High-Yield Production of Fatty Nitriles by One-Step Vapor-Phase Thermocatalysis of Triglycerides

Abstract

Fatty nitriles are widely used as intermediate molecules in the pharmaceutical and polymer industries. In addition, hydrogenation of fatty nitriles produces fatty amines that are common surfactants. In the conventional fatty nitrile process, triglycerides are first hydrolyzed and the resulting fatty acids are catalytically reacted with NH₃ in a liquid-phase reaction. In this study, we report a simpler one-step fatty nitrile production method that involves a direct vapor-phase reaction of triglycerides with NH₃ in the presence of heterogeneous solid acid catalysts. The reactions were performed in a tubular reactor maintained at 400 °C into which triglycerides were injected through an atomizer to allow rapid volatilization and reaction; NH₃ was fed as a gas. Several metal oxide catalysts were tested, and reactions in the presence of V₂O₅ resulted in near-theoretical fatty nitrile yields (84 wt % relative to the feed mass). In general, catalysts with higher acidity such as V₂O₅, Fe₂O₃, and ZnO showed higher fatty nitrile yields compared to low acidity catalysts such as ZrO, Al₂O₃, and CuO. Energy balance calculations indicate that the one-step reaction described here would require significantly lower energy than the conventional process primarily because of the elimination of the energy-intense triglyceride hydrolysis.

1. Introduction

Triacylglycerols (commonly known as triglycerides) from oilseeds or algae have the potential to, at least partially, displace petroleum-derived fuels and chemicals.¹⁻⁵ Their built-in oxygenated functional groups can be utilized for value-added chemicals, which are traditionally produced from the functionalization and oxidation of petroleum-derived hydrocarbons.^6,7 For instance, fatty nitriles (platform molecules in pharmaceutical and polymer industries)⁸ and fatty amines (common cationic surfactant precursor),⁶ which are nitrogen derivatives of fatty acids, olefins, or alcohols, can be produced from fats and oils instead of petrochemical raw materials. Owing to their high affinity to natural surfaces, fatty amines have been used in surface modification applications such as fabric softening, hair conditioning, corrosion inhibition, mineral flotation, and bactericides.^9,10 Fatty amines also have good oil solubility and lubricity and are thereby particularly useful in friction modification. In addition, the emulsification and dispersion properties of fatty amines are exploited in numerous cleaning and agricultural formulations.^9,10 A recent global market study on fatty amines projected that global industrial consumption of fatty amines is expected to reach >650 000 metric tons with a revenue of more than 2 billion US dollars by 2020.¹¹

Figure 1 shows the conventional routes for fatty amine production from petroleum as well as oleaginous biomass feedstocks. Although petroleum is still the main resource for amine production, the bio-based route is potentially more environmentally sustainable. In the conventional bio-based process, the first step is the conversion of triglycerides (extracted from oilseeds) to fatty acids through hydrolysis. Subsequently, fatty nitriles are produced through the “nitrile process”¹² wherein the fatty acids are reacted with NH₃ in the presence of metal oxide catalysts such as alumina or zinc oxide (see purple box in Figure 1). The nitrile reaction is performed at 280–360 °C in the liquid state for several hours. Excess ammonia (2–4 times of the required stoichiometric value) is added, and water is continually removed from the reactor to move the equilibrium reaction toward nitrile production. Downstream, the fatty nitriles are hydrogenated, typically in the presence of a nickel catalyst (e.g., RANEY Ni), to produce fatty amines.¹³

Figure 1.

Pathways to produce fatty amines.

In batch implementations of the conventional nitrile process, fatty acids stay at a high temperature and in a liquid state for extended periods. Consequently, several side reactions such as isomerization, polymerization, Piria, Diels–Alder, or peroxidation reactions can occur, particularly in the presence of unsaturated fatty acids.^8,14 Moreover, in batch-reactor systems, the produced fatty nitriles could also be hydrated to fatty amides. Thus, at the end of the reaction, the product contains unreacted fatty acids, fatty nitriles, fatty amides, and other undesired byproducts, thereby lowering fatty nitrile yields and complicating product recovery. To improve the fatty nitrile yields, some studies have investigated nitrile reactions in continuous flow reactors packed with catalysts and fed with fatty acids and NH₃.¹⁴⁻¹⁶ The continuous nitrile process has the potential to decrease undesired side products (compared to batch reactors) owing to a more efficient contact between NH₃ and fatty acids, lower reaction time (10–60 min), and continuous products removal.¹⁶ Theoretical fatty nitrile yields (70–90%) are reported from the continuous nitrile process, and the product stream is reported to contain unreacted fatty acids, fatty amides, and dimers of fatty nitrile. Higher yields are possible with additional processing such as dehydration of the intermediate fatty amide.¹⁴ Purification/distillation is used as a final step to produce high-purity fatty nitriles required for surfactant applications.¹⁴

An additional variation of the nitrile process involves gas-phase reactions of fatty acids with NH₃ at a temperature range of 300–600 °C and in the presence of a solid catalyst bed. Gas-phase reactions can potentially prevent undesired products owing to a short reaction time (few seconds)^17,18 and better contact between NH₃ and fatty acids.¹⁹ Also, gas-phase nitrile reactions significantly decrease the tendency for dimerization/polymerization owing to a greater distance between molecules (reactant and intermediate products), particularly when unsaturated fatty acids are used as feedstock.^5,20,21 Additionally, the process does not require a catalyst recovery since a heterogeneous catalyst bed is used.⁸

In contrast to the several published studies on the conventional liquid-phase nitrile process, there are relatively few reports on the gas-phase nitrile reactions of fatty acids (or its derivatives, e.g., fatty esters). From the published literature, the important reaction parameters that govern fatty nitrile yields in gas-phase reactions are temperature, NH₃/fatty acid molar ratio, and catalyst type. For instance, Wortz et al.²² used silica gel as a catalyst for nitrile reaction of several fatty acids (e.g., lauric, palmitic, and stearic acid) at reaction temperatures of 425–450 °C. They reported high yields of long-chain fatty nitriles (up to 98% of the theoretical yield) and only ∼1% short chain fatty nitriles. Ralston et al.²³ also carried out gas-phase nitrile reaction of long-chain fatty esters in the presence of Al₂O₃ supported by activated charcoal. In their process, the reaction was performed at a much higher temperature (above 500 °C) to eliminate the production of any heavy byproducts such as polymers and resins. However, higher reaction temperatures favor the formation of short-chain fatty nitriles owing to a C–C cleavage of the original fatty acid (feedstock). Thus, in their study, the products were mainly composed of short-chain fatty nitriles and hydrocarbon gases.

Various metal oxide catalysts such as ZnO, Zr₂O, and Al₂O₃ have been used to convert fatty acids (or its derivatives, e.g., fatty ester) into fatty nitrile. It has been reported that oxides of metals from groups III and IV of the periodic table are especially effective for fatty acid ammonization reaction.^17,23 Mekki-Berrada et al.¹⁷ studied several basic (e.g., MgO), amphoteric (e.g., ZnO, Zr₂O, and Al₂O₃), and acidic (WO₃) solid catalysts for gas-phase nitrile reaction of lauric acid methyl ester at reaction temperature of 300 °C. They showed that catalysts with amphoteric character provide higher fatty nitrile yields compared to basic or strong acidic catalysts. Although, acidic catalysts such as Nb₂O₅ yielded as high as 97% fatty nitriles, stronger acidic catalysts (e.g., bentonite and WO₃) promoted side products (e.g., methyl lauramide) and decreased the fatty nitriles yields.

Overall, the gas-phase nitrile reaction has the potential to produce high yield of fatty nitrile from fatty acids. Previous studies (both gas- and liquid-phase reaction) have used fatty acid, which are typically produced from the hydrolysis of triglycerides, as a reactant for nitrile reactions. However, hydrolysis is an energy-intense process owing to high temperature (250–350 °C) and pressure (45–60 bar) and requires long reaction times.^24,25 Moreover, fatty acid purification/distillation and catalyst recovery is required after hydrolysis. Furthermore, polymerization and degradation can occur during the hydrolysis step, particularly, in presence of unsaturated fatty acids. If nitrile reaction can be performed in the vapor phase by the direct reaction of triglycerides with ammonia in the presence of a catalyst bed, the hydrolysis step as well as catalyst recovery can be avoided. In the present work, we investigated a one-step vapor-phase nitrile reaction to directly convert triglycerides into fatty nitrile using diverse solid acid catalysts. Product yields were correlated to catalyst properties. Effects of triglyceride/NH₃ ratios on product yield were assessed for reactions in the presence of V₂O₅ catalyst.

2. Results and Discussion

2.1. Feedstock Characterization

The coconut oil used in this study was composed of 5% caprylic (C8:0), 5% capric (C10:0), 46% lauric (C12:0), 20% myristic (C14:0), 12% palmitic (C16:0), and 12% stearic and oleic (C18:0 and C18:1) acids (Figure S2). As expected, the coconut oil was mainly composed of fatty acids in the C12–C18 range, which are important in soaps, surfactants for cosmetics, pharmaceuticals, and foodstuffs because they carry a biocompatible lipophilic group.²⁶

2.2. Catalyst Characterization

Table 1 shows the physiochemical properties of the tested catalysts. The texture properties were determined from N₂ adsorption isotherms at 77 K, and the acid properties were measured by the ammonia temperature-programmed desorption (NH₃-TPD) method. As shown in Table 1, most of the metal oxide catalysts have similar Brunauer–Emmett–Teller (BET) surface areas (∼2–6 m² g^–1), with the exception of HZSM-5 that has a much higher surface area (337 m² g^–1) and pore volume (0.2 cm³ g^–1). Figure S3 shows the NH₃ desorption profile obtained during catalyst heating. As observed, most metal oxide catalysts (with the exception of V₂O₅) showed two distinct peaks at different temperatures. The first peak corresponds to weak/medium acidity and the second peak is from strong acid sites on the catalyst. A summary of the catalyst acid properties is also shown in Table 1. ZrO₂, V₂O₅, Fe₂O₃, and ZnO contained more acid sites than Al₂O₃ and CuO. HZSM-5 showed a significantly higher acidity of 459 μmol g^–1 owing to strong Brønsted acid sites.

Table 1. Texture-Properties and Acidity of Tested Catalysts.

properties	texturea			acidityb
catalyst	surface areac (m2 g–1)	pore sized (nm)	pore volumee (cm3 g–1)	total acidity (μmol g–1)	1st acidityf (μmol g–1)	2nd acidityg (μmol g–1)	Tmax.1 (K)	Tmax.2 (K)
V2O5	6	18.7	0.028	14.1	8.6	5.5	378.7	441.3
Fe2O3	4.5	16.2	0.014	12.2	3.7	8.6	485.7	703.1
ZrO2	5.5	4.7	0.007	23.8	23.3	0.4	468.5	748.7
ZnO	4.6	4.4	0.006	8.5	5.9	2.7	490.6	586.8
Al2O3	2.6	12.8	0.009	3.1	2.6	0.5	448.4	615.6
CuO	1.7	20.3	0.009	1.0	1.0	ND	432	ND
HZSM-5	337.7	0.5	0.204	459.1	171.9	287.2	397.1	503

^aTexture properties were measured by ASAP 2020 instrument.

^bAcidity was measured by NH₃-TPD method.

^cBET surface area.

^dBarrett–Joyner–Halenda adsorption average pore diameter.

^eCumulative pore volume.

^fMeasured from the area of first peak; corresponds to medium/weak acid sites.

^gMeasured from area of second peak; corresponds to strong acid sites. T_max.1: temperature where maximum NH₃ desorption was observed in the first peak. T_max.2: temperature where maximum NH₃ desorption was observed in the second peak. ND: not detected.

2.3. Product Yields and Compositions

Figure 2a shows the liquid product yields (relative to feed mass) from the one-step vapor-phase nitrile reaction of triglycerides in the presence of tested catalysts. For the reaction in the absence of catalysts, glass beads (1 mm o.d.) were packed (5 cm length) inside the reactor to maintain a similar residence time (10 s) as the reactions in the presence of catalysts. The dashed horizontal line represents the theoretical maximum fatty nitrile yield. For quantitative coconut oil conversion and 100% selectivity toward fatty nitriles (see the generalized reaction pathways in Figure S4), the mass of fatty nitriles produced would be 598 g per mol of coconut oil fed into the reactor. Using an average coconut oil molecular weight of 693 g mol^–1 (based on measured fatty acid composition of coconut oil; Figure S2), the weight-based theoretical yield of fatty nitriles equals 86 wt % relative to coconut oil.

Figure 2.

(a) Products composition and (b) fatty nitrile composition from the one-step vapor-phase nitrile reaction over tested catalysts. The sum of the individual fatty nitrile weight fractions in the product represents the total fatty nitrile selectivity. C8–C18 represents the carbon numbers in fatty nitriles.

From Figure 2a, it can be seen that the liquid product from the reaction performed in the absence of a catalyst comprises 28 wt % fatty acids, 26 wt % fatty amides, and only 22 wt % fatty nitriles (all values are relative to feed mass). The product also contained 6 wt % unreacted triglyceride feed (not shown in Figure 2a). The fatty nitrile yields significantly increased in the presence of catalysts (Figure 2a) owing to greater ammonization of the triglycerides and/or produced fatty acids and dehydration of the fatty amides (second and third reactions in Figure S4). It must be noted that while the reaction scheme in Figure S4 shows the reactions of fatty acids with ammonia, it is also possible that the triglycerides react directly with NH₃. Our results show that the composition of liquid products was catalyst-dependent. For example, a near-theoretical yield of 84 wt % fatty nitrile (relative to feed mass) was achieved in the presence of V₂O₅. Yields were also high with HZSM-5 and Fe₂O₃—nearly 81 wt % fatty nitriles. Moderate yields of 77 and 73% were obtained with ZrO₂ and ZnO, respectively. Reactions in the presence of Al₂O₃ yielded only 50% fatty nitriles, and the fatty nitrile yields were even lower (34 wt %) when CuO was used as a catalyst. The gas chromatography–flame ionization detection (GC–FID) chromatograms of products from the reactions in the absence of catalysts, with Al₂O₃ and V₂O₅ (to compare products from low- and high-yield catalysts), are shown in Figure S5.

Figure 2b shows the composition of the fatty nitriles in the product from the catalytic reactions (=mass of nitrile/total product mass). The sum of the individual fatty nitrile weight fractions in the product represents the total fatty nitrile selectivity. We observed a mixture of C8–C18 fatty nitriles that corresponded to the C8–C18 fatty acids esters of coconut oil (see Figure S2). In the presence of V₂O₅, the selectivity was as high as 97% such that the fatty nitrile composition was nearly the same as the coconut oil fatty acid composition (see Figure S2). Table S1 shows the detailed composition of liquid products from all of the reactions. Fatty nitrile products smaller than C8 or larger than C18 were not observed indicating that excessive cracking or polymerization reactions were prevented because of the combination of low reaction temperature and short residence time. Although the liquid product from V₂O₅ contained a mixture of fatty nitriles with different carbon chain length, sufficient boiling point differences among the products would allow for the purification of specific fatty nitriles via distillation, if desired.

2.4. Effects of Catalyst Acidity on Fatty Nitrile Yield

Vapor-phase nitrile reaction in the presence of metal oxide catalysts improved the fatty nitrile yields; however, some catalysts such as V₂O₅ and Fe₂O₃ showed much better fatty nitrile yields and selectivity compared to others such as Al₂O₃ and CuO. To understand the difference in performance of the tested catalysts, we further investigated the differences in catalyst properties. Figure 3 shows that catalyst acidity and fatty nitrile yields were positively correlated. For instance, V₂O₅ showed desorption of 14 μmol NH₃ per gram of catalyst and resulted in ∼84% fatty nitriles; however, Al₂O₃ has 8 μmol g^–1 acidity and showed only 50% fatty nitrile yields (acidity values are given in Table 1). Moreover, the products from the one-step vapor-phase nitrile reaction in the presence of V₂O₅ did not contain any measurable fatty amide content likely owing to fast dehydration of fatty amide in the presence of Lewis acid sites on the catalyst (see the reaction steps in Figure S4). On the other hand, liquid products in the presence of Al₂O₃ contained up to 12% fatty amide (see Figure 2a) possibly because of the lower acidity of Al₂O₃ compared to V₂O₅. Furthermore, it is likely that the acid sites of catalysts improved the ammonization reaction of triglyceride and/or produced fatty acid in addition to increasing the rate of the fatty amide dehydration. As a result, there was no evidence of triglyceride and/or fatty acid in the liquid products when V₂O₅ was used. However, the products in the presence of Al₂O₃ and CuO contained fatty acid (see Figure 2a) likely owing to less ammonization of the produced fatty acids as a result of lower catalyst acidity.

Figure 3.

Correlation of catalyst acidity and fatty nitrile yields from vapor-phase nitrile reaction of coconut oil.

2.5. Effect of NH₃/Triglyceride Molar Ratio

Figure 4 shows the fatty acid and fatty nitrile yields from one-step vapor-phase nitrile reactions at various NH₃/triglyceride molar ratios, and Figure S6 shows the GC–FID chromatograms of the liquid products at various reactant molar ratios. The reactions were performed at 400 °C using V₂O₅. The molar ratios were adjusted by changing the mass flow rates of coconut oil into the reactor while maintaining a constant NH₃ flow rate. As shown in Figure 4, in the absence of NH₃, the liquid products were mainly composed of fatty acids (76 wt % of C8–C18 fatty acids) along with some hydrocarbons (mainly C11–C14). These results are consistent with our previous observations of the triglyceride pyrolysis.⁵ In the absence of NH₃, triglycerides degrade into fatty acids and release glycerol backbone; then, the produced fatty acids can undergo deoxygenation and produce hydrocarbons. When stoichiometric amount of NH₃ was introduced into the reactor, the liquid products contained a mixture of fatty nitriles and acids. The fatty acid yields decreased from 76% in the absence of NH₃ to 45% at NH₃/triglyceride molar ratio of 3 (stoichiometric value). Moreover, under these conditions, a 41% fatty nitrile yield was obtained. The fatty nitrile yields consistently increased when the NH₃/triglyceride molar ratio increased up to 7.5 (2.5 times of stoichiometric value) and reached the maximum nitrile yields of ∼97%. Interestingly, there was no evidence of fatty amide productions even when the stoichiometric amount of NH₃/triglyceride molar ratio was used. This indicates that dehydration of fatty amides in the presence of V₂O₅ acid sites is a fast reaction step.

Figure 4.

Effects of NH₃/triglyceride molar ratio on product yields (relative to the total mass of liquid products) from one-step vapor-phase nitrile reaction.

2.6. Proposed Reaction Mechanism of Fatty Nitrile Production

Figure 5 shows the mechanism for the production of fatty nitrile from triglyceride over Al₂O₃. Although the yields were lower with Al₂O₃ than with other catalysts, the reaction mechanism is described with this catalyst because its structure is well-known. Partially dehydroxylated metal oxides provide acidic sites, and the density of the Brønsted/Lewis acidic sites on the surface highly depends on the dehydroxylation temperature and gas flow. In our reactor system, high temperature and flow of inert gas (N₂) were applied prior to the exposure of the metal oxide to ammonia and triglyceride to generate active sites. We propose that the adsorption of ammonia over the surface acidic sites and its activation to form a bridging amide (−NH₂) is a plausible initiation step in this process (see Figure 5a). Then, we hypothesize that a series of acid–base reactions occur as shown in Figure 5b. The nucleophile O on carboxylic group coordinates with the electrophile Al and the electrophile C on the carboxylic group undergo nucleophilic attack by strongly basic −NH₂ to generate the active complex i. Transfer of alkoxy group from the intermediate ii to the Lewis acidic site yields a bridging alkoxide and the desired fatty amide product. The produced fatty amide undergoes dehydration (Figure 5c) and yields fatty nitrile and water. While this reaction can occur at high temperatures even in the absence of a catalyst, the acidic catalyst sites facilitate the dehydration reaction. The alkoxide-bridged surface species is active and can act as an active site and regenerate the catalyst (Figure 5d).

Figure 5.

Possible vapor-phase nitrile reaction mechanism over metal oxide catalysts. (a) Catalysts activation step, (b) fatty amide formation in the presence of the active catalyst, (c) fatty amide dehydration, and (d) active catalyst regeneration. M_xO_y: metal oxide and Δ: high temperature.

2.7. Comparison between Conventional and One-Step Vapor-Phase Nitrile Process

Table 2 shows the energy consumption during the conventional and one-step vapor-phase nitrile process. A detailed energy calculation and assumptions are given in the Supporting Information (see Figure S7 and accompanying description). In the conventional nitrile process, the first step is the hydrolysis of triglycerides to produce fatty acids. A triglyceride/water ratio of 1/5, temperature of 250 °C, pressure of 45 bar, and 100% triglyceride conversion were used to calculate the energy required in the hydrolysis step.^24,25 As shown in Table 2, hydrolysis of 1 kg of triglyceride requires a minimum energy of 5652 kJ. Moreover, conversion of the fatty acid into fatty nitrile through the conventional nitrile process requires a minimum energy of 387 kJ kg^–1 triglyceride. Assuming that 75% of the consumed energy can be recovered via appropriate heat exchange, the conventional method requires more than 1500 kJ to convert 1 kg of triglyceride into fatty nitriles.

Table 2. Comparison of Energy Requirements for Conventional Nitrile Process and the Proposed One-step Vapor Phase Nitrile Production. Detailed Energy Calculations and Assumptions are Given in the Supplementary Information.

	process methods
energy components	conventional (kJ kg–1 triglyceride)	proposed one-step vapor phase (kJ kg–1 triglyceride)
hydrolysis	5652b	n/aa
nitriles production	387c	925d
total energy consumed	6039	925
energy that can be recoverede	4529	694
total net energy required	1510	231

^an/a: not applicable.

^bT = 250 °C, P = 45 bar, and triglyceride/water mass ratio of 1/5.

^cT = 300 °C and P = 1 bar.

^dT = 400 °C and P = 1 bar.

^eAssuming 75% of the consumed energy assumed that can be recovered.

The one-step vapor-phase nitrile reaction requires energy for increasing the feedstock temperature from ambient to the boiling point (∼400 °C) and also vaporization. From Table 2, one can see that the one-step vapor-phase nitrile process requires only 925 kJ kg^–1 triglyceride for this step. This energy requirement would even decrease to 230 kJ kg^–1 triglyceride if 75% energy recovery could be achieved through heat exchange from the hot product stream. The lower energy consumption in our system is because of the elimination of the hydrolysis step. Furthermore, the one-step vapor-phase nitrile process eliminates (or at least significantly decreases) the complicated distillation/purification steps required in the hydrolysis and fatty nitrile production steps through the conventional process. Overall, the one-step vapor-phase nitrile reaction is able to directly transform triglyceride into fatty nitrile at nearly theoretical yields with lower energy consumption and potentially much simpler reactors and separation units.

3. Conclusions

The one-step vapor-phase fatty nitrile production was performed in a single reactor, in which triglyceride reacted with ammonia. In the absence of a catalyst, the products contained 28% fatty acid, 26% fatty amide, and only 21% fatty nitrile (relative to the feed). However, the fatty nitrile yields significantly increased when a catalyst was used. A near-theoretical fatty nitrile yield of 84% (relative to feed mass) was achieved from the nitrile process in the presence of V₂O₅. Catalysts such as Fe₂O₃, ZrO₂, ZnO, and HZSM-5 also showed high fatty nitrile yields. However, the fatty nitrile yields were low in the presence of Al₂O₃ and CuO, likely because of the low acidity of these catalysts. Overall, nitrile yields positively correlated with catalyst acidity. The results showed that a minimum NH₃/triglyceride molar ratio of 7.5 is required to achieve a near-theoretical fatty nitrile yield. Energy assessments suggest that the one-step vapor-phase nitrile process would require lower energy inputs than the conventional nitrile process. Furthermore, owing to the high purity of the nitrile product, purification/distillation steps are expected to be simple. Finally, catalyst recovery from the hydrolysis reactions and the nitrile production steps in the conventional nitrile process is eliminated owing to the vapor-phase reaction of triglyceride.

4. Experimental Section

4.1. Materials

Coconut oil was obtained from Zoye Oil (Zeeland, MI, USA). The following catalysts were purchased from Strem Chemicals, Inc. (Newburyport, MA, USA)—V₂O₅, Fe₂O₃, ZrO₂, ZnO, Al₂O₃, and CuO. NH₄-ZSM-5 powder with an SiO₂/Al₂O₃ molar ratio of 23 was purchased from Zeolyst International (Conshohocken, PA, USA). The purchased NH₄-ZSM-5 was calcinated in a muffle furnace for 5.5 h at 550 °C to obtain HZSM-5. Hexane, chloroform, methanol, and sulfuric acid were purchased from Fisher Scientific (Pittsburgh, PA, USA). Analytical standards of lauric nitrile, lauramide, lauric acid, glycerides (triolein, diolein, and monolein), and fatty acid methyl ester (FAMEs; mixtures of C8–C22) were purchased from Sigma-Aldrich (St. Louis, MO, USA).

4.2. Experimental Setup

All experiments were performed in a continuous vapor-phase reactor system that is schematically shown in Figure 6, as described previously.⁵ Prior to each experiment, the reactor was packed with catalysts, placed in the furnace, and heated while being purged with pure N₂ to remove any traces of H₂O and O₂ from the system. After the reactor reached the set point temperature, the N₂ purge was stopped. Then, NH₃ gas was introduced into the reactor at a controlled flow rate measured by a mass flow meter (Alicat Scientific, USA). Thereafter, feed coconut oil was introduced into the reactor in the form of micron-size droplets through an atomizer (Sonazop, Farmingdale, NY, USA; model: HTNS40). The atomizer consists of an ultrasonic probe with a 4 mm diameter, which was operated at a constant frequency of 40 kHz. The atomizer used in our reactor system creates extremely large surface area for heat transfer and allows the rapid evaporation of feedstock. The system was operated in continuous mode for 3 h, and the products were condensed in a liquid N₂ trap. At the end of the experiment, feed flow (NH₃ and coconut oil) was stopped and the reactor furnace was turned off and allowed to cool to room temperature. Thereafter, the collected liquid products were weighed on an analytical balance (Mettler Toledo, USA) with ±0.1 mg accuracy. The liquid product was separated in a separatory funnel into aqueous and organic phases.

Figure 6.

Schematic diagram of the one-step vapor-phase nitrile reaction system.

The conversion of coconut oil to nonglyceride products was calculated as

1

where, (W_TG)_in is the mass of the oil introduced into the reactor and (W_TG)_out is the mass of glycerides in the collected liquid.

The yield of fatty acid, amide, and nitrile in the liquid products was calculated as

2

where w_i is the mass of fatty acids, amides, or nitriles in the product.

The reaction residence time (τ) was calculated using eq 3 as reported in previous studies.^17,18

3

where V is the catalyst bed volume and V̇ is the NH₃ volumetric flow rate at reaction conditions (400 °C and 1 atm pressure).

4.3. Experimental Conditions

Table 3 shows the experimental conditions for the reaction. The reactor temperature was set at 400 °C to allow vaporization of feedstock and simultaneously prevent extensive cracking reactions.²³ Previous studies on the ammonization of fatty acids reported a residence time range of 3–13 s as the sufficient contact time for fatty nitrile production.¹⁷ As such, the residence time in our experiments was set to 10 s by introducing 8.9 mL min^–1 NH₃ (at room temperature, 20 °C) into the reactor that was packed with 3.4 cm³ of catalyst. An NH₃/triglyceride molar ratio of 12 (4× stoichiometric amount) was used to allow the reversible reactions to proceed in the forward direction.

Table 3. Operating Conditions for the One-Step Vapor-Phase Nitrile Reaction of Triglycerides.

feedstock	coconut oil
reactor temperature (°C)	400
catalyst vol. (cm3)	3.4
NH3 flow rate (mL min–1)	8.9a
NH3 molar rate (mmol min–1)	0.369
triglyceride molar rate (mmol min–1)	0.092
NH3/triglyceride molar ratio	4
residence timeb (s)	10
NH3 linear velocity (mm s–1)	5

^aNH₃ flow rate at room temperature and atmospheric pressure conditions.

^bCalculated from eq 3.

4.4. Catalyst Characterization

Texture properties of catalysts were measured using N₂ adsorption–desorption isotherms at 77 K by a Micromeritics ASAP 2020 instrument. Prior to the analysis, the samples were pretreated at 350 °C for 3 h under vacuum to remove any adsorbed compounds. Multipoint adsorption isotherms were obtained on the samples using N₂ as the adsorbate. The BET equation was used to calculate the surface area, and the micropore volume was determined by the t-plot method.^27,28

The NH₃-TPD technique is a useful method to measure the catalyst acidity and the strength of acid sites.²⁹ In this study, a custom dynamically pumped TPD system was used as shown schematically in Figure S1. Prior to each TPD measurement, ∼130 mg of the catalyst was degassed at 350 °C under high vacuum (<10^–6 Torr) for at least 5 h. Then, 100 mg of the degassed catalyst was weighed and transferred into a borosilicate glass tube with an inner diameter of 7 mm and placed inside the TPD furnace. Thereafter, the system was evacuated for 30 min to remove trace air and H₂O. Subsequently, the valve connected to the main chamber (gas detector) was closed, and the NH₃ gas was introduced into the sample tube that was placed inside the furnace at 80 °C. During this step, NH₃ was allowed to adsorb on the catalyst surface for 1 h. Thereafter, the NH₃ injection was stopped, the main chamber valve (see Figure S1) was opened, and the sample tube was again pumped down to high vacuum for ∼1 h to reduce the NH₃ partial pressure to the background level (<10^–6 Torr). In other words, all “free NH₃” in the system was pumped out with the exception of the NH₃ adsorbed on the catalyst surface. Afterward, the catalyst was heated to 600 °C at a constant ramp rate of 5 °C min^–1. The NH₃ gas desorbing from the catalysts was monitored by a mass spectrometer (RGA 300, Stanford Research System) connected to the dynamically pumped high-vacuum TPD chamber (see Figure S1). Mass spectra in the range 1–50 atomic mass units (amu) were recorded in 3 s intervals. The concentration of the desorbed NH₃ was measured by the integrated-signal approach as reported previously.³⁰

4.5. Feedstock and Products Analysis

The coconut oil was transesterified to FAMEs and analyzed by gas chromatography using mass spectrometry (GC–MS) and GC–FID to identify and quantify the fatty acid constituents in the feedstock, as explained elsewhere.⁵

The GC–MS (Bruker, 450-GC equipped with 300-MS) analysis was also performed to identify the chemical constituents in the liquid products from the reaction. An Agilent DB-5MS-fused silica capillary column (length: 30 m, i.d.: 0.25 mm, and film thickness: 0.25 μm; Agilent Technologies, Santa Clara, CA) was employed, where the column was initially set at 30 °C for 10 min, then heated at 10 °C min^–1 to 300 °C, and finally held at 300 °C for 10 min. The injector, transfer line, ion source, and manifold were maintained at 300, 300, 150, and 40 °C, respectively.

To quantify the chemical compounds in the products, GC–FID (Shimadzu 2012 plus) with an RTX-biodiesel (Restek, Bellefonte, PA, USA) column (15 m length, 0.32 mm i.d., and 0.1 μm film thickness) was used. The column temperature was programmed as follows: initially set at 60 °C for 1 min, followed by a temperature ramp rate of 10 °C min^–1 to 370 °C and a final hold for 5 min. The injector and FID temperatures were maintained at 370 °C. The FID detector response was first calibrated using standards of known concentrations of FAME, fatty acid, fatty amide, fatty nitrile, monoolein, diolein, and triolein. The chemical compound concentrations in the samples (liquid products and feedstock) were estimated from the calibration curves as described previously.^5,31−33

Supporting Information Available

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsomega.7b01502.

• TPD system schematic; coconut oil composition; TPD desorption results; reaction scheme; GC–FID of products; detailed product composition; and energy consumption calculations (PDF)

The authors declare no competing financial interest.