Dear Chad, as the energy barrier is very very very small maybe the use of freezing causes the algorithm to have more instabilities.<br><br>bests<br><br>Layla<br> <br><div class="gmail_quote">2011/12/29 Chad Junkermeier <span dir="ltr"><<a href="mailto:junkermeier@yahoo.com">junkermeier@yahoo.com</a>></span><br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div><div style="color:#000;background-color:#fff;font-family:times new roman,new york,times,serif;font-size:12pt"><div>
<div style="border-collapse:separate;color:rgb(0,0,0);font-family:Helvetica;font-style:normal;font-variant:normal;font-weight:normal;letter-spacing:normal;line-height:normal;text-transform:none;white-space:normal;word-spacing:0px;font-size:medium;word-wrap:break-word">
Hello,<br> I have a problem that I keep encountering when performing NEB calculations. I will briefly describe what is happening here, and then give a more detailed explanation below. In short, the NEB calculations appear to run well for some number of iterations (usually on the order of 40 or 50 iterations) and
then all of the sudden weird image energies start showing up, and the activation energy starts jumping around with each iteration.<br><br></div><div style="border-collapse:separate;color:rgb(0,0,0);font-family:Helvetica;font-style:normal;font-variant:normal;font-weight:normal;letter-spacing:normal;line-height:normal;text-transform:none;white-space:normal;word-spacing:0px;font-size:medium;word-wrap:break-word">
<br></div><div style="border-collapse:separate;color:rgb(0,0,0);font-family:Helvetica;font-style:normal;font-variant:normal;font-weight:normal;letter-spacing:normal;line-height:normal;text-transform:none;white-space:normal;word-spacing:0px;font-size:medium;word-wrap:break-word">
<br></div><div style="border-collapse:separate;color:rgb(0,0,0);font-family:Helvetica;font-style:normal;font-variant:normal;font-weight:normal;letter-spacing:normal;line-height:normal;text-transform:none;white-space:normal;word-spacing:0px;font-size:medium;word-wrap:break-word">
More detailed explanation:</div><div style="border-collapse:separate;color:rgb(0,0,0);font-family:Helvetica;font-style:normal;font-variant:normal;font-weight:normal;letter-spacing:normal;line-height:normal;text-transform:none;white-space:normal;word-spacing:0px;font-size:medium;word-wrap:break-word">
<br>Physically what I am trying to do is model the diffusion of a small molecule across a graphene surface. This molecule is only weakly absorbed (physisorbed) onto the surface, thus, the activation energy is going to be rather small. Our relaxation calculations show that van der Waals type interactions are important in this system, as should be expected for something that is weakly absorbed. In the NEB calculation the molecule is going from one lowest energy site to a neighboring lowest energy site that is only a few angstroms away, all of which is well within the bounds of the supercell.</div>
<div style="border-collapse:separate;color:rgb(0,0,0);font-family:Helvetica;font-style:normal;font-variant:normal;font-weight:normal;letter-spacing:normal;line-height:normal;text-transform:none;white-space:normal;word-spacing:0px;font-size:medium;word-wrap:break-word">
<br></div><div style="border-collapse:separate;color:rgb(0,0,0);font-family:Helvetica;font-style:normal;font-variant:normal;font-weight:normal;letter-spacing:normal;line-height:normal;text-transform:none;white-space:normal;word-spacing:0px;font-size:medium;word-wrap:break-word">
I am using Quantum Espresso 4.3.2 on a Cray XE6. </div><div style="border-collapse:separate;color:rgb(0,0,0);font-family:Helvetica;font-style:normal;font-variant:normal;font-weight:normal;letter-spacing:normal;line-height:normal;text-transform:none;white-space:normal;word-spacing:0px;font-size:medium;word-wrap:break-word">
<br></div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium"><div style="word-wrap:break-word"><br></div>My input file looks like this:</div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">
<br></div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">BEGIN<br>BEGIN_PATH_INPUT<br>&PATH<br> restart_mode = 'from_scratch'<br> string_method = 'neb',<br> ds = 2.D0,<br>
opt_scheme = "broyden",<br> num_of_images = 45,<br> k_max = 0.3D0,<br> k_min = 0.2D0,<br> CI_scheme = "auto",<br> nstep_path=3000,<br> use_freezing = .TRUE.<br>
first_last_opt =
.FALSE.<br>/<br>END_PATH_INPUT<br>BEGIN_ENGINE_INPUT<br>&CONTROL<br>prefix="TW2_u_B3_B4",<br>pseudo_dir = "/espresso/pseudo",<br>outdir = "/workspace/TW2_u_B3_B4/tmp",<br>/<br>&SYSTEM<br>
ibrav = 4,<br> celldm(1) = 37.2672889,<br> celldm(3) = 0.760623,<br> nat = 144,<br> ntyp = 4,<br>ecutwfc = 50.0,<br>ecutrho = 400,<br>occupations = "smearing",<br>smearing = "methfessel-paxton", <br>
degauss = 0.01, <br>london = .TRUE.<br>/<br>&ELECTRONS <br>conv_thr = 0.0000001, <br>mixing_beta = 0.3, <br>electron_maxstep =
150,<br>/<br>ATOMIC_SPECIES<br>C 12.0 C.pbe-van_ak.UPF<br>H 1.0 H.pbe-van_ak.UPF <br>O 18.0 O.pbe-van_ak.UPF<br>P 31.0 P.pbe-van_ak.UPF <br>BEGIN_POSITIONS<br>FIRST_IMAGE<br>ATOMIC_POSITIONS angstrom <br>C 0.000000000 0.000000000 0.000000000 0 0 0<br>
C 1.234000000 0.710000000 0.000000000 0 0 0<br>C <a href="tel:2.465000000" value="+12465000000" target="_blank">2.465000000</a> 0.000000000 0.000000000 0 0 0<br>C 3.699000000 0.710000000 0.000000000
0 0 0<br>C 4.930000000 0.000000000 0.000000000 0 0 0</div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">.</div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">
.</div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">.</div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">LAST_IMAGE <br>ATOMIC_POSITIONS angstrom <br>.</div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">
.</div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">.<br>H 12.789295543 12.155855132 2.858948537<br>H 13.430079977 11.427859045 4.385644094<br>H 14.299090315 11.190357788 2.820979411<br>
END_POSITIONS<br>K_POINTS {gamma} <br>END_ENGINE_INPUT<br>END<br><br></div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">While I didn't include all of the atoms in here (since there are a lot, and I don't think this is necessarily to the resolution problem), I will note that the majority of the atoms will
not move appreciably during the course of the calculation, and thus to speed up the calculation I tell the code to fix the positions of the atoms that aren't moving.</div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">
<br></div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">I started off with values that I thought would help the code run quicker: </div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">
conv_thr = 0.00001</div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">ecutwfc = 30.0,<br>ecutrho =
280,</div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">num_of_images = 15,<br>electron_maxstep = 100,</div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium"><br></div>
<div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">After finding the problem that I am writing here about, and reading everything I could find on the QE website, and within pw_forum, I slowly made changes trying to work through the problem. Having the number of images set to 45 in the above input file has many more images than I think I need, but one of the things that I kept reading is that having more images will decrease the likelihood of problems arising.</div>
<div style="word-wrap:break-word;font-family:Helvetica;font-size:medium"><br></div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium"><br></div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">
After the first few NEB iterations a nice potential barrier curve appears and is slowly refined. An example of the standard output for one of the well behaved images is given below:</div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">
<br></div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">
------------------------------ iteration 44 ------------------------------<br><br> tcpu = 61713.9 self-consistency for image 10<br> tcpu = 61932.1 self-consistency for image 21<br> tcpu = 62113.6 self-consistency for image 28<br>
tcpu = 62294.4 self-consistency for image 31<br><br> activation energy (->) = 0.043939 eV<br> activation energy (<-) = 0.044011 eV<br><br> image energy (eV) error (eV/A) frozen<br>
<br> 1 -21959.2382129
0.081337 T<br> 2 -21959.2388966 0.020923 T<br> 3 -21959.2381194 0.123092 T<br> 4 -21959.2363069 0.164123 F<br>
5 -21959.2352170 0.044025 T<br> 6 -21959.2331629 0.220840
F<br> 7 -21959.2301578 0.115291 T<br> 8 -21959.2277497 0.199382 F<br> 9 -21959.2256241 0.037564 T<br> 10 -21959.2216078 0.257680 F<br>
11 -21959.2173790 0.220445 F<br>
12 -21959.2159181 0.195525 F<br> 13 -21959.2122799 0.189871 F<br> 14 -21959.2093498 0.070035 T<br> 15 -21959.2075466 0.156424 F<br>
16 -21959.2037818 0.237338 F<br> 17 -21959.2013877
0.059399 T<br> 18 -21959.2006768 0.168875 F<br> 19 -21959.1986806 0.049577 T<br> 20 -21959.1962718 0.054413 T<br>
21 -21959.1966851 0.022074 T<br> 22 -21959.1954724 0.053047
T<br> 23 -21959.1942736 0.054956 T<br> 24 -21959.1944583 0.140762 F<br> 25 -21959.1964535 0.032502 T<br> 26 -21959.1950797 0.210304 F<br>
27 -21959.1987196 0.051561 T<br>
28 -21959.2008371 0.031058 T<br> 29 -21959.2005944 0.183532 F<br> 30 -21959.2045923 0.059183 T<br> 31 -21959.2074746 0.148483 F<br>
32 -21959.2094894 0.139970 F<br> 33 -21959.2129334
0.033060 T<br> 34 -21959.2156972 0.237590 F<br> 35 -21959.2188749 0.065209 T<br> 36 -21959.2223341 0.020410 T<br>
37 -21959.2256989 0.041502 T<br> 38 -21959.2278845 0.206609
F<br> 39 -21959.2309458 0.064352 T<br> 40 -21959.2333476 0.228702 F<br> 41 -21959.2354058 0.034633 T<br> 42 -21959.2369634 0.053540 T<br>
43 -21959.2383146 0.179708 F<br>
44 -21959.2390413 0.052827 T<br> 45 -21959.2382847 0.084999 T<br><br> climbing image = 23<br><br> path length = 18.663 bohr<br> inter-image distance = 0.424 bohr</div>
<div style="word-wrap:break-word;font-family:Helvetica;font-size:medium"><br></div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium"><br></div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">
If you were to plot the image number versus the energy that is given above you would find what appears to be a very well behaved curve.</div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium"><br></div>
<div style="word-wrap:break-word;font-family:Helvetica;font-size:medium"><br></div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">But after a while, all of the sudden, there is a problem with the output.</div>
<div style="word-wrap:break-word;font-family:Helvetica;font-size:medium"><br></div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium"><br></div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">
<br> ------------------------------ iteration 45 ------------------------------<br><br> tcpu = 62502.4 self-consistency for image 4<br> tcpu = 62758.1 self-consistency for image 6<br> tcpu = 63006.7 self-consistency for image 8<br>
tcpu = 63287.9 self-consistency for image 10<br> tcpu = 63556.1 self-consistency for image 11<br> tcpu = 63876.3
self-consistency for image 12<br> tcpu = 64126.9 self-consistency for image 13<br> tcpu = 64409.9 self-consistency for image 15<br> tcpu = 64661.2 self-consistency for image 16<br> tcpu = 64941.7 self-consistency for image 18<br>
tcpu = 65190.6 self-consistency for image 24<br> tcpu = 65480.1 self-consistency for image 26<br> tcpu = 65780.1 self-consistency for image 29<br> tcpu = 66077.9 self-consistency for image 31<br>
tcpu
= 66317.9 self-consistency for image 32<br> tcpu = 66580.4 self-consistency for image 34<br> tcpu = 66857.5 self-consistency for image 38<br> tcpu = 67137.2 self-consistency for image 40<br>
tcpu = 67386.9 self-consistency for image 43<br><br> activation energy (->) = 0.287582 eV<br> activation energy (<-) = 0.287654 eV<br><br> image energy (eV) error (eV/A) frozen<br>
<br> 1 -21959.2382129
0.081337 T<br> 2 -21959.2388966 0.020929 T<br> 3 -21959.2381194 0.128767 T<br> 4 -21959.1561960 2.016407 T<br>
5 -21959.2352170 0.054181 T<br> 6 -21959.1233419 3.440925
F<br> 7 -21959.2301578 0.119857 T<br> 8 -21959.1295895 3.858443 F<br> 9 -21959.2256241 0.031368 T<br> 10 -21959.2007906 0.830489 T<br>
11 -21958.9506305 4.210877 F<br>
12 -21959.1226848 2.774980 F<br> 13 -21959.1255613 3.306542 F<br> 14 -21959.2093498 0.063904 T<br> 15 -21959.1425146 2.576069 F<br>
16 -21959.0686534 4.452463 F<br> 17 -21959.2013877
0.051986 T<br> 18 -21959.1237018 2.915400 F<br> 19 -21959.1986806 0.070271 T<br> 20 -21959.1962718 0.055066 T<br>
21 -21959.1966851 0.022251 T<br> 22 -21959.1954724 0.053511
T<br> 23 -21959.1942736 0.155528 T<br> 24 -21959.0321128 1.669573 T<br> 25 -21959.1964535 0.033436 T<br> 26 -21958.9641160 2.813188 F<br>
27 -21959.1987196 0.170191 T<br>
28 -21959.2008371 0.119487 T<br> 29 -21959.0410354 2.100854 T<br> 30 -21959.2045923 0.217405 T<br> 31 -21959.2006761 0.946770 T<br>
32 -21959.1091999 2.022318 T<br> 33 -21959.2129334
0.033922 T<br> 34 -21959.0832736 3.736941 F<br> 35 -21959.2188749 0.136602 T<br> 36 -21959.2223341 0.033429 T<br>
37 -21959.2256989 0.038111 T<br> 38 -21959.1226548 4.018138
F<br> 39 -21959.2309458 0.068887 T<br> 40 -21959.1163356 3.540681 F<br> 41 -21959.2354058 0.069677 T<br> 42 -21959.2369634 0.053677 T<br>
43 -21959.1651795 2.854320 F<br>
44 -21959.2390413 0.041340 T<br> 45 -21959.2382847 0.084999 T<br><br> climbing image = 11<br><br> path length = 19.063 bohr<br> inter-image distance = 0.433 bohr</div>
<div style="word-wrap:break-word;font-family:Helvetica;font-size:medium"><br></div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium"><br></div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">
Where if you plot the image number versus the energy given above you will find that many of the points lay on the same curve as in the proceeding plot, but now almost half of them are jumping to higher energies. When I look at the xyz output in VMD, there doesn't seem to be any change in what the images look like pre/post the start of iteration 45.</div>
<div style="word-wrap:break-word;font-family:Helvetica;font-size:medium"><br>I will now list some of the activation energies produced towards the end of the run (I stopped the code once I saw this happening again).</div>
<div style="word-wrap:break-word;font-family:Helvetica;font-size:medium"><br></div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium"> activation energy (->) = 0.043939 eV<br> activation energy (->) = 0.043939 eV<br>
activation energy (->) = 0.043939 eV<br> activation energy (->) = 0.046128 eV<br> activation energy (->) = 0.043939 eV<br> activation energy (->) = 0.044786 eV<br> activation energy (->) = 0.043951 eV<br>
activation energy (->) = 0.043939 eV<br> activation energy (->) = 0.043939 eV<br> activation energy (->) = 0.043939 eV<br> activation energy (->) = 0.043939 eV<br> activation energy
(->) = 0.044436 eV<br> activation energy (->) = 0.043939 eV<br> activation energy (->) = 0.043939 eV<br> activation energy (->) = 0.043939 eV<br> activation energy (->) = 0.043939 eV<br>
activation energy (->) = 0.043939 eV<br> activation energy (->) = 0.043939 eV<br> activation energy (->) = 0.043939 eV<br> activation energy (->) = 0.287582 eV<br> activation energy (->) = 0.206100 eV<br>
activation energy (->) = 0.601468 eV<br> activation energy (->) = 0.231222 eV<br> activation energy (->)
= 0.216514 eV<br> activation energy (->) = 0.252992 eV<br> activation energy (->) = 0.289003 eV<br> activation energy (->) = 0.107865 eV<br> activation energy (->) = 0.154660 eV<br>
activation energy (->) = 0.302177 eV<br> activation energy (->) = 0.154660 eV<br> activation energy (->) = 0.290079 eV<br><br></div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">
<br></div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium"><br></div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">You can see that the activation energy is fairly constant at about 0.04 eV for a while and then after a certain point bedlam.</div>
<div style="word-wrap:break-word;font-family:Helvetica;font-size:medium"><br></div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium"><br></div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">
Any ideas would be welcome.</div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium"><br></div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium"><br></div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">
<br></div><div style="word-wrap:break-word;font-family:Helvetica;font-size:medium">Chad Junkermeier, Ph.D.<br>NRC Post-Doctoral Associate<br>U. S. Naval Research Laboratory</div></div></div></div><br>_______________________________________________<br>
Pw_forum mailing list<br>
<a href="mailto:Pw_forum@pwscf.org">Pw_forum@pwscf.org</a><br>
<a href="http://www.democritos.it/mailman/listinfo/pw_forum" target="_blank">http://www.democritos.it/mailman/listinfo/pw_forum</a><br>
<br></blockquote></div><br>