Guiding <i>ab initio</i> calculations by alchemical derivatives

Baben, Moritz to; Achenbach, Jan-Ole; Lilienfeld, O. Anatole von

doi:10.1063/1.4943372

Cited by 30 publications

(32 citation statements)

References 31 publications

(36 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…First order terms are firmly established for all variables through the Hellmann-Feynman theorem for changes in nuclear positions (to relax or run ab initio molecular dynamics 42 ), and charges. 22,27,32,33,36,37,[43][44][45] The derivative with respect to N is related to ionization potential and electron affinity by virtue of Koopman's and Janak's theorem, 46 and exhibits the well established derivative discontinuity at integer N , 47,48 so important for the construction of improved exchangecorrelation approximations. 49 Some elements in the Hessian, molecular vibrational normal modes, or the second order derivative of the electronic energy with respect to the number of electrons is the chemical hardness, introduced by Parr and Pearson.…”

Section: Theorymentioning

confidence: 99%

Alchemical Normal Modes Unify Chemical Space

Fias

Chang

Lilienfeld

2018

J. Phys. Chem. Lett.

View full text Add to dashboard Cite

In silico design of new molecules and materials with desirable quantum properties by high-throughput screening is a major challenge due to the high dimensionality of chemical space. To facilitate its navigation, we present a unification of coordinate and composition space in terms of alchemical normal modes (ANMs) which result from second order perturbation theory. ANMs assume a predominantly smooth nature of chemical space and form a basis in which new compounds can be expanded and identified. We showcase the use of ANMs for the energetics of the isoelectronic series of diatomics with 14 electrons, BN doped benzene derivatives (C 6−2x (BN) x H 6 with x = 0, 1, 2, 3), predictions for over 1.8 million BN doped coronene derivatives, and genetic energy optimizations in the entire BN doped coronene space. Using Ge lattice scans as reference, the applicability ANMs across the periodic table is demonstrated for III-V and IV-IV-semiconductors Si, Sn, SiGe, SnGe, SiSn, as well as AlP, AlAs, AlSb, GaP, GaAs, GaSb, InP, InAs, and InSb. Analysis of our results indicates simple qualitative structure property rules for estimating energetic rankings among isomers. Useful quantitative estimates can also be obtained when few atoms are changed to neighboring or lower lying elements in the periodic table. The quality of the predictions often increases with the symmetry of system chosen as reference due to cancellation of odd order terms. Rooted in perturbation theory the ANM approach promises to generally enable unbiased compound exploration campaigns at reduced computational cost. arXiv:1809.03302v2 [physics.chem-ph]

show abstract

Section: Theorymentioning

confidence: 99%

Alchemical Normal Modes Unify Chemical Space

Fias

Chang

Lilienfeld

2018

J. Phys. Chem. Lett.

View full text Add to dashboard Cite

show abstract

“…The rigorous definition of the chemical compound space (CCS), using insights from the grand‐canonical treatment of open systems and the so‐called conceptual density functional theory (DFT), has opened a fruitful avenue of research in the field of rational compound design . This allows us, in principle, to optimize a given property by analyzing its variation in response to changes in the molecular composition, that is, by navigating in CCS.…”

Section: Introductionmentioning

confidence: 99%

“…To conclude on an optimistic note, we should mention that there are nontrivial (exact) results of linear interpolating Hamiltonians within certain small families of systems, for example, the result of Von Lilienfeld for Hydrogenic atoms . Additionally, recent studies have demonstrated that one can find rules according to which the predictions made using the alchemical framework can reach significant accuracy . This shows that it is likely that for small families of molecules, analogous (approximate, but accurate) alchemical transformation formulas can be found by numerical fitting.…”

Section: Introductionmentioning

confidence: 99%

Interpolating Hamiltonians in chemical compound space

Miranda‐Quintana

Ayers

2017

Int J Quantum Chem

View full text Add to dashboard Cite

We analyze the Hamiltonian-interpolating schemes in chemical compound space. We show that if one allows for system-dependent information there are trivial solutions to this problem that allow to linearly interpolate between any two given iso-electronic Hamiltonians. If, on the other hand, one do not use any information about the system's Hamiltonians, we prove that there are no system-independent multiplicative interpolating procedures that linearize the energy. K E Y W O R D S alchemical changes, chemical compound space, conceptual density functional theory, interpolating Hamiltonians Int J Quantum Chem. 2017;117:e25384.

show abstract

“…It seems natural that the exploration of chemical space has improved by merging these tools with visual data analysis techniques; or the combination of quantitative structure–activity relationship (QSAR) methods with computational chemistry calculations . Inverse quantum chemical approaches are becoming popular to find a chemical structure departing from a set of desired properties; an outstanding example of these kind of methodologies is the alchemical potential approximation …”

Section: Introductionmentioning

confidence: 99%

Study of the Chemical Space of Selected Bacteriostatic Sulfonamides from an Information Theory Point of View

et al. 2016

View full text Add to dashboard Cite

The relative structural location of a selected group of 27 sulfonamide-like molecules in a chemical space defined by three information theory quantities (Shannon entropy, Fisher information, and disequilibrium) is discussed. This group is composed of 15 active bacteriostatic molecules, 11 theoretically designed ones, and para-aminobenzoic acid. This endeavor allows molecules that share common chemical properties through the molecular backbone, but with significant differences in the identity of the chemical substituents, which might result in bacteriostatic activity, to be structurally classified and characterized. This is performed by quantifying the structural changes on the electron density distribution due to different functional groups and number of electrons. The macroscopic molecular features are described by means of the entropy-like notions of spatial electronic delocalization, order, and uniformity. Hence, an information theory three-dimensional space (IT-3D) emerges that allows molecules with common properties to be gathered. This space witnesses the biological activity of the sulfonamides. Some structural aspects and information theory properties can be associated, as a result of the IT-3D chemical space, with the bacteriostatic activity of these molecules. Most interesting is that the active bacteriostatic molecules are more similar to para-aminobenzoic acid than to the theoretically designed analogues.

show abstract

Guiding ab initio calculations by alchemical derivatives

Cited by 30 publications

References 31 publications

Alchemical Normal Modes Unify Chemical Space

Alchemical Normal Modes Unify Chemical Space

Interpolating Hamiltonians in chemical compound space

Study of the Chemical Space of Selected Bacteriostatic Sulfonamides from an Information Theory Point of View

Contact Info

Product

Resources

About