2888 Anal. Chem. 1986, 58, 2886-2888

Comparison of Jarrell-Ash, Perkin-Elmer, and Modified Perkin-Elmer Nebulizers for Inductively Coupled Plasma Analysis

David E. Nixon*

Department of Laboratory Medicine, Mayo Clinic, Rochester, Minnesota 55905

Gordon A. Smith

Department of Engineering, Mayo Clinic, Rochester, Minnesota 55905

Performance characteristics of pneumatic nebulizers used with inductively coupled plasmas (ICP) have been critically reviewed in the recent past (1-9). Although the pneumatic nebulizer remains the most popular form of sample intro- duction because of its simplicity and expense, it also remains as the single most limiting factor in the transition of ICP from a laboratory curiosity to a routine tool (2-4,9). Although the original crossflow nebulizer designed by Kniseley et al. (11) gave adequate detection limit performance, sensitivity to solution composition or particulate matter and difficult routine maintenance have plagued this design (2,3,5). The nebulizer, whether concentric or crossflow design, must have a gas orifice small enough to produce a high velocity jet stream past the sample inlet (3 ,5 , 7, 9) and a sample orifice large enough to allow the free aspiration of material such as urine or diluted serum without clogging (3,5,9). In the case of the crossflow design, the proximity of the gas orifice behind the sample orifice is also critical (9).

Second generation fixed crossflow nebulizers with large sample orifices and smaller gas inlets have now appeared (3, 5) . One model, available from Jarrell-Ash Division of Allied Analytical, produces excellent detection limits and is capable of aspirating most urine or serum samples without clogging. When a blockage does occur, however, the plasma must be stopped and the nebulizer unclogged or a replacement in- stalled. While this device represents an improvement over previous designs, it precludes easy routine repair and a number of expensive replacements must be kept on hand.

An alternate approach has been taken by Perkin-Elmer. Their crossflow nebulizer is completely demountable, with inexpensive gas and sample needles of molded plastic easily replaced. New needles simply fit prebored holes in the neb- ulizer body. Sample needle penetration and, therefore, alignment are predetermined by a stop ring molded on the needle. The proximity of the gas needle to the rear of the sample orifice is also preset. The disadvantage of this system is that when samples are pumped into the uptake orifice, signal to background ( S I B ) ratios and detection limits are signifi- cantly poorer than those attained with the Jarrell-Ash design.

In order to retain the performance of the Jarrell-Ash design and the serviceability of the Perkin-Elmer unit, the original gas orifice was replaced by a precision glass capillary with a significantly smaller diameter orifice. In this paper, we de- scribe this modification in detail. Detection limit and signal to background ratio data for this device are compared with experimentally derived values for the stock Perkin-Elmer nebulizer and the Jarrell-Ash. Results for the analysis of urine and serum control specimens are compared to the certificate values.

EXPERIMENTAL SECTION Multielement ICP System. The components that com-

prise our six-element direct reading system are outlined in Table I. A complete description of this scanning/direct reading plasma facility can be found elsewhere (12). Operating parameters used in this study are listed in Table 11. Elements and their wavelengths fixed on the exit slitframe of the direct reader are listed in Table 111. All of the parameters listed

Table I. Inductively Coupled Plasma Facilities

plasma Plasma-Therm, Inc., Kresson, NJ HF generator Model HFS-2500E torch one piece quartz construction

18 mm i.d. plasma tube 14 mm i.d. auxiliary tube 1.2 mm i.d. aerosol tip

flow meters Linde 201-4335; 201-4334 monochromater McPherson Model 216 (1 m focal length)

grating 1200 grooves/mm; 300 nm blaze slits entrance, 25 km X 4 mm

exit, ARL 6910-2 50 pm X 4 mm optics 25 mm X 200 mm focal length planoconvex filter Acton Research Model 395/BR photomultipliers Hamamatsu R760 13 m diameter

electronics computer interface, Mayo design multichannel readout, Mayo design HV and signal boards, Mayo design

Perkin-Elmer Model N058-0368 Gilson Minipuls 2; 10 roller

nebulizers Jarrell-Ash Model 90-790

Pump

Table 11. Plasma System Operating Conditions

forward power reflected power gas flow rates

plasma auxiliary aerosol back pressure

sample pump rate

observation height sienal integration

1300 W <5 w 15 L/min normally off varies with nebulizer 124 kPa (Perkin-Elmer) 207 kPa (Jarrell-Ash) 345 kPa (modified Perkin-Elmer) 4.05 mL/min (Perkin-Elmer) 2.92 mL/min (Jarrell-Ash) 2.92 mL/min (modified Perkin-Elmer) 11 mm 10 s

Table 111. Elements and Wavelengths on the Exit Slit Frame

element wavelength, element wavelength, and state nm and state nm

Zn I 213.856 Ca I1 315.887 Fe I1 259.940 c u I 324.754 Mg I1 293.654 Y I1 360.073

in these three tables remained virtually constant for all ex- periments with the exception of aerosol flow rate (pressure on each nebulizer) and aspiration rate when the nebulizers were pumped.

Reagents. The aqueous acidic calibrating standard was prepared from serial dilutions of Baker Instra-Analyzed At- omic Spectral Standards obtained from J. T. Baker Chemical Co., Phillipsburg, NJ. Concentrated nitric acid was distilled in glass, then redistilled in a Teflon, subboiling still and stored in Teflon bottles. Water was triple distilled and delivered to the laboratory in carboys dedicated to this use.

Reagents and their composition are listed in Table IV. The routine analysis of sera requires a diluent containing only triple

0003-2700/86/0358-2886$01.50/0 0 1986 American Chemical Society

ANALYTICAL CHEMISTRY, VOL. 58, NO. 13, NOVEMBER 1986 2887

Table IV. Composition of Calibrating Standards and Blanks

reagent component and composition

blank 1% aqueous HNO, urine diluent 2% aqueous HNO,

urine standard 1% aqueous HNO, 100 pg of Y/L

50 mg of Ca/L 50 mg of Mg/L 500 pug of Zn/L 500 pg of Fe/L 50 pg of Cu/L 50 pg of Y/L

12.22 mg of Ca/dL 7.17 mg of Mg/dL 2.85 mg of Zn/L 2.18 mg of Fe/L 1.89 mz of Cu/L

serum diluent triple distilled water serum standard

distilled water to avoid protein precipitation. Plasma protein calibrating standard and control specimens were obtained from American Dade, Division of American Hospital Supply Corp., Miami, FL. Elemental concentrations in both plasma ma- terials were confirmed in our laboratory by atomic absorption spectrophotometry. A urine control specimen, Urichem Level 11, Lot No. 523-022, was obtained from Fisher Diagnostics, Division of Fisher Scientific Co., Orangeburg, NY.

Nebulizers. All three nebulizers evaluated here are of the crossflow design. The Jarrell-Ash fixed crossflow nebulizer (Catalogue No. 90-790) is the standard nebulizer heretofore used for routine urine and serum analyses on our ICP system. The Perkin-Elmer (N058-0368) with a double pass Scott type spray chamber (N058-0360), evaluated in this study, was provided by Perkin-Elmer. This same Perkin-Elmer nebulizer, modified by replacing the stock gas orifice with an orifice of precision glass capillary, was also evaluated with the two types of spray chambers on hand (Jarrell-Ash barrel type and Perkin-Elmer double pass Scott type). Since the stock Per- kin-Elmer nebulizer was designed to have samples pumped to the sample orifice, we retained this feature.

The original molded plastic gas orifice in the Perkin-Elmer nebulizer was replaced with a 2.54 0.d. mm X 13.5 mm length of precision bore (0.1778 mm) tubing from Wilmad Glass Co., Inc., Buena, NJ. A stainless steel stop ring 2.21 mm X 5.23 mm in diameter was attached with epoxy glue to the end of the capillary. The replacement capillary had essentially the same outside dimensions as the original Perkin-Elmer molded capillary with the exception of overall length. The length of penetration into the nebulizer body was increased to bring the gas orifice closer to the backside of the sample needle.

Test Design. The objective of this investigation was to produce a nebulizer with performance characteristics equiv- alent to the Jarrell-Ash while retaining the ease of service of the Perkin-Elmer unit. Although nebulizer efficiency has been cited as an evaluation criteria ( I , 5 , 7), the most widely ac- cepted criteria appears to be detection limits based on three standard deviations of the blank readings (13). Two additional criteria were applied here: (1) the nebulizer must be capable of routine daily aspiration of urines or diluted sera, and (2) the nebulizer-plasma system must produce detection limits low enough to perform the intended analyses. To obtain optimized detection limits and signal to background ratios for these three nebulizers, intensity measurements of blanks and standards or samples were recorded after nebulizer pressure, observation height, and solution pump rate were changed.

The test design was further restricted by our choice of copper as the element to be optimized. Normal urinary copper excretion ranges from 10 to 60 fig/24 h void or approximately

Table V. Critical Parameters for Several Crossflow Nebulizers

spacing, gas orifice, sample orifice, manufacturer mm mm mm

Kniseley (11) <<0.5 0.23 0.30 Perkin-Elmer 1 0.24 0.24 Jarrell- Ash <<0.5 0.18 0.42 Novak, Jr. (5) a 0.13 0.30 modified P-E <<0.5 0.18 0.24

(1 Not stated in text. Appears <<0.5 mm in figure.

~~

Table VI. Comparison of Detection Limits

detection limits" nebulizer Ca Mg Fe Zn Cu

Jarrell-Ash (not pumped) 24 90 4 3 6 Jarrell-Ash (pumped) 60 120 8 3 6 Perkin-Elmer (Scott) 114 172 10 6 2 modified Perkin-Elmer 50 147 6 2 2 modified Perkin-Elmer (Scott) 33 80 4 2 2 Winge (9) b 6 0 6 2 5

"All values in uelL. *Value for this calcium line not listed.

Table VII. Comparison of Signal to Background Ratios

S,,R" Ca Mg Fe Zn Cu

Jarrell-Ash (not pumped) 40.1 12.6 4.2 8.6 1.3 Jarrell-Ash (pumped) 39.8 10.6 4.0 8.2 1.2 Perkin-Elmer (Scott) 8.1 7.4 1.8 2.8 1.1 modified Perkin-Elmer 51.1 12.4 4.7 10.1 1.3 modified Perkin-Elmer (Scott) 39.3 10.9 4.1 8.5 1.2

Aqueous acidic urine standard (Table IV) used for data acqui- sition.

~ ~

8-50 fig/L. Normal levels of Ca, Mg, Zn, and Fe in urine are significantly higher; thus detection limit and sensitivity for copper are more critical. With the optimization of copper complete, data were accumulated for the other four elements.

RESULTS AND DISCUSSION Nebulizer Dimensions. I t is useful to examine some

critical dimensions of several second generation crossflow nebulizers and compare them to the original crossflow de- signed by Kniseley et al. (11). These dimensions together with those for the stock and modified Perkin-Elmer are shown in Table V. From this table, it is clear that the Jarrell-Ash and Novak (5) nebulizers utilize significantly smaller gas orifices than the stock Perkin-Elmer or the Kniseley designs. The smaller gas orifices assure a high velocity stream (>500 m/s) of argon past the sample orifice (5) and thus a large pressure drop across the sample orifice (7,9). It has been shown that the particle size distribution of the aerosol shifts to smaller droplets as the pressure drop increases (9) and thus more efficient sample transfer to the plasma (5) .

The pressure drop or velocity of gas streaming past the sample orifice is also influenced by the proximity of the gas orifice to the sample orifice in two dimensions (5,9). Not only should the sample orifice be aligned with the diameter of the gas orifice but the gas orifice should be very close to the rear of the sample orifice to maintain high gas velocity a t a par- ticular flow rate. As the gas orifice is moved toward the sample orifice, the pressure drop increases and the particle size dis- tribution shifts to smaller droplet diameters (9). As a result of the above measurements and the foregoing observations, the Perkin-Elmer nebulizer was modified by not only reducing the gas orifice inside diameter by 75 % , but also moving this

2888 ANALYTICAL CHEMISTRY, VOL. 58, NO. 13, NOVEMBER 1986

Table VIII. Results for the Analysis of Serum and Urine Controls with the Modified Perkin-Elmer Nebulizer

Ca Mg Fe Zn c u

Fisher Urichem Urine Control

ICP 94 f 4 159 f 10 0.08 f 0.02 0.42 f 0.04 0.20 f 0.02 certificate 104 f 11 163 f 10 0.09 f 0.02" 0.48 f 0.08 0.19 f 0.05

Dade Lab-trol Plasma Protein Control

ICP 9.84 f 0.21' 2.52 f 0.22b 0.64 f 0.09 0.35 f 0.05 0.92 f 0.08 certificate 9.90 f 0.30b 2.50 f 0.16b 0.63 f 0.11 0.31 f 0.04" 0.91 f 0.08"

a Values not certified. Determined by flame atomic absorption. Values in mg/dL. All other values are mg/L.

orifice very close to the sample uptake orifice. Detection Limits and SIB Ratios. Detection limits were

tabulated for each nebulizer-spray chamber combination by following the test design outlined above. With the plasma system optimized for copper, multiple signal integrations were recorded on at least two separate operating sessions for both aqueous acidic standard and blank solutions (Table IV). From this body of data, average detection limits and SIB ratios were calculated and are presented in Tables VI and VII.

The following observations and conclusions were drawn from the tabular data:

(1) The stock Perkin-Elmer nebulizer produced the poorest detection limits and S I B ratios of the ones tested. This relatively poor performance in our system is a direct result of the large gas orifice and its extreme setback from the sample orifice. When the Scott chamber was replaced with a barrel type chamber, the nebulization of diluted serum produced a visible stream of iridescent particles traversing the aerosol channel. With the Scott chamber in place, these unatomized particles disappeared. We hypothesize that the Scott chamber effectively eliminates the large particles and with them a large portion of the analyte aerosol. The modified Perkin-Elmer and Jarrell-Ash nebulizers did not produce this stream of unatomized particles.

(2) While detection limits for the modified Perkin-Elmer nebulizer (with a Scott chamber) are only half those obtained with the stock Perkin-Elmer, the overall average S I B ratio is 3 times that produced with the stock unit. The change in S I B ratio is not surprising because it has been shown that a decrease in spacing between the two orifices creates a larger pressure drop across the sample orifice and therefore a higher volume percent of small particles (9). I t has also been shown that a direct relationship exists between the signal magnitude and the total number and volume of small (<8 pm) particles (8). Detection limits are not 3 times better because of stability (or noise) problems produced when the sample orifice is sig- nificantly larger than the gas orifice (9).

(3) When compared with the barrel type spray chamber, the Scott chamber, in use with the modified Perkin-Elmer nebulizer, produced improved detection limits for three of the five elements tested. Average SIB ratio was reduced by about 14%. The small reduction in S I B ratio with the addition of the Scott chamber suggests that the modified Perkin-Elmer

nebulizer produces only a small fraction of large aerosol particles not capable of traversing the double pass chamber. In addition, the signal to noise ( S I N ) ratio is reduced somewhat by culling the large particles. (4) S I B ratios and detection limits produced by the mod-

ified Perkin-Elmer nebulizer are as good as or better than those produced by the Jarrell-Ash fixed crossflow.

(5) Pumping sample into the Jarrell-Ash nebulizer had little effect on its performance.

Control Specimen Analysis. Average results for the analysis of serum and urine controls are presented in Table VIII. Reconstituted urine control was analyzed after a 2-fold dilution with 2% nitric acid containing yttrium internal standard. Serum control was diluted 10-fold with triple distilled water prior to analysis. Fresh dilutions of both serum and urine were analyzed on at least five separate days; therefore, the two standard deviation values listed represent the total analytical variance. Excellent agreement with the certificate values or atomic absorption results was attained by using the modified Perkin-Elmer nebulizer and the Scott type spray chamber.

Registry No. Ca, 7440-70-2; Mg, 7439-95-4; Fe, 7439-89-6; Zn, 7440-66-6; CU, 7440-50-8.

LITERATURE CITED Maessen, F. J. M. J.; Covert, P.; Balke, J. Anal. Chem. 1984, 56,

Barnes, R. M. ICP In f . Newsl. 1981, 6 , 459. Wohlers, C. C.; Hoffman, C. J. ICf In f . Newsl. 1981, 6 , 500. Ebdon, L.; Cave, M. R. Analyst (London) 1982, 107, 172. Novak, J. W., Jr.; Lillie, D. E.; Boorn. A. W.; Browner, R. F. Anal. Chem. 1980, 52, 576. Dobb, D. E.; Jenke, D. R. Appl. Spectrosc. 1983. 37, 379. Gustavsson, A. Spectrochim. Acta, Part B 1984, 398, 743. Olsen, S. D.; Strasheim, A. Spectrochim. Acta, Part B 1983, 388, 973. Fujishiro. Y.; Kubota, M.; Ishida, R. Spectrochim. Acta, Part B 1984. 398, 617. Winge, R. K.; Peterson, V. J.; Fassel, V. A. Report No. EPA-600/4-79- 017, 1979; Ames Laboratory, USDOE: Ames, 1A. Kniseley, R. N.; Amenson, H.; Butler, C. C.; Fassel, V. A. Appl. Spec- trosc. 1974, 28, 285. Nixon, D. E.; Moyer, T. P.; McCall, J. T.; Ness, A. 6.; Fjerstad, W. H.; Johnson, P.; Wehde, M. 8. Clin. Chem. (Winston-Salem, N . C . ) 1988, 32, 1660. Kaiser, H. Anal. Chem. 1970, 42, 26A. Morrison, G. H. CRC Crit. Rev. Anal. Chem. 1979, 8 , 287.

898.

RECEIVED for review March 31, 1986. Accepted July 7,1986.