Symposium Organizers
F. (Shadi) Shahedipour-Sandvik State University of New York-Albany
Kenneth Jones Army Research Lab – SEDD
L.Douglas Bell Jet Propulsion Laboratory
Blake Simpkins Naval Research Laboratory
Daniel Schaadt Karlsruhe Institute of Technology
Miguel Contreras National Renewable Energy Laboratory
Symposium Support
General Electric - Global Research
JLN Solar
D5: Poster Session: Alternative III-V Materials
Session Chairs
Tuesday PM, April 26, 2011
Exhibition Hall (Moscone West)
D8: Poster Session: Nano-Materials and Devices
Session Chairs
Tuesday PM, April 26, 2011
Exhibition Hall (Moscone West)
D1: Nitrides: Strain and Defects
Session Chairs
Tuesday PM, April 26, 2011
Room 2003 (Moscone West)
9:30 AM - **D1.1
Contacts to Group III Nitride Semiconductors.
Suzanne Mohney 1
1 , Penn State, University Park, Pennsylvania, United States
Show AbstractContacts to group III nitride semiconductors are required for the efficient operation of devices for energy conservation (light emitting diodes for white lighting) and conversion (photovoltaic devices). In this presentation, the status of contacts to GaN and related alloy semiconductors will be reviewed, with a particular focus on the contact metallurgy. Ohmic contacts that provide low resistance and satisfy other device constraints will be highlighted. Most work on ohmic contacts has been performed on the metal face of the semiconductor, but the role of the crystallographic orientation and polarity of the semiconductor will also be discussed. Finally, research on contacts to GaN nanowires will be described. These experiments further point to the influence of crystallographic orientation on the resistance of ohmic contacts.
10:00 AM - D1.2
Strain Relaxation in Semipolar Nitride Materials for Optoelectronic Device Applications.
Ingrid Koslow 1 , Erin Young 1 , Matthew Hardy 1 , Michael Cantore 1 , Stuart Brinkley 2 , Shuji Nakamura 1 2 , Jim Speck 1 , Steven DenBaars 1 2
1 Materials Department, University of California Santa Barbara, Santa Barbara, California, United States, 2 Electrical and Computer Engineering, University of California Santa Barbara, Santa Barbara, California, United States
Show AbstractPerformance of devices grown on partially strain-relaxed InGaN layers is reported. The (In)(Al)GaN material system is highly attractive for many devices, including light emitters from the UV to the visible regions, photovoltaics, power electronics, and thermoelectrics. In particular, InxGa1-xN layers with x > 0.2 are necessary for emitters and absorbers in the longer-wavelength visible region, while AlyGa1-yN layers with y > 0.4 are desired for deep ultraviolet (UV) emitters. However, strain in such layers can lead to significant defect generation, as well as high piezoelectric fields, which can be detrimental to device performance.Partial strain relaxation via misfit dislocation (MD) formation was recently reported for the first time in semipolar InGaN and AlGaN films, on both (11-22) and (20-21) orientations[1,2], where the basal c-plane acts as the primary slip system. Given the high lattice mismatch of the (Al,In)GaN material system, the availability of a slip system for dislocation glide has significant consequences for semipolar III-nitride devices. However, thus far the critical thickness for relaxation of InxGa1-xN layers has not been investigated in the literature beyond x ~ 0.05. In this report, strain relaxed InxGa1-xN layers with 0.05 < x < 0.25 have been grown on (11-22) bulk GaN substrates by MOCVD, and characterized by x-ray diffraction, atomic force microscopy, cathodoluminescence, and transmission electron microscopy. We report here for the first time on devices grown on strain-relaxed layers. Long-wavelength light emitting diode (LED) structures were grown on top of partially relaxed InGaN layers, resulting in reduced strain in the active region, which shows promising performance.[1] A. Tyagi, F. Wu, E. C. Young, A. Chakraborty, H. Ohta, R. Bhat, K. Fujito, S. P. DenBaars, S. Nakamura, and J. Speck: Appl. Phys. Lett. 95 (2009) 251905
[2] E. C. Young, C. S. Gallinat, A. E. Romanov, A. Tyagi, F. Wu, and J. S. Speck, Appl. Phys. Expr. 3 (2010) 111002
10:15 AM - D1.3
Indium Rich III-Nitrides on Germanium by Molecular Beam Epitaxy.
Ruben Lieten 1 2 , Wei-Jhih Tseng 2 , Maarten Leys 2 , Jean-Pierre Locquet 1 , Johan Dekoster 2
1 Physics and Astronomy, K.U. Leuven, Heverlee Belgium, 2 , IMEC, Heverlee Belgium
Show AbstractGroup III-nitrides show physical properties interesting to many electronic and optoelectronic devices. Nitride materials are predominantly grown by heteroepitaxy on foreign substrates. Sapphire, SiC, and Si substrates are most commonly used for heteroepitaxial growth of nitrides. We have shown that heteroepitaxial growth of GaN on Ge(111) by plasma assisted MBE (PAMBE) can lead to GaN layers of high crystal quality. The use of Ge substrates for III-Nitrides growth, and in particular InGaN, could be advantageous for devices in which vertical conduction is required. A direct photo electrolysis cell, using photocurrent to split H2O into H2 and O2, is an example of a device which would benefit from vertical conduction. Using a back contact can much simplify such a design. Another promising application of InGaN on conducting Ge substrates is a high-efficiency solar cell. InGaN can absorb the UV and visible part of the solar spectrum and germanium the infrared part.Previous structural characterization revealed that the GaN layers grown on germanium consisted of misoriented domains. The formation of rotated domains of GaN on Ge can be suppressed by enhancing step flow growth with respect to 2D nucleation. As a GaN nucleation layer has a higher bandgap than InGaN, the thickness should be kept to a minimum. We have chosen for 50 nm of this improved GaN buffer layer for subsequent growth of InGaN with layer thickness of ~0.8 μm. Thinner GaN buffer layers can be used, however complicates characterization by X-ray diffraction (XRD).The influence of growth parameters on the crystal quality and composition of InGaN layers on germanium substrates has been investigated by XRD measurements. At a fixed Ga and N flux, the In flux was incrementally increased. For both nitrogen as metal (Ga+In) rich growth conditions the indium incorporation increases for increasing indium flux. Only for metal rich growth conditions segregation of metallic indium is observed. An optimum in crystal quality is obtained for a Ga+In to N flux ratio close to unity.For 14 and 20 % indium composition, the symmetric (0002) ω XRD full-width at half- maximum (FWHM) is around 0.5 ° (1800 arc sec). An optical bandgap of 2.7 eV (459 nm) was deduced from spectroscopic ellipsometry for the sample with 20 % indium. SIMS measurements are used to investigate the background impurity concentration and the In composition in function of depth. Photoluminescence studies are in progress.Additionally we have grown InGaN layers by MBE on GaN-on-sapphire templates, grown by metal organic vapour phase epitaxy (MOVPE). These layers are used as comparison for the layers deposited on germanium.
10:30 AM - D1.4
Crack-free III-nitride Structures (> 3.5 µm) on Silicon.
Mihir Tungare 1 , Jeffrey Leathersich 1 , Neeraj Tripathi 1 , Puneet Suvarna 1 , Yudhishthir Kandel 1 , Fatemeh (Shadi) Shahedipour-Sandvik 1
1 College of Nanoscale Science and Engineering, University at Albany, SUNY, Albany, New York, United States
Show AbstractIII-nitride structures on Si are of great technological importance due to the availability of large area, epi ready Si substrates and the ability to heterointegrate with mature silicon micro and nanoelectronics. High voltage, high power density, and high frequency attributes of GaN make the III-N on Si platform the most promising technology for next-generation power devices. However, the large lattice and thermal mismatch between GaN and Si (111) introduces a large density of dislocations and cracks in the epilayer. Cracking occurs along three equivalent {1-100} planes which limits the useable device area. Hence, efforts to obtain crack-free GaN on Si have been put forth with the most commonly reported technique being the insertion of low temperature (LT) AlN interlayers. However, these layers tend to further degrade the quality of the devices due to the poor quality of films grown at a lower temperature using metal-organic chemical vapor deposition (MOCVD). Our substrate engineering technique shows a considerable improvement in the quality of 2 µm thick GaN on Si (111), with a simultaneous decrease in dislocations and cracks. Dislocation reduction by an order of magnitude and crack separation of > 1 mm has been achieved. Here we combine our method with step-graded AlGaN layers and LT AlN interlayers to obtain crack-free structures greater than 3.5 µm on 2” Si (111) substrates. A comparison of these film stacks before and after substrate engineering is done using atomic force microscopy (AFM), x-ray diffraction (XRD) and optical microscopy. Although there is degradation in the quality of GaN upon the insertion of LT AlN interlayers, this degradation is less prominent in the stack grown on the engineered substrates. Also, this methodology enables a crack-free surface with the capability of growing thicker layers. High electron mobility transistor (HEMT) and Schottky device structures are developed on Si using the above approach.
10:45 AM - D1.5
Unraveling the Role of Defects in InGaN Luminescence with High Ppressure Spectroscopy.
Marius Millot 1 , Kin Man Yu 2 , Lothar Reichertz 3 , Iulian Gherasoiu 3 , Todd Williamson 4 , Joshua Williams 4 , Mark Hoffbauer 4 , Wladek Walukiewicz 2 , Raymond Jeanloz 1
1 Department of Earth and Planetary Science, University of California Berkeley, Berkeley, California, United States, 2 Materials Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, California, United States, 3 , RoseStreet Labs Energy, Phoenix, Arizona, United States, 4 Chemistry Division, Los Alamos National Laboratory, Los Alamos, New Mexico, United States
Show Abstract
InxGa1-xN alloys with the direct gap tunable from 0.7 eV in InN to 3.4 eV in GaN have a great potential for full solar spectrum solar cells and broad spectral range light emitting devices[1].
However, the synthesis of InGaN crystals remains challenging due to a large difference between InN and GaN optimum growth temperatures. In addition, and partly as a consequence of the difficulties in
maintaining uniform film composition as well as stoichiometry and good crystalline properties the photoluminescence (PL) efficiency of these InGaN alloys is found to drop dramatically when the Indium concentration x
approaches 0.3-0.4 and the bandgap lies in the green-red region [2]. It is important to note that at the composition x=0.35 the conduction band of the InGaN alloys falls close to the Fermi level stabilization energy (EFS)
which represents an average energy of highly localized intrinsic defects. Localized defects level which crosses the InGaN conduction band edge may act as nonradiative recombination centers and play a major role in the observed
quenching of the PL efficiency and thus also the efficiency of light emitting devices. In order to understand the role of native defects we studied effects of high hydrostatic pressure on optical properties of InGaN alloys.
High hydrostatic pressures up to 15-20 GPa are well known to induce (i) a large (20-70) meV/GPa blue shift of the conduction band minimum in Gamma (ii) a small (10-30) meV/GPa red shift of the valence band maximum in most of the
common semiconductors but has a very small influence on a localized defect energy [3]. Pressure dependent optical spectroscopy appears then a perfect tool to investigate the correlation between the relative position of the defect
line with respect to the conduction band and the PL intensity.We will present and discuss the results of a comprehensive study by means of PL and absorption spectroscopies on several InxGa1-xN crystals with
0.25 < x <0.45 i.e. close to the compositions at which the conduction band crosses EFS. We will focus in particular on the PL intensity evolution exhibiting in some case an unusual pressure induced enhancement.
Altogether this study yields a better insight on the role of the defects in the InGaN luminescence properties.[1] J. Wu, W. Walukiewicz, K.M. Yu, J.W. Ager et. al., Appl. Phys. Lett. 80, 3967 (2002).[2] J. Phillip, M. Coltrin,
M. Crawford, A. Fischer, M. Krames, R. Mueller-Mach, G. Mueller, Y. Ohno, L. Rohwer, J. Simmons, and J. Tsao, Laser & Photonics Rev. 1, 307 (2007).[3] A. R. Goni, and K. Syassen in Semiconductors and Semimetals Vol. 54, ed.
by T. Suski, and W. Paul (Academic, New York 1998), pp. 247-425.
D2: Modeling Defects
Session Chairs
Tuesday PM, April 26, 2011
Room 2003 (Moscone West)
11:30 AM - **D2.1
Exploiting Nanostructure-based Scattering Effects in High-efficiency Photovoltaic Devices.
Edward Yu 1 , Claiborne McPheeters 1 , Daniel Derkacs 1 , Swee Lim 1
1 Microelectronics Research Center, University of Texas at Austin, Austin, Texas, United States
Show AbstractIntegration of metal and dielectric nanostructures with semiconductors offers new opportunities for engineering photovoltaic devices by influencing the transmission and propagation of light within the device. We discuss device structures in which plasmonic and related scattering effects are exploited to enable efficient coupling of incident photons into optically confined modes of semiconductor photovoltaic devices, including quantum-well, quantum-dot, and related solar cell structures. Such devices have been predicted to be capable of providing power conversion efficiencies substantially higher than those possible with single pn homojunction solar cells, but present major challenges in achieving simultaneously high efficiency in photon absorption and photogenerated carrier collection. We discuss initial approaches in which we exploit the elevated refractive index within the multiple-quantum-well region of a quantum-well solar cell to enable coupling of incident light into lateral propagation paths for which efficient optical absorption can occur even in very thin multiple-quantum-well layers. We then describe work on structures in which substrate removal processes are used to produce ultrathin solar cells with metallic nanostructures fabricated on the rear of the device, enabling highly efficient scattering of long-wavelength photons into guided optical modes within the device. Optimization of device designs, fabrication processes, and device characterization will be described.
12:00 PM - D2.2
Defects and Doping in Nitrides: Effects on Electrical and Optical Properties.
Anderson Janotti 1 , John Lyons 1 , Qimin Yan 1 , Luke Gordon 1 , Chris Van de Walle 1
1 , UCSB, Santa Barbara, California, United States
Show AbstractNitride semiconductors are currently used in commercially available light-emitting diodes and laser diodes that operate in the blue-violet spectral region, as well as in high-power, high electron mobility transistors. Still, the role of some defects and their interaction with common impurities, incorporated during growth or processing, are not fully understood. Using cutting-edge computational methods we investigate the electronic and structural properties of hydrogen and oxygen impurities, and their interaction with cation and anion vacancies in GaN, AlN, and InN. Previously published studies of defects and impurities in nitrides were mainly based on density functional theory within the local density approximation, and hence carry significant uncertainties due to the large band-gap error that is inherent to this approach. Here we overcome this problem by using a hybrid functional method that allows for an accurate description of band gaps and structural properties. Using this method we calculate formation energies as a function of Fermi-level position and atomic chemical potentials, as well as configuration coordinate diagrams, allowing us to extract thermodynamic transition levels and optical excitation and emission energies. We will discuss the role of cation and anion vacancies in the optical properties of AlN and GaN, and the effects of hydrogen and oxygen impurities on the electrical and optical properties of GaN, AlN, and InN. Emphasis is placed on relating the calculated properties to the available experimental data. This work was supported by NSF under award No. DMR-0906805 and by the UCSB Solid State Lighting and Energy Center.
12:15 PM - D2.3
Shallow or Deep Character of Acceptor Impurities in Nitride Semiconductors.
John Lyons 1 , Anderson Janotti 1 , Chris Van de Walle 1
1 , UCSB, Goleta, California, United States
Show AbstractControl over conductivity has proven to be one of the greatest challenges in the development of the nitride semiconductors. Most as-grown nitrides exhibit unintentional n-type conductivity, and doping efforts have focused on lowering the Fermi level to create semi-insulating or p-type material. Magnesium and carbon are two of the most widely used acceptor dopants in GaN. Magnesium is the only proven effective p-type dopant, and is thus essential for all optoelectronic devices. Carbon is used to create high-resistivity layers for use in high electron mobility transistors. In addition, carbon is often unintentionally incorporated, particularly in metal-organic chemical vapor deposition (MOCVD). Despite the importance of these impurities and the considerable research into their properties, many open issues remain regarding their electrical and optical behavior. Carbon is often assumed to behave as a shallow acceptor when incorporating on the nitrogen site in GaN, but this has never been established in electrical measurements, and carbon doping has not led to p-type conductivity. Carbon has also long been associated with the frequently observed yellow luminescence (YL) in GaN; however, no credible carbon-related configuration was ever suggested that could explain the YL. We have used cutting-edge first-principles calculations based on density functional theory using hybrid functionals to elucidate both the electrical and optical behavior. We find that carbon on a nitrogen site is a very deep rather than a shallow acceptor, with an ionization energy of 0.9 eV. This immediately explains the semi-insulating behavior of carbon-doped layers, without the need to invoke compensation. Our calculated configuration-coordinate diagram also shows that this substitutional carbon atom, by itself, gives rise to YL [1].Magnesium doping of nitride semiconductors consistently leads to p-type behavior, though a post-growth annealing step is often required to break Mg-H complexes. While the electrical properties of Mg-doped GaN are seemingly well established, its optical properties are still under debate. Photoluminescence studies have shown that emission signals depend strongly on growth method, Mg concentration, and the thermal history of the sample. Recent electron paramagnetic resonance work has also suggested that there are multiple forms of the magnesium acceptor responsible for the observed near-band-edge emission signals. Our hybrid functional calculations provide a consistent explanation for this rather bewildering array of experimental observations.[1] J. L. Lyons, A. Janotti, and C. G. Van de Walle, Appl. Phys. Lett. 97, 152108 (2010). This work was supported by the NSF under Award No. DMR-0906805 and by the UCSB Solid State Lighting and Energy Center.
12:30 PM - D2.4
Role of Nitrogen Vacancies and Related Complexes in Compensation and Luminescence of Mg-doped GaN.
Qimin Yan 1 , Anderson Janotti 1 , Matthias Scheffler 1 2 , Chris Van de Walle 1
1 Materials Department, University of California, Santa Barbara, California, United States, 2 , Fritz-Haber-Institut der Max-Planck-Gesellschaft, Berlin Germany
Show AbstractNitrogen vacancies act as compensating acceptors in GaN, and Mg-doped GaN will therefore contain a certain concentration of vacancies that reduce the conductivity and may lead to optical absorption or luminescence lines. In addition to isolated vacancies, complexes between Mg and nitrogen vacancies may also be formed. Using first-principles calculations with the hybrid functional approach, we investigate the effects of nitrogen vacancies and related complexes on the electrical and optical properties of Mg-doped GaN. By calculating the formation energies of vacancies and Mg-vacancy complexes we obtain information about their expected concentration, as well as about stable charge states and defect levels. The 3+ state of the nitrogen vacancy and the 2+ state of the complex are found to be most stable when the Fermi level is located near the valence-band maximum (VBM). Our calculations also enable us to study the role of these defects in luminescence. The red (1.8 eV) photoluminescence (PL) band often observed in Mg-doped GaN has been suggested to be due to a recombination process involving vacancy-related deep defects. Based on indirect experimental evidence, vacancy-dopant complexes (including MgGa-VN) have been proposed as origins of deep defects that are responsible for the red luminescence [1,2]. We investigate the optical absorption and emission energies by calculating the configuration coordinate diagram for the vacancy and for the MgGa-VN complex. The emission, in which an electron in the conduction band is transferred to (MgGa-VN)2+, resulting in (MgGa-VN)+, peaks at 1.81 eV, with a Franck–Condon shift of 0.80 eV. Our calculated emission lines thus indicate that MgGa-VN is a likely source for the red luminescence that is often observed in Mg-doped GaN. This work was supported by the UCSB Solid State Lighting and EnergyCenter and by the Center for Energy Effcient Materials, an Energy Frontier Research Center funded by the U.S. DOE-BES under Award No. DE-SC0001009.[1] D. M. Hofmann et al., phys. stat. sol. (a) 180, 261 (2000).[2] S. Zeng et al.., Appl. Phys. Lett. 89, 022107 (2006).
12:45 PM - D2.5
The Role of Copper Interstitials in CIGS: New Insights from First-principles Calculations.
Johan Pohl 1 , Karsten Albe 1
1 Institute for Materials Science, TU Darmstadt, Darmstadt Germany
Show AbstractCopper migration is an important process during CIGS solar cell production. It has been proposed in the literature that copper migration may explain structural changes at the buffer-absorber interface. Moreover, it is also considered to be responsible for some of the observed meta-stabilities in CIGS. The atomistic mechanisms of copper diffusion in CIGS, however, are still not well understood and experiments have shown widely varying copper diffusion coefficients. In this contribution, we present calculations based on density functional theory of the formation and migration energies of Cu defects and Frenkel pairs in CIGS. Our results reveal that not only the vacancy, but also the copper interstitials exhibit sufficiently low formation enthalpies in order to be considered as an important defect in CIGS. While the copper vacancies show comparatively slow migration, the interstitial diffusion is much faster because of the particularly low migration barrier. We finally discuss the implications of these results for interpreting the experimental findings.
D3: Nano and Colloidal Materials
Session Chairs
Tuesday PM, April 26, 2011
Room 2003 (Moscone West)
2:30 PM - **D3.1
Theory of Bandstructure and Defect Properties in Compound Semiconductors for Energy Applications.
Stephan Lany 1 , K. Biswas 1 , J. Vidal 1 , A. Zunger 1
1 , National Renewable Energy Laboratory, Golden, Colorado, United States
Show AbstractThe functioning of semiconductor devices in energy applications is controlled in a convoluted and complex way by a variety of different materials properties, including band-structure properties (band gap, absorption, effective-masses), defect properties (doping, carrier traps, recombination centers), and surface/interface properties (grain boundaries, heterojunctions). In experiment, one often sees only the "end result" with little hint at what is weak link that determines the performance. Here, theory can help to is able to separate the effects of the many, otherwise interdependent variables, thus enabling better control as well as design of systematic improvements. A recent research focus is the search for photovoltaic materials based on earth-abundant elements. Many of the proposed materials contain multivalent elements, like, for example, the In-free Cu2ZnSnS4 PV absorber, where Sn can exist in two oxidation states, +IV and +II. On the other hand it is well known that multivalent transition metal impurities in semiconductors usually create deep levels inside the band gap that are associated with changes in the oxidation state. This raises the possibility that the performance of such materials that contain multivalent elements as part of their structural skeleton may be affected changes of the oxidation state in response to charge-altering perturbations such as illumination or doping. We show here that detrimental deep levels in the gap can occur due to multivalency of Sn, and provide guidance how to circumvent this potential limitation [1]. We further present recent theory results for the band structure and defect properties in the related compound SnS, which is recently receiving attention as an alternative PV absorber material. Even in more conventional materials, which are in many regards well characterized experimentally, there remain puzzling observations on some properties that call for theoretical clarification. A striking example is the nature of acceptor dopants in wide gap semiconductors like GaN or ZnO where one can observe either shallow or deep behavior depending on how the experiment is performed. We find that the metal-site acceptors Mg, Be, and Zn in GaN and Li in ZnO bind holes in deep levels that are largely localized at single anion ligand atoms. In addition to this deep ground state, we observe an effective-masslike delocalized state that can exist as a short lived shallow transient state. The Mg dopant in GaN represents the unique case where the ionization energy of the localized deep level exceeds only slightly that of the shallow effective-mass acceptor, which explains why Mg works so exceptionally well as an acceptor dopant [2].This work was funded by the U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy under Contract No. DE-AC36-08GO28308 to the National Renewable Energy Laboratory. [1] K. Biswas, S. Lany, A. Zunger, Appl. Phys. Lett. 96, 201902 (2010).[2] S. Lany, A. Zunger, Appl. Phys. Lett. 96, 142114 (2010).
3:00 PM - D3.2
Top-down Approach in Preparing Nanostructured Template for III-V Solar Cell.
Yijie Huo 1 , Anjia Gu 1 , Dong Liang 1 , Erik Garnett 2 , Evan Pickett 1 , Jia Zhu 2 , Ching-Mei Hsu 2 , Shu Hu 2 , Shuang Li 1 , Paul McIntyre 2 , Yi Cui 2 , James Harris 1
1 Solid State and Photonics Laboratory, Stanford University, Stanford, California, United States, 2 Department of Materials Science & Engineering, Stanford University, Stanford, California, United States
Show AbstractNanostructures have been proposed as a novel means to produce high-efficiency and low-cost solar cells. Up to now, the most popular way to produce a nanostructured III-V solar cell is by bottom-up approach, e.g. the Au-assisted vapor- liquid-solid (VLS) growth of GaAs nanowires (NWs). Vertically oriented GaAs NWs have been reported using this method. One of the shortcomings of this approach for solar cell devices is low coverage and poor uniformity; and the other drawback is the use of gold as catalyst because gold creates deep level defect which is a life time killer in semiconductor devices. Here we report a more practical method of top-down approach to prepare nano-patterned template for fabricating nanostructured III-V solar cell, to improve the coverage and uniformly and to avoid using any metal catalyst. We developed a top-down process, e.g. plasma enhanced dry etching to pattern a GaAs wafer by using monolayer SiO2 nano-spheres as an artificial mask to achieve large wafer-scale coverage. We obtained nano-pyramids with clear facets on GaAs (100) substrate. We use wet etching as a final surface cleaning, damage removal and pyramid shaping step before growth. The final coverage rate is above 95%, and uniformity of the covered areas is significantly improved. Each nanopyramid is identical and periodically arranged. Both the density, fill factor, pitch of the nanopyramid arrays and height, width, shape of each nanopyramid can be well controlled. These nano-pyramids have superior optical properties compared to their nanowire counterpart, and are more favorable for multijunction configuration. These also provide better templates for further expitaxial growth of III-V materials due to the clear facets while the nanowire or nanocone structures have many irregular and randomly exposed facets that can be problematic for epitaxy. Based on our previous work of growing single-crystal GaAs conformally on Ge nanostructures, we successfully grew high-quality single-crystal III-V solar cell structures on these nanostructured GaAs substrates. I-V characteristics and photovoltaic conversion will be presented at the meeting.
3:15 PM - D3.3
Controlled Synthesis, Characterization and Growth Mechanism of CuInS2 Nanoparticles for Organic-inorganic Hybrid Solar Cells.
Marta Kruszynska 1 , Holger Borchert 1 , Juergen Parisi 1 , Joanna Kolny-Olesiak 1
1 Department of Physics, University of Oldenburg, Oldenburg Germany
Show AbstractSemiconductor nanoparticles have found various technological applications due to their size and shape dependent properties. The distinct and well-defined morphology of these inorganic nanoparticles plays an important role to determine their physical and chemical characteristics. One class of semiconductors very well studied in the bulk, but having only recently attracted interest as nanocrystalline materials is the I-III-VI2 family, comprising CuInS2 (CIS). This material is a direct band gap semiconductor (1.45 eV), having a high absorption coefficient and photoconductivity. The absorption and emission of CuInS2 nanocrystals spans the visible and near IR range, which makes them interesting as a potential electron accepting material for polymer-based hybrid solar cells. Up to now, there are only few reports on polymer/CuInS2 nanoparticle devices.In this work, we present a shape-controlled synthesis of Cu2S-CuInS2 and CuInS2 nanocrystals by a reaction of copper (I) acetate and indium (III) acetate with tert-dodecanethiol as a source of sulfur, and trioctylphosphine oxide, oleylamine and 1-dodecanethiol as ligands. By changing the reaction parameters various shapes (triangular, nanorods, hexagonal, and P-shape) of nanoparticles can be obtained. The final materials were studied by transmission electron microscopy (TEM), powder X-ray diffraction (XRD), energy dispersive X-ray analysis (EDX), and UV-vis absorption spectroscopy. It is found that a copper sulfide phase plays a crucial role in controlling the shape of the CIS nanocrystals. Next, we fabricated photovoltaic devices based on colloidally prepared CIS nanocrystals and regioregular poly(3-hexylthiophene) (P3HT) as an electron donor material. Moreover, we modified the surface of the nanocrystals with various organic compounds, and investigated the impact on charge separation in the CIS/P3HT blends by photo-induced absorption spectroscopy (PIA).
3:30 PM - D3.4
Axially Anisotropic CdSxSe1-x Nanorods: Fabrication and Controlled Metal Photo-deposition.
Javier Vela 1 2 , Purnima Ruberu 1 2
1 Chemistry, Iowa State University, Ames, Iowa, United States, 2 , US DOE Ames Laboratory, Ames, Iowa, United States
Show AbstractOne-dimensional semiconductor and semiconductor-metal nanorods are capable of harvesting solar energy, converting it to potential energy via charge-separation and subsequently to chemical energy. They become redox-active upon illumination and remain active after several-hour dark-storage (Costi et al. NanoLett. 2008, 8, 637). We have found reproducible and controllable routes for making colloidal CdS1-xSex nanorods and their metal hybrids. The nanorods are axially anisotropic, containing a thick-CdSe-rich head and a slim-CdS-rich tail. We observe a strong dependence between the site-selectivity of metal-photo-deposition and the irradiance profile used. The irradiance profile controls whether photo-initiated reduction occurs via homogeneous or heterogeneous-nucleation, leading to free-solution-particles vs. nanorod-surface-bound particles, respectively. The irradiance profile also controls whether metal-photodeposition occurs at the nanorods CdSe-rich head or CdS-rich tail. Our new routes use solution-phase organometallic-precursors and widely available fluorescent-lamps, opening new avenues for greater synthetic control and larger throughput of these materials for their fundamental study and application.
3:45 PM - D3.5
Broadening the Scope of Applications for Colloidal GaAs Nanocrystals.
Jannika Lauth 1 , Tim Strupeit 1 , Andreas Kornowski 1 , Horst Weller 1
1 Institute for Physical Chemistry, University of Hamburg, Hamburg Germany
Show AbstractIII-V semiconductor compounds can be classified as basic materials for modern electronic applications like optoelectronic devices and solar cells. GaAs in this manner is one of the most important direct bandgap semiconductors[1][2]. With a bandgap of 1.42 eV it is widely used in infrared LEDs and LASERs[3]. It has a higher electron saturation velocity and higher electron mobilities leading to uprate operation frequencies above 250 GHz of devices. Due to this properties GaAs finds applications in satellite commmunication and radar systems and has some major advantages to commonly used Si in high efficiency solar cells. The possibility to synthesize III-V semiconductor nanoparticles and nanostructures which are highly crystalline, monodisperse and readily to obtain by the use of wet chemical and cost-effective methods still is a major synthetical problem. The covalent bond share in the atomic lattice of III-V semiconductors causes problems in applicable syntheses of these compounds. It is not possible to use ionic precursor reactions like in well established II-VI nanocrystal syntheses. The few existing ways to obtain GaAs nanostructures are mainly based on the dehalosylation reaction of a metal salt with tris(trimethyl)silylarsine in a high boiling solvent, but the GaAs structures obtained are neither continously crystalline nor monodisperse[4]. We present for the first time the synthesis of colloidal GaAs nanocrystals which are highly crystalline and have a narrow size distribution by the use of wet chemical methods. Different gallium(III) halides acting as precursors are reduced to elementary gallium clusters by n-butyllithium and react with magnesium arsenide to form GaAs crystals. Elongated GaAs structures such as needles and wires are formed by the use of triethylarsine as arsenic precursor. Thin films of the different GaAs structures are spin coated and first efforts on transport measurements on the material are presented.[1]Bar-Lev, A., Semiconductors and Electronic Devices, Prentice Hall, New York, 1984[2]Bosi, M; Pelosi, C., Prog. Photovolt: Res. Appl., 15, 51-68, 2007[3]Katz, A.,Indium phosphide and related materials: processing, technology and devices, Artech House Publishers, Fitchburg, USA, 1992[4]Wells, R.L, Chem. Mater., 7, 793-800, 1995
D5: Poster Session: Alternative III-V Materials
Session Chairs
Tuesday PM, April 26, 2011
Exhibition Hall (Moscone West)
6:00 PM - D5.2
Modeling and Simulation of Hydrogen Diffusion and Impurity Passivation in n- and p-doped III-V Semiconductor Photonic Materials.
Joshua Levinson 1 , Kenneth Glogovsky 2 , Prajesh Adhikari 1
1 Chemical & Biomolecular Engineering, Lafayette College, Easton, Pennsylvania, United States, 2 , Cyoptics, Inc., Breinigsville, Pennsylvania, United States
Show AbstractHydrogen has been identified as an effective passivating species for n- and p-type III-V semiconductor photonic materials. Briefly, upon exposure to hydrogen plasma, hydrogen atoms penetrate the surface of a semiconductor substrate and diffuse through the various thin-film layers. These atoms then react with n- and/or p-type impurities to form a passivated complex, thereby causing electrical or photonic isolation of these regions. As this technique is presently used in photonic device fabrication, it is desired to have a robust, predictive reaction-diffusion model to more efficiently develop passivation techniques with respect to time and cost. The purpose of this work is to develop a fundamental model of the relevant kinetic and transport mechanisms involved in the passivation of photonic materials (with particular emphasis on Zn-doped, p-type InP), to create a predictive computer simulation, and to validate it against experimental data. To meet these objectives, a theoretical model was developed that incorporates concentration-based and potential field-enhanced diffusion of hydrogen in n- and p-type materials while also considering the passivation reactions between the diffusing hydrogen and photonically active impurities. Experiments have been performed to follow the diffusion and/or reaction of deuterium (as a model for hydrogen) in undoped and Zn-doped InP substrates and heterostructures, with deuterium and impurity concentrations determined using SIMS and with the optoelectonic activity of Zn determined using Polaron measurements. This data has been used to determine key parameter values included in the model as a function of temperature, which are subsequently used to predict diffusion-passivation profiles in other thin-film substrate configurations. Computational results have shown excellent agreement between experiment and theory, predicting the passivation profiles and penetration depths for various heterostructures. Present work is focused on expanding the simulation to another computational platform, with development of n-type, p-type, and merged n/p-type versions to address different impurities and their positions in the crystal structure of the substrate. Future work will seek to validate these simulations and to more fully explore the behavior of this reaction-diffusion system under a wider array of thin-film structures.
D8: Poster Session: Nano-Materials and Devices
Session Chairs
Tuesday PM, April 26, 2011
Exhibition Hall (Moscone West)
6:00 PM - D8.1
Nanoheteroepitaxy of III-nitride Nanopyramid LEDs by OMVPE.
Isaac Wildeson 1 3 , David Ewoldt 2 3 , Robert Colby 2 3 , Zhiwen Liang 2 3 , Eric Stach 4 2 3 , Timothy Sands 1 2 3 , R. Edwin Garcia 2 3
1 School of Electrical and Computer Engineering, Purdue University, West Lafayette, Indiana, United States, 3 Birck Nanotechnology Center, Purdue University, West Lafayette, Indiana, United States, 2 School of Materials Engineering, Purdue University, West Lafayette, Indiana, United States, 4 , Brookhaven National Laboratory, Upton , New York, United States
Show AbstractOf the current research challenges in the area of light emitting diodes (LEDs), creating an efficient green LED is among the most pertinent for wide-spread solid state lighting adoption. Today, efficient blue and red LEDs can be fabricated from the III-nitride and III-phosphide material systems, respectively. The addition of an efficient green LED would allow for practical RGB warm white light generation of high color rendering without the use of phosphor down-conversion. The III-nitride material system is the leading candidate for realizing efficient green LEDs, however, there are numerous challenges specific to the III-nitrides that are currently limiting efficiency. Such challenges include the lattice mismatch between (In,Ga)N quantum wells and GaN cladding layers, high dislocation densities deriving from non-ideal substrates, a miscibility gap between GaN and InN, and strong polarization induced electric fields within the quantum well that decrease internal quantum efficiency. We address these challenges with the use of nanoheteroepitaxy. Specifically, nanopatterned dielectric growth masks are used to selectively grow nanopyramid heterostructures via organometallic vapor phase epitaxy [1]. The dielectric growth mask directly controls the dimensions of the nanopyramids and affectively filters threading dislocations, as confirmed by transmission electron microscopy (TEM). Dielectric growth masks are patterned through the use of porous anodic alumina as a reactive ion etch mask and by E-beam lithography, where greater control/uniformity over nanopyramid diameter and spacing can be achieved. Each nanopyramid is outlined by six semipolar {1-101} planes, on which (In,Ga)N quantum wells are grown. Following the nanopyramid heterostructure growth, high temperature p-GaN epitaxy is performed to complete the LED structure. Finite element method (FEM) simulation shows lower polarization-induced electric fields within nanopyramid heterostructures as compared to quantum wells grown on thin-film c-plane GaN, thus providing greater overlap of the electron and hole wave functions in the nanopyramid devices. Furthermore, electroluminescence results corroborate with simulated data indicating that strain relaxation within the nanopyramids allows for higher InN incorporation before the introduction of misfit dislocations, as compared to a thin film. We will discuss in detail the nanopyramid LED fabrication processes, TEM analysis of nanopyramid structure, optical properties of nanopyramid heterostructures, and FEM simulation results on polarization (including both spontaneous and piezoelectric) within the nanopyramid heterostructures. For practical comparison, similar analysis will be conducted on thin-film c-plane heterostructures for many of the results.[1]. I. Wildeson et al., J. Appl. Phys. 108 (2010) 044303.
6:00 PM - D8.2
Comparison of Aluminum Nitride Nanowire Growth with and without Catalysts via Chemical Vapor Deposition.
Kasif Teker 1 , Joseph Oxenham 1
1 Physics and Engineering, Frostburg State University, Frostburg, Maryland, United States
Show AbstractOne-dimensional aluminum nitride (AlN) nanowires are important not only for understanding fundamental concepts underlying the observed electronic, optical, and mechanical properties of materials, but also for the superior potential applications in many fields including power transistors, heat sinks, surface acoustic wave filters, resonators, sensors, and piezoelectric nanogenerators. Free-standing catalyst-free nanowire films provide great opportunities due to enabling easy transfer of these into any substrate for the development of new devices for many technological applications including piezoelectric electricity generators.This paper presents a systematic investigation of AlN nanowire synthesis by chemical vapor deposition using Al, NH3, and a mixture of Al-Al2O3 as source materials on SiO2/Si substrate. A wide variety of catalyst materials, in both discrete nanoparticle and thin film forms, have been used (Co, Au, Ni, and Fe). The growth runs have been carried out at temperatures between 800 and 1100oC under two different carrier gases; H2 and Ar. It is interesting to note that the presence of H2 is required to achieve nanowire networks compared to Ar as carrier gas. The synergistic effect of hydrogen in synthesizing nanowires can be attributed to creation of reducing ambient at high temperatures. It was found that the most efficient catalyst was 20-nm Ni film. In fact, the AlN nanowire diameters are about 20-30 nm, about the same thickness as the Ni-film. Further studies with a mixture of Al and Al2O3 source materials has yielded in ultra-high density free-standing nanowire networks at 1100oC without any catalyst. It was observed that catalyst-free nanowires were significantly longer than that of with catalysts. The nanowires were polycrystalline as determined by x-ray diffraction. The analysis of the grown nanowires has been carried out by scanning electron microscopy, transmission electron microscopy, atomic force microscopy, Fourier transform infrared spectroscopy, and x-ray diffraction.
6:00 PM - D8.4
Synthesis and Characterization of Copper Iron Disulfide Nanoparticles.
Jay Yamanaga 1 , Xin Ai 2 , Teya Topuria 2 , Philip Rice 2 , John Bass 2 , Ho-Cheol Kim 2 , Campbell Scott 2 , Robert Miller 2 , Qing Song 2 , Gregory Young 1
1 Chemical and Materials Engineering, San Jose State University, San Jose, California, United States, 2 , IBM Almaden Research Center, San Jose, California, United States
Show AbstractIn this presentation, we report the successful solution-based synthesis of ternary compound semiconductor copper iron disulfide nanoparticles by using simple environment-friendly precursors and solvents. The crystallographic structure is confirmed by XRD to be tetragonal phase and the stoichiometric composition is confirmed by ICP-MS and RBS measurements. By controlling the reaction temperature and time, and using different precursors and solvent combinations, the size can be varied from 5 to 10 nm. Interestingly, the optical properties based on UV-Vis absorption spectroscopy are independent on the size of nanoparticles, presumably due to the lack of quantum confinement. In light of this, semiconductors composed of earth abundance and low-toxicity elements, such copper iron disulfide nanoparticles may have potential applications in PV.
Symposium Organizers
F. (Shadi) Shahedipour-Sandvik State University of New York-Albany
Kenneth Jones Army Research Lab – SEDD
L.Douglas Bell Jet Propulsion Laboratory
Blake Simpkins Naval Research Laboratory
Daniel Schaadt Karlsruhe Institute of Technology
Miguel Contreras National Renewable Energy Laboratory
Symposium Support
General Electric - Global Research
JLN Solar
D13: Solar Cells I: Arsenide/Antimonide Materials and Devices
Session Chairs
Wednesday PM, April 27, 2011
Room 2003 (Moscone West)
2:30 PM - **D13.1
Metamorphic III-V Multi-junction Solar Cells.
Tobias Roesener 1 , Wolfgang Guter 1 , Vera Klinger 1 , Jan Schoene 1 , Simon Philipps 1 , Rene Kellenbenz 1 , Marc Steiner 1 , Eduard Oliva 1 , Alexander Wekkeli 1 , Elke Welser 1 , Andreas Bett 1 , Frank Dimroth 1
1 , Fraunhofer Institute for Solar Energy Systems (ISE), Freiburg Germany
Show AbstractThe highest conversion efficiencies of sunlight into electricity are achieved with III-V multi-junction solar cells. The combination of a number of pn-junctions of different bandgap energy enables an efficient utilization of the solar spectrum. Presently, Ga0.50In0.50P/Ga0.99In0.01As/Ge triple-junction solar cells grown lattice-matched on Ge reach the highest conversion efficiencies of 41.6 % [1] under concentrated sunlight. But, the lattice-matched growth of the solar cells on substrates such as Ge or GaAs involves constraints which can be circumvented by changing the lattice constant in metamorphic structures. Metamorphic meaning the growth of completely relaxed but lattice-mismatched crystal layers on a common substrate.The metamorphic growth adds important flexibility to the choice of available band gap combinations. According to detailed-balance calculations a combination of Ga0.35In0.65P/Ga0.83In0.17As/Ge pn-junctions promises a relative 11 % higher efficiency (under AM1.5d spectral conditions, 298 K, 500 suns) compared to its previously mentioned lattice-matched counterpart. A difference in lattice constant of 1.2 % between Ge bottom and Ga0.83In0.17As middle cell has to be accomplished in this device structure. Graded buffer layers made of Ga1-xInxAs, Ga1-yInyP or AlGa1-xInxAs have therefore been developed at Fraunhofer ISE for more than a decade. It is important to relax the lattice within these buffer layers and simultaneously prevent the formation of threading dislocations. Based on the experience gained on metamorphic growth, Ga0.35In0.65P/Ga0.83In0.17As/Ge triple-junction solar cells were realized on an industrial AIX2600-G3 MOVPE reactor. The best solar cell devices yield excellent conversion efficiencies of up to 41.1 % under concentrated irradiation (AM1.5d, 298 K, 434 suns).Further research at Fraunhofer ISE is related to the metamorphic growth of III-V multi-junction solar cells on Si. Si is an advantageous substitution for Ge as substrate material. It offers lower mass density and cost, higher crystal hardness and thermal conductivity, larger wafer diameters and abundant availability. However, elaborate challenges arise from the metamorphic growth of III-V semiconductors on Si. The undesirable formation of anti-phase domains in polar III-V material on non polar Si is successfully controlled by nucleation of anti-phase domain free GaP layers. We have been able to demonstrate high quality GaP growth on Si with only 0.36 % lattice mismatch. The remaining 4.1 % difference in lattice constant between Si and GaAs are overcome by metamorphic GaAsxP1-x or Ga1-yInyP growth. These crystal structures for the integration of III-V multi-junction solar cells on Si are currently developed on an industrial feasible AIXTRON 300 mm CRIUS closed coupled showerhead MOVPE reactor. Results of the material development are presented during the conference.[1] R.R. King, et al. Proceedings 24th EUPVSEC, 2009.
3:00 PM - D13.2
Molecular Beam Growth of Amorphous and Crystalline GaNAs and GaNBi Alloys for Solar Energy Conversion.
Sergei Novikov 1 , C. Foxon 1 , K. Yu 2 , A. Levander 2 3 , R. Broesler 2 , M. Hawkridge 2 , Z. Liliental-Weber 2 , O. Dubon 2 3 , J. Wu 2 3 , J. Denlinger 4 , I. Demchenko 4 , F. Luckert 5 , P. Edwards 5 , R. Martin 5 , W. Walukiewicz 2
1 School of Physics & Astronomy , University of Nottingham, Nottingham United Kingdom, 2 Materials Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, California, United States, 3 Department of Materials Science & Engineering, University of California, Berkeley, California, United States, 4 Advanced Light Source, Lawrence Berkeley National Laboratory, Berkeley, California, United States, 5 Department of Physics, Strathclyde University, Glasgow United Kingdom
Show AbstractIn order to effectively convert solar energy into electric power or any other form of usable energy we need to develop novel semiconductor materials, which will absorb over the full solar energy spectrum. Our research concentrates on highly mismatched alloys (HMAs) of GaN with different group V anions. In the current presentation we will discuss and compare details of the growth and properties of GaN1-xAsx and GaN1-xBix alloys.The electronic structure of the conduction and valence bands of alloys with anions of very different electronegativity can be described well by the band anticrossing (BAC) model. The BAC model predicts an energy gap ranging from 0.7eV to 3.4eV for GaN1-xAsx alloys. An even stronger modification of the band structures is anticipated for the GaN1-xBix alloys with extremely mismatched anions. The large band gap range and controllable conduction and valence bands of these HMAs make them promising materials for solar energy conversion devices. However, the synthesis of these GaN-based alloys is difficult due to the large differences between the anions.We have succeeded in growing GaN1-xAsx alloys over a large composition range x(0-0.8) by plasma-assisted molecular beam epitaxy (PA-MBE) on different substrates including sapphire, silicon, Pyrex and glass. The enhanced incorporation of As was achieved by growing the layers under N-rich conditions at extremely low growth temperatures. Although GaN1-xAsx alloys become amorphous for x>0.1, optical absorption measurements show a progressive shift of the optical band gap to lower energy, from ~3.4eV to ~0.8eV, with increasing As content. We observe a continuous downward shift of the conduction band minimum and a upward shift of the valence band maximum energies for amorphous GaN1-xAsx alloys with increasing As concentration. The results strongly suggest that amorphous GaN1-xAsx alloys have short-range order resembling random crystalline GaN1-xAsx alloys. The large band gap range of the amorphous GaN1-xAsx covers much of the solar spectrum making this material system a good candidate for full spectrum multi-junction solar cells. The amorphous nature of the GaN1-xAsx alloys is particularly advantageous since low cost substrates such as glass and Pyrex glass can be used for solar cell fabrication. In addition, we have also explored the substitution of N by Bi in GaN1-xBix alloys. Using PA-MBE at extremely low temperature (~100°C), we have synthesized amorphous GaN1-xBix alloys with compositions up to x~0.11. In contrast to the growth of GaN1-xAsx, for GaN1-xBix to incorporate Bi we need to use Ga-rich growth conditions. Compared to GaN1-xAsx a much more rapid decrease in the optical absorption edge to energies as low as ~1.2eV is observed for GaN1-xBix with x~0.11.
3:15 PM - D13.3
Flexible III-V Photovoltaic Films on Metal Substrates and Studies of Defect Mitigation Strategies.
Venkat Selvamanickam 1 , Senthil Sambandam 2 , Aarthi Sundaram 1 , Akhil Mehrotra 1 , Alex Freundlich 1
1 Mechanical Engineering, University of Houston, Houston, Texas, United States, 2 , SuperPower, Schenectady, New York, United States
Show AbstractRecently, a technique has been developed to fabricate virtual single-crystalline GaAs films on flexible nickel alloy substrates [1]. GaAs was epitaxially grown by molecular beam epitaxy (MBE) on virtual single-crystalline germanium films which was possible by means of templates created by ion beam assisted deposition (IBAD). IBAD is an enabling technique to create virtual single-crystalline films of MgO on flexible, polycrystalline metal substrates and epitaxial germanium films have been grown on these substrates by means of an intermediate layer with a fluorite structure, such as cerium oxide [2]. X-ray and optical measurements of the germanium films reveal properties that are comparable to that single crystal Ge. All layers up to the GaAs film were grown by reel-to-reel IBAD and magnetron sputtering. While excellent epitaxial growth has been achieved in GaAs on flexible metal substrates, the defect density of the films as measured by High Resolution X-ray Diffraction and etch pit experiments showed a high value of 5 * 10^8 per cm^2. Cross sectional transmission electron microscopy of the multilayer architecture showed concentration of threading dislocations near the germanium-ceria interface. The defect density was found decrease as the Ge films were made thicker. The defects appear to originate from the MgO layer presumably because of large lattice mismatches between the various layers. The defect density in GaAs was reduced by a factor of five by adding a step of in-situ deposition of Ge by MBE on the sputtered Ge prior to GaAs growth. A simple non-intrusive photoresist based lithographic process has been used to define high quality patterns on GaAs coated on flexible substrates in a single mesa etch and metal evaporation/resist lift-off. Pattern resolution of few microns with well-defined grid line of 30 µm has been realized on flexible templates. Solar cells of GaAs have been fabricated and defects caused by various mechanisms including diffusion of germanium have been observed. Outcomes of various strategies being pursued to mitigate defects in the III-V photovoltaic films on flexible substrates will be discussed in this presentation.[1] A. Freundlich et al. Proc. 35th IEEE PVSC, Honololu, Hawaii, (2010)[2] V. Selvamanickam et al., J. Cryst. Growth 311, 4553 (2009).
3:30 PM - D13.4
Misfit Dislocation and Strain Relaxation at GaSb/GaAs Interface Versus Substrate Surface Treatment.
Yi Wang 1 , Pierre Ruterana 1 , Ludovic Desplanque 2 , Xavier Wallart 2
1 CIMAP, UMR 6252 CNRS-ENSICAEN-CEA-UCBN, Caen France, 2 IEMN, UMR 8520 CNRS, 59652 Villeneuve d’Ascq France
Show AbstractMetamorphic epitaxy of high lattice-mismatch (7.8%-14.6%) Sb-based materials on GaAs is attracting much attention for potential applications in high-speed and low-power electronic and optoelectronic devices due to their unique band-structure alignments, small electron effective mass and high electron mobility. However, the stress and high density of defects due to the large lattice mismatch between the epitaxial layers and the commonly used GaAs substrate has until now plagued both the electrical and optical properties of the devices. Recently, the possibility to fabricate more efficient devices by a tight control of the dislocation densities (< 106 cm-2) at the substrate/epilayer interface has been reported. In this work, the influence of the substrate surface preparation on the arrangement of the misfit dislocations and the relaxation of GaSb are investigated. We analyze the threading dislocation density all along the hetero-structure using the plan view as well as cross sectional Transmission Electron Microscopy (TEM) observations. The atomic structures of the Lomer, isolated 60o and pairs of 60o dislocations at the interface are determined. The strain distribution as well as the core distribution of the Lomer and 60o dislocation pairs are analyzed by high resolution TEM combined with geometric phase analysis (GPA) method. Weak beam investigations were used to analyze the accommodation of lattice mismatch and the threading mechanism of dislocations in GaSb. The preliminary results indicate that the Sb-rich GaAs surface promotes a flat interface and compact cores of Lomer and 60o dislocation pairs that are more efficient for GaSb relaxation, thus reducing the threading dislocation density.Acknowledgment: This work is supported by the National Research Agency under project MOS35, No.: ANR-08-NANO-022.
3:45 PM - D13.5
Investigation of Defects Influencing Performance of Type-II Superlattice Based Infrared Detectors.
Serguei Maximenko 1 , Nabil Bassim 2 , Edward Aifer 2 , Eric Jackson 2 , Jill Nolde 2 , Chaffra Affouda 2 , Chadwick Canedy 2 , Igor Vurgaftman 2 , Jerry Meyer 2
1 , Global Defense Technology& Systems, Inc., Crofton, Maryland, United States, 2 , Naval Research Laboratory, Washington, District of Columbia, United States
Show AbstractType II InAs/GaInSb superlattices (T2SLs) are a promising material system for mid-, long- and very long-wave infrared detection, with theoretical performance limits well beyond those of incumbent technologies based on InSb and HgCdTe. Despite significant gains over the last five years, however, T2SL based infrared photodiodes still perform well below the predicted levels. Devices exhibit greater than expected dark currents associated with excess recombination due to the presence of defects created during material growth and processing. In this work we investigated the effects of MBE growth-related crystallographic defects on dark currents of mid infrared T2SL InAs/GaInSb devices. Electron beam induced current (EBIC) analysis was performed on processed T2SL photodiodes, in a specially modified scanning electron microscope (SEM) equipped with micro probes and a cryogenic sample stage. As the electron beam was scanned across a device, the EBIC signal was monitored to reveal and map defects via enhanced recombination activity. Correlation of the EBIC data with device I-V characteristics then showed that the presence of electrically-active defects always led to increased dark current. Optical and Atomic Force Microscopy revealed that the same electrically-active defects also directly correlate with the surface morphological defects known as “hillocks”. The nature and origin of the defects were identified by focused ion beam microscopy and transmission electron microscopy, which showed that the majority of the defects corresponding to surface hillocks originate at the GaSb substrate interface. It was also determined that electrically-active defects penetrate an entire device structure, while other types of defects tend to terminate prior to the active device area.
4:00 PM - D13: AAMD
BREAK
D14: Solar Cells II: CdTe/CdS
Session Chairs
Wednesday PM, April 27, 2011
Room 2003 (Moscone West)
4:30 PM - **D14.1
PV Technology and its Role in the Growing Solar Market.
Danielle Merfeld 1 , Todd Tolliver 1
1 , GE Global Research Center, Schenectady , New York, United States
Show AbstractEnergy demands around the world are growing at an incredible rate, likely to nearly double by 2030. Populations and economies are expanding around the globe, both in developed and developing economies. The standard of living is rising in many parts of the world through increased access to housing, clean water, education, medicine, transportation, entertainment, all of which is enabled through greater electrification. The 100+ year old power generation industry has been the backbone of innovation and industrialization and will continue to be vital to the future of our society and our planet. However there are many challenges and uncertainties need to be worked through to meet the needs.All energy technologies must rise to meet the needs of the future, not only existing generation methods but new ones as well. Of the new modalities, renewable energy can play a major role in helping supply electricity globally. Of all the renewable energy sources – hydro, biomass, geothermal, wind and solar (photovoltaics, PV) – PV has the most to gain in terms of technology development. Despite being under investigation for decades and having made significant strides in the last ten years, there is much more work to be done to make PV cost competitive with conventional energy generation without subsidies.Three areas of development are needed: materials/modules, balance of system (BOS) and integration. PV module cost of manufacture has dropped greatly in the last few years driving heavy adoption around the world. However, all module technology costs will need to fall even farther in order to allow PV systems to reach grid parity. Hence significant development in low-cost manufacturing and increased scale is required. Also, there are a number of opportunities at the system level to reduce cost as well, whether through more efficient and reliable inverters, smarter plant design and grid friendly features. Lastly, improved grid integration technology will enable widespread installation of PV. Given the intermittent nature of sunlight, system controls, monitoring and prediction are critical in making these systems a dependable power source for energy providers.In this presentation we outline GE’s approach to addressing the development needs of the PV industry to enable low-cost utility scale solutions for the power generation needs of the future.
5:00 PM - D14.2
Influence of Surface Preparation on Scanning Kelvin Probe Microscopy and Electron Backscattering Diffraction Analysis of Cross Sections of CdTe/CdS Solar Cells.
Helio Moutinho 1 , Ramesh Dhere 1 , Chun-Sheng Jiang 1 , Mowafak Al-Jassim 1
1 , National Renewable Energy Laboratory, Golden, Colorado, United States
Show AbstractScanning Kelvin Probe Microscopy (SKPM) provides measurements of the electrical potential distribution on the sample surface. When applied to cross sections of CdTe/CdS solar cells, it reveals the location of the pn junction, the distribution of the depletion region on the CdTe film, and allows for the study of the interdiffusion layer between CdTe and CdS. Electron Backscattering Diffraction (EBSD) reveals the crystalline structure of the CdTe film, showing small grains at the CdTe/CdS interface, and the subsequent columnar growth until the interface with the back contact. Furthermore, it provides information on crystalline orientation, misorientation between adjacent grains, and special boundaries, such as twins. Although these techniques measure different properties, they share few characteristics: they provide high spatial resolution data, which comes from regions close to the surface, making surface preparation a major requirement for meaningful measurements.EBSD is performed inside a scanning electron microscope, with the sample tilted by 70°, which requires a flat surface, to avoid surface features from shading diffracted electrons to reach the EBSD detector. We tried different ways to produce flat and good-quality samples of CdTe/CdS solar cells, and found that methods that work for plan view do not work for cross sections. Furthermore, shading caused by steps between the different layers can, in some cases, completely shade the CdTe film from the detector. We were able to develop a procedure using a combination of polishing and ion-beam milling that resulted in good-quality data with minimum shading effects. We will show examples of cross sections of CdTe/CdS devices, and the information that can be obtained with this technique.SKPM requires a flat cross section because the probe has a limited vertical movement range, and to avoid convolution between the topographic and SKPM signals. We prepared CdTe/CdS cross sections using polishing, and noticed that the theoretical and experimental data for an unbiased cell are different because of Fermi level pinning at the surface, which is enhanced by the defects created during the polishing process. We combined polishing with ion-beam milling to decrease the defect concentration, which reduced Fermi level pinning. However, ion-beam milling increases the roughness of the cross section. We studied several scenarios, to determine which condition result in the best SKPM data. In this work we will show examples of the distribution of the electrical potential and electric field inside different CdTe/CdS devices, polarized at different conditions, and how the surface condition affects the data.
5:15 PM - D14.3
CdSxTe1-x Alloying in CdS/CdTe Solar Cells and Its Effects on Device Characteristics.
Joel Duenow 1 , Ramesh Dhere 1 , Joel Pankow 1 , Darius Kuciauskas 1 , Timothy Gessert 1
1 , National Renewable Energy Laboratory, Golden, Colorado, United States
Show AbstractInterdiffusion of CdS and CdTe results in formation of a CdSxTe1-x layer during thin-film CdTe photovoltaic (PV) device fabrication. For PV devices deposited in the conventional superstrate structure, high-temperature processing steps such as the close-spaced sublimation (CSS) of CdTe and the post-deposition CdCl2 heat treatment contribute to formation of this alloy. The CdSxTe1-x layer is thought to be crucial in fabricating high-performance CdTe devices because it relieves strain at the CdS/CdTe interface that would otherwise exist due to the large lattice mismatch between these two materials. Of additional note, our previous work has indicated that the electrical junction is located in this interdiffused CdSxTe1-x region. There has, however, been limited research on the properties of CdSxTe1-x. Further understanding is essential to predict the role of this CdSxTe1-x layer in the operation of the CdS/CdTe device. In this study, we will investigate the alloy layer by depositing CdSxTe1-x films onto glass/SnO2:F/SnO2/CdS structures using different methods. In the first approach, we will deposit CdSxTe1-x films directly by RF magnetron sputtering using targets containing a range of compositions from 10 to 60 wt.% CdS. We will also co-evaporate the CdSxTe1-x alloy from CdTe and CdS sources to further vary the composition. Additionally, we will attempt deposition by CSS. The substrate temperature will be varied for each deposition method. We will also investigate the effects of thermal annealing and CdCl2 heat treatments on the alloy. We will use: (1) X-ray diffractometry to examine the structure of the CdSxTe1-x films; (2) spectrophotometry to obtain bandgap values of the alloy; (3) time-resolved photoluminescence to investigate minority-carrier lifetime; and (4) Auger electron spectroscopy to compare the stoichiometry of these directly deposited films to the CdSxTe1-x layers found in high-performance CdTe devices. Preliminary performance results for PV devices incorporating these engineered CdSxTe1-x films will also be presented.
5:30 PM - D14.4
Deposition of II-VI Buffer Layers for Applications to Thin Film Photovoltaics.
Jonathan Bakke 1 , Stacey Bent 1
1 Chemical Engineering, Stanford University, Stanford, California, United States
Show AbstractThe pn junction for thin film photovoltaics is very important to efficient operation of devices, and small changes in the composition of the buffer layer (typically 30 – 100 nm thick) may have significant consequences on device performance. The classical n-type semiconductor buffer layer material is CdS deposited by chemical bath deposition; however, much effort has been put into developing alternative materials such as CdxZn1-xS, ZnOxS1-x, ZnxMg1-xO, and In2S3 by vacuum methods since it simplifies in-line vacuum processing. Moreover, because the 2.4 eV band gap of CdS leads to some parasitic absorption, using Zn-based buffers can strongly increase the quantum efficiency of a cell in the blue segment of the spectrum leading to a gain in JSC since the band gaps of ZnO and ZnS are 3.3 eV and 3.6 eV, respectively. Manipulation of the band alignments by varying the stoichiometry also allows one to change the conduction band offset at the pn junction, which may mitigate recombination.In this report, we demonstrate the ability to grade buffer layers for CIGS and CZTS thin film photovoltaics as a function of thickness by atomic layer deposition (ALD), which is among the best deposition methods for controlling thickness and stoichiometry of materials. We demonstrate for the first time ALD of polycrystalline CdS and CdZnS films with large columnar grains using dimethylcadmium, diethylzinc, and hydrogen sulfide, and we show that mixing of the materials yields a well defined alloy1. The crystal structure and growth rate are controllable by both temperature and pulsing sequence of the precursors. We also show that the bandgap can be tuned from 2.4 eV to 3.6 eV with the alloy; thus, we have fine control over the conduction and valence band offsets and can mitigate long wavelength absorption by the buffer.With these materials, we study the role that Cd plays at the interface by depositing uniform CdxZn1-xS films with compositions varying from CdS to ZnS, and by depositing graded films from CdS to ZnS (or ZnOxS1-x) within one buffer layer. Further, we compare the effect of depositing small amounts of CdS by ALD with the result obtained by treating the surface with Cd electrolytes, followed by deposition of CdxZn1-xS or ZnOxS1-x. We analyze the effects of these various treatments and deposition materials on important operating parameters such as VOC, FF, and JSC to better understand the effect of interface formation on thin film devices. Our results show that using ALD we are able to increase JSC compared to CdS CBD, especially with the higher band gap CdZnS alloy. Further, we can affect individual parameters, such as graded layers from CdS to ZnOxS1-x showing improvements in FF.1J. R. Bakke, J. T. Tanskanen, H. J. Jung, R. Sinclair and S. F. Bent, Journal of Materials Chemistry, 2010, In Press.
5:45 PM - D14.5
Influence of Annealing in H2 Atmosphere on the Electrical Properties of Thin Film CdS.
Jaan Hiie 1 , Natalia Maticiuc 1 2 , Aleksei Gavrilov 1
1 Department of Materials Science, Tallinn University of Technology, Tallinn Estonia, 2 Department of Applied Physics and Informatics, Moldova State University, Chisinau Moldova (the Republic of)
Show AbstractChemical bath-deposited (CBD) thin film CdS has been widely used as a buffer and n-type window layer in CdS/CIGS and CdS/CdTe thin film solar cells.Annealing of CBD CdS assigns to the layers required concentration and mobility of electrons, crystallinity, structural stability and perfect ohmic front contact in TCO/CdS interface.But always annealing reduces band gap (Eg) of solution-deposited CdS and lowers current density of the CdS/CdTe PV device due to optical absorption within the CdS layer.We have studied systematically dynamics of changes in CBD CdS/glass thin film structural, optical and electrical properties in annealing process in H2 ambient at normal pressure in pre-heated ceramic tubular furnace.We will present electrical, x-ray diffraction (XRD), photoluminescence and visible optical spectroscopy characterization results of annealed CBD CdS/glass thin films, 300 nm thick. The films were deposited with thiourea from ammoniacal 1 mM dilute solution of CdSO4 and 0.001 at. % of NH4Cl relative to Cd for Cl doping.We found high concentration of electrons 1-3E19cm-3 in the layers annealed at 200 - 450 C, while for 200 C the long time of annealing over 60 min is needed, but for high temperature region 350 - 450 C only for short 10 min annealings this concentration region of electrons was achieved. In the high temperature region very fast decrease of electron concentration will go on with increasing annealing temperature. Mobility of electrons will decrease from 9 to 5 cm2/V.s in the annealing region 200-300 C, which is connected with destruction (cracking) of primary crystallites, detected by XRD. We demonstrate close correlation between electrical, optical and structural properties.On the basis of acquired results we propose an hypothesis about substitutional incorporation of OH group on S site in CdS lattice in deposition process and that (OH)s complex defect acts as a donor defect like Cl and we believe that the both defects are responsible for changes of thin film CdS electrical, optical and structural properties in the annealing process.This work has been supported by EU 7th FP project FLEXSOLCELL GA-2008-230861, by Estonian Science Foundation grants 7241 and 7608, and by Estonian National Target Financing Nos SF0140092s08, and SF0140099s08.