Reviewing my final from last semester to prep for comps:


A piston of mass M can move freely in a tube with cross-section area A filled with ideal monoatomic gas with molecular mass m ≪ M and density n at temperature T.

enter image description here

The first part of the question asks:

Calculate the rate of molecular collisions with the piston (both sides).

I found an equation in one of my fav SM books (Blundell and Blundell) that I think would help here:

enter image description here

However, my professor's solution works entirely in 1-dimension and so using the 1-d maxwell-boltzmann distrubiton. And so I can justify to myself taking the $1/2 cos(\theta) sin(\theta) d\theta $ out of equation 6.12 to match what my professor has in his solution.


$$ N= A \cdot v \cdot dt \cdot n \cdot f(v) \cdot dv $$

where n = number density (N/V), A is the area of the wall/piston, v is velocity, dt is some time interval, and f(v) is the 1-d maxwell boltzmann distribution.

My question is where my professors integration comes from in his provided solution:

$$ \frac{d N}{dt} = 2 \cdot \bigg(\frac{m}{2 \pi T}\bigg)^{1/2}\cdot n \cdot A \cdot \int_0^\infty v e^{\frac{-mv^2}{2 T}} dv = n A \sqrt{\frac{2T}{\pi m}} $$

I have almost all these ingredients from the Blundell and Blundell equation except the integration on the right hand side, and the dN as opposed to just the N on the left hand side.

The only integration of the distribution I am familiar with is for the average velocity,

$$ \bar{v} = \int_0^\infty v f(v) dv $$

What am I reading wrong regarding the missing integration in that Blundell and Blundell equation?

An instance of integrating to get N I do know about is in the case of the degenerate fermi gas at a small but non-zero temperature from equation 7.53 of Schroeder

$$ N= \int_0^\infty g(\epsilon) \bar{n}_{FD}(\epsilon) d\epsilon $$


The 1 dimensional distribution you are referring to is speed (in contrast to velocity).

$$ p(\vec v)d^3v$$ is really

$$p(v_x, v_y, v_z)dv_xdv_ydv_z\rightarrow p(v)v^2dv$$

since the volume element is a shell in velocity space.

For this problem, you're considering the flux across an area:

$$ p(\vec v)\vec v \cdot \hat A = p(v_x)v_xA_x + p(v_y)v_yA_y + p(v_z)v_zA_z $$

where $\hat A$ is the normal to the surface. Since $A_y=A_z=0$, and $A_x=1$, you're just using:

$$p(v_x)v_x $$

  • $\begingroup$ Im totally comfortable with this, my question is regarding the discrepancy in integration versus no integration $\endgroup$
    – Lopey Tall
    Jan 20 at 14:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.