Counter electrode dependence of germanium-sensitized thermal cells

Semiconductor-sensitized thermal cells (STCs), which generate electricity by converting the photoexcitation of dyes in a dye-sensitized solar cell (DSSC) into thermal excitation in a semiconductor, have attracted attention as a new thermal energy conversion technology. This paper examines the role of the counter electrode (CE) on the STC battery characteristics. The results suggest that, similar to DSSCs, the chemical stability, surface resistance, and electron transfer resistance at the CE/electrolyte interface affect the performance of STCs. The similarities between STCs and DSSCs partly shown in this manuscript indicate that the scientific arguments of DSSCs may be applicable to the STC discussion.


Introduction
The uses of renewable energy, such as solar, 1) wind, 2) and tidal 3) energy, have recently attracted considerable attention 4) due to both environmental and resource issues.Among these appealing energies, we focus on the thermal energy always present in the human range of action. 5)emiconductor-sensitized thermal cells (STCs) [6][7][8][9][10][11][12][13][14] have previously been proposed as a new thermal energy conversion technology based on the mechanism of dye-sensitized solar cells (DSSCs).[15][16][17][18] STCs generate electricity by the redox reaction of electrolyte ions by thermally excited charges in a semiconductor (Fig. 1).19,20) The structure of this device is straightforwardwedged between the semiconductor and counter electrodes (CEs) is an electrolyte.The power generation stops when the chemical reactions reach thermal equilibrium at the set temperature. After he termination, the reduced reaction at the CE is stopped and moved to another equilibrium state by opening the electric circuit.As a result, the STCs can be discharged again.6,14) Germanium (Ge), 6,11,13,14,20) β-FeSi 2 , 10,19) CuFeS 2 , 12) Ag 2 S, 7) and organic perovskites 8,9) have been investigated as semiconductor electrodes for generating thermally excited charges in STCs.On the other hand, the counterpart, the CE materials, have not been discussed.
Let us explain the difference between the DSSC and STC in detail.The DSSC, a known chemical-type solar cell, also called a 'Graetze cell', 21) generates electricity by photonexcited carriers from the photo-sensitizer, such as dye molecules or quantum dots.In the DSSC, the n-Si/Ge electrode part in Fig. 1 is typically a TiO 2 /photo-sensitizer electrode.The roles of the other partsthe electrolytes and CEare theoretically the same.
A favorable CE for DSSCs has low sheet resistance and excellent catalytic activity for reducing electrolyte ions. 22,23)f the same tendency could be found in STCs, it would lead us to an analogy of DSSCs and STCs and construct the STC principle.
In this paper, to investigate the CE, STCs were fabricated using nine different CE materials, n-Si/Ge, Al, n-Si, SUS304, Cu, fluorine-doped tin oxide substrate (FTO), indium-doped tin oxide substrate (ITO), Au, and Pt in Ge-STCs, 6,11,13,14) in which Ge was used as the semiconductor electrode.Their power generation performances were evaluated using electric and electrochemical measurements at 80 °C.Here, the Ge-STC is the most-studied STC because of its stable battery characteristics thanks to the polymer electrolyte, including CuCl and CuCl 2 as the electrolyte ions.As a result, we were able to identify a CE role in STCs similar to that in DSSCs.

Materials
A commercial n-Si/Ge wafer (Ge layer, 2 μm) supplied by Tohnic Corporation, Japan, was used as the semiconductor electrode.A Cr layer (20 nm) was deposited on the n-Si wafer (0.02 Ω cm, 15 × 20 × 0.5 mm) to improve the adhesion of the n-Si/Ge interface.At RT, the resistivity of the n-Si and Ge sides were 3.18 and 3.27 Ω/sq, respectively, measured by the four-terminal method (Loresta GP MCP-T600, Mitsubishi Chemical.Co.) applying a voltage of 90 V.
The electrolyte was prepared in a flow-type glove box, coupled with a circulation/purification machine, and filled with Ar (oxygen concentration 0.2-0.5 ppm).Before the reagents were used in the glove box, they were vacuumed at 80 °C for 15 h in a side box.Furthermore, molecular sieves (Fujifilm Wako Pure Chemical Industries, Ltd) were heated in an electric furnace for 3 and 4 h at 180 °C and 350 °C, respectively.After introducing the reagent in the glove box, CuCl, CuCl 2 , and NaCl electrolytes were prepared at a concentration of 0.25, 0.25, and 0.6 mmol g −1 , respectively, using PEG600 as a solvent.The components were mixed for 10 min using a mortar and pestle.

Cell assembly
A double-sided insulating tape (114 μm thickness) with a hole (6 mm diameter) was attached to the CE.The n-Si/Ge substrate was immersed in a solution of 5% HF for 5 min.Next, it was rinsed three times with deionized water for 1 min to remove oxides on the surface.The CEs and Ge substrates were inserted in the glove box, where all the cell fabrication processes were performed.The electrolyte was dropped in the hole of the insulating tape on the CE.Finally, the n-Si/Ge substrate was placed to cover the hole [Fig.2(a)].The Ge side was in contact with the electrolyte, and the obtained cells were named based on their corresponding CE (FTO cell, ITO cell, n-Si cell, n-Si/Ge cell, Al cell, Cu cell, SUS304 cell, Au cell, and Pt cell).
Reference cells, CE/electrolyte/CE cells using the same electrolyte mentioned above, were also fabricated.

Battery characteristics
Cyclic voltammetry (CV) measurements were conducted on each cell using a VSP-300 instrument (Bio-Logic Science Instruments).The cells were set in a constant temperature bath at 80 °C and connected to the VSP-300.The n-Si side of the n-Si/Ge was connected as the WE.The conductive surface of the CE was connected as the CE.The scan rate was 10 mV s −1 , and the sweep direction was from the open-circuit voltage (V OC ) to 0 V.If the thermally excited electrons of the semiconductor move to the CE and reduce ions as assumed in Fig. 1, we would observe a negative open-circuit voltage and a positive short-circuit current using this connection method.
V OC and chronopotentiometry (CP) measurements were taken to obtain long-term operation measurements.In the CP method, the setting current was selected from the CV results.The discharge capacity was derived from the CP measurements.When the voltage dropped to 0, V OC was recorded every 30 s.
Potentiostat electrochemical impedance spectroscopy (PEIS) was performed to obtain the charge transfer resistance (R ct ) at the CE/electrolyte interface in a temperature bath set at 80 °C.The equivalent circuit [Fig.2(b)] was determined using reference cells.In PEIS, the amplitude of the AC was set to 10 mV.The measurement frequency ranged from 7 Hz to 50 mHz.ZView software (Scribner Associates) was used for the fitting.

Results and discussion
The results of power generation and the first discharge capacity are summarized in Table I.Power generation was confirmed in the ITO, FTO, Au, and Pt cells but not in other cells.

Non-dischargeable electrodes
First, let us discuss the cells that did not generate power.Those CVs are summarized in supplementary material Fig. 1.In the n-Si/Ge cell, the n-Si side of the n-Si/Ge CE is in contact with the electrolyte, and the circuit connection is on the n-Si side.In this case, both the V OC and short-circuit current (J sc ) were almost zero.It can be considered that less redox reaction occurred at the CE/electrolyte interface since the bandgap of n-Si is located on the redox potential of the electrolyte ions. 29)The electronic band of n-Si [VB = + 0.40 V and conduction band = −0.70V versus normal hydrogen electrode (NHE)] was not located at the reduction level of Cu 2+ (+ 0.20 V versus standard hydrogen electrode (SHE) at 80 °C, 30) almost equal versus NHE).
In the case of the n-Si cell, for the same reason as the n-Si/ Ge cell, J sc was almost zero, while V OC could be detected since the WF was different between the WE and CE (the discharge observed in Table I may be an electric double-layer discharge caused by the different WF).Table I.Power generation characteristics dependence on the CE material.

CE material
Power generation Discharge capacity (mA •h/g) In the case of the Al, Cu, and SUS304 cells, opposite-sign voltage and current were observed, i.e. positive V OC and negative J sc were observed.This suggests the elution of the CE materials, Al, Cu, and SUS304, which have a higher ionization tendency (i.e. a smaller WF) than the electrolyte ions, Cu + /Cu 2+ (Fig. 1).
This result confirms that STC generates electricity by reducing Cu 2+ at the CE/electrolyte.

Dischargeable electrodes
Next, the four cells that could generate electricity are discussed.Each CV is shown in Figs.3(a)-3(d).For V OC , there was no significant difference between the CEs [Fig.4(a)].Our previous report 11) and the present results confirm that V OC in STC is between the Fermi level of the WE and the redox reaction level at the electrolyte/CE interface.It should be noted that V OC was measured using almost zero discharge current and does not consider the voltage drop that depends on the material resistance.
Let us discuss J sc .STCs are similar to DSSCs regarding the reduction of ions at the CE.As mentioned in the Introduction, a favorable CE for a DSSC has low sheet resistance and excellent catalytic activity for reducing electrolyte ions. 22,23)Therefore, we plotted J sc versus the sheet resistance and electron transfer resistance R ct at the cathode/ electrolyte interface [Figs.4(b) shows the relationship between J sc and the sheet resistance of each CE.As the sheet resistance decreased, J sc tended to increase, similar to DSSCs.This was because of the current-resistance drop and suggests that the power generation was not caused by the CE dissolution.
To measure R ct , we obtained Cole-Cole plots for each cell (Supplementary Material Fig. 2).All plots comprise one resistance-capacitor half-circle and one semifinite ion-diffusion element suggested by the monotonical resistance increase in the low-frequency range. 31)Thus, we fabricate the equivalent circuit as shown in Fig. 2(b) composed of the circuit series resistance (R s ), diffusion Warburg impedance (W o ), constant phase element (CPE), and R ct .By fitting the Cole-Cole plots to the equivalent circuit, we obtained each R ct value.Figure 4(c) shows the relationship between J sc and R ct .The results show that as R ct decreases, J sc increases.
Let us discuss the relationship between R ct and the catalytic activity for reducing the electrolyte ions.R ct is related to the magnitude of DC at the electrode/electrolyte interface, i.e. the ease of electron exchange.The larger the R ct , the more challenging it is for electrons to exchange, i.e. the harder it is for the reduction reactions to occur.Thus, in a broad sense, it is related to the catalytic activity.In other words, this R ct is directly related to the catalytic activity for reducing the redox electrolyte ions, so these results suggest that the CE role in STCs is the same as in DSSCs.In general, the amount of d-orbital electrons is related to the catalytic activity, but the catalytic activity is completely dependent on the respective chemical reaction.

Conclusions
This paper investigates the influence of the CE material on STCs.The results suggest that similar to DSSCs, the chemical stability, surface resistance, and R ct of the CE material affect the performance of STCs.The similarities between STCs and DSSCs partly shown in this manuscript indicate that the scientific arguments of DSSCs may apply to the STC discussion.

Fig. 1 .
Fig. 1.Schematic of the STC and the CEs examined in this paper.The direction of electron flow is indicated by the arrows.

Fig. 2 .
Fig. 2. Schematic of the fabricated Ge-STC (a) and the equivalent circuit for the analysis (R S : circuit series resistance, R ct : electron transfer resistance, W o : diffusion Warburg impedance, CPE) is shown in (b).

Fig. 3 .
Fig. 3. CVs of the cell using FTO (a), ITO (b), Au (c), and Pt (d) as the CE material.The CE served as the standard in this measurement.

Figure
Figure 4(b) shows the relationship between J sc and the sheet resistance of each CE.As the sheet resistance decreased, J sc tended to increase, similar to DSSCs.This was because of the current-resistance drop and suggests that the power generation was not caused by the CE dissolution.To measure R ct , we obtained Cole-Cole plots for each cell (Supplementary Material Fig.2).All plots comprise one resistance-capacitor half-circle and one semifinite ion-diffusion element suggested by the monotonical resistance increase in the low-frequency range.31)Thus, we fabricate the equivalent circuit as shown in Fig.2(b) composed of the circuit series resistance (R s ), diffusion Warburg impedance (W o ), constant phase element (CPE), and R ct .By fitting the Cole-Cole plots to the equivalent circuit, we obtained each R ct value.Figure4(c)shows the relationship between J sc and R ct .The results show that as R ct decreases, J sc increases.Let us discuss the relationship between R ct and the catalytic activity for reducing the electrolyte ions.R ct is related to the magnitude of DC at the electrode/electrolyte interface, i.e. the ease of electron exchange.The larger the R ct , the more challenging it is for electrons to exchange, i.e. the harder it is for the reduction reactions to occur.Thus, in a broad sense, it is related to the catalytic activity.In other words, this R ct is directly related to the catalytic activity for reducing the redox electrolyte ions, so these results suggest that the CE role in STCs is the same as in DSSCs.In general, the amount of d-orbital electrons is related to the catalytic activity, but the catalytic activity is completely dependent on the respective chemical reaction.