Tailoring the cooperative acid-base effects in silica-supported amine catalysts: Applications in the continuous gas-phase self-condensation of n-butanal

Article abstract of DOI:10.1002/cctc.201301087

A highly efficient solid-base organocatalyst for the gas-phase aldol self-condensation of n-butanal to 2-ethylhexenal was developed by grafting site-isolated amines on tailored silica surfaces. The catalytic activity depends largely on the nature of amine species, the surface concentration of amine and silanol groups, and the spatial separation between the silanol and amine groups. In situ FTIR measurements demonstrated that the formation of nucleophilic enamines leads to the enhanced catalytic activity of secondary amine catalysts, whereas the formation of imines (stable up to 473 K) leads to the low activity observed for silica-supported primary amines. Blocking the silanol groups on the silica support by silylation or cofeeding water into the reaction stream drastically decreased the reaction rates, demonstrating that weaker acidic silanol groups participate cooperatively with the amine groups to catalyze the condensation reaction. This work demonstrates that the spatial separation of the weakly acidic silanols and amines can be tuned by the controlled dehydration of the supporting silica and by varying the linker length of the amine organosilane precursor used to graft the amine to the support surface. A mechanism for aldol condensation was proposed and then analyzed by DFT calculations. DFT analysis of the reaction pathway suggested that the rate-limiting step in aldol condensation is carbon-carbon bond formation, which is consistent with the observed kinetics. The calculated apparent activation barrier agrees reasonably with that measured experimentally. Secondary amines come first: A solid-base organocatalyst achieved by grafting amines onto silica surfaces is applied to the gas-phase aldol self-condensation of n-butanal to 2-ethylhexenal. Silica-supported secondary amine catalysts demonstrate a much higher catalytic activity than the primary amine analogues, owing to the respective formation of enamines as shown by in situ FTIR analysis. The reaction pathway is analyzed by DFT calculations.

Full text of DOI:10.1002/cctc.201301087

CHEMCATCHEM
FULL PAPERS
DOI: 10.1002/cctc.201301087
Tailoring the Cooperative Acid–Base Effects in Silica-
Supported Amine Catalysts: Applications in the
Continuous Gas-Phase Self-Condensation of n-Butanal
[
a]
[a]
[a]
[a]
Sankaranarayanapillai Shylesh, David Hanna, Joseph Gomes, Siddarth Krishna,
[b]
[b]
[a]
Christian G. Canlas, Martin Head-Gordon, and Alexis T. Bell*
A highly efficient solid-base organocatalyst for the gas-phase
aldol self-condensation of n-butanal to 2-ethylhexenal was de-
veloped by grafting site-isolated amines on tailored silica surfa-
ces. The catalytic activity depends largely on the nature of
amine species, the surface concentration of amine and silanol
groups, and the spatial separation between the silanol and
amine groups. In situ FTIR measurements demonstrated that
the formation of nucleophilic enamines leads to the enhanced
catalytic activity of secondary amine catalysts, whereas the for-
mation of imines (stable up to 473 K) leads to the low activity
observed for silica-supported primary amines. Blocking the sila-
nol groups on the silica support by silylation or cofeeding
water into the reaction stream drastically decreased the reac-
tion rates, demonstrating that weaker acidic silanol groups par-
ticipate cooperatively with the amine groups to catalyze the
condensation reaction. This work demonstrates that the spatial
separation of the weakly acidic silanols and amines can be
tuned by the controlled dehydration of the supporting silica
and by varying the linker length of the amine organosilane
precursor used to graft the amine to the support surface. A
mechanism for aldol condensation was proposed and then an-
alyzed by DFT calculations. DFT analysis of the reaction path-
way suggested that the rate-limiting step in aldol condensa-
tion is carbonꢁcarbon bond formation, which is consistent
with the observed kinetics. The calculated apparent activation
barrier agrees reasonably with that measured experimentally.
Introduction
Enzymes accelerate chemical reactions by the cooperative in-
teraction between two or more functional groups accurately
various carbonꢁcarbon bond formation such as the Knoevena-
[3,5]
gel, Henry, and aldol condensation reactions.
As organic
[
1]
positioned within their active sites. For example, serine pro-
amines (base sites) supported on silica surfaces (acid sites) can
activate nucleophiles and electrophiles, respectively, several
synthetic approaches have been reported for tailoring the spa-
12
tease significantly enhances (ꢀ10 times) the hydrolysis of
amides through the cooperative catalytic triad nucleophilic al-
[2]
[6]
cohol, basic imidazole, and acidic carboxylic acid interactions.
tial separation between the antagonist functional groups.
To achieve high activity, mutually incompatible groups, e.g.,
acids and bases, must be positioned to suppress direct proton
transfer; however, the distance between the two active moiet-
ies must be close enough so that they can function synergisti-
The later feature is crucial for cooperative catalysis because the
functional groups must be close enough to each other to ach-
ieve cooperativity without mutual neutralization. For example,
Kubota and co-workers have shown that higher rates of aldol
condensation can be achieved if homogeneous amines are
used with mesoporous silica containing weakly acidic silanol
[
3]
cally. This paradigm of biological transformations has inspired
the synthesis of a large library of homogeneous acid–base bi-
functional catalysts that exhibit excellent activity in many or-
[7]
groups. Katz and co-workers have shown the pivotal role of
silanol groups in the acid–base cooperativity between the sila-
nol and amines groups by demonstrating that the activity of
such acid–base pairs is greatly reduced when the silanol
[
4]
ganic transformations.
Solid-base catalysts synthesized by functionalization of orga-
nosilanes or nitrogen doping of inorganic oxides constitute
a new class of heterogeneous catalysts capable of catalyzing
[8]
groups are passivated. The structure of the support can also
influence the performance of acid–base pairs. Shimizu and co-
workers have observed that amines supported on mesoporous
silica (FSM-16) are nearly twice as active as amines supported
on amorphous silica and five times more active than homoge-
[
a] Dr. S. Shylesh, D. Hanna, J. Gomes, S. Krishna, Prof. A. T. Bell
Department of Chemical and Biomolecular Engineering
University of California
Berkeley, 94720 (USA)
Fax: (+1)510-642-477
E-mail: bell@cchem.berkeley.edu
[9]
nous amines. The superior performance of mesoporous silica
was attributed to the enrichment of reactants inside the or-
dered channels of this support. The influence of acidity of the
acid groups in acid–base pairs has been explored by Davis and
[b] Dr. C. G. Canlas, Prof. M. Head-Gordon
Department of Chemistry
University of California
Berkeley, CA 94720 (USA)
[10]
co-workers. These authors noted that the activity of the bi-
functional materials increased with increasing pK of the organ-
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cctc.201301087.
a
ic acid component, and the material containing weakly acidic
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 9
&
1
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
1
3
carboxylic acid and amine groups exhibited the highest activity
for aldol condensation. By contrast, Jones and co-workers have
recently shown that an organosilane possessing carboxylic acid
and amine groups are less active than surface silanols and
5 s recycle delay at a spinning rate of 7–13 kHz. All C spectra
were referenced against the chemical shifts of adamantane at
29
1
3
8.48 and 29.45 ppm. The Si with H decoupling MAS NMR spec-
tra were acquired at 99.37 MHz, by using a Si 908 pulse width of
.5 ms, recycle delay of 600 s, and spinning rate of 10–11 kHz. All
Si spectra were referenced against polydimethylsiloxane at
22 ppm (relative to TMS at 0 ppm). The resolution obtained in
the Si NMR spectra was sufficient for accurate peak assignments,
and the relative peak area of each site was obtained by the curve-
29
7
[
6,11]
amines.
29
The present study was focused on identifying the factors
governing the catalytic activity of silica-supported primary, sec-
ondary, and tertiary amine catalysts for the gas–phase self-con-
densation of n-butanal to 2-ethylhexenal. We report here ex-
periments conducted with the aim of establishing the roles of
amine groups and their surface density, the size of the pores in
the mesoporous silica support, the surface density of silanol
groups on the surface of the support, and the influence of the
spatial separation between the acid and base groups to the
catalytic activity of silica-supported amine catalysts for aldol
self-condensation of n-butanal. The mechanism and kinetics of
n-butanal condensation were also investigated in detail.
ꢁ
29
[14]
fitting, using a series of Gaussian peaks.
Gas-phase self-condensation of n-butanal was performed in
a 6.35 mm OD (ꢀ4 mm ID) quartz tube containing an expanded
section (ꢀ12.7 mm OD, ꢀ20 mm length). The reactor was packed
with quartz wool above and below the catalyst bed to hold the
catalyst in place. The feed to the reactor consisted of n-butanal
0.02 cm h ) and He. The catalysts were pretreated at 423 K
before passing the feed. Experiments were performed at 343 K,
total gas pressures of 1 atm, total gas flow rate of 150 cm min .
Under these conditions, the conversion of n-butanal was less than
3
ꢁ1
(
3
ꢁ1
1
0%. Reaction products were analyzed by using an Agilent 6890N
gas chromatograph containing a bonded and cross-linked (5%-
phenyl)-methylpolysiloxane capillary column (Agilent, HP-1) con-
nected to a flame ionization detector.
Experimental Section
The preparation of silica-supported amine catalysts was performed
in the following manner. One gram of silica (Silicycle, surface area:
2
ꢁ1
5
4
00 m g , average pore diameter 6 nm) was dried in vacuum at
23 K for 24 h, or pretreated at 823 K and 973 K for 5 h, and then
Results and Discussion
stored in vacuum prior to use. The silylating agents used were 3-
aminopropyltrimethoxysilane (APTS), 3-(N-methylaminopropyl)tri-
methoxysilane (MAPTS) and 3-(N,N-diethylaminopropyl)trimethoxy-
silane (DEAPTS) and methyltrimethoxysilane (MMTS) obtained from
Gelest, Inc. Typically, grafting of organoamines onto silica was
done by stirring a 1 g portion of silica with the required amount of
amino organosilanes in ethanol (100 mL) at 343 K for 15 h under
inert atmosphere. The hot solution was then cooled to RT, filtered,
washed with copious amounts of ethanol, and then dried in
a vacuum oven at 373 K overnight. Silica-supported primary (SiO2–
NH ), secondary (SiO –NHR) and tertiary amines (SiO –NR ) were
The successful grafting of the respective organosilanes onto
13
the silica supports was confirmed by C CP MAS NMR (see
Supporting Information). The observed chemical shifts of the
organic amine groups agree well with those of the corre-
[15]
sponding organosilane precursors measured in solution.
Small peaks were also observed at d=62 and 20 ppm and at-
tributed to ethoxy carbon atoms formed by partial esterifica-
[16] 29
tion of surface silanol groups by ethanol.
Si MAS NMR fur-
ther confirmed the presence of organic functional groups
grafted on the silica support (Supporting Information). Peaks
at d=ꢁ110, ꢁ100, ꢁ90, ꢁ65, and ꢁ55 ppm were assigned to
2
2
2
2
ꢁ2
prepared with nearly identical amine loadings (ꢀ0.4 aminenm ,
according to elemental analysis). A similar synthesis procedure was
used for grafting of a secondary amine onto mesoporous silicas
4
3
2
3
the Q (Si(OSi) ), Q (Si(OH)(OSi) ), Q (Si(OH) (OSi) ), T (SiR(OSi) ),
4
3
2
2
3
2
2
ꢁ1
and T (Si(OH) R(OSi) ) sites, respectively. The presence of
2
(
2
MCM-41, SBA-15, surface area: 800–1300 m g , pore diameter
.5–12 nm).
3
[12,13]
a sharp peak for T sites confirms the formation of a strong co-
valent linkage between the organic amine groups and the sup-
Silylation of amine-supported silica catalysts was
performed by dispersing the catalyst (1 g) in 25 mL of dry toluene
followed by the addition of methyl trimethoxysilane (3.75 mmol)
and stirring at 373 K for 15 h under an inert atmosphere. The final
[12]
port surface.
The catalytic activity of silica-supported amines was evaluat-
ed for the self-condensation of n-butanal to 2-ethylhexenal. As
shown in Figure 1, the catalyst prepared with secondary amine
was approximately five times more active than that prepared
with primary amine, whereas the catalyst prepared with terti-
ary amine showed negligible activity. Control experiments with
silica displayed no catalytic activity, indicating that the surface-
bound organic amine functional groups are responsible for the
observed catalytic activity. The only products observed under
the conditions investigated were the condensation product, 2-
ethylhexenal. Despite the stronger basicity of alkyl-substituted
tertiary amines, the poor catalytic activity observed with silica-
supported tertiary amine suggests that aldol condensation
does not occur according to a general base-catalyzed mecha-
nism. Instead, the order of the activity (secondary@primary>
tertiary) suggests that the reaction proceeds through an enam-
material is referred to as Sil-SiO –NHR.
2
Infrared spectra were acquired by using a Thermo Scientific Nicolet
6
700 FTIR spectrometer equipped with a liquid-nitrogen-cooled
MCT detector. Each spectrum was obtained by averaging 32 scans
ꢁ1
taken with 1 cm resolution. A 0.05 g portion of silica-supported
amine was pressed into a 20 mm-diameter pellet (<1 mm thick)
and placed into a custom-built transmission cell equipped with
CaF2 windows, a K-type thermocouple for temperature control,
and resistive cartridge heaters. Unless otherwise mentioned, all
scans were acquired at 343 K. The spectrum of the catalyst under
He was subtracted from the results reported.
1
3
29
Solid-state C CP MAS NMR and Si MAS NMR experiments were
performed on a Bruker Avance I 500 MHz spectrometer equipped
with a H/X double resonance magic-angle spinning probe that
13
uses 4 mm O.D. rotors. C cross-polarization, tuned to 125.79 MHz,
1
MAS NMR experiments were obtained by using a H 908 pulse
width of 4.2 ms, 2 ms contact time, 60 kHz decoupling field and 2–
[8,17]
ine mechanism involving a carbinolamine intermediate.
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 9
&
2
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
to iminium species, the formation of which is frequently re-
[17]
ported for homogenous secondary amine catalysts. By con-
trast, the spectrum of SiO –NH (spectrum c) during contact
2
2
ꢁ
1
with n-butanal exhibits sharp peaks at 1705 and 1670 cm .
The first of these peaks is assigned to carbonyl vibrations of n-
butanal adsorbed on the silanol groups, whereas the second
peak is assigned to the formation of imine (ꢁHC=Nꢁ) spe-
[19]
cies. This species was found to be stable up to 473 K. The
formation of stable imines was further confirmed using UV/Vis
13
spectroscopy and solid-state C CP MAS NMR of the spent cat-
alysts (see Supporting Information). The only feature observed
in the spectrum of SiO –NR (spectrum a) was a band at
2
2
ꢁ
1
1
705 cm , attributable to interactions of n-butanal with the si-
lanol groups of the support.
Figure 1. The effect of amine composition on the rate of 2-ethylhexenal syn-
thesis over silica-supported amine catalysts Ptotal =1 atm; Pn-butanal =0.2 kPa,
The low catalytic activity observed for SiO –NH is attributed
2
2
3
ꢁ1
to the formation of stable imines, which are known to be in-
balance He; total gas flow rate=150 cm min ; T=343 K; amine load-
ing=0.4 ꢁaminenmꢁ , TOF=turnover frequency.
2
[5,19]
hibitors for enamine catalysis.
The absence of imines or en-
amines in the spectrum of SiO –NR suggests that the reaction
2
2
proceeds through an ion-pair mechanism, with the formation
of enolate species in the rate-determining step. The ion-pair
mechanism involves the abstraction of a proton from the
methylene carbon by a base, forming a carbanion, which then
attacks the carbonyl carbon of another aldehyde to form the
CꢁC bond-forming step with the release of a proton from the
base to form the product (see Supporting Information). Corma
and co-workers have shown that the ion-pair mechanism give
rise to kinetics that are first order in aldehyde if proton ab-
The apparent activation energies calculated from Arrhenius
plots for primary, secondary and tertiary amines in the temper-
ꢁ
1
ature interval 333–353 K are 97, 82, and 115 kJmol , respec-
tively. The partial pressure dependence on the rate of forma-
tion of 2-ethylhexenal was measured by varying the partial
pressure of n-butanal while keeping the total gas flow rate and
3
ꢁ1
pressure constant at 150 cm min and 1 atm. The order of re-
action rate with respect to the partial pressures of n-butanal
for primary and secondary amine catalysts was 0.40 and 0.43,
respectively. By contrast, a first-order dependence in n-butanal
was noted for the silica-supported tertiary amine catalysts (see
Supporting Information).
[20]
straction is assumed to be the rate-limiting step. This finding
is consistent with the first-order partial pressure dependency
on n-butanal observed with the tertiary amine catalysts and
suggests that aldol condensation on the tertiary amine cata-
lysts proceeds through a general base-catalyzed mechanism.
Having established that silica-supported secondary amines
exhibit the highest activity for self-condensation reaction of
n-butanal, the density of secondary amine groups on the silica
surfaces was varied to investigate the effect of amine loading
on the catalytic activity. As seen from Figure 3, increasing the
amine surface density on the silica surface progressively de-
creased the catalytic activity. Even though an increase in the
amine surface density leads to a decrease in the silanol surface
density, the least active amine catalyst, which contains 0.6 ꢁ
In situ IR spectroscopy was used to obtain information
about the mechanism of n-butanal self-condensation on silica-
supported amine catalysts and to identify differences in the
chemistry of primary, secondary and tertiary amine catalysts. In
Figure 2 the spectra of the catalysts exposed to n-butanal and
recorded at 343 K are shown. SiO –NHR (spectrum b) exhibits
2
ꢁ1
bands at 1705, 1650, and 1630 cm . The first of these peaks is
attributable to carbonyl vibrations of n-butanal adsorbed on
the silanol groups, whereas the second two peaks are attribut-
able to the C=O and C=C vibrations of the product, 2-ethyl-
[
18]
ꢁ1
ꢁ2
hexenal. The shoulder at 1680 cm is tentatively assigned
NHR nm still possess a ten-fold excess of silanol relative to
Figure 3. The effect of secondary amine loading on the rate of 2-ethylhexe-
Figure 2. In situ FTIR spectra of n-butanal adsorbed at 343 K on silica-sup-
ported amine catalysts: (a) SiO –NR ; (b) SiO –NHR, and (c) SiO –NH .
2 2 2 2 2
nal synthesis: Ptotal =1 atm; SiO
gas flow rate=150 cm min ; T=343 K.
2
-423K; Pn-butanal =0.2 kPa, balance He; total
3
ꢁ1
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 9
&
3
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
amine groups. This suggests that at low amine loadings, effi-
cient amine–silanol catalytic cooperativity exists because the
amine groups are isolated from one another and cannot inter-
act through hydrogen bonding as a consequence of clustering
of the amine groups near the pore mouth in the silica sup-
performed by cofeeding water with n-butanal into the reaction
stream. A 20-fold decrease in the rate of aldol condensation
was noted after 6 h time on stream relative to that observed
without the addition of water. IR spectra of SiO –NHR catalysts
2
in the presence of co-fed water with n-butanal confirmed that
water interferes with the adsorption of n-butanal on the silanol
group. The loss in catalytic activity with co-fed water can thus
be related to the passivation of silanols with water molecules,
where the aldehyde need to adsorb and get activated during
the CꢁC bond forming step (see Supporting Information).
The role of silanol groups on the amorphous silica support
surface was further probed by changing the surface density of
silanol groups by varying the temperature at which dehydra-
tion of the silica support was performed, prior to grafting the
secondary amines. Dehydration temperatures of 423 K, 823 K
and 973 K were used. The apparent surface concentration of
[
6]
port.
To determine how the spatial confinement of amine groups
enhances the catalytic activity, secondary amines were sup-
ported on high-surface-area mesoporous silicas having pore di-
ameters of 2.5 nm to 12 nm. The catalytic activity increased for
mesoporous silicas up to a pore diameter of approximately
6
nm, after which the activity remained nearly constant
(Figure 4). Remarkably, a three-fold increase in catalytic activity
29
the silanols determined from Si MAS NMR spectroscopy and
the BET surface areas reveal that the concentration of silanol
ꢁ
2
groups decreases from 4 OHnm for SiO -423K to 2.8 OH
2
ꢁ
2
ꢁ2
[14]
nm for SiO -823K and to 1.6 OHnm for SiO -973K.
As
2
2
shown in Figure 5, the catalytic activity passed through a maxi-
Figure 4. The effect of the pore size of mesoporous-silica-supported secon-
dary amine catalysts on the rate of 2-ethylhexenal synthesis: Ptotal =1 atm;
3
ꢁ1
P
n-butanal =0.2 kPa, balance He; total gas flow rate=150 cm min ; T=343 K;
ꢁ
2
amine loadingꢀ0.4 ꢁaminenm
.
was noted if the pore size of the support was changed from
.5 nm to 6 nm. These results suggest that the rate of reaction
2
is limited by molecular diffusion of the products if the reaction
is conducted with mesopores of less than 6 nm. Comparison of
Figure 5. The effect of silanol density on the rate of 2-ethylhexenal synthesis
over silica-supported silica secondary amine catalysts: Ptotal =1 atm; Pn-buta-
ꢁ
2
3
ꢁ1
the activities of secondary amines (0.4 ꢁNHR nm ) supported
mesoporous SBA-15 and amorphous silica (average pore diam-
eterꢀ6 nm) revealed that the mesoporous SBA-15 has a two-
fold higher catalytic activity than the amorphous silica. The
higher activity obtained with SBA-15 may be attributed to the
nal =0.2 kPa, balance He; total gas flow rate=150 cm min ; T=343 K;
amine loading=0.4 ꢁamine nmꢁ
2
.
mum as the concentration of silanol groups decreased. Secon-
dary amine supported on silica pretreated at 823 K was nearly
two times more active than that supported on silica pretreated
at 423 K, and four times more active than that supported on
silica pretreated at 973 K.
ꢁ
2
(
a) higher silanol content of SBA-15 (5 ꢁOHnm ) than of
ꢁ2
amorphous silica (3.8 ꢁOH nm ) and/or (b) to the difference in
acidity of silanols between mesoporous and amorphous sili-
[
21]
cas. To better understand whether the high density of the si-
lanol groups or the acidity of these groups contributes to the
high activity of the secondary amine supported SBA-15, each
of these factors was evaluated separately.
It is known that an increase in the calcination temperature
of silica not only reduce the concentration of silanol groups on
the silica surface but also increase the acidity of the remaining
To assess the influence of silanol group concentration, the
surface concentration of silanol groups was reduced by graft-
ing methyltrimethoxysilane onto the amine-supported silica. Ir-
respective of the support material (SBA-15 or amorphous
silica), blocking the silanol groups by silylation decreased the
rate seven-fold, confirming that silanols act cooperatively with
silanol groups. For instance, the pK of isolated silanols is re-
a
3
ported to be approximately 2–4 (Q sites) whereas the pK of
a
2
[22]
geminal silanols (Q sites) is approximately 6–8. To ascertain
whether the acidity of the silanol groups enhances the rate,
13
C CP MAS NMR spectra were taken of secondary amine
groups supported on silica pretreated at different tempera-
29
13
amines in the aldol condensation of n-butanal. Si MAS NMR
spectroscopy corroborated that silylation decreased the total
concentration of silanol groups by approximately 35%. To fur-
ther ascertain the role of silanol groups, an experiment was
tures. As shown in Figure 6, C CP MAS NMR spectra did not
display an upfield shift in the terminal carbon peak position in-
dicating that amines do not interact with the acidic silanol
groups. This result was corroborated by the adsorption of ho-
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 9
&
4
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
1
3
Figure 6. Effect of pretreatment temperature on the solid-state C CP MAS
NMR spectrum of silica-supported secondary amine catalysts: (a) SiO -423K;
b) SiO -823K; and (c) SiO -973K.
2
Figure 7. Effect of linker length on the catalytic activity of silica-supported
secondary amine catalysts (solid bars) and comparison with their silylated
analogs (striped bars). Ptotal =1 atm; Pn-butanal =0.2 kPa, balance He; total gas
(
2
2
3
ꢁ1
2
flow rate=150 cm min ; T=343 K; SiO -823K; amine loading=
2
0
.4 ꢁaminenmꢁ
.
mogenous triethylamine (NEt ) on silica. Irrespective of the pre-
3
treatment procedures, an upfield shift of the terminal carbon
atoms of organic amine catalysts did not occur confirming that
acidity of the silanol groups does not contribute to the en-
hanced catalytic activity of the catalyst prepared on silica de-
shown that the activity of acid–base-pairs catalysts can be
tuned by altering the linker length of the amine.
[
23]
hydrated at 823 K (see Supporting Information). Further indi-
cation that the dehydration temperature has no effect on the
acidity of the silanols was obtained from the absence of a shift
The acid-base cooperativity was further demonstrated by si-
lylating the silanol groups on the SiO –C –NHR and SiO –C –
2
3
2
1
NHR. Interestingly, silylation decreased the catalytic activity of
the SiO –C –NHR by ten times, whereas the activity of silylated
ꢁ
1
in the position of the 1705 cm peak for n-butanal adsorbed
on silanol groups. All these results taken together indicate that
the temperature of silica dehydration does not alter the acidity
of the silanol groups for the conditions used in this study.
The influence of the spatial separation between the amine
and silanol groups was examined to determine its effect on
the activity of acid–base pairs. The average distance between
silanols on silica pretreated at 423 K is approximately 0.3–
2
3
SiO –C –NHR remained similar to that of nonsilylated SiO –C –
2
1
2
1
NHR (Figure 7). These results suggest that silica-supported pro-
pylamines exhibit cooperative interactions with the silanol
groups for rate enhancements whereas the constrained meth-
ylamine lacks the acid–base cooperativity. Alternatively, the
low activity observed with SiO –C –NHR suggests that the re-
2
1
action does not proceed through a silanol-assisted acid–base
bifunctional mechanism, rather this may proceed through the
general base-catalyzed mechanism. A first-order dependence
on n-butanal partial pressure and a higher activation barrier
0
.4 nm and the propylamine linker has a chain length of ap-
[
24]
proximately 0.5 nm. This means that amines can reach over
the silanols in SiO -423K, making all the amine groups unavail-
2
ꢁ1
able for the catalysis. By contrast, for the SiO -823K support,
(122 kJmol ) support the proposal that aldol condensation on
SiO –C –NHR proceeds through general base-catalyzed
2
the silanols are separated by an average distance of approxi-
mately 0.6–0.7 nm, which means that the amine groups
cannot interact strongly with silanol groups, leading to an en-
hanced catalytic activity. Consequently, for secondary amine
a
2
1
mechanism (see Supporting Information). Thus, short alkyl link-
ers like methyl inhibits the beneficial amine–silanol cooperativi-
ty on silica surfaces resulting in lower catalytic activity.
supported SiO -973K catalysts, moving the silanols farther
The accumulated data suggest that the self-condensation of
n-butanal over SiO –C –NHR proceeds through a concerted
2
apart lead to an inactive catalyst, because the high catalytic ef-
ficiency depends on the cooperative activation of reactants by
the amines and the silanols. These studies clearly demonstrate
how the spatial separation between acid and base groups in-
fluences their catalytic activity. In an alternate way to confirm
that an accurate spatial separation between the amines and
the silanols leads to an enhanced catalytic activity, secondary
amine catalyst having a methyl linker (SiO –C –NHR) were
2
3
mechanism involving the basic amine groups and the weakly
acidic silanol groups, as shown in Scheme 1. A theoretical anal-
ysis of the aldol condensation reaction pathway was per-
formed to test the plausibility of this mechanism and to deter-
mine the rate-limiting step. The active-site model for these
steps contains both a grafted amine group (C H NHCH ) and
3
6
3
a silanol group (SiꢁOH), as represented by the cluster shown in
Figure 8. A model of amorphous silica containing four tetrahe-
dral silicon atoms (T4) was obtained from the crystal structure
of the zeolite H-MOR [S1], with all silicon atoms terminated in
oxygen. The cleaved structure gives an O–O silanol separation
of 0.5 nm, consistent with the experimental measurement from
the most active SiO –C –NHR catalyst. To create the dual-site
2
1
grafted onto SiO -823K support, instead of the propyl linker
2
[
25]
(
SiO –C –NHR).
Remarkably, constrained methyl linker
2
3
showed a ten-fold decrease in activity relative to that for the
propyl linker demonstrating that an accurate distance between
the silanols and propyl amine groups, on SiO -823K support
2
leads to an active and stable catalyst for the self-condensation
of n-butanal (Figure 7). These results are consistent with those
reported recently by Jones and co-workers in which it was
2
3
model containing the grafted amine adjacent to a silanol
group, two terminal SiꢁO fragments were substituted for
SiꢁOH and SiꢁC H NHCH , respectively. The remaining terminal
3
6
3
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 9
&
5
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
theory are shown in Figure 9
and the corresponding free-
energy (343.15 K) changes are
shown in Figure 10. The adsorp-
tion of n-butanal to the bare cat-
alyst surface is exothermic (DH=
ꢁ
1
ꢁ
25 kJmol ) and is character-
ized by a weak hydrogen bond-
ing interaction between n-buta-
nal and the grafted amine. The
formation of
a carbinolamine
species proceeds through a con-
certed CꢁN bond formation and
proton transfer reaction. The
transition-state structure for car-
binolamine formation is similar
to one found for proline-cata-
[26]
lyzed aldol condensation. The
enamine state is produced by in-
tramolecular water elimination.
The iminium state can be
formed by the protonation of
the enamine by the silanol
group, but is omitted from this
diagram because of its instability
relative to the enamine state
ꢁ
1
(
DH= +35 kJmol ). The de-
sorption of water exhibits a fa-
vorable free-energy change at-
tributable to the additional gain
in entropy. The adsorption of
n-butanal in the presence of the
enamine resting state is exother-
Scheme 1. Proposed mechanism for aldol self-condensation of n-butanal over silica-supported secondary amine
catalysts.
ꢁ
1
mic (DH=ꢁ32 kJmol ) and is promoted by hydrogen bond-
ing between the silanol group and the carbonyl group of n-bu-
ꢀ
ꢁ1
tanal. The formation of the CꢁC bond (DH = +116 kJ mol )
proceeds by the protonation of the n-butanal carbonyl group
by the silanol group and CꢁC bond formation between the en-
amine and the alpha-carbon of n-butanal in a concerted fash-
ion, and produces a second iminium species. The hydrolysis
proceeds from the second iminium by the attack of water on
the complex. This process leads to a concerted proton transfer
from the water molecule to the deprotonated silanol group,
leaving the final product, 2-ethyl-3-ol-hexenal, in the chemisor-
bed state. The product then desorbs to reform the catalyst
(
SiO –C –NHR) and leave 2-ethyl-3-ol-hexenal in the physisor-
2 3
bed state. The final step of the reaction pathway is desorption
of 2-ethyl-3-ol-hexenal into the gas phase followed by its dehy-
Figure 8. Model active site for the SiO
: C; *: N.
2 3
–C –NHR catalyst. *: Si, *: O, *: H,
*
[18]
dration to form the final product, 2-ethylhexenal. The theo-
retical analysis of the n-butanal self-condensation pathway
leads to the conclusion that the transition state for CꢁC bond
formation is the rate-limiting step along the free energy path-
O atoms were replaced by H and fixed in their crystallographic
positions during subsequent optimization. The resulting cluster
model has the stoichiometry of C H NO Si . (Further details
[26–28]
way.
The resting state of the catalyst during reaction is the
4
19
4
4
concerning the computational method are provided in the
Supporting Information).
enamine state, since it is the lowest-lying intermediate on the
free energy surface, consistent with the characterization of the
The changes in enthalpy (343.15 K) along the reaction path-
way calculated at the wB97X-D/6-311+ +G(3df,3pd) level of
spent SiO -NHR catalysts. The apparent activation barrier is cal-
2
culated as the difference in enthalpy between the rate-limiting
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 9
&
6
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
is the partial pressure of species
j, and L and L1 represent the
total number of surface silanol
and amine groups, respectively.
A detailed derivation of Equa-
tion (1) is provided in the Sup-
porting Information.
Equation (1) can be used to
rationalize the experimental re-
sults. Under the experimental
conditions used in this study, the
partial pressure of H O was neg-
2
ligible because the conversion
was less than 10%. Therefore,
Equation (1) simplifies to Equa-
tion (2).
2 3
Figure 9. Enthalpy (343.15 K) diagram for aldol condensation of n-butanal over a SiO –C –NHR catalyst calculated
ꢁ
1
at the wB97X-D/6-311+ +G(3df,3pd) level of theory (kJmol ). The elementary steps are labeled according to the
K k P
LL =ðL þ LÞ
5
6
butanal
1
1
ꢀ
ꢁ1
rP ¼
ð2Þ
reaction mechanism in Scheme 1. The calculated apparent activation energy (DHapp ) is 85 kJmol
.
1
þ K5P
butanal
The form of Equation (2) is
consistent with the observation
that the rate of 2-ethylhexenal
formation is 0.4 order in n-buta-
nal, suggesting that both terms
in the denominator contribute
to the overall rate.
The PH2O terms of Equation (1)
in the denominator describe the
inhibitory effect of water ob-
served experimentally. Water
vapor inhibits the catalytic reac-
tion by hydrolyzing the carbon–
carbon double bond of the en-
amine, thereby decreasing con-
centration of the enamine inter-
mediate by shifting equilibrium
of reaction 3 to the left. Addi-
tionally, water vapor competes
with aldehyde for vacant silanol
groups (see Supporting Informa-
tion).
Figure 10. Free-energy (343.15 K) diagram for aldol condensation of n-butanal over a SiO
2
–C
lated at the wB97X-D/6-311+ +G(3df,3pd) level of theory (kJmol ). The elementary steps are labeled according
to the reaction mechanism in Scheme 1. The difference in free energy between the resting state (IV, ꢁH O) and
the rate limiting transition state (5 ) is denoted as DGapp
3
–NHR catalyst calcu-
ꢁ
1
2
ꢀ
ꢀ
.
step and the resting state of the catalyst. The calculated appar-
The proposed mechanism for n-butanal condensation dem-
onstrates that for efficient catalysis, the basic amine groups
must work in concert with the acidic silanol groups. Strong
acids are undesirable because they can protonate the amine
group, thereby lowering their nucleophilicity, which is essential
for formation of enamine species. It is thus essential that the
distance between the acid and base sites should be large
enough to avoid mutual neutralization but close enough so
that the acid and base groups can
ꢁ
1
ent activation energy for this process is 85 kJmol , in good
ꢁ1
agreement with the value of 82 kJmol determined experi-
mentally.
Based upon the proposed reaction mechanism and consider-
ing that reactions prior to the rate-limiting CꢁC bond-forming
step are quasi-equilibrated, the rate of 2-ethylhexenal forma-
tion can be expressed as follows:
2
aP
LL =ðL þ LÞ
interact with the reactant molecule
butanal
2
1
1
rP ¼ PH2O
ð1Þ
2
2
2
þ bPbutanal þ gPH2OPbutanal þ K4P þ dP
þ eP Pbutanal þ lPH2O
P
for cooperative catalysis to occur.
This effect is accurately modeled by
the two-site rate equation indicating
H2O
butanal
H2O
butanal
where a, b, g, d, e, and l represent lumped equilibrium and
rate parameters, K is the equilibrium constant for reaction 4, P
that a lone silanol or amine site cannot exclusively catalyze the
reaction. By increasing the density of amine, the value of the
4
j
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 9
&
7
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
L’=LL /(L+L ) term rises to a maximum and then decreases.
[2] S. Kraut, Annu. Rev. Biochem. 1977, 46, 331–345.
1
1
[
3] E. L. Margelefsky, R. K. Zeidan, M. E. Davis, Chem. Soc. Rev. 2008, 37,
118–1126.
4] W. Notz, F. Tanaka, C. F. Barbas, Acc. Chem. Res. 2004, 37, 580–591.
This trend qualitatively explains the effect of amine site density
on the rate of aldol condensation reaction (see Supporting In-
formation).
1
[
[5] S. Shylesh, W. R. Thiel, ChemCatChem 2011, 3, 278–287 and references
therein.
[
[
6] N. A. Brunelli, C. W. Jones, J. Catal. 2013, 308, 60–72 and references
therein.
7] Y. Kubota, K. Goto, S. Miyata, Y. Goto, Y. Fukushima, Y. Sugi, Chem. Lett.
2003, 32, 234–235.
Conclusions
The results presented in this study provide a comprehensive
understanding of the factors influencing the acid–base cooper-
ativity of amine-supported silica catalysts for carbonꢁcarbon
bond formation during the self-condensation of n-butanal. Sec-
ondary amine functionalized silica is nearly five times more
active than that functionalized with primary amines; tertiary
amine functionalized silica exhibits negligible catalytic activity.
Formation of nucleophilic enamines leads to enhanced catalyt-
ic activity of secondary amine catalysts, whereas formation of
stable imines (up to 473 K) leads to the low activity of primary
amine supported silica catalysts. Secondary amines in combi-
nation with the isolated silanol groups provide the best acid–
base pair and the spatial separation between acid–base
groups can be tailored by simple calcination treatment of the
silica support or by varying the linker length of the organosi-
lane for catalytic enhancements. Detailed experimental as well
as theoretical analysis of the n-butanal self-condensation path-
way leads to the conclusion that carbonꢁcarbon bond forma-
tion is the rate-limiting step in the aldol condensation of n-bu-
tanal over silica-supported secondary amines.
[8] J. D. Bass, A. Solovyov, A. J. Pascall, A. Katz, J. Am. Chem. Soc. 2006, 128,
737–3747.
3
[
9] K. Shimizu, E. Hayashi, T. Inokuchi, T. Kodama, H. Hagiwara, Y. Kitayama,
Tetrahedron Lett. 2002, 43, 9073–9075.
[10] R. K. Zeidan, M. E. Davis, J. Catal. 2007, 247, 379–382.
[11] N. A. Brunelli, K. Venkatasubbaiah, C. W. Jones, Chem. Mater. 2012, 24,
2433–2442.
[
[
12] S. Shylesh, A. P. Singh, J. Catal. 2006, 244, 52–64.
13] M. Kruk, L. Cao, Langmuir 2007, 23, 7247–7254.
[14] S. Shylesh, D. Hanna, S. Werner, A. T. Bell, ACS Catal. 2012, 2, 487–493.
[15] S. A. Didas, A. R. Kulkarni, D. S. Sholl, C. W. Jones, ChemSusChem 2012,
5, 2058–2064.
[
16] S. Farsadpour, L. T. Ghoochany, S. Shylesh, G. Dorr, A. Seifert, S. Ernst,
W. R. Thiel, ChemCatChem 2012, 4, 401–407.
[17] B. List, Acc. Chem. Res. 2004, 37, 548–557.
[18] M. Singh, N. Zhou, D. Paul, K. Klabunde, J. Catal. 2008, 260, 371–379.
19] K. Kandel, S. M. Althaus, C. Peeraphadit, T. Kobayashi, B. G. Trewyn, M.
Pruski, I. I. Slowing, J. Catal. 2012, 291, 63–68.
20] A. Corma, R. M. Martin-Aranda, J. Catal. 1991, 130, 130–137.
[21] T. Yamamoto, T. Tanaka, T. Funabiki, S. Yoshida, J. Phys. Chem. B 1998,
102, 5830–5839.
22] J. M. Rosenholm, T. Czuryszkiewicz, F. Kleitz, J. B. Rosenholm, M. Linden,
Langmuir 2007, 23, 4315–4323.
23] K. Motokura, M. Tada, Y. Iwasawa, Chem. Asian J. 2008, 3, 1230–1236.
24] E. Papier, Adsorption on Silica Surfaces, Marcel Dekker Inc., 2000.
25] N. A. Brunelli, S. A. Didas, K. Venkatasubbaiah, C. W. Jones, J. Am. Chem.
Soc. 2012, 134, 13950–13953.
[
[
[
[
[
[
Acknowledgements
[
[
[
26] S. Bahmanyar, K. N. Houk, J. Am. Chem. Soc. 2001, 123, 11273–11283.
27] J. L. Reymond, J. Mol. Catal. A 1998, 5, 331–337.
28] M. Wiesner, G. Upert, G. Angelici, H. Wennemers, J. Am. Chem. Soc.
2
This work was supported by the XC program funded by BP.
2
010, 132, 6–7.
Keywords: acidity · aldol reaction · amines · organocatalysis ·
silica
Received: December 18, 2013
[1] D. E. Cane, Chem. Rev. 1990, 90, 1089–1103.
Published online on && &&, 0000
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 9
&
8
&
ÞÞ
These are not the final page numbers!

FULL PAPERS
Secondary amines come first: A solid-
base organocatalyst achieved by graft-
ing amines onto silica surfaces is ap-
plied to the gas-phase aldol self-con-
densation of n-butanal to 2-ethylhexe-
nal. Silica-supported secondary amine
catalysts demonstrate a much higher
catalytic activity than the primary amine
analogues, owing to the respective for-
mation of enamines as shown by in situ
FTIR analysis. The reaction pathway is
analyzed by DFT calculations.
S. Shylesh, D. Hanna, J. Gomes,
S. Krishna, C. G. Canlas, M. Head-Gordon,
A. T. Bell*
&& – &&
Tailoring the Cooperative Acid–Base
Effects in Silica-Supported Amine
Catalysts: Applications in the
Continuous Gas-Phase Self-
Condensation of n-Butanal
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 9
&
9
&
ÞÞ
These are not the final page numbers!