Supported nanosized gold catalysi the influence of support morphology and reaction mechanism 3

Chapter 3 Oxidation of Carbon Monoxide over Nanogold Catalysts supported on various iron Oxides -- Effect of Preparation Conditionson Catalytic Performances--In this chapter, four kinds

Chapter 3 Oxidation of Carbon Monoxide over Nanogold Catalysts supported on various iron Oxides Effect of Preparation Conditions

on Catalytic

Performances In this chapter, four kinds of gold/iron oxide catalysts, including AuCH, AuCM , AuCP,AuDP were prepared and used for CO oxidation reactions in the absence/presence of H2respectively It is found that the catalytic performance of the Au/iron oxide catalysts inthe CO oxidations was influenced by a number of factors such as the calcinationstemperature and the pre-reduction treatment, etc For the AuDP and AuCP samples, COconversion decreases with increase of the calcination temperature, while for the AuCMand AuCH catalysts, the influence of the calcination was less evident Also, for all thetested catalysts, a higher reduction temperature resulted in a lower CO conversion Theeffect of the reduction treatment on catalysts activity even exceeded that of thecalcination temperature On all the catalysts, the selectivity towards CO oxidation in H2-rich environment (PROX) decreased appreciably with increase of the reactiontemperature; however, the dependence of this selectivity on the pretreatment temperaturewas negligible On the Au/iron oxide system the PROX reaction occurs through a Mars-van Krevelen type reaction mechanism, which involves lattice oxygen of the iron oxideand CO and H2adsorbed on gold particles.The XRD, TPR, XPS and SIMS studies showthe presence of OH and COO groups on the AuCH samples The performance of Au/ironoxide catalysts towards PROX reaction was found to be strongly affected by catalyst

preparation and post-treatment The colloid-based method can better control Au particlesize and distribution.

3.1 Introduction

The high activity of nanoparticulate gold (Au) catalysts supported on metal oxide (ironoxide and titanium oxide in particular) at room temperature or even lower temperaturemay lead to a revolution in our traditional understanding of heterogeneous catalysis,because it is very rare for synthesized catalysts to work at ambient environment, which isusually the case for enzymes Despite that various reaction routes and mechanisms havebeen proposed, it has been generally agreed that the size of the Au particles and theinteraction between gold nanoparticles and supports are very important factors thatcontribute to the extraordinary catalytic performance of supported nano-gold catalysts

Au particle exhibits good catalytic performances under mild conditions only when thegold particle size is smaller than 5 nm.1-6 It is also reported that the optimum Au particlesize for catalyzing CO oxidation reaction should be 2-4 nm.7-8Oxide supports may alsomodify the Au electronic structure via metal-support interaction Moreover they mayparticipate in activation of oxygen via adsorption at oxide vacancies Therefore theinteraction between Au nanoparticles and metal oxide support is very complicated, andhas attracted a lot of attention Different metal oxide supports seem to interact with Aunanoparticles differently, and the mechanism of one system might not be the same as theother system Even for the same oxide support, preparation method may affect the Auparticle sizes, the oxide support morphology and structure, and the presence of impurities

in the system, all of which would possibly change the catalytic activity

In this chapter iron oxide is selected as the support of the Au catalysts, and the effect ofpreparation methods on the oxide crystalline structure and oxidation state is studiedcarefully Iron oxides are often classified as easily reducible oxides Quite clearcorrelation between the reducibility of the support and the activity was found Ausupported on iron oxides is highly active in CO oxidation, better than ZrO2 (less easilyreducible oxide) and Al2O3 (non-reducible oxide) It is reported that the catalyticperformance of the Au/iron oxide system in the CO oxidation is related both to the goldstate and the iron oxide phase 9-12 Among three different phases of iron oxides, i.e 

Fe2O3, γ-Fe2O3and Fe3O4,Haruta has shown that co-precipitated -Fe2O3 is more activethan impregnated -Fe2O3 and impregnated γ-Fe2O3 due to the smaller size of goldparticles.13-15 Šmit et al indicated that the surface –OH group plays an important role in

the CO activity over gold/ iron oxide system.16 CO may react with –OH groups formingvery reactive adsorbed formates, HCOO(ad), which can be oxidized to carbon dioxide andwater by lattice oxygen Thus, the catalyst preparation and post-treatment conditionsmay affect the catalytic activity of the gold/ iron oxide markedly due to the change in theiron oxide crystalline structure, the amount of surface OH group and oxide ion vacancies.Supported Au catalysts were generally prepared by standard methods, namelycoprecipitation (CP), deposition precipitation (DP), and colloid-based method (CB).17Coprecipitation method involves simultaneous precipitation of HAuCl4and metal nitrate

by Na2CO3 (or NH4OH) Deposition-precipitation technique requires aging of anaqueous solution of HAuCl4 at temperature 50-90oC and a fixed pH value in the range of6-10, which is selected based on the isoelectric point of metal oxide support, to enableselective deposition of Au(OH)3 only on the surface of the metal oxide, without

precipitation in the liquid phase The third method, colloid-based method, is toimpregnate oxide support with mono dispersed Au colloids stabilized by organic ligands

or capping agents The Au/oxides prepared by the above methods usually undergosubsequent drying and calcinations in air to obtain gold particles dispersed on oxides.The catalyst preparation as well as post-treatment conditions, e.g precipitationtemperature/pH value/time, aging temperature/time, calcination temperature/time etc, areimportant factors which may change Au particle size and the contact structure betweenthe Au particles and the support Providing that gold particles are small enough in theAu/iron-oxide prepared by various methods to be able to activate CO and O2, the COoxidation activity is found to remarkably depend on the iron oxide structures The COoxidation activity of different iron oxide species was in the order: ferrihydrite > hematite >magnetite

3.2 Experimental

3.2.1 Materials and catalysts preparation

Au/iron oxide catalysts were prepared by co-precipitation (CP) or precipitation (DP), using HAuCl4 (sigma-aldrich) and Fe(NO3)3·9H2O (sigma-aldrich) asprecursors In the case of the co-precipitation (CP) method, an aqueous mixture of theHAuCl4and Fe(NO3)3precursors was poured into an aqueous solution of Na2CO3(0.25M)which was maintained at 70oC under vigorous stirring (500 rpm) The precipitate waswashed, dried, and calcined in air at 110oC for 12 hrs This co-precipitation sample iscoded AuCP In the deposition-precipitation method, Au nanoparticles were deposited

deposition-on irdeposition-on oxide support by keeping the pH value of the aqueous solutideposition-on of HAuCl4at pH =

8 using 0.1M NaOH The Fe2O3support was generated, prior to the DP process, from 1.0

M Fe(NO3)3 solution Excessive amount of 1.0M NaOH solution was added to theFe(NO3)3solution drop-wisely till all the iron ions in the solution were deposited Thenthe mixed solution was thoroughly washed using DI water by centrifugation The slurryafter centrifuge was dried in 110oC oven for 48 hours The above prepared sample wasthen calcined at 500oC for 5 hour The as-prepared iron oxide was mainly presented in-

Fe2O3phase, with small amount of γ-Fe2O3phase detectable by XRD This self-preparediron oxide sample was used as the support for the AuDP catalyst (The deposition-precipitation sample is coded AuDP) Two other samples, AuCH and AuCM wereprepared using colloid-based method with assistance of the ultrasound irradiation17 Thesupport used for AuCH was commercial Fe2O3 (hematite, Sigma-Aldrich), while that forAuCM was commercial Fe3O4 (Magnetite, Sigma-Aldrech) In colloid-based method L-lysine was added as a capping agent, which has better control on gold particle sizecompared to conventional DP method used in literature HAuCl4 (1mM) was reduced byNaBH4 (0.1M) During the reduction period, colloid-based method was applied Thenano-Au particles were deposited on iron oxide supports The slurry was dried at 70ºCafter centrifuge four times using DI water As chloride ions is a poison to the catalyticreaction and may affect the activity of catalyst, the addition of capping agent andreduction agent and the followed washing procedure are able to remove almost ofchlorine in the solution

3.2.2 Evaluation of catalysts

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

the catalysts were pre-treated in situ with a flow of air (100 ml min-1) for 1 h at 200oCand 300oC respectively For CO oxidation reactions, the feed gas was a mixture of90%He + 5%CO + 5%O2, which was introduced into the reactor at a gas hourly spacevelocity (GHSV) of 60,000 cm3 g-1h-1 For preferential oxidation of CO in the presence

of hydrogen, the feed gas was a 70%H2+ 1%CO + 2%O2mixture balanced with helium,and was introduced into the reactor at a GHSV of 60,000 cm3 g-1h-1 For both reactions,the reaction products were analyzed on-line using Shimadzu GC-2010 gaschromatography equipped with a thermal conductivity detector (TCD) The catalysts

were evaluated for activity (in terms of CO

2 conversion) and CO productivity in atemperature range of 25-200oC Measurement readings were taken after the system hadbeen stabilized for at least 15mins for every designated reaction temperature TheConversion and Selectivity are calculated in terms of concentration:

CO conversion (%) = Inlet CO concentration – Outlet CO concentration x 100%

Inlet CO concentration

CO 2 conversion (%) = 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-specificreaction rates were calculated for the average concentration of each component ċi, at the

in- and outlet of the reactor; mAu, mass of Au in the reactor bed; V, total molar flow rate;

XCO, conversion of CO on the basis of CO2 formation; ċCO, concentration of CO in gas

mixture, equal to pi/p0; pi, partial pressure of reactants; p0, total pressure in the system

(3.1) The mean particle size (D) gets from (3.2).

Powder X-ray diffraction (XRD) patterns were recorded at room temperature on a Bruker

D8 Advance Diffractometer using a Cu Kα radiation source Diffraction angles were

Transition electron microscope (TEM) measurements were performed on a Tecnai TF 20

S-twin instrument with a Lorentz lens The samples were ultrasonically dispersed inethanol solvent and then were dried over a carbon grid for measurements The averagesize of Au particles and its distributions were estimated by counting about 300 Auparticles The Au and Fe contents of prepared catalysts were determined by X-ray

fluorescence multi-elemental analyses (XRF) on a Bruker AXS S4 Explorer.

Temperature programmed reduction (TPR) studies were performed in a continuous-flow

fixed-bed quartz micro-reactor (I.D 4 mm) with 50 mg of samples The catalyst was firstheated in a flow air at 200, 300 or 400 oC for 60 min After cooling to room temperature,the feed gas was switched to 5%H2/Ar After the baseline had been stabilized, thetemperature was increased to 600oC at a heating rate of 10oC/minute The amount of H2

consumed was measured as a function of temperature by means of a thermal conductivitydetector (TCD).

Thermogravimetric Analysis was conducted using TA Instruments SDT 2960

Simultaneous (DTA-TGA), under nitrogen (flow rate= 70ml/min) at a heating rate of

20

o

C/min X-ray photoelectron spectroscopy (XPS) was performed on a VG ESCALAB

XPS, ESCA MK II using Mg Kα (1254.6 eV) light source under UHV better than 3×10-9

torr The in-situ XPS experiments were performed in a UHV chamber at the SINS

beamline of the Singapore synchrotron light source (SSLS) at National University ofSingapore.18 XPS spectra were measured using a hemispherical electron energy analyzer(EA 125, Omicron NanoTechnology GmbH) The XPS experiments were done at normalemission, and the photon energy resolution for the experiments was about 0.5 eV XPSmeasurements were done at constant pass energy mode with overall energy resolution

Table 3.1 summarizes the experimental procedure for CO oxidation in-situ XPS study.

The same scan time on each sample was maintained

Table 3.1 Experimental procedure for CO oxidation in-situ XPS study

and 2%CO + 2%O2in He doses was injected into pretreatment chamber

with the chamber pressure at 1*10-4Torr for 10min



CO + O2does was pumped out and sample was outgas for 1 hour then

transferred back to analysis chamber



Scan for C1s, O1s, Fe 2p and Au4f

Time-of-flight (TOF) secondary ion mass spectrometry (SIMS) analysis was performed

on VG SIMSLAB incorporating a duoplasmatron ion gun and a VG M12-12 quadrupolemass spectrometer, in the mass range (m/z) from 1 to 800 The VGX 900 software wasused to control the experiments and analyze the data During the analysis samples wereneutralized by an electron flood gun of 500 eV energy with a maximum current of 1A

A positive bias potential of around 10 V was applied to obtain a maximum secondary ioncount Ar+ion was used for analysis with 10keV energy and 2 pA beam current to ensurethat the operation is in static SIMS mode The scan size was 300m and scan time 800s

3.3 Results and Discussions

3.3.1 Crystalline structure of various Au on iron oxide samples: XRD Characterization

The effect of catalyst preparation and post-treatment on the crystalline structure of

Au/iron oxide samples has been studied by XRD Figure 3.1 compares the XRD patterns

of the four kinds of the Au/iron-oxide (AuCH, AuCM, AuCP and AuDP) samples aftercalcination at 300◦C The AuDP, AuCH and AuCM samples show well crystallized iron

oxide structures after calcined at 300oC, while the co-precipitate AuCP sample remainspoorly crystallized It is noticed that AuCP is the only sample that does not exhibit the Aupeak at c.a 2=38o, and its α-Fe2O3peaks are weak in intensity even after a calcination at

300oC Pattern a in Figure 3.1 shows five distinct peaks which are identified as magnetite,

Fe3O4, for AuCM catalyst The AuCH (line b) and the AuDP (line d) samples containmainly -Fe2O3 (hematite) and iron oxide hydrate (γ-FeO(OH), Lepidocrocite), The Ausignal is strong for AuCH and rather weak for AuDP To investigate the effect ofcalcinations to the crystalline structure of our samples, XRD patterns before and after

calcinations are presented in Figure 3.2 and Figure 3.3 for AuCP and AuDP samples respectively As can be seen from Figure 3.2, the “as-prepared” co-precipitate AuCP

catalyst is poorly crystallized, while after calcinations at 400oC for 1 hour, four peaks

with quite low signal to noise ratio are visible and identified as the presence of α-Fe2O3(hemitate) A peak that is attributed to metal gold at ca 38o was observed at the AuCPsample after 400oC calcinations Note that the AuCP sample did not exhibit Au peakseven after calcined one hour at 300oC, the Au reflection was identified only after 400oCcalcination The FeO(OH) peak, Lepidocrocite, at ~64ois detectable after the calcinations,though it is not very strong

30 40 50 60 700

c d

a: AuCM b: AuCH c: AuCP d: AuDP

hematitemagnetitegold

Figure 3.1 X-ray powder diffraction patterns of four kinds of gold supported iron oxide samples after calcinations at 300 o C

30 35 40 45 50 55 60 65 70 0

5 10

Figure 3.2 XRD patterns of AuCP sample (a) “as-prepared” and (b) after calcine at 400°C •

Figure 3.3 displays X-ray diffraction patterns of AuDP samples (as-prepared and after

400oC calcinations for 1 hour) The “as prepared” AuDP with three broad peaks reveals

the presence of α-Fe2O3.After calcination at 400°C for 1 hour, more distinct peaks are

observed in Figure 3.3 (b) There are also peaks suggesting the presence of γ-Fe2O3(magnetite) and FeO(OH), Lepidocrocite Au peak becomes observable though it is veryweak

Figure 3.3 XRD patterns of AuDP catalyst (a) “as-prepared” and (b) after calcine at 400°C

Figure 3.4 compares the patterns of the AuCH catalyst after calcination at three different

temperatures (200oC, 300oC and 400oC) In all three AuCH samples calinced at differenttemperatures typical reflection of α-Fe2O3 hematite phase can be observed Calcination

at 300 oC and 400 oC increases its crystallinity The 300 oC and 400oC calcinations alsoresult in the appearance of γ-Fe2O3 magnetite and FeO(OH) lepidocrocite phase It isnoteworthy that all three samples of AuCH calcined at different temperature show peaksattributed to metallic gold, which remain almost the same band width while the peakintensity of iron oxide support increases as the pretreatment temperature increases Thisindicates that the size of the gold nano particles did not change with the increase ofcalcinations temperature when the calcinations temperature are equal or lower than 400o

C, which is also in good agreement with the results of transmission electron microscopy

Similar results were also reported in a paper from our group, 19 using lysine as cappingagent The AuCP and AuDP samples do not show this unique character.

hematite magnetite Lepidocrocite

Figure 3.4 XRD patterns of the AuCH sample after calcined at different pretreatment

temperature

3.3.2 Size of Au particles on various iron oxide supports: TEM characterization

Figure 3.5 shows the TEM image of the four Au/iron oxide samples calcined in air for 1

hour at 300oC The size distribution of Au nano particles for these four Au/iron samples

is summarized in Figure 3.6 in the form of bar graph. It is very clear that

co-precipitation method did not produce very small sized Au particles, and the particle sizedistribution is also poorer than other samples, particularly worse than AuCM and AuCH.

Figure 3.5 TEM micrograph for gold supported samples on four kinds of iron oxide supports

(A) TEM for the AuCM sample pre-treated in air for 1 hour at 300oC

(B) TEM for the AuCH sample pre-treated in air for 1 hour at 300oC

(C) TEM for the AuCP sample pre-treated in air for 1 hour at 300oC

(D) TEM for the AuDP sample pre-treated in air for 1 hour at 300oC

3.3.3 Specific surface area of Au/iron oxides: XRF and BET Characterizations

Detailed information obtained from XRF and BET results of these for gold/ iron oxide

samples are listed in Table 3.1 The gold wt% content of these Au/iron samples are all

around 3.0-3.9% according to the x-ray fluorescence (XRF) results It is noted that thesurface area of the four as-prepared samples is very different, decreasing in the order:AuCP > AuDP> AuCH > AuCM Nevertheless after calcination in air at 300oC for 1

B A

hour the specific surface area of the three samples AuCP, AuDP and AuCH are notdistinctly different, while BET of AuCM is ~4 times lower.

Table 3.2 Au wt% in three kinds of gold iron oxide samples from XRF, BET results of four kinds of

gold supported iron oxide samples

3.3.4 Reducibility of iron oxide supports: H 2 -TPR Characterization

The effect of catalyst heating treatment on the nature of Au/iron oxide samples wasstudied by H2-TPR As a reference, Figure 3.7 presents the H2-TPR results of pure

commercial iron oxides, FeO (curve a), Fe2O3 (curve b) and Fe3O4 (curve c) The peak

at 508oC for FeO is attributed to the reduction of FeO to Fe0.20 The peak at 387 and 620o

C can be assigned to the reduction of Fe2O3Fe3O4 and Fe3O4FeOFe

respectively.21,22

200 400 600 0

3000 6000 9000

a

b

c 387

508 620

Figure 3.8 displays TPR profiles of AuCP (left) and AuCH (right) samples calcined at

different temperatures On the AuCP sample calcined at 200oC (Figure 3.8 line a) two

main reduction peaks are at about 200 ◦C and 650 ◦C From the XRD study of the AuCP

sample (see Figure 3.2 line (a) ), it is known that the as-prepared sample is mainly

amorphous in structure Since the sample was prepared from the co-precipitation of Au

C f: AuCH calcined at 400 o

C

Figure 3.8 TPR profiles of the AuCP (left) and AuCH (right) samples at three different calcinations temperatures 200, 300 and 400 o C.

and Fe-containing precursors in aqueous solution, the hypothesis could be made that theamorphous iron species were present in FeO(OH) Therefore the TPR peaks at 220oC can

be attributed to the reduction of OH to H2O (this peak should be mainly due to thetransition from Fe2O3to Fe3O4 Au cationic ions are usually associated with OH groups.23Probably these TPR peaks in low temperature range are overlapped with that thereduction peak of Au ion to metallic Au) For the sample calcined at 400oC, it should besimply attributed to the transformation from Fe2O3to Fe3O4,

2Fe3+O - OH + 2H2 Fe3+

O + Fe2++ 3H2OThis is consistent with literature 19in which the TPR peak is located at 150-350oC Thepeak at 650oC in Figure 3.8 (left) is obviously due to the reduction of Fe3O4FeOFe

as that happens in pure Fe2O3 sample For the AuCP which was calcined at 300-400oC,the 200oC peak shifts to 250-300oC This is easy to understand by referring to the

TG/DTA profiles in Figure 3.9, in which three TG peaks correspond to the dehydration

of various OH groups at 173, 231, 285 and 330oC respectively

100 200 300 400 500 600 700 800 80

85 90 95 100

-35 -30 -25 -20 -15 -10 -5 0 5

Figure 3.9 Thermogravimetry (TG) and differential thermal analysis (DTA) of the AuCP sample

The H2-TPR profiles of the AuCH catalysts calcined at different temperatures are

presented in Figure 3.8 (right). There is extra peak at ~150oC for AuCH calcined at

200oC This may indicate that the AuCH sample contains loosely bonded OH groupswhich can be removed by hydrogen reduction at 150oC The strong low-temperaturepeak shifts from 200oC for AuCP to 250oC for AuCH sample This appears to mean thestronger bonding of OH groups in AuCH samples Actually XRD data also indicated thepresence of high concentration of OH groups in the AuCH samples even after 400oC,forming -FeO(OH) Lepidocrocite crystals The presence of OH groups on the AuCHcatalysts may play important role in CO oxidation on the catalysts

3.3.5 Catalytic study for CO oxidation over gold supported on various Iron Oxide catalysts

Figure 3.10 Conversion of CO as a function of reaction temperature over four kinds of gold/ iron oxide samples as prepared (a), calcined in air at 200 o C for 1hour (b), and calcined in air at 300 o C for 1hour (c) Reaction conditions: 5%CO+ 5%O 2 in He, GHSV: 60,000 cm 3 g -1 h -1

In Figure 3.10 the conversion of carbon monoxide to CO2 by O2 (without H2) over fourkinds of gold/ iron oxide samples is displayed as function of reaction temperature TheAuCH sample shows 100% CO conversion in the entire range of reaction temperaturesfrom RT to 80oC The AuDP sample is very active at RT, and its activity increased withincreasing temperature before it reaches 100% CO conversion, after that it experiences aslight decrease in CO conversion with increasing reaction temperature The AuCP andAuCM samples are not so active at RT, and gradually increase their CO oxidationactivity with increasing temperature Calcination can improve the activity of AuDP, butdoes not help AuCM These results show that AuCH is the best catalyst for lowtemperature CO oxidation AuDP is active but less stable

20 30 40 50 60 70 80 90 20

40 60 80

60 80 100

The reaction ratercofor the CO+O2CO2at 30oC is 3.36, 7 10, 10.1 and 11.1 (in terms

of 10-4 molco* gAu-1* S-1), respectively, on AuCM, AuCP, AuDP and AuCH catalystscalcined at 200oC for 1 hour These results are comparable with or better than thatreported in literature, i.e 4.87 * 10-4molCOg-1AuS-1 for 1 wt % Au/-Fe2O3 provided bythe World Gold Council (WGC).20The AuCH sample is the best catalyst, while AuCM isthe worst The preparation method and conditions for AuCM and AuCH are identicalexcept that-Fe2O3is used in the preparation of the AuCH sample, whereas smaller BETsurface area Fe3O4 was used as the support for the AuCM sample It can be concludedthat the preparation method, and hence the iron oxide support crystalline structure hasgreat influence on the CO oxidation activity of gold supported iron oxide samples

3.3.6 Selective oxidation of CO in H 2 over gold catalysts on Iron Oxides

20 30 40 50 60 70 80 90 100 110 120 20

30 40 50 60 70 80 90 100

Figure 3.11 exhibits the CO conversion of selective oxidation of carbon monoxide in

hydrogen over the four kinds of the as-prepared gold/iron oxide samples The trend in

sample exhibits the highest CO oxidation activity in the presence of H2 in thetemperatures between RT and 100oC The AuDP sample shows increasing activity withincreasing temperature from 25oC to 50°C But the activity of the AuDP sample dropswith increasing temperature from 50oC to 100 oC, being around 90% at 100oC TheAuCM sample is the worst among these four catalysts The CO conversion of the AuCPsample increases with increasing temperature before the reaction temperature reached

50oC, then, the CO conversion begins to drop with increasing temperature Figure 3.12

to Figure 3.16 report catalytic activity results of the AuCP , AuDP, AuCH and AuCM

samples calcined at different temperature for one hour, (i.e 200oC , 300oC and 400oC) interms of conversion of CO (full lines) and selectivity of O2towards CO oxidation (dottedlines) It is interesting to note that heating treatment normally reduces the activities ofthese catalysts in selective CO oxidation This may be related to the possible reduction ofcationic gold to metallic gold during the heating pretreatments as reported in literature.24