The enhanced photo absorption and carrier transportation of InGaN/GaN Quantum Wells for photodiode detector applications

We have conducted a series of measurements of resonantly excited photoluminescence, photocurrent and photovoltage on InGaN/GaN quantum wells with and without a p-n junction under reverse bias condition. The results indicate that most of the resonantly excited photo-generated carriers are extracted from the quantum wells when a p-n junction exists, and the photon absorption of quantum wells is enhanced by the p-n junction. Additionally, the carrier extraction becomes more distinct under a reverse bias. Our finding brings better understanding of the physical characteristics of quantum wells with p-n junction, which also suggests that the quantum well is suitable for photodiode detectors applications when a p-n junction is used.

Scientific RepoRts | 7:43357 | DOI: 10.1038/srep43357 Two samples with 10-period InGaN/GaN multiple quantum wells grown on c-plane sapphire substrates by Metal Organic Vapor Phase Epitaxy (MOVPE) are prepared. The multiple quantum wells covered by p-GaN is named sample A (with p-n junction) and the one covered by n-GaN is named sample B (without p-n junction), as showed in Fig. 1. A laser with a wave-length of 405 nm, i.e. the photon energy of 3.06 eV is used to resonantly excite the carriers in multiple quantum wells-due to the photon energy is between the band gap of GaN barriers and InGaN wells 17,18 . Fig. 2 Fig. 2(a), sample A demonstrates good rectifying properties with a low circuit current under reverse bias. Hence, it indicates that the photocurrent is generated from the multiple quantum wells in the p-n junction because most of the photo-generated carriers are extracted from wells rather than relax to ground state for recombination when illumination is applied. Figure 2(b) shows the power-dependent photocurrent and ratio of integrated PL intensity of sample A under 3 V reverse bias. The linear dependent of photocurrent with the incident power further confirms that the photocurrent comes from the quantum wells. The ratio of integrated PL intensity under 3 V reverse bias condition to that under open circuit changes from 1.46% to 3.88% with the incident power, which means that only minority of the photo-carriers are confined in quantum wells for recombination under 3 V reverse bias condition. The above photo-electric characteristics of the quantum wells with a p-n junction show that even though the incident photon energy is lower than the barrier band gap, most of the photo-generated carriers escape from wells rather than relax to ground confined state for recombination when the p-n junction is under reverse bias. Furthermore, the photo-electric characteristics of sample A under varied reverse bias are studied in details. As Fig. 3 shows, the photocurrent increases from 1.42 mA to 1.48 mA when the reverse bias changes from 0 V to 7 V with an excitation power of 9 mW, the corresponding ratio of integrated PL intensity under reverse bias condition to that under open circuit decreases from 3.5% to 2.2%. It is worth noting that a photocurrent of 1.42 mA generates even under 0 V bias, which means most of the excited carriers escape from the quantum wells when the p-n junction is under short circuit. The increasing of photocurrent and decreasing of integrated PL intensity with reverse bias shows that an extra reverse bias enhances the carriers escaping effect further, since it enhances the electric field in the p-n junction.

PL measurement of sample A under open circuit and 3 V reverse bias condition are taken and shown in
The same comparison, with and without an external bias, is applied to sample B (a quantum well structure without p-n junction) to study the role of p-n junction in our experiment. For an excitation power of 9 mW, the integrated PL intensity under 3 V bias condition decrease by 1% compared with that under open circuit, and the peak positions are nearly identical (463.9 nm and 463.7 nm) as shown in Fig. 4(a). The InGaN quantum well layer in our samples suffered a piezoelectric polarization field pointed to the direction of sapphire substrate. When sample B is under bias condition, the external field is opposite with the polarization field of quantum wells. Hence, the external field should counteract the polarization field first and the PL peak showed a little blue-shift accordingly shown as Fig. 4(b). The ratio of integrated PL intensity under bias condition to that under open circuit changes slightly with the bias, which only decreases to 78% when the bias is up to 7 V. This means only 22% of the photocarriers lost (comparing with photoluminescence under open circuit), much less than that of the p-n

Discussion
The quantum efficiency defines as the number of electron hole pairs extracted per incident photon. It is equal to the current responsivity times the photon energy of the incident radiation 15 : where P inc is the incident optical power in watts, I ph is the photocurrent, h is the Plank constant and ν is the radiation frequency. Sample A showed a photocurrent of 1.46 mA under 3 V reverse bias with an incident power of 9 mW, thus the quantum efficiency is calculated to be 49.6% accordingly. Considered the total thickness about 25 nm of quantum wells, an absorption coefficient of 2.7 × 10 5 cm −1 is obtained, which is larger than the usually reported of 1 × 10 5 cm −1 for InGaN film 20,21 . Hence the photo absorption of quantum wells was enhanced by the p-n junction. Taking a more comparison of sample A with sample B, it can be concluded that the p-n junction makes most of the resonant excited photo-generated carriers escape from quantum wells. From our experiments, comparing with open circuit, nearly 95% of the photo-carriers lost (considering the carriers for photoluminescence) when the p-n junction is under short circuit. Some documents have attributed this lost to the two main mechanisms of the carriers' extraction from quantum wells, i.e. thermionic electron emission and tunneling mechanism 22,23 . However, the equivalent effect of these two mechanisms on carriers' extraction should be measurable in multiple quantum wells with and without p-n junctions. This disagrees with our measurements when an external bias used to simulate the inner field of p-n junction is applied on multiple quantum wells without p-n junction. In our experiment, applying a 7 V external bias to the quantum well structure without a p-n junction only cause 22% lost of the photo-carriers. According to this, the carriers that are extracted from multiple quantum wells both with and without p-n junction by means of thermionic electron emission and tunneling mechanism are not more than 22 percent. So, the extra carriers' lost can be attributed to the role of p-n junction, which means the carriers transportation was enhanced by p-n junction. This shows that applied an external bias to the quantum well structure without a p-n junction can't implement the same effect of p-n junction. The p-n junction must have played some special role in carriers' escaping from the quantum wells, thus the carrier absorption and transportation process should be discussed further. We will elaborate this phenomenon and mechanism in detail in another article. Here, the result of our experiments shows that the photo absorption and carriers transportation of InGaN/ GaN quantum wells ware enhanced by p-n junction. Additionally, a reverse bias makes the carriers extraction much more distinct. This provides a new knowledge of absorption efficiency of quantum wells with p-n junction and demonstrates an innovative method of detector design.
In summary, we find that the photo excited carriers can escape from the quantum wells when the quantum wells are sandwiched in a p-n junction and the photon absorption of quantum wells is enhanced. This brings a better understanding of the physical characteristics of quantum well with p-n junction, and also gives us a new insight into the design of photodiode detector based on quantum wells.

Methods
Epitaxial growth and chip fabrication. The samples used in this work are grown on c-plane sapphire substrates by a Veeco GaN MOVPE system. The trimethylgallium (TMGa), triethylgallium (TEGa), trimethylindium (TMIn) and ammonia (NH 3 ) are used as precursors. Silane (SiH 4 ) and biscyclopentadienyl magnesium (Cp 2 Mg) were used as n-type and p-type dopants, respectively. After the thermal treatment of sapphire wafers in H 2 at 1050 °C, a 25-nm-thick GaN nucleation layer was deposited at 530 °C, followed by a 3.5-um-thick n-GaN(3 × 10 18 cm −3 ) deposited at 1050 °C. Then 10 periods of InGaN/GaN (2.5 nm InGaN wells and 14 nm GaN barriers) MQWs were growth at 720 °C. The last layer for sample A (with p-n junction) is a 200-nm-thick p-GaN (5 × 10 17 cm −3 ), and for sample B (without p-n junction) is a 200-nm-thick n-GaN (3 × 10 18 cm −3 ). After growth, the epi-wafers were fabricated to 1 mm × 1 mm area size chips for later measurements, using Ti/Cr/Ni and Ni/Au as the n-type contact and p-type contact, respectively.
Photo-electric measurement. A 405 nm semiconductor laser diode was selected as the resonant excitation source, the output power could be adjusted from 25 mW to 1 mW using a neutral optical attenuator. The PL spectra under open circuit condition and bias condition, open circuit voltage, photocurrent under bias condition are measured simultaneously. When sample A is under reverse bias condition, the p-type contact is connecting with the negative electrode of a constant voltage source and the n-type contact is connecting with the corresponding positive electrode. For sample B, the n-type contact above the quantum wells is connecting with the negative electrode and the n-type contact below the quantum wells is connecting with the positive electrode.