1 Introduction

Fuel cells convert chemical energy directly into electrical energy with several advantages, such as simple construction, low cost, light weight, high efficiency, and low-pollution. Direct methanol fuel cells (DMFCs), which uses methanol directly as the fuel, are one of the promising fuel cell technologies due to their high energy density, relatively low operating temperature, absence of reformer unit, and the easy handling of methanol [1–4]. There has been considerable research effort directed towards methanol electro-oxidation [5, 6]. Apart from methanol, which is relatively toxic and volatile, other short chain organic alcohols such as ethanol, 1-propanol, and 2-propanol can also be used as a fuel [7, 8]. Ethanol is particularly interesting due to such advantages as renewability, safety, availability (from biomass products), and high energy density [9–11]. Therefore, ethanol may be more attractive than methanol for direct alcohol fuel cells. The complete oxidation of ethanol is not without problems such as the production of many adsorbed intermediates and by-products, and the difficulty in cleaving the C–C bond at low temperatures. C3-alcohols, such as 1-propanol and 2-propanol, may also provide alternatives to methanol and ethanol [12–15]. Even though C1–C3 alcohols have a high energy content per unit mass, they do not exhibit any electrochemical activity, so a catalyst is needed to activate them. Platinum is a well-known electrocatalyst for the oxidation of alcohols on the anode of direct alcohol fuel cells (DAFCs) [16, 17]. The performance of the catalyst can be enhanced by using Pt nanoparticles instead of bulk Pt, due to their large surface area. Various surfactants can be used to produce stable platinum nanoparticles, and previous studies have shown that the surfactant can affect the size of metal particles and the catalytic performance of the catalysts [18]. In this study, six platinum nanoparticle catalysts have been prepared using PtCl4 as a starting material and 1-decylamine, N,N-dimethyldecylamine, 1-dodecylamine, N,N-dimethyldodecylamine, 1-hexadecylamine, and 1-octadecylamine as surfactants—for the first time. The size and activity of the catalysts towards C1–C3 alcohol oxidations, was measured in order to determine the effect of the chain length and the primary and tertiary structure of the surfactants. X-ray diffraction (XRD), X-ray photoelectron microscopy (XPS), atomic force microscopy (AFM), and transmission electron microscopy (TEM) were used to characterize the amine-stabilized platinum nanoparticles. Cyclic voltammetry (CV) and chronoamperometry (CA) measurements were performed to investigate the electrochemical activities and stabilities of all catalysts in the oxidation of methanol, ethanol, 1-propanol, and 2-propanol. Catalyst II, which was stabilized by the tertiary amine surfactant with the shortest chain length (N,N-dimethyldecylamine), is the most active catalyst. Its activity is 25.5, 49.6, 23.2, and 30.6 times higher for methanol, ethanol, 1-propanol, and 2-propanol oxidation reactions, respectively, than commercially available ETEK Pt catalysts.

2 Experimental

2.1 General

PtCl4 (99 % Alfa Aesar), tetrahydrofuran (THF) (99.5 %, Merck), methanol (≥99.5 %, Merck), ethanol (99.9 %, Merck), 1-propanol (≥99.0 %, Merck), 2-propanol (≥99.5 %, Merck), and HClO4 (60 %, Merck), lithium triethylborohydride (1.0 M dissolved in THF, Sigma Aldrich), 1-decylamine (Sigma Aldrich), N,N-dimethyldecylamine (Sigma Aldrich), 1-dodecylamine (Sigma Aldrich), N,N-dimethyldodecylamine (Sigma Aldrich), 1-hexadecylamine (Sigma Aldrich) and 1-octadecylamine (Sigma-Aldrich) and Carbon XC-72 (Cabot Europa Ltd.) were used as-received. Water was deionized by a Millipore water purification system (18 MΩ) analytical grade. All glassware and Teflon-coated magnetic stir bars were cleaned with aqua-regia before use, followed by washing with distilled water.

2.2 Instrumentation

The TEM micrographs of the catalysts were recorded using a JEOL 200 kV microscope. A drop of the Vulcan carbon–Pt nanoparticle solution was placed on 400-mesh carbon-coated copper grids and these were dried, under vacuum, at room temperature, before analysis. More than 300 particles were examined in order to obtain average particle size distributions of all prepared catalysts.

Specs spectrometer was used to perform XPS analyses using the Kα lines of Mg (1253.6 eV, 10 mA) as an X-ray source. A Cu double-sided tape (3 M Inc.) was used to deposit the samples. The C 1s line at 284.6 eV was chosen as a reference point and peak fittings were done using a Gaussian function.

A Rigaku diffractometer with an Ultima + theta–theta high-resolution goniometer, equipped with a Cu Kα source operating at 40 kV and 40 mA, was used to acquire XRD patterns of all samples. The scan range was 10°–90° and the scan rate was 5° min−1.

A Digital Instruments Multi Mode AFM Nanoscope IV a (Veeco Ins., Santa Barbara, CA) was employed to visualize the surface topographies of the prepared catalysts at ambient temperature. It was operated in ‘tapping’ mode using 0.01–0.025 Ω cm antimony-doped silicon AFM probes (Ultrasharp TESP with about 2 nm tip radius curvature, cantilever spring constants of 20–80 N m−1 and resonance frequencies of 328–379 kHz). The images were analyzed using the Nanoscope 6.13 (version 6.13.r1, Veeco) software package. To prepare the AFM samples, 2.5 μL of the catalyst solution (diluting 300-fold or more with DI water) was placed directly on a freshly cleaved mica disk (supporting material) and allowed to dry in a vacuum at room temperature for at least 12 h.

All electrochemical experiments (CV and CA) were performed in a conventional three-electrode cell using a microcomputer-controlled potentiostat/galvanostat (Solartron 1285) at room temperature. The saturated calomel electrode (SCE), glassy carbon, and prepared catalysts served as the reference, counter, and working electrodes, respectively. The ink solutions on the working electrodes, including our nanocatalysts, were prepared as described in our previous study [18]. The amount of Pt in these ink solutions was determined by a Leeman Lab inductively coupled plasma spectroscope (ICP). The alcohol solutions that were used in these experiments, were fresh and unused, in order to obtain reproducible and reliable results.

2.3 Catalyst preparation

All of the prepared catalysts are shown in Table 1 and were prepared by an ethanol–superhydride reduction method. In this method, superhydride and ethanol are used to reduce a mixture of 0.25 mmol (0.0808 g) of PtCl4 dissolved in small amount of anhydrous tetrahydrofuran and 0.25 mmol of surfactant. The observation of a brown color in the solution indicates the formation of the amine-stabilized platinum nanoparticles. Excess surfactants are removed by washing with ethanol, followed by drying in a vacuum. The prepared platinum nanoparticles and carbon XC-72, in a 1:10 ratio, were vigorously mixed in ethanol for 2 days. It was dried under vacuum at room temperature.

Table 1 Catalysts and surfactants
Full size table

3 Results and discussion

3.1 XRD, AFM, and TEM characterization of the prepared catalysts

All the prepared catalysts were characterized by XRD; this indicates the crystalline structure, orientation, lattice constants, and crystallite sizes. The X-ray patterns for our Pt catalysts are shown in Fig. 1. The diffraction peaks at 2θ = 39.90, 46.60, 67.50, and 81.20 can be attributed to the (111), (200), (220), and (311) planes of the face-centered cubic (fcc) crystal lattice of platinum. The lattice parameter (a Pt) value of the prepared catalysts was calculated as 3.922 Ǻ using the Pt (220) diffraction peak from the following equation [19]

$$\sin \,\theta = \frac{{\lambda \sqrt {h^{2} + k^{2} + l^{2} } }}{2a}\quad \left ({{\text{for}}\;{\text{a}}\;{\text{cubic}}\;{\text{structure}}} \right)$$

where h, k, and l are adjacent lattice planes, θ is the angle at the position of peak maximum, λ = the wavelength of X-ray used (1.54056 Å), and a is lattice parameter. The average crystallite sizes of the catalysts I, II, III, IV, V, and VI were calculated as 3.20, 3.40, 3.03, 3.25, 2.86, and 2.59 nm, respectively, using the full width half-maximum of the (220) peaks in the Scherrer equation [20]:

$$d \left( {\text{\AA}} \right) = \frac{k\lambda }{\beta \,\cos \,\theta }$$

where k = a coefficient (0.9), λ = the wavelength of the X-ray used (1.54056 Ǻ), β = the full width half-maximum of respective diffraction peak (rad), and θ = the angle at the position of peak maximum (rad). These data indicate that as the chain length of surfactant increases, the size of platinum crystallites decreases, which is in good agreement with previous studies [21–23]. Additionally, the size of the platinum crystallites that were prepared by tertiary amine surfactants is larger.

Fig. 1
figure 1

XRD patterns of catalysts I (a), II (b), III (c), IV (d), V (e), and VI (f)

Full size image

The particle size distributions and atomic lattice fringes of the prepared catalysts were analyzed using high-resolution electron microscopy (HRTEM). The electron micrograph and particle size histogram for catalyst II is shown in Fig. 2. The platinum nanoparticles are uniformly distributed on the carbon support and the average particle sizes were found to be ~3.30 ± 0.85, ~3.52 ± 0.70, ~3.10 ± 0.82, 3.33 ± 0.92, 3.00 ± 0.96, and ~2.70 ± 0.84 nm for catalysts I, II, III, IV, V, and VI, respectively. These results are in good agreement with the XRD data. Occasionally, a small number of large particles, ~50–100 nm in diameter, are observed. This is most probably due to the accumulation of the smaller individual particles. Atomic lattice fringes were also examined for all catalysts, e.g., left inset of Fig. 2. The spacing for the Pt (111) planes was found to be 0.227 nm, which is very close to nominal Pt (111) spacing of 0.228 nm [24, 25].

Fig. 2
figure 2

High-resolution transition electron micrograph and particle size histogram of catalyst II

Full size image

These catalysts were also analyzed by AFM to obtain height and lateral diameter distributions of the Pt nanoparticles. The AFM images, height, and lateral diameter distributions, for our catalysts are shown in Fig. 3a–c, respectively. The height of the particles are 3.35, 3.55, 3.23, 3.38, 3.02, and 2.76 nm for catalysts I, II, III, IV, V, and VI, respectively. The height of the particles of our catalysts is in good agreement with XRD and TEM results (Table 2). The lateral diameter of the Pt nanoparticles was found to be 48.5, 44.3, 47.4, 45.7, 45.4, and 42.3 nm for catalysts I, II, III, IV, V, and VI, respectively. These results are very different from the particle size determined by XRD or TEM and are most probably due to tip contamination and/or tip convolution. Tip deconvolution was calculated to try to improve the AFM lateral diameter data with the assistance of a deconvolution formula given below [26, 27]. The mean half angle of the tip, 17°θ, instead of the apex radius, is used because of the larger lateral dimensions of the particles compared to the tip radius.

$$r_{\text{c}} = \, r\left[ {{ \cos }\,\theta + \, ({ \cos }^{ 2} \theta + ( 1+ { \sin }\,\theta )( - 1+ ({ \tan }\,\theta /{ \cos }\,\theta ) + { \tan }^{ 2} \theta ))^{ 1/ 2} } \right],$$

where r and r c are the real particle diameter and the observed particle diameter, respectively.

Fig. 3
figure 3figure 3

a AFM images of catalysts. b Histogram of the height of the particles obtained from the AFM data. c Histogram of the lateral diameter of the particles obtained from the AFM data

Full size image
Table 2 Average crystallite platinum particles size determined by (a) X-ray line broadening, (b) transmission electron microscopy, (c) particle height, (d) observed lateral diameter, and (e) deconvoluted lateral diameter determined by AFM
Full size table

The deconvoluted lateral size of the particles is 37.3, 34.1, 36.5, 35.1, 34.9, and 32.5 nm for catalysts I, II, III, IV, V, and VI, respectively (Table 2). The large value of the ratio of deconvoluted lateral size to height (e c−1) may be associated to the aggregation of particles on the freshly cleaved smooth mica surface owing to capillary forces, and/or tip indentation.

3.2 XPS characterization of the prepared catalysts

The effects of the oxidation state of platinum on the catalytic activity of the prepared catalysts, was probed using XPS. For this purpose, the Pt 4f region of the spectrum was fitted using a Gaussian–Lorentzian method after background subtraction using Shirley’s method [28] and analyzed in terms of relative peak area and chemical shift of Pt. The Pt 4f spectra of the catalysts are shown in Fig. 4 and each spectrum consists of two pairs of doublets. The ratio of the 4f7/2 to 4f5/2 signals for all catalysts was 4:3, which is in good agreement with the literature [28, 29]. The more intense doublet at about 71.0 and 74.3 eV is a signature of metallic platinum [30, 31] and the other doublet at about 74.3 and 77.6 eV is most likely caused by a Pt(IV) species such as PtO2 and/or Pt(OH)4 [32–36] on the surface as shown in Table 3. The ratios of Pt (0) to Pt(IV) for the prepared catalysts were calculated from the relative peak area of the Pt 4f spectrum. Catalyst II has the highest metallic character (the ratio of Pt(0) to Pt(IV) is highest) potentially providing a higher electrochemical active area and improved catalytic performance. The Pt(0) peak position is also important in defining the metallic character of the platinum, i.e., the lower the binding energy of Pt(0) the higher the metallic character of platinum. Catalyst II therefore has the highest metallic character as it has the lowest Pt 4f peak position (70.7 eV).

Fig. 4
figure 4

Pt 4f electron spectra of all catalysts

Full size image
Table 3 Pt 4f7/2 core binding energies, eV, of the prepared catalysts
Full size table

3.3 Electrochemical activity of the prepared catalysts

The cyclic voltammogram of catalyst I, in 0.1 M HClO4, showing typical hydrogen and oxygen adsorption/desorption regions is shown in Fig. 5. The cyclic voltammograms of the other catalysts are very similar. The electrochemical active surface areas (ECSA) of the catalysts were also measured in 0.1 M HClO4 using the hydrogen desorption region, under an Ar atmosphere—before and during the measurement—to deaerate the electrolyte [37]. The ECSA (m2 g−1) were calculated using following formula:

$${\text{ECSA}} = \frac{{Q \;({\text{mC}})}}{{{\text{Pt }}\;{\text{loading }}\;{\text{on }}\;{\text{electrode }} \times 0.21 \,{\text{mC}}\,{\text{cm}}^{ - 2} }}$$

where Q is the electric charge for hydrogen desorption, and assumes a correspondence value of 0.21 mC cm−2 Pt, which indicates the charge required to oxidize a monolayer of H2 on platinum. The chemical surface areas (CSA) of these nanocatalysts were calculated using the following formula:

$${\text{CSA}} = \frac{{6 \times 10^{3} }}{\rho \times d}$$

where d is the mean Pt crystalline size in Ǻ (from the XRD results) and ρ is the density of Pt metal (21.4 g cm−2) [38]. Pt utilization efficiency of the catalysts is calculated using ECSA and CSA (% Pt utility = (ECSA/CSA) × 100) and all these data are shown in Table 4. In addition, the roughness factor, R f (m2 g−1 Pt cm−2), of all catalysts was calculated by using the real ECSA and the geometric area A g (cm2) [39] as in the following formula:

Fig. 5
figure 5

Cyclic voltammogram of catalyst I in 0.1 M HClO4 at room temperature. Scan rate is 50 mV s−1

Full size image
Table 4 The comparisons of particle size, ECSA, CSA, and Pt utilization for all prepared catalysts and a commercial Pt catalyst
Full size table
$$R_{\text{f}} = \frac{\text{ECSA}}{{A_{\text{g}} }}$$

From Table 4 it can be seen that in terms of ECSA, % Pt efficiency and R f values catalyst II > IV > I > III > V > VI. In other words, catalyst II has the highest active surface area, % Pt utility, and roughness factor, which gives an indication that it may have the highest catalytic performance.

When methanol, ethanol, 1-propanol, or 2-propanol was added to the 0.1 M HClO4 electrolyte solution the classical alcohol oxidation response was observed for all catalysts. Only the anodic parts of the CV are shown in Fig. 6a–d. The CV results indicate that catalyst II (stabilized by the tertiary amine with the shortest chain length surfactant, N,N-dimethyldecylamine) has the highest catalytic performance: ~1.95 A (mg Pt)−1 at 0.68 V for methanol, ~1.30 A (mg Pt)−1 at 0.65 V for ethanol, 0.82 A (mg Pt)−1 at 0.62 V for 1-propanol and 2.35 A (mg Pt)−1 at 0.45 V for 2-propanol (Table 5). The enhancement of the catalytic performance of this catalyst, compared to commercially available ETEK Pt catalyst, is 25.5, 49.6, 23.2, and 30.6 times with regards to the methanol, ethanol, 1-propanol, and 2-propanol oxidation reactions, respectively [40–46]. This highest catalytic performance can be explained high: Pt(0) to Pt(IV) ratio, ECSA, percent platinum utility, and roughness factor.

Fig. 6
figure 6

a Anodic part of the cyclic voltammogram of the catalysts in 0.1 M HClO4 + 0.5 M CH3OH at room temperature. Scan rate is 50 mV s−1. b Anodic part of the cyclic voltammogram of the catalysts in 0.1 M HClO4 + 0.5 M CH3CH2OH at room temperature. Scan rate is 50 mV s−1 c Anodic part of the cyclic voltammogram of the catalysts in 0.1 M HClO4 + 0.5 M 1-propanol at room temperature. Scan rate is 50 mV s−1. d Anodic part of the cyclic voltammogram of the catalysts in 0.1 M HClO4 + 0.5 M 2-propanol at room temperature. Scan rate is 50 mV s−1

Full size image
Table 5 The anodic peak potential and maximum currents for all prepared catalysts in methanol, ethanol, 1-propanol, and 2-propanol
Full size table

CA techniques were utilized in order to investigate the long-term stability of catalysts in the C1–C3 alcohol oxidations. The measurements were done at a potential of 0.6 V (vs. SCE) for 3600 s. Figure 7 shows the curves for the methanol oxidation reaction and is representative of the other alcohol oxidations. A rapid current decay was observed which indicates the poisoning of the electrocatalysts. Nevertheless, the CA results indicate that all catalysts have long-term stability, and catalyst II was found to be the most stable due to its higher oxidation current after a 3600 s; this is consistent with the CV data.

Fig. 7
figure 7

Chronoamperometric curves of all prepared catalysts towards the methanol oxidation at 0.6 V (vs. SCE) in 0.1 M HClO4 + 0.5 M CH3OH

Full size image

4 Conclusion

Carbon-supported platinum nanoparticle catalysts were synthesized, characterized and their catalytic activity towards C1–C3 alcohol oxidation determined. The surfactants 1-decylamine, N,N-dimethyldecylamine, 1-dodecylamine, N,N-dimethyldodecylamine, 1-hexadecylamine, 1-octadecylamine were used, for the first time, in the synthesis of these catalysts. The data indicates that the catalyst which is stabilized by a tertiary amine with a shorter chain is larger in size and has increased catalytic activity, i.e., catalyst II (N,N-dimethyldecylamine). This is most probably due to the high ECSA, percent platinum utility, roughness factor, and Pt(0) to Pt(IV) ratio. This most active catalyst (catalyst II) has activities of 25.5, 49.6, 23.2, and 30.6 times higher, than the commercially available catalyst, towards methanol, ethanol, 1-propanol, and 2-propanol oxidation reactions, respectively.