This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: M (15636” 2000 300 50 #istep, istable, iprint 1 0 1 #ini_new, iscale_x, iscale_v {a was
0.0 1.557 #req_temp,xlcd
8 8 8 # ncell(3) v” m meme y 37 El 0 100 0.5 #isquench,iquench,coeff_quench istep is the # of time steps, istable specifies after which step we
start calculating properties
iprint means the program will print properties per so many steps. iniﬂnew (1 if generating new system, 0 if reading in from file)
iscale_x (1 if scaling x, 0 no scaling), iscale,v(1 if scaling v, 0 no scaling) reqmtemp is the preset temperature and xlcd is the lattice constant.
They determine the initial x and v scale if iscaleﬂx = 1 or iscale_v = 1. ncell(3) are the numbers of unit cells in x, y, z direction.
Their product shouldn't exceed 2500 and it has no effect when ini__new=0. isquench (1 it quenching, 0 no quenching), iquench is the # of steps
between quenching,
and coeff__quench is the quenching coefficient. 277* 3'55? —— 0200013®3£>ct 050 ® Assignment 1:
With xlcd = 1.557, compute the following: E .vs. T P versus T Cv versus T » BO (Mean square displacement for solid) or D(Diffusion constant for liquid) versus
T Detailed procedure (with the program already compiled and a.out in the same
folder with md.in): a) Create a new folder called x0. Copy a.out and md.in here. Generate a new
system using the following parameters in md.in: x
94:? 200W V
N Q_
JT/éjoia 1.557 \) 53‘“
S\ X / 7 7 7
O 100 O.54\)aw..L .
L m
b) Create within folder x0 a new directory called 0.3 (or any number that
an, represents the new temperature you will scale the system to), and copy a.out, md.in and xvconf from last run into that directory. Note the new directory is
necessary so that some important output files will not get overwritten. Now scale the system's temperature using the following parameters (here those
numbers that need to be changed from last step are underlined): \ When this run is finished, thermalize the system using the following parameters: 10000 5000 50 O O Q
0.3 1.557
7 7 7 O 100 0.5 Take down the final temperature and pressure and energy after the run is over.
Now without changing anything, run the same thing again. After this run is over, compare the final temperature, pressure and energy with
last run. if the difference is rather large (> 0.1%), run again with the same
parameters until the results converge. Otherwise, these are your final results for
a thermalized state. You have one data point of (T1, E1, P1, Cv1). c) Now create within x0 a new directory called 0.6 (or any number that represents
the new temperature you will scale the system to), and copy a.out, md.in and
xvconf from last run’s directory (0.3) into the new directory. Now scale the system to the new temperature using the following parameters: 10000 5000 50 0 0 1 Q._6_ 1.557 7 7 7 0 100 0.5 When this run is finished, thermalize the system using the following parameters: 10000 5000 50 0 0 0 0.6 1.557 7 7 7 0 100 0.5 Take down the final temperature and pressure and energy after the run is over.
Now without changing anything, run the same thing again. After this run is over, compare the final temperature, pressure and energy with
last run. if the difference is rather large (> 0.1%), run again with the same
parameters until the results converge. Otherwise, these are your final results for
a thermalized state. You have another data point of (T2, E2, P2, CV2). d) Repeat the same process while increase the corresponding T by 0.3 or 0.4
each time. i.e., change the 0.6 in last step to 0.9, 1.2, etc... e) After you repeat the procedure with T finally reaches 3.0, you have got about
10 data points. Now you can plot the graphs of E, P, Cv .vs. T using these data.
About B .vs. T, go to each directory we saved along the way, and copy the data
saved in file ‘meansqdisp’. Inside, the first column is time, and the second column
is mean square displacement. You can plot the mean square displacements for F_____ VVVVVV a , different temperatures as functions of time in one graph (see the gnuplot tutorial
Kenichi sent you or use Excel and merge different data series into a single chart).
Also you can determine the diffusion coefficients from the slope of each curve’s
linear region, as a function of temperature. Assignment 2:
To determine the thermal expansion coefficient of the system:
We already know that P=0 for xlcd=1.557 at T=0. f) Now increase the lattice constant by 0.03. Create a new directory called x1 on
the same level with x0. Repeat what we did in assignment 1 from step a, except
that all the 1.557s in the md.in file are replaced by 1.587, and you will only need
to try 4 or 5 different Ts. What we want is P as a function of T. T doesn’t have to
go up to 3.0 but it should be high enough to give positive pressure values (the
closer P is to 0 the better). After we have some data points, we can extrapolate
the point where the curve intersects P=0 with either linear extrapolation (using
two closest points to the P=0 axis) or quadratic extrapolation (using 3 or more
points). Take down the xlcd and T values that correspond to that P=0 point. g) increase the lattice constant by 0.03 again. Repeat last step and get another
P=0 point. Take down the corresponding xlcd and T values again. h) Repeat the last step by increasing xlcd by 0.03 each time. After you have
enough data point, draw a plot of xlcd?’ .vs. T, and you would be able to
determine the expansion coefficient from the graph. ...
View
Full Document
 Spring '07
 Vashishta

Click to edit the document details