2.1. Catalyst Characterizationm
shows the physicochemical properties of the catalysts and supports used in this work. Zirconia (Z) is a commercial support with a pre-established monoclinic phase and a low surface area (8 m2
/g). On the other hand, the hydrous zirconium oxide prepared in our laboratory, ZrO2
O, is a high surface area material (340 m2
/g). When this hydrated oxide is thermally treated as precursors are, at 600 °C for 2 h, the loss of water and crystallization occur, and consequently the surface area decreases.
The addition of gold and cobalt phases does not generate any changes in the BET surface area of the commercial support (Z). Cobalt addition to the hydrogel and the subsequent thermal treatment cause an important and expected surface area loss even greater than that observed in the support without impregnation (Zt 600). Conversely, catalysts containing gold and cobalt present specific surface areas similar to those observed in supported cobalt catalysts. These results suggest that gold addition has not modified the catalyst textural properties.
Scanning electron microscopy– energy dispersive spectroscopy (SEM-EDS) tests were carried out with the purpose of obtaining a semi-quantitative analysis of cobalt content; the calculated Co/Z ratios are listed in Table 1
.Gold element concentration could not be evaluated since the emission lines overlap with those of zirconium element. The results obtained (Table 1
) indicate that gold addition using the deposit-precipitation technique causes a decrease in cobalt concentration. The atomic content of surface cobalt is 3.78 times greater in the CoZ catalyst than in the AuDpCoZ catalyst. These results could indicate that gold species have been deposited on oxidic cobalt species.
exhibits energy dispersive spectroscopy elemental mapping images showing the distribution of cobalt in AuDpCoZ and AuDpCoZt catalysts. The shining points, corresponding to cobalt, are homogeneously dispersed.
Powder X-ray diffraction (XRD) analyses were performed to identify the crystalline phases of both the support and the supported species. Figure 2
shows the XRD patterns of CoZ, AuZ, AuCoZ, AuDpZ, and AuDpCoZ catalysts with diffraction lines corresponding to monoclinic phase (M) of zirconia located at 2θ= 28.2, 31.5, 34.2, and 50.2° (PDF No. 03-065-1025). This result is expected, given that the commercial zirconia used has a pre-established crystalline structure.
In the case of CoZt, AuZt, AuDpZt, AuCoZt, and AuDpCoZt catalysts, which were prepared from zirconia hydrogel, the diffraction lines located at 2θ =30.0, 35.0, 50.0, and 60.0° (PDF No. 01-089-7710) correspond to tetragonal metastable phase (T) of zirconia (Figure 3
). Evidently, in most samples, there is some interaction between doping species and the zirconia hydrogel, generating the metastable tetragonal phase [27
] during the calcination process. The different behavior presented by the AuZt catalyst is worth noting. Its XRD pattern shows diffraction lines of two different crystalline phases of zirconia, i.e., monoclinic and tetragonal, the first being the most intense. This result suggests that the interaction between gold and the hydrogel is not favored when this metal is added using the traditional method.
Subsequently, the crystalline species generated by the supported phases were analyzed. The diffraction patterns of AuZ and AuZt catalysts, where gold was incorporated with the traditional technique, present diffraction lines corresponding to metallic gold species 2θ = 38.3, 44.5° (PDF No.00-004-0784). These signals are more intense in the AuZt diffraction pattern. It is evident that gold, which is added as HAuCl4, is then reduced in the thermal treatment of precursors. Besides, Au0lines are not clearly observed in the diffraction patterns of samples prepared with the DP method. It is evident that the gold addition method conditions both the formation of the support structure and the size of supported crystalline phases.
The XRD patterns of cobalt catalysts promoted with gold (AuCoZ, AuCoZt, AuDpCoZ, AuDpCoZt) do not clearly exhibit signals associated with the presence of metallic gold.
The presence of Co3
is evidenced in all cobalt-containing catalysts with characteristic diffraction lines located at 2θ= 31.2, 36.8, 59.3, and 65.1° (PDF No. 01-080-1533).In order to see whether gold addition modifies the characteristics of cobalt spinel crystals, Co3
, the crystal size was analyzed with the Scherrer equation from the diffraction line located at 2θ=36.8°, and the results obtained are reported in Table 1
. All catalysts exhibit similar cobalt oxide crystal size, between 15 and 22 nm. The addition of gold produces slight changes in the supported cobalt oxide crystal size. The addition of gold to CoZ catalysts, using both the traditional and the DP methods, led to a slight increase in the Co3
crystal size. Instead, the addition of gold to CoZt catalyst led to a slight decrease in the Co3
crystal size. These results evidence an interaction between gold and cobalt species.
In order to examine whether gold was in a metallic state, Au0
, transmission electron microscopy (TEM) analyses were performed. TEM micrographs (see Supplementary Material Figure S1
) revealed the presence of metallic particles in all gold-containing catalysts. The size distribution of gold particles for each sample was obtained from the micrographs; the histograms are shown in Figure 4
and the average values are listed in Table 1
.Gold-containing catalysts obtained by the DP method and/or those containing cobalt exhibit a relatively narrow gold particle size distribution, mainly consisting of particles having diameters between 2 and 6 nm, except for AuZ and AuZt catalysts, which exhibit a bimodal distribution of crystal size, crystals of a small size, and some crystals of a large particle size (> 20nm). The latter results are consistent with those obtained by XRD. It is evident that both the gold addition method and the presence of cobalt species influence the size distribution of gold particles.
According to the results, it can be concluded that the gold particle sizes in samples where zirconia adopts tetragonal structure are slightly smaller than those in zirconia with a monoclinic structure. Bimetallic catalysts prepared by the DP technique do not show significant differences in gold particle size with respect to catalysts containing only the precious metal in their formulation.
In order to analyze the gold effect on reducibility, temperature programmed reduction (TPR) analyses were performed. Catalysts containing only gold show no signals of reduction, evidencing that gold is found as a metallic species. Figure 5
A and 5B show H2
consumption as a function of temperature for catalysts supported on monoclinic and hydrated zirconia, respectively.
The pure Co3
TPR profile is characterized by two reduction steps, assigned to Co(III)-Co(II) and Co(II)-Co(0) (with maxima at 324°C and 397 °C) [29
]. The TPR profiles of the catalysts under study show a series of signals in the range 300–400 °C for catalysts supported on monoclinic zirconia, and in the range 130–550 °C for catalysts supported on hydrated zirconia. This indicates the presence of various types of cobalt oxide species, or species with different interactions with the support.
Moreover, it can be observed that gold addition using the traditional impregnation technique generates a slight shift towards higher reduction temperatures for CoZ and CoZt. Instead, the incorporation of gold by the DP method shifts the maxima of the reduction bands to lower temperatures. Notoriously, the AuDpCoZt catalyst begins to reduce at a lower temperature than those containing only cobalt. The precious metal presence favors the reduction in the catalyst prepared by DP method; this effect is not observed with the addition of gold by the traditional method. Furthermore, there are reduction signals at very low temperatures (130 °C), suggesting the presence of species with very low interaction with the support. The lower reduction temperature can be explained based on the small size of Co3
particles, as was observed by XRD, and/or on the interaction of transition metal species with the zirconia surface [30
]. For this catalyst, both the support and the gold addition method favor the surface dispersion of the cobalt oxide, therefore increasing the reducibility. This fact also was reported by Solsona et al., who proposed that this reducibility enhancement is probably due to the formation of Co3
domains at the gold–CoOx interface perimeter [32
In general, it can be observed that catalysts that preferably adopt a tetragonal structure are more reducible than catalysts prepared on a monoclinic support. In a research work using CuZ [33
], it was reported that copper species reducibility increases in ZrO2
, especially in tetragonal phase. It is suggested that during the calcination of hydrated zirconia to give anhydrous ZrO2
, the OH-
groups are responsible for the reduction of Zr4+
ions, while the presence of a lower Zr oxidation state generates an increase in anionic vacancies. The vacancy effect of zirconia network affects the redox properties of supported oxide species, increasing the reducibility because these supported particles are in a reducing environment. It is worthwhile remarking that AuDpCoZt is the most reducible catalyst due to an earlier start in reducing the supported cobalt species. In this catalyst, two effects can be associated with the increase of the reducibility, the first one associated with the source zirconia and the second associated with the “disperser effect” of the gold added by the DP technique on the cobalt oxide.
X-ray photoelectron spectroscopy (XPS) was used to investigate the chemical state of surface species of catalysts. The multiregion XPS spectra of all gold-containing catalysts show binding energies of the doublet Au 4f7/2
and Au 4f5/2
at 83.6 and 87.2 eV, respectively, corresponding to metallic gold (Au0
), as reported in the literature [34
] (the spectra of AuCoZt and AuDpCoZt catalysts are shown as Supplementary Material Figure S2
). The presence of reduced gold species is evident in all samples promoted with gold, even those where Au diffraction lines could not be observed by XRD, probably due to their small size.In order to delve into the gold oxidation state a little further, the Au 4f7/2
signal was deconvoluted for all catalysts under study. It is important to remark that as a component around 85.8eV and another one at 85.1 eV corresponding to Au3+
are not observed, the presence of ionic gold in all the samples under study is discarded [34
The Co 2p
core level cobalt spectra show the doublet Co 2p3/2
and Co 2p1/2
at about 781 and 795 eV, respectively, with their respective shake-up satellites (Figure 6
shows the binding energies of peaks found in the Co 2p
core level spectra of bimetallic catalysts. The existence of satellite peaks is associated with the presence of paramagnetic species and, for this reason, this signal is associated with the presence of Co2+
. The Co3
presents core level 2p3/2
signals at 779.6 eV of Co3+
and at 780.7 eV of Co2+
shows that in all bimetallic catalysts, the binding energy is associated with the presence of Co3
species. The appearance of satellite peaks at about 6 eV from the high energy component is further evidence for Co2+
]. It can also be observed that the difference in binding energy between Co 2p3/2
and Co 2p1/2
(ΔE) is about 15 eV, indicating the coexistence of Co2+
].From the results presented, it is not appreciated that the gold addition method generates a change in the BE of the surface cobalt species.
summarizes the surface atomic ratios, Co/Z, Au/Z, Au/Co, and Co2+
. When analyzing the Au/Z and Au/Co ratios, the solids synthesized by the DP method have a higher surface Au content compared to the ones synthesized by the traditional method. The same behavior is observed with the Co/Z ratio. The catalytic activity can be influenced by the differences between cobalt and Au surface concentrations [40
].It is noteworthy that the catalysts promoted with gold using the DP method show the highest Co/Z and Au/Z surface atomic ratios. The AuDPCoZt catalyst also exhibits the highest surface Au/Co ratio, indicating that this catalyst has the highest concentration of surface gold. All catalysts present a similar Co2+
2.2. Catalytic Activity
2.2.1. Catalytic Combustion of Propane in the Presence of O2/He
Propane combustion activity results for all catalysts presented in this paper are shown in Figure 7
and Figure 8
. These experiments were carried out using an O2
/He mixture as oxidizing agent. Table 4
summarizes T50 and T100 values (temperature at which 50% and 100% conversions are reached, respectively) obtained in each experiment.
Catalysts containing only gold in their formulation have low activity for the reaction, but it is noteworthy that they are more active than the pure support. Additionally, it can be observed that catalysts prepared from zirconia hydrogel (AuZt and AuDpZt) have greater activity (T50) than those prepared using commercial monoclinic zirconia. This fact is probably related to the larger surface area and smaller gold crystal size in catalysts supported on Zt zirconia.
Gold addition to the CoZ catalyst using the traditional impregnation technique causes an approximate shift of 40 °C towards higher temperatures, decreasing the activity of this catalyst for propane combustion. A similar effect occurs when gold is added to CoZt catalyst, but only T50 is modified in this case, maintaining T100 value. These results may be associated with the reducibility decrease shown in TPR profiles in which a shift of the reduction signals towards higher temperature is observed for AuCoZ and AuCoZt.
Instead, gold addition to the CoZ catalyst employing DP technique leads to a slight increase in the catalyst activity, evidenced only at high conversions (>70%) with a decrease of the T100. On the other hand, the presence of gold nanoparticles clearly produces an increase in the activity of the CoZt catalyst and the T50 decreases (~30 °C). In this case, the noble metal addition promotes the activity, and the synergic effect between Au-Co3O4-ZrO2 is more evident.
It should be noted that the post-reaction catalysts were analyzed using the XPS technique, and no changes were found at the surface level with regard to the surface atomic ratios or the BE of the elements of the supported species. These results are shown in Supplementary Material Table S1
From the results presented previously, it can be suggested that the catalytic cycle involves various factors that must be taken into account. On the one hand, the high activity of oxide cobalt, Co3O4, is evident, and this result is consistent with those widely reported in the literature. On the other hand, the role of the support is very important because it can participate in the redox cycles and, last but not least, the synergic effect of gold addition. Some aspects related to the physicochemical properties and the activity of the catalysts promoted with gold using the DP technique are discussed below.
The results must be analyzed taking into account that monometallic gold catalysts do not show good catalytic activity at low temperatures (T <400 °C). Changes in the nature of redox sites and higher availability of Co3
supported phases may explain differences in the activity of AuDpCoZ and AuDpCoZt catalysts. In the AuDpCoZt catalyst, gold addition generates a promoting effect, reaching a T50 at 265 °C, which is a good propane combustion temperature and, in this catalyst, gold generates a greater beneficial synergic effect than in AuDpCoZ. According to XPS results, it is observed that the AuDpCoZ catalyst presents higher surface cobalt content than the AuDpCoZt catalyst. Despite this, the AuDpCoZt catalyst is the one with the highest surface content of Au of the series; this sample also presents a higher content of exposed gold according to the Au/Zr and Au/Co ratios (Table 3
). Therefore, the best catalytic performance observed for AuDpCoZt can be attributed to the presence of a high surface concentration of cobalt and gold. There is also a beneficial interaction between oxide cobalt and gold nanoparticles that promotes Co3
dispersion and consequently increases the system reducibility. It is important to remark that this synergism is not observed when gold was added using the traditional impregnation technique. Solsona et al. [41
] reported the gold promoting effect of cobalt catalysts supported on mesoporous silica in the total oxidation of propane and toluene. They concluded that the presence of gold generates higher reducibility of certain cobalt species, which facilitates the redox cycle.
In this context, it is also important to highlight the influence of the support. The most active catalyst is that obtained from the high surface area hydrogel used as zirconia source. In this catalyst, zirconia adopts the metastable tetragonal structure and oxygen vacancies are generated. The presence of zirconia oxygen vacancies favors these processes because the reducibility of species increases, as was observed by TPR results. Au-Co3O4 domains intimately interacting with the support can promote oxygen lattice mobility, which can increase the activity. Moreover, the greater metal–zirconia interaction favors the formation of vacancies in the bulk, and consequently, facilitates the movement of oxygen in the catalytic system from the bulk to the surface.
It has been reported in the literature that one of the most accepted mechanisms in the catalytic oxidation of hydrocarbons is the Mars-van Krevelen (MVK) mechanism [42
]. In this mechanism, it is postulated that the metallic oxide provides the active oxygen for the reaction and generates a redox cycle where Co(III) can be reduced to Co(II) [17
]. Solsona et al. [43
] studied propane oxidation in the absence of oxygen using a CuOx-CeO2
catalyst, and observed CO2
as reaction product only, indicating lattice oxygen participation. Another accepted mechanism for the catalytic combustion of hydrocarbons is the Langmuir-Hinshelwood mechanism [44
]. This one implies the existence of active adsorption sites where the reagents interact and become products. As the zirconia has oxygen vacancies, they can be adsorption sites of reactants (oxygen or hydrocarbons). The O2
adsorption can be explained by the involvement of Zr3+
redox cycles and the formation of O2-ads
]. According to the results obtained in this work, both mechanisms can contribute to the catalytic activity. In this context, and in agreement with what is reported in the literature [18
], the role of gold can be associated with an enhancement of the cobalt species reducibility, which facilitates the redox cycle. The role of gold has been associated with a promoting effect of reoxidation or recovery of the cobalt oxide during the redox cycle [18
2.2.2. Catalytic Combustion of Propane in the Presence of O2/NO/He
The activity of catalysts in the presence of nitrogen oxides was also measured as this pollutant is commonly found in emissions from mobile sources. The results obtained using the same previous operating conditions with the addition of 1000 ppm of NO to the feed are shown in Figure 9
and Figure 10
The results of the catalysts supported on monoclinic zirconia show that the presence of NO in gold catalysts favors the activity, using T50 as a comparison parameter. On the other hand, in the case of monometallic catalysts containing cobalt, the presence of NO poses a slight disadvantage (Table 5
). In bimetallic systems, these two opposite effects could coexist and lead to an increase or decrease in the activity.
The greatest positive effect of gold occurs in the AuCoZ catalyst, leading to a catalyst with increased activity due to the addition of NO to the feed. On the other hand, in the AuDpCoZ catalyst, the cobalt depressor effect prevails, leading to a less active catalyst in the presence of NO (T50).
It should be noted that in the presence of NO, the only system that shows a slight increase in activity when gold is added is CoZt (T50= 320 °C) promoted by the DP method (T50= 305 °C).
In the case of catalysts prepared from the hydrogel (Figure 10
), the effect of NO is negative for the activity of both monometallic gold and cobalt catalysts. This effect is transferred to bimetallic catalysts regardless of the gold addition method.
Differences in catalytic behavior in the presence of NO could be attributed to prior oxidation of NO to NO2 (which is a better oxidant), or to competitive NO adsorption at the sites where the reactants are adsorbed (hydrocarbon or oxygen).
2.2.3. Catalytic Combustion of Naphthalene
The results for naphthalene combustion are presented in Table 6
. Naphthalene is oxidized at a very high temperature in the absence of catalyst with T50= 440 °C and T100= 490 °C. The catalysts decrease these temperatures considerably, indicating that they are active for naphthalene catalytic combustion. According to catalytic results, it can be observed that monometallic cobalt catalysts are more active than gold catalysts. In addition, by comparing gold catalysts, it can be seen that AuDpZ and AuDpZt are much more active than AuZ and AuZt. Once again, it can be seen that the catalysts prepared using the DP method show higher activity than the catalysts prepared by the traditional method.
In general terms, gold addition using the traditional method does not lead to an increase in the activity of the CoZ and CoZt catalysts. The CoZ catalyst is very active, and gold addition does not increase its activity. However, the high activity of gold deposited by the DP technique on the CoZt catalyst is noticeable. When the species are supported on tetragonal zirconia, of greater surface area, gold and cobalt present a synergic effect and this catalyst, AuDpCoZt, is the most active, achieving total naphthalene conversion to CO2 at 260 °C, which is a very good combustion temperature.Again, as was observed for propane combustion, the synergic effect between gold and cobalt species present in the AuDpCoZt catalyst is noteworthy.
It should be noted that the presence of carbonaceous residues in the catalytic processes studied (XPS not shown) is discarded, and the only product observed is CO2, which indicates a selectivity greater than 99%.
There is almost no literature research on total naphthalene oxidation using supported gold-Co3
catalysts, even though these systems were used for the combustion of other aromatic hydrocarbons such as toluene and benzene [45