Analyst, April 1996, Vol. I21 (483488) 483

Slurry Preparation by High-pressure Homogenization for Cadmium, Copper and Lead Determination in Cervine Liver and Kidney by Electrothermal Atomic Absorption Spectrometry

Yanxi Tan=, William D. Marshall",* and Jean-Simon Blaisb a Department of Food Science and Agricultural Chemistry, Macdonald Campus of McGill, 21 11 1 Lakeshore Road, Ste.-Anne-de-Bellevue, Que'bec, Canada, H9X 3V9 b Agriculture Canada, Food Research and Development Center, 3600 Casavant Blvd., St.-Hyancithe, Qutbec, Canada, J26 8E3

Homogenization with a flat valve homogenizer in combination with high-speed blending was evaluated for the preparation of slurries suitable for the ETAAS determination of cadmium, copper and lead concentrations in six SRMs and in frozen cervine liver and kidney. Fresh tissue (approximately 2 g) or powdered SRM (approximately 0.1 g) was dispersed, at high speed, in 20 ml of ethanol-water (1 + 9 v/v) containing 0.25% m/m tetramethylammonium hydroxide. The resulting suspension was passed through a high-pressure flat valve homogenizer. Determinations performed on the resulting homogenate, provided estimates for Cd, Pb and Cu concentrations that were within 27,23 and 18% of the certified values, respectively, for the six SRMs. In all instances, the experimental results did not differ significantly from the certified values. For frozen tissues there was good agreement between the concentrations as determined by slurry homogenization-ETAAS and conventional digestion-ICP-MS. In addition, no significant differences were detected between the slopes of the calibration curves for external standards and standard additions to homogenized sample (SRMs or fresh tissue). Moreover, replicate determinations of analyte concentrations in slurries at various times post-preparation did not detect any segregation of the homogenates during 6 d. For these matrices at least, short-term sample storage had no discernible effect on the analyte apparent concentrations. The applicability of the process was limited only by the levels of contaminating Pb and Cu introduced into the sample by the homogenizer. Keywords: Homogenization; slurry introduction-electrothermal atomic absorption spectrometry; liver; kidney; cadmium, lead and copper determination

Introduction There is continued interest in sample preparation and introduc- tion procedures that avoid classical digestion methods prior to atomic spectroscopy. Frequently, the time required for sample preparation can exceed the actual instrument time by two or more orders of magnitude. Moreover, the more labour intensive the sample pre-treatment, the more prone to errors the analysis becomes. Within the general field of solid sampling analysis,l" the introduction of the sample as a slurry into the atomic

* To whom correspondence should be addressed.

spectrometer continues to be actively studied because of the relative ease with which slurries can be prepared from a variety of sample types, including foods and biological materials. Many foodstuffs can be freeze-dried and ground to a fine powder with relative ease using pertrifluoroethylene-coated beads.6 Direct atomization from the solid state offers the possibility of excellent sensitivity but can suffer from molecular absorption and/or scattering from the matrix, which can produce suffi- ciently large background signals to overwhelm the compensa- tion capabilities of common deuterium background correction systems. Additional difficulties can include sample inhomo- geneities, the requirement for repeated microweighings and the lack of suitable solid calibration standards and techniques. By contrast, a slurry prepared from a solid sample can often be manipulated with conventional liquid sampling devices and techniques.

A variety of sample pre-treatment procedures have been described and evaluated for the production of quasi-stable suspensions of samples prior to analysis by atomic spec- trometry. Included are the addition of a thixotropic thickening agent (Viscalex)7 or a non-ionic surfactant (Triton X- 100)8-10 to a water carrier, or the use of a glycerine-water or glycerine- methanol11 or propan-2-0112 dispersing medium. Wetting and antifoaming agents have been demonstrated to improve the dispersion of certain slurries.13 Alternatively, suspensions with a tendency to segment rapidly have been reproducibly sampled by using ultrasonic agitation,14 air or argon15 bubbling, vortex mixing16 or magnetic stirring.17 Partial digestion procedures to produce carbonaceous slurries have also been successfully applied to the analysis, by ICP-AES, of a series of standard reference materials of biological origin. 18 Various alkylammon- ium hydroxide formulations have been used extensively to solubilize tissues,l9-22 particularly those of zoological origin.

The generic term 'homogenizer' has been applied to any piece of equipment that disperses and/or emulsifies (including a turbine blade mixer, an ultrasonic proble, a high-shear mixer, a colloid mill, a blender or even a mortar and pestle). A more precise definition of a homogenizer is a device consisting of a positive displacement pump and a homogenizing valve that forms a restricted orifice through which product flows. The successful use of a homogenizer to prepare emulsions and/or dispersions of meats suitable for transmission FTIR spec- trometry has been reported23 recently. The prerequisites for successful FTIR analysis of emulsions are that the globule size distribution must be narrow, with a mean diameter less than the spectroscopic wavelength (2-8 pm) to minimize incident radiation scattering. Moreover, large globules are known to

Publ

ishe

d on

01

Janu

ary

1996

. Dow

nloa

ded

by U

nive

rsity

of

Wes

tern

Ont

ario

on

26/1

0/20

14 0

0:49

:17.

View Article Online / Journal Homepage / Table of Contents for this issue

484 Analyst, April 1996, Vol. 121

accelerate coalescence. To our knowledge, high-pressure homogenization has not been studied as a rapid preparatory technique for sample preparation prior to slurry introduction atomic spectrometry.

The objective of this study was to evaluate high-pressure homogenization as a rapid sample treatment procedure to provide a quasi-stable emulsion/dispersion of soft biological tissues or powdered SRMs which could be analysed for cadmium, copper or lead by ETAAS.

Experimental Sample Preparation Frozen (-20 "C) cervine (deer or moose) liver or kidney tissue was thawed and a 1 cm wide transverse section of the organ was excised with a stainless-steel scalpel, accurately weighed (approximately 2 g) and transferred into a 50 ml beaker together with 20 ml of ethanol-water (1 + 9 v/v) containing 0.25% m/m tetramethylammonium hydroxide (TMAH). For SRMs, accu- rately weighed sample (approximately 0.1 g) was added directly to 20 ml of the same solvent combination [20 ml of ethanol- water (1 + 9 v/v) containing 0.25% m/m TMAH] in the beaker. The resulting mixture was maceratedblended (20 000 rpm, 60 s; SDT Tissuemizer, Tekmar, Cincinnati, OH, USA) and the resulting suspension was then processed through the 20 ml capacity flat valve homogenizer (EmulsiFlex Model EF-B3, Avestin, Ottawa, ON, Canada) capable of developing 34.58 MPa. Each slurried sample was re-processed through the homogenizer three more times. Dilutions, when required, were performed with solvent mixture which had not been hom- ogenized.

Prior to ICP-MS determinations of Cd, Pb and Cu, frozen tissues were thawed, cut into thin strips and ground to a paste in a meat grinder. Accurately weighed aliquots of paste (6 g of liver or 3 g of kidney) were digested at room temperature with 70% HN03 (25 ml) until gas evolution had ceased, then heated at 80°C until a clear yellow solution was obtained. The resulting digests were diluted prior to analysis.

ETAAS Analyses for cadmium, copper or lead were performed using a hot injection technique on a Varian Model 300 GF-AAS system equipped with an autosampler, conventional hollow-cathode lamps and Zeeman-effect background correction. Analytical operating parameters for each analyte element are presented in Table 1.

Calibration Quantification was performed by both the method of external standards (ES) and by standard additions (SA), ES consisting of appropriately diluted processed reagent blank, with up to four levels of standard being prepared automatically by the sample introduction device. The background-corrected peak-area re- sponse, resulting from three replicate injections of each diluted standard, was used to define the best-fit regression equation. For SA calibrations, 10 pl aliquots of each emulsion/dispersion were amended with 2,5,10 or 20 1.11 of aqueous standard, chosen to result in a range of peak areas including signals that were at least twice the signal for the unamended sample. The data were modelled by least-squares linear regression. Quantification was performed by dividing the y-intercept of the regression equation by the slope of that equation and the over-all standard error of the estimate (SE,,,) was calculated from

Student's t-test was used to identify significant differences between the slopes or between the y-intercepts of regression for different sample matrices. The F-test was used to detect significant differences between regression models.

Results and Discussion Conventional approaches to protein solubilization have often relied on a combination of high ionic strength and a strong base (alcoholic KOH or aqueous NaOH). For certain volatile heavy metals (As, Cd, Cu, Hg, Mn, Pb, Se and Zn) it was envisaged that the slurry preparation technique would be useful for both flow injection-quartz T-tube atomization24725 and for ETAAS. The appreciably lower limits of detection for these analytes (relative to conventional flame AAS and plasma optical emission spectrometry) result in less severe constraints in the maximum dilution which can be incurred during sample preparation and still provide reliable analytical data. However, a limitation of the quartz interface devices is that the presence of alkali and alkaline earth metal cations in the mobile phase must be minimized or avoided to prolong their useful lifetime. The success of alkylammonium hydroxides as tissue solubilizers and the miscibility of these compounds with organic media suggested that these bases might be substituted for alkali metal hydroxides as effective solubilizers for a variety of biological matrices.

Conventional automated homogenizers and microfluidizers typically employ a pressure-driven inlet and outlet ball check valve and seat design to pump the liquid feed stock through the homogenizing valve, which makes them incompatible with

Table 1 Furnace operating parameters for the determination of Cd, Pb or Cu in test slurries

Analyte Modifier Furnace programme Cd* 5 pl of 0.1 mol 1-1

N&H2P04 Ar purge gas (3 1 min-1) on, 70 "C, ramp to 120 "C

over 50 s, to 300°C over 25 s, isothermal at 600 "C for 25 s, Ar off, 1800 OC, read for 2 s, Ar on, 2600 "C for 2 s, 60 "C for 13 s

Ar purge gas (3 1 min-1) on, 70 "C, ramp to 120 "C over 50 s, to 300°C over 75 s, isothermal at 950 "C for 30 s, Ar off, 2400 "C, read for 2 s, Ar on, 60°C for 13 s

Ar purge gas (3 1 min-1) on, 70 "C, ramp to 120 "C over 50 s, to 300°C over 25 s, isothermal at 900 "C for 25 s, Ar off, 2300 "C, read for 2 s, Ar on, 2600 "C for 2 s, 60 "C for 13 s

Pb+ 5 p1 of 3.0 moll-1 Pd(N03)Z

c u t 5 pl of 1% m/m N&N03

* Analyses for Cd were performed at 228.8 nm (0.5 nm slit width) with a partition tube. + Analyses for Pb were performed at 283.3 nm (0.5 nm slit width) with platform. * Analyses for Cu were performed at 324.8 nm (0.1 nm slit width) with a partition tube.

Publ

ishe

d on

01

Janu

ary

1996

. Dow

nloa

ded

by U

nive

rsity

of

Wes

tern

Ont

ario

on

26/1

0/20

14 0

0:49

:17.

View Article Online

Analyst, April 1996, Vol. 121 485

~~~ ~

particulate matter. A simplified homogenizing plug valve featuring a flat-faced valve stem to control the aperture between the valve stem and seat coupled to a pneumatic pressure multiplier circumvents the need for check valve assemblies. With this design the sample chamber is filled with the suspension, the sample inlet entry port is sealed and the sample is processed by actuating the three-way pneumatic flow controller. After increasing the aperture, the homogenate can then be aspirated back into the sample chamber by reversing the flow controller. Subsequent passes were typically performed with a slightly larger gap setting.

The envisaged strategy was to macerate the sample matrix with aqueous tetramethylammonium hydroxideethanol with a polytron-type device and then to subject the blended mixture to homogenization. In initial studies, which are the subject of this paper, determinations were restricted to ETAAS. It was considered that the capacity to vary both the duration and the temperature of the ashing cycle offered greater possibilities to volatilize selectively components of the sample matrix that might interfere in the subsequent analyte atomization stage of the determination. A serious limitation of the procedure became evident when solvent blanks were processed. Prior to determi- nation, these blanks had been subjected to either (i) the high- speed blending procedure followed by homogenization or (ii) the blending procedure alone or (iii) homogenization alone or (iv) were untreated. As summarized in Table 2, appreciable amounts of Pb and lesser amounts of Cu were mobilized/ solubilized during the mixing-homogenization processes. These concentrations were calculated by additions of analyte standard to the processed solvent mixture. For convenience, the observed analyte levels were expressed as if they had been present in either 2 g of fresh tissue or in 0.1 g of SRM. Whereas the background levels of analyte Cd, Pb and Cu in the aqueous TMAH-ethanol solvent mixture were acceptably low, blending at high speed for 60 s increased the levels of Pb only (fourfold). Processing the solvent mixture through the homogenizer appreciably increased analyte levels in all cases. Moreover, the net effect of the two procedures was approximately additive. Despite the high analyte levels introduced into the samples by the homogenizer, it was judged worthwhile to assess the stability of the resulting homogenates and their suitability for ETAAS. It was envisaged that if the feasibility trials were successful, modifications could subsequently be made to the homogenizer design to reduce the contamination appreciably.

Six powdered SRMs were chosen as substrates for the preliminary evaluation of the suitability of the slurry prepara- tion procedures. The variations in the certified concentrations of Cd, Pb and Cu among the six reference materials were 2023-fold (0.013-26.3 pg g-l), 77-fold (0.37-10.4 pg g-1) and 196-fold (2.34-439 pg g-l), respectively. Despite the high analyte backgrounds introduced by the slurry preparation, there was good agreement (Table 3) between the determined concentrations of cadmium, lead or copper in these materials, as estimated by the method of standard additions, and the

corresponding certified reference values. No statistically sig- nificant differences between the measured analyte levels and the certified values were evident. The largest discords between the determined concentration and the certified value were < 27% for Cd, < 23% for Pb and < 18% Cu. The mean uncertainty (as measured by the mean relative standard error of the estimate) for the different reference materials was +7.5% for Pb, f8.7% for Cu and +8.7% for Cd. Longer term repeatability of sample preparation was assessed by repeating the procedure with fresh sample after 1 month (Table 3, DORM- 1, DORM-2 and bovine muscle). Only for lead determinations in the DORM-2 SRM (certified concentration 0.065 pg g-1) were the data not sufficiently repeatable to be reported.

The slurry preparation procedure was also applied to frozen fresh cervine tissue. Three liver and three kidney samples from ring tail deer and two samples of each tissue from moose were analysed. These tissues had been analysed previously by conventional acid digestion followed by ICP-MS. The measured concentrations of Cu, Pb and Cd in these tissues are summarized in Table 4. Duplicate entries for the same matrix represent separate analyses performed at approximately 1 month intervals. Again, there was good general agreement between the two sample preparation and determination tech- niques. Only single estimates of analyte concentrations by ICP- MS were available; however, duplicate preparationddetermina- tions of the Cd, Pb and Cu levels in two reference materials (DORM- 1 and DOLT- 1 SRMs) had previously demonstrated that the techniques were being performed reliably. The degree of discord between the two techniques for Cu and Cd analyte was < 22% in all cases and < 15% in all but one case. For Pb, the discord was greater, < 3 1 %, in part because of number of significant digits used to report the data.

In Table 5, the means of the slopes of the best-fit regression lines for standard additions of each analyte to liver, to kidney, to SRMs or to homogenized solvent mixture are compared. There were no significant differences in the mean slopes of the regression lines among the different sample matrices for the same analyte provided that the analyses were performed on the same day. Moreover there were no significant differences among the slopes for the six different reference materials which, in turn, were not significantly different from the slope of an external standards calibration curve for the same analyte. Collectively, these observations suggest that, for these matrices at least, a single standard additions calibration could have been used to determine analyte concentrations in any of the samples. An external standards calibration technique might also have been equally effective. However, for SRMs containing lower analyte concentrations ([Cd] < 0.04 and [Pb] < 0.4 pg g-1) this was not the case. Arbitrarily for SRMs, 0.1 g samples had been slurried. For [Cu] in SRMs, the level of discord between the measured and reference values was up to 31% and appreciably larger than the corresponding discord between [Cu] measured by standard additions and the reference value ( < 22%).

~ ~ ~ ~ ~~ ~~~~ ~ ~ ~~ ~ ~ ~~~~~~~~~~~~~~~~~

Table 2 Analyte concentrations in solvent mixture (pg g- *, calculated as if the 20 ml of solvent had contained either 0.1 g or 2 g of sample) following various mixing treatments

Cadmium Copper Lead

Treatment 2 g sample 0.1 g sample 2 g sample 0.1 g sample 2 g sample 0.1 g sample None n.d. n.d. 0.0005 0.0 10 0.0049 0.0620 High-speed blending

(60 s) n.d. n.d. 0.0005 0.0100 0.0198 0.395 Homogenization

(4 sequential passes) 0.001 0.020 0.0 15 0.300 0.069 1.38 High-speed blending

(60 s) +homogenization 0.002 k 23% 0.042 k 23% 0.0154 f 15.6% 0.324 k 7.0% 0.076 k 9.2% 1.58 f 4.9%

Publ

ishe

d on

01

Janu

ary

1996

. Dow

nloa

ded

by U

nive

rsity

of

Wes

tern

Ont

ario

on

26/1

0/20

14 0

0:49

:17.

View Article Online

486 Analyst, April 1996, Vol. 121

For fresh tissues, 2 g of sample were slurried. The degree of discord for [Cd], as determined by ICP-MS and external standards ETAAS (Table 4), was < 21 % in all instances (mean relative difference 7.1 & 6.1%) and < 15% in 13 out of 14 cases. Similarly for [Cu] (Table 4), the relative difference between the two techniques was < 3 1% in all instances (mean relative difference 12 k 9.4%) and S17% in 13 out of 14 cases. By contrast, for Pb analyte the discord was excessive. This is not surprising given that the [Pb] in eight of the ten samples, as determined by ICP-MS, was less than the background level introduced by the homogenization procedure. Collectively, the results determined by external standards calibration were

reliable, provided that the analyte levels in the sample were appreciably greater than the contamination introduced by the slurry preparation procedure.

The stabilities of the slurries with time were assessed by performing analyses (by the method of standard additions) for Cd, Pb and Cu in five reference materials and four cervine tissues. Aliquots were removed from the undisturbed prepara- tions after 2 and 6 d post-preparation and analysed. The results as determined by standard additions calibration are recorded in Table 6. For convenience, the values representing 0 d post preparation from Tables 3 and 4 are reproduced in column 3 or Table 6. For SRMs there was no convincing evidence for any

Table 3 Cd, Pb and Cu concentrations (pg g-1) in SRMs as determined by slurry introduction-ETAAS with calibration by standard additions or external standards

Analyte SRM Cd DORM- 1

DORM-2

Bovine Liver Bovine Muscle

Oyster Tissue Lobster Hepatopancreas

Pb DORM- I

Bovine Liver Bovine Muscle Oyster Tissue Lobster Hepatopancreas

c u DORM- 1

DORM-2

Bovine Liver Bovine Muscle

Oyster Tissue Lobster Hepatopancreas

Standard additions (f 1 RSEE") 0.109f 13% 0.107 f 6.2% 0.053 f 21% 0.040 f 4.5% 0.48 f 6%

0.010 f 15% 0.016 f 6.0%

3.98 f 3.6% 26.7 * 3% 0.31 f 8% 0.43 f 9.0% 0.15 +7% 0.32 f 7.4% 0.35 f 6.9%

9.79 f 4.6% 4.48 f 14% 4.44 f 5.0% 2.21 k 12% 2.01 f 6.6%

124.4 f 7.5% 2.46 f 7.6% 2.35 It 8.4% 54.9 f 15%

425.2 f 2.8%

11.79 f 8%

External standards (f 1 S,+)

0.1 13 f 6.0% 0.094 f 14% 0.046 f 22% 0.040 f 6.9% 0.53 f 5.6%

0.046 f 22% 0.006 rt 16.7% 3.94 f 2. I % 27.2 f 3.7%

0.12 f 9.2%

0.53 f 10% 0.37 f 12%

12.03 f 2.8% 10.1 f 2.6% 6.29 f 10% 4.25 k 0.8% 2.78 f 3.4% 2.43 f 4.2%

116.1 +3.1% 2.80 f 0.7% 2.94 f 4.7% 77.8 f 5.7%

41 1.8 f 2.8%

Certified value/

0.086 f 14% I.18 g-'

0.043 f 19%

0.44 f 14% 0.013 f 85%

4.15 f 9% 26.3 k 8% 0.4 k 30%

0.135 f 11% 0.38 f 63% 0.37 f 3.8% 10.4 f 19%

5.22 _+ 6.3%

2.34 _+ 6.8%

158 f 4.4% 2.24 f 20%

66.3 f 0.4% 439 f 5.0%

* Relative standard error of estimate. + Relative standard deviation based on three replicate determinations of unamended slurry.

Table 4 Cadmium, lead and copper concentrations (pg g-1) in cervine liver or kidney as determined by slurry introduction-ETAAS and calibration by standard additions or by external standards

Sample B2-liver B2-kidney B4-liver

B4-kidney

B6-liver B6-kidney B 1 O-liver

B l0-kidney

B 1 1 -liver B 1 1 -kidney

Cadmium

Std. addns.* (k 1 RSE) 0.304 f 2.3% 2.20 rt 4.1 % 0.59 f 2.1% 0.58 f 3.4% 3.04 k 4.8% 3.29 f 3.4% 1.36 k 3.7% 8.69 f 3.9% 10.9f5.1% 10.5 f 4.9% 17.1 f4.6% 16.5 f 3.7% 3.35 f 2.9% 11.8 k 8.0%

Ext. stds.+ (+ 1 sr) 0.257 f 4.8% 2.15 f 2.5% 0.57 k 2.5% 0.57 f 5.2% 3.12 rt 0.09% 3.23 f 2.0% 1.44 f 2.4% 8.41 f 1.4% 11.3k 1.3% 10.3 f 1.4% 17.5 f 1.1% 16.3 f 1.2% 3.45 f 2.1% 12.2 f 1.2%

ICP-MSt 0.285 1.78 0.57

3.48

1.27 8.47

10.7

16.3

3.03 11.4

Std. addns.* (f 1 RSE) 0.036 f 8% 0.146 f 5% 0.032 f 9% 0.033 f 4.5% 0.048 f 8% 0.059 f 7.9% 0.029 f 7% 0.065 f 8% 0.345 f 12% 0.3 12 rt 4.4% 0.059 f 7% 0.053 f 7.7% 0.020 rt 5% 0.043 f 9%

Lead

Ext. stds.+ (+ 1 S J

0.005 k 3.7% 0.123 f 3.9% 0.005 k 4.9% 0.027 f 10.1% 0.052 f 3.4% 0.048 k 16.9% 0.005 k 2.4% 0.035 k 3.9% 0.338 f 2.9% 0.306 + 4.0% 0.073 f 1.2% 0.043 3- 12.9% n.d. 0.005 3- 2.4%

Copper

ICP-MS* 0.03 0.12 0.04

0.07

0.03 0.07 0.32

0.07

0.03 0.04

Std. addns.* (f I RSE) 49.4 f 4.8% 3.04 f 6.8% 44.0 f 3.1 % 40.6 k 1.7% 3.10f 8.1% 3.04 f 4.9% 83.3 f 7.0% 4.93 f 2.8% 3.42 f 3.3% 3.54 _+ 7.4% 4.56 k 6.7% 4.48 f 6.7% 37.0 f 2.3% 3.14 f 2.1%

Ext. stds.?

45.0 f 0.7% 2.14 f 10.0% 49.2 f 1.0% 41.5 f 2.3% 2.69 f 8.2% 3.25 f 3.3% 83.0 f 0.1% 4.77 f 1.9% 2.83 f 5.3% 3.08 f 12% 4.02 f 6.0% 4.02 k 6.0% 32.9 f I .4% 4.02 & 6.0%

(f 1 Sr) ICP-MSZ 47.1

40.5 3.11

3.17

93.8 4.66 3.30

4.39

43.3 3.28

* Calibration by standard additions (f1 relative standard error of estimate). + Calibration by external standards (+I relative standard deviation). * Determinations were performed by the Research and Productivity Council of New Brunswick, Fredericton, NB.

Publ

ishe

d on

01

Janu

ary

1996

. Dow

nloa

ded

by U

nive

rsity

of

Wes

tern

Ont

ario

on

26/1

0/20

14 0

0:49

:17.

View Article Online

Analyst, April 1996, Vol. 121 487

significant change in the apparent analyte concentrations over the 6 d trial. For each sample matrix, if the results for each analyte from the three different times were treated as replicates, the mean s, was, in 12 out of 14 cases, lower than the mean relative standard error of the estimate (RSEE) associated with the individual determinations of that group. In the other two cases, s, (22 and 20%) was appreciably smaller than the uncertainty associated with the reference value (30 and 63% s,). In only one instance (Cu in DORM-2) did the apparent analyte levels decrease with time to the extent that there was a 22% difference between the determined concentration at day 6 and the reference value. However, this decrease was not statistically significant at the 95% level of confidence.

For the cervine tissues, again there was no appreciable change in the apparent analyte concentrations with time. The degree of discord between the ICP-MS result and the deter- mined value at day 6 was less than 12% for all cases but one (Pb in the B4-liver sample). The apparent changes in [Pb] in the B4- liver sample with time were erratic (mean [Pb] = 0.047 f 28% versus 0.04 by ICP-MS) reflecting, in part, the high background levels introduced by the slurry procedure relative to the Pb analyte level in this sample. By contrast, the mean s, for all 12 analyte-sample matrix combinations was 9.6%. Collectively, these results indicate no segregation of the analyte within the sample over the 6 d trial. It seems certain that at least a high proportion of the analyte was extracted into the liquid phase during sample processing. For these matrices at least, short-term

sample storage had no discernible effect on the analyte apparent concentrations.

Conclusions The results indicate that high-pressure homogenization is capable of generating emulsions/dispersions of SRMs of animal origin or of fresh soft tissues which can be reliably sub-sampled and analysed by ETAAS for Cd, Pb or Cu. This is true provided that the levels of the analyte in these matrices are appreciably greater than the levels of contamination generated by the homogenization procedure. Attractive features of this approach include the speed (60 s blending followed by four passes through the homogenizer, total less than 3 min) and simplicity of sample processing and the apparent stability of the prepara- tion. The repeatability of the determinations performed (over 6 d) on the same homogenate indicate that at least a high proportion of the analytes is extracted into the liquid phase during sample processing. The cost of the homogenizing device is appreciably less than the cost of a commercial microwave digester. The homogenizer is easily cleaned between samples by processing fresh solvent. However, for the technique to become more widely applicable, background levels of analyte introduced by the homogenizer will have to be reduced substantially. It is expected that modifications to the flat-faced valve stem of the homogenizer will reduce levels of contamina- tion appreciably.

Table 5 Means of slopes (f 1 s,) of the best-fit linear regression models for standard additions of Cd, Pb or Cu to liver, kidney, SRMs or processed blank solvent mixture

Liver +kidney + Analyte Liver Kidney SRMs Solvent blank SRM Cd 1.002f0.55% 1.013 f3.06% 1.010f 1.02% 1.005 f2.49%' 1.009f 1.38% Pb 0.1855f0.91% 0.1819+0.78% 0.18585 1.55% 0.1837f 1.33%' 0.1847f 1.46% cu 0.2572 f 1.54% 0.2576 f 0.77% 0.2593 f 0.54% 0.2574 f 1.00%' 0.2585 f 0.82%

* f l relative standard deviation of the slope.

Table 6 Variations in apparent analyte concentrations (pg g-' f 1 RSEE) with time in unmixed slurries

Sample DORM- 1

DORM-2

Bovine Muscle

Oyster Tissue

Lobster HepPn

B4-liver

B4-kidney

B 1 O-liver

B 1 0-kidney

Analyte Cd Pb cu Cd cu Cd Pb cu Cd Pb cu Cd Pb c u Cd Pb c u Cd Pb c u Cd Pb c u Cd Pb c u

Od 0.109f 13% 0.31 f 8% 4.48 f 14%

0.053 f 2 1 % 2.21 f 12%

0.010 f 50% 0.32 f 7.4% 2.46 f 7.6% 3.98 f 3.6% 0.35 f 6.9% 54.9 f 15% 26.7 f 3% 9.79 f 4.6%

425.2 f 2.8% 0.59 f 2.1%

0.033 f 4.5% 40.6 f 1.7% 3.04 f 4.8%

3.10 f 8.1% 10.9 f 5.1%

0.312 rf: 4.4% 3.42 f 3.3% 17.1 f 4.6%

0.053 f 7.7% 4.56 f 6.7%

0.048 f 8%

2 d 0.109 f 5.7% 0.46 f 6.7% 4.59 f 4.9%

0.053 f 4.6% 2.002 10.1%

0.015 +42% 0.36 f 7.4% 2.5 1 f 8.8% 3.85 f 3.0% 0.38 f 7.8% 57.8 +. 14% 25.8 f4.7% 10.2 f 8.8%

415.2 f 3.1% 0.49 f 4.4%

0.050 f 7.7% 41.6 f 2.3% 3.39 f 2.9%

0.064 f 5.1% 3.64 f 9.2% 10.6 f 3.4%

0.302 f 3.1 % 3.64 f 9.2% 21.4f4.6%

0.057 f 7.7% 4.60 f 7.3%

6 d 0.107 f 8.3% 0.33 f 10% 4.73 f 4.8%

0.054 f 8.3% 1.82 f 7.1%

0.007 f 37% 0.47 f 8.8% 2.42 f 8.1% 3.87f4.1% 0.3 1 f 9.7% 58.2 f 7.8% 28.3 f 3.8% 10.5 f 4.7% 406 f 3.6%

0.52 f 5.4% 0.059 i 8.8% 42.7 f 1.8% 3.41 f 2.4%

0.066 f 6.4% 2.78 f 10.6% 10.4 f 3.8%

0.307 f 3.8% 3.71 f 7.4% 15.8 f4.8%

0.068 f 6.6% 4.65 f 6.6%

Reference value 0.086 f 14%

0.4 f 30% 5.22 f 6.3%

0.043 f 19% 2.34 f 6.8%

0.013 f 85% 0.38 f 63% 2.84 f 20% 4.15f9% 0.37 f 3.8% 66.3 f 6.5% 26.3 f 8% 10.5 It 19% 439 f 5%

ICP-MS

0.57 0.04

3.48 0.07 3.17

0.32 3.30

0.07 4.39

40.5

10.7

16.3

Publ

ishe

d on

01

Janu

ary

1996

. Dow

nloa

ded

by U

nive

rsity

of

Wes

tern

Ont

ario

on

26/1

0/20

14 0

0:49

:17.

View Article Online

488 Analyst, April 1996, Vol. 121

Cervine tissues and the results of ICP-MS determinations of their Cd, Pb and Cu contents were generously supplied by D. J. Ecobichon, Pharmacology and Therapeutics, McGill University.

References 1 2 3 4

5

6 7

8 9

10

11 12

Langmyhr, F. J., Analyst, 1979, 104, 993. Langmyhr, F. J., Prog. Anal. At. Spectrosc., 1985,8, 193. Miller-Ihni, N. J., Anal. Chem., 1992, 64, 965A. de Benzo, Z. A., Velosa, M., Ceccarelli, C. , de la Guardia, M., and Salvador, A., Fresenius’ J. Anal. Chem., 1991, 339, 235. Bendicho, C., and de Loos-Vollebregt, M. T. C. , J . Anal. At. Spectrom., 1991, 6, 353. Miller-Ihli, N. J., Fresenius’ J . Anal. Chem., 1993, 345, 482. Stephen, S. C., Littlejohn, D., and Ottaway, J. M., Analyst, 1985,110, 1147. Thompson, D. D., and Allen, R. J., At. Spectrosc., 1981, 2, 53. Madrid, Y., Bonilla, M., and C h a r a , C., J . Anal. At. Spectrom., 1989, 4, 167. L6pez Garcia, I., Ortiz Sobejano, F., and Hernandez Cbrdoba, M., Analyst, 1991, 116, 517. Hoenig, M., and Hoeyweghen, P. V., Anal. Chem., 1986,58,2614. Albers, D., and Sacks, R., 1987, Anal. Chem., 1987, 59, 593.

13

14 15

16 17 18

19 20

21

22

23

24

25

~ ~~~~

Tsalev, D. L., Slaveykova, V. I., and Mandjunkov, P. B., Spec- trochim. Acta, Rev., 1990, 13, 225. Miller-Ihli, N. J., J . Anal. At. Spectrom., 1989, 4, 295. Bendicho, C., and de Loos-Vollebregt, M. T. C., Spectrochim. Acta, Part B , 1990,45,679. Miller-Ihli, N. J., J. Anal. At. Spectrom., 1988, 3, 73. Lynch, S., and Littlejohn, D., J. Anal. At. Spectrom., 1989, 4, 157. Fagioli, F., Landi, S . , Locatelli, C., Righini, F., and Settimo, R., J. Anal. At. Spectrom., 1990, 5, 519. Hansen, D. L., and Bush, E. T., Anal. Biochem., 1967, 18, 320. Jackson, A. J., Michael, L. M., and Schumacher, H. J., Anal. Chem., 1972,44, 1064. Murthy, L., Menden, E. E., Eller, P. M., and Petering, H. G., Anal. Biochem., 1973,53, 365. Uchida, T., Isoyama, H., Yamada, K., Oguchi, K., Nakagawa, G., Sugie, H., and Iida, C., Anal. Chim. Acta, 1992, 256, 277. Dion, B., Ruzbie, M., van de Voort, F. R., Ismail, A. A., and Blais, J. S . , Analyst, 1994, 119, 1765. Tan, Y ., Momplaisir, G.-M., Wang, J., and Marshall, W. D., J . Anal. At. Spectrom., 1994,9, 1153. Momplaisir, G.-M., Lei, T., and Marshall, W. D., Anal. Chem., 1994, 66, 3533.

Paper 5f06966A Received October 23,1995 Accepted January 2,1995

Publ

ishe

d on

01

Janu

ary

1996

. Dow

nloa

ded

by U

nive

rsity

of

Wes

tern

Ont

ario

on

26/1

0/20

14 0

0:49

:17.

View Article Online