From owner-chemistry@ccl.net Fri Aug 21 04:26:01 2009 From: "Serdar Badošlu sbadoglu * gazi.edu.tr" To: CCL Subject: CCL: ERROR FREQUENCY CALCULATIONS Message-Id: <-40043-090821041553-14537-j8vfNZdFNQKVuSPJuCmdEw[]server.ccl.net> X-Original-From: "Serdar Badošlu" Date: Fri, 21 Aug 2009 07:49:28 -0000 Sent to CCL by: "Serdar Badošlu" [sbadoglu(_)gazi.edu.tr] Hi Carlos, This looks like a typical memory problem to me. I suppose you should increase the "%mem=" value. Regards. -- Serdar BADOGLU http://docs.google.com/View?id=dckx8v2c_286bpvdsdt From owner-chemistry@ccl.net Fri Aug 21 06:20:03 2009 From: "Van Dam, Huub (STFC,DL,CSE) huub.van-dam[a]stfc.ac.uk" To: CCL Subject: CCL: Gromacs Error Message-Id: <-40044-090820194638-5508-GqiSxiF+vfF553cNNeK8bw+*+server.ccl.net> X-Original-From: "Van Dam, Huub (STFC,DL,CSE)" Content-class: urn:content-classes:message Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="iso-8859-1" Date: Thu, 20 Aug 2009 23:46:16 +0100 MIME-Version: 1.0 Sent to CCL by: "Van Dam, Huub (STFC,DL,CSE)" [huub.van-dam()stfc.ac.uk] Dear Cintia, As you can see your file 'dppc128.top' contains various include = statements. I suspect that grompp uses cpp to include the files your = .top file is referring to. Clearly in your case there is a problem with = that. Either you don't have cpp at all or it is installed in an unusual = place. To answer this question type: which cpp If cpp is installed (and in your path) the system will tell where it is. = If it turns out that your cpp is not /usr/bin/cpp but something else = then you might be able to tell grompp about this in the .mdp file (see = gromacs-4.0.5/share/tutor/water/grompp.mdp for example). If you don't = have cpp at all (how did GROMACS get compiled on your system?) then it = seems you would have to use the -pp option on grompp. I am not sure = exactly what this entails but it seems that you might have to provide = grompp in that case with an .mdp file that have all includes resolved = already. If that is so you either need to find an alternative = pre-processor to resolve the include statements for you or you would = have to include the files by hand. I hope this helps. Best wishes, Huub =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D Huub van Dam (huub.van-dam(a)stfc.ac.uk, +44-1925-603933) =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D -----Original Message----- > From: owner-chemistry+huub.van-dam=3D=3Dstfc.ac.uk(a)ccl.net = [mailto:owner-chemistry+huub.van-dam=3D=3Dstfc.ac.uk(a)ccl.net] On Behalf = Of "C=EDntia C. Vequi-Suplicy" cintiadevequi=3D=3D=3Dgmail.com Sent: 19 August 2009 16:18 To: Van Dam, Huub (STFC,DL,CSE) Subject: CCL: Gromacs Error Sent to CCL by: = "=3D?ISO-8859-1?Q?=3D22C=3DEDntia_C=3D2E_Vequi-Suplicy=3D22?=3D" = [cintiadevequi=3D=3D=3Dgmail.com] Hello, I am trying to simulate a lipid membrane with Gromacs. I am trying to=20 generate the *.tpr file o run the simulation, but every time I use the=20 command grompp I got the following error: checking input for internal consistency... ...ling /usr/bin/cpp : No such file or directory cpp exit code: 32512 dppc128.top > gromppzJXISa'cpp ' command is defined in the .mdp file ************************************************************* * Important: if you do not have a c-preprocessor installed, * * you MUST run grompp with the "-pp" command line option * * in place of the "-p" option. * ************************************************************* processing topology... processing coordinates... ------------------------------------------------------- Program grompp, VERSION 3.3.3 Source code file: ../../../../src/kernel/grompp.c, line: 469 Fatal error: number of coordinates in coordinate file (dppc128.gro, 17365) does not match topology (dppc128.top, 0) ------------------------------------------------------- "One Cross Each" (Monty Python) I don=B4t know if I have problems with the cpp or my input file for=20 topology. In the topology file I have the following lines: ; ; File 'dppc128.top' was generated ; By user: onbekend (0) ; On host: onbekend ; At date: Tue Aug 18 10:02:52 2009 ; ; This is your topology file ; Pure DPPC bilayer with 128 lipids and 3655 water molecules ; ; The force field files to be include #include "ffoplsaa.itp" ; Include DPPC topology #include "dppc.itp" #include "lipid.itp" #include "spc.itp" [ molecules ] ; Compound #mol DPP 128 SOL 3655 If anyone can help It will be very nice, Thank you very much so far, C=EDntia C. Vequi-Suplicy Biophysics Group Physics Institute S=E3o Paulo University -=3D This is automatically added to each message by the mailing script = =3D-http://www.ccl.net/cgi-bin/ccl/send_ccl_messageSubscribe/Unsubscribe:=20Job: http://www.ccl.net/jobs=20http://www.ccl.net/spammers.txt-- Scanned by iCritical. From owner-chemistry@ccl.net Fri Aug 21 08:23:01 2009 From: "Carlos Pereira carlosps1985[-]gmail.com" To: CCL Subject: CCL: ERROR FREQUENCY CALCULATIONS Message-Id: <-40045-090821080129-3970-MctfRSDE1MLE+QRot1W6wg * server.ccl.net> X-Original-From: Carlos Pereira Content-Type: multipart/alternative; boundary=001636ed761acef3bb0471a58c75 Date: Fri, 21 Aug 2009 08:55:52 -0300 MIME-Version: 1.0 Sent to CCL by: Carlos Pereira [carlosps1985]^[gmail.com] --001636ed761acef3bb0471a58c75 Content-Type: text/plain; charset=ISO-8859-9 Content-Transfer-Encoding: quoted-printable OK, Thanks... I am testing.. 2009/8/21 Serdar Bado=F0lu sbadoglu * gazi.edu.tr > > Sent to CCL by: "Serdar Bado=F0lu" [sbadoglu(_)gazi.edu.tr] > Hi Carlos, > > This looks like a typical memory problem to me. I suppose you should > increase > the "%mem=3D" value. > > Regards. > > > -- > Serdar BADOGLU > http://docs.google.com/View?id=3Ddckx8v2c_286bpvdsdt > > > > -=3D This is automatically added to each message by the mailing script = =3D-> > > --=20 Carlos Pereira da Silva http://Carlosps.110mb.com carlosps1985*gmail.com (86)99223852 --001636ed761acef3bb0471a58c75 Content-Type: text/html; charset=ISO-8859-9 Content-Transfer-Encoding: quoted-printable
OK, Thanks...
=A0
I am testing..

2009/8/21 Serdar Bado=F0lu sbadoglu * gazi.edu.tr <owner-chemistry*ccl.net>

Sent to CCL by: "Serdar= Bado=F0lu" [sbadoglu(_)gazi.edu.tr]
Hi Carlos,

This looks like a typical memory problem to me. I suppose= you should increase
the "%mem=3D" value.

Regards.
<= br>
--
Serdar BADOGLU
http://docs.google.com/View?id=3Ddck= x8v2c_286bpvdsdt



-=3D This is automatically added to each message by the mailing= script =3D-
To recover the email address of the author of the message, = please change
the strange characters on the top line to the * sign. You = can also
E-mail to subs= cribers: CHEMISTRY*ccl.net or use:=
=A0 =A0 =A0http://www.ccl.net/cgi-bin/ccl/send_ccl_message

E-mail to administrators: = CHEMISTRY-REQUEST*ccl.net or use
=A0 =A0 =A0http://www.ccl.net/cg= i-bin/ccl/send_ccl_message

Subscribe/Unsubscribe:
=A0 =A0 =A0http://www.ccl.net/chemistry/sub_u= nsub.shtml

Before posting, check wait time at: http://www.ccl.net

Job: http://www.c= cl.net/jobs
Conferences: http://server.ccl.net/chemist= ry/announcements/conferences/

Search Messages: http://www.ccl.net/chemistry/searchccl/index.sht= ml
=A0= =A0 =A0http:= //www.ccl.net/spammers.txt

RTFI: http://www.ccl.net/chemistry/aboutccl/instructions/




--
Carlos Pereira da Silva
http://Carlosps.110mb.com
carlosps1985*gmail.com
(86)99223852
--001636ed761acef3bb0471a58c75-- From owner-chemistry@ccl.net Fri Aug 21 09:26:00 2009 From: "Andreas Uhe andiuhe]|[gmx.de" To: CCL Subject: CCL:G: TS energy slightly lower than minimum energy Message-Id: <-40046-090821043304-18601-G3/oCO6gvKBz9uz1yNCquQ ~~ server.ccl.net> X-Original-From: "Andreas Uhe" Date: Fri, 21 Aug 2009 04:32:59 -0400 Sent to CCL by: "Andreas Uhe" [andiuhe..gmx.de] Dear CCLers, I hope you can help me explain the following result: Using Gaussian03 I calculated a transition state and the respective energetic minima connected by the transition state (I checked with an IRC calculation). As one expects the electronic energy of the TS is higher than the electronic energy of the minimum. The problem is that this is reversed for the Gibbs free energy obtained by frequency calculations: E Ezpe H G transition state -1624,071411 -1623,310413 -1623,269941 -1623,378578 following minimum -1624,075219 -1623,311174 -1623,270443 -1623,378231 Is there a chemical or physical explanation for this? Or is it an artefact of the calculations? How can I interpret this result? Thanks for your help! Andi From owner-chemistry@ccl.net Fri Aug 21 10:01:00 2009 From: "Tom de Greef t.f.a.d.greef.**.tue.nl" To: CCL Subject: CCL:G: TDDFT problem Message-Id: <-40047-090821043725-20376-Zay2eQ5fXC72GWGpVUTeMA**server.ccl.net> X-Original-From: "Tom de Greef" Date: Fri, 21 Aug 2009 04:37:19 -0400 Sent to CCL by: "Tom de Greef" [t.f.a.d.greef#%#tue.nl] Dear all, Recently, I have run into some problems with TD-DFT calculations using the Gaussian 03 program. I have optimized a molecule of around 132 atoms, of which half are carbons, nitrogens and oxygen atoms. Optmization and frequency analysis revealed that the molecule has been optmized properly (no imaginary frequencies). Now I would like to calculate the vertical excitations using TD-DFT. However, when I use TDDFT with td=(nstates=30,root=1) almost all my oscillator strengths are close to zero. My first question is, is there a way in which I can tell Gaussian to skip finding states of which the oscillator strength is almost zero? My second question relates to time issues. Does anyone know how my calculation time scales with the number of states I am interested in? I use the PBE0 functional and a 6-31G+(d,p) basis set. How much additional time would it take to find lets say 50 states? Thank you for your help. Best regards, Tom From owner-chemistry@ccl.net Fri Aug 21 13:24:01 2009 From: "Sebastian Kozuch kozuchs^^^yahoo.com" To: CCL Subject: CCL:G: TS energy slightly lower than minimum energy Message-Id: <-40048-090821131517-24105-GYEzD94zrQ6QUxQqPA3vZg===server.ccl.net> X-Original-From: Sebastian Kozuch Content-Type: multipart/alternative; boundary="0-1788365986-1250871301=:95978" Date: Fri, 21 Aug 2009 09:15:01 -0700 (PDT) MIME-Version: 1.0 Sent to CCL by: Sebastian Kozuch [kozuchs:_:yahoo.com] --0-1788365986-1250871301=:95978 Content-Type: text/plain; charset=us-ascii Dear Andi, The fact that you found a TS in the energy profile does not mean that it is a TS in the free energy profile. Usually the two states are in a slightly different position, although most of the times this will not give you significant errors. In your case, the energy difference of the intermediate and the TS is small, so adding ZPE, entropy and extras on the same geometry inverted the order. It is theoretically possible that there is no TS at all in the G profile. The ways to find a TS in the free energy profile are not easy, as it has to take into account frequency calculations on each step of the optimization. This is only recommended for high accuracy jobs. There is also the fact that in a solvent the entropy is overestimated by the usual gas phase calculations. Life is not easy. The question that arises is: do you really need to be that accurate? Is it an important step on the reaction? Diffusion steps can be slower than the step you have calculated, and most of the times we don't pay attention to that also. In resume, it is an artifact since you are looking at the free energy but on the geometry calculated with only internal energy. But most of us don't care of those subtleties since we have bigger errors coming from other sources. xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx ..........Sebastian Kozuch........... xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx http://yfaat.ch.huji.ac.il/kozuch.htm xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx ________________________________ > From: Andreas Uhe andiuhe]|[gmx.de To: "Kozuch, Sebastian " Sent: Friday, August 21, 2009 11:32:59 AM Subject: CCL:G: TS energy slightly lower than minimum energy Sent to CCL by: "Andreas Uhe" [andiuhe..gmx.de] Dear CCLers, I hope you can help me explain the following result: Using Gaussian03 I calculated a transition state and the respective energetic minima connected by the transition state (I checked with an IRC calculation). As one expects the electronic energy of the TS is higher than the electronic energy of the minimum. The problem is that this is reversed for the Gibbs free energy obtained by frequency calculations: E Ezpe H G transition state -1624,071411 -1623,310413 -1623,269941 -1623,378578 following minimum -1624,075219 -1623,311174 -1623,270443 -1623,378231 Is there a chemical or physical explanation for this? Or is it an artefact of the calculations? How can I interpret this result? Thanks for your help! Andihttp://www.ccl.net/cgi-bin/ccl/send_ccl_messagehttp://www.ccl.net/chemistry/sub_unsub.shtmlhttp://www.ccl.net/spammers.txt--0-1788365986-1250871301=:95978 Content-Type: text/html; charset=us-ascii
Dear Andi,
The fact that you found a TS in the energy profile does not mean that it is a TS in the free energy profile. Usually the two states are in a slightly different position, although most of the times this will not give you significant errors. In your case, the energy difference of the intermediate and the TS is small, so adding ZPE, entropy and extras on the same geometry inverted the order. It is theoretically possible that there is no TS at all in the G profile.
The ways to find a TS in the free energy profile are not easy, as it has to take into account frequency calculations on each step of the optimization. This is only recommended for high accuracy jobs. There is also the fact that in a solvent the entropy is overestimated by the usual gas phase calculations. Life is not easy.
The question that arises is: do you really need to be that accurate? Is it an important step on the reaction? Diffusion steps can be slower than the step you have calculated, and most of the times we don't pay attention to that also.

In resume, it is an artifact since you are looking at the free energy but on the geometry calculated with only internal energy. But most of us don't care of those subtleties since we have bigger errors coming from other sources.
 
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
..........Sebastian Kozuch...........
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
http://yfaat.ch.huji.ac.il/kozuch.htm
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx


From: Andreas Uhe andiuhe]|[gmx.de <owner-chemistry*ccl.net>
To: "Kozuch, Sebastian " <kozuchs*yahoo.com>
Sent: Friday, August 21, 2009 11:32:59 AM
Subject: CCL:G: TS energy slightly lower than minimum energy


Sent to CCL by: "Andreas  Uhe" [andiuhe..gmx.de]
Dear CCLers,

I hope you can help me explain the following result:
Using Gaussian03 I calculated a transition state and the respective energetic minima connected by the transition state (I checked with an IRC calculation). As one expects the electronic energy of the TS is higher than the electronic energy of the minimum. The problem is that this is reversed for the Gibbs free energy obtained by frequency calculations:

                        E            Ezpe            H            G
transition state  -1624,071411  -1623,310413  -1623,269941  -1623,378578
following minimum  -1624,075219  -1623,311174  -1623,270443  -1623,378231


Is there a chemical or physical explanation for this? Or is it an artefact of the calculations? How can I interpret this result?

Thanks for your help!
Andi



E-mail to subscribers: CHEMISTRY*ccl.net or use:
      http://www.ccl.net/cgi-bin/ccl/send_ccl_message

E-mail to administrators: CHEMISTRY-REQUEST*ccl.net or use
      http://www.ccl.net/cgi-bin/ccl/send_ccl_message
      http://www.ccl.net/chemistry/sub_unsub.shtml

Before posting, check wait time at: http://www.ccl.net

Job: http://www.ccl.net/jobs
Conferences: http://server.ccl.net/chemistry/announcements/conferences/

Search Messages: http://www.ccl.net/chemistry/searchccl/index.shtml
      http://www.ccl.net/spammers.txt

RTFI: http://www.ccl.net/chemistry/aboutccl/instructions/



--0-1788365986-1250871301=:95978-- From owner-chemistry@ccl.net Fri Aug 21 14:27:00 2009 From: "Johannes Johansson johjo76%gmail.com" To: CCL Subject: CCL:G: TS energy slightly lower than minimum energy Message-Id: <-40049-090821142505-20899-4gWRgmCBOB/1Vwmj+bPcIA- -server.ccl.net> X-Original-From: Johannes Johansson Content-Type: multipart/alternative; boundary=000e0cd59cb0a135bd0471aafb9f Date: Fri, 21 Aug 2009 20:24:46 +0200 MIME-Version: 1.0 Sent to CCL by: Johannes Johansson [johjo76!=!gmail.com] --000e0cd59cb0a135bd0471aafb9f Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Dear Andreas, this is quite common for very low barriers like the one you show here (whic= h is less than 3 kcal/mol). Regarding the way we search for TS, i.e. at the electronic energy "level" and the approximations involved in the ideal gas statistical mechanics used to obtain "corrections" for the thermodynamic potentials (by for example Gaussian) this is not strange at all, actually i= t is quite common in the literature when the barrier is low. (Also, you can see in your values that it is in the entropy term that the picture changes and that part is known to be the least accurately described by the ideal ga= s stat mech used.) For your case this probably means nothing and the quality of your predictio= n based on the calculated "barrier" is most likely as good as for any other barrier you calculated the same way, i.e. with an uncertainty of 3-5 kcal/mol for a TM containing system. Also, the same thing could happen when you calculate dielectric effects for a low barrier TS located in gas phase. If you are interested in this issues you can read Donald Truhlars papers on "Generalized TST", for example J. Chem. Phys. *119*, 5828 (2003).Best regards/Johannes, RWTH ;-) 2009/8/21 Andreas Uhe andiuhe]|[gmx.de > > Sent to CCL by: "Andreas Uhe" [andiuhe..gmx.de] > Dear CCLers, > > I hope you can help me explain the following result: > Using Gaussian03 I calculated a transition state and the respective > energetic minima connected by the transition state (I checked with an IRC > calculation). As one expects the electronic energy of the TS is higher th= an > the electronic energy of the minimum. The problem is that this is reverse= d > for the Gibbs free energy obtained by frequency calculations: > > E Ezpe H G > transition state -1624,071411 -1623,310413 -1623,269941 -1623,378578 > following minimum -1624,075219 -1623,311174 -1623,270443 -1623,378231 > > > Is there a chemical or physical explanation for this? Or is it an artefac= t > of the calculations? How can I interpret this result? > > Thanks for your help! > Andi > > > > -=3D This is automatically added to each message by the mailing script = =3D-> > > --=20 Adam Johannes Johansson Ph.D., M.Sc. Institut f=C3=BCr Organische Chemie RWTH Aachen University Landoltweg 1 52056 Aachen Germany Cell phone:0708178487 E-mail:johjo76-,-gmail.com Research web page:www.physto.se/~johjo --000e0cd59cb0a135bd0471aafb9f Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable
Dear Andreas,=C2=A0

<= /div>
this is quite common for very low barriers like the one you show here (whic= h is less than 3 kcal/mol). Regarding the way we search for TS, i.e. at the= electronic energy "level" and the approximations involved in the= ideal gas statistical mechanics used to obtain "corrections" for= the thermodynamic potentials (by for example Gaussian) this is not strange= at all, actually it is quite common in the literature when the barrier is = low. (Also, you can see in your values that it is in the entropy term that = the picture changes and that part is known to be the least accurately descr= ibed by the ideal gas stat mech used.)

For your case this probably means nothing and the quali= ty of your prediction based on the calculated "barrier" is most l= ikely as good as for any other barrier you calculated the same way, i.e. wi= th an uncertainty of 3-5 kcal/mol for a TM containing system.

Also, the same thing could happen when you calcul= ate dielectric effects for a low barrier TS located in gas phase.

If you = are interested in this issues you can read Donald Truhlars papers on "= Generalized TST",=C2=A0for=C2=A0example=C2=A0

J. Chem. Phys.=C2=A0119, 5828=C2=A0(2003).

Best r= egards/Johannes, RWTH ;-)



2009/8/21 Andreas Uhe andiuhe]|[= gmx.de <owner-chemistry-,-ccl.net>

Sent to CCL by: "Andreas =C2=A0Uhe" [andiuhe..gmx.de]
Dear CCLers,

I hope you can help me explain the following result:
Using Gaussian03 I calculated a transition state and the respective energet= ic minima connected by the transition state (I checked with an IRC calculat= ion). As one expects the electronic energy of the TS is higher than the ele= ctronic energy of the minimum. The problem is that this is reversed for the= Gibbs free energy obtained by frequency calculations:

=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2= =A0 =C2=A0E =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0Ezpe =C2=A0 =C2=A0 =C2= =A0 =C2=A0 =C2=A0 =C2=A0H =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0G
transition state =C2=A0 -1624,071411 =C2=A0-1623,310413 =C2=A0-1623,269941 = =C2=A0-1623,378578
following minimum =C2=A0-1624,075219 =C2=A0-1623,311174 =C2=A0-1623,270443 = =C2=A0-1623,378231


Is there a chemical or physical explanation for this? Or is it an artefact = of the calculations? How can I interpret this result?

Thanks for your help!
Andi



-=3D This is automatically added to each message by the mailing script =3D-=
E-mail to subscribers: CHEMISTRY-,-ccl.n= et or use:
=C2=A0 =C2=A0 =C2=A0http://www.ccl.net/cgi-bin/ccl/send_ccl_message=

E-mail to administrators: CHEM= ISTRY-REQUEST-,-ccl.net or use
=C2=A0 =C2=A0 =C2=A0http://www.ccl.net/cgi-bin/ccl/send_ccl_message=

Subscribe/Unsubscribe:
=C2=A0 =C2=A0 =C2=A0http://www.ccl.net/chemistry/sub_unsub.shtml

Before posting, check wait time at: http://www.ccl.net

Job: http://www.ccl.n= et/jobs
Conferences: http://server.ccl.net/chemistry/announcements/co= nferences/

Search Messages: http://www.ccl.net/chemistry/searchccl/index.shtml
=C2=A0 =C2=A0 =C2=A0http://www.ccl.net/spammers.txt

RTFI: http://www.ccl.net/chemistry/aboutccl/instructions/





--
Adam Johannes Johansson=
Ph.D., M.Sc.
Institut f=C3=BCr Organische Chemie
RWTH Aachen Univ= ersity
Landoltweg 1
52056 Aachen
Germany
Cell phone:0708178487<= br> E-mail:johjo76-,-gmail.com<= br>Research web page:www.physto.se/= ~johjo

--000e0cd59cb0a135bd0471aafb9f-- From owner-chemistry@ccl.net Fri Aug 21 16:34:00 2009 From: "Kalju Kahn kalju|,|chem.ucsb.edu" To: CCL Subject: CCL: Atomic Oxygen 3P starting to make sense Message-Id: <-40050-090821163235-8128-oZK/3J9WR4n2ph98FALhQQ**server.ccl.net> X-Original-From: "Kalju Kahn" Content-Transfer-Encoding: 8bit Content-Type: text/plain;charset=iso-8859-1 Date: Fri, 21 Aug 2009 13:32:21 -0700 MIME-Version: 1.0 Sent to CCL by: "Kalju Kahn" [kalju*_*chem.ucsb.edu] Dear All, I had posted yesterday a question about calculating atomic oxygen, and Prof. Alex Granovsky kindly pointed out that the discrepancy arises due to the spherical symmetry constraints on this non-spherical atom. > http://classic.chem.msu.su/cgi-bin/ceilidh.exe/gran/gamess/forum/?C34df668afbHW-7073-1127+00.htm If I get it correctly, atomic SCF codes typically enforce spherical symmetry while general purpose codes use whatever symmetry group the user specifies (or default to C1). From variational principle, it follows that if one lifts symmetry constrains, one gets lower energy, and I see this with NWChem's ROHF calculations: Group name O E(SCF)= -74.8084926724 Group name Td E(SCF)= -74.8084926724 .......................................... Group name D2h E(SCF)= -74.8113776132 Group name C1 E(SCF)= -74.8113776132 .......................................... Partridge reported -74.809396 MCSCF w/ PC GAMESS/Firefly -74.8093961 What I am still confused is what is the correct value here. Oxygen atom is non-spherical, so I fail to see how enforcing spherical symmetry is justified. What would a correct symmetry group be here? Second, does this mean that it is dangerous to use ROHF or UHF calculations of non-spherical atoms when evaluating atomization energies with common QM programs? Does one need to pay a close attention to a symmetry group of an atom when defining the reference wave function for, say CCSD? Or are single reference UHF/ROHF calculations of atoms with two unpaired electrons just an incorrect thing to do, and one should use multi-reference methods exclusively for triplet oxygen and carbon atoms? Is this how the basis set exponents are determined for such atoms? Thank you for your input, Kalju ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Dr. Kalju Kahn Department of Chemistry and Biochemistry UC Santa Barbara, CA 93106 From owner-chemistry@ccl.net Fri Aug 21 17:09:01 2009 From: "Gerald Knizia knizia#,#theochem.uni-stuttgart.de" To: CCL Subject: CCL:G: Triplet (3P) Oxygen HF energy lower than the reported HF limit? Message-Id: <-40051-090821164040-11270-j3OzxHNopRRcfLrNx3yGWA(~)server.ccl.net> X-Original-From: Gerald Knizia Content-Disposition: inline Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset="iso-8859-1" Date: Fri, 21 Aug 2009 22:00:46 +0200 MIME-Version: 1.0 Sent to CCL by: Gerald Knizia [knizia%a%theochem.uni-stuttgart.de] On Friday 21 August 2009 01:29, Kalju Kahn kalju-.-chem.ucsb.edu wrote: > When I ran ROHF or UHF calculations of atomic oxygen in the triplet state, > I get energies quite a bit lower than the reported SCF limit (-74.809398). > > For example, I use Partridge uncontracted 18s13p basis, for which > Partridge in 1988 has reported -74.809398, in a good agreement with > numeric HF value. Several authors who have looked into large Gaussian > basis sets (e.g. Jorge's generator coordinate GTOs) give similar results. > When I run what I think is a triplet oxygen atom, I get: > > E(ROHF) -74.811377613235 > E (UHF) -74.814047051515 This is likely a question of enforced spherical symmetry. For example, Molpro says that the ROHF/uncontracted AV6Z energy for O is -74.81237835, quite similar to what you report. But you need to consider that the spherical symmetry is broken in this calculation: One p orbital is different than the other ones. Some people don't like that and enforce equivalence. A state-averaged MCSCF with spherical symmetry (i.e., where all three different triplet states are forced to use the same orbitals) results in -74.80938176, which is rather close to Partridge's value. -- Gerald Knizia From owner-chemistry@ccl.net Fri Aug 21 20:02:00 2009 From: "Gerald Knizia knizia|-|theochem.uni-stuttgart.de" To: CCL Subject: CCL: Atomic Oxygen 3P starting to make sense Message-Id: <-40052-090821195801-4590-Ciysd3D8KLD8RIvOT7bgyQ*server.ccl.net> X-Original-From: Gerald Knizia Content-Disposition: inline Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset="iso-8859-1" Date: Sat, 22 Aug 2009 01:57:45 +0200 MIME-Version: 1.0 Sent to CCL by: Gerald Knizia [knizia::theochem.uni-stuttgart.de] On Friday 21 August 2009 22:32, Kalju Kahn kalju|,|chem.ucsb.edu wrote: > If I get it correctly, atomic SCF codes typically enforce spherical > symmetry while general purpose codes use whatever symmetry group the user > specifies (or default to C1). The main problem here is that most general purpose programs (for correlated calculations, at least) are not able to handle any point groups which are not subgroups of D2h. This is the largest commutative point group which is easy to handle without complex numbers. Consequently, the programs have only intrinsic support for combinations of reflections (x->-x, y->-y, z->-z, [x->-x & y->-y], ...) as symmetry elements -- a set which is not sufficient to represent 90 degree rotations, and thus admits inequivalent p orbitals. > What I am still confused is what is the correct value here. Oxygen atom > is non-spherical, so I fail to see how enforcing spherical symmetry is > justified. What would a correct symmetry group be here? The /geometry/ of the O atom is spherically symmetric. Consequently, the Hamilton operator H of the system is supposed to commute with every symmetry element of the point group. As commuting operators can be diagonalized together, in principle each hypothetical exact wave function can be chosen in such a way that it is at the same time an eigenvector to both H and every element of a minimal generator set of the symmetry group[1]. This symmetry-operator-eigenvector property means that applying any symmetry operator to the wave function should reproduce the original wave function, except for an (irrelevant) phase factor. Thus, an exact wave function of the O atom /should/ be spherically symmetric. [1] E.g., C3h has two generators: z-reflection and 180deg/3 rotation in the xy-plane. The group is spanned by combining these in all possible ways. > Second, does this mean that it is dangerous to use ROHF or UHF > calculations of non-spherical atoms when evaluating atomization energies > with common QM programs? This is certainly done commonly, both in DFT and correlated communities. Nevertheless, we're dealing with symmetry-broken solutions, and not everyone is happy with that. The common way of dealing with this issue is to ignore it. > Does one need to pay a close attention to a symmetry group of an > atom when defining the reference wave function for, say CCSD? For CCSD this is unlikely to be significant. CCSD doesn't care much about the orbitals it is given. There is also no reason to believe that the non-equivalent orbitals are worse than the equivalent ones -- at least as long as the symmetry is not enforced in the CCSD, either. > Or are single reference UHF/ROHF calculations of atoms with > two unpaired electrons just an incorrect thing to do, and one should use > multi-reference methods exclusively for triplet oxygen and carbon atoms? MR is the best approach, in principle. But the problem here is that both, programs for highly correlated multi-reference treatments, and general point group support in QC programs are sparse. There is no natural extension of CCSD(T) to MR cases, and very few of the many, many suggested MR-CCSD algorithms have evolved into production level programs. And certainly no method is generally agreed on to be ``the way to go''. MRCI+q/MR-ACPF/MR-CEPA is basically the best you have for MR. Considering that, e.g., CCSD is very insensitive to the actual choice of orbitals, the somewhat broken symmetry of atoms is not generally considered to be a significant problem. Just like UHF programs are commonly used, although they break spin symmetry. > Is this how the basis set exponents are determined for such atoms? I'm not quite sure on this, but I think the Dunning-type cc-pVnZ sets were optimized with enforced spherical symmetry, while, for example, the Turbomole sets def2-nZVPP were not. It does not appear to make much of a difference in practice. -- Gerald Knizia