I have a function, an equation of state named the Birch-Murnaghan's equation of state, in which the Energy ( $E$ ) is a function of the Volume ($V$), where $E_{0}$, $V_{0}$, $B_{0}$ and $B_{0}^{'}$ are parameters that can be obtained from the fit:

\begin{equation} \label{BM-EOS-Energy-Volume} E\left ( V \right )=E_{0}+\frac{9V_{0}B_{0}}{16}\left \{ \left [ \left ( \frac{V_{0}}{V} \right )^\frac{2}{3} -1 \right ]^3 B_{0}^{'}+\left [ \left ( \frac{V_{0}}{V} \right )^{\frac{2}{3}}-1 \right ]^{2}\left [ 6-4\left (\frac{V_0}{V} \right )^{\frac{2}{3}} \right ]\right \}, \tag{1} \end{equation}

The pressure P may be written as a function of the volume V as:

\begin{equation} \label{BM-EOS-Energy-Volume-derivative} P\left ( V \right )=-\left ( \frac{\partial E}{\partial V} \right )_{S}, \tag{2} \end{equation}

so that we can get the expression of $P\left ( V \right )$ : \begin{equation} \label{BM-EOS-Pressure-Volume} P\left ( V \right )=\frac{3B_{0}}{2}\left [ \left ( \frac{V_{0}}{V} \right )^\frac{7}{3} - \left ( \frac{V_{0}}{V} \right )^\frac{5}{3}\right ]\left \{ 1+\frac{3}{4}\left ( B_{0}^{'}-4 \right )\left [ \left ( \frac{V_{0}}{V} \right )^\frac{2}{3}-1 \right ] \right \} \tag{3} \end{equation}

Now, I would like to get an expression of $E(P)$, but I do not find the way to accomplish that goal starting from the expression of $E(V)$ (Equation 1) and $P(V)$ (Equation 3).

I would appreciate if you could help me.


I have data of $E$ vs $V$. After plotting, I have fitted this data using Eq. (1).

I have transformed each $V$ data to $P$ using Eq. (3).

I have now plotted $P$ vs $V$, and fitted using Eq. (3).

Now I have plotted $E$ vs $P$ and I need to fit this data using a function. And this function would have to be the expression of $E(P)$, which I don't know a possible way of working it out.

  • $\begingroup$ You could try to manually eliminate $V$. $\endgroup$
    – Karlo
    Jun 18, 2016 at 12:54
  • $\begingroup$ Whoops ... comment deleted. $\endgroup$
    – garyp
    Jun 18, 2016 at 13:08
  • $\begingroup$ @Karlo In the hypothetical case you were to eliminate $V$ from Eq. 1, you would not end up with an expression of $E(P)$ $\endgroup$
    – DavidC.
    Jun 18, 2016 at 13:52
  • 1
    $\begingroup$ @DavidC - you misunderstood Karlo's obvious recipe. You clearly can't eliminate $V$ just from equation 1. You need to elimininate $V$ from the set of equations 1,3. These are two equations involving $E,P,V$, if you eliminate $V$, i.e. find a combination of 1,3 that has no $V$, you will have an equation with $E,P$ only which is an implicit (or explicit, if you are lucky) prescription for $E=E(P)$. $\endgroup$ Jun 18, 2016 at 15:09
  • $\begingroup$ @LubošMotl There is no possible way to describe a combination of Equations 1,3 that have no $V$ $\endgroup$
    – DavidC.
    Jun 18, 2016 at 15:50

1 Answer 1


This is the purpose of the Legendre transform, which lies at the core of hamiltonian mechanics. You have a function $E(V,S)$, which is convex for certain interval of $V$, and want to find $E'(P,S)$, where $P \equiv -\left.\frac{\partial E}{\partial V}\right|_S$. This function is given by the Legendre transform of $E$: $$E'(P,S) = PV + E(V,S)$$ for the interval of $V$ in which $E$ is convex.

  • $\begingroup$ Does that mean that $E(P)$ is simply $E(P) = PV + E_{0}+\frac{9V_{0}B_{0}}{16}\left \{ \left [ \left ( \frac{V_{0}}{V} \right )^\frac{2}{3} -1 \right ]^3 B_{0}^{'}+\left [ \left ( \frac{V_{0}}{V} \right )^{\frac{2}{3}}-1 \right ]^{2}\left [ 6-4\left (\frac{V_0}{V} \right )^{\frac{2}{3}} \right ]\right \}$ ? $\endgroup$
    – DavidC.
    Jun 19, 2016 at 14:47
  • $\begingroup$ @DavidC. Yes, but there is the problem that $E(V,S)$ is convex in a small interval and may not be what you need. $\endgroup$
    – auxsvr
    Jun 19, 2016 at 15:03
  • $\begingroup$ @DavidC. Also, I know that many physicists (me included) abuse notation and write $E(P)$, when in fact they mean "a function that takes values of pressure and returns energy", but in mathematical terms $E(P)$ means $E(V)$ if you replace $V\mapsto P$, which doesn't make sense and may confuse you. $\endgroup$
    – auxsvr
    Jun 19, 2016 at 15:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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