The calcined nanotubes (fig. S7B) show peaks
at–102 ppm (Q^3 ) and–110 ppm (Q^4 ), with a Q^3
fraction of 0.17. This value is similar to the as-
made nanotubes and indicates no significant
condensation of surface silanols after calcina-
tion. Normalized FTIR spectra for the nano-
tubes and a beta zeolite of similar Si/Al ratio
(fig. S9) show a similar nature of silanol peaks
in both materials, with higher silanol peak in-
tensity in the nanotubes. Isolated (3745 cm−^1 ),
terminal (3710 cm−^1 ), internal (3670 cm−^1 ), Al-
bridged (3610 cm−^1 ), and H-bonded (3520 cm−^1 )
silanols ( 38 , 39 ) are recognizable, but the pres-
ence of multiple broad O–H stretch bands at
lower wave numbers precludes the identifica-
tion of any other peaks ( 40 ).
The acid site densities were estimated by
temperature-programmed desorption of am-
monia (NH 3 -TPD) and FTIR measurements of
pyridine adsorption on the nanotubes in their
proton-exchanged form. Figure S10A shows
NH 3 -TPD traces and acid site concentrations
for the proton-exchanged nanotubes and, for
comparison, a conventional MFI (ZSM-5) cata-
lyst with a similar Si/Al ratio (20). The strength
and density of weak acid sites for the two ma-
terials are quite comparable, but ZSM-5 has
more strong acid sites. The ZSM-5 catalyst
would have a theoretical acid site density of
794 mmol/g, close to the measured sum of weak
andstrongacidsites(721mmol/g) shown in fig.
S10A. The nanotubes have a measured total
acid site density of 479mmol/g (the theoretical
value is 988mmol/g based on the Al content).
The sum of Brønsted (B) and Lewis (L) acid
site densities measured by pyridine adsorption
is 151mmol/g (fig. S10B), much lower than
the NH 3 -TPD result. The nanotube material
thus has an acid site accessibility factor (AF =
pyridine acid site density/NH 3 acid site density)
of 0.31 and a B/L site ratio of 0.93. Although
(^27) Al NMR provides no evidence for extra-
framework Al (fig. S8) that is sometimes hy-
pothesized to provide Lewis acid sites, it has
been shown ( 41 ) that there is no correlation
between such Al sites and Lewis acid site den-
sities in zeolites. The moderate AF may relate
to the high aggregation tendency of the nano-
tubes, making a considerable fraction of acid
sites inaccessible to the larger pyridine mole-
cules. Discrepancies between NH 3 -TPD and
pyridine infrared are well known in zeolites
and other materials ( 42 , 43 ), but detailed anal-
ysis of the nanotube acid site behavior should
be performed in the future.
The crystal structures of most polycrystal-
line 3D periodic zeolites have been determined
either from 3D electron diffraction or PXRD.
For materials lacking 3D periodicity, such as
2D zeolites, high-resolution TEM imaging has
been key to structure elucidation. These tech-
niques are particularly challenging in the pre-
sent context because of the reduction of the
zeolite to a hollow cylindrical 1D form with a
very thin (~1-nm) wall. A region of the sample
consisting of one or two aligned nanotubes
was used for selected area electron diffraction
(SAED) and 3D continuous rotation electron
diffraction (cRED) (Fig. 2, D and E). The pat-
terns show characteristic features similar to
those of carbon nanotubes ( 44 ) and imogolite
nanotubes ( 45 ). Both the SAED pattern and
reconstructed 3D reciprocal space based on
cRED data in Fig. 2, D and E, reveal a distinct
periodicity of 12.5 Å along the nanotube di-
rection (denotedc), with no apparent perio-
dicity observed perpendicular to thecaxis.
High-resolution annular dark-field scanning
TEM (ADF-STEM) and integrated differential
phase contrast (iDPC) images were obtained
both perpendicular to and along the nanotube
direction after sectioning the nanotubes by
ultramicrotomy (Fig. 3). Images acquired
along the nanotube direction of individual
and fused nanotubes (Fig. 3, A and B, and fig.
S11) confirm a tubular structure with an ~5-nm
outer diameter and an ~3-nm inner diameter.
Ten identical repeating units with square-like
features are frequently observed around the
circumference of the nanotubes, and the dis-
tance between adjacent units is 12 to 13 Å.
Occasionally, nanodomains with micropores
of ~6-Å diameter and an arrangement resem-
bling 3D zeolite beta (*BEA) ( 27 ) are observed
(figs. S12 and S13, A to E). The square-like
feature is found in both the nanotubes and
the zeolite beta-like domains. A structural
model of the circumferential building unit of
the nanotube (fig. S13E) could be deduced
from the image of an incomplete nanotube
(fig. S13, A to C) and the beta structure. Images
acquired perpendicular to the nanotube di-
rection reveal the projected wall structure in
more detail (Fig. 3, C and D). The Fourier trans-
form of the image (inset in Fig. 3C) confirms
the periodicity of ~12.5 Å along the nanotube
direction and a lack of long-range periodic-
ity perpendicular to the nanotube direction,
consistent with the electron diffraction data.
Isolated dark features of ~6-Å diameter are
observed on the nanotube wall surface (Fig.
3D), revealing the presence of micropores on
the nanotube wall. The pore size range cor-
responds to 10MRs to 12MRs perforating the
wall. The micropores are arranged at an oblique
angle of ~108° (fig. S14) with respect to the
nanotube channel axis at a distance of 12 Å,
similar to that in zeolite beta.
On the basis of the iDPC STEM images and
the axial periodicity from cRED, the structural
model of the nanotube is deduced (Fig. 4). The
circumferential building unit (Fig. 3 and fig.
S13) is depicted in fig. S15A. Repetition of 10
such building units leads to the circumfer-
ential cross-section of the nanotube (Fig. 3A
and fig. S15B). In the nanotube circumference,
these building units are connected through a
5MR (figs. S13 and S15) rather than through a
6MR as in zeolite beta. Although the connec-
tion through a 6MR in zeolite beta retains the
orientation of the building units, the connec-
tion through a 5MR in the nanotube enforces
a ~36° rotation of the building unit relative to
its neighbors (fig. S15, B and C). This leads to
closure of the cylindrical sheet (nanotube)
with 10 building units (fig. S15B). The terminal
T sites in the walls can act as branching points
to form fused nanotubes. Branching occurs
between two circumferential building units,
as observed in the ADF and iDPC-STEM im-
ages (Fig. 3B and fig. S16). The Q^3 fraction of
T atoms in the structural model is 0.23, which
SCIENCEscience.org 7 JANUARY 2022•VOL 375 ISSUE 6576 63
Fig. 2. Zeolite nanotube morphology and diffraction patterns.(A) TEM image showing the tubular
morphology of the calcined zeolitic nanotube material. (BandC) PXRD patterns of the calcined nanotube
material showing low-angle (B) and wide-angle (C) regions. (D) Selected area electron diffraction pattern
from a nanotube (marked in the inset) showing typical tubular features with periodicity along the nanotube
direction and characteristic diffraction streaks perpendicular to the nanotube direction. (E) Reconstructed
3D reciprocal lattice from cRED data collected from a bundle of zeolitic nanotubes with the nanotube
directionc*marked.
RESEARCH | REPORTS