Back to courses
Foundations and Mathematical Physics
FOUNDATION
12 units

Mathematical Methods for Quantum Research

A complete mathematical laboratory for quantum research. Twelve sequential units develop proof literacy, complex linear algebra, spectral theory, Hilbert spaces, analytic functions, contour methods, Fourier analysis, Green operators, stochastic dynamics, distributions, and unbounded operators. Each unit includes explicit hypotheses, derivations, checked worked examples, independent exercises, computational investigations, and mastery questions. Expanded with original in-page technical chapters, mathematical proofs, independently checked calculation baselines and directly embedded credited original instructor videos where relevant.

12 units158 hoursFOUNDATION

Published syllabus

Course Units

Unit 1
645 min5 references

Mathematical reasoning for quantum models

Develop quantum models as explicit mathematical hypotheses, derive two-state predictions, and learn to distinguish proofs from numerical evidence.

What mathematics is doing in quantum research
1. Models, idealizations, and dimensions
2. Statements, quantifiers, proof, and counterexample
3. The two-state model from complex amplitudes
4. A solvable Hamiltonian and a falsifiable prediction
Research case study: What a mathematical model does and does not prove
5. Reproduce and stress-test the model
Dimensional analysis of a driven qubit
A guided lecture and derivation
6. Exercises and solutions
Further learning and checkpoint
3 objectives
Unit 2
670 min7 references

Vectors, bases, and coordinate changes

Build a precise theory of complex vector spaces, orthonormal projections and unitary coordinate changes, with worked qubit examples.

Why coordinate independence matters
1. Span, independence, and dimension
2. Complex inner products and bra–ket notation
3. Orthogonal projection and resolution of identity
4. Unitary basis changes: the Hadamard example
5. Coordinates versus physical rotations
Research case study: The geometry hidden in a change of basis
6. A computational coordinate audit
Nonorthogonal coordinates and the dual basis
Visual intuition and the full university lecture
7. Problems, solutions, and readings
3 objectives
Unit 3
809 min5 references

Linear maps, matrices, and adjoints

Linearity makes superposition calculable: T(aψ+bφ)=aTψ+bTφ. An operator is not only a rectangle of numbers; a matrix represents it after choosing ordered bases. A map between spaces of different dimension may have a matrix that is not square, and nothing in the definition of linearity requires invertibility.

How operators act on quantum states
1. Matrices are coordinate descriptions
2. Kernel, image and rank–nullity
3. The adjoint and complex conjugation
4. Hermitian, unitary, normal and positive maps
5. Commutators, compatibility and uncertainty
6. Worked spin example
7. Practice and errors
Deep worked derivation: Operator rank and positivity
Research case study: Why singular-value decomposition is a research tool
Reproducible computation and error analysis
Nonnormal matrices, transient gain, and pseudospectra
Visual intuition and the full university lecture
Independent problems, reading, and assessment
3 objectives
Unit 4
812 min6 references

Eigenvalues, spectra, and observables

An energy eigenstate evolves by an overall phase when the Hamiltonian is time independent. For a superposition of different eigenvalues, relative phases evolve and generate interference. Spectral theory connects these statements to projectors, degeneracy, and experimental probabilities.

From matrix equations to observable outcomes
1. Eigenvectors, spectra and characteristic polynomials
2. Proof of the finite-dimensional spectral theorem
3. Two-level avoided crossings
4. Projectors and degeneracy
5. Measurements and variance
6. Numerical eigensystems
7. Exercises and solutions
Deep worked derivation: Exact spectral projectors for a two-level Hamiltonian
Research case study: A proof of adiabatic mixing in the simplest Hamiltonian
Reproducible computation and error analysis
The Hellmann–Feynman identity as a response formula
Visual intuition and the full university lecture
Independent problems, reading, and assessment
3 objectives
Unit 5
787 min5 references

Hilbert spaces and orthonormal expansions

Wavefunctions inhabit spaces of functions, not merely vectors with a very large number of entries. For square-integrable functions, two functions that differ on a set of measure zero represent the same L² element. This is one reason quantum wavefunctions should not be treated solely as pointwise graphs.

Why infinite dimensions demand completeness
1. From Euclidean vectors to L² wavefunctions
2. Cauchy sequences and completion
3. Orthogonal projection theorem
4. Bessel, Parseval, and completeness
5. Linear functionals and Riesz representation
6. Variational quantum calculations
7. Practice and pitfalls
Deep worked derivation: Bessel's inequality and the completeness criterion
Research case study: The variational principle and controlled approximation
Reproducible computation and error analysis
Approximation spaces, Galerkin equations, and convergence
A guided lecture and derivation
Independent problems, reading, and assessment
3 objectives
Unit 6
786 min4 references

Complex functions and analytic structure

A real-differentiable function of two real variables need not be holomorphic. Requiring the derivative to be independent of the complex direction of approach imposes coupled constraints on real and imaginary parts. These constraints explain why contour integrals carry so much information.

Why complex differentiability is exceptional
1. Complex differentiability and the Cauchy–Riemann equations
2. Analyticity versus real smoothness
3. Cauchy's theorem and contour independence
4. Cauchy's integral formula
5. Taylor series and the nearest singularity
6. Analytic continuation and branch cuts
7. Practice and numerical diagnostics
Deep worked derivation: Cauchy's integral formula with one and three powers
Research case study: Analytic continuation and why branches affect physics
Reproducible computation and error analysis
Cauchy estimates and the rigidity of analytic functions
Reading the proof alongside the lecture notes
Independent problems, reading, and assessment
3 objectives
Unit 7
827 min6 references

Contour integrals and residues

Residue calculus translates an integral around a closed contour into a sum of local coefficients at enclosed singularities. It does not absolve us from proving that the chosen contour is valid. For integrals on the real axis, contributions from closing arcs and any singularities on the axis require separate treatment.

A residue as local information controlling a global integral
1. Laurent series and isolated singularities
2. A first residue calculation
3. Contour orientation and winding
4. Real integral by upper-half-plane closure
5. Jordan's lemma and oscillatory factors
6. Retarded versus advanced poles
7. Practice and verification
Deep worked derivation: Residue theorem and a real improper integral
Research case study: Contour deformation and a causal propagator
Reproducible computation and error analysis
Argument principle and spectral root counting
A guided lecture and derivation
Independent problems, reading, and assessment
Advanced mathematical methods: functional analysis and approximation
3 objectives
Unit 8
898 min7 references

Fourier series and transforms

A localized signal cannot be sharply concentrated in both position and spatial frequency. Fourier transform pairs make this statement quantitative and connect the mathematics of quantum momentum to the mathematics of waves, filters, and diffusion.

The same physical state in position and frequency space
1. Fourier series as orthogonal projections
2. Convergence and Gibbs oscillations
3. The Fourier transform pair
4. Gaussians and uncertainty
5. Convolution and translation-invariant filters
6. The heat equation and smoothing
7. Sampling, FFT, and aliasing
8. Independent problems and solutions
Deep worked derivation: Gaussian Fourier transform and the heat equation
Research case study: A Fourier derivation of the quantum uncertainty principle
Reproducible computation and error analysis
Sampling, aliasing, and the limits of an FFT
Visual intuition and the full university lecture
Independent problems, reading, and assessment
Advanced mathematical methods: functional analysis and approximation
3 objectives
Unit 9
873 min5 references

Green functions and differential operators

A Green function is the response to a unit point source under a specified differential operator and specified boundary conditions. Superposition then constructs the response to an extended source. This viewpoint unifies electrostatic potentials, quantum resolvents, heat kernels, and linear response.

Solving an equation by solving for a point source
1. Operators, forcing and boundary values
2. The one-dimensional Dirichlet construction
3. Jump conditions and distributional verification
4. A solved forcing problem
5. Eigenfunction expansion of the same kernel
6. Why Neumann data changes the answer
7. Resolvents, causality, and retarded Green functions
8. Exercises, computation and checks
Deep worked derivation: A Green kernel derived from boundary conditions
Research case study: The resolvent as a spectral diagnostic
Reproducible computation and error analysis
Method of images and the half-line Green function
Green's formula in a second representation
Independent problems, reading, and assessment
Advanced mathematical methods: functional analysis and approximation
3 objectives
Unit 10
869 min6 references

Probability and stochastic processes

An experiment produces a distribution of outcomes, and a noisy device produces a distribution of histories. The mathematics of random variables, conditional expectation, and Markov processes distinguishes independent trials from correlated noise and separates stationary distributions from rapid equilibration.

Dynamics of probabilities, not just averages
1. Sample spaces and random variables
2. Conditional probability and Bayes' theorem
3. Expectation, variance and covariance
4. Finite Markov chains
5. Solving the two-state stationary model
6. Continuous-time generators and master equations
7. Random walks and diffusion limits
8. Error models and experimental inference
Deep worked derivation: Stationarity and mixing in a two-state Markov chain
Research case study: From finite Markov chains to quantum noise
Reproducible computation and error analysis
Autocorrelation and memory in the two-state chain
A guided lecture and derivation
Independent problems, reading, and assessment
Advanced mathematical methods: functional analysis and approximation
3 objectives
Unit 11
867 min4 references

Distributions and Dirac delta functions

Plane waves and point sources lead to objects that cannot be described as normalizable functions. Distributions resolve the issue by defining generalized functions through their action on smooth test functions. Their derivatives are defined by integration by parts, allowing discontinuities to produce delta terms.

Why the delta is not an infinitely tall ordinary function
1. Why point evaluation fails for ordinary L² functions
2. Test functions and distributions
3. Weak derivatives and the step function
4. The delta under a change of variables
5. Gaussian approximations to identity
6. Fourier representations and plane waves
7. Compositions, products, and caution
8. Problems with solutions
Deep worked derivation: The derivative of Heaviside is a delta distribution
Research case study: Weak solutions: making nonsmooth objects calculable
Reproducible computation and error analysis
Weak formulations as finite-element matrix equations
Impulse responses and generalized derivatives
Independent problems, reading, and assessment
Advanced mathematical methods: functional analysis and approximation
3 objectives
Unit 12
868 min4 references

Functional analysis bridge to quantum theory

In finite dimensions, a Hermitian matrix automatically defines a self-adjoint operator. On spaces of wavefunctions, differential operators are often unbounded. Their domains and boundary conditions are mathematically essential: merely writing a formally symmetric differential expression does not guarantee unique unitary evolution.

An observable is an operator together with its domain
1. Closed and densely defined operators
2. Symmetric is not necessarily self-adjoint
3. Momentum on an interval
4. Energy spectra and spectral measures
5. Unitary evolution and Stone's theorem
6. Commutators of unbounded operators
7. Variational and numerical spectral approximations
8. Capstone synthesis
Deep worked derivation: Boundary conditions determine momentum spectra
Research case study: The mathematical boundary of finite-dimensional intuition
Reproducible computation and error analysis
Strongly continuous groups are not norm continuous
Operator domains in university lecture notes
Independent problems, reading, and assessment
Advanced mathematical methods: functional analysis and approximation
3 objectives