larger than the bandwidth of the PDs. The fundamental frequencies used were 10 GHz and 10.1 GHz, and the result of the measurement of the UTC-PD at 40 mA is shown in Figure 3 (right inset). The two tone measurements were repeated for different photocurrents, and the results are presented as the OIP3 versus the photocurrent in Figure 4(a). The OIP3 of the UTC-PD was found to be increasing with bias voltage, and therefore a bias À6 V was chosen, which was found to be the largest safe voltage. The increase of the OIP3 with voltage is consistent with results of [8]. Because of the dynamic range limitations of the ESA, the measurements of PIN-PD and the UTC-PD were limited to photocurrents between 5 and 13 mA, and 15 and 40 mA, respectively. The UTC-PD had about 15-20 dB larger OIP3 in this measurement range. The measurements of the linearity and noise can be added together under the assumption that the photodetector is limiting the link, as mentioned in the introduction, and thus, the SFDR can be calculated and is plotted as a function of photocurrent in Figure 4(b). As a consequence of the high OIP3 of the UTC-PD, the SFDR is more than 10 dB higher than for the PIN-PD.