This protocol describes the synthesis and solution deposition of inorganic nanocrystals layer by layer to produce thin film electronics on non-conductive surfaces. Solvent-stabilized inks can produce complete photovoltaic devices on glass substrates via spin and spray coating following post-deposition ligand exchange and sintering.
We demonstrate a method for the preparation of fully solution processed inorganic solar cells from a spin and spray coating deposition of nanocrystal inks. For the photoactive absorber layer, colloidal CdTe and CdSe nanocrystals (3-5 nm) are synthesized using an inert hot injection technique and cleaned with precipitations to remove excess starting reagents. Similarly, gold nanocrystals (3-5 nm) are synthesized under ambient conditions and dissolved in organic solvents. In addition, precursor solutions for transparent conductive indium tin oxide (ITO) films are prepared from solutions of indium and tin salts paired with a reactive oxidizer. Layer-by-layer, these solutions are deposited onto a glass substrate following annealing (200-400 °C) to build the nanocrystal solar cell (glass/ITO/CdSe/CdTe/Au). Pre-annealing ligand exchange is required for CdSe and CdTe nanocrystals where films are dipped in NH4Cl:methanol to replace long-chain native ligands with small inorganic Cl– anions. NH4Cl(s) was found to act as a catalyst for the sintering reaction (as a non-toxic alternative to the conventional CdCl2(s) treatment) leading to grain growth (136±39 nm) during heating. The thickness and roughness of the prepared films are characterized with SEM and optical profilometry. FTIR is used to determine the degree of ligand exchange prior to sintering, and XRD is used to verify the crystallinity and phase of each material. UV/Vis spectra show high visible light transmission through the ITO layer and a red shift in the absorbance of the cadmium chalcogenide nanocrystals after thermal annealing. Current-voltage curves of completed devices are measured under simulated one sun illumination. Small differences in deposition techniques and reagents employed during ligand exchange have been shown to have a profound influence on the device properties. Here, we examine the effects of chemical (sintering and ligand exchange agents) and physical treatments (solution concentration, spray-pressure, annealing time and annealing temperature) on photovoltaic device performance.
Due to their unique emerging properties, inorganic nanocrystal inks have found applications in a wide range of electronic devices including photovoltaics,1–6 light emitting diodes,7,8 capacitors9 and transistors.10 This is due to the combination of the excellent electronic and optical properties of inorganic materials and their solution compatibility on the nanoscale. Bulk inorganic materials are typically not soluble and are therefore limited to high temperature, low pressure vacuum depositions. However, when prepared on the nanoscale with an organic ligand shell, these materials can be dispersed in organic solvents and deposited from solution (drop-, dip-, spin-, spray- coating). This freedom to coat large and irregular surfaces with electronic devices reduces the cost of these technologies while also expanding possible niche applications.6,11,12
Solution processing of cadmium(II) telluride (CdTe), cadmium(II) selenide (CdSe), cadmium(II) sulfide (CdS) and zinc oxide (ZnO) inorganic semiconductor active layers has led to photovoltaic devices reaching efficiencies (ƞ) for metal-CdTe Schottky junction CdTe/Al (ƞ = 5.15%)13,14 and heterojunction CdS/CdTe (ƞ = 5.73%),15 CdSe/CdTe (ƞ = 3.02%),16,17 ZnO/CdTe (ƞ = 7.1%, 12%).18,19 In contrast to vacuum deposition of bulk CdTe devices, these nanocrystal films must undergo ligand exchange following deposition to remove native and insulating long-chain organic ligands which prohibit efficient electron transport through the film. Additionally, sintering Cd- (S, Se, Te) must occur during heating in the presence of a suitable salt catalyst. Recently, it was found that non-toxic ammonium chloride (NH4Cl) can be used for this purpose as a replacement for the commonly used cadmium(II) chloride (CdCl2).20 By dipping the deposited nanocrystal film in NH4Cl:methanol solutions, the ligand exchange reaction occurs simultaneously with exposure to the heat-activated NH4Cl sintering catalyst. These prepared films are heated layer-by-layer to build the desired thickness of the photo-active layers.21
Recent advances in transparent conductive films (metal nanowires, graphene, carbon nanotubes, combustion processed indium tin oxide) and conductive metal nanocrystal inks have led to the fabrication of flexible or curved electronics built on arbitrary non-conductive surfaces.22,23 In this presentation, we demonstrate the preparation of each precursor ink solution including the active layers (CdTe and CdSe nanocrystals), the transparent conducting oxide electrode (i.e., indium doped tin oxide, ITO) and the back metal contact to construct a completed inorganic solar cell entirely from a solution process.24 Here, we highlight the spray process and the device layer patterning architectures on non-conductive glass. This detailed video protocol is intended to aid researchers who are designing and building solution processed solar cells; however, the same techniques described here are applicable to a wide range of electronic devices.
Note: Please consult all relevant materials safety data sheets (MSDS) before use. Many of the precursor solutions and products are hazardous or carcinogenic. Special consideration should be directed to nanomaterials due to unique safety concerns that arise compared to their bulk counterparts. Proper protective equipment should be worn (safety goggles, face shield, gloves, lab coat, long pants and closed-toed shoes) at all times during this procedure.
1. Synthesis of Nanocrystal Precursor Inks
2. ITO Patterning
3. Solution Processing of CdSe, CdTe and Au Films
Small angle X-ray Diffraction Patterns are used to verify the crystallinity and phase of the annealed nanocrystal film (Figure 1A). If crystallite sizes are below 100 nm, their crystal diameter can be estimated with the Scherrer equation (Eq. 1) and verified with Scanning Electron Microscopy (SEM),
where d is the mean crystallite diameter, K is the dimensionless shape factor for the material, β is the full width half maximum of the X-ray Diffraction (XRD) peak at the Bragg angle θ.
Scanning Electron Microscopy (SEM) is used to monitor the extent of grain growth in the annealed films (Figure 2B, C and Figure 3C-F). After depositing a single layer of CdTe or CdSe and heating in the presence of NH4Cl, grain size can be optimized by adjusting the temperature and duration of heating as well as the ink concentration, spray pressure/duration or spin speed. Typically, larger grains indicate devices with higher short circuit currents.12 For profile images, the glass side of the device can be scored with a diamond scribe and cracked to produce a straight edge and mounted in the SEM vertically (Figure 1B).
UV/Vis Spectroscopy is used to estimate nanocrystal size based on absorbance peak correlation with quantum confinement effects (Figure 1C-D). Crystal size can be tuned by modifying the concentration of precursors, the reaction temperature and the duration of the ink synthesis.
Optical Profilometry is used to measure film thickness and roughness. This can be conducted on a single layer of each material and on completed devices (Figure 3G-J).
Fourier Transform Infrared (FTIR) Spectra are taken to monitor the degree of ligand exchange during the NH4Cl:methanol treatment as measured by the disappearance of the C–H alkyl stretching bands at 2,924 and 2,852 cm-1 (Figure 2A).20
Current-Voltage (I-V) characteristics can be obtained in the dark and under simulated one sun illumination from a calibrated solar simulator (Figure 2D, E). Attaching the probe tips to the anode (Au) and the cathode (ITO), a photocurrent can be measured with a digital multimeter/source meter. By scanning from negative to positive potential (Ex. -1.5 V to +1.5 V), an I-V curve is produced and provides data such as the open circuit voltage (VOC) at 0.0 amps, the short circuit current (ISC) at 0.0 volts, the fill factor (FF, Eq. 2) and the efficiency (ƞ, Eq. 3),
where JMP and VMP are the current density and voltage at the maximum power point, respectively. If the software does not provide the FF, find the maximum power point by plotting the product of J and V as a function of V. For efficiency use,
where Pde is the power input per unit area from solar irradiance (100 mW/cm2). By accounting for the device area (ex. 0.1 cm2), the units cancel leaving a unitless fraction. Special consideration must be taken to mask the other devices on the substrate during measurement to avoid an excess photocurrent contribution from adjacent devices.
Figure 1. Film Characterization. X-ray diffraction patterns of each individual device layer as a single film and a completed device (A) including a cross-section SEM image of the device build from nanocrystal inks (B). UV/Vis spectra of commercial ITO (light blue) and ITO-sol (purple) on glass and absorption of CdSe-sol (red), CdTe-sol (brown) and CdSe-sol/CdTe-sol films together (black) on commercial ITO glass substrates (D), and absorption of nanocrystal precursor solutions of CdSe (red), CdTe (brown), Au (gold), and ITO (purple) prior to annealing (C). Adapted from Ref. 24 with permission from The Royal Society of Chemistry.24 Please click here to view a larger version of this figure.
Figure 2. Ligand Exchange Catalyst and Device Properties. FTIR spectra of pyridine exchanged CdTe nanocrystal films (A) dipped in NH4Cl:methanol solution (green) and in pure methanol (red) including corresponding SEM images of these films (B and C, respectively) after annealing at 380 °C for 25 sec. Current-voltage curves of an all solution processed CdSe/CdTe heterojunction device measured under 1 sun illumination (D) and a comparison of spin coated (—) and spray coated (-) Schottky devices (E) under 1 sun illumination (red) and in the dark (black). Reprinted with permission from Ref. 12. Copyright 2014 American Chemical Society and adapted from Ref. 20 and 24 with permission from The Royal Society of Chemistry.20,24 Please click here to view a larger version of this figure.
The XRD patterns exhibit clear diffraction peaks at angles corresponding to the crystal lattice dimensions for each material and the completed device (Figure 1A). Scherrer size analysis estimates crystallite sizes on the order of 100 nm for CdTe films compared to the as-synthesized nanocrystals (3-5 nm). This transformation from quantum confined nanocrystals of CdSe and CdTe to red shifted bulk-scale grains in the annealed films is shown in the UV/Vis spectra of Figure 1C-D. The thickness of the deposited films can be increased by raising the concentration of the ink or increasing the number of layers for both spin coating and spray coating. The thickness and uniformity of the film is monitored by optical profilometry (Figure 3B, G-J). Spray coated films are typically rougher (51 ± 14 nm spray vs. 22 ± 12 nm spin), although this can be reduced with higher delivery pressures and less concentrated inks.12 Once a target thickness and roughness is obtained on a single film on glass, the procedure can be applied to device fabrication. Cross-section images of the device display film thicknesses of each layer and verify intact interfaces between them (Figure 1B).24
As-synthesized nanocrystals contain a shell of long-chain native oleate ligands that interfere with film quality, leaving behind insolating organic material during heating. Pyridine exchange reactions were used to remove the oleate shell; however, as many have observed, this process is incomplete.16,26,27 Following a 18 hr pyridine exchange, residual oleate ligands remain attached to the nanocrystals as observed by their characteristic infrared stretching frequencies of the C-H alkyl groups at 2,924 and 2,852 cm-1. FTIR spectra in Figure 2A show the absence (green) and presence (red) of the native oleate ligand bound to the CdTe nanocrystals in the as-deposited pre-annealed film treated with the NH4Cl:methanol ligand exchange catalyst and methanol only, respectively. This salt treatment simultaneously replaces the residual long-chain oleate ligands with small inorganic chloride anions, while aiding in the sintering reaction. In this situation, which is unique to nanocrystals, the ligand exchange agent must remove the native ligand while also providing excess adequate sintering catalyst on the surface. Both of these processes are key components of a successful CdTe device. Previous research demonstrated that the common usage of CdCl2 can be replaced with non-toxic NH4Cl for this purpose. The resulting average grain growth of 136 ± 39 nm after annealing is shown in Figure 2B for NH4Cl treated CdTe films whereas no growth is observed for the methanol control (Figure 2C). Monitoring ligand exchange is a unique component of many nanocrystal electronic films compared to bulk scale vacuum deposition due to the inherent nature of bottom-up synthetic routes.3,30 These involve the formation of organic ligand shells that provide solution solubility for the inorganic core, although this insulating shell does not typically contribute to the optoelectronic function of the film.
Solar cell devices measured under 1 sun illumination (Figure 2D, E) show current-voltage curves from 0.1 cm2 devices. A characteristic device shown here produces VOC = 0.52 ± 0.02 V, JSC = 9.42 ± 3.2 mA cm-2, FF (%) = 43.3 ± 2.9 and ƞ (%) = 2.37 ± 0.23 under simulated sunlight. However, due to the strong link between grain growth and processing methods, small changes in annealing temperature and heating time of CdTe films can lead to large variation in the open circuit voltages and short circuit currents of these nanocrystal films leading to reported Jsc values ranging from 0.7 mA/cm2 to 25 mA/cm2 and efficiencies above 10%.12,31 Higher efficiencies are expected following enhancement of the quality and combination of materials for solution processed photovoltaics as well as other electronic devices and functional surfaces.
Compared to traditional spin-coating of nanocrystal films, spray-coating requires additional considerations due to the inherent freedoms of using an airbrush with adjustable delivery pressure, distance from substrate, angle of spray and duration. When maintaining constant CdTe ink concentrations (4 mg/ml) and nozzle distance to substrate (60 mm), increasing pressures were found to systematically decrease film roughness producing smoother, higher quality layers. Figure 3 summarizes the effect of adjusting spray pressure on the film morphology and optical properties. As a result of increasing pressure from 15 psi to 40 psi, CdTe nanocrystal films showed higher optical transmittance (Figure 3A) as a result of being physically thinner (30 nm vs 95 nm per layer, Figure 3B). At higher pressures, the spray material is dispersed into a larger area around the target substrate and less material is deposited on the device. After annealing at 380 °C, the film of nanocrystals condense with a higher packing density as ligand molecules are released and the surface areas of individual nanocrystals are reduced to larger consolidated crystal grains. Therefore, thinner films of as-deposited nanocrystals undergo a smaller change in volume, leading to fewer cracks that appear after heating. This effect produces smoother films that are virtually identical to those deposited via spin-coating. This can be observed in the SEM images and corresponding optical profilometry maps (Figure 3C-J). After optimization of the spray parameters to achieve the desired film qualities, devices can be fabricated and tested under simulated sunlight. Figure 2E shows a comparison between spin-coated and spray-coated glass/ITO/CdTe/Ca/Al Schottky devices, where the CdTe nanocrystal layer was solution processed, demonstrating minimal differences between device performance (efficiency = 2.2% for both spin-coated and spray-coated devices).
Figure 3. Nanocrystal Spray Pressure and Film Morphology. (A) Transmission of light through CdTe device films annealed at 380 °C for 25 s after spray-coated deposition at 15 (―), 20 (- -), 30 (- – -), and 40 psi (···) with a spin-coated device (blue―) for comparison. Average film thickness as a function of the spray pressure (B). SEM images split with low magnification of CdTe device films spray-coated at 15 (C), 20 (D), 30 (E), and 40 psi (F) including corresponding optical profilometry scans showing relative surface roughness (G–J). Reprinted with permission from Ref. 12. Copyright 2014 American Chemical Society. Please click here to view a larger version of this figure.
In summary, this protocol provides guidelines for the key steps involved with building a solution processed electronic device from a spray- or spin-coating deposition. Here, we highlight new methods for solution processing transparent conductive indium tin oxide (ITO) films onto non-conductive glass substrates. After a facile etching procedure, individual electrodes can be formed before spray-depositing the photo-active layers. Using a layer-by-layer technique, CdSe and CdTe nanocrystals can be deposited in air under ambient conditions from an airbrush. After ligand exchange and heat treatment, the final non-transparent conductive metal electrode can be spray-coated onto the device and heated to remove native organic ligands. This layer can also be patterned by using a masking pattern during deposition. The resulting fully solution processed, all-inorganic devices can be characterized and tested.
Particular attention should be directed to using fresh reagents as outdated materials can lead to impure or undesired products. Additionally, the conductivity of the top and bottom electrodes should be tested during device preparation. The ITO film should have a sheet resistance of at least 500 Ohms per square and the top metal film should be at least 20 Ohms per square. If the sheet resistance is higher, apply more layers of this electrode. This becomes particularly important if devices are to be connected in series or parallel as each device needs to be inter-connected electronically. Layer thickness and roughness should be carefully controlled by monitoring the effects of changing air pressure and ink concentration. Profilometry scans of these films can provide valuable feedback on the spray- or spin-coating parameters. Typically, thin rough films (>100 nm root mean squared) can lead to device shorting and inactive devices. In order to avoid shorting, deposit thicker smoother active layers, and never touch the actual device during fabrication or when measuring.
Compared to existing vacuum deposition of single crystalline materials and common lithographic cleanroom fabrication techniques, ink-based deposition of nanocrystals is less expensive and affords more freedoms to deposit on large areas or irregular surfaces. However, the quality of the interfaces between individual nanocrystals is reduced due to the presence of native organic ligands and the inherent multicrystalline nature of the film. This leads to a higher densities of impurities and defects within the film and consequently, higher electron hole recombination rates. This can be mitigated by using ligand exchange and sintering agents (e.g., NH4Cl) to enhance crystallinity throughout the film; however, this remains a fundamental issue for inorganic nanocrystal devices. Although, for material systems with a large Bohr-exciton radius like lead sulfide, PbS (~ 20 nm), sintering is not required for effective charge transport between nanocrystals. Additionally, the area of single devices is dependent on the thickness and lateral dimensions of the masking pattern. Large area (>1 cm2) devices are attainable with macroscale masking patterns; however, microscale or nanoscale patterns would be necessary for micro or quantum dimensional electronic devices.
This video protocol describes methods for the fabrication of ink-based thin film photovoltaic devices from a spray/spin coating process. However, due to the ambient air deposition, without the requirements of vacuum or controlled atmosphere, topics covered here could also be modified for ink-jet printing of inorganic devices. The lower cost of ink-based deposition compared to conventional vacuum deposition and solar cell module packaging could also lower the price of solar power by reducing the fabrication and installation costs. Additionally, this method can be applied to other materials systems and architectures, including organic semiconductors. In addition to photovoltaics, the techniques we describe for solution processing of inorganic materials could be used to construct other electronic devices such as light-emitting diodes (LEDs), capacitors and transistors.
The authors have nothing to disclose.
The Office of Naval Research (ONR) is gratefully acknowledged for financial support. A portion of this work was conducted while Professor Townsend held a National Research Council (NRC) Postdoctoral Fellowship at the Naval Research Laboratory and is grateful for internal support from St. Mary’s College of Maryland.
Oleic acid, 90% | Sigma Aldrich | 364525 | |
1-octadecene, 90% | Sigma Aldrich | O806 | Technical grade |
Trioctylphosphine (TOP), 90% | Sigma Aldrich | 117854 | Air sensitive |
Trimethylsilyl chloride, 99.9% | Sigma Aldrich | 92360 | Air and water sensitive |
Se, 99.5+% | Sigma Aldrich | 209651 | |
NH4Cl, 99% | Sigma Aldrich | 9718 | |
CdCl2, 99.9% | Sigma Aldrich | 202908 | Highly toxic |
CdO, 99.99% | Strem | 202894 | Highly toxic |
Te, 99.8% | Strem | 264865 | |
In(NO3)3.2.85H2O, 99.99% | Sigma Aldrich | 326127-50G | |
SnCl2.2H2O, 99.9% | Sigma Aldrich | 431508 | |
NH4OH | Sigma Aldrich | 320145 | Caustic |
NH4NO3, 99% | Sigma Aldrich | A9642 | |
HAuCl4.3H2O, 99.9% | Sigma Aldrich | 520918 | |
Tetraoctylammonium bromide (TMA-Br) | Sigma Aldrich | 294136 | |
Toluene, 99.8% | Sigma Aldrich | 244511 | |
Hexanethiol, 95% | Sigma Aldrich | 234192 | |
NaBH4, 96% | Sigma Aldrich | 71320 | |
Hexanes, 98.5% | Sigma Aldrich | 650544 | |
Ethanol, 99.5% | Sigma Aldrich | 459844 | |
Methanol, anhydrous, 99.8% | Sigma Aldrich | 322415 | |
1-propanol, 99.5% | Sigma Aldrich | 402893 | |
2-propanol, 99.5% | Sigma Aldrich | 278475 | |
Pyridine, > 99% | Sigma Aldrich | 360570 | Purified by distillation |
Heptane | Sigma Aldrich | 246654 | |
chloroform > 99% | Sigma Aldrich | 372978 | |
Acetone | Sigma Aldrich | 34850 | |
Glass microscope slides | Fisher | 12-544-4 | Cut with glass cutter |
Gravity Fed Airbrush | Paasche | VSR90#1 | |
Syringe needle | Fisher | CAD4075 | |
Solar Simulator Testing Station | Newport | PVIV-1A | |
Software | Oriel | PVIV 2.0 | |
Round bottom flask | Sigma Aldrich | Z723134 | |
Round bottom flask | Sigma Aldrich | Z418668 | |
Polytetrafluoroethylene (PTFE) syringe filter | Sigma Aldrich | Z259926 | |
Polyamide tape | Kapton | KPT-1/8 | |
Cellophane tape | Scotch | 810 Tape | |
Polypropylene centrifuge tube | Sigma Aldrich | CLS430290 | |
Silver epoxy | MG Chemicals | 8331-14G |