Lubricating and Waxy Esters II: Synthesis ... - ACS Publications

Oct 19, 2012 - Laziz Bouzidi†, Shaojun Li†, Steve Di Biase‡, Syed Q. Rizvi‡, Peter Dawson†, and Suresh S. Narine*†. † Trent Centre for B...
1 downloads 0 Views 1000KB Size
Download PDF
Article pubs.acs.org/IECR

Lubricating and Waxy Esters II: Synthesis, Crystallization, and Melt Behavior of Branched Monoesters Laziz Bouzidi,† Shaojun Li,† Steve Di Biase,‡ Syed Q. Rizvi,‡ Peter Dawson,† and Suresh S. Narine*,† †

Trent Centre for Biomaterials Research, Departments of Physics & Astronomy and Chemistry, Trent University, 1600 West Bank Drive, Peterborough, Ontario K9J 7B8, Canada ‡ Elevance Renewable Sciences, 175 E Crossroads Parkway, Suite F Bolingbrook, IL, U.S.A., 60440 S Supporting Information *

ABSTRACT: A comprehensive study of branched derivatives of four pure jojoba wax-like esters (JLEs), having 36, 40 (two isomers), and 44 carbons was conducted to elucidate their crystallization and melting behavior. Crystallization and melting characteristics depended strongly on the number of branches, molar mass, and symmetry. The derivatives demonstrated a very strong tendency to form glassy liquids rather than crystal phases and remain liquid-like at very low temperatures. As the number of branches increased, their crystallinity decreased drastically while their glassy phase increased concomitantly with the depression of the onset of crystallization and/or glass transition temperature. A variety of possible transformation paths, ranging from very little polymorphic activity to extremely polymorphic behavior, depending on number of branches, mass, and symmetry were revealed. It is shown that asymmetry plays a large role in the low temperature behavior, rendering the branched derivatives of the asymmetrical JLEs much better candidates for lubricant formulations.

1. INTRODUCTION Increasing environmental and sustainability concerns are driving research efforts to develop high-performance nontoxic biodegradable lubricants, waxes, and gels from renewable sources. Due to their structure, vegetable oils and their derivatives are suitable raw materials for the formulation of such functional materials. In particular, vegetable oil based lubricants exhibit good boundary friction lubricity and wear protection with relatively stable viscosity−temperature behavior.1 However, compared to conventional mineral oil based lubricants, most vegetable oil based lubricants exhibit poor hydrolytic stability, poor thermal and oxidation stability, poor low-temperature fluidity, and inadequate aging resistance. Furthermore, ester oils derived from vegetable oils are significantly more expensive, additionally limiting their widespread usage.2 The oil from the jojoba plant is a particularly interesting material due to its unusual composition and unique characteristics.3−8 Jojoba oil is suitable for lubrication as well as for a variety of other high-end wax and gel applications such as cosmetic and medical formulations and foods.9 Jojoba oil has very good lubricity, particularly suitable in high-grade lubricating oil formulations and additives.6,7 However, its low temperature performance, acid value, and oxidative stability are not the best.10,11 Its high cost and lack of ready availability also constitute a severe obstacle to its large scale use.12 Therefore, the production of cheap jojoba-like wax esters and their derivatives from inexpensive renewable resources which deliver optimal physical properties is necessary for commercialization to be viable.13 The production of jojoba wax ester analogues from oleic acid has been explored, and physical and chemical methods were used to mitigate the limiting factors inherent to the base esters.12,14−18 © 2012 American Chemical Society

The Trent Centre for Biomaterials Research is conducting a comprehensive fundamental study of the syntheses, crystallization, melt, polymorphism, solid fat content, viscosity, and modeling of structure and function of jojoba wax ester analogues: linear mono-, di, and triesters and their branched derivatives. Following our work on pure jojoba-like monoesters (JLEs),19 a comprehensive study of branched derivatives of the JLEs was conducted to elucidate the influence of structure: chain length, branching, symmetry, and functional groups on the physical properties. In the present work, branched derivatives of the JLEs, with 2, 3, and 4 branches of 3 carbons each, were synthesized by epoxidation of the double bonds, followed by ring-opening esterification of the epoxides with propionic acid. The compounds were characterized by 1H NMR, and their purities were measured with HPLC. The base JLEs have carbon chain lengths of 36, 40 (two isomers), and 44, and were prepared by Steglish esterification from fatty acids and fatty alcohols.19 Differential scanning calorimetry (DSC) and wide-line pulsed nuclear magnetic resonance (pNMR) were used to gain insight into the physical properties of the branched JLEs. The phase transition behavior and solid fat content (SFC) are reported as functions of carbon chain length, number of branches, and symmetry.

2. MATERIALS AND METHODS 2.1. Materials. Oleic acid (90%), erucic acid (90%), oleyl alcohol (85%), oleoyl chloride (80%), propionic acid, pyridine, chloroform, dichloromethane, N,N′-dicyclohexylcarbodiimide Received: Revised: Accepted: Published: 14892

June 21, 2012 October 16, 2012 October 19, 2012 October 19, 2012 dx.doi.org/10.1021/ie3016472 | Ind. Eng. Chem. Res. 2012, 51, 14892−14902

Industrial & Engineering Chemistry Research

Article

Scheme 1. Synthesis of Branched Compoundsa

a

n and m values are listed in Supporting Information Table S1.

Scheme 2. Epoxidation of Alkene21

(DCC), 4-dimethylaminopyridine (DMAP), and lithium aluminum hydride (AlLiH4) were purchased from SigmaAldrich. Hexane, ethyl ecetate, and terahydrofuran were purchased from ACP Chemical Int. (Montreal, QC, Canada). The structure of the JLEs previously studied and the branched derivatives prepared in this work is provided in the Supporting Information (Scheme S1). The nomenclature used for the JLEs and their branched derivatives is presented in the Supporting Information (Table S1). 2.2. Preparation of Branched Derivatives of Jojobalike Esters. The jojoba-like esters with 2, 3, and 4 branches (listed in Supporting Information Table S1) were synthesized by reactions including epoxidation and ring open esterification of the epoxides with propionic acid (Scheme 1). The double bonds on the JLEs shown in Scheme 1 were epoxidized following the method published by Moser et al.,20 using peroxyacid as a catalyst, formed in situ from formic acid and hydrogen peroxide as shown in Scheme 2. CH2Cl2 was used as the solvent so as to improve the solubility of reagents and intermediatesepoxidation reactions with CH2Cl2 were faster (∼5 h) and produced fewer side-products. Note that CH2Cl2 is considered carcinogenic in most countries and that that ethyl acetate can be used as an alternative. The compounds with 2 branches were prepared by introducing carboxylic acid (propionic acid) to the epoxides

with a ring-opening esterification without the need for either a catalyst or solvent. In this reaction, the carboxylic acid served as both reactant and solvent. A plausible mechanism for the ringopening esterification of epoxide to afford branched esters (Scheme 3) has been reported in the literature.20 On the basis of this mechanism, two potential alcohols can be formed at the site of one epoxide ring. To prepare compounds with two branches as the main products, the reactions were carried out at 95 and at 120 °C for the synthesis of 3- and/or 4-branched compounds. In the latter case, it was necessary to partially remove the water produced in the esterification reactions. Effectively, the 2-, 3-, and 4branched compounds can all be prepared by a one-pot reaction by controlling reaction time and temperature. No effort was made to distinguish the regio-chemistry (for example in 2-branched 9-alkanonate-10-hydroxy-oactadecanoate versus the equally likely alkyl 10-alkanoate-9-hydroxyoctadecanoate regio-isomer) or the stereochemistry (for 14893

dx.doi.org/10.1021/ie3016472 | Ind. Eng. Chem. Res. 2012, 51, 14892−14902

Industrial & Engineering Chemistry Research

Article

USA). The mobile phase was a mixture of 40% chloroform:60% ACN with a flow rate of 1 mL/min. Migration was studied using a 250 × 4 mm Betasil Diol-100 column (Thermo Hypersil-Keytone Inc., Bellefonte, PA, USA) in normal-phase and isocratic mode. The temperature of the column was maintained at 30 °C (Alliance Column Heater, Waters Corporation, Milford, MA, USA). ELSD nitrogen flow was set at 25 psi with nebulization and a drifting tube maintained at 12 and 60 °C, respectively. Gain was set at 500. The mobile phase was heptane:isopropyl alcohol (90:10) v run for 30 min at a flow rate of 0.5 mL/min. All solvents were HPLC grade and obtained from VWR International (Mississauga, ON, Canada). 2.3.3. Differential Scanning Calorimetry (DSC). DSC analysis was carried out on a Q200 model (TA Instruments, New Castle, DE, USA) equipped with a refrigerated cooling system (RCS 90, TA Instruments) under a nitrogen flow of 50 mL/min. Approximately 5.0−10.0 (±0.1) mg of fully melted and homogenously mixed sample was placed and hermetically sealed in an aluminum DSC pan. The samples were cooled from the melt (50 °C) to −90 °C and subsequently reheated to 70 °C at the same constant rate of 3.0 K/min to obtain the crystallization and melting profiles. “TA Universal Analysis” software together with a method developed in our group22 was used to analyze the data and extract the main characteristics of the peaks. 2.3.4. Solid Fat Content (SFC) Determination. The SFC measurements were performed on a Bruker Minispec mq 20 (Milton, Ontario, Canada) pNMR spectrometer equipped with a combined high and low temperature probe, and a temperature controller (Bruker BVT3000). Approximately 0.57 ± 0.05 mL of fully melted sample was pipetted into the bottom of an NMR tube. The sample was subjected to the same thermal protocol as in the DSC analysis (described above in section 2.3.3). The system was calibrated with highly unsaturated canola oil (with heat capacity similar to that of the saturated oils, which do not crystallize until −20 °C). The accuracy of the measurement was determined to be better than ±0.1 °C. In the low temperature region, the probe is supplied with N2 gas, obtained by evaporating liquid nitrogen. The temperature was controlled to better than ±0.1 °C. SFC data was collected using Bruker’s Minispec plus V1.1 Rev. 05 software.

Scheme 3. Mechanism for the Ring-Opening Esterification of Epoxides

example, S or R at C9 and C10 in branched JLE-36S esters) of any of the polyol esters, due to the laborious chromatography required and the economics involved at potentially larger commercial scales. 2.2.1. Preparation of the Epoxidized Jojoba-like Esters (Scheme 1). To a stirred solution of JLE (10 mmol) and formic acid (60 mmol) in 10 mL CH2Cl2 at 4 °C, H2O2 (44 mmol) was slowly added. The reaction proceeded at room temperature with vigorous stirring for 4−36 h. After removal of the aqueous phase, additional CH2Cl2 (30 mL) was added to the organic phase, which was washed sequentially with water (2 × 20 mL), saturated aqueous NaHCO3 (2 × 10 mL), and brine (2 × 20 mL), and then dried with Na2SO4. After filtration, the filtrate was concentrated and the residue was purified by column chromatography with ethyl acetate/hexane mixture. The chromatographic methods used to purify the epoxides are specified in Table S2, Supporting Information. 2.2.2. Oxirane Ring-Opening Reactions with Propionic Acid (Scheme 1). To a solution of the epoxidation products listed above (10 mmol), 220 mmol propionic acid was added. The reaction was carried out under a N2 atmosphere at a temperature of 95 °C. The reaction mixtures were stirred at that temperature for 16−24 h. To obtain compounds with 3 or 4 branches, the reaction temperature was raised to 120 °C for 24−36 h. The resulting products were poured into 200 mL of water and extracted with ethyl acetate (2 × 50 mL). The organic phase was washed sequentially by water (2 × 100 mL), saturated aqueous NaHCO3 (2 × 100 mL), and brine (2 × 200 mL), and then dried on Na2SO4. After filtration, the filtrate was concentrated and the residue was purified by column chromatography with ethyl acetate and hexane. The chromatographic methods used to purify the branched derivatives are specified in Table S3, Supporting Information. 2.3. Analytical Methods. The physical measurements were run at least in triplicate. The reported values and uncertainties attached are the mean and associated calculated standard deviations, respectively. 2.3.1. 1H-NMR. 1H NMR spectra were acquired on a Bruker Avance III 400 spectrometer (ν(1H) = 400.22 MHz; Bruker BioSpin MRI GmbH, Karlsruhe, Germany) equipped with a 5 mm Broad-Band Observe (BBO) probe. Spectra were acquired at 25 °C over a 16 ppm spectral window with a 1s recycle delay and 32 transients. NMR spectra were Fourier transformed, phase corrected, and baseline corrected. Window functions were not applied prior to Fourier transformation. Chemical shifts were referenced relative to residual chloroform d1 solvent peaks. 2.3.2. High-Performance Liquid Chromatography (HPLC). Purities of compounds were evaluated on a Waters e2695 HPLC (Waters Limited, Mississauga, Ontario) fitted with a Waters 2424 ELS Detector and an X-Bridge column (C18, 150 mm × 4.6 mm, 5.0 μm, Waters Corporation, Milford, MA,

3. RESULTS AND DISCUSSION JLE-36S and JLE-44S were prepared in the symmetrical conformation from C18:1 and C22:1 acids and C18:1 and C22:1 alcohols, respectively. The JLE-40 was prepared in both the C18:1/C22:1 acid/alcohol and C22:1/C18:1 acid/alcohol arrangements. Arbitrarily, we have labeled the C22:1/C18:1 acid/alcohol JLE (sample JLE-40S) as being “symmetrical” and the C18:1/C22:1 acid/alcohol (JLE-40A) as being “asymmetrical” (although of course both of these samples are asymmetrical) for the ease it allows in discussion, and because the thermodynamic behavior of JLE-40S falls within a linear trend with the other, symmetric JLEs studied. 3.1. Products of Epoxidation and Branched Derivatives of Jojoba-like Esters (Scheme 1). Purity of all epoxides and branched derivatives of the JLEs as determined by HPLC was higher than 99%. The pure branched derivatives of JLEs were given as colorless oils by column chromatography. Column chromatography method and 1H NMR data for epoxides of JLEs and of branched derivatives of JLEs are 14894

dx.doi.org/10.1021/ie3016472 | Ind. Eng. Chem. Res. 2012, 51, 14892−14902

Industrial & Engineering Chemistry Research

Article

Figure 1. DSC cooling (3.0 K/min) thermograms of branched derivative of (a) JLE-36S, (b) JLE-40S, (c) JLE-44S, and (d) JLE-40A. The base esters are represented in the left side panels with a right side heat flow (W/g) axis for comparison purposes.

provided in the Supporting Information listed in Tables S2 and S3, respectively. 3.2. Crystallization Behavior of the Branched Derivatives of Jojoba Wax-like Esters. The crystallization thermograms of the branched derivatives of the JLEs were all characterized by a single exotherm (Figures 1a−d). As can be seen, the heat evolved decreases dramatically with increasing branching for the same JLE, suggesting drastic decreases in crystallinity and also in the strength of the interactions between molecules in the crystal lattices. The height of the crystallization signal of the 2-branched JLEs, for example, is about 7−10 times that of the 4-branched JLEs, in all cases. The crystallinity increases with increasing mass for the same number of branches as is evident in Figure 1c to a. However, unlike JLE40S-2, JLE44S-2, and JLE40A-2 which exhibit similarly shaped strong and relatively narrow crystallization peaks, JLE36S-2 has a weak and broad DSC cooling signal reminiscent of a glass transitionindeed a calorimetric glass transition temperature (Tg) was readily measured. The Tg values measured on cooling are listed in Table 1. The onsets and peak maxima of phase change due to cooling of the 2-branched compounds occur at particularly low temperatures, ranging from −21 °C to about −65 °C for peak maxima (Tp). As more branches are added to the same JLE, the transitions become less indicative of crystallization and

Table 1. DSC Calorimetric Glass Transition Temperature (Tg) of the Different Branched Derivatives of the JLEs Measured on Cooling (3.0 K/min) Tg/°C base ester JLE-36S JLE-40S JLE-40A JLE-44S

number of side branches 2 −57.37 ± 0.19

3 −68.27 ± 0.17 −61.47 ± 0.03 −61.49 ± 0.03

4 −73.12 ± −72.81 ± −72.81 ± −70.19 ±

0.32 0.17 0.17 0.16

more suggestive of glass transitions. As mentioned above, a Tg was measured for all the branched derivatives of JLE-36S (Figure 1a). In the case of JLE-40S (Figure 1b), JLE-40S-2 records a very clear crystallization exotherm, but the transition recorded by the DSC thermogram of JLE-40S-3 begins to appear more glass-transition-like, albeit with the crystallization exotherm still remaining clearly defined (Figure 1b). This suggests some amount of phase transition that is related to crystallization and some that is related to the formation of an amorphous solid or very viscous liquid. The thermogram of JLE-40S −4 does not record a crystallization exotherm, but does allow a Tg to be measured unambiguously. In the case of JLE-44S, very clear crystallization exotherms are recorded for the 2- as well as the 3-branched derivatives, with a glass 14895

dx.doi.org/10.1021/ie3016472 | Ind. Eng. Chem. Res. 2012, 51, 14892−14902

Industrial & Engineering Chemistry Research

Article

Figure 2. DSC crystallization characteristics (cooling 3.0 K/min) of branched derivative of (a) JLE-36S, (b) JLE-40S, (c) JLE-44S, and (d) JLE-40A. Symbols represent: ○, peak; ▼, onset; ■, offset; and ★, glass temperature.

JLE-36S-2 was cooled at two lower rates (1.0 and 0.1 K/min) in order to further investigate its glass transition/crystallization behavior. Not surprisingly, the DSC signal decreased dramatically when JLE-36S-2 was cooled at 1.0 K/min and did not appear in the temperature range available for our DSC (down to −90 °C) when cooled at 0.1 K/min (not shown). This can be explained by the shift of the calorimetric glass transition as well as crystallization to lower temperatures with decreasing cooling rate. This is a further indication of the strong tendency of the materials to remain amorphous at very low temperatures, favoring formation of glassy rather than crystalline phases even when cooled very slowly. There is a marked direct and predictive relationship between onset, offset as well as peak maximum of crystallization and number of branches (Figure 2a−d). Where there is a measurable glass transition, the Tg also decreases with increasing branches (stars in Figures 2a−d). Certainly, there is an increasing benefit in terms of suppressing crystallization as additional branches are added to the molecule and as hydroxyl functionality is removed from the molecule. This obviously allows one to suppress crystallization of potential lubricants. Note that the “onset” of crystallization is a good indicator of the “pour point” of the sample; onset would occur before the pour point is recorded.23−26 The enthalpy of crystallizationthe area under the crystallization exotherm (Figure 1a−d)decreases dramatically as branches are added, suggesting clearly that as branches are added and hydroxyl groups removed the crystallinity of the sample decreases drastically. However, there is an exception to the degree of crystallinity displayed by the branched derivatives of sample JLE-44S; although the enthalpy decreases from JLE44S to JLE-44S-4, there is appreciable peak area from JLE-44S to JLE-44S-3 (Figure 1a−d). Notwithstanding, the enthalpy of crystallization of the JLE-44S-derived esters decreases linearly (R2 = 0.983).

transition event being recorded for the 4-branched derivative (Figure 1c). For the branched structures of JLE-40A (Figure 1d), clear crystallization exotherms were recorded for JLE-40A2 and JLE-40A-3, with JLE-40A-4 demonstrating a clear glass transition. Clearly, the behavior of the branched derivatives of JLE-40S and JLE-40A isomers differ significantly; this will be discussed below. JLE-40S and JLE-40A being asymmetric monoesters, it would seem that the symmetry and longer chain length of JLE-44S (a full 8 carbons longer than JLE-36S) contributes to a greater propensity for packing into crystalline units, at least with regards to the JLE-44S-2 and JLE-44S-3 samples. DSC cooling curves which present a clear Tg suggest evidence of competing transitionscrystallization vs glass transitions. In reality, the samples which still have a measurable exotherm and evidence of a Tg, probably contain both crystalline as well as glassy domains. The derivatives with a greater number of branches clearly are increasingly inhibited in terms of mass transfer, due to the number of degrees of freedom and the structural complexity of the molecule (the addition of each new branch is accompanied by the addition of an ester group in the molecule, as well as the protruding branch). This leads to shorter inelastic mean free paths in the melt and, therefore, significantly restricted participation of molecules at the growing crystal surface. Crystallization is clearly promoted by the intermolecular attractions which exist due to the hydrogen bonding developed from the hydroxyl groups in 2-branched and 3-branched derivatives. Two hydroxyl groups are present in the 2-branched derivatives and one in the 3-branched derivatives. The decrease of the crystal phase and increase of the amorphous phase in the branched derivatives can be observed in the shape of the cooling thermograms when the base JLE mass for a given number of side branches is decreased or when the number of side branches increases for a given base JLE. 14896

dx.doi.org/10.1021/ie3016472 | Ind. Eng. Chem. Res. 2012, 51, 14892−14902

Industrial & Engineering Chemistry Research

Article

suggesting a predictive relationship between the length of the base monoester chain and the degree of crystallinity. 3.2.2. Effect of Symmetry on Crystallization. The effect of having 18:1/22:1 alcohol/acid vs 18:1/22:1 acid/alcohol in the structure of the base JLE-40S and JLE-40A was shown to be significant in terms of phase transformation behavior.19 Despite the fact that these molecules are isomers, JLE-40S followed a linear trend compared to the symmetrical JLEs (JLE-36S and JLE-44S), to which JLE-40A does not adhere. The behavior of the branched derivatives of JLE-40A shows trends very similar to those of JLE-40S (by arbitrary convention, we refer to JLE40S as symmetrical, see above) as the number of branches increase. However, there are distinct differences in the crystallization behavior when the same number of branches in the derivatives of JLE-40S and JLE-40A are directly compared. As concluded from the study of linear esters,19 the observation of these differences suggests that the carbon number is not the only factor governing the behavior of the linear esters and their branched derivatives. The asymmetrical JLE-40A-2 with the same carbon number as JLE-40S-2 presented a wider crystallization peak (fwhm 6.22 ± 0.16 vs 2.91 ± 0.12 °C) and crystallized later (onset of crystallization −37.90 ± 0.04 vs −34.66 ± 0.07 °C, peak maximum −42.72 ± 0.08 vs −36.75 ± 0.09 °C). Additionally, the enthalpy of crystallization recorded by JLE-40S-2 was significantly higher than that of JLE-40A-2, suggesting a lower crystallinity and weaker intermolecular interactions in the crystalline lattice formed by JLE-40A-2. The crystallization data suggest that adding two branches to the linear JLEs does not suppress the effect of symmetry seen in the linear JLEsindeed, the effect is enhanced. This trend is maintained by the 3-branched derivatives; JLE-40S-3 records higher temperatures for onset and peak maximum of crystallization and a higher enthalpy of crystallization compared to JLE-40A-3. Interestingly, the differences in the temperatures of onset and peak maxima and the enthalpy are lower between JLE-40S-3 and JLE-40A-3 compared to those between JLE-40S2 and JLE-40A-2. This relates to the fact that in the 2-branched derivatives, there are more hydroxyl groups, which represent a greater degree of hydrogen bonding, as discussed below. It is worth noting that there is no statistically significant measurable difference between the crystallization parameters of JLE-40S-4 and JLE-40A-4, suggesting that the presence of hydroxyl groups in 2- and 3-branched derivative compounds is more important to an expression of crystallization differences as a result of symmetrical considerations than the location of the actual branches and their associated ester groups. The structural differences between the branched derivatives of JLE-40S and JLE-40A lie in the position of the branches and their associated ester groups (and hydroxyl groups, for 2-branched and 3branched derivatives) with respect to the ester bond on the base JLE. In the case of JLE-40S, the C22 moiety is derived from a fatty acid, while in the case of JLE-40A, the C22 moiety is derived from a fatty alcohol; this effectively means that the hydroxyl groups nearest to the ester bond linking the fatty acid and alcohol in the case of 2-branched derivatives are either at carbon lengths of 13/9 or 9/13, depending on whether the monoester is JLE-40S or JLE-40A, respectively. This clearly also extends to the ester groups introduced at the branches for both the 2-branched as well as 3- and 4-branched derivatives. The hydrogen-bonding introduced by the hydroxyl groups and the ester groups (bearing in mind that the hydrogen bonding introduced by the hydroxyl groups would be significantly

A careful polarized light microscopy (PLM) study did not reveal any crystal-like structure for any of the glass forming compounds indicating that the nature of the phase obtained at such low temperature is amorphous, in which very small crystallites (possibly smaller than the resolution limit of the PLM used, approximately 1 nm) may be enclosed. Clearly, it becomes hard for the esters to pack into regular lattices as branches are added (each branch introducing the addition of an ester group) and hydroxyl groups removed, requiring lower levels of supercooling for crystallization and therefore leading to a drastically lowered fraction of the sample actually participating in crystallization. 3.2.1. Crystallization Behavior of the Symmetrical Branched EstersComparing JLEs with the Same Number of Side Branches. The crystallization peak height increases and its fwhm decreases with increasing mass from JLE-36S derivatives to JLE-44S derivatives for any number of branches (follow same branched derivatives from Figure 1a−c). Except for the onset of crystallization of JLE-36S-2, crystallization values (peak, onset, and offset temperature) of the symmetrical compounds scale linearly (R2 > 0.845) as a function of carbon number for any given number of branches. The decrease in the crystallization values as a function of decreasing number of carbon atoms in the base JLE is most significant for derivatives with 2 branches, followed by those with 3 branches, with the decrease being very small for the 4-branched derivatives. Peak temperature change with number of carbons in the base JLE is a good illustration of the trends observed (Figure 3a). The variation of the enthalpy of crystallization with number of carbon atoms in the base JLE (Figure 3b) defines a linear trend,

Figure 3. Crystallization characteristics as a function of number of carbons in the base ester. (a) Peak temperature of crystallization of 2-, 3-, and 4-branched derivatives of symmetrical JLEs and (b) enthalpy of crystallization of 2-, 3-, and 4-branched derivatives of symmetrical JLs. 14897

dx.doi.org/10.1021/ie3016472 | Ind. Eng. Chem. Res. 2012, 51, 14892−14902

Industrial & Engineering Chemistry Research

Article

Figure 4. DSC heating (3.0 K/min) thermograms of branched derivative of (a) JLE-36S, (b) JLE-40S, (c) JLE-44S, and (d) JLE-40A. The base esters and JLE-40S-2 are represented in the right side panels with a right side heat flow (W/g) axis for comparison purposes.

stronger compared to the ester groups) and their relative positioning results in significant differences in packing of the isomers into regular crystal lattices. Clearly, the branched derivatives of the asymmetric JLE experiences a greater barrier to crystallization compared to the branched derivatives of the symmetric JLE, when there are hydroxyl groups present at the site of branching (i.e., the branching is not “saturated”). These relationships extend the usefulness of engineered structure to the utility of the branched derivatives of the JLEs; this is very important for this class of compounds, as the use of a longer acid or alcohol moiety in the synthesis of the esters is equally feasible given that all the esters are produced initially from fatty acids. 3.3. Melting Behavior of the Branched Derivatives of Jojoba Wax-like Esters. Heating behavior is also important, as it gives an indication of how the sample will recover from a freeze. Figure 4a−d clearly demonstrates the wide variety of transformation paths possible for the branched derivatives of JLEs. The heating profile of sample JLE-36S-2, for example, suggests very little polymorphic activity, whereas samples JLE40S-2 and JLE-44S-2, also with two branches, are extremely polymorphic. JLE-36S-2 would tend to recrystallize identically on heating, a characteristic which may be useful for wax-related developments, whereas, JLE-40S-2 and JLE-44S-2 go through an annealing process as they melt, converting to solids of everincreasing thermodynamic stabilities. Overall, the melting behavior of branched samples of JLE44S (Figure 4c) is similar to that of the JLE-40S-derived esters

(Figures 4b). Even though sample JLE-40S-2 has an onset of crystallization in the −40 °C range; on melting, it will continue to be solid-like until around 50 °C (see Figure 4b). Similarly, JLE-44S-2 transforms to increasingly stable thermodynamic states until around 20 °C, although it begins to melt around −60 °C (see Figure 4c). It is clear that these samples would therefore make poor low-temperature lubricants. From a structural perspective, however, this behavior warrants further investigation, if only to avoid similar behavior in a lubricant base-stock. Interestingly, although there is some limited polymorphic activity recorded in the melting of the branched derivatives of JLE-40A, this is in marked contrast to the extensive polymorphic activity recorded by the branched derivatives of JLE-40S. This will be discussed in more detail below. The ability of branched derivatives of JLE-40S and JLE-44S to form multiple polymorphs compared to those of JLE-36S and JLE40A relates on the one hand to symmetry considerations (JLE40S and JLE-40A) and on the other hand to chain length differences (JLE-36S and JLE-44S). Samples JLE-36S-2 to -4, JLE-40A-3, JLE-40A-4, and JLE40S-4 melt with a more or less single transition, a useful behavior in lubricants. JLE-36S-2 to -4, JLE-40S-3, JLE-40S-4, JLE-44S-4, JLE-40A-3, and JLE-40A-4 exhibit glass transitions which may be superimposed on the melting of a small crystalline phase. The DSC heating thermograms of the JLE40S-2, JLE-40A-2, and JLE-44S-2 branched esters did not present a glass transition, which is not surprising as there was 14898

dx.doi.org/10.1021/ie3016472 | Ind. Eng. Chem. Res. 2012, 51, 14892−14902

Industrial & Engineering Chemistry Research

Article

with increasing number of branches for the same base ester (ΔHJLE‑36S‑2 > ΔHJLE‑36S‑3 > ΔHJLE‑36S‑4) and decreases with mass of the base ester for the same number of branches (ΔHJLE‑36S‑4 < ΔHJLE‑40S‑4 > ΔHJLE‑44S‑4). The thermal protocol being the same, one can safely assert that the short-range order decreases with increasing number of branches for the same base ester (SROJLE‑36S‑2 > SROJLE‑36S‑3 > SROJLE‑36S‑4), and decreases with mass of the base ester for the same number of branches (SROJLE‑36S‑4 < SROJLE‑40S‑4 < SROJLE‑44S‑4). It is reasonable to suggest that the fractional free volume follows the same trend. This order may roughly correlate with the molecular mass and hydrogen bond strength. Note that there will a limit to densification and therefore to the free volume reduction. The two last thermal events in JLE-40S-3 and JLE-40S-4 and most probably in JLE-44S-4 are associated with the growth of crystalline embryos that might have nucleated at low temperatures and their subsequent melting. This is not unusual as many substances crystallize upon warming at temperatures about 10−30 °C above Tg as the crystal growth rate increases rapidly above Tg due to the decreasing viscosity of the liquid.35 Note however that the enthalpy associated with both the exotherm and the endotherm is less than 1 J/g, an indication of the very small amount of crystal phase involved. It is worth noting that very little crystalline material is needed to freeze a large portion of amorphous material into a rigid amorphous phase.36 Oils containing this phase would be a very interesting subject for research work as it would significantly affect their rheology. Onset of melt and temperature at maximum heat flow of the leading melting peak for all the samples increase significantly with increasing number of carbon atoms of the base ester. Compared to the 2-branched derivatives, the 3-branched derivatives melt with relatively little polymorphic activity with small latent heat involved, except for JLE-44S-3, which displayed a relatively large and narrow endotherm (Figure 4c). The 4-branched derivatives, with significantly lowered melting points, melted with a more or less single transition and had practically the same very small enthalpy of melting. When the JLE-40A samples are excluded, peak temperature and onset of melting (not shown) of the leading event vary linearly with the number of carbon atoms of the base ester for any given number of branches. Conversely, as the number of carbon atoms decreases in the base ester, the melting behavior relatively simplifies for any given number of side branches as less crystalline material is present with a sharp decrease of the enthalpy of melting. 3.3.1. Effect of Symmetry on the Melting Behavior of the Branched Derivatives of JLEs. Similar to the behavior displayed by the JLE-40S branched derivatives, the melting behavior of branched samples of JLE-40A simplifies as the number of branches increases (Figure 4b and d). However, the branched derivatives of JLE-40A have lower melting characteristics such as onset, peak of melt, and Tg compared to the branched derivatives of JLE-40S, despite having the same total amount of carbon. Perhaps the most surprising behavior is the presence of a large exotherm in the melting profile of JLE-40S2, indicating a strong recrystallization from the melt (Figure 4b), compared to the behavior witnessed with JLE-40A-2 (Figure 4d) which indicates, if any, very weak polymorphism. JLE-40A-3 melts in a similar fashion to JLE-40S-3, with a glass transition followed by the melting of a crystalline phase, but with a lower Tg and Tp (Figure 4b and 4d).

no evidence of such a transition in their DSC cooling trace also. JLE-40S-3 (similar to JLE-40A-3) begins to melt with a glass transition (arrow 1 in Figure 4b) closely followed by a relatively large endotherm characteristic of the melting of a crystal phase (arrow 2 in Figure 4b). This endotherm is followed by a faint exotherm and then a very small endotherm (arrow 3 and 4, respectively in Figure 4b). The glass transition observed in JLE40S-4 was immediately followed by an exotherm rather than an endotherm (indicated by arrows in Figure 4d). The JLE-44S-4 sample also seems to have the same transformation path as JLE40S-4 with two very distinct melting peaks separated by a broad exotherm (indicated by arrows in Figure 4d). The transition points recorded during the DSC heating cycle of the samples are dramatically lowered as the number of branches increases. Furthermore, when excluding the base ester, all the characteristic temperatures such as peak temperature (Tp), onset (Ton), and offset (Toff) scale linearly with the number of branches (R2 > 0.996). The Tg measured in the heating cycle also decreased linearly as a function of number of branches (not shown). As additional branches are added, the melting behavior becomes relatively simpler as less crystalline material is present; this is indicated by the decrease of the enthalpy of melting. Free volume models27 can be used to describe the experimental observations and changes in the transport properties of glassy liquids and interpret the effect of molecular mass, hydrogen bonds, composition, and pressure on Tg. It is also generally recognized that the dramatic changes in transport properties at the glass transition normally accompanies a reduction of free volume. The balance between the disruption of chain packing resulting from the introduction of the branches and the order due to the change in hydrogen bonds is likely important. The variation in Tg with increasing molecular mass can be attributed to the rapid reduction of the rate of diffusion mobility and free volume, competing with an increase of free volume due to the decrease of hydrogen bonds.28 DSC heating cycles of the branched samples undergoing a glass transition (JLE-36S-2 to -4, JLE-40S-3 to -4, JLE-40A-3 to -4, and JLE-44S-4) show a relatively deep endothermic peak at the offset of the glass transition (Figure 4a−d). Such endotherms are routinely recorded in the heating trace of glass transitions.29 They do not represent the heat of melting, but rather the endothermic enthalpy of relaxation at Tg which arises from the cooperative motions in the Tg region.30−32 There is an increase in the density and decrease in entropy and free volume during the glass transition toward the supercooled liquid.33 When the glass is heated through Tg, the excess enthalpy is regained and this event can be observed as an endotherm associated with the glass transition in the DSC profile. The amorphous system reaches the metastable equilibrium state by losing the enthalpy and decreasing the free volume. Molecular mobility increases dramatically at Tg, leading to a loss of short-range order (SRO), which is an endothermic process along the glass line reminiscent of the loss of longrange order due to melting. Consequently, the higher the SRO of the amorphous phase, the greater the enthalpy of relaxation and the greater the endotherm. The phenomenon is particularly exacerbated in a standard DSC if the heating rate is different from the cooling rate.34 For the present compounds, the enthalpy of relaxation, measured as the area of the relaxation endotherm, decreases 14899

dx.doi.org/10.1021/ie3016472 | Ind. Eng. Chem. Res. 2012, 51, 14892−14902

Industrial & Engineering Chemistry Research

Article

Figure 5. (a) SFC/% versus temperature curves of the 4-branched derivatives of JLEs obtained during cooling (3.0 K/min). (b) SFC/% versus time curves of the branched derivatives of JLE-36S obtained during cooling (3.0 K/min). The curves are shifted by 10 min from each other for clarity. (c) SFC/% versus temperature curves of the 2-branched derivatives of JLEs obtained during heating (3.0 K/min). (d) SFC induction temperature versus side branches of symmetrical JLEs obtained during cooling (3.0 K/min). Symbols represent branched derivatives of ●, JLE-36S; ▲, JLE-40S; and ■, JLE-44S.

relatively complex shapes, the SFC versus temperature curves of all the branched compounds are reminiscent of single sigmoids. As illustrated with the case of JLE-36S (Figure 5b), the variances observed within the growth curve of the base JLEs diminish significantly as these esters are branched. As branches are added to the linear esters, their SFC behavior tends to be more closely related. SFC versus temperature curves of the JLEs with 4 branches look almost identical, while those with 3 branches are more similar than those with 2 branches. The SFC data indicate that overall, the branched compounds solidify in a similar fashion but with subtle differences, corroborating the findings from the DSC cooling experiments. The sigmoid-shaped SFC curves can be associated with an overall solidification occurring in one main step. In the case of samples demonstrating a glassy transition, this one step solidification process can be safely related to the formation of the glassy phase. The increasing similarity of the SFC curves as the number of branches increases can be safely associated with the increasing amount of the glassy phase involved in the solidification process. The peculiar shape of the SFC curves of the branched derivatives of JLE-44S suggests that more than one process takes place during their solidification and is attributable to its high crystalline phase content. Obviously, one can expect differences in the processes involved in the crystal and glassy phase formation, and consequently, that the SFC cooling curves reflect such differences. Besides the possible involvement of multiple crystallization processes, competition between crystal

Note that even for the 4-branched derivatives, although JLE40A-4 records almost the same Tg as JLE-40S-4, there is still a significantly smaller amount of crystalline phase formed from the melt (arrows 3 in Figure 4b and 4d). This supports the importance of the role that asymmetry of the base JLEs plays in the crystallization and melting behavior of the branched derivatives, highlighted above. Certainly, from a practical perspective, this melting behavior suggests that the branched samples from JLE-40A are much better candidates for lubricant formulations than JLE-40S, particularly evident in the melting profile of JLE-40A-2 compared to JLE-40S-2. This is behavior that can be exploited in both waxes as well as lubricants. 3.4. Solid Fat Content of the Branched Derivatives of Jojoba Wax-like Esters. The most significant finding in the SFC behavior of the derived esters is the recording of almost 100% final SFC for the cooling cycle. This is true even for compounds where no crystalline phase was detected by DSC, PLM, or XRD, confirming the glassy nature of the phases reached at very low temperature. The effect is illustrated in Figure 5a, which shows the SFC (%) versus temperature cooling (3.0 K/min) curves of the 4-branched derivatives of JLEs. In fact, the SFC data correlated very well with DSC for both the cooling and heating cycles. The trends for onset and offset temperature, as well as polymorphic transformation observed in the heating cycle of JLE-40S-2 and JLE-44S-2 were confirmed. The solidification mechanism and consequently the kinetics, as evidenced by SFC of the branched derivatives, are dramatically simplified compared to the base JLEs.19 Except for the branched derivatives of JLE-44S which displayed 14900

dx.doi.org/10.1021/ie3016472 | Ind. Eng. Chem. Res. 2012, 51, 14892−14902

Industrial & Engineering Chemistry Research

Article

which are very promising in the formulation of renewable lubricants.

and glass formation processes are at play during the cooling of many of the samples, as highlighted above. The heating SFC cycle of the branched esters mirrored their DSC melting behavior. Again, branched derivatives of sample JLE-44S are different from the other samples, explainable by its tendency to crystallize. Most noticeable is the very distinct dip in the SFC (%) heating cycle of JLE-40S-2 and JLE-44S-2 (Figure 5c), revealing a recrystallization event which corroborates the formation of a solid phase mediated by melting in these two samples. The temperature at the minimum of the dip corresponds to the temperature at which a recrystallization peak was observed in their respective DSC heating thermograms. As was observed during the cooling cycle, the SFC heating behavior of the branched derivatives tends to become very similar as branches are added to the base JLEs. SFC heating curves of the derivatives with two branches are less similar than the derivatives with three branches which in turn are less similar than the derivatives with four branches, which look almost identical. The SFC data indicate that overall, the branched compounds having a large glassy phase melt in a similar fashion. The central role that is played by the presence of the hydroxyl groups in the branched derivatives of the JLEs, highlighted above, is supported by both the SFC cooling as well as heating curves. The SFC induction temperature, Ti, determined as the temperature at which the SFC deviates from the baseline, Ti, correlates very well with number of branches for any given base JLE (Figure 5d). Linear regression of the Ti versus branches relationship demonstrates that the addition of 1 branch, quantified by the slope of the line, reduces the tendency for solidification for branched derivatives of JLE-36S and JLE-40S to a greater degree than for JLE-44S. This confirms that the glassy phase which was more readily formed by the branched derivatives of JLE-36S is less prone to dramatic changes than the crystalline phase which is prevalent in those of JLE-44S. Conversely, Ti increases linearly (R2 > 0.998) with carbon number for any given number of branches (not shown). The relationships among solid fat content, temperature, number of branches, and carbon number for symmetric JLEs elucidated here provide the formulator with specific information on the required parameters for particular applications related to waxes, lubricants, and cosmetic chemicals. The dramatic effect of symmetry can be appreciated in the strong polymorphism observed in the SFC heating cycle of JLE-40S-2 compared to the straight melting observed for JLE40A-2 (Figure 5c). Significant differences in SFC heating curves were also observed for the 3-branched compoundsfor example, the symmetrical samples required a longer induction time (higher temperature) for melting to begin.

■

ASSOCIATED CONTENT

S Supporting Information *

Scheme S1: structure of the jojoba-like mono-esters (JLE) and their branched derivatives. Table S1: nomenclature used for the jojoba-like mono-esters (JLE) and their branched derivatives prepared in this work. The column headed “Structure” refers to the generalized structures shown in Scheme S1. Table S2: synthesis and 1H-NMR characterization data of the epoxides of JLE-36S [8-(3-octyloxiran-2-yl)octyl 8-(3-octyloxiran-2-yl)octanoate], JLE-40S [8-(3-octyloxiran-2-yl)octyl 12(3-octyloxiran-2-yl)dodecanoate], JLE-44S [12-(3-octyloxiran2-yl)dodecyl 12-(3-octyloxiran-2-yl)dodecanoate], and JLE40A [12-(3-octyloxiran-2-yl)dodecyl 8-(3-octyloxiran-2-yl)octanoate]. EA/HE: ethyl acetate/hexane. Table S3: synthesis and 1H-NMR characterization data of the branched derivatives of JLEs. EA/HE: ethyl acetate/hexane. The structure and nomenclature of branched derivatives of JLEs are presented in Scheme S1 and Table S1. Table S4: enthalpy of crystallization and melting of the branched derivatives of JLEs. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Tel.: 1-705-748-1011. Ext. 6105 Fax: 1-705-748-1652. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The financial support of Elevance Renewable Sciences, NSERC, Grain Farmers of Ontario, GPA-EDC, Industry Canada, and Trent University is gratefully acknowledged.

■

REFERENCES

(1) Wagner, H.; Luther, R.; Mang, T. Lubricant base fluids based on renewable raw materials - Their catalytic manufacture and modification. Appl. Catal. A−Gen. 2001, 221 (1−2), 429−442. (2) Erhan, S. Z.; Adhvaryu, A.; Sharma, B. K., Chemically Functionalized Vegetable Oils. In Synthetics, Mineral Oils, and BioBased Lubricants. Chemistry and Technology; Rudnick, L. R., Ed.; CRC Press: Boca Raton, FL, 2005; Vol. 111, pp 364−390. (3) Le Dréau, Y.; Dupuy, N.; Gaydou, V.; Joachim, J.; Kister, J. Study of jojoba oil aging by FTIR. Anal. Chim. Acta 2009, 642 (1−2), 163− 170. (4) Wisniak, J. Potential uses of jojoba oil and meal – a review. Ind. Crops Prod. 1994, 3 (1−2), 43−68. (5) Miwa, T. Jojoba oil wax esters and derived fatty acids and alcohols: Gas chromatographic analyses. J. Am. Oil Chem. Soc. 1971, 48 (6), 259−264. (6) Busson-Breysse, J.; Farines, M.; Soulier, J. Jojoba wax: Its esters and some of its minor components. J. Am. Oil Chem. Soc. 1994, 71 (9), 999−1002. (7) Tobares, L.; Frati, M.; Guzmán, C.; Maestri, D. Agronomical and chemical traits as descriptors for discrimination and selection of jojoba (Simmondsia chinensis) clones. Ind. Crops Prod. 2004, 19 (2), 107− 111.

4. CONCLUSIONS Branched derivatives of jojoba wax-like monoesters were synthesized by epoxidation and ring-opening esterification with propionic acid. Their crystallization and melt behavior was elucidated as a function of chain length, branching, symmetry, and functional groups. Growing restriction to the participation of molecules in crystallization due to the increased structural complexity of the molecule was observed. The phase of the highly branched compounds obtained at low temperatures is mostly amorphous, in which very small crystal domains may be enclosed. The findings highlight relatively simple, inexpensive lipid-derived molecules with astonishingly low onsets of crystallization and predictive structure−property relationships 14901

dx.doi.org/10.1021/ie3016472 | Ind. Eng. Chem. Res. 2012, 51, 14892−14902

Industrial & Engineering Chemistry Research

Article

(8) Binman, S.; Belfer, S.; Shani, A. Functionalization at the doublebond region of jojoba oil. 7. Chemical binding of jojoba liquid wax to polystyrene resins. J. Am. Oil Chem. Soc. 1996, 73 (9), 1075−1081. (9) Patel, S.; Nelson, D. R.; Gibbs, A. G. Chemical and physical analyses of wax ester properties. J. Insect. Sci. 2001, No. 1.4, ; available online insectscience.org/1.4. (10) Bhatia, V. K.; Chaudhry, A.; Sivasankaran, G. A.; Bisht, R. P. S.; Kashyap, M. Modification of jojoba oil for lubricant formulations. J. Am. Oil Chem. Soc. 1990, 67 (1), 1−7. (11) Tobares, L.; Guzmán, C.; Maestri, D. Effect of the extraction and bleaching processes on jojoba (Simmondsia chinensis) wax quality. Eur. J. Lipid Sci. Technol. 2003, 105 (12), 749−753. (12) Sanchez, N.; Martinez, M.; Aracil, J.; Corma, A. Synthesis of oleyl oleate as a jojoba oil analog. J. Am. Oil Chem. Soc. 1992, 69 (11), 1150−1153. (13) Kalscheuer, R.; Stoveken, T.; Luftmann, H.; Malkus, U.; Reichelt, R.; Steinbuchel, A. Neutral Lipid Biosynthesis in Engineered Escherichia coli: Jojoba Oil-Like Wax Esters and Fatty Acid Butyl Esters. Appl. Environ. Microbiol. 2006, 72 (2), 1373−1379. (14) Bautista, L. F.; Martinez, M.; Aracil, J. Optimization of a process for the synthesis of a sperm whale oil analogue. Chem. Eng. Technol. 1997, 20 (4), 287−292. (15) Martinez, M.; Torrano, E.; Aracil, J. Synthesis of esters of high molecular-weight - an analog of jojoba oil - a statistical approach. Ind. Eng. Chem. Res. 1988, 27 (11), 2179−2182. (16) Coteron, A.; Sanchez, N.; Martinez, M.; Aracil, J. Optimization of the synthesis of an analog of jojoba oil using a fully central composite design. Can. J. Chem. Eng. 1993, 71 (3), 485−488. (17) Garcia, T.; Martinez, M.; Aracil, J. Enzymatic-synthesis of an analog of jojoba oil - optimization by statistical-analysis. Enzyme Microb. Technol. 1993, 15 (7), 607−611. (18) Campanella, A.; Rustoy, E.; Baldessari, A.; Baltanás, M. A. Lubricants from chemically modified vegetable oils. Bioresour. Technol. 2010, 101 (1), 245−254. (19) Bouzidi, L.; Li, S.; Di Biase, S.; Rizvi, S. Q.; Narine, S. S. Lubricating and waxy esters, I. Synthesis, crystallization, and melt behavior of linear monoesters. Chem. Phys. Lipids 2012, 165 (1), 38− 50. (20) Moser, B. R.; Sharma, B. K.; Doll, K. M.; Erhan, S. Z. Diesters from oleic acid: Synthesis, low temperature properties, and oxidation stability. J. Am. Oil Chem. Soc. 2007, 84 (7), 675−680. (21) Lynch, B. M.; Pausacker, K. H. The oxidation of olefins with perbenzoic acids. A kinetic study. J. Chem. Soc. (Resumed) 1955, 0, 1525−1531. (22) Bouzidi, L.; Boodhoo, M.; Humphrey, K. L.; Narine, S. S. Use of first and second derivatives to accurately determine key parameters of DSC thermographs in lipid crystallization studies. Thermochim. Acta 2005, 439 (1−2), 94−102. (23) Dunn, R. O.; Bagby, M. O. Low-temperature filterability properties of alternative diesel fuels from vegetable oils. Fuel Energy Abst. 1997, 38 (5), 330−330(1). (24) Dunn, R. O. Thermal analysis of alternative diesel fuels from vegetable oils. J. Am. Oil Chem. Soc. 1999, 76 (1), 109−115. (25) Foglia, T.; Nelson, L.; Dunn, R.; Marmer, W. Low-temperature properties of alkyl esters of tallow and grease. J. Am. Oil Chem. Soc. 1997, 74 (8), 951−955. (26) Lee, I.; Johnson, L.; Hammond, E. Use of branched-chain esters to reduce the crystallization temperature of biodiesel. J. Am. Oil Chem. Soc. 1995, 72 (10), 1155−1160. (27) Haward, R. N. Occupied Volume of Liquids and Polymers. J. Macromol. Sci., Part C 1970, 4 (2), 191−242. (28) McGonigle, E. A.; Cowie, J. M. G.; Arrighi, V.; Pethrick, R. A. Enthalpy relaxation and free volume changes in aged styrene copolymers containing a hydrogen bonding co-monomer. J. Mater. Sci. 2005, 40 (8), 1869−1881. (29) Mark, J. E. Physical Properties of Polymers Handbook, 2nd ed.; American Institute of Physics: Woodbury, NY, 2007; Vol. XIX, p 1076.

(30) Boller, A.; Schick, C.; Wunderlich, B. Modulated differential scanning calorimetry in the glass transition region. Thermochim. Acta 1995, 266 (0), 97−111. (31) Hakeyama, T.; Quinn, F. X. Thermal Analysis: Fundamentals and applications to polymer science, 2nd ed.; John Wiley & Sons: Chichester, 1999; p 190. (32) Wunderlich, B. Thermal Analysis; Academic Press: New York, NY, 1990; p 464. (33) Surana, R.; Pyne, A.; Rani, M.; Suryanarayanan, R. Measurement of enthalpic relaxation by differential scanning calorimetryeffect of experimental conditions. 2005, 433 (1−2), 173−182. (34) Badrinarayanan, P.; Zheng, W.; Li, Q.; Simon, S. L. The glass transition temperature versus the fictive temperature. J. Non-Cryst. Solids 2007, 353 (26), 2603−2612. (35) Koop, T.; Bookhold, J.; Shiraiwa, M.; Poschl, U. Glass transition and phase state of organic compounds: dependency on molecular properties and implications for secondary organic aerosols in the atmosphere. Phys. Chem. Chem. Phys. 2011, 13 (43), 19238−19255. (36) Wunderlich, B. Reversible crystallization and the rigid− amorphous phase in semicrystalline macromolecules. Prog. Polym. Sci. 2003, 28 (3), 383−450.

14902

dx.doi.org/10.1021/ie3016472 | Ind. Eng. Chem. Res. 2012, 51, 14892−14902