Phosphorus-modified ZSM-5 for conversion of ethanol to propylene

Article abstract of DOI:10.1016/j.apcata.2010.06.035

In this work, phosphorus-modified ZSM-5 zeolites were used to transform ethanol to propylene. The selectivity of propylene formation depended strongly on the phosphorus content in the zeolites; the highest propylene yield (32%) was observed over H-ZSM-5(80) modified with phosphorus at a P/Al molar ratio of 0.5. The enhancement of propylene selectivity with increasing phosphorus content was attributed to reduction of strong acid sites on the H-ZSM-5. Modification of the zeolite with phosphorous also improved the material's catalytic stability.

Full text of DOI:10.1016/j.apcata.2010.06.035

Applied Catalysis A: General 384 (2010) 201–205
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Phosphorus-modified ZSM-5 for conversion of ethanol to propylene
Zhaoxia Songa,b, Atsushi Takahashi , Isao Nakamura , Tadahiro Fujitani
a
a
a,∗
a
Research Institute for Innovation in Sustainable Chemistry, National Institute of Advanced Industrial Science and Technology (AIST) 16-1,
Onogawa, Tsukuba, Ibaraki 305-8569, Japan
College of Life Science, Dalian Nationalities University, Dalian 116600, China
b
a r t i c l e i n f o
a b s t r a c t
Article history:
In this work, phosphorus-modified ZSM-5 zeolites were used to transform ethanol to propylene. The
selectivity of propylene formation depended strongly on the phosphorus content in the zeolites; the
highest propylene yield (32%) was observed over H-ZSM-5(80) modified with phosphorus at a P/Al molar
ratio of 0.5. The enhancement of propylene selectivity with increasing phosphorus content was attributed
to reduction of strong acid sites on the H-ZSM-5. Modification of the zeolite with phosphorous also
improved the material’s catalytic stability.
Received 5 February 2010
Received in revised form 17 June 2010
Accepted 17 June 2010
Available online 25 June 2010
Keywords:
ZSM-5
©
2010 Elsevier B.V. All rights reserved.
Phosphorus
Ethanol conversion
Propylene
Dealumination
1
. Introduction
to the best of our knowledge, the conversion of ethanol to
propylene over phosphorus-modified ZSM-5 has not been studied
extensively to date. In this study, phosphorus-modified ZSM-
5 catalysts were prepared and were used in the conversion
of ethanol to propylene. The results show that phosphorus-
modified ZSM-5 exhibited enhanced catalytic stability relative to
unmodified zeolites and also effectively produced propylene from
ethanol.
The petroleum crisis has driven many researchers to seek new
routes for converting coal or natural gas into gasoline and other
chemicals. In particular, methanol-to-hydrocarbon, methanol-to-
gasoline (MTG), and methanol-to-olefins (MTO) reactions have
been regarded as especially attractive reactions for this purpose.
Many papers and reviews about MTO processes have been pub-
lished [1–7]. However, in view of decreasing fossil fuel resources
and increasing CO2 emissions, the production of olefins from
ethanol has attracted much attention in recent years [8–13], since
ethanol can be obtained from renewable biomass sources. At the
same time, the demand for propylene is increasing, and new pro-
cesses to yield large quantities of propylene are being sought.
Recently, we reported the conversion of ethanol to propylene
on H-ZSM-5 and ZSM-5 zeolites modified with various metals [14].
Our results suggested that moderate surface acidity was optimum
for the production of propylene. Furthermore, the catalytic stabil-
ity of the zeolites was improved by modification of the zirconium
with Zr/Al = 0.4. However, the zeolites poor hydrothermal stability
and resistance to coke formation render this process unviable for
industrial applications.
2. Experimental
2.1. Catalyst preparation
H-ZSM-5(80) and H-ZSM-5(30) were obtained by calcination of
as-received zeolites (Zeolyst, CBV3024E and CBV8014 with Si/Al2
molar ratios of 30 and 80, respectively) at 873 K in air for 4 h. The
phosphorus-modified ZSM-5 samples were prepared by an impreg-
nation method. NH -ZSM-5(80) powder (3 g) was suspended in
4
a 40 ml solution of 1 M (NH ) HPO , and the solvent was slowly
4
2
4
evaporated from the suspension in a rotary vacuum evaporator
at 323 K. Then, the residue was dried at 373 K for 5 h and cal-
cined at 873 K in air for 4 h. The molar ratios of P to Al were 0.25,
0.5, 0.7, and 1.0 (corresponding to P contents of 0.3, 0.6, 0.9, and
1.3 wt%, respectively). The obtained samples were designated as
PZ(80)-x, where x is the molar ratio of P/Al. Phosphorous-modified
ZSM-5(30) samples were prepared by means of the same method
described above, and these samples were designated as PZ(30)-x,
where x = 0.5 or 0.7 (corresponding to P contents of 1.7 and 2.3 wt%,
respectively).
Many researchers have reported that phosphorus-modified
ZSM-5 shows improved catalytic performance and hydrother-
mal stability relative to unmodified zeolites [2,15–29]. However,
∗ Corresponding author. Tel.: +81 298 61 8454; fax: +81 298 61 8172.
E-mail address: t-fujitani@aist.go.jp (T. Fujitani).
0
926-860X/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2010.06.035

2
02
Z. Song et al. / Applied Catalysis A: General 384 (2010) 201–205
Fig. 1. XRD patterns of unmodified and phosphorus-modified ZSM-5 samples.
Fig. 2. NH3-TPD profiles of unmodified and phosphorus-modified ZSM-5 samples.
2.2. Catalyst characterization
the typical MFI structure and no additional phases. It has been
reported previously that the XRD crystallinity of zeolites decreases
after phosphorus modification owing to the framework defects
caused by dealumination [25,29]. Compared to H-ZSM-5(80),
the phosphorus-modified ZSM-5(80) samples showed decreased
crystallinity. In contrast, for phosphorus-modified ZSM-5(30), no
observable reduction in crystallinity was seen.
X-ray diffraction (XRD) patterns of the samples were obtained
with a diffractometer (Rigaku, RINT2000) operated at 40 kV and
4
0 mA with Cu K␣ monochromatized radiation (ꢀ = 0.154178 nm).
NH -temperature-programmed desorption (TPD) was per-
3
formed in a flow system (BEL-CAT-32). The catalyst sample (0.1 g)
was placed in a small quartz tube and heated at 773 K in a He
−1
flow (30 ml min ) for 1 h. The sample was cooled to 373 K, and
The NH -TPD profiles of the unmodified and phosphorus-
3
NH3 adsorption was performed under a flow of 5 vol% NH /He
3
modified samples are shown in Fig. 2. All the samples showed two
desorption peaks: one centered at 473 K and the other at 673 K,
corresponding to weak and strong acid sites, respectively. With
increasing phosphorus loading, the intensity of the strong acid des-
orption peak decreased markedly, while the intensity of the weak
acid desorption peak increased. Furthermore, the desorption peaks
shifted slightly toward lower temperatures as the phosphorus con-
tent increased. These results clearly indicate that impregnation of
phosphorus in H-ZSM-5 not only decreased the concentration of
strong acid sites but also increased the concentration of weak acid
sites on the zeolite.
−
1
(
30 ml min ) for 1 h. The sample was flushed with He gas at a rate
−
1
of 30 ml min for 1 h to remove NH3 that was physically adsorbed
on the sample surface. The desorption of NH3 species was mea-
−
1
sured from 373 to 873 K at a constant heating rate of 10 K min . A
thermal conductivity detector was used to monitor NH desorption.
3
2.3. Catalytic testing
Catalytic reactions were carried out at atmospheric pressure in
a fixed-bed, continuous-flow reactor equipped with a quartz tube
inner diameter 12 mm). The samples were pelletized, crushed, and
(
then sieved to 14–22 mesh. In a typical run, 0.3 g of catalyst was
placed in the central zone of the reactor. A thermocouple reaching
into the center of the catalyst bed was used to measure the temper-
ature during the reaction. The catalyst was activated at 873 K for 1 h
in flowing N2 before each reaction. Ethanol (Wako, purity >99.5%)
3
.2. Catalytic activity
H-ZSM-5 and the phosphorus-modified ZSM-5 samples showed
00% conversion of ethanol at temperatures above 673 K. Fig. 3
1
shows the product distribution for the initial reaction over
phosphorus-modified ZSM-5(80) and ZSM-5(30) samples com-
pared with their unmodified counterparts at 823 K. The product
distribution depended strongly on the phosphorus content. For H-
ZSM-5(80) and phosphorus-modified ZSM-5(80) samples, when
the phosphorus content increased, C1–4 paraffins as well as C5+
aliphatic and aromatic compounds were significantly suppressed,
and the ethylene yield increased. The propylene yield increased
with increasing phosphorus loading to a maximum value of 32% at
P/Al = 0.5; further increase in the phosphorus loading resulted in
a decrease in the propylene yield. Samples with high phosphorus
content, such as PZ(80)-1, exhibited marked decreases in propylene
yield, with ethylene becoming the predominant product. Therefore
we concluded that the optimum molar ratio of P to Al was around
0.5. Compared with H-ZSM-5(80) samples, H-ZSM-5(30) samples,
which had a higher Si/Al2 ratio, produced more paraffins, C5+
was pumped into the vaporizer and mixed with N at a total flow
2
−1
rate of 30 ml min (PC2H5OH = 50 kPa). The products were analyzed
on-line with gas chromatographs (Shimadzu GC-14A) equipped
with a flame ionization detector and thermal conductivity detec-
®
tors. An Rt -Alumina capillary column (0.32 mm × 60 m) was used
to detect ethanol and C –C hydrocarbons. A MS-5A packed column
1
4
(
3 mm × 2 m) was used to detect N2 and H2.
3
. Results and discussion
3.1. Characterization
The XRD patterns of all the samples are shown in Fig. 1.
Phosphorus-modified ZSM-5 samples exhibited diffraction pat-
terns that were similar to those of the unmodified samples, showing

Z. Song et al. / Applied Catalysis A: General 384 (2010) 201–205
203
Fig. 3. Initial product distribution over ZSM-5 samples at 823 K. Reaction conditions:
−
1
0
.3 g catalyst; 0.1 MPa pressure; total flow rate 30 ml min , PC2 5OH = 50 kPa; time-
H
on-stream, 30 min. The category “Others” includes C5+ aliphatics and C5+ aromatics.
.
aliphatics, and C5+ aromatics. After H-ZSM-5(30) was modified with
phosphorus, the catalyst’s selectivity for light olefins increased, and
the selectivity for C1–4 paraffins, C5+ aliphatics, and C5+ aromatics
decreased. The propylene yield was 26% for PZ(30)-0.5, which was
lower than that obtained for PZ(80)-0.5. Under identical reaction
conditions, unmodified and phosphorus-modified ZSM-5 exhibited
a different product distribution: A higher selectivity toward propy-
lene was observed for the phosphorus-modified samples. As noted
Fig. 4. Effect of reaction temperature on product yields over PZ(80)-0.5. Reac-
−
1
tion conditions: 0.3 g catalyst; 0.1 MPa pressure; total flow rate 30 ml min
,
H
PC2 5OH = 50 kPa. The category “Others” includes C5+ aliphatics and C5+ aromatics.
.
the basis of the above results, we concluded that a W/F of ca.
above in the NH -TPD experiments, the phosphorus-modified sam-
3
−
1
0
.01–0.02 g ml min at 823 K was most favorable for propylene
production.
Fig. 6 shows the time course of ethanol conversion over H-
ZSM-5(80) and over phosphorus-modified ZSM-5(80) samples at
23 K. Under these reaction conditions, the propylene yield over H-
ples featured reduced acidity relative to the unmodified samples.
The strong acid sites on the zeolites are responsible for hydro-
gen transfer reactions, and thus the presence of reduced acidity
in turn decreased the conversion of olefins to paraffins. In addi-
tion to lower amounts of aromatics and long-chain hydrocarbons,
less coke formation was observed over the phosphorus-modified
samples, along with higher selectivity toward light olefins. How-
ever, excessive loading of phosphorus on the samples resulted in
greatly reduced acidity, which decreased the catalysts’ selectivity
toward propylene. These results demonstrate that acidity played an
important role in the product distribution over these catalysts, with
high selectivity toward propylene being associated with moderate
surface acidity.
Fig. 4 shows the effect of reaction temperature on prod-
uct yield over PZ(80)-0.5. The yield of ethylene increased with
increasing temperature, whereas the yields of C1–4 paraffins, C5+
aliphatics, and aromatics gradually decreased. The propylene yield
increased with temperature to a maximum value at 823 K; fur-
ther increases in the reaction temperature decreased the propylene
yield. These results indicate that the conversion of ethanol to propy-
lene is a complicated process involving oligomerization, hydrogen
transfer, isomerization, aromatization, and cracking reactions. Of
these reactions, oligomerization and the conversion of ethylene
to longer-chain hydrocarbons were favored at lower temperature
8
ZSM-5(80) was 23.4% at 30 min; however, after 6 h, the propylene
yield dropped to 4.6%, and the yield of ethylene increased rapidly
from 24.5 to 90.8%. These results show that H-ZSM-5(80) was
greatly deactivated over time, showing very poor catalytic stabil-
ity. Compared with H-ZSM-5(80), the phosphorus-modified sample
(P/Al = 0.25) exhibited improved catalytic stability. Phosphorus-
modified samples containing even greater P content, that is,
PZ(80)-0.5 and PZ(80)-0.7, showed excellent catalytic stability with
no deactivation during this reaction period. During ethanol-to-
propylene (ETP) conversion, the observed deactivation of H-ZSM-5
could have been caused by two mechanisms: (1) the deposition
(
673 K); at a higher temperature (873 K) thermal cracking was
favored, resulting in an increase in ethylene production. In other
words, at low temperatures ethylene was primarily produced by
dehydration of ethanol, whereas at higher temperatures ethylene
was a product of secondary cracking. These observed trends were
in agreement with the thermal equilibrium calculation. Therefore,
the distribution of products should have been limited by the ther-
modynamic calculation.
Fig. 5 shows the effect of reaction contact time on the yield
of products over PZ(80)-0.5 at 823 K. At low W/F, the dehydra-
tion of ethanol to ethylene was predominant owing to the short
contact time, whereas at higher W/F, the long contact time led
to the subsequent conversion of ethylene to products including
propylene, butene, paraffins, C5+ aliphatics, and C5+ aromatics. On
Fig. 5. Effect of reaction contact time on the yield of products over PZ(80)-0.5. Reac-
tion conditions: T = 823 K; 0.1 MPa pressure; PC2 5OH = 50 kPa. The category “Others”
H
includes C5+ aliphatics and C5+ aromatics.

2
04
Z. Song et al. / Applied Catalysis A: General 384 (2010) 201–205
Table 1
The catalytic performance of H-ZSM-5(80) and PZ(80)-0.5 for conversion of ethanol
to propylene after steam treatment.
Steaming time/ha
Yield of propylene/%b
H-ZSM-5(80)
PZ(80)-0.5
0
2
4
23.4
6.4
1.9
32.4
32.3
32.2
32.2
1
0
0.0
a
Steam treatment in H2O/N2 (mol/mol = 1/2) flow at 873 K.
b
The data were obtained at 30 min on stream. Reaction conditions: T = 823 K,
−
1
C2H5OH/N2 (mol/mol) = 1/1, W/F = 0.01 g ml min.
0
.5, the yield to propylene remained nearly constant (near 32%)
after the stream treatment. The results indicate that aluminum was
easily released from the framework of the unmodified ZSM-5 at
high temperature and in aqueous environments, but the presence
of phosphorus protected the catalyst from dealumination. In fact,
modification with phosphorus is a widely used method for improv-
ing the hydrothermal stability of ZSM-5 [2,15–29]. The structures
of aluminum and phosphorus in phosphorus-modified ZSM-5 have
been extensively identified by solid-state magic angle spinning
nuclear magnetic resonance (NMR). The results of NMR showed
that the enhancement of structure stability by phosphorous modifi-
cation is due to the suppression of dealumination [17,19,22,25–29].
However, controversy exists as to how phosphorus interacts with
the zeolite and whether phosphorus occupies framework positions
within the zeolite. Some researchers have proposed that Si–O–Al
bonds are broken under hydrothermal conditions while phospho-
rus atoms simultaneously occupy framework silicon positions to
form (SiO)xAl(PO)4 species [25]. Others have proposed that phos-
−x
phorus interacts with tetrahedral-framework aluminum to form
aluminum phosphates external to the framework [19,22,26–29]. In
other words, modification of zeolites with phosphorus results in the
formation of new species that appears as distorted tetrahedral alu-
minum atoms. These distorted species exhibit higher hydrothermal
resistance than that of the original framework species. Thus, the
phosphorus modification effectively stabilizes zeolite-framework
aluminum.
Fig. 6. Time course of ethanol conversion over H-ZSM-5(80) phosphorus-modified
ZSM-5 samples. Reaction conditions: 0.3 g catalyst; T = 823 K; 0.1 MPa pressure; total
flow rate 30 ml min , PC2
H
5OH = 50 kPa.
−1
of carbonaceous residue (coke) on H-ZSM-5 or (2) an irreversible
loss of activity due to dealumination of the zeolite structure. The
strong acidity of H-ZSM-5 encourages the dehydration of ethanol
to form oligomers responsible for coke formation. Coke would
already have been formed in the early stages of the ethanol con-
version reaction and would have covered the strong acid sites, thus
reducing the conversion of ethylene to propylene. However, when
the H-ZSM-5 was modified with phosphorus, the number of these
sites as well as their acidic strength decreased as discussed above.
This decrease in acidity can successfully suppress the oligomer-
ization and cyclization reactions that lead to coke formation. For
regeneration of deactivated H-ZSM-5(80), the coke in the zeolite
4. Conclusions
Impregnation of phosphorus in H-ZSM-5 decreased the number
of strong acid sites in the zeolite. Acidity played an important role
on the product distribution for ethanol conversion over unmodi-
fied and phosphorus-modified ZSM-5; the catalyst’s high selectivity
toward propylene was associated with the moderate surface acid-
ity observed on the phosphorus-modified catalysts. The optimal
propylene yield was 32%, which was observed on H-ZSM-5(80)
modified with phosphorus (P/Al = 0.5) at 823 K.
−
1
was burned off under a mixed flow of O (15 ml min ) and N2
(
2
−
1
60 ml min ) at 823 K for 1 h. The regenerated H-ZSM-5(80) was
then used for another reaction cycle, for which the yield of propy-
lene increased from 4.6% (for the deactivated zeolite) to 10.9%; the
yield of ethylene decreased from 90.8% to 79.6%. Compared to the
catalysts’ initial activity in the first reaction cycle (propylene and
ethylene yields of 23.4% and 24.6%, respectively), the activity of the
regenerated H-ZSM-5(80) was not recovered completely. Since the
catalytic activity was not fully recovered by coke combustion alone,
these results suggest that, in addition to coke deposition, dealumi-
nation in water vapor was another major cause of H-ZSM-5(80)
deactivation. To investigate the hydrothermal stability of unmodi-
fied and phosphorus-modified ZSM-5, we treated H-ZSM-5(80) and
Unmodified ZSM-5 was deactivated rapidly owing to both dea-
lumination and, to a lesser extent, coke formation. In contrast, both
coke deposition and dealumination were inhibited in the presence
of phosphorus. Thus, both propylene selectivity and catalytic sta-
bility were greatly improved by phosphorus modification of ZSM-5
zeolites.
References
[1] C.D. Chang, Catal. Rev.-Sci. Eng. 26 (1984) 323–345.
[
[
2] W.W. Kaeding, S.A. Butter, J. Catal. 61 (1980) 155–164.
3] I.M. Dahl, S. Kolboe, Catal. Lett. 20 (1993) 329–336.
PZ(80)-0.5 samples with a flow of 33% steam with N at 873 K before
2
[4] M. Stöcker, Micropor. Mesopor. Mater. 29 (1999) 3–48.
[5] M. Bjorgen, S. Svelle, F. Joensen, J. Nerlov, S. Kolboe, F. Bonino, L. Palumbo, S.
Bordiga, U. Olsbye, J. Catal. 249 (2007) 195–207.
reaction with ethanol. Table 1 shows the catalytic performance of
H-ZSM-5(80) and PZ(80)-0.5 after steam treatment. For H-ZSM-
[
[
6] T. Zhao, T. Takemoto, N. Tsubaki, Catal. Commun. 7 (2006) 647–650.
7] C. Mei, P. Wen, Z. Liu, H. Liu, Y. Wang, W. Yang, Z. Xie, W. Hua, Z. Gao, J. Catal.
258 (2008) 243–249.
5
(80), the yield of propylene decreased from 23.4 to 0% when the
catalyst was treated with steam for 10 h. In contrast, for PZ(80)-

Z. Song et al. / Applied Catalysis A: General 384 (2010) 201–205
205
[
[
8] H. Oikawa, Y. Shibata, K. Inazu, Y. Iwase, K. Murai, S. Hyodo, G. Kobayashi, T.
Baba, Appl. Catal. A: Gen. 312 (2006) 181–185.
9] M. Inaba, K. Murata, M. Saito, I. Takahara, Green Chem. 9 (2007) 638–646.
[20] M. Kojima, F. Lefebvre, Y.B. Taarit, Zeolites 12 (1992) 724–727.
[21] B. Viswanathan, A.C. Pulikottil, Catal. Lett. 22 (1993) 373–379.
[22] W. Kolodziejski, V. Fornes, A. Corma, Solid State Nucl. Magn. Reson. 2 (1993)
121–129.
[
10] M. Inaba, K. Murata, M. Saito, I. Takahara, React. Kinet. Catal. Lett. 88 (2006)
1
35–142.
[23] I. Hannus, I. Kiricsi, P. Fejes, A. Fonseca, J.B. Nagy, W.O. Parker Jr., Z. Szendi,
Zeolites 16 (1996) 142–148.
[24] P. Tynjala, T.T. Pakkanen, S. Mustamaki, J. Phys. Chem. B 102 (1998) 5280–5286.
[25] J.Q. Zhuang, D. Ma, G. Yang, Z.M. Yan, X.C. Liu, X.M. Liu, X.W. Han, X.H. Bao, P.
Xie, Z.M. Liu, J. Catal. 228 (2004) 234–242.
[26] T. Blasco, A. Corma, J. Martinez-Triguero, J. Catal. 237 (2006) 267–277.
[27] G. Caeiro, P. Magnoux, J.M. Lopes, F. Ramoa Ribeiro, S.M.C. Menezes, A.F. Costa,
H.S. Cerqueira, Appl. Catal. A: Gen. 314 (2006) 160–171.
[28] N. Xue, X. Chen, L. Nie, X. Guo, W. Ding, Y. Chen, M. Gu, Z. Xie, J. Catal. 248 (2007)
20–28.
[
[
[
[
[
[
[
11] D. Zhang, R. Wang, X. Yang, Catal. Lett. 124 (2008) 384–391.
12] K. Ramesh, L. Hui, Y. Han, A. Borgna, Catal. Commun. 10 (2009) 567–571.
13] K. Murata, M. Inaba, I. Takahara, J. Jpn. Petrol. Inst. 51 (2008) 234–239.
14] Z. Song, A. Takahashi, N. Mimura, T. Fujitani, Catal. Lett. 131 (2009) 364–369.
15] J.A. Lercher, G. Rumplmayr, Appl. Catal. 25 (1986) 215–222.
16] A. Rahman, G. Lemay, A. Adnot, S. Kaliaguine, J. Catal. 112 (1988) 453–463.
17] J. Caro, M. Bulow, D. Derewinski, J. Haber, M. Hunger, J. Karger, H. Pfeifer, W.
Storek, B. Zibrowius, J. Catal. 124 (1990) 367–375.
[
[
18] G. Seo, R. Ryoo, J. Catal. 124 (1990) 224–230.
19] G. Lischke, R. Eckelt, H. Jerschkewitz, B. Parlitz, E. Schreier, W. Storek, B. Zibrow-
ius, G. Ohlman, J. Catal. 132 (1991) 229–243.
[29] G. Zhao, J. Teng, Z. Xie, W. Jin, W. Yang, Q. Chen, Y. Tang, J. Catal. 248 (2007)
29–37.