1 Introduction

The water–gas shift (WGS) reaction has been extensively studied for removing CO and/or producing hydrogen. The reaction is carried out at 1 atm. It is an exothermic and equilibrium-limited reaction, with approximately 20–95 % of CO conversion reported in the literature [1]. There are two types of the commercial WGS catalysts: (1) a mixture of iron (Fe) and chromium (Cr) oxides, operated in the temperature range of 400–500 °C, with ~10 % (v/v) CO concentration in the feed gas and (2) a mixture of copper (Cu), zinc oxide (ZnO) and alumina (Al2O3), operated in the temperature range of 200–400 °C, with ~2–5 % CO concentration. Generally, the reaction is performed in a two stage-reactor, at high and low temperatures, with interstage cooling. There are few drawbacks associated with the operation of the WGS catalysts. They are pyrophobic and prone to sintering. Further, the activation of the catalysts is time-consuming. They are also sensitive to the start-up and shut-down cycles [2]. At present, the development of highly active, cost effective, thermally stable and non-pyrophobic shift catalysts is of considerable interests.

Among the several salient studies performed on the oxidation of CO over the metal oxides-supported noble metal catalysts, Luengnaruemitchai et al. [3] compared the performances of Pt/CeO2, Au/CeO2, and Au/Fe2O3 in the temperature range of 120–360 °C and found that Pt/CeO2 exhibited a high catalytic activity. The CO conversion was measured to be ~80 % at 360 °C. Deng et al. [4] achieved approximately the same conversion on Au–CeO2 at 350 °C. Si and Flytzani-Stephanopoulos [5] studied the effects of shape and crystal-plane of nanoscale ceria on the activity of Au/CeO2 in the temperature range of 150–350 °C. The Au-deposited nano-polyhedra CeO2 yielded the maximum CO conversion of ~80 % at 350 °C.

A few studies showed that the addition of an alkali metal improved the activity of noble metal dispersed in different oxides such as TiO2 [6] and SiO2 [7]. Zhu et al. [6] observed that the addition of sodium (Na) as a promoter in the Pt/TiO2 inhibited the sintering of the Pt particles. Wang et al. [7] showed that the addition of Na to the Pt/SiO2 catalyst not only promoted the oxidation of CO, but also provided structural stability to the Pt catalyst. Recently, a carbon-supported Pt catalyst was activated using Na and the catalyst was used for the WGS reactions in the temperature range of 100–400 °C [8]. The maximum CO conversion achieved was ~82 % at 350 °C. The WGS reaction has also been performed using oxygen (O2) over different noble metal catalysts, for example, Pt–Cu [9], Au–Co3O4 [10] and Pt/CeO2 [2]. These studies indicated that O2 had promotional effects on CO oxidation.

The WGS reaction has also been efficiently performed on the transition metal-based catalysts. To mention a few studies, Andreev et al. [11] studied the reaction over the alkali-treated nickel hydroxides [Ni(OH)2] dispersed in charcoal and the high catalytic activity was attributed to the presence of α-phase Ni(OH)2·xH2O in the carbon matrix. Li et al. [12] studied the reaction over the La2O3-stabilized Cu- or Ni-CeO2 in the temperature range of 150–600 °C. The CO conversion was observed to be ~78 % at 350 °C in each case. Approximately the same catalytic activity of Cu was reported in the studies of Tanaka et al. [13], Djinovic et al. [14], Zhang et al. [15], Si et al. [16] and Budiman et al. [17]. With regard to the WGS reaction studied using bimetal catalysts, for example, Ni–Mo [18], Cu–Ni [19], Fe–Cu [20, 21] approximately 60–80 % conversion was reported in these studies. Recently, Shinde et al. [22] have studied the oxidation of CO over the nanostructured Pd-modified Ni/CeO2 catalyst over the temperature range of 60–500 °C and achieved ~89 % conversion at 350 °C.The WGS reaction has also been studied using O2 over transition metals (Cu or Ni) [23–25]. Approximately 5–10 % increase in the conversion was achieved using O2 over the temperature range of 150–350 °C. Figure 1 shows a bibliographic compilation of CO conversion achieved at different oxidation temperatures, using various metal catalysts discussed above. In these studies, CO concentration was between 0.2 and 8 %, whereas the molar ratio of steam to CO was between 1:1 and 5:1. The pressure was kept constant at 1 atm.

Fig. 1
figure 1

Bibliographic compilation of CO conversion achieved at different oxidation temperatures, using various metal catalysts. CO concentration was between 0.2 and 8 %, whereas the molar ratio of steam to CO was between 1:1 and 5:1. P = 1 atm in all cases

Full size image

In continuation with advancement in the WGS catalysts, this study describes the development of an efficient and stable CeO2- and Ni NPs-based catalyst asymmetrically dispersed in the matrix of activated carbon micro-nanofibers (CNFs/ACF) for the removal of CO. The CNFs were grown on an ACF substrate, using catalytic chemical vapor deposition (CCVD) with Ni as the CVD catalyst and acetylene (C2H2) as the carbon source. The present study has several novelties. (1) Ni NPs played a dual role of the catalyst for the CNF growth and the oxidation of CO at 350 °C. (2) The reactions were performed without requiring O2 or a promoter. CeO2 provided the necessary nascent oxygen required for the oxidation of CO. (3) The reactions were performed in a perforated tubular reactor wrapped with the catalyst, without channeling and maldistribution of the gas.

2 Materials and Methods

2.1 Materials

The phenolic resin precursor-based ACFs were provided free of cost by Nippon Kynol Inc., Japan. Nickel nitrate hexahydrate [Ni(NO3)2·6H2O] (purity >99.99 %), cerium nitrate hexahydrate [Ce(NO3)3·6H2O] (purity >99 %), sodium dodecyl sulfate (SDS) (purity >99 %) were purchased from Merck, Germany. Nitrogen (purity >99.99 %), hydrogen (purity >99.99 %), CO (purity >99.99 %), and C2H2 (AAS grade) gases were supplied by Sigma Gases, Delhi, India.

2.2 Pretreatment of ACF

The volatile impurities and particulate matters present in the as-received ACF samples were removed by pretreating the samples with 0.05 M-aqueous HNO3 solution at 80 °C for 2 h. The samples were washed several times with deionized (DI) water until the pH of the ACF surface was approximately 7. The samples were then dried in a vacuum oven at 120 °C for approximately 12 h to remove the entrapped gases from the pores of the ACFs.

2.3 Catalyst Preparation

Figure 2 shows a schematic illustration of the steps for the catalyst preparation. The steps primarily included the impregnation of the ACFs with the metal salts, followed by calcination, reduction, and CVD. Approximately 1 g of the dried ACF sample was wrapped over a quartz tubular reactor (ID = 0.6 cm and L = 8 cm). One end of the tube was sealed. The tube walls were perforated with 0.5 mm diameter-holes. A 100 mL-mixture containing 0.4 M-aqueous solutions of Ce- and Ni-nitrate salts in 1:3 volume-ratios was used as the impregnating medium. The detailed experimental set-up used for the impregnation may be found in the study [26]. Prior to the impregnation, SDS (0.3 % w/w), an anionic surfactant, was added to the impregnating solution to provide the monodispersion of the metal salts in the solution and transfer maximum amount of the salts to the ACFs, without agglomeration. Some samples were also prepared without SDS for the comparison purposes. The impregnation was performed for 24 h, using a peristaltic pump.

Fig. 2
figure 2

Schematic illustration of the steps for the catalyst preparation

Full size image

After impregnation, the samples were dried in a static air for approximately 3 h and then transferred to the vacuum oven for further drying at 120 °C for 12 h to remove the unbound moisture. The dried samples were then calcined in a programmable horizontal furnace at 400 °C for 4 h to convert Ni(NO3)2 dispersed in the ACFs to Ni oxides. Calcination was performed in N2 to minimize the burn-off of the ACFs. The calcined samples were treated with H2 at 360 °C for 2 h in the reduction step to convert the oxides of Ni into its metallic state. At this temperature, Ce(NO3)3 was not converted to CeO2. The N2 and H2 gas flow rates during calcination and reduction were fixed at 150 and 120 sccm, respectively. The required calcination and reduction temperatures were selected based on the TPR analysis data, as is discussed later. Next, the CNFs were grown on the ACF using CVD at 550 °C for 10 min. As shown later in the SEM images, the Ni NPs were attached at the tips of the CNFs. At the CVD temperature of 550 °C, Ce(NO3)3 was converted to CeO2−x which remained dispersed in the ACFs. The prepared materials are denoted as CeO2–Ni-CNFs/ACF for the reference purposes in this study. For the comparison purposes, some samples of Ni-ACFs were prepared without CeO2 and used as the substrate to grow the CNFs. Such samples are denoted as Ni-CNFs/ACF. Some samples of Ni-CNFs/ACF were subjected to ultra-sonication under mild conditions (0.05 M-HNO3 and 10 min) to dislodge the Ni NPs from the tips of the CNFs, without affecting the nanofibers. Such sonicated samples are denoted as Ni-CNFs/ACF-S. Further, some CeO2-ACF samples were prepared by impregnating the ACFs with Ce(NO3)3 without Ni(NO3)2, followed by calcination at 450 °C.

2.4 Experimental Set-Up for CO Oxidation

The experimental set-up used for the oxidation of CO was similar to the set-up previously used for studying the catalytic activity of Zn NPs [27]. Briefly, the set-up consisted of a specially designed, vertical tubular stainless steel- (SS) shell (L = 140 mm, ID = 14 mm) mounted inside a vertical furnace. The shell contained a perforated SS tubular reactor (L = 50 mm, ID = 4 mm). The CNF/ACF catalyst was wrapped over the perforated section of the reactor and N2 was continuously purged through the reactor to prevent the oxidation of the metal NPs dispersed in the CNFs/ACFs. In such perforated tube-shell assembly, channeling and maldistribution of the gas were not observed.

CO and N2 flow rates were adjusted using mass flow controllers (model FSFIC-1, Bronkhorst, Netherlands). The total flow rate of the reactant gaseous mixture was fixed at 100 sccm. The CO to steam ratio was maintained at 1:3. A programmable PID controller module (Fuji Electric Co., Japan) was used to heat the reactor to reaction temperatures (250–450 °C). The reaction lasted for 3 h. In a typical oxidation reaction, approximately 1 g of the test sample was used. During the experiment, the gaseous mixture flowed down inside the tubular reactor, along with steam, and then radially outward through the perforations into the test samples, before exiting the shell. The residence time of the gaseous mixture in the reactor was ~7 s. The gaseous products containing CO2, N2, unconverted CO and steam were sent to a water chilled condenser for cooling before analyzing through a gas chromatography (Nucon, India, model no. 5700) with the thermal conductivity detector (TCD). The CO conversion was calculated using the formula, XCO = (Cinlet − Cexit)/Cinlet, where Cinlet and Cexit represent the CO concentrations at the inlet and at the outlet of the reactor under steady-state conditions, respectively. The CO oxidation was performed on the different catalysts, namely, ACF, CeO2-ACF, Ni-CNFs/ACF, CeO2–Ni-ACF and CeO2–Ni-CNFs/ACF to investigate the individual roles of Ni, CeO2, and CNFs in the oxidation of CO.

3 Surface Characterizations

Several analytical and spectroscopic techniques were used to characterize the prepared materials. The surface area and pore size distribution (PSD) were determined using the Autosorb-1C Quantachrome instrument (USA). The surface area and PSD were determined using the multi-point BET method. The Barrete-Joyner-Halenda (BJH) method and the density functional theory (DFT) were used to determine the meso- and micro-pore volumes, respectively. The atomic absorption spectroscopy (AAS) (Varian AA240, USA) with an air-acetylene flame was used to determine the Ni-loading in the prepared samples. X-ray diffractometer (XRD) (X-Pert Pro, USA) was used with a Cu Kα radiation (k = 1.54178 Å) to determine the different phases and average crystallite size of the Ni NPs. XRD analysis was performed within the 2θ-range of 18–70° at the scan rate of 3° per min. The average crystallite sizes of the Ni NPs were calculated using Scherrer formula (r = kλ/BCosθ), where the symbols have usual meaning. The Raman spectra were obtained from the Raman spectrometer (WiTec, Germany), using the laser light of 532 nm-wavelength. Scanning electron microscopic (SEM) images were recorded using the field-emission SEM (Supra 40 VP, Zeiss, Germany). TPR analysis was performed to determine the reduction temperature of the metal oxides dispersed in the ACFs, using the Autosorb-1C instrument.

4 Results and Discussion

4.1 Ni-Loadings in ACFs

Table 1 summarizes the Ni-loadings in the impregnated ACF samples with and without the SDS surfactant. The impregnation with 0.4 M-aqueous solution of Ni(NO3)2 without SDS yielded a Ni-loading of 160 mg/g in the ACFs. The loading was approximately doubled when SDS was used, indicating that SDS prevented the agglomeration of large size-metal salt particles in the solution, thereby improving monodispersion. The Ni-loading in the ACFs impregnated with the bimetal Ce(NO3)3- and Ni(NO3)2-salt solution with SDS was determined to be 239 mg/g.

Table 1 AAS analysis data of the Ni-loading
Full size table

4.2 TPR Analysis

Figure 3 shows the TPR profiles of the Ni(NO3)2–Ce(NO3)3-ACF samples calcined at different temperatures (100–400 °C). A major reduction peak was observed at ~295 and 230 °C in the samples calcined at 100 and 200 °C, respectively. The samples calcined at 300 and 400 °C exhibited two peaks each, with the major peak appearing at ~410 and 360 °C, respectively. Such profiles (two peaks in the samples calcined at high temperatures) were in agreement with the TPR data for NiO dispersed in CeO2 [28]. In this study, the authors attributed the peak at low temperature to the reduction of free NiO NPs, whereas that at high temperature was attributed to the reduction of NiO strongly bound with CeO2. Further, the intensity of the reduction peaks increased with increasing calcination temperatures from 100 to 400 °C. A relatively broad peak also appeared at ~530 °C, attributed to the partial decomposition of the various oxygen containing surface functional groups present in the ACF. It is mentioned that NiO-ACF (without CeO2) exhibited the reduction peak at ~490 °C [29]. Therefore, the shift in the reduction peak for NiO toward low temperatures, observed in the mixture of NiO and CeO2, indicated that the introduction of CeO2 improved the reducibility of NiO. Based on the TPR profiles, the optimum calcination and reduction temperatures were selected as 400 and 360 °C, respectively.

Fig. 3
figure 3

TPR profile of NiO–Ce(NO3)3-ACF samples calcined at various temperatures

Full size image

4.3 BET and PSD

The BET surface area (SBET, m2/g) and PSD of the prepared materials were determined using N2-adsorption isotherm over the relative pressure (P/P0) range between 0.05 and 0.35. The total pore volume (VT, cm3/g) was calculated based on the amount of N2 adsorbed at the relative pressure close to unity. Figure 4 shows the N2-adsorption isotherms of the prepared materials. The total amount (volume) of N2 adsorbed in the ACF, CeO2–Ni-ACF, and CeO2-ACF samples increased rapidly until the relative pressure range was ~0.1. Thereafter, the adsorbed volume gradually increased before leveling off to a constant value. The adsorption isotherm of this kind is categorized as Type I and signified the monolayer coverage of the adsorbate over the adsorbent. In the Ni-CNFs/ACF and CeO2–Ni-CNFs/ACF samples, the total volume of N2 adsorbed increased sharply near the relative pressure close to unity, indicating the formation of narrow pores or ultra microporosity in the CNF samples [30, 31].

Fig. 4
figure 4

N2-adsorption isotherms of the prepared materials

Full size image

Table 2 summarizes the SBET and PSD of the prepared materials. The SBET of the as-received ACF samples was determined to be 1269 m2/g and the material primarily contained micro-pores. Upon impregnation with the bimetallic salts, and thereafter, calcination, the SBET and VT values decreased in the ACFs because of the blockage of the pores with the metal NPs. However, an improvement in SBET and VT was observed in the samples reduced with H2. Further, a decrease of the SBET and VT in the CeO2–Ni-CNF/ACF samples was observed after growing the CNFs. The mesopore contents, however, increased. It may be mentioned that the calculated value of SBET (598 m2/g) for the CeO2–Ni-CNF/ACF sample may be underestimated by ~40 %, because the narrow pores are penetrated by the N2 molecules during the physisorption analysis [30]. The BET area of the single metal (Ni)-produced samples was larger than that of the bimetal samples.

Table 2 BET and PSD of the prepared materials
Full size table

4.4 Surface Morphology

SEM images were recorded to observe the surface morphology of the prepared materials. Figure 5 shows the representative SEM images of the ACF, CeO2–Ni-ACF, and CeO2–Ni-CNFs/ACF samples, at low (5 K×) and high (100 K×) magnifications. Figure 5a shows that the external surface of the ACFs was smooth. The images at high magnification showed that the surface contained pores of different sizes (micro-, meso-, and macro-pores) (Fig. 5a′). Figure 5b, b′ shows the SEM images of CeO2–Ni-ACFs. The NPs were uniformly distributed on the external surface of the ACFs. Figure 5c, c′ shows the SEM images of the CeO2–Ni-CNFs/ACF samples. An approximately uniform and dense growth of the CNFs across the ACF surface was observed. The average fiber diameter ranged between 10 and 30 nm. Figure 5d, d′ shows the SEM images of the sonicated samples. The ultra-sonication under mild conditions dislodged many Ni NPs from the tips of the CNFs. However, some NPs remained at the tips.

Fig. 5
figure 5

SEM images of (a, a′) ACF, (b, b′) CeO2–Ni-ACF, (c, c′) CeO2–Ni-CNF/ACF, (d, d′) Ni-CNFs/ACF-S

Full size image

Figure 6a shows the EDX spectra of the CeO2–Ni-CNF/ACF sample, indicating the presence of Ni, Ce and O. The spectra confirmed the shiny metal particles, located at the tip of the fibers, to be a Ni metal. Figure 6b shows the EDX elemental mapping of CeO2–Ni-CNF/ACF. The results show an approximately uniform distribution of Ni, Ce and O in the prepared materials. The spectra also confirmed the absence of any foreign impurities in the prepared material. The SEM images (not included here for brevity) of the sonicated samples showed the absence of the shiny Ni NPs in the materials.

Fig. 6
figure 6

EDX spectra and elemental mapping of CeO2–Ni-CNF/ACF

Full size image

4.5 XRD and Raman Analysis

Figure 7 shows the XRD pattern of the as-received ACF and Ni-CNF/ACF samples. In the ACFs, two major diffraction peaks were observed at 2θ values of ~25° and 43°, corresponding to the crystallographic indices of (0 0 2) and (1 0 1) planes, respectively. The former peak was attributed to the amorphous coke like substances present in the ACF. The broad and relatively less intense peak at ~43° corresponded to the presence of hexagonal graphite like substances. In Ni-CNF/ACF, two major characteristic peaks were observed at the 2θ values of ~44° and 52°, attributed to the crystallographic indices of (1 1 1) and (2 0 0) planes, respectively [29]. The average crystallite size was determined to be in 18 to 29 nm-ranges.

Fig. 7
figure 7

XRD pattern of the prepared materials

Full size image

Figure 8 shows the Raman spectra of the CeO2–Ni-CNFs/ACF sample. The peak at ~1336 cm−1 is attributed to the disordered phase present in the material and is termed as D-band. Another peak observed at ~1592 cm−1 is termed as G-band and attributed to the ordered (graphitic) phase present in the material. The ratio ID/IG is calculated to be 1.03 and reflects the predominance of the disordered phase over the ordered phase. The peak at ~460 cm−1 is attributed to the F2g mode of CeO2 and can be viewed as the symmetric breathing modes of oxygen atoms around the cerium ions [31].

Fig. 8
figure 8

Raman spectra of the prepared materials

Full size image

The Raman spectra of the CeO2–Ni-CNFs/ACF sample did not show any band at ~550 cm−1. Further, Raman analysis was also performed on the samples [Ce(NO3)3–Ni-ACF] reduced at 360 °C (before the CVD step) to confirm the absence of NiO in the samples. NiO exhibits an absorption band at ~550 cm−1 [32]. The spectra of the sample reduced at 360 °C also did not show any band at ~550 cm−1, re-confirming the absence of NiO in the reduced sample.

4.6 Material Activity Tests

Figures 9, 10, 11 and 12 show the experimental data of the tests performed on the oxidation of CO over different materials. Figure 9 shows the comparative performances of the different single- and bimetals-doped materials, namely, CeO2-ACF, Ni-CNFs/ACF, CeO2–Ni-ACF, and CeO2–Ni-CNFs/ACF. The reactions were carried out at 350 °C in each case, with the molar CO/H2O ratio was maintained constant at 1:3 and inlet CO concentration at 2 %. Figure 9 also includes the data for the oxidation reaction performed on the ACF substrate (without metals).

Fig. 9
figure 9

CO conversion over different materials: a ACF, b CeO2-ACF, c Ni-CNF/ACF-S, d Ni-CNF/ACF, e CeO2–Ni-ACF, and f CeO2–Ni-CNF/ACF (W = 1 g, CO concentration = 2 %, CO:H2O = 1:3, T = 350 °C)

Full size image
Fig. 10
figure 10

Reaction rates calculated from the differential reactor-data, with the tests performed on CeO2–Ni-CNFs/ACF (CO concentration = 2 %, CO:H2O = 1:3, T = 350 °C). The inset shows the plot of fraction conversion (X) with respect to the variable W/FCO

Full size image
Fig. 11
figure 11

Effect of reaction temperatures on conversion over CeO2–Ni-CNF/ACF (W = 1 g, CO concentration = 2 %, CO:H2O = 1:3, total flow rate = 100 sccm). The inset shows the equilibrium CO conversion data

Full size image
Fig. 12
figure 12

Effect of CO concentration on conversion over CeO2–Ni-CNFs/ACF (W = 1 g, CO:H2O = 1:3, T = 350 °C, total flow rate = 100 sccm)

Full size image

The ACF substrate exhibited a lowest CO conversion (~4 %). The conversion was attributed to the interaction (reaction) of the adsorbed CO with the different oxygen containing surface functional groups present in the ACFs [31, 33]. In CeO2-ACF, the CO conversion increased to ~26 %. CeO2 produces the CeO2−x (0 < x < 0.5) nonstoichiometric oxides in the presence of CO, which is reoxidized into CeO2 in an oxidizing environment such as H2O [34]. In Ni-CNFs/ACF, the CO conversion increased to ~37 %. The relatively higher conversion was commonly attributed to (1) the catalytic activity of the Ni NPs, (2) the presence of additional free edges of hexagon with several hetero atoms in the graphitic CNFs [31, 35], and (3) relatively increased exposure and interaction of the Ni NPs to CO, attached at the tips of the CNFs. The interaction of the Ni NPs with CO is as follows [36]:

$${\text{Ni}}_{\text{s}} + {\text{CO}} \to {\text{Ni}}_{\text{s}} {-}{\text{CO}}$$
$${\text{Ni}}_{\text{s}} {-}{\text{CO}} + {\text{Ni}}_{\text{s}} \to {\text{Ni}}_{\text{s}} {-}{\text{C}} + {\text{Ni}}_{\text{s}} {-}{\text{O}}$$
$${\text{Ni}}_{\text{s}} {-}{\text{CO}} + {\text{Ni}}_{\text{s}} {-}{\text{O}} \to 2{\text{Ni}}_{\text{s}} + {\text{CO}}_{ 2}$$

The improved reactivity of the Ni NPs and CNFs towards CO was reconfirmed by performing the oxidation study on the Ni-CNFs/ACF-S (sonicated) samples. The reduced Ni content in such samples adversely affected the catalytic activity of the material. The reduction in the catalytic activity was ~9 %, with the CO conversion decreasing to ~28 %. The higher conversion achieved in the sonicated samples than that in the ACF is attributed the reactive CNFs with their hexagonal edges. Furthermore, the active role of the Ni NPs can also be observed by comparing the activity of (b) CeO2-ACF and (e) CeO2–Ni-ACF samples, shown in Fig. 9. There was a significant increase (more than 35 %) in the conversion using the CeO2–Ni-ACF samples.

The CO conversion remarkably increased to ~65 % in the bimetallic CeO2–Ni-ACF catalyst, attributed to the provision of lattice oxygen in ceria for the reaction, combined with the catalytic effects of the Ni NPs. The mechanistic steps involved in the CO oxidation over CeO2–Ni-ACFs comprised of (1) the adsorption of CO on the active sites of Ni, (2) reaction with the nascent oxygen provided by CeO2, and (3) reoxidation of Ce2O3 by H2O [37].

$${\text{Co}} + \sigma {\text{Co}} \to {\text{CO}}_{\text{ads}} \left( {\sigma {\text{ represents the active surface sites of Ni}}} \right)$$
$${\text{CO}}_{\text{ads}} + 2{\text{CeO}}_{ 2} \to {\text{CO}}_{ 2} + {\text{Ce}}_{ 2} {\text{O}}_{ 3} + \sigma$$
$${\text{Ce}}_{ 2} {\text{O}}_{ 3} + {\text{H}}_{ 2} {\text{O}} \to 2{\text{CeO}}_{ 2} + {\text{H}}_{ 2}$$

The steady state CO conversion further increased from 65 to ~73 % in the CeO2–Ni-CNFs/ACF. Therefore, the catalytic activity of the best performing CeO2–Ni-CNFs/ACFs material prepared in this study was attributed to the combined effects of the Ni NPs, CeO2 and the CNFs. During the activity study, the deactivation of some of the catalysts (except ceria-containing samples) was observed, attributed to thermal degradation and sintering under the reaction conditions [14]. Another reason of the catalyst deactivation was the deposition of carbon on the catalyst surface. The similar reasons have been cited in the literature for the WGS-bimetal catalysts [12, 38, 39]. The comparative performances of the various prepared materials are the following: CeO2–Ni-CNFs/ACF > CeO2–Ni-ACF > Ni-CNFs/ACF > CeO2-ACF > ACF. It is important to mention that CeO2–Ni-CNFs/ACF was stable without burn-off at the oxidation temperatures at or below 450 °C.

Figure 10 shows the comparative rates for the ACF substrate and the best performing CeO2–Ni-CNFs/ACF material in the study. The rates were calculated from the experiments performed under differential reactor-mode. Expectedly, the highest reaction rate [~0.75 µmol CO/g-s] was observed in CeO2–Ni-CNF/ACF, which was approximately more than two orders of magnitude larger than that in the ACF substrate. The inset of the figure shows the plot of the fractional conversion versus W/FCO, with the rate calculated from the slope of the respective line: rate (r) = FCO X/W or X/(W/FCO), where FCO represents the molar flow rate (mol/sec) of the gas, X is the fractional conversion, and W represents the weight of the catalyst in g.

Figure 11 shows the effects of oxidation temperatures on the steady-state CO conversion over CeO2–Ni-CNF/ACF. The remaining operating conditions were kept fixed. The data showed that CO conversion increased with increasing reaction temperatures from 250 to 450 °C and reached 83 % at 450 °C. The conversion was ~50 and 80 % at 250 and 400 °C, respectively. It may be mentioned that some burn-off of the ACFs started occurring at 450 °C and above. Therefore, approximately 400 °C was recommended as a safe oxidation temperature for the CO oxidation using CeO2–Ni-CNFs/ACF, yielding 80 % CO conversion. The apparent activation energy for the CO conversion was calculated to be 51 kJ/mol, using the Arrhenius equation. Similar levels of activation energy are reported in the literature for the CO oxidation over different catalysts, for example, 38 kJ/mol on Ni–Ce(La)Ox [12], 54 kJ/mol on Au–Ce(La)Ox [4] and 30 kJ/mol on Cu–Ce(La)Ox [12].

Figure 12 shows the effect of CO concentrations on the steady state CO conversion over CeO2–Ni-CNFs/ACF at 350 °C. The CO conversion decreased with increasing CO concentrations. The conversion decreased from ~80 at 1 to ~73 % at 2 %-CO concentration. Further, CO conversion decreased to ~55 % at the inlet CO concentration of 3 %, indicating the relative suitability of the material at concentrations less than 3 %. The decrease in conversion at high CO concentrations is attributed to the poisoning of Ni by CO [3].

5 Conclusions

The bimetal CeO2 and Ni NPs-dispersed CNFs/ACF was developed as an efficient catalyst for the oxidation of CO by the WGS reaction. CeO2 and Ni were in situ incorporated in the carbon matrix during the impregnation step. The CO oxidation reactions were performed under different CO concentrations and reaction temperatures without using external O2. The steady state CO conversion at 400 °C was determined to be ~80 %. The high catalytic activity of the material was attributed to the combined roles of the catalytic Ni NPs, CeO2 providing oxygen for the reaction, reactive CNFs, and the oxygen containing surface functional groups present in the ACF. The prepared bimetals-doped material is a potential candidate for the CO oxidation catalyst.