Advances in Photoelectrochemical Fuel Cell Research

1. Introduction

Fuel cells are electrochemical devices which can convert chemical energy into electrical power. They have the advantages of quiet in operation, high efficiency and low pollutant emissions. Photoelectrochemical fuel cells(PEFCs or PECs) are special fuel cells. PEFCs are used in organic waste degradation (Patsoura A et al.,2006),solar energy utilization (BakT et al.,2002), gaseous product decomposition(Ollis DF et al., 2000), aqueous pollutants removal (Sakthivel S et al., 2004) and photocatalytic sterilization(Fujishima A et al., 1972). A PEFC or PFC consumes fuels and utilizesluminous energyto generate electricity power when the photoanode is excited by radiation. (LianosP et al.,2010).

Fig. 1 shows a typical two-compartment photo fuel cell separated by a silica frit (Antoniadou M et al., 2010). The electrolyte is NaOH. The anode is nanocrystallinetitania. The cathode is a carbon black deposited with Pt as the catalyst. This device works under UV irradiation. The open circuit voltage was 0.88V without ethanol and 1.22 V with ethanol.

2. Mechanisms of Photoelectrochemical Fuel Cells (PEFCs)

PEFCs normally consist of a semiconductor photoanode, metal cathode and electrolyte which could be an acid, base or just water. Light excites electrons at the photoanode if the light energy is larger than the material energy bandgap. The photoanode generates electrons (e^-) and holes (h⁺). At the anode, production of oxygen happens. Hydrogen generates at the water/cathode interface. The reactions are shown as follows (Chang C et al., 2012):

$$Lightenergy:2hv\to 2h^{+}+2e^{-}$$E1

$$Atanode:2h^{+}+H_{2}O\to 1/2O_2+2H^{+}$$E2

$$Atcathode:2e^{-}+2H^{+}\to H_2$$E3

$$Overallreaction:2hv+H_{2}O \to 1/2O_2+H_2$$E4

Figure 1.

Thesketch of a two-compartment PEFC. (Antoniadou M et al., 2010).

3. Photoanode materials

Fig. 2 shows light absorption and electron transport on a photo sensitive material. The light energy is absorbed by the photo sensitive material. Electrons and holes generate. The electrons flow to cathode. The holes decompose water to produce oxygen. Nanostructured materials may be added to substrates such as Ti, glass, copper etc.

Figure 2.

Schematic of a typical nanostructured photoanode. (Chakrapani V et al., 2009)

Figure 3.

Some types of photoelectrode (PE) commonly used. (Minggu L et al., 2010).

Semiconductor is widely used asphotoelectrode which including n-type (TiO₂), p-type (lnP) and n-p type (n-GaAs/p-InP). They can be combined together to form multi-layered structures to tune the band gaps (Minggu L et al., 2010).In Fig. 3, SC stands for a semiconductor and M stands for a metal which is usually used as a substrate. Nanoporous materials are widely used in fuel cells. There are a number of transparent conductive oxides (TCOs) used as photoanode materials including indium-tin-oxide and fluorine-doped tin oxide. Some non-transparent conductive oxides (NTCOs) including nanocrystallinetitania TiO₂, n-type semiconductor ZnO, Fe₂O_3,SrTiO₃ etc. can also be used as photoanode materials. Among them, TiO₂is the most commonly used one due to its stability and high photo activity.

Fig. 4 shows the design of photoelectrode(Miller EL et al., 2003). Fig. 4a shows the first stage of design using p-type silicon. The catalyst layeris on the left side and the platinum catalyst is deposited on the right side. The arrow indicates the direction of light illumination. In Fig. 4b, the right side is coated with a Shottky barrier metal. Fig. 4c illustrates a three-junction structure consisting of Si-Ge-glass. The photo-hydrogen conversion efficiency is up to 7.8%. This design needs an external connection. Fig. 4d has no external connection, as compared with Fig. 4c. Fig. 4e is the latest integrated planar photoelectrode design. On the right side, there is a highly transparent and corrosion-resistant film to keep the high efficiency. This new design can connect single cells in series, which can generate large power.

Figure 4.

Photoelectrodedesigns. (Miller EL et al., 2003).

TiO₂ is an effective photocatalysis (PC). It is often used as the anode of PFC (GratzelM et al.,2001). The reaction of TiO₂ under UV illumination is follows (Park KW et al., 2007):

$$Ti{O}_2+hv\left(UV\right)\to Ti{O}_2+e^{-}+h^{+}$$E5

This formula is applicable for any metal oxide as the anode in a photo fuel cell. When the metal oxide absorbs photons from any light sources, electron-hole pairs are produced. The photo-generated holes react with fuels.

4. Fuels

There are many types of fuels for PECs including pure water, alcohols (MeOH,EtOH, PrOH), polyols (glycerol, xylitol, sorbito, glucose, fructose, lactose), organic pollutants (urea, ammonia, triton X-100, SDS, CTAB). Alcohols have larger efficiencies than others do (Antoniadou M et al., 2009). In polyols, glycerol has the highest current density. Pure water has the lowest efficiency. Fuels are decomposed in the ways as described below.

Methanol (Lianos P, 2010):

Anode electrode in acidic media：

$${\mathrm{C}\mathrm{H}}_3\mathrm{O}\mathrm{H}+2{\mathrm{h}}^{+}\to {\mathrm{C}\mathrm{H}}_3{\mathrm{O}}^{·}+2{\mathrm{H}}^{+}$$E6

$${\mathrm{C}\mathrm{H}}_3{\mathrm{O}}^{·}\to {\mathrm{C}\mathrm{H}}_2{\mathrm{O}}^{·-}+{\mathrm{H}}^{+}$$E7

$${\mathrm{C}\mathrm{H}}_2{\mathrm{O}}^{·-}\to {\mathrm{C}\mathrm{H}}_2\mathrm{O}+{\mathrm{e}}^{-}$$E8

Anode electrode in base media:

$$6\mathrm{O}{\mathrm{H}}^{-}+6{\mathrm{h}}^{-}\to 6\mathrm{O}{\mathrm{H}}^{·}$$E9

$${\mathrm{C}\mathrm{H}}_3\mathrm{O}\mathrm{H}+6\mathrm{O}{\mathrm{H}}^{·}\to \mathrm{C}{\mathrm{O}}_2+5{\mathrm{H}}_2\mathrm{O}$$E10

Under the photo illumination, PFC absorbs light energy and the TiO₂ is excited to release electrons. By this method, higher electric potential can be generated compared with other fuel cells. The completely reaction of TiO₂ with methanol’s shown as:

$$\mathrm{T}\mathrm{i}{\mathrm{O}}_2\left(\mathrm{U}\mathrm{V}\right)+{\mathrm{C}\mathrm{H}}_3\mathrm{O}\mathrm{H}+6{\mathrm{H}}_2\mathrm{O}\to \mathrm{T}\mathrm{i}{\mathrm{O}}_2+\mathrm{C}{\mathrm{O}}_2+6{\mathrm{e}}^{-}+{6\mathrm{H}}^{+}$$E11

Ethanol:

The completely reaction of TiO₂ with ethanol is as follows:

$$\mathrm{T}\mathrm{i}{\mathrm{O}}_2\left(\mathrm{U}\mathrm{V}\right)+{{\mathrm{C}}_2\mathrm{H}}_5\mathrm{O}\mathrm{H}+3{\mathrm{H}}_2\mathrm{O}\to \mathrm{T}\mathrm{i}{\mathrm{O}}_2+2\mathrm{C}{\mathrm{O}}_2+12{\mathrm{e}}^{-}+{12\mathrm{H}}^{+}$$E12

Reber JF et al., (1984) stated that a common formula could be:

$${\mathrm{C}}_{\mathrm{x}}{\mathrm{H}}_{\mathrm{y}}{\mathrm{O}}_{\mathrm{z}}+\left(2\mathrm{x}-\mathrm{z}\right){\mathrm{H}}_2\mathrm{O}\to {\mathrm{x}\mathrm{C}\mathrm{O}}_2+(2\mathrm{x}-\mathrm{z}+\frac{\mathrm{y}}{2}){\mathrm{H}}_2$$E13

Several types of biomass used in fuel cells are reported by Kaneko M et al., (2006), and shown in Table 1. The experimental condition is in acid solutions contain 0.1M Na₂SO₄. The anode of PEC is TiO₂nanoporous film and the cathode is Pt black on Pt foil. The light intensity is 503 mW/cm² and ambient temperature is 25 °C. The results of open circuit voltage show that acetic acid is the best. Ammonia, glycine, phenylalanine and glutamic acid also show good performances. The short circuit current of methanol has the highest value. The fill factor (FF) as defined by the ratio of maximum obtainable power to the product of the open circuit voltage and short circuit current was calculated. Ammonia has the maximum FF of 0.63.

LiuY et al.,(2011), did similar research on various fuels with a self-organized TiO₂ nanotube array (STNA) as the photoanodeof the photo fuel cell (Table 2). Multiply fuels were tested but each fuel’s concentration was smaller than what Kaneko et al. used. By comparing the data in these two tables, we can see that the open circuit voltage and short circuit current obtained by Liu et al. are slightly larger, which means that they got higher efficiencies from the PFC system they built. When they varied the concentration of Na₂SO₄ from 0 to 0.5M, V_oc and J_screached the peak values at 0.1 M and the FF has the maximum value at 0.05 M. All the experiments were done under solar light illumination.

Fuel (conc./M)	Solvent (pH)	Voc/V	Jsc/ mA cm^-2	FF
Methanol	None	0.54	0.8	0.23
Methanol (50 vol.%)	Water (not controlled)	0.44	0.76	0.28
Ethanol	None	0.49	0.52	0.25
Glucose (0.5)	Water (5)	0.64	0.5	0.32
Urea (5)	Water (5)	0.6	0.3	0.26
Ammonia (10)	Water (12)	0.84	0.53	0.63
Acetic acid (2 wt.%)	Water (not controlled)	0.94	0.47	0.37
Glycine (0.5)	Water (5)	0.76	0.45	0.45
Glutamic acid (0.5)	Water (1)	0.9	0.64	0.42
Tyrosine (0.5)	Water (13)	0.86	0.43	0.36
Phenylalanine (0.5)	Water (13)	0.9	0.61	0.53
Agarose (0.2 wt.%)	Water (5)	0.6	0.12	0.26
Gelatin (2 wt.%)	Water (1)	0.64	0.23	0.32
Collagen (3 mg/ml)	Water (l)	0.62	0.16	0.34
Cellulose sulfate (2 wt.%)	Water (not controlled)	0.56	0.29	0.34
Lignosulfonic acid (0.5 wt.%)	Water (not controlled)	0.57	0.02	0.51
Polyethylene glycol (2 wt.%)	Water (5)	0.6	0.28	0.27
Poly(acrylamide) (2 wt.%)	Water (5)	0.6	0.23	0.24

Table 1.

PFC performances by using different fuels in 0.1M Na₂SO₄ with a TiO₂photoanode and Pt/Pt black cathode. (Kaneko M et al., 2006).

	Organic compounds	Voc (V)m2)	Jsc; (mA cm^-2)	Jvmax; (mW cm^-2)	FF
Model compound	Na2SO4 (0.1 mol/L)	1.13	0.35	0.12	0.31
	Glucose (0.05 mol/L)	1.28	0.83	0.38	0.36
	Glutamic acid (0.05 mol/L)	1.34	1.08	0.51	0.35
	Nicotinic acid (0.05 mol/L)	1.39	0.61	0.3	0.35
	Acetic acid (0.05 mol/L)	1.48	1.42	0.67	0.32
	Urea (0.05 mol/L)	1.41	0.91	0.51	0.4
	Ammonia (0.05 mol/L)	1.24	0.72	0.37	0.41
Actual wastewater	Pharmaceutical wastewater (COD =24572 mg/L)	0.88	1.36	0.43	0.36
	Petroleum exploiting wastewater (COD =19087 mg/L)	1.34	0.98	0.34	0.26
	Dying wastewater; (COD =10842 mg/L)	1.53	1.21	0.5	0.27
	Chemical plant wastewater (COD =11700 mg/L)	1.11	0.99	0.3	0.27
	Original urine solution; (COD =9642mg/L)	0.93	0.61	0.19	0.34

Table 2.

PFC performances by using different fuels. (Liu Y et al., 2011).

5. Cathode materials

As compared with multiple choices of photoanodes, the materials for the cathode of photo fuel cells are limited. Normally a Pt wire or a Pt foil is used. Another option is to use Pt-black. The Pt black powders can be cast, sprayed or hot-pressed on the surface of a Pt (Kaneko M et al., 2006). The surface area becomes larger when the Pt-black powers were deposited ontoPt wires or foils. In addition to platinumcathodes including platinum wire, non-platinized platinum foil, platinized platinum foil, platinized SnO₂ with F, metal nanoparticles deposited on a TiO₂/SnO₂ with F doping are made into electrodes. Pt/TiO₂/SnO₂, Pd/TiO₂/SnO₂, Au/TiO₂/SnO₂, Ag/TiO₂/SnO₂, and Ni/TiO₂/SnO₂) are some of the examples. A platinum-loaded carbon cloth has also been used as a cathode material. The platinized SnO₂ with F electrode has better performance than others. It speaks current, voltage and efficiency are 1.15 mA/cm², 1340 mV and 12.3%, respectively. The platinum-loaded carbon cloth has the maximum efficiency of 32.3%. Thin layer of Si-H film photo cathode can be made by plasmaassisted chemical vapordeposition (PECVD). A Si-H cathode deposited organic or inorganic protective layer or coating with catalytic platinum can enhance the stability for long time use. The best thickness of the polymer protective layer is 5 nm. The optimized thickness of Pt coating is 2 nm.

CuO is a cheap material. CuOnanoparticles and films prepared by flame spray pyrolysis (FSP) were used as photocathodes by Chiang C et al., (2011). The optical band gap was decreased from 1.68 eV to 1.44 eV with the annealing temperature increasing from room temperature to 600°C. The nanoparticle size is from 50 nm to 150 nm, as shown in Fig. 5.The best photocurrent density is 1.2 mA/cm² obtained from CuO particles which were annealed at 600 °Cfor 3 hour. The bias voltage is 0.55 V in 1M KOH. The total conversion efficiency is 1.48% and the hydrogen generation efficiency is 0.91%.

Figure 5.

SEM images of CuO photocathodesprepared under different conditions: (a) 450 °C, 1 h, (b) 450 °C, 3 h, (c) 600 °C, 1 h, (d) 600 °C, 3 h. (Chang C et al., 2011).

6.Terminologies associated with the photo fuel cells

6.1. Optical absorption coefficient for band gap determination

The optical absorption coefficient, α, is related to the wavelength, transmittance, reflectance of the light illuminating on a material. Low absorption coefficient means low photo absorption ability. The following equation holds (Pihosh Y et al., 2009)

$$\alpha \left(hv\right)=d^{-1}\mathrm{l}\mathrm{n}\left(\frac{1-R}{T}\right)$$E14

where T is the transmittance, R the reflectance, and d the thickness of the material. The term h$\upsilon$ refers to the photon energy.

The optical coefficient is used to obtain the band gap E_gfollowing

$$constant\mathrm{ }\times \mathrm{ }{(hv-E_g)}^2=\mathrm{ }\alpha \left(hv\right)\times \left(hv\right)$$E15

(15)

6.2. Roughness factor

Roughness factor is related to the surface area of an electrode. For nanotubes, the geometry roughness is calculated as (Shankar K et al., 2007)

$$\mathit{G}=\left[\frac{4\mathit{\pi }\mathit{L}(\mathit{D}+\mathit{W})}{\sqrt{3{(\mathit{D}+2\mathit{W})}^2}}\right]$$E16

whereDis theinner diameter,Wthewall thicknessand L the tube length of the nanotubes. From the experiment on titania nanotubes by Isimjan TT et al., (2012), a higher surface area (roughness) was obtained at higher processing voltages. At a constant voltage, the pore size of nanotubes is dependent of distance between anode and cathode in the electrochemical process.

6.3. Photo conversion efficiency

The photo conversion efficiency is the overall efficiency of a PEC which can be defined by the following equation

$$\eta \left(\mathrm{\%}\right)=\left[\frac{\left(totalpoweroutput-electricalpoweroutput\right)}{lightpowerinput}\right]\times 100$$E17

(17)

7. Nanostructures photoanodematerials processing

7.1. TiO₂nanotube (TNT)photoanode

TiO₂ nanotubes on the surface of Ti as shown in Fig. 6 demonstrate a self-organized nanostructure. The advantage of the nanobutes is the high surface/volume ratio. TiO₂nanotubes have active photo catalysis characteristic, good corrosion resistance, thermal stability and good operation stability as described by Mahajan V et al., (2008). TiO₂ nanotubes can be made by various ways including hydro/solvothermal method (Kasuga T et al., 1998), template-assisted approach (carbon nanotube, alumina or monocrystal as the template), sol–gel method (Kasuga T et al., 1998), microwave irradiation (Zhao Q et al., 2009),and direct electrochemicalanodization. The advantage of the hydro/solvothermal method is easy to operate. The disadvantage is that only disordered and twisted TiO₂ nanotubes can be obtained. For the template-assisted method, the size of the nanobutes is uniform. For the electrochemical anodic oxidation method, it has the advantage of easy to operate and the obtained nanotubes are highly ordered. Therefore, many researchers prefer the electrochemical method.

Figure 6.

Self-organized TiO₂ nanotubes via anodization. (Shankar K et al., 2007).

7.1.1. Hydrothermal treatment

Hydrothermal method is one of the popular approaches to prepare TNTs. The first group having successfully fabricated TiO₂ nanotubes by hydrothermal method is Kasuga T et al., in 1998. During the process, titaniananopowdersare placed in alkaline aqueous solutions held in high pressure steel vessels. The temperate should be between 50-180°C. The process continues for 10 to 20 hours. Some post treatment can be applied, for example, washing with acid or alkaline solutions for 10 hours, drying at 80 °C and annealing at 500 °C. The reaction process is divided into four steps (Hafez H et al., 2009) i.e. (1) synthesis of TiO₂ nanotubes in alkaline aqueous solutions, (2) protons replacing alkali ions in the reaction, (3) drying, (4) acid washing (post treatment). There is controversy about the necessity of the acid washing. Some researchers (Liu S et al., 2009) think acid washing is a necessary procedure to form TNTs, but other researchers (Chen X et al., 2007) think hydrothermal is more important than the acid washing step as sketched in Fig. 7. The step of washing by acid is not even necessary to form TNTs.

Figure 7.

Hydrothermal method for fabricating TiO₂ nanotubes (Chen XB et al., 2007).

7.1.2. The effects factors of material and solution

With different raw materials and reaction solutions, the different morphology of TiO₂was obtained by hydrothermal method (Yuan ZY et al., 2004). When crystalline TiO₂ react with NaOH under 100-160 °C, the TiO₂ nanotubes was obtained. When amorphous TiO₂ be used under same conditions, the TiO₂nanofibers are fabricated. Either crystalline or amorphous TiO₂ can be used reaction with NaOH can result TiO₂nanoribbons when temperature rise to 180 °C. If the solution used by KOH, the nanowires morphology is formed. The pH value of solution also plays an important role in morphologyof TiO₂nanomaterials (Xu YM et al., 2010). Fen LB et al., (2011) used anataseTiO₂nanopowders (Aldrich 637254-50G, 99.7%) with NaOH solution fabricated TNTs. The inner diameter is 3-6 nm and wall thickness is 1.9 nm. LanY et al., (2005) used rutile nanopowders with 10 M NaOH solution obtained TNTs which inner diameter 2-3 nm and wall thickness is 7-8 nm, besides the length is 200-300 nm. The inner diameter is smaller but the wall thickness is larger than the TNTs made by Fen LB et al., (2011).

Hydrothermal treatment temperature and time are significant factors during the formation of TNTs. The temperature range should be from 100°C to 180 °C and the time range should be from 1 to 24 hours. Sreekantan S et al., (2010), selected the temperatures at 90, 110, 130, 150 °Cand time for 3, 6, 9, 15, 18, 24 hours. The NaOH/TiO₂ solution was used. At 90 °C, the TiO₂ particles form sheets. When the temperature was set at 110 °C, the sheets were transformed into nanotubes because the thermal energy increases with temperature (Seo HK et al., 2008). With the temperature increasing to 130, 150 °C, there is no change of the outer diameter (10 nm) of nanotubes but the TNTs transform to anatase phase. For the effect of reaction time, particles begin to form sheet at 3 hours. SreekantanS et al.,(2010) indicated that Ti-O-Ti bond is replaced by Ti–O–Na and Ti–OH bonds at this time. After 6 and 9 hours, more and more sheets form nanotubes (10 nm). After 15 hours, TNTs form completely. They found that 150 °Cis the best temperature for making TNTs with the highest photocatalyticactivity.

SeoHK et al., (2008), studied the phase transformation of TNTs at different hydrothermal temperatures. They used a 10 M NaOH solution and the temperature range was from 70°C to 150 °C. A 0.1 M HCl solution was used for washing the TNTs. They founded that at 70 °C, the particles begun forming nanosheets. Nanosheets and nanofibers co-existed at 90 °C. At 110 °C, the nanosheets were transformed into nanotubes. This conclusion is also reported by Sreekantan S et al., (2009).Hydrothermal processing can also produce nanoribbons instead of nanotubes if the reaction temperature is higher than 180 °C.

7.2. Synthesis of self-organized TiO₂ nanotubes via electrochemical anodization

In 1999, Zwilling V et al. first used electrochemical anodization method for synthesis of TiO₂ nanotubesin the solution containing chromic acid and hydrofluoric acid. Later many researchers (e.g. MacakJM et al., 2005) showed that using different applied potentials, electrolytes, pH values (much longer nanotubes at neutral pH electrolytes) and anodization time can control the lengths, thickness, diameters and morphology of TiO₂ nanotubes.Zeng X et al., (2011), reported electrochemical oxidation of Ti in a1.0 M H₃PO₄ and 0.25 M NaF solution. With the increasing in the potential, TiO₂ experienced three forms. When the potential was very low, Ti dissolved into the solution. With the increasing of potential, Ti was oxidized toform TiO₂. When the potential was less than 2.5 V, TiO₂ film was obtained. Between 2.5 V and 6.0 V, the TiO₂ porous structure formed. When potential was higher than 6, the self-organized TiO₂ nanotubes were obtained (Fig. 9b).

Figure 8.

Morphology of self-organized anodic TiO₂ nanotubes formed at different temperature and voltage levels. (Liu H et al., 2011).

Figure 9.

a) Sketches for electrochemical oxidation of Ti. (b) effectof voltage level on the morphology of TiO₂. (Zeng X et al., 2011).

Before 2005, all of these researches were exclusively using inorganic solutions as electrolytes, such as HF (Varghese OK et al., 2003), KF, NaF (Cai QY et al., 2005). Macak JM et al., (2005), investigated TiO₂ nanotube formation in Na₂SO₄ electrolytes with NaF. The maximum length of nanotubes was up to 2.4 µm. It takes about 6 hours. But longer than this time, the irregular morphologyshowed up. As compared with HF, NaF can thicker the porous layers. The use of organic electrolytes is a milestone for the TiO₂ nanotubes fabrication. Liu H et al., (2011), studied the temperature effect on morphology of TiO₂ nanotubes. The specified temperatures are -5, 0, 5, 10, 15 °Cand the applied potentials are 10, 30, 50V. It helped control the nanotube size and structure under the complex condition as show in Fig. 8. In summary, there are two types of electrolytes in TiO₂andoization, one is aqueous-based electrolytes, and the other is organics-based electrolytes. Aqueous electrolytes allow the nanotubes to form more quickly because of the low electrical resistance. Besides, lower voltage is enough. However, it is hard to form longer nanotubes because of the dissolution of the nanotubes in the solutions. The organic electrolytes, for example, ethylene glycol and glycerol, have higher electrical resistances. They can slow down the ion transfer. Therefore, higher voltages and longer times are needed. In organic electrolytes, it is easier to form long nanotubes.

7.2.1. Anodizationmechanisms

During anodization, a constant voltage in the range from 1V to 150V is applied. The electrolytes containing fluorides have the concentration range from 0.05 to 0.5M.The processing time ranges from a few minutes to a couple of days.

There are two main reactions with the anodization of Ti (Macak JM et al., 2005):

$${\mathrm{T}\mathrm{i}}^{4+}+2{\mathrm{H}}_2\mathrm{O}\to \mathrm{T}\mathrm{i}{\mathrm{O}}_2+4{\mathrm{H}}^{+}$$E18

$$\mathrm{T}\mathrm{i}{\mathrm{O}}_2+6{\mathrm{F}}^{-}+4{\mathrm{H}}^{+}\to {\left[\mathrm{T}\mathrm{i}{\mathrm{F}}_6\right]}^{2-}+2{\mathrm{H}}_2\mathrm{O}\mathrm{ }$$E19

Figure 10.

Sketches of Ti anodization (a) without F^-, (b) with F^-. (Macak JM et al., 2007).

First, titanium in the electrolyte produces Ti⁴⁺. Then Ti⁴⁺ reacts with water to formTiO₂and hydrogen ion (Eq. 18).TiO₂ becomes oxide film on the surface of the titanium as a barrier layer. Meantime, TiO₂ is etched by F^- and many holes form in the film (Eq. 19). With the processing time increasing, the holes become deeper and form nanotubes. When the anodization rate of Ti is equal to the etching rate ofTiO₂, the process reaches to a steady-state. The length of nanotubes keeps unchanged.

F ion plays an important role in synthesizingTiO₂ nanotubes. Fig. 10 shows the results of NTs obtained from different solutions with and without F^-. Without F^- the TiO₂ is flat without porous structure. With F^-, reaction (Eq.19) occurs. F ion generates TiF₆^2- which is the driving force of etchingTiO₂. H⁺ can enhance the etching ability of F^-. TiF₆^2-ions owing the small diameter can easily move through TiO₂ crystal lattice. Comparing the electrolytes containing Cl^- and Br^- (Chen X et al., 2007),TiO₂ nanotubes arrays fabricated in electrolytescontaining F^- have better quality. Fluoride concentration can affect the electrochemical characteristics (BeranekR et al., 2003). If the fluoride concentration is low (less than 0.05 wt. %), there are almost no fluoride ions. If the fluoride concentration is high (1 wt. %), no oxide formation can be observed.Ti⁴⁺ reacts with F^- immediately to form TiF₆^2-. The maximum nanotube length is about 500 nm synthesized in HF electrolytes. The maximum length is several micron meters using NaF andNH₄F electrolytes.

Figure 11.

Self-organization of TiO₂ nanotubes in F^- containing solutions. (Gan Y et al., 2011).

The mechanism of TiO₂ growth can be shown in Fig. 11. TiO₂ grows on the Ti substrate gradually. With the TiO₂ film being thicker and thicker, TiO₂ has the function of a protecting film to slow down the Ti dissolution. With the development of F^- etching TiO₂, the self-organized TiO₂ nanotubes form as illustrated in Fig.11d.

7.2.2. Synthesis of TiO₂ nanotubes using organic electrolytes

Organic electrolytes containing F^- have some advantages. In 2005, RuanCM et al. used dimethyl sulfoxide (DMSO) and ethylalcohol (1:1) as electrolytes for fabricating TiO₂ nanotubes with a length of 2.3μm.MacakMJ et al., (2005) used glycerinumsynthesized TiO₂ nanotubes with 7μm length. The maximum length could over 1000μm. Prakasam HE et al.,(2007), using ethylene glycol with 1%-3% H₂O volume and 0.1% to 0.5 % wt of NH₄F solution, anodized Ti foil for 17 h at 20, 40, 50, 60 and 65 V. The result showed that with the increase in the voltage, the inner diameter, outer diameter and length of nanotubes were increased. The maximum values are 135 nm, 185 nm and 105 µm, respectively. The nanotubes grow rate is15 µm/h. The important factor to affect the length of the nanotubes is the water content. The water volume content should be under 5% for obtain good quality of nanotubes. The morphology of TiO₂ nanotubesformed in organic electrolytesis more smooth and orderly. Besides, the nanotubes have higher photocatalytic efficiencies.

Non-F^- electrolytes were also used (AllamN et al., 2007) for the environmental protection purpose. Pulse anodization (ChanmaneeW et al., 2008)generated TiO₂ nanotubeswith good photoelectrochemical property. Glancing angle deposition (GLAD) was used to obtain Ti films. The anodization of the Ti films produced nanotubes and nanorods (NRs) on a glass substrate. Even brush type nanostructures (BTNs) were obtained (Pihosh Y et al., 2009) as shown in Fig. 12. As compared with the plate counterparts, the TiO₂ NRs, NTs and BTNs have significantly higher photocatalytic activity under Vis-light and UV illumination. The NTs and BTNs have better photocatalytic activity than the NRs because of their larger surface areas. The BTNs can be obtained by andoization of NRs in base.

Figure 12.

SEM images of brush type nanostructures (BTNs). (Pihosh Y et al., 2009).

7.3. Post-treatmentof TiO₂ nanostructures

7.3.1. Annealing

The purpose of annealing is to change the morphology of TiO₂nanotubes from amorphous to crystalline (anatase or rutile). Over the past 10 to 15 years, there were a large number of researchers focusing on annealing. Stem N et al., (2011), thermal treated TiO₂ at 1000 °Cin wet N₂ for 2 hours, which enhances the photocatalytic performance.Wang MC et al., (2012), showed that annealing temperature affected photocatalytic capability of N-doped TiO₂ thin films. The temperature ranges from 250 °C to 550 °C. The time lasts for 1 hour. Below 350 °C, the surface roughness is low. The photocatalytic activity is the highest after the 350 °C annealing. Lin JY et al., (2011), applied the rapid thermal annealing (RTA) method. Thetemperatures been used that were 700, 800, 900, 1000 and 1100 °C. The temperate increasing rate was 5 °C/s. The total annealing time was 30 s in oxygen. Through the X-ray diffraction (XRD) examination, it was found that oxygen-related defects were reduced when the TiO₂ nanotubes changed from amorphous to anatase phase. Fang D et al., (2011), studied high temperate calcinations.TiO₂ nanotubeswere exposed atthe temperatures of 450,600, 800,and 900°C. The results (Fig. 13) showthat 450 °Chelps generate a pure anatasephase. At 600 °C, a mixed phase of anatase and rutile can be got. At 800°C, pure anatase phase grows into large crystallites. As the conclusion, 450°C is the best calcination temperature. Bauer S et al., (2011), showed that the nanotube’s size affects the crystal phase. When the nanotube’s diameter is smaller than 30 nm, it is more likely to form rutile. In contrary, when the diameter is larger than 30 nm,anatase can be obtained. Not only the temperate, the type of gases used also affects the properties of the nanotubes (Sang LX et al., 2011). Annealing in air (TNT-A), nitrogen (TNT-N) and 5% hydrogen/nitrogen (TNT-H)generate the similar morphology and band gap. But the difference in the UV absorption photocurrent density exists. The maximum photocurrent density is 0.60 mA/cm² for the nanotubes named as TNT-H. The minimum is 0.27 mA/cm² for TNT-A. TNT-H has more surface defects. The more surface defects, the higher the photocurrent was generated.

On the contrary to traditional annealing process, Liu JM et al., (2011), performed vacuum annealing and multi-cycle annealing on theNb-doped TiO₂ thin film. During the three-cycle vacuum annealing, the TiO₂ was heated up to 550 °C(0.05 Pa air pressure) for 1 hour in one cycle. This process was repeated for three times. In another experiment, the TiO₂ film was held at 550 °Cat 5 Pa air pressure for 1 hour. Then the annealing was repeated for three times. These two different procedures both can improve the conductivity of the Nb-doped TiO₂ thin film. At different annealing temperatures, the TiO₂ nanotubes showed different photoelectrochemical characteristics (Tang Y et al., 2008). The treatment temperatures are in the range from 300 °C to 550 °C. Again, the sample annealed at 450 °C showed better performance under UV light. With the UV on, the nanotube electrode showed good photoelectric current stability. When the UV was off, the photocurrent quickly decreased to the initial value.

7.3.2. Ultrasonic clean

During anodization synthesis, TiO₂ nanotubes formed, but unexpected deposits may also be on the nanotubes. They can be cleaned by ultrasonic waves (Cai QY et al., 2005).Xu H et al., (2011), applied ultrasonic waves to clean the surface of TiO₂ nanotubes for different time periods. They employed the ultrasonic wave with the power of 80 W at40 kHz. 9 min is the best treatment time for cleaning the nanotubes. When the time was extended to 40 min, the nanotubes were broken. The nanotubes were peeled off from the Ti completely at 60min.

Figure 13.

Annealing treatment of TiO₂ nanotubes (a) with Ti substrate, (b) free standing nanotubes. (Fang D et al.,2011).

7.3.3. Doping

There are two main limitations of pure TiO₂ nanotubes

1. Pure TiO₂ can only absorb UV light of wavelength shorter than 400 nm because the band gap of TiO₂ is 3.2 eV, which means that pure TiO₂ can only utilize 6% solar energy. The visible light has the energy band gaps from 1.8 eV to 3.1 eV.

2. High electrical resistance of pure TiO₂ at the room temperature results in very low electron transfer rate. This causes electric energy loss. The converted heat energy dissipates into ambient. At 20 °C, TiO₂ is not a conductor. Only when the temperature rises to 400 °C, the resistance of TiO₂ becomes lower.

Direct doping is one way to overcome the limitations of pure titania. Another method of doping is to stack different materials which have different band gaps. That could make hybrid photoanode (HPE)as first reported by Morisaki H et al., (1976).

Some noble metals doped to TNTs such as gold (Malwadkar SS et al., 2009), silver (Guo GM et al., 2009), platinum can improve the photocatalytic activity of the TNTs. This is because these noble metals can inhibit recombination of electron (Chan SC et al., 2005). Metal doped-TNTs for photo fuel cell applications are reported.

In addition to using noble metals, Macak JM et al., (2007) showed doping copper by electrodeposition. First, synthesis of TiO₂ nanotubes which have low conductivity especially at the bottom of the nanotubes was carried out. Second, using an aqueous electrolyte, about 1% of Ti⁴⁺ in the TiO₂ outer layer was converted into Ti³⁺ (Ti⁴⁺+ e^-+ H⁺= Ti³⁺H⁺ and 2H⁺+ 2e^-= H₂).With Ti³⁺, the mobility gap of TiO₂ was reduced from 3.2 eV to 2.4 eV and the bottom of the nanotubes become highly conductive. The third step is to dope the nanotubes with Cu. Since the bottom has a high conductivity with Ti³⁺ and H^-, Cu is easily be doped in the nanotubes by a current pulsing electroplating approach.

Sun L et al., (2009), fabricated Fe-doped TNTs using Fe ion containing electrolytes. This study shows that the content of Fe³⁺ is a significant factor affecting the photo catalysis capacity. They used three Fe(NO₃)₃ solutions for comparison. The concentrations are 0.05M, 0.1M and 0.2M. The result shows that 0.1M Fe(NO₃)₃doped TNTs have the maximum photo current and photocatalytic degradation rate.0.15M Fe(NO₃)₃ doped nanotubes have the maximum absorbance under UV-Vis. Different application need different Fe(NO₃)₃ contents. Tu YF et al., (2010), employed template-based LPD method to dopeFe to TNTs. Redshift of the absorption was found. The best Fe content was 5.9 at %. The doped-TNTs achieved the best efficiency of photo catalysis under visible light. Wu Q et al., (2012), fabricated Fe-doped (Fe₂O₃ and Fe³⁺) with ultrasoundassisted impregnating calcinations method. Results showed that Fe₂O₃ went into the TNTs and Fe³⁺ into the TiO₂ lattice. The operation time and temperatures affect the photo responses of Fe-doped TNTs. Ultrasound treating for 5 min following by annealing at 500 °C provides NTNs the highest photo catalysis efficiency.

C and N doping are non-metal doping examples. B, P, and other nonmetallic dopants are also used. Nitrogen is the earliest, most effective and most studied doping element for TNTs. There are many methods to dope nitrogen into TNTs such as annealing TNTs in ammonia (Vitiello RP, et al., 2006), ion implantation (Ghicov A et al., 2006) etc. Asahi R et al., (2001), doped TiO₂ with nitrogen using a solution method. Vitiello RP et al., (2006), showed a simple method for making N-doped TNTs, which is treating TNTs at 300-600 °C in NH₃ atmosphere. Results showed that 500 °C-600 °C is the best annealing temperate range at which TNTs transfer to anatase and have the most effective photoresponse.Xu JJ et al., (2010), showed difference in photo catalytic activity between N-doped and Non-doped TiO₂ nanotubes under Vis-light. The photocurrent density of N-doped nanotubes was twice as that of the non-doped nanotubes under visible light illumination.Yuan J et al., (2006), synthesized N-doped TiO₂ by heating urea with TiO₂ at 300-700°C. The doped TiO₂ can absorb light with the wavelength up to 600 nm. The result shows that urea changes to chemisorbed N₂ and substituted N staying in the TiO₂.

Liu ZY et al., (2009), doped carbon into TNTs for solar photochemical cellhydrogen generation. Shaban YA et al., (2007), studied the fabrication time and temperature effects on grooved and non-grooved Ti metal sheet doped with carbon for photochemical catalysis. The result shows that the grooved sample has higher photocurrent density than the non-grooved one. The grooved simple with a depth of 0.005 inch has the maximum photo conversion efficiency of 11.37 % (treated at 820 °C, 18 minutes, thermal flame oxidation, tested in 5.0 M KOH, illuminated by a 150 W Xenon lamp).The non-grooved simple, 0.003 inch grooved one, and 0.001 inch groove done have the maximum photo conversion efficiency of 9.08 %, 8.68 %, 7.20 %, respectively under the same treatment condition.

Co-doping multiple elements was also applied to TiO₂ nanotubes. Tungsten and nitrogen co-doping is a typical example (Shen YF et al., 2008). Nitrogen and sulfur (Yan GT et al., 2011), fluorine and boron (Su YL et al., 2008), Pt and N (Huang LH et al., 2007) co-doping has also been studied. Liu SH et al., (2009), developed a carbon and nitrogen co-doping method by adding 5 mg polyvinyl alcohol (PVA) and 20 mg urea. Then calcination was performed in nitrogenat 600 °C. The photocurrent density is 3 times, 2 times, and 1.2 times compared with the non-doped, C-doped and N-doped TiO₂ nanotubes under solar light and 0.2 V bias-potential combined excitation. He HC et al., (2011), doped Pt-Ni into NTNs using pulsedelectrodeposition method. The photo catalytic activity is better than that of only Pt-doped NTNs. Pt-Ni doped NTN is a good anode material for photo fuel cell (direct methanol type). The performance of co-doped TNT is better than that of Pt doped one. Huang LH et al., (2007), synthesized Pt-N doped TNTs by two steps. First, they obtained N-doped TNTs. Second, they used H₂PtCl₆ solution to supply Pt, resulting in Pt-N co-doping. N-doping can enhance the photo response activity and Pt-doping can strengthen the electron separation from holes. Ag can be deposited into N-doped TNTs (Zhang SS et al., 2011), via electrochemical deposition in a0.2 g/L AgNO₃ solution. The result shows that the average photocurrent density of the Ag/N-doped TiO₂ nanotubes is 6 times higher.

Li XQ et al., (2011), developed the CdS nanoparticle and CuTsPc molecule co-doped TNTs. The I-V curve shows that CdS-CuTsPc has the maximum photocurrent density as compared with CdS-CuTsPc, CdS–CuPc, CdS, and CuTsPc doped TNTs. Jia FZ et al., (2012), successful processedZnS-In₂S₃-Ag₂S doped TiO₂-_xS_xby a two-step(anodization and solvothermal) approach.Zhang X et al. (2009), doped PW₁₂O₄₀³⁻ and Cr³⁺ into TNTs through the anodization and impregnation methods. The function of Cr³⁺ is narrowing the band gap of TiO₂. They have found that the synergetic factor is 1.42. Su Y et al., (2008), doped nitrogen and fluorine into TNTs. They simply usedanodization (20 V) of Ti in the C₂H₂O₄ 2H₂O+NH₄F electrolyte through TiO₂ self-organization. After annealing at 400°C, N-F-doped TNTs showed very good photocatalytic ability and stability. The efficiency of methyl orange (MO) decomposition test is higher than 97%. This method avoids using ammonia which is hazardous. With CeO₂ nanoparticles being doped into TNTs, enhanced charge storage capacity of TNTs was achieved (Wen H et al., 2011). Wang J et al., (2012), reported a C₃N₄ doped TiO₂nanorod. The UV-Vis absorbance ability of this modified material is as twice as that of the TiO₂.

8. Conclusions

Photoelectrochemical fuel cells have experienced fast development recently because of the progress in nanomaterials. Using various materials processing techniques, it is possible to obtain various nanostructure forms such as nanoparticles, nanorods, nanothin film and nanotubes for photo fuel cell applications. There are many ways for fabricating nanostructures including hydro/solvothermal method, template-assisted method, sol–gel method, microwave irradiation method and electrochemical direct anodizaiton method. Electrochemical anodization becomes a popular method in recent years because it is easy to control the size of nanotubes. Typical photo sensitive materials such as TiO₂, WO₃, Fe₂O₃, CuO and ZnO have been studied. These materials have different band gaps and many researchers reported how to enhance the photo response of them. Doping is a significant and efficient method for improving the photo response of nanomaterials. Metal doping and Non-metal doping are two major types. Besides, organic doping, co-doping alloys and muti-component materials also result in good performance of PEFCs. In summary, PEFCs represent promising energy conversion systems. Future studies should focus on increasing the photoelectric energy conversion efficiency.