ‘fast pump’ sequence lasted 20 s then the microprocessor reset the sample

pump to the standard sampling speed of 1 mlrmin.Data was accumulated throughout the blank dilu-ent, sample, washout sequence in this mode also.

2.10. CVAAS facility and method Ž The Perkin-Elmer Perkin-Elmer Corp., Nor-.walk, CT ffow injection cold vapor mercury anal-ysis system consists of the following components:Model 3100 spectrometer equipped with a 19-cmffow cell maintained at 100C; FIAS 200 pumpsystem with a 1-ml sample loop; AS 90 autosam-

pler; EDL System 2 lamp power supply. Thesystem is controlled with a Digital Equipment

DECstation 325c computer and Perkin-Elmercontrolling software. Results are printed on a



Texas Instruments Omni 800rModel 855 printer.A complete description of this instrument andw xmethod is described elsewhere 9 but, the methodwill be described brieffy here.

To every Pyrex 125=16-mm digestion tube, 0.2g K S O , 4 ml of concentrated HNO , and 0.52 283ml H SO were added with thorough mixing.24Blanks, controls, and blood samples were pre-

pared by pipetting 0.5 ml of water, control, orsample into a digestion tube containing the oxida-

tive mixture. Standards were prepared by pipet-ting 0.0, 0.05, 0.10, 0.20, 0.50, 1.0, and 2.5 ml ofthe 10 ng Hgrml standard into separate digestiontubes containing the oxidative mixture. In addi-tion, 0.5 ml whole blood was added to each stan-dard tube as a matrix. For urine, a sample size of1.0 ml was used. Standards, blanks, controls,

bloods, and urines were placed in the dry blockheater at 95C for 30 min. Digestion was con-sidered complete when foaming stopped andfumes containing oxides of nitrogen appearedabove the digestion mixture. Allowing the diges-tion to proceed longer than 30 min was not detri-mental to the final result. All blanks, standards,controls, and samples were diluted to 9 ml finalvolume with deionized water. Standard CVAASanalysis was performed on these samples using5% stannous chloride prepared in 3% HCl.ICP-MS mercury results were compared withCVAAS results for whole blood and urine sam-

ples.

3. Results and discussion 3.1. Mercury memory effects The normal Elan 5000A sample pump-operat-ing mode is a constant speed irrespective of thesample probe position. Typically the speed is setto deliver 1 mlrmin to the nebulizer. In the ‘fast

pum p’ mode the pump is microprocessor con-trolled so that the pump speed is increased to 2.5mlrmin for 20 s each time the probe enterseither a sample tube or the rinse solution. This‘fast pump’ mode of sampling was first tested with

urine matrix-based solutions containing 2% HClwith and without gold. We evaluated gold first



because the effects of gold on memory have beenw xmuch discussed at scientific meetings 13 and the

Page 6

( ) D.E. Nixon et al.r Spectrochimica Acta Part B: Atomic Spectroscopy 54 1999 1141 1153 1146method has been the subject of literature fromw xinstrument manufacturers 15 and a USEPA en-w xvironmental method 20. Our results are dis-

played in Fig. 1. for ‘fast pump’ vs. normal sample delivery and for the addition of gold to a urinematrix containing 5 g Hgrl. Gold was added at1 grml. Gold does enhance the overall mercurysignal somewhat and the rise time of the mercurysignal is shorter with gold irrespective of the

pump speed. The presence of gold in the blankrinse solution does increase the mercury blank signal but, the mercury enhancement isgreater than the blank increase. In the washout

part of the experiment it is clear that the pres-ence of gold offers no advantage over HCl alone.

The use of ‘fast pump’ to ffood the s pray chamberwith blank-rinse solution is more effective at sig-nal reduction than the presence of gold. The ‘fast

pump’ mode reduces the washout time by half. In the normal sample delivery mode the signal re-turned to near blank levels at 140 s. With the ‘fast

pump’ only 70 s was required. Use of nebulizer spray chamber wash to reduce memory is notnew. The development of a rapid clearing systemusing a combination of wash solution and highw x

volume gas ffush has been described 24. Ironand zinc signals were reduced by 50000= inw xapproximately 20 s using this system 24 . While a‘fast pump’ feature is standard on an ELAN 6000 ICP-MS, the system described here was designedfor the ELAN 5000A by the Mayo Clinic Sectionof Engineering. The data in Fig. 1 is a compositeof at least three experiments for each diluentcombination.In the next experiment HCl, HClqAu, HClq

Cr O2y, and HClqAu and C O2y were evalu-2



727ated in a urine matrix with mercury at 1.4 and13.6 grl. ‘Fast pump’ sample delivery was used

Fig. 1. Rise time, equilibrium, and washout for mercury ICP-MS signal with and without goldusing normal sampling pump speedand ‘fast pump’.

Page 7

( ) D.E. Nixon et al.r Spectrochimica Acta Part B: Atomic Spectroscopy 54 1999 1141 1153 1147Ž .Ž .Ž . Ž . Ž . Fig. 2. A, B Rise time, equilibrium, and washout for mercury in a urine matrix at 1.4 grl Aand 13.6 grl B in B HCl, qŽ . Ž . HClqAu;HClqdichromate, andHClqdichromateqAu using ‘fast pump’ only.

Page 8

( ) D.E. Nixon et al.r Spectrochimica Acta Part B: Atomic Spectroscopy 54 1999 1141 1153 1148for all experiments. Fig. 2A,B show the raw sig-nals with time for each diluent and blankrinsecombination. When dichromate is present, themercury signal reaches steady state within 50 s ofsample introduction whether gold is present or

not. With HCl alone or with gold present themercury signal continually increases. This is lessevident in Fig. 2B with a higher mercury concen-tration. With HCl and gold, the washout takes atleast 220 s. Even at 220 s some residual mercurysignal can be seen in both figures. Memory andwashout times are significantly less with dichro-mate present. When dichromate is present thesignal drops to near baseline in 30 s or less at 1.4g Hgrl and approximately 60 s with 13.6 gHgrl. Memory effects are reduced whether goldis present with the dichromate or not. It appears



that dichromate is the active agent for mercurymemory reduction.Dichromate was chosen for evaluation becauseit is routinely used to prevent titer change inw x

mercury standard solutions 16 . It has been shownthat acidification of solutions prevents loss ofmercury by adsorption; however, acidificationw xalone does not prevent loss by volatilization 16 .When dichromate alone was added to mercurysolutions loss by volatilization was prevented but,w xloss by adsorption was not 16 . The combinationof dichromate and acid prevented loss by bothw x

mechanisms 16 . It should be added that gold has been shown to be as effective as dichromate inw xsolution loss prevention 16 . In our trials usingaerosols however, dichromate is much more ef-fective at controlling mercury memory effects thangold.Dichromate was chosen for evaluation becauseit has been used in the oxidative digestion ofw x

biomedical samples prior to CVAAS 21. Theaddition of dichromate to digestion solutions al-lowed the complete solubilization of insolubleŽ . Ž . Fig. 3. Washout for mercury and lead with and without HCl-dichromate added to the argonaerosol gas. B 10 g Hgrl. q 10g Pbrl.

Page 9

(

) D.E. Nixon et al.r Spectrochimica Acta Part B: Atomic Spectroscopy 54 1999 1141 1153 1149Ž . w xmercury compounds such as cinnabar HgS 21 .Dichromate is also listed as a matrix modifier inthe Zeeman graphite furnace atomic absorptionŽ .spectrometric ZGFAAS determination of mer-

w xcury 22 . With dichromate present a char tem-



perature of 250C was used without mercury lossw x22 . In effect, dichromate prevents the reductionof mercury to the metal or mercury precipitates such as sulfides and calomel. We assume thatdichromate prevents reductive volatilization and does not allow appreciable mercury vapor or

insoluble compounds to form when mercury solutions are sprayed into a spray chamber.In our last experiment Hg and Pb memory were compared after introduction of 10 grl of eachanalyte. Fig. 3 shows the data with the baseline expanded so that the washout data is morevisible. In this experiment, a urine matrix containing 10 g Hg and Pbrl was aspirated afterwhich the liquid ffow was stopped. Only argon carrier gas was ffowing through the sampleintroduction system. Data was collected for 100 s before the ffow of blank-rinse solution wasresumed. It can be seen that the lead signal dropped to near blank level within 20 s after theliquid ffow was stopped.Mercury signal did drop when the liquid ffow was stopped, but not to the blank level. Someresidual signal remained until the HCldichromate liquid ffow was resumed. Within 10 s afterthe liquid ffow was resumed, the mercury signal dropped significantly. Within 50 s the signal

dropped to the lead signal level. With resumption of liquid ffow, the lead signal wasessentially unchanged.While this experiment does not tell us whether mercury is bound to the spray chambersurface or whether mercury simply resides in the dead volume, it does show that dichromateis effective at removing the residual mercury and reducing memory.In summary, no data for nitric acid is presentedw x

because of poor recovery 19 . We also observedextensive memory effects when 1% nitric acid wasused in the diluent. Gold has been used effec-w xtively in mercury solution preservation 16 but, inspray chamber memory reduction, gold is not aseffective as dichromate. Washout times do varywith mercury concentration. At 13.6g Hgrlwashout was 60 s. At 1.4 g Hgrl washout was 30s or less. Others have reported washout times ofw xw xw x

45 s 14 , 130 s 18 , and 180 s 19 . For quantita-tive analysis 90 s between samples seems ap- propriate.3.2. Mercury quantitation 3.2.1. Detection limits Detection limits were determined on 4 separatedays in routine analytical runs. The standard usedcontained 5g Hgrl in a urine matrix. Thedetection limit is defined as the mercury concen-tration calculated to produce a signal three times

the standard deviation of the blank readings. Inthis experiment 12 blank readings were used to



calculate the standard deviation of the back-ground. The average detection limit was 0.15 gHgrl in the original sample before dilution. Ouraverage detection limit is comparable to otherICPMS methods using solution nebulization

Table 2Day-to-day results for the analysis of QC samples by ICPMS and CVAASaaaControlCertified valueICPMS resultsCVAAS results

bŽ .

NIST 2670 Normal21.7"0.6

Not Analyzed NIST 2670 Elevated105"898.6"4.986.6"1.9cŽ .Bio-Rad Lot 4450274.1 range: "14.867.9"8.671.2"2.8cŽ .Bio-Rad Lot 6090272.5 range: "14.869.3"3.4

Not AnalyzedBlood-inorganic spike20 Nominal added23.2"1.622.6"1.9Blood-methyl mercury20 Nominal added29.5"3.028.6"3.5aAll values are microgram per liter; all other"values are 1 S.D.

b Information value only.

cAcceptable range listed, not S.D.



Page 10

( ) D.E. Nixon et al.r Spectrochimica Acta Part B: Atomic Spectroscopy 54 1999 1141 1153 1150

Ž .Ž . Ž . Fig. 4. A, B Comparison of ICP-MS results for whole blood A and urine B with targetconcentrations for Quebecinterlaboratory comparison samples.

Page 11

(

) D.E. Nixon et al.r Spectrochimica Acta Part B: Atomic Spectroscopy 54 1999 1141 1153 1151wx14,15,18,19 .3.2.2. Quality control Day-to-day results for several urine and whole

blood controls are shown in Table 2. Results for both ICPMS and CVAAS methods compare fa-vorably with certificate values for urines. Inor-

ganic and methyl mercury blood controls were prepared in house. Whole rabbit blood was sup- plemented with inorganic mercury and methylmercury chloride at 20 grl. The whole bloodcontrols were digested with nitric acidsulfuricw xacidpersulfate before CVAAS analysis 9 . ForICPMS analyses whole blood controls were di-luted 1:20 with the HCldichromate diluent andvortexed to thoroughly mix the blood with dilu-ent. The diluted controls were aspirated directly

into the nebulizer without further manipulation. No CVAAS results are shown for the more recentBio-Rad urine control. The CVAAS method wasno longer in use when this Bio-Rad lot was evalu-ated. NIST 2670 Normal Level was analyzed bythe ICPMS only.3.2.3. Interlaboratory comparisons Urine and whole blood interlaboratory compar-ison samples from the Centre de Toxicologie duQuebec were analyzed by ICP-MS and CVAAS.The results for ICP-MS are shown in Fig. 4A forwhole bloods and Fig. 4B for urines. High and



low acceptable analysis limits are also drawn inthe figures. ICP-MS gives results that correlatewell with the certified concentrations. In general,

blood results are close to the identity line buturine results are somewhat low. In all cases few

results are outside the acceptable limits.3.2.4. Patient samples Our evaluation of the ICP-MS method in-cluded the analysis of routine whole blood andurine samples by ICP-MS and CVAAS. For urine,456 samples with concentrations from - 1grspecimen to 25 grspecimen were analyzed

by both techniques. The resulting regressionequation is: ys0.93 xq1.0 with a correlationcoefficient of 0.8763. For whole blood, 241 sam-

ples with concentrations up to 35

grl wereFig. 5. Comparison of ICP-MS with CVAAS results for urine samples positive for mercury.

Page 12

( ) D.E. Nixon et al.r Spectrochimica Acta Part B: Atomic Spectroscopy 54 1999 1141 1153 1152analyzed and the equation is: ys1.1 xq0.2 witha correlation of 0.9357.

A group of 30 urines found to be in the normalrange by CVAAS were sequestered and analyzedin duplicate by CVAAS and ICP-MS in a 2-week

period. The average deviation in duplicates byCVAAS was 1.2 grl. The largest difference inresults was 4grl. Results ranged from -1grl to 10 grl. The average deviation for ICP-MS duplicates was 0.7 grl with the largest dif-ference of 3 grl. Results for the same samplesranged from -1 grl to 14 grl. Correlation of

ICP-MS with CVAAS was ys1.01 xq0.3 with acorrelation coefficient of 0.8316.In a final experiment a group of urines positivefor mercury by CVAAS were analyzed by ICP-MS.The results are shown in Fig. 5. The figure showsthat ICP-MS compares very well with CVAAS.

4. Conclusion We conclude that dichromate plus hydrochloricacid as a diluent and in the wash solution iseffective at minimizing mercury memory. We fur-ther conclude that the ‘fast pump’ mode used in

the wash cycle facilitates mercury washout. The‘fast pump’ mode used at the beginning of a



sample reading cycle facilitates mercury signalequilibrium. The use of dichromate is compatiblewith heavy metal screening for Cd, Pb, Tl, and As.Mercury can be incorporated into the screeningmethod for these other metals.

References w x1 G. Lindstedt, I. Skare, Microdetermination of mercuryin biological samples. Part II. An apparatus for rapidautomatic determination of mercury in digested sam-Ž .

ples, Analyst 96 1971 223229.w x2 W.R. Hatch, W.L. Ott, Determination of submicrogramquantities of mercury by atomic absorption spectropho-Ž .

tometry, Anal. Chem. 40 1968 20852087.w x3 M.P. Stainton, Syringe procedure for transfer ofnanogram quantities of mercury vapor for ffamelessatomic absorption spectrophotometry, Anal. Chem. 43Ž .1971 625627.w x4 A.A. El-Awady, R.B. Miller, M.J. Carter, Automatedmethod for the determination of total and inorganicmercury in water and wastewater samples, Anal. Chem.Ž .48 1976 110116.w x5 C.P. Hanna, J.F. Tyson, Determination of total mercuryin water and urine by ffow injection atomic absorptionspectrometry. Procedures involving on- and off-line oxi-Ž .dation of organomercury species, Anal. Chem. 65 1993653656.w x6 B. Welz, D.L. Tsalev, M. Sperling, On-line microwave

sample pretreatment for the determination of mercuryin water and urine by ffow-injection cold-vapor atomicŽ .absorption spectrometry, Anal. Chim. Acta 261 199291103.w x7 D.E. Nixon, G.V. Mussmann, T.P. Moyer, Inorganic,organic, and total mercury in blood and urine: coldvapor analysis with automated ffow injection sampleŽ .

delivery, J. Anal. Toxicol. 20 1996 1722.w x8 L. Magos, Selective atomic-absorption determination of



inorganic mercury and methylmercury in undigested bio-Ž .logical samples, Analyst 96 1971 847853.w x9 L. Magos, T.W. Clarkson, Atomic absorption determina-

tion of total, inorganic, and organic mercury in blood, J.Ž .AOAC55 1972 966971.w x10 D.E. Nixon, T.P. Moyer, Heavy metal screening by in-Ž .ductively coupled plasma mass spectrometry ICP-MS :routine clinical determination of lead, arsenic, cadmium,and thallium in urine and whole blood, Spectrochim.

Ž .Acta 51 1996 1325.w x11 A. Bohlke, W. Bahmet, C. Thomas, M. Hamilton, Win-ter Conference on Plasma Spectrochemistry. Scottsdale,AZ, January 8, 1998w x12 D.R. Wiederin, F.G. Smith, R.S. Houk, Direct injectionnebulization for inductively coupled plasma mass spec-Ž .trometry, Anal. Chem. 63 1991 219225.w x13 D.E. Dobb, J.T. Rowan, D. Cardenas, Determination ofMercury by ICP-MS, Proceedings, The Tenth AnnualWaste Testing and Quality Assurance Symposium, Ar-lington, VA, July 1115, 1994.w x14 E. Bakowska, D. Potter, The Determination ofCadmium, Lead and Mercury in Whole Blood by Induc-

tively Coupled Plasma Mass Spectrometry, Hewlett-Packard Application Note 228-301, 1995.w x15 Determination of Mercury by ICP-MS, Varian Ultra-Mass Application Report, Chic008, 1996.w x16 J.M. Lo, C.M. Wal, Mercury loss from water duringstorage: mechanisms and prevention, Anal. Chem. 47Ž .1975 1869.

w x17 E. Debrah, E.R. Denoyer, J.F. Tyson, Flow injection



determination of mercury with preconcentration byamalgamation on a gold-platinum gauze by inductivelycoupled plasma mass spectrometry, J. Anal. Atom. Spec-Ž .

trom. 11 1996 127132.w x18 J.A. Moreton, H.T. Delves, Simple direct method for thedetermination of total mercury levels in blood and urineand nitric acid digests of fish by inductively coupled

plasma mass spectrometry, J. Anal. Atom. Spectrom. 13Ž .1998 659665.

Page 13

( ) D.E. Nixon et al.r Spectrochimica Acta Part B: Atomic Spectroscopy 54 1999 1141 1153 1153w x19 R. Kalamegham, K.O. Ash, A simple ICP-MS procedurefor the determination of total mercury in whole bloodŽ .and urine, J. Clin. Lab. Anal. 6 1992 190193.

w x20 Determination of Trace Elements in Waters and wastes by Inductively Coupled Plasma-Mass Spectrometry, EPAMethod 200.8, Revision 5.4, EMMC Version, 1994.w x21 S. Landi, F. Fagioli, The adaptation of the dichromatedigestion method for total mercury determination bycold-vapor atomic absorption spectrometry to the analy-sis of soils, sediments and sludges, Anal. Chim. Acta 298Ž .

1994 363374.w x22W. Slavin, G.R. Carnrick, D.C. Manning, E.Pruszkowska, Recent experiences with the stabilizedtemperature platform furnace and Zeeman backgroundŽ .correction, At. Spectrosc. 4 1993 6986.w x23 D. Beauchemin, K.W.M. Siu, S.S. Berman, Determina-

tion of organomercury in biological reference materials by inductively coupled plasma mass spectrometry using



Ž .ffow injection analysis, Anal. Chem. 60 1988 25872590.w x24 G. Legere, E.D. Salin, A fast clearing spray chamber for

Ž .ICP-AES, Appl. Spectrosc. 48 1994 761765.

mercury in blood _02.docx

Documents

Transcript of mercury in blood _02.docx