Session A33: Physics of Photovoltaics

Search of novel photovoltaic absorbers from first-principles spectroscopic screening of hundreds of materials

Screening for candidate PV absorbers from numerous existing materials would require a good selection criterion. Initial selection criteria generally rest on the intrinsic material properties and abundance, postponing defect and contact issues to after the list of candidates has been narrowed down. The currently available Shockley-Queisser efficiency formula gives a universal efficiency vs band gap curve (no matter whether the gap is direct or indirect) and ignore all radiative recombination loss, and has proven over the years to be insufficient. Here we propose a calculable selection criterion of ``spectroscopic limited maximum efficiency (SLME)'' which considers the type of band gap (direct allowed, direct forbidden and indirect), the shape of absorption spectra and material-dependent radiative recombination loss by a simple model. First-principles quasiparticle calculations of SLME for $\sim$300 generalized I-III-VI and $\sim$500 I-V-VI materials identify over 40 candidates with higher SLME than currently used best thin-film absorbers. Analysis of the electronic structure of the top candidates reveals an interesting mechanism for high absorptivity and shows that some indirect gap materials can even be better than direct gap materials.   

First-principles electronic structure of $\beta $-FeSi$_{2}$ and FeS$_{2}$

Applying density functional theory in the framework of the full-potential linearized augmented plane-wave (FLAPW) method [1], we investigated electronic structure of potential future photovoltaic materials -$\beta $-FeSi$_{2}$ and FeS$_{2}$ in their bulk phases and for selected surface orientations and terminations. Their band gaps are examined using hybrid functionals as well as many-body perturbation theory in the GW-approximation to get insight of their photovoltaic performance. The gap nature in $\beta $-FeSi$_{2}$ changes from direct to indirect as suitable stain field is induced in the structure by epitaxially matching with Si substrate. Furthermore, we also studied the atomic and electronic structure of $\beta $-FeSi$_{2}$ and FeS$_{2}$ thin films for different orientations with different terminations. The most stable orientations are determined by comparing their cohesive energy. Detailed electronic structure calculations show that surface states originating from Fe play an important role and might determine their photovoltaic properties. \\[4pt] [1] www.flapw.de   

The effect of surface stoichiometry on the band gap of the pyrite FeS$_2$(100) surface

Iron pyrite ($FeS_2$) is experiencing a resurgence of interest for use in solar photovoltaic and photoelectrochemical cells. The main hurdle to the use of pyrite is the low open-circuit voltage of pyrite devices, which may result from gap states created by surface and bulk defects. Recently, systematic spin-polarized DFT calculations were performed for a series of pyrite $FeS_2$(100) surfaces to clarify the effect of surface stoichiometry on stability, electronic structure, and band gap. It was found that while stoichiometric and S-poor $FeS_2$(100) surfaces are semiconductors with band gaps of 0.56-0.72 eV, S-rich surfaces are small-gap semiconductors ($E_g < 0.3$ eV) or metals. The stoichiometric FeS2(100) surface is spin polarized in the topmost layer (2 $\mu_B$ per Fe) and displays a band of Fe $d_{z^2}$ gap states centered $~0.2$ eV above the valence band edge. Our calculations suggest that the low open-circuit voltage of pyrite solar cells may result from a narrowed surface band gap. S-poor surfaces may provide larger photovoltages than S-rich surfaces. The segregation process of sulfur vacancy under different surface conditions are also being studied, so as to provide useful guidelines for the design and fabrication of better pyrite photovoltaic materials and devices.   

Band gap engineering of Zn based II-VI semiconductors through uniaxial strain

The electronic structure of bulk wurtzitic Zn$X$ ($X$=O, S, Se, and Te) under uniaxial strain along the [0001] direction is investigated using hybrid density functional theory calculations and many-body perturbation theory. It is found that uniaxial tensile and large compressive strains decrease the band gap, similar to what has been predicted by semilocal density functional theory (DFT) calculations [Yadav et. al, Phys. Rev. B, \textbf{81}, 144120 (2010)]. Moreover, the change in the band gap under uniaxial strains predicted by semilocal DFT is in good quantitative agreement with the present results at all strains considered, thereby bringing a measure of redemption to conventional (semi)local DFT descriptions of the electronic structure of at least this class of insulators. The present results have important implications for band gap engineering through strain, especially for complex systems containing a large number of atoms (e.g., nanowires) for which higher-level calculations may be too computationally intensive.   

Searching Stable CuxS Structure for Photovoltaic Application

Employing the density-functional theory (DFT) methods, we systematically search for the most energetically favorable CuxS compounds in a wide range of 1.2 $<$ x $\le $ 2.0 guided by the experimentally identified four minerals, i.e., the chalcocite (x = 2), djurleite (x = 1.94), digenite (x = 1.80), and and anilite (x = 1.75) compounds. For the chalcocite Cu2S, all its three phases have direct band gaps of 1.3-1.4 eV, with the chalcocite low phase being more stable than other two phases, i.e., the chalcocite high and the chalcocite cubic. According to our calculation, the poor durability of the Cu2S is mainly due to the energetically favorable formation of Cu vacancies. The calculated formation heat as a function of x shows that the anilite Cu1.75S is the most stable structure. Unfortunately, this material is not a good light absorber because of its metallic feature. We propose that doping of the anilite Cu1.75S with interstitial Sn atoms may result in a compound Cu1.75Sn0.125S with an optimum direct band gap of 1.37 eV. Such a material has the ability of light absorption similar to the chalcocite Cu2S.   

Si Nanoparticles embedded in solid matrices for solar energy conversion: electronic and optical properties from first principle calculations

In device applications for solar energy conversion, nanoparticles are often embedded in a solid matrix, either crystalline or amorphous. At present a detailed understanding of the influence exerted by the embedding matrix on absorption of sunlight by the nanoparticle, and the role of the nanoparticle-matrix interface remains elusive. Building on a previous study of Si nanoparticles embedded in SiO2 [1], we investigate Si nanoparticles embedded in ZnS, used in recent experiments as a charge transport layer. A realistic model of the nanoparticle-matrix interface is created by performing ab-initio molecular dynamics simulations, and electronic and optical properties of the embedded Si nanocrystals, are obtained by first principles.\\[4pt] [1] T.Li, F.Gygi, G.Galli, Phys, Rev. Lett. 107, 206805 (2011)   

Multiple exciton generation in silicon nanoparticles

Multiple Exciton Generation (MEG) in semiconductor nanocrystals is considered to be a promising path to improve the efficiency of solar energy conversion. Recent experimental and theoretical studies indicate that MEG is more efficient in nanoparticles (NPs) than in the bulk only on a relative energy scales in units of the gap. The primary cause of this is that quantum confinement increases the energy gap substantially in NPs. MEG will be a true breakthrough when nanocrystals are identified whose impact ionization rate is enhanced even on the absolute energy scale. For this search, we calculated impact ionization rates of silicon NPs with diameters up to 2 nm using density functional theory. Our calculations clearly demonstrate that surface reconstruction creates classes of new states at low energies, de facto lowering the NP gap and thus holding the promise of inreasing MEG even on the absolute energy scale. Our calculations include static screening within the random-phase approximation. We show that a full treatment of the transition matrix elements is essential to obtain correct results due their strong energy dependence.   

All-Carbon Photovoltaics

We present an alternative scheme for nanostructured solar cells, where carbon nanomaterials are the only constituents of a bulk-heterojunction acrive layer and fulfill the role of absorbers, donors and acceptors, in the absence of conjugated polymers. Ab-initio simulations were employed to calculate the band alignment for interfaces between carbon nanotubes, fullerene derivative PCBM and reduced graphene oxide, showing the presence of Type-II and Schottky heterojunctions useful for charge separation in the active layer. Accordingly, we prepared all-carbon solar cells with optimized proportions of these three components that achieved AM1.5 efficiencies up to 1.5{\%}, with fill factors up to 70{\%} and increased thermal stability and lifetime compared to polymer-based devices. Our results show the potential of all-carbon solar cells as an alternative to polymer based ones: the key combination of high carrier mobility, visible and IR absorption and stability under illumination makes them suitable for next-generation flexible photovoltaics.   

Porous Silicon Applications for Photovoltaic Solar Cell Devices

The photovoltaic industry searches for low cost, energy competitive solar cell modules, and the usual material used is crystalline silicon, which has been covering 90{\%} of the solar module market for a long time. Porous silicon (P-Si) is used in photovoltaic applications as an ultra efficient anti-reflection coating, while a graded layer with an expanded band gap offers increased absorption in the visible spectrum regions. We have built P-Si solar cell devices having considered different physical Si wafer parameters such as crystal orientation, resistivity, and doping levels; which crucially affect the device efficiency. Porous Si wafers, prepared after etching crystalline silicon in high HF concentrations, exhibit fluorescence in the purple wavelength region of the visible spectrum under UV illumination. We are now in the process of improving the efficiency of the device by modulating the structure of the P-Si wafer, and studying its photovoltaic characteristics.   

Quantum-dot nanostructures for effective harvesting, detection, and conversion of IR radiation

Our novel approach to improve harvesting, detection, and conversion of IR radiation is based on engineering of three-dimensional potential barriers introduced by quantum dots with built-in charge due to inter-dot doping. The barriers around dots exponentially suppress capture processes and increase the photoelectron lifetime. The built-in-dot charge also strongly enhances the coupling of QD structures to IR radiation. Both effects radically improve the responsivity of IR photodetectors and photovoltaic efficiency of quantum-dot solar cells. Here we report a 50{\%} increase in photovoltaic efficiency in quantum-dot solar cells as well as 25 times increase of the photoresponse of quantum-dot infrared photodetectors when the built-in-dot charge increases up to six electrons per dot. We also present results of modeling of photoelectron kinetics and discuss perspectives of IR photodetectors and solar cells based on quantum dots with built-in charge.   

Device analysis of vertically aligned single-core nanocoaxial solar cells

Analytical expressions for device transport are derived and numerically calculated for an array of vertically aligned single-core nanocoaxial solar cells. Total current of the device is derived as a function of the geometrical configuration of the photovoltaic junction, and expressions for rectifying current behavior are subsequently solved for and analyzed. Fundamental differences and similarities in the physics of device performance are inferred based on the analytical expressions for planar, nanowire, and nanocoaxial solar cells. To emphasize the physical difference of device performance pertaining to geometrical configuration, a comparison between planar and nanocoaxial device performance is analyzed for an amorphous silicon p-i-n junction solar cell.