**3. Artificial photosynthesis**

Since the pioneering work of Inoue et al., 1979, CO2 is being reduced with H2O photocatalytically to mainly one carbon molecules like methane and methanol in the

Artificial Photosynthesis from a Chemical Engineering Perspective 23

���� + �ℎ� � �� + ���

��� + ��� + ��� � ����� + ���

��� + ��� + ��� � ��� + ���� In photocatalysis; surface adsorbed species should have appropriate redox potentials regarding flat band positions of the semiconductor for thermodynamic favorability of the reactions, or vice versa. It required that; semiconductors should have conduction bands located at a more negative potential than the reduction potential of CO2 to hydrocarbons, and valence bands located at a more positive potential than the oxidation potential of H2O. In Table 3.1 oxidation and reduction reactions taking place in photocatalytic CO2 reduction is listed with their electro-potentials at pH=7, vs NHE. In order to provide thermodynamic favorability, large band gap semiconductors such as TiO2 are mostly utilized in photocatalytic CO2 reduction reactions which render UV light illumination obligatory for photo-activation of the catalysts. Visible light utilization in carbon dioxide reduction is the ultimate goal in photocatalytic studies for a complete carbon free energy generation. There are material modification studies in literature conducted for efficient visible light utilization

�� �� �� + ℎ�

�������������

such as dye sensitization (Ozcan et al., 2007) and anion doping (Asahi et al., 2001).

Table 3.1. Half cell reactions and their electro-potentials at pH=7 vs NHE (Jitaru, 2007)

Presence of that much of water on TiO2 surface would inhibit CO2 activation (CO2+e-

indicating that surface is mainly covered with water.

E0

(Ti+3-O-

One disadvantage of the realization of reduction and oxidation reactions at the same surface is the interaction between reactants on the surface. In one of the gas phase photocatalytic CO2 reduction experiments realized on Cu/TiO2 surfaces, when Langmuir- Hinshelwood surface reaction mechanism was selected with competitive adsorption of H2O and CO2, it was found that adsorption constant of H2O dominates that of CO2 (Wu et al., 2005)

redox= -1.9 V vs. NHE at pH 7), which is considered as the essential step in CO2 reduction (Solymosi & Tombacz, 1994), by oxidizing defect structure of CO2. Electron affinity of CO2 molecule is related to the position of lowest unoccupied molecular orbital of CO2 and conduction band of TiO2, assuming that electron is transferred from excited state of TiO2

) to CO2. A decrease in lowest unoccupied molecular orbital (LUMO) of CO2 was reported with lower bond angles that could result from the interaction of the molecule with the surface (Freund & Roberts, 1996). According to Indrakanti et al., 2009, CO2 gains electrons from oxygen deficient TiO2 via the formation of bent CO2 molecules near Ti+3 sites, whereas electron transfer is not favorable with defect free TiO2 due to high LUMO of CO2.

**Reactions E0 (V)** 

→CO2•-

���� + �ℎ� � �� + ��� +0.82 ��� + ��� + ��� � ����� -0.61 ��� + ��� + ��� � �� + ��� -0.52 ��� + ��� + ��� � ���� + ��� -0.48 ��� + ��� + ��� � ����� + ��� -0.38 ��� + ��� + ��� � ��� + ���� -0.24

presence of a photo-activated semiconductor. Electrons, generated upon illumination of the semiconductor, are trapped at the electron trap centers and utilized directly in reduction centers without a complex transportation system involving intermediate charge carriers. Similarly holes, generated upon illumination are utilized in oxidation reactions on the catalyst surface. Since photocatalysis lacks a specialized transportation system for generated electrons and holes, majority of the charge carriers recombine at the catalyst surface or in the bulk volume of the catalyst, lowering photocatalytic rates. (Figure 3.1)

Semiconductors, having a band gap, ensure a life-time for generated electrons and holes; however, this lifetime is limited to 10-7 s, which is the characteristic time of recombination (for bare TiO2) (Carp et al., 2004). In order to increase this lifetime of photo-generated electrons and holes, some modifications on materials such as metal addition to semiconductors (Anpo et al., 1997, Tseng et al., 2002, Ozcan et al., 2007) or formation of solid-solid interfaces in composite catalysts (Chen et al., 2009, Woan et al., 2009) were proposed in literature.

Metal addition to semiconductors is suggested to show charge separation effect on photocatalysis by the Schottky Barrier Formation. When metals are brought into contact with semiconductors, electrons populate on metals if Fermi level of the metal is lower than the conduction band of the semiconductor. Hence, metals act like 'charge carrier traps', increasing lifetime of electron hole pairs with charge separation effect.

Fig. 3.1. Illustrating scheme of electron/hole pair generation and realization of redox reactions in photocatalysis

The other modification that can hinder recombination of generated electrons and holes is formation of solid-solid interfaces in composite photocatalysts having different band gap energies. To illustrate; commercial TiO2 catalysts; Degussa P-25, is composed of anatase and rutile crystal phases of TiO2, having band gap energies of 3.2 eV and 3.0 eV respectively. Mixed phase TiO2, tends to exhibit higher photocatalytic activity than pure phases, because it allows transfer of the photogenerated electron from rutile to anatase, resulting in charge separation (Carp et al., 2004, Chen et al., 2009).

In photocatalysis, reduction and oxidation reactions occur at similar sites on the same catalyst surface. There are no different reaction centers with certain distances apart.

The half reactions of the photosynthesis; water oxidation and CO2 reduction, are also realized in photocatalysis.

presence of a photo-activated semiconductor. Electrons, generated upon illumination of the semiconductor, are trapped at the electron trap centers and utilized directly in reduction centers without a complex transportation system involving intermediate charge carriers. Similarly holes, generated upon illumination are utilized in oxidation reactions on the catalyst surface. Since photocatalysis lacks a specialized transportation system for generated electrons and holes, majority of the charge carriers recombine at the catalyst surface or in the

Semiconductors, having a band gap, ensure a life-time for generated electrons and holes; however, this lifetime is limited to 10-7 s, which is the characteristic time of recombination (for bare TiO2) (Carp et al., 2004). In order to increase this lifetime of photo-generated electrons and holes, some modifications on materials such as metal addition to semiconductors (Anpo et al., 1997, Tseng et al., 2002, Ozcan et al., 2007) or formation of solid-solid interfaces in composite catalysts (Chen et al., 2009, Woan et al., 2009) were

Metal addition to semiconductors is suggested to show charge separation effect on photocatalysis by the Schottky Barrier Formation. When metals are brought into contact with semiconductors, electrons populate on metals if Fermi level of the metal is lower than the conduction band of the semiconductor. Hence, metals act like 'charge carrier traps',

Fig. 3.1. Illustrating scheme of electron/hole pair generation and realization of redox

The other modification that can hinder recombination of generated electrons and holes is formation of solid-solid interfaces in composite photocatalysts having different band gap energies. To illustrate; commercial TiO2 catalysts; Degussa P-25, is composed of anatase and rutile crystal phases of TiO2, having band gap energies of 3.2 eV and 3.0 eV respectively. Mixed phase TiO2, tends to exhibit higher photocatalytic activity than pure phases, because it allows transfer of the photogenerated electron from rutile to anatase, resulting in charge

In photocatalysis, reduction and oxidation reactions occur at similar sites on the same

The half reactions of the photosynthesis; water oxidation and CO2 reduction, are also

catalyst surface. There are no different reaction centers with certain distances apart.

bulk volume of the catalyst, lowering photocatalytic rates. (Figure 3.1)

increasing lifetime of electron hole pairs with charge separation effect.

proposed in literature.

reactions in photocatalysis

realized in photocatalysis.

separation (Carp et al., 2004, Chen et al., 2009).

������������� �� �� �� + ℎ� ���� + �ℎ� � �� + ��� ��� + ��� + ��� � ����� + ��� ��� + ��� + ��� � ��� + ����

In photocatalysis; surface adsorbed species should have appropriate redox potentials regarding flat band positions of the semiconductor for thermodynamic favorability of the reactions, or vice versa. It required that; semiconductors should have conduction bands located at a more negative potential than the reduction potential of CO2 to hydrocarbons, and valence bands located at a more positive potential than the oxidation potential of H2O. In Table 3.1 oxidation and reduction reactions taking place in photocatalytic CO2 reduction is listed with their electro-potentials at pH=7, vs NHE. In order to provide thermodynamic favorability, large band gap semiconductors such as TiO2 are mostly utilized in photocatalytic CO2 reduction reactions which render UV light illumination obligatory for photo-activation of the catalysts. Visible light utilization in carbon dioxide reduction is the ultimate goal in photocatalytic studies for a complete carbon free energy generation. There are material modification studies in literature conducted for efficient visible light utilization such as dye sensitization (Ozcan et al., 2007) and anion doping (Asahi et al., 2001).



One disadvantage of the realization of reduction and oxidation reactions at the same surface is the interaction between reactants on the surface. In one of the gas phase photocatalytic CO2 reduction experiments realized on Cu/TiO2 surfaces, when Langmuir- Hinshelwood surface reaction mechanism was selected with competitive adsorption of H2O and CO2, it was found that adsorption constant of H2O dominates that of CO2 (Wu et al., 2005) indicating that surface is mainly covered with water.

Presence of that much of water on TiO2 surface would inhibit CO2 activation (CO2+e-→CO2•- E0 redox= -1.9 V vs. NHE at pH 7), which is considered as the essential step in CO2 reduction (Solymosi & Tombacz, 1994), by oxidizing defect structure of CO2. Electron affinity of CO2 molecule is related to the position of lowest unoccupied molecular orbital of CO2 and conduction band of TiO2, assuming that electron is transferred from excited state of TiO2 (Ti+3-O- ) to CO2. A decrease in lowest unoccupied molecular orbital (LUMO) of CO2 was reported with lower bond angles that could result from the interaction of the molecule with the surface (Freund & Roberts, 1996). According to Indrakanti et al., 2009, CO2 gains electrons from oxygen deficient TiO2 via the formation of bent CO2 molecules near Ti+3 sites, whereas electron transfer is not favorable with defect free TiO2 due to high LUMO of CO2.


$$\begin{aligned} CO\_2 + 3H\_2 &\leftrightarrow \cdot CH\_3OH + H\_2O \\ CO + H\_2O &\leftrightarrow \cdot CO\_2 + H\_2 \end{aligned}$$



$$\frac{A f}{A r} = \frac{z\_{\mathbb{C}} z\_{\mathbb{B}}}{z\_{\mathbb{A}} z\_{\mathbb{B}}} = \exp\left(\frac{\mathbb{A} \mathbb{S}^0}{R}\right) \text{ for the reaction } A + B \to \mathbb{C} + D \tag{2}$$

Artificial Photosynthesis from a Chemical Engineering Perspective 29

concentration on the surface for methanol formation rates. Since water splitting reaction is the only source of H in photocatalytic CO2 reduction mechanism, it could be said that water oxidation rates are rate limiting in artificial photosynthesis systems whereas water oxidation

Analogy between photosynthesis and artificial photosynthesis is in the similar tools and methods utilized in both systems. Collecting solar energy for triggering chemical reactions by chlorophyll pigments packed in thylakoid membrane or by semiconductors; oxidizing water into molecular oxygen and protons and reducing CO2 with transported electrons and H+s are among the similarities of the two systems. However, the gap between the design of the systems and number of reaction sites and intermediate molecules result in more sophisticated and simpler products in photosynthesis ((CH2O)6) and in photocatalysis (CH4

In photosynthesis, there are three major reaction centers in light dependent reactions, regulating electron and proton transport together with the intermediate charge carriers (redox components). In photocatalysis, on the other hand, design of the system is limited to the presence of a pool of charges wandering on the semiconductor/metal surface in an unregulated fashion, increasing the chance of recombination of charge carriers. In addition, realization of oxidation and reduction reactions on the same catalyst surface results in interactions between the surface adsorbates which is proven to be inhibitory on reaction

In photosynthesis, CO2 diffusion from atmosphere to stromal phase in chloroplasts is controlled by stomata activities and permeability of chloroplast membranes. Photosynthetic rate is limited with the CO2 concentration in stromal phase for values lower than a saturation value; i.e., the photosynthetic rate is linearly increasing with CO2 concentration. For CO2 concentrations above the saturation value, photosynthetic rate stays constant, limited by the rate of the enzyme system. Since CO2 concentration in the stromal phase is

In photocatalysis, diffusion of dissolved CO2 and other reactants/products to/from the catalyst surface is largely dependent upon the reactor types, reaction media and stirring rates. The photocatalytic experiment parameters are not standardized in literature, resulting in confusion about the proper comparison of the real kinetic data. For photocatalytic tests performed in liquid media, which constitute the majority of the tests reported in literature, presence of gas-liquid-solid interfaces imposes non negligible mass transfer limitations in observed rates. A study performed to reveal the effect of stirring rates on photocatalytic hydrogen evolution rates indicated the importance of boundary layer and gas-liquid

Increase in photo catalytic rates with increasing stirring rates (from 350 rpm to 900 rpm) up to a certain hydrogen concentration could be interpreted as decreased mass transfer limitation effects due to thinning of the boundary layer surrounding the catalyst particles whereas after that concentration, hydrogen seems to accumulate in the gas phase with the same limiting liquid-gas transfer rate. The limitation at the gas-liquid interface could also be inferred from the overlapping hydrogen amounts accumulated in the gas phase regardless

related to CO2 diffusion, photosynthetic rate is dependent upon diffusion rates.

equilibrium in liquid phase photocatalytic experiments (Figure 4.1) (Ipek, 2011).

of the catalyst amount or the reaction mixture volume (Figure 4.2).

rate surpass carbon dioxide reduction rates at photosynthesis (Table 2.1).

rates as studied with the microkinetic model in Section 3.1.

**photosynthesis** 

or CH3OH) respectively.

**4. Similarities and differences between photosynthesis and artificial** 


Table 3.5. Methanol formation rates (mol/gcat/s) at different reaction conditions with or without initial water concentration

### **3.2 Rate determining step of methanol formation on Cu (111) surface**

Calculation of degree of rate controls for elementary reaction steps in the microkinetic model allows revealing rate limiting steps in methanol formation from CO2 hydrogenation and water gas shift reaction. Degree of rate control is defined by Campbell such as; the degree of change of the overall rate by a change in rate constant of a single step (Equation 2) (Campbell, 2001). Campbell proposed that steps where degree of rate control is positive be called rate-limiting steps and negative be inhibition steps. The larger the numeric value of degree of rate control, *X*rc,i, the bigger is the influence of its rate constant on the overall reaction rate.

$$X\_{\rm rc,i} = (\mathbf{k}\_i / \mathbf{\hat{8}k})^\* (\mathbf{\hat{8}R/R}) \tag{3}$$

When degree of rate control values were calculated for the microkinetic model of methanol synthesis on Cu (111) surface, the results indicated that H supply (Step 2) to Cu surface as well as formate hydrogenation step (step 7) is essential especially at 300 K, at which artificial photosynthesis occurs (Table 3.6). This study underlines the importance of H supply and


Table 3.6. The degree of rate control values with respect to rf10 (����� ∗� ����� +∗) found by finite difference method at t= 5.18\*10-7 s

**With initial H2O** 

75 bar 4.53\*10-5 3.58\*10-5 6.1\*10-5 (with Cu/ZnO/Al2O3)

1 bar 1.9\*10-11 2.07\*10-12 Photocatalytic: 6\*10-9 (with

Calculation of degree of rate controls for elementary reaction steps in the microkinetic model allows revealing rate limiting steps in methanol formation from CO2 hydrogenation and water gas shift reaction. Degree of rate control is defined by Campbell such as; the degree of change of the overall rate by a change in rate constant of a single step (Equation 2) (Campbell, 2001). Campbell proposed that steps where degree of rate control is positive be called rate-limiting steps and negative be inhibition steps. The larger the numeric value of degree of rate control, *X*rc,i, the bigger is the influence of its rate constant on the overall

 *X*rc,i = (ki/δki)\*(δR/R) (3) When degree of rate control values were calculated for the microkinetic model of methanol synthesis on Cu (111) surface, the results indicated that H supply (Step 2) to Cu surface as well as formate hydrogenation step (step 7) is essential especially at 300 K, at which artificial photosynthesis occurs (Table 3.6). This study underlines the importance of H supply and

> *75 atm 523 K*

�� +∗� �� ∗ ~0 0 0 0 �� ∗ + ∗� �� ∗ 0.36 2.4 3.53 3.7 ��� +∗� ��� ∗ 0 0 ~0 0 �� +∗� �� ∗ 0 0 -3.67 -0.21 �� ∗ +� ∗� ��� ∗+∗ 0 0 0 0 ��� ∗ +� ∗� ���� ∗ + ∗ 0 0 0 0 ���� ∗ +� ∗� ���� ∗ +� ∗ 0.997 ~1 ~1 1 ���� ∗ +� ∗� ���� ∗ + ∗ ~0 ~0 0.71 0.57 ���� ∗ +� ∗� ����� ∗ + ∗ 0 0.56 ~1 1 ����� ∗� ����� +∗ --- - � ∗ +� ∗� �� ∗ + ∗ 0 0 0 0 �� ∗ +� ∗� ���∗+∗ 0 0 0 0 ��� ∗� ��� +∗ 0 0 0 0 Table 3.6. The degree of rate control values with respect to rf10 (����� ∗� ����� +∗)

Table 3.5. Methanol formation rates (mol/gcat/s) at different reaction conditions with or

**3.2 Rate determining step of methanol formation on Cu (111) surface** 

**Microkinetic Modeling Literature** 

(Sahibzada et al., 1998)

Cu/TiO2) (Tseng et al., 2002)

*X***rc,i**

*75 atm 300 K* 

*1 atm 300K* 

*75 atm 423 K* 

**Reaction Conditions** 

523 K

300 K

300 K

reaction rate.

without initial water concentration

 **Elementary reactions** 

found by finite difference method at t= 5.18\*10-7 s

**Without initial H2O** 

75 bar 1.07\*10-9 4.58\*10-12

concentration on the surface for methanol formation rates. Since water splitting reaction is the only source of H in photocatalytic CO2 reduction mechanism, it could be said that water oxidation rates are rate limiting in artificial photosynthesis systems whereas water oxidation rate surpass carbon dioxide reduction rates at photosynthesis (Table 2.1).
