Anal. Chem. 1984, 56, 1285-1288
students and peers. His wit and humor will be sorely missed by his friends and students.
Registry No. I, 116-06-3;11,1646-75-9;111,10074-86-9;water, 7732-18-5. LITERATURE CITED (1) Guerrera, A. A. J.-Am. Water Works Assoc. 1981, 73 (4), 190-199. (2) Knaak, J. 6.;Tallant, M. J.; Sulllvan, L. J. J. Agric. Food Chem. 1966, 14,573-578. (3) Metcalf, R . L.; Fukuto, T. R.; Collins, C.; Borck, K.; Burk, J.; Reynolds, H. T.; Osman, M. F. J. Agric. Food Chem. 1966, 14,579-584. (4) Wright, L. h.; Jackson, M. D.; Lewis, R. G. Bull. Environ. Contam. Toxicol. 1982, 28, 740-747. (5) Sparaclno, C. M.; Hines, J. M. J. Chromatogr. Scl. 1976, 14, 549-556. (6) Move. H. A,: Scherer, S. J.: St. John, P. A. Anal. Left. 1977, IO, 1049-1 073. (7) Krause, R. T. J . Chromatogr. 1979, 185,615-824. (8) . . Galoux, M.; Van damme, J.-C.; Bernes, A.; Potvin, J. J . Chfomatogr. 1979, 177,245-253. I
.
1285
(9) Muszkat, L.; Aharonson, N. Inf . J. Mass Spectrom. Ion Phys . 1983, 4. 8_. .323-328. .~~ Moye, H. A. J. Agric. FoodChem. 1975, 23,415-418. Yost, R. A.; Fetterolf, D. D.; Hass, J. R.; Harvan, D. J.; Weston, A. F.; ~~
Skotnicki, P. A.; Simon, N. A,, submitted to Anal. Chem. Riva, M.; Carisano, A. J. Chromafogr. 1969, 42,464-469. United States Food and Drug Administration Pesticide Analytical Manual; US. Government: Rockvilie, MD, 1970; Volume 11, Pesticide Reg. Sectlon 120.269 Aldicarb. (14) Rouchaud, J.; Moons, C.; Meyer, J. A. Pesfic. Sci. 1980, 1 1 ,
483-492. (15) Andrawes, N. R.; Bagiey, W. P.; Herrett, R. A. J. Agric. Food Chem.
1971, 19,731-737. (16) Andrawes, N. R.; Bagley, W. P.; Herrett, R . A. J . Agric. FoodChern. 1971, 19,727-730. (17) Jones, A. S. J. Agric. Food Chem. 1976, 24, 115-117. (18) Payne, L. K.; Stansbury, H. A., Jr.; Weiden, M. H. J. J. Agric. Food Chem. 1966, 14,356-365.
RECEIVED for review December 27,1983. Accepted February 21, 1984. This project was supported by a State of Florida STAR grant. Funds for the GC/MS/DS were provided by the National Science Foundation.
Determination of Elemental Iodine by Gas Chromatography with Electron Capture Detection S. J. Fernandez,* L. P. Murphy, and R. A. Rankin Exxon Nuclear Idaho Co., P.O. Box 2800, Idaho Falls, Idaho 83401
A gas chromatography/electron capture detector (GC/ECD) technique has been developed to determine elemental iodine In toluene and cyclohexane solvents. The retentlon Index based on the alkyl Iodide and n-paraffin series was determined. Sufflclent resolution was obtalned to resolve I, from alkyl Iodides. A detection iimlt of 39 ng of I, was obtalned. A dlssoclatlve electron capture mechanism was proposed for the ECD response to 12. Radlotracer studles established the I, was transmitted through the column without conversion to an organlc lodlde. The utlllty of the developed technlque was demonstrated on a AgI in ",OH mass spectrometry standard solutlon. An accuracy of 23% and a preclsion of 12% were demonstrated.
Potentially, the most environmentally significant emissions from nuclear facilities are the radioactive isotopes of iodine. The many alkyl iodides, aromatic iodides, and inorganic iodides that may be formed during nuclear and chemical processes complicate the measurement of liquid and gaseous iodine compounds at the submicrogram level. Castello et al. ( I ) published the gas chromatographic separation and identification of several alkyl iodides using glass columns filled with tricresyl phosphate on DMCS treated Chromosorb W. Castello et al. (1)also proposed a homologous series of iodoalkanes as a reference for the calculation of the retention indexes of ECD sensitive substances. Corkill and Giese (2)used fused silica capillary columns coated with either SE52 or SE54 to analyze the iodothyronines. A gas chromatographic separation of Iz and methyl iodide prior to mass spectrometric analysis was developed in this laboratory (3). The analysis of I, by packed column gas chromatography is
more difficult than the analysis of organic iodides because I2 is less volatile and more reactive than the organic iodides of environmental significance. In this article the analysis of 12 by GC/ECD is investigated as a function of detector, injector, and column temperatures. The calculated retention indexes are compared to the results of Patte, Echeto, and Laffort ( 4 ) .
EXPERIMENTAL SECTION Apparatus. The separation and measurement of iodine compounds were performed with a Varian Model 3700 gas chromatograph equipped with a 63Nielectron capture detector. Two columns were used: a 30 cm X 0.3 cm stainless steel column of 10% OV-101on Chromosorb W-Hp 80/100 mesh obtained from Varian Instrument Group (Palo Alto, CA); and a 300 cm X 0.3 cm nickel column of 5% SE-30 on Chromosorb W-HP, SO/lOO mesh obtained from Alltech Associates (Arlington Heights, IL). All experiments were performed with a 30 cm3/min flow N2 or 90% Ar/lO% CHI carrier gas. The column was operated isothermally at 100 "C for the retention index experiments. Reagents. The methyl iodide, ethyl iodide, 1-iodopropane, and 2-iodopropane were obtained from Aldrich Chemical Co. (Milwaukee, WI) and used without further purification. The toluene, benzene, and cyclohexane were "distilled-in-glass"grade obtained from Burdick and Jackson (Muskegon, MI) and used without further purification. The elemental iodine was resublimed grade obtained from Fisher Scientific Co. (Pittsburgh, PA). Iodine Compound Mixture Preparation. Alkyl iodide solutions were prepared by volumetric dilution; the I2 solutions were prepared by weighing I2 crystals to fO.l mg, dissolving the crystals in toluene or cyclohexane, and preparing subsequent serial dilutions volumetrically. RESULTS AND DISCUSSION Retention Index. The retention index was determined in the manner of Castello et al. ( I ) and Patte et al. ( 4 ) at 100
0003-2700/84/0356-1285$01.50/00 1984 American Chemical Society
1286
ANALYTICAL CHEMISTRY, VOL. 56, NO. 8, JULY 1984 2.408
i
3.51
3.0
1 -1-Propyl
"C
i
Iodine
2-Propyl lodlne
'V
Benzene
0.401 0
1.5
0
t 1I
L
300
10
1
1
20
30
40
50
80
70
,
I
80
90
I
1
100
ng 12
i
OS
1
, / I
,
I l l
I
1
1
,
.
i
,
,
Flgure 3. Typical callbratlon curve for nanogram sized standards.
I2 on the SE-30 column using
400 1 500 1 600 700 1 800 900 lob0 100 200 300 400 500 600 '
Retenflon Index
Flgure 1. Adjusted retention time vs. retention index for the OV-101 column. The upper abscissa is the n-paraffin homologous series and the lower abscissa is the alkyl iodide homologous series. 3.0
1
1
I
I
ha
2.5
1
.
t Propyl lodlde
i
-
1
/
/Jg 12 Cyclohexane
Flgure 4. Calibration curve for I2 on the SE-30 column using mlcro~ 1.8. gram sized standards. The regression equation is y = 1 3 . 2 -
-
-
1.0-
??
o
300
I
I
0
I
400
I
'
100
1
500
1
I
200
, 600
I
I
1 700
1
1
300 400 Retentlon Index
1 800
1 1 500
I 900
I I
600
e
J
3
1000
=
Flgure 2. Adjusted retention time vs. retention index for the SE-30 Column. The upper abscissa is the n-paraffin homologous series and the lower abscissa is the alkyl iodide homologous series.
"C for both OV-101 and SE-30 by comparing the retention time of the Iz peak with methyl iodide, ethyl iodide, 2-propyl iodide, 1-propyliodide, benzene, toluene, and cyclohexane on the OV-101 column or with methyl iodide, 1-propyl iodide, and cyclohexane on the SE-30 column. The I, peak was identified by comparing the chromatograms of blank solvents with serial additions of a 13 mg/L 1, standard solution. I, was then assigned to the peak that increased in area in proportion to the amount of added I2 Plots of adjusted retention time (retention time of compound - retention time of 0,) vs. retention index for the two column types are shown in Figures 1 and 2. From these plots, a retention index on OV-101 of 172 f 14 for the alkyl iodide homologus series and 510 f 42 for the n-alkane homologous series was calculated. For SE-30, a retention index of 160 f 16 for the alkyl iodide homologous series and 500 f 52 for the n-alkane homologous series was calculated. Resolution. The height equivalent to an effective theoretical plate (HEETP) was calculated by using the equation
8
0.75-
c
1 X 8
3
0.5-
-n
0.25-
N
L
5.54(t'R/
wl,,)2
where L is the length of the column, tlR is the adjusted re-
0
-
-
,/,,,,,
HEETP =
a
8
I
I
i
I
I
I
Injecior Temperalure (QC)
Figure 5.
I2 response as a function of injector temperature.
tention time, and Wllzis the full width at half maximum of the Iz peak. For OV-101 at 100 "C, the HEETP for I, was 1.1mm; for SE-30 at 100 "C, the HEETP for I, was 39 mm. This was significantly more than reported by Castello et al. (1)for alkyl iodides (0.4-0.5 mm); but the I, peak was sufficiently resolved (resolution = 2.3) from CHJ for analysis. The broad Izpeak may be caused by the greater reactivity of I, compared to the alkyl iodides. S e n s i t i v i t y . A plot of ECD response vs. quantity of 1, injected is shown in Figure 3 (column, SE-30; temperature, 130 "C). From this plot, a lower limit of -39 ng of Iz for the quantitative method was observed. Below 39 ng of 12 significant departures from linearity were observed. The reason for the nonzero intercept of the linear portion of the calibration curve is unknown. Possible explanations include decompo-
ANALYTICAL CHEMISTRY, VOL. 56, NO. 8, JULY 1984
1287
Table I. lZsITracer Experiment Results fraction of transmitted iodine: 90 fraction of total iodine found that reached found as organic detector: % as I, iodine solvent cyclohexane toluene
81 t 10 88t10
99 i 2 9Oi2
It 2 lot2
$ 2 70
no
Column Temperature (OC)
Flgure 6.
a 210.
as determined by counting statistics.
I2 response as a function of column temperature. Table 11. AgI Analysis Results 33.0
makeup value, pg of I/mL 118 a +lo
aliquot l,a pg of I/mL 157
k
19
aliquot 2 ,a of I/mL
pg
133 i 1 6
as determined by precision of triplicate analyses.
1/T x 1000
of In (response X detector temperature. Figure 7. Plot
T3I2)
as a function of reciprocal
siton of the I2 in the GC at low levels (see tracer study) and instability of dilute Iz standards. An intercept not significantly different from zero was obtained when microgram quantities of I, were used to generate the calibration curve (Figure 4). These larger samples are more stable and can be prepared more accurately. The ECD response was studied as a function of injector temperature (Figure 5), column temperature (Figure 6), and detector temperature (Figure 7). From Figure 5 it appears the sensitivity of the GC/ECD technique is independent of injector temperature in the range of 190-220 "C, and sensitivity decreases with decreasing temperature below 190 "C. This decrease in sensitivity may be caused by an unknown reaction on the cooler injector surfaces. The insensitivity of the technique to minor column temperature changes is shown by the constant response when the column temperature was varied in the range of 100-130 "C (Figure 6). The mechanism of the ECD response to I2 was studied by plotting the In (response X W 2 vs. ) reciprocal detector tempeature, where 2' is the detector temperature. As shown in Figure 7, a peak in this plot is observed at about 300 "C. Corkill and Giese (2) also observed a peak in the response vs. reciprocal detector temperature plot for the iodothyronines. They attributed the decreasing response with increasing temperature to decomposition of the iodine compounds on the narrow bore glass inlet and the stainless steel and ceramic in the electron capture detector before the iodine compounds reach the collector region. The same phenomenon was also observed by Miller and Grimsrud (5). If one considers only the region between 250 "C and 300 "C the response increases with increasing temperature indicating a dissociiative mechanism. For these reasons, it appears likely that a dissociative
c Figure 8. Typical chromatogramsof I2 extraction of mass spectrometric standard: (a) 3 KL toluene blank; (b) 3 pL aliquot 1 ; (c) 3 p L aliquot 2.
mechanism is responsible for the ECD response to I,. Measurement of I2Decomposition. A tracer experiment was performed to confirm that I2 was not converted to an organic iodide in the injector or on the column and to measure directly the decomposition of Iz in the column and injector. I, in toluene and cyclohexane was tagged with IZsI. The effluent of the column was routed to an iodine species selective adsorbent sampler in the manner of Ktller et al. (6). Samples containing -40 ng of I2 were injected. The amount of lZ5I injected and collected by each selected adsorbent was then measured by direct X-ray spectrometry of the 27-keV Te KCY X-ray using a hyperpure germanium low-energy photon spectrometer. Collection periods corresponded to the time required to produce a chromatogram (10 min). The results of these experimentsare shown in Table I. From these results it appears that only 20% of the Is decomposed either in the injector or on the column. In addition, more than 80% of the I, injected onto the column was transmitted through the column without conversion to an organic iodide. Demonstration of Quantitative Method. The applicability of the developed GC technique to quantitatively measure 1, in organic solutions was demonstrated by analyzing a AgI in concentrated ",OH mass spectrometric standard. This standard was stored in a stoppered polypropylene test tube. Iodine-125 tracer was added to the samples and allowed to isotopically exchange overnight. Two l-mL aliquots of the solution were rapidly acidified to pH -2 with "Os. The nitric acid was covered with a 10-mL layer of toluene at all times to prevent I2 volatilization losses. One milliliter of
1288
Anal. Chem. 7984, 56, 1288-1293
saturated NaNO, was added to oxidize any free I- to I,; the resulting 1, was extracted into the toluene. Three microliters of the toluene was injected onto the GC to measure the formed 1,; one milliliter of the toluene was counted with a hyperpure germanium low-energy photon spectrometer to determine the chemical yield. Typical gas chromatograms are shown in Figure 8. The results of these analyses are shown in Table 11. The precision of triplicate determinations was 12% at one standard deviation. Although the determinations were within the 95% confidence interval, the analyses were 23 % higher than the nominal makeup value. The cause of this bias is unknown. On the basis of these results, it appears the technique has an accuracy of 23% and a precision of 12% at the one standard deviation level. Registry No. I,, 7553-56-2;AgI, 7783-96-2; CH31, 74-88-4;
CH3CH21,75-03-6; 2-propyl iodide, 75-30-9; 1-propyl iodide, 107-08-4.
LITERATURE CITED (1) Castello, G.; D'Amato, G.; Biagini, E. J. Chromatogr. 1989, 41, 313-324. (2) Corkill, J. A.; Giese, R. W. Anal. Chem. 1981, 53, 1667-1672. (3) Fernandez, S.J.; Rankln, R. A.; McManus, G. J.; Nielsen, R. A,; Deimore, J. E.; Hohorst, F. A.; Murphy, L. P. "Determination of Low Specific Activity Iodine-129 Off-gas Concentrations by GC Separation and Negative Ionization Mass Spectrometry"; ENICO-1134, 1983, 11-29. (4) Patte, F.; Echeto, M.; Laffort, P. Anal. Chem. 1982, 54, 2239-2247. (5) Miller, D. A.; Grimsrud, E. P. J. Chromafogr. 1980, 190, 133-135. (6) Keller, J. H.; Duce, F. A.. Maeck, W. J. "A Selective Adsorbent Sampling System for Differentiating Airborne Iodine Species"; CONF700816, Eleventh AEC Air Cleaning Conference, 1970; Vol. 2, pp 62 1-625.
RECEIVED for review June 15,1983. Accepted March 19,1984. Resubmitted February 21, 1984.
Collection and Determination of Volatile Organic Mercury Compounds in the Atmosphere by Gas Chromatography with Microwave Plasma Detection David S. Ballantine, Jr.,*and William H. Zoller'
Department of Chemistry, University of Maryland, College Park, Maryland 20742
A method for the collection of two volatile organlc mercury compounds In the atmosphere Is described, uslng Chromosorb 101 as a collection substrate. The analytlcal method Involves direct eiutlon of the organic mercury compounds from the collectlon substrate onto a gas chromatographic column prlor to detection wlth a mlcrowave plasma detector. Methylmercury chloride (MMC) Is collected at ambient temperatures, and dlmethylmercury (DMM) Is collected by use of a cryogenic trap at -80 O C . Coilectlon efflclencles for MMC and DMM are 95 f 3 % and 96 f 2 %, respectlvely. The absolute detection llmlt of the system Is 0.05 ng, wlth a detectlon limit for real atmospheric samples of 0.1 ng/m'. Posltlve Identification of collected compounds Is achieved by comparison of sample elution volumes with standards.
A growing public interest in environmental quality has led to the development of analytical techniques for the monitoring of environmental pollutants. Due to its acute toxicity and its tendency to bioaccumulate, mercury is of prime interest. Being extremely volatile in the organic and elemental forms, mercury is well dispersed in the atmosphere. The activity of certain bacteria, molds, and enzymes in the soil or sediment can produce methylated mercury from elemental or inorganic mercury (1-4). The organic mercury compounds produced, primarily dimethylmercury and methylmercury halides, are potentially more toxic than inorganic mercury forms. Therefore, recent studies of environmentalmercury have been concerned with its chemical speciation to determine not only the amounts of mercury present but the chemical forms as well. More extensive data in this area will assist in deterCurrent address: INC-7, MSJ514,Los Alamos National Labo-
ratory Los Alamos, NM 87544.
mining the role of organic mercury in the global cycling of the element. Previous studies have been performed by using a method of selective preconcentration followed by pyrolysis and cold vapor atomic fluorescence detection to determine different mercury species collected from the atmosphere (5). Other studies have been performed by using sequential specific absorption tubes which separate different chemical forms of mercury by selective collection (6-8). In these latter studies, mercury compounds were thermally desorbed and re-collected on gold surfaces prior to elution into an emission detector. While these methods represent a significant advance in atmospheric mercury sampling by achieving the separation of volatile species of mercury, the analytical methods prevent positive identification of the compounds by converting all forms to elemental mercury prior to detection. Chromatographic substrates have been used successfully for the collection of organics and of organic mercury (5-10). By logical extension, a chromatographic method of analysis would permit the positive identification of organic mercury compounds by comparison of sample elution times/volumes with standard compounds. For this study, a collection method has been developed that is compatible with a chromatographicmethod of analysis and is capable of detecting levels of organic mercury in the atmosphere as low as 0.1 ng/m3. Since they are known to be produced by biogenic activity (1-4), dimethylmercury and methylmercury chloride were selected as standard compounds during the laboratory studies. These compounds were also used as model compounds in the development of the selective absorption tube system of Braman and Johnson (6),but because of their choice of analytical method, any organic compounds detected could not be positively identified and could only be operationally defined to be methylmercury chloride or dimethylmercury.
0 1984 American Chemical Society 0003-2700/84/0356-1288$01,50/0