Chapter 1
Development of Epitaxial Oxide Ceramics Nanomaterials Based on Chemical Strategies on Semiconductor Platforms

A. Carretero-Genevrier1*, R. Bachelet1, G. Saint-Girons1, R. Moalla1, J. M. Vila-Fungueiriño2, B. Rivas-Murias2, F. Rivadulla2, J. Rodriguez-Carvajal3, A. Gomez4, J. Gazquez4, M. Gich4 and N. Mestres4

1Institut des Nanotechnologies de Lyon (INL) CNRS-Ecole Centrale de Lyon, Ecully, France

2Centro de Investigación en Química Biológica y Materiales Moleculares (CIQUS), Universidad de Santiago de Compostela, Santiago de Compostela, Spain

3Institut Laue-Langevin, Grenoble Cedex 9, France

4Institut de Ciència de Materials de Barcelona ICMAB, Consejo Superior de Investigaciones Científicas CSIC, Campus UAB Catalonia, Spain

*Corresponding author: adrien.carretero-genevrier@ec-lyon.fr

Abstract

The technological impact of combining substrate technologies with the properties of functional advanced oxide ceramics is colossal given its relevant role in the development of novel and more efficient devices. However, the precise control of interfaces and crystallization mechanisms of dissimilar materials at the nanoscale needs to be further developed. As an example, the integration of hybrid structures of high-quality epitaxial oxide films and nanostructures on silicon remains extremely challenging because these materials present major chemical, structural and thermal differences. This book chapter describes the main promising strategies that are being used to accommodate advanced oxide nanostructured ceramics on different technological substrates via chemical solution deposition (CSD) approaches. We will focus on novel examples separated into two main sections: (i) epitaxial ceramic nanomaterials entirely performed by soft chemistry, such as nanostructured piezoelectric quartz thin films on silicon or 1D complex oxide nanostructures epitaxially grown on silicon, and (ii) ceramic materials prepared by combining soft chemistry and physical techniques, such as epitaxial perovskite oxide thin films on silicon using the combination of soft chemistry and molecular beam epitaxy. Consequently, this chapter will cover cutting-edge strategies based on the potential of combining epitaxial growth and CSD to develop oxide ceramics nanomaterials with novel structures and improved physical properties.

Keywords: Epitaxial growth, thin-film growth, silicon, perovskites, solution chemistry, molecular beam epitaxy, oxide nanostructures, magnetic oxide nanowires, quartz thin films, octahedral molecular sieves

1.1 Introduction

Single-crystalline thin films of functional oxides exhibit a rich variety of properties such as ferroelectricity, piezoelectricity, superconductivity, ferro- and antiferro-magnetism, and nonlinear optics that are highly appealing for new electronic, opto-electronic and energy applications [1, 2]. Over the past few years, tremendous progress has been achieved in the growth of functional oxides on oxide substrates (such as LaAlO3, SrTiO3, Al2O3, MgO, and scandates) [3, 4]. As a result, to date, it is possible to control the epitaxial growth at the unit cell level, which has led to new phenomena arising from the engineering of novel interfaces [5-8]. However, to fully exploit their properties, functional oxides should be effectively integrated on a semiconductor platform like silicon, germanium or III/V substrates, which are compatible with the electronics industry. The controlled epitaxial growth of functional oxide layers on semiconductor substrates is a challenging task as a result of the strong structural, chemical, and thermal dissimilarities existing between these materials. In spite of the difference in lattice parameters and thermal expansion coefficients, the major difficulty to engineer epitaxy is linked to the necessity of preventing the formation of an amorphous interfacial layer during the first stages of the growth (e.g. SiO2 or silicates on Si, depending of the atmosphere), which hinders any further epitaxy. Additionally, the cations of most oxide compounds can easily interdiffuse into the silicon substrate giving rise to the formation of spurious phases at the interface [9]. To overcome these major challenges, it is required to use a stable buffer layer, which can act simultaneously as a chemical barrier preventing ionic inter-diffusion and as a structural template favoring epitaxy.

In this context, McKee et al. [10] demonstrated the possibility to grow epitaxial SrTiO3 (STO) films on Si(001) by molecular beam epitaxy (MBE) with Sr passivation strategy. This work sets the basis to integrate STO and related perovskites on silicon for monolithic devices. Consequently, most of the research on crystalline functional oxides such as STO [11], lead zirconate titanate PbZr0.52Ti0.48O3(PZT) [12], BaTiO3 (BTO) [13-17], LaCoO3 (LCO) [18], and La0.7Sr0.3MnO3 (LSMO) [19] integrated with Si has been based on an STO buffer layer epitaxially grown on Si(001) by MBE.

For decades, the integration of functional oxides onto a silicon platform has been identified as an important route to improve and widen the performances of microelectronics and nanoelectronics devices. A clear example is the successful preparation of two-dimensional electron gas at interfaces between LaAlO3 and SrTiO3 (STO) on Si(001). In this case, the STO film acts simultaneously as a buffer layer and as an active part of the functional heterostrucuture [20]. Moreover, 2D electron gases at the interface have also been demonstrated using LaTiO3 [21] and GdTiO3 [22, 23] grown on STO-buffered Si. Functional non-volatile BTO-based ferroelectric tunnel junctions (FTJ) on Si(001) substrates with a tunneling electroresistance (TER) ratio over 10,000% have been recently demonstrated by pulsed laser deposition (PLD) [24] and MBE [25] growth methods. In both cases, this was accomplished by including a thin layer of STO as an epitaxial template on silicon. In addition, concomitant ferroelectric and antiferromagnetic behaviors were demonstrated on single-crystal BiFeO3 (BFO) films grown on STO on Si(100) using PLD [26] and MBE [27].

Integration of self-assembled vertical epitaxial nanocomposites thin films on Si substrates has been reported for multiferroic or magnetic memory and logic devices. The growth of La0.7Sr0.3MnO3-ZnO perovskite-wurtzite and CeO2-BTO fluorite-perovskite vertical nanocomposites on a Si substrate by PLD was described using a TiN/SrTiO3 bilayer buffer layer [28, 29]. The respective magnetoresistance and ferroelectric properties matched those of similar films grown on single-crystal STO. In addition, perovskite-spinel magneto-electric BFO-CFO vertical nanocomposites were successfully integrated on Si using two different buffered substrates: Sr(Ti0.65Fe0.35)O3/CeO2/YSZ/Si and 8 nm STO/Si [30].

The integration of functional oxides on germanium has recently received a great attention for high-speed and low-power device applications [31], as a result of the higher electron and hole mobility of germanium over silicon [32]. Indeed, a germanium-based ferroelectric field effect transistor was produced recently [33]. In this case, an ultrathin (20 Å) STO layer was first deposited on the Ge substrate. This layer imposes an in-plane compressive strain on BTO to overcome the tensile strain caused by the thermal expansion mismatch between both materials, therefore providing BTO films on Ge with out-of-plane polarization.

The development of freestanding oxide devices based on microelectromechanical systems (MEMS) technologies using standard silicon micromachining techniques was possible from SrTiO3/Si structures. Thus, the fabrication of integrated free-standing LSMO microbridges for low-power consumption pressure sensors [34] and uncooled bolometers [35] was recently demonstrated.

The direct growth of functional oxide film on silicon has proved to be also an effective way of integration without epitaxy. In this context, a field effect transistor preserving magnetoelectric functionality on a silicon-integrated device based on a La0.825Sr0.175MnO3/Pb0.2Zr0.8TiO3 (LSMO/PZT) bilayer directly grown by PLD on non-processed Si substrate has been demonstrated by Fina et al. [36]. The measured modulation of the magnetic and transport properties of LSMO upon PZT ferroelectric switching is large, despite the polycrystalline nature of the structure.

Yttrium-stabilized zirconia (YSZ) has also shown to be a very effective buffer layer to integrate functional oxide layers on Si(001) despite a lattice mismatch of about 5% and because it scavenges the native oxide on the substrate surface and reduces the native SiO2 oxide layer, with controlled oxygen partial pressure. These characteristics favor the formation of an epitaxial relation with the silicon substrate [37-39], thus making possible the integration of functional ferromagnetic spinel oxides [40-42] and ferroelectric perovskite oxides [43]. The use of an YSZ template substrate has also permitted the fabrication of all-oxide, free-standing, heteroepitaxial, and piezoelectric MEMS on silicon by using PbZr0.52Ti0.48O3 as the active functional material [44]. Recently, optimized growth conditions and subsequent functional oxides deposition have been shown on a silicon wafer scale (>4") using PLD [45].

The opportunities of combining functional oxides with integrated photonic devices and circuits are equally...