C Nanocatalysts from Solid

This article is made available for a limited time sponsored by ACS under the ACS Free to Read License, which permits copying and redistribution of the article for non-commercial scholarly purposes.

Article www.acsanm.org

Cite This: ACS Appl. Nano Mater. 2018, 1, 265−273

Low-Pressure Plasma Synthesis of Ni/C Nanocatalysts from Solid Precursors: Influence of the Plasma Chemistry on the Morphology and Chemical State Emile Haye,† Yan Busby,*,† Mathieu da Silva Pires,† Florian Bocchese,† Nathalie Job,§ Laurent Houssiau,† and Jean-Jacques Pireaux† †

Laboratoire Interdisciplinaire de Spectroscopie Electronique, Namur Institute of Structured Matter, University of Namur, 61 Rue de Bruxelles, 5000 Namur, Belgium § Department of Chemical Engineering−Nanomaterials, Catalysis, Electrochemistry, University of Liège, Building B6a, Sart-Tilman, B-4000 Liège, Belgium ABSTRACT: Nanocatalyst materials based on metal nanoparticles (NPs) deposited on mesoporous carbon substrates are widely used in catalysis and energy storage; however, conventional wet-chemical deposition methods based on the reduction of metal salts are not always the best choice when looking for a process ensuring easy scalability and low environmental impact. Moreover, additional surface functionalization steps, such as the addition of nitrogen- or oxygencontaining groups, are more and more explored to increase the activity or the chemical stability of catalysts. In this work, we investigate a new methodology for the fabrication of nickel/ carbon nanocatalysts relying on a low-pressure radio frequency plasma treatment of solid (powder) precursors. A mesoporous carbon xerogel is used as support for nickel NPs synthesized through the decomposition of an organometallic nickel precursor in a plasma discharge. Different plasma treatment conditions and chemical environments are applied by varying the plasma power and the gas mixture injected into the plasma chamber (Ar, N2, NH3, and O2). The nucleation kinetics of nickel NPs, their morphology evolution, and chemical state were fully characterized by combining analytical techniques such as in situ optical emission spectroscopy, transmission electron microscopy, X-ray diffraction, and X-ray photoelectron spectroscopy. Results indicate that the plasma chemistry and conditions strongly influence the organometallic compound decomposition as well as the size and the oxidation state of the homogeneously dispersed nickel NPs. We compare the organometallic precursor degradation efficiency for each plasma by defining a rational “activation power” associated with each plasma chemistry. Moreover, simultaneous carbon substrate functionalization is obtained through plasma treatment, which demonstrates the high versatility of the plasma fabrication for developing green and efficient catalysts and energy materials. KEYWORDS: nickel nanoparticles, nanocatalysts, low-pressure plasma, organometal, surface functionalization, optical emission spectroscopy, photoelectron spectroscopy

■

stability, high electron conductivity, and abundance.8,9 Conventional methods to deposit NPs on a carbon support are mainly based on wet (chemical) methods. The NP synthesis requires multiple oxidation/reduction/washing/filtering steps, thus involving an extensive use of solvents and energy, leading to a high environmental impact of the catalyst. Moreover, solvent and impurities may result in the metal NPs contamination with agents that may lower the electrochemical activity and lifetime of the catalyst.6,10,11 Recently, solvent-free (dry) low-pressure plasma treatments have been applied to process platinum/ carbon (Pt/C) nanocatalysts for proton exchange membrane

INTRODUCTION Catalysts based on metal nanoparticles (NPs) supported on nanostructured materials are used in a multitude of applications, including electrochemical and energy devices. Nanocatalysts are extensively used in fuel cells, for the hydrogen transfer reaction, or for water splitting.1−5 In such applications, noble metal NPs, such as platinum or gold, are known to provide the highest electrochemical mass activity; however, noble metals are rare and expensive, so their use needs to be reduced for large-scale applications. Low-cost green catalysts based on metal-free or non-noble materials have recently attracted intense research; within them, nickel-based nanomaterials exhibit promising properties.2,6,7 Regarding the NP substrates, high surface area carbon materials (HSAC) have been extensively used thanks to their low cost, chemical © 2017 American Chemical Society

Received: October 31, 2017 Accepted: December 27, 2017 Published: December 27, 2017 265

DOI: 10.1021/acsanm.7b00125 ACS Appl. Nano Mater. 2018, 1, 265−273

Article

ACS Applied Nano Materials fuel cells: in particular, solid precursors (such as platinum acetylacetonate and carbon black) were treated in a lowpressure oxygen plasma discharge under stirring conditions. The authors suggested that the resulting Pt NPs were strongly anchored to the carbon black thanks to oxygen bridging bonds formation during the plasma.3,5 In the present work, we investigate the synthesis of Ni/C nanomaterials following a similar strategy based on low-pressure plasma treatments of solid precursors, to shine light on the nucleation and growth mechanisms of Ni NPs. Compared with previous works on Pt/ C plasma catalysts, here we have explored the decomposition nickel acetylacetonate organometallic precursor on a new carbon xerogel support and under different plasma compositions and powers. The plasma chemistry is varied by exploring three different gas mixtures comprising an inert, an oxidizing, or a reducing gas mixed with a carrier gas (argon) at different discharge power conditions. For each treatment, we compare the plasma efficiency in decomposing the Ni precursor, and we compare the morphology and chemical composition of the Ni/ C composite, showing that nickel nitrides are formed under ammonia-based plasma treatment. The simultaneous functionalization of the carbon xerogel surface by species contained in the plasma is also discussed. The plasma processing is demonstrated as a feasible and flexible alternative method for the NPs deposition by controlling (fairly independently) the particles morphology and their oxidation state to possibly meet the requirements for specific catalytic applications. Furthermore, the plasma decomposition kinetics of the organometallic precursor is discussed together with the possibility to simultaneously add functional groups (C−O, CO, C−N, etc.) to the carbon substrate. The carbon functionalization is particularly relevant, as it has been recently explored to enhance the catalyst efficiency and lifetime.12−15

Figure 1. Sketch summarizing the plasma synthesis process of the Ni/ C nanocatalysts. spectra are calibrated by fixing the C 1s main peak at 284.8 eV. Peak fitting is performed with the Avantage software (Thermo Scientific). The nanocatalyst morphology is characterized by transmission electron microscopy (TEM/STEM, Tecnai OSIRIS, FEI): for the analysis, a small amount of powder is dispersed in isopropanol, ultrasonicated, and finally deposited on a copper grid. Bright field (BF) and high-angle annular dark field (HAADF) TEM and STEM analyses are performed at an acceleration potential of 200 kV.

2. RESULTS AND DISCUSSION 2.1. Study of the Plasma Decomposition of the Organometallic Precursor. The RF plasma decomposition of Ni acetylacetonate and the formation of Ni NPs is investigated under three different plasma chemistries: (i) a gas mixture containing two inert gases (Ar/N2, 5:2 sccm), (ii) a mixture comprising one reducing gas species (Ar/NH3, 5:4 sccm), and (iii) a mixture comprising one oxidizing gas species (Ar/O2, 5:2 sccm). Argon is added to promote ionization mechanisms, improve the plasma stability, and help the decomposition of the organic part of the precursor. The ammonia flow rate is doubled with respect to N2 to compare plasma compositions having the same amount of nitrogen species. Depending on the injected gas flow, the reactor pressure during the treatment ranges from 3 to 7 mTorr; a higher working pressure generally resulted in the lowering of the plasma intensity (as is clearly seen in OES signals intensity); however, within the explored pressure range the plasma intensity is not sensibly different. The RF power is varied between 90 and 200 W, and the treatment is maintained until the complete decomposition of the precursor, which occurred between 45 and 60 min depending on the plasma chemistry and the discharge power. The precursor decomposition is followed by in situ OES measurements by monitoring specific emission lines associated with the precursor decomposition (as described in discussion below). Optical emission spectra are acquired in a wavelength range between 200 and 900 nm; the typical spectra obtained at the beginning of the treatment with each plasma composition are reported in Figure 2a. In the range between 680 and 895 nm and 395−435 nm, emission spectra are dominated by argon lines (Ar I), emission lines between 505 and 518 nm are ascribed to C2 Swan bands, and when nitrogen species are present (N2 or NH3), additional lines appear between 310 and 390 nm, corresponding to the N2 C−B lines and CN violet system (emission lines at 386.7 and 388.2 nm).20,21 When O2 is injected into the plasma discharge, OH-related emission lines are observed at 306.7, 309.2, and 313.6 nm (A−X lines20).

1. MATERIALS AND METHODS Ni/C nanocatalysts are prepared from powder precursors containing nickel acetylacetonate (Ni(acac)2, Strem Chemicals) and a carbon xerogel with a pore size of ∼100 nm, obtained following the procedure described elsewere.16 Carbon xerogel is chosen for (i) being a highpurity material17 and (ii) having a controllable porous texture consisting of mesopores and macropores,17 which limit the material transfer,18 and micropores, allowing a large anchoring surface for nanoparticles. The precursor loading in the powder mixture (containing C, O, and Ni species) is fixed to reach 20 wt % of Ni. Plasma treatments are performed in an inductively coupled radio frequency (RF) plasma reactor (13.56 MHz) described elsewhere,19 including a matching unit to minimize the reflected power from the plasma. The powder mixture (∼0.2 g) is homogeneously dispersed in a Petri box and placed into the discharge area of the plasma reactor. The reactor is pumped down by a turbo molecular pump, and masscontrolled gases (Ar, O2, N2, or NH3) are flown into the chamber before igniting the plasma. During the treatment, optical emission spectroscopy (OES) measurements were performed using an Ocean Optics USB4000XR spectrometer by connecting an optical fiber to a quartz window. The nanocatalyst synthesis process is schematized in Figure 1. The crystal structure of the produced Ni/C nanocatalysts is analyzed by X-ray diffraction (XRD, X’Pert PRO Panalytical) using the Cu Kα radiation (1.5406 Å). The composition, carbon surface functionalization, and the chemical state of the Ni NPs were characterized by X-ray photoelectron spectroscopy (XPS, K-Alpha Thermo Scientific) using a monochromatic Al Kα radiation (1486.68 eV). The X-rays spot size was 250 μm; survey spectra were acquired at a pass energy of 200 eV, and high-resolution spectra (C 1s, N 1s, O 1s and Ni 2p) were obtained at 30 eV. The scan number is adjusted (between 5 and 30 scans) to the specific element to get similar signalto-noise ratios. A flood gun is used for charge compensation, and 266


Article

ACS Applied Nano Materials

Figure 2. (a) OES spectra acquired in situ at the initial stage of the nitrogen-based (Ar/N2 and Ar/NH3) and oxygen-based (Ar/O2) plasma treatments. Characteristic lines from the organometallic precursor decomposition are identified and followed in situ. (b) Evolution of the selected OES signals associated with the decomposition of the organometallic precursor in the different plasma environments: for Ar/N2 and Ar/NH3 plasmas, the OES signal at 388.2 nm (CN emission lines) is selected, while for the Ar/O2 plasma we follow the signal at 309.2 nm (OH emission lines).

Figure 3. (a) XRD patterns of the untreated precursor and 200 W plasma-treated powder showing the disappearance of nickel acetylacetonate crystal peaks attesting its decomposition as soon as the selected OES signal drops. (b) XRD patterns obtained in Ar/NH3 treatments at different discharge power showing the formation of Ni3N domains with a size increasing with the discharge power (the average size is estimated from the Debye− Scherrer formula (inset)).

3b). When the precursors are treated with Ar/O2 and Ar/N2 plasmas, only a weak NiO phase is observed for 200 W treatments (Figure 3a), suggesting that amorphous oxidized Ni particles are formed by these plasma treatments. The efficiency and kinetics of a given plasma chemistry in decomposing the Ni acetylacetonate is estimated from OES measurements; namely, an activation “energy” is associated with each specific plasma decomposition process. To do so, the organometallic precursor (OM) decomposition in the discharge is schematically described by the following reaction:

The intensity of the Ar emission lines is fairly constant during the treatment, while other lines exhibit a clear intensity rise and fall along the plasma treatment. These lines are therefore associated with the Ni precursor decomposition: among them, we selected the most intense lines to monitor this decomposition in situ. In nitrogen- and ammonia-based plasmas, we selected the CN line at 388.2 nm, and for Ar/O2 plasma, we selected the OH line at 309.2 nm. The rise and fall of these signal intensities in a 140 W plasma treatment are shown in Figure 2b. The direct correlation between the drop of these specific OES signals and the precursor decomposition is confirmed by acquiring ex situ XRD spectra at regular treatment time steps, showing the progressive disappearance of the XRD pattern related to the (crystalline) Ni(acac)2 precursor (Figure 3a; only the initial and final time are presented). For Ar/NH3 plasma (Figure 3b), the XRD pattern clearly indicates the formation of crystal domains attributed to nickel nitride phase (Ni3N). The average Ni3N domain size increases with the plasma treatment power as estimated by the Debye− Scherrer formula from the average full width at half-maximum (fwhm) of the three most intense (110), (002), and (111) Ni3N diffraction peaks.22 For the Ar/NH3 plasma, a linear correlation is found between the Ni3N domains size (from 5.8 to 8.1 nm) and the discharge power (see the inset in Figure

k OM → F1* + F2

(1)

where F1* represents a generic volatile organic fragment (produced in an excited state), and F2 represents other intermediate decomposition products, containing nickel atoms. In this framework, the fast relaxation of F1* to the ground state generates the CN or OH lines observed by OES. In the second step, F1 is pumped down, while F2 fragments decompose progressively to generate Ni NPs. For a generic reaction such as A → B + C, we can define the consumption rate r of the organometallic precursor as r=− 267

d[OM] = k × [OM]n dt

(2) DOI: 10.1021/acsanm.7b00125 ACS Appl. Nano Mater. 2018, 1, 265−273

Article


Figure 4. Derivation of the plasma activation power (Pa, W) from the OES analysis: (a) the selected OES line intensity as a function of the treatment time; (b) resulting signal after the integration and normalization; (c) determination of the rate constant k (y-axis intercept) and the reaction order n (linear fit slope); (d) analogy of an Arrhenius plot for deriving the activation power (linear fit slope); (e) reaction rate as a function of the plasma composition and power; (f) reaction order as a function of the plasma composition and power. The error bars on n and k are derived from the upper and lower values of the linear fit slope in (c).

where [OM] is the organometallic concentration at the reaction time t, and n and k are the reaction order and reaction rate, respectively. Equation 1 allows relating [OM] to the concentration of F1* and thus to the selected OES line intensity (IRAW, Figure 4a). In other words, by assuming that [F1*] is directly proportional to IRAW, [OM] can be followed by the expression

The eq 5 is a straight line equation if ln(IRAW) is plotted

(

versusln 1 −

∫0 IRAW dt ∞

∫0 IRAW dt

(3)

The eq 3 simply states that [OM] is proportional to the normalized integral of the OES signal intensity (IRAW) and varies from 1 (initial state, t = 0) to 0 (end of the plasma treatment), following a decreasing sigmoid function shown in ∞ Figure 4b. The expression ∫ IRAW dt is a constant for each 0 treatment, called A for simplicity. Equation 2 becomes n t ⎡ ∫0 IRAW dt ⎤⎥ IRAW ⎢ = k × ⎢1 − ⎥ A A ⎣ ⎦

(4)

From this, it is possible to determine n and k, by applying the natural logarithm on both sides of the previous equation: ⎛ 1 ln(IRAW ) = ln A + ln k + n × ln⎜1 − × ⎝ A

∫0

t

t

)

× ∫ IRAW dt , where the reaction rate and the 0

reaction order are determined from the y-axis intercept and the slope, respectively (Figure 4c). Their values allow comparing the ability of the plasma to decompose the organometallic and give a reasonable parameter to evaluate its ability to form Ni NPs. The reaction order n (Figure 4f, from 0.4 to 1) shows a weak dependence on the discharge power and the plasma chemistry except for the Ar/NH3 treatment at 200 W, indicating a possibly different decomposition mechanism. A value of n close to 1 indicates that the reaction rate is proportional to the [OM]. The rate coefficient k depends on the energy provided to the reaction and on the specific reaction mechanism. For all plasma compositions, k increases with the discharge power between 90 and 200 W (Figure 4e); however, for NH3/Ar plasmas, the rate increase at 200 W is much more pronounced than for other plasma compositions. This trend agrees with the higher reaction order observed in Figure 4f suggesting that a different precursor decomposition mechanism is activated at high discharge power. This behavior is attributed to the presence of reducing species (H) promoting new decomposition paths at high power. To get more insights on the “activation energy” that is necessary to trigger the organometallic compound decomposition in a specific plasma chemistry, we introduce an approach based on Arrhenius plots. In Arrhenius plots, the temperature is the variable used to estimate the energy that is provided to the reaction; however, low-pressure plasmas are

t

[OM] = 1 −

1 A

⎞ IRAW dt ⎟ ⎠ (5) 268


Article


Figure 5. HAADF STEM micrographs showing nickel-based NPs (bright regions) deposited on a carbon xerogel. Images refer to the Ar/N2 treatment at 200 W (a) and to the Ar/NH3 plasma treatment at both 90 (b) and 200 W (c). The brighter (with deliberate excess of contrast) shells in (c) suggests the formation of an inhomogeneous (radial) composition profile in NPs formed in Ar/NH3 plasmas, with a “brighter” shell thickness estimated to 2−3 nm.

Figure 6. Evolution of (a) O, (b) Ni, and (c) N surface composition (atom % derived from XPS survey spectra) as a function of the plasma treatment power for Ar/O2, Ar/N2, and Ar/NH3 plasmas.

NPs are clearly observed for each plasma treatment with a particle size affected by both the plasma chemistry and the power. The size distribution is unimodal for each treatment condition. STEM images of Ni/C composites treated in a 90 W Ar/NH3 plasma (Figure 5b) show an average particle size of ∼3 nm (±0.5 nm), while, at a higher power (200 W, Figure 5c), the average particle size increases up to 5−6 nm (±1 nm). The particle size increase at high power is attributed to higher mobility of metal atoms and small clusters leading to the formation of larger particles to minimize the surface energy. At 200 W Ar/NH3 plasmas, we observe a clear contrast between the surface and the core of the NPs (see the inset in Figure 5c): this suggests a radial composition profile or core−shell structuring of Ni NPs. For Ar/O2 and Ar/N2 treatments, the synthesized NPs are smaller with respect to Ar/NH3 treatments: at 200 W, Ar/O2 and Ar/N2 treatments result in average Ni particle sizes of 4 and 2 nm (±0.5 nm), respectively (Figure 5a,b). The NP size estimated by TEM and XRD are slightly different: an overestimation of the NPs size by XRD has been frequently reported; moreover, the Ni3N crystal domains estimated by XRD can also be overestimated by the possible presence of few aggregates. 2.3. Chemical Analysis on Nickel NPs. The plasma treatment conditions were correlated to the surface composition and oxidation state of the Ni NPs and the carbon support, by performing XPS analysis. The evolution of the Ni, O, and N content for each set of plasma composition as a function of the treatment power is shown in Figure 6. As expected, the maximum oxygen content (∼35 atom %) is obtained with Ar/ O2 treatments; in Ar/N2 and Ar/NH3 plasmas, the oxygen comes mainly from the organometallic precursor and from the

nonequilibrium systems for which the temperature is not a welldefined parameter, since the electron temperature differs from the ion temperature.23 A more convenient and accessible energy parameter to derive the activation energy is the RF discharge power (PRF). By analogy with Arrhenius plots, we thus define the “activation power” (Pa) as ⎛ P ⎞ k = A exp⎜ − a ⎟ ⎝ PRF ⎠

(6)

The Pa value is graphically derived from the slope of the (linear) fit of ln(k) as a function of 1/PRF (Figure 4d). The linear dependence observed between 90 and 200 W justifies our choice of PRF as an empirical energy parameter. The derived activation power values for nitrogen-based plasmas (Pa = 317 W for Ar/N2 and Pa = 313 W for Ar/NH3) are significantly higher than for Ar/O2, (Pa = 228 W), suggesting that oxygen plasma may be more effective in decomposing the Ni precursor at low discharge power. However, since the plasma-induced precursor decomposition mechanisms f and the NPs nucleation are complex, this interpretation should only be taken as a general guideline. 2.2. Morphology of the Ni/C Composites. This paragraph discusses the effect of the plasma chemistry and discharge power on the morphology and chemical state of Ni NPs deposited on a carbon xerogel. The treatment duration is defined by the drop of specific OES line intensity and checked by the disappearance of the characteristic XRD peaks from Ni acetylacetonate. The Ni/C composite morphology is investigated by STEM (Figure 5), from which we evaluated the average particle size distribution. Homogeneously dispersed Ni 269


Article


Figure 7. (a) High-resolution Ni 2p3/2 spectra obtained at 90 W (bottom) and 200 W (top) plasma treatments. (b) Example of the peak-fitting procedure of the Ni 2p3/2 spectrum (Ar/NH3, 200 W conditions) to identify the 0, +2, and +3 oxidation states.

preferential sputtering of oxygen from Ni oxide by high-energy Ar+ species.27 Finally, in Ar/O2 treatments, nickel NPs are fully oxidized both at low and high power. The XPS analysis allows evidencing that the Ni3+/Ni2+ ratio decreases at high power going from 6.2 (at 90 W) to 3.8 (at 200 W); this suggests that, to a lower extent, Ni oxide NPs are also reduced at high power by Ar+ ions. 2.4. Functionalization of the Carbon Material. Highresolution C 1s and N 1s spectra (Figure 8) are curve-fitted to explore the ability of the plasma to add O- and/or N-based functional groups on the carbon support. The C 1s spectrum from the untreated powder mixture shows the expected chemical groups from nickel acetylacetonate (Figure 8a). The Ar/O2 treatment at 200 W (Figure 8b) results in the strongest functionalization/oxidation of the carbon surface with C−O, CO, and OCO groups representing altogether ∼70% of the C 1s peak area (see Figure 9). The lowest degree of functionalization of the carbon support is obtained for Ar/NH3 plasmas (Figure 8c), where functional groups represent only 45% of the C 1s peak area. It is important to stress the fact that the measured surface composition is the result of two possible contributions: (i) the possible presence of organic residuals coating coming from the incomplete decomposition of the organometallic precursor and (ii) the substitution or the bonding of N and O species to the carbon xerogel atoms. An additional difficulty in peak-fitting comes from the fact that, in nitrogen-based plasma treatments (Figure 8c,d), the discrimination between C−O and C−N functional groups by XPS is complicated by their very close binding energy. Despite this complexity, by comparing Ar/NH3 and Ar/N2 plasmas, the latter leads to a higher incorporation of N (12 atom %) forming more C−N bonds (Figure 9). A further analysis of N 1s spectra (Figure 8e,f) allows separating the three contributions coming from pyridinic, pyrrolic functions and graphitic nitrogen groups, respectively, occurring at 398.5, 400.0, and 401.0 eV binding energies,28 while the fourth component, observed at 397.5 eV, is assigned to nickel nitride.29 For 200 W plasma treatments, the Ar/NH3 composition favors the formation of pyridinic-N

carbon xerogel. The highest nitrogen content (12 atom %), is obtained with Ar/N2 treatments, while it is almost halved (7 atom %) with Ar/NH3 treatments (Figure 6c). It is useful to stress here that the composition quantification with XPS is made by assuming a homogeneous material surface (over the first 5−10 nm); for a heterogeneous surface containing nanoscale domains, the domain size and morphology strongly affects the value of the derived atomic percentages and possibly the peak binding energy.24,25 Here, the highest Ni content (18 atom %) is obtained for high-power Ar/O2 and Ar/NH3 plasma treatments; this is explained by the larger average NP size (>4 nm) as compared to Ar/N2 treatments (in which the NP size lies between 1 and 3 nm). High-resolution Ni 2p3/2 XPS spectra (Figure 7b) clearly show the presence of three chemical components corresponding to the Ni0, Ni2+, and Ni3+ oxidation states and their shakeup satellite peaks (Figure 7b).26 From the comparison between the spectra from 90 and 200 W treatments (Figure 7a), ammonia-based plasmas result in the lowest Ni oxidation with a Ni0 component representing up to 23% of total Ni 2p3/2 peak area with a slight difference between lowand high-power treatments. The previously discussed XRD and TEM results indicate that Ni3N NPs are formed with a size increasing when increasing the discharge power, with an outer shell producing an enhanced contrast in HAADF-STEM images. We attribute this radial structuring to the enhanced reduction of the Ni NPs surface by direct exposure to the reducing ammonia-based plasma. Interestingly, in Ar/NH3 treatments the RF power affects the NPs size but not their chemical composition: however, because of the too-thin outer shell, the radial structuring could not be clearly distinguished in STEM images of catalysts made 90 and 140 W plasmas. Nickel NPs formed at 90 and 140 W Ar/N2 plasma are fully oxidized, while a clear metallic component (accounting for ∼20% of the Ni 2p peak area) appears at 200 W treatments (Figure 7a). This suggests that the decomposition of the organometallic precursor at high power in Ar/N2 plasmas leads to metallic Ni particles, although fairly amorphous, since no Ni peaks are observed in XRD spectra. We ascribe this effect to the 270


Article


Figure 8. Examples of peak-fitting of high-resolution C 1s for (a) untreated powder mixture, (b) after Ar/O2 200 W, (c) Ar/NH3 200 W, and (d) Ar/N2 200 W treatment. N 1s XPS spectra after (e) after Ar/N2 200 W and (f) Ar/NH3 treatment.

groups (representing 64% of the N 1s signal), while Ar/N2 chemistry resulted in a relatively higher amount of pyrrolic-N

and graphitic-N (representing 30% and 8% of the N 1s peak area, respectively). Figure 9 graphically summarizes the carbon 271


Article


Figure 9. Graphical overview of XPS analysis results obtained from the fitting of the (a) C 1s, (b) Ni 2p, and (c) N 1s spectra from the untreated and the 200 W plasma-treated samples.

functionalization and the influence of the plasma composition on the N-functionalization. These results may have relevant applications when looking for a controlled N-functionalization with pyrrole and pyridinic defects, which have been shown to promote different catalytic reactions. The C-functionalization and the final NPs chemical state are clearly linked in the plasma synthesis; however, as Ni NPs are formed with all the explored plasma compositions, if a specific carbon functionalization should be avoided as required by a specific application, then one could simply select a different gas mixture. Moreover, the carbon substrate can be pretreated with a “functionalizing plasma”, and then the metal precursor can be added and treated with a second inert plasma treatment that will allow to maintain (to a certain extent) the first functionalization.

NH3 plasmas allow the functionalization of the carbon substrate mainly with pyridinic nitrogen functions. (2) Inert Ar/N2 plasmas resulted in intermediate activation power, while the NP size is limited to 3 nm; however, the Ar/N2 composition allows for the highest incorporation of N species into the carbon xerogel, with the formation of pyridinic, pyrrole, and graphitic functions. While nickel NPs were completely oxidized at 90 W, the NPs reduction occurs when increasing the discharge power so that a metallic component appears in the Ni 2p XPS spectrum for 200 W Ar/N2 treatments. This behavior is attributed to the presence of Ar+ ions, which help to decompose the organometallic precursor and reduce Ni oxide. (3) Oxidizing Ar/O 2 plasmas resulted in the lowest activation power and fully oxidized NPs with intermediate sizes (3−5 nm); as expected, this plasma chemistry allows for the strongest functionalization of the carbon support with carboxyl and carbonyl groups.

■

CONCLUSIONS We have investigated the synthesis of Ni/C composites following a simple, eco-friendly, solvent-free plasma processing method in which nickel acetylacetonate decomposes into nickel NPs anchored to mesoporous carbon support. The organometallic precursors were decomposed by different plasma compositions including Ar/NH3, Ar/N2, and Ar/O2 RF plasma chemistries, at a discharge power varying from 90 to 200 W. For each treatment condition, we have derived the precursor decomposition kinetics, as well as the nanocatalyst surface composition and morphology, including the chemical analysis on the Ni NPs and the carbon functionalization. Overall results are (1) Reducing Ar/NH3 plasma resulted in the highest process activation power; however, the activation turns out to be much more effective at 200 W plasma power, due to the activation of additional decomposition paths promoted by the reducing H species. Interestingly, the presence of ammonia allows the formation of Ni3N NPs with a tunable size up to 6 nm (in the explored conditions), characterized by a radial composition distribution with a more metallic shell of ∼3 nm thickness. Moreover, Ar/

To sum up, the illustrated plasma synthesis route appears to be a versatile process to deposit controlled Ni/C composites while simultaneously functionalizing the carbon support. These materials can thus be tuned for specific applications in catalysis or energy storage.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Emile Haye: 0000-0001-9162-561X Yan Busby: 0000-0002-6826-6142 Notes

The authors declare no competing financial interest. 272


Article


■

(17) Job, N.; Pirard, R.; Marien, J.; Pirard, J.-P. Porous Carbon Xerogels with Texture Tailored by PH Control during Sol−gel Process. Carbon 2004, 42, 619−628. (18) Job, N.; Marie, J.; Lambert, S.; Berthon-Fabry, S.; Achard, P. Carbon Xerogels as Catalyst Supports for PEM Fuel Cell Cathode. Energy Convers. Manage. 2008, 49, 2461−2470. (19) Felten, A.; Bittencourt, C.; Pireaux, J. J.; Van Lier, G.; Charlier, J. C. Radio-Frequency Plasma Functionalization of Carbon Nanotubes Surface O2, NH3, and CF4 Treatments. J. Appl. Phys. 2005, 98, 074308. (20) Machala, Z.; Janda, M.; Hensel, K.; Jedlovský, I.; Leštinská, L.; Foltin, V.; Martišovitš, V.; Morvová, M. Emission Spectroscopy of Atmospheric Pressure Plasmas for Bio-Medical and Environmental Applications. J. Mol. Spectrosc. 2007, 243, 194−201. (21) Cvelbar, U.; Krstulović, N.; Milošević, S.; Mozetič, M. Inductively Coupled RF Oxygen Plasma Characterization by Optical Emission Spectroscopy. Vacuum 2007, 82, 224−227. (22) Debye, P.; Scherrer, P. Interferenzen an regellos orientierten Teilchen im Röntgenlicht. I. Nachrichten Von Ges. Wiss. Zu Gött. Math.Phys. Kl. 1916, 1916, 1−15. (23) Belmonte, T.; Noël, C.; Gries, T.; Martin, J.; Henrion, G. Theoretical Background of Optical Emission Spectroscopy for Analysis of Atmospheric Pressure Plasmas. Plasma Sources Sci. Technol. 2015, 24, 064003. (24) Wang, Y.-C.; Engelhard, M. H.; Baer, D. R.; Castner, D. G. Quantifying the Impact of Nanoparticle Coatings and Nonuniformities on XPS Analysis: Gold/Silver Core−Shell Nanoparticles. Anal. Chem. 2016, 88, 3917−3925. (25) Busby, Y.; Pireaux, J. J. Metal Nanoparticle Size Distribution in Hybrid Organic/Inorganic Films Determined by High Resolution XRay Photoelectron Spectroscopy. J. Electron Spectrosc. Relat. Phenom. 2014, 192, 13−18. (26) Payne, B. P.; Grosvenor, A. P.; Biesinger, M. C.; Kobe, B. A.; McIntyre, N. S. Structure and Growth of Oxides on Polycrystalline Nickel Surfaces. Surf. Interface Anal. 2007, 39, 582−592. (27) Steinberger, R.; Walter, J.; Greunz, T.; Duchoslav, J.; Arndt, M.; Molodtsov, S.; Meyer, D. C.; Stifter, D. XPS Study of the Effects of Long-Term Ar+ Ion and Ar Cluster Sputtering on the Chemical Degradation of Hydrozincite and Iron Oxide. Corros. Sci. 2015, 99, 66−75. (28) Wang, H.; Maiyalagan, T.; Wang, X. Review on Recent Progress in Nitrogen-Doped Graphene: Synthesis, Characterization, and Its Potential Applications. ACS Catal. 2012, 2, 781−794. (29) Zhang, B.; Xiao, C.; Xie, S.; Liang, J.; Chen, X.; Tang, Y. Iron− Nickel Nitride Nanostructures in Situ Grown on Surface-RedoxEtching Nickel Foam: Efficient and Ultrasustainable Electrocatalysts for Overall Water Splitting. Chem. Mater. 2016, 28, 6934−6941.

ACKNOWLEDGMENTS The authors thank the Wallonia Region for financial support (Project HYLIFE No. 1410135). N.J. also thanks the Fonds de Bay for funding. The Synthesis, Irradiation & Analysis of Materials platform of the Univ. of Namur is acknowledged for XPS measurements, and SERMA Technologies is acknowledged for TEM/STEM analyses.

■

REFERENCES

(1) Shalom, M.; Ressnig, D.; Yang, X.; Clavel, G.; Fellinger, T. P.; Antonietti, M. Nickel Nitride as an Efficient Electrocatalyst for Water Splitting. J. Mater. Chem. A 2015, 3, 8171−8177. (2) Metin, Ö .; Mazumder, V.; Ö zkar, S.; Sun, S. Monodisperse Nickel Nanoparticles and Their Catalysis in Hydrolytic Dehydrogenation of Ammonia Borane. J. Am. Chem. Soc. 2010, 132, 1468−1469. (3) Laurent-Brocq, M.; Job, N.; Eskenazi, D.; Pireaux, J.-J. Pt/C Catalyst for PEM Fuel Cells: Control of Pt Nanoparticles Characteristics through a Novel Plasma Deposition Method. Appl. Catal., B 2014, 147, 453−463. (4) Alonso, F. Nickel Nanoparticles in the Transfer Hydrogenation of Functional Groups. RSC Catal. Ser. 2014, 17, 83−98. (5) Busby, Y.; Stergiopoulos, V.; Job, N.; Pireaux, J. J. Low Pressure Plasma Synthesis of Pt/C Catalysts for Fuel Cells Applications. Proceedings of the 22nd International Symposium on Plasma Chemistry, Antwerp, Belgium, July 5−10, 2015; ISPC, 2016; pp 1−2. (6) Yu, D.; Zhang, X.; Wang, K.; He, L.; Yao, J.; Feng, Y.; Wang, H. Sawtooth-Shaped Nickel-Based Submicrowires and Their Electrocatalytic Activity for Methanol Oxidation in Alkaline Media. Int. J. Hydrogen Energy 2013, 38, 11863−11869. (7) Carnes, C. L.; Klabunde, K. J. The Catalytic Methanol Synthesis over Nanoparticle Metal Oxide Catalysts. J. Mol. Catal. A: Chem. 2003, 194, 227−236. (8) Yang, H.; Vogel, W.; Lamy, C.; Alonso-Vante, N. Structure and Electrocatalytic Activity of Carbon-Supported Pt-Ni Alloy Nanoparticles toward the Oxygen Reduction Reaction. J. Phys. Chem. B 2004, 108, 11024−11034. (9) Merche, D.; Dufour, T.; Baneton, J.; Caldarella, G.; Debaille, V.; Job, N.; Reniers, F. Fuel Cell Electrodes From Organometallic Platinum Precursors: An Easy Atmospheric Plasma Approach. Plasma Processes Polym. 2016, 13, 91−104. (10) Carpenter, M. K.; Moylan, T. E.; Kukreja, R. S.; Atwan, M. H.; Tessema, M. M. Solvothermal Synthesis of Platinum Alloy Nanoparticles for Oxygen Reduction Electrocatalysis. J. Am. Chem. Soc. 2012, 134, 8535−8542. (11) Park, S.; Kheel, H.; Sun, G.-J.; Hyun, S. K.; Park, S. E.; Lee, C. Ethanol Sensing Properties of Au-Functionalized NiO Nanoparticles. Bull. Korean Chem. Soc. 2016, 37, 713−719. (12) Chen, Y.; Wang, J.; Liu, H.; Banis, M. N.; Li, R.; Sun, X.; Sham, T.-K.; Ye, S.; Knights, S. Nitrogen Doping Effects on Carbon Nanotubes and the Origin of the Enhanced Electrocatalytic Activity of Supported Pt for Proton-Exchange Membrane Fuel Cells. J. Phys. Chem. C 2011, 115, 3769−3776. (13) Gong, K.; Du, F.; Xia, Z.; Durstock, M.; Dai, L. Nitrogen-Doped Carbon Nanotube Arrays with High Electrocatalytic Activity for Oxygen Reduction. Science 2009, 323, 760−764. (14) Jeon, I.-Y.; Yu, D.; Bae, S.-Y.; Choi, H.-J.; Chang, D. W.; Dai, L.; Baek, J.-B. Formation of Large-Area Nitrogen-Doped Graphene Film Prepared from Simple Solution Casting of Edge-Selectively Functionalized Graphite and Its Electrocatalytic Activity. Chem. Mater. 2011, 23, 3987−3992. (15) Shao, Y.; Sui, J.; Yin, G.; Gao, Y. Nitrogen-Doped Carbon Nanostructures and Their Composites as Catalytic Materials for Proton Exchange Membrane Fuel Cell. Appl. Catal., B 2008, 79, 89− 99. (16) Job, N.; Théry, A.; Pirard, R.; Marien, J.; Kocon, L.; Rouzaud, J.N.; Béguin, F.; Pirard, J.-P. Carbon Aerogels, Cryogels and Xerogels: Influence of the Drying Method on the Textural Properties of Porous Carbon Materials. Carbon 2005, 43, 2481−2494. 273
