Applying the Crystalline Sponge Method to Agrochemicals: Obtaining X-ray Structures of the Fungicide Metalaxyl-M and Herbicide S-Metolachlor

The crystalline sponge method is a technique that provides the ability to elucidate the absolute structure of noncrystalline or hard to crystallize compounds through single-crystal X-ray diffraction by removing the need to obtain crystals of the target compound. In this study the crystalline sponges {[(ZnX2)3(2,4,6-tris(4-pyridyl)-1,3,5-trazine)2].x(solvent)}n (X = I, Br) were used to obtain X-ray structures of the agrochemical active ingredients metalaxyl-M and S-metolachlor. The effect of the temperature used during guest uptake and the influence of changing the host framework ZnX2 nodes on guest encapsulation were investigated. Additionally, three compounds containing chemical fragments similar to those of metalaxyl-M and S-metolachlor (phenylacetaldehyde, N-ethyl-o-toluidine, and methyl phenylacetate) were also encapsulated. This allowed for the effect of guest size on the position that guests occupy within the host frameworks to be examined. The disorder experienced by the guest compounds was documented, and an analysis of the intermolecular host–guest interactions (CH···π and π ···π) used for guest ordering within the host frameworks was also undertaken in this study.


S1. Experimental Considerations
To obtain the best quality crystal diffraction data it is imperative to maintain the highest possible quality of the host MOF crystals. To this end all the guest encapsulation experiments were performed in temperature controlled incubators. Controlling the temperature is important to ensure the temperature fluctuations that occur naturally during the day do not potentially create cracks and damage the MOF crystals. The crystalline quality of the MOFs was monitored through visual inspection of the crystal through a microscope and through examination of the diffraction frames produced during SCXRD analysis. Visual inspection of the crystals was performed to look for any obvious signs of crystal degradation, for example cracks on the surface of the crystals such as those seen in Figure  S1. After the visual inspection of the crystals was completed, they were subjected to preliminary SCXRD examination and the diffraction pattern assessed. Good quality diffraction peaks should be observed, these peaks should be sharp and diffract to a resolution at least 0.84 Å ( Figure S2). During the preliminary diffraction scan and the pre-experiment that follows the crystals unit cell parameters are calculated from the diffraction pattern collected. For a good quality crystal, the unit cell parameters should show a high percentage match with the obtained diffraction pattern.
During guest encapsulation the potential inclusion complexes were analysed via SCXRD to assess whether the guest encapsulation had been successful. For each target compound the amount of time required for encapsulation into the host framework is different. Therefore, crystals from each guest encapsulation experiment had to be analysed on several occasions with increasing encapsulation times. This allowed more time for a sufficient quantity of the target compound to enter the hosts pores increasing the possibility for the guest to be located within the hosts framework (Table S1).
Chloroform was used as the host frameworks pore solvent due to it being present in the pores of the as-synthesised MOF and is compatible with all of the target compounds in this investigation. Chloroform is also a labile solvent as it is only possible for a small number of weak CH···π intermolecular interactions to be formed with the host framework. This helps the process of guest encapsulation by allowing the chloroform solvent to leave the hosts pores creating room for the guest molecules to enter. The guest molecules should then out compete the solvent molecules by forming stronger and more numerous intermolecular interactions (CH···π and π···π) with the host framework. The lability of the chloroform solvent also reduces the chance of a guest molecule and a solvent molecule occupying the same average position. If this occurs it greatly increases the challenge of Figure S1. A damaged crystal shown under polarised light of the host framework 1 after incubation in D at 50 °C for 21 days.
solving the crystal structures. Neat liquid guests were also used during guest encapsulation as this produces the highest possible concentration gradient for the guest to enter the MOF.

S5. Guest Position Comparison
The colours of the guest positions can be viewed in Figure 2.
Violet guest molecules Figure S4. Guest

S7. General Crystal Structure Refinement Details
Unit cell determinations, absorption corrections and data reduction were performed using the program CrysAlisPro. 1 Crystal structures were solved within the OLEX2 GUI 2 using direct methods in the program SHELXS 3 and refined by full matrix least squares on the basis of F 2 using the program SHELXL. 4 First the successful location and anisotropic refinement of the host MOF framework was always performed then, the guest and solvent molecules were located. Guest and solvent molecules were initially refined without the use of any restraints or constraints, if a stable refinement could not be achieved then the required restraints and constraints were applied. All guest molecules were constrained using the AFIX 66 command to maintain the known structure of the aromatic ring and all non-hydrogen atoms were anisotropically refined unless otherwise specified. Hydrogen atoms were refined using a riding model. Guest and solvent molecule occupancies were all refined towards the end of refinement using free variables before being fixed to the values quoted. In some structures it was not possible to assign all of the residual density peaks in a way that makes chemical sense, these peaks belong to heavily disordered guests or solvent molecules and were taken into account by the use of SQUEEZE function within PLATON. 5 The individual refinement details for each inclusion complex are given separately.

S8. Crystallographic Tables and Individual structure refinement details
Encapsulation Complex 2.A (guest: phenylacetaldehyde) CCDC: 2046167 One molecule of phenylacetaldehyde was located within the asymmetric unit. The guest molecule shows no disorder and sits in a general position. The occupancy of the guest molecule was refined freely before being assigned to 90%. The DFIX command was employed to maintain realistic bond lengths. The RIGU and SIMU restraints were used to maintain sensible atomic displacement parameters. The bromine atoms Br1 and Br2 attached to Zn1 are disordered over two positions where both disordered parts were refined at 50% occupancy.
At the end of refinement several electron density peaks were not able to be refined in a way that made any chemical sense, these peaks were taken into account by the use of the SQUEEZE function within PLATON. 5 The solvent mask located one significant void within the asymmetric unit of size 851 Å 3 containing 207 electrons.  Four crystallographically independent molecules of N-ethyl-o-toluidine were located within the asymmetric unit. Three of these guests sit in general positions and their occupancies were refined freely at 100%. The fourth guest molecule occupies a site with a centre of inversion positioned next to the carbon atom C72, this means the guest molecule is disordered over two positions related by a centre of inversion and each position was refined with 50% occupancy. Hydrogen atoms were not refined on this guest molecule as the hydrogen atoms attached to C72 where not able to be located in the difference Fourier map, this is likely partly due to the close proximity to a centre of inversion. The DFIX restraint was applied to maintain appropriate bond lengths and the restraints RIGU and SIMU were employed to maintain sensible atomic displacement parameters as well as an EADP constraint to constrain the atomic displacement parameters to similar values. Another guest molecule displays disorder, the carbon atoms C37 and C44 are disordered over two positions with occupancies of 70% and 30% for disordered parts 1 and 2 respectively. The DFIX restraint was applied to maintain appropriate bond lengths and the restraints RIGU and SIMU were employed to maintain sensible atomic displacement parameters. One pyridine ring of the TPT linker of the host framework also displays positional disorder where the ring is disordered over two positions with the atoms C11 and N5 in common, the occupancies of the disordered rings were refined freely before being assigned to 50% each. The FLAT and DFIX restraints were used to model the disorder as well as an EADP constraint to constrain the atomic displacement parameters to similar values. The bromine atom Br1 attached to Zn1 is disordered over two positions where both disordered parts were refined at 50% occupancy.  Four molecules of methyl phenylacetate were located within the asymmetric unit. The complete structures of the four guest molecules were fully refined with occupancies of 50%, 52%, 58% and 59%. All occupancies were refined using free variables. One of the fully refined guest molecules displays disorder, specifically the oxygen atom O6 is disordered over two positions with an occupancy of 25% each. The FLAT command was used to ensure C61 remains flat with the phenyl ring. The DFIX restraint has been used to maintain realistic bond lengths and the RIGU and SIMU restraints used to maintain appropriate atomic displacement parameters. The bromine atoms Br1 and Br2 attached to Zn1 also Br5 and Br6 attached to Zn3 are disordered over two positions where both disordered parts were refined at 50% occupancy.
At the end of structural refinement not all of the residual electron density peaks were able to be assigned in a way that made any chemical sense. These peaks were accounted for by the use of the SQUEEZE function within PLATON. 5 One void was located within the asymmetric unit of size 66 Å 3 containing 12 electrons.  One molecule of metalaxyl-M was located and refined within the asymmetric unit. The guest sits in a general position. The guest occupancy was refined freely at 52%. One molecule of chloroform was also located and refined with an occupancy of 25%, the chloroform molecule displayed no signs of disorder. DFIX and SADI restraints were used to maintain realistic bond lengths, RIGU and SIMU restraints were also employed to maintain appropriate atomic displacement parameters and the FLAT restraint was also employed to keep N13 flat with the phenyl ring. Two EADP constraints were used on two of the pyridine rings of the TPT linker molecule of the host framework to constrain the atomic displacement parameters to similar values. One ZnI 2 comprised of the atoms Zn1, I1 and I2 shows disorder over two positions with both disordered parts refined at 50% occupancy. The iodine atoms I3, I5 and I6 all also show positional disorder over two positions where each disordered part has been refined at 50% occupancy.
The SQUEEZE function was used to take into account of electron density peaks that could not be assigned in a way that makes any chemical sense. One significant void was found with a size of 491 Å 3 which contains 44 electrons.  One molecule of metalaxyl-M and one molecule of chloroform was located and refined within the asymmetric unit of the inclusion complex 2.D. The occupancies of both molecules were calculated freely at 58% and 30% respectively. Neither the molecule of metalaxyl-M or the molecule of chloroform displayed any disorder. DFIX and SADI restraints were employed to maintain realistic bond lengths, RIGU and SIMU restraints were used to maintain suitable atomic displacement parameters. The FLAT restraint was used to keep the atom N13 flat with the phenyl ring of metalaxyl-M. All three ZnBr 2 substituents of the host framework are disordered over two positions, two of the ZnBr 2 substituents were refined with 50% occupancy for each disordered part. The ZnBr 2 substituent with the atoms Zn3, Br5 and Br6 was refined with occupancies of 60% and 40% for parts 1 and 2 respectively.
At the end of refinement several residual electron density peaks were unable to be assigned in a way that would make any chemical sense, these were taken into account using the SQUEEZE function of PLATON. 5 One significant void was found of size 509 Å 3 containing 109 electrons.  One molecule of S-metolachlor was located and refined within the asymmetric unit of the inclusion complex 1.E. The occupancy of the guest molecule was freely refined at 54% and the guest molecule sits in a general position. The molecule of S-metolachlor displays significant disorder where the chloroacetamide is disordered over two positions of the phenyl ring, both disordered parts were refined to 27% occupancy. The nitrogen atoms of the disordered chloroacetamide groups (N12A and N12B) share the same position as the carbon atoms C38B and C49A which represent the disordered ethyl and methyl groups respectively. The second carbon atom of the disordered ethyl group (C37B) and the second methyl group carbon atom (C49B) were unable to be located and refined ( Figure S14). As the guest molecule is incomplete the hydrogen atoms of the guest were not modelled. DFIX, RIGU and SIMU restraints were employed to ensure realistic bond lengths and suitable atomic displacement parameters. The iodine atom I1 attached to Zn1 is disordered over two positions where the disordered parts were refined at 66% and 34% occupancy for parts 1 and 2 respectively. The iodine atoms I3 and I4 attached to Zn2 are disordered over two positions where the disordered parts were refined at 54% and 46% occupancy for parts 1 and 2 respectively. Several electron density peaks remained at the end of structural refinement that could not be assigned in a way that made any chemical sense, these peaks were taken into account by the use of the SQUEEZE function within PLATON. 5 One void was located within the asymmetric unit of size 640 Å 3 which contains 166 electrons.   One molecule of S-metolachlor was located and refined within the asymmetric unit of the inclusion complex 2.E. The guest molecule sits in a general position with an occupancy freely refined at 33%. Four atoms of S-metolachlor could not be located or refined within the crystal structure , three carbon atoms and an oxygen atom, which are identified in Figure S16. As the guest structure is incomplete the hydrogen atoms were not modelled. A FLAT restraint was used on the molecule of S-metolachlor to ensure the atom N13 remains flat with the aromatic plane of the phenyl ring. The restraints RIGU and SIMU were employed on the guest molecule to maintain appropriate atomic displacement parameters and the DFIX restraint used to maintain realistic bond lengths on the guest molecule. All three ZnBr 2 substituents of the host framework display disorder over two positions , the ZnBr 2 substituent containing atoms Zn1, Br1 and Br2 was refined at 50% occupancy for both disordered parts. The ZnBr 2 substituents containing the atoms Zn2, Br3, Br4 and Zn3, Br5, Br6 were refined with occupancies of 60% and 40% for disordered parts 1 and 2 respectively.
Multiple electron density peaks could not be assigned in a way that made any chemical sense by the end of refinement, these peaks were taken into account by the use of the SQUEEZE function within PLATON. 5 One void of size 777 Å 3 was located which contains 180 electrons. Figure S18. A structure diagram of S-metolachlor with the atoms that could not be located in the crystal structure of inclusion complex 2.E identified within red circles.