Effect of polymer layer deposition and annealing on photovoltaic properties of CuIn[S.sub.2]/polymer structures/polumeerikihi ja termilise kasitluse moju uurimine struktuuri CuIn[S.sub.2]polumeer fotovoltomadustele.

INTRODUCTION

Polycrystalline thin film technology is one of the most promising technologies for the future. The chalcopyrite I-III-[VI.sub.2] semiconductors are now established as effective absorbers in thin film photovoltaic cells. The copper-indium chalcogenide materials, e.g. CuIn[Se.sub.2] (CISe), CuIn[S.sub.2] (CIS), Cu(InGa)[Se.sub.2] (CIGSe), and CuIn[[Se.sub.2] (CISSe), have been developed most extensively because they allow tailoring of the energy band gap and other material properties to enhance device performance [1-4].

Our previous studies [5,6] have shown that the polymer layer deposition onto the free surface of photo-absorber layers of CIS leads simultaneously to a small decrease in the surface photovoltage ([V.sub.s]) and to a significant increase in the short-circuit photocurrent (I.sub.sc]). This can be explained by the decrease in the potential barrier height and increase in the efficiency of charge carriers transport through the CIS/polymer interface. The values of the [V.sub.s] and [I.sub.sc] change depend strongly on the molecular structure of the polymer [6]. Nevertheless, this does not lead to changes in spectral dependences of [V.sub.s] and [I.sub.sc]. Therefore a high concentration of the centres of charge carriers trapping and recombination remains on the CIS/polymer interface. These centres reduce the photosensitivity of CIS/polymer heterostructures in the energy (E) region of 2.0-2.5 eV [6].

According to our assumption, the main reasons for the high concentration of charge carriers trapping and recombination centres are the adsorption of oxygen molecules and solvent molecules in the process of polymer film formation. It is known [7] that both residual gases and solvent molecules are most effectively removed at the glass transition temperature of the polymer ([T.sub.g]). Therefore, the annealing of structures at temperatures near or higher than [T.sub.g] can result in a strong decrease in the concentration of the centres of charge carriers trapping and recombination on the CIS/polymer interface and a drastic rise in [V.sub.s].

Thus, the aim of this paper is to study how the annealing of CIS-based structures with deposited layers of different polymers affects photovoltaic properties of these structures.

EXPERIMENTAL

CIS film preparation on a copper tape (CISCuT) by the method of indium electrodeposition, with following sulphurization is described in detail in [8].

To prepare CISCuT/polymer structures, we used the following polymers: polystyrene (PS), co-polymer of styrene with octylmetacrylate (TPN-10) containing 80% styrene, poly-N-epoxypropylcarbazole (PEPC), and poly(3-iodine-9-vinylcarbazole) octylmetacrylate (P-31-9VC:OMA) (Fig. 1). PS, TPN-10, and PEPC polymers were produced by the 'Azot' plant (Severodonetsk, Ukraine). The molecular weight of PEPC is 900-1000 g mol-1, of PS--2 x [10.sup.5] g [mol.sup.-1]. P-319VC:OMA was synthesized in the Polymer Department of Kyiv National University, as reported in [6]. This polymer has a molecular weight of 5000-7000 g mol-1 and contains 70 mol% 3I-9VC monomer in the copolymer structure [6].

The values of [T.sub.g] for as-deposited polymers are as follows: PS -100 [degrees]C, TPN--75 [degrees]C, PEPC--70 [degrees]C, and P-3I-9VC:OMA--75 [degrees]C [9].

Polymer films were deposited by the method of spin casting from the solution of the respective polymer in 1,2-dichloroethane. Samples were annealed in the drying box in the temperature range of 50-1300C during 12 h.

In the study of the surface morphology of CISCuT substrates and the polymer layer deposited onto CISCuT, a serial atomic force microscope (AFM) of NanoScope IIIa type was used in the periodical contact mode. The technique of [V.sub.s] measurements and experimental setup for these measurements is described in [5,6].

[FIGURE 1 OMITTED]

RESULTS AND DISCUSSION

Our studies show that the deposition of the polymer layer onto the free surface of CISCuT film in all cases leads to a small decrease in [V.sub.S] (except for PEPC, where [V.sub.S] is slightly higher, but in the range of the experiment accuracy of [+ or -] 2.5%), while the annealing causes an increase in [V.sub.S] for all studied structures. Figure 2 shows [V.sub.S] spectra for the CIS film (curves 1) and for the CIS/TPN-10 (Fig. 2a) and CIS/P-31-9VC:OMA (Fig. 2b) structures before (curves 2) and after (curves 3) annealing at 130[degrees]C.

[FIGURE 2 OMITTED]

It can be seen that the change in the [V.sub.S] value after annealing significantly depends on the molecular structure of the polymer. The data in Fig. 2a and Fig. 2b show that [V.sub.S] increase after annealing is a few times greater for CIS/P-3I-9VC:OMA structures (polymer with high flexibility of macromolecules) than for the structures with the PS layer, where the flexibility of macromolecules is lower. It should be noted that after annealing the [V.sub.S] value for CIS/P-3I-9VC:OMA structures becomes even higher than [V.sub.S] of the free surface of the CIS layer. The above results and comparison with analogous data for PEPC and TPN-10 polymers show that the change in [V.sub.S] as a result of annealing can be associated with the elastic properties of the polymer layers. These properties are defined by the mobility of polymer macromolecules, which strongly increases at temperatures higher than [T.sub.g] due to an increase in the flexibility of polymer macromolecules.

[FIGURE 3 OMITTED]

The increase in the mobility of polymer macromolecules (that can be the elasticity measure) of used polymers has the following trend: PS, PEPC, TPN-10, P-3I-9VC:OMA [7].

In order to clarify reasons for such correlation, we studied the surface morphology of the prepared structures. Figure 3 shows the AFM plots of CIS layers before (a) and after deposition of the P-3I-9VC:OMA layer (b), and the PS layer before (c) and after (d) annealing.

Comparison of the images in Fig. 3 shows that the deposition of more elastic P-3I-9VC:OMA onto a polycrystalline CIS surface (Fig. 3a) leads to surface smoothing in the case of an amorphous polymer film (Fig. 3b). The picture was similar for the PEPC polymer layer. On the other hand, a significant amount of micropores was observed on a relatively smooth surface when we deposited the PS layer with a smaller flexibility of polymer macromolecules (Fig. 3c), which are filled during the process of annealing (Fig. 3d). Also, micropores disappeared after the annealing of the structures with the TPN-10 polymer layer. The changes in [V.sub.S] and the surface roughness parameters are summarized in Table 1.

In Table 1 the value [V.sub.1] is the value of [V.sub.S] at E =1.5 eV (the region of weak absorption), [V.sub.max] is the value of [V.sub.S] at ca 1.65 eV, defined mainly by the rate of surface recombination of charge carriers on shallow surface centres. To characterize the efficiency of trapping onto deep surface centres, the value of the slope [DELTA][V.sub.S] / [V.sub.S] =([V.sub.2]- [V.sub.3])/ [V.sub.2], where [V.sub.2] is the value of [V.sub.S] at E = 2.0 eV (the beginning of a drastic decrease in [V.sub.S] (E) dependence), and [V.sub.3] is the value of [V.sub.S] at E = 2.5 eV. Surface roughness parameters ([Z.sub.range] peak-to-valley difference in height values and RMS--root mean square average of height deviations) were analysed using fragments with 5 x 5 [micro][m.sup.2] dimensions by the help of routine AFM tools and software.

After the annealing of the structures, the change in [V.sub.S] most probably occurs owing to the change in the barrier height on the CISCuT/polymer interface and to the change in the rate of the direct recombination of charge carriers on superficial centres. These centres are formed, for example, due to the adsorption of molecules of oxygen or solvent and effective trapping of non-equilibrium charge carriers on deep captures with their following recombination.

To estimate the contributions of these processes, we used the basic regularities of photovoltage spectrum formation in the region of weak absorption, where VS is defined mainly by the value of bands bending. On the other hand, the high rate of direct recombination of charge carriers on the surface centres should give VS (E) dependence on saturation in the region of strong absorption [10].

A drastic decrease in the short-wavelength region can only be caused by the trapping of non-equilibrium charge carriers onto deep surface centres, i.e. so-called trapping limited photovoltage rise [10].

Comparison of the obtained data (Table 1) shows that during annealing all structure bands bend and the decrease in the recombination rate is magnified at the CISCuT/polymer interface, due to the desorption of oxygen and solvent molecules during the annealing process. It should be noted that the concentration of deep trapping centres proportional to the parameter [DELTA] [V.sub.S] /[V.sub.S] is practically not changed [5,6,10].

To determine the range of temperatures where the process of desorption is most effective, we studied the properties of structures with PEPC and P-3I-9VC:OMA polymer layers after annealing at different temperatures. We found that in the temperature range of 75-100[degrees]C the influence of annealing is greater and magnification of [V.sub.max] is higher than the values of VS for a free CIS surface. The spectral dependences of VS for structures with the layers of PEPC and P-3I-9VC:OMA after annealing at different temperatures are given in Figs 4 and 5.

Figures 4 and 5 show clearly that VS changes essentially in the temperature range of 75-100[degrees]C, i.e. the process begins at the point of [T.sub.g] at which the effective desorption of oxygen and solvent molecules begins. When the annealing temperature is higher than 100[degrees]C, the change in [V.sub.S] is insignificant and frequently surface roughness is magnified and even defects on the polymer surface are observed (Fig. 3d). A characteristic feature of structures with a P-3I-9VC:OMA layer is an essential decrease in [V.sub.S] after annealing at 50[degrees]C. For the structures with other polymer layers the indicated changes are small and the process is ineffective in the applied temperature range. Measurements on AFM have shown that the surface of a film P-3I-9VC:OMA becomes rougher only after annealing at 500C. It becomes smoother before annealing and after annealing at temperatures higher than 75 [degrees]C. Probably, at this temperature the pores for the desorption of gases are already being formed, but the efficiency of desorption is still low. At higher temperatures not only desorption of gases is observed, but the mobility of molecules is already sufficient for filling pores.

[FIGURE 4 OMITTED]

[FIGURE 5 OMITTED]

CONCLUSIONS

The process of thermal annealing has significant influence on the morphology and photovoltaic parameters of hybrid CISCuT/polymer structures. Thus, active desorption of air (oxygen) and solvent from the polymer layer and subsequent filling of pores by molecules of a polymer take place during the annealing of the CISCuT/polymer structures at temperatures close to [T.sub.g] (75-100[degrees]C for different polymers). This leads to significant decrease in surface roughness, increase in the height of a barrier, and reduction of the recombination rate at shallow surface levels, caused by the adsorption of molecules of oxygen and solvent.

However, annealing practically does not influence the concentration of deep surface trap centres of charge carriers, which define the decrease in photovoltage in the E region from 2 eV to 3 eV. The efficiency of the processes described above increases in the polymer range PS, PEPC, TPN-10, P-3I-9VC:OMA, i.e. it is connected with the flexibility of polymer macromolecules of a polymeric compound.

ACKNOWLEDGEMENTS

The authors thank Dr J. Penndorf and Dr I. Konovalov for the submitted CISCuT substrates. The work was supported by INTAS grant Ref. No. 03-51-4561 and Estonian Science Foundation (grants 7595 and 7669).

Received 14 May 2008, revised 9 September 2008, accepted 10 September 2008

REFERENCES

[1.] Catalano, A. Polycrystalline thin-film technologies: status and prospects. In Proceedings of the 1st IEEE World Conference on Photovoltaic Energy Conversion (Hawaii), 1994,52-59.

[2.] Schock, H. W. Solar cells based on CuInSe2 and related compounds: recent progress in Europe. Sol. Energy Mater. Sol. Cells, 1994, 34, 19-26.

[3.] Miles, R. W., Zoppi, G., and Forbes, I. Inorganic photovoltaic cells. Mater. Today, 2007, 10, 11, 20-27.

[4.] Shah, A., Torres, P., Tscharner, R., Wyrsch, N., and Keppner, H. Photovoltaic technology: the case for thin-film solar cells. Science, 1999, 285, 692-698.

[5.] Verbitsky, A., Vertsimakha, Ya., Lutsyk, P., Studzinsky, S., Bereznev, S., and Kois, J. Properties of CuInSZ free surface and effect of organic layers. Semiconductors, 2006, 40, 197-201.

[6.] Verbitsky, A., Vertsimakha, Ya., Lutsyk, P., Studzinsky, S., Bereznev, S., Kois, J., Opik, A., and Mellikov, E. Properties of CuInSZ free surface and effect of conductive polymer layers on these properties. Proc. Estonian Acad. Sci. Chem., 2006, 55, 111-119.

[7.] Kargin, V. A. (ed.). The Encyclopedia of Polymers (3 volumes). Soviet Encyclopedia, Moscow, 19721977. (Vol. 1.--1972, Vol. 2--1974, Vol. 3--1977) (in Russian).

[8.] Penndorf, J., Winkler, M., Tober, O., Roser, D., and Jacobs, K. CuInSZ thin film formation on a Cu tape substrate for photovoltaic application. Sol. Energy Mater. Sol. Cells, 1998, 53, 285-298.

[9.] Lipatov, Yu. S. (ed.). The Handbook of Polymers Physical Chemistry (3 volumes). Naukova Dumka, Kiev, 1984 (in Russian).

[10.] Dmitruk, N. L., Kruchenko, Yu. V., Litovchenko, V. G., and Popov, V. G. Surface photovoltage in semiconductors at deviation from quasi-equilibrium. Poverkhnost'Fiz., khim., mekh., 1986, 5, 56-62 (in Russian).

Anatoly B. Verbitsky [a], Yaroslav Vertsimakha [a], Sergei Studzinski [b], Sergei Bereznev [c] *, Igor Golovtsov [c], Julia Kois [c], Andres Opik [c], and Oksana Lytvyn [d]

[a] Institute of Physics of Ukrainian NAS, Prosp. Nauki 46, 03028, Kyiv, Ukraine

[b] National Taras Shevchenko University, Volodymyrska Sir. 64, 01033, Kyiv, Ukraine

[c] Department of Materials Science, Tallinn University of Technology, Ehitajate tee 5, 19086 Tallinn, Estonia

[d] Lashkarev Institute of Semiconductor Physics of Ukrainian NAS, Prosp. Nauki 41, 03028, Kyiv, Ukraine

* Corresponding author, [e-mail]

Table 1. The effect of annealing onto the surface and
photo-voltaic parameters of CISCuT/polymer structures

                      [V.sub.1]   [V.sub.max]   [DELTA][V.sub.s]/
Structure                mV           mV            [V.sub.s]

CISCuT                  10.62        16.47            0.409
PEPC/CISCuT              9.87        16.7             0.383
PEPC/CISCuT after       11.93        20.11            0.396
  annealing
P-31-9VC:OMA/CISCuT      9.06        15.57            0.455
P-31-9VC:OMA/CISCuT     13.84        22.33            0.417
  after annealing
PS/CISCuT                9.2         13.49            0.373
PS/CISCuT after          9.62        16.03            0.485
  annealing
TPN-10/CISCuT            8.3         13.48            0.487
TPN-10/CISCuT after     10.14        16.46            0.453
  annealing

                      [Z.sub.range],    RMS,
                            nm           nm
Structure

CISCuT                   872.14       128.21
PEPC/CISCuT                6.76         0.61
PEPC/CISCuT after          5.29         0.64
  annealing
P-31-9VC:OMA/CISCuT        4.48         0.47
P-31-9VC:OMA/CISCuT        2.02         0.25
  after annealing
PS/CISCuT                 22.51         1.71
PS/CISCuT after            9.62         0.68
  annealing
TPN-10/CISCuT             24.07         2.68
TPN-10/CISCuT after       11.39         1.12
  annealing