On the Light-Driven Isomerization of a Model Asymmetric Molecular Rotor:Conformations and Conical Intersections of 2-Cyclopentylidene-tetrahydrofuran

Mariana Assmann, Guillermo Perez-Hernandez, and Leticia Gonzalez*Institut fur Physikalische Chemie, Friedrich-Schiller-UniVersitat Jena, Helmholtzweg 4, 07743 Jena, Germany

ReceiVed: May 28, 2010; ReVised Manuscript ReceiVed: July 16, 2010

The ground state potential energy surface of the model molecular rotor 2-cyclopentylidene-tetrahydrofuran(CPTHF) has been characterized by calculating minimum energy conformations, racemization pathways, androtational barriers with high level ab initio electronic structure calculations. Two conformers with theircorresponding enantiomers are found. The activation barriers for racemization are negligible, therefore thermalracemization takes place at room temperature. Torsional transition states, calculated using multiconfigurationalCASSCF calculations, show twisted and pyramidalized biradical structures. Additionally, the photochemistryof CPTHF has been investigated using the accurate MS-CASPT2/CASSCF methodology. In the UV spectrumit is found that the spectroscopic state is the S1, which corresponds to a ππ* transition within the ethylenemoiety. To understand light-triggered isomerization around the CdC bond, five conical intersections betweenthe S0 and S1 have been located for each conformer of CPTHF, which allow the system to rapidly decay tothe electronic ground state.

Introduction

An important branch of nanoscience is the construction ofmolecules capable of mimicking mechanical machines. Theoperating principle of a molecular machine is the rearrangementof electrons or nuclei induced by external chemical, photo-chemical, or electrochemical stimulation.1,2 A particularlypromising class of molecular machines are those that exploitphotochemical processes. Powering a machine with light hasimportant advantages, for example, the fact that waste productsare avoided and that spectroscopic methods can be used tofollow the state of the machine itself. As a consequence, aconsiderable effort has been done in the last two decades todesign photochemically driven molecular machines.3

Among the components that comprise a molecular machine,an essential part is the motor because it allows to convert energyinto mechanical motion. One of the most exciting directions ofconstructing motors at a molecular level is to functionalize chiralovercrowded alkenes.4 The first example of an artificial mo-lecular motor based on a chiral overcrowded alkene was reportedin 1999.5 This was a symmetric biaryl overcrowded alkene inwhich repetitive unidirectional 360° rotation around a carbon-carbon double bond is achieved in four steps: two light-inducedcis/trans isomerizations are each followed by a thermal-inducedhelicity inversion. Since 1999, a large number of molecularrotors that exploit the same photochemical and thermal schemehave been synthesized.6 One of the smallest (and latest)molecular rotors based on overcrowded alkenes is shown inScheme 1, where also the four steps of the rotary cycle are linedout. In this rotor, A, the upper and lower halves consist ofdimethyl-substituted phenyl groups that are easy to synthesizeand provide faster rotations than the first and second generationrotors designed by the same authors.7

Other successful rotors have been synthesized, with distinctupper and lower halves as well as with heteroatoms.8 Suchasymmetric molecular rotors are particularly interesting since

they could be recently anchored on gold surfaces,9 goldnanoparticles,10 and quartz surfaces.11 The upper half acts as apropeller and it is free, while the lower half of the rotor, thestator, is immobilized to the surface by two legs terminated withthiol groups. In this way, uncontrolled thermal rotation of theentire motor can be avoided, and the rotor is prepared to belinked to macroscopic parts.

Regardless of whether the molecules are free or anchored,two features stand out as mandatory for unidirectional rotationwithin the mentioned variety of rotors. First, when photoisomer-izing the central double bond with unshaped UV-light, therotation must occur only toward one of the two possibledirections, namely the positive or negative variation of thedihedral angle. This is granted by the rotors’ built-in asymmetryaround the central double bond. Different steric impedimentsarise in both directions, and net rotation will occur along the

* To whom correspondence should be addressed: E-mail: leticia.gonzalez@uni-jena.de.

SCHEME 1: Rotary Cycle of A. Adapted from Ref 7

J. Phys. Chem. A 2010, 114, 9342–93489342

10.1021/jp104898t 2010 American Chemical SocietyPublished on Web 08/12/2010

least energetic pathway. The second criterion concerns thethermal steps of the rotatory cycle. The stability differencebetween the two helicity isomers must be large enough for thethermal equilibrium to be totally displaced to the product side.Again, this stability difference has its origin in the specificinterplay of the moieties crowding the fjord-region (see Scheme1). Just by warming up the sample, the rotors irreversiblyundergo the helicity-inversion. This helicity-inversion bears anassociated twist of the central double bond that keeps the samedirection as the torsion of the photoinduced isomerization, sothat the overall unidirectionality is conserved.

In this paper we investigate a reduced model for a molecularrotor: 2-cyclopentylidene-tetrahydrofuran (abbreviated CPTHF),see Figure 1. Rotation around the CdC bond connects twoisomers that are indistinguishable. Several reasons have beenconsidered for choosing the present system. Inspired by therotors based on alkenes, CPTHF shares the central bicyclopen-tylidene scaffolding of the rotor shown in Scheme 1, but oneof the carbon atoms has been substituted by an oxygen atom. Itlacks, however, the steric crowding of A. Nevertheless, weemphasize that our long-term interest is the ignition of unidi-rectional rotations in molecular rotors using ultrafast laserpulses,12-16 where, as opposed to A, rotation does not involvethermal steps. In this respect, in our model the control parameteris not the difference in stability, but rather the dipole moment,or more precisely, its asymmetry along the torsion coordinate.This asymmetry is a desired feature to induce direct and efficientlaser transitions from the ground to the excited state in the caseof symmetric potential energy surfaces (PES).16 This is thereason why the bicyclopentylidene scaffolding has been dopedwith an oxygen atom. Finally, given its reduced size, CPTHFallows for accurate ab initio electronic structure calculationsand laser-driven quantum dynamical simulations.

As a first step toward the goal of inducing unidirectionalrotations using laser pulses, in this paper we investigate the PESof CPTHF both in the ground and in the first singlet electroni-cally excited state. CPTHF has different enantiomeric forms,both in the local minima of the PES as well as in the transitionstates (TSs). A complex PES arises, and different pathways forrotation are possible. In this paper, we investigate the landscapeof the electronic ground state of CPTHF and try to establishpossible rotation mechanisms. Because the rotation can betriggered with light via an excited state, the photochemistry ofCPTHF is also characterized. The UV-vis absorption spectrumis computed and critical points in the excited state PES areoptimized. Of particular interest will be the existence of conicalintersections (CIs) that play a key role in the relaxation to theground state and thus in the overall rotation mechanisms in theexcited state.

Computational Methods

The equilibrium structures of CPTHF in the electronic groundstate were optimized at the MP2/6-31G(d,p) level of theory.The five-membered rings adopt nonplanar puckered conforma-tion and thus, several minima are possible depending whether

the carbon atoms of the ring are below or above the molecularplane defined by the ethylene backbone. This out-of-planemovement provides a very flat ground state PES. The TSsconnecting the different minima are also optimized at MP2/6-31G(d,p) level of theory.

The vertical absorption spectrum of CPTHF was calculatedusing the complete active space self consistent field17 (CASSCF)method with the 6-31G(d) basis set. Different active spaces,ranging from (2,2) to (8,6), have been tested. The minimumactive space CAS(2,2) consists of the πCCπCC* molecular orbitalssituated at the central CdC bond. The larger active spacessystematically add the πOCC orbital located at the oxygen atom(CAS(4,3)), the lone pair of the oxygen (CAS(6,4)), and aσCOσCO* pair (CAS(8,6)). As it will be shown in Section 3, theintermediate active space (4,3) is enough to describe the systemadequately, and therefore, the rest of the calculations are carriedout with this active space. The πOCC orbital and the pair ofπCCπCC* orbitals are shown in Figure 2. Depending on the sizeof the active space, different numbers of states have been equallyaveraged, so that the wave functions are properly balanced. Theresulting energies have been corrected with complete activespace second-order perturbation theory in its multistate version,18

MS-CASPT2, and the same 6-31G(d) basis set. The level shifttechnique19 with a parameter value of 0.2 au was employed tocorrect for intruder states. This parameter was chosen after acareful examination of the stability of the excitation energiesand the comparison of the excited states reference weight withthat of the ground state.

The rotational motion of CPTHF around the CdC bondinvolves twisted TSs of biradical nature, for which the use ofmulticonfigurational methods is mandatory. Following theanalysis of the active space to calculate vertical excitations, thebiradical TSs have been optimized using CASSCF(4,3)/6-31G(d)in the ground state. For consistency, the energies of the minimain the electronic ground state were also recalculated at theCASSCF(4,3)/6-31G(d) level of theory.

The optimization of CIs between the ground and firstelectronically excited singlet states (S0 and S1, respectively) isperformed using state-averaged wave functions of the twoinvolved states (SA2) at the CASSCF(4,3) level with the6-31G(d) basis set.

Starting at the Franck-Condon geometries in the S1, aminimum energy path (MEP) calculation was carried out in the

Figure 1. CPTHF with used atom labels and coordinate system.

Figure 2. (a) CPTHF. (b-d) Orbitals included in the CASSCF(4,3)/6-31G* calculations.

Ground State Potential Energy Surface of CPTHF J. Phys. Chem. A, Vol. 114, No. 34, 2010 9343

excited state surface, leading to the geometry of one CI, whichwas then used as starting point to obtain further nine CIs. Theso obtained CIs were then connected with the correspondingminima in the ground state using intrinsic reaction coordinate20,21

(IRC) calculations. TS geometries were also associated to theirminima through IRC calculations.

Except for the MS-CASPT2/CASSCF calculations of theUV-vis absorption spectrum and the MEP calculation, whichare done with the MOLCAS 7.0 package,22 the rest of optimiza-tions and single point calculations are done with the Gaussian03 package.23

Results and Discussion

Conformational Studies in the Ground State. As shownin Figure 3, both five-membered rings of CPTHF are puckeredin a half-chair conformation. As a consequence, CPTHFpossesses two stereoisomers, 1 and 2, each of them with acorresponding enantiomer, 1′ and 2′. The four minima can bedistinguished by the atoms C4, C5, C8, and C9 (recall atomnumbering in Figure 1) which are differently above or belowthe molecular plane. The molecular plane is described by thexz-plane of the coordinate system in Figure 1, and it correspondsto the plane formed by the ethylenic backbone. Then, aboveand below the plane is positive and negative y-direction. Withthis definition, stereoisomer 1 has the carbons C4 and C8 abovethe plane, whereas C5 and C9, at the other side of the doublebond, are below the plane. Stereoisomer 2 has the carbons C5

and C8 above the plane and C4 and C9 below the plane. Theposition of the four carbons is reversed for the mirror images1′ and 2′. Thus, the stereoisomers 1 and 2 have a similardisposition of the ring containing exclusively carbon atoms(henceforth called carbon-ring) and differ in the five-memberedring containing the oxygen atom (henceforth called oxygen-ring).

In Table 1 we specify the position of the carbon atoms of allthe structures here reported and corresponding enantiomers. Theposition of the relevant C atoms is labeled with “d” for down,that is, below the molecular plane, and with “u” for up, that is,above the plane. The stereoisomers 1 and 2 are related by afloppy skeleton motion of the carbons and therefore, within theaccuracy of the employed method, they can be considereddegenerated. For pragmatic reasons, the rest of the energies willbe given with respect to 1.

The geometrical parameters of 1 and 2 are summarized inTable 2. Here, the bond lengths of the C1-C2 double bond and

the C2-C3 bond as well as some selected angles and dihedralsare given. It can be seen that the length of the double bond isnearly the same in 1 and 2; the same applies to the C2-C3 bond.The C1-C2 bond with values around 1.339 Å is in the range ofa typical C-C double bond, whereas the C2-C3 bond withvalues of 1.390 Å is shorter than the C4-C3 bond, which isabout 1.438 Å in all stereoisomers, and is shorter than in freeTHF (1.430 Å, value calculated at MP2/6-31G(d,p) level oftheory). The shortening of the C2-C3 bond is a hint for theparticipation of the p-orbital of the oxygen atom in the doublebond system. The dihedrals C6-C2-C1-C7 are all very closeto 180°, showing that the system formed by the double bondand the neighbored atoms is nearly planar. The presence of theoxygen atom obviously makes the oxygen-ring not symmetric.The other selected dihedrals show that the carbon-ring is alsonot symmetric.

The minima 1 and 2 are linked by TSs in which twoneighbored carbon atoms, either C4 and C5 or C8 and C9, areboth above (denoted by u) or below (denoted by d) themolecular plane, so that the five-membered ring takes anenvelope form, see Figure 4. Note that in the minima these atomsare never on the same side of the ethylenic plane. The TSs arenamed after the minima they connect, as confirmed with thehelp of IRC calculations. The corresponding geometricalparameters are listed in Table 2.

The interconversion between 1 and 2 proceeds via twopossible TSs: TS12u and TS12d. Both have similar very smallactivation energies, 2.81 and 2.88 kcal/mol, respectively. Asshown in Table 1, TS12u has the C4 and C5 atoms up, so thatthe oxygen-ring takes an envelope form, while the carbon-ringis twisted in a half-chair as in the minima. In TS12d, the C4

and C5 atoms also take an envelope form, but they are down.The imaginary frequencies of both TSs are 153i and 154i cm-1,respectively, and as expected, they correspond to the movementof C4 and C5 in opposite directions, from envelope to the half-chair form.

The racemization pathway of 1 into 1′ has also beencalculated. For this purpose, the TSs connecting 2 and 1′ arenecessary. The corresponding saddle points, TS1′2u and TS1′2dhave relative energies of 2.82 and 2.94 kcal/mol at MP2/6-31G(d,p) level of theory, respectively, see Figure 4. Both TSshave the atoms C8 and C9 simultaneously up or down so thatthe carbon-ring is in envelope form, whereas the oxygen-ringis twisted in half-chair conformation. The corresponding

Figure 3. Front and side view of the ground state isomers (1 and 2)and enantiomers (1′ and 2′) of CPTHF. MP2/6-31G(d,p) and CASS-CF(4,3)/6-31G(d) (in brackets) energies in kcal/mol.

TABLE 1: Position of the Atoms C4, C5, C8, and C9 in AllGeometries Including the Enantiomersa

structure C4 C5 C8 C9 enantiomer C4 C5 C8 C9

1 u d u d 1′ d u d u2 d u u d 2′ u d d uTS12u u u u d TS1′2′d d d d uTS12d d d u d TS1′2′u u u d uTS1′2u d u u u TS12′d u d d dTS1′2d d u d d TS12′u u d u uTS11L/R u d u d TS1′1′R/L d u d uTS22L/R d u u d TS2′2′R/L u d u dCI1-C1 L/R u d u d CI1′-C1R/L d u d uCI1-C2 L/R u d u d CI1′-C1R/L d u d uCI1-C1 Le u d u u CI2′-C1Re d u u uCI2-C1 L/R d u u d CI2′-C1R/L u d d uCI2-C2 L/R d u u d CI2′-C2R/L u d d uCI2-C1Re d u d d CI1′-C1Le u d d d

a “u” (up) means the C atom is above and “d” (down) means it isbelow the plane defined by the ethylene moiety. The labels u and dcorrespond to the unrotated and unpyramidalized structures.

9344 J. Phys. Chem. A, Vol. 114, No. 34, 2010 Assmann et al.

imaginary frequencies of ca. 111i cm-1 show a similar move-ment of these C atoms as in the case of TS12u and TS12d.Therefore, the racemization of 1 goes first to 2 through eitherTS12u or TS12d and then to 1′ via TS1′2u or TS1′2d. The

racemization of 1 to 1′ is also possible through 2′. This pathwayincludes the enantiomers of the TSs considered above, seeFigure 4.

For completeness, we report that CPTHF with both five-membered rings planar corresponds to a saddle point of fourthorder, SP4, with an MP2/6-31G(d,p) energy of 7.72 kcal/mol.Its four imaginary frequencies show movements of C4 and C5

or C8 and C9 in the same or the opposite direction, leading tothe TSs described above.

For further investigations of the PES we shall consider onlygeometries connected to the stereoisomers 1 and 2. Thecorresponding enantiomers of these and other critical points(listed in Table 1) will not be shown.

The TSs reported so far keep the double bond planar.However, a rotation of a double bond in the electronic ground

TABLE 2: Selected Bond Lengths (in Å), Angles, and Dihedral Angles (Both in Degrees) of the Geometries of CPTHFOptimized with MP2/6-31G(d,p) (Minima and Planar TSs) or CASSCF(4,3)/6-31G(d) (Twisted TSs and CIs)a

structure C1-C2 C2-C3 C6-C2-C1-C7 C1-C2-C3-C4 C1-C2-C6-C5 C2-C1-C7-C8 C2-C1-C10-C9 C2-C1-X25 C1-C2-X26

1 1.339 1.390 -180 -175 -161 170 1651′ 1.339 1.390 180 175 161 -170 -1652 1.338 1.390 -180 176 160 168 1672′ 1.338 1.390 180 -176 -160 -168 -167TS12u 1.338 1.384 180 -156 161 164 171TS12d 1.339 1.384 180 157 -162 173 162TS1′2u 1.339 1.391 180 180 157 -142 -143TS1′2d 1.334 1.373 180 -172 166 -147 148TS11L 1.472 1.379 128 -155 169 -178 155 165 150TS11R 1.472 1.377 74 167 -140 155 -178 165 150TS22L 1.471 1.377 124 -166 140 156 -179 166 150TS22R 1.472 1.378 71 155 -169 154 -177 164 149CI1-C1L 1.405 1.269 -58 -179 -167 120 -145 127 177CI1-C1R 1.400 1.286 -120 -176 -171 -146 121 126 173CI1-C1Le 1.410 1.279 -55 -178 -166 166 -166 127 177CI1-C2L 1.451 1.523 161 -115 109 169 168 175 106CI1-C2R 1.451 1.510 21 116 -111 155 -178 172 106CI2-C1L 1.405 1.270 -62 179 166 125 -148 127 177CI2-C1R 1.400 1.287 -126 176 170 -145 121 124 173CI2-C1Re 1.403 1.295 -128 176 170 -165 166 126 173CI2-C2L 1.460 1.490 165 -109 68 165 173 172 108CI2-C2R 1.452 1.509 21 112 -71 157 178 173 107

a The corresponding labels of the atoms in the different structures can be seen in Figures 1 and 5.

Figure 4. Interconversion and racemization pathways from 1 to 1′ via 2 and 2′. MP2/6-31G(d,p) energies in kcal/mol.

Figure 5. Biradical transition states of CPTHF. CASSCF(4,3)/6-31G(d)energies in kcal/mol. In panel (a), the atom numbering with dummyatoms X25 and X26 is given.

Ground State Potential Energy Surface of CPTHF J. Phys. Chem. A, Vol. 114, No. 34, 2010 9345

state involves a twisted and biradical TS. These TSs are muchhigher in energy than that of the planar TSs, since the rotationinvolves the rupture of the CdC bond. We have found twotwisted TSs for each of the stereoisomers, 1 and 2, see Figure5. Their imaginary frequencies are between 260i and 270i cm-1

and correspond to the torsion around the CdC bond.As it can be seen in Table 2, the C1-C2 bond of all the twisted

TSs is as expected elongated to a C-C single bond (1.47 Å).Moreover, the twisted TSs are slightly pyramidalized in C1 andC2. This is best appreciated in the angles C2-C1-X25 andC1-C2-X26, which indicate the grade of pyramidalization withthe aid of the dummy atoms X25 and X26. The atom X25 issituated in the middle of an imaginary line between C7 and C10.Likewise, X26 is in the middle of the line joining C3 and C6,see Figure 5a. Compared with 180°, which correspond to anonpyramidalized structure, the angles C2-C1-X25 andC1-C2-X26 indicate a pyramidalization at C1 and C2 of ca. 165and 150°, respectively. In all twisted TSs, the degree ofpyramidalization of C2 is higher than the one of C1.

The TSs corresponding to the torsion of 1 or 2 are namedTS11 or TS22, respectively. These saddle points of first ordershow the same configurations of the C4, C5, C8, and C9 atomsas the stereoisomers they come from (cf. Table 1). Moreover,we indicate with L (left) or R (right) in which direction alongthe x-axis the oxygen-ring is pyramidalized. We define left tobe in negative and right to be in positive x-direction. TheCASSCF(4,3)/6-31G(d) energies amount to ca. 74 and 62 kcal/mol for the left- and right-pyramidalized TSs, respectively.Therefore, the activation barriers are ca. 10 kcal/mol lower forthe saddle points TS11/22R.

UV-vis Absorption Spectrum of CPTHF. Table 3 collectsthe CASSCF and MS-CASPT2 vertical excitation energiesobtained for CPTHF with different active spaces. The smallestactive space (2,2), including the πCC and the πCC* orbitalslocalized mainly at the CdC bond (see Figure 2), predicts anintense bright S1 state of ππ* character at 6.61 eV (SA2-CASSCF: 8.36 eV) with an oscillator strength f ) 0.475. Thisenergy is typical of olefinic systems. The valence state ofethylene itself is obtained experimentally at 7.65 eV,24 andtheoretically at MS-CASPT2 level of theory at 7.98 eV.18 Inrelated systems, such as bicyclohexylidene25 or 4-(methylcy-clohexylidene)fluoromethane,26 similar levels of theory predictthe bright π f π* transition at ca. 6.5-6.8 eV, in accordanceto available experimental values, see, for example, ref 27.

The inclusion of the p-orbital of the oxygen (πOCC) in theactive space (4,3) allows for resonance within the π-systemformed by the p-orbitals of the atoms C1, C2, and C3. Asexpected, the S1 ππ* excited state is stabilized and thus, theexcitation energy decreases to 6.57 eV. The oscillator strengthis very similar to that obtained with CAS(2,2) (f ) 0.483). Toinvestigate the effect of the nonbonding electron pair of theoxygen, the nO-orbital was included and SA4-CASSCF(6,4)

calculations were performed. The S1 state corresponds to theππ* excitation at 6.39 eV. The S2 state is ca. four times weakerthan the ππ* S1 state and corresponds to a πOCCf πCC* transitionwith an excitation energy of 9.13 eV. The S3 is very weak, andit is a nO f πCC* excitation predicted at 9.22 eV.

The influence of σ-orbitals was investigated using the activespace (8,6) by adding the σCO σCO* pair to the (6,4) active space.In this case, six states need to be included in the averagedcalculation, otherwise the πOCCf πCC* transition does not appearwithin the calculated wave functions. In the SA6-CASSCF(8,6)calculations, the πCC f πCC* transition at 6.46 eV is still thebrightest state with an oscillator strength of 0.433ssimilar tothe results obtained with other active spaces. However, the S2

state, predicted at 8.15 eV, corresponds now to a dark πCC fσCO* transition. This state has intercalated between the ethylenicππ* state and the πOCCπCC* , which now is the S3 state at 9.56eV. The S4 state corresponds to the weak nO f πCC* state (9.71eV), and the S5 state is a double excitation from the πCC to theπCC* -orbital (11.79 eV).

As we can see, all the calculations are consistent with thefact that CPTHF should show a band peaking around 6 eV anda shoulder at around 9 eV. The S1 is the brightest state, andcorresponds to a ππ* (HOMO-LUMO) transition within thecarbon-carbon double bond. Because of the presence of theoxygen atom, an additional ππ* state weaker than the previousone is responsible for the shoulder at higher energies. To thetail of the band contributes a weak state of nπ* character.

The excitation energy of the ππ* S1 state is robust regardlessof the active space, so are the associated oscillator strengths.However, at MS-CASPT2/SA6-CASSCF(8,6) level of theory,the πOCCπCC* transition is predicted much weaker than at MS-CASPT2/SA4-CASSCF(6,4), and it is destabilized by 0.4 eVbecause a πσ* state intercalates in between. Obviously, the SA-CASSCF procedure is more accurate when few states arecalculated, and it can deteriorate very quickly if a large numberof states are calculated with the same set of averaged orbitals.Since the πσ* as well as nπ* states are very weak and do notplay a role in the photochemistry of CPTHF, we consider thatthe (4,3) active space provides a satisfactory description of thebright states. From those, the spectroscopic one is the S1 of ππ*character, in analogy to the valence state of ethylene.28,29

Following this analysis, we shall employ SA2-CASSCF(4,3)to study the radiative and nonradiative (via CIs) relaxationpathways from the S1 excited state of CPTHF.

Relaxation Pathways and Conical Intersections. Five CIscoupling the S1 and S0 states have been located for 1 and fiveanalogous for 2, see Figure 6. Their labels CI1 and CI2 referto the stereoisomers 1 and 2, respectively. All 10 CIs are twistedand pyramidalized at either C1 or C2. Accordingly, they aredenoted by CI-C1 or CI-C2 and R or L, with the samemeaning as in the TSs. Six CIs were found pyramidalized atC1, these are CI1-C1L, CI1-C1R, CI1-C1Le, CI2-C1L,

TABLE 3: Excitation Energies ∆E (in eV) Obtained with MS-CASPT2/CASSCF/6-31G(d)a

MS-CASPT2(2,2) MS-CASPT2(4,3) MS-CASPT2(6,4) MS-CASPT2(8,6)

configuration c ∆E (eV) f c ∆E (eV) f c ∆E (eV) f c ∆E (eV) f

(πCC)2 0.989 -0.981 -0.980 -0.975(πCC)1(πCC* )1 0.998 6.61(8.36) 0.475 0.991 6.57(8.46) 0.483 -0.964 6.39(8.29) 0.482 0.921 6.46(8.27) 0.433(πOCC)1(πCC* )1 -0.857 9.13(10.55) 0.123 0.845 9.56(10.69) 0.061(nO)1(πCC* )1 0.936 9.22(9.93) 0.015 -0.925 9.71(10.12) 0.014(πCC)1(σCO* )1 0.921 8.15(9.16) 0.007(πCC)0(πCC* )2 0.900 11.79(12.85) 0.012

a CASSCF energies are given in parentheses. Also given are the configuration interaction coefficients, c, of the contributing wave functionand the oscillator strength, f.

9346 J. Phys. Chem. A, Vol. 114, No. 34, 2010 Assmann et al.

CI2-C1R and CI2-C1Re. Among them, CI1-C1Le andCI2-C1Re show a carbon-ring in an envelope form (e), whilethe others adopt half-chair conformations in all the rings. Acloser inspection on the configurations (recall Table 1) revealsthat CI1-C1Re is the enantiomer form of CI2-C1Re, andlikewise CI2-C1Le is not reported in Figure 6 and Table 2since it is the enantiomer of CI1-C1Le. The remaining fourCIs, CI1-C2L, CI1-C2R, CI2-C2L, and CI2-C2R, arepyramidalized at C2.

At CASSCF(4,3) level of theory, the energies of the CI-C1sare around 105 kcal/mol (ca. 4.5 eV), whereas the CIspyramidalized at C2 (CI-C2s) are higher in energy, at ca. 128kcal/mol (ca. 5.5 eV). The geometrical parameters whichcharacterize the CIs are listed in Table 2. Similar to the twistedTSs, the CIs have biradical character and hence, the C1-C2

distance is elongated to a single bond. In the case of the CI-C2

structures the C1-C2 bond is slightly longer than in the CI-C1

ones. Furthermore, the CI-C1 structures show a contraction ofthe C2-C3 bond (around 1.27 Å) toward a CdO bond. In theCI-C2s the C2-C3 bond (around 1.5 Å) is longer than a typicalC-O single bond. Observing the dihedrals C6-C2-C1-C7, onecan see that in the CI-C2s the torsion differs only around 20°from a planar structure but the C2-C3 bond arranges nearlyperpendicular to the molecular plane. Therefore, the displace-ment of the oxygen-ring with respect to the planar structure ishigher than in the CI-C1 s and the pyramidalization of C2 ismuch larger than the one of C1 in the CI-C1s. This is bestobserved comparing the angles C1-C2-X26 and the anglesC2-C1-X25. As in the twisted TSs, X25 and X26 are dummyatoms, situated as shown in Figure 5a. In the CI-C1s the angleC2-C1-X25 is around 126°, which corresponds to a deviationfrom the planar structure (180°) of about 54°. In the CI-C2sthe angle C1-C2-X26 is around 107°, indicating a largerdeviation from the planar structure of ca. 73°.

An IRC calculation along the gradient difference vector ofCI1-C1L in the ground state leads to 1 (in one direction) andto 2 (in the other direction). Furthermore, the optimization ofthe Franck-Condon geometry of 1 in the S1 as well as a MEPcalculation beginning at the same point ends in CI1-C1L. Thus,it appears that this CI is the most important one to relax to theelectronic ground state S0.

Conclusions

In this paper, the ground and first excited state potentialenergy surfaces of the model molecular rotor CPTHF have beeninvestigated. In the electronic ground state, two degeneratedstereoisomers are possible. The two stereoisomers appear dueto the distortion suffered in cyclopentane and tetrahydrofuranrings to release strain. Both stereoisomers have enantiomeric

counterparts (giving rise to four stereoisomers), related by theposition up or down of the carbons of the ring with respect tothe plane defined by the ethylene backbone. The racemizationpathways have been calculated, finding that the activationbarriers are negligible and that thermal racemization takes placeat room temperature. The rotational profiles have also beeninvestigated, finding activation energies of 63.25 kcal/mol atCASPT2 level of theory.

The UV-vis absorption spectrum of CPTHF has beencalculated. The spectroscopic state is the first electronicallyexcited state, which is characterized by a ππ* transition withinthe ethylene moiety. This state is predicted at ca. 6.5 eV, whichis characteristic of olefinic systems. This value is shifted to lowerenergies in the molecular rotors based on overcrowded alkenes(e.g., 4 eV in ref 7). Because of the oxygen atom, anothervalence state is present in CPTHF, but it is higher in energy(ca. 9 eV) and well separated from the S1. Therefore, we canconsider that the relevant state to describe the photochemistryof CPTHF is the S1. Five different conical intersections havebeen located for each of the stereoisomers of CPTHF. All ofthem are twisted and pyramidalized at the ethylenic carbons,as ethylene does. Among the CIs with lower energy, one CI(CI1-C1L) is expected to be first accessed from the S1 to relaxto the S0.

The low isomerizations barriers in the ground state for thereactions 1T 2′ and 2T 1′ point toward no distinction between1 and 2′ as well as 2 and 1′ at room temperature. Thus,unidirectionality solely by warming it up is only possible ifCPTHF is functionalized by crowding the ring moieties.However, ultrafast laser pulses can be used to achieve unidi-rectional rotation in systems like CPTHF, where the PES aresymmetric around the minima, solely with the use of light andthus avoiding the thermal step, see, for example, ref 16. Workalong these lines is in progress.

Acknowledgment. The authors thank the Deutsche For-schungsgemeinschaft for financial support (GO 1059/2-1). Weare grateful to A. Hauser and M. Brautigam for their preliminarycalculations on the UV spectrum of CPTHF. Most of thecalculations have been performed in the HP computers of theTheoretical Chemistry group at the Friedrich-Schiller-University(FSU) of Jena. The computer center of the FSU is alsoacknowledged for allocation of computer time.

References and Notes

(1) Feringa, B. L. Molecular Switches, 1st ed.; Wiley-VCH: Weinheim,2001.

(2) Balzani, V.; Credi, A.; Venturi, M. Molecular DeVices andMachines: A Journey into the Nanoworld, 1st ed.; Wiley-VCH: Weinheim,2003.

(3) Credi, A.; Venturi, M. Cent. Eur. J. Chem. 2008, 6, 325.(4) Feringa, B. L.; van Delden, R. A.; Koumura, N.; Geertsema, E. M.

Chem. ReV. 2000, 100, 1789.(5) Koumura, N.; Zijlstra, R. W. J.; van Delden, R. A.; Harada, N.;

Feringa, B. L. Nature 1999, 401, 152.(6) Feringa, B. L. J. Org. Chem. 2007, 72, 6635.(7) Pollard, M. M.; Meetsma, A.; Feringa, B. L. Org. Biomol. Chem.

2008, 6, 507.(8) Koumura, N.; Geertsema, E. M.; van Gelder, M. B.; Meetsma, A.;

Feringa, B. L. J. Am. Chem. Soc. 2002, 124, 5037.(9) van Delden, R. A.; ter Wiel, M. K. J.; Pollard, M. M.; Vicario, J.;

Koumura, N.; Feringa, B. L. Nature 2005, 437, 1337.(10) Pollard, M. M.; ter Wiel, M. K. J.; van Delden, R.; Vicario, J.;

Koumura, N.; van den Brom, C.; Meetsma, A.; Feringa, B. Chem.sEur. J.2008, 14, 11610.

(11) Pollard, M. M.; Lubomska, M.; Rudolf, P.; Feringa, B. L. Angew.Chem., Int. Ed. 2007, 46, 1278.

(12) Yamaki, M.; Nakayama, S.; Hoki, K.; Kono, H.; Fujimura, Y. Phys.Chem. Chem. Phys. 2009, 11, 1662.

Figure 6. CIs of CPTHF. CASSCF(4,3)/6-31G(d) energies in kcal/mol.

Ground State Potential Energy Surface of CPTHF J. Phys. Chem. A, Vol. 114, No. 34, 2010 9347

(13) Fujimura, Y.; Gonzalez, L.; Kroner, D.; Manz, J.; Mehdaoui, I.;Schmidt, B. Chem. Phys. Lett. 2004, 386, 248.

(14) Hoki, K.; Sato, M.; Yamaki, M.; Sahnoun, R.; Gonzalez, L.; Koseki,S.; Fujimura, Y. J. Phys. Chem. B 2004, 108, 4916.

(15) Alfalah, S.; Kinzel, D.; Gonzalez-Vazquez, J.; Gonzalez, L. Chem.Phys. 2010, 369, 138.

(16) Perez-Hernandez, G.; Pelzer, A.; Gonzalez, L.; Seideman, T. NewJ. Phys. 2010, 12, 075007.

(17) Roos, B. O. AdV. Chem. Phys. 1987, 69, 399.(18) Finley, J.; Malmqvist, P.-Å.; Roos, B. O.; Serrano-Andres, L. Chem.

Phys. Lett. 1998, 288, 299.(19) Roos, B. O.; Andersson, K. Chem. Phys. Lett. 1995, 245, 215.(20) Fukui, K. Acc. Chem. Res. 1981, 14, 363.(21) Dykstra, C. E.; Frenking, G.; Kim, K. S.; Scuseria, G. Theory and

Applications of Computational Chemistry: The First 40 Years, 1st ed.;Elsevier: Amsterdam, 2005.

(22) Andersson, K. et al. MOLCAS, Release 7.0; Department ofTheoretical Chemistry, Lund University.

(23) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.;Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.;Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.;Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.;Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li,

X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.;Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.;Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.;Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich,S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.;Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.;Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz,P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.;Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson,B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03,ReVision D.01; Gaussian, Inc.: Wallingford, CT, 2004.

(24) Merer, A. J.; Mulliken, R. S. Chem. ReV. 1969, 69, 639.(25) Perez-Hernandez, G.; Gonzalez, L.; Serrano-Andres, L. Chem. Phys.

Chem. 2008, 9, 2544.(26) Schreiber, M.; Barbatti, M.; Zilberg, S.; Lischka, H.; Gonzalez, L.

J. Phys. Chem. A 2006, 111, 238.(27) Gedanken, M. D.; Huang, J.; Rachon, J.; Walborsky, H. M. J. Am.

Chem. Soc. 1988, 110, 4593.(28) Levine, B. G.; Martınez, T. J. Annu. ReV. Phys. Chem. 2007, 58,

613.(29) Ben-Nun, M.; Martınez, T. J. Chem. Phys. 2000, 259, 237.

JP104898T

9348 J. Phys. Chem. A, Vol. 114, No. 34, 2010 Assmann et al.