Synchrotron radiation and atomic layer deposition for advanced materials

Resume : Electron enhanced atomic layer deposition (EE-ALD) can drastically reduce the temperatures required for film growth through electron stimulated desorption (ESD) of surface species. The desorption process creates highly reactive “dangling bond” surface sites. Precursors can then adsorb efficiently on the dangling bonds. EE-ALD enables deposition of thin films on thermally sensitive substrates and growth of laminate materials where alloying would occur at traditional processing temperatures. We have demonstrated the EE-ALD of polycrystalline GaN, BN, and amorphous Si at room temperature (27˚C). Film growth was performed using alternating exposures of the appropriate precursors and low energy electrons (~100 eV). X-ray reflectivity and spectroscopic ellipsometry measurements monitored linear film growth versus number of reaction cycles. Additionally, we observed some dependence of the growth rate on electron flux and electron energy. Growth rates varied from 0.2 Å/cycle for Si films to 3 Å/cycle for BN films. Depth-profiling using x-ray photoelectron spectroscopy demonstrated clean BN (1.3:1) films with < 3 at.% C and O impurities. Crystallites as large as 10 nm were observed by high resolution transmission electron microscopy and grazing angle x-ray diffraction for GaN films. EE-ALD should facilitate the deposition of a variety of ALD films at low temperature. Materials grown with hydride or halide precursors are all possible candidates due to their large ESD cross-sections.

Resume : In the last year, all-inorganic perovskite quantum dots (CsPbX3 QDs with X=Br,I,Cl) have emerged as a new class of semiconductor nanocrystals with improved optical properties compared to the well-studied cadmium based QDs.[1,2] One of the issue that still remain to be addressed for implementation in optoelectronic devices and for the exploration of their properties under ambient conditions is the lack of stability with respect to temperature, moisture and light exposure. We have successfully developed a low temperature atomic layer deposition (ALD) process for the deposition of a thin layer of metal oxide on a film of CsPbBr3 QDs that 1) protects the perovskite QDs from water and high temperature and 2) prevents dissolution of the bottom layer or anion exchange in a multi-layering process.[3] The nucleation and growth process of AlOx on the QD surface was thoroughly investigated by a miscellanea of techniques which highlighted the importance of the interaction between the ALD precursors and the QD surface to uniformly coat the QDs while preserving the optoelectronic properties. These nanocomposites show an exceptional stability against exposure to air (for at least 45 days), irradiation under simulated solar spectrum (for at least 8h), to thermal treatment (at least up to 200ºC in air), and finally against immersion in water. The method was extended to assembly CsPbBrxI3-x QD/AlOx and CsPbI3 QD/AlOx nanocomposites which were more stable compared to the pristine QD films.[3] 1) D. Bera, et al., Materials, 2010, 3, 2260. 2) L. Portesecu, et al., Nano Letters, 2015, 15, 3692. 3) A. Loiudice, et al., Angew. Chem. Int. Ed. 2017, accepted.

Resume : Atomic-layer deposition (ALD) is an attractive method for low-temperature growth of nitride semiconductors. However, carbon contamination remains a concern during ALD growth, since cation precursors frequently contain methyl groups. While plasma-assisted ALD can aid in lowering carbon incorporation, the optimal growth conditions for achieving minimal carbon concentrations are unknown. Using density functional theory (DFT) we have calculated the behavior of methyl groups on InN, GaN, and AlN surfaces. We determine the surface binding energies of these species, their interactions with atomic hydrogen (present in plasma-enhanced ALD), and the mechanisms for incorporation into the nitride lattices. We incorporate these DFT results into a kinetic Monte Carlo model, allowing us to predict the amount of carbon impurity incorporation as a function of growth conditions.

Resume : Zinc blende-structure copper(I) halide materials (CuHa) are direct gap semiconductors with band energies in the ultra-violet region. They have high exciton and bi-exciton binding energies which have the potential for laser action with very low lasing threshold. Thin films and nanocrystallites of CuHa have been deposited by evaporation, sputtering and molecular beam epitaxy. Recently, preliminary work has shown that ALD has the ability to deposit thin films and nanocrystallite arrays which exhibit the characteristic photoluminescence of zinc blende CuCl. This paper investigates the growth processes and crystal structure and shows that CuCl films can be grown within an ALD window which ranges from 50oC to 150oC using [bis(trimethylsilyl)acetylene] (hexafluoroacetylacetonato)-copper(I) and HCl in butanol as Cu and Cl precursors, respectively. Initial nucleation depends on the length of the post-Cl purge times with short purge time leading to faster nucleation. After nucleation the deposition rate is similar in both cases. In addition, the results of film deposition of CuCl and CuBr using alternative halogen precursors will be presented. The use of in-situ deposition of capping layers to prevent degradation of the CuHa due to atmospheric moisture and their effect on the stability and structure of the films will be described.

Resume : Integration of III-V compounds with Si is of great interest for development of optoelectronics devices like light emitting diodes, solar cells and optical IC. However, high quality and long term stable devices require rather a lattice matched growth of III-V materials. Thus GaP is the best candidate to be grown on Si substrates because it has the smallest lattice mismatch with Si (less 0.4 %) beyond all binary III-V alloys. GaP has an indirect band with a gap of 2.26 eV being a promising material for barrier layers but it is unsuitable for active layers of top junction for multijunction solar cells based on Si. Lattice-matched GaP/Si superlattice is proposed to be used as an active material. According to theoretical estimation GaP/Si superlattice for Si(100) plane becomes a direct-gap [1]. However, the growth of III-V and IV layers with sharp interfaces and controllable doping level is difficult to realize using conventional epitaxial technology of such as MOCVD and MBE. In this paper a new approach to growth GaP/Si quantum wells at temperatures not exceeding 400 °C is explored. The layers of GaP are grown by plasma enhanced atomic layer deposition, where the growth of Ga atom monolayer alternates with the growth of one monolayer of P atoms. While Si is grown using conventional PECVD mode with high hydrogen dilution. [1] P. J. Lin-Chung. Electronic Structures for (Si)m (GaP)n Superlattices. MRS Proceedings, 220 (1991) 589

Resume : In recent years, it has been realized that the unique features of atomic layer deposition (ALD) can also be employed to face processing challenges for various types of solar cells. With this, ALD for photovoltaics (ALD4PV) has attracted great interest in academic and industrial research and it has even been introduced in high-volume manufacturing. In this presentation, the status of the use of ALD nanolayers in various solar cell technologies will be reviewed and their future prospects will be discussed. The presentation will focus particularly on (i) the application of ALD oxides, most prominently Al2O3, for the passivation of surfaces of high-efficiency silicon solar cells; (ii) the preparation of highly transparent conductive oxides (doped ZnO films and In2O3:H); and (iii) the use of ALD-prepared nanolayers in perovskite solar cells. Also some underlying aspects of the ALD processes and material systems will be addressed on the basis of results obtained by advanced characterization techniques and by in situ and in operandi studies that have been carried out.

Resume : Atmospheric pressure spatial atomic layer deposition (AP-SALD) has emerged as a competitive and appealing technique for rapidly depositing semiconducting oxides for electronics. The growth rates achievable have been demonstrated to be up to two orders of magnitude higher than conventional ALD, and yet the films have comparable quality, uniformity and conformality to the substrate. We will discuss the operating principles of AP-SALD and the variety of laboratory- and industry-based reactors that have been developed. These will include the reactors from our laboratory and how the design has evolved over time to achieve improved uniformity and conformal coverage of high-aspect ratio nanostructures. A selection of the broad range of n-type metal oxides that have been grown by AP-SALD will be explored, with particular emphasis on the effect of doping on tuning materials properties, such as bandgap and carrier concentration. We will also explore the narrower range of p-type oxides grown by AP-SALD, and highlight the growing importance of expanding upon the number of p-type materials explored for thin film devices. Finally, we will show how the rapid growth of oxides with controllable electronic properties by AP-SALD has been critical in enabling improvements in the performance of emerging thin film photovoltaic materials and perovskite light emitting diodes.[1,2] [1] R. L. Z. Hoye, et al., Adv. Mater., Accepted. DOI: 10.1002/adma.201702176 [2] R. L. Z. Hoye, et al., Adv. Mater., 27(8), 1414-1419 (2015)

Resume : Recently tin oxide was employed as electron contact in perovskite solar cells and in silicon-perovskite tandem solar cells. The best silicon heterojunction-perovskite tandem employed SnO2 deposited by PE-ALD in the internal tunnel-recombination junction between two subcells. As a first step for the further optimization, we conducted a study of the growth of SnO2 on silicon using in-system XPS. We varied the SnO2 PE-ALD process by deposition temperature (120-250 °C), oxygen plasma step duration (2-8 seconds) and plasma power (200 to 500 W). Samples were grown on c-Si substrates, and in-system XPS was applied to measure the layer stoichiometry and to quantify carbon and nitrogen contamination levels. Additionally, optical (n,k) and electrical (conductivity) properties were measured. The main focus of our study is the dependency optoelectronic properties of SnO2 layers on their elemental composition. XPS analysis showed that all samples contained Sn, O, C and N peaks. O, C and N peaks contained signals originating from residuals of the metal-organic precursor. We find that OH residuals are detrimental to the conductivity of SnO2 and reached a resistivity of 1mOhm*cm. All SnO2 films contained carbon residuals, which are partly assigned to surface contamination and partly to carbon incorporated into the volume. In total the carbon content of the films varied from 13-30%. Increasing carbon content in the layer leads to an increase of the refractive index from 1.6 up to 1.71.

Resume : The evaluation of the electronic structure and intrinsic defect mechanisms in thin films is essential for their effective use in applications with desired functionality such as surface passivation schemes for solar cells [1,2]. We present a comparative study of Al2O3, HfO2, and TiO2 films grown by atomic layer deposition (ALD) [3,4]. These films were characterized by resonant photoelectron spectroscopy and by electrical measurements. For all films investigated intrinsic defect states within the electronic band gap were observed including excitonic, polaronic, and charge-transfer defect states, where their relative abundance is subject of the choice of ALD parameters and of the used substrate. The spectroscopic assigned in-gap defect states are related with electronic charges as determined in the electrical measurements. We determine the ionicity in these systems and find that, in particular, the polaronic charges correspond to the covalent fraction. We attribute them to quasi-free charge carriers. The excitonic defects are located around the Fermi energy. Their existence and abundance depends on structural defects within the ALD films. [1] G. Dingemans and W.M.M. Kessels, J. Vac. Sci. Technol. A 30 (2012) 040802. [2] M. Kot et al.: ChemSusChem 9 (2016) 3401. [3] S. Alberton Corrêa, S. Brizzi, D. Schmeisser, J. Vac. Sci. Technol. A 34 (2016) 01A117. [4] K. Henkel, M. Kot, D. Schmeißer, J. Vac. Sci. Technol. A 35 (2017) 01B125.

Resume : The discovery of ferroelectric (FE) behavior in 10 nm thin Si:HfO2 films [1] in 2011 has raised hope for further discoveries of up-to-now unknown FE material classes and for overcoming integration and scaling issues of ferroelectric memories. The proof of the originally claimed Pca21 phase succeeded just in 2015 [2]. However, results obtained by transmission electron microscopy might not sufficiently reflect the structure of hafnia in the thin film capacitors (poor statistics, structure relaxation in the lamella). X-ray diffraction (XRD) provides improved sampling of the structure for a larger averaged description. However, the existence of multiple phases with overlapping and broadened XRD peaks and insufficient sensitivity to the oxygen ions hamper structural studies [3]. Here, x-ray absorption spectroscopy comes into play. First, powder samples of different phases were measured at different temperatures to understand the different polymorphs under well-defined conditions and separate from each other. After that, thin-film capacitors of different compositions were studied and the differences in structure compared to electrical data and the structural data obtained for the bulk references. [1] T.S. Böscke et al. Appl. Phys. Lett. 2011, 99, 102903, DOI: 10.1063/1.3634052 [2] X. Sang et al., Appl. Phys. Lett. 2015, 106, 162905, DOI: 10.1063/1.4919135 [3] E. D. Grimley, T. Schenk et al., Adv. Electron. Mater. 2016, 2, 1600173, DOI: 10.1002/aelm.201600173

Resume : Piezo- and ferroelectric thin films become more and more important in modern technology, with the ever increasing need of higher performing materials on smaller and smaller devices. Atomic layer deposition (ALD) can be used to deposit ferroelectric films, with routes for materials such as Pb(Zr,Ti)O3, BiFeO3, LiNbO3 and KxNa1-xNbO3 already available [1]. Understanding the structural behavior of such films upon perturbation by an electric field is crucial to design materials with the necessary functional and mechanical properties for industrial use. This type of in situ probing of structure is not very common, but could make an interesting contribution to the field. In this work, we present deposition of ferroelectric LiNbO3 and KxNa1-xNbO3 by ALD, and point out some important advantages and disadvantages of using ALD to make these materials. Evidence of ALD-type growth is presented, and piezo- and ferroelectricity is shown using piezoelectric force microscopy and electrical measurements (hysteresis loops). Furthermore, together with SNBL@ESRF, we are currently working on an in situ probe for measuring crystal structure of the thin films upon perturbation by an electric field. This setup is currently under commissioning, but preliminary results on some powder samples are available. We can follow the elongation of the unit cell of a PMN-PZT piezoelectric in situ while varying the applied electric field. This way we can correlate structural changes to the hysteresis loop, giving us important insight in the structural changes around the loop. Changes in diffuse scattering around the peaks are also visible, possibly increasing the amount of information we can get out of these measurements. This setup already allows for reflection geometry, and can be used for thin films as soon as a sample stage that allows for electrical measurements have been constructed. In this talk we would like to present the possibilities of such a setup, and discuss what can be learned by doing this type of measurements. We also point out some important constraints of both the setup and the deposited materials. We foresee the possibility to do measurements on thin films come fall 2017. [1] H. H. Sønsteby et al., Adv. Mat. Interfac., 10.1002/admi.201600903 (2017)

Resume : Redox-based memristive devices (ReRAM) are considered for future high-density non-volatile data storage due to significant advantages in speed, power and scalability. In contrast to standard crossbar and pillar-structured cells, most of the vertical 3D ReRAM approaches require the resistive switching oxide layer to be deposited on the chemically active non-noble metal electrode. In this study, dense Hf, Ta and Pt metal layers deposited by off-axis sputtering were transferred through an UHV tunnel into the atomic layer deposition system (ALD) for the growth of ultrathin layers of Al2O3 and TiO2. Angle dependent X-ray photoelectron spectroscopy with different take-off angles was utilized to determine the thickness of the interfacial metal oxide layer formed as well as the valence states of the involved metal cations. No PtOx was found at the oxide/Pt interface. In contrast, applying identical processes to Hf and Ta metal surfaces/films resulted in oxidized interfacial layers with a thickness of up to 6 nm depending on the materials and process parameters. A direct comparison between TiO2 and Al2O3 capping layers clearly verified that Al2O3 acts effectively as a diffusion barrier resulting in thinner HfO2 and Ta2O5 layers compared to the TiO2 deposition. Additionally performed hard X-ray photoelectron spectroscopy experiments allowed a more detailed analysis of the deeper interfaces up to the pure metal bottom layer to from a complete picture of the layer stack.

Resume : Low energy ion scattering (LEIS) provides an analytical tool for probing surface composition and structure on the angstrom to nanometre. These length scales are central to the processing of thin films in electronic devices such as resistive RAM (RRAM), insulated – gate power transistors and transparent electronics. We will report on the application of LEIS to the elucidation of ALD deposition processes and how it can provide information about four factors, namely: the growth per cycle; adatom incorporation into the growing film; the nature of ligand leaving groups; and film-substrate interfaces. The complementary nature of the information from ion scattering will be compared to that provided by photoemission and X-ray reflectivity measurements, as a means of high-lighting the synergies and challenges of exploiting synchrotron radiation to understand the structure of ultrathin films.

Resume : ALD has gained an important and rapidly increasing role in microelectronics. The main drivers for this growth have been the increasing needs of high conformality, uniformity and repeatability that derive from the device development toward increasingly three-dimensional structures and smaller layer thicknesses. In addition, introduction of new materials and novel patterning techniques have opened for ALD fast lanes to production. ALD technology consists of two main components: reactors and processes. As the reactors have reached a level meeting industry specifications, the future of ALD will depend largely on development of new processes. Indeed, the success of ALD is built on chemistry: the unique benefits of ALD can be exploited only when proper precursors undergoing saturative surface reactions can be identified for the material of an interest. Development of ALD precursors and processes has therefore been in a central role over the entire history of the ALD technique. The review of Miikkulainen et al. [1] lists 780 ALD processes for 160 materials with 300 metal precursors, yet there are lots of important materials that lack ALD processes and also many of the existing ones call for improvement. This presentation summarises our recent results on ALD process development for materials with expected interest in the future generations of microelectronics, like MoS2, ReS2, Au, Cu and Re. Further directions and challenges in ALD process and precursor development as well as needs for reaction mechanism studies will also be outlined. [1] V. Miikkulainen, M. Leskelä, M. Ritala and R. L. Puurunen, J. Appl. Phys. 113 (2013) 021301.

Resume : Sensing properties of novel optical sensors strongly depend on properties of the thin film materials - such as electrical permittivity or optical properties and size or thickness – that these devices the most often require for initiating or modifying their sensorial response. In this work we studied structure, composition and optical properties of TaxOy nanolayers as a promising dielectric material that can be easily deposited by low-temperature (100C) atomic layer deposition (ALD) technique, for application in optical fiber sensor technology. The TaxOy planar test layers were deposited at the temperature 100C on glass and silicon using TFS-200 (Beneq) ALD system, with tantalum pentachloride and deionized water as chemical precursors, and 6N-purity argon as carrier/purging gas. Their properties were investigated using x-ray diffractometry, Ratherford-backscattering spectroscopy, x-ray photoelectron spectroscopy (XPS) and spectroscopic ellipsometry (SE) performed in visible-to-near-infrared range. Experiments revealed total amorphousness of the layers and slight differences between compositions when deposited on different substrates, with lower O/Ta atomic ratio for glass-related layers. The compositions measured from layers’ surfaces towards the layer-substrate interface were not constant, with decreasing O/Ta atomic ratio in the near-surface range, indicated by XPS. The layers’ refractive indices were according to SE close to 2.08 (at 632 nm) with extinction coefficients negligible in the visible range. The TaxOy thickness ensuring maximum sensitivity was about 63 nm, when tested as an overlay for sensors based on optical fibers with induced long-period-grating patterns.

Resume : Beryllium being one of the most transparent to X-ray radiation elements has become the material of choice for X-ray optics instrumentation, applicable at synchrotron radiation sources and free electron laser facilities. However, there is a significant drawback in utilizing Be optics due to material oxidation and subsequent degradation under synchrotron irradiation [1]. Also there are serious health risks associated with a powder BeO which can be originated because of X-ray irradiation [2]. Improved properties of X-ray optics in terms of their surface quality and higher stability in the powerful X-ray beams are required as the high-brilliant fourth generation of synchrotrons are being developed now. In this work we propose to apply protective coatings in order to seal off beryllium from the ambient atmosphere. Al2O3 was chosen as a protective layer due to its passivation properties and low X-Ray absorption coefficient. To provide a good adhesion of the thin film towards the substrate and to make conformal deposition on objects with complicated shape we opted for atomic layer deposition (ALD) technique. Samples of beryllium coated with ALD Al2O3 were tested under monochromatic, pink and white beams at the European Synchrotron Radiation Facility (ESRF), in order to establish the conditions which the samples could tolerate without radiation damage. X-ray treatment was implemented at various environments – vacuum, helium, nitrogen, argon and dry air of different pressure. Mechanical study, post-process visualization and chemical analysis were performed to check the coatings efficiency. 1. Gmur N. (1992). Nucl. Instrum. Methods Phys. Res., A319, 228-232. 2. Gmur N., (1988). Nucl. Instrum. Methods Phys. Res., A266, 362-274.

Resume : The current trend towards miniaturized and autonomous electronic devices requires innovative energy storage solutions. For instance, autonomous micro-sensor networks or implantable medical devices would need a robust power source with high cyclability and a large power density, which might be out of the scope of conventional battery technologies. For such applications, microsupercapacitors (µSCs) are promising alternatives, and their integration “on-chip” could allow significant innovations. However, finding a suitable “on-chip” µSCs technology implies addressing key challenges, such as temperature resistance, silicon industry compatibility and good electrochemical performances on a small footprint. Nanostructures such as SiNWs and SiNTrs demonstrated excellent cyclability with more than 1 million cycles of galvanostatic charge/discharge under a 4 V wide electrochemical windows in EMI-TFSI ionic liquid, with large power densities and good capacitance values. Moreover, the use of silicon for electrode material allows extremely interesting developments towards “on-chip” integration and potential scale-up production using standard silicon industry processes for small micro-sized energy storage devices. Furthermore, we have also investigated the impact of the addition of a high-k dielectric layer, such as Al2O3 as protective films on silicon nanotrees. The electrochemical performances was enhanced, allowing symmetric 2 electrodes device to reach an unprecedented cell voltage of 5.5 V, improving energy and maximum power densities compared to unmodified nanostructured silicon. The cyclability was also largely enhanced, with only 3% capacitance fade after 1 million galvanostatic charge/discharge cycles at 4 V, and no degradation even after several 100000 cycles over 5 V.

Resume : One of the attractive features of atomic layer deposition (ALD) is that it works at much lower temperatures (e.g. 150-300°C) than comparable chemical vapour deposition processes, and this enables deposition onto thermally-sensitive substrates. In some cases, ALD can be achieved at even lower temperatures, but such temperatures often require plasma, as the thermal process may show too low a growth rate to be useful. The prototypical ALD process of trimethylaluminium (TMA) and water is one such example. Here the amount of alumina deposited in an ALD cycle drops steadily as the growth temperature is lowered below 200°C, whereas the growth rate is maintained even down to room temperature if TMA+O2-plasma is used. It has long been supposed that the reason lies in thermal activation being required for some reaction step within the overall ALD process. As outlined in this talk, recent experiments in various laboratories and calculations in our group have revealed what reaction step is involved, and this has modified our view of ALD surface chemistry. We first review experimental results from Eindhoven University of Technology that detect persistent methyl groups at the end of the H2O pulse in low temperature alumina ALD [1] and analogous results for zinc and tin oxide ALD from Stanford University [2]. We then highlight the particular surface reactions that were previously computed to show coverage-dependent kinetics for alumina and hafnia ALD and what types of surface species can consequently be expected to persist at the surface [3]. The theory could be confirmed if these species and their local environment could be detected by surface-sensitive techniques, such as those based on synchrotron X-ray radiation. We expect that these effects may be even more acute in the case of thermal ALD of nitrides with the less reactive co-reagent NH3. Finally, we contrast the thermal H2O mechanism with that predicted by DFT for TMA+O2-plasma. [1] V. Vandalon et al., Appl. Phys. Lett. 108, 011607 (2016). [2] A.J.M. Mackus et al., J. Chem. Phys. 146, 052802 (2017). [3] M. Shirazi et al., Nanoscale 7, 6311 (2015).

Resume : The good intrinsic conductivity of ALD-ZnO in the range of ~5 mΩcm is widely known. Furthermore, these films are highly transparent with only ~2.5% light absorption. This leads to the question if it is possible to further increase the conductivity by variations in precursor choice, deposition temperature and ALD parameters (pump & dose times). We compared the conductivity of ALD ZnO films at different deposition temperatures for the precursors diethylzinc (DEZ) and dimethylzinc (DMZ). DEZ shows a lower refractive index, lower deposition rate and lower conductivity over the whole investigated deposition temperature range. The variation of pump and dose times shows promising results for further increasing the conductivity and decreasing the process time. Additionally, the long-term stability of the conductivity of ALD-ZnO was investigated(1). Finally, the origin of the intrinsic conductivity itself is identified to be a donor state at the ZnO conduction band created by hydrogen incorporation during deposition(2). (1) Beh et al., Journal of Vacuum Science and Technology A, 35, 01B127 (2017) (2) Beh et al., J. Appl. Phys (under review)

Resume : InGaAs is a III-V semiconductor of great interest for high-electron-mobility transistors or high-speed electronics. A critical issue in such applications is reducing the contact resistance between the metal and the active semiconductor. One solution is to decrease the Sсhottky barrier height, by insertion of an ultrathin (~1 nm) tunneling insulator layer (such as ZnO) in between the metal and semiconductor [1], but this requires precise control over growth of the insulating layer. We have studied the incipient stages of ZnO growth by Atomic Layer Deposition (ALD) on In0.53Ga0.47As (001) substrates by in situ synchrotron X-ray characterization, using a custom-built ALD reactor that can be installed to different synchrotron centers [2]. We will show that the Zn fluorescence yield measured as a function of number of ALD cycles shows two growth regimes. One corresponds to the completion of an ultrathin ZnO (amorphous) layer (~1 nm) on atomically flat InGaAs, and the second to the growth of ZnO on top of this layer. These results are consistent with in situ X-ray absorption spectra (XAFS) and ex situ AFM images. Modeling of the film mean growth (FMG) rate helps to elucidate the different FMG curves as a function of the growth temperature. We will also demonstrate the existence of an amorphous to crystalline transition when the ZnO layer thickness is increased and provide results on contact resistance measurements. [1] dx.doi.org/10.1063/1.4813881 [2] 10.1021/acs.cgd.6b00844

Resume : Organic materials are key components of many emerging electronic devices such as solar cells and transistors. Their printability and flexibility make them attractive as carrier transport layers, insulators and for patterning. Inorganic materials such as ZnO are often used in combination with organics as semiconductors, and can also act as barriers to e.g. water vapour, increasing stability. Atomic layer deposition (ALD) is capable of growing ZnO at temperatures compatible with organics, but nucleation and growth varies with substrate due to the surface chemistry. Studies have previously reported examples of growth on polymer fibres, within bulk heterojunctions and on top of select materials. In this work we build on this by using a simple method to produce thin films of a large range of polymers and small molecules, and we employ top view transmission electron microscopy (TEM) to investigate the early ZnO growth behaviour. Analysis of crystallinity, nucleation density and element mapping enable us to draw comparisons, for example, growth on poly-4-vinylphenol is similar to on amorphous alumina due to the abundance of hydroxyl groups, producing a continuous, crystalline ZnO film with an orientation dependence on thickness. In contrast, materials such as PMMA and PCBM lead to island growth, resulting in porous ZnO layers with inferior barrier properties. TEM is combined with other characterisation techniques and the impact of growth behaviour on device performance is discussed.