TCM
UoC crest

History

Although the Cavendish was founded as a laboratory of experimental physics, it was amongst the first to establsh a dedicated theory group with a focus on the quantum physics of condensed matter. The first 40 years of its history have been documented by Volker Heine, a member of the Cavendish for more than 60 years, and a founding member of the theory group. A different, and narrower, perspective on TCM's history can be found on our page of our computing history. Separately we list particularly eminent alumni.

Nevill Mott was our father. He arrived in 1954 from Bristol as Cavendish Professor, and brought John Ziman in the same year as a new lecturer to start a group on Solid State Theory (or SST for short). Of course the Cavendish had been traditionally the Department of Experimental Physics, with just one or two isolated theoreticians in experimental groups. The main effort in theoretical physics in Cambridge was centred in what is now called the Department of Applied Mathematics and Theoretical Physics (or DAMTP), but no one there had any interest in ("microscopic") solid state physics.

However, experience before and during World War II, and developments in the USA, had shown Mott the importance of a strong theory group allied to the experimental effort. Furthermore solid state physics was becoming a leading area of research in the Cavendish, with the work of David Shoenberg and Brian Pippard on mapping out the Fermi surfaces of metals for the first time, Pippard introducing the concept of coherence length in superconductivity, David Tabor's work on surface physics, crystallography (including the recent work of Francis Crick and James Watson on DNA), Peter Hirsch turning the electron microscope onto seeing dislocations, and more. So it made sense to start a group on solid state theory in the Cavendish.

Mott was Cavendish Professor (1954-71) and was never really part of the theory group, though of course his presence, research interests, visitors, and an occasional student including Nick Rivier (housed within SST) were very important. He was not a "groupie" so that he is hardly mentioned further in this history, but his influence was always there. In fact, when Volker Heine arrived in 1954 to undertake his PhD with Mott, he did not realise that Mott did not really like having research students, preferring to have his own ideas in conjunction with more senior experimentalists. Mott tried to palm him off onto others around the Cavendish! But after thinking about it for a day, Heine felt he had not come half way around the world to be put off as easily as that. He therefore promised to be no problem as long as he could have access when he needed it, and so Mott gave him his project, which was "Why don't you go down to the Low Temperature Group and see if you can make yourself useful." That was Mott's concept of a theoretician, and it accorded with Heine's New Zealand culture. Mott's research interests at this time centred on dislocations, electrons in disordered materials, and the metal-to-insulator transition. Notable visitors that he brought to the Cavendish (or who came to him) included Hans Bethe, Philippe Nozières, and Phil Anderson.

In Oxford Ziman had been working on magnetism, but in Cambridge he turned his attention towards the electron theory of solids. In order to learn the subject thoroughly, he started writing his monumental tome Electrons and Phonons: The theory of transport phenomena in solids, finding there many research topics as he went along, mainly on transport theory. Research on electron transport in disordered media was very topical, and led to Ziman's calculation of the resistivity of molten sp-electron bonded metals, which was amazingly successful for reasons we understand better now.

A theory of the electrical properties of liquid metals. I: The monovalent metals J. M. Ziman, Philosophical Magazine 6 1013-1034 (1961)

Alongside Ziman, the other senior member of the group was Heine himself, who had gained a fellowship in 1957 and a Junior Lectureship in 1958. Among the research students in the group were Lu Sham (from Hong Kong), Neil Ashcroft (from New Zealand), Maurice Rice (from the Irish Republic), Leo Falicov (from Argentina) and Federico Garcia-Moliner (from Spain), whose subsequent contributions are now well known. What is noticeable is that none of them came from Great Britain, as was true of almost all the group at that time, including Ziman and Heine. Young Cambridge theorists wanted to pursue field theory, Green function methods, and nuclear physics.

Ziman was active on a broader front. He introduced the first systematic graduate lectures in the Cavendish, which later evolved into his book, Principles of the Theory of Solids, and he organised a colloquium series across the whole of the solid state effort. Indeed graduate students at that time were a rather neglected group in Cambridge, and Ziman at King's College became the first ever "tutor" (having administrative responsibility in College) specifically for graduate students. The late 1950s and early 60s were a time of considerable change in Cambridge (and indeed the whole of the University sector in Britain), with many adjustments needed in the era that followed World War II. Indeed there were threats of another Royal Commission "to clean up the mess in Oxbridge". Ziman, and his friend Jasper Rose, edited The Cambridge Review, turning it from a largely literary journal into a mouthpiece for their policies of reform. With every issue, another "sacred cow" slaughtered, seemed to be their motto. And it was great fun! Of course the University and Colleges did gradually embrace change, with revisions of admissions policy, the founding of graduate colleges, and much else; and there was no Royal Commission. I believe that Ziman and Rose are owed a great debt of gratitude for helping to focus the need for change when such issues did not always seem very clear at the time. Ziman's time of influence in Cambridge ended with his move to Bristol in 1964.

From the late 1950s Mott began to build up many-body theory at the Cavendish. The first step in this process was an invitation to Bethe to spend 1955-56 in Cambridge. His lectures, and the presence of Jeffrey Goldstone as a graduate student, led to considerable discussion. The next step was an invitation to Anderson to spend a year (1961-62) in Cambridge. Anderson gave the lectures that eventually, with the help of Sham's notes, led to the primer, Concepts in Solids. In these lectures Anderson repeated the orthodox remark that the phase of the superconducting wavefunction could not be measured. This was challenged by the young Brian Josephson, who had already acquired a formidable reputation. Anderson soon saw the strength of Josephson's arguments that, despite the skepticism of some authoritative members of the community, were verified experimentally, and rewarded by Josephson's 1974 Nobel Prize. (Brian Josephson joined the group in 1967, having previously been with Brian Pippard in Low Temperature Physics, then called the Mond Laboratory.) Anderson also led David Thouless, who was then (1961-64) a Lecturer in DAMTP, but housed on the top floor of the Austin Wing of the Cavendish Laboratory, to develop the theory of the nuclear magnetism of solid 3He, and the two maintained an intermittent but fruitful collaboration over the next twenty years.

Supercurrents through barriers B. D. Josephson, Advances in Physics 14 419-451 (1965) Exchange in solid 3He and the Heisenberg Hamiltonian D. J. Thouless, Proc. Phys. Soc. 86 893 (1965)

The continuing rise of many-body theory: The next step in Mott's plan to build up many-body theory involved a grant that brought several well-known theorists for extended visits. Among those who came under this programme were Philippe Nozieres, Leo Kadanoff, John Wilkins, and Karl Bennemann. These visits produced a lot of lively discussions, and were particularly helpful to those students who were already using many-body methods to study electrons in real solids. At this time Kadanoff was working on the application of renormalization techniques to study critical phenomena, work which prefigured the spectacular breakthrough made by Ken Wilson and Michael Fisher six or seven years later. When Wilkins was talking, his voice could be heard clearly from any point on the top floor!

Thouless left DAMTP to become a lecturer in Physics and an official member of SST (1964-65), but left just one year later to go to Birmingham University. Later, in the 1970s and 80s, he made several longer and shorter visits as a Royal Society Research Professor. The process of building up many-body theory in the Cavendish culminated in the appointment of Anderson as a Professor in 1967. One of Mott's and Pippard's actions was to persuade the University to create the professorship "Physics 1966" for solid state theory. P.W. (Phil) Anderson arrived in 1967 on a half-time basis, shuffling between Cambridge and Bell Labs. Anderson was formally Head of the Group, but Heine ran it on an administrative basis. They had a pub lunch every Wednesday during term time to coordinate, talking also about everything else under the sun.

Quantised adiabatic charge transport in the presence of substrate disorder and many-body interaction Q. Niu and D. J. Thouless, J. Phys. A: Math. Gen. 17 2453 (1984)

Recalling that time, Phil Anderson writes: "In 1967 I arrived to find the brilliant, hyperactive Gideon Yuval as one of my first students. He and I worked on my new idea of the "infrared catastrophe", and together we battled through to the solution of the Kondo problem by '69, with encouragement from others, including especially John Hopfield. That work has reverberations all over theoretical physics: including the 1D Coulomb gas, the Kosterlitz-Thouless transition, X-ray edge phenomena, and much modern stuff! In the course of this research, we were the first to use renormalization group techniques in condensed matter theory. Gideon returned to the Mossad in Israel, and then went on to Microsoft. John Armytage's thesis carried on from where Yuval left off, and calculated numbers, not quite up to Wilson's later and famous ones, but good enough. Later he joined the paint industry. In fact he and Yuval were among the first of our graduates to move out of basic research into industry and commerce.

Exact Results in the Kondo Problem: Equivalence to a Classical One-Dimensional Coulomb Gas P. W. Anderson and G. Yuval, Phys. Rev. Lett. 23 89 (1969) A poor man's derivation of scaling laws for the Kondo problem P. W. Anderson, J. Phys. C: Solid State Phys. 3 2436 (1970) More is different: Broken symmetry and the nature of the hierarchical structure of science P. W. Anderson, Science 177 393-396 (1972) Fluctuation Effects at a Peierls Transition P. A. Lee, T. M. Rice, and P. W. Anderson, Phys. Rev. Lett. 31 462 (1973) Conductivity from charge or spin density waves P. A. Lee, T. M. Rice, and P. W. Anderson, Solid State Communications 14 703-709 (1974) Model for the Electronic Structure of Amorphous Semiconductors P. W. Anderson, Phys. Rev. Lett. 34 953 (1975)

John Lekner worked some with Josephson, including producing some good stuff on ions in He with a successful student, Roger Bowley, and at the end of his Assistant Lectureship (called Demonstratorship in those days) in Cambridge, he returned to New Zealand. Josephson did important pre-Wilsonian theory on critical phenomena. He joined the group in 1967, moving from the Low Temperature Physics group, where of course it was he that invented the Josephson Effect.

I (PWA) was still interested in superconductivity, and from that period I had fundamental thoughts about the physics of the phonon interaction, including writing a much-reviled but correct paper with Marvin Cohen on the maximum Tc. John Inkson carried on, when he arrived in about 1970, in formal many-body theory, and later wrote a book based on the lectures I gave in Part III Mathematics on that subject for the first few years. To solve the problem I set him on metal-semiconductor interfaces, he developed the so-called G-W method, later used by the Swedes for correcting LDA energy gaps. He was not as good a publicist as a physicist, but got a lectureship in the Cavendish and then had a good career as professor and Deputy Vice-Chancellor at the University of Exeter.

While we were still in our "rainbow" or intercontinental phase, a student from Malaysia, Wai-Choo Kok, and I began to puzzle about new experiments on random magnetic alloys, which I named "spin glass" in our papers of 1969 and 1970. She's now a professor in Singapore. Attending a meeting to report on that work, I picked up the striking experimental work of Mydosh, which led to my stimulating Sam Edwards in 1974-75 to help me with this problem. It led to the famous replica theory, and then with Richard Palmer to the famous paper by Thouless, Anderson, and Palmer on spin glasses in 1977.

Anomalous low-temperature thermal properties of glasses and spin glasses P. W. Anderson, B. I. Halperin and C. M. Varma, Philosophical Magazine 25 1-9 (1972) Theory of spin glasses S. F. Edwards and P. W. Anderson, J. Phys. F: Met. Phys. 5 965 (1975) Solution of 'Solvable model of a spin glass' D. J. Thouless, P. W. Anderson, and R. G. Palmer, Philosophical Magazine 35 593-601 (1977)

Patrik Fazekas belongs also to our rainbow phase. He came with Hungarian money and I put him on the RVB idea of Resonating Valence Bonds, which has recently become of interest again. Our papers of 1972-73 were stimulating but inconclusive.

Resonating valence bonds: A new kind of insulator? P. W. Anderson, Materials Research Bulletin 8 153-160 (1973) On the ground state properties of the anisotropic triangular antiferromagnet P. Fazekas and P. W. Anderson, Philosophical Magazine 30 423-440 (1974)

An interest that coincided a bit with Heine's and Roger Haydock's during those years was my "chemical pseudopotential" idea, where the attempt was to abstract the parts of chemistry which correspond to the naive "chemical intuition" about bonds, ions, and ligand complexes. The main point is that these are all LOCAL properties, coming out of the electronic theory, which is not local. I did this by deriving equations for ultralocalized, non-orthogonal Wannier functions, based on the Cohen-Heine ideas about pseudopotentials. A marvellous student from Northern Ireland, Dave Bullett, took this over and ran very far with it, but had difficulty publishing in Physical Review from the powerful LDA forces. Bullett, Haydock, Heine and Mike Kelly wrote a series of review articles filling Volume 35 of the Seitz-Turnbull series. Dave died tragically after serving manfully as Head of Department at Bath for many years, giving it the reputation of the happiest department there!

Electronic structure based on the local atomic environment for tight-binding bands R. Haydock, V. Heine and M. J. Kelly, J. Phys. C: Solid State Phys. 5 2845 (1972)

Around 1970, I began to be reconverted to localization and amorphousness, and from that period "two-level centers", the Fermi glass, etc., emerged. I began to talk again to Mott! In the Summer of 1975, I brought out the "negative U" idea, published in Nature but not as far as I recall associated with any Cantabridgians other than Mott and Morrel Cohen (as a visitor). From the mid 1960s we began to get really first-rate UK students including Richard Palmer, Alan Bishop, John Armytage, Mike Cross, Duncan Haldane, Roger Bowley, and John Inkson on the many-body side, with Dave Bullett, John Inglesfield, John Pendry, Denis Weaire on "one-electron" theory.

A self-consistent theory of localization R. Abou-Chacra, D. J. Thouless and P. W. Anderson, J. Phys. C: Solid State Phys. 6 1734 (1973) Theory of extended X-ray absorption fine-structure P. A. Lee and J. B. Pendry, Phys. Rev. B 11 2795-2811 (1975)

Palmer started work on neutron stars with a thesis on equations of state in about 1974, but it was a 'rainbow coalition' visitor, Naoki Itoh, who stimulated me to come up with the 'superfluid glitch' idea, which was then Ali Alpar's thesis, finished at Princeton after I left Cambridge. Alpar is well-situated in the neutron star world and in Turkish physics now. But Hopfield recognized Palmer's quality and brought him to Princeton; thence he went to Duke where he was one of the originators of "econophysics". A few years ago, at the peak of his career, he had a stroke.

The wind-up of my stay in Cambridge was two major projects. In the fall of 1973 I took a sabbatical to stay in our new cottage in Cornwall and spent the time totally reworking the course on many-body theory that I had been giving in Part III of the Mathematics Tripos, basing the notes on what I really thought were the key notions we actually use in research: modelling and analytic continuation, broken symmetry, renormalization group, etc. Thereafter my course was based on these new notes, which eventually, in 1983, became the book Basic Notions of Condensed Matter Physics , which has been influential. (The long delay was due to my reluctance to leave out strongly correlated physics, but I was right that I didn't understand it then.)

The second project was superfluid 3He, which I worked on over the winter 1972-73, with visits from Chandra Varma and Bill Brinkman. I suggested to Cross that he not get involved because of the difficulty and competitiveness of the field, but over the summer at Bell he began working with Bill and me, and we wrote several important papers, which I quoted in our review article of 1974. He went on to Bell and then to Caltech, where he is an important figure in the field of fluid dynamics.

A generalized Ginzburg-Landau approach to the superfluidity of helium 3 M. C. Cross, Journal of Low Temperature Physics 21 525-534 (1975)

Haldane's thesis (finished in Princeton) was not one of my more successful ideas, but it started him off in a direction that he could exploit. He went on to really be trained while a postdoc with Nozieres, and ended up here in Princeton, as a brilliant and incomprehensibly mathematical as ever."

Simple model of multiple charge states of transition-metal impurities in semiconductors F. D. M. Haldane and P. W. Anderson, Phys. Rev. B 13 2553 (1976)

Electronic structure and other things: After Ziman moved to Bristol, he initiated an annual Bristol-Cambridge 'mingle'. But in other ways Heine felt academically isolated in Britain and indeed Europe. Academic visitors from America were very important, including (some of these before 1964) Jim Phillips, Morrel H Cohen, Bill McMillan, Marvin L Cohen, Walter Harrison and John Hopfield. All of these were or had been at Bell Labs or Chicago, except Harrison who was nevertheless part of the Chicago circle through Morrel Cohen's consulting at General Electric Research Labs. They kept inviting Heine over (and paying) most years, so that he worked almost as an offshoot of the USA until into the 1970s. In fact it became very much a two-way arrangement, with many of the best Cambridge PhD graduates going over to work at Bell Labs or with others of these people in USA.

On electronic side, the visit of Igor Abarenkov gave birth to the 'model pseudo-potential', the list of pseudopotentials by Alex Animalu becoming one of the Science Citation's 'best sellers'. Dennis Weaire applied the pseudopotentials to explain a whole raft of facts about the structure and properties of sp-bonded metals. When Heine gave an early talk about this while visiting Bell labs, Conyers Herring remarked "Never have I learnt so much physics in one hour." Three review articles by Heine, Cohen (Marvin) and Weaire on all this and more filled Volume 24 of the Seitz Turnbull series on Solid State Physics.

Other work included David Pettifor on transition metals, John Pendry on low energy electrons, John Inglesfield on alloys and surfaces, Mike Finnis on metallic structures including surfaces, and Bob (R.O.) Jones doing the first realistic calculations of surface states on semiconductors. Roger Haydock, later appointed as an Assistant Lecturer, developed the recursion method with Chris Nex as Computer Officer, an expert on computational mathematics. This was designed to study electronic structure from the point of view of the local atomic environment. They distributed their computer code freely by magnetic tape, at a time when secrecy and competitiveness was the norm, and it set new standards in clear documentation and user friendliness, with worked examples. This local point of view formed the theme of Volume 35 of the Seitz Turnbull series written by Haydock, Kelly, and Heine.

There was far too much good work to mention all here, but important Post-docs, visitors and PhD students from overseas include (besides those already mentioned above) Bob Shaw, Itoh, Bob White, Chandra Varma, Denis Newns (actually from UK), Michel van Hove, Risto Nieminen, Erio Tosatti, and Abhijit Mookerjee.

Electronic structure and the start of ab initio computer simulation: After Anderson left in 1975, Heine was appointed to his professorship. Anderson, with his US Air Force grant, had supported the whole group regarding travel and visitors, and therefore left a big hole when he left. So Heine's first job was to engage in battle with EPSRC over the need to have financial support for visitors and to travel. At that time experimentalists got support to travel to big facilities, which could take in some other places en route, but theoreticians were deemed not to need to travel. The funding structure of EPSRC, at least on the physics side, seemed to be run by experimentalists for experimentalists! However, new concepts are frequently carried around by theoreticians so that, if one wants to be up to date and participate at the leading edge, travel and hosting visitors are vital. Anyway, Heine's arguments eventually prevailed and support was eventually provided in the form of a flexible Rolling Grant came about, to support the work of the whole group, and affording a considerable degree of flexibility!

Haydock was appointed to an Assistant Lectureship, and this was the time to cash in on his and Nex's recursion method to calculate electronic structure and lattice vibrations in amorphous materials, and also in iron with the local magnetic moments pointing randomly or at least not aligned. No one else had the capability to do calculations like that at the time. In the late 1970s Richard Martin and Marvin Cohen had shown that one could do amazingly good computer simulations of properties and processes in solids with pseudopotentials and a plane wave basis set. Heine had never been a friend of big computing, but he saw that this development could probably open many new doors that had been so frustratingly closed.

Around 1980 Heine complained that EPSRC was not supporting electronic structure calculation just when it was taking off, and the Physics Committee chair (Mike Hart) said he would consider a proposal if a senior figure would also make it a major priority. The following year the New Blood scheme came along for young lecturers to renew ageing university departments. Heine put in a proposal for someone to start this type of calculation, which was ranked very low by the powers-that-be in the Cavendish. But in that first year of the New Blood scheme it was the EPSRC committees that made the selection. Much to the chagrin of the universities, those committees were independent minded (and hence not allowed to make the decisions in subsequent years), and so Heine's proposal was funded!

There was a lack of outstanding people experienced in electronic structure suitable for the appointment. However, a young researcher by the name of Richard Needs said he was interested and wanted to change from the polymer simulations he had been doing for his PhD with Edwards. We arranged a special leave of absence for his first year (something unheard of at that time) to learn the new technique of electronic structure calculation with Richard Martin in USA. He came back with a working code written by Karel Kunc, and it all took off from there with Neville Churcher, Dominic King-Smith, Ching Cheng, Abdallah Qteish, Jyh-Shin Lin and others. A whole slew of applications followed, including the phase diagram of SiC polytypes, surface stress and reconstruction, electron field emission, hydrolytic weakening of quartz, and many more.

First-principles calculations of the electronic properties of silicon quantum wires A. J. Read, R. J. Needs, K. J. Nash, L. T. Canham, P. D. J. Calcott, and A. Qteish, Phys. Rev. Lett. 69 1232 (1992)

Heavier computing needed heavier computing equipment, and this was another battle with the university and EPSRC. The dogma at that time was that under the 'dual support system', universities were responsible for providing computing resources for which they got a new mainframe every seven years or so. EPSRC also had central supercomputers on which one could get a bit of time through research grants. In particular, EPSRC never funded any small local computers (except for High Energy Particle researchers who were privileged). But with the advent of workstations, that system was becoming grossly inefficient. For example the Rutherford Lab ran a computing centre with a staff of over 100 people, but the time we could get on their machine was quite inadequate. So we applied for an EPSRC grant for a Floating Point Systems (FPS) vector processor, to be driven by a small Vax (partly from University funds), which we got after battling for over a year. It gave us much more computing power than we could have gotten from the Rutherford centre. The Computing Service under the direction of David Hartley was extremely helpful, and they were the first in British universities to drive the change to networking and distributed computing.

The FPS machine was based on new technology, and TCM successfully insisted on acceptance criteria involving the machine running for several days without error. In industry the norm was to expect error-free running during working hours, but for our codes, which often took several days to complete, reliability had to be higher. During the trial year the reliability of the FPS computer had improved, but had not reached our acceptance criteria. FPS was very keen on a sale, and we were very keen to have the machine, but the Government advisor who needed to sign off the purchase was not minded to do so. So Heine suggested that FPS add a second vector processor at a notional cost. All three parties agreed to this, so TCM got two vector processors, not one.

Heine's complaint about the lack of support for electronic structure from EPSRC co-incided with the formation of a national computing centre at Rutherford designed to support more than just High Energy Physics. To co-ordinate the use of this centre, EPSRC encouraged the formation of the CCPs, and Heine was instrumental in establishing Computational Collaborative Project No 9 (CCP9) on the "Electronic Structure of Solids" in 1981. This UK-based network worked well, and in 1994 Heine, in conjuction with Walter Temmerman, another founding member of CCP9, was able to initiate the European equivalent, Psi-k (or Ψk).

Oh, and there was so much other good physics going on (in addition to the statistical mechanics part of the group) with Kiyo Terakura, Bernard Buxton, Pedro Echenique, Angela Lahee, Michael Gunn, Matthew Foulkes, James Annett, the Churchers and many others. For example Rex Godby came back from USA with the GW method applied to calculating better band gaps than could be done with Density Functional Theory, and he greatly developed it.

Polymers and statistical mechanics: Sam Edwards came to the Cavendish in 1972 as a Humphry Plummer Professor, and it was decided to rename the group TCM (Theory of Condensed Matter), since polymers were in many ways different from conventional solids. Almost immediately Edwards became Head of the Science Research Council (now EPSRC), then located in London, and so he supervised research students on the train and worked out multi-dimensional integrals while chairing meetings, filling successive 'little red books'. It was however a most productive period. He used his recent invention in polymers of the replica trick to construct (with Anderson) a celebrated theory of spin glasses. They found how to define an order parameter, and hence were able to develop a mean field theory of the spin glass transition. A second paper quantised the spin glass. A whole industry on spin glasses and then neural networks developed. Polymer work turned to snake-like motion (reptation; with Masao Doi), where entangled long chain molecules can wiggle as if in a tube, extending the tube at one end and retracting at the other end. The release of constraints and thus molecular distortions was very successfully described and underpins the huge, industrially important field of rheology. The standard book (by Doi and Edwards), The theory of polymer dynamics also emerged.

Dynamics of concentrated polymer systems. Part 1. Brownian motion in the equilibrium state M. Doi and S. F. Edwards, J. Chem. Soc., Faraday Trans. 2 1789-1801 (1978) Dynamics of concentrated polymer systems. Part 2. Molecular-motion under flow M. Doi and S. F. Edwards, J. Chem. Soc., Faraday Trans. 2 1802-1817 (1978) Dynamics of concentrated polymer systems. Part 3. Constitutive equation M. Doi and S. F. Edwards, J. Chem. Soc., Faraday Trans. 2 1818-1832 (1978) Dynamics of concentrated polymer systems. Part 4. Rheological properties M. Doi and S. F. Edwards, J. Chem. Soc., Faraday Trans. 75 38-54 (1979) The Surface Statistics of a Granular Aggregate S. F. Edwards and D. R. Wilkinson, Proc. R. Soc. Lond. A381 17-31 (1982)

Subsequently, new staff members, Robin Ball, Mark Warner and Michael Cates, obtained posts for work in statistical areas. Ball was a pioneer in diffusion limited aggregation (how do fluffy objects like snow crystals and blobs of soot form?) Once there is a little nucleus to start with, individual molecules or tiny specks of soot arrive to stick onto it, and then more and more. Obviously they are likely to hit first one of the outer parts of the incipient cluster and so it grows in a fluffy form known as a fractal. Of course there are lots of wrinkles, because not all snow crystals and soot smuts are alike, and Ball's work during the 80s and 90s sorted this out. In particular he developed a computer code that was vastly cleverer and ran vastly faster than other people's, so that he could simulate what actually happens under different assumptions as well as trying to explain it. Ball moved in 1998 to a professorship at Warwick University.

Elasticity of entangled networks R. C. Ball, M. Doi, S.F. Edwards, and M. Warner, Polymer 22 1010-1018 (1981) Fractal growth of copper electrodeposits R. M. Brady and R. C. Ball, Nature 309 225-229 (1984) Universality in colloid aggregation M. Y. Lin, H. M. Lindsay, D. A. Weitz, R. C. Ball, R. Klein, and P. Meakin, Nature 339 360-362 (1989)

Warner discovered many new phenomena in "quasi-solids" after creating the first theories of liquid crystal elastomers. Predictions include solids that change their lengths by factors of 4 or 5 on slight heating or exposure to light, strange solids that change their shape without energy cost, can tell the difference between right and left, lase when lightly pumped, and change emission colour on stretching, etc.

Nematic elastomers - A new state of matter? M. Warner and E.M. Terentjev, Progress in Polymer Science 21 853-891 (1996) A New Opto-Mechanical Effect in Solids H. Finkelmann, E. Nishikawa, G. G. Pereira, and M. Warner, Phys. Rev. Lett. 87 015501 (2001)

Cates questioned whether chains composed of assemblies of soap-molecules can break and reform rather than wiggling out of entanglements (a special form of a reptation). He became a specialist in complex phases in general made up of these amphiphiles to form onions, worms and flapping sheets (sheety-phases). Cates was appointed to a professorship in Edinburgh in 1995 and was later appointed to a Royal Society Research Chair.

Statics and dynamics of worm-like surfactant micelles M. E. Cates and S. J. Candau, J. Phys. Cond. Matt. 2 6869-6892 (1990)

Edwards had excellent relations with industry. He obtained the initial funding for Warner's post and then also funds from Unilever for a large experimental/theoretical programme in Polymers and Colloids (P&C). A separate P&C group (now renamed Biology and Soft Matter) under Athene Donald, a long-term associate of TCM, was founded. Edwards moved to P&C on retirement, and Ball, Cates and Warner all had theoretical post-docs in that group. Edwards became the Cavendish Professor in 1984 until his retirement in 1995, and did his seven year stint as Head of Department. A memorable meeting of his current and former colleagues was held on his retirement, but he continues seminal research, largely on granular materials. Equally famous were Edwards' dinners for "top people" in Caius College, ably organised by Doreen Alderton. These ranged from captains of industry (the stimulated emission of money, as an unnamed colleague termed them) to close friends of the group such as Pierre-Gilles de Gennes.

Quantum Monte Carlo (or QMC) is a much more accurate way of doing ab initio simulations, with less approximations (though still with one, its Achilles heel!). It also takes much more computing power, so that it was regarded around 1990 as still a far-off hope as far as useful applications go. However it was also correspondingly challenging as a new development, which is what proved attracted to Needs. So the baton of doing Density Functional calculations passed to Michael Payne (see below), with his new CASTEP code, and Needs soon became a world expert on QMC. In the meantime the operation of Moore's law concerning the inexorable growth of computing power means that now, the QMC approach to modelling real condensed matter are no longer a distant hope but a practical reality. Guna Rajagopal played a leading role in the early days of the development of QMC in Cambridge, and more recently Mike Towler, and a variety of graduate students and post-docs, including Steve Kenny, Andrew Williamson, Paul Kent, John Trail, Neil Drummond and Pablo Lopez Rios, led its transformation into the powerful CASINO code.

CASTEP/UKCP and ONETEP After completing his PhD in Cambridge in the early 80s, Payne spent a year at MIT in John Joannopoulos' group in 1985/6. Joannopoulos was one of the first to hear about the Car-Parrinello method when he met Roberto Car at IBM's Yorktown Heights Lab. He decided that this was the future of first principles calculations and decided that Payne would develop a set of codes for his research group to use. Payne continued to work in this area after his return to TCM with a gradually expanding group of excellent students and post- docs including Alessandro De Vita, Ivan Stich, Victor Milman, Ruben Perez, Ian Robertson, Graham Francis, Guy Makov and Carla Molteni. Heine persuaded CCP9 to support the development of this technique as their flagship project and the result of this support was the CASTEP (CAmbridge Sequential Total Energy Package) code, which was widely distributed to other academic groups in the UK and Europe.

In 1991, an SERC initiative lead to the purchase of a 64 node Meiko Parallel Computer, which was located in Edinburgh. The machine was predominantly used for Quantum Chromodynamics (QCD) calculations, but in order to spread the cost to other research areas within SERC, 25% of the time was allocated by 'higher authority' to Car-Parrinello simulations. Rather rapidly, the UK Car-Parrinello Consortium was created, involving groups in Oxford, Edinburgh, Bath (David Bird) and Cambridge, and they wrote a proposal for this funding without any of us having a clue whether Car-Parrinello simulations could be run on parallel computers. It was a real worry until at one meeting Payne stunned everyone by coming up with the answer. It is obvious in retrospect, namely distribute the expansion coefficients of the Bloch wave functions, but at the time, parallelising was in its infancy and a wave function was thought of as a single entity, so that people's imagination had not got much beyond distributing different k-points to different processors. The consortium did successfully develop a parallel version of the CASTEP code to be used on this machine and the resulting code was called CETEP (Cambridge-Edinburgh Total Energy Package), acknowledging the contribution of the Edinburgh Parallel Computing Centre (EPCC) in the creation of the code. The CETEP code was truly world leading, and allowed density functional theory calculations to be applied to systems containing many hundreds of atoms.

In 1994, CASTEP was licensed to Molecular Simulations (which has subsequently been renamed Accelrys) and the free academic distribution of the code ended, to the derision of much of the academic electronic structure community. The decision became inevitable when the research council refused any postdoc money to support the code at a time of growing user base, development of new features, and the advent of new platforms to run it on. There was no way TCM could cope at that time with all those demands. CASTEP has now generated sales exceeding 6 million dollars and Accelrys employ a good number of former members of TCM. CASTEP was re-written from scratch between 1999 and 2002 by Matt Segall, Chris Pickard, Matt Probert, Phil Hasnip (all members of TCM) Stewart Clarke in Durham and Keith Refson at the Rutherford Appleton Laboratory. Philip Lindan who is now at the University of Kent played a major role in getting the re-write started. The new CASTEP code is written in a version of the Fortran computer language that neither Payne nor Heine would claim to understand.

The computational cost of CASTEP, as with most conventional electronic structure approaches, scales as the cube of the system size. To Payne it was clear that this unfavourable scaling (much worse than that of the recursion method for example) would prevent the application of first principles calculations to very large systems, such as the biological systems that he was becoming increasingly interested in. For the last 10 years, Haynes has worked on the development of a first principles technique whose computational cost scales linearly with the number of atoms in the system. Drs. Arash Mostofi and Chris Kriton-Skylaris have worked with Haynes for the last 4 years, and together they have developed the ONETEP code (Order N Electronic Total Energy Package), which offers the same accuracy as CASTEP but can be applied to systems containing thousands of atoms. The ONETEP code puts TCM ahead of the rest of the world in electronic structure, yet again.

Iterative minimization techniques for ab initio total-energy calculations: molecular dynamics and conjugate gradients M. C. Payne, M. P. Teter, D. C. Allan, T. A. Arias, and J. D. Joannopoulos, Rev. Mod. Phys. 64 1045 (1992) First-principles simulation: ideas, illustrations and the CASTEP code M. D. Segall, P. J. D. Lindan, M. J. Probert, C. J. Pickard, P. J. Hasnip, S. J. Clark, and M. C. Payne, J. Phys.: Condens. Matter 14 2717 (2003) First principles methods using CASTEP S. J. Clark, M. D. Segall, C. J. Pickard, P. J. Hasnip, M. I. J. Probert, K. Refson, and M. C. Payne, Zeitschrift fur Kristallographie 220 567-570 (2005)

In more recent years the atmosphere of TCM has been greatly enriched by the presence of David Khmel'nitskii, elected to a prestigious Title B Fellowship at Trinity College. David was one of several outstanding physicists nurtured at the Landau Institute in Chernogolovka, indeed the first student to graduate from this prestigious institution. His PhD work was ground breaking, anticipating the celebrated work of Wilson in USA on the application of the renormalisation group. After joining the faculty at the Landau Institute, he was responsible for several achievements in areas ranging from localisation phenomena and the integer quantum Hall effect to the broad area of mesoscopic physics, a field that he helped to pioneer. His years spent as Editor of JETP Letters have endowed him with an unsurpassed knowledge of the literature, while he still maintains the energy and distinction to police the integrity of the subject of which he is so fond. David helped to develop a graduate lecture programme, and is particularly well-known for his idiosyncratic 'Fairy Tales' spanning a wide range of physics.

From around 1990 handed over most of the running of the group to the 'Young Turks'. Peter Littlewood was appointed as Volker's successor as chair and head of group. Peter added new dimensions to the work of TCM with research in many directions, including, in particular, on materials with strongly correlated electrons and on condensation phenomena in semiconductor optics. But already trends in that direction had come with David Khmel'nitskii and Ben Simons. Analogies with polariton condensation led Peter into the growing field of strong-correlation phenomena in ultra-cold atomic gases, in which the group now has broad and established interests, through the work of Ben Simons and with the appointment of Nigel Cooper in 2000.

Quantum Phases of Vortices in Rotating Bose-Einstein Condensates N. R. Cooper, N. K. Wilkin, and J. M. F. Gunn, Phys. Rev. Lett. 87 120405 (2001) Electronically soft phases in manganites G. C. Milward, M. J. Calderon, and P. B. Littlewood, Nature 433 607-610 (2005) Bose-Einstein condensation of exciton polaritons J. Kasprzak, M. Richard, S. Kundermann, A. Baas, P. Jeambrun, J. M. J. Keeling, F. M. Marchetti, M. H. Szymanska, R. Andre, J. L. Staehli, V. Savona, P. B. Littlewood, B. Deveaud, and Le Si Dang, Nature 443 409-414 (2006)

During his PhD, Ben Simons was supervised externally by Michael Gunn, then at RAL and formerly himself a member of TCM. After a short stint as a Research Fellow of Gonville and Caius College working in collaboration with Mike Cates, he transferred to a fellowship at MIT, returning a year later to a Royal Society position and lectureship at Imperial College. Subsequently he returned to the Cavendish Lab and TCM as a lecturer in 1995. He is now the Herchel Smith Chair in Physics and is the current head of group.

Tom Duke, then a research fellow, started work in TCM on biophysics, producing an important theory of hearing. It had always been a puzzle how the very faint sound levels could produce a response in the nervous system of the ear, and the same was true of other sensory receptors. Tom showed that the nervous system of the ear is in fact close to what in physics we would call a phase transition where the susceptibility goes infinite, which results in a large amplification factor on the incoming disturbance. He showed that the hearing system could maintain itself in a steady state so close to that transition. In 2002 he was appointed to a lectureship in the Cavendish, and formed a new group in Biological Physics. Around 2007 he transferred to a chair at the London Centre for Nanotechnology at UCL. Later Tom tragically passed away at the young age of just 48.

Auditory sensitivity provided by self-tuned critical oscillations of hair cells S. Camalet, T. Duke, F. Julicher, and J. Prost, PNAS 97 3183-3188 (1997)

In 2003-04 Mike Payne took over as Head of Group while Peter was away on sabbatical, and stayed in that position in 2004-05 while Peter prepared to take over as Head of Department from October 2005. Later, Peter transferred to Argonne National Laboratory in the US recently becoming its Director. These days, the group remains firmly anchored in the traditional arena of condensed matter physics, and is probably as broad in its outlook in styles of research and in subject matter as it has ever been. TCM remains a dominant force in electronic structure, and is pushing into new realms from biological processes on the one hand to quantum Monte Carlo on another. Mark Warner is a leader in the soft matter community (recognised by the Agilent Prize in 2003). Ben, Nigel, and David together with the recent appointments of Claudio Castelnovo and Austen Lamacraft several years ago, form an active nucleus for "quantum matter" - spanning strongly correlated electron systems, mesoscopics, and utracold atom physics.

High-pressure phases of group-IV, III-V, and II-VI compounds A. Mujica, Angel Rubio, A. Munoz, and R. J. Needs, Rev. Mod. Phys. 75 863 (2003)