Page 254 - Organic Electronics in Sensors and Biotechnology
P. 254

An Intr oduction to Or ganic Photodetectors     231

                      both small; state-of-the-art op-amps have noise values in the
                      nV/  Hz and fA/  Hz  range.
                    •  Use as large a feedback resistor as possible; although this
                      increases the thermal noise from the feedback resistance, this
                      drawback is outweighed by the improved signal gain.
                    •  Minimize the measurement bandwidth B, ideally matching it
                      to the signal bandwidth to eliminate extraneous noise at other
                      frequencies; this can be done by adding a bandpass filter after
                      the op-amp to reject unwanted frequencies.
                    •  Maximize the shunt resistance of the photodiode, preferably
                      ensuring it is at least as large as the feedback resistor, in order
                      to minimize the effect of the amplifier voltage noise; fortui-
                      tously, maximizing the shunt resistance also minimizes ther-
                      mal noise from the photodiode.
                    •  Minimize the capacitance of the photodiode (again to
                      minimize the effects of amplifier voltage noise).
                   To determine what can feasibly be detected using an organic pho-
               todiode, we estimate the output noise voltage based on typical per-
               formance characteristics for a high-precision low-noise op-amp and
                                               2
               for an organic PV device of area 1 mm . We assume a noise voltage of
               8 nV/  Hz  and a noise current 0.6  fA/  Hz—the data sheet specifi-
               cations for the state-of-the-art AD795 op-amp from Analog Devices.
               We also assume a feedback resistance of 1 GΩ, which is about the
               highest that can be realistically used, and a shunt resistance of 1 GΩ
               and a capacitance of 400 pF for the photodiode. Inserting these values
               into Eq. (6.57) and assuming a modest measurement bandwidth of
               1 kHz, we obtain a root mean squared noise voltage at the output of
               0.4 mV. This is equivalent to a root mean squared current noise at the
               input of 0.4 pA, implying, on the basis of the amplifier performance
               alone, it should be possible to measure currents of ~1 pA and above.
               The photodiode has a thermal noise of 0.1 pA and shot noise of 0.02 pA
               (assuming a photocurrent of 1 pA). The amplifier characteristics are
               therefore the main determiner of sensitivity. The measurement noise is
               lowered to approximately 0.1 pA if the photodiode capacitance is
               reduced by a factor of 10. Increasing the shunt resistance by a factor of
               10 leads to only a marginal reduction in the measurement noise.


               Amplifier Stability
               A final issue to note when using transimpedance amplifiers is stability.
               The photodiode capacitance and the feedback resistance together act as
               a low-pass filter that introduces a phase lag into the feedback loop,
               which approaches 90° at high frequencies. The op-amp itself introduces
               an extra phase lag that also approaches 90° at high frequencies, leading
               to a combined “lag” of almost 180°. Since the signal is fed back into the
   249   250   251   252   253   254   255   256   257   258   259