Preparation of the target palladium alkyl complex proceeds first by reaction of an acetonitrile solution of the bis(phosphino)borate ligand, [ASN][Ph2BP2] (ASN = 5- azonia-spiro[4.4]nonane), with a benzene solution of (tmeda)PdMe2 (tmeda = N,N,N’,N’- tetramethylethylenediamine). The resulting anionic product, [ASN][(Ph2BP2)PdMe2] (A.1), is then protonated by the ammonium salt [HNiPr2Et][BPh4] to cleanly generate the target zwitterion, [Ph2BP2]PdMe(THF) (A.2), as a light peach solid (Scheme A.1).
With the solvento adduct A.2 in hand, we examined the scope of its reactivity under CO and ethylene gas. A light yellow solution results when a dilute THF solution (ca. 10 μM) of A.2 is exposed to a CO atmosphere. The unstable product, [Ph2BP2]Pd(C(O)Me)(CO) (A.3), is related to a cationic derivative, [(dppp)Pd(C(O)Me)(CO)][B(3,5-(CF3)2-C6H3)4], recently reported by Brookhart and coworkers.11,12 At 30 psi of CO and 30 psi of ethylene, an appreciable quantity of strictly alternating polyketone material rapidly precipitates from the yellow solution (established by 13C NMR and MALDI-TOF). The activity of catalyst A.2 was examined at higher pressures (200 psi, vide infra), and calculated TON values (kg polymer mol-1 catalyst h-1) establish that it is a very active catalyst for polyketone production at room temperature (Table A.1).
Table A.1. Copolymerization results for catalysts [Ph2BP2]PdMe(THF), A.2;
[(dppp)Pd(Me)(THF)][B(C6F5)4], A.7; and [(PhSiP2)Pd(Me)(THF)][B(C6F5)4], A.8.a Catalyst activity (TON)b yield (g polymer)c Mw (103)d Mn (103)d Mw/Mn
A.2 39 ± 1 0.36 ± 0.01 138 ± 1 112 ± 3 1.3
A.7 35 ± 2 0.32 ± 0.02 130 ± 1 99 ± 2 1.3
A.8 45 ± 1 0.42 ± 0.01 190 ± 1 130 ± 5 1.5
a Conditions: 9.3 x 10-6 mol Pd catalyst in 10 mL THF; 100 psi CO; 100 psi ethylene; 23 ºC; 1 h. b TON values are expressed as kg polymer per mol catalyst per h. c Average mass of polymer obtained from eight independent runs. d Determined by GPC using polystyrene standards for calibration (1,3-cresol, 1 mL/min, 120 ºC, duplicate runs).
Figure A.2. 50% thermal ellipsoid representation of {[Ph2BP2]Pd}2 (A.4).
Hydrogen atoms and three CH2Cl2 molecules have been omitted for clarity.
Selected bond distances (Å) and angles (deg) for A.4: Pd1–Pd2 2.7281(6), Pd1–
P1 2.331(1), Pd1–P2 2.232(1), Pd1–P4 2.695(1), Pd2–P3 2.351(1), Pd2–P4 2.235(1), Pd2–P2 2.781(1), P1–Pd1–P2 92.83(5), P3–Pd2–P4 92.33(5).
Interestingly, the reactivity of catalyst A.2 with CO gas, in the absence of ethylene, is concentration-dependent. At high concentrations of A.2 (ca. 0.01M), exposure to an atmosphere of CO produces an orange solution (31P NMR, singlet at 22 ppm),13 which turns red (two doublets at 11 and 30 ppm) upon replacing the CO atmosphere with dinitrogen. The final red product is accessible by other routes (vide infra), and its crystallographic characterization establishes it to be the dimeric palladium(I) complex, {[Ph2BP2]Pd}2 (A.4),14,15 shown in Figure A.2 and Scheme A.2.
An alternative method for cleanly generating dimeric A.4 is the exposure of A.2 to ethylene gas in the absence of CO. In this case, the color change to red is rapid.
Monitoring this transformation under ethylene at low temperature (from -78 °C to -10 °C) establishes one observable intermediate, assigned as an insertion product, [Ph2BP2]Pd(C2H5)(H2C=CH2) (A.5). A related cationic species, [(dppp)Pd(C2H5)(H2C=CH2)][B(3,5-(CF3)2-C6H3)4], has also been reported by Brookhart and coworkers. Storage of intermediate A.5 at -10 °C under excess ethylene effects the catalytic production of butenes. On warming, the only new species detectable by 31P NMR is the red dimer A.4. The expected product of β-hydride elimination, [Ph2BP2]Pd(H)(L) (6) (L = THF or ethylene), was not observed. Furthermore, an independent attempt to generate the hydride A.6 by addition of H2 to A.2, while generating dimeric A.4 quantitatively, offered no evidence for a detectable hydride intermediate. Considered collectively, the data suggest that the conversion of A.2 to A.4 under ethylene may occur as follows (Scheme A.2): initial ethylene insertion and rapid β-hydride elimination, followed by a second ethylene insertion, generates the observable intermediate A.5. Intermediate A.5 can undergo further chemistry with ethylene to
generate butenes. At higher temperatures, a bimolecular path competes in which the unobservable hydride A.6, generated by a β-hydride elimination step, reacts with a palladium alkyl, such as A.5, to produce alkane and the dimeric palladium(I) species A.4.
Notably, gas analysis of a reaction mixture from the conversion of A.2 to A.4 under ethylene showed no evidence for hydrogen production. Bimolecular loss of H2 from a hydride intermediate such as A.6 does not occur.
Having established some of the comparative reaction chemistry between charge neutral A.2 and its cationic dppp counterpart, [(dppp)Pd(Me)(solv)][B(C6F5)4], we sought to compare their relative activities with respect to the CO/ethylene copolymerization reaction of interest. The THF adduct complex, [(dppp)Pd(Me)(THF)][B(C6F5)4] (A.7), was thus prepared, and its copolymerization activity was measured (Table A.1). To our
surprise, under analogous conditions, zwitterionic A.2 proved to be a slightly better copolymerization catalyst from the perspective of total amount of polyketone produced per unit time (Table A.1). To examine whether the slightly increased activity of the [Ph2BP2] system was perhaps due to the difference in relative charge between the palladium centers in A.2 and A.7, we sought a second comparison. A cationic complex more structurally related to neutral A.2, [(Ph2SiP2)Pd(Me)(THF)][B(C6F5)4] (A.8), was prepared using the neutral phosphine chelate (Ph2Si(CH2PPh2)2 (abbreviated (Ph2SiP2)).16 Under analogous conditions, cationic A.8 proved to be a slightly better catalyst than zwitterionic A.2 (Table A.1), indicating that the phenyl substituents incorporated within the ligand backbones of A.2 and A.8 might also contribute to slight differences in reactivity by comparison to the dppp system A.7.
Because differences in bulk activity are difficult to interpret, a closer examination of the elementary steps in the polymerization mechanism is desirable. Indeed, Brookhart and coworkers have measured the rates and calculated the activation barriers for migratory insertion reactions in the dppp system, specifically for the complexes:
(dppp)PdMe(CO)+, (dppp)PdC(O)Me(C2H4)+, and (dppp)PdMe(C2H4)+.12a The latter represents an olefin misinsertion step during the polymerization. We attempted to study all three migratory insertions for the zwitterionic analogues, but thus far, only the kinetics of [Ph2BP2]PdMe(C2H4) (A.9) have yielded fruitful results. Compound A.9 is generated in situ by the addition of ethylene to A.2 at low temperatures. The migratory insertion showed no dependence on ethylene concentration. The rate constant at -45.8 ºC, kobs = 1.16 x 10-4 s-1 (with ΔG‡ = 17.3(1) kcal/mol), is ca. four times slower than that observed for the dppp system (kobs = 4.9 x 10-4 s-1 and ΔG‡ = 16.6(1) kcal/mol at -45.6 ºC). The
temperature dependence of the reaction rate was also investigated from -58 to -41 ºC. An Eyring plot of this data is shown in Figure A.3, from which the activation parameters, ΔH‡ = 15.4(4) kcal/mol and ΔS‡ = -8.4(1.9) eu, were extracted. Surprisingly, these parameters are identical within error to those reported for the dppp system, ΔH‡ = 15.2(7) kcal/mol and ΔS‡ = -6.2(2.9) eu.
-17 -16 -15 -14 -13
4.2 4.3 4.4 4.5 4.6 4.7
1000/T
ln(k/T)
Figure A.3. Eyring plot for the migratory insertion reaction in [Ph2BP2]PdMe(C2H4) (A.9), from -58 to -41 ºC, ΔH‡ = 15.4(4) kcal/mol and ΔS‡ = -8.4(1.9) eu.