Promoter Effects on Catalyst Selectivity and Stability for Propylene Partial Oxidation to Acrolein

Article abstract of DOI:10.1007/s10562-019-02969-3

Abstract: Highly dispersed silica-supported CuO_x/SiO₂ catalysts were synthesized via solution-phase deposition and studied for their activity, selectivity, and stability in catalyzing the selective oxidation of propylene to acrolein. Strategies for ensuring high metal dispersion included controlling the surface density of silanols (via covalent silanol-capping) or by pre-installing different “promoter” transition metals at submonolayer coverages. A comparison of the effect of first row transition metal promoters showed that V and Cr significantly boost catalyst performance and stabilize CuO_x sites against aggregation. Graphic Abstract: [Figure not available: see fulltext.].

Full text of DOI:10.1007/s10562-019-02969-3

Catalysis Letters
https://doi.org/10.1007/s10562-019-02969-3
Promoter Efects on Catalyst Selectivity and Stability for Propylene
Partial Oxidation to Acrolein
Patricia Anne Ignacio‑de Leon1 · Magali Ferrandon2 · Louisa M. Savereide3 · Scott L. Nauert3 · Jorge Moncada4 ·
Rachel Klet2 · Karena Chapman5 · Massimiliano Delferro2 · Jefrey Camacho‑Bunquin2 · Carlos A. Carrero4 ·
Justin M. Notestein3 · SonBinh Nguyen6
Received: 1 August 2019 / Accepted: 14 September 2019
© Springer Science+Business Media, LLC, part of Springer Nature 2019
Abstract
Highly dispersed silica-supported CuOx/SiO2 catalysts were synthesized via solution-phase deposition and studied for their
activity, selectivity, and stability in catalyzing the selective oxidation of propylene to acrolein. Strategies for ensuring high
metal dispersion included controlling the surface density of silanols (via covalent silanol-capping) or by pre-installing difer-
ent “promoter” transition metals at submonolayer coverages. A comparison of the efect of frst row transition metal promoters
showed that V and Cr signifcantly boost catalyst performance and stabilize CuOx sites against aggregation.
Graphic Abstract
O
Cu
M
O
Si
Cu
O
O
O
Si
O
O
OH
Si
O
O
Si
Si
O
O
Si
O
O
O
O
O
O
O
O
O
O
O
M = V, Cr, Zr
-- lower activity and selectivity
-- facile active site agglomeration
-- higher activity and selectivity
-- suppressed active site agglomeration
Keywords Acrolein · Copper · Oxidation · Promoter · Propylene
1 Introduction
The catalytic partial oxidation of propylene to acrolein
has become one of the most studied reactions in hydro-
carbon oxidation [1–6]. Various catalyst design elements
have been investigated in the design of supported catalysts
for selective partial oxidation of propylene to acrolein.
Specifcally, the efects of active site and oxide dispersion,
Electronic supplementary material The online version of this
article (https://doi.org/10.1007/s10562-019-02969-3) contains
supplementary material, which is available to authorized users.
* Magali Ferrandon
Ferrandon@anl.gov
Extended author information available on the last page of the article
Vol.:(0123456789)
1 3

P. A. Ignacio-de Leon et al.
presence of promoter ions and nature of solid supports
(i.e., acidity and basicity) have been studied in the devel-
opment of catalytic materials that activate C–H bonds and
sequentially mediate oxygen atom transfer to form C–O
bonds [7]. Despite the breadth of catalytic systems devel-
oped and studied for this reaction, gains in performance
and fundamental understanding are still needed [8]. Het-
erogeneous catalysts for propylene-to-acrolein oxidation
include monometallic materials like CuOx–SiO2 and multi-
metallic/multicomponent materials like promoted bismuth
molybdates [9–11]. The relative compositional simplicity
of the CuOx/SiO2 systems renders them more amenable to
structure–activity correlation studies. Recent studies on
CuOx/SiO2-catalyzed propylene oxidation to acrolein sug-
gest that site dispersion is key to achieve high selectivity
to acrolein; Schüth and coworkers [8] reported that highly
diluted CuOx on SiO2 aforded higher selectivity to acr-
olein and reduced combustion to carbon oxides and C1–C2
cracking products. Site dispersion has also been shown
to be important in other transition metal-based oxidation
catalysts; sub-monolayer VOx–SiO2 enhanced propylene
oxidation selectivity to oxygenates, particularly acrolein
[12].
2 Experimental
2.1 Preparation of Trimethylsilyl‑Capped SiO2
Commercially available mesoporous silica support (Selecto
Scientifc, 30–200 μm size, pore size ca. 6 nm, surface area
520 m2/g) was used (Fig. S1a, see Supporting Information).
To 20 g SiO2 in 250 mL hexane in a 500-mL round bottom
fask, 15 mL of hexamethyldisilane (HMDS, Me3Si-SiMe3)
was added with stirring. The mixture was refuxed over-
night, and the solids fltered after allowing to cool to room
temperature. After rinsing with three 200-mL portions of
hexane, the solids were dried in vacuo at 120 °C for com-
plete removal of any unreacted and by-product silanes and
solvent. Thermal gravimetric analyses (TGA) of dried sol-
ids indicated the presence of a trimethylsilyl groups at a
coverage of ca. 0.6 moieties per nm2, thus reducing silanol
surface density to about 1.3 SiOH/nm2 (Fig. S2, Table S1).
The trimethylsilyl-capped SiO2 is labelled TMS-SiO2.
2.2 Preparation of M/TMS‑SiO2 (M = Fe, Co, Cu)
High active site dispersion could be achieved via two
general synthetic strategies. The frst approach involves the
purposeful decrease in SiO2 surface hydroxyl concentra-
tion via thermal treatment or covalent surface modifcation
(e.g., –SiMe3 capping), while the second approach involves
modifying oxide surfaces with cationic ‘anchors’ that sta-
bilize highly dispersed catalytic sites [13–15]. In the lat-
ter approach, the cations can also act as catalyst promoters
[16–19]. The promoter efect was thought to be responsi-
ble for the enhanced oxygenates selectivity of Fe catalysts
[20–22]. Enhancement in CuOx/SiO2 selectivity to acrolein
was observed in the presence of Au promoters; Au was
proposed to stabilize low-valent Cu1+ species which are
more selective to oxygenate formation (e.g., acrolein) [18].
Similarly, the efects of transition metal promoter cations
on the activity and selectivity of supported CuOx systems
for hydrocarbon partial oxidation have been demonstrated
in the use of redox-active V [23–26] and Cr [27–30] sites.
One limitation of prior studies on these promotion efects is
that initial surface densities are typically high, presenting a
range of surface structures and a relatively low number of
the isolated sites proposed to be the most selective.
Metal ions known for their catalytic oxidation activ-
ity (specifically Fe, Co and Cu) were installed
via grafting of molecular precursors on TMS-
SiO2, using molecular bis(hexamethyldisilazido)
iron(II), bis(hexamethyldisilizado)cobalt(II) and
bis(dimethylamino-2-propoxy)copper(II) precursors in
toluene at room temperature for 24 h. Two loadings (0.1
and 1.0 metal atoms/nm2) of Cu, Fe, and Co supported on
TMS-SiO2 were chosen.
2.3 Preparation of M‑SiO2 Supports (M =V, Cr, Zn,
Co, Zr)
SiO2 treated at 200 °C under vacuum overnight was chosen
as support. The 2.5 wt% VOx on SiO2 and 0.7 wt% CrOx
on SiO2 supports were synthesized via incipient wetness
technique following published methods from an aqueous
solution of ammonium metavanadate (NH4VO3) mixed
with oxalic acid (H2C2O4) [31], and a methanolic solution
of chromium(III) nitrate nona-hydrate (Cr(NO3)3·9H2O),
respectively [32]. The 8.3 wt% Zn2+/SiO2 and 9.5 wt%
Co2+/SiO2 supports were prepared by strong electrostatic
attraction (SEA), employing literature methods from aque-
ous solutions at pH 11 of zinc(II) nitrate hexa-hydrate [33]
(Zn(NO3)2·6H2O) and hexamminecobalt(III) chloride [34]
In this paper, we frst investigate how these two synthetic
strategies (TMS-capped SiO2 and promoter-modifed SiO2)
can be employed to increase the dispersion and stabilization
of active sites for propylene oxidation, especially at very low
surface coverages where statistical isolation is expected, at
least for the initial catalyst state. Then we focus on a few
promising promoters which lead to higher yield and higher
stabilisation of active sites.
(Co(NH3)6Cl3), respectively. In addition, 3.9 wt% Zr4+
/
SiO2 support was prepared via solution-phase deposition
of tetrakis(isopropoxy)zirconium(IV). The samples were
1 3

Promoter Efects on Catalyst Selectivity and Stability for Propylene Partial Oxidation to…
calcined at various temperatures (See supporting informa-
tion for detailed synthesis procedures).
a quartz U-tube reactor and calcined in air at 500 °C for
15 min. The samples were then cooled down to 30 °C in
Ar. The H2-TPR profles were obtained by fowing H2/Ar
mixture at 50 mL/min from 30 to 700 °C at 10 °C/min. The
hydrogen consumption was monitored quantitatively using a
thermal conducutivity detector. High energy X-ray total scat-
tering data suitable for pair distribution function (PDF) anal-
ysis were collected at beamline 11-ID-B of the Advanced
Photon Source at Argonne National Laboratory. Data were
collected in 3–5 min exposures using an amorphous-silicon
based area detector at an X-ray wavelength of~0.211 Å. The
images were reduced to one-dimensional difraction patterns
within GSAS-II and PDFs calculated within PDFgetX2 and
XPDFsuite, following subtraction of the background. Dif-
ferential PDFs were calculated by subtracting the PDF meas-
ured for the support from that measured with the catalytic
functionalization.
2.4 Deposition of Copper onto SiO2 and M‑SiO2
Supports via High‑Throughput Synthesis
A protocol was developed for high-throughput, automated
synthesis of catalysts that included support and solution
dispensing, stirring/shaking and washing. Using this meth-
odology, a wide range of copper loadings (nominally 0.05
to 0.25 metal atom/nm2), using bis(dimethylamino-2-pro-
poxy)copper(II) as precursor in toluene, was deposited on
SiO2 pre-treated at 200 °C under vacuum overnight and on
M-SiO2 (M=V, Cr, Zn, Co, Zr) supports. There was no fur-
ther treatment after deposition. Details can be found in SI.
2.5 Catalyst Characterization
A JEOL JEM-2100F transmission electron microscope
(TEM) was used for all bright feld imaging at 200 kV.
Samples were prepared by dispersing solids in ethanol via
sonication for 30 s. Colloidal suspensions were drop cast
onto either Mo grids or lacey C on Ni grids (300 mesh)
from Electron Microscopy Sciences at Argonne National
Laboratory. TEM was used to probe for the presence of
metal clusters on the supports; however, it should be noted
that sample preparation for TEM analyses (sonication in
ethanol) as well as imaging conditions (a 200 kV electron
beam) could change material composition in situ (e.g., metal
agglomeration). Difuse refectance UV–Vis spectra were
collected with a Shimadzu UV-3600 spectrophotometer
equipped with a Harrick Praying Mantis difuse refectance
accessory. Polytetrafuoroethylene was used as the base-
line white standard, and refectance data were transformed
to pseudo-absorbance using the Kubelka–Munk function,
F(R∞). Elemental analysis was performed using a Thermo
iCap7600 ICP-OES instrument. Approximately 30 mg of
sample was digested in 1 mL of HF and subsequently diluted
to 11 mL with a 0.9 wt% HNO3 aqueous solution. Standards
of known concentration were prepared to quantify the sam-
ples’ metal content. Raman spectroscopy studies were per-
formed using a Renishaw InVia Qontor Raman Spectrometer
equipped with 785, 532, and 405 nm solid-state lasers, ×5,
×20, ×50, and ×100 objectives, and a MS 20 Encoded Stage.
All measurements used a 785 nm laser and 1200 L mm−1
grating and were taken with a range of 200–1500 cm−1 using
a 50× long distance objective, 30 s of exposure time, and six
accumulations under ambient conditions. The laser power
used was tuned to get the maximum Raman spectra intensi-
ties without damaging the diferent metal oxide supported
catalysts. H2-TPR studies were carried out in an Altamira
(AMI-100) system. A mixture of 3%H2/N2 at atmospheric
pressure was used. Samples of 25 mg were loaded into
2.6 High‑Throughput Catalyst Testing
High-throughput catalyst testing was carried out to iden-
tify optimum conditions (temperature, propylene to oxidant
ratio) for further testing and to screen a large number of
catalysts. Propylene oxidation to acrolein was performed in
a 16 fxed bed reactor system (Flowrence, Avantium). Typi-
cally, 50 mg of as-prepared catalyst was diluted with 100 mg
nonporous silica gel (Davisil grade 646) and loaded into a
quartz reactor (ID=2 mm, OD=3 mm, L=300 mm). The
silica diluent showed no catalytic activity in control experi-
ments. Reactions were performed between 200 and 500 °C
at nominal atmospheric pressure. All gases were purchased
from Airgas. 5% propylene in Ar was used together with dry
air, N2 (UHP) and He (UHP) as the standard. Propylene:O2
ratios were varied (1:1, 3:1 and 5:1) and total fow rates
varied from 10 to 30 mL/min. Reactions were performed at
a space velocity of 140–180 mol propylene/min mol Cu and
a GHSV of 36,200 mL/gcat/h. The efuent of each reactor
was analyzed sequentially by a gas chromatograph (7890B,
Agilent Technologies), equipped with a thermal conductiv-
ity detector (TCD) and 2 fame ionization detectors (FID).
A simplifed reaction diagram is presented in Scheme 1.
However, methanol, ethanol, propanal, 1- and 2-propanol,
O
O2
O2
O2
CO2 and CO
Scheme 1 Reaction scheme of propylene oxidation to acrolein or
COx
1 3

P. A. Ignacio-de Leon et al.
propionic acid, acetaldehyde and propylene oxide were also
detected in small quantities. Selectivities were calculated
on a C3 basis.
yields of CO2 (i.e., complete oxidation of propylene) with
limited selectivity for acrolein. Propylene conversion is low-
est with the use of CoOx systems, and while appreciable
conversions were also observed for FeOx systems, CuOx was
clearly the best-performing among the three metals chosen
in this study in terms of selectivity to acrolein. In contrast,
FeOx and CoOx systems favored deep oxidation products
such as CO and CO2. Furthermore, our data is consistent
with previous literature reports that lower metal loadings
improve the selectivity to partial oxidation at any given
conversion, most likely due to more dispersed active sites
[8]. While Cu loadings of 0.1 atom/nm2 lead to lower con-
version compared to 1 atom/nm2 at low temperature, it is
worth noting that higher conversion is achieved at 500 °C.
The highest acrolein yield, 4.8% was obtained for 0.1 Cu at
400 °C at 36,000 mL/gcat/h. We note for 1 atom/nm2 Cu
that propylene conversion failed to increase with increasing
temperature (i.e. conversions at 400 and 500 °C) which is
indicative of deactivation via agglomeration of the CuOx
sites. TEM shows signifcant agglomeration of CuOx in these
materials (see Fig. S4), and this will be discussed further
below. DR UV–Vis analyses of Cu- and Fe-based TMS-SiO2
confrm a lesser degree of agglomeration for the samples at
low loadings after reaction at various temperatures (Fig. S5).
It should also be noted any deactivation with TOS is convo-
luted in with all this activity data, which can limit the quan-
titative utility of the rates determined from high-throughput
screening. Nevertheless, this screening does allow for quali-
tative comparisons of large numbers of catalysts. Based on
the higher selectivity seen in these initial results, we nar-
rowed our scope to low loadings of CuOx/SiO2.
2.7 Microreactor Studies
Microreactor experiments were used for kinetic studies and
long-term testing to evaluate catalyst deactivation. Propylene
oxidation to acrolein was performed at atmospheric pres-
sure in a packed-bed quartz microreactor. Typically 50 mg
catalyst, previously calcined at 400 °C for 4 h in air, was
diluted in 100 mg non-porous silica gel to achieve a bed
height of 2–3 mm to minimize pressure and temperature
gradients. The silica diluent showed no catalytic activity in
control experiments. Catalysts were pre-treated at 250 °C
with 33 mL/min of 6% O2/balance He for 30 min to remove
residual water before reaction and cooled to 200 °C. The
temperature was ramped in 25 °C intervals up to 400 °C.
The catalyst was held at each intermediate temperature for
approximately 1 h. Reactions were performed at a constant
space velocity of 14.2 mol propylene/min mol Cu, a total
fow rate of 33 mL/min, and a GHSV of 39,600 mL/gcat/h.
The ratio of propylene:O2 was 1:1. Products were analyzed
using an Agilent GC 6890 with FID and TCD in parallel.
Instantaneous turnover frequencies for each GC injection
are calculated from the diferential region of product yield
divided by contact time, and selectivities were calculated as
shown below:
Acrolein
CO
SC3
=
Acrolein + CO2∕3 +
+ C6-olefin × 2
3
3.2 Efect of CuOx on TMS‑SiO2 versus CuOx on M–
3 Results and Discussion
SiO2 (M = Zr, Zn, V, Cr, Co)
3.1 Selection of Propylene:O2 ratio and Active
Metal in M/TMS‑SiO2 (M = Fe, Co, Cu)
To further enhance the activity of CuOx catalysts and com-
pare the efect of TMS capping with that of doping, we
synthesized CuOx on doped-SiO2 with a variety of promot-
ers. Promoter ions were chosen based on two hypotheses.
Redox-active, oxophilic promoters (V, Cr, Co) may facili-
tate oxidant transfer from the active metal site by weaken-
ing the active metal–oxygen bond and acting as a chemical
promoter. Alternately, promoter ions (Zr, Zn) that are more
acidic than silanols may be better able to anchor the active
CuOx domain, thus improving catalyst stability against
aggregation and other structural changes [35].
Initial experiments were undertaken with CuOx, FeOx, and
CoOx supported on TMS-SiO2 at target metal loadings of
1.0 and 0.1 metal atoms/nm2. The initial catalyst testing was
carried out on a 16-reactor parallel fxed bed reactor system
to identify propylene to oxygen ratio. The ratio of propylene
to O2 was frst investigated using 0.1 and 1.0 Cu atoms/nm2;
results of these experiments indicate that lower substrate-to-
oxidant ratios favor acrolein selectively (Fig. S3). A ratio of
propylene:O2 to 1:1 was then chosen.
Figure 2 summarizes the results of the evaluation of a
series of CuOx/M-SiO2 (0.05 to 0.25 Cu atoms/nm2, M=Zr,
V, Cr, Co, Zn) catalyst candidates for propylene oxidation
to acrolein at 250 °C. With the notable exception of CuOx/
Zr–SiO2, acrolein yields decreased with CuOx loading due to
rapidly decreasing selectivity to acrolein. At a low loading of
0.05 Cu atoms/nm2, the trend in acrolein per-pass yield was:
Figure 1a–c shows a comparison of the catalytic activ-
ity of FeOx, CoOx and CuOx catalysts on TMS-SiO2 at two
diferent loadings measured at temperatures between 200
and 500 °C. The experiments revealed, as expected, increas-
ing conversion of propylene with increasing temperature;
however, increasing temperature generally led to increasing
1 3

Promoter Efects on Catalyst Selectivity and Stability for Propylene Partial Oxidation to…
A
B
50
40
30
20
10
0
50
40
30
20
10
0
50
40
30
20
10
0
50
40
30
20
10
0
dashed
→
dashed
→
0.1 Co
1 Co
0.1 Fe
1 Fe
150
250
350
450
550
150
250
350
450
550
Temperature / °C
Temperature / °C
C
50
50
dashed
→
0.1 Cu
1 Cu
40
30
20
10
0
40
30
20
10
0
150
250
350
450
550
Temperature / °C
Fig. 1 Comparison of propylene conversions and acrolein selectivities of catalytic systems with target loadings of 0.1 and 1 Co (a), Fe (b) or Cu
(c) atoms/nm2 onto TMS-SiO2 supports as a function of temperature (50 mg catalyst, propylene:O2 1:1)
CuOx/Cr–SiO2 >CuOx/V–SiO2 >CuOx/TMS-SiO2 >CuOx/
SiO2 >CuOx/Zn–SiO2 >CuOx/Zr–SiO2 >CuOx/Co–SiO2. It
can be noted that CuOx/TMS-SiO2 gives a higher acrolein
7
yield compared to CuOx on bare SiO2 (vacuum treated at
200C SiO2
200 °C), which indicates that a higher site isolation is favora-
TMS-SiO2
6
ble for the production of acrolein.
Cr-SiO2
5
4
3
2
1
0
Co-SiO2
Zn-SiO2
Zr-SiO2
V-SiO2
Following screening, the two of the highest-performing
promoted catalysts (CuOx/Cr–SiO2 and CuOx/V–SiO2) as
well as CuOx/Zr–SiO2 were selected for more detailed stud-
ies in a microreactor. Loadings of Cu and the promoters are
presented in Table 1.
Activity and selectivity (to acrolein) data collected
from the micro-reactor study are presented in Fig. 3. Data
points represent an average over 60 min time on stream at
a given temperature. Materials prepared with just the pro-
moter ion and no copper (V–, Cr–, or Zr–SiO2) were not as
active as the heterobimetallic catalysts (Table 2, Fig. S6).
Total propylene conversion reached as high as 32% under
these conditions, and conversion rates normalized to the
total amount of catalyst are reported in Fig. 3a. The most
0.05
0.1
0.25
Nominal Cu loading / atoms nm-2
Fig. 2 Acrolein yield per Cu atom as a function of Cu loadings for Cu
onto various supports (SiO2, TMS-SiO2, Cr–, Co–, Zn–, Zr–, and V–
SiO2). (50 mg catalyst, propylene: O2 1:1, 250 °C)
1 3

P. A. Ignacio-de Leon et al.
Table 1 Characteristics of CuOx/M-SiO2 materials selected for
300 °C (Fig. 3a). However, the acrolein yield of CuOx/
Cr–SiO2 is roughly equal to the sum of the acrolein yield
of CuOx/SiO2 and Cr–SiO2 (Table 2), so there is no evi-
dence of synergistic interaction between CrOx and CuOx.
Alternatively, although CuOx/Zr–SiO2 had higher propylene
conversion than CuOx–SiO2, the yield of acrolein is less than
that of the un-promoted copper catalyst. The addition of Zr
lowers the selectivity of the catalyst to acrolein by dispro-
portionately increasing the rate of formation of CO2. Finally,
CuOx/V–SiO2 maintained its selectivity to the desired acr-
olein product as temperatures increased, and above 300 °C,
it had the highest selectivity, (Fig. 3b, d), and the highest
rate of acrolein production. Moreover, the acrolein yield
was higher than that of CuOx–SiO2 and V–SiO2 combined,
suggesting that synergistic efect exists between Cu and V.
The acrolein yield obtained at 400 °C for CuOx/V–SiO2
was 3.4%, which is higher compared to that obtained from
other Cu/V systems [24, 25]. When comparing with precious
metal-free Cu-based catalyst [36], it is also in the same range
as the one reported by Tüyüz et al. 3.7% [8] but they used a
smaller GHSV (15,000 mL/gcat/h at the same O2:C=1:1),
while we use almost 40,000 mL/gcat/h.
microreactor studies
Catalyst
M (wt%) Cu (wt%) Cu (μmol/gcat) Cu (atoms/nm2)
CuOx/SiO2
–
3.40
0.45
0.47
70
75
0.085
0.090
CuOx/Zr–
SiO2
CuOx/Cr–
SiO2
0.70
2.37
0.45
0.35
69
55
0.082
0.066
CuOx/V–
SiO2
prevalent products were CO and CO2, not shown here.
There were no signifcant side products formed, although
dimerization products were observed in small quantities at
low temperatures. The catalysts displayed linear Arrhenius
behavior under diferential conversion conditions (325 °C
and below) (Fig. 3c). Acrolein yield and apparent activa-
tion energies are reported in Table 2.
Of the materials studied, CuOx/Cr–SiO2 displayed the
highest overall propylene conversion at temperatures above
A
B
10-3
CuO /200°C-SiO
CuO /200°C-SiO
12
x
2
x
2
80
60
40
20
CuO /Cr-SiO
CuO /Cr-SiO
x
2
x
2
10
8
CuO /V-SiO
CuO /V-SiO
x
2
x
2
CuO /Zr-SiO
CuO /Zr-SiO
x
2
x
2
6
4
2
200
250
300
350
400
200
250
300
350
400
Temperature(°C)
Temperature(°C)
C
D
CuO /200°C-SiO
0
-2
-4
-6
-8
x
2
80
60
40
20
CuO /Cr-SiO
x
2
CuO /V-SiO
x
2
CuO /Zr-SiO
x
2
CuOx/200°C-SiO2
CuOx/Cr-SiO2
CuOx/V-SiO2
CuOx/Zr-SiO2
5
10
15
20
25
30
1.6
1.8
2
2.2
-1
Propylene Conversion (%)
1000/T (K )
Fig. 3 Performance comparison of the promoted and unpromoted
catalysts listed in Table 1. a Rate of propylene conversion normalized
to catalyst mass at increasing temperatures, b Selectivity to acrolein
on a C3 basis at increasing temperatures, c Arrhenius plot of rate of
propylene conversion of copper supported on modifed and unmodi-
fed silica, d Selectivity to acrolein at increasing extents of propylene
conversion
1 3

Promoter Efects on Catalyst Selectivity and Stability for Propylene Partial Oxidation to…
Table 2 Rates of acrolein production and apparent activation energy
300 oC (Fig. 4a). However, the yield of acrolein remained
for M–SiO2 and CuOx/M–SiO2 materials (M=Cr, V and Zr)
Acrolein yield
nearly constant with time on stream (Fig. 4b). This fnd-
ing suggests that at high temperatures, the rearrangement
of metal oxides on the silica surface results in a decrease in
the number of sites that are active for combustion while sites
that are responsible for selective oxidation are preserved or
possibly even generated.
(µmolAcr
mgcats×103)
A,app (kJ/mol) 200 °C 300 °C
/
Catalyst
E
400 °C
CuOx/SiO2
98
1.1
0.19
<0.01 90
23
33
17
82
56
190
660
38
120
240
CuOx/Zr–SiO2 137
CuOx/Cr–SiO2 86
CuOx/V–SiO2 90
3.3 CuOx/M‑SiO2 Characterization Pre‑
and Post‑Catalysis (M = Zr, V, Cr)
120
Zr–SiO2
Cr–SiO2
V–SiO2
92
47
102
<0.01
9.4
2
3
55
27
The Raman spectra of the as-prepared catalysts (0.05 CuOx/
SiO2, Cr–SiO2, V–SiO2, Zr–SiO2, 0.05 CuOx/Cr–SiO2, 0.05
CuOx/V–SiO2 and 0.05 CuOx/Zr–SiO2) as well as bulk
CuO, monoclinic ZrO2, SiO2 and V2O5/SiO2 are presented
in Fig. 5. In agreement with previous reports, features at
296 and 345 and 660 cm−1 are observed for bulk crystalline
CuO reference sample [37]. In the case of CuOx/SiO2, CuOx/
Cr–SiO2, CuOx/V–SiO2 and CuOx/Zr–SiO2, there is a vibra-
tion around 976 cm−1 that is not observed in the bulk phase
Micro-reactor studies were also used to provide more
insight into the deactivation of the high-performing cata-
lysts. All of the materials tested showed a gradual deactiva-
tion and decrease in conversion (Fig. S7) and in the forma-
tion of CO2 with time on stream for isotherms at or above
Fig. 4 Comparison for catalytic
systems with nominal loading
of 0.05 Cu atoms/nm2 onto
various supports of a CO2 and
b acrolein yield with time on
stream. Propylene conversion is
shown in Fig. S7
1
0.8
0.6
0.4
0.2
0
A
CuOx/200 °C-SiO2
CuOx/Cr-SiO2
CuOx/V-SiO2
CuOx/Zr-SiO2
350 °C
300 °C
325 °C
375 °C
400 °C
8
9
10
11
12
13
14
15
Time on Stream (h)
B
0.04
0.035
0.03
0.025
0.02
0.015
0.01
0.005
0
CuOx/200°C-SiO2
CuOx/Cr-SiO2
CuOx/V-SiO2
CuOx/Zr-SiO2
300 °C
325 °C
350 °C
375 °C
400 °C
8
9
10
11
12
13
14
15
Time on Stream (h)
1 3

P. A. Ignacio-de Leon et al.
Fig. 5 Raman spectra of as-pre-
pared Cr–SiO2, 0.05 CuOx/SiO2
and 0.05 CuOx/Cr–SiO2 a, V–
SiO2, 0.05 CuOx/SiO2 and 0.05
CuOx/V–SiO2 (b) and Zr–SiO2,
0.05 CuOx/SiO2 and 0.05 CuOx/
Zr–SiO2 (c). Reference spectra
of SiO2, ZrO2, CuO, and V2O5/
SiO2 added for comparison
B
A
488
488
1034
CuOx/V-SiO2
CuOx/Cr-SiO2
CuOx/SiO2
689
607
982
CuOx/SiO2
1034
V-SiO2
Cr-SiO2
CuO
975
CuO
SiO2
V/SiO2-bulk
SiO2
200
400
600
800
1000
1200
1400 200
400
600
800
1000
1200
1400
Raman shift (cm-1)
Raman shift (cm-1)
C
488
976
CuOx/Zr-SiO2
CuOx/SiO2
Zr-SiO2
CuO
ZrO2
SiO2
200
400
600
800
1000
1200
1400
Raman shift (cm-1)
of CuO, but is due to silica aerogel [38, 39]. For Cr–SiO2,
the Raman spectrum is dominated by one strong, broad
band at 890 cm−1 and shoulders around 982, and 689 cm−1.
The band at 890 cm−1 is typical for dichromate [40]. The
Raman band at around 982 cm−1 is characteristic of a dehy-
drated chromium oxide species [41, 42]. However, previous
Raman studies concluded that this chromium oxide specie
is predominantly present as a polychromate specie [40]. An
interesting observation was that the peak around 689 cm−1
observed in the binary and ternary chromium systems is not
observed in the literature [40, 42–44] for reported Cr–SiO2
catalysts. However, no isolated surface chromate species
were observed (Cr=O vibrations at 1030 cm−1). On CuOx/
Cr-SiO2, the large peak at 890 cm−1 disappears and the
vibration at 689 cm−1 dampens. The presence of polychro-
mate species in CuOx/Cr–SiO2 may explain the high conver-
sion of propylene into CO2, as seen in Fig. 4a. For vanadium
based catalysts a vibration was found at 1034 cm−1 char-
acteristic of molecularly dispersed 2D V=O [44, 45]. The
absence of 995 cm−1 vibration (3D V=O) [45] in compari-
son with bulk V2O5 indicates the V–SiO2 is well-dispersed.
Since the partial oxidation reaction follows a redox mech-
anism, the reducibility of the samples, was investigated by
the temperature-programmed reduction (TPR) technique.
The TPR profle of the 0.05 CuOx/SiO2, Cr–SiO2, 0.05
CuOx/Cr–SiO2, V–SiO2, 0.05 CuOx/V–SiO2, Zr–SiO2 and
0.05 CuOx/Zr–SiO2 after treatment in air at 500 °C are
shown in Fig. S8. There is a small peak at 252 °C for CuOx/
SiO2 which corresponds to highly dispersed CuO and a
broad peak up to 630 °C which corresponds to dispersed
Cu (II) species onto silica which is hard to reduce [46, 47].
There is one main sharp peak for Cr–SiO2 and for V–SiO2
while for Zr–SiO2, there is one broad peak. For V–SiO2,
the peak reduction temperature is at 568 °C, which does
correspond to either isolated or oligomeric VOx species in
p-tetrahedral and square-pyramidal coordination [48]. The
peak at 463 °C for Cr–SiO2 corresponds to the reduction
of Cr(VI) to Cr2O3 [49]. The temperature for reduction of
these promoter oxides are lowered in the presence of copper
indicating close interactions, as reported for various load-
ings of Cu onto V-HMS [50]. In addition, for CuOx/V–SiO2,
the copper species reduce at higher temperature indicating a
stronger interaction of Cu with vanadium than Cu with the
1 3

Promoter Efects on Catalyst Selectivity and Stability for Propylene Partial Oxidation to…
SiO2 support. This may explain the enhanced stability of
copper species when supported on V–SiO2. In addition the
hydrogen consumption is enhanced for copper species on
V–SiO2, which also indicates more lattice oxygen is avail-
able which is responsible for the enhanced reactivity of allyl
oxidation [47].
3.5
2.1
1.7
5.9
4.0
5.3
4.3
2.7
2.4
1.4
Analysis of the spent CuOx/SiO2 catalysts (from micro-
reactor studies) via DR UV–Vis provided a rationale for
the diferences in the efects of various promoters (Fig. 6).
The spent CuOx/SiO2 catalyst displays increased absorp-
tion>350 nm, consistent with aggregation to form extended
CuO sheets, and>550 nm indicative of aggregation to form
larger CuO crystallites [51, 52], as compared to the spectra
after preliminary calcination. The CuOx/Cr–SiO2 and CuOx/
Calcined Cu/Cr/SiO2
Calcined Cu/V/SiO2
Calcined Cu/Zr/SiO2
Zr–SiO2 show slightly increased absorption between 350
0
1
2
3
5
6
7
r / Å 4
and 550 nm consistent with some growth of CuO sheets. In
contrast, CuOx/V–SiO2 catalysts do not display increased
absorption>350 nm which indicates the discrete VOx phases
Fig. 7 Diferential pair distribution data for the calcined samples
particularly stabilize a highly dispersed CuO phase [53]. It
should be acknowledged that CrOx and VOx both contribute
to the total absorption spectra in addition to CuOx, which
may obscure precisely which domains are changing under
reaction conditions.
(SiO2 data was subtracted.)
be attributed to deformation of the strained CuO structure
indicating highly dispersed CuO phase onto support [54].
The main peak at 3.5 Å can also be attributed to Cu–O–M
for instance in Cu2[μ-(η2:η2-O2)], in agreement with a pre-
vious report by Vilella and co-workers on the stability of
oxo-bridged Cu species supported on zeolites [55]. The cal-
cined CuOx/V–SiO2 has short range ordered features, while
the CuOx/Cr–SiO2 and CuOx/Zr–SiO2 show well defned
peaks to longer distance, which supports the better stability
of CuOx/V–SiO2 against formation of aggregates.
PDF characterization of the local structure of calcined
(400 °C) CuOx/M–SiO2 systems was carried out in order to
elucidate structural factors responsible for the stability of
CuOx. Positions of maxima correspond directly to intera-
tomic distances; while areas are related to coordination
numbers. Figure 7 shows the data for the calcined samples.
Some of the interatomic distances do not correspond to CuO,
Cu2O or Cu0, for instance 2.7, 4.0, 5.3 and 5.9 Å but could
The stabilizing efect of a promoter can also be confrmed
visually via TEM. Figure 8 compares TEM images of post-
catalysis CuOx/SiO2 and CuOx/V–SiO2 (Fig. 8b, c). On
CuOx/SiO2 there is evident formation of crystalline CuOx
phases larger than 10 nm in diameter. TEM images of as-
prepared V–SiO2 (Fig. 8a) showed the presence of clusters
of less than 5 nm. Raman (Fig. 5) confrmed these clus-
ters to be 2D V=O. Post-catalysis CuOx/V–SiO2 maintains
these VOx islands and there is no evidence of CuOx or VOx
aggregating further, which is also in agreement with the DR
UV–Vis (Fig. 6). This is consistent with microreactor stud-
ies in which the CuOx/V–SiO2 undergoes less deactivation
over the course of reaction than CuOx/Cr–SiO2 and CuOx/
Zr–SiO2 or (non-promoted) CuOx/SiO2 catalysts (see Fig.
S7). On the other hand, the Zr–SiO2 likely does not stabi-
lize the CuOx phase well due to its poor dispersion leaving
large areas of SiO2 surface without ZrOx to stabilize the
CuOx phase (Fig. S11), which is consistent with increased
absorption>350 nm (Fig. 6) indicative of sintering of the
CuOx phase during reaction on the CuOx/Zr–SiO2 catalyst.
It is possible that the increased absorption in the tail is due
Fig. 6 Absorption spectra of 0.05 CuOx/SiO2 and CuOx/M–
SiO2 (M=V, Cr, and Zr) catalysts. M–SiO2 (black), freshly cal-
cined (400 °C 4 h air) 0.05 CuOx/M-SiO2 (blue) and after micro-
reactor studies (red). Spectra are normalized to freshly calcined
CuOx/M-SiO2. LMCT maximum for comparison
1 3

P. A. Ignacio-de Leon et al.
strategies, which included (i) the decrease in SiO2 surface
hydroxyl concentration via covalent surface modifcation
(silanol capping), or (ii) the modifcation of the oxide sur-
faces with cationic ‘anchors’ that stabilize highly dispersed
catalytic sites. In a frst part, a series of highly dispersed
metal (Fe3+, Co2+, Cu2+) sites on trimethylsilyl-capped SiO2
via solution-phase synthesis was tested. Evaluation of cata-
lyst candidates reveals minimal activity of CoOx/SiO2 cata-
lysts while FeOx/SiO2 systems produced more of undesirable
deep oxidation product, CO2. Appreciable propylene conver-
sion and selectivity for acrolein were observed for CuOx/
SiO2 systems which, however, sufered from rapid deactiva-
tion due to CuOx agglomeration under reaction conditions.
The second part included synthesizing a series of highly
dispersed copper of various surface densities, onto modifed-
SiO2 with promoters (V, Cr, Zn, Co, Zr). Several promising
promoters, V and Cr, were identifed and subjected to more
rigorous activity and durability testing. CuOx/V–SiO2 was
ultimately the catalyst with the highest acrolein yield. Across
a range of temperatures, the yield was much higher than the
combined yield of CuOx/SiO2 and control VOx/SiO2 cata-
lysts, suggesting a synergistic efect between the two metals
that was not observed for any other promoter. Although all of
the materials tested showed a gradual decrease in total pro-
pylene conversion at high temperatures, the majority of the
observed deactivation came from a decrease in COx produc-
tion. Acrolein selectivity remained nearly constant and even
improved for both CuOx/V–SiO2 and CuOx/Cr–SiO2, which
implies that at high temperatures, the rearrangement of metal
oxides on the silica surface results in a decrease in the num-
ber of sites that are active for combustion while the sites that
are responsible for selective oxidation are preserved. Finally,
it was seen via Raman, TPR-H2, DR UV–Vis, PDF and TEM
that one potential reason for the high activity and stability
of CuOx/V–SiO2 is that the addition of V, which formed 2D
islands, appears to help minimize the aggregation of CuOx
into large clusters before, during calcination and under reac-
tion conditions.
A
B
C
AggregatedCu
sites
VOx islands
maintained
Fig. 8 TEM images of as-prepared V–SiO2 (a), post-catalysis of
CuOx/SiO2 (b) and CuOx/V–SiO2 (c)
to sintering of CuOx that sits directly on the SiO2 support
rather than ZrO2 crystallites.
Several literature reports on selective oxidation also
report CuOx promotion and stabilization by VOx or CrOx.
Yang and co-workers reported synergistic efects between
copper and vanadium, attributing the increase in catalytic
epoxidation of propylene by oxygen to enhanced dispersion
of copper by vanadium [23]. In a separate work, Bøyesen
and others also identifed a synergistic Cu/V redox pair
prepared via co-deposition into AlPO-5 with selective for-
mation of acrolein from propylene oxidation as a result of
extralattice oxygens generated by reactive Cu/V pairs linked
by bridging oxygens (Cu–O–V) [24–26]. Dekker and co-
workers also reported the CO oxidation by O2 and N2O using
Cu–Cr/Al2O3 by co-impregnation from an aqueous solution
of Cu(II) and Cr(III) nitrate, with Cr thought to stabilize Cu
against sintering by limiting Cu reduction via the formation
of a CuCr2O4 phase [29]. This promoter efect of Cr on Cu
via a CuCr2O4 spinel was previously reported by Khanma-
medov’s group in their studies of Al–Cu–Cr catalysts for
H2S and CO oxidation [30].
Acknowledgements This work is supported by the Department
of Energy, Laboratory Directed Research and Development fund-
ing at Argonne National Laboratory under Contract No. DE-AC02-
06CH11357. This research used resources of the Advanced Photon
Source and the Center for Nanoscale Materials, U.S. Department of
Energy (DOE) Ofce of Science User Facilities operated for the DOE
Ofce of Science by Argonne National Laboratory under Contract No.
DE-AC02-06CH11357.
4 Conclusions
References
Highly dispersed silica-supported catalysts were studied
for their activity, selectivity, and stability in catalyzing the
selective oxidation of propylene to acrolein. High active
site dispersion, could be achieved via two general synthetic
1. Belin S, Bracey CL, Briois V, Elllis PR, Hutchings GJ, Hyde TI,
Sankar G (2013) Catal Sci Technol 3:2944–2957
2. Baussart H, Delobel R, Le Bras M, Leroy JM (1979) J Chem Soc,
Faraday Trans 1(75):1337–1345
1 3

Promoter Efects on Catalyst Selectivity and Stability for Propylene Partial Oxidation to…
31. Keranen J, Guimon C, Liskola E, Auroux A, Niinisto L (2003) J
3. Liu CH, Lai NC, Lee JF, Chen CS, Yang CM (2014) J Catal
316:231–239
Phys Chem B 107(39):10773–10784
32. Liu YM, Feng WL, Wang LC, Cao Y, Dai WL, He HY, Fan KN
(2006) Catal Lett 106(3–4):145–152
4. Imachi M, Kuczkowski RL, Groves JT, Cant NW (1983) J Catal
82:355–364
33. Schweitzer NM, Hu B, Das U, Kim H, Greeley J, Curtiss LA, Stair
PC, Miller JT, Hock AS (2014) ACS Catal 4:1091–1098
34. Hu B, Getsoian A, Schweitzer NM, Das U, Kim H, Niklas J, Pol-
uektov O, Curtiss LA, Stair PC, Miller JT, Hock AS (2015) J Catal
322:24–37
5. Choi HS, Lin JT, Kuczkowski RL (1986) J Catal 99:72–78
6. Horváth B, Soták T, Hronec M (2011) Appl Catal A 405:18–24
7. Albonetti S, Cavani F, Trifrò F (1996) Catal Rev 38(4):413–438
8. Tüysüz H, Galilea JL, Schüth F (2009) Catal Lett 131:49–53
9. Schuh K, Kleist W, Høj M, Trouillet V, Beato P, Jensen AD, Grun-
waldt JD (2015) Catalysts 5(3):1554–1573
35. Camacho-Bunquin J, Ferrandon M, Sohn H, Yang D, Liu C,
Ignacio-de Leon PA, Perras FA, Pruski M, Stair PC, Delferro M
(2018) J Am Chem Soc 140:3940–3951
10. Zhai Z, Wütschert M, Licht RB, Bell AT (2016) Catal Today
261:146–153
36. Song W, Perez Ferrandez DM, van Haandel L, Liu P, Nijhuis TA
(2015) hensen EJM. ACS Catal 5:1100–1111
11. Pudar S, Oxgaard J, Chenoweth K, van Duin ACT, Goddard WA
(2007) J Phys Chem C 111:16405–16415
37. Deng Y, Handoko AD, Du Y, Xi S, Yeo BS (2016) ACS Catal
6:2473–2481
12. Fierro JLG, Gambaro LA, Cooper TA, Kremenić G (1983) Appl
Catal 6:363–378
38. Lgarashi K, Tajiri K, Tai Y, Tanemura S (1993) Suppl Z Phys
D26:S207
13. Panyad S, Jongpatiwut S, Sreethawong T, Rirksomboon T, Osu-
wan S (2011) Catal Today 174:59–64
39. Owens L, Tillotson TM, Hair IM (1995) J Non-Cryst Sol
186:177–183
14. Labaki M, Lamonier JF, Sifert S, Zhilinskaya EA, Aboukaïs A
(2003) Coll Surf A 227:63–75
40. Jehng JM, Wachs IE, Weckuysen BM, Schoonheydt RA (1995) J
Chem Soc Faraday Trans 91(5):953–961
15. Inui T, Ueda T, Suehiro M (1980) J Catal 65:166–173
16. Grasselli RK (2002) Top Catal 21(1–3):79–88
17. Bettahar MM, Costentin G, Savary L, Lavalley JC (1996) Appl
Catal A 145:1–48
41. Kim DS, Tatibouet JM, Wachs IE (1992) J Catal 136:209–221
42. Weckhuysen BM, Wachs IE (1996)
100:14437–14442
J Phys Chem
18. Bracey CL, Carley AF, Edwards JK, Ellis PR, Hutchings GJ
(2011) Catal Sci Technol 1:76–85
43. Strunk J, Baňares MA, Wachs IE (2017) Top Catal 60:1577–1617
44. Lee EL, Wachs IE (2007) J Phys Chem C 111:14410–14425
45. Xie S, Iglesia E, Bell AT (2001) J Phys Chem B 105:5144–5152
46. Niu X, Zhao T, Yuan F, Zhu Y (2015) Sci Rep 5:9153
47. Chu H, Yang L, Zhang Q, Wang Y (2006) J Catal 241:225–228
48. Arena F, Fruteri F, Parmaliana A (1999) Catal Lett 60:59–63
49. Bugrova TA, Litvyakova NN, Mamontov GV (2015) Kinet Catal
56(6):746–752
19. Wang X, Zhang Q, Guo Q, Lou Y, Yang L, Wang Y (2004) Chem
Commun 1396:1397. https://doi.org/10.1039/B402839B
20. Krenzke LD, Keulks GW (1980) J Catal 61:316–325
21. Forzatti P, Villa PL (1982) J Catal 76:188–207
22. Allen M, Betteley R, Bowker M, Hutchings GJ (1991) Catal
Today 9:97–104
23. Yang L, He J, Zhang Q, Wang Y (2010) J Catal 276:76–84
24. Bøyesen KL, Kristiansen T, Mathisen K (2014) Phys Chem Chem
Phys 16:20451–20463
50. Hu L, Yue B, Chen X, He H (2014) Catal Commun 43:179–183
51. Nauert SL, Schax F, Limberg C, Notestein JM (2016) J Catal
341:180–190
25. Bøyesen KL, Kristiansen T, Mathisen K (2015) Catal Today
254:21–28
52. Barton DG, Shtein M, Wilson RD, Soled SL, Iglesia E (1999) J
Phys Chem B 103:630–640
26. Bøyesen KL, Mathisen K (2014) Catal Today 229:14–22
27. Wojciechowska M, Haber J, Łomnicki S, Stoch J (1999) J Mol
Catal A 141:155–170
53. Gao X, Wachs IE (2000) J Phys Chem B 104:261–1268
54. Pakharukova VP, Moroz EM, Zyuzin DA, Ishchenko AV, Dolgikh
LY, Strizhak PE (2015) J Phys Chem C 119:28828–28835
55. Vilella L, Studt F (2016) Eur J Inorg Chem 2016:1514–1520
28. Xanthopoulou G, Vekinis G (1998) Appl Catal B 19:37–44
29. Dekker NJJ, Hoorn JAA, Stegenga S, Kapteijn F, Moulijn FA
(1992) Am Inst Chem Eng J 38(3):385–396
30. Khanmamedov TK, Kalinkin AV, Kundo NN, Novopashina VM
(1988) React Kinet Catal Lett 37(1):83–88
Publisher’s Note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional afliations.
Afliations
Patricia Anne Ignacio‑de Leon1 · Magali Ferrandon2 · Louisa M. Savereide3 · Scott L. Nauert3 · Jorge Moncada4 ·
Rachel Klet2 · Karena Chapman5 · Massimiliano Delferro2 · Jefrey Camacho‑Bunquin2 · Carlos A. Carrero4 ·
Justin M. Notestein3 · SonBinh Nguyen6
4
Justin M. Notestein
Department of Chemical Engineering, Auburn University,
Auburn, AL 36849, USA
j-notestein@northwestern.edu
5
6
X-ray Sciences Division, Advanced Photon Source, Argonne
National Laboratory, Argonne, IL 60439, USA
1
2
3
Advanced Materials Division, Argonne National Laboratory,
Argonne, IL 60439, USA
Department of Chemistry, Northwestern University,
Evanston, IL 60208, USA
Chemical Sciences and Engineering Division, Argonne
National Laboratory, Argonne, IL 60439, USA
Department of Chemical and Biological Engineering,
Northwestern University, Evanston, IL 60208, USA
1 3