Transition Metal Catalysed Oxidation Reactions and Ligand Effects in Aprotic Solvents

전이금속 촉매작용의 산화반응과 리간드 효과

Abstract

Cobalt oxygen carrier complex N,N'-ethylenebis(3-methoxysalycylideneiminato)cobalt(II), Co(3MeOsalen) was prepared at $25{\circ}C$. UV and visible absorption spectra of the complex and hydrazobenzene were studied in non-aqueous solvent methanol in the range of wavelength 200-600 nm. The oxidation of hydrazobenzene by oxygen in non-aqueous solvent is catalysed by Co(3MeOsalen). In the presence of triphenylphosphine($PPh_3$), the rate decreases in methanol. This is presumably attributable to the coordination of $PPh_3$ to the Co(3MeOsalen), resulting in the catallytically inactive compound. The initial rates of the oxidation of hydrazobenzene with the ligand triphenylphosphine were measured by the theoretical values of the rates, Rate=$k_1+k_2K_1[P]/1+K_1[P]+K_1K_2[P]^2$. This fact would be a poorer σ-donor ligand than methanol.

코발트 산소 운반체인 N,N''-ethylenebis(3-methoxysalycylideneiminato)cobalt(II), Co(3MeOsalen)을 $25{\circ}C$에서 합성하였다. 이 착물과 하이드라조벤젠의 자외선 및 가시부분 광스펙트럼은 파장 범위 200-600 nm에서 비수용매 메탄올을 사용하여 연구하였다. 하이드라조벤젠의 산소와의 산화반응은 메탄올에서 Co(3MeOsalen) 촉매로 사용하였다. 트라이페닐포스핀($PPh_3$) 존재하에서, 반응속도는 감소하였으며 이는 촉매가 리간드 트라이페닐포스핀과 배위화합된 것으로 추정되며 촉매가 비활성인 Co(3MeOsalen)$(PPh_3)_2$으로 되어 속도가 급격히 떨어지는 것으로 생각된다. 리간드 트라이페닐포스핀과 하이드라조벤젠의 초기산화속도는 이론속도식, Rate=$k_1+k_2K_1[P]/1+K_1[P]+K_1K_2[P]^2$으로 측정되었다. 이것은 리간드가 메탄올 분자보다 더 좋지않은 σ-주게일 것으로 간주된다.

INTRODUCTION

It has been known since the observation of Pfeiffer and his cowokers in 19331 that cobalt(Ⅱ) schiff’s base complexes such as Co(salen), oxygen carrier, form reversible complex with oxygen. There has been considerable recent interest in these compounds because of their realtionship to the natural iron-containing oxygen carrier hemoglobin and myoglobin.2

Hemoglobin and myoglobin consists of an ironporphyrin complex containing the heme group and imidazole group. On oxygenation the sixth coodination site of the iron accepts the dioxygen ligand. Accordingly, the ligands around the iron in oxyhemoglobin are approximagtely octahedral.3

Co(salen) and Co(3MeOsalen) are square planar with a low spin d7 electron configuration, Fig. 1. The binding of an axial, fifth ligand, e.g. triphenylphosphine(PPh3) leads to a ground state with the unpaired eletron in the dz2 orbital. This eletron configuration is a necessary prerequisite for the binding with dioxygen which is called cobalt oxygen carriers.

Fig. 1.Cobalt(Ⅱ)-Schiff base complexes, e.g. N,N’-ethylenebis-(salicylideneiminato)Cobalt(Ⅱ), Co(Ⅱ)(salen), R=H, Co(salen); R=CH3O, Co(3MeOsalen).

Cobalt(Ⅱ) Schiff’s base complexes with added axial ligands have been shown to catalyse the oxidation by oxygen of secondary alcohols5 to ketones, and of phenols6 to quinones. Kinetic studies5,6 have shown that in the oxidation of alchols and phenols the transition state is a ternary complex of dioxygen, cobalt catalyst (including axial base) and substrate. Thus the reaction resemble an enzymecatalysed process in which the two substrate, dioxygen and the organic molecule, are brought together by the catalyst.

The effects of bases coordinated trans to dioxygen have been investigated by measurement of the equilibrium constants.7,8 Pure σ-bonding8 can occur by interaction of metal dz2 orbital with non-bonding electron pairs of the N-ligand (e.g. n-BuNH2). Complexes with strong σ-donor ligands in the axial position, such as primary aliphatic amines form the strongest complexes with dioxgen. This can be attributed to the more stable molecular orbital formed between the oxygen π* orbital and the coblat dz2 orbital when the energy of the latter is raised towards the energy of the π* orbital by the axial ligand.

π-donor ligands will enhance the bonding from cobalt to dioxygen. It follows that imidazoles, which are better π-donors9-11 than pyridines,8 produce an increase in the oxygen binding constants of the dioxygen complex of Co(benacen). The strength of the cobalt-oxygen bond should reflect the extent of back-donation of electron density from the filled 3dxz (or 3dyz) orbital of the cobalt atom into the π* orbital of the oxygen moiety.8 Therefore, a good π-donating axial ligand forces more electron density on to the coblat and enhances the π-donating between cobalt and oxygen. However, π-acceptor ligand12(CO for instance) trans to O2 will decrease π-electron density on the metal, resulting in a weaker Co-O2 bond by lowering the dxz, dyz orbital. Axial ligands such as P(C4H9)3, AsPh3 and S(CH3)2 are considered to have not only σ-donor, but also π-acceptor properties.

In the paper we show that the oxidation of hydrazobenzene catalysed by Co(3MeOsalen) has been studied in the non-aqueous solvent methanol. When a σ- and π-doner ligand, PPh3 was added th the reaction mixture of hydrazobenzene, the effect of added axial ligand will be enabled to be investigated.

 

EXPERIMENTAL

Co(3MeOsalen) was prepared under dry nitrogen according to the method of Diehl et al.13,14 The reaction mixture was prepared by mixing the catalyst solution, O2 saturated methanol, and the substrate solution using a syringe. The catalyst(5.20×10-4 M) was dissolved in methanol by stirring magnetically under dry nitrogen gas. The catalyst solution was injected by syringe into a 10 ml volumetric flask and diluted with O2 saturated methanol to 9.0 ml. The substrate hydrazobenzene(3.33×10-3 M) was easily dissolved in the methanol after passing nitrogen gas continuously and strring magenetically for a fex minuets. The solution was moved with a syringe and put into a 10 ml volumetric flask and diluted under nitrogen to 1.0 ml.

The catalyst solution (9.0 ml) and substrate solution (1.0 ml) were warmed in a themostant for about 15 minuets at 25.0℃. The two solutions were then mixed, shaken, a portion transferred to a UV cell and the initial gradients, (dA/dt)o were measured in the UV-visible spectrophotometer thermostatted at 25.0℃. The time was taken to be zero when the catalyst and substrate were mixed.

 

RESULTS AND DISCUSSION

Initial rates R0 were employed and calculated from the slopes of tangents at time t=0:

where 1 is the path length, Δε the diffrence of the extinction coefficients of product and reaction, (dA/dt)0 initial tangents to the absorbance/time traces.15 The reation was followed with Co(3MeOsalen) (7.43×10-5 M) and the appropriate ligand: triphenylphosphine (1.91×10-2 M). When the initial rate Ro is plotted against the hydrazobenzene concentration, the plots are linear, Fig. 2. We have demonstrated that the autoxidation of hydrazobenzene in anhydrous methanol was found to be first order in air or O2, Table 1. The rate constants of the reaction mixture were calculated to be 3.85×10-6 sec-1 in The O2 saturated methanol and 1.98×10-6 sec-1 in the air saturated methanol.

Fig. 2.Plot of initial rates Ro against[H2AB]. O2(a) and air (b) saturated methanol at 25.0℃, λ=315 and 437nm. The rate constants are k(O2)=3.85×10-6 sec-1 and k(air)=1.98×10-6 sec-1.

Table 1.λ=315 nm(ε315=1.61×104 M-1·Cm-1) (NO. 1-3) and λ=437 nm(ε437=5.09×102 M-1·Cm-1) (NO. 4-8). Δε-1315 = 7.04×10-5 M·cm. R0 = Δε-1315 · (dA/dt)0.

As the ligand triphenylphosphine, Table 2. the initial rate was decreased, Fig. 3, which is in good agreement with the experimental values. The destruction of catalyst system Co(3MeOsalen)(O2) could be due to the catalyst becoming coordination with triphenylphosphine in both Z axis sites of the catalyst.

The ligand could affect dioxygen binding to the catlaytic system Co(3MeOsalen)(O2) as triphenylphosphine is usually considerd to have not only σ-donor but also π-acceptor properties.7 The initial rates, Ro, the oxidation of hydrazobenzene with the ligand triphenylphosphine (PPh3) were measured, Table 2. Triphenylphosphine is an interesting ligand as it decrease the rate of the oxidation of hydrazobenzene in methanol. This may be due to the destruction of the catalytic activity because the catalyst becoming coordinated with triphenylphosphine at both Z axis sites of the complex e.g.Co(3MeOsalen)-(PPh3)2. The seconed triphenylphosphine molecule may prevent oxygen binding to the complex. It seems more likely that it is due to the specific solvent effect methanol through coordination to the cobalt catalyst.

The initial rates, R0, of the oxidation of hydrazobenzene with the ligand triphenylphosphine (PPh3) were measured, Table 2. As the triphenylphosphine concentrations increased, the initial rates were found to decrease. When he initial rates were plotted against the triphenylphosphine concentrations, a curve was formed, Fig. 3. Theroretical values of the rates, Table 3, were calculated from Eq. 2:

Table 2.k=(R0-k0[s][O2])/[S][O2][C]T=1.11×1010(R0=7.51×10-10)M-2·sec-1, R'0=R0-k0[s][O2], k0[S][O2]=7.51×10-10M·sec-1 from Fig. 1.,[S]=[H2AB]=1.95×10-4M,[O2]=9.63×10-3M,[C]T=[Co(3MeOsalen)]=4.78×10-5M, ε315=1.61×104M-1·cm-1,λ=315 nm, k1=1.10×103M-2·sec-1, k2=2.69×102M-2·sec-1,[P]=[PPh3],K1=4.56×104dm3·mol-1, K2=5.39×102dm3·mol-1

Fig. 3.Plots of k against [PPh3]. O2 saturated methanol, 25.0℃, λ=315nm.[Co(3MeOsalen)]=4.78×10-5M, [H2AB]=1.95×10-4M, solid curve: experimental value; dotted curve: theoretical value

where C and P are catalyst and triphenylphosphine. The Eq. 2 was derived from Eq. 3, 4, and 5 when k3=0

Table 3.k=(k1+k2K1[P]/1+K1[P]+K1K2[P]2.[P]=[PPh3]. k1=1.10×103M-2·sec-1,k2=2.69×102M-2·sec-1, K1=4.56×104dm3·mol-1,K2=5.39×102dm3·mol-1.

The rate k0[S][O2] is the autoxidation of hydrazobenzene whose observed value of 7.51×10-10M·sec-1, Fig. 2, is independent of triphenylphosphine concentration. The second term k1[S][O2][C] is the rate of oxidation catalysed by Co(3MeOsalen) without an axial ligand. The third term is the rate of oxidation of hydrazobenzene catalysed by Co(3MeOsalen)(PPh3). The fourth term k3=0 because it is assumed that the complex Co(3MeOsalen)(PPh3)2 is catalytically inactive, the secnd triphenylphosphine molecule preventing oxygen binding to the complex. Eq. 5 is rearranged to Eq. 6.

The total concertration of the catalyst [C]T is given by Eq. 7.

Substituting Eq. 3 and 4 into 7 gives:

Eq. 8 gives ris to

Substituting Eq. 9 into 6 gives Eq. 2. The constants k2, K1, and K2 were determined by reciprocal plots from Eq. 2, and k1 is known to be 1.10×103M-2sec-1 when [P]=0, Table 2.

When [P] is small, Eq. 2 reduces to Eq. 10 as K1[CP]>>K1K2[CP2]2.

Subtracting k1 from both sides,

When P/(k-k1) is plotted against P from Eq. 10, K1 and k2 were obtained graphically, Table 2.

When [P] is large, Eq. 2 reduces to Eq. 11 as k1 ≪ k2K1[CP2]and 1≪K1[P]. 1/k is plotted against P from Eq. 11 and K2 was

determined graphically, Table 2.

As the theortical values of the initial rates agreed faily well with experimental observations, Fig. 3, it is suggested that the catalyst Co(3MeOsalen) was associated with triphenylphosphine which would be a poorer σ donor ligand compared with metanol molecule and the reaction rate therefore decreased. The large K1 value (K1=4.56×104dm3mol-1)means that the equlibrium to Co(3Mesalen)(PPh3) was soon achieved and was ready to associated with triphenylphoslhine and became Co(3MeOsalen)(PPh3)2 which is catalytically inactive and the rates were then expected to decrease quickly