Article pubs.acs.org/JPCB
Solvent Effects on Nonradiative Relaxation Dynamics of Low-Energy Ligand-Field Excited States: A CuCl42− Complex Andrey S. Mereshchenko,*,† Olesya S. Myasnikova,† Maxim S. Panov,† Vladimir A. Kochemirovsky,† Mikhail Yu. Skripkin,† Darya S. Budkina,‡,§ and Alexander N. Tarnovsky*,‡,§ †
Saint-Petersburg State University, 7/9 Universitetskaya nab., St. Petersburg 199034, Russia Department of Chemistry and ‡The Center for Photochemical Sciences, Bowling Green State University, Bowling Green, Ohio 43403, United States
§
ABSTRACT: Nonradiative relaxation dynamics of CuCl42− complexes photoexcited into the highest-energy ligand-field electronic state (2A1) is studied in acetonitrile, dichloromethane, and chloroform solvents, as well as in acetonitrile−water and in acetonitrile−deuterated water mixtures. Due to ultrafast internal conversion, this excited state directly converts to the electronic ground state in dichloromethane and chloroform. The nonradiative relaxation constant is similar in anhydrous acetonitrile. Addition of water to acetonitrile solutions efficiently quenches the excited ligand-field 2A1 state. The quenching is proposed to be due to the diffusion-controlled formation of an electronically excited pentacoordinated [CuCl4H2O]2− encounter complex or a short-lived exciplex of similar structure, in which the electronic excitation energy transfers into the O−H stretch of the coordinated H2O molecule. This is followed by the dissociation of the pentacoordinated species, resulting in the reformation of the ground-state CuCl42− and free H2O molecules.
■
INTRODUCTION Copper(II) plays a significant role in the metabolism of living organisms, for example, a blue copper protein plastocyanin is an electron donor species in Photosystem I during photosynthesis in plants and bacteria (ref 1 and the references therein). While a better understanding of how copper(II) complexes react photochemically in solution is of interest, these complexes, especially simple copper(II) chlorocomplexes, are known as model compounds to understand the photochemistry of more complex copper complexes, and they can potentially be useful for better understanding of copper protein redox reactions. Copper(II) complexes possess two types of electronic transitions: ligand-to-metal charge transfer (LMCT) transitions from ligand-localized orbitals to d-orbitals of the copper(II) ion and ligand-field (LF) transitions between d-orbitals.1−3 Excitation of copper(II) complexes into LF states usually results in fast internal conversion of the excited species into the ground state.4−9 As a result, the LF excited states are usually nonreactive, “chemically inert” states.4 In contrast, LMCT states of copper(II) complexes have very efficient and fast photochemical reaction path and, therefore, undergo extremely fast sub-100 fs relaxation into lower-lying LF states and photochemical product formation.1,4−13 Typically, photochemical transformations upon LMCT excitation involves ligand elimination, which sometimes is accompanied by the reduction of copper(II) to copper(I) and the oxidation of the ligand.1,10,11,13−17 The solvent environment plays an important role in the photochemistry of copper(II) complexes. Solvent structure, © XXXX American Chemical Society
polarity, and donor−acceptor properties all can significantly affect the photoreaction mechanism. Thus, dissociation quantum yields of copper(II) monochlorocomplexes in methanol are more than 10 times smaller than those in acetonitrile due to a more pronounced cage effect in methanol.11 For the CuCl42− complex in particular, the photoproduct recombination mechanisms are different in coordinating solvents (such as acetonitrile) and noncoordinating solvents (like chloroform).18 Moreover, electron-donor solvents are able to form electronically excited complexes (exciplexes) with photoexcited copper(II) complexes having organic ligands, usually resulting in quenching of intraligand excited states.19−22 Generally, however, the solvent effects on the relaxation mechanisms and dynamics of LMCT and LF excited states are poorly studied. Therefore, in this work, we have investigated the solvent effects on the nonradiative relaxation dynamics of LF excited states in a model copper(II) chlorocomplex, CuCl42−. Specifically, relaxation dynamics of CuCl42− photoexcited directly into the Laporte-forbidden highest-energy 2A1 LF excited state are studied in different solvents, such as acetonitrile, dichloromethane, and chloroform as well as in acetonitrile−water solvent mixtures. Received: March 2, 2017 Revised: April 3, 2017 Published: April 6, 2017 A
DOI: 10.1021/acs.jpcb.7b02015 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
Figure 1. Steady-state absorption spectra of the CuCl42− complex in acetonitrile (black line), acetonitrile-2.2% vol. water (blue line), dichloromethane (red line), and chloroform (green line). Intense absorption in the UV−vis range is due to LMCT transitions; weak absorption in the vis-near-IR range is due to LF transitions. For convenience of the reader, the weak absorption between 600 and 1600 nm was rescaled. Thus, the molecular decadic extinction coefficient (ε) values between 240 and 600 nm are plotted on the left axis, whereas the ε values between 600 and 1600 nm are plotted on the right axis. In the inset, the energy diagram of the LF excited states is presented. The arrows in the inset correspond to the LF excitations corresponding to the broad band in the near-IR part of the spectrum. For clarity, the lowest-energy 2B2 → 2E transition and the corresponding arrow are not shown in the spectrum.
■
copper(II) is present in the form of CuCl42− complexes in all solvent environments studied in this work. The CuCl42− complex has D2d symmetry in acetonitrile.24,25 In all other solutions studied in this work, the geometry and the electronic structure of CuCl42− complexes is thought to be similar because of the great similarity of their steady-state absorption spectra. In the ligand field of D2d symmetry, five dorbitals of a copper(II) ion split into four molecular orbitals with the symmetries a1, b1, e, and b2. For a Cu2+ ion in d9 configuration, three low-lying orbitals (a1, b1, and e) are fully occupied, whereas the highest-energy b2 orbital is half-occupied. Therefore, the electronic transitions from a1, b1, and e to b2 orbitals are possible. Such electronic transitions are called d-d or ligand-field transitions. They are responsible for the weak absorption band in the near-IR region of the steady-state absorption spectrum, Figure 1. As a result of these electronic transitions, three possible LF excited states can be formed: 2E, 2 B1, and 2A1 with the energy of 4500, 6000, and 8300 cm−1, respectively, above the 2B2 ground state (the transition energies are obtained from Gaussian deconvolution of the steady-state absorption spectrum26). The above-mentioned LF transitions are Laporte-forbidden. However, vibronic coupling relaxes the Laporte prohibition rule, so the LF transitions become observable in the absorption spectrum, albeit with small extinction coefficients. Another type of electronic excitation for the CuCl42− complex is ligand-to-metal charge transfer transitions. The LMCT transitions responsible for the absorption bands in the UV−vis region correspond to electron promotion from either fully occupied chloride-ion localized orbitals (nCl), or σCu−Cl orbitals, or πCu−Cl orbitals to the b2 orbital of the copper(II) ion. LMCT transitions are symmetry allowed, therefore, LMCT absorption bands have significantly larger extinction coefficients relative to symmetry-forbidden LF bands. In the transient absorption experiments carried out in this work, CuCl42− complexes were promoted into the highestenergy LF excited state (2A1)26 upon 1300 nm excitation. The 2 A1 LF excited state lies conveniently at significantly higher energies relative to the other two LF states and, therefore, can be predominantly excited. Ultrafast ΔA spectra for 0.5 mM Cu(ClO 4) 2 −100 mM NEt 4Cl solutions in acetonitrile, dichloromethane, chloroform, and the 2.2% vol. acetonitrile− water mixture are shown in Figure 2. Within the first 200 fs following excitation, the negative ΔA band at 405 nm due to ground-state bleaching, weak ΔA signals between 340 and 370 nm, and two intense ΔA bands with the maxima at 460 and 585
EXPERIMENTAL SECTION Copper(II) perchlorate hexahydrate (>98%), tetraethylammonium chloride (>98%), phosphorus pentoxide (>98%), deuterium oxide (99.9%), and acetonitrile (>99.8%) were purchased from Sigma-Aldrich. Dichloromethane (HPLC grade) and chloroform (HPLC grade) were purchased from EMD Millipore Chemicals. Anhydrous solvents were obtained by distilling acetonitrile, dichloromethane, and chloroform with phosphorus pentoxide. Transient absorption (ΔA) spectra were measured using the experimental setup based on a regeneratively amplified Ti:sapphire laser system (800 nm, 100 fs, 1 kHz) described previously. 23 The signal output of a TOPAS (LightConversion) optical parametric amplifier was used to generate 1300 nm excitation (“pump”) pulses with the energy of 40 μJ pulse−1. A white-light continuum in the 340−760 nm spectral range was used for probing. The relative polarization of the pump and probe beams was set at the magic angle (54.7°). The sample solutions were circulated through a 1 mm Spectrosil UV quartz flow cell. The probe beam was focused onto the sample to a 160-μm diameter spot and overlapped at an angle of 6° with the pump beam focused to a 460-μm diameter spot. A linear dependence of transient absorption signals on excitation energy was ensured. The UV−vis-near-IR absorption spectra were measured using a Varian Cary 50 UV−vis, a Perkin Elmer Lambda 750 UV−vis-NIR, and a Perkin Elmer Lambda 1050 spectrophotometer. All experiments were performed at 21 °C.
■
RESULTS AND DISCUSSION To prepare CuCl42− solutions, a large excess (100 mM) of tetraethylammonium chloride was added to copper(II) perchlorate (0.5 mM) solutions in acetonitrile, dichloromethane, chloroform, and acetonitrile−water mixtures. Based on the previously published values of overall stability constants of copper(II) chlorocomplexes,10 more than 99% of copper(II) exist as tetrachlorocomplexes in the acetonitrile solutions prepared in this way. The electronic absorption spectrum of CuCl42− complexes in acetonitrile consists of two highly intense UV bands, and a broad near-IR band of much lower intensity. Chloroform and dichloromethane are poorer electron donors in comparison with acetonitrile. Therefore, CuCl42− complexes are expected to be more stable in chlorinated alkane solutions. The UV−vis-near-IR steady-state spectra of 0.5 mM Cu(ClO4)2-100 mM NEt4Cl solutions in different solvents (Figure 1) are almost identical, confirming that the majority of B
DOI: 10.1021/acs.jpcb.7b02015 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
Figure 3. Transient absorption kinetic traces (symbols) at a 590 nm probe wavelength measured for 0.5 mM Cu(ClO4)2 − 100 mM NEt4Cl solutions in anhydrous acetonitrile (black squares), dichloromethane (red open circles), and chloroform (green triangles) following 1300 nm excitation. Solvent contribution was subtracted. The best fits (lines) of the data for delays times >0.8 ps are given by a single-exponential decay function.
Figure 2. Transient absorption spectra of 0.5 mM Cu(ClO4)2 − 100 mM NEt4Cl solutions in anhydrous acetonitrile (panel A), acetonitrile2.2% vol. water (panel B), dichloromethane (panel C), and chloroform (panel D) following 1300 nm excitation. Under identical excitation conditions, neat solvents contributed signals for delay times between 0.2 and 0.5 ps, which were subtracted from the shown ΔA data.
nm, are observed. From 0.2 to 0.8 ps, the 340−370 nm ΔA signals fade away, whereas the 460 and 585 nm ΔA bands do not change in intensity, but slightly blue-shift. Between 0.8 and 20 ps, the 405, 460, and 585 nm ΔA bands all decay without changing their spectral shape. The transient absorption spectra exhibit no noticeable ΔA amplitude at delay times longer than 20 ps. The analysis of transient absorption kinetic traces reveals that the decay of the ΔA signals can be well approximated (for delay times >0.8 ps) by the monoexponential decay law ΔA(t) = ΔA0·e−t/τ, Figures 3 and 4. The decay time constants obtained from the fits, τ, differ slightly across the solvent environments investigated and are equal to 5.15 ± 0.06, 5.46 ± 0.17, and 3.80 ± 0.12 ps in anhydrous acetonitrile, dichloromethane, and chloroform, respectively, Figure 3. In acetonitrile−water mixtures, the decay time constants are observed to decrease with increasing water concentration, though remaining firstorder. The τ values are equal to 4.88 ± 0.12, 4.33 ± 0.10, 3.83 ± 0.10, 3.06 ± 0.09, and 2.01 ± 0.07 ps for the mixtures containing 0.2, 0.7, 1.2, 2.2, and 5.2% vol. of water, respectively, Figure 4. Thus, the maximum amount of water added (5.2% vol.) shortens the excited state lifetime by a factor of 2.5. The decay rate constants, kd, defined as 1/τ, are found to depend linearly on the water concentration in the acetonitrile−aqueous mixtures, kd = k0 + kq·[H2O]. The y-intercept and slope of the Stern−Volmer plot yield k0 = (1.93 ± 0.02)·1011 s−1, where k0 is a decay rate constant in the absence of water (anhydrous acetonitrile), and kq = (1.06 ± 0.02)·1011 M−1 s−1, where kq is a quenching rate constant, Figure 5. In acetonitrile−deuterated water mixture (5.0 vol% of D2O), the decay time constant was equal to 3.12 ± 0.05 ps, which is 1.5 times slower than that in CH3CN-H2O mixed solvent with the same molar concentration of water (2.89 M), Figure 4.
Figure 4. Transient absorption kinetic traces (symbols) measured for 0.5 mM Cu(ClO4)2 − 100 mM NEt4Cl acetonitrile−water mixtures at 590 nm following 1300 nm excitation. Water concentration is shown in the legend. Solvent contribution was subtracted. The best singleexponential fits of the data are shown as lines.
The 460 and 585 nm ΔA bands are due to the absorption from the 2A1 LF excited state.6 Indeed, these ΔA bands cannot be attributed to reaction products, such as either CuCl3− and Cl− or CuCl32− and Cl·,10,26−28 because the absorption spectra of these species are inconsistent with the observed transient absorption. Also, the observed stability of the complex for at least 24 h upon continuous 1300 nm irradiation and the fast decay (
