Titanium(IV) oxide having a copper co-catalyst: a new type of semihydrogenation photocatalyst working efficiently at an elevated temperature under hydrogen-free and poison-free conditions

Abstract

A copper-loaded titanium(IV) oxide photocatalyst exhibited perfect selectivity in hydrogenation of alkynes to alkenes in an alcohol solution at 298 K under hydrogen-free and poison-free conditions. A slight elevation in the reaction temperature to 323 K greatly increased the reaction rate with the selectivity being preserved and the formation of an H₂ by-product being suppressed. The apparent activation energy of 4-octyne semihydrogenation was determined to be 54 kJ mol⁻¹, indicating that the rate determining step of this photocatalytic reaction was not an electron production process but a thermocatalytic hydrogenation process under light irradiation.

---

When titanium(IV) oxide (TiO₂) is irradiated by UV light, electrons in the valence band (VB) are excited to the conduction band (CB), leaving positive holes in the VB. The thus-formed electrons and positive holes cause reduction and oxidation, respectively, of compounds adsorbed on the surface of a photocatalyst.^1,2 Since positive holes have strong oxidation power, total or partial oxidation of organic compounds has been extensively studied. On the other hand, applications of photocatalytic reduction have been less frequently reported because the reduction potential of many organic compounds is more negative than the reduction potential of the CB. Therefore, target compounds in photocatalytic reduction (or hydrogenation) are basically limited to those having a carbonyl group³ and a nitro group,⁴ and other photocatalytic reductions of organic compounds have scarcely been reported. Recently, we reported photocatalytic reduction (hydrogenation) of a cyano group (benzonitrile) to an amino group (benzylamine) using palladium-loaded TiO₂ (Pd–TiO₂) even though the reduction potential of benzonitrile is higher than the potential of the conduction band of TiO₂,⁵ indicating that the applicability of photocatalytic reduction is not limited by the CB position of semiconductor photocatalysts and that a new photocatalytic reduction can be developed if metal co-catalysts are introduced on photocatalysts. Subsequently, we found that difficult reductions can be achieved by combining the TiO₂ photocatalyst with silver (Ag) and copper (Cu) co-catalysts. (2,3-Epoxypropyl)benzene and 4-octyne were successfully converted to allylbenzene and cis-4-octene in alcohol suspensions of Ag-loaded TiO₂ and Cu-loaded TiO₂, respectively.^6,7 The excellence of Ag–TiO₂ and Cu–TiO₂ is that these photocatalysts showed complete chemoselectivity, i.e., no subsequent hydrogenation of alkenes to alkanes occurred over these photocatalysts. In the latter case, photocatalytic hydrogenation consists of two processes: (1) formation of an active hydrogen species over metal particles loaded on TiO₂ under light irradiation and (2) reduction (or hydrogenation) of the target compound over metal particles (Scheme 1).

Scheme 1 Semihydrogenation of alkynes to cis-alkenes over a Cu–TiO₂ photocatalyst.

In these processes, process (2) is not a photocatalytic process but a catalytic process. If process (1) is the rate determining step, the overall reaction rate can be increased by decreasing electron–hole recombination and increasing carrier trapping, which could be achieved by preparation of photocatalysts having high crystallinity and a large surface area. On the other hand, if process (2) is the rate determining step, acceleration of process (2) would be more effective than acceleration of process (1). In this study, we focused on the hydrogenation of alkynes in an alcohol suspension of a Cu–TiO₂ photocatalyst⁷ because we think that the latter hydrogenation process is the rate determining step. Here we briefly show that the overall reaction rate can be increased by acceleration of the catalytic (thermal) process at an elevated temperature.

By using the photodeposition method, 0.5 wt% Cu as a co-catalyst was loaded on TiO₂. TiO₂ powder (P 25 supplied by Nippon Aerosil Co., Ltd) was suspended in 10 cm³ of an aqueous methanol solution (10 vol%) containing copper chloride in a test tube. The test tube was sealed with a rubber septum under argon (Ar) and then photoirradiated for 90 min at λ > 300 nm by a 400 W high-pressure mercury arc (Eiko-sha, Osaka) with magnetic stirring in a water bath continuously kept at 298 K. The color of Cu–TiO₂ after the photodeposition process was purple, typically due to surface plasmonic resonance of metallic Cu nanoparticles. The resulting powder was washed repeatedly with distilled water and dried for 1 h in vacuo. The color turned faint blue when Cu–TiO₂ was exposed to air.

In a typical run, Cu–TiO₂ (50 mg) was suspended in 5 cm³ of methanol containing 50 μmol of 4-octyne in a test tube, which was sealed with a rubber septum and then photoirradiated under Ar with a xenon lamp (Optical Modulex, Ushio, Tokyo) in a water bath. The temperature of the water bath was changed to evaluate the effects of reaction temperature on the rate, product selectivity and chemoselectivity of the photocatalytic reaction. During the photocatalytic reaction, the color changed to gray, that is a typical color of fine metal particles. The amounts of 4-octyne, cis-4-octene, trans-4-octene and octane were determined with an FID-type gas chromatograph (GC-2025, Shimadzu, Kyoto) equipped with a DB-1 column. The amount of H₂ as the reduction product of protons (H⁺) was determined with a TCD-type gas chromatograph (GC-8A, Shimadzu) equipped with an MS-5A column. The color of Cu–TiO₂ again changed to faint blue when exposed to air after the reaction.

Fig. 1(a) shows the time course of photocatalytic conversion of 4-octyne in a methanolic suspension of Cu–TiO₂ at 298 K. 4-Octyne was consumed along with photoirradiation, while cis-4-octene was produced instead, indicating that 4-octyne was hydrogenated by active hydrogen species that were formed over Cu by reduction of protons by photogenerated electrons. Material balance (MB) was calculated to be almost unity from eqn (1):

where n(4-octyne) and n(cis-4-octene) are the amounts of 4-octyne and cis-4-octene during the photocatalytic reaction, respectively, and n₀(4-octyne) is the amount of 4-octyne before the photocatalytic reaction. This indicates that chemoselective and diastereoselective hydrogenation of 4-octyne to cis-4-octene occurred. Actually, other hydrogenated compounds such as trans-4-octene and octane were not observed (Fig. S1, ESI†). We noted that there was an induction period in the first hour and that there was a linear increase after the induction period. Changes in the color of Cu–TiO₂ and the induction period suggest that the Cu species loaded on TiO₂ was partly oxidized to Cu(OH)₂-like species under air and that reduction of the oxidized parts to the metallic state by photogenerated electrons is necessary to show the activity for semihydrogenation. The yield of cis-4-octene after 3 h including the induction period for regeneration of the Cu species was 50% at 298 K. In addition to semihydrogenation of 4-octyne, active hydrogen species over Cu were coupled, resulting in H₂ production (Fig. S2, ESI†). This means that semihydrogenation and H₂ evolution are competitive reactions in utilization of photogenerated electrons. As clearly shown in Fig. 1(a), the yield of H₂ reached 150 μmol, which was much larger than that of the hydrogenated product (cis-4-octene, 25 μmol), suggesting that activation energy for H₂ formation is smaller than that for the hydrogenation process. This means that selectivity in electron utilization is decreased by H₂ formation in this system; therefore, some ingenuity for suppressing H₂ evolution is required to improve selectivity in electron utilization for semihydrogenation in this system.

Fig. 1 Time courses of the amounts of 4-octyne, cis-4-octene and H₂ and the material balance (MB) in a methanolic suspension of 0.5 wt% Cu–TiO₂ photocatalyst under deaerated conditions at (a) 298 K and (b) 323 K.

Fig. 1(b) shows the time course of the same reaction operated at 323 K. The reaction rate for semihydrogenation drastically increased at 323 K, and 4-octyne was almost completely consumed at 1 h and 40 min with chemoselectivity and diastereoselectivity being preserved as they were at 298 K, i.e., 4-octyne was almost quantitatively hydrogenated to cis-4-octene within a short time. We noted that the formation of H₂ was completely suppressed during the production of cis-4-octene and that H₂ evolved only after consumption of 4-octyne. The results of photocatalytic reaction at 323 K were in contrast to those of the reaction at 298 K shown in Fig. 1(a). After consumption of 4-octyne, no cis-4-octene was consumed, indicating that Cu–TiO₂ possessed complete chemoselectivity for semihydrogenation of 4-octyne, i.e., no cis-4-octene was hydrogenated and isomerized to trans-4-octene over Cu–TiO₂ even at 323 K.

To understand the semihydrogenation of 4-octyne under light irradiation, we carried out additional experiments under different conditions: dark reactions at 323 K under argon (Ar, 1 atm) and H₂ (1 atm, 1.3 mmol). Since the Cu species was partly oxidized as shown in Fig. 1(a), Cu–TiO₂ was used for the dark reactions after reduction of the Cu species in methanol under light irradiation in the absence of 4-octyne. No reactions occurred in the dark under Ar and H₂ at 323 K. The result under Ar indicates that no thermal reaction occurred over Cu metal at 323 K and that light irradiation is indispensable for inducing hydrogenation of 4-octyne. The result under H₂ is an interesting feature of this system, i.e., an excess of H₂ (1.3 mmol) in the gas phase did not contribute to the hydrogenation of 4-octyne over Cu–TiO₂ at 323 K, although semihydrogenation of alkynes with H₂ over Cu supported on silica (Cu/SiO₂) has been reported in both liquid-batch and gas-flow systems.⁸

There are some possible reasons for the differences between the present and previously reported results. First, the reaction temperature (323 K) is still low for dissociative adsorption of H₂ on Cu metal in a methanolic suspension. Second, H₂ pressure (1 atm) is insufficient for H₂ to be dissolved in methanol or to be adsorbed on Cu–TiO₂ suspended in methanol. Third, the hydrogen active species formed by dissociative adsorption of H₂ and that formed photocatalytically are somehow different.

Since the reaction temperature showed a great effect on the reaction rate of photocatalytic semihydrogenation of 4-octyne to cis-4-octene, photocatalytic reactions were carried out at various temperatures. Fig. 2(a) shows the amounts of 4-octyne and cis-4-octene and the material balance calculated by eqn (1) after 1 h irradiation. The reaction rate for semihydrogenation of 4-octyne to cis-4-octene increased with elevation of the reaction temperature. The results were used for estimation of the apparent activation energy (E_a) for this reaction over Cu–TiO₂, and an Arrhenius plot is shown in Fig. 2(b). A linear correlation was observed in the Arrhenius plot between 298 K and 323 K. From the slope of the plot, E_a in this reaction system was estimated to be 54 kJ mol⁻¹.

Fig. 2 (a) Effect of reaction temperature on semihydrogenation of 4-octyne to cis-4-octene in a methanolic suspension of 0.5 wt% Cu–TiO₂ photocatalyst for 1 h. (b) Arrhenius plot (logarithm of k vs. reciprocal of T).

For comparison, photocatalytic reactions in a methanolic suspension of Cu–TiO₂ in the absence of 4-octyne were examined at various temperatures (Fig. 3(a)). In this case, H₂ was solely evolved as the reduced product because there is no other electron acceptor. The results were used for estimation of E_a for H₂ evolution over Cu–TiO₂, and an Arrhenius plot is shown in Fig. 3(b). A linear correlation was observed in the Arrhenius plot between 298 K and 323 K, and E_a of H₂ evolution under this condition was estimated from the slope of the plot to be 12 kJ mol⁻¹. The value of E_a was close to values of photocatalytic evolution of H₂ from aqueous solutions of ethylene glycol over gold-loaded TiO₂ (15 kJ mol⁻¹)⁹ and 2-propanol over platinum-loaded TiO₂ (20 kJ mol⁻¹).¹⁰ These results indicate that semihydrogenation of 4-octyne on Cu metal should overcome the relatively large E_a. Due to the large E_a of semihydrogenation, the reaction rate was small at temperatures around room temperature, resulting in a large yield of H₂. The reaction rate of semihydrogenation became large with elevation of the reaction temperature and finally exceeded the reaction rate of H₂ evolution at 323 K. Switching of the predominant reaction from H₂ formation to semihydrogenation at 323 K over Cu–TiO₂ in this system is attributed to the large frequency factor of the Arrhenius equation in semihydrogenation. These results indicate that the rate determining step is not a photocatalytic process but a thermocatalytic process. If the rate determining step is not a photocatalytic process, the reaction rate is independent of the light intensity. Photocatalytic reactions of 4-octyne in methanol suspensions of Cu–TiO₂ under photoirradiation of different intensities were examined (Table 1). Almost constant rates show that the light intensity is not a decisive factor in this reaction and that a photocatalytic process is not the rate determining step.

Fig. 3 (a) Time courses of H₂ evolution from a methanolic suspension of 0.5 wt% Cu–TiO₂ photocatalyst at various temperatures. (b) Arrhenius plot (logarithm of k vs. reciprocal of T).

Table 1 Effect of UV light intensity on semihydrogenation of 4-octyne to cis-4-octene in a methanolic suspension of Cu–TiO₂ photocatalyst^a

Light source^b	UV intensity^c/mW cm^−2	Temp./K	Rate^d/μmol h^−1
a 4-octyne: 50 μmol, methanol: 5.0 cm^3, 0.5 wt% Cu–TiO2: 50 mg. b Xenon lamp: Optical Modulex (Ushio, Tokyo), UV LED: PJ-1505-2CA (CCS, Kyoto). c Determined using a spectroradiometer USR-45D (Ushio, Tokyo). d Rate of cis-4-octene formation.
Xenon lamp	31.7	298	11.5
UV LED	4.7	296	11.0
UV LED	2.3	296	11.1

---

Since a slight elevation in the reaction temperature drastically increased the reaction rate and improved selectivity in electron utilization in diastereoselective semihydrogenation of 4-octyne, this effect was evaluated in another system, i.e., hydrogenation of 5-hexynenitrile having another functional group (Scheme 2).

Scheme 2 Photocatalytic semihydrogenation of 5-hexynenitrile.

A previous study revealed that only the C≡C triple bond was hydrogenated to the C=C double bond over a Cu–TiO₂ photocatalyst with the nitrile group being preserved.⁷Table 2 shows results of photocatalytic conversion of 5-hexynenitrile in a methanol suspension of Cu–TiO₂ at 298 K and 323 K. At 298 K, 5-hexynenitrile was almost completely consumed and 5-hexenenitrile was obtained in 93% yield after 7 h. When the reaction was carried out at 323 K, 5-hexynenitrile was almost quantitatively converted to 5-hexenenitrile only after 1.5 h. These results clarify that a slight elevation in the reaction temperature over a Cu–TiO₂ photocatalyst is effective for drastically increasing a reaction rate without the change in the chemoselectivity.

Table 2 Photocatalytic conversion of 5-hexynenitrile in a methanol suspension of Cu–TiO₂ under a deaerated condition^a

T^b/K	Time/h	Conv.^c/%	Yield^d/%	H2/μmol
a 5-Hexynenitrile: 48 μmol, methanol: 5.0 cm^3, Cu–TiO2: 50 mg, light source: UV LED (CCS). b Reaction temperature. c 5-Hexynenitrile conversion. d 5-Hexenenitrile yield.
298	7.0	98	93	150
323	1.5	99	99	80

---

In summary, a slight increase in the reaction temperature greatly increased the reaction rate for photocatalytic semihydrogenation of alkynes to alkenes in a methanolic suspension of Cu–TiO₂. A large amount of H₂ simultaneously evolved at 298 K, while only semihydrogenation of 4-octyne occurred at 323 K. These results indicate that the rate determining step is not a photocatalytic process but a thermocatalytic process. Switching of the predominant reaction from H₂ formation to semihydrogenation at 323 K over Cu–TiO₂ in this system is attributed to a large E_a and a large frequency factor for semihydrogenation and a small E_a for H₂ formation over a Cu co-catalyst. The results obtained in this study show that photocatalytic reactions can be thermally accelerated if a thermocatalytic process is the rate determining step in the photocatalytic reaction and E_a of the process is relatively large.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was partly supported by JSPS KAKENHI Grant Numbers 17H03462, 15J11412, 17H04967, and 16K18292. This work was also supported by MEXT-Supported Program for the Strategic Research Foundation at Private Universities 2014-2018, subsidy from MEXT and Kindai University. K. N. is grateful to the Japan Society for the Promotion of Science (JSPS) for a Research Fellowship for young scientists. A. T. is grateful for financial support from the Faculty of Science and Engineering, Kindai University.

References

1. J. Colmenares and Y.-J. Xu, Heterogeneous Photocatalysis: From Fundamentals to Green Applications, Springer-Verlag, Berlin, Heidelberg, 2016.
2. D. Friedmann, A. Hakki, H. Kim, W. Choi and D. Bahnemann, Green Chem., 2016, 18, 5391.
3. (a) C. Joyce-Pruden, J. K. Pross and Y. Li, J. Org. Chem., 1992, 57, 5087; (b) J. W. Park, M. J. Hong and K. K. Park, Bull. Korean Chem. Soc., 2001, 22, 1213; (c) Y. Matsushita, S. Kumada, K. Wakabayashi, K. Sakeda and T. Ichimura, Chem. Lett., 2006, 35, 410; (d) S. Kohtani, E. Yoshioka, K. Saito, A. Kudo and H. Miyabe, Catal. Commun., 2010, 11, 1049.
4. (a) F. Mahdavi, T. C. Bruton and Y. Li, J. Org. Chem., 1993, 58, 744; (b) J. L. Ferry and W. H. Glaze, Langmuir, 1998, 14, 3551; (c) O. V. Makarova, T. Rajh, M. C. Thurnauer, A. Martin, P. A. Kemme and D. Cropek, Environ. Sci. Technol., 2000, 34, 4797; (d) V. Brezová, P. Tarábek, D. Dvoranová, A. Staško and S. Biskupič, J. Photochem. Photobiol., A, 2003, 155, 179; (e) H. Tada, T. Ishida, A. Takao and S. Ito, Langmuir, 2004, 20, 7898; (f) T. Zhang, L. You and Y. Zhang, Dyes Pigm., 2006, 68, 95; (g) S. O. Flores, O. R. Bernij, M. A. Valenzuela, I. Córdova, R. Gómez and R. Gutiérrez, Top. Catal., 2007, 44, 507; (h) H. Kominami, S. Iwasaki, T. Maeda, K. Imamura, K. Hashimoto, Y. Kera and B. Ohtani, Chem. Lett., 2009, 15, 410; (i) K. Imamura, S. Iwasaki, T. Maeda, K. Hashimoto, B. Ohtani and H. Kominami, Phys. Chem. Chem. Phys., 2011, 13, 5114; (j) S. Chen, H. Zhang, X. Yu and W. Liu, Chin. J. Chem., 2010, 28, 21; (k) K. Imamura, K. Hashimoto and H. Kominami, Chem. Commun., 2012, 48, 4356; (l) K. Imamura, T. Yoshikawa, K. Hashimoto and H. Kominami, Appl. Catal., B, 2013, 134-135, 193.
5. K. Imamura, T. Yoshikawa, K. Nakanishi, K. Hashimoto and H. Kominami, Chem. Commun., 2013, 49, 10911.
6. H. Kominami, S. Yamamoto, K. Imamura, A. Tanaka and K. Hashimoto, Chem. Commun., 2014, 50, 4558.
7. H. Kominami, M. Higa, T. Nojima, T. Ito, K. Nakanishi, K. Hashimoto and K. Imamura, ChemCatChem, 2016, 8, 2019.
8. (a) J. T. Wehrli, D. J. Thomas, M. S. Wainwright, D. L. Trimm and N. W. Cant, Appl. Catal., 1991, 70, 253; (b) A. Fedorov, H.-J. Liu, H.-K. Lo and C. Copéret, J. Am. Chem. Soc., 2016, 138, 16502.
9. G. R. Bamwenda, S. Tsubota, T. Kobayashi and M. Haruta, J. Photochem. Photobiol., A, 1994, 77, 59.
10. F. H. Hussein and R. Rudham, J. Chem. Soc., Faraday Trans. 1, 1984, 80, 2817.

---

Footnotes

† Electronic supplementary information (ESI) available: Fig. S1 and S2. See DOI: 10.1039/c8cp02316f

‡ Present address: Research Laboratory of Hydrothermal Chemistry, Faculty of Science and Technology, Kochi University, Akebono-cho 2-5-1, Kochi City, Kochi 780-8520, Japan.

---