Radiofluorination of a Pre-formed Gallium(III) Aza-macrocyclic Complex: Towards Next-Generation Positron Emission Tomography (PET) Imaging Agents

As part of a study to investigate the factors influencing the development of new, more effective metal-complex-based positron emission tomography (PET) imaging agents, the distorted octahedral complex, [GaCl(L)]⋅2 H2O has been prepared by reaction of 1-benzyl-1,4,7-triazacyclononane-4,7-dicarboxylic acid hydrochloride (H2L⋅HCl) with Ga(NO3)3⋅9 H2O, which is a convenient source of GaIII for reactions in water. Spectroscopic and crystallographic data for [GaCl(L)]⋅2 H2O are described, together with the crystal structure of [GaCl(L)]⋅MeCN. Fluorination of this complex by Cl−/F− exchange was achieved in high yield by treatment with KF in water at room temperature over 90 minutes, although the reaction was complete in approximately 30 minutes if heated to 80 °C, giving [GaF(L)]⋅2 H2O in good yield. The same complex was obtained by hydrothermal synthesis from GaF3⋅3 H2O and Li2L, and has been characterised by single-crystal X-ray analysis, IR, 1H and 19F{1H} NMR spectroscopy and ESI+ MS. Radiofluorination of the pre-formed [GaCl(L)]⋅2 H2O has been demonstrated on a 210 nanomolar scale in aqueous NaOAc at pH 4 by using carrier-free 18F−, leading to 60–70 % 18F-incorporation after heating to 80 °C for 30 minutes. The resulting radioproduct was purified easily by using a solid-phase extraction (SPE) cartridge, leading to 98–99 % radiochemical purity. The [Ga18F(L)] is stable for at least 90 minutes in 10 % EtOH/NaOAc solution at pH 6, but defluorinates over this time scale at pH of approximately 7.5 in phosphate buffered saline (PBS) or human serum albumin (HSA). The subtle role of the Group 13 metal ion and co-ligand donor set in influencing the pH dependence of this system is discussed in the context of developing potential new imaging agents for PET.


Introduction
Fluorine-18, a positron-emitting isotope with a half-life of 109.8 minutes, is readily produced using a cyclotron and has become the radioisotope of choice for many medical imaging applications. The use of metal chelate-based complexes as a route towards new types of radioimaging agents for positron emission tomography (PET) by using 18 F offers an alternative strategy towards new PET agents from the widely studied organo-fluorine-based agents. A consequence of the relatively short half-life, t1 = 2 , of 18 F is that for medical applications rapid, late-stage radiolabelling is particularly desirable, ideally this should be the final step of the synthesis. The ability to introduce the radiolabel in water is also attractive in simplifying the procedure.
In addition to very elegant recent work towards organo-fluorine-based agents, [1] there has been a surge of research activity targeted for developing new inorganic 18 F agents, including those centred upon BÀF, [2] SiÀF systems [3] and also metal coordination complexes, based on AlÀF and GaÀF species. [4][5][6] The strength of the fluorine-element bond being formed during the radiofluorination is one of several key parameters in determining the suitability of a particular agent. Within the metalchelate-based systems, formation of a strong MÀF bond on a labile metal ion can allow reactions to proceed quickly and under relatively mild conditions. The strength of the other metal-co-ligand interactions are also important depending upon the mechanism that prevails for the introduction of the F À , and also for the stability of the final metal-fluoride complex under physiological conditions. The trivalent Group 13 metal ions, Al, Ga, In, are redox inactive, have relatively low toxicities and have well-defined coordination numbers for particular ligand sets. [7] These are important considerations in sim-plifying the solution chemistry of potential imaging agents in vivo. Other important factors include the nature of the chelating ligand to provide stability and scope for functionalisation to allow conjugation to the relevant biomolecules. Macrocyclic ligands offer advantages, because they tend to form very robust complexes with metal ions, which are often resistant to demetallation.
The work of McBride and co-workers [4] has demonstrated that careful choice of ligand type and metal allows easy access to 18 F-containing compounds founded upon Al III -triaza-macrocyclic complexes (Scheme 1). Recently Wan and co-workers have reported that this "Al-18 F" system has been translated into the clinic, by the simple addition of carrier-free [ 18 F] fluoride to a "pre-formed kit" containing an RGD NOTA (RGD = arginine-glycine-aspartic acid; NOTA = 1,4,7-triazacyclononane-1,4,7-triacetic acid) conjugate and AlCl 3 ·6 H 2 O buffered to pH 4, and heating to 100 8C. [8] The success of McBride's approach has stimulated efforts to develop the Al chemistry further to create new generation agents by using a coordination chemistry approach. [5] We recently reported an alternative route towards stable "Ga-18 F" labelled compounds by using a pre-formed chloridecontaining Ga III precursor complex that undergoes rapid halide (Cl/F) exchange within the metal coordination sphere under mild conditions, that is, at room temperature in weakly acidic aqueous MeCN solution. [6] The resulting [Ga 18/19 F 3 (BzMe 2 -tacn)] (BzMe 2 tacn = 1-benzyl-4,7-dimethyl-1,4,7-triazacyclononane), containing the neutral tridentate tacn macrocycle, showed very good stability in phosphate buffered saline (PBS) at pH 7.5 for at least two hours. As well as demonstrating the viability of this approach by using pre-formed Group 13 metal complexes, this study also revealed some important differences in the chemistry and properties of the corresponding Al III , Ga III and In III systems.
To develop this approach towards new PET imaging agents, there is considerable scope for exploring the underlying chemistry and for optimising the metal-ligand system. An optimal target for clinical use is a pre-formed complex (agent), provided in kit form, that undergoes rapid 18 F incorporation under mild conditions and at nanomolar concentration. This should ideally be a single synthetic step in the clinic that requires no purification post labelling, or, if purification is necessary, in which a simple cartridge-based method can be applied.
Herein, we demonstrate the successful radiofluorine-labelling of a pre-formed gallium chloride chelate complex based on the anionic pendant arm azamacrocyclic ligand, 1-benzyl-1,4,7-triazacyclononane-4,7-dicarboxylate (L 2À , the dianion of H 2 L in Scheme 2) on a nanomolar scale in water by using carrier-free 18 F À , in which the resulting radiocomplex has been purified by using a solid-phase extraction (SPE) cartridge.

Results and Discussion
Preparative scale synthesis of [GaX(L)] (X = Cl or F) The success of the room temperature halide-exchange reaction of the pre-formed [MCl 3 (R 3 -tacn)] complexes in aqueous solution [6] confirms that the GaÀF bonds are sufficiently stable to be considered for imaging applications and prompted further work to determine whether the same general method could be employed with the dicarboxylate pendant-arm ligand, 1benzyl-1,4,7-triazacyclononane-4,7-dicarboxylate (tacn; L 2À ). At first sight, this ligand may offer some advantages over the neutral BzMe 2 -tacn; it is potentially pentadentate with an N 3 O 2 donor set, leaving only the one coordination site required for 18 F À incorporation, while also contributing to the stability of the metal complex through a combination of both the macrocyclic and chelate effects. This moiety may also allow 18 F to be incorporated without the need for any 19 F and provides a direct comparison with the work of Jeong and co-workers. [5a] The ligand has been synthesised as both the dilithium salt (Li 2 L) [9] and as the carboxylic acid (H 2 L·HCl). [5a] Preparation of the precursor complex [GaCl(L)] was initially undertaken through reaction of GaCl 3 in anhydrous MeCN with a solution of Li 2 L in dry MeOH, giving a yellow orange crude solid that was poorly soluble in common solvents (H 2 O, MeOH, MeCN). Mass spectrometry, IR and 1 H NMR spectroscopic data are consistent with the target complex, indicating pentadentate coordination of the macrocyclic ligand, with a single Cl À ligand completing the distorted octahedral coordination envi- ronment at Ga III . However, purification and separation of the inorganic by-product (LiCl) proved to be challenging. Furthermore, the use of GaCl 3 , which is both very readily hydrolysed and reactive, as the source of the GaÀCl unit in the product was undesirable considering the goal to be able to create and radiolabel sub one milligram quantities of the pre-formed complex for imaging. Therefore, we sought an alternative source of Ga III , which would be better suited for use in aqueous solution. The compound Ga(NO 3 ) 3 ·x H 2 O is commercially available (Aldrich), and has been identified as the nona-hydrate in the solid state. [10] Reaction of Ga(NO 3 ) 3 ·9 H 2 O with one molar equivalent of H 2 L·HCl on a preparative scale in aqueous solution gave a yellow solid after work-up. Spectroscopic data are consistent with the formulation [GaCl(L)]. The solid was recrystallised from MeCN/Et 2 O to remove impurities. The chelation reaction is slow at room temperature in concentrated solution (ca. 12 h), but proceeds significantly faster at 85 8C. The reaction may also be performed upon heating at reflux in MeOH from which the desired product precipitates as a pale yellow solid. It is stable for many months in the solid state, and the 1 H NMR spectrum (D 2 O) is unchanged after four weeks in solution. The [GaCl(L)] is soluble in H 2 O and MeCN, and is poorly soluble in chlorocarbons.
The 1 H NMR spectrum (in D 2 O) of the [GaCl(L)] is significantly shifted and more complex than that of the H 2 L·HCl. The CH 2 protons of the benzyl group are split into an AB quartet ( 2 J HH 13.7 Hz) indicating that the tacn-dicarboxylate ligand is locked by N 3 O 2 coordination, leading to diastereotopic inequivalence of the CH 2 protons in the carboxylate groups. The tacn protons also appear as second-order multiplets. ESI + mass spectrometry gave m/z 402.1 (100 %), with an associated 69/71 Ga isotope pattern, corresponding to the monocation, [Ga(L)] + . The IR spectrum (Nujol mull) showed the expected carboxylate CO bands of the ligand were shifted to low frequency by approximately 50 cm À1 compared with H 2 L·HCl itself. The IR spectrum also showed the expected GaÀCl stretching vibration (375 cm À1 ) and evidence for H-bonded water (n(OH) = 3750, d(HOH) = 1648 cm À1 ). On the basis of these data, we concluded that the NO 3 À groups from the Ga(NO 3 ) 3 ·9 H 2 O precursor are not retained in the product, being replaced by the two carboxylate pendant arms of the macrocycle and one chloride anion, the latter derived from the H 2 L·HCl, in the distorted octahedral complex. Confirmation of this formulation came from a structure determination on a very weakly diffracting crystal obtained by Before the Cl/F halide exchange reaction was attempted by using the [GaCl(L)], we sought a method to obtain the corresponding GaÀF complex as a model compound to provide a well-defined spectroscopic fingerprint in preparation for the exchange studies. Given the success of hydrothermal synthesis in our previous work, [6] a similar method was employed. Reaction of GaF 3 ·3 H 2 O with Li 2 L (to ensure exclusion of Cl À ) under hydrothermal conditions resulted in the formation of the complex [GaF(L)]·2 H 2 O as a yellow solid after work-up. The precipitation of LiF provides a driving force towards the complexation. The 1 H NMR spectrum of the product showed a complex pattern consistent with ligand coordination. The pattern showed small chemical-shift differences from the chloride analogue, [GaCl(L)], and significant differences from the spectrum of the Li 2 L. The 19 F{ 1 H} NMR spectrum revealed a single resonance at d = À184.2 ppm, consistent with a GaÀF containing compound. [6] The ESI + mass spectrum of the compounds showed m/z 402.1 (100 %), with the expected isotope pattern for [Ga(L)] + . The IR spectrum showed a single, broad GaÀF stretching band (ñ = 568 cm À1 ). Evidence for H-bonded water was also evident from the IR spectrum. Slow evaporation of water from the hydrothermal reaction solution gave small crystals suitable for single crystal X-ray diffraction.
X-Ray structural characterisation confirmed the expected distorted octahedral coordination at gallium, through a pentadentate L 2À ligand and one terminal F À ligand (Figure 2 a). The GaÀ F bond length was found to be 1.821(2) , one of the shortest GaÀF bonds observed crystallographically, and even slightly shorter than d(GaÀF) in [GaF 3 (3) ), [6] although the latter exhibited extensive F···HÀOH hydrogen bonding to lattice water, and these may be responsible for longer GaÀF bond lengths in this complex. [11,12] Comparison with the analogous AlÀF complex, [AlF(L)]·2 H 2 O [5a] showed d(AlÀF) = 1.7090 (14) , leading to a difference in d(MÀF) of 0.11 . This is larger than would be expected based on the difference in covalent radii from Al 3 + to Ga 3 + . [13] Similarly, the difference between d(AlÀO) and d(GaÀO) is approximately 0.10 . Comparison of the MÀN bond strengths showed that d(GaÀN) lies in the range from 2.082(4) to 2.146(3) , whereas d(AlÀN) lies between 2.0497(18) and 2.1125(18) , that is, there is a smaller effect on the MÀN distances between the aluminium and gallium complexes. The presence of the macrocyclic ring may also play a role here.
Similar to the chloride analogue (ESI), the [GaF(L)] complex crystallises as a dihydrate; although they are isostructural, they are not isomorphous due to differences in the H-bonding arrangement. In the fluoride complex, both lattice waters are involved in H bonding (Figure 2 b).
The fluoride complex is stable in the solid state, and for several weeks in aqueous solution (in which the pH was measured to be ca. 4) and in other acidic media. Addition of aqueous KOH to an aqueous solution of the [GaF(L)] complex to bring the solution to pH 7 led to release of F À within 5-10 minutes. This is supported by the loss of the 19 F{ 1 H} NMR resonance ob-served for the complex and the growth of a resonance due to F À (d = À123.7 ppm).
Reaction of [GaCl(L)]·2 H 2 O in aqueous MeCN (unbuffered) with one molar equivalent of aqueous KF leads to complete conversion to the corresponding fluoride complex. The exchange proceeds to completeness at a moderate rate (ca. 3 h) at room temperature, but is significantly accelerated (requiring ca. 45 min) if heated to 80 8C. Spectroscopic analysis of the resulting product matched to that observed for [GaF(L)] synthesised hydrothermally. The Cl/F halide exchange may also be performed in buffered NaOAc solution (pH 4).

F Radiolabelling
Having demonstrated the ability to successfully perform halide exchange on the pre-formed [GaCl(L)] complex on a preparative scale by using 19 F À , radiofluorination was attempted ( Figure 3).
Carrier-free 18 F/ 18 OH 2 (500-1000 MBq) was added to 0.1 mg of the chloride precursor (210 nmol) dissolved in NaOAc buffer (pH 4) and left to react at room temperature for 30 minutes. HPLC analysis of the crude reaction mixture showed a radiopeak at R t 6-6.2 minutes corresponding to [Ga 18 F(L)] integration of the radiopeak indicated approximately 30 % incorporation of 18 F into the gallium complex at room temperature after 30 minutes ( Figure S2 in the Supporting Information). Heating the reaction solution to 80 8C for 30 minutes led to a significant increase in the incorporation to 65-70 % ( Figure S3 in the Supporting Information).
A small sample of the crude product (R t = 6 min) was removed and injected directly onto an ESI + mass spectrometer. Similar to the model compound, the parent ion was not observed; however, peaks attributed to the monocation, The crude 18 F-labelled compound was then purified by trapping it on an hydrophilic lipophilic branched (HLB) cartridge and eluted with EtOH/H 2 O. This purification process is very efficient, giving radioactive concentrations (RACs) of up to 100 MBq mL À1 . The purified radiochemical product is stable for at least 180 minutes when formulated in 10 % EtOH/NaOAc at   Figure S8 a-c in the Supporting Information). However, when formulated in 10 % EtOH/PBS (pH 7.5), the product is unstable, with the initial radiochemical purity (RCP) of 95-98 %, decreasing to 40 % after 20 minutes and 2 % after 90 minutes ( Figure S9 a-c in the Supporting Information). This is consistent with our observation that the non-radioactive [GaF(L)] complex is unstable in aqueous KOH at pH 7. Compound [Ga 18 F(L)] also liberates 18 F À when formulated into human serum albumin (HSA; pH 7.4; Figure S10 a-c in the Supporting Information). These data suggest that the stability is strongly pH dependant (Table S2 in the Supporting Information), although the presence of competing ions in the PBS and HSA formulations may also play a role.
The trend in stability with pH can be replicated on a preparative scale when K 19 F was used as the fluorinating agent. The fluoride liberated at pH 7.5 (in PBS) was readily observed by 19 F{ 1 H} NMR spectroscopy. It is notable that the [GaF 3 (BzMe 2tacn)] is stable for several hours (with no evidence of defluorination) in PBS, [6] both on a tracer scale concentration (micromolar) and on a preparative scale (mmol), and hence, it is clear that neither the GaÀF bonds nor the GaÀN(tacn) bonds are inherently unstable at pH 7.5. Therefore, it is likely that the pHdependent instability of [GaF(L)] is associated primarily with the GaÀO(carboxylate) bonds, with subsequent loss of F À . It is known from previous work on F À /H 2 O exchange on Al III complexes [14] that small changes in the steric environment can change the substitution mechanism (e.g., from D to I), hence, the steric and/or electronic changes at Ga III caused by the different donor sets in [GaF(L)] compared with [GaF 3 (R 3 -tacn)] may be the basis for the observed instability of the former at pH 7.5. Because Ga III is less Lewis acidic than Al III , [12,15] the introduction of the GaÀO (carboxylate) bonds in [GaF(L)] may lead to initial cleavage of the GaÀO bond at pH 7.5 (e.g., by H 2 O or an anion present in the formulation), destabilising the Ga III coordination sphere to cause the observed decomposition. Hence, the experimental observations from this study suggest that modifications to the macrocyclic pendant groups either, for example, by increasing the steric bulk at the carboxylate functions, or by changing the carboxylate functions to other anionic donor groups could be important, leading to improved stability of MÀ 18 F complexes.
The [Ga 18 F(L)] may also be synthesised by a one-pot method. Reaction of Ga(NO 3 ) 3 ·9 H 2 O and Li 2 L (1:1) in NaOAc (pH 4) with 18 F/ 18 OH 2 resulted in up to 80 % incorporation of 18 F into the gallium complex after heating at 80 8C for 30 minutes. This product may also be purified by SPE, as was described above. The purified product showed similar defluorination in 10 % EtOH/PBS over 90 minutes at pH 7.4.

Conclusion
This work describes a method for the preparation of [GaCl(L)]·2 H 2 O, containing Ga III in a distorted octahedral environment provided by the pentadentate L 2À ligand and one Cl À (derived from H 2 L·HCl), using the commercially available and easy-tohandle Ga(NO 3 ) 3 ·9 H 2 O as a convenient source of Ga III in water. The corresponding fluoride complex, [GaF(L)]·2 H 2 O, has been synthesised by a hydrothermal route to provide crystallographic and spectroscopic data, and also on a bulk scale by Cl/F exchange from [GaCl(L)]·2 H 2 O with K 19 F in (unbuffered) water, as a model for the radiofluorination reaction.
Direct radiofluorination at nanomolar concentration by treatment of this pre-formed [GaCl(L)]·2 H 2 O with carrier-free 18 F À in water at pH 4 (NaOAc) at room temperature has also been demonstrated, while heating to 80 8C for 30 minutes increases the yield of [Ga 18 F(L)] giving 65-70 % 18 F incorporation. The crude product was readily purified using an SPE cartridge, and showed excellent radiochemical stability at pH 6 (10 % EtOH/ NaOAc). Defluorination was observed when [Ga 18 F(L)] was formulated in PBS and HSA (pH 7.5 and 7.4, respectively). The instability of [GaF(L)] at high pH is attributed primarily to the presence of the coordinated carboxylate groups in L 2À . This also demonstrates subtle, but important differences in the behaviour of Ga III versus Al III with this dicarboxylate co-ligand; the higher Lewis acidity of Al III is more manifested towards the anionic ligand groups in [AlF(L)]. This is in accord with structural data, which showed that the MÀO and MÀF bond lengths are longer (by ca. 0.1 ) for Ga over Al, whereas the MÀN bond lengths are more similar. The significance of the precise metalco-ligand coordination on the stability of the radiofluoride Ga III complex with pH demonstrated in this work suggests that careful tuning of the steric and electronic properties of the anionic pendant groups on the tacn should allow optimisation of the design of new improved metal-complex-based imaging agents.

Experimental Section
Reactions were performed in standard lab glassware when appropriate. Water was freshly distilled before use. All other solvents used were of HPLC grade quality. ESI mass spectrometry was performed by using a Waters (Manchester, UK) ZMD mass spectrometer equipped with a single quadrupole analyser. Samples were introduced to the mass spectrometer by flow injection using a Waters 600 pump (flow rate 0.1 mL min À1 MeCN) and Waters 2700 autosampler. 1 H and 19 F{ 1 H} NMR spectra were recorded in solution in deuterated H 2 O or methanol on a Bruker DPX-400 or AV-400 spectrometers and are referenced to the residual solvent protons ( 1 H) and CCl 3 F ( 19 F) at 298 K. IR spectra were recorded neat (oils) or as Nujol mulls (solids) between CsI plates by using a Perkin-Elmer Spectrum 100 spectrometer over the range 4000-200 cm À1 . Microanalyses were undertaken by Stephen Boyer at London Metropolitan University. Compounds Bz(CH 2 CO 2 H) 2 -tacn·HCl (H 2 L·HCl) [5a] and Li 2 [Bz(CH 2 CO 2 ) 2 -tacn] (Li 2 L) [9] were prepared by using the literature methods; Ga(NO 3 ) 3 ·9 H 2 O and GaF 3 ·3 H 2 O were obtained from Aldrich and used as received.

Synthesis of [GaCl(L)]·2 H 2 O
A solution of H 2 L·HCl (0.111 g, 0.332 mmol) in freshly distilled H 2 O (3 mL) was added dropwise to a solution of Ga(NO 3 ) 3 ·9 H 2 O (0.139 g, 0.332 mmol) in H 2 O (3 mL). The yellow solution was heated to 85 8C for 2 h. The solution was then cooled to RT, and the volatiles were removed under high vacuum upon gentle heating (ca. 40 8C). The resulting yellow solid was washed with MeCN, and the solution was filtered to remove undissolved particulates, before concentrating this to approximately 50 % volume in vacuo.
Treatment of the solution with diethyl ether led to precipitation of a yellow orange solid, which was isolated by filtration and dried under high vacuum. Yield: 0.032 g, 21 % (orange solid.  MeOH. The insoluble particulates were removed by filtration, and the solution was concentrated in vacuo to give the complex as a yellow solid. Yield: 0.016 g, 58 %; spectroscopic data are as those reported for method 1. Method 4: As was described for method 3, but using Ga(NO 3 ) 3 ·9 H 2 O (0.025 g, 0.060 mmol), Li 2 L (0.021 g, 0.060 mmol) and KF (0.003 g, 0.060 mmol). Yield: 0.012 g, 44 % (yellow solid.); spectroscopic data are as those reported for method 1. Method 5: A solution of KF (0.005 g, 0.086 mmol) in H 2 O (1 mL) and Li 2 L (0.033 g, 0.086 mmol) in H 2 O (5 mL) was added simultaneously to powdered GaCl 3 (0.015 g, 0.086 mmol). Addition of the aqueous solutions resulted in an exothermic reaction and the formation of an orange solution. The mixture was stirred at RT for 90 min, over which time the solution darkened. The volatiles were removed under high vacuum upon gentle heating (ca. 40 8C). The yellow orange solid was washed with MeOH. The insoluble particulates were removed by filtration, and the solution was concentrated in vacuo to give the desired complex as a yellow solid. Yield: 0.026 g, 50 %; spectroscopic data are as those reported for method 1.
ESI + mass spectra were recorded from direct injection of the products onto a Thermo Finnigan mass spectrometer fitted with an LCQ advantage MS pump. NaOAc buffer solutions were prepared by combination of the appropriate volumes of 2 mm NaOAc and 2 mm HOAc.

HLB Purification of [Ga 18 F(L)]
The crude reaction mixture was diluted into NaOAc buffer (10 mL), loaded onto a pre-conditioned HLB cartridge and washed with H 2 O (3 1 mL). The product was eluted with EtOH/H 2 O (2 0.2 mL 1:1). Cartridge purification gave approximately 50 % yield of desired compound in up to 100 MBq mL À1 radioactive concentration.

Stability studies
The HLB-purified product was formulated into a number of solvent compositions of various pH so that the total formulated volume was 1 mL. Aliquots (100 mL) of the formulated product were taken and diluted further prior to injection on to the analytical HPLC.

X-Ray crystallography
Crystals were obtained as described above. Details of the crystallographic data collection and refinement are in Table 1. Rigaku AFC12 goniometer equipped with an enhanced sensitivity (HG) Saturn724 + detector mounted at the window of an FR-E + Super-Bright molybdenum rotating anode generator (l 1 = 0.71073 ) with VHF Varimax optics (70 mm focus). Cell determination, data collection, data reduction, cell refinement and absorption correction: CrystalClear-SM Expert 2.0 r7. [16] Structure solution and refinement were routine by using WinGX and software packages within, [17]