Computational Chemistry with Python

Computational chemistry offers a peculiar hospitality. It admits the student with a modest computer, gives the experienced investigator access to elaborate theories, and presents both with numbers whose assurance may considerably exceed their entitlement. The numbers are not at fault. They have answered the question they were given. Our difficulty is to decide whether that question was the one we intended to ask. This book is written for those who wish to take responsibility for that distinction. It joins quantum chemistry, molecular simulation and machine learning through a common discipline: define the chemical question, construct a representation, state the approximation, perform the calculation, and examine the grounds on which its conclusion may be believed. Python supplies the working language. Chemistry supplies the object. Neither is permitted to excuse carelessness in the other. The argument proceeds from principles towards practice, but it does not ask the newcomer to wait until every advanced theorem has been proved before doing useful work. A definition is introduced as a definition, an empirical assumption as an empirical assumption, and a derived relation with the steps that support it. Physical constants are identified through their definitions or measurement; they are not conjured by algebra. A result that requires a theory developed later is signposted rather than smuggled into an introductory formula. This distinction allows the exposition to remain honest without becoming motionless. The mathematical apparatus is intended to assist reasoning, not to decorate it. Displayed equations are numbered, their symbols are defined, and the transitions between them are explained. Worked examples show what the formalism does to actual numbers. The chapter exercises carry full solutions because an unexplained answer is an oddly economical form of assistance: it saves the author the very labour for which the student bought the book. Our practical emphasis is on accessible tools and small systems. Accessibility does not mean that all useful chemistry can be performed cheaply, or that an ordinary computer can imitate a research cluster by displaying sufficient optimism. It means that the central lessons should be reproducible without proprietary software or exceptional hardware. More demanding methods are placed in their proper relation to the questions they answer. The circumstances under which a calculation becomes expensive are part of the science, not an embarrassment to be concealed in an installation note. Black-and-white copies are used to make these books available to a broader audience. Where feasible, a full-colour edition is available in hardcover, although technical and economic conditions do not always permit it. The illustrations are consequently designed to retain their meaning when colour is absent. A blue curve may be handsome; a dashed curve can still be recognised when the printer has declined to share our enthusiasm.The intended audience includes advanced undergraduates, postgraduate students, experimental chemists, physicists entering molecular science, and independent scholars whose previous education has left a few uneven stones in the path. No embarrassment attaches to returning to a definition. The skilled practitioner does so more frequently than the confident novice, although usually with less ceremony. Readers should be comfortable with elementary chemistry, algebra, functions and introductory calculus. Please search for extra material at www.scottishsciencesociety.org