Canonical Quantization

Introduction

Hamiltonian field theory gave us a classical phase-space picture for fields. A field configuration $\phi(\mathbf{x},t)$ and its conjugate momentum $\pi(\mathbf{x},t)$ play the same structural role that coordinates and momenta play in ordinary mechanics, except now there is one canonical pair for every point in space. That classical framework is not yet quantum field theory. It is only the stage on which quantization will happen.

Canonical quantization is the step where the classical field variables stop being ordinary commuting quantities and become operators acting on a Hilbert space of quantum states. This is the exact analogue of what happens in ordinary quantum mechanics when one replaces classical variables $q$ and $p$ by operators $\hat q$ and $\hat p$ satisfying a commutation relation. In field theory, the same idea survives, but because the system has infinitely many degrees of freedom, the canonical relations become local relations between operators at different points in space.

That is the first core idea of this lesson:

$$\phi(\mathbf{x},t), \ \pi(\mathbf{x},t) \quad\longrightarrow\quad \hat\phi(\mathbf{x},t), \ \hat\pi(\mathbf{x},t).$$

The second core idea is that a free field can be decomposed into modes, and each mode behaves like a harmonic oscillator. That means quantizing the field can be reduced, in a very precise sense, to quantizing an infinite collection of oscillators. This is where creation and annihilation operators enter. A single field mode carries quanta of excitation, and those quanta are interpreted as particles.

So canonical quantization is not just a rule for adding hats to symbols. It is the bridge from:

• classical field profiles, to
• operator-valued fields, and then from:
• operator-valued fields, to
• particle states in Fock space.

This lesson develops that bridge carefully. The goal is to understand why the equal-time commutators are chosen as they are, why Fourier modes are the natural basis for a free field, how the harmonic oscillator reappears in every momentum mode, and how the vacuum and excited states arise from ladder operators. By the end, the phrase “a particle is an excitation of a field” should feel like a concrete statement rather than a slogan.

Learning Objectives

• Explain how canonical quantization generalizes from ordinary mechanics to field theory.
• State the equal-time canonical commutation relations for a real scalar field.
• Understand why field operators are not wavefunctions.
• Decompose a free field into Fourier modes and interpret each mode as a harmonic oscillator.
• Define creation and annihilation operators for field modes.
• Explain how the vacuum and excited states arise in field quantization.
• Understand the connection between field quantization and particle interpretation.

Prerequisite Knowledge

• Hamiltonian field theory
• Classical real scalar field and its Hamiltonian formulation
• Quantum harmonic oscillator
• Linear algebra with operators, eigenstates, and commutators
• Basic Fourier analysis
• Familiarity with the difference between states and operators in quantum mechanics

1. From classical canonical variables to quantum operators

Canonical quantization begins by copying the logic of ordinary quantum mechanics as faithfully as possible.

In classical mechanics, a system is described by canonical variables $q_i$ and $p_i$. In quantum mechanics, these become operators $\hat q_i$ and $\hat p_i$, and the Poisson bracket structure is replaced by commutators. In its simplest form, the canonical rule is

$$[q_i,p_j]_{\text{classical}} \ \leadsto \ [\hat q_i,\hat p_j] = i\delta_{ij},$$

with units chosen so that $\hbar = 1$.

Field theory follows the same pattern, except that the label $i$ becomes the continuous spatial label $\mathbf{x}$. So instead of finitely many pairs $(q_i,p_i)$, we now have the canonical field pair

$$\phi(\mathbf{x},t), \qquad \pi(\mathbf{x},t).$$

Canonical quantization promotes these to operators:

$$\phi(\mathbf{x},t) \to \hat\phi(\mathbf{x},t), \qquad \pi(\mathbf{x},t) \to \hat\pi(\mathbf{x},t).$$

This is the point where the theory becomes genuinely quantum. The field is no longer an ordinary numerical function on spacetime. It is now an operator-valued object acting on a Hilbert space of states.

That sentence has to be taken literally. The field itself is not a state. It is not a wavefunction. It is not a probability amplitude. It is an operator. It can act on the vacuum. It can create excitations. It can appear inside commutators. It can have matrix elements between states.

This is one of the most important mental corrections in QFT. In ordinary quantum mechanics, students are used to seeing $\psi(x)$, which depends on position and represents a state in the position basis. In field theory, $\hat\phi(\mathbf{x},t)$ also depends on position, but it is not a state representation. It is a local operator acting on states.

Canonical quantization is therefore not just about quantizing “the values” of a field. It is about reclassifying the field as a quantum operator with a local spacetime label.

2. Equal-time canonical commutation relations

Once the classical fields have been promoted to operators, the next step is to impose the quantum canonical structure.

For a real scalar field, the equal-time canonical commutation relations are

$$[\hat\phi(\mathbf{x}, t), \hat\pi(\mathbf{y}, t)] = i\delta^3(\mathbf{x} - \mathbf{y}),$$

together with

$$[\hat\phi(\mathbf{x}, t), \hat\phi(\mathbf{y}, t)] = 0, \qquad [\hat\pi(\mathbf{x}, t), \hat\pi(\mathbf{y}, t)] = 0.$$

These are the field-theoretic analogues of

$$[\hat q_i,\hat p_j] = i\delta_{ij}.$$

The Kronecker delta $\delta_{ij}$ of finite-dimensional systems has been replaced by the Dirac delta function $\delta^3(\mathbf{x}-\mathbf{y})$, because the label is now continuous.

This structure has a clear meaning. The field operator and the momentum operator fail to commute only when they refer to the same spatial point. That is exactly what one expects when a continuum of canonical pairs is being quantized locally across space.

The phrase equal-time matters a lot. These commutation relations are imposed at the same time $t$. They are not generic relations between arbitrary spacetime-separated points. Once the Hamiltonian is specified, time evolution determines how operators at different times are related, but the canonical structure is imposed at equal time.

This matters because beginners often write the equal-time commutator carelessly as if it held for all pairs of spacetime points. It does not. Equal-time quantization is a specific canonical slice through spacetime.

3. Why equal-time quantization is natural

At first, it may seem strange that time is treated differently from space in the canonical formalism. After all, relativity teaches us to treat spacetime in a unified way. So why impose commutators at equal time?

The answer is that canonical quantization is built from the Hamiltonian formalism, and the Hamiltonian formalism already singles out time as the evolution parameter. At each fixed time, one specifies a field configuration $\phi(\mathbf{x},t)$ and a momentum configuration $\pi(\mathbf{x},t)$, and the Hamiltonian tells you how those data evolve forward.

So equal-time quantization is the quantum version of the classical phase-space setup:

• choose a time slice,
• specify canonical data on that slice,
• impose canonical relations there,
• evolve in time with the Hamiltonian.

This does not mean relativity is lost forever. It means that the canonical formalism is not manifestly symmetric between space and time. Covariance can later reappear in the final physical theory, but the canonical starting point is built around time evolution.

This is one of the recurring themes in QFT: a formalism may hide some symmetry at an intermediate stage while still describing a symmetric theory in the end.

4. From Poisson brackets to commutators

It is useful to see canonical quantization as the field-theoretic continuation of a familiar replacement rule from ordinary quantum mechanics.

In classical Hamiltonian mechanics, canonical variables satisfy Poisson bracket relations such as

$$\{q_i,p_j\} = \delta_{ij}.$$

For fields, the classical Poisson bracket becomes

$$\{\phi(\mathbf{x},t),\pi(\mathbf{y},t)\} = \delta^3(\mathbf{x}-\mathbf{y}).$$

Quantization then replaces the classical bracket by a commutator:

$$\{\cdot,\cdot\} \quad\longrightarrow\quad \frac{1}{i}[\cdot,\cdot].$$

So the quantum relation

$$[\hat\phi(\mathbf{x}, t), \hat\pi(\mathbf{y}, t)] = i\delta^3(\mathbf{x}-\mathbf{y})$$

is not arbitrary. It is chosen so that the quantum theory inherits the canonical structure of the classical one.

This is conceptually important because canonical quantization is not being invented from scratch. It is a systematic translation of the classical phase-space structure into operator language.

5. The free scalar field as the main example

The cleanest example is the free real scalar field, whose classical Lagrangian density is

$$\mathcal{L} = \frac{1}{2}\dot\phi^2 -\frac{1}{2}(\nabla\phi)^2 -\frac{1}{2}m^2\phi^2.$$

From Hamiltonian field theory, the conjugate momentum is

$$\pi = \dot\phi,$$

and the Hamiltonian is

$$H = \int d^3x \left[ \frac{1}{2}\pi^2 + \frac{1}{2}(\nabla\phi)^2 + \frac{1}{2}m^2\phi^2 \right].$$

Canonical quantization now tells us to replace these classical quantities by operators:

$$\phi \to \hat\phi, \qquad \pi \to \hat\pi,$$

with the equal-time commutators imposed.

The free scalar field is the perfect training ground because its equations are linear. Linear equations can be solved by decomposing the field into modes, and that decomposition reveals the harmonic oscillator structure immediately.

This is the key practical reason why free fields are taught first. They are simple enough to solve exactly, but rich enough to show the whole architecture of QFT.

6. Why mode decomposition matters

A field written as $\hat\phi(\mathbf{x},t)$ is a local object depending on position and time. That is natural for discussing locality, equations of motion, and interactions. But for solving the free theory, another basis is more useful: momentum space.

The reason is that the free field equations are translationally invariant. That means plane waves are natural building blocks. A Fourier expansion separates the field into modes labeled by momentum $\mathbf{k}$. For a free field, those modes evolve independently, which is exactly what one wants when quantizing.

So the field is expanded as a superposition of plane-wave modes. Schematically,

$$\hat\phi(\mathbf{x},t) = \int \frac{d^3k}{(2\pi)^3}\, \hat\phi_{\mathbf{k}}(t) e^{i\mathbf{k}\cdot \mathbf{x}}.$$

The crucial point is that for the free theory, each momentum mode behaves like an independent oscillator. This is one of the most important structural facts in all of QFT.

It means that field quantization reduces to:

• decompose the field into modes,
• quantize each mode as a harmonic oscillator,
• assemble the whole field back together.

That is why the quantum harmonic oscillator is not just a side topic in preparation for QFT. It is the local algebraic core of the free field.

7. Each Fourier mode is a harmonic oscillator

Let us make the previous idea more concrete.

For a free real scalar field, the classical equation of motion is the Klein-Gordon equation:

$$\ddot\phi - \nabla^2\phi + m^2\phi = 0.$$

If we Fourier transform in space, the Laplacian becomes multiplication by $-\mathbf{k}^2$. So each momentum mode obeys an ordinary differential equation of the form

$$\ddot\phi_{\mathbf{k}}(t) + \omega_{\mathbf{k}}^2 \phi_{\mathbf{k}}(t) = 0,$$

where

$$\omega_{\mathbf{k}} = \sqrt{\mathbf{k}^2 + m^2}.$$

But this is exactly the equation of motion of a harmonic oscillator with frequency $\omega_{\mathbf{k}}$.

That is the breakthrough. A free field is not one complicated quantum object that needs a brand-new quantization method. It is a continuum of oscillator modes, one for each momentum $\mathbf{k}$, each with its own frequency $\omega_{\mathbf{k}}$.

Once you know that, the rest is inevitable. Every oscillator mode should have:

• a ground state,
• excited states,
• ladder operators,
• and quantized energy levels.

The field is therefore a giant many-oscillator quantum system, and its particle interpretation comes from counting excitations of these modes.

8. Ladder operators for field modes

In the quantum harmonic oscillator, the most convenient operators are not $\hat q$ and $\hat p$ themselves, but the annihilation and creation operators $\hat a$ and $\hat a^\dagger$. They diagonalize the oscillator Hamiltonian structure and make the state space transparent.

The same thing happens for field modes.

For each momentum $\mathbf{k}$, one defines annihilation and creation operators

$$\hat a_{\mathbf{k}}, \qquad \hat a_{\mathbf{k}}^\dagger,$$

which obey commutation relations of the form

$$[\hat a_{\mathbf{k}}, \hat a_{\mathbf{k}'}^\dagger] = (2\pi)^3 \delta^3(\mathbf{k}-\mathbf{k}'),$$

up to convention-dependent normalization factors.

These operators allow the field to be written in its standard mode expansion. For a free real scalar field, the usual form is

$$\hat\phi(\mathbf{x},t) = \int \frac{d^3k}{(2\pi)^3} \frac{1}{\sqrt{2\omega_{\mathbf{k}}}} \left( \hat a_{\mathbf{k}} e^{-i\omega_{\mathbf{k}} t + i\mathbf{k}\cdot\mathbf{x}} + \hat a_{\mathbf{k}}^\dagger e^{i\omega_{\mathbf{k}} t - i\mathbf{k}\cdot\mathbf{x}} \right),$$

again depending on convention.

This formula is central. It tells you that the field operator contains both annihilation and creation pieces. Acting on states, those pieces lower or raise occupation numbers in momentum modes.

That is where the particle picture enters. A one-quantum excitation created by $\hat a_{\mathbf{k}}^\dagger$ is interpreted as a one-particle state of momentum $\mathbf{k}$.

9. Vacuum and Fock space

Once ladder operators exist, the state space can be organized exactly as in the harmonic oscillator, but now for infinitely many modes.

The vacuum state $|0\rangle$ is defined by

$$\hat a_{\mathbf{k}} |0\rangle = 0 \qquad \text{for all } \mathbf{k}.$$

This is the state with no quanta in any mode. It is the ground state of the free field.

From the vacuum, one builds excited states by acting with creation operators:

$$\hat a_{\mathbf{k}}^\dagger |0\rangle$$

is a one-particle state with momentum $\mathbf{k}$.

Similarly,

$$\hat a_{\mathbf{k}_1}^\dagger \hat a_{\mathbf{k}_2}^\dagger |0\rangle$$

is a two-particle state, and so on.

The full Hilbert space built in this way is called Fock space. It is the natural state space for a quantum field because particle number is not fixed in advance. Different sectors of Fock space correspond to zero particles, one particle, two particles, and so on.

This is one of the deep reasons QFT can describe processes that ordinary fixed-particle-number quantum mechanics cannot. The formalism does not hard-code a fixed number of particles. It starts from the field, and particle number emerges as an occupation-number property of field modes.

10. Why the field operator is not a wavefunction

At this stage, one common confusion becomes especially dangerous, so it is worth addressing directly.

The field operator $\hat\phi(\mathbf{x},t)$ depends on spacetime coordinates, and a one-particle wavefunction $\psi(\mathbf{x},t)$ also depends on spacetime coordinates. That superficial similarity causes trouble.

But they are not the same type of object.

A wavefunction is a representation of a state in a chosen basis. It tells you how a state vector looks in the position basis.

A field operator is an operator acting on states. It changes the state it acts on. It can create or annihilate quanta. It appears inside operator algebra.

So although both carry labels like $\mathbf{x}$ and $t$, they play completely different roles.

This distinction is one of the foundations of understanding QFT properly. If you mistake the field operator for a many-component wavefunction, the logic of quantization, vacuum structure, and particle creation becomes badly distorted.

11. The Hamiltonian after quantization

For a free field, once the mode expansion is inserted into the Hamiltonian, the Hamiltonian becomes a sum or integral over independent oscillator Hamiltonians:

$$\hat H = \int \frac{d^3k}{(2\pi)^3} \, \omega_{\mathbf{k}} \left( \hat a_{\mathbf{k}}^\dagger \hat a_{\mathbf{k}} + \frac{1}{2}\,[\text{zero-point term}] \right),$$

again up to normalization conventions.

This is the field-theoretic version of the harmonic oscillator energy spectrum.

Each momentum mode contributes:

• an occupation-number term,
• plus a zero-point energy contribution.

The occupation-number term is what gives the particle interpretation. The vacuum energy term is what later leads to important physical effects such as vacuum fluctuations and, in more advanced settings, the Casimir effect.

So the mode expansion does more than solve the free theory. It exposes the energy structure of the quantized field directly.

12. Why free fields are the starting point

At this point, it might seem that canonical quantization works smoothly only because the theory is free. That is basically true, and it is not a weakness. It is the right place to start.

In a free field theory:

• the equations are linear,
• momentum modes decouple,
• each mode is an oscillator,
• the Hilbert space can be built exactly.

In interacting theories, things become more complicated:

• modes no longer evolve independently,
• the oscillator decomposition is no longer the whole story,
• particle interpretation can become subtler.

But the free theory is still foundational because it provides the language in which more complicated theories are described. Creation operators, annihilation operators, Fock space, and equal-time commutators remain basic ingredients even when interactions are added.

So canonical quantization of the free field is not a toy model to forget later. It is the conceptual skeleton of the full subject.

Worked Examples

Example 1: Equal-time commutator

Canonical quantization imposes

$$[\hat{\phi}(\mathbf{x}, t), \hat{\pi}(\mathbf{y}, t)] = i\delta^3(\mathbf{x} - \mathbf{y}),$$

with

$$[\hat{\phi}(\mathbf{x}, t), \hat{\phi}(\mathbf{y}, t)] = 0, \qquad [\hat{\pi}(\mathbf{x}, t), \hat{\pi}(\mathbf{y}, t)] = 0.$$

This is the local continuum version of the canonical quantum-mechanical relation

$$[\hat q_i,\hat p_j] = i\delta_{ij}.$$

The Dirac delta function appears because the canonical degree of freedom is indexed by continuous position rather than a discrete label.

Example 2: Oscillator mode

For a free scalar field, each momentum mode satisfies

$$\ddot\phi_{\mathbf{k}} + \omega_{\mathbf{k}}^2 \phi_{\mathbf{k}} = 0, \qquad \omega_{\mathbf{k}} = \sqrt{\mathbf{k}^2 + m^2}.$$

This is the equation of motion of a harmonic oscillator. So the mode labeled by $\mathbf{k}$ can be quantized exactly like an oscillator with frequency $\omega_{\mathbf{k}}$. The full quantum field is therefore a continuum of quantized oscillators.

Intuition

Canonical quantization tells you to stop picturing a field as a classical waveform that just happens to exist in a quantum world. Instead, the field itself becomes quantum. Its value at each point is no longer an ordinary number but part of an operator algebra. The field fluctuates, the vacuum has structure, and excitations are quantized.

A good physical picture is this: the free field is like a huge instrument made of infinitely many normal modes. Quantization says each normal mode can only be excited in discrete steps, exactly like a harmonic oscillator. A particle is one quantum of one of those modes. Many-particle states are just multi-mode excitation patterns of the same underlying field.

That is why QFT does not begin with particles and then try to explain interactions. It begins with the field, quantizes it, and particles appear as the natural language for its excitations.

Common Mistakes

• Writing canonical commutators without specifying that they are equal-time relations.
• Treating $\hat\phi(\mathbf{x},t)$ as if it were a wavefunction rather than an operator.
• Forgetting that the Dirac delta function replaces the Kronecker delta in the continuum theory.
• Missing the point that each free momentum mode behaves like an independent harmonic oscillator.
• Thinking the creation and annihilation operators belong to one oscillator only, rather than one for every momentum mode.
• Forgetting that the vacuum is defined by annihilation operators killing the state for all momenta.
• Confusing the field itself with a particle, instead of seeing particles as excitations of the field.
• Assuming the harmonic-oscillator picture automatically solves interacting field theories in the same simple way.

Short Summary

Canonical quantization promotes the classical field $\phi(\mathbf{x},t)$ and its conjugate momentum $\pi(\mathbf{x},t)$ to operators acting on a Hilbert space. Their quantum structure is imposed through the equal-time commutation relations

$$[\hat\phi(\mathbf{x}, t), \hat\pi(\mathbf{y}, t)] = i\delta^3(\mathbf{x}-\mathbf{y}),$$

with field-field and momentum-momentum equal-time commutators vanishing. For a free scalar field, a Fourier decomposition shows that each momentum mode satisfies the equation of a harmonic oscillator with frequency

$$\omega_{\mathbf{k}} = \sqrt{\mathbf{k}^2 + m^2}.$$

This allows the field to be quantized mode by mode using creation and annihilation operators. The vacuum is the state annihilated by all annihilation operators, and excited states are built by creation operators acting on the vacuum. In this way, canonical quantization turns a classical field into an operator-valued quantum field and reveals particles as quantized excitations of its modes.

Practice Problems

1. State the equal-time canonical commutation relations for a real scalar field and explain why the Dirac delta function appears.

2. Why is canonical quantization formulated at equal time rather than as a commutator relation between arbitrary spacetime points?

3. Explain carefully why $\hat\phi(\mathbf{x},t)$ is an operator and not a wavefunction.

4. Starting from the Klein-Gordon equation, show that a spatial Fourier mode satisfies the harmonic-oscillator equation

$$\ddot\phi_{\mathbf{k}} + \omega_{\mathbf{k}}^2 \phi_{\mathbf{k}} = 0.$$

5. Derive the expression

$$\omega_{\mathbf{k}} = \sqrt{\mathbf{k}^2 + m^2}$$

for the frequency of a free scalar field mode.

6. Explain why a free field can be viewed as a continuum of independent oscillators.

7. Define the vacuum state in terms of annihilation operators and explain why this gives the no-particle state.

8. What is Fock space, and why is it a natural Hilbert space for a quantum field?

9. A student says, “The field operator is the wavefunction of the particle.” Explain why this is wrong.

10. Why is canonical quantization of the free field considered foundational even though realistic theories include interactions?