Prof. C. Richard A. Catlow FRS recently turned 65, and I attended a very stimulating symposium held in his honour yesterday at University College London.

The initial session combined talks from Richard's third PhD student (Prof. Steve Parker), experimental "sidekick" (Prof. Alan Chadwick), and a selection of his current research stars (Scott Woodley, Alexey Sokol, David Scanlon and Nikolaos Dimitratos). Richard ended the afternoon with a public science lecture covering five career-spanning topics (with a stream of publications in Proc. Roy. Soc. and Nature; the former used to me more prestigious than the latter!).

1. Forces between atoms
- The development and validation of interatomic potentials, beginning with a fluoride potential in 1972.
2. Fascination of disorder
- Non-stoichiometric oxides and the utility of EXAFS.
3. Facilitates and facilitation
- The development of synchrotron diffraction at Daresbury Laboratories.
4. Transforming molecules
-  Heterogeneous catalysis: modelling of barriers and pathways.
5. Impact and interdisciplinary
- How models of fission products; zeolite templating and planetary formation had wide-ranging impact.

The message of the day was "If the model has the right physics, you can't go too wrong".

Unless you want all of your non-existent free time soaked up for six months, don't buy a house.  The mountains of paperwork involved definitely make me feel better about university bureaucracy. More regular updates coming soon...

It has been a busy year in terms of university and network activities, but chemistry comes first. The first third of the year has been a mix of code developments, research reviews and new science.

• “Introducing k-Point Parallelism into VASP” A. Maniopoulou, E. R. M. Davidson, R. Grau-Crespo, A. Walsh, I. J. Bush, C. R. A. Catlow and S. M. Woodley, Computer Physics Communications 183, 1696 (2012).

VASP is the most used electronic structure code on the national supercomputer, but it cannot take full advantage of the thousands of available processors. This paper is the result of a collaboration between the HPC Materials Chemistry Consortium  and NAG  to add a new mode of parallelism into the code. It works, and will hopefully be adopted into the mainstream version of VASP soon. The  approach makes hybrid density functional theory calculations less painful to run!

• “Structural and electronic properties of CuSbS2 and CuBiS2: potential absorber materials for thin-film solar cells” J. T. R. Dufton, A. Walsh, P. M. Panchmatia, L. M. Peter, D. Colombara and M. S. Islam, Physical Chemistry Chemical Physics 14, 7729 (2012).

Solar cells work, but to enable widespread adoption of thin-film technologies, new absorber materials need to be developed that are made from earth abundant elements. Copper based compounds are the current hot topic, and this work presents a theoretical understanding of the structure and bonding in two candidate materials that have recently been synthesised in the Department of Chemistry at Bath.

• “Kesterite Thin-Film Solar Cells: Advances in Materials Modelling of Cu2ZnSnS4” A. Walsh, S. Chen, S.-H. Wei and X. G. Gong, Advanced Energy Materials 4, 400 (2012).

People are excited about CZTS solar cells, and with good reason. While the elements are abundant and low cost, the challenge is in the complex solid-state chemistry associated with a four-component system. This paper reviews our work performed over the past five years exploring the materials chemistry and physics of the kesterite system.

• “Synthesis, Characterization, and Calculated Electronic Structure of the Crystalline Metal–Organic Polymers [Hg(SC6H4S)(en)]n and [Pb(SC6H4S)(dien)]n” D. L. Turner, K. H. Stone, P. W. Stephens, A. Walsh, M. P. Singh, and T. P. Vaid, Inorganic Chemistry 51, 370 (2012).

The development of semiconducting metal-organic frameworks is the primary subject of my recently funded European grant. To complement our predominately theoretical research, we are collaborating with the group of Tom Vaid at the University of Alabama. Novel lead and mercury sulfide networks that show some real promise were synthesized and characterised.

•  “Surface structure of In2O3(111) (1×1) determined by density functional theory calculations and low energy electron diffraction” K. Pussi, A. Matilainen, V. R. Dhanak, A. Walsh, R. G. Egdell, K. H. L. Zhang, Surface Science 606, 1 (2012).

It may be increasingly rare and expensive, but indium makes my favourite oxide material. So simple yet so complex. Last year we predicted the surface structure of In2O3 based on my calculations, and this year we managed to get experimental validation by one of the largest LEED studies performed to date.

Academic genealogy is definitely easier to trace back than family, especially when most of the people involved have entries on Wikipedia. How much knowledge gets passed down the chain?

I often find myself at the interface between chemistry and physics, which is mirrored by the split between Richard Catlow and Marshall Stoneham. One path leads to the birth of computational chemistry in the United Kingdom (Lennard-Jones and his ubiquitous inter-atomic potential), while the other includes some of the pioneers of scientific computing (von Neumann) and quantum physics (Pauli). Tracing the links back even further towards the 19th century leads to Sommerfeld (theory of electrons in metals) and towards the 18th century to the polymath Gauss (Gaussian distributions and Gauss' Law).

This group would make for an interesting Christmas dinner, although, I think I'd end up feeling like the country cousin.

For my research, Mac offers a nice compromise between the functionality of Windows and the power of Linux.  Chemistry comfort software like Chemdraw, Endnote, Mendeley, MS Office and LaTEX are all available. For scientific computing, there are a few quirks that you need to get around:

(i) Where is .bashrc? It hides in /etc/bashrc. For a splash of colour, add export TERM=xterm-color to your .bash_profile.

(ii) Package managers: no Yum, but Macports and Fink work well.

(iii) How to hide folders from Finder? Away from the command line, some folders really should remain hidden from view. One good example is the 'Microsoft_User_Data’ folder. Solution: SetFile -a V [folder]

(iv) Free useful applications? Mendeley (references); Vesta (visualisation); Inkscape (vector art); Gimp (bitmap art); Open Babel (structure conversion);  GeoGebra (geometric fun); Texmaker (LaTex GUI).

(v) Free C and Fortran compilers? You would expect to have standard compilers installed by default. Well no, but fortunately the good people at Mac HPC and Mac Research are there to help. First you need to install XCode (from the AppStore, or directly from here), and for the newer versions you need to install the command line tools add-on separately. This sets up a basic gcc compiler. Binaries for the latest versions of gcc/gfortran are available here. Be sure to download both binaries, and remember to add the path to your shell: export PATH=/usr/local/bin:$PATH

(vi) Materials modelling codes (FHI-AIMS, VASP, GULP)? For GULP, an OSX binary is available. Once gfortran is installed, you can download and compile standard Unix distributions of GotoBLAS and LAPACK. For FHI-AIMS, the makefile is straightforward, and the binary runs perfectly:

FC = gfortran -m64
FFLAGS = -O2 -ffree-line-length-none
F90FLAGS = $(FFLAGS)
ARCH = arch_generic.f90
LAPACKBLAS = /progs/lapack-3.2.1/lapack_LINUX.a /progs/GotoBLAS2/libgoto2-r1.13.a

For VASP 5, the following gives a serial binary that works quite well:

FC = gfortran -m64
OFLAG  = -O2
FFLAGS = -ffree-form -ffree-line-length-none
BLAS = /progs/GotoBLAS2/libgoto2-r1.13.a
LAPACK = /progs/lapack-3.2.1/lapack_LINUX.a

In my experience OpenMPI works well, with efficient use of four cores on i7 iMacs, and as expected for the premium price, the Intel fortran compiler (and MKL libraries) result in a significant (~20%) performance boost.  Unfortunately, no non-commercial version of ifort is available for Macs.

Two hydrogen atoms walk into a bar.
One says “I think I’ve lost an electron.”
The other says, “Are you sure?”
The first replies, “Yes, I’m positive…”
Anon.

Due to the loss of Apple's MobileMe service, my former blog Computational Alchemist is now reborn as a proton after a six month hiatus.