Direct current-voltage measurements of the Mo/CuInSe/sub 2/ contact on operating solar cells

Abstract

Many high efficiency CulnSe, based solar cells show blocking or non-ohmic contact behavior in their currentvoltage characteristic which has often been attributed to the Mo/CulnSe, back contact.A novel device configuration is presented which allows the currentvoltage characteristic of the Mo/CulnSe, junction to be analyzed separately from the rest of the operating solar cell. Direct measurements of the back contact on operating CulnSe, based solar cells which demonstrate this blocking behavior show that the Mo/CulnSe, contact is ohmic with negligible contact resistance compared to the total series resistance of the device. I NT R 0 DUCT IO NIMany high efficiency CulnSe, based solar cells have shown blocking or non-ohmic contact behavior in their current-voltage (J-V) characteristic [1][2][3][4]. This "roll-over" of the J-V curve in forward voltage bias is seen at normal operating temperatures in some devices, but is more common and more pronounced at reduced temperatures. The shape of this roll-over behavior in the J-V curve and the temperature at which it appears depend on the processing and electronic properties of the CulnSe,. In many cases this affects the J-V curve in the power quadrant, reducing the fill factor. However, it has been shown that the parasitic J-V behavior does not have any effect on the open circuit voltage of the device [4] and is therefore assumed to occur in series with the primary diode. Because of these properties, the blocking behavior has been attributed to the Mo/CulnSe, back contact. Measurements of the Mo contact to single crystal CulnSe, showed that the contact behavior depended on the defect state concentration of the CulnSe, [5].However, there has also been evidence that the blocking behavior can be light-dependent [4] and may specifically depend on the deposition of the CdS or ZnO layers [6]. In this case, the behavior may originate at one of the interfaces between the ZnO, CdS, and CulnSe,. Numerical modeling has shown that the qualitative features of the blocking behavior could be simulated by a non-ohmic Mo/CulnSe, contact or changes in the electrical properties of the CulnSeJCdS junction [7].In this paper, a novel device configuration is presented which has allowed the Mo/CulnSe, junction to be analyzed separately from the rest of the operating solar cell. Direct measurements of the contact have been completed on a device which shows a non-ohmic blocking contact in its J-V behavior. The results show that the nonohmic behavior does not occur at the Mo/CulnSe, contact. 917 0-7803-3166-4/96/$5 .OO 0 1996 IEEE! EXPERIMENT The CulnSe, in this work was formed by the selenization of Cu/ln precursor layers in a H, S e atmosphere [8]. The MO was deposited by sputtering, the CdS by chemical bath deposition, and the ZnO by rf sputtering. Details are giver1 in [9]. The direct measure of the back contact J-V behavior requires that the CulnSe, is deposited across a gap in the MO back contact. A schematic of the configuration is shown in Fig. 1. The MO has been prepared either by laser scri...

The importance of back contact modification in Cu2ZnSnSe4 solar cells: The role of a thin MoO2 layer

“…Nevertheless, it has been also reported that a MoSe 2 layer could act as buffer layer between the Mo and the absorber, promoting an ohmic contact, and thereby improving the electrical transport. By looking at the absorber/Mo back contact interface of images b and c, it seems obvious that bi-layers might led to a too small MoSe 2 layer, insufficient to promote a good band alignment to avoid hole blocking transport and recombination of minority carriers at this interface [24][25][26]. The MoSe 2 layer can barely be seen by looking at the SEM images, which correlates also with the small values of MoSe 2 thickness obtained by XRF in Table 3.…”

“…Additionally, a thicker MoSe 2 layer can be formed by the uncontrolled reaction between the chalcogen, Mo and CZTSe known as overselenization, and thereby affecting as well the solar cells performance [21]. Nevertheless, despite it is commonly agreed that MoSe 2 can be detrimental for solar cells when present in large amount, it seems also clear that is necessary to allow for a good ohmic contact at the back region [24][25][26]. So far, several layers have been introduced in addition to the Mo back contact with the aim of coping with its instability during the thermal annealing treatments.…”

Spectroscopic investigation of the deeply buried Cu(In,Ga)(S,Se)2∕Mo interface in thin-film solar cells

“…2 One main demand for the back contact is a low series resistance, and usually an Ohmic contact to the absorber layer is believed to be optimal. For the CIGSSe/ Mo interface some authors find an Ohmic contact [3][4][5] whereas others find a Schottky barrier. 6,7 One of the main reasons for this discrepancy is that it is difficult to investigate the real CIGSSe/ Mo interface as it is found in the cell device.…”

“…Therefore the experiments described in the literature have been performed either on idealized systems ͑CuInSe 2 single crystal and evaporated Mo͒ 3,7 or by using indirect techniques. [4][5][6] However, the real CIGSSe/ Mo interface is probably much more complicated than can be assessed by such approaches. In particular, it is a challenge to investigate the chemical and electronic properties of thin deeply buried interfaces by surface-sensitive techniques.…”

Characterization of the Electronic and Chemical Structure at the Thin Film Solar Cell Interfaces: June 2005 -- June 2009

“…However, it is not clear whether the Cu-poor surface phase at the CIGSSe Back Surface is present from the beginning or whether its formation is induced by the lift-off process and subsequent exposure to N 2 atmosphere in the glovebox/glovebag or UHV, respetively. In the past it was discussed whether the chalcopyrite/Mo interface results in an Ohmic contact [23,28,29] or whether a Schottky contact [30,31] is formed. Kohara et al even suspected the observed MoSe 2 layer to be responsible for an Ohmic contact at the chalcopyrite/Mo interface [23].…”