rticle history:eceived 30 May 2012eceived in revised form 9 August 2012ccepted 10 August 2012vailable online 20 August 2012
eywords:
a b s t r a c t
This review outlines the synthesis and properties of mono and oligonuclear polypyridyl ruthenium com-plexes incorporating a range of S3 and S4 thiacrown ligands, focusing on the mixed valence complexesthat have been produced in these studies. This work has revealed that the chemical, electrochemical,and electronic properties of the metal centers are modulated by the nature of the coordinated thiacrownligand. In particular the back-bonding properties of the ligands mean that they strongly stabilize theruthenium(II) state and also produce metal centers that are relatively kinetically labile. These distinctive
lectron transferutheniumhiacrownsixed valence
electronic properties have been exploited in the construction of mixed valence systems with unusualelectrochemical properties and also facilitated the self-assembly of oligonuclear molecular architectureswith multiple accessible oxidation states. Computational studies on the isovalent and mixed valencecomplexes reflect the experimental data by indicating that, within a series of S4 coordinated dinuclearcomplexes, the properties of the mixed valence state are subtly dependent on the cavity size of thecoordinated thiacrown.
. Introduction
Since the original report on the synthesis and optical proper-ies of the Creutz–Taube (CT) ion [1] [(NH3)5Ru(�-pyz)Ru(NH3)5]5+
where pyz = pyrazine), a huge number of papers on rutheniumased mixed valence, MV, systems have been published [2]. Thisork has yielded a variety of molecular devices including molec-
lar wires and switches [3–7] and also provided insights intohe kinetics and thermodynamics of long-range electron transfer8,9]. Although these studies have encompassed a wide range of
bridging and ancillary ligands, in the vast majority of cases, coor-dination to ruthenium centers has exclusively involved nitrogendonor sets such as NH3 and 2,2′-bipyridine (2,2′-bpy). Given thatmany biological electron transfer (ET) systems are based on tran-sition metals coordinated to sulfur donor sites, my research groupdecided to investigate how the incorporation of sulfur donatingligands would modulate the ET properties of MV complexes whichwere analogous to the conventional N-donor systems. In partic-ular, we focused on using thiacrowns shown in Fig. 1, as – apartfrom allowing us to control the number of S-donors in the ligand
set – the size of the macrocycle ring provides a method to “finetune” the coordination geometry around metal centers [10] whenthey are linked by pyridyl based bridging ligands. Our first step insuch studies was to investigate the chemical and electrochemical
Fig. 1. Thiacrown ligands relevant to this review.
roperties of the RuII(9-S3) fragment, in order to assess whetherhis moiety would be a suitable building block towards the synthe-is of mixed valence systems.
. Mononuclear complexes with monodentate pyridyligands
The complex [RuCl2(DMSO)(9-S3)], 1, which was first reportedy the Sheldrick group [11], seemed to be a suitable starting mate-ial for such studies; an investigation into its reaction with a varietyf pyridyl-based ligands led to distinctive results – Scheme 1 [12].
The simple addition of monodentate pyridyl ligands to solu-ions of 1 led to the isolation of monocationic products suchs [RuCl(py)2(9-S3)]+, 2+. Presumably, the formation of theseonocations is due to the restricted ability of thioethers to neu-
ralize positive charges through �-donation [13,14]. However,
re-treatment of 1 with Ag+ to remove coordinated chloride ligandsollowed by addition of the relevant pyridyl ligand resulted in thesolation of the dications such as [Ru(py)3(9-S3)]2+, 32+ [12]. It wasound that, although the dicationic complexes are stable in poorly
cheme 1. Reactions involving monodentate nitrogen donor ligands and the RuII(9-3) fragment.
eviews 257 (2013) 1555–1563
coordinating solvents such as nitromethane and dichloromethane,they are unstable towards solvolysis in more coordinating solvents,as illustrated by the fact that over a period of two weeks stirred ace-tonitrile solutions of 32+ cleanly converted to the dipyridyl complex[Ru(py)2(NCMe)(9-S3)]2+, 42+. These substitution reactions precedefurther under more forcing conditions: for example, after threedays refluxing in acetonitrile complex 32+ is cleanly convertedinto the mono-pyridyl complex [Ru(py)(NCMe)2(9-S3)]2+, 52+. Stri-kingly, related nitrile-based complexes do not display the samereactivity, complexes [Ru(NCR)3(9-S3)]2+ (R = Me, Ph) are inert tosolvolysis by other coordinating solvents, even over a period of1 month. Ruthenium(II) complexes incorporating pyridine ligandsare often synthesized via nitrile intermediates, but clearly in thiscase the opposite is true. This unusual reactivity is due to two fac-tors. The crystal structure of 32+ reveals that one face of the complexis steric congested due to interactions between the ortho hydro-gens of the coordinated pyridine ligands. It was reasoned that theseunfavorable steric interactions are one of the causes of the rela-tive instability of 32+ and explains why the substitution of pyridineligands stops after the replacement of only two of these ligands.However, the related facially capped complex [Ru(tpm)(py)3]2+
(tpm = tris(1-pyrazole)methane) is stable towards solvolysis [15]and a comparison of its crystal structure with that of 32+ showsthat the Ru-py bond lengths in the tpm-based complex are almost5 pm shorter than those for 32+. These observations indicate thatthe unusual reactivity of 32+ can also be partially attributed to elec-tronic factors – it is now well established that thiacrown ligandsare good �-acceptor ligands with C–S �* orbitals of appropriatesymmetry acting as the acceptors [16]. However, while coordi-nated monodentate pyridyl ligands are labilized by the RuII(9-S3)moiety, we found that by employing bidentate chelating ligandswe could synthesize oligonuclear complexes that displayed muchgreater kinetic stability, allowing their mixed valence states to bestudied.
3. Dinuclear mixed valence RuII/III systems containing the9-S3 ligand
To make a detailed comparison between conventional MVsystems built up from [Ru(NH3)4]2+/3+ and [Ru(2,2′-bpy)2]2+/3+
units, dinuclear complexes bridged by three different bridg-ing ligands were synthesized [17,18]. The bridging ligands2,3-bis(2-pyridyl)pyrazine (bpp), 2,2′-bipyrimidine (bpym), and3,6-bis(2-pyridyl)-1,2,4,5-tetrazine (bpta) were selected to varythe intensity of coupling between the metal centers. While the firstmember of this series is a typical, much employed, bridging ligandin energy and electron transfer systems [19], previous studies haveshown that bpym bridged systems often display comparativelyweaker electronic and electrochemical coupling [20], whilst studieson bptz-based RuIII/II complexes demonstrated that the low lyingLUMO of the central tetrazine bridge can facilitate very intenseinteractions [21,22].
The reaction of 1 with the relevant ligands yielded the isovalentdinuclear complexes, [{RuCl(9-S3)}2(bpym)]2+, 62+, [{RuCl(9-S3)}2(bpp)]2+, 72+, and [{RuCl(9-S3)}2(bpta)]2+, 82+ – Fig. 2 [17,18].The electrochemical interaction between the two ruthenium(II)centers in these complexes was first assessed through cyclicvoltammetry.
It was found that although both 72+ and 82+ were oxidized in twoclearly defined and reversible one-electron processes, oxidation of62+ produced two overlapping waves. In all cases, the potentials
for the first oxidation of the thiacrown complexes were intermedi-ate to the values for their RuII(NH3)4 and RuII(2,2′-bpy)2 analogues,but closer to the figures obtained for the latter systems. For exam-ple, the first oxidation of 82+ was observed at 1.36 V (vs. Ag/AgCl),
J.A. Thomas / Coordination Chemistry Reviews 257 (2013) 1555–1563 1557
warioasRio1Kc
aiaSsabtceetp
tstia7t
td–e6adewt
�
Iet
all these processes were chemically irreversible and thus any MVstates could not be investigated. However, related studies on trinu-clear metallomacrocycles assembled through a similar approachled to more successful outcomes.
Fig. 2. Structure of complexes 62+–82+.
hile the potentials of the equivalent couples for the RuII(NH3)4nd RuII(2,2′-bpy)2 complexes were observed at 0.69 V and 1.52 Vespectively [21,22]. These observations indicate that, despite thenclusion of a �- and �-donor chloride ligand and the lower chargef the [RuCl(9-S3)]+ unit, the �-accepting properties of the thi-crown ligand still leads to a great degree of RuII oxidation statetabilization. A consideration of the difference in the Ru(1) andu(2) oxidation potentials, �E1/2, for the complexes revealed a sim-
lar trend. The smallest value of �E1/2 (120 mV) is observed for thexidation of 62+, with the equivalent values for 72+ and 82+ being55 mV and 480 mV respectively. Comproportionation constants,c, derived from these data revealed more subtle differences whenompared to the N-donor systems.
Kc values for the thiacrown-based complexes are, in all cases,ppreciably smaller than those for their Ru(NH3)4 analogues; thiss not unexpected as the inclusion of �-donor ancillary NH3 lig-nds will enhance the metal/bridging-ligand electronic overlap.omewhat surprisingly, while the large comproportionation con-tant for 82+ (1.4 × 108 [17]) is virtually identical to its Ru(2,2′-bpy)2nalogue, Kc values for 62+ and 72+ are both lower than their Ru(2,2′-py) equivalents [18]. By itself the electrochemical data indicateshat interaction between metal centers is weakest in the thiacrownoordinated systems; however, previous studies have shown thatlectrochemical data do not always give a true indication of actuallectronic coupling [23–25]. In fact, more definitive information onhis issue can be revealed through a consideration of the opticalroperties of the MV state.
The absorption spectra of 63+–83+ were probed by spectroelec-rochemistry. For all three complexes, generation of the RuIII/II MVtate leads to the growth of an near infra-red, intervalence chargeransfer (IVCT) band. As might be expected from the electrochem-cal data, 83+ displayed a relatively intense IVCT band; and whilebroader – but nonetheless well-defined – band was observed for3+, for 63+, a very broad, weak low energy band was only revealedhrough subtraction of the 62+ spectrum – Fig. 3.
These bands were then analyzed through a comparison withhe theoretical model for localized, “electron hopping,” MV systemseveloped by Hush [26–28]. It was found that electronic couplingas estimated by HAB – is considerably larger than indicated by thelectrochemical data. Indeed, the analysis revealed that although3+ is a Robin and Day Class II (valence localized) system [29], 73+
nd 83+ are in fact Class III, (valence delocalized) systems, with 83+
isplaying coupling comparable to the CT-ion itself [1]. The appar-nt contradictions between the electrochemical and optical dataere resolved through a consideration of the factors that contribute
o the free energy of comproportionation �Gc:
Gc = �Gr + �Ge + �Gs + �Gi (1)
n this expression �Gr is the free energy of resonance, �Ge is anlectrostatic term that takes into account the mutual repulsion ofhe two cationic metal centers, �Gs is an entropic factor that takes
Fig. 3. The [63+–62+] difference spectrum for the NIR region.Reprinted with permission from Ref. [18]. Copyright 2002 American Chemical Soci-ety.
the value (1/2)RT ln 1/4, and �Gi is a synergistic effect due to thestabilization of RuII by RuIII or vice versa [30]. Only �Gr is a measureof true electronic coupling. Previous electrochemical studies havedemonstrated that �Ge is dependent on the charge of the systemstudied [31]. From these arguments, it would be expected that �Ge
for RuIII/IICl(9-S3) based complexes will be less than that for morehighly charged RuIII/II(NH3)4 and RuIII/II(diimine)2 systems. It wasalso suggested that the combination of �-acceptor and �-donorligands within the RuCl(9-S3) moiety leads to an enhancement of�Gi thus stabilizing both RuII and RuIII oxidation states and favoringthe mixed valence state.
The enhanced kinetic lability of the RuII(9-S3) moiety has facili-tated the thermodynamic self-assembly of oligonuclear complexes.In such conditions entropic and enthalpic arguments indicate thatdiscrete architectures are favored [32]. In our first report on thisapproach, we used nitromethane as the non-coordinating solventand reacted 1 with 4,4′-bpy in an 8:12 stoichiometry. This led to theisolation of an octanuclear cube complex, 916+[33] – Fig. 4. Unfor-tunately, although the RuII centers display stepwise oxidizations,as for the analogous mononuclear complex [Ru(4,4′bpy)3(9-S3)]2+,
Fig. 4. Structure of self-assembled cube 916+.
1558 J.A. Thomas / Coordination Chemistry Reviews 257 (2013) 1555–1563
Ft
tF(
ftIRtitont
BastatPtcta[
tcfv
[
Fig. 6. dI1/dE vs. E deconvolution [38] of the cyclic voltammogram for the oxidation
Table 1Comproportionation constants derived from electrochemistry data for the mixedvalence states of Ru(9-S3)-based macrocyclic bowls.
Complex �E1/2 (mV) Kc
4+
ig. 5. Structure of metallomacrocycle 103+, reported by Fish and co-workers, andhiacrown RuII complexes 112+, 123+, and 133+.
Inspired by the elegant studies of Fish on trinuclear bowls con-aining cp*RhIII bridged by adenine-based ligands [34], such as 103+,ig. 5, we investigated the reaction of 1 with 9-methyl adenine9-MA) in a pH7 buffered aqueous solution [35].
Initially we found that the only product of the reaction was inact the mononuclear complex 112+, where the exocyclic NH2 ofhe coordinated adenine-based ligand has not been deprotonated.t seems that, compared to the cp*RhIII center, the coordinateduII(9-S3) moiety is not a strong enough Lewis acid to cause depro-onation of the amine group. Consequently, the N1 position of 9-MAs an insufficiently good donor to compete for the metal coordina-ion site occupied by DMSO in 112+ and self-assembly of higherrder structures does not occur. Since it is established that coordi-ated thiacrowns are susceptible to cleavage by nucleophiles [36]he deprotonation by a sterically hindered base was investigated.
In the presence of pentamethylpiperidine, PMP – a very strongrØnsted base but a weak nucleophile – the reaction of 1 with 9-MAffords the trinuclear bowl 123+ as the only product. A consequenttudy showed that, for suitably hindered adenine bridging ligands,his procedure appears to be general; for example reaction withdenosine yields macrocycle 133+ [37]. Oxidation of these struc-ures occurs in three reversible processes between 0.8 and 1.4 V.resumably these complexes now display reversible oxidations dueo the increased stability of the chelated bridging ligands. The threeouples are most clearly revealed by dI1/dE vs. E deconvolutions ofhe voltammetry data; this transform produces traces equivalent todifferential pulse polarogram with an additional reverse sweep
38] – Fig. 6.A comparison of data for 123+ and 133+ revealed that the
hree oxidation steps occur at almost identical potentials in bothomplexes. Thus the analysis confirms that the macrocycles haveour accessible oxidation states, two of which are formally mixed
alence:
RuII3]
3+−e−�
+e−[RuII
2RuIII]4+−e−
�+e−
[RuIIRuIII2 ]
5+−e−�
+e−[RuIII
3 ]6+
(2)
processes of complex 133+.
As shown in Table 1, the comproportionation constant calcu-lated for the [RuIIRuIII
2 ] states are several orders of magnitude largerthan that of the [RuII
2RuIII] states, suggesting that the intensity ofelectronic coupling is different in the two MV states; an infer-ence that was confirmed by absorption spectroscopy studies onthe macrocycles.
The spectra of 124+ and 134+ are dominated by structured, lowenergy IVCT bands typically seen in oligonuclear MV systems:low symmetry coordination geometries, extensive orbital mixingand spin orbit coupling lead to the d�5/6 states being split intoKramer doublets [2]. The relative low intensity and broadness ofthese bands, as well as a comparison to their theoretical widthat half-height, indicated that the [RuII
2RuIII] valence state for bothcomplexes is in an electron hopping state. Contrastingly, electro-chemical generation of 125+ or 135+ leads to the IVCTs, sharpening,shifting to lower energy, and increasing in intensity. An analysis ofthese new bands clearly reveals that that the [RuIIRuIII
2 ] state is infact valence delocalized.
These observations are in contrast with data obtained for con-ventional trinuclear mixed valence complexes bridged by tritopicligands complexes [39], where the spacing of electrochemical cou-ples and charge delocalization are approximately constant, aswould be expected if there was no variation in the individualmetal–metal interactions across the redox series. It was suggestedthat the anomalous properties of the macrocycles arise from theirdistinctive molecular architectures. In conventional systems witha central rigid tritopic bridging ligand, redox-induced changes inindividual metal–ligand bond lengths and angles are not directlymechanically coupled to the other metal centers within the com-plex. However, connectivity in the macrocycles is more complexas the metal centers are connected through three peripherallyarranged ditopic bridging ligands; hence changes in the bonds andangles at one metal center may be mechanically coupled to theother two. Therefore structural changes can, in principle, be propa-gated throughout the macrocycle – Scheme 2, thus affecting orbitalmixing within the whole system [37].
12 170 760134+ 165 625125+ 380 2.75 × 106
135+ 370 1.85 × 106
J.A. Thomas / Coordination Chemistry Reviews 257 (2013) 1555–1563 1559
Scheme 2. Schematic showing effect of oxidizing a single metal center in: (A) Atrinuclear complex bridged by a single central ligand. (B) A trinuclear complex wherettn
5c
itt
1rdtocaiiwtc
grisbpCbs–
i
solid state – both complexes take up conformation b, which is thelowest energy conformation for coordinated 14-S4 in mononuclearcomplex 152+ and only slightly higher in energy than conformationc for the coordinated 12-S4 in 142+. The most obvious differences
he metal ions are connected by three ditopic ligands. In both cases, it is assumedhe final mixed valence state is a Class II system, where bond lengths and angles areot averaged.
. Synthesis and NMR fluxionality of mononuclear Ru(n-S4)omplexes
To explore the effect of increasing the number of S-donor sitesn the ligand set of the ruthenium centers, we set out to synthesizeetrathiacrown-based analogues of 1 as suitable starting materialsowards this aim.
The reaction of Ru(DMSO)4Cl2 with n-S4 (where n = 12, 14, and6) yields monocationic complexes [RuCl(DMSO)(n-S4)]+, whicheact readily with pyridyl based ligands, but – as with 1 – the coor-inated chloride ligand of these complexes is only substituted afterreatment with Ag+ [40,41]. We initially investigated the propertiesf simple mononuclear dicationic complexes 142+, 152+, and 162+
ontaining 2,2′bpy – Fig. 7. Crystallographic studies revealed thatlthough complex 142+ shows considerable distortion away fromdeal octahedral coordination geometry [40] due to the small cav-ty size of 12-S4 complexes 152+, and 162+ are much less distorted
ith 152+ being the closet to an ideal geometry [41]. This suggestedhat the orbital overlap between the components of the complexesould be tuned through thiacrown ligand selection.
However, initial NMR studies on the three complexes sug-ested that they may not be stable in coordinating solvents as theyevealed line broadening suggestive of chemical exchange involv-ng cleavage of the Ru N bonds with 2,2′bpy [40]. Consequenttudies on 152+ showed that at temperatures below −43 ◦C theroadened signals split to give two sets of signals indicating theresence of two isomers in a ratio of 1.00:0.16. Further detailedOSY and EXSY experiments revealed that the observed exchangeroadening is, in fact, due to coordinated sulfur lone pair inver-ion – that is interconversion between three possible invertomers
Fig. 8 [41].
Furthermore, a comparison of the data for three complexesndicated that whilst the c-conformer is the most stable inver-
Fig. 7. Structures of mononuclear complexes, 142+–162+.
Fig. 8. The three possible isomers that can arise from the orientation of macrocyclicsulfur lone pair in 152+.
tomer for 142+, for complexes 152+ and 162+ the b-conformer ispreferred. DFT calculations provided a rationalization for theseobservations as they indicated that the relative stability of the threeisomeric forms are related to the number and strength of unfavor-able intramolecular contacts within the coordinated macrocycleand between the macrocycle and coordinated 2,2′bpy. Havingestablished the stability of this class of complex, we went on toinvestigate systems involving ditopic bridging ligands containingboth bidentate and monodentate coordination sites.
6. Dinuclear Ru(n-S4)-based complexes using bridgingligands with bidentate coordination sites
When dinuclear ruthenium(II) analogues were synthesized itwas found that their electrochemical and electronic propertieswere dependent on the nature of the coordinated thiacrown. Stri-kingly, CV studies reveal that the back-bonding interactions withthe tetrathiacrown leads to greater stabilization of the RuII state (byup to ∼400 mV) compared to that observed in equivalent RuII(bpy)2based systems. While the �E1/2 (120–140 mV) and hence Kc val-ues for the three complexes were similar – suggesting that the MVstates are Class II – ligand effects were revealed in a closer studyof the first oxidizations of 174+–194+ [42] – Fig. 9. A comparison ofthis first couple revealed that the most anodic is observed for 184+
while the most easily oxidized complex is 194+, with the oxidationfor 174+ being intermediate between these values; these data agreewith the optical properties of the complexes.
Complexes 174+–194+ all display RuII→bridging ligand-basedmetal-to-ligand charge-transfer, MLCT, bands. Since each of thesecomplexes has the same bridging ligand (and thus a similar LUMO)any difference in the energy of their MLCT bands can be attributedto perturbation of their metal-based HOMO. Thus, as indicated bythe electrochemical data, 194+ has the lowest energy MLCT of threecomplexes, whilst 184+ has the highest energy MLCT. To investigatesuch effects in more detail complexes 204+ and 214+ were synthe-sized [43] so that direct comparisons with 82+ could be made –Fig. 10.
Crystal structures of 204+ and 214+ showed that – at least in the
Fig. 9. Structures of dinuclear complexes, 174+–194+.
1560 J.A. Thomas / Coordination Chemistry R
bpaiccwlM
scaTt>tiiacc
csalci
242+ confirmed that these complexes take up a cis-geometry. Their
A
Fig. 10. Structures of dinuclear complexes, 204+ and 214+.
etween 82+ and its S4 analogues were in their electrochemicalroperties. Although 82+ clearly displays two RuII-based oxidationsnd a low-lying reduction of the tetrazine ring of bpta which – ass typically observed for dinuclear RuII complexes – is at very lowathodic voltages (−0.049 V vs. Ag/AgCl), for both 204+ and 214+ aomparable bpta/bpta•− reduction is not observed. What is more,hile the first of the two observed anodic couples is at an unusually
ow potential of 0.25–0.45 V, the second is just at the edge of theeCN redox window at 1.8–1.9 V.Given the �-acceptor properties of the thiacrown ligand it
eemed unlikely that this first oxidation is due to a RuII-based pro-ess, especially as this would also imply that �E1/2(1,2) for 204+
nd 214+, would take an unprecedentedly large value of over 1.5 V.herefore, it was concluded that the first oxidation in the n-S4 sys-ems is actually due to the bpta/bpta•−-based couple, shifted by0.5 V compared to previously reported RuII-based systems, andhat the second observed anodic process is actually the first metalon oxidation. These data imply that, in the conditions employedn the CV experiments and at 0 V vs. Ag/AgCl, the metal complexesre actually in the 203+ and 213+ state and contain a reduced radi-al anion bridging ligand; they also indicate that the second RuIII/II
ouple occurs at a potential beyond the MeCN redox window.When solutions of either 204+ or 214+ are placed in anaerobic
onditions at 0 V vs. Ag/AgCl distinctive changes in their absorptionpectra are observed, most notably intense bands at 600–800 nmssigned to Ru → bpta 3MLCT transitions almost completely col-
apse – Fig. 11. On application of a potential above the first anodicouple for the complexes, processes that exactly mirror image thenitial changes occur, with the original low energy bands growing
Fig. 11. Absorption changes observed when 214+ is placdapted with permission from Ref. [43]. Copyright 2006 American Chemical Society.
eviews 257 (2013) 1555–1563
back in. These observations offer further proof that under an inertatmosphere and at 0 V vs. Ag/AgCl, the complexes contain two RuII
centers, bridged by the radical anion, bpta•−.At potentials above the second oxidation observed in the CV
of 204+ the most noticeable change was, again, a reduction in theintensity of the MLCT band, but on closer inspection the growth of abroad, lower intensity band at low energy was observed. Althoughthe lower energy side of the band extends outside the range ofthe spectrometer, the band appeared to be Gaussian with a maxi-mum (ε ≈ 3700 dm3 mol−1 cm−1) centered at very low energy, ca.2850 nm (∼3500 cm−1). The energy, intensity and form of this bandare consistent with intervalence charge-transfer, IVCT. A compar-ison with the theoretical Hush model suggested that the complexis a Robin and Day Class II (valence trapped/electron hopping) sys-tem, perhaps verging on class II/III behavior. Given that the previousstudies demonstrated that 82+/3+ is a fully delocalized system, it isclear that substitution of the [RuCl([9]-ane-S3)] unit with [Ru([n]-ane-S4)] moieties results in a reduction in intermetallic coupling.The substitution of the �-donor chloride ligand with a �-acceptingthiacrown-based sulfur ligand decreases the electron density on themetal center, thus stabilizing the RuII oxidation state. This resultsin the RuII metal centers accepting more electron density from thealready electron deficient tetrazine ring of the bpta bridge, whichdestabilizes the oxidized form of the ligand with respect to the radi-cal anion form. Increased back bonding also lowers electron densityavailable for delocalization over the bridging ligand leading to alowering of intermetallic electronic coupling. Surprisingly, no IVCTcould be discerned when similar studies were carried out on 214+,suggesting that the IVCT is still further shifted outside the windowof the spectrometer used.
7. Dinuclear RuCl(n-S4)-based complexes using bridgingligands with monodentate coordination sites
With the aim of creating systems that were closer analogues ofthe original CT ion we synthesized complexes 222+–242+ in whichRuCl(n-S4) centers are bridged by pyrazine [44] – Fig. 12. NMR dataon all the three complexes and the crystal structures of 222+ and
oxidation was found to take place into two steps and again thedI1/dE vs. E deconvolutions of the CV data facilitated an analysis ofthe data: as for complexes 174+–194+, the first RuIII/II couple for all
ed in an inert atmosphere at +0.00 V vs. Ag/AgCl.
J.A. Thomas / Coordination Chemistry R
ttKc2sbcao
taWfhtatmci
�t–Iod
the corresponding figures are −0.588 and −0.588; although these
R
Fig. 12. Structures of dinuclear complexes, 224+–244+.
he complexes displays good reversibility, while the second oxida-ion is not completely chemically reversible – Fig. 13. Calculatedc values obtained from these data suggest that electrochemicaloupling within 233+ (Kc = 75) is slightly higher than for 223+ and43+ (Kc = 35 for both), however – like the 9-S3-based mixed donoret systems 62+–82+ – the apparent electrochemical interactionetween the metal centers of the pyrazine bridged systems areonsiderably lower than those reported for N-donor systems suchs the Creutz–Taube ion, where Kc > 106; but, again like 63+–83+,ptical studies revealed contrasting results.
The first one electron oxidation of all three complexes results inhe growth of characteristic IVCT bands whose energy and shapere dependent on the identity of the coordinated thiacrown ligand.hile the IVCT bands for 233+ and 243+ are symmetrical, the band
or 223+ is asymmetric with narrower half bandwidths on theigh-energy side of the IVCT. Although IVCT band asymmetries areypical of the Class II/III interface where solvent interactions areveraged, these are observed at low energy [2]. Therefore it seemshat – like the trinuclear systems discussed above – the asymmetry
ay arise from multiple IVCTs produced from Kramer doublets. Aomparison of the IVCT data with the Hush model revealed dispar-ties between experimental and calculated figures.
Complex 223+ is the only one of the three systems with a�1/2(calc) larger than the experimentally derived figure, indeed
he value of ��1/2(expt) is probably slightly overestimated becauseas outlined above – it is probably composed of several overlapping
VCTs. Contrastingly, for both 233+ and 243+ the experimentallybserved IVCT bands are appreciably wider than theoretical pre-ictions, with this discrepancy being larger for 233+ (∼20%). This
Fig. 13. Metal based oxidation couples for 212+. Cyclic voltammoeprinted with permission from Ref. [44]. Copyright 2008 American Chemical Society.
eviews 257 (2013) 1555–1563 1561
analysis suggests that although 223+ is a Class III (valence delocal-ized), both 233+ and 243+ are valence localized Class II systems;it also indicates that – as in their RuCl(9-S3) analogues – thelow charge and combination of coordinated �-acceptor and �-donor ancillary ligands results in higher electronic coupling in theMV state than expected from the electrochemical data. Neverthe-less, electronic coupling in these systems is still lower than thatobserved in the original CT-ion, which is a consequence of the�-accepting ability of the thiacrowns. Not only are the RuII oxida-tion couples for 223+–243+ anodically shifted compared to the CTion, they are shifted compared to systems that incorporate typical�-accepting pyridyl-based ligands. Clearly, by accepting electrondensity from the metal centers, the coordinated thiacrown ligandslower delocalization across the bridging pyrazine; leading to anincreased electron-transfer barrier and decreased electron transferrate. To further investigate the experiment differences in electroniccoupling between thiacrown-based systems computational studieson selected complexes were carried out.
8. Computational studies
Using experimental X-ray data as a starting point, complexes82+, 174+, 222+, and 242+ were fully optimized and compared at theDFT (B3LYP) level of theory [44]. As would be expected, for all thecomplexes, the HOMOs are out-of-phase metal centered orbitalsand the LUMOs are �* orbitals localized on the bridging ligand –Fig. 14. The most obvious difference between the systems occurson replacing monodentate bridging ligands of 222+ and 242+ withthe bidentate bridges of 82+, 174+ which leads to a >20% increasedparticipation of the metal ion the LUMO of the complex. A closeranalysis also reveals differences in MO compositions due to thenature of the thiacrown.
A comparison of 222+ with 242+ shows that exchanging coor-dinated 12-S4 with 16-S4 causes a ∼17% to ∼13% decrease in themetal ion participation in the LUMO, furthermore the calculatedruthenium charges for 222+ are −0.600 and −0.602, whilst for 242+
differences are small they suggest that the metal centers of 222+
are slightly more electron rich than those of 32+. DFT optimizationswere also carried out on the 222+ and 242+ MV states, and the optical
gram (fine line) and dI1/dE vs. E deconvolutions (bold line).
1562 J.A. Thomas / Coordination Chemistry Reviews 257 (2013) 1555–1563
Fig. 14. HOMOs (left) and LUMOs (right) of complexes 222+, 242+, 82+, and 174+.Adapted with permission from Ref. [44]. Copyright 2008 American Chemical Society.
ns ofA y.
pt
cbdttlicF
Fig. 15. Kohn–Sham molecular orbitals involved in the electronic transitiodapted with permission from Ref. [44]. Copyright 2008 American Chemical Societ
roperties of the isovalent and MV systems were then investigatedhrough TD-DFT.
For both the isovalent states, the two strongest low-energy cal-ulated transitions are RuII → pyrazine MLCT-based absorptionsetween 400 and 515 nm and, whilst the experimental data onlyisplays a single absorption, it is consistent with an imposition ofhe calculated transitions. Strikingly, the calculated TD-DFT spec-rum for the corresponding MV states leads to a different much
ower energy excitation. As shown in Fig. 15, this transition, whichnvolves ruthenium-based orbitals on both metal centers, is clearlyonsistent with the postulated metal-to-metal charge-transfer.urther evidence of charge delocalization came from a Mulliken
isovalent 212+ (left) and its oxidized mixed valence analogue 213+ (right).
Population Analysis of spin densities: for both complexes the spindensities at each ruthenium center were equal. This latter analysisis slightly inconsistent with the experimental data indicating thatonly 213+ is fully delocalized, however it is known that DFT overesti-mates delocalization of the unpaired electron due to self-repulsioneffects [44].
9. Conclusions and future work
Due to their characteristic donor capabilities, thiacrown ligandsproduce distinctive effects on coordinated RuII centers, with bothelectronic and chemical properties being modulated relative to
istry R
NrTiapstalaaturla
rlcshtt
A
wfofiPR
R
[
[[[[[
[
[[
[
[
[
[
[[
[[[[[[[[[
[[
[
[
[
[
[
[
J.A. Thomas / Coordination Chem
-donor ligand sets. The electrochemical studies outlined in thiseview confirm that these macrocycles are strongly �-accepting.his can lead to dinuclear complexes in unusual redox states, asllustrated by the facile generation of the radical anion bridged 203+
nd 213+ systems. The studies on the S4 thiacrown-based MV com-lexes also illustrate how the nature of the coordinated ligand canubtly modulate the electrochemical and electronic properties ofhese systems. While the strong back bonding properties of the thi-crown ligands generally lower delocalization across the bridgingigand, perhaps the most interesting effects on electronic couplingrise in the lower charged complexes containing a mixture of thi-crown and �-donor ligands; for both S3 and S4 based complexeshese conditions produce systems in which electrochemical datanderestimates the true extent of electronic coupling. In futureeports the effect of coordination to non-innocent redox activeigands in mono- and oligonuclear complexes involving relate thi-crown based centers will be explored.
Electronic effects involving coordinated 9-S3 also explain whyuthenium(II) centers coordinated to this ligands are anomalouslyabile for low-spin d6-systems. This effect has been exploited toonstruct trinuclear supramolecular bowls possessing four acces-ible redox states. The distinctive connectivities in these bowlsave resulted in MV states with unique electron-transfer proper-ies. Studies on the host-guest properties of these systems will formhe basis of forthcoming reports.
cknowledgements
My interest in MV systems was sparked through my PhD workith Prof. Jon McCleverty and Dr. Chris J. Jones. I thank them both
or planting the seeds of inspiration. The quality of the researchutlined herein has been greatly enhanced through extremely fruit-ul collaborations with the groups of Prof. Mike D. Ward heren Sheffield and Prof. Vitor Felix at the Universidade de Aveiro,ortugal. I gratefully acknowledge the financial support from Theoyal Society, The British Council, Santander Bank, and EPSRC.
eferences
[1] (a) C. Creutz, H. Taube, H.J. Am, Chem. Soc. 91 (1969) 3988;(b) C. Creutz, H. Taube, H.J. Am, Chem. Soc. 95 (1973) 1086.
[2] (a) Selected reviews: C. Creutz, Prog. Inorg. Chem. 30 (1) (1983);(b) R.J. Crutchley, Adv. Inorg. Chem. 41 (1994) 273;(c) W. Kaim, A. Klein, M. Glöckle, Acc. Chem. Res. 33 (2000) 755;(d) K.D. Demadis, C.M. Hartshorn, T.J. Meyer, Chem. Rev. 101 (2001) 2655;(e) B.S. Brunschwig, C. Creutz, N. Sutin, Chem. Soc. Rev. 31 (2002) 168;(f) J.T. Hupp, in: T.J. Meyer, J.A. McCleverty (Eds.), Comprehensive CoordinationChemistry II, vol. 2, Elsevier, Oxford, 2004, p. 709;(g) D.M. D’Alessandro, F.R. Keene, Chem. Soc. Rev. 35 (2006) 424;(h) D.M. D’Alessandro, F.R. Keene, Chem. Rev. 106 (2006) 2270;(i) W. Kaim, B. Sarkar, Coord. Chem. Rev. 251 (2007) 584;(j) W. Kaim, G.K. Lahiri, Angew. Chem. Int. Ed. 46 (2007) 1778.
[3] M.D. Ward, Chem. Soc. Rev. 24 (1995) 121.
[4] K. Prassides (Ed.), Mixed Valency Systems—Applications in Chemistry, Physics
and Biology, Kluwer Academic Publishers, Dordrecht, 1991.[5] V. Balzani, A. Credi, M. Venturi, Molecular Devices and Machines: Concepts and
Perspectives for the Nanoworld, Wiley-VCH, Weinheim, 2003.[6] (a) S.B. Braun-Sand, O. Wiest, J. Phys. Chem. B 107 (2003) 9624;
[[
[
eviews 257 (2013) 1555–1563 1563
(b) H. Tannai, K. Tsuge, Y. Sasaki, Inorg. Chem. 44 (2005) 5206.[7] J. Otsuki, T. Akasaka, K. Araki, Coord. Chem. Rev. 252 (2008) 32.[8] J.R. Bolton, N. Mataga, G. McLendon (Eds.), Electron Transfer in Inorganic,
Organic, and Biological Systems, Advances in Chemistry Series, 228, AmericanChemical Society, Washington, DC, 1991.
[9] J.P. Launay, Chem. Soc. Rev. 30 (2001) 386.10] See for example: A.J. Blake, G. Reid, M. Schröder, J. Chem. Soc. Dalton Trans.
(1989) 1675.11] C. Landgrafe, W.S. Sheldrick, J. Chem. Soc. Dalton Trans. (1994) 1885.12] S. Roche, H. Adams, S.E. Spey, J.A. Thomas, Inorg. Chem. 39 (2000) 2385.13] I.R. Young, L.A. Ochrymowycz, D.B. Rorabacher, Inorg. Chem. 25 (1986) 2576.14] K.K. Brandt, W.S. Sheldrick, J. Chem. Soc. Dalton Trans. (1996) 1237.15] (a) F. Laurent, E. Plantalech, B. Donnadieu, A. Jiménez, F. Hernández, M.
Martínez-Ripoll, M. Biner, A. Llobet, Polyhedron 18 (1999) 3321;(b) X. Sala, I. Romero, M. Rodríguez, A. Llobet, Inorg. Chem. 43 (2004) 5403.
16] (a) G.E.D. Mullen, M.J. Went, S. Wocaldo, A.K. Powell, P.J. Blower, Angew. Chem.Int. Ed. Engl. 36 (1997) 1205;(b) G.E.D. Mullen, T.F. Fassler, M.J. Went, K. Howland, B. Stein, P.J. Blower, J.Chem. Soc. Dalton Trans. (1999) 3759.
17] S. Roche, L.J. Yellowlees, J.A. Thomas, Chem. Commun. (1998) 1429.18] C.S. Araújo, M.G.B. Drew, V. Félix, L. Jack, J. Madureira, M. Newell, S. Roche, T.M.
Santos, J.A. Thomas, L. Yellowlees, Inorg. Chem. 41 (2002) 2250.19] (a) C.H. Braunstein, A.D. Baker, T.C. Strekas, H.D. Gafney, Inorg. Chem. (23)
(1984) 857;(b) R.R. Ruminski, T. Cockroft, M. Shoup, Inorg. Chem. 27 (1988) 4026.
21] (a) J.E.B. Johnson, C. de Groff, R.R. Ruminski, Inorg. Chem. Acta 187 (1991) 73;(b) J. Poppe, M. Moscherosch, W. Kaim, Inorg. Chem. 327 (1993) 2640.
22] (a) S. Ernst, V. Kasack, W. Kaim, Inorg. Chem. 27 (1988) 1146;(b) W. Kaim, S. Ernst, V. Ksack, J. Am. Chem. Soc. 112 (1990) 173.
23] B.D. Yeomans, D.G. Humphrey, G.A. Heath, Dalton Trans. (1997) 4153.24] F. Barriere, N. Camire, W.E. Geiger, U.T. Mueller-Westerhoff, R. Sanders, J. Am.
1429.34] See: R.H. Fish, Coord. Chem. Rev. 185–186 (1999) 569, and references therein.35] N. Shan, S.J. Vickers, H. Adams, M.D. Ward, J.A. Thomas, Angew. Chem. Int. Ed.
43 (2004) 3938.36] (a) M.A. Bennett, L.Y. Goh, A.C. Willis, J. Chem. Soc. Chem. Commun. (1992)
1180;(b) K. Brandt, W.S. Sheldrick, J. Chem. Soc. Dalton Trans. (1996) 1237.
37] N. Shan, J.D. Ingram, T.L. Easun, S.J. Vickers, H. Adams, M.D. Ward, J.A. Thomas,Dalton Trans. (2006) 2900.
38] (a) J.-C. Imbeau, J.-M. Saveant, J. Electroanal. Chem. 44 (1973) 169;(b) A.J. Bard, L.R. Faulkner, Electrochemical Methods. Fundamentals and Appli-cations, 2nd ed., Wiley, New York, USA, 2001.
39] (a) S. Kar, T.A. Miller, S. Chakraborty, B. Sarkar, B. Pradhan, R.K. Sinha, T. Kundu,M.D. Ward, G.K. Lahiri, Dalton Trans. (2003) 2591;(b) S. Patra, B. Sarkar, S. Ghumaan, J. Fiedler, W. Kaim, G.K. Lahiri, Dalton Trans.754 (2004);(c) D.M. D’Alessandro, M.S. Davies, F.R. Keene, Inorg. Chem. 45 (2006) (1656);(d) C.S. Grange, A.J.H.M. Meijer, M.D. Ward, Dalton Trans. 39 (2010) 201.
40] T.M. Santos, B.J. Goodfellow, J. Madureira, J. Pedrosa de Jesus, V. Felix, M.G.B.Drew, New J. Chem. 23 (1999) 1015.
41] H. Adams, A.M. Amado, V. Félix, B.E. Mann, J. Antelo-Martinez, M. Newell, P.J.A.Ribeiro-Claro, S.E. Spey, J.A. Thomas, Chem. Eur. J. 11 (2005) 2031.
42] M. Newell, J.A. Thomas, Dalton Trans. (2006) 705.43] M. Newell, J.D. Ingram, T.L. Easun, S. Vickers, H. Adams, M.D. Ward, J.A. Thomas,
Inorg. Chem. 45 (2006) 821.44] H. Adams, P.J. Costa, M. Newell, S.J. Vickers, M.D. Ward, V. Félix, J.A. Thomas,
