Synthesis, Isotopic Enrichment, and Solid-State NMR Characterization of Zeolites Derived from the Assembly, Disassembly, Organization, Reassembly Process

The great utility and importance of zeolites in fields as diverse as industrial catalysis and medicine has driven considerable interest in the ability to target new framework types with novel properties and applications. The recently introduced and unconventional assembly, disassembly, organization, reassembly (ADOR) method represents one exciting new approach to obtain solids with targeted structures by selectively disassembling preprepared hydrolytically unstable frameworks and then reassembling the resulting products to form materials with new topologies. However, the hydrolytic mechanisms underlying such a powerful synthetic method are not understood in detail, requiring further investigation of the kinetic behavior and the outcome of reactions under differing conditions. In this work, we report the optimized ADOR synthesis, and subsequent solid-state characterization, of 17O- and doubly 17O- and 29Si-enriched UTL-derived zeolites, by synthesis of 29Si-enriched starting Ge-UTL frameworks and incorporation of 17O from 17O-enriched water during hydrolysis. 17O and 29Si NMR experiments are able to demonstrate that the hydrolysis and rearrangement process occurs over a much longer time scale than seen by diffraction. The observation of unexpectedly high levels of 17O in the bulk zeolitic layers, rather than being confined only to the interlayer spacing, reveals a much more extensive hydrolytic rearrangement than previously thought. This work sheds new light on the role played by water in the ADOR process and provides insight into the detailed mechanism of the structural changes involved.


S3
The PXRD pattern for the as-made 29 Si-enriched Ge-UTL (acquired as described in the main text) is shown in Figure S1.1. Figure S1.2 shows the PXRD pattern for the Ge-UTL (natural abundance) after calcination, along with the simulated pattern for a typical UTL framework from the IZA database. S2 Figure S1.1. PXRD pattern for as-made 29 Si-enriched Ge-UTL. To minimize the amount of 17 O-enriched water that would ultimately be required, the hydrolysis reaction had to be scaled down, and was carried out (initially with naturalabundance zeolite and H 2 O) using the conditions given in Table S1.3, resulting in a total acid concentration of 6 M, for reaction times between 4 and 48 h. Reactions were carried out using a 10 ml round-bottomed flask topped with a condensing tube in refluxing conditions at 95 °C. The amount of washing water was also minimized in an attempt to S4 avoid possible loss of 17 O. Calcination of the hydrolysed products was carried out to remove any remaining water, and to allow the condensation of the framework. Typically, the zeolite was heated to 575 °C at a rate of 1 °C / min, held for 6 hours and cooled to room temperature at a rate of 2 °C / min under an atmosphere of air.  Table S1.4.
Calcination of the hydrolysed products was carried out to remove any remaining water from hydrolysis, and to allow the condensation of the framework. PXRD patterns (expanded to show only the region corresponding to the 200 peak) of the calcined hydrolysed materials are shown in Figure S1.4, with the d spacings (d 200 ) given in Table   S1.4.

S7
The repeatability of the results obtained from the small volume ADOR hydrolysis reactions was tested by carrying out a second hydrolysis for 16 h. The Q 4 /Q 3 ratio of the hydrolysed material determined from the 29 Si MAS NMR spectrum (as described in the main text) was 5.7 for the second sample, in good agreement with the 5.4 obtained previously. Good agreement was also obtained for calcined samples, where this ratio was 5.8. The PXRD data showed consistent d 200 spacings for the two samples (12.0 Å and 11.9 Å, for the first and second as made samples, respectively, and 11.7 Å and 11.8 Å for the calcined materials).
For the calcined samples the average micropore volume was 0.18 ± 0.02 cm 3 g -1 , characteristic of the IPC-2 structure. A typical adsorption isotherm is shown in Figure S1.5.
To determine whether the small amount of washing water had an effect on the final porosity of the structures, the 16 h ADOR hydrolysis was repeated using 40 ml of washing water. The d 200 spacing observed for this sample in its hydrolysed form was 12.0 Å, and the micropore volume 0.17 cm 3 g -1 . These values are in good agreement with those obtained previously. Figure S1.5. 77 K N 2 adsorption isotherm (for the Ge-UTL sample hydrolysed for 4 h and subsequently calcined).

S8
Hydrolysis using (41% enriched) H 2 17 O was carried out on natural-abundance Ge-UTL and 29 Si-enriched Ge-UTL for 16 h using 6 M HCl (as shown in Table S1. shown in Figures S1.6 and S1.7, respectively. The d 200 spacings observed for these two samples are 11.9 Å and 11.8 Å, in good agreement with the data given in Table S1.4. Figure S1.6. Capillary PXRD pattern of the as-made 17 O-enriched zeolite hydrolysed for 16 h, using the conditions given in Table S1.3. Reflection positions match that for IPC-2P.  Table S1.3. Reflection positions match that for IPC-2P. spectra (acquired for another sample that was prepared in the same way) shown in Figure   S2.2 reveal this peak exhibits little field dependence, suggesting it has a negligible C Q and confirming it can be attributed to H 2 O. Average NMR parameters <P Q > and <δ iso >, extracted from the position of the centre of gravity of the lineshape seen in the MQMAS spectra, are given in Table S2.1. This resonance can be assigned to Si-O-Si species, expected to be found primarily in the bulk of the zeolitic layers. The Si-OH signal is not observed in MQMAS spectra unless 1 H decoupling is applied.

S2. Variable-field 17 O MAS and MQMAS spectra
Note that P Q is a combined quadrupolar parameter that depends on both the magnitude and asymmetry of the quadrupolar interaction, with P Q = C Q (1 + η Q 2 /3) 1/2 . This parameter can be determined from the position of the centre of gravity of the lineshapes within an MQMAS spectrum, whilst the determination of C Q and η Q individually requires the fitting of cross sections through the lineshapes. This latter process can be difficult for disordered materials and P Q (and <P Q >) is often used in preference.   Figure S2.1 (c) and (d).  shown in Figure S3.1a (black line) indicates the presence of a distribution of Q sites with approximate intensity ratios of 0.05% : 2.92% : 73.5% : 23.53% for Q 1 : Q 2 : Q 3 : Q 4 species.

B0 / T δ iso (ppm) CQ / MHz
This distribution is confirmed by the 1 H-29 Si CP MAS spectrum shown in Figure S3.1a (red line) and appears to show a greater proportion of Q 3 than spectra reported for other amorphous silica samples obtained using a similar synthetic procedure. S5 Figure S3.1b shows the 1 H MAS NMR spectrum recorded 10 days after synthesis, where a broad line is observed, indicating the presence not only of isolated and hydrogen-bonded silanols, but also of physisorbed water with varying degrees of hydrogen bond strengths (signals in the range 3-8 ppm). S6 The 17 O MAS NMR spectrum is shown in Figure S3  magnetization was generated using a rf pulse with ν 1 = 50 kHz, before a spin-locking pulse was applied, with ν 1 = 8 kHz or 45 kHz in Figures S3.2a and S3.2b, respectively. In Figure   S3.2a, after a sharp drop in signal intensity (owing to initial dephasing of terms that do not commute with the spin-locking Hamiltonian, as described in Ref. S8) a reasonable spin lock is observed. The behaviour seen is characteristic of the sudden limit and the adiabaticity parameter, α (= ν 1 2 /2ν Q PAS 2ν R ) is 4.3 × 10 −3 . When the spin lock is applied with a higher rf field, the drop in signal intensity is more significant and a pronounced oscillation with the rotor period confirms the system is between the intermediate and adiabatic regimes (α ≈ 0.127).  acquired with a CP rf field strength, ν 1 = 10 kHz, and contact times of 100 µs (blue) and 500

S16
µs (red). The 17 O MAS NMR spectrum is also shown for comparison (black).

S18
S4. T1 ρ measurements and spin-locking behaviour Figure S4.1 shows the intensity resulting from a 1 H spin-lock experiment (or T 1ρ measurement) for a hydrolysed Ge-UTL zeolite. Only a small drop in intensity is observed over a spin-lock duration of 1 ms, demonstrating that rapid 1 H T 1ρ relaxation is not responsible for the poor 1 H-17 O CP efficiency observed in Figure 6 of the main text. Figure   S4.    From the 2 H MAS NMR spectrum in Figure S5.2, it can be estimated that the S22 intensity ratio of the Si-O-D: D 2 O signal is ~1:4 (suggesting a 1:2 ratio of Si-O-H groups to molecular water in the interlayer spacing). It should be noted, however, that the level of water in the hydrolysed zeolites varies with hydrolysis duration and storage time and conditions. From samples studied in this work it is estimated that the ratio of Si-O-H groups to molecular water in the interlayer spacing ranges between 1:2 and 1:4.  Similar fits were also carried out for the 17 O MAS NMR spectra shown in Figure 4 of the main text. Although these spectra are not truly quantitative (as a short flip angle pulse was not used), the fits do provide insight into the changes in the lineshape observed over a S24 time period of 30 days. As shown in Figure S6.2 and Table S6.1, the relative proportion of the three components changes with time, with an increase in the proportion of Si-O-H groups and a relative decrease in Si-O-Si signal. This reflects a low level of ongoing hydrolysis, most likely as a result of small amount of acid remaining between the layers, owing to the reduced volume of (unenriched) washing water used. Si-OH species.  Figure S6.2 (and Figure 4 of