Menu
▾
▴
nlopredict-users — Nonlinear optical prediction tool
You can subscribe to this list here.
| 2008 |
Jan
|
Feb
|
Mar
(1) |
Apr
(9) |
May
(4) |
Jun
(1) |
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2009 |
Jan
(4) |
Feb
(1) |
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
(1) |
Nov
|
Dec
|
| 2021 |
Jan
(1) |
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|
From: Kerry J. W. A. <ker...@ut...> - 2021-01-14 06:03:26
|
Dear I am looking for some manual for NLOPredict. I have access to Gaussian09 and Gaussia16, but I don't know if NLOpredicts can use these output files. For any information please let me know Kind regards K Wrighton-Araneda Dr. Kerry Jeffry Wrighton Araneda Investigador Postdoctoral Programa Institucional de Fomento a la I+D+i (PIDi) Universidad Tecnológica Metropolitana Fono (+56 9) 8420 4894 Perfiles: <https://orcid.org/0000-0003-4872-8976> <https://publons.com/researcher/3009816/kerry-jeffry-wrighton-araneda/> <https://scholar.google.com/citations?user=vY5GTB8AAAAJ&hl=en> <https://www.researchgate.net/profile/Kerry_Wrighton-Araneda> |
|
From: 伟奇 李 <lw...@ya...> - 2009-10-23 02:29:10
|
Hello! Dear Prof:
I am a theoretic computional doctor. I am interesting macropic NLO properties of materials . As in your web said that NLOPredict gives a rapid means of predicting molecular orientation at surfaces using the second nonlinear optical coefficient and experiment result.
But if I calculate second nonlinear optical coefficient of single and structure of polymer crystal, can I predict macropic NLO properties of crystal using NLOpredict? If answer is yes, how I acheive it? Thank you.
Your sincerely
WeiQi
Ocb 21 2009
___________________________________________________________
好玩贺卡等你发，邮箱贺卡全新上线！
http://card.mail.cn.yahoo.com/ |
|
From: Randy H. <he...@in...> - 2009-02-12 02:09:46
|
Dear All, For now, I've put up quick & dirty instructions here: http://poincare.uits.iupui.edu/~heiland/nlopredict/ Warning - I've only tested on OSX (Intel/10.5.6) so far and will try to test on Linux & Windows (XP anyway) in the coming days. I know there are still some layout issues in the plots that aren't perfect, but what I've seen is quite close to the "old days". Let me know of problems and sorry it's taken so long... A big thanks to the Chimera developers! -Randy |
|
From: Randy H. <he...@in...> - 2009-01-17 13:02:17
|
Kramer, Just a brief update... thanks to the hard work of the Chimera team, we should (hopefully) be very close to an OSX (daily) build that will support NLOPredict. Basically, what they've done is (through our requests and maybe others) provide access to matplotlib, a dependency for NLOPredict. We still have a problem to resolve, but when it's ready to go, I was wondering if you'd mind being a guinea pig and do some testing? -Randy On Jan 13, 2009, at 9:10 AM, Kramer Campen wrote: > Dear Randy: > > I (and at least one other colleague at my institution) would like it > on > OS X. I don't know if the issues are relevant to the particular build > of OS X (either through x windows or native) but we would obviously be > happy to do either. > > Kramer > > > > Randy Heiland wrote: >> Hi Kramer, >> >> Unfortunately, the answer, in general, is no. I've been >> talking/working w/ the Chimera dev team about some of the issues and >> they're not resolved yet. What platform do you use - Windows, Linux, >> or OSX? We can try to make your port a priority. >> >> -Randy >> >> On Jan 13, 2009, at 7:23 AM, Kramer Campen wrote: >> >>> Dear all: >>> >>> I wonder if there is a version of NLOPredict floating around that >>> would >>> work with the current version of Chimera (1.3). It's proving to >>> be a >>> little tricky to track down older versions. >>> >>> Thanks in advance for any suggestions. >>> >>> Kramer >>> >>> -- >>> Biosurface Spectroscopy Group >>> FOM Institute for Atomic and Molecular Physics [AMOLF] >>> 407 Kruislaan >>> 1098 SJ Amsterdam >>> The Netherlands >>> >>> >>> ------------------------------------------------------------------------------ >>> >>> This SF.net email is sponsored by: >>> SourcForge Community >>> SourceForge wants to tell your story. >>> http://p.sf.net/sfu/sf-spreadtheword >>> _______________________________________________ >>> Nlopredict-users mailing list >>> Nlo...@li... >>> https://lists.sourceforge.net/lists/listinfo/nlopredict-users > > -- > Biosurface Spectroscopy Group > FOM Institute for Atomic and Molecular Physics [AMOLF] > 407 Kruislaan > 1098 SJ Amsterdam > The Netherlands > |
|
From: Kramer C. <ca...@am...> - 2009-01-13 14:10:22
|
Dear Randy: I (and at least one other colleague at my institution) would like it on OS X. I don't know if the issues are relevant to the particular build of OS X (either through x windows or native) but we would obviously be happy to do either. Kramer Randy Heiland wrote: > Hi Kramer, > > Unfortunately, the answer, in general, is no. I've been > talking/working w/ the Chimera dev team about some of the issues and > they're not resolved yet. What platform do you use - Windows, Linux, > or OSX? We can try to make your port a priority. > > -Randy > > On Jan 13, 2009, at 7:23 AM, Kramer Campen wrote: > >> Dear all: >> >> I wonder if there is a version of NLOPredict floating around that would >> work with the current version of Chimera (1.3). It's proving to be a >> little tricky to track down older versions. >> >> Thanks in advance for any suggestions. >> >> Kramer >> >> -- >> Biosurface Spectroscopy Group >> FOM Institute for Atomic and Molecular Physics [AMOLF] >> 407 Kruislaan >> 1098 SJ Amsterdam >> The Netherlands >> >> >> ------------------------------------------------------------------------------ >> >> This SF.net email is sponsored by: >> SourcForge Community >> SourceForge wants to tell your story. >> http://p.sf.net/sfu/sf-spreadtheword >> _______________________________________________ >> Nlopredict-users mailing list >> Nlo...@li... >> https://lists.sourceforge.net/lists/listinfo/nlopredict-users -- Biosurface Spectroscopy Group FOM Institute for Atomic and Molecular Physics [AMOLF] 407 Kruislaan 1098 SJ Amsterdam The Netherlands |
|
From: Randy H. <he...@in...> - 2009-01-13 13:40:14
|
Hi Kramer, Unfortunately, the answer, in general, is no. I've been talking/ working w/ the Chimera dev team about some of the issues and they're not resolved yet. What platform do you use - Windows, Linux, or OSX? We can try to make your port a priority. -Randy On Jan 13, 2009, at 7:23 AM, Kramer Campen wrote: > Dear all: > > I wonder if there is a version of NLOPredict floating around that > would > work with the current version of Chimera (1.3). It's proving to be a > little tricky to track down older versions. > > Thanks in advance for any suggestions. > > Kramer > > -- > Biosurface Spectroscopy Group > FOM Institute for Atomic and Molecular Physics [AMOLF] > 407 Kruislaan > 1098 SJ Amsterdam > The Netherlands > > > ------------------------------------------------------------------------------ > This SF.net email is sponsored by: > SourcForge Community > SourceForge wants to tell your story. > http://p.sf.net/sfu/sf-spreadtheword > _______________________________________________ > Nlopredict-users mailing list > Nlo...@li... > https://lists.sourceforge.net/lists/listinfo/nlopredict-users |
|
From: Kramer C. <ca...@am...> - 2009-01-13 12:44:13
|
Dear all: I wonder if there is a version of NLOPredict floating around that would work with the current version of Chimera (1.3). It's proving to be a little tricky to track down older versions. Thanks in advance for any suggestions. Kramer -- Biosurface Spectroscopy Group FOM Institute for Atomic and Molecular Physics [AMOLF] 407 Kruislaan 1098 SJ Amsterdam The Netherlands |
|
From: Antonio De C. <ant...@ut...> - 2008-06-24 17:44:55
|
Hello All, Does UCSF Chimera/NLO Predict support PC GAMESS output files? I notice there is no file association type for PC GAMESS. Alternatively, I could convert the GAMESS output to xyx coordinates and open that in Chimera/NLO Predict. Thanks, Anthony |
|
From: Antonio De C. <ant...@ut...> - 2008-05-28 23:08:56
|
Hello, When viewing a single molecule in NLO predict, the Z axis is point towards the viewer, but which direction does the light (laser) source project? Coming out towards the viewer or going into the plane (screen)? Thanks, Anthony |
|
From: Antonio De C. <ant...@ut...> - 2008-05-22 16:16:33
|
Hi I was wondering if you can refer some papers that explain how the sum-over-states 'method' can be used to calculate THG from computations. With best regards, Anthony University of Toronto |
|
From: Randy H. <he...@in...> - 2008-05-13 15:21:56
|
Just forwarding for the archive... Begin forwarded message: > From: "Antonio De Crisci" <ant...@ut...> > Date: November 29, 2007 2:03:30 PM EST > To: "Simpson, Garth J" <gsi...@pu...>, "Andrew Moad" > <and...@ni...>, <he...@in...>, <rwa...@pu...> > Cc: "me2" <ant...@ut...> > Subject: Re: NLOPredict / Gaussian question > > > Your exactly right, Prof. Simpson. > > See this new output file attached. This has many input frequencies > (shown below) and you will see beta(w,w,-2w) along with beta(-w;w, > 0) in the output: > > Using perturbation frequencies: 0.000000 1.000000 1.000000 > 1.000000 1.000000 > Using perturbation frequencies: 1.000000 1.000000 1.000000 > 2.000000 2.000000 > Using perturbation frequencies: 2.000000 3.000000 3.000000 > 4.000000 4.000000 > Using perturbation frequencies: 5.000000 5.000000 6.000000 > 6.000000 7.000000 > Using perturbation frequencies: 7.000000 8.000000 8.000000 > > Im not exactly sure why there are so many replicate frequencies (i > didn't input them), but I suspect it may just be a syntax error, > i.e. I didn't leave an empty line in the input file and it took > some of the molecular coordinates as a frequency inputs. > Nonetheless, the output generated SHG tensors, along with Pockels > effect tensors.as shown in the attached out file. > > Thanks for your insight! > > Cheers, > Anthony > > > > > ----- Original Message ----- From: "Simpson, Garth J" > <gsi...@pu...> > To: "Antonio De Crisci" <ant...@ut...>; "Andrew > Moad" <and...@ni...>; <he...@in...>; > <rwa...@pu...> > Sent: Thursday, November 29, 2007 1:25 PM > Subject: RE: NLOPredict / Gaussian question > > >> Anthony, >> >> From the output file, it looks like the beta you are calculating >> (whether you intend it or not) is for the Pockels effect, as it is >> given by B(-w;w,0). This tensor describes the change in the linear >> polarizability as a function of an applied external field. The >> values in parentheses correspond to the exigent frequency (first >> value) and the two incident frequencies (second and third values). >> Since the last frequency is zero, it indicates a DC electric >> field. Although I have not used the algorithm you describe for >> calculating hyperpolarizabilities, it may be a simple matter of >> including more than one frequency in the input. >> >> Hope this helps, >> >> Garth >> >> -----Original Message----- >> From: Antonio De Crisci [mailto:ant...@ut...] >> Sent: Thursday, November 29, 2007 11:47 AM >> To: Andrew Moad; he...@in...; rwa...@pu...; >> Simpson, Garth J >> Cc: me2 >> Subject: Re: NLOPredict / Gaussian question >> >> Thanks you all for your replies! You have clarified the my >> situation. Thanks for the ACR paper, this is something that I will >> definitely be digging into. >> >> But in terms of my calculation, Andrew, I will have to try your >> command line for my Gaussian calculations. >> >> Just to let you know what I have been setting up my calculations, >> attached is the out file of one of my calculations (urea). I used >> keywords polar and CPHF=RdFreq and set a frequency of 1064nm >> (there is a line in the out that >> says "Using perturbation frequencies: 0.000000 0.042823" in >> the out >> file) because i want to calculate the frequency-dependent >> polarizabilities and hyperpolarizabilities, as a in the presence >> of a 1064nm laser for the tensors in NLOPredict. >> >> Gaussian out file does give me mu, alpha and beta at TWO >> frequencies (0 and >> 1064nm) but the beta output was confusing me at first. >> >> Please see my Gaussian out file (txt format) of urea.What do you >> think? >> >> Sorry if some of the questions seem trivial, im actually an >> organometallic chemistry and just recently have become familiar >> with NLO properties of some of the materials I have made and would >> like to compliment the results with computations and NLOPredict >> for publication. >> >> Thanks, >> Anthony >> >> >> >> ----- Original Message ----- >> From: "Andrew Moad" <and...@ni...> >> To: <ant...@ut...> >> Sent: Thursday, November 29, 2007 11:09 AM >> Subject: NLOPredict / Gaussian question >> >> >>> Dear Anthony, >>> >>> The answer to your question is yes and it depends. If you are >>> talking >>> about electronic transitions (such as those common for second >>> harmonic >>> generation) the answer is yes. We use Gaussian to calculate >>> transition >>> dipoles between ~100 excited states and perform a sum-over-states >>> (outside >>> of Gaussian) calculation. Here we can indeed calculate all 27 tensor >>> elements. If you are talking about vibrational transitions then >>> the story >>> is a little more complicated. Gaussian will give you a “symmetric” >>> transition polarizability for each vibrational transition, but >>> you are >>> right that it contains only the unique tensor elements. Internally >>> Gaussian is applying finite fields and doing derivatives to do this >>> calculation. To get all vibrational tensor elements you will have >>> to apply >>> the finite fields yourself and do the derivatives yourself (as >>> far as I >>> know). It’s not that bad. I started to write a program to >>> automate this, >>> but decided to just use the symmetric tensors instead. Actually, >>> we like >>> to use the b = am approximation which is valid if you are near >>> resonance >>> with a real transition in the molecule. See my J. Phys. Chem. >>> papers for a >>> detailed explanation. This simplifies computation because you >>> just need to >>> calculate a 9 element a tensor and a one transition dipole m and >>> then do a >>> direct product between them to get a 27 element tensor (there are >>> some >>> equalities so you would not have 27 unique tensor elements here). >>> >>> There is more about Gaussian that I don’t know than I know, but >>> it does >>> concern me that I have never seen CPHF=ReFreq. We usually do a >>> different >>> calculation to get the vibrational transition polarizabilities. I >>> usually >>> use this: >>> >>> #p freq=raman rb3lyp/6-311++g(d,p) int=ultrafine iop(7/33=1) polar >>> >>> to get the transition a and m for each vibration. What is the >>> full command >>> line (line the one above) that you are using? >>> >>> I’m sorry that I could not give you the answer that you wanted. Try >>> emailing Gaussian with one very specific question. They are >>> pretty good at >>> answering those. >>> >>> Andy >>> >>> =================================================== >>> | Im a grad student at the University of Toronto investigating >>> the NLO >>> properties of single molecules. I came across your fascinating >>> | paper J. Comp. Chem. 2007, 28, 1996 using NLOPredict to >>> visualize the >>> process. In your computations (using G98), were you >>> | able to calculate all 27 beta tensors? Im running into an issue >>> were >>> using Gaussian 03 to calculate the beta tensors, but the output file >>> | only contains 18 tensors using the keywords 'polar' and >>> 'CPHF=ReFreq' at >>> 1064nm. >>> | >>> | Would you be able to tell me how you found all 27 tensors from >>> calculation? >>> =================================================== >>> >>> -- >>> Dr. Andrew J. Moad >>> Research Chemist >>> National Institute of Standards and Technology >>> Surface and Microanalysis Science Division >>> 100 Bureau Drive, Stop 8372 >>> Gaithersburg, MD 20899-8372 >>> http://www.nist.gov >>> Tel: (301) 975-8532 >>> Fax: (301) 417-1321 >>> Email: and...@ni... >>> >>> >>> >>> > > > [The attachment urea_hyperpolarizbility.out has been manually removed] > |
|
From: Antonio De C. <ant...@ut...> - 2008-05-08 02:40:43
|
-------------------------------------------------- From: <he...@ga...> Sent: Tuesday, January 01, 2008 10:38 PM To: "Antonio De Crisci" <ant...@ut...> Subject: Re: beta tensors (hyperpolarizibility) from a polar CPHF=RdFreq calculation > > Dr. De Crisci, > > My apologies for the delay. > > This is a linear representation of the matrix hyperpolarizability > matrix. Due to the presence of the external field it is no longer > symmetric along the third axis so the first 6 elements correspond to > elements (i,j,x) where (i,j) range over (x,x), (x,y), (y,y), (x,z), > (y,z) and (z,z). Then then next 6 have y in the third index and the > last 6 have z. This is different from the static cases, W=0, where > permutation symmetry can be applied to all three coordinates. > > So element 4 is beta(x,z,x), element 11 is beta (y,z,y) and element > 18 is beta(z,z,z). > > > > On Thu, Nov 29, 2007 at 02:57:39PM -0500, Antonio De Crisci >> Below is the result of your feedback form. It was submitted by >> Antonio De Crisci (ant...@ut...) on Thursday, >> November 29, 2007 at 13:43:58 >> --------------------------------------------------------------------------- >> >> >> message: >> Hi, im trying to calculate the beta tensors (hyperpolarizibility >> calculation) for a molecule using multiple frequencues using this input >> line: >> >> # polar cphf=rdfreq HF/6-311+g(d,p) >> >> Now here is PART of the output file for a beta value: >> >> ------------------ >> >> Property number 6 -- Beta(w,w,-2w) frequency 8 1.000000: >> 1 >> 1-0.147073D-01 >> 2 0.463795D+02 >> 3 0.788382D-01 >> 4 0.192524D+01 >> 5 0.155267D-01 >> 6 0.110622D+00 >> 7-0.472364D+02 >> 8-0.696973D-03 >> 9 0.124584D+03 >> 10-0.779190D-01 >> 11 0.327279D+00 >> 12-0.801925D+03 >> 13 0.313352D+00 >> 14 0.691959D-01 >> 15-0.333249D+01 >> 16-0.324354D-01 >> 17 0.742661D+03 >> 18-0.112394D+02 >> ----------------------- >> >> My question is, what is the order of these 18 beta tensor values? >> i.e. XXX,XXY,XXZ,YXX,YXZ....etc, is it alphabetical? >> >> Best Regards, >> Anthony > |
|
From: Randy H. <he...@in...> - 2008-04-07 19:59:02
|
is this appearing?? |
|
From: Randy H. <he...@in...> - 2008-04-07 19:37:14
|
Begin forwarded message: > From: "Simpson, Garth J" <gsi...@pu...> > Date: March 30, 2008 9:28:50 PM EDT > To: "Randy Heiland" <he...@in...>, "Levi Haupert" > <lha...@pu...> > Subject: RE: 回复： Re: > =BB=D8=B8=B4=A3=BA=20Re:=20=E5=9B=9E=E5=A4=8D=EF=BC=3FRe:=20Help > > Dear Wei-Qi, > > NLOPredict is a software program. As such, the quality of the > outputs depend on the quality of the inputs. If you trust our > inputs for the amide chromophores of proteins, we have some that > you can upload from the NLOPredict installation page for the NV1 pi- > pi* electronic transition, and several common amide vibrational > modes. If you Import these files (as *.cfg in NLOPredict), you > should be getting outputs for the net NLO tensors of proteins for > the corresponding transition. > > For the electronic properties, we’ve done ZINDO calculations and > rotated the tensor to coincide with the coordinate system for the > amide unit used by NLOPredict. However, I encourage you to perform > your own calculations using whatever means you feel best recovers > the NLO tensors of the monomers and incorporate them into > NLOPredict. Just make sure the tensors are expressed in the > appropriate local coordinates before writing a config file. > > Hope this helps, > > Garth > > From: Randy Heiland [mailto:he...@in...] > Sent: Sunday, March 30, 2008 10:05 AM > To: Simpson, Garth J; Levi Haupert > Subject: Fwd: 回复： Re: > =BB=D8=B8=B4=A3=BA=20Re:=20=E5=9B=9E=E5=A4=8D=EF=BC=3FRe:=20Help > > In case you'd like to try to answer this person's questions... > -Randy > > Begin forwarded message: > > > From: 伟奇 李 <lw...@ya...> > Date: March 30, 2008 4:06:26 AM EDT > To: Randy Heiland <he...@in...> > Subject: 回复： Re: > =BB=D8=B8=B4=A3=BA=20Re:=20=E5=9B=9E=E5=A4=8D=EF=BC=3FRe:=20Help > > Dear Prof: > > Thank you for your help in these days. When my computer is > connected with internet, the NLOpred can be installed. I want to > konw if the ZINDO must be calculated before using NLOpred to > anlysis. In your paper, you mentioned the ZINDO calculation. > Because I haven't ZINDO program, I need konw it. > Thank you very much. > > Your Sincerely > > Wei-Qi > May 30 2008 > > 雅虎邮箱，您的终生邮箱！ > |
|
From: Randy H. <he...@in...> - 2008-04-05 20:29:33
|
Begin forwarded message:
> From: "Simpson, Garth J" <gsi...@pu...>
> Date: July 19, 2007 11:37:01 PM EDT
> To: "Stefano Tonzani" <to...@gm...>
> Cc: <rwa...@pu...>, <he...@in...>
> Subject: RE: NLOpredict info
>
>
> Hello,
> thanks for the answers!
> What I am trying to do at the moment is to use NLOPredict to gauge
> if there is something wrong with my code. At the moment (after
> doing the in-plane averaging and the Fresnel factors math) I get
> for some modes an sps signal where the IR vector should be
> orthogonal to the surface (for the admittedly complicated case of
> DNA, so things might not be as clear cut as I think they are), so I
> was trying to use NLOPredict to validate my procedure with
> something simple.
> First of all, the sps polarization combination is generally no more
> selective to transitions with the ir vectors parallel to the
> surface than the ssp combination, because the role of the Raman
> tensor must also be included.
>
> So I tried stacked thymines, for which I know what one should get,
> and tried to get the same results with both codes, that's when I
> figured out that yours worked differently than I thought, when you
> have multiple chromophores in a pdb file.
> To tell you the truth, I have validated the single molecule
> results, including rotations, and I am quite confident that I get
> the same results as NLOPredict, so now it's the sums of many
> monomers that worry me.
> Basically my idea was to modify a pdb file of DNA in such a way
> that it has only one N,C,O per residue, the other atoms called
> something else (say CA, N1...). I thought that would do the trick,
> but I can see now that I was wrong. I am not sure why, though: 3
> atoms should define the frame, and if you pick the same 3 atoms in
> the pdb residue you should be set.
> I've copied Ron Wampler on this email. He is currently working on
> creating a more general parser to read input files and "trick"
> NLOPredict into treating them as amide monomers. He can probably
> best advise you about how to implement such an algorithm.
>
> The orientation of the Gaussian calculation (the reference frame
> for the input alpha, mu) is not necessarily identical to the first
> residue in the pdb file, since the structure has to be optimized,
> which in general implies some rotations (to the principal axes
> frame). So what I do in my code is first a rotation of the alpha,
> mu (from standard Gaussian orientation) to the frame of the first
> residue of the pdb (in which the z-axis corresponds to the lab z-
> axis).
> It is not yet clear to me that this is the procedure in NLOPredict.
> In fact, it would be great if you could tell me what series of
> transformations are performed on alpha, mu, when you introduce a
> pdb file, say in the case of a protein.
> In NLOPredict, as I understand it, we do no transformations of mu
> and alpha. It turns out to be far simpler and faster to transform
> the full beta tensor. We only use mu and alpha to generate the beta
> tensor in the initial coordinates. In subsequent versions of the
> program, we will be generalizing the code to processes of arbitrary
> order, such that mu and alpha can also be independently transformed.
>
> If I cannot use NLOPredict for this kind of validation that's fine,
> but at least I'm happy I figured it out!
>
> Thanks again,
>
> Stefano
>
> Hope this helps,
> Garth
>
>
>
> On 7/18/07, Simpson, Garth J < gsi...@pu...> wrote:
> Stefano,
>
> I will try to answer your questions as best I can. First, I would
> caution you against redefining the C, O, and N positions if at all
> possible, and certainly not for the calculations you describe.
> Those definitions should only come in when you are reading in a
> protein. They tell the program how to define the coordinate system
> of the amide chromophore and the corresponding orientation angles
> for each amide monomer in a polyamide. The code is not yet general
> enough for this component to be applied straighforwardly to other
> polymeric systems. We provided access to allow changes to the
> definitions of C, O, and N solely for internal purposes in order to
> piggy-back the existing algorithms to other polymeric systems, but
> the implemenation of this capability is nontrivial and confusing,
> becuase you basically have to write a separate parser for every
> polymer in order to tell the code what is meant by a "chromophore".
> We put this capability in so we could start some preliminary
> development to make this procedure more broadly applicable and
> transparent, but it is still at an early stage.
>
> You will have to refresh my memory, but I believe you are trying to
> ultimately run NLOPredict calculations in combination with
> molecular dynamics calculations of DNA polymers in order to
> calculate the beta tensors in each frame, is this correct? If so,
> my advice would be to work with us in the development stage of the
> code to meet this challenge.
>
> However, if you are simply trying to validate the outputs of
> NLOPredict as an early step, don't mess with the definitions of C,
> O, and N!!!! Instead, run all your calculations in single molecule
> mode, where the definitions of C, N, and O are irrelevant. In the
> files you included, the net beta tensor is trivial to calculate
> because no coordinate transformations are required or should be
> performed. Simply add the two monomer tensors together, doubling
> each element in the monomer tensor. I'm a bit confused by your
> comments about transforming from the standard Gaussian orientation
> to the pdb orientation of the first - are they not identical? After
> all, the standard orientations are based on the moments of inertia.
> Translating along one of these principal axes and adding a dimer
> will affect the magnitudes, but not the orientations of the
> eigenvectors of the moment of intertia tensor, such that the
> standard orientation of the dimer and trimer should not change.
>
> HOWEVER, you are unlikely to run across such an ideal circumstance
> in a more general model system. Try rotating one monomer relative
> to the other some arbitrary angle, or translating along something
> other than a principal moment of inertia and you will no longer
> recover the simple summation of the betas when you pull up the
> dimer or trimer in the standard Gaussian orientations. Here's where
> NLOPredict becomes useful.
>
> Hope this helps,
>
> Garth
>
>
>
> From: Stefano Tonzani [mailto:to...@gm...]
> Sent: Wed 7/18/2007 6:41 PM
> To: Simpson, Garth J
> Subject: Re: NLOpredict info
>
> Hello,
> There is still something that mystifies me in the functioning of
> NLOPredict.
> Namely: I tried, for a check against my code, the simple case of
> two thymines stacked at a distance of 4 Angstroms apart on the z-
> axis. The first thymine is in exactly the same geometry and
> orientation as the standard orientation of the Gaussian calculation.
> Therefore, what I do is input the C, O, N positions of the first
> thymine, and then the pdb of the stack. What I get is pretty
> strange, basically a reorientation of the beta tensor of a single
> thymine (if I reorient the arrow accordingly to the calculation for
> a single thymine I get the same beta components). Instead I should
> get the same relative intensities of the different components but
> doubled, since there are 2 thymines now, a thing that I recover
> from my code. Am I still missing something? If I have a stack of 3
> thymines, the signal trebles (after reorientation) so it looks like
> the first thymine is neglected.
> I thought that the C,O,N positions were used to transform to the
> pdb orientation, but in this case the C,O,N have the same exact
> positions as their counterparts in the first residue of the stack
> in the pdb file, so there should be no rotation. Once again, what
> does the transformation to the principal axes accomplish? The
> Gaussian standard orientation is already in the principal axes frame.
> The transformations I perform in my code are from the standard
> Gaussian orientation to the pdb orientation of the 1st residue
> (which is going to be the final orientation), then I sum up the
> various betas using the relations between their pdb coordinates.
> Thanks again for your patience,
>
> Stefano
>
> On 6/12/07, Simpson, Garth J <gsi...@pu... > wrote:
> Stefano,
>
> That works for me. If you have any troubles finding my office, call
> me at 765-496-3054. See you Thursday morning.
>
> Garth
> --------------------------
> Garth Simpson
> Associate Professor of Chemistry
> Purdue University
>
>
>
>
> ----- Original Message -----
> From: Stefano Tonzani <to...@gm... >
> To: Simpson, Garth J
> Sent: Tue Jun 12 10:15:32 2007
> Subject: Re: NLOpredict info
>
> Hello,
> so is it fine for me to come on Thursday? I should be getting there
> around 10:30 or so. Your office is BRWN 4170C, right?
> Let me know if this does not work for you,
>
> Thanks again,
>
> Stefano
>
>
> On 6/8/07, Stefano Tonzani < to...@gm...> wrote:
>
> Hello,
>
> Thanks a lot for the help!
> It would be probably best for me to meet at Purdue say
> Wednesday, Thursday or Friday, let me know what works best for you.
> I have, in fact, only tried planar diNMA, which I assumed
> was the one in the Supporting information, is this correct?
> My results for diNMA do not look so identical to the ones
> in your paper, but that might be a problem of linearly combining
> the two Amide A modes or something like that,
>
> Later,
>
> Stefano
>
>
>
> On 6/8/07, Simpson, Garth J < gsi...@pu...> wrote:
>
> Stefano,
>
>
>
> First, I would be happy to meet with you to discuss
> your scientific problem and the possible approaches in which
> NLOPredict could help address them. I will be in town all of next
> week – can you come down then? Alternatively, I'll be at the Mol-
> Spec conference. If you are attending as well, we could meet there
> to talk shop.
>
>
>
> Second, it sounds like your results are actually
> matching up pretty well qualitatively with those published in the
> supporting information of the JPC-B paper for the Amide A mode
> (Figure S1 of the attached), apart from a potential units
> conversion issue.
>
>
>
> Finally, assembly of the predicted NLO properties
> of di-NMA is not the best place to start because of the large
> degrees of freedom when selecting a coordinate system (nonplanar
> molecule of low symmetry). If my memory serves (Andy can correct me
> on this), we used the sum and differences of the monomer transition
> moments to define the coordinates in the twisted form of di-NMA,
> which is not trivial to reproduce. The planar structure is
> considerably simpler, but still significantly more complex than DNA
> (where the assumption of C(infinity) symmetry is reasonable). We
> had calculation of the NLO properties of DNA on the list of things
> to do at the Chautauqua, but we never quite got around to it. If
> you come down with the vibrational calculations in hand, we should
> be able to hammer it out in a day for all the isoforms of DNA no
> problem. That would certainly be easier than trying to do this by
> long distance.
>
>
>
> Garth
>
>
>
> ________________________________
>
> From: Stefano Tonzani
> [ mailto:to...@gm...]
> Sent: Thursday, June 07, 2007 12:22 PM
> To: Simpson, Garth J
> Subject: NLOpredict info
>
>
>
> Hello,
> sorry I couldn't come to the nonlinear optics
> Chautauqua, but I was out of the country at the time,
>
> I had some exchange with Andy Moad about the NMA/
> diNMA calculations you guys did a while ago, to validate the code I
> am developing. Here are the questions:
> 1) I did some Gaussian calculations like in a
> sample file I got from Andy for diNMA, to compare to the results in
> your JPCB 109 20009 ('05) in the supporting information. In
> particular I look at the Amide A mode (there are two of them, modes
> 50 and 51 in the file I am sending you) and basically I plug in the
> numbers from Gaussian (lines 9043-9056) and reorient the molecule
> as in your paper, with the dipole pointing at me and the principal
> molecular axis horizontal, in other words:
> theta=90, phi=139, psi=270
> This gives me a beta tensor for both modes with
> large xxz and zzz components (-15 and -18 for mode 50) and little
> else, which is very different from what is in the paper. Am I doing
> something wrong? Andy said that you guys probably added the two
> modes together in some way, but he did not remember the details.
> 2) Aside from that, I am trying to assemble the
> diNMA using the spectra of two separate NMA molecules. I use the
> pdb file attached and for the Gaussian calculations I also attach
> the output file. The orientation of one of the monomers in the
> "assembled" diNMA is identical to the Gaussian calculations for
> NMA, the 2nd is rotated like in the "real" diNMA molecule. This is
> the way it should be done, correct?
>
> Then I have a request: would it be possible for me
> to come to Purdue for a day so we could talk about these and other
> issues?
>
>
> Thank you very much again,
>
> Stefano
>
> On 4/6/07, Stefano Tonzani < to...@gm... <
> mailto:to...@gm...> > wrote:
>
> Hello,
> thanks for the heads-up. I will be more than happy
> to talk when you get back,
>
> Later,
>
> Stefano
>
> On 4/6/07, Andrew Moad < and...@ni... <
> mailto:and...@ni...> > wrote:
>
> I am leaving for the MRS conference in San
> Francisco. When I get back
> on April 16th I will be able to give your
> questions some better answers.
> In brief:
> 1) Don't use it. It is something I wrote
> just for me and am not
> comfortable with anyone else using it. I
> didn't realize that it was
> included in the program. It works, but
> requires a very specific gaussian
> calculation. I would be glad to explain it
> to you later. It is useless
> for vibrational SFG.
>
> 2) No script is needed. I get these from
> gaussian directly via the
> IOP(33/7)=1 keyword. This will give you the
> transition dipole and
> transition polarizability directly (lines
> 9439-9647 in NMA_002.log). I
> can give you my procedure when I get back.
>
> 4) There are several routes. One is to load
> a PDB file that has a helix
> in it. I'll attach mine ( alphahelix03.pdb).
>
> Sorry that I don't have more time this week.
> Andy
>
> Stefano Tonzani wrote:
> > Dear Andrew,
> > my name is Stefano Tonzani, I work at
> Northwestern University and I am
> > developing an approach that would use
> molecular dynamics calculations
> > to predict nonlinear optical spectra.
> > Since I am interested in modeling NLO
> experiments of DNA on surfaces,
> > and I recently met Garth Simpson and
> talked about the software you
> > guys have developed (NLOpredict) I wanted
> to ask you some questions
> > about it, if possible.
> > In particular, your software would be
> useful for me to validate what I
> > am currently doing, since there are no
> other computational apporaches
> > one can compare to. So the questions are:
> > 1) I saw that the package includes a
> script called gaussianParser.py
> > for SHG (or electronically resonant SFG).
> Does that work well? Garth
> > said he hasn't used it very much, but it
> would be useful for me, since
> > what I am developing is kind of similar,
> and of course it is much
> > easier to get stuff directly from ab
> initio files than calculate and
> > input them one by one, by hand, into a
> graphic interface.
> > 2) Is there an equivalent of that script
> for vibrationally resonant
> > SFG (since that is what I am interested in)?
> > 3) On that note, how did you obtain the
> results in your J. Phys. Chem
> > B 109, 20009 (2005) in particular for the
> vibrational SFG in the
> > supporting info? Is it possible to have
> some of the files used for the
> > dimer (not perturbation) calculation
> (like ab initio or scripts you
> > use to parse the ab initio)?
> > 4) I haven't really understood how to
> combine chromophores in that
> > code, for example if you have performed
> calculations for a molecule
> > and you want to build a helix of
> identical molecules.
> >
> > Thank you very much in advance,
> >
> > Best regards,
> >
> > Stefano Tonzani
>
>
>
>
> --
> Stefano Tonzani
> Department of Chemistry
> Northwestern University, 2145 Sheridan Rd.,
> Evanston, IL. 60208-3113 USA
> Phone: (847) 467 4989
> Fax: (847) 467 4996
> Office location: Room 4037, Center for
> Nanofabrication and Molecular Self Assembly, 2190 Campus Dr.
> Home Page: http://ucsu.colorado.edu/~tonzani/
> Newstyle/Newstyle.html <http://ucsu.colorado.edu/%7Etonzani/
> Newstyle/Newstyle.html>
>
|
|
From: Randy H. <he...@in...> - 2008-04-05 20:27:57
|
Begin forwarded message:
> From: "Simpson, Garth J" <gsi...@pu...>
> Date: March 20, 2008 1:05:02 PM EDT
> To: <cl...@la...>
> Cc: "Randy Heiland" <he...@in...>, "Levi Haupert"
> <lha...@pu...>
> Subject: RE: questions about NLOpredict
>
> Claire,
>
>
>
> Levi Haupert in my group has been working recently on simplifying
> the input and output of NLOPredict in text file format. You may
> want to check with him regarding access to this modified form of
> the program. In this version, you can write a file containing the
> microscopic tensor and ensemble averaged chi(2) tensor for any
> molecular orientation you choose.
>
>
>
> Regarding units, we have tried to keep the program as general as
> possible. If you are using our calculated values for the tensor of
> NMA (uploaded as the NMA.cfg configuration file), the units are in
> atomic units of e^2 *a0^2/Eh. To convert to units of
> hyperpolarizability, you multiply this by the experimental
> lineshape function, expressed in units of 1/Eh. I believe the
> microscopic tensor for the electronic pi->pi* transition of NMA is
> also the default when the program is installed, but I could be wrong.
>
>
>
> You don’t need to change the coordinates of the reference atoms NCO
> – those are used to define a local coordinate system that is
> consistent with the definition of the monomer tensor (e.g.,
> NMA.cfg). We put that option in just in case we chose to perform
> the summations over different monomeric units other than amides.
>
>
>
> Something you may want to try is just to pull up a pdf file
> corresponding to an alpha-helix, and make sure that you recover the
> trends reported in our previous paper to double-check the operation.
>
>
>
> Hope this helps, and again Levi should be able to provide you with
> a updated version of NLOPredict that facilitates exporting the
> calculation outputs, and importing cfg files derived from quantum
> chemistry.
>
>
>
> Garth
>
>
>
> From: Randy Heiland [mailto:he...@in...]
> Sent: Thursday, March 20, 2008 9:13 AM
> To: Simpson, Garth J; Levi Haupert
> Subject: Fwd: questions about NLOpredict
>
>
>
> Let me forward these questions from our new French user to you guys
> also...
>
>
>
> -Randy
>
>
>
> Begin forwarded message:
>
>
>
>
> From: Loison Claire <cl...@la...>
>
> Date: March 20, 2008 8:47:41 AM EDT
>
> To: Randy Heiland <he...@in...>
>
> Subject: questions about NLOpredict
>
>
>
> Randy,
>
>
>
> I have been playing a bit with NLOPredict and
>
> I permit myself to ask you about it.
>
>
>
> One thing I tried is just to use a PBD structure into chimera,
>
> and get the microscopic tensor (beta_{ijk}, in the molecular frame).
>
> The reply log of chimera give some ouput tensors, but could you
> precise me
>
> where to find the microscopic tensor ? And the units
>
> Is the microscopic tensor of the backbone already loaded in the
> NLOPredict,
>
> or do I have to input it in the tensor frame first ?
>
> Can it be done by opening the NMA cfg file first ?
>
>
>
> How is it with the cooridnates of the reference atoms NCO ?
>
> Do I have to change them depending on the PDB structure ?
>
>
>
> Thanks in advance for your understanding,
>
> Claire
>
>
>
>
>
>
>
>
>
>
>
> Randy Heiland wrote:
>
> Claire,
>
>
>
> Very glad to hear that. Regarding a user guide, we're probably
> guilty of not having such a thing, but I'll cc Garth in case he's
> aware of something along those lines. In addition to the JCC
> article, there's also the Video Preview on the NLOPredict web page.
>
>
>
> We'd be interested in learning more about your type of problems
> and, hopefully, figure out how NLOPredict might be able to help.
>
>
>
> -Randy
>
>
>
>
>
> On Mar 13, 2008, at 9:36 AM, Loison Claire wrote:
>
>
>
> Hi Randy,
>
>
>
> It works ! Thank you very much for your rapid help.
>
>
>
> Is there a kind of user-guide for NLOPredict ? I have seen the
> example cases on the website, and read the article about NLOPredict
>
> in J. Comput. Chem. but I could not guess the step by step
> procedure to correctly
>
> use it for my own problems.
>
>
>
> Have a nice day,
>
> Claire Loison
>
>
>
>
>
>
>
>
>
>
>
> Randy Heiland wrote:
>
> Claire,
>
>
>
> When you get a chance, could you try to do the following (on your
> 32-bit installation of Chimera):
>
>
>
> cd /usr/local/chimera/lib
>
> ln -s libtk8.4.so libtk8.4.so.0
>
> ln -s libtcl8.4.so libtcl8.4.so.0
>
>
>
> This will simply create symbolic links (new files) with the ".0"
> suffix appended, but simply point to the other .so's. This is what
> NLOPredict was complaining about.
>
>
>
> Just so you know, Chimera comes as a self-contained application,
> i.e. it doesn't look in your system directories for any library
> dependencies - it comes with everything that it needs. Therefore,
> any attempts you make to install a lib on your system, assuming
> that Chimera will look there, will be in vain.
>
>
>
> let me know how it goes,
>
> -Randy
>
>
>
> On Mar 12, 2008, at 10:42 AM, Loison Claire wrote:
>
>
>
>
>
> Yes, I use a linux pc (suse10.1). Here is the answer when I type
> "uname -a"
>
>
>
> Linux pclasim77 2.6.16.13-4-smp #1 SMP Wed May 3 04:53:23 UTC 2006
> x86_64 x86_64 x86_64 GNU/Linux
>
>
>
> I have to fetch my kids now, but I'll check my emails this evening
> (at least the european evening...)
>
>
>
> Claire
>
>
>
> Randy Heiland wrote:
>
> Thanks for supplying the error log - that should help me debug it
> quicker. Based on that, I take it that you're running Linux as
> your operating system? Can you send me the output from entering
> the command "uname -a"?
>
>
>
>
>
> On Mar 12, 2008, at 9:54 AM, Loison Claire wrote:
>
>
>
> Dear M. Heiland,
>
>
>
> Thank you very much for you rapid answer.
>
>
>
> I am not running Vista on my PC ( at least not that I know).
>
>
>
> When I click in the chimera window "Tools > Indiana/Purdues/
> NLOPredict",
>
> a new window appears with the error at the end of this email
> (sorry it's long...).
>
>
>
> I installed then python-configobj, python-numpy and matplotlib for
> 64 bits on my pc,
>
> but I suppose NLOPredict does not search in the right place
>
> ( for example /usr/lib64/python2.4/site-packages/matplotlib ).
>
> How can i tell NLOPredict to use the right matplotlib ? Is that
> possible ?
>
>
>
> To avoid this problem I installed chimera in 32 bit (as advised on
> chimera website),
>
> then i get the same kind of problem, this time with libtk8.4 not found
>
> (libtk8.4 which i have yet not installed in 32 bits).
>
>
>
> Thank you in advance for you kind help,
>
>
>
> Claire
>
>
>
> ERROR LOG :
>
> ----------------------------------------------------------------------
> --------------
>
> The import of the numpy version of the _transforms module,
>
> _ns_transforms, failed. This is is either because numpy was
>
> unavailable when matplotlib was compiled, because a dependency of
>
> _ns_transforms could not be satisfied, or because the build flag for
>
> this module was turned off in setup.py. If it appears that
>
> _ns_transforms was not built, make sure you have a working copy of
>
> numpy and then re-install matplotlib. Otherwise, the following
>
> traceback gives more details:
>
>
>
> Traceback (most recent call last):
>
> File "/usr/local/chimera/lib/python2.4/site-packages/
> NLOPredict-1.0.7-py2.4.egg/NLOPredict/__init__.py", line 3, in ?
>
> from gui import showNLOPredictDialog
>
> File "/usr/local/chimera/lib/python2.4/site-packages/
> NLOPredict-1.0.7-py2.4.egg/NLOPredict/gui.py", line 30, in ?
>
> from matplotlib.figure import Figure
>
> File "/usr/local/chimera/lib/python2.4/site-packages/
> matplotlib-0.90.0-py2.4-linux-i686.egg/matplotlib/figure.py", line
> 5, in ?
>
> import artist
>
> File "/usr/local/chimera/lib/python2.4/site-packages/
> matplotlib-0.90.0-py2.4-linux-i686.egg/matplotlib/artist.py", line
> 4, in ?
>
> from transforms import identity_transform
>
> File "/usr/local/chimera/lib/python2.4/site-packages/
> matplotlib-0.90.0-py2.4-linux-i686.egg/matplotlib/transforms.py",
> line 223, in ?
>
> from _transforms import Value, Point, Interval, Bbox, Affine
>
> File "/usr/local/chimera/lib/python2.4/site-packages/
> matplotlib-0.90.0-py2.4-linux-i686.egg/matplotlib/_transforms.py",
> line 17, in ?
>
> from matplotlib._ns_transforms import *
>
> ImportError: /usr/local/chimera/lib/python2.4/site-packages/
> matplotlib-0.90.0-py2.4-linux-i686.egg/matplotlib/
> _ns_transforms.so: wrong ELF class: ELFCLASS32
>
> ImportError Exception in Tk callback
>
> Function: <function activate at 0x2b446351b2a8> (type: <type
> 'function'>)
>
> Args: ()
>
> Traceback (innermost last):
>
> File "CHIMERA/lib/python2.4/site-packages/Pmw/Pmw_1_2/lib/
> PmwBase.py", line 1747, in __call__
>
> None
>
> File "CHIMERA/share/chimera/extension/base.py", line 469, in activate
>
> None
>
> File "CHIMERA/share/chimera/extension/base.py", line 91, in
> menuActivate
>
> None
>
> File "/usr/local/chimera/lib/python2.4/site-packages/
> NLOPredict-1.0.7-py2.4.egg/NLOPredict/ChimeraExtension.py", line
> 17, in activate
>
> self.module().showNLOPredictDialog()
>
> File "CHIMERA/share/chimera/extension/base.py", line 113, in module
>
> None
>
> File "/usr/local/chimera/lib/python2.4/site-packages/
> NLOPredict-1.0.7-py2.4.egg/NLOPredict/__init__.py", line 7, in ?
>
> raise exc
>
> ImportError: /usr/local/chimera/lib/python2.4/site-packages/
> matplotlib-0.90.0-py2.4-linux-i686.egg/matplotlib/
> _ns_transforms.so: wrong ELF class: ELFCLASS32
>
>
>
> ------------------------------------------------------------
>
>
>
>
>
>
>
>
>
>
>
>
>
> Randy Heiland wrote:
>
> My guess is that the 64-bit issue of matplotlib is indeed the
> problem, but I will look into it and get back to you.
>
>
>
> Dr. Loison, are you running Vista on your PC? And does the
> NLOPredict GUI/panel actually appear from inside Chimera, but it's
> just that something is not displayed properly (matplotlib-generated
> plot perhaps), or does some action in NLOPredict actually crash
> Chimera?
>
>
>
> -Randy
>
>
>
> On Mar 12, 2008, at 9:10 AM, Simpson, Garth J wrote:
>
>
>
> Dear Dr. Loison,
>
>
>
> I'm copying Randy Heiland on this email. He is the contact person
> on technical aspects of the program related to installation and
> basic operation. We focus primarily on algorithms and applications.
>
>
>
> I sincerely hope you find it helpful!
>
>
>
> All the best,
>
>
>
> Garth
>
>
>
>
>
>
>
> -----Original Message-----
>
> From: Loison Claire [mailto:cl...@la...]
>
> Sent: Wednesday, March 12, 2008 6:29 AM
>
> To: Simpson, Garth J
>
> Subject: problem installing/using NLOpredict
>
>
>
> Dear Prof. Simpson,
>
>
>
> The possibility of using NLOpredict seems very interesting for my
>
> research, so
>
> I wanted to install NLOpredict on my 64-bit pc.
>
>
>
> Because of the built of my local libraries, I used the 64 bits
> chimera
>
> executable.
>
> The installation of NLOpredict goes without errors, but when i try to
>
> use it,
>
> there seems to be a problem with the matplotlib. Is teh library of the
>
> package in 32 bits ? in 64 bits ?
>
>
>
> Is there a package to install the NLOpredict compatible with 64 bits
>
> architechture ?
>
> Do I have to compile matplotlib in 64 bit ?
>
>
>
> Thank you in advance for your understanding !
>
> Yours,
>
> Claire Loison
>
>
>
> --
>
> ----------------------------------------------------
>
> Claire LOISON
>
> Theoretical Physical-Chemistry Group in LASIM
>
> tel : 00 33 4 72 43 12 57
>
> cellular : 00 33 6 31 30 11 72
>
> ----------------------------------------------------
>
> Laboratoire de Spectrométrie Ionique et Moléculaire
>
> UMR 5579
>
> Domaine Scientifique de la Doua
>
> Université Claude Bernard Lyon 1
>
> Bâtiment Alfred Kastler
>
> 43, bd du 11 Novembre 1918
>
> 69622 Villeurbanne cedex
>
> ---------------------------------
>
>
> [The attachment 2005 JPCB NLO of Proteins.pdf has been manually
> removed]
>
> [The attachment alphahelix03.pdb has been manually removed]
>
|
|
From: Randy H. <he...@in...> - 2008-04-05 20:27:57
|
Begin forwarded message: > From: "Simpson, Garth J" <gsi...@pu...> > Date: March 21, 2008 10:07:25 AM EDT > To: <cl...@la...> > Cc: "Levi Haupert" <lha...@pu...>, "Randy Heiland" > <he...@in...> > Subject: FW: NLOPredict questions > > Claire, > > You raise a good point, at some point, we really need to sit down > and prepare a user guide. > > Regarding your questions, the only two inputs you need to provide > NLOPredict are the tensor for the amide monomer (e.g., from the > NMA.cfg file, calculated for the pi-pi* transition at 190 nm) and > the molecular structure as read directly from the pdb file of the > protein. The algorithm running behind the scenes then starts at one > end of the protein chain, finds the first amide unit, determines > its orientation relative to the lab frame (defined with z > perpendicular to the screen), and redefines the molecular tensor > into the lab frame. NLOPredict coherently adds all these > contributions together to generate the net beta tensor displayed in > the bar. You do not need to worry about changing orientations or > reference frames - NLOPredict takes care of that all. The cfg files > we provide were all written in the local amide coordinate system > expected by NLOPredict. > > Regarding the difference between mu, alpha, and beta, if you are > using the cfg files we provide, the beta tensor is generated as the > direct product of the transition moment and either the Raman or Two- > photon absorption tensor (alpha). Alternatively, you can enter your > own values for those, or for the full beta tensor. That option is > there simply to improve generalizability in the program. > > There are some tricks you can do to visualize individual amino > acids. If you select individual residues in Chimera, you can then > save them as a separate pdb file and load them up again. > > Hope this helps, > > > Garth > > -----Original Message----- > From: lha...@pu... [mailto:lha...@pu...] > Sent: Friday, March 21, 2008 9:13 AM > To: Simpson, Garth J > Subject: Fwd: NLOPredict questions > > > > ----- Forwarded message from Loison Claire <cl...@la...- > lyon1.fr> ----- > Date: Fri, 21 Mar 2008 10:39:14 +0100 > From: Loison Claire <cl...@la...> > Reply-To: Loison Claire <cl...@la...> > Subject: NLOPredict questions > To: Levi Haupert <lha...@pu...> > > Dear Levi Haupert, > > I got your email adress from Garth Simpson, when I asked him > some questions about NLOpredict. > > I am a new user to NLOPredict, and try to understand the basics, and I > have a lot of questions ... > if you have a kind a small user guide, it would be great ! > > * Concerning the input, and the way I should proceed to get > the molecular beta tensor of a protein using its PDB file. > > * If i understood right, the beta tensor that one have to input > should be > defined in a molecular frame consistent with the CNO positions also > defined in the input. > Is that right ? If I would like to use the NMA tensor published on the > internet NMA.cfg. > Then the frame of the NMA calculation is surely different from the > frame > of the PDB of the protein. > I suppose this is not a problem, is it ? What is the order : load the > NMA.cfg in NLOPredict, > then the PDB in chimera ? The mu and alpha tensors look like they are > computer from a quantum chemical code, > but the beta tensor is just a model one. How should i understand which > is used in the calculation ? > > * Garth Simpson told me you develop a version of NLOPredict > where it is possible to have an ascii input file where > the tensors and rotations of the different sources are written. > Of course it would be great to have the possibility to define several > different sources, > to take into account the aromatic amino-acids or completely different > molecules. Is that possible yet ? > > * Concerning the outputs. Where is the molecular beta tensor ? > in the reply log of chimera? I get lots of beta tensors in the reply > log but i am not > sure what they are, and in which frame. > > * Is there a possibility to output the values of the tensor seen > on the > screen into an ascii file ? > > * Is it possible to see the output of the rotation for each amino- > acids ? > I suppose there is a tool doing such a conversion ( PDB-> > rotations). > Could it be available as an independant tool inside NLOPredict ? > > I thank you in advance for your help and understanding. > Yours, > Claire > > > > -- > ---------------------------------------------------- > Claire LOISON > Theoretical Physical-Chemistry Group in LASIM > tel : 00 33 4 72 43 12 57 > cellular : 00 33 6 31 30 11 72 > ---------------------------------------------------- > Laboratoire de Spectrométrie Ionique et Moléculaire > UMR 5579 > Domaine Scientifique de la Doua > Université Claude Bernard Lyon 1 > Bâtiment Alfred Kastler > 43, bd du 11 Novembre 1918 > 69622 Villeurbanne cedex > ---------------------------------------------------- > > > > ----- End forwarded message ----- > > |
|
From: Randy H. <he...@in...> - 2008-04-05 20:27:54
|
Begin forwarded message: > From: "Simpson, Garth J" <gsi...@pu...> > Date: November 29, 2007 3:30:20 PM EST > To: <he...@in...> > Cc: "Andrew Moad" <and...@ni...>, <rp...@pu...> > Subject: FW: Chis for NLOPredict algorithm > > Randy, > One of the new directions we'd like to see NLOPredict head is > toward extension to higher and lower order optical processes. The > next step in this is to tackle four-wave mixing, in which the > molecular and macroscopic processes are described in terms of 3 by > 3 by 3 by 3 tensors. We were thinking we could upgrade the list of > beta tensors to include a scroll-bar so that all 81 elements of the > molecular tensor could be shown on the screen. We can probably fit > all 21 of the unique chi tensor elements on the bottom panel. The > unique chi tensor elements are given on the first page of the > attached ppt file, with the algorithms used to connect the values > for beta and chi on the second page. > This is probably all quite new to you - I don't think you've > tinkered much with this part of the code. Feel free to contact me > with questions. > ªí-----Original Message----- > From: Ryan Davis [mailto:rp...@pu...] > Sent: Thursday, November 29, 2007 3:21 PM > To: Simpson, Garth J > Subject: Chis for NLOPredict algorithm > Here they are. I put the two Chis (XXXX and YXXX) in parentheses > that we took out to avoid being redundant. I figured we'd still > want to incorporate those into the algorithm since they're not > zero; just not unique. > Although, would we want the algorithm designed to where those two > are calculated separately just like the rest of them? I don't know > if that would be more efficient than having it do a sum of 3 other > Chis...I don't know too much about programming. > -- > Ryan Davis > Purdue University > Department of Chemistry > > > [The attachment Chis_for_algorithm.ppt has been manually removed] > |
|
From: Randy H. <he...@in...> - 2008-04-05 20:27:54
|
Begin forwarded message: > From: "Simpson, Garth J" <gsi...@pu...> > Date: November 29, 2007 1:25:26 PM EST > To: "Antonio De Crisci" <ant...@ut...>, "Andrew > Moad" <and...@ni...>, <he...@in...>, > <rwa...@pu...> > Subject: RE: NLOPredict / Gaussian question > > Anthony, > > From the output file, it looks like the beta you are calculating > (whether you intend it or not) is for the Pockels effect, as it is > given by B(-w;w,0). This tensor describes the change in the linear > polarizability as a function of an applied external field. The > values in parentheses correspond to the exigent frequency (first > value) and the two incident frequencies (second and third values). > Since the last frequency is zero, it indicates a DC electric field. > Although I have not used the algorithm you describe for calculating > hyperpolarizabilities, it may be a simple matter of including more > than one frequency in the input. > > Hope this helps, > > Garth > > -----Original Message----- > From: Antonio De Crisci [mailto:ant...@ut...] > Sent: Thursday, November 29, 2007 11:47 AM > To: Andrew Moad; he...@in...; rwa...@pu...; Simpson, > Garth J > Cc: me2 > Subject: Re: NLOPredict / Gaussian question > > Thanks you all for your replies! You have clarified the my > situation. Thanks for the ACR paper, this is something that I will > definitely be digging into. > > But in terms of my calculation, Andrew, I will have to try your > command line for my Gaussian calculations. > > Just to let you know what I have been setting up my calculations, > attached is the out file of one of my calculations (urea). I used > keywords polar and CPHF=RdFreq and set a frequency of 1064nm > (there is a line in the out that > says "Using perturbation frequencies: 0.000000 0.042823" in > the out > file) because i want to calculate the frequency-dependent > polarizabilities and hyperpolarizabilities, as a in the presence of > a 1064nm laser for the tensors in NLOPredict. > > Gaussian out file does give me mu, alpha and beta at TWO > frequencies (0 and > 1064nm) but the beta output was confusing me at first. > > Please see my Gaussian out file (txt format) of urea.What do you > think? > > Sorry if some of the questions seem trivial, im actually an > organometallic chemistry and just recently have become familiar > with NLO properties of some of the materials I have made and would > like to compliment the results with computations and NLOPredict for > publication. > > Thanks, > Anthony > > > > ----- Original Message ----- > From: "Andrew Moad" <and...@ni...> > To: <ant...@ut...> > Sent: Thursday, November 29, 2007 11:09 AM > Subject: NLOPredict / Gaussian question > > >> Dear Anthony, >> >> The answer to your question is yes and it depends. If you are talking >> about electronic transitions (such as those common for second >> harmonic >> generation) the answer is yes. We use Gaussian to calculate >> transition >> dipoles between ~100 excited states and perform a sum-over-states >> (outside >> of Gaussian) calculation. Here we can indeed calculate all 27 tensor >> elements. If you are talking about vibrational transitions then >> the story >> is a little more complicated. Gaussian will give you a “symmetric” >> transition polarizability for each vibrational transition, but you >> are >> right that it contains only the unique tensor elements. Internally >> Gaussian is applying finite fields and doing derivatives to do this >> calculation. To get all vibrational tensor elements you will have >> to apply >> the finite fields yourself and do the derivatives yourself (as far >> as I >> know). It’s not that bad. I started to write a program to automate >> this, >> but decided to just use the symmetric tensors instead. Actually, >> we like >> to use the b = am approximation which is valid if you are near >> resonance >> with a real transition in the molecule. See my J. Phys. Chem. >> papers for a >> detailed explanation. This simplifies computation because you just >> need to >> calculate a 9 element a tensor and a one transition dipole m and >> then do a >> direct product between them to get a 27 element tensor (there are >> some >> equalities so you would not have 27 unique tensor elements here). >> >> There is more about Gaussian that I don’t know than I know, but it >> does >> concern me that I have never seen CPHF=ReFreq. We usually do a >> different >> calculation to get the vibrational transition polarizabilities. I >> usually >> use this: >> >> #p freq=raman rb3lyp/6-311++g(d,p) int=ultrafine iop(7/33=1) polar >> >> to get the transition a and m for each vibration. What is the full >> command >> line (line the one above) that you are using? >> >> I’m sorry that I could not give you the answer that you wanted. Try >> emailing Gaussian with one very specific question. They are pretty >> good at >> answering those. >> >> Andy >> >> =================================================== >> | Im a grad student at the University of Toronto investigating the >> NLO >> properties of single molecules. I came across your fascinating >> | paper J. Comp. Chem. 2007, 28, 1996 using NLOPredict to >> visualize the >> process. In your computations (using G98), were you >> | able to calculate all 27 beta tensors? Im running into an issue >> were >> using Gaussian 03 to calculate the beta tensors, but the output file >> | only contains 18 tensors using the keywords 'polar' and >> 'CPHF=ReFreq' at >> 1064nm. >> | >> | Would you be able to tell me how you found all 27 tensors from >> calculation? >> =================================================== >> >> -- >> Dr. Andrew J. Moad >> Research Chemist >> National Institute of Standards and Technology >> Surface and Microanalysis Science Division >> 100 Bureau Drive, Stop 8372 >> Gaithersburg, MD 20899-8372 >> http://www.nist.gov >> Tel: (301) 975-8532 >> Fax: (301) 417-1321 >> Email: and...@ni... >> >> >> >> |
|
From: Randy H. <he...@in...> - 2008-04-05 20:24:31
|
Begin forwarded message:
> From: "Simpson, Garth J" <gsi...@pu...>
> Date: July 19, 2007 11:37:01 PM EDT
> To: "Stefano Tonzani" <to...@gm...>
> Cc: <rwa...@pu...>, <he...@in...>
> Subject: RE: NLOpredict info
>
>
> Hello,
> thanks for the answers!
> What I am trying to do at the moment is to use NLOPredict to gauge
> if there is something wrong with my code. At the moment (after
> doing the in-plane averaging and the Fresnel factors math) I get
> for some modes an sps signal where the IR vector should be
> orthogonal to the surface (for the admittedly complicated case of
> DNA, so things might not be as clear cut as I think they are), so I
> was trying to use NLOPredict to validate my procedure with
> something simple.
> First of all, the sps polarization combination is generally no more
> selective to transitions with the ir vectors parallel to the
> surface than the ssp combination, because the role of the Raman
> tensor must also be included.
>
> So I tried stacked thymines, for which I know what one should get,
> and tried to get the same results with both codes, that's when I
> figured out that yours worked differently than I thought, when you
> have multiple chromophores in a pdb file.
> To tell you the truth, I have validated the single molecule
> results, including rotations, and I am quite confident that I get
> the same results as NLOPredict, so now it's the sums of many
> monomers that worry me.
> Basically my idea was to modify a pdb file of DNA in such a way
> that it has only one N,C,O per residue, the other atoms called
> something else (say CA, N1...). I thought that would do the trick,
> but I can see now that I was wrong. I am not sure why, though: 3
> atoms should define the frame, and if you pick the same 3 atoms in
> the pdb residue you should be set.
> I've copied Ron Wampler on this email. He is currently working on
> creating a more general parser to read input files and "trick"
> NLOPredict into treating them as amide monomers. He can probably
> best advise you about how to implement such an algorithm.
>
> The orientation of the Gaussian calculation (the reference frame
> for the input alpha, mu) is not necessarily identical to the first
> residue in the pdb file, since the structure has to be optimized,
> which in general implies some rotations (to the principal axes
> frame). So what I do in my code is first a rotation of the alpha,
> mu (from standard Gaussian orientation) to the frame of the first
> residue of the pdb (in which the z-axis corresponds to the lab z-
> axis).
> It is not yet clear to me that this is the procedure in NLOPredict.
> In fact, it would be great if you could tell me what series of
> transformations are performed on alpha, mu, when you introduce a
> pdb file, say in the case of a protein.
> In NLOPredict, as I understand it, we do no transformations of mu
> and alpha. It turns out to be far simpler and faster to transform
> the full beta tensor. We only use mu and alpha to generate the beta
> tensor in the initial coordinates. In subsequent versions of the
> program, we will be generalizing the code to processes of arbitrary
> order, such that mu and alpha can also be independently transformed.
>
> If I cannot use NLOPredict for this kind of validation that's fine,
> but at least I'm happy I figured it out!
>
> Thanks again,
>
> Stefano
>
> Hope this helps,
> Garth
>
>
>
> On 7/18/07, Simpson, Garth J < gsi...@pu...> wrote:
> Stefano,
>
> I will try to answer your questions as best I can. First, I would
> caution you against redefining the C, O, and N positions if at all
> possible, and certainly not for the calculations you describe.
> Those definitions should only come in when you are reading in a
> protein. They tell the program how to define the coordinate system
> of the amide chromophore and the corresponding orientation angles
> for each amide monomer in a polyamide. The code is not yet general
> enough for this component to be applied straighforwardly to other
> polymeric systems. We provided access to allow changes to the
> definitions of C, O, and N solely for internal purposes in order to
> piggy-back the existing algorithms to other polymeric systems, but
> the implemenation of this capability is nontrivial and confusing,
> becuase you basically have to write a separate parser for every
> polymer in order to tell the code what is meant by a "chromophore".
> We put this capability in so we could start some preliminary
> development to make this procedure more broadly applicable and
> transparent, but it is still at an early stage.
>
> You will have to refresh my memory, but I believe you are trying to
> ultimately run NLOPredict calculations in combination with
> molecular dynamics calculations of DNA polymers in order to
> calculate the beta tensors in each frame, is this correct? If so,
> my advice would be to work with us in the development stage of the
> code to meet this challenge.
>
> However, if you are simply trying to validate the outputs of
> NLOPredict as an early step, don't mess with the definitions of C,
> O, and N!!!! Instead, run all your calculations in single molecule
> mode, where the definitions of C, N, and O are irrelevant. In the
> files you included, the net beta tensor is trivial to calculate
> because no coordinate transformations are required or should be
> performed. Simply add the two monomer tensors together, doubling
> each element in the monomer tensor. I'm a bit confused by your
> comments about transforming from the standard Gaussian orientation
> to the pdb orientation of the first - are they not identical? After
> all, the standard orientations are based on the moments of inertia.
> Translating along one of these principal axes and adding a dimer
> will affect the magnitudes, but not the orientations of the
> eigenvectors of the moment of intertia tensor, such that the
> standard orientation of the dimer and trimer should not change.
>
> HOWEVER, you are unlikely to run across such an ideal circumstance
> in a more general model system. Try rotating one monomer relative
> to the other some arbitrary angle, or translating along something
> other than a principal moment of inertia and you will no longer
> recover the simple summation of the betas when you pull up the
> dimer or trimer in the standard Gaussian orientations. Here's where
> NLOPredict becomes useful.
>
> Hope this helps,
>
> Garth
>
>
>
> From: Stefano Tonzani [mailto:to...@gm...]
> Sent: Wed 7/18/2007 6:41 PM
> To: Simpson, Garth J
> Subject: Re: NLOpredict info
>
> Hello,
> There is still something that mystifies me in the functioning of
> NLOPredict.
> Namely: I tried, for a check against my code, the simple case of
> two thymines stacked at a distance of 4 Angstroms apart on the z-
> axis. The first thymine is in exactly the same geometry and
> orientation as the standard orientation of the Gaussian calculation.
> Therefore, what I do is input the C, O, N positions of the first
> thymine, and then the pdb of the stack. What I get is pretty
> strange, basically a reorientation of the beta tensor of a single
> thymine (if I reorient the arrow accordingly to the calculation for
> a single thymine I get the same beta components). Instead I should
> get the same relative intensities of the different components but
> doubled, since there are 2 thymines now, a thing that I recover
> from my code. Am I still missing something? If I have a stack of 3
> thymines, the signal trebles (after reorientation) so it looks like
> the first thymine is neglected.
> I thought that the C,O,N positions were used to transform to the
> pdb orientation, but in this case the C,O,N have the same exact
> positions as their counterparts in the first residue of the stack
> in the pdb file, so there should be no rotation. Once again, what
> does the transformation to the principal axes accomplish? The
> Gaussian standard orientation is already in the principal axes frame.
> The transformations I perform in my code are from the standard
> Gaussian orientation to the pdb orientation of the 1st residue
> (which is going to be the final orientation), then I sum up the
> various betas using the relations between their pdb coordinates.
> Thanks again for your patience,
>
> Stefano
>
> On 6/12/07, Simpson, Garth J <gsi...@pu... > wrote:
> Stefano,
>
> That works for me. If you have any troubles finding my office, call
> me at 765-496-3054. See you Thursday morning.
>
> Garth
> --------------------------
> Garth Simpson
> Associate Professor of Chemistry
> Purdue University
>
>
>
>
> ----- Original Message -----
> From: Stefano Tonzani <to...@gm... >
> To: Simpson, Garth J
> Sent: Tue Jun 12 10:15:32 2007
> Subject: Re: NLOpredict info
>
> Hello,
> so is it fine for me to come on Thursday? I should be getting there
> around 10:30 or so. Your office is BRWN 4170C, right?
> Let me know if this does not work for you,
>
> Thanks again,
>
> Stefano
>
>
> On 6/8/07, Stefano Tonzani < to...@gm...> wrote:
>
> Hello,
>
> Thanks a lot for the help!
> It would be probably best for me to meet at Purdue say
> Wednesday, Thursday or Friday, let me know what works best for you.
> I have, in fact, only tried planar diNMA, which I assumed
> was the one in the Supporting information, is this correct?
> My results for diNMA do not look so identical to the ones
> in your paper, but that might be a problem of linearly combining
> the two Amide A modes or something like that,
>
> Later,
>
> Stefano
>
>
>
> On 6/8/07, Simpson, Garth J < gsi...@pu...> wrote:
>
> Stefano,
>
>
>
> First, I would be happy to meet with you to discuss
> your scientific problem and the possible approaches in which
> NLOPredict could help address them. I will be in town all of next
> week – can you come down then? Alternatively, I'll be at the Mol-
> Spec conference. If you are attending as well, we could meet there
> to talk shop.
>
>
>
> Second, it sounds like your results are actually
> matching up pretty well qualitatively with those published in the
> supporting information of the JPC-B paper for the Amide A mode
> (Figure S1 of the attached), apart from a potential units
> conversion issue.
>
>
>
> Finally, assembly of the predicted NLO properties
> of di-NMA is not the best place to start because of the large
> degrees of freedom when selecting a coordinate system (nonplanar
> molecule of low symmetry). If my memory serves (Andy can correct me
> on this), we used the sum and differences of the monomer transition
> moments to define the coordinates in the twisted form of di-NMA,
> which is not trivial to reproduce. The planar structure is
> considerably simpler, but still significantly more complex than DNA
> (where the assumption of C(infinity) symmetry is reasonable). We
> had calculation of the NLO properties of DNA on the list of things
> to do at the Chautauqua, but we never quite got around to it. If
> you come down with the vibrational calculations in hand, we should
> be able to hammer it out in a day for all the isoforms of DNA no
> problem. That would certainly be easier than trying to do this by
> long distance.
>
>
>
> Garth
>
>
>
> ________________________________
>
> From: Stefano Tonzani
> [ mailto:to...@gm...]
> Sent: Thursday, June 07, 2007 12:22 PM
> To: Simpson, Garth J
> Subject: NLOpredict info
>
>
>
> Hello,
> sorry I couldn't come to the nonlinear optics
> Chautauqua, but I was out of the country at the time,
>
> I had some exchange with Andy Moad about the NMA/
> diNMA calculations you guys did a while ago, to validate the code I
> am developing. Here are the questions:
> 1) I did some Gaussian calculations like in a
> sample file I got from Andy for diNMA, to compare to the results in
> your JPCB 109 20009 ('05) in the supporting information. In
> particular I look at the Amide A mode (there are two of them, modes
> 50 and 51 in the file I am sending you) and basically I plug in the
> numbers from Gaussian (lines 9043-9056) and reorient the molecule
> as in your paper, with the dipole pointing at me and the principal
> molecular axis horizontal, in other words:
> theta=90, phi=139, psi=270
> This gives me a beta tensor for both modes with
> large xxz and zzz components (-15 and -18 for mode 50) and little
> else, which is very different from what is in the paper. Am I doing
> something wrong? Andy said that you guys probably added the two
> modes together in some way, but he did not remember the details.
> 2) Aside from that, I am trying to assemble the
> diNMA using the spectra of two separate NMA molecules. I use the
> pdb file attached and for the Gaussian calculations I also attach
> the output file. The orientation of one of the monomers in the
> "assembled" diNMA is identical to the Gaussian calculations for
> NMA, the 2nd is rotated like in the "real" diNMA molecule. This is
> the way it should be done, correct?
>
> Then I have a request: would it be possible for me
> to come to Purdue for a day so we could talk about these and other
> issues?
>
>
> Thank you very much again,
>
> Stefano
>
> On 4/6/07, Stefano Tonzani < to...@gm... <
> mailto:to...@gm...> > wrote:
>
> Hello,
> thanks for the heads-up. I will be more than happy
> to talk when you get back,
>
> Later,
>
> Stefano
>
> On 4/6/07, Andrew Moad < and...@ni... <
> mailto:and...@ni...> > wrote:
>
> I am leaving for the MRS conference in San
> Francisco. When I get back
> on April 16th I will be able to give your
> questions some better answers.
> In brief:
> 1) Don't use it. It is something I wrote
> just for me and am not
> comfortable with anyone else using it. I
> didn't realize that it was
> included in the program. It works, but
> requires a very specific gaussian
> calculation. I would be glad to explain it
> to you later. It is useless
> for vibrational SFG.
>
> 2) No script is needed. I get these from
> gaussian directly via the
> IOP(33/7)=1 keyword. This will give you the
> transition dipole and
> transition polarizability directly (lines
> 9439-9647 in NMA_002.log). I
> can give you my procedure when I get back.
>
> 4) There are several routes. One is to load
> a PDB file that has a helix
> in it. I'll attach mine ( alphahelix03.pdb).
>
> Sorry that I don't have more time this week.
> Andy
>
> Stefano Tonzani wrote:
> > Dear Andrew,
> > my name is Stefano Tonzani, I work at
> Northwestern University and I am
> > developing an approach that would use
> molecular dynamics calculations
> > to predict nonlinear optical spectra.
> > Since I am interested in modeling NLO
> experiments of DNA on surfaces,
> > and I recently met Garth Simpson and
> talked about the software you
> > guys have developed (NLOpredict) I wanted
> to ask you some questions
> > about it, if possible.
> > In particular, your software would be
> useful for me to validate what I
> > am currently doing, since there are no
> other computational apporaches
> > one can compare to. So the questions are:
> > 1) I saw that the package includes a
> script called gaussianParser.py
> > for SHG (or electronically resonant SFG).
> Does that work well? Garth
> > said he hasn't used it very much, but it
> would be useful for me, since
> > what I am developing is kind of similar,
> and of course it is much
> > easier to get stuff directly from ab
> initio files than calculate and
> > input them one by one, by hand, into a
> graphic interface.
> > 2) Is there an equivalent of that script
> for vibrationally resonant
> > SFG (since that is what I am interested in)?
> > 3) On that note, how did you obtain the
> results in your J. Phys. Chem
> > B 109, 20009 (2005) in particular for the
> vibrational SFG in the
> > supporting info? Is it possible to have
> some of the files used for the
> > dimer (not perturbation) calculation
> (like ab initio or scripts you
> > use to parse the ab initio)?
> > 4) I haven't really understood how to
> combine chromophores in that
> > code, for example if you have performed
> calculations for a molecule
> > and you want to build a helix of
> identical molecules.
> >
> > Thank you very much in advance,
> >
> > Best regards,
> >
> > Stefano Tonzani
>
>
>
>
> --
> Stefano Tonzani
> Department of Chemistry
> Northwestern University, 2145 Sheridan Rd.,
> Evanston, IL. 60208-3113 USA
> Phone: (847) 467 4989
> Fax: (847) 467 4996
> Office location: Room 4037, Center for
> Nanofabrication and Molecular Self Assembly, 2190 Campus Dr.
> Home Page: http://ucsu.colorado.edu/~tonzani/
> Newstyle/Newstyle.html <http://ucsu.colorado.edu/%7Etonzani/
> Newstyle/Newstyle.html>
>
|
|
From: Randy H. <he...@in...> - 2008-04-05 20:16:33
|
Begin forwarded message: > From: "Simpson, Garth J" <gsi...@pu...> > Date: November 29, 2007 9:51:02 AM EST > To: <ant...@ut...> > Cc: <he...@in...>, <rwa...@pu...> > Subject: FW: NLOPredict and Computations > > Anthony, > > > > In its current form, NLOPredict is designed to help visualize and > interpret the molecular tensors generation from computational and/ > or experimental measurements. We purposely separated the methods > used to generate the molecular tensor from the software program. > > > > With that being said, we have used alternative approaches to > calculate the resonant hyperpolarizabilities for frequency doubling > and mixing, based on sum over states expressions that are not built > in to Gaussian. As such, I am not as familiar with the method you > describe. Ron Wampler may know more, and he is copied on this > email. If you are calculating the tensors for frequency doubling, > the last two indices are interchangeable, such that only 18 values > need to be reported. This may very well be the reason only 18 > elements of the tensor are generated in the output file. You can > use NLOPredict to visualize this tensor by directly entering all 27 > elements through Parameters/Tensors/Adjust Beta, and visualize them > by selecting Parameters/Multiplication Order/Full Beta along with > View/AlphaMuVisibility/Surface. You may need to adjust the Tensor > Visual Scale to see the full surface. > > > > Hope this helps! > > > > Garth > > > > From: Randy Heiland [mailto:he...@in...] > Sent: Thursday, November 29, 2007 8:24 AM > To: Simpson, Garth J > Cc: Andrew Moad > Subject: Fwd: NLOPredict and Computations > > > > Hi Anthony, > > > > Your question falls more into the Chemistry (Gaussian) side of > things, so I'll fwd it on to the PI and original PhD student who > implemented things. > > > > I'd welcome being copied on any replies, > > -Randy > > > > > > > Begin forwarded message: > > > > > From: "Antonio De Crisci" <ant...@ut...> > > Date: November 28, 2007 3:38:48 PM EST > > To: <he...@in...> > > Subject: NLOPredict and Computations > > > > Hello Prof. Heiland, > > > > Im a grad student at the University of Toronto investigating the > NLO properties of single molecules. I came across your fascinating > paper J. Comp. Chem. 2007, 28, 1996 using NLOPredict to visualize > the process. In your computations (using G98), were you able to > calculate all 27 beta tensors? Im running into an issue were using > Gaussian 03 to calculate the beta tensors, but the output file only > contains 18 tensors using the keywords 'polar' and 'CPHF=ReFreq' at > 1064nm. > > > > Would you be able to tell me how you found all 27 tensors from > calculation? > > > > Thanks for your time, > > Anthony > > > > > > [The attachment 2007 ACR Visualization of NLO.pdf has been manually > removed] > |
|
From: Randy H. <he...@in...> - 2008-03-22 22:58:38
|
Please ignore. |
