Journal of Catalysis 394 (2021) 387–396

Contents lists available at ScienceDirect

Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat

Creating Brønsted acidity at the SiO2-Nb2O5 interface

Andrew T.Y. Wolek a, M. Alexander Ardagh a, Hien N. Pham c, Selim Alayoglu b,
Abhaya K. Datye c, Justin M. Notestein a,⇑
a
Department of Chemical and Biological Engineering, Northwestern University, 2145 Sheridan Rd., Evanston, IL 60208, United States
b
Center for Catalysis and Surface Science, Northwestern University, 2145 Sheridan Rd., Evanston, IL 60208, United States
c
Department of Chemical and Biological Engineering and Center for Micro-Engineered Materials, University of New Mexico, Albuquerque, NM 87131, United States

a r t i c l e i n f o a b s t r a c t

Article history: Catalytically active acid sites associated with the silica-niobia interface were probed with a series of over-
Received 7 July 2020 coated SiO2 on Nb2O5 (SiO2/Nb2O5) mixed oxide materials prepared by deposition of tetraethyl orthosil-
Revised 5 October 2020 icate onto niobic acid (Nb2O5
nH2O) or calcined niobia (Nb2O5). NH3 TPD and pyridine DRIFTS studies
Accepted 25 October 2020
indicated that the speciation of acid sites in the materials evolved as a function of SiO2 loading, impacting
Available online 9 November 2020
the quantity and stability of Brønsted sites. Catalyst activity was highly dependent on SiO2 loading in the
liquid phase hydroalkoxylation of dihydropyran with n-octanol. At SiO2 surface densities corresponding
Keywords:
to approximately 1 Si per 2 surface Nb, the activity of these catalysts passed through a maximum approx-
Overcoated materials
Hydroalkoxylation
imately 20 times higher than the activity of calcined Nb2O5. Apparent reaction barriers measured over the
Tetrahydropyranylation most active SiO2/Nb2O5 catalysts were 10 kJ/mol lower than those measured over niobic acid, suggesting
Mixed oxides that the OH features unique to the SiO2-Nb2O5 interface were slightly more reactive than those on niobic
Brønsted acids acid.
Niobic acid Ó 2020 Elsevier Inc. All rights reserved.

1. Introduction (e.g. niobia on silica) and NbOx/Al2O3 materials, with strong

Brønsted sites forming at sub-monolayer coverages of NbOx
The catalytic properties of niobium-based oxides have received [13,14]. A complicating factor in understanding NbOx/SiO2 materi-
increasing interest since pioneering studies by Tanabe et al. in the als is that NbOx domains grow and change structure (e.g. from iso-
19800 s reported the use of NbOx as a promoter, SMSI support, and lated sites to monolayers) with increasing surface density, while
acid catalyst [1]. In particular, hydrated niobium oxides, also the niobia-silica interface also changes structure. This can make
referred to as niobic acid (Nb2O5
nH2O), have been extensively it challenging to separate the importance of these two factors.
studied due to their strong Brønsted acid character, equivalent to The acidity of the reverse materials, consisting of SiO2 deposited
70% sulfuric acid, and catalytic stability in the presence of water on niobium oxides (SiO2/Nb2O5), is not currently known. The cat-
[2–4]. While these properties make niobic acid a promising cata- alytic dehydration and isomerization activity of comparable SiO2/
lyst for liquid phase reactions, the material loses acidity when ZrO2 and SiO2/TiO2 materials have been studied as a function of
heated to elevated temperatures such as might be needed for Si surface density [15], and we and others have examined Brønsted
vaporized reactants or to calcine the catalyst during a regeneration acidity at the SiO2/Al2O3 interface [16–18]. Beyond improved per-
step, limiting its commercial applications [1,5]. formance metrics, these materials have yielded additional under-
The inclusion of secondary oxide phases, such as WO3, ZrO2, standing of the new active sites that arise at the mixed oxide
Al2O3, and SiO2, is one strategy to improve the Brønsted acidic interfaces because the deposited SiO2 is, on its own, catalytically
properties of bulk niobium oxides [6–8]. Among these, niobia- inert, and because the bulk structure of the underlying metal oxide
silica oxides in particular have been successfully synthesized by a does not depend on loading, unlike their supported oxide counter-
wide variety of techniques. Mixed oxides have been synthesized by parts. In this work, we prepare a series of overcoated silica on nio-
routes including sol–gel, solution thermolysis, and co- bia (SiO2/Nb2O5) materials, tracking the evolution of surface
precipitation, and they have exhibited increased thermal stability species as a function of surface SiO2 density and measuring their
and decreased deactivation in the presence of water [9–12]. activity in the hydroalkoxylation of dihydropyran with n-octanol,
Brønsted acidity has also been identified in supported NbOx/SiO2 which is used here as a probe reaction that requires Brønsted
acidity.
⇑ Corresponding author.

https://doi.org/10.1016/j.jcat.2020.10.027
0021-9517/Ó 2020 Elsevier Inc. All rights reserved.
Andrew T.Y. Wolek, M.A. Ardagh, H.N. Pham et al. Journal of Catalysis 394 (2021) 387–396

2. Experimental methods Transmission electron microscopy (TEM) images were acquired

on a JEOL-2010F with an accelerating voltage of 200 kV. Samples
2.1. Synthesis were prepared by dispersing solids in ethanol then mounted on a
holey carbon TEM grid. Images were recorded in both bright field
SiO2 was grafted onto niobic acid (Nb2O5
nH2O, HY-340, (BF) and high angle annular dark field (HAADF) modes.
CBMM) by adapting a base-catalyzed methodology previously BET surface area and BJH pore volumes were calculated from N2
used by some of us to deposit thin films of SiO2 onto Al2O3 physisorption isotherms collected at 77 K on a Micromeritics
and TiO2 supports [16,19]. This synthesis method was suitable 3FLEX instrument in the REACT core facility at Northwestern
for multi-gram syntheses without specialized equipment, facili- University. All materials were degassed prior to treatment by hold-
tating subsequent characterization efforts. In summary, 2 g of ing at 250 °C under dynamic vacuum (<5 mm Hg) overnight. BET
niobic acid was taken as received and added to a solution of surface areas were estimated over a relative pressure range of
6.7:1 parts ethanol (200 proof; Fisher) and NH4OH (28% in 0.05 to the maximum of the Rouquerol plot [21], typically ~0.50
H2O, Fisher). The niobic acid was then suspended in the solution P/P0. BJH pore volumes were calculated from the desorption branch
by sonicating at room temperature for 30 min. Following this 70, of the isotherms. Pore size distributions were calculated using
140, or 260 lL of tetraethyl orthosilicate (TEOS, 99%, Aldrich), Non-Local Density Functional Theory (NLDFT) with a N2 cylindrical
corresponding to 0.6, 1.2, or 2.4 molecules / nm2 was injected pore model for oxides and a regularization value of 0.2000 [22,23].
into the mixture and allowed to react for 1 h while shaking on X-ray diffraction scattering patterns were acquired on a Rigaku
a gyratory plate. These conditions correspond to samples 0.6- Ultima PXRD in the Jerome B. Cohen X-Ray Diffraction Facility at
SiO2/Nb2O5, 1.2-SiO2/Nb2O5, 2.4-SiO2/Nb2O5, which are named Northwestern University. Samples were scanned with a Cu K-a
according to the surface loading of deposited metal normalized source from 5° to 90° 2-h in glass sample pans and the diffraction
relative to the surface area of niobic acid in units of atoms / patterns were identified using MDI JADE 9 fitting software (Mate-
nm2. The SiO2 overcoat in these samples is sub-monolayer as rials Data Inc., USA).
the Nb surface density on niobic acid is ~5 Nb atoms / nm2 [20]. Catalyst acid properties were studied with NH3 temperature pro-
For higher SiO2 loadings, the above procedure was carried out grammed desorption (TPD), propylamine temperature programmed
using 280 lL TEOS, corresponding to 2.5 molecules / nm2. After 1 reaction (TPRx), and diffuse reflectance infrared Fourier transform
hr shaking, the samples were re-sonicated for 30 min to re- spectroscopy (DRIFTS) using pyridine and NH3 as probe molecules
suspend, then an additional 120 lL TEOS was injected and allowed in the REACT core facility at Northwestern University. Typical NH3
to react for 1 h while shaking, to give sample 3.6-SiO2/Nb2O5. The TPD experiments involved pretreating the catalyst at 300 °C in 30
cycles of re-sonication and addition of 120 lL TEOS were repeated sccm of 10% O2 / 90% He for 30 min before the sample was saturated
3 and 14 additional times for samples 8.6-SiO2/Nb2O5 and 28.5- with 10% NH3 / 90% He at 100 °C. Following this, UHP He was flown at
SiO2/Nb2O5, respectively. A smaller amount of TEOS was injected 30 sccm to purge physisorbed NH3 from the system, and the samples
during these higher cycles to compensate for the lower OH density were heated to 600 °C at 10 °C / min under the same flow NH3 des-
expected on the SiO2 overcoat compared to the underlying niobic orption (m/z = 17) was monitored by inline MS.
acid. Note that we were unable to precisely determine the OH den- Propylamine TPRx (Hoffmann elimination reaction) was used to
sity of the niobic acid by thermogravimetric analysis due to its high selectively quantify Brønsted acid sites. In a typical experiment,
H2O content and thermal sensitivity. samples were pretreated at 300 °C in 30 sccm of UHP He for
After completing the deposition process, samples were washed 30 min before cooling to 100 °C for propylamine chemisorption.
twice with 200 mL ethanol and once with 200 mL hexanes (Fisher) Samples were saturated with propylamine (>99%, Alfa) by flowing
before being collected by vacuum filtration. Samples were allowed 10 sccm UHP He through a bubbler held at 3 °C, providing ~17%
to dry with vacuum applied overnight before being collected. Cal- propylamine flow. Physisorbed proyplamine was purged by flow-
cined niobic acid, referred hereafter as calcined Nb2O5, was pre- ing 30 sccm UHP He before samples were heated to 550 °C at
pared by calcining as-obtained Nb2O5
nH2O at 550 °C in static air 10 °C / min to catalytically decompose the propylamine to form
for 3 h with a 3 °C / min ramp rate. A set of SiO2 on calcined propylene and NH3. Desorption of propylamine (m/z = 30) and
Nb2O5 controls were prepared by following the SiO2 deposition reaction products were monitored by inline MS, with NH3 (m/
procedure outlined above, but starting with calcined Nb2O5. These z = 17) and propylene (m/z = 41) desorption peak areas found to
controls are named using the previously described nomenclature be equivalent within error. Propylamine decomposition occurs at
with an additional Nb2O5–c label. All materials were activated by ~400 °C regardless of Brønsted strength, producing a sharp propy-
calcining at 550 °C in static air for 3 h. A 3 °C / min ramp rate lene elution peak and limiting the possibility that propylamine
was used, as we have previously identified this parameter to influ- desorbs and reacts on a secondary Brønsted site [24]. We have used
ence SiO2 overcoat properties [16]. Samples were stored in a desic- the propylene signal for quantification since NH3 desorption was
cator and were reactivated by calcining at 300 °C overnight if not significantly broadened by the chromatographic effect.
used within 30 days. Pyridine DRIFTS experiments were performed on a Thermo
Nicolet 6700 FTIR spectrometer with a praying mantis diffuse
reflectance sample holder. Sample spectra as shown are the aver-
2.2. Characterization age of 64 scans each with a resolution of 0.5 cm1. Samples were
diluted ~10 wt% in KBr and pretreated at 300 °C under 40 sccm
SiO2 loadings for the series of catalysts were determined by 10% O2 / 90% Ar for 30 min. Following cooling to 150 °C, back-
inductively coupled plasma optical emission spectroscopy (ICP- ground spectra were obtained and 100 uL of pyridine (99.0%,
OES) conducted on a Thermo iCAP 7600 in the Quantitative Bioele- Sigma) was injected into a heated sample loop and introduced to
ment Imaging Center (QBIC) at Northwestern University. Samples the system with 40 sccm of UHP Ar carrier gas. After the pyridine
were digested in 48 wt% HF (Fisher), shaken at 200 rpm for 24 h, had saturated the samples, vacuum was pulled for 1 h and the final
then diluted with 1 wt% HNO3 (Fisher). [CAUTION: Store and han- spectra were acquired. Pyridine desorption was studied by heating
dle HF with extreme care]. Optical emission spectra were cali- samples to 200 °C or 250 °C for 30 min under vacuum before cool-
brated against a standards curve prepared by dilution of ing to 150 °C for spectra collection. The acquired diffuse reflectance
commercial Si standard (Sigma-Aldrich). Reported values are aver- spectra were processed by applying the Kubelka-Munk transfor-
ages of triplicates conducted for each sample. mation with OMNIC FTIR software (Thermo Scientific).
388
Andrew T.Y. Wolek, M.A. Ardagh, H.N. Pham et al. Journal of Catalysis 394 (2021) 387–396

2.3. Catalytic reactions Sample N2 isotherms and NLDFT pore size distributions are
located in Supporting Information Figs. S3 and S4, respectively. In
Liquid phase hydroalkoxylation of dihydropyran with n-octanol agreement with literature, calcining niobic acid at 550 °C results
was studied with batch reactions run in excess 3,4-dihydro-2H- in a ~85 m2 / g decrease in surface area [25]. The calcined, over-
pyran (DHP, molar ratio DHP:n-octanol = 2). Prior to reaction, pre- coated materials show a monotonic drop in the BJH pore volume
viously calcined catalysts were re-activated by drying at 300 °C with increasing SiO2 loading, but for most SiO2 loadings, they
overnight in static air. In a typical reaction, 15 mg catalyst was retain more surface area than does the uncoated catalyst. The pore
weighed hot into 20 mL glass vials to minimize readsorption of size distributions (Supporting Information Fig. S4) show that SiO2
moisture. 7 mL of preheated reaction solution (0.6 mmol 1- deposition partially stabilizes the 2–4 nm pores in the original nio-
octanol (99%, Sigma), 1.2 mmol DHP (97%; Aldrich), 0.6 mmol bic acid, which otherwise disappear when niobic acid is calcined.
mesitylene internal standard (98%, Sigma) in heptane (99%, X-ray diffraction patterns, shown in Fig. 2, confirm that the cal-
Sigma)) was then dosed and allowed to react on a heated shaker cination conditions are enough to crystallize uncoated niobic acid
plate at 500 rpm and 70 °C. Samples (50 lL) were taken periodi- into Nb2O5. With major peaks at 22.7°, 28.6°, and 36.7° 2-h, the
cally via syringe equipped with a Whatman 0.7 um glass microfiber crystal phases present in the calcined Nb2O5 sample were identi-
filter and analyzed on a Shimadzu GC 2010-plus equipped with a fied as 80 wt% TT-Nb2O5 (hexagonal, PDF# 00-028-0317) and
MS detector and a 30 m Zebron ZB-624 capillary column. Triplicate 20 wt% T-Nb2O5 (orthorhombic, PDF# 01-071-0336). We note the
reactions were run for each catalyst and initial reaction rates were mixed phase composition of calcined Nb2O5 is consistent with
calculated by quantifying the production of the tetrahydropyranyl the Nb2O5 phase progression put forth by other studies [25,26].
ether over the first 15 min of the reaction. Mass balances were As expected, XRD of calcined SiO2/Nb2O5 samples does not
>95% based on octanol and octyl-containing products. detect any crystalline SiO2 phases, since the calcination tempera-
tures used are significantly lower than those needed for the amor-
phous silica to cristobalite phase transformation [27]. However,
3. Results and discussion crystallization of the Nb2O5 core to TT-Nb2O5 and T-Nb2O5 phases
is observed in low SiO2 content samples 0.6-SiO2/Nb2O5 and 1.2-
3.1. SiO2 overcoating of Nb2O5
nH2O SiO2/Nb2O5. At higher SiO2 loadings, the major crystalline Nb2O5
reflections at 22.7° and 28.6° are still present but secondary to
Previously reported techniques were adapted to study the SiO2 / the amorphous character of the samples, indicating incomplete
bulk Nb2O5 interface which has not been investigated as thor- crystallization of the core oxide, which is consistent with the
oughly as its inverse, supported NbOx on bulk SiO2 [16,19]. A sum- TEM images of 3.6-SiO2/Nb2O5 in Fig. 1.
mary of physical characterization for the series of prepared SiO2/ Other reports have indicated that thin films deposited by
Nb2O5 materials and associated controls is shown in Table 1. Sam- atomic layer deposition have similar impacts on phase transitions
ple information for additional SiO2/Nb2O5 materials used primarily in Fe2O3 and Nb2O5 systems [28–30]. We hypothesize that limited
in catalyst testing is located in Supporting Information Table S1. Nb mobility at the Nb-O-Si interface is slowing Nb2O5 crystalliza-
Thermal treatment of niobic acid at temperatures above 500 °C tion [31], rather than increasing the phase transition temperature.
has been previously shown to result in extensive structural To investigate further, we calcined samples 2.4-SiO2/Nb2O5 and
changes, crystallization, and a near complete loss of Brønsted acid- 28.5 SiO2/Nb2O5 for an additional 24 h, and indeed, XRD patterns
ity [2,5,25]. In contrast, HAADF-STEM images of calcined 3.6-SiO2/ of both samples obtained after the additional 24 h hold show com-
Nb2O5 (Fig. 1b) show the material primarily retains the original plete crystallization to T-Nb2O5 and TT-Nb2O5 (Supporting Infor-
niobic acid morphology (Fig. 1a). While we did observe some local- mation Fig. S5). In summary, the SiO2 deposition process creates
ized crystallization in the sample, (Supporting Information Fig. S2), an overcoat of sub- to multi-layer amorphous SiO2. The presence
the majority of the sample remained amorphous, as shown by of this overcoat reduces the extent of the structural changes that
absence of a regular lattice in Fig. 1d. When comparing HRTEM niobic acid undergoes at temperatures above 500 °C and slows
images of niobic acid (Fig. 1c) and 3.6-SiO2/Nb2O5 (Fig. 1d), we the core oxide’s crystallization to T- and TT-Nb2O5.
see the SiO2 overcoat is close to 1 nm in thickness. SiO2 deposition
did not yield a perfect monolayer, which would be largely invisible 3.2. NH3 desorption and propylamine decomposition studies
at this scale, but neither did it yield large regions of bulk SiO2.
Deposition of much larger amounts of SiO2 (shown in Supporting After identifying the physical effects associated with SiO2 depo-
Information Fig. S2) results in a less homogeneous catalyst, where sition, we set out to quantify the surface acidity of the materials.
all niobic acid is encapsulated in SiO2, but some particles have sub- We first probed the strength of Brønsted sites via chemisorption
stantially more SiO2 deposition than others. The structural hetero- of Hammett indicators (Supporting Information Fig. S7). This
geneity of the high-loaded samples is not a cause for significant method has been extensively used to characterize the acidity of
concern, as these materials are poorly active for catalysis. niobic acid and Nb2O5 based materials [10,32–34]. These

Table 1
Physical properties of SiO2/Nb2O5 materials.

material SiO2 content (wt%)a calculated Si surface density BET surface area BJH pore volume included analysis
(Si atom/ nm2) (m2 / g)b (cc / g)
niobic acid n.d. – 151 0.16 c,d,e,f,h
calcined Nb2O5 n.d. – 67 0.16 c,d,e,f,h
0.6-SiO2/Nb2O5 0.9 0.6 78 0.16 c,h
1.2-SiO2/Nb2O5 1.9 1.2 81 0.14 c,h
2.4-SiO2/Nb2O5 3.6 2.4 78 0.13 d,g
3.6-SiO2/Nb2O5 5.4 3.6 85 0.13 c,d,e,f,h
8.6-SiO2/Nb2O5 13.0 8.6 55 0.07 e,f,h
28.5-SiO2/Nb2O5 43.0 28.5 27 0.04 c,d,e,f,h

a Accurate to ±10% relative; b Accurate to ±5 m2/g; c TEM; d XRD; e NH3 TPD and propylamine TPRx; f pyridine and NH3 DRIFTS; g Hammett indicators; h octanol

389
Andrew T.Y. Wolek, M.A. Ardagh, H.N. Pham et al. Journal of Catalysis 394 (2021) 387–396

Fig. 1. Representative HAADF-STEM images of a) niobic acid, b) calcined 3.6-SiO2/Nb2O5, and HRTEM images of c) niobic acid d) calcined 3.6-SiO2/Nb2O5. Visible in d), the
deposited SiO2 is ~1 nm in thickness. Additional TEM images are located in Supporting Information Fig. S2.

Table 2
NH3 and propylamine temperature programmed desorption.

NH3 uptake Propylamine reacted

material lmol/ molec./ lmol/ molec./
gNb2O5 nm2102 gNb2O5 nm2102
niobic acid 83 33 25 10.
calcined Nb2O5 25 22 2 1.8
0.6-SiO2/Nb2O5 50 39 9 6.9
3.6-SiO2/Nb2O5 41 31 3 2.1
8.6-SiO2/Nb2O5 22 28 2 2.2
28.5-SiO2/Nb2O5 3 10 1 2.2

experiments led us to next quantify surface acidity via gas phase

desorption of basic probe molecules. NH3 was chosen as an appro-
priate starting probe due to its small kinetic diameter (0.26 nm)
that allows it to probe most structurally hindered sites [16,35–
37]. Temperature programmed desorption of NH3 was conducted
for select SiO2/Nb2O5 samples to study the evolution of acid sites
as SiO2 loading was increased. Total acidity counts are listed in
Table 2 on a mass and surface area basis.
Fig. 2. X-Ray diffraction patterns of SiO2/Nb2O5 materials.
Although niobic acid has been reported to be a promising acid
catalyst at low to moderate temperatures, heating it to elevated
temperatures is known to eliminate both Lewis and Brønsted acid
experiments suggest that SiO2/Nb2O5, niobic acid, and calcined sites [1,5]. The loss of sites from moderate thermal treatment
Nb2O5 contain Brønsted sites of similar strength between (<300 °C) can be reversed by exposure to water vapor but becomes
2.4 < pKa (aq.) < 6.2. However, the limited resolution of these irreversible once niobic acid crystalizes to Nb2O5 [32]. We observe

390
Andrew T.Y. Wolek, M.A. Ardagh, H.N. Pham et al. Journal of Catalysis 394 (2021) 387–396

a decrease of ~70% (mass basis) or ~30% (surface area basis) in total Propylamine decomposition was used as an additional basic
NH3 uptake after calcining niobic acid to form crystalline T- and probe to quantify the number of Brønsted acid sites [24,35,42–
TT-Nb2O5, consistent with previous reports. Interestingly, the 44]. Propylamine adsorbs on both Lewis and Brønsted sites, but
SiO2/Nb2O5 materials adsorb more NH3 than calcined Nb2O5 at all it only decomposes to NH3 and propylene on Brønsted acid sites
but the highest SiO2 loadings, and at moderate SiO2 loadings, the and at ~400 °C [24], meaning that the method only counts strong
SiO2/Nb2O5 materials adsorb comparable or even slightly higher sites able to retain propylamine at these elevated temperatures
amounts of NH3 per unit surface area than niobic acid. As SiO2 [45]. In qualitative agreement with previous reports and our with
loadings are increased, the NH3 absorption generally decreases, NH3 desorption results, there is a ~80% (surface area basis)
leading to minimal detectable acidity over 28.5-SiO2/Nb2O5, as decrease in the number of Brønsted acid sites quantified by propy-
expected for a purely siliceous surface. lamine decomposition when niobic acid is calcined [1,5]. Relative
The shape of NH3 desorption profiles provides some insight into to calcined Nb2O5 we observe the number of Brønsted sites in
the speciation of acid sites, since NH3 desorbs from stronger acid the SiO2/Nb2O5 materials increase at low SiO2 loadings before con-
sites at higher temperatures. In Fig. 3a, the raw TPD data for each verging to comparable values. At the local maximum centered at
sample are fit to a series of three Gaussians centered at approxi- 0.6 Si / nm2, the Brønsted site density is 3.8 times (surface area
mately 200 °C, 260 °C, and 370 °C. We tentatively assign these fea- basis) greater than that of calcined Nb2O5.
tures to weakly acidic surface hydroxyls, medium Lewis sites, and a The thermal stability of these materials was tested by repeating
combination of strong Lewis + Brønsted sites, respectively [38–41]. the propylamine decomposition experiments for a total of three
Fig. 3b shows the relative contributions from each population as a cycles. Each cycle was treated as an individual experiment with
function of the Si surface density. Looking at the relative contribu- its own pretreatment at 300 °C, propylamine saturation, and TPRx
tions, it is apparent that the significant decrease in total NH3 ramp. Fig. 4 shows a progressive loss in acidity for both niobic acid
uptake in the calcined Nb2O5 sample is mainly due to the loss of and 0.6-SiO2/Nb2O5, indicating both materials are metastable. Raw
medium and strong acid sites. Interestingly, we observe an TPRx data and tabulated data for additional samples are located in
increased amount of both medium and strong acid sites relative Supporting Information Figs. S8 and S9, respectively. Niobic acid
to calcined Nb2O5 in the series of SiO2/Nb2O5 materials. The quan- saw the largest decrease after the first TPRx cycle, with a loss of
tity of these acid sites is maximized at sub-monolayer SiO2 load- 8.2  102 molec. / nm2, corresponding to ~82% the original
ings, with the number of strong acid sites in the 0.6-SiO2/Nb2O5 Brønsted sites. The loss of sites is due to the evolution of moisture
sample being roughly double that of calcined Nb2O5 (mass basis). from the sample, since we observed a moderate water signal dur-
As the SiO2 loading is further increased we observe a decrease in ing the first desorption cycle. The SiO2/Nb2O5 samples also showed
medium and strong acid sites until only contributions from weak progressively fewer Brønsted sites with sequential cycles, although
acid sites are detected over 28.5-SiO2/Nb2O5, again as we expect none exhibited as severe a decrease as niobic acid. Critically, the
for a siliceous surface. We note that we were only able to obtain 0.6-SiO2/Nb2O5 sample was more thermally stable than niobic acid,
a good fit for niobic acid when fitting the strong acid feature retaining a larger absolute number of Brønsted sites in the second
~60 °C higher (430 °C) than we did in the SiO2/Nb2O5 samples, sug- and third TPRx cycles.
gesting unique populations of acid sites on the two types of Together, the NH3 TPD and propylamine TPRx experiments
surfaces. show that the number and distribution of acid sites in the SiO2/

Fig. 3. (Left) Raw NH3 TPD profiles for select SiO2/Nb2O5 materials and controls and data fit overlays. Profiles were fit to m/z = 17 NH3 fragment signal. (Right) Relative
contributions of fit components as function of SiO2 loading (mass basis). The relative fit contributions are tabulated on both mass and surface area bases in Supporting
Information Table S6.

391
Andrew T.Y. Wolek, M.A. Ardagh, H.N. Pham et al. Journal of Catalysis 394 (2021) 387–396

DRIFTS spectra are discussed in Supporting Information Fig. S11.

Pyridine has strong gas-phase basicity and distinct Lewis and
Brønsted acid binding modes [46], and is discussed in depth here.
IR spectra were collected in diffuse reflectance mode since signifi-
cant absorbance of IR light by Nb2O5 precluded operation in trans-
mission mode. Fig. 5 shows the hydroxyl and ring stretching
regions of the pyridine difference spectra for niobic acid and select
SiO2/Nb2O5 materials. The full pyridine DRIFTS spectra are located
in Supporting Information Fig. S10.
In the OH region of the niobic acid spectra, we observe a nega-
tive feature at 3705 cm1 assigned to the loss of isolated Nb-OH
species following pyridine adsorption. This negative feature is no
longer present in the difference spectrum after the sample is cal-
cined to Nb2O5 at 550 °C [5,47]. The loss of this feature in the cal-
cined Nb2O5 spectra is again consistent with the evolution of water
from surface hydroxyls during thermal treatment. This feature is
also largely absent in the SiO2/Nb2O5 materials and is instead
replaced by a negative feature at 3745 cm1, whose absolute inten-
sity is maximized at a loading of 3.6 Si / nm2. The shift from 3705
to 3745 cm1 indicates that the surface OH species have been mod-
ified by the SiO2 deposition. We assign the feature at 3745 cm1 to
Fig. 4. Cyclic propylamine decomposition for SiO2/Nb2O5 materials.
pyridine adsorbing on Si-OH perturbed by proximity to the niobia
surface [16,48]. We separately note that materials with high SiO2
Nb2O5 materials is highly dependent on the SiO2 loading, with low loadings likely have high absolute numbers of Si-OH groups, but
loadings increasing the number of medium and strong acid sites that these Si-OH groups resemble those of bulk amorphous SiO2,
relative to calcined Nb2O5 and higher loadings leading to a which do not appreciably adsorb pyridine and thus do not appear
decrease in the number of these acid sites. Additionally, our cyclic in the difference spectrum.
propylamine TPRx experiments suggest that Brønsted acid sites in In the ring-stretching region of the spectra, we observe peaks at
the SiO2 overcoated materials have increased thermal stability 1445 cm1 and 1606 cm1 assigned to pyridine bound to Lewis
compared to those in niobic acid. sites [5,49]. Peaks at 1540 cm1 and 1639 cm1 are indicative of
protonated pyridinium ions on Brønsted sites. The band at
3.3. DRIFTS studies of surface acid sites 1488 cm1 has been assigned to a combination of Brønsted and
Lewis binding modes [49]. As SiO2 is deposited, the Brønsted peak
IR experiments with NH3 and pyridine probes were used to at 1540 cm1 grows in size, reaching a maximum at 3.6 Si / nm2
spectroscopically differentiate the types of acid sites present in before decreasing in intensity at higher loadings. In contrast, SiO2
samples. Both probes gave qualitatively similar results. The NH3 deposition decreases the Lewis binding features at 1445 cm1

Fig. 5. Pyridine DRIFTS difference spectra for SiO2/Nb2O5 materials at 150 °C. Hydroxyl bending region (left) and pyridine stretching and ring vibration region (right). *
indicates half scale.

392
Andrew T.Y. Wolek, M.A. Ardagh, H.N. Pham et al. Journal of Catalysis 394 (2021) 387–396

Scheme 1. Proposed pseudo-bridging silanol structure. The acid site is originally in a ‘‘open” configuration with distinct Si-OH and Nb Lewis sites that ‘‘closes” upon exposure
to pyridine.

and 1606 cm1 relative to niobic acid and calcined Nb2O5, suggest- in fine chemical and pharmaceutical synthesis [55,56]. Recent
ing they are either consumed or blocked by the deposition process. work has identified several Brønsted acid catalysts capable of
Finally, pyridine begins to desorb from Brønsted and Lewis sites on obtaining high yields for both intermolecular addition and
these materials at temperatures as low as 250 °C (Supporting Infor- intramolecular cyclization routes [57–59]. Molecular complexes
mation Fig. S12), indicating that pyridine is adsorbed on a range of of Ru, Pt, La, and Ce have also been proven to be active for the reac-
acid site strengths. tion. They are hypothesized to work by either directly activating
In aggregate, NH3 TPD, propylamine TPRx, and pyridine DRIFTS the alkene or promoting protonolysis of the reaction media [60–
spectra suggest that new acid sites are formed at the SiO2-Nb2O5 63].
interface as SiO2 is deposited. In niobic acid, Lewis sites are attrib- The hydroalkoxylation of alcohols with 3,4-dihydro-2H-pyran,
uted to undercoordinated NbOx species while Brønsted sites are also known as tetrahydropyranylation (Scheme 2), is used in mul-
attributed to Nb-OH groups [32]. In SiO2/Nb2O5, we attribute Lewis tistep syntheses to protect vulnerable alcohol functional groups.
character to the fraction of the same undercoordinated NbOx spe- The reaction has been extensively employed since its discovery
cies that remain uncovered by SiO2. In contrast, the shift of the in the late 1960s and is commonly catalyzed by solid acids, includ-
OH stretch from 3705 cm1 to 3745 cm1 indicates the formation ing silica-alumina, protic ion-exchange resins, and zeolites, among
of new Brønsted sites in SiO2/Nb2O5. other materials [55,56,64]. The reaction is easily carried out in par-
In amorphous silica alumina materials, recent DFT calculations allel batch reactions, which for this work makes it very useful for
have provided evidence that Brønsted acidity originates from screening catalyst materials with acid character. Furthermore,
pseudo-bridging silanols formed from neighboring Si-OH and Al the reaction has been shown elsewhere to be sensitive to acid
Lewis sites [50,51]. These structures exist in an ‘‘opened” state strength [65,66]. The reaction is first order with respect to acid
with distinct Si-OH and Al Lewis sites that ‘‘close” to form a Si-O concentration over relevant ranges (Supporting Information
(H)-Al bridge when exposed to a base. We hypothesize that analo- Fig. S13).
gous structures are the source of Brønsted acidity in SiO2/Nb2O5. We do not observe any conversion over Lewis acidic oxides such
Here, Si-OH species from deposited SiO2 domains may interact as Al2O3 and TiO2, suggesting that Brønsted acidity is a necessary –
with neighboring NbOx Lewis sites to ‘‘close” and form Brønsted but perhaps not sufficient – requirement for the reaction under the
acidic Si-O(H)-Nb bridges (shown in Scheme 1) [52]. We note that tested conditions. Cooperativity between Brønsted and Lewis sites
the negative 3745 cm1 Si-OH feature is maximized at the same may also possible, since their combination has been shown to
SiO2 loadings as the positive 1540 cm1 Brønsted feature, suggest- enhance reactivity in this reaction elsewhere [55,59,67]. We have
ing that these pseudo-bridging sites remain ‘‘open” until exposed plotted the initial reaction rate vs Si surface density in Fig. 6 to
to pyridine. Finally, we comment that a definitive assignment of highlight activity trends across the series of SiO2/Nb2O5 catalysts.
the acid site structure in these materials remains challenging given A tabulated summary of catalytic activities on surface area and
the continuing difficulty in characterizing amorphous mixed oxi- mass bases is located in Supporting Information Table S14. An
des [53,54]. alternate form of Fig. 6 plotted on a mass basis is found in Support-
These pseudo-bridging structures would be predicted to be ing Information Fig. S15.
most abundant at sub-monolayer SiO2 loadings where the SiO2 – Niobic acid, calcined Nb2O5, and SiO2 were tested as moderate
Nb2O5 interface is significant and where undercoordinated NbOx Brønsted acid catalysts and controls. Niobic acid showed moderate
still persist. As discussed in the previous section, the number of activity, reaching 45% conversion after 1 h, with an initial rate of
medium and strong acid sites as determined by NH3 TPD was max- 0.20 mmol / m2 – h. Initial rates measured over calcined Nb2O5
imized at moderate, sub-monolayer SiO2 loadings. Likewise, pyri- were an order of magnitude lower (0.02 mmol / m2 - h) as we
dine DRIFTS indicates that the SiO2/Nb2O5 materials retain Lewis expected based on the loss of 82% of its Brønsted sites (surface area
sites after overcoating and annealing at 550 °C while the Brønsted basis) during thermal treatment. Negligible activity was measured
sites corresponding to the peak at 1540 cm1 are maximized at a over SiO2. Since it appears neither the crystalline Nb2O5 surface nor
SiO2 loading of 3.6 Si / nm2 due to the interaction of a Si-OH with the SiO2 surface are highly active for this reaction, we attribute any
an adjacent, uncovered NbOx Lewis site. Finally, propylamine TPRx
shows a continuous decrease in the strongest Brønsted sites as SiO2
loading increases, suggesting that these sites may not be associated
with pseudo-bridging structures at the SiO2/Nb2O5 interface.

3.4. Hydroalkoxylation of dihydropyran to probe Brønsted acidity

Hydroalkoxylation, or the addition of alcohols across alkenes

and alkynes, is a synthetic route that has found wide application Scheme 2. Alcohol Tetrahydropyranylation.

393
Andrew T.Y. Wolek, M.A. Ardagh, H.N. Pham et al. Journal of Catalysis 394 (2021) 387–396

increases in activity in the SiO2/Nb2O5 samples to the newly

formed silica-niobia interface.
Deposition of SiO2 onto niobic acid resulted in a rapid increase
in activity from 0.02 mmol / m2 - h for calcined Nb2O5 to a maxi-
mum 0.51 mmol / m2 – h in 3.6-SiO2/Nb2O5. A similar trend is
observed in the mass-normalized rates, shown in Supporting Infor-
mation Fig. S15. Notably, this value is 25.5 times greater than the
0.02 mmol / m2 - h observed over calcined Nb2O5. It is also
~150% greater than the 0.20 mmol / m2 – h measured over the par-
ent niobic acid, even though that sample had a greater quantity of
total acid sites as measured by NH3 TPD and strong Brønsted sites
as measured by propylamine TPRx. We measured similarly high
reaction rates over a series of SiO2/Nb2O5-c controls (Supporting
Information Table S14) which were synthesized by SiO2 deposition
on pre-calcined Nb2O5. The significant enhancement in activity for
sub-monolayer loadings of SiO2, independent of the phase of Nb2O5
that SiO2 is deposited on, strongly suggests new active sites are
being formed at the silica-niobia interface. Then, as SiO2 loadings
approach and exceed ~5 Si / nm2, the activity steadily decreases,
consistent with the formation of an inactive silica shell at high
SiO2 loadings. This further suggests that active sites are found at
the SiO2-Nb2O5 interface and rules out the formation of an active Fig. 7. Proportionality between tetrahydropyranylation rates (as in Fig. 6) and a
proxy for the Brønsted acid site density from pyridine DRIFTS. The latter is defined
overlayer formed by diffusion of Nb into the SiO2, or vice versa, as the fractional area of the Bronsted acid feature at 1540 cm1 relative to the sum
during thermal treatments. of the Brønsted and Lewis features (1445 cm1). That fraction is multiplied by the
We determined activation energies over niobic acid and 3.6- total acidity as determined by NH3 TPD (Table 2). See Table S11 for values.
SiO2/Nb2O5 by fitting initial rates collected at 50–70 °C to an Arrhe-
nius model (Supporting Information Fig. S16). The apparent energy
barrier is slightly lower over 3.6-SiO2/Nb2O5 than over niobic acid and strong sites determined from NH3 TPD are also ineffective in
(68 kJ/mol vs. 78 kJ/mol ±5 kJ/mol), indicating that although there normalizing tetrahydropyranylation rates across the range of
may be fewer acid sites in the SiO2/Nb2O5 materials as measured SiO2 loadings. As shown in Fig. 7, tetrahydropyranylation rates do
by NH3 TPD and propylamine TPRx, those present are more active appear to be proportional to the Brønsted sites determined from
for the tetrahydropyranylation reaction. pyridine DRIFTS. Absent the ability to quantify absolute numbers
Finally, we calculated catalytic turnover frequencies (TOFs) of sites directly from pyridine DRIFTS, the number of these acid
based on a number of different approaches to defining an active sites is assumed to be proportional to the total acid sites from
site (Fig. S17). First, TOFs based on propylamine TPRx vary signifi- NH3 TPD multiplied by the relative fraction of Brønsted sites seen
cantly from material to material, arguing against the relevance of in the pyridine DRIFTS spectra. Critically, two trend-lines emerge,
this metric for determining catalytic active sites. The medium where the specific activity (the slope of the trendline) of the
SiO2/Nb2O5 materials is higher than that of niobic acid or calcined
Nb2O5, consistent with the lower apparent activation barrier for
the SiO2/Nb2O5 materials. The structure of the putative SiO2/
Nb2O5 Brønsted site (Scheme 1) appears to be more proficient at
adsorbing and activating dihydropyran than is an acid site on bare
Nb2O5.

4. Conclusions

In this study, a series of overcoated SiO2 on Nb2O5 materials

were synthesized to create and probe acidity at the silica-niobia
interface. The deposition of SiO2 impacts the total number and
the distribution of acid sites, which are increased relative to cal-
cined Nb2O5 at low SiO2 loadings and decrease as the surface
becomes siliceous. The deposition of SiO2 also generates new fea-
tures in the pyridine DRIFTS spectra, notably silanol groups at
3745 cm1 whose depletion is correlated with the formation of
protonated pyridine at 1540 cm1. These features are both most
prominent at sub-monolayer SiO2 loadings, suggesting that the
Brønsted sites formed at the silica-niobia interface are pseudo-
bridging silanols formed between surface Si-OH and residual
undercoordinated, Lewis acidic Nb atoms on the uncovered the
oxide surface.
The activity of SiO2/Nb2O5 in the hydroalkoxylation of 2,4-
dihydropyran with n-octanol is also maximized at intermediate
Fig. 6. Tetrahydropyranylation rates for SiO2/Nb2O5 materials. Conditions: 70 °C, SiO2 loadings, with a maximum activity of 0.51 mmol / m2 – h at
500 rpm, ~15 mg catalyst in 7 mL of preheated reaction solution (0.6 mmol 1-
octanol, 1.2 mmol DHP). Red line indicates activity of as-obtained niobic acid and
3.6 Si / nm2, or a Si / surface Nb ratio of ~0.5, that is significantly
dashed black line indicates activity of calcined Nb2O5. Plots of activity normalized higher than the activities we measured over niobic acid (0.20 mmol
to mass and different measures of surface acidity are located in S15 and S17. / m2 – h) and calcined Nb2O5 (0.02 mmol / m2 – h). Similar reaction
394
Andrew T.Y. Wolek, M.A. Ardagh, H.N. Pham et al. Journal of Catalysis 394 (2021) 387–396

rates over the series of SiO2/Nb2O5 –c controls provide strong evi- Appendix A. Supplementary material
dence that this enhanced activity is due to newly formed Brønsted
sites at the silica-niobia interface. Decreasing activity observed at Additional SiO2/Nb2O5 sample information, N2 physisorption
higher SiO2 loadings is likely due to the silica-niobia interface isotherms, NLDFT micropore distributions, comparison XRD pat-
becoming inaccessible. The hydroalkoxylation rates scale in pro- terns after secondary calcination, raw first cycle propylamine TPRx
portion to the Brønsted acidity revealed by pyridine DRIFTS, and data, Hammett Indicator experiments, cyclic propylamine TPRxn
they remain higher for SiO2/Nb2O5 than for niobic acid or Nb2O5 values, full pyridine DRIFTS spectra, full NH3 DRIFTS spectra, pyri-
even on this basis. Lastly, slightly lower apparent hydroalkoxyla- dine desorption DRIFTS spectra, preliminary tetrahydropyranyla-
tion activation energies over SiO2/Nb2O5 provide further evidence tion over SiO2/Al2O3 catalyst, summary of catalytic data, mass-
that the sites located at the silica-niobia interface are more active normalized Tetrahydropyranylation rates, Arrhenius plots for
than those found on niobic acid. SiO2/Nb2O5 and niobic acid, NH3 and propylamine TOF plots. Sup-
In this work we have extended our use of overcoated materials, plementary data to this article can be found online at https://doi.
previously used to study SiO2/Al2O3 and SiO2/TiO2, to demonstrate org/10.1016/j.jcat.2020.10.027.
the platform’s utility in studying the chemistry of SiO2/Nb2O5
materials. In this application, the highly tunable overcoating pro-
cess has provided additional understanding of the active sites that References
form at the SiO2 – Nb2O5 interface, showing that the SiO2 loading
can be controlled to influence the properties of the composite [1] K. Tanabe, Application of niobium oxides as catalysts, Catal. Today. 8 (1990) 1–
material. Future work is aimed toward applying this synthesis 11.
[2] T. Iizuka, K. Ogasawara, K. Tanabe, Acidic and catalytic properties of niobium
approach to study the acidity of additional SiO2-MOx interfaces, pentaoxide, Bull. Chem. Soc. Jpn. 56 (1983) 2927–2931.
expand the reaction scope to higher temperature gas phase reac- [3] K. Nakajima, J. Hirata, M. Kim, N.K. Gupta, T. Murayama, A. Yoshida, N. Hiyoshi,
tions, and to probe the structure of surface species with solid state A. Fukuoka, W. Ueda, Facile formation of lactic acid from a triose sugar in water
over niobium oxide with a deformed orthorhombic phase, ACS Catal. 8 (2018)
NMR. 283–290, https://doi.org/10.1021/acscatal.7b03003.
[4] K. Nakajima, Y. Baba, R. Noma, M. Kitano, J.K. Kondo, S. Hayashi, M. Hara,
Nb2O5*3nH2O as a heterogeneous catalyst with water-tolerant lewis acid
Funding sources sites, J. Am. Chem. Soc. 133 (2011) 4224–4227, https://doi.org/
10.1021/ja110482r.
This material is based upon work supported by the U.S. Depart- [5] G.S. Foo, D. Wei, D.S. Sholl, C. Sievers, Role of Lewis and Brønsted acid sites in
the dehydration of glycerol over niobia, ACS Catal. 4 (2014) 3180–3192,
ment of Energy, Office of Science, Office of Basic Energy Sciences https://doi.org/10.1021/cs5006376.
under Award Number DOE DE-FG02-03-ER154757 for support of [6] P. Lauriol-Garbey, G. Postole, S. Loridant, A. Auroux, V. Belliere-Baca, P. Rey, J.
M.A.A., S.A. and the National Science Foundation under Cooperative M. Millet, Acid – base properties of niobium-zirconium mixed oxide catalysts
for glycerol dehydration by calorimetric and catalytic investigation, Appl.
Agreement No. EEC-1647722 for support of A.T.Y.W. Catal. B Environ. 106 (2011) 94–102, https://doi.org/10.1016/j.
apcatb.2011.05.011.
[7] C. Tagusagawa, A. Takagaki, A. Iguchi, K. Takanabe, J.N. Kondo, K. Ebitani, S.
Declaration of Competing Interest Hayashi, T. Tatsumi, K. Domen, Highly active mesoporous Nb – W oxide solid-
acid catalyst, Angew. Chem. Int. Ed. 49 (2010) 1128–1132, https://doi.org/
The authors declare that they have no known competing finan- 10.1002/anie.200904791.
[8] M. Ziolek, Niobium-containing catalysts — the state of the art, Catal. Today. 78
cial interests or personal relationships that could have appeared (2003) 47–64.
to influence the work reported in this paper. [9] M.S.P. Francisco, R. Landers, Y. Gushikem, Local order structure and surface
acidity properties of a Nb2O5 / SiO2 mixed oxide prepared by the sol – gel
processing method, J. Solid State Chem. 177 (2004) 2432–2439, https://doi.
Acknowledgment org/10.1016/j.jssc.2004.03.035.
[10] C.G. Lugmair, T.D. Tilley, Single source molecular precursors to niobia – silica
and niobium phosphate materials, Monatshefte Fur Chemie. 137 (2006) 557–
A.T.Y.W. would like to thank Vincent Cheng for assistance with 566, https://doi.org/10.1007/s00706-006-0445-9.
running tetrahydropyranylation reactions. A.T.Y.W thanks Mihir [11] A. Aronne, E. Marenna, V. Califano, E. Fanelli, P. Pernice, M. Trifuoggi, A.
Bhagat for assistance with method development for GC-MS Vergara, Sol – gel synthesis and structural characterization of niobium-silicon
mixed-oxide nanocomposites, J. Sol-Gel Sci. Technol. 43 (2007) 193–204,
analyses.
https://doi.org/10.1007/s10971-007-1563-5.
This material is based upon work supported by the U.S. Depart- [12] P. Carniti, A. Gervasini, M. Marzo, Silica – niobia oxides as viable acid catalysts
ment of Energy, Office of Science, Office of Basic Energy Sciences in water: effective vs. intrinsic acidity, Catal. Today. 152 (2010) 42–47, https://
under Award Number DOE DE-FG02-03ER15457 for support of doi.org/10.1016/j.cattod.2009.07.111.
[13] J. Datka, A.M. Turek, J.M. Jehng, I.E. Wachs, Acidic properties of supported
M.A.A. and S.A. and the National Science Foundation under Cooper- niobium oxide catalysts: an infrared spectroscopy investigation, J. Catal. 135
ative Agreement No. EEC-1647722 for support of A.T.Y.W. (1992) 186–199, https://doi.org/10.1016/0021-9517(92)90279-Q.
Metal analysis was performed at the Northwestern University [14] P.A. Burke, E. Ko, Acidic properties of oxides containing niobia on silica and
niobia in silica, J. Catal. 46 (1991) 38–46.
Quantitative Bio-element Imaging Center. This work made use of [15] M. Niwa, N. Katada, Y. Murakami, Generation of acid sites by SiO2 deposition
the EPIC facility of Northwestern University’s NUANCE Center, on groups IVB metal oxides, J. Catal. 134 (1992) 340–348.
which has received support from the Soft and Hybrid Nanotechnol- [16] M.A. Ardagh, Z. Bo, S.L. Nauert, J.M. Notestein, Depositing SiO2 on Al2O3: a
route to tunable bronsted acid catalysts, ACS Catal. 6 (2016) 6156–6164,
ogy Experimental (SHyNE) Resource (NSF ECCS-1542205); the https://doi.org/10.1021/acscatal.6b01077.
MRSEC program (NSF DMR-1720139) at the Materials Research [17] A.R. Mouat, C. George, T. Kobayashi, M. Pruski, R.P. Van Duyne, T.J. Marks, P.C.
Center; the International Institute for Nanotechnology (IIN); the Stair, Highly dispersed SiO x / Al 2 O 3 catalysts illuminate the reactivity of
isolated silanol sites, Angew. Commun. 54 (2015) 13346–13351, https://doi.
Keck Foundation; and the State of Illinois, through the IIN. This org/10.1002/anie.201505452.
work made use of the Jerome B.Cohen X-Ray Diffraction Facility [18] A.R. Mouat, T. Kobayashi, M. Pruski, T.J. Marks, P.C. Stair, Direct spectroscopic
supported by the MRSEC program of the National Science Founda- evidence for isolated silanols in SiO x /Al 2 O 3 and their formation mechanism,
J. Phys. Chem. C. 121 (2017) 6060–6064, https://doi.org/10.1021/acs.
tion (DMR-1720139) at the Materials Research Center of North-
jpcc.6b11196.
western University and the Soft and Hybrid Nanotechnology [19] Z. Bo, T.R. Eaton, J.R. Gallagher, C.P. Canlas, J.T. Miller, J.M. Notestein, Size-
Experimental (SHYNE) Resource (NSF ECCS-1542205.) The REACT selective synthesis and stabilization of small silver nanoparticles on TiO2
Core facility acknowledges funding from the Department of Energy partially masked by SiO2, Chem. Mater. 27 (2015) 1269–1277, https://doi.org/
10.1021/cm504243f.
(DE-FG02-03ER15457) used for the purchase of the Altamira AMI- [20] Z. Yang, Y. Li, Q. Wu, N. Ren, Y. Zhang, Z. Liu, Y. Tang, Layered niobic acid with
200 and the Nicolet 6700 FT-IR. self-exfoliatable nanosheets and adjustable acidity for catalytic hydration of

395
Andrew T.Y. Wolek, M.A. Ardagh, H.N. Pham et al. Journal of Catalysis 394 (2021) 387–396

ethylene oxide, J. Catal. 280 (2011) 247–254, https://doi.org/10.1016/j. [45] K. Zhang, D.M. Zimmerman, A. Chung-phillips, C.J. Cassady, Experimental and
jcat.2011.03.026. ab initio studies of the gas-phase basicities of polyglycines, J. Am. Chem. Soc.
[21] J. Rouquerol, P. Llewellyn, F. Rouquerol, Is the BET equation applicable to 115 (1993) 10812–10822, https://doi.org/10.1021/ja00076a044.
microporous adsorbents ?, in: Stud. Surf. Sci. Catal., Elsevier B.V., 2007: pp. 49– [46] J. Briggs, R. Yamdagni, P. Kebarle, Intrinsic basicities of ammonia,
56. https://doi.org/10.1016/S0167-2991(07)80008-5. methylamines, anilines, and pyridine from gas-phase proton exchange
[22] G. Kupgan, T.P. Liyana-arachchi, C.M. Colina, NLDFT pore size distribution in equilibria, J. Am. Chem. Soc. 94 (1972) 5128–5130, https://doi.org/
amorphous microporous materials, Langmuir. 33 (2017) 11138–11145, 10.1021/ja00769a081.
https://doi.org/10.1021/acs.langmuir.7b01961. [47] H. Shima, M. Tanaka, H. Imai, T. Yokoi, T. Tatsumi, J.N. Kondo, IR observation of
[23] S.H. Madani, L.H. Diaz, M.J. Biggs, P. Pendleton, Uncertainty in pore size selective oxidation of cyclohexene with H2O2 over mesoporous Nb2O5, J.
distribution derived from adsorption isotherms: II. Adsorption integral Phys. Chem. C. 113 (2009) 21693–21699.
approach, Microporous Mesoporous Mater. 214 (2015) 217–223, https://doi. [48] L.T. Zhuravlev, The surface chemistry of amorphous silica. Zhuravlev model,
org/10.1016/j.micromeso.2015.04.030. Colloids Surf. Physicochem. Eng. Asp. 173 (2000) 1–38, https://doi.org/
[24] O. Kresnawahjuesa, R.J. Gorte, D. De Oliveira, L.Y. Lau, A simple, inexpensive, 10.1016/S0927-7757(00)00556-2.
and reliable method for measuring Bronsted- acid site densities in solid acids, [49] Z. Xue, Y. Zhang, G. Li, J. Wang, W. Zhao, T. Mu, Niobium phytate prepared from
Catal. Lett. 82 (2002) 155–160. phytic acid and NbCl5: a highly efficient and heterogeneous acid catalyst,
[25] V. Lebarbier, M. Houalla, T. Onfroy, New insights into the development of Catal. Sci. Technol. 6 (2016) 1070–1076, https://doi.org/10.1039/c5cy01123j.
Brønsted acidity of niobic acid, Catal. Today. 192 (2012) 123–129, https://doi. [50] C. Chizallet, P. Raybaud, Pseudo-bridging silanols as versatile brønsted acid
org/10.1016/j.cattod.2012.02.061. sites of amorphous aluminosilicate surfaces, Angew. Chemie - Int. Ed. 48
[26] D. Pradhan, A.W. Wren, S.T. Misture, N.P. Mellott, Investigating the structure (2009) 2891–2893, https://doi.org/10.1002/anie.200804580.
and biocompatibility of niobium and titanium oxides as coatings for [51] F. Leydier, C. Chizallet, A. Chaumonnot, M. Digne, E. Soyer, A.A. Quoineaud, D.
orthopedic metallic implants, Mater. Sci. Eng. C. 58 (2016) 918–926, https:// Costa, P. Raybaud, Brønsted acidity of amorphous silica-alumina: the
doi.org/10.1016/j.msec.2015.09.059. molecular rules of proton transfer, J. Catal. 284 (2011) 215–229, https://doi.
[27] O. Yamaguchi, T. Kanazawa, K. Shimizu, Crystallization of amorphous silica org/10.1016/j.jcat.2011.08.015.
into quartz, J. Chem. Soc. Dalt. Trans. (1982) 1005–1007, https://doi.org/ [52] B. Chakraborty, B. Viswanathan, Surface acidity of MCM-41 by in situ IR
10.1017/CBO9781107415324.004. studies of pyridine adsorption, Catal. Today. 49 (1999) 253–260.
[28] K. Kraffert, M. Karg, R. Schmack, G. Clavel, C. Boissiere, T. Wirth, N. Pinna, R. [53] M. Valla, A.J. Rossini, M. Caillot, P. Raybaud, M. Digne, A. Chaumonnot, A.
Kraehnert, Stabilization of mesoporous iron oxide films against sintering and Lesage, L. Emsley, J.A. Van Bokhoven, Atomic Description of the Interface
phase transformations via atomic layer deposition of alumina and silica, Adv. between Silica and Alumina in Aluminosilicates through Dynamic Nuclear
Mater. Interfaces. 5 (2018) 1800360, https://doi.org/10.1002/admi.201800360. Polarization Surface- Enhanced NMR Spectroscopy and First-Principles
[29] J.N. Kondo, K. Domen, Crystallization of mesoporous metal oxides, Chem. Calculations, (2015). https://doi.org/10.1021/jacs.5b06134.
Mater. 20 (2008) 835–847. [54] F.A. Perras, Z. Wang, T. Kobayashi, A. Baiker, J. Huang, M. Pruski, Shedding light
[30] J.M.R. Gallo, D. Wang, H.N. Pham, J.A. Libera, C.L. Marshall, W. Elam, A.K. Datye, on the atomic-scale structure of amorphous silica – alumina and its Brønsted
J.A. Dumesic, Synthesis of highly ordered hydrothermally stable mesoporous acid sites, PCCP. 21 (2019) 30–33, https://doi.org/10.1039/c9cp04099d.
niobia catalysts by atomic layer deposition, ACS Catal. 1 (2011) 1234–1245, [55] I. Rodriguez, M.J. Climent, S. Iborra, V. Fornes, A. Corma, Use of delaminated
https://doi.org/10.1021/cs200367t. zeolites (ITQ-2) and Mesoporous molecular sieves in the production of fine
[31] J.G. Weissman, E.I. Ko, P. Wynblatt, Study of the morphology and structure of chemicals: preparation of dimethylacetals and tetrahydropyranylation of
niobia-silica surface oxides using model thin films, J. Catal. 108 (1987) 383– alcohols and phenols, J. Catal. 192 (2000) 441–447.
400, https://doi.org/10.1016/0021-9517(87)90187-4. [56] N. Miyashita, A. Yoshikoshi, P.A. Grieco, Pyridinium p-toluenesulfonate. A mild
[32] K. Tanabe, Niobic acid as an unusual acidic solid material, Mater. Chem. Phys. 7 and efficient catalyst for the tetrahydropyranylation of alcohols, J. Org. Chem.
(1987) 217–225. 42 (1977) 3772–3774, https://doi.org/10.1021/jo00443a038.
[33] J. He, Q.J. Li, Y.N. Fan, Dispersion states and acid properties of SiO2-supported [57] M. Shibuya, S. Fujita, M. Abe, Y. Yamamoto, Brønsted acid/silane catalytic
Nb 2O5, J. Solid State Chem. 202 (2013) 121–127, https://doi.org/10.1016/j. system for intramolecular hydroalkoxylation and hydroamination of
jssc.2013.03.040. unactivated alkynes, ACS Catal. 7 (2017) 2848–2852, https://doi.org/
[34] S.-H. Chai, H.-P. Wang, Y. Liang, B.-Q. Xu, Sustainable production of acrolein: 10.1021/acscatal.7b00403.
gas-phase dehydration of glycerol over Nb2O5 catalyst, J. Catal. 250 (2007) [58] N. Tsuji, J.L. Kennemur, T. Buyck, S. Lee, S. Prévost, P.S.J. Kaib, D. Bykov, C. Farès,
342–349, https://doi.org/10.1016/j.jcat.2007.06.016. B. List, Activation of olefins via asymmetric Brønsted acid catalysis, Science
[35] S.A. Bates, W.N. Delgass, F.H. Ribeiro, J.T. Miller, R. Gounder, Methods for (80-) 359 (2018) 1501–1505.
NH3titration of Brønsted acid sites in Cu-zeolites that catalyze the selective [59] G. Sartori, R. Ballini, F. Bigi, G. Bosica, R. Maggi, P. Righi, Protection (and
catalytic reduction of NOx with NH3, J. Catal. 312 (2014) 26–36, https://doi. deprotection) of functional groups in organic synthesis by heterogeneous
org/10.1016/j.jcat.2013.12.020. catalysis, Chem. Rev. 104 (2004) 199–250, https://doi.org/10.1021/cr0200769.
[36] M. Hiyoshi, B. Lee, D. Lu, M. Hara, J.N. Kondo, K. Domen, Supermicroporous [60] H. Qian, X. Han, R.A. Widenhoefer, Platinum-catalyzed intramolecular
niobium oxide as an acid catalyst, Catal. Letters. 98 (2004) 1–6. hydroalkoxylation of c - and d -hydroxy olefins to form cyclic ethers, JACS
[37] H.N. Pham, Y.J. Pagan-torres, J.C. Serrano-ruiz, D. Wang, J.A. Dumesic, A.K. Commun. 126 (2004) 9536–9537, https://doi.org/10.1021/ja0477773.
Datye, Improved hydrothermal stability of niobia-supported Pd catalysts, [61] X. Yu, S. Seo, T.J. Marks, Effective, selective hydroalkoxylation / cyclization of
Appl. Catal. A Gen. 397 (2011) 153–162, https://doi.org/10.1016/j. alkynyl and allenyl alcohols mediated by lanthanide catalysts, JACS Commun.
apcata.2011.02.026. 129 (2007) 7244–7245, https://doi.org/10.1021/ja071707p.
[38] C.P. Kumar, K.R. Reddy, V.V. Rao, Vapour phase ammoxidation of toluene over [62] Y. Oe, T. Ohta, Y. Ito, Ruthenium-catalyzed addition reaction of alcohols across
vanadium oxide supported on Nb2O5-TiO2, Green Chem. 4 (2002) 513–516, olefins, Synlett. 1 (2005) 179–181, https://doi.org/10.1055/s-2004-836058.
https://doi.org/10.1039/b206581a. [63] E. Marotta, E. Foresti, T. Marcelli, F. Peri, P. Righi, N. Scardovi, G. Rosini, CeCl3,
[39] M. Anilkumar, W.F. Hoelderich, New non-zeolitic Nb-based catalysts for the 7H2O  NaI catalyzed hydrooxacyclization of unsaturated 3-hydroxy esters,
gas-phase Beckmann rearrangement of cyclohexanone oxime to caprolactam, Org. Lett. 4 (2002) 4451–4453, https://doi.org/10.1021/ol026955z.
J. Catal. 293 (2012) 76–84, https://doi.org/10.1016/j.jcat.2012.06.007. [64] T. De Baerdemaeker, W. Vandebroeck, H. Gies, B. Yilmaz, U. Müller, M. Feyen,
[40] Z. Zhang, P. Wang, Z. Wu, C. Yue, X. Wei, J. Zheng, M. Xiang, B. Liu, Efficient D. De Vos, Shape-selective organic – inorganic zeolitic catalysts prepared via
synthesis of niobium pentoxide nanowires and application in ethanolysis of interlayer expansion, Catal. Today. 235 (2014) 169–175, https://doi.org/
furfuryl alcohol y, RSC Adv. 10 (2020) 5690–5696, https://doi.org/10.1039/ 10.1016/j.cattod.2014.02.035.
d0ra00085j. [65] G.P. Romanelli, J.C. Autino, M.N. Blanco, L.R. Pizzio, Tungstosilicate salts as
[41] Y. Duan, J. Zhang, D. Li, D. Deng, L. Ma, Y. Yang, Direct conversion of catalysts in phenol tetrahydropyranylation and depyranylation, Appl. Catal. A
carbohydrates to diol by the combination of niobic acid and a hydrophobic Gen. 295 (2005) 209–215, https://doi.org/10.1016/j.apcata.2005.08.019.
ruthenium catalyst y, RSC Adv. 7 (2017) 26487–26493, https://doi.org/ [66] M.V. Shamzhy, C. Ochoa-Hernández, V.I. Kasneryk, M.V. Opanasenko, M.
10.1039/c7ra03939e. Mazur, Direct incorporation of B, Al, and Ga into medium-pore ITH zeolite:
[42] J.G. Tittensor, R.J. Gorte, D.M. Chapman, Isopropylamine adsorption for the synthesis, acidic, and catalytic properties, Catal. Today. 277 (2016) 37–47,
characterization of acid sites in silica-alumina catalysts, J. Catal. 720 (1992) https://doi.org/10.1016/j.cattod.2015.10.013.
714–720. [67] N. Azizi, M. Abdoli-Senejani, F. Abbasi, An efficient Brønsted-Lewis acidic ionic
[43] W.E. Farneth, R.J. Gorte, Methods for characterizing zeolite acidity, Chem. Rev. liquid catalyzed tetrahydropyranylation of alcohols, Tetrahedron Lett. 57
95 (1995) 615–635, https://doi.org/10.1021/cr00035a007. (2016) 5009–5011, https://doi.org/10.1016/j.tetlet.2016.09.099.
[44] D.J. Parrillo, A.T. Adamo, G.T. Kokotailo, R.J. Gorte, Amine adsorption in H-ZSM-
5, Appl. Catal. 67 (1990) 107–118.

396