We derive a useful expression for the matrix elements [∂f[A(t)]/∂t]ij of the derivative of a function f[A(t)] of a diagonalizable linear operator A(t) with respect to the parameter t at t0. The function f[A(t)] is supposed to be an operator acting on the same space as the operator A(t) which is assumed to have a nondegenerate, pure point spectrum. We use the basis which diagonalizes A(t0), i.e., [A(t0]ij=λiδij, and obtain [∂f[A(t)]/∂t|t=t0]ij=[∂A/∂t|t=t0]ij{[f(λj)−f(λi)]/(λj−λi)}. In addition to this, we show that further elaboration on the (not necessarily simple) integral expressions given by Wilcox (who basically considered f[A(t)] of the exponential type) and generalized by Rajagopal [who extended Wilcox results by considering f[A(t)] of the q-exponential type where expq(x)≡[1+(1−q)x]1/(1−q) with q∈R; hence, exp1(x)=exp(x)] yields these same expressions. Some of the lemmas first established by the above authors are easily recovered.
We derive a useful expression for the matrix elements [∂f[A(t)]/∂t]ij of the derivative of a function f[A(t)] of a diagonalizable linear operator A(t) with respect to the parameter t at t0. The function f[A(t)] is supposed to be an operator acting on the same space as the operator A(t) which is assumed to have a nondegenerate, pure point spectrum. We use the basis which diagonalizes A(t0), i.e., [A(t0]ij=λiδij, and obtain [∂f[A(t)]/∂t|t=t0]ij=[∂A/∂t|t=t0]ij{[f(λj)−f(λi)]/(λj−λi)}. In addition to this, we show that further elaboration on the (not necessarily simple) integral expressions given by Wilcox (who basically considered f[A(t)] of the exponential type) and generalized by Rajagopal [who extended Wilcox results by considering f[A(t)] of the q-exponential type where expq(x)≡[1+(1−q)x]1/(1−q) with q∈R; hence, exp1(x)=exp(x)] yields these same expressions. Some of the lemmas first established by the above authors are easily recovered.
Elementary parameter-differentiation techniques are developed to systematically derive a wide variety of operator identities, expansions, and solutions to differential equations of interest to quantum physics. The treatment is largely centered around a general closed formula for the derivative of an exponential operator with respect to a parameter. Derivations are given of the Baker-Campbell-Hausdorff formula and its dual, the Zassenhaus formula. The continuous analogs of these formulas which solve the differential equation dY(t)/dt = A(t) Y(t), the solutions of Magnus and Fer, respectively, are similarly derived in a recursive manner which manifestly displays the general repeated-commutator nature of these expansions and which is quite suitable for computer programming. An expansion recently obtained by Kumar and another new expansion are shown to be derivable from the Fer and Magnus solutions, respectively, in the same way. Useful similarity transformations involving linear combinations of elements of a Lie algebra are obtained. Some cases where the product eAeB can be written as a closed-form single exponential are considered which generalize results of Sack and of Weiss and Maradudin. Closed-form single-exponential solutions to the differential equation dY(t)/dt = A(t) Y(t) are obtained for two cases and compared with the corresponding multiple-exponential solutions of Wei and Norman. Normal ordering of operators is also treated and derivations, corollaries, or generalization of a number of known results are efficiently obtained. Higher derivatives of exponential and general operators are discussed by means of a formula due to Poincaré which is the operator analog of the Cauchy integral formula of complex variable theory. It is shown how results obtained by Aizu for matrix elements and traces of derivatives may be readily derived from the Poincaré formula. Some applications of the results of this paper to quantum statistics and to the Weyl prescription for converting a classical function to a quantum operator are given. A corollary to a theorem of Bloch is obtained which permits one to obtain harmonic-oscillator canonical-ensemble averages of general operators defined by the Weyl prescription. Solutions of the density-matrix equation are also discussed. It is shown that an initially canonical ensemble behaves as though its temperature remains constant with a ``canonical distribution'' determined by a certain fictitious Hamiltonian.
Ψ=Ψ Iˆ In view of the expansion (29.28), which holds for an arbitrary state function in the form Ψ = Ψ Ψ Ψ Ψ = Ψ Ψ Ψ∑ ∑ we may define the unit or identity operator as i iI ΨΨ=∑ =1 ˆ (33.4) Conversely, for every linear operator we can form a square matrix, in terms of a complete set of functions by multiplying the left-hand side of Eq.(33.3) by and integrating, which gives fˆ ,iΨ ∗Ψi ΨΨ=ΨΨ=ΦΨ=ΦΨ∫ ∗ Iffd iiii ˆˆˆV ΨΨ=ΨΨΨΨ= n ˆˆ Since ii b=ΦΨ this result is identical with that given by Eq.(33.1), provided we set Vˆˆ dfff kikiik ΨΨ=ΨΨ= ∫ ∗ (33.5) which are called the matrix elements of the linear operator in terms of the particular complete set of functions It is a straightforward matter to show that the rules of matrix algebra for addition, multiplication, associativity and distributivity (see Appendix II) apply to the matrix elements (33.5), following from their definition. Two properties of fˆ .iΨ matrix elements that are needed in quantum mechanics can be readily derived from Eq.(33.5). Using Eq.(29.19) we obtain ∗∗ =ΨΨ=ΨΨ= kiikkiik ffff ˆˆ (33.6) which means that if is a Hermitian operator, then its matrix representation in a complete set of functions is Hermitian. Hence, the elements in the principal diagonal of a Hermitian matrix must be real. When all the elements are real, the matrix must be symmetric. From Eq.(29.20) we have fˆ ikiikiik f δλλ =ΨΨ=ΨΨ ˆ (33.7) that is: if the functions are the eigenfunctions of the Hermitian operator the matrix representation is diagonal, and the matrix elements are the eigenvalues of the operator. A quantum mechanical problem can be stated in a differential equation form or in a matrix form. In both cases the solution consists of finding the eigenvalues and the eigenfunctions of an operator associated with a physical observable. In view of a matrix representation in any complete set of functions, the eigenvalue equation (29.20) must be formulated as iΨ ,fˆ (33.8) Ψ=Ψ λfˆ where the arbitrary eigenfunction Ψ can be expanded, as given by Eq.(29.4), in terms of a complete set of functions If we multiply Eq.(33.8) by and integrate, we obtain .iΨ ∗Ψi ˆ ˆ ˆV V ori i if d d f Iλ λ∗ ∗Ψ Ψ = Ψ Ψ Ψ Ψ = Ψ Ψ∫ ∫ i or also, in view of Eq.(33.4) k kki f λ ˆ This can be rewritten as = n 1 λ which reads (33.9) ⎟⎟ ⎟⎟ ⎟ ⎠ ⎞ ⎜⎜ ⎜⎜ ⎜ ⎝ ⎛ = ⎟⎟ ⎟⎟ ⎟ ⎠ ⎞ ⎜⎜ ⎜⎜ ⎜ ⎝ ⎛ ⎟⎟ ⎟⎟ ⎟ ⎠ ⎞ ⎜⎜ ⎜⎜ ⎜ ⎝ ⎛ a a a a a a fff fff fff …… … ……… … … λ Equation (33.6) is called the matrix eigenvalue equation and consists of n linear homogenous equations for the given by ,ka ( ) 0... 1212111 =+++− nnafafaf λ ( ) 0... 2222121 =++−+ nnafafaf λ (33.10) ……… ( ) 0...2211 =−+++ nnnnn afafaf λ This system of simultaneous equations has a nontrivial solution for the if and only if the characteristic determinant vanishes ( ) ( ) ( ) ( ) 01ˆˆdet = ⎟⎟ ⎟⎟ ⎟ ⎠ ⎞ ⎜⎜ ⎜⎜ ⎜ ⎝ ⎛ − − − =− λ λ λ λ nnnn fff fff fff f … ……… … … (33.11) If expanded, the determinant gives a polynominal of nth degree called the characteristic polynominal of fˆ with n roots which are the desired eigenvalues. For each eigenvalue Eqs.(33.10) can be solved for the giving the corresponding eigenfunction that is a column matrix.
It is shown that repeated differentiation, with respect to the perturbation parameter λ, of the equation 〈ψm (λ)‖H0+λV‖ψn(λ)〉 =En(λ)δmn, and methodical application of the closure relation ∑‖ψn〉〈ψn‖=1, provide a painless technique for deriving the energy corrections in the Rayleigh–Schrödinger perturbation theory. The procedure is compared with one based on solving the secular equation D(E)=0, where D is the determinant whose elements are 〈ψi‖H0+V‖ψj〉−Eδij.
Abstract When a mathematical operation (such as multiplication, division, or differentiation) is performed on a function, the result is generally some different function. For instance, differentiation of x2 yields the different function 2x. However, for some combinations of operations and functions the same function is regenerated, but perhaps multiplied by a number (Fig. E.l). Thus differentiation of e2x gives 2e2X, which is the same function multiplied by 2. In such cases, the function is said to be an eigenfunction of the operator (in this case the differential operator d/dx), and the numerical factor (2 in the example) is called the eigenvalue of the operator. (Eigen is the German word meaning ‘own’ or ‘characteristic’.) If an operator Q operates on a function f and produces wf, where w is a number, f is an eigenfunction of Q and w is the corresponding eigenvalue. An expression of the form Qi= wf is called an eigenvalue equation.
We study the statistics of matrix elements of local operators in the basis of energy eigenstates in a paradigmatic, integrable, many-particle quantum theory, the Lieb-Liniger model of bosons with repulsive delta-function interactions. Using methods of quantum integrability, we determine the scaling of matrix elements with system size. As a consequence of the extensive number of conservation laws, the structure of matrix elements is fundamentally different from, and much more intricate than, the predictions of the eigenstate thermalization hypothesis for generic models. We uncover an interesting connection between this structure for local operators in interacting integrable models and the one for local operators that are not local with respect to the elementary excitations in free theories. We find that typical off-diagonal matrix elements ⟨ μ | O | λ ⟩ in the same macrostate scale as exp ( − c O L ln ( L ) − L M μ , λ O ) , where the probability distribution function for M μ , λ O is well described by Fréchet distributions and c O depends only on macrostate information. In contrast, typical off-diagonal matrix elements between two different macrostates scale as exp ( − d O L 2 ) , where d O depends only on macrostate information. Diagonal matrix elements depend only on macrostate information up to finite-size corrections. Published by the American Physical Society 2024