Session H33: Focus Session: Scaleable Technologies for Photovoltaics

Solar Glitter -- Microsystems Enabled Photovoltaics

Many products have significantly benefitted from, or been enabled by, the ability to manufacture structures at an ever decreasing length scale. Obvious examples of this include integrated circuits, flat panel displays, micro-scale sensors, and LED lighting. These industries have benefited from length scale effects in terms of improved performance, reduced cost, or new functionality (or a combination of these). In a similar manner, we are working to take advantage of length scale effects that exist within solar photovoltaic (PV) systems. While this is a significant step away from traditional approaches to solar power systems, the benefits in terms of new functionality, improved performance, and reduced cost for solar power are compelling. We are exploring scale effects that result from the size of the solar cells within the system. We have developed unique cells of both crystalline silicon and III-V materials that are very thin (5-20 microns thick) and have very small lateral dimensions (on the order of hundreds of microns across). These cells minimize the amount of expensive semiconductor material required for the system, allow improved cell performance, and provide an expanded design space for both module and system concepts allowing optimized power output and reduced module and balance of system costs. Furthermore, the small size of the cells allows for unique high-efficiency, high-flexibility PV panels and new building-integrated PV options that are currently unavailable. These benefits provide a pathway for PV power to become cost competitive with grid power and allow unique power solutions independent of grid power.   

Perspective on the Practical Limits and Roadmap for Nanostructured Photovoltaics

The practical efficiency limits for nanostructured photovoltaics including organic small molecule, dye-sensitized, polymer, and colloidal-quantum-dot architectures are assessed \textit{a posterori}. Five decades since Shockley and Queisser derived the theoretical power conversion efficiency limit of single-junction photovoltaic cells, researchers have still not demonstrated such high performance for any photovoltaic device system. Hence, in evaluating the achievable performance of a comparatively new photovoltaic technologies, such as nanostructured PVs, it is prudent to estimate the upper limit of achievable efficiencies based on trends of the best technical demonstrations across the nanostructured platforms. This analysis is utilized to give a clear perspective on the potential market viability of these technologies in the near future and outline the challenges necessary to overcome this threshold. These technologies are compared and contrasted to provide an overview for the potential of each for reducing thermal losses with ``Third Generation'' concepts accessible to nanostructured PVs that can subsequently impact cost structures.   

Role of radiative recombination in 1 eV GaInNAs solar cells

High quality GaInNAs p-i-n solar cells with depletion widths in excess of 1$\mu $m for material absorbing in the practically important 1eV band gap regime are presented [1]. This is achieved through optimization of post-growth rapid thermal annealing at a temperature of $\sim $ 910\r{ } C. Despite the improvements in material quality evidenced by a low background impurity concentration and improved minority carrier collection, the external quantum efficiency remains limited to $\sim $ 50{\%}. This is attributed to losses due to efficient radiative recombination in the bulk GaInNAs intrinsic region enhanced via localization of carriers in alloy fluctuations. \\[4pt] [1] Sellers \textit{et al}. Applied Physics Letters \textbf{99}, 151111 (2011)   

Device architectures for efficient photovoltaics from hard-to-dope semiconductors

In the search for cheap and efficient next-generation solar cells, much attention and effort has focused on semiconductor absorber materials which are difficult to dope both p and n, because of self-compensation, trap creation, or other effects. Such materials include oxides, sulfides, nanoparticles, organics, and so on. Even when the material itself has desirable electrical and optical properties for photovoltaic performance, the lack of a p-n homojunction architecture hampers device efficiency. Heterojunctions are a frequent solution, but compatible semiconductors are often unavailable or suboptimal. Therefore, we have explored new, cost-effective device architectures that promise performance comparable to a p-n homojunction, but which require neither bipolar doping nor compatible materials for heterojunctions. These architectures have the potential to bring emerging hard-to-dope semiconductors into technological and commercial relevance.   

Efficiency Analysis and Demonstration of Split-Junction Photovoltaic Solar Cells

Recently, it has been proposed that separating solar radiation with a split-junction solar cell can result in higher overall conversion efficiencies, than are possible for monolithic designs. This hypothesis is investigated by simulating and analyzing 2+1 split junction cells for efficiency comparisons with theoretical and actual multi-junction cells. Ideal band-gaps for simultaneously operating photovoltaic and thermophotovoltaic cells have been determined. With the new configuration, it is shown that the efficiency achievements previously set by Ge/InGaAs/InGaP cells can be surpassed. A total increase in power output is observed during field tests using a Cassegrain split-junction concentrator with a dichroic lens (1.1 micron cutoff wavelength). Proposed benefits such as reduced heat load on the solar cell and ease of lattice constant matching in cell design are also validated. Additionally, with the flexibility of the concentrator assembly, it is shown that similar split-junction configurations with matching dichroic lenses allow for significant improvements in high efficiency solar cell technology.   

Electronic structure and optical spectra of Bulk and Nanocrystalline CuInS$_2$

Chalcopyrite CuInS$_2$ is a promising candidate for semiconductor photovoltaic devices. Here we study the behavior of the electronic structure and optical spectra of nanocrystalline CuInS$_2$. We determine the bulk band structure using the HSE06 Hybrid Density Functional as implemented in the Vienna Ab-initio Simulation Package (VASP). We find an equilibrium structure in good agreement with experiment, and a direct band gap of 1.2~eV, as compared to the experimental value of 1.5~eV. The band gap is extremely sensitive to the position of the sulfur atoms, which suggests that it can be controlled in part by the doping of the In site with Ga. CuInS$_2$ is nearly cubic, so we fit the first-principles band structure to a k.p Hamiltonian with invariants consistent with the departure for cubic symmetry. This Hamiltonian is then used to describe the hole spectra in a CuInS$_2$ nanocrystal. We show the symmetry breaking inherent in the chalcopyrite structure can activate the optically passive ground hole state in the nanocrystal, and discuss the resulting optical behavior of the nanocrystal.   

Impact of a high resistivity transparent (HRT) layer on ultra-thin CdTe devices

Pushing the limits of the absorber layer thickness to submicron levels without compromising the efficiency in CdTe devices is important for several reasons. Reducing the thickness of the CdTe layer can increase the manufacturing speed and lower the material usage, production time, cost and energy needed for production. Thickness fluctuations in these submicron devices can lead to shunting when these fluctuations are comparable to the absorber layer thickness. Introducing a high resistivity transparent (HRT) layer in the device structure can reduce such shunting. We find that the impact of the HRT layer is higher for ultra-thin absorber layers. In this study, the effect of using sputtered ZnO as the HRT layer was investigated for devices with a 0.2 $\mu $m CdTe layer. Our results show a 40{\%} increase in efficiency in the CdTe devices with the HRT layer compared to the devices without the HRT layer. The increase in efficiency is primarily from improved fill factor due to an increase in shunt resistance from 0.15 kohm-cm$^{2}$ to 0.32 kohm-cm$^{2}$. As a result of increased shunt resistance there is an increase in the efficiency as well as a substantial increase in the yield for small dot cells.   

Solution-based synthesis of crystalline silicon thin films from liquid silane inks

Silicon (Si) dominates the photovoltaics industry and there is a need for new approaches that can significantly reduce fabrication cost. In this context, we report a non-vacuum, solution-based process for the synthesis of crystalline silicon (c-Si) thin films from liquid cyclohexasilane (CHS) in a platform that is readily applicable to large-area flexible devices. UV-polymerization during spin coating leads to the formation of thin films, which were crystallized via thermal and laser annealing. Structural changes in the films were examined using SEM, AFM and Raman spectroscopy. Subsequent chemical annealing through atmospheric-pressure hydrogen plasma treatment led to a four-decade enhancement in film conductivity, which we attribute to a disorder-order transition in a bonded Si network.   

Device modeling for organic solar cells

Organic solar cells (OSCs) are expected to play an important role in addressing our future energy needs due to their low cost and low processing requirements compared to inorganic solar cells (ISCs). However the efficiency of OSCs is still too low in comparison with ISCs for widespread applications. The biggest loss of quantum efficiency (QE) in OSCs is due to the limited free carrier generation occurring at the donor-acceptor (D-A) interface. Excitons (bound electron-hole pairs) are generated in the bulk by photo-absorption, but only a portion of them reach the D-A interface where they can dissociate into free charge carriers. Therefore, better understanding and control of exciton diffusion, free carrier generation and recombination are critical in order to improve QE for OSCs. To elucidate the physics of OSCs and aid in experimental studies, we developed a drift-diffusion model to describe the dynamics of excitons and free charge carriers. Our model predicts the performance of OSC devices by calculating their QE and current-voltage curves (I-V), as well as many other important physical quantities, such as the internal electric field, and the concentration and flux of excitons and free carriers. The effect of exciton and free carrier mobility, device temperature, and layer thickness, will be discussed. Furthermore, the exciton dissociation mechanism widely described by Onsager's model, will be investigated in detail.   

Combined Photo and Thermionic Energy Conversion with Doped Diamond Electron Emitters

Conversion of heat into electrical energy has been demonstrated using low effective work function diamond films achieved with n-type doping and surface hydrogen termination. Recently, visible light photo-electron emission has been demonstrated from the same diamond, and this work suggests that this effect could be utilized for a new approach to solar energy conversion namely combined photo and thermionic energy conversion. This work presents a spectroscopic study of photo- and thermionic electron emission from nitrogen doped diamond films on silicon substrates. In this experiment the diamond samples are heated from 100\r{ }C to 500\r{ }C, while being illuminated with light from 240 to 600 nm. The emission spectra show a significant increase of photo-emission intensity with elevated temperature and a lowering of the effective work function. The results are discussed in terms of the photo and thermal excitation, the carrier transport and the electron statistics. The results indicate the potential of diamond films in a combined photo and thermionic energy conversion solar cell. This research is supported through the Office of Naval Research.   

Metal nanoparticle-graphene superstructures as electrodes for solar cells

We present metal nanoparticle and graphene superstructures as a potential electrode for solar cells. We show the effect of various metallic nanoparticles on the work function, sheet resistance, and optical properties of graphene layers. The geometries studied include nanoparticles on single layer graphene, and embedded in a graphene sandwich. We discuss the application of these electrodes to organic and silicon solar cells.   

Study of the crystalline phases in paste coating deposition of CIGS

One or two microns of CIGS can absorbs most of the incident solar radiation because CuIn$_{1-x }$Ga$_{x}$Se$_{2}$ have a direct band gap with a high absorption coefficient. The theoretical predictions explain that the optimum photovoltaic performance should be provided by a high gallium concentration, but experimentally is observed that above 30{\%} of Ga the efficiency is reduced. This contradictory behaviour is not completely understood. Using paste coating technique for CIGS deposition, we have recently shown a strong correlation among Ga concentration, structural properties and compound stoichiometry. The desired stoichiometric compound will be obtained varying the concentration of the basic elements in the paste. Our diffraction data show that the maximum of the crystalline phase is reached when CIGS have a Ga concentration higher. Furthermore the SEM EDX quantitative analysis performed on the same samples have shown the presence of different phases. Such a phases separation find a theoretical explanation as ODC (Ordered Defect Compound) in Cu-poor system. X-ray absorpition spectroscopy reveals the structure around a specific atomic species and allow us to understand the type of phases in chalcopyrite strucuture.   

A large-area, LED-based spectral response measurement system for solar PV device characterization

Accurate and reliable measurement of the spectral responsivity (SR) of a solar cell is an important step in evaluating the electrical performance of competing photovoltaic (PV) technologies. We have investigated ways to measure the spectral responsivity, and hence the external quantum efficiency, of solar cells using measurement techniques that employ light emitting diodes (LEDs). Our setup includes one or more plates of compactly-installed, high-powered LEDs each containing up to 32 different LEDs that span the wavelength range of 375 nm to 1200 nm. Each LED plate is placed at the entrance of a tapered, highly reflective light guide for light mixing and large-area projection. Two unique measurement techniques have been investigated at NIST. The first technique consists of an LED sweep algorithm where a pulsed signal is applied to a given LED and the photogenerated current from the device under test is recorded using a lock-in technique. In the second SR technique, 32 variable-frequency, pulsed signals are applied to all LEDs at the same time, while recording the photogenerated current by a spectrum analyzer in the frequency domain. We will describe the uniqueness and advantages offered by each technique in detail and compare the accuracy of the two methods. A scheme for providing light bias and its impact on the SR measurements will be reported.