Improving the Performance of PbS Quantum Dot Solar Cells by Optimizing ZnO Window Layer
Comparing with hot researches in absorber layer, window layer has attracted less attention in PbS quantum dot solar cells (QD SCs). Actually, the window layer plays a key role in exciton separation, charge drifting, and so on. Herein, ZnO window layer was systematically investigated for its roles in QD SCs performance. The physical mechanism of improved performance was also explored. It was found that the optimized ZnO films with appropriate thickness and doping concentration can balance the optical and electrical properties, and its energy band align well with the absorber layer for efficient charge extraction. Further characterizations demonstrated that the window layer optimization can help to reduce the surface defects, improve the heterojunction quality, as well as extend the depletion width. Compared with the control devices, the optimized devices have obtained an efficiency of 6.7% with an enhanced Voc of 18%, Jsc of 21%, FF of 10%, and power conversion efficiency of 58%. The present work suggests a useful strategy to improve the device performance by optimizing the window layer besides the absorber layer.
KeywordsZnO Window layer Thin film solar cells PbS quantum dots Physical mechanism
The efficiencies of PbS solar cells was significantly improved from 4.3% to 6.7% by optimizing ZnO window layer.
Optimized ZnO window layer can reduce the surface defects, extend thedepleted-heterojunction width and align with energy band of absorber layer.
Colloidal quantum dots (CQDs) have attracted significant attention for potentially wide applications in optoelectronic devices such as solar cells [1, 2, 3], photodetectors [4, 5, 6], and light-emitting diodes [7, 8] due to low-temperature fabrication, solution-based processing, and their peculiar optoelectronic properties [9, 10, 11]. For solar cell applications, the QDs’ bandgap can be conveniently tuned via the quantum size effect in order to match the wide absorption of solar spectra. Furthermore, recently the multi-exciton generation (MEG) effect in CQD-based solar cells (SCs) was reported, which can efficiently utilize high energy photons . The above superior properties enable them as a promising light-absorbing material. In terms of device architecture, depleted-heterojunction ZnO–PbS SCs have achieved the state-of-art highest efficiency and demonstrated the outstanding atmosphere stability [12, 13, 14].
In depleted-heterojunction CQD SCs, there were numerous researches for optimizing absorber layers. In contrast, the window layer attracts less attention in spit that it plays the key roles in extracting and transporting charge carriers in heterojunction. As an n-type window layer, ZnO is an ideal candidate due to its relatively high electron mobility, environment stability, and high transparency . Even utilizing the same window layer of ZnO, different groups utilized varied thickness and obtained over 8% conversion efficiency [12, 14, 16, 17]. Bawvendi et al. utilized 120 nm ZnO layer to achieve 8.5% certified efficiency . Recently, Sargent group adopted 80 nm ZnO layer as n-layer and molecular-halide-passivated PbS QDs as absorber to obtain 9.9% certified efficiency . Considering the optoelectronic function of the window layer, the varied thickness of ZnO layer needs further optimization for CQD SCs.
For ZnO layer fabrication, a sol–gel method was commonly used to prepare ZnO layer due to its low cost and simplicity [18, 19, 20]. However, the quality of solution-based ZnO film suffers from the surface defects or dangling bonds, which may act as charge trap sites or recombination centers [21, 22, 23]. To solve the above-mentioned problems, several strategies such as surface passivation or doping were reported to control the interfacial properties of heterojunction [24, 25, 26, 27, 28]. All of them have made promising progresses in the improvement of interface quality.
Herein, we adopted a layer-by-layer (LBL) sol–gel method to optimize the ZnO window layer. The modified sol–gel method could hold stronger capability to obtain smooth junction interface and finely control film processing. On the other hand, each layer deposition was followed one time of annealing. Thus, different ZnO layer thicknesses were corresponding to varied thermal treatment time as well as varied doping concentration [29, 30]. The performance of ZnO–PbS-QD solar cells was improved by optimizing ZnO window layer. The physical mechanism was also systematically investigated. Our work was expected to support an efficient routine for device performance improvement.
3 Experimental Section
3.1 Synthesis of PbS Quantum Dots
PbS CQDs were synthesized according to the modified literature method . In this work, 0.9 g lead oxide (PbO, 99.9%) and 3 mL oleic acid (OA, 90%) were mixed with 20 mL 1-octadecene (ODE, 90%) in a 50-mL three-neck flask. The mixture was stirred and degassed at room temperature for 8 h and heated to 90 °C for 2 h. The obtained solution was then heated to 100 °C under nitrogen for 5 min, followed by injection of TMS (hexamethyldisilathiane (bis (trimethylsilyl) sulfide) solution (300 μL TMS mixed with 10 mL pre-degassed ODE) at 90 °C. After the reaction, the resulting solution was cooled to room temperature naturally. The obtained product was washed and purified 4 times by dispersion/precipitation in hexane/acetone. Finally, the cleaned QDs were dispersed in hexane and octane (vol:vol = 4:1) mixed solvents with ~15 mg mL−1 to be ready for use.
3.2 Layer-by-Layer Sol–Gel Method Deposition of ZnO Film
The ZnO precursor was prepared by dissolving 1.5 g zinc acetate dehydrate (Zn(Ac)2·2H2O, sinopharm, 99%) and 400 μL ethanolamine (NH2CH2CH2OH, sinopharm, 99%) in 20 mL 2-methoxyethanol (CH3OCH2CH2OH, sinopharm, 99%) under vigorous stirring at 60 °C for 10 h for the hydrolysis reaction in air. On a precleaned ITO/glass substrate, ZnO precursor solution was spin-coated at 4000 r min−1 for 30 s and annealed at 400 °C for 15 min, followed by repeating this process some times to reach the required thickness.
3.3 Device Fabrication
PbS CQD films were fabricated by layer-by-layer spin-coating according to the published reports . For tetrabutylammonium iodide (TBAI) ligand exchange process, QDs dispersed in hexane/octane mixed solvents was dropped on ZnO-coated substrate and then immediately spinned at 2500 r min−1 for 10 s. The obtained film was soaked in TBAI (10 mg mL−1 in methanol) solution for 1 min, followed by two-time methanol rinsing. This process obtained a TBAI-treated QD layer and the number of layers was 10–12. For PbS-EDT (1,2-ethanedithiol) layer, 0.01 vol% EDT/acetonitrile solution was used and spinned after 30 s soaking, which was followed by a 3-time acetonitrile rinsing. This process was repeated two times. The total thickness of PbS CQD film was ~240 nm. Finally, 100 nm Au was evaporated on PbS film to complete the device fabrication. The active device area (9 mm2) was defined by shadow mask. It is noted that majority of high-efficiency PbS QDSCs reported so far were obtained based on small area (<5 mm2) which was almost half of our device area.
The ZnO films were investigated by X-ray diffraction (XRD) with Cu Ka radiation (Philips, X pert pro MRD, Netherlands), UV–Vis absorption spectra (Cary, Lambda 950, America), Hall effect (Ecopia, HMS-5500, Korea), photoluminescence (PL, LabRAM HR800, France), and X-ray photoelectron spectroscopy (XPS, EDAX Inc. Genesis, America). The device cross-section was obtained from using scanning electron microscopy (FEI Nova 450, America). The J–V characteristics were measured by a Keithley 2400 source unit with Xenon lamp (Newport, 3A solar simulator, 94023A-U, Germany) as the light source with simulated air mass (AM) 1.5G irradiation at 100 mW cm−2. The external quantum efficiency (EQE) measurements were taken by a home-made setup containing a Keithley 2400 Source Measure unit and Newport monochromator. The output power was also calibrated by Si photodetectors. The work function of various ZnO films was measured by using a Scanning Kelvin Probe microscopy (SKPM, UHV-KP, KP technology, Britain) in air at dark condition. The C–V measurements were acquired with an Agilent 4200A at a frequency of 10 kHz and AC signal of 50 mV, scanning from −1 to +0.6 V, with a step size of 50 mV. The EIS of the QD SCs was performed on an electrochemical workstation (Autolab PGTSAT302N, Metrohm Autolab, Utrecht, Netherlands) in the dark with the frequency ranging from 0.1 to 106 Hz.
4 Results and Discussion
Thickness-dependent electrical properties for varied thickness of ZnO layer
Thickness of ZnO film (nm)
Carrier concentration (cm−3)
Mobility (cm2 v−1 s−1)
Conductivity (S cm−1)
1.05 × 1016
8.7 × 10−3 a
2.25 × 10−3
1.02 × 1018
3.64 × 10−1 b
5.92 × 10−2
1.70 × 1018
1.04 × 10−1 b
2.83 × 10−2
Device performance parameters obtained from Fig. 2c
μ (cm2 (v s)−1)
Jsc (mA cm−2)
Rs (Ω cm2)
Rsh (Ω cm2)
J0 (mA cm−2)
30-nm (C-ZnO) SCs
1.3 × 10−3
90-nm (O-ZnO) SCs
3.64 × 10−1 b
1.4 × 10−4
150-nm (T-ZnO) SCs
1.04 × 10−1 b
7.8 × 10−4
To study the origin of Jsc improvement, EQE spectra of three types of ZnO film-based devices are shown in Fig. 2d. There are three characteristic regions from EQE comparison. In ultra-violet region (300–400 nm), the response in control devices is highest, which agreed well with aforementioned absorption measurement results of ZnO films. Thus, the response loss for devices based on O-ZnO and T-ZnO film is mainly caused by window layer absorption. In visible region (500–800 nm), O-ZnO devices demonstrate higher and broader response. This result demonstrates that O-ZnO devices could more efficiently extract electrons from PbS QD layers. In infrared region, all three EQE values are similar among these devices, which confirm the efficient back field in PbS-TBAI/PbS-EDT device structure . To investigate the contribution of Jsc for various ZnO–PbS devices, Fig. S3a shows the integrated short-currents for C-ZnO and O-ZnO devices. Compared with C-ZnO film devices, although a loss of light absorption is found in first region (Region I, UV spectrum) for O-ZnO device, the more contributions of short-currents can be obtained from second and third regions (Region II and III, visible and infrared regions). The current density variations corresponding to Region I–III are 0.25, 2.21, and 1.09 mA cm−2, respectively. Consequently, the O-ZnO devices could more efficiently convert visible and infrared spectra into photocurrent. On the other hand, UV spectra energy only takes 4% while the visible and infrared spectra energy takes more than 90% in solar spectra energy distribution. Based on the above analysis, the improved quantum efficiency in visible and infrared regions is the main contribution to achieve the higher short-currents of ZnO–PbS SCs.
Device performance parameters extracted from the Mott–Schottky analysis
WD, ZnO (nm)
1.6 × 1016
~4 × 1016
1.0 × 1018
4.8 × 1016
1.7 × 1018
4.5 × 1016
The WPbS for C-ZnO-based devices (<151 nm, referring to Table S1) is much narrower than the other two devices. The WPbS for O-ZnO-based devices extends to 185 nm (Table 3). The higher carrier concentration of ZnO films could help to extend the WPbS and enhance the electrical field resulting in the improvement of the charge-collection efficiency.
For FF enhancement analysis, EIS was measured to investigate the interfacial properties. Figure 4c shows the Nyquist plots of varied thickness of ZnO film-based devices. Only one semicircle is obtained in these devices regardless of the ZnO film thickness. From their equivalent circuit diagrams and intercept with the horizontal axis, the O-ZnO-based devices extract a smaller series resistance. Thus, the higher FF in O-ZnO PbS QDSCs is ascribed to the decreased Rs .
The XPS spectra provide more details in terms of the surface component of ZnO. The O1s core level spectra of the C-ZnO and O-ZnO are shown in Fig. 5b. In general, the peak for ZnO are deconvoluted into three peaks: the lower-binding-energy peak (530.2 eV) is associated with the oxygen atoms in a ZnO matrix, the higher-binding-energy peaks (532.17 and 531.43 eV) are attributed to the oxygen-deficient defects such as oxygen vacancies and hydroxyl OH groups (Fig. 5c, d) . After increasing the thickness (annealing time), the relative intensities of higher-binding-energy components decreased (Fig. 5b, d), suggesting that the oxygen-deficient defects in the ZnO films are suppressed. These results, together with PL analyses, indicate that the thicker film could help to passivate window layer defects.
In the present work, we have successfully demonstrated an obvious improvement in the performance of ZnO–PbS QD SCs via optimizing the window layer. The optimized O-ZnO window layer-based PbS QD SCs showed an enhanced PCE of 58% compared with control devices. The physical mechanism for enhanced parameters (Voc, Jsc, and FF) was also systematically illustrated. It demonstrated that the O-ZnO could reduce the surface defects, extend the depleted width in heterojunction, and align with energy band of absorber layer. The above effects could be conveniently implemented by optimizing the ZnO film thickness and its parasitic thermal treatment. The present simple and reliable optimizing strategy may provide a viable reference for depleted-heterojunction solar cells.
This work was financially supported by the National Natural Science Foundation of China (61306137, 51602114), the Research Fund for the Doctoral Program of Higher Education (20130142120075) and the Fundamental Research Funds for the Central Universities (HUST:2016YXMS032), the Guangdong-Hong Kong joint innovation project (Grant No. 2016A050503012), and the Guangdong Natural Science Funds for Distinguished Young Scholars (Grant No. 2015A030306044). The authors also thank Testing Center of HUST and the Center for Nanoscale Characterization and Devices, Wuhan National Laboratory for Optoelectronics (WNLO) for facility access.
- 19.L.K. Jagadamma, M. Abdelsamie, A. El Labban, E. Aresu, G.O. Ngongang Ndjawa, D.H. Anjum, D. Cha, P.M. Beaujuge, A. Amassian, Efficient inverted bulk-heterojunction solar cells from low-temperature processing of amorphous ZnO buffer layers. J. Mater. Chem. A 2(33), 13321–13331 (2014). doi:10.1039/C4TA02276A CrossRefGoogle Scholar
- 24.H. Zhang, R.C. Shallcross, N. Li, T. Stubhan, Y. Hou, W. Chen, T. Ameri, M. Turbiez, N.R. Armstrong, C.J. Brabec, Overcoming electrode-induced losses in organic solar cells by tailoring a quasi-ohmic contact to fullerenes via solution-processed alkali hydroxide layers. Adv. Energy Mater. 6(9), 1502195 (2016). doi:10.1002/aenm.201502195 CrossRefGoogle Scholar
- 26.R. Azmi, H. Aqoma, W.T. Hadmojo, J.-M. Yun, S. Yoon, K. Kim, Y.R. Do, S.-H. Oh, S.-Y. Jang, Low-temperature-processed 9% colloidal quantum dot photovoltaic devices through interfacial management of p-n heterojunction. Adv. Energy Mater. 6(8), 1502146 (2016). doi:10.1002/aenm.201502146 CrossRefGoogle Scholar
- 39.S. Christoulakis, M. Suchea, E. Koudoumas, M. Katharakis, N. Katsarakis, G. Kiriakidis, Thickness influence on surface morphology and ozone sensing properties of nanostructured ZnO transparent thin films grown by PLD. Appl. Surf. Sci. 252(15), 5351–5354 (2006). doi:10.1016/j.apsusc.2005.12.071 CrossRefGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.