Predictive Analytical Methods - Intertek Caleb Brett

ASG Analytical Sciences Group Laboratory

Computational Chemistry

Support of Analytical Results through Predictive Methods

Predicting Physical Properties

Calculation of pKa, logP (octanol:water partition coefficient), logD (dissociative octanol:water partition coefficient) and aqueous solubility values can be carried out for single molecules or for large chemical libraries. These predictions are based on experimental results which are held in an extensive database resulting in industry-leading degrees of accuracy.

The databases can also be searched to find literature values, results for similar molecules or molecules with similar properties. In addition, other physical properties can be calculated including, Hansen Solubility Parameters, bioconcentration factor (BCF) and electronic substitution constants (?, Hammett sigma values).

Predicting Magnetic Properties

NMR spectrometry is an important tool in the experimental characterisation of molecular systems and structures, as the chemical shifts obtained describe the environments of the atoms. Equivalent information can be obtained computationally from first-principles magnetic property calculations. These can produce high accuracy results for the entire range of molecular systems that can be studied experimentally via both NMR and ESR techniques. Chemical shifts and coupling constants can be predicted for all atom types and these can be used to help interpret the often complex patterns found in experimental data.

Predicting Vibrational Properties

Vibrational spectroscopy is a powerful tool for analytical chemists as it allows non-destructive analysis of samples in all physical states. Infrared and Raman techniques are complementary and provide information on the nature of the bonding within the molecule. Using computational methods, infrared and Raman frequencies and intensities can be predicted. The nature of the atomic movements giving rise to these vibrational modes can be interpreted through animated visualisations allowing complete and unambiguous assignment of experimental vibrational spectra to be made.

Predicting the Electronic Transitions of a Molecule

Understanding the nature of an excited state can be important in areas of chemistry dealing with coloured or electronic materials. Experimental methods such as UV/vis spectroscopy and various electrochemical techniques can provide some information on processes such as absorption of light or oxidation but to start to characterise the excited state more fully is usually very hard and often expensive. Using a combination of state of the art ab initio packages and other software developed in-house, individual excited states can be probed and the nature of the transition between the ground state and an excited state of interest can be described.

This approach can be used to predict the energy of the transition (measured experimentally as the wavelength of absorption) and its associated oscillator strength (directly related to the area under the experimental UV/vis curve). The effects of the transition on the electronic properties of the molecule can be investigated by visualising the electron density difference between the two states.