Article pubs.acs.

org/IECR

The Phase Behavior Eect on the Kinetics of Transesterication Reactions for Biodiesel Production
Stephen N. Csernica and James T. Hsu*
Deparment of Chemical Engineering, Lehigh University, 111 Research Drive, Bethlehem, Pennsylvania 18015, United States ABSTRACT: The phase behavior eect on the kinetics of transesterication reactions is investigated in this study. First, the partition coecient of the catalyst, KOH, was determined. It was found that the vast majority of the catalyst resides in the methanol phase; thus this phase is taken as the primary reaction volume. Mass transfer studies of soybean oil triglycerides show that triglycerides are in fact soluble in methanol. Additionally, it is shown that the rate of mass transfer and the equilibrium concentration of triglycerides in methanol increase with increasing methyl ester concentration. Establishing the methanol phase as the primary reaction volume facilitated the thorough investigation of the rate-limiting step. First, the terms mass transfer control and reaction control are nally dened for transesterication reactions. Once properly dened, it is shown that the sigmoidal, or Sshaped, methyl ester production curve that is so often observed is not necessarily caused by a transition from a mass transfer controlled regime to a reaction controlled regime, as is so often speculated in the literature. Rather, it is shown that the cause of the acceleration, regardless of the rate-limiting step, is caused by a transition from a two-phase system to a single-phase system.

INTRODUCTION Biodiesel, also known as fatty acid methyl esters, is a possible alternative to petroleum-based diesel fuel due to its numerous ecological advantages. Biodiesel is often produced by the transesterication of triglycerides from vegetable oils with a short-chained alcohol, most commonly methanol, in the presence of a catalyst. Many dierent catalysts have been reported, ranging from the common homogeneous alkaline and acidic catalysts, to various heterogeneous catalysts.14 Initially, the transesterication reaction system contains two nearly immiscible phases: the triglyceride-rich oil phase and the alcohol phase. It has been hypothesized that the reaction is initially mass transfer controlled due to the low miscibility of the two reactant phases and subsequently slow initial reaction rate.37 Noureddini and Zhu hypothesized that the reaction rate transitions from the mass transfer controlled regime to a faster reaction controlled regime since the methyl ester production curve displays a sigmoidal shape.4 They noted that when samples were extracted during the kinetically slow regime, i.e., what they called the mass transfer controlled regime, the reactants separated into two distinct phases. However, when samples were extracted during the kinetically fast regime, i.e., what they called the reaction controlled regime, the components existed as a single phase. They suggested that the methyl esters act as a mutual solvent for the reactants and allow for the formation of single phase; other authors have suggested similar hypotheses.49 Similar studies have been carried out to determine the relationship between the reaction rate and agitation intensity.4,6,10 It is generally observed that the eect of mixing is most signicant during the initial stages of the reaction. When the mixing is strong, the initial lag in methyl ester production disappears.4,6 This presumably is due to the increased mass transfer that results from high agitation. Boocock et al. also showed that the initial lag in the methyl ester curve disappears when butanol is selected as the alcohol source.7 This is because
2012 American Chemical Society

the butanolsoybean oil system is completely miscible. Additionally, many authors have reported the use of cosolvents, such as tetrahydrofuran, dimethyl ether, and propane, to obtain single-phase systems.7,1113 It should be noted that while successfully increasing the mixing between the oil and alcohol, the addition of cosolvents adds additional costs to the nal purication of the methyl esters. It should be noted, however, that most of the hypotheses put forward attempting to describe the phase behavior of transesterication reactions have been qualitative in nature. Unfortunately, the terms mass transfer control and reaction control have never been clearly dened for transesterication reactions, and thus much ambiguity remains with regard to the rate-limiting step. The hypothesis that suggests that the overall rate transitions from a mass transfer controlled regime to a reaction controlled regime is based on incomplete and insucient physical evidence. The majority of the studies describing the phase behavior of these reactions assume the transition occurs simply because the two reactant phases are highly immiscible and an initial lag in the methyl ester production is often observed. High immiscibility and a lagging reaction rate do not necessarily imply mass transfer control, as will be shown in this study. For example, a sigmoidal methyl ester production curve, i.e., one that displays an initially slow rate followed by a fast acceleration, can be obtained for the transesterication reaction between the highly immiscible components soybean oil and methanol while the rate is exclusively under reaction control, even with minimal mixing. Not surprisingly, there have also been many dierent kinetic studies reported for transesterication reactions.4,9,1416 The most common kinetic system found in the literature appears to
Received: Revised: Accepted: Published:
6340

January 12, 2012 March 8, 2012 April 14, 2012 April 14, 2012
dx.doi.org/10.1021/ie300116p | Ind. Eng. Chem. Res. 2012, 51, 63406349

Industrial & Engineering Chemistry Research be that presented by Freedman et al.14 This kinetic model is based on three consecutive, reversible, second-order reactions. A triglyceride rst reacts with methanol to form a diglyceride and a methyl ester; the diglyceride reacts with more methanol to form a monoglyceride and a second methyl ester; nally, the monoglyceride reacts with more methanol to form the byproduct glycerol and a third methyl ester. Overall, 3 mol of methyl esters are produced for every mole of triglyceride consumed. Freedman et al. also proposed a fourth possible reaction termed a shunt mechanism, in which three molecules of methanol react simultaneously with a triglyceride to form three molecules of methyl esters without forming the intermediate species. Noureddini and Zhu reported that the inclusion of the shunt mechanism in the transesteri cation kinetics is unnecessary, however.4 In contrast, Darnoko and Cheryan reported kinetics that were second order with respect to the tri-, di-, and monoglyceride species.9 A variety of other kinetic studies have been reported that consider the eect of mass transfer limitations.5,15 Other studies even consider the eect of the formation of the methoxide ion in the reaction between methanol and the catalyst species.16 Based on the complicated biphasic nature of the transesterication reaction systems, as well as the ambiguity surrounding the rate-limiting step, it is not surprising that so many diering kinetic studies exist. Unfortunately, there still exists much uncertainty when it comes to the physical understanding of these systems. The objective of this study, therefore, is to develop simple and fundamental experiments that provide a better understanding of the complicated nature of the two-phase system and its eects on the transesterication reaction kinetics. A signicant portion of this study will be spent on properly dening mass transfer control and reaction control for transesteri cation reaction systems and to demonstrate what the denitions physically signify within the system. An experimental setup was designed specically for obtaining the necessary data required to show what it means to be under mass transfer control or reaction control and how these controlling mechanisms aect the overall reaction rate.

Article

Figure 1. Experimental setup for transesterication reactions.

MATERIALS AND METHODS Materials. Rened soybean oil, methanol, and HPLC-grade 2-propanol were purchased from Sigma-Aldrich (St. Louis, MO). Potassium hydroxide (KOH) pellets (88%) were obtained from local chemical storage. Experimental Setup. All transesterication experiments in this study were carried out in a thin graduated cylinder maintained at 60 C as shown in Figure 1. This experimental setup was carefully designed so that the phases would exist as distinct layers, with the methanol phase resting on top of the triglyceride-rich oil phase. The mixing that was selected was unique in that each phase was mixed independently so that the contents of each phase could be assumed homogeneous and yet the phases would still remain as distinct layers. This setup allowed for the interfacial area between the two phases to be held constant; it is simply the cross-sectional area of the cylinder. It should be stressed that the experimental design was constructed to gain a fundamental understanding of transesterication systems, not for a commercialized process. HPLC Instrumentation. HPLC analyses were performed with an Agilent Technologies (Santa Clara, CA) 1200 Series HPLC equipped with a vacuum degasser, quaternary pump, autosampler, temperature controlled column compartment, and UV detector.17 The analysis of the transesterication reaction
6341

samples was carried out using isocratic elution with a mobile phase consisting of 50% methanol + 50% 2-propanol. Separation was achieved using a ZORBAX Eclipse XDB-C18 column (150 4.6 mm, 5 m i.d.). The column temperature was maintained at 35 C, and detection was carried out using a wavelength of 205 nm. Determination of K+ Partition Coecient. The presence of the K+ ion was determined by use of a Vernier (Beaverton, OR) conductivity probe (Model CON-BTA). To calibrate the probe, solutions of 0.125, 0.25, 0.5, and 1.0 M KOH in methanol were prepared. The samples were diluted by a factor of 10 prior to measurement due to the range of the conductivity probe (020 000 S/cm). The transesterication experiments were conducted in the experimental setup shown in Figure 1. First, a known quantity of soybean oil was preheated to the reaction temperature. Potassium hydroxide (1 wt % of the soybean oil) was dissolved in the appropriate amount of methanol, which depended on the molar ratio. Methanol-to-oil molar ratios of 6:1, 10:1, 20:1, and 40:1 were studied; the methanol/KOH mixture was also preheated to the reaction temperature. Once at the desired temperature, the methanol was added to the soybean oil. The contents were brought to complete mixing for 30 s, and then the phases were allowed to separate for 10 min. After 10 min, a sample was withdrawn from each phase for analysis. The analysis of the methanol phase proceeded as follows: 0.5 mL was diluted in 4.50 mL of fresh methanol and the conductivity of this solution was measured with the conductivity probe. The analysis of the soybean oil phase proceeded as follows: rst, a 1 mL sample of the phase was washed with 5 mL of fresh methanol to leach out any potassium ions dissolved in the oil phase. The mixture was centrifuged for 10 min at 10 000 rpm using a Beckman-Coulter Avanti J-E centrifuge (Brea, CA). The conductivity of the methanol wash was then measured. Any conductivity reading in the resulting methanol wash was attributed to potassium ions leached from the oil phase. A second wash of the oil phase was conducted to ensure all ions had been leached out in the rst wash (it was observed that one wash was sucient to leach out all ions). The
dx.doi.org/10.1021/ie300116p | Ind. Eng. Chem. Res. 2012, 51, 63406349

Industrial & Engineering Chemistry Research conductivity of the oil phase was measured indirectly this way because of the nonconducting properties of the oil. Even with some ions present, direct measurement of the oil phase provides zero conductivity reading. Determination of Mass Transfer Coecient. To determine the mass transfer coecient, kL, of soybean oil triglycerides into methanol, the experimental setup shown in Figure 1 was used. As mentioned, the surface area, a, between the two phases is constant and known, thus allowing the mass transfer coecient to be calculated. Various agitation speeds were studied; however, the maximum speed was limited by the phase boundary: the agitation speed was increased only up to the point at which the phases would remain as two distinct layers with a constant interfacial area. First, a known quantity of soybean oil and methanol were added to the cylinder at room temperature. A methanol-tosoybean oil molar ratio of 40:1 was selected for ease of experimentation. They were brought to 60 C by immersing the cylinder in a constant-temperature water bath. During this preheating period, in which there was no mixing, it was inevitable that a small quantity of soybean oil triglycerides would diuse into the methanol phase. A sample of the methanol phase was therefore withdrawn after the desired temperature was reached; the concentration of triglycerides in this sample was taken as the initial data point, cTG,0. The convective mass transfer equation, as dened by Seader and Henley, can be written according to
dNTG(t ) * c TG(t )) = kLa(c TG dt

Article

Table 1. Soybean Oil and Methanol LiquidLiquid Binary Phase Mixtures: Methanol Phase Compositions at Increasing Temperatures
temp (C) 20 40 60 methanol (g/g) 0.996 0.992 0.988 triglycerides (g/g) 0.004 0.008 0.012 triglycerides (mol/m3) 3.66 6.80 10.03

where NTG is the total moles of triglycerides, cTG * is the concentration of triglycerides at the soybean oil/methanol interface, and cTG(t) is the concentration of triglycerides in the bulk of the methanol phase at a given time t.18 It was experimentally determined that the equilibrium concentration of soybean oil triglycerides in methanol at T = 60 C, cTG,eq, is 10.03 mol/m3. A binary phase diagram of soybean oil and methanol can be found in Figure 2, and the methanol phase compositions can be found in Table 1. It is assumed that the concentration of triglycerides at the interface, cTG * , comes to equilibrium with the methanol rapidly; therefore, the value of c* TG will be set equal to cTG,eq. A plot of ln((c* TG cTG(t))/(c* TG cTG,0)) versus t should yield a straight line with a slope =

kLa/VM, where VM is the methanol volume. At any given time t, the concentration of triglycerides in methanol, cTG(t), can be measured by the HPLC. Determination of Rate-Limiting Step. As mentioned, due to the high immiscibility of soybean oil and methanol, it is often hypothesized that the reaction is initially mass transfer controlled.37 There is, however, still much ambiguity with regard to the rate-limiting step in these reactions. Therefore, experiments were again carried out in the experimental setup shown in Figure 1 to determine whether the rate of reaction is mass transfer or reaction controlled. The most important variables that aect the rate-limiting step are the catalyst concentration and mixing rate. The e ect of catalyst concentration at constant agitation speed will be the predominant focus of this paper. The eect of agitation speed will be briey discussed at the end of the study. For any given experiment, a known quantity of soybean oil and an appropriate amount of methanol with dissolved KOH were rst preheated separately to 60 C. Once at the desired temperature, the methanol/KOH mixture was added to the soybean oil. Unless otherwise stated, the agitation speed was selected so that the initial conditions of each experiment were such that each reactant phase existed as a distinct liquid layer with a constant initial interfacial area between the phases.

RESULTS K Partition Coecient. The calibration of the conductivity probe with various KOH/methanol solutions was rst carried out. Additionally, since glycerol forms as a byproduct of the transesterication reaction and remains predominantly in the methanol phase due to strong hydrophilic interactions, it must be determined if the presence of glycerol has any eect on the conductivity reading. Methanol solutions containing equal amounts of potassium ions with various amounts of glycerol were prepared, and the conductivity of these solutions was measured. It was observed that the conductivity measurements were linear within the potassium ion concentration range studied. It was also determined that glycerol has a negligible eect of the conductivity measurement of the potassium ions in methanol. Next, the partitioning of the potassium ion between the oil and methanol phases during the transesterication reactions was determined. The partition coecient of the potassium ions is dened as the ratio of the concentration of potassium ions in the methanol phase to the concentration of ions in the soybean oil phase: c K+,MeOH K K+ = c K+,SBO
+

Figure 2. Binary phase diagram of soybean oil and methanol.

6342

As mentioned, four dierent methanol-to-soybean oil molar ratios were tested: 6:1, 10:1, 20:1, and 40:1. The conductivity measurements of the methanol phase, the soybean oil phase, and the resulting partition coecients for each molar ratio studied can be found in Table 2. From the data it is evident that
dx.doi.org/10.1021/ie300116p | Ind. Eng. Chem. Res. 2012, 51, 63406349

Industrial & Engineering Chemistry Research Table 2. Partition Coecient of Potassium Catalyst during Transesterication Reactionsa
cK+ (mol/m3) MeOH:SBO molar ratio 6:1 10:1 20:1 40:1
a

Article

MeOH phase 928 466 221 110

SBO Phase 31.27 11.56 0.800 0.145

KK+ 29.67 40.29 275.64 755.00

At 540 rpm; mixing time = 30 s; settling time = 10 min; T = 60 C; KOH = 1.0 wt % of SBO.

the potassium ions are predominantly found in the methanol phase. This is not surprising considering the polar nature of methanol and glycerol. This is consistent with previous studies on catalyst distributions.19,20 From this point on, the methanol phase will be taken as the primary reaction volume since the catalyst remains predominantly in this phase. Therefore, in order for the transesterication reaction to proceed, triglyceride molecules must be transferred to this phase. To better understand this part of the reaction, simple experiments were designed to investigate the mass transfer between soybean oil and methanol. Determination of Mass Transfer Coecient. As mentioned, experiments measuring the mass transfer coecient of soybean oil triglycerides into methanol were carried out using the experimental setup shown in Figure 1. For these experiments, all agitation speeds employed were low enough that the system consisted of two distinct liquid phases with the methanol phase sitting on top of the soybean oil triglyceride phase. As mentioned, the maximum agitation speed studied was limited to the point just before the phases began to mix and the surface area was no longer equal to the cross-sectional area of the cylinder. The concentration of triglycerides in methanol as a function of time and agitation speed can be found in Figure 3.

Figure 4. Plot of ln((cTG * cTG(t))/(cTG * cTG,0)) vs t to determine kLa/VM for triglycerides into methanol as a function of agitation speed (MeOH:SBO = 40:1; T = 60 C).

in Table 3. As expected, the values of the mass transfer coecient increase with increasing agitation. Table 3. Mass Transfer Coecient, kL, of Soybean Oil Triglycerides into Methanol at Various Mixing Speeds (T = 60 C)
rpm 100 140 180 kL (m/s) 1.46 105 3.87 105 4.86 105

Figure 3. Concentration of triglycerides in methanol as a function of time at dierent mixing speeds (MeOH:SBO = 40:1; T = 60 C).

As mentioned, a plot of ln((cTG * cTG(t))/(cTG * cTG,0)) versus t should yield a straight line with slope = kLa/VM. These plots can be found in Figure 4. Using the data in Figure 4, along with the known values of a and VM, the mass transfer coecient, kL, for the transfer of soybean oil triglycerides into methanol was calculated for each agitation speed studied; the values are given
6343

It is important to note that these mass transfer studies were conducted between essentially pure triglycerides and methanol. During transesterication reactions, however, biodiesel methyl esters are continually accumulating in the system. It has been reported that methyl esters can enhance the solubility and mixing of the phases, but very little physical data exist. Therefore, the mass transfer of triglycerides into methanol at increasing methyl ester compositions was studied. It was rst observed that the presence of methyl esters increased the equilibrium concentration of triglycerides in methanol. Next, experiments were conducted to measure the mass transfer cocient in the presence of methyl esters. It was observed that fatty acid methyl esters increase the overall rate of mass transfer, as can be seen from Figure 5. Values for the equilibrium concentration of triglycerides in methanol, along with values of the mass transfer coecient, both as functions of increasing methyl ester concentration, can be found in Table 4. From these data, it is evident that the presence of methyl esters increases both the solubility of triglycerides in methanol and the mass transfer coecient. This supports the claim that methyl esters act to enhance the mixing between the two reactant phases. From these experiments it was determined that the maximum agitation speed employable in the experimental setup shown in Figure 1 while still maintaining a constant interfacial surface area between pure methanol and pure soybean oil triglycerides was 180 rpm. Unless otherwise stated, 180 rpm will therefore be employed for the remainder of the experiments in this study. This ensures that the initial conditions of each experiment will be such that there exist
dx.doi.org/10.1021/ie300116p | Ind. Eng. Chem. Res. 2012, 51, 63406349

Industrial & Engineering Chemistry Research

dc TG,M(t ) dc TG,M(t ) < dt dt MT R

Article

(3)

Figure 5. Plot of ln((c* TG cTG(t))/(c* TG cTG,0)) vs t to determine kLa/VM for triglycerides into methanol as a function of fatty acid methyl ester concentration (MeOH:SBO = 40:1; T = 60 C; 180 rpm).

To experimentally compare the rates of mass transfer and reaction, the primary reaction volume, i.e., the methanol phase, was sampled and its contents were analyzed. Mass transfer control is signied by the absence of soybean oil triglycerides in the methanol phase. When this result is observed, it is because triglycerides react to form biodiesel methyl esters as soon as they enter the methanol phase; thus only methyl esters are detected. The term mass transfer control is thus used to signify that the overall rate of methyl ester production is physically limited by the rate of triglyceride mass transfer into the methanol phase. On the other hand, the transesterication reaction rate is dened as being under reaction control when the rate of mass transfer of triglycerides into methanol is greater than the reaction rate of triglycerides to form methyl esters:
dc TG,M(t ) dc TG,M(t ) > dt dt R MT

Table 4. Equilibrium Triglyceride Concentration in Methanol, c* TG, and Mass Transfer Coecient of Triglycerides into Methanol, kL, at Increasing Methyl Ester Concentrationsa
FAME (wt % MeOH) 0.0 5.0 10.0
a

(4)

cTG * (mol/m3) 10.03 12.43 14.04

kL (m/s) 4.86 105 9.66 105 1.44 104

MeOH:SBO = 40:1; 180 rpm; T = 60C.

two distinct reactant phases with a constant interfacial surface area between them. Mass Transfer Control vs Reaction Control. Now that information regarding the mass transfer of triglycerides into the reaction volume has been obtained, the rate-limiting step can be investigated more thoroughly. Before the results are presented, it is prudent to dene what is means for the transesterication reaction rate to be mass transfer controlled and reaction controlled. First, the rates of the triglyceride mass transfer (MT) into the methanol phase and the triglyceride reaction (R) inside the methanol phase will be dened as
dc TG,M(t ) dt k a * c TG,M(t )) = L (c TG,M VM

MT

(1)

dc TG,M(t ) = kc TG,M(t ) dt
R

(2)

Again, cTG,M refers to the triglyceride concentration in the methanol phase; this is not necessarily the same value as the total concentrations of triglycerides in the system. Dierentiating between the rates of triglyceride mass transfer into the methanol phase and the triglyceride reaction to form methyl esters facilitates dening mass transfer control and reaction control as well as the ability to identify which one limits the rate at any given time. The transesterication reaction rate is dened as being under mass transfer control when the mass transfer rate into the methanol phase is less than the reaction rate of triglycerides to form methyl esters; thus
6344

When this occurs, triglycerides will be detected in the methanol phase because the rate of mass transfer is now greater than the rate of reaction. This allows triglycerides to accumulate in the reaction volume. Thus, the term reaction control is used to signify that the overall rate of methyl ester production is physically limited by the instrinsic reaction rate between the two reactant species. This does not mean that the rate of mass transfer has no eect on the overall rate; rather, it emphasizes the fact that the instrinsic reaction rate is not great enough to consume all of the triglycerides as they enter the reaction volume; thus an accumulation of reactants in the reaction volume occurs. To experimentally investigate the rate-limiting step, the transesterication reaction was again carried out in the experimental setup shown in Figure 1. A series of experiments were carried out in which the catalyst concentration in the methanol phase was at 1.0, 0.5, 0.25, 0.125, 0.0625, and 0.03125 wt % of the soybean oil. These weight percents correspond to methanol phase molar concentrations of 96, 48, 24, 12, 6, and 3 mM, respectively. The goal of the experiment was to keep reducing the catalyst concentration until triglyceride species were detected in the methanol phase. The catalyst concentration at which the triglyceride species are rst detected is the point at which the rate-limiting step transitions from mass transfer control to reaction control. It should be noted that when a sample was withdrawn from the reaction volume it was immediately neutralized with 1.0 M HCl to stop the reaction. Thus, any triglycerides present in the methanol phase will not continue to react once withdrawn from the reaction vessel. The chromatograms shown in Figure 6 visually illustrate the transition of the rate-limiting step. It can be seen that when the catalyst concentration in the methanol phase is 1.0 wt % of the soybean oil at a 40:1 methanol-to-oil molar ratio, no triglycerides are detected in the methanol phase, only methyl esters. This clearly indicates that at this catalyst concentration the overall rate is under mass transfer control. As the catalyst concentration is reduced down to 0.0625 wt % of the soybean oil, triglycerides are clearly present in the methanol phase. This indicates that the mass transfer rate is now greater than the reaction rate, thus allowing triglycerides to accumulate in the
dx.doi.org/10.1021/ie300116p | Ind. Eng. Chem. Res. 2012, 51, 63406349

Industrial & Engineering Chemistry Research

Article

Figure 6. Transition of the transesterication reaction rate from mass transfer control to reaction control as a function of decreasing catalyst concentration, cKOH (MeOH:SBO = 40:1; T = 60 C; tR = 1 min; 180 rpm; TG = triglycerides; DG = diglycerides; MG = monoglycerides; FAME = fatty acid methyl esters).

methanol phase. Under these conditions the overall rate is clearly under reaction control. Mass Transfer Controlled Regime. Now that the terms mass transfer control and reaction control for transesterication reactions have been clearly dened, along with the development of a simple experimental method for determining which mechanism is limiting, the transesterication reaction rates under both mass transfer control and reaction control can be analyzed and compared. To reiterate, the agitation speed employed here was 180 rpm, ensuring that the initial conditions are such that two distinct liquid phases with a constant interfacial area are present. First, the reaction rate under mass transfer control will be analyzed. Since no triglycerides are detected in the methanol phase under mass transfer control, data from two separate experiments will be compared, as shown in Figure 7. The triglyceride data points were obtained when no catalyst was present in the methanol phase; thus the slope of this line is the rate of the triglyceride mass transfer (these data are also present in Figure 3, 180 rpm). The methyl ester data points were obtained when potassium hydroxide at a concentration of 1.0 wt % of the soybean oil was present in the methanol phase. From Figure 6, it is clear that at this catalyst concentration the reaction is under mass transfer control and thus no triglycerides will be detected in the methanol phase. The slope of the line through the methyl ester data points is thus the initial rate of methyl ester production when the reaction is under mass transfer control. As can be seen, the observed rate of methyl ester production is
6345

Figure 7. Comparison of initial rate of mass transfer of triglycerides into methanol and initial rate of methyl ester production (MeOH:SBO = 40:1; T = 60 C; 180 rpm).

approximately 3 times the rate of the triglyceride mass transfer (slopes of 1.32 and 0.41 mol m3 min1), as is predicted by the reaction stoichiometry. The discrepancy between the absolute values of the triglyceride and methyl ester concentrations is due to the initial mixing that occurs when methanol with dissolved catalyst is added to soybean oil.
dx.doi.org/10.1021/ie300116p | Ind. Eng. Chem. Res. 2012, 51, 63406349

Industrial & Engineering Chemistry Research As the reaction progresses under mass transfer control, the observed rate of methyl ester production in the methanol phase begins to accelerate, as shown in Figure 8. The rst change in

Article

Figure 8. Concentration of triglycerides and methyl esters in the methanol phase during the transesterication reaction under mass transfer control (MeOH:SBO = 40:1; T = 60 C; KOH = 1.0 wt % of SBO; 180 rpm; , transition from mass transfer control to reaction control; |, transition from two-phase system to single-phase system).

reached when the two-phase system passes through the emulsion stage and completes a phase transition to a singlephase system. Thus, as mentioned, the mass transfer rate still has an aect on the overall rate under reaction control; as the mass transfer increases, so too does the overall rate. This is because the methyl ester production rate is directly proportional to the concentration of triglycerides in the methanol phase. The reason for the use of the term reaction control, however, has to do with the fact that the mass transfer rate is now greater than the rate of reaction; i.e., the mass transfer rate no longer limits the overall rate. Rather, the overall rate is limited by the instrinsic reaction rate between the two reactant species, as is evident by the fact that triglyceride species begin to accumulate in the reaction volume after t = 105 min. Reaction Controlled Regime. Now that it has been veried that when the transesterication reaction rate is initially under mass transfer control it transitions to reaction control, it can be compared to the behavior of the rate exclusively under reaction control. When a catalyst concentration is selected so that the rate starts under reaction control (see Figure 6), the initial concentration of triglycerides in the methanol phase is no longer zero. Rather, the triglycerides begin to accumulate in the methanol phase during the initial stages of the reaction, as shown in Figure 9 when a catalyst concentration of 0.0625 wt %

the reaction rate, as signied by an increase in the slope of the methyl ester production rate, occurs at 105 min. At this time triglycerides become detected in the methanol phase due to the increased rate of mass transfer, which has been shown to increase with increasing methyl ester concentration. Thus it can be said that under the present conditions the overall rate has transitioned from mass transfer control to reaction control. Interestingly, however, at t = 105 min the system still remains as two distinct layers, i.e., phases. However, as the time increases beyond t = 105 min, small droplets of the triglyceride phase begin to form in the methanol phase and the reaction volume begins to appear as an emulsion with an ever-increasing quantity of triglycerides. It should be reinforced that the agitation speed was maintained constant throughout; the increased mixing between the phases results purely from chemical changes in the system. It should be stressed that the fact that the rate-limiting step has transitioned from mass transfer control to reaction control does not mean that mass transfer no longer has an eect on the overall rate. Rather, it means that the overall rate is now limited by the instrinsic reaction rate instead of the mass transfer rate. No longer is it possible for the catalyst to consume all of the triglycerides as they enter the reaction volume, and thus there is an accumulation of reactants in the reaction volume. This trend can be observed by the data in Figure 8 after t = 105 min. After this time, the overall rate can be seen to be slowly accelerating until nally, after an additional 45 min under reaction control, the rate reaches its maximum value at t = 150 min. It is clearly evident that the beginning of the acceleration in the overall rate directly corresponds to the time when the triglyceride concentration in the methanol phase begins to increase, i.e., after t = 105 min, and that the time of the maximum rate corresponds to the time when the triglyceride concentration in the methanol is at a maximum, i.e., at t = 150 min. As mentioned, at times beyond t = 105 min the reaction volume begins to appear as an emulsion with an ever-increasing quantity of triglycerides in the form of small droplets surrounded by the bulk methanol. The maximum rate is
6346

Figure 9. Concentration of triglycerides and methyl esters in the methanol phase during the transesterication reaction under reaction control (MeOH:SBO = 40:1, T = 60 C; KOH = 0.0625 wt % of SBO; 180 rpm; |, transition from two-phase system to single phase system).

of SBO is selected. The triglyceride concentration in the methanol phase is dependent on the competing rates of mass transfer into the methanol phase and the reaction to form methyl esters. If the catalyst concentration were to be increased, the triglyceride concentration in methanol would eventually reach zero and the reaction rate would be back under mass transfer control. When the reaction-controlled rate is allowed to run to completion, a similar acceleration in the overall rate of methyl ester production is observed, as also shown in Figure 9. The fact that the system has been under reaction control the entire time and still displays a sigmoidal methyl ester production curve proves that the accelerating rate is not necessarily caused by a transition from mass transfer control to reaction control. The maximum acceleration in the methyl ester rate occurs again as the system transitions from a two-phase system to a single-phase system, in this case at t = 270 min. As previously mentioned, when this phase transition occurs, the triglyceride
dx.doi.org/10.1021/ie300116p | Ind. Eng. Chem. Res. 2012, 51, 63406349

Industrial & Engineering Chemistry Research concentration in the methanol phase surges and the reaction rate increases proportionally. As emphasized for the data in Figure 8, it should be stressed that the reason that the data in Figure 9 are dened as reaction control is because the intrinsic reaction rate is limiting the overall rate of methyl ester production. Clearly, the rate of mass transfer aects the overall rate throughout the entirety of the reaction in Figure 9; however, at no point is the rate of mass transfer limiting the overall rate. There is always an accumulation of triglycerides in the reaction volume, and thus the term reaction control is applied throughout. Aside from starting under reaction control, the only other dierence between the data in Figures 8 and 9 is the time at which the systems transition from two phases to a single phase, which is due to the di erent catalyst concentrations used. Lower catalyst concentrations will take a greater amount of time to produce a sucient quantity of methyl esters to cause the phase transition. Eect of Mixing on the Rate-Limiting Step. It has now been clearly established that at constant agitation speed the catalyst concentration has a direct eect on the rate-limiting step. High catalyst concentrations result in mass transfer control, while low catalyst concentrations result in reaction control. The catalyst concentration, however, is not the only variable that aects the rate-limiting step. Whereas the catalyst concentration directly aects the reaction rate, the agitation speed directly aects the mass transfer rate; this will have some eect on the rate-limitng step. The rate of mass transfer is proportional to the surface area between the two phases. As agitation is increased, the surface area between the two phases increases. This results in an increase in the mass transfer rate of triglycerides into methanol. As a result, the overall rate is forced into the direction of reaction control. This is shown in Figure 10, where it can be seen that, at any given catalyst

Article

denitions of mass transfer control and reaction control, the rate is under reaction control. As was seen in Figures 8 and 9, high concentrations of triglycerides in the methanol phase result in a correspondingly high reaction rate. Phase Behavior Eect on the Kinetics of Transesterication Reactions. The fact that the transesterication reaction accelerates under both mass transfer control and reaction control is intriguing for a variety of reasons. Most importantly, it suggests that the acceleration of the methyl ester production is not necessarily caused by a change in the ratelimiting step, which is often claimed in the literature. Rather, the acceleration is caused by a phase transition from two phases to a single phase. If the concentrations of the intermediate species are assumed to remain relatively low throughout the reaction, the methyl ester rate can be approximated according to
dc FAME,M dc TG,M 3 dt dt

(5)

The rate of change of the triglyceride concentration within the methanol phase with time, dcTG,M/dt, has already been dened for the competing cases of mass transfer control and reaction control, as found in eqs 1 and 2, respectively. Therefore, the methyl ester production rates under both mass transfer and reaction control can be approximated accordingly:
kLa dc FAME,M * c TG,M) (c TG,M = 3 dt MT VM
dc FAME,M = 3kc TG,M dt R

(6)

(7)

Figure 10. Eect of agitation speed on the initial concentration of triglycerides present in the methanol phase (MeOH:SBO = 40:1; T = 60 C; KOH = 1.0 wt % of SBO; tR = 1 min).

concentration, the amount of triglycerides in the methanol phase increases with increasing agitation. At agitation speeds that result in complete emulsication, as seen by the 540 rpm data points in Figure 10, it is very clear that oil triglycerides are present in the methanol phase right from the very beginning of the reaction, and thus, based on the
6347

Again, cFAME,M and cTG,M refer to the concentrations in the methanol phase. Equation 7 can only be used to approximate the reaction rate constant, k, when the system is under reaction control. As has been shown, reaction control may exist for the entire length of the reaction or only a portion of the reaction; it depends on the conditions selected. The reaction rate constant was determined using eq 7 for the data collected during the reaction controlled regime of Figure 8 (i.e., after t = 105 min); a value of k = 1.43 min1 was obtained. Since the mass transfer coecient, kL, has previously been determined, eqs 6 and 7 can then be used to predict the methyl ester production curve as the rate-limiting step changes from mass transfer control to reaction control. The comparison between the experimental and theoretical methyl ester production rates when catalyst concentration is at 1.0 wt % of SBO can be seen in Figure 11. The reaction rate constant for the data in Figure 9, in which the catalyst concentration is 0.0625 wt % of SBO, was similarly determined using eq 7; a value of k = 0.57 min1 was obtained. At this catalyst concentration, the entire data set was under reaction control. The comparison between the experimental and theoretical methyl ester rates for this data set is shown in Figure 12. Since the maximum reaction rate, regardless of the initial rate-limiting step, occurs as the system transitions from two phases to a single phase, it is important to analyze this phase transition. In doing so, the solubility parameter of the methanol phase will be examined. By denition, the solubility parameter of pure component i, i, is the square root of the cohesive energy density of pure component i in the pure state.21
dx.doi.org/10.1021/ie300116p | Ind. Eng. Chem. Res. 2012, 51, 63406349

Industrial & Engineering Chemistry Research Table 5. Solubility Parameters for Components in Transesterication Reactions
component glycerol methanol methyl oleate soybean oil i ((cal/cm3)1/2) 16.822 14.422 8.322 7.923

Article

j = 1 jVjxj j = 1 Vjxj
R

Figure 11. Comparison of experimental and theoretical methyl ester concentrations when the transesterication reaction starts in mass transfer control and transitions to reaction control (MeOH:SBO = 40:1; T = 60 C; KOH = 1.0 wt % of SBO; 180 rpm; , transition from mass transfer control to reaction control).

where j is the solubility parameter of pure component j, Vj is the molar volume of component j, and xj is the mole fraction of component j. As the average solubility parameter of the methanol phase falls, its value moves closer to that of soybean oil. Additionally, there is likely some methanol and methyl ester species that transfer from the methanol phase into the soybean oil phase; this would raise the volume average solubility parameter of this phase. Eventually the solubility parameters of the two phases become close enough that the system transitions from two phases to a single phase.

Figure 12. Comparison of experimental and theoretical methyl ester concentrations when the transesterication reaction is under reaction control the entire time (MeOH:SBO = 40:1; T = 60 C; KOH = 0.0625 wt % of SBO; 180 rpm).

(Ev)i 1/2 i = Vi
The cohesive energy density is a measure of the strength of intermolecular forces holding molecules together in the liquid state per unit volume of liquid. It is dened as the ratio of the latent energy of vaporization of a pure component, (Ev)i, to the molal volume of that component, Vi. According to King, a general rule is that components or phases with solubility parameters diering by greater than 3 (cal/cm3)1/2 will form two separate liquid phases. As can be seen from Table 5, the solubility parameters of the initial reactant phases dier by 6.5 (cal/cm3)1/2 and thus two liquid phases are observed. This behavior is similarly observed when analyzing biodiesel, glycerol, and methanol ternary phase diagrams.24 As the transesterication reaction proceeds, however, triglycerides and especially methyl esters continue to accumulate in the methanol phase. This has the eect of lowering the volume average solubility parameter, , of the methanol phase, given by
6348

CONCLUSION The phase behavior eect on the kinetics of transesterication reactions for biodiesel production has been thoroughly investigated. First, the partition coecient of the potassiumbased catalyst was determined. These results clearly indicate that the majority of the catalyst resides in the methanol phase. Thus, the methanol phase was established as the primary reaction volume for transesterication reactions with homogeneous potassium based catalysts. Once the methanol phase was taken as the primary reaction volume, it was established that triglycerides must transfer into this phase to react. It was shown that both the equilibrium concentration of soybean oil triglycerides in methanol and the rate at which the triglycerides were transferred into methanol increase with increasing methyl ester concentration. Next, mass transfer control and reaction control were clearly dened for transesterication reactions. The reaction was dened as being under mass transfer control when the rate of mass transfer of triglycerides into the methanol phase was less than the rate of reaction of triglycerides to form methyl esters; i.e., the mass transfer rate limits the overall rate. On the other hand, the reaction was dened as being under reaction control when the rate of mass transfer of triglycerides into the methanol phase is greater than the rate of reaction of triglycerides to form methyl esters; i.e., the intrinsic reaction rate limits the overall rate. Once these terms were properly dened, the overall transesterication rate under mass transfer and reaction control were compared. This comparison revealed the overall rate undergoes an acceleration regardless of the rate-limiting step. This proves that the acceleration is not caused by a change in the rate-limiting step. Rather, the acceleration is caused by a phase transition from two phases to a single phase. When this occurs, the triglyceride concentration in the methanol phase surges, which directly increases the rate of methyl ester production. It is believed that the transition from two phases to a single phase occurs when the volume-average solubility parameter of the methanol phase is reduced by the increasing concentration of methyl esters.
dx.doi.org/10.1021/ie300116p | Ind. Eng. Chem. Res. 2012, 51, 63406349

Industrial & Engineering Chemistry Research

Article

AUTHOR INFORMATION

Corresponding Author

*Tel.: (610) 758-4257. Fax: (610) 758-5057. E-mail: jth0@ lehigh.edu.

Notes

The authors declare no competing nancial interest.

ACKNOWLEDGMENTS This project has been partially nanced by a grant from the Commonwealth of Pennsylvania Department of Community and Economic Development through the Pennsylvania Infrastructure Technology Alliance (PITA). REFERENCES

(19) Zhou, W.; Boocock, D. G. B. Phase Distributions of Alcohol, Glycerol, and Catalyst in the Transesterication of Soybean Oil. J. Am. Oil Chem. Soc. 2006, 83, 1047. (20) Chiu, C. W.; Go, M. J.; Suppes, G. J. Distribution of Methanol and Catalysts Between Biodiesel and Glycerin Phases. AIChE J. 2005, 51, 1274. (21) King, C. J. Separation Processes, 2nd ed.; McGraw-Hill: New York, 1980. (22) Yaws, C. L. Chemical Properties Handbook; McGraw-Hill: New York, 1999. (23) King, J. W. Determination of the Solubility Parameter of Soybean Oil by Inverse Gas Chromatography. Lebensm.-Wiss. Technol. 1995, 28, 190. (24) Csernica, S. N.; Hsu, J. T. Reverse-Phase Ternary Phase Diagram, Tie Lines, and Plait Point for Commercial Biodiesel GlycerolMethanol. Ind. Eng. Chem. Res. 2011, 50, 1012.

(1) Suppes, G. J.; Dasari, M. A.; Doskocil, E. J.; Mankidy, P. J.; Go, M. J. Transesterication of Soybean Oil with Zeolite and Metal Catalysts. Appl. Catal., A 2004, 257, 213. (2) Lopez, D. E.; Goodwin, J. G., Jr.; Bruce, D. A.; Lotero, E. Transesterication of Triacetin with Methanol on Solid Acid and Base Catalysts. Appl. Catal., A 2005, 295, 97. (3) Lotero, E.; Liu, Y.; Lopez, D. E.; Suwannakarn, K.; Bruce, D. A.; Goodwin, J. G., Jr. Synthesis of Biodiesel via Acid Catalysis. Ind. Eng. Chem. Res. 2005, 44, 5353. (4) Noureddini, H.; Zhu, D. Kinetics of Transesterication of Soybean Oil. J. Am. Oil Chem. Soc. 1997, 74, 1457. (5) Liu, X.; Piao, X.; Wang, Y.; Zhu, S. Model Study on Transesterication of Soybean Oil to Biodiesel with Methanol Using Solid Base Catalysts. J. Phys. Chem. A 2010, 114, 3750. (6) Stamenkovic, O. S.; Lazic, M. L.; Todorovic, Z. B.; Veljkovic, V. B.; Skala, D. U. The Eect of Agitation Intensity on Alkali-Catalyzed Methanolysis of Sunower Oil. Bioresour. Technol. 2007, 98, 2688. (7) Boocock, D. G. B.; Konar, S. K.; Mao, V.; Sidi, H. Fast One-Phase Oil-Rich Process for the Preparation of Vegetable Oil Methyl Esters. Biomass Bioenergy 1996, 11, 43. (8) Lam, M. K.; Lee, K. T. Accelerating Transesterication Reaction with Biodiesel as Co-Solvent: A Case Study for Solid Acid Sulfated Tin Oxide Catalyst. Fuel 2010, 89, 3866. (9) Darnoko, D.; Cheryan, M. Kinetics of Palm Oil Transesterication in a Batch Reactor. J. Am. Oil Chem. Soc. 2000, 77, 1263. (10) Ma, F.; Clements, D.; Hanna, M. A. The Eect of Mixing on Transesterication of Beef Tallow. Bioresour. Technol. 1999, 69, 289. (11) Kuramochi, H.; Maeda, K.; Osako, M.; Nakamura, K.; Sakai, S. Superfast Transesterication of Triolein Using Dimethyl Ether and a Method for High-Yield Transesterication. Ind. Eng. Chem. Res. 2008, 47, 10076. (12) Cao, W.; Han, H.; Zhang, J. Preparation of Biodiesel from Soybean Oil Using Supercritical Methanol and Co-Solvent. Fuel 2005, 84, 347. (13) Mao, V.; Konar, S. K.; Boocock, D. B. G. The Pseudo-Single Phase, Base-Catalyzed Transmethylation of Soybean Oil. J. Am. Oil Chem. Soc. 2004, 81, 803. (14) Freedman, B.; Buttereld, R. O.; Pryde, E. H. Transesterication Kinetics of Soybean Oil. J. Am. Oil Chem. Soc. 1986, 63, 1375. (15) Huang, K.; Xu, Q.; Zhang, S.; Ren, Z.; Yan, Y. Multi-Step Controlled Kinetics of the Transesterication of Crude Soybean Oil with Methanol by Mg(OCH3)2. Chem. Eng. Technol. 2009, 32, 1595. (16) Komers, K.; Skopal, F.; Stloukal, R.; Machek, J. Kinetics and Mechanism of the KOH-Catalyzed Methanolysis of Rapeseed Oil for Biodiesel Production. Eur. J. Lipid Sci. Technol. 2002, 104, 728. (17) Csernica, S. N.; Hsu, J. T. Simple and Ecient Method for the Analysis of Transesterication Reaction Mixtures for Biodiesel Production by Reversed-Phase High Performance Liquid Chromatography. Energy Fuels 2010, 24, 6131. (18) Seader, J. D.; Henley, E. J. Separation Process Principles, 2nd ed.; John Wiley and Sons: New York, 2006.
6349
dx.doi.org/10.1021/ie300116p | Ind. Eng. Chem. Res. 2012, 51, 63406349