Solvatochromic shifts vs nanosolvation patterns: Uracil in water as a test case Costantino Zazza a,b,⇑ , Jógvan Magnus Olsen c , Jacob Kongsted c a Supercomputing Center for University and Research, CASPUR, via dei Tizii 6/b, 00185 Rome, Italy b Centro di Calcolo della Scuola Normale Superiore di Pisa, Palazzo del Castelletto, via del Castelletto, 11 56126 Pisa, Italy c Department of Physics and Chemistry, University of Southern Denmark, Campusvej 55, DK-5230 Odense, Denmark article info Article history: Received 1 June 2011 Received in revised form 18 July 2011 Accepted 18 July 2011 Available online 4 August 2011 Keywords: Density functional theory Molecular dynamics Solvation effects abstract We address the role played by a different subset of nanosolvation patterns on the lowest p–p / and n–p / solvatochromic shifts of uracil in water at 298 K. To this end, a computational investigation which com- pares perturbed electronic properties based on correlated electronic wavefunctions – computed either at the DFT or CCSD level of theory – in conjunction with two nanoseconds time-scale classical trajectories is presented. The simulations are based on the use of either an implicit or an explicit description of polar- ization. The predicted p–p / solvatochromic shifts based on these two simulations are very similar but slightly underestimate the experimental values. This finding reinforce the idea that for a proper estima- tion of the p–p / transition local interactions beyond those of purely electrostatic origin are mandatory. On the other hand, the n–p / excitation is seen to increase its shift upon inclusion of explicit polarization, thereby bringing the predicted shift in closer agreement with experimental data. Ó 2011 Elsevier B.V. All rights reserved. 1. Introduction Chemistry is ‘‘naturally’’ devoted to deduce from single-mole- cule-level observations made on the basis of a statistically aver- aged ensemble characterized by a huge number of interacting and fluctuating molecules. For this reason, theoretical/computa- tional modelling of real chemical systems has always represented an intimately complicated task. In this respect, theoretical chemis- try has reduced the object of the observations thus making the connections with usual evidences detected in the laboratory not al- ways possible. On the other hand, in the last years the general trend is that of trying to model chemical species maintaining as much as possible the basic features of the real system in silico. This advancement has become possible through the development of first-principles methodologies designed to achieve spectroscopic properties in condensed phase matter. Therefore, we are now in the position of theoretically addressing, at the molecular level, important observables of large systems which are strongly modu- lated by the conformational space accessible via thermal fluctua- tions. This fascinating topic obviously requires extended simulations which plausibly have to be carried out with force fields of relatively high accuracy. In all these cases, however, a full quan- tum mechanics (QM) approach is still limited to systems of up to hundred of atoms, or even smaller when the highest levels of theory are applied. This is basically the reason why the most pop- ular methods for simulating molecular system in silico conditions in a feasible way are the so called quantum-mechanics/molecu- lar-mechanics (QM/MM) ones [1–4]; QM/MM algorithms allow us to practically combine ab initio electronic structure calculations on a selected molecular target with a molecular mechanics treat- ment of the surrounding environment. This decomposition is obvi- ously needed because empirical MM force fields are not able to describe bond-breaking along chemical reaction pathways, charge transfer processes and/or electronic excitations. Such events unequivocally require a full QM picture. In this context, much attention has been paid to biologically relevant molecules in dilute water solution, and in particular to a better understanding of sol- vent effects in different contexts [5–11]. This is witnessed by sev- eral investigations dealing with statistical mechanically averaged molecular properties of different solutes in liquid water [12–17]. Among chemical properties under physiological conditions, accurate prediction of solvatochromic shifts has represented a very attractive step in this research area and it can be considered as one of the greatest challenges to modern quantum chemistry. Uracil (Ur), because of its key role in DNA [18–20] and RNA [21,22] in aqueous solution, has been the object of a series of experimental [23–27] and computational investigations [13,28–36] mainly de- voted to rationalize changes in the electronic properties when it passes from gas-phase to a water environment. The experimental UV spectrum of aqueous Ur is characterized by a broad peak at 4.8 eV and a second one above 6.0 eV [23–27]. The first signal is assigned to a 1 p ? p / transition undergoing in water solution a red shift in the range of 0.2–0.3 eV. On the contrary, the weak 1 n ? p / electronic excitation is known to be shifted in the opposite 2210-271X/$ - see front matter Ó 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.comptc.2011.07.020 ⇑ Corresponding author at: Centro di Calcolo della Scuola Normale Superiore di Pisa, Palazzo del Castelletto, via del Castelletto, 11 56126 Pisa, Italy. E-mail address: costantino.zazza@sns.it (C. Zazza). Computational and Theoretical Chemistry 974 (2011) 109–116 Contents lists available at SciVerse ScienceDirect Computational and Theoretical Chemistry journal homepage: www.elsevier.com/locate/comptc