Royal Society of Chemistry Theoretical Chemistry Group

25th Anniversary Graduate Student Meeting

University of Nottingham, April 13th, 2005

The annual meeting for presentations by final year graduate students in theoretical chemistry will be held on Wednesday 13th April in room C17 of the School of Chemistry, University of Nottingham, starting at 11 a.m. Travel details can be found here and a map of the campus here.

This year marks the 25th anniversary of the first graduate student meeting organised by the Theoretical Chemistry Group. To celebrate this occasion, in addition to the talks by final year postgraduate students in theoretical/computational chemistry, we will be having a poster session for second year postgraduate students and there will be plenary lectures by three of the speakers from the first TCG graduate student meeting in 1980: Professor Patrick Fowler (Exeter); Professor Sally Price (UCL); and Professor Jonathan Tennyson (UCL).


11:00Prof.Ian H. WilliamsIntroduction
11:00Prof. Patrick W. Fowler (Exeter)Fullerenes and their exotic friends
11:30Andy May (Bristol)Density fitting in explicitly correlated electronic structure theory
11:55Thomas W. Keal (Durham)Semi-empirical fitting of exchange-correlation functionals
12:20LUNCH(including poster session and tour of computing facility)
14:00Prof. Sally L. Price (UCL)Organic crystal structure prediction: an application of computational chemistry that can make useful predictions?
14:30Mahesh Sundararajan (Manchester)Structure and function of the iron-sulfur protein Rubredoxin, studied by QM/MM methods
14:55Bob Barber (UCL)A new high accuracy water line list: computation and applications
15:20Susan Larkin (Imperial College)Excited states, conical intersections and the ONIOM method
16:15Adam Kirrander (Oxford)H2: a simple molecule with complex dynamics
16:40James Munro (UCL)H3+ to dissociation and beyond
17:05Mohammad Noh Daud (Bristol)Photodissociation of N2O
17:30Prof. Jonathan Tennyson (UCL)Molecular physics and the mass of the neutrino
18:00Announcement of Coulson Prize 2005  


Fullerenes and their exotic friends

Patrick W. Fowler (Exeter)

Fullerenes are n-atomic carbon polyhedra with 12 pentagonal and (n/2-10) hexagonal faces. Their electronic configurations and many of their properties are usefully predicted by purely graph-theoretical arguments. Fullerenes are essentially spherical, but why should Nature stop at the sphere? Fullerene-like graphs can be embedded as maps in four closed surfaces (sphere, torus, Klein bottle, projective plane), all with characteristic Huckel rules derived by topological double-cover arguments, and some with experimental realisations. An extended argument shows the general relationship between energy levels/orbitals of Moebius and cylindrical conjugated systems, with some surprising results.

Organic crystal structure prediction: an application of computational chemistry that can make useful predictions?

Sarah (Sally) L. Price (University College London)

When 25 years ago, I was researching the intermolecular forces between hydrogen molecules, it seemed unlikely that accurate intermolecular potentials would ever be developed for organic molecules. However, now model potentials, partially derived from the molecular charge density, form a major part of the input into searches for energetically feasible crystal structures. The practical importance of predicting organic crystal structures and polymorphism means that theoreticians working in this area have been challenged to predict crystal structures given only the chemical diagram. Whilst there have been some successes in these blind tests, and indeed some crystal structures have been predicted before they have been found experimentally, there is still a long way to go before we have a quantitative theory that can predict the crystallisation of organic molecules reliably.

Molecular physics and the mass of the neutrino

Jonathan Tennyson (University College London)

It is now established that neutrinos have a mass but only an upper limit for this mass is known. The best prospect for a laboratory determination of the mass of a neutrino appears to be tritium endpoint experiments. The KATRIN collaboration is preparing new, high sensitivity measurement based on precise measurements of tritium molecule decays. These measurements can only be interpreted using detailed information on the (molecular) products of the decay. Progress on calculations which (a) characterise the decay processes, (b) allow for errors in both the temperature and isotopic composition of the sources, and (c) consider the effects of other possible processes in the tritium source will be presented.

A new high accuracy water line list: computation and applications

Robert J. Barber (University College London)

Water is the third most common molecule in the Universe. Its spectrum is particularly rich and H2O lines have been detected in locations as diverse as: the ISM, comets, Mira variables, sun spots, brown dwarfs and red giants. Nevertheless, only a very small fraction of the lines in the water spectrum are known experimentally, and if for no other reason than sheer number, it seems inevitable that astronomers will continue to use theoretical data when exploring the nature of these objects. All previous water line lists suffer from incompleteness and inaccuracy (particularly at higher energies). We have produced a new list, BT2, that addresses these shortcomings. Our ab initio computation has employed the best available potential energy and dipole moment surfaces to perform quantum mechanical nuclear motion (rotation-vibration) calculations using the DVR3D suite of codes. The preliminary calculations which established the optimum parameters for the basis function and the ro-vibrational basis set were also fundamental in producing accurate line positions. The BT2 water line list contains in excess of 500 million transitions. It will be published shortly [1], but it is already finding laboratory [2] and astronomical applications [3,4].

[1] R. J. Barber, J. Tennyson et al. (in prep.).
[2] P.-F. Coheur et al. J. Chem. Phys. 122 (2005).
[3] N. Dello Russo et al. ApJ 621 (2005).
[4] D. P. K. Banerjee, R. J. Barber et al. (in prep.).

Click here for a pdf version.

Semi-empirical fitting of exchange-correlation functionals

Thomas W. Keal (Durham)

We describe our continuing work to improve the quality of semi-empirical exchange-correlation functionals in DFT by fitting to a diverse range of properties with a large number of training molecules. This highly empirical approach can result in significant improvements in accuracy over competing generalised gradient approximation (GGA) and hybrid functionals. Previous studies [1,2] have shown that a simple gradient correction to the local density approximation can provide NMR shielding constants for challenging main-group nuclei that are 2-3 times more accurate than those of conventional GGAs and hybrids. These functionals were denoted KT1 and KT2. By adding one additional exchange term, the KT3 GGA functional was developed [3], which retained the quality of shielding constants while substantially improving the performance for other properties. In particular, it gives atomisation energies that are as accurate as the best competing GGA functionals and equilibrium molecular bond lengths that are as accurate as the best hybrid functionals. However, its performance for classical reaction barriers is relatively weak. We describe our attempts to develop new GGA and hybrid functionals that give improved reaction barriers while retaining the high quality performance of KT3 for other properties.

[1] T.W. Keal and D. J. Tozer, J. Chem. Phys. 2003, 119, 3015.
[2] M.J. Allen, T.W. Keal and D.J. Tozer, Chem. Phys. Lett. 2003, 380, 70.
[3] T.W. Keal and D.J. Tozer, J. Chem. Phys. 2004, 121, 5654.

Click here for a pdf version.

H2: a simple molecule with complex dynamics

Adam Kirrander (Oxford)

The hydrogen molecule may be the simplest of molecules, but its highly complex and non-adiabatic dynamics is far from simple. Due to the low mass of the nuclei, separation into electronic, vibrational and rotational energy becomes difficult for excited states and the energy levels are much better described as mixed rovibronic levels. Furthermore, highly excited rovibronic levels couple to the ionisation and dissociation continua. These states have a limited lifetime, which gives them an inherent line width in spectral measurements. We apply Multichannel Quantum Defect Theory (MQDT) and R-matrix theory to calculate and interpret preionised and predissociated hydrogen spectra. Furthermore, we show how to obtain branching ratios for the dissociative reactions. These branching ratios are highly relevant to astrophysical processes in diffuse interstellar clouds. Finally, we intend to emphasise how our theoretical treatment gives a physically intuitive understanding of the dynamics of highly excited molecules in general.

Click here for a pdf version.

Molecular dynamics simulation of peptide-protein interactions

Susan Larkin (Imperial College London)

The use of ab initio methods to find conical intersections can be very time-consuming and if the molecule is too large, it is impossible. Using a hybrid method, ONIOM, we have been able to successfully locate conical intersections geometries in a fraction of the time of the ab intio method. As this presentation illustrates, one ONIOM method in particular was successful at reproducing accurate relative energies between the conical intersection geometry and the s1 minima.

Density fitting in explicitly correlated electronic structure theory

Andy May (Bristol)

Accurate quantum chemical calculations are limited to small molecules partly owing to the poor convergence of orbital based methods. One solution to this problem is to use explicitly correlated electronic structure theories [1]. These methods give very accurate energies for small molecules, but become computationally difficult for larger systems. The problem arises from the need to compute three- and four-electron integrals, which are both complicated and extremely numerous. The R12 methods developed by Kutzelnigg and Klopper [2-4] use Resolution of Identity (RI) to approximate these integrals and have allowed calculations on chemically significant problems.

In this work we develop a version of MP2-R12 theory for an arbitrary correlation factor and derive recurrence relations for the integrals needed when the correlation factor is a frozen Gaussian geminal [5]. We show that the accuracy of MP2-R12 methods can be greatly improved by replacing the conventional r12 factor with a function of r12 better able to describe the shape of the correlation hole. It is shown that the errors associated with using linear r12 (rather than something better) are at least an order of magnitude greater than the errors associated with RI or other approximations in the MP2-R12 class of methods. The impact on our aim to do accurate quantum chemistry in reasonably compact orbital basis sets is assessed.

[1] E. A. Hylleraas, Z. Phys. 54, 347 (1929).
[2] W. Kutzelnigg, Theor. Chim. Acta 68, 445 (1985).
[3] W. Klopper and W. Kutzelnigg, Chem. Phys. Lett. 134, 17 (1987).
[4] W. Kutzelnigg and W. Klopper, J. Chem. Phys. 94, 1985 (1991).
[5] A. J. May and F. R. Manby, J. Chem. Phys. 121, 4479 (2004).

Click here for a pdf version.

H3+ to dissociation and beyond

James Munro (University College London)

In the 1980's Carrington et al [1] recorded a C02 laser predissociation spectrum of H3+. The spectrum proved to be remarkable and quite unique. Over 20 years later and some progress has been made but the spectrum remains completely unassigned. Recently we have made progress in understanding the dissociative behaviour of H3+ by performing very large calculations on a supercomputer using the PDVR3D program [2]. These improve on previous calculations [3,4] in a number of ways. We have discovered a series of asymptotic vibrational states, states with very long range behaviour which could have important consequences for the spectroscopy of H3+.

[1] A. Carrington I.R. McNab, Acc. Chem. Res. 22 (1989) 218-222.
[2] H.Y. Mussa, J. Tennyson, Comp. Phys. Comm. 128 (2000) 434-445.
[3] M.A. Kostin, O.L. Polyansky, J. Tennyson, J. Chem. Phys. 118 (2003) 3538-2542.
[4] V.A. Mandelshtam, H.S. Taylor, J. Chem. Soc. Fara. Trans. 93 (1997) 847.

Photodissociation of N2O

Mohammad Noh Daud (Bristol)

Click here for pdf abstract.

Structure and function of the iron-sulfur protein Rubredoxin, studied by QM/MM methods

Mahesh Sundararajan (Manchester)

Iron-Sulfur proteins are an important class of electron transfer protein acting as electron sinks for many biological reactions. One such iron-sulfur protein is Rubredoxin (Rd) which has a high spin iron centre, tetrahedrally coordinated to four cysteine sulfurs. We have explored the redox chemistry of Rd using both density functional theory (DFT) and semi-empirical methods. These latter methods require quite modest computing resources, but their accuracy is often suspect, particularly when the molecule contains a transition metal atom. However, a possible way forward is to develop parameters for use in semi-empirical MO studies that are tailored for a particular chemical situation, the so-called specific reaction parameters SRP [1]. We first describe a possible strategy to extend this approach to transition metal complexes [2]. Based on fitting to DFT data for redox site analogue, a PM3 parameter set for iron has been developed, which is appropriate for the active site of iron sulfur proteins having a single iron atom, and further tested on similar molecules. The use of these parameters within a two layer ONIOM treatment of the protein Rd, yields accurate predictions of the effect of the protein on both Fe-S bond lengths and inner sphere reorganization energies [3]. We also describe calculations on similar proteins in which the central metal atom or the cysteine ligands are mutated.

[1] Rossi, I., and Truhlar, D.G. Chem. Phys. Lett. 1995, 233, 231-236.
[2] Mohr, M., McNamara, J.P., Wang, H., Rajeev, S.A., Ge, J., Morgado, C.A., and Hillier, I.H. Faraday Discuss. 2003, 124, 413-428.
[3] Sundararajan, M., McNamara, J.P., Hillier, I.H., Wang, H., and Burton, N.A. Chem. Phys. Lett. 2005, 404, 9-12.

Click here for a pdf version.