ZEOLITES

Accelerated crystallization of zeolites

via hydroxyl free radicals
Guodong Feng,1* Peng Cheng,1* Wenfu Yan,1 Mercedes Boronat,2 Xu Li,1 Ji-Hu Su,3
Jianyu Wang,1 Yi Li,1 Avelino Corma,2 Ruren Xu,1 Jihong Yu1
In the hydrothermal crystallization of zeolites from basic media, hydroxide ions (OH)
catalyze the depolymerization of the aluminosilicate gel by breaking the Si,AlOSi,Al
bonds and catalyze the polymerization of the aluminosilicate anions around the hydrated
cation species by remaking the Si,AlOSi,Al bonds. We report that hydroxyl free radicals
(OH) are involved in the zeolite crystallization under hydrothermal conditions. The
crystallization processes of zeolitessuch as NaA, NaX, NaZ21, and silicalite-1can
be accelerated with hydroxyl free radicals generated by ultraviolet irradiation or
Fentons reagent.

eolites are microporous crystalline aluminosilicates that are used as ion-exchangers in

the detergent industry, catalysts in the petrochemical and chemical industry, and adsorbents in air separation through pressure
swing adsorption (1, 2). Zeolite materials are typically synthesized in a strong basic medium in
which a high concentration of hydroxide ions
(OH) assists in the mineralization of silicate and
aluminate species in the reactant gels (3). The
crystallization process can be described through
the following steps: (i) polymerization, forming
an amorphous gel via making Si,AlOSi,Al bonds;
(ii) depolymerization, forming soluble aluminosilicates and silicates via breaking Si,AlOSi,Al
bonds; and (iii) repolymerization, remaking Si,
AlOSi,Al bonds around the hydrated cation
speciesthat is, the structure-directing agents
via condensation reaction (4, 5). These steps comprise the nucleation and crystal growth stage of
the crystallization, which are catalyzed by OH
involving multiple equilibria. However, the crystallization mechanism of zeolites is still unclear,
and the understanding of their formation at a
molecular level has not yet been achieved.
Under basic conditions, OH ions increase the
coordination of tetrahedral silicon (Si) atoms to
pentahedral or octahedral, which weaken and
break the SiOSi bonds (6). Theoretical calculations show that the dissociation of the SiOSi
bonds is more favorable when catalyzed by hydroxyl free radicals (OH) (7, 8); OH species are
highly active in organic synthesis (9), polymerization (10), and modification of proteins (11).
Generation of OH in solution can be achieved
by means of, for example, electron pulse radiolysis,
ultraviolet (UV) irradiation, Fenton reactions,

chemical reactions, and high-voltage electrical

discharge (11). We show that hydrothermal zeolite crystallization can be accelerated by introducing OH via physical or chemical methods.
To investigate the effect of OH on zeolite crystallization, we used a modified multiparallel reactor that allowed for UV irradiation and controlled
heating (fig. S1). The hydrothermal syntheses
conducted under UV irradiation are referred to
as UV condition, and control syntheses conducted
in the oven without the UV irradiation are referred to as dark condition. We initially studied
the UV and dark-condition syntheses in the Na2O
Al2O3SiO2H2O system at 298 K. Under UV conditions for 24 hours, the experimental and simulated
x-ray diffraction (XRD) patterns of NaX (SiO2/
0.21Al2O3/9.36Na2O/85H2O) (Fig. 1A), NaZ21 (SiO2/
0.32Al2O3/10.05Na2O/85H2O) (Fig. 1B), and NaA
(SiO2/0.46Al2O3/4.4Na2O/60H2O) (fig. S2A) and
the scanning electron microscopy (SEM) images
of the corresponding products (fig. S3) indicate
that the products were already crystallized. Under
dark conditions for 24 hours, the corresponding

XRD patterns (Fig. 1, C and D, and fig. S2B) show

that the materials were still primarily amorphous.
These results demonstrate that the crystallization
rate of zeolites was accelerated by UV irradiation.
The crystallization process of zeolite NaA with
the starting molar composition of SiO2/0.46Al2O3/
4.4Na2O/60H2O was further investigated at 298 K
under dark and UV conditions with different irradiances that varied the OH concentration (reported as crystallization curves in Fig. 1E). Under
dark conditions, long-range ordering of zeolite
NaA, as confirmed with XRD, began to be observed after 40 hours (fig. S4A). After 60 hours,
highly crystalline zeolite NaA was obtained, as
confirmed with the SEM and high-resolution transmission electron microscopy (HRTEM) images
(fig. S5). In contrast, the XRD patterns (fig. S4, B,
C, and D) show that the long-range ordering of
zeolite NaA was already observed at 32, 20, and
16 hours, and highly crystalline zeolite Na-A was
obtained at 52, 40, and 36 hours for irradiances
of 2.0, 4.0, and 8.0 mW/cm2, respectively, as shown
in the SEM and TEM images (figs. S6 to S8).
Reducing the OH concentration could slow
down the crystallization of zeolites. We further
studied the crystallization behavior of the initial
reaction mixture with a reduced Na2O/SiO2 molar
ratio (3.08 versus 4.4) at 298 K. Under the dark
conditions, the long-range ordering of zeolite Na
A was not observed until 45 hours, and the crystals have been well developed when the reaction
time reached 55 hours (fig. S9A, XRD; fig. S10,
SEM and TEM images), whereas long-range ordering was already observed at 40 hours when the
Na2O/SiO2 ratio was 4.4 (fig. S4A). In contrast,
when the initial reaction mixture was irradiated
with a UV lamp, highly crystalline Na-A was formed
at 40 hours (fig. S9B, XRD; fig. S11, SEM and TEM
images). The results presented above clearly demonstrate that reducing OH decreases the crystallization rate of zeolite, whereas OH can promote
the crystallization at the same time. Consequently,
the crystallization rate of zeolite NaA under UV
with a Na2O/SiO2 ratio of 3.08 is even faster than

State Key Laboratory of Inorganic Synthesis and Preparative

Chemistry, College of Chemistry, Jilin University, 2699
Qianjin Street, Changchun, 130012, China. 2Instituto de
Tecnologia Quimica, Universidad Politecnica de Valencia,
Consejo Superior de Investigaciones Cientificas, Valencia,
46022, Spain. 3Department of Modern Physics, University of
Science and Technology of China, 96 Jinzhai Road, Hefei,
230026, China.
*These authors contributed equally to this work. Corresponding
author. E-mail: jihong@jlu.edu.cn

1188

11 MARCH 2016 VOL 351 ISSUE 6278

Fig. 1. Crystallization processes under UV irradiation and dark conditions at 298 K. (A to D) The
experimental XRD patterns of NaX and NaZ21 synthesized under the UV conditions [(A) and (B), black,
4.0 mW/cm2] and dark conditions [(C) and (D), black] for 24 hours, respectively. The corresponding
simulated XRD patterns are plotted for comparison [(A) and (B), red]. (E) Crystallization curves of zeolite
NaA under dark conditions and under UV conditions with irradiance of 2.0, 4.0, and 8.0 mW/cm2, with a
Na2O/SiO2 ratio of 4.4.

sciencemag.org SCIENCE

Downloaded from on March 27, 2016

R ES E A RC H | R E PO R TS

RE S EAR CH | R E P O R T S

that under dark with a Na2O/SiO2 ratio of 4.4. The

accelerating effect on crystallization upon UV with
the reduced alkalinity is further enhanced as compared with that upon UV with normal alkalinity.

Upon UV irradiation, water can generate OH

(11). We used electron paramagnetic resonance
(EPR) to characterize OH and the derived species formed in the zeolite reaction system. We

Fig. 2. Radicals identification from UV irradiation. (A to C) EPR spectra of the initial reaction mixture
containing DMPO under the UV irradiation for (A) 90 s, (B) 30 s, and (C) 0 s. (D to F) Comparison of the
experimental and the simulated EPR spectra of (D) DMPOOH adduct, (E) DMPOSi adduct, and (F) oxidized
DMPO radicals. (G) EPR spectra of the initial reaction mixture containing the spin-trapping agent of BMPO
after 10 hours of dark incubation. The EPR signals are marked as following: red circles, hydroxyl free
radicals; green rectangles, oxidized DMPO radicals; and blue arrows, silicon-based radicals.

added 5,5-dimethylpyrroline-N-oxide (DMPO) as

the spin-trapping agent of OH into the initial
reaction mixture, and we recorded the EPR signals in situ after the reaction mixture was irradiated for 0, 30, and 90 s. For comparison, we
also characterized the initial reaction mixture
and the pure water under dark conditions. As
anticipated, the EPR signals from OH captured
by the DMPOa 1:2:2:1 quartet pattern, with a
splitting of 1.5 mT characteristic of a DMPO-OH
adductwere observed after irradiation for 90 s
(Fig. 2A) and 30 s (Fig. 2B). The characterized
hyperfine coupling constants (HFCs) were aN =
aHb = 1.5 mT. Because DMPO was also photolyzed, the EPR signals of oxidized DMPO radicals
featured by a three-line spectrum were also observed. Sextet EPR signals (Fig. 2, A and B) with
aN = 1.59 mT and aHb = 4.5 mT were observed.
None of these signals were observed in the spectrum of the initial reaction mixture without the
UV irradiation (0 s) in the presence of DMPO
(Fig. 2C), indicating that the new sextet EPR signals were not from the inorganic impurities of
the initial reaction mixture but from the radicals
generated by the UV irradiation.
Because of the large aHb value (4.5 mT), these
sextet EPR signals cannot be attributed to the
carbon (C) and oxygen (O)centered radicals.
For the C-centered radical, the aHrb is at the range
of 1.8 to 2.8 mT, and for the O-centered radical,
the aHb is at the range of 0.7 to 1.4 mT (12, 13). In
fact, the large HFCs of these sextet EPR signals are
similar to that of the DMPO-P(O)(OC2H5)2 adduct (12). Because no phosphorous (P) species was
involved in the initial reaction mixture, these signals might be attributed to the DMPO-Si adduct.
If this is so, then the slight change in the intensity

Fig. 3. Acceleration processes by Fentons reagent. (A to C) Crystallization processes of silicalite-1 at 343 K under (A) Fenton conditions, (B) UV conditions
(4.0 mW/cm2), and (C) dark conditions. (D to F) EPR spectra of the TPAOHTEOSH2O system under (D) Fenton conditions, (E) UV conditions (4.0 mW/cm2),
and (F) dark conditions.The EPR signals are marked as following: red circles, OH; green rectangles, oxidized DMPO radicals; blue arrows, silicon-based radicals;
asterisks, ethanol radicals; and red rectangles, oxidized BMPO radicals.

SCIENCE sciencemag.org

11 MARCH 2016 VOL 351 ISSUE 6278

1189

R ES E A RC H | R E PO R TS

of these signals in the spectra after 30 and 90 s of

irradiation indicated that these Si-based radicals
were generated during the zeolite crystallization.
Previous studies show that fumed silica has an
intrinsic population of planar three-memberedrings (3MRs) formed at high temperature and
frozen-in by rapid quenching, which can undergo cleavage forming Si and SiO radicals because
of the strain in the SiOSi bonds (1416). Aqueous
alkaline silicate solutions also contain 3MRs,
as confirmed by previous 29Si29Si correlation

spectroscopy nuclear magnetic resonance studies

(17). Thus, the Si-based radicals in zeolite synthesis probably form through homolytic cleavage
of the strained SiOSi bonds in planar 3MRs in
the initial reaction mixture. The reason why no
SiO radicals (another type of oxygen-centered
radical) were observed might be caused by their
high activity; they can easily react with water to
form OH. Shown in Fig. 2, D to F, is a comparison of the experimental and the simulated
EPR spectra of DMPOOH adduct, DMPOSi

adduct, and oxidized DMPO radicals, respectively,

which confirms the assignment.
Considering that the concentration of radicals
in the nonUV-irradiated initial reaction mixture and in the pure water might be too low to be
detected by DMPO, we used a recently developed
spin-trapping agent, 5-tert-butoxycarbonyl-5-methyl1-pyrroline-N-oxide (BMPO), that allows a long-time
accumulation of the EPR signals (18). We also observed OH in the initial synthesis mixture under
dark conditions for 10 hours (Fig. 2G). We recorded

Fig. 4. Reaction of [SiO2(OH)OSiO3]Na5

system and Gibbs free-energy calculation. (A
and B) Reaction of [SiO2(OH)OSiO3]Na5 system
with (A) OH and (B) OH. (C) Calculated Gibbs
free-energy profiles for the reaction of [SiO2(OH)
OSiO3]Na5 model with OH (blue) and OH (red).
Electronic (DE) and Gibbs free energies (DG) are
given in kilocalorie per mole.

1190

11 MARCH 2016 VOL 351 ISSUE 6278

sciencemag.org SCIENCE

RE S EAR CH | R E P O R T S

the EPR signals that arose from OH captured by

BMPO, a characteristic 1:2:2:1 quartet pattern of a
BMPOOH adduct (18). The complicated spectrum is due to the spectral overlapping of two
isomers of BMPO (Fig. 3F, as well). Pure water
under dark conditions did not give any EPR
signals (fig. S12), suggesting that OH came from
the zeolite reaction system. The observation of
OH in the initial mixture under dark conditions
suggests that in addition to OH, OH was also
involved in the zeolite crystallization as mineralizers. Although the reactivity of the OH is much
greater than that of OH, the concentration difference favors the OH mechanism.
The Fentons reagent can also produce OH,
and we investigated the crystallization process
for silicalite-1 in the reaction system tetrapropylammonium hydroxide (TPAOH) tetraethyl orthosilicate (TEOS)H2O with the addition of the
Fentons reagent (9TPAOH/25SiO2/480H2O/
100EtOH/0.015H2O2/0.001FeSO418H2O, 343 K).
The XRD patterns of the solid products (Fig. 3A)
indicate that the long-rang ordering of silicalite1 was already observed when the reaction time
reached 40 hours, which is 10 hours shorter than
that under UV conditions (Fig. 3B) and 20 hours
shorter than that under dark conditions (Fig. 3C).
The greater acceleration effect of the Fenton conditions than UV conditions indicates the higher
concentration of OH generated by the Fentons
reagent than by the UV irradiation. The SEM and
TEM images of silicalite-1 recovered at the end of
crystallization under Fenton conditions as compared with the dark and UV conditions are shown
in fig. S13. The yields of silicalite-1 at different
crystallization periods under Fenton and UV
conditions, which are also remarkably improved
as compared with the dark conditions, are given
in fig. S14. The composition and N2 adsorption
data for all samples are included in table S1.
To identify the radicals in the crystallization of
silicalite-1, we characterized the initial reaction
mixtures at ambient temperature under Fenton,
UV, and dark conditions by means of EPR spectroscopy (Fig. 3, D to F). Under dark conditions,
signals from the BMPOCH(CH3)OH adduct
(aN = 1.51 mT, aHb = 2.1 mT) (Fig. 3F) imply the
existence of alkane radicals, which may be formed
from the reaction of OH with the ethanol generated from the hydrolysis of TEOS. This reaction would consume the very limited OH formed
under the dark conditions. Under UV conditions
(Fig. 3E), DMPOCH(CH3)OH (aN = 1.59 mT,
aHb = 2.27 mT), DMPOOH (aN = aHb = 1.50 mT),
DMPOSi (aN = 1.59 mT, aHb = 4.5 mT), and oxidized DMPO radicals were observed and indicate
that more OH formed. The initial reaction mixture under the Fenton conditions (Fig. 3D) contained OH (aN = aHb = 1.50 mT), ethanol carbon
radicals (aN = 1.59 mT, aHb = 2.27 mT), Si-based
radicals (aN = 1.59 mT, aHb = 4.5 mT), and oxidized DMPO radicals.
Inhibiting OH may slow down the zeolite crystallization. Because ethanol is an effective OH
scavenger, we crystallized silicalite-1 at 343 K
under UV conditions by evaporating the ethanol
in the mixture. We prepared two batches of the
SCIENCE sciencemag.org

initial mixtures with the same molar composition; the batches were stirred overnight at ambient temperature under dark conditions in sealed
and open quartz tubes, respectively. The weight
loss of the initial mixture in the open quartz tube
suggests that most of the ethanol hydrolyzed from
TEOS was evaporated from the mixture. Subsequently, the quartz tubes were surely sealed and
heated at 343 K for 48 hours under UV irradiation.
As anticipated, the crystallization process was accelerated when the ethanol was removed, as indicated by the corresponding XRD patterns (fig. S15).
The zeolite crystallization is a two-step process, including nucleation and crystal growth. The
sigmoidal crystallization curves shown in Fig. 1E
show that the induction period under UV conditions is much shorter than that under dark conditions, which suggests that OH generated by
UV irradiation plays an important role in accelerating the nucleation stage during the crystallization. However, the slope of the crystallization
curve for the UV conditions is similar to that for
the dark conditions, implying similar crystal growth
rates. To further confirm which stage was accelerated by OH, we investigated the synthesis of
zeolite NaA at 298 K by means of UV pretreatment in the induction period for 18 hours, followed by crystallization under dark conditions. The
XRD patterns of the products (fig. S16) show that
the crystallization of the synthetic system under
such conditions is similar to that under UV conditions, indicating that OH mostly influences the
nucleation stage but not the crystal growth stage.
Theoretical calculations provide more insight
into how OH accelerates the nucleation stage.
Under the very basic conditions used in the synthesis, anionic [Si2O1+x(OH)6x]x species stabilized by Na+ cations will be preferentially formed
via successive and clearly exothermic deprotonation steps (fig. S17). We first considered the depolymerization of the gel by breaking of the SiOSi
bonds under very basic conditions and studied
the attack of either OH or OH to a highly deprotonated [SiO2(OH)OSiO3]Na5 model of the
gel. In the first case (Fig. 4A), water reacts with
the dimeric silicon species, forming a pentacoordinated intermediate I1, this being the most
energy-demanding step (Fig. 4C, blue line). A H
transfer step converting intermediate I1 into I2 is
necessary to have one OH group on each Si atom,
which facilitates the rapid dissociation of the SiO
bond, generating the monomeric species P. Similar structures are involved in the reaction of the
[SiO2(OH)OSiO3]Na5 model with a OH radical
(Fig. 4B). However, all the activation energies obtained in the reaction with OH (Fig. 4C, red line)
are much lower, indicating that the presence of
OH in the reaction medium will considerably increase the rate of SiO bond-breaking and depolymerization of the gel.
The second step in the nucleation process is
the formation of new SiOSi bonds, which could
be considered to be the reverse of the process
investigated above, starting from products P and
ending with reactants R. The high stability of the
products when using the [SiO2(OH)OSiO3]Na5
model of the gel results in very high activation

energies: 30 kcal/mol in the presence of only

OH, and somewhat lower23 kcal/molwhen
OH is also present in the media.
In order to check the influence of decreased
alkalinity on the mechanism, a monodeprotonated
[Si(OH)2OSi(OH)3]Na model was also considered.
The theoretical results clearly prove the enhanced
positive effect of OH as compared with OH in
breaking of the SiOSi bonds (activation barrier,
29 versus 4 kcal/mol) and promoting the formation of new SiOSi bonds (activation barrier, 17
versus 8 kcal/mol) in the gel with reduced alkalinity (figs. S18 and S19). This is consistent with
our experimental results of the enhanced accelerating effect upon UV with reduced alkalinity.
The discovery that zeolite synthesis mechanism
can be promoted through free radicals sheds a
new light on zeolite crystallization and will open
new perspectives for the synthesis of zeolite materials that are largely demanded in the chemical
industry.
REFERENCES AND NOTES

1. Y. Li, J. Yu, Chem. Rev. 114, 72687316 (2014).

2. J. ejka, A. Corma, S. Zones, Eds., Zeolites and CatalysisSynthesis, Reactions and Applications (Wiley-VCH Verlag GmbH
& Co. KGaA, 2010).
3. C. S. Cundy, P. A. Cox, Chem. Rev. 103, 663702 (2003).
4. C. S. Cundy, P. A. Cox, Micropor. Mesopor. Mater. 82, 178 (2005).
5. A. I. Lupulescu, J. D. Rimer, Science 344, 729732 (2014).
6. R. K. Iler, The Chemistry of Silica: Solubility, Polymerization,
Colloid and Surface Properties, and Biochemistry (John Wiley &
Sons, 1979).
7. Y. T. Xiao, A. C. Lasaga, Geochim. Cosmochim. Acta 60,
22832295 (1996).
8. R. Konecny, J. Phys. Chem. B 105, 62216226 (2001).
9. S. Z. Zard, Radical Reactions in Organic Synthesis (Oxford Univ.
Press, 2003).
10. G. Moad, D. H. Solomon, The Chemistry of Radical
Polymerization (Elsevier, 2006).
11. G. Xu, M. R. Chance, Chem. Rev. 107, 35143543 (2007).
12. G. R. Buettner, Free Radic. Biol. Med. 3, 259303 (1987).
13. A. Angelo, M. Dante, in Electron Paramagnetic Resonance:
A Practitioners Toolkit, M. Brustolon, E. Giamello, Eds.
(John Wiley & Sons, 2008), vol. 4, pp. 285324.
14. D. L. Griscom, C. J. Brinker, C. S. Ashley, J. Non-Cryst. Solids
92, 295301 (1987).
15. H. Zhang et al., J. Am. Chem. Soc. 134, 1579015804 (2012).
16. W. L. Warren, P. M. Lenahan, C. Jeffrey Brinker, J. Non-Cryst.
Solids 136, 151162 (1991).
17. C. T. G. Knight, R. J. Balec, S. D. Kinrade, Angew. Chem. Int. Ed.
Engl. 46, 81488152 (2007).
18. C. H. Huang et al., Chem. Commun. (Camb.) 49, 64366438
(2013).
AC KNOWLED GME NTS

This work was supported by the 973 Project (grants

2014CB931802 and 2013CB921802) and the National Natural
Science Foundation of China (grants 21320102001, 91122029, and
21571075). A.C. thanks the Program Severo Ochoa for financial
support and ERC-AdG-2014-671093SynCatMatch. J.Y. designed
and supervised the project; W.Y., A.C., and R.X. involved the
design of the experiments; G.F., P.C., and J.W. performed the
experiments; J.-H.S. performed the EPR analyses; M.B., X.L., and
Y.L. contributed to the calculations; J.Y. and W.Y. analyzed the
data; G.F. wrote the first draft; and J.Y. and W.Y. deeply revised the
manuscript. A Chinese patent about the method for UV-assisted
synthesis of zeolite materials has been applied for.
SUPPLEMENTARY MATERIALS

www.sciencemag.org/content/351/6278/1188/suppl/DC1
Materials and Methods
Figs. S1 to S19
Table S1
References (1922)
24 December 2015; accepted 28 January 2016
10.1126/science.aaf1559

11 MARCH 2016 VOL 351 ISSUE 6278

1191

Accelerated crystallization of zeolites via hydroxyl free radicals

Guodong Feng et al.
Science 351, 1188 (2016);
DOI: 10.1126/science.aaf1559

This copy is for your personal, non-commercial use only.

If you wish to distribute this article to others, you can order high-quality copies for your
colleagues, clients, or customers by clicking here.

The following resources related to this article are available online at

www.sciencemag.org (this information is current as of March 27, 2016 ):
Updated information and services, including high-resolution figures, can be found in the online
version of this article at:
/content/351/6278/1188.full.html
Supporting Online Material can be found at:
/content/suppl/2016/03/09/351.6278.1188.DC1.html
This article cites 16 articles, 1 of which can be accessed free:
/content/351/6278/1188.full.html#ref-list-1
This article appears in the following subject collections:
Chemistry
/cgi/collection/chemistry

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the
American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright
2016 by the American Association for the Advancement of Science; all rights reserved. The title Science is a
registered trademark of AAAS.

Downloaded from on March 27, 2016

Permission to republish or repurpose articles or portions of articles can be obtained by

following the guidelines here.