quantization

References

Summary

Here's the idea:

Heisenberg: We try to use the Heisenberg picture from quantum mechanics: operators should evolve in time, and states remain time-independent. Because quantum field theory is meant to be relativistic, we can't just take time evolution, we need our operators to depend on both time and space: ϕ(x0, x1, x2, x3)

Uncertainty: But as it turns out, we can't really define our quantum fields pointwise over the manifold — because of Heisenberg's uncertainty principle, knowing precise points would correspond to infinite momentum and energy, which takes the form of ultraviolet divergences. So really the proper thing to do is to take operator-valued distributions.

Quantum field: Then a quantum field has the form ϕ(f) = ∫f(x)ϕ(x) — the test function f acts as a kind of mollifier. Then the action is defined to be ϕ(f)|v⟩ = ∫f(x)ϕ(x)|v⟩ for any state in the Hilbert space.

Evaluation: But how do we actually evaluate this integral? The easiest way is to take our state v to be a mutual eigenvector of a basis of operators for ϕ(x). In other words, if we have some operators {ai(x)} such that ϕ(x) = ∑igi(x)ai(x) and ai(x)|v⟩ = λi|v⟩, then we can write

ϕ(f)|v⟩ = (∑i∫f(x)gi(x)λi)|v⟩ and evaluate this interval of scalar functions.

Vacuum: What physicists decide to do is to take v to be the vacuum state, and our basis to be creation/annihilation operators on v

Also we don't just want to write ϕ(x) as a sum; we need it to be an integral

Quantization

The quantum version of the classical field is taken by assumption to just be the classical version's form, but we use creation and annihilation operators