The structure of 3,5-dimethylpyrazole/carboxylic acids co-crystals

Mixtures prepared either by mechanical grinding or by evaporation of equimolar amounts of 3,5-dimethylpyrazole ( 1 ) and five carboxylic acids, four benzoic acids ( 2 - 5 ) and a pyrazole-4-carboxylic acid ( 6 ), were studied by 13 C and 15 N CPMAS NMR spectroscopy. In the cases corresponding to 1 and 2,4,6-trimethylbenzoic acid ( 2 ) or 1 and 2,6-dimethylbenzoic acid ( 3 ) the spectrum of the mixture is different from those of its components and we interpret them in terms of co-crystals formation through donor-acceptor hydrogen bonds. The remaining pairs behave as physical mixtures of both components, the spectrum of the mixture being the sum of the individual spectra. The origin of the differences is the much higher acidity of o,o -disubstituted benzoic acids.


Introduction
The use of hydrogen bonding (HB) as a structural steering force can be considered as the most important strategy in crystal engineering. 1 The patterns formed by carboxylic acids -cyclic dimers and open catemers-2 or by azole derivatives -cyclic dimers, trimers, tetramers and open catemers-have been widely investigated. 3Now we have focused our interest on how the presence of competitive hydrogen bond donor-acceptor groups can modify the HB patterns.
3,5-Dimethylpyrazole (1), a compound having both a hydrogen bond donor [HBD, the N(1)-H] and a hydrogen bond acceptor site [HBA, the -N(2)=] forms trimers 1 3 in the solid state (Figure 1).This structure confers to 1 the ability to transfer, in a concerted manner, the three N-H protons along the hydrogen bond (a phenomenon we named SSPT, solid-state proton transfer). 4ur present interest concerns the structure of the derivatives that result from mixing 3,5dimethylpyrazole (1) with other double hydrogen bonding compounds (HBD and HBA).Recently, we have proved 5 that in the case of imidazoles [HBD, the N(1)-H and HBA, the -N(3)= sites] only the 4,5-dimethylimidazole was able to disrupt the trimer 1 3 formed by 3,5dimethylpyrazole in the solid state.As a continuation of our work concerning the determination of the X-ray molecular structure of 3,5-dimethylpyrazole (1)-2,4,6-trimethylbenzoic acid (2) cocrystal, 6 in the present paper we report a study concerning carboxylic acids (HBD, the O-H and HBA, the O= sites).

N
This complex crystallizes forming a tetramer 1 2 2 2 (Figure 2) and, although a quadruple proton transfer is possible, such SSPT was not observed by solid-state 15 N NMR.The hydrogenbonded network, as determined by crystallography, shows an additional O-H•O-H bond that breaks the symmetry.The chemical shifts of the nitrogen atoms of the 1 moiety appears at -181.7 (NH) and -115.3 ppm (-N=) while in trimer 1 3 they appear at -171.3 and -96.8, 7 this difference was attributed to the replacement of N-H•N by N-H•O=C hydrogen bonds.It is important to note that no proton transfer has occurred, that is, that the tetramer is not a pyrazolium benzoate salt.When searching within the Cambridge Structural Database (November 2003 version) only another paper concerning this problem can be found.In this publication, 8 also from our group, the crystal and molecular structures formed by 1,1'-binaphthyl-2,2'-dicarboxylic acid (BNDA) with 3,5-dimethylpyrazole (1) and pyrazole (7) are described.The most significant result is that the structure depends on the stoichiometry, when it is 1:1, the resulting structure is a salt and when it is 1:2, a neutral complex is obtained both for 1 and for 7.This result proves that the neither the acidity of the acid nor the basicity of the pyrazole alone determines the structure (salt or complex) of the resulting mixture.

Results and Discussion
When trying to prepare equimolar mixtures of 3,5-dimethylpyrazole (1) with the five carboxylic acids (2-6) depicted in Figure 3, three main questions arise: 1. Is the new compound a molecular complex or a salt (pyrazolium carboxylate)?2. If the compound is a neutral complex, does the SSPT occur? 3. Is the stoichiometry of the complex a 1:1 or a higher one (for instance, 2:2)?The 13 C NMR data of 2,4,6-trimethylbenzoic acid (2), 2,6-dimethylbenzoic acid (3) and its sodium salt 3 -, p-methoxybenzoic acid (4), p-chlorobenzoic acid (5) and 3,5-dimethyl-4-pyrazole carboxylic acid ( 6) are gathered in Table 1, while in Table 2 the data of 3,5-dimethylpyrazole (1) are reported.Table 1. 13 C chemical shifts of carboxylic acids 2-6 (*, † , ‡ : pairs of unassigned signals For eventual occurrence of proton transfer, we have measured the anion of 3 (Table 1) and the cation of 1 (Table 2, for an NMR study of the protonation of pyrazoles 11 ).This allowed to determine the following most important chemical shift effects: The [1 + 4], [1 + 5] and [1 + 6] mixtures are physical mixtures and the CPMAS spectra are the sum of the spectra of individual components, therefore, they were only studied by 13 C NMR at 50.32 MHz.
The fact that the 13 C CPMAS spectrum of the [1 + 6] mixture shows no difference from the individual spectra of 3,5-dimethylpyrazole (1) and 3,5-dimethylpyrazole-4-carboxylic acid ( 6) is of particular relevance because pyrazole-4-carboxylic acid (8) also shows SSPT although by a different mechanism. 12The data reported in Table 1, namely the asymmetry of the positions 3 and 5, seems to indicate that a similar phenomenon occurs in 6.
The [1 + 2] mixture (1:1), actually 1 2 2 2 , was studied again.In 15 N CPMAS NMR (40.60 MHz) we obtain a signal at -182.8 ppm (-181.7,N-H [6]) and two signals at -116.8 and -113.9 (-115.3,-N= [6]).The better resolution of the new instrument allows the observation of a splitting of the N(2) signal, possibly due to the small asymmetry of the tetramer (additional O-H•O-H bond).The 13  The [1 + 3] mixture (1:1) is very interesting.First of all it has a 1:1 (or 2:2) stoichiometry.By analogy of the 1 2 2 2 neutral complex previously reported (Figure 2) we assume that it correspond to a 1 2 3 2 tetramer.Its 13 C and 15 N CPMAS NMR spectra reported in Figure 5   In Figure 7 we have represented four possible tetramers corresponding to 1 2 3 2 , all of them without the supplementary O-H•O-H HB present in 1 2 2 2 (Figure 2) together with the predicted 13    None of these structures explain the observed chemical shifts for 1 2 3 2 .Therefore, we tentatively propose a rapid equilibrium (SSPT) between two identical d structures, d1 and d2 (Figure 8).Although the observed asymmetry at the (2 & 6) and (3 & 5) positions of the benzoic acids might be the result of many different sorts of crystalline environments, we cannot find any alternative explanation for the 15 N chemical shifts.Unfortunately, the [1 + 3] mixture is a viscous oil that solidifies on standing after several weeks but not in a crystalline form, preventing to determine its structure by X-ray crystallography.The reason why both 2,4,6-trimethylbenzoic acid (2) and 2,6-dimethylbenzoic acid (3) destroy the 3,5-dimethylpyrazole trimer 1 3 to form new supramolecular structures while the other carboxylic acids 4-6 did not is almost certainly related to the much higher acidities of the first two.Here are the experimental pK a s of some benzoic acids 13 : benzoic acid 4.19, pmethoxybenzoic acid (4) 4.47, p-chlorobenzoic acid (5) 3.98, 2,4,6-trimethylbenzoic acid (2) 3.45 and 2,6-dimethylbenzoic acid (3) 3.35, p-nitrobenzoic acid 3.44.We predict that pnitrobenzoic acid would behave similarly to 2 and 3 when mixed with 1.

Experimental Section
Sample Preparation.The mixtures were prepared using two different methods: i) both components were mixed in a ceramic mortar and ground with a pestle for 10 min until a homogeneous mixture was obtained, ii) the compounds were dissolved in ethanol and the solvent was removed in vacuum.All compounds were commercially available, except 3,5dimethylpyrazole-4-carboxylic acid (6), which was prepared by hydrolysis of the corresponding ethyl ester. 9NMR Spectroscopy Solid state 13 C (50.32 MHz) and 15 N (20.28MHz) CPMAS NMR spectra have been obtained with a Bruker AC-200 spectrometer at 298 K and a 7-mm Bruker DAB-7 probe head, which achieves rotational frequencies of ca.3.5-4.5 kHz.Samples were carefully packed in a ZrO 2 rotors with Kel-F end-caps and the standard CPMAS pulse sequence was used.To observe only the quaternary C-atoms, we run the NQS (Non-Quaternary Suppression) experiments by conventional cross-polarization (CP) at different contact times and with the dipolar dephased technique. 10 13C spectra were originally referenced to a glycine sample and then the chemical shifts were recalculated to the Me 4 Si (for the carbonyl atom δ (glycine) = 176.1 ppm) and 15 N spectra to 15 NH 4 Cl and then converted to nitromethane scale using the relationship: δ 15 N(nitromethane) = δ 15 N(ammonium chloride) -338.1 ppm.Some spectra were recorded again using a 400 MHz instrument: solid state 13 C (100.73 MHz) and 15 N (40.59MHz) CPMAS NMR spectra have been obtained on a Bruker WB-400 spectrometer at 300 K using a 4 mm DVT probehead at rotational frequencies of ca.5-10 kHz.

3 Figure 1 .
Figure 1.Trimer structure of 1 in the solid state.

Figure 4 .
Figure 4.Chemical shift effects in ppm due to the deprotonation of 3 and the protonation of 1.
and 6 are in agreement with this hypothesis.