Microporous and Mesoporous Materials 335 (2022) 111821

Contents lists available at ScienceDirect

Microporous and Mesoporous Materials

journal homepage: www.elsevier.com/locate/micromeso

Synthesis of high-surface area tungstated zirconia by atomic layer

deposition on mesoporous silica
Ching-Yu Wang , Ohhun Kwon , Raymond J. Gorte , John M. Vohs *
Department of Chemical and Biomolecular Engineering, University of Pennsylvania, Philadelphia, PA, 19104, USA

A R T I C L E I N F O A B S T R A C T

Keywords: Atomic layer deposition (ALD) was used to prepare ZrO2 films on the surface of the mesoporous silica, SBA-15,
Atomic layer deposition (ALD) and to modify the surface of these films with WO3 in order to form tungstated zirconia. Adsorption-desorption
Mesoporous material isotherms, pore size distributions, and transmission electron microscopy demonstrated that the ALD synthesis
Solid acid catalysis
produced zirconia films that were conformal to the SBA-15 pores. DRIFT spectroscopy of pyridine adsorbed on
Brønsted acid site
the tungstated-zirconia SBA-15 samples showed adsorbed pyridinium ions, confirming the presence of Brønsted-
acid sites on this material, consistent with what has been reported for bulk tungstated zirconia. The ALD-
synthesized, tungstated-zirconia SBA-15 was also shown to be active in the acid-catalyzed H-D exchange be­
tween toluene and D2O.

1. Introduction m2/g [6].

It should be noted that SBA-15 has previously been studied as a
Tungstated zirconia has properties characteristic of a solid Brønsted support for tungstated zirconia with the active phase being produced by
acid and is active for the isomerization of small alkanes, such as butane wet infiltration of zirconium acetylacetonate and ammonium meta­
and n-propane, at relatively low temperatures [1–3]. While the structure tungstate [7]. While this synthesis method produces an active catalyst, it
of the active form of the catalyst is still an open question, it is clear that provides less control over the structure of the WO3/ZrO2 deposits, and
intimate contact between the tungsta and zirconia is required for the pore blockage will likely occur at relatively low weight loadings [8]. In
formation of the acid sites. It is also known that each Brønsted site contrast, ALD growth of WO3/ZrO2 in SBA-15 has the potential to pro­
consists of multiple W atoms in the form of a (WOx)n cluster, with several duce a much more uniform material while preserving the linear pore
papers suggesting a minimum of four W per site [4,5]. structure of the support. Indeed, previous studies have shown that ALD
An active tungstated zirconia catalyst can be produced using atomic can be an effective means to grow conformal metal oxide films in the
layer deposition (ALD) to grow ultrathin WO3 films on a zirconia support pores of mesoporous oxides, including SBA-15 [9–17]. For example,
as demonstrated by Wang et al. [5]. This work demonstrated that both Zaera and coworkers used ALD to grow TiO2 and Al2O3 films on SBA-15
Brønsted-acid sites and redox sites were generated following deposition [15,16], while Pagan-Torres et al. used this approach to grow Nb2O3
of WO3 and that the density of each type of site could be controlled by films on this support [11]. The Zaera group did extensive character­
varying the number of WO3 ALD cycles. The maximum Brønsted-site ization of the growth process and the structure of the resulting films and
concentration was achieved after 5 ALD cycles. The ability to prepare demonstrated that the ALD oxide films were conformal to the SBA-15,
tungstated zirconia with active Brønsted-acid sites using ALD presents with the pore diameter decreasing systematically with the number of
the exciting possibility producing high-surface-area, solid-acid catalysts ALD cycles.
in well-defined pore structures that would be difficult to produce by Here, we have explored the use of ALD to grow tungstated zirconia
more conventional methods. This has motivated the present study in films in the pores of SBA-15. The deposition rate per ALD cycle for each
which we have explored the use of ALD to synthesize tungstated zirconia oxide was measured and the structure of the resulting films was char­
on a mesoporous silica, SBA-15. SBA-15 features uniform acterized using a combination of temperature programmed desorption-
one-dimensional pores with a diameter up to 10 nm, depending on thermogravimetric analysis (TPD-TGA) of 2-propanol, transmission
synthesis conditions, and typically has a surface area greater than 600 electron microscopy (TEM), and N2 adsorption isotherms. The formation

* Corresponding author.
E-mail address: [email protected] (J.M. Vohs).

https://doi.org/10.1016/j.micromeso.2022.111821
Received 28 February 2022; Accepted 8 March 2022
Available online 10 March 2022
1387-1811/© 2022 Elsevier Inc. All rights reserved.
C.-Y. Wang et al. Microporous and Mesoporous Materials 335 (2022) 111821

of active Brønsted acid sites in the ALD-synthesized WO3/ZrO2/SBA-15 electron microscopy (TEM), scanning transmission electron microscopy
materials was confirmed using diffuse reflectance infrared Fourier (STEM), and energy dispersive X-ray spectra (EDS) were performed with
transform (DRIFT) spectroscopy of adsorbed pyridine and the rate of H- a JEOL JEM-F200 STEM operated at 200 kV. For these measurements,
D exchange between D2O and toluene. The thermal stability of the the samples were diluted in ethanol and then deposited onto carbon
mesoporous structure was also shown to improve after coating the support films on copper grids (Electron Microscopy Sciences).
structure with ZrO2. TPD-TGA was used to characterize both the structure and reactivity
of the various samples. The TPD-TGA system consisted of a vacuum
2. Experimental section microbalance (CAHN 2000) equipped with a quadrupole mass spec­
trometer (Stanford Research Systems, RGA-100) and allowed the mass
SBA-15 was synthesized following a process reported in the literature change and desorbing products to be measured simultaneously. In a
[6,11]. Initially, 4.0 g of Pluronic P-123 (Sigma-Aldrich) was dissolved TPD-TGA measurement, 20 mg of sample was loaded into the micro­
in 120 g of 2 M HCl solution and 30 g deionized water with stirring at balance and then pretreated by heating to 823 K in a vacuum of ~10− 8
298 K for 20h. Tetraethoxysilane (TEOS, Sigma-Aldrich, 99%) was then Torr, then cooling to 298 K. After this, the sample was exposed to 2-
added dropwise with stirring at 308 K until 8.5 g had been added, at propanol (Fisher Scientific, 99.9%) vapor, followed by evacuation at
which point the solution was stirred at this temperature for 20 h. The room temperature. The sample was then heated from 298 to 823 K at 10
solution was then maintained at 373 K for 24 h to allow hydrolysis of the K/min while measuring the mass and the desorbing species.
TEOS. The resulting SBA-15 precipitate was filtered, washed with DRIFT spectroscopy was performed using an FTIR (Mattson, Galaxy)
deionized water, and dried at 353 K for 8 h in air. To remove the P-123 equipped with a diffuse-reflectance cell (Pike Technologies, DiffusIR). In
surfactant, the sample was calcined in flowing air at 773 K for 6 h with each DRIFT measurement, the sample was pretreated at 523 K with
an initial ramping rate of 1 K/min. As shown in Fig. S1a, the low-angle flowing He for 30 min before cooling to 423 K. The sample was exposed
X-ray diffraction (XRD) pattern of the as-synthesized SBA-15 contains to pyridine (Fisher Chemical, 99.9%) vapor with He as the carrier gas for
the 100, 110, and 200 peaks that are characteristic of this material [6]. 10 min and then flushed with He for 10 min. All DRIFT spectra were
ALD of ZrO2 and WO3 onto the SBA-15 was performed using a collected with the sample at 423 K.
homebuilt ALD system that has been described in detail previously H-D exchange between toluene (Fisher Chemical, 99.9%) and
[18–20]. Tetrakis(2,2,6,6-tetramethyl-3,5-heptanedionato)zirconium deuterium oxide (D2O, Cambridge Isotope Laboratories, 99.9%) was
(Zr(TMHD)4, Strem Chemicals, 99%) was used as the Zr precursor and used to confirm the formation of active Brønsted acid sites in the 15-
tungsten carbonyl (W(CO)6, Strem Chemicals, 99%) was used as the W ZrO2/SBA-15 and 5-WO3-15-ZrO2/SBA-15 samples, as described previ­
precursor. For each ZrO2 ALD cycle, the sample was evacuated, then ously [25]. Steady-state rate measurements were made using a micro­
exposed to Zr(TMHD)4 vapor at 523 K for 10 min, followed by oxidation flow flow reactor packed with 30 mg of catalyst that was initially
at 773 K in flowing air for 40 min. Note that this high-temperature pretreated at 523 K in He for 30 min. The reaction was carried out by
oxidation step was required to fully oxidize the adsorbed Zr(TMHD)4 co-feeding 1% toluene and 1% D2O in 20 mL/min of He at various
precursor. While it is common to use ozone or an O2 plasma for this temperatures. The products were analyzed by an online GC-MS (Shi­
purpose, these oxidants are ineffective on highly porous substrates madzu, GC-17A).
where rapid quenching of radicals occurs on the external surfaces
[21–23]. Throughout the paper, we denote ZrO2-modified SBA-15 3. Results and discussion
samples as XX-ZrO2/SBA-15, where XX corresponds to the number of
ZrO2 ALD cycles. 3.1. ALD of ZrO2 and WO3 on SBA-15
In a WO3 ALD cycle, the evacuated sample was exposed to W(CO)6
vapor at 473 K for 3 min, followed by evacuation and then oxidation at The amount of ZrO2 and WO3 deposited on the SBA-15 support was
the same temperature in air for 6 min. We use YY-WO3-XX-ZrO2/SBA-15 determined gravimetrically. Fig. 1a shows a plot of the mass of ZrO2
to denote samples with YY ALD cycles of WO3 on the XX-ZrO2/SBA-15. deposited, normalized to the mass of the SBA-15, versus the number of
The ALD growth rates of ZrO2 and WO3 were determined gravimetri­ ALD cycles. Note that this ratio increased nearly linearly with the
cally. The low-angle XRD pattern of a 15-ZrO2/SBA-15 sample displayed number of ALD cycles. The ZrO2 deposition rate was 0.06 g of ZrO2 per g
in Fig. S1a contains the peaks that are characteristic of SBA-15, of SBA-15 per ALD cycle, corresponding to 4.9 × 1017 Zr atom­
demonstrating that the SBA-15 structure was maintained following s⋅m− 2⋅cycle− 1. This growth rate is close to that reported previously for
ZrO2 ALD. Fig. S1b shows the wide angle XRD patterns of SBA-15, 15- ALD of TiO2 and Nb2O3 on SBA-15 [11,16]. After 15 ZrO2 ALD cycles,
ZrO2/SBA-15, and 5-WO3-15-ZrO2/SBA-15. No characteristics peaks the Zr coverage was 7.3 × 1018 atoms⋅m− 2. To determine the Zr atom
were detected in these diffraction patterns for either ZrO2 or WO3 which coverage to ZrO2 monolayers, one way is to use the surface density of Zr
is consistent with the formation of uniform ultra-thin films. on the cubic ZrO2(100) plane, 7.5 × 1018 atoms⋅m− 2. Based on this
For comparison purposes, we also synthesized a WO3-modified ZrO2 calculation, the ZrO2 coverage in the 15-ZrO2/SBA-15 sample is essen­
powder. The ZrO2 powder was synthesized by dissolving zirconyl hy­ tially 1 monolayer. Another way is to assume that two Si–OH groups on
drate (ZrO(NO3)2⋅xH2O, Sigma-Aldrich, 99%) in deionized water, fol­ SBA-15 react with each Zr(TMHD)4. On the basis of the Si–OH group
lowed by drying in air at 333 K for 12 h and calcining in air at 773 K for density on the mesoporous silica, 4.0 × 1018 Si–OH⋅m− 2, a monolayer of
5 h [24]. The resulting ZrO2 powder had a surface area of 77 m2/g. WO3 ZrO2 contains 2.0 × 1018 Zr atoms⋅m− 2 and the resulting coverage in the
was deposited on the ZrO2 powder using 5 ALD cycles using the pro­ 15-ZrO2/SBA-15 sample is 3.7 monolayers [14]. An analogous plot for
cedure described above. As mentioned above, an earlier study showed the ALD growth of WO3 on a 15-ZrO2/SBA-15 sample is shown in the
that 5 ALD cycles of WO3 maximized the Brønsted -site concentration of supporting information (Fig. S2). The amount of WO3 deposited per ALD
tungstated zirconia prepared by ALD [5]. cycle was also constant at 3.8 × 1017 W atoms⋅m− 2⋅cycle− 1 and is similar
N2 adsorption-desorption isotherms were measured at 78 K using a to that reported in a previous study [5].
Micromeritics, TriStar II Plus surface area and porosity analyzer. For Fig. 1b shows a plot of the surface area both normalized by the total
these measurements, the samples were pretreated at 573 K under mass and by the mass of only the SBA-15 as a function of the number of
evacuation. Surface areas were determined using the BET method and ZrO2 ALD cycles. Note that the total mass normalized surface area
pore-size distributions were determined from the adsorption isotherms decreased from 610 m2/g to 245 m2/g with 15 ALD cycles of ZrO2. The
using the Barrett-Joyner-Halenda (BJH) method. XRD patterns for each SBA-15 mass-normalized values show, however, that 75% of the SBA-15
sample were obtained using Xenocs Xeuss 2.0 and Rigaku MiniFlex 6G x- surface area was retained. These results demonstrate that the decrease of
ray diffractometers for 0–5 and 10–70◦ 2Θ, respectively. Transmission surface area is mainly associated with the increase in mass and argues

2
C.-Y. Wang et al. Microporous and Mesoporous Materials 335 (2022) 111821

Fig. 1. (a) Mass of ZrO2/mass of SBA-15 and (b) surface areas and normalized surface areas versus the number of ALD cycles.

against any pore blockage during ALD. The composition and surface between the average pore diameters of SBA-15 and that of the ALD-
areas of the ALD samples that were used throughout the remainder of modified SBA-15 or calculate the thickness from the mass of added
the study, namely, 15-ZrO2/SBA-15 and 5-WO3-15-ZrO2/SBA-15, are ZrO2, using the BET surface area and assuming the film has the same
tabulated in Table 1. For comparison purposes, data for SBA-15, bulk density as the bulk oxide, 5.68 g/cm3. Based on the change in the pore
ZrO2 powder, and 5-WO3/bulk ZrO2 are also included in the table. diameter, the ZrO2 film thickness for the 15-ZrO2/SBA-15 sample was
1.2 nm; however, it was only 0.25 nm based on the mass of the deposited
ZrO2 and its bulk density. One possible contributing factor to this
3.2. Structural characterization discrepancy is that the BJH method, which was used to calculate the
pore diameters, can be imprecise for pore sizes less than ~10 nm. This is
N2 adsorption-desorption isotherms were measured to characterize well documented in the literature where it has been shown that this
the changes in the SBA-15 pore size as a function of the number of ZrO2 method underestimates the pore size for many mesoporous materials,
ALD cycles. An adsorption-desorption isotherm was also measured for including SBA-type materials [26,27]. If one uses the BJH pore di­
the 5-WO3-15-ZrO2/SBA-15 sample. These data are displayed in Fig. 2a ameters, the film density would be less than 25% of that of the bulk
and b and show the adsorption-desorption isotherms and pore size dis­ value. Since for our growth procedure the film was annealed to 773 K in
tributions, respectively. The isotherms for all samples contain a hys­ each ALD cycle, a temperature that should be sufficient to produce a
teresis loop that results from a difference in the onset of capillary dense ZrO2 layer, this result seems unlikely.
condensation and evaporation of the N2 adsorbent in the linear pores in It should be noted, however, that in a study of the ALD growth of
the SBA structure. Note that, in addition to the linear pores indicative of TiO2 on SBA-15 Ke et al. also observed a large discrepancy between the
SBA-15, the pore size distribution plots show the presence of some less film thickness determined gravimetrically assuming a bulk density for
well-defined pores at small diameters, which we attribute to some the film, and that based on BJH pore size measurements [16]. They
amorphous silica that is present in our samples. attributed this difference to the film nucleation process resulting in a
As shown in Fig. 2b, the pristine SBA-15 had a unimodal pore size low-density film during the initial stages of growth. While we cannot
distribution centered at 9.0 nm which is consistent with that reported in rule out this possibility, dense oxide films are routinely grown on flat
the literature for this material [6]. The narrow peak width in this substrates including oxidized silicon wafers and it is not clear why a
isotherm is consistent with the SBA-15 having a relatively uniform pore different result would be obtained for a mesoporous silica, especially in
structure. Fig. 2a shows that, with increasing number of ZrO2 ALD cy­ the initial ALD cycles. It is noteworthy that, in at least some cases,
cles, the onset of capillary condensation shifted to lower P/P0 which can ALD-deposited films as thin as 0.5 nm can show diffraction patterns
be attributed to a decrease in pore diameter. Note that, after 15 ZrO2 consistent with bulk lattice parameters [28], a result that suggests that
ALD cycles (15-ZrO2/SBA-15), the average pore size decreased to 6.6 the ALD film densities are often similar to those of the bulk material.
nm; after subjecting this sample to 5 WO3 ALD cycles (5-WO3-15-Z­ The BET surface areas of the various samples are listed in Table 1.
rO2/SBA-15), it was only 5.4 nm. While the pore size decreased upon The surface area of the as-synthesized SBA-15 was 610 m2/g and
oxide film growth, the shape of the hysteresis loop in the adsorption decreased to 245 m2/g and 170 m2/g for the 15-ZrO2/SBA-15 and 5-
isotherms remained fairly constant, indicating that the pore structure WO3-15-ZrO2/SBA-15 samples, respectively. Much of the decrease upon
was maintained and that the pores were uniformly covered by the metal ALD oxide film growth can be attributed to the increase in the sample
oxide film. mass but some of the decrease must also be associated with a decrease in
To estimate the ZrO2 film thickness, one can either use the difference pore size. Even so, it is noteworthy that the surface area of the ALD-
synthesized 15-ZrO2/SBA-15 (245 m2/g) was still 3 times higher than
Table 1 that of the bulk ZrO2 (77 m2/g) sample.
Sample compositions and surface areas. TEM and STEM were also used to characterize the structure of the
Material Composition (wt%) Surface Area (m2/g) SBA-15 sample and demonstrate that the Zr in the 15-ZrO2/SBA-15
SBA-15 SiO2: 100 610
sample was in the pores. Fig. 3a and b shows representative TEM images
15-ZrO2/SBA-15 ZrO2: 48 245 of SBA-15 and 15-ZrO2/SBA-15, respectively. In the SBA-15 TEM image,
SiO2: 52 both the linear pores (left region) and their arrangement in a hexagonal
5-WO3-15-ZrO2/SBA-15 WO3: 15 170 pattern (upper right region), which are characteristic of this material,
ZrO2: 41
are readily apparent. As shown in Fig. 3b, consistent with the BET re­
SiO2: 44
Bulk ZrO2 ZrO2: 100 77 sults, this structure is maintained in the 15-ZrO2/SBA-15 sample. Fig. 3c
5-WO3/bulk ZrO2 WO3: 6 65 shows a representative STEM image of the 15-ZrO2/SBA-15 sample
ZrO2: 94 along with EDS maps of Si and Zr from the indicated region. The Zr EDS

3
C.-Y. Wang et al. Microporous and Mesoporous Materials 335 (2022) 111821

Fig. 2. (a) N2 adsorption-desorption isotherms and (b) pore size distributions for 0 to 15 ALD cycles of ZrO2 on SBA-15 and 5 ALD cycles of WO3 on 15-ZrO2/SBA-15.

Fig. 3. TEM images of (a) SBA-15 and (b) 15-ZrO2/SBA-15. (c) STEM image of 15-ZrO2/SBA-15 with EDS maps of Si and Zr.

map shows that the Zr is dispersed uniformly throughout the SBA-15 distribution, surface area, and volume of the linear pores in the SBA-15
which is again consistent with the adsorption results presented above. and 15-ZrO2/SBA-15 samples were measured after annealing each
sample in air for 24 h at a series of successively higher temperatures
3.3. Thermal stability between 1073 and 1273 K. These data are displayed in Fig. 4. We
defined the volume of linear pores as the pores with width larger than 4
The effect of coating the pores with ZrO2 on the thermal stability of nm. The corresponding adsorption-desorption isotherms for each sam­
the SBA-15 was also investigated. In these studies, the pore size ple are provided in the supporting information (see Fig. S3).

4
C.-Y. Wang et al. Microporous and Mesoporous Materials 335 (2022) 111821

Fig. 4. Pore size distributions of (a) SBA-15 and (b) 15-ZrO2/SBA-15 under 773–1273 K treatment. (c) Surface areas and linear pore volumes of SBA-15, 15-ZrO2/
SBA-15, and bulk ZrO2 under 773–1273 K treatment.

As shown in Fig. 4, there were significant differences in the thermal disappearance of the hysteresis loop that is characteristic of the linear
stability of the SBA-15 and 15-ZrO2/SBA-15 samples. For SBA-15 pores in the adsorption isotherm (see Fig. S3a). In contrast, while some
annealing at 1073 K produced 26% and 26% decreases in the surface degradation did occur upon heating the 15-ZrO2/SBA-15 sample to
area and linear pore volume, respectively, while the corresponding 1173 K, as evidenced by decreases in surface area and pore volume (35%
values for 15-ZrO2/SBA-15 were only 6% and 7%. For the SBA-15, and 29%, respectively relative to the as-synthesized sample), it is
further significant decreases in surface area and pore volume occurred noteworthy that, for this sample, much of the linear pore structure
upon heating to 1173 K and complete collapse of the SBA-15 pore remained intact, even after heating to 1273 K, as evidenced by the
structure occurred upon heating to 1273 K, as evidenced by the persistence of the well-defined hysteresis loop in the adsorption

Fig. 5. 2-propanol TPD-TGA from (a) SBA-15, (b) 15-ZrO2/SBA-15, and (c) 5-WO3-15-ZrO2/SBA-15. m/e = 41: propene; m/e = 41 and 45: 2-propanol.

5
C.-Y. Wang et al. Microporous and Mesoporous Materials 335 (2022) 111821

isotherm. Somewhat surprisingly, the sample annealed at this temper­

ature still had a linear pore volume that was 55% of that in the as-
synthesized sample. Even after this very harsh treatment, the 15-ZrO2/
SBA-15 sample still had a surface area that was 1.5 times higher than
that of the untreated, bulk ZrO2 powder. These results demonstrate that
coating the pores in SBA-15 with an ALD-grown ZrO2 film produces a
thermally stable, high surface area ZrO2.

3.4. Characterization of Lewis and Brønsted acid sites

Since 2-propanol undergoes dehydration to produce propene and

water on acid sites on many metal oxide surfaces [29,30], TPD-TGA of
2-propanol was used to characterize the presence and strength of such
sites on the SBA-15, 15-ZrO2/SBA-15, and 5-WO3-15-ZrO2/SBA-15
samples. These experiments also provided additional insight into the
uniformity of the ALD-deposited oxide films, as discussed elsewhere [5].
The TPD-TGA results for 2-propanol in SBA-15, shown in Fig. 5a, indi­
cated that most of the alcohol (m/e = 41 and 45) desorbs intact in a large
peak centered at 390 K, with a small fraction desorbing as propene (m/e
= 41) in a peak centered at 770 K. This is consistent with reports in the
literature for purely siliceous materials that indicate 2-propanol reacts
with surface silanols to form a silyl-isopropyl ether species whose
decomposition produces propene between 700 and 800 K [31,32]. Based
on the TGA data, the density of sites on which this reaction takes place
on SBA-15 was 130 μmol/g (1.3 × 1017 molecule/m2). This low
coverage suggests that only a small fraction of the silanol groups on the Fig. 6. DRIFT spectra with pyridine dosed SBA-15, 15-ZrO2/SBA-15, 5-WO3-
surface react with the alcohol. 15-ZrO2/SBA-15, and 5-WO3/bulk ZrO2. B and L represent Brønsted and Lewis
Significantly different results were obtained for the 15-ZrO2/SBA-15 acid sites, respectively.
(Fig. 5b). For this sample, the 770 K propene peak observed for the bare
SBA-15 was replaced by a much larger peak centered at 505 K. The 3.5. H-D exchange
absence of the 770 K peak is consistent with the ALD ZrO2 film
completely covering the SBA-15 surface. The large propene peak at 505 The H-D exchange reaction between D2O and toluene was used to
K indicates that the ZrO2-coated sample has a higher density of active further characterize the active sites on the surface of the 5-WO3-15-
sites, in this case, exposed Zr2+ cations which act as Lewis acids. Based ZrO2/SBA-15 sample. H-D exchange between Brønsted sites and D2O is
on the amount of propene produced, the Lewis-acid site density on the known to be facile at all temperatures, but the temperature at which
15-ZrO2/SBA-15 sample was 355 μmol/g. This corresponds to 9 × 1017 proton transfer occurs between a Brønsted site and toluene is strongly
molecules/m2, a value close to that expected for a close-packed mono­ dependent on the strength of the Brønsted site [25]. H-D exchange is
layer of the alcohol. Similar 2-propanol TPD-TGA results were obtained expected to occur at much lower rates on Lewis sites. Fig. 7 plots the
for the 5-WO3-15-ZrO2/SBA-15 sample, although the propene desorp­ steady-state conversion for the H-D exchange reaction as a function of
tion feature shifted down to 420 K suggesting either stronger Lewis temperature for the SBA-15, 15-ZrO2/SBA-15, and 5-WO3-15-Z­
acidity compared to 15-ZrO2/SBA-15 or reaction on Brønsted sites that rO2/SBA-15 samples. These data show that the SBA-15 was inactive and
are associated with the WO3. The acid site density, 390 μmol/g, how­ the 15-ZrO2/SBA-15 sample was nearly inactive, requiring temperatures
ever, was similar to that of the 15-ZrO2/SBA-15 sample.
Since WO3-decorated ZrO2 is reported to have Brønsted acid sites [4,
33], DRIFT spectroscopy of adsorbed pyridine was used as a probe for
the presence of such sites on the 5-WO3-15-ZrO2/SBA-15 sample. Fig. 6
displays a DRIFT spectrum obtained after exposing this sample to pyri­
dine at 423 K. For comparison purposes, spectra for pyridine-dosed
SBA-15, 15-ZrO2/SBA-15 and 5-WO3/bulk ZrO2 are also included in
the figure. The spectrum for 5-WO3/bulk ZrO2 is similar to that reported
in the literature for tungstated zirconias [5,34,35] and contains promi­
nent adsorption peaks centered at 1600 cm− 1 and 1540 cm− 1 which
have been attributed to pyridine adsorbed on Lewis acid sites and pro­
tonated pyridine (pyridinium ions) adsorbed on Brønsted-acid sites,
respectively. For SBA-15, the spectrum shows no observable peak
associated with Brønsted acid sites and low peak intensity for Lewis acid
site, implying that SBA-15 possesses no acid site characteristics. While a
prominent peak due to adsorption on Lewis acid sites is present in the
spectrum for 15-ZrO2/SBA-15, there is relatively little adsorption in the
region characteristic of the pyridinium species, indicating the lack of
Brønsted acid sites on this material. In contrast, a prominent peak at
1540 cm− 1 due to pyridinium ions adsorbed on Brønsted-acid sites is
again present in the spectrum for 5-WO3-15-ZrO2/SBA-15, confirming
the presence of Brønsted-acid sites on this sample. Thus, the ALD
deposited WO3/ZrO2 layer on SBA-15 appears to have surface properties Fig. 7. Conversion for H-D exchange between toluene and D2O as a function of
similar to that of bulk WO3/ZrO2. temperature for SBA-15, 15-ZrO2/SBA-15, and 5-WO3-15-ZrO2/SBA-15.

6
C.-Y. Wang et al. Microporous and Mesoporous Materials 335 (2022) 111821

in excess of 600 K to produce minimal conversion. This is consistent with however, the ability to prepare materials with well-defined composi­
the lack of Brønsted acid sites on this material. Significantly different tions and pore structures is intriguing. The application to mesoporous
results were obtained for the 5-WO3-15-ZrO2/SBA-15 sample, which materials, demonstrated here, provides one more example of how this
was highly active for the D2O/toluene H-D exchange reaction. For this approach can be used to prepare interesting catalytic materials.
sample, the reaction onset temperature, as defined by the temperature at
which 10% conversion is obtained, was 393 K; and greater than 80% 5. Conclusions
conversion was obtained at 475 K. This result is consistent with the
DRIFT spectra results described above and further confirms the presence In this study, we demonstrated that ALD can be used to grow tung­
of Brønsted acid sites on this catalyst. stated zirconia on the surface of the pores in SBA-15. The observed
It is useful to compare the results obtained for 5-WO3-15-ZrO2/SBA- systematic decrease in the pore diameter with the number of ALD cycles,
15 with those reported previously for bulk WO3/ZrO2 and transition along with 2-proponal TPD results demonstrated that the oxide films
metal substituted H-ZSM-5s by Wang et al. who used reaction conditions were conformal to the pore surfaces. Zirconia-coated SBA-15 was shown
nearly identical to those used in the present study [5,25]. They reported to enhance the thermal stability of the SBA-15 support and to have
that the reaction onset temperature for a 10 wt % WO3/ZrO2 was 373 K surface Lewis acid sites that are typical of zirconia. In contrast, a
which is close to that obtained here for 5-WO3-15-ZrO2/SBA-15. This tungstated-zirconia modified SBA-15 that was synthesized using 15 ZrO2
demonstrates that the ALD-grown WO3/ZrO2 layer in SBA-15 contains ALD cycles followed by 5 WO3 ALD cycles exhibited properties consis­
Brønsted acid sites that are similar in strength to those in a bulk tung­ tent with the presence of surface Brønsted acid sites, including the for­
stated zirconia. In agreement with the previous study, the Brønsted sites mation of adsorbed pyridinium upon exposure to pyridine. These
in 5-WO3-15-ZrO2/SBA-15 are slightly less reactive than those in H-[Al] Brønsted sites exhibited reactivity similar to that of bulk tungstated
ZSM-5, which has a reaction onset temperature of ~335 K, or even zirconia in that they were highly active for the acid-catalyzed H-D ex­
H-[Fe]ZSM-5, 345 K. change between D2O and toluene.
The results of this study also further demonstrate the viability of
4. Discussion using ALD to synthesize specific active sites on the surface of meso­
porous oxides. The pore structure of these materials can be maintained
In this study we demonstrated the use of ALD to grow films of ZrO2 in while growing oxide films on their surfaces and at least in the case of
the pores of SBA-15. Upon modification with WO3 this material con­ zirconia, the ALD oxide film can also significantly enhance the thermal
tained Brønsted-acid sites that exhibited catalytic properties similar to stability of the mesoporous support.
those of bulk tungstated zirconia. The ALD growth process resulted in
conformal films with a ZrO2 growth rate was 4.9 × 1017 Zr atom­ CRediT authorship contribution statement
s⋅m− 2⋅cycle− 1 which remained constant for the first 15 ALD cycles. Due
to the decrease in pore diameter and surface area which occurs upon Ching-Yu Wang: Writing – original draft, Investigation. Ohhun
ALD film growth, one might expect the growth rate per cycle to decrease Kwon: Investigation. Raymond J. Gorte: Supervision. John M. Vohs:
with increasing cycle number once the oxide film completely covers the Supervision, Writing – review & editing.
SBA-15 support. Note, however, that since 15 ZrO2 ALD cycles produces
a ZrO2 film that is only slightly beyond monolayer coverage, it is likely Declaration of competing interest
we did not perform enough cycles to see the expected decrease in growth
rate. The authors declare that they have no known competing financial
In addition to demonstrating the formation of active Brønsted sites interests or personal relationships that could have appeared to influence
for WO3-modified 15-ZrO2/SBA-15, this study also showed that the the work reported in this paper.
thermal stability of SBA-15 was greatly enhanced by coating the pores
with an ALD-grown ZrO2 film. As shown in Fig. 4, the linear pore Acknowledgements
structure was maintained after heating 15-ZrO2/SBA-15 to 1273 K,
while for the unmodified SBA-15 the pore structure completely This work was funded by the Department of Energy, Office of Basic
collapsed upon heating to this temperature. This suggests that using ALD Energy Sciences, Chemical Sciences, Geosciences and Biosciences Divi­
to coat the pores of mesoporous materials with a refractory oxide, such sion, Grant No. DE-FG02-13ER16380. The STEM work was carried out in
as zirconia, may be a viable strategy for producing a high-surface-area part at the Singh Center for Nanotechnology, part of the National
catalyst support material that can be used in applications that require Nanotechnology Coordinated Infrastructure Program, which is sup­
high temperatures. ported by the National Science Foundation grant NNCI-1542153. The
Finally, in addition to providing a method to produce a thermally authors acknowledge Dr. Daeyeon Lee and Dr. Jaehyun Kim at the
stable, high-surface-area, tungstated zirconia (170 m2g-1), a result which University of Pennsylvania for the help on SBA-15 synthesis. The authors
cannot be obtained using a bulk ZrO2 support, these materials may have acknowledge use of the Dual Source and Environmental X-ray Scattering
advantages for other applications. One such application is suggested by facility operated by the Laboratory for Research on the Structure of
the recent study of Tennakoon et al. on the use of Pt-modified meso­ Matter at the University of Pennsylvania (NSF MRSEC 17-20530). The
porous silica for the selective hydrogenolysis of individual polymer equipment purchase was made possible by a NSF MRI grant (17-25969),
chains in high-density polyethylene [36]. The authors of that study a ARO DURIP grant (W911NF-17-1-0282), and the University of
showed that catalytic sites that can only be accessed by diffusion of the Pennsylvania.
polymer chains into mesopores were more selective to forming higher
value, larger alkanes. Since Pt/WOx/ZrO2 has been shown to exhibit
Appendix A. Supplementary data
interesting properties for hydrocracking of polymers [37], the ability to
produce these sites in mesopores is intriguing. Finally, recent work has
Supplementary data to this article can be found online at https://doi.
shown that polymer-surface interactions can be changed significantly by
org/10.1016/j.micromeso.2022.111821.
modifying the surface composition of pores using ALD [38]. It is not yet
clear what impact this will have on the catalytic properties but the
References
ability to tune adsorption properties represents another variable in the
catalytic toolbox. [1] D.G. Barton, S.L. Soled, E. Iglesia, Solid acid catalysts based on supported tungsten
ALD as a means to prepare catalytic materials is still in its infancy; oxides, Top. Catal. 6 (1998) 87–99.

7
C.-Y. Wang et al. Microporous and Mesoporous Materials 335 (2022) 111821

[2] M. Scheithauer, T.K. Cheung, R.E. Jentoft, R.K. Grasselli, B.C. Gates, H. Knozinger, supported on high-surface-area LaFeO3 prepared by atomic layer deposition,
Characterization of WOx/ZrO2 by vibrational spectroscopy and n-pentane J. Am. Chem. Soc. 140 (2018) 4841–4848.
isomerization catalysis, J. Catal. 180 (1998) 1–13. [21] G. Dingemans, C.A.A. van Helvoirt, D. Pierreux, W. Keuning, W.M.M. Kessels,
[3] N. Soultanidis, W. Zhou, A.C. Psarras, A.J. Gonzalez, E.F. Iliopoulou, C.J. Kiely, I. Plasma-assisted ALD for the conformal deposition of SiO2: process, material and
E. Wachs, M.S. Wong, Relating n-pentane isomerization activity to the tungsten electronic properties, J. Electrochem. Soc. 159 (2012) H277–H285.
surface density of WOx/ZrO2, J. Am. Chem. Soc. 132 (2010) 13462–13471. [22] J. Musschoot, J. Dendooven, D. Deduytsche, J. Haemers, G. Buyle, C. Detavernier,
[4] W. Zhou, N. Soultanidis, H. Xu, M.S. Wong, M. Neurock, C.J. Kiely, I.E. Wachs, Conformality of thermal and plasma enhanced atomic layer deposition on a non-
Nature of catalytically active sites in the supported WO3/ZrO2 solid acid system: a woven fibrous substrate, Surf. Coating. Technol. 206 (2012) 4511–4517.
current perspective, ACS Catal. 7 (2017) 2181–2198. [23] H.C.M. Knoops, J.W. Elam, J.A. Libera, W.M.M. Kessels, Surface loss in ozone-
[5] C. Wang, X.Y. Mao, J.D. Lee, T.M. Onn, Y.H. Yeh, C.B. Murray, R.J. Gorte, based atomic layer deposition processes, Chem. Mater. 23 (2011) 2381–2387.
A characterization study of reactive sites in ALD-synthesized WOx/ZrO2 catalysts, [24] Y.C. Ji, A. Lawal, A. Nyholm, R.J. Gorte, O.A. Abdelrahman, Dehydra-decyclization
Catalysts 8 (2018) 15. of tetrahydrofurans to diene monomers over metal oxides, Catal. Sci. Technol. 10
[6] D.Y. Zhao, J.L. Feng, Q.S. Huo, N. Melosh, G.H. Fredrickson, B.F. Chmelka, G. (2020) 5903–5912.
D. Stucky, Triblock copolymer syntheses of mesoporous silica with periodic 50 to [25] C. Wang, S. Li, X.Y. Mao, S. Caratzoulas, R.J. Gorte, H-D exchange of simple
300 angstrom pores, Science 279 (1998) 548–552. aromatics as a measure of brønsted-acid site strengths in solids, Catal. Lett. 148
[7] T. Li, S.T. Wong, M.C. Chao, H.P. Lin, C.Y. Mou, S.F. Cheng, N-pentane (2018) 3548–3556.
isomerization over platinum-promoted W/Zr mixed oxides supported on [26] C.G. Sonwane, S.K. Bhatia, Characterization of pore size distributions of
mesoporous silica, Appl. Catal. A-Gen. 261 (2004) 211–219. mesoporous materials from adsorption isotherms, J. Phys. Chem. B 104 (2000)
[8] W. Lee, C.C. Yang, S. Cheng, Isomerization of n-pentane over platinum promoted 9099–9110.
tungstated zirconia supported on mesoporous SBA-15 prepared by supercritical [27] J.C. Groen, L.A.A. Peffer, J. Perez-Ramirez, Pore size determination in modified
impregnation, J. Chin. Chem. Soc. 68 (2021) 409–420. micro- and mesoporous materials. Pitfalls and limitations in gas adsorption data
[9] Z. Gao, Y. Qin, Design and properties of confined nanocatalysts by atomic layer analysis, Microporous Mesoporous Mater. 60 (2003) 1–17.
deposition, Accounts Chem. Res. 50 (2017) 2309–2316. [28] T.Y. Cao, O. Kwon, C. Lin, J.M. Vohs, R.J. Gorte, Two-dimensional perovskite
[10] P. Cop, E. Celik, K. Hess, Y. Moryson, P. Klement, M.T. Elm, B.M. Smarsly, Atomic crystals formed by atomic layer deposition of CaTiO3 on gamma-Al2O3,
layer deposition of nanometer-sized CeO2 layers in ordered mesoporous ZrO2 films Nanomaterials (2021) 11.
and their impact on the ionic/electronic conductivity, ACS Appl. Nano Mater. 3 [29] S. Roy, G. Mpourmpakis, D.Y. Hong, D.G. Vlachos, A. Bhan, R.J. Gorte, Mechanistic
(2020) 10757–10766. study of alcohol dehydration on gamma-Al2O3, ACS Catal. 2 (2012) 1846–1853.
[11] Y.J. Pagan-Torres, J.M.R. Gallo, D. Wang, H.N. Pham, J.A. Libera, C.L. Marshall, J. [30] K. Larmier, C. Chizallet, N. Cadran, S. Maury, J. Abboud, A.F. Lamic-Humblot,
W. Elam, A.K. Datye, J.A. Dumesic, Synthesis of highly ordered hydrothermally E. Marceau, H. Lauron-Pernot, Mechanistic investigation of isopropanol conversion
stable mesoporous niobia catalysts by atomic layer deposition, ACS Catal. 1 (2011) on alumina catalysts: location of active sites for alkene/ether production, ACS
1234–1245. Catal. 5 (2015) 4423–4437.
[12] Y.Q. Li, S.C. Zhao, Q.M. Hu, Z. Gao, Y.Q. Liu, J.K. Zhang, Y. Qin, Highly efficient [31] R.J. Huang, Y. Cheng, Y.C. Ji, R.J. Gorte, Atomic layer deposition for preparing
CoOx/SBA-15 catalysts prepared by atomic layer deposition for the epoxidation isolated Co sites on SiO2 for ethane dehydrogenation catalysis, Nanomaterials 10
reaction of styrene, Catal. Sci. Technol. 7 (2017) 2032–2038. (2020) 13.
[13] J. Dendooven, B. Goris, K. Devloo-Casier, E. Levrau, E. Biermans, M.R. Baklanov, K. [32] M.T. Aronson, R.J. Gorte, W.E. Farneth, The influence of oxonium ion and
F. Ludwig, P. Van der Voort, S. Bals, C. Detavernier, Tuning the pore size of ink- carbenium ion stabilities on the Alcohol/H-ZSM-5 interaction, J. Catal. 98 (1986)
bottle mesopores by atomic layer deposition, Chem. Mater. 24 (2012) 1992–1994. 434–443.
[14] J.E. Herrera, J.H. Kwak, J.Z. Hu, Y. Wang, C.H.F. Peden, J. Macht, E. Iglesia, [33] V. Lebarbier, G. Clet, M. Houalla, A comparative study of the surface structure,
Synthesis, characterization, and catalytic function of novel highly dispersed acidity, and catalytic performance of tungstated zirconia prepared from crystalline
tungsten oxide catalysts on mesoporous silica, J. Catal. 239 (2006) 200–211. zirconia or amorphous zirconium oxyhydroxide, J. Phys. Chem. B 110 (2006)
[15] Z.H. Weng, Z.H. Chen, X.D. Qin, F. Zaera, Sub-monolayer control of the growth of 13905–13911.
oxide films on mesoporous materials, J. Mater. Chem. 6 (2018) 17548–17558. [34] J. Rorrer, Y. He, F.D. Toste, A.T. Bell, Mechanism and kinetics of 1-dodecanol
[16] W. Ke, Y. Liu, X.L. Wang, X.D. Qin, L.M. Chen, R.M. Palomino, J.P. Simonovis, etherification over tungstated zirconia, J. Catal. 354 (2017) 13–23.
I. Lee, I. Waluyo, J.A. Rodriguez, A.I. Frenkel, P. Liu, F. Zaera, Nucleation and [35] C. Wang, T. Xie, P.A. Kots, B.C. Vance, K. Yu, P. Kumar, J. Fu, S. Liu,
initial stages of growth during the atomic layer deposition of titanium oxide on G. Tsilomelekis, E.A. Stach, W. Zheng, D.G. Vlachos, Polyethylene Hydrogenolysis
mesoporous silica, Nano Lett. 20 (2020) 6884–6890. at Mild Conditions over Ruthenium on Tungstated Zirconia, JACS Au, 2021.
[17] M.H. Wang, Y. Li, X.L. Gao, C.R. Liu, F. Xing, TiO2 ultrathin coating of SBA-15/Au [36] A. Tennakoon, X. Wu, A.L. Paterson, S. Patnaik, Y.C. Pei, A.M. LaPointe, S.
by atomic layer deposition for enhanced catalytic performance in epoxidation of C. Ammal, R.A. Hackler, A. Heyden, Slowing II, G.W. Coates, M. Delferro,
styrene, J. Phys. Chem. C 124 (2020) 10530–10540. B. Peters, W.Y. Huang, A.D. Sadow, F.A. Perras, Catalytic upcycling of high-density
[18] R.J. Huang, O. Kwon, C. Lin, R.J. Gorte, The effects of SMSI on m-Cresol polyethylene via a processive mechanism, Nat. Catal. 3 (2020) 893–901.
hydrodeoxygenation over Pt/Nb2O5 and Pt/TiO2, J. Catal. 398 (2021) 102–108. [37] S.B. Liu, P.A. Kots, B.C. Vance, A. Danielson, D.G. Vlachos, Plastic waste to fuels by
[19] T.M. Onn, S.Y. Zhang, L. Arroyo-Ramirez, Y.C. Chung, G.W. Graham, X.Q. Pan, R. hydrocracking at mild conditions, Sci. Adv. 7 (2021) 9.
J. Gorte, Improved thermal stability and methane-oxidation activity of Pd/Al2O3 [38] T. Ren, R. Huang, R.J. Gorte, D. Lee, Modulating interactions between molten
catalysts by atomic layer deposition of ZrO2, ACS Catal. 5 (2015) 5696–5701. polystyrene and porous solids using atomic layer deposition, Langmuir 37 (2021)
[20] T.M. Onn, M. Monai, S. Dai, E. Fonda, T. Montini, X.Q. Pan, G.W. Graham, 14520–14526.
P. Fornasiero, R.J. Gorte, Smart Pd catalyst with improved thermal stability