RhCl(PPh3)3 and RhH(PPh3)4 Catalyzed Hydrogenation of

Dec 4, 1996 - Detailed kinetic and selectivity data, collected according to a statistical design framework, have clearly defined the effect of varying...
7 downloads 11 Views 189KB Size
Download PDF
Ind. Eng. Chem. Res. 1996, 35, 4417-4423

4417

RhCl(PPh3)3 and RhH(PPh3)4 Catalyzed Hydrogenation of Acrylonitrile-Butadiene Copolymers J. Scott Parent, Neil T. McManus, and Garry L. Rempel* Department of Chemical Engineering, University of Waterloo, Waterloo, Ontario, Canada N2L 3G1

Homogeneous catalyst precursors of the forms RhCl(PPh3)3 and RhH(PPh3)4 are applied commercially for the selective hydrogenation of acrylonitrile-butadiene copolymers. This report details a study of their catalytic behavior at reaction conditions approaching those found in industrial settings. Detailed kinetic and selectivity data, collected according to a statistical design framework, have clearly defined the effect of varying process conditions. The kinetic results suggest the behavior observed under severe conditions is consistent with that reported for pressures and temperatures near ambient. Dilute solution viscosity data are used to demonstrate the uniform selectivity of both RhCl(PPh3)3 and RhH(PPh3)4 catalyzed hydrogenations. A probable reaction mechanism for the RhCl(PPh3)3 system has been derived using analysis of variance model discrimination procedures. Introduction In many respects the commercial demand for more robust elastomers has extended beyond the limit of traditional butadiene-based polymers. Residual carboncarbon unsaturation in the polymer structure, being reactive toward thermal and oxidative degradation, compromises the material’s mechanical integrity when exposed to harsh operating conditions. Chemical modification of these copolymers has greatly improved their performance, creating a new class of specialty elastomers with an expanded useful range (McGrath et al., 1995). A commercially important example is the catalytic hydrogenation of acrylonitrile-butadiene copolymer to yield its tougher and more stable derivative, hydrogenated nitrile butadiene rubber (HNBR) (Hashimoto et al., 1984). To be an efficient HNBR synthesis, a catalytic process must selectively hydrogenate olefin in the presence of the copolymer’s nitrile unsaturation as illustrated in Figure 1. Any reduction of the polar nitrile groups to imine or amine comprises the oil resistant properties for which the material is designed. Furthermore, such undesirable nitrile hydrogenations are suspected to promote polymer cross-linking which reduces the elastomer’s processability (McManus and Rempel, 1995). Two homogeneous catalysts, RhCl(PPh3)3 and RhH(PPh3)4, are known to exhibit the activity and selectivity required to produce HNBR economically on a large scale (Oppelt et al., 1976; Rempel and Azizan, 1984). However, despite their commercial significance, little attention has been paid to the fundamental hydrogenation chemistry of these catalysts at conditions that are employed industrially. In addition to a brief report by Weinstein (1984), Mohammadi and Rempel (1987b) have carried out a mechanistic study of RhCl(PPh3)3/NBR system at the mild reaction conditions of 65 °C and ambient pressures. Bhattacharjee et al. (1991) extended these process conditions to 100 °C and 56 bar in an effort to optimize the process and characterize the final product. However, little is known of the kinetic behavior of RhCl(PPh3)3 at these elevated conditions, and to our knowledge, no reports on the RhH(PPh3)4/NBR system are present in the open literature. This paper is an attempt to appreciate the underlying chemistry of both the RhCl(PPh3)3 and RhH(PPh3)4 catalyzed hydrogenations of NBR at conditions that are S0888-5885(95)00668-3 CCC: $12.00

Figure 1. Selective hydrogenation of acrylonitrile-butadiene copolymer. Table 1. Range of Process Factors Analyzed Rh concentration, [Rh]T system pressure, PH2 additional PPh3, [PPh3] nitrile loading, [CN] temperature, T

19.6-140 µM 4.85-76.2 bar 0.02-5.60 mM 47.6-255 mM 115-155 °C

relevant to industrial applications of the technology (Table 1). Carefully obtained kinetic and selectivity data are presented within the context of fundamental reaction mechanisms. Model discrimination techniques are used to support a likely catalytic pathway for the RhCl(PPh3)3 catalytic system. Experimental Section RhCl(PPh3)3 and RhH(PPh3)4 were synthesized according to literature preparations (Osborn et al., 1966; Ahmad et al., 1974) from RhCl3‚3H2O obtained from Engelhard and recrystallized PPh3 from Aldrich. Oxygen-free hydrogen with a purity of 99.99% was obtained from Linde-Union Carbide Canada Ltd. All reported kinetic and selectivity experiments employed monochlorobenzene as a solvent. Limited trials with 2-butanone and cyclohexanone were undertaken for comparison to other literature reports. In all cases these reagent grade solvents were used as received. The acrylonitrile-butadiene copolymer (NBR) contained 62% butadiene by weight (78% trans, 12% cis, 10% vinyl isomerization) and had a molecular weight of approximately 200 000. This rubber (Krynac 38.50) was used as received from Bayer Rubber Inc. A styrenebutadiene copolymer (Solprene 308 from Shell) having © 1996 American Chemical Society

4418 Ind. Eng. Chem. Res., Vol. 35, No. 12, 1996

Figure 2. Representative olefin conversion profiles for the hydrogenation of NBR. [Rh]T ) 80 µM; [CN] ) 172 mM; [PPh3] ) 4.0 mM; PH2 ) 23.7 bar; T ) 433.2 K. b RhCl(PPh3)3, [ RhH(PPh3)4, s regression.

an 18% styrene content and a molecular weight of 160 000 was used in control experiments designed to monitor the response of the process in the absence of nitrile functionality. Raw kinetic data consisted of time-resolved measurements of the amount of hydrogen consumed by the reaction. The hydrogenation apparatus, a high-pressure variation of that developed by Mohammadi and Rempel (1987a), successfully maintained isothermal and isobaric conditions throughout the hydrogenation. Polymer solutions were prepared in the dark under an argon atmosphere and vigorously degassed in the autoclave before the introduction of catalyst. Once the experimental conditions were established, microcrystalline RhCl(PPh3)3 or RhH(PPh3)4 was dispersed in the solution using an agitation rate of 1200 rpm. Each experiment proceeded until gas consumption ceased, after which the reactor was quickly cooled and the product isolated by precipitation with ethanol and drying under vacuum. The final degree of olefin conversion measured by gas uptake was confirmed by infrared analysis (Marshall et al., 1990). Spectra of solvent cast films were collected on a Nicolet 520 FT-IR spectrophotometer. Select samples were analyzed by both 1H and 13C{1H} NMR on a Bruker 200 MHz spectrometer. Viscosities of dilute solutions (1 g of HNBR/100 mL of 2-butanone) of fully saturated HNBR were measured at 35 °C using a Ubellohde capillary viscometer. The data are reported as the viscosity relative to pure solvent. The solubility of hydrogen in chlorobenzene/NBR solutions was measured over the range of pressures and temperatures employed in the study (Parent and Rempel, 1996). Results and Discussion Selectivity of NBR Hydrogenation. Over the range of process conditions studied, both RhCl(PPh3)3 and RhH(PPh3)4 functioned as efficient catalyst systems for the quantitative hydrogenation of NBR in chlorobenzene. Representative hydrogen uptake profiles corresponding to the saturation of olefin (see Figure 1) are presented in Figure 2. That the reaction profiles are well represented by a simple first-order regression model would indicate that neither system is appreciably selective to the cis/trans/vinyl butadiene isomerization of the copolymer under the reaction conditions employed. This is in contrast to behavior observed under mild reaction

conditions where reduction of terminal unsaturation is strongly favored (Mohammadi and Rempel, 1987b; Hjortkjaer, 1974). The influence of high reaction temperatures and pressures on the behavior of RhCl(PPh3)3 is also noted in its response to using ketone solvents. At 100 °C and 60 atm, Oppelt et al. (1976) report that 2-butanone and cyclohexanone promote the selective reduction of vinyl versus internal butadiene isomers. Under the conditions employed in this study (145 °C, 23 bar), no greater than 5% conversion was achieved with either solvent. In each case, an uncharacterized black precipitate was noted to be dispersed in the reaction solution. Mohammadi and Rempel (1987b) report no such effect at 65 °C, suggesting that catalyst deactivation in these solvents is promoted by excessive reaction temperatures. Infrared and NMR analysis of the hydrogenated product revealed that no detectable reduction of nitrile unsaturation results from the use of either catalyst. The 1H and 13C {1H} spectra of select HNBR samples were consistent with those of Mohammadi and Rempel (1987b) and Bhattacharjee et al. (1991). This would suggest that the oil resistance of the material is not compromised by the hydrogenation process. However, trace levels of nitrile hydrogenation, well below the detection threshold of spectroscopic analysis, can lead to levels of polymer cross-linking that adversely effect the processability of the product (McManus and Rempel, 1995). Therefore, an assessment of the amount of this cross-linking is also required as a measure of the overall impact of the hydrogenation process. While difficult to measure directly, the degree of cross-linking induced by nitrile reduction may be inferred from measurements of the dilute solution viscosity of the product. The utility of the technique lies in its sensitivity to changes in molecular weight that result from the combination of two or more polymeric molecules. Changes in the solubility of the polymer make the technique sensitive to the degree of olefin conversion as well. However, comparisons between fully saturated HNBR samples provide a means of assessing the selectivity of the reaction over a range of process conditions. The relative viscosities (ηrel) reported in Table 2a,b show no statistically significant difference between RhCl(PPh3)3 and RhH(PPh3)4 systems. The lack of any systematic variation in ηrel suggests that irrespective of process conditions, the product attributes are uniform. This favorable attribute of the rhodium catalysts distinguishes them from ruthenium-based systems that are known to promote gelation (McManus and Rempel, 1991). It has implications not only for process design and operation but also, as will be demonstrated later, for the coordination chemistry of nitrile to catalytically active rhodium complexes. Kinetics of NBR Hydrogenation. The conversion profiles presented in Figure 2, along with all those observed in this study, adhere to a first-order rate model with respect to olefin according to eq 1. Regression

-

d[CdC] ) k′[CdC] dt

(1)

estimates of k′, the pseudo-first-order rate constant, therefore summarize the essential information regarding the hydrogenation rate. The functional relationship between the rate of hydrogenation and [H2], [Rh]T, [CtN], [PPh3], and temperature has been explored by measuring the response of k′ to specific combinations

Ind. Eng. Chem. Res., Vol. 35, No. 12, 1996 4419 Table 2. Kinetic Results expt [Rh]T, [CN], [PPh3], no. µM mM mM

PH2, bar

[H2], mM

temp, k′ × K 103, s-1 ηrel

38 39 40 SRBa SBRb

200.4 19.6 49.9 110.4 140.2 80.3 79.6 80.4 80.2 79.9 80.4 80.6 80.0 80.5 80.3 79.8 80.5 80.4 80.2 80.5 79.8 79.7 80.4 80.3 80.0 79.6 80.4 80.2 80.4 80.1 80.2 80.1 79.9 79.9 80.5 80.2 80.5 80.0 79.9 80.4 80.0 80.1

(a) RhCl(PPh3)3/NBR System 150.6 7.60 54.5 190 373.2 171.8 1.00 23.7 101 418.2 172.1 2.50 23.7 101 418.2 171.2 5.51 23.7 101 418.2 171.6 7.01 23.7 101 418.2 47.6 4.00 23.7 101 418.2 85.8 4.00 23.7 101 418.2 116.0 4.00 23.7 101 418.2 190.0 4.00 23.7 101 418.2 255.4 4.00 23.7 101 418.2 172.0 0.02 23.7 101 418.2 171.6 0.94 23.7 101 418.2 171.4 1.60 23.7 101 418.2 171.8 2.40 23.7 101 418.2 171.6 5.60 23.7 101 418.2 85.7 2.41 23.7 101 418.2 85.7 5.59 23.7 101 418.2 172.1 3.98 4.85 20.6 418.2 172.0 4.02 8.28 35.2 418.2 172.0 4.00 13.4 57.1 418.2 171.5 4.00 33.9 144 418.2 171.4 4.01 46.9 199 418.2 171.8 4.01 76.2 324 418.2 171.8 4.01 23.7 101 418.2 171.4 4.00 23.7 101 418.2 171.8 4.01 23.7 101 418.2 171.8 4.00 23.7 98.7 403.2 171.8 4.07 23.7 102 423.2 171.8 4.00 23.7 103 433.2 171.7 4.00 23.7 105 443.2 131.0 2.81 15.1 64.4 418.2 212.3 5.21 15.1 64.4 418.2 212.4 5.19 15.1 64.4 418.2 131.0 5.22 15.1 64.4 418.2 212.3 2.81 15.1 64.4 418.2 130.9 2.81 32.3 137 418.2 130.9 5.21 32.3 137 418.2 131.2 5.20 32.3 137 418.2 212.2 2.80 32.3 137 418.2 212.5 5.20 32.3 137 418.2 4.02 23.7 101 418.2 4.01 23.7 101 418.2

0.95 0.79 2.42 5.14 6.67 7.92 6.03 4.87 3.42 2.64 4.29 4.44 4.17 4.11 3.47 7.12 6.07 1.05 1.87 2.96 4.75 5.78 6.34 3.59 3.85 3.56 1.38 4.95 7.31 10.2 3.58 2.49 2.37 3.32 2.67 6.03 5.38 5.29 4.30 3.96 4.96 4.93

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

50.4 110.0 138.4 80.3 79.9 79.8 80.3 80.0 79.9 80.2 80.4 79.7 80.3 79.9 79.6 80.0 80.4 80.3 80.4

(b) RhH(PPh3)4/NBR System 172.0 2.51 23.7 101 418.2 171.9 5.49 23.7 101 418.2 171.5 7.01 23.7 101 418.2 172.0 4.00 4.90 20.6 418.2 171.6 4.01 11.7 49.8 418.2 171.5 4.01 174 418.2 4.49 191.9 4.01 68.9 293 418.2 172.0 0.80 23.7 101 418.2 172.0 2.41 23.7 101 418.2 172.0 5.61 23.7 101 418.2 47.7 4.01 23.7 101 418.2 85.5 4.00 23.7 101 418.2 249.2 4.01 23.7 101 418.2 171.9 4.01 23.7 101 418.2 171.8 4.00 23.7 101 418.2 171.5 4.00 23.7 101 418.2 171.6 4.00 23.7 96.4 388.2 171.7 4.01 23.7 98.7 403.2 172.0 4.01 23.7 103 433.2

1.81 4.04 5.27 1.01 2.27 3.52 4.95 3.53 3.41 2.85 4.66 3.46 2.21 3.06 2.81 2.93 0.74 1.86 4.92

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

a

3.54 3.71 3.66 3.74 3.75 3.57 3.93 3.61 3.82 3.67 3.54 3.35 3.59 3.48 3.67 3.52 3.65

3.52 3.67 3.65 3.63 3.48 3.63 3.56 3.36 3.59 3.48 3.64

28.3 g/L. b 14.2 g/L.

of factor levels. For the RhCl(PPh3)3 system, these combinations were assigned on the basis of a central composite design structure (Box et al., 1978). This consists of a two-level factorial design coupled with univariate experiments which inspect the influence of a single factor. Together, these experiments provide the most efficient means of studying the influence of factors acting alone or in combination. The univariate factor combinations have been duplicated for RhH(PPh3)3 to

Figure 3. Influence of catalyst loading on the hydrogenation rate. [CN] ) 172 mM; [PPh3] ) 4.0 mM; PH2 ) 23.7 bar; T ) 418.2 K. b RhCl(PPh3)3, [ RhH(PPh3)4, s model.

gain a rudimentary understanding of the kinetic behavior of this system. The rate constants observed over the specified range of process conditions are contained in Table 2a for RhCl(PPh3)3 and Table 2b for RhH(PPh3)4. On the basis of experiment 1 of Table 2a, it would appear that the hydrogenation rates reported in this work are in conflict with those cited by Bhattacharjee et al. (1991). While both studies confirm a first-order olefin dependence for RhCl(PPh3)3, the cited data yield a rate constant nearly 1/10th of that observed in the present study. Such a rate corresponds to a reaction time of 11 h to reach 99% conversion, whereas only 1.3 h was required using our experimental technique. Although a reliable rate constant cannot be derived from their limited data, the patent granted to Oppelt et al. (1976) supports the assertion that the rates reported by Bhattacharjee et al. (1991) are not representative of the system. Operating at 100 °C and 60 atm, they reportedly achieved 100% conversion within 4.5 h. Barring any misquotation of the amount of catalyst used by Bhattacharjee et al. (1991), it is possible that the addition of RhCl(PPh3)3 to the polymer solution prior to the purging of oxygen could have had a deleterious effect on the catalytic activity (Baird et al., 1996; Strohmeier, 1977). The univariate experiments designed to study the influence of each process factor acting alone are illustrated in Figures 3-7. Interestingly, the behavior of RhH(PPh3)4 paralleled that of RhCl(PPh3)3 in each case, though at a lesser rate. Figure 3 demonstrates the linear response of k′ with changes in the concentration of catalyst precursor ([Rh]T). Note that each experiment of this series was carried out under a constant Rh:PPh3 ratio to safeguard the stability of the catalyst. In spite of varying [PPh3], a first-order dependence holds throughout the data set. With respect to [H2], both systems exhibit a first- to zero-order dependence as the system pressure is increased (Figure 4). The conversion from PH2 to [H2] was accomplished by direct measurement of the solubility of hydrogen in chlorobenzene/NBR solutions (Parent and Rempel, 1996). The extent of the influence of phosphine dissociation equilibria on the hydrogenation rate has been determined by varying the amount of added PPh3 from a PPh3:[Rh]T ratio of 10:1 to 70:1. The data presented in Figure 5 suggest that at the reaction conditions employed, free PPh3 has a marginal effect on hydrogenation activity. It is therefore likely that equilibria

4420 Ind. Eng. Chem. Res., Vol. 35, No. 12, 1996

Figure 4. Influence of hydrogen pressure on the hydrogenation rate. [Rh]T ) 80 µM; [CN] ) 172 mM; [PPh3]d 4.0 mM; T ) 418.2 K. b RhCl(PPh3)3, [ RhH(PPh3)4, s model.

Figure 6. Influence of nitrile unsaturation on the hydrogenation rate. [Rh]T ) 80 µM; [PPh3] ) 4.0 mM; PH2 ) 23.7 bar; T ) 418.2 K. b RhCl(PPh3)3, [ RhH(PPh3)4, s model.

Figure 5. Influence of added PPh3 on the hydrogenation rate. [Rh]T ) 80 µM; 86 mM; PH2 ) 23.7 bar; T ) 418.2 K. b RhCl(PPh3)3, [CN] ) 172 mM; O RhCl(PPh3)3, [CN] ) 86 mM; [ RhH(PPh3)4, [CN] ) 172 mM; s model.

Figure 7. Arrhenius plot for the hydrogenation of NBR. [Rh]T ) 80 µM; [CN] ) 172 mM; [PPh3] ) 4.0 mM; PH2 ) 23.7 bar. b RhCl(PPh3)3, [ RhH(PPh3)4, s regression.

between unbound PPh3 and catalytic species favor the dissociated products. Our experience has shown that although both catalysts are capable of quantitative hydrogenation, their behavior becomes increasingly erratic below a [PPh3]:[Rh]T ratio of less than 20:1. The inhibiting effect of nitrile unsaturation on the olefin hydrogenation rate has been identified by Mohammadi and Rempel (1987b). Figure 6 illustrates this influence as the polymer loading is varied while holding all other variables constant. That the coordination of nitrile to a catalytic intermediate is responsible is supported by an absence of the effect for styrenebutadiene rubber (Table 2a). A similar inhibition by nitrile is reported by Schrock and Osborn (1976) who identified a strong coordination of acetonitrile as being responsible for lessening the catalytic activity of a cationic rhodium complex. At the temperature and pressures used in this work, identifying the influential modes of nitrile complexation is difficult. However, the solution viscosity data suggest that although nitrile coordinates to species involved in olefin hydrogenation, it is not reduced concurrently. Any Rh-CN complex must therefore lack a hydride ligand (in the case of a chloro-rhodium complex) or not favor insertion of nitrile into the metal-hydride bond. The Arrhenius plot provided in Figure 7 illustrates the influence of temperature on the rate of hydrogenation. Over the range of 130-170 °C for RhCl(PPh3)3 and

115-160 °C for RhH(PPh3)4, a linear response is observed, from which apparent activation energies of 73.5 kJ/mol for the former and 57.4 kJ/mol for RhH(PPh3)4 are derived. These estimates show that the experiments were carried out without mass transfer limitation, else the reaction rates would be relatively insensitive to temperature variations. For RhCl(PPh3)3, Bhattacharjee et al. (1991) report a value of 22 kJ/mol without providing sufficient data to rationalize the discrepancy. Nevertheless, the values derived from this study are more consistent with homogeneous catalytic processes. A thorough kinetic study improves our knowledge not only of how each factor effects k′ alone (as probed by the univariate experiments) but also of whether or not factors act in combination. The 23 factorial component of the experimental design (experiments 31-40, Table 2a) provides a means of assessing the significance of such joint factor interactions. The results of an analysis of variance listed in Table 3, prove that in addition to strong main group influences ([H2], [CtN], and [PPh3]), a highly significant [H2]*[CtN] interaction operates within the kinetic mechanism. The influence of [H2]*[PPh3] is somewhat weaker, as may be expected from the univariate studies discussed earlier. The existence of other two- and three-factor interactions is not supported and may be dismissed from a RhCl(PPh3)3 reaction pathway. Catalytic Pathways of the RhCl(PPh3)3 and RhH(PPh3)4 Systems. The catalytic chemistry of

Ind. Eng. Chem. Res., Vol. 35, No. 12, 1996 4421 Table 3. 23 Factorial Design; Analysis of Variance source

sum-ofsquares

DF

mean square

[H2] [CN] [PPh3] [H2]*[CN] [H2]*[PPh3] [CN]*[PPh3] [H2]*[CN]*[PPh3]

8.30 × 10-6 3.45 × 10-6 3.37 × 10-7 2.43 × 10-7 4.09 × 10-7 2.01 × 10-7 1.60 × 10-7

1 1 1 1 1 1 1

8.30 × 10-6 1476 3.44 × 10-6 611.0 3.37 × 10-7 59.8 2.43 × 10-7 43.2 4.09 × 10-7 7.27 2.01 × 10-7 3.57 1.60 × 10-7 2.85

error

1.12 × 10-8

2

5.62 × 10-9

F-ratio

P 0.0007 0.002 0.016 0.022 0.114 0.199 0.233

Rh(I) phosphine complexes has been extensively researched, resulting in a greater understanding of trace intermediates that bring about the observed kinetic behavior (Jardine, 1981; James, 1973). While an extrapolation of this knowledge to severe reaction conditions may not be straightforward, it is proposed that strong parallels exist between the chemistry underlying this work and that documented at mild reaction conditions. Figure 8 illustrates a catalytic mechanism that is consistent both with the kinetic data of Table 2a and our understanding of the coordination chemistry of RhCl(PPh3)3 in solution. That RhCl(PPh3)3 oxidatively adds molecular hydrogen to form the five-coordinate dihydride according to eq 2 is well established (Halpern, 1973). K

Figure 8. Proposed reaction mechanism for the RhCl(PPh3)3/NBR system.

As nitrile likely coordinates by σ donation of its lone pair, it experiences little of the steric hinderance effecting the coordination of olefin. It may therefore compete effectively with olefin for coordination to unsaturated complexes. Equations 5 and 6 illustrate two possible modes of nitrile coordination that could be capable of bringing about the observed inhibitory behavior. K2

RhClH2(PPh3)2 + CtN y\z RhClH2(CtN)(PPh3)2 (5) K5

(2)

y z RhCl(CtN)(PPh3)2 (6) RhCl(PPh3)2 + CtN \

Mohammadi and Rempel (1987b) provide evidence that at 65 °C under 1 bar of H2 the reaction is quantitative toward formation of the dihydride. The comparative ease of hydrogen activation compared to addition of the sterically hindered olefin favors the “hydride pathway” shown in the catalytic mechanism. The dissociation of phosphine from RhClH2(PPh3)3 according to eq 3, while not appreciable at room temperature, is likely to be encouraged by the temperatures used in this study.

Having few means to probe the nitrile coordination chemistry at the conditions employed, the relative importance of these reactions may only be assessed using the kinetic data. Statistical measures of the agreement between the observed data and derived catalytic mechanisms, while not proving the validity of a model, may aid in the discrimination between various proposals. When applied to a rate expression based solely on eq 5, an analysis of variance and residuals proved the simple model ill-equipped to account for the degree of nitrile inhibition. While an expression founded on eq 6 alone proved to be satisfactory, the propensity of RhCl(CN)(PPh3)2 to oxidatively add H2 (Ohtani et al., 1979) would suggest an overall mechanism as illustrated in Figure 8. Provided eq 4 adequately represents the rate-determining step, eq 1 may be written in the form of eq 7.

RhCl(PPh3)3 + H2 y\z RhClH2(PPh3)3

K1

RhClH2(PPh3)3 y\z RhClH2(PPh3)2 + PPh3

(3)

As the coordination of olefin to RhClH2(PPh3)2 is not facile, a direct assignment of the rate-determining step as the insertion of olefin into the Rh-H bond is not warranted. It is equally probable that the initial coordination of the substrate is rate limiting in this case. Therefore, we have chosen to represent the ratedetermining step with eq 4. k4

RhClH2(PPh3)2 + CdC 98 RhClH2(CdC)(PPh3)2 (4) Given that coordinatively unsaturated complexes are often presumed to associate with solvent, the fact that a potential ligand such as nitrile may inhibit the hydrogenation cycle is not surprising. [Rh(CO)(MeCN)(PPh3)2]+ClO4- has been prepared (Booth et al., 1976), and [Rh(PPh3)3(MeCN)][BF4] has been analyzed crystallographically by Pimblett et al. (1985). As cited earlier, Schrock and Osborn (1976) report that the use of acetonitrile as a solvent has a deleterious effect on the hydrogenation activity of [Rh(diene)(PPh3)2]+A-. To date a detailed study of the propensity of nitrile to associate with complexes derived from RhCl(PPh3)3 is lacking, although Ohtani et al. (1979) have presented some spectrophotometric data on the system.

-

d[CdC] ) k4[RhClH2(PPh3)2][CdC] dt

(7)

The conversion of [RhClH2(PPh3)2] to the total amount of rhodium [Rh]T used in each experiment may be accomplished using the mass balance of eq 8.

[Rh]T ) [RhClH2(PPh3)3] + [RhClH2(PPh3)2] + [RhCl(PPh3)2] + [RhCl(CtN)(PPh3)2] + [RhClH2(CtN)(PPh3)2] (8) Applying the equilibrium relations defined in Figure 8, the concentration of RhClH2(PPh3)2 may be substituted into eq 7 to provide the functional relationship between k′ and the process factors studied (eq 9). k′ ) k4K1K3[Rh]T[H2] K1 + K1K3[H2] + K3[H2][PPh3] + K1K5[CtN] + K1K2K3[H2][CtN]

(9)

4422 Ind. Eng. Chem. Res., Vol. 35, No. 12, 1996

Conclusions

Table 4. Model Analysis of Variance Results source

sum-of-squares

DF

mean-square

regression residual

7.092 × 10-4 1.512 × 10-6

5 30

1.418 × 10-4 5.042 × 10-8

total corrected

7.107 × 10-4 9.588 × 10-5

35 34

Table 5. Model Parameter Estimates 〈95%〉 parameter s)-1

k4, (mM K1, mM K2, mM-1 K3, mM-1 K5, mM-1 K6, mM-1

estimate

ASE

lower

upper

1.19 1.44 3.98 × 10-2 3.41 × 10-3 2.71 × 10-2 6.28 × 10-3

0.17 0.38 5.94 × 10-2 2.06 × 10-4 3.20 × 10-3 3.78 × 10-4

0.85 0.67 2.76 × 10-2 2.99 × 10-3 8.12 × 10-3 5.51 × 10-3

1.53 2.21 5.19 × 10-2 3.83 × 10-3 4.16 × 10-2 7.05 × 10-3

Both RhCl(PPh3)3 and RhH(PPh3)4 function as efficient catalysts systems for the selective hydrogenation of acrylonitrile-butadiene copolymer under severe reaction conditions. Observed influences of process conditions on the rate of hydrogenation are consistent with behavior reported at mild temperatures and pressures near ambient. This knowledge has led to a straightforward extrapolation of a mild reaction condition mechanism for the RhCl(PPh3)3 system that is supported by rate and selectivity data. Uniform product HNBR sample viscosities suggest that while the copolymer’s nitrile functionality undoubtedly complexes with coordinatively unsaturated rhodium species, this coordination does not lead to detectable levels of nitrile reduction. Acknowledgment Support from the Natural Science and Engineering Research Council (NSERC) and funding from the Ontario Centre for Materials Research to J.S.P. is gratefully acknowledged. Literature Cited

Figure 9. Residual plot of (k′actual - k′model) versus k′.

The results of an analysis of variance summarized by Table 4 suggest that the derived rate expression complies with the observed hydrogenation kinetics over the range of process conditions studied. Although this result is insufficient to definitively certify the model, it indicates that its fit is superior to all other options explored. The parameters derived from the rate equation are provided in Table 5 along with their associated error estimates. The residual plot shown in Figure 9 demonstrates none of the systematic patterns produced by other possible kinetic pathways. Actual model predictions relative to the experimental data are plotted in Figures 3-6. In contrast to RhCl(PPh3)3, the RhH(PPh3)4 system has received relatively little attention. The results of this investigation would suggest that the underlying chemistry of the two systems is similar. Indeed, a catalytic mechanism derived using RhH3(PPh3)2 as the active species is equipped to account for the kinetic data. However, ascertaining whether coordination of hydrogen precedes or follows the addition of olefin is less straightforward than for RhCl(PPh3)3, while the use of model discrimination techniques is limited by the similarity of rate expressions derived from either assumption. Therefore, given the complexity of the system, an assignment of a reaction mechanism is not warranted on the basis of kinetic data alone. Detailed research directed at identifying important reaction intermediates is required before such a process may be unequivocally assigned.

Ahmad, N.; Levison, J. J.; Robinson, S. D.; Uttley, M. F. Triphenylphosphine ComplexessHydridotetrakis(triphenylphosphine)rhodium(I). Inorg. Synth. 1974, 58-59. Baird, M. C.; Lawson, D. N.; Mague, J. T.; Osborn, J. A.; Wilkinson, G. Novel Addition Reactions of Chlorotris(triphenyl phosphine)rhodium(I). J. Chem. Soc., Chem. Commun. 1966, 129-130. Bhattacharjee, S.; Bhowmick, A. K.; Avasthi, B. N. High Pressure Hydrogenation of Nitrile Rubber: Thermodynamics and Kinetics. Ind. Eng. Chem. Res. 1991, 30, 1086-1092. Booth, B. L.; Haszeldine, R. N.; Holmes, R. G. G. Transition Metal Complexes Containing the Bis(trifluoromethyl)nitroxy Ligand. J. Chem. Soc., Chem. Commun. 1976, 13, 489-490. Box, G. E.; Hunter, W. G.; Hunter, J. S. Statistics for Experimenters; Wiley: New York, 1978. Halpern, J. Hydrogenation of Tris(triphenylphosphine)chlororhodium(I). J. Chem. Soc., Commun. 1973, 629-630. Hashimoto, K.; Watanabe, N.; Yoshioka, A. Highly Saturated NitrilesA New High temperature, chemical resistant elastomer Rubber World 1984, 190 (2), 32-47. Hjortkjaer, J. The Hydrogenation of 1-Hexene Catalyzed by Solutions of Hydridotetrakis(Triphenylphosphine)Rhodium(I). Adv. Chem. Ser. 1974, 132, 133-144. James, B. R. Homogeneous Hydrogenation; Wiley: New York, 1973. Jardine, F. Chlorotris(triphenylphosphine)rhodium(I): Its Chemical and Catalytic Reactions. Prog. Inorg. Chem. 1981, 29, 63202. Marshall, A. J.; Jobe, I. R.; Dee, T.; Taylor, C. Determination of the Degree of Hydrogenation in Hydrogenated Nitrile-Butadiene Rubber (HNBR). Rubber Chem. Technol. 1990, 63, 244-255. McGrath, M. P.; Sall, E. D.; Tremont, S. J. Functionalization of Polymers by Metal-Mediated Processes. Chem. Rev. 1995, 95, 381-398. McManus, N. T.; Rempel, G. L. U.S. Patent 5 075 388, 1991. McManus, N. T.; Rempel, G. L. Chemical Modification of Polymers. Catalytic Hydrogenation and Related Reactions. J. Macromol. Sci. Rev. Macromol. Chem. Phys. 1995, C35, 239-285. Mohammadi, N. A.; Rempel, G. L. Control, Data Acquisition and Analysis of Catalytic Gas-Liquid Mini Slurry Reactors using a Personal Computer. Comput. Chem. Eng. 1987a, 11, 27-35. Mohammadi, N. A.; Rempel, G. L. Homogeneous Selective Catalytic Hydrogenation of CdC in Acrylonitrile-Butadiene Rubber. Macromolecules 1987b, 20, 2362-2368. Ohtani, Y.; Yamagishi, A.; Fujimoto, M. Substitution Reactions of (Acetonitrile)chlorobis(triphenylphosphine)rhodium(I) in Benzene. Bull. Chem. Soc. Jpn. 1979, 52, 2149-2150. Oppelt, D.; Schuster, H.; Thormer, J.; Braden, R. Brit. Patent 1 558 491, 1976.

Ind. Eng. Chem. Res., Vol. 35, No. 12, 1996 4423 Osborn, J. A.; Jardine, F. H.; Young, J. F.; Wilkinson, G. The Preparation and Properties of Tris(triphenylphosphine(halogenorhodium(I). J. Chem. Soc. A. 1966, 1711-1732. Parent, J. S.; Rempel, G. L. Solubility of Hydrogen in Chlorobenzene. J. Chem. Eng. Data, 1996, 41, 192-194. Pimblett, B.; Garner, C. D.; Clegg, W. Isolation and Crystal Structure of [Rh(PPh3)3(MeCN)][BF4], Acetonitriletris(triphenyl phosphine)rhodium(I) Tetrafluoroborate. J. Chem. Soc., Dalton Trans. 1985, 1977-1980. Rempel, G. L.; Azizan, H. U.S. Patent 4 464 515, 1984. Schrock, R. R.; Osborn, J. A. Catalytic Hydrogenation Using Cationic Rhodium Complexes. I. Evolution of the Catalytic System and the Hydrogenation of Olefins. J. Am. Chem. Soc. 1976, 98 (8), 2134-2143.

Strohmeier, W.; Hitzel, E.; Kraft, B. Comparison Between Homogeneous Catalysts and their Heterogenised Counterparts. J. Mol. Catal. 1977, 3, 61-69. Weinstein, A. H.; Elastomeric Tetramethylene-EthylethyleneAcrylonitrile Copolymers. Rubber Chem. Technol. 1984, 57, 203-215.

Received for review November 6, 1995 Revised manuscript received June 14, 1996 Accepted August 19, 1996X IE9506680 X Abstract published in Advance ACS Abstracts, November 1, 1996.