Effects of acid type and concentration on the ... - ACS Publications

for technical assistance, Lew Rubin of EG&G/Princeton. Applied Research for use of equipment,Ray Pepin and. Wallace M. Rogers of S.D. Warren Co. for ...
0 downloads 0 Views 1MB Size
1242

Anal. Chem. 1985, 57, 1242-1252

Note Added in Proof. Improvements in fiber end preparation have allowed us to increase the laser power incident upon the fiber end to levels beyond that reported here. This observation is consistent with the manufacturer's reported damage threshold. ACKNOWLEDGMENT We thank Timothy M. Woudenberg and George Scherer for technical assistance, Lew Rubin of EG&G/Princeton Applied Research for use of equipment, Ray Pepin and Wallace M. Rogers of S.D. Warren Co. for providing leachate samples, and Karyn Moskowitz of the Harvard Museum of Comparative Zoology for preparation of the electron micrographs. Registry No. Phenol, 108-95-2; o-cresol, 95-48-7; toluene, 108-88-3; o-chlorophenol, 95-57-8; p-nitrophenol, 100-02-7; 2,4dinitrophenol, 51-28-5; xylene, 1330-20-7; water, 7732-18-5.

LITERATURE CITED (1) Peterson, J. I.; Fitzgerald, R. V.; Buckhoid, D. K. Anal. Chem. 1984, 56, 62-67. (2) McCreery, R. L.; Fleischmann, M.; Hendra, P. Anal. Chem. 1983, 55, 146- 148. (3) Hirschfeid, T.; Deaton, T.; Milanovich, F.; Kiainer, S. Opt. Eng. 1983, 22 (5),527-531. (4) Seitz, W. R. Anal. Chem. 1984, 56, 16A-34A. (5) Keith, S. J.; Wilson, L. G.; Fitch, H. R.; Esposito, D. M. Ground Water Monif. Rev. 1983, 3 (2) 21-32. (6) Hirschfeld, T.; Deaton, T.; Milanovich, F.; Klainer, S. M.; Fitzsimmons, C. Project Summary, "The Feasibility of Using Fiber Optics for Monitoring Groundwater Contaminants", Environmental Monitoring Systems Laboratory, US. E.P.A., January, 1984. (7) Dovichi, N. J.; Martin, J. C.; Jett, J. H.; Trkula, M.; Keiier, R. A. Anal. Chem. 1984, 56, 348-354.

M. A. H., Eds. "Standard Methods for the Examination of Water and Wastewater", 15th ed.; American Public Health Association, American Water Works Association, and Water Pollution Control Federation: Washington, DC, 1981; p 774. Beriman, I.B. "Handbook of Fluorescence Spectra of Aromtic Molecules", 2nd ed.; Academic Press: New York, 1971. Optical Systems and Components Catalogs, Oriel Corp., Stratford, CT 06497. Fowles, G. R. "Introduction to Modern Optics", 2nd ed.; Holt, Rinehart and Winston: New York, 1975; Chapter 2. 518A-521A. Josephson, J. Environ. Scl. Technol. 1983, 17 (ll), Hager, D. G.; Smith, C. E.; Loven, C. G.; Thompson, D. W. I n "Proceedings of the Third National Symposium on Aquifer Restoration and Ground Water Modeling"; Nielsen, D. M., Ed.; National Weii Water Association: Worthington, OH, 1983; pp 123-124. Willard, H. H.; Meritt, L. L., Jr.; Dean, J. A,; Settle, F. A,, Jr. Instrumental Methods of Analysis", 6th ed.; D. Van Nostrand Co.: New York, 1981; Chapter 4. Ratziaff, E. R.; Harfmann, R. G.; Crouch, S. R. Anal. Chem. 1984, 56, 342-347. Russo, V.; Righimi, G. C.; Sottini, S.; Trigari, S. Appl. Opt. 1984, 23. 3277-3283.

( 8 ) Greenberg, A. E., Conners, J. J., Jenkins, D., Franem,

(9) (10) (11) (12) (13)

(14) (15) (16)

RECEIVED for review December 26, 1984. Accepted February 14, 1985. This research was supported by a grant from Research Corp., NSF Grant PRM-8114621; and a Tufts Faculty Research Award. Part of this work was performed while we were visiting scientists a t the M.I.T. Laser Research Center, Massachusetts Institute of Technology, Cambridge, MA 02139, which is a National Science Foundation Regional Instrumentation Facility. The Mallinckrodt Corp. provided a Graduate Research Fellowship for M.M.C. during part of this work. Part of this work was presented a t the Northeast Regional Meeting of the American Chemical Society, June, 1984, in the Environmental Chemistry session.

Effects of Acid Type and Concentration on the Determination of 34 Elements by Simultaneous Inductively Coupled Plasma Atomic Emission Spectrometry Shane S. Que Hee,* Timothy J. Macdonald,' and James R. Boyle Department of Environmental Health, University of Cincinnati Medical Center, 3223 E d e n Avenue, Cincinnati, Ohio 45267-0056

A mlxed acld conslstlng of 11.6% HC1/2.8% HNO, proved superlor to 2 to 10% HCI, HNO,, and H2S04alone in chemical compatlblllty and storage characteristics for slmultaneous inductively coupled plasma atomlc emlsslon spectrometric (ICP-AES) determlnatlon of 33 elements admlxed up to concentrations of 100 pg/mL each. A 2 % aqua regla solution appeared to be adequate below 10 pg/mL of all these admixed elements plus sllver. Use of the mlxed acld generally also allowed for more reproduclble lnterelementalk factors. Less sensltlve elements and elements whose llnes were in the vacuum ultravlolet were not as reproducible. A two-point standardlratlon procedure was adequate, and k factor values agreed wlthln 10 % only over a speclflc concentratlon range. A practical procedure to deflne the range of determination was developed using the 11.6 % HCV2.8 % HNO, acid solvent.

'Present

ID 83415.

address:

EG&G

I d a h o Inc.,

P.O. Box 1625, I d a h o Falls,

Few investigators have studied the effects of different acids on the chemical compatibility, interelemental interferences, and spectral background for a large number of elements when analyzed by inductively coupled plasma atomic emission spectrometry (ICP-AES). In contrast, there are many publications dealing with the effect of argon flow rate, rf power, and torch height (1-5), sample uptake (4, 5 ) , and determination of interelemental interferences (6-9). The importance of matrix matching to obtain accurate levels has been recognized (6,10,11). Investigators have attempted to obtain accurate answers in various ways: standard addition experiments can be performed; the sample can be diluted to match the matrix used for standards (at the sacrifice of sensitivity); or the matrix of the sample can be simulated artificially once the sample is screened. The matrix consists of contributions both from the acid and from the sample, and since the sample composition is often unique, many investigators have utilized the dilution method for screening purposes often with excellent results. In such a method, the acid solvent should be optimum in terms of chemical compatibility and

0003-2700/85/0357-1242$01.50/00 1985 American Chemical Society

ANALYTICAL CHEMISTRY, VOL. 57, NO. 7, JUNE 1985 ~~

1243

~

Table I. Spectral Lines, Orders, and Spex Solvents for the Elements for the ICAP-9000 Utilized in This Study channeld Ag

A1 As Au B Ba Be Ca1 Ca2 Cd co Cr cu Fe Hg I1 I2 In K Li Mg, Mgz Mn Mo Na1 Na2 Ni P Pb Pt S Sb Se Si, Siz Sn Sr1 Sr, Ti T1 V Zn

wavelength, nm

order

338.2891 308.2150 193.6960 367.5950 249.6778 493.4040 313.0402 393.3660 317.9332 228.8020 228.6169 205.5520 324.7540 259.9402 184.9602 183.0000 206.1903 451.1323 766.4907 670.7844 279.5530 383.2310 257.6102 202.0302 589.5983 330.2370 231.6040 178.2870 220.3530 203.6460 182.0402 217.5810 196.0260 251.6120 288.1510 189.9890 421.5524 460.7331 334.9406 377.5720 292.4020 2 13.8560

1 2 2

2 3 1 2 2 1 2

3 2 2 2 2

3 3 1 1 1 2 1 2 2 1 1

3 2 2 2 2 2 2 2 2

3 1 1 1 1 2 2

acid for Spexn 1000 pg/mL ref (%) v/v 2% HN03 10% HN03 10% HN03 10% HC1 water 10% HN03 10% HN03 10% HN03 10% HN03 10% HN03 10% HN03 10% HN03 10% HN03 10% HN03 10% HN03 waterb waterb 2% HCl 10% "03* 10% HN03 10% HN03 10% HN03 10% HN03 water 10% "OBc 10% HN03C 10% HN03 10% HN03 2% HN03 10% HCl water 20% HCl(g)e 10% HN03 waterb waterb 20% HCl 10% HN03 10% HN03 20% HCl(g)e 2% HN03 10% HN03 10% HN03

OSpex Industries, Inc. 10000 pg/mL. c20000 pg/mL. d l , first channel; 2, second channel. e (g) glass container; polyethylene containers were used for the rest of the elements. solvating power. Most prior studies on the effect of acid type and concentration have dealt only with one acid at a time and the depression of line intensities with increasing acid strength has been widely noted (9, 12-18). Little data for mixed acid systems exist. One study has investigated the effect of HC104 on the response characteristics of elements in HC1 (19)and here it was concluded that 5 % and 10% HC104 increased F e line intensities linearly relative t o HC1 itself. This was also true for such other lines as A1 and Mn (12) and for Ca, Mg, and P (18). For 2.5% HC104, depression in Fe line intensity was observed for HC1 concentrations below 10% and above 20%, with the concentration range between 10 and 20% HC1 showing no effect. These results confirmed those of Dahlquist and Knoll (9). This study considers the effect of the acid matrix on chemical compatibility, interelemental interferences, determination ranges, and background measurements. E X P E R I M E N T A L SECTION Reagents. The nitric and sulfuric acids were Ultrex distilled in Vycor (G. F. Smith), and the hydrochloric acid was Fisher Scientific A-144C; National Bureau of Standards (NBS) water was NBS 1643a; all aqueous metal standards were at least 99.999%

Table I1 std no.

solvent, elements

% (v/v)

A. Spectrally and Chemically Compatible Mixtures of 10 pg/mL Elements Recommended by Allied Analytical Systems water 10% HCl 10% HCl 10% HCl 10% HCl 10% HC1 water

Ba, Ca, Cd, Co, Cu, Mg, Mn, Pb, Sr, Zn Al, Be, Fe, Li, Mo, Na, Ni, Sb, Ti, T1 As, Cr, P, S, Se, Si Hg, V Au, Pt, Sn I

10%

Ag

HN03 10% HC1

K (100 ppm)

B. Spectrally and Chemically Compatible Mixtures of 1000 rg/mL Elements Recommended by Spex Industries, Inc.

I, Mo, S, Si Sb, Sn, Ti 2% HN03 Al, As, Ba, Cd, Co, Cr, Cu, Fe, K, Li, Na, Pb, V Be, Ca, Hg, Mg, Mn, Ni, P, Se, Sr, Zn 10% HN03

1 water 2 20% HCl

3 4

pure (Spex Industries, see Table I); Milli-Q deionized water was used in all dilutions. The liquid argon utilized was of 99.999% nominal purity. Instrument. The Jarrell-Ash (J/A) 9000 utilized in this study was a vacuum instrument equipped with 40 channels for simultaneous detection. An (N 1)channel accessory (J/A Model 82.000 0.5 m Ebert scanning spectrometer) was also available. The analytical wavelengths, orders, and the elements for the array are given in Table I. The instrument was equipped with a spectrum shifter with a maximum range of f0.25 nm on either side of the analytical line. The peristaltic pump was a Rainin Rabbit. Two-point standardization is recommended (0 pg/mL and usually 10 rg/mL) for each channel. Three kinds of background correction can be performed (plus or minus a given spectrum shifter value or no correction, all of which must be specified before analysis). Instrumental Optimization. The quality control procedures for the spectrometer (20) (e.g., dark current, white light, integrator card tests) were performed with the temperature, humidity, and atmospheric pressure noted. The instrument was profiled with a mercury lamp source. The 1000 fig/mL Spex standards were then aspirated (20, 21). The spectrum shifter was utilized to ensure that the intensity maximum was in the middle of each scan with the instrument in profile. After creating an Analytical Control Table (ACT) which contained no background or interelemental interferences ( k values) corrections (20), the initial standardization was performed a t the conditions suggested (30 psi argon pressure, 1.10 kW rf forward power, no peristaltic pump), using the mixed elements recommended by the manufacturer in Table IIA. Standard number 1 in Table IIA was replaced by 10% HCI. Optimization using the intensity mode and aspirating 10 pg/mL CdClz in 10% HCl and 10% HC1 alone ( 4 , 5 , 2 0 ) for 10-s exposure times was then done (20). The optimum rf, torch height, argon pressure, and peristaltic pump flow rate were 1.20 kW, 16 mm above the top edge of the top of the plasma cooling coil, 29 psi, and 1.6 mL/min, respectively. The scandium visualization test was then performed (20). Since the top of the inner pink cone of the plasma was difficult to define, the plasma flow was further manipulated to set the Si value in 10% HC1 to be around 0.050 pg/mL. This procedure allowed reproducible S I N ratios (between 240 and 250) and a longer life for torches. Later on in the investigation after evaluating several torches, the rf power was set at 1.10 kW to allow the power amplifier of the rf generator to last longer at fully resonant conditions (20). S I N ratios were around 300 at a torch height of 14 mm with the other conditions being the same. Over a period of time, the inside of the quartz window of the purged optical path system adjacent to the torch became coated owing to a photochemical reaction. This was caused by variable argon quality and markedly affected SIN ratios, as well as leading

+

1244

ANALYTICAL CHEMISTRY, VOL. 57, NO. 7, JUNE 1985

to frequent cleaning of the window. No trouble was found once a direct indicating silica gel scrubber (Wilkerson X03-02.000), and a charcoal bed filter (Koby Junior No. 3,169,112 Air Purifier/Flow Equalizer) were placed downstream in that sequence from the argon gas cylinder. S I N ratios were very constant after this change. A t ambient conditions of around 23-25 “C, relative humidities 0.1 pg/mL status are summarized in Tables I11 and IV for the elements segregated into nine mixtures selected on the basis of chemical and spectral compatibility. HC1, aqua regia, and HNO, solutions showed no precipitation. Chemical incompatibility was observed for sulfuric acid and was attributable to barium, lead, and strontium. Silver precipitated in 5 and 10% HCl but not in 2 and 5% aqua regia. The channels exhibiting behavior not meeting the 10% percentage accuracy criterion are given in Table III; the number of channels in each category are provided in Table IV. Since the instrument was originally standardized using 10% HC1, the latter should be the best acid medium. This was true for almost all parameters (Table IV). However, 5% HC1 is matched or surpassed overall by aqua regia and by 5 and 10% “OB. Since I,, I,, Naz, and Sn had LQL values >0.1 pg/mL for all the acids as indicated in Table 111,these channels were intrinsically less sensitive. The next most insensitive channels were As, P, and S according to dilution behavior. Thus HC1 was not the optimal solvent for Ca, HNOBwas not for As, Mg, Na, Pt, S, and Si, and aqua regia was not for As, Pt, S and Se. Sulfuric acid gave the same results as 10% HCl only for Au, Cd, K, and Mn, showing that sulfuric acid was inadequate. Increasing aqua regia strength decreased sensitivity (LQL) for Ag, Al, 11,I,, S, Sb, and Si channels, for Naz in HCl, and for Al, Nal, Na2, Sb, and Se in “OB. This could reflect the influence of background since this should be enhanced as acid strength increases. The reverse effect in aqua regia was shown by Cal, Ca,, Nal, Pb, and Srz channels, As, Cal, Caz, I, and I,, Mg,, Na,, P, and Sn in HC1, and Caz, Cu, 11,I,, Naz, P, S, Sil, Siz, Srz,and T1 in “OB. Therefore, high ionic strengths were required to keep the ions corresponding to these channels in solution. As can be seen from Table 111, the slope behavior was unacceptable for P in aqua regia, Ba, Be, I, P, Pb, Se, and Zn in H2S04,Na in “OB, and no elements in HC1, with Ag, however, precipitating in 5 and 10% HCl. The correlation coefficients for the linear regression of observed values vs. expected values were all above 0.9900. Aqua regia was the best solvent in terms of 10 pg/mL status, followed by HNO, (I,), HC1 (I,, P), and H2S04(Ba, Be, I,, P, Pb, Se, Zn), the channels with inadequate characteristics being denoted in parentheses. In terms of dilution behavior, the same elements which were inadequate in HC1 were also inadequate in aqua regia (except for Pt) and in “OB. After 10 pg/mL standards were stored for 4 days, I and P were lower (>lo%)in 2% aqua regia. Similarly, Ag and I were lower in 10% HC1, Be, I, and Ag in 5% HC1, and Hg in 2% HC1. In the same manner, As, Ni, Pt, Se, Sn, and T1 were lower in 2% H N 0 3 and Be in 5% “OB. Similarly Ag, Cu, Sb, Sr, TI, and Zn were also lower in both 2 and 10% HzS04. Thus the media chosen for further study when the elements were to be mixed together were HC1, HNO,, and aqua regia. Linear Range Study. The apparent linear range, slope, standard deviation (SD), and relative standard deviation (RSD) when the observed values were compared with expected values after a known dilution of single elements in 2% aqua regia are provided in Table V. Twenty-four of the channels showed unacceptable linearity (>lo% percentage accuracy from the expected slope of 1)over the range 2 to 1000 pg/mL. The % RSD of the slope over this same range was 0.1 pg/mL Have Larger Than 10% Performance Accuracies for Diluted Segmented Elementsu acid concn, % (v/v)

acid aqua regia

“03

HCl

10 pg/mL status

all concnsb

Ag, As, Cal, Caz, 11, 12,

10

Ag,

2

Al, Au, Cr, Hg, K, Mg,, Mo, Ni, Pt, Sb, Si,, Si,, Sr,, T1, V Au, Be, Li, Nal, Ni, Pt, Si,, Si,, Sr,, Ti, T1 Na1

all concns*

Ag, Al, As, Au, Ba, Ca,, Caz, Hg, 11, I,, K, Mgz,

Ag, As, Ba, Be,

5

HZS04

dilution behavior

slope behavior P

As, 11, Iz, Naz, Ni, P, Pt, S, Se, Sn, T1 Ag, AI, Mg,, Nal, Si,

Ag, Be, P

Gal, Caz, Mgz, Pb, Sr,

Ba, Be, 11,P, Pb, Se, Zn

Al, Ba, Ca,, Caz, Hg, 11, 12, Mgz,

Mg,, Na2, P, S,Se, Sn

10

Nal, Na,, Ni, P, Pb, Pt, Sn, Srz, T1 Cu, Fe, Sil, Siz,Ti, V

2

Cr, Li, Sb, Se

all concnsb

As, 11, Iz, K, Mgz, Na1, Naz, Pt, Pt, S, Se, Si,, Sn

10 5

11, Iz, Naz,

LQL > 0.1 pg/mL

P

12

K

I,, Naz, P, Pb, Se

Al, Au, Ca,, Caz, Co, Cr, Cu, Fe, Hg, 12, Li, Mg1, Mg,, Mn, Mo, Na,, Ni, Pt, Sil, Siz, Sn, Sr,, Ti, V Zn

Ag, As

Na,, P, Pb, Sn, T1 Cr, Si2

Gal, Caz, Co, Cr, Cu, Fe, Hg, 12, Li, Mgi, Mg,, Mn, Mo, Nal, Ni, Pt, Sb, Si,, Si,, Sn, Sr,, SI,,Ti, T1, V

Ag, As, Au, K, Nal, Pt, Si,, Sr,

IZ

Na1

Ag, Ca,, Co, Cu, Ni, Pb,

I,, Na,, P

Li

Sb, Siz,Srz, T1 Hg

Be, Hg

Be, Hg, IZ

2

Al, Ca,, Ca,, Cd, Co, Hg, Li, Ni, Pb, Sb, T1

Be, Hg, K, Na1

Be, Hg, 12, K

As, 11, 12, Mgz, Na,, Naz, Pt, S, Si,, Sn Ca,, Cu, P, Pb, Si,, Srz, T1 Caz, K, Se, Si, Al, P, Pb, Sb, Se, T1

all concnsb

As, Gal, II,Iz,Nal, Naz, P, Pt, S,Se, Sn

I,, p

10

Ag, SI2 Ag, Caz, Cd, Fe, Li, Mg1, Mg,, Ni, Pb, Sb, Si,, Siz, Srz, T1

Ag, Se

Ag Ag, Be,

2

Ba, Ca2, Cd, Co, Hg, K, Li, Mg,, Mgz, Mn, Ni, Pb, Sb, Si, Sr,, T1, V, Zn

Gal, Caz, Cd, CO, Hg, Mgi, Pb, Sb, Se, Sr,, Zn

Hg, Mgi, Sb, Se, Srp, V, Zn

5

Gal, I,, Iz, Nal, Naz, Snz S,Sr, Ag, Al, As, Au, Caz, Mgl, Pb, S, Si,, Sn, Sr,, T1 As, Caz, Mg,, P

all As, 11, Iz, Naz, P, S, I,, Iz, Naz, concnsb Sn Sn “The channels in the 10, 5, and 2% concentration categories are the channels in addition to the channels in the “all concns” category. * Concns concentrations. all acids

all elements except for Naz and Sn. Since the Naz channel was supposed to be only accurate a t concentrations greater than lo00 pg/mL, this behavior was expected. The Sn channel gave acceptable accuracy (510% RSD from the expected slope of 1) in the range 10 t o 250 pg/mL. A concentration of 100 pg/mL was the highest concentration in the linear range for all channels which were most sensitive for a particular element. Similarly, the linear range common to all channels, excluding I that required no background correction, was 10-100 pg/mL of the element. k Factors and Linear Ranges. Table VI contains the interchannel k factors for 2% aqua regia solutions for the indicated linear range for the appropriate affected channel

and affecting channel combinations. The elemental concentration a t which the interference was first observed is also provided. The latter did not coincide with the lower limit of the linear range if background was important. Only three interferences were linear over the 2-1000 pg/mL range (A1 on As; Mg, on Ni; V on Be) and 25 were not linear up to 1000 pg/mL of affecting element. The highest concentration a t which linearity (510% variation in k factor value) was observed for nearly all interferences was 250 pg/mL of affecting element concentration; the 100 pg/mL concentration was also suitable. Thus, k factors for each new torch could be found by simply aspirating 100 pg/mL concentrations of each affecting element and calculating k factors by dividing the

1248

ANALYTICAL CHEMISTRY, VOL. 57, NO. 7, JUNE 1985

Table IV. Numbers of Channels Not Meeting the 10% Percentage Accuracy Criteria for 10 pg/mL Status, Dilution, and Slope Behavior and LQL > 0.1 rg/mL Characteristics for Segmented Elements

HCl

aqua regia 10% 5%

parameter

all

10%

5%

2%

all

di 1ution 10 rg/mL status slope behavior LQL >0.1

11 2

13 4

25

29 13

12

2

0

27 5

0

1

2

7

1

6

8

18

10

totals rg/mL grand totals for 10 and 2%

19

26

47

60 83

hz504

HNOB 10% 5%

2%

all

2%

all

10%

2%

23

13

13

1

1

1

23 4

14 3

24 5

21

9

27 33

25 10

1

1

4

1

2

4

5

7

9

32

11

16

11

16

10

17

14

16

13

15

21

24

49

36

34 83

25

46

35

50 96

50

84

88 172

Table V. Linear Ranges, Qualitative "k"Factor Information for Spectrometer Channels and Using 2% Aqua Regia Solvent for Each Single Element

channels with k factors >0.1%" >l%*

apparent linear range, rLg/mL

slopec

std dev

RSD, %

B Ba Be Cal Ca2 Cd co

2-1000 2-250 2-250 2-1000 0.1-1000 2-1000 2-500 2-100 2-1000 2-250 2-250

1.065 0.921 0.940 1.007 1.02 1.047 1.088 1.062 1.078 0.931 0.944

0.031 0.0154 0.0304 0.022 0.048 0.037 0.025 0.080 0.096 0.017 0.017

2.90 1.68 3.23 2.20 4.71 3.51 2.28 7.52 8.92 1.78 1.81

1 11 1 1 2 11

Cr

2-100

1.067

0.035

3.26

8

cu Fe Hg I1 In K Li Mg1 Mgz Mn Mo Na1 Na2 Ni P Pb Pt S Sb Se Si, Siz Sn Sr1 Sr2 Ti

2-500 2-1000 2-1000 50-1000 1-500 10-1000 2-500 2-100 2-1000 2-250 2-1000 10-500 2-500 2-1000 2-1000 2-1000 2-1000 2-1000 2-1000 2-500 10-1000 10-1000

0.025 0.055 0.030 0.042 0.040 0.09 0.045 0.027 0.037 0.063 0.058

2.53 5.43 3.07 4.00 3.87 9.71 4.73

6 8

2.71

0.0274 0.064 0.054 0.016 0.027 0.046 0.042 0.028 0.135 0.011 0.017 0.028 0.015 0.041

3.79 6.32 5.91 6.00 3.63 6.51 5.44 1.56 2.72 4.43 4.41 2.86 1.41 1.13 8.98 2.85 1.53 4.36

7 7 10 17

2-100 2-1000 2-500

0.976 1.002 0.981 1.061 1.03 0.936 0.956 0.996 0.965 1.002 0.982 0.996 0.755 0.985 1.100 1.040 1.005 1.027 0.962 0.960 0.957 0.971 1.191 0.985 0.954 0.941

TI V

2-1000 2-500

0.977 1.019

0.026 0.033

2.69 3.20

0 12

Zn

2-250

1.040

0.017

1.61

1

channel Ag A1 As AU

10-1000

0.006

0

0

9

As

2 1

3

1 1 2 1 0

0 0

7 0 2

4 1

4 0

4 4 5 2 2

14

'0.1 % (1ng/mL affected channel/rg/mL affecting channel). 1% (10 ng/mL affected channel/pg/mL affecting channel). The equation is: observed value = slope X expected value for the specific element + constant.

pg/mL observed for the affected element by 100 since the observed concentrations for all affecting channels will agree with the specified concentration for each element a t 100

pg/mL. The manufacturer's procedure involves aspirating 1000 pg/mL standards (20). Though linearity up to 1000 pg/mL affecting element was observed for the majority of

ANALYTICAL CHEMISTRY, VOL. 57, NO. 7, JUNE 1985

interferences, more elements were still in the linear range a t 250 pg/mL than a t 1000 pg/mL for k factors greater than 0.088%. In addition, the lowest concentration at which the interference was still linear for all elements in the array was 100 pg/mL a t around a k factor of 0.1 %. I t is also possible t o utilize 250 pg/mL concentrations for most elements t o obtain acceptable accuracy in k factors. A summary of the number of k factors >0.1 and >1% is provided in Table V. Table VI1 shows how the k factors above 1% vary in Table I acids, 2% aqua regia, and 11.6% HC1/2.8% H N 0 3 and also between different torches. The solutions containing both HCl and H N 0 3 appeared to allow more reproducible results for different torches than Table I acid. The k factors for Iz and Naz channels were excluded, since their practical lower analytical limit was above 1000 pg/mL. The greatest acid effects apart from those arising from the background should be observed for the most severe interferences. The affecting element/affected channel pairings most susceptible t o the type of acid were As/Il, Co/I,, Mn/I,, Mo/Sn, and Ti/I1. The channels with the highest lower analytical limit (I,, Sn) tended to be the most variable in k factor. Table VI1 also shows that k factors must be found directly for each torch. Nevertheless, the k factors for Fe/I,, Mo/I1, Ti/Ag, V/Il, and V/A1 are relatively acid independent (i.e., they agree to within 10%). Chemical Compatibility Studies. Since four mixtures of elements were available a t 1000 pg/mL concentrations (Table IIB), to make up a equi-pg/mL 200 pg/mL concentrate and then dilute stepwise is practical. All the solvents (nitric, sulfuric, hydrochloric, aqua regia) gave rise to precipitates when all the elements were mixed together a t 200 pg/mL levels. The 10% aqua regia solution, however, had the smallest precipitate and this was largely AgC1. When silver was omitted, only a very fine precipitate appeared. Empirical experimentation showed that this fine precipitate disappeared a t an HC1/HN03 composition of 11.6% HC1/2.8% H N 0 3 (v/v). In none of the other solvents (HC1 or "OB) did the precipitates disappear on increase of acid strength. Since all the elements may be present in samples after any digestive treatment, it is clearly desirable that the final analyte be in a medium in which precipitation of inorganics does not occur and which is of high enough ionic strength to solvate high concentrations of ions. Using the 1000 Kg/mL individual standards, I, P, Na, S, Si, Sn, and T1 coprecipitated out when all the elements were mixed in 10% nitric acid a t 1pg/mL; a t 0.1 pg/mL, Cu and P b also coprecipitated. The elements which coprecipitated in 10% HC1 were Ag, As, Mo, and V. Evidently, the aqua-regia-like media allowed compromise conditions so that no chemical incompatibility occurred when concentrated equi-pg/mL mixtures were involved. No precipitates were observed for the 10 pg/mL Allied Chemical/ Jarrell Ash mixtures and for 1000 pg/mL Spex standards 1 and 2 (see Table 11) even after a year of storage a t room temperature. Precipitates were noted in the 1000 pg/mL standards 3 and 4 of Spex standard mixtures less than 1month after receipt. This again confirms that 2 and 10% H N 0 3 are not adequate acid media when the elements in these standards are to be combined a t high equi-pg/mL levels. I t was also demonstrated t h a t 2% aqua regia allowed dissolution of all elements in equi-pg/mL proportions including silver a t levels less than or equal to 10 Wg/mL, but not above. In contrast, high concentrations of single elements were generally compatible in 2% aqua regia. Though high solids nebulizers are now available, a heterogeneous medium must be perfectly mixed to obtain accurate answers. In addition, nebulizers may clog and tubing exposed to such heterogeneous media may show unpredictable memory effects. I t is also possible t h a t solid particles may not give rise to the same population of excited states as initially sol-

1247

vated chemical species. A high solids content will also increase background contributions, which change unpredictably as the degree of homogeneity of the nebulized solution varies. k Factors/Background Corrections. The channels for all the elements mixed in 2% aqua regia in 10, 1, and 0.1 pg/mL concentrations were scanned and the background correction positions then selected relative to the scans of these channels for the solvent. These spectrum shifter scans were superimposed as necessary. Backgrounds a t +15 or -15 positions (half maximum scan range) were adequate for nearly all elements in the equi-pg/mL mixtures. Solutions containing 200 pg/mL of each element made from Spex mixtures (Table IIB) were then diluted stepwise to concentrations of 100, 10, 1,and 0.1 pg/mL in 11.6% HCl/2.8% "OB. The channels for the 0.1 pg/mL concentration were scanned to assess the background corrections. Again backgrounds a t +15 or -15 positions were generally adequate. After the background corrections and k factors were entered into the data station, standardization was performed in the appropriate solvent using the nine segregated standards, and each of the above mixtures was then analyzed in the concentration mode for the four Spex standards, and for all the elements mixed together. The results for 2% aqua regia are given in Table VIII. The analytical error was always less than h 5 %, so that standard deviations are not provided. The results for the segregated elements and for the admixed elements are basically equivalent over the 0.1 to 10 pg/mL range. This implies that the correct interelemental and background corrections were utilized. In fact, the mixture gave somewhat better results than the segregated mixtures in certain cases, e.g., for Ag, Fe, K, Mg,, Mo, Pb, S, Sb, Sil, Siz, and Ti. The reverse was so for As, 11,Iz, and Nal. The dilution results are provided in Table I X for 11.6% HC1/2.8% H N 0 3 using the four Spex mixtures in Table IIB, and comparing these when all the elements except silver were mixed. The results in both segregated and combined elements are a t least as good as or better than those obtained in 2% aqua regia. Since standardization was performed with 10 pg/mL concentrations using the nine mixtures in Table IIA, this concentration should be the most accurate. The 11.6% HC1/2.8% H N 0 3 medium gives more accuracy here than 2% aqua regia. The results for 2% aqua regia are not adequate above levels of 10 pg/mL since the segregated elements do produce better results than do the combined elements. This is suggestive of insufficient solvating power when all the elements are mixed. However, nearly all the results for the combined elements for concentrations a t or below 100 pg/mL in 11.6% HC1/2.8% H N 0 3 were within 10% of the value observed for the segregated elements. The obvious exceptions again were I,, Naz, and Caz, which are not very sensitive channels. On the basis of these results, the Iz channel was replaced by an indium channel to act as an internal standard channel. Since the Si, and Siz channels were almost equivalent, the Siz channel was replaced by a boron channel. Determination Ranges. To determine the lower limits of determination as defined by percentage accuracies of 110%, the 33 element mixture in 11.6% HC1/2.8% HN03 was diluted stepwise to concentrations of 100, 10, 1,0.1,0.01, 0.005, and 0.001 pg/mL and analyzed with background corrections and k factors taken into account. The results are given in Table X and are almost equivalent to least quantifiable levels (defined as a concentration equivalent to ten times background). The detection limits (a concentration equivalent to three times background) are much lower but are analytically meaningless. The above dilution procedure allows the investigator to define the range of determina4ion quickly. Thus for 11.6% HC1/ 2.8% "OB, this extends over a t least 5 orders of magnitude for Be, Cr, Li, Sr, Ti, and V and a t the most, 3 orders of

1248

ANALYTICAL CHEMISTRY, VOL. 57, NO. 7, JUNE 1985

Table VI. Interchannel k Factors for Single Elements in 2% Aqua Regia Solvent Using No Background Corrections Over the Concentration Range 2-1000 rg/mL

affecting channel Ag A1

AS

affected channel

k factor X 10-3

std dev X 10-3

k factor Linear range, rLg/mL

a,

MI mL

AS I1 Pt Se Pb S Sb Sn P

15.2 8.41 4.52 3.55 1.48 1.46 1.38 1.38 1.31

0.56 0.16 0.54 0.061 0.091 0.26 0.093 0.15 0.035

2-1000 100-1000 50-1000 250-1000 50-1000 50-1000 50-1000 250-1000 250-1000

2 2 2 10 2 2 2 2 2

I1 Cd

23.7 19.3

4.3 1.2

100-1000 10-500

100

Si,

0.977

0.065

100-1000

100

Ba

K

3.97

0.22

50-1000

50

Be

Sn Na, Sb Ni

2.8 8.3 0.74 0.70 0.32 0.23 0.19 0.33 0.042 0.056 0.029

250-1000 500-1000

P A1 S AS Se Cr

49.6 34.3 8.31 3.66 3.48 3.45 2.80 2.36 2.15 1.55 1.49

Ca1

A1

1.03

Ca, Cd

A1

co

50-1000 250-500 500-1000

2

100-1000 250-1000 500-1000 500-1000 50-500

50 2 2 2 10

0.12

10-1000

10

0.880

0.164

10-500

10

P co

3.93 1.14

0.38 0.055

50-1000 50-1000

2 2

Pt P Ni AS S Na2 I1 Se Sn Pb Au

34.4 16.9 6.77 6.29 3.53 2.92 2.62 2.02 1.83 1.61 1.32

3.1 1.7 0.10 0.42 0.50 0.66 0.46 0.16 0.46 0.086 0.025

50-1000

2 2 2 2 2 100 2 2 2 2 2

Cr

11

Sn Naz Ag S Pt P Sb Au AS Si2 cu

Na* Zn Pt Sb Pb P

Fe

11

Pt P S Se AS Sn Ni

310

28.8 21.2 10.5

4.62 3.73 3.21 2.69 3.05 2.03 1.50 12.6 5.40 3.10 1.84 1.67 1.30 212 9.71 6.75 3.96 2.95 2.72 2.23 1.69

27 3.5 3.3 0.99 0.16 0.27 0.27 0.25 0.20 0.057 0.14 0.78 0.18 0.26 0.15 0.15 0.36 21

0.57 0.23 0.35 0.12 0.85 0.19 0.093

100-1000 50-1000

250-1000 100-1000 250-500 10-1000 250-1000 250-1000 50-1000 50-1000 50-1000 5-100 250-500 50-500 100-1000 50-1000 100-1000 5-1000 50-1000 100-1000 100-1000

Se AS A1 Pt

1.68 1.67 1.37 1.07

0.068 0.085 0.092 0.15

250-1000 250-1000 500-1000 50-1000

I1 Nap Sn Pt A1 Sil Mg, P S Na1 AS T1 Pb Se Siz Cr V

93.6 28.4 18.6 7.22 6.86 5.69 3.32 2.66 2.62 2.29 2.26 2.06 1.90 1.56 1.32 1.01

10.1 1.5 2.73 0.27 0.18 0.097 0.35 0.23 0.25 0.43 0.15 0.18 0.093 0.076 0.017 0.066 0.064

10-1000 250-500 10-250 50-1000 50-500 50-500 50-500 100-1000 50-1000 250-1000 50-1000 250-1000 50-1000 50-1000 50-500 50-1000 50-1000

T1 P Sb Zn Pb Cr Cd

53.6 10.8 4.19 3.08 1.79 1.61 0.460

1.74 0.027 0.31 0.018 0.091 0.026 0.016

50-1000 50-1000 50-1000 50-1000 100-1000 50-1000 250-1000

5 50 50 50

Pb

S AS

5.06 1.36

0.33 0.16

100-1000 500-1000

2 2

Pt

AS

60.3 4.49 3.10 1.87

0.15 0.21 0.24 0.054

50-1000 50-1000 100-1000 10-1000

2 2 2 2

Mn

Mo

10

250 2 2 250

11

10-3

affected channel

2

Au

50-1000

10-3

k factor linear range, MImL

affecting channel

5 5 250 50 2 2 5 2 2 5 2

250-1000 50-1000 250-1000 50-1000 250-1000 250-1000

100

50-1000 250-1000 250-1000 250-1000 250-1000 250-1000 100-1000 250-1000

10

5 2 2 100 2

2 50 2

10 2

50 2

k factor

X

1.17

std dev X

a,

PLPl

mL 2 2 10

10 10 50 2 2 2 2 10 2

10 100 2 50 10

50 2 2 2

Nal Na, Ni

10 10 10

P

Hg P Cd

S

11

1.85

0.40

500-1000

250

Sb

Si, Si2 Ca, Ca1

6.27 6.14 1.04 0.950

0.12 0.21 0.11 0.085

50-1000 100-1000 500-1000 500-1000

10 250 250

Hg Na1 P Sn

1.99 1.63 1.50 1.08

0.046 0.45 0.21 0.16

250-1000 500-1000 500-1000 250-1000

Hg Na1 Sn

1.94 1.58 1.46 1.05

0.065 0.42 0.22 0.17

250-1000 500-1000 500-1000 500-1000

Si, Si, Na1 Ca1 Ca,

4.27 4.27 2.50 0.967 0.938

0.14 0.15 0.015 0.015

250-1000 250-1000 250-1000 250-1000 100-1000

2 2 2 2 10

Sr1

Pt

6.10 4.60

0.74 0.078

50-250 500-1000

100

Sr2

Nap Pt Na2

5.86 3.37

1.11

0.36

50-250 250-1000

2 100

Ti

Na,

2

Se Sil

Si,

P

Sn

I1

433 131

0.14

11

7.0

50-500 50-500

2

250 2 2

2

250 2 2

2

2

10

ANALYTICAL CHEMISTRY, VOL. 57, NO. 7, JUNE 1985 ~

1249

~~

Table VI (Continued) k

affecting channel

affected channel

Hg

As

I

P

K Li

I1

Mg1

11 Sn P Ni S As Pt

Mg,

factor 10-3

0.867 595

11

Sn P Ni S Pt As

Mn

X

std dev X 10-3 0.038 13

k factor linear range. rglmL

250-1000

50

50-1000

10

2.17

0.86

250-1000

2

349 49.8 9.92 4.39 2.37 2.04 1.93

23 7.01 1.50 0.41 0.12 0.38 0.11

10-1000 10-100 50-1000 2-1000 10-1 000 100-1000 250-1000

10 2 2 2 2

329 64.9 6.75 4.44 2.26 1.33 1.29

19 4.7 0.89 0.19 0.22 0.29 0.34

10-250 2 50-1 000 250-500 2-100 10-250 250-1000 250-1000

10 2 2 2 10 2 2

73.3 8.79 7.53 4.76 4.16 2.18

4.2 0.12 0.31 0.35 0.22 0.14

500-1000 50-1000 50-500 250-1000 100-1000 250-1000

50 10 2 2

1,

Hg Ni

P S Sn

std dev X

10-3

10-3

Ag T1 Si, Si, A1 Na1 Ca, Sn co P Ca1

11.5 9.91 5.60 5.30 4.55 2.34 1.74 1.65 1.62 1.60 1.10 1.05

0.53 0.58 0.30 0.29 0.27 0.28 0.047 0.087 0.076 0.11 0.053 0.057

k factor linear range, rg/mL 50-500 10-500 50-250 50-250 50-250 10-500 50-500 50-500 50-500 250-500 10-500 250-500

Sn

All Sb Se S

458 148 27.5 10.1 8.68 3.13 2.87 1.85 1.78 1.52 1.01 0.95

22 3.06 4.9 0.28 0.36 0.076 0.074 0.21 0.050 0.029 0.090 0.071

50-1000 250-1000 10- 1000 50-1000 250-1000 2-1000 100-1000 100-1000 50-1000 100-1000 250-1000 250-1000

Na,

2.60

k

a,

wgl affecting mL channel

affected channel

S

factor

X

V

A1 As T1 Be Si,

P

50

Zn

50

0.20

10-500

Table VII. Comparison of k Factors above 1 % for Different Torches and Acidse k factor, 70

affecting element

affected channel

Table I solvent"

Table I solventb

2% aqua regia"

2% aqua regiab

HC1/HN03C 11.6%/2.8%

11.6% 12.870

A1

As

1.27

1.60*

1.50

1.52

1.35

1.15

As

1,

Cd

2.71 1.74

2.61 1.68

2.76 2.01

2.37 1.93

2.30 2.79

3.42* 3.03

Be

Sn

3.87

1.61*

4.81

4.96

3.40

2.47*

co

Pt I1 P

2.72 2.24 1.43

2.16* 1.89 1.48

3.18 1.34 1.48

3.44 2.62* 1.69

3.73 2.50 0.910

5.13* 0.14* 0.41*

Cr

Sn

5.76*

3.15

2.88

3.89

3.46

HCl/HN03d

Fe I

I1 P

21.1

22.5

21.2

23.2

20.8

64.0

56.3

59.3

59.5

63.6

43.4*

Mg

1, Sn P

18.0 5.38 0.437

36.1* 5.75 1.08*

36.1 5.75 1.08

34.9 4.98 0.992

31.1 4.53 1.07

29.8 3.88 0.40*

Mn

11

5.73

6.11

4.63

7.33*

8.30

7.45

Mo

11 Sn

9.31 2.96

9.85 2.19*

9.85 2.16

9.36 1.86

9.90 2.32

9.67 2.07

Ni

T1 P

4.96 0.994

5.99 1.83*

5.48 1.09

5.36 1.08

6.10 2.90

3.40* 0.93*

Pt

As

5.80

5.76

5.90

6.03

10.3

13.1 1.07

15.1 1.20

11.9 1.08

13.1 1.15

13.6 1.28

42.6 13.4 2.49 0.905

44.1 14.0 2.51 0.956

47.4 14.5 2.75 0.991

Ti

11

V

Ag Sn 11

A1 As OTorch 1. bTorch 2. "Torch

21.5

mL 10 2 2

2 2 10 10 50 2 2 10 2

T1

2

2

" Lowest concentration of affecting element to cause interference.

3.76

a,

@I

45.8 27.3 14.8 13.8 2.75 2.98 1.01 1.09 3. dTorch 4. eAsterisk signifies >lo% deviation from mean of values of two torches.

3.89* 13.7 1.15 26.0 15.8 2.62 0.70*

2 10 2

10 10 2 50 2 2 2 50 10 2

1250

ANALYTICAL CHEMISTRY, VOL. 57, NO. 7, JUNE 1985

Table VIII. Dilution Behavior of Segregated and Mixed Elements in 2% Aqua Regia Using the Spex Standards in Table IIB and Using Background Corrections and k Factorsc

segregatedb channel Ag A1 As AU

Ba Be Ca1 Ca2

Cd co Cr

cu

Fe Hg 1, I2 K Li Mg, Mgz

Mn Mo

Na1 Naz Ni

P Pb Pt S Sb Se Si, Si2 Sn Srl Sr, Ti T1

v

Zn

100 8.66* 10.1 10.5 -d

9.91 9.80 10.0 9.90 10.0

10.1 9.79 10.2 9.87 8.66* 9.67 10.0 9.46 10.6 9.94 10.1 9.94 9.68 13.6* 7.32* 10.0 19.0* 10.2 -

9.58 10.1 10.0 9.99 9.96 10.8 9.95 10.5 9.72 -

10.3 9.98

together 0.1”

I” O*

O*

0.936 1.23*

0.0530* 0.339*

-d

-d

0.975 0.957 0.981 0.910 0.965 0.966 0.961 0.975 0.957 0.768* bd 5.17* 0.682* 0.997 0.980 0.850* 0.959 0.874* 1.30* bd* 1.01 1.03 0.946

0.0941 0.0957 0.104 0.0547* 0.0907 0.0913 0.0921 0.0936 0.0881* 0.109 bd 6.17* bd* 0.0984 0.0994 bd* 0.0943 0.0819* 0.0621* bd* 0.104 bd* 0.0498* 0.0983 0.186* 0.0863* 0.106 0.199* 0.178*

1.01

0.970 1.13* 1.11* 1.07 0.422* 0.974 0.957 0.924

O*

0.0967 0.0259* 0.0849*

-

-

1.00

0.0904 0.104

0.995

10” 9.66 9.64 11.7* 10.1 9.47 9.34 9.70 9.57 9.70 9.68 9.51 9.67 9.49 8.16* 24.7* 1706* 9.01 9.82 9.52 9.66 9.53 9.60 12.7* 35.2* 9.70 19.1* 9.85 10.6 9.71 10.4

9.55 10.0 10.0 17.8* 9.44 9.95 9.57 10.1 9.93 9.73

10

0.826* 1.01 1.34* 1.06 1.01 1.01 1.06 1.01 1.02 1.01 1.01 1.03 1.02 0.883*

1.58* 180*

0.816* 1.08 1.03 0.945 1.01

0.988 1.60* 1.06 1.08 1.91* 1.02

1.30* 1.04 1.10 1.08 1.03 1.04 0.859* 1.03 1.03 1.02 0.876* 1.05 1.03

0.1“

bd* 0.0542* 0.242* 0.0891* 0.0929 0.0933 0.104 0.0714* 0.0924 0.0900 0.0905 0.0982 0.0961 0.116* bd* 15* bd* 0.0993 0.0982 bd* 0.0943 0.0957 0.353* 0.650* 0.929 0.238* 0.919 0.0965 0.0938 0.106 0.149* 0.0951 0.0911 0.0970 0.0952 0.114* 0.0960 0.175* 0.0991 0.0994

Expected concentration in bg/mL based on known dilutions; standardization performance with solutions of Table IIA with standard 1 being 2% aqua regia. bAs in Table IIB but in 2% aqua regia. “Asterisk indicates 210% error from expected value. dNot analyzed. magnitude for I and Sn. All the results for the other elements in this particular array lie between these two extremes. NBS Water Standard. The analyses results for NBS water obtained by using the developed techniques are given in Table XI. Only the values for Na, Ni, and P b were significantly different from the target NBS values. Since Si values were also evident, the high sodium level may have been caused by contact with glass before the NBS water was placed in its original plastic container. Some of the more imprecise values, e.g., As and Co are because of the proximity of the lowest analytical limit. The analytical limits for Ag, Hg, and Se were above the concentrations present in the standard.

DISCUSSION The work concerning the range of linearity in the present study validated the manufacturer’s recommended method (20) for standardization (Table V). Most elements, however, were not acceptably linear from low concentrations to 1000 pg/mL. These results agreed partially with those of Maessen and Balke (5),who found nonlinearity at low concentrations. The present results a t low concentrations were found to be due to background effects and not to intrinsic nonlinearity. The lower limit of determination here was defined as the lowest concentration of an element that still agreed within 10% of the true value, the 10% figure being selected since it represented

the reasonable day-to-day variation of line intensities, and because the linear range was selected to tolerate percentage accuracies and relative standard deviations of less than 10%. I t was recommended that the k factor be calculated from k factor = concn in affected channel (pg/mL) concn of affecting element (250 or 100 pg/mL) The 100 and 250 yg/mL concentrations were selected from linearity considerations relative to standardization of the affecting element and to the interference that the affecting element caused in the affected channel. The k factors for new torches were easily found by nebulizing 250 or 100 pg/mL solutions since these concentrations allow k factors >0.088% to be detected, and interchannel k factors are equivalent to interelemental k factors. An alternative method is to set a constant Cu/Mn intensity ratio (21) to achieve constant interelemental corrections, but this proved more tedious than just aspirating 250 or 100 pg/mL standards a t the optimized conditions for each torch. In addition, full advantage of “sensitive” torches was possible, and in any case, “less sensitive” torches had to be subjected to the other procedure. Another alternative technique is selective spectral line modulation (ZZ), which, however, is limited only to channels monitoring emission from excited atoms, and so is not ap-

ANALYTICAL CHEMISTRY, VOL. 57, NO. 7, JUNE 1985

1251

Table IX. Dilution Behavior of Segregated and Mixed Elements in 11.6% HC1/2.8% HNO, Using the Spex Standards in Table IIB and Using Background Corrections and k Factorsc

together

segregatedb channel Ag

A1 As Au Ba Be Ca1 Ca2 Cd co

Cr

cu

Fe Hg I1 12 K Li Mg1 Mgz Mn Mo Nal Na2

Ni P Pb Pt S

Sb Se Si, Si, Sn Sr1 Sr2 Ti T1 V Zn

100" -d

A

100"

-d

-d

-

-

1.00 1.00

0.0954 0.0745*

94.1 99.6 90.0 92.2 103 80.0* 77.2 85.8* 79.8* 77.9* 90.4 76.2* 78.0* 94.6 70584* 86.3* 101 77.5* 86.9* 77.5* 79.3* 105 18.1* 81.9* 79.0* 81.1* 84.0* 87.31 89.1* 94.3 86.6* 86.6* 102 83.7* 99.2 85.9* 88.6* 85.8* 89.2*

10.3 10.6 10.6 9.59 10.2 9.64 9.62 9.38 9.55 9.82 9.52 9.92 8.51* 10.6 2634* 9.26 10.1 9.59 9.47 9.55 9.86 13.4* 2.15* 10.1 7.92* 9.33 10.1 9.82 10.5 10.0 10.1 10.0 10.2 9.68 9.79 10.0 10.2 9.94 9.53

1"

90.2 87.9*

10.5 10.1

89.8* 96.7 81.2* 93.4 92.3 90.0 87.3* 86.7* 87.2* 84.0* 104 32.0* 80.7* 84.0* 88.3* 93.6 90.0 94.5 116* 73.7* 94.5 96.1 90.7

10.0 10.3 9.55 9.62 9.80 9.98 10.2 9.87 10.4 8.52* 8.91* -5.00* 9.58 10.3 9.65 9.51 9.61 100 13.5* 8.83* 10.2 8.29* 9.80

1.02 1.04 1.01 0.989 0.982 0.996 1.00 0.986 1.01 0.869* 1.45* -7.29* 0.860* 1.02 1.03 0.985 0.998 0.947 1.65* -0.150* 1.07 1.05 0.988

0.101 0.100 0.0993 0.0894* 0.0983 0.102 0.102 0.0962 0.0990 0.0850* 0.400* 2.00* 0.0608* 0.101 0.102 0.0905 0.0972 0.0867* 0.255* -1.13* 0.103 0.0589* 0.0968

93.0 104 102 115* 115* 134* 85.2* 95.0 94.4

9.83 10.8 9.77 13.2* 13.2* 10.7 9.66 9.68 10.0

0.978 1.01 0.980 1.06 1.05 0.688* 1.01 1.01 0.990

0.108 0.105 0.0700* 0.107 0.106 0.106 0.0975 0.109 0.0990

91.3 97.3

10.3 9.52

-

-

-

-

-

-

10"

0.1"

10'

-

-

-

-

-

-

1.03 1.00

0.0987 0.100

0.1"

1"

-

1.03 1.07 1.06 1.03 1.05 1.03 1.01 0.989 1.04 1.02 0.993 1.01 0.885* 1.59* 270* 0.918 1.03 1.04 0.996 1.01 0.965 1.66* 0.319* 1.07 1.00 1.01 1.02 1.11

1.07 0.994 1.40 1.40. 1.00

1.02 1.03 1.04 1.09 1.04 1.02

-

0.104 0.099 0.0993 0.0990 0.0988 0.101 0.0993 0.0950 0.0931 0.103 0.0944 0.0968 0.0888* 0.0871* 32.6* 0.0136* 0.0976 0.0996 0.0868* 0.0966 0.0810* 0.410* 0.818* 0.0935 0.102 0.0893* 0.0927 0.0726* 0.109 0.112* 0.263* 0.267* 0.0955 0.0958 0.101 0.0979 0.0973 0.0970 0.0926

Expected concentrations in wg/mL based on known dilutions; standardization performed with solutions of Table IIA except for standard 1 being 11.6% HC1/2.8% HN03. bAs in Table IIB but in 11.6% HC1/2.8% HN03. "Asterisk indicates 210% error from expected value. Not analyzed. Table X. Lower Determination Limit Ranges (110% Percentage Accuracy) for 33 Elements" Mixed Together in equi-wg/mL Quantities in 11.6% HC1/2.8% HNOs

lower determination limit range, rg/mL 0.1-1.0 0.01-0.1 0.005-0.010 0.001-0.005