In our study of photochemistry of TiO2/Pt, we found that the photocurrent density increasing as the applied bias gradually became negative. Generally, bias is used to consume the yeilded photogenerated holes. Why would a more negative bias result in an increase in photocurrent density?