Article Cite This: Langmuir XXXX, XXX, XXX−XXX
pubs.acs.org/Langmuir
Boosting the Efficiency of Quantum Dot-Sensitized Solar Cells through Formation of the Cation-Exchanged Hole Transporting Layer Sourav Maiti,†,§ Farazuddin Azlan,† Pranav Anand,† Yogesh Jadhav,§ Jayanta Dana,∥,† Santosh K. Haram,§ and Hirendra N. Ghosh*,†,‡ †
Radiation and Photochemistry Division, Bhabha Atomic Research Centre, Mumbai 400085, India Institute of Nano Science and Technology, Mohali, Punjab 160062, India § Department of Chemistry, Savitribai Phule Pune University, Ganeshkhind, Pune 411007, India ∥ Homi Bhabha National Institute, Mumbai 400094, India ‡
S Supporting Information *
ABSTRACT: In search of a viable way to enhance the power conversion efficiency (PCE) of quantum dot-sensitized solar cells, we have designed a method by introducing a hole transporting layer (HTL) of p-type CuS through partial cation exchange process in a postsynthetic ligand-assisted assembly of nanocrystals (NCs). Highquality CdSe and CdSSe gradient alloy NCs were synthesized through colloidal method, and the charge carrier dynamics was monitored through ultrafast transient absorption measurements. A notable increase in the short-circuit current concomitant with the increase in open-circuit voltage and the fill factor led to 45% increment in PCE for CdSe-based solar cells upon formation of the CuS HTL. Electrochemical impedance spectroscopy further revealed that the CuS layer formation increases recombination resistance at the TiO2/NC/electrolyte interface, implying that interfacial recombination gets drastically reduced because of smooth hole transfer to the redox electrolyte. Utilizing the same approach for CdSSe alloy NCs, the highest PCE (4.03%) was obtained upon CuS layer formation compared to 3.26% PCE for the untreated one and 3.61% PCE with the conventional ZnS coating. Therefore, such strategies will help to overcome the kinetic barriers of hole transfer to electrolytes, which is one of the major obstacles of high-performance devices.
1. INTRODUCTION
unwanted exciton trapping and recombination at the TiO2/ QD/electrolyte interface, which is the focus of many recent investigations.22,24−29 In this regard, a passivating layer of a wide band gap semiconductor such as ZnS24,25,30,31 together with SiO2 has proved to be advantageous in improving the efficiency of the TiO2/QD/ZnS/SiO2 assembly.22 As discussed in the literature, one of the main limiting factors of the performance of QDSSCs is the hole transfer to the redox electrolyte which is orders of magnitude slower than the electron injection to TiO2.26,32,33 Previously, an inorganic capping layer has been used to facilitate hole transfer to the electrolyte from the NC valence band (VB).34−36 One of the limitations was the use of NCs grown by a successive ionic layer adsorption and reaction (SILAR) technique which suffers from corrosion of the device and limited control over NC size distribution, resulting in uncontrolled surface passivation and trap state density leading to poor solar cell performance.7,14,37
Quantum dot-sensitized solar cells (QDSSCs) are promising candidates for 3rd generation solar cells because of their easy solution processability and low-cost fabrication technology, broad absorption spectra, high extinction coefficient, size, shape, and composition-tunable band gap, high photostability, and possibility of multiple exciton generation.1−12 However, till date, the QDSSC has reached about 13% efficiency which is way beyond the theoretical limits.13 In QDSSCs, exciton trapping and recombination are the major two processes that inhibit the charge carrier extraction, thus reducing their efficiency.14,15 Moreover, proper choice of the counter electrode and electrolyte is also very important and the use of the Cu2S or CuxSe electrode along with a polysulfide electrolyte is well-accepted in the literature for QDSSCs.10,15−20 As for QD sensitizers, alloy nanocrystals (NCs) having excellent opto-electronic properties and photostability compared to their individual counterparts, improve solar cell performance. The alloy structure helps to absorb in a broader range and exhibits better charge separation.21−23 Therefore, the biggest challenge in QDSSCs is to control and reduce the © XXXX American Chemical Society
Received: July 31, 2017 Revised: December 4, 2017 Published: December 8, 2017 A
DOI: 10.1021/acs.langmuir.7b02659 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
Figure 1. (A) UV−vis absorption and PL spectra, (B) TA spectra after 400 nm pump excitation at different delay times, and (C) ground-state bleach dynamics of (a) CdSe, (b) CdSSe_a, and (c) CdSSe_b NCs. methane, and ferrocene (Sigma-Aldrich) were used for cyclic voltammetric (CV) measurements. 2.2. Synthetic Procedure. The CdSe and CdSSe NCs were synthesized through a high-temperature hot injection method following previously reported procedures, with amendments as necessary, and it is described in detail in the Supporting Information.45−47 2.3. Solar Cell Fabrication. For solar cell fabrication, QDsensitized TiO2, polysulfide solution, and copper selenide (CuxSe) were used as the photoanode, electrolyte, and counter electrode, respectively. The device fabrication protocols were adopted from previous reports with required modifications, as described in detail in the Supporting Information.16,22,23,38,39 2.4. Ultrafast Transient Absorption Spectroscopy. The transient absorption (TA) measurements were carried out in films either NCs deposited on glass or on loaded on TiO2. The NCs were excited with a 400 nm pump beam, and the resulting carrier dynamics was probed in the 450−750 nm regions. The pump fluence was kept low enough so that the number of excitons per NC is less than 0.5 to avoid multiexciton-related complications. Details of the experimental setup are described elsewhere.48 2.5. IPCE-JV Measurements. For the measurements of photovoltaic performance and PCE, current density versus voltage (J−V) curves were obtained for the solar cell assembly under 1 sun illumination (100 mW/cm2) using a G-short arc Xe lamp solar simulator (Peccell, model PEC-L01) with a Keithley 2400 source meter. Using an action spectrum measurement system (Peccell, model: PEC-S20) with a xenon lamp (150 W) as the light source, incident photon-to-electron conversion efficiency (IPCE) measurements were carried out in the region of 300−800 nm. 2.6. Electrochemical Impedance Spectroscopic Measurements. The EIS measurements were performed on a BioLogic SP 300 instrument. For impedance measurements, the solar cell device was kept at a voltage close to VOC (0.5 V), and a 10 mV AC perturbation signal with a frequency of 100 KHz to 0.1 Hz was applied under dark conditions.
To the contrary, in the postsynthesis assembly, high-quality presynthesized colloidal NCs with a well-defined size, morphology, and surface properties are deposited on the TiO2 film.13,22,37−40 So far, linker-assisted deposition using short chain thiols as linkers results in the best efficiencies reported for QDSSCs because of controlled opto-electronic properties of the NCs and excellent loading on TiO2.21,37,41 Although wide band gap materials have been used as a passivating layer to enhance power conversion efficiency (PCE), the effect of a hole transporting layer (HTL) after linker-assisted deposition of NCs has not been elucidated in detail. In this paper, we have obtained significant improvement in PCE after depositing a partial cation-exchanged CuS p-type layer on CdSe and CdSSe alloy QDSSCs in the liquid junction solar cell assembly FTO/TiO2/QDs/CdS/CuS/CuxSe using a polysulfide electrolyte. Recently, organic hole transporting materials have also been used to enhance the PCE of QDSSCs where the organic layer facilitates hole extraction and prevents electron recombination between the QD (PbS) and the metal electrode (Au).42−44 However, organic hole transporting materials are costly and require rigorous synthetic protocols, whereas inorganic hole transporting CuS is easy to deposit, cost-effective, and efficient. Interestingly, after the formation of CuS at the surface, the PCE increased to 3.57% compared to the initial 2.56% PCE for CdSe NCs. The highest PCE obtained was for CdSSe NCs (4.03%) after CuS deposition, which is about 25% higher compared to the untreated one and ∼12% higher than with ZnS coating. Electrochemical impedance spectroscopic (EIS) measurements unveil that the CuS layer increases the charge collection efficiency by augmenting the recombination resistance at the TiO2/QD/ electrolyte interface, acting as a hole transporting buffer layer, indicating viability of this methodology to design highly efficient QDSSCs.
3. RESULTS AND DISCUSSION OA-capped CdSe and CdSSe alloy NCs were synthesized utilizing a high-temperature hot injection method through required modifications of the reported procedure.45−47 Figure 1A shows the absorption and photoluminescence spectra of the CdSe and CdSSe alloy NCs. The sharp first excitonic band ∼564, 14 nm Stokes shift, and symmetric photoluminescence spectra clearly ensure the good size distribution and the high quality of the CdSe NCs. The size of the CdSe NCs was determined to be ∼3.36 nm from the sizing curve provided by Peng et al.49 For CdSSe alloy NCs, two different sizes were obtained as confirmed from transmission electron micrographs (Figure S1A,B, Supporting Information). Energy-dispersive X-
2. EXPERIMENTAL SECTION 2.1. Chemicals. Cadmium oxide (CdO, 99.5%), selenium powder (Se, 99.99%), sulfur powder (S, 99.99%), oleic acid (OA, 90%), trioctyl phosphine (90%), octadecene (90%), cadmium nitrate tetrahydrate (Cd(NO3)2·4H2O), sodium sulfide (Na2S), zinc nitrate hexahydrate (Zn(NO3)2·6H2O), potassium chloride (KCl), potassium hydroxide (KOH), and mercaptopropionic acid (MPA) were purchased from Sigma-Aldrich and used as received. Copper chloride dihydrate (CuCl2·2H2O, 99%) was purchased from Alfa Aesar. ARgrade methanol, ethanol, acetone, and chloroform were used for cleaning. Deionized (DI) water was used for phase transfer and electrolyte preparation. Tetrabutyl ammonium perchlorate, dichloro B
DOI: 10.1021/acs.langmuir.7b02659 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
the TA spectra of all three NCs at different delay times. For CdSe, the ground-state 1S excitonic bleach ∼565 nm resulting from 1S(e)−1S3/2(h) transition matches well with the optical absorption. As the NCs are excited with energy higher than the band−edge transitions, in the bleach dynamics, we see an electron cooling component of 0.4 ps signifying the population of the 1S(e) state followed by triexponential recovery of the bleach signal attributed to the depopulation of the 1S(e) state through carrier trapping and recombination (Table S1, Supporting Information). For CdSSe_a and CdSSe_b alloy NCs, additional electron cooling components of 5.0 and 8.0 ps, respectively, were observed (Figure 1C). Electron cooling slows down because of the quasi type-II structure in the alloy NCs, decoupling the electronic wavefunction from the hole.47,52,53 During the synthesis of CdSSe alloy NCs due to higher reactivity of Se compared to S, the gradient structure can form with a CdSe-rich core and a CdS-rich shell.46,47,54 Therefore, the electrons spread throughout the NC, confining the hole to the core, essentially decoupling the charge carriers forming a quasitype-II alignment. In our case, both CdSSe NCs have a gradient-like structure which has been further confirmed through carrier-quenching experiments (both steady-state and time-resolved measurements), as discussed in Figure S2, Supporting Information.55 Moreover, the recombination process manifested by the long >400 ps component is much slower compared to pure CdSe in both alloy NCs. Therefore, alloy NCs are superior to their binary counterparts, having longlived electrons, both hot and thermalized, available for PCE.23,56 To obtain significant increments in PCE, the recombination of electrons between TiO2/QD interfaces with holes in the NC VB (process 3, Scheme 1) also needs to be minimized, which was obtained by putting p-type hole transporting CuS through cation exchange. At first, the conditions for cation exchange were optimized for CdSe NCs. After transferring to the aqueous phase through a ligand exchange with MPA, CdSe NCs were dropcasted on a TiO2 photoanode for 1 h followed by washing them with DI water to remove the excess NCs. The CdSe NCs deposited on TiO2 photoanodes were coated with 6 cycles of CdS through a SILAR technique. The CdS quasishell also helps to passivate the NCs and also the TiO2 surface
ray (EDS) analysis confirms the presence of all three elements, as shown in Figure S1C,D, Supporting Information. Both CdSSe alloy NCs show a high photoluminescence quantum yield and improved photoluminescence lifetime (Figure S1E, Supporting Information) compared to CdSe alloy NCs because of minimized trap states, which is consistent with the literature.46,47,50 All the optical parameters are summarized in Table 1. The X-ray diffraction measurements Table 1. Optical Properties in Terms of First Excitonic Absorption Maxima (λExciton ), Photoluminescence Maxima abs (λExciton ), Stokes Shift (Δλ), Quantum Yield (QY) and em Photoluminescence Lifetime (τPL) for CdSe, CdSSe_a, and CdSSe_b NCs material
λExciton abs
λExciton em
Δλ (nm)
QY (%)
τPL (ns)
CdSe CdSSe_a CdSSe_b
564 552 586
578 572 604
14 20 18
25 70 73
16.4 22.9 21.5
as depicted in Figure S1F, Supporting Information, further suggest a zinc blend structure for all the NCs and confirm the formation of alloy NCs, as all the three major diffraction peaks, (111), (220), and (311), shift toward a higher angle because of the incorporation of smaller-sized sulfur in the CdSe lattice. CdSSe_a NCs of 3.20 ± 0.05 nm size have Se/S ≈ 39:61 compared to CdSSe_b NCs of 4.54 ± 0.08 nm size have Se/S ≈ 29:71. Furthermore, we have performed CV experiments on CdSSe_a and CdSSe_b NCs to determine their band positions (Figure S1G, Supporting Information). For CdSSe_a and CdSSe_b NCs, the conduction band (CB) edge was found to be ∼−3.44 eV and ∼−3.48 eV (vs vacuum), respectively, which are well above the CB of TiO2 (∼−4.0 eV) suitable for electron injection. TA spectroscopy was utilized to understand the carrier dynamics in the NCs, which is one of the crucial factors influencing PCE. The NCs deposited on glass were excited with a 400 nm pump beam (number of excitons/NC was maintained
