Adsorption Synthesis of Iron Oxide-Supported Gold Catalyst ... - MDPI

catalysts Article

Adsorption Synthesis of Iron Oxide-Supported Gold Catalyst under Self-Generated Alkaline Conditions for Efficient Elimination of Carbon Monoxide Feng Pan, Weidong Zhang, Yuxiao Ye, Yixuan Huang, Yanzhe Xu, Yufeng Yuan, Feng Wu and Jinjun Li * School of Resources and Environmental Sciences, Wuhan University, Wuhan 430079, China; [email protected] (F.P.); [email protected] (W.Z.); [email protected] (Y.Y.); [email protected] (Y.H.); [email protected] (Y.X.); [email protected] (Y.Y.); [email protected] (F.W.) * Correspondence: [email protected]; Tel.: +86-27-6877-8936

Received: 31 July 2018; Accepted: 25 August 2018; Published: 27 August 2018

Abstract: Goethite- and hematite-supported highly dispersed gold catalysts for carbon monoxide oxidation were synthesized by gold precursor adsorption onto the support materials in self-generated alkaline solutions. The support materials were prepared by reacting iron nitrate with excess sodium hydroxide. The residual minor alkali incorporated into the support could provide suitable alkaline conditions at approximately pH 8 for the hydrolysis of tetrachloroaurate anions and the subsequent adsorption process. Gold species underwent autoreduction to achieve activation during the synthesis. An increase in pH or temperature to 80 ◦ C decreased the gold loading of the catalysts. The optimal catalysts could achieve complete oxidation of carbon monoxide at −20 ◦ C. Keywords: adsorption; iron oxide; gold; carbon monoxide; catalytic oxidation

1. Introduction Since Haruta’s group discovered that gold, which is traditionally considered to be inert in catalysis, exhibits high activity in low-temperature oxidation of carbon monoxide (CO) when it is well-dispersed on metal oxides as fine particles [1], extensive research has been conducted in this area. Studies have reported that gold could catalyze a number of reactions, such as complete oxidation of air pollutants [2,3], epoxidation of propylene [4], selective oxidation of alcohols [5], synthesis of hydrogen peroxide, and C–C coupling reactions [6]. For some reactions, gold catalysts work more efficiently when their nanoparticles are smaller than 5 nm [7,8]. Therefore, much work has been conducted in nanogold synthesis. The impregnation method, which is simple and often used to produce other noble metal catalysts, appears to be unsuitable for nanogold catalyst synthesis [7]. Coprecipitation, which involves alkali-induced simultaneous precipitation of gold and support precursors, may bury gold in the support oxides [7]. Deposition–precipitation (DP) is the most popular method [9–12], which involves hydrolyzing gold precursors in a suspension that contains metal oxides and the deposition of hydrolyzed gold species on the surface of metal oxides. Iron oxide-based materials are widely used in catalysis [13,14]. Among them, hematite (α-Fe2 O3 ) is often used to support gold because of its easy reducibility, low cost and nontoxicity [7]. When preparing Fe2 O3 -supported gold catalysts by DP, some alkali is usually added to the Fe2 O3 and HAuCl4 -containing suspension to promote HAuCl4 hydrolysis and subsequent anchoring on Fe2 O3 surfaces. Some works have suggested that the pH values should be maintained at ~9–10 to achieve optimal activity [9]. In addition, a relatively higher temperature, typically 80 ◦ C, is preferred in the DP process [15,16]. In the process of Fe2 O3 synthesis, excess alkali is required to produce Fe2 O3 because a stoichiometric OH/Fe ratio of three is incapable of precipitating Fe3+ completely. Some excess alkali can become Catalysts 2018, 8, 357; doi:10.3390/catal8090357

www.mdpi.com/journal/catalysts

optimal activity [9]. In addition, a relatively higher temperature, typically 80 °C, is preferred in the DP process [15,16]. In the process of Fe2O3 synthesis, excess alkali is required to produce Fe2O3 because a Catalysts 2018, 8, 357 of 11 stoichiometric OH/Fe ratio of three is incapable of precipitating Fe3+ completely. Some excess 2alkali can become entrapped in the resultant hydroxide precipitate owing to the strong flocculating ability of hydrolyzed iron species. In this paper, we will show that residual alkali can help synthesize entrapped in the resultant hydroxide precipitate owing to the strong flocculating ability of hydrolyzed highly dispersed gold catalysts for efficient low-temperature CO oxidation. iron species. In this paper, we will show that residual alkali can help synthesize highly dispersed gold catalysts for efficient low-temperature CO oxidation. 2. Results and Discussion 2. Results and Discussion pH control is essential to obtain active gold catalysts, since species variation in tetrachloraurate ions pH in aqueous solution is to pH-dependant Acatalysts, pH increase of species aqueous HAuCl4in solution promotes control is essential obtain active [7]. gold since variation tetrachloraurate − the hydrolysis tetrachloraurate ions and[7]. formation of AuClof x(OH)4−x complexes [7]. Some chlorine ions in aqueousofsolution is pH-dependant A pH increase aqueous HAuCl4 solution promotes − atoms remain coordinated to gold atoms at pH < 7, whereas Au(OH) the[7]. dominant species the hydrolysis of tetrachloraurate ions and formation of AuClx (OH)4−4x −becomes complexes Some chlorine at pHremain > 7 [7].coordinated In general,tochlorine adsorbed noble metal is detrimental to catalytic activities − atoms gold atoms at pH 7 [7]. In general, chlorine adsorbed on noble metal is detrimental to catalytic activities [17,18]. negative effect of chlorine. Wolf and Schüth showed that a pH of ~7.8–8.8 yielded the most active Therefore, alkaline conditions are usually used to deposit gold on iron oxides to avoid the negative gold catalysts [19]. Wolf In some as-prepared goldofcatalysts washed with ammonia to effect of chlorine. andstudies, Schüth the showed that a pH ~7.8–8.8were yielded the most active gold substitute the chlorine with hydroxyl [9,20]. In this work, the pH of the synthesis system containing catalysts [19]. In some studies, the as-prepared gold catalysts were washed with ammonia to substitute ironchlorine oxide and acid increased to ~8 because the release residual the with tetrachloraurate hydroxyl [9,20]. In thisgradually work, the pH of the synthesis systemof containing ironofoxide and alkali incorporated into the support material without extra addition of any other base (Figure 1). The tetrachloraurate acid gradually increased to ~8 because of the release of residual alkali incorporated self-generated alkaline conditions were favorable for generating catalysts that were free of chlorine into the support material without extra addition of any other base (Figure 1). The self-generated [7]. alkaline conditions were favorable for generating catalysts that were free of chlorine [7].

8.0 7.5

pH

7.0 6.5 6.0 5.5 5.0 4.5 0

20

40

60

80

100

Time (min) Figure 1. 1. pH pH variation variation of of synthesis synthesis system system with with time time on on stream. stream. Figure

Nitrogen sorption sorption experiments experiments carried carried out out at at liquid liquid nitrogen nitrogen temperature temperature revealed revealed the the pore pore Nitrogen structure and and properties properties of ofthe theas-dried as-driediron ironhydroxide hydroxide(FeOOH) (FeOOH)and andthe thecalcined calcinedFe FeO 2O3. As shown structure 2 3 . As shown in Figure 2, on isotherm of FeOOH, there was a hysteresis loop that was close to type H3, does in Figure 2, on isotherm of FeOOH, there was a hysteresis loop that was close to type which H3, which not display any limiting adsorption at high P/P 0, indicating that FeOOH had narrow pore size does not display any limiting adsorption at high P/P0 , indicating that FeOOH had narrow pore size distribution and and slit-shaped slit-shaped pores pores [21]. [21]. The The pore pore size size distribution distribution curve curve showed showed that that FeOOH FeOOH had had distribution pore sizes in the range of 3–4 nm. Nevertheless, isotherm of Fe 2O3 exhibited an obviously lower N2 pore sizes in the range of 3–4 nm. Nevertheless, isotherm of Fe2 O3 exhibited an obviously lower N2 uptake, and and the 2O3 had rare pore in the mesoporous and uptake, the hysteresis hysteresis loop loopwas wassmaller. smaller.Meanwhile, Meanwhile,FeFe 2 O3 had rare pore in the mesoporous microporous scale. A summary of nitrogen sorption experiment findings is listed in Table 1. BET and microporous scale. A summary of nitrogen sorption experiment findings is listed in Table 1. surface area and total pore volume of FeOOH were significantly higher than those of Fe 2O3. In BET surface area and total pore volume of FeOOH were significantly higher than those of Fe2 O3 . In general,large largespecific specificsurface surface area was beneficial for gold adsorption and allowed the reactant to general, area was beneficial for gold adsorption and allowed the reactant to easily easily diffuse and undergo surface reactions. diffuse and undergo surface reactions.

Catalysts 2018, 8, 357

3 of 11

140 120 100 80

dV/dD (cm3/g nm)

3

Volume adsorbed (cm /g)

Catalysts 2018, 8, x FOR PEER REVIEW

3 of 11

0.03

FeOOH Fe2O3

0.02 0.01 0.00 2

60

4 6 8 Pore diameter (nm)

40

10

FeOOH Fe2O3

20 0 0.0

0.2

0.4

0.6

0.8

1.0

Relative pressure Figure 2. N2 adsorption–desorption isotherms and Barrett–Joyner–Halenda (BJH) pore size

Figure 2. N2 adsorption–desorption isotherms and Barrett–Joyner–Halenda (BJH) pore size distribution distribution curves of the support materials. curves of the support materials. Table 1. Textural properties of the support materials.

Table 1. Textural properties of the support materials.

Catalysts FeOOH Catalysts Fe2O3

Micropore Area External Surface Area Total Pore Volume BET Surface Area (m2/g) (m2/g) (m3/g) (m2/g) BET 102 Surface Area Micropore Area External Surface Total Pore Volume 16.7 85.2 0.19 2 /g) 2 /g) 2 /g) (m (m Area (m (m3 /g) 19 3.5 15.4 0.13

FeOOH 102 16.7 85.2 0.19 Fe2Both O3 FeOOH and Fe 192O3 were used as supports 3.5 15.4 for gold in this work. The nominal 0.13 loading of gold was 2.5 wt % in the synthesis, whereas the actual gold loading in Au/FeOOH and Au/Fe2O3 was 2.3 wt % and 1.4 wt %, respectively, as determined by atomic absorption spectroscopy (AAS) (Table Both FeOOH and Fe2 O3 were used as supports for gold in this work. The nominal loading of 2), suggesting FeOOH was capable of adsorbing more gold species from solution. This can be gold attributed was 2.5 wt the synthesis, whereas loading in for Au/FeOOH andspecies Au/Fe2 O3 to% theinhigher specific surface areathe of actual FeOOH,gold which allowed increased gold was 2.3 wt % and 1.4 wt %, respectively, as determined by atomic absorption spectroscopy (AAS) adsorption and a higher gold loading. In alkaline aqueous solutions, gold species exist as soluble (Tablecomplex 2), suggesting FeOOH was capable of adsorbing more gold species from solution. This anions instead of Au(OH)3 precipitant [22]. The actual reaction occurring during synthesis can could be to thethe adsorption of complex anions the support by allowed reaction for withincreased surface hydroxyl be attributed higher specific surface areaonto of FeOOH, which gold species groups;and therefore, the gold adsorption capacity of the support was critical forgold the loading gold. adsorption a higher loading. In alkaline aqueous solutions, speciesof exist asThe soluble actual gold loading on Au/FeOOH (pH 9) and Au/FeOOH (pH 10) was 2.1 wt % and 1.5 wt %, complex anions instead of Au(OH)3 precipitant [22]. The actual reaction occurring during synthesis respectively, which suggests that a further increase in pH decreased gold loading under the could be the adsorption of complex anions onto the support by reaction with surface hydroxyl groups; experimental conditions used. The isoelectric point (IEP) of iron oxides was at pH ~8–9 [23]. Above therefore, the adsorption capacity of the support was critical for the loading of gold. The actual gold the IEP, the particle surface would be negatively charged [23], and higher pH would lead to more loading on Au/FeOOH (pH 9) and Au/FeOOH (pH 10) was 2.1 wt % and 1.5 wt %, respectively, abundant surface negative charges. Negative charges on particle surfaces would disfavor adsorption whichofsuggests a further increase inrepulsion pH decreased goldwould loading under the experimental conditions − because Au(OH)4that of electrostatic [24], which result in decreased Au loading. In used.addition, The isoelectric point (IEP) of iron oxides was at pH ~8–9 [23]. Above the IEP, the particle surface the temperature of synthesis solution also had an effect on gold loading onto the support. would be negatively charged [23], and higher pH would more abundantsynthesized surface negative Au/FeOOH (80 °C) exhibited a lower gold loading of 1.9 lead wt %tothan Au/FeOOH at − because of ambient temperature. Aton high temperature, molecule wasadsorption intensified, which also disfavored charges. Negative charges particle surfaces wouldmotion disfavor of Au(OH) 4 gold adsorption onto thewhich support surface. electrostatic repulsion [24], would result in decreased Au loading. In addition, the temperature

of synthesis solution also had an effect on gold loading onto the support. Au/FeOOH (80 ◦ C) exhibited a lower gold loading of 1.9 wt % than Au/FeOOH synthesized at ambient temperature. At high temperature, molecule motion was intensified, which also disfavored gold adsorption onto the support surface.


4 of 11


4 of 11

Table 2. 2. Actual Actual loading loading and and oxidation oxidation state state of of gold gold on on various various catalysts. catalysts. Table

Catalysts AuAu Loading Catalysts Loading(%) (%) Au/FeOOH 2.3 Au/FeOOH 2.3 Au/Fe 2O3 2 O3 1.41.4 Au/Fe Au/FeOOH (80 ◦ C) Au/FeOOH (80 °C) 1.91.9 Au/FeOOH (pH 9) Au/FeOOH (pH 9) 2.12.1 Au/FeOOH (pH 10) 1.5 Au/FeOOH (pH 10) 1.5

Percentage ofAu(0) Au(0)(%) (%) Percentage Percentage of Au(I) Percentage of of Au(I) (%) (%) 55.7 44.3 55.7 44.3 00 100100 66.4% 33.6% 66.4% 33.6% 41.5 58.558.5 41.5 42.0 58.0 42.0 58.0

Figure 33 shows shows the the X-ray X-ray diffraction diffraction (XRD) (XRD) patterns patterns of of various various supported supported gold gold catalysts. catalysts. Figure FeOOH has a goethite structure (PDF No.29-0713), whereas Fe O shows a hematite structure 3 FeOOH has a goethite structure (PDF No.29-0713), whereas Fe2O3 2shows a hematite structure (PDF (PDF No.72-0469). all samples, no diffraction peaks associated with goldspecies specieswere were observed, observed, No.72-0469). In all Insamples, no diffraction peaks associated with gold although the AAS been loaded onto thethe support. TheThe absence of peaks of gold although AASindicated indicatedthat thatgold goldhad had been loaded onto support. absence of peaks of indicated that gold species were highly dispersed on the support materials as tiny clusters that were gold indicated that gold species were highly dispersed on the support materials as tiny clusters that too small to induce defined X-rayX-ray diffractions. were too small to induce defined diffractions. * : Goethite

° °

Intensity

°

°

* *

°

* * **

*

**

°

°

* *

: Hematite °°

**

Au/Fe2O3 Au/FeOOH

Au/FeOOH (pH 10) Au/FeOOH (pH 9) Au/FeOOH (80°C)

10

20

30

40

50

60

70

80

2 theta (degree) Figure Figure 3. 3. X-ray X-ray diffraction diffraction patterns patterns of of gold gold catalysts. catalysts.

In the transmission electron microscopy (TEM) images of Au/FeOOH and Au/Fe2O3 (Figure 4), In the transmission electron microscopy (TEM) images of Au/FeOOH and Au/Fe2 O3 (Figure 4), the supports consisting of rod-like particles with lattice fringes of goethite and hematite could be the supports consisting of rod-like particles with lattice fringes of goethite and hematite could be identified, and little gold particles were well dispersed in the support. However, gold particles in the identified, and little gold particles were well dispersed in the support. However, gold particles in high-resolution TEM images were seldom found. As we know, due to the resolution limit of the high-resolution TEM images were seldom found. As we know, due to the resolution limit of conventional high-resolution electron microscopy, particles less than 1–2 nm may not be detected. conventional high-resolution electron microscopy, particles less than 1–2 nm may not be detected. The invisibility of the supported gold in the high-resolution TEM images may indicate that the gold The invisibility of the supported gold in the high-resolution TEM images may indicate that the gold species were dispersed as tiny clusters, probably in the subnanometer scale, which is in agreement species were dispersed as tiny clusters, probably in the subnanometer scale, which is in agreement with the XRD patterns. Figure 5 shows the scanning transmission electron microscopy (STEM) with the XRD patterns. Figure 5 shows the scanning transmission electron microscopy (STEM) images images of Au/FeOOH and Au/Fe2O3. It suggests that the gold species were homogeneously of Au/FeOOH and Au/Fe2 O3 . It suggests that the gold species were homogeneously distributed distributed across the supports; however, it was still hard to calculate the gold particles sizes from across the supports; however, it was still hard to calculate the gold particles sizes from these images. these images.

Catalysts 2018, 8, x FOR PEER REVIEW Catalysts 2018, 8, 357 Catalysts 2018, 8, x FOR PEER REVIEW

5 of 11 5 of 11 5 of 11

Figure 4. Transmission electron microscopy images of Au/FeOOH (a,b) and Au/Fe2O3 (c,d).

Figure O33 (c,d). (c,d). Figure4.4.Transmission Transmissionelectron electronmicroscopy microscopyimages imagesofofAu/FeOOH Au/FeOOH(a,b) (a,b)and andAu/Fe Au/Fe22O

Figure 5. Scanning transmission electron microscopy images of Au/FeOOH (a) and Au/Fe2O3 (b).

Figure 6 displays XP spectra of the gold catalysts. The deconvolution of the Au 4f spectra of Figure microscopy ofofAu/FeOOH (a) (b). Figure5.5.Scanning Scanning transmission electron microscopyimages images Au/FeOOH (a)and and Au/Fe22O 33 (b). Au/FeOOH revealedtransmission the presenceelectron of two components; the binding energy of Au 4f 7/2 Au/Fe around 83.8 eV was attributed to metallic gold, while the other at around 84.5 eV was assigned to Au(I) [9,25]. In the Figure displays XP spectra spectra of the gold gold catalysts. deconvolution 4fwhich spectra of of Au 4f spectra of Au/Fe 2O3, the of binding energy of Au The 4f7/2 clearly centered of at the ~84.7Au eV, Figure 66 displays XP the catalysts. 4f spectra indicated that the Au atoms in this sample were in the form of Au(I) [9]. The binding energy of the Au/FeOOH revealed the the presence presence of of two two components; components;the thebinding bindingenergy energyofofAu Au4f4f7/2 7/2 around 83.8 83.8 eV eV Au/FeOOH Au 4f7/2 line for Au(III) should bethe around 86.6 eV [9].84.5 Although Au(III) wastopresent in the gold wasattributed attributed metallic gold, while the other around 84.5 assigned to Au(I) [9,25]. In Au the was totometallic gold, while other at at around eVeV waswas assigned Au(I) [9,25]. In the precursor, itofwas not present the prepared catalysts, as 4f indicated by the XP spectra. It has beenwhich Auspectra 4f spectra Au/Fe 2O3binding , thein binding 7/2 centered clearly centered at ~84.7 4f of Au/Fe energyenergy of Au 4fof7/2Au clearly at ~84.7 eV, whicheV, indicated 2 O3 , the reported that gold in the reduced state is an active component in catalytic reactions [25,26]. In some indicated the in Authis atoms in this sample in Au(I) the form [9]. energy The binding ofline the that the Authat atoms sample were in the were form of [9]. of TheAu(I) binding of the energy Au 4f7/2 literature regarding the preparation of Au/Fe2O3 catalysts, when HAuCl4 was used as gold Au 4f 7/2 line for Au(III) should be around 86.6 eV [9]. Although Au(III) was present in the gold for Au(III) should around 86.6 eV [9].underwent Although Au(III) present in the gold temperatures precursor, it was precursor, the be as-prepared catalysts thermal was decomposition at mild precursor, it was not present in the prepared catalysts, as indicated by the XP spectra. It has been not present in the prepared catalysts, as indicated by the XP spectra. It has been reported that gold in (generally 300–400 °C) or reduction treatment using reducing agents to obtain reduced gold [6,7,26]. reported that gold in the reduced state is an active component in catalytic reactions [25,26]. In some the reduced state is an active component in catalytic reactions [25,26]. In some literature regarding the Such treatments appear to be necessary to activate the catalysts by forming metallic or partially literature regarding the preparation Au/Fe 2Oas-prepared 3 used catalysts, when HAuCl was as gold oxidized [6,25]. in thisof work, samples underwent noneused of these preparation ofgold Au/Fe catalysts, when HAuCl was as gold precursor, the4 as-prepared catalysts 2O 3 However, 4the

precursor, thermal the as-prepared catalysts underwent thermal decomposition mild temperatures underwent decomposition at mild temperatures (generally 300–400 ◦ C) at or reduction treatment (generally 300–400 °C) or treatment reducing to obtain reduced gold [6,7,26]. using reducing agents to reduction obtain reduced goldusing [6,7,26]. Such agents treatments appear to be necessary to Such treatments appear to be necessary to activate the catalysts by forming metallic or partially activate the catalysts by forming metallic or partially oxidized gold [6,25]. However, in this work, oxidized gold [6,25]. However, in this work, the as-prepared samples underwent none of these


6 of 11

treatments. This indicates that autoreduction of gold species occurred during synthesis. Figure 1 shows that during the adsorption synthesis, the solution pH increased to a maximum value of ca. 8.2 at 20 min and then decreased gradually to a plateau of ca. 8.0, which indicated that some hydronium ions were produced in the process. We propose that after the Au(III) hydroxo complex (such as6 of 11 Catalysts 2018, 8, 357 Au(OH)4−) was adsorbed onto the support materials by inner-sphere surface complexation, electron transfer occurred from the oxygen in the hydroxo ligand to the central Au(III) atom, which caused the reduction of Au(III) along with the generation of gaseous This oxygen and thethat release of hydronium the as-prepared samples underwent none of these treatments. indicates autoreduction of gold ions into the during solution. In addition, a previous revealed that very small gold particles species occurred synthesis. Figure 1 showswork that has during the adsorption synthesis, the solution supportedtoona TiO 2 could be readily reduced. Near-complete reduction of tiny gold nanoparticles pH increased maximum value of ca. 8.2 at 20 min and then decreased gradually to a plateau of (5 nm) occurred [27]. after the Au(III) hydroxo complex (such as Au(OH)4− ) was adsorbed onto the support materials by The XP spectra of Au/FeOOH (80 °C), Au/FeOOH (pH 9), and Au/FeOOH (pH 10) were also inner-sphere surface complexation, electron transfer occurred from the oxygen in the hydroxo ligand recorded and are shown in Figure S1. The deconvolution of the Au 4f spectra of these catalysts to thesuggests central that Au(III) atom, caused thepresent. reduction Au(III) along withand theAu(I) generation of gaseous Au(0) andwhich Au(I) were both The of percentage of Au(0) are shown in oxygen and2.the release of hydronium ions intosynthesis the solution. In addition, a previous work revealed Table It demonstrates that the adsorption at 80 °C produced more metallic goldhas than at that very small gold particles supported TiO2 couldfavored be readily reduced.ofNear-complete reduction ambient temperature, suggesting higheron temperature the reduction gold. In contrast, the adsorption synthesis at (5 nm) occurred [27].

Intensity (a.u.)

Au 4f

Au(I)

Au(0) Au/FeOOH

Au 4f5/2

Au 4f7/2 Au/Fe2O3

90

88

86

84

82

80

Binding Energy (eV) Figure6.6.Au Au4f 4f XP XP spectra spectra of Figure ofthe thecatalysts. catalysts.

Gold-catalyzed low-temperature oxidation of CO has attracted significant attention because of

The XP spectra of Au/FeOOH (80 ◦ C), Au/FeOOH (pH 9), and Au/FeOOH (pH 10) were also its potential applications in air purification, safety masks, gas sensors, closed-cycle CO2 lasers, and recorded and[28]. are In shown in Figure S1. The deconvolution thecatalysts Au 4f spectra of thesewere catalysts fuel cells this work, the catalytic activities of preparedof gold for CO oxidation suggests that Au(0) and Au(I) were both present. The percentage of Au(0) and Au(I) are shown in tested. It should be mentioned that both FeOOH and Fe2O3 did not shown any activity for CO ◦ C produced more metallic gold than Tableoxidation 2. It demonstrates that the adsorption synthesis at 80 below 100 °C (Figure S2). The conversion curves of supported gold catalysts are shown in Figure temperature, 7. Au/FeOOH suggesting could fully higher convert temperature CO into carbon dioxide ca. −20of°C to ambient at ambient favored thefrom reduction gold. In contrast, temperature.synthesis CO couldatbepH oxidized at ambient temperature on at Au/Fe 2O3, whereasalkaline its the adsorption 9 and completely pH 10 produced more Au(I) than self-generated low-temperature activity was lower and theofconversion −20 °C was 64%. FeOOH has a much conditions, suggesting further enhancement alkalinityatinhibited theca. reduction of gold. higher specific surface area than Fe2O3, which ensured better gold dispersion and more active sites Gold-catalyzed low-temperature oxidation of CO has attracted significant attention because of its on the former. Au/FeOOH (80 °C), Au/FeOOH (pH 9), and Au/FeOOH (pH 10) appeared to be potential applications in air purification, safety masks, gas sensors, closed-cycle CO2 lasers, and fuel inferior to Au/FeOOH, which could be correlated with their relatively lower gold loading. The XRD cells [28]. In this work, the catalytic activities of prepared gold catalysts for CO oxidation were tested. patterns of the spent Au/FeOOH and Au/Fe2O3 catalysts (Figure S3) still did not show any It should be mentioned that both FeOOH Fe2 O3 did activitywell for CO oxidation diffraction peak associated with gold, and suggesting that not goldshown speciesany remained dispersed afterbelow 100 ◦ C (Figure S2). The conversion curves of supported gold catalysts are shown in Figure 7. Au/FeOOH catalytic reactions. The Au 4f spectra of the spent catalysts (Figure S4) were similar to the fresh ones, could fully convert CO into carbon dioxide from ca. −20 ◦ C to ambient temperature. CO could be oxidized completely at ambient temperature on Au/Fe2 O3 , whereas its low-temperature activity was lower and the conversion at −20 ◦ C was ca. 64%. FeOOH has a much higher specific surface area than Fe2 O3 , which ensured better gold dispersion and more active sites on the former. Au/FeOOH (80 ◦ C), Au/FeOOH (pH 9), and Au/FeOOH (pH 10) appeared to be inferior to Au/FeOOH, which could be correlated with their relatively lower gold loading. The XRD patterns of the spent Au/FeOOH and Au/Fe2 O3 catalysts (Figure S3) still did not show any diffraction peak associated with gold, suggesting that gold species remained well dispersed after catalytic reactions. The Au 4f spectra of the spent catalysts (Figure S4) were similar to the fresh ones, suggesting the oxidation state of gold were stable during

Catalysts 2018, 8, 3578, x FOR PEER REVIEW Catalysts 2018,

7 of 117 of 11

suggesting the oxidation state of gold were stable during the reactions. It is meaningful that the

the reactions. It is meaningful thatheating the catalyst prepared without extra heatingwas under self-generated catalyst prepared without extra under self-generated alkaline conditions the most active alkaline conditions the most active as its synthesis procedure is the simplest. as its synthesis was procedure is the simplest.

CO conversion (%)

100 80 60

Au/FeOOH Au/Fe2O3

40

Au/FeOOH (pH 9) Au/FeOOH (pH 10) Au/FeOOH (80 °C)

20 0

-20

-10

0

10

20

Temperature (°C) Figure Conversioncurves curves of of CO Figure 7. 7. Conversion CO on ongold goldcatalysts. catalysts.

There is continuing interest in the nature of the active sites of nanogold catalysts, and no

There is continuing interest in What the nature of ascertained the active is sites nanogold catalysts, consensus has yet been reached. has been thatofthe gold catalyst activityand is no consensus has yet been reached. What been ascertained is that gold good catalyst activity size-dependent, and the gold particle sizehas should be controlled within 5 nmthe to ensure activity [7,8,29]. A low-coordination of nanogold is the be most popular explanation for to its ensure excellentgood is size-dependent, and the goldstate particle size should controlled within 5 nm catalytic performance [30,31]. Because atoms at the solid surface have fewer neighbors than in activity [7,8,29]. A low-coordination state of nanogold is the most popular explanation for those its excellent the performance bulk, they have unsaturated bonds and bindhave foreign atoms and molecules morein the catalytic [30,31]. Because atoms at can the therefore solid surface fewer neighbors than those tightly. A small particle size implies a high concentration of exposed surface atoms in bulk, they have unsaturated bonds and can therefore bind foreign atoms and molecules more tightly. low-coordination state. Therefore, nanogold particles can adsorb CO and oxygen molecules readily A small particle size implies a high concentration of exposed surface atoms in low-coordination state. [32,33], which facilitates oxidation reaction. Some other researchers have emphasized the special role Therefore, nanogold particles can adsorb CO and oxygen molecules readily [32,33], which facilitates of the Au/oxide interface [2,24,34]. Oxygen vacancies are present at the Au/oxide interface, which oxidation reaction. Some other researchers have emphasized the also special of the Au/oxide can activate oxygen molecules for CO oxidation [35]. There are somerole other postulated interface [2,24,34]. Oxygen vacancies are present at the Au/oxide interface, which can activate mechanisms, including charge donation from the support, strain effects, and metal–insulator oxygen molecules foraCO oxidation [35].[36]. There aremechanisms also some other postulated mechanisms, including transition below certain cluster size These need not be mutually exclusive, and it likelyfrom that the activity inherently A smallertransition gold particle sizeaiscertain desirable chargeseems donation thecatalytic support, strainiseffects, andcomplex. metal–insulator below cluster regardless the mechanism. Recent experimental theoretical have suggested thatcatalytic the size [36]. Theseofmechanisms need not be mutually and exclusive, andstudies it seems likely that the fraction of low-coordinated Au atoms reaches ~90% as the cluster size decreases to 0.5 nm, and the activity is inherently complex. A smaller gold particle size is desirable regardless of the mechanism. efficiency of CO oxidation continues to increase as the gold particle size decreases to subnanometer Recent experimental and theoretical studies have suggested that the fraction of low-coordinated Au level [37]. In this work, the absence of gold peaks in the XRD patterns and the invisibility of gold in atoms reaches ~90% as the cluster size decreases to 0.5 nm, and the efficiency of CO oxidation continues the TEM images suggest that the gold particles may exist in the subnanometer scale, which could to increase as the gold particle size decreases to subnanometer level [37]. In this work, the absence of contribute to their perfect activity. gold peaks in the patterns and the invisibility of gold in the imagesasuggest that Since theXRD durability of gold catalyst is of great importance in itsTEM application, durability testthe forgold particles exist in the subnanometer scale, which could contribute to their perfect CO may catalytic oxidation on Au/FeOOH at ambient temperature was conducted (Figure 8).activity. During the Since of gold is ofofgreat importance in itswhich application, durability test for CO entirethe 315durability h test period, 100%catalyst conversion CO was maintained, indicatesa that the long-term stability of the on catalyst is excellent. catalytic oxidation Au/FeOOH at ambient temperature was conducted (Figure 8). During the entire 315 h test period, 100% conversion of CO was maintained, which indicates that the long-term stability of the catalyst is excellent.


8 of 11

CO conversion (%)


8 of 11

100.0

99.5

99.0

0

50

100

150

200

250

300

Time on stream (h) Figure 8. Conversion CO overAu/FeOOH Au/FeOOH at temperature with timetime on stream. Figure 8. Conversion of of CO over atambient ambient temperature with on stream.

3. Experimental

3. Experimental

3.1. Chemicals

3.1. Chemicals

Ferric nitrate nonahydrate, sodium hydroxide and chloroauric acid were purchased from Ferric nitrate nonahydrate, sodium Limited, hydroxide and chloroauric acid were Sinopharm Chemical Reagent Company Shanghai, China. All reagents were purchased of analytical from Sinopharm Chemical Reagent Company Limited, Shanghai, China. All reagents were of analytical grade and were used as received.

grade and were used as received. 3.2. Catalyst Preparation

3.2. Catalyst Preparation In a typical synthesis of support materials, 200 mL of 0.2 mol/L ferric nitrate was mixed with 100 of 1.6 synthesis mol/L sodium hydroxide (with a OH/Fe molar of 4). The resultant precipitate was with In amL typical of support materials, 200 mL of ratio 0.2 mol/L ferric nitrate was mixed separated by centrifugation and washed with deionized water. Washing did not allow for the was 100 mL of 1.6 mol/L sodium hydroxide (with a OH/Fe molar ratio of 4). The resultant precipitate complete removal of the incorporated sodium hydroxide and further washing would induce separated by centrifugation and washed with deionized water. Washing did not allow for the complete extensive dispersion of the ferric hydroxide, which formed a stable colloid that was difficult to removal of the incorporated sodium hydroxide and further washing would induce extensive dispersion precipitate by centrifugation. The collected solid was dried at 100 °C in air to yield FeOOH. Some of of thethe ferric hydroxide, whichatformed a stable that wasFedifficult to precipitate by centrifugation. hydroxide was heated 500 °C for 2 h tocolloid form iron oxide 2O3. FeOOH and Fe2O3 were used as ◦ The collected solid was at 100 loading. C in air to yield FeOOH. Some of the hydroxide was heated at support materials for dried gold catalyst 500 ◦ C forIn2ahtypical to form iron oxide Fe2 O3 . gold FeOOH and 0.5 Fe2gOof3 FeOOH were used materials for gold synthesis of supported catalyst, or Feas 2Osupport 3 was added into 50 mL of aqueous catalyst loading.HAuCl4 solution containing 12.5 mg of gold and then shaken on a platform shaker at ambient temperature h. The solid was separated with deionized In a typical synthesisfor of21 supported gold catalyst, 0.5 by g ofcentrifugation, FeOOH or Fewashed added into 50 mL 2 O3 was water, and dried atsolution 100 °C tocontaining yield the final catalyst, which was labeled Au/FeOOH orplatform Au/Fe2O3shaker . of aqueous HAuCl 12.5 mg of gold and then shaken on a at 4 For comparison, we also synthesized FeOOH-supported gold catalyst at 80 °C, and the catalyst ambient temperature for 21 h. The solid was separated by centrifugation, washed with deionized was labeled Au/FeOOH (80 °C). In some cases, the synthesis system pH was adjusted to 9 and 10 by water, and dried at 100 ◦ C to yield the final catalyst, which was labeled Au/FeOOH or Au/Fe2 O3 . sodium hydroxide addition, and the resultant products were labeled Au/FeOOH (pH 9) and For comparison, we also synthesized FeOOH-supported gold catalyst at 80 ◦ C, and the catalyst Au/FeOOH (pH 10), respectively. was labeled Au/FeOOH (80 ◦ C). In some cases, the synthesis system pH was adjusted to 9 and 10 by sodium addition, and the resultant products were labeled Au/FeOOH (pH 9) and Au/FeOOH 3.3.hydroxide Catalyst Characterization (pH 10), respectively. XRD patterns were recorded on a Panalytical diffractometer (PAN-alytical, Almelo, The Netherlands) using Cu-Kα radiation at a generator voltage of 40 kV and a tube current of 40 mA. 3.3. Catalyst Characterization TEM images were taken on a JEOL-2100 (JEOL, Tokyo, Japan) transmission electron microscope at an acceleration voltage of 200 Samples were crushed, dispersed in ethanol, and deposited on a XRD patterns were recorded onkV. a Panalytical diffractometer (PAN-alytical, Almelo, The Netherlands) microgrid prior to observation. using Cu-Kα radiation at a generator voltage of 40 kV and a tube current of 40 mA.

TEM images were taken on a JEOL-2100 (JEOL, Tokyo, Japan) transmission electron microscope at an acceleration voltage of 200 kV. Samples were crushed, dispersed in ethanol, and deposited on a microgrid prior to observation.


9 of 11

The specific surface areas were determined using an ASAP 2020 (Micromeritics, Norcross, GA, USA) gas sorption analyzer. Before measurement, Fe2 O3 was degassed under vacuum at 250 ◦ C for 5 h, FeOOH was degassed under vacuum at 100 ◦ C overnight, and the BET specific surface areas were calculated based on the linear part of the BET plot (P/P0 = 0.05–0.25). The Barrett–Joyner–Halenda (BJH) pore distributions were calculated based on desorption branches, and the total pore volumes were calculated based on the quantities of adsorbed nitrogen at the maximum relative pressure (P/P0 = 0.99). X-ray photoelectron spectroscopy (XPS) spectra were recorded on a XSAM800 X-ray photoelectron spectrometer (Kratos Analytical, Manchester, UK) with Mg-Kα monochromatic excited radiation (1253.6 eV) at a vacuum of less than 7 × 10−7 Pa. Fixed analyzer transmission mode was used in the test. The electron binding energy was calibrated by the C 1 s (Eb = 284.6 eV) spectrum. Gold loading on the catalysts was measured by atomic absorption spectroscopy using a TAS-990 spectrophotometer (Beijing Purkinje General Instrument Company Limited, Beijing, China). After adsorption, the solution containing support was centrifuged, and 1 mL of the supernatant was transferred into a 50 mL volumetric flask and diluted with 20% hydrochloric acid solution to volume, and the residual gold content in the solution was then determined by AAS. The actual gold loading was calculated based on the gold concentration change after adsorption. 3.4. Activity Evaluation Tests for catalytic CO oxidation were performed in a continuous-flow U-shaped glass reactor (4 mm inner diameter), and the reactor temperature was controlled using an ice bath, which contained 13 g ammonium chloride, 37.5 g sodium nitrate, and 100 g crushed ice. In each test run, 0.1 g of catalyst (~40–80 mesh) was loaded in the U-shaped glass reactor. The CO concentration was 10,000 ppm, balanced by air, and the total flow rate was controlled at 50 mL/min by a mass flow controller, which generated a gas hourly space velocity of ca. 30,000 h−1 . An online gas chromatograph equipped with a flame ionization detector and methanizer was used to measure the CO and CO2 concentrations in the feed and effluent gas. Pre-experiment suggested that FeOOH and Fe2 O3 in the reactor showed no activity for CO oxidation below 300 ◦ C. 4. Conclusions FeOOH- and Fe2 O3 -supported gold catalysts were prepared by adsorption of gold precursors onto the support under self-generated alkaline conditions. Since the support materials were prepared by reacting iron salt with alkali, minor alkalis were incorporated into the support materials, which could be released gradually into aqueous solution and basify the solution to pH 8. This favored gold precursor hydrolysis and their deposition onto the support materials. XPS spectra indicated that autoreduction of gold species occurred during synthesis. XRD and HRTEM characterizations suggested that the gold was highly dispersed on the support. A further increase in pH by alkali addition or elevating the synthesis solution temperature would decrease the gold loading. Au/FeOOH showed the best activity in the catalytic oxidation of CO, which could achieve complete oxidation of CO from −20 ◦ C to 20 ◦ C. Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4344/8/9/357/s1, Figure S1: XP spectra of the catalysts prepared under different conditions, Figure S2: Conversion curves of CO on the support materials, Figure S3: XRD patterns of the used catalysts, Figure S4: XP spectra of the used catalysts. Author Contributions: J.L. conceived and designed the experiments; F.P. and W.Z. performed the experiments and the characterizations; all authors contributed to the data interpretations; F.P., W.Z., and J.L. wrote the manuscript. Acknowledgments: This work was financially supported by the National Natural Science Foundation of China (21477092), and the Fundamental Research Funds for the Central Universities of China (2042017kf0185). Conflicts of Interest: The authors declare no conflicts of interest.


10 of 11

References 1. 2. 3. 4.

5. 6.

7. 8. 9.

10.

11.

12.

13. 14. 15.

16. 17. 18. 19. 20. 21.

Haruta, M.; Kobayashi, T.; Sano, H.; Yamada, N. Novel gold catalysts for the oxidation of carbon monoxide at a temperature far below 0 ◦ C. Chem. Lett. 1987, 16, 405–408. [CrossRef] Centeno, M.; Ramírez Reina, T.; Ivanova, S.; Laguna, O.; Odriozola, J. Au/CeO2 catalysts: Structure and CO oxidation activity. Catalysts 2016, 6, 158. [CrossRef] Xue, W.J.; Wang, Y.F.; Li, P.; Liu, Z.T.; Hao, Z.P.; Ma, C.Y. Morphology effects of Co3 O4 on the catalytic activity of Au/Co3 O4 catalysts for complete oxidation of trace ethylene. Catal. Commun. 2011, 12, 1265–1268. [CrossRef] Feng, X.; Duan, X.; Qian, G.; Zhou, X.; Chen, D.; Yuan, W. Au nanoparticles deposited on the external surfaces of TS-1: Enhanced stability and activity for direct propylene epoxidation with H2 and O2 . Appl. Catal. B Environ. 2014, 150–151, 396–401. [CrossRef] Keshipour, S.; Mirmasoudi, S.S. Cross-linked chitosan aerogel modified with Au: Synthesis, characterization and catalytic application. Carbohydr. Polym. 2018, 196, 494–500. [CrossRef] [PubMed] Fu, H.; Zhang, L.; Wang, Y.; Chen, S.; Wan, Y. Thermally reduced gold nanocatalysts prepared by the carbonization of ordered mesoporous carbon as a heterogeneous catalyst for the selective reduction of aromatic nitro compounds. J. Catal. 2016, 344, 313–324. [CrossRef] Barakat, T.; Rooke, J.C.; Genty, E.; Cousin, R.; Siffert, S.; Su, B.L. Gold catalysts in environmental remediation and water-gas shift technologies. Energy Environ. Sci. 2013, 6, 371–391. [CrossRef] Scirè, S.; Liotta, L.F. Supported gold catalysts for the total oxidation of volatile organic compounds. Appl. Catal. B Environ. 2012, 125, 222–246. [CrossRef] Smolentseva, E.; Simakov, A.; Beloshapkin, S.; Estrada, M.; Vargas, E.; Sobolev, V.; Kenzhin, R.; Fuentes, S. Gold catalysts supported on nanostructured Ce-Al-O mixed oxides prepared by organic sol-gel. Appl. Catal. B Environ. 2012, 115–116, 117–128. [CrossRef] ˚ M.; Matolín, V.; Nehasil, V. Study of the character Zahoranová, T.; Mori, T.; Yan, P.; Ševˇcíková, K.; Václavu, of gold nanoparticles deposited onto sputtered cerium oxide layers by deposition-precipitation method: Influence of the preparation parameters. Vacuum 2015, 114, 86–92. [CrossRef] Murayama, T.; Haruta, M. Preparation of gold nanoparticles supported on Nb2O5 by deposition precipitation and deposition reduction methods and their catalytic activity for CO oxidation. Cuihua Xuebao/Chin. J. Catal. 2016, 37, 1694–1701. [CrossRef] Sandoval, A.; Louis, C.; Zanella, R. Improved activity and stability in CO oxidation of bimetallic Au-Cu/TiO2 catalysts prepared by deposition-precipitation with urea. Appl. Catal. B Environ. 2013, 140–141, 363–377. [CrossRef] Ma, X.; Sun, Q.; Feng, X.; He, X.; Guo, J.; Sun, H.; Cao, H. Catalytic oxidation of 1,2-dichlorobenzene over CaCO3 /α-Fe2 O3 nanocomposite catalysts. Appl. Catal. A Gen. 2013, 450, 143–151. [CrossRef] Wang, Y.; Sun, H.; Ang, H.M.; Tadé, M.O.; Wang, S. Magnetic Fe3 O4 /carbon sphere/cobalt composites for catalytic oxidation of phenol solutions with sulfate radicals. Chem. Eng. J. 2014, 245, 1–9. [CrossRef] Tran, N.D.; Besson, M.; Descorme, C. TiO2 -supported gold catalysts in the catalytic wet air oxidation of succinic acid: Influence of the preparation, the storage and the pre-treatment conditions. New J. Chem. 2011, 35, 2095–2104. [CrossRef] Tang, Z.; Zhang, W.; Li, Y.; Huang, Z.; Guo, H.; Wu, F.; Li, J. Gold catalysts supported on nanosized iron oxide for low-temperature oxidation of carbon monoxide and formaldehyde. Appl. Surf. Sci. 2016, 364, 75–80. [CrossRef] Kamal, M.S.; Razzak, S.A.; Hossain, M.M. Catalytic oxidation of volatile organic compounds (VOCs)—A review. Atmos. Environ. 2016, 140, 117–134. [CrossRef] Yang, P.; Li, J.; Cheng, Z.; Zuo, S. Promoting effects of Ce and Pt addition on the destructive performances of V2 O5 /γ-Al2 O3 for catalytic combustion of benzene. Appl. Catal. A Gen. 2017, 542, 38–46. [CrossRef] Wolf, A.; Schüth, F. A systematic study of the synthesis conditions for the preparation of highly active gold catalysts. Appl. Catal. A Gen. 2002, 226, 1–13. [CrossRef] Cárdenas-Lizana, F.; Keane, M.A. The development of gold catalysts for use in hydrogenation reactions. J. Mater. Sci. 2013, 48, 543–564. [CrossRef] Tiya-Djowe, A.; Laminsi, S.; Noupeyi, G.L.; Gaigneaux, E.M. Non-thermal plasma synthesis of sea-urchin like α-FeOOH for the catalytic oxidation of Orange II in aqueous solution. Appl. Catal. B Environ. 2015, 176–177, 99–106. [CrossRef]


22. 23. 24. 25.

26.

27. 28.

29.

30.

31. 32.

33. 34. 35.

36. 37.

11 of 11

Mironov, I.V.; Afanas’eva, V.A. Gold(III) amine complexes in aqueous alkali solutions. Russ. J. Inorg. Chem. 2010, 55, 1156–1161. [CrossRef] Ji, Y. Ions removal by iron nanoparticles: A study on solid-water interface with zeta potential. Colloids Surf. A Physicochem. Eng. Asp. 2014, 444, 1–8. [CrossRef] Zhang, W.; Lu, X.; Zhou, W.; Wu, F.; Li, J. Mesoporous iron oxide-silica supported gold catalysts for low-temperature CO oxidation. Chin. Sci. Bull. 2014, 59, 4008–4013. [CrossRef] Kotolevich, Y.; Kolobova, E.; Mamontov, G.; Khramov, E.; Cabrera Ortega, J.E.; Tiznado, H.; Farías, M.H.; Bogdanchikova, N.; Zubavichus, Y.; Mota-Morales, J.D.; et al. Au/TiO2 catalysts promoted with Fe and Mg for n-octanol oxidation under mild conditions. Catal. Today 2016, 278, 104–112. [CrossRef] Meire, M.; Tack, P.; De Keukeleere, K.; Balcaen, L.; Pollefeyt, G.; Vanhaecke, F.; Vincze, L.; Van Der Voort, P.; Van Driessche, I.; Lommens, P. Gold/titania composites: An X-ray absorption spectroscopy study on the influence of the reduction method. Spectrochim. Acta Part B At. Spectrosc. 2015, 110, 45–50. [CrossRef] Roldan Cuenya, B.; Behafarid, F. Nanocatalysis: Size- and shape-dependent chemisorption and catalytic reactivity. Surf. Sci. Rep. 2015, 70, 135–187. [CrossRef] Huang, J.; Xue, C.; Wang, B.; Guo, X.; Wang, S. Gold-supported tin dioxide nanocatalysts for low temperature CO oxidation: Preparation, characterization and DRIFTS study. React. Kinet. Mech. Catal. 2013, 108, 403–416. [CrossRef] He, Y.; Liu, J.; Luo, L.; Wang, Y.; Zhu, J.; Du, Y.; Li, J.; Mao, S.X.; Wang, C. Size-dependent dynamic structures of supported gold nanoparticles in CO oxidation reaction condition. Proc. Natl. Acad. Sci. USA 2018, 115, 7700–7705. [CrossRef] [PubMed] Sarvesh, K.S.; Ryosuke, Y.; Ogino, C.; Akihiko, K. Biogenic synthesis and characterization of gold nanoparticles by Escherichia coli K12 and its heterogeneous catalysis in degradation of 4-nitrophenol. Nanoscale Res. Lett. 2013, 8, 70. [CrossRef] Taherkhani, F.; Akbarzadeh, H.; Rezania, H. Chemical ordering effect on melting temperature, surface energy of copper-gold bimetallic nanocluster. J. Alloys Compd. 2014, 617, 746–750. [CrossRef] Sudarsanam, P.; Mallesham, B.; Reddy, P.S.; Großmann, D.; Grünert, W.; Reddy, B.M. Nano-Au/CeO2 catalysts for CO oxidation: Influence of dopants (Fe, La and Zr) on the physicochemical properties and catalytic activity. Appl. Catal. B Environ. 2014, 144, 900–908. [CrossRef] Sun, K.; Kohyama, M.; Tanaka, S.; Takeda, S. Theoretical study of atomic oxygen on gold surface by Hückel theory and DFT calculations. J. Phys. Chem. A 2012, 116, 9568–9573. [CrossRef] [PubMed] Green, I.X.; Tang, W.; Neurock, M.; Yates, J.T. Spectroscopic observation of dual catalytic sites during oxidation of CO on a Au/TiO2 catalyst. Science 2011, 333, 736–739. [CrossRef] [PubMed] Liu, Y.; Deng, J.; Xie, S.; Wang, Z.; Dai, H. Catalytic removal of volatile organic compounds using ordered porous transition metal oxide and supported noble metal catalysts. Chin. J. Catal. 2016, 37, 1193–1205. [CrossRef] Liu, X.Y.; Wang, A.; Zhang, T.; Mou, C.Y. Catalysis by gold: New insights into the support effect. Nano Today 2013, 8, 403–416. [CrossRef] Liu, C.; Tan, Y.; Lin, S.; Li, H.; Wu, X.; Li, L.; Pei, Y.; Zeng, X.C. CO self-promoting oxidation on nanosized gold clusters: Triangular Au3 active site and CO induced O–O scission. J. Am. Chem. Soc. 2013, 135, 2583–2595. [CrossRef] [PubMed] © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).