A Synthetically Tunable System To Control MLCT Excited-State

Subscriber access provided by UNIV OF REGINA

Article

A Synthetically-Tunable System to Control MLCT ExcitedState Lifetimes and Spin States in Iron(II) Polypyridines Steven M Fatur, Samuel G Shepard, Robert F. Higgins, Matthew P. Shores, and Niels H. Damrauer J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.7b00700 • Publication Date (Web): 01 Mar 2017 Downloaded from http://pubs.acs.org on March 1, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of the American Chemical Society is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

A Synthetically-Tunable System to Control MLCT ExcitedState Lifetimes and Spin States in Iron(II) Polypyridines Steven M. Fatur,

†‡

Samuel G. Shepard,†‡ Robert F. Higgins,§ Matthew P. Shores,§ Niels H.

Damrauer†* †

Department of Chemistry and Biochemistry, University of Colorado Boulder, Boulder,

Colorado, 80309, United States §

Department of Chemistry, Colorado State University, Fort Collins, Colorado 80523, United

States

Abstract: 2,2′:6′,2′′-terpyridyl (tpy) ligands modified by fluorine (dftpy), chlorine (dctpy), or bromine (dbtpy) substitution at the 6- and 6′′- positions are used to synthesize a series of bishomoleptic Fe(II) complexes. Two of these species, [Fe(dctpy)2]2+ and [Fe(dbtpy)2]2+, which incorporate the larger dctpy and dbtpy ligands, assume a high-spin quintet ground state due to substituent-induced intramolecular strain. The smaller fluorine atoms in [Fe(dftpy)2]2+ enable spin crossover with a T1/2 of 220 K and a mixture of low-spin (singlet) and high-spin (quintet) populations at room temperature. Taking advantage of this equilibrium, dynamics originating from either the singlet or quintet manifold can be explored using variable wavelength laser excitation. Pumping at 530 nm leads to ultrafast non-radiative relaxation from the singlet metalto-ligand charge transfer (1MLCT) excited state into a quintet metal centered state (5MC) as has been observed for prototypical low-spin Fe(II) polypyridine complexes such as [Fe(tpy)2]2+. On the other hand, pumping at 400 nm excites the molecule into the quintet manifold (5MLCT ← 5

MC) and leads to the observation of a greatly increased MLCT lifetime of 14.0 ps. Importantly,

this measurement enables an exploration of how the lifetime of the 5MLCT (or 7MLCT, in the event of intersystem crossing) responds to the structural modifications of the series as a whole. We find that increasing the amount of steric strain serves to extend the lifetime of the

5,7

MLCT

from 14.0 ps for [Fe(dftpy)2]2+ to the largest known value at 17.4 ps for [Fe(dbtpy)2]2+. These data support the design hypothesis wherein interligand steric interactions are employed to limit conformational dynamics and/or alter relative state energies, thereby slowing non-radiative loss of charge-transfer energy. 1

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 33

Introduction Recent advances in the use of transition-metal-based photocatalysts have highlighted their capabilities at performing complex and energetically-demanding organic transformations.1-3 Originally, many of these strategies relied on precious transition metals such as ruthenium2,4 and iridium,5,6 but recently, there has been an increasing amount of work concerning more abundant transition metals such as copper,7,8 chromium,9-11 zirconium,12 and molybdenum.13 Despite the growing body of photoredox catalysis research, iron, the most abundant transition metal in the earth’s crust, has proven more difficult to utilize due to subpicosecond non-radiative relaxation of the metal-to-ligand charge transfer (MLCT) state into a low-lying metal-centered (MC) state.14-16 While some groups have reported organic transformations using Fe(II) polypyridyl systems as photocatalysts,17-19 the conversion of the optically bright 1MLCT to a much lower energy high-spin ligand-field state on an ultrafast time scale (vide infra) suggests that the mode of action must be different compared to related Ru(II) systems. This point has been recently noted by Collins et al. in describing their use of a [Fe(phen)3]2+ photocatalyst.19 For general purposes the rapid MLCT relaxation in Fe(II) polypyridines remains an impediment to their widespread adoption for catalyzing organic transformations. It has been established that MLCT loss in prototypical [Fe(2,2’-bipyridine)3]2+ is ultrafast and occurs via intersystem crossing from the 1MLCT to the 5MC, which carries < 25% of the original excitation energy.16,20-23 It is expected that states such as the intermediates

22-24

or coupling pathways

16,21

3

MC serve as

25

(or both ) in these dynamics, having both

intermediate energies and structures (notably along a metal-ligand bond coordinate) in the transformation from

1,3

MLCT to 5MC. Because of the complexity of the rearrangements that

occur between the initial and final excited states, theoretical analyses of relevant geometric changes26 and state couplings27 within parent complexes, as well as electronic structure predictions of new ligand sets28-30 can inform synthetic alterations that slow or shut down decay pathways. For example, increasing the ligand field strength may be expected to change the energetic accessibility of metal-centered states whose conformations engender their participation in non-radiative decay. McCusker and coworkers have explored the use of an alternate tridentate ligand that supports geometries more closely approaching true octahedral coordination.31 While effective at switching the energetic ordering of 3T1 and 5T2 MC states due to enhanced ligand

2


Page 3 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


field, the deactivation of the MLCT manifold is still rapid. Focusing specifically on strengthening metal-ligand bonds, the groups of Gros as well as Sundström and Wärnmark have used N-heterocyclic carbene ligands to dramatically increase Fe(II) MLCT excited state lifetimes to 26 ps32 and 18 ps,33 respectively. Furthermore, both groups have shown high electron injection rates in dye-sensitized TiO2 photoelectrodes,33,34 thereby demonstrating an approach to utilizing these high potential excited states. More recently, a heteroleptic complex with a combination of cyano ligands and 2,2′-bipyridine ligands has been studied as a way of separately controlling the 3,5

MC and 1,3MLCT state energies to achieve a 20 ps lifetime.35 This strong ligand-field strategy

may be applicable in other d6 systems as suggested by Wenger and coworkers who reported a > 2 ns MLCT excited state lifetime in a tris-diisocyanide Cr(0) complex.36 Recent work in our group has introduced a complimentary strategy for extending MLCT lifetimes, albeit from high-spin states. This approach uses a sterically demanding ligand, 6,6′′dichloro-2,2′;6′2′′-terpyridine (dctpy). In the homoleptic complex [Fe(dctpy)2]2+, the halogen atoms of each of the ligands repulsively interact with the opposing ligand thereby impacting the metal-ligand coordination. This reduction of the ligand field serves to destabilize the closed shell 1

A (and open shell 3MC) resulting in a quintet ground state (5MC) and an optically accessible

5

MLCT. That research found a charge transfer lifetime (5MLCT or 7MLCT, hereafter

5,7

MLCT)

of 16 ps corresponding to a > 300 fold increase in the longevity of the MLCT compared to prototypical species such as [Fe(bpy)3]2+.37 The hypothesis that emerges has its basis in two ideas. First, it is conceivable that compared to the 1MLCT or 3MLCT of prototypical low-spin Fe(II) polypyridyl complexes, the 3

5,7

MLCT happen to be more poorly coupled to states such as

MC that can drive non-radiative decay processes. Second, it is conceivable that that the size of

the substituents driving the interligand steric interactions serve to diminish the conformational freedom of the system within the excited state manifold, thus preventing dynamics necessary for intersystem crossing events within the relaxation process. This current work seeks to explore these ideas by virtue of exploring the sensitivity of high-spin MLCT lifetimes to the size of substituents that drive the interligand steric interactions in the first place. In an effort to minimize other potential perturbations, we explore a halogen series of molecules (see Scheme 1), bookending the aforementioned [Fe(dctpy)2]2+ with bromine- and fluorine-containing analogues [Fe(dbtpy)2]2+ and [Fe(dftpy)2]2+, respectively. Within this work, it is important to recognize that the parent [Fe(tpy)2]2+ has a singlet ground state (1A), suggesting that for substituent sizes

3



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

smaller than Cl, the interconversion from 5MC to 1A may become thermally allowed. Should this occur, it would be an interesting addition to the body of literature on light-induced excited-state spin trapping (LIESST)38,39 where a low driving force and a high-barrier for 5MC→1A interconversion (as would be expected given the steric constraints in dftpy) results in extended lifetimes of the 5MC. Given that [Fe(dctpy)2]2+ has an optically bright 5MLCT←5MC transition, but relaxes back to the 5MC state in 16 ps, we became curious about the possibility that a lowerlying 1A in [Fe(dftpy)2]2 might alter that pathway, perhaps enabling a reverse-LIESST40 effect and creating a long-lived transient singlet population. Therefore, the exploration herein considers optical, crystallographic, and magnetic data, which then enables a comparative exploration of excited state dynamics and the impact of halogen size.

Scheme 1. General Structure of Ligands and Complexes

Experimental General Synthesis. All solvents and reagents were obtained from Sigma-Aldrich and used without further purification unless otherwise noted. Deuterated solvents were obtained from Cambridge Isotope Laboratories, Inc. NMR spectra were acquired on a Bruker Avance-III 300 MHz spectrometer, mass spectrometry was performed at the University of Colorado Central Analytical Laboratory, and elemental analysis was performed by Robertson Microlit Laboratories Inc., Ledgewood, NJ. The syntheses of 6,6''-dichloro-2,2':6',2''-terpyridine (dctpy) and [Fe(dctpy)2](BF4)2 have been described previously.37 6,6''-difluoro-2,2':6',2''-terpyridine (dftpy). Using standard Schlenk techniques, 2,6dibromopyridine (168 mg, 0.71 mmol, 1.0 eq.) was combined with 6-fluoropyridine-2-boronic acid (Combi-Blocks, 300 mg, 2.13 mmol, 3.0 eq.), and palladium(0) tetrakis-triphenylphosphine (123 mg, 15 mol. %) in 72 mL of dinitrogen-sparged 1,4-dioxane. An additional solution of 375

4


Page 4 of 33

Page 5 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


mg (5 eq.) of sodium carbonate in 24 mL of deionized water (0.15 M) was sparged with dinitrogen and added to the reaction vessel. The reaction mixture was heated to reflux for 24 hours with stirring until the 2,6-dibromopyridine was consumed, as monitored by TLC. ~75 mL of 2:1 water:ethyl acetate was added and the organic layer was washed with water three times with water, then dried with magnesium sulfate, filtered, and evaporated to dryness. The crude solid was purified by column chromatography using chloroform on silica gel (Rf = 0.7). Yield: 151 mg, 79%. 1H NMR (300 MHz, chloroform-d) δ 8.49 (ddd, J = 7.6, 2.4, 0.8 Hz, 2H), 8.40 (d, J = 7.8 Hz, 2H), 8.01-7.89 (m, 3H), 6.98 ppm (ddd, J = 8.1, 3.0, 0.8 Hz, 2H).

13

C NMR (75

MHz, chloroform-d): δ 163.15 (d, J = 238.7 Hz), 154.77 (d, J = 13.0 Hz), 153.88, 141.78 (d, J = 7.6 Hz), 138.14, 121.70, 118.17 (d, J = 4.0 Hz), 109.61 ppm (d, J = 37.4 Hz). HRMS (ESI-TOF) m/z: [M+Na]+ Calcd for C15H9N3F2 292.0662 [(M+Na)+]; Found 292.0660. [Fe(dftpy)2](BF4)2. Using standard Schlenk techniques, Fe(II) tetrafluoroborate hexahydrate (29 mg, 0.086 mmol, 1.0 eq.) was combined with dftpy (45 mg, 0.17 mmol, 2.0 eq.) in 24 mL of dinitrogen-sparged 1,4-dioxane. The solution was heated to reflux with stirring for 24 hours, whereupon the resulting mixture was evaporated to dryness, then dissolved in acetonitrile and filtered through glass wool paper. The filtrate was concentrated in vacuo and crystallized from slow diffusion of diethyl ether into acetonitrile where the product was collected as red-orange blocks. Crystals suitable for x-ray diffraction were obtained via slow evaporation of acetonitrile. Yield: 20 mg, 30%. 1H NMR (300 MHz, acetonitrile-d3) δ 70.79 (br. s, 4H), 70.37 (br. s, 4H), 52.06 (br. s, 4H), 14.96 (br. s, 4H), -24.60 ppm (br. s, 2H). A reproducible sub-spectrum was observed corresponding to a diamagnetic species which is assigned to the low-spin state of the same molecule. The integration values reported for this state have been scaled by 12.5 times relative to the integration values listed above in order to facilitate comparison of the diamagnetic peaks to one another. Assigned to singlet ground state: δ 8.88 – 8.67 (m, 4H), 8.64 – 8.46 (m, 4H), 8.42 – 8.25 (m, 4H), 8.19 – 7.99 (m, 2H), 7.32 – 7.05 ppm (m, 4H). Anal. calcd for C30H18N6FeB2F12: C, 46.92; H, 2.36; N, 10.94. Found: C, 46.64; H, 2.21; N, 10.68. HRMS (ESITOF) m/z: M2+ Calcd for C30H18N6Br4FeB2F8 297.0439; Found 297.0436. [Fe(dbtpy)2](BF4)2. The same synthetic procedure was used as for [Fe(dftpy)2](BF4)2. Fe(II) tetrafluoroborate hexahydrate (34 mg, 0.10 mmol, 1.0 eq.) was combined with dbtpy (86 mg, 0.22 mmol, 2.2 eq.) to produce orange crystalline plates. Crystals suitable for x-ray diffraction were obtained by vapor diffusion of diethyl ether into an acetonitrile solution. Yield: 40 mg,

5



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

40%. 1H NMR (300 MHz, acetonitrile-d3) δ 71.19 (br. s, 4H), 60.60 (br. s, 4H), 47.05 (br. s, 4H), 11.09 (br. s, 4H), -28.39 ppm (br. s, 2H). Anal. Calcd. for C30H18N6Br4FeB2F8: C, 35.62; H, 1.79; N, 8.31. Found: C, 35.90; H, 2.01; N, 8.20. HRMS (ESI-TOF) m/z: M2+ Calcd for C30H18N6Br4FeB2F8 418.8820; Found 418.8823. Transient Absorption (TA) Spectroscopy. Data collected on timescales ranging from 1 ps to 1600 ps used methods described previously.37,41 For TA data collected on longer time scales out to 400 ns, the following setup was used. Samples were prepared in HPLC-grade acetonitrile (EMD Millipore) in a 1 cm pathlength cuvette which was sparged with argon prior to the experiment and then sealed. A Continuum Surelite II-10 laser equipped with third harmonic generation optics pumped a Surelite optical parametric oscillator (OPO) with 355 nm light to produce a 520 nm excitation source with an ~4 ns pulse width at a 10 Hz repetition rate. This pump was attenuated to 5.2 mJ and focused with a cylindrical lens to a size of 4 mm (width) by 1.1 mm (height) in the sample. Concurrently, a 100 W xenon arc lamp output (Newport/Oriel) was collimated and then softly focused (100 mm focal length) into the sample with a beam diameter of 0.7 mm at a 90° angle relative to the pump path. After the sample, the probe was directed into a monochromator where spectral regions of ~ 2 nm could be dispersed onto a photomultiplier tube (Hamamatsu R-928). Data were collected on a digital oscilloscope (Lecroy LC574AM) and processed in Labview using a home-built program. X-ray Diffraction. Crystals were mounted under a stream of dinitrogen on a cryoloop with Paratone-N oil. Data collection was performed at 100 K, 120 K, or 300 K as noted using a Bruker D8-Quest Eco diffractometer, Mo Kα radiation, and a graphite monochromator. Data integration and absorption correction were performed using the APEX 3 software suite (Bruker) and structures were solved by direct methods and refined in OLEX242 using SHELXS, SHELXT, and SHELXL.43 All non-hydrogen atoms were refined anisotropically by full matrix leastsquares on F2. Hydrogen atoms were placed in calculated positions and refined using a riding model. Additional experimental details can be found in the Supporting Information. Magnetic Susceptibility. Data were collected using a Quantum Design MPMS XL SQUID magnetometer. All sample preparations were performed on the bench top. Powdered microcrystalline samples were dissolved in acetonitrile and placed into a straw, whereupon the straw was sealed. The compounds [Fe(dftpy)2](BF4)2 and [Fe(dctpy)2](BF4)2 were measured at concentrations of 0.077 and 0.053 M, respectively. For solid state measurements, powdered

6


Page 6 of 33

Page 7 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


microcrystalline samples were placed into polyethylene bags, sealed and inserted into a straw. Ferromagnetic impurities were probed through a variable field (H) analysis (0 to 10 kOe) of the magnetization (M) at 100 K: no curvature was observed in the M vs H plot (Figures S4 and S5), indicating that there are no significant ferromagnetic impurities in these samples. Magnetic susceptibility data were collected at temperatures ranging from 2 to 315 K for [Fe(dctpy)2](BF4)2 in acetonitrile, 100 to 315 K for [Fe(dftpy)2](BF4)2 in acetonitrile, and 2 to 315 K for

[Fe(dftpy)2](BF4)2 in the solid state. The samples that were dissolved in acetonitrile were inserted into the magnetometer at ~30-50 K essentially flash-freezing them. They were then warmed to 100 K for the M vs H scan, and then cooled to 2 K to begin the susceptibility sequence. After warming to 315 K, the samples were immediately cooled back down to 2 K. Data were corrected for the diamagnetic contributions of the sample holder and solvent by subtracting empty containers; corrections for the sample were calculated from Pascal’s constants.44 Other Physical Measurements. UV-Visible absorption spectra were collected on a HewlettPackard diode array spectrophotometer (HP8452A). Samples were prepared using HPLC-grade acetonitrile (EMD Millipore) in a 1 cm pathlength cuvette. Temperature control of samples was performed using a refrigerated constant temperature circulator (VWR 1145). Detailed procedures for cyclic voltammetry and Evans’ NMR measurements have been described elsewhere.37 Electronic Structure Calculation. Geometry minimized structures were generated using the Q-Chem software package.45 For halopyridines, the Hartree-Fock method was used along with the 6-31G* basis set for all atoms. For metal complexes, the B3LYP functional and the 6-31G* basis set were used for all atoms. Although this level of theory is insufficient for absolute energies, it was chosen because it was found to yield accurate trends in relative energies of related systems.46 Orbital energies were extracted from the diagonalized Fock matrix. Restricted, closed-shell calculations were performed on the pyridine and ground state zinc systems (model complexes [Zn(dftpy)2]2+, [Zn(dctpy)2]2+, and [Zn(dbtpy)2]2+) and unrestricted, open-shell calculations were performed on the iron systems and the singly-reduced zinc-containing systems. In the open-shell systems, β MO energies are reported.

Results and Discussion [Fe(dbtpy)2] 2+ electronic absorption. The UV-Vis absorption spectrum of [Fe(dbtpy)2]2+ is displayed in Figure 1, where the features closely match those of the previously investigated

7



[Fe(dctpy)2]2+.37 For both molecules, there are two transitions appearing at ~ 220 nm and ~ 297 nm that are also observed in the free ligand spectra (see Figure S1). These are expected to be ligand-based π* ← π in nature. Moving to the red, intense transitions are observed at ~ 350 nm, that by contrast are absent in the free ligand spectra. On the basis of this observation as well as TD-DFT explorations of [Fe(dctpy)2]2+, observations concerning the model [Zn(dctpy)2]2+, and spectroelectrochemical data for [Fe(dctpy)2]2+, the feature in [Fe(dbtpy)2]2+ is assigned as being ligand-to-ligand charge transfer (LLCT) in nature.37 From our perspective, the most important spectral quality that is observed in [Fe(dbtpy)2]2+ is a long tail that extends from 400-500 nm. We have previously argued that a similar feature in the [Fe(dctpy)2]2+ spectrum is evidence for MLCT within the quintet manifold (5MLCT ← 5GS).37 There, the transition is blue-shifted and less absorptive relative to the parent [Fe(tpy)2]2+ (562 nm, 12,000 M-1cm-1),47 consistent with the assignment of a ground state quintet (in the halogen-substituted molecule) where longer Fe-N bond lengths reduce metal-ligand orbital overlap and raise the energy of the charge transfer state by Coulombic arguments. Given that the larger size of the bromine substituents in [Fe(dbtpy)2]2+ will only serve to increase interligand repulsions compared to [Fe(dctpy)2]2+, we readily make the same spectral assignment for this new molecule.

80x103

60

40

2000 Molar Absorptivity

Molar Absorptivity / M-1 cm-1

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Fe(dftpy)2]2+ Fe(dctpy)2]2+ Fe(dbtpy)2]2+

1000

0 300

400 500 600 Wavelength / nm

20

200

300 400 Wavelength / nm

500

600

Figure 1. UV-vis absorption spectra in room-temperature acetonitrile for [Fe(dftpy)2]2+, [Fe(dctpy)2]2+, and [Fe(dbtpy)2]2+ with inset to highlight redder transitions.

8


Page 8 of 33

Page 9 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Data from other experiments are consistent with the assignment of a high-spin ground state in [Fe(dbtpy)2]2+. From Evans’ measurements, a magnetic susceptibility value of χMT = 3.3 cm3 K mol-1 (µeff = 5.1) is obtained, which is a typical value for a high-spin Fe(II) complex with four unpaired electrons, and it is consistent with our previous measurement of µeff = 5.3 in [Fe(dctpy)2]2+.37 Further evidence of a pure high-spin state is found in the 1H spectrum of [Fe(dbtpy)2]2+ (Figure S14), which shows paramagnetically shifted peaks between 80 and -30 ppm and no evidence of low-spin aromatic hydrogen resonances. Structural Data for both [Fe(dctpy)2] 2+ and [Fe(dbtpy)2] 2+. X-ray diffraction data were collected for single crystals of both [Fe(dctpy)2]2+ and [Fe(dbtpy)2]2+ at 120 K and 100 K, respectively. The resultant structural models display geometries that are consistent with a quintet ground state, while at the same time highlighting the steric strain that encourages the electronic configuration of four unpaired electrons in metal-centered orbitals, two of which are antibonding in nature. This latter point about steric strain is highlighted in Figure 2, which presents thermal ellipsoid plots for [Fe(dctpy)2]2+ (the corresponding structure for [Fe(dbtpy)2]2+ is very similar and is provided in Figure S11). Viewed along the axis containing the central Fe-N bonds (lefthand image), the distortion of the terpyridine planes is immediately noticeable. The vertical dctpy ligand is twisted about the two inter-ring C-C bonds, so that the two chlorine atoms are pointing towards opposite sides of the central pyridyl group with the C-Cl bonds at a pseudodihedral angle of 28.0(2)°. A similar effect is seen in the horizontal dctpy ligand but is less pronounced at only 10.1(2)°. The overall twisting distortion is likely occurring because of repulsion between the chlorine atoms and the π-system of the central pyridine on the opposing ligand, especially the more electronegative nitrogen atom. This steric interaction is undoubtedly critical in the determination of the ground spin state in both [Fe(dctpy)2]2+ and [Fe(dbtpy)2]2+. To limit its energy-raising effect, these molecules adopt an electronic configuration where Fe–N bonds are elongated and where the terpyridine Nterminal-Ncentral-Nterminal (N-N-N) angle is consequentially increased, thus serving to extend the distance of the halogen from the π-system of the second ligand relative to what is seen in low-spin analogues such as [Fe(tpy)2]2+. In [Fe(dctpy)2]2+ and [Fe(dbtpy)2]2+ respectively, the Fe-Ncentral bond elongates to 2.080(2) and 2.074(2) Å compared to a typical low-spin value of ~1.9 Å (Table 1). Overall, there is a substantial deviation from a true octahedral coordination geometry as evidenced by the average

9



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 33

Nterminal-Fe-Ncentral bond angle of ~75° in the two halogenated complexes, even smaller than the ~81° seen in [Fe(tpy)2]2+.48

Figure 2. Crystal structure of the Fe-containing complex in [Fe(dctpy)2](BF4)2 shown from two distinct points of view with thermal ellipsoids rendered at 50% probability. C = gray, Cl = green, Fe = orange, N = blue. Hydrogen atoms have been removed for clarity. Despite the Fe–N elongation, the [Fe(dctpy)2]2+ and [Fe(dbtpy)2]2+ structures show short distances between the halogen substituent and the central pyridyl ring of the opposing ligand. The values are tabulated in Table 1 (X-pyridyl) and are ~0.35 Å shorter than the sum of the Bondi radii of the halogen and carbon.49,50 The fact that these distances are still short despite the significant geometric distortions described above underscores the unfavorable steric interactions that would persist in a low-spin complex. A final tabulated metric for these complexes is the socalled “rocking angle” Θ, which represents the angular deviation from 180° for Ncentral-Fe-Ncentral’.26 We note this quantity by way of connecting to work by Jakubikova and coworkers who have shown that the quintet state of [Fe(tpy)2]2+ exhibits a substantial (10°) value for Θ and a soft potential in this rocking direction whereas the singlet ground state is locked into a linear arrangement.26 In [Fe(dctpy)2]2+ and [Fe(dbtpy)2]2+, we observe values of 4.1(1)° and 11.65(8)°, with the larger halogen leading to the larger value.

Table 1. X-Ray Structural Data – Selected Parameters complex

Fe-Ncentral (Å) Fe-Nterminal (Å)

∠ N-N-N (°)

Θ (°)

X-pyridyl (Å)

[Fe(tpy)2]2+ a

1.890(4)

1.988(3)

102.8(2)

1.5(3)

N/A

[Fe(dctpy)2]2+

2.080(2)

2.272(1)

111.4(1)

4.1(1)

3.074(1)

10


Page 11 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


[Fe(dbtpy)2]2+

2.074(2)

2.304(2)

112.65(8)

11.65(8)

3.179(2)

a

Structural data from reference.48 Data taken at 295 K for [Fe(tpy)2]2+, 120 K for [Fe(dctpy)2]2+ and 100 K for [Fe(dbtpy)2]2+. See Supporting Information for additional details. [Fe(dftpy)2] 2+ electronic absorption. The absorption spectrum of [Fe(dftpy)2]2+ shown in Figure 1 exhibits marked differences from those belonging to [Fe(dctpy)2]2+ and [Fe(dbtpy)2]2+. We first note the higher energy peaks at 288 nm and 332 nm. These mirror features seen in the bromine and chlorine-substituted complexes, but exhibit a blue shift of ~ 10 nm. Given their respective assignments as π* ← π and LLCT in nature, we anticipate energy shifts in ligandcentered occupied and unoccupied orbitals. To gain insight into this blue shift, we have considered zinc(II) complexes within the halogen series using density functional theory (DFT). This d10 metal was chosen because the metalcentered orbitals of the complexes are stabilized (and filled) such that HOMO and LUMO orbitals are ligand π and π* in nature. These ground state calculations therefore afford a cursory view into the origins of energy shifts in bands assigned to LLCT. Calculated energies of the HOMO and LUMO for [Zn(dftpy)2]2+, [Zn(dctpy)2]2+, and Zn(dbtpy)2]2+ are listed in Table S3. As seen computationally, fluorine substitution within the series serves to stabilize the HOMO while destabilizing the LUMO, suggesting that this complex would be more difficult to oxidize as well as more difficult to reduce compared to its chlorine- and bromine-containing analogues. These orbital energy trends are also observed in mono-substituted halogenated benzenes51 and halogenated pyridines (see also Table S2). Overall, it appears that stabilization of occupied π orbitals has its origin in the increased electronegativity of the F substituent, relative to Cl or Br, which acts through an inductive effect. On the other hand, the fluorine substituent is more effective at destabilizing the LUMO through π-donation than the larger halogen atoms.52,53 It is noted that HOMO and LUMO orbital energy trends are consistent across the full series of Zn(II) complexes although the perturbations are significantly more subtle in going from Cl to Br then they are upon going from F to Cl. We next note an even more dramatic spectroscopic change that was first encountered by visual inspection of acetonitrile solutions of [Fe(dftpy)2]2+, which have a much redder appearance than the other halogenated complexes (orange-yellow solutions). UV-Vis absorption spectroscopy reveals the emergence of a new prominent transition at 506 nm for the fluorinated

11



complex, with a molar absorptivity on the order of 103 M-1cm-1 (Figure 1, inset). This absorptivity is too high to be assigned to a metal-centered d-d transition. Given the smaller size of the fluorine substituents in [Fe(dftpy)2]2+ and therefore the potential for increased metal-ligand interactions, we suspected that at room temperature there would be partial population of a singlet ground state and a concomitant 1MLCT ← 1GS transition. To test this further, we have collected UV-Vis absorption spectra as a function of temperature over the range 277 K – 334 K and find the expected result that the visible absorption feature that is peaked at 506 nm increases in intensity as the sample is cooled (Figure 3). This temperature dependence heralds an equilibrium between high- and low-spin populations, where the singlet state is favored enthalpically and where the quintet increases in prominence as temperature increases due to entropic contributions to the Gibbs free energy. This same effect on UV-Vis absorption has been observed in Fe(II) bis-4,6-diphenylterpyridine, where temperaturedependence is attributed to spin crossover (SCO).54

277 K 279 K 281 K 283 K 285 K 294 K 304 K 314 K 324 K 334 K

1.0 Absorbance

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 33

0.5

0 400

450

500 550 Wavelength / nm

600

650

Figure 3. UV-visible absorption spectra of a solution of [Fe(dftpy)2]2+ in acetonitrile at temperatures between 277-334 K. SQUID Magnetometry Data. In order to investigate thermodynamic parameters within the spin-crossover equilibrium, we performed magnetic property measurements that could be applied over a wide temperature range. Magnetic susceptibility data for flash-frozen acetonitrile solutions of [Fe(dftpy)2](BF4)2 and [Fe(dctpy)2](BF4)2, collected via SQUID magnetometry, are

12


Page 13 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


shown in Figure 4. The chlorine-containing compound [Fe(dctpy)2](BF4)2 displays a χMT value of 3.30 cm3 K mol-1 at 315 K, in agreement with the expected spin only value for a noninteracting high-spin Fe(II) species (3.00 cm3 K mol-1 for g = 2).55 Upon cooling, the χMT product decreases gradually until ~225 K, at which point a discontinuity in χMT coincides with the freezing of the acetonitrile solvent. Between 200 K and 18 K, the monotonic decrease in χMT may be interpreted as temperature-independent paramagnetism (TIP), very gradual spin crossover, or possibly (weak) antiferromagnetic coupling between high-spin Fe(II) complexes. The sharper decrease observed below 18 K is likely due to magnetic anisotropy of any remaining high-spin Fe(II) fraction.55

Figure 4. Temperature dependence of magnetic susceptibility data collected at 1000 Oe for [Fe(dftpy)2](BF4)2 (red circles, warming and blue circles, cooling) and [Fe(dctpy)2](BF4)2 (green squares) in the presence of acetonitrile. The cyan line corresponds to the best fit of the data to equation S3. The parameters of the best fit are: gHS = 2.11, gLS = 2.00, θHS = + 2.18 K, θLS = 0.001 K, TIPHS = 3.0 × 10-3 cm3 mol-1, TIPLS = 3.4 × 10-3 cm3 mol-1 and ∆HSCO = 24.5 kJ/mol assuming T1/2 = 220 K. Interestingly, and in alignment with our expectations considering the electronic absorption measurements, [Fe(dftpy)2](BF4)2 shows quite different temperature-dependent magnetic susceptibility behavior. At 315 K, the complex exhibits a χMT value of 3.13 cm3 K mol-1, also in agreement with other non-interacting high-spin Fe(II) complexes. Upon cooling, however, the χMT product decreases, first gradually, then more sharply until ~ 190 K (χMT = 0.87 cm3 K mol-1) where the change in susceptibility slows. The large change in the χMT value between 300 K and 190 K is attributed to a non-hysteretic spin crossover (SCO) event wherein the system converts 13



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 33

from S = 2 at high temperatures to S = 0 at low temperatures. Discontinuities observed in the χMT product in the temperature range 200 – 225 K are similar to those observed for the chloride analogue, and perhaps relate to the freezing/thawing of acetonitrile. The non-zero χMT values below 190 K show a similar temperature dependence as [Fe(dctpy)2](BF4)2. If this is attributed entirely to TIP, then the admixture of excited states is very large (χTIP ~3.4 × 10-3 cm3mol-1). We note that large TIP has been reported in systems where other excited spin-states are accessible,5658

and is evident in frozen solution for Fe(II) triazacyclononane complexes.59 As depicted in Figure 4, this SCO event can be fit to an ideal solution model.60,61 Assuming a

single spin-crossover event (S = 0 → S = 2) and no cooperativity, the model provides T1/2 = 220 K and thermodynamic parameters ∆HSCO = 24.5 kJ mol-1 and ∆SSCO = 111 J K-1mol-1 (calculated using ∆SSCO = ∆HSCO/T1/2). These thermodynamic parameters are at the higher end of the range reported for Fe(II) SCO complexes, but similar to values measured in pyridine-toluene solutions for salen-type [LFeII(py)2] complexes which bear a N4O2 coordination geometry.62 The large ∆SSCO value suggests a significant entropic benefit due to populating the high-spin state. The large TIP values and χMT discontinuities in the SCO regime prompted us to perform solid state magnetic measurements on [Fe(dftpy)2](BF4)2 (Figure S6). From these measurements, we can conclude that a full SCO event is possible in the solid state (T1/2 ~ 160 K), and uncommonly large TIP is not observed. However, these data also show a complicated two-step transition in the temperature regime 175-250 K, overlapping with the “solution” data described above. The origins of this behavior are under investigation and the results will be disclosed in a future report. Deconvolving the T-Dependent Absorption. Based on the thermodynamic parameters extracted from SQUID, we can return to the temperature dependent UV-Vis absorption data for [Fe(dftpy)2]2+ that were introduced earlier (Figure 3). In order to determine molar extinction (ε) values for the high- and low-spin states, we have globally modeled the absorbance of the sample as a function of temperature, T, and wavelength, λ, with a modified Beer’s law (see Supporting Information for details). Using the ∆H and ∆S values extracted from SQUID measurements, a global analysis of the T-dependent absorption data (Figure 3; between 277 K and 334 K) shows that two basis spectra cleanly reproduce the observed behavior (Figure S2). These basis spectra are shown in Figure 5.

14


Page 15 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 5. Molar absorptivities of the high-spin and low-spin states of [Fe(dftpy)2]2+ in acetonitrile determined from global analysis of the T-dependent UV-Visible data (Figure 3). Also shown for comparison is the measured molar absorptivity of [Fe(tpy)2]2+ in room temperature acetonitrile. Discussing first the low-spin basis spectrum, we note that it bears remarkable similarity in band shape to that of the low-spin parent [Fe(tpy)2]2+, although it is blue-shifted by ~50 nm. Care must be taken assessing this shift using electrochemical data (Table S6) given that both oxidation and reduction events in [Fe(dftpy)2]2+ are positively shifted by similar amounts (0.32 V and 0.35 V, respectively) relative to corresponding potentials measured for the parent [Fe(tpy)2]2+. This leads to comparable values of ∆E1/2, a quantity that is generally considered to be a useful proportional indicator of MLCT energy.63 However, as we have noted previously for [Fe(dctpy)2]2+,37 the quantity ∆E1/2 is expected to underestimate the MLCT values when there is significant high-spin character in the sample as is the case for [Fe(dftpy)2]2+ at room temperature (97 %) because of a suppression of the 3+/2+ couple due to the orbital origins of the oxidation event: nominally eg* in high-spin systems rather than t2g expected for low-spin complexes. With this in mind, we can understand the blue shift of the MLCT in low-spin populations of [Fe(dftpy)2]2+ versus [Fe(tpy)2]2+ as being driven by the orbital origins of the electronic transition. In addition to having π-donor character, the fluorine substituents are σ-withdrawing. This reduces electron density at the metal center and in turn makes the formal oxidation of the metal center that accompanies MLCT more positive.

15



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 33

The high-spin basis spectrum for [Fe(dftpy)2]2+ is shown in Figure S3 along with molar absorptivities measured for the larger members of the halogen series [Fe(dctpy)2]2+ and [Fe(dbtpy)2]2+. These data highlight a trend of increasing absorptivity for the broad shoulder (~ 500 nm) on going from a bromine substituent to a chlorine and finally to fluorine. We can rationalize this in terms of how halogen size impacts metal ligand distances across the series (in particular Fe-Nterminal), which then affects 5MLCT ← 5GS intensity through orbital overlap (see Tables 1 and 2). The broadness of the shoulder precludes determination of 5MLCT energies and whether there is any systematic variation. However, the similarity of band shapes does suggest that 5MLCT energy shifts within the series must be subtle. Structural Data for [Fe(dftpy)2](BF4)2. As a final exploration of SCO in [Fe(dftpy)2]2+, we have directly explored structural changes that accompany temperature variation using X-ray diffraction. Two data sets were collected for a single crystal at two different temperatures: 120 K and 300 K. The higher temperature data set was collected first, followed by slow cooling (60 K/hour) to 120 K for collection of the second data set. In previous experimental attempts that reversed this order (first data collection at low temperature followed by raising the temperature) we observed crystal fracturing upon warming. Accompanying the temperature change was a significant color change from light red to a deep purple, consistent with the solution-phase temperature-dependent absorption spectra seen in Figure 3. In addition to the color change, we observed a change in the geometric parameters of the unit cell. As can be seen in Table S1, while the space group remains as P21/c at both temperatures, there is a volume decrease from 3425.9(5) to 3212.2(5) Å3 upon cooling the crystal driven by a contraction from 10.9455(9) to 10.1127(9) Å along one axis. Table 2. Temperature-dependent structural data for [Fe(dftpy)2](BF4)2 T (K)

Fe-Ncentral (Å)

Fe-Nterminal (Å)

∠ N-N-N (°)

Θ (°)

X-pyridyl (Å)

120

1.904(2)

2.027(1)

103.86(8)

1.1(1)

2.680(2)

300

2.116(1)

2.206(1)

108.80(7)

3.89(9)

2.848(2)

Superimposed ball and stick plots for [Fe(dftpy)2]2+ at both temperatures are shown in Figure 6 to highlight structural differences. While the viewpoint focuses on a single dftpy ligand and its metal coordination environment, comparable changes are evident for the second ligand. The first striking differences are the elongation of the three Fe–N bonds and the increase in the N-N-N 16


Page 17 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


angle as the temperature is changed from 120 K to 300 K. These are, of course, hallmarks of a low-spin to high-spin transition. Examining the structures quantitatively (Table 2), we see that upon conversion to the high-spin structure, Fe-Ncentral increases by 0.212(2) Å while Fe-Nterminal increases by 0.179(1) Å, consistent with a weakening of the ligand field and population of metalligand antibonding orbitals. This weakening of the metal-ligand bonds allows the N-N-N angle to relax into a more obtuse conformation by 4.94(5)°. The combination of these two motions allows the fluorine atoms to shift 0.168(1) Å farther from the central pyridine on the other ligand, yet in both configurations, this distance is smaller than the sum of the Bondi radii of the two species (3.17 Å for C and F).

Figure 6. Overlaid crystal structures of the Fe-containing complex in [Fe(dftpy)2]2+ collected at 300 K and 120 K (semi-transparent). Both structures are plotted as ball-and-stick models due to the thermal ellipsoid differences at the two temperatures. C = gray, F = purple, Fe = orange, N = blue, H = white. Excited-state dynamics. Having established an understanding of ground-state electronic properties within this series of Fe(II) complexes using UV-Vis spectroscopy, magnetic susceptibility measurements, and single crystal x-ray diffraction, we turn to time-resolved absorption spectroscopies to characterize excited-state dynamics including ground-state recovery. Our starting point is the complex with the largest halogen in the series, [Fe(dbtpy)2]2+, whose ground state at room temperature is the quintet, much like what is seen for the previously studied [Fe(dctpy)2]2+.28 Transient absorption spectral data resolved on a picosecond time scale 17



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 33

are shown in Figure 7 (left) following excitation with laser pulses centered at 400 nm and with a duration of ~ 150 fs. Reiterating, this excitation wavelength accesses a broad tail in the absorption spectrum (Figure 1) that is expected to be associated with 5MLCT ← 5GS excitation. As may be seen in the 2D intensity plot of ∆A versus time (Figure 7, top left), as well as in spectral slices at specific values of ∆t (Figure 7, bottom left), there are two main features: a bright, more sharply-peaked UV absorption (in the vicinity of 370 nm) and a weaker, significantly broader visible absorption. There is evidence for subtle spectral shifting in the UV feature at early times (< 4 ps; vide infra), but the dominant character of the data is decay of both UV and visible features to baseline on a 10s of picoseconds time scale. We emphasize that all spectral and temporal features bear a close resemblance to what has been previously reported for the chlorinated analogue [Fe(dctpy)2]2+.37 In that work, with strong support from computational and spectroelectrochemical experiments, we concluded that the UV and visible TA features have redox origins and can be attributed to the formally reduced ligand and oxidized metal center of a high-spin

5,7

MLCT state. The thermalization of this state is the

origin of the early-time spectral shifting and the decay of this state on a longer time scale heralds recovery of the high-spin ground state with no evidence for population of an intermediate. Given the high degree of similarity between the data collected and shown within this current work for [Fe(dbtpy)2]2+ (Figure 7, left) and the data previously presented for [Fe(dctpy)2]2+,37 we are confident in drawing the same mechanistic conclusion. Details about the kinetic analysis will be presented below, after introducing TA results for [Fe(dftpy)2]2+.

18


Page 19 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 7. TA spectroscopy of [Fe(dbtpy)2]2+ (left) and [Fe(dftpy)2]2+ (right) in room temperature acetonitrile following 400 nm excitation. Top: full 2D surface plot of data. Bottom: selected spectra (pump scatter is not shown; data are interpolated between 392 nm and 405 nm). TA spectral data for [Fe(dftpy)2]2+ collected with similar experimental conditions are shown in Figure 7 (right) and exhibit markedly different behavior than what is seen for [Fe(dbtpy)2]2+ or previously for [Fe(dctpy)2]2+.37 Most notable is a long-lived bleach feature centered at 504 nm that is invariant over the 50 ps timescale of the experiment and whose shape closely matches the inverse of the absorption feature that grows in for [Fe(dftpy)2]2+ as the sample is cooled. These observations are consistent with TA studies of low spin Fe(II) polypyridyl systems where

19



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 33

evolution from the 1MLCT to 5MC is ultrafast14-16 (see also the discussion below where the laser excitation wavelength is shifted to the red). At the same time, however, there are UV and visible features flanking the bleach, whose kinetic behavior and shape bear strong resemblance to features attributed to the

5,7

MLCT in [Fe(dbtpy)2]2+ and [Fe(dctpy)2]2+ within similar spectral

regions. Given our findings (vide supra) that room temperature samples of [Fe(dftpy)2]2+ consist of a 3:97 admixture of low- and high-spin states, it appears clear that 400 nm pulses excite both states. In another Fe(II) spin crossover system, [Fe(2-Me-phen)3]2+, Gallé et al. selectively excited the low- or high-spin MLCT state using a blue or near-UV pump, respectively.64 They propose a bifurcating kinetic model following excitation into the 5MLCT that results in subpicosecond formation of low energy 1A and 5T states, with the transient population of the singlet persisting for several nanoseconds.65 We are able to rule out such a mechanism in high-spin excitation of [Fe(dftpy)2]2+ due to our observation of the near-UV feature that matches what is seen in [Fe(dctpy)2]2+ and [Fe(dbtpy)2]2+, which is diagnostic of an MLCT state. Further, there is no evidence of early decay of the 1MLCT←1A bleach as would be expected if 5MLCT excitation was followed by transient population of the singlet state. Despite this result, low-T measurements would be needed to assess the potential for reverse-LIESST40 using this molecule. Attempting a similar selective excitation experiment of our own, we shifted the pump wavelength in our TA spectrometer to 515 nm. At this wavelength, the molar extinction of the low spin state in [Fe(dftpy)2]2+ is 35 times larger than that of the high-spin state at room temperature. Four transient spectra are shown in Figure S7 collected at times ranging from 6 ps – 730 ps. Again the same bleach feature mentioned above is seen peaked at 504 nm, but it is now the dominant feature and its magnitude is unchanged over the time delays sampled. Any component due to excitation in the high-spin manifold is within the noise in less than 6 ps. In addition to the bleach of the 1MLCT ← 1A absorption, there is a small excited state absorption in the near-UV. We attribute this to increased LLCT in this spectral region due to the transient increase in the 5MC population. A nanosecond transient absorption experiment was used to determine the lifetime of the bleach feature. At room temperature, the lifetime is 22 ns as can be seen in Figure 8, which represents a substantial elongation compared to the reported value of 5.35 ns for [Fe(tpy)2]2+ under comparable conditions.31 In order to better understand this lifetime increase via the interplay between low-energy singlet and quintet surfaces in [Fe(dftpy)2]2+, the temperature

20


Page 21 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


dependence of bleach recovery was measured over the range of 279 to 316 K. As can be seen in Figure 8, these data are well modeled with an Arrhenius expression and reveal an activation barrier of 34 kJ/mol for the inter-conversion between 5MC and 1A states. This barrier is substantial when compared to the literature value of 6.4 kJ/mol measured for [Fe(tpy)2]2+.66 There are two likely contributing factors. First, inter-ligand steric interactions exacerbated by the halogen substituents in [Fe(dftpy)2]2+ may be expected to increase the reorganization energy in going from a geometry relevant for the 5MC to one relevant for the 1A. Second, the 5MC/1A inter-conversion driving force is reduced (the origin of the near room temperature spin crossover behavior) and this would lead to an increase in the activation barrier even under circumstances where the reorganization energy is comparable. It is noted that the pre-exponential determined from the Arrhenius analysis (A = 3.6×1013 s-1) is surprisingly large for a direct conversion between a quintet and a singlet, suggesting that intermediate spin states may be involved in the rate-limiting step.

21



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 33

Figure 8. Top: Arrhenius plot of [Fe(dftpy)2]2+ fit to an activation barrier of 34 ± 3 kJ/mol and a natural log of the pre-exponential of 31.2 ± 1.3. Bottom: Single-wavelength kinetics at 510 nm following 520 nm excitation. Lines indicate fits to Gaussian-convoluted single exponential decay. In this final section, we address the kinetic behavior of the full halogen series and ask whether trends emerge as the ligand substituent is varied. Within the series, the two largest species [Fe(dctpy)2]2+ and [Fe(dbtpy)2]2+ are again the simplest to analyze. As was discussed in our previous work, the transient spectral data for [Fe(dctpy)2]2+ (362 nm – 610 nm) for ∆t = 1 ps and beyond may be cleanly reproduced using a global kinetic model inclusive of two exponential components with time constants τ1 = 3.6 ps and τ2 = 16.0 ps. The same is true for [Fe(dbtpy)2]2+ 22


Page 23 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


leading to a determination of τ1 = 3.7 ps and τ2 = 17.4 ps. The high quality of the twoexponential model may be seen in kinetic slices taken from the full spectral data and shown in Figure S8. The transient spectral data for [Fe(dftpy)2]2+ may be similarly modeled with the caveat that the 400 nm laser excitation also accesses some of the singlet ground state population in this spin crossover system. This simply means that the kinetic expression necessary to globally model the data is inclusive of a third exponential component whose decay is fixed to the 22 ns that were measured for the compound at room temperature using the 515 nm excitation. With this in place, τ1 and τ2 are measured for [Fe(dftpy)2]2+ at 3.2 ps and 14.0 ps, respectively (see Figure S8 for representative kinetic slices modeled this way). All measured time constants are shown in Table 3 with values and reported error based on three independent sets of data collected for each species in the halogen series.

Table 3. Global Fit Lifetimes collected in Room Temperature Acetonitrile. complex [Fe(dftpy)2]2+ [Fe(dctpy)2]2+ [Fe(dbtpy)2]2+

τ1 / ps (2σ)a 3.2 (1.0) 3.6 (0.6) 3.7 (1.0)

τ 2 / ps (2σ)a 14.0 (1.5) 16.0 (1.0) 17.4 (2.0)

a

Average tau values and standard deviations calculated from the independent data sets based on a global fit of the full spectral range beginning 1 ps after excitation

As previously discussed, the faster time component (τ1) in these molecules reflects thermalization within the

5,7

MLCT while the slower component (τ2) reflects the state’s lifetime

prior to decay to the high-spin ground state. These conclusions are supported using τ1 and τ2 basis spectra generated from the respective global fits, focusing on features in the near-UV (Figure S9). For example, the τ2 spectrum for each molecule in the series points to the excited state absorption of the thermalized 5,7MLCT state corresponding to π* ← π* of the ligand-based radical anion as is seen in other d6 terpyridyl and bipyridyl MLCT complexes.35,67,68 The systematic blue shift in this feature across the halogen series, from 370 nm for [Fe(dbtpy)2]2+ to 365 nm for [Fe(dctpy)2]2+ to 355 nm for [Fe(dftpy)2]2+, can be qualitatively understood in the context of calculated HOMO/LUMO gaps within [Zn(dftpy)2]1+, [Zn(dctpy)2]1+, and [Zn(dbtpy)2]1+ which are used to model ligand-based radical anion π* ← π* transitions (see Table S5).

23



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 33

The τ1 basis spectrum for each species resembles the derivative of the respective τ2 spectrum. This has the effect of blue-shifting and narrowing the band in the first few ps following excitation. Similar observations were made for the early evolution of [Fe(CN)4(bpy)]2- 35 and [Os(bpy)3]2+ 69 and attributed to vibrational redistribution and cooling. In the latter system, Papanikolas and coworkers observed the shift continuing beyond the redistributional timescale, so we feel confident attributing this to vibrational cooling of the excited state. There is some evidence in Table 3 for a systematic trend in cooling within our halogen series, where the largest member [Fe(dbtpy)2]2+ is slowest and where the smallest member [Fe(dftpy)2]2+ is fastest. This trend is corroborated by fits of λmax versus time (Figure S10). This halogen-series cooling trend could be explained if final dissipation of vibrational energy occurs along low-frequency modes such as metal-ligand bond stretching coordinates and/or ligand distortions and where the weakest metal-ligand bond of the series (in [Fe(dbtpy)2]2+) and/or the lowest-frequency distortion (in dbtpy) is more poorly coupled to the solvent bath. However, the series effect is small and a more detailed analysis (exploring, for example, dynamics in different solvent isotopomers) would be needed to characterize the effect. Finally, we can use the derivative shape of the τ1 basis spectra within the halogen series (in particular their zero-valued-crossing) to determine the wavelength at which it is possible to isolate τ2 dynamics corresponding to

5,7

MLCT loss and high-spin ground state recovery. In

Figure 9 we plot on a log scale the normalized ∆A decay for each molecule isolated from the corresponding 2D data sets at 369 nm for [Fe(dbtpy)2]2+, at 366 nm for [Fe(dctpy)2]2+, and at 356 nm for [Fe(dftpy)2]2+. Clearly from the fits, the isolation of τ2 dynamics is reasonable. Most importantly, the data indicate a trend across the halogen series where the

5,7

MLCT lifetime

increases with halogen size. Given the picosecond lifetimes of these species, we do not expect a direct transition between

5,7

MLCT and 5MC. These states should be nested (both have

configurations with population of two metal-ligand antibonding orbitals) and firmly in the Marcus inverted regime given the ~ 2.5 eV driving force. Thus, it appears unlikely that potential variations in reaction free energy as the halogen changes is the origin of the τ2 trend. An interpretation we favor is that the observation supports the original design strategy where larger halogen substituents are impacting surface crossings between states during non-radiative decay. We anticipate that the 3MLCT (with no occupancy of metal-ligand antibonding orbitals) is not of importance given its significant conformational difference. On the other hand, if the decay 24


Page 25 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


pathway from the

5,7

MLCT to the 5MC within which it is nested is facilitated by transient

population of a triplet with a configuration that now has one metal-ligand antibonding electron, such as the 3MC,70 it would be reasonable to encounter slower non-radiative decay in cases where there is a more significant energy penalty (i.e., a higher reorganization energy) upon decreasing Fe–N bond distances, as would be the case for larger halogen atoms. Further, we strongly suspect that the energy of the 3MC relative to the 5,7MLCT is impacted by the size of the halogen substituent. This idea is informed by the observation of SCO in the smaller [Fe(dftpy)2]2+ and not in the larger species. As the halogen substituent increases in size, those states with more metal-ligand bonding character – such as 1A that we observe through SCO but also 3MC where there is only a single electron in an antibonding ligand field orbital – are destabilized according to interligand steric interactions relative to states bearing configurations with two electrons in antibonding ligand field orbitals (5MC and

5,7

MLCT). One would then

encounter slower non-radiative decay in the limit that this increases the barrier for crossing between 5,7MLCT and 3MC.71 Both the reorganization energy and the state energy shift ideas are presented as a cartoon in Figure 10. Of note, it would not be unreasonable for both processes to contribute to changes in decay rates by impacting a rate-limiting crossing between 5,7MLCT and 3

MC (see the barrier increase highlighted within the circle of Figure 10). Alternatively, it would

not be impossible for the two effects to oppose each other, and this will ultimately depend on the true (as opposed to hypothetical) placement of the relevant states in an energetic and conformational landscape. In this context, theory is needed to gain further insight. As a final point, it should be emphasized that while steric effects do appear to impact relaxation rate constants, the overall effect in this halogen series is subtle and appears to have come close to reaching its limit. There is a 24% increase in lifetime from [Fe(dftpy)2]2+ to [Fe(dbtpy)2]2+ (Table 3) but for the last two members, from [Fe(dctpy)2]2+ to [Fe(dbtpy)2]2+, the increase is only 9%. Overall, this suggests that new strategies are needed beyond halogen steric effects in order to realize significant additional gains in high-spin MLCT lifetime.

25



: 14.1 ps : 16.7 ps : 18.1 ps

A

10 0

Normalized

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 33

10 -1

5

10

15

20

25

30

35

Time Delay / ps

Figure 9. Single-exponential fit at the peak wavelength of the thermalized TA spectrum: [Fe(dftpy)2]2+ (blue) at 356 nm, [Fe(dctpy)2]2+ (green) at 366 nm, and [Fe(dbtpy)2]2+ (red) at 369 nm. In the Fe(dftpy)2]2+ data, the positive feature due to the τ3 component has been subtracted.

Figure 10. Cartoon of potential energy surfaces and orbital occupancy diagrams for states used to describe the dynamics in the series of Fe(II) dihaloterpyridine complexes. The reorganization energy argument is captured by the increase in curvature (grey dashed lines) as steric hindrance is increased, while the 3MC energy argument is captured by the increase in the energy of the potential energy curve (red dashed lines) as steric hindrance is increased. The grey circle indicates the significant surface crossing (the rate-limiting step) in determining non-radiative decay of the MLCT state.

26


Page 27 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Conclusions X-ray diffraction, magnetic measurements, and static and time-resolved optical spectroscopies were used to characterize molecules and explore systematic variations within a series of Fe(II) bis-6,6′′-dihaloterpyrdine complexes, [Fe(dbtpy)2]2+, [Fe(dctpy)2]2+, and [Fe(dftpy)2]2+, that incorporate bromine, chlorine, and fluorine substituents, respectively. For the larger members of the series, [Fe(dbtpy)2]2+ and [Fe(dctpy)2]2+, structural and magnetic susceptibility data provided strong evidence for a high-spin quintet ground state. Optical absorption measurements indicate that the lowest energy feature is a broad visible shoulder that has been assigned to 5MLCT ← 5GS. Although both molecules appear to be high spin with very similar absorption features and band energies, a modest increase in molar extinction is observed from [Fe(dbtpy)2]2+ to [Fe(dctpy)2]2+, consistent with an enhancement in metal-ligand electronic interaction as the halogen atom size is decreased from Br to Cl. When the halogen size decreases further in [Fe(dftpy)2]2+, complexity emerges in the structural, optical, and magnetic data that manifests in significant temperature dependence. Notably, a spin crossover equilibrium is uncovered where the singlet state, 1A, is enthalpically favored and where the quintet state, 5MC, is populated due to entropic contributions to the free energy. At room temperature, a sample of [Fe(dftpy)2]2+ is 97:3 of quintet to singlet. Variations in optical features within this spin equilibrium are useful for disentangling excited state dynamics attributable to excitation origins in two different ground state populations. For example, using a redder pump wavelength (520 nm) the singlet manifold may be selectively addressed (1MLCT ← 1A), ultimately leading to transient excess in the 5MC population. In reestablishing the 1A/5MC spin equilibrium, the time constant for 5MC → 1A is measured to be 22 ns at room temperature, a value that is four times longer than what is observed for the parent molecule [Fe(tpy)2]2+. Turning to bluer excitation (400 nm), the quintet manifold in [Fe(dftpy)2]2+ may be selectively addressed (5MLCT ← 5MC), an experiment that is useful because it enables measurement of the

5,7

MLCT lifetime needed for comparison. Across the

halogen series as a whole, we observe an increase in high-spin MLCT lifetime from 14.0 ps ([Fe(dftpy)2]2+) to 16.0 ps ([Fe(dctpy)2]2+) to 17.4 ps ([Fe(dbtpy)2]2+). On one hand these values are significant in magnitude, indicative of some of the longest lived MLCT states known in Fe(II) complexes. The 17.4 ps measured for [Fe(dbtpy)2]2+ is the longest known lifetime for a high-spin (quintet or septet) MLCT and to our knowledge is exceeded only by the 18 ps33 and 26

27



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 33

ps32 3MLCT lifetimes that were recently measured for high ligand-field N-heterocyclic carbene complexes of Fe(II) and the 20 ps MLCT lifetime measured for [Fe(CN)4(bpy)]2-.35 On the other hand, these lifetimes are short enough to argue against direct nonradiative passage between

5,7

MLCT and the ground state 5MC. Both states have configurations with two

electrons in metal-ligand antibonding orbitals with the expected consequence of being nested and unable to interconvert on sub-nanosecond timescales given the relaxation driving force. These lifetimes then suggest that the decay pathway involves participation by additional states. Triplets with a single electron in a metal-ligand antibonding orbital, such as the 3MC, are likely candidates. Certainly the relative energy of participating states could play a role in our observation of lifetime control in the halogen series. Judging from comparative absorption spectroscopy there is little reason to believe that the

5,7

MLCT energy is significantly perturbed

relative to the 5MC as substitution varies from F to Cl to Br. On the other hand, the 3MC, with shorter metal-ligand bond lengths, is expected to be sensitive to the structural and steric differences explored in the halogen series. If the rate-limiting barrier to interconversion from 5,7

MLCT to 3MC increases with the size of the halogen substituent, the lifetime variation would

track our observations of increasing from 14.0 ps to 17.4 ps on going from [Fe(dftpy)2]2+ to [Fe(dbtpy)2]2+. A complementary mechanism involves variations in reorganization energy. The larger the halogen atom, the more difficult it is energetically for the 5,7MLCT to adopt geometries needed for conversion to the 3MC in its pathway to the ground state 5MC. These ideas are at this stage based on observations tied to structural modifications and will require additional support, presumably through theory and subsequent measurements of temperature dependent lifetimes. A final conclusion to be made is that improvements to Fe(II)

5,7

MLCT lifetimes, with

halogen substitution at 6 and 6′′ positions of terpyridine, appear to be approaching a limit with bromine. At this stage a change in design strategy is needed. We believe that it should take advantage of the structural control principles that have been elucidated, perhaps even seeking more rigidity through substituents that make multiple points of contact (per substituent) with the opposing ligand. At the same time, new strategies should seek to change the energetic relationship between states involved in nonradiative decay and they should seek to reduce the coupling between MLCT and metal-centered states using electronic perturbations. Efforts along these lines are currently underway.

28


Page 29 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


ASSOCIATED CONTENT Supporting Information Additional UV-vis spectra, magnetic susceptibility data, X-ray diffraction data, NMR spectra, TA spectroscopy data, electrochemical data, and computational results. The crystallographic data is also available as a CIF file. This material is available free of charge via the Internet at http://pubs.acs.org. AUTHOR INFORMATION Corresponding Author *[email protected] Author Contributions ‡These authors contributed equally. ORCID Steven M. Fatur: 0000-0003-2115-6277 Samuel G. Shepard: 0000-0002-6783-5403 Robert F. Higgins: 0000-0001-6776-3537 Notes The authors declare no competing financial interests. ACKNOWLEDGMENT This work was supported by the NSF and EPA through the Catalysis Collaboratory for LightActivated Earth Abundant Reagents (C-CLEAR), a Network for Sustainable Molecular Design and Synthesis (CHE-1339674). This work utilized the Janus supercomputer, which is supported by the National Science Foundation (award number CNS-0821794) and the University of Colorado Boulder. The Janus supercomputer is a joint effort of the University of Colorado Boulder, the University of Colorado Denver and the National Center for Atmospheric Research. We would also like to thank Professor Tarek Sammakia for useful discussions regarding Hammett parameters.

29



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 33

REFERENCES (1) (2) (3) (4) (5)

Prier, C. K.; Rankic, D. A.; MacMillan, D. W. C. Chem. Rev. 2013, 113, 5322-5363. Yoon, T. P.; Ischay, M. A.; Du, J. Nat. Chem. 2010, 2, 527-532. Shaw, M. H.; Twilton, J.; MacMillan, D. W. C. J. Org. Chem. 2016, 81, 6898-6926. Lin, S.; Ischay, M. A.; Fry, C. G.; Yoon, T. P. J. Am. Chem. Soc. 2011, 133, 19350-19353. Yayla, H. G.; Peng, F.; Mangion, I. K.; McLaughlin, M.; Campeau, L.-C.; Davies, I. W.; DiRocco, D. A.; Knowles, R. R. Chem. Sci. 2016, 7, 2066-2073. (6) Huo, H.; Shen, X.; Wang, C.; Zhang, L.; Röse, P.; Chen, L.-A.; Harms, K.; Marsch, M.; Hilt, G.; Meggers, E. Nature 2014, 515, 100-103. (7) Kainz, Q. M.; Matier, C. D.; Bartoszewicz, A.; Zultanski, S. L.; Peters, J. C.; Fu, G. C. Science 2016, 351, 681684. (8) Pirtsch, M.; Paria, S.; Matsuno, T.; Isobe, H.; Reiser, O. Chem. Eur. J. 2012, 18, 7336-7340. (9) Stevenson, S. M.; Shores, M. P.; Ferreira, E. M. Angew. Chem., Int. Ed. 2015, 54, 6506-6510. (10) Stevenson, S. M.; Higgins, R. F.; Shores, M. P.; Ferreira, E. M. Chem. Sci. 2017, 8, 654-660. (11) Higgins, R. F.; Fatur, S. M.; Shepard, S. G.; Stevenson, S. M.; Boston, D. J.; Ferreira, E. M.; Damrauer, N. H.; Rappé, A. K.; Shores, M. P. J. Am. Chem. Soc. 2016, 138, 5451-5464. (12) Zhang, Y.; Petersen, J. L.; Milsmann, C. J. Am. Chem. Soc. 2016, 138, 13115-13118. (13) Büldt, L. A.; Guo, X.; Prescimone, A.; Wenger, O. S. Angew. Chem., Int. Ed. 2016, 55, 11247-11250. (14) McCusker, J. K.; Walda, K. N.; Dunn, R. C.; Simon, J. D.; Magde, D.; Hendrickson, D. N. J. Am. Chem. Soc. 1992, 114, 6919-6920. (15) Cannizzo, A.; Milne, C. J.; Consani, C.; Gawelda, W.; Bressler, C.; van Mourik, F.; Chergui, M. Coord. Chem. Rev. 2010, 254, 2677-2686. (16) Aubock, G.; Chergui, M. Nat. Chem. 2015, 7, 629-633. (17) Gualandi, A.; Marchini, M.; Mengozzi, L.; Natali, M.; Lucarini, M.; Ceroni, P.; Cozzi, P. G. ACS Catal. 2015, 5, 5927-5931. (18) Zhang, J.; Campolo, D.; Dumur, F.; Xiao, P.; Fouassier, J. P.; Gigmes, D.; Lalevée, J. J. Polym. Sci. A Polym. Chem. 2016, 54, 2247-2253. (19) Parisien-Collette, S.; Hernandez-Perez, A. C.; Collins, S. K. Org. Lett. 2016, 18, 4994-4997. (20) McCusker, J. K.; Walda, K. N.; Dunn, R. C.; Simon, J. D.; Magde, D.; Hendrickson, D. N. J. Am. Chem. Soc. 1993, 115, 298-307. (21) Monat, J. E.; McCusker, J. K. J. Am. Chem. Soc. 2000, 122, 4092-4097. (22) Zhang, W.; Alonso-Mori, R.; Bergmann, U.; Bressler, C.; 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, 345-348. (23) Cammarata, M.; Bertoni, R.; Lorenc, M.; Cailleau, H.; Di Matteo, S.; Mauriac, C.; Matar, S. F.; Lemke, H.; Chollet, M.; Ravy, S.; Laulhé, C.; Létard, J.-F.; Collet, E. Phys. Rev. Lett. 2014, 113, 227402. (24) Sousa, C.; de Graaf, C.; Rudavskyi, A.; Broer, R.; Tatchen, J.; Etinski, M.; Marian, C. M. Chem. Eur. J. 2013, 19, 17541-17551. (25) Moguilevski, A.; Wilke, M.; Grell, G.; Bokarev, S. I.; Aziz, S. G.; Engel, N.; Raheem, A. A.; Kuhn, O.; Kiyan, I. Y.; Aziz, E. F. ChemPhysChem 2017, in press. (26) Nance, J.; Bowman, D. N.; Mukherjee, S.; Kelley, C. T.; Jakubikova, E. Inorg. Chem. 2015, 54, 11259-11268. (27) Pápai, M.; Penfold, T. J.; Møller, K. B. J. Phys. Chem. C 2016, 120, 17234-17241. (28) Dixon, I. M.; Alary, F.; Boggio-Pasqua, M.; Heully, J.-L. Inorg. Chem. 2013, 52, 13369-13374. (29) Dixon, I. M.; Boissard, G.; Whyte, H.; Alary, F.; Heully, J.-L. Inorg. Chem. 2016, 55, 5089-5091. (30) Bowman, D. N.; Bondarev, A.; Mukherjee, S.; Jakubikova, E. Inorg. Chem. 2015, 54, 8786-8793. (31) Jamula, L. L.; Brown, A. M.; Guo, D.; McCusker, J. K. Inorg. Chem. 2014, 53, 15-17. (32) Liu, L.; Duchanois, T.; Etienne, T.; Monari, A.; Beley, M.; Assfeld, X.; Haacke, S.; Gros, P. C. Phys. Chem. Chem. Phys. 2016, 18, 12550-12556. (33) Harlang, T. C. B.; Liu, Y.; Gordivska, O.; Fredin, L. A.; Ponseca, C. S., Jr.; Huang, P.; Chábera, P.; Kjær, K. S.; Mateos, H.; Uhlig, J.; Lomoth, R.; Wallenberg, R.; Styring, S.; Persson, P.; Sundström, V.; Wärnmark, K. Nat. Chem. 2015, 7, 883-889. (34) Pastore, M.; Duchanois, T.; Liu, L.; Monari, A.; Assfeld, X.; Haacke, S.; Gros, P. C. Phys. Chem. Chem. Phys. 2016, 18, 28069-28081.

30


Page 31 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(35) 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, 515-523. (36) Büldt, L. A.; Guo, X.; Vogel, R.; Prescimone, A.; Wenger, O. S. J. Am. Chem. Soc. 2017, 139, 985-992. (37) Shepard, S. G.; Fatur, S. M.; Rappé, A. K.; Damrauer, N. H. J. Am. Chem. Soc. 2016, 138, 2949-2952. (38) Gütlich, P.; Hauser, A.; Spiering, H. Angew. Chem.-Int. Edit. Engl. 1994, 33, 2024-2054. (39) Létard, J.-F. J. Mater. Chem. 2006, 16, 2550-2559. (40) Marino, A.; Chakraborty, P.; Servol, M.; Lorenc, M.; Collet, E.; Hauser, A. Angew. Chem., Int. Ed. 2014, 53, 3863-3867. (41) Vallett, P. J.; Damrauer, N. H. J. Phys. Chem. A 2013, 117, 6489-6507. (42) Dolomanov, O. V.; Bourhis, L. J.; Gildea, R. J.; Howard, J. A. K.; Puschmann, H. J. Appl. Crystallogr. 2009, 42, 339-341. (43) Sheldrick, G. M. Acta Cryst. A 2007, 64, 112-122. (44) Bain, G. A.; Berry, J. F. J. Chem. Educ. 2008, 85, 532-536. (45) Shao, Y. H.; Gan, Z. T.; Epifanovsky, E.; Gilbert, A. T. B.; Wormit, M.; Kussmann, J.; Lange, A. W.; Behn, A.; Deng, J.; Feng, X. T.; Ghosh, D.; Goldey, M.; Horn, P. R.; Jacobson, L. D.; Kaliman, I.; Khaliullin, R. Z.; Kuś, T.; Landau, A.; Liu, J.; Proynov, E. I.; Rhee, Y. M.; Richard, R. M.; Rohrdanz, M. A.; Steele, R. P.; Sundstrom, E. J.; Woodcock, H. L.; Zimmerman, P. M.; Zuev, D.; Albrecht, B.; Alguire, E.; Austin, B.; Beran, G. J. O.; Bernard, Y. A.; Berquist, E.; Brandhorst, K.; Bravaya, K. B.; Brown, S. T.; Casanova, D.; Chang, C. M.; Chen, Y. Q.; Chien, S. H.; Closser, K. D.; Crittenden, D. L.; Diedenhofen, M.; DiStasio, R. A.; Do, H.; Dutoi, A. D.; Edgar, R. G.; Fatehi, S.; Fusti-Molnar, L.; Ghysels, A.; Golubeva-Zadorozhnaya, A.; Gomes, J.; Hanson-Heine, M. W. D.; Harbach, P. H. P.; Hauser, A. W.; Hohenstein, E. G.; Holden, Z. C.; Jagau, T. C.; Ji, H. J.; Kaduk, B.; Khistyaev, K.; Kim, J.; Kim, J.; King, R. A.; Klunzinger, P.; Kosenkov, D.; Kowalczyk, T.; Krauter, C. M.; Lao, K. U.; Laurent, A. D.; Lawler, K. V.; Levchenko, S. V.; Lin, C. Y.; Liu, F.; Livshits, E.; Lochan, R. C.; Luenser, A.; Manohar, P.; Manzer, S. F.; Mao, S. P.; Mardirossian, N.; Marenich, A. V.; Maurer, S. A.; Mayhall, N. J.; Neuscamman, E.; Oana, C. M.; Olivares-Amaya, R.; O'Neill, D. P.; Parkhill, J. A.; Perrine, T. M.; Peverati, R.; Prociuk, A.; Rehn, D. R.; Rosta, E.; Russ, N. J.; Sharada, S. M.; Sharma, S.; Small, D. W.; Sodt, A.; Stein, T.; Stuck, D.; Su, Y. C.; Thom, A. J. W.; Tsuchimochi, T.; Vanovschi, V.; Vogt, L.; Vydrov, O.; Wang, T.; Watson, M. A.; Wenzel, J.; White, A.; Williams, C. F.; Yang, J.; Yeganeh, S.; Yost, S. R.; You, Z. Q.; Zhang, I. Y.; Zhang, X.; Zhao, Y.; Brooks, B. R.; Chan, G. K. L.; Chipman, D. M.; Cramer, C. J.; Goddard, W. A.; Gordon, M. S.; Hehre, W. J.; Klamt, A.; Schaefer, H. F.; Schmidt, M. W.; Sherrill, C. D.; Truhlar, D. G.; Warshel, A.; Xu, X.; Aspuru-Guzik, A.; Baer, R.; Bell, A. T.; Besley, N. A.; Chai, J. D.; Dreuw, A.; Dunietz, B. D.; Furlani, T. R.; Gwaltney, S. R.; Hsu, C. P.; Jung, Y. S.; Kong, J.; Lambrecht, D. S.; Liang, W. Z.; Ochsenfeld, C.; Rassolov, V. A.; Slipchenko, L. V.; Subotnik, J. E.; Van Voorhis, T.; Herbert, J. M.; Krylov, A. I.; Gill, P. M. W.; Head-Gordon, M. Mol. Phys. 2015, 113, 184-215. (46) Bowman, D. N.; Jakubikova, E. Inorg. Chem. 2012, 51, 6011-6019. (47) Braterman, P. S.; Song, J. I.; Peacock, R. D. Inorg. Chem. 1992, 31, 555-559. (48) Baker, A. T.; Goodwin, H. A. Aust. J. Chem. 1985, 38, 207-214. (49) Bondi, A. J. Phys. Chem. 1964, 68, 441-451. (50) Rowland, R. S.; Taylor, R. J. Phys. Chem. 1996, 100, 7384-7391. (51) Crews, P.; Rodríguez, J.; Jaspars, M. Organic structure analysis; 2nd ed.; Oxford University Press: New York, 2010. (52) The difference between the Hammett parameter, σ, for meta versus para substitution can be used to isolate resonance from inductive effects. The Hammett parameter of a particular atom or group measures the relative change of acidity when this substituent is placed on benzoic acid. More specifically, σ is the logarithm of the ratio of the acid dissociation constants of a substituted and unsubstituted benzoic acid. The electronegativity of the halogen substituents increases the acidity of the benzoic acid by stabilizing the conjugate base, as evidenced by the positive value of σ. Halogens are also resonance-donating, however only when one is placed ortho or para to the benzoic acid is it able to donate pi electron density to the carboxylic acid, reducing acidity. Assuming inductive effects are comparable for meta and para substitution, the difference between Hammett parameters for meta and para substitution subtracts the inductive contribution, leaving only resonance effects. The change in the Hammett parameter is more pronounced with fluorine than it is with chlorine, so fluorine is the better pi donor. (53) Ritchie, C. D.; Sager, W. F. In Progress in Physical Organic Chemistry; John Wiley & Sons, Inc.: New York, 2007; Vol. 2, p 323-400.

31



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 32 of 33

(54) Constable, E. C.; Baum, G.; Bill, E.; Dyson, R.; van Eldik, R.; Fenske, D.; Kaderli, S.; Morris, D.; Neubrand, A.; Neuburger, M.; Smith, D. R.; Wieghardt, K.; Zehnder, M.; Zuberbühler, A. D. Chem. Eur. J. 1999, 5, 498508. (55) Kahn, O. Molecular magnetism; VCH: New York, NY, 1993. (56) Fryzuk, M. D.; Leznoff, D. B.; Ma, E. S. F.; Rettig, S. J.; Young, V. G. Organometallics 1998, 17, 2313-2323. (57) Schweinfurth, D.; Demeshko, S.; Sommer, M. G.; Dechert, S.; Meyer, F.; Sarkar, B. Eur. J. Inorg. Chem. 2016, 2016, 2581-2585. (58) van der Meer, M.; Rechkemmer, Y.; Breitgoff, F. D.; Dechert, S.; Marx, R.; Dörfel, M.; Neugebauer, P.; van Slageren, J.; Sarkar, B. Dalton Trans. 2016, 45, 8394-8403. (59) Turner, J. W.; Schultz, F. A. Inorg. Chem. 2001, 40, 5296-5298. (60) Halcrow, M. A. Spin-crossover materials : properties and applications; Wiley: Chichester, West Sussex, United Kingdom, 2013. (61) Berry, J. F.; Cotton, F. A.; Lu, T. B.; Murillo, C. A. Inorg. Chem. 2003, 42, 4425-4430. (62) Weber, B.; Walker, F. A. Inorg. Chem. 2007, 46, 6794-6803. (63) Lever, A. B. P. Inorg. Chem. 1990, 29, 1271-1285. (64) Gallé, G.; Jonusauskas, G.; Tondusson, M.; Mauriac, C.; Létard, J.-F.; Freysz, E. Chem. Phys. Lett. 2013, 556, 82-88. (65) In light of our interpetation of [Fe(dftpy)2]2+ dynamics (bluer excitation) in the context of the halogen series as a whole (vide infra) we believe it may be interesting to reassess the interpretation of dyamics in [Fe(2-Mephen)3]2+ including the conclusion of high-spin to low-spin photo-switching. Addional TA spectral data that seeks to better resolve and characterize near-UV TA features would be valuable. (66) McCusker, J. K.; Rheingold, A. L.; Hendrickson, D. N. Inorg. Chem. 1996, 35, 2100-2112. (67) Brown, A. M.; McCusker, C. E.; McCusker, J. K. Dalton Trans. 2014, 1-12. (68) Hewitt, J. T.; Vallett, P. J.; Damrauer, N. H. J. Phys. Chem. A 2012, 116, 11536-11547. (69) Shaw, G. B.; Styers-Barnett, D. J.; Gannon, E. Z.; Granger, J. C.; Papanikolas, J. M. J. Phys. Chem. A 2004, 108, 4998-5006. (70) Another possibility is an alternate 3MLCT having a configuration with a single electron in an metal-ligand antibonding orbital. (71) Note that the observed lifetime trend would not be expected if the intermediate state in the non-radiative decay was a quintet metal-centered excited state (5MC*) akin to the 5E for true octahedral systems where there are three electrons in metal-ligand antibonding orbitals. As the halogen size increases, the 5MC* will stabilize relative to the 5,7MLCT (where there are two electrons in metal-ligand antibonding orbitals). Because 5MC* would be expected to be in the normal region with respect to the 5,7MLCT, the lifetime should decrease with increasing halogen size.

32


Page 33 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


For Table of Contents Only:

33


A Synthetically Tunable System To Control MLCT Excited-State

Recommend Documents