From owner-chemistry@ccl.net Sun Nov 29 03:31:01 2009 From: "Ol Ga eurisco1===pochta.ru" To: CCL Subject: CCL: Problem SCF convergence(energy oscillation) in extended pi sy Message-Id: <-40807-091129031602-29424-gMJ4SwB2nBbgmQXD9fisEQ#,#server.ccl.net> X-Original-From: "Ol Ga" Date: Sun, 29 Nov 2009 03:15:54 -0500 Sent to CCL by: "Ol Ga" [eurisco1:-:pochta.ru] Dear Ronald C. Bakus, I suppose keyword XQC can help. Sincerely, OL Ga ****************************** Sent to CCL by: "Ronald C. Bakus" [rbakus!=!chem.ucsb.edu] Im having some issues with a problem convergence case for a extended pi system. Without diffuse functions, convergence is fine, however, upon addition of diffuse functions the optimization will get close to completing, but on one of the last scf cycles, the energy will decrease and then oscillate around convergence. The homo-lumo gap is small, slightly less than .2H. I have already tried various combinations of the standard ways of dealing with problem scf (qc, vshift=x (x=150,300,500,1000), nodiis, novaracc, optimization+freq calc without diffuse as starting point for diffuse, etc.) with no success. I was hoping to perhaps get some input on other ways to deal with the problem. The diffuse functions will be important to later phases of the calculations. Sample Input (I have tried both syn and anti conformers of the dimethylamino groups) ------ %chk=dsb.chk %mem=20MW #p opt b3lyp/6-31+g(d,p) geom=connectivity int=ultrafine cphf=ultrafine formcheck dsb14 opt 0 1 C C 1 B1 C 2 B2 1 A1 C 3 B3 2 A2 1 D1 C 4 B4 3 A3 2 D2 C 5 B5 4 A4 3 D3 H 1 B6 2 A5 3 D4 H 2 B7 1 A6 6 D5 H 4 B8 3 A7 2 D6 H 5 B9 4 A8 3 D7 C 3 B10 2 A9 1 D8 H 11 B11 3 A10 2 D9 C 11 B12 3 A11 2 D10 H 13 B13 11 A12 3 D11 C 13 B14 11 A13 3 D12 C 15 B15 13 A14 11 D13 C 15 B16 13 A15 11 D14 C 16 B17 15 A16 13 D15 H 16 B18 15 A17 13 D16 C 17 B19 15 A18 13 D17 H 17 B20 15 A19 13 D18 C 18 B21 16 A20 15 D19 H 18 B22 16 A21 15 D20 H 20 B23 17 A22 15 D21 C 22 B24 18 A23 16 D22 H 25 B25 22 A24 18 D23 C 25 B26 22 A25 18 D24 H 27 B27 25 A26 22 D25 C 27 B28 25 A27 22 D26 C 29 B29 27 A28 25 D27 C 29 B30 27 A29 25 D28 C 30 B31 29 A30 27 D29 H 30 B32 29 A31 27 D30 C 31 B33 29 A32 27 D31 H 31 B34 29 A33 27 D32 C 32 B35 30 A34 29 D33 H 32 B36 30 A35 29 D34 H 34 B37 31 A36 29 D35 N 6 B38 5 A37 4 D36 N 36 B39 32 A38 30 D37 C 40 B40 36 A39 32 D38 H 41 B41 40 A40 36 D39 H 41 B42 40 A41 36 D40 H 41 B43 40 A42 36 D41 C 40 B44 36 A43 32 D42 H 45 B45 40 A44 36 D43 H 45 B46 40 A45 36 D44 H 45 B47 40 A46 36 D45 C 39 B48 6 A47 5 D46 H 49 B49 39 A48 6 D47 H 49 B50 39 A49 6 D48 H 49 B51 39 A50 6 D49 C 39 B52 6 A51 5 D50 H 53 B53 39 A52 6 D51 H 53 B54 39 A53 6 D52 H 53 B55 39 A54 6 D53 B1 1.38590600 B2 1.40879299 B3 1.40676154 B4 1.38827161 B5 1.41292513 B6 1.08298050 B7 1.08583864 B8 1.08754627 B9 1.08285660 B10 1.45944010 B11 1.08912863 B12 1.35168559 B13 1.08891058 B14 1.46036881 B15 1.41101301 B16 1.40955110 B17 1.38649352 B18 1.08745064 B19 1.38650055 B20 1.08563046 B21 1.40957879 B22 1.08563312 B23 1.08745466 B24 1.46038208 B25 1.08891491 B26 1.35170500 B27 1.08912993 B28 1.45944392 B29 1.40677707 B30 1.40881876 B31 1.38826824 B32 1.08755006 B33 1.38589746 B34 1.08584524 B35 1.41292195 B36 1.08285687 B37 1.08298484 B38 1.38847165 B39 1.38841623 B40 1.45260430 B41 1.09097759 B42 1.10130937 B43 1.09581331 B44 1.45285452 B45 1.09097407 B46 1.09576437 B47 1.10121929 B48 1.45286246 B49 1.09098248 B50 1.10122787 B51 1.09575774 B52 1.45262017 B53 1.09098468 B54 1.09579664 B55 1.10131866 A1 122.07360923 A2 116.25323549 A3 122.52104756 A4 120.89016562 A5 118.44193028 A6 118.09314223 A7 118.89576382 A8 118.63570829 A9 124.25260979 A10 114.09091838 A11 127.48313332 A12 118.61687356 A13 127.13426723 A14 119.07045476 A15 124.12563764 A16 122.03442793 A17 118.76061142 A18 121.16247464 A19 119.98708840 A20 121.16222854 A21 118.85028804 A22 119.20697504 A23 124.12861653 A24 114.24477657 A25 127.13686366 A26 118.42466374 A27 127.48931804 A28 119.49046853 A29 124.25717650 A30 122.52172015 A31 118.89429417 A32 122.07417140 A33 119.83267571 A34 120.89096115 A35 118.63484888 A36 118.44040512 A37 121.62197526 A38 121.62236490 A39 119.34902804 A40 109.13960300 A41 112.95155437 A42 111.01772561 A43 119.49930105 A44 109.10238405 A45 111.07362683 A46 112.96782739 A47 119.49517384 A48 109.10177009 A49 112.97116251 A50 111.06987287 A51 119.34521560 A52 109.13771153 A53 111.01169226 A54 112.95688048 D1 -0.35125191 D2 0.32638204 D3 0.40233558 D4 179.48620457 D5 -179.90506437 D6 -179.42169103 D7 -179.52106898 D8 180.00000000 D9 178.40429122 D10 -1.82780326 D11 -0.31757066 D12 179.89196744 D13 178.73985841 D14 -1.28157819 D15 -179.92043875 D16 -0.00000000 D17 179.91842635 D18 -0.19884610 D19 -0.05473033 D20 179.90530662 D21 179.97520249 D22 -179.96404995 D23 -179.65549459 D24 0.42839517 D25 0.19711844 D26 -179.94034630 D27 -179.19548240 D28 1.14931637 D29 -179.94453526 D30 -0.22485722 D31 179.95511206 D32 0.34504623 D33 -0.42074545 D34 179.50779600 D35 -179.50941383 D36 178.51739236 D37 -178.52921489 D38 171.94094710 D39 178.43928000 D40 -61.14986272 D41 60.18975765 D42 11.63000000 D43 -178.68286485 D44 -60.46741792 D45 60.95449961 D46 170.08048307 D47 178.67201365 D48 -60.96543365 D49 60.45777828 D50 9.82506217 D51 -178.40320889 D52 -60.16058276 D53 61.18077952 From owner-chemistry@ccl.net Sun Nov 29 10:19:01 2009 From: "=?ISO-8859-1?Q?Ulf_Ekstr=F6m?= ulfek+/-few.vu.nl" To: CCL Subject: CCL:G: Problem SCF convergence(energy oscillation) in extended pi system/G03 Message-Id: <-40808-091129064817-3643-ByP0tplzuoeCMywe/Azamg!=!server.ccl.net> X-Original-From: =?ISO-8859-1?Q?Ulf_Ekstr=F6m?= Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset=ISO-8859-1 Date: Sun, 29 Nov 2009 11:58:25 +0100 MIME-Version: 1.0 Sent to CCL by: =?ISO-8859-1?Q?Ulf_Ekstr=F6m?= [ulfek+/-few.vu.nl] > Im having some issues with a problem convergence case for a extended pi s= ystem. > Without diffuse functions, convergence is fine, however, upon addition of= diffuse functions > the optimization will get close to completing, but on one of the last scf= cycles, the energy will > decrease and then oscillate around convergence. In my experience (with a different code than Gaussian) you sometimes need to turn off any kind of integral screening or linear scaling method in cases like this. Your system is not that big, but screening is probably used. The reason this helps is that with integral screening the gradient (off-diagonal Fock matrix elements) is not in fact the true gradient of the system, but an approximation that contains noise. Most algorithms we use are not designed to handle "unphysical" noise. Another thing you might try is to check for linear dependence in your diffuse basis set. Sometimes certain combinations (i.e. diffuse s functions with opposite sign on neighboring atoms) have huge molecular orbital coefficients. These huge coefficients magnify numerical errors in the one and two electron integrals, but I would guess Gaussian would print a warning in this case. If your problems are due to linear dependencies then you might consider a single center diffuse basis set, which can be made very large without problems. Also check if your DFT integration grid is good enough for the diffuse functions. Sincerely, Ulf Ekstr=F6m, VU University Amsterdam From owner-chemistry@ccl.net Sun Nov 29 12:43:00 2009 From: "xunlei ding dingxunlei*gmail.com" To: CCL Subject: CCL:G: Gaussian03 TZVP/B3LYP and ghost atoms Message-Id: <-40809-091128044004-4542-DOKVGB+aveaRdBMiOQ+3lA(~)server.ccl.net> X-Original-From: xunlei ding Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset=GB2312 Date: Sat, 28 Nov 2009 17:10:45 +0800 MIME-Version: 1.0 Sent to CCL by: xunlei ding [dingxunlei++gmail.com] Dear Bjoern, In g03, there is a keyword "counterpoise" to do the bsse calculation. I think you can try it at first. Or can you provide the input file? Best wishes, Ding 2009/11/27 Bjoern Baumeier baumeier()mpip-mainz.mpg.de : > Hi all, > > I have a problem with Gaussian03. For geometry consisting of two > molecules, I have to run calculations for each individual molecule using > a "counterpoise" basis set, i.e. putting ghost atoms at the positions > of the second molecule. > > If I perform these calculations using the 6-311++G** basis set, > everything works fine. When I switch to TZVP (or TZV), the calculations > never converge. In fact, the energy drops after two or three cycles from > the typical single molecule value (~1670 a.u.) to around the double. > > Has anyone ever observed something similar? > > Cheers, > Bjoern > > > --=20 ------------------------------------------------------ Xun-Lei Ding (=B6=A1=D1=B8=C0=D7) Associate Research Professor of Physical Chemistry, Ph.D State Key Lab for Struct. Chem. of Unstable and Stable Species Institute of Chemistry, Chinese Academy of Sciences Zhongguancun North First Street 2=A3=AC Beijing 100190, P. R. China Phone 86-10-62568330 Fax 86-10-62559373 From owner-chemistry@ccl.net Sun Nov 29 13:18:00 2009 From: "Mariusz Radon mariusz.radon(a)gmail.com" To: CCL Subject: CCL: Energy difference Message-Id: <-40810-091128191733-17734-chjIumZ7QNyLJ1SK6RPr/g%server.ccl.net> X-Original-From: Mariusz Radon Content-Type: text/plain; charset=UTF-8 Date: Sun, 29 Nov 2009 01:09:57 +0100 MIME-Version: 1.0 Sent to CCL by: Mariusz Radon [mariusz.radon*|*gmail.com] On Sat, Nov 28, 2009 at 3:12 PM, S. Bill s_bill36.##.yahoo.co.uk wrote: > I was wondering, is 8 kcal/mol a big energy difference between two isomers? could you direct me to any articles or books concer this point? This is because of Boltzmann distribution of probability as a function of energy: P(E) = const * exp(-E/(RT)) (I would suggest any thermodynamical textbook). Assuming that Boltzmann formula applies (i.e., assuming an equilibrium between both isomers), we can calculate that the isomer of 8 kcal/mol higher energy would be 1.4 million (!) less abundant than the one of lower energy for T=298K (room temperature). You may ask what is "small energy difference" between isomers. Maybe the difference for which Boltzmann exponent exp(-E/(RT)) yield a much larger number than 1.4e-6, let say 0.1 (10%). This would yield the energy difference of 1.5 kcal/mol. Most people would, indeed, say that 1-2 kcal/mol is a "small" energy difference between isomers, wheras 8 kcal/mol is "large". Greetings, Mariusz Radon From owner-chemistry@ccl.net Sun Nov 29 23:19:00 2009 From: "Green Power powergreen-#-gmail.com" To: CCL Subject: CCL: effective nuclear charges Message-Id: <-40811-091129185411-5122-QUMMMr75MM7wzm02Pim1ig],[server.ccl.net> X-Original-From: Green Power Content-Type: multipart/alternative; boundary=00c09f972535f4f59e04798b2472 Date: Sun, 29 Nov 2009 18:47:14 -0500 MIME-Version: 1.0 Sent to CCL by: Green Power [powergreen..gmail.com] --00c09f972535f4f59e04798b2472 Content-Type: text/plain; charset=ISO-8859-1 Dear All, Could you suggest a way to construct an atom X with same number of nuclear charge and electron as Hydrogen(ie. Z=1,e=1), but with higher electronegativity than H. The reason I do this is that I want to know how the electronegtivity of X in a molecule affect the chemical properties. Thank you. Regards Roy --00c09f972535f4f59e04798b2472 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable 
Dear All,

Could you suggest a way to construct an atom X with s=
ame number of nuclear charge and electron as Hydrogen(ie. Z=3D1,e=3D1), but=
 with higher electronegativity  than H. The reason I do this is that I want=
 to know how the electronegtivity of X in a molecule affect the chemical pr=
operties.

Thank you.

Regards

Roy
--00c09f972535f4f59e04798b2472--