NOTE: This page will be updated soon with information about our new machine learning-augmented DFT method (DU8ML)

Relativistic Force Field (RFF) is our method for fast and accurate calculations of nuclear spin-spin coupling constants (SSCCs). Selected examples of SSCCs calculated with RFF (DU8c) are found below with links to interactive structures and spectra.

This project in large part is supported by the National Science Foundation National Science Foundation

RFF: origin of the term

There has been growing consensus that the Fermi contact mechanism dominates nuclear spin scalar coupling in organic molecules. Other contributors, such as dia- and paramagnetic components of spin-orbit coupling are generally small (in organic compounds), yet they require the lion share of computational time. Our approach to fast and accurate computations of SSCCs is to augment the readily computed Fermi contacts with a parametric empirical Relativistic Force Field designed to fake these missing SOC contributions and improve the accuracy of SSCCs calculations, not unlike the ubiquitous Force Fields in molecular mechanics utilize sets of empirical parameters to fake quantum mechanical effects and obtain accurate molecular structures.

Our new (and very light) basis set, DU8, is designed to further accelerate the DFT computations of Fermi contacts. The NBO-assisted parametric scaling occurs in a fraction of a second. The net result is a new RFF DU8c method ("c" means that the parameterization is implemented not only for the proton-proton but also for the 13C-1H SSCCs) which offers an equal or better accuracy of SSCCs computations in a small fraction of time compared with the existing state of the art methods (the acceleration for relatively large organic molecules like strychnine could be as dramatic as several orders of magnitude).

References:

JSMol is best viewed with Firefox, Chrome, or Safari
(IE and Opera are slow)

Note that rff is designed to accurately compute spin-spin coupling constants. The chemical shifts - in the examples below - are calculated with modest accuracy using existing methods, such as GIAO at the mPW1PW91/6-311+G(d,p) level of theory. (For more information on the state of the art in computing chemical shifts see CHESHIRE - CHEmical SHIft REpository maintained by Dean Tantillo).

The J scaling is being refined and tested on complex organic molecules. The Unsigned Error Histrogram shows the test set as of April 2016, which contains >2400 experimental constants with rmsd = 0.28 Hz. (red line shows the ideal gaussian distribution for σ=0.28).

Links to predicted NMR spectra of a few prominent natural products and other interesting molecules in the test set are found below (In parentheses after the name: rmsd(Hz) / number of reported Jexp)


1-oxo-9-hydroisochatancin
(0.31Hz/8)
11-epi-rabdocoestin A [ JHH 0.27Hz/24 ] - [ 1H 0.25ppm/21 ] - [ 13C 1.65ppm/20 ]
12-epoxyobtusallene IV
(0.30Hz/21)
12a-hydroxydaphniyunnine A
(0.52Hz/16)
 17-nor-pimarane
(0.38Hz/12)
 2-oxo-34-dehydroneomajucin
(cmpd 9)
(0.22Hz/6)
2S-hydroxy-34-dehydroneomajucin
(cmpd 6)
(0.26Hz/7)
2-α-hydroxyartepestrin F
(0.21Hz/9)
 6a-acetoxyanopterine
7-hydroxy-trichothecolone acetate
(0.18Hz/6)
 P.acaciicola
(compound 25)
(0.34Hz/13)
 P.acaciicola
(compound 26)
(0.33Hz/8)
PR-toxin
(0.32Hz/5)
alsmaphorazine D
(0.38Hz/17)
 alstoscholarisine E
(0.17Hz/10)
ambiguine P
(0.34Hz/13)
 aplydactone
(0.37Hz/7)
 aplydactone-epi8
(0.44Hz/9)
aquatolide
(0.31Hz/23)
 artemisinin
(0.24Hz/12; 1H 0.09ppm/16; 13C 0.94ppm/15)
artepestrin G
(0.26Hz/13)
aza[6]cycloparaphenylene
 bacaryolane B
(0.21Hz/15)
bacaryolane C
(0.29Hz/17)
berkelic acid
(0.30Hz/11)
briarellinJ
(synthetic - Crimmins)
(0.29Hz/8)
calyciphylline N
(0.31Hz/16)
ciliatonoid A
(0.29Hz/11)
 cornolactone B
(0.18Hz/10)
 dendrobine
(0.25Hz/11)
discorhabdin A [prianosin A]
(0.33Hz/5)
echinopine A
(0.29Hz/15)
 fischambiguine B
(0.47Hz/11)
fusarisetin A
(0.22Hz/9)
 galiellalactone
(0.33Hz/17)
gelsemine
(0.16Hz/15)
gibberellic acid
(0.35Hz/10)
ginkgolide B
(0.14Hz/6)
 grisemycin
(0.44Hz/11)
hawaiinolide A
(0.23Hz/10)
 hexacyclinol
(0.32Hz/9)
 hippolachnin A
(0.14Hz/3)
hypodoratoxide
(revised)
(0.13Hz/12)
 incarviatone A
(0.39Hz/14)
 indolenine S10 - Wang
(0.33Hz/15)
isopalhinine A
(0.24Hz/27)
isoschizogamine
(0.30Hz/26)
jiadifenolide
(0.24Hz/12)
kadcoccinine A
(0.27Hz/3)
 lancolide C
(0.24Hz/14)
lancolide D
(0.20Hz/14)
lancolide E
(0.18Hz/14)
laxitextine B
(0.29Hz/19)
leonuketal
(0.30Hz/18)
marilzabicycloallene D
(0.46Hz/23)
 marilzanin
(0.28Hz/19)
 meridane
(0.10Hz/4)
mitchellene A
(0.28Hz/22)
 mitchellene B
(0.30Hz/20)
 myrtenal
(0.29Hz/10)
nocardioazineA
(0.21Hz/20)
obtusallene V revised
(0.51Hz/20)
 omphadiol
(0.35Hz/4)
pallambin A
(0.22Hz/12)
palmarumycin CE3
(0.31Hz/12)
paucidactine D
(0.28Hz/11)
paucidactine E
(0.16Hz/22)
penicilleremophilane B
(0.34Hz/18)
 penifulvin A
(0.21Hz/12)
pepluacetal
(0.23Hz/12)
pepluanol A
(0.27Hz/11)
pepluanol B
(0.25Hz/14)
(-)-pericosine D - revised
(0.22Hz/5)
 phomarol
 phyllostacin J [ JHH 0.29Hz/11 ] - [ 1H 0.26ppm/18 ] - [ 13C 1.36ppm/20 ]
picrinine
(0.19Hz/13)
 protoilludanes dimer 18
(0.44Hz/10)
 protoilludanes alkene 13
(0.38Hz/12)
protoilludanes epoxide 15
(0.29Hz/10)
 protoilludanes lactone 19
(0.32Hz/9)
 purpurogenolide D
(0.28/8)
radudiol
(0.42Hz/15)
 salvilenone
(0.14Hz/3)
salvileucanthsin A [ JHH 0.48Hz/8 ] - [ 1H 0.18ppm/16 ] - [ 13C 2.01ppm/20 ]
schincalide A [ JHH 0.27Hz/13 ] - [ 1H 0.13ppm/21 ] - [ 13C 1.82ppm/29 ]
(+)-strictamine
(0.33Hz/13)
 strychnine
(0.18Hz/27)
tetrahydropicrotoxin sulfamate
(0.26Hz/4)
 tetrapetalone A-Me aglicon
(0.29Hz/8)
 trichothecolone-acetate
(0.28Hz/9)
(7S,15R)-7,15,25-Trihydroxy-acta-(16S,23R,24S)-16,23;16,24-binoxol
(0.21Hz/19)
uladane
(0.22Hz/2)
 wenyujinin H
(0.26Hz/12)
triterpene endoperoxide
(Keith Woerpel)(0.13Hz/8)