From owner-chemistry@ccl.net Sun May 30 05:58:00 2010 From: "sobereva sobjubao-*-yahoo.com.cn" To: CCL Subject: CCL:G: Laplacian of the total electron density in G09(03) Message-Id: <-42026-100530055438-16181-NDaxASCS/q4W9rW4NM6U1Q[*]server.ccl.net> X-Original-From: sobereva Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset=iso-8859-1 Date: Sun, 30 May 2010 17:54:27 +0800 (CST) MIME-Version: 1.0 Sent to CCL by: sobereva [sobjubao||yahoo.com.cn] Hi, You should use "cube=3D(laplacian,full)" keyword in route section. Without = "full" option, Gaussian deletes core orbital automately. Tian Lu --- On Sat, 5/29/10, Emmanuel Aubert emmanuel.aubert!A!crm2.uhp-nancy.fr wrote: > From: Emmanuel Aubert emmanuel.aubert!A!crm2.uhp-nancy.fr > Subject: CCL:G: Laplacian of the total electron density in G09(03) > To: "Lu, Tian " > Date: Saturday, May 29, 2010, 11:32 PM >=20 > Sent to CCL by: "Emmanuel=A0 Aubert" > [emmanuel.aubert__crm2.uhp-nancy.fr] > Dear all, >=20 > does anyone knows how to get the Laplacian of the total > electron density=20 > with Gaussian09 (or 03) program ? > (I tried to create a cub from the .fchk file, with the > Laplacian keyword=20 > : then there are five record in the cub file, the first is > the density,=20 > the next three are the gradient (I not sure of that) and > the last the=20 > Laplacian, but it seems to me that's only the Laplacian of > the valence=20 > density). >=20 > Best regards, > Emmanuel. >=20 >=20 >=20 > -=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>=20 > E-mail to subscribers: CHEMISTRY.[*].ccl.net > or use: > =A0 =A0 =A0>=20 > E-mail to administrators: CHEMISTRY-REQUEST.[*].ccl.net > or use > =A0 =A0 =A0>=20 > Subscribe/Unsubscribe:=20 > =A0 =A0 =A0>=20>=20 > Job: http://www.ccl.net/jobs=20>=20>=20> =A0 =A0 =A0>=20>=20 >=20 > =0A=0A=0A From owner-chemistry@ccl.net Sun May 30 06:32:00 2010 From: "vandestreek() avmatsim.de" To: CCL Subject: CCL: Crystal Structure vs optimized structure Message-Id: <-42027-100530055511-18848-o5moCMx3dJUWM+94kI7o/g^server.ccl.net> X-Original-From: vandestreek%%avmatsim.de Content-Disposition: inline Content-Transfer-Encoding: 8bit Content-Type: text/plain; charset=ISO-8859-1; DelSp="Yes"; format="flowed" Date: Sun, 30 May 2010 11:55:00 +0200 MIME-Version: 1.0 Sent to CCL by: vandestreek(_)avmatsim.de Quoting "William Flak williamflak__yahoo.com" : > High level of calculation means MP2/6-311++G**, it consumes a lot of > time and I want to test a tens of molecules, also most of them > didn't optimize to the crystal structure.Let me ask you again, May I > do my calculation on the crystal structure without optimization? > Could you support me with a literature done this approach, please? 1. Minimising the crystal structure with the energy potential that you intend to use for the energy calculations also serves the purpose of checking if the energy potential is appropriate for your system. Structures (atomic coordinates + unit-cell parameters) should be easier to reproduce than energies. If your potential cannot even reproduce your structure, how reliable are your energies going to be anyway? (I'm assuming you are using MP2/6-311++G** to energy-minimise your crystal structure and to calculate your dimer interaction energies.) 2. There is an important reason to optimise crystal structures before calculating interaction energies. In the crystal, your dimer is surrounded by 10^23 molecules at distances greater than the Van der Waals distance. The interaction energies between these 10^23 molecules are slightly attractive. As a result, they exert a nett inward force on your dimer. Because forces must balance in order for your crystal structure to be in equilibrium, the nett force between the two molecules in your dimer must be repulsive (the energy can still be negative, so the interaction can still be stabilising; I am assuming that the molecules in your dimer are nearest neighbours in your crystal). Because the repulsion between two molecules tends to be exponential, minor changes to your atomic coordinates should lead to substantial changes in your energies (I have never seen this confirmed, but I have never seen it refuted either). Because quantum-mechanical methods tend to underestimate the attractive Van der Waals contribution (the main attractive contribution in molecular crystals), your crystals probably want to expand slightly in your energy potential (I don't have experience with MP2, perhaps you're OK), which means that if you use the experimental crystal structure without optimising, all your interactions will be very repulsive (not a problem, they should be repulsive), and may have positive energies (that may point to a problem). That your energies will be very sensitive to small changes to the intermolecular distances is also a problem if you want to compare the interaction energies of several systems and if some crystal structures were measured at 120 K whereas others were measured at room temperature: the thermal expansion changes your interaction energies, presumably with some exponential dependency. 3. What does "didn't optimize to the crystal structure" mean? What happened? 4. For large systems (and crystals are infinite), the polarisation functions can lead to numerical instabilities. If "didn't optimize to the crystal structure" means that some of the calculations did not finish/converge, I'd try again without the polarisation functions. But the polarisation functions may be necessary if you are working on molecules with formal charges (molecular salts or zwitterions). 5. The poorly determined hydrogen atoms can lead to very large initial forces, which can upset your minimisation. You may have ended up in a different minimum in your energy potential. First energy-minimise the hydrogen atoms with the unit-cell parameters and the positions of all non-hydrogen atoms kept fixed, then relax all atoms but keep the unit cell fixed, then energy-minimise everything together. 6. Which elements are present in your crystal structures? Where do your crystal structures come from? Are they room-temperature X-ray structures, low-temperature neutron structures, structures determined > from powder diffraction data? What are their R-values? What is their quality (which journal were they published in, what year?)? Are you sure all hydrogen atoms have been added correctly? 7. You are asking for a literature reference where people have done the same, presumably so that you can refer to it in an attempt to get your troublesome calculations past some refereeing process. Science is not about finding justifications for dodgy calculations, it is about finding reliable alternatives for dodgy calculations. But to answer your question proper, my guess is that the following paper, published this month in one of the best crystallographic journals, does what you propose to do: Acta Cryst. (2010). B66, 396-406 [ doi:10.1107/S0108768110008074 ] "The lines-of-force landscape of interactions between molecules in crystals; cohesive versus tolerant and `collateral damage' contact" A. Gavezzotti I am not sure whether or not the crystal structures are energy-minimised first in this paper (working in industry, I cannot read papers unless I buy them), but it is my guess that they are not. 8. You should at the very least energy-optimise the hydrogen atoms. They have very low X-ray scattering factors and their coordinates are therefore generally not determined as well as the non-hydrogen atoms. That is one reason to energy-minimise their positions before calculating energies. The second reason is that because hydrogen atoms are monovalent, they cover the surface of the molecule, they therefore determine the close contacts, and when you energy-optimise their positions, you will have solved most of the problems I mentioned under point 2. Best wishes, -- Dr Jacco van de Streek Senior Scientist Avant-garde Materials Simulation Freiburg im Breisgau, Germany From owner-chemistry@ccl.net Sun May 30 07:42:01 2010 From: "Vincent Leroux vincent.leroux * loria.fr" To: CCL Subject: CCL: Crystal Structure vs optimized structure Message-Id: <-42028-100530072315-28837-b1pUvKohaEuN4Nq+ymXzng[*]server.ccl.net> X-Original-From: Vincent Leroux Content-Transfer-Encoding: 8bit Content-Type: text/plain; charset=ISO-8859-1 Date: Sun, 30 May 2010 13:23:05 +0200 MIME-Version: 1.0 Sent to CCL by: Vincent Leroux [vincent.leroux=-=loria.fr] Hi William, It depends on what you want to analyze. If you just use the crystal coordinates without any kind of molecular mechanics processing then, as was already said here, you might encounter high energy terms due to local differences between the energy function used for refining the crystal structure and the one you will be using yourself. But if your goal is to simply analyze your system *qualitatively*, you should be more interested into the nature and repartition of the interactions between your two molecules than into the total free energy of binding. Then in such a situation the possibility of a limited number of local forcefield-dependent artifacts is not so much a problem and you could almost use directly coordinates derived from X-ray structures, provided those are of good quality of course. A good example is from my last paper: http://dx.doi.org/10.1111/j.1747-0285.2009.00895.x This paper identifies binding modes by computing the repartition of CHARMM protein/ligand interaction energies (non-bonded terms of the FF, simply Coulomb+vdW) from crystal structures. The structures were completed in order to get defined in the forcefield (protonation, addition of missing residues by homology modeling) and then only slightly minimized. This is by no means something I would call "high level optimization" (and this was on purpose as to remain as close as possible to the starting X-ray data). Upon looking at the pre-minimization details (data not shown in the paper), there were indeed clues of the presence of a limited number of local high-energy terms in the interaction energies distribution according to CHARMM. This might originate from incompatibilities between CHARMM and the X-ray refinement process, but also from the protonation process making wild guesses upon adding hydrogens at the protein/ligand interface. In any case the slight minimization appeared to correct that, resulting into a far-from-perfect representation of enthalpy, but enough for its distribution to make sense in the context of identifying ligand binding modes. Regards, VL Le 28/05/10 13:25, William Flak williamflak---yahoo.com a écrit : > Sent to CCL by: "William Flak" [williamflak]^[yahoo.com] > Dear CCL > I was wondering if I can calculate an interaction properties between two > molecules based on crystal structure without optimization? Optimization with > high level couldn't give me the crystal structure. > Any comment would be appreciated. > W. Flak > > > From owner-chemistry@ccl.net Sun May 30 11:44:01 2010 From: "Jim Kress ccl_nospam#kressworks.com" To: CCL Subject: CCL:G: Crystal Structure vs optimized structure Message-Id: <-42029-100530102751-28722-H0tu+uaeHM0I1qxhCdz/Pw|server.ccl.net> X-Original-From: "Jim Kress" Content-Language: en-us Content-Type: multipart/alternative; boundary="----=_NextPart_000_0001_01CAFFE2.B5494490" Date: Sun, 30 May 2010 10:27:29 -0400 MIME-Version: 1.0 Sent to CCL by: "Jim Kress" [ccl_nospam]_[kressworks.com] This is a multi-part message in MIME format. ------=_NextPart_000_0001_01CAFFE2.B5494490 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Why don't you use code specifically designed for crystalline systems, like cp2k, CPMD, or Siesta (Quantum-Espresso)? Even Gaussian (ugh!) has a periodic system capability built in as do a number of other QM packages like NWCHEM. Look at this list as a starter: http://en.wikipedia.org/wiki/List_of_quantum_chemistry_and_solid_state_physi cs_software Many (if not most) of these packages will allow mixed plane-wave and LCAO basis sets which will allow you to perform the LCAO calculation at the points of interest on your crystal while making sure the rest of the system is properly represented using plane-waves. Most also allow integration with Molecular Mechanics/ Molecular Dynamics programs (like gromacs) that allow the simulation and analysis of arbitrarily large systems (i.e. via QM/MM). These software packages and methodologies have met with continual success, as can be seen from a relevant literature search. As an aside, I have performed many calculations comparing 6-311++G** basis sets with a number of others. IMHO, there are a number of basis sets that are vastly superior to (without being more expensive than) the 6-311++G** basis sets. Dunning's correlation consistent basis sets are among the best, both for HF, MP2, CI, and MCSCF calculations. Their DFT optimized variants are also far superior to 6-311++G** for any kind of DFT calculations. I would suggest you explore them since, if I we refereeing any paper that used 6-311++G** basis sets, I would immediately criticize it for that reason. Good luck! Jim > From: owner-chemistry+ccl_nospam==kressworks.com.|.ccl.net [mailto:owner-chemistry+ccl_nospam==kressworks.com.|.ccl.net] On Behalf Of William Flak williamflak__yahoo.com Sent: Saturday, May 29, 2010 3:03 PM To: Kress, Jim Subject: CCL: Crystal Structure vs optimized structure Dear Dr. Radoslaw Thanks for your reply. I read your comment dated 2008. High level of calculation means MP2/6-311++G**, it consumes a lot of time and I want to test a tens of molecules, also most of them didn't optimize to the crystal structure. Let me ask you again, May I do my calculation on the crystal structure without optimization? Could you support me with a literature done this approach, please? Thanks in advance. W. Flak --- On Fri, 5/28/10, Radoslaw Kaminski rkaminski.rk-#-gmail.com wrote: > From: Radoslaw Kaminski rkaminski.rk-#-gmail.com Subject: CCL: Crystal Structure vs optimized structure To: "Flak, William " Date: Friday, May 28, 2010, 12:53 PM Hi William, I don't really get the point why are you worried. Of course you can do it. However, you have to remember that such approach will lead to neglecting of intermolacular interactions, and crystal field in general. If you look in archive I was commenting on this some time ago (a year or so). The think I could suggest to do is to take the best possible coordinates from the crystal (preferably from high-resolution measurement), keep the heavy atoms constant and optimize hydrogen atom positions. Also, if you look at the literature these could lead to different interaction energies (because the system is relaxed or not etc.). Always take care while analysing the results! Short question: what does it mean "high level" for you? Hope this helps somehow. Cheers, Radek 2010/5/28 William Flak williamflak---yahoo.com Sent to CCL by: "William Flak" [williamflak]^[yahoo.com] Dear CCL I was wondering if I can calculate an interaction properties between two molecules based on crystal structure without optimization? Optimization with high level couldn't give me the crystal structure. Any comment would be appreciated. W. FlakE-mail to subscribers: CHEMISTRY#ccl.net or use:E-mail to administrators: CHEMISTRY-REQUEST#ccl.net or usehttp://www.ccl.net/chemistry/sub_unsub.shtmlhttp://www.ccl.net/spammers.txt------=_NextPart_000_0001_01CAFFE2.B5494490 Content-Type: text/html; charset="us-ascii" Content-Transfer-Encoding: quoted-printable

Why don’t you use code specifically designed for crystalline systems, like cp2k, CPMD, or Siesta = (Quantum-Espresso)?  Even Gaussian (ugh!) has a periodic system capability built in as do a number = of other QM packages like NWCHEM.  Look at this list as a = starter:

 

http://en.wikipedia.org/wiki/List_of_quantum_chem= istry_and_solid_state_physics_software

 

Many (if not most) of these packages will allow mixed = plane-wave and LCAO basis sets which will allow you to perform the LCAO calculation = at the points of interest on your crystal while making sure the rest of the = system is properly represented using plane-waves.  Most also allow = integration with Molecular Mechanics/ Molecular Dynamics programs (like gromacs) that = allow the simulation and analysis of arbitrarily large systems (i.e. via = QM/MM).  These software packages and methodologies have met with continual = success, as can be seen from a relevant literature search.

 

As an aside, I have performed many calculations comparing = 6-311++G** basis sets with a number of others.  IMHO, there are a number of = basis sets that are vastly superior to (without being more expensive than) the 6-311++G** basis sets.  Dunning’s correlation consistent = basis sets are among the best, both for HF, MP2, CI, and MCSCF calculations.  = Their DFT optimized variants are also far superior to 6-311++G** for any kind = of DFT calculations.  I would suggest you explore them since, if I we = refereeing any paper that used 6-311++G** basis sets, I would immediately criticize = it for that reason.

 

Good luck!

 

Jim

 

From:= = owner-chemistry+ccl_nospam=3D=3Dkressworks.com.|.ccl.net [mailto:owner-chemistry+ccl_nospam=3D=3Dkressworks.com.|.ccl.net] On = Behalf Of William Flak williamflak__yahoo.com
Sent: Saturday, May 29, 2010 3:03 PM
To: Kress, Jim
Subject: CCL: Crystal Structure vs optimized = structure

 

Dear Dr. Radoslaw

Thanks for your reply. I read your comment dated = 2008.

High level of calculation means MP2/6-311++G**, = it consumes a lot of time and I want to test a tens of molecules, also = most of them didn't optimize to the crystal structure.

Let me ask you again, May I do my calculation on = the crystal structure without optimization? Could you support me = with a literature done this approach, please?

Thanks in advance.

W. Flak

--- On Fri, 5/28/10, Radoslaw Kaminski rkaminski.rk-#-gmail.com = <owner-chemistry . ccl.net> wrote:


From: Radoslaw Kaminski rkaminski.rk-#-gmail.com <owner-chemistry . ccl.net>
Subject: CCL: Crystal Structure vs optimized structure
To: "Flak, William " <williamflak . yahoo.com>
Date: Friday, May 28, 2010, 12:53 PM

Hi William,

I don't really get the point why are you worried. Of course you can do = it. However, you have to remember that such approach will lead to = neglecting of intermolacular interactions, and crystal field in general. If you look = in archive I was commenting on this some time ago (a year or so). The = think I could suggest to do is to take the best possible coordinates from the = crystal (preferably from high-resolution measurement), keep the heavy atoms = constant and optimize hydrogen atom positions. Also, if you look at the = literature these could lead to different interaction energies (because the system = is relaxed or not etc.). Always take care while analysing the = results!

Short question: what does it mean "high level" for you?

Hope this helps somehow.

Cheers,

Radek

2010/5/28 William Flak williamflak---yahoo.com <owner-chemistry#ccl.net>


Sent to CCL by: "William  Flak" [williamflak]^[yahoo.com]
Dear CCL
I was wondering if I can calculate an interaction properties between = two
molecules based on crystal structure without optimization? = Optimization with
high level couldn't give me the crystal structure.
Any comment would be appreciated.
W. Flak



-=3D This is automatically added to each message by the mailing script = =3D-

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
=
Subscribe/Unsubscribe:
     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/conferenc= es/

Search Messages: http://www.ccl.net/chemistry/searchccl/index.shtml<= br>
     http://www.ccl.net/spammers.txt

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

 =

------=_NextPart_000_0001_01CAFFE2.B5494490-- From owner-chemistry@ccl.net Sun May 30 12:18:00 2010 From: "ros rodrigogalindo(a)gmail.com" To: CCL Subject: CCL:G: Laplacian of the total electron density in G09(03) Message-Id: <-42030-100530114538-18468-gsdV3XHebrPatZRHza6LGw,+,server.ccl.net> X-Original-From: ros Content-Type: multipart/alternative; boundary=001485e2ebde1937b10487d1a1d7 Date: Sun, 30 May 2010 10:45:27 -0500 MIME-Version: 1.0 Sent to CCL by: ros [rodrigogalindo.:.gmail.com] --001485e2ebde1937b10487d1a1d7 Content-Type: text/plain; charset=ISO-8859-1 Hello! You can use the AIMALL software ( http://aim.tkgristmill.com/ ) to calculate the electron density (rho) and the laplacian (delta-rho) from an .fchk file. The program can give you quite a lot information regarding the electron density and you can visualize it too. Other method is using the ext94b from the original Bader AIM site ( http://www.chemistry.mcmaster.ca/aimpac/ ). Its more work to use it but its free. Hope this helps! Regards, Rodrigo. 2010/5/29 Emmanuel Aubert emmanuel.aubert!A!crm2.uhp-nancy.fr < owner-chemistry%%ccl.net> > > Sent to CCL by: "Emmanuel Aubert" [emmanuel.aubert__crm2.uhp-nancy.fr] > Dear all, > > does anyone knows how to get the Laplacian of the total electron density > with Gaussian09 (or 03) program ? > (I tried to create a cub from the .fchk file, with the Laplacian keyword > : then there are five record in the cub file, the first is the density, > the next three are the gradient (I not sure of that) and the last the > Laplacian, but it seems to me that's only the Laplacian of the valence > density). > > Best regards, > Emmanuel.> > > --001485e2ebde1937b10487d1a1d7 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Hello!

You can use the AIMALL software ( http://aim.tkgristmill.com/ ) to calculate the electron de= nsity (rho) and the laplacian (delta-rho) from an .fchk file.=A0 The progra= m can give you quite a lot information regarding the electron density and y= ou can visualize it too.

Other method is using the ext94b from the original Bader AIM site ( http://www.chemistry.mcma= ster.ca/aimpac/ ).=A0 Its more work to use it but its free.

Hope= this helps!

Regards,

Rodrigo.

2010/5/29 Em= manuel Aubert emmanuel.aubert!A!crm2.u= hp-nancy.fr <owner-chemistry%%ccl.net>

Sent to CCL by: "Emmanuel =A0Aubert" [emmanuel.aubert__crm2.uhp-nanc= y.fr]
Dear all,

does anyone knows how to get the Laplacian of the total electron density with Gaussian09 (or 03) program ?
(I tried to create a cub from the .fchk file, with the Laplacian keyword : then there are five record in the cub file, the first is the density,
the next three are the gradient (I not sure of that) and the last the
Laplacian, but it seems to me that's only the Laplacian of the valence<= br> density).

Best regards,
Emmanuel.



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

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

Subscribe/Unsubscribe:
=A0 =A0 =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
=A0 =A0 =A0h= ttp://www.ccl.net/spammers.txt

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



--001485e2ebde1937b10487d1a1d7--