Microporous and Mesoporous Materials 294 (2020) 109908

Contents lists available at ScienceDirect

Microporous and Mesoporous Materials
journal homepage: />
Theoretical study on
CHA zeolites

31

P NMR chemical shifts of phosphorus-modified

Pei Zhao a, Bundet Boekfa b, Toshiki Nishitoba c, Nao Tsunoji d, Tsuneji Sano d, Toshiyuki Yokoi c,
Masaru Ogura e, Masahiro Ehara a, f, *
a

Research Center for Computational Science, Institute for Molecular Science, Okazaki, 444-8585, Japan
Department of Chemistry, Faculty of Liberal Arts and Science, Kasetsart University, Kamphaengsaen Campus, Nakhonpathom, 73140, Thailand
Institute of Innovative Research, Tokyo Institute of Technology, 4259 Nagatsuta, Midori-ku, Yokohama, Kanagawa, 226-8503, Japan
d
Department of Applied Chemistry, Graduate School of Engineering, Hiroshima University, 1-4-1 Kagamiyama, Higashi-Hiroshima, Hiroshima, 739-8527, Japan
e
Institute of Industrial Science, The University of Tokyo, 4-6-1 Komaba, Meguro, Tokyo, 153-8505, Japan
f
Element Strategy Initiative for Catalysts and Batteries (ESICB), Kyoto University, Kyoto, 615-8245, Japan
b
c

A R T I C L E I N F O

A B S T R A C T

Keywords:
P-CHA zeolite
31
P NMR chemical shift
Density functional theory calculation

The 31P MAS NMR spectra of phosphorus-modified chabazite (P-CHA) zeolites have been observed during the
hydrothermal treatment to probe the structural changes of phosphorus species in zeolites. Characteristic changes
of the spectra were observed in the range of À 27 ~ À 42 ppm, which correlates to the hydrothermal structure
changes in P-CHA zeolites. Theoretical calculations on the 31P and 27Al NMR chemical shifts have been sys­
tematically performed to disclose the possible phosphorus species of intra- and extra-framework and the struc­
tural changes during the hydrothermal treatment. The shift of the 31P resonances toward higher field were
identified for condensed phosphates and aluminophosphates. Based on the calculated 31P NMR chemical shifts,
the peak with increased intensity at À 42 ppm in the initial stage of the hydrothermal treatment is mainly
assigned to the formation of siliconoaluminophosphate (SAPO) species in the 6-membered ring of the zeolite
framework, while the peak with increased intensity at À 29 ppm in the later stage of treatment is ascribed to the
accumulation of the extra-framework condensed phosphate or aluminophosphate species after the partial
framework decomposition. The dominant peak at À 33 ppm in all 31P NMR spectra is assigned to the phosphates
in the framework.

1. Introduction
Zeolites are inorganic microporous crystalline materials composed
by SiO4 and AlO4 tetrahedra that link to form channels and cavities of
molecular dimensions [1], which are important heterogeneous catalysts
widely used in the modern chemical and petrochemical industries [2–4].
The substitution of tetrahedrally coordinated framework Si atoms with
the trivalent aluminum introduces bridging hydroxyl groups in the or­
dered structure, which significantly affect the catalytic activity.

The modification of zeolites with phosphorus is a widely adopted
method to tune the acidity and consequently the catalytic properties in
terms of activity, shape selectivity, and hydrothermal stability [5,6].
Two intriguing impacts of phosphorus introduction in zeolites have been
demonstrated. (1) The interaction between phosphorus and the zeolite
framework can effectively stabilize tetrahedrally coordinated

framework aluminum, presenting enhanced stability against deal­
umination in the presence of steam [7,8]. (2) The Brønsted acidity is
reduced upon the introduction of phosphorus species [9], which can be
ascribed to the bonding interaction of phosphorous with bridging hy­
droxyl groups and phosphorus-induced dealumination of tetrahedrally
coordinated lattice aluminum together with the simultaneous formation
of water insoluble amorphous extra-framework aluminum phosphates
[8,10].
In particular, regarding selective catalytic reduction (SCR) of NOx
with NH3, it was found that the phosphorus-modified CHA (P-CHA)
zeolites show high hydrothermal stability. The systematic examinations
of wide range of P-CHA zeolites have been performed [11–13]. P-CHA
could exhibit NO purification ability in the SCR process even after severe
hydrothermal treatment at 900 � C. As in general, the role of phosphorus
species in zeolites has been intensively analyzed by various

* Corresponding author. Research Center for Computational Science, Institute for Molecular Science, Okazaki, 444-8585, Japan.
E-mail address: (M. Ehara).
/>Received 24 September 2019; Received in revised form 29 October 2019; Accepted 18 November 2019
Available online 21 November 2019
1387-1811/© 2019 The Authors.
Published by Elsevier Inc.
This is an open access

( />
article

under

the

CC

BY-NC-ND

license


P. Zhao et al.

Microporous and Mesoporous Materials 294 (2020) 109908

spectroscopic measurement including 31P Magic Angle Spinning (MAS)
NMR. However, the origin of the above advantages has still not been
clear partly because the fundamental theoretical characterization or
identification of the 31P NMR chemical shift of P-CHA zeolites has not
been satisfactory.
Up to now, the phosphorus modification has been extensively stud­
ied on H-ZSM-5 stemmed from a well-ordered MFI type framework [9].
Based on various spectroscopy information especially 27Al, 29Si, and 31P
MAS NMR, several models have been proposed to understand the
interaction of the ZSM-5 zeolite framework with phosphorus compounds
and the species generated by this modification [7,14–16]. However, the
preferred chemical models during the severe hydrothermal treatment

are still elusive due to the absence of compelling evidences. Meanwhile,
the difference in experimental conditions during and after the intro­
duction of phosphorus also makes the interpretation of data and spectra
ambiguous. Thus, theoretical studies play a significant role in disclosing
the feasible species. However, systematic studies of the 31P NMR spectra
regarding the phosphorus-modified zeolites have not been performed so
far. Note that zeolitic materials based on the chabazite topology have
high symmetry and only one symmetry-distinct tetrahedral site in the
framework [17], which is an advantage of theoretical calculations.
Herein, the 31P MAS NMR spectra of P-CHA zeolites under the hy­
drothermal treatment have been observed. The 31P MAS NMR is sensi­
tive to phosphorus species involved and therefore, is an effective method
to simulate or probe its structural change. To give the identification of
the phosphorus species and to disclose the structural changes of the PCHA zeolites during hydrothermal treatment, we performed systematic
theoretical calculations on the 31P and 27Al NMR chemical shifts.
Different phosphates and aluminophosphates were first considered to
evaluate the 31P NMR chemical shifts of different types of phosphorus
species. The 31P and 27Al NMR chemical shifts of possible chemical
structures for the phosphorus-zeolite interaction system were studied to
uncover the preferred species during the hydrothermal treatment of the
P-CHA zeolite.

NMR chemical shifts were referenced to 1.6 ppm, ammonium dihy­
drogen phosphate, respectively, and samples were spun at 15 kHz by
using a 4 mm ZrO2 rotor. For 31P MAS NMR spectra, which were
recorded by using a single pulse, the pulse width was set at 4.25 μs and
1000 scans were accumulated at a sample spinning rate of 15 kHz. A 5 s
relaxation delay was determined so as to be long enough to permit
quantitative analysis of zeolite samples.
3. Computational details

For the identification of 31P NMR spectrum of P-CHA zeolites under
the hydrothermal treatment, 31P NMR chemical shifts of both extra- and
intra-framework phosphorus species as well as 27Al NMR chemical shifts
have been systematically studied by the density functional theory (DFT)
calculations. For the extra-framework phosphorus species, phosphates
and aluminophosphates were examined. Regarding the phosphateframework interacting systems, both monophosphates and di­
phosphates bound to P-CHA were considered.
The 42T quantum cluster where T stands for the tetrahedral Si or Al
atoms was used to represent the H-CHA zeolite. The quantum cluster
was adopted from the crystallographic structure of CHA zeolite [18].
The chemical formula of the 42T cluster can be expressed as
(Al2Si40O66H2)H36 (H2: Brønsted protons; H36: terminal hydrogen
atoms). The model was carefully treated to represent the effect of the
Brønsted acid at the six-membered ring (6-MR) and the framework effect
inside the cavity where the phosphorus species dominate. Two Si atoms
at the 6-MR are replaced by two Al atoms to generate the proton
Brønsted acid site by adding a proton to the O atom which is adjacent to
€wenstein’s rule [19], i.e., the
each Al atom. According to the Lo
–Al–O–Al– bond formation is forbidden, two different positions at the
6-MR were considered to place two Al atoms: 3NN (third nearest
neighbor Al site, Al–O–Si–O–Si–O–Al) and 2NN (second nearest
neighbor Al site, Al–O–Si–O–Al) [20]. Meanwhile, two Brønsted protons
were attached to the bridging O atom of 6-MR and 8-MR to locate their
favorable position (Table S4). At the boundary, the terminal Si–H bonds
were treated at the bond length of 1.47 Å with the same direction of Si–O
bond from the crystallographic data. Only the terminal hydrogen atoms
were kept fixed with X-ray structure while other atoms were allowed to
relax.
The adsorption or reaction energy (ΔEr) was calculated as the energy

difference between the considered chemical structure (Ephosphorus-zeolite)
and the sum of the H3PO4 (Ephosphorus) and water (Ewater) molecules as
well as single zeolite (Ezeolite), ΔEr ¼ Ephosphorus-zeolite - (Ephosphorus ỵ
Ewater ỵ Ezeolite). Note that the amounts of the H3PO4 and water mole­
cules depend on the chemical structure under consideration.
The M06-L functional [21] with the basis set of 6-31G(d,p) was
utilized in the geometry optimizations, and vibrational frequency ana­
lyses were also conducted at the same level of theory to confirm the
considered structures to be local minima. The series of M06 functional
can well describe the electrostatic and van der Waals interaction [22].
Particularly, the M06-L functional has improved performance for
calculating NMR chemical shielding constants [23]. NMR chemical
shielding values were evaluated employing the gauge-invariant atomic
orbital (GIAO) method [24] at the optimized geometry. The calculated
31
P and 27Al chemical shifts in the considered structures are referenced
to H3PO4 and Al(NO3)3, respectively, which are often used as reference
molecules [16,25,26]. The following formula is used to obtain the
calculated NMR chemical shift (δ):

2. Experimental details
Phosphorus-modified CHA zeolite was prepared according to the
previous report [13]. Sodium hydroxide, dealuminated FAU zeolite, N,
N,N-trimethyl-1-adamantammonium hydroxide (TMAda), tetraethyl
phosphonium hydroxide (TEP), and distilled water were mixed to obtain
a gel with the molar composition 1:0.0625:0.25–0.03:0.05–0.27:0.1:7.5
of Si/Al/TMAda/TEP/NaOH/H2O. The resulting gel was transferred
into a 100 cm3 Teflon-lined stainless-steel autoclave (stirring-type hy­
drothermal synthesis reactor, R-100, Hiro Company, Japan) and heated
at 150 � C for seven days with tumbling at 10 rpm. After crystallization,

the solid product was collected by centrifugation, washed thoroughly
with distilled water until it was almost neutral, and then dried overnight
at 70 � C. To remove organic molecules, the obtained as-synthesized
P-CHA zeolite was calcined in air at 600 � C for 10 h.
Phosphorus-modification degree (P/Al) was controlled by tuning the
amount of TMAda and TEP in the synthesis gel. The obtained Na-form
P-CHA sample was converted to NH4-form by ion-exchange method; 1
g of Na-form P-CHA sample was stirred in 100 ml of 1.0 M NH4NO3
aqueous solution at 60 � C for 6 h. The treatment was repeated three
times. Then, Copper ions were introduced into the NH4-form P-CHA by
ion-exchanged method; 1 g of the NH4-form P-CHA was stirred in 100
mL of 0.012 M Cu(CH3COO)2 aqueous solution at room temperature for
24 h. The sample was filtered, washed with distilled water, dried over­
night at 100 � C. Thus, obtained Cu ion exchanged P modified CHA-type
zeolite was denoted as “Cu/P-CHA”. The Si/Al, P/Al and Cu/Al atomic
ratios of Cu/P-CHA were found to be 11, 0.20 and 0.25, respectively,
which were estimated by inductively coupled plasma-atomic emission
spectrometer (ICP-AES, Shimadzu ICPE-9000). The high-resolution 31P
MAS NMR spectra were obtained on a JEOL ECA-600 spectrometer
(14.1 T) equipped with an additional 1 kW power amplifier. The 31P

δ ¼ σref - σ
where σref is the calculated chemical shielding of the P or Al atom in
H3PO4 or Al(NO3)3, and σ is the calculated chemical shielding of the
atom under consideration.
The functional dependence was also examined using the B3LYP
functional [27] with the basis set of 6-31G(d,p). All DFT calculations
2



P. Zhao et al.

Microporous and Mesoporous Materials 294 (2020) 109908

of the framework of CHA is decomposed after the 15-h treatment
(Fig. S1).
These spectroscopic signatures provide valuable information for the
interactions between phosphorus species and the zeolite framework.
Subsequently, in this work, theoretical studies on 31P NMR chemical
shifts have been carried out to disclose the possible formed species for
the P-CHA zeolites. The phosphorus sites are denoted as QPnm , where n
and m represent the number of P–O–P and P–O–Al connectivity,
respectively, with given tetrahedrally coordinated phosphorus.
4.2. Phosphate species
Previous studies on the 31P NMR chemical shifts of orthophosphates,
short chain polyphosphates and other condensed phosphates suggested
that QP00 resonances are observed at 0 � 5 ppm, QP10 at À 7 � 10 ppm,
QP20 at À 17 � 10 ppm and QP30 at À 40 � 8 ppm [29–32], presenting
considerable overlaps in chemical shifts caused by different types of
phosphorous. In order to improve spectral assignments, various phos­
phates including linear and branched polyphosphates as well as cyclic
phosphates as shown in Fig. 2 were first considered to evaluate their 31P
NMR chemical shifts.
As shown in Table 1, the calculated 31P NMR chemical shift of the
end groups (QP10 ) in pyrophosphate H4P2O7 are around À 2 ppm, while
the chemical shift of the middle group (QP20 ) in H5P3O10 is À 34 ppm.
Three configurations were considered for the polyphosphate H6P4O13:
the linear one has the lowest energy and the relative energies of two

Fig. 1. 31P MAS NMR spectra of the Cu ion exchanged P modified CHAtype zeolites.


were conducted using Gaussian 09 suite of programs version E.01 [28].
4. Results
4.1. 31P MAS NMR spectra of P-CHA zeolites under hydrothermal
treatment
Fig. 1 presents the 31P MAS NMR spectra of the Cu/P-CHA zeolites
before and after the hydrothermal treatment, the introduction of P into
the zeolite leads to several characteristic peaks around À 27, À 33, and
À 40 ppm in the fresh state, accompanied by the weak shoulders of the
peaks centered at À 14 and À 22 ppm. Apparently, the peak around À 33
pm is dominant, presenting the highest intensity in the considered range
of chemical shift. Meanwhile, the peaks around À 27 and À 40 ppm also
exhibit considerable intensities. After the hydrothermal treatment at
900 � C, changes in the position and intensity of 31P NMR peaks can be
observed, indicating the structural changes of the formed species during
the hydrothermal treatment. The peaks around À 33 ppm are always
dominant during the 1–15 h treatment. It should be noted that the peak
intensity around À 42 ppm drastically increases after the 1-h treatment,
which is still strong after 7-h but decreases after the 9–15 h treatment.
Meanwhile, the intensity of peaks around À 29 ppm increases after the
7–9 h treatment. The XRD and 29Si MAS NMR spectra reveal that a part

Table 1
The 31P NMR chemical shifts (in ppm) and Mulliken charges of phosphorus in
phosphates.
Phosphate

P type

NMR


Charge

H3PO4

QP00

0.0

1.216

2.4

1.286

QP20 a

À 34.0

1.343

32.4
27.3
30.4
16.2
36.6
24.4

1.355
1.340

1.353
1.381
1.378
1.337

H4P2O7
H5P3O10
H3P3O9

QP10
QP20

L-H6P4O13

QP20

A-H6P4O13

QP30

B–H6P4O13
P4O10

QP30
QP30

a
b
c
b

c
a

À
À
À
À
À
À

a

À 37.6

1.405

À 52.9

1.387

Fig. 2. Optimized structures of different phosphate species. The lowercase letters (a–d) label distinct phosphorus atoms. The values in parentheses represent relative
energies for H6P4O13 with different configurations (in kcal⋅molÀ 1). Color code: O red; H white; P orange. (For interpretation of the references to colour in this figure
legend, the reader is referred to the Web version of this article.)
3


P. Zhao et al.

Microporous and Mesoporous Materials 294 (2020) 109908


Fig. 3. Optimized structures of monodentate aluminophosphates (six coordinate). The lowercase letters label (a–d) distinct phosphorus atoms, and the distances are
in Å. The values in parentheses represent relative energies in three H6P4O13–AlO3H3(H2O)2 configurations (in kcal⋅molÀ 1). Color code: O red; H white; Al blue; P
orange. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

branched configurations are very close. Two middle groups (QP20 ) in LH6P4O13 resonate at À 16.2 and À 36.6 ppm for b and c, respectively. AH6P4O13 and B–H6P4O13 exhibit different 31P NMR chemical shifts for
QP30 , that is, À 24.4 and À 37.6 ppm, indicating the chemical shift is
sensitive to the difference in configuration, which may originate from
the discrepancy in charge distribution. Chemical shifts of the end groups
in H5P3O10 and H6P4O13 fall into the range of À 2.1~-10.1 ppm
(Table S1).
As for the species after the intense dehydration, the QP20 atoms
resonate at À 32.4, À 30.4, and À 27.3 ppm in the cyclic phosphate
H3P3O9, and four QP30 atoms in P4O10 resonate at around À 53 ppm. In
addition, water and chloroform were taken into account to evaluate the
solvent effect on chemical shifts, which was found to be negligible
within the polarizable continuum model (PCM) (Table S1). Overall, the
31
P resonances are shifted towards higher field as the number of P–O–P
increases, which exhibits the similar trend to previously experimental
studies [26,33]. However, except for one QP20 (b) in L-H6P4O13 and QP30
in P4O10, other QP20 and QP30 atoms possess rather close chemical shifts in
the range of À 24 ~ À 37 ppm. Based on these results, the dominant peak
observed at around À 33 ppm in Fig. 1 may have some contributions
from QP20 or QP30 groups, in particular after the partial decomposition of
the zeolite framework in the present work. Additionally, as for the
considered phosphates, the P atoms with higher positive charge tend to
exhibit higher-field 31P NMR chemical shifts (Table 1), indicating the
charge distribution may affect the 31P chemical shift. The similar trends
can also be found at the B3LYP/6-31G(d,p) level of theory (Table S1).


Table 2
The 31P and 27Al chemical shifts (in ppm) and Mulliken charges of P and Al in
monodentate aluminophosphates.
Aluminophosphate

P type

H3PO4–AlO3H3(H2O)2

QP01

H4P2O7–AlO3H3(H2O)2
H5P3O10–AlO3H3(H2O)2
A-H6P4O13–AlO3H3(H2O)2
B–H6P4O13–AlO3H3(H2O)2
L-H6P4O13–AlO3H3(H2O)2

31
P
NMR

Charge

27

Al
NMR

Charge


9.2

1.332

5.5

1.028

QP11

a

À 1.9

1.378

2.0

1.101

QP21

a

À 20.0

1.389

5.6


1.096

QP31

a

À 25.8

1.432

1.5

1.090

QP31

a

À 26.2

1.490

4.9

1.094

QP11

a


À 2.5

1.493

8.7

1.057

QP20

b
c

À 16.9
À 32.1

1.380
1.340

atoms tend to migrate to the hydroxyl connected to aluminum, which
leads to the formation of hydrogen bonds. As shown in Table S2, except
for the Al atom bound to H5P3O10 and H4P2O7, the 31P resonance dif­
ference between the tetrahedral and octahedral coordinate aluminum is
small in other structures, while the Al atoms exhibit typical resonances
for the tetrahedral or octahedral coordinate structure [7,16]. This work
focuses on the six-coordinate octahedral aluminum phosphates.
As shown in Table 2, the 31P NMR of aluminophosphates tend to shift
to lower field compared to the corresponding phosphates. The P atom of
QP01 in H3PO4–Al(OH)2(H2O)2 resonates at 9.2 ppm. The QP21 , and QP31
(B–H6P4O13–AlO3H3) species are also shifted to lower field, i.e., À 20.0

and À 26.2 ppm, respectively. The QP11 in H4P2O7–AlO3H3(H2O)2, and
QP31 in A-H6P4O13–AlO3H3(H2O)2 have similar chemical shifts (À 1.9 and
À 25.8 ppm) to the corresponding phosphoric acid (À 1.8 and À 24.4
ppm). The QP10 chemical shifts fall into the range of 14.2 to ỵ1.7 ppm
(Table S2). It is found that the linear structure L-H6P4O13–AlO3H3(H2O)2
is more unstable (22.4 kcal molÀ 1) than the branched one. Note that the
QP11 (-2.5 ppm) and QP20 (-32.1 ppm) chemical shifts in LH6P4O13–AlO3H3(H2O)2 are also shifted to lower field after attaching the
Al atom.
When the hydroxyl group (P–OH) of the phosphate is connected to
the Al atom, the attachment of two hydroxyl groups on the aluminum
atom results in the bidentate aluminophosphate species, as shown in
Fig. 4 and Fig. S3. Note that the attachment of three hydroxyl groups

4.3. Aluminophosphate species
Aluminophosphate species considered in the previous study [26]
were assessed since the amorphous extra-framework aluminophosphate
species play a significant role during framework dealumination. Two
types of aluminophosphates were studied via the interaction of
– O) or the hydroxyl group (P–OH).
aluminum with the double bond (P–
Fig. 3 shows the structures with the octahedral coordinate aluminum
– O bond, which results in the
in the first coordination sphere via the P–
monodentate aluminum phosphates. Note that the corresponding
structures with the tetrahedral coordinate aluminum were also consid­
ered (Fig. S2). These optimized structures revealed that some hydrogen
4


P. Zhao et al.


Microporous and Mesoporous Materials 294 (2020) 109908

Fig. 4. Optimized structures of bidentate aluminophosphates. The lowercase letter label (a–d) distinct phosphorus atoms. Color code: O red; H white; Al blue; P
orange. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

aluminophosphate species are slightly shifted to lower field compared to
the corresponding monodentate structures, as shown in Table 3.
Meanwhile, it is found these species possess low-field 27Al-NMR chem­
ical shifts compared to other aluminumphosphate species.
As for bidentate aluminophosphate species containing a hexagon, the
aluminum atom is connected to a hydroxyl group of one phosphate and a
double bond of another phosphate. The QP11 chemical shift exhibits a
high-field value of À 20.1 ppm, and the QP21 resonance in the condensed
phosphate is higher-field (À 27.4 ppm). Additionally, the formed
hexagon-structure is much more stable than the tetragon-structure,
which may play an important role in aluminophosphate species.

Table 3
The 31P and 27Al chemical shifts (in ppm) and Mulliken charges of P and Al in
bidentate aluminophosphates.
Aluminophosphate

P type

H2PO4–AlO2H2 (H2O)2

QP02

A-H3P2O7–AlO2H2(H2O)2

H4P3O10–AlO2H2(H2O)2
B–H3P2O7–AlO2H2(H2O)2
H5P4O13–AlO2H2(H2O)2

31

P
NMR

Charge

27

Al
NMR

Charge

13.4

1.497

14.7

0.958

QP12

a


3.0

1.517

17.5

0.986

a

À 14.0

1.538

18.1

0.976

QP11

a
b
a
b

À
À
À
À


20.1
2.3
27.4
18.2

1.392
1.339
1.448
1.405

3.5

1.042

3.2

1.069

QP22

QP21

4.4. Geometry structure of the CHA zeolites
As for the CHA framework, the 3NN structure with two protons
located on 6-MR and 8-MR has the lowest relative energy (Table S4),
which is considered in the following calculations of P-CHA zeolite. As
shown in Fig. 5, as for the Al/Si–O bonds without the Brønsted proton on
the O atom, the distances of Al–O bonds fall into the range of 1.69–1.76
Å, while those of Si–O bonds are around 1.59–1.63 Å, which are
consistent with the experimental data (1.70–1.73 Å and 1.58–1.64 Å for

Al–O and Si–O bond distances, respectively) [34]. Nevertheless, the
Si–O and Al–O bonds in the vicinity of Brønsted acid sites are much
longer, that is, 1.895 and 1.926 Å for the Al–O bonds; 1.706 and 1.721 Å
for the Si–O bonds. The O–H bond distances are around 0.97 Å. Exper­
imentally, the 27Al resonance of the CHA zeolite is detected at 58 ppm
[35]. However, the calculated 27Al chemical shifts are 33.3 and 26.0
ppm for α and β positions, respectively, whose Mulliken charges are
1.056 e (α) and 1.109 e (β). The confinement of the finite computational
model as well as the methodology errors should contribute to the
observed difference between theoretical and experimental values.
Consequently, a constant shift of ỵ24.7 ppm is applied to the following
27
Al chemical shifts, which is a simple way to improve the calculated
values. Therefore, the corrected 27Al chemical shifts in the pristine
framework are 58.0 and 50.7 ppm for the α and β Al atoms, respectively.

Fig. 5. The optimized structure (left) of the most stable structure (3NN-6R8R)
and the geometrical parameters in the 6-MR (right, distances are in Å). Color
code: O red; Si tan; H white; Al blue. (For interpretation of the references to
colour in this figure legend, the reader is referred to the Web version of
this article.)

4.5. Interaction of phosphorus with the Brønsted acid sites of the CHA
framework

leads to the open-shelled electronic structure of aluminophosphates,
which is not considered in the present calculations. The six-coordinate
phosphates attached with two extra H2O molecules are shown in
Fig. 4. It is found that the 31P NMR chemical shifts in bidentate


Previous studies have proposed various structures for the in­
teractions between phosphates and the Brønsted acid site as illustrated
5


P. Zhao et al.

Microporous and Mesoporous Materials 294 (2020) 109908

influence the catalytic activity [9]. The hydrothermal stability, on the
other hand, is enhanced by the intra-framework phosphates [36,37].
In M0 structure, the phosphate approaches the Brønsted acid site
with the distance of 1.707 Å and the bridging O–H bond being slightly
lengthened, indicating the formation of a hydrogen bond. The calculated
interaction energy (À 29.0 kcal molÀ 1) also suggests the formation of M0
is favorable. The 31P NMR chemical shift is À 6.0 ppm, which is in

Table 4
Adsorption or reaction energies (ΔEr, in kcal⋅molÀ 1), 31P and 27Al chemical
shifts (in ppm) and Mulliken charges in structures shown in Fig. 6.
Charge

27

À 6.0

1.338

α


QP00

4.6

1.359

α

1.2

QP01

À 5.7

1.445

α

M2

13.8

QP01

À 17.2

1.423

α


M3

38.7

QP00 Si2

À 23.5

1.512

α

M4

À 36.6

QP00

7.3

1.430

α

M5

À 8.8

QP01


À 31.9

1.590

α

Model

ΔEr

M0

Scheme 1. Chemical structures (M0-M5) of the interaction models of phos­
phorus with the Brønsted acid sites proposed by (M0) Abubakar et al. [36],
(M1) Kaeding et al. [6], (M2) Lercher et al. [15], (M3) Xue et al. [38], (M4, 5)
Blasco et al. [7].

in Scheme 1, and their corresponding optimized structures are shown in
Fig. 6. The 31P NMR chemical shifts of these structures were investigated
for the identifications of the present hydrothermal treatment and also
for the reference data of the future related works. The calculated reac­
tion energies, Mulliken charges, and NMR chemical shifts of the P and Al
atoms are listed in Table 4. It has been recognized that these formed
species are essential for decreasing the zeolite acid strength and

31

P NMR

À 29.0


QP00

M0-b

À 35.7

M1

a

β
β
β
β
β
β
β

Al NMRa
57.4
51.9
58.1
50.6
54.1
49.4
57.0
57.0
50.4
49.6

60.7
50.2
53.2
49.2

Charge
1.073
1.103
1.055
1.107
1.073
1.104
1.086
1.116
1.103
1.110
1.056
1.098
1.082
1.107

Values corrected by 24.7 ppm.

Fig. 6. Optimized structures of the interaction models of phosphorus with the Brønsted acid sites. The α and β represent two Al atoms. Color code: O red; Si tan; H
white; Al blue; P orange. The terminal H atoms are omitted for clarity. (For interpretation of the references to colour in this figure legend, the reader is referred to the
Web version of this article.)
6


P. Zhao et al.


Microporous and Mesoporous Materials 294 (2020) 109908

Scheme 2. Chemical structures (M6-M10) of the interaction models of diphosphates with the Al sites in the CHA zeolite.

higher-field than free H3PO4 due to the increased charge. The adsorption
– O group in­
energy is more negative (À 35.0 kcal molÀ 1) when the P–
teracts the Brønsted acid site (M0-b) via a hydrogen bond, as shown in
Fig. S4. However, the M0-b structure exhibits a low-field 31P resonance
of 4.6 ppm. The chemical shifts of two Al atoms in these two structures
are almost the same as those in the pristine zeolite.
The phosphorous can bound to framework via the formation of tetư
rahydroxy phosphonium cation P(OH)4ỵ [7], M4 and M5, having
different H3PO4 configurations and coordination numbers of the P atom
after protonation. M4 leads to the most stable structure (À 36.6 kcal
molÀ 1). A recent study proposed that M4 is an effective phosphorous
species for improving the framework stability in P-modified ZSM-5 ze­
olites [37]. Meanwhile, the calculated 31P NMR chemical shift is 7.3
ppm, which suggests that M4 is not the dominant species observed here.
For M5, the adsorption of the protonated orthophosphoric acid results in
a smaller adsorption energy of À 8.8 kcal molÀ 1 compared to M0 and M4.
It is found that the 31P resonance in M5 is significantly shifted to high
field, i.e., À 31.9 ppm, which suggests this structure may have a signif­
icant contribution to the 31P NMR MAS spectra observed here and
therefore, is a possible candidate in the initial stage of P-CHA zeolite.
Compared to the pristine framework, the 27Al resonances (α) in M4 and
M5 are shifted to lower and higher field, respectively.
In M1 structure, the interaction of orthophosphate with the bridging
hydroxyl leads to the dehydration of one water molecule and generates a

slightly less stable structure. The calculated 31P NMR chemical shift is
À 5.7 ppm, which is similar to that in M0 even though the P atom in M1 is
more positively charged. M2 structure has the Si–O–P–O–Al structure by
breaking an Al–O bond, resulting in a positive reaction energy of 13.8
kcal molÀ 1. Note that the addition to the Brønsted acid site in 8-MR was
also considered, but the geometry optimization leads to the more stable
structure like M1, indicating the steric repulsion between the phosphate
and the 4-MR is an advantage of the formation of M2. It is found that M2
shows high-field 31P resonance of À 17.2 ppm than M1. In M3, the P
atom is connected with two oxygen atoms by breaking two Al–O bonds,
and the dehydration caused by the orthophosphate and two Brønsted
acid hydrogen atoms leads to a less stable structure with the energy of
38.7 kcal molÀ 1. The corresponding 31P resonance of À 23.5 ppm is

shifted higher field compared to M2, which is also a promising species in
P-CHA in terms of the chemical shift.
Overall, the high-field 31P NMR chemical shift was found in M5
(À 31.9 ppm), followed by M3 (À 23.5 ppm) and M2 (À 17.2 ppm) among
the considered models. M0, M4, and M5 described the interactions of
phosphate with the framework via hydrogen bonds, which result in
negative adsorption energies that may be indicative of the formation
upon the impregnation of phosphoric acid. In contrast, M1, M2, and M3
were unstable because of the dehydration or breakage of the Si–O–Al
bond, which should be difficult to revert by single cationic exchange or
hot-water washing [7]. Compared to the pristine framework, the small
changes in the 27Al chemical shift also reveal that the interaction be­
tween phosphate groups and framework has a minor effect on the 27Al
resonances.
To further disclose the 31P NMR chemical shifts of condensed phos­
phates in the zeolite framework, the pyrophosphoric acid was also

entrained in the mentioned models (Fig. S5 and Table S5). When the
pyrophosphoric acid is attached, all the structures M00 -M50 except for
M30 exhibit more favorable reaction energies than orthophosphoric acid,
especially for M5’ (À 20.3 kcal molÀ 1), indicating there is a high possi­
bility to trap pyrophosphate in the P-modified zeolites. As for M30 , the
reaction energy (39.1 kcal molÀ 1) is comparable to that for orthophos­
phate. Importantly, the 31P NMR chemical shifts are shifted to higher
field, and the shifted values depend on the models. The chemical shift in
M50 is shifted to À 54.0 ppm from À 31.9 ppm, showing the largest
change. M30 exhibits the smallest change of 4.6 ppm. The shifted values
for other models fall into the range of 7.1–13.5 ppm. The 31P NMR
chemical shifts labeled as a are À 19.5, À 54.0, À 26.7, and À 28.1 ppm for
M00 , M50 , M20 , and M30 , respectively. The end group (b) exhibits the
relatively high-field resonance in M4’ (À 19.2 ppm), M1’ (À 16.3 ppm),
and M2’ (À 13.6 ppm).
4.6. Interaction of phosphorus with the aluminum atom in the CHA
framework
Previous studies proposed that the presence of phosphorous around
the tetrahedral Al atom in the framework would lead to the distorted
7


P. Zhao et al.

Microporous and Mesoporous Materials 294 (2020) 109908

Fig. 7. Optimized structures of the interaction models of diphosphates with the Al sites. The α and β represent two Al atoms. Color code: O red; Si tan; H white; Al
blue; P orange. The terminal H atoms are omitted for clarity. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web
version of this article.)


phosphates were considered to construct the intra-framework structures
bound to framework Al atom (Scheme 2 and Fig. 7) [26].
As shown in Table 5, the structures with phosphates bound to the
framework tetrahedral aluminum have large reaction energies in the
range of -32.8 ~ À 68.5 kcal molÀ 1, indicating that these intraframework structures of M6-M10 are thermodynamically favorable.
Note that M7 leads to the largest reaction energy of À 68.5 kcal molÀ 1,
followed by M8 with a reaction energy of À 58.7 kcal molÀ 1. The results
also further prove that the P-CHA zeolite frameworks prefer to be sta­
bilized by expelling the framework Al.
– O bond of diphosphate and two monophosphates at­
When the P–
tacks the Al atom in the framework, respective M6 and M7 structures are
generated. The shortest Al–O distances between phosphate and the Al
atom are 1.946 and 1.904 Å for M6 and M7, respectively. Note that the
other H3PO4 molecule in M7 has a longer distance with the framework
tetrahedral aluminum (3.550 Å), which should be ascribed to the steric
repulsion. It is found that the QP11 atom in M6 and the QP01 atom in M7
have higher-field chemical shifts (À 11.5 and À 2.1 ppm) compared to
free H4P2O7–AlO3H3 (7.2 ppm) and H3PO4–AlO3H3 (8.1 ppm).
M9 and M10 represent two possible structures that the framework
aluminum is attacked by the hydroxyl group of phosphates. Reaction
energies of M9 and M10 are À 37.7 and À 32.8 kcal molÀ 1, respectively,
indicating the former one is slightly stable. In the case of M10, there is a
– O bond and the other bridging hydroxyl.
hydrogen bond between the P–
1
The QP2 chemical shift in M9 is 11.8 ppm, which is comparable with the
isolated one (10.1 ppm). The QP11 and QP10 chemical shifts in M10 are in

Table 5

Absorption or reaction energies (ΔEr, in kcal⋅molÀ 1), 31P and 27Al NMR chemical
shifts (in ppm) and Mulliken charges in structures shown in Fig. 7.
Model

ΔEr

M6

À 46.9

M7
M8
M9
M10

a

À 68.5
À 58.7
À 37.7
À 32.8

31

P NMR

Charge

27


Al NMRa

Charge

a

À 11.5

1.363

α

57.9

1.076

QP10

b

À 5.8

1.470

β

44.1

1.084


a

À 2.1

1.512

α

55.7

1.079

QP00

b

6.8

1.306

β

42.5

1.103

a

À 16.9


1.455

α

57.4

1.063

QP11

b

À 7.5

1.444

β

36.2

1.191

a

11.8

1.530

α


56.0

1.095

QP10

b

À 7.2

1.345

β

45.8

1.078

a

À 18.8

1.420

α

55.9

1.067


QP10

b

À 16.3

1.405

β

52.6

1.157

QP11
QP01
QP11
QP12
QP11

Values corrected by 24.7 ppm.

O–Al–O structures, which was caused by the hydrolysis of the tetrahe­
dral Al atoms that can be facilitated by acidic conditions [10,26].
Meanwhile, it was pointed out that the formation of Al–O–P bonds may
take place within the framework in a manner similar to the onset of
dealumination process during steaming, leading to the formation of the
local SAPO interfaces [25,26]. Eventually, the process can result in the
formation of amorphous extra-framework aluminophosphate species
discussed above [26]. In this work, monodentate and bidentate of

8


P. Zhao et al.

Microporous and Mesoporous Materials 294 (2020) 109908

Fig. 8. Optimized structures of phosphorus-framework interaction based on M8. (A) dealumination caused by diphosphates; (B) phosphates containing three P
atoms; (C) phosphates containing four P atoms. Color code: O red; Si tan; H white; Al blue; P orange. The terminal H atoms are omitted for clarity. (For interpretation
of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
Table 6
Reaction energies (ΔEr, in kcal⋅molÀ 1), 31P and 27Al NMR chemical shifts (in
ppm) and Mulliken charges in structures shown in Fig. 8.
Structure

ΔEr

A

À 59.3

B

C

a

À 64.9

À 65.7


31

P NMR

Charge

27

Al NMRa

Charge

a

À 21.8

1.454

α

56.5

1.071

QP11

b

À 5.8


1.427

β

64.0

1.135

a

À 27.8

1.489

α

57.8

1.059

QP11

b

À 12.0

1.453

β


36.6

1.181

QP10

c

À 15.2

1.389

QP21

a

À 33.0

1.492

α

55.8

1.063

QP11

b


À 3.9

1.443

QP20

c

À 34.8

1.462

β

35.3

1.204

QP10

d

À 4.2

1.336

QP11
QP21


Values corrected by 24.7 ppm.

slightly higher-field than the isolated structure (À 14.4 and À 13.9 ppm,
Fig. S6).
– O group simultaneously attack the Al
One hydroxy group and one P–
atom, resulting in the formation of a hexagon as shown in M8. Compared
to resonances of the isolated structure, B–H3P2O7–AlO2H2 (À 22.1 and
3.1 ppm for a and b), one QP11 (a: -16.9 ppm) is slightly low-field, while
the other QP11 (b: -7.5 ppm) is high-field. Fig. 8(A) shows the structure of
M8 after expelling the framework Al atom, in which the QP11 (a, À 21.8
ppm) is close to the corresponding value in the isolated structure, and
the Q P11 (b, À 5.8 ppm) is close to the corresponding value in M8. Due to
the high-field 31P NMR chemical shift in M8 as well as the large reaction
energy, the condensed phosphates in the framework based on M8 are
also studied as shown in Fig. 8. As summarized in Table 6, the chemical
shift of the P atom (a) is further shifted to the lower chemical shifts of
À 27.8 ppm (B) and À 33.0 ppm (C).
In addition, it is found that the 27Al-NMR chemical shifts (β) in M6-9
are shifted to higher field due to the distorted tetrahedral aluminum in
the framework [39]. Particularly, M8 exhibits the smallest 27Al reso­
nance (36.2 ppm), and chemical shifts of other models are in the order of
M7 (42.5 ppm) < M6 (44.1 ppm) < M9 (45.8 ppm). Interestingly, the
corresponding 27Al NMR chemical shift in M10 is slightly shifted to
lower field (52.6 ppm), which may suggest a small distortion for the Al
atom.

Fig. 9. Optimized structures of the SAPO species based on the 42 T structure.
Color code: O red; Si tan; H white; Al blue; P orange. The terminal H atoms are
omitted for clarity. (For interpretation of the references to colour in this figure

legend, the reader is referred to the Web version of this article.)

4.7. Silicoaluminophosphate (SAPO) species in the CHA framework
In the present preparation procedure of P-CHA [11,13] and consid­
ering the contents of aluminum and phosphorus in the sample (Si/Al ¼
11 and P/Al ¼ 0.2), it is assumed that only one-phosphorus species exist
in one-unit cell of zeolite cage. To simulate this P-CHA cage, the possible
SAPO species are constructed in the 6-MR of the framework, in which
one framework Si atom is replaced by one P atom and one or two Al
atoms are placed in the 6-MR, as shown in Fig. 9. Note that one proton is
9


P. Zhao et al.

Microporous and Mesoporous Materials 294 (2020) 109908

sites (M0-M5), M3 and M5 generate significant 31P resonances at À 23.5
and À 31.9 ppm, respectively, while other models exhibit relatively lowfield resonances, as shown in Fig. 10b. When one more H3PO4 molecule
is condensed on these models by dehydration, 31P resonances are shifted
to higher field, for example, the QP11 resonance in M2 (À 26.7 ppm) and
M5 (À 54.0 ppm); the QP10 Si2 resonance in M3 (À 28.1 ppm); the QP10
resonances in M0 (À 19.5 ppm), M4 (À 19.2 ppm), and M1 (À 16.3 ppm).
In the case of dealuminated structures caused by phosphates (M6M10), the QP11 resonances in M8 (À 16.9 ppm) and M10 (À 18.8 ppm)
may play a significant role, while other models exhibit relatively lowfield 31P resonances, as shown in Fig. 10b. As for the condensed phos­
phates in M8, higher-field resonances were found for QP21 (-27.8 and
À 33.0 ppm for B and C in Fig. 8).
The SAPO species in the framework may be generated by substituting
the Si atom with the P atom in the hydrothermal treatment. The for­
mation of SAPO species in the 6-MR zeolite framework results in sig­

nificant resonances at the range of À 35.8 ~ À 48.8 ppm.
Based on the theoretical results and the contents of Si, Al, and P
atoms in the sample, the possible assignments of the 31P MAS NMR
spectra of P-CHA zeolites before and after the hydrothermal treatment
are considered. The SAPO species generated in the framework can be
candidates for the peaks with increased intensity at À 42 ppm after 1-h
treatment, and the intensity-decrease at À 42 ppm after the long-time
treatment should be caused by the partial framework decomposition
with the SAPO species. Meanwhile, the highly condensed phosphates
may also contribute to the peak at À 42 ppm. Considering the decrease in
the crystallinity of zeolite by the hydrothermal treatment with long time
(Fig. S1), the peaks with increased intensity at around À 29 ppm after
treatment for over 7–9 h should be due to the accumulation of amor­
phous extra-framework phosphate and aluminophosphate species
caused by the framework decomposition, such as the QP20 resonances in
H3P3O9, the QP31 resonances in monodentate, and the QP11 and QP21
resonances in bidentate with a hexagon. As for the dominant peak at
À 33 ppm in all 31P NMR spectra, the phosphates in the framework like
M5 and M8 may contribute to the peak.

Table 7
Relative energies (ΔE, in kcal⋅molÀ 1), 31P and 27Al chemical shifts (in ppm) and
Mulliken charge in structures shown in Fig. 9.
model

ΔEa

31

A1

A2
B1

0.0
0.3
0.0

À 48.8
À 40.1
À 40.5

B2

5.5

P NMR

À 35.8

Charge

27

1.505
1.518
1.558

54.2
50.2


1.537

Al NMRb

α
β

α
β

a
b

57.0
51.9
61.0
51.1

Charge
1.088
1.099
1.059
1.121
1.064
1.114

Relative energy with respect to the most stable species.
Values corrected by 24.7 ppm.

added to compensate the negative charge in the model with two Al

atoms. Based on the L€
owenstein’s rule [19], the possible positions for
the Al atom are considered to determine the stable structures (Fig. S7),
and the structures with the Al atom next to the P atom are much more
stable. As listed in Table 7, the 31P resonances of these stable SAPO
species locate at the much higher field as À 40.1 ppm (A2), À 48.8 ppm
(A1), À 40.9 ppm (B1), and À 35.8 ppm (B2), species of which should
contribute to the observed 31P NMR chemical shifts at À 42 ppm in the
initial stage of the hydrothermal treatment. The chemical shifts of the
framework tetrahedral aluminum are slightly smaller than the pristine
framework.
5. Discussion
Based on the present theoretical results, the possible structures are
qualitatively clarified for the observed 31P resonances in the range of
À 27 ~ À 42 ppm. Fig. 10 summarizes the 31P NMR chemical shifts of the
species examined in this work. For the phosphates without aluminum,
QP20 and QP30 mainly fall into the range of À 24 ~ À 37 ppm, as shown in
Fig. 10a. After attaching the aluminum atom to phosphates, some 31P
resonances are shifted to the lower field. One QP21 and two QP31 reso­
nances are found at À 20.0, À 25.8 and À 26.2 ppm for monodentate
aluminophosphate species. Note that bidentate aluminophosphate spe­
cies caused by attaching the Al to the hydroxyl group of the phosphate
are slightly shifted to lower field compared to the corresponding mon­
odentate. The bidentate aluminophosphate species forming a hexagon
exhibit a significant resonance at À 20.1 ppm for QP11 , which possesses
higher-field chemical shifts for condensed species (À 27.4 ppm for QP21 ).
For the possible interactions of phosphorus with the Brønsted acid

6. Conclusions
The 31P MAS NMR spectra of P-CHA zeolites were observed under the

hydrothermal treatment. Characteristic changes of the spectra were
identified in the range of À 27~-42 ppm. It was observed that the

Fig. 10. The 31P NMR chemical shifts of different types of phosphorus species considered in this work. (a) Phosphates (orange) and aluminophosphates (blue:
monodentate; magenta: bidentate); (b) P-CHA models (purple diamond: structures shown in Figs. 6, 7 and 9; green triangle: structures shown in Fig. S5; circle:
structures A (dark cyan), B (pink), and C (wine) shown in Fig. 8. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web
version of this article.)
10


P. Zhao et al.

Microporous and Mesoporous Materials 294 (2020) 109908

resonance at À 42 ppm significantly increases after 1-h treatment, while
the resonance at À 29 ppm increases after further treatment for 7-h. The
31
P NMR chemical shifts of a wide range of phosphates have been
theoretically studied by the DFT calculations to propose the identifica­
tion of phosphate species generated by the hydrothermal treatment of PCHA zeolite, that may correlate to the structural changes of P-CHA
zeolite.
The results revealed that the 31P resonances in phosphates and alu­
minophosphates are shifted to higher field in the condensed species.
Interestingly, the bidentate aluminophosphate species forming a hexa­
gon exhibit a rather high-field resonance at À 20.1 ppm for QP11 . Note
that the 31P resonances in aluminophosphates are slightly shifted to
lower field compared to the corresponding phosphates without
aluminum, particularly, the shifted values in bidentate structures are
larger than those in monodentate structures.
Six possible models have been considered to study the interactions

between phosphorus with the Brønsted acid sites. M3 and M5 show the
31
P resonances at À 23.5 and À 31.9 ppm and therefore, these structures
are one of the candidates of the P-CHA prepared here. Five models are
also considered for the interactions of phosphorus with the framework
aluminum, in which the QP11 resonances in M8 (À 16.9 ppm) and M10
(À 18.8 ppm) may play a significant role, while other models exhibit
relatively low-field 31P resonances.
The increased intensity at À 42 ppm in the initial stage of the hy­
drothermal treatment is mainly attributed to the SAPO species in the
framework, while the increased intensity at À 29 ppm in the later stage
of treatment is ascribed to the accumulation of the extra-framework
condensed phosphate and aluminophosphate species caused by the
partial framework decomposition. The dominant peak at À 33 ppm in all
31
P NMR spectra is assigned to the phosphates in the framework like M5
and M8. Overall, the present theoretical results provide useful infor­
mation regarding the 31P NMR chemical shifts in the P-CHA, which will
significantly improve the 31P NMR assignments in future experiments.

[6] W.W. Kaeding, S.A. Butter, Production of chemicals from methanol: I. Low
molecular weight olefins, J. Catal. 61 (1980) 155–164.
[7] T. Blasco, A. Corma, J. Martínez-Triguero, Hydrothermal stabilization of ZSM-5
catalytic-cracking additives by phosphorus addition, J. Catal. 237 (2006) 267–277.
[8] G. Lischke, R. Eckelt, H.G. Jerschkewitz, B. Parlitz, E. Schreier, W. Storek,
€
B. Zibrowius, G. Ohlmann,
Spectroscopic and physicochemical characterization of
P-modified H-ZSM-5, J. Catal. 132 (1991) 229–243.
[9] H.E. van der Bij, B.M. Weckhuysen, Phosphorus promotion and poisoning in

zeolite-based materials: synthesis, characterisation and catalysis, Chem. Soc. Rev.
44 (2015) 7406–7428.
[10] J. Caro, M. Bülow, M. Derewinski, J. Haber, M. Hunger, J. K€
arger, H. Pfeifer,
W. Storek, B. Zibrowius, NMR and IR studies of zeolite H-ZSM-5 modified with
orthophosphoric acid, J. Catal. 124 (1990) 367–375.
[11] S.V. Priya, T. Ohnishi, Y. Shimada, Y. Kubota, T. Masuda, Y. Nakasaka,
M. Matsukata, K. Itabashi, T. Okubo, T. Sano, N. Tsunoji, T. Yokoi, M. Ogura,
A collective case screening of the zeolites made in Japan for high performance
NH3-SCR of NOx, Bull. Chem. Soc. Jpn. 91 (2017) 355–361.
[12] Y. Kakiuchi, T. Tanigawa, N. Tsunoji, Y. Takamitsu, M. Sadakane, T. Sano,
Phosphorus modified small-pore zeolites and their catalytic performances in
ethanol conversion and NH3-SCR Reactions, Appl. Catal. Gen. 575 (2019) 204–213.
[13] Y. Yamasaki, N. Tsunoji, Y. Takamitsu, M. Sadakane, T. Sano, Synthesis of
phosphorus-modified small-pore zeolites utilizing tetraalkyl phosphonium cations
as both structure-directing and phosphorous modification agents, Microporous
Mesoporous Mater. 223 (2016) 129–139.
[14] J.C. V�edrine, A. Auroux, V. Bolis, P. Dejaifve, C. Naccache, P. Wierzchowski, E.
G. Derouane, J.B. Nagy, J.-P. Gilson, J.H.C. van Hooff, J.P. van den Berg,
J. Wolthuizen, Infrared, microcalorimetric, and electron spin resonance
investigations of the acidic properties of the H-ZSM-5 zeolite, J. Catal. 59 (1979)
248–262.
[15] J.A. Lercher, G. Rumplmayr, Controlled decrease of acid strength by
orthophosphoric acid on ZSM5, Appl. Catal. 25 (1986) 215–222.
[16] N. Xue, R. Olindo, J.A. Lercher, Impact of forming and modification with
phosphoric acid on the acid sites of HZSM-5, J. Phys. Chem. C 114 (2010)
15763–15770.
[17] L.S. Dent, J.V. Smith, Crystal structure of chabazite, a molecular sieve, Nature 181
(1958) 1794–1796.
[18] />[19] W. Loewenstein, The distribution of aluminum in the tetrahedra of silicates and

aluminates, Am. Mineral. 39 (1954) 92–96.
[20] E. Borfecchia, P. Beato, S. Svelle, U. Olsbye, C. Lamberti, S. Bordiga, Cu-CHA – a
model system for applied selective redox catalysis, Chem. Soc. Rev. 47 (2018)
8097–8133.
[21] Y. Zhao, D.G. Truhlar, A new local density functional for main-group
thermochemistry, transition metal bonding, thermochemical kinetics, and
noncovalent interactions, J. Chem. Phys. 125 (2006), 194101.
[22] Y. Zhao, D.G. Truhlar, Density functionals with broad applicability in chemistry,
Acc. Chem. Res. 41 (2008) 157–167.
[23] Y. Zhao, D.G. Truhlar, Improved description of nuclear magnetic resonance
chemical shielding constants using the M06-L meta-generalized-gradientapproximation density functional, J. Phys. Chem. A 112 (2008) 6794–6799.
[24] K. Wolinski, J.F. Hinton, P. Pulay, Efficient implementation of the gaugeindependent atomic orbital method for NMR chemical shift calculations, J. Am.
Chem. Soc. 112 (1990) 8251–8260.
[25] H.E. van der Bij, B.M. Weckhuysen, Local silico-aluminophosphate interfaces
within phosphated H-ZSM-5 zeolites, Phys. Chem. Chem. Phys. 16 (2014)
9892–9903.
[26] K. Damodaran, J.W. Wiench, S.M. Cabral de Menezes, Y.L. Lam, J. Trebosc, J.
P. Amoureux, M. Pruski, Modification of H-ZSM-5 zeolites with phosphorus. 2.
Interaction between phosphorus and aluminum studied by solid-state NMR
spectroscopy, Microporous Mesoporous Mater. 95 (2006) 296–305.
[27] A.D. Becke, Density-functional thermochemistry. III. The role of exact exchange,
J. Chem. Phys. 98 (1993) 5648–5652.
[28] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman,
G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, Gaussian 09, Revision A.01,
Gaussian, Inc., Wallingford, CT, 2009.
[29] E.C. de Oliveira Lima, J.M. Moita Neto, F.Y. Fujiwara, F. Galembeck, Aluminum
polyphosphate thermoreversible gels: a study by 31P and 27Al NMR spectroscopy,
J. Colloid Interface Sci. 176 (1995) 388–396.
[30] L. Zhang, H. Eckert, Multinuclear NMR studies on the sol-gel preparation of sodium
aluminophosphate glasses, Solid State NMR 26 (2004) 132–146.

[31] L. Montagne, G. Palavit, M. Draoui, Mechanism of polyphosphate gel formation in
the Na2O Al2O3 P2O5 system, J. Non-Cryst. Solids 155 (1993) 115–121.
[32] A. Grimmer, G. Wolf, 31P-MAS-NMR studies on crystalline and vitreous
polymorphs of phosphorus pentoxide P2O5, Eur. J. Solid State Inorg. Chem. 28
(1991) 221–232.
[33] M. Derewinski, P. Sarv, X. Sun, S. Müller, A.C. van Veen, J.A. Lercher, Reversibility
of the modification of HZSM-5 with phosphate anions, J. Phys. Chem. C 118 (2014)
6122–6131.
[34] R.F. Lobo, Introduction to the Structural Chemistry of Zeolites, Handbook of
Zeolite Science and Technology, CRC Press, 2003, pp. 92–124.
[35] Q. Zhu, J.N. Kondo, R. Ohnuma, Y. Kubota, M. Yamaguchi, T. Tatsumi, The study
of methanol-to-olefin over proton type aluminosilicate CHA zeolites, Microporous
Mesoporous Mater. 112 (2008) 153–161.
[36] S.M. Abubakar, D.M. Marcus, J.C. Lee, J.O. Ehresmann, C.-Y. Chen, P.W. Kletnieks,
D.R. Guenther, M.J. Hayman, M. Pavlova, J.B. Nicholas, J.F. Haw, Structural and

Declaration of competing interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
Acknowledgements
This work was supported by the Research Association of Automotive
Internal Combustion Engines (AICE) project, Japan. M. E. acknowledges
the financial support from a Grant-in-Aid for Scientific Research, Japan
Society for the Promotion of Science (JSPS), JP16H04104 and
JP16H06511. The computations were partially performed at the
Research Center for Computational Science, Okazaki, Japan.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.micromeso.2019.109908.

References
[1] R.M. Barrer, Zeolites and their synthesis, Zeolites 1 (1981) 130–140.
[2] J. Xu, Q. Wang, F. Deng, Metal active sites and their catalytic functions in zeolites:
insights from solid-state NMR spectroscopy, Acc. Chem. Res. 52 (2019)
2179–2189.
[3] M. Dusselier, M.E. Davis, Small-Pore zeolites: synthesis and catalysis, Chem. Rev.
118 (2018) 5265–5329.
[4] I. Yarulina, A.D. Chowdhury, F. Meirer, B.M. Weckhuysen, J. Gascon, Recent trends
and fundamental insights in the methanol-to-hydrocarbons process, Nat. Catal. 1
(2018) 398–411.
[5] N.Y. Chen, W.W. Kaeding, F.G. Dwyer, Para-directed aromatic reactions over
shape-selective molecular sieve zeolite catalysts, J. Am. Chem. Soc. 101 (1979)
6783–6784.

11


P. Zhao et al.

Microporous and Mesoporous Materials 294 (2020) 109908

mechanistic investigation of a phosphate-modified HZSM-5 catalyst for methanol
conversion, Langmuir 22 (2006) 4846–4852.
[37] Y. Chu, X. Gao, X. Zhang, G. Xu, G. Li, A. Zheng, Identifying the effective
phosphorous species over modified P-ZSM-5 zeolite: a theoretical study, Phys.
Chem. Chem. Phys. 20 (2018) 11702–11712.

[38] N. Xue, X. Chen, L. Nie, X. Guo, W. Ding, Y. Chen, M. Gu, Z. Xie, Understanding the
enhancement of catalytic performance for olefin cracking: hydrothermally stable
Acids in P/HZSM-5, J. Catal. 248 (2007) 20–28.

[39] Z. Yan, D. Ma, J. Zhuang, X. Liu, X. Liu, X. Han, X. Bao, F. Chang, L. Xu, Z. Liu, On
the acid-dealumination of USY zeolite: a solid state NMR investigation, J. Mol.
Catal. A Chem. 194 (2003) 153–167.

12