Single photon avalanche diodes (SPADs) are avalanche photodiodes (APD) operating in Geiger mode with bias higher than the breakdown voltage (V_br). Due to the internal gain from avalanche effect, a single incident photon can trigger off a measurable current, which represents the ultimate sensitivity for a photodetector.^[1,2] Silicon SPAD is one of the most qualified semiconductor photodetector by now for sensing a weak signal with wavelength ranging from 300 nm to 1100 nm due to its intact crystalline quality and large ratio of electron and hole ionization coefficients. Due to the advantages of ruggedness, compactness, suitability, and easiness to build integrated systems, Si SPADs are more competitive for the deep space applications.^[3]

However, in the space environment, radiation may cause permanent performance degradation due to the energy deposition on the sensitive region of the device. The radiation dose rate is associated with obit and orbital inclination. It varies from 100 rad(Si)/year to 10 krad(Si)/year in low Earth orbit (LEO), 100 krad(Si)/year in medium Earth orbit (MEO), and over 10 krad(Si)/year in high Earth orbit (HEO).^[4] Gamma-ray (γ-ray), which consists of the shortest wavelength electromagnetic waves and has a large penetration depth, is a kind of most energetic photons indicating highly destructive power.^[5,6] In the case of γ-ray radiation, taking ⁶⁰Co γ-ray source as an example, the radiation effect is mainly through Compton scattering.^[3] It mainly causes electron ionization and atomic displacement.^[7,8] The former can generate positive charges and oxide trapped charges, as well as interface traps at the SiO₂/Si interface, which act as generation centers for a surface current if exposed to an electric field.^[9] The displacement damage is mainly through Compton scattering by the photon with energy far greater than the binding energy of the atomic electrons.^[10] The energetic Compton electron interacts with the Si lattice, and then Si atoms in the bulk material will be displaced from their host lattice positions and defects will be generated if the energy from the scattered electrons is high enough. Defects in the device play an important role for the increase of the leakage current and DCR, and then the reduction of the sensitivity. Hence, it is mandatory to assess the performance of a SPAD after exposed in the γ-ray radiation environment.

This paper reports the results of γ-ray radiation hardness test for Si SPADs with a total dose of 100 krad(Si) and a dose rate of 50 rad(Si)/s by using ⁶⁰Co as the radiation source. The purpose of this experiment is to assess the laser time transfer payloads equipped on the BeiDou navigation satellite system. It is found that the V_br, photocurrent, and gain have no measurable changes after radiation. However, both leakage current and DCR of the device after radiation increase by about one order of magnitude. The carrier transport dynamics after radiation, which is verified by the annealing experiment, is analyzed by the temperature-dependent current–voltage (I–V) measurement.

An epitaxial-based planar n⁺–p–p⁻–p⁺ SPAD structure with a diameter of 200μm in this work is fabricated by complementary metal oxide semiconductor (CMOS) technology. Si epitaxial is grown on the heavily-doped p⁺-Si substrate with 〈 111〉 crystallographic orientation. The p-type multiplication layer is then formed by implantation method, followed by annealing process. The guard ring, p⁺ stopper ring, and n⁺ ohmic contact layer are also formed by ion implantation. The active region is surrounded by the p⁺ stopper ring at the edge of the chip which stops space charge at the SiO₂/Si interface. An optimized SiO₂ antireflecting coating is applied to increase the light incident efficiency. Finally, the chips are all sealed in a standard TO-46 case with a glass window on the frontside.

The radiation test of the Si SPADs has been conducted at room temperature with γ-rays, which were produced by a ⁶⁰Co source. The SPAD chips under test were sealed in a standard TO-46 case with a glass window on the frontside. Then they were exposed to the ⁶⁰Co γ-ray source directly with a distance about 3 cm to ensure the dose rate of 50 rad(Si)/s. The total dose was 100 krad(Si). The SPADs were divided into three groups with different bias conditions (unbiased, 0.9 V_br, 1.1 V_br) during the radiation test. The leakage current was measured both before and after radiation (within three hours) using Keithley 6517B Elctrometer. A tungsten-halogen lamp (Zolix LSH-T75) and a grating monochromator (Zolix Omni-λ 3005) were employed in the measurement of the photocurrent under the wavelength of 532 nm. The SPAD was operated in passive-quenching Geiger mode, while the bias voltage on the SPAD was higher than V_br. The avalanche pulse would be triggered and discriminated, while photons or dark noise occurred. Then the avalanche would be quenched by the large resistor. The DCR was recorded by an oscilloscope (Iwatsu SS7840H).

The photocurrent and gain of the devices show negligible changes after radiation. Figure 1(a) shows the photocurrent and corresponding gain results under the incident wavelength of 532 nm, the sample biased at 1.1 V_br under radiation is taken for example. The slight rising of leakage current could not make obvious change of the photocurrent and gain. From the electrical measurement, the negligible change of V_br, which is about 33 V, has been observed for all the samples in the inset of Fig. 1(b), indicating that the doping level of the whole structure has not been remarkably changed by the defects caused by the γ-ray radiation. However, the leakage current is really affected by the γ-ray radiation, as shown in Fig. 1(b). Compared with the value before radiation, the leakage current still remains at a low level of about 10⁻¹⁰ A. The increase of the leakage current is around one order of magnitude for all the devices under test. As shown in Table 1, the leakage current at 29.7 V is 351 pA, 667 pA, and 151 pA, respectively, for the devices under the biases of 0.9 V_br (29.7 V), 1.1 V_br (36.3 V), and unbiased during the radiation test. It can be seen that the degradation degree of leakage current becomes severe with increasing bias during the radiation test. Table 2 shows the radiation effect on the DCR performance, which is measured by keeping PDE at about 5 % under the wavelength of 532 nm. DCR of the devices before radiation is about 300 cps/μ m², where cps is short for counts per second. Correspondingly, DCR also increases by about one order of magnitude after the radiation, and also strongly depends on the bias condition during the γ-ray exposure.

Table 1.

Table 1.

Table 1. Leakage current changes at 29.7 V before and after γ-ray radiation test. . Condition Leakage current before radiation 28 pA unbiased 151 pA 0.9 Vbr 351 pA 1.1 Vbr 667 pA

Table 1. Leakage current changes at 29.7 V before and after γ-ray radiation test. .

Table 2.

Table 2.

Table 2. Comparison of DCR before and after radiation while keeping PDE at 5 %. . Condition DCR before radiation 300 cps/μm2 0.9 Vbr 5.1× 103 cps/μm2 1.1 Vbr 11.5× 103 cps/μm2

Table 2. Comparison of DCR before and after radiation while keeping PDE at 5 %. .

Fig. 1.

Figure Option ViewDownloadNew Window

Fig. 1. (a) The measured photocurrent and corresponding gain at 532 nm of the devices biased at 1.1 Vbr, which show negligible variations compared with the values before radiation. (b) Leakage current as a function of reverse voltage before and after radiation at room temperature. During the radiation, the samples are biased at 0.9 Vbr, 1.1 Vbr and unbiased, respectively, with the total radiation dose of 100 krad(Si) and dose rate of 50 rad(Si)/s. The inset shows the negligible changes of Vbr after radiation.

To further gain the effect of radiation on carrier transport dynamics, the temperature-dependent forward and reverse I–V characteristics are investigated from 60 K to 300 K before and after radiation, respectively. The Arrhenius plots of the forward and reverse I–V curves after radiation are shown in Figs. 2(a) and 2(b), respectively. V_br decreases as the temperature decreases from 300 K to 60 K, with a positive temperature coefficient (0.02 V/K), indicating a true avalanche breakdown occurred.^[11] According to the forward current equation

Fig. 2.

	Figure Option ViewDownloadNew Window
	Fig. 2. Temperature-dependent (a) forward and (b) reverse I–V curves. (c) Comparison of measured temperature-dependent ideality factor n and their fitting results before and after radiation. (d) Arrhenius plot of leakage current measured at 20 V and the fitting results as a function of temperature before and after radiation.

The ideality factor fits the model of tunneling enhanced interface recombination very well at lower temperature (< 210 K), which is given by the following equation:

Figure 2(d) depicts the Arrhenius plot of leakage current at 20 V as a function of temperature before and after radiation, respectively. According to the reverse current (I_R) equation

In order to verify the radiation effect on the Si SPAD chips, annealing experiment was carried out. Figure 3 monitors the leakage current of a Si SPAD after radiation at room temperature without bias. As shown by the red squares, it increases at first 300 hours from 50 pA to 151 pA firstly, which is owing to the increased density of interface traps, whose long build-up time is determined by the transport of a positively charged ion, probably H⁺, through the oxide.^[20] The decrease of the leakage current in the following 268 hours may be due to the recovery of vacancy-interstitial pairs nearby. However, the leakage current is still as high as 40 pA due to the existing oxide trapped charges and interface traps, and also traps inside the avalanche junction. After annealing at 200 °C for 2 hours, an obvious reduction of the leakage current is observed. The leakage current (14 pA) approaches its initial value before radiation (black square). Correspondingly, DCR also decreases to 1.3× 10³ cps/μm² at 33 V after annealing at 200 °C. During this process, hole de-trapping may occur by thermal emission of holes from the traps in the oxide or hop to the interface and into silicon.^[21] Besides, the originally displaced atoms caused by Compton scattering could gain enough energy to move back to their host lattice positions, which results in the reduction of the traps or defects. The reduction of the trap number both in the bulk and interface is determined by the temperature and time. The recovery of the noise level of the weakened device by annealing and the anti-radiation performance reach the space availability.

Fig. 3.

	Figure Option ViewDownloadNew Window
	Fig. 3. Time evolution of leakage current after radiation. Leakage current of the unbiased device is traced at 29.7 V at room temperature for 568 hours (red square) and annealed at 200 °C for 2 hours (blue square), after which the leakage current approaches its initial value before radiation (black square).

Total dose test results with γ-ray for silicon SPADs are presented. The breakdown voltage, photocurrent, and gain demonstrate their stability even after exposure to γ-rays, which guarantees a stable operation of Si SPADs. While the leakage current and DCR increase by about one order of magnitude compared with the values before radiation. Temperature-dependent I–V measurement results reveal the trap-assisted generation carrier transport dynamics inside the device. The leakage current and DCR can be almost recovered after annealing at 200 °C for about 2 hours, which further verifies the radiation damage effect on the devices. The recovery of the original noise level after annealing indicates a reasonable method to fulfill its space application.