TM

Trace Analyses in MetalMatrices Using the ELAN DRC II

ww

w.p

erk

ine

lme

r.c

om

AP

PL

IC

AT

IO

N

NO

TE

Introduction

Analyses of matrices containinghigh levels of metals present achallenge for ICP-MS. First, theconcentrations of metals beingintroduced into the instrumentmay be high – up to 1,000 ppm.These high concentrations can causeinstrumental drift resulting fromdeposition on the interface or othercomponents. Also, the matricesmay be complex due to both thecomposition of the materials (i.e.,alloys) and the digestion solutions.

The usual interferences encounteredwith metal matrices are oxidesand argides of the matrix species.However, the resulting interferencesmay be unusual compared to whatis normally encountered with ICP-MS. For example, chloride isusually the major interferingspecies on As+ (ArCl+), but in acobalt matrix, CoO+ is the majorinterfering species for As+. Also,interferences may form on elementswhich are normally interference-free, such as Rh which suffersfrom the CuAr+ interference in acopper matrix. Additionally,because the matrix species arepresent at high concentrations, theinterferences are usually significant.

Another problem posed by metalmatrices is that the desired analyte

levels are usually very low. Afterdigestion and dilution, desiredanalyte levels are often low ppb (µg/L) to ppt (ng/L) levels. Theselow analyte levels, in combinationwith large interferences, present a challenge for the analysis of highmetal matrices.

In this work, we show how DynamicReaction Cell™ (DRC™) ICP-MScan successfully perform low-levelanalyses in samples containinghigh metal concentrations.

Experimental

The instrument used in this workwas the ELAN® DRC II. Specificoperating conditions appear inTable 1. A highly energetic plasmais desirable to break down thematrix species; this condition was achieved by using high Radio Frequency (RF) power incombination with a low uptakenebulizer. Method-specific condi-tions for different matrices are discussed later.

All matrix solutions were madefrom 1,000 or 10,000 mg/L (ppm)single-element standards; analytesand internal standards werespiked into the matrix solutions at appropriate levels.

Authors:

Kenneth R. NeubauerMatthew A. KnoppLeon DavidowskiLuc Dionne

PerkinElmer Life and Analytical Sciences710 Bridgeport AvenueShelton, CT 06484 USA

Table 1. Instrumental Operating Conditions

Parameter Setting

Nebulizer PFA-100 (ESI, Inc., Omaha, NE)

Sample Uptake Rate ≈175 µL/min (self aspiration)

Spray Chamber Cyclonic

RF Power 1500 W

Nebulizer Flow Set for ≤1.5% oxides (CeO+/Ce+)

Results and Discussion

Cleanliness of the Reaction Cell

It may be expected that high-levelmetal matrices would contaminatethe reaction cell. However, this is notthe case – the reaction cell remainsclean and does not become contami-nated. The primary reason for this isthe presence of an electrically neutralshadow stop positioned in-linebetween the orifice of the skimmercone and the opening to the reactioncell, as shown in Figure 1. Becauseof this placement, non-ionizedspecies collide with and deposit onthe shadow stop, thus keeping thereaction cell clean. Neutral species,which expand around the shadowstop, deposit on the lens and differential aperture located beforethe entrance to the reaction cell.Because of the design and placementof these components, the reaction cellremains clean, and the reaction cellparameters remain constant.

Cadmium in a Molybdenum Matrix

Determining trace levels of cadmiumin the presence of high molybdenumlevels is difficult due to the formationof molybdenum oxide, whose isotopesoverlap all the cadmium isotopes.This type of analysis is important inthe metallurgical industry becausemolybdenum is a major componentin steels and alloys.

To solve this problem, O2 is used as a reaction gas in combination withthe appropriate bandpass setting.Figure 2 shows a reaction gas profilewhich demonstrates the removal ofthe MoO+ interference on 114Cd+, themajor cadmium isotope. In this figure114Cd+ intensity is plotted versus O2gas flow. The blue line represents thesignal resulting from a solution of 28 ppm Mo, while the green lineshows the signal from a solution of28 ppm Mo + 0.5 ppb Cd. At O2flows < 1 mL/min, the lines overlapindicating an interference. However,at higher flows, the lines diverge;

the difference between the lines represents the signal due to Cd inthe presence of Mo. At an O2 flow of2 mL/min, the signal due to Mo(MoO+) is about 50 cps, while thatresulting from 0.5 ppb Cd is about700 cps. This difference results fromthe different reactivities of thespecies present: MoO+ reacts readilywith O2 to form MoO2

+, while Cd isunreactive with O2 and is thus unaf-fected by the gas flow.

The reaction profile in Figure 2 wasacquired with a bandpass setting(RPq) of 0.75. At this setting, Mo+ isexcluded from the reaction cell sothat the only Mo species entering the cell is MoO+ generated from theplasma. This latter species reactswith O2 in the cell to form MoO2

+.

Figure 1. Schematic diagram of the ELAN DRC II showing how the reaction cell remains clean.

Figure 2. Cell gas optimization plot for Cd in a Mo matrix. The y-axis represents intensity at114Cd+; the x-axis represents O2 gas flow in mL/min. The green line represents the signals froma solution of 28 ppm Mo. The blue line represents the signal from a solution of 28 ppm Mospiked with 0.5 ppb Cd. The red line is the estimated detection limit (EDL). This data wastaken with a bandpass parameter (RPq) of 0.75.

Neutralspecies

from plasma

AutoLens DRC

Differential aperture

2

GroundedShadow stop

These reaction cell conditions(O2=2.0 mL/min; RPq=0.75) wereapplied to the determination of Cd ina cobalt alloy containing Mo as partof the matrix. After digestion anddilution, the Mo concentration enteringthe instrument was 28 ppm, whilethe desired Cd analytical level (aftersample preparation) was 0.2 ppb. An external calibration curve in 1%HNO3 (no matrix matching) wasused, and the internal standard wasRh (10 ppb).

Table 2 shows the results of thisanalysis. First, the sample matrixwas analyzed 3 consecutive times inboth standard and DRC modes. Instandard mode, the Cd concentrationreads about 30 ppb, while in DRCmode, Cd reads 0.055 ppb. Theseresults show that a large interferenceexists, which is eliminated with theDRC. A 0.2 ppb Cd spike was thenanalyzed, and the recoveries calculated.In DRC mode, the recoveries are 95%. These results demonstrate thesuccessful determination of low-levelCd analysis in the presence of a highMo level.

Molybdenum in an Iron Matrix

In matrices containing high quantitiesof iron, such as steels and alloys,low-level molybdenum determinationis difficult due to the presence ofArFe+ which overlaps the major Moisotopes. To remove the effect of thisinterference, Mo+ can be moved to anew analytical mass. In the previoussection, it was seen that Mo reactswith O2 to form MoO2

+ (m/z 130);therefore, this reaction was used tomove Mo to an analytical mass awayfrom the interference.

Figure 3 shows a cell gas optimizationplot of MoO2

+ in the presence of ironwith RPq=0.45. The matrix is 1,000ppm Fe, and the spike level is 1 ppb

Figure 3. Cell gas optimization plot for Mo in an Fe matrix. The y-axis represents intensity atMoO2

+(m/z 130); the x-axis represents O2 gas flow in mL/min. The green line represents thesignals from a solution of 1,000 ppm Fe. The blue line represents the signal from a solution of1,000 ppm Fe spiked with 1 ppb Mo. The red line is the estimated detection limit (EDL). Thisdata was taken with a bandpass parameter (RPq) of 0.45.

Mo. The formation of MoO2+ is

observed as O2 is introduced into thereaction cell, yet the signal from theiron matrix remains low. MaximumMoO2

+ sensitivity occurs at an O2flow of 0.35 mL/min.

These reaction cell conditions wereapplied to the quantitative determi-

Table 2. Quantitative Determination of Cadmium in a Molybdenum Matrix

Matrix 0.2 ppb Cd Spike Recovery

Sample # DRC Std Mode DRC Std Mode DRC Std Mode(ppb) (ppb) (ppb) (ppb) (ppb) (ppb)

1 0.054 32.0 0.242 32.1 94% 50%

2 0.059 31.6 0.259 32.0 100% 200%

3 0.056 32.3 0.236 32.3 90% ---

Average 0.056 32.0 0.246 32.1 95% ---

nation of Mo in an iron matrix. A10,000 ppm Fe standard containing 1ppb Mo was diluted ten times to give a sample consisting of 100 pptMo and 1,000 ppm Fe. Quantitativeanalysis was performed using externalcalibration standards in 1% HNO3(no matrix matching) with In (10 ppb)as an internal standard.

3

Table 4. Reaction Cell Conditions for Multielement Analysis in a Copper Matrix

Analyte m/z Reaction Gas Gas Flow RPq(mL/min)

As 75 --- --- 0.25Ag 107 --- --- 0.25Cd 111 --- --- 0.25Sb 121 --- --- 0.25Pb 208 --- --- 0.25Bi 209 --- --- 0.25Rh 103 NH3 0.60 0.75Te 128 NH3 0.60 0.60

Figure 4 displays the calibrationcurve for 0.1, 0.5, and 1 ppb Mostandards while monitoring MoO2

+

at m/z 130. The linearity of the curvedemonstrates that the formation ofMoO2

+ is reproducible over an orderof magnitude. This is expected sincethe reaction Mo+ + O2 ➝ MoO2

+ isgoverned by kinetics and thermo-dynamics and, therefore, will always be constant. The quantitative resultsappear in Table 3. The data showsthat 0.1 ppb Mo can be consistentlydetermined in the presence of 1000 ppm Fe.

Multielement Analysis in a Copper Matrix

Determining trace contaminants inhigh purity metals is important sothat final products made from thesemetals function properly. Majorinterferences in these matrix typesresult from matrix oxides, matrixargides, and matrix doubly-chargedspecies. Elements commonly determined in a copper matrixinclude As, Ag, Cd, Sb, Pb, Bi, Rh,and Te. After sample preparation, the desired analytical level for theseanalytes is 100 ppt in a 1,000 ppmCu matrix. Expected matrix-relatedinterferences are CuAr+ on Rh andCu2

+ on Te; other elements are interference-free. Table 4 shows thereaction cell conditions used for thisanalysis. NH3 is used to eliminate the interferences on Rh and Te, whilethe interference-free elements areanalyzed in standard mode. All elements are measured using onemethod and one analytical run.

Table 3. Quantitative Determination ofMolybdenum in an Iron Matrix

Sample Mo Concentration(ppb)

1 0.1142 0.1113 0.111

Figure 4. Calibration for 0.1, 0.5, and 1 ppb Mo monitoring MoO2+ (m/z 130).

Reaction cell conditions: O2 = 0.35 mL/min; RPq=0.45.

4

Calibration was performed withexternal standards prepared in 1%HNO3. Ten ppb each of Ga, Y, and Irwere used as internal standards. The1,000 ppm Cu matrix was made bydilution of a 10,000 ppm Cu stan-dard, and 100 ppt of each elementwas spiked into the matrix solution.

Table 5 shows the quantitativeresults for the unspiked 1,000 ppmCu matrix, both at the beginning ofthe run and 65 minutes later after thespiked solution had been run 10times. By comparing the twocolumns, it is evident that the measured concentrations did not change. This result shows thestability of the method and indicatesthat the reaction cell is not beingcontaminated during the analysis. By comparing the black and bluerows for Rh and Te, it is evident that interferences exist on both thesespecies in the standard mode, yet the DRC mode removes the interfering species.

Spike recovery data is presented inFigure 5. The spiked sample was run10 times over 1 hour after theunspiked sample had been run.These plots further demonstrate thestability of the instrument andmethod by showing that 100 pptspike recoveries vary by less than ± 15% over one hour. These recoveries were calculated relative to the unspiked sample analyzed atthe beginning of the run (Sample 1 in Table 5). A plot of spike recoveriescalculated relative to the unspikedsample analyzed at the end of therun (Sample 11 in Table 5) looksidentical and is not shown.

Figure 5. 100 ppt spike recoveries over 1 hour for several elements in 1000 ppm Cu. Rh and Tewere analyzed in DRC mode (NH3 as reaction gas); the other elements were run in standard mode.

Table 5. Quantitative Results for Multielement Analysis in a Copper Matrix

Analyte Sample 1 Sample 11(ppb) (ppb)

As 0.000 0.000Ag 0.041 0.041Rh 0.001 0.000Rh 21.8 21.7Cd 0.002 0.002Sb 0.009 0.008Te 0.004 0.006Te 0.078 0.095Pb 0.356 0.365Bi 0.020 0.020

Blue = DRC modeBlack = Standard mode

Per

cent

Rec

over

y

5

0 10 20 30 40 50 60Time (min)

160

140

120

100

80

60

40

Ag

Cd

Sb

As

Pb

Bi

Rh

Te

Multielement Analysis of a Cobalt Alloy

Alloy analysis is more complex than high-purity metal matricesbecause of the multiple componentspresent which lead to more possibleinterferences. Table 6 displays thecomposition of a cobalt alloy, as well as the desired analytes and theirlevels, both in the solid and aftersample preparation. From this list,expected interferences are CoO+ onAs+ and MoO+ on Cd+.

Table 7 shows the reaction cell conditions used. Most of the elementsare run in standard mode becausethey are interference-free; the effectsof the interfering species on As+ andCd+ are eliminated using O2 as areaction gas. For this method, O2 isused in two ways: first, to convertAs+ to a new analytical species(AsO+) away from the CoO+ interfer-ence, and second, to remove theMoO+ interference on Cd+.

External calibrations using standardsmade in 1% HNO3 were used. Sc,Rh, In, and Ir were used as internalstandards. The matrix solution wasmade in the laboratory by appropriatedilution of 10,000 ppm standardsand then was spiked with the analytesat their desired analytical levels, asshown in Table 6.

Table 8 displays the quantitativeresults of analyzing the cobalt alloymatrix, both at the beginning of therun and 65 minutes later after thespiked solution had been run 10times. By comparing the two samplecolumns in Table 8, it is evident thatthe measured concentrations do notchange. This result shows the stability of the method and indicatesthat the reaction cell remains cleanduring the analysis.

Table 6. Composition of Cobalt Alloy

Component Solid After Sample Prep (ppm) (ppb)

Bi, Cd, Tl 0.4 0.2Pb, Sb 2 1Te 4 2Sn, As 10 5Mg 200 100

Matrix Solid After Sample Prep (%) (ppm)

Co 60 300Cr 28 140Mo 5.5 28C, Mn, Si, Ni, Fe ≤ 1 ≤ 10

Table 7. Reaction Cell Conditions for Analysis of a Cobalt Alloy

Analyte m/z Reaction Gas Gas Flow RPq(mL/min)

Mg 24 --- --- 0.25Sn 120 --- --- 0.25Sb 121 --- --- 0.25Te 125 --- --- 0.25Tl 205 --- --- 0.25Pb 208 --- --- 0.25Bi 209 --- --- 0.25AsO 91 O2 2.0 0.70Cd 114 O2 2.0 0.75

Table 8. Quantitative Results for Analysis of a Cobalt Alloy

Analyte Sample 1 Sample 11(ppb) (ppb)

Mg 2.47 2.61Sn 0.190 0.171Sb 0.031 0.032Te 0.292 0.263Tl 0.004 0.003Pb 0.328 0.335Bi 0.178 0.188Cd 0.061 0.049As 1.14 0.95

Blue = DRC modeBlack = Standard mode

6

Figure 6. Spike recoveries over 1 hour for several elements in a cobalt alloy matrix. Cd andAsO were analyzed in DRC mode (O2 as reaction gas); the other elements were run in standardmode.

Table 9. Stability of Spike Recoveries in a Cobalt Alloy Over 1 Hour

Analyte Spike Level RSD(ppb) (%)

Mg 100 1.5Sn 5 0.9Sb 1 0.8Te 2 2.9Tl 0.2 1.6Pb 1 0.9Bi 0.1 1.3Cd 0.2 2.5As 5 3.2

Blue = DRC modeBlack = Standard mode

Figure 6 shows spike recovery dataover one hour; the spiked samplewas run 10 times after the unspikedsample had been run. These plotsfurther demonstrate the stability ofthe instrument and method by showing that the spike recoveriesvary by less than ± 15% over onehour. These recoveries were calculated relative to the unspikedsample analyzed at the beginning ofthe run (Sample 1 in Table 8). A plotof spike recoveries calculated relativeto the unspiked sample analyzed atthe end of the run (Sample 11 inTable 8) looks identical and is notshown. Further evidence of the stability is presented in Table 9which shows that the RSD’s (RelativeStandard Deviation) for the spikesdisplayed in Figure 6 are less than or equal to 3%.

Conclusions

This work demonstrates the ability of the ELAN DRC II to perform low-level analyses in samples containinghigh levels of metal species. It hasbeen shown, both conceptually andwith stability data, that the reactioncell does not become contaminateddespite the high-level matrix samplesbeing introduced to the instrument.The flexibility of the DRC has beendemonstrated by showing that theeffects of matrix-derived interferencescan be eliminated by either removingthe interference or by converting theanalytes to new analytical massesaway from the interferences. By utilizing these schemes, ppt levels of many analytes can be successfullymeasured in high metal matrices.

Per

cent

Rec

over

y

7

0 10 20 30 40 50 60 70

160

140

120

100

80

60

40

Mg

Sn

Sb

Te

Tl

Pb

Bi

Cd

AsO

Time (min)

TM

© 2003 PerkinElmer, Inc. All rights reserved. PerkinElmer is a registered trademark of PerkinElmer, Inc. SCIEX and ELAN are registered trademarks of MDS SCIEX, a division of MDS, Inc.All trademarks depicted are the property of their respective holders or owners. PerkinElmer reserves the right to change this document at any time and disclaims liability for editorial, pictorial, or typographical errors.

006765_01 070300 Printed in USA

PerkinElmer Life and Analytical Sciences710 Bridgeport AvenueShelton, CT 06484-4794 USAPhone: (800) 762-4000 or (+1) 203-925-4602www.perkinelmer.com