CNT nanoparticles in alkaline medium

Journal Pre-proof Glycerol electro-oxidation to dihydroxyacetone on phosphorousdoped Pd/CNT nanoparticles in alkaline medium

Muhammad Sheraz Ahmad, Sharanjit Singh, Chin Kui Cheng, Huei Ruey Ong, Hamidah Abdullah, Maksudur Rahman Khan, Suwimol Wongsakulphasatch PII:

S1566-7367(20)30040-6

DOI:

https://doi.org/10.1016/j.catcom.2020.105964

Reference:

CATCOM 105964

To appear in:

Catalysis Communications

Received date:

27 November 2019

Revised date:

8 February 2020

Accepted date:

17 February 2020

Please cite this article as: M.S. Ahmad, S. Singh, C.K. Cheng, et al., Glycerol electrooxidation to dihydroxyacetone on phosphorous-doped Pd/CNT nanoparticles in alkaline medium, Catalysis Communications (2020), https://doi.org/10.1016/ j.catcom.2020.105964

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2020 Published by Elsevier.

Journal Pre-proof

Glycerol electro-oxidation to dihydroxyacetone on phosphorous-doped Pd/CNT nanoparticles in alkaline medium Muhammad Sheraz Ahmad a, *, Sharanjit Singh b, Chin Kui Cheng a, c, **, Huei Ruey Ong d, Hamidah Abdullah

a

, Maksudur Rahman khana, Suwimol Wongsakulphasatche a

Faculty of Chemical and Natural Resources Engineering, Universiti Malaysia Pahang, Lebuhraya Tun Razak,

Innovative Catalysis Program, Key Laboratory of Organic Optoelectronics & Molecular Engineering,

oo

b

f

26300 Gambang, Pahang, Malaysia. Department of Chemistry, Tsinghua University, Beijing, 100084, China Centre of Excellence for advanced Research in fluid flow, Universiti Malaysia Pahang, Lebuhraya Tun Razak,

pr

c

26300 Gambang, Pahang, Malaysia.

Faculty of Engineering and Technology, DRB-HICOM University of Automotive Malaysia, 26607

e-

d

Pekan, Pahang, Malaysia. Center of Ecomaterials and Cleaner Technology, Department of Chemical Engineering, Faculty of

Pr

e

Engineering, King Mongkut's University of Technology North Bangkok, Bangkok 10800, Thailand

Jo u

Abstract

rn

al

Corresponding *E-mail: [email protected], **E-mail: [email protected]

In this communication report, a comparative study between P-doped Pd/CNT and bare Pd/CNT catalyst was carried out to decouple the effects of phosphorous-doping on electro-oxidation of glycerol. The initial characterization results suggested that Pd and its oxides were successfully incorporated within the pore channels of CNTs support for both catalysts by using hydrazine-assisted hydrothermal technique. The XPS results revealed that the amount of Pd2+ for bare Pd/CNT were 1.4 times higher than P-doped electrocatalysts (about 70.1% and 48.7%, respectively) which confirms that phosphorus facilitates the reduction of Pd2+ to metallic Pd (Pd0). The electrochemical results showed that the electrochemical surface area (392.22 m2 gPd-1) and current density (26 mA/cm2) for P-doped Pd/CNT catalyst were 2.84 and 1.6 times, respectively, higher than Pd/CNT catalyst. The Pdoped catalyst was found to suppress the formation of carbonaceous intermediates; thus, improved the glycerol oxidation reaction. Small quantities of deep oxidation side products such as mesoxalic acid ( < 2 % ) and tartronic acid (< 0.1 % ) were found along with the dihydroxyacetone (DHA), a major product of glycerol electrooxidation. The best performing catalyst exhibited 1.4 folds higher DHA selectivity (90.8%) compared to the Pd/CNT.

1

Journal Pre-proof

Keywords: P-doped Pd/CNT, electrocatalyst, glycerol electro-oxidation reaction, HPLC, dihydroxyacetone.

1. Introduction Recently, biodiesel has gained substantial interests due to the finite fossil fuel resources, leading to a simultaneous increase in the crude glycerol (a biodiesel waste that originates from transesterification of animal fat and vegetable oils) [1]. The projected production of biodiesel is 42 giga L by 2020. This will contribute to massive glycerol production as a by-product (100 g glycerol / 1 kg of biodiesel produced) [2]. The electro-oxidation of glycerol has been touted as the most promising pathway for value-added chemicals generation from biodiesel waste [3]. This reaction has been mainly studied over Pt and Au-based catalysts due to their high activity for oxidation of alcohols

oo

f

such as methanol, ethanol, and formic acid [4]. For example, Zhao et al. [5] studied the methanol electro-oxidation over electrodeposited Au-Pt alloy nanocatalysts. These nanaocatalyst materials showed superior catalytic activity as compared to Au-core Pt shell and Pt-core Au-shell at the lowest working potential of 0.25 V. In addition, it was

pr

suggested that the alloy formations enhanced the catalytic activity of both Pt (low activity) and Au (no activity for

e-

methanol in alkaline media). Recently, Li et al. [6] reported an efficient trimetallic CoPtAu catalyst for electrooxidation of formic acid, methanol, and ethanol. Their results further revealed that the high activity of ternary

Pr

catalyst is due to stability of Co through L10- structure, which facilitates the interaction among substrate molecules, and the active sites of the catalyst. Nevertheless, high cost, low availability, and rapid poisoning of active sites of existing catalysts are major impediments towards commercialization [7]. Some earlier studies found that the less

rn

[8, 9].

al

expensive and more abundant Pd based catalysts showed better performance towards electro-oxidation of alcohols Various palladium-based alloy catalysts such as Pd-Ag [10], Pd-Ni [11], Pd-Au [12], PdCuPb/C [13], and

Jo u

PdAu/C [14], have been investigated for electro-oxidation of alcohols. However, it was reported that the electrocatalytic activity and stability of the catalyst can be further enhanced by introducing promoter (heteroatoms) to the catalyst [15]. In this respect, Kulesza et al. [16] reported that the substitution of lattice sites with heteroatoms (N, Fe or Co) can remarkably improve the stabilization and electrocatalytic activity of graphene based catalysts for both alkaline as well as in acidic media. Likewise, Dioati et al. [17] studied the interaction between Pt with Ir, Rh and Ni, and also interaction of Ni with Ir and Au over carbon nitride support for ORR performance. The results revealed that the incorporation of group VI metals significantly amplified the onset potential (from ca. 300 to ca.780 to ca.900 mV) vs RHE, respectively, for Au, Ir and Pt. Moreover, their work further elaborated the role of secondary metals in promoting the selectivity of ORR reactions via 4-electrons system. In addition, Sun et al. doped phosphorous into Pd based catalyst for formic acid electro-oxidation. Their results revealed that the phosphorous significantly enhanced the activity of catalyst by facilitating better dispersion of palladium nanoparticles due to electronic effect [18]. More recently, it was found that the doping of carbon support with palladium can enhance the oxygen vacancies on catalyst surface. It contributed to the oxidative conversion of intermediate during ethanol oxidation reaction [19]. Similarly, another study reported that the addition of

2

Journal Pre-proof phosphorous has accelerated the oxidation of COads to CO2 in ethanol electro-oxidation via bifunctional mechanism [20]. So far, there is no study exists which discusses adequately about the effects of phosphorus (P) doping on electro-oxidation of glycerol. Therefore, in the present work, phosphorous-doped and bare Pd nanoparticles on MWCNT support were prepared using hydrothermal method and investigated for electro-oxidation of glycerol. The as prepared catalysts were characterised by XRD, XPS, FESEM and TEM for crystallography, oxidation states and particle size measurements. A strong interaction of P with Pd was achieved by carefully monitoring the catalyst preparation and was further confirmed by XPS analysis. The effect of P-doping into Pd/CNT was studied for electrochemical performance and influence on GOR products distribution was compared with bare Pd/CNT.

Synthesis of catalysts

pr

2.1.

oo

f

2. Experimental

The P-doped Pd/CNT catalysts were synthesized by using hydrazine reduction approach in a stainless-steel

e-

hydrothermal reactor. It is shown in supplementary data (Figure. S1). Briefly, 4.67 ml (0.1 mmol) of palladium precursor (Pd (NO3)2. H2O - Alfa-Aesar, USA, 99.9% Pd) aqueous solution was mixed with 0.45 g (aqueous

Pr

solution) of phosphorous pentoxide (Sigma-Aldrich, USA, 98 %). This was followed by sonication for 30 min at ambient temperature. Thereafter, hydrazine was added drop-wise to the solution and re-sonicated for 30 min. Then, 0.8 g of MWCNTs (Sigma-Aldrich, USA, 98% carbon basis) activated prior to experiment using HNO3 were

al

added to the mixture. Subsequently, the reduced mixture was transferred to a hydrothermal reactor that was kept at

rn

120 ºC for 5 h. The solid residue was washed with distilled water using centrifugation followed by vacuum-drying at 80 ○C. Finally, the dried sample was ground and stored for physicochemical and electrochemical studies.

Physiochemical Characterization

Jo u

2.2

To study the crystal structure of catalysts, XRD analysis was performed (Rigaku Miniflex diffractometer), with the X-ray generator (Cu Ka radiation = 1.54 Å) set at 30 kV and 15 mA. The crystallite size of the catalyst was calculated using the Scherer’s equation based on 111 diffracted peak. The elemental analysis was carried out using an inductively-coupled plasma-atomic emission

spectrometer (ICP-AES, 8300 Optima, Perkin Elmer). The sample preparation was carried out by using MILSTONE microwave digester by taking 35 mg of each catalyst into aqua regia solution for 40 min. After the digestion, the sample was diluted to 50 ml with 2 % HNO3 solution. An aliquot of sample was used to run in the ICP-OES. In addition, the weight percentage of carbon was measured by CHNOS analyser (Elemeter, vario Macro cube, 2012). To examine the morphology of electrocatalyst, TEM of the brand Tecnai G2 20S-TWIN, with accelerating voltage of 220 kV was used. The SEM-EDX analysis was conducted on a tabletop Hitachi TM 3030 for morphology and elemental mapping of the electrocatalysts. The X-ray photon spectroscopy (XPS) analysis was

3

Journal Pre-proof performed on a Versa probe II (ULVAC- OHI/PHI 5000) spectrometer, with Al K α source (E = 1486.6 eV) working at 15 kV, Epress = 117.400 eV, 20 V energy step, and 0.2 s per point acquisition time. The binding energies of the P 2p, Pd 3d were referred to C 1 s peak, at 284.5 eV (accuracy within ± 0.2 e V).

2.3.

Electrochemical characterization of catalysts

The cyclic voltammetry (CV) and chronoamperometry (CA) analyses were conducted in a microcomputercontrolled Potentiostat/Galvanostat NOVA 1.10. Electrochemical reaction was conducted in a three-electrode cell. The electrocatalyst was deposited on glassy carbon (GC) as a working electrode. A high surface area Pt wire as a counter electrode and Ag/AgCl | KCl (sat) was used as a reference electrode. The alkaline electrolyte KOH (0.5 M)

f

and glycerol (0.5 M)/KOH (0.5 M) were employed for electrochemical surface area (ESA) and GOR analyses.

oo

The CV curves were obtained by multiple-cycle scans with a sweep rate of 50 mVs-1 under oxygen-free condition

e-

pr

(N2 purging). The ESA for both catalysts were calculated using Equation (1).

Where S is the area calculated by integration of the total charge corresponding to the PdO reduction peak in CV

Pr

curve, m is the mass of the Pd/CNT catalysts (0.16 mg based on the 0.3 μL ink used), C is the charge density (424 µC/cm2) associated with the reduction of the formed PdO monolayer and V is the scan rate used for current sweep. The working electrode is prepared by pasting the catalyst ink on glassy carbon. In a typical procedure, a

al

22 mg of catalyst powder was mixed with 140 µl of Nafion solution (5 wt. %) and 280 mL of isopropyl alcohol (99.5%), followed by sonication for 30 min. For each experiments, 0.3 µL of resulting ink was transferred onto

rn

0.785 cm2 working electrode, which is equal to 0.1571 mg of Pd on electrode tip. The electrode was finally dried

Jo u

for 1 h at 80 °C. Prior to each runs, the electrode surface was cleaned by cyclizing through (- 0.8 to - 0.2 V) potential range vs Ag/AgCl for several times until a stable voltammograms was recorded. To calculate the ESA of the catalyst, the electrolyte solution (0.5 M KOH) was bubbled with N2 for 30 min for oxygen removal. Finally, the glycerol was introduced to the cell and CV was conducted within potential window of -0.8 to -0.2 V @ 50 mVs-1. The stability was measured by performing CA analysis at -0.13 V and -0.05 V for 1800 s for P-doped Pd/CNT and Pd/CNT, respectively, and samples were taken for subsequent high performance liquid chromatography (HPLC) analysis for separation and quantification of GOR products.

2.4.

High performance liquid chromatography (HPLC)

The HPLC (Agilent 1260 series) instrument, equipped with a diode-array detector (DAD) and refractive index detector (RID), was used to analyse the composition of products after GOR. The HPLC column was purchased from Phenomenex, Model: Rezex ROA®, having dimensions of 4.6 mm × 250 mm, 5 µm, flow rate of 0.5 ml/min, wavelength of 210 nm and column and detector temperature of 75 °C and 35 °C, respectively. The column was heated to 65 °C and the products were separated using 3 mmol/L H2SO4 as an eluent. Authentic standards of

4

Journal Pre-proof tartronic acid (TAT), mesoxalic acid (MOXA), glyceraldehyde (GALD), glyceric acid (GLY), dihydroxyacetone (DHA) and oxalic acid (OXA) were used for the quantification of reaction products. Calibration curves for each standard were taken by replicating the analyses for at least three times and shown in supporting data S7. The selectivity of the products obtained from GOR was calculated using Equation (2).

where nOA is the number of moles of the individual product (organic acid = OA) and ntotal is the total number of moles of all the products (organic acids) produced during GOR.

oo

3.1

f

3. Results and discussion XRD analysis

pr

The XRD pattern of Pd/CNT and P-doped Pd/CNT are shown in Figure. 1. A sharp diffraction peak at 2θ of 26.2°

e-

which corresponds to strong diffraction from (002) plane for graphitic carbon can be clearly seen in all the samples. In addition, the three diffraction peaks at 2θ of 39.91º, 46.19° and 67.59° corresponding to (111), (200), and (220)

Pr

lattice planes, respectively, can be assigned to the face centred cubic crystal structure of palladium (JCPDS 050681). However, the diffraction peaks of phosphorous cannot be observed, which can be ascribed either due to the insufficient amount of phosphorous or due to displacements of lattice positions [21]. Furthermore, a peak shift

al

from higher degree region (40.16°) to lower degree region (39.91°) was observed for P-doped Pd/CNT for (111)

rn

lattice plan (inset Figure. 1). This can be due to the lattice size expansion of Pd after the incorporation of P, which also led to the peak displacement [22]. The larger crystallite size (6.48 nm) for P-doped Pd/CNT as compared to

Jo u

Pd/CNT (5.41 nm) was observed, revealing the alloy formation between P and Pd/CNT. In one of the studies by Zhang et al., phosphorous was doped in Pd nanoparticles also showed similar shift to lower degree region after Pdoping due to the partial substitution of the Pd with P, hence confirming the alloy formation [23].

5

pr

oo

f

Journal Pre-proof

XPS analysis

Pr

3.2

e-

Figure. 1: XRD patterns of (a) P-doped Pd/CNT, and (b) Pd/CNT catalysts.

Figure. 2 represents XPS spectra of P-doped Pd/CNT while the XPS spectra for Pd/CNT catalyst is presented in

al

Figure. S2. The XPS survey scan shows the presence of P, Pd and C elements in P-doped Pd/CNT catalyst

rn

(Figure. 2 (a) – (b)). The detailed spectra of P2p (Figure.2a) found two distinct peaks at 133.7 eV and 130.1 eV corresponds to the PV, P0. The first peak represents the surface oxidized phosphorous species, which originates

Jo u

from the dissociation of P2O5, whereas the comparatively smaller peak at 130.1 eV indicates the incorporation of phosphorous into palladium (P-Pd) [22]. The binding energy of nominal P, in nanoalloys shifts negatively by 0.3 eV from that of red phosphorous (130.4 eV). This shifting favours P0 to withdraw electrons density from surrounding donor metals (Pd in our work). A similar shifting was also reported for P with Pd and Cr [24].

6

oo

f

Journal Pre-proof

Figure 2: XPS spectra of P-doped Pd/CNT: (a) P d3d, and (b) P2p.

pr

For both doped and undoped catalysts, two pairs of doublet with spin orbit separation of 5.3 eV were observed for Pd 3d (Figure (2b) & (S2)). The Pd 3d spectrum of P-doped Pd/CNT found two deconvoluted peaks

e-

at binding energy of 342.4 eV (Pd3/2) and 337.1 eV (Pd5/2), correspond to the PdII whilst the peaks at 341.2 eV

Pr

(Pd3/2) and 335.9 eV (Pd5/2) represent the metallic Pd (Pd0) [25] (Figure. 2(b)). In Pd/CNT, the peaks 340.7 eV (Pd3/2) and 335.4 eV (Pd5/2) correspond to (Pd0), whilst the peaks at 341.9 eV (Pd3/2) and 336.6 eV (Pd5/2) correspond to PdII. Interestingly, the peaks for both PdII and Pd0 were negatively shifted (0.5 eV) for bare Pd/CNT

al

catalyst. This is due to the fact that the phosphorous withdraws electron from Pd [26] and hence decreased the electron density of 3d orbital in P-doped Pd/CNT (Figure. S2). The downshifting in d-band center of Pd weakened

rn

the reactants and intermediates (CO, CO2) adsorption over catalyst surface. Consequently, this favours the electro-

Jo u

oxidation of glycerol (GOR). The modification in Pd electronic structure changed the reaction species adsorption/desorption capability in P-dopedPd/CNT catalyst which mainly affect the catalytic activity of the catalyst. This also confirms the incorporation of P in Pd which is in line with literature [27]. In addition, the fraction of Pd0 for P-doped-Pd/CNT and bare Pd/CNT catalyst was higher than PdII, revealing that most of Pd exist as Pd0, which confirms that the incorporation of phosphorus facilitates the reduction of Pd2+ to metallic Pd (Pd0) [9].

3.3

Surface morphology and chemical composition TEM images and EDX of Pd/CNT and P-doped Pd/CNT catalysts are shown in Figure. 3 and (Fig S4),

respectively. It can be clearly seen that the Pd nanoparticles were encapsulated inside the channels of MWCNTs and possess uniform distribution for both catalysts (Figure. 3). The average particle size was 2.99 nm for Pd/CNT and 3.7 nm for P-doped Pd/CNT as calculated by using the Image J software. However, the particle size calculated from TEM was larger than XRD analysis. The particle size was higher for P-doped Pd/CNT than Pd/CNT catalyst. It is possibly due to incorporation of phosphorous in Pd, whereas P was not detected in the TEM images. The

7

Journal Pre-proof values of d-spacing increased with the doping of phosphorus in Pd/CNT catalyst, which further confirms the

oo

f

incorporation of P in Pd/CNT lattice (inset Figure.3).

pr

Figure 3: (a) TEM images of Pd/CNT, and (b) P-doped Pd/CNT.

For compositional analysis, the EDX analysis was conducted on P-doped Pd/CNT catalyst (Figure. S4).

e-

The percentage composition of 31.9 % and 27.3 % for P and Pd, respectively, was detected. The results not only

Pr

confirms the presence of phosphorus, but also in-line with the employed formulation. The elemental compositions were further confirmed by using ICP-OES and CHNOS analysis. The results obtained are presented in the Table S1. The results are in close agreement with the amounts of metals used for preparation and with EDX analysis. It

al

is noteworthy that P was distributed on the lattice sites of Pd and encapsulated in the pore channels of MWCNT as shown in Figure. 3. In addition, the decrease in concentration of Pd in P-doped catalyst (10.30 %) compared to

rn

Pd/CNT (15.87 %) evinced the incorporation of P onto palladium lattice sites covering the Pd sites. Moreover, the

Jo u

change in the overall concentration of carbon (C) in both the catalysts remarkably changed after the addition of P.

3.4 Electrochemical characterization 3.4.1 Cyclic Voltammetry

The cyclic voltammetry curves (CV) of undoped Pd/CNT and P-doped Pd/CNT catalysts are shown in Figure. 4. The ESA calculation was carried out by taking the CV curves in 0.5 M KOH (Figure 4 (inset)), while glycerol-oxidation reaction (GOR) studies were carried out in 0.5 M glycerol/0.5 M KOH (Figure 4 (main)). It can be seen from Figure 4 (inset), during forward sweep, the Pd/CNT catalyst shows one shoulder (I) peak in the potential range -0.8 V to -0.6 V which can be ascribed to adsorption/desorption of hydrogen on the surface of Pd. Second shoulder (II) peak found in between -0.4 V to 0.0 V and last peak (III) in potential range 0.0 V to 0.2 V which can be linked to PdO formation on Pd surface. For the backward scan, there are more pronounced reduction peaks (IV), between -0.4 V to -0.1 V which can be ascribed to reduction of palladium oxides produced during forward scan [28]. Moreover, the reduction potential of the Pd/CNT has shifted (0.02 V) to the higher region with

8

Journal Pre-proof the doping of phosphorous (P). This demonstrates that the P-doped Pd/CNT catalysts exhibited weaker oxophilicity and thus lower hydroxyl surface coverage. The results are in consensus with XPS analysis, in which a reduced oxide state was observed for P-doped Pd/CNT compared to Pd/CNT catalyst [29]. The electrochemically active surface area (ESA) for P-doped Pd/CNT (392.22 m2 gPd-1) was found to be 2.84 times higher than the Pd/CNT (137.74 m2 gPd-1). The GOR activity of catalysts is analysed in 0.5 M glycerol/0.5 M KOH at a scan rate of 50 mV/s and presented in Figure. 4 (main). It was found that the current density of P-doped Pd/CNT (26 mA/cm2) was 1.6 folds higher than the bare Pd/CNT (16 mA/cm2). In addition, during GOR only anodic peaks are founds in both forward and backward sweeps, which revealed that the electro-oxidation of glycerol followed an irreversible process. Besides, the onset potential for P-doped Pd/CNT was slightly lower than the Pd/CNT catalyst. The high ESA and

oo

f

current density for P-doped Pd/CNT signifying the superior electrocatalytic activity of the catalyst for GOR, can be attributed to synergetic effect of P and Pd [30]. In addition, mass activities were also measured for both catalysts. The P-doped Pd/CNT showed 1.16 folds higher mass activity compared to Pd/CNT. The values are presented in

pr

Table S1. This showed an enhanced GOR activity for P-doped Pd/CNT, which can be attributed to the P-doping. It decreased the 3d electron density of Pd and favoured desorption of reaction intermediates (CO /CO2) as well as

e-

synergetic effects of Pd and P [31].

Pr

Figure. S5 (a) represents CV plots at different scan rates for P-doped Pd/CNT in the range of 10 - 100 mV -1

s . The results disclosed that the positive shifting for potential was noticed by increasing scan rate, while the peak current density was also correspondingly increased with scan rate. Besides, a linear trend [7] was found for the

al

square root of scan rate as well as forward (If) and backward current (Ib). The results obtained suggested that the GOR followed irreversible diffusion controlled process [32]. The operational stability of P-doped Pd/CNT is

rn

shown in Figure. S5 (b). It can be clearly seen that the activity of P-doped Pd/CNT catalyst has slightly decreased

Jo u

at the end of 500 cycles, which indicate the high stability of the catalyst.

9

Pr

e-

pr

oo

f

Journal Pre-proof

Figure. 4: Electrochemical study of Study of Pd/CNT and P-dopedPd/CNT: CV in 0.5 M glycerol /0.5 M

Products Analysis by HPLC

rn

3.6

al

KOH @ 50 mV/s, and ESA in 0.5 M KOH (inset).

Jo u

Figure. S6 depicts the chronoamperometry curves of Pd/CNT and P-doped Pd/CNT catalysts after 30 min of GOR in 0.5 M glycerol/0.5 M KOH at 0.05 V and -0.13 V, respectively. The current densities of both the catalysts decreased owing to the poisoning effect of adsorbed reaction intermediates, such as CO, during GOR [33]. The current density of P-doped Pd/CNT catalyst was found to be 1.7 times greater than Pd/CNT, confirming the superior activity for P-doped Pd/CNT catalyst. The liquid samples for product analysis were also collected using chronoamperometry and analysed using HPLC. The product distribution for both the catalysts is shown in Figure. 5. The concentrations (mg/L) of the products obtained through glycerol electro-oxidation reaction are presented in Table S1. All the product are obtained in their salt form due to reaction occurring in alkaline medium. For simplicity however, it was stated in acid form. Five main products that were separated include mesoxalic acid (MOXA), dihydroxyacetone (DHA), glyceraldehyde (GALD), tartronic acid (TAT) and oxalic acid (OXA). Interestingly, formic acid was not detected among the products from both sets of catalysts. According to the recent works by Ahmad et al. [34], the formic acid converted to carbonaceous intermediates (CO and CO2) following the easiest route for CO formation. The results found that the P-doping did not change the reaction pathway whilst the yield of desired products

10

Journal Pre-proof (dihydroxyacetone and mesoxalic acid) was remarkably enhanced for P-doped Pd/CNT catalyst. The final concentration (mg/L) of DHA and MOXA was enhanced by15.91 and 1.32 times over P-doped Pd/CNT catalyst compared to undoped Pd/CNT. The selectivity was calculated for all the products using Equation (2) and the values are presented in Figure.5 (inset). It was found that P-doped Pd/CNT has selectivity of 90.82% and 7.88% for DHA and MOXA. This was 1.4 folds higher than DHA selectivity over the Pd/CNT. In contrast, the selectivity for MOXA was 2.7 times lower than Pd/CNT. The enhanced activity, higher concentration and better selectivity of as synthesized P-doped Pd/CNT can be attributed to doping of P. It is expected that the doping of P suppressed hydrogen and CO adsorption via decrease in 3d electron density of Pd which in turns increased the availability of the reduced Pd with more active sites resulting enhanced performance of GOR and this justification is in line with the results reported by Yang et al. [19]. In addition, the higher pH (12.35 in our work) for alkaline media also

Jo u

rn

al

Pr

e-

pr

oo

f

favourable for the deprotonation of H ions [35].

Figure. 5: P-doped Pd/CNT & Pd/CNT; (a) Product distribution, and selectivity (inset).

4. Conclusions

In the present work, P-doped Pd/CNT catalyst was successfully synthesized by hydrothermal reduction technique. The formation of face-centred cubic crystal structure of palladium was confirmed for both the P-doped Pd/CNT and bare Pd/CNT catalyst. It was observed that the phosphorus was successfully incorporated inside the channels of MWCTs and enhanced the reduction of Pd in P-doped Pd/CNT catalyst. This led to weaker oxophilicity and thus lower hydroxyl surface coverage. In addition, the ESA, current density and selectivity were also significantly improved for P-doped Pd/CNT. The interaction between P and Pd/CNT suppressed the formation of intermediates and increased the OH-1 concentration on Pd surface, which eventually improved the activity and stability of Pdoped Pd/CNT for GOR.

11

Journal Pre-proof Acknowledgements We would like to acknowledge Universiti Malaysia Pahang, (PGRS180335) for financial support. Chin Kui Cheng acknowledges Ministry of Education, Malaysia for the Trans-Disciplinary Research Grant Scheme with vot. no. of RDU191802-1.

References

Jo u

rn

al

Pr

e-

pr

oo

f

[1] Y.T. Algoufi, G. Kabir, B.H. Hameed, Synthesis of glycerol carbonate from biodiesel by-product glycerol over calcined dolomite, Journal of the Taiwan Institute of Chemical Engineers, 70 (2017) 179187. [2] P.U. Okoye, A.Z. Abdullah, B.H. Hameed, Glycerol carbonate synthesis from glycerol and dimethyl carbonate using trisodium phosphate, Journal of the Taiwan Institute of Chemical Engineers, 68 (2016) 51-58. [3] M.S. Ahmad, C.K. Cheng, R. Kumar, S. Singh, K.A. Saeed, H.R. Ong, H. Abdullah, M.R. Khan, Pd/CNT Catalysts for Glycerol Electro-oxidation: Effect of Pd Loading on Production of Valuable Chemical Products, Electroanalysis, (2020). [4] X.G. Sun, K.W. Gao, X.L. Pang, H.S. Yang, A.A. Volinsky, Electrochemical Oxidation of Methanol on Pt-SnOx/C Catalysts Characterized by Electrochemistry Methods, Journal of the Electrochemical Society, 162 (2015) F1540-F1548. [5] L. Zhao, J.P. Thomas, N.F. Heinig, M. Abd-Ellah, X. Wang, K. Leung, Au–Pt alloy nanocatalysts for electro-oxidation of methanol and their application for fast-response non-enzymatic alcohol sensing, Journal of Materials Chemistry C, 2 (2014) 2707-2714. [6] J. Li, S.Z. Jilani, H. Lin, X. Liu, K. Wei, Y. Jia, P. Zhang, M. Chi, Y.J. Tong, Z. Xi, Ternary CoPtAu Nanoparticles as a General Catalyst for Highly Efficient Electro‐ Oxidation of Liquid Fuels, Angewandte Chemie, 58 (2019) 11527-11533. [7] M. Wang, M. Chen, Z.Y. Yang, Y.T. Wang, Y.R. Wang, G.C. Liu, J.K. Lee, X.D. Wang, A study on fuel additive of methanol for room temperature direct methanol fuel cells, Energy Conversion and Management, 168 (2018) 270-275. [8] A. Reina, A. Serrano-Maldonado, E. Teuma, E. Martin, M. Gomez, Palladium nanocatalysts in glycerol: Tuning the reactivity by effect of the stabilizer, Catalysis Communications, 104 (2018) 22-27. [9] K. Wu, X.B. Mao, Y. Liang, Y. Chen, Y.W. Tang, Y.M. Zhou, J. Lin, C.N. Ma, T.H. Lu, Multiwalled carbon nanotubes supported palladium-phosphorus nanoparticles for ethanol electrooxidation in alkaline solution, Journal of Power Sources, 219 (2012) 258-262. [10] J.P. Liu, H.H. Zhou, Q.Q. Wang, F.Y. Zeng, Y.F. Kuang, Reduced graphene oxide supported palladium-silver bimetallic nanoparticles for ethanol electro-oxidation in alkaline media, Journal of Materials Science, 47 (2012) 2188-2194. [11] F.J. Miao, B.R. Tao, Methanol and ethanol electrooxidation at 3D ordered silicon microchannel plates electrode modified with nickel-palladium nanoparticles in alkaline, Electrochimica Acta, 56 (2011) 6709-6714. [12] L.D. Zhu, T.S. Zhao, J.B. Xu, Z.X. Liang, Preparation and characterization of carbon-supported sub-monolayer palladium decorated gold nanoparticles for the electro-oxidation of ethanol in alkaline media, Journal of Power Sources, 187 (2009) 80-84. [13] C.H. Liu, X.L. Cai, J.S. Wang, J. Liu, A. Riese, Z.D. Chen, X.L. Sun, S.D. Wang, One-step synthesis of AuPd alloy nanoparticles on graphene as a stable catalyst for ethanol electro-oxidation, International Journal of Hydrogen Energy, 41 (2016) 13476-13484.

12

Journal Pre-proof

Jo u

rn

al

Pr

e-

pr

oo

f

[14] Y.Y. Huang, Y.L. Guo, Y.B. Wang, J.N. Yao, Synthesis and performance of a novel PdCuPb/C nanocatalyst for ethanol electrooxidation in alkaline medium, International Journal of Hydrogen Energy, 39 (2014) 4274-4281. [15] Y.J. Dong, W. Wang, Y.H. Wang, P.D. Wang, L.W. Zhang, Z.Q. Lei, PdAu nanoparticles anchored on P and Se codoped carbon support as an efficacious electrocatalyst towards glycerol electrooxidation, Journal of the Taiwan Institute of Chemical Engineers, 93 (2018) 500-508. [16] P.J. Kulesza, J.K. Zak, I.A. Rutkowska, B. Dembinska, S. Zoladek, K. Miecznikowski, E. Negro, V. Di Noto, P. Zelenay, Elucidation of role of graphene in catalytic designs for electroreduction of oxygen, Current Opinion in Electrochemistry, 9 (2018) 257-264. [17] S. Diodati, E. Negro, K. Vezzù, V. Di Noto, S. Gross, Oxygen reduction reaction and X-ray photoelectron spectroscopy characterisation of carbon nitride-supported bimetallic electrocatalysts, Electrochimica Acta, 215 (2016) 398-409. [18] H.J. Sun, J.F. Xu, G.T. Fu, X.B. Mao, L. Zhang, Y. Chen, Y.M. Zhou, T.H. Lu, Y.W. Tang, Preparation of highly dispersed palladium-phosphorus nanoparticles and its electrocatalytic performance for formic acid electrooxidation, Electrochimica Acta, 59 (2012) 279-283. [19] G.X. Yang, Y. Chen, Y.M. Zhou, Y.W. Tang, T.H. Lu, Preparation of carbon supported Pd-P catalyst with high content of element phosphorus and its electrocatalytic performance for formic acid oxidation, Electrochemistry Communications, 12 (2010) 492-495. [20] J.C.M. Silva, I.C. de Freitas, A.O. Neto, E.V. Spinacé, V.A. Ribeiro, Palladium nanoparticles supported on phosphorus-doped carbon for ethanol electro-oxidation in alkaline media, Ionics, (2017) 19. [21] R. Jiang, D.T. Tran, J.P. McClure, D. Chu, A class of (Pd–Ni–P) electrocatalysts for the ethanol oxidation reaction in alkaline media, Acs Catalysis, 4 (2014) 2577-2586. [22] G. Yang, Y. Chen, Y. Zhou, Y. Tang, T. Lu, Preparation of carbon supported Pd–P catalyst with high content of element phosphorus and its electrocatalytic performance for formic acid oxidation, Electrochemistry Communications, 12 (2010) 492-495. [23] K. Zhang, C.Q. Wang, D. Bin, J. Wang, B. Yan, Y. Shiraishi, Y.K. Du, Fabrication of Pd/P nanoparticle networks with high activity for methanol oxidation, Catalysis Science & Technology, 6 (2016) 6441-6447. [24] J. Wu, Z.R. Yang, X.W. Li, Q.J. Sun, C. Jin, P. Strasser, R.Z. Yang, Phosphorus-doped porous carbons as efficient electrocatalysts for oxygen reduction, Journal of Materials Chemistry A, 1 (2013) 9889-9896. [25] R.K. Singh, R. Rahul, M. Neergat, Stability issues in Pd-based catalysts: the role of surface Pt in improving the stability and oxygen reduction reaction (ORR) activity, Physical chemistry chemical physics : PCCP, 15 (2013) 13044-13051. [26] Z.L. Xin, S.H. Wang, J. Wang, X. Huang, X.B. Ji, Y.Y. Yao, L.D. Shao, Nanosizing low-loading Pd on phosphorus-doped carbon nanotubes for enhanced HCOOH oxidation performance, Electrochemistry Communications, 67 (2016) 26-30. [27] Y. Wang, F.-F. Shi, Y.-Y. Yang, W.-B. Cai, Carbon supported Pd–Ni–P nanoalloy as an efficient catalyst for ethanol electro-oxidation in alkaline media, Journal of Power Sources, 243 (2013) 369-373. [28] J.B. Xu, T.S. Zhao, Y.S. Li, W.W. Yang, Synthesis and characterization of the Au-modified Pd cathode catalyst for alkaline direct ethanol fuel cells, International Journal of Hydrogen Energy, 35 (2010) 9693-9700. [29] J.D. Li, Q.F. Tian, S.Y. Jiang, Y. Zhang, Y.X. Wu, Electrocatalytic performances of phosphorus doped carbon supported Pd towards formic acid oxidation, Electrochimica Acta, 213 (2016) 21-30. [30] E. Leal da Silva, A. Cuña, M. Rita Ortega Vega, C. Radtke, G. Machado, N. Tancredi, C. de Fraga Malfatti, Influence of the support on PtSn electrocatalysts behavior: Ethanol electro-oxidation performance and in-situ ATR-FTIRS studies, Applied Catalysis B: Environmental, 193 (2016) 170-179.

13

Journal Pre-proof

Jo u

rn

al

Pr

e-

pr

oo

f

[31] J. Zhang, Y. Xu, B. Zhang, Facile synthesis of 3D Pd–P nanoparticle networks with enhanced electrocatalytic performance towards formic acid electrooxidation, Chemical Communications, 50 (2014) 13451-13453. [32] W. Hong, C.S. Shang, J. Wang, E.K. Wang, Bimetallic PdPt nanowire networks with enhanced electrocatalytic activity for ethylene glycol and glycerol oxidation, Energy & Environmental Science, 8 (2015) 2910-2915. [33] T.B. Zhou, H. Wang, J. Key, S. Ji, V. Linkov, R.F. Wang, Highly dispersed ultrafine Pt nanoparticles on hydrophilic N-doped carbon tubes for improved methanol oxidation, Rsc Advances, 3 (2013) 16949-16953. [34] M.S. Ahmad, C.K. Cheng, H.R. Ong, H. Abdullah, C.S. Hong, G.K. Chua, P. Mishra, M.M. Rahman Khan, Electro-oxidation of waste glycerol to tartronic acid over Pt/CNT nanocatalyst: study of effect of reaction time on product distribution, Energy Sources, Part A: Recovery, Utilization, and Environmental Effects, (2019) 1-17. [35] Y. Kwon, S.C. Lai, P. Rodriguez, M.T. Koper, Electrocatalytic oxidation of alcohols on gold in alkaline media: base or gold catalysis?, J Am Chem Soc, 133 (2011) 6914-6917.

14

Journal Pre-proof Highlights 

Phosphorus doping facilitated the reduction of Pd2+ to Pd.



ESA and current density increased 2.84 and 1.4 times, respectively for P-doped electrocatalysts compare to counterpart.

rn

al

Pr

e-

pr

oo

f

A high DHA selectivity of 90.8% was achieved for P-doped electrocatalyst.

Jo u



15