From owner-chemistry@ccl.net Mon May 7 04:17:00 2012 From: "Kalaivanan Nagarajan kalaichemis(0)gmail.com" To: CCL Subject: CCL: ECD Calculation Message-Id: <-46864-120507041611-31648-npaEEgkIzIjXbLvJfIMzTQ=-=server.ccl.net> X-Original-From: "Kalaivanan Nagarajan" Date: Mon, 7 May 2012 04:16:08 -0400 Sent to CCL by: "Kalaivanan Nagarajan" [kalaichemis::gmail.com] Dear all, I am trying to simulate the ecd spectrum for my bichromophoric molecule. I am using TD-DFT, B3LYP, 6-311G method to do the calculation. Is it the best computational method to do ecd simulation, or is there any better method to do the job. Please help me, i got stuck here... Thanks in advance... From owner-chemistry@ccl.net Mon May 7 10:27:01 2012 From: "=?ISO-8859-1?Q?J=E9r=F4me?= Kieffer jerome.kieffer!^!terre-adelie.org" To: CCL Subject: CCL: Effect of using SSD for scratch Message-Id: <-46865-120506161834-8559-bjC6uj7NpIFy8q5EzADyIw*o*server.ccl.net> X-Original-From: =?ISO-8859-1?Q?J=E9r=F4me?= Kieffer Content-Transfer-Encoding: 8bit Content-Type: text/plain; charset=ISO-8859-1 Date: Sun, 6 May 2012 22:18:24 +0200 Mime-Version: 1.0 Sent to CCL by: =?ISO-8859-1?Q?J=E9r=F4me?= Kieffer [jerome.kieffer*|*terre-adelie.org] On Tue, 24 Apr 2012 17:42:18 -0400 "John McKelvey jmmckel[]gmail.com" wrote: > There is always the issue of cpu speed, > but for large systems disk-io can be a significant issue, even for the > usual scratch file. Has anyone done any direct or reasonable indirect > evaluations around this issue? Hi, I have a couple of comments on that (+ramdisks): -SSD are very fast but especially for fragmented chunks of data. For large chunks the difference in speed is less than 2x. - scratch should be used only if data do not fit in memory (obvious!) ; if data fit in memory it's your operating system's work to do the caching. If your QC program goes via scratch and flushes the disk all the time... it is time to change your QC program. Hence those who clame tricking the system by using ramdisks are just wasting memory. On my linux system with 8G of RAM 4G are used for cache. Cheers, -- Jérôme Kieffer From owner-chemistry@ccl.net Mon May 7 19:51:01 2012 From: "Dr. Lars Goerigk lars.goerigk]=[chem.usyd.edu.au" To: CCL Subject: CCL:G: ECD Calculation Message-Id: <-46866-120507192927-11409-C85t6km/arj/+S/oavk9/g%a%server.ccl.net> X-Original-From: "Dr. Lars Goerigk" Content-Type: multipart/alternative; boundary="------------030709030600090107030603" Date: Tue, 08 May 2012 09:29:12 +1000 MIME-Version: 1.0 Sent to CCL by: "Dr. Lars Goerigk" [lars.goerigk||chem.usyd.edu.au] This is a multi-part message in MIME format. --------------030709030600090107030603 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Hello, the most thorough study about the simulation of TD-DFT to ECD spectra was carried out by Diedrich and Grimme in 2003: J. Phys. Chem. A 2003, 107, 2524. The most accurate functionals for ECD spectra are currently double-hybrid functionals as shown in: J. Phys. Chem. A 2009, 113, 767 Double hybrid functionals for excited states are currently only available in ORCA (which is for free), but only in the so-called Tamm-Dancoff approximation (TDA-DFT), A comparison between TD-double hybrids and TDA-double hybrids is shown in: J. Chem. Phys. 2007, 127, 154116 If you are generally interested in the accuracy of excitation energies of TD-DFT you can also have a look at these works by the Jacquemin or Grimme groups (here are only some examples): JCTC 2009, 5, 2420 PCCP 2009, 11, 4611 J. Chem. Phys. 2010, 132, 184103 JCTC 2011, 7, 3272 Note that the simulation of ECD spectra is more challenging than for UV/Vis spectra because not only the position of the bands, but also their signs and their relative distances to each other are important. A bad description of any of these things can lead to artifical bands or wrong cancellation of bands. Therefore it is crucial not just to compare the excitation energies to the positions of the experimental bands (which is unfortunately still done sometimes), but to actually simulate the spectrum by overlapping Gaussian or Lorentzian functions (as is also done in the papers cited above), and then to compare the simulated with the experimental spectrum! You mention a bichromophoric molecule. If you expect charge-transfer or exciton-coupling, then TD-B3LYP is definitely NOT the best choice. In terms of CT, range-separated functionals work well, but they do not always get the exciton coupling right. In this latter case, double-hybrids can again be recommended: ChemPhysChem 2008, 9, 2467 "Theoretical ECD spectroscopy of large organic and supramolecular systems", in "Comprehensive Chiroptical Spectroscopy" Vol.1, pp.643-673, 2012, Wiley-VCH (in general a very good book about simulation ECD phenomena). You should always compare results for several functionals with each other to rule out any artifacts. These functionals should differ in their amount of Fock-exchange, as this has decisive influence on spurious (artifical) states, which are often found and might influence the analysis of your spectra. I would not recommend GGA or meta-GGA functionals because of that problem. You also have to check if you have several conformers in your system and maybe do an Boltzmann-averaging of the separate ECD spectra for each conformer and combine these to a total, averaged spectrum. Note again that pure B3LYP does not give the correct energetic order of conformers of organic systems, you need definitely dispersion corrections for that. I hope these points could help you. Best wishes, Lars -- Dr. Lars Goerigk School of Chemistry (Building F11) The University of Sydney, NSW 2006 Australia www.researcherid.com/rid/D-3717-2009 --------------030709030600090107030603 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Hello,

the most thorough study about the simulation of TD-DFT to ECD spectra was carried out by Diedrich and Grimme in 2003: 
J. Phys. Chem. A 2003, 107, 2524.

The most accurate functionals for ECD spectra are currently double-hybrid functionals as shown in:
J. Phys. Chem. A 2009, 113, 767

Double hybrid functionals for excited states are currently only available in ORCA (which is for free), but only in the so-called Tamm-Dancoff approximation (TDA-DFT),
A comparison between TD-double hybrids and TDA-double hybrids is shown in:
J. Chem. Phys. 2007, 127, 154116

If you are generally interested in the accuracy of excitation energies of TD-DFT you can also have a look at these works by the Jacquemin or Grimme groups (here are only some examples):
JCTC 2009, 5, 2420
PCCP 2009, 11, 4611
J. Chem. Phys. 2010, 132, 184103
JCTC 2011, 7, 3272


Note that the simulation of ECD spectra is more challenging than for UV/Vis spectra because not only the position of the bands, but also their signs and their relative distances to each other are important. A bad description of any of these things can lead to artifical bands or wrong cancellation of bands. Therefore it is crucial not just to compare the excitation energies to the positions of the experimental bands (which is unfortunately still done sometimes), but to actually simulate the spectrum by overlapping Gaussian or Lorentzian functions (as is also done in the papers cited above), and then to compare the simulated with the experimental spectrum!

You mention a bichromophoric molecule. If you expect charge-transfer or exciton-coupling, then TD-B3LYP is definitely NOT the best choice.
In terms of CT, range-separated functionals work well, but they do not always get the exciton coupling right. In this latter case, double-hybrids can again be recommended:
ChemPhysChem 2008, 9, 2467
"Theoretical ECD spectroscopy of large organic and supramolecular systems", in "Comprehensive Chiroptical Spectroscopy" Vol.1, pp.643-673, 2012, Wiley-VCH (in general a very good book about simulation ECD phenomena).

You should always compare results for several functionals with each other to rule out any artifacts. These functionals should differ in their amount of Fock-exchange, as this has decisive influence on spurious (artifical) states, which are often found and might influence the analysis of your spectra. I would not recommend GGA or meta-GGA functionals because of that problem.
You also have to check if you have several conformers in your system and maybe do an Boltzmann-averaging of the separate ECD spectra for each conformer and combine these to a total, averaged spectrum. Note again that pure B3LYP does not give the correct energetic order of conformers of organic systems, you need definitely dispersion corrections for that.

I hope these points could help you.

Best wishes,
Lars


--
Dr. Lars Goerigk
School of Chemistry (Building F11)
The University of Sydney, NSW 2006
Australia
www.researcherid.com/rid/D-3717-2009
--------------030709030600090107030603-- From owner-chemistry@ccl.net Mon May 7 22:53:00 2012 From: "Kalaivanan kalaichemis * gmail.com" To: CCL Subject: CCL:G: ECD Calculation Message-Id: <-46867-120507225123-2537-pGiIiKfwzWb63gvSFJXC3Q#%#server.ccl.net> X-Original-From: Kalaivanan Content-Type: multipart/alternative; boundary=20cf303b40e3d61dd304bf7d751f Date: Tue, 8 May 2012 08:21:15 +0530 MIME-Version: 1.0 Sent to CCL by: Kalaivanan [kalaichemis-$-gmail.com] --20cf303b40e3d61dd304bf7d751f Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Hi Lars, Thanks a lot for ur detailed answer. All these details are unknown to me. Definitely this will be very helpful to go ahead in my research. I will go through all the literature you suggested. I will contact you, i need of any help... On Tue, May 8, 2012 at 4:59 AM, Dr. Lars Goerigk lars.goerigk]=3D[ chem.usyd.edu.au wrote: > Hello, > > the most thorough study about the simulation of TD-DFT to ECD spectra was > carried out by Diedrich and Grimme in 2003: > J. Phys. Chem. A 2003, 107, 2524. > > The most accurate functionals for ECD spectra are currently double-hybrid > functionals as shown in: > J. Phys. Chem. A 2009, 113, 767 > > Double hybrid functionals for excited states are currently only available > in ORCA (which is for free), but only in the so-called Tamm-Dancoff > approximation (TDA-DFT), > A comparison between TD-double hybrids and TDA-double hybrids is shown in= : > J. Chem. Phys. 2007, 127, 154116 > > If you are generally interested in the accuracy of excitation energies of > TD-DFT you can also have a look at these works by the Jacquemin or Grimme > groups (here are only some examples): > JCTC 2009, 5, 2420 > PCCP 2009, 11, 4611 > J. Chem. Phys. 2010, 132, 184103 > JCTC 2011, 7, 3272 > > > Note that the simulation of ECD spectra is more challenging than for > UV/Vis spectra because not only the position of the bands, but also their > signs and their relative distances to each other are important. A bad > description of any of these things can lead to artifical bands or wrong > cancellation of bands. Therefore it is crucial not just to compare the > excitation energies to the positions of the experimental bands (which is > unfortunately still done sometimes), but to actually simulate the spectru= m > by overlapping Gaussian or Lorentzian functions (as is also done in the > papers cited above), and then to compare the simulated with the > experimental spectrum! > > You mention a bichromophoric molecule. If you expect charge-transfer or > exciton-coupling, then TD-B3LYP is definitely NOT the best choice. > In terms of CT, range-separated functionals work well, but they do not > always get the exciton coupling right. In this latter case, double-hybrid= s > can again be recommended: > ChemPhysChem 2008, 9, 2467 > "Theoretical ECD spectroscopy of large organic and supramolecular > systems", in "Comprehensive Chiroptical Spectroscopy" Vol.1, pp.643-673, > 2012, Wiley-VCH (in general a very good book about simulation ECD > phenomena). > > You should always compare results for several functionals with each other > to rule out any artifacts. These functionals should differ in their amoun= t > of Fock-exchange, as this has decisive influence on spurious (artifical) > states, which are often found and might influence the analysis of your > spectra. I would not recommend GGA or meta-GGA functionals because of tha= t > problem. > You also have to check if you have several conformers in your system and > maybe do an Boltzmann-averaging of the separate ECD spectra for each > conformer and combine these to a total, averaged spectrum. Note again tha= t > pure B3LYP does not give the correct energetic order of conformers of > organic systems, you need definitely dispersion corrections for that. > > I hope these points could help you. > > Best wishes, > Lars > > > -- > Dr. Lars Goerigk > School of Chemistry (Building F11) > The University of Sydney, NSW 2006 > Australia > www.researcherid.com/rid/D-3717-2009 > --=20 Thanks and Regards, Kalaivanan.N, Ph. D Scholar, IISER, Thiruvanandapuram. Contact No : (+91) 9746126684 "The main lesson that I have learned over time is that a good scientist is a humble and listening scientist and not one that is sure 100 percent in what he reads in the textbooks.=E2=80=9D - Dr. Shechtman --20cf303b40e3d61dd304bf7d751f Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Hi Lars,

Thanks a lot for ur detailed answer. All these = details are unknown to me. =C2=A0Definitely this will be very helpful to go= ahead in my research. I will go through all the=C2=A0literature=C2=A0you s= uggested. I will contact you, i need of any help...

On Tue, May 8, 2012 at 4:59 AM, Dr. Lars Goe= rigk lars.goerigk]=3D[chem.usyd.edu.au<= /a> <owner-chemistry|a|ccl.net> wrote:
=20 =20 =20
Hello,

the most thorough study about the simulation of TD-DFT to ECD spectra was carried out by Diedrich and Grimme in 2003:=C2=A0
J. Phys. Chem. A 2003, 107, 2524.

The most accurate functionals for ECD spectra are currently double-hybrid functionals as shown in:
J. Phys. Chem. A 2009, 113, 767

Double hybrid functionals for excited states are currently only available in ORCA (which is for free), but only in the so-called Tamm-Dancoff approximation (TDA-DFT),
A comparison between TD-double hybrids and TDA-double hybrids is shown in:
J. Chem. Phys. 2007, 127, 154116

If you are generally interested in the accuracy of excitation energies of TD-DFT you can also have a look at these works by the Jacquemin or Grimme groups (here are only some examples):
JCTC 2009, 5, 2420
PCCP 2009, 11, 4611
J. Chem. Phys. 2010, 132, 184103
JCTC 2011, 7, 3272


Note that the simulation of ECD spectra is more challenging than for UV/Vis spectra because not only the position of the bands, but also their signs and their relative distances to each other are important. A bad description of any of these things can lead to artifical bands or wrong cancellation of bands. Therefore it is crucial not just to compare the excitation energies to the positions of the experimental bands (which is unfortunately still done sometimes), but to actually simulate the spectrum by overlapping Gaussian or Lorentzian functions (as is also done in the papers cited above), and then to compare the simulated with the experimental spectrum!

You mention a bichromophoric molecule. If you expect charge-transfer or exciton-coupling, then TD-B3LYP is definitely NOT the best choice.
In terms of CT, range-separated functionals work well, but they do not always get the exciton coupling right. In this latter case, double-hybrids can again be recommended:
ChemPhysChem 2008, 9, 2467
"Theoretical ECD spectroscopy of large organic and supramolecular systems", in "Comprehensive Chiroptical Spectroscopy" Vo= l.1, pp.643-673, 2012, Wiley-VCH (in general a very good book about simulation ECD phenomena).

You should always compare results for several functionals with each other to rule out any artifacts. These functionals should differ in their amount of Fock-exchange, as this has decisive influence on spurious (artifical) states, which are often found and might influence the analysis of your spectra. I would not recommend GGA or meta-GGA functionals because of that problem.
You also have to check if you have several conformers in your system and maybe do an Boltzmann-averaging of the separate ECD spectra for each conformer and combine these to a total, averaged spectrum. Note again that pure B3LYP does not give the correct energetic order of conformers of organic systems, you need definitely dispersion corrections for that.

I hope these points could help you.

Best wishes,
Lars


--
Dr. Lars Goerigk
School of Chemistry (Building F11)
The University of Sydney, NSW 2006
Australia
www.researcherid.com/rid/D-3717-2009



--
Thanks and R= egards,

Kalaivanan.N,
Ph. D Scholar,
IISER,=C2=A0
= Thiruvanandapuram.=C2=A0

Contact No : (+91) 9746126684


"The main lesson th= at I have learned over time is that a good scientist=20 is a humble and listening scientist and not one that is sure 100 percent in what he reads in the textbooks.=E2=80=9D =C2=A0 =C2=A0 -<= span style=3D"font-family:comic sans ms,sans-serif"> Dr. Shechtman

--20cf303b40e3d61dd304bf7d751f--