4.1

Bandwidth measurements

The bandwidth measurements of the CoC have been performed using an Agilent 86032A light wave component analyzer. The CoC is contacted using an MPI 67 GHz Titan RF Probe needle and is biased through an Agilent 11612N bias-T. The average optical power is adjusted using an Agilent N7752A optical attenuator. The photodiodes frequency response has a lowpass characteristic with a 3 dB bandwidth of approximately 21.4 GHz at low optical input power of - 10 dBm and reverse bias voltage of 5 V (see Figure 2). This demonstrates the capability of the photodiode chip to operate at low optical input power without relying on self-biasing effects. The measurement at 95°C show no degradation in bandwidth, which means, that temperature alone is not affecting the carrier transport as long as sufficient large electric field and no optical saturation effect is present.

Figure 2

Frequency response of CoC at - 10 dBm optical input power and 5 V reverse bias voltage, Left: at 25°C, Right: at 95°C.

4.2

Optical saturation current measurements

Linearity is an ambiguous term. In the context of this work, we are determining the linear regime of the photodiode by measuring the 1 dB compression point. This is determined as a function of the RF output vs. the RF input power where the RF output power at the fundamental frequency decreases by 1 dB or 10%, respectively from its linear relationship with respect to the RF input power. At this point the device goes into optical saturation and generates increased higher order harmonics or non-linearities. The optical RF input power is a function of its extension ratio (ER) and is proportional to the average optical input power which is itself proportional to the average photocurrent generated by the photodiode. Therefore, for a fixed ER, it is common to describe the linearity of the photodiode by the average photocurrent at the 1 dB compression point. This is called the optical saturation current.

The optical saturation current measurements (see Figure 3) are obtained using a Keysight E8361A PNA, an iXblue 70 GHz 1550 nm VNA Modbox with ER~14 dB, an IPG Photonics EAD-500-C EDFA and an Agilent N7752A optical attenuator for average optical power measurements. The CoC is mounted on a RF mount and biased through the internal bias-T of the PNA. The bias voltage is carefully adjusted to guarantee the voltage at the anode and cathode of the CoC. The sample is mounted on a 3D motorized stage allowing accurate optical alignment with a standard single mode fiber (SMF), see Figure 4. The measurement is done at different distances (d) between the light entry of the CoC and the tip of the SMF. By increasing the distance between the fiber and the chip, the optical power density on the detector is decreased. The results are obtained at room temperature without any active cooling of the device.

Figure 3

Experimental setup for RF measurement of CoC on 3D motorized stage.

Figure 4

SMF alignment on the light entry of the CoC using 3D motorized stage.

The relative RF output power at 20 GHz as function of average photo-current and at different distances of the SMF from the backside of the photodiode chip is shown in Figure 5. The results show that the optical power density on the active area impacts the optical saturation current. At 150 μm distance between the SMF tip and the backside of the photodiode chip, the optical saturation current of the photodiode is 70 mA at 8 V. This corresponds to an average optical input power of +20.8 dBm. Furthermore, no significant degradation of coupling efficiency has been detected up to a distance of 150 μm, but we observed an increased responsivity at higher photocurrent which we attribute to the increased device temperature.

Figure 5

Relative RF output power measurement at 20 GHz and 8 V reverse bias voltage on the left y-axis and RF power compression on the right y-axis as a function of measured average photocurrent for the CoC. The different curves show the behavior of the photodiode as a function of the distance between the optical fiber tip and the backside light entry of the photodiode.

5.1

Bandwidth measurements

The bandwidth measurements of the packaged photodiode have been performed as schematically represented in Figure 6. The optical modulation is provided by the interference of two CW lasers. The wavelength of the first lasers (PPCL550, Pure Photonics) is fixed at 1559 nm while the wavelength of the second laser (CTL 1550, Toptica Photonics) slowly scans between 1559.0 nm and 1559.3 nm. In this manner, the interference of the two lasers produces a modulated optical signal (beatnote) at frequencies between 0 and approximately 37 GHz.

Figure 6

Experimental setup for high-power S12 measurement.

The lasers are independently amplified by two home-built EDFAs with adjustable pump power and thereafter combined in a polarization maintaining 50/50 fiber combiner. One of the combiner outputs is monitored with an optical power meter (S145C and PM110D, Thorlabs) whereas the other output is connected to a second PM fiber combiner with a 90/10 splitting ratio used to sample 10% of the power onto an optical spectrum analyzer (OSA, AQ6317B, Ando Electric).

The pump power of the EDFAs is first roughly adjusted to obtain the desired photocurrent in the Device-Under-Test (DUT) and then fine-tuned to ensure that the two signals have the same amplitude, which results in a 100% modulation depth of the optical signal. This power-balance adjustment is performed by monitoring the signals in the OSA.

The electrical spectrum analyzer (N9030A PXA, Keysight) is set to max-hold acquisition mode recording the peak value of the beatnote signal as it sweeps across the desired frequency range. The losses of the cable and the DC-block connected to the RF spectrometer have been measured and compensated for.

Using this setup, the S12 frequency response of the broadband (PQW20B) and narrowband (PQW20N) packaged photodiode have been measured. The results for the broadband module show a high 3 dB bandwidth of around 24 GHz over a large range of average photocurrent, see Figure 7. The 3 dB bandwidth of the PQW20B is slightly higher than for the CoC. The difference can be explained by the internal 50 Ω termination resistor used in the package, which increases the RC time limiting bandwidth of the photodiode. Increasing the bias voltage improves the bandwidth at higher photocurrent as optical saturation effects are suppressed by the higher electric field. At a bias voltage of 10 V the bandwidth remains above 22 GHz up to 50 mA. The bandwidth decreases at lower photocurrent than we would expect from the CoC measurement. Despite the fact, that the photodiode within the package is illuminated with a distance of approximately 100-150um between the SMF tip and the backside of the photodiode. For the CoC measurement, the bias is applied through a bias-T which includes a DC block isolating the DC signal from the 50 Ω load in the PNA. In case of the PQW20B, the internal bias-T includes a 10 Ohm series resistor for ESD protection on the biasing pin. In conjunction with the matching, this leads to an internal serial resistance which reduces the bias voltage on the photodiode at higher photo current. As a result the packaged photodiode saturates at significantly lower photocurrent than the CoC.

Figure 7

Left: -3 dB bandwidth for different photocurrent and reverse bias voltage of 10 V for a broadband packaged photodiode. Right: Frequency response of the RF output for different optical powers at reverse bias voltage of 10 V for a narrow band packaged photodiode

The frequency response measurements of the narrowband packaged photodiode (PQW20N) show the RF output power peaking at 19 GHz which is the result of the specific matching at this frequency, see Figure 7.

5.2

Optical saturation current measurements

For the saturation current measurements of the packaged photodiode the same setup shown in Figure 6 has been used. The RF output power as a function of the photocurrent at 19 GHz (see Figure 8) shows that the narrowband module (PQW20N) offers higher RF output power of up to +15 dBm compared to +12 dBm for the broadband module.

Figure 8

RF output power at 19 GHz as a function of photocurrent and bias voltage for the broadband (PQW20B) and narrowband module (PQW20N).

6.1

Gamma irradiation testing

The gamma irradiation testing has been performed in the ESA-ESTEC Co60 facilities, Nordwijk, Netherlands. The test procedure is shown in Figure 9. The sample consists of four hermetically packaged photodiodes (PQW20B). One additional sample (S/N Q1804006) was stored without being exposed to irradiation and acts as reference sample. Two samples have been biased at 5 V during irradiation testing while two samples remained unbiased. Before irradiation testing all packaged photodiodes under test have been characterized with respect to responsivity, dark current, electrical return loss (S11) and 3 dB bandwidth (S₁₂). The total dose of 106 krad which corresponds to approximately 15 years exposure to geostationary orbit (GEO) has been applied in four runs whereas dark current has been measured after each run at ESTEC. After the last irradiation step, the component has been relaxed for 24 h before another dark current measurement has been performed at ESTEC. After the samples have been returned to Albis dark current, responsivity and S-parameter have been measured again and thereafter curing at 100°C for one week.

Figure 9

Gamma Irradiation Test Procedure

The dark current values before and after each irradiation step as well as after curing are shown in Figure 10. Additionally, Figure 11 depicts the -3 dB bandwidth measurements and responsivity values before, after irradiation and after curing. The dark current, responsivity as well as the S-parameter measurements do not show any degradation compared to the initial measurement.

Figure 10

Dark current at room temperature and 5V bias voltage as a function of Gamma irradiation dose in krad, 24h and 7d curing. USL stands for upper specification limit of the photodiode.

Figure 11

Responsivity at wavelength of 1550 nm and bias voltage of 5 V before and after irradiation as well as after one week curing at 100°C. LSL stands for lower specification limit for the packaged photodiodes.

The results of the gamma irradiation testing are summarized in Table 1. None of the tested devices shows significant degradation.

Table 1

Gamma irradiation result summary

6.2

Proton irradiation testing

The proton irradiation testing was conducted at the Proton Irradiation Facility (PIF) of the Paul-Scherrer Institute (PSI), Villingen, Switzerland. The test plan is depicted in Figure 12. Before irradiation testing, all packaged photodiodes have been tested with regards to dark current, responsivity and electrical return loss (S₁₁) and 3 dB bandwidth (S₁₂).

Figure 12

Proton irradiation test procedure.

Intermediate measurements of S-parameters and thereby assess a potential change in capacitance of the device have not been possible. Therefore, it was decided to expose the sample to different levels of fluence and verify the S-parameters in post-irradiation measurements. Two modules have been exposed to moderate irradiation while two packaged photodiodes have been exposed to radiation largely exceeding any reasonable space mission in LEO. An overview of the applied dose is given in Table 2. The equivalent LEO mission time is estimated based on an averaged proton flux of approximately 1e2 protons/(cm²s) for ISS low-earth-orbit (LEO), condition of solar maximum and no shielding⁸.

Table 2

Proton Irradiation Test Results.

Due to the limited proton beam width (5 cm), only two modules have been simultaneously exposed (see Figure 13).

Figure 13

Two packaged photodiodes mounted in the proton beam line at PSI.

During proton irradiation the dark current increases as a function of the accumulated dose (see Figure 14). The dark current does not reach the upper specification limit of 100 nA. Considering the main applications for the high-power packaged photodiode are operation at high optical input power thus an increase of the dark current to around 100 nA is not expected to affect the overall system performance.

Figure 14

Room temperature dark current at 5 V reverse bias voltage, before and directly after proton irradiation, as well as after 40 weeks storage and followed by one week curing at 100°C., D1: S/N Q1901001, D2: S/N Q1901003, D3: S/N Q1901008, D4: S/N 1901010, D5: reference sample without irradiation.

The 3 dB bandwidth and responsivity (see Figure 15) do not show significant alteration. Same applies for the electrical return loss (S11).

Figure 15

-3 dB bandwidth measurement and responsivity at wavelength of 1550 nm and 5 V bias voltage before proton irradiation testing, after irradiation and 40 weeks relaxation as well as one week curing at 100°C.

The results of the proton irradiation testing are summarized in Table 2. Due to a fiber breakage as result of handling error, responsivity and 3 dB bandwidth measurement of module Q1901010 could not be performed.