You are on page 1of 9

Journal of Molecular Catalysis A: Chemical 363–364 (2012) 371–379

Contents lists available at SciVerse ScienceDirect

Journal of Molecular Catalysis A: Chemical


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

Lewis and Brönsted acidic sites in M4+ -doped zeolites (M = Ti, Zr, Ge, Sn, Pb) as
well as interactions with probe molecules: A DFT study
Gang Yang a,c,∗ , Lijun Zhou a , Xiuwen Han b
a
Engineering Research Center of Forest Bio-preparation, Ministry of Education, Northeast Forestry University, Harbin 150040, China
b
State Key Laboratory of Catalysis, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, China
c
Department of Chemical Engineering and Chemistry, Eindhoven University of Technology, P.O. Box 513, NL-5600 MB Eindhoven, The Netherlands

a r t i c l e i n f o a b s t r a c t

Article history: Tetravalent-ion (M4+ )-doped zeolites show excellent performances for a variety of catalytic processes,
Received 23 January 2012 including the focusing biomass conversions. In this work, density functional calculations were per-
Received in revised form 8 July 2012 formed to probe the Lewis and Brönsted acidities of various M4+ -doped zeolites as well as to study
Accepted 12 July 2012
interactions with probe molecules. The Lewis and Brönsted acidities increase in the orders of Silicalite-
Available online 21 July 2012
1  Ge < Ti < Pb < Sn < Zr and Silicalite-1  Ti < Ge < Zr ≈ B < Pb < Sn < Al, respectively. The Lewis acidities
should be defined as the local sites around the M4+ ions, explaining why the adsorption energies give a
Keywords:
more consistent order with LUMO energies and absolute electronegativity rather than fukui functions.
Brönsted acidity
Lewis acidity
The formation of Brönsted acidic sites is facilitated by doping with M4+ ions. Albeit the Brönsted acidities
Density functional calculations of these M4+ -doped zeolites change greatly with Sn being the strongest, their strengths are far below
Adsorption energy that of Al3+ . The interactions of five probe molecules of changing basicities with the Brönsted acidic sites
Probe molecules indicate that the formations of covalent and/or ionic structures are the proton-competing results: the
covalent and ionic structures co-exist only for trimethyphosphine and pyridine of comparable basicity;
otherwise, proton transfer will take place and result in only the ionic or covalent structures. The proton
affinity fails to predict the Brönsted acidity of Zr and is evidenced, especially by formation of the covalent
structure during pyridine adsorption. Thus, this work presents a dynamic image of acid–base interactions
and aids our understanding toward the catalysis of solid-state acids.
© 2012 Elsevier B.V. All rights reserved.

1. Introduction zeolites [7–10]. Instead, the studies with incorporation of tetrava-


lent ions (M4+ ) lag far behind [11–17]. The discovery and
In the past few decades, the isomorphous substitutions of var- application of titanium silicalite-1 (TS-1 zeolite) in 1980s have been
ious ions into silica lattices have become an emerging topic in considered as one milestone in heterogeneous catalysis [11,12].
materials and catalysis fields [1–22]. The incorporation of trivalent Recently, Sn-doped zeolites have shown excellent catalytic perfor-
ion (M3+ ), such as Al, B or Fe, will create one negative charge in the mances for biomass conversions [25,26], which spurs our interest
zeolite lattice, which is balanced by a counterion, usually proton to investigate the acidity and catalysis of M4+ -doped zeolites.
(H+ ). The bridging hydroxyl group ( Si OH M3+ ) is the source of In the previous works [27–30], we studied the Lewis and Brön-
Brönsted acidity. As a matter of fact, zeolites are known to be the sted acidities of TS-1 and Ti-MCM-22 zeolites, see Scheme 1. The
earliest solid-state acids [23]. The largest user of zeolites is fluid cat- Lewis acidities of Ti4+ -doping zeolites have also received attention
alytic cracking (FCC), where the global consumption approximates from other groups [31–35], while few reports have been given on
five million tones per year [24]. other tetravalent ions (M4+ ). To the best of our knowledge, only
A number of experimental and theoretical efforts have devoted several publications from Pal’s group are related with the Lewis
to clarifying the acidities of M3+ -doped zeolites [4–10]. Calcula- acidity of Sn-Beta zeolite [35,36]. Obviously, it lacks a clear and
tions of various cluster models and theoretical levels indicated that systematic understanding on the Lewis acidities of tetravalent-ion
the Brönsted acidities increase in the order B < Fe < Ga < Al-doped (M4+ )-doping zeolites, which will be attempted here by considering
the Ti4+ , Zr4+ , Ge4+ , Sn4+ and Pb4+ ions. Then interactions of them
with ammonia (NH3 ) were studied, helping us to clarity the catal-
ysis of Lewis acids and the differences among the various doping
∗ Corresponding author at: Engineering Research Center of Forest Bio-
zeolites.
preparation, Ministry of Education, Northeast Forestry University, Harbin 150040,
The other object of this work is to clarify the Brönsted acid-
China. Tel.: +86 451 82192223; fax: +86 451 82102082.
E-mail address: dicpgy@yahoo.com (G. Yang). ity of zeolites with incorporation of tetravalent ions (M4+ ), which

1381-1169/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molcata.2012.07.013
372 G. Yang et al. / Journal of Molecular Catalysis A: Chemical 363–364 (2012) 371–379

Scheme 1. The Lewis acidity (LA) and Brönsted acidity (BA) created by doping the M4+ ion in Silicalite-1.

remains largely unknown in contrast to trivalent ions (M3+ ). The 3. Results and discussion
Brönsted acidities of zeolites with incorporation of Ti4+ , Zr4+ , Ge4+ ,
Sn4+ and Pb4+ ions were calculated and compared with each other. 3.1. The Lewis acidity
These Brönsted acidic sites were then to interact with five different
probe molecules (pyridine, trimethyphosphine, water, ammonia The Lewis acidic sites in tetravalent-ion (M4+ )-doped zeolites
and acetone). To aid the understanding of the Brönsted acidity of are shown in Scheme 1a and Fig. 1a, where the M4+ ions are
zeolites with incorporation of tetravalent ions (M4+ ), parallel cal-
culations were performed on two M3+ ions (B and Al). The basicities
of the five probe molecules change within a wide range and thus
present a dynamic image of acid–base interactions.

2. Computational

In agreement with the previous studies [27–29], the cluster


models containing 17-T sites were chosen to represent the local
structures of MFI-type zeolite (Fig. 1). The doped tetravalent ions
(M4+ ) occupied one of the T7 site, which were supported by the neu-
tron diffraction experiments of TS-1 zeolite [15,16] and partially by
the calculations [33,37–39]. The boundary Si atoms were saturated
with H atoms with the Si H distances of 1.500 Å and kept at their
crystallographic coordinates.
All the calculations were performed using B3LYP density
functional [40,41], Gaussian03 program suite [42]. The standard
6-31G(d) basis set was applied to the elements with atomic
numbers below 21. As to Ti, Ge, Sn, Pb and Zr, their core
electrons were represented by LANL2DZ effective core poten-
tial (ECP) [43,44] and valence electrons by LANL2DZ basis set,
respectively [17,27–29]. The combined basis sets are denoted
as B1 (default). For the adsorption of NH3 on the Lewis acidic
sites (Scheme 1a), the Wiberg bond indices (bond orders)
were derived from the natural bond orbital (NBO) program
[45], helping us to judge the strengths of the Lewis acidity
[46].
The computational methods and cluster models were validated:
(1) larger basis set (B2). In B2, the LanL2TZ(f) and LanL2DZ(d,p)
ECP basis sets [44,47,48] were used for the Ti, Zr and Ge, Sn,
Pb atoms, respectively, whereas the 6-311G(d,p) basis set for the
other atoms, (2) larger cluster model models (37-T) with com-
plete straight channels (Fig. 2). The two-layer ONIOM methodology
(ONIOM) [49,50] was used for the calculations. The [(HO)3 SiO]4 M
fragment and probe molecules were defined as the high-level
region and treated with the B3LYP/B1 method. The rest as
the low-level region was treated at B3LYP/3-21G level [30,51].
The boundary Si H distances were set to 1.500 Å. The atoms
of the low-level region were retained at their crystallographic Fig. 1. The local structures of the Lewis acidity (LA) in the M4+ -doped zeolite as well
coordinates. as of the Lewis acidic site with adsorption of NH3 .
G. Yang et al. / Journal of Molecular Catalysis A: Chemical 363–364 (2012) 371–379 373

ions (M4+ ) are not thermodynamically favored [36]. Generally, the


larger the M4+ radii, the more unfavorable the substitutions.
The LUMO energies, fukui functions (fM ), absolute electroneg-
ativity () have been used to characterize the Lewis acidities of
zeolites [28,30–32,35,36,55]. The electrophilic fukui functions (fM + )
are approximated to be [56–58],
+
fM = qM (N0 + 1) − qM (N0 ) (3)

where qM (N0 ) and qM (N0 + 1) are the atomic charges of the M atom
with the zeolites at the neutral and anionic states, respectively.
The absolute electronegativity () is defined,

I+A
= (4)
2
where I and A represent the ionization potential and electron affin-
ity of the systems, respectively.
Meanwhile, the absolute hardness () is defined,

I−A
= (5)
2
The LUMO energies, fM + ,  and  values are calculated for
the Ge-, Ti-, Sn-, Pb- and Zr-doping zeolites as well as Silicalite-
1, see Table 2. The Lewis acidities predicted by LUMO energies
and  change in the same order, Silicalite-1  Ge < Zr < Sn < Ti < Pb.
Although the fM + values based on the Hirshfeld population anal-
ysis (HPA) are non-negative and thus physically realistic, there
are problems regarding to the HPA charge partitioning technique
[56–59]. On the other hand, the Mulliken population analysis (MPA)
can be used to rank the Lewis acidities of the various tetrava-
lent ions (M4+ ) albeit it may generate negative fM + values. The
fM + data based on MPA show that the Lewis acidity increases
Fig. 2. The Lewis acidic site of the M4+ -doped zeolite as well as the structure with as Silicalite-1  Ti < Zr ≤ Pb < Ge ≤ Sn, exactly the same as that of
adsorption of NH3 , which are calculated by the two-layer ONIOM methodology. The
the Natural population analysis (NPA). All the three parameters
high-level region is displayed in ball and stick whereas the low-level region in tube,
respectively. (LUMO energies,  and fM + ) confirm the experimental observations
that Silicalite-1 has the least Lewis acidity. Sn is known to have
higher Lewis acidity than Ti [61–63] and correctly predicted by fM +
coordinated to four O atoms. The deviations of [MO4 ] from regular
instead of LUMO energies or . The energy gap between LUMO and
tetrahedron can be characterized by the mean square deviation ()
HOMO (Egap ) and the absolute hardness () are correlated with the
[30,52] defined,
hardness and softness of zeolitic systems [60,62]. Hard molecules

 6 correspond to large Egap and  values. As Table 2 shows, Silicalite-1
1
=
has the largest Egap and  values and weakest Lewis acidity, while
¯ 2
(˛i − ˛) (1)
6 Ti, Sn and Pb have smaller Egap and  values and stronger Lewis
i=1
acidities.
In Eq. (1), ˛i represents the ith O M O angle, and is the aver-
age of the six O M O angles. The  values of Ge, Ti, Pb, Sn 3.2. Adsorption of NH3 on the Lewis acidic sites
and Zr are calculated to be 3.12◦ , 3.70◦ , 4.28◦ , 3.57◦ and 5.87◦
respectively, indicating the different deviation degrees from regu- The adsorption structures of ammonia (NH3 ) on the Ge-, Ti-,
lar tetrahedron. Compared with Silicalite-1 ( = 2.81◦ ), the doping Sn-, Pb-, Zr-doped zeolites are given in Fig. 1b. The M N dis-
of tetravalent ions (M4+ ) enlarges the deviations from regular tetra- tances (bond orders) are equal to 2.339 (0.217), 2.360 (0.266), 2.340
hedron, probably due to the radius increase by ion doping. The (0.228), 2.412 (0.228) and 2.471 Å (0.207) for Ge, Ti, Sn, Pb and
M O distances are listed in Table 1 and show elongations by each Zr, respectively (Table 2). In each doping case, the NH3 molecule
doping case. The average M O distances increase in the order forms direct bond with the M4+ ion; that is, all these doping zeo-
Ge (1.708 Å) < Ti (1.774 Å) < Sn (1.851 Å) < Pb (1.912 Å) < Zr (1.935 Å). lites show Lewis acidity [32,35]. Instead, for Silicalite-1 the Si N
There have been computational reports on the Ti, Ge and Sn- distance (bond order) is calculated at 3.109 Å (0.033), much larger
doped zeolites, wherein the M O distances are consistent with the (less) than any of the doping cases. The NH3 molecule should be
present work [15–18,31–36,53,54]. physisorbed on Silicalite-1, in agreement with the absence of Lewis
The substitution energies (Esub ) of Si4+ by tetravalent metal ions acidity predicted by LUMO energies, fM + ,  and . The adsorption
(M4+ ) are evaluated, energies of NH3 (Ead ) amount to −4.6, −8.7, −15.1, −20.7, −21.0
and −22.7 kcal mol−1 for Silicalite-1 and Ge-, Ti-, Pb-, Sn-, Zr-doped
Esub = E(LA) + E(Si4+ ) − E(Silicalite − 1) − E(M4+ ) (2)
zeolites, respectively (Table 2). There are H-bonds forming between
where E(Silicalite-1) and E(LA) are the energies of Silicalite-1 and the NH3 molecule and O1, O3 of Silicalite-1 with lengths of about
Lewis acidic sites with doped tetravalent ions (M4+ ), whereas 2.8 Å, which are mainly responsible for the negative adsorption
E(Si4+ ) and E(M4+ ) are the energies of isolated Si4+ and M4+ ions, energy. The data in Table 1 show that the Si O distances remain
respectively. The Esub values are given in Table 2. Owing to the almost intact upon NH3 adsorption, whereas the M O distances
larger radii than Si4+ , the isomorphic substitutions of tetravalent are significantly elongated, by 0.019–0.028 Å [28,32,35,62].
374 G. Yang et al. / Journal of Molecular Catalysis A: Chemical 363–364 (2012) 371–379

Table 1
The (average) M O distances, deviations of [MO4 ] from regular tetrahedron () and Mulliken/NPA charges of the Lewis acidic sites in tetravalent-ion (M4+ ) doped zeolites
and Silicalite-1, before and after adsorption of NH3 .a

M O1 M O2 M O3 M O4 M Ob b q(M)c q(NH3 )c


4+
Si Bare 1.605 1.616 1.618 1.609 1.612 (1.607, 1.613) 2.81 (2.78, 2.35)
NH3 1.604 1.625 1.622 1.609 1.615 (1.610, 1.617) 4.19 (4.17, 4.61)
Ge4+ Bare 1.700 1.715 1.715 1.702 1.708 (1.709, 1.710) 3.12 (3.37, 2.64) 1.917 (2.569)
NH3 1.716 1.764 1.742 1.723 1.736 (1.733, 1.745) 11.06(11.61,13.05) 1.921 (2.536) 0.129 (0.126)
Ti4+ Bare 1.758 1.783 1.785 1.771 1.774 (1.771, 1.776) 3.70 (4.00, 3.40) 1.054 (1.712)
NH3 1.783 1.810 1.813 1.784 1.797 (1.793, 1.803) 11.39(12.03,13.23) 1.070 (1.594) 0.177 (0.163)
Pb4+ Bare 1.905 1.926 1.918 1.899 1.912 (1.912) 4.28 (4.89) 1.765 (2.529)
NH3 1.927 1.950 1.941 1.917 1.934 (1.933) 13.51(14.80) 1.823 (2.533) 0.155 (0.141)
Sn4+ Bare 1.844 1.862 1.858 1.842 1.851 (1.856, 1.853) 3.57 (4.00, 3.77) 2.032 (2.714)
NH3 1.867 1.896 1.886 1.863 1.878 (1.883, 1.885) 12.65(13.92,14.48) 2.053 (2.708) 0.143 (0.127)
Zr4+ Bare 1.915 1.949 1.949 1.927 1.935 (1.921, 1.935) 5.87 (6.55, 5.83) 1.309 (2.324)
NH3 1.947 1.959 1.971 1.939 1.954 (1.940, 1.960) 12.72(13.36,14.94) 1.341 (2.272) 0.167 (0.110)
a
Units of distances, and charges are in angstrom and |e|, respectively.
b
The average distances in parentheses are calculated at the B3LY/B2 level, 17-T cluster model and at the ONIOM(B3LY/B1:B3LYP/3-21G) level, 37-T cluster model
(underlined), respectively.
c
The NPA charges are given in parentheses.

Table 1 shows that the Mulliken/NPA charges of the M4+ ions have even smaller deviations. As a result, their adsorption energies
do not decrease but increase, which seems to contradict the con- of the B3LYP/B2 (17-T) or ONIOM (37-T) schemes are exactly the
ception of Lewis acidity wherein the empty d orbitals of the M4+ same order as the default method (B3LYP/B1, 17-T), in the order of
ions should accept electrons from the incoming molecules. In fact, Silicalite-1  Ge < Ti < Pb < Sn < Zr.
the Mulliken/NPA charge analyses reveal that electron transfers
are conducted from the NH3 molecules toward the zeolitic sys- 3.3. The Brönsted acidity
tems; that is, the M4+ ions accept electrons but disperse them
rapidly around, which is supported by the delocalization of the As shown in Scheme 1b, the M4+ ions of the Brönsted acidic
LUMOs in Fig. 3. Accordingly, the Lewis acidities of the M4+ -doped sites are directly coordinated with five O atoms instead of four in
zeolites should be more accurately described as the local sites the Lewis acidic sites. The exact coordination numbers of the M4+
around the M4+ ions rather than merely the M4+ ions. The adsorp- ions can be calculated [27],
tion energies, LUMO energies and  are global descriptors and
more suitable for predicting Lewis acidity than the fukui func- r0 r0 r0 r0 r0
nM = + + + + (6)
tions (fM + ) that are based solely on the charges of the M4+ atoms r1 r2 r3 r4 r5
[63]. In addition, the adsorption energies instead of the LUMO
where r1 –r5 are the five M O distances of the Brönsted acidic sites,
energies and  correctly predict that Sn has higher Lewis acid-
and r0 is the average M O distances of the corresponding Lewis
ity than Ti [61,62,64]. Hence, the adsorption energies are superior
acidic sites. The coordination numbers of the M4+ ions (M = Ge, Ti,
in characterizing the Lewis acidities of zeolites, which increase as
Pb, Sn, Zr) fall within 4.63–4.74 (Table 3). It is in agreement with
Silicalite-1  Ge < Ti < Pb < Sn < Zr.
the experimental observations that the M4+ coordination numbers
The present results have been testified by the calculations with
are larger than four [65]. The Brönsted acidic sites of the M4+ -doped
larger basis set (B3LYP/B2) and cluster models (37-T), see the details
zeolites are the hydrolysis products of the Lewis acidic sites, and
in Section 2. Their geometries are in good agreement with those of
their formation energies (FBA ) from the Lewis acidic sites (LA) are
the default method (B3LYP/B1, 17-T). Table 1 lists the average M O
evaluated,
distances and deviations of [MO4 ] from regular tetrahedron () cal-
culated with different computational methods and cluster models, FBA = E(BA) − E(LA) − E(H2 O) (7)
which indicate that for the three computational methods and clus-
ter models, close structural distortions are caused by interactions of The FBA values are calculated to be −18.9, −30.9, −40.2,
these Lewis acidic sites with NH3 . In the case of Silicalite-1 charac- −42.9 and −52.7 kcal mol−1 in the cases of Ge, Ti, Sn, Pb and Zr,
teristic of weak interactions, the M-N distances of the B3LYP/B2 respectively. All except Ge have obviously larger FBA values than
(17-T) or ONIOM (37-T) schemes deviate not more than 0.07 Å Silicalite-1 (Table 3), indicating that the formation of the Brönsted
from the default values; for the doping cases, the M N distances acidic sites will be promoted by ion doping.

Table 2
The radii of the M4+ ions (rM ) and LUMO energies, energy gaps of LUMO and HOMO, electrophilic fukui functions (fM + ), absolute electronegativity (), absolute hardness (),
NH3 adsorption energies (Ead ) as well as M–N distances and bond orders (BOM N ) for the tetravalent-ion (M4+ )-doped zeolites and Silicalite-1.

Si4+ Ge4+ Ti4+ Pb4+ Sn4+ Zr4+

rM /Å 0.42 0.53 0.68 0.84 0.71 0.79


Esub /kcal mol−1 208.6 222.3 475.5 474.1 598.1
LUMO/eV −0.394 −0.585 −1.949 −2.491 −1.239 −0.999
Egap /eV 7.700 7.469 6.090 5.481 6.765 6.999
fM + a −0.036 (−0.038) −0.614 (−0.655) −0.208 (−0.189) −0.501 (−0.613) −0.623 (−0.650) −0.501 (−0.552)
/kcal mol−1 99.0 110.4 122.0 133.9 120.8 114.0
/kcal mol−1 112.1 100.5 88.8 76.0 89.6 95.7
M−N/Å´ 3.109 2.339 2.360 2.412 2.340 2.471
BOM −N 0.033 0.217 0.266 0.228 0.228 0.207
Ead /kcal mol−1 −4.6 −8.7 −15.1 −20.7 −21.0 −22.7
a
Based on the Mulliken (NPA) charges.
G. Yang et al. / Journal of Molecular Catalysis A: Chemical 363–364 (2012) 371–379 375

Fig. 3. The LUMO diagrams of Silicalite-1 and the M4+ -doped zeolites.

As the data in Tables 1 and 3 show, the M O dis- The Brönsted acidity strengths of the M4+ - and M3+ -doped
tances are increased due to the formation of Brönsted acidic zeolites can be evaluated with the aid of proton affinity (PA)
sites. The two M O3 and M O5 bonds (Fig. 4), which are [5,7–10,27,66,67],
potentially related with the acidic protons (H+ ), are obviously
PA = E(BA− ) − E(BA) (8)
longer than the other three (M O1, M O2, M O4). Espe-
cially the M O3 distances are 0.353–0.520 Å larger than the where E(BA) and E(BA− )
represent the energies of the Brönsted
average values of the other three; however, the M O3 dis- acidic sites (BA) and corresponding deprotonated states (BA− ),
tances are dramatically reduced in the deprotonated states respectively.
(BA− ), where the protons are deprived by the incoming probe The proton affinities (PA) of the various M4+ - and M3+ -
molecules. doped zeolites decrease in the order Silicalite-1 (324.9) Ti
376 G. Yang et al. / Journal of Molecular Catalysis A: Chemical 363–364 (2012) 371–379

Table 3
The M O distances, coordination numbers of the doped ions (nM ), Mulliken charges (q) in the Brönsted acidic (BA) and deprotonated (BA− ) states as well as proton affinities
(PA) and formation energies of the Brönsted acidic sites (FBA ) for the M4+ -, M3+ -doped MFI-type zeolites and Silicalite-1.a

M Ob M O3 M O5 nM qH3 qO3H3 FBA PA

Si 4+
BA 1.623 3.511 1.628 4.43 0.466 −0.279 −20.2 324.9
BA− 1.713 1.757 1.719
Ti4+ BA 1.801 2.272 1.854 4.69 0.478 −0.283 −30.9 307.4
BA− 1.857 1.899 1.904
Ge4+ BA 1.742 2.362 1.765 4.63 0.464 −0.299 −18.9 306.4
BA− 1.805 1.846 1.811
B 3+
BA 1.369 2.195 0.450 −0.277 303.6
BA− 1.462 1.467
Pb4+ BA 1.952 2.367 1.980 4.71 0.482 −0.307 −42.9 300.9
BA− 1.995 2.019 2.031
Zr4+ BA 1.964 2.363 2.014 4.74 0.485 −0.296 −52.7 300.2
BA− 2.016 2.056 2.050
Sn4+ BA 1.891 2.244 1.922 4.73 0.481 −0.331 −40.2 299.2
BA− 1.935 1.968 1.956
Fe 3+
BA 1.786 2.019 0.506 −0.273 292.8
BA− 1.823 1.832
Al3+ BA 1.694 1.884 0.491 −0.242 288.2
BA− 1.726 1.735
a
Units of distances charges and energies are in angstrom, |e| and kcal mol−1 , respectively.
b
The average distances of M O1, M O2 and M O4.

(307.4) > Ge (306.4) > B (303.6) > Pb (300.9) ≥ Zr (300.2) ≥ Sn The Brönsted acidities of M4+ -doped zeolites are not closely related
(299.2) > Fe (292.8) > Al (288.2), units in kcal mol−1 (Table 3 with and fail to be predicted by the charges of the OH groups.
and Fig. 5). The acidic sites with higher PAs are more dif-
ficult to donate protons and thus have lower Brönsted 3.4. Interactions of the Brönsted acidic sites with probe molecules
acidities. Accordingly, the acidities increase as Silicalite-
1  Ti4+ < Ge4+ < B3+ < Pb4+ ≤ Zr4+ ≤ Sn4+ < Fe3+ < Al3+ . The PA values Five probe molecules of changing basicities are used to char-
of Pb, Zr and Sn are close to each other and their acid strengths acterize the Brönsted acidities of M4+ -doped zeolites (M = Ge,
will be differed by further interactions with probe molecules. Ti, Pb, Sn, Zr). The basicities decrease in the order of ammo-
Silicalite-1 has no Brönsted acidity [27,60], and Ti and Ge have nia (NH3 ) > trimethyphosphine (TMP) > pyridine > acetone > water
weaker Brönsted acidity than B, the weakest Brönsted acidity [68]. For comparisons, two M3+ ions B and Al are studied as well.
among the three trivalent ions. Among the various tetravalent ions The results are summarized in Tables 4 and 5. Some adsorption
(M4+ ), Sn has the strongest Brönsted acidity, but is much weaker structures are illustrated in Figs. 6–8. In the covalent structures
than Al. It seems not likely for a tetravalent ion (M4+ ) to have (Figs. 7a and 8), the acidic protons remain on the zeolites and
comparable Brönsted acidity to Al3+ . Nonetheless, the Brönsted do not transfer to probe molecules. Instead, the acidic protons
acidities of M4+ -doped zeolites change within a wide range, and have been transferred to probe molecules in the ionic structures
their acidities are stronger than those of certain M3+ ions; e.g., (Figs. 6 and 7b).
Pb4+ , Zr4+ and Sn4+ vs. B3+ . The formations of covalent and/or ionic structures are the
Fig. 5 shows that the Brönsted acidities of M3+ -doped zeolites proton-competing results between the probe molecules and Brön-
increase with decrease of the negative charges of the bridging OH sted acidic sites. If the probe molecule has strong enough basicity,
groups. It is not applicable to the M4+ -doped zeolites, however. the acidic proton will be definitely on it; that is, even if starting
from the covalent mode, the proton will gradually transfer toward
to the probe molecule (PM) and result in the PMH+ and BA− species.

Fig. 5. The ranks of Brönsted acidities and Mulliken charges for the various M4+ -
Fig. 4. The local structure of the Brönsted acidity (BA) in the M4+ -doped zeolite. and M3+ -doped zeolites as well as Silicalite-1.
G. Yang et al. / Journal of Molecular Catalysis A: Chemical 363–364 (2012) 371–379 377

Table 4 Table 5
Distances and interaction energies (Ein ) for the ionic structures due to adsorption of Distances and interaction energies (Ein ) for the covalent structures due to adsorption
probe molecules on the Brönsted acidic sites in M4+ - and M3+ -doped zeolites.a of probe molecules on the Brönsted acidic sites in M4+ - and M3+ -doped zeolites.a

Ti4+ Ge4+ B3+ Pb4+ Zr4+ Sn4+ Al3+ Ti4+ Ge4+ B3+ Pb4+ Zr4+ Sn4+ Al3+

NH3 Pyridine
O3 H3 1.526 1.557 1.604 1.622 1.584 1.602 2.494 O3 H3 1.031 1.040 1.055
N H3 1.092 1.083 1.078 1.072 1.080 1.075 1.037 N H3 1.644 1.615 1.561
Ein −22.2 −24.5 −25.4 −28.4 −26.2 −30.0 −36.2 Ein −12.0 −11.4 −14.6
Pyridine TMP
O3 H3 1.492 1.545 1.616 1.557 1.523 1.560 1.640 O3 H3 1.006 1.003 0.990 1.024 1.020 1.026 1.044
N H3 1.094 1.079 1.065 1.070 1.084 1.074 1.060 P H3 2.242 2.271 2.431 2.127 2.146 2.128 2.102
Ein −12.2 −13.8 −19.2 −18.4 −16.7 −19.7 −29.0 Ein −5.6 −4.5 −5.9 −7.2 −8.1 −7.4 −11.8
TMP Acetone
O3 H3 1.702 1.751 2.235 1.725 1.745 1.733 1.996 O3 H3 1.006 1.005 1.003 1.015 1.011 1.108 1.024
P H3 1.432 1.421 1.398 1.425 1.426 1.424 1.408 O H3 1.642 1.658 1.681 1.599 1.617 1.589 1.583
Ein −2.7 −3.3 −5.7 −9.8 −7.9 −10.5 −20.0 Ein −10.0 −9.6 −10.2 −12.3 −12.7 −12.7 −15.1
H2 O
a
Units of distances and energies in angstrom and kcal mol−1 , respectively.
O3 H3 1.014 1.020 1.020 1.023 1.018 1.028 1.046
O H3 1.608 1.591 1.612 1.594 1.607 1.569 1.511
It is applicable to NH3 where only the ionic structure has formed in Ein −18.3 −20.5 −18.7 −22.5 −22.6 −23.3 −22.8

each doping case (Table 4 and Fig. 6). Owing to the lower basicity, a
Units of distances and energies in angstrom and kcal mol−1 , respectively.
TMP can result in both covalent and ionic structures; that is, nei-
ther of the probe molecules and Brönsted acidic sites can deprive
protons from the other sides (Tables 4 and 5). As the basicities of the ionic and covalent structures, see Ti, Ge and Zr in Fig. 7 and
probe molecules continue to decrease, the deprotonated Brönsted Tables 4 and 5. The strengths of the Brönsted acidities are thus dif-
acidic sites (BA− ) are strong enough to seize the protons from the fered. Such differences can also be observed from the geometric
protonated probe molecules (PMH+ ), causing the absence of the analyses and interaction energies (Ein ). In the ionic structures, the
ionic modes. It is applicable to acetone and water, wherein only the H3 protons form H-bonds with the zeolite-O3 atoms. Generally,
covalent structure exists (Table 5 and Fig. 7) [52,69,70]. For pyridine the stronger the Brönsted acidity, the longer the O3 H3 H-bond,
(Py) adsorption, stronger Brönsted acidic sites result in only ionic and the shorter the N H3 bond (Table 4). In the covalent struc-
structures, and weaker Brönsted acidic sites cause co-existence of tures, the H3 atoms form H-bonds with probe molecules, and the

Fig. 6. The optimized structures of NH3 adsorption on the Brönsted acidic site (BA) Fig. 7. The optimized structures of acetone adsorption on the Brönsted acidic site
of zeolites with the doping of M4+ (a) and M3+ (b) ions. (BA) of zeolites with the doping of M4+ (a) and M3+ (b) ions.
378 G. Yang et al. / Journal of Molecular Catalysis A: Chemical 363–364 (2012) 371–379

Fig. 9. The energy differences of the ionic and covalent structures for TMP adsorp-
tion on the Brönsted acidic site of the M4+ - and M3+ -doped zeolites as well as
Silicalite-1.

4. Conclusions

The M4+ -doped zeolites are excellent catalysts for a lot of chem-
ical processes, such as the currently focusing biomass conversions.
Through the studies of the Lewis and Brönsted acidities of various
M4+ -doped zeolites (M = Ge, Ti, Pb, Sn, Zr) as well as their interac-
tions with probe molecules, this work presents a dynamic image of
acid–base interactions and aids our understanding to the catalysis
of solid-state acids.
Compared with Silicalite-1, the doping of the M4+ ions causes
Fig. 8. The covalent and ionic structures due to pyridine (Py) adsorption on the more deviations to [MO4 ] tetrahedrons and different degrees of
Brönsted acidic site (BA) in the M4+ -doped zeolites (M = Ti, Ge, Zr). M O bond elongations. The charges of the M4+ ions do not decrease
as expected, due to that the M4+ ions transfer the electrons,
obtained from the adsorbed molecules, quickly to the adjacent
stronger Brönsted acidity generally leads to the shorter O H3 bond atoms. The Lewis acidities of the M4+ -doped zeolites should be
(Table 5). more accurately defined as the local sites around the M4+ ions
The interaction energies between the probe molecules (PM) and rather than merely the M4+ ions. It also explains why the adsorp-
Brönsted acidic sites (BA) can be calculated, tion energies give a more consistent order of Lewis acidities with
LUMO energies and absolute electronegativity rather than fukui
Ein = E(Product) − E(BA) − E(PM) (9) functions. According to the adsorption energies, the Lewis acidities
should increase as Silicalite-1  Ge < Ti < Pb < Sn < Zr, in agreement
where E(Product) is the energy of the optimized structure, in with the experimental data available.
the covalent or ionic form. TMP, often used to probe the zeo- The doping of M4+ ions into silica lattices facilitates the forma-
lite acidities [17,27,71–73], is the only one that in all the doping tion of Brönsted acidic sites and expands the coordination numbers.
cases, the covalent and ionic modes will co-exist. The energy According to the proton affinity (PA), the Brönsted acidities of M4+ -
differences between the ionic and covalent structures (Ediff ) doped zeolites change within a wide range. Sn has the strongest
decrease in the order Silicalite-1 (22.4) Ti (2.9) > Ge (1.2) > Zr Brönsted acidity among the five M4+ -doped zeolites and the acidi-
(0.2) = B (0.2) > Pb (−2.6) > Sn (−3.1) > Al (−8.2), units in kcal mol−1 ties of the M4+ ions can be stronger than those of certain M3+
(Fig. 9). It suggests that the Brönsted acidity increases as Silicalite- ions. It seems not likely for a tetravalent ion (M4+ ) to have com-
1  Ti < Ge < Zr ≈ B < Pb < Sn < Al, in general agreement with the parable Brönsted acidity to Al3+ . Unlike the M3+ ions, the Brönsted
order of proton affinities (PA). The main deviation between Ediff acidities of M4+ are not closely related with the charges of the bridg-
and PA is the Zr4+ ion. The Ediff (TMP) rather than PA data suggest ing OH groups. The interactions between five probe molecules of
a lower Brönsted acidity for Zr, supported by the changing ten- changing basicities and various Brönsted acidic sites of M4+ - and
dencies of the O3 H3 and H3 P distances (Tables 4 and 5). The M3+ -doped zeolites indicate that the formations of covalent and/or
Ein data in Table 4 show that in all cases, the protonated probe ionic structures are the results of proton competitions. NH3 has
molecules (PMH+ ) with Zr have less interaction energies than with strong enough basicity and the resulting structure is exclusively
Pb. Especially, only Ti, Ge and Zr form covalent complexes by pyri- ionic, even if starting from the covalent mode. Acetone and water
dine adsorption and should have lower Brönsted acidities (Table 5). has weak basicities and even if protonated, the protons will trans-
Accordingly, the Brönsted acidities should increase as Silicalite- fer to the Brönsted acidic sites, forming the covalent structures.
1  Ti < Ge < Zr ≈ B < Pb < Sn < Al. Neither of TMP (of medium bacisity) and Brönsted acidic sites can
G. Yang et al. / Journal of Molecular Catalysis A: Chemical 363–364 (2012) 371–379 379

deprive protons from the other sides, and the covalent and ionic [30] G. Yang, L.J. Zhou, X.C. Liu, X.W. Han, X.H. Bao, Chem. Eur. J. 17 (2011)
structures co-exist with each other. For pyridine adsorption, only 1614–1621.
[31] G. Sastre, A. Corma, Chem. Phys. Lett. 302 (1999) 447–453.
ionic structures in the cases of stronger Brönsted acidic sites, and [32] A. Damin, S. Bordiga, A. Zecchina, C. Lamberti, J. Chem. Phys. 117 (2002)
will co-exist with covalent structures by reducing the Brönsted 226–237.
acidity. The different acidity strengths are thus observed. Accord- [33] T. Atoguchi, S.T. Yao, J. Mol. Catal. A: Chem. 191 (2003) 281–288.
[34] S. Bordiga, F. Bonino, A. Damin, C. Lamberti, Phys. Chem. Chem. Phys. 9 (2007)
ing to the energy differences, interaction energies and geometric 4854–4878.
parameters, the Brönsted acidities should increase as Silicalite- [35] B.S. Kulkarni, S. Krishnamurty, S. Pal, J. Mol. Catal. A: Chem. 329 (2010) 36–43.
1  Ti < Ge < Zr ≈ B < Pb < Sn < Al. The main difference with proton [36] S. Shetty, S. Pal, D.G. Kanhere, A. Goursot, Chem. Eur. J. 12 (2006) 518–523.
[37] R.C. Deka, V.A. Nasluzov, E.A.I. Shor, A.M. Shor, G.N. Vayssilov, N. Rösch, J. Phys.
affinities (PA) is the prediction of a lower ranking for the Zr4+ ion,
Chem. B 109 (2005) 24304–24310.
supported by the geometric and energetic results, especially in the [38] J.D. Gale, Solid State Sci. 8 (2006) 234–240.
case of pyridine adsorption wherein the covalent structures exist [39] G. Pirc, J. Stare, Acta Chim. Slov. 55 (2008) 951–959.
[40] A.D. Becke, J. Chem. Phys. 88 (1988) 2547–2551.
only on Zr and Ti, Ge.
[41] C. Lee, W. Yang, R.G. Parr, Phys. Rev. B 37 (1988) 785–789.
[42] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman,
Acknowledgements J.A. Montgomery, K.N. Vreven, T. Kudin Jr., J.C. Burant, J.M. Millam, S.S. Iyengar,
J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G.A. Petersson,
H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T.
We gratefully acknowledged the financial supports from the Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J.E. Knox, H.P. Hratchian,
National Natural Science Foundation (No. 20903019) and Min- J.B. Cross, C. Adamo, J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J.
istry of Science and Technology of the Peoples’ Republic of China Austin, R. Cammi, C. Pomelli, J.W. Ochterski, P.Y. Ayala, G.A. Morokuma, G.A.
Voth, P. Salvador, J.J. Dannenberg, V.G. Zakrzewski, S. Dapprich, A.D. Daniels,
(2003CB615806). M.C. Strain, O. Farkas, D.K. Malick, A.D. Rabuck, K. Raghavachari, J.B. Foresman,
Q. Ortiz, Q. Cui, A.G. Baboul, S. Clifford, J. Cioslowski, B.B. Stefanov, G. Liu, A.
References Liashenko, P. Piskorz, I. Komaromi, R.L. Martin, D.J. Fox, T. Keith, M.A. Al-Laham,
C.Y. Peng, A. Nanayakkara, M. Challacombe, M.W.P. Gill, B. Johnson, W. Chen,
M.W. Wong, C. Gonzalez, J.A. Pople, Gaussian 03, Revision D. 01, Gaussian, Inc,
[1] R. Fricke, H. Kosslick, G. Lischke, M. Richter, Chem. Rev. 100 (2000) 2303–2406. Wallingford, CT, 2004.
[2] J.M. Thomas, R. Raja, G. Sankar, R.G. Bell, Acc. Chem. Res. 34 (2001) 191–200. [43] T.H.J. Dunning, P.J. Hay, in: H.F. Schaefer III (Ed.), Modern Theoretical Chemistry,
[3] J.B. Nagy, R. Aiello, G. Giordano, A. Katovic, F. Testa, Z. Kónya, I. Kiricsi, Isomor- Plenum Press, New York, 1976, pp. 1–35.
phous substitution in zeolites, in: H.G. Karge, J. Weitkamp (Eds.), Molecular [44] P.J. Hay, W.R. Wadt, J. Chem. Phys. 82 (1985) 270–283.
Sieves, vol. 5, Springer/Heidelberg, Berlin, 2007, pp. 365–478. [45] E.D. Glendening, A.E. Reed, J.E. Carpenter, F. Weinhold, NBO version 3.1, 1988.
[4] H. Kosslick, V.A. Tuan, B. Parlitz, R. Fricke, C. Peuker, W. Storek, J. Chem. Soc., [46] K.B. Wiberg, Tetrahedron 24 (1968) 1083–1096.
Faraday Trans. 89 (1993) 1131–1138. [47] P.J. Hay, W.R. Wadt, J. Chem. Phys. 82 (1985) 284–298.
[5] H.V. Brand, L.A. Curtiss, L.E. Iton, J. Phys. Chem. 97 (1993) 12773–12782. [48] P.J. Hay, W.R. Wadt, J. Chem. Phys. 82 (1985) 299–310.
[6] A. Chatterjee, A.K. Chandra, J. Mol. Catal. A 119 (1997) 51–56. [49] S. Dapprich, I. Komaromi, K.S. Byun, K. Morokuma, M. Frisch, J. Mol. Struct.
[7] A. Chatterjee, T. Iwasaki, T. Ebina, A. Miyamoto, Micropor. Mesopor. Mater. 21 Theochem. 461–462 (1999) 1–21.
(1998) 421–428. [50] M. Lundberg, T. Kawatsu, T. Vreven, M.J. Frisch, K. Morokuma, J. Chem. Theory
[8] S.P. Yuan, J.G. Wang, Y.W. Li, H.J. Jiao, J. Phys. Chem. A 106 (2002) 8167–8172. Comput. 5 (2009) 222–234.
[9] Y. Wang, D.H. Zhou, G. Yang, S.J. Miao, X.C. Liu, X.H. Bao, J. Phys. Chem. A 108 [51] G. Yang, L.J. Zhou, C.B. Liu, J. Phys. Chem. B 113 (2009) 10399–10402.
(2004) 6730–6734. [52] G. Yang, Y. Wang, D.H. Zhou, X.C. Liu, X.W. Han, X.H. Bao, J. Mol. Catal. A: Chem.
[10] Y. Wang, G. Yang, D.H. Zhou, X.C. Liu, X.H. Bao, J. Phys. Chem. B 108 (2004) 237 (2005) 36–44.
18228–18233. [53] G. Sastre, A. Corma, J. Phys. Chem. C 114 (2010) 1667–1673.
[11] B. Notari, Adv. Catal. 41 (1996) 253–334. [54] G. Sastre, J.A. Vidal-Moya, T. Blasco, J. Rius, J.L. Jordá, M.T. Navarro, F. Rey, A.
[12] P. Ratnasamy, D. Srinivas, H. Knozinger, Adv. Catal. 48 (2004) 1–169. Corma, Angew. Chem. Int. Ed. 41 (2002) 4722–4726.
[13] P.E. Sinclair, G. Sankar, C.R.A. Catlow, J.M. Thomas, T. Maschmeyer, J. Phys. [55] P.G. Pearson, Proc. Natl. Acad. Sci. U.S.A. 83 (1986) 8440–8441.
Chem. B 101 (1997) 4232–4237. [56] R.K. Roy, S. Pal, K. Hirao, J. Chem. Phys. 110 (1999) 8236–8245.
[14] S. Bordiga, A. Damin, F. Bonino, G. Ricchiardi, C. Lamberti, A. Zecchina, Angew. [57] R.K. Roy, K. Hirao, S. Pal, J. Chem. Phys. 113 (2000) 1372–1379.
Chem. Int. Ed. 41 (2002) 4734–4737. [58] R.K. Roy, K. Hirao, S. Krishnamurty, S. Pal, J. Chem. Phys. 115 (2001) 2901–2907.
[15] C.A. Hijar, R.M. Jacubinas, J. Eckert, N.J. Henson, P.J. Hay, K.C. Ott, J. Phys. Chem. [59] S. Saha, R.K. Roy, P.W. Ayers, Int. J. Quantum Chem. 109 (2009) 1790–1806.
B 104 (2000) 12157–12164. [60] W.O. Parker, Magn. Reson. Chem. 37 (1999) 433–436.
[16] C. Lamberti, S. Bordiga, A. Zecchina, G. Artioli, G. Marra, G. Spanò, J. Am. Chem. [61] A. Corma, M.E. Domine, L. Nemeth, S. Valencia, J. Am. Chem. Soc. 124 (2002)
Soc. 123 (2001) 2204–2212. 3194–3195.
[17] J.Q. Zhuang, G. Yang, D. Ma, X.J. Lan, X.M. Liu, X.W. Han, X.H. Bao, U. Müller, [62] S. Shetty, B.S. Kulkarni, D.G. Kanhere, A. Goursot, S. Pal, J. Phys. Chem. B 112
Angew. Chem. Int. Ed. 43 (2004) 6377–6381. (2008) 2573–2579.
[18] G. Sastre, A. Pulido, R. Castañeda, A. Corma, J. Phys. Chem. B 108 (2004) [63] R.C. Deka, R. Vetrivel, S. Pal, J. Phys. Chem. A 103 (1999) 5978–5982.
8830–8835. [64] R.S. Assary, L.A. Curtiss, J. Phys. Chem. A 115 (2011) 8760–9754.
[19] C.M. Zicovich-Wilson, A. Corma, J. Phys. Chem. B 104 (2000) 4134–4140. [65] C. Lamberti, S. Bordiga, D. Arduino, A. Zecchina, F. Geobaldo, G. Spanó, F. Genoni,
[20] A. Corma, U. Diaz, M.E. Domine, V. Fornś, J. Am. Chem. Soc. 122 (2000) G. Petrini, A. Carati, F. Villain, G. Vlaic, J. Phys. Chem. B 102 (1998) 6382–6390.
2804–2809. [66] A. Redondo, P.J. Hay, J. Phys. Chem. 97 (1993) 11754–11761.
[21] S. Shetty, S. Pal, D.G. Kanhere, A. Goursot, Chem. Eur. J. 12 (2005) 518–523. [67] G. Yang, X.C. Liu, X.W. Han, X.H. Bao, J. Phys. Chem. B 110 (2006) 23388–23394.
[22] K. Szczodrowski, B. Prélot, S. Lantenois, J.M. Douillard, J. Zajac, Micropor. Meso- [68] J.H. Lunsford, Top. Catal. 4 (1997) 91–98.
por. Mater. 124 (2009) 84–93. [69] H.J. Fang, A.M. Zheng, Y.Y. Chu, F. Deng, J. Phys. Chem. C 115 (2010)
[23] D.T. Win, AU J. Technol. 11 (2007) 36–41. 12711–12718.
[24] T.F. Degnan Jr., Top. Catal. 13 (2000) 349–356. [70] B. Boekfa, P. Pantu, M. Probst, J. Limtrakul, J. Phys. Chem. C 115 (2010)
[25] M.S. Holm, S. Saravanamurugan, E. Taarning, Science 328 (2010) 602–605. 15061–15067.
[26] M. Moliner, Y.R. Leshkov, M.E. Davis, Proc. Natl. Acad. Sci. U.S.A. 107 (2010) [71] J.H. Lunsford, W.X. Shen, W.P. Rothwell, W.X. Shen, J. Am. Chem. Soc. 107 (1985)
6164–6168. 1540–1547.
[27] G. Yang, X.J. Lan, J.Q. Zhuang, D. Ma, L.J. Zhou, X.C. Liu, X.W. Han, X.H. Bao, Appl. [72] J.O. Ehresmann, W. Wang, B. Herreros, D.P. Luigi, T.N. Venkatraman, W.G. Song,
Catal. A: Gen. 337 (2008) 58–65. J.B. Nicholas, J.F. Haw, J. Am. Chem. Soc. 124 (2004) 10868–10874.
[28] G. Yang, J.Q. Zhuang, D. Ma, X.J. Lan, L.J. Zhou, X.C. Liu, X.W. Han, X.H. Bao, J. [73] Y.Y. Chu, Z.W. Yu, A.M. Zheng, H.J. Fang, H.L. Zhang, S.J. Huang, S.B. Liu, F. Deng,
Mol. Struct. 882 (2008) 24–29. J. Phys. Chem. C 115 (2011) 7660–7667.
[29] G. Yang, L.J. Zhou, X.C. Liu, X.W. Han, X.H. Bao, Struct. Chem. 18 (2007) 353–356.

You might also like