GTM-3, an Extra-Large Pore Enantioselective Chiral Zeolitic Catalyst

The development of chiral zeolitic catalysts possessing extra-large pores and endowed with the capability of enantioselectively processing bulky products represents one of the greatest challenges in chemistry. Here, we report the discovery of GTM-3, an enantio-enriched extra-large pore chiral zeolite material with -ITV framework structure, obtained using a simple enantiopure organic cation derived from the chiral pool, N,N-ethyl-methyl-pseudoephedrinium, as the chiral-inductor agent. We demonstrate the enantio-enrichment of GTM-3 in one of the two enantiomorphic polymorphs using the two enantiomers of the organic cation. Interestingly, we prove the ability of this zeolitic material to perform enantioselective catalytic operations with very large substrates, here exemplified by the catalytic epoxide aperture of the bulky trans-stilbene oxide with alcohols, yielding unprecedented product enantiomeric excesses up to 30%. Our discovery opens the way for the use of accessible chiral zeolitic materials for the catalytic asymmetric synthesis of chiral pharmaceutical compounds.


S3
. XRD patterns of GTM-3 materials obtained with a Si/Ge ratio of 8 in the gel, in the absence or in the presence of seeds. Supplementary Fig. S14. Repetition of enantioselective crystallization: XRD patterns of samples prepared with SS-EMPS or RR-EMPS as SDA in the gel, using 5 % of GTM-3 seeds previously prepared with SS-EMPS or RR-EMPS. Supplementary Fig. S15. Molecular simulations for the selection of the epoxide. Table S2. Molecular dimensions (expressed as averaged molecular shadow lengths during MD simulations) of the different epoxides studied computationally. Table S3. Averaged interaction energies (in kcal/mol epoxide) of one enantiomer of the epoxide molecule loaded in the two enantiomorphic -ITV frameworks.

B) Details of the synthesis of GTM-3
In a typical preparation, 6.36 g of an aqueous solution of (1S,2S)-N,N-ethyl-methyl-pseudoephedrinium hydroxide (28.8 wt %) and 0.57 g of GeO 2 are mixed and stirred for 30 minutes, after which 5.76 g of tetraethylorthosilicate (TEOS) are added and stirred until all the ethanol coming up from the hydrolysis of TEOS and the required amount of water to achieve the desired composition are evaporated. 0.34 g of HF (48 %) are then added and manually stirred (with the help of a spatula) until a homogenous thick gel is obtained, giving a gel composition of 0.25EMPS:0.83SiO 2 :0.17GeO 2 :0.25HF:6.5H 2 O. The gels are introduced into 60 ml Teflon lined stainless steel autoclaves and heated statically at 100 ºC under autogenous pressure for 6 days. The resulting solids were separated by filtration, thoroughly washed with ethanol and water and dried at room temperature overnight. GTM-3 materials could be obtained with Si/Ge ratios of 3 and 5 in the gel. On the other hand, Al-containing GTM-3 catalysts were obtained from gels with (Si+Ge)/Al ratios around 70 with the following molar composition 0.25EMPS:0.007Al 2 O 3 :0.75SiO 2 :0.25GeO 2 :0.20HF:6.5H 2 O; Aluminium isopropoxide was added together with TEOS. For these materials, 10% of GTM-3 seeds were added to the synthesis gels to facilitate crystallization. Calcination of GTM-3 materials was carried out at 550 ºC in air flow for 12 hours; XRD confirmed the resistance of the material to elimination of the organics ( Supplementary Fig. S11). After calcination, GTM-3 samples were manipulated under inert atmosphere to avoid contact with ambient humidity.

C) Details of the general physico-chemical characterization
The obtained solids were characterized by powder X-Ray Diffraction (XRD), using a Philips X´PERT diffractometer with CuK α radiation with a Ni filter. Thermogravimetric analyses (TGA) were registered using a Perkin-Elmer TGA7 instrument (heating rate = 20 ºC/min) under air flow. CHN chemical analyses were carried out with a Perkin-Elmer 2400 CHN analyzer. Liquid NMR spectra were recorded with a Bruker Avance III-HD Nanobay 300MHz spectrometer, using a 5mm HBO 1H/X probe. Solid State MAS-NMR spectra of the solid samples were recorded with a Bruker AV 400 WB spectrometer. 1 H to 13 C CP MAS NMR spectra were recorded using π/2 rad pulses of 2.75 s for 1 H, a contact time of 3 ms and a recycle delay of 4 s. The spectra were recorded while spinning the samples at ca 11.2 kHz. 19 F MAS NMR spectra were recorded using π/2 rad pulses of 2 , and a recycle delay of 120 s, while spinning the samples at ca 20 kHz. Nitrogen adsorption/desorption isotherms were measured at -196 ºC in a Micromeritics instrument ASAP 2420 device after degassing the sample at 200 ºC for 2 hours. Field Emission Scanning Electron Microscopy (FE-SEM) was performed using a Philips XL30 S-FEG. Transmission electron microscopy (TEM) images were obtained (after the samples were well ground and dispersed in ethanol) in a JEOL 2000FX microscope operated with an accelerating voltage of 200 kV. Cs-corrected STEM data was collected in a Titan XFEG spherical aberration corrected microscope. This equipment is fitted with a monochromator which facilitates the control of the electron dose and with a CEOS corrector for the electron probe. Because of the low stability and small size of the GTM-3 crystallites, graphene Cu grids were used for image analysis.

D) Details of the polarimetry study
Retention of the enantiopurity of the SDAs during crystallization was analyzed by determining the optical rotation of the SDAs occluded within the zeolite after the hydrothermal treatment. GTM-3 materials prepared with (S,S)-EMPS (GTM-3/SS-EMPS) or (R,R)-EMPS (GTM-3/RR-EMPS) were dissolved in HF and subjected to a polarimetry study. The optical rotation of the SDA solutions (after zeolite dissolution) was determined by using an optical polarimeter (P1000 LED, A. Krüss-Optronic).

GTM-3/SS-EMPS
1.88 g of GTM-3/SS-EMPS are suspended in 4.03 g of water, and 3.5 g of HF wt 48 % are carefully added with magnetic stirring until complete dissolution of the zeolite; 13 C NMR confirmed the integrity of the released SS-EMPS after this dissolution process. 7 g of NaOH aqueous solution (8 M) are carefully added, yielding a solid precipitate. The solids are removed by centrifugation. The supernatant is taken, and water is added until reaching ~11 g of aqueous solution. Further NaOH aqueous solution is then added until reaching a pH of ~13. 13 C NMR verified the integrity of EMPS obtained by this protocol (Supplementary Fig. S9). The concentration of EMPS hydroxide is then determined by UV-VIS spectroscopy (after building a calibration curve, and 1:50 dilution to reach the lineal regime) as 2.34 wt %. Finally, the specific rotation of the organic salt (normalized to the concentration) is measured with the polarimeter as +68º, very similar in magnitude and sign to the original value of the pristine SS-EMPS hydroxide (+66º), evidencing the retention of enantiopurity of SS-EMPS during crystallization of GTM-3/SS-EMPS.

GTM-3/RR-EMPS
1.83 g of GTM-3/RR-EMPS are suspended in 4.03 g of water, and 3.5 g of HF wt 48 % are carefully added with magnetic stirring until complete dissolution of the zeolite. 8 g of NaOH aqueous solution (8 M) are carefully added, yielding a solid precipitate. The solids are removed by centrifugation. The supernatant is taken, and water is added until reaching ~11 g of aqueous solution. Further NaOH aqueous solution is added until reaching a pH of ~13. The concentration of the organic hydroxide is then determined by UV-VIS spectroscopy as 3.48 wt %. Finally, the specific rotation of the organic salt (normalized to the concentration) is measured with the polarimeter as -58º, S6 similar in magnitude and sign to the original value of the pristine RR-EMPS hydroxide (-64º), evidencing the retention of enantiopurity of RR-EMPS during crystallization of GTM-3/RR-EMPS.

E) Details of the asymmetric catalytic reaction experiments: ring-aperture of trans-stilbene oxide with 1hexanol
Trans-stilbene oxide (trans-2,3-diphenyloxirane) is a chiral reactant and has two enantiomers, (R,R) and (S,S). Acid-catalyzed aperture of trans-stilbene oxide with nucleophiles can proceed via two possible mechanisms : 1,2 i) a concerted S N 2-like mechanism where protonation of O and attack by the nucleophile (1-hexanol) from the opposite side occurs in a concerted way, giving place to (R,S) and (S,R)-u-products (with inversion of configuration) from (R,R) and (S,S)-reactants, respectively, or ii) a S N 1-like mechanim where the epoxide is first protonated, giving place to the corresponding carbocation with a C planar geometry where attack from the nucleophile can occur from both sides, giving place to a mixture of u-[(R,S) and (S,R), with inversion of configuration] or l-[(R,R) and (S,S), with retention of configuration] products. U-species are usually the most abundant products with homogeneous acid catalysts. Catalytic experiments were performed with 20 wt % of catalytst, and 10 mg/ml of racemic trans-stilbene oxide dissolution in 1-hexanol; catalytic reaction was carried out at 30 ºC, and aliquots were extracted at different time intervals. Calcined GTM-3 samples were manipulated under inert atmosphere. Evolution of the reaction was followed by HPLC with chiral stationary phases; (Supplementary Fig. S17). The evolution of the reactants, inversion (u) and retention (l) products was followed by HPLC (Agilent 1200 Infinity Series, with Quad Pump and DAD detector) using chiral chromatographic columns Chiralpak IG and Chiralcel OD-H (Daicel); purity of the peaks was analyzed by UV-VIS spectroscopy. Reactants (RR and SS) and retention (l, RR and SS) products were analyzed with Chiralcel OD-H column, using mobile phases of 97(hexane):3(isopropanol) (flow 1.2 ml/min) at 0 ºC for reactants (λ = 235 nm), and 93(hexane):7(isopropanol) (flow 1.2 ml/min) at 25 ºC for retention products (λ = 215 nm). Inversion (u, RS and SR) products were analyzed with Chiralpac IG column, using a mobile phase of 75(hexane):25(isopropanol) (flow 1.2 ml/min) at 0 ºC (λ = 215 nm) ( Supplementary Fig. S17). Standards for the uand l-products were prepared from the reaction of 1-hexanol with trans-stilbene oxide or cis-stilbene oxide, respectively, catalyzed by homogeneous sulfuric acid; u/l products were characterized by liquid NMR (Supplementary Fig. S16). These standards were used to identify the corresponding peaks in the HPLC analyses.

F) Details of the characterization of acid sites by FTIR spectroscopy analysis of adsorbed pyridine
Fourier transform infrared (FTIR) spectra were recorded with a Thermo Nicolet Nexus 670 FTIR spectrophotometer, equipped with an MCT cryodetector. The as-made sample SS-GTM-3 was pressed into a selfsupporting wafer (thickness ca. 10 mg/cm 2 ), calcined in synthetic air at 550°C for 12 h, transferred under dry nitrogen atmosphere to a transmission IR cell made of glass and provided with CaF 2 windows, and degassed in vacuum (ca. 10 -3 mbar) at 450°C for 12 h. The pre-treated sample wafer was kept in contact with excess pyridine vapor at room temperature for 10 min and then evacuated for 30 min at increasing temperatures from 50 to 300°C. FTIR spectra were acquired in the mid-IR range at 4 cm -1 resolution by averaging 250 scans. The spectra of adsorbed pyridine were computed by subtracting the spectrum of the calcined sample from the spectra collected after pyridine adsorption and thermodesorption. Density of acid sites was calculated from the integral intensity of the band at 1543 cm -1 (Brønsted sites) and at 1452 and 1458 cm -1 (Lewis sites) using the extinction coefficients determined by Emeis. 3 Overlapping bands were deconvoluted into Gaussian components using Microcal OriginPro v9.8 software.  Although ITV was the only crystalline phase observed, XRD patterns of GTM-3 showed the presence of some remaining amorphous material in the samples, which increased upon reduction of the crystallization temperature (compare Figure S5 with Figure S3 and S4). Such unreacted material might be a consequence of the low crystallization temperature required to avoid the thermal decomposition of the EMPS cations (below 110 ºC), which are relatively unstable under hydrothermal conditions. Supplementary Tab. S1. CHN elemental analysis of GTM-3 (two analyses of the same sample are shown); *theoretical C/N ratios are shown between brackets.

% C % H % N C/N* GTM-3_1
14 Thermogravimetric ( Figure S8-A) and CHN elemental analysis (Table S1) of GTM-3 showed a very large organic content of 19.3 %, corresponding to a packing value of 18 EMPS per ITV unit cell; such amount of organic cations is very close to the number of D4R units in the ITV framework (20), and therefore of the negative charges provided by Fanions occluded within those D4R units. TGA (in He) coupled to mass spectrometry ( Figure S8-B,C,D) showed the release of the expected decomposition products of EMPS upon Hofmann elimination. CHN analysis showed a C/N ratio of 12.8 (Table S1) Table S2. Molecular dimensions (expressed as averaged molecular shadow lengths during MD simulations) of the different epoxides studied computationally.
Molecular shadow lengths 1,2-epoxybutane 1,2-epoxyoctane Styrene oxide Trans-stilbene oxide X 7.6 11.5 9.5 13.5 Y 5.3 6.4 6.8 6.9 Table S3. Averaged interaction energies (in kcal/mol epoxide) of one enantiomer of the epoxide molecule loaded in the two enantiomorphic -ITV frameworks; enantiodifferentiation energies (ΔE) are expressed as the difference between the interaction energy of the R enantiomer minus that of the S enantiomer for a given -ITV polymorph. Molecular simulations based on the Dreiding force-field 4 were used in order to guide our selection of a proper chiral epoxide substrate whose chirality is able to be recognized by the enantiomorphic polymorphs of -ITV (P4 1 32 and P4 3 32). Four chiral epoxides with different size (including aliphatic and aromatic chains) were studied as potential reactants: 1,2-epoxybutane, 1,2-epoxyoctane, styrene oxide and trans-stiblene oxide (A). Dimensions of the epoxides are expressed as the three molecular shadow lenghts averaged along 100 ps NVT MD simulations (B). The chiral-recoginition ability of the -ITV framework for the chiral epoxides was estimated by loading one epoxide molecule (each enantiomer separately) on each of the two enantiomorphic ITV polymorphs (P4 1 32 and P4 3 32) and running 1000 cycles of simulated annealing calculations (only hydroxyl groups of the ITV framework and the epoxide molecules were allowed to relax). The interaction energies were calculated by subtracting to the final energy the energy of both the ITV framework and the isolated epoxide molecule, and were averaged along the 1000 annealing cycles (C). Trans-stilbene oxide S17 The condition for ITV to recognize the chirality of the epoxide molecule involves the enantiodifferentiation energies of the two polymorphs to be similar in magnitude but opposite in sign; the higher the energy, the higher the enantio-discrimination ability. For our set of epoxides, the aliphatic 1,2-epoxybutane and 1,2-epoxyoctane did not satisfy such enantio-discrimination condition (the enantiodifferentiation energy were very close to 0), probably because of their size (compared to that of the ITV channels) and the high conformational flexibility. In contrast, both aromatic epoxides, styrene and trans-stilbene oxides, satisfied such condition, and a notably higher enantio-differentiation energy was observed for the latter (~0.7 kcal/mol), probably because of its larger size (B) and the higher rigidity provided by the presence of two aromatic rings. Thus, we selected trans-stilbene oxide as substrate for the catalytic asymmetric reaction. S18 Supplementary Fig. S16. NMR characterization of the stilbene oxide aperture with 1-hexanol. Standards of the products of the trans-stilbene oxide ring aperture with 1-hexanol were prepared using sulfuric acid as catalyst.
A. 13 C NMR of products of the cis-stilbene oxide ring aperture with 1-hexanol catalyzed by sulfuric acid, where the products have l-configuration (R,R and S,S), and the corresponding HPLC chromatogram in the conditions used for its determination (A1). B. 13 C NMR of products of the trans-stilbene oxide ring aperture with 1-hexanol catalyzed by sulfuric acid, where the main products have u-configuration (R,S and S,R), although in this case a minor amount of l-products are also observed (u/l ratio ~ 9), and the corresponding HPLC chromatogram in the conditions used for its determination (B1).

Characterization of acid sites by FTIR spectroscopy analysis of adsorbed pyridine.
The spectra show bands characteristic of pyridine coordinatively bonded to Lewis acid sites at 1613, 1489 and 1452 cm -1 (which evolve into two additional components at 1621 and 1458 cm -1 after evacuation at 100 °C), pyridinium ions bonded to Brønsted sites at 1637, 1543 and 1489 cm -1 and H-bonded pyridine at 1593, 1489 and 1444 cm -1 . 8 Bands corresponding to H-bonded pyridine and pyridinium ions are fully removed after evacuation at 100 and 200 °C, respectively, whereas those corresponding to pyridine coordinatively bonded to Lewis sites are still present after evacuation at 300 °C although strongly decrease at temperatures above 150 °C. Furthermore, additional bands are found at 1591, 1489, 1438 (shoulder) and 1429 cm -1 which retain most of their initial intensity even after evacuation at 300 °C. This later group of bands that, to the best of our knowledge, have not been previously reported for other germano-silicates, should correspond to pyridine more strongly bonded to other acid sites that seem to be characteristic of this material. Supplementary Fig. S22. Enantiomeric excesses (%) of reactants (top), u-products (middle) and l-products (bottom) as a function of conversion, using Al-containing GTM-3 catalysts obtained with SS-EMPS (red), RR-EMPS (blue) or an artificial racemic mixture of both (grey), for the trans-stilbene epoxide aperture with 1hexanol.

S23
Minor deviations from ee = 0 expected for racemic-GTM-3 are due to the fact that the racemic EMPS was artificially prepared from the enantiomerically-pure hydroxide EMPS solutions, and minor errors coming from the determination of the concentration might involve a slight enrichment in one of the enantiomers.