On the occasion of the 2015 edition of the Proceedings of the 55th Sanibel meeting, a few thoughts come to mind.
In January 1961, a scientific meeting organised by the University of Florida's Quantum Theory Project was held on Sanibel Island, FL, off Ft. Myers, FL, in the Gulf of Mexico. It was on the topic of the quantum theory of the electrons in atoms, molecules, solids, and bio-systems. Now 56 years later, that meeting remains the largest, annual meeting in the field, and will re-convene on Valentine's day, 14–19 February 2016, at St. Simons Island, GA. The ‘Sanibel’ meeting is known as the ‘Theory Meeting for Theoreticians’ as it emphasises the cutting edge of new developments in the electronic structure and dynamics of atoms, molecules, solids, bio-systems, and the critical understanding and interpretation that only a knowledge of ‘what the electrons are really doing’ offers.
Few institutions have had such an influential effect on a field as has the Sanibel meeting. In 1961, the state of theoretical and computational chemistry was in its infancy, along with electronic structure studies in solid-state physics. Few, if any calculations could be made that were deemed useful by the larger chemistry and solid-state physics community, and those that were almost invariably of a semi-empirical nature. In fact, in his famous talk [Revs. Mod. Phys. 32, 170–177 (1960)], C.A. Coulson alluded to the two cultures: those that aspired to first-principle, ab initio calculations; and the much larger group of semi-empirical theorists. In essence, he condemned the former as having little to offer. Yet, it was exactly this group that became the core of the Sanibel meetings, and today, it is well appreciated that because its adherents require nothing but the atoms involved to completely describe a system, for many problems, the most direct route to providing essential information about atoms, molecules, solids, and even bio-systems is offered by first-principle calculations that solve the Schrödinger or Dirac equation to high accuracy.
Furthermore, applications of theory drives computer development, with the current avowed world-wide objective to make such calculations at the exascale, in turn, driving massively parallel program development for applications of the theory. Theory has thus become an equal partner with experiment in building a detailed understanding and an ability to ‘predict’ phenomena involving electrons, from molecular electronics, to the detonation of explosives, to the action of RNA. One only has to glance at the array of programs available today that make it possible, indeed, easy, to do ab initio, correlated wavefunction theory (WFT) or first-principle (but not ab initio!) density functional theory (DFT) applications for molecules, polymers, and crystals; and there is still semi-empirical theory though the distinction between that and DFT is less apparent.
Today, there are still two cultures in our field. First, there is the culture devoted to developing new and more powerful theories for the solution of the Schrödinger and Dirac equations and for associated dynamics problems. These advances make it possible to study new phenomena and provide the critical accuracy that is not accessible to today's program implementations. These scientists are ‘theoreticians’.
The other group consists of scientists that use the existing software to make applications to molecular and condensed phase problems, whose answers aid in defining devices or species that will serve critical purposes in catalysis, solar energy storage, drug design, and atmospheric characterisation, among a myriad of other applications. To do the latter successfully, requires scientists that are very-well versed in the problem to be solved, who know what questions to ask, and how to intelligently use the available computer software to help get that information. These I will call ‘computational scientists’. The task of the theoreticians is to invent new descriptions of phenomena and the computational tools that produce the numerical results demanded and often provided by the computational scientists. Unlike the latter, theoreticians are never limited by existing software, as they will write or add new things to evolving software to attack a new range of problems, and in this way the field advances its capabilities and applications.
Obviously all good theoreticians and computational scientists would expect to do some of both, but in the interest of a well-focused scientific meeting of limited size and duration, one has to choose. As the focus of the Sanibel meeting has always been on enabling theoreticians to formulate the methods and tools required for their computational application, we retain the emphasis on theory. Hence, the Sanibel meeting is the ‘Theory Meeting for Theoreticians’, among the meetings that are regularly organised today. And this emphasis on theory is what enables it to offer essential cross-fertilisation among physicists, chemists, materials scientists, and bio-scientists. Electrons are the glue of the whole field, so being able to extract their secrets from quantum mechanics is paramount to all those disciplines. For the computational chemists, a visit to Sanibel enables them to know the new theory, and software that can be applied to their problems. In turn, they can inform the theoreticians about demanding problems that require their expertise to invent new solutions.
I write this while at the PacifiChem meeting in Honolulu composed of the combined chemical societies from all the countries bordering the Pacific Ocean. There are 18,000 people here. Among these 18,000, my count shows about 600 are theoretical or computational chemists who are attending about 40 parallel sessions. Even that number of theoretical and computational chemists pales by comparison to the recent World Association of Theoretically Oriented Chemists in Santiago de Compostella, Spain that had ∼1300 attendees. At such meetings with many parallel sessions, it is very hard to attend the right talk, or be able to have the conversations and exchange of insight that can accelerate progress in one's work. Sanibel, on the other hand, offers exactly such a forum with no parallel sessions and the most expert audience in theory, who understands the equations and the limits of a method. Their expertise can be brought to bear to help accelerate the progress of an investigator and the field. In many cases, our attendees are just developing the exact tool required and can make those programs available. This exchange of information among knowledgeable investigators is a hallmark of the 55 years of the Sanibel meeting, and plenty of time is devoted to informal ‘bar’ or ‘beach’ discussions for this purpose.
Nearly all of the most important developments in the field of electronic structure theory were first presented at the Sanibel meeting. These range from DFT, one of the core methods of the meeting from its inception to the present, including even pre-DFT methods like Xα, to gradient corrected and hybrid Kohn–Sham. The first report of ‘constrained-search’ was at Sanibel as was E.B. Wilson's famous alternative ‘proof’ of the Hohenberg–Kohn theorem. The other cornerstone of the meeting has been the correlation problem, and the most popular electron correlation methods from configuration interaction, many-body perturbation theory (or Moeller–Plesset PT), coupled-cluster theory, including its equation-of-motion CC excited state extensions, and electron and polarisation propagator theory were first discussed there. Sanibel even offered the first report of semi-empirical methods like CNDO/INDO, plus such crucial developments as analytical gradients and hessians for ab initio methods, quantum Monte Carlo, along with methods that include explicit correlation (R12 and F12), vibrational perturbation theory, and density matrices. In condensed matter, besides DFT, Sanibel offered the forum to develop the tools required to treat the electrons in infinite systems like polymers and crystals, including the now widely popular density fitting method, and how to do ‘real-world’ simulations for dynamics. Sanibel was the meeting where a core of people then called ‘quantum biologists’ met every year and argued their points, one of whom won the 2013 Nobel Prize. Another regular attendee won the Nobel Prize in 1998.
Continuing problems first raised at Sanibel include the n-representability problem, spin-algebras including the graphical unitary group (GUGA), justification for linear scaling, and the theory for all spectroscopies, photo-electron, electronic, vibrational, NMR, ESR, NLO, and a wealth of other magnetic field and electric field phenomena just to name a few. Along with all of these theory advances was an emphasis on software development for the fastest computers in existence. The special issue [Mol. Phys. 108, 21–23 (2010)] offers interesting historic insight along with many scientific papers that celebrate the 50th anniversary of the Sanibel meeting.
The schedule of talks at the 55th meeting in 2014 listed in the Appendix shows the mix that we have at every Sanibel meeting. The 55th meeting was especially exciting in showing the attendees what is really happening at the forefront whether it be strong correlation, coupled-cluster theory for a protein, or the interplay between DFT and WFT that is beginning to pay important computational dividends. Some of the speakers in the plenary, invited, oral, and poster presentations have contributed excellent papers to this Sanibel Issue of Molecular Physics. I hope you will read their papers with the same enthusiasm we experienced during the editorial review.
Thank you.