# Colloidal networks Authors: Yoldas, Frank, Laura ## Plans ### NVT Monte Carlo Simulation We will explore the behavior of colloidal networks of patchy particles using simulations and Wertheim theory. ## To Do :::success green is for finished tasks ::: ### Simulations :::success * Get the simulation running * Implement visualization in Python * Plot behavior of the energy and pressure as a function of time * Plot behavior of the energy and pressure as a function of T * Plot behavior of the energy and pressure as a function of denstiy ::: ### 21/05: Notes * vizualisation code is done. Some notes on the code * python keeps the location in mind when the patch file is read. * new command learned: **np.dot** is for the matrix multiplication of numpy arrays. * An example image that seems to match the linux code: ![](https://i.imgur.com/GlUjird.png) ### 24/05: Notes * plotted potential and pressure with regards to time. * it is clear high epsilon needs more time to reach equilibria --> skip first 200 rows * run the simulation for different epsilon and plot * as $\epsilon$ increases the pressure has higher fluctuations, even reaching negative values. * ![](https://i.imgur.com/asmFjm3.png) * as $\epsilon$ increases pot energy goes down but fluctuation isn't as high. * ![](https://i.imgur.com/eJ1YREB.png) * numerically for epsilon 10 * potential= -13.535601753333333 std = 0.09570560566061856 pressure= -0.03288044333333333 std= 0.07562040064039237 * the std for pressure is so high * for epsilon 0.1, 5.5 and 10 ![](https://i.imgur.com/YVspygl.png) ![](https://i.imgur.com/mUvWQNg.png) ![](https://i.imgur.com/zGAVjfB.png) ### 02/06: Notes **To Do** - [x] plot $\Delta F$ with different values of epsilon and density * Derivation of Gamma $d\Gamma_{12} = (1/2\pi)^2 d\vec{r}_{12} d\theta_1 d\theta_2$ $d\Gamma_{12} = (1/2\pi)^2 \int_{\sigma}^{r}\int_{0}^{2\pi}rsin\theta dr d\theta \int_{-\theta_{max}}^{\theta_{max}} d\theta_1\int_{-\theta_{max}}^{\theta_{max}} d\theta_2$ $d\Gamma_{12} = \frac{(r_{max}^2-\sigma^2)\theta_{max}^2}{\pi}$ ![](https://i.imgur.com/XRzNa73.png) ![](https://i.imgur.com/aLYa2ot.png) * Seems like there is an issue? Because delta F is positive? I expected the difference to get larger as we have higher $\epsilon$ and $n$ but it is positive because the term inside log is too small. - [x] from $\delta F$ calculate the bonding probability, the number of bonds in the system and so the potential energy. - [x] compare with data from the simulation - ![](https://i.imgur.com/7qF0sg5.png) - theory underestimates the number of bonds, probably becuase we took g{r} to be 1. In reality the particles are more densely packed. - the full code is in name theory_changinT ### 03/06 Notes - [x] incorporate g_r from Yuste to the above case * It is done. Gives more matching results than the first order approximation. ![](https://i.imgur.com/k3aoQvs.png) * changed the code in a way to find $p_b$ analytically. `fsolve` function choose the wrong solution occasionally. Since it works depending on the initial guess. ### 04/06 Notes - [x] calculate pressure using Bianchi(2008) paper as a function of density and compare with simulation - [ ] Find expression for free energy as a funciton of density to investigate phase behaviour The expression for the free energy as a function of density. $$F(\rho)=F^\mathrm{id}(\rho) + N \int_0^\rho \mathrm{d}\rho' \frac{P(\rho') - \rho' k_B T}{\rho'^2}$$ $$F^\mathrm{id}(\rho) = k_B T N\left[\log(\rho \Lambda^2) - 1\right]$$ Here, $\rho$ is the number density, $P(\rho)$ is the pressure at density $\rho$, and $\Lambda$ is the thermal wavelength. In practice, $\Lambda$ does not affect the phase behavior, so you can simply set it equal to $\sigma$. $P(\rho) = -\frac{d(F)}{dV} = \frac{d(F/N)}{d\rho} \rho^2 = \frac{1}{N}\frac{dF}{d\rho} \rho^2$ (Check my math) $\frac{d(F/V)}{d\rho} = \frac{d(F)}{d\rho}/V + F(\rho)\frac{d(1/V)}{d\rho}$