3
$\begingroup$

This book on photoelectrochemistry calculates the emission current from a semiconductor as a function of changing frequency (see Equation 1.13, but reproduced below). The expression for the photoemission current $I$ directed perpendicularly to emitter's surface is given by:

$$ I=e \ \rho_{0} \int j_x \left[e^{\left(E_i-\mu\right) / k T}+1\right]^{-1} d E_{\mathrm{i}} d \mathbf{p}_{\|} $$

where the $E_i$ and $\mathbf{p}_{\|}=\left\{p_y, p_z\right\}$ variables are the energy and parallel (to the surface) components of the momentum of the initial electrons in the metal, respectively. The density of states, $\rho_{0}$ is assumed constant. The electron current, $j_x$ is dependent on energy $E = p^{2}/2m$.

The limits of integration with respect to $E$ and $p_{\|}^2$ are obtained from the law of conservation of energy:

$$ p=\sqrt{2 m (E_{i}+\hbar \omega)-p_{\|}^2}, $$

where $p$ and $m$ denote the $x$-component of the momentum and the mass, respectively, of the emitted electron. Since the partial photocurrent $J$ differs from zero only for real values of $p$, the integration must extend over all values of $E$ and $p_{\|}$which satisfy the condition

$$ \begin{gathered} 2 m(E_{i}+\hbar \omega) \geqslant p_{\|}^2 \text { or } \\ 0 \leqslant p_{\|}^2 \leqslant 2 m(E_{i}+\hbar \omega),-\hbar \omega<E<\infty . \end{gathered} $$

$p$ depends only on $\mathbf{p}_{\|}{ }^2=\left|\mathbf{p}_{\|}\right|^2$. Therefore, the integral can be immediately integrated once to obtain $d \mathbf{p}_{\|}=2 \pi\left|\mathbf{p}_{\|}\right| d\left|\mathbf{p}_{\|}\right|$.

Introducing, in place of $E_i$ and $\left|\mathbf{p}_{\|}\right|$, the new variable $E \equiv p^2 / 2 m$, we can integrate over $E_i$. Thee final result, the absolute value of $I$ is given by:

$$ I=2 \pi e \rho_0 m k T \int_0^{\infty} j_x(\sqrt{2 m E}) \ln \left[1+e^{(\mu+\hbar w-E) / k T}\right] d E $$

I do not understand this step. I do not see or work out the substitution for the exponential.

There is a reference for this step, a Russian textbook - Brodskii and Gurevich, 1973 - Brodskii, A. M., and Gurevich, Yu. Va. (1973). Theory of Electronic Emission from Metals. Nauka, Moscow. But it isn't available anywhere I can find. Apparently page 7 has the solution.


An Attempt:

I think - but I'm not sure since I don't get the correct result, but, taking the first expression,

$$ I=e \rho_0 \int j_x\left[e^{\left(E_i-\mu\right) / k T}+1\right]^{-1} d E_{\mathrm{i}} d \mathbf{p}_{\|} $$

using $k_{y}^2+k_{z}^2=p_{\|}^2 $ and the fact that $ \int_0^{\infty} \left(1+e^x / b\right)^{-1} \mathrm{d} x = \ln (1+b)$.

Hence

$$ \int j_x\left[e^{\left(E_i-\mu\right) / k T}+1\right]^{-1} d \mathbf{p}_{\|}\\ = \int j_x\left[e^{\left(\left[\hbar^2\left(k_x^2+k_y^2+ k_z^2\right) /(2 m)\right]-\mu\right) / k T}+1\right]^{-1} d \mathbf{p}_{\|}\\ = \pi \int_{0}^{\infty} j_x\left[e^{\left(\left[\hbar^2\left(p_{\|}^2+ k_x^2\right) /(2 m)\right]-\mu\right) / k T}+1\right]^{-1} dp_{\|}^{2}\\ = j_x \frac{2 \pi m k_{\mathrm{B}} T}{\hbar^2} \ln \left[1+e^{\frac{\mu-\hbar^2 k_{z}^2 /(2 m)}{k_{\mathrm{B}} T}}\right]\\ $$

and so the integral becomes

$$ I=e \rho_0 \frac{2 \pi m k_{\mathrm{B}} T}{\hbar^2} \int j_x \ln \left[1+e^{\frac{\mu-\hbar^2 k_{x}^2 /(2 m)}{k_{\mathrm{B}} T}}\right] d E_{\mathrm{i}} $$

$\hbar^2 k_{x}^2 /(2 m)$ is now the energy of the remaining component.

$\endgroup$
4
  • 1
    $\begingroup$ This is not a question for a chemistry forum. It should be in physics- $\endgroup$
    – Maurice
    Commented Sep 20 at 19:40
  • 2
    $\begingroup$ The textbook says chemistry on the front? $\endgroup$
    – Tomi
    Commented Sep 20 at 19:46
  • 2
    $\begingroup$ There is no solid boundary between physics and chemistry. $\endgroup$
    – ACR
    Commented Sep 20 at 21:34
  • 2
    $\begingroup$ It's on topic (ChemPhys-PhysChem) but it might be true that there are more people at phys SE ready to answer such a math-ladden q. $\endgroup$
    – Buck Thorn
    Commented Sep 21 at 0:48

1 Answer 1

2
$\begingroup$

I've found a copy of the textbook (in Russian) from Princeton University library, who have very kindly offered to send it to me. I will learn russian, then read the textbook and then report back.

But, until then, I have something that resembles a solution, but has some (obvious) inconsistencies.

Starting with

$$ I=\int j\left[e^{(E-\mu) / k_B T}-1\right]^{-1} \rho_0 d E d \vec{{p}_{\|}} $$

Integrating $d \vec{{p}_{\|}}$ over all angles

$$ \begin{aligned} & \int ^\theta d \vec{{p}_{\|}} \Rightarrow 2 \pi\left|{p}_{\|}\right| d\left|{p}_{\|}\right|=d \vec{{p}_{\|}} \\ \Rightarrow & I=2 \pi \int j\left|\vec{{p}_{\|}}\right|\left[e^{(E-\mu) / k_B T}-1\right]^{-1} \rho_0 d E d\left|{p}_{\|}\right| \\ \end{aligned} $$

Suppose that ${p}_{\|}$ is always positive $$ \left|p_{11}\right|={p}_{\|} \\ \Rightarrow I=2 \pi \int j {p}_{\|}\left[e^{(E-\mu) / k_B T}\right]^{-1} \rho_0 d E d {p}_{\|}$$

Now substitute

$$ \begin{gathered} E=\frac{1}{2 m}\left({p}_{\|}^2+p_z^2\right)-\hbar \nu \\ \Rightarrow I=2 \pi \int j {p}_{\|}\left[e^{\left(\frac{1}{2 m}\left({p}_{\|}^2+p_z^2\right)-\hbar \nu-r\right) / k_B^T}{ }-1\right]^{-1} \rho_0 d E d {p}_{\|} \end{gathered} $$

Now integrate over $d{p}_{\|}$ (from $0$ to $\infty$ ?!?!), gives

$$ \begin{aligned} & =2 \pi \rho_0 \frac{k_B T_m}{\hbar^2} \int j \ln \left[1+e^{\left(\mu+\hbar v-p_{z}^{2} / 2 m\right) / h_B T}\right] d E \end{aligned} $$

Now changing integration variables to $p_{z}$, $$ \begin{aligned} & E=\frac{1}{2 m}\left({p}_{\|}^2+p_z^2\right)-\hbar \nu \\ & \frac{d E}{\partial p_z}=\frac{p_z}{m} \Rightarrow d E=\frac{p_z}{m} d p_z \\ & =2 \pi \rho_0 \frac{k_B T m}{\hbar^2} \int j \ln \left[1+e^{\left.\left(\mu+\hbar r-p_z^2 / 2 m\right) / k_B T\right]} \frac{p_z}{m} d p_z\right. \\ & =2 \pi \rho_0 \frac{k_B T}{\hbar} \int_0^{\infty} j \ln \left[1+e^{\left(\mu+\hbar r-p_z^2 / 2 m\right) / h_B T}\right] p_{z} d p_z \end{aligned} $$

Now, this looks very much like the above expression, apart from if you change to $p_{z}\rightarrow E or E_{z}$, you'll lose the $p_{z}$ and there is no way to get the $\sqrt{E_{z}2m}$ in the integral.

$\endgroup$
2
  • $\begingroup$ No need to learn Russian, try machine translation, esp. if it is only one page. Machine translation of scientific /symbolically heavy text is possible via ChatGPT/DeepL if the equations are fed properly. See the answer to my post SE Science History of Science and Mathematics hsm.stackexchange.com/a/17888/6085 $\endgroup$
    – ACR
    Commented Sep 22 at 12:49
  • $\begingroup$ You can also post a high-res (600 dpi) image and I can try it too. $\endgroup$
    – ACR
    Commented Sep 22 at 12:50

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.