ARTICLE pubs.acs.org/JPCA
Photochemistry of (η6-Anisole)Cr(CO)3 and (η6-Thioanisole)Cr(CO)3: Evidence for a Photoinduced Haptotropic Shift of the Thioanisole Ligand, a Picosecond Time-Resolved Infrared Spectroscopy and Density Functional Theory Investigation Ian P. Clark,‡ Michael W. George,§ Gregory M. Greetham,‡ Emma C. Harvey,† Conor Long,†,* Jennifer C. Manton,† Hazel McArdle,† and Mary T. Pryce† †
School of Chemical Sciences, Dublin City University, Dublin 9, Ireland Central Laser Facility, Science & Technology Facilities Council, Research Complex at Harwell, Rutherford Appleton Laboratory, Didcot, OX11 0QX, United Kingdom § School of Chemistry, University of Nottingham, Nottingham, NG7 2RD, United Kingdom ‡
bS Supporting Information ABSTRACT: The photochemistry of (η6-anisole)Cr(CO)3 and (η6-thioanisole)Cr(CO)3 was investigated by picosecond time-resolved infrared spectroscopy in n-heptane solution at 298 K. Two independent excited states are populated following 400 nm excitation of each of these complexes. An excited state with some metal-to-CO charge-transfer character is responsible for the CO-loss process, which is slow compared to CO-loss from Cr(CO)6. Observed first order rate constants of 1.8 1010 s 1 and 2.5 1010 s 1 were obtained for the anisole and thioanisole complexes, respectively. The second excited state has metal-to-arene charge transfer character and results in a haptotropic shift of the thioanisole ligand. DFT calculations characterized the excited states involved and the nature of the haptotropic shift intermediate observed for the thioanisole species.
’ INTRODUCTION We have been investigating the photochemical properties of (η6-Arene)Cr(CO)3 complexes in low temperature matrixes and alkane solution at ambient temperatures.1 4 Measurements using picosecond transient infrared spectroscopy (ps-TRIR) have greatly enhanced our understanding of the photophysical processes following optical excitation of these systems. For instance, the formation of two independent excited state-species of either Metal-to-Arene Charge-Transfer (MACT) or Metalto-CO Charge-Transfer (MCCT) character has been proposed to explain both the wavelength and arene dependence of the COloss quantum yield.5 Population of the MCCT excited state results, both in the slow expulsion of one CO ligand (over approximately 150 ps), and the relaxation to the ground state. The latter process reduces the overall quantum yield of CO-loss (0.7 in the case of arene = benzene λexc. = 400 nm). Optical excitation of (η6-naphthalene)Cr(CO)3, populates a MACT state which forms a reduced hapticity intermediate species before reforming the parent η6-complex over about 50 ps.6 Matrix isolation experiments have shown that the photochemistries of (η6-arene)Cr(CO)3 complexes are temperature dependent.3 The CO-loss process involves both optical excitation and overcoming a small thermal barrier in the excited state. The thermal barrier slows the CO-loss process enabling the dynamics to be r 2011 American Chemical Society
observed using ps-TRIR. Low temperature annealing of CH4 matrixes containing (η6-aniline)Cr(CO)3 provided evidence for a haptotropic shift processes which was promoted by phase changes in the isolating matrix material.7 This work demonstrated the availability of multiple coordination modes for the aniline ligand with respect to the Cr(CO)3 unit. It is clear from these studies that (η6-arene)Cr(CO)3 complexes exhibit a rich and complex photochemistry. Recently, the possibility of constructing molecular switches based on substituted (η6-arene)Cr(CO)3 or (η5-cyclopentadienyl)Mn(CO)3 complexes has been investigated by Burkey and coworkers.8 13 In these systems, the polyene ligand contains pendent donor sites, either a group 15 or 16 element, or alternatively an alkenyl substituent. Initial photoexpulsion of a CO ligand from the metal tricarbonyl unit is followed by formation of a chelate species with the pendent donor site coordinating to the metal. Photochromic systems can be constructed using substituents which possess two types of donor sites. Photoinduced linkage isomerization can produce photochromic behavior.8 However, with this approach, it is necessary to so design the Received: December 6, 2011 Revised: December 24, 2011 Published: December 28, 2011 962
dx.doi.org/10.1021/jp211726j | J. Phys. Chem. A 2012, 116, 962–969
The Journal of Physical Chemistry A
ARTICLE
Figure 1. (a) The IR spectra of (η6-anisole)Cr(CO)3 (blue) and (η6-thioanisole)Cr(CO)3 (red) obtained in n-heptane solution; the absorbance scale is normalized using the high-energy (symmetric) band at 1980 and 1981 cm 1 for the anisole and thioanisole complex, respectively, showing the difference in the relative intensities of the symmetric and asymmetric absorptions: (b) the UV vis spectra of (η6-anisole)Cr(CO)3 (0.4 mM, blue) and (η6-thioanisole)Cr(CO)3 (0.08 mM, red) obtained in cyclohexane (the arrow indicates the excitation wavelength used in ps-TRIR experiments).
substituent, as to prevent the competitive coordination of a solvent molecule to the coordinatively unsaturated intermediate produced following CO-loss. This represents a significant design problem. An alternative approach is to construct systems in which the coordination mode of a ligand can be altered in a controlled way without requiring ligand loss. The resulting species would then form the basis of a photochromic system, provided the spectroscopic properties of the various coordination isomers differ significantly. As mentioned above, we have obtained evidence for the formation of a reduced hapticity intermediate following population of the MACT excited state of (η6-naphthalene)Cr(CO)3.6 The quantum yield of CO loss in this system is just 0.02 compared to 0.70 for (η6-benzene)Cr(CO)3 following 400 nm excitation at room temperature. Populating the MACT state selectively precludes CO-loss which only occurs following population of the MCCT excited state. Consequently, our target is to design systems based on a (η6-arene)Cr(CO)3 unit in which the arene is capable of coordinating in at least two distinct modes that can be selected photochemically without the added complication of the photoinduced CO-loss process. In this work, we report the results of DFT calculations and psTRIR studies on two model systems (η6-anisole)Cr(CO)3 and (η6-thioanisole)Cr(CO)3. These results demonstrate how the coordination mode of the thioanisole ligand can be altered following optical excitation. The DFT calculations provide a model for the various coordination modes of thioanisole which is consistent with the experimental results.
Figure 3. The energy change calculated in moving the Cr(CO)3 unit along the E to C4 vector in (η6-thioanisole)Cr(CO)3 (E = S; red plot) and (η6-anisole)Cr(CO)3 (E = O; blue plot), indicating that both (η6-thioanisole)Cr(CO)3 and (η4-S-thioanisole)Cr(CO)3 exist in local minima on the potential energy hypersurface and that no equivalent η4-O coordination mode is evident on the (η-anisole)Cr(CO)3 surface.
’ RESULTS The IR spectra of (η 6 -anisole)Cr(CO)3 (blue) and 6 (η -thioanisole)Cr(CO)3 (red) are presented in Figure 1(a). A comparison of the spectra in Figure 1(a) shows a significant difference in the relative intensities of the symmetric and asymmetric bands (with respect to the plane of symmetry assuming Cs symmetry) for the anisole complex while the bands have similar intensities for the thioanisole. The low energy band in (η6-thioanisole)Cr(CO)3 is slightly broader than the equivalent band of (η6-anisole)Cr(CO)3. The UV vis spectra of the two complexes
are presented in Figure 1(b). Both complexes exhibit an asymmetric low energy absorption with λmax = 315 and 320 nm for the anisole and thioanisole complexes respectively (note the different concentrations used to obtain these spectra). All ps-TRIR experiments in this report used an excitation wavelength of 400 nm (black arrow in Figure 1(b)). The potential energy (PE) surfaces around the two arene ligands in this study (anisole and thioanisole) were determined using a relaxed potential energy scan method. These calculations involved defining the location of the Cr(CO)3 unit on a 0.5 0.3 Å grid relative to the plane of the aromatic ring and permitting all
Figure 2. The potential energy surface presented to the Cr(CO)3 unit by thioanisole ligand (indicated by the sigma-bond framework) showing the locations of the η6 and η4-S coordination sites with isoenergetic contours indicated in kJ mol 1 relative to the calculated energy of (η6-thioanisole)Cr(CO)3 set to zero.
963
dx.doi.org/10.1021/jp211726j |J. Phys. Chem. A 2012, 116, 962–969
The Journal of Physical Chemistry A
ARTICLE
Figure 4. An energy level (eV) diagram for (η6-thioanisole)Cr(CO)3 and (η4-S-thioanisole)Cr(CO)3 indicating the principal contributions from the thioanisole and Cr(CO)3 fragments, the green arrows indicate the HOMO LUMO gap for each coordination isomer, red energy levels have predominant Cr(CO)3 character while blue are predominantly thioanisole based.
structural parameters to refine including the Cr to arene meanplane distance. All calculations used the B3LYP functional14,15 and the Tzvp basis set.16 A similar approach was used in mapping the PE surface around the arene ligand (arene = naphthalene or aniline) in other (arene)Cr(CO)3 complexes.6,7 The energy of the optimized structure at each point on the grid constituted one point on the PE surface. This surface is symmetric with respect to the plane of symmetry orthogonal to the plane of the arene ligand and containing E, C1 and C4 (E = O or S). The methyl group can be neglected because of the free rotation about the E C1 bond. Therefore, only half of the surface was calculated and the remaining half generated by reflection in this plane of symmetry (Figure 2). Further details of these calculations are available in the Supporting Information. The anisole surface exhibited a minimum near the center of the aromatic ring as expected for an η6 coordination to the Cr(CO)3 fragment (Figure 3). In addition, a shallow higher energy minimum was also observed for the (k-O-anisole)Cr(CO)3 species (Figure 3). This species lies at an energy 60 kJ mol 1 above that of the (η6-anisole)Cr(CO)3 species. The surface obtained for the thioanisole complex is more complicated. These calculations suggest that potential energy minima exist where the thioanisole ligand coordinates to the Cr(CO)3 unit as both (η4-S-thioanisole)Cr(CO)3 and (η6-thioanisole)Cr(CO)3 complexes (Figure 3). A very shallow potential energy well also exists for (η4-S-thioanisole)Cr(CO)3. The transition state (TS) between (η4-S-thioanisole)Cr(CO)3 and (η6-thioanisole)Cr(CO)3 was located (Figure 2) and an intrinsic reaction path calculation (IRC) along the vibrational mode corresponding to the single imaginary frequency of TS confirmed that it connected (η4-Sthioanisole)Cr(CO)3 and (η6-thioanisole)Cr(CO)3 complexes. Detailed energy calculations on the TS provided the activation enthalpies of both the forward and reverse reactions. The (η4-Sthioanisole)Cr(CO)3 species lies 75 kJ mol 1 above (η6-thioanisole)Cr(CO)3 while the TS energy is higher again by 25 kJ mol 1 (Figure 3). Thus the thioanisole complex represents a better candidate for a bistable system than the anisole complex because the coordination mode of the ligand can be altered in a more controlled manner. Nature of Bonding in (η6-Thioanisole)Cr(CO)3 and (η4-SThioanisole)Cr(CO)3. Orbital energy level diagrams based on the Cr(CO)3 and thioanisole units were constructed for both the (η6-thioanisole)Cr(CO)3 and (η4-S-thioanisole)Cr(CO)3 complexes. The results of these calculations which again used the B3LYP functional and the Tzvp basis set are presented in Figure 4 (further details are available in the Supporting Information).
In this diagram, contributions greater than 4% are indicated by lines to the fragment orbitals. The most obvious difference between the orbital energies in these two species is the smaller HOMO LUMO (H L) energy gap in the case of the (η4-S-thioanisole)Cr(CO)3 complex. In both isomers the HOMO is substantially derived from the Cr(CO)3 fragment (98% in (η4-S-thioanisole)Cr(CO)3 and 83% in the (η6-thioanisole)Cr(CO)3 complex). The increased contribution from the thioanisole ligand in (η6-thioanisole)Cr(CO)3 has the effect of stabilizing the HOMO (H) ( 5.700 eV) compared to the H of (η4-S-thioanisole)Cr(CO)3 ( 5.383 eV). The LUMO (L) orbital in (η6-thioanisole)Cr(CO)3 ( 1.330 eV) is mainly derived from the thioanisole ligand (82%). The L in (η4-S-thioanisole)Cr(CO)3 contains a large contribution from the L+1 on the Cr(CO)3 fragment (42.2%) which stabilizes this orbital ( 2.366 eV). A combination of these two factors explains the smaller HOMO LUMO gap in (η4-S-thioanisole)Cr(CO)3 compared to (η6-thioanisole)Cr(CO)3. Time-Dependent DFT calculations were undertaken to provide estimates of the nature of the optically accessible excited states for these systems. Again fuller details of these calculations are available in the Supporting Information. For the (η6-thioanisole)Cr(CO)3 complex, the lowest energy excited state is a MACT state and this weak transition corresponds to an absorption at 391 nm. Further stronger absorptions are calculated to occur at 376, 372, 367, and 357 nm and while these transitions can be classified as substantially MACT in character, they contain some MCCT character. Absorptions between 354 and 291 nm are predominantly MCCT in nature. In general terms, the excited state produced following optical excitation progresses from MACT character, acquiring greater MCCT character as the excitation energy increases. TDDFT calculations on (η4-S-thioanisole)Cr(CO)3 present quite a different picture. The smaller HOMO LUMO energy gap is reflected by a smaller excitation energy to the lowest energy excited state which occurs at 802 nm. This excited state has substantial MCCT character because of the contribution of the Cr(CO)3 fragment to the LUMO of (η4-S-thioanisole)Cr(CO)3 as mentioned above. Time Resolved Infrared Studies on (η6-Anisole)Cr(CO)3 in n-heptane. The spectroscopic changes observed following 50 fs pulse photolysis of (η6-anisole)Cr(CO)3 at 400 nm are presented in Figure 5. The two bands of the (η6-anisole)Cr(CO)3 parent complex at 1980 and 1907 cm 1 are depleted within the excitation pulse. Over the subsequent 350 ps, these two bands recover leaving a residual depletion of approximately 46%, 964
dx.doi.org/10.1021/jp211726j |J. Phys. Chem. A 2012, 116, 962–969
The Journal of Physical Chemistry A
ARTICLE
Figure 5. (a) ps-TRIR spectra obtained following 400 nm excitation of (η6-anisole)Cr(CO)3 in n-heptane solution at 292 K, spectra were obtained at 2.12, 2.5, 3, 3.5, 4.36, 6.74, 9.26, 11.96, 14.84, 17.96, 21.34, 25.02, 29.08, 33.6, 38.7, 44.57, 51.4, 59.66, 70.1, 84.26, 106.38, 160, 250, 350, 500, 750, and 1000 ps after the 50 fs excitation pulse: (b) selected offset spectra obtained at the times indicated.
representing an upper limit of 0.46 for the quantum yield of COloss which is the only photochemical process observed for this system. Upon initial excitation, three new features (1880, 1963, and 1988 cm 1) are produced. Two of these features (1880 and 1963 cm 1) become narrow and increase in absorbance within 4 ps of the excitation (this is indicated in Figure 5 by curved arrows) which indicates vibrational relaxation. The feature at 1988 cm 1 lies close to the parent band at 1980 cm 1 and consequently it is difficult to accurately analyze the time dependence of this band. However, the band does not appear to narrow significantly, suggesting that this species is vibrationally relaxed within a few ps of the excitation pulse. This behavior has been observed previously for the (η6-methylbenzoate)Cr(CO)3 system where two independent excited states were also identified and characterized as MACT and MCCT. The MCCT excited state of (η6-methylbenzoate)Cr(CO)3 exhibited νCO bands (1981 and 1912 cm 1) to the low energy side of each parent absorption while the MACT excited state exhibited two bands one to the high energy side of the higher energy parent absorption and the other to the low energy side of the lower energy parent absorption. Consequently, the transient bands at 1963 and 1880 cm 1 observed following photolysis of (η6-anisole)Cr(CO)3, are assigned to a MCCT excited state. The feature at 1988 cm 1 and a second feature, possibly contributing to the shoulder on the high energy side of the 1880 cm 1 band, i.e., to the low energy side of the lower energy parent band, are assigned to the MACT excited state. The band at 1988 cm 1 decays following first order kinetics with an observed first order rate constant (kobs) of 1.2 ((0.6) 1011 s 1 (all kinetic analyses were conducted at 292 K). The band at 1963 cm 1, assigned to the MCCT excited state, initially increases in absorbance over about 4 ps through vibrational relaxation, but then decays with a
Scheme 1. Reaction Scheme Describing the Photophysics and Photochemistry of (η6-Anisole)Cr(CO)3 Following 400 nm Excitation in n-Heptane at 292 K
kobs = 1.9 ((0.2) 1010 s 1. This rate constant is over an order of magnitude smaller than the kobs for the decay of the 1988 cm 1 band. Concomitant with the decay of the MCCT state, bands at 1921 and 1861 cm 1 are produced (kobs = 1.8 ((0.3) 1010 s 1) which are assigned to the CO-loss species. This assignment is made by comparison with the νCO band positions of similar 16 electron species, for instance (η6-benzene)Cr(CO)2(n-heptane) 1930 and 1870 cm 1 and (η6-methyl benzoate)Cr(CO)2(n-heptane) 1941 and 1887 cm 1.2,6 The overall reaction scheme depicting the photophysical and photochemical behavior of (η6-anisole)Cr(CO)3 is presented as Scheme 1. Kinetic data 965
dx.doi.org/10.1021/jp211726j |J. Phys. Chem. A 2012, 116, 962–969
The Journal of Physical Chemistry A
ARTICLE
Table 1. Observed First Order Rate Constants (kobs) at 292 K for the Decay of MACT and MCCT Excited States, The Parent Recovery, And Formation of the CO-Loss Species for (η6-Anisole)Cr(CO)3 and (η6-Thioanisole)Cr(CO)3 Systems in n-Heptane Solution (η6-anisole)Cr(CO)3
complex process
(η6-thioanisole)Cr(CO)3
MACT decay MCCT decay CO-loss formation parent recovery MACT decay MCCT decay CO-loss formation parent recovery 1
monitored (cm )
1988
1962
1920
1979
1994
1967
1926
1981
kobs( 1010 s 1)
12.0 ((0.6)
1.9 ((0.2)
1.8 ((0.2)
6.2 ((0.5)
14.2 ((0.6)
2.4 ((0.1)
2.5 ((0.6)
6.6 ((0.5)
Figure 6. (a) The ps-TRIR spectra obtained following pulsed photolysis (50 fs) at 400 nm of (η6-thioanisole)Cr(CO)3 in n-heptane solution at 292 K, spectra were recorded at 2.12, 2.5, 3, 3.5, 4.36, 6.74, 9.26, 11.96, 14.84, 17.96, 21.34, 25.02, 29.08, 33.6, 38.7, 44.54, 51.4, 59.66, 70.1, 84.26, 106.38, 160, 250, 350, 500, 750, and 1000 ps after the excitation pulse, the curved arrows indicate initial vibrational relaxation followed by decay of the feature (b) Selected ps-TRIR spectra obtained the bands of (η6-thioanisole)Cr(CO)3 exhibit negative Δ absorbance values and the ordinate scale is as indicated in the inset.
reform the parent (η6-thioanisole)Cr(CO)3 and the CO-loss species(η6-thioanisole)Cr(CO)2(n-heptane) (1927and1870cm 1). This behavior mimics that of the MCCT excited state of (η6-anisole)Cr(CO)3. Kinetic data for this process are presented in Table 1. The MACT excited state decays with a similar rate to the equivalent state in the (η6-anisole)Cr(CO)3 system albeit slightly faster (Table 1). The features appearing as shoulders on the high energy side of the (η6-thioanisole)Cr(CO)2(n-heptane) bands require a fuller explanation best supported by the stepped spectra presented in Figure 6(b). These features have been assigned to a (η4-S-thioanisole)Cr(CO)3 (ring-slip) species as predicted by the potential energy surface calculations outlined above. These spectra suggest that the feature at 1933 cm 1 is produced faster than the feature at 1927 cm 1, i.e., (η6-thioanisole)Cr(CO)2(n-heptane) . The 1933 cm 1 feature appears concomitantly with the decay of the MACT excited state (1994 cm 1). The spectra obtained up to 22 ps after excitation suggest that the ringslip species also has a band at about 1870 cm 1 overlapping the low energy band of (η6-thioanisole)Cr(CO)2(n-heptane) which has yet to be formed to any significant extent on this time scale.
for decay and formation of selected spectroscopic features are presented in Table 1. Time Resolved Infrared Studies on (η6-Thioanisole)Cr(CO)3 in n-Heptane. The spectroscopic changes observed following 50 fs pulsed photolysis at 400 nm of a solution of (η6-thioanisole)Cr(CO)3 in n-heptane at 292 K are presented in Figure 6. The general shape of these changes is similar to that obtained for the anisole complex (Figure 5) with the exception of pronounced shoulders on the photoproduct features at 1927 and 1870 cm 1. In addition the broad absorption close to the high energy absorption of the parent complex is more pronounced than the equivalent absorption in the anisole complex (Figure 5). The depletion of the parent absorptions recover to approximately 54% of the initial depletion after 1000 ps. These spectra provide evidence for the formation of two excited states, the MACT absorbing at approximately 1994 and 1889 cm 1 and the MCCT excited state absorbing at 1967 and 1889 cm 1. Qualitatively, the relative yield of the MACT excited state appears to be greater in (η6-thioanisole)Cr(CO)3 compared to (η6-anisole)Cr(CO)3. The MCCT bands increase in absorbance within 4 ps of the excitation pulse (vibrational relaxation) before decaying to 966
dx.doi.org/10.1021/jp211726j |J. Phys. Chem. A 2012, 116, 962–969
The Journal of Physical Chemistry A
ARTICLE
Scheme 2. Reaction Scheme Describing the Photophysics and Photochemistry of (η6-Thioanisole)Cr(CO)3 Following 400 nm Excitation in n-Heptane at 292 K
Figure 7. Idealized potential energy profiles for two extreme conditions (a) populating the metal-to-ligand charge-transfer (MLCT) excited state surface at a point further along the internal coordinate than the avoided crossing with the metal centered (MC) state which results in rapid CO-loss (b) populating the excited state at a point before the avoided crossing which results in slow CO-expulsion over a small thermal barrier.
The final band of the ring-slip species occurs at 1880 cm 1 and is only revealed after the MCCT excited state has decayed. Thus, the three features at approximately 1933, 1880, and 1870 cm 1 correspond to the same species, obviously containing the intact Cr(CO)3 fragment. DFT calculations provide an estimate of the νCO band positions for (η4-S-thioanisole)Cr(CO)3 at 1962, 1884, and 1872 cm 1 which, apart from the highest energy band, are close to the experimental features described above. [These values were corrected for zero point energy using the following formula (calculated value) 1.134 350 cm 1 based on a calibration curve obtained from calculations on (η6-thioanisole)Cr(CO)3 (see Supporting Information).] Thus, the ps-TRIR results provide experimental support for the availability of two coordination modes for the thioanisole ligand in (η-thioanisole)Cr(CO)3 complexes and that the (η4-S-thioanisole)Cr(CO)3 can be accessed photochemically from (η6-thioanisole)Cr(CO)3. The overall photophysical and photochemical processes following irradiation of (η6-thioanisole)Cr(CO)3 are presented in Scheme 2.
wavelength reduces.2 This is in contrast with the Mo system where the quantum yield increases with decreasing excitation wavelength.23,24 Our current description of the photophysics of CO-loss from (η6-arene)Cr(CO)3 complexes suggest that two independent excited states exist, classified as MCCT and MACT, but only the MCCT excited state is involved in CO-loss by way of an avoided crossing with a metal centered (MC) state not directly accessed in the excitation process. The metal centered excited state exists at higher energy and cannot be populated directly following excitation at 400 nm (see Supporting Information for Mullikan population analysis). The avoided crossing occurs at a point on the M-CO internal coordinate greater than that of the Franck Condon state (Figure 7). This generates a small thermal barrier to CO-loss which results in the slow release of CO (160 ps) compared to Cr(CO)6 (88 fs).21 Thus, the quantum efficiency at which the MCCT state is populated upon optical excitation controls the quantum yield of the CO-loss process. The MACT excited state is populated efficiently when the arene ligand contains a substituent which reduces the formal symmetry from C3v for arene = benzene to Cs or perhaps C1 for arene = substituted-benzene. Population of a second (MACT) excited state in the substituted systems reduces the quantum yield of CO-loss (ΦCO = 0.7 for arene = benzene but 0.3 for arene = methyl benzoate).6 This is because the excitation energy is partitioned between two excited states (MCCT and MACT) rather than just one (MCCT). One important question remains to be answered, will population of the MACT excited state induce other photochemical processes such as a haptotropic shift or even arene loss? A photoinduced haptotropic shift has been reported following visible irradiation of (η6-pyridine)Cr(CO)3.25 Matrix isolation experiments in N2 matrixes confirmed the formation of (k-N-pyridine)Cr(CO)3(N2)2 as the ring-slip product. Photolysis of either (η6-anisole)Cr(CO)3 or (η6-thioanisole)Cr(CO)3 at 400 nm produces two excited state species (MCCT and MACT). The MACT excited state in both complexes exhibit two IR bands one of which is on the high energy side of the high energy parent band and the second is on the low energy side of the low-energy parent band. This suggests that population of this excited state results in a significant change to the structure of the Cr(CO)3 fragment. In the case of the anisole complex, this
’ DISCUSSION For many years, it has been proposed that M(CO)6 (M = Cr, Mo, or W) complexes represent prototypical systems to describe the photophysical processes involved in CO-loss.17 20 In these systems, photoinduced CO-loss occurs following direct population of an unbound excited state which converts efficiently the photon energy into nuclear kinetic energy expelling a CO ligand within 100 fs.21 Sophisticated calculations suggest that the unbound state results from an avoided crossing of a MCCT excited state and a metal centered state, explaining why CO-loss is both efficient and independent of the excitation wavelength.22 Thus, fast and efficient photoinduced CO-loss from homoleptic metal carbonyl complexes has become the accepted and expected behavior following excitation. However, our work on “half sandwich” complexes in particular (η6-arene)M(CO)3 (M = Cr or Mo) systems has revealed a more complex behavior. Despite early suggestions that the quantum yield of CO-loss from (η6-benzene)Cr(CO)3 was independent of the excitation wavelength,5 more recent measurements suggest that there is a slight wavelength dependence, with the quantum yield reducing as the excitation 967
dx.doi.org/10.1021/jp211726j |J. Phys. Chem. A 2012, 116, 962–969
The Journal of Physical Chemistry A excited state simply relaxes to the parent complex. However, in the case of the thioanisole complex a long-lived intermediate species is formed. By mapping the potential energy surface over the arene ligands, an additional reduced hapticity coordination mode for the thioanisole complex was revealed. It is interesting to note that a (k-C-benzene)Cr(CO)3 species was located on the singlet potential energy surface by Weitz and co-workers.26 All calculations in this work were conducted on the singlet hypersurface and there was no evidence for significant spin contamination. The calculated νCO bands of the reduced hapticity intermediate located in the potential energy mapping (i.e., (η4-S-thioanisole)Cr(CO)3) are similar to the νCO bands observed in the ps-TRIR experiments. As a result the long-lived intermediate species is assigned to the (η4-S-thioanisole)Cr(CO)3 species. As outlined in Table 1 single exponential fitting methods were used to determine the rate constants for the decay of the MACT and MCCT excited states and also the rate constants for the formation of the CO-loss and recovery of the parent species. The decay of the MACT excited state is almost an order of magnitude faster than the decay of the MCCT excited state for both complexes. However, the observed rate constant for the parent recovery (obtained by single exponential curve fitting) has a value between those for the decay of the two excited states. This confirms that both excited states relax to the parent complex. For the thioanisole system, the formation of the reduced hapticity intermediate from the MACT excited state must compete with the process of relaxation to the ground state (Scheme 2). The similarity of the rate constants for the relaxation of the MACT excited state in both systems would suggest that the relaxation to the ground state is the rate determining step and that the reduced hapticity species simply intercepts the relation process. Unfortunately, because of substantial overlap of the νCO bands of the various species, it is not possible to estimate the relative yields of the intermediates produced. The results of TDDFT calculations on the (η4-S-thioanisole)Cr(CO)3 species confirm that this species will absorb toward the red end of the visible spectrum. Attempts to detect this absorption using nanosecond UV vis flash photolysis were unsuccessful. This was because the CO-loss species ((η6-thioanisole)Cr(CO)2(cyclohexane)) also exhibits a broad absorption across the visible spectrum and its presence would mask any visible absorptions of (η4-S-thioanisole)Cr(CO)3. In addition, the excitation wavelength used in the UV vis flash photolysis experiments (355 nm) would populate the MCCT excited state more efficiently than the 400 nm excitation pulse used in the ps-TRIR experiments. The lowest energy absorption of (η4-S-thioanisole)Cr(CO)3 has significant MCCT character (see Supporting Information). However, the population of this excited state would not result in CO-loss because the photon energy would not be sufficient to break a Cr-CO bond. This work demonstrates that both the MCCT and MACT excited states of (η6-arene)Cr(CO)3 complexes are photochemically active. The MCCT excited state is responsible for the wellknown CO-loss process, while the MACT is responsible for the formation of reduced hapticity intermediates. However, to observe both types of photochemical behavior the arene must possess a substituent or substituents which reduce the overall symmetry of the molecule. This is required to permit the direct population of the MACT excited state following optical excitation as the transition to the MACT state is symmetry forbidden for the symmetric (η6-benzene)Cr(CO)3 complex. In addition,
ARTICLE
the substituent(s) must be capable of trapping the reduced hapticity intermediate formed during the relaxation of the MACT excited state. This work suggests that the models developed to explain the photoinduced behavior of symmetric homoleptic metal carbonyl complexes are not necessarily applicable to heteroleptic systems. Indeed, within the subclass of “half-sandwich” complexes, symmetry considerations are also important in explaining photochemical behavior.
’ EXPERIMENTAL SECTION TRIR Apparatus. The ULTRA laser system used in these TRIR studies is a cryogenically cooled Ti:Sapphire laser amplifier (Thales Laser ALPHA 10000) providing 50 fs duration pulses at 800 nm with a repetition rate of 10 kHz.27 The 0.8 mJ, 800 nm output is divided to generate 400 nm pump and mid-IR probe beams through second harmonic generation and difference frequency generation of signal and idler produced in an optical parametric amplifier (Light Conversion TOPAS), respectively. The pump beam was attenuated to 0.5 μJ at the sample and focused to overlap with the probe beam in the sample. The pump and probe spot sizes in the sample were 100 and 50 μm, respectively. After the sample, the IR probe beam was dispersed onto two linear 128 element MCT detector arrays acquiring spectra at 10 kHz. The signals from these two detector arrays were then averaged. By modulating the pump laser output at 5 kHz, difference spectra could be generated.28 A portion of the IR probe taken before the sample was dispersed on to a second 64 element MCT detector array, which with interpolation was used to generate reference spectra used to remove energy and spectral instabilities of the laser. All complexes in this study were obtained from Ellanova Laboratories29 and were used as supplied. Solutions were prepared in spectroscopic grade n-heptane and the concentration determined by UV vis. spectroscopy. A typical concentration of (η6-arene)Cr(CO)3 in these experiments was 6 10 3 M. The solutions were then passed through a flow-through cell fitted with CaF2 windows and a 200 μm spacer resulting in a typical absorbance of 0.1 A.U. at 400 nm. Signals were averaged for 8 s representing 40,000 measurements (pump on and pump off) at each delay time. The sample cell position was subjected to a twodimensional raster motion during the experiments to avoid multiple excitation of the same sample volume resulting in sample degradation. Computational Details. DFT methods were used in all calculations presented in this work. These employed a three parameter hybrid functional (B3),14 and the Lee Yang Parr correlation functional (LYP)15 i.e. the B3LYP functional. The Tzvp basis set was used for all calculations including the TDDFT calculations. Structures were optimized to tight convergence criteria (OPT=Tight) without constraints using the Gaussian 09 package,30 at Ireland’s High-Performance Computing Facility (http:// www.ichec.ie/). Orbital composition analyses were derived from accurate self-consistent field calculations using the AOMix software package.31,32
’ ASSOCIATED CONTENT
bS
Supporting Information. Details of the PE surfaces of both anisole and thioanisole complexes along with details of TDDFT calculations for (η6-anisole)Cr(CO)3, (η6-thioanisole)Cr(CO)3,
968
dx.doi.org/10.1021/jp211726j |J. Phys. Chem. A 2012, 116, 962–969
The Journal of Physical Chemistry A
ARTICLE
and (η4-S-thioanisole)Cr(CO)3, fragment contributions to the valence orbitals of (η6-thioanisole)Cr(CO)3, and (η4-S-thioanisole)Cr(CO)3, and IR calibration details. This material is available free of charge via the Internet at http://pubs.acs.org.
(20) Kosma, K.; Trushin, S. A.; Fuss, W.; Schmid, W. E.; Schneider, B. M. R. Phys. Chem. Chem. Phys. 2010, 12 (40), 13197–13214. (21) Paterson, M. J.; Hunt, P. A.; Robb, M. A.; Takahashi, O. J. Phys. Chem. A 2002, 106 (44), 10494–10504. (22) Villaume, S.; Strich, A.; Daniel, C.; Perera, S. A.; Bartlett, R. J. Phys. Chem. Chem. Phys. 2007, 9 (46), 6115–6122. (23) Breheny, C. J.; Kelly, J. M.; Long, C.; O’Keefe, S.; Pryce, M. T.; Russell, G.; Walsh, M. M. Organometallics 1998, 17 (17), 3690–3695. (24) George, M. W.; Long, C.; Pryce, M. T.; Sun, X.-Z.; Vuong, K. Q. Organometallics 2012, 31 (1), 268 272. (25) Breheny, C. J.; Draper, S. M.; Grevels, F. W.; Klotzbucher, W. E.; Long, C.; Pryce, M. T.; Russell, G. Organometallics 1996, 15 (17), 3679–3687. (26) Cohen, R.; Weitz, E.; Martin, J. M. L.; Ratner, M. A. Organometallics 2004, 23 (10), 2315–2325. (27) Greetham, G. M.; Cao, Q.; Clark, I. P.; Codd, P. S.; George, M. W.; Farrow, R. C.; Matousek, T.; Parker, A. W.; Pollard, M.; Robinson, D. A.; Xin, Z. J.; Towrie, M. Appl. Spectrosc. 2010, 64 (12), 1311–1319. (28) Towrie, M.; Grills, D. C.; Dyer, J.; Weinstein, J. A.; Matousek, P.; Barton, R.; Bailey, P. D.; Subramaniam, N.; Kwok, W. M.; Ma, C. S.; et al. Appl. Spectrosc. 2003, 57 (4), 367–380. (29) Ellanova Laboratories. http://www.ellanovalabs.com/. (30) Frisch, J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; et al. Gaussian 09, Gaussian, Inc.: Wallingford CT, 2009. (31) Gorelsky, S. I. AOMix program, http://www.sg-chem.net/, version 6.4; University of Ottawa: Ottawa, 2010. (32) Gorelsky, S. I.; Ghosh, S.; Solomon, E. I. J. Am. Chem. Soc. 2006, 128, 278–290.
’ AUTHOR INFORMATION Corresponding Author
*Tel.: +353 1 7008001; fax: +353 1 7005503; e-mail: conor.long@ dcu.ie.
’ ACKNOWLEDGMENT The authors gratefully acknowledge helpful discussions with Anthony W. Parker of CLF. M.T.P., E.C.H., and J.C.M. acknowledge funding from Science Foundation Ireland (08/RF/PHY082) and the Environmental Protection Agency (2008-ET-MS-3-52). The authors also thank the Central Laser Facility for granting access to the ULTRA system under EU Access Grant No.92004. M.W.G. gratefully acknowledges receipt of a Royal Society Wolfson Merit Award. ’ REFERENCES (1) Clark, I. P.; George, M. W.; Greetham, G. M.; Harvey, E. C.; Long, C.; Manton, J. C.; Pryce, M. T. J. Phys. Chem. 2010, 114, 11425– 11431. (2) Alamiry, M. A. H.; Boyle, N. M.; Brookes, C. M.; George, M. W.; Long, C.; Portius, P.; Pryce, M. T.; Ronayne, K. L.; Sun, X. Z.; Towrie, M.; Vuong, K. Q. Organometallics 2009, 28 (5), 1461–1468. (3) Alamiry, M. A. H.; Brennan, P.; Long, C.; Pryce, M. T. J. Organomet. Chem. 2008, 693 (17), 2907–2914. (4) Creaven, B. S.; George, M. W.; Ginzburg, A. G.; Hughes, C.; Kelly, J. M.; Long, C.; McGrath, I. M.; Pryce, M. T. Organometallics 1993, 12 (8), 3127–3131. (5) Wrighton, M. S.; Haverty, J. L. Z. Naturforsch. 1975, 30(b) (3 4), 254–258. (6) Clark, I. P.; George, M. W.; Greetham, G. M.; Harvey, E. C.; Long, C.; Manton, J. C.; Pryce, M. T. J. Phys. Chem. A 2011, 115 (14), 2985–2993. (7) Alamiry, M. H.; Long, C.; Fidgeon, P. P.; Pryce, M. T. J. Organomet. Chem. 2010, 695 (12 13), 1634–1640. (8) To, T. T.; Heilweil, E. J.; Duke, C. B.; Ruddick, K. R.; Webster, C. E.; Burkey, T. J. J. Phys. Chem. A 2009, 113 (12), 2666–2676. (9) To, T. T.; Duke, C. B.; Junker, C. S.; O’Brien, C. M.; Ross, C. R.; Barnes, C. E.; Webster, C. E.; Burkey, T. J. Organometallics 2008, 27 (2), 289–296. (10) To, T. T.; Heilweil, E. J.; Duke, C. B.; Burkey, T. J. J. Phys. Chem. A 2007, 111 (30), 6933–6937. (11) To, T. T.; Heilweil, E. J.; Burkey, T. J. J. Phys. Chem. A 2006, 110 (37), 10669–10673. (12) Yeston, J. S.; To, T. T.; Burkey, , T. J.; Heilweil, E. J. J. Phys. Chem. B 2004, 108 (15), 4582–4585. (13) To, T. T.; Barnes, C. E.; Burkey, T. J. Organometallics 2004, 23 (11), 2708–2714. (14) Becke, A. D. J. Chem. Phys. 1993, 98 (7), 5648–5652. (15) Lee, C. T.; Yang, W. T.; Parr, R. G. Phys. Rev. B 1988, 37 (2), 785–789. (16) Schafer, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys. 1994, 100 (8), 5829–5835. (17) Trushin, S. A.; Fuss, W.; Schmid, W. E.; Kompa, K. L. J. Phys. Chem. A 1998, 102 (23), 4129–4137. (18) Trushin, S. A.; Fuss, W.; Schmid, W. E. Chem. Phys. 2000, 259 (2 3), 313–330. (19) Trushin, S. A.; Kosma, K.; Fuss, W.; Schmid, W. E. Chem. Phys. 2008, 347 (1 3), 309–323. 969
dx.doi.org/10.1021/jp211726j |J. Phys. Chem. A 2012, 116, 962–969
