DaltonTransactions

Dynamic Article Links

Cite this: DOI: 10.1039/c2dt12016j

www.rsc.org/dalton PAPER

A new class of thermo- and solvatochromic metal–organic frameworks basedon 4-(pyridin-4-yl)benzoic acid†

Gift Mehlana, Susan A. Bourne* and Gaëlle Ramon

Received 24th October 2011, Accepted 12th February 2012DOI: 10.1039/c2dt12016j

Using 4-(pyridin-4-yl)benzoic acid, 44pba (1) as a ligand, two new metal-coordination networks[Co4(44pba)8]n·[(DMF)3·(EtOH)0.25·(H2O)4]n (2) and [Ni4(44pba)8]n·[(DMF)3.5·(EtOH)·(H2O)1.5]n (3)were synthesized by solvothermal methods and structurally characterized. Compounds 2 and 3 areisostructural but differ in their solvent content. Each is a 2D-network which forms a spiral parallel to[001], giving rise to three distinct large channels, accounting for some 47% of the unit cell volume. Both2 and 3 display water-induced phase transformations with chromotropism, which has been confirmed byTGA and XRPD analysis. Solvatochromism in 2 is also evident with crystals exhibiting a range ofcolours depending on the solvent included. This phenomenon has been characterized using TGA, XRPDand UV-vis spectrophotometry.

Introduction

Dynamic materials, capable of responding to their environment,are currently receiving a great deal of attention. Such systemsrequire flexibility which may be achieved using weak inter-actions such as hydrogen bonding1,2 or through the use of suit-able metals and ligands in coordination compounds.3–5 Metal–organic frameworks (MOFs) are examples of the latter, and areof particular interest for their potential use as storage devices oras sensors.6–8

Solvatochromism is the ability of a chemical species tochange colour on exposure to or following inclusion of differentsolvents. It arises as a result of a change in the energy gapbetween ground and excited states, which in turn is caused byinteractions between chromophores and solvent molecules. Sol-vatochromism has been observed in organic conjugated poly-mers,9 which change conformation in certain solvents, insupramolecular systems as a result of changes in hydrogen-bonding interactions,10 and in transition metal complexes,usually as a result of changes in the coordination centre whichleads to changes in the visible d–d energy transitions.11–13 Thelatter is often accompanied by large changes in the structure ofthe complex.14,15 While most solvatochromic effects are reportedin solution and require large amounts of solvent, recent interesthas focussed on the use of MOFs as solvatochromic

systems.16–18 As MOFs frequently contain solvent-accessiblevoids or channels, they are ideally suited for use as solvatochro-mic sensors.

In this study 4-(pyridin-4-yl)benzoic acid (44pba) was used asa ditopic spacer ligand in coordination networks with first rowtransition metals. 44pba contains several interaction sites: boththe pyridyl nitrogen and the carboxylic acid can act as eithercoordination sites or hydrogen bonding centres and the aromaticrings allow intermolecular π–π stacking interactions. A search ofthe Cambridge Structural Database (CSD, Version 5.31, October2011 update)19 found 38 crystal structures containing 44pba.Most are discrete coordination complexes, but there are severalextended coordination polymers, linked by hydrogen-bondingbetween the carboxylic acid moieties.20–22 Previously reportedcomplexes with CoII have focussed on magnetic properties.23,24

Here we report the structures of 4-(pyridin-4-yl)benzoic acid,44pba (1) and its complexes [Co4(44pba)8]n·[(DMF)3·(EtOH)0.25·(H2O)4]n (2) and [Ni4(44pba)8]n·[(DMF)3.5·(EtOH)·(H2O)1.5]n(3). 2 and 3 exhibit significant solvatochromatism with reversiblesolvent-induced phase transitions in the solid state.

Experimental

Materials and preparative methods

All chemicals were obtained from commercial sources and wereused without further purification.

Analytical methods

C, H, N elemental analyses were performed on a LECOCHNS-932 analyzer. IR spectra were recorded as KBr pellets ona PerkinElmer spectrometer. Solid-state (diffuse reflectance)electronic spectra were measured on polycrystalline samples on

†Electronic supplementary information (ESI) available: Tables S1–S2:crystal information and hydrogen bonding; Fig. S1–S10: packingdiagram, thermal analysis, XRPD patterns, kinetic curves. CCDC850071–850073. For ESI and crystallographic data in CIF or other elec-tronic format see DOI: 10.1039/c2dt12016j

Centre for Supramolecular Chemistry Research, Department ofChemistry, University of Cape Town, Rondebosch, 7701, South Africa.E-mail: susan.bourne@uct.ac.za; Fax: +27 21 6505195; Tel: +27 216502563

This journal is © The Royal Society of Chemistry 2012 Dalton Trans.

Dow

nloa

ded

by U

NIV

ER

SIT

Y O

F C

APE

TO

WN

on

06 M

arch

201

2Pu

blis

hed

on 0

5 M

arch

201

2 on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/C

2DT

1201

6JView Online / Journal Homepage

an Analtikjena Specord 210-plus UV-Vis spectrophotometerover the range 300–800 nm, at a scanning rate of 5 nm s−1.

C12H9NO2 (1)

Single crystals of 4-(pyridin-4-yl)benzoic acid, 44pba (1) weregrown by slow evaporation from dimethylformamide (DMF)over two weeks. Elemental analysis found: C 72.26; H 4.68; N6.81%. Calcd for C12H9NO2: C 72.35; H 4.55; N 7.03%. IR(cm−1): 2410(br), 1705(s), 1603(s), 1542(m), 1516(m), 1403(s),1317(s), 1296(s), 1178(m), 1072(m), 1005(s), 828(s), 774(s),741(s), 658(m), 519(m).

Synthesis of [Co4(44pba)8]n·[(DMF)3·(EtOH)0.25·(H2O)4]n (2) and[Ni4(44pba)8]n·[(DMF)3.5·(EtOH)·(H2O)1.5]n (3)

80 mg of 44pba (0.4 mmol) was dissolved in 5 mL DMF withheating and stirring. 29 mg of Co(NO3)2·6H2O (0.1 mmol) wasdissolved in 2 mL ethanol with stirring. The two solutions weremixed and sealed in a 15 mL Telfon-lined autoclave and heatedat 105 °C for three days, followed by cooling to room tempera-ture. Purple prismatic crystals of (2) were obtained. Elementalanalysis found: C 59.40; H 4.06; N 6.82%. Calcd forCo4C105.5H94.5N11O23.25: C 59.15; H 4.45; N 7.19%. IR (cm−1):3248(br), 1658(m), 1610(s), 1557(m), 1518(m), 1376(vs), 1177(m), 1035(m), 1007(m), 828(m), 781(s), 751(m).

Green crystals of 3 were prepared by the same method as 2,using Ni(NO3)2·6H2O. As only a few crystals were isolated froma mixture of crystals and powdered material, full characterizationwas not possible.

Preparation of [Co4(44pba)8]n (2d) and[Co4(44pba)8]n·[(H2O)18]n (2d-w)

Drying crystals of 2 under vacuum for 10 h at 200 °C affordedpurple crystalline material 2d. TGA analysis of 2d shows nomass loss before decomposition above 400 °C. Elemental analy-sis is consistent with 2d being the desolvated framework of 2:Found: C 62.89; H 4.08; N 5.84%. Calcd for Co4C96H64N8O16:C 63.31; H 3.54; N 6.15%. IR (cm−1): 3245(br), 1610(s), 1553(m), 1521(m), 1377(vs), 1189(m), 1035(m), 1007(m), 828(m),780(s), 751(m), 706(m).

On exposure of 2d to water vapour at 25 °C for 24 h, the crys-talline material changes colour to yellow (2d-w). The TGA of2d-w has a mass loss of 14.96% over the temperature range50–105 °C, corresponding to 4.5 water molecules per cobalt ion,or to a proposed structure [Co4(44pba)8]n·[(H2O)18]n (calculatedmass loss 15.22%). Elemental analysis Found: C 54.12; H 4.54;N 5.46%. Calcd for Co4C96H100N8O34: C 53.74; H 4.70; N5.22%. IR (cm−1): 3180(br), 1610(s), 1557(m), 1518(m), 1428(m), 1376(vs), 1190(m), 1080(m), 1039(m), 1007(m), 826(m),780(s), 751(m), 693(m).

Hot-stage microscopy

The thermal behaviour of compounds 1–3 was observed using aNikon SMZ-10 stereomicroscope fitted with a LinkamTHMS600 hot stage and a Linkam TP92 temperature control

unit. Samples were placed under silicone oil on a cover slip andheated at 10 °C min−1. Thermal events were monitored with aSony Digital Hyper HAD colour video camera and capturedusing the Soft Imaging System program analysis.25

Thermogravimetric analysis (TGA)

Thermogravimetric analyses (TGA) were performed with a TAInstruments TA-Q500. The sample weight was approximately1–2 mg, and heating rates of 10 °C min−1 under a nitrogen gasflow of 50 mL min−1 were used.

Differential scanning calorimetry (DSC)

Differential scanning calorimetry (DSC) thermograms wererecorded using a TA Instruments DSC-Q200. Samples weighing1–2 mg were heated in closed aluminium cells with a pinhole inthe lid at a rate of 10 °C min−1 under a nitrogen gas flow of50 mL min−1.

X-ray powder diffraction (XRPD)

For powder X-ray diffraction (XRPD) traces, a HUBER ImagingPlate Guinier Camera 670 was used with Ni-filtered Cu-Kα-radiation (λ = 1.5406 Å) generated by a Philips X-ray generatorat 40 kV and 25 mA. The sample crystals were either applieddirectly onto Mylar® film or placed in a capillary and exposed tothe radiation for 40 min with 10 multiscans to collect the data. AHuber High-Temperature Controller HTC 9634 unit was usedwith a capillary rotation device in order to obtain powder X-raydiffraction patterns at elevated temperatures.

X-ray crystallography

Intensity data collection for compounds 1–3 was performed on aBruker KAPPA APEX II DUO diffractometer with graphite-monochromated Mo-Kα radiation (λ = 0.71073 Å). In each case,the crystal was maintained at 173(2) K using a Cryostreamcooler (Oxford Cryosystems UK). Unit cell refinement and datareduction were performed using the program SAINT.26 Datawere corrected for Lorentz-polarization effects and for absorption(program SADABS27). Structure solution was achieved by directmethods (program SHELXS28) and refined by full-matrix least-squares on F2 with anisotropic thermal parameters for all non-hydrogen atoms using SHELXL29 within the X-SEED30 inter-face. Hydrogen atoms bound to carbon and nitrogen atoms andall hydrogen atoms on ethanol guests were placed in calculatedpositions and refined isotropically using a riding model. Hydro-gen atoms could not be located for water molecules in 2 and 3.The details of crystal data, intermolecular hydrogen bonds andselected bond lengths and angles for compounds 1–3 are listedin Table 1 and in the ESI.†

Kinetics

Kinetic studies were carried out to analyse the uptake of watervapour by dried samples of 2, codename 2d. A balance, placedin a thermostatically controlled oven (precision ca. 0.5 °C), is

Dalton Trans. This journal is © The Royal Society of Chemistry 2012

Dow

nloa

ded

by U

NIV

ER

SIT

Y O

F C

APE

TO

WN

on

06 M

arch

201

2Pu

blis

hed

on 0

5 M

arch

201

2 on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/C

2DT

1201

6J

View Online

connected to a computer which tracks the experiment. A sampleholder is contained within a brass cylinder which has fivesolvent chambers evenly distributed around the central samplecontainer. The brass container fits over the sample on thebalance and can be sealed. The solvent was placed in the fivesolvent chambers, the container sealed and the vapour pressureallowed to equilibrate. The scale was zeroed, the seal betweensample and solvent vapour removed and the monitoring compu-ter program (PB303S31) started. The increase in mass with timewas recorded and then fitted to analyse the extent of reactionwith time. Standard kinetic models were fitted to the data todetermine the rate constant for water uptake at a giventemperature.

Results and discussion

Structural and thermal analysis

C12H9NO2 (1). Compound 1 can be crystallized from solventssuch as dimethylformamide, dimethylacetamide and dimethyl-sulfoxide. In all cases, the same form was obtained, crystallizingin the orthorhombic system, space group P212121. Details of thecrystal structure are given in Table 1. The two independent mol-ecules (labelled A and B) pack in anti-parallel layers along[010], with hydrogen-bonds between the carboxylic acid andpyridyl moieties of crystallographically related molecules (O⋯Ndistances are 2.619(3) and 2.611(3) Å for molecules labelled Aand B respectively). A weak hydrogen bond between C9A andO14B connects the layers. Fig. 1 shows the atomic labelling.The TGA of compound 1 was featureless until decomposition,indicating that no solvent was included. The DSC trace shows adouble endotherm with maxima at 342 and 371 °C, which corre-sponds to the melting and decomposition of 1.

[Co4(44pba)8]n·[(DMF)3.25·(EtOH)0.25·(H2O)4.25]n (2) and[Ni4(44pba)8]n·[(DMF)3.6·(EtOH)·(H2O)1.75]n (3). The frameworkin 2 and 3 is identical, although their solvent content differs

slightly. Crystals of 3 diffracted relatively poorly and the struc-ture could not be refined to an acceptable level of precision (seeTable 1). For this reason, only the structure of 2 is fullydescribed here. The crystal structure of compound 2 containsfour crystallographically independent Co(II) ions. Each Co2+ ionis 6-coordinate and bound to four 44pba ligands, via two Co–Nand four Co–O bonds, as shown in Fig. 2. The octahedron is dis-torted by the bidentate binding of the carboxylate moieties. Theenvironment around each metal ion is very similar with Co–Nbond lengths of between 1.969(12) and 2.118(10) Å and Co–Obond lengths between 2.002(10) and 2.352(10) Å. More signifi-cant differences arise in the relative arrangements of the two aro-matic rings of the 44pba ligands and the orientation of thecarboxylate groups; these differences account for the four crys-tallographically independent Co(44pba)2 units which make upthe framework. The relevant torsion angles are listed in Table 2.

The overall network structure is a 2D-network which forms aspiral parallel to [001], thus giving rise to the three-dimensionalchannels seen in Fig. 3. There are three distinct channels, whichdiffer in their solvent contents. The solvent-accessible volume inthe unit cell was estimated, using the SQUEEZE routine inPLATON,32 as 12 246 Å3, which is approximately 47% of the

Fig. 1 Atom labelling in molecule A of 1.

Table 1 Crystallographic data for compounds 1–3

1 2 3

Formula C12H9NO2 [Co4(C12H8NO2)8]n·[(C3H7NO)3.25·(C2H6O)0.25·(H2O)4.25]n

[Ni4(C12H8NO2)8]n·[(C3H7NO)3.6·(C2H6O)·(H2O)1.75]n

Mr 199.20 2146.92 2152.13T/K 173 173 173Crystal size/mm 0.13 × 0.08 × 0.03 0.26 × 0.26 × 0.25 0.30 × 0.26 × 0.25Crystal system Orthorhombic Tetragonal TetragonalSpace group P212121 I4 I4a/Å 10.7482 (12) 31.9819 (12) 31.8199 (19)b/Å 11.7997 (13) 31.9819 (12) 31.8199 (19)c/Å 14.7279 (15) 25.3323 (11) 25.1149 (15)V/Å3 1867.9 (3) 25 910.9 (18) 25 429 (3)Z 8 8 8ρ (calcd) (Mg m−3) 1.417 1.101 1.124μ(Mo-Kα)/mm−1 0.098 0.565 0.646Theta range for data collection (°) 2.21–27.17 1.64–27.34 1.64–28.29Reflections collected 6844 47 724 90 225No. unique data [R(int)] 4121 [0.0344] 21 374 [0.0455] 31 528 [0.0556]No. data with I > 2σ(I) 2564 11 366 17 448Final R (I > 2σ(I)) 0.0547 0.0764 0.1397Final wR2 (all data) 0.1177 0.2718 0.4313Max, min e density (e A−3) 0.197, −0.204 0.756, −1.013 2.067, −1.942

This journal is © The Royal Society of Chemistry 2012 Dalton Trans.

Dow

nloa

ded

by U

NIV

ER

SIT

Y O

F C

APE

TO

WN

on

06 M

arch

201

2Pu

blis

hed

on 0

5 M

arch

201

2 on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/C

2DT

1201

6J

View Online

unit cell volume (25 910.9 Å3). Channel A, centred on (0 12 z)

contains an ethanol molecule (with refined site occupancy of0.25) and several disordered water molecules. Channel B,centred at (14

14 z) contains DMF and water while channel C,

centred at (1212 z) contains only water. The total solvent content

was estimated using TGA and this was used to model the rel-evant site occupancy factors for solvent molecules. CHN analy-sis corresponds to the bulk material having composition[Co4(44pba)8]n·[(DMF)3·(EtOH)0.25·(H2O)4]n. One might expectthat solvent content would differ from one crystal to another in acompound of this type and, for the crystal selected, a total of3.25 DMF molecules are modelled in 2, disordered over fourpositions, one ethanol position is modelled with site occupancy0.25 while fifteen positions are modelled for a total of 4.25 watermolecules per asymmetric unit. There are no intermolecularinteractions between the solvent molecules and the[Co4(44pba)8]n framework, but extensive hydrogen bondingexists between the solvent molecules.

DSC analysis of compound 2 shows a complex endothermbetween 50 and 175 °C which corresponds to the multistep massloss seen in the TGA (Fig. 4). The observed mass loss of16.36% is a reasonable match to the solvent content modelled inthe crystal structure, see Table 3. Hot stage microscopy indicatesthat the purple crystals become opaque with the loss of the guestmolecules and decomposition occurs above 400 °C. Variable

temperature XRPD of compound 2 over the temperature range25–350 °C (Fig. 5) shows only minor changes in peak intensitiesuntil 200 °C and a loss of crystallinity by 300 °C. The XRPDpattern of the desolvated sample (above 120 °C) is in reasonableagreement with the simulated XRPD pattern of 2 (Fig. 5), indi-cating that the porous framework is retained in the absence ofguest solvent molecules. The slight shift observed in peak pos-itions for the as-made sample 2 at 25 °C compared to the simu-lated pattern (calculated from the crystal structure measured at−100 °C) is consistent with the shrinking of the unit cell at lowtemperature.

Thermal analysis of 3 is similar to that of 2, except for thesolvent content in the channels. In the structure of the crystalselected for analysis, 3.6 DMF molecules could be located, dis-ordered over three positions, four ethanol positions were mod-elled each with site occupancy 0.25 and seven positions weremodelled for a total of 1.75 water molecules per asymmetricunit. TGA analysis (Table 3) indicates that further solvent ispresent, but the diffraction data was of insufficient quality toallow us to model this.

Water-induced phase transformation with chromotropism

Single crystals of 2 exposed to air for several hours show an ir-reversible phase transformation to a powder crystalline phasewhich is accompanied by a colour change from purple to yellow-brown. Fig. 6 is a photograph of a single crystal of 2 in the

Fig. 2 CoII environment in 2, showing atom labels used. Furtherligands are labelled with suffices E–H.

Fig. 3 Crystal packing in 2, viewed down [001]. The framework andDMF and ethanol guests are shown in stick form while water moleculesare shown as red spheres.

Table 2 Torsion angles in 44pba in compounds 1–2

Atom labels A B C D E F G H

Compound 1

C(5)–C(4)–C(7)–C(8) (°) −140.8(3) −145.3(3)O(14)–C(13)–C(10)–C(11) (°) −4.0(5) −4.2(5)O(15)–C(13)–C(10)–C(9) (°) −4.8(4) −7.5(4)

Compound 2

C(5)–C(4)–C(7)–C(8) (°) 153(2) −149(2) 159(2) −177(2) 144(2) −137(2) 153(2) −158(2)O(14)–C(13)–C(10)–C(11) (°) 176(2) 179(1) 176(2) 177(2) 169(1) 176(2) 176(2) 168(2)O(15)–C(13)–C(10)–C(9) (°) 180(1) 176(2) 168(2) 179(2) −179(1) 170(2) 172(2) 176(1)

Dalton Trans. This journal is © The Royal Society of Chemistry 2012

Dow

nloa

ded

by U

NIV

ER

SIT

Y O

F C

APE

TO

WN

on

06 M

arch

201

2Pu

blis

hed

on 0

5 M

arch

201

2 on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/C

2DT

1201

6J

View Online

process of undergoing this phase transformation. The processyields a powder crystalline phase precluding single crystal analy-sis. However, we were able to replicate the powder diffractionpattern of the yellow material by first drying crystals of 2 undervacuum at 200 °C for 10 h (2d) and then exposing the driedmaterial to water vapour at 25 °C for 24 h (2d-w). The TGA of2d-w has a mass loss of 14.96% over the temperature range50–105 °C, corresponding to 4.5 water molecules per cobalt ion,or to a proposed structure [Co4(44pba)8]n·[(H2O)18]n (calculatedmass loss 15.22%). Intriguingly, when yellow 2d-w is heated,the water loss from the channels is accompanied by a colourchange to a deep purple, very similar to that of the dry com-pound, 2d (Fig. 7). This is accompanied by a complete loss ofcrystallinity by 150 °C.

Similarly, crystals of 3 change colour from green to white onexposure to moist air or to water vapour. When 3 is dried (3d)and then exposed to water for 24 h, the TGA shows a mass lossof 13.96% between 126 and 177 °C, corresponding to a pro-posed structure [Ni4(44pba)8]n·[(H2O)16]n (calculated mass loss13.67%).

Kinetics of water sorption

The general equation used to study the kinetics of a solid-stateprocess under isothermal conditions can be represented by eqn(A), or by the integral form eqn (B)

dα=dt ¼ kf ðαÞ ðAÞ

gðαÞ ¼ kt ðBÞ

where α is the fractional mass change [(mt − m0)/(m∞ − m0)] atreaction time t, k is the rate constant and f(α) and g(α) are func-tions that describe a given reaction mechanism.

The rate of uptake of water vapour by 2d was followed usingthe apparatus described in the experimental section. The sorptionexperiment was repeated on fresh samples of 2d at temperatures20, 25, 26, 27 and 30 °C. The sorption curves were converted toα-time curves and then fitted to standard kinetic models.33 Thebest fit was obtained for the 3-dimensional diffusion mechanism,D4 (eqn (C)), with rate constant (k) values ranging from 2.23 ×10−3 (at 20 °C) to 7.06 × 10−3 min−1 (at 30 °C). When fitted toan Arrhenius plot (Fig. 8) these yielded an activation energy of84.5 kJ mol−1.

kt ¼ ½1� 2α=3� � ð1� αÞ2=3 ðCÞThe D4 model applies to a diffusion reaction which takes

place at several nucleation sites forming phase interfaces

Fig. 4 TGA (green) and DSC (blue) of compound 2.

Table 3 Thermal analysis results for compounds 1–3

TGA mass lossobserved (%)

TGA mass losscalculated (%)

DSC peak 1(Ton) (°C)

DSC peak 2(Ton) (°C)

DSC peak 3(Ton) (°C)

1 — — Endo, 342 Endo, 3712:Co4(44pba)8]n·[(DMF)3·(EtOH)0.25·(H2O)4]n

16.36 DMF: 10.32 Endo, 50–175 Exo, 311 Endo, 419EtOH: 0.54H2O: 3.39Total: 14.25

3:[Ni4(44pba)8]n·[(DMF)3.5·(EtOH)·(H2O)1.5]n

18.55 DMF: 11.90 Endo, 80–195 Exo, 311 Endo, 419EtOH: 2.14H2O: 1.26Total: 15.30

Fig. 5 Variable temperature XRPD of compound 2.

Fig. 6 A single crystal of 2 (purple) after exposure to air. A surfacelayer of yellow polycrystalline material has formed.

This journal is © The Royal Society of Chemistry 2012 Dalton Trans.

Dow

nloa

ded

by U

NIV

ER

SIT

Y O

F C

APE

TO

WN

on

06 M

arch

201

2Pu

blis

hed

on 0

5 M

arch

201

2 on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/C

2DT

1201

6J

View Online

between the product (2d-w) and the reactant (2d). For the analo-gous nickel(II) system 3d forming 3d-w, rate constants rangedfrom 9.85 × 10−4 at 20 °C to 2.77 × 10−3 at 28 °C, and yieldedan activation energy of 87.1 kJ mol−1. During the process ofsorption, the water molecules diffuse into the crystal surface.The activation energies obtained for the cobalt (2) and nickel (3)compounds are almost identical and suggest that the kinetics ofthis process is similar for these related structures. Based on theknown crystal structures of the solvated forms (2 and 3), and thesimilarity of the XRPD pattern of 2d to 2, the D4 model is con-sistent with the crystal packing. The structures contain largechannels and there are no obvious directional constraints on thediffusion of guest water molecules. Comparison of the valuesobtained for the activation energies here to those of othersystems is not simple, as there are not many comparative studies,perhaps owing to the complexities in measuring kinetics ofenclathration processes. However, activation energies rangingfrom 20 to 140 kJ mol−1 have been reported for the uptake ofsmall solvent molecules by organic and coordination net-works.34,35 The kinetics of desolvation processes have beenmore thoroughly studied and recent reviews have reported valuesof activation energies of between 75 and 135 kJ mol−1.36,37 Thevalues obtained in this study fall within these ranges.

Solvatochromic effects in 2

The remarkable colour changes observed when 2 absorbs watervapour led us to study the behaviour of this compound onexposure to several organic solvents. 2 was dried under vacuumfor 10 h at 200 °C to give 2d. This material was then immersedin dry methanol, ethanol, DMF, DMSO or benzene to give 2d-

m, 2d-e, 2d-dmf, 2d-dmso and 2d-benzene respectively. Theresulting inclusion compounds have different colours dependingon the solvent included. In none of the cases studied here wasthere any evidence of compound 2d dissolving. 2d also respondsto solvent vapours but at a much slower rate. While 2d-dmf and2d-w appear to be stable at room temperature for extendedperiods, the alcohol inclusion compounds 2d-m and 2d-e beganto change colour within minutes of being removed from thesolvent. The change to a yellow material in each case suggeststhat these compounds were losing guest alcohol in favour ofwater vapour being absorbed from the atmosphere.

The same crystalline phase as in the as-made material 2 is pro-duced on exposure to DMF, DMSO or benzene (Fig. 9) butdifferent phases result from exposure to methanol, ethanol orwater. It is notable that the structure of 2 is regenerated onexposure to DMF, but not to the other solvents (ethanol, water)used in the synthesis of 2. This suggests that DMF may play arole in templating the formation of the framework in 2. Theother solvents do produce materials that vary in colour, structureand crystallinity. That formed by exposure to water shows a sig-nificant loss of crystallinity compared to either 2 or 2d. Asimilar effect was observed by Stylianou et al.38 for a 3D frame-work [In2(OH)2(TBAPy)] (guests) where TBAPy is 1,3,6,8-tetrakis(p-benzoic acid)pyrene. Those authors suggested that theabsorption of water led to significant structural disorder and dis-tortion in the network, causing a reduction in crystallinity due tothe accompanying strain.

Fig. 9 shows the changes in the XRPD patterns which can beclustered into two groups: those produced by loading of the non-polar or polar aprotic solvents (benzene, DMF, DMSO) andthose of the polar protic solvents (ethanol, methanol, water). Themost important changes are the decreased intensities of thepeaks at 7.5, 17 and 18° and the shift to lower θ of the peak at16° on exposure to polar protic solvents.

The structural changes indicated by the XRPD patterns inFig. 9 correlate with the colours observed and the UV-vis spectrashown in Fig. 10. The purple colour of the as-made material 2 isreplicated when 2d is loaded with DMF, DMSO or benzene(2d-dmf, 2d-dmso and 2d-benzene respectively), all either non-polar or polar aprotic solvents and none capable of acting as

Fig. 7 (a) single crystals of 2, as synthesized. (b) 2 after exposure towater vapour for 24 h (yellow crystals). On heating, water is releasedfrom the channels and the sample reverts to a purple colour.

Fig. 8 Arrhenius plot for sorption of water by 2d.

Fig. 9 XRPD of solvated forms of 2 after exposure of 2d to a range ofsolvents.

Dalton Trans. This journal is © The Royal Society of Chemistry 2012

Dow

nloa

ded

by U

NIV

ER

SIT

Y O

F C

APE

TO

WN

on

06 M

arch

201

2Pu

blis

hed

on 0

5 M

arch

201

2 on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/C

2DT

1201

6J

View Online

hydrogen-bond donor. This is seen as a broad absorption(450–750 nm) in the UV-vis spectrum. In contrast, polar proticsolvents which may act as hydrogen-bond donors and acceptors(ethanol, methanol, water) cause the loaded material to changecolour to pink (methanol), orange-brown (ethanol) or yellow(water). The UV-vis spectra of these samples show a narrowerabsorption band in the region 450–500 nm.

Infrared spectra of the solvated compounds are identical toone another, as well as to 2 and 2d, thus the metal coordinationgeometry is not affected by solvation. Elemental analysis of thewater-loaded sample (2d-w) is consistent with TGA analysis,indicating that there are 17.8 water molecules per [Co4(44pba)8]formula unit. This is rather lower than the maximum loadingvalue of 38 (calculated from the crystallographically derivedvoid volume32 and the liquid density of water), but it was alsoconsistently obtained on water sorption experiments from the

vapour (see preceding section on kinetics). Possibly theinclusion of a number of water molecules blocks access toregions of the pores. For the other solvents used, loading valueswere calculated from TGA analysis and compared to maximumtheoretical values (Table 4). All samples show a lower loadingvalue than is theoretically possible. Table 4 also lists somesolvent properties, including their Hansen solubility parameters(HSP) which have been suggested as useful measures of inter-molecular interactions with other solvents and surfaces.39 Thevalues indicate a correlation between δH, H-bonding capabilityand solvatochromic effects.

Conclusions

Using the versatile 44pba ligand (1), two isostructural metal–organic networks, 2 and 3, were synthesized. These comprise 2Dnetworks which form 3D channels by spiral assembly parallel to[001]. Three distinct channels are formed, which differ in theirguest solvent content.

Water sorption–desorption processes trigger a reversible chro-motropic transformation in both 2 and 3. The water sorptionprocess was followed kinetically to give activation energies of84.5 and 87.1 kJ mol−1 respectively.

Drying compound 2 yields a crystalline material with retentionof several of the largest diffraction peaks. Exposure of thismaterial to various solvents results in the formation of severalsolvated phases. These are similar to that of the as-made material2 when non-H-bond donor solvents are used. However theinclusion of H-bond donating solvents gives rise to new struc-tural phases whose colours are different to the former phases.These solvatochromic effects were confirmed by solid-state UV-vis diffuse reflectance spectra. These effects are fully reversible,a desirable property for a potential sensor material.

We have considered the nature of the colour changes. Themechanism most likely derives from molecular distortions in theframework induced by solvent hydrogen-bonding. Similareffects have been reported previously.16,17,40 Such frameworkdistortions may give rise to changes in the d–d transitions in thevisible region, although MLCT transitions and π–π* transitionsin the ligand may also play a role.

Fig. 10 (a) Photographs of crystalline powders of selected solvatedforms of 2. (b): UV-vis spectra of 2 (blue), dried form 2d (brown) andseveral solvated forms (2d-w: yellow, 2d-m: pink and 2d-dmf: purple).

Table 4 Solvatochromic parameters for compound 2

Solvent(code)

TGAmass loss(%)

Loading value(x in proposed formula):[Co4(44pba)8]n·[x(solvent)]n

Maximumloadingcapacity Colour

Solventproperties

Solventproperties:H-bondinga

HSPb

δdHSPδP

HSPδH

Water (2d-w) 14.96 17.8 38 Yellow Polar protic HBA-D 15.5 16.0 42.3methanol (2d-m) 13.80 9.0 10.3 Pink Polar protic HBA-D 14.7 12.3 22.3ethanol (2d-e) 13.27 6.1 16.0 Orange-brown Polar protic HBA-D 15.8 8.8 19.4DMF (2d-dmf) 21.95 7.0 11.9 Purple Polar

aproticHBA 17.4 13.7 11.3

DMSO (2d-dmso) 29.00 9.5 13.0 Purple Polaraprotic

HBA 18.4 16.4 10.2

benzene (2d-benzene) 15.02 4.1 10.4 Purple Non-polar NHB(HBA)

18.4 0 2.0

aHBA-D: Hydrogen bond acceptor and donor; HBA: H-bond acceptor; NHB(HBA): formally non H-bonding but have been reported to accept H-bonds. bHansen solubility parameters: δd (energy from dispersion bonds between molecules), δP (energy from intermolecular force betweenmolecules), δH (energy from hydrogen bonds between molecules).

This journal is © The Royal Society of Chemistry 2012 Dalton Trans.

Dow

nloa

ded

by U

NIV

ER

SIT

Y O

F C

APE

TO

WN

on

06 M

arch

201

2Pu

blis

hed

on 0

5 M

arch

201

2 on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/C

2DT

1201

6J

View Online

This work is valuable in expanding the field of coordinationcrystal engineering by providing insights into the effect ofweak interactions on the physical properties of coordinationnetworks.

Acknowledgements

Funding for this project was received from the South AfricanNational Research Foundation (CPR20100409000010269, grantnumber 73676). GM thanks the University of Cape Town Chem-istry Equity Development Programme for financial support. Wethank the University of Stellenbosch for the use of the UV-visspectrophotometer and the University of the Witwatersrand forperforming CHN analyses.

Notes and references

1 S.-R. Keum, S.-J. Roh, S.-E. Kim, S.-H. Lee, C.-H. Cho, K. Sung-Hoonand K.-N. Koh, Bull. Korean Chem. Soc., 2006, 27, 187.

2 Y. Sakaino, T. Takizawa, Y. Inouye and H. Kakisawa, J. Chem. Soc.,Perkin Trans. 2, 1986, 1623.

3 B. Moulton and M. J. Zaworotko, Chem. Rev., 2001, 101, 1629.4 M. J. Rosseinksy, Microporous Mesoporous Mater., 2004, 73, 15.5 C. L. Cahill, D. T. de Lill and M. Frisch, CrystEngComm, 2007, 9, 15.6 L. Ma, C. Abney and W. Lin, Chem. Soc. Rev., 2009, 38, 1248.7 M. Yoshizawa, J. K. Klosterman and M. Fujita, Angew. Chem., Int. Ed.,2009, 48, 3418.

8 K. Davies, S. A. Bourne, L. Öhrström and C. L. Oliver, Dalton Trans.,2010, 39, 2869.

9 J. Aimi, Y. Nagamine, A. Tsuda, A. Muranaka, M. Uchiyama andT. Aida, Angew. Chem., Int. Ed., 2008, 47, 5153.

10 J. L. Scott, A. Almesaker, Y. Sumi and K. Tanaka, Cryst. Growth Des.,2007, 7, 1049.

11 R. W. Soukup and K. Sone, Bull. Chem. Soc. Jpn., 1987, 60, 2286.12 W. Linert, R. F. Jameson and A. Taha, J. Chem. Soc., Dalton Trans.,

1993, 3181.13 O. Hallale, S. A. Bourne and K. R. Koch, New J. Chem., 2005, 29, 1416.14 M. Du, X.-G. Wang, Z.-H. Zhang, L.-F. Tang and X.-J. Zhao, CrystEng-

Comm, 2006, 8, 788.15 L. Dobrzanksa, G. O. Lloyd, C. Esterhuysen and L. J. Barbour, Angew.

Chem., 2006, 118, 5988.

16 Z.-Z. Lu, R. Zhang, Y.-Z. Li, Z.-J. Guo and H.-G. Zheng, J. Am. Chem.Soc., 2011, 133, 4172.

17 Y. Gong, Y. Zhou, J. Li, R. Cao, J. Qin and J. Li, Dalton Trans., 2010,39, 9923.

18 J. Boonmak, M. Nakano, N. Chaichit, C. Pakawatchai and S. Youngme,Dalton Trans., 2010, 39, 8161.

19 F. H. Allen, Acta Crystallogr., Sect. B: Struct. Sci., 2002, 58, 380.20 X.-L. Jia, J. Zhou, S.-T. Zheng and G.-Y. Yang, J. Cluster Sci., 2009, 20,

555.21 P. Teo, L. L. Koh and T. S. A. Hor, Inorg. Chem., 2008, 47, 6464.22 R. Sekiya, S. Nishikiori and K. Ogura, Inorg. Chem., 2006, 45,

9233.23 M.-H. Zeng, Y.-L. Zhou, M.-C. Wu, H.-L. Sun and M. Du, Inorg. Chem.,

2010, 49, 6436.24 X.-M. Zhang, Y.-Q. Wang, K. Wang, E.-Q. Gao and C.-M. Liu, Chem.

Commun., 2011, 47, 1815.25 Soft Imaging System GmbH: Digital solutions for Imaging and

Microscopy, Version 3.1 for Windows, © 1987–2000.26 SAINT, Version 7.60a, Bruker AXS Inc, Madison, WI, USA, 2006.27 G. M. Sheldrick, SADABS, Version 2.05, 2007.28 G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Crystallogr., 2008,

64, 112.29 G. M. Sheldrick, SHELXL-97, Program for Crystal Structure Solution,

1997.30 L. J. Barbour, J. Supramol. Chem., 2001, 1, 189.31 L. J. Barbour, private communication.32 A. L. Spek, Acta Crystallogr., Sect. D: Biol. Crystallogr., 2009, 65,

148.33 M. E. Brown, D. Dollimore and A. K. Galway, in Comprehensive Chemi-

cal Kinetics, Vol. 22, Reactions in the Solid State, ed. C. H. Bamford andC. F. H. Tipper, Elsevier, Amsterdam, 1980, p. 340.

34 A. Coetzee, L. R. Nassimbeni and K. Achleitner, Thermochim. Acta,1997, 298, 81.

35 A. J. Fletcher, E. J. Cussen, T. J. Prior, M. J. Rosseinsky, C. J. Kepert andK. M. Thomas, J. Am. Chem. Soc., 2001, 123, 10001.

36 A. Jacobs, N. L. Z. Masuku, L. R. Nassimbeni and J. H. Taljaard, Crys-tEngComm, 2008, 10, 322.

37 A. Jacobs, L. R. Nassimbeni, K. L. Nohako, H. Su and J. H. Taljaard,Cryst. Growth Des., 2008, 8, 1301.

38 K. C. Sylianou, R. Heck, S. Y. Chong, J. Bacsa, J. T. A. Jones, Y.Z. Khimyak, D. Bradshaw and M. J. Rosseinsky, J. Am. Chem. Soc.,2010, 132, 4119.

39 C. Hansen, Hansen Solubility Parameters: A User’s Handbook, CRCPress, Boca Raton, 2nd edn, 2007.

40 J. Zhang, Y.-S. Xue, J. Bai, M. Fang, Y.-Z. Li, H.-B. Du and X.-Z. You,CrystEngComm, 2011, 13, 6010.

Dalton Trans. This journal is © The Royal Society of Chemistry 2012

Dow

nloa

ded

by U

NIV

ER

SIT

Y O

F C

APE

TO

WN

on

06 M

arch

201

2Pu

blis

hed

on 0

5 M

arch

201

2 on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/C

2DT

1201

6J

View Online