**5.1 Conductivity, transparency, and workfunction of polyaniline and PEDOT:PSS**

The electrical conductivity was evaluated using a current-voltage (*I*-*V*) measurement setup under dark conditions. The conductivities estimated from the result of the *I*-*V* measurements were 3.4×102 S/cm and 5.7×10-1 S/cm for PANI and PEDOT:PSS, respectively.

The optical transmittance was evaluated using a UV-visible-near-infrared spectrophotometer (UV-3150, Shimadzu Co., Ltd.). Figure 3(a) shows the optical transmittance spectra of the PEDOT:PSS and PANI films. Both of the films exhibited transmittance greater than 80% within the wavelength region between 250 and 1500 nm. This is superior to conventional transparent contact materials such as transparent conductive oxides or semi-transparent metals (Kim et al., 2002; Satoh et al., 2007), which exhibit significant drops in transparency particularly near the UV region, as seen in Figure 3.

Transparent Conducting Polymer/Nitride Semiconductor Heterojunction Solar Cells 315

heterojunctions had a Schottky contact property comparable to that exhibited by

The depletion width, *WD*, in the n-type GaN of the TCP/epi.-GaN heterojunction is

*D Built in D W VV qN* 

*<sup>S</sup>* is the relative dielectric constant of GaN and equals 8.9 (Wu, 2009),

given by *Q qN W SC D D* , thus, the depletion layer capacitance *CD* is obtained by

2

*D*

Equation (5) can also be written in the following form:

required to attain good Schottky contact with a

ambient temperature. The observed

> *<sup>m</sup>*

difference

TCP/epi.-GaN heterojunction.

vacuum dielectric constant, *Vbuilt-in* is the built-in voltage formed in GaN, *V* is the bias voltage, and *ND* is the donor concentration. The space charge, *QSC*, in the depletion layer is

*SC S D*

*VVV* 

0

*<sup>B</sup>*of the TCPs were much lower than expected from the energy

. There are various possibilities for the lower barrier heights including the

the good Schottky contact properties in the TCP/epi.-GaN heterojunction were achieved with convenient spin coating of a water-dispersed TCP solution onto the GaN layer in air at

Schottky effect, which is caused by the electronic mirror force, interface dipole effect, surface defects of GaN, inhomogeneous workfunctions in the TCP film, and/or residual contamination (Sze, 1981; Kampen, 2006). However, the major candidates for the modification of the barrier height have been discussed and are still controversial even in conventional metal/semiconductor Schottky heterojunctions (Tung, 2001). Further detailed investigation is required to determine which effects dominate in lowering the barrier in the

*V V*

1 2( ) *built in D S D*

Equation (5)' suggests that if 1/*CD*2 exhibits linear plots against *V*, *Vbuilt-in* can be obtained at the *V*-intercept of extrapolated fit-line of the plots. Figure 4(b) shows the plot of 1/*CD*2 as a function of the applied voltage. The frequencies used for the capacitance measurements were 100 Hz and 1 KHz for the PANI/epi.-GaN and PEDOT:PSS/epi.-GaN heterojunctions, respectively. The frequency for measurement was chosen within a range that was sufficiently lower than the cut-off frequency, which is described in section 5.4. In Figure 4(b), both the data sets are linear and straight lines were successfully fitted to the data. The determined diode characteristics of the TCP/epi.-GaN heterojunction determined from the *J*-*V* characteristics and capacitance measurements are summarized in Table 2. The observed barrier height was comparable to that obtained by conventional metal Schottky contacts (Tracy et al., 2003). In the case of the conventional metal Schottky contacts, elaborate surface cleaning processes and moderate metal deposition in ultra-high-vacuum conditions are

*C q N* 

*<sup>Q</sup> q N <sup>C</sup>*

<sup>0</sup> <sup>2</sup> *<sup>S</sup>*

0 2( )

*built in*

(4)

. (5)

. (5)'

*<sup>B</sup>* of more than 1 eV. It is worth noting that

0 is the

conventional metal Schottky contacts.

expressed by

where 

The workfunctions of the TCPs were estimated using an ultraviolet photoelectron emission spectrometer (AC-3, Riken Keiki Co., Ltd.). Figure 3(b) depicts the photoelectron emission spectra of the PEDOT:PSS and PANI films. The spectra consist of two parts: one with a constant slope and another that linearly increases against the photon energy. The workfunction of PEDOT:PSS and PANI were found to be 5.3 and 5.2 eV, respectively, from Figure 3(b) by assuming that the threshold energy for photoelectron emission is located at the intersection point of the two straight lines that are fitted to the constant-slope and linearly-increasing-slope regions of the plots. These workfunction values show good agreement with those reported previously (Brown et al., 1999; Jang et al., 2008).

Fig. 3. (a) Optical transmittance spectra of PEDOT:PSS and PANI films. The transmittance spectra of ZnO-SnO2 (Satoh et al., 2007) and semi-transparent Pt (Kim et al., 2002)thin films are also shown for comparison. (b) Photoelectron emission yield spectra of PEDOT:PSS and PANI films.

### **5.2 Diode characteristics of transparent conducting polymer/nitride structures**

Figure 4(a) shows the current density-voltage (*J*-*V*) characteristics of the TCP/epi.-GaN samples. The diode ideality factor, *n*, and the SBH, B, were evaluated by fitting the theoretical values obtained using the following equation based on the thermionic emission theory (Crowell, 1965):

$$J = A \, \text{\*} \, T^2 \exp\left(-\frac{\phi\_\text{B}}{kT}\right) \cdot \left[\exp\left(\frac{qV}{nkT}\right) - 1\right] \tag{3}$$

where *q* is the electronic charge, *A\** represents the effective Richardson constant, which is defined as \*2 3 \* 4 *A mk h <sup>e</sup>* (26.4 A/(cm2K2) for GaN), *T* is the absolute temperature, *k* is the Boltzmann constant, *V* is the applied bias, *m\** is the effective electron mass (0.2 *me* for GaN), and *h* is Planck's constant. The n and B values derived using the *J*-*V* characteristics were 3.0 and 0.90 eV, respectively, for PEDOT:PSS/epi.-GaN, and 1.2 and 0.97 eV, respectively, for PANI/epi.-GaN. The low reverse leakage current, which ranged between 10-8 and 10-9 A/cm2 at a reverse bias voltage of -3 V, indicates that the TCP/epi.-GaN

The workfunctions of the TCPs were estimated using an ultraviolet photoelectron emission spectrometer (AC-3, Riken Keiki Co., Ltd.). Figure 3(b) depicts the photoelectron emission spectra of the PEDOT:PSS and PANI films. The spectra consist of two parts: one with a constant slope and another that linearly increases against the photon energy. The workfunction of PEDOT:PSS and PANI were found to be 5.3 and 5.2 eV, respectively, from Figure 3(b) by assuming that the threshold energy for photoelectron emission is located at the intersection point of the two straight lines that are fitted to the constant-slope and linearly-increasing-slope regions of the plots. These workfunction values show good

(Photoelectron yields)

(a) (b)

*mk h <sup>e</sup>* (26.4 A/(cm2K2) for GaN), *T* is the absolute temperature, *k* is

Fig. 3. (a) Optical transmittance spectra of PEDOT:PSS and PANI films. The transmittance spectra of ZnO-SnO2 (Satoh et al., 2007) and semi-transparent Pt (Kim et al., 2002)thin films are also shown for comparison. (b) Photoelectron emission yield spectra of PEDOT:PSS and

**5.2 Diode characteristics of transparent conducting polymer/nitride structures** 

Figure 4(a) shows the current density-voltage (*J*-*V*) characteristics of the TCP/epi.-GaN

theoretical values obtained using the following equation based on the thermionic emission

<sup>2</sup> \* exp exp 1 *<sup>B</sup> qV J AT kT nkT* 

where *q* is the electronic charge, *A\** represents the effective Richardson constant, which is

the Boltzmann constant, *V* is the applied bias, *m\** is the effective electron mass (0.2 *me* for

were 3.0 and 0.90 eV, respectively, for PEDOT:PSS/epi.-GaN, and 1.2 and 0.97 eV, respectively, for PANI/epi.-GaN. The low reverse leakage current, which ranged between 10-8 and 10-9 A/cm2 at a reverse bias voltage of -3 V, indicates that the TCP/epi.-GaN

(arb. units)

1/2

4.5 5.0 5.5 6.0

B, were evaluated by fitting the

B values derived using the *J*-*V* characteristics

(3)

(b)

Photon energy (eV)

PANI

PEDOT:PSS

agreement with those reported previously (Brown et al., 1999; Jang et al., 2008).

PEDOT:PSS

PANI

Pt (Kim et al., 2002)

(Satoh et al., 2007)

0 500 1000 1500 <sup>0</sup>

Wavelength (nm)

samples. The diode ideality factor, *n*, and the SBH,

ZnO-SnO2

20

PANI films.

theory (Crowell, 1965):

defined as \*2 3 \* 4 *A*

GaN), and *h* is Planck's constant. The n and

40

60

Transmittance (%)

80

100

heterojunctions had a Schottky contact property comparable to that exhibited by conventional metal Schottky contacts.

The depletion width, *WD*, in the n-type GaN of the TCP/epi.-GaN heterojunction is expressed by

$$\mathcal{W}\_{D} = \sqrt{\frac{2\varepsilon\_{S}\varepsilon\_{0}}{qN\_{D}}(V\_{Built-in} - V)}\tag{4}$$

where *<sup>S</sup>* is the relative dielectric constant of GaN and equals 8.9 (Wu, 2009), 0 is the vacuum dielectric constant, *Vbuilt-in* is the built-in voltage formed in GaN, *V* is the bias voltage, and *ND* is the donor concentration. The space charge, *QSC*, in the depletion layer is given by *Q qN W SC D D* , thus, the depletion layer capacitance *CD* is obtained by

$$\mathbf{C}\_{D} = \frac{\left| \partial \mathbb{Q}\_{SC} \right|}{\partial V} = \sqrt{\frac{q \varepsilon\_{S} \varepsilon\_{0} N\_{D}}{2(V\_{bulk-in} - V)}} \,. \tag{5}$$

Equation (5) can also be written in the following form:

$$\frac{1}{\left|\mathcal{C}\_{D}\right|^{2}} = \frac{\mathcal{D}(V\_{bulk-in} - V)}{q\varepsilon\_{S}\varepsilon\_{0}N\_{D}}.\tag{5}$$

Equation (5)' suggests that if 1/*CD*2 exhibits linear plots against *V*, *Vbuilt-in* can be obtained at the *V*-intercept of extrapolated fit-line of the plots. Figure 4(b) shows the plot of 1/*CD*2 as a function of the applied voltage. The frequencies used for the capacitance measurements were 100 Hz and 1 KHz for the PANI/epi.-GaN and PEDOT:PSS/epi.-GaN heterojunctions, respectively. The frequency for measurement was chosen within a range that was sufficiently lower than the cut-off frequency, which is described in section 5.4. In Figure 4(b), both the data sets are linear and straight lines were successfully fitted to the data. The determined diode characteristics of the TCP/epi.-GaN heterojunction determined from the *J*-*V* characteristics and capacitance measurements are summarized in Table 2. The observed barrier height was comparable to that obtained by conventional metal Schottky contacts (Tracy et al., 2003). In the case of the conventional metal Schottky contacts, elaborate surface cleaning processes and moderate metal deposition in ultra-high-vacuum conditions are required to attain good Schottky contact with a *<sup>B</sup>* of more than 1 eV. It is worth noting that the good Schottky contact properties in the TCP/epi.-GaN heterojunction were achieved with convenient spin coating of a water-dispersed TCP solution onto the GaN layer in air at ambient temperature.

The observed *<sup>B</sup>*of the TCPs were much lower than expected from the energy difference*<sup>m</sup>* . There are various possibilities for the lower barrier heights including the Schottky effect, which is caused by the electronic mirror force, interface dipole effect, surface defects of GaN, inhomogeneous workfunctions in the TCP film, and/or residual contamination (Sze, 1981; Kampen, 2006). However, the major candidates for the modification of the barrier height have been discussed and are still controversial even in conventional metal/semiconductor Schottky heterojunctions (Tung, 2001). Further detailed investigation is required to determine which effects dominate in lowering the barrier in the TCP/epi.-GaN heterojunction.

Transparent Conducting Polymer/Nitride Semiconductor Heterojunction Solar Cells 317

deterioration of *VOC* and *FF*. The optimization of the deposition process of TCP and introduction of a metal comb-shaped electrode on the TCP layer will improve *VOC* and *FF*. Figure 5(b) depicts external quantum efficiency of the PANI/epi.-GaN heterojunction solar cell. In order to visualize the capabilities of the photovoltaic device, the transmittance of

> 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35

(a) (b)

heterojunction solar cells. (b) External quantum efficiency of PANI/epi.-GaN heterojunction

*JSC*

PANI 170 7.1 0.73 0.41 0.42 0.13 21.2 310.3 PEDOT:PSS 420 3.0 0.80 0.25 0.54 0.11 36.8 17.4

In this study, we found that the capacitance of the TCP/epi.-GaN heterojunction exhibits significant dependence on the frequency of measurement. Figure 6 shows the capacitancefrequency (*C*–*f*) characteristics of the samples. The characteristics were measured under zero-bias conditions. As seen in the graph, the capacitance is constant at a lower frequency; however, it starts to drop at a specific frequency and then rapidly decreases towards the higher frequencies (cut-off). The frequencies at which the capacitance begins to drop are located at ~20 Hz and ~6 kHz for the PEDOT:PSS/epi.-GaN and PANI/epi.-GaN samples, respectively. It is obvious that the difference in the specific frequencies between the two

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6

*Rsh* (k/cm2)

*Rs* (/cm2)

T

SLI

Transmittance (T) (%)

200 250 300 350 400 450

Photovoltaic characteristics Resistivity

(mW/cm2)

(mA/cm2) *FF Pmax*

Wavelength (nm)

EQE

Solar light intensity (SLI)

(Wm-2

nm

)


PANI and the solar light intensity are also plotted as a function of wavelength.

External quantum efficiency

(EQE)

Fig. 5. (a) Photovoltaic characteristics of PANI/epi.-GaN and PEDOT:PSS/epi.-GaN

solar cell, transmittance of PANI (T), and solar light intensity (SLI) as a function of

*VOC*  (V)

Table 3. Photovoltaic characteristics of PEDOT:PSS/epi.-GaN and PANI/epi.-GaN.

**5.4 Frequency-dependent capacitance and its application to deep-level optical** 

0.0 0.2 0.4 0.6 0.8 1.0

PANI/n-GaN PEDOT:PSS/n-GaN

Voltage (V)

Polymer thickness (nm)

**spectroscopy (DLOS)** 

Schottky contact area (mm2)

0.0

wavelength.

0.1

0.2

Current density (mA/cm2

)

0.3

0.4

0.5

Fig. 4. (a) *J*-*V* characteristics and (b) Capacitance-voltage plots of TCP/GaN heterojunction solar cells.


Table 2. Diode characteristics of PEDOT:PSS/epi.-GaN (0001) and PANI/epi.-GaN.

### **5.3 Photovoltaic characteristics of transparent conducting polymer/nitride semiconductor heterojunction solar cells**

Figure 5(a) shows the photovoltaic characteristics (*J*-*V* measurements under AM1.5 light irradiation) of the PANI/epi.-GaN and PEDOT:PSS/epi.-GaN samples. Table 3 represents a summary of the resulting photovoltaic and resistivity characteristics, which include opencircuit voltage (*VOC*), short-circuit current density (*JSC*), maximum output power (*Pmax*), fill factor (*FF*), shunt resistivity, and series resistivity. Note that the *VOC* exhibited high values (>0.5 V), which was much higher than the photovoltage observed in metal Schottky contacts on n-type GaN (Zhou et al., 2007) or PEDOT:PSS Schottky contacts on ZnO (Nakano et al., 2008). The superior photovoltages of the TCP/epi.-GaN heterojunctions are attributed to the following advantages conveyed by our process and substance properties: the ambient temperature fabrication resulted in less process damage and GaN exhibits less electron affinity (3.3 eV) than ZnO (4.4 eV) (Wu et al., 1999).

However, the rather small shunt resistivity and large series resistance that are observed, especially in the PANI/epi.-GaN heterojunction solar cell, are clearly due to the

0

(a) (b) Fig. 4. (a) *J*-*V* characteristics and (b) Capacitance-voltage plots of TCP/GaN heterojunction

*n* 

Schottky contact area (mm2)

PANI 170 7.1 1.2 0.97 39 0.94 PEDOT:PSS 420 3.0 3.0 0.90 40 0.95

Figure 5(a) shows the photovoltaic characteristics (*J*-*V* measurements under AM1.5 light irradiation) of the PANI/epi.-GaN and PEDOT:PSS/epi.-GaN samples. Table 3 represents a summary of the resulting photovoltaic and resistivity characteristics, which include opencircuit voltage (*VOC*), short-circuit current density (*JSC*), maximum output power (*Pmax*), fill factor (*FF*), shunt resistivity, and series resistivity. Note that the *VOC* exhibited high values (>0.5 V), which was much higher than the photovoltage observed in metal Schottky contacts on n-type GaN (Zhou et al., 2007) or PEDOT:PSS Schottky contacts on ZnO (Nakano et al., 2008). The superior photovoltages of the TCP/epi.-GaN heterojunctions are attributed to the following advantages conveyed by our process and substance properties: the ambient temperature fabrication resulted in less process damage and GaN exhibits less electron

However, the rather small shunt resistivity and large series resistance that are observed, especially in the PANI/epi.-GaN heterojunction solar cell, are clearly due to the

Table 2. Diode characteristics of PEDOT:PSS/epi.-GaN (0001) and PANI/epi.-GaN.

**5.3 Photovoltaic characteristics of transparent conducting polymer/nitride** 

0

B (eV)

1

1/

*C*

2 (1013 F-2 cm4

)

*D*

2

3

4

5


Voltage (V)

*WD* (nm) *VBuilt-in* (V)

*J*-*V C*-*V*

 PEDOT:PSS PANI


Polymer thickness (nm)

**semiconductor heterojunction solar cells** 

affinity (3.3 eV) than ZnO (4.4 eV) (Wu et al., 1999).

PANI/n-GaN PEDOT:PSS/n-GaN

Voltage (V)

10-10 10-9 10-8 10-7 10-6 10-5 10-4 10-3 10-2 10-1

Current density (A/cm2

solar cells.

)

1

2

3

Current density (mA/cm2

)

4

5

deterioration of *VOC* and *FF*. The optimization of the deposition process of TCP and introduction of a metal comb-shaped electrode on the TCP layer will improve *VOC* and *FF*. Figure 5(b) depicts external quantum efficiency of the PANI/epi.-GaN heterojunction solar cell. In order to visualize the capabilities of the photovoltaic device, the transmittance of PANI and the solar light intensity are also plotted as a function of wavelength.

Fig. 5. (a) Photovoltaic characteristics of PANI/epi.-GaN and PEDOT:PSS/epi.-GaN heterojunction solar cells. (b) External quantum efficiency of PANI/epi.-GaN heterojunction solar cell, transmittance of PANI (T), and solar light intensity (SLI) as a function of wavelength.


Table 3. Photovoltaic characteristics of PEDOT:PSS/epi.-GaN and PANI/epi.-GaN.

### **5.4 Frequency-dependent capacitance and its application to deep-level optical spectroscopy (DLOS)**

In this study, we found that the capacitance of the TCP/epi.-GaN heterojunction exhibits significant dependence on the frequency of measurement. Figure 6 shows the capacitancefrequency (*C*–*f*) characteristics of the samples. The characteristics were measured under zero-bias conditions. As seen in the graph, the capacitance is constant at a lower frequency; however, it starts to drop at a specific frequency and then rapidly decreases towards the higher frequencies (cut-off). The frequencies at which the capacitance begins to drop are located at ~20 Hz and ~6 kHz for the PEDOT:PSS/epi.-GaN and PANI/epi.-GaN samples, respectively. It is obvious that the difference in the specific frequencies between the two

Transparent Conducting Polymer/Nitride Semiconductor Heterojunction Solar Cells 319

follows. The residual electrons in the deep levels were excluded by applying a reverse bias (-2 V) and extending the depletion layer. Then, the bias was removed for 1 second to fill the deep levels with electrons in the dark. After that, the same reverse bias was again applied to form the depletion layer followed by monochromatic light illumination that excites electrons in the deep levels up to the conduction band. The difference in the capacitance between the

capacitance that is obtained in the filled state in the dark. Figure 7 shows the resulting DLOS spectra. Interestingly, both the spectra acquired at 1 and 10 kHz bias frequency show no characteristic peaks; however, when the bias frequency was increased to 100 kHz, several peaks appeared in the spectrum. This specific frequency, 100 kHz, corresponds to the point where the total capacitance dropped down to a negligible level compared to the capacitance

0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5

**G2**

100kHz -2V

**T3 T2**

NBE

**Incident Photon Energy (eV)**

In Figure 7, five photoemission states are clearly revealed with onsets at ~1.40, ~1.70, ~2.08, ~2.64, and 2.90 eV below the conduction band, which are denoted as T1, G1, G2, T2, and T3, in addition to the near-band-edge (NBE) emissions of GaN at 3.3–3.5 eV. For all the deep levels, electron emission to the conduction band is a dominant process due to their positive photocapacitance transients. The T1, T2, and T3 levels are identical to the deep-level defects that have been commonly reported for GaN, whereas the G1 and G2 levels look like the specific deep levels characteristic of AlGaN/GaN heterointerfaces that were reported recently (Nakano et al., 2008). Using the TCP Schottky contact, we successfully revealed the deep-level states in the near-surface region of the n-GaN layer. These experimental results and further detailed investigations can provide important information on the electronic properties that is needed to improve the performance of the device in optical and electronic

**5.5 Future perspective of TCP/nitride semiconductor heterojunction solar cells** 

In order to increase the output power of TCP/nitride semiconductor heterojunction solar cells, the nitride portion is required to be substituted from GaN to InxGa1-x N. The presumed difficulty in developing the TCP/n-InxGa1-x N heterojunction is the lowering of the barrier height since the electron affinity significantly increases with an increase of the In content.

Incident photon energy (eV)

*C/C*i, where *ND* is the donor concentration and *C*i is the initial

1kHz 10kHz

10 kHz

1 kHz

1kHz10kHz100kHz

100 kHz

*C*. The density of the deep-

*C*, thus, 2*ND*

*C/C*i is

filled state and post-excited states (discharged) was detected as

at 1 and 10 kHz. This means that the *C*i became smaller comparable to

0 V


0V 1.0s

1.0 s

**G1 T1**

0

Fig. 7. DLOS spectra of PANI/epi.-GaN heterojunction solar cell.

4x1016

4

2

6

**2NdC/C0 (cm )**

2

*ND*

C/Ci

(×1016 cm-3)

8x1016

8

effectively enhanced enough to be detectable.

levels is estimated by 2*ND*

fields.

samples is due to differences in the intrinsic properties of the TCPs. Conductivity in TCPs is generated by a polaron in the -conjugated bond; this polarized state causes a Debye-type dielectric dispersion response against an applied alternating electric field (Cole et al., 1941).

Fig. 6. C-f characteristics of TCP/epi.-GaN heterojunction solar cells.

Referring to a previous study on the frequency-dependent capacitance of PANI film (Mathai et al., 2002), the characteristics can be analyzed by assuming an equivalent circuit consisting of a frequency-independent capacitive element, *C*0, in parallel with a resistive element, *R*, both in series with a constant low-value resistance. Based on this model, the frequencydependent capacitance of TCP, *Cp*, is given by the following equation:

$$\mathbf{C}\_p = \mathbf{C}\_0 + \frac{1}{\left(2\pi f \mathcal{R}\right)^2 \mathbf{C}\_0} \tag{6}$$

where *f* is the applied bias frequency.

Furthermore, considering that the capacitance of the depletion layer, *Cd*, is in series with *CP*, then the measured total capacitance of the sample, *Ctotal*, can be expressed by

$$\mathbf{C}\_{total} = \frac{\mathbf{C}\_p \cdot \mathbf{C}\_d}{\mathbf{C}\_p + \mathbf{C}\_d} \,. \tag{7}$$

The solid lines shown in Figure 5 represent the results of the least-square fit of the analytical curve produced based on Equations (6) and (7). The excellent fitting results indicate that the assumed model is adequate. The values of *R* and *C*0, which were derived from the fitting, were 5.3×102 and×10-9 F·cm-2, respectively, for PANI/epi.-GaN and 8.4×104and ×10-9 F·cm-2, respectively, for PEDOT:PSS/epi.-GaN. The large difference in the *R* values between the two samples is reasonable if we take into account the large difference in the conductivity between PEDOT:PSS (5.7×10-1 S/cm) and PANI (3.4×102 S/cm).

We describe below that the transparent Schottky contact fabricated by TCP is applicable not only to the photovoltaic device but also to defect density investigation. Nakano et al. applied deep-level optical spectroscopy (DLOS) to the PANI/epi.-GaN samples (Nakano et al., 2010, 2011a, 2011b). DLOS allows the deep-level density in semiconductors to be estimated by detecting the change in capacitance, which is caused by discharging the deeplevels by exciting electrons with monochromatic light. The measurement process was as

samples is due to differences in the intrinsic properties of the TCPs. Conductivity in TCPs is generated by a polaron in the -conjugated bond; this polarized state causes a Debye-type dielectric dispersion response against an applied alternating electric field (Cole et al., 1941).

Fit-line

<sup>10</sup>-6 PANI/n-GaN

10<sup>0</sup> 10<sup>1</sup> 10<sup>2</sup> 10<sup>3</sup> 10<sup>4</sup> 10<sup>5</sup>

Frequency (Hz)

Referring to a previous study on the frequency-dependent capacitance of PANI film (Mathai et al., 2002), the characteristics can be analyzed by assuming an equivalent circuit consisting of a frequency-independent capacitive element, *C*0, in parallel with a resistive element, *R*, both in series with a constant low-value resistance. Based on this model, the frequency-

0 2

Furthermore, considering that the capacitance of the depletion layer, *Cd*, is in series with *CP*,

*fR C*

*p d*

*p d C C*

*C C*

The solid lines shown in Figure 5 represent the results of the least-square fit of the analytical curve produced based on Equations (6) and (7). The excellent fitting results indicate that the assumed model is adequate. The values of *R* and *C*0, which were derived from the fitting, were 5.3×102 and×10-9 F·cm-2, respectively, for PANI/epi.-GaN and 8.4×104and ×10-9 F·cm-2, respectively, for PEDOT:PSS/epi.-GaN. The large difference in the *R* values between the two samples is reasonable if we take into account the large difference in the

We describe below that the transparent Schottky contact fabricated by TCP is applicable not only to the photovoltaic device but also to defect density investigation. Nakano et al. applied deep-level optical spectroscopy (DLOS) to the PANI/epi.-GaN samples (Nakano et al., 2010, 2011a, 2011b). DLOS allows the deep-level density in semiconductors to be estimated by detecting the change in capacitance, which is caused by discharging the deeplevels by exciting electrons with monochromatic light. The measurement process was as

1 (2 )

0

(6)

. (7)

PEDOT:PSS/n-GaN

10-8

Fig. 6. C-f characteristics of TCP/epi.-GaN heterojunction solar cells.

dependent capacitance of TCP, *Cp*, is given by the following equation:

where *f* is the applied bias frequency.

*C C <sup>p</sup>*

then the measured total capacitance of the sample, *Ctotal*, can be expressed by

*total*

*C*

conductivity between PEDOT:PSS (5.7×10-1 S/cm) and PANI (3.4×102 S/cm).

10-7

Capacitance (F cm-2

)

follows. The residual electrons in the deep levels were excluded by applying a reverse bias (-2 V) and extending the depletion layer. Then, the bias was removed for 1 second to fill the deep levels with electrons in the dark. After that, the same reverse bias was again applied to form the depletion layer followed by monochromatic light illumination that excites electrons in the deep levels up to the conduction band. The difference in the capacitance between the filled state and post-excited states (discharged) was detected as *C*. The density of the deeplevels is estimated by 2*NDC/C*i, where *ND* is the donor concentration and *C*i is the initial capacitance that is obtained in the filled state in the dark. Figure 7 shows the resulting DLOS spectra. Interestingly, both the spectra acquired at 1 and 10 kHz bias frequency show no characteristic peaks; however, when the bias frequency was increased to 100 kHz, several peaks appeared in the spectrum. This specific frequency, 100 kHz, corresponds to the point where the total capacitance dropped down to a negligible level compared to the capacitance at 1 and 10 kHz. This means that the *C*i became smaller comparable to *C*, thus, 2*NDC/C*i is effectively enhanced enough to be detectable.

Fig. 7. DLOS spectra of PANI/epi.-GaN heterojunction solar cell.

In Figure 7, five photoemission states are clearly revealed with onsets at ~1.40, ~1.70, ~2.08, ~2.64, and 2.90 eV below the conduction band, which are denoted as T1, G1, G2, T2, and T3, in addition to the near-band-edge (NBE) emissions of GaN at 3.3–3.5 eV. For all the deep levels, electron emission to the conduction band is a dominant process due to their positive photocapacitance transients. The T1, T2, and T3 levels are identical to the deep-level defects that have been commonly reported for GaN, whereas the G1 and G2 levels look like the specific deep levels characteristic of AlGaN/GaN heterointerfaces that were reported recently (Nakano et al., 2008). Using the TCP Schottky contact, we successfully revealed the deep-level states in the near-surface region of the n-GaN layer. These experimental results and further detailed investigations can provide important information on the electronic properties that is needed to improve the performance of the device in optical and electronic fields.

### **5.5 Future perspective of TCP/nitride semiconductor heterojunction solar cells**

In order to increase the output power of TCP/nitride semiconductor heterojunction solar cells, the nitride portion is required to be substituted from GaN to InxGa1-x N. The presumed difficulty in developing the TCP/n-InxGa1-x N heterojunction is the lowering of the barrier height since the electron affinity significantly increases with an increase of the In content.

Transparent Conducting Polymer/Nitride Semiconductor Heterojunction Solar Cells 321

Chen, X., Matthews, K. D., Hao, D., Schaff, W. J., & Eastman, L. F. (2008). Growth,

Crowell, C. R. (1965). The Richardson constant for thermionic emission in Schottky barrier

Cole, K. S. & Cole, R. H. (1941). Dispersion and absorption in dielectrics I. Alternating

Dahal, R., Pantha, B., Li. J., Lin, J. Y., & Jiang, H. X. (2009). InGaN/GaN multiple quantum

Grabowski, S. P., Schneider, M., Nienhaus, H., Mönch, W., Dimitrov, R., Ambacher, O., &

Green, M. A., Emery, K., Hishikawa, Y., & Warta, W. (2010). Solar cell efficiency tables

Green, M. A. (2001). Third generation photovoltaics: Ultra-high conversion efficiency at low

Ha, Y. H., Nikolov, N., Pollack, S. K., Mastrangelo, J., Martin, B. D., & Shashidhar, R. (2004).

Heeger J. A. (2001). Nobel Lecture: Semiconducting and metallic polymers: The fourth

Jang J., Ha, J., & Kim, K. (2008). Organic light-emitting diode with polyaniline-poly(styrene

Kampen, T. U. (2006). Electronic structure of organic interfaces – a case study on perylene

Kane, E. O. (1962). Theory of Photoelectric Emission from Semiconductors, *Physical Review*,

Kim, J. K., Jang, H. W., Jeon, C. M., & Lee, J.-L. (2002). GaN metal–semiconductor–metal

Kuwahara, Y., Takahiro F.; Yasuharu, F.; Sugiyama, T.; Iwaya, M.; Takeuchi, T.; Kamiyama,

Vol. 127, No. 1, (July 1962), pp. 131-141, ISSN 1943-2879

No. 24, (December 2002), pp. 4655-4657, ISSN 0003-6951

No. 6, (February 2009), pp. 063505-1-063505-3-1693, ISSN 0003-6951

*Letters,* Vol. 78, No. 17, (April 2001), pp. 2503-2505, ISSN 0003-6951

(February 2003), pp. L21-L23, ISSN 1361-6641

(December 2010), pp. 84-92, ISSN 1099-159X

2010), pp. 123-125, ISSN 1099-159X

2001), pp. 681-700, ISSN 0034-6861

pp. 111001-1-111001-3, ISSN 1882-0778

pp. 3152-3156, ISSN 1616-301X

1101

301X

0947-8396

0021-9606

205, No. 5, (July 2003), pp. 1103-1105, ISSN 1862-6300

prepared by RF sputtering, *Semiconductor Science and Technology*, Vol. 18, No. 4,

fabrication, and characterization of InGaN solar cells, *Physica Status Solidi (a)*, Vol.

diodes, *Solid-State Electronics*, Vol. 8, No. 4, (April 1965), pp. 395-399, ISSN 0038-

current characteristics, *J. Chem. Phys.*, Vol. 9, No. 4, (April 1941), pp. 341-351, ISSN

well solar cells with long operating wavelengths, *Applied Physics Letters*, Vol. 94,

Stutzmann, M. (2001). Electron affinity of AlxGa1-xN(0001) surfaces, *Applied Physics* 

(version 37). *Progress in Photovoltaics: Research and Applications*, Vol. 19, No. 1,

cost. *Progress in Photovoltaics: Research and Applications*, Vol. 9, No. 2, (December

Towards a transparent, highly conductive poly(3, 4-ethylenedioxythiophene), *Advanced Functional Materials*, Vol. 14, No. 6, (June 2004), pp. 615-622, ISSN 1616-

generation of polymeric materials, *Review of Modern Physics*, Vol. 73, No. 3, (July

sulfonate) as a hole injection layer, *Thin Solid Films*, Vol. 516, No. 10, (August 2007),

derivatives, *Applied Physics A*, Vol. 82, No. 3, (September 2005), pp. 457-470, ISSN

ultraviolet photodetector with IrO2 Schottky contact, *Applied Physics Letters*, Vol. 81,

S.; Akasaki, I.; & Amano, H. (2010). Realization of nitride-based solar cell on freestanding GaN substrate, *Applied Physics Express*, Vol. 3, No. 11, (October 2010),

One of the most plausible solutions for this issue is to insert a several-tens-nanometer-thick GaN or AIN layer between TCP and n-InxGa1-x N. With this device structure, it is expected that the barrier height at the TCP/nitride semiconductor interface will be maintained at a high value and an internal electric field should be formed.

The cost of the sapphire substrate will become a high barrier for reducing the production cost of III-nitride based solar cells. Matsuki et al. have shown that high quality GaN can be grown on mica plates (Matsuki et al., 2005), which are inexpensive and flexible. Applying such a novel alternative to sapphire for the epitaxial growth substrate will be effective for developing large area TCP/nitride semiconductor heterojunction solar cells.

TCPs have a high transparency from 250 nm to the visible wavelength region, as described in section 5.1. Thus TCP/nitride semiconductor heterojunction photovoltaic devices also have a high potential for applications in ultraviolet sensors.
