Supported nanosized gold catalysi the influence of support morphology and reaction mechanism 4

The surface areaand morphology of the titanium oxide support influence the catalytic performance ofthe supported gold nanoparticles.. In preferential oxidation of carbon monoxideover the

Chapter 4

Catalysts -Effect of Hydrothermal Process of TiO2Support on Nano

Gold

catalysis-In this chapter, 6 kinds of TiO2, including bulk TiO2 (from Merck), nanotubular (NT)TiO2, and 4 other commercial TiO2 products, were utilized as the support for goldnanoparticles The catalytic performance and kinetics of the carbon monoxideoxidation reaction over the Au/TiO2 catalysts were investigated, and roles of variousfactors that can influence the catalytic performances were identified The surface areaand morphology of the titanium oxide support influence the catalytic performance ofthe supported gold nanoparticles The gold supported on TiO2nanotubes, which havethe largest surface area among the 6 TiO2supports, showed better catalytic activitythan the Au/TiO2catalysts Also, in the Au/TiO2 system, various preparation methodsand treatments play more important role than the support crystalline structure in terms

of the CO oxidation catalytic performance The catalysts calcined at 300oC showedmuch better catalytic activities for CO oxidation than the catalysts calcined at 200oC,and the hydrothermal processing of the TiO2oxide supports can enhance the nanogoldcatalytic activity Further studies showed that some structural factors contribute to thecatalytic performances, e.g., the hydrothermal processing of the TiO2oxide supportscan introduce more oxygen vacancies and chemically bonded OH/H2O to thesupport/catalyst, and the OH/H2O are indeed involved in the CO oxidation reaction.Both DRIFT and XPS results confirmed the existence of Aun+species in Au/TiO2

system during CO oxidation reaction More Aun+species were found in Au/TiO2(NT)

sample than in Au/TiO2 (CB) sample In preferential oxidation of carbon monoxideover the Au/TiO2 catalysts, the Au/TiO2(NT) catalyst still showed the best activity interms of the CO conversion, while the Au/TiO2(NT), Au/TiO2(CB) samples showedsimilar H2selectivity.

4.1 Introduction

The high activity of nano-sized gold catalysts supported on metal oxides for thecarbon monoxide oxidation reaction at or below room temperature has attracted greatinterest and attention from scientists worldwide Despite that various reaction routesand mechanisms have been proposed, it has been generally agreed that the size of thegold particles and the interaction between gold nanoparticles and supports are veryimportant factors that contribute to the extraordinary catalytic performance of thesupported nanogold catalysts, for example, Au particle exhibits good catalyticperformances under mild conditions only when the gold particle size is smaller than5nm; Oxide supports not only play role in supporting Au catalysts and keeping them

in well-dispersed condition, but may also modify the Au electronic structure viametal-support interaction; Moreover, the catalyst supports may participate inactivation of oxygen via adsorption at oxide vacancies Therefore the structural andelectronic properties of oxide supports are significantly correlated to the catalyticperformance of the nanogold catalysts

The oxide catalyst supports for the gold nanoparticles are usually classified into threegroups: (i) easily reducible oxides such as FeOx and CuO; (ii) less easily reducibleoxides including TiO2 and CeO2 and (iii) non-reducible oxides like Al2O3 and SiO2etc In Chapter 3, iron oxide support was selected for studying the effect ofpreparation methods on the nanogold catalysts Colloids-based (CB) method using

Lysine as capping agent is found to be more effective in deposition of gold onto thecatalyst supports as compared to Co-Precipitation (CP) and Deposition-Precipitation(DP) Therefore, in Chapter 4, we will employ the same method for gold depositionbut mainly focus on the TiO2 support, and will study how the support surface area,crystalline structure, morphology, pre-treatment conditions, presence of defects,presences of water, can affect the nanogold catalysts in low temperature oxidations.1-6Though Mayfair et al 7 believe that, unlike most of traditional catalysts, the effect ofsupport surface area might not be significant on supported gold nanoparticles Theirconclusion is based on the fact that when nanosized gold particles were supported onmetal oxide with high surface area (normally > 50 m2/g) and low loading, whichcould also result in poor catalytic performance.8-9

Titanium oxide supported nanogold system is currently one of the most investigatedsystems among all the metal oxide supported gold nanocatalysts.10-17The most widelyused preparation method of Au/TiO2 is deposition–precipitation, and this methodproduces gold particles that are quite uniform.18 The effects of different crystallinestructures of TiO2on the activity of Au catalysts have been addressed.18TiO2exists inthree main different crystalline forms: anatase, brookite, and rutile W.F Yan et al.compared supported gold nanocatalysts on anatase, brookite, rutile, and P25polymorphs of TiO2 for catalytic oxidation of CO, and concluded that brookite-supported gold catalyst sustains the highest catalytic activity But not many groupshave concentrated on researching the effect of support morphology on the catalyticperformance of Au/TiO2 catalysts, though some group had prepared Au supportedTiO2 nanotube catalyst For example, Idakiev et al have reported the usage of TiO2

nanotube supported gold nanoparticles as catalysts for low temperature water-gas shiftreaction.19

4.2 Experimental

4.2.1 Preparation of TiO 2 Nanotubes via Hydrothermal Process

Firstly, a systematic hydrothermal preparation was carried out In a typicalpreparation, 1g of commercially available bulk TiO2(Merck, anatase) was dispersed

in 100ml of 10M NaOH and the slurry was put in a screw-capped autoclavecontaining a Teflon vessel The autoclave was then placed in a furnace at 393K for 48hrs All acid washing processes were carried out with 0.1M HCl at room temperature.The TiO2-derived nanotubes were washed with copious amount of 0.1M HCl acid(2L), then pickled in fresh 500ml 0.1M HCl overnight and followed by washing withdistilled water, and finally recovered by centrifugation The acid washing wasconsidered as a proton exchange process in our experiments All the samples werethen air dried overnight at 80oC Subsequent air annealing was carried out in a well-ventilated furnace

Figure 4.1(A) refers to the starting TiO2 materials After 6-10 hrs of 10M NaOHhydrothermal treatment at 120oC (Figure 4.1(B), 4.1(C), 4.1(D)), step-like structures

and nanosheets exfoliated from bulk TiO2were observed This strongly suggests that

a sheet-folding process is involved Curling and scrolling of the nano sheets can be

observed in Figure 4.1(E) and Figure 4.1(F) (18hr) In Figure 4.1(I) and Figure

4.1(J) (24hr) of the hydrothermal process, formation of TiO2-derived nanotubes is

observable Figure 4.1(K) and Figure 4.1(L) show the multi-layered nanotubular

structure after TiO2-derived nanotubes were washed with 0.1M HCl

Figure 4.1 (A-L) TEM Observation of sodium titanate nanotube Formation.

H G

I

L J

K

4.1(A): Starting material, commercial TiO 2 sample.

4.1(B): Commercial TiO2after 10 hours of 10 M NaOH hydrothermal processing at 120oC.

4.1(C): Layered structure around TiO 2 after 10 hours of 10 M NaOH hydrothermal processing at 120oC 4.1(D): TiO2after 10 hours of 10M NaOH hydrothermal processing at 120oC layered structure peel off from the

surface.

4.1(E): TiO2after 18 hours in 10 M Na10 hours of 10 M NaOH hydrothermal process at 120 o C, curving formed 4.1(F): TiO 2 after 18 hours of 10 M NaOH hydrothermal processing at 120oC.

4.1(G): TiO2after 20 hours 10 hours of 10 M NaOH hydrothermal processing at 120 o C, end of tube.

4.1(H): TiO 2 after 20 hours of 10 M NaOH hydrothermal processing at 120oC.

4.1(I): TiO 2 after 24 hours of 10 M NaOH hydrothermal processing at 120oC.

4.1(J): TiO 2 after 24 hours of 10 M NaOH hydrothermal processing at 120oC.

4.1(K): TiO 2 after 48 hours of 10 M NaOH hydrothermal processing at 120oC.

4.1(L): TiO2after 48 hours of 10 M NaOH hydrothermal processing at 120oC.

4.2.2 Preparation of Au/TiO 2 via Colloids-Impregnation Procedure

Commercial TiO2(Merck, anatase), TiO2 nanotubes, and four other TiO2 supportsfrom ISHIHARA SANGYO KAISHA, LTD are labeled as CB, NT, MC-50, TTO-D-

1, TTO-S-1 and MC-150 respectively These six kinds of titanium oxide were thenused as support for the preparation of TiO2 supported gold nano particle samplesusing colloids-based method HAuCl4(1mM) is used as a precursor, NaBH4(0.1M) as

a reducing agent and lysine as a capping agent During the reduction period,sonication was applied The slurry was dried at 70ºC after centrifuged for four timesusing DI water.20-24

4.2.3 Evaluation of Catalytic Activity

Catalytic runs were carried out at atmospheric pressure in a continuous-flow bed quartz micro-reactor (I.D 4 mm) packed with samples and quartz wool Before

fixed-testing, the catalysts were pre-treated in situ with a flow of air (100 ml/min) for 1 h at

200 or 300oC For CO oxidation reactions, the feed gas was a mixture of 90%He +5%CO + 5%O2, introduced at a gas hourly space velocity (GHSV) of 60,000 cm3g-1h-

1

For preferential oxidation of CO in the presence of hydrogen, the feed gas was a70%H2+ 1%CO + 2%O2 mixture, introduced into the reactor at a GHSV of 60,000

cm3g-1h-1 For both reactions, the reaction products were analyzed on-line usingShimadzu GC-2010 gas chromatography equipped with a thermal conductivitydetector (TCD) The catalysts were evaluated for activity (in terms of CO conversionand CO2 productivity) in a temperature range of 20-300oC Data were collected afterthe system had stabilized for at least 15mins at every set reaction temperature TheConversion and Selectivity are calculated in terms of concentration:

CO conversion (%) =

Inlet CO concentration – Outlet CO concentration

x 100% Inlet CO concentration

O 2 selectivity (%) =

Inlet CO concentration – Outlet CO concentration

x 100%

2 x (Inlet O 2 concentration – Outlet O 2 concentration)

For kinetics study, the catalyst was diluted with SiC powder Absolute mass-specific

reaction rates were calculated using Eq (4.1) Detail calculation of CO Conversion, Selectivity and Kinetic data can refer to Chapter 3 (3.2.2 )

a Lorentz lens Before measurement, all samples were ultrasonically dispersed inethanol solvent and then dried over a carbon grid The average size of the Au particlesand its distributions were estimated by counting about 300 Au particles The Au and

Ti contents of prepared catalysts were determined by X-ray fluorescence elemental analyses on a Bruker AXS S4 Explorer The in-situ Diffusion ReflectanceInfrared Fourier Transform spectroscopy of CO adsorption study was carried out on aBio Rad FTIR 3000 MX spectrometer equipped with a reaction cell (modifiedHarricks model HV-DR2) The Au/TiO2catalyst was loaded into the DRIFT cell with1:1 weight ratio with KBr The spectra were acquired with a resolution of 4cm-1typically averaging 150 scans.The sample was purged using helium flow (20 ml/min)for 2 hrs before exposure to the reaction gas For CO adsorption experiments, 2.0%

multi-CO were used to investigate the relative surface reaction rate And as for DRIFTstudy on surface species during CO oxidation reaction, 2%CO + 2%O2(He as balance

gas) were used as the reactant gas The detailed experimental procedure implemented for CO adsorption and CO oxidation DRIFT study is presented in Table 4.1.

Table 4.1 Experimental procedure for CO adsorption and oxidation DRIFT study

Purge in He flow for at least 30 mins

remove gas phase CO and physisorbed

Purge in He flow for at least 30 mins

remove gas phase CO and physisorbed

Purge in He flow for at least 30 mins

remove gas phase CO and physisorbed

Take spectra of 2%CO adsorption Take spectra o 2% oxidation

X-ray photoelectron spectroscopy was performed on a VG ESCALAB XPS, ESCA

MK II using Mg Kα (1254.6eV) source under UHV better than 3 × 10-9 torr Thebackground contribution B (E) (obtained by the Shirley method) caused by inelasticprocesses was subtracted, while the curve-fitting was performed with a Gaussian-Lorentzian profile, and binding energies (BEs) were calculated using the Origin 7.0

The in-situ XPS experiments were performed in a UHV chamber at the SINSbeamline of the Singapore synchrotron light source (SSLS) at National University ofSingapore.25 XPS spectra were measured by a hemispherical electron energy analyzer(EA 125, Omicron NanoTechnology GmbH) The XPS experiments were done atnormal emission, and the photon energy resolution for the experiments was set toabout 0.5 eV XPS measurements were done at constant pass energy mode with

overall energy resolution The experimental procedure implemented for CO adsorption and oxidation in-situ XPS study are summarized in Table 4.2

Table 4.2 Experimental procedure for CO adsorption and oxidation in-situ XPS study

As-prepared catalyst in pretreatment chamber,

degas for 30 min then transfer to analysis chamber

As-prepared catalyst in pretreatment chamber, degas for 30 min then transfer to analysis chamber

Transfer the samples back to pre-treatment

chamber and 2% CO in He does was injected into

pretreatment chamber with the chamber pressure

at 1*10-4Torr for 10min

Transfer the samples back to pre-treatment chamber and 2% CO + 2% O2in He does was injected into pretreatment chamber with the chamber pressure at 1*10-4Torr for 10min

CO does was pumped out and sample was outgas

for 1hour then transferred back to analysis

4.3.1 Characterization of catalysts

Figure 4.2 shows the TEM micrographs of the six Au/TiO2 samples after treatment at 300oC in air for 1 hour The morphology of the oxide supports does notchange after the gold deposition.

pre-Figure 4.2 (A-F) TEM micrograph of six Au/TiO 2 samples pre-treated in air for 1 hour at 300 o C

A: Au/MC-150 ( Anatase, 286.9m2/g) ; B: Au/MC-50 ( Anatase, 210.4m2/g) C: Au/TTO-S1( Rutile 99.4m2/g); D: Au/TTO-D1( Rutile 117.3m2/g )

E : Au/TiO2(NT) (Anatase 287m2/g ); F : Au/TiO2(CB) Anatase 57m2/g)

D C

Figure 4.3 shows the size distribution of the Au nanoparticles in the six Au/TiO2

samples after pre-treated at 300oC in air for 1 hour Detailed information for XRF andBET results of TiO2supported gold samples are shown in Table 4.3.

Figure 4.3 Bar graph of six kinds of Au/TiO 2 samples

a: Au/TiO 2 (CB) b: Au/TiO 2 -TTO-S1 c: Au/TiO 2 -TTO-D1 d:Au/TiO 2 -MC50 e: Au/TiO 2 -MC150 f: Au/TiO 2 (NT)

Table 4.3 Au wt% in six kinds of Au/TiO 2 samples from XRF and their BET results

Au/MC-150 Au/MC-50 Au/TTO-S1 Au/TTO-D1 Au/TiO 2 (NT) Au/TiO 2 (CB)

XRF

(Au wt %)

As prepared

BET

m 2 /g

As prepared

The gold loadings of these Au/TiO2samples were between 1.06 and 0.73% according

to the x-ray fluorescence (XRF) results It is noticed that the six samples havedifferent BET surface area, decreasing in the order: Au/M-150 ~ Au/TiO2(NT) >Au/MC-50 > Au/TTO-D1 > Au/TTO-S1 > Au/TiO2(CB) and there are not manydifferences for the surface area of samples before and after being pre-treated

structures with the six Au/TiO2 samples The TiO2 in Au/MC-50, Au/MC-150,Au/TiO2CB and Au/TiO2NT exist in anatase phase, while the TiO2in Au/TiO2-TTO-S1 and Au/TiO2-TTO-D1 in rutile phase Note that the (110) diffraction at 25o israther weak in TiO2(NT) compared to other bulk TiO2samples

Before the reaction, all six samples were purple-pink color powder, and no changes incolor after the reactions were observed Also, no obvious changes for the Aunanoparticles were found in the TEM micrographs of these samples before and afterthe CO oxidation reaction The CO oxidation on the six TiO2 samples without Audeposition were also conducted, showing almost no catalytic activities for COoxidation at temperatures lower than 200oC when the GHSV value was set at 60,000

cm3g-1h-1

Au/TiO2-TTO-D1 Au/MC-50, Au/MC-150, Au/TiO2(CB) and Au/TiO2 (NT) afterbeing pretreated in air for 1 hour at 200oC, as a function of the reaction temperature

Figure 4.5 Conversion of CO as a function of reaction temperature over six TiO 2 samples treated at 200 o C Reaction conditions: 5%CO+5%O 2 in He, GHSV: 60,000 cm 3 g -1 h -1 .

pre-Among the 6 catalysts, Au/TiO2(NT) is the best catalyst, which converted 100% CO

to CO2 at 40oC Note that the six samples have similar Au loading and particle sizedistribution but different BET surface area The activity of the catalysts changed in

0 20

the same order of surface area: Au/MC(150) (279m2/g) > Au/MC(50) (192m2/g)>Au/TTO(D1) (106m2/g)> Au/TTO-S1 (90m2/g) > Au/CB (39m2/g) Actually the COoxidation activity on Au/TTO-S1 and Au/TTO-D1, in which the supports have rutilestructure, was basically the same as that on anatase Au/MC-50, indicating that thesupport crystalline structure does not affect the Au catalysis significantly in thisexample, disagreed with some reports in literature.28Au/TiO2(NT) is obviously betterthan Au/MC150, though both of them have very similar BET surface area, and are inthe anatase form TiO2 nanotubes are derived from commercial TiO2(CB), but theiractivities as the support of Au catalysts are remarkably different (100% conversion at

40oC vs 120oC) All these will be further investigated in the sections below

The influence of the gas hourly space velocity over the catalytic performance ofAu/TiO2(CB) sample is illustrated in Figure 4.6

Figure 4.6 Conversion of CO as a function of reaction temperature over Au/TiO 2 (CB) samples at different input gas flow rate Sample pre-treated at 200 o C Reaction conditions: 5%CO+5%O 2 in He, GHSV: 60,000 cm 3 g -1 h -1

0 50 100 150 200 250 0

20 40 60 80 100

100mlmin-1 120mlmin-1

As stated by BL Zhu, the pre-treatment temperature can also affect the catalyticactivity of gold supported samples.26 In order to study the effect of pre-treatmenttemperature on Au/TiO2samples’ for CO oxidation (reaction gas: 5%CO + 5%O2 inHe) catalytic activity, Au/TiO2 (CB) and Au/TiO2 (NT) were pre-treated in air for 1hour at 200, 300 and 400oC respectively Figure 4.7 show CO oxidation activities

over the Au/TiO2 (CB) and Au/TiO2 (NT) samples pre-treated at 200, 300 and 400oCrespectively Both these two samples showed the same catalytic activity trendaccording to pretreatment condition, i.e A300oC>A200oC(A=activity) The samples thatwere calcined at 300oC are better catalysts for CO oxidation than the samples thatwere pre-treated at 200

Figure 4.7 Conversion of CO as a function of reaction temperature over Au/TiO 2 (CB) and

Au/TiO 2 (NT) samples pre-treated at different temperature; Reaction conditions: 5%CO+5%O 2 in He, GHSV: 60,000 cm 3 g -1 h -1

0 20 40 60 80 100

Kinetics of CO Oxidation.

Table 4.4 lists the kinetic data of the Au/TiO2 catalysts for the CO oxidation(5%CO+5%O2 in He) at temperatures from 25 to 70oC All the samples were pre-treated in air for 1 hour at 300oC The activation energy of our samples is compatible

to other Au/TiO2 catalysts reported in literature.27 In particular Au/TiO2(NT) andAu/MC-150 are the most active catalysts On the basis of this study, it is inferred thatsurface modification of TiO2support may dramatically influence the catalytic activity

[mol/gcats] [TOF s-1]

Au (2.1wt %)/TiO 2 (CB)

Au (1.8wt %)/TiO 2 (NT)

Au(2.6wt%) TiO 2 TTO-S1

Au(2.3wt%) TiO 2 TTO-D1

3.5 3.1 2.7 2.5 2.9

CB CB CB CB CB CB

DP DP DP DP DP

300 300 300 300 300 300

300 300 300 300 300

23 16 28 27 18 17

27 19 18 20 27

27 27 27 27

27

a

Preparation methods: CB collide based method; DP, deposition participation;

Preferential Oxidation of Carbon Monoxide in the Presence of Hhydrogen (PROX)

function of reaction temperature over the Au/TiO2(NT) and Au/TiO2(CB) samples inthe PROX reaction Experiments were conducted under flow of 1%CO + 2%O2 +

70%H2balanced with helium at a GHSV of 6,000 cm3 g-1h-1 The samples were treated at 300oC in air for 1 hour The Au/TiO2 (NT) sample shows better COconversion than Au/TiO2(CB), but there is not much difference for these two samples

pre-in terms of O2selectivity

0 20 40 60 80 100

Au/TiO2(NT)

32 36 40 44 48 52 56 60 64 68 72 76

Figure 4.9 CO conversion and O 2 selectivity as a function of reaction temperature over Au/TiO 2 (NT) samples Reaction conditions: 1%CO+2%O 2 +70%H 2 in He, GHSV: 6,000 cm 3 g -1 h -1 Sample pre-treated at 300 o C in air for 1 hour

15 30 45 60 75 0

20 40 60

4.3.3 DRIFTs Study on CO Adsorption and CO Oxidation over Gold Supported

on TiO 2 Catalysts

DRIFTS study on CO adsorption and CO oxidation over the supported gold catalysts

were conducted, following the procedure listed in Table 4-1 Only parts of the results

are shown here for convenience of analysis and brevity

the Au/TiO2 (CB) catalyst at room temperature after exposed to a flow of CO (0.1%

or 0.2% or 2% in sequence) for 10 minutes, followed by helium purging for at least 30min The catalyst was pre-treated in air flow (20mlmin-1, 14.7 psi) for 1 hour at

300oC, then cooled down to room temperature Five minutes after the introduction of

CO, a sharp band was seen at 2102 cm-1, which has been assigned to CO adsorbed on

Au Bands were also seen in the 1700-1000 cm-1 region Five minutes after stopping

CO, the main adsorbed CO peak at 2105 cm-1was reduced in intensity as CO desorbsfrom the surface, with an accompanying shift to a higher wavenumber of 2108 cm-1atlow CO concentration (0.1% and 0.2% CO) Similar phenomenon (i.e the shifter of

CO adsorption peak to higher wavemnumber) was observed as CO concentration wasincreased (2% CO), except that the peak at 2102 cm-1shifted to 2106 cm-1 The band

at this range was universally assigned to a carbon monoxide molecule linearlyadsorbed on top of a single gold atom.28-29 From our results, we can tell that withincreased CO partial pressure, the Au0—CO band shifted to lower frequency Figure

4.11 (d) shows the DRIFT spectra recorded after soaking in the same sample in a flow

of 2%CO + 2%O2 for 10 minutes, followed by helium purge for at least 30 minutesbefore the measurements The DRIFT spectra of Au/TiO2 (CB) sample after CO

adsorption (Figure 4.11 (c)) and CO+O2 adsorption (Figure 4.11(d)) are of great

similarity except that Au/TiO2 (CB) sample after CO oxidation had greater bandsintensity in this range This result confirmed the assumption of non-competitiveadsorption, i.e the adsorption of carbon monoxide was not affected by the presence ofoxygen Though it is rather weak the IR band at 2141 cm-1 appears in the 4 spectra,and may be due to CO molecules adsorbed at Au+ A few more weak bands between

2210 and 2150 cm-1 after the introduction of CO + O2 may be assigned to CO-(O2)Au3+and CO—(HO)Ti4+, i.e the vibration of CO bonded on Ti4+ions site.30-34 Twokinds of cations can be detected on anatase surface by CO adsorption at roomtemperature: α sites and β sites were used to denote the more strongly bonding sitesand more weakly bonding sites individually CO adsorption on α sites leads to a band

-at a higher wavenumber And it is assumed th-at α sites are four-coordin-ated Ti4+ions