From smori@chem.s.u-tokyo.ac.jp Fri Dec 8 00:12:01 1995 Received: from utsc.s.u-tokyo.ac.jp for smori@chem.s.u-tokyo.ac.jp by www.ccl.net (8.6.10/950822.1) id AAA28015; Fri, 8 Dec 1995 00:03:53 -0500 Received: from [133.11.6.112] (smori.chem.s.u-tokyo.ac.jp) by utsc.s.u-tokyo.ac.jp (5.67+1.6W/TISN-1.3/R2) id AA10142; Fri, 8 Dec 95 14:03:34 JST Message-Id: <9512080503.AA10142@utsc.s.u-tokyo.ac.jp> Date: Fri, 8 Dec 1995 14:08:02 +0900 To: chemistry@www.ccl.net From: smori@chem.s.u-tokyo.ac.jp (Mori Seiji) X-Sender: smori@chem.s.u-tokyo.ac.jp (Unverified) Subject: Summary for NMR chemical shift calculations Mime-Version: 1.0 Content-Type: text/plain; charset=iso-2022-jp X-Mailer: Eudora-J(1.3.8.5-J13) Dear netters, Thank you for several replies about my questions for the method of NMR chemical shift, and so on. I think the reliability of IGAIM , CSGT in G94 packages were not known very much thus we have to check these methods comparing with known methods. The systematic review of solvent effects, polarity and coordination with donating solvent such as ether have still be awaited. As for the concept of anitrosopy and isotropy, thank you for sending me many replies. I show two original questions and corresponding replies. Thank you very much for your help. Sincerely yours, Seiji Mori ---my original question 1 --- Dear everybody, Recently, many letters about NMR chemical shift calculation were posted, and I am also interested in NMR shifts in organic and organometallic complexes in solution. I have several questions, 1. How is the reliablities of 1.1 methods (GIAO, IGAIM, CSGT, IGLO) 1.2 theory (HF,DFT,MP2...CC) and basis-sets 1.3 How is the solvent effect of polarity on the NMR shift, coupling constant (I think it is not related calcluation , probably it is a fundamental question about NMR)? 2. Do you know the program or references which one can calculate not only chemical shift but also the coupling constant ( for example, 1H-1H and 13C-13C) and NOE in quantum chemistry other than G94? As to 1H-1H coupling constant, many programs such as PCMODEL was supported but its calculation is not QC. If you have comments and indicate the references, would you please send me ? I will summarize replies. Thanks in advance, Seiji Mori ------------- Replies ---1--- Received: by zinc.chem.ucalgary.ca (AIX 3.2/UCB 5.64/4.03) id AA10813; Fri, 24 Nov 1995 15:29:30 -0700 From: schrecke@zinc.chem.ucalgary.ca Hi Seiji, you have quite a lot of questions in your posting ... I shall try to give you a few hints, with no intention to be comprehensive. > > 1. How is the reliablities of > 1.1 methods (GIAO, IGAIM, CSGT, IGLO) > 1.2 theory (HF,DFT,MP2...CC) and basis-sets > 1.3 How is the solvent effect of polarity on the NMR shift, coupling > constant > (I think it is not related calcluation , probably it is a fundamental > question about NMR)? Good references to start with are three nice reviews, one new, two are older: D.B.Chesnut: Annual Reports on NMR Spectroscopy, published by Academic Press, Vol.21,1989, p. 51 and ibid, Vol. 29, 1994, p.71. H.Fukui: Magn. Res. Rev. 1987, vol. 11, 205 Further, very comprehensive reviews are contained in an annual series, since 1980 written every year by C.Jameson: in Specialist Periodical Report on Nuclear Magnetic Resonance, Vol.8,1980-... (edited by G.A.Webb, and published by the Royal Society of Chemistry, Cambridge) Now more specific comments on your various questions. 1.1 Methods with distributed gauge origins (IGLO, LORG, GIAO,...) are generally preferable (for a given basis set) over methods with a common gauge origin. For examples, see the various reviews. GIAO seems to converge slightly faster with the size of the basis set then IGLO (see Wolinski, Hinton, Pulay, JACS 1990, 112, 8251). This would also make sense from a theoretical point of view. I don't know off hand about IGAIM, CSGT. 1.2 Correlation. there are cases where correlation is necessary, e.g., M.Buehl et al. J.Phys.Chem. 99, 4000 (1995), and M.Buehl et al. Chem.Phys.Lett. 241 (1995), 248. These papers contain examples where MP2 fails completely to predict the shift while DFT and CCSD are able to achieve quantitative results. There are however many other cases where the HF level works just fine (all the respective literature up to -- say -- five years ago). Basis sets. See about any paper with calculated shifts. E.g., the one by Pulay et al. mentioned above, or various papers by the Kutzelnigg group, or the various reviews. 1.3 Solvation effects can be considerable, indeed. Look for experiments that were done in the gas phase (in addition to solution studies). A review of gas phase NMR was written by C. Jameson: Chem.Rev.1991, 91, 1375. > 2. Do you know the program or references which one can calculate not only > chemical shift but also > the coupling constant > ( for example, 1H-1H and 13C-13C) and NOE in quantum chemistry other than > G94? As to 1H-1H > coupling constant, > many programs such as PCMODEL was supported but its calculation is not QC. Coupling constants are MUCH tougher to calculate than chemical shifts. The reason is (in my opinion) that you are dealing with operators that evaluate the electron density in essentially just one point of space rather than over an entire region of space. Annual reviews of calculations are contained in the same series of books that was metioned before: Specialist Periodical Report on Nuclear Magnetic Resonance Calculations have been done recently, e.g., by Erikson/Malkin/Salahub et al. but I don't have the references on hand right now. I am looking forward to your summary! Yours, Georg ---2--- From: "Steve Gwaltney" Date: Mon, 27 Nov 1995 10:51:20 -0500 > 2. Do you know the program or references which one can calculate not only > chemical shift but also > the coupling constant > ( for example, 1H-1H and 13C-13C) and NOE in quantum chemistry other than > G94? As to 1H-1H ACES II can calculate chemical shifts using GIAO's and MBPT(2) wavefunctions. ACES II can also calculate CCSD coupling constants. For information about ACES II contact aces2@qtp.ufl.edu. I believe TURBOMOLE can also calculate chemical shifts, but you should check with them to make sure. Steve ---3--- From: wagenert@Mailer.Uni-Marburg.DE (Wagener Thomas) Subject: CCL:NMR shift and coupling constant calculation Date: Mon, 27 Nov 1995 14:12:47 +0100 (CET) Dear Seiji Mori, since I have finished my thesis on the ab-initio calculation a few month ago, I hopefully can help you with some of your questions. > 1. How is the reliablities of > 1.1 methods (GIAO, IGAIM, CSGT, IGLO) > 1.2 theory (HF,DFT,MP2...CC) and basis-sets > 1.3 How is the solvent effect of polarity on the NMR shift, coupling > constant I had only the opportunity to work with GIAO and IGLO so I can only make a statement on these methods. Basically IGLO and GIAO are quite reliable for the calculation of "standard" organic compounds though the accuracy of the methods is not as high as the experimental accuracy. A rough estimate would be an accuracy of about 5ppm for 13C chemical shifts - probably better. Taking into account that you are comparing gas phase values at absolute zero with values taken from a sample in solution at a temperature of, say, 200-300K that is probably as close as you can get. The computational effort for the IGLO method is somewhat higher than for the GIAO calculation (at HF-level) though the information you get from IGLO is more extensive. Because IGLO works with localized molecular orbitals (LMOs) you get an anlysis of the contribution of the LMOs to the chemical shift while you only get the shielding constant out of the GIAO calculation. IGLO is available at HF level and there is a DFT version using the IGLO code to calculate chemical shifts. The GIAO method in MO theory is available up to CC level though the computational effort is extremely high so the calculation of the shielding constant at such a highly coorelated level is only possible for small molecules. The GAUSSIAN94 manual states that the GIAO module of the program can cope with DFT but I have no experience with that. > 2. Do you know the program or references which one can calculate not only > chemical shift but also > the coupling constant > ( for example, 1H-1H and 13C-13C) and NOE in quantum chemistry other than > G94? As to 1H-1H coupling constant, As far as I know (and I really may not be up to date) the only programs capable of calculating coupling constants is DeMon with DFT-IGLO and ACESII (at MP2 level). Some useful references are: - IGLO: Kutzelnigg, W.; Fleischer, U.; Schindler, M.:"The IGLO-Method: Ab- initio Calculation and Interpretation of NMR chemical Shifts and magnetic Susceptibilities", NMR Basic Princ. Prog., Vol. 23, Springer-Verlag, Berlin Heidelberg, 1990, 165 - DFT-IGLO: Malkin, M. G.; Malkina, O. L.; Casida, M. E.; Salahub, D. R., JACS, 116(1994), 5898 - HF-GIAO: Wolinski, K.; Hilton, J. F.; Pulay, P., JACS, 112(1990), 8251 - MP2-GIAO: Gauss, J., Chem. Phys. Lett., 191(1992), 614 Hope, this was of some help Thomas Wagener ---my original questions 2 about isotropies...--- Dear Sirs, I posted to CCL about the methods of NMR calculation a fey days ago , and had an other question. In G94, I tried to calculate the chemical shifts , there are magnetic shieldings in the output, however, no chemical shift but isotropic and anitrosopy. I examine meaning of these terms in the user's manual of Gaussian 94 and several original papers,for example, M. Schindler, et al. JCP, 1982, 76, 1919 and Pulay et al. JACS, 1990, 112, 8251, and so on which there is replies of my previous questions , but I only know that isotropic terms is "likely" to be corresponding to the chemical shifts which was experimentally observed. I had had experiments in organic chemistry, and I am now carrying out the ab initio calculations but I amn't much familiar to know the theory of NMR spectroscopy. Would you please tell me the physical meaning of isotropic and anitrosopy, and which is an equivalent of the experimental chemical shift? It is better that you show the references. Sincerely yours, Seiji Mori ---- replies---- ---1--- Date: Tue, 28 Nov 1995 11:10:24 +0100 (MEZ) From: Alexander Christian Backes Konichiwa Mori-san, the chemical shift is always a tensor (a symmetric 3 x 3 matrix). So, this tensor describes the interchange between several NMR-active nuclei. The outerdiagonal elements are only important in spectra of the solid phase because there the position of the molecules and therefore the observed nuclei are defined and the shift is therefor anisotropic (CSA = Chemnical shift anisotropy). In solution the motion of the molecules is randomized, so the NMR-spectrometer observes the middled shift. This means, the trace of the matrix is relevant (the trace is the sum of the diagonal elemnts of the matrix, divided by three, in case of the 3 x 3 matrix). So, in G94 the calculated isotropic shift is the relevant one for solution. Of course you need a standard for the chemical shift; G94 doesn't know to which reference substance the shifts are related to. So you have to calculate TMS at exactly the same level as the calculated molecule, and then you have to subtract the isotropic data of your molecule from the isotropic data of TMS. (This you have to do each time you use another method and basis set for geometry optimization and NMR-calculations). But be careful - the input geometry of TMS must have Td-symmetry. And therefore you have to precalculate TMS at MP2-level to make sure it gets Td-Symmetry. This geometry you can use as standard input for any of the following calculations of geometry and NMR-parameters. References to the first point you can find in any book dealing with the theory of NMR-spectroscopy. The second point has references first our own experience and second papers which will be published soon. Greetings from Germany, Alexander. ---2--- From: lohrenz@oci.unizh.ch Date: Tue, 28 Nov 1995 13:41:54 +0100 (MET) Hi Seiji, your are probably interested in the isotropic shielding. To convert from shielding to chemical shift (in delta scale) you have to calculate the shielding of a reference compound, say for example TMS for 1H-, 13C- or Si NMR and substract the calculated shielding values of your compound. In principle it is also possible to use smaller references like CH4. In this case you have to take care of the delta-value of this compound (eg. -2.3 ppm for 13C in CH4). Just make sure that you use identical basissets for the calculations. The anisotropic values can sometimes be measured in CPMAS (solid state nmr). They give the shielding along the princible axis of your system. For spherical homogeneous compounds like CH4 you will notice that the anisotropic shieldings are the same as the isotropic. Since solution nmr gives an average of all possible orientations you alway get the isotropic shieldings. Only in cases where the molecule can be oriented (no rotation) like in the solid state, it is possible to measure the three anisotropic values independently. John ---3--- Date: 28 Nov 1995 14:31:34 +0000 From: "E.A.Moore (Elaine Moore)" G94 and all other ab initio programs used for calculating chemical shifts calula te chemical shielding. Experiment measures the chemical shift which is the chemi cal shielding relative to that of a reference molecule such as TMS (tetramethyls ilane). The relative isotropic shielding is the quantity normally observed, for example in liquid samples. The anisotropy can be obtained from solid state measu rements and can be important in relaxation. Chemical shielding is a tensor quant ity and G94 gives you all 9 components in the frame you have chosen. The isotrop ic shielding is (sigma xx + sigma yy + sigma zz)/3. The anisotropy is sigma 33 - (sigma 11 + sigma 22)/2 where 1,2,3 refer to the principal axis frame and sigm a 33> sigma 22 >sigma 11. The chemical shift is sigma (reference)- sigma (compou nd) where sigma is the isotropic shielding. Hope this is some help Elaine A. Moore Chemistry Dept. The Open University UK ---4--- From: evaldera@inti.ivic.ve (Elmer Valderrama) Date: Tue, 28 Nov 1995 09:28:48 +0000 (GMT) Hi, Anisotropy only shows up in ordered media (crystalls, liquid-crystals). Organic compounds may be in solid state. As such, -e.g. not disolved in any solvent-, they may be "measured" in a RMN experiment. If, to this end, a direction of the crystal have been chosen, it may happen that the shift measured in the perpendicular to this chosen direction shows different. Then you'll have a diference which is called anisotropy. Moreover, -and for this it's better to read the reference below-, since the chem. shift actually is a tensor you'd have to compute three components (the diagonal) relative to the chosen direction in the crystal. In isotropic media (a mix of crystalls, no direction can be especified, or in liquid/solved sample) the shift is just the common measured quantity which physically is equal to one third of the trace of the above mencioned tensor. ..my two cents, hope it helps. Elmer Valderrama [1] K. Eichele, et al. "Phosphorus-31 Chemical Shift of Phosphinidene Ligands in Ruthenium Carbonyl Cluster Compounds: A 31-P Single Crystal and CP/MAS-NMR Study" J.Am.Chem.Soc. 1995,117,6961-6969 [and ref there in] (This work includes an orbital study using CACAO and a qualitative application of Ramsey's theory) ---5--- From: willsd@APPSTATE.BITNET To: Mori Seiji Date: Tue, 28 Nov 1995 10:51:13 -0400 (EDT) Seiji: I ran into this proble recently and have a suggestion: 1) You need the isotropic part of the shielding tensor. 2) Chemical shifts can be calculated from this isotropic part using some simple theory: I recommend the information in chapter 2 of : Multinuclear NMR, J. Mason Ed., Plenum Press, New York, 1987. You will also need the isotropic shielding for the chemical shift standard for your nucleus. (In my case it was (EtO)2.BF3 for 11B NMR). Steve Williams Chemistry Appalachian State University Boone, NC 28608 willsd@appstate.edu ---6--- Date: Wed, 29 Nov 1995 12:08:48 +1000 To: smori@chem.s.u-tokyo.ac.jp (Mori Seiji) From: hughc@extro.ucc.su.OZ.AU (Hugh Capper) Dear Mori, isotropic = having the same physical arrangement of moieties around the atom of interest which in this case is the proton ( if you are doing proton specroscopy) anitrosopy = is the reverse ie having different moities which cause diffences around the proton of interest. anitrosopy is reflected in the chemical shift. As for a reference a good basic text is Fundamentals of NMR Spectroscopy by Derome. best wishes, Hugh Capper Institute for Magnetic Resonance Research University of Sydney NSW 2006 Australia ---7--- From: mjf@biosym.com (Mark J Forster ) Date: Tue, 28 Nov 1995 10:03:11 -0800 ( You can see in CCL archives.) --- #################################################### Seiji Mori Graduate student in Nakamura Laboratory Department of Chemistry The University of Tokyo Hongo 7-3-1, Bunkyou-ku, Tokyo 113, JAPAN. email:smori@chem.s.u-tokyo.ac.jp #################################################### From smori@chem.s.u-tokyo.ac.jp Fri Dec 8 00:12:12 1995 Received: from utsc.s.u-tokyo.ac.jp for smori@chem.s.u-tokyo.ac.jp by www.ccl.net (8.6.10/950822.1) id AAA27998; Fri, 8 Dec 1995 00:02:05 -0500 Received: from [133.11.6.112] (smori.chem.s.u-tokyo.ac.jp) by utsc.s.u-tokyo.ac.jp (5.67+1.6W/TISN-1.3/R2) id AA10121; Fri, 8 Dec 95 14:01:38 JST Message-Id: <9512080501.AA10121@utsc.s.u-tokyo.ac.jp> Date: Fri, 8 Dec 1995 14:06:07 +0900 To: chemistry@www.ccl.net From: smori@chem.s.u-tokyo.ac.jp (Mori Seiji) X-Sender: smori@chem.s.u-tokyo.ac.jp (Unverified) Subject: Summary of NMR chemical shift calculatons Cc: smori@utsc.s.u-tokyo.ac.jp Mime-Version: 1.0 Content-Type: text/plain; charset=iso-2022-jp X-Mailer: Eudora-J(1.3.8.5-J13) Dear netters, Thank you for several replies about my questions for the method of NMR chemical shift, and so on. I think the reliability of IGAIM , CSGT in G94 packages were not known very much thus we have to check these methods comparing with known methods. The systematic review of solvent effects, polarity and coordination with donating solvent such as ether have still be awaited. As for the concept of anitrosopy and isotropy, thank you for sending me many replies. I show two original questions and corresponding replies. Thank you very much for your help. Sincerely yours, Seiji Mori ---my original question 1 --- Dear everybody, Recently, many letters about NMR chemical shift calculation were posted, and I am also interested in NMR shifts in organic and organometallic complexes in solution. I have several questions, 1. How is the reliablities of 1.1 methods (GIAO, IGAIM, CSGT, IGLO) 1.2 theory (HF,DFT,MP2...CC) and basis-sets 1.3 How is the solvent effect of polarity on the NMR shift, coupling constant (I think it is not related calcluation , probably it is a fundamental question about NMR)? 2. Do you know the program or references which one can calculate not only chemical shift but also the coupling constant ( for example, 1H-1H and 13C-13C) and NOE in quantum chemistry other than G94? As to 1H-1H coupling constant, many programs such as PCMODEL was supported but its calculation is not QC. If you have comments and indicate the references, would you please send me ? I will summarize replies. Thanks in advance, Seiji Mori ------------- Replies ---1--- Received: by zinc.chem.ucalgary.ca (AIX 3.2/UCB 5.64/4.03) id AA10813; Fri, 24 Nov 1995 15:29:30 -0700 From: schrecke@zinc.chem.ucalgary.ca Hi Seiji, you have quite a lot of questions in your posting ... I shall try to give you a few hints, with no intention to be comprehensive. > > 1. How is the reliablities of > 1.1 methods (GIAO, IGAIM, CSGT, IGLO) > 1.2 theory (HF,DFT,MP2...CC) and basis-sets > 1.3 How is the solvent effect of polarity on the NMR shift, coupling > constant > (I think it is not related calcluation , probably it is a fundamental > question about NMR)? Good references to start with are three nice reviews, one new, two are older: D.B.Chesnut: Annual Reports on NMR Spectroscopy, published by Academic Press, Vol.21,1989, p. 51 and ibid, Vol. 29, 1994, p.71. H.Fukui: Magn. Res. Rev. 1987, vol. 11, 205 Further, very comprehensive reviews are contained in an annual series, since 1980 written every year by C.Jameson: in Specialist Periodical Report on Nuclear Magnetic Resonance, Vol.8,1980-... (edited by G.A.Webb, and published by the Royal Society of Chemistry, Cambridge) Now more specific comments on your various questions. 1.1 Methods with distributed gauge origins (IGLO, LORG, GIAO,...) are generally preferable (for a given basis set) over methods with a common gauge origin. For examples, see the various reviews. GIAO seems to converge slightly faster with the size of the basis set then IGLO (see Wolinski, Hinton, Pulay, JACS 1990, 112, 8251). This would also make sense from a theoretical point of view. I don't know off hand about IGAIM, CSGT. 1.2 Correlation. there are cases where correlation is necessary, e.g., M.Buehl et al. J.Phys.Chem. 99, 4000 (1995), and M.Buehl et al. Chem.Phys.Lett. 241 (1995), 248. These papers contain examples where MP2 fails completely to predict the shift while DFT and CCSD are able to achieve quantitative results. There are however many other cases where the HF level works just fine (all the respective literature up to -- say -- five years ago). Basis sets. See about any paper with calculated shifts. E.g., the one by Pulay et al. mentioned above, or various papers by the Kutzelnigg group, or the various reviews. 1.3 Solvation effects can be considerable, indeed. Look for experiments that were done in the gas phase (in addition to solution studies). A review of gas phase NMR was written by C. Jameson: Chem.Rev.1991, 91, 1375. > 2. Do you know the program or references which one can calculate not only > chemical shift but also > the coupling constant > ( for example, 1H-1H and 13C-13C) and NOE in quantum chemistry other than > G94? As to 1H-1H > coupling constant, > many programs such as PCMODEL was supported but its calculation is not QC. Coupling constants are MUCH tougher to calculate than chemical shifts. The reason is (in my opinion) that you are dealing with operators that evaluate the electron density in essentially just one point of space rather than over an entire region of space. Annual reviews of calculations are contained in the same series of books that was metioned before: Specialist Periodical Report on Nuclear Magnetic Resonance Calculations have been done recently, e.g., by Erikson/Malkin/Salahub et al. but I don't have the references on hand right now. I am looking forward to your summary! Yours, Georg ---2--- From: "Steve Gwaltney" Date: Mon, 27 Nov 1995 10:51:20 -0500 > 2. Do you know the program or references which one can calculate not only > chemical shift but also > the coupling constant > ( for example, 1H-1H and 13C-13C) and NOE in quantum chemistry other than > G94? As to 1H-1H ACES II can calculate chemical shifts using GIAO's and MBPT(2) wavefunctions. ACES II can also calculate CCSD coupling constants. For information about ACES II contact aces2@qtp.ufl.edu. I believe TURBOMOLE can also calculate chemical shifts, but you should check with them to make sure. Steve ---3--- From: wagenert@Mailer.Uni-Marburg.DE (Wagener Thomas) Subject: CCL:NMR shift and coupling constant calculation Date: Mon, 27 Nov 1995 14:12:47 +0100 (CET) Dear Seiji Mori, since I have finished my thesis on the ab-initio calculation a few month ago, I hopefully can help you with some of your questions. > 1. How is the reliablities of > 1.1 methods (GIAO, IGAIM, CSGT, IGLO) > 1.2 theory (HF,DFT,MP2...CC) and basis-sets > 1.3 How is the solvent effect of polarity on the NMR shift, coupling > constant I had only the opportunity to work with GIAO and IGLO so I can only make a statement on these methods. Basically IGLO and GIAO are quite reliable for the calculation of "standard" organic compounds though the accuracy of the methods is not as high as the experimental accuracy. A rough estimate would be an accuracy of about 5ppm for 13C chemical shifts - probably better. Taking into account that you are comparing gas phase values at absolute zero with values taken from a sample in solution at a temperature of, say, 200-300K that is probably as close as you can get. The computational effort for the IGLO method is somewhat higher than for the GIAO calculation (at HF-level) though the information you get from IGLO is more extensive. Because IGLO works with localized molecular orbitals (LMOs) you get an anlysis of the contribution of the LMOs to the chemical shift while you only get the shielding constant out of the GIAO calculation. IGLO is available at HF level and there is a DFT version using the IGLO code to calculate chemical shifts. The GIAO method in MO theory is available up to CC level though the computational effort is extremely high so the calculation of the shielding constant at such a highly coorelated level is only possible for small molecules. The GAUSSIAN94 manual states that the GIAO module of the program can cope with DFT but I have no experience with that. > 2. Do you know the program or references which one can calculate not only > chemical shift but also > the coupling constant > ( for example, 1H-1H and 13C-13C) and NOE in quantum chemistry other than > G94? As to 1H-1H coupling constant, As far as I know (and I really may not be up to date) the only programs capable of calculating coupling constants is DeMon with DFT-IGLO and ACESII (at MP2 level). Some useful references are: - IGLO: Kutzelnigg, W.; Fleischer, U.; Schindler, M.:"The IGLO-Method: Ab- initio Calculation and Interpretation of NMR chemical Shifts and magnetic Susceptibilities", NMR Basic Princ. Prog., Vol. 23, Springer-Verlag, Berlin Heidelberg, 1990, 165 - DFT-IGLO: Malkin, M. G.; Malkina, O. L.; Casida, M. E.; Salahub, D. R., JACS, 116(1994), 5898 - HF-GIAO: Wolinski, K.; Hilton, J. F.; Pulay, P., JACS, 112(1990), 8251 - MP2-GIAO: Gauss, J., Chem. Phys. Lett., 191(1992), 614 Hope, this was of some help Thomas Wagener ---my original questions 2 about isotropies...--- Dear Sirs, I posted to CCL about the methods of NMR calculation a fey days ago , and had an other question. In G94, I tried to calculate the chemical shifts , there are magnetic shieldings in the output, however, no chemical shift but isotropic and anitrosopy. I examine meaning of these terms in the user's manual of Gaussian 94 and several original papers,for example, M. Schindler, et al. JCP, 1982, 76, 1919 and Pulay et al. JACS, 1990, 112, 8251, and so on which there is replies of my previous questions , but I only know that isotropic terms is "likely" to be corresponding to the chemical shifts which was experimentally observed. I had had experiments in organic chemistry, and I am now carrying out the ab initio calculations but I amn't much familiar to know the theory of NMR spectroscopy. Would you please tell me the physical meaning of isotropic and anitrosopy, and which is an equivalent of the experimental chemical shift? It is better that you show the references. Sincerely yours, Seiji Mori ---- replies---- ---1--- Date: Tue, 28 Nov 1995 11:10:24 +0100 (MEZ) From: Alexander Christian Backes Konichiwa Mori-san, the chemical shift is always a tensor (a symmetric 3 x 3 matrix). So, this tensor describes the interchange between several NMR-active nuclei. The outerdiagonal elements are only important in spectra of the solid phase because there the position of the molecules and therefore the observed nuclei are defined and the shift is therefor anisotropic (CSA = Chemnical shift anisotropy). In solution the motion of the molecules is randomized, so the NMR-spectrometer observes the middled shift. This means, the trace of the matrix is relevant (the trace is the sum of the diagonal elemnts of the matrix, divided by three, in case of the 3 x 3 matrix). So, in G94 the calculated isotropic shift is the relevant one for solution. Of course you need a standard for the chemical shift; G94 doesn't know to which reference substance the shifts are related to. So you have to calculate TMS at exactly the same level as the calculated molecule, and then you have to subtract the isotropic data of your molecule from the isotropic data of TMS. (This you have to do each time you use another method and basis set for geometry optimization and NMR-calculations). But be careful - the input geometry of TMS must have Td-symmetry. And therefore you have to precalculate TMS at MP2-level to make sure it gets Td-Symmetry. This geometry you can use as standard input for any of the following calculations of geometry and NMR-parameters. References to the first point you can find in any book dealing with the theory of NMR-spectroscopy. The second point has references first our own experience and second papers which will be published soon. Greetings from Germany, Alexander. ---2--- From: lohrenz@oci.unizh.ch Date: Tue, 28 Nov 1995 13:41:54 +0100 (MET) Hi Seiji, your are probably interested in the isotropic shielding. To convert from shielding to chemical shift (in delta scale) you have to calculate the shielding of a reference compound, say for example TMS for 1H-, 13C- or Si NMR and substract the calculated shielding values of your compound. In principle it is also possible to use smaller references like CH4. In this case you have to take care of the delta-value of this compound (eg. -2.3 ppm for 13C in CH4). Just make sure that you use identical basissets for the calculations. The anisotropic values can sometimes be measured in CPMAS (solid state nmr). They give the shielding along the princible axis of your system. For spherical homogeneous compounds like CH4 you will notice that the anisotropic shieldings are the same as the isotropic. Since solution nmr gives an average of all possible orientations you alway get the isotropic shieldings. Only in cases where the molecule can be oriented (no rotation) like in the solid state, it is possible to measure the three anisotropic values independently. John ---3--- Date: 28 Nov 1995 14:31:34 +0000 From: "E.A.Moore (Elaine Moore)" G94 and all other ab initio programs used for calculating chemical shifts calula te chemical shielding. Experiment measures the chemical shift which is the chemi cal shielding relative to that of a reference molecule such as TMS (tetramethyls ilane). The relative isotropic shielding is the quantity normally observed, for example in liquid samples. The anisotropy can be obtained from solid state measu rements and can be important in relaxation. Chemical shielding is a tensor quant ity and G94 gives you all 9 components in the frame you have chosen. The isotrop ic shielding is (sigma xx + sigma yy + sigma zz)/3. The anisotropy is sigma 33 - (sigma 11 + sigma 22)/2 where 1,2,3 refer to the principal axis frame and sigm a 33> sigma 22 >sigma 11. The chemical shift is sigma (reference)- sigma (compou nd) where sigma is the isotropic shielding. Hope this is some help Elaine A. Moore Chemistry Dept. The Open University UK ---4--- From: evaldera@inti.ivic.ve (Elmer Valderrama) Date: Tue, 28 Nov 1995 09:28:48 +0000 (GMT) Hi, Anisotropy only shows up in ordered media (crystalls, liquid-crystals). Organic compounds may be in solid state. As such, -e.g. not disolved in any solvent-, they may be "measured" in a RMN experiment. If, to this end, a direction of the crystal have been chosen, it may happen that the shift measured in the perpendicular to this chosen direction shows different. Then you'll have a diference which is called anisotropy. Moreover, -and for this it's better to read the reference below-, since the chem. shift actually is a tensor you'd have to compute three components (the diagonal) relative to the chosen direction in the crystal. In isotropic media (a mix of crystalls, no direction can be especified, or in liquid/solved sample) the shift is just the common measured quantity which physically is equal to one third of the trace of the above mencioned tensor. ..my two cents, hope it helps. Elmer Valderrama [1] K. Eichele, et al. "Phosphorus-31 Chemical Shift of Phosphinidene Ligands in Ruthenium Carbonyl Cluster Compounds: A 31-P Single Crystal and CP/MAS-NMR Study" J.Am.Chem.Soc. 1995,117,6961-6969 [and ref there in] (This work includes an orbital study using CACAO and a qualitative application of Ramsey's theory) ---5--- From: willsd@APPSTATE.BITNET To: Mori Seiji Date: Tue, 28 Nov 1995 10:51:13 -0400 (EDT) Seiji: I ran into this proble recently and have a suggestion: 1) You need the isotropic part of the shielding tensor. 2) Chemical shifts can be calculated from this isotropic part using some simple theory: I recommend the information in chapter 2 of : Multinuclear NMR, J. Mason Ed., Plenum Press, New York, 1987. You will also need the isotropic shielding for the chemical shift standard for your nucleus. (In my case it was (EtO)2.BF3 for 11B NMR). Steve Williams Chemistry Appalachian State University Boone, NC 28608 willsd@appstate.edu ---6--- Date: Wed, 29 Nov 1995 12:08:48 +1000 To: smori@chem.s.u-tokyo.ac.jp (Mori Seiji) From: hughc@extro.ucc.su.OZ.AU (Hugh Capper) Dear Mori, isotropic = having the same physical arrangement of moieties around the atom of interest which in this case is the proton ( if you are doing proton specroscopy) anitrosopy = is the reverse ie having different moities which cause diffences around the proton of interest. anitrosopy is reflected in the chemical shift. As for a reference a good basic text is Fundamentals of NMR Spectroscopy by Derome. best wishes, Hugh Capper Institute for Magnetic Resonance Research University of Sydney NSW 2006 Australia ---7--- From: mjf@biosym.com (Mark J Forster ) Date: Tue, 28 Nov 1995 10:03:11 -0800 ( You can see in CCL archives.) --- ********************************************************** 　 　　森　聖治 東京大学大学院理学系研究科 物理有機化学講座中村研究室 D1 電話番号：03-3812-2111　内線4369 FAX：　　03-3812-8099 email:smori@chem.s.u-tokyo.ac.jp ********************************************************** From smori@chem.s.u-tokyo.ac.jp Fri Dec 8 01:56:45 1995 Received: from utsc.s.u-tokyo.ac.jp for smori@chem.s.u-tokyo.ac.jp by www.ccl.net (8.6.10/950822.1) id BAA29686; Fri, 8 Dec 1995 01:54:13 -0500 Received: from [133.11.6.112] (smori.chem.s.u-tokyo.ac.jp) by utsc.s.u-tokyo.ac.jp (5.67+1.6W/TISN-1.3/R2) id AA10957; Fri, 8 Dec 95 15:53:53 JST Message-Id: <9512080653.AA10957@utsc.s.u-tokyo.ac.jp> Date: Fri, 8 Dec 1995 15:58:22 +0900 To: chemistry@www.ccl.net From: smori@chem.s.u-tokyo.ac.jp (Mori Seiji) X-Sender: smori@chem.s.u-tokyo.ac.jp (Unverified) Subject: Summary for NMR isotropies Cc: smori@utsc.s.u-tokyo.ac.jp Mime-Version: 1.0 Content-Type: text/plain; charset=iso-2022-jp X-Mailer: Eudora-J(1.3.8.5-J13) Dear netters, In the meanwhile, as for the concept of anitrosopy and isotropy, thank you for sending me many replies. I WILL study these concepts and theory. Thank you very much for your help. Sincerely yours, Seiji Mori ---my original questions about isotropy and anisotropy--- Dear Sirs, I posted to CCL about the methods of NMR calculation a fey days ago , and had an other question. In G94, I tried to calculate the chemical shifts , there are magnetic shieldings in the output, however, no chemical shift but isotropic and anitrosopy. I examine meaning of these terms in the user's manual of Gaussian 94 and several original papers,for example, M. Schindler, et al. JCP, 1982, 76, 1919 and Pulay et al. JACS, 1990, 112, 8251, and so on which there is replies of my previous questions , but I only know that isotropic terms is "likely" to be corresponding to the chemical shifts which was experimentally observed. I had had experiments in organic chemistry, and I am now carrying out the ab initio calculations but I amn't much familiar to know the theory of NMR spectroscopy. Would you please tell me the physical meaning of isotropic and anitrosopy, and which is an equivalent of the experimental chemical shift? It is better that you show the references. Sincerely yours, Seiji Mori ---- replies---- ---1--- Date: Tue, 28 Nov 1995 11:10:24 +0100 (MEZ) From: Alexander Christian Backes Konichiwa Mori-san, (S. Mori wrote: In correct in Japanese, Kon-nichiwa. ^^:)) the chemical shift is always a tensor (a symmetric 3 x 3 matrix). So, this tensor describes the interchange between several NMR-active nuclei. The outerdiagonal elements are only important in spectra of the solid phase because there the position of the molecules and therefore the observed nuclei are defined and the shift is therefor anisotropic (CSA = Chemnical shift anisotropy). In solution the motion of the molecules is randomized, so the NMR-spectrometer observes the middled shift. This means, the trace of the matrix is relevant (the trace is the sum of the diagonal elemnts of the matrix, divided by three, in case of the 3 x 3 matrix). So, in G94 the calculated isotropic shift is the relevant one for solution. Of course you need a standard for the chemical shift; G94 doesn't know to which reference substance the shifts are related to. So you have to calculate TMS at exactly the same level as the calculated molecule, and then you have to subtract the isotropic data of your molecule from the isotropic data of TMS. (This you have to do each time you use another method and basis set for geometry optimization and NMR-calculations). But be careful - the input geometry of TMS must have Td-symmetry. And therefore you have to precalculate TMS at MP2-level to make sure it gets Td-Symmetry. This geometry you can use as standard input for any of the following calculations of geometry and NMR-parameters. References to the first point you can find in any book dealing with the theory of NMR-spectroscopy. The second point has references first our own experience and second papers which will be published soon. Greetings from Germany, Alexander. ---2--- From: lohrenz@oci.unizh.ch Date: Tue, 28 Nov 1995 13:41:54 +0100 (MET) Hi Seiji, your are probably interested in the isotropic shielding. To convert from shielding to chemical shift (in delta scale) you have to calculate the shielding of a reference compound, say for example TMS for 1H-, 13C- or Si NMR and substract the calculated shielding values of your compound. In principle it is also possible to use smaller references like CH4. In this case you have to take care of the delta-value of this compound (eg. -2.3 ppm for 13C in CH4). Just make sure that you use identical basissets for the calculations. The anisotropic values can sometimes be measured in CPMAS (solid state nmr). They give the shielding along the princible axis of your system. For spherical homogeneous compounds like CH4 you will notice that the anisotropic shieldings are the same as the isotropic. Since solution nmr gives an average of all possible orientations you alway get the isotropic shieldings. Only in cases where the molecule can be oriented (no rotation) like in the solid state, it is possible to measure the three anisotropic values independently. John ---3--- Date: 28 Nov 1995 14:31:34 +0000 From: "E.A.Moore (Elaine Moore)" G94 and all other ab initio programs used for calculating chemical shifts calulate chemical shielding. Experiment measures the chemical shift which is the chemical shielding relative to that of a reference molecule such as TMS (tetramethylsilane). The relative isotropic shielding is the quantity normally observed, for example in liquid samples. The anisotropy can be obtained from solid state measu rements and can be important in relaxation. Chemical shielding is a tensor quantity and G94 gives you all 9 components in the frame you have chosen. The isotropic shielding is (sigma xx + sigma yy + sigma zz)/3. The anisotropy is sigma 33 - (sigma 11 + sigma 22)/2 where 1,2,3 refer to the principal axis frame and sigma 33> sigma 22 >sigma 11. The chemical shift is sigma (reference)- sigma (compound) where sigma is the isotropic shielding. Hope this is some help Elaine A. Moore Chemistry Dept. The Open University UK ---4--- From: evaldera@inti.ivic.ve (Elmer Valderrama) Date: Tue, 28 Nov 1995 09:28:48 +0000 (GMT) Hi, Anisotropy only shows up in ordered media (crystalls, liquid-crystals). Organic compounds may be in solid state. As such, -e.g. not disolved in any solvent-, they may be "measured" in a RMN experiment. If, to this end, a direction of the crystal have been chosen, it may happen that the shift measured in the perpendicular to this chosen direction shows different. Then you'll have a diference which is called anisotropy. Moreover, -and for this it's better to read the reference below-, since the chem. shift actually is a tensor you'd have to compute three components (the diagonal) relative to the chosen direction in the crystal. In isotropic media (a mix of crystalls, no direction can be especified, or in liquid/solved sample) the shift is just the common measured quantity which physically is equal to one third of the trace of the above mencioned tensor. ..my two cents, hope it helps. Elmer Valderrama [1] K. Eichele, et al. "Phosphorus-31 Chemical Shift of Phosphinidene Ligands in Ruthenium Carbonyl Cluster Compounds: A 31-P Single Crystal and CP/MAS-NMR Study" J.Am.Chem.Soc. 1995,117,6961-6969 [and ref there in] (This work includes an orbital study using CACAO and a qualitative application of Ramsey's theory) ---5--- From: willsd@APPSTATE.BITNET To: Mori Seiji Date: Tue, 28 Nov 1995 10:51:13 -0400 (EDT) Seiji: I ran into this proble recently and have a suggestion: 1) You need the isotropic part of the shielding tensor. 2) Chemical shifts can be calculated from this isotropic part using some simple theory: I recommend the information in chapter 2 of : Multinuclear NMR, J. Mason Ed., Plenum Press, New York, 1987. You will also need the isotropic shielding for the chemical shift standard for your nucleus. (In my case it was (EtO)2.BF3 for 11B NMR). Steve Williams Chemistry Appalachian State University Boone, NC 28608 willsd@appstate.edu ---6--- Date: Wed, 29 Nov 1995 12:08:48 +1000 To: smori@chem.s.u-tokyo.ac.jp (Mori Seiji) From: hughc@extro.ucc.su.OZ.AU (Hugh Capper) Dear Mori, isotropic = having the same physical arrangement of moieties around the atom of interest which in this case is the proton ( if you are doing proton specroscopy) anitrosopy = is the reverse ie having different moities which cause diffences around the proton of interest. anitrosopy is reflected in the chemical shift. As for a reference a good basic text is Fundamentals of NMR Spectroscopy by Derome. best wishes, Hugh Capper Institute for Magnetic Resonance Research University of Sydney NSW 2006 Australia ---7--- From: mjf@biosym.com (Mark J Forster ) Date: Tue, 28 Nov 1995 10:03:11 -0800 ( You can see in CCL archive list.) --- end--- #################################################### Seiji Mori Graduate student in Nakamura Laboratory Department of Chemistry The University of Tokyo Hongo 7-3-1, Bunkyou-ku, Tokyo 113, JAPAN. email:smori@chem.s.u-tokyo.ac.jp #################################################### From smori@chem.s.u-tokyo.ac.jp Fri Dec 8 01:56:53 1995 Received: from utsc.s.u-tokyo.ac.jp for smori@chem.s.u-tokyo.ac.jp by www.ccl.net (8.6.10/950822.1) id BAA29688; Fri, 8 Dec 1995 01:54:17 -0500 Received: from [133.11.6.112] (smori.chem.s.u-tokyo.ac.jp) by utsc.s.u-tokyo.ac.jp (5.67+1.6W/TISN-1.3/R2) id AA10962; Fri, 8 Dec 95 15:54:00 JST Message-Id: <9512080654.AA10962@utsc.s.u-tokyo.ac.jp> Date: Fri, 8 Dec 1995 15:58:27 +0900 To: chemistry@www.ccl.net From: smori@chem.s.u-tokyo.ac.jp (Mori Seiji) X-Sender: smori@chem.s.u-tokyo.ac.jp (Unverified) Subject: Summary for NMR chemical shift calculations Cc: smori@utsc.s.u-tokyo.ac.jp Mime-Version: 1.0 Content-Type: text/plain; charset=iso-2022-jp X-Mailer: Eudora-J(1.3.8.5-J13) Dear netters, Thank you for several replies about my questions for the method of NMR chemical shift, and so on. I think the reliability of IGAIM , CSGT in G94 packages were not known very much thus we have to check these methods comparing with known methods. The systematic review of solvent effects, polarity and coordination with donating solvent such as ether have still be awaited. As to couplling constant, I understood the these computaitonal costs since huge of integrals. Recently, I found the application of coupling constant (29Si-7Li) calculation for silyllithiums. At the calculated values at HF/MIDI-4*( truncated ) methods is close to experimental results. ( O, Kikuchi, et al. Organometallics, 1995, 14, 4018 and references cited therein). I show original question and corresponding replies. Thank you very much for your help. Sincerely yours, Seiji Mori ---my original question --- Dear everybody, Recently, many letters about NMR chemical shift calculation were posted, and I am also interested in NMR shifts in organic and organometallic complexes in solution. I have several questions, 1. How is the reliablities of 1.1 methods (GIAO, IGAIM, CSGT, IGLO) 1.2 theory (HF,DFT,MP2...CC) and basis-sets 1.3 How is the solvent effect of polarity on the NMR shift, coupling constant (I think it is not related calcluation , probably it is a fundamental question about NMR)? 2. Do you know the program or references which one can calculate not only chemical shift but also the coupling constant ( for example, 1H-1H and 13C-13C) and NOE in quantum chemistry other than G94? As to 1H-1H coupling constant, many programs such as PCMODEL was supported but its calculation is not QC. If you have comments and indicate the references, would you please send me ? I will summarize replies. Thanks in advance, Seiji Mori ------------- Replies ---1--- Received: by zinc.chem.ucalgary.ca (AIX 3.2/UCB 5.64/4.03) id AA10813; Fri, 24 Nov 1995 15:29:30 -0700 From: schrecke@zinc.chem.ucalgary.ca Hi Seiji, you have quite a lot of questions in your posting ... I shall try to give you a few hints, with no intention to be comprehensive. > > 1. How is the reliablities of > 1.1 methods (GIAO, IGAIM, CSGT, IGLO) > 1.2 theory (HF,DFT,MP2...CC) and basis-sets > 1.3 How is the solvent effect of polarity on the NMR shift, coupling > constant > (I think it is not related calcluation , probably it is a fundamental > question about NMR)? Good references to start with are three nice reviews, one new, two are older: D.B.Chesnut: Annual Reports on NMR Spectroscopy, published by Academic Press, Vol.21,1989, p. 51 and ibid, Vol. 29, 1994, p.71. H.Fukui: Magn. Res. Rev. 1987, vol. 11, 205 Further, very comprehensive reviews are contained in an annual series, since 1980 written every year by C.Jameson: in Specialist Periodical Report on Nuclear Magnetic Resonance, Vol.8,1980-... (edited by G.A.Webb, and published by the Royal Society of Chemistry, Cambridge) Now more specific comments on your various questions. 1.1 Methods with distributed gauge origins (IGLO, LORG, GIAO,...) are generally preferable (for a given basis set) over methods with a common gauge origin. For examples, see the various reviews. GIAO seems to converge slightly faster with the size of the basis set then IGLO (see Wolinski, Hinton, Pulay, JACS 1990, 112, 8251). This would also make sense from a theoretical point of view. I don't know off hand about IGAIM, CSGT. 1.2 Correlation. there are cases where correlation is necessary, e.g., M.Buehl et al. J.Phys.Chem. 99, 4000 (1995), and M.Buehl et al. Chem.Phys.Lett. 241 (1995), 248. These papers contain examples where MP2 fails completely to predict the shift while DFT and CCSD are able to achieve quantitative results. There are however many other cases where the HF level works just fine (all the respective literature up to -- say -- five years ago). Basis sets. See about any paper with calculated shifts. E.g., the one by Pulay et al. mentioned above, or various papers by the Kutzelnigg group, or the various reviews. 1.3 Solvation effects can be considerable, indeed. Look for experiments that were done in the gas phase (in addition to solution studies). A review of gas phase NMR was written by C. Jameson: Chem.Rev.1991, 91, 1375. > 2. Do you know the program or references which one can calculate not only > chemical shift but also > the coupling constant > ( for example, 1H-1H and 13C-13C) and NOE in quantum chemistry other than > G94? As to 1H-1H > coupling constant, > many programs such as PCMODEL was supported but its calculation is not QC. Coupling constants are MUCH tougher to calculate than chemical shifts. The reason is (in my opinion) that you are dealing with operators that evaluate the electron density in essentially just one point of space rather than over an entire region of space. Annual reviews of calculations are contained in the same series of books that was metioned before: Specialist Periodical Report on Nuclear Magnetic Resonance Calculations have been done recently, e.g., by Erikson/Malkin/Salahub et al. but I don't have the references on hand right now. I am looking forward to your summary! Yours, Georg ---2--- From: "Steve Gwaltney" Date: Mon, 27 Nov 1995 10:51:20 -0500 > 2. Do you know the program or references which one can calculate not only > chemical shift but also > the coupling constant > ( for example, 1H-1H and 13C-13C) and NOE in quantum chemistry other than > G94? As to 1H-1H ACES II can calculate chemical shifts using GIAO's and MBPT(2) wavefunctions. ACES II can also calculate CCSD coupling constants. For information about ACES II contact aces2@qtp.ufl.edu. I believe TURBOMOLE can also calculate chemical shifts, but you should check with them to make sure. Steve ---3--- From: wagenert@Mailer.Uni-Marburg.DE (Wagener Thomas) Subject: CCL:NMR shift and coupling constant calculation Date: Mon, 27 Nov 1995 14:12:47 +0100 (CET) Dear Seiji Mori, since I have finished my thesis on the ab-initio calculation a few month ago, I hopefully can help you with some of your questions. > 1. How is the reliablities of > 1.1 methods (GIAO, IGAIM, CSGT, IGLO) > 1.2 theory (HF,DFT,MP2...CC) and basis-sets > 1.3 How is the solvent effect of polarity on the NMR shift, coupling > constant I had only the opportunity to work with GIAO and IGLO so I can only make a statement on these methods. Basically IGLO and GIAO are quite reliable for the calculation of "standard" organic compounds though the accuracy of the methods is not as high as the experimental accuracy. A rough estimate would be an accuracy of about 5ppm for 13C chemical shifts - probably better. Taking into account that you are comparing gas phase values at absolute zero with values taken from a sample in solution at a temperature of, say, 200-300K that is probably as close as you can get. The computational effort for the IGLO method is somewhat higher than for the GIAO calculation (at HF-level) though the information you get from IGLO is more extensive. Because IGLO works with localized molecular orbitals (LMOs) you get an anlysis of the contribution of the LMOs to the chemical shift while you only get the shielding constant out of the GIAO calculation. IGLO is available at HF level and there is a DFT version using the IGLO code to calculate chemical shifts. The GIAO method in MO theory is available up to CC level though the computational effort is extremely high so the calculation of the shielding constant at such a highly coorelated level is only possible for small molecules. The GAUSSIAN94 manual states that the GIAO module of the program can cope with DFT but I have no experience with that. > 2. Do you know the program or references which one can calculate not only > chemical shift but also > the coupling constant > ( for example, 1H-1H and 13C-13C) and NOE in quantum chemistry other than > G94? As to 1H-1H coupling constant, As far as I know (and I really may not be up to date) the only programs capable of calculating coupling constants is DeMon with DFT-IGLO and ACESII (at MP2 level). Some useful references are: - IGLO: Kutzelnigg, W.; Fleischer, U.; Schindler, M.:"The IGLO-Method: Ab- initio Calculation and Interpretation of NMR chemical Shifts and magnetic Susceptibilities", NMR Basic Princ. Prog., Vol. 23, Springer-Verlag, Berlin Heidelberg, 1990, 165 - DFT-IGLO: Malkin, M. G.; Malkina, O. L.; Casida, M. E.; Salahub, D. R., JACS, 116(1994), 5898 - HF-GIAO: Wolinski, K.; Hilton, J. F.; Pulay, P., JACS, 112(1990), 8251 - MP2-GIAO: Gauss, J., Chem. Phys. Lett., 191(1992), 614 Hope, this was of some help Thomas Wagener ---end-- #################################################### Seiji Mori Graduate student in Nakamura Laboratory Department of Chemistry The University of Tokyo Hongo 7-3-1, Bunkyou-ku, Tokyo 113, JAPAN. email:smori@chem.s.u-tokyo.ac.jp #################################################### From S.R.Kilvington@soton.ac.uk Fri Dec 8 11:56:58 1995 Received: from beech.soton.ac.uk for S.R.Kilvington@soton.ac.uk by www.ccl.net (8.6.10/950822.1) id LAA07309; Fri, 8 Dec 1995 11:48:32 -0500 Received: from willow.soton.ac.uk (willow.soton.ac.uk [152.78.128.20]) by beech.soton.ac.uk (8.6.12/hub-8.5a) with ESMTP id QAA03425 for ; Fri, 8 Dec 1995 16:46:22 GMT Received: (from srk@localhost) by willow.soton.ac.uk (8.6.10/client-8.8) id QAA05533 for CHEMISTRY@www.ccl.net; Fri, 8 Dec 1995 16:42:34 GMT Message-Id: <199512081642.QAA05533@willow.soton.ac.uk> Subject: Re: Problem in reading PDB files To: CHEMISTRY@www.ccl.net Date: Fri, 8 Dec 1995 16:42:34 +0000 (GMT) From: "Simon Kilvington" X-Mailer: ELM [version 2.4 PL23] Content-Type: text Content-Length: 1436 > > Have you ever used SYBYL? During the last few days, I was astonished to > find SYBYL does have problem in reading PDB files. Let me elaborate now. > [troubles with aromatic atoms deleted] the problem is PDB files don't have any data explicitly stored in them about bond orders and atom hybridisations. As the PDB format was initially designed to store protein structures it was assumed that all the molecules would be composed of only atoms in the twenty common amino acids. Therefore the residue field ie (TRP, PRO, etc) is used to look up a template which describes the connectivity [and bond type?]. now people are using PDB as a general molecule format and this causes problems. there should be no reason to use PDB files for storing general molecular structures because there are already *far* too many chemical structure file formats and most of these allow you to store connectivity, bond order, and hybridisation information. if I were you I would use a file format like MOL2 - this is text based, free format, and capable of storing a wide variety of information. even if you need to import your structure files into a program that doesn't support MOL2 files you can use the invaluable program BABEL to convert it. I think a standard molecule file format is needed. What happened to the chemical MIME type? is this still being designed or is it all agreed on yet? -- Simon Kilvington, srk@soton.ac.uk From smori@chem.s.u-tokyo.ac.jp Fri Dec 8 18:57:00 1995 Received: from utsc.s.u-tokyo.ac.jp for smori@chem.s.u-tokyo.ac.jp by www.ccl.net (8.6.10/950822.1) id SAA16235; Fri, 8 Dec 1995 18:52:20 -0500 Received: from [133.11.6.112] (smori.chem.s.u-tokyo.ac.jp) by utsc.s.u-tokyo.ac.jp (5.67+1.6W/TISN-1.3/R2) id AA16205; Sat, 9 Dec 95 08:52:01 JST Message-Id: <9512082352.AA16205@utsc.s.u-tokyo.ac.jp> Date: Sat, 9 Dec 1995 08:56:30 +0900 To: chemistry@www.ccl.net From: smori@chem.s.u-tokyo.ac.jp (Mori Seiji) X-Sender: smori@chem.s.u-tokyo.ac.jp (Unverified) Subject: Summary for NMR isotropy and anitrosopy Cc: smori@utsc.s.u-tokyo.ac.jp Mime-Version: 1.0 Content-Type: text/plain; charset=iso-2022-jp X-Mailer: Eudora-J(1.3.8.5-J13) Dear netters, In the meanwhile, as for the concept of anitrosopy and isotropy, thank you for sending me many replies. I WILL study these concepts and theory. Thank you very much for your help. Sincerely yours, Seiji Mori ---my original questions about isotropy and anisotropy--- Dear Sirs, I posted to CCL about the methods of NMR calculation a fey days ago , and had an other question. In G94, I tried to calculate the chemical shifts , there are magnetic shieldings in the output, however, no chemical shift but isotropic and anitrosopy. I examine meaning of these terms in the user's manual of Gaussian 94 and several original papers,for example, M. Schindler, et al. JCP, 1982, 76, 1919 and Pulay et al. JACS, 1990, 112, 8251, and so on which there is replies of my previous questions , but I only know that isotropic terms is "likely" to be corresponding to the chemical shifts which was experimentally observed. I had had experiments in organic chemistry, and I am now carrying out the ab initio calculations but I amn't much familiar to know the theory of NMR spectroscopy. Would you please tell me the physical meaning of isotropic and anitrosopy, and which is an equivalent of the experimental chemical shift? It is better that you show the references. Sincerely yours, Seiji Mori ---- replies---- ---1--- Date: Tue, 28 Nov 1995 11:10:24 +0100 (MEZ) From: Alexander Christian Backes Konichiwa Mori-san, (S. Mori wrote: In correct in Japanese, Kon-nichiwa.^^:)) the chemical shift is always a tensor (a symmetric 3 x 3 matrix). So, this tensor describes the interchange between several NMR-active nuclei. The outerdiagonal elements are only important in spectra of the solid phase because there the position of the molecules and therefore the observed nuclei are defined and the shift is therefor anisotropic (CSA = Chemnical shift anisotropy). In solution the motion of the molecules is randomized, so the NMR-spectrometer observes the middled shift. This means, the trace of the matrix is relevant (the trace is the sum of the diagonal elemnts of the matrix, divided by three, in case of the 3 x 3 matrix). So, in G94 the calculated isotropic shift is the relevant one for solution. Of course you need a standard for the chemical shift; G94 doesn't know to which reference substance the shifts are related to. So you have to calculate TMS at exactly the same level as the calculated molecule, and then you have to subtract the isotropic data of your molecule from the isotropic data of TMS. (This you have to do each time you use another method and basis set for geometry optimization and NMR-calculations). But be careful - the input geometry of TMS must have Td-symmetry. And therefore you have to precalculate TMS at MP2-level to make sure it gets Td-Symmetry. This geometry you can use as standard input for any of the following calculations of geometry and NMR-parameters. References to the first point you can find in any book dealing with the theory of NMR-spectroscopy. The second point has references first our own experience and second papers which will be published soon. Greetings from Germany, Alexander. ---2--- From: lohrenz@oci.unizh.ch Date: Tue, 28 Nov 1995 13:41:54 +0100 (MET) Hi Seiji, your are probably interested in the isotropic shielding. To convert from shielding to chemical shift (in delta scale) you have to calculate the shielding of a reference compound, say for example TMS for 1H-, 13C- or Si NMR and substract the calculated shielding values of your compound. In principle it is also possible to use smaller references like CH4. In this case you have to take care of the delta-value of this compound (eg. -2.3 ppm for 13C in CH4). Just make sure that you use identical basissets for the calculations. The anisotropic values can sometimes be measured in CPMAS (solid state nmr). They give the shielding along the princible axis of your system. For spherical homogeneous compounds like CH4 you will notice that the anisotropic shieldings are the same as the isotropic. Since solution nmr gives an average of all possible orientations you alway get the isotropic shieldings. Only in cases where the molecule can be oriented (no rotation) like in the solid state, it is possible to measure the three anisotropic values independently. John ---3--- Date: 28 Nov 1995 14:31:34 +0000 From: "E.A.Moore (Elaine Moore)" G94 and all other ab initio programs used for calculating chemical shifts calulate chemical shielding. Experiment measures the chemical shift which is the chemical shielding relative to that of a reference molecule such as TMS (tetramethylsilane). The relative isotropic shielding is the quantity normally observed, for example in liquid samples. The anisotropy can be obtained from solid state measu rements and can be important in relaxation. Chemical shielding is a tensor quantity and G94 gives you all 9 components in the frame you have chosen. The isotropic shielding is (sigma xx + sigma yy + sigma zz)/3. The anisotropy is sigma 33 - (sigma 11 + sigma 22)/2 where 1,2,3 refer to the principal axis frame and sigma 33> sigma 22 >sigma 11. The chemical shift is sigma (reference)- sigma (compound) where sigma is the isotropic shielding. Hope this is some help Elaine A. Moore Chemistry Dept. The Open University UK ---4--- From: evaldera@inti.ivic.ve (Elmer Valderrama) Date: Tue, 28 Nov 1995 09:28:48 +0000 (GMT) Hi, Anisotropy only shows up in ordered media (crystalls, liquid-crystals). Organic compounds may be in solid state. As such, -e.g. not disolved in any solvent-, they may be "measured" in a RMN experiment. If, to this end, a direction of the crystal have been chosen, it may happen that the shift measured in the perpendicular to this chosen direction shows different. Then you'll have a diference which is called anisotropy. Moreover, -and for this it's better to read the reference below-, since the chem. shift actually is a tensor you'd have to compute three components (the diagonal) relative to the chosen direction in the crystal. In isotropic media (a mix of crystalls, no direction can be especified, or in liquid/solved sample) the shift is just the common measured quantity which physically is equal to one third of the trace of the above mencioned tensor. ..my two cents, hope it helps. Elmer Valderrama [1] K. Eichele, et al. "Phosphorus-31 Chemical Shift of Phosphinidene Ligands in Ruthenium Carbonyl Cluster Compounds: A 31-P Single Crystal and CP/MAS-NMR Study" J.Am.Chem.Soc. 1995,117,6961-6969 [and ref there in] (This work includes an orbital study using CACAO and a qualitative application of Ramsey's theory) ---5--- From: willsd@APPSTATE.BITNET To: Mori Seiji Date: Tue, 28 Nov 1995 10:51:13 -0400 (EDT) Seiji: I ran into this proble recently and have a suggestion: 1) You need the isotropic part of the shielding tensor. 2) Chemical shifts can be calculated from this isotropic part using some simple theory: I recommend the information in chapter 2 of : Multinuclear NMR, J. Mason Ed., Plenum Press, New York, 1987. You will also need the isotropic shielding for the chemical shift standard for your nucleus. (In my case it was (EtO)2.BF3 for 11B NMR). Steve Williams Chemistry Appalachian State University Boone, NC 28608 willsd@appstate.edu ---6--- Date: Wed, 29 Nov 1995 12:08:48 +1000 To: smori@chem.s.u-tokyo.ac.jp (Mori Seiji) From: hughc@extro.ucc.su.OZ.AU (Hugh Capper) Dear Mori, isotropic = having the same physical arrangement of moieties around the atom of interest which in this case is the proton ( if you are doing proton specroscopy) anitrosopy = is the reverse ie having different moities which cause diffences around the proton of interest. anitrosopy is reflected in the chemical shift. As for a reference a good basic text is Fundamentals of NMR Spectroscopy by Derome. best wishes, Hugh Capper Institute for Magnetic Resonance Research University of Sydney NSW 2006 Australia ---7--- From: mjf@biosym.com (Mark J Forster ) Date: Tue, 28 Nov 1995 10:03:11 -0800 ( You can see in CCL archive list.) --- end--- #################################################### Seiji Mori Graduate student in Nakamura Laboratory Department of Chemistry The University of Tokyo Hongo 7-3-1, Bunkyou-ku, Tokyo 113, JAPAN. email:smori@chem.s.u-tokyo.ac.jp #################################################### From smori@chem.s.u-tokyo.ac.jp Fri Dec 8 18:57:07 1995 Received: from utsc.s.u-tokyo.ac.jp for smori@chem.s.u-tokyo.ac.jp by www.ccl.net (8.6.10/950822.1) id SAA16196; Fri, 8 Dec 1995 18:45:51 -0500 Received: from [133.11.6.112] (smori.chem.s.u-tokyo.ac.jp) by utsc.s.u-tokyo.ac.jp (5.67+1.6W/TISN-1.3/R2) id AA16193; Sat, 9 Dec 95 08:45:31 JST Message-Id: <9512082345.AA16193@utsc.s.u-tokyo.ac.jp> Date: Sat, 9 Dec 1995 08:50:01 +0900 To: chemistry@www.ccl.net From: smori@chem.s.u-tokyo.ac.jp (Mori Seiji) X-Sender: smori@chem.s.u-tokyo.ac.jp (Unverified) Subject: Summary for NMR chemical shift calculations Cc: smori@utsc.s.u-tokyo.ac.jp Mime-Version: 1.0 Content-Type: text/plain; charset=iso-2022-jp X-Mailer: Eudora-J(1.3.8.5-J13) Dear netters, Thank you for several replies about my questions for the method of NMR chemical shift, and so on. I think the reliability of IGAIM , CSGT in G94 packages were not known very much thus we have to check these methods comparing with known methods. The systematic review of solvent effects, polarity and coordination with donating solvent such as ether have still be awaited. As to couplling constant, I understood the these computational costs since huge of integrals. Recently, I found the application of coupling constant (29Si-7Li) calculation for silyllithiums. At the calculated values at HF/MIDI-4*( truncated ) methods is close to experimental results. ( O, Kikuchi, et al. Organometallics, 1995, 14, 4018 and references cited therein). I show original question and corresponding replies. Thank you very much for your help. Sincerely yours, Seiji Mori ---my original question --- Dear everybody, Recently, many letters about NMR chemical shift calculation were posted, and I am also interested in NMR shifts in organic and organometallic complexes in solution. I have several questions, 1. How is the reliablities of 1.1 methods (GIAO, IGAIM, CSGT, IGLO) 1.2 theory (HF,DFT,MP2...CC) and basis-sets 1.3 How is the solvent effect of polarity on the NMR shift, coupling constant (I think it is not related calcluation , probably it is a fundamental question about NMR)? 2. Do you know the program or references which one can calculate not only chemical shift but also the coupling constant ( for example, 1H-1H and 13C-13C) and NOE in quantum chemistry other than G94? As to 1H-1H coupling constant, many programs such as PCMODEL was supported but its calculation is not QC. If you have comments and indicate the references, would you please send me ? I will summarize replies. Thanks in advance, Seiji Mori ------------- Replies ---1--- Received: by zinc.chem.ucalgary.ca (AIX 3.2/UCB 5.64/4.03) id AA10813; Fri, 24 Nov 1995 15:29:30 -0700 From: schrecke@zinc.chem.ucalgary.ca Hi Seiji, you have quite a lot of questions in your posting ... I shall try to give you a few hints, with no intention to be comprehensive. > > 1. How is the reliablities of > 1.1 methods (GIAO, IGAIM, CSGT, IGLO) > 1.2 theory (HF,DFT,MP2...CC) and basis-sets > 1.3 How is the solvent effect of polarity on the NMR shift, coupling > constant > (I think it is not related calcluation , probably it is a fundamental > question about NMR)? Good references to start with are three nice reviews, one new, two are older: D.B.Chesnut: Annual Reports on NMR Spectroscopy, published by Academic Press, Vol.21,1989, p. 51 and ibid, Vol. 29, 1994, p.71. H.Fukui: Magn. Res. Rev. 1987, vol. 11, 205 Further, very comprehensive reviews are contained in an annual series, since 1980 written every year by C.Jameson: in Specialist Periodical Report on Nuclear Magnetic Resonance, Vol.8,1980-... (edited by G.A.Webb, and published by the Royal Society of Chemistry, Cambridge) Now more specific comments on your various questions. 1.1 Methods with distributed gauge origins (IGLO, LORG, GIAO,...) are generally preferable (for a given basis set) over methods with a common gauge origin. For examples, see the various reviews. GIAO seems to converge slightly faster with the size of the basis set then IGLO (see Wolinski, Hinton, Pulay, JACS 1990, 112, 8251). This would also make sense from a theoretical point of view. I don't know off hand about IGAIM, CSGT. 1.2 Correlation. there are cases where correlation is necessary, e.g., M.Buehl et al. J.Phys.Chem. 99, 4000 (1995), and M.Buehl et al. Chem.Phys.Lett. 241 (1995), 248. These papers contain examples where MP2 fails completely to predict the shift while DFT and CCSD are able to achieve quantitative results. There are however many other cases where the HF level works just fine (all the respective literature up to -- say -- five years ago). Basis sets. See about any paper with calculated shifts. E.g., the one by Pulay et al. mentioned above, or various papers by the Kutzelnigg group, or the various reviews. 1.3 Solvation effects can be considerable, indeed. Look for experiments that were done in the gas phase (in addition to solution studies). A review of gas phase NMR was written by C. Jameson: Chem.Rev.1991, 91, 1375. > 2. Do you know the program or references which one can calculate not >only > chemical shift but also > the coupling constant > ( for example, 1H-1H and 13C-13C) and NOE in quantum chemistry >other than > G94? As to 1H-1H > coupling constant, > many programs such as PCMODEL was supported but its calculation is >not QC. Coupling constants are MUCH tougher to calculate than chemical shifts. The reason is (in my opinion) that you are dealing with operators that evaluate the electron density in essentially just one point of space rather than over an entire region of space. Annual reviews of calculations are contained in the same series of books that was metioned before: Specialist Periodical Report on Nuclear Magnetic Resonance Calculations have been done recently, e.g., by Erikson/Malkin/Salahub et al. but I don't have the references on hand right now. I am looking forward to your summary! Yours, Georg ---2--- From: "Steve Gwaltney" Date: Mon, 27 Nov 1995 10:51:20 -0500 > 2. Do you know the program or references which one can calculate not > only > chemical shift but also > the coupling constant > ( for example, 1H-1H and 13C-13C) and NOE in quantum chemistry > other than > G94? As to 1H-1H ACES II can calculate chemical shifts using GIAO's and MBPT(2) wavefunctions. ACES II can also calculate CCSD coupling constants. For information about ACES II contact aces2@qtp.ufl.edu. I believe TURBOMOLE can also calculate chemical shifts, but you should check with them to make sure. Steve ---3--- From: wagenert@Mailer.Uni-Marburg.DE (Wagener Thomas) Subject: CCL:NMR shift and coupling constant calculation Date: Mon, 27 Nov 1995 14:12:47 +0100 (CET) Dear Seiji Mori, since I have finished my thesis on the ab-initio calculation a few month ago, I hopefully can help you with some of your questions. > 1. How is the reliablities of > 1.1 methods (GIAO, IGAIM, CSGT, IGLO) > 1.2 theory (HF,DFT,MP2...CC) and basis-sets > 1.3 How is the solvent effect of polarity on the NMR shift, coupling > constant I had only the opportunity to work with GIAO and IGLO so I can only make a statement on these methods. Basically IGLO and GIAO are quite reliable for the calculation of "standard" organic compounds though the accuracy of the methods is not as high as the experimental accuracy. A rough estimate would be an accuracy of about 5ppm for 13C chemical shifts - probably better. Taking into account that you are comparing gas phase values at absolute zero with values taken from a sample in solution at a temperature of, say, 200-300K that is probably as close as you can get. The computational effort for the IGLO method is somewhat higher than for the GIAO calculation (at HF-level) though the information you get from IGLO is more extensive. Because IGLO works with localized molecular orbitals (LMOs) you get an anlysis of the contribution of the LMOs to the chemical shift while you only get the shielding constant out of the GIAO calculation. IGLO is available at HF level and there is a DFT version using the IGLO code to calculate chemical shifts. The GIAO method in MO theory is available up to CC level though the computational effort is extremely high so the calculation of the shielding constant at such a highly coorelated level is only possible for small molecules. The GAUSSIAN94 manual states that the GIAO module of the program can cope with DFT but I have no experience with that. > 2. Do you know the program or references which one can calculate not > only > chemical shift but also > the coupling constant > ( for example, 1H-1H and 13C-13C) and NOE in quantum chemistry >other than > G94? As to 1H-1H coupling constant, As far as I know (and I really may not be up to date) the only programs capable of calculating coupling constants is DeMon with DFT-IGLO and ACESII (at MP2 level). Some useful references are: - IGLO: Kutzelnigg, W.; Fleischer, U.; Schindler, M.:"The IGLO-Method: Ab- initio Calculation and Interpretation of NMR chemical Shifts and magnetic Susceptibilities", NMR Basic Princ. Prog., Vol. 23, Springer-Verlag, Berlin Heidelberg, 1990, 165 - DFT-IGLO: Malkin, M. G.; Malkina, O. L.; Casida, M. E.; Salahub, D. R., JACS, 116(1994), 5898 - HF-GIAO: Wolinski, K.; Hilton, J. F.; Pulay, P., JACS, 112(1990), 8251 - MP2-GIAO: Gauss, J., Chem. Phys. Lett., 191(1992), 614 Hope, this was of some help Thomas Wagener ---end-- ********************************************************** 　 　　森　聖治 東京大学大学院理学系研究科 物理有機化学講座中村研究室 D1 電話番号：03-3812-2111　内線4369 FAX：　　03-3812-8099 email:smori@chem.s.u-tokyo.ac.jp **********************************************************