From owner-chemistry@ccl.net Thu Oct 11 07:53:01 2007 From: "Jing Tao Lu tower.lu()gmail.com" To: CCL Subject: CCL:G: cubic force constant in gausian03 Message-Id: <-35368-071011074915-12553-NPx1DXQQUWOG/YkQg/p54A,+,server.ccl.net> X-Original-From: "Jing Tao Lu" Date: Thu, 11 Oct 2007 07:49:12 -0400 Sent to CCL by: "Jing Tao Lu" [tower.lu a gmail.com] Hi, I have one question on the numerical energy third derivatives that gaussian outputs when I use freq=cubic. For 2nd derivatives which is the force constants, I am sure that gaussian output is in atomic unit, without mass-weighted: (\partial E/\partial u_i \partial u_j with dimension Hartree/bohr^2). Is the 3rd derivatives in dimension Hartree/bohr^3 ? Below is part of the outputs: --------------------------------------------------------------------------- Force constants in Cartesian coordinates: 1 2 3 4 5 1 0.301449D+00 2 -0.206943D+00 0.261212D+00 3 0.000000D+00 0.000000D+00 0.350959D-01 4 0.126056D-01 0.290608D-01 0.000000D+00 0.106035D+00 5 -0.144983D-01 -0.111484D-01 0.000000D+00 0.111810D+00 0.456627D+00 6 0.000000D+00 0.000000D+00 0.398419D-02 0.000000D+00 0.000000D+00 7 -0.293610D+00 0.197957D+00 0.000000D+00 -0.112098D+00 -0.102742D+00 ... Third derivatives in cartesian coordinates (Sum=-8.21871419D-01): K= 1 block: 1 1 -0.228906D-01 K= 2 block: 1 2 1 -0.273269D-01 2 0.773555D-02 0.652875D-01 K= 3 block: 1 2 3 1 -0.150169D-01 2 0.287761D-02 -0.622560D-01 3 -0.378423D-01 0.380952D-01 -0.888412D-02 K= 4 block: 1 2 3 4 1 0.372761D-02 2 0.864463D-02 -0.729166D-02 3 -0.237464D-02 -0.491702D-02 -0.299589D-01 4 0.167156D-01 0.288205D-02 0.488097D-01 0.187014D-01 ----------------------------------------------------------------------------- Thank you. Jing Tao Lu Email: tower.lu%a%gmail.com Postdoc Research Fellow Department of Physics National University of Singapore From owner-chemistry@ccl.net Thu Oct 11 08:49:00 2007 From: "Sina T reli sina.tureli ~~ boun.edu.tr" To: CCL Subject: CCL: Questions Regarding Orbitals/UV-Visible Absorption && CI Spaces Message-Id: <-35369-071011063955-7134-WTQQpNJ4JzBSzJn1pLO+xA|*|server.ccl.net> X-Original-From: "Sina T reli" Date: Thu, 11 Oct 2007 06:39:51 -0400 Sent to CCL by: "Sina T reli" [sina.tureli,boun.edu.tr] I have some questions regarding qualitative reasoning about how to corralate orbital localizations with uv-visible absorption and also size CI spaces to be used in detecting the bands of bacteriochlorophylls. 1. It is known that bacteriochlorophylls usually have qx bands around 700-800nm. And that can be accounted to effect of both conjugation and the jumping of electrons from nitrogen orbitals to the conjugated pi orbitals of the system. How ever still knowing that I am not able to qualitatively reason on how the orbitals localization and size effect the osscilator strength and absorption. Given below are the orbitals that contirbute to the Qx (830nm) of bacteriochlorophyll a. The transition L->H is the dominant one. This calculation is performed on the nonoptimized chlorophyll taken from a pdb with arguslab using (5,5) CI space. The qx absorbance was seen to be about 833 nm. So in short, why would the third transition be the one with the higher oscillator strength or why would these orbitals specifically contirbute to 830nm absorbance, how can be make qualitative reasoning about orbital-absorbance relatedness. Any sources of study as an answer is also welcome. http://img518.imageshack.us/img518/4318/qyyc5.jpg 2. In literature and in some discssions that Mark Thompson (creator of the arguslab program) has participated in this list, it is said that a CI space of (20,20) is optimal for such large molecules... But 20,20 gives about 950 nm while 5,5 gives about 830 nm (which is very close). And also according to the four orbital model (which can also be visiualized by plotting the energies of the orbitals), 4,4 CI space shuld be enough to visiualize the Qx reagion, shouldnt it be? Keep in mind that I performed the calculations only the chlorophyll not the whole complex. 3. Finally, would performing CI calculations on a chlorophyll taken directly from pdb give reliable results or would some kind of optimization of the chlorophyll molecule required to get reliable results? Thanks for your kind answers... From owner-chemistry@ccl.net Thu Oct 11 12:20:00 2007 From: "Igor Avilov avilovi{}averell.umh.ac.be" To: CCL Subject: CCL: Questions Regarding Orbitals/UV-Visible Absorption && CI Spaces Message-Id: <-35370-071011120231-2358-jGEZnnk3xm75zeNV9uBfjA..server.ccl.net> X-Original-From: "Igor Avilov" Content-class: urn:content-classes:message Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="us-ascii" Date: Thu, 11 Oct 2007 18:02:44 +0200 MIME-Version: 1.0 Sent to CCL by: "Igor Avilov" [avilovi::averell.umh.ac.be] Dear Sina, The 4-orbital model is often used to explain the trends in the spectra of tetrapyrrolic compounds. It is based on the fact that in many cases 2 HOMOs and 2 LUMOs are separated from the lower/higher energy orbitals by the considerable energy gaps. Thus, one-electron excitation between these 4 frontier MOs are likely to play a dominant role in the formation of the lowest transitions in the absorption spectrum. But it does not mean that the other one-electron excited configurations cannot mix with them. This admixture is larger, of cause, for the B-states, as they have higher energy than the Q-states. So one must use a sufficiently large basis for CI calculations. On my personal experience, 10 HOMOs x 10 LUMOs is enough for the calculation of the excited states of the porphyrin monomers. If you will use minimal bases like 2 HOMOs x 2 LUMOs (4-orbital model), you could get into trouble even if you are interested only in the several lowest excited states. By the way, I suspect that in this basis you would get the energies of the B-states completely wrong (even if the energies of the Q-states seem to be good). Once I performed such calculations (with 2 x 2 basis) and at the end I had to simply throw away the results. The fact that semi-empirical methods like CNDO/S and INDO/S give the energy of the Q-states lower than the experimental values, it's not a big problem, on my opinion. What is important - is to get the trends right, and here the semi-empirical methods do a very nice job. TD-DFT, for example, also does not assure quantitative agreement with the experiment (look at the paper of D. Sundholm "A density-functional-theory study of bacteriochlorophyll b", PHYSICAL CHEMISTRY CHEMICAL PHYSICS 5 (19): 4265-4271 OCT 1 2003).=20 On my opinion you can use the structure from the pdb-file (containing the experimental crystal structure of the molecule, I suppose), although in solution your molecule may have slightly different geometry. The results should not be dramatically different if you perform geometry optimization first. Anyway, the most important thing is not to get theoretical spectrum very-very close to the experimental one, which could be (or even usually is) just due to the lucky cancellation of errors. I hope I helped you a little bit... Igor Avilov.=20 -----Original Message----- > From: owner-chemistry^^ccl.net [mailto:owner-chemistry^^ccl.net]=20 Sent: jeudi 11 octobre 2007 12:40 To: Igor Avilov Subject: CCL: Questions Regarding Orbitals/UV-Visible Absorption && CI Spaces Sent to CCL by: "Sina T reli" [sina.tureli,boun.edu.tr] I have some questions regarding qualitative reasoning about how to corralate orbital localizations with uv-visible absorption and also size CI spaces to be used in detecting the bands of bacteriochlorophylls. 1. It is known that bacteriochlorophylls usually have qx bands around 700-800nm. And that can be accounted to effect of both conjugation and the jumping of electrons from nitrogen orbitals to the conjugated pi orbitals of the system. How ever still knowing that I am not able to qualitatively reason on how the orbitals localization and size effect the osscilator strength and absorption. Given below are the orbitals that contirbute to the Qx (830nm) of bacteriochlorophyll a. The transition L->H is the dominant one. This calculation is performed on the nonoptimized chlorophyll taken from a pdb with arguslab using (5,5) CI space. The qx absorbance was seen to be about 833 nm. So in short, why would the third transition be the one with the higher oscillator strength or why would these orbitals specifically contirbute to 830nm absorbance, how can be make qualitative reasoning about orbital-absorbance relatedness. Any sources of study as an answer is also welcome. http://img518.imageshack.us/img518/4318/qyyc5.jpg 2. In literature and in some discssions that Mark Thompson (creator of the arguslab program) has participated in this list, it is said that a CI space of (20,20) is optimal for such large molecules... But 20,20 gives about 950 nm while 5,5 gives about 830 nm (which is very close). And also according to the four orbital model (which can also be visiualized by plotting the energies of the orbitals), 4,4 CI space shuld be enough to visiualize the Qx reagion, shouldnt it be? Keep in mind that I performed the calculations only the chlorophyll not the whole complex. 3. Finally, would performing CI calculations on a chlorophyll taken directly from pdb give reliable results or would some kind of optimization of the chlorophyll molecule required to get reliable results? Thanks for your kind answers... -=3D This is automatically added to each message by the mailing script = =3D-http://www.ccl.net/cgi-bin/ccl/send_ccl_messageSubscribe/Unsubscribe:=20http://www.ccl.net/spammers.txt