† These authors contributed equally to this work.
‡ Corresponding author. E-mail:
Project supported by the National Natural Science Foundation of China (Grant Nos. 11574362, 61210014, 11374340, and 11474205) and the Innovative Clean-Energy Research and Application Program of Beijing Municipal Science and Technology Commission, China (Grant No. Z151100003515001).
According to the well-established light-to-electricity conversion theory, resonant excited carriers in the quantum dots will relax to the ground states and cannot escape from the quantum dots to form photocurrent, which have been observed in quantum dots without a p–n junction at an external bias. Here, we experimentally observed more than 88% of the resonantly excited photo carriers escaping from InAs quantum dots embedded in a short-circuited p–n junction to form photocurrent. The phenomenon cannot be explained by thermionic emission, tunneling process, and intermediate-band theories. A new mechanism is suggested that the photo carriers escape directly from the quantum dots to form photocurrent rather than relax to the ground state of quantum dots induced by a p–n junction. The finding is important for understanding the low-dimensional semiconductor physics and applications in solar cells and photodiode detectors.
Practically, most photovoltaic devices including solar cells and photovoltaic detectors incorporate a p–n junction in a semiconductor utilizing photovoltaic effect to convert light into electricity.[1–3] Large optical absorption coefficients and great carrier extraction capability are conductive to solar cells[4] and photovoltaic devices.[5] In order to fully absorb the light, solar cells and photodiode detectors mainly make use of bulk materials.[6] Furthermore, to obtain high crystalline quality, low-dimensional semiconductors have been applied in solar cells[7,8] and photovoltaic detectors.[9]
Recently, quantum dots (QDs) have garnered extensive interest for applications in optoelectronic devices because of their strong zero-dimensional confinement effects with a δ-function density of states.[10–12] These characteristics make the QDs to form their distinct quasi-Fermi level and to restrict photo-generating carriers to the ground energy level.[13,14] Consequently, without additional excitation, carriers in the quantum restriction level cannot escape to external circuit,[15,16] so QDs as an intermediate-band have been suggested to utilizing low energy photon in the solar spectrum.[17,18] The principle is that one photon pumps an electron from the valence band (VB) to conduction band of a QD, namely intermediate band (IB), while the second one pumps an electron from the IB to the conduction band (CB) of the barrier.[16,19] which can produce photocurrent from photon absorption by QDs.[20] However, all investigations have failed to preserve the open-circuit voltage of an IBSC compared to that of a standard solar cell,[16,19,21] which is explained by tunneling and thermal emission processes.[22,23]
It is reported that the mechanism of IBSC has been confirmed when the photocurrent increases under two sub-band-gap energy lasers shone on IBSC.[24,25] However, the experimental results cannot rule out the possibility that the photocurrent separately comes from the VB to IB transition or IB to CB transition,[26,27] respectively. Namely, it cannot confirm that the increase of photocurrent is caused by two successive jumps. Thus, the carrier transport process is not clear so far. In order to solve this problem, we use only one laser to pump an electron from the VB of a QD to its CB, then find that a p–n junction induces the most of photo-excited carriers to directly escape from the QDs to generate the photocurrent rather than relax to the ground state under the short-circuit condition. The results cannot be well explained by the current light-to-electricity conversion theory.[20]
Two sample structures with n–n and p–n structures (Fig.
The photoluminescence (PL) spectra of Sample A were measured at 150 K under the open-circuit condition and with a 0.7-V bias to provide an external electrical field (Fig.
The resonant PL spectra of Sample B were measured at 150 K under open- and short-circuit conditions (Fig.
To deeply investigate the phenomenon of PL intensity quenching of Sample B under open and short-circuit condition, we varied the incident laser powers from 5 mW to 60 mW to probe the PL spectra (Fig.
To clarify the escape pathway of photon-generated carriers, we connected a variable resistor to the circuit to adjust the circuit current (Fig.
The resonant excited photocarriers directly escape from the QDs to generate the photovoltaic effect in Sample B (with a p–n junction) under the short-circuit condition. However, in Sample A (without a p–n junction), the resonant excited photocarriers relax to the ground state and then recombine to emit light under a 0.7-V bias. Because Samples A and B have the same barrier height and thickness, thermionic emission and tunneling processes cannot be used to explain the photo-excited carriers escaping from the quantum dot.[21,22,34,35] It can be deduced that the p–n junction induces the photocarriers escaping from the QDs.
According to the established photon absorption model of QD heterojunction, the basic absorption processes for different values of energy hν are shown in Fig.
Firstly, we compared the time of carrier drifting out of the QD and the relaxation time to the ground state to check the possibility that photocarriers can directly escape from the QDs instead of relaxing to the energy level. It has been reported that the time of an electron relaxing to the ground state is approximately several hundred picoseconds.[31] For the short-circuit condition in Sample B (with a p–n junction), the competition between the free excited-state carrier transits over the InAs quantum and relaxes to the ground state, which is critical to account for the experimental phenomenon. The time of the free excited-state carrier transiting over the QD in the depletion region is given by
Secondly, we explains why photo-carriers can be extracted from the QDs by a p–n junction rather than by an external voltage bias. For Sample A, when the QD structure is applied with an external bias, the electric-field distribution is nearly uniform at the undoped region. Carrier drift velocity is proportional to the electric field and carrier mobility. Electron mobility is faster than the hole mobility, which leads to the electron drift velocity larger than the hole drift velocity. Furthermore, the electrons drifted from QD region are more than the holes drifted from QD region. Thus, it will lead to excess holes accumulated in the device and forming a large positive electric field, in turn preventing carriers from escaping. However, for Sample B, in the depletion region, the electric field is a linear function of the distance through the junction and the tendency reverses at the position of abrupt junction. The slope of electric field can be adjusted by changing the dopant concentration of p- or n-type region, respectively. The electric-field distribution in the p–n junction can balance the hole velocity to match the electron velocity, which diminishes the blocking effect of current. The electric-field distribution in the p–n junction facilitates the hole to be extracted and in turn reaches a better electrically neutral condition. The unique characteristic of p–n junction results in continuous carriers escaping from the QDs.
When the incident laser-beam power is 65 mW, we measure that the Jsc of Sample B is 2.82 mA at 150 K. The incident-photon-to-current conversion efficiency is calculated as 5.9% under the short-circuit condition. The effective absorption layer is ten layers of InAs QDs, which is approximately 60 nm in total, and the absorption coefficient of the InAs QDs is calculated to be approximately 104 cm− 1, which far exceeds the theoretical calculation of 100 cm− 1 for the sample without a p–n junction.[38,39] The absorption coefficient we measured is underestimated because the surface reflection is not considered and the extraction efficiency of the p–n junction is assumed as 100%. Such value is comparable to the absorption coefficient of bulk GaAs near its band gap.[40] Considering the lower density of state and the localized wavefunction of zero-dimensional semiconductor material, it is reasonable to deduce the p–n junction dramatically to enhance the absorption of InAs QDs, which demonstrates that the QDs can be easily used in solar cells and photodetectors.
The results in this study indicate that the final state of the carriers is a free excited state in the QDs within a p–n junction under the short-circuit condition. However, the absorption final state would be a localized ground state, like the QDs under the short-circuit condition in our experiments. Based on the perturbation theory of quantum mechanics, the absorption coefficient is related to the density of states and the wavefunction of the final state of the carriers.[41] The changes of the physical nature increase the absorption coefficient when the QDs within a p–n junction operate under the short-circuit condition. The results deviate from the established theories of solar cells and photodiode detectors, where the absorption coefficient is assumed to be a constant for different conditions.
In summary, we have observed that the most photo-excited carriers escaped from the InAs QDs with a p–n junction under the short-circuit condition, this finding challenges the current light-to-electricity conversion theory. A new mechanics applying InAs QDs incorporate a p–n junction in solar cells and photodetectors is deduced. It is found that the InAs QDs can not only make the design of solar cells more flexible, but also extend the spectrum response range of photodetectors. It is visualized that our mechanism of photo-excited carriers escaping from the QDs can be extended to other materials since the fundamental band-alignment and light absorption process are essentially the same.
1 | |
2 | |
3 | |
4 | |
5 | |
6 | |
7 | |
8 | |
9 | |
10 | |
11 | |
12 | |
13 | |
14 | |
15 | |
16 | |
17 | |
18 | |
19 | |
20 | |
21 | |
22 | |
23 | |
24 | |
25 | |
26 | |
27 | |
28 | |
29 | |
30 | |
31 | |
32 | |
33 | |
34 | |
35 | |
36 | |
37 | |
38 | |
39 | |
40 | |
41 |