Introduction

The oversupply of glycerol from biodiesel production drives research into finding value-added applications or products from glycerol1,2,3,4. One such approach is the production of triacetin (TA), which can increase the economic value of glycerol. Triacetin (TA), also known as glyceryl triacetate, is a versatile compound with a wide range of applications as a fuel additive, plasticizer, solvent, food additive, and emulsifier in various industries5. As a fuel additive, TA can enhance fuel performance by improving cold flow properties and viscosity, reducing engine knock, and increasing thermal efficiency in gasoline engines. These properties contribute to better energy conversion and lower emissions of nitrogen oxides (NOx), carbon monoxide (CO), and particulate matters6. Based on the various applications, the global market for TA was expected to grow at a compound annual growth rate (CAGR) of 5.42% from 2022 to 2030, potentially reaching a value of $473.79 million by 20307.

Glycerol esterification with acetic acid is the most common route for industrial TA production. The reaction involves three reaction steps, forming the intermediate products of monoacetin (MA) and diacetin (DA) for the 1 st and 2nd reaction steps, respectively, and the product TA in the final step as illustrated in Eq. (1) to (3). Water is generated as a by-product for each step. Kale et al.8 reported that the Gibbs free energy of formation of MA, DA and TA were 19.2, 17.80, and 55.6 kJ/mol, respectively. This indicated that TA production was limited due to the highest TA’s Gibbs free energy of formation compared to other products. The TA production via this reaction often encountered low TA yield although the complete glycerol conversion could be achieved.

(1)
(2)
(3)

The most common catalyst used in the TA industrial process was concentrated sulfuric acid9. However, the use of concentrated sulfuric acid as a homogeneous catalyst poses the significant drawbacks, including equipment corrosion, side reactions such as polymerization and dehydration10,11,12difficult product purification, and environmental risks. For these reasons, heterogeneous catalysts have attracted considerable interest in enhanced sustainable development specifically for this process. Examples of promising heterogeneous catalysts for the efficient TA production included the mixed metal oxide catalysts of CeO2-ZrO2 for unsulfated and sulfated mixed oxide forms12. The highest yields of about 57% of DA and 21% of TA were achieved in 3 h. An activated carbon/UiO-66 composite catalyst also showed the great stability for at least three consecutive cycles, and the selectivity of TA at 180 min was about 18% using an acetic acid to glycerol molar ratio of 6 to 1 and 90°C13. Nevertheless, this catalyst required a complicated synthesis procedure and was not available on a commercial scale. Interestingly, Purolite C160 appeared as a potential cation exchange resin commonly used in esterification since it was inexpensive and conducive for the applications with different scales of production, e.g., from laboratory to the industrial levels14. The high acid site density of Purolite C160 accelerated the glycerol conversion without requiring additional solvent, leading to a potential catalyst candidate for DA and TA production. At 150 min, this catalyst gave 95% glycerol conversion with MA, DA and TA yields of 25, 57 and 13%, respectively at 110 °C using an acetic acid to glycerol molar ratio of 6 to 115.

Apart from the research efforts on catalyst development, the success of the TA production is also focused on the reactor and process design. The glycerol esterification with acetic acid consists of three reversible reactions, especially the chemical equilibrium is limited to produce the final TA product. Typically, in the conventional process, excessive acetic acid and inter-stage water removal are required to drive the reaction forward to produce TA. Some multifunctional reactors, such as an ultrasound-assisted reactor16a microwave reactor17 and a reactive distillation (RD)18 have been proposed to enhance TA production performance.

RD combines reaction and separation into a single unit (Fig. 1), representing a promising process intensification technology to overcome the chemical equilibrium issue19,20,21. It allows to reduce the excessive reactant and number of purification units, and improved process efficiency. Unlike the traditional reactors, the simultaneous reaction and separation in the RD provides the continuous removal of products from the reactive zone, promoting raw material conversion and product selectivity22,23.

Fig. 1
figure 1

Proposed RD configuration of glycerol esterification for TA production.

Full size image

High reaction efficiency for liquid-phase reversible reactions, especially an esterification, can be achieved in the RD24. Hasabnis and Mahajani25 proposed the TA production in a RD combined with an addition of ethylene dichloride as an entrainer and Amberlyst-15 as a catalyst using an acetic acid to glycerol molar ratio of 3 to 1. The large total number of stages of 45 with 32 reactive stages was required to achieve 100% selectivity of TA. The entrainer was required for this TA production via esterification to enhance the TA selectivity. However, this RD system with an entrainer was complicated for industrial operation due to the need for feeding of acetic acid, entrainer, and glycerol at different stages, as well as splitting the reflux to remove the entrainer and water. The separation of water from the entrainer in the distillate stream and reflux operations at the top of the column presented the realistic operational challenges26. Similarly, Li et al.18 used NKC-9 to catalyze glycerol esterification for TA production in a pilot-scale RD column. However, the experimental results could not achieve high TA purity because it would require a large number of reactive stages. This research also extended the number of stages in the simulation results to achieve about 99.5% TA purity which required more than 60 stages (50 reactive stages, 10 rectifying stages and at least 1 striping stage) with an acetic acid to glycerol molar ratio of 3:1 under the vacuum pressure of 35 kPa18.

Based on the reviewed information, it was evident that the utilization of RD significantly enhances the efficiency of TA production. However, the previous RD process design typically requires a large number of stages and an entrainer to facilitate the operation. These factors can negatively impact the overall efficiency of glycerol esterification for TA production. Consequently, optimizing the process to reduce these operating requirements has become a promising strategy to improve the feasibility and attractiveness of TA production, particularly for potential investors seeking cost-effective and scalable solutions.

The novelty aspect of this work is to simulate the RD column using Purolite C160 as a catalyst to gain economic benefits. Furthermore, a comprehensive parametric analysis, specifically for Purolite C160 utilized in the RD column has never been undertaken. Herein, the RD column packed with the Purolite C160 would be thoroughly designed using Aspen Plus. Various operating and design parameters, including the acetic acid to glycerol molar ratio, feed stage, reflux ratio, bottom to feed ratio, and the number of stages, were examined to assess the RD column’s performance and to optimize the TA production yield. The evaluation of process performances along with their comparative study can provide insights into the potential of utilizing RD for TA production in practical applications.

Methodology

Aspen Plus software (Aspen plus v14.0) was employed to model RD for the esterification of glycerol with acetic acid in this study (Table 1). The reaction mechanism as illustrated in Eq. (1) to (3) included the six reactions with six constituents: acetic acid (AA), glycerol (G), water (H2O), monoacetin (MA), diacetin (DA), and triacetin (TA). Table 1 shows all chemicals information and their boiling point at 1 and 0.7 bar, respectively as obtained from Aspen plus. In addition, to ensure the system was in vapor–liquid equilibrium (VLE), the Gibbs free energy of glycerol and acetic acid mixing for all mole fraction was provided. Negative values observed within the pressure range of 0.7–1 bar indicated that the system remained homogeneous in the liquid phase, as illustrated in Figure S1 in the Supplementary Materials. The operating pressure was insignificant effect on the Gibbs free energy of mixing because these liquids are nearly incompressible25,27. Therefore, the NRTL was selected as thermodynamic model for the conventional reactor, distillation and RD for this simulation.

Table 1 Basic information of the chemical components in this study.
Full size table

Reactive distillation simulation

RADFRAC model was used to simulate the RD column. Due to the strongly non-ideal behavior of glycerol esterification, the NRTL model was selected to predict the thermodynamic properties of the non-ideal reaction mixture28. The basis for RD column design was set at a glycerol feed rate of 1,000 kg/h to evaluate process performance, including conversion efficiency, product distribution, and energy requirements, ensuring optimal operating conditions for TA production.

Glycerol was introduced into the RD column at the first reactive stage, while acetic acid was fed into the bottom section of the reactive stage, as illustrated in Fig. 1. The column operated with partial reflux conditions, and Purolite C160, a solid acid catalyst, was packed within the reactive stages to catalyze esterification of glycerol with acetic acid. Unreacted acetic acid and water were vaporized and exited the column at the top, while all products, monoacetin (MA), diacetin (DA), and triacetin (TA,) were in the bottom stream of the RD column. The preliminary RD process for glycerol esterification required at least 30 theoretical stages to achieve complete glycerol conversion and high TA selectivity. Therefore, the RD column was designed with 32 stages: 1 stripping stage, 1 rectifying stage, and 30 reactive stages and operated at atmospheric pressure. A liquid hold-up time of 60 min for the reactive stages in RD was specified to ensure sufficient residence time for the reaction.

Kinetic data validation for purolite C160 catalyzed glycerol esterification with acetic acid

Purolite C160, a resin catalyst with a high acid-site density, was selected for TA production due to its commercial availability, cost-effectiveness, and suitable lifespan for industrial-scale production. Regarding its reaction performance, this resin catalyst provided high selectivity toward the desired product, leading to a higher yield of TA with minimal by-product formation. Natália et al.15 proposed the kinetics of Purolite C160 catalyzed esterification of glycerol as homogeneous pseudo-first-order as illustrated in Eq. (4) to (7).

$$\:{-r}_{g}={k}_{1}\left[GLY\right]-{k}_{-1}\left[MA\right]$$
(4)
$$\:{r}_{MA}={k}_{1}\left[GLY\right]-{k}_{2}\left[MA\right]-{k}_{-1}\left[MA\right]+{k}_{-2}\left[DA\right]$$
(5)
$$\:{r}_{DA}={k}_{2}\left[MA\right]-{k}_{3}\left[DA\right]-{k}_{-2}\left[DA\right]+{k}_{-3}\left[TA\right]$$
(6)
$$\:{r}_{TA}={k}_{3}\left[DA\right]-{k}_{-3}\left[TA\right]$$
(7)

Where \(\:{-r}_{g},{r}_{MA},\:{r}_{DA}\) and \(\:{r}_{TA}\) are the reaction rates of consumption of glycerol and formation of MA, DA and TA (mol/L.min).

The kinetic constants for the reactions were calculated using the Arrhenius equation as shown below.

$$\:{k}_{i}={A}_{i}exp(-{E}_{A,i}/R)$$
(8)

Where \(\:{k}_{i}\) is the reaction rate constant (min− 1), \(\:{\:A}_{i}\) is the pre-exponential factor (min− 1), \(\:{E}_{A,i}\) is the activation energy (kJ/mol). The subscript i denotes the chemical reaction number corresponding to Eqs. (1)-(3).

The values of the pre-exponential factors and activation energies for the esterification of glycerol with acetic acid are summarized in Table 2. Since the esterification is reversible, the subscripts with negative numbers, e.g., −1, −2, and − 3 signify the reverse reactions. To validate the kinetic parameters for this study, the simulation results obtained from these kinetic parameters were compared with experimental data from the literature15. Additionally, the process performance was assessed based on the glycerol conversion and product yield as given in Eqs. (9) and (10), respectively.

$$\:{Conv}_{G}={(F}_{G,in}-{F}_{G,out})/{F}_{G,in}\times\:100\%$$
(9)
$$\:{Yield}_{i}={(F}_{i,out}/{F}_{G,in}\:)\times\:100\%$$
(10)

Where \(\:{F}_{G,in}\) is the glycerol molar flow rate in the feed stream, \(\:{F}_{G,out}\) is the glycerol molar flow rate in the product stream, and \(\:{F}_{i,out}\) is the molar flow rate of the i component in the product stream.

Table 2 Values of pre-exponential factor (Ai) and activation energy (EA, i) of glycerol esterification with acetic acid based on Eqs. (1)-(3)15.
Full size table

Conventional TA production process

The conventional TA process using Sn-DTP/K-10 catalyst was purposed by Pandit et al.29. This catalyst was synthesized by impregnating K-10 montmorillonite clay with tin (II) chloride (SnCl2) and dodecylthiophosphonic acid (DTP) or a similar organophosphonic acid, followed by drying and calcination30. A special synthesis procedure was required for this catalyst, which could lead to high production costs, potentially increasing overall production expenses as compared to the commercial catalyst. Additionally, Sn-DTP/K-10 was more likely unstable or less effective under certain operating conditions in the presence of significant amount of water29. In contrast, Purolite C160 is a polystyrenic macroporous, strong acid cation exchange resin in the sodium form, known for its high ion exchange capacity and excellent resistance to osmotic and thermal shocks. These properties make it highly suitable for a wide range of industrial applications. For instance, demineralization processes to produce high-purity water31serving as an effective acid catalyst in esterification and transesterification, removing impurities to purify biodiesel, and softening water by eliminating hardness ions like calcium and magnesium in both domestic and industrial systems to prevent scaling and corrosion in equipment29. These diverse applications highlight the versatility and reliability of Purolite C160 in various sectors that require efficient ion exchange and catalytic performance. In addition, a similar catalyst should be used to compare the process performance. Therefore, the Purolite C160 also used as a heterogenous exchange resin catalyst for glycerol esterification to produce TA as the conventional process simulation.

To assess the feasibility of industrial scale production, since the current biodiesel production typically yields approximately 80,000 to 100,000 tonnes per year32with glycerol constituting about 10 wt% of the total biodiesel output, the process was initially designed using a conventional method, targeting a glycerol mass flow rate of 1,000 kg/h (10.86 kmol/h)33. A feed containing an acetic acid to glycerol molar ratio of 6:1 was preheated to 60℃ before feeding to the first esterification reactor (R-101) operated at 110℃ to produce MA, DA, and TA as presented in Fig. 2. The product stream was fed to the distillation column (C-201) with 5 theoretical stages to remove water, a by-product from the 1 st esterification. Water was vaporized to the top column stream (110℃) with unreacted acetic acid. The distillation column (C-501) was further used to recover unreacted acetic acid back to mix with the main product stream (MA, DA, TA, and unreacted glycerol) in the mixer (M-201) and preheated at E-301 at 60℃ for feeding to the 2nd esterification reactor (R-301) operated at 110℃ (similar to 1 st reactor). Based on the evaluation of the T-x-y diagram (Figure S2A), the binary mixture of acetic acid and water exhibited the potential for azeotropic distillation, which could hinder the achievement of the desired separation purity. However, the composition of the stream fed into the distillation column (C-501) included MA, DA, TA, and glycerol. These components effectively acted as entrainers, altering the vapor–liquid equilibrium and thereby suppressing the azeotropic behavior between acetic acid and water, as shown in Figure S2A.

After this process, the stream was sent to the distillation column (C-401) to purify TA. In addition, the distillation column (C-401) was required to increase purity of unreacted acetic acid (stream 13) and decrease acetic acid concentration in the wastewater (stream 8). The distillation column operation in this process was optimized through Aspen Plus simulation to maximize production yield and minimize energy consumption in the system.

Fig. 2
figure 2

Conventional process for TA production via glycerol esterification using Purolite C160 catalyst.

Full size image

Production of TA in the RD column

The RD process for esterification of glycerol with acetic acid was developed as depicted in Fig. 3. The initial design, a feed stream with the similar acetic acid to glycerol molar ratio of 6:1 was introduced to the RD column (C-101) under atmospheric pressure. The distillate stream from the RD column, consisting of water and unreacted acetic acid (90℃), was sent to the distillation column (C-501) to separate and recycle the unreacted acetic acid back into the process. This is a crucial step because the accumulation of water can shift the equilibrium of the reversible glycerol esterification, potentially lowering the overall product yield. Additionally, the optional distillation column (C-502) was utilized, particularly for scenarios of the intermediate products (MA and DA) production to enhance the flexibility and efficiency of the process.

Fig. 3
figure 3

TA production using Purolite C160 catalyst in the RD process.

Full size image

In contrast, this research focuses on the optimization of glycerol esterification process from an industrial perspective, with the goal of minimizing operational complexity and emphasizing on process simplification such as avoiding the use of external entrainers such as ethylene dichloride25 or hexane34 to provide a more practical and economically viable solution. The approach centers on leveraging the inherent advantages of RD, such as simultaneous reaction and separation, to achieve high conversion rate and high product purity without the need for additional chemicals. The key solution to simplify RD with minimize the number of total stages was relied on the operating parameters in terms of acetic acid to glycerol molar ratio, feed position, bottom to feed ratio, reflux ratio and column pressure. To simplify the process in this study, the system was designed for optimal operability by selecting a basic RD configuration. This decision was because previous research on triacetin production still involved complex process designs, which made production control more challenging.

Results and discussion

Kinetic data validation for glycerol esterification using purolite C160

The accuracy of the kinetic data for glycerol esterification with acetic acid using Purolite C160 was verified with experimental data obtained using a batch reactor15. Figure 4 compares the glycerol conversion from simulations with the experimental data15 across various temperatures and reaction times at a fixed acetic acid to glycerol molar ratio of 6:1.

Fig. 4
figure 4

Kinetic model validation for glycerol esterification using Purolite C160 in a batch reactor (data adapted from15 using an acetic acid to glycerol molar ratio of 6 to 1 at atmospheric pressure. (A) glycerol conversion lines (-) represent simulation results and symbols represent experimental results) and (B) comparison product yields between experimental results and simulation results.

Full size image

Figure 4 (A) shows the validation of glycerol conversion with the calculated root mean square error (RMSE) for glycerol conversion of 4.42, 4.36, and 3.82 at the reaction temperatures of 90, 100, and 110 °C, respectively. These small RMSE values demonstrated that the kinetic model provided a reliable prediction of glycerol conversion. Furthermore, the predicted yields of MA, DA, and TA at 110 °C were 25, 60, and 15%, respectively as closely aligned with the experimental yields of 25, 57, and 13% as depicted in Fig. 4 (B). This strong agreement between predicted and experimental results validates the kinetic parameters listed in Table 1 and the rate expressions provided in Eqs. (4)-(7). Consequently, these validated parameters were employed to simulate the performance of glycerol esterification using Purolite C160 in both conventional and RD processes, ensuring accurate and reliable process design and optimization. In the design, the amount of catalyst was specified to be consistent with the quantity proposed in the kinetic data. Accordingly, a Purolite C160 catalyst loading of 1.24 kg per stage was employed, in line with a catalyst to reaction mixture volume ratio of 0.05 g/mL (corresponding to 23 wt% of glycerol), in agreement with the experimental results15. In this study, the stage height was assumed to be 0.4 m per stage, following a standard design parameter. Accordingly, the liquid height was taken as approximately 10% of the tray height, or 0.04 m per stage, corresponding to a liquid holdup volume of 0.0248 m³ per stage. The calculated column diameter was approximately 0.89 m, which is suitable for a semi-industrial (pilot-scale) application35. The dimensions of the RD column were designed to accommodate the catalyst loading appropriately and to provide a liquid hold-up time of 60 min for the reactive stages.

TA production via purolite C160-catalyzed glycerol esterification of acetic acid in conventional and RD processes

Two esterification reactors based on the conventional process flow diagram (Fig. 3) were required to achieve a glycerol conversion of 93.13% and a TA selectivity of only 55.02%. In addition, the evaluation of the x–y diagram for the binary mixture of acetic acid and water at 1 bar revealed the potential for an azeotropic mixture at an acetic acid concentration of approximately 36–38 wt%, as shown in Figure S2A. This could hinder the achievement of the desired separation purity. However, the composition of the stream fed into the distillation column (C-501) also included MA, DA, TA, and glycerol. These components effectively acted as entrainers, altering the vapor–liquid equilibrium and thereby suppressing the azeotropic behavior between acetic acid and water, as shown in Figure S2B. As a result, the acetic acid content in the distillate stream (stream 8) was 7.69 wt%, as shown in Table S2 of the Supplementary Materials.

The superior process performance was observed in the RD process. The simulation results indicated that the higher glycerol conversion of 99.99% and TA selectivity of 99.94% were achieved in the RD column with a configuration comprising 32 total stages, 1 stripping stage, 30 reactive stages, and 1 rectifying stage using an acetic acid to glycerol molar ratio of 6:1. To support this evidence, the values of the reaction rate constants for each step was calculated at the selected reaction temperatures of 50, 100 and 150 °C using the Arrhenius equation, as presented in Table S1 in the Supplementary Materials. The reaction rate constant for backward reaction was higher that of forward reaction rate constant for 3rd step indicated that the high TA yield could not be achieved in the conventional batch system. In contrast, owing to the unique characteristics of RD, where vaporized components (especially by-product water) move to the upper stages, the reaction equilibrium can shift forward accordingly. This facilitates a more favorable environment for TA product formation and enhances overall reaction performance. In addition, the distillate stream from the RD column contained 4.18 wt% acetic acid, which was far from the azeotropic composition. Therefore, efficient separation of acetic acid from water in C-501 was achieved, with the water stream containing less than 0.01 wt% acetic acid, as shown in Table S3 of the Supplementary Materials.

The inclusion of reactive stages within the column allowed for simultaneous reaction and separation, which minimized the need for additional separation units, thus enhancing the overall efficiency of the process. The stripping and rectifying stages further ensured the removal of light and heavy components, respectively, leading to high TA product purity and separation of the lighter components of acetic acid and water mixture, respectively. This present design offered a practical and efficient solution as compared to the conventional process. Souza et al.34 also proposed a glycerol esterification process using a RD column with a lower number of stages, specifically 16 stages. However, their approach required the addition of hexane as an entrainer to enhance production efficiency leading to additional complexity of the process. This complexity arose from the need to handle, separate, and recover the entrainer, increasing both operational and capital costs.

Therefore, it was observed that replacing the reactor and separator in the conventional process with an RD significantly improved the reaction efficiency at the same acetic acid to glycerol feed ratio. This change was implemented to reduce production costs. Further optimization was performed with the additional parameters such as temperature profiles, feed location, reflux ratio, and column pressure. The total number of stages was selected at 32, since it was slightly improved in the glycerol conversion and TA purity when the reactive stages greater than 32 stages as can be seen in Figure S3 in Supplementary Materials.

This should be noted that water was produced as a by-product for glycerol esterification, which was dissolved in the acetic acid phase. Therefore, the water was separated from the unreacted acetic acid stream before recycling. The unreacted acetic acid with a purity of up to 95% was recovery back to RD with simultaneously discharging the water by-product. This again stressed the benefit of RD in overcoming an azeotrope that occurred during distillation36. This study also aimed to minimize the total number of stages in the RD column to reduce complexity and potential inefficiencies. Although rectifying stages typically assist in the separation of acetic acid and water, a well-designed RD process leaves only a small amount of unreacted acetic acid. Other operating parameters can be optimized to maintain the highest possible TA production performance. Therefore, increasing the number of rectifying stages was deemed unnecessary. The same rationale applies to the stripping section, as the design resulted in a high concentration of TA at the bottom of the RD column, eliminating the need for additional stripping stages.

Effect of acetic acid and glycerol molar feed ratio

The feed rate ratio of 6:1, equivalent to that of the conventional process, was used to simulate the preliminary design of the RD process for the TA production, and it could achieve the complete glycerol conversion. Therefore, the effect of acetic acid to glycerol molar ratio, ranging from 1:1 to 9:1, on the process performance in the RD column configured with 30 reactive stages, 1 rectifying stage, and 1 striping stage, was investigated as depicted in Table 3.

Table 3 The simulation results of RD process using different acetic acid to glycerol molar ratios.
Full size table

The almost complete glycerol conversion (> 99.93%) was achieved when feeding with an acetic acid to glycerol molar ratio higher than 5:1. However, the significant amount of unreacted acetic acid was higher compared to the theoretical acetic acid to glycerol molar ratio of 3:1 (Table 3). Additionally, the higher acetic acid to glycerol molar ratio increased the energy requirement of the RD process, as indicated by the remarkable increase in the reboiler duty with the excess acetic acid. Furthermore, the excessive acetic acid to glycerol molar feed ratio led to decrease in TA purity. For instance, when the acetic acid to glycerol molar ratio of 5:1 ratio was applied, the low TA purity of 70.53%, was observed, resulting in the complicated separation and purification of TA, and increased process complexity and operational costs. Consequently, the acetic acid to glycerol molar ratio of 3:1 was selected, providing 93.72% glycerol conversion, 65.08% TA selectivity and 93.02% TA purity, balancing conversion efficiency with process feasibility and cost-effectiveness.

The optimum feed ratio was a critical design parameter that needed to be considered from an industrial perspective dealing with the operation cost. In the RD process, it was possible to adjust the feed ratio close to the theoretical value, thereby optimizing efficiency. The optimum feed ratio of 3:1 in this study aligned with previous research18,25which provided the high glycerol conversion and TA purification efficiency. The identical molar ratio of 3:1 was also used for the conventional process. The simulation results indicated that the TA yield in the conventional process was significantly reduced. At reaction temperature of 110 °C, the glycerol conversion was 88% while the product yields were 36% for MA, 46% for DA, and 6% for TA. This should be because the rate backward for glycerol esterification with acetic acid using Purolite C160 in 3rd step was dominated compared to the forward rate. The other desired parameters were further investigated to determine the optimum condition for TA production in the RD.

Effect of acetic acid feed location

The feed location is crucial as it affects both the extent of reactions and the energy used in RD operations. To evaluate these effects, the locations of acetic acid were investigated using an RD column configured with 30 reactive stages, 1 rectifying stages, and 1 stripping stage for the production of TA via glycerol esterification, as illustrated in Fig. 5.

Fig. 5
figure 5

Conversion of glycerol, TA purity and TA selectivity at various feed locations of acetic acid (A) and unreacted acetic acid in distillate at various feed locations of acetic acid (B) (1 rectifying stage, 30 reactive stages, 1 stripping stage, an acetic acid to glycerol molar feed ratio of 3:1, a reflux ratio of 3:1 and a column pressure of 1 bar).

Full size image

Modifying the feed location of glycerol which was the heavier component at the top of the distillation column could increase the likelihood of reaction occurrence37. The glycerol feed position in the appropriate upper or middle section of the column promoted a favorable concentration gradient, enabling more effective esterification38. As heavier phase glycerol flowed downward, it reacted with acetic acid along the column, leading to increase in the conversion. Therefore, the glycerol feed was fixed at the 2nd stage, while the acetic acid feed locations were varied from the 2nd to the 31 st stage (from the beginning to the end of the reactive section). The acetic acid feed position slightly influenced the glycerol conversion which approached complete conversion. While the TA selectivity was increased from 48.52 to 65.07% when the acetic acid feed position was changed from 2nd to 31 st stage, as illustrated in Fig. 5A.

Although the RD could facilitate the reversible glycerol esterification to complete glycerol conversion, the TA selectivity was limited by the glycerol to acetic acid molar ratio as indicated in Table 3. On the other hand, acetic acid feed position significantly influenced the TA purity and the reboiler duty. When acetic acid feed location was changed from 2nd to 31 st stage, the TA purity was increased from 57.26 to 93.02% (Fig. 5B), as well as an increase in the reboiler duty from 276 to 378 kW. Additionally, the acetic acid feed location also had a considerable effect on the acetic acid content in the distillate product. When acetic acid was fed at the 31 st stage (last stage of reactive stage) at the bottom section, the amount of unreacted acetic acid rising to the top was significantly lower at 4.08 kmol/h compared to the acetic acid feeding at the 2nd stage (8.69 kmol/h). The lower amount of acetic acid content in the distillate stream led to the lower separation duty of the subsequent distillation column. This resulted in the decrease of energy consumption for the separation of the unreacted acetic acid and discharge water from the system (C-501 in Fig. 3). This is because the low-boiling component (acetic acid) readily vaporized and rose through the column, promoting longer contact with the descending high-boiling component (glycerol) from the top can enhance the reaction efficiency of glycerol esterification as well as minimize the acetic acid content in the distillate stream (stream 4 in Fig. 3). Although this approach required higher energy consumption, the operation of the RD in terms of control and reactant feed management, the feeding at the top and final reactive stage is more practical in an industrial scale. Therefore, the optimal feed positions for glycerol and acetic acid were selected to be the 2nd and 31 st stages, respectively to achieve 93.72% glycerol conversion, 75.76% TA purity.

Effect of bottom to feed ratio

The bottom to feed ratio in the RD column is also a critical parameter which has significant influence on the balance between reaction performance and separation efficiency. A higher bottom to feed ratio can increase the liquid flow in the product stream, while the lower ratio can restrict separation efficiency and reduce reactant residence time. Therefore, the optimizing bottom to feed ratio was investigated to ensure efficient reaction conversion, enhance product purity, and minimize energy consumption.

Figure 6 illustrates the influence of the bottom to feed ratio on bottom temperature in the RD column, and glycerol conversion and product selectivity. When the bottom to feed ratio was increased, the bottom temperature of the column was decreased which was attributed to the lower energy requirement for vaporizing components to the top of the column (Fig. 6A).

Fig. 6
figure 6

Bottom temperature at various bottom to feed ratio (A) and conversion of glycerol, TA purity, TA selectivity yield and bottom of RD column temperature at various bottom to feed ratio (B) (1 rectifying stage, 30 reactive stages, 1 stripping stage, an acetic acid to glycerol molar feed ratio of 3:1, a reflux ratio of 3:1 and a column pressure of 1 bar).

Full size image

The temperature variation significantly influenced the reaction efficiency and the purity of the distillation process. Figure 6B revealed that the operation at a higher bottom to feed ratio of 0.90, as referred to the larger bottom product fraction, lowered the glycerol conversion to 82.68%. In addition, the selectivity of TA (target product) was relatively low at 49.30%, indicating that a high bottom to feed ratio might not be suitable for the TA production under this RD configuration. Conversely, the lower bottom to feed ratio of 0.25 significantly enhanced both glycerol conversion, TA selectivity and TA purity, reaching 99.82, 94.52 and 97.84%, respectively. Nevertheless, the substantial heat duty of 641.65 kW was required to vaporize the chemical compounds within the column under this condition. In addition, the purity of TA remarkably depended on the amount of reboiler duty for RD operation. At lower bottom to feed ratios, higher reboiler duty was required. Due to the fact that RD needed more energy to heat and evaporate all components within the column, corresponding to the operating ratio which tended to increase the column temperature (Fig. 7B). These findings suggested a trade-off between product selectivity, conversion efficiency, and energy consumption. While a lower bottom to feed ratio was favorable for achieving high TA selectivity and glycerol conversion, it demanded significantly higher heat duty. This was achieved by facilitating the vaporization of heavier components to the top of the column, which enhanced the contact time and elevated the temperature between glycerol and acetic acid in the presence of the catalyst39. In addition, this prolonged interaction improved the efficiency of the esterification process, while the removal of heavier products from the bottom of the column ensured a more uniform distribution of volatile reactants throughout the column40. Previous research proposed the RD process for TA production with different constraints. A high bottom to feed ratio18 required a large number of reactive stages because the heat energy distribution from a high bottom to feed ratio results in lower column temperatures, slowing the reaction compared to a lower ratio. Despite this, the complete glycerol conversion (98.99%) was achieved but lower TA purity of 85.88% was observed with the required large number of reactive stages. This could be due to the high bottom to feed ratio resulted in the lower temperatures distribution to inhibit the reaction rate compared to a lower bottom to feed ratio. Therefore, to minimize the number of reactive stages, a lower bottom to feed ratio should be used to compensate by increasing the operating temperature and enhancing reaction efficiency. The bottom to feed ratio of 0.25 was selected for further RD improvement.

Effect of column pressure

In the RD process, column pressure is a critical operating parameter. It directly influences the temperature distribution within the column, thereby affecting both the reaction rate and the reaction equilibrium. Additionally, the setting of column pressure is constrained by the thermal stability of the catalysts used and product decomposition.

Figure 7 indicates that the reducing pressure inside the RD column resulted in a gradual decrease in the temperature within the column. At a pressure of 1 bar, the bottom temperature of the column was 242 °C, which was relatively high and undesirable as TA product could initially degrade at the temperature of 215°C41. Therefore, it is necessary to lower the bottom temperature.

Fig. 7
figure 7

Conversion of glycerol, TA purity, TA selectivity and bottom column temperature at various pressures (1 rectifying stage, 30 reactive stages, 1 stripping stage, an acetic acid to glycerol molar feed ratio of 3:1 and a reflux ratio of 3:1).

Full size image

One effective approach is by reducing the pressure within the RD column. When the RD pressure was reduced to 0.7 bar, the bottom temperature decreased to 209 °C, which was below the threshold at which TA decomposition occurs. This phenomenon is explained by the relationship between pressure and temperature. When the pressure was reduced, the bubble point of a mixture was decreased, allowing it to evaporate at a lower temperature. Consequently, the temperature inside the distillation column also decreased. Therefore, the vacuum pressure of 0.7 bar was selected to achieve glycerol conversion of 99.83%, TA selectivity of 96.96% and TA purity of 98.7%.

Effect of reflux ratio

The reflux ratio plays a significant role in both the progression of the slow reaction and the reboiler duty in a RD process. The effect of the reflux ratio on the TA purity and reboiler duty was investigated in a range of 1 to 9.

Figure 8 demonstrates that increasing the reflux ratio in the RD column enhanced the purity of the TA product. It was observed that a higher reflux ratio facilitated a greater amount of unreacted acetic acid to be recycled back into the column, promoting its participation in the chemical reactions. In addition, the higher reflux could promote higher efficiency for the consecutive reactions in the 2nd and 3rd steps to further produce DA and TA, respectively. As can be seen, there was a slight increase in the TA selectivity when reflux ratio was increased. As a result, the purity of TA in the bottom section of the RD column significantly increased with reducing the presence of impurities such as acetic acid, DA and MA.

Fig. 8
figure 8

Conversion of glycerol, TA purity, TA selectivity and column reboiler duty at various column reflux ratio (1 rectifying stage, 30 reactive stages, 1 stripping stage, acetic acid to glycerol molar feed ratio of 3:1 and column pressure 0.7 bar).

Full size image

At a reflux ratio of 7.0, the TA product achieved a purity of 99%. Although increasing the reflux ratio improved the product purity, it also raised the energy demand of the system. Specifically, a higher reflux ratio required a larger reboiler duty to evaporate more amount of recycled material. It should be noted that the use of a high reflux ratio caused both acetic acid and water to be recycled back into the RD column. This resulted in the dilution of the reactants, leading to a slight decrease in glycerol conversion. In this study, a reflux ratio of 7.0 was selected as the optimum condition for producing TA with 99% purity, with an associated energy requirement of 1,367 kW. This balance between product purity and energy efficiency represents a key finding of the research.

Production of high purity of TA in the RD column process

Based on the study of the design parameters, the production of TA using RD for glycerol esterification demonstrated various modifications, as outlined in Table 4. Each modification of the design parameters was interrelated and significantly influenced the production efficiency. The optimization of the RD operation aimed to achieve lower production costs while maintaining high TA purity and ensuring a high glycerol conversion rate. This approach was investigated by enhancing the economic feasibility of the process without compromising product quality or reaction efficiency.

Table 4 Summary of optimized design parameters for the RD process in glycerol esterification.
Full size table

The RD column consists of a total of 32 stages, including 1 rectifying stage, 30 reactive stages, and 1 stripping stage was configurated for maximum production efficiency. Acetic acid was fed at the 31 st stage (the final stage of the reactive zone) with an acetic acid to glycerol feed molar ratio of 3:1. The bottom to feed ratio was optimized to the lowest feasible value for the reaction, which was 0.25. The composition profiles are illustrated in Figs. 9. High TA purity was observed at bottom stream (32nd) while acetic acid-water mixture was removed at the top with 4.18 wt% of acetic acid. Since the RD pressure column was selected at 0.7 bar, the maximum operating temperature was 209 °C to avoid the degradation of TA as the desired product as depicted in Fig. 10. In this study, it was also found that the efficiency of the glycerol esterification could be further enhanced by recycling unreacted acetic acid from the top of the column back into the reaction zone. Specifically, increasing the reflux ratio leads to improved process efficiency. At a reflux ratio of 7, the highest TA yield and purity were achieved. Under these optimum conditions, the RD process achieved a glycerol conversion rate of 99.20% and a TA purity of 99.00% as summarized in Table S2 of the supplementary material. Therefore, this RD configuration with the optimum condition was used for the preliminary techno-economic analysis in the next section.

Fig. 9
figure 9

Composition profiles of the constituents in the reactive distillation column for the glycerol esterification.

Full size image
Fig. 10
figure 10

Temperature profile in the reactive distillation column for the glycerol esterification.

Full size image

Performance comparison of different processes for glycerol esterification to produce TA

After optimizing the glycerol esterification process for both conventional and RD processes, a preliminary techno-economic analysis was conducted to quantitatively assess the commercial feasibility of the process. In the economic analysis, the catalyst cost was excluded from the operating expenses, as the same amount of catalyst was utilized in both processes, ensuring a fair and consistent comparison.

The capital costs and the total capital investment were calculated as the sum of the costs for all process equipment required to process 1,000 kg/h of glycerol, based on the data provided in Table 5. The equipment cost was determined using Aspen Plus software, with the Aspen Process Economic Analyzer (APEA) integrated tool employed for the accurate cost estimation. This approach provided a comprehensive evaluation of the economic viability of the optimized processes. The results indicated that the conventional process required higher energy consumption than the RD process, both in terms of equipment energy demands and the number of unit operations involved. Specifically, acetic acid recovery in the conventional process relied on distillation, which required more severe operating conditions and higher energy input due to the greater concentration of unreacted acetic acid. In contrast, the RD process yielded only a trace amount of unreacted acetic acid in the water stream, enabling easier separation and reducing the potential for azeotrope formation, as illustrated in Figure S2A. As a result, the distillation column (C-501) in the conventional process consumed more energy than that in the RD process.

Table 5 Process design parameters for different processes using purolite C160 catalyst for esterification of glycerol to product TA.
Full size table

Figure 11 presents the comparative analysis of the capital cost between the conventional process and the RD process for TA production. The evaluation was performed using the APEA under the conditions expressed in Table 4, incorporating equipment costs, indirect costs, and installation costs. The conventional process, comprising six unit operations, incurred the estimated capital cost of approximately USD 5,805,622. These results were consistent with prior economic assessments of comparable conventional processes, where the total capital investment reached USD 5,488,647.23 of the total equipment cost33. In contrast, the RD process integration of reaction and separation into a single column can reduce the number of unit operations from six to three. Remarkably, the RD process achieved a glycerol conversion rate of 99% and produced TA with a high purity of 99%, utilizing only 32 total stages under 0.7 bar of column pressure without the need of the entrainer. This represents a significant improvement over previous studies19,20which required more complex setups. Furthermore, the streamlined design of the RD process resulted in a substantial reduction in capital costs, estimated at USD 1,123,942 a reduction of approximately 79.16% compared to the conventional process. These findings underscore the economic superiority of the RD process for TA production, demonstrating its potential as a cost-effective and efficient alternative. The significant reduction in capital investment, coupled with high conversion rates and product purity, suggested that the RD process as a highly attractive option for industrial-scale implementation and investment.

Fig. 11
figure 11

Breakdown of capital costs in the conventional process and RD process.

Full size image

Moreover, the APEA was utilized to estimate the operating costs required to achieve the desired economic efficiency for both processes, as outlined earlier. The operating costs were calculated by summing the utility costs across all operating units, waste treatment, maintenance and repairs as well as supplies. The results revealed that the conventional process incurred an annual operating cost of approximately USD 3,365,000 which was consistency to the previous work33. USD 3,248,378.53 was reported based on the similar operating consideration in case of utilization acetic acid as feedstock. Remarkably, the RD process required a significantly lower annual operating cost of about USD 1,873,000. This stark contrast highlights the economic inefficiency of the conventional process compared to the RD process.

Table 6 presented a comparison of process performance, including the total number of unit operations and the corresponding production rate. The higher operating costs of the conventional process can be attributed to its reliance on multiple utilities to adjust temperatures at various stages, as evidenced by the greater number of unit operations involved (6 equipment). For instance, certain equipment in the conventional process required the cooling unit to prepare for the reaction, followed by reheating, which significantly increased the overall energy consumption and costs. In contrast, the RD process integrated reaction and separation into a single column, assisted the primarily temperature adjustment within the column itself to minimize energy demands and operational complexity, leading to substantially lower operating costs.

Additionally, the specific production cost was calculated using Eq. (11), further emphasizing the economic advantages of the RD process. These findings underscore the RD process as a more cost-effective solution for TA production, making it a highly attractive option for industrial implementation and investment. The significant reduction in both capital and operating costs, coupled with the process’s efficiency and simplicity, positions RD as a superior alternative to conventional methods.

$$\:Specific\:production\:cost=\:\frac{Operating\:cost}{\left(Production\:time\right)\times\:\left(Product\:in\:Mass\right)}$$
(11)
Table 6 Process performance comparison for TA production.
Full size table

To evaluate and compare the production efficiency of both processes, this study established a consistent production timeframe of 20 h per day and 365 days per year for each process42. Simulations were conducted to estimate the production costs, enabling a direct comparison of economic feasibility at the same glycerol feed rate (1,000 kg/h). The results revealed that the conventional process required a specific production cost of USD 0.347/kg of TA, while the RD process demonstrated a significantly lower specific production cost of USD 0.094/kg of TA. Notably, this cost was even more competitive than the RD process proposed by Li et al.18which reported a specific production cost of USD 0.101/kg of TA. The substantial cost reduction achieved in this study can be attributed to two key factors: (1) the reduction in the number of theoretical stages in the RD column and (2) the optimization of the bottom to feed ratio for glycerol esterification. These improvements not only streamline the process but also enhance its economic viability. The findings highlight the potential of the optimized RD process as a cost-effective and efficient method for TA production, offering significant advantages over both conventional and previously reported RD processes. This makes it a highly promising option for industrial-scale implementation and investment.

Conclusions

This study simulated the TA production using a RD column packed with Purolite C160 catalyst for the esterification of glycerol with acetic acid using Aspen Plus software. The optimized RD design features a total of 32 stages (1 rectifying stage, 30 reactive stages, and 1 stripping stage), operating under the stoichiometric ratio of acetic acid to glycerol molar ratio of 3:1. Key design parameters include glycerol feed at the 2nd stage, acetic acid feed at the 31 st stage, and a bottom to feed ratio of 0.25. This configuration not only enhanced reaction efficiency, but also reduced excess acetic acid consumption, simplified purification challenges, and lower raw material costs. The optimized RD process achieved exceptional results, including a 99.20% glycerol conversion 99.41% TA selectivity and 99.00% TA purity. The RD process demonstrated a significant reduction in the capital costs to lower the number of unit operations from six to three, compared to the conventional process. The capital cost was reduced by an estimated 79.16% for a 1,000 kg/h glycerol feed rate compared to the conventional process. These findings underscore the RD process as a highly efficient, cost-effective, and sustainable solution for TA production. The process not only addresses economic and environmental challenges but also offers a scalable and attractive investment opportunity for large-scale industrial applications, setting a new benchmark for glycerol esterification processes.