Introduction

Polymorphism is the ability of a pure compound to adopt more than one packing arrangement in the solid state [1]. The occurrence of polymorphic modifications in molecular compounds is manifested not just as a consequence of minimum free energy of the crystalline phases but also by kinetics of crystal nucleation and growth [2, 3]. This phenomenon is of particular importance in the pharmaceutical industry since the activity of a drug might change abruptly from one polymorph to another [4, 5]. The preferential growth of a particular from can often be rationalized by consideration of conformational equilibria or changes in patterns of weak intermolecular interactions which are able to strongly influence the final packing structure [2, 6, 7]. The comparison between different polymorphs could provide an ideal approach for understanding the structure–property relationships in molecular solids because property differences between polymorphs result from different molecular packing and/or conformation and not from different chemical compositions [8, 9]. Therefore, controlling crystal formation of a specific polymorph has recently become one of the major activities and of high practical interest in the field of crystal engineering, particularly the characterization of polymorphs, the conversion of one into another, and the ways in which solid state properties of polymorphs vary with structure [10–12].

Unlike purely organic systems, it is rather unusual to observe polymorphism in inorganic compounds, in particular if the two or more polymorphs are stable at the same temperature. The difference in crystal packing may be accompanied by a change in colour. For example, the simple complex Pt(2,2′-bipyridine)Cl2 exists as two different crystal forms under ambient conditions depending on the solvent employed: a red form [13] that crystallises in space group Cmcm and a yellow form [14] that adopts space group Pbca.

Thiazolecarboxylic acids have attracted attention as potential ligands to first row transition metals. When deprotonated, thiazole-2-carboxylic acid and thiazole-4-carboxylic acid can form five-membered chelating rings through the nitrogen atom and the carboxylate. Thiazole-5-carboxylic acid cannot form [N,O] chelates because of the separation of the nitrogen and the carboxylic acids, but might form [S,O] chelates. The chemistry of these and similar ligands has been reviewed recently [15] and several complexes have been reported [16, 17].

Herein, we describe the syntheses of five molecular structures with chemical formula [M(4-tza)2(H2O)2] (where M is Co for complex I, Ni for complex II or Cu for complex III, respectively) and a complex of [Cu(4-tza)2]·4H2O (IV) with thiazole-4-carboxylic acid (4-tza) and the selective formation of polymorphs for [Co(4-tza)2(H2O)2] (I) and [Ni(4-tza)2(H2O)2] (II). Interestingly, the results reveal a rare example of an inorganic compound that exhibits two stable but distinct crystal structures under the same conditions without ready interconversion. The crystal packing is dictated by synthetic conditions including solvent compositions and temperature and the extremely stable nature of the two polymorphs does not allow interconversion between the two different crystal structures once formed within processing parameters investigated in this study.

Experimental

Materials and instruments

Chemical reagents were purchased commercially and were used as received without further purification; CoSO4·7H2O (Alfa Aesar, 98 %), Ni(SO4)·6H2O (BDH Chemicals, 97 %), Cu(NO3)2·2.5H2O (Alfa Aesar, 99 %), thiazole-4-carboxylic acid (C4H3NO2S, Apollo Scientific), l-tryptophan (C11H12N2O2, Merck, 99.0 %), ethanol (VWR Chemicals, 99.97 %).

Elemental analysis was performed on a Fisons—EA1108 CHNS/O Element Analyzer. IR spectra were collected in a range of 4000–600 cm−1, using a Thermo Scientific Nicolet iS5 FT-IR Spectrometer. Thermogravimetric-differential thermal analysis (TG/DTA) was conducted using a TGA/DSC 1 thermogravimetric analyzer instrument with a heating rate of 20 °C min−1 operating under air. The X-ray powder diffraction (PXRD) experiments were conducted using a PANalytical Empyrean X-ray diffractometer.

Preparation of [Co(4-tza)2(H2O)2] (I)

Co(SO4)·7H2O (0.0281 g, 0.100 mmol) was dissolved in 1.0 mL of distilled water and 4-Htza (0.0129 g, 0.1 mmol) was dissolved in 4.0 mL of distilled water (for I- α polymorph) or 4.0 mL of a mixed solvent of 40 % ethanol in distilled water (for I- β polymorph). The metal solution was slowly added to the ligand solution with constant stirring. The mixture was left at room temperature for several days from which orange crystals were formed. Anal. Calcd for [Co(4-tza)2(H2O)2] (I- α) (%): C, 27.36; H, 2.30; N, 7.98; S, 18.26. Found (%): C; 27.08; H, 2.30; N, 7.85; S, 18.00. Anal. Calcd for [Co(4-tza)2(H2O)2] (I- β) (%): C, 27.36; H, 2.30; N, 7.98; S, 18.26. Found (%): C; 27.55; H, 2.19; N, 7.81; S, 18.31. Yield based on Co: 90 % for I- α and 95 % for I- β. IR (cm−1) for I- α: 3301br, 3133m, 1649vs, 1522m, 1429m, 1323m, 1295s, 1185m, 965w, 891w, 814w, 792m, 625w. IR (cm−1) for I- β: 3132br, 3118m, 1600vs, 1517m, 1433m, 1340s, 1307s, 1197w, 964m, 855w, 779w, 622w.

Preparation of [Ni(4-tza)2(H2O)2] (II)

To 2.5 mL of an aqueous solution of 4-Htza (0.0129 g, 0.100 mmol) in a small glass bottle with a screw cap was added a solution of Ni(SO4)·6H2O (0.0262 g, 0.100 mmol) in 2.5 mL of distilled water. After the mixture was stirred for an hour, the bottle was left at room temperature. Light blue crystals of II- α polymorph crystallized after several days. II- β polymorph can be prepared in the same way as described for II- α but the bottle was sealed and was placed in an oven maintained at 80 °C for 48 h. The products were washed with water and isolated for X-ray diffraction experiments. Anal. Calcd for [Ni(4-tza)2(H2O)2] (II- α) (%): C, 27.38; H, 2.30; N, 7.98; S, 18.27. Found (%): C; 27.59; H, 2.24; N, 7.74; S, 17.99. Anal. Calcd for [Co(4-tza)2(H2O)2] (II- β) (%): C, 27.38; H, 2.30; N, 7.98; S, 18.27. Found (%): C; 27.64; H, 2.17; N, 7.82; S, 18.00. Yield based on Ni: 95 % for II- α and 97 % for II- β. IR (cm−1) for II- α: 3309br, 3131m, 1648vs, 1523m, 1431m, 1326m, 1296s, 1187m, 967w, 890m, 813m, 792m, 626w. IR (cm−1) for II- β: 3190br, 3136m, 3114m, 1599vs, 1518m, 1435m, 1343s, 1311s, 1195w, 969m, 855w, 778w, 623w.

Preparation of [Cu(4-tza)2(H2O)2] (III) and [Cu(4-tza)2]·4H2O (IV)

To obtain crystals of III, a solution of Cu(NO3)2·2.5H2O (0.0242 g, 0.100 mmol) in 8.0 mL of distilled water was slowly added to a solution of l-tryptophan (0.0204 g, 0.100 mmol) in 3:1 volumetric ratio of methanol and distilled water with constant stirring. To this mixture, a solution of 4-tza (0.0129 g, 0.100 mmol) in 3.0 mL of ethanol was gently added with stirring. The mixture was then transferred to a Teflon bomb which was then placed in a water bath of approximate 1 L capacity. The reaction was carried out in a domestic microwave oven (Prestige SM23, medium high, 626 W) for 30 min. Anal. Calcd for [Cu(4-tza)2(H2O)2] (III) (%): C, 27.00; H, 2.27; N, 7.87; S, 18.02. Found (%): C; 27.18; H, 1.94; N, 7.63; S, 17.89. Yield based on Cu: 94 %. IR (cm−1) for III: 3354br, 3140m, 3103m, 1610vs, 1524m, 1444s, 1336m, 1309s, 1195w, 972m, 850m, 779m, 619w.

Crystals of IV were obtained in similar way as described for III. However, Cu(NO3)2·2.5H2O (0.0242 g, 0.100 mmol) was added to a solution of l-tryptophan (0.0204 g, 0.100 mmol) in 9:3 volumetric ratio of methanol and distilled water with constant stirring. To this mixture, a solution of 4-tza (0.0129 g, 0.100 mmol) in 9.0 mL of ethanol was gently added with stirring. Anal. Calcd for [Cu(4-tza)2]·4H2O (IV) (%): C, 24.52; H, 3.09; N, 7.15; S, 16.37. Found (%): C; 25.19; H, 2.62; N, 7.15; S, 18.19. Yield based on Cu: 90 %. IR (cm−1) for IV: 3236br, 3114m, 3089m, 1600vs, 1533w, 1432s, 1347s, 1324w, 976w, 859w, 767w, 623w.

Preparation of dehydrated compound of III and rehydrated compound of III′

A crystalline sample of III was heated to 120 °C in a vial in an oven. Well-shaped crystals of a deep blue solid of III′ were obtained after 1 h of heating. For the redydration process, a crystalline sample of III′ was placed in a closed double-vial system with saturated water vapour for a week.

Temperature-controlled crystallization experiment for I and II

In each experiment, an aqueous solution (2.50 mL) of Co(SO4)·7H2O (0.0281 g, 0.100 mmol) (or Ni(SO4)·6H2O (0.0296 g, 0.100 mmol)) was slowly added to an aqueous solution (2.50 mL) of 4-Htza (0.0129 g, 0.100 mmol) with constant stirring. The mixture was then left in oven at a constant temperature from room temperature to 80 °C. The orange crystals obtained in each experiment were isolated from mother liquor for analysis by X-ray diffraction.

Solvent-controlled crystallization experiment for I

Water, methanol and ethanol were used to explore the crystallization behaviour of I. In a typical experiment, Co(SO4)·7H2O (0.0281 g, 0.100 mmol) was dissolved in 1.0 mL of distilled water and 4-Htza (0.0129 g, 0.100 mmol) was dissolved in 4.0 mL of water or a mixed solvent of 20, 40, 60 and 80 % of alcohol in water. The metal solution was slowly added to the ligand solution. The mixture was then left at room temperature. Orange crystals were obtained from the solution after several days through solvent evaporation. The crystals were isolated for analysis by X-ray diffraction.

Crystal structure determination

The sets of X-ray diffraction intensity data from I- β, II- α, III and IV were collected in series of ω-scans using a Stoe IPDS2 image plate diffractometer operating with Mo Kα radiation at 150(2) K while the data set for II- β was collected in ω-scans using a Bruker APEX-II CCD diffractometer operating with Mo Kα radiation at 273(2) K. The collected data were corrected for absorption using the Tompa method in the case of large well-formed crystals [18]. In other cases, multi-scan absorption corrections were applied [19]. The structures were solved using dual-space methods within SHELXT and full-matrix least squares refinement was carried out within SHELXL-2014 via the WinGX program interface [20, 21]. All non-hydrogen positions were located in the direct and difference Fourier maps and refined using anisotropic displacement parameters. Hydrogen atoms attached to the tza were refined using a riding model. Hydrogen atoms of water were refined using a rigid body model for simplicity (I- β and III) but this approach led to unreasonably close approach of the water to adjacent tza in two cases (II- α and IV) hence the hydrogen atoms were refined freely subject to sensible bond angle and distance restraints. Crystal structure data for the compounds reported here are summarized in Table 1. A preliminary check on the crystals of I- α revealed this to be isostructural with II- α. The sample of I- α produced aggregates of crystals and we examined many crystallites, but did not find a crystal suitable for full data collection. Therefore, the full data set of I- α was, therefore, not collected, and a discussion is based on the data of II- α.

Table 1 Crystallographic data of I, II, III and IV
Full size table

Results and discussion

The 4-tza ligand is found to act as a chelating ligand to ions of the first row of the transition series. However, the packing of the resulting coordination complexes leads to polymorphic forms in the solid state.

Description of crystal structures

[Co(4-tza)2(H2O)2] (I) and [Ni(4-tza)2(H2O)2] (II)

Two polymorphs of [Co(4-tza)2(H2O)2] (I), hereafter referred to as I- α and I- β have been identified. Crystallographic data reveal that these polymorphs are isostructural with those of [Ni(4-tza)2(H2O)2] (II) and, therefore, the description of the crystal structures will be based on those of II in the following section.

Molecular structure of II

The molecular structure of the discrete complex [Ni(4-tza)2(H2O)2] (II) found in each polymorph (α and β) is illustrated below in Fig. 1. The nickel(II) cation is hexacoordinated with two ligands of 4-tza occupying the equatorial plane displaying κ 2-N,O coordination mode producing chelating five membered rings. The two remaining axial coordination sites are occupied by two water ligands directly bonding to the metal centre which lies on the crystallographic inversion centre. The coordination geometry for II can be described as a roughly regular octahedral geometry in which the Ni–N and Ni–O distances are not significantly different from each other, ranging from 2.023(10) to 2.092(10) Å (for II- α form) and from 2.0322(13) to 2.0925(13) Å (for II- β form) and the axial O3–Ni1–O3i is almost perpendicular to the coordination square-plane of the nickel(II) ion found in both polymorphs. It is possible to examine the difference in orientation of the ligands by overlaying the molecules of II- α and II- β polymorphs. The differences are strikingly small; there appears to be a small change in the location and orientation of the water ligand (Fig. 2) but the hydrogen-bonding interactions formed by this water appear very similar. It is, therefore, the different packing of these units that leads to different polymorphs.

Fig. 1
figure 1

Molecular structure of [Ni(4-tza)2(H2O)2] in II- α and II- β polymorphs drawn at 50 % probability ellipsoids with non-hydrogen atom labeling scheme. Symmetry code: (i) −x + 1, −y + 1, −z + 1

Full size image
Fig. 2
figure 2

View of overlaying asymmetric units of II- α (blue) and II- β (red) polymorphs. (Color figure online)

Full size image

Molecular packing of II- α and II- β polymorphs

The differences in the molecular packings of II- α and II- β polymorphs are remarkable, as shown in Fig. 3. II- α form crystallizes in monoclinic cell with space group P2 1 /n. Two O–H···O hydrogen bonds between the coordinating water ligands and the carboxylate oxygen atoms and a C–H···O weak hydrogen bonding interaction between the thiazole ring and the adjacent carboxylate oxygen atom bring the mononuclear species into a layered structure, as shown in Fig. 4a, b and listed in Table 2. If the [Ni(4-tza)2(H2O)2] unit of II- α is taken as a node, the hydrogen bonding network can be simplified as a two-dimensional 6-connected uninodal hxl/Shubnikov (3,6) plane net with Schläfli symbol 36·46·53 (Fig. 4c) [22]. These layers are further connected by inter-plane thiazole-carbonyl interactions and packed along the c-axis in ABAB stacking fashion (Fig. 5). Considering the inter-plane S···O contact with the non-bonded S1···O2i atomic distance of 3.290(9) Å, the distance which is significantly less than the sum of the van der Waals radii of S and O atoms (3.35 Å) indicates a possible interaction of S1 and O2 atoms with an S1···O2=C bond angle of 130.8(8)° resulting in the formation of the three-dimensional supramolecular network [23, 24]. This non-bonded S···O interaction is only observed in II- α possibly causing a less close-packed structure and a longer c-axis compared to II- β (Table 1).

Fig. 3
figure 3

Diagram showing the molecular packing of polymorphs in II

Full size image
Fig. 4
figure 4

Views of hydrogen bonding interactions in (a) II- α and (d) II- β. Diagrams depicting hydrogen bonding frameworks in (b) II- α with 6-connected uninodal hxl/Shubnikov (3,6) plane net (c) and in (e) II- β with 10-connected uninodal bct net (f)

Full size image
Table 2 Lists of hydrogen bonding interactions in the complexes
Full size table
Fig. 5
figure 5

View of non-bonded S···O interactions (dash lines) between two adjacent planes in II- α

Full size image

The polymorph II- β crystallizes in a different unit cell from II- α; each is monoclinic but these are different setting of the same space group of P2 1 /c. It is important to note that although each polymorph crystallizes in space group 14, the relationship between them is not a trivial resetting of the axes. The hydrogen bonding interactions found in II- β form are significantly different from that of II- α form. The discrete [Ni(4-tza)2(H2O)2] units of II- β are held together through intramolecular hydrogen bonding interactions to form a three-dimensional network structure, as illustrated in Fig. 4d, e and listed in Table 2. Each coordinating water ligand acts as both hydrogen bonding donor and acceptor involved in three hydrogen bonding interactions; two strong O–H···O hydrogen bonds with the carboxylate oxygen atoms of adjacent mononuclear molecules and one weak C–H···O hydrogen bonding interaction involving the thiazole ring. A weak C–H···O hydrogen bonding interaction between the thiazole ring and uncoordinated carboxyl oxygen atom is also observed. If the [Ni(4-tza)2(H2O)2] is taken as a node, then the hydrogen bonding network in II- β polymorph can be simplified to a three-dimensional 10-connected uninodal bct net with Schläfli symbol 312·428·55 (Fig. 4f) [22].

[Cu(4-tza)2(H2O)2] (III)

Complex III crystallizes in monoclinic cell with space group P2 1 /c. The molecular structure of III is almost identical to that of I- β and II- β polymorphs where the copper(II) cation is bonded to two 4-tza ligands in trans positions with five-membered chelate rings in the equatorial plane and two aqua ligands in the axial positions, in which the crystallographic inversion centre is located at the copper centre. Compared with those two structures, the Cu–OH2 bond length of 2.3759(16) Å found in III is significantly longer that of Co–OH2 (2.1054(12) Å) and Ni–OH2 (2.0925(13) Å) found in I- β and II- β polymorphs, respectively. This can be explained by a Jahn–Teller effect of the copper(II) ion resulting in an elongated octahedral geometry for the copper centre. This difference in coordination geometry of the metal centre does not affect the hydrogen bonding interactions among these isostructures. The hydrogen bonds found in III are similar to that of I- β and II- β polymorphs (Table 2) giving rise to the same three-dimensional hydrogen bonding net topology of 10-connected uninodal bct net with Schläfli symbol 312·428·55 [22].

[Cu(4-tza)2]·4H2O (IV)

Complex IV crystallizes in the same crystal system and space group as III. However, the asymmetric unit consists of one copper(II) cation which is located on an inversion centre and is N,O-chelated by two 4-tza ligands in the equatorial plane and further coordinated by two bridging carboxylate O atoms from another two 4-tza ligands in the axial positions, with an elongated octahedral geometry with the Cu–O bond length of 2.6857(19) Å and two lattice water molecules (Fig. 6). The 4-tza ligand in IV is found to act as chelating and bridging to the copper(II) ions. Therefore, instead of forming a discrete molecule like III, a two-dimensional layer with 4-connected uninodal sql/Shubnikov tetragonal plane net (Schläfli symbol 44·62) is found [22], as depicted in Fig. 7a, b and each layer is held together by strong hydrogen bonding interactions between carboxyl O3 atoms and lattice water molecules and between lattice water molecules themselves (Table 2).

Fig. 6
figure 6

Molecular structure of IV drawn at 50 % probability ellipsoids with non-hydrogen atom labeling scheme. Symmetry code: (i) −x + 1, −y + 2, −z + 1; (ii) −x + 1, y + 1/2, −z + 1/2; (iii) x, −y + 3/2, z + 1/2

Full size image
Fig. 7
figure 7

a View of the two-dimensional framework in IV and b diagram depicting 4-connected uninodal sql/Shubnikov tetragonal plane net

Full size image

Thermal properties of the complexes of II

Thermogravimetric analysis (TGA) reveals that II- α and II- β polymorphs possess high thermal stability up to 180 and 215 °C, respectively (Fig. 8a). On heating from room temperature to 900 °C, two major weight losses were observed. The first weight loss of approximately 12.35 % (calc. 10.26 %) occurring in a temperature range of 180–280 °C is related to the loss of two coordinating water ligands. At approximately 300 °C, the second weight loss of 66.58 % began which corresponds to the decomposition of the organic ligands (calc. 73.58 %). At the end of analysis, 21.07 % of ash remained. There are notable differences in the TGA traces here; II- β retains water to a higher temperature (loss begins around 215 °C) but II- α appears to convert to a phase that is moderately stable between 250 and 300 °C. As expected the decomposition of the two phases yields the same final product. The DSC trace (Fig. 8b) is featureless around 50–60 °C where the II- β polymorph might be expected to form. This confirms that there is no interconversion from II- α to II- β upon thermal annealing.

Fig. 8
figure 8

a TGA and b DSC curves for II- α (solid line) and II- β (dash line) polymorphs

Full size image

Temperature-controlled crystallization and polymorphic formation of I and II

The crystal packing can be controlled by the temperature of the solution from which the crystals are grown, and the stable nature of the two polymorphs does not allow interconversion between the two different crystal structures once formed. We have been unable to interconvert the two forms by annealing with and without solvent present. However, it is possible to control the formation of the different polymorphs by controlling the reaction conditions.

The effect of crystallization temperature on crystal growth behaviour and polymorphic formation of I and II was investigated. The overlay of the experimental powder diffraction patterns with the calculated one (Fig. 9a, b) shows good correlation of the crystal structures with the bulk material, indicating the purity of the bulk phase in each case.

Fig. 9
figure 9

PXRD patterns showing the study of temperature-controlled crystallization of I (a) and II (b)

Full size image

It is obvious from the PXRD patterns that crystallization temperature does not have an effect on the formation of polymorphs in I. At any studied temperature conditions, only I- β polymorph was obtained (Fig. 9a). Interestingly, in the case of II, we observe that when the crystallization occurs at near room temperature, only II- α polymorph is obtained. The temperature between 50 and 60 °C is a transition state in which a mixture of II- α and II- β polymorphs can be observed. When the temperature is raised to and over 80 °C, only II- β polymorph is formed (Fig. 9b). It can be noted that the temperature plays a key role in determining type of the derived polymorphs in II. II- α and II- β polymorphs can be, therefore, synthesized in bulk quantities through a temperature-control strategy.

Solvent-controlled crystallization and polymorphic formation of I

Water, methanol and ethanol were used to explore the crystallization behaviour of I. The solvent affects the crystal growth and the formation of polymorphs in I as seen by Fig. 10a, b. I- β polymorph can be readily obtained in the bulk in a wide range of solvents, e.g. water and mixed solvents of 20, 40 and 60 % methanol in water (Fig. 10a) at room temperature. In addition, I- β polymorph can be formed quickly from an aqueous solution at 80 °C. In contrast, I- α polymorph can only be obtained from a mixed solvent of 40 % ethanol with water (Fig. 10b). Furthermore, when using 20 and 60 % of ethanol in water as solvents, a mixture of I- α and I- β polymorphs can be observed. By using 80 % of ethanol in water, a poorly crystalline unknown phase which is not corresponding to either I- α nor I- β phases is observed. These results indicate a solvent-selective crystallization process of I. It can be concluded that I- α and I- β polymorphs can be synthesized in bulk quantities through a solvent-control strategy. The fact that seemingly small changes in solvent composition lead to different polymorphs is unexpected. Presumably during crystallization, interactions between the solvent and metal complexes are crucial in determining the packing present.

Fig. 10
figure 10

X-ray powder diffraction patterns showing crystallization behaviour of I in a mixed solvent of methanol with water (a) and ethanol with water (b)

Full size image

Solvent-dependent crystallization of III and IV

The microwave-assisted solvothermal technique has proved to be an effective way to obtain crystals of suitable quality for single crystal X-ray diffraction experiments. The syntheses of III with and without l-tryptophan were investigated. Strangely, it was found that III can only be obtained in the presence of l-tryptophan although this was not present in the product. We do not understand the role of this additive in the synthesis.

An attempt to study the polymorphic properties of III has been made by varying the solvent compositions. However, it was found that III does not exhibit polymorphism in a similar way to I and II. Interestingly, we observed that the solvent compositions play an important role in the crystallization behaviour of copper(II) complex with 4-tza. III was obtained through a mixed methanol/ethanol/water (3:3:9 mL) solvent whereas IV crystallized in a mixed methanol/ethanol/water (9:9:3 mL) solvent. These results indicate the selective crystallization of III and IV. In copious water, this acts as a ligand, but in limited water, it is present unbound but forming hydrogen bonds.

Thermal-induced irreversible structure transformation of III

Although III is isostructural with I and II, only a single polymorph of III was observed. However, we observe an irreversible structural transformation upon removal of coordinating water ligands. When a crystalline sample of [Cu(4-tza)2(H2O)2] (III) is heated at 80 °C for an hour, all of the aqua ligands are released from the structure giving rise to a dehydration product (III′) which is the known complex [Cu(4-tza)2] [17], confirmed by the PXRD patterns (Fig. 11) and accompanied by a colour change from turquoise to deep blue. Removal of water happens with minimal structural change (Fig. 12). The crystal of [Cu(4-tza)2] bears a close relationship to that of III; these have the same crystal system and space group, monoclinic P2 1 /c, and very similar unit cell parameters (see Table 1). Upon standing for seven days in air that is saturated with water vapour, form III′ does not return back to the hydrated form III, but the crystallinity of the sample improves.

Fig. 11
figure 11

X-ray powder diffraction patterns showing structural transformation upon removal of coordinating water molecules

Full size image
Fig. 12
figure 12

View of molecular structures of III (a) and III′ (b) in the unit cells

Full size image

Compared with III, the formation of an additional Cu–O4-tza coordination bond with the bond length of 2.8863(4) Å takes place along with removal of the the Cu–OH2 coordination bond with the bond length of 2.3759(16) Å forming a one dimensional chain running along a-axis. Thus, the coordination number of the CuII centre does not change. However, this change in coordination bond affects the hydrogen bonding network in the dehydrated complex compared to the pristine complex. Taking [Cu(4-tza)2(H2O)2] in III as a node, the hydrogen bonding network can be simplified as a 10-connected uninodal bct net with Schläfli symbol 312·428·55 whereas that of the dehydrated form can be simplified to a 6-connected uninodal pcu alpha-Po primitive cubic net with Schläfli symbol 412·63 by considering [Cu(4-tza)2] as a node [22].

Conclusion

Thiazole-4-carboxylic acid provides a predictable chelating mode of coordination to first row transition metal ions. We have shown the existence of several complexes with the same bis-chelate and shown the existence of different crystal packings for apparently identical complexes. Thus far, the sulfur atom of the thiazole ring has been almost entirely innocent in the chemistry described. We are currently investigating whether the sulfur atom can coordinate to another metal ion in a similar manner to the coordination of thiophene. The presence of hard donor (carboxylate) and soft donor (sulfur) might allow us to build extended frameworks of hard and soft metal ions for a variety of applications.

Supplementary materials

CCDC 1486988-1486992 contains the supplementary crystallographic data for this paper. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.