Thursday, November 29, 2012

A Hierarchy of Methods for the Energetically Accurate Modeling of Isomerism in Monosaccharides

Sameera, W. M. C.; Pantazis, D. A. J. Chem. Theory Comput. 2012, 8, 2630-2645 (Paywall)
Contributed by Steven Bachrach.
Reposted from Computational Organic Chemistry with permission


A comprehensive evaluation of how different computational methods perform in predicting the energies of monosaccharides comes to some very interesting conclusions. Sameera and Pantazis1 have examined the eight different aldohexoses (allose, alltrose, glucose, mannose, gulose, idose, galactose and talose), specifically looking at different rotomers of the hydroxymethyl group, α- vs. β-anomers, pyranose vs. furanose isomers, ring conformations (1C4 vs skew boat forms), and ring vs. open chain isomers. In total, 58 different structures were examined. The benchmark computations are CCSD(T)/CBS single point energies using the SCS-MP2/def2-TZVPP optimized geometries. The RMS deviation from these benchmark energies for some of the many different methods examined are listed in Table 1.
Table 1. Average RMS errors (kJ mol-1) of the 58 different monosaccharide structures for
different computational methods.
method
average RMS error
LPNO-CEPA
0.71
MP2
1.27
SCS-MP2
1.55
mPW2PLYP-D
2.02
M06-2x
2.03
PBE0
3.62
TPSS
4.78
B3LYP-D
4.79
B3LYP
5.06
HF
6.69
B97D
7.66
Perhaps the most interesting take-home message is that CEPA, MP2, the double hybrid methods and M06-2x all do a very good job at evaluating the energies of the carbohydrates. Given the significant computational advantage of M06-2x over these other methods, this seems to be the functional of choice! The poorer performance of the DFT methods over the ab initio methods is primarily in the relative energies of the open-chain isomers, where errors can be on the order of 10-20 kJ mol-1 with most of the functionals; even the best overall methods (M06-2x and the double hybrids) have errors in the relative energies of the open-chain isomers of 7 kJ mol-1. This might be an area of further functional development to probe better treatment of the open-chain aldehydes vs. the ring hemiacetals.

References

(1) Sameera, W. M. C.; Pantazis, D. A. "A Hierarchy of Methods for the Energetically Accurate Modeling of Isomerism in Monosaccharides," J. Chem. Theory Comput. 20128, 2630-2645, DOI:10.1021/ct3002305

Tuesday, November 27, 2012

Polyoxometalates are cationic, not anionic

N.V. Izarova, N. Vankova, A. Banerjee, G.B. Jameson, T. Heine, F. Schinle, O. Hampe, U. Kortz, Angewandte Chemie International Edition 2010, 49, 7807-7811 (Paywall)
Contributed by Marcel Swart

Polyoxometalates are clusters of metals connected together through oxygens, and can be giant molecules such as {(MoVI)MoVI5O21}12(MoV2O4)30]12- (also known as Mo132) as shown by Bo and Miró in Dalton Transactions recently[1]. These clusters bear a total anionic charge, for instance -12 in the aforementioned example. However, this is not the complete picture.

Ulrich Kortz (Univ. Bremen) visited Girona a couple of months ago and reported an interesting example of how theory can be useful. His group was working on a "A Noble-Metalate Bowl", and when trying to reproduce the 51V NMR spectrum computationally, it was impossible to get good agreement. By introducing Na/K cations they did get more and more reasonable results, and only obtained good agreement after having introduced 7 potassiums (or alternatively 6 potassiums and 1 sodium).



At first, the experimental people did not believe the calculations, because they had their mind set on the fact that polyoxometalates carry a large negative charge, and not a positive one. However, after doing electrospray mass spectrometry, indeed they observed mainly two peaks:

1) a singly charged molecular cation {K7[Pd7V6O24(OH)2]}+ (with seven potassiums)
2) a singly charged molecular cation {Na1K6[Pd7V6O24(OH)2]}+ (with six potassiums)

Therefore, contrary to popular belief, polyoxometalates are cationic!

[1] C. Bo, P. Miró, Dalton Trans. 2012, 41, 9984-9988

Creative Commons License
This work is licensed under a Creative Commons Attribution 3.0 Unported License

Sunday, November 25, 2012

Predicting pKa of Amines for CO2 Capture: Computer versus Pencil-and-Paper


K Z. Sumon, A. Henni, and Allan L. L. East. Industrial & Engineering Chemistry Research 2012 51 (37), 11924-11930
Contributed by Thomas Cundari

I tell my students, as was told tome, that the title is the shortest abstract for a paper. The title should becarefully chosen to be as succinct as possible, and also to catch the reader'sattention and make them want to read the rest of the paper. I mention this inconnection with a recent paper[1]by Sumon, Henni and East of the University of Regina, entitled"Predicting pKa of Aminesfor CO2 Capture: Computer versus Pencil-and-Paper."  This title has three factors to recommend it - an important andpopular application (carbon dioxide capture), one of the most fundamental concepts in chemistry (pKa andhence Brønsted acidity), and the title itselfinvokes images of a titanic struggle between a giant pencil and a multi-headedLinux cluster beast, perhaps akin to the Leviathans from recent Avengers movie.

Thus, suitably intrigued, I read on.The authors reported a new pKa calculational method, entitled SHE - one assumes thatthe abbreviation denotes theauthor's last names. The rationale for the various energetic contributionsneeded for ab initio computed pKavalues like the free energy of the proton in aqueous solution were given. Themethod was then applied to organic amines of interest in carbon capture, morespecifically the amines used to strip carbon dioxide from power plant effluent.

The underlying electronic energiesneeded to compute thedifference in free energy between an amine base (B) and its Brønsted conjugate acid (BH+)were computed at the MP2/6-31G(d) level of theory. The basis set settled uponby the authors was compared to larger basis sets - triple zeta, withpolarization functions on H, inclusion of diffusion functions, etc. - and the smaller basis setreproduced a larger, extended basis set reasonably well for a subset ofprimary, secondary, tertiary and cyclic amines. Moreover, I assumed that anydifferences introduced by the use of a less complete basis set could be accounted for with theempirical constant C (more about that below). Obviously, smaller basis setswill permit access to larger amines, as the authors pointed out.

For the ab initio calculations, the authors also investigated the role ofthe atomic radii (used in the continuum solvation calculations) as well as the impactof conformational flexibility. The latter issue is also one of the majordifficulties in developing and applying paper-and-pencil methods, in that thefunctional groups of a molecule are the same whether or not it is in a low orhigh energy conformation, and thus so-called 2D methods cannot distinguish onefrom the other, and both would thus be predicted to have identical properties.

 Interestingly, only electronic energies (thusobviating the need for a potentially expensive Hessian calculation) wereutilized as current and previous analyses by the author indicated that the zeropoint, enthalpic and entropic corrections to convert electronic energiesto free energies were nearconstant given the chemical similarity of the conjugate acid-base pairs (B andBH+ or B-OH2 and BH+-OH2) that makeup the amine systems of interest. For the latter conjugate pair, an extra watermolecule was added to the continuum solvent cavity to mimic what some mightrefer to as microsolvation (Model II in the author's jargon).

An empirical constant of 1.7 pKaunits for cyclic amines and 0.7 pKa units for acyclic amines was subtracted toput computed values into better correspondence with experimentalvalues for a test set of 32 molecules, and presumably to account for some ofthe uncertainty and approximations made elsewhere in the SHE procedure.

Hence, the authors presented acompact, computationally efficient procedure for computing the pKa's of amines.Arrayed against this approach is a more old school, parametric method. For those of us old enoughto remember the days before computers were all powerful, the use of parametricmethods for estimating important quantities was a popular researchendeavor for understanding and estimating important physicalquantities of novel materials. Benson's rules for predicting heats offormation,[2]and Drago's EC model for acid-base reaction enthalpies,[3]are but two examples of parametric methods. Today, the parametric approach is embodied in QSPR and QSARanalyses, many of which have at their heart an assumption of group additivity,which assumes that the whole (molecule) is the sum of its parts (functionalgroups).

The pencil-and-paper method utilizedby the authors was the PDS (Perrin-Dempsey-Sarjeant) technique. Not requiringany MP2 computations,the method is, ofcourse, much faster than "computer" methods. Like all methods, it islimited by the functional group approximation. Therefore, if one would like touse PDS to estimate thepKa of a new material, which contains a functional group not in the originalparameterization, then one is faced with the daunting task of extending theparameterization to the new functional group. Of course, any attempts to go “outsidethe box,” will always going to suffer from the uncertainty that arises fromextrapolation versus interpolation, which has always been a problem forparametric methods.

The authors presented an updated PDSmethod by introducing new base values and functional group corrections, whichessentially halved the RMS error for amines of interest in carbon dioxidecapture.

Finally, with this framework, theepic battle between computer and paper/pencil was on! And, it was a true battleroyale, with parametric and computational methods squaring off among themselvesand with each other. The old and new PDS methods were compared for a test setof 32 amines, and the latter, as alluded to above, outperformed the former by afactor of two. The new PDS method yielded a pKa error of only 0.18 pKa units.The SHE method had an RMS error of 0.28, hence only marginally higher. Theinclusion of the C empirical constant reduced the error in the SHE method bynearly 1 pKa unit. At the end of the struggle, the SHE method stood alone inthe ring. It had an RMS error less than the other approaches investigated bythe authors in this paper, and given the attention the authors put into makingit as computationally efficient as possible, one would expect this recipe tohave significant upside for larger organic amines. However, the updated PDSmethod fought valiantly, and would seem to be an interesting candidate for furtherresearch to diversify it to alarger functional grouplibrary.

Likewise, it would also be ofinterest to diversify the SHE method in a similar manner to investigate a greater range of possible CO2sorbents. Other possible extensions include assessing the impact of lessexpensive abs initio methods likeHartree-Fock (yes, good old Hartree-Fock!) or perhaps DFT, instead of MP2, tobring the SHE method to larger and more complex organics amines and indeed evennon-amines.



[1]      Predicting pKa of Amines for CO2 Capture: Computer versusPencil-and-Paper.  K Z. Sumon, A. Henni, and Allan L. L. East. Industrial& Engineering Chemistry Research 2012 51 (37),11924-11930
[2]      Estimation of heats of formation of organic compounds by additivitymethods. N. Cohen and S. W.Benson. Chemical Reviews 1993 93 (7), 2419-2438
[3]      A Double-Scale Equation for Correlating Enthalpies of Lewis Acid-BaseInteractions. R. S. Dragoand B. B. Wayland. Journal of the American Chemical Society 1965 87(16), 3571-3577

Thursday, November 15, 2012

Experimental and Theoretical Investigation of 1JCC and nJCC Coupling Constants in Strychnine

Williamson, R. T.; Buevich, A. V.; Martin, G. E. Org. Letters 2012, 14, 5098-5101 (Paywall)
Contributed by Steven Bachrach.
Reposted from Computational Organic Chemistry with permission

The use of computed NMR coupling constants is starting to grow. In a previous post I discussed a general study by Rablen and Bally on methods for computing JHH coupling constants. Now Williamson reports methods to experimentally obtain 1 JCC and 3JCC coupling constants.1 These were obtained for strychnine. He then computed the coupling constants in two steps. Using the B3LYP/6-31G(d) optimized geometry, first the Fermi contact contribution was computed at B3LYP/6-31+G(d,p) by uncontracting the basis set and adding an additional tighter set of polarization functions. Second, the remaining terms (spin-dipolar, paramagnetic spin-orbit and diamagnetic spin-orbit coupling) were computed with the 6-31+Gd,p) set without modifications. The two computed terms were added to give the final estimate.

A plot of the experimental vs. the DFT computed 1 JCC and 3JCC coupling constants shows an excellent linear relation, with correlation coefficient of 0.9986 and a slope of 0.98. The mean absolute deviation for the computed and experimental 1 JCC and 3JCC coupling constants is 1.0 Hz and 0.4 Hz, respectively, both well within the experimental error. I expect that computed NMR spectra will continue to be a growth area, especially for structural identification.

References

(1) Williamson, R. T.; Buevich, A. V.; Martin, G. E. "Experimental and Theoretical Investigation of 1JCC andnJCC Coupling Constants in Strychnine," Org. Letters 201214, 5098-5101, DOI: 10.1021/ol302366s

InChIs

strychnine:
InChI=1S/C21H22N2O2/c24-18-10-16-19-13-9-17-21(6-7-22(17)11-12(13)5-8-25-16)14-3-1-2-4-15(14)23(18)20(19)21/h1-5,13,16-17,19-20H,6-11H2/t13-,16-,17-,19-,20-,21+/m0/s1
InChIKey=QMGVPVSNSZLJIA-FVWCLLPLSA-N

Sunday, November 11, 2012

Benchmarking Semiempirical Methods for Thermochemistry, Kinetics, and Noncovalent Interactions: OMx Methods Are Almost As Accurate and Robust As DFT-GGA Methods for Organic Molecules

Martin Korth and Walter Thiel Journal of Chemical Theory and Computation 2011, 7, 2929-2936 (Paywall)
Highlighted by Jan Jensen

This paper presents a thorough benchmark of six semiempirical method: AM1, PM6, SCC-DFTB, OM1, OM2, and OM3, as well as their corresponding dispersion-corrected versions.  The OMx methods, developed by Thiel, include orthogonalization corrections and are available in the MNDO99 code.

The methods are benchmarked using a 370-entry subset of the GMTKN24 database (containing only H, C, N, and O) constructed by Goerigk and Grimme for the evaluation of the "true" performance of quantum mechanical methods.  This database is comprised of 24 chemically different subsets of experimental or CCSD(T)/CBS (or similar) data, such as barrier heights, conformational energies, ionization potentials, etc.

Overall OM3 performs best with a mean absolute deviation (MAD) of 7.9 kcal/mol, which is not too different from MAD obtained with DFT: PBE/TZVP (6.6 kcal/mol) and B3LYP/TZVP (4.8) kcal/mol.  Surprisingly, PM6 performs significantly worse than AM1: 18.2 vs 14.5 kcal/mol.

The error is largest for the MB08-165 subset consisting of decomposition energies of randomly generated artificial molecules, specifically designed to test robustness and general applicability. This proves a real challenge for PM6 with a MAD of 128.4 kcal/mol.  If this subset, as well triplets and quartets, are removed, PM6 now outperforms AM1 (though not by much: 10.2 vs 12.2 kcal/mol) and OM3 is now within 1.5 kcal/mol of the DFT results!

A previous highlight indicated that PM6 out-performs AM1 and PM3 when it comes to barrier heights.  I was therefore surprised to see that, in this study, the PM6-MADs for the two barrier-height subsets (BH76 and BHPERI) were higher or comparable to AM1.  OM3 (or OM2) performs best again and in the case of BH76 out-performs PBE!  It is worth noting that the benchmarking consists of single-point energies using the structures in the database.  It is possible that computing the barriers using stationary points obtained using the respective methods will change the picture.

Creative Commons License
This work is licensed under a Creative Commons Attribution 3.0 Unported License.  


Friday, November 9, 2012

Experimental Verification of the Homoaromaticity of 1,3,5-Cycloheptatriene and Evaluation of the Aromaticity of Tropone and the Tropylium Cation by Use of the Dimethyldihydropyrene Probe

Williams, R. V.; Edwards, W. D.; Zhang, P.; Berg, D. J.; Mitchell, R. H. J. Am. Chem. Soc. 2012, 134, 16742-16752 (Paywall)
Contributed by Steven Bachrach.
Reposted from Computational Organic Chemistry with permission

Assessing the degree of aromaticity in a novel compound has been a much sought after prize, and is the topic of much of Chapter 2 in my book. An interesting approach is described in a recent JACS paper by Williams and Mitchell.1 The interior methyl groups of 1 sit above and below the ring plane of the aromatic dihydropyrene and provide an interesting magnetic probe of the aromaticity; the chemical shift of these methyl groups are δ -4.06ppm, far upfield as they sit in the shielded region above the aromatic plane. Annelation of a benzene ring to give 2 should reduce the ring current, thereby reflecting a reduced aromatic character. In fact, the chemical shifts of the methyls in 2 are δ -1.58 ppm. This relatively large chemical shift difference provides a means for measuring the aromatic influence of other fused rings.

1

2
Suppose a different (non-benzene) ring were fused onto 1. Williams and Mitchell examined two such cases 3 and 4 (among others). These two compounds were prepared and studied by 1H NMR and also by B3LYP/6-31G* computations. The optimized structures of 3 and 4 are shown in Figure 1.

3

4
The experimental chemical shifts of the interior methyl groups are δ -3.32 ppm. This downfield shift of the methyls relative to their position in 1 reflects some homoaromatic character of the cycloheptatrienyl ring. If we take the difference in the methyl chemical shifts in 1 and 2 to reflect the aromatic character of benzene (2.48 ppm), then the difference in the chemical shifts of 3 and 1 (0.74 ppm) indicates that the cycloheptatrienyl ring has 0.74/2.48*100 = 30% the (homo)aromatic character of benzene! Similarly, the methyl chemical shifts in 4 are δ -3.56 ppm, leading to an estimate of the aromatic character of the tropone ring of 20%.

3

4
Figure 1. B3LYP/6-31G* optimized geometries of 3 and 4.
In using NICS to estimate the aromatic character, they make use of the average value of the NICS in the four rings of the dihydropyrene fragment. The baseline comparison is then in the average NICS value of 1compared to that in 5, a compound that has a similar geometry but without the aromatic character of the fused benzene ring. This difference is 11.42ppm. The analogous relationship is then 3 with 6 (a NICS difference of 3.81 ppm) and 4 with 7 (a NICS difference of 2.77ppm). This gives an estimate of the (homo)aromatic character of cycloheptatriene of 33% and the aromatic character of tropolone of 24%. This NICS estimates are in great agreement with the experimental values from the proton chemical shifts.

5

6

7

References

(1) Williams, R. V.; Edwards, W. D.; Zhang, P.; Berg, D. J.; Mitchell, R. H. "Experimental Verification of the Homoaromaticity of 1,3,5-Cycloheptatriene and Evaluation of the Aromaticity of Tropone and the Tropylium Cation by Use of the Dimethyldihydropyrene Probe," J. Am. Chem. Soc. 2012134, 16742-16752, DOI: 10.1021/ja306868r.

InChIs

1: InChI=1S/C26H32/c1-23(2,3)21-13-17-9-11-19-15-22(24(4,5)6)16-20-12-10-18(14-21)25(17,7)26(19,20)8/h9-16H,1-8H3/t25-,26-
InChIKey=SEGNSURCRWXVRS-DIVCQZSQSA-N
2: InChI=1S/C30H34/c1-27(2,3)21-15-19-13-14-20-16-22(28(4,5)6)18-26-24-12-10-9-11-23(24)25(17-21)29(19,7)30(20,26)8/h9-18H,1-8H3/t29-,30-/m1/s1
InChIKey=JQXZCWYPFGGVNF-LOYHVIPDSA-N
3: InChI=1S/C33H40/c1-20-13-21(2)15-27-26(14-20)28-18-24(30(3,4)5)16-22-11-12-23-17-25(31(6,7)8)19-29(27)33(23,10)32(22,28)9/h11-12,14-19H,13H2,1-10H3/t32-,33-/m1/s1
InChIKey=MUBRTBMPRBJVOC-CZNDPXEESA-N
4: InChI=1S/C33H38O/c1-19-13-25-26(14-20(2)29(19)34)28-18-24(31(6,7)8)16-22-12-11-21-15-23(30(3,4)5)17-27(25)32(21,9)33(22,28)10/h11-18H,1-10H3/t32-,33-/m1/s1
InChIKey=OWIRQHSHEVFPPM-CZNDPXEESA-N