# Representation of conjugate momentum

1. Oct 9, 2015

### AlonsoMcLaren

We know that in Cartesian position basis the representation of momentum is -ihbar (d/dx)

Consider a cylindrical/spherical/whatever curvilinear coordinates. To make life simple, consider a particle constrained to move on a circle so that its position can described by θ only. Suppose we express the wavefunction as a function of θ, not x. The system has an Lagrangian from which we can find the conjugate momentum pθ

Can we thus declare that pθ can be represented by -ihbar (d/dθ) in the θ basis?

2. Oct 9, 2015

### Matterwave

Canonical quantization sort of only works well in rectilinear coordinates iirc. Quantizing generalized coordinates is a messy subject. However, in the case of cylindrical coordinates in one dimension, All you have to do is turn x into a periodic coordinate and it should work fine.

3. Oct 10, 2015

### vanhees71

The momentum operator in position representation is always
$$\hat{\vec{p}}=-\mathrm{i} \hbar \vec{\nabla}.$$
Since $\vec{\nabla}$ is a vector operator, it can be expressed in any coordinates you like. It always has the same meaning, but perhaps I don't understand the question right.

4. Oct 11, 2015

### moss

Quantization will work in any coordinate system as long as there exists coordinate transformation between different system, that is the whole pt. and the physics does not depend on a specific coordinate system. Generalized coordinates are used all the time in QFT, QM and Stat.Physics, you name it.

On a circle, theta can be replaced by x, just consider the circle is very large and is almost a straight line btw 2 pts. A & B. Why not? There is only one basis vector in theta-basis, since 1-dim. If the radius of the circle is changing then use r and theta as basis.

You can also transform the coordinates from x-quantized system to theta system with simple coordinate transformation and thats it, you have a theta-quantized system.

Bottom line is that you quantize in one system of coordinates and just transform into any coordinates, the system remain quantized.

Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook