Determination of aliphatic primary and secondary ... - ACS Publications

Anal. Cham. 1987, 59,480-484. Determination of Aliphatic Primary and Secondary Amines and. Polyamines in Air by High-Performance Liquid Chromatography...
0 downloads 0 Views 566KB Size
Anal. Chem. 7907, 59, 480-484

400

Determination of Aliphatic Primary and Secondary Amines and Polyamines in Air by High-Performance Liquid Chromatography Patrice Simon* a n d Clotilde Lemacon Institut National de Recherche et de S&curit&,Dgpartement Environnement Chimique, Service Chirnie toxicologique, Avenue de Bourgogne, B.P. No. 27, 54501 Vandoeuure, France

An easy and rapid method lnvolvlng derlvatlzatlon was developed for low concentratlon analysls of aliphatic amlnes (prknary amines, secondary amines, and polyamines) In alr. Sampling was effected wlth sllka gel tubes. The derorptlon of the collected amlnes was performed simultaneously wlth the derivatizatlon. The resulting m -toluoyi derlvatlves were dlrectly analyzed by high-pertonnance llqukl chromatography (HPLC) wlthout an extractlon procedure. Analyses were performed by use of a reversed-phase chromatography system wlth ultraviolet detectlon.

Aliphatic amines and polyamines are industrial chemicals with a wide variety of applications. They are used as raw materials or at an intermediate stage in the production of other chemicals, pharmaceuticals, polymers, pesticides, dyestuffs, and corrosion inhibitors ( I ) . Amines are sensitizers of and irritants to the skin, mucous membrane, and respiratory tract; some are precursors of N-nitrosamines, which are carcinogenic substances ( 2 ) . Consequently, it would be useful to determine not only trace amounts of nitrosoamines but also the unconverted amines in the atmospheric and human environment (3). Previous analytical methods developed for these substances have involved colorimetry ( 4 ) and polarography ( 5 ) . Gas chromatography (6-10) and particularly high-performance liquid chromatography (10-14) have more recently been used for this purpose and have several advantages over other analytical methods, including high specificity and sensitivity. However, few reports describe an easy method of sampling amines in air and of analyzing them by HPLC (15-19). The procedure used in this work was applied to 16 amines a t various air concentration levels. Simultaneous desorption and derivatization of amines and a minimum manipulation of samples as a result of the suppression of the extraction procedure-in spite of the derivatization step-account for the singularity of this method. A gradient on two different reverse-phase octadecylsilane columns, with acetonitrilewater as a mobile phase, was needed for optimal separation. A CN polar bonded phase and silica were successfully used to separate rn-toluamides by normal phase chromatography. MATERIALS A N D METHODS Reagents. Chlorosilanes and solvents were purchased from Merck (Darmstadt, West Germany) and were of reagent grade and chromatographic grade, respectively. Amines and m-toluoyl chloride were obtained from Aldrich. Silica Spherosil XOA 600, 5 pm (Rh6ne-Poulenc),was used in the preparation of the bonded phases. Silica Nucleosil CIB,5 pm, and a column (25 cm X 4.5 mm id.) packed with Nucleosil CN 5 p m (Macherey,Nagel, FRG), were purchased from Interchim (France). Water for the HPLC runs was purified in a Millipore Milli R/Q water purifier. mToluoyl derivatives were prepared according t o the procedure described by Chen and Farquharson (12). HPLC Support Preparation. The reaction was performed at room temperature according to the procedure described by

Evans et al. (20). Silica gel (2.5 g) was reacted with octadecyltrichlorosilane (1.41 g, 3.65 X mol) in carbon tetrachloride (50 mL) for 12 h. The bonded phase was washed consecutively with dry CC4to remove unreacted silane, dry methanol, and dry methylene chloride. After being suction dried, the prepared material was “end capped” with an excess of trimethylchlorosilane in the conditions previously described. Column Preparation. The 15 cm X 4.5 mm i.d. stainless steel columns were prepared by the balanced density slurry packing method using toluene, 2-propanol, and 95% ethanol (l/l/l)(21). Pressure was set at approximately 40 000 kPa. Preparation of Derivatives. m-Toluoyl derivatives were prepared according to the following method: 0.75 g (0.5 X mol) of m-toluoyl chloride was diluted in 30 mL of acetonitrile and an excess of amine (0.6 X lo-* mol), 5 mL of 20% NaOH was then added, and the heterogeneous mixture was shaken for 10 min. The acetonitrile was evaporated and the samples were then extracted with 20 mL of methylene chloride and washed with HCl 0.1 N and distilled water. After drying on MgSO,, the samples were filtered, the methylene chloride was evaporated, and the derivatives isolated. Preparation of Sampling Tube. The sampling tubes, made of Pyrex glass (5 cm length, 6 mm i.d.), were filled with about 0.350 g of silica gel (Merck Kieselgel60,30-70 mesh) held between two glass wool plugs. Standards. Amines were dissolved in acetonitrile and further diluted to form a series of four dilutions (Table I). Ten-microliter aliquots of each amine solution were placed on the glass wool plug, and air with a known flow rate (1 L/min) and humidity content (35% and 63%) was pumped through the silica gel tube for a period of 1 h. Desorption and Derivatization of Samples. The adsorbent was transferred into glass vials with PFTE-lined screwcaps, and 5 mL of acetonitrile with m-toluoyl chloride (4 X 10” M) and 0.2 mL of NaOH or KOH 5 M were added, after which the mixture was mechanically shaken for 10 min. A 0.2-mL portion of concentrated NHIOH was then added to destroy the excess of reagent and the sample was again shaken for 10 min. Further extraction was not necessary. Desorption and Derivatization of Diisopropylamine and Diethylenetriamine. Either the adsorbed material on the silica gel or 1O-wL aliquots of standard solution were placed into glass vials with PFTE-lined screwcaps. A 5 mL portion of KOH 5 M was added with rn-toluoyl chloride (30.1 mg, 2 X lo4 mol) and the mixture was submitted to a vigorous manual shaking for 2 min. Then 5 mL of acetonitrile and 0.2 mL of concentrated NHIOH were added to destroy the excess reagent and the heterogeneous mixture was shaken mechanically for 10 min. On completion of this procedure, the sample was ready for analysis. Apparatus. HPLC analysis was carried out with a Waters Associates Model 590 chromatographypump in conjunction with a IO-pL loop sample injector (Rheodyne Model 7125) and a LCUV variable wavelength adsorbance detector (Pye Unicam PU 4020). An integrator (Spectra Physics SP 4100) or a recorder (Sefram) was used for the quantification of the results. The identity of the m-toluamide derivatives was checked by IR spectrophotometry and mass spectrometry (Nermag/Sidar) with and without chemical ionization (NH,). Chromatographic Analysis. Samples free of particulate matter were injected directly into the chromatographic system. The separating columns ( L = 15 cm X 4.5 mm i.d.) used were packed with commercial and homemade ODS silica phases which

0003-2700/87/0359-0480$01.50/00 1987 American Chemical Society

ANALYTICAL CHEMISTRY, VOL. 59, NO. 3, FEBRUARY 1, 1987

481

Table I. Recovery of the Amines from the Silica Gel % recovery'

% recove@

amine

std solutions, g/L

avb f SDc

122.13 61.06 28.74 5.75 177.20 70.88 47.60 9.55 179.65 89.83 44.91 8.98 180.70 90.35 14.08 50.33 25.20 12.50 2.52 22.14 11.07 5.54 55.52 29.80 5.60

98 f 2 96 f 1 94 f 3 93 f 3 97 f 2 100 f 1 95 f 1 94 f 4 99 f 2 96 f 1 97 f 2 98 f 2 92 f 1 91 f 2 95 f 1 91 f 1 91 f 2 91 f 1 96 f 1 95 f 1 94 f 0 91 f 1 95 f 4 97 f 1 95 f 1 98 f 4

202.20 101.08 50.40 10.10 148.00 74.00 37.00 7.40

96 f 2 93 f 2 92 f 2 95 f 1 95 f 6 96 f 3 98 f 8 97 f 3

methylamine

dimethylamine

ethylamine

diethylamine n-propylamine

dipropylamine

1.11

isopropylamine

diisopropylamine

n-butylamine

amine tert-butylamine

dibutylamine

allylamine

diallylamine cyclohexylamine

ethylenediamine

diethylenetriamine

morpholine

3-(dimethylamino)propylamine

std solutions, g/L

avb f SDc

139.20 69.60 34.80 6.96 153.40 76.70 38.35 7.67 4.57 2.29 1.14 9.23 9.88 4.70 2.47 398.82 208.08 100.04 26.00 251.72 125.86 71.92 14.40 38.20 19.10 9.55 1.91 350.00 175.00 35.00

98 f 2 96 f 1 99 f 3 96 f 1 100 f 1 100 f 1 100 dz 1 100f4 97 f 2 94 f 2 96 f 3 93 3 92 f 0 94 f 1 97 f 1 104 f 0 100 f 1 94 f 3 94 f 3 91 f 2 97 f 3 100 f 4 85 f 0 95 f 3 91 f 0 94 f 2 95 f 3 92 f 4 97 f 2

81.20

97 f 3

100 f 1

*

Recovery for 10 pL of amine standard solution placed in the silica gel tube. *Average of four runs. Standard deviation. Table 11. Collection and Recovery of Amines Using the Dynamic U-Tube System compounds

amt added, mg

amt found, mg

recovery, %

amt added, mg

amt found, mg

methylamine dimethylamine ethylamine diethylamine n-propylamine dipropylamine isopropylamine diisopropylamine n-butylamine tert-butylamine allylamine diallylamine cyclohexylamine morpholine

0.108 0.180 0.180 0.283 0.049 0.021 0.110 0.202 0.148 0.153 0.004 0.0094 0.382 0.700

0.101 0.175 0.169 0.275 0.050 0.019 0.105 0.194 0.139 0.148 0.004 0.0089 0.355 0.625

94 97 94 97 102 90 96 96 94 97

1.08 1.80 1.80 2.83 0.49 0.21

1.04 1.70 1.78 2.72 0.49 0.20 1.07 0.41 1.45 1.45 0.042 0.088 3.61 6.74

had been previously prepared. The eluent, consisting of acetonitrile and water, was pumped (1 mL/min) either by isocratic means or by using gradients, depending on whether the amount or the quality of the amines was to be determined. The peaks were usually monitored at 235-255 nm. The optimized gradient, composed of three linear steps, is presented in Figure 1. Qualitatively the retention times were compared with those of known standards, and quantitatively the method of peak height measurement with an external standard was used. For normal-phase chromatography, a silica column and a Nucleosil CN bonded silica column (5 pm; 25 cm X 4.6 mm i.d.) were used with isooctane 90%, methylene chloride 9%, and 2-propanol 1%as an eluent.

100

95 93 90

1.11

0.41 1.48 1.53 0.045 0.094 3.82 7.00

recovery, % 96 93 99 96

100

94 96

100

98 95 93 94 95 96

RESULTS AND DISCUSSION The poor ultraviolet absorptivity of aliphatic amines necessitates chemical derivatization t o detect small amounts. m-toluoyl chloride has been proved to be a good and practical derivatizing reagent for amines (IO,11). This chromophor reagent was selected in preference to a fluorophor reagent, because the W absorptivity of n-toluoyl derivatives is more than sufficient to detect any trace of aliphatic amines in air at levels of less than 1 pg/m3. The following equation describes the reaction of m-toluoyl chloride with ammonia and aliphatic amines:

482

ANALYTICAL CHEMISTRY, VOL. 59, NO. 3, FEBRUARY 1, 1987 6

A

The amine adsorption-desorption by the static method was evaluated for 16 amines at various amine levels (Table I). The amounts of the different amines corresponded to a concentration range of 0.24-150 mg/m3 in a 50-L air sample. The amines recovered from silica are listed in Table I. Humidity was not observed to have an effect on collection efficiency, and storing the spiked tubes in the dark at 4 "C for 1 month did not impair recovery (16). The efficiency of sampling and the adopted analytical procedure were investigated by means of the dynamic generation of atmospheres. These atmospheres were generated in a dynamic U-tube system (22)as well as in 200-L (stainless steel) inhalation chambers designed for sustaining dynamic and adjustable flows of air (10-12 m3/h). Amines vapors were produced by bubbling an additional air flow through a vial containing the liquid compound. The vaporized amines were carried into a mixing device for further dilution with air to the final concentration before entering the exposure chamber. All the detailed information of the shape, the dimensions, and the operating of chambers has been given previously (23). The results of collection efficiency obtained from the dynamic U-tube system are summarized in Table 11. For a vapor concentration equal to twice the ACGIH standard level, no breakthrough of any of the compounds was observed when 120 L of air of 100% relative humidity was sampled and the capacity of the silica gel (0.350 g) exceeded the amount of amine collected in this concentration. These results are in concordance with the observations related by Taylor (24,25) and Bouyoucos et al. (16). The chamber concentrations were monitored continuously. A sample loop was swept through the cell atmosphere and samples of chamber atmosphere were analyzed by gas-liquid chromatography; the variations were no more than 5 % . Performance of sampling with silica gel and analysis by HPLC was examined for different amines and in more detail for diisopropylamine. From determination performed on 40 specimens over a 4-h sampling period, variations of 6% and 7% were recorded for 2-L and 5-L samples, respectively, indicating that the method is reliable. During the assessment of the applicability of the method to various aliphatic amines, it was observed that special conditions were required to obtain complete derivatization of diisopropylamine and of diethylenetriamine; this difference in behavior was probably due to steric hindrance, as previously observed by Hamano et al. (6) in the course of the reaction of diisopropylamine with benzenesulfonyl chloride. rn-Toluamide derivatives were stable with time, as were standard solutions in acetonitrile; it is not, however, advised to use these solutions after 2 months or later. Figure 1 shows a good separation of m-toluoyl derivatives of amines, using two different reverse-phase octadecylsilane columns with the same gradient elution conditions. With Nucleosil CIS (commercial RP) as a support, neither diallylamine and cyclohexylamine nor ethylamine and dimethylamine were separated (Figure 1). The resolution was better with Spherosil XOA 600 C18 (homemade RP) with the exception of propylamine and isopropylamine, which were poorly

I 0

P

10

20 Rstintion time

30

40

min

6 9 B

L

0

w

10

20 Retenlion t i m e

30

40

min

Figure 1. Separation of m-toluoyi derivatives of amines on (A) the Nucleosii C,, column and (B)the Spherosil C,, column (wavelength 240 nm): (1) ammonia, (2) methylamine, (3) morpholine, (4) ethyiamine, (5) dimethylamine, (6) allylamine, (7) Isopropylamine, (8) propylamine, (9) ethylenediamine, (10) diethylamine, (1 1) n-butylamine, (12) tertbutylamine, (13) cyclohexylamine,(14) dialiylamine, (15) diisopropylamine, (16) dipropyiamine.

separated. Figure 2 shows a partial separation of m-toluamides with a silica phase and CN bonded phase. The sample was eluted in the isocratic mode. In cases where the identity of a peak was ambiguous, the use of two different R P columns or two chromatography techniques (normal LC and RPLC) resulted in an unequivocal

ANALYTICAL CHEMISTRY, VOL. 59, NO. 3, FEBRUARY 1, 1987

483

I

+ 2 + 3 9 + +

A B

Rstenlion

time

min

Flgwe 3. Revers&phasellquid chromatograms of amine vapors from vuicanlratlon fumes (wavelength 235 nm): (A) sample volume 60 L,

I

D

IO

0

20

40

30

60/40

water/

where the presence of amines is suspected. The derivatization procedure may also be used successfully to monitor amines in bactericides. This method may also be of value for determining trace amines in industrial samples such as oils, in biological samples such as urines, and in cosmetics such as shampoos.

5

8 9

I+

dlmethylamlne, (2) isopropylamine, mobile phase

acetonitrile at 1 mumin; (B) same sample, (3) cyclohexylamine, mobile phase 50/50 water/acetonltrile.

+ 4

(1)

13

Registry No. Methylamine, 7489-5;dimethylamine, 124-40-3; ethylamine, 75-04-7;diethylamine, 109-89-7;n-propylamine, 107-10-8; dipropylamine, 142-84-7; isopropylamine, 75-31-0;ditert-butylamine, isopropylamine,108-189;n-butylamine, 109-73-9; 75-64-9;dibutylamine, 111-92-2; allylamine, 107-11-9; diallylamine, 124-02-7; cyclohexylamine, 108-91-8; ethylenediamine, 107-15-3; diethylenetriamine, 111-40-0;morpholine, 110-91-8;34dimethylamino)propylamine, 109-55-7.

LITERATURE CITED 0

IO

30

20 Rilintion

timi ,

40

50

rnin

Figure 2. Separation of m-toluoyl derivatives of amines on (A) the CN polar bonded phase column and (B) the silica column (wavelength 240

nm): (1) tert-butylamine, (2) diallylamine, (3) dlisopropylamlne, (4) dipropylamlne, (5) diethylamine, (6) cyclohexylamine, (7) hexylamine, (8) isobutylamlne, (9) isopropylalmlne, (10) n-butylamlne, (11) dimethylamine, (12) morphdine, (13) n-propylamlne, (14) allylamine, (15) ethylamine, (16) methylamine. identification. However, for routine and rapid measurement of amines, the use of a reversed phase is suitable. The derivative of 3-(dimethy1amino)propylamine was not eluted under the experimental conditions, and the addition of N(Et), (1%) to the eluent was necessary to desorb this compound from the column. Aliphatic amines were also determined as they occurred in vulcanization fumes; Figure 3 shows typical liquid chromatograms of m-toluoyl derivatives of amines from this source. Cyclohexylamine, dimethylamine, and isopropylamine were detected a t a level as low as 0.1 mg/m3 in air samples (60 L pumped). The identity of the amines was checked by IR spectrophotometry and mass spectrometry on postcolumn trapping of each component. The proposed method, which is both easy and sensitive, may be useful in routine analysis of large numbers of air samples

Kirk-Othmer €ncyc/oped& of Chemical Technologv, 3rd ed.;Wiley-Intersclence: New York and London, 1978; Vol. 2. pp 272-283. paw’s hdUStnkrl end TOXlcology, 3rd ed.;Wlley-Iflt6fsCbnCS: New York and London, 1981; Vol. 28, pp 3135-3173. Blome, H.; Hennig, M.; Augustln, S. Staub-Relnhan. Luft 1084, 4 4 , 27-32. Stanley, E. L.; Baum, H.;Gave J. L. Anal. Chem. 1051, 23, 1779. English, F. L. Anel. Chem. 1051, 23, 344. Hamano, T.; Hasegawa, A.; Tanaka, K.; Matzuki, Y. J. Chromatogr. 1070, 779.346-350. Wood, 0. O., Nlckols, J. W. Energy Res. Abstr. 1979, 4(1). Abstr. No. 1318. Hamano, T.; Mlsuhashi, Y.; Matsuki, Y. J. Chromatogr. 1980, 790, 462-465. Kuwata, K.; Yamazakl, Y.; Ueborl, M. Anal. Chem. 1980, 52, 1980-1982. Kuwata. K.; Aklyama, E.; Yamazaki, Y.; Yamasaki, H.; Kuge, Y. Anal. Chem. 1083, 55, 2199-2201. Wellons, S. L.; Carey, M. A. J. Chromatogr. 1978, 754, 219-255. Chen, E. C. M.; Farquharson, R. A. J. Chromatogr. 1979. 778, 358-363. Lln, J. K.; Lai, C. C. Anal. Chem. 1980, 52, 630-635. Bjorkqvist, B. J. Chromatogr. 1981, 204, 109-114. Lepage, J. N.; Rocha, E. M. Anal. Chem. 1083, 55, 1360-1364. S. A.; Melcher, R. G.; Am. Ind. WQ. BOUYOUCOS, . _ Assoc. J . 1083, 44(2), 119-122. Andersson, K.; Hallgren, C.; Lenin. J. 0.; Nllsson, C. A. J. Chromatour. 1084. 3 1 2 . 482-488. N&hikawa,’Y.; Kuwata, K. Anal. Chem. 1084, 58, 1790-1793. Andersson, K.; Hallgren, C.; Levin, J. 0.; Nllsson, C. A. Am. Ind. W g . ASSOC. J . 1085, 46(4), 225-229. Evans, M. B.; Dab, A. D.; Little, C. J. Chromatographla 1080. 73(1), 5-10. Keller, H. P.; Emi, F.; Llnder, H. R.;Frel, R. W. Anal. Chem. 1077, 49, 1958-1963.

484

Anal. Chem. 1987, 59, 484-486

(22)Severs. L. W.; Mekher, R. G.; Kocsis, M. J. Am. Ind. Hyg. Assoc. J . 1978, 39(4),321-326. (23)Gradiski, D.; Bonnet, P.; RaouA, G.; Magadur, J. L. Arch. Mal. Prof. Med. Trav. Secur. SOC.1978,39, 249-257. (24) Taylor, D. G.;Kupel, R. E.; Bryant, J. M. Documentstion of the NIOSH vaihtion Tests; NO. s 139,s 142,s 144,s 147,s 150; US. ~ 0 ernment Printing Office: Washington, DC, 1977.

(25) Taylor. D. 0. NIOW Menuel of Analytcal Methods; No., 221; US. Government Printing Office: Washington, DC, 1977.

~

-RECEIVED for review June

25,1986. AccepM October 7,1986. This is Publication No. 1285 from the INRS.

Comparison of Anion-Exchange Methods for Preconcentration of Trace Aluminum Corrado Sarzanini, Edoardo Mentasti,* Valerio Porta, and Maria Carla Gennaro Department of Analytical Chemistry, University of Torino, Via P. Giuria, 5, 10125 Torino, Italy

Alumhum( I I I ) traces have been preconcentrated through formam of anfonlc compkxee whkh are r e t a w by the anlon exchanger. The ligand was PyrocatcrcholVldet (PV) whose structure contains a sutronato group (not Involved In AI( I1 I ) coordlnatlon) which Is rerpondMe for the metal uptake by the resin. Two d#tcHent prooedues were hves@atd and compared. One Involves the fonnatlon of ACPV complex(es) followed by elutlon through a column contahdng the anlon exchanger. The second procecbe lnvotves the kadlng of the PV llgand on the redn and subsequent ehdlon of the sample. The results obtalned wHh the two dlfferent procedures have been compared wHh reference to percent recovery yield as a functh of pH and effect of foreign compommts (eurfactants, Inorgsnk: SelQ such as phosphates and chkrkleq organlc compombsuch as hunlc ack!8, and NTA). The fint procedure was superb and hae been UNlZedfor the analysis of synthetic samples contalnlng AI at the parts-perbllllon level and for drlnklng water.

Ion exchangers have been used for the preconcentration of trace elements, speciation studies in complex mixtures, removal of interfering components from analyzed solutions, and separations of ions (1-4). Resins, both cationic or anionic, ion-exchange papers (5) or membranes (6) have been shown to be suitable for collecting and separating trace quantities of ions (7,8). Aromatic complexing agents containing sulfonic acid groups are particularly useful for the separation of metal ions on anion-exchange resins (9-14). Two methods are compared here employing Pyrocatechol Violet in preconcentration techniques for removal of aluminum traces. This chelating agent gives highly stable complexes with aluminum and shows, owing to its very large molecular size, an increase of affinity with a macroporous anion exchanger. The use of Pyrocatechol Violet for metal-matrix separations and preconcentration is based mainly on two methods (i) complex formation coupled with an anion-exchange resin and (ii) column chelation procedure. In the first case the solutions of the metal are added of the chelating agent and fluxed through the exchanger at the proper pH. In the latter the samples are eluted through the macroporous anion-exchange resin previously loaded with an appropriate amount of ligand. Synthetic and natural samples have been investigated in the presence of interfering agents. The final determination of the preconcentrated element was accomplished by dc argon

plasma emission spectrometry.

EXPERIMENTAL SECTION Apparatus and Materials. All laboratory glassware and polyethylene and polypropylene equipment were cleaned in 6 M nitric acid and repeatedly rinsed with high-purity water (HPW). Thermostated (water-jacketed)borosilicate glaas columns 8 mm i.d. and 30 cm high) were employed and a slurry of 0.5-1.0 g of resin was supported. A rotary vacuum pump with a bypass flowmeter ensured a 1-5 mL/min constant flow of samples through the column. A peristaltic pump equipped with a variable-speed control ensured reduced flows, generally within 0.5-1.0 mL/min. Metal concentration measurements were performed by dc argon plasma emission spectrometry (Spectraspan IV, SMI, Andover, MA) at X 396.15 nm where the detection limit for Al is about 5-10 ng/mL. Two-point calibration was used. The pH measurements were performed with a Orion 811 pH meter equipped with a combined glass calomel electrode. Adjustable Eppendorf pipets were used to prepare the solutions. Analytical grade Bio Rad AG MP 1 macroporous anion-exchange resin, 1W200 mesh, was used in chloride form. acid, PyrocatecholViolet (3,3f,4f-trihydro~chsone-2"-sulfonic PV, Merck) was an analytical grade reagent. A standard aluminum (Merck) stock solution (loo0mg/L metal concentration) was diluted to the desired concentrations. Other reagents were of analytical grade and all solutions were prepared by using deionized water further purified by a Milli-Q water purification system (Millipore, Bedford, Ma). Procedures. Precomplexation Anion Exchange (PAEJ. The borosilicate column filled with 0.5-1.0 g of AG MP 1 resin was rinsed and preconditionedwith HPW to the proper pH. Solutions (100 mL and lo00 mL) each containing 10.0 pg of Al(II1) were added to 2.0 and 7.0mL of 0.05 M Pyrocatechol Violet, respectively, and were brought to the desired pH. The samples were fluxed through the column, which was then washed with HPW after the elution. The metal was recovered by acidic elution (10.0 mL). Blanks were periodically run and the results of all experiments were unaffected by flow rate in the employed range, 1.0-5.0 mL/min. Chelating Agent Loaded Resin (CALR). The chelating agent loaded resin was prepared by running 4.0 mL of 0.05 M Pyrocatechol Violet solution through the column packed with 1.0 g of AG MP1 resin, chloride form. The loaded resin bed was washed with HPW at the pH to be used in the next experiment. The samples were fluxed at 1.0 mL/min and the metal was recovered as in the PAE method. In no case did release of chelating agent from the resin occur. Metal Recouery as Function of p H . The metal recovery as a function of pH was evaluated for both the procedures. In the PAE procedure 100.0-mL samples at 1.0 mg/L metal concentration were used. Solutions, added to 2.0 mL of 0.05 M

0003-2700/87/0359-0484$01.50/00 1987 American Chemical Society