830 (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22) (23) (24)
Anal. Chem. 1982, 5 4 , 830-832 Crouch, W. H., Jr. Anal. Chem. 1982, 34,1698. Swoboda, P. A. T.; Lea, C. H. Chem. Ind. (London)1958, 1090. Schnepfe, M. M. Anal. Chlm. Acta 1972. 58, 83. Hlnze, W. L.; Humphrey, R. E. Anal. Chem. 1973, 45. 814. Hlnze, W. L.; Humphrey, R. E. Anal. Chem. 1973, 45, 385. Taylor, S. L.; Flna, L. R.; Lambert, J. L. Appl. Mlcroblol. 1970, 20, 720. Lambert, J. L.; Flna. G. T.; Flna. L. R. Ind. Eng. Chem., Prod. Res. Dev. 1980. 79, 258. Hatch, G. L.; Lambert, J. L.; Flna, L. R. Ind. Eng. Chem., Prod. Res. Dev. 1980, 79, 259. Smith, C. A.; Spleth, H. H. Anal. Chem. 1973, 45, 422. Sugawara, K.; Koyama, T.; Terada, K. Bull. Chem. SOC.Jpn. 1955, 28,494. Novick. R. P. Blochem. J . 1982, 8 3 , 236. Lambert, J. L. Anal. Chem. 1951, 23, 1251. Hanes, C. S. New Phytol. 1937, 36,189. Rundle, R. E.; Baldwin, R. R. J . Am. Chem. SOC. 1943, 65, 544. Rundie, R. E.; French, D. J . Am. Chem. Soc. 1943, 65, 558.
(25) Rundle, R. E.; Foster, J. F.; Ealdwln, R. R. J . Am. Chem. Soc. 1944, 66. 2116. (28) Gilbert, G. A.; Marriott, J. V. R. Trans. Faraday Soc. 1948, 4 4 , 84. (27) Watanabe, T.; Ogawa, K.; Ono, S. Bull Chem. Soc. Jpn. 1970, 43, 950. (28) Thompson, J. C.; Hamorl, E. J . Phys. Chem. 1971, 75, 272. (29) Lambert, J. L.; Rhoads, S. C. Anal. Chem. 1958, 28, 1629. (30) Krishnaswamy, K. 0.; Sreenlvasan, A. J . Blol. Chem. 1948, 778, 1253. (31) Schoch, T. J. "Advances In Carbohydrate Chemlstry"; Pigman, W. W., Wolfram, M. L., Eds.; Academic Press: New York, 1945; Vol. 1. p 259. (32) Dlkeman, E. A. M.S. Thesis, Kansas State University, 1972.
RECEIVED for review January 16,1981. Accepted December 28,1981. This work was supported in part by National Science Foundation Grant CHE79-15217.
Indirect Determination of Cyanide by Single-Column Ion Chromatography D.
L. DuVal, J. S. Fritz," and D. T. Gjerde'
Ames Laboratory USDOE,Iowa State lJnivers/?y, Ames, Iowa 5001 1
Many different methods for determining cyanide have been developed over the years. Most colorimetric methods are based on the Konig reaction of cyanogen halide with an aromatic amine and pyridine and are sensitive from about 0.01 to 2 ppm cyanide (I). Two common titrimetric methods are Liebig's and Volhard's methods (I). Both use the formation of a very stable silver cyanide complex as the basis for their methods and are used above 20 ppm cyanide. The sample must also be free from interferences. The chromatographic method described in this article was developed by modifying a pneumatoamperometric method developed by Beran and Bruckenstein (2) that used the reaction of cyanide with iodine listed below.
I2
+ HCN e H++ I- + ICN
K,, = 0.73
They describe the equilibrium as being well-known, pH dependent, and quantitative from a pH of 2 to 7. Since the final concentration of iodide is proportional to the initial concentration of cyanide, the new chromatographic method analyzes iodide using a single-column ion chromatograph and a conductivity meter to determine cyanide concentrations. Excess iodine in the solution is removed by adsorption onto a short precolumn containing unfunctionalized XAD-4 resin (Rohm and Haas). After a series of injections the precolumn is removed from the chromatograph and the iodine is desorbed by flushing with a nitric acid-acetone mixture. The precolumn is then placed back in the injection line of the chromatograph. EXPERIMENTAL SECTION Chemicals. All chemicals were of reagent grade quality from Fisher Chemical Co. and were used as received. A 0,100 M iodine solution in 95% ethanol was prepared. The container used to store the iodine solution was wrapped in aluminum foil to slow the decomposition by W and visible radiation. Cyanide samples for calibration and for testing the procedure under various conditions were prepared from a stock solution of potassium cyanide that was prepared fresh weekly and contained approximately 100 ppm of cyanide. The stock solution was made up in approximately 0.001 M sodium hydroxide. Cyanide samples of varying concentration were prepared by dilution of the stock solution just before analysis. 'Present address: Exxon Research and Engineering Co., P.O. Box
121, Linden, NJ 07036.
0003-2700/82/0354-0630$01.25/0
A 0.05 M acetate buffer solution with a pH of 4.75 was made with glacial acetic acid and 0.1 M sodium hydroxide. A 0.05 M nitric acid solution was prepared in acetone and the 100 ppm iodide stock solution used in these experiments was made from potassium iodide. Interferences were made from stock solutions of nitrate salts for the cations and sodium or potassium salts for the anions. Apparatus. A single-column ion chromatograph described by Gjerde et. al. (3) was used with a Model 213A Wescan conductivity detector and cell. The chromatograph was modified by adding a glass precolumn as described below. The precolumn was fitted with a female leur adapter on one end so that samples could be injected by hand directly into the precolumn using a syringe. The other end of the precolumn was connected to the injector valve, containing the sample loop, with a short length of Teflon tubing. This arrangement allows the operator to load the sample loop and remove the excess iodine from the sample at the same time. The precolumn should be glass so that the operator can tell when the precolumn needs to be flushed. When the resin in the precolumn at the end closest to the injector valve starts to turn yellow, the iodine needs to be flushed from the precolumn. The precolumn, 75 X 2 mm i.d., was packed with unfunctionalized Rohm and Haas XAD-4 resin, 100-150 mesh. The analytical column, 500 X 2 mm i.d., was packed with anion exchange resin prepared from 325 to 400 mesh Rohm and Haas XAD-1 (3). The resin had an anion exchange capacity of 0.025 mequiv/g. Other specific conditions were as follows: eluent pumping rate, 1.5 mL/min; sample loop size, 100 pL; detector full scale, 1 pmho; detector output, 10 mV; recorder full scale, 10 mV. Procedure. If necessary, dilute the sample so that the expected cyanide concentration will be in the range of 0.5-10 ppm. Pipet accurately 40 mL of sample into a 50-mL volumetric flask and add 1 mL of 5 x low2M acetate buffer (pH 4.75). Mix well and then add 0.3 mL of 0.1 M iodine in 95% ethanol. A yellow color should remain after mixing, indicating that an excess of iodine has been added. The excess iodine is necessary for the reaction of cyanide and iodine to be quantitative. Dilute the solution to 50 mL, mix, and allow the solution to sit at least 5 min. Then slowly inject a 2-3 mL aliquot of the sample into the ion chromatograph through the XAD-4 precolumn. The slow injection time, approximately 5-10 s, through the precolumn allows the resin more time to adsorb the excess iodine. Separate the iodide ion, resulting from reaction of cyanide with iodine, chromatoM sodium or potassium graphically by elution with 2 X phthalate, pH 6.25. The retention time for iodide on the column 0 1982 American Chemical Soclety
ANALYTICAL CHEMISTRY, VOL. 54, NO. 4, APRIL 1982
831
~~
Table 11. Cation Interference Study
Table I. pH Stu'dy pH
4.10 5.18 6.26 a
Ineight,'l cm 8.45 8.35 8.20
pH 7.88 8.90
height,a cm 11.60 14.30
cation NiZ+ Ni2+
Done with 10 ppm cyanide.
MgZ Mg2
+
ppm
2
4
CN-
+
8
6
1
0
1
2
co2+
co2+ Hg2+ Hg2+ CdZ CdZ+ Znz+ Zn2+ cu2+ cu2+ +
approx molar ratio (CN-/Mnt)
% %
recoverya 98.5 95.2 98.2 96.1 15.8
1O:l 1:l 1O:l 1:l 1O:l
1:l
0
1O:l 1:1 1O:l 1:l
14.0
re1 dev 2.1 1.4 2.2 1.4
0
100.2 96.8 96.4 94.2 103.6 124.2
1O:l 1:l 1O:l 1:l
4.9 2.7 3.1
2.5 4.0 4.2
a Ni", MgZ+,HgZ+, Co2+,and Cuz+were measured by peak height, the other cations were measured by area because of increased broadening of peaks at higher interference concentration.
Table 111. Anion Interference Study
anions c1c1-
so,2s0,z-
0
I0
20
30 ppm
40
50
60
1-
Figure 1. Calibration plots for cyanide based on iodide qnd on cyanide standards.
used was approximately 6 min. Measure either the height or the area of the iodide peak and calculate the amount of cyanide from a calibration plot. Prepare the calibration plot by running several cyanide standards, freshly prepared, by the same procedure used for samples and plotting peak height (or area) against the micrograms of cyanide in the standards. After four to six sample injections, the iodine must be desorbed from the XAD-4 precolumn. Do this by flushing the precolumn with 2-3 mL of 0.05 M nitric acid in acetone, followed by 3-6 mL of water. This flushing is facilitated by removing the precolumn from the system to prevent any accidental contamination. RESULTS AND DISCUSSION pH. The sample pH Rhould be adjusted and buffered before adding iodine because side reactions involving iodine are likely at highly alkaline pH values. The results of pH studies, summarized in Table I, show that the iodide peak height is constant between pH 4.0 and 6.5 but that the peak height increases rapidly as the pH increases above pH 6.5. The pH range 4.0-6.5 thus appears to be satisfactory for the cyanide determination. Because some actual samples are highly alkaline, some neutralization in addition to that obtained by adding the buffer may be required to place the sample in the desired pH range for analysis. The peak heights listed in Table I for 10 ppm C N - (-8.3 cm) do not correspond with the calibration curve a t Figure 1 (-15.5 cm) because different recorder attenuations were used for the two experiments. Removal of Unreacted Iodine. Earlier studies by Jarosz (4) showed that methylene chloride could be used to extract unreacted iodine before chromatographic separation of the iodide. However, the use of a small resin precolumn is con-
Cr0,2CrO,*NO; NO; ClO, ClO,
so,2-
so32szo,2szo32SCNSCN-
PPm ratio (CN-/An-) % recoverya 1:10 1:l 1:lO 1:l 1:lO
1:l 1:lO 1:l 1:lO 1:l 1:lO 1:l 1:lO 1:l 1:lO
%
re1 dev
101.8 99.6 95.8 99.1 96.1 101.4 98.9 99.2 102.5 100.9 >400
2.9 0.8 2.8 3.3 2.9 4.1 1.5 1.9 2.0 0.5
101.0
1.7
> 500 >125 >700
>250
1:l
a Cl-, NO;, and C10, were measured by peak height, the other anions were measured by area because of increased broadening of peaks at higher interference concentration.
venient and is more time-efficient. A preliminary separation of cyanide by distillation of hydrogen cyanide has been used (5) to avoid interferences in the determination of cyanide. Generally a sodium hydroxide solution is used to trap the distilled HCN. We tried an iodine trap that initially was unsuccessful but still appears to be a sound idea. Calibration. I t should be possible to base the cyanide calculation on the stoichiometric amount of iodide produced by the reaction
-
I2 + CN-
ICN + I-
Figure 1 shows two calibration plots, one based on iodide standards and the other based on cyanide standards. Both of the plots are essentially linear and are similar enough to indicate that the iodine-cyanide reaction is approximately stoichiometric. However, the difference in slopes is large enough that quantitative results should be based on cyanide standards. That the difference in slopes may possibly be due to the formation of 1,- has not been verified. The calibration plot for cyanide standards is linear for points between 0.4 and 10 ppm cyanide with a correlation coefficient (for linear regression analysis) of 0.998.
832
Anal. Chem.
1982,5 4 , 832-833
complexes with cyanide and thus interfere with the new procedure for cyanide. The results in Table I11 indicate no interference from common anions such as chloride, nitrate, sulfate, perchlorate, chromate, and small amounts of sulfite. Thiosulfate and thiocyanate react readily with iodine and cause serious interference. An actual chromatogram of a sample containing both cyanide and chromate is shown in Figure 2. The initial positive peak is called the “pseudopeak” and is always encountered in single-column ion chromatography because of differences in the conductances of the sample and the eluent. The negative peak immediately following the pseudopeak is caused by the buffer, sodium acetate. The next positive peak is iodide; the final positive peak is chromate. The sensitivity of the method is improved by reacting cyanide to produce iodide, because the chromatographic detection of iodide is several times more sensitive than it would be for the cyanide ion directly. LITERATURE CITED
and 20
(1) Chambers, W. E.; et al. “Treatise on Analytical Chemistry Part 11: Analytical Chemlstry of Inorganic and Organic Compounds”; Kolthoff. I. M., Elving, P. J., Eds.; Wiley: New York, 1978; Vol. 10. (2) Beran, P.; Bruckensteln, S. Anal. Chern. 1880, 52. 1183-11813. (3) Gjerde, D. T.; Fritz, J. S.; Schmuckler, G. J . Chrornatogr. 1878, 786, 509-519 (4) Jarosz, M.; Fritz, J. S., unpublished work. (5) “Handbook for Analytical Quallty Control in Water and Wastewater Laboratorles”; Analytlcal Quallty Control Laboratory, National Environmental Research Center: Clnclnnatl, OH, 1972.
Interference Study. The effects of various metal ions and foreign anions on the results for cyanide were studied. The results for metal ions reported in Table I1 show essentially no interference from nickel(II), magnesium(II), cadmium(II), or zinc(I1). Mercury(I1) and cobalt(I1) form more gtable
RECEIVED for review November 16,1981. Accepted January 18,1982. Operated for the US. Department of Energy by Iowa State University under Contract No. W-7405-Eng-82. This research was supported by the Director of Energy Research, Office of Basic Energy Sciences.
I
I
I
I
0
3
6
9
Minutes
Figure 2. Chromatogram for a sample containing 4 ppm
CN-
ppm CrOd2-.
Integration of a Gas Flow Proportional Counter into a Commercial Gas Chromatagraph/Mass Spectrometer/Data System D. L. Doerfler, G. T. Emmons, and I . M. Campbell” Department of Biological Sciences, University of Pittsburgh, Parran Hall, 130 DeSoto Street, Pittsburgh, Pennsylvania
Coupling a gas flow proportional conter (PC) to a commercial gas chromatograph/mass spectrometer (GC/MS) is a straightforward undertaking (1-3). Development of software to collect and process data from the radiogas chromatograph/mass spectrometer so formed is also straightforward (4,5).Integration of a data channel derived from a PC into a data system that was provided with a commercial GC/MS respresents a more formidable challenge. This report describes how this challenge can be met. EXPERIMENTAL SECTION Hardware. A Packard Model 894 gas flow PC and an Hew1etbPackard (HP) Model 5985B GC/quadrupole-MS/data system were used. The former converts a portion of the effluent of a GC column to COzand H20 (copper oxide furnace), reduces the H20 to H2 (iron chip furnace, used if tritium is involved) or removes it from the gas flow (magnesium perchlorate or sodium sulfate), and countv the processed gases in a proportional tube following addition of a quench gas (e.g., propane). The computer of the HP 5985B data system is an HP 21MX-E, configured with 32k words of core, all standard features plus a fast Fortran processor and four unused 1 / 0 slots. Plumbing the PC into the GC/MS. The gas flow in the GC module of the GC/MS was broken at the effluent end of the 0003-2700/82/0354-0832$0 1.25/O
1526 1
column and a fiied ratio splitter was inserted (Packard Instrument Co., part 5060723,3:1). The splitter resides in an aluminum block of the same dimensions as the flame detector housing and is positioned where that housing normally resides. The temperature sensing and heating circuitry provided for the flame detector is used to heat the splitter. The majority of the gas flow (75%) is fed to the copper oxide furnace of the PC by means of a heated 1/16 in. nickel line while the minority (25%) is led to the molecule separator of the MS. The drying tube (we have worked so far exclusively with 14C)and PC counting tube/shield assembly sits atop the GC. The PC electronics have been repackaged to fit on the GC/MS main frame. Construction of a Pulse Counter/Computer Interface. For each nuclear disintegration that the Model 894 PC detects, a 5 V/3 ps pulse can be picked up at its scaler terminal. In our interface this pulse is narrowed in width to 600 ns (with a 74C221 CMOS IC) and is fed to a 16-bit synchronous pulse counter (four 74C161 CMOS IC’s) that increments on a positive edge. The maximum permissible count rate of the pulse counter is in excess of 1MHz, a figure that comfortably encompasses the maximum count rate of the PC (20 kHz). The running sum of all counts collected in an RGC/MS run (maximum range, 0-177 7778,065 53510)is stored in the interface; the current value is displayed on a 6-digit LED display on the interface cabinet. The interface communicates with the HP MX21-E by means of 16-bit parallel 0 1982 American Chemical Society
