Functional films - Thin film technological breakthrough
In industry, there is a demand for coatings to provide wear resistance, as well as films that impart new and enhanced functionality (conductivity – both electronic and ionic, ferroelectric, and dielectric). This is a direct consequence of recent scientific and technological breakthroughs in microelectronics, optics and nanotechnology. Research and development of thin films is a key platform for the production of new materials and devices at the London Centre for Nanotechnology, University College London, UK.
Thin films are material layers that range in thickness from fractions of a nanometre to several micrometres. The distinguishing feature compared with thick films is that their growth is expected to be epitaxial with the chosen substrate and they are cheaper to manufacture, owing to reduced material, energy, handling and capital costs.
Technologies are being developed to substantially reduce the weight and dimensions of integrated systems. Certain properties are greatly enhanced in thin form because of the epitaxy – the surface resistance and the critical current density of a high temperature superconductor in which properties of the thin film are orders of magnitude better than in a bulk counterpart. Usually the deposition is performed in vacuum, which prevents contamination. Thin film deposition techniques fall into three general groups – chemical deposition (chemical vapour deposition, plating), physical deposition (evaporation, sputtering, ablation), and a mixture of both (reactive sputtering, molecular beam epitaxy).
Pulsed laser deposition
Pulsed laser deposition (PLD) is a versatile technique where a pulsed laser rapidly evaporates a target material forming a thin film that retains target stoichiometry. At Imperial College London, UK, the Physical Electronics and Thin Film Materials Group uses a Neocera PLD system, consisting of two fully automated vacuum chambers with a 248nm KrF Lambda Physik laser shared between them. Each chamber has six positioned target carousels enabling in situ deposition of up to six different materials.
One of the disadvantages of barium strontium titanate (BSTO) films is the temperature dependence of their properties. The Curie temperature of a BSTO film is determined mainly by the barium:strontium ratio. So by altering the stoichiometry ratio it is possible to tailor a composition with specific properties (relative dielectric constant, tuning ratio and dielectric loss) for set temperature ranges. Barium strontium titanate deposition usually involves a single target with the required stoichiometry. However, this approach becomes expensive, in terms of the number of targets used, when a multilayered structure with different barium:strontium ratios is required and is impossible for BSTO-based graded structures.
A new method has been developed that enables the stoichiometry of the thin films to be engineered during deposition. The advantage of this was shown by a Ba0.75Sr0.25TiO3 thin film deposited on lanthanum aluminate substrate using a single target with corresponding stoichiometry. Another sample with the same stoichiometry was deposited as a periodic structure of BaTiO3 (BTO) and SrTiO3 (STO) layers.
The layer thickness was chosen so as not to exceed the critical thickness for epitaxial growth. To obtain a specific barium:strontium stoichiometric ratio in the sample, the thicknesses of the BTO and STO layers were adjusted. Both samples were made under the same conditions, maintaining the same total thickness.
High-angle annular dark field imaging and energy dispersive X-ray spectroscopy were carried out using a Titan 80-300 scanning transmission electron microscope (STEM). The result is presented below left.
Good separation is shown between the STO (dark part, thickness ~1nm) and BTO (bright part, thickness ~4nm) layers, which was confirmed by elemental analysis – the maxima of strontium corresponds with the minima of barium. Compared to a conventional BSTO film with corresponding stoichiometry, the multilayered structure exhibits broader phase-transition and improved thermo-stability. The electrical tunability (n = C(0)/C(V)) and the dielectric loss were also improved.
Magnetoelectrics and artificial supercells
Research has also focused on optimising magnetoelectric coupling. The first approach uses the growth template (substrate) as an active material rather than just a growth template. The substrate must respond to external stimuli in accordance with the film response, and to benefit film properties. This suppresses clamping of the film and substrate, which affects the film/substrate interface that inhibits the film’s functional properties.
The second approach is to develop multiferroic films to act as an artificial supercell. Ferroic properties alter dramatically when the dimensions of each layer or the total film thickness are reduced below sub-micron level. Strain, surface energy and structural defects are driving forces that, by influencing ferroic domain structure and spin and dipole polarisation, govern the behaviour of ferroic properties in ultra-thin films (<20nm). Interfacial defects that cause relaxation of the film growth, substitution altering the ordering due to off-centering, and pinning or growth of domains causing alteration in switching mechanisms, can all be used in this thin-film engineering.
Work is being conducted on the magneto-electric thin film system with piezoelectric substrates such as barium titanate or PZT, which responds to an external electric field and reduces the normal clamping effect of the substrate. For CoFe2O4 films deposited on BaTiO3 substrate, the strain developed in the substrate when an electric field is applied will be directly transferred to the magnetostrictive CoFe2O4 layer without any reduction in electric displacement, which would occur if any other non-active substrate was chosen.
To realise the appropriate condition for domain switching in the magnetostrictive layer, coercivity of the CoFe2O4 needs to be reduced. Epitaxial thin films of CoFe2O4 have been grown on BaTiO3 and it was found that magnetisation can be kept high, as in bulk values, but with a dramatically reduced coercive field when approaching film thicknesses of 13nm. Despite the enormous lattice mismatch between CoFe2O4 and BaTiO3, it is possible to grow a highly ordered epitaxial film when the parameters of the PLD are adjusted. This unexpected epitaxy, with almost no strain (<0.2%), is governed by interfacial defects which relax the structure. In addition, the unit cell of a highly flexible crystal structure of CoFe2O4 can be compressed in all three axes while retaining the cubic symmetry. In future, even thinner films will be examined, below 10nm, to investigate spin-coupling responses on the unit-cell compression and the magnetic response. How the parallel top and bottom interfaces influence the ferroic properties also needs to be explored.
The aim is to continue taking the cobalt ferrite into an artificial multiferroic supercell system, which will couple two or more switchable states such as polarisation and magnetisation. We would like to develop thin films with highly periodic nano-layered structure, in which the alternating ferroelectric and ferromagnetic layers will form an artificial single phase superstructure.
Attractive alternatives to inorganics for optoelectronic uses are molecular materials, due to their low cost, weight, and potential to modify their properties by inserting functional groups by chemical synthesis. Their semiconducting properties are due to delocalised π-orbitals, more often C-C double bonds, which lower the bandgap. Work at Imperial focuses on commercially available polyaromatic molecules, such as phthalocyanines (Pcs) and perylene derivatives.
Those molecules can be sublimed in the vapour phase, leading to the formation of high purity films. Two main techniques are used. In organic molecular beam deposition (OMBD) – the source material is enclosed in a crucible (knudsen cell) in a high vacuum chamber (10-6-10-10mbar depending on applications) and heated until it reaches its sublimation point. Although epitaxy can be observed when strongly interacting crystalline substrates are used, this rarely leads to practical applications, contrary to PLD. Molecular thin films can be deposited onto any substrate, paving the way for transparent flexible lightweight plastic electronics (see Materials World, February 2006, p18).
Films are usually amorphous or polycrystalline, and the typical morphology of a CuPc film deposited on a glass substrate at room temperature form a herringbone structure, where their molecular planes are stacked parallel to each other due to π-π interactions. Growth conditions can be modified and, after annealing of the film, the crystallites become elongated and the structure is modified, with the angle between the stacking axis and the molecular pane decreasing from 65 to 45º. This phase transition has a profound influence on the film’s properties.
Another method that has recently been developed is organic vapour phase deposition. This addresses the high cost associated with OMBD as it is performed in a vacuum between 1-10-2mbar, avoiding the need for expensive pumps and fittings. Due to the larger choice of parameters that can be modified in this technique, new structural and morphological properties can be obtained, where highly anisotropic crystallites of CuPc are observed.
Applications of thin films have focused on photovoltaics and, more recently, magnetism. Organic solar cells based on CuPc have attained promising conversion efficiencies, and values of five per cent have been reported. The control and versatility of sublimation methods has allowed us to concentrate on modifying device architecture to improve efficiencies. For example, we have doubled the efficiencies of a simple donor/acceptor bilayer structure by moving towards a gradient cell, where the stoichiometries of donor and acceptor vary continuously. This formed a composition gradient between the electrodes, creating a large interface that improves photon harvesting. Imperial’s studies in magnetism are based on metal-Pc crystals, which display an unpaired electron and, therefore, spin on their central metal. The magnetic couplings between the metals can be switched depending on the polymorph adopted by the crystals. In the case of CuPc, a switching between antiferromagnetic and paramagnetic is observed. Although these effects only occur at low temperatures, this opens up exciting perspectives on information technologies at the molecular level.
Thanks to co-authors Sandrine Heutz, Hai Wang, Peter Petrov, Anna-Karin Axelsson and Matjaz Valant from Imperial College.
Professor Neil Alford, Chair in Materials, Imperial College London Exhibition Road, London, SW7 2AZ. Tel: +44 (0)20 7594 6724. Fax: +44 (0)20 7594 6757. Email: firstname.lastname@example.org.
To view detailed images, graphs and diagrams referred to in this feature, please download the PDF of the original article using the link below.