Recent advances in polymer/silicon heterojunction solar cells

Figure 2: J-V characteristics of two BackPEDOT solar cells with two different SiOx tunneling layer thicknesses. ‘Thin’ (< 1 nm) tunneling oxide is required to achieve a low contact resistance
Table I: Measured one-sun solar cell parameters of thetwo BackPEDOT cells shown in Fig. 2.
Finally, the entire rear surface was metallized with silver by e-beam evaporation. Table I and Fig. 2 show the solar cell performances of the two best cells with a ‘thin’ native SiOx (red triangles) and a ‘thick’ native SiOx (blue squares) tunneling layer, respectively [11]. The Jsc and
Voc values of the two solar cells in Fig. 2 are almost identical. However, we observe a large deviation in the fill factors FF. The solar cell with ‘thick’ SiOx shows a drastically increased series resistance Rs of 2.15 cm2, leading to a low FF of only 73.1% compared to 80.6% and an Rs of 0.61 cm2 for the cell featuring the ‘thin’ (< 1 nm) SiOx tunneling layer. We hence conclude that a too thick SiOx interface layer hampers the transport of holes from the c-Si wafer into the hole-conducting PEDOT:PSS layer and in order to achieve high efficiencies, a sufficiently thin SiOx interface layer has to be chosen to obtain low contact resistances between
PEDOT:PSS and c-Si
3 TWO-SIDES HETEROJUNCTION CELLS
We have recently introduced [11] an attractive alternative cell design with a much higher Voc potential compared to our standard BackPEDOT cell design as shown in Fig. 2, based on combining the PEDOT:PSS hole-selective layer with an a-Si:H(n) electron-selective layer on the other surface of the silicon wafer. Aschematic cross section of such a solar cell is shown inthe inset of Figure 3

Figure 3: J-V characteristic of a two-sides heterojunction BackPEDOT solar cell. In the cell structure shown in Fig. 3, the electroncollector is realized at the cell front using n-type a-Si:H
with an interfacial intrinsic a-Si:H layer to achieve the best possible interface passivation (Voc potential > 730 mV has been demonstrated on reference cells). The front is then coated with a sputtered ITO layer and contacted by a screen-printed low-temperature silver paste. Note that it is also possible to achieve very high Voc values (~700 mV) with an a-Si:H(n) layer without interfacial a- Si:H(i) layer, which is not the case for a-Si:H(p). Hence, it might be an attractive solution to combine a-Si:H(n) as electron-selective layer with PEDOT:PSS as holeselective layer, which would only require one single PECVD process step. The performance of a
representative solar cell of our first batch ofPEDOT:PSS/c-Si/a-Si:H(n) cells is show in Fig. 3.
Although the reached cell efficiency of 15.2% is far below optimal, this is a clear proof-of-principle.However, the full potential of this cell type could not be exploited in our first batch due to a non-optimal front side preparation (too thick a-Si:H and ITO layers limit Jsc, front screen-printed contact limits Rs, non-optimal a-Si passivation limits Voc). Further optimized cell runs are currently under way in our lab and the results will be presented in the near future. 4 LONG-TERM STABILITY
One major challenge of solar cells with polymer/silicon heterojunctions are their long-term stability. As PEDOT:PSS is known to be very hygroscopic, humidity has been found to be a main cause of efficiency instabilities during storage in ambient atmosphere [3]. We have now performed a comprehensive study of the fundamental degradation mechanisms which will be published soon [12]. The main result is that due to a humidity-related drifting of the work function of PEDOT:PSS, the barrier height for holes transported from the silicon into the PEDOT:PSS increases with time and hence the contact resistance increases, decreasing fill factor and efficiency.