Reduction of charge recombination in PbS colloidal quantum dot solar cells at the quantum dot/ZnO interface by inserting a MgZnO buffer layer

Abstract

Interfacial charge recombination occurring at the interface between a colloidal quantum dot (CQD) solid and the electron collecting layer (ECL) in CQD solar cells significantly affects the charge carrier collection, therefore limiting the device photovoltaic performance. In this work CQD solar cells with an improved performance are reported by employing MgZnO as a buffer layer (BL) with tunable electronic energy levels in the solar cells to reduce interfacial charge recombination and hence improve the solar cell photovoltaic performance. The effect of the BL on the solar cell performance is experimentally investigated and compared to theoretical calculations. Incorporation of a BL with favorable electronic energy levels forming a suitable band alignment with the CQD layer in solar cells diminishes the interfacial charge recombination and an increased photovoltage can be obtained. A CQD solar cell with a BL shows a power conversion efficiency of up to 9.3%, compared to that of 8.2% for a solar cell without any BLs. The unsealed solar cells are also rather stable under ambient conditions both in the dark and under continuous illumination. This work suggests that a MgZnO BL with energy level tunability provides a potential strategy to improve the interfacial properties of CQD photovoltaic devices.

---

1. Introduction

Colloidal quantum dots (CQDs) have attracted much attention for use in optoelectronic and photovoltaic devices because of their low cost and high stability and the possibility for solution based preparation methods.^1–5 The band-gap energy of lead chalcogenide (PbS or PbSe) CQDs can be tuned from 0.5 eV to 1.6 eV by controlling their particle size, which gives them great potential for the construction of semitransparent or tandem solar cells through tailoring the absorption across the solar spectrum.^6–9 CQD solar cells are also attractive due to the possibility for low temperature manufacturing, which together with solution processability makes them compatible with flexible substrates.^10–13 Impressive progress has been achieved by engineering device architectures and the surface chemistry of CQDs, and power conversion efficiencies (PCEs) exceeding 10% have been obtained.^14,15

Most of the highly efficient CQD solar cells are constructed with a p–i–n heterojunction device architecture, where a CQD solid is sandwiched between a wide bandgap n-type metal oxide (such as TiO₂ or ZnO) and a p-type electron blocking layer (such as MoO₃, CuI or p-type CQDs).^1,12,16,17 The photoinduced electron–hole pair is separated due to the built-in electric field within the solar cell and the charges drift/diffuse to the corresponding charge collecting electrode.^18,19 However, charge collection crucially competes with charge recombination during solar cell operation. Charge recombination in CQD solar cells simultaneously occurs within the CQD solid and at the interfaces between the different materials, significantly affecting the charge collection and therefore the photovoltaic performance.^19,20 To reduce the trap-assisted charge recombination within the CQD solid, different passivation strategies were developed to decrease the trap density in the CQD solid, which therefore increased the charge collection.^6,20–23

Interfacial charge recombination at the interface of a CQD/metal oxide electron collecting layer (ECL) may occur with faster kinetics than the charge recombination within the CQD, limiting the electron extraction and the quasi-Fermi level separation.^24,25 Introducing a buffer layer (BL) between the CQD and ECL layers can therefore be essential to reduce interfacial charge recombination losses in solar cells and to improve the photovoltaic performance. With the aim to reduce this recombination and to enhance charge extraction, several BL materials have been exploited, such as ZnO,²⁶ fullerenes,²⁵ molecular monolayers (such as 4-aminobenzoic acid)¹⁴ and polymers,²⁴ and it was concluded that the BL should be finely controlled to give the appropriate energy band alignment for the heterojunction without affecting the depletion region in the CQD solid to ensure that the incoming electrons can be efficiently injected into the ECL from the CQD solid.²⁵ However, simple processing of the BL for large areas and finely controlling the electronic energy levels of the BL with respect to that of the CQD layer and the ECL remain challenges.

In this work, with the purpose of reducing interfacial charge recombination and enhancing charge extraction from the CQDs, a MgZnO (MZO) layer with an appropriate energy level between the ZnO layer and the CQD layer, working as a BL, was introduced into the CQD solar cell, forming a suitable band alignment at the ZnO/MZO/CQD interfaces. The MZO thin films were prepared using a sol–gel method under an ambient atmosphere and the energy levels were tuned by controlling the molar ratio of magnesium in the film. The effect of the MZO-BL on the energy band alignment within the CQD solar cell and related photovoltaic performance were modeled, and experimental studies were performed for comparison and photovoltaic optimization. Incorporation of the MZO-BL into the CQD solar cell significantly reduced the interfacial charge recombination at the ECL/CQD interface resulting in enhanced charge extraction without affecting the depletion region in the CQD solid. By coupling with extensive theoretical modeling and experimental investigations, the photovoltaic performance was dramatically improved that a PCE as high as 9.3% was achieved for a solar cell with a MZO-BL, compared to that of 8.2% for a solar cell without any BLs. Importantly, the solar cell with a MZO-BL also shows promising stability, and almost no performance degradation was observed in 37 days when the solar cells were stored under ambient conditions.

2. Results and discussion

The device architecture of the CQD solar cell with a MZO-BL is shown in Fig. 1a. The bandgap of ZnO can be increased by cationic substitution of zinc with magnesium forming a MZO alloy, which can be applied in a variety of optoelectronic devices through tunable electronic energy levels.^27,28 A higher concentration of magnesium in MZO will result in a higher energy of the conduction band minimum (CBM),^28,29 which can be advantageous in photovoltaic applications to match the energy levels of the light absorbing material and the electron collecting electrode. ZnO and MZO layers were prepared using a sol–gel method under an ambient atmosphere according to the literature,²⁹ and the energy level shift as a function of molar ratio of Mg in the MZO film is shown in Fig. S1.† A compact ZnO layer with a thickness of ∼40 nm was firstly deposited on the top of a cleaned FTO glass and subsequently a MZO layer with a thickness of ∼20 nm is deposited on the ZnO layer (Fig. S2†). PbS CQDs with a bandgap of ∼1.3 eV were synthesized according to the literature.²¹ The light absorption and photoluminescence spectra of the CQD solution are shown in Fig. S3.† The CQD solid working as a light absorber in the solar cell was deposited on the top of the MZO layer using sequential layer-by-layer spin-coating under ambient conditions. The CQDs capped with oleic acid ligands were treated with tetrabutylammonium iodide (TBAI, CQD-TBAI) for ligand exchange during the deposition of the initial eight layers. Subsequently, two CQD layers working as electron blocking layers were treated with 1,2-ethanedithiol (EDT, CQD-EDT) for ligand exchange. Finally, the device was completed by thermal deposition of an Au electrode on top of the CQD-EDT layer (Fig. 1a).

Fig. 1 (a) Device architecture of the CQD solar cell with a BL between CQDs and the ZnO electron transporting layer. (b) Simulated J–V curves of the solar cell without any BLs and with a BL. The CBM of the BL shifts either −0.1 eV or +0.1 eV referring to the ZnO layer. The simulated photovoltaic parameters (c) V_oc, (d) J_sc, (e) FF and (f) PCE of CQD solar cells as a function of CBM shift of the BL in comparison to the ZnO layer.

One of the critical functions of the BL is to minimize the energy barrier between the CQD layer and the ECL. The energy level of the CBM of the BL should be properly aligned with the CBM of the CQD layer and the CBM of the ZnO to minimize charge losses during electron collection.^25,30 Depending on the molar ratio of magnesium in MZO, the energy levels of the MZO BL can be shifted,^29,31 to achieve a good alignment and to reduce the losses at the interface between the CQD layer and the ECL.

To simulate the effect of BLs with different energy levels on the solar cell performance, we built an optoelectronic model using the one dimensional program SCAPS to simulate the photovoltaic performance and energy bands within the solar cells.³² For the performance simulation, the material's parameters, such as electronic energy levels, the density of states, light absorption, and conductivity, as well as interfacial properties within the solar cell were all taken into account and the simulation was performed under AM1.5G 100 mW cm⁻² illumination.^14,15,33,34Fig. 1b shows the photocurrent density–voltage (J–V) curves from the solar cell simulations as a function of CBM shift of the BL. In the model, the CBM of the BL was shifted either −0.1 eV or +0.1 eV compared to the ZnO layer, and the solar cell without any BL was also simulated for comparison. The simulated device photovoltaic parameters, such as open-circuit voltage (V_oc), short-circuit current (J_sc), fill factor (FF) and PCE, as a function of the CBM shift of the BL are summarized in Fig. 1c–f. The results suggest that the CBM shift of the BL can significantly affect the device performance. A proper upward shift of the CBM in the BL may provide a better energy level alignment between the CQD and the ECL maximizing electron extraction. By incorporating a BL with a CBM +0.1 eV higher than the ZnO layer, a simulated PCE of 10.6% was obtained, whereas 10.1% was calculated for simulating a solar cell without any BL. However, lower efficiency was obtained in the simulation by introducing a BL with a further higher or lower CBM into the solar cells, suggesting reduced electron injection from the CQD solid to the ECL due to the poor band alignment in these devices. The simulations therefore show that the solar cell performance can be optimized by tailoring the energy level of the BL and that an appropriate band alignment within the solar cell is necessary for maximizing the photovoltaic performance.

To obtain more details regarding the energy bands in the solar cell, the energy band diagram within the solar cell without any BLs and with a BL was also simulated for operation under short-circuit conditions (Fig. 2). When different materials are connected together forming a complete solar cell under illumination, a uniform electric field is generated within the CQD solid and under such a built-in electric field the photocarriers are separated and drift to the corresponding charge collecting electrodes.

Fig. 2 Energy band diagram of the CQD solar cell (a) without any BLs and (b) with a BL CBM shift of +0.1 eV in comparison to the ZnO layer under short-circuit conditions.

By incorporating a BL between the CQD and the ZnO layer, a smoother band alignment of the ZnO/BL/CQD interfaces is formed that may aid electrons sweeping out from CQDs to the ZnO electrode, diminishing the interfacial charge recombination at the interface of ZnO/CQDs (Fig. 1d). However, it should be noted that incorporation of a BL with CBM shifted upwards may shorten the depletion region in the CQD solid, especially at the maximum power point (MPP). To confirm this, the band diagram of the solar cell with a BL CBM shift of +0.1 eV referring to the ZnO layer was further modeled at the MPP (Fig. S4†), which indicates that even at the MPP the solar cell is also fully depleted without forming any quasi-neutral region in the CQD solid.

To verify the effect of MZO-BL on the photovoltaic performance and to compare the results with those from the simulations, MZO-BLs with different molar ratios of magnesium were prepared and incorporated into CQD solar cells. Fig. 3a shows the cross-sectional scanning electron microscopy (SEM) image of a complete solar cell with the device architecture of FTO/ZnO/MZO/CQD-TBAI/CQD-EDT/Au. The energy level diagram including the different materials in the solar cell is shown in Fig. 3b. For this band structure, including the MZO-BL with a favorable energy band alignment with the CQD solid, the photocarriers are likely to be quickly swept out from the CQD layer to the corresponding electrode.

Fig. 3 (a) Cross-sectional SEM image of the CQD solar cell with a MZO-BL between the CQD solid and the ZnO layer. (b) Energy level diagrams of the materials for the solar cells. Top-surface SEM images of the (c) ZnO layer and (d) MZO layer. (e) XRD patterns of the ZnO layer and MZO layer.

Fig. 3c and d display the SEM images of the surface morphology of the ZnO film and the MZO film, respectively. After magnesium doping of the ZnO film, the MZO film exhibits good uniformity over a large area and small crystal grains. Meanwhile, the MZO film has no micropores in the film, and no observed pin-holes through the film. It is notable that MgO is an insulator and therefore may affect the electron transport in the MZO layer if MgO is formed in the film.^35,36 To investigate this, X-ray diffraction (XRD) was further applied to analyze the crystal properties, which reveals that no diffraction peaks from MgO were observed in the XRD patterns, which suggests that Mg instead dopes the ZnO (Fig. 3e).

As previously discussed, with the molar proportion of magnesium in the MZO film increasing, a larger bandgap energy with CBM up-shift can be obtained, which may affect the electron injection from CQDs to the MZO layer. A series of MZO-BLs were prepared with different molar proportions of magnesium in the MZO film. Fig. 4a shows the J–V curves of the best solar cells without any BLs and with a MZO-BL, under AM1.5G 100 mW cm⁻² illumination. The solar cell with only the ZnO layer as the ECL shows a PCE of 8.2% with a J_sc of 24.2 mA cm⁻², a V_oc of 0.55 V and a FF of 0.61.

Fig. 4 Photovoltaic characteristics of CQD solar cells prepared without any BLs and with a MZO-BL. (a) J–V curves of the solar cells without any BLs and with a MZO-BL under AM1.5G 100 mW cm⁻² illumination. (b) Corresponding IPCE spectra of the solar cells without any BLs and with a MZO-BL. (c) Statistics of device efficiency as a function of molar ratio of magnesium in the MZO-BL. The molar ratio of magnesium in the MZO layer is 0–15%. (d) Steady-state efficiency and photocurrent density of the solar cell with a MZO-BL at the MPP under continuous AM1.5G 100 mW cm⁻² illumination. The illumination was blocked with a shutter at ∼2100 s to measure the solar cell under dark conditions, and the shutter was removed at ∼2250 s letting light reach the solar cell again. The molar ratio of magnesium in the MZO layer is 10% in (a), (b) and (d).

Through incorporation of a favorable MZO-BL between CQDs and the ZnO layer, the device performance was dramatically improved and a PCE as high as 9.3% was achieved with a J_sc of 25.3 mA cm⁻², a V_oc of 0.57 V and a FF of 0.65. The molar ratio of magnesium in the MZO film was 10% for this device. The V_oc and FF are generally affected by the competition between charge recombination and charge extraction.³⁷ Therefore, we suggest that the increased efficiency results both from the change in energy levels (which was shown in the theoretical model) and diminished interfacial charge recombination at the interface of ZnO/CQD, which will be discussed in the following part.

To verify the improved J_sc, the corresponding incident photon-to-current efficiency (IPCE) spectra of the solar cell without any BLs and with a MZO-BL were measured, as shown in Fig. 4b. Both solar cells exhibit broad IPCE spectra throughout the visible to the near IR wavelength region (350–1150 nm). The integrated photocurrent from the IPCE results is 23.6 mA cm⁻² and 24.8 mA cm⁻² for the solar cells without any BLs and with a MZO-BL, respectively, in good agreement with J_sc measured from the J–V results.

The dependence of the photovoltaic parameters on the molar ratio of magnesium in the MZO-BL is summarized in Table 1 and the corresponding J–V curves are displayed in Fig. S5.†Fig. 4c shows statistics over the device efficiency as a function of magnesium in the MZO-BL. The results obtained from the experimental observations are in rather good accordance with the simulated predictions from the optoelectronic model (Fig. 1f). The relatively lower efficiency in the real solar cells than the results obtained from the optoelectronic model may mainly be due to the assumption of ideal interfaces in the model, as demonstrated in the literature.^33,34 Comparing the simulations with experimental investigations we conclude that the concentration of magnesium in the MZO-BL largely influences the solar cell performance and the optimized concentration of magnesium in the MZO-BL in the solar cell reported here is ∼10 mol%. It should be noted that the MZO thin film with 10 mol% of magnesium in the film has a slightly higher conductivity (1.39 × 10⁻⁴ cm² V⁻¹ s⁻¹) compared with the ZnO film (9.0 × 10⁻⁵ cm² V⁻¹ s⁻¹), which may partly result in higher performance in the solar cells, although the difference in efficiency due to this probably is very small.³⁸ The solar cell with only MZO as the ECL (without a ZnO ECL) was also fabricated for comparison, and the solar cell exhibited a relatively lower efficiency of 8.6% (Fig. S6†), whereas it is slightly higher than the reported 8.4% in the literature with MZO as the ECL.²⁹ Compared with the solar cell with a MZO-BL, we believe that the lower performance of the solar cell with only MZO as the ETL may be due to the band alignment inside the solar cell. Using a combination of MZO and ZnO results in a stepwise energy alignment aiding the electron transfer to the contact. However, for a high concentration of magnesium in the MZO-BL, the CBM of MZO will be closer to or higher than that of CQDs, which will affect the electron collection by the ZnO/MZO electrode negatively.

Table 1 Photovoltaic parameters of the solar cell with a MZO-BL as a function of molar ratio of magnesium in the MZO-BL

Molar ratio (%)	J sc (mA cm^−2)	V oc (V)	FF	PCE (%)	Average PCE^a (%)
a The average PCE was calculated from 6–8 solar cells.
0	24.2	0.55	0.61	8.2	7.8 ± 0.3
5	24.4	0.56	0.64	8.7	8.6 ± 0.2
10	25.3	0.57	0.65	9.3	8.9 ± 0.4
15	24.6	0.56	0.64	8.8	8.6 ± 0.2

---

The solar cell stability is also an important factor for future applications. Therefore, the device stability was evaluated by placing the solar cell under a solar simulator with a continuous AM1.5G 100 mW cm⁻² illumination. Fig. 4d shows the photocurrent density and the efficiency of the solar cell with a MZO-BL under continuous illumination at the maximum power point (MPP). The illumination was blocked with a shutter at ∼2100 s letting the solar cell under dark conditions, and the shutter was removed at ∼2250 s letting the solar cell under illumination again. The photocurrent density at the MPP (J_mpp) shows almost no decrease under continuous illumination for ∼2100 s. With the continuous illumination, a slight decrease in the device efficiency was observed and ∼90% of its original efficiency was maintained after illumination for ∼2100 s, and the decreased efficiency mainly results from decreased photovoltage. During the stability test, the illumination was blocked by using a shutter so that the device was in the dark for ∼150 s and thereafter illuminated again. The photocurrent shows a response back to the original value of J_mpp, after the time in the dark. It is interesting that the device efficiency almost recovered to the original value after a short time of ∼150 s under dark conditions, which indicates that the lowered efficiency under the continuous illumination may stem from the heating of the device rather than from the degradation of the materials in the device. Unsealed solar cells were also stored under an ambient atmosphere under dark conditions and a high level performance and no large degradation were observed in 37 days (Fig. S7†). These results suggest that the CQD solar cells demonstrated here exhibit decent stability with potential towards high efficiency and stable photovoltaic devices.

Capacitance–voltage (Cap–V) analysis was performed to study the depletion region and built-in potential (V_Bi) in the solar cells without any BLs and with a MZO-BL. Fig. 5a shows the Mott–Schottky (1/C²–V) curves of both solar cells, which indicate that both of the solar cells are fully or close to fully depleted, but the higher slope in the curve for the solar cell with a MZO-BL at negative voltages suggests that the depletion is more complete in the solar cell with a MZO-BL. The V_Bi was determined from the curves,^29,30 which reveals that a slightly higher V_Bi of 0.61 V is obtained in the solar cell with a MZO-BL, whereas the solar cell without any BLs has a value of 0.59 V. The slightly increased V_Bi in the solar cell with a MZO-BL will result in a stronger built-in electric field facilitating the charge extraction, and the improved V_Bi is further confirmed by the dark J–V curves where the diode turn-on voltage, depending on the built-in potential, is higher in the solar cell with a MZO-BL (Fig. S8†).

Fig. 5 (a) Mott–Schottky curves of the solar cell without any BLs and with a MZO-BL. (b) Photocurrent density as a function of internal voltage of the solar cell without any BLs and with a MZO-BL. (c) Electron lifetime as a function of V_oc for the cells without any BLs and with a MZO-BL. (d) J_sc and (e) V_oc of the solar cells without any BLs and with a MZO-BL as a function of light intensity.

To investigate the effect of the MZO-BL on the charge recombination and charge collection, the photocurrent density (J_ph) as a function of internal voltage (V_int) was analyzed. The J_ph was calculated from the difference between the photocurrent density under illumination (J_ill) and the dark current (J_d),³⁹ as shown in Fig. S7.† The V_int is calculated by using V_Bi − V_app, where V_app is the applied voltage during the J–V measurement. Fig. 5b shows the J_ph–V_int curves, which imply that J_ph in both solar cells strongly depends on V_int so that at higher V_int more charge carriers can be swept out. However, when the V_int is greater than 0.3 V, the J_ph reaches saturation, which implies that the internal electric field is strong enough to sweep out almost all charge carriers so that they can be collected by the corresponding electrodes. However, it is notable that a higher J_ph is achieved in the solar cell with a MZO-BL.

The electron lifetime (τ) was determined from the V_oc transient decay at different light intensities with a modulated light perturbation.^40,41 A single exponential function was used to fit the transient voltage decay, and although decay processes at the very fast timescale (<microseconds) cannot be measured with this method, a clear trend was observed. The τ was measured as a function of the light intensity, which also changes the V_oc, and a decrease in τ was observed at higher light intensity (and therefore higher V_oc) which is due to the relatively faster charge recombination, as shown in Fig. 5c. We also observed that the solar cell without any BLs shows a lower τ, which may be explained by faster charge recombination from the CQDs to the ZnO layer. After incorporation of the MZO-BL into the solar cells, an increased τ was obtained, and therefore a slower charge recombination time.

The reduced interfacial charge recombination by incorporation of a MZO-BL into the solar cell can also be verified with the dependence of J_sc and V_oc on the light intensity. The J_sc strongly depends on the light intensity in solar cells and can be expressed as J_sc ∝ I^α, where I and α are the light intensity and exponential factor, respectively.^39,42,43 An α value close to unity in a solar cell suggests that the recombination during the charge carrier collection is not dependent on light intensity. As expected, α = 0.91 was observed for the solar cell with a MZO-BL, whereas the α value is 0.87 for the solar cell without any BLs (Fig. 5d), which indicates that the charge recombination is reduced under the short-circuit conditions in the solar cell with a MZO-BL, especially for light intensities close to one sun.

We also observed a similar effect on V_oc commonly measured in organic solar cells that the logarithmic plot of V_oc depends linearly on the light intensity with a slope of kT/q, where k, T and q are the Boltzmann constant, temperature in Kelvin and elementary charge, respectively.^38,44,45 The solar cell with a MZO-BL exhibits a different dependence of V_oc on the light intensity with a slope of 1.43kT/q, whereas in the solar cell without any BLs the slope is 1.58kT/q, as shown in Fig. 5e. The light intensity dependence of the photocurrent and photovoltage for the different devices therefore suggest that there is a higher recombination of charges in the device without the MZO-BL. These results suggest that the MZO-BL in the solar cell may reduce the density of interfacial traps between the CQD solid and ECL and hence reduce interfacial charge recombination and improve the charge extraction.^25,37

Incorporating a BL with favorable electronic energy levels into the CQD solar cells to enhance the charge carrier collection and V_oc is therefore an efficient approach to enhance the performance of the solar cell but the BL needs to be finely controlled for a favorable band alignment. Compared with other organic or fullerene applied as BLs to reduce the interfacial charge recombination,^24,25 the MZO-BL reported here can be easily prepared by a sol–gel method under an ambient atmosphere, which is important for future large scale fabrication of devices. Furthermore, the MZO allows for tailoring of the energy levels to obtain an energy band alignment that is advantageous in the solar cells.

3. Conclusions

In conclusion, improved performance in CQD solar cells was demonstrated by introducing a MZO-BL into the device. The tunability of the electronic energy levels in the MZO-BL by changing the concentration of Mg allows for an optimization of the energy level alignment. By introducing a MZO-BL with favorable electronic energy levels into the solar cells, higher photovoltage is obtained and the interfacial charge recombination at the ECL/CQD interface is significantly diminished, resulting in improved electron extraction from the CQDs. Benefiting from the reduced interfacial charge recombination, the solar cell with a MZO-BL shows as high as 9.3% efficiency, compared to that of 8.2% for the solar cell without any BLs. The solar cells without any sealant exhibit decent stability under ambient conditions both in the dark and under illumination. These results clearly define a potential avenue for the fabrication of solution-processed solar cells with high efficiency and good stability by introducing a MZO-BL with the possibility of band alignment engineering.

4. Experimental section

ZnO and MZO layer preparation

ZnO and MZO layers were prepared by a sol–gel method under an ambient atmosphere. The precursor for the ZnO layer preparation was prepared by dissolving zinc acetate dihydrate (0.5 M) and ethylendiamine (0.5 M) into 1-methoxy-2-propanol under stirring. The precursor for MZO layer preparation was prepared by dissolving different molar ratios (0–15%) of magnesium acetate tetrahydrate into the ZnO precursor. The precursors were stirred for 24 h at room temperature, and filtered with a filter with a pore size of 0.2 μm. The precursor was spin-coated on cleaned FTO glass substrates and annealed at 200 °C for 15 min. For ZnO as the ECL, three ZnO layers were deposited. For the band alignment of the ZnO/MZO electrode, two ZnO layers and one MZO layer were deposited in sequence. Finally, the films were sintered at 400 °C for 30 min under ambient conditions.

Device fabrication

PbS CQDs were synthesized according to the literature.²¹ The CQDs were finally dispersed in octane with a concentration of 50 mg mL⁻¹ and filtered with a filter with a pore size of 0.2 μm. CQD solution was spin-coated on top of the ZnO or the MZO films at 2500 rpm for 15 s. The CQD film was treated with 10 mg mL⁻¹ TBAI in methanol for 30 s and then spun again at 2500 rpm. The treated film was rinsed with methanol two times. These processes were repeated 8 times for CQD-TBAI layer deposition. Another two CQD layers were deposited on the top of the CQD-TBAI film and treated with 0.01% EDT in acetonitrile for 30 s. The film was rinsed 3 times with acetonitrile. Finally, the device was completed by thermal deposition of an Au film with a thickness of 80 nm on top of the CQD solid.

Device characterization

A Newport solar simulator (model 91160) was used to provide a light intensity of AM1.5G 100 mW cm⁻² (1 sun) illumination that was calibrated using a certified reference Si solar cell (Fraunhofer ISE) prior to measurement. The J–V curves were measured using a Keithley model 2400 digital sourcemeter under a nitrogen atmosphere. The device working area was defined by a black mask with an aperture area of 6.8 mm². IPCE spectra were recorded according to a previous method that uses a computer-controlled setup consisting of a xenon lamp (Spectral Products ASBXE 175), a monochromator (Spectral Products CM110), a LabJack U6, and a potentiostat (PINE, AFRDE 5).⁴⁶ The setup was calibrated using a certified reference solar cell (Fraunhofer ISE) prior to measurement. Electron lifetime, J_sc and V_oc as a function of light intensity were measured according to previous methods.⁴⁷

Cap–V measurements

Capacitance–voltage (Cap–V) curves were measured using an Autolab Potentiostat Galvanostat with a modulation frequency and amplitude of 1 kHz and 10 mV under dark conditions. The voltage sweep was performed between −1 V and 1 V with an AC signal.

Material characterization

Light absorption spectra were measured using an Ocean Optics HR2000 spectrometer with a Micropack DH-2000-BAL light source. The photoluminescence spectrum was recorded using a Fluorolog Spectrophotometer (HORIBA JOBIN YNON) with a Xe lamp. SEM images were obtained using scanning electron microscopy (SEM, Zeiss LEO1550) at an accelerating voltage of 5 kV. XRD measurements were carried out using a diffractometer D5000 (SIEMENS) with Cu Kα radiation (λ = 0.154059 nm) for diffraction angle 2θ in the range of 10–70°.

Simulation

The energy band diagram of the solar cell and photovoltaic performance are simulated using SCAPS3301 software under AM1.5G 100 mW cm⁻² illumination. The parameters for the simulation are obtained from the literature^14,15 and summarized in Fig. S9 and Table S1,† except for the thicknesses of the layers, which were 230 nm for the CQD-TBAI layer and 50 nm for the CQD-EDT layer. The band edge was calculated from the light absorption edge of the exciton peak.

Acknowledgements

This work was funded by the Göran Gustafsson Foundation, the Swedish Energy Agency, the Swedish Research Council FORMAS, ÅForsk, and the Swedish Research Council (VR).

References

1. G. H. Carey, A. L. Abdelhady, Z. Ning, S. M. Thon, O. M. Bakr and E. H. Sargent, Chem. Rev., 2015, 115, 12732–12763.
2. Z. Ning, X. Gong, R. Comin, G. Walters, F. Fan, O. Voznyy, E. Yassitepe, A. Buin, S. Hoogland and E. H. Sargent, Nature, 2015, 523, 324–328.
3. R. Saran and R. J. Curry, Nat. Photonics, 2016, 10, 81–92.
4. J. Tang, K. W. Kemp, S. Hoogland, K. S. Jeong, H. Liu, L. Levina, M. Furukawa, X. H. Wang, R. Debnath, D. K. Cha, K. W. Chou, A. Fischer, A. Amassian, J. B. Asbury and E. H. Sargent, Nat. Mater., 2011, 10, 765–771.
5. I. J. Kramer, J. C. Minor, G. Moreno-Bautista, L. Rollny, P. Kanjanaboos, D. Kopilovic, S. M. Thon, G. H. Carey, K. W. Chou, D. Zhitomirsky, A. Amassian and E. H. Sargent, Adv. Mater., 2015, 27, 116–121.
6. D. K. Ko, A. Maurano, S. K. Suh, D. Kim, G. W. Hwang, J. C. Grossman, V. Bulovic and M. G. Bawendi, ACS Nano, 2016, 10, 3382–3388.
7. M. L. Bohm, T. C. Jellicoe, M. Tabachnyk, N. J. Davis, F. Wisnivesky-Rocca-Rivarola, C. Ducati, B. Ehrler, A. A. Bakulin and N. C. Greenham, Nano Lett., 2015, 15, 7987–7993.
8. X. Zhang, G. E. Eperon, J. Liu and E. M. J. Johansson, Nano Energy, 2016, 22, 70–78.
9. T. Kim, Y. Q. Gao, H. L. Hu, B. Y. Yan, Z. J. Ning, L. K. Jagadamma, K. Zhao, A. R. Kirmani, J. Eid, M. M. Adachi, E. H. Sargent, P. M. Beaujuge and A. Amassian, Nano Energy, 2015, 17, 196–205.
10. Z. Yang, A. Janmohamed, X. Lan, F. P. Garcia de Arquer, O. Voznyy, E. Yassitepe, G. H. Kim, Z. Ning, X. Gong, R. Comin and E. H. Sargent, Nano Lett., 2015, 15, 7539–7543.
11. R. Wang, Y. Shang, P. Kanjanaboos, W. Zhou, Z. Ning and E. H. Sargent, Energy Environ. Sci., 2016, 9, 1130–1143.
12. M. J. Choi, S. Kim, H. Lim, J. Choi, D. M. Sim, S. Yim, B. T. Ahn, J. Y. Kim and Y. S. Jung, Adv. Mater., 2016, 28, 1780–1787.
13. X. Zhang, J. Zhang, J. Liu and E. M. J. Johansson, Nanoscale, 2015, 7, 11520–11524.
14. G. H. Kim, F. P. Garcia de Arquer, Y. J. Yoon, X. Lan, M. Liu, O. Voznyy, Z. Yang, F. Fan, A. H. Ip, P. Kanjanaboos, S. Hoogland, J. Y. Kim and E. H. Sargent, Nano Lett., 2015, 15, 7691–7696.
15. M. Liu, F. P. de Arquer, Y. Li, X. Lan, G. H. Kim, O. Voznyy, L. K. Jagadamma, A. S. Abbas, S. Hoogland, Z. Lu, J. Y. Kim, A. Amassian and E. H. Sargent, Adv. Mater., 2016, 28, 4142–4148.
16. X. Lan, S. Masala and E. H. Sargent, Nat. Mater., 2014, 13, 233–240.
17. D. K. Ko, P. R. Brown, M. G. Bawendi and V. Bulovic, Adv. Mater., 2014, 26, 4845–4850.
18. J. Tang, H. Liu, D. Zhitomirsky, S. Hoogland, X. Wang, M. Furukawa, L. Levina and E. H. Sargent, Nano Lett., 2012, 12, 4889–4894.
19. D. Zhitomirsky, O. Voznyy, L. Levina, S. Hoogland, K. W. Kemp, A. H. Ip, S. M. Thon and E. H. Sargent, Nat. Commun., 2014, 5, 3803.
20. C. H. Chuang, A. Maurano, R. E. Brandt, G. W. Hwang, J. Jean, T. Buonassisi, V. Bulovic and M. G. Bawendi, Nano Lett., 2015, 15, 3286–3294.
21. X. Lan, O. Voznyy, A. Kiani, F. P. Garcia de Arquer, A. S. Abbas, G. H. Kim, M. Liu, Z. Yang, G. Walters, J. Xu, M. Yuan, Z. Ning, F. Fan, P. Kanjanaboos, I. Kramer, D. Zhitomirsky, P. Lee, A. Perelgut, S. Hoogland and E. H. Sargent, Adv. Mater., 2016, 28, 299–304.
22. A. H. Ip, S. M. Thon, S. Hoogland, O. Voznyy, D. Zhitomirsky, R. Debnath, L. Levina, L. R. Rollny, G. H. Carey, A. Fischer, K. W. Kemp, I. J. Kramer, Z. Ning, A. J. Labelle, K. W. Chou, A. Amassian and E. H. Sargent, Nat. Nanotechnol., 2012, 7, 577–582.
23. Z. Fan, L. C. Lin, W. Buijs, T. J. Vlugt and M. A. van Huis, Nat. Commun., 2016, 7, 11503.
24. R. Azmi, H. Aqoma, W. T. Hadmojo, J.-M. Yun, S. Yoon, K. Kim, Y. R. Do, S.-H. Oh and S.-Y. Jang, Adv. Energy Mater., 2016, 6, 1502146.
25. M. Yuan, O. Voznyy, D. Zhitomirsky, P. Kanjanaboos and E. H. Sargent, Adv. Mater., 2015, 27, 917–921.
26. K. W. Kemp, A. J. Labelle, S. M. Thon, A. H. Ip, I. J. Kramer, S. Hoogland and E. H. Sargent, Adv. Energy Mater., 2013, 3, 917–922.
27. A. Ohtomo, M. Kawasaki, T. Koida, K. Masubuchi, H. Koinuma, Y. Sakurai, Y. Yoshida, T. Yasuda and Y. Segawa, Appl. Phys. Lett., 1998, 72, 2466–2468.
28. B. A. MacLeod, P. Schulz, S. R. Cowan, A. Garcia, D. S. Ginley, A. Kahn and D. C. Olson, Adv. Energy Mater., 2014, 4, 1400073.
29. L. Hu, D. B. Li, L. Gao, H. Tan, C. Chen, K. H. Li, M. Li, J. B. Han, H. S. Song, H. Liu and J. Tang, Adv. Funct. Mater., 2016, 26, 1899–1907.
30. C. Z. Li, C. Y. Chang, Y. Zang, H. X. Ju, C. C. Chueh, P. W. Liang, N. Cho, D. S. Ginger and A. K. Y. Jen, Adv. Mater., 2014, 26, 6262–6267.
31. Z. Yin, Q. Zheng, S.-C. Chen, D. Cai and Y. Ma, Adv. Energy Mater., 2016, 6, 1501493.
32. K. Decock, P. Zabierowski and M. Burgelman, J. Appl. Phys., 2012, 111, 043703.
33. M. Burgelman, K. Decock, S. Khelifi and A. Abass, Thin Solid Films, 2013, 535, 296–301.
34. M. Burgelman, P. Nollet and S. Degrave, Thin Solid Films, 2000, 361, 527–532.
35. C. T. Chi, I. C. Cheng and J. Z. Chen, J. Alloys Compd., 2012, 544, 111–117.
36. Y. Wang, X. Zhao, L. Duan, F. Wang, H. Niu, W. Guo and A. Ali, Mater. Sci. Semicond. Process., 2015, 29, 372–379.
37. K. F. Jeltsch, M. Schadel, J. B. Bonekamp, P. Niyamakom, F. Rauscher, H. W. A. Lademann, I. Dumsch, S. Allard, U. Scherf and K. Meerholz, Adv. Funct. Mater., 2012, 22, 397–404.
38. J. Song, E. Zheng, L. Liu, X. Wang, G. Chen, W. Tian and T. Miyasaka, ChemSusChem, 2016, 9, 2640–2647.
39. A. K. Kyaw, D. H. Wang, D. Wynands, J. Zhang, T. Q. Nguyen, G. C. Bazan and A. J. Heeger, Nano Lett., 2013, 13, 3796–3801.
40. X. Zhang, Y. Justo, J. Maes, W. Walravens, J. Zhang, J. Liu, Z. Hens and E. M. J. Johansson, J. Mater. Chem. A, 2015, 3, 20579–20585.
41. A. Hagfeldt, G. Boschloo, L. Sun, L. Kloo and H. Pettersson, Chem. Rev., 2010, 110, 6595–6663.
42. I. Riedel, J. Parisi, V. Dyakonov, L. Lutsen, D. Vanderzande and J. C. Hummelen, Adv. Funct. Mater., 2004, 14, 38–44.
43. J. K. J. van Duren, X. N. Yang, J. Loos, C. W. T. Bulle-Lieuwma, A. B. Sieval, J. C. Hummelen and R. A. J. Janssen, Adv. Funct. Mater., 2004, 14, 425–434.
44. H. Azimi, A. Senes, M. C. Scharber, K. Hingerl and C. J. Brabec, Adv. Energy Mater., 2011, 1, 1162–1168.
45. C. R. McNeill, J. J. M. Halls, R. Wilson, G. L. Whiting, S. Berkebile, M. G. Ramsey, R. H. Friend and N. C. Greenham, Adv. Funct. Mater., 2008, 18, 2309–2321.
46. X. Zhang, C. Hägglund and E. M. J. Johansson, Adv. Funct. Mater., 2016, 26, 1253–1260.
47. X. Zhang, J. Liu, J. Zhang, N. Vlachopoulos and E. M. Johansson, Phys. Chem. Chem. Phys., 2015, 17, 12786–12795.

---

Footnote

† Electronic supplementary information (ESI) available: SEM image, light absorption and photoluminescence spectra, band diagram, stability test, and J–V curves. See DOI: 10.1039/c6ta07775g

---