Catalytic performance of AuIII supported on SiO2 modified activated carbon

Article abstract of DOI:10.1039/c4ra06068g

Silica was deposited onto activated carbon through TEOS hydrolysis and this composite was used as a support for the Au catalyst in acetylene hydrochlorination. Silica content and the catalyst synthesis process were both optimized. It was found that 1Au/5SiO₂/AC showed improved stability, while being as active as 1Au/AC. Thermogravimetric analysis (TGA) quantitatively revealed the synchronism of TEOS hydrolysis and Au deposition, which was believed to be the linchpin in 1Au/5SiO₂/AC synthesis. Scanning electron microscopy (SEM) and X-ray diffraction (XRD) demonstrated that amorphous silica dispersed on the carbon surface uniformly as spherical particles. Silica deposition brought down the surface area of the catalyst while leading to better distribution of gold nanoparticles. Higher gold distribution degree guaranteed the catalytic activity of 1Au/5SiO₂/AC despite surface area loss. A higher level of resistance to acetylene, excellent surface property stability and less carbonaceous deposition were determined as the origin of the improved stability of 1Au/5SiO₂/AC. This journal is the Partner Organisations 2014.

Full text of DOI:10.1039/c4ra06068g

RSC Advances
View Article Online
View Journal | View Issue
PAPER
III
Catalytic performance of Au supported on SiO
2
modified activated carbon
Cite this: RSC Adv., 2014, 4, 36316
Xiaohui Tian, Guotai Hong, Ying Liu, BinBo Jiang* and Yongrong Yang
Silica was deposited onto activated carbon through TEOS hydrolysis and this composite was used as a
support for the Au catalyst in acetylene hydrochlorination. Silica content and the catalyst synthesis
process were both optimized. It was found that 1Au/5SiO
active as 1Au/AC. Thermogravimetric analysis (TGA) quantitatively revealed the synchronism of TEOS
hydrolysis and Au deposition, which was believed to be the linchpin in 1Au/5SiO /AC synthesis. Scanning
2
/AC showed improved stability, while being as
2
electron microscopy (SEM) and X-ray diffraction (XRD) demonstrated that amorphous silica dispersed on
the carbon surface uniformly as spherical particles. Silica deposition brought down the surface area of
the catalyst while leading to better distribution of gold nanoparticles. Higher gold distribution degree
Received 21st June 2014
Accepted 7th August 2014
guaranteed the catalytic activity of 1Au/5SiO
acetylene, excellent surface property stability and less carbonaceous deposition were determined as the
origin of the improved stability of 1Au/5SiO /AC.
2
/AC despite surface area loss. A higher level of resistance to
DOI: 10.1039/c4ra06068g
www.rsc.org/advances
2
III
slow deactivation, which can be correlated to reduction of Au ,
Introduction
10
carbonaceous deposition, loss of surface area and sintering of
12
The manufacture of the vinyl chloride monomer (VCM) via Au nanoparticles. Past researches comprise two main ways in
HgCl
catalyzed acetylene hydrochlorination is an important solving this problem. Firstly, Hutchings, Conte and coworkers
2
industrial route, contributing nearly twenty percent of global have paid general attention on prolongation of the life time of
poly-vinyl chloride production and more than sixty percent of gold catalyst through regeneration of it. Unlike mercury-based
1
that in China. Nowadays, rekindled enthusiasm in coal-derived catalyst, gold based catalyst in acetylene hydrochlorination
feedstock makes this route economically more advantageous never deactivates due to run off of gold element. Only very slight
9
than ethylene oxychlorination, the major route for VCM gold loss (less than 1%) was detected by Hutchings and
13
production around the globe. Unfortunately, manipulation of Zhang, which is considered as the basis of regeneration.
the industrial catalysts is quite complicated and dangerous, as Furthermore, it was demonstrated that treatment of deactivated
the catalysts, mainly consist of mercuric chloride and activated gold catalyst by aqua regia could restore its activity without any
carbon, deactivate rapidly owing to the volatility of highly toxic gold loss, thus making multiple cycles of deactivation/
2,3
14,15
mercuric compounds.
Extensive environmental concerns regeneration possible.
In account of these ndings, Conte
arise and the quest for alternatives never ceases.
and coworkers proposed aqua regia as an effective and
Topics of gold nanostructured catalysts have augmented economical regenerating reagent. However, despite its effec-
exponentially in the last 20 years due to their potential appli- tiveness, the offline regeneration process was not preferred for
cability to various reactions of industrial and environmental application, let alone the complexity of aqua regia in operation.
4–8
III
interest. In 1990s, carbon supported Au has been identied Besides of this offline way, more attention was given to addi-
as the most active catalyst for acetylene hydrochlorination and tives, metal salts mostly, that can improve the catalytic
9–11
12,13,16,17
the investigation continues.
High price of gold makes two stability.
Gold in combination with additional metal
functionalities, the catalytic activity and stability, particularly salts sometimes resulted in higher catalytic activity as well as
important in acetylene hydrochlorination. Despite its high better stability. Among various alternate elements, it seems that
1
2
price, gold catalyst is feasible in industry since it is easy to be copper, congener of gold, is the best one. In addition, Co and
1
3
III
retrieved as other noble metals (Pd and Pt, for instance). Actu- La showed benets in Au reduction and coke deposition
ally, lifetime of the catalyst was much more crucial for appli- resistance. Generally speaking, the screening of a second metal
cation. Gold catalyst in acetylene hydrochlorination undergoes is relatively inefficient and in lack of instructive theories.
By far, activated carbon is determined as the unique support
for gold species in acetylene hydrochlorination. Although oxide
State Key Laboratory of Chemical Engineering, Department of Chemical and supports, such as SiO
2 3 2
, Al O and TiO , are playing an
2
Biochemical Engineering, Zhejiang University, Hangzhou 310027, P. R. China.
increasingly important role in modern catalysis owing to their
E-mail: jiangbb@zju.edu.cn
36316 | RSC Adv., 2014, 4, 36316–36324
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
versatile properties, they are unviable for hydrochlorination added dropwise to the activated carbon support under shaking.
because TOFs (turnover frequency) are very low aer gold being Aer that, the wet product was kept at room temperature for 2 h
9
,12
ꢁ
supported on them.
application of gold catalyst, low temperature CO oxidation, before being collected for use.
oxide supports are crucial and versatile. It is known that the SiO modied activated carbon (SiO
function of different oxide supports in CO oxidation varies. For impregnation of tetraethyl orthosilicate (TEOS) dissolved in
example, Au/SiO
exhibits very low activity and good stability; ethanol (TEOS : ethanol ¼ 1 : 15, mass ratio). The volume of the
meanwhile Au/TiO is highly active but suffers deactivation solution was approximately 3.2 times that of the total pore
By contrast, in the most attractive to gain a better impregnation, and then dried at 140 C for 18 h
2
2
/AC) was prepared by
2
2
18
upon calcination. To unite the superiorities of different oxide volume of the activated carbon. The solution was further added
materials, composite oxide supports have been nely designed dropwise to the acid-washed carbon under shaking. Aer that,
for CO oxidation catalysts, leading to a more profound under- the wet product was kept at room temperature for 1 h in vacuum
standing of gold–support interaction, and also rendering a (ꢀ95 kPa) and then dried under the same vacuum degree at
ꢁ
ꢁ
reference method for other catalytic systems as well. Plenty of 60 C for 6 h. The product was dried in air under 70 C for 12 h.
literatures have reported improved performance of noble metal During these processes, ethanol evaporated as well as TEOS
19–23
catalysts on composite supports.
This prompted us to perform a systematic study of the
hydrolyzed to SiO
2
on activated carbon.
Au/SiO /AC was prepared by using collected SiO /AC (before
2
2
function of composite supports for hydrochlorination of acety- exposure in air longer than 5 min) as support. Wetness
lene, since support is one of the most important aspects in impregnation was still the technique for gold supporting.
heterogeneous catalysis. In this contribution, we report the Preparation of SiO /Au/AC involved the preparation of Au/AC in
2
potential of composite support in acetylene hydrochlorination the rst step and SiO modication of Au/AC in the second, took
2
for the rst time. Silica modied activated carbon was used as the same procedures of activated carbon modication as
support for gold to investigate effect of carbon modication by introduced.
oxides. Properties of these catalysts were characterized by
scanning electron microscopy (SEM), nitrogen adsorption AC and Au/SiO
isotherms (BET), transmission electron microscopy (TEM), these two samples with Au/SiO
As a variation, two samples denoted as SiO
(ammonia)/AC were prepared. Difference of
/AC and SiO /Au/AC was that
2
(ammonia)/Au/
2
2
2
ꢀ1
thermal gravimetric analysis (TGA) and X-ray diffraction (XRD). ammonia liquor (0.1 mol L ) was added dropwise to SiO /AC
2
ꢁ
This work provides an example for catalyst design and may be and SiO /Au/AC aer vacuum drying under 60 C respectively.
2
helpful to broaden researchers' methodological horizons in Then, ammonia contained samples were kept at room temper-
ꢁ
hydrochlorination.
ature for 1 h before being dried at 70 C for 12 h. The volume of
ammonia liquor was equal to the total pore volume of the
support.
Experimental
Materials and reagents
Catalyst testing
HAuCl $4H O (assay 49%) and C H OH were purchased from
4
2
2
5
Hydrochlorination reaction of acetylene was carried out in a
Guoyao Chemical Reagent Company (Shanghai, China); acti-
vated carbon (marked as AC, 20–40 mesh) was obtained from
Jiangsu Yonghua Fine Chemicals Co., Ltd.; tetraethyl orthosi-
licate (TEOS) was purchased from Hangzhou Shuanglin
Chemical Reagent Factory. Hydrogen Chloride (99.998%) was
provided by Shanghai Weichuang Standard Gas Analytical
Technology Co., Ltd.; acetylene (99.5%) was purchased from
Jiaxing Tianli Gas Co., Ltd.

xed bed laboratory microreactor. Catalysts were tested using a
stainless steel reactor tube (i.d. of 8 mm). The reaction zone
consisted of 1.0 g fresh sample of catalyst. The reactor was
ꢁ
operated at atmospheric pressure, and maintained at 180 C, in
down-ow mode. N
the reactants, HCl and C
2
was used as purging gas. The pressure of
, was chosen for safety and in
2
H
2
accordance with industrial operating condition.
ꢁ
Aer being heated to 180 C, highly puried hydrogen
chloride was regulated by mass ow controllers and fed into the
reactor alone for 1 h to activate the catalyst, according to the
Catalyst preparation
The activated carbon was initially washed with dilute aqueous industrial process. Then, acetylene was introduced into
ꢀ1
ꢁ
HCl (1 mol L ) at 70 C for 5 h to remove residual alkali species, concentrated sulfuric acid to remove trace poisonous impurities
which may affect the distribution of nanoparticles on the such as acetone, moisture, S, P and As, and subsequently fed
support. This mixture was ltered, washed with distilled water into the reactor to start the reaction.
ꢁ
ꢀ1
till pH ¼ 7 and then dried at 140 C for 12 to 18 h.
A total GHSV of 530 h was chosen to determine the optimal
In all samples introduced in the following paragraphs, Au catalyst preparation procedure. Catalytic stability and the effect
content was controlled at 1% in weight, denoted as 1Au. SiO
2
2
of SiO content on catalytic performance were tested under a
ꢀ
1
content of the catalyst was denoted as xSiO (e.g. 5SiO means higher GHSV of 1060 h . Conversion of acetylene was not too
2
2
SiO mass ratio of the sample was 5%).
high in both situations. Thus all results obtained were in the
2
The activated carbon-supported gold catalyst (Au/AC) was kinetic regime. The gas product was passed through a vessel
prepared according with the incipient wetness impregnation lled with 15% sodium hydroxide solution and a drying tube in
technique. Proper amount of HAuCl aqueous solution was sequence to remove the remaining hydrogen chloride and
4
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 36316–36324 | 36317

View Article Online
RSC Advances
Paper
moisture. The composition of the effluent was determined for each catalyst is plotted against the time on stream (TOS)
immediately using a gas chromatography equipped with a ame in Fig. 1.
ionization detector. Two catalytic functionalities of the cata-
It was observed that whilst the highest catalytic activities of
lysts, conversion of acetylene (X ) and the deactivation rate 1Au/5SiO /AC and 1Au/5SiO (ammonia)/AC were both close to
A
2
2
(DR), were dened as follows:
2 2
activity of 1Au/AC, 5SiO /1Au/AC and 5SiO (ammonia)/1Au/AC
showed relatively low activities. For 5SiO
the case that modication of 1Au/AC resulted in coverage of
active species by SiO , which made them not readily available to
2
/1Au/AC, might it be
X
A
¼ (1 ꢀ 4
A
) ꢂ 100%
(1)
(2)
2
L
H
A
DR ¼ (X ꢀ X )/t
A
reactants. Additionally, ammonia liquor added during synthesis
of 5SiO (ammonia)/1Au/AC promoted the hydrolysis of TEOS,
2
A
where 4 is the volume fraction of remaining acetylene in gas
hence resulting in more SiO2 covering gold nanoparticles.
L
H
product, X
respectively during the experiment. Time span from X
denoted as t, in units of hour. Selectivity of catalyst was dened
A A A
and X refer to the highest and the last X obtained
III
Ammonia liquor could reduce Au to metallic gold in the
H
A
L
A
to X is
meantime. Synergetic effects of active site coverage and reduction
III
of Au brought the catalytic activity of 5SiO
2
(ammonia)/1Au/AC
/1Au/AC, as
13
as published reports.
down to a signicantly lower level than 5SiO
2
shown in Fig. 1. Furthermore, it seemed that the induction
period of catalyst was also related to catalyst synthesis proce-
dures. Both of 5SiO /1Au/AC and 5SiO (ammonia)/1Au/AC
Catalyst characterization
2
2
Morphology of different catalysts were obtained by scanning
electron microscopy (SEM) using a Hitachi TM-1000 scanning
electron microscope (low magnication) and a Hitachi S-4700
showed a markedly slower approach to steady-state activity,
which was not preferred, whilst 1Au/AC, 1Au/5SiO /AC and 1Au/
2
5
SiO
Another important catalytic property affected by SiO
cation was stability. According to Fig. 1, distinction of 1Au/AC,
/AC and 1Au/5SiO (ammonia)/AC was more obvious
in stability than in activity. As dened by eqn (2), deactivation
2
(ammonia)/AC reached to steady-state activity in two hours.

eld emission scanning electron microscope (high magnica-
2
modi-
tion). The samples were deposited on carbon holders and
evacuated at high vacuum before micrographs were taken.
The texture properties of the catalysts were derived from N
adsorption–desorption measurements carried out at liquid

1Au/5SiO
2
2
2
ꢀ1
rates of 1Au/AC and 1Au/5SiO (ammonia)/AC were ꢀ0.88% h
2
nitrogen temperature using an ASAP2020 instrument. Prior to
ꢀ1
and ꢀ0.84% h respectively. By contrast, 1Au/5SiO /AC deac-
2
any adsorption measurements, each sample was outgassed at
ꢀ1
tivated at ꢀ0.35% h , a signicantly slower rate in comparison
ꢁ
2
00 C for 6 h to eliminate air and vapor from the capillaries of
with the previous two samples.
the pore structures of the solids. Specic surface areas and pore
volume of the samples were calculated applying BET and T-plot
models respectively.
The synthesis process of 1Au/SiO
chosen as the best one and a question immediately emerged:
following this process, what value of SiO content was
2
/AC was accordingly
2
The crystal structures were studied with a Philips PW3050/60
vertical goniometer using Ni-ltered Cu Ka radiation (l ¼
preferred? To answer this question, further experiments were
performed and results are displayed in Fig. 2. 1Au/1SiO /AC and
˚
ꢁ
2
1
.5406 A). A proportional counter and a 0.02 step size in the 2q
1
1
Au/10SiO /AC were prepared and tested under a total GHSV of
2
ꢁ
range from 5 to 80 . The assignment of the various crystalline
ꢀ1
060 h . In comparison, 1Au/AC and 1Au/5SiO /AC were
2
phases was based on the JPDS powder diffraction le cards.
To determine SiO content and the content of carbonaceous
2
deposits on catalysts aer reaction, samples were analyzed
prepared and tested again under this higher GHSV. As can be
seen in Fig. 2, all three 1Au/xSiO /AC samples were more stable
than 1Au/AC. According to eqn (2), 1Au/AC, 1Au/1SiO /AC,
2
2
thermogravimetrically (TA Instruments, TGA2050). Analysis was
ꢁ
ꢁ
ꢁ
performed under owing air from 50 C to 750 C with a ow
3
ꢀ1
rate of 50 cm min , and a heating rate of 10 C. Differential
thermogravimetric analysis (DTGA) curves were obtained from
the TGA data by numerically differentiating the later with
respect to temperature.
Transmission electron microscopy (TEM) studies were per-
formed on Tecnai F20 instrument operating at 120 kV to study
the distribution of Au nanoparticles. The ground catalysts were
dispersed in ethanol and then supported on a copper grid.
Results and discussion
Effect of synthesis process and SiO content on catalytic
2
performance
Fig. 1 Conversion of acetylene over 1Au/AC (C), 1Au/5SiO
SiO /1Au/AC (;), 1Au/5SiO (ammonia)/AC (O) and 5SiO (ammonia)/
Au/AC (:). C flow rate was 12.5 mL min , HCl/C
¼ 1.13, GHSV
2
/AC (P),
Catalytic performance of different samples, namely 1Au/AC,
5
1
2
2
2
ꢀ1
1Au/5SiO
2 2 2
/AC, 5SiO /1Au/AC, 1Au/5SiO (ammonia)/AC and
2
H
2
2 2
H
ꢀ1
5SiO
2
(ammonia)/1Au/AC, were tested. Conversion of acetylene ¼ 530 h
.
36318 | RSC Adv., 2014, 4, 36316–36324
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
Table 1 Inert residue mass ratio of support and silica doped samples
2
Mean residue SiO Content
1
2
3
4
wt% wt%
AC
2.7
2.72 3.12 2.53 2.77
0
5
1
SiO
Au/5SiO
2
/AC
4.55 4.45 4.68 4.51 4.55
/AC 7.73 6.24 8.80 7.06 7.46
1.78
3.69
2
respectively to gain an average residue mass ratio for each
sample, as shown in Table 1 and Fig. 3. As listed in Table 1,
fresh activated carbon contained 2.77 wt% of non-combustible
materials, while it was 1.78 wt% higher for 5SiO
deposition of SiO . Moreover, the mass ratio of inert residue
grew to 7.46 wt% for 1Au/5SiO /AC, which consisted of non-
2
/AC due to
2
Fig. 2 Acetylene Conversion over 1Au/AC (C), 1Au/1SiO /AC (,),
2
1
Au/5SiO
2
/AC (P), 1Au/10SiO
2
/AC (9), C
H
2 2
flow rate was 25 mL
ꢀ
1
ꢀ1
min , HCl/C
2
H
2
¼ 1.13, GHSV ¼ 1060 h
.
2
combustible materials on activated carbon (2.77 wt%), gold
loaded on the carbon (1.0 wt%) and deposited SiO . Thus, SiO2
2
ꢀ
1
content on 1Au/5SiO /AC was calculated as 3.69 wt%, with 1.91
wt% of that formed during Au impregnation (subtract the
2
1
ꢀ
Au/5SiO
2
/AC and 1Au/10SiO
2
/AC deactivated at ꢀ1.09 h
,
III
ꢀ
1
ꢀ1
ꢀ1
0.40 h , ꢀ0.48 h and ꢀ0.57 h respectively. Length of the
mass ratio of SiO
SiO (ammonia)/AC, we believe that it might be this part of SiO
that enhanced stability of 1Au/5SiO /AC. Moreover, no signi-
cant difference on TGA and DTGA proles between 1Au/AC and
Au/5SiO /AC is observed from Fig. 3, indicating that SiO
doesn't interact with carbon strongly.
2 2
on 5SiO /AC). In comparison with 1Au/
induction period of all samples was close to each other. The
most active sample was 1Au/5SiO /AC as well 1Au/1SiO /AC the
5
2
2
2
2
2
least active one. Although more stable than 1Au/5SiO /AC, 1Au/
SiO /AC showed a much lower activity, which was not
2
1
2
1
2
2
preferred in the comparative study of this article. On the other
hand, all catalysts showed high selectivity toward VCM
>99.7%). Considering all these factors, namely, catalytic
activity, stability, selectivity and induction period, we concluded
that the optimal procedure in SiO modication is the one fol-
It is intriguing to note the advantage of this synchronism over
(
cases that SiO deposited before or aer Au impregnation. It
2
seems that SiO
tune the nal distribution of gold species, through interaction
among TEOS, SiO and aqueous solution. Probably, besides
2 4
formation during HAuCl impregnation could
2
lowed to synthesize 1Au/5SiO /AC. The rest of this contribution
was performed to gure out differences between 1Au/AC and
2
2
selection of proper modiers, of particular importance for cata-
lyst design is how depositing gold species interact with the
modiers at carbon surface. This inclination is the same as
1Au/5SiO /AC.
2
Hydrolyzing time of TEOS and amount of SiO
with gold impregnation
2
formed along
strong acid (such as aqua regia), which favors support modifying
14
(by acid) simultaneously with gold impregnation, implying the
importance of “synchronism effect” in gold catalyst exploration.
It is worth noting that large stability variation between 1Au/
SiO /AC and 1Au/5SiO (ammonia)/AC exists, which might root
5
2
2
in their difference of TEOS hydrolyzing time. Reasonably, we
Morphology of SiO
The morphology of 1Au/AC, 5SiO
characterized by scanning electron microscopy. SEM images are
2
on the catalysts and its crystal form
assumed that during the synthesis process of 5SiO /AC TEOS
2
2
/AC and 1Au/5SiO /AC was
2
was not completely hydrolyzed in absence of hydrolytic agent.
Moreover, there must be some TEOS remained on 5SiO
2
/AC
ꢁ
aer heating under 70 C because of its high boiling point
ꢁ
(
165.5 C). Addition of ammonia liquor to 5SiO
2
/AC might help
accomplish TEOS hydrolysis. Consequently, during the
synthesis process of sample 1Au/5SiO (ammonia)/AC, hardly
2
any SiO formed when gold nanoparticles deposition took place
2
in the subsequent step and weak interaction between them
happened. On the contrary, during the synthesis process of
sample 1Au/5SiO
2
4
/AC, addition of acidic HAuCl solution onto
5
SiO /AC promoted formation of SiO
2
2
simultaneously with gold
nanoparticles deposition, thus facilitating more immediate
impaction between them.
This was testied quantitatively by analysis of TGA results
through calculating mass ratio of inert (non-combustible) solid
ꢁ
residue aer burning under 750 C. Fresh activated carbon,
ꢁ
5SiO
2
/AC (immediately analyzed aer 12 h drying at 70 C) and
2
Fig. 3 TGA and DTGA profiles of AC, 1Au/AC, 5SiO /AC and 1Au/
corresponding 1Au/5SiO
2
/AC were analyzed for four times 5SiO
2
/AC.
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 36316–36324 | 36319

View Article Online
RSC Advances
Paper
presented in Fig. 4 at different magnications. Low magnica-
tion (1000ꢂ) images reveal that SiO distributes as micro-sized
2
particles on the surface of activated carbon (Fig. 4(a)). Image of
1
Au/5SiO /AC (Fig. 4(c)) clearly demonstrates that SiO particle
2
2
has a dense and uniform dispersion, which surely will alter the
property of activated carbon. In the meantime, a sparser
distribution was observed on the surface of 5SiO
2
/AC (Fig. 4(b)),
indicating a lower content of SiO . To our knowledge, TEOS
2
slowly hydrolyzed along with ethanol's evaporation during the
preparation of 5SiO /AC. Without promotion of water, acid or
2
base, hydrolysis of TEOS in 5SiO /AC was incomplete. The
2
remained TEOS hydrolyzed quickly aer addition of aqueous
solution of HAuCl
of SiO particles on the surface of 1Au/5SiO
cation (15Kꢂ) SEM micrograph (Fig. 4(d)) depicted the specic
appearance of SiO particles. Most of SiO particles immobi-
4
, hence resulting in the denser distribution
Fig. 5 XRD patterns of different samples.
2
2
/AC. High magni-

2
2
lized on the surface of activated carbon had a near-spherical
shape, with diameters of 100 nm–1 mm.
Inuence of SiO
stability of the 1Au/5SiO /AC
2
modication on carbon surface and surface
2
X-ray diffraction was conducted to determine crystal form of
SiO on the catalyst. The powder XRD patterns of AC, 1Au/AC, BET results of different samples are listed in Table 2. The total
2
2
ꢀ1
5
SiO
2
/AC, 1Au/5SiO
2
/AC, 10SiO
2
2
/AC and 1Au/10SiO /AC are surface area of activated carbon was 1180 m g . It increased
2
ꢀ1
displayed in Fig. 5.
slightly to 1201 m g aer impregnation of HAuCl , which
4
ꢁ
The broad peak in the range of 2q ¼ 15–30 in fresh activated might be resulted from bridging of HAuCl crystals in 1Au/AC.
4
carbon XRD pattern reects the amorphous structure of the Aer being modied by SiO
2
2
, the surface area of 5SiO
2
/AC
ꢀ1
support. Slight intensity growth of that broad peak can be slightly dropped to 1111 m
observed from the sample containing no SiO2 to 10% SiO2 possessed a signicantly lower surface area of 963.7 m g
g
. Meanwhile, 1Au/5SiO
2
/AC
2
ꢀ1
,
deposited sample, providing evidence for the formation of which was only 80% of 1Au/AC and 88% of 5SiO /AC. The
2
amorphous silica only as no peaks for other SiO crystal forms difference of surface area between 5SiO /AC and 1Au/5SiO /AC
2
2
2
were observed. Compared 1Au/5SiO
peak increase again testied a SiO content elevation aer Au
impregnation. In the XRD of 1Au/AC, 1Au/5SiO /AC and 1Au/
0SiO
2 2
/AC to 5SiO /AC, the slight is in coincident with results introduced above since more SiO2
III
2
formed on 1Au/5SiO /AC.
2
2
Ratio of micropore area was 46.1% for activated carbon.
Deposited gold nanoparticles on 1Au/AC reduced this ratio to
ꢁ
ꢁ
ꢁ
1
2
/AC, intensive diffraction patterns at 2q ¼ 38 , 44.8 , 65
ꢁ
and 77.5 agree well with metallic gold peaks. These peaks, 42.8%, implying micropores as slightly preferred position for
being assigned to the (111), (200), (220) and (311) planes of the anchoring of nanoparticles. Higher micropore area ratio of 1Au/
III
crystalline FCC structure of gold, imply that reduction of Au
happens in the process of catalyst preparation.
5SiO /AC (45.2%) hinted the formation of micropores on/
2
among SiO particles, or between SiO spheres and activated
2
2
carbon framework.
Beside of the difference of surface area and micropore area
of fresh samples, signicant disparity between 1Au/AC and
1
Au/5SiO
2
/AC was observed aer hydrochlorination reaction.
As shown in Table 2, surface area of 1Au/AC greatly reduced to
2
ꢀ1
1
071 m g aer 14 h on stream, about 11% lower than that
of fresh sample. Micropore area followed almost the same
trend. On the contrary, only 2.88% loss of surface area
happened within the same time for 1Au/5SiO /AC. Loss of
2
micropore area was only 2.76% in the meantime. Similar
trend was observed for pore volume, micropore volume and
average pore diameter. As is well known, the loss of surface
area is mainly because of collapse and sealing of pores of the
catalyst. This blockage isolated active sites in those pores
from reactants and made the catalyst less active. According to
results listed in Table 2, superior surface stability of 1Au/
5
SiO /AC is considered to be one of the most important
2
factors in maintaining catalytic activity compared to 1Au/AC.
This might stem from the strengthening of surface channels
by uniformly distributed SiO
2 2
Fig. 4 SEM images of AC (a), 5SiO /AC (b) and 1Au/5SiO /AC (c and d). from collapse.
2
particles, which protected pores
36320 | RSC Adv., 2014, 4, 36316–36324
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
a
2 2
Table 2 BET results of AC, 1Au/AC, 5SiO /AC, 1Au/5SiO /AC and corresponding used samples
1Au/AC
5SiO
2
/AC
2
1Au/5SiO /AC
Catalysts
AC
Fresh
Used
Fresh
Fresh
Used
Surface area (a)
1184.8
546.7
0.461
1.093
0.28
1201
513.9
0.428
1.124
0.26
1071.8
445
0.415
1.028
0.22
1111
452.2
0.407
1.039
0.23
963.7
436
0.452
0.915
0.22
935.9
424
0.453
0.903
0.22
Micropore area (a)
Ratio of micropore area
Pore volume (b)
Micropore volume (b)
Ave. pore diameter (c)
3.69
3.74
3.84
3.74
3.80
3.86
a
2
ꢀ1
3
ꢀ1
Units in the table: (a) m g , (b) cm g , and (c) nm.
1
1
Gold dispersion in catalysts and its thermal resistance ability
As aforementioned, catalytic activity of 1Au/5SiO /AC is nearly
the same with 1Au/AC while being grossly inferior in surface
area. This phenomenon can be explained by TEM observed
obvious sintering was observed, possibly because of the
difference between selected activated carbon supports.
2
Severe sintering made the statistics of particle size for spent
catalysts impossible and undependable because reliability of
TEM statistical results is dependent on the uniformity of bulk
samples. Accordingly, two experiments were performed to esti-
mate the difference of sintering between 1Au/AC and 1Au/5SiO /
2
AC. In both experiments, fresh samples of 1Au/AC and 1Au/
2
smaller nanoparticles on 1Au/5SiO /AC, as presented in Fig. 6.
No large particles, several tens of nanometers in diameter,
were observed on fresh 1Au/5SiO /AC and 1Au/AC. Nano-
2
particles on 1Au/AC are in the range between 5.0 to 17.3 nm,
with an average diameter of 11.1 nm. In the case of 1Au/5SiO2/
AC, gold particles are in the range between 2.0 and 9.3 nm,
mean size of which is 5.4 nm. Standard deviation of particle
ꢁ
5
SiO /AC were placed in an oven at 180 C under N protection
2
2
ꢀ
1
for 32 h respectively before activity testing under 1060 h . It can
be reasonably assumed that no surface property change took
place during this period on account of that the production of
activated carbon involves calcination of raw materials under
greatly higher temperature (around 800 C) for hours. Moreover,
Conte has demonstrated that simple thermal treatment at 180 C
diameter is 2.7 nm for 1Au/AC and 1.5 nm for 1Au/5SiO
2
/AC
separately. It is clear that SiO modication leads to improved
2
ꢁ
distribution of active sites and hence compensates its disad-
vantage in surface area.
ꢁ
2 2
in absence of C H /HCl reactants could not alter the valence state
Spent catalysts of 1Au/AC and 1Au/5SiO /AC aer 14 h on-
2
16
of gold. And of course, carbonaceous deposition did not happen
since no reaction took place. Accordingly, activity loss aer
thermal treatment can be solely ascribed to sintering of gold
nanoparticles. Steady-state activity of both fresh 1Au/AC and 1Au/
stream testing (Fig. 2) have also been analyzed to observe the
particle sintering. Intensive sintering of gold nanoparticles was
observed in both samples. Whilst much less number of small
nanoparticles could be discerned, particles with a diameter of
ꢃ100 nm or more were observed. Obvious sintering observed in
our samples is quite different with Conte's work, in which no
5
2
SiO /AC without thermal treatment during this test were 48.3%
and 46.7% respectively (Fig. 7(a)), slightly lower than in Fig. 2 as
catalyst for each test was newly prepared and there do exist some
labile factors during pore volume impregnation, which cannot be
eliminated totally because of its shortcoming. However, activity
disparity between 1Au/AC and 1Au/5SiO /AC changed little. Thus
2
the results were believed to be reliable. As shown in Fig. 7(b),
compared to analogous sample without 32 h thermal treatment,
highest conversion of acetylene aer thermal treatment was
2
45.6% for 1Au/AC, dropped by 5.6% relatively. For 1Au/5SiO /AC,
the conversion was about 42.7% and dropped by 9.4% relatively.
This discrepancy indicated that smaller gold nanoparticles in
1
Au/5SiO /AC that possess higher surface energy were more
2
tended to sinter through migration and coalescence, or Ostwald
ripening, than that in 1Au/AC. Obvious sintering of nanoparticles
were observed too when we tried to take TEM images of the
catalysts aer 32 h thermal treatment. In comparison, the biggest
particles observed are ꢃ60 nm in diameter, much smaller than
the biggest particles observed in spent catalysts, hinting that the
reactants play roles in sintering of nanoparticles.
It is worth noting that modication of SiO led to shorter
2
induction period of fresh 1Au/5SiO /AC. Aer 32 h thermal
2
Fig. 6 TEM images and particle size distributions of fresh 1Au/AC
treatment, induction period of 1Au/5SiO /AC was totally
2
(
2
a and b) and 1Au/5SiO /AC (c and d).
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 36316–36324 | 36321

View Article Online
RSC Advances
Paper
ꢀ1
GHSV of 1060 h . Flow rate of HCl or C
2 2 2 2
H during HCl or C H
exposure step was set as the total reactants ow rate in an
ꢀ
1
ordinary test (53 mL min ) without dilution of inactive gases.
The steady-state catalytic activity aer HCl or C H exposure was
2
2
measured and compared to that before the exposure, for
quantitative interpretation of deactivation. We show the results
in Fig. 8. It was found that before cutoff of C
2 2
H the conversion
of acetylene was 46.94% for 1Au/AC and 45.46% for 1Au/5SiO
2
/
AC. This activity was maintained aer 1 h pure HCl treatment
for both samples (46.51% and 45.23% respectively), without
obvious activation or deactivation effect being observed. By
comparison, aer being exposed in pure C H , the steady-state
2
2
conversion of acetylene declined to 25.54% for 1Au/AC, signif-
icantly dropped by 43.1% relatively. In the meantime, the
conversion of acetylene over 1Au/5SiO
exposure and only dropped by 38.0% relatively, lower than
that of 1Au/AC.
Difference between 1Au/AC and 1Au/5SiO /AC aer C H
2
2
/AC was 26.86% aer
2 2
C H
2
2
exposure showed that, to some extent, SiO2 strengthened
C H resistance of 1Au/5SiO /AC in comparison with 1Au/AC,
2
2
2
thus helping maintain its catalytic activity in hydro-
chlorination of acetylene. These results partly veried the
conclusions of Conte about effectiveness of different reac-
tants on gold catalyst for acetylene hydrochlorination, and are
helpful for catalyst design and industrial process design.
Moreover, we observed that two hours were needed to reac-
tivate each sample aer C H treatment, whereas, HCl treated
2
Fig. 7 Catalytic activity of 1Au/AC and 1Au/5SiO /AC: (a) fresh, (b) after
ꢀ1
3
2 h thermal treatment. C
2
H
2
flow rate was 25 mL min , HCl/C
H
2 2
¼
ꢀ1
1
.13, GHSV ¼ 1060 h
.
2
2
eliminated while being unchanged for 1Au/AC. This might be
related to smaller gold nanoparticles in 1Au/5SiO /AC, since
samples reached steady state immediately. This phenomenon
was not reported in Conte's work and implied acetylene
during the exposure as a reagent that inuences the catalyst
2
smaller gold nanoparticles provided more coordinatively
unsaturated atoms for reactant activation. This is an issue that
lacks in-depth mechanism understanding, which still needs
investigation.
III
not just by reducing ionic Au , which led to decreased activity
(
steady-state activity), but also by deteriorating the real active
III
sites (Au containing structure that really exists in catalysis)
to inactive ones (for example, coordinatively saturated Au
atoms), which led to re-emergence of induction period. We
envisioned this to help nd a convenient way to uncover the
intrinsic nature of gold catalysts for acetylene hydro-
chlorination in forthcoming work.
Acetylene resistance and roles of acetylene played in catalysis
III
0
It has been demonstrated that the reduction of Au to Au in
the strongly reducing conditions accompanies hydro-
chlorination is one of the most critical factor that deactivates
the catalyst. However, Conte and coworkers showed that loca-
III
tion of Au is more important than its content recently, which
III
illustrates that measurement of Au content through XPS was
ineffective in foreseeing activity and stability of the catalyst.
Therefore, a more effective way is needed to evaluate the relative
III
contribution of Au reduction to the loss of catalytic activity. As
has been testied by Conte, acetylene is the causation of
III
reduction of Au , whereas HCl plays a regeneration/activation
role. According to this, contrast experiments were carried out
to accomplish the evaluation. The catalysts, namely 1Au/AC and
1
Au/5SiO /AC, were respectively exposed to switching gas
2
circumstances in the following operating sequence:
C
H
2 2
/HCl (5 h) / HCl (1 h) / C
H
2 2
/HCl (2 h)
/
C
2
H
2
(1 h) / C
H
2 2
/HCl (4 h).
ꢁ
Temperature of catalyst bed was kept at 180 C all along,
2
Fig. 8 Control experiments over Au/AC (C) and 1Au/5SiO /AC (P)
meanwhile each step of the sequence was carried out under catalysts.
36322 | RSC Adv., 2014, 4, 36316–36324
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
revealed that amorphous SiO
2
spheres distributed on the
surface of 1Au/5SiO /AC in a very dense mode, which substan-
2
tially brought down surface area of the catalyst. However,
despite this disadvantage, we found that smaller gold nano-
particles were obtained in this catalyst, thus keeping its catalytic
activity as high as 1Au/AC. Furthermore, in comparison with
1
Au/AC, benets on acetylene resistance, carbonaceous depo-
sition resistance and especially surface area stability were
obtained aer SiO modication, and these nally improved the
2
overall stability of 1Au/5SiO /AC. In summary, modication of
2
activated carbon by SiO could bring about stability enhance-
2
ment on gold catalyst for acetylene hydrochlorination. Most
2
signicantly, 1Au/5SiO /AC showed advantages in acquisition of
2
Fig. 9 TGA profiles of fresh and used 1Au/AC and 1Au/5SiO /AC
catalysts.
smaller nanoparticles and surface area stability, indicating that
composite supports might have great potential in catalyst
exploring for acetylene hydrochlorination.
Resistance of carbonaceous deposition aer SiO
modication
2
Notes and references
Catalyst deactivation caused by site coverage and pore blockage
may also be induced by growth of the carbonaceous deposits
besides collapse of the channel. The amount of carbonaceous
materials deposited during the reaction was determined by
TGA, through subtraction of mass loss ratio between fresh and
1
China Chloro-Alkali Industry Association. The 12th Five-Year
Plan of chloro-alkali industry in China. China Chem. Ind.
News 2011, 7 (in Chinese).
2
W. Ren, L. Duan, Z. W. Zhu, W. Du, Z. Y. An, L. J. Xu,
C. Zhang, Y. Q. Zhuo and C. H. Chen, Environ. Sci.
Technol., 2014, 48, 2321–2327.
ꢁ
12,13
used catalysts in the temperature range 140–400 C.
The TGA
proles of fresh/used 1Au/AC and 1Au/5SiO /AC are showed in
2
3
4
I. Trotu ¸s , T. Zimmermann and F. Sch u¨ th, Chem. Rev., 2013,
114, 1761–1782.
A. S. Hashmi and G. J. Hutchings, Angew. Chem., Int. Ed.,
Fig. 9. In the inset gure, it can be observed that weight loss is
ꢁ
gradual between 140 and 400 C. Slight growth of mass loss rate
of used samples can be observed too, indicating the formation
of deposits. Aer subtracting the values get from corresponding
fresh samples, the amount of deposits formed over 1Au/AC is
2006, 45, 7896–7936.
5
6
G. J. Hutchings and M. Haruta, Appl. Catal., A, 2005, 291, 2–5.
S. A. C. Carabineiro, L. M. D. R. S. Martins, M. Avalos-Borja,
J. G. Buijnsters, A. J. L. Pombeiro and J. L. Figueiredo, Appl.
Catal., A, 2013, 467, 279–290.
clearly much higher (1.593 wt%) than that over 1Au/5SiO
2
/AC
(0.719 wt%). This denitely is one of the reasons for improved
stability of 1Au/5SiO /AC. The advantage in resistance of
2
7
8
9
E. G. Rodrigues, S. A. C. Carabineiro, J. J. Delgado, X. Chen,
deposits formation may originate from smaller nanoparticles
on 1Au/5SiO /AC, since molecules of unsaturated reactants and
2
products adsorbed on closely contacted active sites have the
tendency to polymerize with each other.
´
˜
M. F. R. Pereira and J. J. M. Orfao, J. Catal., 2012, 285, 83–91.
¨
Y. Onal, S. Schimpf and P. Claus, J. Catal., 2004, 223, 122–
33.
1
B. Nkosi, N. J. Coville, G. J. Hutchings, M. D. Adams, J. Friedl
and F. E. Wagner, J. Catal., 1991, 128, 366–377.
1
0 B. Nkosi, M. D. Adams, N. J. Coville and G. J. Hutchings, J.
Catal., 1991, 128, 378–386.
Conclusion
In this article we reported a composite support as a new 11 M. Conte, A. F. Carley, C. Heirene, D. J. Willock, P. Johnston,
orientation for gold catalysts applied in the acetylene hydro-
chlorination reaction. It was found that varied modication
A. A. Herzing, C. J. Kiely and G. J. Hutchings, J. Catal., 2007,
250, 231–239.
procedures of activated carbon with SiO could inuence cata- 12 H. Y. Zhang, B. Dai, X. G. Wang, W. Li, Y. Han, J. J. Gu and
2
lytic performance in signicantly different ways. Evidently, if
J. L. Zhang, Green Chem., 2013, 15, 829–836.
III
Au
impregnation took place synchronously with TEOS 13 H. Y. Zhang, B. Dai, X. G. Wang, L. L. Xu and M. Y. Zhu, J.
hydrolysis at the surface of activated carbon, good catalytic
Ind. Eng. Chem., 2012, 18, 49–54.
activity and stability can be achieved. The nding suggests that 14 M. Conte, C. J. Davies, D. J. Morgan, T. E. Davies, D. J. Elias,
besides careful selection of modiers, of particular importance
for catalyst design is how depositing gold species interact
A. F. Carley, P. Johnston and G. J. Hutchings, J. Catal., 2013,
297, 128–136.
with the modiers at carbon surface—just as strong acid 15 M. Conte, C. J. Davies, D. J. Morgan, T. E. Davies, A. F. Carley,
functions.
During the modication process, proper amount of SiO2
P. Johnston and G. J. Hutchings, Catal. Sci. Technol., 2013, 3,
128–134.
deposition was needed. Optimized catalyst, 1Au/5SiO
2
/AC, was 16 M. Conte, A. Carley, G. Attard, A. A. Herzing, C. J. Kiely and
studied thoroughly in comparison with 1Au/AC. SEM images
G. J. Hutchings, J. Catal., 2008, 257, 190–198.
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 36316–36324 | 36323

View Article Online
RSC Advances
Paper
1
1
1
2
7 S. J. Wang, B. X. Shen and Q. L. Song, Catal. Lett., 2010, 134, 21 K. Nakagawa, T. Okayama, Y. Tanimoto, K. Sotowa,
1
02–109.
8 H. G. Zhu, Z. Ma, S. H. Overbury and S. Dai, Catal. Lett., 2007,
16, 128–135.
S. Sugiyama, T. Moriga, S. Takenaka and M. Kishida, Appl.
Catal., A, 2012, 419–420, 13–21.
22 K. Qian, W. X. Huang, Z. Q. Jiang and H. X. Sun, J. Catal.,
2007, 248, 137–141.
23 R. J. H. Grisel, C. J. Weststrate, A. Goossens, M. W. J. Craj ´e ,
A. M. van der Kraan and B. E. Nieuwenhuys, Catal. Today,
2002, 72, 123–132.
1
9 S. M. Park, H. Jang, E. S. Kim, H. Han and G. Seo, Appl.
Catal., A, 2012, 427–428, 155–164.
0 Y. Tai, J. Murakami, K. Tajiri, F. Ohashi, M. Dat ´e and
S. Tsubota, Appl. Catal., A, 2004, 268, 183–187.
36324 | RSC Adv., 2014, 4, 36316–36324
This journal is © The Royal Society of Chemistry 2014