Amorphous silica-aluminas (ASAs) are widely used in acid-catalyzed C-H activation reactions and biomass conversions in large scale, which can be promoted by increasing the strength of surface Brønsted acid sites (BAS). Here, we demonstrate the first observation on a synergistic effect caused by two neighboring Al centers interacting with the same silanol group in flame-made ASAs with high Al content. The two close Al centers decrease the electron density on the silanol oxygen and thereby enhance its acidity, which is comparable to that of dealuminated zeolites, while ASAs with small or moderate Al contents provide mainly moderate acidity, much lower than that of zeolites. The ASAs with enhanced acidity exhibit outstanding performances in C–H bond activation of benzene and glucose dehydration to 5-hydroxymethylfurfural, simultaneously with an excellent calcination stability and resistance to leaching, and they offer an interesting potential for a wide range of acid and multifunctional catalysis.
Silica-alumina materials, particularly crystalline zeolites and amorphous silica-aluminas (ASAs), are among the most popular solid acids that have been widely commercialized as efficient and environmentally friendly catalysts in the petrochemical industry1, and in bio-refinery2. These materials can provide Brønsted acid sites (BAS) with tunable density and strength, which facilitates the optimization of the surface acidity to promote a series of important industrial chemical reactions, through the formation of surface complexes or transition states by proton transfer from BAS to reactants3,4,5,6, such as to initialize C–H activation for hydrocarbon conversions7,8,9,10,11,12. Zeolites with strong Brønsted acidity, are of increasing importance in various sustainable processes, in the fields of biomass conversion, CO2 capture and conversion, air-pollution remediation, and water purification13. For instance, zeolites can efficiently catalyze the redox disproportionative conversion of biomass-derived sugars into α-hydroxy acids14, and they are more active than ASA catalysts, albeit the latter facilitate improved molecular diffusion in the porous network15. The lower performance of ASA in many catalytic applications is widely attributed to their moderate Brønsted acidity1, which fostered recent works on the discovery of ASAs with increased Brønsted acidity5, aiming at expanding their applications in a broad range of fields.
The formation of BAS in silica-aluminas is based on aluminum centers distributed in the silica framework or network6,16,17,18,19,20, as (i) a tetra-coordinated aluminum species (AlIV), replacing a Si4+ atom in the zeolite framework, can introduce a negatively charged framework oxygen to be balanced by a proton, resulting in bridging OH groups6,17,18, and (ii) the interaction between AlIV atoms and neighboring silanols in the silica network can provide acidic OH groups, acting as BAS in ASA5,21,22,23,24. In crystalline zeolites, it is well accepted that increasing the Si/Al ratio can enhance the BAS strength by increasing the overall electronegativity25. However, the amount of BAS (e.g. bridging SiOHAl groups) is then significantly reduced, owing to the fewer Al sites in the framework. Alternatively, the introduction of extra-framework cations, such as Al3+ and La3+, via dealumination or ion exchange can significantly improve the BAS strength due to a synergistic effect between Lewis acid sites (LAS) and nearby BAS4,17. Those solid acids have been widely applied in gas-phase cracking, such as fluid catalytic cracking processes26. However, extra-framework cations can easily leach out from solid acids during the liquid-phase reactions27. Additionally, the synergy between two nearby Al sites in the zeolite framework is impossible due to the absence of Al-O-Al linkage based on Löwenstein’s rule28.
Enhancing the BAS strength in ASAs still remains a significant challenge. Although a surface bridging SiOHAl model has been proposed for the formation of BAS on ASA21,29, the strength of the BAS on ASA is generally lower than that on crystalline zeolites1, since the amorphous structure of ASAs weakens the Al-O bonds (2.94–4.43 Å)24, compared to those in the crystalline zeolite framework (1.88–2.0 Å)30. Two models have been proposed to account for BAS generation in ASA: (i) a flexible coordination between the Al atom and the neighboring silanol oxygen atom5,23, and (ii) a pseudo-bridging silanol (PBS) with a nearby Al atom24,31,32. In both models, it was proposed that a Lewis acidic Al center interacts with a nearby silanol group, withdrawing electron density from the hydroxyl O atom to enhance the acid strength of the hydroxyl proton. Notably, these models mainly account for the formation of moderate BAS. Nevertheless, the presence of an additional Al species in the vicinity of the SiOH site may further enhance the acid strength of ASA via a synergistic effect, which has not yet been evidenced to the best of our knowledge.
In this work, the synergy between Al species in the ASA network has been studied using solid-state NMR spectroscopy and atom probe tomography (APT). The 27Al double-quantum single-quantum (DQ-SQ) through-space homonuclear correlation (D-HOMCOR) NMR experiments allow us to probe 27Al-27Al proximities by applying recoupling sequences that restore the dipolar interaction between neighboring 27Al spins33,34,35,36. Unlike for crystallized materials, the investigation of the location and distribution of Al atoms is impossible in ASAs by routine characterization methods, such as X-ray diffraction (XRD) and high-resolution transmission electron microscopy (HRTEM), due to their amorphous structure. Atom probe tomography (APT) can provide quantitative three-dimensional (3D) information on elemental distributions in catalyst nanoparticles at the atomic scale37,38. It has recently been employed with sub-nanometer-scale resolution on zeolite-based catalysts to establish their structure–composition–property relationships39,40,41,42,43. Here, we apply APT for the same purpose in ASAs44,45,46,47. The combined investigations of APT on Al distribution and 27Al and 1H DQ-SQ NMR experiments reveal, for the first time, the existence of a synergy between Al species in the ASA network. This synergy can significantly enhance the acid strength of ASA, as demonstrated by the H/D exchange with deuterated benzene. The beneficial effect of the enhanced acidity is demonstrated by an example of the liquid-phase catalytic dehydration of glucose to 5-hydroxymethylfurfural, which is an important building block in the production of various valuable chemicals, such as liquid alkanes, biofuels, and furan derivatives48.
Local structure of SA/10
The ASA materials were prepared by flame spray pyrolysis as described in Supplementary Methods and they are designated as SA/x, where x = 10 or 50 represents the percentage of Al atoms with respect to the total amount of Al and Si atoms in the precursor. APT has been applied to show the distribution of Al within the SA/10 as shown in Supplementary Fig. 1 and Supplementary Movie 1 in the Supplementary Information. The tomographic reconstruction qualitatively shows a homogeneous distribution of Al, Si and O (Supplementary Fig. 1b, c, d), where each sphere represents the 3D position of an individual atom. In agreement with recent energy-dispersive X-ray atom mapping investigations49, the APT reveals a homogeneous dispersion of Al species in the silica network, similar to that observed in well-developed crystalline zeolites42,43. Due to the lack of evidence for the existence of zeolite-like bridging OH groups in ASA, Al atoms in SA/10 can be expected either bridged to SiO (SiOAl) or located nearby the silanol groups (SiO(H)∙∙∙Al), as reported in literature16,21.
Supplementary Fig. 2a displays the single-pulse 1D 27Al MAS spectrum of SA/10. It exhibits three peaks at 50, 30, and 4 ppm, assigned to tetra- (AlIV), penta- (AlV) and hexa- (AlVI) coordinated Al sites, respectively16,50. In the DQ-SQ 2D spectra, as shown in Fig. 1, autocorrelation diagonal peaks resonating at the frequencies (2ν, ν) along the indirect and direct dimensions, respectively, indicate the proximities between nuclei with identical isotropic shifts, ν, whereas the cross-peak pairs resonating at frequencies (νa + νb, νa) and (νa + νb, νb) specify proximities between nuclei resonating at distinct frequencies, νa and νb, in the 1D spectra. The 27Al DQ-SQ 2D spectrum (Fig. 1a) exhibits a weak diagonal peak at (60, 30) ppm, which indicates proximities between AlV sites in SA/10. No other obvious 27Al correlation signal could be detected, which was confirmed by the 27Al slices that were extracted at the corresponding shifts of these correlation signals (Supplementary Fig. 2e–i). The 1H MAS spectrum of SA/10 shown in Supplementary Fig. 2b exhibits a single peak resonating at 1.9 ppm, which is ascribed to the silanol protons. The 1H DQ-SQ 2D spectrum shows a single strong signal at (3.8, 1.9) ppm (Fig. 1b). This peak indicates the spatial proximity of many silanol groups on the surface.
The presence of BAS has been further confirmed by the adsorption of a basic molecular probe, ammonia. The 1H spectrum of SA/10 loaded with ammonia (see Supplementary Fig. 2d) exhibits a peak at δ1H = 6.7 ppm, which is ascribed to ammonium ions6. The presence of an auto-correlation peak at (13.4, 6.7) ppm in the 1H DQ-SQ spectrum of Fig. 1d indicates that the ammonium protons are dipolar coupled and hence, do not exhibit isotropic motion in the sample. The cross-peaks at (8.6, 1.9) and (8.6, 6.7) ppm in the same spectrum demonstrate mainly that some ammonia ions stay adsorbed near the SiOH groups. The peak at δ1H = 2.6 ppm is assigned to ammonia adsorbed at LAS51,52,53,54. The autocorrelation peak at (5.2, 2.6) ppm in Fig. 1d indicates that the protons of adsorbed ammonia are dipolar coupled and hence that these molecules do not exhibit isotropic motions. Conversely, the 27Al 1D (Supplementary Fig. 2c) and 2D D-HOMCOR (Fig. 1c) spectra are not significantly modified by the adsorption of ammonia. In particular, we only detect a weak AlV diagonal peak in the 2D D-HOMCOR spectrum, and no other correlation signal could be detected as observed with dehydrated SA/10 (Fig. 1a).
The 13C NMR signal of SA/10 loaded with CH313COCH3 probe molecule resonates at δ13C = 213 ppm5. This value is similar to those observed in most ASA samples22, but is much smaller than that of zeolite H-ZSM-5 (δ13C = 223 ppm)55. This chemical shift δ13C value is commonly utilized to evaluate the strength of acid sites in solid acids, e.g., a larger δ13C value indicates a higher acid strength6. Hence, the BAS of SA/10 exhibits a moderate acidity. These sites have been described as SiOH groups in the proximity of one AlIV or AlV site16.
Local structure of SA/50
The APT reconstructions of SA/50 in Fig. 2, Supplementary Fig. 3, and Supplementary Movie 2 show the 3D distributions of Al, Si and O species. Visually, the Si and O species are homogeneously distributed, similarly to those observed in Supplementary Fig. 1 for SA/10, however, the Al atoms are distributed rather heterogeneously. The comparison of these APT reconstructions with those in SA/10 (Supplementary Fig. 1) shows a higher Al density in SA/50. Additionally, a radial distribution function (RDF)56,57 was calculated to further evaluate the clustering tendency of Al (Supplementary Note 1). The RDF of Al in SA/10 shows no significant positive or negative correlation, which is in a good agreement with the random Al distribution in the silica network without significant clustering in SA/10. The RDF of Al in SA/50 has a high positive correlation indicating that more and more Al species are close to each other. Indeed, at high Al concentration, more Al species in the network lead to shorter average Al-Al distance in SA/50 than in SA/10. Hence, more than one Al center can be expected in the proximity of a SiOH group in SA/50.
The 27Al 1D NMR spectrum of SA/50 (Supplementary Fig. 4a) indicates the presence of AlIV, AlV and AlVI species in that sample. Moreover, the relative amount of AlV sites is higher in SA/50 than in SA/10 (compare Supplementary Figs. 2a and 4a). Various correlation peaks were detected in the 27Al DQ-SQ 2D spectrum of dehydrated SA/50 as shown in Fig. 3a, providing information about the proximities between the different Al sites. The most intense peak is the diagonal one of AlV site at (60, 30) ppm (Fig. 3d). When normalized by the number of transients and the Al molar fraction, the intensity of that peak is 3-fold higher for SA/50 than for SA/10 in the spectrum of Fig. 1a. This higher intensity indicates a shorter average distance between the closest AlV sites in SA/50 than in SA/10, in line with the APT data (Supplementary Note 1). The pair of intense cross-peaks at (85, 55) and (85, 30) ppm also indicates that a significant amount of AlV species (δ27Al = 30 ppm) is close to AlIV ones (δ27Al = 55 ppm) (Fig. 3e). An AlV-AlVI correlation is also detected at (34, 4) and (34, 30) ppm (Fig. 3c). The weak cross-peaks at (59, 55) and (59, 4) ppm detected in Fig. 3d also indicate the proximity between AlIV and AlVI species. The weak diagonal signals at (8, 4) or (110, 55) ppm point to proximities between two AlVI or two AlIV sites, respectively. However, owing to the low density of AlVI and AlIV species, these peaks are very weak, as shown in the corresponding slices of Fig. 3b, f, respectively.
The 1H 1D NMR spectrum of dehydrated SA/50 (Supplementary Fig. 4b) is dominated by the signal of SiOH groups resonating at 1.9 ppm. The shoulder at 1.1 ppm is assigned to non-acidic terminal AlVIOH protons16. The 1H DQ-SQ spectrum in Supplementary Note 2 exhibits an intense autocorrelation peak at (3.8, 1.9) ppm, indicating close proximity between silanol protons. As mentioned above, the RDF data in Supplementary Note 1 indicates that almost all network Si atoms are close to Al ones. Furthermore, it has been shown for SA/50 using 1H-27Al through-space correlation experiments at 18.8 T that the silanol protons are mostly close to AlIV and AlV sites16. Hence, the intense autocorrelation peak for silanol protons in the 1H DQ-SQ 2D spectrum is consistent with the intense AlIV-AlV and AlV-AlV correlations detected in the 27Al DQ-SQ 2D spectrum (Fig. 3d, e).
As already observed for SA/10, the loading of SA/50 with ammonia leads to the appearance of a signal of ammonium protons at 6.7 ppm in the 1H 1D NMR spectrum. The relative intensity of this signal with respect to that of silanol is higher for SA/50 than for SA/10, which indicates the increased amount of BAS in SA/50 in agreement with previous studies5. The 1H DQ-SQ 2D spectrum of SA/50 loaded with ammonia is shown in Supplementary Note 2. The significant increase of the broad auto-correlation signal at (2.6, 5.2) ppm with increasing Al/Si ratio from 1/9 to 5/5 (compare Fig. 1d and Supplementary Note 2) allows assigning it to ammonia adsorbed on LAS51,52,53,54. Such assignment is supported by previously reported 2D 27Al-1H through-space correlations of SA/10 and SA/50 loaded with ammonia16. In those spectra, the 1H signal resonating at 2.6 ppm is mainly correlated with AlV sites, acting as LAS. After ammonia adsorption, the 27Al DQ-SQ 2D spectrum is broadened (Supplementary Fig. 5). However, the majority of the AlIV-AlV and AlV-AlV correlations are still observed. In the 1H DQ-SQ spectrum of dehydrated SA50 (Supplementary Note 2) no signal of Al(OH)Al groups at ca. 1.7–2.7 ppm could be detected6, indicating a low probability of Al-Al correlations originating from AlIV-OH-AlV and AlV-OH-AlV groups in alumina domains. Furthermore, these sites do not protonate ammonia and hence, the corresponding 27Al correlation will not be significantly broadened in the presence of ammonia. Conversely ammonia can be protonated by surface BAS (SiOH with nearby Al) on ASAs, which explains the peak broadening observed in the 27Al DQ-SQ 2D spectrum after ammonia adsorption (Supplementary Fig. 5). This broadening has been proposed by the synergy of AlIV-AlV and AlV-AlV spin pairs in the local structure of the same SiOH group (Al-SiOH-Al) with enhanced acid strength.
Acidity enhancement by the synergy of nearby Al species
In SA/10, as in most ASAs, only a moderate acidity strength (δ13C = 213 ppm probed with CH313COCH3) was detected5. Figure 1 and Supplementary Fig. 1 show that the AlIV and AlV species are well-distributed on ASA without obvious correlations, e.g. SA/10. Moreover, bridging OH groups (δ1H = 3.6–5.2 ppm) were not detected in 1H NMR 1D experiments (Supplementary Fig. 2b). Therefore, the moderate BAS strength in SA/10 is proposed to be generated by one Al center interacting with the neighboring silanol group and in that way decreasing the electron density of the O atom, leading to the formation of one acid SiOH site (BAS), as shown in Fig. 4a. This arrangement is similar to the PBS model proposed in previous theoretical calculation studies, where a SiOH group electrostatically interacts with an acceptor Al center (AlIV or AlV), but is not covalently bonded as bridging OH groups in zeolites24,31.
In SA/50, two Al centers (two AlV or one AlV and one AlIV) can be in the proximity of the same SiOH group. In a previous study, based on DNP (dynamic nuclear polarization) and first-principle calculations, several possible models for two or more Al centers in the vicinity of a SiOH group on ASA, prepared by chemical liquid deposition of SiO2 on Al2O3, have been proposed by Valla and coworkers21. In ASAs, BAS are formed at the surface consisting of mixed alumina and silica and thus a wide distribution of Al species in the local structure of Si species can be expected, particularly in the Al-rich phase, as exemplified for ASAs prepared by chemical liquid deposition of SiO2 on Al2O321. However, these models do not account for the enhanced Brønsted acidity of flame-made ASAs with high Al content, such as SA/50. Therefore, here we propose a structural model of BAS, in which AlIV and AlV sites or two AlV sites interact with the same SiOH group (see Fig. 4b), most likely via pseudo-bridging OH groups as often proposed24,32, and withdraw electrons from the oxygen of the near SiOH group, thus enhancing its Brønsted acidity. The structural models shown in Fig. 4 bear similarities with oxygen tri- and tetra-coordinated clusters, which have been proposed earlier in ASA prepared by chemical liquid deposition of SiO2 on Al2O3 and aluminosilicates glasses21,58. The formation of oxygen tri- or tetra- coordinated clusters is driven by the increased ionicity of these materials at higher Al content due to the difference in electronegativity between Si and Al atoms31,59.
However, one or more Al centers solely covalently bound to the Si atoms of the silanols are unable to protonate adsorbed ammonia, since the silanolate cannot be efficiently stabilized after deprotonation31,60. Proton transfer from silanol to the guest molecule can be promoted by the stabilization of the conjugated base (silanolate). The silanolate can be stabilized by a neighboring unsaturated Al center via PBS model31,60,61,62, such as proposed in Fig. 4a, which is often characterized by moderate acidity.
Here, we propose a model as shown in Fig. 4b, for the formation of BAS on SA/50 with an acidity comparable to zeolites. Compared to Fig. 4a for SA/10, the second neighboring unsaturated Al center involved as an acceptor is able to further stabilize the formed silanolate, and thus to promote the proton transfer. Besides, the extra Al center(s) may further withdraw electrons from the O atom from neighboring SiOH that significantly enhances the BAS strength (up to δ13C = 227 ppm)5, to a higher extent than that of zeolite H-ZSM-5 (δ13C = 223 ppm)55, similar to that of dealuminated zeolite H-Y (δ13C = 228 ppm)17. The proposed model is similar to those reported for zeolites where an ionic effect induced by extra-framework Al species can enhance the Brønsted acid strength of bridging OH groups17. Therefore, the synergy of two AlV or one AlIV and one AlV centers with nearby SiOH groups is expected to significantly enhance the acid strength of BAS in the ASA.
In situ 1H MAS NMR study on acidity enhancement of BAS
In this work, we show that the proximity between one SiOH group and two Al sites can remarkably enhance the Brønsted acidity in ASAs5,16. The activation of the C-H bond in hydrocarbon conversion often requires solid acids containing strong BAS. The activation of the C-H bond in benzene, the simplest aromatic compound, has been extensively studied using H/D exchange experiments, which is of great importance to understand the alkylation processes of aromatic compounds8,9,10,11,12. The H/D exchange carried out between C6D6 and surface BAS (bridging OH on zeolite H-ZSM-5 and acidic SiOH groups on ASA) was confirmed by 1H solid-state NMR (Supplementary Fig. 6).
On H-ZSM-5 zeolite, the three 1H signals of H-ZSM-5 zeolite at 7.5, 4.0 and 1.8 ppm (see Supplementary Fig. 6a) are assigned to hydrogen atoms bound to aromatic rings, bridging OH (SiOHAl) groups and terminal SiOH groups, respectively. After H/D exchange reaction, the intensity of SiOH groups remained unchanged while that of bridging OH groups decreased with increasing intensity of the aromatic hydrogens. This indicates that the H/D exchange occurred between the benzene-d6 and the bridging OH groups (e.g. SiOHAl), rather than with the terminal SiOH groups (δ1H = 1.8 ppm).
Conversely, no bridging OH groups could be observed at 3.5–5.2 ppm on SA/50 (see Supplementary Fig. 6b)6, while the signal of protons bound to the aromatic rings was observed at 7.3 ppm. As shown in the stack plot of the 1H MAS spectra recorded during H/D exchange of benzene-d6 loaded over dehydrated SA/50 (Fig. 5a), the intensity of the terminal SiOH groups (δ1H = 1.8 ppm) decreased as a function of time, while that of protons bound to aromatic rings (δ1H = 7.3 ppm) increased. This demonstrates that the H/D exchange occurred between C6D6 and the acidic SiOH groups of ASA. It must be reminded that Lewis acid aluminum sites cannot exchange H with C6D6 (see Supplementary Note 3).
The reaction rate, k, for the C6D6–SiOH exchange can be utilized to evaluate the relative strength of BAS in silica-alumina catalysts under the same conditions. As described in the Methods section, one molecule of C6D6 per Brønsted acid site was quantitatively loaded, and thus, a higher k value indicates a higher acid strength63,64,65. The rate of the H/D exchange reaction over different catalysts upon loading one C6D6 per BAS at 313 K was determined by fitting the evolution of the signal intensity of aromatic protons as a function of reaction time9,12. The obtained k values are shown in Fig. 5b. Evidently, SA/10 with mainly moderate BAS was virtually inactive in the reaction, resulting in a small k value, while zeolite H-ZSM-5 with strong BAS provided a much higher reaction rate (k = 0.0072 min−1). A similar value, k of 0.0078 min−1, was obtained with SA/50 under the same conditions, hinting to the existence of strong BAS in SA/50 with a strength comparable to that of zeolite H-ZSM-5. Considering the structural difference between SA/10 and SA/50, it shows that the proximity between more than one Al center and SiOH group in ASA with high Al content, such as SA/50, gives rise to BAS with zeolitic strength.
Stability test of ASA catalysts
Besides their enhanced surface Brønsted acidity, the high stability of ASAs under various conditions is crucial for efficient catalysis. Current ASAs were thus firstly tested in liquid-phase glucose dehydration to 5-hydroxymethylfurfural (HMF), requiring LAS for glucose isomerization to fructose and BAS for fructose dehydration48. The catalytic reaction results are summarized in Supplementary Note 4, 5, and Supplementary Fig. 7. The reusability of ASA was tested with SA/50, which exhibited the best performance under the same conditions. After five recycle runs, no significant loss of catalytic activity could be observed (Supplementary Note 5). This is attributed to the high stability of SA/50 in the liquid phase glucose dehydration. In liquid phase reactions under heating, dealumination could cause the modification of surface acid sites. The comparison of Supplementary Fig. 12a, b demonstrates that aluminum species in SA/50 are stable without dealumination upon water treatment at 433 K for 2 h (Supplementary Note 6). Therefore, the synergy of Lewis acidic AlV and enhanced Brønsted acidity renders these ASAs promising bifunctional catalysts for Brønsted-Lewis acid-catalyzed reaction, such as the glucose dehydration to HMF.
It is noteworthy to point out the high stability of the BAS with enhanced strength when they are exposed to high temperature (1073 K) calcination (regeneration temperature in fluid catalytic cracking, Supplementary Fig. 12c) or to a liquid-phase conversion of glyceraldehyde in ethanol (Supplementary Fig. 12d). Two or more Al centers nearby SiOH groups in the amorphous silica network did not leach out from the catalysts after five recycle uses in a batch reaction with ultrasonic washing (Supplementary Fig. 12d), as confirmed by the lack of detectable Al species in the reaction mixture. Conversely, a similar treatment caused strong leaching of extra-framework aluminum species in zeolites, which exhibited enhanced acidity of BAS via the synergy of extra-framework aluminum and BAS. This leaching led to a significant activity loss of these zeolites (Supplementary Note 7) as confirmed by 1D 27Al NMR experiments (Supplementary Note 8).
In conclusion, a remarkable synergy between two Al centers (AlV-AlIV or AlV-AlV) close to the same SiOH group has been evidenced in flame-made amorphous silica-alumina (ASA) by 2D 27Al and 1H DQ-SQ NMR experiments, and analysis of the 3D spatial elemental distribution of Al and Si by APT. The study revealed that compared to the widely accepted model of one Al center per SiOH group with moderate strength (δ13C = 213 ppm)5, two proximate Al centers can strongly decrease the electron density from a neighboring silanol oxygen and thereby can significantly boost its acid strength (with δ13C = 227 ppm for CH313COCH3) to a value higher than that of H-ZSM-5 (δ13C = 223 ppm)55, or even reaching that of dealuminated zeolite HY (δ13C = 228 ppm)17. These BAS with zeolitic strength have been evidenced by comparative H/D exchange experiments with C6D6. Furthermore, the synergy between BAS with zeolitic strength and LAS afforded a much higher HMF yield (38%) than catalysts with moderate BAS strength (e.g. SA/10 and [Al]MCM-41). The achieved yield was comparable to that realized with metal-doped zeolites (33%) at a higher temperature. The present study highlights a promising route for generating BAS with zeolitic strength and high stability on ASAs, which could facilitate improved catalytic performances in a wide range of applications, including acid and multifunctional catalysis.
APT sample preparation and measurement method
A drop of the diluted dispersed ASA nanoparticles in methanol (≈0.01 mol/L) was placed onto a Si flat wafer, which was covered by a 150-nm thick protective Cr layer in Leica EM ACE600. Needle-shaped APT specimens were prepared from the Si flat sample by a site-specific lift-out procedure using a FEI G4 CX focused ion beam (FIB)/scanning electron microscope66. The APT experiments were conducted on a CAMECA LEAP 5000 XR instrument equipped with an ultraviolet laser with a spot size of 2 µm and a wavelength of 355 nm. The detection efficiency of this state-of-the-art microscope is ca. 54%. Data were acquired in laser pulsing mode at a specimen temperature of 50 K, with a target evaporation rate of 3 ions per 1000 pulses, a pulsing rate of 200 kHz, and a laser pulse energy of 50 pJ. The APT data were reconstructed and analyzed using the commercial IVAS 3.6.14™ software.
NMR experimental details
Before each experiment, the samples in glass tubes were dehydrated at 723 K for 12 h at a pressure lower than 10−2 bar. Subsequently, the samples were transferred into the MAS NMR rotors under dry N2 inside a glove box. These ammonia-loaded samples were prepared by dehydrated samples loaded with ammonia on a vacuum line, followed by evacuation at 393 K for 1 h to remove weakly physisorbed molecules.
All 1H and 27Al NMR spectra were recorded on a Bruker Avance III 800 MHz spectrometer equipped with 3.2 mm MAS rotors spinning at 20 kHz. For 1H DQ-SQ 2D experiments, the 1H DQ coherences were excited and reconverted by applying the symmetry-based R\(12_2^5\) scheme67, which reintroduces the 1H-1H dipolar interactions under MAS. The 1H radio frequency (rf) amplitudes for the π/2 pulse and R\(12_2^5\) scheme were equal to ν1 = 75 and 60 kHz, respectively. The length of the excitation recoupling scheme was equal to that of the reconversion and ranged from 250 to 300 μs, depending on the experiment. 1H DQ-SQ 2D spectra resulted from averaging 32 to 128 transients with recycle delay of 1 to 5 s, resulting in a total experimental time of 2 to 4 h. During 27Al DQ-SQ 2D experiments, selective central transition (CT) π/2 and π-pulses of 8 and 16 μs, that is, an rf amplitude of about 10 kHz, were applied. The 27Al two-spin DQ coherences were excited and reconverted by applying the BR\(2_2^1\) pulse sequence36, which reintroduces the 27Al−27Al dipolar interactions under MAS. The lengths of the excitation and reconversion periods were equal and ranged from 800 to 1200 μs, depending on the experiment. The rf amplitude applied during the BR\(2_2^1\) pulse sequence was 6.6 kHz, which corresponds to a nutation frequency of 20 kHz for the 27Al CT. Furthermore, the Hyper–Secant (HS) scheme was applied before the BR\(2_2^1\) excitation68, in order to enhance the 27Al CT polarization by saturating the satellite transitions69,70. HS employed a shaped pulse lasting 4 ms with an rf field amplitude of 16 kHz and a frequency sweep of 20 kHz around an offset of 200 kHz with respect to the CT. 27Al DQ-SQ 2D spectra resulted from averaging 14,400 and 3200 transients for SA/10 (Fig. 1) and SA/50 (Fig. 3) with recycle delay of 0.2 s, resulting in a total experimental time of 25.3 and 5.7 h, respectively. The 1H isotropic chemical shifts were referenced to tetramethylsilane using the resonance of adamantane (1.83 ppm) as a secondary reference, whereas the 27Al ones were referenced to 1 M solution Al(NO3)3.
In situ 1H MAS NMR Spectroscopy of H/D exchange with C6D6
1H MAS NMR spectra of H/D exchange with C6D6 was carried out on a Bruker Avance III 400 WB spectrometer at the Larmor frequency of 400.1 MHz with 4 mm MAS rotors spinning at 8 kHz. Spectra were recorded after single-pulse π/2 excitation with repetition times of 20 s and 8 scans. Prior to measurements, all samples were dehydrated at 723 K in vacuum (pressure <10−2 bar) for 12 h in glass tubes. The density of BAS on all dehydrated samples was determined by quantitative 1H MAS NMR experiments using NH3 as probe molecules. The total number of BAS was calculated based on the BAS density and weight of the sample. A known amount of dehydrated samples was transferred into the MAS rotors under dry nitrogen gas inside a glove box, sealed and utilized for in-situ loading on a vacuum line. The loading pressure of benzene-d6 (99.6%, Cambridge Isotope Laboratories, Inc.) was calculated and controlled according to the total number of BAS and known volume of the vacuum line to ensure one molecule of benzene-d6 per BAS. Then the sample was cooled down by liquid nitrogen till nearly no pressure could be detected. Subsequently, the loaded samples in the MAS rotors were kept 10 min at room temperature under dry nitrogen gas inside a glove box for better diffusion. The H/D experiments were performed by heating the MAS rotor at 313 K in a variable-temperature probe for 1H MAS NMR investigations. The concentration of protons bound to the aromatic rings was calculated as the ratio between the integrated intensity of the aromatic 1H signal and the number of Brønsted acid sites. The rate k of the H/D exchange between the deuterated molecules and the acidic OH groups (BAS) is described by an exponential relationship9
where I(t) and I(\(\infty\)) denote the intensities of the 1H MAS NMR signal of the aromatic rings at the observation time t and t → + \(\infty\) in the equilibrium state, respectively. The b parameter describes the exchange at t = 0, which corresponds to the start of the H/D exchange experiment, i.e. when the temperature was increased from ca. 293 K to that of the reaction.
Raw data are available from the authors upon reasonable request.
Corma, A. Inorganic solid acids and their use in acid-catalyzed hydrocarbon reactions. Chem. Rev. 95, 559–614 (1995).
Huber, G. W., Iborra, S. & Corma, A. Synthesis of transportation fuels from biomass: chemistry, catalysts, and engineering. Chem. Rev. 106, 4044–4098 (2006).
Wang, Z. et al. Acidity enhanced [Al]MCM-41 via ultrasonic irradiation for the Beckmann rearrangement of cyclohexanone oxime to ɛ-caprolactam. J. Catal. 358, 71–79 (2018).
Wang, Z., Wang, L., Jiang, Y., Hunger, M. & Huang, J. Cooperativity of Brønsted and Lewis acid sites on zeolite for glycerol dehydration. ACS Catal. 4, 1144–1147 (2014).
Huang, J., van Vegten, N., Jiang, Y., Hunger, M. & Baiker, A. Increasing the Brønsted acidity of flame-derived silica/alumina up to zeolitic strength. Angew. Chem. Int. Ed. 49, 7776–7781 (2010).
Jiang, Y., Huang, J., Dai, W. & Hunger, M. Solid-state nuclear magnetic resonance investigations of the nature, property, and activity of acid sites on solid catalysts. Solid State Nucl. Magn. Reson. 39, 116–141 (2011).
Fang, H. et al. New insights into the effects of acid strength on the solid acid-catalyzed reaction: theoretical calculation study of olefinic hydrocarbon protonation reaction. J. Phys. Chem. C. 114, 10254–10264 (2010).
Huang, J., Jiang, Y., Marthala, V. R. R., Ooi, Y. S. & Hunger, M. Regioselective H/D exchange at the side-chain of ethylbenzene on dealuminated zeolite H-Y studied by in situ MAS NMR-UV/Vis spectroscopy. ChemPhysChem 9, 1107–1109 (2008).
Huang, J. et al. In situ H-1 MAS NMR investigations of the H/D exchange of alkylaromatic hydrocarbons on zeolites H-Y, La,Na-Y, and H-ZSM-5. Microporous Mesoporous Mater. 99, 86–90 (2007).
Huang, J., Jiang, Y., Marthala, V. R. R. & Hunger, M. Insight into the mechanisms of the ethylbenzene disproportionation: Transition state shape selectivity on zeolites. J. Am. Chem. Soc. 130, 12642–12644 (2008).
Beck, L. W., Xu, T., Nicholas, J. B. & Haw, J. F. Kinetic NMR and density-functional study of benzene H/D exchange in zeolites, the simplest aromatic-substitution. J. Am. Chem. Soc. 117, 11594–11595 (1995).
Mildner, T. & Freude, D. Proton transfer between Brønsted sites and benzene molecules in zeolites H-Y studies by in situ MAS NMR. J. Catal. 178, 309–314 (1998).
Li, Y., Li, L. & Yu, J. H. Applications of Zeolites in Sustainable Chemistry. Chem 3, 928–949 (2017).
Coppola, G. M., Schuster, H. F. α-Hydroxy acids in enantioselective synthesis (WileyVCH Verteg GmbH, 1997).
Wang, Z., Jiang, Y., Baiker, A. & Huang, J. Efficient acid-catalyzed conversion of phenylglyoxal to mandelates on flame-derived silica/alumina. ACS Catal. 3, 1573–1577 (2013).
Wang, Z. C. et al. Brønsted acid sites based on penta-coordinated aluminum species. Nat. Commun. 7, 13820 (2016).
Li, S. et al. Brønsted/Lewis acid synergy in dealuminated HY zeolite: A combined solid-state NMR and theoretical calculation study. J. Am. Chem. Soc. 129, 11161–11171 (2007).
Weitkamp, J., and Hunger M. Acid and base catalysis on zeolites. In Studies in Surface Science and Catalysis, Ejka, J., van Bekkum, H., Corma, A., Schüth, F., ed. (Elsevier: Amsterdam, 2007), pp. 787-835.
Mouat, A. R. et al. Highly dispersed SiOx/Al2O3 catalysts illuminate the reactivity of isolated silanol sites. Angew. Chem. Int. Ed. 54, 13346–13351 (2015).
Mouat, A. R., Kobayashi, T., Pruski, M., Marks, T. J. & Stair, P. C. Direct spectroscopic evidence for isolated silanols in SiOx/Al2O3 and their formation mechanism. J. Phys. Chem. C. 121, 6060–6064 (2017).
Valla, M. et al. Atomic description of the interface between silica and alumina in aluminosilicates through dynamic nuclear polarization surface-enhanced NMR spectroscopy and first-principles calculations. J. Am. Chem. Soc. 137, 10710–10719 (2015).
Wang, Z. et al. One-step room-temperature synthesis of [Al]MCM-41 materials for the catalytic conversion of phenylglyoxal to ethylmandelate. ChemCatChem 5, 3889–3896 (2013).
Luo, Q. et al. Using trimethylphosphine as a probe molecule to study the acid states in Al-MCM-41 materials by solid-state NMR spectroscopy. J. Phys. Chem. B 107, 2435–2442 (2003).
Chizallet, C. & Raybaud, P. Pseudo-bridging silanols as versatile Brønsted acid sites of amorphous aluminosilicate surfaces. Angew. Chem. Int. Ed. 48, 2891–2893 (2009).
Mortier, W. J. Zeolite electronegativity related to physicochemical properties. J. Catal. 55, 138–145 (1978).
Kalirai, S., Paalanen, P. P., Wang, J., Meirer, F. & Weckhuysen, B. M. Visualizing dealumination of a single zeolite domain in a real-life catalytic cracking particle. Angew. Chem. Int. Ed. 55, 11134–11138 (2016).
Dapsens, P. Y., Mondelli, C. & Perez-Ramirez, J. Highly selective Lewis acid sites in desilicated MFI zeolites for dihydroxyacetone isomerization to lactic acid. ChemSusChem 6, 831–839 (2013).
Loewenstein, W. The distribution of aluminum in the tetrahedra of silicates and aluminates. Am. Min. 39, 92–96 (1954).
Hensen, E. J. M. et al. Acidity characterization of amorphous silica-alumina. J. Phys. Chem. C. 116, 21416–21429 (2012).
Eichler, U., Brandle, M. & Sauer, J. Predicting absolute and site specific acidities for zeolite catalysts by a combined quantum mechanics interatomic potential function approach. J. Phys. Chem. B 101, 10035–10050 (1997).
Chizallet, C. & Raybaud, P. Acidity of amorphous silica-alumina: from coordination promotion of Lewis sites to proton transfer. ChemPhysChem 11, 105–108 (2010).
Perras, F. A. et al. Shedding light on the atomic-scale structure of amorphous silica-alumina and its Brønsted acid sites. Phys. Chem. Chem. Phys. 21, 19529–19537 (2019).
Yu, Z. et al. Insights into the dealumination of zeolite HY revealed by sensitivity-enhanced Al-27 DQ-MAS NMR spectroscopy at high field. Angew. Chem. Int. Ed. 49, 8657–8661 (2010).
Lee, D. et al. Enhanced solid-state NMR correlation spectroscopy of quadrupolar nuclei using dynamic nuclear polarization. J. Am. Chem. Soc. 134, 18491–18494 (2012).
Li, S. H. et al. Extra-framework aluminium species in hydrated faujasite zeolite as investigated by two-dimensional solid-state NMR spectroscopy and theoretical calculations. Phys. Chem. Chem. Phys. 12, 3895–3903 (2010).
Wang, Q. et al. Double-quantum homonuclear NMR correlation spectroscopy of quadrupolar nuclei subjected to magic-angle spinning and high magnetic field. J. Magn. Reson. 200, 251–260 (2009).
Li, T. et al. Atomic imaging of carbon-supported Pt, Pt/Co, and Ir@Pt nanocatalysts by atom-probe tomography. ACS Catal. 4, 695–702 (2014).
Eley, C. et al. Nanojunction-mediated photocatalytic enhancement in heterostructured CdS/ZnO, CdSe/ZnO, and CdTe/ZnO nanocrystals. Angew. Chem. Int. Ed. 53, 7838–7842 (2014).
Schmidt, J. E., Peng, L. Q., Poplawsky, J. D. & Weckhuysen, B. M. Nanoscale chemical imaging of zeolites using atom probe tomography. Angew. Chem. Int. Ed. 57, 10422–10435 (2018).
Schmidt, J. E. et al. Isolating clusters of light elements in molecular sieves with atom probe tomography. J. Am. Chem. Soc. 140, 9154–9158 (2018).
Schmidt, J. E. et al. Coke formation in a zeolite crystal during the methanol-to-hydrocarbons reaction as studied with atom probe tomography. Angew. Chem. Int. Ed. 55, 11173–11177 (2016).
Schmidt, J. E., Oord, R., Guo, W., Poplawsky, J. D. & Weckhuysen, B. M. Nanoscale tomography reveals the deactivation of automotive copper-exchanged zeolite catalysts. Nat. Commun. 8, 1666 (2017).
Perea, D. E. et al. Determining the location and nearest neighbors of aluminium in zeolites with atom probe tomography. Nat. Commun. 6, 7589 (2015).
Xu, W. et al. A high-specific-strength and corrosion-resistant magnesium alloy. Nat. Mater. 14, 1229–1235 (2015).
Moutanabbir, O. et al. Colossal injection of catalyst atoms into silicon nanowires. Nature 496, 78–82 (2013).
Luo, Z. et al. Atomic gold–enabled three-dimensional lithography for silicon mesostructures. Science 348, 1451–1455 (2015).
Sun, Z. et al. Dopant diffusion and activation in silicon nanowires fabricated by ex situ doping: a correlative study via atom-probe tomography and scanning tunneling spectroscopy. Nano Lett. 16, 4490–4500 (2016).
Corma, A., Iborra, S. & Velty, A. Chemical routes for the transformation of biomass into chemicals. Chem. Rev. 107, 2411–2502 (2007).
Wang, Z. et al. High population and dispersion of pentacoordinated AlV species on the surface of flame-made amorphous silica-alumina. Sci. Bull. 64, 516–523 (2019).
Wang, Z. et al. Palladium-doped silica–alumina catalysts obtained from double-flame FSP for chemoselective hydrogenation of the model aromatic ketone acetophenone. J. Catal. 302, 10–19 (2013).
Lang, S., Benz, M., Obenaus, U., Himmelmann, R. & Hunger, M. Novel approach for the characterization of Lewis acidic solid catalysts by solid-state NMR spectroscopy. ChemCatChem 8, 2031–2036 (2016).
Wang, Z., Jiang, Y., Hunger, M., Baiker, A. & Huang, J. Catalytic performance of Brønsted and Lewis acid sites in phenylglyoxal conversion on flame-derived silica-zirconia. ChemCatChem 6, 2970–2975 (2014).
Yin, F., Blumenfeld, A. L., Gruver, V. & Fripiat, J. J. NH3 as a probe molecule for NMR and IR study of zeolite catalyst acidity. J. Phys. Chem. B 101, 1824–1830 (1997).
Ma, D. et al. An investigation of the roles of surface aluminum and acid sites in the zeolite MCM-22. Chem. Eur. J. 8, 162–170 (2002).
Zheng, A. M. et al. Relationship between H-1 chemical shifts of deuterated pyridinium ions and Brønsted acid strength of solid acids. J. Phys. Chem. B 111, 3085–3089 (2007).
De Geuser, F., Lefebvre, W. & Blavette, D. 3D atom probe study of solute atoms clustering during natural ageing and pre-ageing of an Al-Mg-Si alloy. Philos. Mag. Lett. 86, 227–234 (2006).
Haley, D., Petersen, T., Barton, G. & Ringer, S. P. Influence of field evaporation on Radial Distribution Functions in Atom Probe Tomography. Philos. Mag. 89, 925–943 (2009).
Stebbins, J. F. & Xu, Z. NMR evidence for excess non-bridging oxygen in an aluminosilicate glass. Nature 390, 60–62 (1997).
Pauling, L. in The nature of the chemical bond: An introduction to modern structural chemistry. 3 ed. (Cornell University Press: Ithaca, NY, 1960).
Leydier, F. et al. Bronsted acidity of amorphous silica-alumina: The molecular rules of proton transfer. J. Catal. 284, 215–229 (2011).
Larmier, K. et al. Isopropanol dehydration on amorphous silica–alumina: synergy of Brønsted and Lewis acidities at pseudo-bridging silanols. Angew. Chem. Int. Ed. 56, 230–234 (2017).
Leydier, F., Chizallet, C., Costa, D. & Raybaud, P. Revisiting carbenium chemistry on amorphous silica-alumina: Unraveling their milder acidity as compared to zeolites. J. Catal. 325, 35–47 (2015).
Ernst, H., Freude, D., Mildner, T., Pfeifer, H. High temperature 1H MAS NMR studies of the proton mobility in zeolites. In Proceeding of the 12th International Zeolite Conference, Treacy, M. M. J., Marcus, B. K., Bisher, M. E., Higgins, J. B. Eds. (Materials Research Society: Warrendale), pp. 2955 (1999).
Goncalves, V. L. C., Rodrigues, R. C., Lorencato, R. & Mota, C. J. A. Assessing the acid strength of solid acid catalysts with the use of linear free energy relationship: H/D exchange with substituted benzene derivatives. J. Catal. 248, 158–164 (2007).
Freude, D., Oehme, W., Schmiedel, H. & Staudte, B. NMR-studies concerning Brønsted acidity of zeolites. J. Catal. 49, 123–134 (1977).
Felfer, P. et al. New approaches to nanoparticle sample fabrication for atom probe tomography. Ultramicroscopy 159, 413–419 (2015).
Mafra, L. et al. High-resolution 1H homonuclear dipolar recoupling NMR spectra of biological solids at MAS rates up to 67 kHz. J. Magn. Reson. 199, 111–114 (2009).
Siegel, R., Nakashima, T. T. & Wasylishen, R. E. Signal enhancement of NMR spectra of half-integer quadrupolar nuclei in solids using hyperbolic secant pulses. Chem. Phys. Lett. 388, 441–445 (2004).
Madhu, P. K., Goldbourt, A., Frydman, L. & Vega, S. Sensitivity enhancement of the MQMAS NMR experiment by fast amplitude modulation of the pulses. Chem. Phys. Lett. 307, 41–47 (1999).
Yao, Z., Kwak, H. T., Sakellariou, D., Emsley, L. & Grandinetti, P. J. Sensitivity enhancement of the central transition NMR signal of quadrupolar nuclei under magic-angle spinning. Chem. Phys. Lett. 327, 85–90 (2000).
J.H. and Z.W. acknowledge the financial supports from Australian Research Council Discovery Projects (DP150103842, DP180104010, and DE190101618). J.H. thanks the University of Sydney SOAR fellowship, Sydney Nano Grand Challenge, and the International Project Development Funding. The authors acknowledge the kindly help from Prof. Michael Hunger at Stuttgart University for H/D exchange experiments during paper revision. O.L., J.L. and J.-P.A. are grateful for funding provided by the Region Nord/Pas de Calais (France), Europe (FEDER), CNRS, Ministère de l’Enseignement Supérieur et de la Recherche, CPER, Chevreul Institute (FR 2638), Infrastructure de Recherche en Résonance Magnétique Nucléaire à Très Haut Champ (IR-RMN, FR 3050), ENSCL, and contracts ANR-17-ERC2-0022 (EOS) and ANR-18-CE08-0015-01 (ThinGlass). This project has also received funding from the European Union’s Horizon 2020 research and innovation program under grant agreement n°731019 (EUSMI). T.L. thanks Deutsche Forschungsgemeinsschaft (DFG) for financial support (project number 407513992) and Zentrumfür Grenzflächendominierte Höchstleistungswerkstoffe (ZGH) at Ruhr University Bochum for the access of infrastructure (FEI Helios G4 CX FIB/SEM and Cameca LEAP 5000 XR).
The authors declare no competing interests.
Peer review information Nature Communications thanks Bert Weckhuysen, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Wang, Z., Li, T., Jiang, Y. et al. Acidity enhancement through synergy of penta- and tetra-coordinated aluminum species in amorphous silica networks. Nat Commun 11, 225 (2020). https://doi.org/10.1038/s41467-019-13907-7
This article is cited by
Nature Reviews Methods Primers (2021)
Direct synthesis of hierarchical zeolites through mesoporous 3D wormhole framework using single-head quaternary ammonium salt as structure-directing agent
Journal of Materials Science (2021)