This mistake has been made by many engineers; the photodiode  may be connected in reverse on a single-ended amplifier. If this happens, the current wants to flow the opposite direction, generating a negative voltage at the output, and since this circuit is single-sided supply (there is no negative power supply), the op-amp output stops at ground and can go no further. If dual power supplies had been used, the designer would have seen the output signal reverse polarity and realize the problem quickly. In any case, flipping the diode around solves the issue. Even with the diode correctly installed, if the diode is in complete darkness and the amplifier type has positive bias currents (out of the pins) and the particular amp in the prototype happens to have an offset voltage with the negative pin above the positive, the amplifier will still be trying to drive the output a little below ground. It can never achieve this, but the internal base or drive currents inside the amplifier will go to maximum values. It also may take a while to get the amplifier to recover from this saturation condition when a light pulse does come along. In this case, a little reverse bias on the diode will ensure that the amp never is driven past ground. Alternatively, just make sure there is a little ambient light to achieve the same result. If at all possible, use a negative supply, the amplifier will operate much better when the input pins are in the middle of the power supply range than all the way at the extremes.