We fabricated photoanodes by depositing QDs into mesoporous TiO2 films using electrophoretic deposition (EPD)57, forming QDs/TiO2 heterostructures. After the ligand exchange and ZnS protection layer via SILAR, several types of QDs based photoanodes were assembled for PEC measurements. For all the anodes, we kept identical parameters for anode preparation, post-treatment and PEC configuration. In general, the photocurrent density (J) gradually increases with the increase of applied voltage until a saturated photocurrent density (Jph) is obtained. As shown in Fig. 5a, a Jph of 8.8 mA/cm2 can be obtained for the photoanode based on CdSe/6ZnSe. Remarkably, the CdSe/CdxZn1−xSe#1 QDs based photoanode reaches Jph values as high as 30.0 ± 1.0 mA/cm2. When the alloyed layers are further optimized by adding gradient layers, the CdSe/CdxZn1−xSe#2 QDs based photoanode exhibits unprecedented champion Jph values of 35 mA/cm2 with average value of 33 ± 2.0 under one sun illumination (AM 1.5 G, 100 mW/cm2), which is the highest photocurrent density among all QDs based PEC cells to date. To confirm the data obtained, we prepared more than 10 samples for PEC measurements. Figure 5b shows the Jph distribution for the three types of QDs and the inset figure shows the Jph of each measurement for CdSe/CdxZn1−xSe#2 QDs.
To confirm the alloy effect on other types of core/shell QDs, we also synthesized CdSe/ZnS, CdSe/CdS, CdSe/CdS/ZnS core/shell QDs and their corresponding alloyed QDs, CdSe/CdZnSe/CdZnS/ZnS, CdSe/CdSeS/CdS, CdSe/CdSeS/CdS/CdZnS/ZnS and spherical CdSe/CdxZn1−xSe QDs. All types of QDs were used 6 monolayers for the shells with different composition via SILAR process. We then used them as sensitizers for TiO2 for PEC H2 generation. The results are shown in Fig. 5c-d and Figure S11. The Jph is observed to increase with alloyed shells compared to pure-shell QDs. However, the highest Jph based on other types of alloyed QDs is only 22 mA/cm2 (CdSe/CdSeS/CdS QDs), much lower than 35 mA/cm2 obtained from CdSe/CdxZn1−xSe#2 QDs.
To investigate the morphology effect, spherical CdSe/CdZnSe/ZnSe QDs with the same structure/composition were compared with rod/egg shaped CdSe/CdZnSe/ZnSe QDs. The Jph based on spherical CdSe/CdZnSe/ZnSe QDs can reach up to 26.5 mA/cm2 which is still lower than that of rod/egg shaped CdSe/CdZnSe/ZnSe (CdSe/CdxZn1−xSe#1 and CdSe/CdxZn1−xSe#2 QDs) QDs. Such unprecedented results of CdSe/CdxZn1−xSe#2 QDs compared to CdSe/ZnSe QDs and other types of alloyed QDs can be attributed to: (i) the enhanced absorption range by adding alloyed gradient shells, leads to more absorbed photons from solar irradiation; (ii) as ZnSe is a wide bandgap semiconductor, it allows a more efficient passivation effect for the QDs surface compared to CdS and the overall band gap of the shell could be reduced by adding Cd in ZnSe, lowering the CB energy offset from the core to the shell; (iii) compared to the spherical shape, asymmetric rod/egg shaped QDs with intermediate alloyed shells improve the electron transfer efficiency (more than 7 times faster than other types of alloyed QDs), and simultaneously separate the carriers in different axis direction, leading to fewer carrier spatial overlap and exciton recombination.
Thus, the asymmetric QDs exhibited an outstanding Jph compared to other types of CdS/ZnS shelled QDs and spherical QDs with similar structure and composition. The obtained Jph in this work is comparable to the state-of-art of all semiconductors as PEC photoelectrodes as shown in Table 2. The highest Jph (35 ± 2.0) obtained using asymmetric QDs is close to the one obtained with photoanodes based on Si, and higher compared to values from photoanodes based on perovskites, and other bulk semiconductors such as metal oxides, nitrides, sulfides, etc.
Table 2
Best current density values for PEC H2 evolution.
Photoelectrode | Reaction electrolyte | Jph (mA/cm2) | Reference electrode | Ref. |
CdSe/CdxZn1−xSe QDs/TiO2 | 0.35 Na2S3/0.25 M Na2S | 35 (0.6 V) | Ag/AgCl | This work |
n-Si/SiOx/Al2O3/Pt/Ni | 1 M KOH | 28.5 (1.6 V) | Hg/HgO | 58 |
n-Si/SiOx/Co/CoOOH | 1 M KOH | 35 (0.6 V) | Ag/AgCl | 59 |
SrTiO3/p-Si | 0.5 M H2SO3 | 35 (0.6 V) | Ag/AgCl | 60 |
b-Si/TiO2/Co(OH)2 | 1 M NaOH | 32.3 (1.48 V) | calomel | 61 |
CoOx/p+n-Si | 1 M NaOH | 30.8 (1.23 V) | Hg/HgO | 62 |
MoSe2/ p+n-Si | 1 M HBr | 30 (0.3 V) | Ag/AgCl | 63 |
n-Si/PEDOT: PSS | hydrogen iodide | 28.8 (0.3 V) | Ag/AgCl | 64 |
n+p-Si microwire/SiO2 | 0.1 M H2SO4 | 34 (0 V) | Ag/AgCl | 65 |
p+n-Si/SiO2 | 1 M KOH | 31.2 (1.23 V) | / | 66 |
np+-Si/SiOx/NiFe | 1.0 M KOH | 30.7 (1.23 V) | / | 67 |
MAPbI3 with proline | 0.5 M H2SO4 | 21.7 (0 V) | calomel | 68 |
CH3NH3PbI3 | 0.5 M H2SO4 | 18 (0 V) | calomel | 69 |
BiVO4/N:NiFeOx | 0.5 M K3BO3 | 6.4 (1.23 V) | Ag/AgCl | 70 |
Cu3BiS3 | 0.2 M Na2HPO4/NaH2PO4 | 7 (0 V) | Ag/AgCl | 71 |
In:GaN/Ta3N5/Mg:GaN | 1 M KOH | 9.3 (1.23 V) | Hg/HgO | 72 |
CdIn2S4 | 0.5 M Na2SO4 | 5.73 (1.23 V) | Ag/AgCl | 73 |
To compare the real PEC H2 evolution process with the theoretical values derived from Jph based on alloyed QDs, we measured the H2 evolution rate using gas chromatography (GC) under one sun irradiation. According to Fig. 6a and the equation for the calculation of the faradaic efficiency (\({\eta }_{FE}\)) (shown in the Experimental section), a final \({\eta }_{FE}\) of ~ 82 % canbe calculated for our system with operation time of one hour. The incident photon-to-electron conversion efficiency (IPCE) was further measured and calculated based on the IPCE equation (shown in supporting information). As shown in Fig. 6b, for CdSe/CdxZn1−xSe#1 and CdSe/CdxZn1−xSe#2 QDs, from 650–700 nm, the IPCE decreases sharply and the values are close to 0, in agreement with the UV-vis absorption spectra. For CdSe/6ZnSe QDs, the initial drop position is blue-shifted to 600 nm which is attributed to the narrower absorption range. From the wavelength of 450 to 650 nm, IPCE value of two alloyed QDs are higher than CdSe/6ZnSe QDs while the IPCE for CdSe/CdxZn1−xSe#2 QDs is higher than that of CdSe/CdxZn1−xSe#1 QDs which illustrates a more efficient carrier transfer efficiency due to more favorable band alignment.
The stability performance is another critical factor for the application of PEC H2 generation. Stability tests were conducted under 0.6 V bias vs RHE with continuous one sun intensity (AM 1.5 G, 100 mW/cm2) solar irradiation (shown in Fig. 6c). The CdSe/ZnSe QDs based photoanode only shows a photocurrent retention of 58% after 2-h continuous illumination. This is mainly due to the photo-oxidation of the QDs where electrons and holes cannot be transferred out of the QDs due to unfavourable band alignment. For two alloyed QDs, the CdSe/CdxZn1−xSe#1 QDs based photoanode can maintain ~ 78% of the initial current density value after 2-h solar irradiation. Such improvement is attributed to the favourable band alignment between core and shell obtained by adding alloyed shells. As a result, self-oxidation and recombination can be largely reduced inside the QDs, leading to better stability performance. For CdSe/CdxZn1−xSe#2 QDs, ~ 96% of the initial photocurrent can be preserved after 2-hour continuous illumination, and even after 10 hrs ~ 82% of the initial Jph is maintained. Such gradient band alignment can largely improve the carrier transfer rate, leading to fewer hole accumulation and accordingly reduced photo-oxidation and photo-recombination. We also compared the stability performance of CdSe/CdxZn1−xSe#2 QDs to other types of alloyed QDs under 2-h illumination which is shown in the inset of Fig. 6c. Due to the fastest electron transfer rate and efficient surface passivation effect, the CdSe/CdZnSe/ZnSe (CdSe/CdxZn1−xSe#2) QDs still exhibit the best stability performance among all alloyed QDs.