Infrared analysis of refined uranium ore - ACS Publications - American

1972, 2,. 95-106. (11) Mizutanl, T.; Matsumoto, M. FoodHyg. Soc. Jpn. 1972, 13, 398-404. (12) Armour ... and U308 were prepared that contained 0% ammo...
0 downloads 0 Views 667KB Size
2134

Anal. Chem. 1985, 57,2134-2138

the Aroclors. Koka Jayasimhulu is thanked for performing the GC/MS analyses. Redstry No. DCB, 2051-24-3;SbC&,7647-18-9; Aroclor 1016, 12674-11-2;Aroclor 1242, 53469-21-9;Aroclor 1254, 11097-69-1; Aroclor 1268, 11100-14-4.

LITERATURE CITED (1) Higuchi, K., Ed. "PCB Poisoning and Pollution"; Academic Press: New York, 1976. (2) Cordle, F.; Corneliussen, P.; Jeiinek, C.; Hackiey, B.; Lehman, R.; McLaughlin, J.; Rhoden, R.; Shapiro, R. EHP, Envlron. Health Perspect. 1978, 2 4 , 157-172. (3) Hutzinger, 0.; Safe, S.; Zitko, V. "The Chemistry of PCBs"; CRC Press: Cleveland, OH, 1974. (4) Calms, T.; Slegmund, E. G. Anal. Chem. 1981, 53, 1183A-1193A. (5) Chen, P. H.; Gaw, J. M.; Wong, C. K.; Chen, C. J. Bull. Environ. Contam. Toxlcol. 1980, 25, 325-329. (6) Zitko, V. Bull. Environ. Contam. Toxlcol. 1970, 5 , 279-285. (7) Brownlow, C. S. M.S. Thesls, University of Cinclnnati, Clncinnati, OH, 1982. (8) Webb, R. G.; McCaii, A. C. J . Chromatogr. Scl. 1973, 7 7 , 366-373. (9) Berg, 0. W.; Diosady, P. L.; Rees, G. A. V. Bull. Envlron. Contam. rox;col. 1972, 7, 338-347. (10) Hutzinger, 0.;Safe, S.; Zitko, V. J . Envlron. Anal. Chem. 1972, 2, 95- 106. (1 1) Mizutani, T.; Matsumoto, M. FoodHyg. SOC.Jpn. 1972, 13, 398-404. (12) Armour, J. A. J . Assoc. Off. Anal. Chem. 1973, 56, 987-993. (13) Huckins, J. W.; Swanson, J. E.; Stalling, D. L. J . Assoc. Off. Anal. Chem. 1974, 57, 416-417.

(14) Trotter, W. J.; Young, S. J. V. J . Assoc. Off. Anal. Chem. 1975, 58, 466-468. (15) Takamiya, K. Bull. Environ. Contam. Toxicol. 1983, 30, 600-605. (16) Lee, R. E.; Bramston-Cook, R.; Tschida, J. Anal. Chem. 1983, 55, 626-629. (17) Dunn, W. J., 111; Stalling, D. L.; Schwartz, T. R.; Hogan, J. W.; Petty, J. D.; Johansson, E.; Wold, S. Anal. Chem. 1984, 56, 1308-1313. (18) "Occupatlonal Exposure to Polychlorinated Biphenyls (PCBs)"; DHEW (NIOSH) 1977; Publication No. 77-225, pp 204-205. (19) "Handbook for Analytical Quality Control in Water and Wastewater Laboratories"; U S . EPA: Cinclnnati, OH, EPA 600/4-79-019, p 2-2. (20) Que Hee, S. S.; Ward, J. A.; Tabor, M. W.; Suskind, R. R. Anal. Chem. 1983, 55, 157-160. (21) Kovats, E. S. Adv. Chromatogr. 1985, 7 , 229-247. (22) Sissons, D.; Welti, D. J . Chromatogr. 1971, 60, 15-32. (23) Mullin, M.; Sawki, G.; Safe, L.; McCrindie, S.; Safe, S. J . Anal. ToxiCOl. 1981, 5 , 138-142. (24) Atkins, P. W. "Physical Chemistry"; W. H. Freeman and Co.: San Francisco, CA, 1978; pp 863-866.

RECEIVED for review March 25,1985. Resubmitted May 20, 1985. Accepted May 20,1985. This work was presented at the 187th American Chemical Society Meeting, April 8-13, 1984, St. Louis, MO, as ENVR 20. ES000159 was partially responsible for financial support. J. M. Lin was the recipient of a graduate assistantship in the Department of Environmental Health.

Infrared Analysis of Refined Uranium Ore Arthur F. Eidson Inhalation Toxicology Research Institute, Louelace Biomedical and Environmental Research Institute, P.O. Box 5890, Albuquerque, New Mexico 87185

Infrared assay of reflned uranlum ore (yellowcake) Is described and the results are related to worker protectlon measures. Eleven standard mixtures of ammonlum dluranate and U308were prepared that contalned 0% ammonium dluranate (pure U308)through 100% ammonium dluranate (no U308)In 10% Intervals. Assay of these mlxtures (0.30% In KBr) showed that ammonium dluranate could be accurately assayed wlthln f7% standard error of the mean (n = 8) and U308to withln 12%. For speclmens that contalned only one of the uranlum forms, the percentage of ammonlum dluranate was overestimated by 16 f 4 % and U,08 was underestlmated by 24 f 2%. Flfty-six commerclal samples from 10 mllls were assayed. The results were applled to the use of urlnalysls data to estlmate the amount of uranlum In the body of a worker after a hypothetical lnhalatlon of dust from an assayed sample. I t was shown that the uncertalnty In body burden estimates could be reduced from a factor of 100 to a factor of 10 wlth 95% confldence. Infrared assay results also showed that the ammonium dluranate and U,08 content of a speclflc yellowcake sample cannot be predicted from the dryer temperature alone.

Uranium ore is refined in uranium mills to produce the commercial product known as yellowcake. Processes vary among mills, but three basic steps are commonly used: leaching of crushed ore, recovery of leached uranium, and drying of product for packaging. Leaching is accomplished with HzS04or Na2CO3/NaHCO3solutions. Uranium dissolved in HzSO4 is removed by solvent extraction or ion exchange and precipitated from solution by ammonia, as ammonium 0003-2700/85/0357-2 134$01.50/0

diuranate. Ammonium diuranate (often written as (NH4)zU207)is actually a variable mixture of UO3.xNH3.yHz0 compounds with their composition dependent on the pH during precipitation. Four stoichiometric forms can be crystallized by shaking under an ammonia atmosphere for 2 weeks (1); but these conditions do not occur in industry. Uranium dissolved in NaZCO3/NaHCO3is precipitated by NaOH. Sodium diuranate is generally dissolved and reprecipitated by ammonia. Prior to packaging, ammonium diuranate precipitate is dried at temperatures chosen to either dehydrate it or convert it to U,Os. Partial or incomplete conversion often occurs. Both ammonium diuranate and UB08are potentially toxic, if inhaled, requiring that dry yellowcake packaging be done under regulations that specify ventilation equipment and other safety measures to protect workers. Because ammonium diuranate is more soluble than U308 (2-5) and dissolved uranium in the form of UOZz+is absorbed into blood and excreted in urine, chemical toxicity to kidney and accumulation in bone are possible. Dissolution and excretion of U308 also occur, but a t a much slower rate, so that gradual accumulation of U308 in lung could also deliver an appreciable radiation dose. Routine urinalysis for uranium is used to monitor protection measures so that workers who might inhale yellowcake dust do not accumulate dangerous amounts of internally deposited uranium. It is necessary, then, to quantitatively measure ammonium diuranate and U308 in yellowcake to interpret urinalysis data, either as part of routine monitoring or as an evaluation technique for accidental exposures. Dissolution studies in vitro are useful for studies of a few selected samples, such as a sample of a lot involved in an accident, but they are too time-consuming for use in a survey of yellowcake samples 0 1985 American Chemical Society

ANALYTICAL CHEMISTRY, VOL. 57, NO. 11, SEPTEMBER 1985

(2-4). X-ray diffraction techniques were considered for yellowcake assay (5);however, industrial products often include large fractions of poorly crystallized or amorphous ammonium diuranate, precluding accurate assay. Infrared absorption measurements were chosen to assay the two uranium forms regardless of crystallinity. The objectives of this study are to describe the assay of mixtures of standard ammonium diuranate and us08 by infrared absorption, to assay commercial samples obtained from operating mills, and to demonstrate application of the results.

2135

'""I

I

EXPERIMENTAL SECTION Ammonium diuranate was prepared in the laboratory by dropwise addition of 15% aqueous ",OH to an aqueous solution of U02(N03)2with stirring at room temperature. When the pH increased to 7.5, NHIOH addition was stopped and the resulting yellow precipitate was stirred for 16 h, fiItered, washed with cold water and acetone, and air-dried at room temperature. Ammonium diuranate prepared as above was ground with Us08 (New Brunswick Laboratory, Argonne, IL) using a Wigl-Bug grinder (Crescent Dental Mfg. Co., Chicago, IL) to prepare 11 mixtures that contained from 0% ammonium diuranate (pure U308)to 100% ammonium diuranate (no U308present) in 10% intervals. Weighed aliquots of the mixtures were added to 1 g of spectral grade KBr to prepare standard mixtures that were 0.30 f 0.01%, 0.50 f 0.01%, and 1.00 f 0.01% in KBr. All masses were measured to within f0.05mg. Duplicate pellets were prepared by pressing 200 mg of the ground mixture at 2000 psi for 5 min. Pellets were 0.052 f 0.001 cm thick. Fifty-six yellowcake samples were obtained from 10 commercial mills and KBr pellets that contained 0.3% sample in KBr were prepared as above. There was no pretreatment of yellowcake samples prior to grinding with KBr. Prior to use, KBr was heated at 100 "C and stored in a desiccator. All pellets were stored in desiccators but were not reheated to avoid decomposition of ammonium diuranate. Mill designations A-G and I-K do not identify the mills. Infrared absorption measurements were obtained with a Perkin-Elmer Model 283B infrared spectrophotometer equipped with a microprocessor control unit and programs for quantitative analyses of mixtures using the Beer-Lambert law. Absorbance measurements were made at wavenumbers assigned in the literature to U308 and to the uranyl moiety of U03 and at an intermediate frequency chosen to represent the remaining sample matrix. Base line points were chosen at 970 f 5 cm-l and 635 f 5 cm-l q the relative minima of the standard spectra. They represent the minimum of several overlapping absorptions, rather than wavelengths of 100% transmittance, but are the best available wavelengths in the 200 cm-' to 1500 cm-' range (Figure 1).

The 11standard ammonium diuranate plus U308mixtures were used to calibrate the instrument. One duplicate standard pellet of each mixture was selected randomly. The absorbances of these pellets were used to obtain the absorptivity array that described the entire range of ammonium diuranate and U308percentages in a mixture. The remaining pellets of each pair were then analyzed as unknowns. The process was then reversed to provide an analysis of all standard pellets as if they were unknowns. The absorptivityarray derived in this manner includes mixture and pellet preparation errors and instrumental errors between individual spectra obtained during the same analysis session. The array does not include instrumental errors introduced by changing settingsfor analysis of other materials and resetting the instrument for yellowcake analysis. Estimates of uranium body burdens for a hypothetical inhalation of an assayed yellowcake sample by a worker were calculated using equations given in ICRP Publication 30 (6). Ninety-five percent confidence limits on the estimates were calculated by using the precision of the assay and the function minimization program EOUAF (Numerical Algorithms Group, Downers Grove, IL).

RESULTS Changes in the infrared transmittance spectrum of a commercial yellowcake specimen upon thermal conversion to U308 are shown in Figure 1. The unheated specimen spectrum (A)

cm-1

Figure 1. Infrared transmittance spectra of a yellowcake specimen from mill B before (A) and after (B) heating at 150 "C for 16 h, showing partial conversion of ammonium diuranate to U30,. The band kfentlfled as UO, represents the uranyl moiety absorptlon ( 7 ) .

u1

I

or

1000

+q-\ 735

I

1

1

1

I

,

800

600

cm-'

Figure 2. Infrared absorbance spectra of pure ammonium dluranate (A), pure U308(B), and a yellowcake from mill D (C). All specimens were 0.30 f 0.01 wt % in KBr. Mill D curve is offset for clarity.

Analyses were done with wavelengths specified. Table 1. Assay of Standard Mixtures of Ammonium Diuranate and U308in KBr Pellets wt % mixture

in KBr fO.O1 0.30 0.50 1.00

deviation (analyzed % - known wt % in mixture) mean f std error (n) ammonium diuranate U308 0.40 f 2.6 (22) 0.070 f 6.6 (8)" 0.068 f 3.4 (22) -0.36 f 3.2 (22)

0.48 f 2.6 (22) -0.13 f 12 (8)o -0.34 f 4.8 (22) 0.22 f 3.8 (22)

O0.3% standard mixtures were assayed four times during the study of commercial samples. These values are considered more reliable for routine assays.

shows the presence of NH3 and HzO and a band assigned to the asymmetric uranyl stretching frequency of UOs at 920 cm-' (7). Spectrum B was measured after the specimen was heated at 150 "C for 16 h. It shows nearly complete loss of NH3 and HzO and the appearance of a band near 740 cm-l corresponding to u308 (8). Absorptions by NH and OH species in the sample are the major interferences in the 600 cm-' to 1000 cm-l region of interest (Figure 1)(9). Sulfate ion and carbonate ion can also interfere (9),but which can represent from 1%to 7% of yellowcake and C032- (I. *No absorption bands fir this form observed. O n

Table 11. Assay of Standard Pellets that Contained Either Ammonium Diuranate or UaOsin KBr deviation (analyzed % - 100%) mean % f std error (n = 8 ) ammonium wt %

component in KBr

diuranate

U308

0.30 f 0.01

15.5 f 3.8

-24.1 f 2.1

ammonium diuranate and U30, are slightly overestimated, but not significantly so when compared with the precision of the estimates. The 22 values arose from measuring duplicate pellets from 11 mixtures. On the basis of these analyses, results of the 0.3%pellets were shown to be more precise than those for the 0.5% or 1.0% pellets. After the 0.3%pellets were chosen for unknown assays, the standards were rescanned and assayed four times during the subsequent work, and the results are shown in Table I with n = 8. The results of these four separate assays of the 0.3% pellets, with error bars representing the standard error of the mean, are shown in Figure 3. The greater standard errors for 8 than for 22 measurements resulted not only from the smaller number of measurements but represent possible changes in the instrument or the standard pellets with time and are considered more reliable for routine work. The accuracy and precision of results are greatest for mixtures containing 10% to 90% ammonium diuranate and 30% to 70% U308. Given the accuracy and precision of standard mixture assays (Table I), unknown specimen assays showing less than ~ 1 3 %ammonium diuranate (f2 standard error) or less than -24% U308might indicate the presence of only U,08or ammonium diuranate. A calibration curve was derived from peak area analyses of 0% (pure KBr), 0.3%, 0.570, and 1.0% ammonium diuranate or U30s and used to assay pellets that contained only 0.3% standard ammonium diuranate or U308. The results (Table 11) indicate that ammonium diuranate was overestimated by 16% and U308was underestimated by 24%. The diminished accuracy of analyses that included 0.5% and 1.0% standards in the working curve was attributed to bias introduced by the sloping base lines of standard spectra and to NH and OH interferences that were more pronounced at the higher concentrations in KBr. The spectrum of a sample from mill D (curve C, Figure 2) illustrates a typical commercial sample assay. The analytical approach used was to assume that an unknown sample was

I

1000

I

800

1

1

600

cm-‘

Flgure 4. Spectra of 0.30 f 0.01 % specimens taken from two drums from lot no. 55 produced by mill E. This Illustrates the variable nature of ammonium diuranate produced during apparently similar conditions.

a mixture of ammonium diuranate, U308,and other interfering milling products. If the results indicated the yellowcake contained only ammonium diuranate or U308 (and the spectrum showed only one form present), the sample was reanalyzed with calibration data from the single component standards (Table 11). Table 111 shows assay results of samples taken from the mills studied to date. Results of analyses of unknown samples containing only one uranium compound were corrected for the systematic errors shown in Table 11. Results that indicated greater than 108% ammonium diuranate in a specimen were reported as containing only that form of uranium. Figure 4 illustrates the most extreme variability observed, which occurred among grab samples from two drums of mill E, lot no. 55. Both spectral bands in the drum no. 42 sample can be assigned to peaks from ammonium diuranate (7) and the spectrum shows no UB08absorbance. Mills C, D, and J produced yellowcake that was partially converted to U308. The mill I sample was a special case, It was collected from the floor of the packaging area and represents a mixture of production lots and other dusts generated in the overall milling process that might be resuspended and inhaled by a worker. Note that, although no U308was present, the specimen was not pure ammonium diuranate. This information would be valuable in assessing the consequences of an accidental inhalation of this material. Ammonium diuranate variability of the type shown in Figure 4 also complicated the analysis of mixtures (Figure 5). Such specimens were analyzed using the same 11 standard pellets used in all other assays, but wavenumbers were chosen to correspond to the U30smaximum, the minimum between the U308and ammonium diuranate peaks, and the wavenumber at the “crossover”between the ammonium diuranate standard and the unknown spectrum. The accuracy and

ANALYTICAL CHEMISTRY, VOL. 57, NO. 11, SEPTEMBER 1985

2137

1,000

,000

100

10 I

1 I

0

1

I

I

I 1

I I

I I 1

I I I

I

Figure 5. Spectra of mill D sample containing variable forms of ammonium diuranate (A) and binary mixture of 50% ammonium diuranate plus 50% U308 standard (6).The instrument was recalibrated for assay of this specimen using absorbances of the 11 standard mixtures at these new wavenumbers. Results were similar to those shown in Table I. precision were similar to those shown in Table I; mean and standard error (n = 11) were ammonium diuranate -2.3 f 2.7% and U308-2.0 f 3.9%. DISCUSSION Infrared analysis can accurately assay ammonium diuranate in the presence of U308to within f 7 % standard error of the mean, and U308content to within =k12%. For specimens suspected to contain only one of the uranium oxide forms and reanalyzed accordingly, the percentage of ammonium diuranate was overestimated by 16 f 4% and U30s was underestimated by 24 f 2%. Urinary Bioassay Interpretation. If a urine specimen from a worker contains uranium, yellowcake might have been accidentally inhaled and the amount remaining in the body must be estimated. Previous research showed a potential role for in vitro dissolution results in the interpretation of urinary bioassay data by estimating the relationship between yellowcake deposited in lung and uranium excretion in urine (4). Infrared assay results for ammonium diuranate and U308in yellowcake could be similarly useful but will require validation by comparison with results of accidental human exposures or inhalation studies using animals. Equations derived for health protection planning purposes (6) relate distribution kinetics of inhaled uranium compounds to their dissolution rates; where U03 is considered to have a dissolution half-time of l o 0 days. Figure 6 shows body burdens estimated assuming 1 hg of U/L of urine detected sometime within 100 days after a hypothetical inhalation of yellowcake. If the estimates are made assuming the inhaled material contained only ammonium diuranate with a dissolution half-time of 1 day (4) or U308 with a dissolution half-time of 100 days, they could be in error by a factor of 100. Results for a specimen from mill C that contained 52.7% ammonium diuranate and 56.0% U308were used to illustrate the effect of their uncertainty on body burden estimates if a worker inhaled this material. The shaded region (Figure 6) shows the result. The maximum and minimum curves were calculated by using the 95% confidence limits for mixture analysis (Table I, n = 8). Interpretation of urinalysis data with the aid of infrared assay results avoids the necessity to assume the worker inhaled only ammonium diuranate or U308 and appreciably reduces the uncertainty of body burden estimates. Yellowcake Composition Variability. Commercial yellowcake composition or its biological behavior cannot be reliably predicted from its drying temperature alone (Figure 7). The range in drying temperatures, obtained from mill

I 20 40 60 80 DAYS AFTER INHALATION

100

Figure 6. Estimated body burdens of a mill C specimen in a hypothetical worker based on infrared assay results compared with ICRP Publication 30 estimates for pure ammonium diuranate or pure U308. Shaded area represents the 95 Yo confidence interval calculated from precision of assay results. Use of infrared assay can reduce the uncertalnty in body burden estimates from a factor of 100 to approximately a factor of 10. Mill F

\I

IS I

\Mill C

w b-

u

60-

z U a 2 n I 402 z 0 I

z I-

z w u

i_ Mill D

20-

Mill J

a W 4.

,