The Calculable Approximate: A Narrative Review of Approximation Methods in Quantum Mechanics, from Rayleigh's Perturbations to the Divergence of Series
Exactly solvable quantum systems are rare—the harmonic oscillator, the hydrogen atom, the square well—and the physics of everything else is the physics of approximation. This article presents a narrative review of the primary literature that built quantum mechanics' approximation methods, from Rayleigh's Theory of Sound, whose treatment of slightly perturbed vibrating systems gave perturbation theory its name and scheme, and Ritz's 1909 variational method, which converted boundary-value problems into energy minimizations, through Schrödinger's 1926 wave mechanics, which equipped the new quantum theory with perturbation theory from the start, Born and Oppenheimer's 1927 adiabatic separation of electronic and nuclear motion, the 1926 papers of Wentzel, Kramers, and Brillouin that independently derived the semiclassical expansion bearing their initials, Hartree's 1928 self-consistent field and Hylleraas's 1929 variational computation of helium, which made the many-electron atom calculable, Kato's 1949 mathematical foundation of the perturbation method's convergence, Landau and Lifshitz's codifying textbook, and Bender and Wu's 1969 demonstration that the anharmonic oscillator's perturbation series diverges—the asymptotic character that defines the method's limits. The synthesis is organized around three themes: the perturbation program, in which small parameters turned the solvable cases into engines for the unsolvable; the variational program, in which upper bounds and trial functions turned minimization into the many-body atom's working instrument; and the semiclassical bridge, in which the WKB expansion connected wave mechanics to action-angle quantization and exposed tunneling. It is concluded that the three programs are complementary solutions to one problem—the gap between what can be solved and what must be calculated—and that their asymptotic character, made precise by Bender and Wu, is not a defect but the field's realistic self-understanding.
Exactly solvable quantum systems are rare—the harmonic oscillator, the hydrogen atom, the square well—and the physics of everything else is the physics of approximation. This article presents a narrative review of the primary literature that built quantum mechanics' approximation methods, from Rayleigh's Theory of Sound, whose treatment of slightly perturbed vibrating systems gave perturbation theory its name and scheme, and Ritz's 1909 variational method, which converted boundary-value problems into energy minimizations, through Schrödinger's 1926 wave mechanics, which equipped the new quantum theory with perturbation theory from the start, Born and Oppenheimer's 1927 adiabatic separation of electronic and nuclear motion, the 1926 papers of Wentzel, Kramers, and Brillouin that independently derived the semiclassical expansion bearing their initials, Hartree's 1928 self-consistent field and Hylleraas's 1929 variational computation of helium, which made the many-electron atom calculable, Kato's 1949 mathematical foundation of the perturbation method's convergence, Landau and Lifshitz's codifying textbook, and Bender and Wu's 1969 demonstration that the anharmonic oscillator's perturbation series diverges—the asymptotic character that defines the method's limits. The synthesis is organized around three themes: the perturbation program, in which small parameters turned the solvable cases into engines for the unsolvable; the variational program, in which upper bounds and trial functions turned minimization into the many-body atom's working instrument; and the semiclassical bridge, in which the WKB expansion connected wave mechanics to action-angle quantization and exposed tunneling. It is concluded that the three programs are complementary solutions to one problem—the gap between what can be solved and what must be calculated—and that their asymptotic character, made precise by Bender and Wu, is not a defect but the field's realistic self-understanding.
The accuracy of the renormalized perturbation series for anharmonic oscillators and other simple quantum-mechanical models increases with the quantum number when the adjustable parameter is determined by a simple variational criterion. In the semiclassical limit of quantum mechanics such an expansion converges towards the result of the Jeffreys–Wentzel–Kramers–Brillouin method to all orders. The hypervirial and Hellmann–Feynman theorems facilitate the derivation of the semiclassical series from Rayleigh–Schrödinger perturbation theory.
The exactly solvable systems---the harmonic oscillator, the hydrogen atom, and the angular momentum algebra---are quantum mechanics' solved ground: the spectra computed in closed form that calibrated the theory, taught its methods, and revealed the symmetries that make solvability itself a theorem. This article presents a narrative review of the primary literature of that ground, from de Broglie's 1924 thesis---the wave hypothesis the solvable problems confirmed---through Heisenberg's 1925 umdeutung, Schrödinger's 1926 eigenvalue problem, whose oscillator and rotator opened the solvable list, Dirac's 1926 theory of quantum mechanics and Pauli's 1926 hydrogen solution by the new algebra, Darwin's 1927 vector waves, Dirac's 1928 electron---the relativistic hydrogen's fine structure---Wigner's 1931 group theory and its application to the atomic spectra, Condon and Shortley's 1935 atomic spectra canon, Fock's 1935 hydrogen symmetry---the O(4) hidden behind the spectrum---to the factorization method's systematization in Infeld and Hull's 1951 review and its supersymmetric completion in Cooper, Khare, and Sukhatme's 1995 physics reports. The synthesis is organized around three themes: the solved spectra, in which the oscillator's and hydrogen's levels fixed the new mechanics' validity; the hidden symmetries, in which Runge-Lenz and spherical symmetry explained the solvabilities; and the factorization program, in which the ladder operators' algebra became a method---the exactly solvable potentials classified. It is concluded that the solvable systems are quantum mechanics' load-bearing examples---and that their symmetries remain the theory's deepest explanation of why some problems close.
Perturbation theory for non-degenerate and degenerate systems, variational methods, and WKB approximation are covered in this chapter. First, second, and kth order corrections to energy and wave function ket of non-degenerate states, energy eigenvalues, and wave functions kets of anharmonic oscillator, first-order perturbation theory for degenerate states and linear stark effect are written under perturbation theory. Second part of the chapter includes the variational methods for computation of the energy for ground state of helium atom, Rayleigh-Ritz variational method and its application to hydrogen molecule ion, calculation of energy of excited states, and its application to 1D harmonic oscillator. The topics which are covered under WKB approximation are the classical and non-classical regions of particle motion, alternative derivation of WKB formulae, connecting formulae, and quantum condition for bound state. The kth order corrections to energy and wave function kets for non-degenerate perturbation theory, treatment of WKB approximation from two different approaches and the detailed derivation of connecting formulae, application of variational method to compute energy for excited states are topics which are not generally found in existing text books on quantum mechanics. A good number of solved examples illustrating the importance of approximation methods in contemporary fields are added at the end of chapter.
One of the basic rules of the universe is that nothing is perfect. Perfection simply doesn't exist…..Without imperfection, neither you nor I would exist —Stephen Hawking (1942–2018) Approximations are made in mathematics, physics, engineering, and life sciences for a variety of reasons such as (a) obtaining an exact solution is not possible and (b) an exact solution is not necessary, subject to a given context. Both of these factors are important in the development of approximations. We have already seen in Chapter 2 that out of the infinite alternative paths accessible to a system due to their entanglement/superposition, the classical path that makes action stationary results from the condition that action is much larger than the PEB constant. When this condition is satisfied, laws of classical physics provide an excellent approximation to quantum mechanics. In Chapter 2, we observed that the interpretation of the wavefunction as probability amplitude became the basis of the path integral formulation of quantum mechanics, in which the propagator's phase is the action . Considering this relationship, it is appropriate to wonder if the Schrödinger equation could have been predicted from classical mechanics, if only the wavefunction (Eq. 6.22, below) was introduced, even just as a definition . In Section 6.1, we demonstrate, in hindsight , how this could have been possible. Of course, historical development of quantum mechanics took place on a different course. We also demonstrate the inverse of this process, viz., how the classical equation of motion is obtainable as an approximation from quantum mechanics in the limit h →0. An approximation scheme often implemented to solve the Schrödinger equation, called the Wentzel–Kramers–Brilloiun–Jeffrey (WKBJ) method, is based on the connections between classical and quantum mechanics. It is introduced in Section 6.1. In Section 6.2, we discuss perturbation methods to obtain approximate solutions to unsolvable problems. Time-independent perturbation methods are discussed in Section 6.2, and time-dependent perturbation methods in Section 6.3. Major non-perturbative approaches to attack complex problems include the adiabatic approximation, the sudden approximation, and the method of variation. These are introduced in Section 6.4. 6.1 Classical Limit of the Schrödinger Equation and the WKBJ Approximation Readers who are well conversant with the Hamilton–Jacobi (HJ) theory in classical mechanics [1] will be able to skip a significant part of the discussion in this section. It is, however, included here, even if only briefly, for the sake of completeness.
As in classical mechanics the number of exactly soluble quantum mechanical problems is very limited and in most cases of practical interest numerical techniques and approximation methods are unavoidable. The availability of fast computers has turned quantum computation into a major industry. This chapter looks at the several approximation techniques. In seeking to solve the Schrodinger equation with the perturbed Hamiltonian, it is convenient to work with series expansions by introducing a dimensionless parameter λ. In many cases the system of interest differs from an exactly soluble system by only a small disturbance, enabling an approximation to be made by expanding in powers of a smallness parameter. The Variational Method provides a very simple method of approximating energy eigenvalues. It is most useful for computing ground state energies and its success depends on our being able to make a reasonable guess for the form of the ground state wave function.
Many physical systems cannot be solved exactly. In this case, we look for approximate methods to gain insight into their properties. Perturbation theory is one such method. When the Hamiltonian of the system doesn't change over time, one approximate method is the Time Independent Perturbation Theory, which is the subject of Chapter 10 . The infinite square well, the harmonic oscillator, and the hydrogen atom are among the few physical systems for which the Schrodinger equation can be solved exactly. For this reason, they are described in detail in most of the introductory quantum mechanics textbooks. However, most physical systems are not exactly solvable, i.e., they are described by the Schrodinger equation, whose solutions cannot be expressed in terms of commonly used functions. In order to study such systems, we try to decompose the corresponding Hamiltonian into a solvable Hamiltonian plus a small perturbation. For such cases, we say that the Hamiltonian is “close enough” to an exactly solvable Hamiltonian to warrant using perturbative methods to approximate its eigenenergies and eigenfunctions.