Studies on the Effects of Imidazole on the Peroxyoxalate Chemiluminescence Detection System for High Performance Liquid Chromatography Kazuhiro I m a i t , Atsuhiko Nishitani, Yukie Tsukamoto, Wei-Hong Wang a nd Susumu Kanda Branch Hospital Pharmacy, University of Tokyo, 3-28-6 Mejirodai, Bunkyo-ku, Tokyo 112 Japan

Kazuichi H a y ak awa and Motoichi Mi y azaki Faculty of Pharmaceutical Sciences, Kanazawa University, Takara-machi 13-1, Kanazawa 920, Japan

The catalytic effect of bases (imidazole, pyridine, Tris and triethylamine) on the peroxyoxalate chemiluminescence (PO-CL) reaction for high performance liquid chromatography (HPLC) was investigated. Imidazole increased PO-CL intensity extraordinarily, whereas the other bases (pyridine, Tris and triethylamine) did not. The peak heights of dipyridamole (coronary vasodilator) obtained using the eluents containing buffers were largest a t pH 7.0, a few times less at pH 6.0 and pH 5.0, 100 times less at pH 4.0 and a few hundred times less at pH 3.0. The eluents containing buffers at pH 3, 4, 5, 6 or 7 each with imidazole increased the peak heights by a few to ten times as compared with those without imidazole, and those peak heights were within one order of magnitude. On the other hand, the eluent containing buffer at pH 2 did not affect the peak heights with or without imidazole. Bis(4-nitro-2(3,6,9-trioxadecyloxycarbonyl)phenyl) oxalate (TDPO) alone and bis(2,4-dinitrophenyI)oxalate (DNPO) plus T D P O were recommended to be used against eluents containing buffers of pH 5 - 7 and pH 3 4, respectively. Dipyridamole and benzydamine hydrochloride (anti-inflamatory drug) were separated on the ODS column and detected by the present system. The detection limits of dipyridamole and benzydamine hydrochloride were 40 amol and 270 fmol, respectively.

-

INTRODUCTION

Since the introduction of the peroxyoxalate chemiluminescence (PO-CL) reaction (Scheme 1) to the 0 0

0

RO-C

iiii c-c

0

I!

I!

-C

I

-OR+HzOz-*

0

D+FluFlu*

-

I

0 D

+2ROH

Flu* (excited \tate)+2CO2 Flu + h u

Flu = Fluorophore R = *NO1

COO(CH~CH~O)SCH~ TDPO 1. Proposed mechanism for t h e peroxyoxalate chemiluminescence reaction. TDPO; bis(4-nitro-2-trioxadecyloxycarbony1)phenyl oxalate.

Scheme

detection system for high performance liquid chromatography (HPLC) (Kobayashi and Imai, 1980a), highly sensitive determination has been achieved (Imai, 1986) of fluorescent compounds such as polycyclic aromatic hydrocarbons (Sigvardson and Birks, 1983; Grayaski and Weber, 1984; Sigvardson et al., 1984; Sigvardson and Birks, 1984; Weinberger et al., 1984) and drugs (Imai et al., 1987b), and fluorescent labeled compounds such as amino acids, amines, 0x0-steroids, t To whom cotlzspondence should be addressed

thiols, aldehyde, carboxylic acid and drugs with DNS-CI (5-N,N-dimethylaminonaphthalene sulfonyl chloride) (Kobayashi and Imai, 1980a; Mellbin, 1983; Miyaguchi et ul., 1984a; Mellbin and Smith, 1984; de Jong et al., 1984; Koziol et al., 1984; Miyaguchi et al., 1986; Nozaki et al., 1988; Hayakawa et al., 1989), DNS-hydrazine (Imai el al., 1989b), NBD-F (4-fluoro-7-nitro-2,1,3benzoxadiazole) (Mellbin and Smith, 1984; Imai, 1986), fluorescamine (Kobayashi et a/., 1981), OPA (o-phthaldialdehyde) (Mellbin and Smith, 1984), NDA (naphthalene-2,3-carboxadialdehyde) (Kawasaki et al., 1988), N-(4-(6-dimethylamino-2-benzofuranyl)phenyl) maleimide (Nakashima e f al., 1989), 3-aminofluoranthene (Mann and Grayaski, 1987) and 7-dimethylamino3-(4-iodoacetylaminophenyl)-4-methylcoumarin (Grayaski and DeVasto, 1987). Optimization of the reaction detection system has also been investigated in terms of solvent delivery system (Mellbin, 1983; Mellbin and Smith, 1984), solvent composition (Kobayashi and lmai, 1980a; Weinberger, 1984), kinds of oxalates (Honda et a/., 1985a; Imai et al., 1986; Poulson et al., 1986a; de Jong et al., 1986; Imai et al., 1987a), stability of oxalates (Poulson et al., 1986b; Imaizumi et a/., 1989; Hanaoka, 1989), fluorescent compounds (Sigvardson and Birks, 1984; Honda et al., 1985b; Imai et al., 1987a,b; van Zoonen et al., 1987), pH (Kobayashi and Imai, 1980a; Weinberger, 1984), salt species (Honda et al., 1983), water content (Kobayashi and Imai, 1980a; Weinberger, 1984; Jennings and Capomacchia, 1988), temperature (Hanaoka et al., 1988), quenching (van Zoonen et al., 1986), mixing (Kobayashi and Imai, 1980b), Row cell volume (de Jong et al., 1984; Weber and

CX’C 0269-3879/90/0100-0104 $2.50 100

BIOMEDICAL CHROMATOGRAPHY, VOL 4, NO. 3, 1990

$3

Heyden & Son Limited, 1990

EFFECTS O F IMIDAZOLE O N THE PO-CL FOR HPLC

Grayaski, 1987) and luminescence detection devices (Mellbin 1983). As to the effect of salt species of the buffer in the eluent, our previous data revealed that halide ions (chloride, bromide and iodide) gave a negative effect on the PO-CL intensity whereas nitrate and phosphate ions had no effect (Honda et al., 1983). Thus, nitrate imidazole buffer which is miscible with both organic and aqueous solvent was used as a component of the eluent for the HPLC-PO-CL detection system (Imai et al., 1985). Since then, many studies have adopted nitrate imidazole buffer as a component of the eluent for HPLC. Recently, kinetic studies of the effects of imidazole on the PO-CL reaction (Hanaoka et al., 1988) have been reported. We have also presented a preliminary report on the effective use of imidazole on the PO-CL reaction detection system in HPLC (Imai et al., 1989a). In this paper, a detailed investigation is described for the practical use of imidazole catalyst in the PO-CL detection system for HPLC.

EXPERIMENTAL Chemicals. Bis(4-nitro-2-(3,6,9-trioxadecyloxycarbonyl)phenyl) oxalate (TDPO) and bis(2,4-dinitrophenyI) oxalate (DNPO) were purchased from Wako Pure Chemical Industires Ltd. (Osaka, Japan). Dipyridamole and DNS-serine (DNSSer) were from Sigma (St. Louis, Mo. USA). Tris(hydroxymethy1)aminomethane was from Daiichi Pure Chemicals Co. (Tokyo, Japan). Hydrogen peroxide (30%) was purchased from Mitsubishi Gas Kagaku (Tokyo, Japan). Imidazole (Merck, FRG) was used after recrystalising once from diethyl ether. Triethylamine (special grade), pyridine (special grade), distilled water (HPLC grade), acetonitrile (HPLC grade), and ethyl acetate (HPLC grade) were from Wako. Benzydamine hydrochloride was donated from Daiichi Pharmaceutical Co. (Tokyo, Japan). All other chemicals were of reagent grade. Time course of PO-CL reaction in a static system. 150 p L 10 p~ of dansylated (DNS) serine in 25 mM phthalate buffer (pH 4.5) with or without imidazole and 150 pL of 1.0 mM hydrogen peroxide in acetonitrile were pre-mixed in a borosilicate glass tube (50 x 6 mm ID) with a vortex mixer for several seconds, then 20 p L of 0.5 mM TDPO in acetonitrile was added with a microsyringe. The tube was immediately placed into a ChemGlow photometer (Aminco, Baltimore, MD, USA) and the time course of the CL intensity was recorded. HPLC-chemiluminescence detection system. The pumps used were a KHP-011 (Kyowa Seimitsu Co., Tokyo, Japan) for the eluent and a LC-5A (Shimadzu Seisakusho Co., Tokyo, Japan)

Eluent

Water Bath Figure 1. Flow diagram for the chemiluminescence reaction detection system for HPLC. TDPO, Bis(4-nitro-2-trioxadecyloxycarbonyl)phenyloxalate; P, pump; C, dummy column; I, injector; HzOz, hydrogen peroxide; MD, rotating flow mixing device; R, recorder; Det, chemiluminescence monitor.

@ Heyden & Son Limited, 1990

for the reagent solution. The injection valve (Rheodyne, Cotati, CA, USA) with a 20 p L loop, an analytical column (TSK ODS SOTM, 150 x 4 mm ID, 5 p ~ Tosoh , Co., Tokyo, Japan) and a dummy column for the reagent solution (TSK ODS-l20A, 100 x 4.6 mrn ID, 5 pM, Tosoh) were used. A chemiluminescence monitor, 825-CL (JASCO, Tokyo, Japan) with a 93 pL spiral type flow cell and a recorder (Technicorder type 3047, Yokogawa Denki, Tokyo, Japan) were used. A 25 p L rotating flow mixing device (Kyowa Seimitsu) was heated at 35 "C in a water-bath. The flow line (PTFE tubing, 50 x 0.5 mm I D and 5 0 x 0.8 mm ID) was connected to the detector next to the mixing device. The final composition of the eluent for dipyridamole and benzydamine hydrochloride was 50 mM imidazole in 50 mM phosphate buffer (pH 6.0) +acetonitrile (1 : 1, v/v), the flow rate of which was 1 mL/min. For the apparent time course studies in the HPLC-PO-CL detection system (Fig. l ) , 0.05% trifluoroacetic acid (pH 2), 50 mM phthalate buffer (K, pH 3, 4, and 5) and 50 mM phosphate buffer (pH 6 and 7) with or without 50mM imidazole or other amines were adopted instead of 50 mM phosphate buffer (pH 6.0) alone in the eluent above. The chemilumigenic reagent solution was 0.5 mM TDPO in acetonitrile and 25 mM hydrogen peroxide in ethyl acetate (1: 1, v/v) with various flow rates ranging from 1.5 to 3.0 mL/min to obtain the apparent time course. The apparent reaction times were calculated by the residence time of dipyridamole between the merging of the chemilumingenic reagent and the entrance of the monitor. Both solutions were mixed just prior to use. For the apparent time course studies utilizing the buffers of pH 2.0,3.0 and 4.0 containing imidazole, 4 m M DNPO with 0.8mM TDPO was used as the chemilumigenic reagent instead of TDPO alone. The stock solutions of dipyridamole in acetonitrile and benzydamine hydrochloride in distilled water were diluted with the eluents to appropriate concentrations, and 20 pL samples were subjected to HPLC analysis.

RESULTS AND DISCUSSION Although the original usage of imidazole nitrate for the PO-CL reaction was intended for the elimination of the quenching effect of inorganic salts such as halide ions and also the solubility in the organic constituent in the eluent, the catalytic effect of imidazole on the PO-CL reaction has been gradually acknowledged. As shown in Fig. 2, the catalytic effect was clearly demonstrated at the lower pH of 4.5 in a static system. The experiment was not performed at pH 6 since at pH 6 the reaction was too fast to follow in the static system although pH 6 buffer was usually adopted in the HPLC-PO-CL detection system (Imai et al., 1985). This effect was also demonstrated by the stopped flow technique (Hanaoka et a[., 1988). The effect of several other bases was also examined in the PO-CL detection system for HPLC. Figure 3 shows that the addition of pyridine, Tris or triethylamine to the pH 3 buffer lowered the relative CL intensity. In order to extend the popular use of PO-CL detection system for HPLC, the C L reaction should afford similar CL intensities to any reversed phase column eluates which constitute the reaction medium. However, as shown in Fig. 4, the peak heights of dipyridamole, an efficient fluorescent drug for PO-CL (Imai et al., 1987b) were different on the pHs of the column eluates. The heights appeared to be almost constant between 3.7 x lo-' and 5.7 x lop1s reaction for every pH, but the BIOMEDICAL CHROMATOGRAPHY, VOL. 4, NO. 3, 1990 101

K. IMAl ET AL.

u)

0) .-c .-

ln W

.I-

.-

W C

Q

P Q

.-

5 .-

E,

L

L

V

‘I

W

.-

-mW

.I-

a

1

11 0

1 .o

0 3.8 4.0

3.0

2.0

4.5 5.0 5.5 Reaction time (s x 10 ’)

6.0

Time (min) Figure 2. Effect of imidazole on the PO-CL reaction in a static system. For experimental conditions are as given in the text. The following buffers were included in the reaction mixture: M, 25 mM phthalate buffer (pH 4.5); A, 0.25 JLM imidazole +25 mM phthalate buffer (pH 4.5); V, 0.5 JLM imidazole +25 mM phthalate buffer (pH 4.5); 0 , 5 JLMimidazole +25 mM phthalate buffer (pH 4.5); 0, 25 JLM imidazole +25 mM phthalate buffer (pH 4.5).

heights at pH 7.0 were the largest, a few times less at pH 6.0 and pH 5.0, 100 times less at pH 4.0 and a few hundred times less at pH 3.0. On the basis of the finding of the catalytic effect of imidazole mentioned above, the effect of imidazole was investigated at various buffers of various pHs. The addition of imidazole to pH 7.0 buffer gave about 10 times increment of the peak height, but accelerated the decomposition of TDPO (Fig. 5). In pH 6.0, 5.0,4.0

Figure 4. Effect of various pHs on the apparent time course of the peak height of dipyridamole in the HPLC-PO-CL detection system utilizing TDPO as a chemilumigenic reagent. For experimental conditions refer to the text. The following buffers were included in the eluent: 0, phosphate buffer (pH 7.0); A, phosphate buffer (pH 6.0); M, phthalate buffer (pH 5.0); V,phthalate buffer (pH 4.0); 0, phthalate buffer (pH 3.0).

and 3.0 buffer, the increment was about 10 times. The peak heights at 3.8 x lo-’ s at pH 7 , 6 and 5 were almost the same, and those at pH 3 and 4 were a few hundred times less. At pH lower than 4, the PO-CL reaction gave very low intensity with TDPO. Therefore, DNPO was used instead (Honda et al., 1985a). On this occasion, 0.8 mM 1O6

1oz .I-

.c

105

0 .W c

E0

Y

.W .c

0

Y

.-

n W

m

-m

.I-

W

n 10

$ lo3

W

..I-

-mQ a

1

0 ” 4.5

5.0

5.5

6.0

6.5

7.0

Reaction time (s x10-’) Figure 3. Effect of various bases on the apparent time course of the peak height of dipyridamole in the HPLC-PO-CL detection system i$tizing TDPO as a chemilumigenic reagent. For experimental conditions refer to the text. The following buffers were included in the eluent: 0,50 mM phthalate buffer (pH 3.0); A,50 mM imidazole +50 mM phthalate buffer (pH 3.0); 0,50 mM triethylamine+50 mM phthalate buffer (pH 3.0); 0, 5 mM pyridine+50 mM phthalate buffer (pH 3.0); V, 50 mM tris(hydroxymethy1)aminomethane+50 mM phthalate buffer (pH 3.0).

102 BIOMEDICAL CHROMATOGRAPHY, VOL. 4, NO. 3, 1990

10

0

3.8 4.0

4.5

5.0

5.5

6.0

Reaction time ( s x 10 ’) Figure 5. Effect of imidazole on the apparent time course of the peak height of dipyridamole in the HPLC-PO-CL detection system utilizing TDPO as a chemilumigenic reagent. For experimental conditions refer to the text. The following buffers were included in the eluent: 0 , 50 mM imidazole +phosphate buffer (pM 7.0); A, 50 mM imidazole +phosphate buffer (pH 6.0); M, 50 mM imidazole+ phthalate buffer’ (pH 5.0); V, 50 mM imidazole +phthalate buffer (pH 4.0); 0, 50 mM imidazole +phthalate buffer (pH 3.0).

@ Heyden & Son Limited, 1990

EFFECTS OF IMIDAZOLE O N THE PO-CL FOR HPLC

TDPO was added to 4 m DNPO ~ solution because of the stability enhancement of DNPO (88% vs 62% activity after 8 h at room temperature) by the addition of TDPO irrespective of the presence of hydrogen peroxide, the reasons for which are not known. As shown in Fig. 6, DNPO gave higher peak heights of dipyridamole at pH 4.0 and pH 3.0 buffers than those using TDPO alone (Fig. 3). The addition of imidazole to the buffers at pH4.0 and 3.0 also increased the peak height by 100 times at both pHs. However, no effect was observed by the addition of imidazole at pH 2.0. The increment of the peak heights observed by the addition of imidazole to pH 4 and 3 buffers was found to be a successful method for the practical use of the PO-CL detection system, since the peak heights were similar to those obtained at pH 5 and 6 without imidazole and about one tenth of those obtained by the addition of imidazole at pH 7.0, 6.0 and 5.0. Considering these data, the effect of imidazole on the PO-CL reaction detection system would be up to pH 3 . Therefore, another catalyst would be desirable for the effective PO-CL detection system for a column eluate containing pH 2.0 buffer. Otherwise, two separate solvent delivery pumps should be adopted, one for TDPO in acetonitrile and the other for hydrogen peroxide in a buffer to adjust the pH of the column eluate.

L dipyridamole

b., 0

2

r'

4 6 8 10 Retention time (min)

Figure 7. Chromatogram of dipyridamole obtained by the HPLC-POCL detection system. For experimental conditions refer to the text. 2 fmol Dipyridamole.

Since, in the present study, a high intensity at pH 6 with TDPO was obtained by use of imidazole buffer, this condition was applied to the HPLC-PO-CL detection system of two fluorescent drugs, dipyridamole and benzydamine hydrochloride. The former gave high CL and the latter gave low CL (Imai et al., 1987b). Figures 7 and 8 show chromatograms of dipyridamole and benzydamine hydrochloride separated on the ODS column and detected. The detection limits for

benzydamine

0

3.8 4.0

4.5 5.0 5.5 Reaction time (s x lo-')

6.0

Figure 6. Effect of imidazole on the apparent time course of the peak height of dipyridamole in the HPLC-PO-CL detection system utilizing DNPO-TDPO as chemilumigenic reagents for experimental conditions refer to the text. The following buffers were included in the eluent: W, 50 mM imidazole +50 mM phthalate buffer (pH 4.0); A, 50 mM imidazole +50 mM phthalate buffer (pH 3.0); 0 , 50 mM 50 mM phthalate imidazole +0.05% trifluoroacetic acid (pH 2.0); 0. buffer (pH 4.0); A,50 mM phthalate buffer (pH 3.0); 0, 0.05% trifluoroacetic acid buffer (pH 2.0).

0 Heyden & Son Limited,

1990

-Aw 0

2 4 6 8 Retention time (min)

10

Figure 8. Chromatogram of benzydamine hydrochloride obtained by the HPLC-PO-CLdetection system. For experimental conditions refer to the text. 2 pmol Benzydamine hydrochloride.

BIOMEDICAL CHROMATOGRAPHY, VOL. 4, NO. 3, 1990 103

K. IMAl ET

dipyridamole and benzydamine hydrochloride (antiinflamatory drug) were 40 atmol and 270 fmol, respectively, which were lower than those (160 atmol and 880 fmol, respectively) in the previous paper (Imai et at., 1987b).

AL.

Acknowledgements The authors express their thanks to Tosoh Co. for the generous gift of TSK ODS ~ O T M columns. A part of the work was supported by a Grant-in-Aid from The Tokyo Biochemical Research Foundation.

REFERENCES de Jong, G. J., Lammers, N., Spruit, F. J., Brinkman, U. A. Th. and Frei, R. W. (1984). Cbromatographia. 18, 129. de Jong, G. J., Lammers, N., Spruit, F. J., Frei, R. W . and Brinkman, U. A. Th. (1986). J. Chromatogr. 353, 249. Grayeski, M. L. and Weber, A. J. (1984). Anal. Lett. 17, 1539. Grayeski. M . L. and De Vasto, J. K. (1987). Anal. Chem. 59, 1203. Hanaoka, N., Givens, R. S., Schowen, R. L. and Kuwana, T. (1988). Anal. Cbem. 60.2193. Hanaoka, N. (1989). Anat. Cbem. 61, 1300. Hayakawa, K., Hasegawa, K., Imaizumi, N., Wong, 0. S a n d Miyazaki, M. (1989). J. Cbromatogr. 464, 343. Honda, K., Sekino, J. and Imai, K. (1983). Anal. Chem. 55, 940. Honda, K., Miyaguchi, K. and Imai, K. (1985a). Anal. Cbim. Acta 177, 103. Honda, K., Miyaguchi, K. and Imai, K. (1985b). Anal. Chim. Acta 177, 111. Imai, K., Miyaguchi, K. and Honda, K. (1985). In 'Bioluminescence and Chemiluminescence: Instruments and Applications', ed. by Knox van Dyke, vol. 11, pp. 65-75, CRC Press, New York. Imai, K. (1986). In 'Methods in Enzymology', ed by M. A. Deluca & W. D. McElroy, vol 133. part B, p p 435-449, Academic Press, New York. Imai, K., Nawa, H.. Tanaka, M. and Ogata, H. (1986). Analyst 110,209. Imai, K.. Matsunaga. Y.. Tsukamoto. Y. and Nishitani. A. (1987a). J. Chromatogr. 400,169. Imai. K.. Nishitani, A. and Tsukamoto, Y. (1987b). Cbromatographia 24, 77. Imai, K., Nishitani, A., Tsukamoto, Y. and Akitomo, H. (1989a). In 'Xenobiotic metabolism and Desposition', ed. by R. Kato, R. W . Estabrook and M. N. Cayen, p. 325, Taylor & Francis, London. Imai, K., Higashidate, S., Nishitani, A. and Tsukamoto, Y. (1989b). Anal. Cbim. Acta, 227, 21. Imaizumi, N., Hayakawa, K., Miyazaki, M. and Imai, K. (1989). Analyst 114, 161. Jennings, R. N. and Capomacchia, A. C. (1988). A n a l Chim. Acta. 205, 207. Kawasaki, T., Wong, 0. S.. Wang, C. and Kuwana, T. (1988). J. Res. National Bureau Stnd. 93, 504.

104 BIOMEDICAL CHROMATOGRAPHY, VOL. 4, NO. 3, 1990

Kobayashi, S. and Imai. K. (1980a). Anal. Cbem. 52,424. Kobayashi, S . and Imai, K. (1980b). Anal. Cbem. 52, 1548. Kobayashi, K., Sekino, J., Honda, K. and Imai, K. (1981). Anal. Biochem. 112, 99. Koziol, T., Grayaski, M. L. and Weinberger, R. (1984). J. Cbromatogr. 317, 355. Mann, B. and Grayeski, M. L. (1987). J. Cbromatogr. 386, 149. Mellbin, G. (1983). J. Liq. Chromatogr. 6, 1603. Mellbin, G. and Smith, B. E. F. (1984). J. Chromatogr. 312, 203. Miyaguchi, K., Honda, K. and Imai, K. (1984a). J. Cbromatogr. 303, 173. Miyaguchi, K., Honda, K. and Imai, K. (1984b). J. Chromatogr. 316, 501. Miyaguchi, K., Honda, K., Toyo'oka, T. and Imai, K. (1986). J. Cbromatogr. 352, 255. Nakashima, K., Umekawa, C., Nakatsuji, S., Akiyama, S. and Givens, R. S. (1989). Biomed. Chromatogr. 3, 39. Nozaki, 0.. Ohba, Y. and Imai, K. (1988). Anal. Chim. Acta 205, 255. Poulsen, J. R., Birks, J. W., Gubitz, G., Van Zoonen, P., Gooijer, C., Velthorst, N. H. and Frei, R. W. (1986a) J. Chromatogr. 360,371. Poulsen, J. R., Birks, J. W.. van Zoonen, P., Gooijer, C., Velthorst, N. H. and Frei. R. W. (1986b). Chromatographia 21. 10. Sigvardson, K. W. and Birks, J. W. (1983). Anal. Chem. 55, 432. Sigvardson, K. W., Kennish, J. M. and Birks, J. W. (1984). Anal. Cbem. 56, 1096. Sigvardson, K. W. and Birks, J. W. (1984). J. Chromatogr. 316, 507. van Zoonen, P., Kamminga, D. A., Gooijer, C., Velthorst, N. H., Frei. R. W. and Gubitz, G. (1986). Anal. Cbem. 58, 1245. van Zoonen. P., Kamminga, D. A,, Gooijer. C., Velthorst, N. H. and Frei, R. W. (1987). J. Liq. Cbromatogr. 10. 819. Weber, A. J. and Grayeski, M. L. (1987). Anai. Cbem. 59, 10. Weinberger, R., Mannan, C. A,, Cerchio, M. and Grayeski, M. L. (1984). J. Cbromatogr. 288, 445. Weinberger, R. (1984). J. Chromatogr. 314, 155.

Received 27 October 1989; accepted 9 November 1989

@ Heyden & Son Limited, 1990