Probing the Impact of Solvation on Photoexcited Spin Crossover

Nov 10, 2017 - Our studies demonstrate that, although the metal center of [FeII(mbpy)3]2+ could have been expected to be rather shielded by the three ...
0 downloads 11 Views 2MB Size
Download PDF
Article Cite This: J. Am. Chem. Soc. XXXX, XXX, XXX-XXX

pubs.acs.org/JACS

Probing the Impact of Solvation on Photoexcited Spin Crossover Complexes with High-Precision X‑ray Transient Absorption Spectroscopy Cunming Liu,† Jianxin Zhang,‡ Latévi M. Lawson Daku,⊥ David Gosztola,∥ Sophie E. Canton,*,§,∇ and Xiaoyi Zhang*,† †

X-ray Science Division, and ∥Center for Nanoscale Materials, Argonne National Laboratory, 9700 S. Cass Avenue, Lemont, Illinois 60439, United States ‡ State Key Laboratory of Hollow Fibre Membrane Materials and Processes, School of Environmental and Chemical Engineering, Tianjin Polytechnic University, Tianjin 300387, China ⊥ Département de Chimie Physique, Université de Genève, Quai E. Ansermet 30, CH-1211 Genève 4, Switzerland § ELI-ALPS, ELI-HU Non-Profit Ltd., Dugonics ter 13, Szeged 6720, Hungary ∇ Attosecond Science Group, Deutsches Elektronen Synchrotron (DESY), Notkestrasse 85, D-22607 Hamburg, Germany S Supporting Information *

ABSTRACT: Investigating the photoinduced electronic and structural response of bistable molecular building blocks incorporating transition metals in solution phase constitutes a necessary stepping stone for steering their properties toward applications and performance optimizations. This work presents a detailed X-ray transient absorption (XTA) spectroscopy study of a prototypical spin crossover (SCO) complex [FeII(mbpy)3]2+ (where mbpy = 4,4′-dimethyl-2,2′bipyridine) with an [FeIIN6] first coordination shell in water (H2O) and acetonitrile (CH3CN). The unprecedented data quality of the XTA spectra together with the direct fitting of the difference spectra in k space using a large number of scattering paths enables resolving the subtle difference in the photoexcited structures of an FeII complex in two solvents for the first time. Compared to the low spin (LS) 1A1 state, the average Fe−N bond elongations for the photoinduced high spin (HS) 5T2 state are found to be 0.181 ± 0.003 Å in H2O and 0.199 ± 0.003 Å in CH3CN. This difference in structural response is attributed to ligand−solvent interactions that are stronger in H2O than in CH3CN for the HS excited state. Our studies demonstrate that, although the metal center of [FeII(mbpy)3]2+ could have been expected to be rather shielded by the three bidentate ligands with quasi-octahedral coordination, the ligand field strength in the HS excited state is nevertheless indirectly affected by solvation effects that modifies the charge distribution within the Fe−N covalent bonds. More generally, this work highlights the importance of including solvation dynamics in order to develop a generalized understanding of the spin-state switching at the atomic level.

■

INTRODUCTION

absorption toggles the electronic state of the complexes between a low-spin (LS) state and a high-spin (HS) state.3−16 Since the first report of unusual magnetism in Fe(III) complexes,17 numerous experimental and theoretical studies have been carried out in order to identify the various physicochemical factors that govern such spin-state transitions.3−5,12,15−21 At the microscopic level, the collective response within SCO bulk materials results from a complicated interplay between intramolecular processes, intermolecular interactions (e.g., between the complex and the solvent molecules or counterions), and crystal packing forces.22 Investigations conducted in the solution phase can circumvent

Advanced photoswitchable materials based on transition-metal molecular complexes display macroscopic features (e.g., electronic, magnetic, optical properties, and conductivity) that can be dynamically tailored through light illumination.1 Understanding the ultrafast dynamics of the photoswitching event is of fundamental and practical importance for controlling light-triggered reactivity and for implementing novel lightdriven operating principles toward applications in data storage, digital processing, and communication.2 As synergistically bistable molecular systems, the large family of 3d4−3d7 transition-metal complexes is intensively developed for their spin crossover (SCO) behavior and for exploiting the lightinduced excited spin state trapping (LIESST) effect, where light © XXXX American Chemical Society

Received: August 31, 2017 Published: November 10, 2017 A

DOI: 10.1021/jacs.7b09297 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Article

Journal of the American Chemical Society

mM were measured at the Center for Nanoscale Materials (CNM) of Argonne National Laboratory in a pump−probe setup described elsewhere.53 The optical length of sample cuvette was 2 mm. The 800 nm, ∼100 fs (full width half-maximum, fwhm), 2.5 kHz output from a Coherent Libra Ti:Sapphire ultrafast regenerative amplifier was split into two beams: the one with the majority of the power was sent to an optical parametric amplifier to produce the 340 nm pump source, which was tuned to a fluence of ∼0.05 mJ/mm2, and the other was focused on a 3 mm sapphire plate to generate white light continuum as the probe beam. The OTA kinetic traces were fit by a singleexponential function, f(t) = A exp(−t/τ), convoluted with a Gaussian instrument response function (IRF) of ∼140 fs (fwhm), where τ represents the relaxation lifetime from the HS excited state (ES) to the LS ground state (GS). X-ray Transient Absorption (XTA) Spectroscopy. The XTA spectra of 4 mM [FeII(mbpy)3]2+ dissolved in H2O and CH3CN, respectively, were collected at the beamline 11-ID-D of the Advanced Photon Source (APS) in Argonne National Laboratory. The 400 nm laser pump pulse was the second-harmonic output from a 10 kHz Coherent Legend Elite Ti:Sapphire ultrafast amplifier laser system. This pulse was further stretched to ∼1 ps (fwhm) by a pair of prisms. The probe was ∼78.8 ps (fwhm) X-ray pulse at 6.5 MHz delivered from the APS storage ring under the 24-bunch operation mode. The laser pump and X-ray probe beams intersected on a free-flowing sample jet of ∼0.6 mm in diameter. The diameter of the laser spot was ∼650 μm at the sample position, yielding a fluence of ∼0.43 mJ/mm2. The X-ray beam size was 450 μm (horizontal) by 100 μm (vertical) at the sample position, hence completely covered by the laser. The X-ray fluorescence photons from the FeII centers were collected at 90° angle on both sides of the incident X-ray beam by two avalanche photodiodes (APDs). A soller slits/Mn filter combination, which was custom-designed for the sample chamber configuration and the distance between the sample and the detector, was inserted between the free sample jet and the APD detectors. The laser_on spectrum was collected using the fluorescence signal from the X-ray pulse at 120 ps after photoexcitation, and the laser_off (GS) spectrum was obtained by averaging 12 X-ray pulses before the laser excitation. In a typical XTA spectrum measurement, each energy point was integrated for 4 s. The XTA kinetic traces were fit by a single-exponential function as for the OTA kinetics analysis but convoluted with a Gaussian IRF of 78.8 ps due to the much broader X-ray pulse. XTA Spectroscopy Data Analysis and Fitting Procedure. Experimental Data Reduction. The XTA spectroscopic measurements of [FeII(mbpy)3]2+ in H2O and CH3CN under the same experimental conditions (i.e., laser power, concentration, jet speed, etc.) were repeated 70 and 40 times, respectively. The averaged spectra were provided as input to the Athena program in order to process the GS X-ray absorption (XA) spectra (μlaser_off) and to extract the normalized oscillation amplitude χexp GS (k), where k is the photoelectron wavenumber defined by k = 2m(E − E0) /ℏ and E0 is the absorption edge energy. Since the GS spectrum (μlaser_off) and laser_on spectrum (μlaser_on) were measured simultaneously, the normalization and background removal parameters of the GS spectrum were used to extract the experimentally measured χexp laser_on(k) om the laser_on spectrum (μlaser_on). Extended X-ray Absorption Fine Structure (EXAFS) Data Fitting. An in-house developed software with perl script incorporating FEFF 9.05 was employed in order to fit the EXAFS data. The theoretical oscillation amplitude of the X-ray absorption spectrum, χ(k) was calculated with the EXAFS equation,54

the intricacies associated with the long-range forces imposed in the solid phase and are therefore particularly valuable for correlating the observed rates of interconversion with specific attributes of the complexes (e.g., the Z number and oxidation state of the transition metal ion or the nature of the ligand scaffold).22,23 At the molecular level, extensive synthetic work has established how steric and electronic effects, as well as ligand conformations can dramatically tune the spin-multiplicity of the molecular ground state.22,24,25 Recent comparative characterizations also suggest that strongly associating solvents usually stabilize the LS state over the HS state.23,26 In other words, the local environment around the metal center, which determines the spin-multiplicity of the state with minimal energy, encompasses not only the ligand coordination shell but also the solvation shell. However, elaborating similar design trends for the photoexcited metastable state has been largely hampered so far by the lack of analytical methods capable of tracking simultaneously and directly the coupled evolution of the spin, electronic, and nuclear degrees of freedom on the ultrafast time scale. The fast-paced development of X-ray spectroscopies and Xray scattering now provides the scientific community with ground-breaking tools uniquely suited for addressing longstanding open questions in the multidisciplinary area of ultrafast structural dynamics.16,22,27−33 In particular, the continuous optimization of X-ray beamline stability, detection schemes, and sample environment at synchrotron facilities has led to drastic improvements in the attainable signal-to-noise ratio over a few years,32,34−36 opening up unprecedented possibilities for reliably measuring very small (∼0.01 Å) bond length changes in photoexcited solutes using high resolution Xray transient absorption (XTA) spectroscopy.37,38 In this work, the resolution of the XTA technique is further improved, enabling for the first time determining with extremely highprecision (0.003 Å) the ultrasmall differences between the photoexcited structures of the HS for [FeII(mbpy)3]2+ (where mbpy = 4,4′-dimethyl-2,2′-bipyridine) in two different solvents, water (H2O) and acetonitrile (CH3CN). Detailed correlation between the average Fe−N bond elongations and the lifetimes of the HS state in H2O and CH3CN evidences stronger ligand− solvent interactions for the HS in H2O. These experimental findings are furthermore discussed in connection to DFT calculations that employ the BLYP-D3 functional and a continuum treatment of solvation with the COSMO model.

■

EXPERIMENTAL SECTION

Samples. [FeII(mbpy)3](PF6)239 and [FeII(mbpy)3]Cl240 soluble in CH3CN and H2O, respectively, were synthesized according to the reported protocols. All reagents were used as received from commercial sources without further purification. DFT Calculations. Density functional theory (DFT)41,42 has been applied to the determination of the geometry adopted by the [FeII(mbpy)3]2+ complex in the LS state and in the HS state. The calculations have been performed with the ADF program package43 using the dispersion-corrected BLYP-D3 functional,44−46 with the Slater-type TZP basis set of triple-ζ polarized quality from the ADF basis set database.47,48 The influence of H2O and CH3CN as solvents has then been taken into account using the conductor-like screening model (COSMO).49−52 UV−Vis Absorption Spectroscopy. The UV−visible absorption spectra of 0.33 mM [FeII(mbpy)3]2+ dissolved in H2O and CH3CN were taken on a Shimadzu UV-2401 PC spectrophotometer. The optical length of sample cuvette was 2 mm. Optical Transient Absorption (OTA) Spectroscopy. The femtosecond OTA spectra of samples with the concentration of 0.33

χ theo (k) =

∑ j

S0 2Nfj j (k) kR j 2

2 2

e−2k σj e−2rj / λ(K ) sin[2kR j + δj(k , rj)] (1)

where j indicates the jth shell with identical backscatters, Nj is the coordination number of the jth shell, f j is the backscattering amplitude, Rj is the average distance between the center atom and backscatters, σj is the mean square variation in Rj, δj is the scattering phase shift, λ(k) B

DOI: 10.1021/jacs.7b09297 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society

■

is the effective mean free path, and S02 is the amplitude reduction factor. Determination of the LS GS Structure. The LS GS structure of [FeII(mbpy)3]2+ from the DFT calculation was taken as a starting point. The parameters f j, Nj, δj, and λ were first calculated with FEFF 9.05.55 The nonstructural parameters, S02, Rj, σj2, and E0 were refined through least-squares minimization of the k3-weighted EXAFS spectra χ(k)*k3 with k ranging from 2.3 to 11.0 Å−1. Since the mbpy ligand is rather rigid, a model with frozen mbpy ligand structure moving by the same amount along the line that bisects the angle N−Fe−N′ (Figure 4) was used to construct the FEFF 9.05 input files during the fitting procedure. Going beyond the conventional analysis schemes that first transform the data into R space before performing the EXAFS fitting, here we directly fitted χ(k) in k space. The contributions from all the scattering paths with more than 4% curved wave amplitude ratio (i.e., a total of 81 paths) were included to calculate χ(k). All the paths shared a common S02 and E0. The σj2 was obtained by fixing the temperature to 300 K (i.e., room temperature) and fitting the Debye temperature. Determination of the HS ES Structure. The experimentally measured and calculated difference absorption spectra are defined by eq 2 and eq 3 respectively, exp exp Δexpχ (k) = χES (k) − χGS (k) =

Article

RESULTS AND DISCUSSION Figure 1a shows a “ball and stick” schematic for the molecular structure of the [FeII(mbpy)3]2+ complex in H2O in its GS

1 exp exp [χ (k) − χlaser (k)] _off η laser_on (2)

theo

Δ

χ (k) =

theo χES (k)

−

theo χGS (k)

(3)

Figure 1. (a) “Ball and stick” schematic for the molecular structure of the [FeII(mbpy)3]2+ complex in its GS obtained from DFT optimization. (b) Normalized UV−vis absorption spectra of 0.33 mM [FeII(mbpy)3]2+ in H2O and CH3CN. (c) General energy level scheme for Fe(II) SCO complexes, indicating the electronic cascade that follows photoexcitation from the LS ground state (1A1). (d) Kinetics traces of [FeII(mbpy)3]2+ in H2O and CH3CN probed at 527 nm (ground state bleach band) by OTA after photoexcitation at ∼340 nm. The black curves are the fits based on a single-exponential function as described in the Experimental Section.

where η is the ES fraction measured in the laser_on spectrum and theo χtheo ES (k) − χGS (k) represents the calculated transient difference spectrum. The structural parameters from the best fit of the experimental GS spectra were used to calculate χtheo GS (k). The optimized GS structure was introduced as the starting structure of the ES state. A similar structural model assuming rigid mbpy ligands moving along the line that bisects the angle N−Fe−N′ (Figure 4) and equal metal− ligand bond elongation was employed to construct the FEFF 9.05 input for the ES. In the difference spectrum fitting, the S02 and E0 from the GS data analysis were used for the ES. The σj2 and E0 of the ES, the bond-length change from GS to ES (ΔRFe−N), and the ES fraction η were refined by least-squares minimization of the k3 weighted χ defined by eq 2 and eq 3. The uncertainties were estimated through variance analysis from a least-squares regression. The final uncertainties reported in Table 1

obtained from the DFT optimization. The coordination symmetry is quasi-octahedral so that the strong ligand field splits the atomic-like 3d levels into t2g and eg molecular orbitals. As for the large majority of Fe(II)-based SCO complexes possessing an [FeIIN6] first coordination shell, the 1A1 (t2g6eg0, S = 0) LS state is stabilized as the GS.3−5 The steady-state optical absorption spectra of [FeII(mbpy)3]2+ in H2O and CH3CN are shown in Figure 1b. Both the spectral shape and the extinction coefficient are nearly identical in the two solvents. Similarly to [FeII(bpy)3]2+, the intense absorption band at ∼355 nm is assigned to a π−π* transition, while the band between 450 and 560 nm arises from the singlet metal-toligand charge transfer (1MLCT) transition.56,57 As shown in Figure 1c, the photoexcited 1MLCT state evolves through successive intersystem crossing (ISC) steps to the HS state 5T2 (t2g4eg2, S = 2), which relaxes nonradiatively back to the LS (1A1) GS.27,30,57,58 Drastic differences between electronic and geometric structures can be expected across the spin-state transition as two electrons from the nonbonding t2g orbitals are promoted to the antibonding eg orbitals. OTA measurements were performed in order to characterize the ultrafast dynamics. Figure S1 displays the two-dimensional femtosecond OTA spectra of [FeII(mbpy)3]2+ in H2O and CH3CN. Both samples exhibit a strong GS bleach band centered at ∼527 nm.57,59,60 Fitting the GS bleach (GSB) recovery kinetics shows that the lifetimes of the solvated HS state are 830 ± 10 ps in H2O and 1240 ± 12 ps in CH3CN (Figure 1d). These values are comparable to the ∼650 ps and

have been scaled by χ 2 , which is equivalent to rescaling the input error to obtain χ2 = 1.

Table 1. Fitting Parameters for the First N Shell of Fe(II) in the LS and HS States of [FeII(mbpy)3]2+ Dissolved in H2O and CH3CN solvent

spin

H2O

LS HS LS HS

CH3CN

RFe−N (Å) (CN = 6)

ΔRFe−N

S02

± ± ± ±

0.181 ± 0.003

0.91 ± 0.02 0.91 0.91 ± 0.02 0.91

1.981 2.162 1.979 2.178

0.002 0.003 0.002 0.003

0.199 ± 0.003

Since the uncertainties on the bond length obtained with the present analysis are extremely small, an alternative error-estimation method based on fitting of replicate measurements was then tested in order to validate the precision of the fitting results. The XTA spectra measured in H2O and CH3CN were both divided into 10 groups. Each group were processed and fitted using the same procedure as the one described above to obtain 10 repetitions of fitting results. The structural parameters extracted from the fitting of individual groups are listed in Tables S1 and S2 of the Supporting Information (SI). The structural parameters and uncertainties obtained from the two approaches are very comparable as summarized in Table S3. C

DOI: 10.1021/jacs.7b09297 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Article

Journal of the American Chemical Society ∼1340 ps reported for the closely related complex [FeII(bpy)3]2+ in H2O27 and in CH3CN, respectively.6 Considering that the OTA spectra of the complex in both solvents show nearly identical GSB spectroscopic features devoid of structural information about the HS state, additional diagnostics have to be obtained with XTA in order to elucidate the microscopic origin of the different dynamics. The μlaser_off GS X-ray absorption spectra of the complexes dissolved in H2O and CH3CN measured at the Fe K-edge are presented in Figure 2a,b (blue curves). They are very similar to the ones reported for photoexcited [FeII(bpy)3]2+ and [FeII(terpy)2]2+ complexes.6,7,11,13,16,22,27

nonbonding t2g orbital (LS) to populate the antibonding eg orbital (HS) (Figure 1c). Figure 2c presents the XTA kinetics of [FeII(mbpy)3]2+ probed at feature B (∼7.125 keV), corresponding to the maximum positive signal in the difference XTA spectra (Figure 2d). The HS lifetimes have been found through single exponential fitting to be 865 ± 11 ps and 1295 ± 10 ps in H2O and CH3CN, respectively. These values are fully consistent with those reported above using OTA spectroscopy as seen in Figure S2 where the XTA and OTA HS decay traces nearly superpose. Figure 2d shows the difference XA spectra Δμ = μlaser_on − μlaser_off for [FeII(mbpy)3]2+ (time delay Δt of 120 ps) in H2O and CH 3 CN, which were collected under the same experimental conditions (i.e., laser power, concentration, jet speed, etc.). Analyzing the transient EXAFS spectra allows quantifying accurately the structural evolution. Table 1 lists the major structural parameters obtained from the EXAFS data analysis (Figure 3a−d) described in the Experimental Section. The extremely high data quality together with the advanced analysis allows determining the average Fe−N bond length with an unprecedented precision as high as 0.002 Å. Figure 3e shows the Fourier transformed (FT) EXAFS spectra (in R space) of the LS GS in H2O and CH3CN. The strongest peaks in these traces correspond to the average bond

Figure 2. Normalized Fe K-edge X-ray absorption (XA) spectra μLaser_off (blue) and μLaser_on (red) of [FeII(mbpy)3]2+ in (a) H2O and (b) CH3CN. The time delay was 120 ps. The inset zooms into the near-edge region. (c) X-ray kinetics probed at 7.125 keV for [FeII(mbpy)3]2+ in H2O (blue) and CH3CN (red). The black lines in panel c represent the single exponential fits. (d) Comparison between the difference traces Δμ = μlaser_on − μlaser_off at the time delay of 120 ps. The inset shows the pre-edge region.

The X-ray absorption fingerprints typical for [FeIIN6] complexes are observed in the near edge region (Figure 2a,b, inset):6,10 (A) is the weak quadrupole-allowed 1s−3d transition at the pre-edge (∼7.112 keV), due to the mixing of 3d−4p orbitals upon loss of centrosymmetry, (B) (∼7.125 keV) and (C) (∼7.130 keV) are assigned to the promotion of the 1s core electron to the lowest unoccupied states hybridized between Fe (4s, 4p) and N (2p), (D) (∼7.142 keV) is due to the multiple scattering of the photoelectron, and (E) (∼7.175 keV) corresponds to the single scattering of the photoelectron. No distinctive spectral alteration (i.e., shift or distortion) of these features could be observed in the GS spectra between H2O to CH3CN within the experimental resolution. Moreover, the GS spectra exhibit the same absorption edge energy, E0. These combined observations imply that the solvated LS GS in the two solvents has identical electronic and geometric structures. The red curves, μlaser_on in Figure 2a,b show the laser_on spectra recorded at a time delay of 120 ps. The transient spectra reflect the coupled electronic and nuclear rearrangements taking place in the HS ES.6,7,36 The intensities of features B, C, and E are enhanced with a slight downshift in energy of their rising edges. This is in line with the expected bond elongation triggered by the unpairing of two electrons from the

Figure 3. Fe K-edge EXAFS spectra and the corresponding fitting (black line) of LS GS [FeII(mbpy)3]2+ in (a) H2O (blue) and (c) CH3CN (red). The difference EXAFS spectra between laser_on (time delay Δt of 120 ps) and laser_off and the corresponding best fits (black line) of [FeII(mbpy)3]2+ in (b) H2O (blue) and (d) CH3CN (red). The Fourier transform (FT) of the EXAFS functions χ(k)*k3 in R-space for [FeII(mbpy)3]2+ in H2O (blue) and CH3CN (red) of the (e) LS GS and (f) constructed HS ES. The spectra in panels e and f are phase-uncorrected, so the distance R shown in those panels are smaller than the actual values. D

DOI: 10.1021/jacs.7b09297 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Article

Journal of the American Chemical Society length R between Fe(II) center and its first N shell in the two solvents. As listed in Table 1, R from the best fit for the LS GS are 1.981 ± 0.002 Å in H2O and 1.979 ± 0.002 Å in CH3CN, so that they can be considered identical within the experimental precision. Surprisingly, the bond-length elongation, ΔR, from the LS state to the HS state is significantly different in the two solvents, namely, 0.181 ± 0.003 Å in H2O and 0.199 ±0.003 Å in CH3CN (Table 1). In addition, the direct fitting of the difference EXAFS data (Figure 3b,d) delivers the ES fraction η in the laser_on spectra: 0.31 ± 0.01 in H2O and 0.35 ± 0.01 in CH3CN. Using these η, the FT EXAFS spectra of the HS states can be constructed from laser_on and laser_off spectra with eq 2 as shown in Figure 3f. It is clearly seen that the average Fe−N bond length in the ES is longer in CH3CN than in H2O. Given the facts that the UV−vis absorbance are identical in the two solvents and that the XTA measurements were conducted under the same experimental conditions, similar values of η could be expected for H2O and CH3CN. However, it should be kept in mind that the actual η is the excited state population averaged over the time span of the ∼78.8 ps X-ray pulse. Based on this consideration, the ratio of excited state population measured at 120 ps delay is η(H2O)/η(CH3CN) = 0.95. This value is in line with the fact that the HS lifetime is shorter in H2O than in CH3CN. The results of the DFT structure optimizations performed with the BLYP-D3 functional and a continuum treatment of solvation (COSMO model) are summarized in Table 2, along

Figure 4. Comparison between the DFT optimized complex molecular structures of the LS GS and the photoinduced HS ES for solvated [FeII(mbpy)3]2+ in water. In the present EXAFS fitting model, the three mbpy ligands are moved rigidly along the lines (arrows) bisecting the angle N−Fe−N′ by the ΔR and the averaged groove angle changes from ∼89.2° to ∼93.7°. The H atoms connected to C atoms are here omitted for clarity.

bond elongation and the concurrent expansion of the average groove angle from 89.2° to 93.7° in the HS state (Figure 4, right)25,48,63 allows closer approach of the smaller, more polar H2O molecules compared to CH3CN. Consequently, a stronger solvent (H2O)−ligand (mbpy) interaction may occur and participate in the local electric fields that modulate the ligand field strength. The solvent influence on the character (covalent versus ironic) of the Fe−N bond in the photoinduced HS state has also been suggested by time-resolved measurements in the soft-X-ray regime for the closely related [Fe(bpy) 3 ] 2+ complex.64 Finally, the subtle differences in HS structure due to solvation have a direct impact on the photoinduced relaxation. The measured effect of solvent on the bond length clearly mediates the spin crossover dynamics. Pioneering work by Hauser has established that for Fe(II) SCO complexes, at low temperature, the HS state decays nonradiatively back to the LS following the so-called inverse-energy-gap law: the lifetime of the HS state increases as ΔEHL decreases.15,65−68 However, until recently, it was difficult to conduct direct experimental investigation to test the applicability of inverse-energy-gap law in solution at room temperature due to the short-lived nature of the metastable HS species. Such measurements are now enabled by ultrafast X-ray techniques that provide structural observables complementary to the energetic observables delivered by other analytical techniques. In the present case, a shorter Fe−N bond length in the HS excited state implies a stronger ligand field strength and therefore a larger energy gap ΔEHL and a shorter lifetime for the HS state if the inverse-energy gap law is obeyed.15,65−69 The OTA and XTA studies on [FeII(mbpy)3]2+ outlined here reveal a shorter Fe−N bond-length and a faster decay of HS state in H2O than in CH3CN. Collectively, these results indicate that the inverse-energy-gap law is indeed holding at room temperature in solution for this important class of photoactive complexes. Disentangling the precise role of the intra- and intermolecular forces that are at play remains very challenging. This problem will be tackled with ongoing picosecond time-resolved wide-angle X-ray scattering studies coupled to ab initio molecular dynamics simulations that are capable of accessing the dynamics of the solvation shell on the atomic length scale. In addition, elucidating the details of the spin, electronic, and

Table 2. DFT Calculated ΔEHL and ΔRHL of [FeII(mbpy)3]2+ in H2O and CH3CN using BLYP-D3 H2O

CH3CN

functional

ΔEHL (eV)

ΔRHL (Å)

ΔEHL (eV)

ΔRHL (Å)

BLYP-D3

1.378

0.188

1.364

0.189

with the calculated HS−LS electronic energy difference, ΔEHL = EHS − ELS, and the average Fe−N bond change (ΔRHL) from HS to LS in the two solvents. The predicted values of ΔRHL agree well with the experimental results but barely depend on the nature of the solvent. Similarly, the values of ΔEHL calculated for the complex in the two solvents are almost identical, with only a slight difference of 14 meV ≈ 82 cm−1. These results indicate that the use of an implicit model of solvation is not sufficient to capture the subtle solvation effects evidenced by the present experiments. It is well-recognized that continuum solvation models indeed tend to overlook the influence of the specific solute−solvent interactions (e.g., H bonding, preferential orientation, charge transfer), which are all short-ranged effects in that they are mostly operative in the first solvation shell.61 Consequently, understanding the microscopic origin of this solvation sensitivity requires an atomistic description of the solute−solvent interactions since the difference in ΔRHL ultimately results from the very sensitive balance between the strength of the ligand field and the magnitude of solute−solvent interactions. From a structural point of view, the LS exhibits shorter Fe− N bond length, hence stronger ligand field, compared to the HS (Figure 4).25,62 The relatively narrow interligand groove in the LS state imposes strong steric hindrance that tends to shield the ligand field from solvent impact (Figure 4, left). Therefore, the LS states have similar Fe−N bond lengths in the two solvents. However, the enlarged interligand grooves caused by the Fe−N E

DOI: 10.1021/jacs.7b09297 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Article

Journal of the American Chemical Society

European Union and cofinanced by the European Regional Development Fund. The laser system at 11-ID-D of APS was funded through New Facility and Midscale Instrumentation grants to Chemical Sciences and Engineering Division, Argonne National Laboratory (PI: Lin X. Chen). This research used resources of the Advanced Photon Source and the Center for Nanoscale Materials, U.S. Department of Energy (DOE), Office of Science User Facilities operated for the DOE Office of Science by Argonne National Laboratory under Contract No. DE-AC02-06CH11357.

nuclear cascade that leads to the photoinduced formation of the HS state on the sub-picosecond time scale will be undertaken in upcoming measurements at X-ray free electron laser facilities in order to reveal the details of the competition between intersystem crossing and vibrational cooling at the onset of the ultrafast photoinduced dynamics.

■

CONCLUSION In summary, by using XTA spectroscopic measurements, we have compared the electronic and structural dynamics of [FeII(mbpy)3]2+ dissolved in H2O and CH3CN, in order to reveal the elusive solvent effects on the photoinduced SCO phenomenon. By combining unprecedented XTA data quality with an analysis approach based on directly fitting the XA difference spectra in k space including a large number of scattering paths, we could determine the local structural response of [FeII(mbpy)3]2+ upon photoexcitation with extremely high precision, thereby resolving the slight structural differences induced by solvation for the first time. The Fe−N bond length from the LS (1A1) GS to HS (5T2) ES in H2O is 0.181 ± 0.003 Å, which is smaller than that in CH3CN (0.199 ± 0.003 Å). In addition, the HS lifetime increases from 865 ± 11 ps in H2O to 1295 ± 10 ps in CH3CN. These observations directly demonstrate that, although the metal center of Fe(II) SCO complexes could have been expected to be rather shielded from solute−solvent interactions by the methylated ligands, solvation effects significantly affect the photoinduced structural dynamics. These combined results about the HS structures and decay kinetics in the two solvents indicate that the relaxation from the HS ES to the LS GS follows the inverse-energy-gap law in solution at room temperature. The presented methodology can be generally applied to the elaboration of an atomistic model that will describe the detailed role of solvation on the photoexcited structures in SCO complexes, thereby paving the way to bridging the description of their spin-switching behavior from the liquid to the solid phase.

■

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.7b09297. 2D OTA spectra, comparison of XTA and OTA kinetics, and error analysis (PDF)

■

REFERENCES

(1) Ichikawa, H.; Nozawa, S.; Sato, T.; Tomita, A.; Ichiyanagi, K.; Chollet, M.; Guerin, L.; Dean, N.; Cavalleri, A.; Adachi, S.-i.; Arima, T.-h.; Sawa, H.; Ogimoto, Y.; Nakamura, M.; Tamaki, R.; Miyano, K.; Koshihara, S.-y. Nat. Mater. 2011, 10, 101. (2) Lewanowicz, A.; Luty, T. Acta phys. Pol. A 2012, 121, 291. (3) Gütlich, P.; Goodwin, H. A. Top. Curr. Chem. 2004, 233, 1. (4) Kepp, K. P. Inorg. Chem. 2016, 55, 2717. (5) Goodwin, H. A. Top. Curr. Chem. 2004, 233, 59. (6) Canton, S. E.; Zhang, X.; Lawson Daku, L. M.; Smeigh, A. L.; Zhang, J.; Liu, Y.; Wallentin, C.-J.; Attenkofer, K.; Jennings, G.; Kurtz, C. A.; Gosztola, D.; Wärnmark, K.; Hauser, A.; Sundström, V. J. Phys. Chem. C 2014, 118, 4536. (7) Zhang, X.; Lawson Daku, M. L.; Zhang, J.; Suarez-Alcantara, K.; Jennings, G.; Kurtz, C. A.; Canton, S. E. J. Phys. Chem. C 2015, 119, 3312. (8) Ohkoshi, S.-i.; Imoto, K.; Tsunobuchi, Y.; Takano, S.; Tokoro, H. Nat. Chem. 2011, 3, 564. (9) Vankó, G.; Bordage, A.; Pápai, M.; Haldrup, K.; Glatzel, P.; March, A. M.; Doumy, G.; Britz, A.; Galler, A.; Assefa, T.; Cabaret, D.; Juhin, A.; van Driel, T. B.; Kjær, K. S.; Dohn, A.; Møller, K. B.; Lemke, H. T.; Gallo, E.; Rovezzi, M.; Németh, Z.; Rozsályi, E.; Rozgonyi, T.; Uhlig, J.; Sundström, V.; Nielsen, M. M.; Young, L.; Southworth, S. H.; Bressler, C.; Gawelda, W. J. Phys. Chem. C 2015, 119, 5888. (10) Khalil, M.; Marcus, M. A.; Smeigh, A. L.; McCusker, J. K.; Chong, H. H. W.; Schoenlein, R. W. J. Phys. Chem. A 2006, 110, 38. (11) Gawelda, W.; Pham, V. T.; Benfatto, M.; Zaushitsyn, Y.; Kaiser, M.; Grolimund, D.; Johnson, S. L.; Abela, R.; Hauser, A.; Bressler, Ch.; Chergui, M. Phys. Rev. Lett. 2007, 98, 057401. (12) Saes, M.; van Mourik, F.; Gawelda, W.; Kaiser, M.; Chergui, M.; Bressler, Ch.; Grolimund, D.; Abela, R.; Glover, T.; Heimann, P.; Schoenlein, R.; Johnson, S.; Lindenberg, A.; Falcone, R. Rev. Sci. Instrum. 2004, 75, 24. (13) Chergui, M. Acta Crystallogr., Sect. A: Found. Crystallogr. 2010, 66, 229. (14) Papanikolaou, P.; Margadonna, S.; Kosaka, W.; Ohkoshi, S.; Brunelli, M.; Prassides, K. J. Am. Chem. Soc. 2006, 128, 8358. (15) Hauser, A. Top. Curr. Chem. 2004, 234, 155. (16) Bressler, Ch.; Chergui, M. Annu. Rev. Phys. Chem. 2010, 61, 263. (17) Cambi, L.; Szego, L. Ber. Dtsch. Chem. Ges. B 1931, 64, 2591. (18) Vef, A.; Manthe, U.; Gütlich, P.; Hauser, A. J. Chem. Phys. 1994, 101, 9326. (19) Halder, G. J.; Kepert, C. J.; Moubaraki, B.; Murry, K. S.; Cashion, J. D. Science 2002, 298, 1762. (20) Zhou, J.; Wang, Q.; Sun, Q.; Kawazoe, Y.; Jena, P. J. J. Phys. Chem. Lett. 2012, 3, 3109. (21) Tao, J.; Wei, R.-J.; Huang, R.-B.; Zheng, L.-S. Chem. Soc. Rev. 2012, 41, 703. (22) Gütlich, P.; Goodwin, H. A. Spin Crossover in Transition Metal Compounds I-III; Springer-Verlag: Berlin Heidelberg, Germany, 2004. (23) Linert, W. Highlights in Solute-Solvent Interactions; SpringerVerlag: New York, 2002. (24) Halcrow, M. A. Crystals 2016, 6, 58. (25) Phan, H.; Hrudka, J. J.; Igimbayeva, D.; Lawson Daku, L. M.; Shatruk, M. J. Am. Chem. Soc. 2017, 139, 6437. (26) Lawson Daku, M. L.; Hauser, A. J. Phys. Chem. Lett. 2010, 1, 1830.

AUTHOR INFORMATION

Corresponding Authors

*[email protected] *[email protected], [email protected] ORCID

Latévi M. Lawson Daku: 0000-0003-1305-6807 Xiaoyi Zhang: 0000-0001-9732-1449 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences under Contract No. DE-AC02-06CH11357. J.Z. gratefully acknowledges the support from NSFC (21302138). S.E.C. acknowledges funding from the Helmoltz Recognition Award. The ELIALPS project (GINOP-2.3.6-15-2015-00001) is financed by the F

DOI: 10.1021/jacs.7b09297 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Article

Journal of the American Chemical Society (27) Chergui, M. Ultrafast Studies of the Light-Induced Spin Change in Fe(II)-Polypyridine Complexes. In Spin-Crossover Materials: Properties and Applications; Halcrow, M. A., Ed.; John Wiley & Sons: Oxford, UK, 2013; Chapter 15. (28) Haldrup, K.; Gawelda, W.; Abela, R.; Alonso-Mori, R.; Bergmann, U.; Bordage, A.; Cammarata, M.; Canton, S. E.; Dohn, A. O.; van Driel, T. B.; Fritz, D. M.; Galler, A.; Glatzel, P.; Harlang, T.; Kjær, K. S.; Lemke, H. T.; Møller, K. B.; Németh, Z.; Pápai, M.; Sas, N.; Uhlig, J.; Zhu, D.; Vankó, G.; Sundström, V.; Nielsen, M. M.; Bressler, Ch. J. Phys. Chem. B 2016, 120, 1158. (29) Zhang, W.; Gaffney, K. Acc. Chem. Res. 2015, 48, 1140. (30) Zhang, W.; Alonso-Mori, R.; Bergmann, U.; Bressler, Ch.; Chollet, M.; Galler, A.; Gawelda, W.; Hadt, R. G.; Hartsock, R. W.; Kroll, T.; Kjær, K. S.; Kubiček, K.; Lemke, H. T.; Liang, H. W.; Meyer, D. A.; Nielsen, M. M.; Purser, C.; Robinson, J. S.; Solomon, E. I.; Sun, Z.; Sokaras, D.; van Driel, T. B.; Vankó, G.; Weng, T. C.; Zhu, D.; Gaffney, K. J. Nature 2014, 509, 345. (31) Zhang, W.; Kjær, K. S.; Alonso-Mori, R.; Bergmann, U.; Chollet, M.; Fredin, L. A.; Hadt, R. G.; Hartsock, R. W.; Harlang, T.; Kroll, T.; Kubiček, K.; Lemke, H. T.; Liang, H. W.; Liu, Y.; Nielsen, M. M.; Persson, P.; Robinson, J. S.; Solomon, E. I.; Sun, Z.; Sokaras, D.; van Driel, T. B.; Weng, T.-C.; Zhu, D.; Wärnmark, K.; Sundström, V.; Gaffney, K. J. Chem. Sci. 2017, 8, 515. (32) Chergui, M. Struct. Dyn. 2016, 3, 031001. (33) Bressler, Ch.; Milne, C.; Pham, V. − T.; ElNahhas, A.; van der Veen, R. M.; Gawelda, W.; Johnson, S.; Beaud, P.; Grolimund, D.; Kaiser, M.; Borca, C. N.; Ingold, G.; Abela, R.; Chergui, M. Science 2009, 323, 489. (34) Chen, L. X.; Zhang, X. J. Phys. Chem. Lett. 2013, 4, 4000. (35) van Bokhoven, J. A.; Lamberti, C. X-Ray Absorption and X-Ray Emission Spectroscopy: Theory and Applications; John Wiley & Sons: Chicago, 2016. (36) Chen, L. X.; Zhang, X.; Shelby, M. L. Chem. Sci. 2014, 5, 4136. (37) Zhang, X.; Canton, S. E.; Smolentsev, G.; Wallentin, C.-J.; Liu, Y.; Kong, Q.; Attenkofer, K.; Stickrath, A. B.; Mara, M. W.; Chen, L. X.; Wärnmark, K.; Sundström, V. J. Am. Chem. Soc. 2014, 136, 8804. (38) Zhang, X.; Pápai, M.; Møller, K. B.; Zhang, J.; Canton, S. E. Molecules 2016, 21, 235. (39) Onggo, D.; Hook, J. M.; Rae, D. A.; Goodwin, H. A. Inorg. Chim. Acta 1990, 173, 19. (40) Jaeger, F. M.; van Dijk, J. A. Zeitschrift Fur Anorganische Und Allgemeine Chemie 1936, 227, 273. (41) Hohenberg, P.; Kohn, W. Phys. Rev. 1964, 136, B864. (42) Kohn, W.; Sham, L. J. Phys. Rev. 1965, 140, A1133. (43) te Velde, G.; Bickelhaupt, F. M.; Baerends, E. J.; Fonseca Guerra, C.; van Gisbergen, S. J. A.; Snijders, J. G.; Ziegler, T. J. Comput. Chem. 2001, 22, 931. (44) Becke, A. D. Phys. Rev. A: At., Mol., Opt. Phys. 1988, 38, 3098. (45) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B: Condens. Matter Mater. Phys. 1988, 37, 785. (46) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. J. Chem. Phys. 2010, 132, 154104. (47) Chong, D. P. Can. J. Chem. 1995, 73, 79. (48) Van Lenthe, E.; Baerends, E. J. J. Comput. Chem. 2003, 24, 1142. (49) Klamt, A. J. Phys. Chem. 1995, 99, 2224. (50) Klamt, A.; Jonas, V. J. Chem. Phys. 1996, 105, 9972. (51) Klamt, A.; Schüürmann, G. J. Chem. Soc., Perkin Trans. 2 1993, 799. (52) Pye, C.; Ziegler, C. Theor. Chem. Acc. 1999, 101, 396−408. (53) Burdett, J. J.; Gosztola, D.; Bardeen, C. J. J. Chem. Phys. 2011, 135, 214508. (54) Bunker, G. Introduction to XAFS: A Practical Guide to X-ray Absorption Fine Structure Spectrsocopy; Cambridge University Press: New York, 2010. (55) Rehr, J. J.; Kas, J. J.; Vila, F. D.; Prange, M. P.; Jorissen, K. Phys. Chem. Chem. Phys. 2010, 12, 5503. (56) Braterman, P. S.; Song, J.-I.; Peacock, R. D. Inorg. Chem. 1992, 31, 555. (57) Auböck, G.; Chergui, M. Nat. Chem. 2015, 7, 629.

(58) Lemke, H. T.; Kjær, S. K.; Hartsock, R.; van Driel, T. B.; Chollet, M.; Glownia, J. M.; Song, S.; Zhu, D.; Pace, E.; Matar, S. F.; Nielsen, M. M.; Benfatto, M.; Gaffney, K. J.; Collet, E.; Cammarata, M. Nat. Commun. 2017, 8, 15342. (59) Monat, J. E.; McCusker, J. K. J. Am. Chem. Soc. 2000, 122, 4092. (60) Gawelda, W.; Cannizzo, A.; Pham, V.-T.; van Mourik, F.; Bressler, Ch.; Chergui, M. J. Am. Chem. Soc. 2007, 129, 8199. (61) Kelly, C. P.; Cramer, C. J.; Truhlar, D. G. J. J. Chem. Theory Comput. 2005, 1, 1133. (62) Hauser, A. Adv. Polym. Sci. 2004, 233, 49. (63) Van Kuiken, B. E.; Cho, H.; Hong, K.; Khalil, M.; Schoenlein, R. W.; Kim, T. K.; Huse, N. J. Phys. Chem. Lett. 2016, 7, 465. (64) Huse, N.; Cho, H.; Hong, K.; Jamula, L.; de Groot, F. M. F.; Kim, T. K.; McCusker, J. K.; Schoenlein, R. W. J. Phys. Chem. Lett. 2011, 2, 880. (65) Hauser, A. Comments Inorg. Chem. 1995, 17, 17. (66) Hauser, A. Top. Curr. Chem. 2004, 233, 49. (67) Hauser, A.; Enachescu, C.; Daku, M. L.; Vargas, A.; Amstutz, N. Coord. Chem. Rev. 2006, 250, 1642. (68) Hauser, A.; Vef, A.; Adler, P. J. Chem. Phys. 1991, 95, 8710. (69) Gutlich, P.; Goodwin, H. A. Top. Curr. Chem. 2004, 233, 1.

G

DOI: 10.1021/jacs.7b09297 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX