Method to reduce CO2 to CO using plasmon-enhanced photocatalysis

Abstract

Described is a method of reducing CO.sub.2 to CO using visible radiation and plasmonic photocatalysts. The method includes contacting CO.sub.2 with a catalyst, in the presence of H.sub.2, wherein the catalyst has plasmonic photocatalytic reductive activity when exposed to radiation having a wavelength between 380 nm and 780 nm. The catalyst, CO.sub.2, and H.sub.2 are exposed to non-coherent radiation having a wavelength between 380 nm and 780 nm such that the catalyst undergoes surface plasmon resonance. The surface plasmon resonance increases the rate of CO.sub.2 reduction to CO as compared to the rate of CO.sub.2 reduction to CO without surface plasmon resonance in the catalyst.

Claims

1. A method of reducing CO.sub.2 to CO, the method comprising: (a) contacting CO.sub.2 with a catalyst, in the presence of H.sub.2, wherein the H.sub.2 is present in a greater concentration than the CO.sub.2, wherein the catalyst has plasmonic photocatalytic reductive activity when exposed to radiation having a wavelength between about 380 nm and about 780 nm, and wherein the catalyst comprises a metallic element selected from the group consisting of calcium, copper, europium, gold, lithium, magnesium, palladium, platinum, potassium, silver, sodium, rubidium, and yttrium, and combinations thereof, has an average particle size no greater than 100 nm, and is deposited on an oxide semiconductor material; and (b) exposing the catalyst, CO.sub.2, and H.sub.2 to non-coherent radiation having a wavelength between about 380 nm and about 780 nm such that the catalyst undergoes surface plasmon resonance, wherein the surface plasmon resonance increases the rate of CO.sub.2 reduction to CO as compared to the rate of CO.sub.2 reduction to CO without surface plasmon resonance in the catalyst.

2. The method of claim 1, comprising, in step (b), exposing the catalyst, CO.sub.2, and H.sub.2 to solar radiation.

3. The method of claim 1, wherein the oxide semiconductor material is selected from the group consisting of oxides of titanium, aluminum, iron, silicon, zinc, and cerium, and combinations thereof.

4. The method of claim 1, wherein the metallic element comprises copper, silver, platinum, or gold, and the semiconductor material comprises titania or ceria.

5. The method according to any one of claim 1, 2, 3, or 4, wherein the surface plasmon resonance in the catalyst increases the rate of CO.sub.2 reduction to CO by a factor of at least 1.8 as compared to the rate of CO.sub.2 reduction to CO in the absence of surface plasmon resonance in the catalyst.

6. The method according to any one of claim 1, 2, 3, or 4, wherein the surface plasmon resonance in the catalyst increases the rate of CO.sub.2 reduction to CO by a factor of at least 3 as compared to the rate of CO.sub.2 reduction to CO in the absence of surface plasmon resonance in the catalyst.

7. The method according to any one of claim 1, 2, 3, or 4, wherein the surface plasmon resonance in the catalyst increases the rate of CO.sub.2 reduction to CO by a factor of at least 4 as compared to the rate of CO.sub.2 reduction to CO in the absence of surface plasmon resonance in the catalyst.

8. The method according to any one of claim 1, 2, 3, or 4, wherein the surface plasmon resonance in the catalyst increases the rate of CO.sub.2 reduction to CO by a factor of at least 5 as compared to the rate of CO.sub.2 reduction to CO in the absence of surface plasmon resonance in the catalyst.

9. A method of reducing CO.sub.2 to CO, the method comprising: (a) contacting CO.sub.2 with a catalyst, in the presence of H.sub.2, wherein the H.sub.2 is present in a greater concentration than the CO.sub.2, wherein the catalyst has plasmonic photocatalytic reductive activity when exposed to non-coherent radiation having a wavelength between about 380 nm and about 780 nm, and wherein the catalyst comprises a metallic element selected from the group consisting of calcium, copper, europium, gold, lithium, magnesium, palladium, platinum, potassium, silver, sodium, rubidium, and yttrium, and combinations thereof, has an average particle size no greater than 100 nm, and is deposited on an oxide semiconductor material; and (b) exposing the catalyst, CO.sub.2, and H.sub.2 to solar radiation such that the catalyst undergoes surface plasmon resonance, wherein the surface plasmon resonance increases the rate of CO.sub.2 reduction to CO as compared to the rate of CO.sub.2 reduction to CO without surface plasmon resonance in the catalyst.

10. The method of claim 9, wherein upon exposing the catalyst, CO.sub.2, and H.sub.2 to solar radiation, the catalyst achieves a light efficiency of at least about 2%.

11. The method of claim 9, wherein upon exposing the catalyst, CO.sub.2, and H.sub.2 to solar radiation, the catalyst achieves a solar light efficiency of at least about 3%.

12. The method of claim 9, wherein upon exposing the catalyst, CO.sub.2, and H.sub.2 to solar radiation, the catalyst achieves a solar light efficiency of at least about 4%.

13. The method of claim 9, wherein the oxide semiconductor material is selected from the group consisting of oxides of titanium, aluminum, iron, silicon, zinc, and cerium, and combinations thereof.

14. The method of claim 9, wherein the metallic element comprises copper, silver, platinum, or gold, and the semiconductor material comprises titania or ceria.

15. The method according to any one of claim 9, 10, 11, 12, 13, or 14, wherein the surface plasmon resonance in the catalyst increases the rate of CO.sub.2 reduction to CO by a factor of at least 1.8 as compared to the rate of CO.sub.2 reduction to CO in the absence of surface plasmon resonance in the catalyst.

16. The method according to any one of claim 9, 10, 11, 12, 13, or 14, wherein the surface plasmon resonance in the catalyst increases the rate of CO.sub.2 reduction to CO by a factor of at least 3 as compared to the rate of CO.sub.2 reduction to CO in the absence of surface plasmon resonance in the catalyst.

17. The method according to any one of claim 9, 10, 11, 12, 13, or 14, wherein the surface plasmon resonance in the catalyst increases the rate of CO.sub.2 reduction to CO by a factor of at least 4 as compared to the rate of CO.sub.2 reduction to CO in the absence of surface plasmon resonance in the catalyst.

18. The method according to any one of claim 9, 10, 11, 12, 13, or 14, wherein the surface plasmon resonance in the catalyst increases the rate of CO.sub.2 reduction to CO by a factor of at least 5 as compared to the rate of CO.sub.2 reduction to CO in the absence of surface plasmon resonance in the catalyst.

19. The method of any one of claim 9, 10, 11, 12, 13, or 14, wherein the method is conducted at a temperature of from about 100° C. to about 400° C., wherein H.sub.2 is present in a greater concentration than CO.sub.2, and the H.sub.2 and CO.sub.2 are present at a pressure of from atmospheric to about 2000 psi.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 is a schematic representation of harnessing solar energy to drive photoelectrochemical (PEC) splitting of water with subsequent plasmonic photocatalytic reduction of CO.sub.2 to value-added products such as formic acid, syn-gas (a mixture of CO and H.sub.2) and hydrocarbons.

(2) FIG. 2 is a schematic representation of plasmonic photocatalysis, illustrating that the plasmonic resonance causes a number of beneficial phenomena that drive catalysis, including intense scattering of the incoming radiation, electron/hole pair generation, and localized heating, all of which impact catalysis.

(3) FIG. 3 is a histogram showing CO.sub.2 conversion rates for the reverse water gas shift reaction (CO.sub.2+H.sub.2.fwdarw.CO+H.sub.2O; ΔH.sub.reaction=41 kJ/mol) using different catalysts in “dark” mode (no added light) and “enhanced” mode (in presence of simulated solar radiation). Dark reaction shown in black bars, with rate depicted on the left-hand Y-axis; reaction with simulated solar radiation shown in open bars, with enhancement factor depicted on the right-hand Y-axis. Reaction conditions: H.sub.2:CO.sub.2=2:1; total gas flow rate=15 sccm, P=110 psi, T=400° C. (The term “sccm” denotes standard cubic centimeters per minute indicating cc/min at 0° C. at 1 atmosphere of pressure. This unit is used to calculate the amount of gas or volume of gas that passes through a given point in a unit time.)

(4) FIG. 4 is a graph depicting CO.sub.2 conversion rates for the reverse water gas shift reaction as a function of temperature. Light reaction (.diamond-solid.); dark reaction (.circle-solid.). Reaction conditions: H.sub.2:CO.sub.2=2:1; total gas flow rate=15 sccm, P=110 psi, T=100° C. to 400° C.

(5) FIG. 5 is a graph depicting enhancement due to light as a function of temperature for the reverse water gas shift reactions whose conversion rates are depicted in FIG. 4. Rate enhancement (light rate/dark rate) depicted on Y-axis. As shown in the figure, the enhancement moves inversely to temperature.

(6) FIG. 6 is a graph depicting rate due to light versus temperature for the reverse water gas shift reactions whose conversion rates are depicted in FIG. 4.

(7) FIG. 7: is a graph depicting light efficiency versus temperature for the reverse water gas shift reactions whose conversion rates are depicted in FIG. 4. Again, Reaction conditions: H.sub.2:CO.sub.2=2:1; total gas flow rate=15 sccm, P=110 psi. Light efficiency is defined as

(8) $\mathrm{Light}\mathrm{efficiency}\left(\%\right)=\frac{{\mathrm{CO}}_2\mathrm{conversion}\mathrm{rate}\mathrm{due}\mathrm{to}\mathrm{light}\times \Delta H_{\mathrm{reaction}}}{\mathrm{Intensity}\times \mathrm{Catalyst}\mathrm{surface}\mathrm{area}}\times 100\%$

(9) FIG. 8 is a plot depicting ln(CO.sub.2 conversion rate) versus 1/Temp(1/K)×10.sup.3 for the reverse water gas shift reactions whose conversion rates are depicted in FIG. 4. If the increased rate for the light reaction were strictly a localized heating effect, the activation energy for the light reaction versus the dark reaction should be the same. However, the dark reaction (.circle-solid.) has an E.sub.a of 47.09+/−0.27 kJ/mol, while the light reaction (.diamond-solid.) has an E.sub.a of 34.93+/−0.47 kJ/mol. The change in E.sub.a indicates that it not solely a localized heating effect that is responsible for the light-induced enhancement of the CO.sub.2 reduction rates.

(10) FIG. 9 is a graph depicting ln(reaction rate) for the CO.sub.2 to CO reduction as a function of the ln(partial pressure of CO.sub.2) for the light reaction (.diamond-solid.) versus the dark reaction (.circle-solid.). [Rate]=Kapp.sub.CO2 [PP.sub.CO2].sup.m. It was found for the light reaction that m.sub.light=1.04; for the dark reaction, m.sub.dark=0.50. Reaction conditions: Total gas flow rate=15 sccm, P=110 psi, T=200° C.

(11) FIG. 10 is a graph similar analogous to FIG. 9, but depicting ln(reaction rate) for the CO.sub.2 to CO reduction as a function of the ln(partial pressure of H.sub.2) for the light reaction (.diamond-solid.) versus the dark reaction (.circle-solid.). [Rate]=Kapp.sub.H2 [PP.sub.H2].sup.n. It was found for the light reaction that n.sub.light=0.17; for the dark reaction, n.sub.dark=0.07. Reaction conditions were the same as noted for FIG. 9.

(12) FIG. 11 is a graph depicting the dependence of light efficiency on H.sub.2:CO.sub.2 ratio in plasmon-enhanced water gas shift reaction over Au/TiO.sub.2 catalyst. Experimental conditions: P=103 psi, T=200° C., Total gas flow rate=15 sccm, catalyst amount=7.9 mg.

(13) FIG. 12 is a graph depicting the dependence of rate enhancement on H.sub.2:CO.sub.2 ratio in plasmon-enhanced water gas shift reaction over Au/TiO.sub.2 catalyst. Experimental conditions: P=103 psi, T=200° C., Total gas flow rate=15 sccm, catalyst amount=7.9 mg.

DETAILED DESCRIPTION

(14) Disclosed herein is a method of reducing CO.sub.2 to CO using hydrogen (H.sub.2) as the reducing agent, and using plasmonic photocatalysts and visible light (preferably solar light to increase the speed of the reaction to unprecedented rates. The method includes the steps of contacting the CO.sub.2 with the plasmonic photo catalyst, in the presence of H.sub.2. The plasmonic photocatalytic is then exposed to non-coherent radiation having a wavelength between about 380 nm and about 780 nm (that is, in the visible range) so that the catalyst undergoes surface plasmon resonance. It has been found that when using mixed catalysts comprising a nanoparticulate metal and a semiconductor, the surface plasmon resonance induced in the catalyst greatly increases the rate of CO.sub.2 reduction reaction.

(15) In particular, a catalyst comprising a noble metal nanoparticle (preferably gold) is fabricated via the sol-gel technique or deposition precipitation technique with an oxide semiconductor material, preferably a titania or alumina semiconductor. The Au/TiO.sub.2 (DP), Au/CeO.sub.2 (DP), Au/Al.sub.2O.sub.3 (DP) were prepared by deposition-precipitation (DP) method.sup.1-3. Degussa P25 TiO.sub.2 (Sigma-Aldrich, St. Louis, Mo., USA >99.5%), CeO.sub.2 (Sigma-Aldrich), Al.sub.2O.sub.3 (Strem Chemicals, Newburyport, Mass., USA) were used as supports, while HAuCl.sub.4.3H.sub.2O (Sigma-Aldrich) and CuSO.sub.4.5H.sub.2O (Sigman-Aldrich) were used as metal precursors for catalyst synthesis. The Au/TiO.sub.2 (SG) catalyst was prepared using sol-gel chemistry.sup.4. The Au/TiO.sub.2 (SG) solutions were then dried to obtain powdered Au/TIO.sub.2 (SG) catalyst. Au/Al.sub.2O.sub.3 (IVO) catalyst was prepared by incipient wetness impregnation. Cu/TiO.sub.2 (I) catalyst was prepared by impregnation (I) method.sup.5,6. The resulting photocatalytic material can then be used, in conjunction with light in the visible spectrum, to photocatalytically reduce CO.sub.2 in the presence of hydrogen via the reverse water gas shift reaction. The reverse water gas shift reaction produces a syn-gas mixture which can then be further converted to liquid fuels using mature existing technologies.

(16) The reverse water gas shift reaction, of course, is endothermic. Thus, the reaction needs to be driven. As described herein, it has been shown that metallic nano-particles absorb light radiation in the visible range. Thus, by coupling a suitable plasmonic catalyst comprising one or more nano-particulate metals that exhibit surface plasmon resonance (SPR) in response to light in the visible range of wavelengths (such as the photon found in solar radiation), solar radiation (a non-coherent radiation) can be used to drive the endothermic reverse water gas shift reaction. In this sense, the plasmonic response of the catalyst has a two-fold benefit: it both derives from solar energy the energy required for the reaction, and also catalyzes the reaction. For the reverse water gas shift reaction, where CO.sub.2 is being reduced to CO in the present of H.sub.2, the data presented herein show that the rate of reaction increases up to 13 times under simulated solar radiation as compared to the corresponding dark reaction. Thus, process is highly useful as a means to use the visible part of sunlight to drive chemical reactions.

(17) As used herein, the term “nanoparticle,” generally refers to a particle that exhibits one or more properties not normally associated with the corresponding bulk material (e.g., quantum optical effects such as surface plasmon resonance). The term also generally refers to materials having an average particle size no larger than about 100 nm. Nanoparticles include particles of any shape or geometry (spheres, rods, other crystalline and non-crystalline shapes, etc.), including individual nanoparticles and clusters of adhered nanoparticles. The nanoparticles can have a variety of shapes, dependent or independent, on their crystalline structure. The preferred nanoparticles for use in the process comprise calcium, copper, europium, gold, lithium, magnesium, palladium, platinum, potassium, silver, sodium, rubidium, and yttrium, and/or combinations thereof, mixtures thereof, and/or alloys containing these metals. The size and/or shape of a nanoparticle can be determined by transmission electron microscopy.

(18) Nanoparticles with well-controlled, highly-uniform sizes, and particle geometries can be fabricated using known techniques. Nanoparticles are widely available commercially from several worldwide suppliers, such as Sigma-Aldrich, St. Louis, Mo., USA. Various shapes of plasmonic nanoparticles can also be obtained by various methods such as those described in the U.S. Pat. No. 7,820,840. Some of these nanoparticles (e.g., metals with free-electron-like valence bands, such as noble metals) exhibit a strong localized surface plasmon resonance due to the nanometer scale spatial confinement, and the metal's inherent electronic structure. For example, the resonance frequency of silver and gold nanoparticles falls in the ultraviolet to visible light range, and can be tuned by changing the geometry and size of the particles. The intensity of resonant electromagnetic radiation is enhanced by several orders of magnitude near the surface of plasmonic nanoparticles. Thus, the catalysts described herein are compositions that exploit the ability of plasmonic nanoparticles to create electron-hole pairs, and simultaneously catalyze the reduction of CO.sub.2 to CO.

(19) Surface plasmon resonance (SPR) or simply plasmon resonance is an optical phenomenon arising from the collective oscillation of conduction electrons in a metal when the electrons are disturbed from their equilibrium positions. When electromagnetic energy (photons) of the proper energy impinge on such a metal, the free electrons of the metal are driven by the alternating electric field to coherently oscillate at a resonant frequency relative to the lattice of positive ions. The plasmon frequencies for most metals occur in the UV region of the electromagnetic spectrum. However several alkali metals and transition metals, including copper, silver, gold, and others have plasmon frequencies in the visible region of the spectrum. A “plasmonic nanoparticle,” therefore, is a nanoparticle having conduction electrons that collectively oscillate when excited by a stream of photon of the appropriate energy (i.e., wavelength).

(20) In the disclosed process, the plasmon resonance of the plasmonic catalyst is induced by non-coherent electromagnetic energy, preferably solar radiation. The solar radiation may be concentrated by any means or device now known or developed in the future. (A host of solar radiation concentrators are known in the art) The frequency and intensity of a plasmon resonance is generally determined by the intrinsic dielectric property of a given metal, the dielectric constant of the medium in contact with the metal, and the pattern of surface polarization. Thus, variations in the shape or size of the nanoparticulate metals in the catalyst can alter the surface polarization and cause a change to the plasmon resonance frequency. This dependence offers the ability to tune the surface plasmon resonance of metal nanoparticles through shape-controlled synthesis. A suitable shape-control synthesis is described in Lu et al. (2009) Annu. Rev. Phys. Chem. 60:167-92.

(21) The radiation applied comprises incoherent radiation in the visible range, approximately 380 nm to approximately 780 nm). The wavelengths of the photons that contact the catalyst may be full spectrum or otherwise attenuated by filters, monochromators, and the like.

(22) In various embodiments, the plasmon-resonating nanostructures include at least one of copper, silver, and gold nanoparticles. These nanoparticles may be copper/silver/gold alloy nanoparticles (e.g., copper-silver nanoparticles, copper-gold nanoparticles, silver-gold nanoparticles, copper-silver-gold nanoparticles). The nanostructures also may include, for example, silica as a core onto which the copper, silver and/or gold are deposited. In another variation, the nanostructures can be particles of substrates, for example silica, platinum, or other metal particles, onto which a plasmon-resonating layer or plasmon-resonating nanoparticle is deposited, e.g., layers or nanoparticles of Cu, Ag, and/or Au. In one preferred embodiment, the nano structures include copper. In another preferred embodiment, the nanostructures include silver. In yet another preferred embodiment, the nanostructures include gold.

(23) There are many advantages to using plasmonic catalysts for driving solar-powered chemical reactions. Notably, plasmonic catalysts, such as Au/TiO.sub.2, operate in the visible wavelength range of the solar radiation spectrum. This is an important consideration because 48% of the solar spectrum of radiation falls within the visible range, while only 6% falls within the ultraviolet range. Photocatalysts that operate only in the UV range are thus incredibly inefficient at converting solar energy into chemical energy. Thus, plasmonic catalysts that operate in the visible range of solar radiation provide higher efficiencies as compared to conventional, heterogeneous catalysts, as well as plasmonic catalysts that do not resonate in response to visible wavelengths of energy. Additionally, surface plasmon resonance itself depends on both the metal substrate selected and its particle size. The particle size dependency of SPR allows for the catalyst to be “tuned” or optimized over the visible range of wavelengths by adjusting the particle size accordingly. Additionally, there is no Shockley-Queisser limit on SPR. That is, the maximum theoretical efficiency of a p-n junction photovoltaic solar cell (as modeled by Shockley and Queisser) is a function of black-body radiation, e.sup.−/h.sup.+ pair recombination (i.e., the opposite of e.sup.−/h.sup.+ pair generation), and spectrum losses due to the wide range of wavelengths present in solar radiation. (That is, a significant portion of solar photons do not have the proper wavelength to generate e.sup.−/h.sup.+ pairs when they strike a PV panel.)

(24) The catalysts described herein are preferably fabricated using the sol-gel technique. This technique is well known to those skilled in the art, so it will not be described in exhaustive detail. Very briefly, in a typical sol-gel process, metal alkoxide and metal chloride precursors are solubilized to form a solution (sol) and then undergo hydrolysis and polycondensation reactions to form a colloid system composed of solid particles dispersed in a solvent. These solid particles continue to coalesce until they define an inorganic network containing a liquid phase (gel). The gel is then dried to remove the liquid phase, thereby yielding a highly porous material. Because of the high porosity, catalysts fabricated by the sol-gel technique typically have very high surface areas. In effect, solid nanoparticles dispersed in a liquid (a sol) agglomerate together to form a continuous three-dimensional network extending throughout the liquid (a gel). The liquid phase is then removed. The term “sol-gel” is sometimes improperly used as a noun to refer to gels made through the sol-gel process. See, for example, Brinker and Scherer, “The Physics and Chemistry of Sol-Gel Processing,” © 1990, Academic Press, Inc. San Diego, Calif., USA; ISBN 978-0-12-134970-7.

(25) Referring now to the figures, FIG. 1 is a schematic illustration showing how to reduce CO.sub.2 to CO (and other downstream products such as formic acid, syn-gas and hydrocarbons) using H.sub.2 produced from solar-powered photoelectrocatalytic (PEC) hydrolysis of water. Starting from the left-hand side of the figure, the box labeled “PV electricity generation” represents a conventional photovoltaic solar cell for producing electricity from sunlight. This electricity is then introduced into a photoelectrocatalytic reaction of water, along with additional, concentrated sunlight, as shown in the middle of FIG. 1, in the box labeled “PEC H.sub.2 production.” In this reactor, the water is split into H.sub.2 and O.sub.2 using a plamonic photoelectrochemical catalyst and solar radiation to induce the plasmonic resonance in the catalyst. Preferred catalysts for the water-splitting reaction include Au/TiO.sub.2 and Ag/TiO.sub.2, as well as Au, Ag, and/or Cu supported on other semiconductors. As noted previously, the preferred semiconductors are oxides of titanium, aluminum, iron, silicon, zinc, and/or cerium.

(26) The molecular hydrogen generated by the water-splitting reaction can then be used to drive the plasmonic photocatalytic reduction of CO.sub.2 (which can be obtained from a myriad of industrial processes, including any process involving the combustion of carbohydrates). This is shown in the box labeled “CO.sub.2+H.sub.2 conversion” in FIG. 1. As noted earlier, the plasmonic catalyst preferably comprises a metallic element have an average particle size no greater than 100 nm in combination with a semiconductor material. The metallic element must exhibit surface plasmon resonance, when the required average particle size range, in response to photons within the visible spectrum (about 380 to about 780 nm). The preferred nanoparticles comprise copper, mercury, ruthenium, rhodium, rhenium, palladium, silver, osmium, iridium, platinum, gold, and/or combinations thereof. The semiconductor is preferably an oxide of titanium and/or an oxide of aluminum. The radiation used to drive the plasmonic photocatalysis is preferably solar radiation.

(27) FIG. 2 is a schematic representation of the basic operation of plasmonic photocatalysis. When visible radiation induces plasmon resonance in the metallic nanoparticle (“Metal NP in FIG. 2) a number of quantum and macro phenomena occur, including intense scattering of the incoming radiation, electron/hole pair generation, and localized heating, all of which impact catalysis and can be harnessed to drive an endothermic reaction such as the reverse water-gas shift reaction.

(28) FIG. 3 is a histogram that demonstrates the considerable reaction rate enhancement that can be achieved using the plasmonic photocatalysis method described herein as compared to traditional heterogeneous catalysis. FIG. 3 is a histogram showing CO.sub.2 reduction rates for the reverse water gas shift reaction (CO.sub.2+H.sub.2CO+H.sub.2O; ΔH.sub.reaction=41 kJ/mol) using different catalysts in “dark” mode (no added light) and “enhanced” mode (in presence of simulated solar radiation that causes plasmon resonance in the catalysts). In the reaction, high-pressure hydrogen and CO.sub.2 (H.sub.2:CO.sub.2=2:1, 110 psi) were flowed into a sealed photocatalytic reaction vessel at 15 sccm and 400° C. Light reactions (using simulated solar radiation) and dark reactions were conducted for six different catalyst compositions, as noted in the figure. The results for the dark reactions are shown in black bars, with the reaction rate (μmol/gm-cat/min) depicted on the left-hand Y-axis. The enhancement of the reaction rate when exposed to simulated solar radiation is shown in open bars, with the enhancement factor depicted on the right-hand Y-axis. As can be seen from FIG. 3, CeO.sub.2, Al.sub.2O.sub.3, TiO.sub.2 on their own, in the dark, is a rather poor catalyst (relative to the others) for the reverse water-gas shift reaction. Even so, when run plasmonically, the reaction rate for TiO.sub.2 improved by almost 100%. Under these conditions, a catalyst comprised of nanoparticulate gold and TiO.sub.2 was a reasonably good catalyst when run in dark mode, and the dark rate was improved by almost 200% when the reaction was run plasmonically. For all of the catalyst combinations depicted in FIG. 3, the plasmonic enhancement of the reaction rate was significant. See also Table 1.

(29) TABLE-US-00001 TABLE 1 CO.sub.2 Conversion Rates under Various Conditions Catalyst CO.sub.2 conversion rate Amount (μmol/gm-cat/min) Enhancement Sr. No. Catalyst (mg) DARK LIGHT (LIGHT-DARK) (LIGHT-DARK) 1 Au—TiO.sub.2 (DP) 7.4 2033.4 2663.4 630.1 1.3 2 Au—CeO.sub.2 12.8 655.9 1416.6 760.8 2.2 (DP) 3 Au—TiO.sub.2 (SG) 12.4 641.2 900.4 259.2 1.4 4 Au-A1.sub.2O.sub.3 16.6 76.5 118.3 41.8 1.5 (DP) 5 Au-A12O.sub.3 13.1 47.3 71.6 24.4 1.5 (IWI) 6 Cu—TiO.sub.2 (I) 8.6 19.5 25.1 5.7 1.3 7 TiO.sub.2 12.2 21.2 18.9 −2.3 0.9 8 CeO.sub.2 23.9 21.0 22.0 1.1 1.1 9 Al2O.sub.3 30.00 67.4 73.8 6.4 1.2

(30) In light of these results, using the reverse water gas shift reaction over an Au/TiO.sub.2 catalyst run plasmonically and in the dark as a means to reduce CO.sub.2 with H.sub.2 was investigated in greater detail. FIG. 4 is a graph that depicts the rate of the reduction reaction as a factor of temperature for both the plasmonic reactions (“light”, .diamond-solid.) and dark reactions, .circle-solid.. Of particular note in FIG. 4 is the enhancement of the plasmonic reaction rates across all temperatures tested. As the temperature rises, the reaction rate predictably rises. The reaction is endothermic, so its rate would be expected to rise with rising temperature. However, the enhancement due to running the reaction plasmonically is not expected, especially at the higher end of the temperature range. That is, at the higher end of the temperature range, the expectation is that the thermal effect on catalysis would dominate and the enhancement due to running the reaction plasmonically would decrease or disappear entirely. However, even at the highest temperature tested, 400° C., FIG. 4 shows that there is a very significant enhancement in the reaction rate between the light reaction and the dark reaction.

(31) FIG. 5 presents the enhancement data between the light reaction versus the dark reaction in isolation—i.e., it is a graph depicting the enhancement in rate due to running the reaction plasmonically as a function of temperature for the reverse water gas shift reactions described above for FIG. 4. Here, the data show that in a direct comparison, the enhancement factor (i.e., the rate of light reaction/rate of dark reaction) is more pronounced at 100° C. and decreases in a smooth curve to approximately a factor of 2 at 400° C. Extrapolated, these data indicate a light enhancement of a factor of 7; i.e., 700%. These same date are presented in FIG. 6 not as a rate enhancement, but rather as the actual difference between CO.sub.2 reduction rate (μmol/gm-cat/min) under light and dark vs. temperature that can be attributed solely to the plasmonic influence of the catalyst (and not temperature). FIG. 6 indicates that the maximum plasmonic-induced enhancement in the reaction rate as a function of temperature-induced increases in reaction rate peaks somewhere between about 300° C. and about 350° C. in the Au/TiO.sub.2 system. FIG. 7 corroborates these findings by showing that the light efficiency versus temperature for this same reaction also reaches a peak between about 300° C. and about 350° C. In FIG. 7, light efficiency is defined as

(32) $\mathrm{Light}\mathrm{efficiency}\left(\%\right)=\frac{{\mathrm{CO}}_2\mathrm{conversion}\mathrm{rate}\mathrm{due}\mathrm{to}\mathrm{light}\times \Delta H_{\mathrm{reaction}}}{\mathrm{Intensity}\times \mathrm{Catalyst}\mathrm{surface}\mathrm{area}}\times 100\%$

(33) The salient point of FIGS. 4 through 7 taken together is that the visible light energy that induces plasmonic activity in the catalyst is the cause of a very marked increase in the reaction rate of the reverse water-gas shift reaction. The enhancement is achieved using simulated, non-coherent solar radiation.

(34) Now, it could be possible that the enhanced catalytic effect is not a photocatalytic effect, per se, but simply a thermal effect due to localized heating caused by the surface plasmon resonance. To investigate this possibility, an Arrhenius plot (ln(rate) v 1/T) was constructed for the light reactions described above for FIGS. 4-7 and the corresponding dark reactions. The results are shown in FIG. 8. Thus, FIG. 8 is a plot depicting ln(CO.sub.2 reduction rate) versus 1/Temp(1/K)×10.sup.3 for the reverse water gas shift reactions. The plots for the light reaction versus the dark reaction clearly show different activation energies. If the increased rate for the light reaction were strictly a localized heating effect, the activation energy for the light reaction versus the dark reaction should be the same. However, the dark reaction (.diamond-solid.) has an E.sub.a of 47.09+/−0.27 kJ/mol, while the light reaction (.circle-solid.) has an E.sub.a of 34.93+/−0.47 kJ/mol. The change in E.sub.a indicates that it not solely a localized heating effect that is responsible for the light-induced enhancement of the CO.sub.2 reduction rates.

(35) FIGS. 9 and 10 are corresponding plots that map the ln(rate of CO.sub.2 reduction) versus the ln(partial pressure of CO.sub.2) (FIG. 9) and ln(rate) versus the ln(partial pressure of H.sub.2) (FIG. 10) for the light (.diamond-solid.) and dark (.circle-solid.) reactions. In both figures, the reaction conditions were identical: Total gas flow rate=15 sccm, P=110 psi, T=200° C., H.sub.2:CO.sub.2=2:1. In FIG. 9, the rate equation sets up as [Rate]=Kapp.sub.CO2 [PP.sub.CO2].sup.m. Thus, the exponent “m” is the reaction order and its value is dependent upon the mechanism that causes the CO.sub.2 reduction. In FIG. 9, which is the data based on the partial pressure of CO.sub.2, it was found for the light reaction that m.sub.light=1.04; for the dark reaction, m.sub.dark=0.50. These data clearly indicate that there is a distinctly different reaction mechanism for the “light,” plasmonically catalyzed reaction as compared to the dark reaction.

(36) The same holds true when ln(rate) versus the ln(partial pressure of H.sub.2) is plotted for the light reaction versus the dark reaction. See FIG. 10. Here, the rate equation sets up as [Rate]=Kapp.sub.H2 [PP.sub.H2].sup.n. It was found for the light reaction that n.sub.light=0.17; for the dark reaction, n.sub.dark=0.07.

(37) FIGS. 11 and 12 is a graph depicting the dependence of light efficiency and rate enhancement on H.sub.2:CO.sub.2 ratio in plasmon-enhanced water gas shift reaction over Au/TiO.sub.2 catalyst. Experimental conditions: P=103 psi, T=200° C., Total gas flow rate=15 sccm, catalyst amount=7.9 mg. As can be seen in FIG. 11, lower the ratio of H.sub.2:CO.sub.2 in the plasmonically catalyzed reaction results in the higher light efficiency of the reaction. That is, at high light efficiencies, the reaction produced increased amounts of H.sub.2 as compared to CO.sub.2. FIG. 12 shows that at low H.sub.2:CO.sub.2 ratio, plasmonic rate enhancement up to 1300% can be achieved.

(38) Suitable catalysts for use in the present method may be fabricated by the following methods. Note that these methods are exemplary and are included solely to provide a more complete disclosure of the method claimed herein. The exemplary catalysts are not limiting.

(39) Preparation of Au/TiO.sub.2 (DP) Catalyst:

(40) The Au/TiO.sub.2 DP was prepared by deposition-precipitation with NaOH (1M).sup.1,2. Titania Degussa P25 was used as support (Sigma-Aldrich, >99.5% trace metal basis) and solid HAuCl.sub.4.3H.sub.2O (Sigma-Aldrich, >99.9% trace metal basis) as the precursor. Before the preparation, TiO.sub.2 was dried in the air at 110° C. overnight. 100 ml of aqueous HAuCl.sub.4 solution (4.2*10.sup.−3 M) was heated to 80° C. and the pH was adjusted to 8 by drop-wise addition of NaOH (1M). Then, 1 g of TiO.sub.2 was dispersed in the solution, and the pH was readjusted to 8 with NaOH. The suspension was thermostated at 80° C. was stirred for 2 h and centrifuged. The solids were then washed, dried, and calcined at 300° C. under the flow of air (30 ml/min) with a heating rate of 2° C./min and maintained for 4 h.

(41) Preparation of Au/CeO.sub.2 DP Catalyst:

(42) The Au/CeO.sub.2 (DP).sup.5 was prepared by deposition-precipitation with NaOH (1M) which is same with Au/TiO.sub.2 (DP).sup.1,2. Cerium (IV) oxide was used as support (Sigma-Aldrich) and solid HAuCl.sub.4.3H.sub.2O (Sigma-Aldrich, >99.9% trace metal basis) as the precursor. Before the preparation, CeO.sub.2 was dried in the air at 110° C. overnight. 100 ml of aqueous HAuCl.sub.4 solution (4.2*10.sup.−3 M) was heated to 80° C. and the pH was adjusted to 8 by drop-wise addition of NaOH (1M). Then, 1 g of CeO.sub.2 was dispersed in the solution, and the pH was readjusted to 8 with NaOH. The suspension was thermostated at 80° C. was stirred for 2 h and centrifuged. The solids were then washed, dried, and calcined at 300° C. under the flow of air (30 ml/min) with a heating rate of 2° C./min and maintained for 4 h.

(43) Preparation of Au/Al.sub.2O.sub.3 (DP) Catalyst:

(44) The Au/Al.sub.2O.sub.3 (DP).sup.3 was prepared by deposition-precipitation with NaOH (1M) which is same with Au/TiO.sub.2 (DP).sup.1,2. Alumina was used as support (Strem Chemicals) and solid HAuCl.sub.4.3H.sub.2O (Sigma-Aldrich, >99.9% trace metal basis) as the precursor. Before the preparation, Al.sub.2O.sub.3 was dried in the air at 110° C. overnight. 100 ml of aqueous HAuCl.sub.4 solution (4.2*10.sup.−3 M) was heated to 80° C. and the pH was adjusted to 8 by drop-wise addition of NaOH (1M). Then, 1 g of Al.sub.2O.sub.3 was dispersed in the solution, and the pH was readjusted to 8 with NaOH. The suspension was thermostated at 80° C. was stirred for 2 h and centrifuged. The solid was then washed, dried, and calcined at 300° C. under the flow of air (30 ml/min) with a heating rate of 2° C./min and maintained for 4 h.

(45) Preparation of Cu/TiO.sub.2 (I) Catalyst:

(46) The Cu/TiO.sub.2 (I) was prepared by impregnating 1 g of titania Degussa P25 (Sigma-Aldrich, >99.5% trace metal basis) with a solution of 53 mg of CuSO.sub.4.5H.sub.2O (Sigma-Aldrich, puriss, meets analytical specification of Ph. Eur., BP, USP, 99-100.5%) in 10 ml of DI water.sup.5,6. The slurry was stirred for 4 h at room temperature, then all liquid was evaporated and the solid was dried at 110° C. overnight. The catalyst was calcined at 300° C. under the flow of air (30 ml/min) with a heating rate of 2° C./min and maintained for 4 h.

REFERENCE CITED

(47) 1. R. Zanella, S. Giorgio, C. H. Shin, C. R. Henry and C. Louis, J. Catal., 2004, 222, 357-367. 2. R. Zanella, S. Giorgio, C. R. Henry and C. Louis, J. Phys. Chem. B, 2002, 106, 7634-7642. 3. D. L. Nguyen, S. Umbarkar, M. K. Dongare, C. Lancelot, J. S. Girardon, C. Dujardin and P. Granger, Catal. Commun., 2012, 26, 225-230. 4. C. Burda, X. Chen, R. Narayanan and M. A. El-Sayed, Chem. Rev., 2005, 105, 1025-1102. 5. F. Sastre, M. Oteri, A. Corma and H. Garcia, Energy Environ. Sci., 2013, 6, 2211-2215. 6. F. Boccuzzi, A. Chiorino, G. Martra, M. Gargano, N. Ravasio and B. Carrozzini, J. Catal., 1997, 165, 129-139.