Driven by this goal, we develop a computational framework that effectively acts like a virtual “stereological microscope” to visualize morphology evolution from early stages of phase separation until the formation of the stable morphology. This multiscale framework is based on a continuum description of evaporation-induced phase-separation in ternary systems and is able to resolve nano-morphological features while simulating device scale domains. Comparison of predicted morphology with experimental measurements shows the validity of the developed framework. We will showcase the potential of such coupled computational and experimental analysis by investigating morphology evolution in a specific class of organic photovoltaics (P3HT:PCBM:solvent). The framework is used to characterize the organization of percolating networks within the active layer as a function of (a) blend ratio, (b) evaporation profile, (c) solvent type, and (d) substrate patterning. We finally establish process-structure-property relationships using tools developed based on recent advances in image processing and computational homology.

Real-time techniques enable us to better study and understand the mechanisms of film growth, the evolution of structures and imperfections, as well as the possible appearance of transient structures and transformations, which can be crucial for the quality of the growing film and which would be missed in post-growth studies.

Particular emphasis is put on our experiments recording the scattered intensity in a broad range of q points simultaneously as a function of time as opposed to only one q point. We will explain how this can be used for an improved analysis of the data and the resulting structure.

We will also demonstrate how the optical properties of the films evolve and how they are related to the structure. Our experimental approach is quite general and applicable to many different materials, but examples will be mostly shown based on our work on organic semiconductors, which are applied in organic electronic and optoelectronic devices

(A. C. Dürr et al., Phys. Rev. Lett. 90 (2003) 016104; B. Krause et al., Europhys. Lett. 65 (2004) 372; S. Kowarik et al., Phys. Rev. Lett. 96 (2006) 125504; D. Zhang et al., Phys. Rev. Lett. 104 (2010), 056601; U. Heinemeyer et al., Phys. Rev. Lett. 104 (2010), 257401; A. Hinderhofer et al., Europhys. Lett., in print).

The crystallinity of simulated Mg_xM_yO_z (M=Al, Cr, or Y) thin films by a molecular dynamics (MD) model is studied with a variation in the stoichiometry of the thin films at operating conditions similar to the experimental conditions of a dual magnetron sputter-deposition system. The Mg metal content in the film ranges from 100% (i.e. MgO film) to 0%, (i.e. M₂O₃ film) [1, 2]. A classical ionic potential with formal charges describes the interactions between atoms. The structure of the simulated films was found to be in excellent agreement with the structure of the experimentally deposited films analyzed by X-ray diffraction (XRD) and transmission electron microscopy (TEM) techniques.

Both simulation and experimental results showed that the structure of the Mg-M-O film was varying from crystalline to amorphous when the Mg concentration decreases. The crystalline Mg-M-O films have a MgO structure (cubic, S.G. Fm3m) with M in solid solution.

The role of the surface diffusion in the growth process was studied by calculating the energy barriers for surface diffusion based on the classical potential used for the MD simulation. Some interesting results have been obtained about the mobility of species. For example, the energy barriers for surface diffusion are calculated to be lower in the Mg-Cr-O system compared to that in the Mg-Al-O system at the same Cr/Al content, which could explain the higher level of crystallinity observed in the Mg-Cr-O thin films compared to Mg-Al-O thin films with the same Mg-M metal ratio. In addition, increasing the other metal (Al, Cr, or Y) amount shifts the maximum of the energy barrier distribution to very low values (below 0.1 eV) which is a clear indication for a transition from an ordered to a disordered system.

[1] V Georgieva, M Saraiva, N Jehanathan, O I Lebedev, D Depla and A Bogaerts, J. Phys. D: Appl. Phys. 42 (2009) 065107.

[2] M. Saraiva, V. Georgieva, S. Mahieu, K. Van Aeken, A. Bogaerts and D. Depla, J. Appl. Phys. 107 (2010) 034902.

Thin films exhibit compressive-tensile-compressive stress states during the nucleation of islands, coalescence of islands and post-coalescence stages of growth. Using an in situ wafer curvature measurement technique, the stress evolution in Fe-Pt and Fe-Cu alloy thin films has been investigated. The stresses were shown to be compositionally dependent. In general, the tensile or compressive stress for the various binary compositions was associated with whichever element enriched the grain boundaries. Under specific growth conditions, a ‘zero-stress’ state could be achieved. The as-deposited alloy stress states do not show significant stress recovery upon ceasing the deposition. For the FePt films, upon annealing, the magnitude of the compressive stress state was reduced with increasing order parameter and has been explained in terms of reduced adatom surface migration. The FeCu films showed phase separation and an increase in surface roughness. Density functional theory calculations were performed to quantify the possible diffusion pathways and binding energies for Fe and Pt on a {111} L1₀ surface. Upon ceasing deposition, the post-growth stress relaxation rate increased with FePt order parameter and is explained in terms of an increase in interfacial energy contribution at the grain boundaries formed by chemically ordered grains. XRD, TEM, and atom probe tomography have been employed to quantify the phase, grain size and grain boundary chemistries, respectively, as they relate to the preferential segregation and thin film stress measurements.