Synthesis and Application of Novel NiMoK/TS-1 for Selective Conversion of Fatty Acid Methyl Esters/Triglycerides to Olefins

Abstract

Methyl palmitate (or triglyceride) was converted into C₁₅ olefin with remarkable selectivity using nickel–molybdenum oxides on the mesoporous titanosilicate support. The olefin has one carbon atom less than the acid portion of the ester. A new catalyst NiMoK/TS-1 was synthesized in which the effect of acidity of supports and molybdenum loading on the decarboxylation conversion along with product selectivity was investigated in methyl palmitate conversion into C₁₅ olefin. The prepared catalysts were analyzed using ammonia-temperature-programmed desorption (NH₃-TPD), thermogravimetric analysis (TGA), X-ray photoelectron spectroscopy (XPS), X-ray diffraction (XRD), Fourier transform infrared (FTIR) spectroscopy, scanning electron microscopy (SEM), transmission electron microscopy (TEM), and Brunauer–Emmett–Teller (BET) techniques. The reaction was carried out using a vapor-phase fixed-bed downflow reactor system at atmospheric pressure. The NiMoK/TS-1 catalyst at a weight hourly space velocity (WHSV) of 5.6/h was found to be selective toward C₁₅ olefin. The catalyst was stable up to 15 h, and it can be regenerated with no considerable decrease in the activity even after fourth reuse. Beyond 653 K, the conversion of methyl palmitate increased but the selectivity for C₁₅ products and C₁₅ olefin was decreased.

1. Introduction

The use of biomass to produce fuels is required to fulfill the energy demands of the world today, due to the high industrialization and urbanization and the rapid depletion of nonrenewable resources. The major part of the world’s energy is consumed by the transportation sector due to the high amount of fuel requirements, which is nearly unavoidable. Recent studies are therefore directed toward the synthesis of biofuels from renewable resources. Naturally occurring oils and fats composed of C₁₂–C₂₄ saturated/unsaturated fatty acid esters can readily reduce the energy crisis and maintain the ecological carbon balance. Biofuels are categorized as first-, second-, third-, or fourth-generation fuels depending on the feedstock from which they are produced. Biofuels made from food crops and manufactured from various biomass derived from different inedible plants and animal materials fall under the first- and second-generation categories.¹ Fatty acid methyl esters (FAME) biodiesels, which are also called as first-generation biofuels, help in decreasing greenhouse gas emissions, decreasing the dependence on fossil fuels, and improving rural agricultural economies. On the other hand, FAME biofuels manufactured from vegetable oils have disadvantages like high freezing point, high cloud point, high viscosity, high oxygen content, and low calorific value.² C₁₅–C₁₈ hydrocarbons produced from saturated/unsaturated fatty acid esters can be divided into first- or second-generation biofuels depending on edible or inedible feedstock used to produce them. These hydrocarbons with high cetane values are used as additives for blending with biodiesels to improve their properties.³ The naturally occurring triglycerides or fatty acids or fatty acid methyl esters undergo the following catalytic deoxygenation reactions: (1) hydrodeoxygenation (HDO) exothermic reaction with removal of oxygen in the form of water, leaving behind n-alkane with the same carbon number as the corresponding fatty acid, (2) decarbonylation (DCO) and decarboxylation (DCO₂) endothermic reactions with elimination of oxygen in the form of CO and water or CO₂, respectively, to give hydrocarbons with one carbon atom less compared to the original fatty acid (Scheme 1).

Scheme 1. Schematic Representation of HDO, DCO, and DCO₂ Reactions.

For the past few decades, various heterogeneous catalysts are being studied for this deoxygenation, where all HDO, DCO, and DCO₂ reactions occur simultaneously in the presence of hydrogen to yield paraffins and olefins. Temperature, pressure, and H₂/substrate ratio affect these reaction pathways. High pressure, H₂/substrate ratio, and low temperatures are favorable for HDO product selectivity.⁴ Sulfided NiMo/γ-alumina is a widely studied catalyst for hydrodeoxygenation, which resulted in around 40% selectivity of HDO products.⁴⁻⁷ DCO₂ and DCO products were obtained parallelly with 10% olefin selectivity.⁸ Ni or NiMo supported on other more acidic supports such as AlSBA-15 and SAPO-11 were studied to yield iso-alkanes.⁹⁻¹¹ CoMo over alumina and MCM-42 were reported to convert rapeseed oil into hydrocarbons by hydrodeoxygenation and hydrodecarboxylation mechanisms.^12,13 Ni/ZrO₂, Ni/ZnO–Al₂O₃, and NiAl₂O₄ spinels have shown good conversion of fatty acids into straight-chain alkanes with different alkane selectivities.¹⁴⁻¹⁶ Precious metals like Pt, Pd, and Ru were studied to improve the selectivity and produce identical alkanes. Platinum supported over alumina or SAPO-11 was reported for deoxygenation of oleic acid with a maximum 35% selectivity of n-heptadecane.¹⁷ Ru/ZSM-5 was used for hydrodeoxygenation of methyl stearate giving 64.3% of n-C₁₇ and 13% of n-C₁₈ alkanes.¹⁸ 5% Pd/C is reported to give the highest conversion of around 100% and selectivity of around 97% for n-heptadecane from stearic acid and n-pentadecane from palmitic acid in the presence of hydrogen.¹⁹⁻²¹ In all of these studies, a huge amount of hydrogen (around 3 moles per mole of feedstock) is used for deoxygenation.

Though HDO has many advantages, this high-pressure H₂ operation seems to be undesirable because it requires high H₂ consumption and expensive facilities to make the process explosive-free and safe. DCO₂, on the other hand, can be produced without hydrogen, and unlike hydrodeoxygenation, it does not produce water, which may deactivate the catalyst. Some of the research works by Murzin et al.^19,20,22,23 studied decarboxylation of stearic acid, oleic acid, or methyl oleate in an inert atmosphere without H₂, which ended up in an increased amount of olefins in the product stream. Na et al.²⁴ introduced hydrotalcites as decarboxylation catalyst, where almost 98% of oleic acid was converted into C₁₇ products, but the selectivity toward n-heptadecene was not focused. Mäki-Arvela et al.²² yielded heptadecene with 51% selectivity during the decarboxylation of methyl stearate. Chiappero et al.²⁵ thoroughly studied the decarboxylation reaction of methyl octanoate and other triglycerides with the PtSnK/SiO₂ catalyst, without H₂ flow, in the He atmosphere yielding almost 70–80% olefin products. Stern et al.²⁶ patented a catalyst consisting of nickel and a metal selected from tin, germanium, and/or lead resulting in 71% selectivity of α-pentadecene obtained from decarboxylation of palmitic acid.

Selective conversion of fatty acid methyl esters or triglycerides to hydrocarbons, particularly α-olefins, is valuable because of their applications, not only as biofuels but also in producing specialty chemicals like polymers and surfactants. Therefore, in the present work, we have focused on the production of identical long-chain hydrocarbons from methyl esters and/or triglycerides via decarboxylation without the use of hydrogen using different supported NiMo catalysts to achieve high selectivity of olefins in particular. The Ni–Mo–O catalyst is selected as it can operate at a low conversion but with a high selectivity in oxidative dehydrogenation reactions (ODH) of alkanes.²⁷ Titanosilicate (TS-1) is explored as support material due to its optimum acidic and basic properties, which helped to control the cracking side products and hence to improve the selectivity of products.

2. Results and Discussion

2.1. Catalyst Characterization

2.1.1. Ammonia-Temperature-Programmed Desorption (NH₃-TPD)

The acidity of support plays an important role in the selectivity of product as acidity is prone to yield more cracking products. Ammonia TPD of all supports and catalysts, which are studied for this conversion were carried out (Figure 1). γ-Alumina has the highest acidity of 0.79 mmol/g, followed by TS (1:10) and TS-1. TS-1 has the least acidity of 0.121 mmol/g. Nickel and molybdenum oxides over TS-1 have an acidity of 0.202 mmol/g (Table S1). As doping of potassium decreases the acidic sites of the catalyst, the acidity is further reduced to 0.176 mmol/g.

Figure 1.

NH₃-TPD curves for different supports and catalysts studied.

2.1.2. X-ray Diffraction (XRD) Analysis

Figure 2 shows the XRD patterns of TS-1, NiMoK/TS-1, as well as regenerated NiMoK/TS-1. The XRD pattern of TS-1 showed diffraction peaks resembling the MFI topology with 2θ values of 7.9, 8.8, 23.0, 23.9, and 24.4 (Figure 2A).^28,29 Impregnation of nickel, molybdenum, and potassium has not changed the topology of the TS-1 (Figure 2B). No extra peaks of Ni or Mo oxides were observed though they were present in significant quantity. Little shifts in the peak positions as well as increase in peak intensities were observed in NiMoK/TS-1. The Ni and Mo oxide species are thus said to be well dispersed inside the pores of TS-1, which can be also supported by the transmission electron microscopy (TEM) image provided (Figure 8). Average crystallite sizes of TS-1 as well as NiMoK/TS-1 were calculated using the Scherrer formula to be 19.89 and 23.38 nm, respectively. The regenerated catalyst also retains its original phase with the said 2θ values as well as an average crystallite size of 23.32 nm.

Figure 2.

XRD spectra of TS-1 (A), NiMoK/TS-1 (B), as well as regenerated NiMoK/TS-1 (C).

Figure 8.

(i) TEM images of TS-1 (A) and NiMoK/TS-1 (B). (ii) TEM images of (A) TS-1 and (B) NiMoK/TS-1 to understand the particle sizes of TS-1 and Ni–Mo oxides.

2.1.3. Fourier Transform Infrared (FTIR) Spectroscopy

FTIR spectra of TS-1 and NiMoK/TS-1 were recorded (Figure 3). The Si–O bending mode and the stretching vibration of the structural double five-membered ring in the TS-1 framework are identified by bands at wavenumbers of 450 and 550 cm^–1, respectively. These two bands confirm the formation of the MFI topology in all samples, which matches with the XRD results. The band at 968 cm^–1 can possibly be assigned to stretching of Si–O–Ti, which can be considered as evidence of the presence of titanium in the TS-1 framework. Metal-oxide-loaded TS-1 does not have any extra peak but differs in intensities of peaks at 801 and 968 cm^–1. The relative intensities of the bands at 968 and 801 cm^–1 (I_968/801) were used to compare the relative contents of framework titanium in different catalyst samples.^29,30I_968/801 is higher (1.04) for TS-1 than for NiMoK/TS-1 (0.70).

Figure 3.

FTIR spectra of TS-1 (A) and Ni₅Mo₁₀K/TS-1 (B).

2.1.4. Surface Area Analysis

The N₂ adsorption–desorption method was used to find the textural properties of TS-1, NiMoK/TS-1, as well as regenerated NiMoK/TS-1 (Figure 4). The adoption curve obtained was elongated S-type and convex to the x-axis. This is type V isotherm indicating mesoporous materials produced by the unrestricted multilayer formation process with weak interactions between adsorbate and adsorbent. The Brunauer–Emmett–Teller (BET) surface area of TS-1 was 381.5 m²/g with a pore volume of 0.42 cm³/g and pore size of 10.5 nm, whereas NiMoK/TS-1 exhibited a BET surface area of 419.2 m²/g, a pore volume of 0.34 cm³/g, and a pore size of 6.9 nm. An increase in surface area can be attributed to the additional surfaces of Ni and Mo oxide particles. Similarly, the decrease in pore volume and pore size was due to the adsorption of Mo and Ni oxides into the pores.

Figure 4.

N₂ adsorption–desorption isotherms of TS-1 (A) and Ni₅Mo₁₀K/TS-1 (B).

2.1.5. X-ray Photoelectron Spectroscopy (XPS)

XPS analysis was performed to ensure the elemental composition, chemical state, and electronic state of the elements existing from the surface of the catalyst material.The high-resolution spectra of Si 2p in TS-1 had a binding energy of 102.4 eV before the impregnation of Ni and Mo oxides, but after the impregnation, the peak was shifted to 100.6 eV (Figure 5I). Similarly, the core spectra of Ti 2p in TS-1 (Figure 5II) showed the binding energies for Ti 2p_3/2 and Ti 2p_1/2 (457.7 and 463.4 eV, respectively). While after loading of Mo and Ni oxides, the curves moved slightly to lower binding energies of 455.9 and 462.2 eV, respectively. This confirmed that the bond strengths of Si–O and Ti–O were changed slightly due to the incorporation of metal oxides. The wide spectra of NiMoK/TS-1 (1:30) are shown in Figure 6. Mo 3d (Mo 3d_5/2) had the highest binding energies at 233 and 229.7 eV, respectively, corresponding to the oxidation states Mo⁺⁶ and Mo⁺⁴. The peaks between of 850 and 880 eV were attributed to Ni 2p, with multiplate splits at 853 and 871.2 eV for Ni 2p_3/2 of Ni⁺² with a satellite peak at 859.5 eV. A satellite peak of Ni 2p_1/2 at 878.7 eV was also observed. High-resolution spectra of Mo 3d and Ni 2p are provided in the Supporting Information (Figures S1 and S2). As potassium was added at a very low concentration, it could not be seen in the XPS spectra. Elemental quantification of Si, Ti, Mo, and Ni provided by the area under the curve method was found to be 27.72, 1.18, 3.05, and 1.55% respectively.

Figure 5.

(I) Si 2p photoelectron signals for TS-1 (A) and NiMoK/TS-1 (B). (II) Ti 2p photoelectron signals for TS-1 (A) and NiMoK/TS-1 (B).

Figure 6.

Wide XPS spectrum of the Ni₅Mo₁₀K/TS-1 catalyst.

2.1.6. Scanning Electron Microscopy (SEM) and Transmission Electron Microscopy (TEM) Analysis

SEM images of TS-1 and NiMoK/TS-1 are shown in Figure 7. The surface morphology of TS-1 (1:30) was seen to be retained after loading of nickel, molybdenum, and potassium. This indicates that the metal oxides were not present on the surface of the support TS-1 but present inside the pores of support. The fact is also concluded from the XRD patterns and TEM images provided. Figure 8i,ii shows the TEM images of TS-1 and Ni₅Mo₁₀K/TS-1, where we can clearly see the loaded nickel and molybdenum oxide particles onto the support TS-1. The images show that the diffraction fringes of the zeolite lattice can be distinguished, which denotes that the reorganized TS-1 materials keep the high crystallinity degree after metal-oxide loading, confirming the results obtained by XRD. The particle size of TS-1 was confirmed by TEM images. It was found to have approximately uniform spheroidal morphology with a particle size of around 150–200 nm. The average particle size of Ni and Mo oxides was 10–20 nm. As lattice fringes of these metal oxides were also observable, they can be said to be present inside the pores of TS-1 supporting the fact that no extra peaks due to crystalline Ni and Mo oxides were observed in XRD.

Figure 7.

SEM images of TS-1 (A) and NiMoK/TS-1 (B).

2.1.7. Thermogravimetric Analysis (TGA)

Thermal stabilities of catalysts were analyzed by thermogravimetric analysis (Figure 9). TS-1 and NiMoK/TS-1 were calcined at 600 °C before TG analysis. After an initial weight loss of ≤8% below 100 °C, the catalyst was found to be stable up to 600 °C. The initial weight loss might be due to the removal of water.

Figure 9.

TGA of TS-1 (A) and NiMoK/TS-1 (B).

2.2. Catalytic Conversion of Methyl Palmitate

2.2.1. Catalyst Screening

MoK/alumina, Ni₅Mo₅/TS (1:10), Ni₅Mo₅/TS-1, Ni₅Mo₁₀/TS-1, and Ni₅Mo₁₀K/TS-1 catalysts were synthesized and screened for methyl palmitate conversion. All reactions were carried out at 380 °C, with a flow rate of 0.2 mL/min, catalyst loading of 1.5 g, and initial feed of 0.2 M. It has been found that as the acidity of the catalyst was decreased, the conversion of methyl palmitate was reduced, but the selectivity toward C₁₅ hydrocarbons, including paraffins and specifically olefins, was increased (Table 1). This was inferred due to the formation of more cracking products when acidic catalysts were used. When MoK/alumina and Ni₅Mo₅/TS (1:10) having comparatively higher acidity were used as catalysts, the selectivity for C₁₅ products was found to be less due to the formation cracking products. The catalysts Ni₅Mo₅/TS-1, Ni₅Mo₁₀/TS-1, and Ni₅Mo₁₀-K/TS-1 had a good selectivity of 88.3, 92.0, and 95.0%, respectively, toward C₁₅ compounds. The selectivities of the α-olefinic C₁₅ compound were 67.7, 82.2, and 87.1%, respectively. It can be seen that doping of base potassium on Ni₅Mo₁₀/TS-1 decreased the conversion from 38.7 to 28.2% but increased the selectivity of C₁₅ products as well as the selectivity of C₁₅ olefin. According to this study, Ni₅Mo₁₀K/TS-1 was found to give the maximum selectivity of 87.1% toward the C₁₅ olefinic compound because of its lowest acidity.

Table 1. Effect of Different Catalysts on the Conversion of Methyl Palmitate and Selectivity of C₁₅ Products.

catalyst	Mo-K/Al	Ni5Mo5/TS (1:10)	Ni5Mo5/TS-1	Ni5Mo10/TS-1	Ni5Mo10-K/TS-1
acidity (mmol/g)	1.24	0.487	0.15	0.202	0.176
conversion (%)	57 ± 2	44 ± 2	41.3 ± 2	38.7 ± 1	28.2 ± 1
selectivity for C15 (%)	39.8 ± 2	63.4 ± 3	88.3 ± 3	92 ± 2	95 ± 2
C15 paraffin/C15 (%)	81.1	59.8	23.3	11.7	9.4
C15 olefin/C15 (%)	18.9	40.2	76.7	89.3	91.7
selectivity of C15 olefin (%)	7.5	25.5	67.73	82.16	87.1
Cn, n < 15	54.6	33.4	10.4	7	3.4
C16 hydrocarbons (%)	1.6	0.5	0.2	0.1	0.1
C16, C17 oxygenates (%)	4.1	3.3	1.4	1	0.6

2.2.2. Effect of Catalyst Loading

Different amounts of catalyst in the range of 1–2.5 g were studied (Figure 10). All reactions were carried out at 380 °C, with a flow rate of 0.2 mL/min and an initial feed of 0.2 M concentration. As the catalyst loading increases, the number of active sites increases, and so the conversion of methyl palmitate increases. To study the effect of catalyst loading, the fixed-bed reactor was packed with 1, 1.5, 2, and 2.5 g of catalyst, and the conversions of methyl palmitate were 23.2, 28.2, 31.6, and 33.5%, respectively. But the selectivities of C₁₅ compounds were decreased as 95.2, 95, 92.4, and 90.7%, respectively. This can be attributed to the formation of cracking products due to the increase in acidic sites present on the catalyst as well as due to the higher residence time of methyl palmitate because of increased catalyst bed height. Consequently, the selectivity of C₁₅ olefinic product was decreased as the catalyst loading was increased and found to be almost similar to 1 and 1.5 g of catalyst loading with a marginal difference. Thus, a catalyst loading of 1.5 g was optimized for further reactions giving the maximum yield of C₁₅ olefinic product.

Figure 10.

Effect of catalyst loading on conversion and selectivity. The reaction is carried out in a fixed-bed reactor using Ni₅Mo₁₀K/TS-1 as catalyst with a fixed flow rate of 0.2 mL/min at 380 °C and hexane as the solvent, where the concentration of methyl palmitate is 0.2 M.

2.2.3. Effect of Weight Hourly Space Velocity (WHSV)

The effect of WHSV (h^–1) was studied by changing the feed flow rates in mL/min (Figure 11). An increase in flow rate decreases the residence time of reactant on the catalyst surface, which results in the decreased conversion of reactant. Different feed flow rates such as 0.15, 0.2, 0.25, and 0.3 mL/min were studied so as to get the maximum selectivity of olefinic C₁₅ products. All reactions were carried out at 380 °C, with the same catalyst loading of 1.5 g and initial feed of 0.2 M concentration. The WHSVs for the said flow rates were 4.2, 5.6, 7.0, and 8.4 h^–1, respectively. With an increase in WHSV, the conversion of methyl palmitate was decreased by 30.4, 28.2, 26.5, and 25.5%. But the C₁₅ selectivity was found to be less at 4.2 h^–1 because of the formation of more cracking products due to the high residence time of methyl palmitate over the catalyst surface. An increase in WHSV to 5.6 h^–1 increased the C₁₅ selectivity to 95% and remained almost constant for all further WHSVs. As a result, the selectivity of the C₁₅ α-olefin product was highest at the WHSV of 5.6 h^–1, and therefore this WHSV was used to carry out further experiments.

Figure 11.

Effect of WHSV (h^–1) on conversion and selectivity. The reaction is carried out using a fixed-bed reactor, using Ni₅Mo₁₀K/TS-1 as catalyst, with a catalyst loading of 1.5 g at 380 °C, using hexane as solvent, where the concentration of methyl palmitate is 0.2 M.

2.2.4. Effect of Feed Concentration

The concentration of methyl palmitate in feed was an important parameter to affect the conversion. The feed concentrations of 0.1, 0.2, 0.3, and 0.4 M were studied at 5.6 h^–1 WHSV, resulting into conversions of 30.7, 28.2, 15.4, and 8.9%, respectively (Figure 12). As the feed concentration increased, with a catalyst loading of 1.5 g at 380 °C, the conversion of methyl palmitate was decreased. At 0.2 M concentration, the selectivity of C₁₅ olefin was found to be maximum, i.e., 87.1%, at higher concentrations, the selectivity remains unaffected. This due to the limited number of active sites present for increasing moles of methyl palmitate. Though increase in concentration has shown a decrease in concentration, it does not conclude that the rate of the reaction also decreased. The rate of reaction remained almost constant for all of the concentrations. Thus, the concentration of 0.2 mol/L in hexane was selected for further study.

Figure 12.

Effect of feed concentration (M) on conversion and selectivity. The reaction is conducted in a fixed-bed reactor using Ni₅Mo₁₀K/TS-1 as catalyst, with a WHSV of 5.6 h^–1 and catalyst loading of 1.5 g at 380 °C, using hexane as solvent.

2.2.5. Effect of Temperature

The temperature was found to be the most important parameter for this reaction so as to result in good selectivity. Temperatures ranging from 340 to 420 °C were studied for better conversion and selectivity (Figure 13). This was due to the formation of cracking products obtained by thermal cracking. All reactions were carried out with a catalyst loading of 1.5 g, feed flow rate of 0.2 mL/min, and initial feed of 0.2 M concentration. As the temperature was increased, thermal cracking of methyl palmitate took place to lower the selectivity of C₁₅ and so the final selectivity of the C₁₅ olefinic compound. Beyond 380 °C, at 400 and 420 °C, the conversion of methyl palmitate was increased from 28.2 to 33.35 and 39.4%, respectively, but the selectivity of C₁₅ products was decreased from 87.1 to 81.6 and 69.8%, respectively. Therefore, 380 °C was considered as the optimum temperature.

Figure 13.

Effect of temperature on conversion and selectivity. The reaction is conducted in a fixed-bed reactor, using Ni₅Mo₁₀K/TS-1 as catalyst, with a WHSV of 5.6 h^–1 using hexane as solvent, where the concentration of methyl palmitate is 0.2 M.

2.2.6. Effect of Hydrogen

An experiment using the same Ni₅Mo₁₀K/TS-1 catalyst with a WHSV of 5.6 h^–1, feed concentration of 0.2 M, and temperature of 380 °C was carried out in the presence of hydrogen with a H₂/feed molar ratio of 5, to give C₁₅ as well as C₁₆ paraffins and olefins. The C₁₅ paraffin and olefin selectivities were 58.2 and 10.7%, respectively, and the C₁₆ paraffin and olefin selectivities were 26.9 and 4.2%, respectively. This concludes that the presence of hydrogen facilitates both deoxygenation and decarboxylation reactions simultaneously.

2.2.7. Time on Stream Study

The activity and stability of the Ni₅Mo₁₀K/TS catalyst were studied by carrying out the reaction till 40 h, with the same catalyst bed throughout (Figure 14). The catalyst was found to be stable and active up to the first 15 h and resulted in 27.3% conversion at the 15th hour. After that, the conversion started to reduce at the 40th hour, and the conversion was found to be only 10.5%. The reduction in conversion after a certain time was because of the deactivation of catalyst due to coking and sintering. Due to deactivation, very few active sites were available for the reactants to adsorb, resulting in very less conversion. The selectivity of C₁₅ products also started to reduce after 15–20 h. Due to the deposition of coke on the catalyst surface, cracking was favored forming C_n (n < 15) products, and also the product composition contains more of paraffins than olefins. Overall, the selectivity of the C₁₅ olefinic product was 86.5% at 15 h, and after that it started reducing; at the 40th hour, it was only 48.2%.

Figure 14.

Time on stream study of a reaction conducted in a fixed-bed reactor using Ni₅Mo₁₀K/TS-1 as catalyst, with a WHSV of 5.6 h^–1, catalyst loading of 1.5 g at 380 °C, using hexane as solvent, where the concentration of methyl palmitate is 0.2 M.

2.2.8. Regeneration and Reusability of the Catalyst

For regeneration of the catalyst, the spent catalyst was first washed with N₂ gas with a flow rate of 100 mL/min for 1 h at 200 °C. The dry catalyst thus obtained was washed with isopropyl alcohol to wash out the adsorbed material from the catalyst. Then, it was calcined at 600 °C for 4 h. The loss of catalyst during this regeneration process was calculated and made up with fresh catalyst so as to get the same catalyst loading. The catalyst can also be regenerated simply by calcining at 600 °C inside the reactor. In this manner, the catalyst reusability study was done four times (Figure 15). It was found that there was no considerable difference in the conversion of methyl palmitate and the selectivity of the C₁₅ olefinic product. Hence, Ni₅Mo₁₀K/TS-1 was a robust and reusable catalyst.

Figure 15.

Time on stream study of a reaction carried out in a fixed-bed reactor using Ni₅Mo₁₀K-TS-1 as a catalyst, with a WHSV of 5.6 h^–1, catalyst loading of 1.5 g at 380 °C, using hexane as solvent, where the concentration of methyl palmitate is 0.2 M.

2.3. Kinetic Model

The possible mechanism of conversion of methyl palmitate over metal-oxide surface to α-pentadecene is shown in Scheme 2. In this mechanism, Mⁿ⁺ denotes metal M with oxidation state n+. According to the stated mechanism, methyl palmitate first gets adsorbed on the active metal M of catalyst, thus increasing the oxidation state of metal to n + 1. Oxygen from metal oxide then abstracts proton from the 14th carbon of methyl palmitate to yield CO₂ and olefinic product. This decarboxylation process leaves the methyl group on M. On further desorption of methane, metal M retains its original oxidation state (i.e., n+). Figure S4 shows the gas chromatograph of gaseous products obtained, which supports the proposed mechanism.

Scheme 2. Possible Mechanism of Methyl Palmitate to C₁₅ α-Olefin Conversion.

The conversion of methyl palmitate to C₁₅ products is a parallel reaction with paraffin and olefin as products. It contains methyl palmitate (M) as a reactant and C₁₅ paraffin (P) and C₁₅ olefin (O) as products. Gaseous co-products (G) can be neglected.

The chemisorption of methyl palmitate on vacant catalytic site (S) is the first step followed by other reactions.

1

The chemisorbed methyl palmitate (MS) then undergoes two parallel reactions to produce C₁₅ paraffin (P) and olefin (O) as depicted below

2

3

The final rate of reaction of methyl palmitate M (mol/h·g-cat) assuming both of the parallel reactions 2 and 3 as rate-controlling is given by

4

where C_S is the concentration of vacant sites (S) calculated from the concentration of total site (C_t), adsorption equilibrium constants (K_i), and concentrations of relevant species (C_i)

5

6

According to eqs 5 and 6, eq 4 becomes

7

Considering all species are weakly adsorbed and if ∑K_iC_i ≪ 1, the above reaction rate equation becomes

8

9

where the overall rate constant was

10

When eq 9 is integrated, we get the following equation for a fixed-bed vapor-phase catalytic reactor

11

A plot of −ln (1 – X_M) vs W/F_Mo was made as a function of temperature to get a very good fit (Figure 16). This confirms the first order of the reaction.

Figure 16.

First-order fixed-bed reactor plot −ln (1 – X_M) vs W/F_Mo for the conversion of methyl palmitate.

Rate constants at different temperatures, 613, 633, 653, and 673 K, were calculated. The Arrhenius plot was made to calculate the activation energy of the reaction (Figure 17). The value of the apparent energy of activation was 10.11 kcal/mol. It can be concluded that this reaction was essentially kinetically controlled.

Figure 17.

Arrhenius plot of ln k vs 1/T for the conversion of methyl palmitate.

To find k₁ and k₂ individually, we have assumed that only one reaction occurs at a time. The rate equations for the synthesis of paraffin and olefin separately using a fixed-bed vapor-phase catalytic reactor are derived from eqs 2 and 3, respectively

12

13

Plots of −ln (1 – X_M) vs W/F_Mo were made as a function of temperature to find out k₁ and k₂ at different temperatures (Figure 18). The values of k′ obtained in Figure 16 were found to be almost matching with k₁ + k₂. The marginal difference obtained may be attributed to the formation of other side products.

Figure 18.

First-order fixed-bed reactor plot −ln (1 – X_M) vs W/F_Mo for the conversion of methyl palmitate to C₁₅ paraffin (A) and olefin (B).

The Arrhenius plots (Figure 19) were made to calculate the activation energies of paraffin as well as olefin formation. The activation energy of paraffin production from methyl palmitate is 10.31 kcal/mol, and the activation energy of olefin production from methyl palmitate is 9.46 kcal/mol. Thus, both the reactions can be said to be kinetically controlled. The activation energy analysis suggested that the production of the C₁₅ olefinic product from methyl palmitate is more temperature-sensitive than the production of the C₁₅ paraffin from methyl palmitate.

Figure 19.

Arrhenius plot of ln k₁ and ln k₂ vs 1/T for the conversion of methyl palmitate to C₁₅ paraffin (A) and olefin (B).

3. Conclusions

Catalytic decarboxylation of fatty acid methyl esters and triglycerides is a promising route to produce “green” diesel fuel or hydrocarbons, which can substitute for petrochemical feedstock or specialty chemicals. The reported catalyst NiMoK/TS-1 is found to be efficient to selectively produce decarboxylated products, especially α-olefins without the use of hydrogen. WHSV of 5.6 h^–1 and temperature of 380 °C were found to be optimum for this reaction when the feed concentration is 0.2 M. Cyclic and aromatic as well as cracking products were eliminated to selectively get 95% C₁₅ hydrocarbons with 87% of 1-pentadecene (i.e., C₁₅ α-olefin). Though the single-pass conversion is less, reactive distillation enables easy removal of products from the unreacted feed, which can be recirculated through the reactor.

4. Experimental Section

4.1. Materials

Methyl palmitate (97% pure) and tetraethyl orthosilicate (TEOS) were obtained from Alfa Aesar, India. Titanium isopropoxide 98% (TTIP), tetrapropylammonuim hydroxide solution 1.0 M (TPAOH), and nickel (II) nitrate hexahydrate (Ni(NO₃)₂·6H₂O) were obtained from Avra Chemicals, India. Ammonium heptamolybdate ((NH₄)₆Mo₇O₂₄) was purchased from Thomas Baker Chemicals, India.

4.2. Catalyst Preparation

TS-1 represents TS (1:30) with a 1:30 (w/w) ratio of TiO₂/SiO₂, which also means a 1:23.5 (w/w) ratio of Ti/Si in the gel. Similarly, TS (1:10) represents a 1:10 (w/w) ratio of TiO₂/SiO₂ and a 1:7.35 (w/w) ratio of Ti/Si in the gel. The procedure of synthesizing crystalline support TS-1 is given below.

Tetraethyl orthosilicate (TEOS) (49 g) was added with 60 mL of aqueous tetrapropylammonium hydroxide (1 M TPAOH) for hydrolysis of the TEOS, resulting in a pH of 12.5. To this mixture, 1.67 g of titanium isopropoxide (TTIP) solution prepared in dry isopropyl alcohol was added dropwise with vigorous stirring. The clear liquid obtained was then stirred for 1 h to hydrolyze TEOS and TTIP completely. Finally, the remaining 20 mL of TPAOH solution was slowly added to the above mixture. This mixture was then stirred for about 3 h at 348–353 K to remove the alcohol. The chemical composition of the initial gel was

The final mixture was transferred into a 250 mL Teflon flask and placed into a stainless steel autoclave. Crystallization was carried out under static hydrothermal conditions at 443 K for 1 day. After crystallization, solids were separated by centrifugation and dried at 373 K for 5 h.

Nickel nitrate hexahydrate and ammonium heptamolybdate were simultaneously impregnated onto TS-1 by the incipient wetness impregnation method in appropriate amounts. The materials were then calcined at 600 °C for 5 h to get catalysts Ni₅Mo₅/TS-1 and Ni₅Mo₁₀/TS-1. During the synthesis of Ni₅Mo₁₀K/TS-1, KNO₃ is impregnated after metallic precursors so as to get 0.5% (w/w) loading of potassium. TS (1:30) is henceforth denoted by TS-1.

4.3. Decarboxylation Reaction

The decarboxylation reactions are carried out in a fixed-bed vapor-phase reactor. A detailed description of the reactor is provided in the Supporting Information. Single-pass experiments were carried out for studying the effects of different parameters such as concentration, catalyst loading, temperature, WHSV, etc. N₂ was used as a carrier gas. Hexane was used as a solvent for methyl palmitate. The used catalyst was regenerated by calcining it at 600 °C for 4 h.

4.4. Analysis of Products and Characterization of Catalysts

Catalyst characterizations were done using different techniques, namely, temperature-programmed desorption (NH₃-TPD), Fourier transform infrared (FTIR) spectroscopy, X-ray diffraction (XRD), thermogravimetric analysis (TGA), Brunauer–Emmett–Teller (BET) surface area, scanning electron microscopy (SEM), transmission electron microscopy (TEM), and X-ray photoelectron spectroscopy (XPS). Detailed specifications of these instruments are provided in the Supporting Information.

The product samples of decarboxylation reactions were analyzed using GC with a TG5MS capillary column with dimensions of 0.25 mm × 30 m (Thermo 1110 model). The product identification was done using GCMS (Thermo Scientific ISQLT 1300; capillary column TG5MS; 0.25 mm × 30 m). The methods used for calculating % concentration and % selectivity are also provided in the Supporting Information.

Glossary

Nomenclature

C_M

concentration of methyl palmitate (mol/g)

C_i

concentration of relevant species (mol/g)

C_MS

concentration of adsorbed M

C_s

concentration of vacant sites (mol/g)

F_Mo

molar flow rate of methyl palmitate (mol/h)

G

other gaseous products

k′

rate constant for methyl palmitate conversion (h^-1)

k₁

rate constant for olefin formation (h-1)

k₂

rate constant for paraffin formation (h-1)

K_A

adsorption coefficient for methyl palmitate

M

reactive species methyl palmitate

MS

chemisorbed M

O

C₁₅ olefin

OS

chemisorbed C₁₅ olefin

P

C₁₅ paraffin

PS

chemisorbed C₁₅ paraffin

r_M

rate of surface reaction of methyl palmitate M (mol/h·g-catalyst)

S

vacant sites

TS

titanosilicate

w

weight of catalyst (g)

WHSV

weight hourly space velocity

X_M

mole fraction of methyl palmitate

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.9b03993.

• NH₃-TPD analysis for different supports and catalysts; high-resolution XPS spectra of Mo 3d and Ni 2p; unprocessed FTIR spectra; gas chromatography data of gaseous products obtained during reaction using TCD; details about the reactor and analytical instruments; and methods used along with some characterization data (PDF)

The authors declare no competing financial interest.