DaltonTransactions

Dynamic Article Links

Cite this: Dalton Trans., 2012, 41, 12577

www.rsc.org/dalton PAPER

Anion dependence in the spin-crossover properties of a Fe(II) podandcomplex†

Christina M. Klug, Ashley M. McDaniel, Stephanie R. Fiedler, Kelsey A. Schulte, Brian S. Newell andMatthew P. Shores*

Received 6th June 2012, Accepted 15th August 2012DOI: 10.1039/c2dt31213a

We report the syntheses, characterisations, and spin state behaviours of salts of the tripodal-ligated Fe(II)complex [FeL6-OH]X2 (L

6-OH = tris{4-[(6-methanol)-2-pyridyl]-3-aza-3-butenyl}amine, X = OTf− (1),Br− (2), I− (3), BPh4

− (4)). Covalent linking of the ligand arms is imperative as a high-spin bis(tridentate)complex (5) is formed when a non-tethered ethyl iminopyridine ligand (L2 = 4-[(6-methanol)-2-pyridyl]-3-aza-3-butenyl) is used. For salts 1–4, thermally-induced spin-crossover (SCO) is observed in the solidstate, with dependence on anion and solvate molecules. Salts with larger anions show more completeSCO centred at higher temperatures (1 > 3 > 2); the triflate salt 1 (T1/2 = 173 K) also shows the strongestcooperativity of the compounds examined. Hydrogen bonding appears to be critical to SCO in this familyof salts: limiting interactions by use of tetraphenylborate produces a high-spin complex down to 5 K. Inprotic solvents such as methanol, spectra of [FeL6-OH]2+ are largely unchanged over a period of threedays, but dissociate when interrogated with strong field bidentate ligands. Compounds 1–3, and 5 remainhigh spin in solution down to 180 K, consistent with the data obtained in the solid state.

Introduction

Octahedral coordination of d4–d7 transition metal centres com-bined with a properly tuned ligand field can facilitate switchingbetween high- and low-spin states (HS and LS, respectively).1–5

This spin-crossover (SCO) phenomenon can be triggered bysmall environmental forcings (E ∼ kBT), and exploitation of thismolecular effect in materials is a major synthetic goal. SCOmaterials have been widely investigated for use in data storageand display devices due to the dramatic magnetic and colourchanges that can accompany spin-state switching.3,5,6

Recent inquiries have focused on combining host–guest inter-actions with the SCO phenomenon, both to probe non-covalentinteractions as well as to develop new signalling pathwaysfor chemosensing.7–14 In the solid state, inclusion of guestmolecules into metal–organic frameworks such as[Fe(bpbd)2(NCS)2] and [Fe2(azpy)4(NCS)4] have been shown toinduce SCO.7,8 The presence and identity of counter anions havealso been shown to have an impact on the spin transition

temperature in both the solid state and in solution.3,9 Investi-gation of a series of Fe(II) triimidazole–halide complexes haveshown SCO in the solid state: large anions with low chargedensity induce SCO at higher temperatures than their smaller,more charge dense counterparts.10 Previous studies using 2,2′-bi-1,4,5,6-tetrahydropyrimidine (H2bip) in homo-11,12 andheteroleptic13–15 complexes with Fe(II), have also shown anion-dependent spin-state switching both in solution and in the solidstate.

In order to exploit the colour change induced by SCO for usein anion sensors, more robust ligand sets must be procured toovercome the lability associated with HS Fe(II) complexes featur-ing bidentate ligands.16 Thus, we have become interested in tri-podal Schiff base ligands based on tris{4-[(6-R)-2-pyridyl]-3-aza-3-butenyl}amine (H3−x(6-R-py)xtren (R = H, Me, X =1–3)).17,18 The parent Fe(II) complexes have been shown toundergo SCO in solution and the solid state, with the SCO temp-erature dependent on the R substituent at the 6-position of thepyridine: steric hindrance associated with substitution at thecarbon adjacent (α) to the pyridyl N donor stabilizes the HSstate, lowering the SCO temperature in the solid state and insolution with each additional substitution.17

More recent work on Fe(II) complexes with pyrazole-contain-ing tripodal ligands has shown anion dependence in the solidstate SCO temperatures and completeness; SCO in solution hasalso been observed.19 Related, hydrogen-bonding interactions inmixed anion–cation salts of Fe(II) complexes with analogousimidazole-containing ligands have been shown to affect the per-centage of HS moieties at room temperature.20

†Electronic supplementary information (ESI) available: Full renderingsof X-ray structures, information for pertinent intermolecular interactions(plots and tables), magnetic data upon grinding for 2 and 2·solv, syn-thesis and characterization of [FeL6-OH](BPh4)1.75Br0.25·MeCN, NMRspectra for 1, 2, 3, and 5, and solution magnetic data for 1, 2, 3, and 5.CCDC 865553–865555, 866021, 866022, 884193, 892044 and 892318for structures listed in Table 2. For ESI and crystallographic data in CIFor other electronic format see DOI: 10.1039/c2dt31213a

Department of Chemistry, Colorado State University, Fort Collins,CO 80523-1872, USA. E-mail: matthew.shores@colostate.edu;Fax: +01-970-491-1801; Tel: +01-970-491-7235

This journal is © The Royal Society of Chemistry 2012 Dalton Trans., 2012, 41, 12577–12585 | 12577

Dow

nloa

ded

by U

nive

rsity

of

Mas

sach

uset

ts -

Am

hers

t on

04 O

ctob

er 2

012

Publ

ishe

d on

07

Sept

embe

r 20

12 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C2D

T31

213A

View Online / Journal Homepage / Table of Contents for this issue

As part of our efforts to combine spin-state changes withanion binding in environmentally relevant media, we have inves-tigated the preparations and magnetic properties of salts of theFe(II) complex with tris{4-[(6-methanol)-2-pyridyl]-3-aza-3-butenyl}amine (L6-OH, Fig. 1) and the non-tethered derivative4-[(6-methanol)-2-pyridyl]-3-aza-3-butenyl (L2, Scheme 1). Thetripodal complex was chosen for several reasons. First, hexaden-tate coordination should increase complex solution stability,especially in protic solvents.16 Second, electrostatic attractions ofanions to the cationic complex trigonal pocket would beenhanced by incorporation of hydrogen-bonding methanolgroups at the 6(α) position. The trigonal pocket formed by chela-tion of Fe(II) may allow for stronger binding and better selectiv-ity for C3 symmetric and/or larger spherical anions compared tobidentate chelation. Combining these factors, tridentate anioninteractions could force the Fe(II) ion to adopt a LS state due tothe contraction of the trigonal pocket. Meanwhile, structural andphysical influences of the bridgehead nitrogen could be investi-gated by removal of the tether, via study of Fe(II) complexeswith L2.

Results and discussion

Syntheses

The ligand L6-OH is obtained by a condensation reaction of2-pyridinecarboxaldehyde-6-methanol21 with tris(2-amino-ethyl)-amine (tren) in methanol at room temperature. The oily ligand is

used with minimal purification, and pure metal-containing pro-ducts are obtained via crystallization. As shown in Scheme 1, theligand is combined with [Fe(MeCN)2(OTf)2]

22 or FeBr2 inmethanol to produce the triflate (1) or bromide (2) salts, respect-ively; subsequent anion exchanges and crystallizations produceanalytically pure samples of the iodide (3) and tetraphenylborate(4) salts. The triflate anion was chosen because of its C3 sym-metry, potentially enabling greater interaction with the trigonalpocket formed by the ligand. Meanwhile, the halides bromideand iodide allow us to explore the effect of anion size andpacking interactions on SCO temperature. The use of tetraphenyl-borate aims to demonstrate the complex cation’s solid-state prop-erties in the absence of significant hydrogen-bondinginteractions.

X-ray quality crystals of 1–3 are obtained by diethyl ether dif-fusion into saturated alcohol solutions of the compounds; crys-tals of 2·solv are formed by diethyl ether diffusion into a 50 : 50ethanol–methanol solution of 2. Based on elemental analysis and1H NMR spectroscopy, a small amount of ethanol co-crystalliseswith the triflate salt to produce 1, but it is not evident in thecrystal structure (vide infra). X-ray quality crystals of 4 are pro-duced by ether diffusion into acetonitrile. Elemental analysesand crystallography data for compound 4 indicate that approxi-mately two acetonitrile molecules co-crystallise with the salt.

The non-tethered derivative of the L6-OH ligand is synthesisedby condensation of ethylamine and 2-pyridinecarboxaldehyde-6-methanol to produce L2. This combines with [Fe(MeCN)2-(OTf)2] to produce the bis(tridentate) complex 5. X-ray qualitycrystals are obtained by diethyl ether diffusion into an ethanolicsolution of the complex. Regardless of reactant stoichiometry(2 : 1 or 3 : 1 ligand–metal), crystals of [Fe(L2)2](OTf)2 areisolated.

Anion dependence in solid state magnetic properties

The variable temperature solid-state magnetic susceptibilities forground crystals of 1–4 are shown in Fig. 2. At room temperature,all salts are HS, with χMT values typical for high-spin octahedralFe(II) complexes (S = 2). For the triflate salt 1, χMT is 3.63 cm3

K mol−1 at 295 K. Upon cooling, a sharp decrease in χMTbegins at 178 K and is virtually complete by 150 K (T1/2 =173 K), ultimately giving a fully LS sample with χMT equal to0.06 cm3 K mol−1 at 4 K. This salt undergoes the highest andsharpest spin transition observed for all the salts studied inthis work. Note that this occurs at 40 K less than the parent[Fe(6-Me-py)3tren](PF6)2 complex reported by Drago and co-workers.17

Similar to 1, the bromide salts 2 and 2·solv show SCO, albeitat a lower temperature. The anion appears to be the major contri-butor to spin-state properties since both 2 and 2·solv show SCOat approximately the same temperature (“T1/2” = 113 and 111 K,respectively‡). However, the presence of solvate molecules influ-ences the completeness of SCO behaviour. In the absence ofsolvate molecules, the SCO is more complete for 2 as χMTdecreases to 0.65 cm3 K mol−1 at 85 K and 0.26 cm3 K mol−1 at

Fig. 1 Depiction of the complex cation: hexadentate chelation aims toincrease stability, while positioning of hydroxyl groups tunes SCO prop-erties and entices anion–cation hydrogen-bonding interactions.

Scheme 1 Reaction sequence for preparing [FeL6-OH]2+ and [Fe-(L2)2]

2+ complex salts; a: 3.1 equiv 2-pyridinecarboxaldehyde-6-metha-nol +1 equiv tren; b: 1 equiv of 2-pyridinecarboxaldehyde-6-methanol+1 equiv ethylamine.

‡Since SCO is incomplete for 2 and 2·solv, “T1/2” is defined as thetemperature of the average χMT value before and after SCO.

12578 | Dalton Trans., 2012, 41, 12577–12585 This journal is © The Royal Society of Chemistry 2012

Dow

nloa

ded

by U

nive

rsity

of

Mas

sach

uset

ts -

Am

hers

t on

04 O

ctob

er 2

012

Publ

ishe

d on

07

Sept

embe

r 20

12 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C2D

T31

213A

View Online

5 K. In contrast, the partially solvated compound 2·solv showsincomplete SCO: χMT is 1.20 cm3 K mol−1 at 85 K, and dropsto 0.66 cm3 K mol−1 at 5 K. The interplay is complex, since wealso find for 2 that varying sample grinding conditions changesSCO completeness (Fig. S4†). Such impacts on Fe(II) SCO prop-erties have been observed elsewhere.23,24 Regardless of theorigins, a comparison of the data for both compounds indicatessignificant intermolecular effects are operative.

Meanwhile, the iodide salt 3 undergoes complete SCO (T1/2 =145 K), albeit more gradually than 1. The χMT value at 295 K of3.52 cm3 K mol−1 gradually decreases to 3.20 cm3 K mol−1 at198 K, then decreases more sharply to 0.44 cm3 K mol−1 at100 K, and levels out to 0.08 cm3 K mol−1 at 5 K.

The tetraphenylborate salt 4 remains HS throughout the temp-erature range probed. A χMT value of 3.81 cm3 K mol−1 at295 K decreases slightly to 3.27 cm3 K mol−1 by 29 K; a down-turn in χMT to 2.07 cm3 K mol−1 at 5 K is ascribed to zero-fieldsplitting of the HS complex.

Similarly, magnetic susceptibility data acquired at 120 K and5 K (3.38 and 2.50 cm3 K mol−1, respectively) indicate that thecomplex in 5 remains HS. The slight decrease in χMT is attribu-ted to zero-field splitting of the HS compound.

Irrespective of structural considerations, a few trends can begleaned from the magnetic susceptibility data. The largest coun-teranions capable of hydrogen bonding (OTf−, I−) favour higher

SCO temperatures as well as more complete transitions. Decreas-ing the size of the anion appears to stabilise the high spin state,as indicated by lower T1/2 values and incomplete conversion inthe case of 2. For the same anion, solvate molecules appear toaffect the degree of completion of SCO. Disruption of hydrogen-bonding interactions also appears to stabilize HS states as seenwith 4. Whereas the magnetic behaviour of the iodide salt 3 canbe considered spin equilibrium, those of 1 and 2 show strongercooperativities. Finally, iminopyridine groups do not offer suffi-cient ligand field strength to overcome the weak field generatedby direct coordination of hydroxyl groups, as exemplified in themagnetic behaviour of 5.

Structural studies

To better understand the subtle differences in magnetic proper-ties, single-crystal X-ray diffraction analyses were performed oncompounds 1, 2, 2·solv, 3, 4, and 5. Relevant bond lengths,angles, and distortion parameters are provided in Table 1; detailsare provided in the ESI.† All cationic complexes (except 5)show the same basic features; the complex in the room temp-erature structure of the triflate salt 1·RT is shown in Fig. 3 as arepresentative example. The first coordination sphere of the Fe–

Fig. 2 Temperature dependence of χMT for 1–4. Lines connecting datapoints are guides to the eye.

Table 1 Selected bond distances (Å) and distortion parameters (°)

Fe–Nim Fe–Npy Fe–Nbridge Fe⋯Feb Σ θ

1·LT 1.948(2) 2.098(2) 3.723(2) 10.937(1) 80.5(3) 119.51·RT 2.147(4) 2.331(4) 3.187(3) 11.323(4) 115.5(3) 196.62 2.128(3) 2.319(7) 3.192(4) 9.0074(3) 107.8(2) 195.32aa 2.151(3) 2.322(3) 3.241(1) 8.8440(1) 114.7(2) 196.63·LT 1.967(4) 2.112(4) 3.537(1) 8.8845(5) 85.0(3) 129.53·RT 2.134(4) 2.325(4) 3.167(1) 9.1254(5) 112.9(3) 195.44 2.159(3) 2.277(3) 3.244(2) 9.327(4) 125.4(2) 196.65c 2.184(2) 2.093(2) c 8.0556(5) 140.1(1) 441.7

a 2a = 2·solv; b Shortest contact in structure; c dFe–OH = 2.174(1) Å.

Fig. 3 Crystal structure of the cationic complex in the structure of1·RT, rendered with 40% thermal ellipsoids. Green, red, blue and greyellipsoids represent Fe, O, N and C atoms, respectively. Non-hydroxyl Hatoms and disordered components have been omitted for clarity.

This journal is © The Royal Society of Chemistry 2012 Dalton Trans., 2012, 41, 12577–12585 | 12579

Dow

nloa

ded

by U

nive

rsity

of

Mas

sach

uset

ts -

Am

hers

t on

04 O

ctob

er 2

012

Publ

ishe

d on

07

Sept

embe

r 20

12 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C2D

T31

213A

View Online

podand complexes comprises six nitrogen atoms, from threeimine and three pyridine groups, in a distorted octahedral geo-metry. Comparing the structures, the average Fe–Nimine bondlengths tend to be shorter than the average Fe–Npyridine distancesfor both spin states, a trend seen in other podand-type com-plexes.17 The average Fe–N bond lengths for 1·RT, 2, 2·solv,3·RT, 4, and 5 are consistent with HS Fe(II) ions, while 1·LT and3·LT are LS,3 as expected from the magnetic susceptibility data.Additionally, the bridgehead nitrogen (N7) does not coordinateto the iron centre; the closest contact observed in all of the struc-tures presented here is 3.167(1) Å for 3·RT. The structural distor-tion parameters Σ25–27 and θ27,28 for 1·RT, 2, 2·solv, 3·RT, 4,and 5 are indicative of large distortion away from perfect octa-hedral coordination, similar to HS Fe(II) complexes containingbidentate ligands. Structures 1·LT and 3·LT show smaller octa-hedral and trigonal distortion values, consistent with conversionto LS configurations.

Comparing high- and low-temperature structures for 1 and 3,no crystallographic phase transitions are observed upon cooling,as indicated by retention of the same space group at 296 K and120 K; unit cell volumes contract by approximately 6% and 5%for 1 and 3, respectively, as expected.

Characteristic of this ligand type,18,19,29 the Fe–N7 distanceshows strong dependence on the spin state (Fig. 4): in the LSstate, the bridgehead nitrogen adopts a quasi-planar confor-mation, pointing slightly away from the Fe(II) centre; uponwarming it assumes a more pyramidal geometry, pointingtowards the metal. The conformation change is a consequence ofthe Fe–N bonds expanding upon the LS → HS conversion. Thechange is less pronounced for the iodide salt 3, as conversion tothe fully LS state is not complete at 120 K.

For the structures of 1–3, the immediate anion–cation inter-actions of each salt are not identical, but comparable: twohydroxyl groups on the same complex (containing O1 and O2)chelate one of the anions, while the third hydroxyl (O3) contortsout of the plane of the pyridine. This facilitates better interactioneither with the second anion (1·RT, 2 and 3) or the bridgeheadnitrogen of an adjacent cation (1·LT). Details of anion–cationinteractions are provided in the ESI.†

Since the local interactions are comparable for 1, 2 and 3, yetthe magnetic properties are quite distinct, more detailed investi-gations of packing and other subtle intermolecular interactionsare warranted. We note a trend between increasing (shortest)

Fe⋯Fe distances (2·solv ≈ 2 < 3·RT < 1·RT, Table 1) andincreasing SCO T1/2 and cooperativity. This is consistent withthe general notion that anion–cation interactions are critical, andsuggests that larger anions may provide more lattice flexibilityfor Fe–ligand bond contraction upon SCO.

Comparing the structures of the bromide salts, small differ-ences in interactions may give some insight into differences inthe completeness of SCO, with the caveat that the magneticproperties have already been shown to be affected by grinding(i.e. bulk interactions may be more important than localinteractions). Although the average Br⋯O distances are shorterin 2·solv than 2 (3.22 versus 3.27 Å respectively§), additionalhydrogen-bonding interactions occur between the anions and thesolvent in 2·solv. Also, complex packing in 2·solv preventshydrogen-bonding interactions between cations, with a columnof solvent molecules parallel to the c axis in 2·solv (Fig. S13†).Similar to what was observed with the spin-state properties ofpartially solvated [(H2bip)2Fe(pic)]Br2 salts,13 we propose thatadditional solvent–anion interactions reduce the strength ofanion–cation interactions, effectively separating metal centresfrom each other and resulting in incomplete SCO. The effectwould be similar to that of exhaustive grinding: smaller particleswith more surface-site Fe(II) ions are less connected to neigh-bouring Fe(II) ions and have higher barriers to SCO, thus stabi-lizing the HS state.23

Comparing the two salts that unambiguously undergo com-plete SCO (1 and 3), the packing of cations in 3 (Fig. S14†) gen-erates unique intermolecular interactions: adjacent cations showhydrogen bonding through the hydroxyl groups containing O1and O2. This interaction lengthens the O2⋯I− contact(Table S1†), making the distance between those atoms greaterthan the sum of their van der Waals radii. Through this contact,it is plausible that if one Fe(II) centre switches from HS to LS, anadjacent cation may be compelled to remain HS to maintain thehydrogen bond. The combination of weaker anion chelation andcation–cation hydrogen-bonding may inhibit productive coopera-tivity between metal centres, resulting in a more gradual SCOprocess.

Meanwhile, cation–cation interactions are also observed in thestructures of 1, as mentioned earlier, but in this case they appearto enhance SCO cooperativity (Fig. 5). The cations of 1·RTpack in a head-to-toe fashion, with the hydroxyl groups all inter-acting with triflate anions. Upon cooling, the hydroxyl group(O3) does not interact with a triflate, but shifts so that it has asignificant interaction with the bridgehead nitrogen of a neigh-bouring complex (Fig. 5, right; also Fig. 4 and Fig. S21–22†).Combined with spin-state induced Fe–N distance changes in thepodand ligand, the temperature-dependent changes in packingevoke a “push–pull” type mechanism: the SCO experienced bythe Fe(II) centre pushes the bridgehead amine out while thehydroxyl orientation changes and subsequent acid–base inter-action pulls on the amine. Although the observed interactionsmay be an artefact of the molecular packing arrangement, theyhighlight a potentially critical spin-state directing role for theunbound (to Fe) bridgehead amine group, one that should beconsidered when seeking to control spin-state changes inFig. 4 Overlay of room temperature (red) and low temperature (blue)

structures for the cations in 1 (a) and 3 (b). All hydrogen atoms and thedisordered components of 1·RT have been omitted for clarity. The circleemphasizes the orientation of O3 in the HS and LS states for 1. §Average values are not weighted for disorder.

12580 | Dalton Trans., 2012, 41, 12577–12585 This journal is © The Royal Society of Chemistry 2012

Dow

nloa

ded

by U

nive

rsity

of

Mas

sach

uset

ts -

Am

hers

t on

04 O

ctob

er 2

012

Publ

ishe

d on

07

Sept

embe

r 20

12 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C2D

T31

213A

View Online

solution with this family of complexes. Note that a previousstudy has demonstrated that pyramidalisation of the tren back-bone via alkylation converts a Fe(II) SCO complex to a LSspecies.30

In comparison, strong anion–cation interactions are notobserved in the structure of 4. The large size of tetraphenylborateprevents the anion from packing within the trigonal pocket,allowing cation–cation hydrogen bonding between O1 and O2 ofadjacent complexes. The cations pack in a toe-to-toe fashion(Fig. S23†). Some subtle OH⋯BPh4

−–π interactions are appar-ent (Fig. S23, Table S1†); however, unlike the other podandstructures with pocket-packing anions, the anion does not pullligand hydroxyl groups and their associated arms towards oneanother. The lack of significant anion–cation interactions and thepresence of cation–cation interactions similar to those seen for 3apparently contribute to the lack of SCO observed in this salt.

In the absence of a tether between the ligand arms, two imino-pyridine-hydroxyl ligand arms coordinate in a meridonal triden-tate fashion to produce 5 (Fig. 6). The coordination environmentof the iron(II) centre is different from the tethered ligand set, asoxygen and nitrogen atoms combine to give a N4O2 first coordi-nation sphere. The Fe–Nim and Fe–O bond lengths (Table 1) areconsistent with a HS species, in accordance with the magnetic

susceptibility data. Interestingly, the Fe–Npy distances are actu-ally more comparable to the LS podand structures; however,these shorter contacts are likely due to steric constraints imposedby the meridonal coordination mode, and is common for planar-terpyridine type ligands.31 The “octahedral” coordinationenvironment is highly strained as indicated by the large calcu-lated Σ and θ values which are comparable to values calculatedfor similar ligand environments.27 A strong hydroxyl stretch at3137 cm−1 is observed in the IR, signifying that the hydroxylhas not been deprotonated. The presence of two triflate countera-nions and C–O bond lengths typical for an sp3 hybridizedcarbon32 indicate that the oxygens are not oxidized/deprotonatedand metal is iron(II). There are hydrogen bonding interactionsbetween the coordinated hydroxyl groups and the triflate coun-teranions, but no intermolecular interactions between cations areobserved (Fig. S24†).

Solution (magnetic) properties

Based on relatively low solid-state T1/2 values, it is unlikely thatSCO would be observed for the [FeL6-OH]2+ complex in solutionwithin the limited temperature range available for most solvents;nevertheless solvent environments are sufficiently distinct fromthe crystal lattice that there are several examples of solid-stateHS compounds which undergo spin equilibrium in solution.12,33

Even if [FeL6-OH]2+ were HS at all temperatures, establishingstability in polar solvents informs our efforts to corral labile ionsin solution.

Crystals of 1, 2, or 3 dissolved in CD3OD solution all showidentical spectra, indicating that the anion has little effect on thesolution characteristics of the cation at room temperature. The[FeL6-OH]2+ complex is HS, with paramagnetic shifting andbroadening of proton resonances (Fig. S25†, bottom). There arefour strong but broad resonances and several smaller peaks (−5to 80 ppm) due to a combination of rapid ligand twisting andlability. Resonances assigned to a small amount of ligand dis-sociation and/or hydrolysis are observed, but these do notincrease over time. In comparison to the tripodal complex, the1H NMR spectrum for [Fe(L2)2]

2+ contains 8 strong paramagne-tically broadened and shifted peaks (−2 to 190 ppm, Fig. S26,†bottom). There is no indication of ligand hydrolysis upondissolution in CD3OD. Based on the solid state coordinationenvironments for 1–4 as 5, several coordination modes are poss-ible for the tripodal complexes, including FeN6 and FeN4O2.Switching between these two modes could account for themuddled 1H NMR spectrum for [FeL6-OH]2+, and dissociation ofthe third arm could make it susceptible to hydrolysis.

To test for SCO and possible anion dependence in solution,the magnetic susceptibilities of 1, 2, 3, and 5 were studied inCD3OD from 193–296 K using Evans’ method (Fig. S33†). Allthree [FeL6-OH]2+ salts and 5 remain HS down to 193 K. Withcooling, the resonances for [FeL6-OH]2+ sharpen and becomemore prolific (24 total), indicative of a low-symmetry species insolution (Fig. S31†). For comparison, the (non-hydroxylated)compound [Fe(6-Me-py)3tren](PF6)2 is HS from room temp-erature to 185 K in acetone, due to the steric hindrance of the6-position modification.17 Complex 5 demonstrates Curie behav-iour upon cooling down to 193 K in solution. The number of

Fig. 5 Comparison of key intermolecular interactions in 1·RT (a) and1·LT (b). Dark green, blue, red, grey, yellow, and light green spheres rep-resent Fe, N, O, C, S and F respectively. Disordered components, nones-sential C atoms and non-hydroxyl H atoms have been removed forclarity. The second triflate in (b) has been removed for clarity sincestrong interactions between the anion and cation are not observed.

Fig. 6 Complex cation of 5, rendered with 40% thermal ellipsoids.Green, red, blue and grey ellipsoids represent Fe, O, N and C atoms,respectively. Non-hydroxyl H atoms have been omitted for clarity.

This journal is © The Royal Society of Chemistry 2012 Dalton Trans., 2012, 41, 12577–12585 | 12581

Dow

nloa

ded

by U

nive

rsity

of

Mas

sach

uset

ts -

Am

hers

t on

04 O

ctob

er 2

012

Publ

ishe

d on

07

Sept

embe

r 20

12 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C2D

T31

213A

View Online

peaks observed remains constant indicating that the complex issymmetric at room temperature (Fig. S32†).

The solution stabilities of 2 and 5 in CD3OD were monitoredby 1H NMR over three days (Fig. S25–S26†). In both cases,small changes occur in the NMR spectra over time, indicative ofa small amount of complex degradation/lability. No significantchanges in χMT values (3.5 and 3.2 cm3 K mol−1 for 2 and 5,respectively) are seen, indicating that any lability impartsminimal effects on the solution magnetic properties. Throughthese stability studies, we can attribute any changes seen in χMTat variable temperatures to changes in the spin state and not dueto complex degradation over the several hours required for thevariable temperature experiment.

As a measure of each complex’s stability in the presence of astrong field ligand, three equivalents of 2,2′-bipyridine (bpy)were added to solutions of 2 and 5, respectively. Within5 minutes, a decrease in χMT to 0.5 cm3 K mol−1 for 2 and0.3 cm3 K mol−1 for 5 occurs, and signals corresponding to freeligands increase in intensity (Fig. S27–S30†), indicating the for-mation of LS [Fe(bpy)3]X2 (X = Br− or OTf−). The lability ofthe HS species allows bpy to easily substitute for the ligands andform LS complexes.

Conclusions

We have prepared several salts of a new tripodal hexadentate[FeL6-OH]2+ complex and studied their temperature-dependentstructural and magnetic properties. We observe that covalentlinkage of iminopyridine-hydroxyl ligand arms is required topoise functional groups for anion binding: without it, thehydroxyl groups bind to the Fe(II) centre. Solid state SCO temp-eratures, completeness and cooperativity depend on severalcomplex and interrelated factors, some of which are extractedfrom structural comparisons. We observe that larger anionscapable of strong hydrogen bonding interactions appear to fitbetter with the hydroxyl-containing arms, and may also providemore flexibility for the complex to change shape upon HS → LSconversion, while weakly interacting anions stabilize the HSstate. Cation–cation and cation–solvent interactions impart some-what capricious effects on SCO, but with the exception of 1 theyappear to work against sharp transitions at higher temperatures.Subtle interactions are not easily controlled in the solid state;nevertheless the observation of anion dependence in SCO prop-erties makes this complex promising for continued efforts inanion-triggered spin-state switching schemes.

In methanol solution, salts of [FeL6-OH]2+ remain HS down to183 K, but higher levels of stability are achieved than tris(biden-tate) Fe(II) complexes with similar ligand fields.

If the SCO temperature can be increased, then anion depen-dence in solution spin-state switching may be observed for tripo-dal Fe(II) complexes. In that vein, the structural analysespresented here suggest that only two arms are needed for anionchelation. Previous work by Steed and co-workers has shownthat only two arms are necessary for strong anion binding inRu(II) organometallic complexes.34,35 Using the third arm forelectronic/steric tuning of the ligand field could drive SCO tohigher temperatures in the solid-state and in solution. Efforts toincorporate such temperature tunability in heteroarm versions of

[FeL6-OH]2+ are underway. This will allow us to address the sig-nificance of anion chelation and/or “push–pull” interactions onthe host–guest tuning of spin-state switching in solution.

Experimental section

The ligand L6-OH was prepared under anaerobic conditions andL2 was prepared under aerobic conditions; compound manipula-tions involving Fe(II) ions were performed inside a dinitrogen-filled glovebox (MBRAUN Labmaster 130). Pentane was dis-tilled over sodium metal and subjected to three freeze–pump–thaw cycles. Acetonitrile (MeCN) and diethyl ether (Et2O) weresparged with dinitrogen, passed over molecular sieves, and sub-jected to three freeze–pump–thaw cycles. All other reagentswere obtained from commercial sources and were used withoutfurther purification; 2-pyridinecarboxaldehyde-6-methanol21 and[Fe(MeCN)2(OTf)2]

22 were synthesized using previouslydescribed procedures.

Synthesis of tris{4-[(6-methanol)-2-pyridyl]-3-aza-3-butenyl}-amine (L6-OH)

A 25 mL flask was charged with one equivalent of tris(2-amino-ethyl)amine (tren, 0.248 g, 1.70 mmol), 3.1 equivalents of2-pyridinecarboxaldehyde-6-methanol (0.785 g, 5.25 mmol),8 mL methanol and 1 g of molecular sieves (3 Å). The mixturewas stirred until the alcohol precursor dissolved. The resultingbrown-yellow oil was triturated with 10 mL of Et2O. The oil wasmixed with NaSO4 in methanol to remove residual water. Thesolution was filtered, the solvent was removed from the filtrateand the product was dried under vacuum for 8 hours to afford0.629 g of yellow oil (80%). The product was used as isolatedwithout any additional purification. IR (ATR) νOH 3273 cm−1,νCH 2845 cm−1, νCvN 1647 cm−1. 1H NMR (400 MHz;CD3OD; Me4Si) δ 8.3 (3H, s), 7.8 (3H, t, J = 7.715, 7.696 Hz),7.72 (3 H, d, J = 7.628 Hz), 7.53 (3 H, d, J = 7.567 Hz), 4.68(6 H, s), 3.77 (6H, t, J = 6.157, 6.003 Hz), 2.98 (6H, t, J =6.207, 6.291 Hz) ppm. 13C NMR (CD3OD) δ 161.4 (s), 159.4(s), 151.4 (s), 135.9 (s), 120.1 (s), 118.2 (s), 62.5 (s), 57.6 (s),53.2 (s) ppm. Anal. MS (CH3OH): m/z 526.253 [L6-OH + Na]+,504.271 [L6-OH + H]+.

Synthesis of [FeL6-OH](CF3SO3)2·0.25 EtOH (1)

A solution of L6-OH (0.250 g, 0.49 mmol) in 6 mL of methanolwas added to a colourless solution of [Fe(MeCN)2(CF3SO3)2](0.255 g, 0.49 mmol) in 4 mL of methanol. The instantly-formed red solution was stirred for 30 minutes. The solvent wasremoved in vacuo and the crude product was triturated with20 mL of Et2O to produce a brick red free flowing powder. Thesolid was isolated by filtration and X-ray quality crystals wereobtained by Et2O diffusion into an ethanolic solution of the com-pound (0.38 g, 91% yield). IR (ATR) νOH 3434, 3238 cm−1.λmax (CH3OH)/nm 500 (1160 M−1 cm−1); (MeCN)/nm 490(1300 M−1 cm−1). 1H NMR (300 MHz; CD3OD; Me4Si) δ 66.1,59.6, 48.6, 33.9, 26.4, 18.5, −9.4 ppm. χMT (SQUID, 295 K) =3.63 cm3 K mol−1 (μeff = 5.39 μB). Anal. Calcd for

12582 | Dalton Trans., 2012, 41, 12577–12585 This journal is © The Royal Society of Chemistry 2012

Dow

nloa

ded

by U

nive

rsity

of

Mas

sach

uset

ts -

Am

hers

t on

04 O

ctob

er 2

012

Publ

ishe

d on

07

Sept

embe

r 20

12 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C2D

T31

213A

View Online

C29.5H34.5N7O9.25FeS2F6: C, 40.8; H, 4.0; N, 11.3. Found: C,40.7; H, 3.8; N, 11.3.

Synthesis of [FeL6-OH]Br2 (2)

A solution of L6-OH (0.189 g, 0.88 mmol) in 6 mL of methanolwas added to a stirring, light brown-coloured solution of FeBr2(0.447 g, 0.88 mmol) in 8 mL of methanol. The instantly formeddark red-coloured solution was stirred for 30 minutes. A dark redpowder was isolated by removing the solvent under vacuum; theproduct was triturated with 5 mL of Et2O. X-ray quality crystalsof 2 were obtained by slow diffusion of Et2O into a methanolicsolution of the compound (0.561 g, 95.2% yield). IR (ATR) νOH3357, 3232 cm−1. λmax (CH3OH)/nm 502 (1140 M−1 cm−1). 1HNMR (300 MHz; CD3OD; Me4Si) δ 64.8, 57.3, 48.3, 33.9, 28.2,18.0, 10.4, −7.5 ppm. χMT (SQUID, 295 K) = 3.45 cm3

K mol−1 (μeff = 5.25 μB). Anal. Calcd for C27H33N7O3FeBr2: C,45.1; H, 4.6; N, 13.6. Found: C, 45.0; H, 4.4; N, 13.5.

Isolation of [FeL6-OH]Br2·0.5 MeOH (2·solv)

X-ray quality crystals of the solvated compound were obtainedby slow diffusion of Et2O into a 50 : 50 methanol–ethanol solu-tion of 2. χMT (SQUID, 295 K) = 3.50 cm3 K mol−1 (μeff =5.32 μB). Anal. Calcd for C27.50H35N7O3.50FeBr2: C, 44.9; H,4.8; N, 13.3. Found: C, 44.6; H, 4.6; N, 13.3.

Synthesis of [FeL6-OH]I2 (3)

A solution of 1 (0.033 g, 0.38 mmol) was added to a solution ofn-Bu4NI (0.057 g, 0.15 mmol) in 2 mL of acetonitrile to formimmediately a red precipitate; the mixture was stirred for oneadditional hour. The light red solid was isolated by filtration andwashed with 10 mL of acetone and 5 mL of pentane. Theproduct was purified by recrystallization via diffusion of Et2Ointo a methanolic solution of the compound (0.024 g, 77%yield); X-ray quality crystals were selected from the product. IR(KBr) νOH 3339, 3202 cm−1. λmax (CH3OH)/nm 502 (1180 M−1

cm−1). 1H NMR (300 MHz; CD3OD; Me4Si) δ 65.6, 58.5, 48.8,34.0, 28.4, 17.7, −8.9 ppm. χMT (SQUID, 296 K) = 3.52 cm3 Kmol−1 (μeff = 5.31 μB). Anal. Calcd for C27H33N7O3FeI2: C,39.9; H, 4.1; N, 12.1. Found: C, 39.7; H, 3.9; N, 11.9.

Synthesis of [FeL6-OH](BPh4)2·2 MeCN (4)

Carried out in a manner similar to the preparation of 3, anionexchange of 2 (0.103 g, 0.144 mmol) into 6 mL of methanolwith NaBPh4 (0.205 g, 0.599 mmol) produced a light pink pre-cipitate. The mixture was stirred for 30 minutes, and then thesolid was isolated by filtration. The resulting solid was trituratedwith methanol (6 mL) and stirred for 16 hours. The solid wasisolated by filtration and washed with 6 mL of Et2O to produce abrick red, free flowing powder. X-ray quality crystals wereobtained by Et2O diffusion into a concentrated acetonitrile solu-tion (0.120 g, 66% yield); IR (ATR) νOH 3509 cm−1. λmax

(MeCN)/nm 489 (1340 M−1 cm−1). 1H NMR (300 MHz;CD3CN; Me4Si) δ 75.3, 56.9, 47.9, 40.7 32.3, 10.4, 8.9 ppm.χMT (SQUID, 296 K) = 3.62 cm3 K mol−1 (μeff = 5.38 μB).

Anal. Calcd for C7H7NO3FeB2: C, 74.1; H, 6.2, N, 9.9. Found:C, 74.1; H, 5.9; N, 9.5.

Synthesis of 4-[(6-methanol)-2-pyridyl]-3-aza-3-butenyl (L2)

In a similar synthetic procedure as used for L6-OH, 2-pyridine-carboxaldehyde-6-methanol (0.208 g, 1.518 mmol) and ethyl-amine (0.103 g, 2.277 mmol) were combined, resulting in a paleyellow oil (0.170 g, 68% yield). The product was used as iso-lated without any additional purification. IR (KBr) νOH3282 cm−1, νCH 2972, 2952 2869 cm−1, νCvN 1650 cm−1. 1HNMR (400 MHz; CDCl3) δ 8.39 (1H, s), 7.87 (1H, d, J =7.79 Hz), 7.73 (1 H, t, J = 7.71, 7.72 Hz), 7.28 (1 H, d, J =7.68 Hz), 4.80 (2H, s), 3.92 (1H, s), 3.70 (2H, q, J = 14.57,7.28, 7.30 Hz), 1.32 (3H, t, J = 7.28, 7.28 Hz) ppm. 13C NMR(CDCl3) δ 16.4 (s), 56.0 (s), 64.4 (s), 120.1 (s), 121.8 (s), 137.6(s), 153.8 (s), 159.4 (s), 161.5 (s). Anal Calcd forC9.075H12.15N2OCl0.15 (L2·0.075 CH2Cl2): C, 63.9; H, 7.2; N,16.4. Found: C, 63.8; H, 7.2; N, 16.5; the compound was dis-solved in CH2Cl2 for preparation of the elemental analysissample.

Synthesis of [Fe(L2)2](OTf)2 (5)

Similarly to 1, L2 (0.100 g, 0.609 mmol) in 4 mL of methanolwas added to a stirring solution of [Fe(MeCN)2(OTf)2] (0.132 g,0.305 mmol) in 4 mL of methanol. The immediately dark redcoloured solution was stirred for an additional 30 minutes. Thesolvent was removed in vacuo and the resulting red oil was tritu-rated in ether (15 mL) for 2 hours. The resulting red free-flowingpowder was isolated by vacuum filtration. X-ray quality crystalswere obtained by slow diffusion of ether into an ethanolic solu-tion of the compound (0.140 g, 55% yield). IR (KBr) νOH3137 cm−1. λmax (MeOH)/nm 496 (1570 M−1 cm−1). 1H NMR(400 MHz; CD3OD; Me4Si) δ 188.6, 162.7, 131.4, 128.9, 73.4,53.4, 8.33, −2.48 ppm. Anal Calcd for C20H24F6FeN4O8S2: C,35.2; H, 3.5; N, 8.2. Found: C, 35.2; H, 3.4; N, 8.2.

Crystallographic data collection and refinement

All single crystals were coated in paratone-N oil prior to removalfrom the glovebox. For structures obtained at 100–120 K, thecrystals were supported on Cryoloops and mounted under astream of cold dinitrogen. For data collection at higher temp-eratures, the crystal was encased in epoxy resin and mounted toa glass fibre. Data were collected using a Bruker Kappa Apex IICCD diffractometer with Mo Kα radiation and a graphite mono-chromator. Data were integrated and corrected for absorptioneffects with the Apex II software package.36 The SHELXTLsoftware package was used for solving the structures by directmethods and for subsequent refinements.37 Unless otherwisenoted, thermal parameters for all non-hydrogen atoms wererefined anisotropically. Hydrogen atoms were added at theideal positions and refined using a riding model where thethermal parameters were set at 1.2 times those of the attachedcarbon atom. Crystallographic information is collected inTable 2.

This journal is © The Royal Society of Chemistry 2012 Dalton Trans., 2012, 41, 12577–12585 | 12583

Dow

nloa

ded

by U

nive

rsity

of

Mas

sach

uset

ts -

Am

hers

t on

04 O

ctob

er 2

012

Publ

ishe

d on

07

Sept

embe

r 20

12 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C2D

T31

213A

View Online

In the structure of 1·LT, one triflate anion was disordered overtwo positions. These components were modelled with partialoccupancies and refined anisotropically. The structure for 1·RTwas found to contain several disordered components. Disorder ofthe triflate anion and O1 were modelled with partial occupanciesand refined anisotropically. The position of O3 was disorderedover three sites; the thermal parameters of each part were refinedisotropically. One bromide in the structure of 2 was disorderedover two positions; the disordered components were restrained tomaintain spherical atoms. Free refinement of disordered solventfor 2·solv resulted in 52% methanol, 45% ethanol, and 3% void;for simplicity, the occupation was constrained to 50% and atomswere refined isotropically due to the large amount of disorder.One hydroxyl group in the structure of 4 was disordered overtwo positions; thermal parameters for each part were refinedanisotropically. CCDC reference numbers 865553–865555,866021, 866022, 884193, 892044 and 892318 contain the sup-plementary crystallographic data for this paper.

Magnetic susceptibility measurements

All samples were prepared under a dinitrogen atmosphere. Crys-tals of compounds 1–3 and 5 were ground into fine powders,loaded into gelatin capsules, encased in six drops of Eicosane,and inserted into straws for analysis; crystals were not ground for4. Measurements were performed using a Quantum Designmodel MPMS-XL superconducting quantum interference device(SQUID) magnetometer in the temperature range of 5 to 300 Kunder a measuring field of 1000 Oe. The data were corrected bysubtracting the measured susceptibility of an empty sampleholder and six drops of Eicosane. Diamagnetic corrections werecalculated by using Pascal’s constants.38

Other physical measurements

Infrared spectra were measured with a Nicolet 380 FT-IR usingan ATR attachment or KBr plates as indicated. Visible absorptionspectra were obtained using an Agilent 8453 UV-visible spectro-meter under air-free conditions using a glass cuvette. 1H NMRspectra were recorded using Varian INOVA instruments operatingat 300 or 400 MHz. Solution magnetic susceptibility measure-ments were obtained by Evans’ method using TMS as the refer-ence at 300 MHz.39–41 Elemental analysis was performed byRobertson Microlit Laboratories in Ledgewood, NJ. High-resol-ution mass spectra were obtained on an Agilent Technologies6220 time-of-flight LC/MS.

Acknowledgements

We thank the National Science Foundation (CHE-1058889) andColorado State University for financial support.

Notes and references

1 P. Gütlich, Y. Garcia and H. A. Goodwin, Chem. Soc. Rev., 2000, 29,419–427.

2 P. Gütlich and A. Hauser, Coord. Chem. Rev., 1990, 97, 1–22.3 J. A. Real, A. B. Gaspar and M. C. Munoz, Dalton Trans., 2005, 2062–2079.

4 M. A. Halcrow, Polyhedron, 2007, 26, 3523–3576.5 P. Gütlich and H. A. Goodwin, Spin Crossover in Transition Metal Com-pounds I–III, Springer, Berlin, 2004.

6 A. Bousseksou, G. Molnar and G. Matouzenko, Eur. J. Inorg. Chem.,2004, 4353–4369.

7 G. J. Halder, C. J. Kepert, B. Moubaraki, K. S. Murray andJ. D. Cashion, Science, 2002, 298, 1762–1765.

8 S. M. Neville, G. J. Halder, K. W. Chapman, M. B. Duriska,B. Moubaraki, K. S. Murray and C. J. Kepert, J. Am. Chem. Soc., 2009,131, 12106–12108.

9 P. Gütlich, A. Hauser and H. Spiering, Angew. Chem., Int. Ed., 1994, 33,2024–2054.

Table 2 Crystallographic data

1·LT 1·RT 2 2·solv 3·LT 3·RT 4 5

Formula C29H33Fe C29H33Fe C27H33Fe C28.50H38Fe C27H33Fe C27H33Fe C79H79Fe C20H24FeN7O9S2F6 N7O9S2F6 N7O3Br2 N7O4Br2 N7O3I2 N7O3I2 N9O3B2 N4O8S2F6

Form wt g−1 mol−1 857.59 857.59 719.27 757.32 813.25 813.25 1197.88 682.40Colour Red Red Red Red Red Red Red RedHabit Block Block Plate Plate Block Block Block PlateT/K 120 296 120 120 120 296 120 100Crystal sys Mono Mono Mono Mono Ortho Ortho Triclinic MonoSpace grp P21/c P21/c P21/c P21/c Pbca Pbca P1̄ P21/cZ 4 4 4 4 8 8 2 4a/Å 10.2032(16) 10.3855(4) 10.2839(5) 10.0980(2) 11.0878(4) 11.3371(4) 13.8564(5) 13.3173(8)b/Å 19.557(3) 20.1767(8) 14.5320(7) 14.0077(3) 18.8979(6) 19.1168(6) 13.9397(5) 15.4235(10)c/Å 17.283(3) 17.4647(8) 19.9842(9) 22.0804(5) 28.2307(9) 28.6754(9) 17.9453(6) 14.8826(8)α/° 90 90 90 90 90 90 94.418(2) 90β/° 96.879(8) 93.046(1) 96.513(2) 94.7740(10) 90 90 103.112(2) 112.692(3)γ/° 90 90 90 90 90 90 100.035(2) 90V/Å3 3423.8(9) 3654.5(3) 2967.3(2) 3112.43(12) 5915.4(3) 6214.8(4) 3299.6(2) 2820.2(3)dcalc/ g cm−3 1.664 1.559 1.610 1.616 1.826 1.738 1.211 1.607Meas. ref 39 875 42 446 45 940 40 279 116 396 104 197 65 015 38 061Ind. ref 6490 6008 6790 9462 6271 7402 16 647 10 522Rint 0.0373 0.0333 0.0793 0.0270 0.0525 0.0445 0.0570 0.0387GOF 1.026 1.046 1.010 1.044 1.032 1.034 1.019 1.042R1/% 4.26 4.70 4.20 3.38 2.34 3.18 4.72 3.59(wR2)/%

a (10.79) (12.51) (9.93) (8.33) (4.99) (6.45) (11.24) (7.96)

a R1 = ∑||Fo| − |Fc||/∑|Fo|; wR2 = {∑[w(Fo2 − Fc

2)2]/∑[w(Fo2)2]}1/2.

12584 | Dalton Trans., 2012, 41, 12577–12585 This journal is © The Royal Society of Chemistry 2012

Dow

nloa

ded

by U

nive

rsity

of

Mas

sach

uset

ts -

Am

hers

t on

04 O

ctob

er 2

012

Publ

ishe

d on

07

Sept

embe

r 20

12 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C2D

T31

213A

View Online

10 G. Lemercier, N. Bréfuel, S. Shova, J. A. Wolny, F. Dahan, M. Verelst,H. Paulsen, A. X. Trautwein and J.-P. Tuchagues, Chem.–Eur. J., 2006,12, 7421–7432.

11 Z. Ni and M. P. Shores, J. Am. Chem. Soc., 2008, 131, 32–33.12 M. G. Burnett, V. McKee and S. M. Nelson, J. Chem. Soc., Dalton

Trans., 1981, 1492–1497.13 Z. Ni and M. P. Shores, Inorg. Chem., 2010, 49, 10727–10735.14 Z. Ni, A. M. McDaniel and M. P. Shores, Chem. Sci., 2010, 1, 615–621.15 Z. Ni, S. R. Fiedler and M. P. Shores, Dalton Trans., 2011, 40, 944–950.16 H. Taube, Chem. Rev., 1952, 50, 69–126.17 M. A. Hoselton, L. J. Wilson and R. S. Drago, J. Am. Chem. Soc., 1975,

97, 1722–1729.18 R. M. Kirchner, C. Mealli, M. Bailey, N. Howe, L. P. Torre, L. J. Wilson,

L. C. Andrews, N. J. Rose and E. C. Lingafelter, Coord. Chem. Rev.,1987, 77, 89–163.

19 H. Z. Lazar, T. Forestier, S. A. Barrett, C. A. Kilner, J. F. Letard andM. A. Halcrow, Dalton Trans., 2007, 4276–4285.

20 G. Brewer, R. J. Butcher, C. Viragh and G. White, Dalton Trans., 2007,4132–4142.

21 N. M. Shavaleev, F. Gumy, R. Scopelliti and J.-C. G. Bünzli, Inorg.Chem., 2009, 48, 5611–5613.

22 K. S. Hagen, Inorg. Chem., 2000, 39, 5867–5869.23 M. S. Haddad, W. D. Federer, M. W. Lynch and D. N. Hendrickson,

J. Am. Chem. Soc., 1980, 102, 1468–1470.24 E. W. Mueller, H. Spiering and P. Guetlich, Inorg. Chem., 1984, 23, 119–

120.25 M. G. B. Drew, C. J. Harding, V. McKee, G. G. Morgan and J. Nelson,

J. Chem. Soc., Chem. Commun., 1995, 1035–1038.

26 P. Guionneau, M. Marchivie, G. Bravic, J.-F. Letard and D. Chasseau,J. Mater. Chem., 2002, 12, 2546–2551.

27 M. A. Halcrow, Chem. Soc. Rev., 2011, 40, 4119–4142.28 M. Marchivie, P. Guionneau, J.-F. Létard and D. Chasseau, Acta Crystal-

logr., Sect. B: Struct. Sci., 2005, 61, 25–28.29 M. Seredyuk, A. B. Gaspar, J. Kusz, G. Bednarek and P. Gütlich, J. Appl.

Crystallogr., 2007, 40, 1135–1145.30 C. Brewer, G. Brewer, G. Patil, Y. Sun, C. Viragh and R. J. Butcher,

Inorg. Chim. Acta, 2005, 358, 3441–3448.31 A. Cabort, A. Michel, B. Therrien, H. Stoeckli-Evans, K. Bernauer,

G. Süss-Fink, A. F. Williams and G. Stupka, Inorg. Chim. Acta, 2003,350, 193–200.

32 P. Mueller, R. Herbst-Irmer, A. L. Spek, T. R. Schneider andM. R. Sawaya, Crystal Structure Refinement: A Crystallographer’s Guideto SHELXL, Oxford University Press, Oxford, 2006.

33 B. Weber and F. A. Walker, Inorg. Chem., 2007, 46, 6794–6803.34 S. J. Dickson, S. C. G. Biagini and J. W. Steed, Chem. Commun., 2007,

4955–4957.35 S. J. Dickson, M. J. Paterson, C. E. Willans, K. M. Anderson and

J. W. Steed, Chem.–Eur. J., 2008, 14, 7296–7305.36 APEX 2, Bruker Analytical X-Ray Systems, Inc., Madison, WI,

2008.37 G. M. Sheldrick, SHELXTL Version 6.14, Analytical X-Ray Systems,

Madison, WI., 1999.38 O. Kahn, Molecular Magnetism, VCH, New York, 1993.39 D. F. Evans, J. Chem. Soc., 1959, 2003–2005.40 D. H. Grant, J. Chem. Educ., 1995, 72, 39.41 D. Ostfeld and I. A. Cohen, J. Chem. Educ., 1972, 49, 829.

This journal is © The Royal Society of Chemistry 2012 Dalton Trans., 2012, 41, 12577–12585 | 12585

Dow

nloa

ded

by U

nive

rsity

of

Mas

sach

uset

ts -

Am

hers

t on

04 O

ctob

er 2

012

Publ

ishe

d on

07

Sept

embe

r 20

12 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C2D

T31

213A

View Online