I am confused by the internal energy of a crystal. Lets say I have a isotropically stressed crystal. The total differential of the internal energy per unit volume V is defined as

$$du = \sigma d\epsilon$$

$\epsilon$ being the strain and $\sigma$ the stress. Because the internal energy is a homogeneous function, this implies

$$u=\sigma \epsilon$$

But if I integrate $du = \sigma d\epsilon$ from $0$ to some strain $\epsilon$ using the constitutive relation $\sigma = C \epsilon$, $C$ being some constant, I arrive at

$$ u = \frac{1}{2} \sigma \epsilon $$

Now I am very confused. Which equation is the correct one, $u=\frac{1}{2}\sigma \epsilon$ or $u=\sigma \epsilon$. Or which equation applies to which situation? Probably I didnt get some really important point here :)

Thanks for the help.


Ok, I think I now have understood my problem and solved it. In case anyone else ever wonders about this I will post my answer her. But first I want to reframe the question and clarify some assumptions. I will then post my answer beneath.

So, let there be a system given with some time dependent volume $V$ that we want to compress or expand isothermically and reversibly. That is $dT = 0$ and $dS=0$. Furthermore, for simplicity of the argument, I will assume a homogeneous crystal and homogeneous deformation, so the work of the deformation is $-pdV$. That said, we have for the change of internal energy

$$ dU = -p dV \tag{1} \label{eq1} $$ After division by the reference volume $V_0$ with the definition $J=V/V_0$ and $u=U/V_0$, this reads $$ du=-pdJ' \tag{2} \label{eq2} $$ We will assume a constitutive relation for the pressure as being $$ p(J) = - K ln(J) \tag{3} \label{eq3} $$ Where $K$ is the bulk modulus. Using the small strain approximation, that is $J\approx 1$, this reads $$ p(J) = - K (J-1) \tag{4} \label{eq4} $$

We now integrate eq. \eqref{eq2} from $J'=1$ to $J'=J$ and get $$ \Delta u = K/2(J-1)^2 \tag{5} \label{eq5} $$ On the other hand, $U$ is a homogeneous function of order $1$, that is $U(\lambda V) = \lambda U(V)$ and therefor, according to Eulers Theorem, we got $$ U=-pV \tag{6} \label{eq6} $$ This holds independently of the function $p(J)$ and after devision by $V_0$ and usage of \eqref{eq4} we end up with $$ \Delta u = K(J-1)^2 \tag{7} \label{eq7} $$ Which, obviously, is not the same as \eqref{eq5}. Why is that so? Which formula is correct?

  • $\begingroup$ First of all, if $\sigma$ is a tensile stress (given by $\sigma=E\epsilon$), then the equation for internal energy change should read $du=+\sigma d\epsilon+Tds$ $\endgroup$ – Chet Miller Jul 1 '18 at 11:57
  • $\begingroup$ sorry,i updated the question accordingly $\endgroup$ – Stefanowitschko Jul 1 '18 at 12:02

Now, I think I understood my missunderstandings and want other people to participate from that :)

First of all, the equation $6$ in the question is not the whole truth. Because $U$ is also a homogeneous function of $S$ and $N$, $N$ being the total number of particles in the system, the function $U$ can be written as $$ U = TS-pV+N\mu \tag{8} \label{eq8} $$ $\mu$ is the chemical potential of the system. Keeping in mind, that $dS=0$ because of reversibility and $dT=0$ and $dN = 0$ we calculate for the change of internal energy between the states $0$ and $1$, $1$ being the finale state of volume $V$ and $0$ the initial state of voume $V_0$ $$ \Delta u = -p_1 J + p_0 + N \Delta \mu = -p(J) J + p(1) + N \Delta \mu = -p(J)J + N \Delta \mu \tag{9} \label{eq9} $$ So we have to calculate $\Delta \mu$ to get the full picture. This can be done by considering the Gibbs-Duhem relation. Since $du = TdS-pdJ+\mu d\rho$, $\rho := N/V_0$, the Gibbs-Duhem relation reads $$ sdT-Jdp+\rho d \mu = 0 \tag{10} \label{eq10} $$ Since $dT=0$ and using equation $4$ from the question, this then leads to $$ d \mu = \frac{J}{\rho} dp = \frac{J}{\rho} p_J dJ' = -K \frac{J}{\rho} dJ' \tag{11} \label{eq11} $$ After integration from $J'=1$ to $J'=J$ this gives $$ \Delta \mu = - \frac{K}{2 \rho} (J^2-1) \tag{12} \label{eq12} $$ Inserting \eqref{eq12} and the equation $4$ from the question into \eqref{eq9} we end up with $$ \Delta u = K/2(J-1)^2 \tag{13} \label{eq13} $$ So, in the end the equations are, of course, consistent.

| cite | improve this answer | |

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.