Magnesium chloride supported high-mileage catalysts for olefin polymerization. I. Chemical composition and oxidation states of titanium

The chemical composition of a MgCl₂-supported, high-mileage catalyst has been determined at every stage of its preparation. Ball milling of MgCl₂ with ethyl benzoate (EB) resulted in the incorporation of 95% of the EB present to give MgCl₂·EB_0.15. A mild reaction with a half-mole equivalent of p-cresol (PC) at 50°C for 1 h resulted in near quantitative retention of p-cresol by the support. The composition is now approximately MgCl₂·EB_0.15P?_0.5. Addition of an amount of AlEt₃ corresponding to half-mole equivalent of p-cresol liberated one mole of ethane per mole of p-cresol, thus signaling quantitative reaction between the two components. The support contains on the average one ethyl group per Al. Further reaction with TiCl₄ resulted in the incorporation of titanium of approximately 8, 38, and 54% in the oxidation states of +2, +3, and +4, respectively. The ratio of Al to Ti in the catalyst lies in the range of 0.5–1.0. Only 19% of all the Ti⁺³ species in the catalyst can be observed by electron paramagnetic resonance (EPR); these are attributable to isolated Ti⁺³ complexes. The remaining EPR silent Ti⁺³ species are believed to be bridged to another Ti⁺³ by Cl ligands. The total Cl content is equal to the sum of 2 × Mg + 3 × Al + 3.5 × Ti. Most of the p-cresol moiety apparently disappeared from the support, leaving much of ethyl benzoate in the catalyst. Activation with AlEt₃/methyl-p-toluate complex reduces 90% of the Ti⁺⁴ in the catalyst to lower oxidation states. The ester apparently moderates the alkylating power of AlEt₃ to avoid excessive formation of divalent titanium sites. There appears to be a constant fraction of 1/4–1/5 of the titanium which is isolated and the remainder is in bridged clusters independent of the oxidation states of titanium.     

High activity magnesium chloride supported catalysts for olefin polymerization. XXIX. Molecular basis of hydrogen activation of magnesium chloride supported Ziegler–Natta catalysts

Hydrogen (pH₂ = 72 torr) increases the rate of propylene polymerization by a MgCl₂/ethyl benzoate/p-cresol/AlEt₃/TiCl₄-AlEt₃/methyl-p-toluate catalyst (CW-catalyst) by two-to three-fold which corresponds closely with the increase in the number of active sites as counted by radiolabeling with tritiated methanol. The oxidation states of titanium in decene polymerizations by the CW-catalyst were determined as a function of time of polymerization (t_p). In the absence of H₂, all [Ti⁺ⁿ] for n = 2, 3, 4 remain constant during a batch polymerization. In the presence of H₂ and within 5 min of t_p, [Ti⁺²] decreases by an amount, corresponding to 15% of the total titanium and [Ti⁺³] increases by the same amount, while [Ti⁺⁴] is not changed. Therefore, three-fourths of the H₂ activation result from oxidative addition processes. The remaining one-fourth of the H₂ activation may be attributed to the activation of previously deactivated Ti⁺³ ions by hydrogenolysis. Monomer converts some of the EPR silent Ti⁺³ sites to EPR observable species resulting in their activation.     

Magnesium chloride supported catalysts for olefin polymerization. XIX. Titanium oxidation states,catalyst deactivation,and active site structure

A precise method for the determinations of Ti⁺², Ti⁺³ and Ti⁺⁴ was developed. The CW-procatalyst before activation contains mostly Ti⁺⁴ ions with 6% Ti⁺³ and 4% Ti⁺² ions. Activation with AlEt₃ alone at room temperature reduced all the titaniums to lower valence states consisting of 71% Ti⁺³ and 29% Ti⁺². Reduction is incomplete when methyl-p-toluate was present as external Lewis base during activation: at 25°C the distribution of Ti⁺⁴ : Ti⁺³ : Ti⁺² is 36% : 25% : 38%; the distribution at 50°C is 37% : 22% : 40%. Aging of the activated catalyst caused little or no changes in the distribution of [Ti⁺ⁿ]; whereas the catalytic activity decays rapidly with aging. The aged catalysts have polymerization activity comparable to the decreased activity of the catalyst during a polymerization. The [Ti⁺ⁿ] was determined for the CW-catalyst during the course of a decene polymerization; they were found to be Ti⁺⁴ : Ti⁺³ : Ti⁺² = 30% : 27% : 43%, which did not change with polymerization time. These results suggest that the reducibility of Ti⁺⁴ species by AlEt₃ or 3AlEt₃/MPT to different valence states is predicated by their structures. These species do not undergo further changes in their oxidation states during either aging or polymerization. Their decays probably involve nonreductive metathesis reactions like those known for zirconium alkyls. Possible structures for the stereospecific and nonspecific sites are proposed.     

Superactive and stereospecific catalysts. III. Definitive identification of active sites by electron paramagnetic resonance

Superactive Ziegler–Natta catalysts have been prepared from a soluble MgCl₂·2-ethyl hexanol adduct in the presence of organic esters through reactions with TiCl₄ and activated with AlEt₃/phenyltriethoxy-silane. Electron paramagnetic spectra (EPR) were used to elucidate the nature and amount of those Ti⁺³ ions not bridged to another Ti⁺³ ion; the chlorine bridged Ti⁺³ ions are EPR silent. The EPR spectra were attributed to two rhombic Ti⁺³ sites with principal values for the g-tensors (1.967, 1.949, 1.915; and 1.979, 1.935, 1.887). The total amount of the EPR species, obtained by double integration of the EPR spectra, is in close agreement with the concentration of isospecific catalytic sites determined by radiotagging. This suggests that the nonspecific sites are EPR silent. When o-phthalic ester was present during the catalyst synthesis, there appears an EPR signal at the free electron g-value. This signal was attributed to a Ti⁺³ phthalate species with resonance stabilization and spin delocalization; it is absent in the catalysts made in the presence of monoesters such as ethyl benzoate. The effects of monomer, O₂, H₂O, and I₂ on the EPR spectra were investigated. The changes in the EPR spectral intensity and the total Ti⁺³ ions, the latter determined by redox titrations during a polymerization or catalyst aging, are described. The results were extensively compared with those observed for supported Ziegler–Natta catalyst prepared with crystalline MgCl₂.     

Magnesium chloride supported high mileage catalysts for olefin (decene-1) polymerization XV. Termination and energetics

Polymerizations of decene-1 were carried out from 0° to 70° at A/T = 167 and [M] = 0.75 M initiated by 0.17, 0.34, and 0.69 mM of Ti contained in the MgCl₂/ethylbenzoate/p-cresol/AlEt₃/TiCl₄-AlEt₃/methyl-p-toluate catalyst. The rate of polymerization is directly proportional to the catalyst concentration. About 12% of the Ti in the catalyst is initially active at 50°; they are 1.4%, 8.8%, and 9.4% at 0°, 25°, and 70°, respectively. The changes of R_p with temperature parallels the variations in the active site concentration. The decline of R_p with time has second-order plots with slopes which are inversely proportional to the catalyst concentration, but the rate constants for these deactivations are nearly the same for decene and propylene polymerizations. These results strongly support a mechanism of deactivation involving two adjacent sites in the catalyst particle surfaces. The rate constants of propagation and of chain transfer to AlEt₃, the energetics for these processes, and MW and MW distribution data have been obtained.     

Magnesium chloride supported high-mileage catalysts for olefin polymerization. IV. FTIR and quantitative analysis of modifiers in the catalysts

Fourier transform infrared (FTIR) spectra were obtained for a typical MgCl₂-supported, high-mileage catalyst for propylene polymerization. When ball-milling MgCl₂ with ethyl benzoate (EB), the latter is incorporated into the support (I) by Lewis acid-base complexation involving both oxygen atoms of the ester. Reaction of (I) with p-cresol (PC) resulted in a material (II) that contains all the characteristic IR bands of PC. The reaction of (II) that contains all the characteristic IR bands of PC. The reaction of (II) with AlEt₃ (TEA) resulted in (III) whose spectrum supports the reaction observed by product analysis and NMR spectroscopy. There was no evidence of any reaction between TEA and EB. Further reaction of (III) with an excess of TiCl₄ caused substantial removal of the p-cresol moiety as shown by the diminution of its characteristic bands. Finally, activation with 3TEA-1MT (methyl-p-toluate) complexes gave spectra that revealed the presence of MT in the activated catalyst without any visage of p-cresol moiety. The nondestructive FTIR method, however, is not quantitative. Quantitative analysis of the organic components in the support materials (I), (II), and (III) and the catalysts was accomplished by hydrolysis of the inorganic components, extraction with ether, and analysis by gas chromatography. The results are in good agreement with composition deducted from elemental analysis and substantiate the FTIR conclusions.     

Probing the Coordinative Unsaturation and Local Environment of Ti3+ Sites in an Activated High‐Yield Ziegler–Natta Catalyst

The typical activation of a fourth generation Ziegler–Natta catalyst TiCl₄/MgCl₂/phthalate with triethyl aluminum generates Ti³⁺ centers that are investigated by multi‐frequency continuous wave and pulse EPR methods. Two families of isolated, molecule‐like Ti³⁺ species have been identified. A comparison of the experimentally derived g tensors and ^35,37Cl hyperfine and nuclear‐quadrupole tensors with DFT‐computed values suggests that the dominant EPR‐active Ti³⁺ species is located on MgCl₂(110) surfaces (or equivalent MgCl₂ terminations with tetra‐coordinated Mg). O₂ reactivity tests show that a fraction of these Ti sites is chemically accessible, an important result in view of the search for the true catalyst active site in olefin polymerization.     

Magnesium chloride supported high mileage catalyst for olefin polymerization. VII. Determinations of concentrations of titanium–polymer and aluminum–polymer bonds

Two methods were used in an attempt to determine by radioquenching the active site concentration, [Ti*], in a MgCl₂ supported high activity catalyst. For the reactions of tritium labelled methanol, the kinetic isotope effects were first determined: k_H/k_T = 1.63 for the total polymer and 1.67 for the isotactic polypropylene fraction. Polymerizations were quenched with an excess of isotopic CH₃OH after various lengths of time, at different A/T (amount of AlEt₃ with 0.33 equivalent of methyl-p-toluate to amount of Ti in the catalyst) ratios, and temperatures. From the known specific activity of tritium in CH₃OH and radioassay of the polymer, value of the total metal polymer bond, [MPB], can be obtained. [MPB] increases linearly with polymerization time. Extrapolation to t = 0 gives [MPB]₀, which should be close to [Ti*] because chain transfer with aluminum alkyls to produce Al–P bonds is negligible during very early stage of the polymerization. The values of [MPB]₀ range from 7–30% of the total Ti; the number of MPB is nearly equally distributed in the amorphous and isotactic fractions of polypropylene in most runs. The rate of incorporation of radioactive CO into polymers produced by the MgCl₂ supported high mileage catalyst is far slower than that claimed by some investigators for TiCl₃ type catalysts. There is an initial rapid phase of incorporation of CO which lasts for about 1 hr of contact time. The subsequent rate of CO incorporation steadily declines, yet there is no constant maximum value of radioactivity even after 48 h of reaction in the absence of monomer. Radioquenching of polymerizations with CO was also performed at several temperatures and A/T ratios. In all cases, the maximum [Ti–P] was reached after 30–40 min of polymerization, whereas the maximum rates of polymerization, R_p,m, occurred within 3–10 min. In fact, the rate of polymerization decays to a small fraction of R_p,m after 30–40 min. Furthermore, this maximum value of [Ti–P] remains constant until the end of polymerization (t = 90 min). Therefore, isotopic CO is not reacting with the initially formed active sites Ti₁*, but only with those sites, Ti₂*, which predominate during the later stage of polymerization.     

Magnesium chloride supported high mileage catalysts for olefin (decene-l) polymerization XIV. Propagation and chain transfer

Decene-l was polymerized with the MgCl₂/ethylebenzoate/p-cresol/AIEt₃/TiCl₄-AlEt₃/methyl-p-toluate catalyst at 50° using an A/T ratio of 167 and a range of monomer concentration. The concentration of the two kinds of active sites are [Ti] = 12% and [Ti] = 4% of the total titanium. The rate constants of propagation are 24 M^?1 s^?1. Chain transfers to AIEt₃, monomer, and by β-hydride elimination have rate constant values of 1.7 × 10^?3 M^?1 s^?1, 1.34 × 10^?2 M^?1 s^?1, and 1.7 × 10^?2 s^?1, respectively. Poly(decene-l) have relatively narrow MW which are unchanged during the course of a polymerization. Therefore, the active site concentrations in the CW catalyst for propylene and decene polymerization are identical and their rate constant values agree within a factor of 2. However, the rate of decene polymerization depends on fractional order of monomer concentration and decreases with the increase of activator concentration. Furthermore, the formation of metal polymer bonds has a rate independent of these concentrations. These kinetic behaviors are a manifestation of absorption processes of these species which are not seen in propylene polymerizations.     

Magnesium chloride supported high-mileage catalysts for olefin polymerizations. XVIII. Effect of hydrogen and lewis bases

Hydrogen has been found earlier to increase the initial rate of polymerization by MgCl₂/EB/PC/AlEt₃/TiCl₄-AlEt₃/MPT, CW-catalyst (+B_i, +B_e) (EB, ethyl benzoate; PC, p-cresol; MPT, methyl-p-toluate), but decays more rapidly as compared to polymerizations in the absence of H₂. In this study the effect of H₂ was studied when either the internal Lewis base, EB B_i, or the external Lewis base, MPT B_e, or both are deleted from the CW-catalyst. H₂ does not affect the stereospecificity of all the catalysts, but causes a slight increase of polymer yield, whereas the yield is virtually unchanged by H₂ for the catalysts activated with B_e. Unlike the catalyst (+B_i, +B_e) where H₂ increases active site concentrations [Ti^*] about threefold, it affects [Ti^*] negligibly when B_e is absent. The rate constants of propagation is about the same with or without H₂ for the CW-catalyst (+B_i, –B_e) or (–B_i, –B_e); the same statement can be said about the rate constant of chain transfer with AlEt₃ or with H₂. Hydrogen increases the rate of catalyst site deactivation for the various catalysts in the order of(+B_i, +B_e) > (–B_i, –B_e) > (+B_i, –B_e).     

Magnesium chloride supported high mileage catalysts for olefin polymerization. XIII. Effect of external Lewis base on ethylene polymerization

The effect of external Lewis base (B_e) on the polymerization of ethylene by the MgCl₂/ethyl benzoate/p-cresol/AlEt₃/TiCl₄ catalyst was studied by activation with AlEt₃ alone without the use of methyl-p-toluate. The initially formed active site concentration, [Ti], is about doubled in the absence of B_e; at 50°C about 93% of the total titanium became catalytic. The same increment of [Ti] was observed without B_e. The rate constants of propagation are not appreciably affected by B_e; the values are the same at 50° with and without B_e. At other temperatures the k_p values are somewhat smaller without B_e. One major effect was the very large k values for chain transfer with aluminum alkyls in the absence of B_e as compared to those with B_e. This can be attributed to the greater monomeric AlEt₃ concentration in the former, but in much smaller amounts in the presence of B_e due to complexation. The rate constants of chain transfer with hydrogen are not much affected by B_e. However, the termination rate constants are generally much smaller when Lewis base is not present.     

Magnesium chloride supported high-mileage catalysts for olefin polymerization. XVII. Effect of lewis base on propylene polymerization

A systematic study has been made on the functions of external Lewis base (B_e, methyl-p-toluate, MPT) and internal Lewis base (B_i, ethyl benzoate, EB) for the CW-catalyst system MgCI₂/EB/PC/AlEt₃/TiCl₄–AlEt₃/MPT (PC, p-cresol). B_i is a nonstereoselective modifier. It increases the active site concentrations and rate constants of propagation, k_p, of both the isospecific and nonspecific sites, and thus the productivities of the stereoregular and irregular polypropylenes by five- to tenfold. It seems that B_i complexes with the MgCl₂ support to lower the electronegativity of the surface Mg atoms. It also acts to lower the rate constant of chain transfer to aluminum alkyl, k, by two- to fourfold. The action of B_e is highly stereospecific. The isotacticity index of polypropylene is ? 95% in the presence of B_e but ? 68% without it. Addition of B_e decreases nonspecific [Ti^*]_a by about (11 ± 2)-fold; there is only about a twofold reduction of the isospecific [Ti^*]_i. It decreases k_p,a about three times but has no effect on k_p,i, so that the latter is (21 ± 4) times the former. B_e decreases k for transfer with aluminum alkyl much more than it does to k; but it does not affect the rates of chain transfer with monomer and by β-hydride elimination or the rate of catalyst deactivation. The number of active sites without B_e is [Ti^*]_i = 15% and [Ti^*]_a = 55% for a total of 70%. In the presence of B_e they are both about 6%. Optimum performance in propylene polymerizations requires both B_i and B_e in the case of the CW-catalyst.     

Kinetics of propylene polymerization in the initial acceleration stage

The kinetics of propylene polymerization catalyzed over a superactive and stereospecific catalyst for the initial build-up period was investigated in slurry-phase. The catalyst was prepared from Mg(OEt)₂/benzoyl chloride/TiCl₄ co-activated with AlEt₃ in the absence or presence of external donor. Despite a very fast activation of the prepared catalyst the acceleration stage of polymerization could be identified by the precise estimation of polymerization kinetics for a very short period of time after the commencement of polymerization (ca. 2 min). The initial polymerization rate, (dR_p/dt)₀ extrapolated to the beginning of the polymerization was second order with respect to monomer concentration. The dependence of initial polymerization rate on the concentration of AlEt₃ could be represented by Langmuir adsorption mechanism. The initial rate was maximum at about Al/Ti ratio of 20. The activation energy for the initiation reaction was estimated to be 14.3 kcal/mol for a short-time polymerization. The addition of a small amount of p-ethoxy ethyl benzoate (PEEB) as an external donor increased the percentage of isotactic polymer, which was obtained after 120 s of polymerization, to 98% and the initial polymerization rate decreased sharply as [PEEB]/[AlEt₃] increased. © 1994 John Wiley & Sons, Inc.     

State of titanium in supported titanium-magnesium catalysts for propylene polymerization

The oxidation state of titanium and the coordination state of Ti³⁺ ions in TiCl₄/D₁/MgCl₂ (D₁ is a phthalate) supported titanium-magnesium catalysts (TMCs) after the interaction with an AlEt₃/D₂ cocatalyst (D₂ is propyltrimethoxysilane or dicyclopentyldimethoxysilane) were studied by chemical analysis and EPR spectroscopy. Different oxidation state distributions of titanium ions were observed in the activated catalyst and mother liquor: Ti³⁺ and Ti²⁺ ions were predominant in the activated catalyst and mother liquor, respectively. The effects of interaction conditions (reaction temperature and time and Al/Ti and D₂/Ti molar ratios) of TMCs with the cocatalyst on the state of titanium in activated samples were studied. The interaction of TMCs with the cocatalyst decreased the titanium content and caused the appearance of aluminum in the activated sample, which was most clearly pronounced at a temperature of 25°C and occurred within the first 10 min of treatment. An increase in the temperature to 70°C and an increase in the interaction time to 60 min only slightly affected the concentrations of titanium and aluminum. The presence of D₂ as a cocatalyst constituent facilitated the removal of titanium compounds and restricted the adsorption of aluminum compounds on the catalyst surface. The main fraction of titanium consisted of Ti³⁺ ions (62–89%), and the rest was Ti⁴⁺ ions (22–35%) under mild interaction conditions (25°C; Si/Ti = 25) or Ti⁴⁺ (0–21%) and Ti²⁺ (9–21%) ions under more severe conditions (50 or 70°C; Si/Ti from 0 to 5). According to EPR-spectroscopic data, at D₂/Ti from 1 to 5, Ti³⁺ ions mainly occurred as associates, whereas they occurred as isolated ions at D₂/Ti = 25. The initial and activated catalysts were similar in activity in the reaction of propylene polymerization, and titanium compounds, which were removed from the catalyst upon interaction with AlEt₃/D₂, were inactive in this process.     

Kinetics and mechanism of acetylene polymerization

The kinetics of acetylene polymerization initiated by Ti(OBu)₄/4AlEt₃ catalyst was studied by radioquenching with C^*O to count the number of active sites [C] and by CH₃OT^* to determine the total metal polymer bonds [MPB] and M?_n of the polymer. The amount of quenching agent and time of reaction required and the kinetic isotope effect for CH₃OT^* were determined. The effects of Al/Ti ratio, catalyst aging, catalyst concentration, temperature, and monomer pressure on the polymerization were investigated. Detailed kinetic data on the variation of rate of polymerization, R_p, [C] [MPB], and M?_n with time were obtained at 298 and 195°K. The results required the assumption that the catalytic species C, is initially active and within less than 30 min all are converted by bimolecular kinetics to a far less active species. Analysis of the data yielded rate constants of propagation and termination and their energies of activation. Estimates of chain transfer efficiency were obtained. The mechanisms for the propagation, termination, and transfer processes were discussed. By drawing on our earlier EPR results we propose probable structures for the catalytically active species.     

Polymer-supported Ziegler–Natta catalyst for the polymerization of isoprene

A polymer-supported Ziegler–Natta catalyst, polystyrene-TiCl₄AlEt₂Cl (PS–TiCl₄AlEt₂Cl), was synthesized by reaction of polystyrene–TiCl₄ complex (PS–TiCl₄) with AlEt₂Cl. This catalyst showed the same, or lightly greater catalytic activity to the unsupported Ziegler–Natta catalyst for polymerization of isoprene. It also has much greater storability, and can be reused and regenerated. Its overall catalytic yield for isoprene polymerization is ca. 20 kg polyisoprene/gTi. The polymerization rate depends on catalyst titanium concentration, mole ratio of Al/Ti, monomer concentration, and temperature. The kinetic equation of this polymerization is: R_p = k[M]^0.30[Ti]^0.41[Al]^1.28, and the apparent activation energy ΔE_act = 14.5 kJ/Mol, and the frequency factor A_p = 33 L/(mol s). The mechanism of the isoprene polymerization catalyzed by the polymer-supported catalyst is also described. © 1993 John Wiley & Sons, Inc.     

Polymerization of tetrahydrofuran by AlEt3–H2O promoter system: Rate of propagation reaction

Kinetic studies were carried out on the polymerization of tetrahydrofuran with catalyst systems of aluminum alkyl–epichlorohydrin. As aluminium alkyl species AlEt₃, AlEt₃–H₂O (1:0.1 to 1:1.0), and “oxyaluminum ethyl” were employed. The polymerizations with these catalysts are characterized by a mechanism of stepwise addition without chain transfer or termination, which is expressed by the kinetic relation R_p = K_p[P*] ([M]–[M]_e), where [M] and [M]_e are the instantaneous and equilibrium concentrations of monomer and [P*] is the concentration of propagating species calculated from the amount and molecular weight of the product polymer. The determination of the rate constant k_p for these catalysts has shown that the polymerization rate varied considerably with the change of aluminum alkyl species, i.e., with the water-to-aluminum ratio, but the propagation rate constant itself varied very little. The variation of polymerization rate was, therefore, attributed primarily to the differences in concentration of the propagating species, i.e. the efficiency of the catalyst in forming propagating species. The catalyst efficiency was closely related to the acid strength of the aluminum alkyl species, which was estimated from the magnitude of shift of the xanthone carbonyl band in the infrared spectrum of its coordination complex with aluminum alkyl. The maximal catalyst efficiency was attained at about [H₂O]/[AlEt₃] = 0.75.     

Magnesium-chloride-supported high-mileage catalysts for olefin polymerization. II. Reactions between aluminum alkyl and promoters

The reactions between AlEt₃ and the modifiers, promoters, and coactivators of a typical magnesium-chloride-supported, high-activity propylene polymerization catalyst were studied. Infrared, MS analysis of the gas evolved, and GC–MS of the hydrolysis products for the reaction between AlEt₃ and p-cresol showed rapid and quantitative reactions with p-cresol either in the support or solution. The reaction products from AlEt₃ and esters were hydrolyzed, acidified, and dehydrated. The resulting carbonyl and olefinic compounds were identified by GC–MS. Proton and carbon nuclear magnetic resonance (NMR) techniques were also used to study these reactions. The expected intermediates were found in the PMR and CMR spectra. The mechanisms of reactions were proposed. The results of this study showed that when AlEt₃ and esters are used as coactivators reaction products that can significantly influence the performance of the catalyst are formed.     

EPR Studies of Photoreduction of Ni/TiO2 Catalysts

Irradiations of Ni/TiO₂ catalyst by UV in hydrogen at 77 K produced not only Ni⁺ ions on the catalyst surface, but also Ni³⁺ and Ti³⁺ species in bulk or near the interface between nickel and titania. These photo-generated species were detected and characterized by low temperature electron paramagnetic resonance (EPR) spectroscopy. Relative spin concentrations of the photogenerated paramagnetic species (Niⁿ⁺ and Ti³⁺) varied with the nickel content in titania. A high nickel content in the sample resulted in a high peak intensity ratio of Niⁿ⁺ to Ti³⁺. It was found that the photoinduced self-redox reaction of Ni²⁺ ions to form Ni⁺ and Ni³⁺ ions has a priority over the photoreduction of Ti⁴⁺ to Ti³⁺ ions. The characteristic EPR spectrum of the Ni³⁺ (3d⁷) ions with g₁ = 2.268, g₂ = 2.237, and g₃ = 2.045 indicates that the Ni³⁺ ions are most likely located in the substitutional sites of TiO₂, possibly near the surface rutile phase. The Ni⁺ species (3d⁹) with g₄ = 2.130 and g₁ = 2.063 are on the surface of TiO₂. Both Ni⁺ and Ni³⁺ ions are quite stable in hydrogen. The Ni³⁺ ions seem to be responsible for anchoring the nickel ions onto titania and stablizing the Ni⁺ species on the surface. The Ni⁺ ions are thus free from oxygen poisoning and still show a high activity toward olefin oligomerization.