Langmuir 1990,6, 1614-1616
1614
An Efficient Iodosobenzoate-FunctionalizedPolymer for the Cleavage of Reactive Phosphates Robert A. MOSS* and Yong-Chan Chung Wright and Rieman Laboratories, Department of Chemistry, Rutgers, The State University of New Jersey, New Brunswick, New Jersey 08903 Received July 2, 1990 The cross-linked, macroreticular, acrylic anion-exchange resin IRA-35 has been readily converted to the iodosobenzoate/hexadecylammonium-functionalizedpolymer 6a. In pH 8 buffer, 6a catalyzes the cleavage of p-nitrophenyldiphenylphosphate with k+ = 0.067 s-l, a rate constant comparableto that obtained with iodosobenzoate in CTACl micellar solution. Catalyst 6a turns over when saturated with substrate, with k,,, = 0.025 s-l at pH 8. A “family”of functional polymers related to 6a is described. Much exploratory research followed reports that l-oxido1,2-benziodoxol-3(1H)-one (iodosobenzoate), 1, is a strong 0-nucleophile that rapidly cleaves reactive phosphates with catalytic turnover in cationic micellar solutions.’ The
alized polystyrene or polyacrylate reagents (e.g., 2)1° and iodosobenzoate covalently immobilized on silica (3),11 titania,12or nylon (4).12 These materials were catalytically @-CONHCH2CH2CH2N+Me2CH2CH20
I
1
continuing search for efficient decontaminants against toxic phosphorus compounds (e.g., insecticides or nerve agents)2 was expressed in studies of alternative nucleophilic reagents such as pyridine N-oxides;& (dimethylamin~)pyridines;~~,~ and copper,lb or lanthanumk chelates. Simultaneously, the continuing interest in iodosobenzoate catalysts focused on alternative media for catalyst delivery, including microemulsions,5 liquid crystals! and bolaform’ or zwitterionic surfactants.* An intensive study has also appeared of iodosobenzoate-catalyzed cleavages of nerve agents in active against the test substrate, p-nitrophenyl diphenyl the “traditional” aqueous micellar cetyltrimethylammophosphate (5, PNPDPP), but chiefly due to the difficulty nium chloride (CTACl) m e d i ~ m . ~ of achieving high iodosobenzoate loading, none of them Solid decontaminants, in contrast to the preceding fluid was as reactive as iodosobenzoate itself in CTACl micelsystems, offer special features of easy handling and lar solution.1g10-12 Now we wish t o report the simple applicability in the flow-through treatment of contaminated preparation of a highly loaded, macroreticular, acrylate water. We therefore prepared iodosobenzoate-functionresin immobilized iodosobenzoate catalyst @a),designed to mimic the l/CTA micellar catalyst, which functions with comparable reactivity in nonmicellar aqueous suspension. (1) (a) Moss,R. A.; Alwis, K. W.; Bizzigotti, G. 0. J.Am. Chem. Soc. 1983,105,681. (b) Moss,R. A.; Alwis, K. W.; shin, J.4. Ibid. 1984,106, For simplicity of preparation and catalytic efficiency, this 2651. (c) Moss, R. A.; Kim, K. Y.; Swarup, S. Ibid. 1986, 108, 788. material is the best “solid” iodosobenzoate derivative yet (2) See the series of reviews: Chem. Br. 1988,24 (7), 657-691. described. (3) (a) Katritzky, A. R.; Duell, B. L.; Rasala, D.; Knier, B.; Durst, H. D. Langmuir 1988, 4, 1118. (b) Katritzky, A. R.; Duell, B. L.; Seiders, Rohm and Haas Amberlite IRA-35 is a cross-linked,macR. P.; Durat, H. D. Ibid. 1987, 3, 976. (c) Katritzky, A. R.; Duell, B. L.; roreticular, weakly basic, acrylic anion exchange resin, with Knier, B. L.; Durst, H. D. Ibid. 1988, 4, 192. a high loading (5.4 mequiv/g by titration or N analysis) (4) (a) GeUman, S. H.; Petter, R.; Breslow, R. J. Am. Chem. SOC.1986, 108,2388. (b) Menger, F. M.;Gan, L. H.; Johnson, E.; Durst, H. D. Ibid. of pendant dimethylamino residues. Its high porosity 1987,109,2800. (c) Hay, R. W.; Govan, N. J . Chem. SOC.,Chem. Compermits facile adsorption and desorption of large organic mun. 1990,714. molecules, while its terminal NMe2 groups are readily func(5) Garlick, S.M.; Durst,H. D.;Mackay, R. A.; Haddaway, K. G.; Longo, F. R. J. Colloid Surf. Sci. 1990, 135, 508. Burnside, B. A.; Szafraniec, tionalized. This resin was converted to 6a in the following L. L.; Knier, B. L.; Durst, H. D.; Mackay, R. A.; Longo, F. R. J . Org. Chem. way. 1988,53, 2009. Burnside, B. A.; Knier, B. L.; Mackay, R. A.; Durst, H. An initial sequential washing with dilute NaOH, H20, D.; Longo, F. R. J. Phys. Chem. 1988,92, 4505. (6) Ramesh, V.; Labes, M. M. J. Chem. Soc., Chem. Commun. 1988, MeOH, EtOH, MeCN, EtOAc, and Et20 was followed by 891. Ramesh, V.; Labes, M. M. J. Am. Chem. SOC.1988,110,738; 1987, drying a t 0.1 mm-Hg for 12 h and grinding of the resin 109, 3228. (7) Bunton, C. A,; Donvin, E. L.; Savelli, G.; Si, V. C. Red. Trao. Chim. 1990, 109, 64. (8) Bunton, C. A.; Mhala, M. M.; Moffatt, J. R. J.Phys. Chem. 1989, 93, 854. (9) Hammond, P. S.; Forster, J. S.; Lieske, C. N.; Durst,H. D. J. Am. Chem. SOC.1989, 111,7860.
(10) Moss, R. A.; Bolikal, D.; Durst, H. D.; Hovanec, J. W. Tetrahedron Lett. 1988, 29, 2433. (11) Moss, R. A.; Chung, Y.-C.; Durst, H. D.; Hovanec, J. W. J.Chem. SOC.,Perkin Trans. 1 1989, 1350. (12) Moss, R. A.; Chung, Y.-C. J. Org. Chem. 1990,55, 2064.
0743-1463190 12406-1614$02.50/0 0 1990 American Chemical Societv
Langmuir, Vol. 6, No. 10, 1990 1615
Letters to a powder.13 Next, 2 g of resin (10.8 mequiv of amine) was swelled in MeOH for 3 h and then refluxed (36 h, under N2) with 2.3 g (6.5 mmol) of methyl 5-(bromomethyl)-2iodobenzoate, 7.12 The resin was filtered, washed (MeOH, EtzO), and dried to afford 4.1 g of a modified resin loaded with 1.3-1.4 mequiv/g of methyl iodobenzoateresidues (by Br- titration, Br analysis, or weight increase), corresponding to 51 % quaternization of available NMe2 moieties, leaving 49% of free NMez residues (1.4 mequiv/g, as established by titration with HC1). Then, 700 mg (0.98 mequiv of NMe2) of modified resin and 1.24 g (4.0 mmol) of n-hexadecyl bromide were maintained in refluxing MeOH for 2 days under N2. Filtering, washing (EtOH, EtnO), and vacuum drying gave 868 mg of an additionally modified resin in which 51’% of the original NMe2 groups were quaternized with methyl iodobenzoate moieties, 47 ’% were quaternized with hexadecyl residues (Br microanalysis), and 270 of the NMez residues were unchanged. Finally, the resin’s methyl iodobenzoate groups were oxidized and cyclized to the iodoxolone form, (protonated) 6a, by excess 35% peracetic acid (4 h, 25 0C).12J4 Hydrolysis, neutralization (saturated NazC03 solution), filtration, washing, and vacuum drying afforded (protona p d ) 6a, loaded with 0.90 mequiv/g of iodosobenzoate moieties (by KI/Na&03 iodometric titration9, corresponding to 91% oxidative conversion of the iodobenzoate to “iodosobenzoate” residues. By very similar chemical manipulations and analytical characterizations, we also prepared the related reagents 6b-d,in which each type of quaternary center also accounts for approximately half of all the pendant residues. Finally, we prepared catalyst 8,where 50% of the original NMe2 groups were deliberately left unquaternized,lg and catalyst 9, in which all of the NMez groups were quaternized with 7 and then oxidized (78% conversion) to iodosobenzoate residues. The final loadings of iodosobenzoate groups (mequiv/g), as determined for the new resins 6a-d,8,and 9,are shown as L values with their structural formulas. 0 COOMe
be, R
-
--
n-C16H33 [ L 0.921 ~ - C I H Q[ L 1.201 6c. R = CHI [ L 1.281 Bd, R I CH2CH2COOH [ L = 1.181
6b,R
I
-
8, R I NMe2 [ L = 0.92) 9 , R is absent ) L 1.481
For reactions in which the iodosobenzoate resins were present in stoichiometric excess relative to substrate, the cleavages of 3.3 X lo4 M (0.10 pmol) of 5 by 10 mg of resin were followed in 3-mL aliquots of 0.02 M pH 8 phosphate buffer ( p = 0.08, KC1) at 25 “C by monitoring the released p-nitrophenylate ions at 400 nm as a function of time. Our methodology for these one vial-one point kinetics experiments has been described in detail.12 Pseudo-firstorder rate constants (kJ were calculated from at least five points and had good correlation coefficients (r 2 0.99). Table I collects k+ values for the new catalysts in both buffer and in 5 x M CTACl micellar solution. Also (13) The particle size by dynamic light scattering was 0.6-3 fim. (14) Sharefkin, J. G.; Saltzman, H. Anal. Chem. 1963, 35, 1428. On the direct oxidative conversion of iodobenzoate esters to iodosobenzoates, see: Moss, R. A.; Scrimin, P.; Bhattacharya, S.; Chatterjee, S. Tetrahedron Lett. 1987, 28, 5005. (15) Lucas, H. J.; Kennedy, E. R.; Formo, M. W. In Organic Synthesis; Horning, E. C . , Ed.; Wiley: New York, 1955; Collect. Vol 3, p 483. (16) HCl titration indicated that peracid oxidation of the iodo to iodoso residues did not simultaneouslyconvert the dimethylaminogroups to dimethylamine oxides.
Table I. Kinetics of Cleavage of PNPDPP by Supported Iodoeobenzoates. no CTACl catalyst 6e 6b 6c
6d 8 9
9 3h 4’
Mequivb k+, s-l 0.0092 0.012 0.013 0.012 0.0092 0.015 0.0032 0.0010 0.0025
0.067 0.016 0.0074 0.0019 0.0052 0.017 0.011 0.018 0.012
kz, M-l s-l 22 4.0 1.7 0.48 0.17 3.4 17 40 14
5 X 10“ M CTACl
k+, s-l e 0.0091 0.0034 0.051 0.071 0.049 g g 0.067’
k2, M-’ s-l e 2.3 0.80 13 23 9.9 g g
Conditions: 0.02 M pH 8 phosphate buffer, p = 0.08 (KCl), 25 OC, [PNPDPP) = 3.3 X 10-6 M. Mequiv of iodosobenzoate in 10 mg of solid catalyst in 3 mL of buffer solution. Pseudo-firstorder rate constant for the release of p-nitrophenylate. d k+/ [catalyst], treating the catalyst as if it were soluble. e Too fast to be followed by one vial-one point method. Reference 10, fi = 0.1. 8 Not done. Reference 11. Reference 12. j In 1 X 10“ M CTACl.
*
*
included are comparable data for the previously studied supported iodosobenzoate catalysts 2,1° 3,11 and 4.12 The tabulated kz values represent “bimolecular”rate constants calculated as if the bound iodosobenzoate residues were dissolved in the solution. The resulting k2 (=k$/ “[iodosobenzoate]”) affords a rough way of comparing the efficiencies of the catalysts in the sense of kinetic potency per available iodosobenzoate. The apparent pK, of the I-OH group of 9 was determined from a k+/pH profile in which 9 was used to cleave 5 in 0.02 M aqueous phosphate buffers ( p = 0.08, KC1) a t six different pH values in the range 7.0-8.5. A plot of log k+ vs pH showed a sharp discontinuity at pH 7.44, which was taken as the pKa of 9. This value is typical of the iodosobenzoate residue in cationic surroundings, either in CTACl solution or on solid supports,l1J2and indicates that -80% of the catalyst’s I-OH groups will be in the reactive 1-0- form at pH 8.” Several important observations emerge from the data. In terms of k+ per 10 mg of catalyst, the 1:l iodosobenzoate/hexadecyl resin 6a is the most reactive of the resins in buffer alone. The hexadecyl chain appears to be essential to the reactivity toward PNPDPP; the fully iodosobenzoate-functionalizedresin, 9, is less reactive than 6a,and analogues of 6a with short auxiliary chains (6bor 6c) are also less reactive, even at higher iodosobenzoate loadings. This suggests that the long alkyl chains provide binding sites on the resins’s surface for the hydrophobic PNPDPP substrate molecules, an idea with precedence in reactions catalyzed by functional polymers.18 Indeed, by judicious choice of the auxiliary chain in 6,one might be able to “fine-tune”the reactivity toward substrates that are less hydrophobic than 5. If we regard 6a as a kind of covalently bound iodosobenzoate-CTA, it is perhaps not surprising to note comparable reactivity toward PNPDPP of solid 6a in buffer and iodosobenzoate 1 in micellar CTACl solution;1bk+ = 0.067 and 0.065 s-l, respectively. (17) We take this result as applicable for all the reagents in Table I except 6d, where 50% of the pendent residues will carry anionic carboxylate moieties at pH 8. These will decrease the aggregate positive surface charge due to the quaternary ammonium ion centers, suppress I-OH ionization, and raise the pK,. (18) Klotz, I. M.; Royer, G. P.; Scarpa,I. S. R o c . Natl. Acad. Sci. U.S.A. 1971,68, 263. Brown, J. M.; Jenkins, J. A. J. Chem. Soc., Chem. Commun. 1976,458. Kunitake, T.; Okahata, Y. J. Am. Chem. Soc. 1976,98, 7793. Meyers, W. E.;Royer, G. P. Ibid. 1977,99,6141. Okahata, Y.; Kunitake, T. J.Mol. Catal. 1979,6, 163. Tanaka, N.; Hosoya, K.; Iwaguchi, K.; Araki, M. J. Am. Chem. SOC.1984,106, 3057.
1616 Langmuir, Vol. 6, No. 10, 1990
f
-5 0
Letters
08-1
oo+
I
f
.
, 5
.
.
, 10
I
15
.
,
.’
20
TIME (MIN)
Figure 1. Burst kinetics for t h e cleavage of P N P D P P (5) by
reagents 6a (A) and 9 (B).Plotted are the absorbances of released p-nitrophenylate vs time for cleavages of 2 X 10-4 M P N P D P P by 4 mg of 6a or 9. T h e values of [p-nitrophenylate] a t zero time (“burst”) are extrapolated from the d a t a points.
Positive charge on the polymer surface enhances catalyst reactivity, possibly by accommodating the negative charge that builds up on the substrate during attack by I-0-. Thus, 6d, carrying a negatively charged propionate auxiliary chain, is the least reactive catalyst in buffer M cationic CTACl micelsolution. However, in 5 X lar solution, its reactivity is strongly augmented, surpassing that of butylammonium catalyst 6b under similar conditions. This suggests that CTACl binds at the polymer surface, providing both positive charge and hydrophobic binding sites, possibly functioning as a kind of solidliquid phase-transfer agent, facilitating substrateiodosobenzoate interaction a t the solid’s surface.I2 CTACl potentiates the reactivity of other catalysts in Table I, including 6a, 8, 9, and 4,11 but the strongest enhancement, a factor of 27, is for 6d. Presumably, this reflects CTACl charge neutralization of the catalyst’s carboxylate groups, and the resultant increases in surface hydrophobicity and aggregate positive charge. It is not clear, however, why 6b and 6c are less reactive in CTACl than in buffer solution. Table I also includes k z , “bimolecular” rate constants that reflect the “concentration” of iodosobenzoateresidues, if these are imagined to be dissolved in solution. Too much significance should not be given to these data because they , by a single are single-point values (i.e., k ~divided “concentration”). Nevertheless, resin 6a is seen to be superior to all the other reagents, excepting only silicasupported iodosobenzoate 3.” On a per-gram basis, 6a is
a more potent catalyst than 3 because of its 9-fold higher iodosobenzoate loading, whereas, on a per-residue basis, the iodosobenzoates in 3 are about twice as reactive as those of 6a. This may be due to the greater “wettability” of the silica surface in 3. Most importantly, 6a is a true catalyst for the cleavage of PNPDPP, turning over rapidly when saturated with PNPDPP. When challenged in pH 8 buffer with substrate in excess of the catalytically active12iodosobenzoate groups, rapid phosphorylation of 6a (or 9) and release of p-nitrophenylate ions were observed, followed by a slower, linear increase of absorbance with time (“burst”kineti~s).~~J2J9 The latter kinetic phase afforded “turnover” rate constants of 0.025 s-l for 6a and 0.0070 s-l for 9 (see Figure 1).20 Control experiments with 9 under turnover conditions showed that it could be recycled at least 4 times without a significant decrease in its reactivity toward PNPDPP. For comparison, silica-supported 3 and nylon-supported reagent 4 gave turnover rate constants of 0.018 and 0.002 s-l, respectively, under comparable conditions. Thus, in terms of turnover also, 6a is a superior catalyst for the destruction of PNPDPP.21 We conclude that iodosobenzoate/hexadecylammoniumfunctionalized resin 6a is an easily prepared, efficient catalyst for the rapid cleavage of reactive phosphates in weakly basic aqueous buffers. It is more reactive than previous iodosobenzoate resins, and its composition (type and loading of auxiliary residues) can readily be varied. At least in principle, the latter feature means that a family of related catalysts can be prepared with their hydrophobicity or wettability adjusted for particular substrates or applications. Acknowledgment. We are grateful to the U.S.Army Research Office and to P P G Industries for financial support. (19) Bender, M. L.; Kbzdy, F. J.; Wedler, F. C. J. Chem. Educ. 1967, 44, 84.
(20) The turnover constants are derived from the slopes of the linear correlations in Figure 1,corrected for the kinetically available iodosobenzoate (indicated by the burst). The available iodosobenzoate, a8 a percentage of the iodometrically titratable moiety, was 7 % for 6a and 4 % for 9. See ref 12 for discussion of this effect. (21) The (turnover) rate constant for OH- (buffer) mediated hydrolysis of 0-phosphorylated 6a is about 3 times less than k l for the phosphorylation by PNPDPP, so that the hydrolysis step is rate-limiting in the presence of excess substrate. Nevertheless, the relatively high value of k m suggests CTACI-like catalysis of dephosphorylation by OH- bound at neighboring ammonium ion sites.
