2007 Winter

MATSCI 316 Term Project

Quantum Dot Solar Cells

Stanford University Cheng-Chieh Chao Kuan-Lin Chen Joong Sun Park Pei-Chen Su ccchao1@stanford.edu kuanlinc@stanford.edu mrjpark@stanford.edu peichen@stanford.edu

3/8/2007

Solar Cells and Classical Limitation

A solar cell is a device that converts photons from light into electricity. Fundamentally, the device needs to fulfill two functions: photogeneration of charge carriers (electrons and holes) in a light-absorbing material, and separation of the charge carriers to a conductive contact that will transmit the electricity. This conversion is called the photovoltaic effect, and the field of research related to solar cells is known as photovoltaics. Due to the nature of photovoltaics, the light-absorbing material will only absorb certain energy level from photon. For example, Silicon has a band gap of 1.14 eV, which means 1.14 eV can excite the electron into conduction band. If photon has energy more than 1.14 eV, excess energy will generate heat instead of generating more electron-hole pairs. This phenomenon limits the overall efficiency of the conventional semiconductor photovoltaic device. Shockley and Queisser have calculated that the maximum thermodynamic efficiency for the single threshold absorber is around 31%. [1] Thus, several approaches to increase the efficiency have been investigated.

Theory of High Efficiency Solar Cells

- Multiple Excition Generation of Carriers One of the potential approaches to overcome this limit is carrier multiplication (CM), or multiple excition generation (MEG). The idea is illustrated in figure 1. In a traditional photon excitation in figure 1A, one phonon can generate only one pair of exciton. The excess photon energy (h-Eg) is dissipated as heat via phonon emission and therefore is wasted. In this case, the quantum efficiency (QE) of photon-to exciton conversion is zero below Eg, the energy gap, and is 100% above it. This corresponds to the Shockley Queisser limitation of the maxima quantum efficiency mentioned earlier. In figure 1B, the number of excitons produced by a single photon is only limited by energy conservation. Photons with energies Eg, 2Eg, and 3Eg produce one, two, and three excitons, respectively. The QE is increased by 100% if photon energy is increased by Eg. When Eg--> 0 the power conversion efficiency approaches 100%. [2] The MEG phenomenon has been known in bulk semiconductors since the 1950s. However, since the restrictions imposed by energy and momentum conservation, the power conversion efficiency improvement due to MEG was less than 1%. The MEG for solar cells was then deeming for a while. [3] Recently, Schaller et al discovered that while being low efficiency in bulk semiconductors, MEG can become extremely efficient in ultrasmall semiconductor nanocrystals. This lead to the quantum dots (QDs) solar cells.

Figure 1. (A) Traditional schematic of solar cell (B) Multiple excitons are generated with one absorption of a single phonon

- Why Quantum Dots Solar Cells? A quantum dot is a semiconductor nanostructure that confines the motion of conduction band electrons, valence band holes, or excitons in all three spatial directions. The dramatic quantization effects are shown when charge carriers are confined by potential barriers to small regions of space where the dimensions of the confinement are less than their deBroglie wavelength; the length scale at which these effects begin to occur ranges from about 10 to 50 nm for typical semiconductors. In general, potential barriers in one, two, or three spatial dimensions can confine charge carriers in semiconductors. These regimes are termed quantum films, quantum wires, and quantum dots, respectively. From confinement in a potential energy level is discretized and as well size decreases, spacing between energy levels increases (En ~ 1/d2 where d is well size). This leads to quantization of electron and phonon modes. In semiconductor nanocrystals i.e. quantum dots, discrete quantized energy levels are formed that affect the relaxation dynamics of hot electrons and carriers, and allow other relaxation channels to compete with phonon emissions. As Schrdinger equation explains, the states are quantized according to angular momentum in quantum dot, and translation momentum conservation is replaced with angular momentum conservation, which is less restrictive.

Figure 2. Quantum confinements in semiconductor and Quantum Dot

As mentioned at limitation of bulk semiconductor solar cell, high-energy electrons (hot electron-hole pair created by absorption of photons larger than band gap) convert their excess energy to heat through phonon emission. Normally the time scale of this relaxation in bulk semiconductor is in the order of sub picosecond scale. However, in quantum dot the generated electron-hole pairs become bound to each other due to strong quantum confinement. And the formation of discrete quantized energy level slows recombination time scale. So the slowed cooling of energetic excitons are able to enhance the photon conversion efficiency by allowing free energy to be extracted from the high energy excitons before they relax to their lowest state and produce heat.

Figure 3. Thermalization losses in solar cells

Limitation of solar cell is also related with recombination of electron-hole pair. Recombination is part of a process to restore equilibrium to a semiconductor that has been perturbed, or disturbed out of equilibrium. Perturbations can be in the form of an applied electric field, a change in temperature or exposure to light. Recombination occurs when there is an excess of carriers and they are destroyed, by recombining. When electron-hole pairs are destroyed, a negatively charged electron is attracted to a positively charged hole, and as they get together, their charges are canceled and the electron is part of a bond once again. Auger recombination is a type of band-to-band recombination that occurs when two carriers collide. The collision transfers the energy released from the recombining carrier to the surviving carrier. In other words, one carrier loses energy and the other gains it. The one that loses it is recombined, and the one that gains it goes to a higher energy level. Eventually, this highly energized carrier "thermalizes" - loses energy in small steps through heat producing collisions with the semiconductor lattice, until it eventually recombines or gains energy once more. And larger hole effective mass leads to rapid thermalization. One of the approaches to overcome those limitations (thermalization and Auger recombination) is MEG or CM. CM is very efficient in quantum-confined semiconductor nanocrystals, whereas it is inefficient in bulk semiconductors (the maximum CM-induced increase in the efficiencies of traditional solar cells is less than 1%). By analyzing Figure 4. Auger recombination in QD dynamical signatures of excitons and multi-excitons in PbSe quantum dot, it is discovered that the absorption of a single photon can produce two or even three electronhole pairs (excitons), which results in internal quantum efficiencies greater than 200% for conversion of light quanta into charge carriers. The schematic of generation/relaxation kinetics in a quantum dot is described in Figure 5. At time t1, a photon with energy greater than the CM threshold results in the generation of is absorbed hot bi-exciton(at t2) on the time scale of CM. After relaxing to its ground state (at t3) with the sub-picosecond time constant r, this bi-exciton recombines on the sub-nanosecond time scale (timeconstant A) by the non-radiative Auger process to produce a single exciton (at t4). Finally, the exciton recombines radiatively on a much slower timescale of tens to hundreds of nanoseconds. A significant difference in the relaxation behavior of bi-excitons (fast decay) and single excitons (slow decay) is the key property of nanocrystals. Therefore considering CM and Auger recombination kinetics, current should be extracted before Auger recombination occurs.

Figure 5. Evolution of the electron-hole system when photon is absorbed

- Impact Ionization To increase the conversion efficiency by reducing the loss caused by the thermal relaxation of photogenerated hot electrons and holes, we can utilize the hot carriers before they relax to the band edge via phonon emission. There are two fundamental ways to do so. One way is to produce and enhanced photovoltage, and the other way is to produce an enhanced photocurrent. [ref] The former requires that the hot carriers be extracted from the photoconverter before they cool, and the latter requires the hot carriers to produce two or more electron-hole pairs. This latter process is know as impact ionization (I.I.) and is well-known in bulk semiconductors. I.I. is an inverse process of Auger process, whereby two electron-hole pairs recombine to produce a single highly energetic electron-hole pair. In order to get the enhanced photocurrent, the rate of I.I. (rI.I.) or e-h pair multiplication needs to be greater than the rate of carrier cooling (rcooling), the rate of electron transfer (rET), and the rate of the forward Auger processes (rAuger). The schematic of the energies are shown in Figure 6. In quantum dots, the rates of hot-carrier cooling rates in bulk semiconductors can be slowed down because of the formation of discrete quantized energy levels. The photogenerated electrons and holes become bound to each other due to strong quantum confinement. The slowed cooling of excitons allows electrical or chemical free energy to be extracted before the excitons relax to lowest Figure 6. Electron multiplication by Impact Ionization state and generate heat. In arbitrarily low light intensity, the slowed cooling is theoretically possible. This is called a phonon bottleneck. As the quantum confinement dimension decreases, the spacing between discrete energy level E is larger. As E gets larger than phonon energies, the cooling of carriers would require multiphonon process. The multiphonon relaxation time was extimated from: Where c is the hot carrier cooling time, is the phonon frequency. For example, for strongly quantized electron levels with E > 0.2eV, c could be > 100ps from the above equation. Recently, reports showed the experiment result for electron cooling dynamics. If the electron and hole remain in the QD core, the e-h interaction leads to fast relaxation by Auger mechanics. However, in the InP QDs, if the surface was modified for hole trapping but not electron trapping, the increased spatial separation inhibits the Auger process and results in slower relaxation. Also, when only electrons were injected into the QD from an external redox molecule (sodium biphenyl) so that holes necessary for the Auger cooling were not present in QD, the relaxation is also slowed by an order of magnitude. Figure 7 summarizes the experimental results of the effects of e-h separation on slow cooling.
Figure 7. Summary of the experimental results of slow cooling by e-h separation

Quantum dot solar cell examples

- TiO2 solar cell In the paper by Istvan Robel et al. [6], the quantum dot solar cell was achieved with CdSe nanocrystal molecularly linked to mesoscopic TiO2 film. The fabrication process involves using bifunctional surface modifier to link the CdSe and TiO2. Details of the process are described as following: - Preparation of CdSe Nanoparticles A one-pot synthesis is used to create colloidal CdSe QDs. In particular, 0.05 g (0.39 mmol) of CdO and 0.30 g (1.1 mmol) of TDPA are mixed in a 1:2-1:3 mole ratio with 4.0 g of TOPO. The mixture is heated under nitrogen at 320 C to coordinate Cd ions with the phosphonic acid. Upon completion, the temperature is lowered to 270 C, and 0.25 mL (1M) TOPSe diluted with 4.0 mL of TOP is injected into the mixture. This causes the solution temperature to drop to 240-250 C. Subsequent growth is carried out at 270 C. On reaching the desired QD size, as determined through UV/visible spectroscopy, the temperature is lowered, stopping the reaction. - TiO2/CdSe Films for Photoelectrochemical Measurements TiO2 colloids were first prepared by hydrolyzing titanium isopropoxide in a glacial acetic acid solution followed by autoclaving the suspension. The suspension was sonicated whereupon a known amount was applied to the conducting glass plates (referred as OTE) and dried. After annealing these electrodes at 673 K, the weight of the film was determined. The average thickness of the film was subsequently determined from the weight, density, and area of the TiO2. The TiO2 films, stored in an oven (373 K), were subsequently dipped into a solution of acetonitrile containing carboxy alkane thiol (TAA, MPA, and MDA). Resulting TiO2 films, functionalized with these bifunctional surface modifiers, were then washed with both acetonitrile and toluene and transferred to a glass vial containing a suspension of CdSe QDs in toluene. Figure 8 shows the sequence of steps followed. The resulting film is referred to as TiO2/L/CdSe films (L ) TAA, MPA, and MDA). Bifunctional linker molecules (HOOC-RSH) facilitate binding CdSe QDs to TiO2. The THF solution containing HOOC-R-SH is added dropwise to a stirred CdSe suspension in THF: EtOH (1:1). CdSe nanoparticles undergo ligand exchange with HOOC-R-SH through the thiol linkage. Controlled amounts of titanium isopropoxide hydrolyzed in 1:1 THF:EtOH are added dropwise to the surface modified CdSe nanoparticles. The resulting suspensions are Figure 8. Fabrication flow of making OTE/ TiO2/CdSe then examined using transient absorption solar cell system. spectroscopy. Modification of the TiO2 film with HOOC-R-SH was carried out by immersing the OTE/TiO2 electrode in an acetonitrile-based HOOC-R-SH solution for 4 h. TiO2 has a strong affinity for the carboxylate group of the linker molecules, as demonstrated previously with a variety of sensitizing dyes. Thiol and amine groups, on the other hand, bind strongly to CdS enanoparticles. The following figure illustrates the principle behind anchoring CdSe QDs onto a nanostructured TiO2 film.

- Results Figure 9 shows that the absorption spectra of CdSe QDs to MPA prefunctionalized TiO2 film with different thicknesses. The nanocrystals retain their optical properties when bound to TiO2. Both the absorption onset and the band edge peak of the CdSe QDs are not changed. Evidence for monolayer coverage is also obtained from the dependence of the CdSe absorbance on the TiO2 film thickness. In particular, increasing the thickness of the TiO2 film causes a Figure.9: Absorption spectra of the OTE/TiO2 films linked corresponding increase in the CdSe with CdSe QDs with different thickness of TiO2 film. absorbance. The increased absorbance can Spectrum a corresponds to blank TiO2 fim. The film thickness of TiO be understood as a consequence of the 2um, respectively. 2 was (b) 0.2 um, (c) 0.5um, (d) 1um, (e) larger number of binding sites available for CdSe QDs with thick TiO2 film. The porous TiO2 matrix has interior surfaces on which CdSe QDs can bind. The linear dependence of the absorption to thickness of film suggests that the QDs are not exclusively deposited on the surface of the film but into the interior surface of the mesoscopic TiO2 film. Based on the principle of sensitizing large band gap with smaller band gap semiconductors, efforts have been made to employ CdSe as sensitizers to extend the photoresponse of TiO2 into the visible. Fig. 10 again shows how composited semiconductor system can minimize the carrier recombination. In this work, 0.5V open-circuit voltage and 0.14mA short-circuit current are obtained. Fig.11 shows the I-V curve of the device fabricated in this work. Result shows great enhancement in terms of photocurrent from TiO2 film alone to TiO2/CdSe composited semiconductor system.

Figure 10: Charge injection from CdSe to TiO2 to prevent recombination.

Figure 11: I-V curves comparison between TiO2 and TiO2/CdSe system.

- QDs Dispersed in Organic Semiconductor Polymer Matrixes One way to realize quantum dot solar cell is to combine polymers with inorganic semiconductors. Charge transfer is favored between high electron affinity inorganic semiconductors and relatively low ionization potential organic molecules and polymers. Charge transfer rates can be remarkably fast in the case of organics that are chemically bound to high density of states nanocrystalline and bulk inorganic semiconductors. In the paper by Istvan Robel et al. [7], by varying the radius of the rods, the quantum size effect can be used to control the band gap; furthermore, quantum confinement leads to an enhancement of the absorption coefficient as compared to the bulk such that devices can be made thinner. One-dimensional nanorods are preferable to quantum dots or sintered nanocrystals in solar energy conversion, since they naturally provide a directed path for electrical transport. The length of the nanorods can be adjusted to the device thickness required for optimal absorption of incident light.
Figure 12. As the nanocrystal length is increased successively from 7 nm by 7 nm to 7nm by 30 nm and to 7 nm by 60 nm, the external quantum efficiency in (A) rises by almost a factor of 3 to 54 %, under illumination of 0.084 mW/cm2 at 515 nm. (B) The corresponding current-voltage characteristics of the 7 nm by 60 nm nanorod device exhibit very high rectification ratios of 105 in the dark, a short circuit current of 0.019 mA/cm2, fill factor of 0.62 and open circuit voltage of 0.52 V under the same illumination conditions. (C) Solar cell haracteristics of this 7 nm by 60 nm CdSe nanorod device, measured with a simulated AM 1.5 Global light source, include a short circuit current of 5.7 mA/cm2, fill factor of 0.42, open circuit voltage of 0.67 V to yield a solar power conversion efficiency of 1.7 %. (D) Tuning of the photocurrent spectrum is shown for 2 devices containing 60 nm long nanorods with diameters 7 nm and 3 nm at 90 % by weight concentration in P3HT.

Method for fabricating an efficient inorganic-organic hybrid solar cell by spin-casting a solution of 90 % by weight CdSe nanorods in P3HT onto an indium tin oxide glass substrate coated with poly(ethylene dioxythiophene) doped with polystyrene sulfonic acid with aluminum as the top contact. With CdSe nanorods 7 nm by 60 nm, a power conversion efficiency of 6.9 % was obtained under 0.1 mW/cm2 illuminations at 515 nm inside an inert atmosphere of flowing argon. At this intensity, the open circuit voltage is 0.5 V, the photovoltage at the maximum power point is 0.4 V and the fill factor is 0.6. Devices were tested in simulated sunlight a power conversion efficiency of 1.7 %, open circuit voltage of 0.7 V, photovoltage at maximum power point of 0.45 V and fill factor of 0.4 were obtained. The maximum open circuit voltage is determined by the difference between the work functions of the electrodes, PEDOT:PSS and aluminum, as well as the difference between lowest unoccupied energy level in the CdSe nanorod and the highest occupied energy level in P3HT. This difference is close to solar intensity.

Figure 12. (A) The structure of regioregular P3HT. (B) The schematic energy level diagram for CdSe nanorods and P3HT showing the charge transfer of electrons to CdSe and holes to P3HT. (C) The device structure consists of a film ~ 200 nm in thickness sandwiched between an aluminum electrode and a transparent conducting electrode of PEDOT:PSS (Bayer AG), which was deposited on an indium tin oxide glass substrate. The active area of the device is 3 mm2. This film was spin-cast from a solution of 90 % by weight CdSe nanorods in P3HT in a pyridine/chloroform solvent mixture.

All of the possible QDorganic polymer photovoltaic cell configurations would benefit greatly if the QDs could be coaxed into producing multiple electron-hole pairs by the inverse Auger/impact ionization process. This is also true for all the QD solar cell systems described in Quantum Dot Solar Cells. The most important process in all the QD solar cells for reaching very high conversion efficiency is the multiple electron-hole pair productions in the photoexcited QDs; the various cell configurations simply represent different modes of collecting and transporting the photogenerated carriers produced in the QDs.

Summary
In this paper, we have discussed the principles of quantum dot solar cell. Different approaches have been introduced and studied. Also, some examples have been shown and explained. CdSe as QDs deposited on TiO2 film has been shown to increase the absorption of the light and increase the photocurrent generated from solar cell. The other example of incorporating nanorods in conducting polymer also showed significant improvements in solar cell efficiency. However, the technology in quantum dot solar cell is not mature, and there is no practical application commercially available yet. Further investigations and investments in the related field is very active and under great attention.

Reference
[1] Shockley, W.; Queisser, H. J. J. Appl. Phys. 1961, 32, 510. [2] Schaller, R. et al, Seven excitons at a cost of one: redefining the limits of conversion efficiency of photons into charge carriers, Nano Letters vol. 6 No. 3 (2006) 424 [3] Wolf, M.; Brendel, R.; Werner, J. H.; Queisser, H. J. J. Appl. Phys.1998, 83, 4213. [4] Ellingson, R. J.; Blackburn, J. L.; Beard, M.; Micic, O. I.; Yu, P.; Murphy, J.; Nozik, A. J. In Proceedings of the ECS Meeting, San Antonio, TX, 2004; Lian, T., Murakoshi, K., Rumbles, G., Eds.; to be published. [5] Nozik, A. J. In The Next Generation PhotoVoltaics: High Efficiency through Full Spectrum Utilization; Marti, A., Luque, A., Eds.; Institute of Physics: Bristol, U.K., 2003. [6] Istvan Robel, Vaidyanathan Subramanian, Masaru Kuno, and Prashant V. Kamat Quantum Dot Solar Cells. Harvesting Light Energy with CdSe Nanocrystals Molecularly Linked to Mesoscopic TiO2 Films [7] W. U. Huynh, J. J. Dittmer, and A. P. Alivisatos, "Hybrid Nanorod-Polymer Solar Cell