An efficient and durable hierarchically porous KLA/TiPO catalyst for vapor phase condensation of lactic acid to 2,3-pentanedione

Article abstract of DOI:10.1039/c9nj00267g

Sustainable production of 2,3-pentanedione from bio-lactic acid via a vapor condensation reaction over KLA/TiPO (KLA: potassium lactate) was investigated in this work. A KNO₃ precursor supported on the surface of TiPO was in situ converted to basic sites in a KLA/TiPO catalyst. KLA together with Ti⁴⁺(Lewis acidic site) make up the acid-base pairs in the KLA/TiPO catalyst, resulting in excellent activity for the condensation of lactic acid to 2,3-pentanedione. The loading amount of KNO₃ was shown to have an important influence on the catalytic performance, since the acid-base properties of the catalysts were found to vary with the addition of KNO₃. Reaction conditions such as lactic acid feed flow rate and lactic acid concentration were also discussed. Both lactic acid conversion and 2,3-pentanedione selectivity increased with elevated lactic acid feed flow rates, indicating the existence of an external diffusion resistance of the lactic acid reactant during the catalytic reactions. However, the lactic acid feed flow rate increased to 1.0 mL h^?1 (corresponding to LA liquid hourly space velocity (LHSV) = 2.6 H^?1), and the external diffusion resistance was efficiently eliminated. Enhancing the LA concentration improved the selectivity of 2,3-pentanedione, suggesting that the reaction order of the lactic acid molecule for lactic acid conversion to 2,3-pentanedione is higher than the other side reactions. Encouragingly, in retaining 30-45% of the lactic acid conversion, the condensation reaction with a 2,3-pentanedione selectivity of around 73% proceeded efficiently for at least 116 h on stream. The long-term stability of the present catalyst was found to be related to its hierarchical pores, which ameliorated the mass transfer effect of the reactant and product, except for the appropriate acid-base properties for lactic acid condensation to 2,3-pentanedione.

Full text of DOI:10.1039/c9nj00267g

NJC
View Article Online
View Journal
PAPER
An efficient and durable hierarchically porous
KLA/TiPO catalyst for vapor phase condensation
of lactic acid to 2,3-pentanedione†
Cite this:DOI: 10.1039/c9nj00267g
ab
a
ab
c
a
Ju Zhang,‡ Xinli Li,‡ Jun Pang, Weixin Zou, Congming Tang * and
Lin Dong
c
Sustainable production of 2,3-pentanedione from bio-lactic acid via a vapor condensation reaction over
3
KLA/TiPO (KLA: potassium lactate) was investigated in this work. A KNO precursor supported on the
4+
surface of TiPO was in situ converted to basic sites in a KLA/TiPO catalyst. KLA together with Ti (Lewis
acidic site) make up the acid–base pairs in the KLA/TiPO catalyst, resulting in excellent activity for the
condensation of lactic acid to 2,3-pentanedione. The loading amount of KNO
important influence on the catalytic performance, since the acid–base properties of the catalysts were
found to vary with the addition of KNO . Reaction conditions such as lactic acid feed flow rate and lactic
3
was shown to have an
3
acid concentration were also discussed. Both lactic acid conversion and 2,3-pentanedione selectivity
increased with elevated lactic acid feed flow rates, indicating the existence of an external diffusion resis-
tance of the lactic acid reactant during the catalytic reactions. However, the lactic acid feed flow rate
À1
À1
increased to 1.0 mL h (corresponding to LA liquid hourly space velocity (LHSV) = 2.6 H ), and the
external diffusion resistance was efficiently eliminated. Enhancing the LA concentration improved the
selectivity of 2,3-pentanedione, suggesting that the reaction order of the lactic acid molecule for lactic
acid conversion to 2,3-pentanedione is higher than the other side reactions. Encouragingly, in retaining
30–45% of the lactic acid conversion, the condensation reaction with a 2,3-pentanedione selectivity of
around 73% proceeded efficiently for at least 116 h on stream. The long-term stability of the present
catalyst was found to be related to its hierarchical pores, which ameliorated the mass transfer effect of
the reactant and product, except for the appropriate acid–base properties for lactic acid condensation
to 2,3-pentanedione.
Received 17th January 2019,
Accepted 12th March 2019
DOI: 10.1039/c9nj00267g
rsc.li/njc
LA has also been converted to a variety of chemicals, for
Introduction
8
–14
example, acrylic acid via LA dehydration,
acetaldehyde via LA
1
5–19
20–23
Nowadays, interesting research is being carried out on the synthesis decarbonylation,
propionic acid via LA deoxygenation,
1
,2
24
of fine/bulk chemicals from biomass or its derivatives. For poly-lactic acid via LA polycondensation, pyruvic acid via LA
2
5
26–29
example, many bio-chemicals such as ethanol, lactic acid (LA) oxidation and 2,3-pentanedione via LA condensation.
3
–5
and furfural can be obtained from corn, corncob and sugars.
Among these chemicals derived from LA, 2,3-pentanedione has
Based on these bio-chemicals as raw materials, developing widespread commercial value and good prospects for research
downstream products is necessary for a sustainable society. As from LA since it is a high value-added fine chemical product.
one of the listed bio-chemicals, LA can be prepared by microbial 2,3-Pentanedione has important applications as a flavoring agent
6
,7
30
fermentation and chemical catalysis, where the raw materials and photoinitiator. In addition, it can also be used as an important
may be cellulose and glycerol, but not corn. In recent years, intermediate in chemical synthesis. For example, 2,3-pentanedione
reacts with phenol to synthesize novel bisphenols, serving as an
3
1
a
b
c
excellent plasticizer for polyethylene terephthalate.
School of Chemistry and Chemical Engineering, Chongqing University of
Technology, Chongqing 400054, P. R. China. E-mail: tcmtang2001@163.com
Chemical Synthesis and Pollution Control Key Laboratory of Sichuan Province,
China West Normal University Nanchong, Sichuan, 637002, P. R. China
Jiangsu Key Laboratory of Vehicle Emissions Control, Center of Modern Analysis,
Nanjing University, Nanjing 210093, P. R. China
The production of 2,3-pentanedione from LA is widely believed
to be a green and potential route that proceeds via catalytic vapor-
phase synthesis, since other routes discharge liquid waste acid.
In LA condensation into 2,3-pentanedione, alkali metal salts
are usually utilized as a precursor, converting into an active
species, and the evaluated supports are carbon, hydroxyapatite,
32,33
†
‡
Electronic supplementary information (ESI) available. See DOI: 10.1039/c9nj00267g
J. Zhang and X. L. Li contributed equally to this work.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019
New J. Chem.

View Article Online
Paper
NJC
a
Si–Al compounds, barium pyrophosphate and silicate, from Table 1 K component loading amount versus catalytic performance
2
6,27,29,33–38
previous work.
been used as a precursor in some studies.
metal hydroxides have strong hygroscopicity, resulting in con-
Alkali metal hydroxides have also
c
Sel. /%
2
9,39–41
However, alkali
b
KNO
3
: TiPO
LA conv. /%
2,3-PD
AD
AC
PA
AA
4.1
18.3
12.2
10.3
9.4
11.1
11.1
11.9
d
e
siderable difficulty in solid catalyst moulding due to the water 0 : 1
73.9
41.5
57.1
72.7
73.4
75.7
87.5
74.8
—
81.2
17.9
2.8
3.7
4.5
3.9
3.8
3.6
1.7
6.5
3.4
2.3
2.3
2.2
1.8
2.7
12.1
34.0
25.0
20.1
11.0
13.2
10.2
13.3
0
1
.5 : 1
: 1
22.9
56.2
63.0
66.5
69.0
72.7
67.9
content in air. As a result, subsequent work used alkali nitrates
to replace the corresponding hydroxides. CsNO
investigated in several alkali nitrates, and offered high initial
3
was fully 2 : 1
3
: 1
4 : 1
: 1
2
6–28
activity.
3 3
In comparison with CsNO , KNO has been rarely
5
studied. Considering that 2,3-pentanedione is a food flavoring, 6 : 1
utilizing the latter instead of the former is safer. Besides this,
exploring the process of the active species is very important
to understand the catalytic mechanism in LA condensation to 1.0 mL h ; reaction temp.: 270 1C; TOS: 7–8 h. LA: lactic acid. 2,3-PD:
a
Catalyst, 0.38 mL, 0.30–0.33 g; particle size: 20–40 meshes; carrier
À1
gas N
2
: 1.2 mL min ; LA feedstock: 20 wt% in water, feed flow rate:
À1
b
c
2
,3-pentanedione, AD: acetaldehyde, AC: acetic acid, PA: propionic acid,
2
,3-pentanedione. However, so far the research is very rare in
this regard.
As for the support, the surface acid–base properties have
d
e
AA: acrylic acid. Bare support, TiPO. ‘‘—’’, trace.
been more focused on, and the pore structure has hardly been abruptly reduced. Meanwhile, another product, propionic acid,
investigated for LA condensation to 2,3-pentanedione. The was extensively formed. These results indicate that it is necessary
hierarchical pore structure of the support has been demonstrated to use KNO as a precursor for LA condensation to produce 2,3-
3
4
2–44
19,47
to have many advantages.
molecules have high mass transfer rate in porous channels due of propionic acid is related to the surface basicity of the catalyst,
to relatively low mass transfer resistance, which can effectively indicating that supporting KNO on TiPO, increases the number
This effect, brought of basic sites in the catalyst. With a continuous increase in the
about by hierarchical pores, is more evident and positive for amount of KNO , LA conversion and 2,3-pentanedione selectivity
catalytic reactions with high inner diffusion resistance. In the also steadily rose. For example, as the mass ratio of KNO
For example, reactant/product pentanedione. According to previous reports,
the formation
3
45,46
eliminate or reduce secondary reactions.
3
3
:
catalytic reaction of LA conversion to 2,3-pentanedione, the three TiPO increased to 5 :1 (by weight), the LA conversion and 2,3-
functional hydroxyl, carboxyl and carbonyl groups in LA and pentanedione selectivity were up to 87.5% and 72.7%, respectively.
2
,3-pentanedione are polar, resulting in molecular polarity. For With a further increase in KNO
that reason, mass transfer resistance of these molecules increases 2,3-pentanedione selectivity hardly rose, even reducing slightly
in the channels of catalysts. It is hoped that catalytic performance (ca., KNO : TiPO = 6 : 1).
can be improved if a catalyst is constructed with a hierarchical
To understand the effects of KNO
3
, the LA conversion and
3
loading amount on the
3
porous structure. TiPO is similar to previous supports such as catalytic performance, several characterization methods, such
hydroxyapatite, with acidic properties. Besides this, it can easily as Fourier-transform infrared spectroscopy (FT-IR), X-ray diffraction
be constructed with a hierarchically porous structure.
2 3
(XRD) and CO (NH )-temperature programmed desorption (TPD),
In this work, we developed a simple and efficient method for were used. From the FT-IR spectra of the fresh catalysts shown
the preparation of hierarchically porous KLA/TiPO (KLA: potassium in Fig. 1a, strong characteristic absorption bands at 1381 and
À1
À
48
lactate) materials, which were utilized to catalyze LA condensation 825 cm can be observed, indexed to NO
3
derived from KNO
3
,
À1
to 2,3-pentanedione. We further revealed that the acid–base while the wide characteristic absorption band at 1070 cm
49,50
properties of the catalysts changed with KNO
3
loading amounts indexed to TiPO decreased.
Besides this, the XRD patterns of
and the effect of the hierarchical porous structure of TiPO on the the fresh catalysts shown in Fig. 2a also reflect similar results. For
catalytic performance for LA condensation to 2,3-pentanedione.
example, the characteristic diffraction peaks at 2y = 23.51, 23.81,
29.41 and 33.81, attributed to the crystal faces (111), (102), (201)
and (211), respectively, match well with the standard sample of
Results and discussion
K component loading amount versus catalytic performance
Table 1 shows the effect of the K component loading amount
on the catalytic performance for the condensation of LA to
2
,3-pentanedione with a LA concentration of 20 wt% and LA flow
À1
rate of 1.0 mL h . With only the bare TiPO support, almost no
2,3-pentanedione was formed, but LA was largely converted to
acetaldehyde with a LA conversion of 73.9% and acetaldehyde
selectivity of 81.2%. When a small quantity of KNO3 was
3
supported on the TiPO support (ca., KNO : TiPO = 0.5 : 1, by
weight), the activity for the formation of 2,3-pentanedione from
Fig. 1 FT-IR spectra of the fresh catalysts (a) and their corresponding
LA drastically increased, while selectivity of the acetaldehyde used catalysts (b).
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019

View Article Online
NJC
Paper
K component, the Lewis acid sites drastically increased to
À1
0.041 mmol g
and the Brønsted acid sites disappeared.
Further combined with the above discussion of the activity,
the active sites for the catalytic condensation of LA to
2,3-pentanedione require an acid–base pair, so the Lewis acidic
sites play an important role. Thus, tuning the Lewis acid–base
pair of catalysts can hopefully result in efficient catalysis for
condensation of LA to 2,3-pentanedione.
Fig. 2 XRD patterns of the fresh catalysts (a) and their corresponding used
catalysts (b).
Effect of LA LHSV (liquid hourly space velocity)
In preliminary experiments, the effect of LA concentration on
KNO
diffraction peaks increased upon an increase in the amount of shown in Table S2 (ESI†). In a fixed LA feed flow rate of 1.0 mL h
KNO , except for the KNO : TiPO = 5 : 1 sample. As a result, this and reaction temperature of 270 1C, the 2,3-pentanedione selectivity
evidence indicates that the method can effectively load KNO onto increased with LA concentration, since the LA conversion to
3
(JCPDS no. 74-2055). The intensities of these characteristic the reaction performance was investigated, and the results are
À1
3
3
3
the surface of TiPO according to a pre-designed loading strategy. It 2,3-pentanedione is a secondary order reaction in terms of the LA
was further observed from the used catalysts that the characteristic concentration, whereas the other accompanying side reactions to
absorption band and diffraction peaks indexed to KNO dis- form acrylic acid and acetaldehyde are first order in terms of the LA
3
À1
27,34,35
appeared, and new absorption bands at 1277–1459 cm
,
concentration.
LA conversion did not increase lineally with an
occurred and increased in intensity (Fig. 1b and 2b). Notably, increase in LA concentration, but a maximum of 88.9% at a LA
no new diffraction peaks were observed in the XRD patterns of concentration of 40 wt% was achieved. Next, the effect of LA LHSV
the used catalysts, suggesting that the new species can be was further evaluated at a LA concentration of 40 wt% and reaction
ascribed as amorphous.
temperature of 270 1C, to understand the catalytic efficiency of the
Due to the change in the KNO precursor during the catalytic KNO /TiPO catalyst with a KNO : TiPO ratio of 4 : 1, and the results
3
3
3
reaction, measuring the surface acid–base properties of the are shown in Table 2. Here, the LA flow rate changed from 0.5 to
À1
À1
catalysts after reaction was by far more important than before 3.0 mL h (corresponding to a LA LHSV of 1.3–7.8 h ). For LA
À1
reaction, to deeply understand the acid–base catalysis for the conversion, 85.4% at a LA LHSV of 1.3 h
catalytic condensation of LA to 2,3-pentanedione. Thus, CO (NH
was achieved,
À1
2
3
)- whereas 87.2% was achieved at a LA LHSV of 2.6 h , indicating
TPD measurements were performed on the used catalysts, and the that enhancing the LA feed flow rate reduces the external
results are shown in Fig. 3. With an increase in the KNO
diffusion resistance and increases the reaction rate. As the LA
precursor, the temperature of the NH -desorption peak moved feed flow rate further increased, the external diffusion resistance
3
3
towards a high temperature (Fig. 3a), suggesting that the acidity did not decrease any more, and the surface catalytic reaction rate
strength increased. At the same time, the area of the NH3- gradually dominated in the conversion of LA to 2,3-pentanedione.
desorption peaks increased, indicating that the density of the As a result, the LA conversion slowly decreased from 87.2% to
À1
À1
acidic sites was enhanced. Similar phenomena can also be observed 60.7% when the LA LHSV increased from 2.6 h to 7.8 h . As for
in the CO -TPD profiles shown in Fig. 3b. The CO (NH )-TPD the LA consumption rate and 2,3-pentanedione formation rate,
2
2
3
measurements showed that acidic and basic sites occurred in pairs these always increased with the LA feed flow rate, as shown in
on the surfaces of the catalysts, and their acid–base properties can Fig. 4. For example, the LA consumption rate and 2,3-pentanedione
À2 À1
À1
À2 À1
be controlled by varying the KNO
3
loading amount (Table S1, ESI†). formation rate were 1.2 mmol m
h
and 0.4 mmol m
h
,
À2 À1
In situ IR-Py (IR spectroscopy of adsorbed pyridine) was performed respectively, at LA LHSV = 1.3 h , while 5.1 mmol m
h
and
À2 À1
À1
to classify the acidic sites, and the results are shown in Fig. S1 1.9 mmol m
h
were achieved, respectively, at LA LHSV = 7.8 h
.
(
ESI†). It can be clearly seen that the bare support of TiPO is rich in
À1
Brønsted acid sites (0.014 mmol g ) and has fewer Lewis acidic
sites (0.006 mmol g ). However, on the modified catalyst with
a
À1
Table 2 Effect of LA LHSV
c
Sel. (%)
À1
b
LA LHSV (h
)
LA conv. (%)
2,3-PD
AD
AC
PA
AA
1
2
3
5
6
7
.3
.6
.9
.2
.5
.8
85.4
87.2
78.2
74.9
64.9
60.7
73.4
79.8
75.5
76.7
75.7
74.8
2.9
3.7
1.8
2.5
1.6
1.4
2.1
1.1
3.7
2.5
3.0
3.4
13.1
10.2
10.3
9.1
9.1
8.8
7.1
4.5
8.1
8.2
9.2
10.6
a
KNO
3
/TiPO catalyst with KNO
3
: TiPO = 4 : 1, 0.38 mL, 0.30 g; particle
À1
size: 20-40 meshes; carrier gas N
2
: 1.2 mL min ; LA feedstock: 40 wt%
b
c
in water; reaction temp.: 270 1C; TOS: 7–8 h. LA: lactic acid. 2,3-PD:
,3-pentanedione, AD: acetaldehyde, AC: acetic acid, PA: propionic acid,
AA: acrylic acid.
2
3 2
Fig. 3 NH -TPD (a) and CO -TPD (b) profiles of the used catalysts.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019
New J. Chem.

View Article Online
Paper
NJC
Fig. 5 LA conversion and 2,3-pentanedione selectivity as a function
Fig. 4 LA consumption and 2,3-pentanedione formation rate versus LA
of time on stream over KNO
3 3 3
/TiPO. KNO /TiPO catalyst with KNO :
LHSV.
TiPO = 4 : 1, 0.38 mL, 0.30 g; particle size: 20–40 meshes; carrier gas
À1
À1
;
N
2
: 1.2 mL min ; LA feedstock: 20 wt% in water, LA feed flow rate: 1.0 mL h
reaction temp.: 270 1C. LA: lactic acid, 2,3-PD: 2,3-pentanedione.
It was also noted that the 2,3-pentanedione selectivity fluctuated
slightly from 73.4% to 79.8% according to a variation in the range of
À1
À1
was observed in the initial 2 h on stream, suggesting no activity
of KNO together with TiPO for the condensation of LA to
,3-pentanedione. Whereas, while the production of 2,3-pentane-
dione gradually increased with time on stream, the KNO was also
LA LHSV from 1.3 h to 7.8 h , suggesting that 2,3-pentanedione
selectivity is slightly influenced by contact time between LA and the
surface of the catalyst. These observations do not agree with the
3
2
2
6,37
3
results obtained in previous work.
This result may be related
2
7,32 gradually converted to KLA (Fig. 5). This evidence indicates that
KLA possibly acts as the active species. The used catalyst after
more than 116 h on stream was characterized by IR spectroscopy
to the catalyst and reaction temperature. In previous work,
the reaction temperature of around 290–300 1C was higher than
that used in this work (270 1C), suggesting that the present
catalyst has better activity than in previous work. At a relatively
lower reaction temperature, the side reactions have lower reaction
rates, and the formation rate of 2,3-pentanedione dominates (as
seen in Table S3, ESI†). For that reason, the choice of an appropriate
LA LHSV mainly affects the LA conversion. In other words, we have
the opportunity to choose a higher LA LHSV to enhance the
(shown in Fig. S4, ESI†) and its activity as a function of time on
stream is shown in Fig. 9. We found that the KLA species was
stable and the corresponding catalyst offered excellent activity,
further increasing the possibility that KLA is the active species.
Next, we removed the KLA from the surface of the catalyst by
washing with water (also shown in Fig. S4, ESI†), and used it to
catalyze the condensation of LA again (Fig. 6). As expected, almost
no activity was observed. Thus, we further confirmed KLA as the
active species.
2,3-pentanedione space-time yield of the present catalyst.
Formation of the active species and its catalytic reaction
mechanism
Fig. 7 shows the FT-IR spectra and XRD patterns of the
catalyst after time on stream. From the FT-IR spectra shown in
Fig. 7a, the characteristic absorption bands indexed to KNO
3
at
It is an important process in understanding catalysis to search
À1
825 and 1381 cm drastically disappeared, and instead, new
for the active species. Over the KNO
3
/TiPO catalyst with a
KNO : TiPO ratio of 4 : 1, we attempted to detect the intermediates
3
to confirm the active species in the condensation of lactic acid
to 2,3-pentanedione. Three sets of contrasting experimental
conditions were used: (1) without LA and pure water, feeding
only a carrier gas, (2) feeding pure water as well as a carrier gas
and (3) feeding a LA aqueous solution as well as a carrier gas.
For the third set of experimental conditions, the formation of a
reddish-brown gas of NO
not observed for the former two sets of experimental conditions
shown in Fig. S2, ESI†). Furthermore, the catalysts used in three
2
was clearly observed, while this was
(
sets of experimental conditions were characterized with XRD
and FT-IR, and the results are shown in Fig. S3 (ESI†). Evidently,
the former two sets of experimental conditions showed almost
the same structure of the catalyst in comparison with the third
set of conditions. These results show that LA reacts with KNO3
Fig. 6 LA conversion and 2,3-pentanedione selectivity as a function of
to form HNO
poses at a high reaction temperature, resulting in the generation
of NO
3 3
and KLA, and the formed HNO further decom-
time on stream over the catalyst removed K component. Carrier gas
À1
N
2
: 1.2 mL min ; LA feedstock: 20 wt% in water, LA feed flow rate:
À1
. It is noted that only a trace amount of 2,3-pentanedione 1.0 mL h ; reaction temp.: 270 1C. LA: lactic acid, 2,3-PD: 2,3-pentanedione.
2
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019

View Article Online
NJC
Paper
acid–base properties of the catalyst can be ascribed to KNO
3
transforming to KLA, since they are fundamentally related to
strong base and weak acid salts, exhibiting alkalinity on stream
4
+
at a high temperature. Acidic sites (Ti , Lewis acid sites)
formed and were further enhanced, ascribed to the combining
+
À
of K with TiPO, which can be verified by location of the NO
3
À1
absorption band at 1381 cm in the FT-IR spectra, as shown in
Fig. 1a. In recent work, hydroxyapatite was doped with Cs , and
+
the acidity was also observed to increase with the addition
2
7
Fig. 7 FT-IR spectra (a) and XRD patterns (b) of the sample after time on of the Cs component. As a result, the K component was
stream.
supported on TiPO and in situ converted to KLA, further
generating acid–base pairs, which can efficiently catalyze the
LA condensation to 2,3-pentanedione.
À1
absorption bands at 500–1800 cm attributed to KLA appeared,
suggesting that KNO reacted with LA to form KLA, accompanied
by the removal of HNO due to decomposition at a high
temperature. In the XRD patterns of samples shown in Fig. 7b,
the fresh sample exhibits clear characteristic diffraction peaks at
Based on the above discussion and previous investiga-
3
26,27,29,36
tions,
a possible mechanism for the catalytic condensation
3
of LA to 2,3-pentanedione was proposed as follows. (1) The
formation of an active species of KLA: LA reacted with the
precursor KNO to form KLA, accompanying by the generation
3
2
y = 23.51, 29.31, 33.71, and 41.01, respectively, that can be
indexed to KNO (JCPDS no. 71-2055). However, after 7 h on
stream in the catalytic reaction, these characteristic diffraction
peaks indexed to KNO disappeared, and the energy-dispersive
X-ray (EDX) spectroscopic analysis verified the existence of the K
element in the catalyst, indicating the transformation of KNO
These results are in good agreement with each other.
of HNO , as shown in chemical reaction eqn (1), then the
3
3
generated HNO3 decomposed quickly at high temperature
according to chemical reaction eqn (2). As a result, the chemical
equilibrium of reaction eqn (1) moved toward the right hand
side direction to form more KLA. In this stage, the 2,3-pentanedione
product was rarely produced and this process is known as the
induction period. (2) Acid–base catalysis (Scheme 1): after the
formation of KLA, this acted as a basic site (nucleophile), and
further attacked the carboxyl group of another LA molecule to form
a C6 compound (i). Next, the acidic sites in TiPO attacked the
hydroxyl group of the C6 compound (i) to form an intermediate (ii),
3
3
.
Previous work had demonstrated that the condensation of
LA to 2,3-pentanedione is related to the acid–base properties of
the catalyst. In this work, NH -TPD and CO -TPD were used to
3 2
2
7
investigate the acid–base properties of the catalysts. According
to the above discussion, it was important for us to monitor the
acid–base properties of the samples with time on stream, to
fully understand the condensation mechanism of LA. From the
measurements on the acid–base properties as shown in Fig. 8, a
drastic change was observed. At the initial 1 h of reaction, the
acidity/basicity rapidly increased. Subsequently, upon further
lengthening the reaction time, the strength of the acidity/
basicity continued to increase until up to 7 h of reaction time
on stream. It was noteworthy that the relative weak-medium
acidic/basic sites concentrated at 359, 358, 414 and 417 1C
decreased, and relatively strong acidic/basic sites were found to
be concentrated at 453 and 456 1C. In addition, acid–base pairs
occurred with reaction time on stream, and after 7 h of reaction
on stream, the acid–base pairs concentrated at 453 and 456 1C,
respectively, stabilized. The main reason for the change in the
+
further accompanied by elimination of a hydroxyl group and K
to form the intermediates (iii) and (iv). Intermediate (iv) was
converted to compound (v) via a decarboxylation reaction, which
can be confirmed by tail gas analysis, as we observed the
emission of carbon dioxide (Fig. S5, ESI†), and further isomerized
into the key intermediate (vi). Another LA molecule attacked the
intermediate (vi) to form the active species KLA, which completed
Fig. 8 NH
3
-TPD (a) and CO
2
-TPD (b) profiles of the sample after time on
Scheme 1 The proposed mechanism of the catalytic condensation of LA
stream.
to 2,3-pentanedione.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019
New J. Chem.

View Article Online
Paper
NJC
the catalytic cycle. At the same time, compound (vii) was also
formed in this step, and was further attacked by the acidic sites in
TiPO to form an enol (viii) via the dehydration reaction. Finally,
the enol form (viii) quickly isomerized into the 2,3-pentanedione
product.
(
1)
2)
(
Stability
From the early investigations, some were short on reports on
2
9,33,34,36
the stability of the catalyst,
others described the poor
2
8
stability of the catalyst, and only a few catalysts were reported
2
6,27
with acceptable stability.
Thus, it is still a crucial step to
Fig. 10
N
2
adsorption–desorption isotherms (A and C) and the corres-
improve the catalyst stability. In this work, we performed
ponding pore size distributions (B and D) of the bare support TiPO (A and B)
and the KNO /TiPO catalyst (C and D).
stability experiments on the catalyst at 270 1C, at a LA flow rate
3
À1
of 1.0 mL h (Fig. 9). Compared with the above experiments,
in the stability testing experiments, the catalyst amount was
reduced by two thirds in order to understand the real stability stability than previous catalysts such as Cs-doped hydroxy-
2
7
of active sites for the catalytic condensation of LA to 2,3- apatite (around 40 h on stream) and Cs-doped barium pyro-
2
6
pentanedione. During the initial 7 h on stream, the LA conversion phosphate (around 30 h on stream).
as well as the 2,3-pentanedione selectivity drastically increased
Why did the KLA/TiPO catalyst offer excellent performance,
with reaction time, which agreed well with the above observation, including activity and stability? According to the reported work,
also demonstrating the formation of the active species during the excellent activity is related to the acid–base properties of the
in situ reaction. As the reaction time exceeded 7 h, the reaction catalyst, since the catalytic condensation of LA to 2,3-pentane-
26,29,32,33
proceeded efficiently, showing excellent stability. It was noted that dione requires concerted catalysis with acidic and basic sites.
LA conversion was maintained at around 40%, indicating the full The results of the CO /NH -TPD of the catalyst with time on stream
2
3
utilization of all of the active sites in the catalytic condensation showed rich acidic and basic sites, favoring synergetic catalysis.
of LA to 2,3-pentanedione. In other words, catalyst stability Besides this, the KLA/TiPO catalyst has a hierarchical porous
reflects the real durability of active sites for catalytic reaction. structure containing mesoporous and macroporous pores (Fig. 10
Encouragingly, the catalytic reaction on the catalyst proceeded and Table S4, ESI†), where the reactant LA can quickly diffuse
efficiently for 116 h on stream, with only a slight decrease in LA toward the active sites and the product 2,3-pentanedione can
conversion observed, while the selectivity of 2,3-pentanedione quickly leave from the active sites due to relatively low mass
remained almost constant (B73%). This catalyst offered better transfer resistance. According to Gaussian calculations at the
b3lyp/6-31g(d) level, the molecular equivalent diameters for LA
and 2,3-pentanedione are 5.90 Å and 6.39 Å, respectively, and
their corresponding molecular polarities represented as dipole
moments are 1.7216 and 0.1407 Debye, respectively (Fig. S6,
ESI†). Due to their relatively big molecular size together with
molecular polarity, mass transfer process of LA and 2,3-pentane-
dione are improved as a result of the hierarchical catalyst
structure. Thus, the hierarchical structure of the catalyst efficiently
enhances the catalytic reaction rate, diminishing the occurrence of
side-reactions while enhancing the selectivity of 2,3-pentanedione.
Experimental
Preparation of the catalysts
The hierarchical support (TiPO) in this work was prepared
Fig. 9 Stability of the catalyst. KNO
3 3
/TiPO catalyst with KNO : TiPO =
51
according to previous work, as follows. At a stirring state under
ambient temperature, 0.21 g of Brij56 (Energy Chemical Company,
abbreviated as ECC) was fully dissolved in 10 mL of ethanol,
4
: 1, 0.10 g, diluted with quartz sand; particle size: 20–40 meshes; carrier gas
À1
À1
;
N
2
:1.2 mL min ; LA feedstock: 20 wt% in water, LA feed flow rate: 1.0 mL h
reaction temp.: 270 1C. LA: lactic acid, 2,3-PD: 2,3-pentanedione.
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019

View Article Online
NJC
Paper
followed by 3.45 g of Ti(OC
4
H
9
)
4
(ECC), to form a homogeneous Technology Department Application Foundation Project
3 4
solution. Next, 0.69 g of H PO (85 wt%, ECC) was added dropwise (2017JY0188).
to the resultant solution, and reacted with the titanium source to
form titanium phosphate at a continuous stirring state under
ambient temperature for 2 h. Then, the resultant solution was
slowly evaporated at 70 1C to remove the solvent, to obtain a white
solid composite. The resulting solid composite was calcined at
Notes and references
1 P. Maki-Arvela, I. L. Simakova, T. Salmi and D. Y. Murzin,
Chem. Rev., 2014, 114, 1909–1971.
2 H. Veisi, S. Azizi and P. Mohammadi, J. Cleaner Prod., 2018,
170, 1536–1543.
4
50 1C for 2 h under an air atmosphere.
KLA/TiPO was prepared using a wet immersion method.
Typically, 2.0 g of KNO (ECC) was fully dissolved in 20 mL of
3
H O at ambient temperature, and next, 0.5 g of TiPO was added
3 D. Esposito and M. Antonietti, ChemSusChem, 2013, 6, 989–992.
4 B. Tabah, I. N. Pulidindi, V. R. Chitturi, L. M. R. Arava,
A. Varvak, E. Foran and A. Gedanken, J. Mater. Chem. A,
2017, 5, 15486–15506.
2
under stirring. Then, the liquid was slowly evaporated at 80 1C
to form a white solid powder. Subsequently, the resultant
sample was used to catalyze the condensation reaction of LA
to 2,3-pentanedione, and KNO
a result, KLA/TiPO was formed.
3
was in situ converted to KLA. As
5 D. Kubicka and O. Kikhtyanin, Catal. Today, 2015, 243, 10–22.
6 L. S. Sharninghausen, J. Campos, M. G. Manas and R. H.
Crabtree, Nat. Commun., 2014, 5, 1–9.
Catalyst characterization
7
Y. L. Wang, W. P. Deng, B. J. Wang, Q. H. Zhang, X. Y. Wan,
Z. C. Tang, Y. Wang, C. Zhu, Z. X. Cao, G. C. Wang and
H. L. Wan, Nat. Commun., 2013, 4, 1–7.
XRD, FT-IR, specific surface area and NH
3 2
-TPD and CO -TPD
measurements were carried out using similar methods to those
2
6,28,52
described in previous work
and the detailed information
8 B. Yan, L. Z. Tao, A. Mahmood, Y. Liang and B. Q. Xu, ACS
Catal., 2017, 7, 538–550.
is shown in the ESI.†
9
Y. Noda, H. Zhang, R. Dasari, R. Singh, C. Ozmeral,
Y. Rom ´a n-Leshkov and R. M. Rioux, Ind. Eng. Chem. Res.,
Activity evaluation
1
5,27
2017, 56, 5843–5851.
Similar to previous work,
the activity of the catalyst was
1
1
1
1
1
1
1
1
0 X. H. Zhang, L. Lin, T. Zhang, H. O. Liu and X. F. Zhang,
Chem. Eng. J., 2016, 284, 934–941.
1 V. C. Ghantani, S. T. Lomate, M. K. Dongare and S. B.
Umbarkar, Green Chem., 2013, 15, 1211–1217.
2 J. F. Zhang, Y. L. Zhao, M. Pan, X. Z. Feng, W. J. Ji and
C. T. Au, ACS Catal., 2011, 1, 32–41.
evaluated using a vertical fixed-bed quartz tubular reactor, and
the detailed information is shown in the ESI.†
Conclusions
The catalytic condensation of LA to 2,3-pentanedione over KLA/
TiPO was investigated. The active sites for LA condensation to
3 J. S. Peng, X. L. Li, C. M. Tang and W. Bai, Green Chem.,
2014, 16, 108–111.
2,3-pentanedione were found to be acid–base pairs, which can
4 Z. Guo, D. S. Theng, K. Y. Tang, L. L. Zhang, L. Huang, A. Borgna
and C. Wang, Phys. Chem. Chem. Phys., 2016, 18, 23746–23754.
5 C. M. Tang, Z. J. Zhai, X. L. Li, L. W. Sun and W. Bai,
J. Catal., 2015, 329, 206–217.
6 C. M. Tang, J. S. Peng, X. L. Li, Z. J. Zhai, W. Bai, N. Jiang,
H. J. Gao and Y. W. Liao, Green Chem., 2015, 17, 1159–1166.
7 M. E. Sad, L. F. G. Pena, C. L. Padro and C. R. Apesteguia,
Catal. Today, 2018, 302, 203–209.
be regulated by varying the KNO loading amount. The catalyst
3
has a hierarchically porous structure, which favors the quick
diffusion of the LA reactant towards the active sites and 2,3-
pentanedione can quickly leave from the active sites due to
relatively low mass transfer resistance. As a result, KLA/TiPO
has excellent stability for at least 116 h on stream.
Conflicts of interest
18 B. Katryniok, S. Paul and F. Dumeignil, Green Chem., 2010,
2, 1910–1913.
1
There are no conflicts to declare.
19 C. M. Tang, J. S. Peng, X. L. Li, Z. J. Zhai, H. J. Gao, W. Bai,
N. Jiang and Y. W. Liao, Korean J. Chem. Eng., 2016, 33,
9
9–106.
Acknowledgements
2
0 Y. Zhang, X. Li, L. Sun and C. Tang, ChemistrySelect, 2016, 1,
The authors thank Professor Jian Liu who calculated molecular
5002–5007.
equivalent diameters for LA and 2,3-pentanedione and their 21 X. L. Li, Z. J. Zhai, C. M. Tang, L. W. Sun, Y. Zhang and
corresponding molecular polarities represented as dipole moments W. Bai, RSC Adv., 2016, 6, 62252–62262.
using Gaussian at the b3lyp/6-31g(d) level. This work was supported 22 Z. B. Huo, J. F. Xiao, D. Z. Ren, F. M. Jin, T. Wang and
by the Initial Scientific Research Foundation of Chongqing G. D. Yao, Green Chem., 2017, 19, 1308–1314.
University of Technology (2017ZD47, 2017ZD48), the Chongqing 23 T. J. Korstanje, H. Kleijn, J. Jastrzebski and R. Gebbink,
Education Commission Youth Project (KJQN201801109), the Green Chem., 2013, 15, 982–988.
Opening Foundation of Jiangsu Key Laboratory of Vehicle Emissions 24 G. Coullerez, C. Lowe, P. Pechy, H. H. Kausch and J. Hilborn,
Control (OVEC 032), and the Sichuan Province Science and J. Mater. Sci.: Mater. Med., 2000, 11, 505–510.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019
New J. Chem.

View Article Online
Paper
NJC
2
2
2
2
2
3
5 K. T. Liu, X. M. Huang, E. A. Pidko and E. J. M. Hensen, 39 J. Zhao, J. Xie, C. T. Au and S. F. Yin, RSC Adv., 2014, 4,
Green Chem., 2017, 19, 3014–3022. 6159–6164.
6 X. L. Li, Y. Zhang, Z. Chen, P. Can, W. X. Zou, C. M. Tang, L. Dong 40 J. Zhao, J. Xie, C. T. Au and S. F. Yin, Appl. Catal., A, 2013,
and Y. Wang, Ind. Eng. Chem. Res., 2017, 56, 14437–14446. 467, 33–37.
7 X. L. Li, L. W. Sun, W. X. Zou, P. Cao, Z. Chen, C. M. Tang 41 J. Xie, L. Chen, C. T. Au and S. F. Yin, Catal. Commun., 2015,
and L. Dong, ChemCatChem, 2017, 9, 4621–4627. 66, 30–33.
8 L. W. Sun, X. L. Li and C. M. Tang, Acta Phys.-Chim. Sin., 42 L. T. Bu, M. R. Nimlos, D. J. Robichaud and S. Kim, Catal.
016, 32, 2327–2336. Today, 2018, 312, 73–81.
9 M. S. Tam, J. E. Jackson and D. J. Miller, Ind. Eng. Chem. 43 P. Chen, L. Chen, Y. Zeng, F. Ding, X. Jiang, N. Liu, C. T. Au
Res., 1999, 38, 3873–3877. and S. F. Yin, Appl. Catal., B, 2018, 234, 311–317.
0 H. Bouzidi, M. Djehiche, T. Gierczak, P. Morajkar, C. Fittschen, 44 Y. Zhang, G. Y. Li, J. M. Liu, T. Wang, X. Wang, B. Liu,
2
P. Coddeville and A. Tomas, J. Phys. Chem. A, 2015, 119,
2781–12789.
Y. L. Liu, Q. S. Huo and Z. Y. Chu, J. Colloid Interface Sci.,
2018, 528, 109–115.
1
3
1 W. Schutyser, S. F. Koelewijn, M. Dusselier, S. Van de Vyver, 45 J. F. Li, H. Bai, X. S. Li, W. T. Li, J. F. Zhai, M. C. Li and
J. Thomas, F. Yu, M. J. Carbone, M. Smet, P. Van Puyvelde, G. C. Xi, J. Mater. Chem. A, 2018, 6, 12153–12161.
W. Dehaen and B. F. Sels, Green Chem., 2014, 16, 1999–2007. 46 Y. H. Zang, X. F. Dong, D. Ping, J. M. Geng and H. F. Dang,
3
3
3
3
3
3
3
2 J. F. Zhang, X. Z. Feng, Y. L. Zhao, W. J. Ji and C. T. Au, J. Ind.
Eng. Chem., 2014, 20, 1353–1358.
3 G. C. Gunter, R. Craciun, M. S. Tam, J. E. Jackson and
D. J. Miller, J. Catal., 1996, 164, 207–219.
Catal. Commun., 2018, 113, 51–54.
47 C. M. Tang, J. S. Peng, G. C. Fan, X. L. Li, X. L. Pu and W. Bai,
Catal. Commun., 2014, 43, 231–234.
48 W. J. Lo, M. Y. Shen, C. H. Yu and Y. P. Lee, J. Mol. Spectrosc.,
1997, 183, 119–128.
49 Y. Sargin, M. Kizilyalli, C. Telli and H. Guler, J. Eur. Ceram.
Soc., 1997, 17, 963–970.
50 D. Ilieva, B. Jivov, D. Kovacheva, T. Tsacheva, Y. Dimitriev,
G. Bogachev and C. Petkov, J. Non-Cryst. Solids, 2001, 293,
562–568.
4 G. C. Gunter, D. J. Miller and J. E. Jackson, J. Catal., 1994,
148, 252–260.
5 G. C. Gunter, R. H. Langford, J. E. Jackson and D. J. Miller,
Ind. Eng. Chem. Res., 1995, 34, 974–980.
6 M. S. Tam, G. C. Gunter, R. Craciun, D. J. Miller and
J. E. Jackson, Ind. Eng. Chem. Res., 1997, 36, 3505–3512.
7 D. C. Wadley, M. S. Tam, P. B. Kokitkar, J. E. Jackson and 51 T. Z. Ren, Z. Y. Yuan, A. Azioune, J. J. Pireaux and B. L. Su,
D. J. Miller, J. Catal., 1997, 165, 162–171. Langmuir, 2006, 22, 3886–3894.
8 M. S. Tam, R. Craciun, D. J. Miller and J. E. Jackson, Ind. 52 C. M. Tang, Z. J. Zhai, X. L. Li, L. W. Sun and W. Bai,
Eng. Chem. Res., 1998, 37, 2360–2366.
J. Taiwan Inst. Chem. Eng., 2016, 58, 97–106.
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019