Research Projects Incoming graduate students are encouraged to learn about as many available projects as possible. Information about available projects can be obtained during special seminars in the first few weeks and through inquiries with individual faculty. The following is a list of projects offered by the faculty. The list is generally fluid and incomplete, so students are encouraged to learn about as many of the research groups in the department as possible and explore all opportunities. Doctoral students are asked to submit a list of their preferences for advisors and projects before the fall semester break (Oct. 23, 2015). Please include three preferences in rank order. Doctoral students are then admitted to individual research groups after the fall semester break. The following projects are Ph.D. Projects “Developing Novel Methods for Characterizing Strain and Defects in crystals by Electron Channeling” Professor Yoosuf Picard This project seeks to leverage recent advances in backscattered electron (BSE) detectors and high brightness sources used in scanning electron microscopy (SEM). Electron channeling is a SEM-based approach using BSE detection in order to directly and non-destructively image defects in crystalline solids. CMU has recently procured a commercial SEM with established beam rocking capabilities. In collaboration with this commercial SEM manufacturer, this project will: comprehensively study defect contrast behavior under various SEM imaging conditions, realize a novel approach for direct strain analysis using enhanced beam rocking capabilities, and explore new image modalities via electron channeling phenomena. "Using machine vision and machine learning to address the microstructural big data challenge” Professor Elizabeth Holm Microstructure is the foundation of materials science, whether in metals, polymers, or ceramics. However, we lack systematic methods to analyze and classify microstructures without expert human intervention. Capitalizing on recent advances in computer science, this project applies a subset of data science concepts – including data harvesting, machine vision, and machine learning – to advance the science of microstructure. The goal is to develop an open-access, automatic, and objective machine learning system for finding relationships between microstructural images. Customers include scientists, publishers, and industry engineers. “Building the Materials Genome: Combinatorial Exploration of Quantum Materials” Professor Vincent Sokalski Fascinating new discoveries in physics have made it possible to use an electron’s spin (rather than its charge) to both probe and manipulate the magnetic configuration of materials at the nanoscale, which has resulted in an emerging field we call spintronics. More and more, it’s become clear that the challenges facing this technology are related to sluggish discovery of new materials – a problem that the Materials Genome Initiative launched by the White House in 2011 aims to overcome. In this effort, we use combinatorial thin film deposition techniques that produce precisely controlled composition gradients contained within strategically selected ternary and quaternary magnetic systems (e.g. Heusler alloys). This enables extremely rapid exploration of materials (>100 compositions/sample) when combined with automated testing and characterization techniques. In addition to experimental work, the student will explore thermodynamic models of phase stability and work to improve on tools for plotting, visualizing, and interpreting data in 3 or 4-dimensional space. “Fatigue in Metals” Professor Anthony Rollett We are expecting funding from the Air Force Office of Scientific Research to support work on microstructurally short fatigue cracks. This means finding cracks that are on the same order as the grain size and measuring where they occur in relation to features such as grain boundaries, triple lines etc. The thrust of the project will be to use high-energy x-rays at a synchrotron (most likely the Advanced Photon Source) to measure cracks, the state of stress around them and slip events associated with crack advance. Advanced micro-mechanical modeling (using image-based methods and Fast Fourier transforms) will be used in conjunction with the measurements in order to predict to the greatest extent possible what we plan to measure. The measurements are intended to be state of the art in terms of attempting to use coherency in the x-rays to try to image down at the scale of dislocations. This project will be jointly advised by Tony Rollett (MSE) and Bob Suter (Physics). “Carbon Nanotube Aerogel based Polymer Composites” Professor Mohammad Islam ** See Professor Islam for abstract. “Controlling charges at the surfaces of centrosymmetric oxides” Professor Greg Rohrer Charges at the surfaces of polar ferroelectrics are known to separate photogenerated charge carriers. We have recently found that charges can also occur on the surfaces of nonpolar materials because of their surface termination, that this charge influences photochemical reactions in a way similar to ferroelectric domains, and that the charge can be controlled. This project will be aimed at controlling the charge at oxide surfaces to optimize photochemical reactivity. The following projects are Ph.D./MS “Elucidating Ligand-Induced Phase Separation Processes In Mixed Colloidal Materials” (PhD preferred, MS possible) Professor Michael Bockstaller Description: The viable fabrication of microstructured multicomponent nanoparticle thin films in which distinct particle species are organized in deliberately shaped uniform domain structures presents a pervasive challenge in technology areas ranging from functional coatings to photovoltaics and solid state lighting. The overarching technical objective of this project is to test the hypothesis that ligand-induced phase separation can facilitate the autonomous organization of mixed polymer-modified colloidal systems into monochromatic domain structures. Electron imaging in conjunction with small-angle Xray scattering will be used to elucidate the governing parameters that control the interactions and phase behavior of mixed polymer-tethered particle systems. In collaboration with researchers in the Chemistry Department as well as industry this project will subsequently explore the application of the concept of ’ligand-induced phase separation’ for the fabrication of luminescent panels for backlighting applications. "Atomistic simulations of grain boundary properties”( One PhD and/or one MS student) Professor Elizabeth Holm Atomistic simulation is a powerful tool to understand the behavior of materials at the scale of the individual atoms. In this project, we will use molecular dynamics simulations to understand the temperature dependent properties (structure, energy, and mobility) of grain boundaries in metals. We are particularly interested in how the structure changes from smooth to rough and from faceted to unfaceted as the temperature increases, because these phenomena strongly affect the annealing behavior of metals. This study will be collaborative with a PhD project in the department. The following projects are MS Projects “Enzyme-Actuated Nanoporous Polymer Membrans” (MS only) Professor Michael Bockstaller Active membranes, that is membranes capable of adjusting permeability in response to external stimuli, hold the potential to transform applications ranging from water purification to protective clothing. This projects is part of a collaboration with researchers at Cornell University as well as industry that is focused on (1) establishing the conditions for incorporation of active enzymes into phase-separated nanoporous polymer membranes and (2) the elucidation of enzyme-actuated phase transitions in nanoporous polymers. Electron imaging, small-angle X-ray as well as light scattering will be applied to established the mechanism of interaction of enzymes and organic substrate materials. “Molecular dynamics simulations of coated carbon nanotubes”(One research MS student) Professor Elizabeth Holm Coating carbon nanotubes (CNTs) with graphene changes their properties. In this project, we are particularly interested in the mechanical properties of graphene-coated CNTs, both individually and in pairs. We will use molecular dynamics simulations to calculate the modulus of single coated CNTs, as well as the interaction potential of pairs of CNTs, and we will use the results to model the behavior of CNT aerogel materials. This study will be collaborative with a PhD project in the department. “Building the Materials Genome: Combinatorial Techniques to Explore Advanced Energy Materials” Professors Vincent Sokalski & Paul Ohodnicki (Adjunct) The Materials Genome Initiative (MGI) was announced in 2011 with the universal goal of doubling the rate at which new, functional materials are discovered and developed. In the area of energy technology, finding materials with the composition & structure necessary to resist degradation in the presence of corrosive gas mixtures is critical for future natural gas distribution technology. In this project, the student will leverage combinatorial techniques to grow thin films that, by design, will have a distribution in composition across the surface of a substrate where each position represents a different composition in a ternary or quaternary system. Such a technique enables extremely rapid exploration of materials that will be analyzed using advanced structural and spectroscopic techniques. Experimental work will be coupled with thermodynamic modeling (e.g. FactSage and ThermoCalc) to explain trends in phase identity / stability. This work will be performed in collaboration with the National Energy Technology Laboratory (NETL) research facility in Pittsburgh, PA. “Strain Induced Anisotropy in Ni-based Nanocomposites” Professor M. E. McHenry Magnetic nanocomposites, annealed under stress, have recently been investigated for inductors, transformers, sensors, motors, and other magnetic applications. We have demonstrated induced magnetic anisotropies, Ku’s, in strain annealed Co-based metal/amorphous nanocomposites (MANCs) with a response over an order of magnitude larger than field annealed Co-based MANCs and response to applied stress twice that of Fe-based MANCs. This project will study strain induced anisotropy in Ni-based MANCs Transverse magnetic anisotropies and switching by rotational processes are expected to impact anomalous eddy current losses at high frequencies. A phenomenological model for multi-phase resistivity in MANCs identifies implications for classical eddy current losses. We will correlate response with concentrations of early transition metal elements (TE) providing virtual bound states (VBS) to Co(Fe) rigid bands, that increase both induced anisotropies and resistivities. This work is aimed at applications in inductive components for power electronics. “Magnetocaloric Effect in Alloys with Distributed Exchange Interactions” Professor M. E. McHenry This project investigates the role of crystallographic disorder and pressure on magnetic exchange interactions and magnetocaloric effects (MCE) in (1) multicomponent, fcc-based High Entropy Alloy (HEA) systems and (2) Low symmetry hexagonal and orthorhombic giant magnetocaloric and barocaloric materials. We will investigate Curie temperature engineering and distributed exchange interactions in multicomponent crystalline solid solutions. We will investigate systems with 2nd and 1st order magnetostructural phase transitions considering: (1) Synthesis of new MCE Materials. (2) Synchrotron X-ray structural characterization and phase transitions under pressure. (3) Structural and microstructural characterization by TEM. (4) Magnetic, thermal and thermomagnetic properties measurements. (5) Modeling of exchange interactions to explain and predict MCE from first principles. (a) J(R) using a parameterized Bethe-Slater curve. (b) Confirmation of J(R) distributions in model systems using first principles calculations (c) Model the change in J(R) distribution with pressure. (d) Use selected J(R) distributions in a mean field theory of the magnetic transition. (6) Assessment of materials performance metrics for applications. MCE materials are of current interest for applications in refrigeration because this technology is energetically more efficient than conventional gas compression refrigeration, by about 20 %. MCE materials are more environmentally friendly because ozone depleting and warming refrigerants are not used. Critical rare earths metals (REs) and compounds have large MCEs and working temperatures close to room temperature, but their scarcity, high price and corrosion limit their commercial use. Transition metalbased alloys will be investigated to replace RE’s. Environmental, materials scarcity and cost considerations speak to the potential of these materials. MCE based refrigerators have been announced for consumer products as early as 2020. “Oscillation mark formation in continuous casting” Professor Chris Pistorius Oscillation marks are ubiquitous on the surface of continuously cast steel, and clearly form as a result of mold oscillation (under normal casting conditions). However, the mechanism(s) of oscillation mark formation remain unclear. Possible mechanisms include solidification along the liquid steel meniscus (related to “hook” formation), and a sticking-healing mechanism (similar to what occurs during horizontal casting). In this project, steel solidified in the mold will be examined to test possible mechanisms. Normally, the detail of initial solidification is obliterated by subsequent oxidation (scaling) of the steel surface, and deformation of the surface by support rolls. However, a number of samples which solidified in the mold during plant upsets are available, and these will be examined metallographically and by surface profiling. “Water content and mold flux heat transfer” Professor Chris Pistorius Mold flux, a high temperature lubricant in continuous casting of steel, typically contains oxide and fluoride compounds. Mold flux is hygroscopic; water absorbed by the mold flux before use can be expected to dissolve in the mold flux when it melts during use. Dissolved water is expected to have at least two opposite effects on the solidified mold flux film that forms on the caster mold: an increased tendency for the mold flux to crystallize (which would increase thermal conductivity), and formation of more porosity (which would tend to lower thermal conductivity). In this project, selected mold fluxes will be exposed to atmospheres of controlled humidity, and then the behavior in the coldfinger heat transfer simulator will be tested to evaluate the opposing effects of crystallization and porosity formation “Using microwave radiation to catalyze low temperature thin film growth at solid-liquid interfaces” Professor B. Reeja Jayan Microwave radiation (MW) provides a “green” low temperature method for materials synthesis. My research previously demonstrated that MW can assist in assembling thin films directly on various substrates (e.g. glass, metal, plastic). Specifically, the selective heating of a conducting layer (e.g. metal, ITO) by MW significantly lowered the solution temperatures (< 200 oC) required for growing crystalline thin films (Figure 1). Films grown by conventional techniques (e.g. sol-gel, hydrothermal) require high post-deposition annealing temperatures (> 400 oC) to crystallize, particularly for inorganic materials like ceramics. These high processing temperatures create incompatibility with microfabrication processes and limit the choice of substrate materials on which these films can be grown, as flexible plastic or polymeric substrates typically decompose at temperatures > 200 oC. Figure 1: Schematic of microwave assisted thin film growth Conventional heating is limited by thermal conduction from the walls of the reaction vessel. In contrast, MW quickly and uniformly heats a solution by directly coupling energy to molecules within the solution through polarization (dielectric heating) and conduction (ohmic heating). Dielectric and ohmic heating in solution are commonplace in rapid, energy efficient synthesis of inorganic nanocrystals. However, the synthesis of thin films by selective ohmic heating of solid materials within a growth solution has been minimally explored. Several open questions remain relating to the existence of a specific or non-thermal MW effect that modifies the effective temperatures and activation energies of molecules at the solid-liquid interface. Accordingly, our research project has three goals: The first is to study by experimental and modelling (electromagnetic, thermal) methods, how MW influences chemistry at an interface to promote film growth. Thin film growth will be selectively localized to a desired region inside the reaction vessel (e.g. the substrate) using the conducting layer, facilitating in-situ monitoring of reaction intermediates and interfacial charge transfer using analytical and spectroscopy techniques. The second goal is to utilize the knowledge gained to identify molecular precursors for energy efficient synthesis of thin films of advanced material systems like perovskites. These materials typically require processing at 400-1000 oC due to inefficient heat transfer in conventional processes. The third goal is to apply the MW-assisted process to synthesize a broad range of organic, inorganic and hybrid materials with unique morphologies (e.g. microporous, patterned surfaces) and properties (e.g. high surface area) for research on energy devices and biosensors. The data generated from the goal 1 spans 100s of time-consuming experimental runs where parameters shown in Figure 2 are sequentially varied and properties like film crytallinity, thickness, and morphology (uniformity, delamination) are evaluated. Analyzing this data to predict experimental conditions and/or pathways to optimize film properties within a reasonable number of (< 10) experimental runs is highly desirable. Towards this end, novel methods to arrange, database and analyze this data set is also required. Figure 2: Summary of film growth optimization conditions. The most uniform and thickest films were grown at a temperature of 150 oC held for 60 minutes. Shorter hold times or lower temperatures led to thinner, more amorphous films. Longer times or higher temperatures led to film delamination at the edges. The power ramp rate had little effect on film formation. “Research and Development of Gallium Oxide for Ultra-Efficient High-Power Electronics” Professor Lisa M. Porter The worldwide increase in the demand for energy places a growing need for highly efficient electronics for energy conversion and transport. In particular, power electronics are needed for electrical switching within the electrical grid and for green modes of transportation, such as in switched-mode power supplies for hybrid electric vehicles. The market for these power electronics components is predicted to reach $15 billion by 2020 [1]. Whereas silicon-based devices are most commonly used in traditional high-power electronics applications, wide bandgap semiconductors are much more efficient and thus useful for future energy applications, because their larger bandgaps allow them to withstand higher electric fields with less material and less energy loss. For example, Toyota recently began trials of a new hybrid system using power electronics based on the wide bandgap semiconductor SiC [2] and claims that power electronic devices based on SiC could increase fuel efficiency of hybrid vehicles by 10% [3]. The main disadvantage of SiC and GaN (another prime wide bandgap candidate), however, is that their substrates are very expensive due to the energy intensive growth methods required. A very promising alternative to SiC and GaN is gallium oxide, Ga2O3, which has an even larger bandgap than that of SiC or GaN; moreover, Ga2O3 bulk single crystals can be grown using the same inexpensive melt-growth methods and equipment that are used to grow sapphire, and have very recently become commercially available in 2-in diameter wafers [4]. The prospect of a ‘new’ wide bandgap semiconductor having both superior properties (Table 1) for many electronic and optoelectronic devices as well as commercially-available substrates is a novelty not seen in the wide bandgap semiconductor field for at least two decades; this situation has opened up new possibilities for disruptive devices and technologies and could therefore translate to even greater fuel efficiencies at lower cost than predicted using traditional wide bandgap semiconductors (e.g., SiC or GaN). Devices based on b-Ga2O3 that have so far been demonstrated (with the vast majority in Japan) include Schottky diodes [5, 6], metalsemiconductor field effect transistors (MESFETs) [7], metal-oxide-semiconductor fieldeffect transistors (MOSFETs) [8], and ultra-violet photodiodes [9-14]. However, with research on b-Ga2O3 as a wide band gap semiconductor in its very early stages, there is little understanding of how to control device-relevant interfaces to this material. In this research program we aim to optimize high-power-device interfaces (e.g., ohmic and Schottky contacts to Ga2O3) to hasten the development of commercially-viable, ultraefficient high-power electronics based on Ga2O3. Table 1. Electrical properties of b-Ga2O3 and other selected semiconductors. (Electrical properties obtained from [15]) Si 4H-SiC GaN Diamond b-Ga2O3 Bandgap (eV) 1.1 3.3 3.4 5.5 4.7–4.9 Breakdown field (MV/cm) 0.3 2.5 3.3 10 8 Relative dielectric constant 11.9 9.7 9.0 5.5 10 Electron mobility (cm2/V s) 1400 1000 1200 2000 300 Power FoM rel. to Si 1 340 870 24660 3440 Melting point (°C) 1412 2830 ~2500 [17] Requires 1740 (decomposes) [16] ultra-high pressures Commercial single- Yes Yes, expensive Very No Recently crystal substrates limited, available readily available? expensive References 1. Stevenson (Ed.), R. Electric vehicles: SiC and beyond. Compound Semiconductor, 2014. 20, 18-19. 2. Aubernon, C. Toyota unveils silicon carbide semiconductor trial. The Truth about Cars 2015 January 30, 2015; Available from: http://www.thetruthaboutcars.com/2015/01/toyota-unveils-silicon-carbidesemiconductor-trial/. 3. Ashley, S. Efficient power electronics for hybrids and EVs. Automotive Engineering Magazine 2014 4 June 2014; Available from: http://articles.sae.org/13244/. 4. Tamura Corporation, Single-crystal gallium oxide substrates, Tokyo, Japan: http://www.tamura-ss.co.jp/en/release/20131122/. 5. Sasaki, K., M. Higashiwaki, A. Kuramata, T. Masui and S. Yamakoshi, MBE grown Ga2O3 and its power device applications. J. Cryst. Growth, 2013. 378: p. 591-595. DOI: 10.1016/j.jcrysgro.2013.02.015. 6. Sasaki, K., M. Higashiwaki, A. Kuramata, T. Masui and S. Yamakoshi, Ga2O3 Schottky barrier diodes fabricated by using single-crystal β-Ga2O3 (010) substrates. IEEE Electron Dev. Lett., 2013. 34(4): p. 493-495. 7. Higashiwaki, M., K. Sasaki, A. Kuramata, T. Masui and S. Yamakoshi, Gallium oxide (Ga2O3) metal-semiconductor field-effect transistors on single-crystal β-Ga2O3 (010) substrates. Appl. Phys. Lett., 2012. 100: p. 013504.1-3. DOI: 10.1063/1.3674287. 8. Higashiwaki, M., K. Sasaki, T. Kamimura, M.H. Wong, D. Krishnamurthy, A. Kuramata, T. Masui and S. Yamakoshi, Depletion-mode Ga2O3 metal-oxidesemiconductor field-effect transistors on β-Ga2O3 (010) substrates and temperature dependence of their device characteristics. Appl. Phys. Lett., 2013. 103: p. 123511.14. DOI: 10.1063/1.4821858. 9. Huang, C.-Y., R.-H. Horng, D.-S. Wuu, L.-W. Tu and H.-S. Kao, Thermal annealing effect on material characterizations of β-Ga2O3 epilayer grown by metal organic chemical vapor deposition. Appl. Phys. Lett., 2013. 102: p. 011119.1-3. DOI: 10.1063/1.4773247. 10. Oshima, T., T. Okuno, N. Arai, N. Suzuki, H. Hino and S. Fujita, Flame detection by a β-Ga2O3-based sensor. Jpn. J. Appl. Phys., 2009. 48: p. 011605.1-7. 11. Oshima, T., T. Okuno, N. Arai, N. Suzuki, S. Ohira and S. Fujita, Vertical solar-blind deep-ultraviolet Schottky photodetectors based on β-Ga2O3 substrates. Applied Physics Express, 2008. 1: p. 011202.1-3. 12. Oshima, T., T. Okuno and S. Fujita, Ga2O3 thin film growth on c-plane sapphire substrates by molecular beam epitaxy for deep-ultraviolet photodetectors. Jpn. J. Appl. Phys., 2007. 46(11): p. 7217-7220. 13. Ravadgar, P., R.-H. Horng, S.-D. Yao, H.-Y. Lee, B.-R. Wu, S.-L. Ou and L.-W. Tu, Effects of crystallinity and point defects on optoelectronic applications of β-Ga2O3 epilayers. Optics Express, 2013. 21(21): p. 24599-24610. DOI: 10.1364/OE.21.024599. 14. Suzuki, R., S. Nakagomi, Y. Kokubun, N. Arai and S. Ohira, Enhancement of responsivity in solar-blind β-Ga2O3 photodiodes with a Au Schottky contact fabricated on single crystal substrates by annealing. Appl. Phys. Lett., 2009. 94: p. 222102.1-3. DOI: 10.1063/1.3147197. 15. Higashiwaki, M., K. Sasaki, A. Kuramata, T. Masui and S. Yamakoshi, Development of gallium oxide power devices. Phys. Status Solidi A, 2014. 211(1): p. 21-26. DOI: 10.1002/pssa.201330197. 16. Madelung, O., U. Rössler and M. Schulz, eds. Silicon carbide (SiC), Debye temperature, density, hardness, melting point, thermodynamic functions. LandoltBörnstein - Group III Condensed Matter Numerical Data and Functional Relationships in Science and Technology. Vol. 41A1b. 2014, SpringerMaterials. 17. Madelung, O., U. Rössler and M. Schulz, eds. Gallium nitride (GaN), Debye temperature, melting point, density. Landolt-Börnstein - Group III Condensed Matter. Vol. 41A1b. 2014, SpringerMaterials. “Particle Pinning of Grain Growth and Grain Size Distributions” Professor Anthony Rollett Joint with Elizabeth Holm, this project will examine the effect of second phase (inert) particles on grain growth using simulation and the Potts model. This will be a continuation of previous work on this topic with the aim of completing and publishing the work. The main result is that particles limit the grain size according to the longestablished Smith-Zener model (1948). The additional result, which will be tbe focus of the project, is that the grain size distribution without particles changes from one that deviates strongly from log-normal, to one with particles that is close to log-normal. At least two different Potts model codes will be used. Data analytics (e.g. principal component analysis) will be used. The simulation results will be compared to experimental data to the extent that it is available. The student will learn about both simulation techniques for microstructural evolution and statistical data analysis. “Simulation of stress fields around cracks using the FFT method” Professor Anthony Rollett This project will use the so-called FFT method for (elasto-)viscoplastic FFT method to investigate stress fields around crack tips. This is a classical problem in linear elastic fracture mechanics, which means that it represents a useful verification test of the imagebased FFT method [Lebensohn, Acta mater. 2001]. Dream3D [http://dream3d.bluequartz.net/] will be used to generate the required 3D microstructures, with the addition of different crack geometries. Some work has already been as a Senior project so the expectation is that results will be rapidly obtained and it will be possible to publish a paper. “Titanium-containing inclusions in steels” Professor Bryan Webler During liquid steel processing, titanium can react with various components to form oxide, nitride, sulfide, or carbide inclusions. These can significantly influence steel properties and performance during subsequent processing. For example, thermal stresses on cooling can sometimes result in cracking in slabs of titanium-stabilized ultralow carbon steels. The evolution and distribution of titanium-containing inclusions plays an important role in these problems. This project will examine the evolution and distribution of titanium-bearing inclusions in several industrially-relevant steel grades. The observed inclusion behavior will be connected to reaction thermodynamics and kinetics as well as implications for subsequent processing steps. “Aerogels of Nanotubes and Graphene as Electrodes for Supercapacitors and Batteries” Professor Mohammad Islam ** See Professor Islam for abstract. “Enzymatic biofuel cells using carbon nanotube-based gels” Professor Mohammad Islam This project involves the development and characterization of enzymatic biofuel cells that utilize carbon nanotube-based gels as electrode materials. We have shown the capacity of these materials to allow 1 to 2 orders of magnitude higher enzyme loadings than the highest performing systems to date, but our systems lack optimization of power output and stability.1 Toward this goal, the student carrying out this project will focus on producing and modifying graphene/SWCNT cogels with materials such as gold nanoparticles and polymer coatings. Fabrication and modification conditions will be varied to produce ideal electrode properties. The end goal of this research is the development of an optimized enzymatic biofuel cell with increased power density and stability to be implemented in the continuous powering of implantable devices. 1 A.S. Campbell, Y.J. Jeong, S.M. Geier, R.R. Koepsel, A.J. Russell and M.F. Islam, ACS Applied Materials and Interfaces 7, (2015) 4056-4065