Surface and interlayer base-characters in lepidocrocite titanate: The adsorption and intercalation of fatty acid

Article abstract of DOI:10.1016/j.jssc.2016.03.030

While layered double hydroxides (LDHs) with positively-charged sheets are well known as basic materials, layered metal oxides having negatively-charged sheets are not generally recognized so. In this article, the surface and interlayer base-characters of O^2- sites in layered metal oxides have been demonstrated, taking lepidocrocite titanate K_0.8Zn_0.4Ti_1.6O₄ as an example. The low basicity (0.04 mmol CO₂/g) and low desorption temperature (50-300 °C) shown by CO₂- TPD suggests that O^2- sites at the external surfaces is weakly basic, while those at the interlayer space are mostly inaccessible to CO₂. The liquid-phase adsorption study, however, revealed the uptake as much as 37% by mass of the bulky palmitic acid (C₁₆ acid). The accompanying expansion of the interlayer space by ~0.1 nm was detected by PXRD and TEM. In an opposite manner to the external surfaces, the interlayer O^2- sites can deprotonate palmitic acid, forming the salt (i.e., potassium palmitate) occluded between the sheets. Two types of basic sites are proposed based on ultrafast ¹H MAS NMR and FTIR results. The interlayer basic sites in lepidocrocite titanate leads to an application of this material as a selective and stable two-dimensional (2D) basic catalyst, as demonstrated by the ketonization of palmitic acid into palmitone (C₃₁ ketone). Tuning of the catalytic activity by varying the type of metal (Zn, Mg, and Li) substituting at Ti^IV sites was also illustrated.

Full text of DOI:10.1016/j.jssc.2016.03.030

Author’s Accepted Manuscript
Surface and interlayer base-characters in
lepidocrocite titanate: The adsorption and
intercalation of fatty acid
Tosapol Maluangnont, Pornanan Arsa, Kanokporn
Limsakul, Songsit Juntarachairot, Saithong
Sangsan, Kazuma Gotoh, Tawan Sooknoi
www.elsevier.com/locate/yjssc
PII:
S0022-4596(16)30103-7
http://dx.doi.org/10.1016/j.jssc.2016.03.030
YJSSC19326
DOI:
Reference:
To appear in:
Journal of Solid State Chemistry
Received date: 31 January 2016
Revised date: 5 March 2016
Accepted date: 18 March 2016
Cite this article as: Tosapol Maluangnont, Pornanan Arsa, Kanokporn Limsakul
Songsit Juntarachairot, Saithong Sangsan, Kazuma Gotoh and Tawan Sooknoi
Surface and interlayer base-characters in lepidocrocite titanate: The adsorption
and intercalation of fatty acid, Journal of Solid State Chemistry
http://dx.doi.org/10.1016/j.jssc.2016.03.030
This is a PDF file of an unedited manuscript that has been accepted fo
publication. As a service to our customers we are providing this early version o
the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting galley proof before it is published in its final citable form
Please note that during the production process errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal pertain

Surface and interlayer base-characters in lepidocrocite titanate:
The adsorption and intercalation of fatty acid
Tosapol Maluangnont,*,†,‡ Pornanan Arsa,‡,§ Kanokporn Limsakul,§ Songsit Juntarachairot,§
Saithong Sangsan,§ Kazuma Gotoh,∥ and Tawan Sooknoi*,‡,§
†College of Nanotechnology, King Mongkut’s Institute of Technology Ladkrabang, Bangkok
10520, Thailand
‡Catalytic Chemistry Research Unit, Faculty of Science, King Mongkut’s Institute of
Technology Ladkrabang, Bangkok 10520, Thailand
§Department of Chemistry, Faculty of Science, King Mongkut’s Institute of Technology
Ladkrabang, Bangkok 10520, Thailand
∥Graduate School of Natural Science & Technology, Okayama University, 3-1-1 Tsushima-naka,
Okayama 700-8530, Japan
*Corresponding author
Email: tosapol.ma@kmitl.ac.th (T.M.); Tel: +66 9 0992 2076; Fax: +66 2 329 8265
Email: kstawan@gmail.com (T.S.); Tel.: + 66 8 1929 8288; Fax: + 66 2 326 4415
1

Abstract
While layered double hydroxides (LDHs) with positively-charged sheets are well known
as basic materials, layered metal oxides having negatively-charged sheets are not generally
recognized so. In this article, the surface and interlayer base-characters of O2- sites in layered
metal oxides have been demonstrated, taking lepidocrocite titanate K0.8Zn0.4Ti1.6O4 as an
example. The low basicity (0.04 mmol CO2/g) and low desorption temperature (50-300oC)
shown by CO2-TPD suggests that O2- sites at the external surfaces is weakly basic, while those at
the interlayer space are mostly inaccessible to CO2. The liquid-phase adsorption study, however,
revealed the uptake as much as 37% by mass of the bulky palmitic acid (C16 acid). The
accompanying expansion of the interlayer space by ~0.1 nm was detected by PXRD and TEM.
In an opposite manner to the external surfaces, the interlayer O2- sites can deprotonate palmitic
acid, forming the salt (i.e., potassium palmitate) occluded between the sheets. Two types of basic
1
sites are proposed based on ultrafast H MAS NMR and FTIR results. The interlayer basic sites
in lepidocrocite titanate leads to an application of this material as a selective and stable two-
dimensional (2D) basic catalyst, as demonstrated by the ketonization of palmitic acid into
palmitone (C31 ketone). Tuning of the catalytic activity by varying the type of metal (Zn, Mg,
and Li) substituting at TiIV sites was also illustrated.
Keyword:
Intercalation, layered materials, alkali titanate, basic sites, ketonization, fatty acid
2

Introduction
Solid base catalysts are capable of activating acidic reactants1,2 in several reactions of
industrial importance, or in the synthesis of fine chemicals3,4 and first-generation biodiesel.5 The
base-characters of metal oxides originate from under-coordinated O2- sites on different locations
(step, corner, terrace, etc.) of the crystals. These sites could behave as the Lewis base by
donating an electron pair to reactants. The formation of surface carbonate (or bicarbonate) on
oxides in contact to atmospheric CO2 is an example of the Lewis base-character of oxides.
Alternatively, the O2- sites act as the Brønsted base if they abstract a proton from reactants. This
is the case for the formation of the surface hydroxyl group on oxides upon contacting with
atmospheric water.
Common oxides such as CaO, MgO, ZrO2 are strongly basic, inexpensive, but typically
have low surface area and random distribution of the atomic planes.6 Their catalytic activity
greatly depends on the exposed facets of the crystals which have to be controlled through a
delicate synthetic procedure. In contrast, a family of solids known as layered metal
7 ,
8
oxides/hydroxides
naturally crystalizes into two-dimensional (2D) plate-like objects,
composed of atomically-thin sheets as elementary units. Such crystals inherently expose the
faces (i.e., the basal plane) and the edges to the environment, resulting in the external active
sites. In addition, the 2D crystals possess the internal interlayer space sandwiched between the
sheets. These sites can be utilized only if they are accessible to reactants. However, the
accessibility to internal basic sites and the thermal/chemical stability do not necessarily coexist
in the same catalyst. For example, well-known 2D metal hydroxides such as layered double
hydroxides (LDHs) or layered hydroxy salts (LHSs) are considered basic largely because they
2-
exhibit high affinity toward the acidic molecule CO2 (in the form of intercalated carbonate CO3
), which can have the electrostatic attraction with the positively-charged sheets. However, these
3

materials transform into (mixed) metal oxides upon contact with reactants at elevated reaction
temperature,8,9,10 thereby losing the two-dimensionality and the associated benefits. Therefore,
there is still a need to develop a new type of active and stable 2D solid base catalysts whose the
external and the internal surfaces could be effectively utilized.
Lepidocrocite titanates (AxMyTi1-yO4 , where A = K, M = Zn, Mg, Ni, Cu, FeIII, MnIII;11,12
A = Cs, M = Zn,13 Ni,14 Mg,15,16 Nb;17 A = K, Rb, and Cs, M = Li;18,19 A = Cs, M = cationic
vacancy20) show great potential as thermally stable, 2D basic catalysts. The basal plane (either
external or internal) is on the ac plane of an orthorhombic unit cell. Two types of bridging basic
oxygen atoms can be found. The first one is O2c coordinating to two TiIV atoms nearby, while the
second one is O4c and coordinates to four TiIV atoms. It is natural to assume that the negative
charges on the sheets would localize at O2c and O4c, contributing to the base characters.
Moreover, the edges of crystals also contain under-coordinated oxygen atoms of which the exact
geometry depends on the extent of hydroxylation. Therefore, the nature of basic sites (O2-) in
layered metal oxides is fundamentally different from that in LDHs/LHSs mentioned above. Still,
with the exception of a simple ion exchange of interlayer alkali cations with aqueous mineral
acids19,21,22,23, 24,25 , reports on base characters of layered metal oxides and their behavior toward
other acidic probe molecules are lacking.
In this work, the surface basicity of lepidocrocite titanate K0.8Zn0.4Ti1.6O4 has been
characterized by CO2 temperature programmed desorption (CO2 TPD). In combination with
1
FTIR and H ultrafast MAS NMR, surface hydroxyl can be revealed. Results were compared to
TiO2 and the well-known basic catalyst such as MgO. Bulky acidic probe molecule, palmitic acid
(C16 acid), was employed to test the accessibility to the interlayer basic sites via liquid phase
adsorption (intercalation). We further demonstrated the application of some lepidocrocite
titanates (K0.8Zn0.4Ti1.6O4, K0.8Mg0.4Ti1.6O4 and K0.8Li0.27Ti1.73O4) as catalysts for the gas-phase
4

ketonization of palmitic acid, in comparison to the mixed Mg/Al oxides from the decomposition
of LDH. The ketonization of carboxylic acids to the respective ketones is of industrial
importance as a processing step in biomass conversion to hydrocarbon fuels.26, 27, 28, 29
Material and Methods
Synthesis. Lepidocrocite titanate K0.8Zn0.4Ti1.6O4 was synthesized11 by the calcination of
the stoichiometric mixture of K2CO3, ZnO, and TiO2 at 900ºC for 20 h. Two other compositions,
K0.8Mg0.4Ti1.6O4 and K0.8Li0.27Ti1.73O4, can be prepared similarly but replacing ZnO with MgO11
and Li2CO3,18,19 respectively. The layered double hydroxide with Mg/Al mole ratio of 2.5 was
synthesized by a coprecipitation method30 and was calcined at 450 ºC for 6 h, giving the mixed
Mg/Al oxide “MgAl2.5”. Potassium palmitate was obtained as the white powder from the
reaction between KOH(aq) and palmitic acid. The PXRD patterns of other materials in addition
to K0.8Zn0.4Ti1.6O4 can be found in Figure S1-S2 in Supporting Information. All chemicals are of
reagent grade and were used as received, except K2CO3, Li2CO3, KOH and palmitic acid which
were dried 120ºC overnight prior to use.
CO2 temperature programmed desorption. Approximately 200-300 mg of the sample was
activated in situ at 450ºC for 2 hours under air zero in a quartz tube reactor, followed by a
cooling to room temperature under N2 gas. Gaseous CO2 was then flown through the sample at
the flow rate 30 mL/min for an hour, followed by a flow of helium at the same flow rate for
another hour. Next, the temperature was raised from room temperature to 600ºC at 5ºC/min using
pure helium as a carrier gas. The CO2 desorption profile was recorded by a thermal conductivity
detector and was normalized by mass of the sample employed. The basicity was expressed as
mmol of CO2 per g of the sample. The peak area obtained was compared to that from a known
5

amount of CaCO3 which decomposes completely and quantitatively to CaO and CO2. All gases
are the commercial products of Praxair with the purity of 99.99% and were used as received.
Adsorption of palmitic acid. An amount of 2.0 g of K0.8Zn0.4Ti1.6O4 was heated with 150
mL of a 5%w/w palmitic acid (Fluka, 97%+) in isopropanol at 60ºC from 1 to 36 h. The liquid
was withdrawn at certain intervals, and the content of adsorbed palmitic acid was determined
based on equation (1), together with an appropriate conversion of concentration to mass. Here,
A0 is the corrected peak area of palmitic before the adsorption, and At is the corrected peak area
of palmitic acid remaining in the mother liquor after the adsorption time t.
%Adsorption = [(A0-At)/A0]×100%
(1)
The solid was filtered, washed with isopropanol, and then dried at 70ºC overnight. The
experiment with decanoic acid was performed similarly. A control experiment was performed by
heating K0.8Zn0.4Ti1.6O4 with isopropanol at 60 ˚C in the absence of any carboxylic acid. Another
control experiment was also performed, where K0.8Zn0.4Ti1.6O4 was magnetically stirred with 1
M HCl (3 times, 12 h each, with the fresh HCl loaded between repetition)24 so as to obtain the
proton-containing lepidocrocite where K+ is ion exchanged with H+·H2O.
Characterization. Powder X-ray diffraction (PXRD) measurements were conducted on a
Rigaku DMAX 2200/Ultima+ diffractometer, employing a Cu-Kα radiation at 40 kV and 30 mA.
The sample was scanned in the range 2 = 5-65º at a 0.05 degree/step and a detection time of 0.6
second/step. BET specific surface area of the sample was determined by a Gas Adsorption
Analyzer (Autosorb-1C, Quantachrome). For TGA (Perkin-Elmer, Pyris 1), the sample was
heated from room temperature to 900ºC (10ºC/min) under nitrogen gas flowing at 20 mL/min.
FTIR spectra of the pelletized samples (prepared by grinding with KBr) were collected on a
6

Spectrum GX (Perkin Elmer) spectrometer. SEM images were acquired on a Zeiss EVO/MA10
scanning electron microscope, while TEM images were acquired on a JEOL JEM 2010
1
transmission electron microscope. H MAS NMR (spinning speed: 100 kHz) of K0.8Zn0.4Ti1.6O4
was recorded using a JEOL ECA-600 spectrometer with a 0.75 mm MAS sample rotor. Depth2
pulse sequence with 1.3 s of 90 pulse length and 5.0 s of repetition time were applied to obtain
a spectrum.
Catalytic activity testing. Catalytic activity testing was conducted in a continuous fixed
bed down-flow reactor made of glass (length, 50 cm; outer diameter, 8 mm; inner diameter, 6
mm) under atmospheric pressure of N2. The catalyst was activated to 800oC in air zero for an
hour, followed by a cooling to the reaction temperature (375oC) under air zero. Then, 5%
palmitic acid in p-xylene was fed into the reactor by an HPLC pump at a flow rate of 0.025
mL/min. The catalyst activity testing was conducted at atmospheric pressure, at the contact time
(W/F) of 1500 g·h/mol, and the flow rate of palmitic acid plus carrier gas (N2) of 30 mL/min, for
a total of 360 min. The liquid products were trapped by an ice bath and collected every 45
minutes before the analysis by an HP6890 gas chromatograph (Hewlett-Packard). The catalyst
recovered after the activity testing at the mentioned condition is hereafter called “spent”.
7

Results and discussion
Calcination of the stoichiometric mixture of K2CO3, ZnO, and TiO2 at 900ºC for 20 h
results in the lepidocrocite titanate K0.8Zn0.4Ti1.6O4 with high crystallinity, as shown by the
PXRD pattern in Figure 1a. This product will be hereafter called “as made” K0.8Zn0.4Ti1.6O4. The
diffraction pattern can be indexed based on the orthorhombic symmetry, with unit cell
parameters (refined by Cellcalc) of a = 0.3809(3), b = 1.567(1), and c = 0.2981(2) nm, in
agreement with the previous work (suggested space group11 Cmc21). The first peak at 2 =
11.34º (d = 0.78 nm) corresponds to the 020 reflection which is the repeating distance of the
edge-shared TiO6 (or (Ti/Zn)O6) sheets stacking along the b-direction. Figure 2a shows that the
plate-like crystals of the as made K0.8Zn0.4Ti1.6O4 with the lateral size up to 2m × 2 m form
aggregates, likely due to the use of high temperature in the synthesis. K0.8Zn0.4Ti1.6O4 had an
extremely low specific surface area of 3 m2/g.
8

Figure 1.
PXRD patterns of (a) as made K0.8Zn0.4Ti1.6O4, (b) the solid product after the
adsorption of the 5%w/w palmitic acid/isopropanol solution at 60oC for 36 h, (c) potassium
palmitate, (d) the solid product heated with isopropanol only, and (e) similar to (b) but with
decanoic acid.
9

Figure 2.
SEM images of (a) as made, and (b) spent K0.8Zn0.4Ti1.6O4.
The CO2 desorption profiles of K0.8Zn0.4Ti1.6O4 and other tested materials are shown in
Figure 3. CO2 desorbs from K0.8Zn0.4Ti1.6O4 in the range 50-300ºC (with the peak temperature Tp
at 89ºC), which is close to the range reported for mixed sodium/potassium titanate nanotubes.31
A relatively low Tp suggests that the majority of CO2 adsorption occurrs physically.
Lepidocrocite titanate K0.8Zn0.4Ti1.6O4 is therefore a relatively weak base solid, similar to other
reported titanate-based materials.31, 32 On the other hand, CO2 desorption from TiO2 anatase (Tp =
120ºC) and “MgAl2.5”, which is the mixed (Mg/Al) oxide from the decomposition of layered
double hydroxide (Tp = 170ºC), span a wider range (both in the range 50-400ºC). CO2 molecules
adsorbed on these two materials could contain a minor fraction which was adsorbed chemically,
in addition to the major fraction which was adsorbed physically. While MgO shows the small
peak at Tp = 196ºC somewhat similar to K0.8Zn0.4Ti1.6O4, TiO2 and MgAl2.5, a very strong CO2
desorption at 400-700ºC is due to the decomposition of MgCO3 which was produced in situ from
the exposure of MgO to CO2. The strength of the basic sites which are accessible to CO2 can be
estimated from Tp, and is in the order: K0.8Zn0.4Ti1.6O4 < TiO2 < MgAl2.5 < MgO.
10

Figure 3.
CO2 desorption profiles of as made K0.8Zn0.4Ti1.6O4, TiO2 (anatase), MgAl2.5
which is a mixed (Mg/Al) oxide from the decomposition of a layered double hydroxide with
Mg/Al = 2.5, and MgO. The y-axis is the signal from the thermal conductivity detector
normalized by the mass of the sample.
The basicity (mmol CO2/g, in parenthesis) is in the order: K0.8Zn0.4Ti1.6O4 (0.04) < TiO2
(0.07) < MgAl2.5 (0.82) < MgO (1.18). Excluding MgO whose the CO2 desorption has a
contribution from the bulk, these values parallel the surface area (m2/g) of each material which is
in the order K0.8Zn0.4Ti1.6O4 (3) < TiO2 (6) < MgAl2.5 (122). Our results imply the
inaccessibility of the O2- sites at the interlayer region of K0.8Zn0.4Ti1.6O4 with respect to gaseous
CO2. Consequently, the adsorption of CO2 should take place at the sites located on the external
surfaces of the crystals only. It is important to note that CO2-TPD detects the basic sites
accessible by CO2 without discriminating whether they are O2- or OH sites. These basic sites
could react with CO2, forming carbonate (for O2-) or bicarbonate (for OH) in various
configurations.33 Interestingly, the basicity of K0.8Zn0.4Ti1.6O4 (0.04 mmol/g) with a low surface
area of 3 m2/g is just slightly different from some of the values reported for titanate nanotubes
11

(0.03-0.05 mmol/g)31 despite a much larger surface area of the nanotubes (generally more than
100 m2/g).31
The FTIR spectrum of K0.8Zn0.4Ti1.6O4 is shown in Figure 4a. The sharp peak at 811 cm-1
is due to Ti-O lattice vibration.18,19 For comparison, the peak at 899 cm-1 was reported34 and
assigned similarly in a lepidocrocite titanate with interlayer Na+ cations. The band at 3418 and
1624 cm-1 can be ascribed to the stretching and bending vibration of adsorbed H2O18,19
respectively. The hydroxyl group attached to the layered titanates can be found generally at 900-
1000 cm-1.18, 19, 35, 36, 37 It is likely that the sharp peak at 1018 cm-1 in K0.8Zn0.4Ti1.6O4 (Figure 4a)
has similar origin. The relatively high wavenumber of this peak in relation to others18, 19, 35, 36, 37
suggests a strong bonding between the oxygen atom (i.e., O2- sites) and the hydrogen atom
(presumably from adsorbed water molecules). In addition to the formation of a hydroxylated
surface, a small band at 1388 cm-1 due to the asymmetric stretching of monodentate carbonate
(vs 1378 cm-1 as calculated by Mino et al.38) can be observed. Here, the O2- sites are the Lewis
basic donating an electron pair to gaseous CO2 in the atmosphere. Other peaks at ~2300-2375
cm-1 are due to the asymmetric stretching of gaseous CO2. Several bands at 2850-3000 cm-1 are
due to CH stretching of the thin organic films coated on an FTIR window.
12

Figure 4.
FTIR spectrum of (a) as made K0.8Zn0.4Ti1.6O4, (b) the solid product after being
heated with isopropanol only, (c) the solid product after the adsorption of the 5%w/w palmitic
acid/isopropanol solution at 60oC for 36 h, and (d) palmitic acid.
The solid state 1H ultrafast MAS NMR spectrum of as made K0.8Zn0.4Ti1.6O4 is shown in
Figure 5. The peak at 4.1 ppm (FWHM = 0.6 ppm) is assigned to protons of physisorbed
water.39, 40 The presence of water molecules as detected by NMR agrees with results by FTIR
just mentioned. The broad signal at ~9 ppm can be deconvoluted into two peaks at 8.2 and 9.0
ppm (FWHMs of both peaks are 2.0 ppm), suggesting that these two protons are in very similar
yet different chemical environments. These signals can be inferred to water molecules interacting
with two types of O2- sites in as made K0.8Zn0.4Ti1.6O4. Since there are two types of structural
oxygen atoms, each bound to different number of TiIV atoms, two partially overlapped H signals
could be expected. For comparison, the peaks at 7.9 and 10.0 ppm have been reported in
protonated titanate nanotubes, 41 and at 6.0 and 8.0 ppm in H2Ti3O7. 42 However, in these
materials41,42 the signals are due to protons introduced via acid ion exchange, while in our case
13

the presence of protons (or hydroxyl groups) is most likely a result of the affinity between O2-
sites and atmospheric water.
Figure 5.
1H MAS NMR spectrum of as made K0.8Zn0.4Ti1.6O4 (black solid line) and the
fitting spectrum (red broken line). The deconvoluted Lorentzian components are shown by blue
dotted lines at the bottom spectrum.
In order to demonstrate the accessibility of interlayer basic sites, we performed a liquid-
phase adsorption study of a 5% w/w palmitic acid in isopropanol on as made K0.8Zn0.4Ti1.6O4 at
60oC. As shown in Figure 6, the uptake of “palmitic acid” gradually increases to 34.5% (12 h) by
weight, and remains constant after that (34.9%, 36 h). Such a large adsorption solely to the
external surfaces is unlikely considering the extremely low surface area of 3 m2/g. Instead, the
internal surfaces must involve in this uptake. We will show later on that “palmitic acid” is
adsorbed as palmitate species, and that the adsorption involves the interlayer space, i.e.,
intercalation. In agreement with %adsorption determined from the mother liquor, the solid after
contact with palmitic acid at 36 h shows 37.0% mass loss in the range 200-400oC (Figure 7a),
ascribed to the loss of the adsorbed “palmitic acid”. The mass loss curve of palmitic acid is
shown for comparison in Figure 7c. Besides, Figure 7b shows a negligible mass loss from the
14

control experiment in the absence of palmitic acid where only isopropanol was in the liquid
phase.
Figure 6. The %adsorption by mass of palmitic acid by K0.8Zn0.4Ti1.6O4 at 60oC,
determined from the corrected peak area of palmitic acid before the adsorption (A0) and at time t
(At) of the mother liquor.
15

Figure 7. Mass loss curves of (a) the solid product after the adsorption of the 5%w/w palmitic
acid/isopropanol solution to K0.8Zn0.4Ti1.6O4 at 60oC for 36 h, (b) the solid product treated with
isopropanol only, and (c) palmitic acid.
In line with above view, PXRD pattern of the sample after palmitic acid adsorption in
Figure 1b shows an expansion of the interlayer. The PXRD pattern can also be indexed based on
the orthorhombic symmetry, but with a = 0.3763(1) nm, b = 1.7190(7) nm, and c = 0.3005(1)
nm. While the orthorhombic unit cell of pristine K0.8Zn0.4Ti1.6O4 is C based-centered11 (hkl
reflections observed following h+k = 2n), such restriction is no longer hold considering the
presence of reflections such as 121, 141, 161, 211 and 251. Instead, the unit cell of this product
is likely primitive orthorhombic. (See also Figure S3 and Table S1.) Unfortunately, the quality of
the data does not allow us to do Rietveld analysis. Two parameters (a and c) are close to those in
as made K0.8Zn0.4Ti1.6O4, suggesting that the host layer was kept unchanged during the
adsorption. The most obvious feature is the shift of the first peak (i.e., d020) to a lower angle 2 =
10.36o (i.e, larger spacing d = 0.85 nm), resulting from the expansion of ~0.1 nm. While this
16

expansion is rather small compared to typical values (0.3-0.4 nm) reported 43 , 44 for the
intercalation compounds having an alkyl chain, the 0.23 nm-expansion due to the pressure-
induced intercalation45 of ethanol into K0.8Ni0.4Ti1.6O4 has been reported. For comparison, the
unit cell parameters a = 0.378(3) nm, b = 1.79(2) nm, and c = 0.298(4) nm can be deduced from
the PXRD pattern (Figure S4) of the proton-containing solid prepared via repeated ion exchange
with aqueous HCl. The relatively large d spacing (d020 = 0.90 nm at 2 = 9.82o) is due to the
intercalation of H+·H2O in agreement with the literature,46 but is in sharp contrast to the small
d020 in Figure 1b.
In K0.8Zn0.4Ti1.6O4 having the C based-centered unit cell (space group Cmc21),11
potassium cations sit at the 4a site within the trigonal prismatic cavity. Since the two adjacent
sites are too small to be occupied simultaneously, the site occupancy11 of only 40% was
achieved. So, there are rooms available for the subsequent intercalation of the guest species into
the interlayer space. Upon intercalation, we observed the change of the symmetry from C-base
centered to the primitive one. Recently, Shirpour et al.47 have reported a similar change as
K0.8Li0.27Ti1.73O4 intercalates the bulky hydrated sodium cations. This change occurs via the
lateral gliding of the lepidocrocite layers along the a axis by a/2, generating a relatively larger
cubic cavity for the guest species. Based on this cited work47, we propose that the cubic cavity
could locally accommodate palmitic acid (palmitate) while limiting the long-range expansion
along the b-direction. It could be deduced that the strong guest-host interaction (i.e., acid-base
interaction between palmitic acid and O2- sites) overcomes the weaker host-host interactions such
that the gliding of the layers is possible. Other examples18,45,48 of the change of symmetry of the
orthorhombic unit cell in lepidocrocite titanate upon the intercalation of guest species have also
been reported.
17

The PXRD pattern of the solid product heated with isopropanol only (Figure 1d) is
similar to that of the as made K0.8Zn0.4Ti1.6O4, suggesting the essential role of palmitic acid in
such topotactic transformation. The FTIR spectrum of the isopropanol-treated K0.8Zn0.4Ti1.6O4 in
Figure 4b is generally similar to as made K0.8Zn0.4Ti1.6O4, with some variations in the relative
peak intensities. We also exclude the possibility that the expanded product be potassium
palmitate, as evidenced from the notable differences between Figure 1c and Figure 1b.
Additionally, replacing palmitic acid (C15H31COOH) by decanoic acid (C9H19COOH) gave the
product with a similar PXRD pattern (Figure 1e) and unit cell parameters (a = 0.3760(4) nm, b =
1.716(2) nm, and c = 0.3002(3) nm). The similar interlayer spacing in these two cases suggests
the orientation where the long molecular axis of the acids is parallel to inorganic sheets.
Further evidence for the interlayer expansion is the TEM image. It can be seen that as
made K0.8Zn0.4Ti1.6O4 shows a repeating distance of ~0.8-0.9 nm (Figure 8a), while the solid
after the adsorption of palmitic acid exhibits a repeating distance of ~1.0 nm (Figure 8b). These
distances agree reasonably with the d-spacing of the first peak in the respective PXRD patterns.
Figure 8 TEM images of (a) as made K0.8Zn0.4Ti1.6O4, and (b) the solid product after adsorption
of palmitic acid at 60°C for 36 h.
18

The characteristics peaks in the FTIR spectrum of palmitic acid in Figure 4d include the
strong C=O stretching at 1701 cm-1, and the OH stretching at 3420 cm-1. The co-presence of
these two peaks is widely accepted as the signature of a carboxylic acid. Several other peaks
serve as a fingerprint especially at 1000-1250 cm-1. Unfortunately, the peaks due to CH2
symmetric and asymmetric stretching (~2850-2960 cm-1) overlap with the thin organic films
coated on the FTIR window. Interestingly, while the OH stretching is still visible, the C=O
stretching is very weak in K0.8Zn0.4Ti1.6O4 refluxed with palmitic acid (Figure 4c). Instead, the
peaks at 1597 and 1385 cm-1 were detected and assigned to the asymmetric and symmetric
stretching of the carboxylate anion, 49 respectively. These results suggest the existence of
palmitate species. Besides, two other peaks (1620 and 1467 cm-1) at the positions slightly shifted
from the previous ones (1597 and 1385 cm-1) indicate that palmitate species are in two
environments. Several vibrations in the range 1000-1250 cm-1 are maintained, confirming the
integrity of the skeleton of the intercalated palmitate species after the intercalation into
K0.8Zn0.4Ti1.6O4.
Basic sites at several positions in K0.8Zn0.4Ti1.6O4 crystals might involve in the uptake of
palmitic acid. Initially, palmitic acid would be adsorbed onto the external basic sites at the basal
plane or the edges. The adsorption at the latter site might specifically lead to the
rearrangement/activation of palmitic acid into the geometry suitable for the subsequent
intercalation into the interlayer region. The internal O2- sites at the interlayer region act as a
Brønsted base, abstracting the proton from palmitic acid. Meanwhile, the charge-balancing
potassium cations at the interlayer region migrate close to the newly-formed palmitate anion, so
as to preserve charge neutrality of the system as shown by equation (2):
K0.8[Zn0.4Ti1.6O4] + 0.8C15H31COOH  H0.8[Zn0.4Ti1.6O4] + 0.8C15H31COOK
(2)
19

where the species in the bracket represents the immobilized layers of lepidocrocite titanate. This
reaction might be driven by the acid-base interactions between interlayer O2- sites and proton of
palmitic acid, and also by the electrostatic attraction between palmitate anion and K+. Our
observation is somewhat different from that reported 50 by Chou et al., where only the
complexation of the carboxylate with divalent (but not monovalent) cations in smectite clays
drives the formation of the occluded salt. Upon the intercalation, an internal hydroxylated
surface would be formed, together with the palmitate species. The peak at 1023 cm-1 in Figure 4c
could be the O-H stretching of the hydroxyl group attached to the layered titanates, similar to the
one in as made K0.8Zn0.4Ti1.6O4 at 1018 cm-1. According to equation (2), and using the formula
mass of 198.02 g/mol (K0.8Zn0.4Ti1.6O4) and 256.42 g/mol (palmitic acid), the content of fully
intercalated palmitic acid would be as high as [(0.8×256.42)/(198.02 + 0.8×256.42)]×100% =
50.9%. So, it is reasonable that the uptake of 37% from GC analysis and TGA (Figure 6 and 7)
virtually originates from the “palmitic acid” adsorbed at the interlayer space of this material.
The fact that K0.8Zn0.4Ti1.6O4 intercalates carboxylic acids into the interlayer space has
prompted us to investigate its use as a catalyst in the gas-phase ketonization of palmitic acid. The
removal of oxygen atom by ketonization (i.e., the transformation of two molecules of the acid
into the ketone) could be coupled with the cracking of the corresponding ketone to C10 to C14
hydrocarbons applicable for use as a diesel fuel. Table 1 compares the conversion of palmitic
acid and the distribution of products obtained over tested materials. The 86% conversion of
palmitic acid was achieved over K0.8Zn0.4Ti1.6O4, giving palmitone C15H31(C=O)C15H31 as the
major product via ketonization process. Other groups of products include heavy ketone (mostly
C17 ketone CH3(C=O)C15H31), C10 to C14 hydrocarbons, and light hydrocarbons (<C10). These
products were formed via the cracking at the -position of palmitone, followed by further
cracking and/or hydrogen transfer. A detailed reaction pathway for the formation of these
20

products is under investigation and will be reported in a near future. Figure S5 shows the PXRD
pattern of spent K0.8Zn0.4Ti1.6O4, confirming that the lepidocrocite structure is preserved under
the heat treatment in air at 800ºC and subsequent contact with palmitic acid at 375ºC (see
Experimental section). The stability of K0.8Zn0.4Ti1.6O4 is in contrast to other layered materials
such as layered double hydroxide (LDH)9 or layered hydroxy salt (LHS)10 which undergo
structural transformation upon contact with liquid fatty acids at 100-140ºC. We also tested the
catalytic activity of the mixed (Mg/Al) oxide “MgAl2.5”, which is known to catalyze the
ketonization of carboxylic acids by its basic sites.51, 52, 53 MgAl2.5 gave an almost complete
conversion of palmitic acid with a smaller yield of palmitone (36%, vs 59% over
K0.8Zn0.4Ti1.6O4), together with the larger yield of hydrocarbon products.
Table 1. Conversion of palmitic acid and the distribution of liquid products over three
compositions of lepidocrocite titanate and MgAl2.5.a
Catalyst
K0.8Zn0.4Ti1.6O4 K0.8Mg0.4Ti1.6O4 K0.8Li0.27Ti1.73O4 MgAl2.5
Conversion
85.6
96.5
97.9
96.9
Product distribution
Palmitone
58.6
7.1
17.7
1.0
41.5
28.1
26.5
1.0
39.2
23.4
34.5
2.9
35.5
37.0
22.6
5.0
Heavy ketoneb
C10 to C14 hydrocarbons
Light hydrocarbons
a
Reaction conditions: 5% palmitic acid in p-xylene; flow rate of feed plus carrier gas (N2) 30
mL·min-1; reaction temperature 375oC, atmospheric pressure; W/F = 1500 g·h/mol. The values
shown were averaged at the time on stream 225-360 min where conversion and product
distributions were stabilized.
b
C17 ketone CH3(C=O)C15H31 and others
21

Besides, an almost complete conversion of palmitic acid over two other compositions of
lepidocrocite titanate was obtained (Table 1). This result can be explained by their larger specific
surface area (K0.8Mg0.4Ti1.6O4, 13 m2/g; K0.8Li0.27Ti1.73O4, 28 m2/g), among several factors. Since
the adsorption of palmitic acid must take place initially before subsequent diffusion and
activation at the interlayer space, the catalysts with higher surface area would facilitate the latter
two processes. The increase in palmitic acid conversion over these two catalysts is hence
accompanied by a decrease in the yield of palmitone, as compared to that over K0.8Zn0.4Ti1.6O4.
Yet, the yields of hydrocarbons over Mg and Li-containing lepidocrocite titanate (26-34%) are
higher than those over MgAl2.5 (22%). This result indicates a better selectivity of Mg- and Li-
containing lepidocrocite titanate toward cracked hydrocarbons, as compared to typical basic
oxides.
The high conversion (86%) of palmitic acid over K0.8Zn0.4Ti1.6O4 with a low specific
surface area (3 m2/g) suggests that the observed activity is a result of the active sites at the
interlayer space. In line with this view, an SEM image (Figure 2b) shows that the surface of
spent K0.8Zn0.4Ti1.6O4 was roughened and the crystals expanded, in comparison to the dense
crystals of as made K0.8Zn0.4Ti1.6O4 with a smooth surface. We propose that as made
K0.8Zn0.4Ti1.6O4 intercalates palmitic acid vapor, as it does in the liquid phase. The intercalated
palmitic acid and the reaction products might exert appreciable mechanical stress to the layers
sandwiching them, such that the layers are deformed and then the crystals expanded. However,
the preservation of the crystal structure and the associated two-dimensionality is clearly
beneficial.
22

Conclusions
The surface and interlayer basic characters in lepidocrocite titanate K0.8Zn0.4Ti1.6O4 was
investigated. K0.8Zn0.4Ti1.6O4 is a weak base as evaluated by CO2 TPD. Besides, the low basicity
implies the major contribution from O2- sites at the external surfaces, while those at the interlayer
region are not accessible to CO2. Yet, the liquid-phase adsorption study suggests the intercalation
of palmitic acid into the interlayer region as palmitate salts. The accessibility of interlayer basic
sites in K0.8Zn0.4Ti1.6O4 was further demonstrated by its use as a catalyst in the gas phase
ketonization of palmitic acid. Despite of small (external) surface area and low basicity, this
lepidocrocite titanate was active and selective for the conversion of palmitic acid to
hydrocarbons via ketonization-cracking. The layered structure was also preserved in contrast to
other 2D basic catalysts reported so far. Moreover, the catalytic activity of lepidocrocite titanate
can be tuned by the variation of the metal MI and MII substituting at TiIV sites.
Acknowledgements
The financial support from the Thailand Research Fund to T.M. (TRG5780160) and T.S.
1
(BRG5680007) is acknowledged. The H ultra-fast MAS NMR experiments were supported by
the Natural Science Center for Basic Research and Development, Hiroshima University.
Supplementary Material: PXRD patterns and additional material and methods available
23

Reference
1 Hattori, H. J. Jpn. Petrol. Inst., 2004, 47, 67-81.
2 Barthomeuf, D. Catal. Rev. Sci. Eng., 1996, 38, 521-612.
3 Ono, Y. J. Catal., 2003, 216, 406-415.
4 Weitkamp, J.; Hunger, M.; Rymsa, U. Microporous Mesoporous Mater., 2001, 48, 255-270.
5 Lee, D. –W. ; Park, Y. –M.; Lee, K. –Y. Catal. Survey Asia, 2009, 13, 63-77.
6 Chen, L.; Zhao, J.; Yin, S. –F.; Au, C. –T. RSC Adv., 2013, 3, 3199-3814.
7 Debecker, D. P.; Gaigneaux, E. M.; Busca, G. Chem. Eur. J., 2009, 15, 3920-3935.
8 Ma, R.; Sasaki, T. Acc. Chem. Res., 2015, 48, 136-143.
9 Cordeiro, C. S.; da Silva, F. R.; Marangoni, R.; Wypych, F.; Ramos, L. P. Catal. Lett., 2012,
142, 763-770.
10 Cordeiro, C. S.; Arizaga, G. G. C.; Ramos, L. P.; Wypych, F. Catal. Commun., 2008, 9, 2140-
2143.
11 Groult, D.; Mercey, C.; Raveau, B. J. Solid State Chem., 1980, 32, 289-296.
12 Reid, A. F.; Mumme, W. G.; Wadsley, A. D. Acta Cryst. 1968, B24, 1228-1233.
13 Gao, T.; Fjellvåg, H.; Norby, P. Chem. Mater., 2009, 21, 3503-3513.
14 Gao, T.; Norby, P.; Okamoto, H.; Fjellvåg, H. Inorg. Chem., 2009, 48, 9409-9418.
15 England, W. A.; Birkett, J. E.; Goodenough, J. B.; Wiseman, P. J. J. Solid State Chem., 1983,
49, 300-308.
16 Gao, T.; Fjellvåg, H.; Norby, P. J. Mater. Chem., 2009, 19, 787-794.
17 Song, H. Y.; Sjastad, A. O.; Vistad, O. B.; Gao, T.; Norby, P. Inorg. Chem., 2009, 48, 6952-
6959.
18 Sasaki, T.; Kooli, F.; Iida, M.; Michiue, Y.; Takenouchi, S.; Yajima, Y.; Izumi, F.;
Chakoumakos, B. C.; Watanabe, M. Chem. Mater., 1998, 10, 4123-4128.
19 Sasaki, T.; Watanabe, M.; Michiue, Y.; Komatsu, Y.; Izumi, F.; Takenouchi, S. Chem. Mater.,
1995, 7, 1001-1007.
20 Grey, I. E.; Li, C.; Madsen, I. C.; Watts, J. A. J. Solid State Chem., 1987, 66, 7-19.
21 Kitano, M.; Nakajima, K.; Kondo, J. N.; Hayashi, S.; Hara, M. J. Am. Chem. Soc., 2010, 132,
6622-6623.
22 Draskovic, T. I.; Wang, T.; Henderson, C. N.; Mallouk, T. E. Int. J. Hydrogen Energy, 2014,
39, 4576-4580.
23 Fukuda, K.; Akatsuka, K.; Ebina, Y.; Ma, R.; Takada, K.; Nakai, I.; Sasaki, T. ACS Nano,
2008, 2, 1689-1695.
24 Maluangnont, T.; Matsuba, K.; Geng, F.; Ma, R.; Yamauchi, Y.; Sasaki, T. Chem. Mater.,
2013, 25, 3137-3146.
25 Sasaki, T.; Watanabe, M. J. Am. Chem. Soc., 1998, 120, 4682-4689.
26 Gaertner, C. A.; Serrano-Ruiz, J. C.; Braden, D. J.; Dumesic, J. A. J. Catal., 2009, 266, 71-78.
27 Deng, L.; Fu, Y.; Guo, Q. X. Energy Fuels, 2009, 23, 564-568.
28 Pham, T. N.; Sooknoi, T.; Crossley, S. P.; Resasco, D. E. ACS Catal., 2013, 3, 2456-2473.
29 Renz, M. Eur. J. Org. Chem. 2005, 979-988.
30 Climent, M. J.; Corma, A.; Iborra, S.; Epping, K.; Velty, A. J. Catal., 2004, 225, 316-326.
31 Grigorieva, A. V.; Yuschenko, V. Y.; Ivanova, I. I.; Goodilin, E. A.; Tretyakov, Y. D. J.
Nanomater., 2012, 920483.
32 Santos-Lopez, I. A.; Handy, B. E.; Garcia-de-Leon, R. Thermochimica Acta 2013, 567, 85-92.
33 Raupp, G. B.; Dumesic, J. A. J. Phys. Chem., 1985, 89, 5240-5246.
34 Hu, W.; Li, L.; Li, G.; Liu, Y.; Withers, R. L. Sci. Rep. 2014, 4, 6582.
24

35 Izawa, H.; Kikkawa, S.; Koizumi, M. J. Phys. Chem., 1982, 86, 5023-5026.
36 Sasaki, T.; Watanabe, M.; Komatsu, Y.; Fujiki,Y. Inorg. Chem., 1985, 24, 2265-2271.
37 Sasaki, T.; Komatsu, Y.; Fujiki, Y. Chem. Mater., 1992, 4, 894-899.
38 Markovits, A.; Fahmi, A.; Minot, C. J. Mol. Struct. Theochem 1996, 371, 219-235.
39 Takagaki, A.; Sugisawa, M.; Lu, D.; Kondo, J. N.; Hara, M.; Domen, K.; Hayashi, S. J. Am.
Chem. Soc., 2003, 125, 5479-5485.
40 Yu, G.; Zhou, Y.; Yang, R.; Wang, M.; Shen, L.; Li, Y.; Xue, N.; Guo, X.; Ding, W.; Peng, L.
J. Phys. Chem. C, 2015, 119, 12325-12334.
41 Bavykin, D. V.; Carravetta, M.; Kulak, A. N.; Walsh, F. C. Chem. Mater., 2010, 22, 2458-
2465.
42 Corcoran, D. J. D.; Tunstall, D. P.; Irvine, J. T. S. Solid State Ionics 2000, 136-137, 297-303.
43 Maluangnont, T.; Bui, G. T.; Huntington, B. A.; Lerner, M. M. Chem. Mater., 2011, 23, 1091-
1095.
44 Maluangnont, T.; Sirisaksoontorn, W.; Lerner, M. M. Carbon 2012, 50, 597-602.
45 Nakano, S.; Sasaki, T.; Takemura, K.; Watanabe, M. Chem. Mater., 1998, 10, 2044-2046.
46 Fujimoto, K.; Tobito, N.; Ito, S. J. Ion Exchange 2015, 25, 12-15.
47 Shirpour, M.; Cabana, J.; Doeff, M. Chem. Mater., 2014, 26, 2502-2512
48 Sasaki, T.; Izumi, F.; Watanabe, M. Chem. Mater., 1996, 8, 777-782
49 Arizaga, G. G. C.; Mangrich, A. S.; da Costa Gardolinski, J. E. F.; Wypych, F. J. Colloid
Interface Sci., 2008, 320, 168-176.
50 Chou, C. –C.; Chiang, M. –L.; Lin, J. –J. Macromol. Rapid Commun. 2005, 26, 1841-1845.
51 Kandare, E.; Hossenlopp, J. M. Inorg. Chem., 2006, 45, 3766-3773.
52 Parida, K.; Das, J. J. Mol. Catal. A 2000, 151, 185-192.
53 Das, J.; Parida, K. React. Kinet. Catal. Lett. 2000, 69, 223-229.
Highlights
 K0.8Zn0.4Ti1.6O4 intercalates palmitic acid, forming the occluded potassium salt.
 The interlayer expansion is evidenced by PXRD patterns and TEM image.
 Two types of basic sites are deduced from ultrafast 1H MAS NMR
 Lepidocrocite titanate catalyses ketonization of palmitic acid to palmitone and
corresponding cracked products.
Graphical Abstract Legend
Interlayer basic sites in lepidocrocite titanate, K0.8Zn0.4Ti1.6O4, lead to an intercalation of palmitic
acid with a layer expansion.
25

26