PHYSICAL CHEMISTRY 2 • PROBLEM-SOLVING & DATA SKILLS

Expectation Value Calculations

Extracting measurable predictions from quantum-mechanical wavefunctions through operator algebra and integration.

Historical Context & Motivation

The concept of an expectation value arose from one of the deepest philosophical shifts in physics: the realization that quantum systems do not possess definite values of all observables simultaneously. In classical mechanics, a particle at a given instant has a precise position and momentum, and any measurement simply reveals those pre-existing values. Quantum mechanics, by contrast, encodes all accessible information about a system in a wavefunction, and measurements of a particular observable generally yield a distribution of outcomes rather than a single deterministic result. The expectation value formalism provides the bridge between this probabilistic description and the experimentally reproducible averages that physical chemists actually record in the laboratory.

1925
Matrix Mechanics
Werner Heisenberg formulated quantum mechanics in terms of matrices of observable quantities, implicitly defining expectation values as diagonal matrix elements averaged over states.
1926
Wave Mechanics & the Born Rule
Erwin Schrödinger published his wave equation, and Max Born interpreted |ψ(x)|² as a probability density, giving the expectation value its modern integral form ⟨Â⟩ = ∫ψ*Âψ dx.
1927
Uncertainty Principle
Heisenberg's uncertainty relations showed that the variances (computed from expectation values) of conjugate observables obey ΔxΔp ≥ ℏ/2, making expectation value calculus essential for quantifying quantum limitations.
1930
Dirac's Bra-Ket Notation
Paul Dirac unified matrix and wave mechanics using ⟨ψ|Â|ψ⟩ notation, making expectation value expressions compact and representation-independent.
1960s–present
Computational Chemistry Era
With the advent of Hartree–Fock and density functional theory, expectation values of energy, dipole moment, and other operators became routine computational outputs used to predict molecular properties.

The central question that expectation values address is deceptively simple: given a quantum state described by ψ, what is the average outcome one would obtain if a particular observable were measured on an infinite ensemble of identically prepared systems? Mastering the computational techniques to answer this question is indispensable for physical chemistry, where molecular energies, bond lengths, dipole moments, and spectroscopic transition frequencies are all fundamentally expectation values or derived from them.

Core Principles & Definitions

Before computing any expectation value, one must internalize several foundational ideas that connect quantum states, operators, and measurement outcomes. These principles form the theoretical scaffolding on which every calculation rests, and misunderstanding any one of them leads to systematic errors in both analytical and computational work.

1

Quantum State (Wavefunction)

The state of a system is fully described by a normalized wavefunction ψ(x, t). The normalization condition ∫|ψ|² dx = 1 ensures that probabilities sum to unity, which is a prerequisite for meaningful expectation values.
2

Hermitian Operators

Every measurable physical quantity (observable) corresponds to a Hermitian operator  satisfying ⟨φ|Âψ⟩ = ⟨Âφ|ψ⟩. Hermiticity guarantees that expectation values are real numbers, as physically required.
3

The Expectation Value Postulate

For a system in state ψ, the expectation value of observable A is ⟨A⟩ = ⟨ψ|Â|ψ⟩. In the position representation this becomes the integral ∫ψ*(x) Â ψ(x) dx over all space.
4

Eigenvalue Connection

If ψ is an eigenstate of  with eigenvalue a, then ⟨A⟩ = a exactly and the variance is zero. For superposition states, ⟨A⟩ is a weighted average of eigenvalues with weights |cₙ|².
5

Variance and Uncertainty

The variance σ² = ⟨²⟩ − ⟨Â⟩² quantifies the spread of measurement outcomes. Its square root σ is the standard deviation, which appears in the Heisenberg uncertainty relation.
KEY TAKEAWAY
Think of computing an expectation value like calculating a weighted grade-point average. Each eigenvalue is like a course grade, and each |cₙ|² is the credit-hour weight assigned to that course. Just as a GPA collapses many individual grades into a single representative number, ⟨Â⟩ collapses the full spectrum of possible measurement outcomes into the single number you would report as the statistical average over many identical experiments.

Visual Explanation — From Wavefunction to Expectation Value

Panel A shows a symmetric wavefunction ψ(x) centered at the origin. Panel B displays the corresponding probability density |ψ(x)|². The dashed pink line marks ⟨x⟩, the position expectation value, which coincides with the peak for this symmetric distribution. For asymmetric wavefunctions, ⟨x⟩ shifts toward the heavier tail of |ψ|².

The diagram above illustrates the conceptual pipeline: one begins with the wavefunction ψ(x), constructs the probability density |ψ(x)|², and then takes the weighted average of the observable (here, position x) over that density. For a general operator Â, the integrand becomes ψ*(x) Â ψ(x) instead of simply x |ψ(x)|², because the operator may involve differentiation (as for momentum) or more complex algebraic forms (as for kinetic energy). The visual intuition, however, remains the same: the expectation value is the "center of mass" of the operator's action weighted by the probability density.

Mathematical Framework

The mathematical machinery for expectation value calculations rests on a small number of key equations. Each equation below is presented in the position representation; analogous expressions hold in momentum space or in the abstract Dirac notation.

GENERAL EXPECTATION VALUE
⟨A⟩ = ∫₋∞^∞ ψ*(x) Â ψ(x) dx
ψ*(x) is the complex conjugate of the wavefunction, Â is the Hermitian operator for observable A, and the integration extends over all space. This is valid for any normalized state.
POSITION EXPECTATION VALUE
⟨x⟩ = ∫₋∞^∞ ψ*(x) · x · ψ(x) dx = ∫₋∞^∞ x |ψ(x)|² dx
The position operator x̂ acts by simple multiplication. This reduces to the familiar formula for the mean of a continuous probability distribution.
MOMENTUM EXPECTATION VALUE
⟨p⟩ = ∫₋∞^∞ ψ*(x) (−iℏ d/dx) ψ(x) dx
The momentum operator p̂ = −iℏ(d/dx) involves differentiation. One must apply the derivative to ψ first, then multiply by ψ* and integrate.
VARIANCE (UNCERTAINTY SQUARED)
σ²_A = ⟨²⟩ − ⟨Â⟩² = ∫ψ* ² ψ dx − (∫ψ*  ψ dx)²
The variance measures the spread of measurement outcomes. If σ²_A = 0, the system is in an eigenstate of  and every measurement yields the same result.
💡 Superposition Shortcut
When ψ is expanded in eigenstates of Â, i.e., ψ = Σ cₙ φₙ with  φₙ = aₙ φₙ, the expectation value simplifies to ⟨A⟩ = Σ |cₙ|² aₙ. This avoids integration entirely and is the preferred approach whenever the eigenstate expansion is known.

Operator Catalog & Integration Strategies

Different observables demand different operators, and the form of the operator determines the integration strategy. A systematic awareness of the common operators encountered in physical chemistry prevents algebraic missteps and streamlines the computation process.

Six common operators are shown with their action on ψ and their classification as multiplicative or differential. The decision tree at the bottom summarizes how operator type dictates the integration approach. Symmetry arguments (e.g., odd integrand over symmetric limits → zero) are among the most powerful time-saving techniques.

A recurring theme in physical chemistry problems is the exploitation of symmetry to evaluate or simplify integrals. If ψ(x) is an even function (symmetric about x = 0), then |ψ(x)|² is also even, and the product x · |ψ(x)|² is odd. Since the integral of an odd function over symmetric limits (−∞ to +∞) is identically zero, one can immediately state ⟨x⟩ = 0 without computing anything. Similarly, the momentum expectation value ⟨p⟩ vanishes for any real-valued wavefunction, because the integrand ψ(−iℏ dψ/dx) is purely imaginary and its real part integrates to zero. Recognizing these shortcuts is the hallmark of efficient problem solving in quantum chemistry.

Summary of common operators and their expectation value integrals in one dimension
ObservableOperator (1D)Typical Integral FormUseful Shortcut
Position ⟨x⟩x̂ = x∫ x |ψ|² dx= 0 if ψ is even or odd
Momentum ⟨p⟩p̂ = −iℏ d/dx−iℏ ∫ ψ* (dψ/dx) dx= 0 if ψ is real
Kinetic Energy ⟨T⟩T̂ = −(ℏ²/2m) d²/dx²−(ℏ²/2m) ∫ ψ* (d²ψ/dx²) dxUse integration by parts → (ℏ²/2m) ∫ |dψ/dx|² dx
Potential Energy ⟨V⟩V̂ = V(x)∫ V(x) |ψ|² dxFor harmonic oscillator: ⟨V⟩ = ⟨E⟩/2 (virial theorem)
Energy ⟨E⟩Ĥ = T̂ + V̂∫ ψ* Ĥ ψ dx= Eₙ if ψ = φₙ (eigenstate)

Worked Example — Particle in a Box

Consider a particle of mass m confined to a one-dimensional box of length L. The normalized wavefunctions are ψₙ(x) = √(2/L) sin(nπx/L) for 0 ≤ x ≤ L. We will compute ⟨x⟩, ⟨x²⟩, and ⟨p⟩ for the ground state (n = 1).

Expectation Values for the Particle-in-a-Box Ground State
1
Step 1 — Write Down the Ground-State WavefunctionFor n = 1, the normalized wavefunction is ψ₁(x) = √(2/L) sin(πx/L), valid on the interval [0, L] and zero outside. This is a real-valued function, which will simplify several results.
2
Step 2 — Compute ⟨x⟩Apply the position expectation value formula: ⟨x⟩ = ∫₀ᴸ ψ₁*(x) · x · ψ₁(x) dx = (2/L) ∫₀ᴸ x sin²(πx/L) dx. Use the identity sin²θ = (1 − cos 2θ)/2, giving (2/L) ∫₀ᴸ x · [1 − cos(2πx/L)]/2 dx = (1/L) [∫₀ᴸ x dx − ∫₀ᴸ x cos(2πx/L) dx]. The first integral gives L²/2. The second integral is evaluated by parts and yields zero (the boundary terms cancel). Therefore ⟨x⟩ = (1/L)(L²/2) = L/2.
⟨x⟩ = L/2 — the particle is found, on average, at the center of the box, as expected from the symmetry of |ψ₁|².
3
Step 3 — Compute ⟨x²⟩⟨x²⟩ = (2/L) ∫₀ᴸ x² sin²(πx/L) dx. Using the same trigonometric identity and integrating by parts twice, the standard result is ⟨x²⟩ = L²(1/3 − 1/2π²). This is less than (L/2)² = L²/4, reflecting the spread of the probability distribution.
⟨x²⟩ = L²(1/3 − 1/(2π²))
4
Step 4 — Compute the Position Uncertainty ΔxThe variance is σ²ₓ = ⟨x²⟩ − ⟨x⟩² = L²(1/3 − 1/2π²) − L²/4 = L²(1/12 − 1/2π²). Taking the square root gives Δx = L√(1/12 − 1/2π²) ≈ 0.181 L.
Δx ≈ 0.181 L
5
Step 5 — Compute ⟨p⟩Since ψ₁ is real-valued, we can immediately state that ⟨p⟩ = 0. To verify: ⟨p⟩ = −iℏ (2/L) ∫₀ᴸ sin(πx/L) · (π/L) cos(πx/L) dx = −iℏ(2π/L²) · [sin²(πx/L)/(2)]₀ᴸ = −iℏ(2π/L²) · (0 − 0) = 0. Physically, the particle bounces back and forth with equal probability of moving left or right, so the average momentum vanishes.
⟨p⟩ = 0
🔍 Physical Check
Always verify results against physical intuition. For a symmetric probability distribution in a symmetric potential, ⟨x⟩ should lie at the center and ⟨p⟩ should be zero. If your calculation gives different values, re-examine the integration.

Strengths & Limitations of Calculation Methods

Expectation values can be computed through several distinct approaches, each with trade-offs in applicability, computational effort, and transparency. Selecting the right method for a given problem is itself a key problem-solving skill.

Comparison of methods for computing expectation values
MethodStrengthsLimitations
Direct IntegrationGeneral and exact; works for any wavefunction and operator. Transparent about assumptions.Integrals can be tedious or analytically intractable for complex potentials or multi-dimensional systems.
Eigenstate ExpansionConverts integral to a discrete sum ⟨A⟩ = Σ|cₙ|²aₙ. Extremely efficient when expansion coefficients are known.Requires knowing the eigenstates and eigenvalues of Â, plus the expansion coefficients cₙ.
Ladder OperatorsElegant for harmonic oscillator and angular momentum. Often avoids integration entirely using algebraic relations.Only applicable to systems with factored ladder-operator algebras (harmonic oscillator, hydrogen atom, spin).
Matrix MechanicsNatural for finite-dimensional systems (spin, truncated basis). Reduces to linear algebra: ⟨A⟩ = c†Ac.Requires truncation of the basis for continuous-variable systems, introducing approximation error.
Numerical / ComputationalHandles arbitrary potentials, many-body systems, and approximate wavefunctions (HF, DFT). Scalable.Subject to basis set incompleteness, numerical noise, and discretization errors. Results require convergence testing.
KEY TAKEAWAY
The relationship between these methods is analogous to choosing tools in engineering: a wrench (direct integration) works on any bolt but is slow, while a power ratchet (eigenstate expansion) is fast but only fits standard sizes. The expert physical chemist selects the most efficient method based on the operator, the wavefunction, and the available algebraic structure of the problem.

Connection to Advanced Theory

Expectation value calculations form the conceptual backbone of many advanced topics in physical chemistry and chemical physics. Understanding how the basic ⟨ψ|Â|ψ⟩ framework generalizes provides a roadmap for continued study.

From basic expectation values to advanced theoretical frameworks
Basic ConceptAdvanced ExtensionKey Difference
⟨Â⟩ for pure states⟨Â⟩ = Tr(ρ̂ Â) for mixed statesThe density operator ρ̂ replaces |ψ⟩⟨ψ| and handles statistical ensembles and decoherence.
Static ⟨Â⟩ (time-independent)Time-dependent ⟨Â⟩(t) and Ehrenfest's theoremd⟨Â⟩/dt = (i/ℏ)⟨[Ĥ, Â]⟩ connects quantum expectation values to classical equations of motion.
Exact wavefunctionsVariational principle: ⟨Ĥ⟩_trial ≥ E₀The energy expectation value of any trial wavefunction provides an upper bound to the true ground-state energy.
Single-particle operatorsReduced density matrices in many-electron theoryOne- and two-electron reduced density matrices allow computing expectation values without the full N-electron wavefunction.
Position-space integrationSecond quantization and field operatorsExpectation values are computed using creation/annihilation operators, essential for spectroscopy and many-body perturbation theory.

Perhaps the most consequential advanced application is the variational principle, which asserts that for any normalized trial wavefunction ψ_trial, the energy expectation value ⟨Ĥ⟩ ≥ E₀, where E₀ is the true ground-state energy. This single inequality underlies Hartree–Fock theory, configuration interaction methods, and essentially all of modern computational chemistry. Whenever you optimize a molecular geometry or compute a reaction barrier in a quantum chemistry software package, you are, at root, minimizing an expectation value of the Hamiltonian operator.

Practice Problems

PROBLEM 1CONCEPTUAL
Explain why the expectation value ⟨p⟩ is zero for any bound, stationary state with a real-valued wavefunction. Is this result still true if the wavefunction is complex? Justify your answer.
PROBLEM 2BASIC CALCULATION
For the particle in a one-dimensional box (0 ≤ x ≤ L) in the state n = 2, compute ⟨x⟩. The normalized wavefunction is ψ₂(x) = √(2/L) sin(2πx/L).
PROBLEM 3INTERMEDIATE
The ground-state wavefunction of the quantum harmonic oscillator is ψ₀(x) = (mω/πℏ)¹ᐟ⁴ exp(−mωx²/2ℏ). Compute ⟨T⟩, the expectation value of the kinetic energy, using the operator T̂ = −(ℏ²/2m) d²/dx².
PROBLEM 4APPLIED
A molecular system is prepared in the superposition state ψ = (1/√3)φ₁ + √(2/3)φ₂, where φ₁ and φ₂ are normalized eigenstates of the Hamiltonian with energies E₁ = 2.0 eV and E₂ = 5.0 eV. (a) What is ⟨E⟩? (b) What is the energy uncertainty ΔE? (c) If a spectroscopist measures the energy of 900 identically prepared molecules, approximately how many will yield E₁?
PROBLEM 5CRITICAL THINKING
Prove that for any Hermitian operator  and normalized state ψ, the variance σ²_A = ⟨( − ⟨A⟩)²⟩ ≥ 0, and explain why equality holds if and only if ψ is an eigenstate of Â. Discuss the physical implications for the uncertainty principle.

Summary — Expectation Value Calculations

Expectation value calculations connect the abstract wavefunction of a quantum system to experimentally measurable averages via the formula ⟨A⟩ = ∫ψ*Âψ dx. Every physical observable — position, momentum, kinetic energy, potential energy — is associated with a Hermitian operator that ensures the computed average is always real. The computational strategy depends on whether the operator is multiplicative or differential, and whether the eigenstate expansion of the state is known.

Key results include: ⟨x⟩ = L/2 for the particle in a box (by symmetry), ⟨p⟩ = 0 for any real stationary-state wavefunction, and the superposition formula ⟨A⟩ = Σ|cₙ|²aₙ, which replaces integration with a weighted sum when the eigenstate decomposition is available. The variance σ² = ⟨²⟩ − ⟨Â⟩² quantifies measurement spread and feeds directly into the Heisenberg uncertainty principle. Mastery of these techniques is essential for the variational method, perturbation theory, and all of computational chemistry.

Varsity Tutors • Physical Chemistry 2 • Expectation Value Calculations