Through-Space Intervalence Charge Transfer as a Mechanism for

Subscriber access provided by Kaohsiung Medical University

Article

Through-Space Intervalence Charge Transfer as a Mechanism for Charge Delocalisation in Metal-Organic Frameworks Carol Hua, Patrick William Doheny, Bowen Ding, Bun Chan, Michelle Yu, Cameron J. Kepert, and Deanna M. D'Alessandro J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.8b02638 • Publication Date (Web): 04 May 2018 Downloaded from http://pubs.acs.org on May 4, 2018

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 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 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.

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 22 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

Through-Space Intervalence Charge Transfer as a Mechanism for Charge Delocalization in Metal-Organic Frameworks Carol Hua,a Patrick W. Doheny,a Bowen Ding,a Bun Chan,b Michelle Yu,a Cameron J. Kepert,a Deanna M. D’Alessandroa* a

School of Chemistry, The University of Sydney, New South Wales, 2006 Australia

b

Graduate School of Engineering, Nagasaki University, Nagasaki 852-8521 Japan

* Corresponding author phone: +61 2 93513777; email: [email protected]

Abstract Understanding the nature of charge transfer mechanisms in 3-dimensional Metal-Organic Frameworks (MOFs) is an important goal owing to the possibility of harnessing this knowledge to design electroactive and conductive frameworks. These materials have been implicated as the basis for the next generation of technological devices for applications in energy storage and conversion, including electrochromic devices, electrocatalysts, and battery materials. After nearly two decades of intense research into MOFs, the mechanisms of charge transfer remain relatively poorly understood, and new strategies to achieve charge mobility remain elusive and challenging to experimentally explore, validate and model. We now demonstrate that aromatic stacking interactions in Zn(II) frameworks containing cofacial thiazolo[5,4-d]thiazole units lead to a mixed-valence state upon electrochemical or chemical reduction. This through-space Intervalence Charge Transfer (IVCT) phenomenon represents a new mechanism for charge transfer in MOFs. Computational modelling of the optical data combined with application of Marcus-Hush theory to the IVCT bands for the mixed-valence framework has enabled quantification of the degree of charge transfer using both in situ and ex situ electro- and spectro-electrochemical methods. A distance dependence for the through-space electron transfer has also been identified on the basis of experimental studies and computational calculations. This work provides a new window into electron transfer phenomena in 3-dimensional coordination space, of relevance to electroactive MOFs where new mechanisms for charge transfer are highly sought after, and to understanding biological light harvesting systems where through-space mixed-valence interactions are operative.

1 ACS Paragon Plus Environment

Journal of the American Chemical Society 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 22

Introduction Metal-organic frameworks (MOFs) are a class of functional materials derived from the reaction of metal nodes with organic linkers to form supramolecular assemblies that exhibit features such as nanoscale porosity, large internal surface areas, well-defined cavities and voids throughout their structures.1 Although MOFs are well known for their adsorptive properties, the incorporation of redox-active metal ions or organic ligands can impart additional functionality via the chemical and structural changes associated with redox transformations.2 The synthesis of electroactive MOFs is often impeded by the intrinsically insulating nature of the organic components and the closed shell nature of the metal ions (i.e., Zn2+, Zr4+) commonly used in the formation of such materials. An additional challenge is the inherent instability of MOFs to oxidation or reduction, where structural changes due to redox manipulation of either the metal or organic components can result in irreversible decomposition and loss of structure.

Despite this, important strides in the experimental studies of electroactive and conductive MOF materials has driven a surge of interest in understanding their mechanisms of charge transfer.3-5 This interest has been driven, in part, by potential applications of these systems in technologically and industrially useful devices for applications in energy storage and conversion, including electrochromic devices,2,6-8 electrocatalysts,9-11 and battery materials.12-14 MOFs also provide a unique platform upon which fundamental electron transfer phenomena can be interrogated in threedimensional coordination space owing to their ordered, crystalline structures, in which the relative spatial orientation and location of the multiple metal and organic units are unambiguously defined.

One means of interrogating the mechanisms of charge transfer in MOFs, with a view to exert greater control over their conductive properties, is through the exploitation of mixed valency. This is a widespread phenomenon in a plethora of physical, chemical and biological systems ranging from gemstones and photosynthetic reaction centres in Nature,15-17 to materials science applications such as molecular electronics for quantum computing.18,19 The common feature of mixed-valence systems is the presence of either metal or organic moieties with different formal oxidation states, giving rise to the rapid oscillation of charge between the centres and, correspondingly, the absorption of light in the electromagnetic spectrum (which explains the highly coloured nature of these compounds). Marcus-Hush theory has been widely employed as a cornerstone of analysis for these systems, providing insight into both through-bond (inner sphere) and through-space (outer sphere) electron transfer through the pioneering experimental and theoretical work of Taube, Marcus and Hush, amongst many others.20-24 2 ACS Paragon Plus Environment

Page 3 of 22 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


To date, mixed valency has been characterised in frameworks that exhibit a through-bond Intervalence Charge Transfer (IVCT) mechanism which has been invoked as the origin of electroactivity and conductivity in these materials. For example, mixed valency is the origin of the through-bond charge hopping mechanism in the historically important pigment Prussian blue (Fe4III[FeII(CN)6]3.xH2O) whose deep colour and intrinsic semiconductivity arises from IVCT between the Fe(II) and Fe(III) nodes mediated by cyanido ligands-.25,26 In contrast to the prototypical metal-to-metal IVCT of Prussian blue, Long and co-workers recently reported an Fe(III) framework that exhibits the first example of purely ligand-based IVCT originating from the through-bond interplay between multiple redox states of the 2,5-dihydroxybenzoquinone (dhbq) ligands.27,28 This material, [(NBu4)2FeIII2(dhbq)3] (where NBu4 = tetrabutylammonium and dhbq = dihydroxybenzoquinone) is the most highly conductive three-dimensional MOF known at the present time, and exhibits ligand-based mixed valency which manifests in an IVCT band due to charge transfer between dhbq2− and dhbq3−. Other examples of through-bond mixed valency in frameworks with (6,3) net-type topologies include Harris’ 2D iron 2,5-dichloro-3,6-dihydroxo-1,4-benzoquinone-based material29,30 and Abrahams and Robson’s (NEt4)2[Fe2(can)3] (where can is the dianion of chloranilic acid), where a strong near-infrared (NIR) IVCT band was observed in the latter case.29,31

By comparison, through-space IVCT behaviour is extremely rare in both discrete and extended coordination materials, with the exception of the elegant rhenium-based discrete molecular rectangles reported by Dinolfo and Hupp.32-34 Through-space mixed valency is well-known in biological systems such as the mixed-valence radical cation or “special pair” in bacterial photosynthesis.15,16,35 The ability to engineer 3D materials which display such through-space electronic delocalization is significant from both a practical and theoretical perspective, the former being a versatile platform from which to fabricate materials that exploit mixed valency to achieve both conductivity and multifunctionality. From a theoretical perspective, the presence of mixed valency leading to IVCT should offer new insights into fundamental aspects of charge transfer in 3D systems.

Marcus-Hush theory has been widely applied to quantify the degree of electron delocalization in biological, mineralogical and synthetic materials.36,37 For mixed-valence systems, the degree of electron delocalization can be classified into three main classes using the Robin and Day classification scheme.38 Class I compounds involve redox centres that do not interact with each other, or where the interaction is too small to be detected; Class II compounds 3 ACS Paragon Plus Environment


Page 4 of 22

have weakly interacting redox centres, where the charge is primarily localised on one of the redox centres; and Class III describes systems where the charge is completely delocalized over the entire system due to the strong electronic coupling between the redox centres. While there has been extensive investigation of electron transfer in short range systems, such as discrete complexes, there has been a distinct lack of exploration of longer range and polymeric systems, such as MOFs. This can be attributed to the difficulties in obtaining experimental data for the often unstable intermediate mixed-valence states of solid-state materials. It has only been recently that in situ solid state UV/Vis/NIR spectroelectrochemical techniques have been developed that allow spectral data of the mixed-valence state of a bulk sample to be obtained.2,39 In addition, quantifying experimental parameters for theoretical analyses of IVCT from the electromagnetic spectra of mixed-valence species have, to date, proven problematic due to difficulties in quantifying the molar extinction coefficient from the electronic spectra of solid materials.

Herein, we describe a new Zn(II) framework, [Zn2(BPPTzTz)2(tdc)2]n, denoted 1 (where BPPTzTz = 2,5-bis(4(pyridine-4-yl)phenyl)thiazolo[5,4-d]thiazole), containing cofacial thiazolo[5,4-d]thiazole units that exhibits the first example of the through-space IVCT phenomenon in a framework material. The origin of the mixed-valence interaction is elucidated using computational density functional theory (DFT) methods, where the close proximity of the electroactive units, facilitated by aromatic stacking interactions, is crucial. Although alluded to in bacterial photosynthetic15,16,35 and discrete molecular systems,32-34 this mechanism has not previously been observed or characterised in framework materials, although aromatic stacking interactions and their associated structural effects have been previously demonstrated to facilitate intermetallic through-bond IVCT in mixed-valence Fe/Co tetranuclear complexes.40 Important considerations for the theoretical analysis of mixed-valence systems in the solid state are addressed, particularly determination of the extinction coefficient (ε) for the IVCT band of the mixed-valence species using Kubelka-Munk analysis, which enables application of Marcus-Hush theory. This connection between experiment and theory represents a significant advance in quantifying charge transfer interactions in MOFs. The combined experimental and computational approach is also used to probe the distance dependence of the throughspace mixed valency, a prediction of Marcus-Hush theory,21 using the closely related framework [Zn4(BDPPTzTz)2(tdc)2]n (2) (where BDPPTzTz = 2,5-bis(4-(pyridine-4-yl)-3,5-dimethyl-phenyl)thiazolo[5,4d]thiazole) in which the close approach of the electroactive ligands is precluded, in addition to computational simulations of the electronic spectra. This work provides a new window into charge delocaliztransfer mechanisms in 3-dimensional materials. 4 ACS Paragon Plus Environment

Page 5 of 22 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


Results and Discussion Synthesis and Structure. The two new ligands, BPPTzTz and BDPPTzTz, were synthesised via two-step reactions. The first step involved a Suzuki coupling between 4-pyridylboronic acid and either 4-bromobenzaldehyde (BPPTzTz) or 4-bromo-3,5-dimethylbenzaldehyde (BDPPTzTz) to yield the dipyridyl compound. The intermediate then underwent a condensation reaction with dithiooxamide in DMF to form the thiazolo-thiazole core.

Two Zn(II) frameworks, [Zn2(BPPTzTz)2(tdc)2]n (1) and [Zn4(BDPPTzTz)2(tdc)2]n (2), were obtained upon heating the BPPTzTz or BDPPTzTz ligands with 2,5-thiophene dicarboxylic acid (tdc) and Zn(NO3)2·6H2O at 120 °C to yield orange and yellow prismatic crystals, respectively. The synthesis of the frameworks was repeated a number of times and able to be obtained reliably. The structure of 1 was solved and refined in the orthorhombic space group Pcc2 (Figure 1). The asymmetric unit consists of two BPPTzTz ligands and two tdc co-ligands. The two Zn(II) centres are coordinated by the carboxylate moiety on the tdc co-ligand, where the oxygen donors are coordinated in a monodentate manner to each of the Zn(II) centres. Importantly, the tdc co-ligand holds two π-stacking BPPTzTz ligands in close proximity resulting in a ligand-to-ligand distance of 3.80 Å. The Zn(II) centres exhibit a trigonal bipyramidal coordination with three oxygen donors in the equatorial plane and two nitrogen donors from the pyridyl rings of the BPPTzTz ligand on the axial positions. The biphenyl moiety in the ligand exhibits a slight rotation of approximately 27° from each of the pyridyl rings, such that they are not coplanar.

The dinuclear Zn(II) centres in 1 are linked by tdc ligands yielding a sheet-like structure, which is then extended into a 3D cubic α-polonium (pcu) network by the dipyridyl BPPTzTz ligand acting as a pillar and the binuclear units serving as six-connecting nodes. 1 consists of two interpenetrated networks (Figures 1a and b), but still exhibits significant channels when viewed down either the a or c axes, which is partly due to the π-stacking of the aromatic moieties and favourable S····N interactions between adjacent BPPTzTz ligands in the structure (Figure 1c). The pore void space was calculated in SQUEEZE41 to be 30%. The TGA of 1 exhibits a gradual mass loss until 300 °C due to liberation of the DMF in the pores of the framework, before a steep mass loss indicates decomposition of the framework (Supplementary Figure S2). Gas adsorption studies with N2, H2 and CO2 were performed on the framework after activation (using a supercritical CO2 wash), following which the structural integrity of the material was maintained according to powder X-ray diffraction measurements (Supplementary Figure S3). The framework displays a small



Page 6 of 22

adsorption of N2 (0.76 mmol/g at 960 mbar), H2 (1.72 mmol/g at 1200 mbar) and CO2 (0.7 mmol/g at 1200 mbar) with a Brunauer–Emmett–Teller (BET) surface area of 53.4 ± 0.2 m2/g (Supplementary Figure S4). a)

b)

c)

Figure 1. Crystal structure of framework 1 showing the cofacial alignment of the BPPTzTz ligands. (a) View down the c axis, (b) view down the a axis and (c) the “cofacial” pair of BPPTzTz ligands shown in the box in (a). Only one of the two interpenetrated networks is shown. Atom labelling: Zn = cyan, O = red, N = blue, S = yellow, C = grey. The hydrogen atoms are omitted for clarity.

To generate a material in which cofacial interactions between ligands were eliminated, the sterically bulky BDPPTzTz ligand was used to synthesise framework 2 which was solved and refined in the triclinic P-1 space group. The asymmetric unit of the framework consists of two BDPPTzTz ligands and four tdc ligands with four crystallographically distinct Zn(II) centres (Supplementary Figure S5). In contrast to 1, the oxygen donors from the carboxylates on four tdc ligands bind in a monodentate fashion to two Zn(II) centres to yield a ‘paddlewheel’ unit. Coordination of the tdc ligand with monodentate binding from each of the two carboxylate groups, as opposed to monodentate coordination from the same carboxylate group, results in a distance of 8.93 Å between the two BDPPTzTz ligands. The adjacent BDPPTzTz ligands are rotated with respect to each other indicating the presence of minimal π-stacking interactions. On the axial positions of the ‘paddlewheel’ are nitrogen donors from the pyridyl ring of the BDPPTzTz ligand. The BDPPTzTz ligand has a 90° orientation of its pyridyl rings with respect to the central core. 6 ACS Paragon Plus Environment

Page 7 of 22 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


Framework 2 consists of two interpenetrating networks with pores down the c axis (Supplementary Figure S5). The SQUEEZE41 function in PLATON42 gave an estimated 42% accessible pore space−a value that is significantly greater than that for 1, presumably due to the influence of the steric bulk afforded by the extra methyl groups in the BDPPTzTz ligand. The TGA of 2 showed a mass loss of 15% under 200 °C due to liberation of the DMF solvent from the pores of the framework (Supplementary Figure S2). It is likely that steric hindrance provided by the methyl groups on the BDPPTzTz ligand prevents the loss of additional solvent from the pores.

Electroactive Properties. Solid-state cyclic voltammograms on the BPPTzTz ligand in [(n-C4H9)4N]PF6/CH3CN revealed two processes in the cathodic region and no processes in the anodic region (Supplementary Figure S6). The processes observed at -1.69 and -2.01 V vs. Fc/Fc+ can be assigned to two sequential one electron reduction processes of the thiazolo-thiazole core in the BPPTzTz ligand, involving a one electron quasi-reversible reduction to the radical anion followed by a further one electron irreversible reduction to the dianion. Similar processes were observed for BDPPTzTz, however, an additional cathodic process was observed. The peaks at -1.49 and -1.82 V vs. Fc/Fc+ were assigned to reduction of the TzTz core, with the peak at -2.40 V vs. Fc/Fc+ due to reduction of the pyridyl moieties on the ligand (Supplementary Figure S7). The reduction of the pyridyl rings for BDPPTzTz within the electrochemical window for acetonitrile may be a consequence of weakly electron donating methyl groups which lower the potential required for reduction.

The electrochemistry of frameworks 1 and 2 is dominated by ligand based processes (Figures S9-S11). In the anodic region, a broad process at ~1.0 V vs. Fc/Fc+ was observed for both frameworks and can be assigned to oxidation of the thiophene ring in the tdc co-ligand. In contrast, the presence of the methyl groups in BDPPTzTz resulted in a marked difference in the electrochemical potentials for reduction of the ligand when compared to BPPTzTz. Two broad reductive process were observed in the solid state cyclic voltammogram for 1 in the cathodic region at -1.38 and -2.03 V vs. Fc/Fc+. These processes can be assigned to the corresponding one and two electron reductions of the BPPTzTz ligand (Supplementary Figure S8). In contrast, three reductive processes were present for 2, where the weakly electron donating methyl groups facilitate reduction of the ligand within the framework (Figure S9). The processes observed at -1.15 and -2.03 V vs. Fc/Fc+ in 2 are due to the thiazolo-thiazole core, whilst the process at -2.32 V vs. Fc/Fc+ is 7 ACS Paragon Plus Environment


Page 8 of 22

assigned to reduction of the pyridyl rings within the BDPPTzTz ligand. These reduction processes correspond well to those observed for the ligand itself.

The electrochemistry of 1 was also examined in [(n-C4H9)4N]PF6/CH2Cl2 electrolyte (Supplementary Figure S10) for use in the solid-state EPR spectroelectrochemical experiment (vide infra) due to the lower dielectric constant of dichloromethane compared with acetonitrile. At scan rates of 100 and 500 mV/s, two quasi reversible processes were observed in the cathodic region at -1.26 and -1.40 vs. Fc/Fc+. These processes are ascribed to one electron reduction of the thiazolo-thiazole core to its radical anion. The framework was also found to have good electrochemical properties in 0.1 M KCl/H2O electrolyte, where the material underwent multiple cycles with minimal changes in the CV, demonstrating its structural stability in both aqueous and non-aqueous media (Supplementary Figure S11).

Modulating Redox State. The neutral BPPTzTz and BDPPTzTz ligands displayed one main band in their electronic absorption spectra above 20000 cm-1 which can be attributed to the π→π* transition of the aromatic phenyl and pyridyl groups in the ligand. As the potential was decreased from 0 to -1.6 V during the solid state UV/Vis/NIR spectroelectrochemical experiment for the BPPTzTz ligand, three bands formed at 14000, 17500 and 23000 cm-1 along with an associated colour change from yellow to dark yellow (Supplementary Figure S12). The bands at ~14000 and 17500 cm-1 can be assigned to the radical anion of the thiazolo-thiazole core.

The features in the solid state UV/Vis/NIR spectrum of 1 predominantly arise from absorptions of the BPPTzTz ligand (Supplementary Figure S12). The bands above 20000 cm-1 can be assigned to π → π* transitions of the aromatic moieties within the BPPTzTz ligands. Similar to the ligands themselves, 1 displays a lower energy absorption band than 2. Upon changing the potential from 0 to -1.75 V for 1, new bands form at 15000, 17000 and 23500 cm-1, with an associated colour change from yellow to dark green (Figure 2a, Supplementary Figure S13). Unexpectedly, a new broad band appeared in the NIR region of the spectrum, which is absent upon reduction of the ligand itself. The broad band in the NIR region can be deconvoluted to give two Gaussian bands centred at 6576 and 8202 cm-1. As the potential is further changed to -1.85 V, the NIR band decreases, supporting its assignment as an Intervalence Charge Transfer (IVCT) band that arises upon formation of a mixed-valence state of the material (Supplementary Figure S13). Upon returning the potential to 0 V, the framework is able to be ‘reset’ to its neutral state, with a colour change from dark green to yellow and a spectrum reminiscent of that for the starting material 8 ACS Paragon Plus Environment

Page 9 of 22 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


demonstrating both electrochemical reversibility and structural stability of the material to redox modulation. An in situ solid state EPR spectroelectrochemical experiment43-45 on 1 (Figure 2b) corroborated the formation of a radical state upon reduction.

Figure 2. Solid state spectroelectrochemistry at room temperature of 1. (a) Vis-NIR at applied potentials of 0 to -1.75 V in [(n-C4H9)4N]PF6/CH3CN electrolyte. Photographs of the framework during the solid state UV/Vis/NIR spectroelectrochemistry experiment demonstrate the marked colour change from yellow (neutral) to dark green (mixed-valence state). (b) X-band EPR at applied potentials of 0 to -2.0 V in [(n-C4H9)4N]PF6/CH2Cl2 electrolyte (applied modulation amplitude of 2 G) showing the progressive evolution of a radical species.

The properties of 2 in which no cofacial interactions are present were of interest to understand the nature of the ligandto-ligand interactions in 1 which give rise to mixed valency. As the potential was reduced from 0 to -1.4 V in 2, a colour change from a pale to dark yellow was observed. This was similar to observations for the BDPPTzTz ligand itself, which darkened from yellow to dark yellow at -2.0 V with increases in intensity of bands at 18690, 21730 and 24290 cm-1. These processes can be assigned to the formation of the radical anion state of the thiazolo-thiazole core, and were fully reversible upon returning the potential to 0 V (Supplementary Figure S14). In a similar manner to the UV/Vis/NIR spectroelectrochemical experiment on the BDPPTzTz ligand, reduction of 2 at -1.4 V led to an increase in intensity of the bands at 15450, 20880 and 24000 cm-1 (Supplementary Figures S16 and S17). The reductive process was spectrally reversible, with the starting spectrum obtained upon returning the applied potential to 0 V. The redox processes in 2 are therefore dominated by those of the ligand. Notably there is the lack of a band in the NIR as observed for 1. This can be explained upon inspection of the solid state structures of each of the zinc frameworks,



Page 10 of 22

where there is an absence of cofacial interactions between the BDPPTzTz ligands in 2 which are located at a distance of 8.93 Å apart, precluding through-space mixed valency and IVCT.

Computational Modelling. Single point TD-DFT calculations using the BMK\6-311G(d) procedure46 on the neutral BPPTzTz ligand revealed a single band at 22700 cm-1 due to a transition between the HOMO and LUMO+1 orbitals (Supporting Information, Figure S18), which is consistent with the experimental spectrum (23500 cm-1, Figure 2a) and supports the assignment as a π → π* transition.

The same calculation was performed on the cofacial dimer pair of BPPTzTz ligands extracted from the X-ray crystal structure of 1. For the singly reduced dimer, using the atomic coordinates of the neutral material, bands centred at 7319 and 9292 cm-1 were predicted for the SOMO to LUMO+1 and SOMO to LUMO+2 transitions, respectively (Figure 3a, Supplementary Figure S18). Visual inspection of the molecular orbitals shown in Figure 3a reveals that the lower energy NIR transition occurs from a SOMO almost fully localised on one BPPTzTz molecule of the dimer pair to the LUMO+1 orbital predominantly localised on the adjacent BPPTzTz, i.e., an IVCT transition. The second transition in the NIR arises from the SOMO orbital localised on one BPPTzTz ligand to the LUMO+2 orbital localised mainly on the same ligand, but also with substantial IVCT character. Gaussian deconvolution of the NIR band obtained from the experimental spectrum of the mixed-valence species in Figure 2a indicated two major underlying components at 6576 and 8202 cm-1 (Figure 3b) which are consistent with the predictions from the DFT calculations (Supplementary Table S4 contains details of the deconvoluted spectrum). Transitions in the visible region at 19107 and 19911 cm-1 were also predicted via computational calculations (vs. 15000 and 17000 cm-1 determined experimentally) due to the generation of an anionic radical with associated SOMO to LUMO+6 and HOMO to LUMO transitions, respectively.


Page 11 of 22 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 3. Comparison of the computationally predicted and experimental IVCT bands in the mixed-valence form of 1. (a) Molecular orbitals involved in the transitions in the NIR region for the cofacial mixed-valence dimer (BPPTzTz0/•−)2 extracted from the crystal structure of 1. (b) The two lowest energy Gaussian components of the NIR band assigned to the mixed-valence form of 1, [Zn2(BPPTzTz0/•−)2(tdc)2]n obtained upon deconvolution of the experimental spectrum . The curvefit (black line) obtained upon deconvolution (dotted lines) is overlayed with the experimental spectrum (red line).

In light of these results, the prominent band observed in the NIR region for the mixed-valence state of framework 1 arises primarily due to IVCT between BPPTzTz in its radical anion state, BPPTzTz•−, and the adjacent, formally neutral BPPTzTz0 ligand. Calculations performed on the discrete neutral and singly reduced ligands reveal that the angular offset of the outer pyridyl rings decreases relative to the planar core upon reduction. We postulate that a similar effect occurs in the framework, whereby reduction of one ligand induces a resonance effect in the other that decreases the stacking offset and increases the overall donor-acceptor orbital overlap, thereby facilitating the IVCT behaviour. Such an effect has previously been postulated for stacked porphyrinic ligands in molecular rectangles displaying similar through-space IVCT properties.34 11 ACS Paragon Plus Environment


Page 12 of 22

The distance dependence of IVCT transitions is well-known in biological systems and in synthetically designed dinuclear complexes which have been used as models for exploring mixed valency.15,16,21,35,47 In the present work, the relative distance between the DPPTzTz ligands in the dimer pair was adjusted computationally to reflect a geometry in which orbital overlap would be increased, and another in which the distance between the pair would be so great that orbital overlap would be non-existent (Figure 4, see also Supplementary Figures S19-S26). TD-DFT calculations revealed that the NIR transitions corresponding to IVCT in the dimer pair did indeed have a strong distance dependence, with the relative intensity of the underlying IVCT transitions increasing at distances closer than the crystallographic distance (see also Supplementary Figures S27-S29). Conversely, when the dimer pair was set at a distance >5 Å, IVCT transitions were non-existent in the calculated spectra, with only an intramolecular transition localised on DPPTzTz appearing in the NIR region of the calculated spectrum (see Supplementary Figure S18). These results are consistent with the well-known distance dependence for IVCT,21 and corroborate the absence of an IVCT band in framework 2 where the ligands are separated by >8 Å.


Page 13 of 22 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 4. Distance dependence of the through-space IVCT interaction. (a) Schematic representation of the cofacial pair of BPPTzTz ligands showing reduction to the mixed-valence state which facilitates a through-space IVCT interaction. (b) Simulations of the UV-Vis-NIR spectra for the cofacial dimer (BPPTzTz0/•−)2 in a -1 reduced state. The distance between the BPPTzTz ligands was initially the crystallographic distance of 3.80 Å, where the two major components of the NIR band were reproduced. Moving the ligands closer (3.5 Å) also confirmed the appearance of two major underlying components, whereas increasing the inter-ligand separation to 6.0 Å led to a loss of the throughspace IVCT transition, with the appearance of an intra-ligand transition only.

Intervalence Charge Transfer. To quantify the extent of electronic delocalization in the mixed-valence framework, Marcus-Hush theory was employed based on the parameters of the two lowest energy NIR bands which were found to have significant IVCT character. The energies (νmax = 6576 and 8202 cm-1) and bandwidths-at-half-height (∆ν1/2 = 1465 and 2939 cm-1) for these two transitions were readily determined from the deconvolution to the experimental spectrum for the mixed-valence form of 1, [Zn2(BPPTzTz0/•−)2(tdc)2]n, as shown in Figure 3. The molar extinction coefficients for the two band maxima were derived by application of an extended form of Kubelka-Munk theory.48 This involved first quantifying the molar extinction coefficient of the UV-Vis π → π* band in the neutral framework using KBr pellets containing various concentrations of the framework (see Supplementary Information and Figures 13 ACS Paragon Plus Environment


Page 14 of 22

S30-S32 for details of the analysis). The extinction coefficients for the two NIR bands at 6576 and 8202 cm-1 were subsequently determined as 54 and 106 M-1cm-1, respectively, and the electronic coupling constants, Hab, as 123 and 273 cm-1, respectively, through application of Marcus-Hush theory.20 Given that both transitions contain some intramolecular charge transfer character localised on a given BPPTzTz ligand, these values should be considered as upper limits. The Robin and Day classification scheme for electron transfer38,49,50 provides a description of the extent of electronic delocalization: since 2Hab

Through-Space Intervalence Charge Transfer as a Mechanism for

Recommend Documents