The Born-Oppenheimer Approximation and its role in the reduction of chemistry

Abstract The article sets out to clarify a number of confusions that exist in connection with the Born–Oppenheimer approximation (BOA) (Born-Oppenheimer, 1927). It is generally claimed that chemistry cannot be reduced to quantum mechanics because of the nature of this commonly used approximation in quantum chemistry, that is popularly believed to require a ‘clamping’ of the nuclei. It is also claimed that the notion of molecular structure, which is so central to chemistry, cannot be recovered from the quantum mechanical description of molecules and that it must be imposed by hand through the BOA. Such an alleged failure of reduction is then taken to open the door to concepts such as emergence and downward causation. Another mistaken view is that chemists have no choice but to use the BOA whereas there is an entire sub-discipline which involves non-Born Oppenheimer calculations, and which regularly and successfully calculates many chemical and biochemical properties of molecules. Yet another misconception, according to the present author, is the view that the application of the BOA represents a violation of the Heisenberg Uncertainty Principle. Many of the claims made in the philosophy of chemistry community are based on the highly technical papers of authors such as Sutcliffe and Wooley, many of which date from about 50 years ago. While these authors remained skeptical of the possibility of recovering molecular structure from quantum mechanics, others maintained that it would eventually possible to do so. Significant progress has now been made in this direction. For example, whereas it is claimed that the full, or Coulombic Hamiltonian, for a molecule precludes the existence of molecular dipoles, some recent calculations have succeeded in obtaining the exact value of dipole moment of the LiH molecule. Even more significantly, a group in Norway has now succeeded in recovering the structure of the D 3 + molecule in a completely ab initio manner without applying the BOA, but through the use of a Monte Carlo approach.