Quantum Computing · Module 1 Prof. (Dr.

) Abhijit Debnath

Elementary Quantum Mechanics

Linear Algebra, Quantum States in Hilbert Space, and the Bloch Sphere

Dr. Abhijit Debnath

University of Engineering & Management, Kolkata

1 General Introduction and Historical Background

Quantum mechanics is the theoretical framework that describes nature at the scale of atoms,
electrons, and photons. Between roughly 1900 and 1930 a sequence of discoveries forced physicists
to abandon the deterministic picture of classical mechanics. Max Planck (1900) introduced energy

h
quantisation to explain black-body radiation; Albert Einstein (1905) used the same idea to explain
the photoelectric effect; Niels Bohr (1913) quantised atomic orbits; and Louis de Broglie (1924)

at
proposed that matter itself has wave character.
The decisive step came in 1925–1926, when two seemingly different theories appeared almost

bn
simultaneously. Werner Heisenberg, working with Max Born and Pascual Jordan, formulated
matrix mechanics, in which physical observables are represented by arrays of numbers that do
not generally commute. Erwin Schrödinger independently formulated wave mechanics, in which
the state of a system is a continuous wave function. Schrödinger and later Paul Dirac and John
De
von Neumann proved that the two formulations are mathematically equivalent: both are concrete
realisations of an abstract structure built on linear algebra over complex vector spaces. Von
Neumann’s 1932 treatise Mathematical Foundations of Quantum Mechanics placed the theory on
the rigorous footing of Hilbert spaces, the language we still use today.
This guide develops three closely connected foundations of that language. First, the linear algebra
jit
of complex vectors, inner products, and operators that supplies the grammar of the theory. Second,
the notion of a quantum state living in a Hilbert space, including superposition, measurement,
and the Born rule. Third, the Bloch sphere, a beautiful geometric picture of the simplest
possible quantum system, the two-level system or qubit, which is the workhorse of modern quantum
hi

information science. Throughout we use Dirac’s bra–ket notation, introduced by Dirac in 1939,
which has become the universal shorthand of the field.
Ab

Two features of the quantum description deserve emphasis at the outset, because they distinguish it
sharply from every classical theory. The first is the essential appearance of complex numbers. In
classical physics complex numbers are a convenience for bookkeeping oscillations, and the physical
answer is always the real part. In quantum mechanics the complex phase of an amplitude is
physically real: it is what produces interference, and removing it would destroy the predictions
of the double-slit experiment and of every quantum algorithm. The second feature is that the
.

theory is irreducibly probabilistic. Given complete knowledge of the state, quantum mechanics
Dr

still predicts only the statistics of measurement outcomes, not the individual results. These two
ingredients — complex amplitudes that interfere and a rule turning amplitudes into probabilities
— are the conceptual heart of everything below, and each of our three topics is, in the end, a way
of organising them.

2 Linear Algebra for Quantum Mechanics

2.1 Introduction and History
Linear algebra is not merely a computational tool in quantum mechanics; it is the structure of
the theory. When Heisenberg wrote down his 1925 multiplication rule for observable quantities,
he had effectively rediscovered matrix multiplication, which Born immediately recognised. The
non-commutativity of matrices, AB ̸= BA, turned out to encode the most famous feature of the
quantum world, the uncertainty principle. Where classical physics uses real numbers and ordinary

1
Quantum Computing · Module 1 Prof. (Dr.) Abhijit Debnath

calculus, quantum physics uses complex vector spaces, linear operators, eigenvalues, and inner
products. Mastering this algebra is the single most important prerequisite for everything that
follows.
Historically, it is worth appreciating how surprising this was. Heisenberg, working on the island
of Helgoland in June 1925 to escape his hay fever, was trying only to compute the intensities of
spectral lines using observable frequencies. He arrived at tables of numbers that multiplied in a
strange order-dependent way; he did not know he had reinvented matrices. Born, on seeing the
manuscript, recalled his student-era linear algebra and, with Jordan, recast the scheme as genuine
matrix mechanics, culminating in the canonical commutation relation [x̂, p̂] = iℏ. That a piece of
pure nineteenth-century mathematics — matrix algebra, developed by Cayley and Sylvester with
no physical application in mind — should turn out to be the exact language of the atom is one of

h
the most striking examples of Eugene Wigner’s “unreasonable effectiveness of mathematics in the
natural sciences.”

at
2.2 Mathematical Formulation

bn
2.2.1 State vectors, bras, and kets
A quantum state is represented by a column vector in a complex vector space, written as a ket
|ψ⟩. For a finite d-dimensional system,
 
c1
c2 
|ψ⟩ =  .  ,
 
De
ci ∈ C.
 .. 
cd
jit
The corresponding bra ⟨ψ| is the conjugate transpose (Hermitian adjoint), a row vector:

⟨ψ| = |ψ⟩† = c∗1 c∗2 · · · c∗d .

hi

2.2.2 Inner product

The inner product of two states is the complex number
Ab

d
X
⟨ϕ|ψ⟩ = ϕ∗i ψi .
i=1

It satisfies ⟨ϕ|ψ⟩ = ⟨ψ|ϕ⟩∗ , is linear in the second argument, and gives a non-negative norm
∥ψ∥2 = ⟨ψ|ψ⟩ ≥ 0. Physical states are normalised, ⟨ψ|ψ⟩ = 1. Two states are orthogonal when
.

⟨ϕ|ψ⟩ = 0.
Dr

2.2.3 Operators

A linear operator  maps kets to kets and is represented by a d × d matrix. Three classes are
central:
• Hermitian (self-adjoint): † = Â. These represent observables; their eigenvalues are real and
represent possible measurement outcomes.

• Unitary: Û † Û = ⊮. These represent time evolution and quantum gates; they preserve inner
products and hence normalisation.

• Projectors: P̂ 2 = P̂ and P̂ † = P̂ . These represent measurement onto a subspace.

2
Quantum Computing · Module 1 Prof. (Dr.) Abhijit Debnath

h
at
Figure 1: Left: a complex amplitude c = reiϕ has a magnitude r and a phase ϕ. Right: the inner

bn
product measures overlap; the projection of |b⟩ onto |a⟩ is the geometric content of ⟨a|b⟩.

2.2.4 Eigenvalues and the spectral picture

De
The eigenvalue equation  |a⟩ = a |a⟩ identifies the special directions left invariant by Â. For a
Hermitian operator the eigenvalues aPare real and the eigenvectors form an orthonormal basis, so
any state can be expanded as |ψ⟩ = a ⟨a|ψ⟩ |a⟩.
The physical content of this purely algebraic fact is profound. Because measurement results are real
jit
numbers, the operators representing observables must have real eigenvalues — and the theorem
that Hermitian operators have real eigenvalues is exactly what guarantees this. Furthermore,
because the eigenvectors of a Hermitian operator are mutually orthogonal, the possible outcomes
of a measurement are perfectly distinguishable: a system found to have energy E1 is in a state
hi

orthogonal to one found to have E2 . Thus a single mathematical property, † = Â, simultaneously
secures real outcomes and a sensible probability rule. This is the sense in which the linear algebra
is not a description of the physics but is the physics, expressed in its native language.
. Ab
Dr

Figure 2: Action of a Hermitian matrix on the unit circle. The circle becomes an ellipse; the
eigenvectors (teal) point along the principal axes and are merely stretched by their eigenvalues
(red).

3
Quantum Computing · Module 1 Prof. (Dr.) Abhijit Debnath

2.2.5 Outer products, projectors, and completeness

While the inner product ⟨ϕ|ψ⟩ produces a number, the outer product |ϕ⟩ ⟨ψ| produces an operator
(a matrix). The most important special case is P̂a = |a⟩ ⟨a|, the projector onto the state |a⟩: acting
on any |ψ⟩ it returns ⟨a|ψ⟩ |a⟩, the component of |ψ⟩ “along” |a⟩. For any orthonormal basis {|a⟩}
the projectors sum to the identity, X
|a⟩ ⟨a| = ⊮,
a

a statement called the completeness relation or resolution of the identity. Inserting it between
P two objects is the workhorse trick of practical quantum mechanics: it lets us expand |ψ⟩ =
any
a ⟨a|ψ⟩ |a⟩ and rewrite operators in a chosen
P
basis at will. The spectral theorem then says that
every Hermitian operator can be written  = a a |a⟩ ⟨a|, a weighted sum of orthogonal projectors

h
— its eigenvalues weighting its eigen-projectors.

at
2.2.6 Commutators and the Pauli matrices

The commutator [Â, B̂] = ÂB̂ − B̂ Â measures the failure of two operators to commute. For

bn
two-level systems the fundamental observables are the Pauli matrices:
     
0 1 0 −i 1 0
σx = , σy = , σz = ,
1 0 i 0 0 −1
De
which satisfy [σx , σy ] = 2i σz (and cyclic permutations) and σj2 = ⊮.

2.2.7 Tensor products

Combining systems uses the tensor product ⊗. Two qubits live in a 2 × 2 = 4-dimensional space
jit
spanned by |00⟩ , |01⟩ , |10⟩ , |11⟩. In coordinates, the tensor product of (a1 , a2 )⊤ and (b1 , b2 )⊤ is the
stacked 4-vector (a1 b1 , a1 b2 , a2 b1 , a2 b2 )⊤ . The exponential growth of dimension with the number of
parts — n qubits require 2n complex amplitudes — is precisely what makes quantum systems hard
hi

to simulate classically and powerful for computation. Crucially, most vectors in the joint space
cannot be written as a single product |u⟩⊗|v⟩; those that cannot are entangled, and entanglement
is the uniquely quantum resource that has no counterpart in classical probability.
Ab

Key identities. ⟨ϕ|ψ⟩ = ⟨ψ|ϕ⟩∗ ; (ÂB̂)† = B̂ † †

 Â ; Hermitian
 ⇒ real eigenvalues; unitary
⇒ | det Û | = 1 and norm-preserving; Tr ÂB̂ = Tr B̂ Â .

2.3 Worked Numerical Problems (Five Types)

.

2


Type 1 — Inner product and orthogonality. Let |a⟩ = ( 1i ) and |b⟩ =
Dr

−i (unnor-
malised). Compute ⟨a|b⟩.

Solution. ⟨a| = (1, −i), so ⟨a|b⟩ = (1)(2) + (−i)(−i) = 2 + i2 = 2 − 1 = 1. Since the result is
non-zero, the states are not orthogonal.

Type 2 — Normalisation. Normalise |ψ⟩ = 3 |0⟩ + 4i |1⟩.

Solution. ⟨ψ|ψ⟩ = |3|2 + |4i|2 = 9 + 16 = 25, so ∥ψ∥ = 5. The normalised state is |ψ⟩norm =
1
5 (3 |0⟩ + 4i |1⟩), with probabilities 9/25 and 16/25.

4
Quantum Computing · Module 1 Prof. (Dr.) Abhijit Debnath

Type 3 — Eigenvalues and eigenvectors. Find the eigenvalues and eigenvectors of

σx = ( 01 10 ).

Solution. det(σx − λ⊮) = λ2 − 1 = 0 ⇒ λ = ±1. For λ = +1: eigenvector √1 ( 1 ) = |+⟩. For

2 1
λ = −1: √12 −1
1


= |−⟩. These are the eigenstates of a spin measured along x.

Type 4 — Commutator. Verify [σx , σy ] = 2iσz .

h
0 −i i 0 −i 0 2i 0





Solution. σx σy = ( 01 10 ) i 0 = 0 −i , and σy σx = 0 i . Hence [σx , σy ] = 0 −2i =
1 0


2i 0 −1 = 2iσz .

at
√1 1 1


Type 5 — Unitarity check / tensor product. (a) Show the Hadamard H = 2 1 −1
is unitary. (b) Write |0⟩ ⊗ |1⟩ as a 4-vector.

bn
Solution. (a) H † = H and H 2 = 21 ( 20 02 ) = ⊮, so H † H = ⊮: unitary. (b) |0⟩ = ( 10 ), |1⟩ = ( 01 ), so
|0⟩ ⊗ |1⟩ = (1 · 0, 1 · 1, 0 · 0, 0 · 1)⊤ = (0, 1, 0, 0)⊤ = |01⟩ .
De
Bonus — Expanding a state in a basis (completeness in action). Write |0⟩ in the
{|+⟩ , |−⟩} basis.
jit
Solution. Insert the identity |0⟩ = |+⟩ ⟨+|0⟩ + |−⟩ ⟨−|0⟩. Since ⟨+|0⟩ = ⟨−|0⟩ = √1 , we get
2
|0⟩ = √1 (|+⟩
+ |−⟩). The same physical state, two different coordinate descriptions — exactly
2
what a change of basis means.
hi

2.4 Detailed Pros and Cons

Ab

Strengths. (i) Linear algebra makes the theory exactly solvable for finite systems and gives
a single unified language for spin, polarisation, and qubits. (ii) Operators encode observables,
dynamics, and measurement in one framework. (iii) Powerful theorems (spectral theorem,
singular-value decomposition) carry over directly. (iv) It scales conceptually to infinite di-
mensions via functional analysis.
Limitations. (i) The exponential growth of dimension under tensor products makes clas-
.

sical computation intractable for many particles. (ii) Infinite-dimensional spaces (position,
Dr

momentum) require care: unbounded operators, domains, and rigged Hilbert spaces. (iii) The
abstraction can obscure physical intuition for beginners. (iv) Linearity alone does not tell you
which operator is the correct Hamiltonian; that input comes from physics, not algebra.

2.5 Exercises
1. Normalise |ψ⟩ = |0⟩ + (1 + i) |1⟩ and give the measurement probabilities.

2. Compute ⟨+|0⟩ and ⟨+|−⟩ where |±⟩ = √1 (|0⟩ ± |1⟩).

2

3. Find the eigenvalues and eigenvectors of σy .

4. Evaluate [σy , σz ] and confirm the cyclic rule.

5
Quantum Computing · Module 1 Prof. (Dr.) Abhijit Debnath

5. Show that any 2 × 2 Hermitian matrix can be written a ⊮ + ⃗b · ⃗σ with a, ⃗b real.

6. Prove that the eigenvalues of any unitary operator have modulus 1.

3 Quantum States in Hilbert Space

3.1 Introduction and History
The phrase “state in Hilbert space” encodes one of the deepest shifts in physics. In classical me-
chanics the state of a particle is a point (⃗x, p⃗) in phase space, and everything about the future is
determined. In quantum mechanics the state is a vector |ψ⟩ in a complex Hilbert space, and the
theory predicts only probabilities of measurement outcomes. The Hilbert space — named after

h
David Hilbert, whose early-1900s work on integral equations and infinite-dimensional spaces sup-
plied the mathematics — was adopted by von Neumann in 1932 as the rigorous arena for quantum

at
theory. Max Born’s 1926 probabilistic interpretation of the wave function (for which he received
the 1954 Nobel Prize) is the bridge between the abstract vector and laboratory measurement.
It is hard to overstate how radical the probabilistic reading was at the time. Schrödinger himself

bn
had hoped his wave ψ(x) described a literal smeared-out charge density; Einstein never accepted
that “God plays dice.” Born’s insight — that |ψ(x)|2 is a probability density for finding the
particle at x, not a physical density of stuff — survived every experimental test and remains the
standard interpretation of the formalism. The Hilbert-space picture is what makes this precise:
De
the inner product supplies the probabilities, the completeness of the space guarantees that limiting
procedures (expansions in energy eigenstates, Fourier transforms) are well defined, and the linearity
of the space is exactly the superposition principle.

3.2 Mathematical Formulation

jit
3.2.1 What a Hilbert space is
A Hilbert space H is a complex vector space equipped with an inner product ⟨·|·⟩ that is com-
plete (every Cauchy sequence converges). Finite-dimensional examples are just Cd ; the infinite-
hi

dimensional example Rrelevant to a particle on a line is L2 (R), the space of square-integrable func-
∞
tions, where ⟨ϕ|ψ⟩ = −∞ ϕ∗ (x)ψ(x) dx.
Ab

3.2.2 The postulates of state and measurement

1. States. A pure state is a unit ray |ψ⟩ ∈ H; |ψ⟩ and eiγ |ψ⟩ describe the same physics (global
phase is unobservable).
P
2. Observables. Each measurable quantity is a Hermitian operator  = a a P̂a , where P̂a
.

projects onto the eigenspace with eigenvalue a.

Dr

3. Born rule. Measuring  yields outcome a with probability P (a) = ⟨ψ| P̂a |ψ⟩ = | ⟨a|ψ⟩ |2 for
a non-degenerate eigenstate.
p
4. Collapse. Immediately after obtaining a, the state becomes P̂a |ψ⟩ / P (a).

5. Dynamics. Between measurements the state evolves unitarily by the Schrödinger equation
iℏ ∂t |ψ⟩ = Ĥ |ψ⟩, giving |ψ(t)⟩ = e−iĤt/ℏ |ψ(0)⟩.

6
 Quantum Computing · Module 1 Prof. (Dr.) Abhijit Debnath

3.2.3 Superposition and expectation values

The defining quantum feature is superposition: if |ψ1 ⟩ and |ψ2 ⟩ are allowed states, so is α |ψ1 ⟩ +
β |ψ2 ⟩. The expectation value (mean of many measurements) of an observable is
D E X
 = ⟨ψ|  |ψ⟩ = a P (a),
a
rD E D E2
and the spread is ∆A = Â2 − Â , which enters the Heisenberg uncertainty relation ∆A ∆B ≥

th
D E
1
2 | [ Â, B̂] |.

Superposition must be carefully distinguished from classical mixing. A coin hidden under a hand

na
is in one definite state — heads or tails — that we merely fail to know; its description is a
classical probability distribution. A superposition α |0⟩ + β |1⟩ is not “secretly |0⟩ or |1⟩.” It is
a genuinely new state with its own definite value of some other observable: |+⟩ = √12 (|0⟩ + |1⟩)
is a perfectly sharp eigenstate of σx , even though it is maximally uncertain in σz . Interference
experiments confirm that the two branches of a superposition are simultaneously present and can

eb
cancel or reinforce, something no classical “ignorance” distribution can do. The distinction between
coherent superposition and incoherent mixture is exactly the distinction between a state on the
surface of the Bloch sphere and one in its interior, the subject of the third part of this guide.

D
t
iji
bh
A

Figure 3: A two-level state |ψ⟩ = α |0⟩ + β |1⟩ as a vector whose components along the orthonormal
r.

basis {|0⟩ , |1⟩} are the probability amplitudes. Squaring the components gives outcome probabil-
ities.
D

3.2.4 The continuous case: wave functions

For a particle moving on a line the Hilbert space is infinite-dimensional. The state is a wave function
ψ(x) = ⟨x|ψ⟩, the “coordinate” of |ψ⟩ along R the (continuum) position basis |x⟩.R Sums become
integrals: the completeness relation reads |x⟩ ⟨x| dx = ⊮, normalisation reads |ψ(x)|2 dx = 1,
and Born’s rule says |ψ(x)|2 dx is the probability of finding the particle in [x, x + dx]. Position
and momentum become operators x̂ and p̂ = −iℏ d/dx obeying [x̂, p̂] = iℏ, the continuum origin
of the uncertainty principle. The finite-dimensional intuition built from qubits carries over almost
verbatim — which is exactly why the abstract Hilbert-space viewpoint is so valuable.

7
Quantum Computing · Module 1 Prof. (Dr.) Abhijit Debnath

3.2.5 Entanglement
For two or more subsystems, a joint state is separable if it factorises, |Ψ⟩ = |u⟩ ⊗ |v⟩, and entangled
otherwise. The canonical example is the Bell state |Φ+ ⟩ = √12 (|00⟩+|11⟩), which cannot be written
as a product of single-qubit states. Measuring the first qubit instantly fixes the outcome of the
second, no matter how far apart they are — the correlation that so troubled Einstein, Podolsky,
and Rosen in 1935 and that Bell’s 1964 theorem later showed is strictly stronger than any classical
(“local hidden-variable”) correlation. Entanglement is the central resource of quantum computing,
teleportation, and cryptography.

3.2.6 Density matrices and mixed states

h
P as the projector ρ = |ψ⟩
A pure state can be written P⟨ψ|. More generally a statistical mixture is
† , Tr ρ = 1, ρ ≥ 0,
the density matrix ρ = k pk |ψk ⟩ ⟨ψk | with pk ≥ 0, kDpk E= 1. Itsatisfies ρ = ρ

at

and is pure iff Tr ρ2 = 1. Expectation values become  = Tr ρ .

bn
De
jit
hi

Figure 4: The Born rule in action: outcome probabilities for measuring three different states in
the computational basis {|0⟩ , |1⟩}.
Ab

3.3 Worked Numerical Problems (Five Types)

3
Type 1 — Born-rule probabilities. For |ψ⟩ = 5 |0⟩ + 45 |1⟩, find P (0) and P (1).

Solution. P (0) = |3/5|2 = 9/25 = 0.36, P (1) = |4/5|2 = 16/25 = 0.64. They sum to 1, as
.

required.
Dr

Type 2 — Expectation value. For the same state, compute ⟨σz ⟩.

Solution. σz |0⟩ = + |0⟩ , σz |1⟩ = − |1⟩, so ⟨σz ⟩ = (+1)(0.36) + (−1)(0.64) = −0.28.

Type 3 — Measurement in a rotated basis. Measure |0⟩ in the {|+⟩ , |−⟩} basis. What
are the probabilities?

Solution. ⟨+|0⟩ = √1 , ⟨−|0⟩ = √1 , so P (+) = P (−) = 1

2 2 2. A definite z-state is maximally
uncertain in x.

8
Quantum Computing · Module 1 Prof. (Dr.) Abhijit Debnath

Type 4 — Time evolution. A state |ψ(0)⟩ = |+⟩ evolves under Ĥ = ℏω

2 σz . Find |ψ(t)⟩.

Solution. e−iĤt/ℏ = e−iωtσz /2 = diag(e−iωt/2 , e+iωt/2 ). Acting on |+⟩ = √1 (|0⟩

2
+ |1⟩) gives
|ψ(t)⟩ = √1 (e−iωt/2 |0⟩ + e+iωt/2 |1⟩): the relative phase ωt rotates the state — Larmor precession.
2

Type 5 — Purity of a density matrix. Is ρ = 21 (|0⟩ ⟨0| + |1⟩ ⟨1|) pure or mixed?

Solution. ρ = 21 ⊮, so ρ2 = 41 ⊮ and Tr ρ2 = 1


< 1. It is the maximally mixed state — complete

h
2
ignorance of the qubit.

at
3.4 Detailed Pros and Cons
Strengths. (i) The Hilbert-space formalism makes spectacularly accurate predictions (e.g. the

bn
electron g-factor to 12 digits). (ii) It unifies discrete (spin) and continuous (position) systems.
(iii) Superposition and entanglement, which have no classical analogue, emerge naturally. (iv)
Density matrices extend the theory cleanly to noise, mixtures, and open systems.
De
Limitations. (i) The theory predicts only probabilities; individual outcomes are irreducibly
random. (ii) The “collapse” postulate is non-unitary and sits uneasily beside smooth Schrödinger
evolution — the measurement problem. (iii) Interpretations (Copenhagen, many-worlds,
Bohmian, etc.) disagree on what the state vector is. (iv) Infinite-dimensional spaces demand
careful functional analysis that elementary treatments gloss over.
jit

3.5 Exercises
q
1. For |ψ⟩ = √1 |0⟩ + 2
eiπ/4 |1⟩, find P (0), P (1), and ⟨σz ⟩.
hi

3 3

2. Show that a global phase eiγ leaves all Born probabilities unchanged.
Ab

3. Compute ⟨σx ⟩ for |ψ⟩ = cos 2θ |0⟩ + sin 2θ |1⟩.

4. Verify the uncertainty relation for σx and σy in the state |0⟩.

3
5. Determine whether ρ = 4 |0⟩ ⟨0| + 14 |1⟩ ⟨1| is pure, and compute ⟨σz ⟩.

6. Evolve |0⟩ under Ĥ = ℏω

2 σx and describe the motion.
.
Dr

4 The Bloch Sphere

4.1 Introduction and History
The Bloch sphere, named after the physicist Felix Bloch (who introduced the closely related “Bloch
vector” in his 1946 work on nuclear magnetic resonance), is the geometric representation of the
state space of a single qubit — the simplest non-trivial quantum system. Any two-level system
(electron spin, photon polarisation, a two-level atom, a superconducting qubit) shares this same
state space. The sphere turns abstract complex amplitudes into a vivid, manipulable picture: every
pure state is a point on the surface, every quantum gate is a rotation, and decoherence shrinks the
radius toward the centre. It is arguably the most important single diagram in quantum information
science.

9
Quantum Computing · Module 1 Prof. (Dr.) Abhijit Debnath

Bloch’s original context was magnetic resonance, where a large ensemble of nuclear spins precessing
in a magnetic field is described by a macroscopic magnetisation vector that traces exactly the same
geometry. The optical analogue — the Poincaré sphere of light polarisation, introduced by Henri
Poincaré in 1892 — is mathematically identical, with the poles representing right- and left-circular
polarisation and the equator representing linear polarisations. The recognition in the 1980s and
1990s that this one picture governs every qubit, whatever its physical embodiment, is part of why
quantum information became a unified field rather than a collection of separate technologies.

4.2 Mathematical Formulation

4.2.1 Parametrising a qubit

h
A normalised qubit state, up to an irrelevant global phase, can always be written with two real
angles:

at
θ θ
|ψ⟩ = cos |0⟩ + eiϕ sin |1⟩ 0 ≤ θ ≤ π, 0 ≤ ϕ < 2π.
2 2
Here θ is a polar angle and ϕ an azimuthal angle. The factor of 2θ (rather than θ) is what makes

bn
orthogonal states (|0⟩ at θ = 0, |1⟩ at θ = π) sit at antipodal points of the sphere.
Two subtleties justify this parametrisation. First, a general qubit has two complex amplitudes, i.e.
four real numbers; normalisation removes one and the unobservable global phase removes another,
De
leaving exactly the two real angles (θ, ϕ) that a sphere provides. Second, although global phase
is invisible, the relative phase ϕ between the two amplitudes is fully physical and is precisely the
longitude on the sphere — rotating ϕ carries |+⟩ to |+i⟩ to |−⟩ around the equator, and these are
genuinely different, distinguishable states. The Bloch sphere is therefore not an arbitrary visual
aid but the exact, faithful picture of the qubit’s true state space once redundant information has
jit
been stripped away.

4.2.2 The Bloch vector

hi

Associate to the state the real 3-vector

⃗r = (sin θ cos ϕ, sin θ sin ϕ, cos θ),

Ab

which has unit length for any pure state. Equivalently, the density matrix is
 
1 1 1 + rz rx − iry
ρ = (⊮ + ⃗r · ⃗σ ) = ,
2 2 rx + iry 1 − rz

and the Bloch components are simply expectation values: rx = ⟨σx ⟩, ry = ⟨σy ⟩, rz = ⟨σz ⟩.
.
Dr

4.2.3 Cardinal states and mixed states

The six cardinal states sit at the axis intercepts: |0⟩ (north pole), |1⟩ (south pole), |±⟩ = √1 (|0⟩ ±
2
|1⟩) on the ±x axis, and |±i⟩ = √1 (|0⟩ ± i |1⟩) on the ±y axis. Mixed states have |⃗r| < 1 and live
2
inside the ball; the centre ⃗r = 0 is the maximally mixed state 12 ⊮.

4.2.4 Gates as rotations

γ
A single-qubit unitary Û = e−i 2 n̂·⃗σ acts on the Bloch vector as a rigid rotation by angle γ
about the axis n̂. Thus X (σx ) is a 180◦ rotation about x, the Hadamard maps |0⟩ → |+⟩ (a
rotation swapping the z and x poles), and phase gates rotate about z. This is why the sphere is
so powerful: algebra of 2 × 2 unitaries becomes geometry of SO(3) rotations.

10
Quantum Computing · Module 1 Prof. (Dr.) Abhijit Debnath

h
at
bn
Figure 5: A general pure qubit state |ψ⟩ = cos 2θ |0⟩ + eiϕ sin 2θ |1⟩ sits on the surface of the unit
sphere, with polar angle θ from the north pole |0⟩ and azimuthal angle ϕ in the equatorial plane.

4.2.5 Dynamics, Rabi oscillations, and decoherence

De
The same rotation picture describes real laboratory control. Driving a two-level atom with a
resonant laser, or a spin with an oscillating magnetic field, makes the Bloch vector rotate steadily
about an axis in the equatorial plane; the population oscillates between |0⟩ and |1⟩ at the Rabi
frequency. A pulse timed to produce a 180◦ rotation (a “π-pulse”) swaps the populations — it is the
jit
physical implementation of the X gate — while a 90◦ (“π/2”) pulse creates an equal superposition
on the equator. Imperfections appear as two distinct kinds of radius loss: energy relaxation (T1 )
pulls the vector toward the |1⟩ pole as the qubit dumps energy into its environment, while dephasing
hi

(T2 ) shrinks the in-plane component as the relative phase randomises. Both drive the state from
the surface into the interior of the ball, which is the geometric meaning of decoherence: the loss
of quantum coherence is literally the loss of Bloch radius.
Ab

4.3 Worked Numerical Problems (Five Types)

Type 1 — State to Bloch angles. Find (θ, ϕ) for |ψ⟩ = √1 (|0⟩ + |1⟩).
2
.

Solution. Matching cos 2θ = √12 gives θ = π

2; the |1⟩ coefficient is real positive so ϕ = 0. The
Dr

point is on the +x axis — this is |+⟩.

Type 2 — Bloch angles to state. Write the state at θ = π2 , ϕ = π2 .

Solution. |ψ⟩ = cos π4 |0⟩ + eiπ/2 sin π4 |1⟩ = √1 (|0⟩

2
+ i |1⟩) = |+i⟩, on the +y axis.

Type 3 — Bloch vector from a density matrix. Find ⃗r for ρ = ( 0.7 0.3
0.3 0.3 ).

Solution. Using ρ = 12 (⊮ + ⃗r · ⃗σ ): rz = ρ00 − ρ11 = 0.4; rx = 2 Re ρ10 = 0.6; ry = 2 Im ρ10 = 0. So

√
⃗r = (0.6, 0, 0.4), |⃗r| = 0.52 ≈ 0.72 < 1 — a mixed state.

11
Quantum Computing · Module 1 Prof. (Dr.) Abhijit Debnath

h
at
bn
Figure 6: The six cardinal states at the intercepts of the coordinate axes. Antipodal points are
orthogonal states.

De
Type 4 — Effect of a gate. Apply σx to |0⟩ and locate the result on the sphere.

Solution. σx |0⟩ = |1⟩. Geometrically, σx is a 180◦ rotation about the x axis, carrying the north
pole |0⟩ to the south pole |1⟩.
jit

Type 5 — Purity / radius. A qubit has ⃗r = (0, 0, 0.5). Compute Tr ρ2 and state whether

it is pure.
hi

Solution. For a single qubit Tr ρ2 = 12 (1 + |⃗r|2 ) = 21 (1 + 0.25) = 0.625 < 1, so the state is mixed;


Ab

it lies halfway between the centre and the north pole.

4.4 Detailed Pros and Cons

Strengths. (i) It turns invisible complex amplitudes into an intuitive, drawable geometry.
(ii) Gates become rotations, so circuit design and intuition become geometric. (iii) Pure vs.
.

mixed states are distinguished by radius, giving an immediate visual measure of decoherence
Dr

and purity. (iv) Expectation values ⟨σi ⟩ are read straight off the coordinates.
Limitations. (i) It represents only one qubit; there is no faithful, low-dimensional Bloch
picture for two or more entangled qubits (an n-qubit pure state needs a much larger space).
(ii) It hides the global phase, which can matter in larger circuits. (iii) Entanglement, the key
resource of quantum computing, is invisible on a single sphere. (iv) For higher-dimensional
systems (qutrits and beyond) the generalisation is far more complicated and far less intuitive.

4.5 Exercises
1. Give (θ, ϕ) for |−⟩ and for |−i⟩.

2. Write the explicit state vector at θ = π3 , ϕ = π.

12
Quantum Computing · Module 1 Prof. (Dr.) Abhijit Debnath

h
at
bn
Figure 7: Left: pure states live on the surface (r = 1); mixed states lie strictly inside (r < 1).
Right: a unitary gate Rx (γ) moves the state along a circular arc — a rotation about the x axis.
De
3. Find the Bloch vector of ρ = 21 ⊮ and interpret it.

4. Where does the Hadamard gate send |0⟩ and |1⟩ on the sphere?

5. A state has ⃗r = (0.6, 0.8, 0). Is it pure? Identify the point.

jit
6. Show that antipodal points on the sphere correspond to orthogonal states.

5 Synthesis and Closing Remarks

hi

These three topics are one idea seen from three angles. Linear algebra supplies the alphabet —
vectors, inner products, Hermitian and unitary operators, eigenvalues. The Hilbert-space for-
malism assembles that alphabet into the grammar of physics: states are unit vectors, observables
Ab

are Hermitian operators, measurement obeys the Born rule, and evolution is unitary. The Bloch
sphere then takes the smallest possible instance of this grammar, the qubit, and renders it as
tangible geometry, where states are points, gates are rotations, and noise is a loss of radius.
A student who can move fluently between the three — computing an inner product, interpreting it
as a probability amplitude, and visualising the corresponding rotation on the sphere — possesses
.

the core conceptual toolkit of modern quantum mechanics and quantum information. The natural
Dr

next steps are the harmonic oscillator and angular momentum (infinite-dimensional and higher-spin
Hilbert spaces), entanglement and the tensor-product structure of many qubits, and the dynamics
of open quantum systems, where the Bloch ball comes fully into its own as a map of decoherence.

One-line summary. Quantum mechanics is linear algebra with a probabilistic interpretation;

a quantum state is a unit vector in a Hilbert space; and for a single qubit that vector is a
point on the Bloch sphere.

13