Lewis acid solid catalysts for the synthesis of methylenedianiline from aniline and formaldehyde

Article abstract of DOI:10.1016/j.jcat.2021.05.031

A catalyst containing Hf⁴⁺ and Zn²⁺ supported on silica has been found to be highly effective for the synthesis of methylenedianiline (MDA), an indispensable precursor in the polyurethane industry. Its performance was further improved when the silica support was replaced by silica-alumina, which resulted in a catalyst that was both active and selective, as indicated by the high MDA yield and high 4,4′–MDA/(2,2′–MDA + 2,4′–MDA) isomer ratio obtained. Furthermore, the catalyst also gave an appreciable oligomeric MDA (OMDA) yield and was noticeably more stable than the zeolites that were used in comparative tests: it could be used in at least five consecutive runs without any significant loss in activity. The combination of Br?nsted and Lewis acidity strongly increases the overall activity and yields a catalyst that represents a remarkably stable and reusable alternative to the commonly studied systems for this reaction.

Full text of DOI:10.1016/j.jcat.2021.05.031

Journal of Catalysis 400 (2021) 114–123
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Lewis acid solid catalysts for the synthesis of methylenedianiline from
aniline and formaldehyde
Ka Yan Cheung a, Carlos Marquez a, Patrick Tomkins a, Andrei-Nicolae Parvulescu b, Alvaro Gordillo b,
Trees De Baerdemaeker b, Dirk De Vos a,
⇑
a Centre for Membrane Separations, Adsorption, Catalysis and Spectroscopy for Sustainable Solutions (cMACS), KU Leuven, Celestijnenlaan 200F, Box 2454, 3001 Leuven, Belgium
b BASF SE, Process Research and Chemical Engineering, Carl-Bosch-Strasse 38, 67056 Ludwigshafen am Rhein, Germany
a r t i c l e i n f o
a b s t r a c t
A catalyst containing Hf4+ and Zn2+ supported on silica has been found to be highly effective for the syn-
thesis of methylenedianiline (MDA), an indispensable precursor in the polyurethane industry. Its perfor-
mance was further improved when the silica support was replaced by silica-alumina, which resulted in a
catalyst that was both active and selective, as indicated by the high MDA yield and high 4,40–MDA/(2,20–
MDA + 2,40–MDA) isomer ratio obtained. Furthermore, the catalyst also gave an appreciable oligomeric
MDA (OMDA) yield and was noticeably more stable than the zeolites that were used in comparative tests:
it could be used in at least five consecutive runs without any significant loss in activity. The combination
of Brønsted and Lewis acidity strongly increases the overall activity and yields a catalyst that represents a
remarkably stable and reusable alternative to the commonly studied systems for this reaction.
Ó 2021 Published by Elsevier Inc.
Article history:
Received 3 February 2021
Revised 13 May 2021
Accepted 27 May 2021
Available online 01 June 2021
Keywords:
Methylenedianiline
Diaminodiphenylmethane
Aminal
Lewis acid
Polyurethane
1. Introduction
diphenylmethylenediamine (also called aminal) and aniline, here-
after referred to as aminal solution. In the second step, the aminal
Methylenedianiline (MDA, or diaminodiphenylmethane) is an
important organic compound predominantly used in the produc-
tion of methylene diphenyl diisocyanate (MDI), a precursor to
polyurethane, a polymer class with sales of over 20 million tons
globally and a market estimated at USD 65.5 billion in 2018 [1,2].
MDA is industrially synthesised in high yields by condensation of
aniline and formaldehyde in a homogeneously catalysed process.
The catalyst employed, hydrochloric acid, presents several draw-
backs. Large quantities of a base, typically sodium hydroxide, are
used to neutralise the resulting hydrochloride salt. Furthermore,
the generated waste water has to be treated to remove organic
contaminants prior to discharge. In light of this, there has been a
strong incentive to search for solid acid catalysts to replace HCl
whilst still fulfilling the following requirements: (i) enable reason-
able yields to be reached with minimal intermediates and by–
products formation, (ii) offer control in the product composition,
(iii) have a long lifetime, and (iv) be regenerated in a feasible
way [3–5]. The heterogeneous process is generally split into two
steps: a non–catalysed condensation reaction first takes place
between aniline and formaldehyde, followed by phase separation
to isolate the organic phase which contains the condensate N,N’–
is consumed in a series of acid–catalysed rearrangement reactions
to form the MDA products via two N–(aminobenzyl)aniline inter-
mediates (o– and p–ABA). The product mixture consists of MDA
isomers (4,40–, 2,40–, and 2,20–MDA), an oligomeric fraction (oligo-
mers of ꢀ 3 aryl rings bridged by methylene groups), and by-
products (N-methyl-MDA, N-formyl-MDA, and quinazolines), out
of which 4,40-MDA, 2,40-MDA, and the oligomeric fraction are of
value. The reaction conditions of the two steps can be adjusted sep-
arately to tailor the product distribution to needs (Scheme 1)[3–5].
A wide array of materials and their modified versions have been
probed over the years as potential catalyst substitutes: ion-
exchange resins were found to be selective to the 4,40-isomer but
lacked sufficient catalytic activity [6]; some clays showed good sta-
bility but were not selective to the 4,40-isomer [7], and while silica-
aluminas did show an increase in activity and a slower rate of cat-
alyst deactivation compared with other materials, the 4,40-isomer
yield was low and eventual catalyst deactivation by fouling was
inevitable [8,9]. Zeolites with their unique properties such as
shape-selectivity and flexible acid strength, also emerged as candi-
dates, but the bulky nature of the aminal reactant and the products
implies that molecular diffusion limitations in the micropores of
the zeolite crystals would severely decrease the overall activity
and longevity of even large pore zeolites such as Beta and Morden-
ite [10–13]. To counteract this, access to the acid sites in zeolites
⇑
Corresponding author.
E-mail address: dirk.devos@kuleuven.be (D. De Vos).
https://doi.org/10.1016/j.jcat.2021.05.031
0021-9517/Ó 2021 Published by Elsevier Inc.

Ka Yan Cheung, C. Marquez, P. Tomkins et al.
Journal of Catalysis 400 (2021) 114–123
Scheme 1. Simplified reaction scheme for MDA synthesis.
Table 1
has been improved by swelling and delaminating precursors of lay-
ered zeolites such as MCM-22 to obtain 2D zeolites like ITQ-2, ITQ-
6, and ITQ-18, which performed better compared with the corre-
sponding zeolites. However, a sizeable quantity of catalyst was
required to achieve moderate 4,40-isomer yields [14,15]. Since all
of the catalytic systems studied so far have their downsides, there
is still much interest in exploring solid materials with a different
set of properties. One option that, to the best of our knowledge,
has not been addressed so far in MDA synthesis is Lewis acid catal-
ysis. Lewis acids can be supported on inorganic solids with high
surface areas such as clays, silicas, silica-aluminas, and zeolites,
in order to facilitate dispersion and accessibility of the active sites
[16]. We expect that catalysts with Lewis acidic character could be
viable candidates in the synthesis of MDA. To this end, various
Lewis acid catalysts were investigated and characterised in the
acid-catalysed synthesis of MDA.
Metal-loaded catalysts tested in MDA synthesis.
Metal-loaded silica samples (precursor salts)
Hafnium-loaded samplesa
La1.5/SiO2 (La(NO3)3ꢂ6H2O)
Ce1.7/SiO2 (Ce(NO3)3ꢂ6H2O)
Sc1.5/SiO2 (Sc(NO3)3ꢂxH2O)
Ta1.3/SiO2 (TaCl5)b
Hf3.0/Al2O3
Hf1.5/SiO2ꢂAl2O3
Hf3.0/SiO2ꢂAl2O3
Hf2.7Zn0.5/SiO2ꢂAl2O3
Hf3.0/Al2MgO4
Hf3.0/HAP
Hf3.0/TiO2
Hf3.0/ZrO2
Hf3.0/MCM-22
Hf3.0/Y
Sn1.5/SiO2 (Sn(CH3CO2)2)
Cr1.9/SiO2 (Cr(acac)3)c
Fe1.3/SiO2 (Fe(NO3)3ꢂ9H2O)
Ni1.4/SiO2 (Ni(NO3)2ꢂ6H2O)
Cu1.4/SiO2 (Cu(NO3)2ꢂ3H2O)
Zn1.7/SiO2 (Zn(NO3)2ꢂ6H2O)
Zr1.4/SiO2 (ZrCl4)
Hf1.5/SiO2 (HfCl4)
Hf2.7/SiO2 (HfCl4)
Hf2.6Zn0.5/SiO2 (HfCl4 and Zn(NO3)2ꢂ6H2O)
Zr1.9Zn0.3/SiO2 (ZrCl4 and Zn(NO3)2ꢂ6H2O)
The subscripts indicate the metal content (wt%) as determined by ICP-OES.
a
HfCl4 was used as the metal precursor salt.
Toluene was used to dissolve TaCl5 in the sample preparation.
Ethanol was used to dissolve Cr(acac)3 in the sample preparation.
b
2. Experimental
c
2.1. Catalyst preparation
Commercial zeolites were used in their proton form without
further modification. Zeolite H–Y was obtained from Zeolyst Inter-
national (CBV 720, Si/Al = 15), H–MCM–22 from China Catalyst
Group (Si/Al = 14), and H–Beta from PQ Corporation (CP811 BL–
25, Si/Al = 12.5) and from ZeoCat (Si/Al = 33). The silica used as a
support was acquired from Cabot (CAB–O–SILÒ Mꢁ5), the alumina
from Condea Chemie (PURALOX NGa–150), the magnesium alumi-
nate spinel (nanopowder, < 50 nm particle size) and hydroxyap-
atite (HAP, nanopowder, < 200 nm particle size) from Sigma-
Aldrich, the titanium(IV) oxide (anatase) and zirconium oxide from
Alfa Aesar, and the silica-alumina from Grace (MS 13/110, 13%
Al2O3). For the metal-loaded catalysts, target amounts of the metal
precursor salts were first dissolved in ultrapure water at room
temperature (unless stated otherwise); then the required amounts
of the supports were added and the mixture was left stirring until
the water evaporated. The resulting powder was dried at 60 °C,
then calcined in air at 500 °C for 2 h at a heating rate of 2 °C/min
to obtain the samples presented in Table 1.
using a Varian 720-ES equipped with a double-pass glass cyclonic
spray chamber, a SeaSpray concentric glass nebuliser and a high
solids torch. The textural properties of the solids were derived from
N2 physisorption isotherms, which were collected at ꢁ196 °C on a
Micromeritics 3Flex surface characterisation analyser. Samples
were outgassed under vacuum at 150 °C for 4 h prior to data col-
lection. The specific surface areas (SBET) were determined using
the BET method, the specific external surface areas (SEXT) and
micropore volumes (Vmicro) were obtained using the t-plot method,
and the mesopore volumes (Vmeso) were estimated by the BJH
method. Temperature programmed desorption of ammonia (NH3-
TPD) was used to determine the number of acid sites in the cata-
lysts. Measurements were conducted using a Gasmet DX4000
Fourier–Transform infrared (FTIR) gas analyser with Calcmet soft-
ware for converting the collected spectral data into ammonia con-
centrations. The sample was activated at 200 °C for 2 h under
nitrogen, then cooled to 100 °C and dosed with an ammonia/nitro-
gen gas mixture until saturation at this temperature. Non–ad-
sorbed ammonia was then flushed away with nitrogen at 100 °C.
The sample was subsequently heated at 5 °C/min under nitrogen
and the amount of desorbed ammonia was monitored with a
downstream Peltier cooled mercury cadmium telluride (MCT)
detector. The nature of the acid sites and the acid site density were
determined by pyridine adsorption followed by FTIR spectroscopy.
To this end, a self-supported wafer was placed in a cell under vac-
uum and activated at 400 °C for 1 h. The cell was subsequently
2.2. Catalyst characterisation
Powder X-ray diffractograms were recorded on a Malvern Pan-
alytical Empyrean diffractometer in Transmission/Debye-Scherrer
geometry (1.3 – 45° 2h range, 0.013° step size) with a PIXcel3D
solid state detector. The metal content was determined by induc-
tively coupled plasma-optical emission spectrometry (ICP-OES)
115

Ka Yan Cheung, C. Marquez, P. Tomkins et al.
Journal of Catalysis 400 (2021) 114–123
cooled and a reference spectrum was recorded at 150 °C. After this,
the cell was further cooled and pyridine (25 mbar) was contacted
with the wafer at 50 °C until the sample material was saturated.
Weakly coordinated pyridine was removed by evacuation for
30 min at 50 °C, prior to collection of the IR spectrum at 150 °C.
Using the difference spectrum generated from the measurements
and using Emeis’ integrated molar extinction coefficients [17],
the Lewis acidity was calculated from the area of the pyridine
absorption band at 1450 cmꢁ1 and the Brønsted acidity from the
area of the band at 1545 cmꢁ1. Samples were imaged by transmis-
sion electron microscopy (TEM) using a Tecnai Osiris machine
operated at 200 keV in High-Angle Annular Dark-field Scanning
(TEM) (HAADF-STEM) mode. Chemical composition maps were
acquired by energy-dispersive x–ray spectroscopy (EDXS) using a
Thermo-Fisher SuperX G1 detector. Prior to imaging, the powder
samples were dispersed in ethanol and applied on an ultra-thin
carbon TEM-grid.
monitored as a function of increasing temperature; therefore,
peaks appearing at higher temperatures indicate
a stronger
aniline-catalyst interaction. Aniline in THF (0.0025 ganiline/mLTHF
)
was added (5 wt% solution with respect to the catalyst) to a pre-
dried sample (dried overnight at 200 °C). The mixture was then
stirred at room temperature until it was dry. The measurements
were carried out at a heating rate of 5 °C/min to 600 °C under a
nitrogen atmosphere. TGA-MS data were collected on a NETZSCH
STA 449 F3 JupiterÒ thermal analyser and a Hiden HPR-20 EGA
gas analysis system.
3. Results and discussion
3.1. Catalytic reactions using commercial zeolites
As mentioned in Section 1, the reaction was carried out in two
steps; the first step being a non–catalysed condensation reaction
between aniline and formaldehyde to form a biphasic mixture of
the aminal condensate and water. The aminal solution was isolated
(water content <1 wt%, Fig. S1) and used in the second step which
involves a series of acid–catalysed rearrangement and condensa-
tion reactions to form the products. The first intermediates formed
are aminobenzylanilines (ABA), which further react to form the
methylenedianiline isomers (MDA) and their higher homologues,
oligomeric MDAs containing at least three aryl rings (OMDA). Typ-
ically, MDA is produced with a sizeable oligomeric fraction and an
excess of para– isomers compared to ortho– isomers, with the 2,20–
isomer being undesired [13]. A profile of how the reaction proceeds
over time is shown in Fig. S2. The 4,40–MDA/(2,20-MDA + 2,40-MDA)
isomer ratio drops as the reaction progresses, since the MDA prod-
ucts are formed from two ABA precursor isomers which differ in
reactivity. The product 4,40–MDA is derived from the rearrange-
ment of the p–ABA intermediate, 2,20–MDA from the o–ABA inter-
mediate, whilst both the p– and o–ABA intermediates contribute to
the formation of the 2,40–MDA product isomer. As can be observed
in Fig. S3, the p–ABA intermediate reacts considerably faster than
the o–ABA, which prompts the decrease in the eventual 4,40–MD
A/(2,20-MDA + 2,40-MDA) isomer ratio.
2.3. Synthesis of aminal
Aminal was the product of the non-catalysed condensation
reaction between aniline and formaldehyde. It was prepared fol-
lowing a modification of a previously reported procedure [15].
Briefly, in a 2 L two-necked round bottom flask, aniline (980 mL,
Acros Organics, 99.8% purity) was stirred and heated at 50 °C.
Formaldehyde (274 mL, VWR Chemicals, 36% aqueous solution,
stabilised with methanol) was added dropwise until the mixture
reached an aniline/formaldehyde (A/F) molar ratio of 3. The mix-
ture was stirred for a further 1 h at 50 °C. The formed aminal
and the residual aniline (generalised as aminal solution) was col-
lected in the organic layer, and the aqueous layer was discarded
after phase separation of the mixture. The remaining water content
in the organic layer was determined by Karl Fischer titration. The
A/F ratio was determined by proton nuclear magnetic resonance
spectroscopy (1H NMR). The NMR spectra were recorded on a Bru-
ker AVANCE III HD 400 MHz Spectrometer and CDCl3 (Sigma-
Aldrich, 99.8 atom % D) was used as the solvent.
2.4. Catalytic reactions
A series of H-form zeolites were preliminarily tested to serve as
a benchmark for this work. The yields of o–ABA and p–ABA were
combined into a single ABA value (ABA), whereas the MDA yield
represented the summed yields of the three MDA isomers. The cat-
alytic activity results in the conversion of the ABA intermediates
(Fig. 1). Zeolites H-Y and H–Beta with comparable Si/Al ratios
showed similar catalytic activity but zeolite H-Y was more
Catalysts were dried overnight at 200 °C before use. The aminal
solution was added to the dried catalyst in a sealed vial, later
purged with nitrogen. Typically, 200 mg of dried catalyst was used
in the experiments, representing 5 wt% with respect to the aminal
solution. The mixture was stirred and maintained at 150 °C (unless
stated otherwise) for the required reaction time. Aliquots
of ꢃ100 mL were taken at different time points, and the spent cat-
alyst was separated from the crude product solution by centrifuga-
tion. The crude product was then analysed by gas chromatography
(GC). Tetrahydrofuran (Acros Organics, 99+% extra pure) was used
to dilute the crude product solution and nitrobenzene (Fluka,
> 99.5% purity) was added as an external standard. Reaction sam-
ples were analysed in a Shimadzu 2010 gas chromatograph
equipped with a 60 m CP–Sil 5 CB column and an FID detector.
Compounds were identified with known compounds and gas chro-
matography–mass spectrometry (GC–MS). For the recycling tests,
after a reaction cycle of 24 h at 150 °C, the crude product was
extracted for analysis, and the catalysts were used in subsequent
reaction cycles without intermediate calcination.
100
80
60
40
20
0
6
5
4
3
2
1
0
1 h 5 h 24 h
1 h 5 h 24 h
1 h 5 h 24 h
1 h 5 h 24 h
MCM-22
(Si/Al = 14)
Y
Beta
(Si/Al = 12.5)
Beta
(Si/Al = 33)
2.5. Strength of aniline-catalyst interaction (temperature programmed
desorption of aniline)
(Si/Al = 15)
ABA
MDA
OMDA
4,4'/(2,2' + 2,4')
The strength of the aniline-catalyst interaction was probed by
dosing the sample with aniline and using Thermogravimetric Anal-
ysis–Mass Spectrometry (TGA-MS). The MS signal at m/z 93 was
Fig. 1. Yields of products obtained in MDA synthesis using various zeolites as
catalysts. A/F = 3.0 M, 5 wt% catalyst, 150 °C, reaction time of 1 h, 5 h, and 24 h.
Isomer ratio indicated by the crosses (ꢄ).
116

Ka Yan Cheung, C. Marquez, P. Tomkins et al.
Journal of Catalysis 400 (2021) 114–123
selective towards 4,40–MDA, as indicated by the much higher 4,40–
MDA/(2,20-MDA + 2,40-MDA) isomer ratio. Salzinger et al. reported
that the catalytic activity of materials with the same zeolite topol-
ogy can be directly correlated to the total number of Brønsted acid
sites, and pore diffusion effects can be disregarded in this case [11].
This is confirmed by the comparison of two Beta zeolites with dif-
ferent Si/Al ratios. Upon increasing the Si/Al ratio from 12.5 to 33,
the number of acid sites decreased and the catalytic activity
dropped. Zeolite H-MCM-22 was the most active catalyst of the
series, as reflected by the complete absence of the ABA intermedi-
ates after 5 h of reaction. However, the selectivity to 4,40–MDA was
very low. These results contradict previous studies that reported
H–MCM–22 to be almost inactive [11]. Nevertheless, it is impor-
tant to note that our reactions were carried out at a higher temper-
ature. Zeolite H-MCM-22 has both 10-MR and 12-MR micropores,
which restrict the diffusion of the bulky reactant and intermedi-
ates. As a result, the reaction is believed to mainly take place on
the external surface of the catalyst and at the pore openings, which
not only results in a suboptimal use of available acid sites, but also
reduces the advantageous shape selectivity effect normally
brought about by zeolites [11,18,19]. N2 physisorption studies
revealed that both the H–Y and the H–MCM-22 samples exhibit
acid-catalysed transformations [21–23], were only able to form
MDA in modest yields (37% MDA yield with Sn1.5/SiO2, entry 7).
In contrast, some of the studied first-row transition metals per-
formed better. Although Cr1.9/SiO2, Ni1.4/SiO2, and Cu1.4/SiO2
showed poor activity, this was noticeably improved when Zn1.7
/
SiO2 was used as catalyst (27% MDA yield and a 4.8 4,40–MDA/(2,
20-MDA + 2,40-MDA) isomer ratio, entry 12) and further enhanced
with Fe1.3/SiO2 (55% MDA yield and a 4.3 4,40-MDA/(2,20-MDA + 2
,40-MDA) isomer ratio, entry 9). In fact, given the availability and
low cost of Fe, this catalyst already emerges as an interesting can-
didate for the synthesis of MDA. Nonetheless, the best results were
attained when other Lewis acids were employed. Both Zr1.4/SiO2
and Hf1.5/SiO2 (entries 13 and 14) were found to be outstanding
potential catalysts for the synthesis of MDA, achieving high MDA
and OMDA yields and a respectable 4,40–MDA/(2,20-MDA + 2,40-M
DA) isomer ratio. Moreover, the activity of such catalysts can be
further improved by increasing the metal loading from 1.5
to ~ 3 wt%, as indicated by the results obtained with the sample
Hf2.7/SiO₂ (entry 15). A further increase in metal loading, however,
did not lead to a noticeable surge in product yields (Fig. S6). In
order to confirm that catalysts with exclusively Lewis acid sites
are capable of catalysing the ABA to MDA rearrangement, a small
amount of Zn was added in the preparation of the Hf– and Zr–
loaded samples. Previous research indeed showed that any residual
Brønsted acidity on Hf/SiO2 can be suppressed by the addition of
Zn, likely because the formed ZnO particles have basic properties
[24]. The acidity characteristics of the catalysts will be studied in
detail in Section 3.3. Remarkably, upon testing the Hf2.6Zn0.5/SiO₂
catalyst, both the OMDA yield and the isomer ratio increased, with
only a minor drop in MDA yield (entry 16 ). A bimetallic catalyst
with double the amount of Zn showed very similar catalytic activ-
ity (Fig. S7). Due to the superior overall MDA and OMDA yields, the
attractive 4,40–MDA/(2,20–MDA + 2,40-MDA) isomer ratio, and the
purely Lewis acidic nature of the material (see also Fig. 8 and
Table 5), the bimetallic system Hf2.6Zn0.5/SiO2 was selected for fur-
ther study.
a
type I isotherm, characteristic of microporous materials
(Figs. S4 and S5) [20], this means they would be more prone to pore
blockage by the bulky molecules and thus to fast deactivation.
Nevertheless, these samples also possess a relatively high external
surface area, as observed in Table S1, which contributes to the
moderate catalytic activity exhibited by the zeolites. Furthermore,
as explained earlier, for almost every studied zeolite, the 4,40–MD
A/(2,20-MDA + 2,40-MDA) isomer ratio drops as the reaction
proceeds.
3.2. Catalytic reactions using Lewis acid catalysts
A series of metal loaded SiO2 catalysts was evaluated to test the
hypothesis that Lewis acid solids can catalyse the formation of
MDA. The results are presented in Table 2 and Table S2. Blank
experiments did not form detectable amounts of any products,
not even ABA intermediates; unloaded SiO2 only produced MDA
in a 10% yield after 24 h reaction time (entries 1 and 2 respec-
tively). Three rare-earth elements were studied; La, Ce, and Sc.
La1.5/SiO2 and Ce1.7/SiO2 only showed a moderate activity (30%
and 25% MDA yield, entries 3 and 4), this activity was improved
with Sc1.5/SiO2 (entry 5). A transition metal like Ta and a post-
transition metal like Sn, which are often used in other Lewis
Using Hf2.6Zn0.5/SiO2 as catalyst, the reaction temperature was
varied to evaluate its influence on the product distribution; the
results are plotted in Fig. 2. At 125 °C, the ABA intermediates frac-
tion remained large at short reaction times; it was still present
even after 24 h (13%) (Fig. 2a). The 4,40-MDA yield increased over
time to 38% after 24 h, which was comparable to values reached
at higher reaction temperatures. Although the yield of the unde-
sired 2,20-MDA was favourably low (<0.2%, 5 h) at 125 °C compared
to all the temperatures tested, and the 4,40-MDA/(2,20-MDA + 2,40-
Table 2
Yields and isomer ratios of the MDA synthesis products over various metal-loaded silica catalysts (5 wt% catalyst, 24 h, 150 °C).
Yield (%)
Entry
Catalyst
ABA
MDA
OMDA
4,40/(2,20+2,40)
1
Blank
/
/
/
/
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
SiO₂
22.8
13.7
16.5
3.8
17.7
4.7
25.2
6.1
34.3
36.2
27.0
2.1
10.0
30.1
25.0
48.8
21.3
37.3
11.2
55.4
5.5
/
3.6
/
10.8
/
5.1
/
10.8
/
/
3.6
4.6
3.8
4.0
4.2
4.1
3.6
4.3
3.5
1.9
4.8
3.4
4.0
3.4
4.0
3.0
La1.5/SiO₂
Ce1.7/SiO₂
Sc1.5/SiO₂
Ta1.3/SiO₂
Sn1.5/SiO₂
Cr1.9/SiO₂
Fe1.3/SiO₂
Ni1.4/SiO₂
Cu1.4/SiO₂
Zn1.7/SiO₂
Zr1.4/SiO₂
Hf1.5/SiO₂
Hf2.7/SiO₂
Hf2.6Zn0.5/SiO₂
Zr1.9Zn0.3/SiO₂
3.3
26.7
53.8
46.5
52.7
51.5
52.3
/
12.9
15.1
14.6
16.5
15.4
5.9
3.1
3.8
3.6
117

Ka Yan Cheung, C. Marquez, P. Tomkins et al.
Journal of Catalysis 400 (2021) 114–123
50
50
40
30
20
10
0
10
(a)
(b)
(d)
125 °C
40
30
20
10
0
8
150 °C
6
175 °C
200 °C
125 °C
150 °C
175 °C
200 °C
4
2
0
0
5
10
15
20
25
0
5
10
15
20
25
Time (h)
Time (h)
(c)
50
40
30
20
10
0
6
5
4
3
2
1
0
0
5
10
15
20
25
0
5
10
15
20
25
Time (h)
Time (h)
Fig. 2. Temperature profiles of MDA synthesis using 5 wt% Hf2.6Zn0.5/SiO2 as catalyst; values at zero were omitted for clarity. (a) ABA yield, (b) 4,40-MDA and 2,20-MDA yields,
(c) OMDA yield, and (d) 4,40-MDA/(2,20-MDA + 2,40-MDA) isomer ratio.
MDA) isomer ratio was at a high average of 5.5, the OMDA yield
was only at a low 7%. At higher temperatures of 175 °C and
200 °C, ABAs were swiftly converted (Fig. 2a). The highest 4,40-
MDA and OMDA yields (Fig. 2b and c) were reached at 175 °C,
but the increased 2,20–MDA formation and decreased 4,40-MDA/(
2,20-MDA + 2,40-MDA) isomer ratio indicate that 175 °C may be
suboptimal for this reaction. At 200 °C, the 4,40-MDA yield was sig-
nificant (36%) after 1 h, but it dropped over time, contrary to the
positive trends observed at lower temperatures. Furthermore, both
2,40-MDA and OMDA yields plateaued after 5 h. These trends sug-
gest that the catalyst is more prone to deactivation at 200 °C, while
also the higher 2,20-MDA yield is undesired. The 4,40-MDA/(2,20-
MDA + 2,40-MDA) isomer ratio also dropped as the reaction pro-
ceeded at 175 °C and 200 °C, unlike the more stable ratio seen at
125 °C and 150 °C. In fact, after 24 h at 150 °C, the 2,20-MDA yield
remained low (<0.7%), the 4,40-MDA and OMDA yields reached
appreciable values of 41% and 17%, and the 4,40-MDA/(2,20-MDA +
2,40-MDA) isomer ratio stayed at a satisfactory average of 4.1. Con-
sequently, 150 °C was chosen as the reaction temperature for sub-
sequent reactions. Similar effects of increased temperature, e.g.
higher OMDA yields, and decreased 4,40-MDA/(2,20-MDA + 2,40-M
DA) isomer ratios, have been reported by others using zeolite cat-
alysts [15,25].
To improve the catalytic activity, alternative supports were
tested (Fig. 3). First, a magnesium aluminate spinel (MgAl2O4)
was studied as a potential basic support [26]. However, no signifi-
cant activity was observed with Hf on spinel, and not even the ABA
intermediates were obtained after 1 h. This is likely due to the
100
80
60
40
20
0
6
5
4
3
2
1
0
1 h 5 h 16 h
Al₂MgO₄
1 h 5 h 16 h
HAP
1 h 5 h 16 h
TiO₂
1 h 5 h 16 h
ZrO₂
1 h 5 h 16 h
Al₂O₃
1 h 5 h 16 h
SiO₂
1 h 5 h 16 h
SiO₂·Al₂O₃
1 h 5 h 16 h
1 h 5 h 16 h
MCM-22
Y
(Si/Al = 14)
(Si/Al = 15)
ABA
MDA
OMDA
4,4'/(2,2' + 2,4')
Fig. 3. Product yields obtained in MDA synthesis using Hf loaded at 3 wt% on various supports; 150 °C, 5 wt% catalyst. 4,40-MDA/(2,20-MDA + 2,40-MDA) isomer ratio indicated
by the crosses (ꢄ).
118

Ka Yan Cheung, C. Marquez, P. Tomkins et al.
100
Journal of Catalysis 400 (2021) 114–123
5
4
3
2
1
0
80
60
40
20
0
0.5 h 1 h
5 h
0.5 h 1 h
SiO₂·Al₂O₃
MDA OMDA
5 h
0.5 h 1 h
Hf₂.₇Zn₀.₅/SiO₂·Al₂O₃
4,4'/(2,2' + 2,4')
5 h
Hf₂.₆Zn₀.₅/SiO₂
ABA
Fig. 4. Yields of products obtained in MDA synthesis using Hf2.6Zn0.5/SiO2, SiO2ꢂAl2O3, and Hf2.7Zn0.5/SiO2ꢂAl2O3. 1 wt% catalyst with respect to starting aminal solution, 150 °C.
Isomer ratio indicated by the crosses (ꢄ).
improvement was seen when supports with some Brønsted acid
Table 3
character such as alumina (Al2O3) and silica-alumina (SiO2ꢂAl2O3)
Initial rates of MDA formation using 1 wt% catalyst with respect to starting aminal
were employed. Particularly silica-alumina produced a marked
solution (conditions same as in Fig. 4).
increase in the catalytic activity, with much higher MDA and
Catalyst
Initial rate of MDA formation (mmolMDA/gcatalyst/h)
OMDA yields obtained after just 1 h of reaction. Complete conver-
sion was achieved as indicated by the absence of the ABA interme-
diates. Lastly, since zeolites like H-MCM-22 and H-Y showed
promising activity (Section 3.1), they were also tested as supports.
For Hf3.0/MCM-22, this resulted in slightly higher ABA and OMDA
contents, similar MDA yields, and a slightly higher 4,40-MDA/(2,20
-MDA + 2,40-MDA) isomer ratio, as opposed to values reached by
the unloaded H-MCM-22 zeolite. For Hf3.0/Y, all values were very
similar to those obtained with the metal-free parent zeolite, sug-
gesting that metal addition, at least in the case of Hf, does not
change the catalytic behaviour of zeolite H-Y. Based on these
results, silica-alumina appears to be the best potential support
for the Lewis acid-catalysed synthesis of MDA.
Hf2.6Zn0.5/SiO2
SiO2ꢂAl2O3
5.7
22
153
Hf2.7Zn0.5/SiO2ꢂAl2O3
basic behaviour of the spinel, suppressing the Lewis acidity.
Similar results were obtained using hydroxyapatite (HAP,
Ca10(PO4)6(OH)2), which has more weakly basic properties [27].
With Hf3.0/HAP as catalyst, the aminal conversion was incomplete,
with an insignificant amount of MDA produced (<13%) even after
16 h. Next, two Lewis acidic metal oxides were tested, TiO2 and
ZrO2 [28]. With TiO2 as the support, a high isomer ratio was
obtained after 5 h. However, the MDA yield was low even after
16 h (<17%). The results obtained over Hf3.0/ZrO2 were similar to
those obtained with Hf3.0/TiO2, but with a lower 4,40–MDA/
Temperature variation for the Hf2.7Zn0.5/SiO2ꢂAl2O3 catalyst
showed similar trends as for Hf2.6Zn0.5/SiO2, with the best 4,40-
MDA and OMDA yields obtained at 150 °C, together with a satisfac-
tory high product isomer ratio (Fig. S8).
(2,20-MDA
+
2,40-MDA) isomer ratio. However,
a
significant
B+L
a
L
b
B
B
L
a
b
c
c
100
200
300
400
500
1700
1600
1500
1400
Wavenumber (cm-1)
Temperature (°C)
Fig. 5. NH3-TPD profiles, normalised to the mass of sample (left) and difference IR spectra of adsorbed pyridine, normalised to 10 mg of sample/cm2 (right) of (a) zeolite H-Y,
Si/Al = 15, (b) Hf2.7Zn0.5/SiO2ꢂAl2O3, and (c) Hf2.6Zn0.5/SiO2. Absorption bands corresponding to Brønsted (B) and Lewis (L) acid sites are marked.
119

Ka Yan Cheung, C. Marquez, P. Tomkins et al.
Journal of Catalysis 400 (2021) 114–123
Table 4
port, and of Lewis acidity provided by Hf is beneficial for creating
an active catalyst for MDA synthesis.
Acidity measurements of selected samples by NH3-TPD and FTIR spectroscopy using
pyridine as probe molecule.
Sample
NH3 desorbeda (mmol/g)
Acidity (mmol/g)b
3.3. Relation of catalytic activity with acidity characteristics and
mechanistic implications
Brønsted
Lewis
H-Y (Si/Al = 15)
Hf2.6Zn0.5/SiO2
Hf2.7Zn0.5/SiO2ꢂAl2O3
0.43
0.16
0.46
200
n.dc
22
57
62
62
Two materials were selected for detailed characterisation, viz.
Hf2.6Zn0.5/SiO2 and Hf2.7Zn0.5/SiO2ꢂAl2O3. FTIR monitoring of
adsorbed pyridine probe molecules shows that Hf2.6Zn0.5/SiO2
indeed exclusively contains Lewis acid sites (Fig. 5, right; Table 4)
[24]. In the NH3-TPD profile (Fig. 5, left), this corresponds to a des-
orption of ammonia at rather low temperatures. In contrast, both
Lewis and Brønsted acid sites are found on the Hf2.7Zn0.5/SiO2ꢂAl2-
O3 catalyst. Comparison of the latter material with H-Y in the NH3-
TPD data (Fig. 5, left) proves that Hf2.7Zn0.5/SiO2ꢂAl2O3 contains a
larger number of weakly acid Brønsted sites than H-Y.
The activity of the purely Lewis acidic Hf2.6Zn0.5/SiO2 for MDA
formation is unusual, considering that the search for solid alterna-
tives for HCl in MDA formation has strongly focused on Brønsted
acid zeolites [10–15]. Clearly, Lewis acids are also capable of cata-
lysing the rearrangement of the ABA intermediates to MDA. Such
rearrangements, bringing an alkyl substituent from the N-atom
of an aniline to the aromatic ring, are known as Hofmann-
a
Number of acid sites determined using NH3-TPD.
Acidity determined by FTIR spectroscopy using pyridine as probe molecule after
evacuation at 150 °C.
b
c
Not detected, detection limit <2 mmol/g.
Table 5
Textural properties of fresh and spent zeolite H-Y (Si/Al = 15) and Hf2.7Zn0.5/SiO2-
ꢂAl2O3 determined by N2 physisorption at ꢁ196 °C. Spent catalysts underwent a 24 h
reaction.
Sample
SBET
Vmicro
Vmeso
(m2/g)a
(cm3/g)b
(cm3/g)c
H-Y (Si/Al = 15)
870
61
303
226
0.262
0.011
0.0086
negligible
0.27
0.12
0.94
0.68
H-Y (Si/Al = 15), spent, washed
Hf2.7Zn0.5/SiO2ꢂAl2O3
Hf2.7Zn0.5/SiO2ꢂAl2O3, spent, washed
a
Martius rearrangements [10]. Notably, using N-a-phenethyl-
BET p/p0 range 0.001 – 0.07 for H-Y, 0.001 – 0.10 for washed H-Y, 0.001 – 0.20
aniline as a reactant, it has been shown that not only HCl, but also
ZnCl2 can effectively promote the rearrangement to ortho- and
for Hf2.7Zn0.5/SiO2ꢂAl2O3, and 0.001 – 0.30 for spent Hf2.7Zn0.5/SiO2ꢂAl2O3.
b
Micropore volume obtained using the t–plot method.
c
Mesopore volume estimated from BJH desorption.
para-
a -phenethylaniline. In fact, higher para-selectivities were
even observed with ZnCl2 than with HCl. Based on the mechanism
by Hart and Kosak [29], a proposal for the mechanism of ABA rear-
rangement on the Hf Lewis acid sites in Hf2.6Zn0.5/SiO2 is shown in
Scheme 2. The first step is coordination of the ABA reactant on the
Lewis acid site (I in Scheme 2), followed by formation of a well-
stabilised benzylic carbenium ion, together with a Hf-coordinated
anilide (II). Subsequently, an electrophilic aromatic substitution
takes place to form complex IV. Aniline-TPD data is used to prove
that Hf2.6Zn0.5/SiO2 is capable of activating anilines, e.g. complex I
in Scheme 2 (Fig. 6). The profile for Hf2.6Zn0.5/SiO2 shows the same
fast desorption of weakly physisorbed aniline as the SiO2 support,
but additionally, a distinct desorption of aniline is seen between
To study in more detail the catalysts combining both Brønsted
and Lewis acid sites, reactions were repeated with reduced catalyst
amounts, and the results are presented as the initial rates of MDA
formation (Fig. 4 and Table 3). The Hf2.7Zn0.5/SiO2ꢂAl2O3 catalyst
produces MDA at a rate which is significantly higher than that of
the individual Hf2.6Zn0.5/SiO2 or SiO2ꢂAl2O3 catalysts. Moreover,
the rate with Hf2.7Zn0.5/SiO2ꢂAl2O3 is also much higher than the
sum of the rates obtained with the Lewis acidic Hf2.6Zn0.5/SiO2
and with the Brønsted acidic SiO2ꢂAl2O3. Evidently, the combined
presence of Brønsted acidity provided by the silica-alumina sup-
Scheme 2. Proposed mechanism for the formation of MDA from p–ABA, using a catalyst with Hf–centred Lewis acidity.
120

Ka Yan Cheung, C. Marquez, P. Tomkins et al.
Journal of Catalysis 400 (2021) 114–123
adsorption of the weakly basic –NH2 groups on the strong acid
sites of the zeolite. Physicochemical proof for the strong binding
of anilines on H-Y was already provided by the aniline-TPD
experiments in Fig. 6. Pore blocking is the major factor con-
tributing to zeolite deactivation, as proven by the physisorption
measurements of Fig. 8 and Table 5: the N2 isotherm shows
that the microporosity is largely lost, with a reduction of the
micropore volume from 0.262 cm3/g to 0.011 cm3/g after
reaction.
Hf Zn/SiO2
Hf Zn/SiO2.Al2O3
SiO2
SiO2.Al2O3
H-Y
In contrast with zeolite H–Y, Hf2.6Zn0.5/SiO₂ was significantly
more stable in consecutive runs (Fig. 7). Despite a gradual drop
in catalytic activity after each run, the extent of deactivation was
far less significant; the isomer ratio hovered around 4.2 during
all five runs. Catalyst stability and recyclability were further
improved when the silica support was replaced by silica-alumina
(sample Hf2.7Zn0.5/SiO2ꢂAl2O3) (Fig. 7). The initial activity was
maintained for five cycles and the isomer ratio remained around
3.4 throughout all five cycles. Complete conversion was obtained
and the OMDA fraction remained high (around 20%) in all cycles.
The isotherm and mesopore volume derived from N2 physisorption
of a solvent-washed, spent Hf2.7Zn0.5/SiO2ꢂAl2O3 catalyst after one
24 h reaction showed retention of the porosity, suggesting that
the mesoporous sample was less prone to catalyst deactivation
(Fig. 8, Table 5). Notably, in the aniline-TPD profile recorded for
Hf2.7Zn0.5/SiO2ꢂAl2O3, the desorption occurs at the highest rates
between 150 and 260 °C. This temperature range is much lower
than that observed for H-Y, which displays maxima in the desorp-
tion profile at 220 and 390 °C (Fig. 6). Thus, in comparison with H-
Y, both the larger pore diameters and the weaker interaction with
anilines contribute to the lower susceptibility of Hf2.7Zn0.5/SiO2-
ꢂAl2O3 to deactivation.
0
100
200
300
400
500
600
Temperature (°C)
Fig. 6. Temperature programmed desorption of aniline for Hf2.6Zn0.5/SiO2, Hf2.7
-
Zn0.5/SiO2ꢂAl2O3, SiO2, SiO2ꢂAl2O3, and zeolite H–Y, investigated by monitoring the
m/z value of 93 (corresponding to aniline, C6H5NH2). MS peak intensity normalised
to the sample mass.
150 and 300 °C, evidencing the interaction of anilines with the
Lewis acid sites.
The mechanism as proposed in Scheme 2 also suggests how the
additional presence of Brønsted acid sites could be beneficial for
the strongly increased activity of Hf2.7Zn0.5/SiO2ꢂAl2O3. Indeed,
the desorption of the ring-alkylated product requires a reprotona-
tion of intermediates IV/V, and the availability of mild proton acid-
ity can enhance this step, leading to the formation of MDA and
regeneration of the LAS.
Finally, the metal contents of fresh and spent Hf2.6Zn0.5/SiO2-
ꢂand Hf2.7Zn0.5/SiO2ꢂAl2O3 catalysts were compared by ICP-OES.
The metal contents were unchanged after a 48 h reaction, implying
that there is no undesired metal leaching. For the most active cat-
alyst, viz. Hf2.7Zn0.5/SiO2ꢂAl2O3, EDXS elemental maps were
recorded, both on a fresh and a spent catalyst. These maps showed
homogeneous dispersions of Hf and Zn on the silica–alumina sup-
port. After 24 h of reaction at 150 °C and subsequent calcination at
500 °C, the two metals remained in close contact with each other
and also with the support, without apparent aggregation of metal
particles. This stable dispersion is believed to contribute to the sta-
bility of this catalyst (Fig. 9).
3.4. Catalyst recycling and stability
The stability and reusability of selected samples were then
investigated by recycling tests (Fig. 7). The product yields were
obtained without intermediate regeneration of the catalyst after
each cycle. Using zeolite H-Y, the catalytic activity plummeted
after just one 24 h cycle; also the high isomer ratio obtained
in the first run was lost in following runs. This is due to the
microporous nature of the zeolite, together with the strong
100
80
60
40
20
0
5
4
3
2
1
0
1
2
3
Y
4
5
1
2
3
4
5
1
2
3
4
5
Hf₂.₆Zn₀.₅/SiO₂
Hf₂.₇Zn₀.₅/SiO₂·Al₂O₃
(Si/Al = 15)
ABA
MDA
OMDA
4,4'/(2,2' + 2,4')
Fig. 7. Catalyst recycling experiments for acid-catalysed MDA synthesis at 150 °C (24 h reaction time).
121

Ka Yan Cheung, C. Marquez, P. Tomkins et al.
Journal of Catalysis 400 (2021) 114–123
400
600
500
400
300
200
100
0
300
200
100
0
0.0
0.2
0.4
0.6
0.8
1.0
0.0
0.2
0.4
0.6
0.8
1.0
Relative pressure (p/p0)
Relative pressure (p/p0)
Fig. 8. Nitrogen physisorption isotherms of (left) zeolite H–Y (Si/Al = 15) (j/h), spent H–Y (Si/Al = 15) (▲/4), (right) Hf2.7Zn0.5/SiO2ꢂAl2O3 (d/s), spent Hf2.7Zn0.5/SiO2ꢂAl2O3
(./5). Filled symbols denote adsorption and open symbols denote desorption.
Fig. 9. HAADF–STEM image (a) and EDXS maps (b), (c), (d) of Hf2.7Zn0.5/SiO2ꢂAl2O3. HAADF–STEM image (e) and EDXS maps (f), (g), (h) of calcined, spent Hf2.7Zn0.5/SiO2ꢂAl2O3
after 24 h reaction of MDA synthesis at 150 °C. Green spots are Hf and blue spots are Zn. (For interpretation of the references to colour in this figure legend, the reader is
referred to the web version of this article.)
of Hf, Zn on silica-alumina contribute to its resistance to
deactivation.
4. Conclusions
In the search for
a viable heterogeneous alternative to
hydrochloric acid in the industrial synthesis of MDA, various Lewis
acid catalysts were tested. Among others, catalysts containing Hf
on silica display satisfactory activity. Addition of a small amount
of Zn to Hf resulted in a catalyst with exclusively Lewis acid sites,
proving that purely Lewis acidic materials also catalyse MDA for-
mation. Rates were strongly increased when SiO2 as a support
was replaced by SiO2.Al2O3. Catalyst stability tests showed that
the Hf, Zn on silica-alumina catalyst could be used in five consecu-
tive runs without any significant drop in performance, while an H-
Y zeolite readily loses its activity under identical reaction condi-
tions. Both the mesoporosity and the moderately strong acidity
Declaration of Competing Interest
The authors declare that they have no known competing finan-
cial interests or personal relationships that could have appeared
to influence the work reported in this paper.
Acknowledgements
Support from BASF SE under the INCOE framework (Interna-
tional Network of Centers of Excellence) is acknowledged. The
122

Ka Yan Cheung, C. Marquez, P. Tomkins et al.
Journal of Catalysis 400 (2021) 114–123
[15] P. Botella, A. Corma, R.H. Carr, C.J. Mitchell, Towards an industrial synthesis of
diamino diphenyl methane (DADPM) using novel delaminated materials: A
breakthrough step in the production of isocyanates for polyurethanes, Appl.
authors thank Philipp Mueller (BASF SE) and Matthias Surau (BASF
SE) for the TEM measurements.
Catal.
A
Gen.
398
(2011)
143–149,
https://doi.org/10.1016/j.
apcata.2011.03.026.
Appendix A. Supplementary data
[16] A. Corma, H. García, Lewis acids: From conventional homogeneous to green
homogeneous and heterogeneous catalysis, Chem. Rev. 103 (2003) 4307–
4365, https://doi.org/10.1021/cr030680z.
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.jcat.2021.05.031.
[17] C.A. Emeis, Determination of Integrated Molar Extinction Coefficients for
Infrared Absorption Bands of Pyridine Adsorbed on Solid Acid Catalysts, J.
Catal. 141 (1993) 347–354, https://doi.org/10.1006/jcat.1993.1145.
[18] A. Corma, C. Corell, F. Llopis, A. Martínez, J. Pérez-Pariente, Proposed pore
volume topology of zeolite MCM-22 based on catalytic tests, Appl. Catal. A,
Gen. 115 (1994) 121–134, https://doi.org/10.1016/0926-860X(94)80382-X.
[19] S.L. Lawton, M.E. Leonowicz, R.D. Partridge, P. Chu, M.K. Rubin, Twelve-ring
pockets on the external surface of MCM-22 crystals, Microporous Mesoporous
Mater. 23 (1998) 109–117, https://doi.org/10.1016/S1387-1811(98)00057-2.
[20] M. Thommes, K. Kaneko, A.V. Neimark, J.P. Olivier, F. Rodriguez-Reinoso, J.
Rouquerol, K.S.W. Sing, Physisorption of gases, with special reference to the
evaluation of surface area and pore size distribution (IUPAC Technical Report),
Pure Appl. Chem. 87 (2015) 1051–1069, https://doi.org/10.1515/pac-2014-
1117.
References
[1] Kunststoffe International, https://en.kunststoffe.de/a/article/article-245437,
accessed April 2020.
[2] Grand
View
Research,
https://www.grandviewresearch.com/industry-
analysis/polyurethane-pu-market, accessed April 2020.
[3] S. Lowenkron, Amines, Aromatic, Methylenedianiline, in: Kirk-Othmer Encycl.
Chem. Technol., John Wiley & Sons, Inc., Hoboken, NJ, USA, 2000. https://doi.
org/10.1002/0471238961.1305200812152305.a01.
[4] A. De Angelis, P. Ingallina, C. Perego, Solid Acid Catalysts for Industrial
Condensations of Ketones and Aldehydes with Aromatics, Ind. Eng. Chem. Res.
43 (2004) 1169–1178, https://doi.org/10.1021/ie030429+.
[5] G. Bellussi, Zeolite catalysts for the production of chemical commodities: BTX
derivatives, in: Stud. Surf. Sci. Catal., 2004: pp. 53–65. https://doi.org/10.1016/
S0167-2991(04)80785-7.
[21] A. Corma, M.E. Domine, S. Valencia, Water-resistant solid Lewis acid catalysts:
Meerwein-Ponndorf-Verley and Oppenauer reactions catalyzed by tin-beta
zeolite, J. Catal. 215 (2003) 294–304, https://doi.org/10.1016/S0021-9517(03)
00014-9.
[22] E. Nikolla, Y. Román-Leshkov, M. Moliner, M.E. Davis, ‘‘One-pot” synthesis of 5-
(hydroxymethyl)furfural from carbohydrates using tin-beta zeolite, ACS Catal.
1 (2011) 408–410, https://doi.org/10.1021/cs2000544.
[23] B. Tang, W. Dai, G. Wu, N. Guan, L. Li, M. Hunger, Improved postsynthesis
strategy to Sn-beta zeolites as Lewis acid catalysts for the ring-opening
[6] J. Nafziger, L. Rader, I. Seward Jr., US Patent 4 554 378 (1985), to The Dow
Chemical Company.
[7] F. Frulla, A. Sayigh, H. Ulrich, P. Whitman, US Patent 4 092 343 (1978), to the
Upjohn Company.
[8] C. Perego, A. de Angelis, A. Carati, C. Flego, R. Millini, C. Rizzo, G. Bellussi,
Amorphous aluminosilicate catalysts for hydroxyalkylation of aniline and
phenol, Appl. Catal. A Gen. 307 (2006) 128–136, https://doi.org/10.1016/j.
apcata.2006.03.013.
[9] T.C. Keller, J. Arras, M.O. Haus, R. Hauert, A. Kenvin, J. Kenvin, J. Pérez-Ramírez,
Synthesis-property-performance relationships of amorphous silica-alumina
catalysts for the production of methylenedianiline and higher homologues, J.
Catal. 344 (2016) 757–767, https://doi.org/10.1016/j.jcat.2016.08.016.
[10] M. Salzinger, J.A. Lercher, Reaction network and mechanism of the synthesis of
methylenedianiline over dealuminated Y-type zeolites, Green Chem. 13 (2011)
149–155, https://doi.org/10.1039/c0gc00428f.
[11] M. Salzinger, M.B. Fichtl, J.A. Lercher, On the influence of pore geometry and
acidity on the activity of parent and modified zeolites in the synthesis of
methylenedianiline, Appl. Catal. A Gen. 393 (2011) 189–194, https://doi.org/
10.1016/j.apcata.2010.11.040.
[12] D. Jin, P. Wu, M. Zhang, Q. Wang, H. Jin, H. Ge, X. Wang, Z. Wang, H. Lou,
Dramatic change of methylenedianiline activity and selectivity in different
pore geometry of zeolites, Micropor. Mesopor. Mater. 233 (2016) 109–116,
https://doi.org/10.1016/j.micromeso.2015.12.012.
hydration of epoxides, ACS Catal.
10.1021/cs500891s.
4 (2014) 2801–2810, https://doi.org/
[24] T. De Baerdemaeker, M. Feyen, U. Müller, B. Yilmaz, F.S. Xiao, W. Zhang, T.
Yokoi, X. Bao, H. Gies, D.E. De Vos, Bimetallic Zn and Hf on silica catalysts for
the conversion of ethanol to 1,3-butadiene, ACS Catal. 5 (2015) 3393–3397,
https://doi.org/10.1021/acscatal.5b00376.
[25] T. Kugita, S. Hirose, S. Namba, Catalytic activity of zeolites for synthesis
reaction of methylenedianiline from aniline and formaldehyde, Catal. Today.
111 (2006) 275–279, https://doi.org/10.1016/j.cattod.2005.10.036.
[26] I. Ganesh, A review on magnesium aluminate (MgAl2O4) spinel: Synthesis,
processing and applications, Int. Mater. Rev. 58 (2013) 63–112, https://doi.org/
10.1179/1743280412Y.0000000001.
[27] K. Mori, T. Hara, T. Mizugaki, K. Ebitani, K. Kaneda, Hydroxyapatite-supported
palladium nanoclusters: A highly active heterogeneous catalyst for selective
oxidation of alcohols by use of molecular oxygen, J. Am. Chem. Soc. 126 (2004)
10657–10666, https://doi.org/10.1021/ja0488683.
[28] M.I. Zaki, M.A. Hasan, F.A. Al-Sagheer, L. Pasupulety, In situ FTIR spectra of
pyridine adsorbed on SiO2-Al2O3, TiO2, ZrO2 and CeO2: General considerations
for the identification of acid sites on surfaces of finely divided metal oxides,
Colloids Surfaces A Physicochem. Eng. Asp. 190 (2001) 261–274, https://doi.
org/10.1016/S0927-7757(01)00690-2.
[13] T.C. Keller, J. Arras, S. Wershofen, J. Pérez-Ramírez, Design of hierarchical
zeolite catalysts for the manufacture of polyurethane intermediates, ACS Catal.
5 (2015) 734–743, https://doi.org/10.1021/cs5017694.
[14] A. Corma, P. Botella, C. Mitchell, Replacing HCl by solid acids in industrial
processes: synthesis of diamino diphenyl methane (DADPM) for producing
polyurethanes, Chem. Commun. (2004) 2008–2010. https://doi.org/10.1039/
b406303a.
[29] H. Hart, J.R. Kosak, Mechanism of rearrangement of N-alkylanilines, J. Org.
Chem. 27 (1962) 116–121, https://doi.org/10.1021/jo01048a029.
123