Catalytic properties of carbon nanotubes-supported heteropolyacids in isopropanol conversion

Article abstract of DOI:10.1016/j.apcata.2017.10.008

The technique of catalytic flow microreactor has been combined with the gas flow-through microcalorimetry to correlate the catalytic activity of supported heteropolyacids with both the acid strength of protons as well as the protons’ accessibility. Multiwall carbon nanotubes (CNT) were used as a support for Keggin (H₃PW₁₂O₄₀) and Wells-Dawson (H₆P₂W₁₈O₆₂) structured heteropolyacids, in order to produce catalysts combining high acidity from the parent acids with the inherent microporosity of the support. Prior to the catalytic tests, the obtained materials were characterized by X-ray diffraction (XRD), scanning electron microscopy (SEM), Fourier transform infrared (FTIR) and Raman spectroscopies as well as by the nitrogen adsorption-desorption analysis (BET). The latter technique confirmed overall improved porosity of the obtained materials. Upon testing for activity in the isopropyl alcohol dehydration, the supported Wells-Dawson catalysts turned out to be superior to both the Keggin-based materials, as well as to the unsupported H₆P₂W₁₈O₆₂. It has been found that the improvement of catalytic performance in the isopropanol conversion is mostly related to the increase of the accessibility of protons, rather than to the changes in the acid strengths.

Full text of DOI:10.1016/j.apcata.2017.10.008

AppliedCatalysisA,General549(2018)254–262
Contents lists available at ScienceDirect
Applied Catalysis A, General
journal homepage: www.elsevier.com/locate/apcata
Research Paper
Catalytic properties of carbon nanotubes-supported heteropolyacids in
isopropanol conversion
A. Kirpsza, E. Lalik, G. Mordarski, A. Micek-Ilnicka⁎
Jerzy Haber Institute of Catalysis and Surface Chemistry, Polish Academy of Sciences, Niezapominajek 8, PL-30239 Krakow, Poland
A R T I C L E I N F O
A B S T R A C T
Keywords:
The technique of catalytic flow microreactor has been combined with the gas flow-through microcalorimetry to
correlate the catalytic activity of supported heteropolyacids with both the acid strength of protons as well as the
protons’ accessibility. Multiwall carbon nanotubes (CNT) were used as a support for Keggin (H3PW12O40) and
Wells-Dawson (H6P2W18O62) structured heteropolyacids, in order to produce catalysts combining high acidity
from the parent acids with the inherent microporosity of the support. Prior to the catalytic tests, the obtained
materials were characterized by X-ray diffraction (XRD), scanning electron microscopy (SEM), Fourier transform
infrared (FTIR) and Raman spectroscopies as well as by the nitrogen adsorption-desorption analysis (BET). The
latter technique confirmed overall improved porosity of the obtained materials. Upon testing for activity in the
isopropyl alcohol dehydration, the supported Wells-Dawson catalysts turned out to be superior to both the
Keggin-based materials, as well as to the unsupported H6P2W18O62. It has been found that the improvement of
catalytic performance in the isopropanol conversion is mostly related to the increase of the accessibility of
protons, rather than to the changes in the acid strengths.
Heteropolyacids
Isopropanol conversion
Carbon nanotubes
Diisopropyl ether
DIPE
Propylene
1. Introduction
by Chimienti et al. [11]. The authors used both the commercial wood-
based activated carbon with a surface area of 960 m2/g and the silica
The mobility of protons in the heteropolyacids (HPAs) is funda-
mental for their performance in catalytic conversion of alcohols,
making possible the activation of the organic molecule by addition of
the H+ species to the oxygen atom and forming methoxonium and
ethoxonium ions [1,2]. Using HPAs in the form of supported catalysts
can improve the accessibility of protons to the alcohol molecules, due to
increased porosity resulting from high surface area of the supporting
materials. However, there appears to be a trade-off between the ac-
cessibility of protons and their acidity in the supported HPAs, since the
interaction between the HPA and the support can in some cases reduce
acidity of the former.
The use of alcohols as reactive probe molecules to investigate acidic
properties has been developed since the 1970s [3–8]. It was reviewed
by Lauron-Pernot in [9], describing the catalytic data of alcohols con-
version on various catalysts. The use of 2-propanol as a probe molecule
in gas phase over solid supported catalysts containing heteropolyacids
(HPA) as active phase has been described in [10]. The authors noticed
that heteropolytungstic acid H3PW12O40 exhibited high catalytic ac-
tivity and selectivity when it was loaded onto activated carbon. They
investigated a broad series of activated carbon supports with large
surface areas of 920–1656 m2/g. These investigations were continued
gel impregnated with either H3PW12O40 or H4SiW12O40 hetero-
polyacids. They found a dependence of their respective content on the
activity in isopropanol dehydration at 150 °C, for samples containing
less than ca. 25%wt. of HPA. However, further increase of the HPA
content did not improve the dehydration activity (constant 2-propanol
conversion). The authors of [8] concluded, that the reaction was cata-
lyzed mainly by the surface protons of heteropolyacid crystallites, ra-
ther than those located deeper in the bulk of crystals, thus confirming
the accessibility of protons to be an important factor for the catalytic
activity of supported HPAs.
Arguably, the latter may also be a function of structure of the HPA
being used. While there is a profusion of data reporting catalytic ac-
tivity of the Keggin-type HPAs, the Wells-Dawson-type HPAs seems to
have been receiving relatively less attention. In this paper we attempt to
compare catalytic properties of the Wells-Dawson-type H6P2W18O62
(HP2W) with the Keggin-type H3PW12O40 (HPW), pure and supported
on the carbon nanotubes (CNT), and to correlate textural properties of
these catalysts, as well as their microcalorimetrically measured acidity,
with their activity for the isopropanol dehydration. HPW is considered
to be the strongest acid among all known HPAs [12]. In spite of this
fact, the Wells-Dawson-type catalysts (HP2W/support) are considered
⁎
Corresponding author.
E-mail address: ncilnick@cyf-kr.edu.pl (A. Micek-Ilnicka).
http://dx.doi.org/10.1016/j.apcata.2017.10.008
Received 4 July 2017; Received in revised form 11 October 2017; Accepted 12 October 2017
Availableonline14October2017
0926-860X/©2017PublishedbyElsevierB.V.

A. Kirpsza et al.
AppliedCatalysisA,General549(2018)254–262
to be more effective in the acid–base type reactions, compared to the
Keggin-type, in both the gas and the liquid phase reactions [13]. In
particular, they appear to be very good and selective catalysts for ter-
tiary ethers synthesis by alcohol addition to olefins [14].
Table 1
Characterization of CNT and modified HPA/CNT catalysts.
Sample
HPA (% wt.)
BET (m2/g)
The CNT may be considered as a useful support in heterogeneous
catalysis because, first, they are inert enough to avoid strong interaction
with active phase and to remain catalytically inactive themselves, and
secondly, because of their high surface area combined with porous
structure. There have been only a handful of reports describing the
application of carbon nanotubes as supports for heteropolyacids,
against the prevalence of active carbon materials or carbon fibers [15]
having been used for the purpose. The catalysts comprising a hetero-
polyacid and CNT or CNT modified with nitrogen atoms were in-
vestigated only by Timofeeva [16] in esterification of n-butanol with
acetic acid, by Qi [17] in ester hydrolysis reactions and by the present
authors [18] in ethanol conversion. A CNT-supported H3PW12O40·6H2O
was investigated as potential material for NOx capture [19]. To our best
knowledge, the Wells-Dawson-type HPA supported on CNT has not
been investigated yet. The aim of the present study is to confirm the
potential of carbon nanotubes as supports for Wells-Dawson-type het-
eropolyacid as an active phase and the comparison of these systems
with the Keggin-type heteropolyacid supported on CNT.
CNT
–
136
58
33
71
32
0.5HPW/CNT
1.0HPW/CNT
0.5HP2W/CNT
1.0HP2W/CNT
27.0
42.5
30.2
46.4
surface area of 136 m2/g has been measured basing on the N2 sorption.
Two series of catalysts were synthesized: HPW/CNT and HP2W/
CNT in which CNT were used as a support. The appropriate amounts of
heteropolyacids HPW and HP2W were dissolved in absolute ethanol
(10–30 cm3) and mixed with carbon nanotubes. The suspensions were
evaporated at room temperature and dried at 333 K for 2 h. The amount
of heteropolyacid used corresponded to the coverage of support with
approximately 0.5 or 1.0 of heteropolyacid monolayer. Hence, the
samples were labeled as 0.5HPW/CNT or 1.0HP2W/CNT, etc. where
the values 0.5 or 1.0 denote the coverage of the surface with hetero-
polyacid. Calculations of the number of layers were carried out based
on the values: 11.7 Å [22] and 12.8 Å [23] as the dimensions of
3−
6−
[PW12O40
]
and [P2W18O62
]
heteropolyanions, respectively, and
the surface covered by a single molecule of HPW was assumed to be
equal to 107 Å2, while that covered by a HP2W molecule to 129 Å2.
Table 1 presents composition of the catalysts. Isopropyl alcohol (an-
hydrous, 99.5%, Sigma-Aldrich) was used as reagent in catalytic tests,
whereas propylene (≥99%, Aldrich) and diisopropyl ether (analytical
standard, Sigma-Aldrich) were used in calibration procedures in the
catalytic experiments.
2. Experimental
2.1. Materials
The Keggin-type heteropolyacid H3PW12O40 (HPW) has been pro-
vided by Aldrich p.p.a. The Wells-Dawson type H6P2W18O62 (HP2W)
was prepared in our lab according to [20]. In order to maintain a
constant content of crystallization water, the obtained HP2W was
stored over saturated solution of Mg(NO3)2 prior to its use in experi-
ments. The samples so treated had the compositions corresponding to
the formulas: H3PW12O40 26.6H2O and H6P2W18O62 28.8H2O. The
commercial multi-wall carbon nanotubes (Sunnano®, China) were
formed by CVD process, resulting in a material of purity > 80% and the
outer diameter of nanotubes ranging, according to the producer’s in-
formation, from 100 to 300 Å. As for the inner diameter, our own es-
timation has been made using a dedicated software (ImageJ 1.51p) for
analysis of our SEM images. The results of a separate estimations at
twenty eight selected points turned out to be falling into three ranges:
38–60 Å, 77–92 Å and 110–123 Å (cf. Fig. 1A). The XRF analysis shows
mainly Fe and Ni impurities, with traces of Zn (cf. Fig. 1B). The former
are likely resulting from using spinels in the synthesis of CNT [21]. The
Fe content was determined based on analysis of a series of three sam-
ples prepared by adding various amounts of iron citrate to the original
CNT. Using the Ni peak as an internal standard to obtain a calibration
curve, the calculated Fe content was less than 1.2% wt. The BET specific
2.2. Methods
Powder X-ray diffraction (XRD) analysis was performed using a
X’Pert PRO MPD (Pananalytical) powder diffractometer with Cu Kα
radiation (40 kV, 30 mA). The measurements were conducted from 5 to
80° (2θ) at an average scanning rate of 0.02°. The XRF spectra were
obtained using Skayray EDXRF 3600H spectrometer equipped with the
tungsten lamp, with energy resolution of 135 +/–5 eV. The FTIR
characterizations were carried out by Excalibur 300 Series Digilab
spectrometer, equipped with DTGS detector and using the standard
transmission KBr techniques (ca. 300 mg of KBr and 2 mg of sample,
pellet diameter 2 cm). The spectra were recorded in the range of
400–4000 cm−1 with a resolution of 2 cm−1. Spectra processing was
performed using TotalChrom software package. Nitrogen adsorption/
desorption analysis was carried out at −196 °C using an Autosorb-1
(Quantachrome) surface area/pore size analyzer. Prior to volumetric
sorption measurement, the samples were preheated and degassed under
Fig. 1. Characteristics of the CNT support: A) the
inner diameters estimations (mean values shown for
each range); B) XRF spectrum.
255

A. Kirpsza et al.
AppliedCatalysisA,General549(2018)254–262
Fig. 2. XRD patterns of A) HP2W and B) HPW supported on CNT.
vacuum at 120 °C for 18 h. Scanning electron microscopy (SEM) images
were recorded by means of a JEOL JSM-7500F field-emission apparatus
operated at 15 kV. Raman spectroscopic measurements were performed
by means of a Witec alpha 300 spectrometer equipped with laser:
532 nm, power 0.5 mW. The Raman spectra were recorded with a re-
solution of 5 cm−1. Twenty scans were accumulated to ensure sufficient
signal-to-noise ratio. The fitting of curves and their deconvolution
process for the determination of spectral parameters were performed
with the software program Origin using pseudo-Voight profile.
Microcalorimetric experiments were performed using a Microscal FMC
4110 microcalorimeter, described in detail in [24]. The instrument
measures the rate of heat evolution accompanying a solid−gas inter-
action. A sample of catalyst is placed in a small microcalorimetric cell
(7 mm in diameter, volume of 0.15 cm3), and the measurement is car-
ried out in the flow-through mode. The FMC microcalorimeter is also
equipped with the downstream detector (DSD), currently a thermal
conductivity detector (TCD), to analyze the NH3 content in the carrier.
The uptake of NH3 is, therefore, determined concurrently with the rate
of heat evolution. The microcalorimetric measurements have been
carried out at 150 °C and under ambient pressure. Typically, the ex-
periment begins with the sample immersed in the N2 carrier flowing at
the rate of 3 cm3/min. Thermal equilibration usually takes several
hours. Once the thermal equilibrium is reached the flow of pure N2
carrier through the cell is replaced by the flow of NH3/N2 mixture (3
vol.% NH3; 3 cm3/min). The isopropanol conversion was used as a test
reaction in a flow differential microreactor (Ø = 10 mm). The reactant
was delivered by passing helium through anhydrous isopropyl alcohol
in a saturator kept at the temperature maintaining alcohol pressure of
11.5 kPa. The weight of samples of either catalyst was adjusted to en-
sure that it always contained exactly 0.4 mmole of H+ (0.38 g HPW and
0.29 g HP2W).
The catalysts (Table 1) were mixed with quartz particles
(Ø = 0.19 mm) to produce the catalyst layer of 0.5 cm thickness. Cat-
alytic tests were carried out at the temperature ranging from 60 to
120 °C and at 11.5 kPa of isopropanol pressure. The flow rate of the
mixture of alcohol and helium was 20 cm3/min. Both the up-stream
analysis of reaction mixture as well as the down-stream product de-
tection have been performed by gas chromatograph (Perkin-Elmer
AutoSystem XL) equipped with a column filled with Porapak Q.
For the temperature-dependent experiments the procedure con-
sisted of two stages: First, prior to the catalytic tests, the catalyst
samples were dehydrated in situ (in helium flow 30 cm3 min−1), first at
50 °C for 20 min, and subsequently at 180 °C for 50 min. In order to
check whether this procedure was sufficient for total dehydration, a
selected sample was removed from the reactor following the dehydra-
tion, and quickly transferred to the cell of the moisture analyzer
(MAC50, Radwag, Poland) for TG analysis. The lack of any loss of mass
over heating to 250 °C indicates that our in situ dehydration procedure
was sufficient for the samples of HPA to be totally dehydrated. The
initial dehydration is crucial to avoid the crystallization water hin-
dering the access of the alcohol molecules to protons [9]. In the second
stage of the temperature-dependent experiments, the sample was
cooled down from 180 °C to 60 °C in the flow of pure He carrier. Fol-
lowing a preliminary 10 min period in the flow of reaction mixture at
60 °C, the analysis of products was carried out. Such preliminary
10 min period has been followed at each successive temperature before
the chromatographic analysis of products. For the stability tests, the
same activation procedure has been applied, and the several cycle of
product analysis were repeated at the same temperature of 80 °C, the
first one after the 10 min period in stream of reagent.
3. Results and discussion
3.1. Characterization of the supported catalysts
Fig. 2 compares XRD patterns of the crystalline parent hetero-
polyacids, HP2W (Fig. 2A) and HPW (Fig. 2B), with those of the sup-
ported heteropolyacids, as well as with the pure CNT. The XRD patterns
for both parent HPAs are consistent with those found in the literature
for HPW [25,26], and for HP2W [27,28]. In spite of a certain decrease
of crystallinity in both the supported heteropolyacids, several reflec-
tions from the crystalline HPW and HP2W are still well resolved in the
1.0HPW/CNT (10.2; 14.4; 20.8; 25.5; 34.7; 37.9°) and 1.0HP2W/CNT
(7.7; 8.7; 24.7; 25.3°) materials. Arguably, in both cases the decrease of
crystallinity indicates that the HPA polyoxyanions are distributed on
the CNT surface in mostly amorphous form, with a fraction of small
crystallites. The patterns of the starting CNT consist of three broad re-
flections (26.1, 43.8, 51.1°) in agreement with [29]. The relatively
sharp reflection at 21.9° and a very weak one at 36.4° in the CNT
pattern seems to be due to an impurity of crystalline α-FeO(OH)
(JCPDS-29-0713).
Fig. 3 compares SEM images of pure CNTs (A) with carbon nano-
tubes covered by heteropolyacid (B). The coverage of the nanotubes by
HPA has been confirmed by EDS analysis. In the HP2W covered na-
notubes (Fig. 3C, D), the EDS analysis (atomic percent of phosphorus
and tungsten) has been performed in four points for a selected nanotube
(Fig. 3D). The EDS results are listed in Table 2. The W/P ratio close to
the expected stoichiometric value of 9 in HP2W is evident in the central
part of the nanotube (Spectrum 3 and 4, Fig. 3D). Deviation from the
W/P ratio of 9 may be a result of steric effects in the HP2W heteropoly
anions located within the nanotube entrances.
FT-IR spectroscopy makes it possible to confirm the retention of
256

A. Kirpsza et al.
AppliedCatalysisA,General549(2018)254–262
Fig. 3. SEM images: A) pure CNT (×50000); B) 0.5HP2W/CNT (×50000); C) 1.0HP2W/CNT (×10000), D) enlarged part of image (C) used for EDS analysis.
catalysts, 1.0HPW/CNT and 1.0HP2W/CNT. This seems to suggest that
the coverage of CNT by HPA blocks the slot-shaped pores, thus pre-
venting the capillary condensation inside them to occur [33]. At the
same time, the total capacity for N2 decreases significantly with in-
creasing of the amount HPA used for impregnation, so that in the 0.5
monolayer materials the uptake of N2 (cf. points marked A in both
panels of Fig. 5) is twice as much as that in the 1.0 monolayer materials
(cf. points B Fig. 5). This pore blockage is further confirmed by the
surface area results listed in Table 1. The studied HPAs themselves show
very low surface areas of around 2–4 cm2/g, with no porosity, whereas
the surface area of the CNT support used in this study is 136 m2/g.
Therefore, the contribution from surface area of HPA in the overall
surface area and porosity in the supported materials is negligible.
Table 1 shows that the BET surface area of supported catalysts de-
creases in proportion of the HPA loading.
Fig. 6 presents the plots of pore volume vs. pore size distribution
(PSD) obtained using DFT method. The sharp peak at ca. 35 Å in the
pure CNT graph may be attributed to the slot-like pores, as mentioned
above. However, it can also be a an artificial feature resulting from the
so called Tensile Strength Effect (TSE) [34]. It typically appears in
certain types of mesoporous materials for which the N2 sorption iso-
therms may feature a hysteresis loop closing at 0.45 p/p0, and such loop
is evident from the N2 isotherms on CNT shown in Fig. 5. The other
irregular peaks in the PSD of CNT in Fig. 6 indicate an occurrence a
broad range of pore diameters up to ca. 200 Å. The three of the maxima,
at ca. 52, 80 and 113 Å, are in a well agreement with the three main
ranges of the SEM-estimated inner diameters of CNTs (38–60 Å,
77–92 Å and 110–123 Å, cf. Fig. 1A). The overall pore volume de-
creases on supporting of HPAs on the CNT, likely due to hindering the
entrances to nanotubes by HPA entities. The carbon structure itself
Table 2
EDS analysis of 1.0HP2W/CNT (the selected points are marked in Fig. 3 D).
No of spectrum
P content, au
W content, au
Ratio W/P
1
2
3
4
21
21
10
12
79
79
90
88
3.8
3.8
9.0
7.3
heteropoly anions in the CNT-supported catalysts by detection of the
four well resolved bands typical for HPAs. These are the bands at
1089 cm−1 (νas[P-Oa]), 962 cm−1 (νas[W = Od]), 912 cm−1 (νas[W-Ob-
W]) and a broad band at c.a. 794 cm−1 (νas[W-Oc-W]) for HP2W [30].
Likewise, the HPW spectrum consists of analogous bands at 1080, 980,
890 and 810 cm−1. Fig. 4 shows that the first three of these original
bands are essentially reproduced in the resultant spectra of the sup-
ported HP2W (Fig. 4A) and the supported HPW (Fig. 4B) alike. How-
ever, the broad features at the lower frequencies appear to be shifted
upward, by 30 cm−1 in the supported HP2W and by 20 cm−1 in the
supported HPW spectrum. In both cases, this shift can be attributed to
the interaction of Oc bridging oxygen atoms within CNT [17,31].
Fig. 5 compares the N2 sorption isotherms of the supported catalysts
with those of the pure CNT support. All the curves are type IV isotherms
featuring an H4 hysteresis loop (according to the Brunauer-Deming-
Deming-Teller classification [32]), indicative of the occurrence of ca-
pillary condensation within a certain kind of slot-shaped pores possibly
formed between parallel walls. The hysteresis is most pronounced in the
pure CNT support, and is barely discernible in the full monolayer
257

A. Kirpsza et al.
AppliedCatalysisA,General549(2018)254–262
Fig. 4. FTIR spectra of the following samples in KBr: (A) pristine HP2W and supported HP2W materials; (B) pristine HPW and supported HPW materials. All the spectra were recorded at
room temperature in air.
appears to be not destroyed in the supported materials (as shown by
Raman spectroscopy, cf. next paragraph), in other words, the nanotubes
are still there, it follows that the access to their interior for the N2
molecules may be blocked by the HPA particles plugging their en-
trances. The reason for this may be that CNTs are highly hydrophobic,
and we used absolute ethanol in our supporting procedure as solvent for
HPAs. However, ethanol is still a very polar solvent and so its pene-
tration into the inner spaces of CNTs is likely to be rather limited.
Raman spectroscopy appears to be one of the primary techniques for
studying fundamental properties of carbon nanotubes. Fig. 7 shows
Raman spectra of the CNT-supported Wells-Dawson heteropolyacid
1.0HP2W/CNT in the region of 1800–400 cm−1. The analogous spectra
for pure CNTs and for 0.5HP2W/CNT have been presented in Supple-
mentary materials (Fig. S1). The spectra of both the pure CNT support
as well as of the HP2W/CNT catalysts all exhibit two significant bands
centered at about 1580 cm−1 (G) and 1345 cm−1 (D). Deconvolution of
these profiles yielded the following Voigh’t-shaped bands: G, D1, D2,
D3 and D4 at about 1580, 1345, 1610, ∼1500 and ∼1200 cm−1 re-
spectively, corresponding to various carbon structures. The G band
corresponds to an ideal graphitic lattice vibration mode with E2g sym-
metry. The D group can be attributed to a distorted lattice, namely: D1
to the disordered graphitic lattice (graphene layer edges with A1g
symmetry); D2 to the disordered graphitic lattice (surface graphene
layers with E2g symmetry); D3 to the amorphous carbon; D4 to the
disordered graphitic lattice with A1g symmetry (such as polyenes or
ionic impurities) [35]. The accuracy of deconvolution may be appre-
ciated by inspection of the “fitting” curve resulting from the summation
of bands G and D1-D4, and plotted against the experimental curve in
Fig. 6 (R2 = 0.997 or 0.996). It may be concluded, that the same forms
of carbon appear in both the supported catalyst and in the pure CNT
support. A degree of disorder in the graphitic lattice can be assessed
from the ratio (ID1 + ID2)/(IG + ID1 + ID2), where the I values represent
intensities of the bands G, D1 and D2. The degree of disorder so cal-
culated turned out to be comparable for all samples, namely: 0.645,
0.674 and 0.651 for the CNT support, 0.5, respectively. A fraction of
amorphous carbon can, in turn, be approximated from the ratio of ID3
/
IG, yielding ca. 0.8 for all the materials studied. It can be therefore
concluded, that the deposition of the HP2W acid on CNT does not in-
crease the fraction of disordered carbon in comparison to the pure
phase.
3.2. Acidic properties of catalysts
Catalytic performance of the tested catalysts is a function of several
factors, such as the proton mobility (acid strength), their accessibility to
reagents, and the dispersion of the active phase. The proton mobility
can be estimated by calorimetric measurement of the energy of inter-
action between protons and ammonia molecules. The enthalpy of this
interaction, i.e., the molar heat of the formation of monomeric am-
monium ion NH4+, may be considered as a measure of catalyst' acid
Fig. 5. Nitrogen sorption isotherms.
258

A. Kirpsza et al.
AppliedCatalysisA,General549(2018)254–262
Fig. 6. DFT pore size distribution of HP2W/CNT catalysts (A) and HPW/CNT catalysts (B) with mono- and half- layer of heteropolyacid. The arrows mark mean values of the inner
diameters of CNTs estimated from SEM images (cf. Fig. 1A).
ammonia sorption temperature [36,37]. For H3PW12O40, Highfield and
Moffat [35] reported that temperature as high as 150 °C was necessary
for the NH3/H+ = 1 to be achieved, whereas at room temperature the
NH3/H+ ratio of 1.7 was observed. It has been later found, however,
that the deviation of the NH3/H+ ratio from unity is mainly a result of
+
the formation of an N2H7+ adduct [36,38]. Thus, the monomeric NH4
ion was found to be solely formed at the temperature higher that 100 °C
+
in H4SiW12O40 [36], while the N2H7 formation was revealed at room
+
temperature. Similar temperature pattern of N2H7 formation has also
been confirmed for Wells-Dawson heteropolyacid [37]. Therefore, the
NH3/H+ ratio higher than one suggests that the sorption of ammonia
results partly in the formation of the N2H7+ adduct, rather than leading
+
solely the NH4 ion, and this is clearly the case for the 1.0HP2W/CNT
material showing the ratio of R = 1.2 (cf. Table 3). For the sake of
comparability, therefore, one needs to determine the molar heats of the
+
monomer NH4 formation for the 1.0HP2W/CNT.
+
Doing this, we have to account for the N2H7 formation, and we
note that the NH3 sorption on the 1.0HP2W/CNT can be describe as the
following reactions:
Fig. 7. The Raman spectra 1.0HP2W/CNT catalyst and their fittings with D1, D2, D3, D4
and G components.
2NH3(g) + H+(s) = N2H7(s)
(1)
(2)
(3)
+
+
+
NH3(g) + NH4(s) = N2H7(s)
strength. Table 3 lists the results of microcalorimetric results of am-
monia sorption on selected catalysts containing heteropolyacids (HPW,
HP2W), supported or otherwise. However, the ratio NH3/H+ is a
measure of both the acid strength and the accessibility of protons to
ammonia.
NH3(g) + H+(s) = NH4(s)
+
The enthalpy of the monomer NH4+ formation may be calculated as
the difference between the reactions (1) and (2)
+
In order to use the heat of NH4 formation as measures of protons
ΔH3 = ΔH1 − ΔH2
(4)
acidity, a great care has to be taken to ensure that the one-to-one
proportion of NH3 to H+ is being upheld in the ammonia sorption ex-
periments. In fact, the NH3/H+ ratio can range from values lower than
one, but can also assume a value as high as two, dependent on the
where ΔH1, ΔH2 correspond to the enthalpies of reactions (1) and (2)
respectively.
However, with the NH3/H+ ratio R = 1.19, the actually measured
value of the enthalpy ΔHexp (ΔHexp = −124 kJ/mol NH3; cf. Table 3)
Table 3
The microcalorimetric results of NH3 sorption-desorption over catalysts at 423 K.
Catalyst
(NH3)total μmol NH3/g
(NH3)irrev μmol NH3/g
R NH3/H+
-ΔHexp kJ/mol NH3
HPW
1.0HPW/CNT
HP2W
845
825
1356
1160
805
459
1307
763
0.77
1.04
0.95
1.19
235
151
159
124
141a
1.0HP2W/CNT
a
+
The value of 141 kJ/mol NH3 approximates the enthalpy of the NH4 formation in the 1.0HP2W/CNT, calculated from Eqs. (4) and (5).
259

A. Kirpsza et al.
AppliedCatalysisA,General549(2018)254–262
Fig. 9. Stability tests of catalytic activity in the isopropanol dehydration at 80 °C and the
isopropanol partial pressure of 11.5 kPa.
Fig. 8. Differential heat of sorption (DHS) of NH3 at 150 °C.
may be expressed as a following function of R:
ΔHexp = x ΔH3 + y (ΔH1/2)
amount of heteropolyacid layers should be among the factors affecting
performance in this process. Isopropanol conversion can form either
propylene and water (dehydration) or acetone and hydrogen (dehy-
drogenation). The last reaction takes place at temperatures 300–600 °C.
In the present paper the catalytic tests were carried out below 120 °C
and only the dehydration process occurred.
(5)
where x = (2-R)/R and y = (2R-2)/R. Note, that x + y = 1, and also
+
we assume that the N2H7 adduct is only formed after the entire po-
pulation of protons has been consumed to form NH4+. Thence, for the
limiting case of R = 2 corresponding to exclusive N2H7+ formation, we
have x = 0 and y = 1, which reduces the relation (5) to
ΔH1 = 2 × ΔHexp, satisfying the stoichiometry of reaction (1). For
R = 1, on the other hand, i.e., for the exclusive NH4+ formation, x = 1
and y = 0, yielding ΔH3 = ΔHexp. For the observed R = 1.19 (cf.
Table 3), x = 0.68 and y = 0.32. The enthalpy of adduct formation in
reaction (2) can be found in [23] to be −65 kJ/mol NH3. We have used
the latter value as an estimation of ΔH2, solving the Eqs. (4) and (5) to
obtain ΔH1 = −206 kJ/mol NH3 and ΔH3 = −141 kJ/mol NH3. It can
be therefore concluded, that deposition of HP2W on CNT decreases the
acidity of protons only slightly, as the heat of NH4+ formation changes
from −159 kJ/mol NH3 in HP2W to −141 kJ/mol NH3 1.0HP2W/
CNT. The analogous difference between the pristine and supported
HPW, on the other hand, is much larger, as the heat of NH4+ formation
changes from −235 kJ/mol NH3 to −151 kJ/mol NH3. Thus, the de-
crease in acidity in CNT-deposited HPW is much more profound.
Fig. 8 represents the differential heats of sorption (DHS) for all four
catalysts plotted vs. the coverage by ammonia. These DHS curves are
bell-shaped for all the catalysts, supported and unsupported alike, in-
dicative of a limited accessibility of protons to NH3 in these materials.
The maxima of the heats of sorption achieved by both unsupported
catalysts are much higher compared to the supported ones, which may
be explained by reduced acidity of the latter. However, a striking dif-
ference appears between the positions of their respective maxima. For
the case of the unsupported catalysts, the maxima appear to be shifted
much towards the end of the sorption process, which seems to point out
to very limited accessibility of protons to the NH3 molecules, to the
extent that the HPW catalysts turned out to be only partially saturated
by NH3 (cf. NH3/H+ ratio of R = 0.77 in Table 3). The accessibility of
protons in the CNT-supported heteropolyacids, however, seem to be
much improved, evidenced both by the maxima in the DHS curves
shifted toward the lower NH3 coverage, as well as by the higher values
of NH3/H+ ratio obtained for these materials.
Fig. 9 represents the stability of catalytic performance in the iso-
propanol dehydration for both the one monolayer catalysts, 1.0HPW/
CNT and 1.0HP2W/CNT, measured at 80 °C for three hours. It is clear
that the conversion does not change significantly over the period.
Fig. 10 shows the temperature dependence of the catalytic activity
for all catalysts within the temperature range 60–120 °C. Diisopropyl
ether (DIPE) and propylene (C3) and were the main products. The pure
CNT support showed no activity at this temperature range. Since we
know that the CNT contain small amounts of iron in the form of FeO
(OH), as well as traces of Ni (cf. Fig. 1B in Materials section), the lack of
activity of the pure support suggests that these impurities do not affect
the catalytic activity of the supported HPAs. In the case of HP2W/CNT
catalysts, aside from the C3 and DIPE products, only traces of acetone
were detected at 100–120 °C (not shown).
For all the catalysts, the activity increases with temperature and
stabilizes eventually after reaching ca. 40% conversion of isopropanol
(cf. Fig. 10A and B). The supported catalysts all reach this level at ca.
80 °C and the unsupported HP2W shows only slightly inferior perfor-
mance, reaching the 40% conversion at 90 °C (cf. Fig. 10 B). However,
the unsupported HPW is the odd one out, as it only reaches 40% con-
version at 110 °C. The curves representing DIPE selectivity are bell-
shaped (Fig. 10C and D), all the supported catalysts attaining maximum
at 80 °C, after which the DIPE yields all drop sharply as the C3 for-
mation takes over (Fig. 10C and D). For both the unsupported HPAs, the
maximum of DIPE is shifted toward higher temperatures, respectively
to 100 °C and to 90 °C for HPW and HP2W. Thus the effect of the
support is evident in both HPAs but is most prominent in HPW. This
may be rationalized in terms of lower accessibility of protons in HPW to
the reagent, also hinted by our microcalorimetric results (vide ultra).
The effect of the increase of numbers of monolayers in the supported
catalysts manifests itself most strongly by the significantly higher DIPE
formation on the 1.0HPW/CNT catalyst, compared to the 0.5HPW/CNT
(cf. Fig. 10C). For the Dawson type, the difference in DIPE yields be-
tween the 1.0HP2W/CNT and 0.5HP2W/CNT materials is less pro-
nounced (cf. Fig. 10D). However DIPE yields on the Dawson type cat-
alysts are noticeably higher than those on the Keggin-type. As for the C3
formation, Figs. 10E and F show that the effect of the number on
monolayers is only discernible in the lower temperature region, below
90 °C, for all except the unsupported HPW catalysts. The latter is clearly
3.3. Catalytic activity in isopropanol conversion
The catalytic activity of catalysts containing different amount of
heteropolyacids (0.5 and 1.0 monolayer of HPA) towards the conver-
sion of isopropanol was determined and the authors believed that the
260

A. Kirpsza et al.
AppliedCatalysisA,General549(2018)254–262
Fig. 10. Temperature-dependent tests: isopropanol convesrtion (A, B), the yield of DIPE (C, D) and the yield of propene (E, F) at the isopropanol partial pressure of 11.5 kPa. The solid
lines have been obtained by cubic spline fitting and are only plotted for guding the eye.
inferior, only “catching up” to the 35% C3 yield at ca. 110 °C, while the
other materials all reach this level of C3 already at 90 °C. Again, this is
consistent with the notion of lower accessibility of protons in the bulk
HPW.
Both the catalytic performance and the microcalorimetric results
confirm the crucial role of accessibility of protons in heteropolyacid
supported and unsupported alike. We note, that the performance of the
unsupported HPW in the isopropanol dehydration is consistently in-
ferior (cf. Fig. 10), in spite of its showing the highest enthalpy of the
ammonia sorption (cf. Table 3 and Fig. 8). Therefore, the acidity of
protons seems to be a secondary factor determining the catalytic ac-
tivity of HPAs. Similar observations have been already noticed in the
literature for the dehydration of alcohols [39]. Recently a novel acti-
vation procedure for the unsupported Keggin-type heteropolyacids was
proposed [40] and ascribed by the authors to increasing proton acces-
sibility within the bulk of HPAs. The novelty consists of pre-exposure of
HPA catalyst to alcohol reagent at low temperature, following its de-
hydration at 300 °C, but before heating it up again to desired reaction
261

A. Kirpsza et al.
AppliedCatalysisA,General549(2018)254–262
[7] W. Tureka, J. Haberb, A. Krowiaka, Appl. Surf. Sci. 252 (2005) 823–827.
[8] J. Schnee, A. Eggermont, E.M. Gaigneaux, ACS Catal. 7 (2017) 4011–4017.
[9] H. Lauron-Pernot, Catal. Rev. 48 (2006) 315–361.
temperature. Possibly, such pre-exposure to alcohol may expand the
HPA bulk structure and further facilitate the contact of protons with
reagents. The proposed method may be, therefore, an interesting al-
ternative to that of using supports, both methods effectively increasing
the accessibility of protons in HPA.
[10] Y. Izumi, R. Hasebe, K. Urabe, J. Catal. 84 (1983) 402–409.
[11] M.E. Chimienti, L.R. Pizzio, C.V. Cáceres, M.N. Blanco, Appl. Catal. A: General 208
(2001) 7–19.
[12] M. Misono, Stud. Surface Scie. Catal. 176 (2013) 97–155.
[13] L.E. Briand, G.T. Baronetti, H.J. Thomas, Appl. Catal. A: General 256 (2003) 37–50.
[14] J. Poźniczek, A. Lubańska, A. Micek-Ilnicka, D. Mucha, E. Lalik, A. Bielański, Appl.
Catal. A: Gen. 298 (2006) 217–224.
4. Conclusions
[15] J. Alcañiz-Monge, G. Trautwein, J.P. Marco-Lozar, Appl. Catal. A: General 468
The multiwall carbon nanotubes have been used as support for both
the Keggin- and Dawson-type heteropolyacids as catalysts for the con-
version of isopropanol. In both cases the supported catalysts are sig-
nificantly more active than the unsupported heteropolyacids, with the
Dawson being superior to the Keggin catalysts. The yields of both
products, propylene and diisopropyl ether, are higher on the supported
catalysts compared to those observed on the unsupported materials.
However, the gas flow-through microcalorimetry of NH3 sorption has
shown that the acidity of the supported catalyst was consistently lower
than that of the pristine HPAs. It can be concluded, therefore, that the
observed improved activity in the isopropanol conversion must be re-
lated to the increase of accessibility of protons to isopropanol in the
case of supported catalysts, rather than to the overall acidity of the
catalysts. The increased availability of protons was also confirmed by
the curves of differential heat of sorption of NH3 which showed a clear
shift in the maximum values of heat in the supported acids towards the
lower ammonia coverage (cf. Fig. 8).
(2013) 432–441.
[16] M.N. Timofeeva, M.M. Matrosova, G.N. Il’inich, T.V. Reshetenko, L.B. Avdeeva,
R.I. Kvon, A.L. Chuvilin, A.A. Budneva, E.A. Paukshtis, V.A. Likholobov, Kinet.
Catal. 44 (2003) 778–787.
[17] Wei Qi, Wei Liu, Songyun Liu, Bingsen Zhang, Xianmo Gu, Xiaoling Guo,
Dangsheng Su, Chem. Cat. Chem. 6 (2014) 2613–2620.
[18] A. Micek-Ilnicka, E. Bielańska, L. Lityńska-Dobrzyńska, A. Bielański, Appl. Catal. A:
General 421–422 (2012) 91–98.
[19] M.A. Gomez-Garcı, V. Pitchon, A. Kiennemann, M. Corrias, Ph. Kalck, Ph. Serp,
Topics Catal. 30/31 (2004) 229–233.
[20] J. Poźniczek, A. Micek-Ilnicka, A. Lubańska, A. Bielański, Appl. Catal. A: General
286 (2005) 52–60.
[21] A. Govindaraj, E. Flahaut, Ch. Laurent, A. Peigney, A. Rousset, J. Mater. Res. 14
(1999) 2567–2576.
[22] M.S. Kaba, M.A. Barteru, W.Y. Lee, I.K. Song, Appl. Catal. A: General 194–195
(2000) 129–136.
[23] S.K. Yun, T.J. Pinnavaia, Inorg. Chem. 35 (1996) 6853–6860.
[24] E. Lalik, R. Mirek, J. Rakoczy, A. Groszek, Catal. Today 114 (2006) 242–247.
[25] C. Trolliet, G. Coudurier, J.C. Védrine, Topics Catal. 15 (2001) 73–81.
[26] H.J. Lee, S. Park, J. Chul Jung, I.K. Song, Korean J. Chem. Eng. 28 (7) (2011)
1518–1522.
[27] G. Baronetti, L. Brianda, U. Sedranb, H. Thomas, Appl. Catal. A: General 172 (1998)
265–272.
[28] J. Poźniczek, A. Lubańska, A. Micek-Ilnicka, D. Mucha, A. Lalik, A. Bielański, Appl.
Catal. A: General 298 (2006) 217–224.
Appendix A. Supplementary data
[29] C.N.R. Rao, R. Seshadri, A. Govindaraj, R. Sen, Mater. Sci. Eng R15 (95) (2017)
209–2622.
[30] C. Rocchiccioli-Deltcheff, M. Fournier, R. Franck, R. Thouvenot, Inorg. Chem. 22
(1983) 207–216.
Supplementary data associated with this article can be found, in the
online version, at http://dx.doi.org/10.1016/j.apcata.2017.10.008.
[31] T. Okuhara, N. Mizuno, M. Misono, Adv. Catal. 41 (1996) 137–252.
[32] S. Brunauer, L. Deming, W. Deming, E. Teller, J. Am. Chem. Soc. 62 (1940) 1723.
[33] W.J. Roth, B. Gil, W. Makowski, B. Marszalek, P. Eliasova, Chem. Soc. Rev. 45
(2016) 3400–3438.
References
[1] A. Bielański, J. Datka, B. Gil, A. Małecka-Lubańska, A. Micek-Ilnicka, Phys. Chem.
Chem. Phys. 1 (1999) 2355–2360.
[2] A. Micek-Ilnicka, J. Mol. Catal. A: Chem. 260 (2006) 170–178.
[3] Z. Li, Z. Fang, M.S. Kelley, B.D. Kay, R. Rousseau, Z. Dohnalek, D.A. Dixon, J. Phys.
Chem. C 118 (2014) 4869–4877.
[34] J.C. Groen, L.A.A. Peffer, J. Perez-Ramırez, Microporous Mesoporous Mat. 60
(2003) 1–17.
[35] A. Sadezky, H. Muckenhuber, H. Grothe, R. Niessner, U. Poschl, Carbon 43 (2005).
[36] J.G. Highfield, J.B. Moffat, J. Catal. 88 (1984) 177–187.
[37] E. Lalik, A. Micek-Ilnicka, A.J. Groszek, A. Bielański, Phys. Chem. Chem. Phys. 5
(2003) 3606–3609.
[38] A. Micek-Ilnicka, B. Gil, E. Lalik, J. Mol. Struct. 760 (2005) 25–29.
[39] M.R. Mostafa, A.M. Youssef, S.M. Hassan, Mater. Lett. 12 (1991) 207–213.
[40] J. Schnee, E.M. Gaigneaux, Catal. Sci. Technol. 7 (2017) 817–830.
[4] S. Golay, L. Kiwi-Minsker, R. Doepper, A. Renken, Chem. Eng. Sci. 54 (1999)
3593–3598.
[5] Y.F. Ho, Y.S. Chu, A.N. Ko, React. Kinet. Catal. L 77 (2002) 189–195.
[6] Heesoo Kima, Pil Kim Kwan-Young Leeb, Sung Ho Yeomc, Jongheop Yia, In
Kyu Song, Catal. Today 111 (2006) 361–365.
262