Reduced Carrier Recombination in PbS - CuInS2 Quantum Dot Solar Cells

Energy loss due to carrier recombination is among the major factors limiting the performance of TiO2/PbS colloidal quantum dot (QD) heterojunction solar cells. In this work, enhanced photocurrent is achieved by incorporating another type of hole-transporting QDs, Zn-doped CuInS2 (Zn-CIS) QDs into the PbS QD matrix. Binary QD solar cells exhibit a reduced charge recombination associated with the spatial charge separation between these two types of QDs. A ~30% increase in short-circuit current density and a ~20% increase in power conversion efficiency are observed in binary QD solar cells compared to cells built from PbS QDs only. In agreement with the charge transfer process identified through ultrafast pump/probe spectroscopy between these two QD components, transient photovoltage characteristics of single-component and binary QDs solar cells reveal longer carrier recombination time constants associated with the incorporation of Zn-CIS QDs. This work presents a straightforward, solution-processed method based on the incorporation of another QDs in the PbS QD matrix to control the carrier dynamics in colloidal QD materials and enhance solar cell performance.


I. Transmission electron microscopy of PbS and Zn-CIS QDs
. TEM characterizations of PbS QDs (a) and Zn-CIS QDs (b). Due to relativly lighter elements in the composition of Zn-CIS QDs compared to PbS, the TEM contrast of Zn-CIS QDs is weaker compared to that of PbS QDs.

II. Fourier transform infrared (FT-IR) spectroscopy
FT-IR spectroscopy was carried out on PbS and Zn-CIS QDs thin films before and after ligand exchanged by 3-mercaptopropionic acid (MPA) via a layer-by-layer spin-coating process (described in the experimental section). For PbS QDs (Figure S2), both oleic acid and MPA-capped QD films show carboxylate absorption signatures at 1400 -1750 cm -1 .
Upon ligand-exchange, the absorption intensity around 2900 cm -1 is significantly reduced, indicating a lost of CH stretching vibrations. This result agrees well with previous studies (K.S. Jeong et al., ACS Nano, 2012, 6(1), 89-99) and reflects the exchange of the oleic acid by MPA. Figure S2. FT-IR absorbance of PbS QD thin films spin-coated on polished silicon substrates. These QDs films are either non-treated (black curve, capped by as-synthesized ligands of oleic acid) or ligand-exchanged by MPA via a layer-by-layer procedure (red curve). FT-IR absorbance was measured in air; the CO2 in air leads to the absorption signature at about 2300 cm -1 and multiple narrow peaks in the 1600 -1900 cm -1 region are associated with the residual water vapor absorption.  . FT-IR absorbance of Zn-CIS QD thin films spin-coated on CaF2 substrates. These QDs films are either non-treated (black curve, capped by as-synthesized ligands of 1dodecanethiol) or ligand-exchanged by MPA via a layer-by-layer procedure (red curve). FT-IR absorbance was measured in vacuum.

III. Cyclic voltammetry of PbS and Zn-CIS QDs
Cyclic voltammograms were recorded under nitrogen atmosphere by a Potentiostat Galvanostat (Uniscan Instruments, PG581), with a circular gold disk (diameter = 3 mm) as the working electrode, a Pt wire as the counter-electrode, and Ag/AgNO3 (Ag wires with 0.01 M AgNO3 in anhydrous acetonitrile) as the anhydrous reference electrode. 0.1 M tetrabutylammonium hexafluorophosphate was dissolved in anhydrous acetonitrile and used as the supporting electrolyte. The reference electrode was calibrated by ferrocene/ ferrrocenium (Fc+/Fc) redox reaction. The samples used for cyclic voltammetry measurements were fabricated under inert atmosphere by coating QDs onto a 3-mmdiameter circular gold disk microelectrode (the working electrode). The QDs on this electrode also went through the ligand exchange procedure by its immersion into a solution of methanol containing 10% (v/v) of MPA for 30 s followed by subsequent solvent rinsing and drying. Multiple coating and ligand exchange iterations were performed in order to reach a compact QD films on the electrode. When the pump energy drops below 2uJ, the transients become energy independent which indicates that there is no more than one excitation per QD. The measurements presented in the paper have been performed using 0.65 uJ excitation energy.

V. Energy-dispersive X-ray (EDX)/SEM analysis of binary QD thin films
(1) Planar sample characterizations The hybrid QD layer of the photovoltaic devices described in this work composes typically ten sub-layers which are deposited by ten times of layer-by-layer spin-coating and ligandexchange process. The results of SEM imaging coupled with EDX elementary mapping and EDX spectroscopy from one sub-layer of hybrid QD film (40% volume fraction of Zn-CIS QDs in PbS QDs) on an ITO/glass substrate are shown below ( Figure S6). EDX mapping reveals a uniform presence of Pb and Cu elements in the whole sample area and thus confirms a uniform mixing of the two QD components in one sub-layer.
Quantitative analysis from the EDX spectra of the same area ( Figure S6 (d)) indicates an atomic percentage of Pb and Cu to be respectively 0.05% and 0.02%. As the QD film is composed of only two types of QDs (PbS and Zn-CIS), from the measured atomic (molar) ratio between these two QD components, combining their molar masses and densities, the volume fraction of Zn-CIS QDs in the QDs layer is calculated to be 40.6%. This result is in excellent agreement with the 40% volume fraction estimated from the formulation of binary QD solutions. Note that it is difficult to perform the same quantitative analysis on one sub-layer of hybrid QD film with 10% of Zn-CIS QDs in PbS QDs due to the scarcity of Cu content. Therefore, we performed similar analysis on a thicker hybrid QD film (with ten sub-layers) with 10% volume fraction of Zn-CIS ( Figure S7). EDX mapping also indicates a uniform mixing of the two QD components in this sample. Quantitative analysis from the EDX spectra of the same area ( Figure S7 (d)) indicates an atomic percentage of Pb and Cu to be respectively 0.45% and 0.03%. From the measured atomic (molar) ratio between the two QD components, combining their molar masses and densities, the volume fraction of Zn-CIS QDs in the QDs layer is calculated to be 10.2%. This result is in excellent agreement with the 10% volume fraction estimated from the formulation of binary QD solutions.

Figure S7. (a) SEM image of a film of hybrid QDs containing 10% of Zn-CIS QDs in PbS
QDs deposited onto an ITO/glass substrate by layer-by-layer spin-coating and ligandexchange process (with ten iterations). The scale bar corresponds to 300 nm; (b) and (c) EDX mapping of Pb (b) and Cu (c) elements on the same sample area of (a); (d) the EDX spectra detected from this area.
(2) Cross-sectional sample characterizations Similar EDX mapping and spectroscopy analysis were carried out a hybrid QD film with 10% volume fraction of Zn-CIS deposited by a layer-by-layer process with ten iterations on a Si substrate ( Figure S8). EDX mapping reveals a uniform presence of Pb and Cu elements along the thickness direction which indicates a uniform mixing of the two QD components.
Quantitative analysis from the EDX spectra of the same area ( Figure S8 (g)) indicates an atomic percentage of Pb and Cu to be respectively 4.84% and 0.26%. From the measured atomic (molar) ratio between the two QD components, combining their molar masses and densities, the volume fraction of Zn-CIS QDs is calculated to be 8.4%. This result is in agreement (within 2% error) with the 10% volume fraction estimated from the formulation of binary QD solutions.

VIII. Experimental details on synthesis
Synthesis of nanocrystals. PbS quantum dots were synthesized by a procedure reported previously [28,30] with a post-synthesis CdCl2 halide treatment. [18] In brief, in a 50 mL threeneck flask, 18 mL of octadecene, 0.45 of g lead(II) oxide, and 1.5 mL of oleic acid were loaded and degassed under vacuum under stirring at 100 °C for the formation of lead oleate.
After the lead oxide was dissolved, the mixture was kept at 125 °C under argon flow. A sulfur precursor was prepared separately inside a glovebox by mixing 10 mL of octadecene (degassed previously) and 0.18 mL of hexamethyldisilathiane. This sulfur precursor was then injected into the lead oleate solution at 125 °C. After injection, the heating was switched off and the reaction mixture was allowed to cool down to 40 °C in about 40 minutes. During this process when the temperature reached ~ 100 °C 1 ml CdCl2 precursor was injected in to the PbS QD flask for halide treatment. [18] For the preparation of CdCl2 precursor, 0.6 g cadmium chloride and 0.066 g tetradecylphosphonic acid were dissolved in 10 ml oleylamine under argon. This mixture was then degassed in vacuum for > 2 h at 100 °C and kept again under argon. Post-synthesis QD precipitation procedures are identical as those published previously [28] and they were finally dispersed in anhydrous octane.
Zn-CIS QDs were synthesized according to a reported procedure. [6] Typically, 1 mmol of copper (I) iodide (190 mg), 1 mmol of indium acetate (291 mg) and 0.1 mmol of zinc acetate (Zn(OAc)2, 18 mg) were introduced in a three-neck flask with 5 ml of dodecanethiol (DDT) and degassed under vacuum at 40 °C for 30 min. The yellow turbid suspension was then placed under argon flow and heated to 120 °C for 10 min. The resulting clear yellow solution was subsequently further heated to 220 °C. After approximately 15 minutes, the mixture was quickly cooled to room temperature. The QDs were then collected by precipitation with acetone and washed three times by successive redispersion and precipitation in hexane/ ethanol, toluene/acetone and toluene/methanol. They are finally dispersed in anhydrous octane.