Photoelectrochemical reduction of CO2

Background/Introduction

Semiconductor material has a band gap and generates a pair of electron and hole per absorbed photon if the energy of photon is higher than band gap of semiconductor. This property of semiconductor materials has been successfully used to convert solar energy into electrical energy by photovoltaic devices. So, semiconductor electrodes can be used for CO₂ photoelectrochemical reduction.

When a semiconductor comes into contact with a liquid (redox species), to maintain electrostatic equilibrium, there will be a charge transfer between the semiconductor and liquid phase if formal redox potential of redox species lies inside semiconductor band gap. At thermodynamic equilibrium, the Fermi level of semiconductor and the formal redox potential of redox species are aligned at the interface between semiconductor and redox species. This introduces a downward band bending in a n-type semiconductor for n-type semiconductor/liquid junction (Figure 1) and upward band bending in a p-type semiconductor for p-type semiconductor/liquid junction (Figure 2). This characteristic of semiconductor/liquid junction is similar to a rectifying semiconductor/metal junction or Schottky junction. Ideally to get a good rectifying characteristics at the semiconductor/liquid interface, the formal redox potential must be close to the valence band of the semiconductor for a n-type semiconductor and close to the conduction band of the semiconductor for a p-type semiconductor. The semiconductor/liquid junction has one advantage over rectifying semiconductor/metal junction in that the light is able to travel through to semiconductor surface without much reflection; whereas most of the light is reflected back from metal surface at semiconductor/metal junction. Therefore, semiconductor/liquid junction can also be used a photovoltaic devices similar to solid state p–n junction devices. Both n-type and p-type semiconductor/liquid junction can be used as photovoltaic devices to convert solar energy into electrical energy and are called photoelectrochemical cell. In addition, a semiconductor/liquid junction could also be used to directly convert solar energy into chemical energy by virtue of photoelectrolysis at the semiconductor/liquid junction.

The catalytic conversion of CO₂ to liquid fuels is a critical goal that would positively impact the global carbon balance by recycling CO₂ into usable fuels. The challenges presented here are great, but the potential rewards are enormous. CO₂ is extremely stable molecule generally produced by fossil fuel combustion and respiration. Returning CO₂ to a useful state by activation/reduction is a scientifically challenging problem, requiring appropriate catalysts and energy input. This poses several fundamental challenges in chemical catalysis, electrochemistry, photochemistry, and semiconductor physics and engineering.

Thermodynamic challenges

Thermodynamic potentials for the reduction of CO₂ to various products is given in the following table versus NHE at pH = 7. Single electron reduction of CO₂ to CO₂^●− radical occurs at E° = −1.90 V versus NHE at pH = 7 in an aqueous solution at 25 °C under 1 atm gas pressure. The reason behind the high negative thermodynamically unfavorable single electron reduction potential of CO₂ is the large reorganization energy between the linear molecule and bent radical anion. Proton-coupled multi-electron steps for CO₂ reductions are generally more favorable than single electron reductions, as thermodynamically more stable molecules are produced.

Kinetic challenges

Thermodynamically, proton coupled multiple-electron reduction of CO₂ is easier than single electron reduction. But to manage multiple proton coupled multiple-electron processes is a huge challenge kinetically. This leads to a high overpotential for electrochemical heterogeneous reduction of CO₂ to hydrocarbons and alcohols. Even further heterogeneous reduction of singly reduced CO₂^●− radical anion is difficult because of repulsive interaction between negatively biased electrode and negatively charged anion.

Is it feasible to photoelectrochemically reduce CO2 on semiconductor surface?

Figure 1(b) shows that in case of a p-type semiconductor/liquid junction photo generated electrons are available at the semiconductor/liquid interface under illumination. The reduction of redox species happens at less negative potential on illuminated p-type semiconductor compared to metal electrode due to the band bending at semiconductor/liquid interface. Figure 3 shows that thermodynamically, some of the proton-coupled multi-electron CO₂ reductions are within semiconductors band gap. This makes it feasible to photo-reduce CO₂ on p-type semiconductors. Various p-type semiconductors have been successfully employed for CO₂ photo reduction including p-GaP, p-CdTe, p-Si, p-GaAs, p-InP, and p-SiC. Kinetically, however, these reactions are extremely slow on given semiconductor surfaces; this leads to significant overpotential for CO₂ reduction on these semiconductor surfaces. Apart from high overpotential; these systems have a few advantages including sustainability (nothing is consumed in this system apart from light energy), direct conversion of solar energy to chemical energy, utilization of renewable energy resource for energy intensive process, stability of the process (semiconductors are really stable under illumination) etc. A different approach for photo-reduction of CO₂ involves molecular catalysts, photosensitizers and sacrificial electron donors. In this process sacrificial electron donors are consumed during the process and photosensitizers degrade under long exposure to illumination.

Solvent effect

The photo-reduction of CO₂ on p-type semiconductor photo-electrodes has been achieved in both aqueous and non-aqueous media. Main difference between aqueous and non-aqueous media is the solubility of CO₂. The solubility of CO₂ in aqueous media at 1 atm. of CO₂ is around ≈ 35 mM; whereas solubility of CO₂ in methanol is around 210 mM and in acetonitrile is around 210 mM.

Aqueous media

Halmann had first shown CO₂ photoreduction to formic acid on p-GaP as photocathode in aqueous media in 1978. Apart from several other reports of CO₂ photoreduction on p-GaP, there are other p-type semiconductors like p-GaAs, p-InP, p-CdTe, and p⁺/p-Si have been successfully used for photoreduction of CO₂. The lowest potential for CO₂ photoreduction was observed on p-GaP. This may be due to high photovoltage excepted from higher band gap p-GaP (2.2 eV) photocathode. Apart from formic acid, other products observed for CO₂ photoreduction are formaldehyde, methanol and carbon monoxide. On p-GaP, p-GaAs and p⁺/p-Si photocathode, the main product is formic acid with small amount of formaldehyde and methanol. However, for p-InP and p-CdTe photocathode, both carbon monoxide and formic acid are observed in similar quantities. Mechanism proposed by Hori based on CO₂ reduction on metal electrodes predicts formation of both formic acid (in case of no adsorption of singly reduced CO₂^●− radical anion to the surface) and carbon monoxide (in case of adsorption of singly reduced CO₂^●− radical anion to the surface) in aqueous media. This same mechanism can be evoked to explain the formation of mainly formic acid on p-GaP, p-GaAs and p⁺/p-Si photocathode owing to no adsorption of singly reduced CO₂^●− radical anion to the surface. In case of p-InP and p-CdTe photocathode, partial adsorption of CO₂^●− radical anion leads to formation of both carbon monoxide and formic acid. Low catalytic current density for CO₂ photoreduction and competitive hydrogen generation are two major drawbacks of this system.

Non-aqueous media

Maximum catalytic current density for CO₂ reduction that can be achieved in aqueous media is only 10 mA cm⁻² based solubility of CO₂ and diffusion limitations. The integrated maximum photocurrent under Air Mass 1.5 illumination, in the conventional Shockley-Quiesser limit for solar energy conversion for p-Si (1.12 eV), p-InP (1.3 eV), p-GaAs (1.4 eV), and p-GaP (2.3 eV) are 44.0 mA cm⁻², 37.0 mA cm⁻², 32.5 mA cm⁻² and 9.0 mA cm⁻², respectively. Therefore, non-aqueous media such as DMF, acetonitrile, methanol are explored as solvent for CO₂ electrochemical reduction. In addition, Methanol has been industrially used as a physical absorber of CO₂ in the Rectisol method. Similarly to aqueous media system, p-Si, p-InP, p-GaAs, p-GaP and p-CdTe are explored for CO₂ photoelectrochemical reduction. Among these, p-GaP has lowest overpotential, whereas, p-CdTe has moderate overpotential but high catalytic current density in DMF with 5% water mixture system. Main product of CO₂ reduction in non-aqueous media is carbon monoxide. Competitive hydrogen generation is minimized in non-aqueous media. Proposed mechanism for CO₂ reduction to CO in non-aqueous media involves single electron reduction of CO₂ to CO₂^●− radical anion and adsorption of radical anion to surface followed by disproportionate reaction between unreduced CO₂ and CO₂^●− radical anion to form CO₃²⁻ and CO.