Assessing the activity of natural food antioxidants - Oxidation in Foods and Beverages and Antioxidant Applications - page 336

14.4 Theoretical prediction
Owing to great advances in computer science over the last decade, theoretical
methods have been recognized as a potentially useful tool in studies regarding
antioxidant activity. 24 In contrast to experimental studies, wherein the evalua-
tion of activity is often based on monitoring substrate changes or probe decay,
theoretical studies can predict radical scavenging activity of a compound of
interest as a function of various physicochemical parameters computed for an
optimized structure.
Various molecular descriptors have been utilized in antioxidant activity studies
of phenols. These include bond length values and order, the eigenvalue of the
highest occupied molecular orbital (HOMO), dipole moment values, differences in
heat of formation (HOF), phenolic O±H bond dissociation enthalpy (BDE), and
the vertical (VIP) or the adiabatic ionization potential (IP). 6,8,40±43 Of these, the
most frequently employed are BDE and IP values. The former is calculated
according to the formula: BDE = H r H h ÿ H p , where H r is the enthalpy of
phenoxyl radical generated after H atom abstraction, H h is the enthalpy of
hydrogen atom and H p is the enthalpy of the parent molecule; adiabatic ionization
potential value is given by the formula: IP = E c ÿE p , where E c is the energy of the
cation radical generated after electron transfer and E p is the energy of the parent
molecule. These two descriptors may well characterize the ability of phenols to
scavenge free radicals via HAT and/or SET mechanism previously described
(R14.1 and R14.2). 42 Low BDE and IP values predict high antioxidant activity.
Nevertheless, very low IP values may also indicate a possible prooxidant activity. 42
BDE and IP values can be computed using various quantum chemical
methods. So far, there is no defined method that is both accurate and efficient.
Semiempirical quantum chemical methods 44,45 that do not require the user to
have a strong theoretical background and training are adequate for experimental
researchers as a high-throughput means to screen and rank antioxidants. 24
Nevertheless, with semi-empirical methods, the theoretical values are typically
not consistent with experimentally determined ones, as the computational
approach utilized is simplified (e.g., the core electrons are not included and some
integrals are omitted). In addition, the solvent effect cannot be considered. In
order to take into account the latter and obtain more accurate data (with an error
of ~1±2 kcal/mole compared with experimental methods), density functional
theory (DFT) is suggested. 46±48 DFT, which has been developed more recently
than other advanced computational methods, usually employing B3LYP (Becke
3 term with Lee, Yang, Parr exchange), is the most widely used approach in
molecular calculations. DFT takes into account all electrons, is less computa-
tionally expensive than other methods (e.g., Mùller-Plesset, MP2) with similar
accuracy. 46 An alternative for saving time withought a significant compromise in
accuracy is the combination of AM1 (structure optimization) and DFT (single
point energy calculation) methods. 4,6,49
BDE and IP values have been successfully applied to structure-activity
relationship studies of various classes of natural phenolic compounds such as the