Cooling Rate Effects on the Microstructure, Solid Content, and

May 22, 2013 - Yaqi Lan , Michael A. Rogers ... Mixtures, Vegetable Oil, Stirring Speed, and Temperature on the Physical Properties of Organogels...
0 downloads 5 Views 2MB Size
Article pubs.acs.org/Langmuir

Cooling Rate Effects on the Microstructure, Solid Content, and Rheological Properties of Organogels of Amides Derived from Stearic and (R)‑12-Hydroxystearic Acid in Vegetable Oil Jorge F. Toro-Vazquez,*,† Juan Morales-Rueda,† Adriana Torres-Martínez,† Miriam A. Charó-Alonso,† V. Ajay Mallia,‡ and Richard G. Weiss‡,§ †

Facultad de Ciencias Químicas-CIEP, Universidad Autónoma de San Luis Potosí, SLP 78210, Mexico Department of Chemistry and §Institute of Soft Matter Synthesis and Metrology, Georgetown University, Washington, D.C. 20057-1227, United States



S Supporting Information *

ABSTRACT: Using safflower oil as the liquid phase, we investigated the organogelation properties of stearic acid (SA), (R)-12-hydroxystearic acid (HSA), and different primary and secondary amides synthesized from SA and HSA. The objective was to establish the relationship between the gelator’s molecular structure, solid content, and gels’ microstructure that determines the rheological properties of organogels developed at two cooling rates, 1 and 20 °C/min. The results showed that the presence of a 12-OH group in the gelator molecule makes its crystallization kinetics cooling rate dependent and modifies its crystallization behavior. Thus, SA crystallizes as large platelets, while HSA crystallizes as fibers forming gels with higher solid content, particularly at 20 °C/min. The addition to HSA of a primary or a secondary amide bonded with an alkyl group resulted in gelator molecules that crystallized as fibrillar spherulites at both cooling rates. Independent of the cooling rate, gels of HSA and its amide derivatives showed thixotropic behavior. The rheological properties of the amide’s organogels depend on a balance between hydrogen-bonding sites and the alkyl chain length bonded to the amide group. However, it might also be associated with the effect that the gelators’ molecular weight has on crystal growth and its consequence on fiber interpenetration among vicinal spherulites. These results were compared with those obtained with candelilla wax (CW), a well-known edible gelling additive used by the food industry. CW organogels had higher elasticity than HSA gels but lower than the gels formed by amides. Additionally, CW gels showed similar or even higher thixotropic behavior than HSA and the amide’s gels. These remarkable rheological properties resulted from the microstructural organization of CW organogels. We concluded that microstructure has a more important role determining the organogels’ rheology than the solid content. The fitting models developed to describe the organogels rheological behavior support this argument.

1. INTRODUCTION Molecular organogels are bicontinuous, frequently colloidal systems that coexist as aggregated low molecular mass organic molecules (LMOGs) and an organic liquid. Organogel formation is based on the spontaneous self-assembly of individual gelator molecules into three-dimensional (3D) networks of randomly entangled fiberlike or platelike structures. This 3D network, resulting from a supramolecular organization, has viscoelastic and thermo-reversible properties and holds micro domains of the liquid in a nonflowing state mainly through surface tension and capillary forces.1 In these systems, the LMOG is only slightly soluble in the liquid at the temperatures where the gel exists. The potential applications of these supramolecular materials rely on the fact that many classes of organic liquids, including vegetable oils, can be gelled efficiently.2−6 This characteristic and the organogels’ physical properties have led to potential applications in cosmetics,7 drug delivery,8,9 and the food industry.3,5,6 © XXXX American Chemical Society

In most cases, the relationship between the gelator chemical structure and its gelling capability is not evident a priori.4,10,11 However, some systematic studies have been performed to provide insights into structure−gelation relationships. For example, Mallia et al.11,12 studied the relationship between molecular structure and the gelation properties of LMOG amides, amines, and ammonium salts based on (R)-12hydroxystearic acid in several organic liquids. Recently,4 we reported the relationships between the molecular structures of (R)-12-hydroxyoctadecanamide (HOA), (R)-N-propyl-12-hydroxyoctadecanamide (PHOA), and (R)-N-octadecyl-12-hydroxyoctadecanamide (OHOA) and the thermo-mechanical properties of their 2% (w/w) organogels, using safflower oil high in oleic acid (HOSFO) as the liquid phase. Received: March 4, 2013 Revised: May 20, 2013

A

dx.doi.org/10.1021/la400809a | Langmuir XXXX, XXX, XXX−XXX

Langmuir

Article

From our perspective, a robust approach to design LMOGs with particular properties ought to establish correlations among the chemical structures of LMOGs, their ability to self-assemble and develop a 3D network with particular topology, and the macroscopic mechanical properties of the material (i.e., its rheology). Such correlations must be supported by the identification and understanding of the factors that control the crystal network formation and topology of the LMOG aggregates.14 These include the establishment of both the nature of the short-range weak forces and the strong solvent dependence on the molecular self-assembling capability of the gelator,13,16 as well as the effect of the thermodynamic and mass transfer conditions applied through external forces (i.e., crystallization temperature, cooling rate, and shearing). These parameters induce molecular alignment, crystal growth, and crystal network organization to engineer the desired macroscopic properties of supramolecular soft materials, such as organogels. Organogelation is, in most cases, easily achieved by heating an LMOG in a low polarity solvent until achieving its full solubility and then cooling the system below the solubility limit temperature, TS. For a given LMOG in solution at a given concentration, its crystallization and subsequent gelation depends mainly on the gel setting temperature, Tset, where Tset < TS. The difference between the melting temperature of the gelator in the solution/sol (TM) and Tset establishes the supercooling (i.e., ΔT = TM − Tset), the thermodynamic driving force for microphase separation. Alternatively, we might define this force in terms of the prevalent supersaturation for the LMOG at a given temperature.17 In any case, during cooling of the LMOG solution/sol to achieve Tset, supercooling (and supersaturation) increases as the cooling rate increases. Within this framework, the primary aim of this study is to establish the effect of cooling rate, using 1 °C/min and 20 °C/ min, to achieve a given Tset for 2% organogels in HOSFO as the liquid phase on the relationships among the molecular structures of LMOGs, the solid content in the gel, the microstructure developed by the LMOGs, and the rheological properties. The LMOGs investigated include primary and secondary amides synthesized from stearic acid (SA) and (R)12-hydroxystearic acid (HSA):1-octadecanamide (OA), HOA, PHOA, and OHOA (Figure 1). Fatty acid amides are found in nature but are seldom encountered in significant quantities in edible fats and oils. However, they are produced on a large scale and used in the production of fiber lubricants, detergents, flotation agents, textile softeners, antistatic agents, mold release agents, and plasticizers for the polymer industry.18−20 To support the achievement of our objective and to establish a more comprehensive evaluation, we have included octacedecanoic acid (i.e., stearic acid, SA), and (R)-12-hydroxystearic acid (HSA) in the study. As in our previous investigation, we will compare the results with these amides with those obtained with 2% candelilla wax (CW) organogels. CW is a well-known edible gelling additive approved by the FDA under regulations 21CFR, 175.105, 175.320, and 176.180.5,6 The main constituent in CW is the long n-alkane, hentriacontane (CH3(CH2)29CH3).5,6 2. Materials and Methods. 2.1. Materials. The LMOG amides were prepared, as described previously.11 SA and HSA had a purity of 98.5% and 99%, respectively, and were obtained from Sigma-Aldrich Company (St Louis, MO). HOSFO (Coral Internacional, San Luis Potosi,́ Mexico) and CW (Multiceras, Monterrey, Mexico) were obtained from local manufacturers.

Figure 1. Structures of the LMOG amides investigated. Hentriacontane is the main component in CW.

The major triacylglycerol in this HOSFO is 63.32% (±0.06%) triolein. The CW had 44−45% of n-alkanes with 28−33 carbon atoms, 6−7.4% of esters of aliphatic acids and alcohols, 15− 18.8% of aliphatic acids with 18−34 carbon atoms, 5−7.6% aliphatic alcohols with 24−34 carbon atoms, 21−23% of alcohols of penta-cyclic triterpenoids, and 1.9−2.2% of esters of alcohols of penta-cyclic triterpenoids. Of the n-alkanes present, hentriacontane was the major constituent (C31H64; 75.9% ± 0.1%), with lower concentrations of nonacosane (4.2% ± 0.1%; C29H60), triacontane (4.2% ± 2.0%; C30H62), dotriacontane (2.8% ± 0.4%; C32H66), and tritriacontane (9.9% ± 0.4%; C33H68).21 We solubilized each LMOG amide and the CW at 2% (w/w) in HOSFO by heating to 140 °C and agitation. 2.2. Calorimetry and Solid Content. Using a differential scanning calorimeter (model Q1000, TA Instruments; New Castle, DE), samples of neat LMOG, SA, HSA, amides, and CW (0.5−7 mg) were heated at 140 °C for 20 min and then cooled to −20 °C at 10 °C/min. After 2 min at −20 °C, the system was heated to 140 °C at 5 °C/min. For the 2% (w/w) gelator solutions in HOSFO, ∼7−10 mg were heated at 120 °C for 20 min, cooled to 10 °C at either 1 °C/min or 20 °C/min, maintained at this temperature for 5 min, and then heated to 25 °C (5 °C/min). After 75 min at this temperature, the samples were melted by heating to 120 °C (5 °C/min). The temperature at the beginning of the crystallization exotherm (TCr), the heat of crystallization (ΔHCr), and during heating, the temperature of maximum heat flow (TM), and the heat of melting (ΔHM) were calculated as described previously.4 For neat gelators, ΔHCr and ΔHM were reported in the molar basis. The percentage of solid content in the 2% organogels was B

dx.doi.org/10.1021/la400809a | Langmuir XXXX, XXX, XXX−XXX

Langmuir

Article

In eq 1 (Figure 2A), G0 is the instantaneous elastic modulus and defines the resistance of the sample to deformation (i.e., G0

determined by low-frequency nuclear magnetic resonance (Minispec Bruker mq20; Bruker Analytik; Rheinstetten, Germany), following the direct method of the AOCS official method Cd 16b-9322 but with the use of similar time− temperature conditions as those for the DSC determinations. The solid content was reported as the solid mass fraction (SMF). For SMF, at least four independent measurements were made and at least two measurements for the reported DSC data. The ΔHCr and ΔHM values for the 2% organogels were normalized for their solid content and are reported as kJ/mol. The analysis of variance and contrast among the mean values of the thermal parameters were calculated with STATISTICA version 9.0 (StatSoft Inc., Tulsa, OK). 2.3. G′/G″ and Creep and Recovery Profiles. The elastic (G′) and loss (G″) moduli of the 2% organogels were determined with a mechanical spectrometer (Paar Physica MCR 301, Stuttgart, Germany), using a steel cone−plate geometry (CP25−1/TG, Anton Paar, Graz-Austria) equipped with a true-gap system. The sample temperature was controlled through a Peltier temperature control located on the base of the geometry and with a Peltier-controlled hood (H-PTD 200). An aliquot of a 2% gelator solution was placed on the base plate of the rheometer, and the cone was set using the true-gap function of the software. We employed the same time−temperature program as in the calorimetry studies. Thus, after 15 min at 25 °C, the G′ and G″ moduli were measured during 60 min, always within the linear viscoelastic region (LVR). At the 0.5 or 1.0 Hz frequencies used, the strain (γ) applied was always within 0.002% and 0.01%, respectively. After the 75 min at 25 °C, the yield stress (σ*) of the organogels was determined by applying a strain sweep between 0% and 100%. σ* was calculated from a log−log plot of shear stress versus γ (%), at the corresponding upper limit of strain. To determine the creep and recovery profile, a new organogel was made in the rheometer and, after the 75 min at 25 °C, a constant stress of 17.5 Pa was applied to the gel for 60 s while measuring the γ (i.e., in the creep stage). After this time, the force was withdrawn while γ measurements continued for an additional 300 s (i.e., in the recovery stage). The creep (i.e., slow and progressive deformation of the material under constant stress) and recovery profiles were determined by plotting the compliance (J) as a function of time. J is the ratio between γ of the sample and the stress applied. The stress applied (17.5 Pa) corresponded to a force equivalent of 50% of the σ* for the 2% CW organogels. This force was selected based on σ* profiles and exploratory creep and recovery measurements with the 2% organogels. In all cases, two independent determinations were made and the mean used for statistical analysis. The creep profile was fitted to Burger’s model (eq 1), and the recovery profile to eq 2.23 The corresponding parameters were obtained using the quasi-Newtonian methodology using STATISTICA version 9.0 (StatSoft Inc.). The % recovery achieved at 300 s after releasing the stress was calculated using eq 3. J = 1/G0 + 1/G1[1 − exp( −tG1/n1)] + t /n0

(1)

J = JC + JKV exp( −Bt C)

(2)

Recovery % = [Jmax − J(t )]/Jmax × 100

(3)

JSM = Jmax − (JC + JKV )

(4)

Figure 2. Description of Burger’s parameters in a generalized (A) creep and (B) recovery profile. The parameters associated with Burger’s equation (eq 1) and the ones calculated using eqs 2 and 3 are defined in the text.

occurs immediately during the deformation profile and is instantaneously recovered when the stress is removed), and G1 is the contribution of the retarded elastic region to the total compliance. n0 is the residual viscosity or viscous flow of the system after suffering deformation, and n1 is the internal viscosity. Within this framework, the rate at which the system has achieved the maximum strain, also known as the retardation or delay time, λret, is equal to n1/G1. If the system is a Hookean solid (i.e., completely elastic), λret = 0 and the maximum strain would be obtained immediately after the application of a constant stress. Viscoelastic materials like gels are characterized by a delay in achieving the maximum strain, the λret. On the other hand, in eq 2 (Figure 2B), JC corresponds to the permanent or residual deformation, and JKV is a slow or retarded recovery due to a decreasing exponential type and tends toward an asymptote when time t → ∞; B and C (not calculated in this investigation) are parameters that define the recovery speed of the sample. The percentage of recovery achieved by the gel at a given time (% Rt) was calculated with eq 3, where JMAX is the maximum compliance achieved during the creep profile and J(t) is the compliance at t = 300 s after the stress force was withdrawn. The initial or instantaneous shear compliance of the system, JSM (Figure 2B), is difficult to determine experimentally and was calculated using eq 4, where C

dx.doi.org/10.1021/la400809a | Langmuir XXXX, XXX, XXX−XXX

Langmuir

Article

Table 1. Thermal parameters (crystallization and melting) for Neat Gelators and for 2% (w/w) Organogels in HOSFOa neat gelators functional group

TCr (°C)

ΔHCr (kJ/mol)

2% organogels

TM (°C)

ΔHM (kJ/mol)

cooling rate (°C/min) 1

SA*

a

a

a

a

65.12 (1.03)

65.15 (1.32)

70.12 (0.47)

64.28 (1.35)

73.46b (0.02)

51.13b (0.57)

79.75b (0.04)

50.45b (0.89)

104.75c (0.43)

50.72b (3.77)

106.98c (0.03)

48.05b (3.29)

109.69d (0.25)

53.17b (1.63)

113.70d (0.81)

53.53c (0.57)

99.99e (0.06)

59.42c (0.27)

106.93c (0.01)

61.24d (0.34)

102.75f (0.28)

92.45d (1.09)

106.74c (0.01)

90.55e (1.13)

76.58g (0.68)

64.37a (0.83)

64.42e (0.23)

65.42a (0.53)

20 1

HSA

20 1

OA

20 1

HOA

20 1

PHOA

20 1

OHOA

20 1

CW

20

TCr (°C)

ΔHCr (kJ/molSolids)

TM (°C)

ΔHM (kJ/molSolids)

solid Mass Fraction × 102

27.49a,a (0.26) 21.52a,b (0.17) 51.65b,a (1.24) 48.18b,b (0.43) 69.49c,a (0.81) 67.11c,b (3.25) 93.06d,a (1.44) 92.16d,a (2.12) 77.44e,a (0.93) 74.85e,b (0.65) 80.54f,a (0.67) 76.42e,b (0.91) 38.32g,a (0.34) 35.83f,b (0.18)

115.4a,a (23.2) 127.6a,a (1.6) 54.1b,a (7.5) 50.1b,a (2.9) 223.0c,a (38.3) 136.0a,b (10.4) 115.5a,a (11.8) 97.9c,a (2.9) 139.9d,a (0.6) 87.8c,b (6.9) 251.9e,a (21.1) 112.3a,b (4.8) 84.6f,a (4.1) 45.4b,b (0.9)

38.78a,a (0.53) 38.20a,a (0.22) 58.43b,a (2.06) 57.13b,a (0.02) 78.81c,a (1.08) 76.67c,b (0.18) 102.18d,a (2.12) 98.75d,b (1.00) 91.56e,a (1.38) 91.22e,a (0.66) 87.41f,a (1.40) 88.69f,a (1.20) 40.44a,a (0.43) 39.14a,a (0.08)

95.0a,a (5.7) 103.5a,a (11.9) 44.5b,a (2.2) 54.8b,a (3.0) 255.8c,a (51.4) 134.5a,b (12.7) 97.7a,a (1.2) 93.8a,a (6.4) 123.3a,a (6.3) 86.2a,b (4.1) 236.4c,a (42.0) 116.8a,b (0.5) 75.9d,a (4.4) 87.8a,a (4.0)

0.62a,a (0.08) 0.52a,a (0.21) 1.49b,a (0.10) 1.71b,b (0.13) 0.45a,a (0.08) 0.60a,a (0.09) 1.21c,a (0.15) 1.40c,b (0.19) 0.98d,a (0.18) 1.50c,b (0.15) 0.84d,a (0.21) 1.87b,b (0.26) 1.22c,a (0.05) 1.21c,a (0.11)

a

The solid mass fractions in the 2% organogels at both cooling rates are included. In the table, superscripts a−g: for the same thermal parameter and cooling rate values with the same first letter indicate no significant difference between gelators. For the same gelator and thermal parameter values, the same second letter indicates no significant effect of the cooling rate. Values with a different letter indicate a significant effect of cooling rate (P < 0.10). The asterisk: the reported thermal parameters correspond to the C polymorph of SA.

2.5. Polarized Light Microscopy. With the use of the same thermal treatment as for the calorimetry and rheological studies, polarized light microphotographs (PLM) of the organogels were obtained with an Olympus BX51microscope (Olympus Optical Company, Ltd., Tokyo, Japan) equipped with a color video camera (KP-D50; Hitachi Digital, Tokyo, Japan) and a Linkam TP94 heating/cooling stage (Linkam Scientific Instruments, Ltd., Surrey, England) connected to an LTS 350 temperature control station (Linkam Scientific Instruments, Ltd.) and a liquid nitrogen tank. PLMs of the organogels were obtained after they had been incubated at 25 °C for 75 min.

JC is the compliance at a given time (t = 300 s). The corresponding contribution of JSM to the gel recovery (% RSM) was calculated using JSM as J(t) in eq 3. 2.4. X-ray Analyses. Powder X-ray diffraction patterns of samples were obtained as previously described,11 using a Rigaku R-AXIS image plate system with Cu Kα X-rays (λ = 1.54 Å) from a Rigaku generator operating at 46 kV and 40 mA with the collimator at 0.5 mm. Data processing and analyses were performed using Materials Data JADE (version 5.0.35) XRD pattern processing software. The samples were sealed in 0.5 mm glass capillaries (W. Müller, Schönwalde, Germany), and diffraction data were collected for 2 h. For the organogels, an Xray diffractometer (Panalytical, X́ Pert Pro; Almelo, Netherlands) with Cu Kα X-rays (λ = 1.54 Å), operating at 35 KV and 30 mA, was employed to record wide-angle X-ray diffraction patterns (WAXS). Angular scans from 3° to 40° were obtained using a step size of 0.016° and a scan speed of 0.06°/s. The small-angle X-ray diffraction patterns (SAXS) were recorded on a Nanostar (Bruker Analytik; Rheinstetten, Germany) diffractometer equipped with a Cu Kα source (λ = 1.54 Å), operating at 45 kV and 30 mA at a distance between the sample and the detector of 106 cm and collecting data during 1800 s. The sample was kept at 25 °C during all measurements. Data processing and analyses were performed using X́ Pert Data collector version 2.0d.

3. RESULTS AND DISCUSSION 3.1. Thermal Parameters and Solid Content. The temperatures and heats of melting (i.e., TM and ΔHM) and solidification (i.e., TCr and ΔHCr) parameters are associated with the polarities of the molecules and the energies of the molecular interactions that establish the crystal structure of the gelators in their neat and organogel states. Thus, in LMOGs, the amount of heat released or adsorbed during self-assembly, ΔHCr, or melting, ΔHM, of the neat crystals or the organogel, as well as TM, depend on the energy of the noncovalent interactions involved during molecular self-assembly. On the other hand, TCr depends mainly on the polarity of the groups D

dx.doi.org/10.1021/la400809a | Langmuir XXXX, XXX, XXX−XXX

Langmuir

Article

Table 2. Comparison of Lattice Spacings (d) Determined by WAX and SAX and Calculated Extended Molecular Lengths (L) for the Gelators Studied in Neat Powders and 2% Organogels Developed Using a Cooling Rate of 1 °C/min and 20 °C/mina d (gel state)b (Å) gelator

d (powder state)b (Å)

SA

44.1, 14.9, 8.9, 6.3, 5.6, 4.9, 4.4, 3.7, 3.2

HSA OA

47.0, 15.7, 6.8, 5.9, 5.2, 4.6, 4.0, 3.8, 3.7 36.5, 17.0, 12.2, 10.2, 9.3, 8.4, 7.3, 6.3, 6.1, 5.1, 4.6, 4.5, 4.4, 4.2, 3.8, 3.6 48.5, 15.7, 4.5, 3.9, 3.8 28.5, 14.7, 10.8, 8.2, 4.7, 4.1, 3.9, 3.6 49.0, 23.8, 16.0, 12.2, 9.5, 8.8, 4.6, 4.1, 3.9, 3.5 42.0, 20.9, 13.9, 8.2, 4.1, 3.9, 3.0, 2.5

HOA PHOA OHOA CW

probable packing arrangement in the gel state

1 °C/min

20 °C/min

Lc(Å)

44.6, 40.1, 20.0, 13.3, 8.0, 4.14, 3.7 48.5, 15.5, 4.5, 3.96 36.5,12.2, 7.3, 4.5, 3.7, 3.3

40.1, 20.3, 13.5, 8.0, 4.13, 3.7 47.5, 15.5, 4.5, 3.95 36.5, 12.1, 7.3, 4.5, 3.7, 3.3 48.5, 15.7, 4.5, 3.9, 3.8 29.2, 19.5, 14.0, 4.0 48.5, 15.0, 4.4, 3.9 43.3, 4.1, 3.7

25.3

bilayer

25.6 25.8

bilayer bilayer

26.4 31.1 50.3 42.5d

bilayer monolayer monolayer monolayer

48.5,15.7, 4.5, 3.9, 3.8 28.5, 14.3, 4.1, 3.9 48.5,15.0, 4.4, 3.9 43.3, 4.1, 3.7

a All experimental values were determined by SAX or WAX at 25 °C. bThe d value was determined by WAX and SAX. cCalculated extended molecular lengths using Chem 3D Ultra 8 software (Cambridge Soft Corporation, USA) and adding the van der Waals radii of the terminal atoms. d Value for hentriacontane.

higher ΔHCr and ΔHM in both the neat and gel states (Table 1). SA and HSA interact through cyclic carboxylic acid dimer motifs. However, in contrast to SA, HSA has a secondary −OH group that increases its polarity and capability to interact efficiently with other molecules through strong hydrogen bonds. This packing arrangement limits the interaction between hydrocarbon chains of parallel HSA molecules. Although intermolecular interactions between parallel hydrocarbon chains of SA, mainly through London dispersion forces, are more efficient that in HSA,24,26 they are insufficient to compensate for the H-bonding in HSA: in the neat, and (particularly) the gel states, HSA had the higher TCr and TM but lower ΔHCr and ΔHM (Table 1; P < 0.025). When comparing the thermal parameters of neat SA and HSA, there is an additional factor to consider. SA has four different polymorphs: A, with a triclinic subcell, B, C, and E with an orthorhombic subcell organization. Form C is the most thermodynamically stable and is the only polymorph crystallizing directly from the melt.26,27 After the SA from the manufacturer was heated to 140 °C for 20 min for “memory erasing” and then cooled to −20 °C, the subsequent melting of the SA showed only the endotherm characteristic of the C polymorph.26−29 On the basis of these results, we concluded that the thermal parameters for neat SA reported in Table 1 are the ones corresponding to crystallization and melting of the C polymorph. CW is not a single component, and its constituent molecules21 are listed in Materials and Methods. Because nalkanes are the major component, the intermolecular stabilizing forces in neat CW are mainly from weaker London dispersion forces through lamellar packing. In contrast with the amides and acids studied, the thermal parameters of CW in its neat and gel states represent the crystallization and melting behavior of a complex mixture of components. Therefore, the behavior of the thermal parameters of CW has no direct comparison with the other gelators investigated. Neat CW was the only neat gelator, showing a TCr > TM, a behavior rarely observed in neat gelators and probably ascribable to its heterogeneous composition. Also, the SMF of CW gels was independent of the cooling rate of its solutions/sols (vide inf ra). The lattice spacings (d, Å) for the neat gelators and the 2% gels are summarized in Table 2. As established previously by Xray measurements, primary amides (OA and HOA) and carboxylic acids (SA and HSA) develop dimeric units stabilized

involved in these interactions, particularly when an LMOG is dissolved in a solution/sol. In some cases, the polymorphic transitions experienced by the gelators also affect the crystallization and melting thermal parameters. In the case of the acids and amides investigated here, the noncovalent interactions include strong hydrogen bonding (from complementary H-bond donating/accepting motifs), dipolar interactions between carboxylic or amide groups, hydrogen bonds between hydroxyl groups, and the weaker London dispersion forces along the parallel hydrocarbon chains. Overall, amides are the most polar of the LMOGs; their large dipole moments result in strong attractive intermolecular forces, and they are able to develop hydrogen bonds on both the oxygen and nitrogen atoms of the amide group. As a result, HOA, PHOA, and OHOA have stronger intermolecular interactions than HSA, their carboxylic acid analogue, in both the neat and gel states (Table 1). Additionally, primary amides have another hydrogen atom attached to the nitrogen atom than do the secondary amides, PHOA and OHOA. Consequently, OA and HOA can develop more extensive H bonding with other molecules. Furthermore, the secondary 12hydroxyl group in HOA increases the molecular polarity and allows the crystal structure to grow along the secondary axis through hydrogen bonds in a similar organization as in HSA (i.e., between parallel but partially interdigitated HOA molecules).24 The result is that HOA in its neat and gel states had the highest TCr and TM of the amides studied (Table 1; P < 0.05). Secondary amides like PHOA and OHOA developed a lamellar molecular packaging, stabilized by dipole−dipole interactions between opposing −CONH− and −OH groups, but mainly through London dispersion forces along the parallel hydrocarbon chains of 12-hydroxyoctadecanamide and the noctadecyl chain, in OHOA, or the n-propyl chain in PHOA.11 As a result, ΔHCr and ΔHM for secondary amides are higher on the molar basis than for primary amides (P < 0.05, Table 1), particularly for OHOA. This result shows the greater involvement of London forces in stabilizing the crystal structure. Using novel L-lysine-based compounds, Suzuki et al.25 observed that under the same hydrogen bonding (i.e., polar) conditions, the self-assembling ability of their gelators depended acutely on the alkyl chain length (i.e., London forces). A similar situation occurs when comparing PHOA and OHOA, where the longer n-octadecyl chain in OHOA in comparison with the n-propyl chain in PHOA resulted in E

dx.doi.org/10.1021/la400809a | Langmuir XXXX, XXX, XXX−XXX

Langmuir

Article

required molecular packing for nucleation. We observed this trend for all of the gelators except HOA, the gelator with the highest polarity and, therefore, the highest saturation temperature in HOSFO (Table 1). Independent of the cooling rate, HOA exhibited the highest TCr of the gelators studied (Table 1). In the hydrophobic environment of the vegetable oil, the high polarity of −CONH2 and −OH groups should drive the self-assembly of HOA. At the low cooling rate, supercooling changes at a lower rate than at 20 °C/min, and thus, gelator molecules have enough time and greater selectivity to develop self-assembled structures than at the high cooling rate. Except the gelator with lowest overall polarity, SA, the thermal parameters associated with the energy of the molecular interactions that establish the crystal structure (ΔHCr and ΔHM) were higher in the organogels formed at 1 °C/min. Nevertheless, the cooling rate effect on ΔHCr and ΔHM was statistically significant just for OA and the secondary amides (P < 0.05; Table 1), the gelator molecules where both polar interactions and London dispersion forces are involved in the self-assembly process.11 At 1 °C/min, OHOA form the organogels with the highest ΔHCr and ΔHM followed by OA > PHOA > HOA = SA > HSA. At 20 °C/min, OHOA, PHOA, and OA showed lower thermal values and ΔHCr and ΔHM were practically the same for most LMOGs, with HSA showing the lowest values (Table 1). In HOA and HSA, the presence of a 12-hydroxyl group provides efficient secondary hydrogen-bonding interactions in partially interdigitated layers.24 Consequently, ΔHCr and ΔHM of HOA and HSA in HOSFO are independent of the cooling rate. SA showed the lowest TCr at both cooling rates and no perceptible dependence of ΔHCr and ΔHM on the cooling rate because of its high solubility in HOSFO. Note that the gelator thermal parameters are not associated directly with gelation efficiency; this property, not measured in the present study, is defined as the minimum LMOG concentration that leads to gels in a given solvent, usually at room temperature. Independent of the type of gelator, higher SMF was obtained at the 20 °C/min than at the 1 °C/min cooling rate (P < 0.01; Table 1) in all of the LMOGs except SA and OA, which lack a 12-hydroxyl group. These two gelators were the most soluble in HOSFO and, thus, had the lowest solid contents in their organogels (Table 1). The solid content of CW organogels also showed no cooling rate dependence, but as previously noted, its complex molecular composition precludes direct comparisons between its gelation results and those of the other LMOGs. In a HSA−canola oil system, Rogers and Marangoni31 showed that the constant rates of nucleation and crystal growth are smaller at cooling rates lower than 5 °C/min than at higher rates. The authors proposed that a time-dependent thermodynamic force (i.e., time of exposition under supercooling conditions) resulted in higher rates of nucleation and crystal growth at cooling rates faster than 5 °C/min. At slower cooling rates, the rates of crystallization showed only a time-dependent behavior.31 These conditions seem to apply to the derivatives of HSA investigated here as well: gels developed at 20 °C/min had higher solid contents than those formed at 1 °C/min. Comparisons between LMOGs with and without a 12-hydroxyl group indicate, as expected, that the presence of the polar group increases the solid content in the HOSFO gels, particularly in those developed at 20 °C/min (P < 0.025; Table 1). Using the same LMOGs, Mallia et al.11 observed similar results when measuring their gelation temperatures and critical gelation concentrations in solvents with different degrees of polarity.

by dipole−dipole longitudinal interactions between opposing −CONH2 groups11or H bonding between carboxylic acid groups.24 In these cases, the gelator molecules are in stacked lamellar structures with thicknesses nearly twice an extended molecular length in both their neat and gel phases.11,24 Our SAXS results agree with that organization. The d values from SAXS for the neat solids and gels formed at both cooling rates for SA (40.1−44.6 Å), HSA (47−48.5 Å), OA (36.5 Å), and HOA (48.5 Å) are slightly shorter than twice the corresponding calculated extended molecular lengths (Table 2). Neat OA was the only gelator that showed two short-angle spacings in the neat solid, 50.2 Å and 36.5 Å. However, after melting to 140 °C for 20 min then cooling to −20 °C and again melting, OA showed only the d value at 36.5 Å. It seems that OA directly from the synthesis and purification process crystallizes in two different polymorphs but just as one polymorph from its melt. The neat solid phases of the secondary amides had only one short-angle diffraction peak with d values of 49.0 Å for OHOA and 28.5 Å for PHOA, which correspond to approximately the calculated extended lengths of the gelator molecules (Table 2). Independent of the cooling rate, the PHOA and OHOA organogels also showed one short-angle diffraction peak with d values consistent with a monolayer lamellar molecular packaging (Table 2).11 The additional d values obtained by WAXS for the acids, and primary and secondary amides correspond to higher order diffractions in the expected progression for a lamella packing arrangement. They also contained characteristic d values for orthorhombic and/or triclinic subcell structures at approximately 3.8 and 4.2 Å and 4.6 Å, respectively (Table 2). These subcell crystal structures are present in the neat and in the organogel state, independent of the cooling rate. In spite of its compositional complexity, the SAXS diffractograms of CW in their neat and gel phases, formed at both cooling rates, contained one major diffraction line corresponding to the d values of 42.0 and 43.3 Å, respectively (Table 2). The interplanar spacing determined by SAXS in CW in its neat and gel states, 42.5 Å, is consistent with the extended molecular length of n-hentriacontane, the major n-alkane present in this gelator. These results and the ones obtained by WAXS (i.e., peaks at 4.1 Å and 3.7−3.9 Å) indicate that nhentriacontane (and perhaps the other components) of CW crystallized in the neat and organogel states in an orthorhombic perpendicular subcell packing, consistent with results obtained by Chopin et al.30 Because the X-ray results showed that cooling rate does not affect the crystal structure achieved by the LMOGs during gelation, it alters only the kinetics of crystallization. Independent of the cooling rate, the thermal parameters in the 2% organogels (Table 1) followed similar behavior as those of the one observed by the neat solids. The noncovalent interactions involved in the self-assembly of the LMOGs can be used to describe most of their thermal behavior in the organogels. However, LMOG solubility in the low-polarity solvent, HOSFO, and the cooling rate are additional factors that must be considered to understand the thermal behavior of the gelator solutions/sols and their organogels. Overall, the TCr of the 2% gelator solutions/sols cooled at 1 °C/min were higher than those obtained upon cooling at 20 °C/min (P < 0.10; Table 1). This phenomenon can be explained considering that gelator molecules have less time to organize in a solid phase at the higher cooling rate. As a result, the gelator molecules require a lower TCr temperature to achieve the F

dx.doi.org/10.1021/la400809a | Langmuir XXXX, XXX, XXX−XXX

Langmuir

Article

Table 3. Yield Stress (σ*), Mean G′, and Burger’s Parameters for the 2% Organogels Developed at 1 °C/min and 20 °C/mina Burger’s parameters gelator

cooling rate (°C/min)

SA

1 20 1 20 1 20 1 20 1 20 1 20 1 20

HSA OA HOA PHOA OHOA CW

σ* (Pa) a

G′ (Pa)h ×104

1.04 (0.00) 0.00 (0.00)a 65.06 (14.23)b 49.73 (1.34)b 1.29 (0.08)c 18.08 (12.60)c 172.04 (8.68)d 174.90 (34.5)d 200.79 (2.13)e 88.10 (5.20)e 223.48 (36.65)f 182.20 (18.60)d 43.54 (3.71)g 26.50 (7.20)c

G0 (Pa) ×104

η0 (Pa s) ×106

G1 (Pa) ×105

η1 (Pa s) ×105

R2 i

− − 0.33 0.49 − − 1.70 6.35 2.23 1.55 4.28 8.66 3.37 3.39

− − 0.68 0.90 − − 4.27 22.26 11.62 6.60 7.28 10.28 10.33 6.47

− − 0.11 0.16 − − 1.51 2.13 1.43 1.51 1.14 1.92 1.84 1.17

− − 0.24 0.15 − − 1.93 11.00 4.69 3.41 1.77 5.71 5.86 3.55

− − 0.999 0.957 − − 0.987 0.979 0.993 0.986 0.990 0.991 0.999 0.991

a

0.03 (0.03) 0.001 (0.001)a 0.41 (0.17)b 0.74 (0.14)b 0.13 (0.01)c 0.60 (0.12)c 4.26 (0.61)d 6.70 (0.20)d 7.52 (0.30)e 1.56 (0.34)e 18.70 (1.13)f 15.75 (0.49)f 2.84 (0.35)g 2.17 (0.32)g

The Burger’s parameters were obtained by fitting the experimental creep profile to the Burger’s equation. In the table, superscripts a−g: σ* and G′ values at the same cooling rate with the same subscript letter indicate no significant difference. Values with a different subscript letter indicate a significant difference (P < 0.05). Superscript h: obtained from the average of the G′ values of the last 30 min. The mean and standard deviation were calculated from the corresponding averaged G′ values of the independent determinations. Superscript i: determination coefficient of the model. a

Table 4. JMAX, JSM, J∞, JkV, the Instantaneous Percentage of Recovery (% RSM), and the Percentage of Recovery at t = 300 s (% Rt) for the 2% (w/w) Organogels Developed at 25°C and a Cooling Rate of 1 °C/min and 20 °C/min gelator SA HSA OA HOA PHOA OHOA CW a

cooling rate (°C/min)

JMAX (1/Pa) × 10−5

JSM (1/Pa) × 10−5

J∞ (1/Pa) × 10 −5

JkV (1/Pa) × 10−5

R2

% RSM

% Rt

mean square error × 10−12

1 20 1 20 1 20 1 20 1 20 1 20 1 20

−a −a 47.9 32.86 −b −b 7.84 2.30 5.68 7.49 3.98 2.24 4.06 4.68

− − 32.67 23.45 − − 5.53 1.80 4.33 6.34 2.36 1.27 3.38 3.27

− − 20.17 14.5 − − 3.34 0.69 1.68 2.22 1.89 1.26 0.98 1.98

− − −4.95 −5.1 − − −1.03 −0.19 −0.33 −0.59 −0.27 −0.30 −0.30 −0.57

− − 0.9312 0.9551 − − 0.9174 0.9835 0.9860 0.9952 0.9777 0.9763 0.9084 0.9900

− − 31.8 28.6 − − 29.5 21.7 23.7 20.6 40.6 43.1 16.7 30.2

− − 83.6 90.0 − − 85.2 107.3 87.1 92.7 72.0 79.2 93.3 91.2

− − 28276 25503 − − 245.80 0.0852 0.7188 2.3235 0.2289 0.4834 0.4472 2.6524

JMAX at 1 °C/min = 485.43 1/Pa; JMAX at 20 °C/min = 720.29 1/Pa. bJMAX at 1 °C/min = 273.94 1/Pa; JMAX at 20 °C/min = 351.80 1/Pa.

3.2. Rheological Properties of the Organogels and Their Relationship with Solid Phase Content and Molecular Structure. The protocol for preparing the organogels for rheological studies was the same as used previously.4,11 The corresponding G′ and σ* values and the parameters describing the creep and recovery profiles for the different organogels at each cooling rate are shown in the Tables 3 and 4. The G′ profiles of the organogels at 25 °C were relatively constant over time, particularly after 30 min and at a cooling rate of 20 °C/min (data not shown). Thus, the average of the G′ values after 30 min was used to calculate the mean and standard deviations reported in Table 3. The creep and recovery profiles of the organogels had the same general behavior (Figure 3 of the Supporting Information). Thus, the creep profiles of the 2% organogels showed progressive deformation of the LMOG network under constant stress. Upon release of the stress, the strain of the samples exhibited an instantaneous recovery followed by a progressive decrease of deformation until attaining a constant

In contrast to the effect of the 12-hydroxyl group, the substitution of a primary amide group for the carboxylic acid of SA did not appreciably change the solid content. However, gels with HSA had higher solid contents than those of the corresponding primary amide, HOA, at both cooling rates and at 1 °C/min, even higher than for the secondary amides (P < 0.10; Table 1). The TCr, solid content, the melting parameters of the gels with OHOA and PHOA, HOA containing a N-alkyl group and, thus, only one N−H amide bond, are different from those of the primary amide. PHOA and OHOA gels had lower solid content (i.e., higher solubility in HOSFO) than HOA at 1 °C/min and similar solid contents at 20 °C/min. Thus, based on the results of Rogers and Marangoni31 and the solid content present in the 2% organogels, we concluded that the presence of a 12-hydroxyl group in the gelator molecules makes their crystallization kinetics in HOSFO dependent on the cooling rate applied to the solutions/sols, and the dependence is more evident in the secondary amides. G

dx.doi.org/10.1021/la400809a | Langmuir XXXX, XXX, XXX−XXX

Langmuir

Article

Figure 3. Polarized light microphotographs of organogels formed by (A and B) 2% SA, (C and D) 2% HSA, and (E and F) 2% OA at 25 °C, using cooling rates of (A, C, and E) 1 °C/min and (B, D, and F) 20 °C/min.

value. The SA and OA gels were the weakest, having σ* lower than the force applied for creep measurements. SA and OA gels formed at both cooling rates showed a creeping profile but suffered permanent deformation with no recovery (Table 4). In accordance with Steffe,32 if samples follow the Burger model and the measurements are made within the LVR, as in the present investigation, 1/G0 from the creep profile must be equal to JSM from the recovery profile. This behavior was followed by organogels formed with HSA, CW, and the amide gelators with a 12-hydroxyl group. At each cooling rate, these organogels exhibited a significant inverse linear relationship

between the log(JMAX) and the corresponding log(G′) and log(G0) (R2 > 0.80; P = 0.05). This behavior is reasonable, considering that both G′ and G0 are parameters directly associated with the resistance of the organogel structure to deformation; the higher is JMAX, the lower is the gel resistance to deformation. The Figures 3 and 4 show microphotographs of the 2% organogels. Additional microphotographs at higher magnification are included in Figure 1 of the Supporting Information. The SA, OA, and CW crystallized as platelets, HSA as fibers, and HOA, PHOA, and OHOA as spherulites with a fibrillar H

dx.doi.org/10.1021/la400809a | Langmuir XXXX, XXX, XXX−XXX

Langmuir

Article

Figure 4. Polarized light microphotographs of organogels formed by (A and B) 2% HOA, (C and D) 2% PHOA, (E and F) 2% OHOA, and (G and H) 2% CW at 25 °C, using cooling rates of (A, C, E, and G) 1 °C/min and (B, D, F, and H) 20 °C/min.

Some gelators are known to be polymorphous.33−36 Lam et al. found that at cooling rates 7 °C/min. The authors showed that at the higher cooling rates, 12-HSA monomers do not effectively dimerize through their carboxyl groups before being incorporated into the crystal lattice and, thus, crystal imperfections impede linear epitaxial crystal growth and enhance the frequency of branched fibers.31,38 Within this framework,

organization (Figures 3 and 4 and Figure 1 of the Supporting Information). At the higher cooling rate, the crystals were smaller. This behavior is common to many gel systems6,17,37,38 as well as to others in which bulk-separated crystals are observed33−36 and is a consequence of nucleating rates being accelerated at the expense of growth rates.14,27,38 In contrast, at low cooling rates, the aggregating molecules have sufficient time to diffuse and to add selectively to a nucleated crystallite surface. Additionally, crystal growth of already nucleated species is favored over additional nucleation events and fewer, but larger crystals are formed.

38

I

dx.doi.org/10.1021/la400809a | Langmuir XXXX, XXX, XXX−XXX

Langmuir

Article

exhibited higher G′ and similar σ* than the gels formed at 1 °C/min (having lower solid contents), while PHOA and OHOA gels formed at 1 °C/min had higher G′ and σ* values than the gels developed at 20 °C/min (Table 3). This, despite PHOA and OHOA gels formed at 1 °C/min, had lower solid content than those formed at 20 °C/min. In the gels formed using the 1 °C/min cooling rate, G′ and σ* followed the order OHOA > PHOA > HOA, while at 20 °C/min, the order for G′ was OHOA > HOA > PHOA and for σ* it was OHOA = HOA > PHOA (P < 0.05; Table 3). At both cooling rates, the OHOA gels had the highest G′ and σ*, and therefore the lowest JMAX. Additionally, the OHOA gels showed the highest instantaneous recovery (% RSM : 41−43%) and the lowest extended thixotropic behavior (% Rt = 72−79%) (Table 4). This rheological behavior is outstanding, particularly given the low solid content of the OHOA gels formed at 1 °C/min (Table 1). Evidently, the gelator microstructure has a more important role in determining the viscoelastic behavior of the organogels than their solid content. As mentioned, CW crystallized as platelets, but their size was smaller than those obtained with SA and OA. Their 3D organization of interacting crystals (Figure 4, panels G and H) is independent of the cooling rate. Their gels had higher G′, lower σ*, and higher JMAX than HSA gels (Tables 3 and 4). Although CW organogels had just 16% to 30% instant recovery, they showed more than 90% of extended recovery (% Rt) (Table 4). This rheological behavior occurred in spite of the lower mass of crystals present in CW organogels than in HSA organogels (P < 0.05; Table 1). The high number of small crystals and the high extent of crystal−crystal interaction, stabilized mainly through London dispersion forces, imbue remarkable rheological properties to the CW organogels developed under static or even shearing conditions.4,21,30 3.3. Models that Describe the Rheological Behavior of the Organogels. The solid content is a primary factor in determining the rheological properties of viscoelastic systems. Usually, the solid mass fraction is related to the elastic properties of fats and crystallized vegetable oils. Narine and Marangoni33 showed that for a particular fat system, the slope of the log(G′) − log(solid volume fraction) plot is associated with the fractal dimension of the crystal network. Within this framework, we found that, independent of the cooling rate, the log(G′) of the organogels studied had a significant linear relationship (R2 = 0.4076; P < 0.001) with the corresponding log(SMF). This relationship has a slope of 4.0, corresponding to a fractal dimension of 2.75. Analyses as a function of cooling rate also resulted in acceptable linear relationships, with R2 = 0.2465 (P < 0.08) and R2 = 0.6307 (P < 0.001) providing relative fractal dimensions of 2.65 and 2.82 for 1 and 20 °C/ min cooling rates, respectively. However, the model of Narine and Marangoni33 was developed to determine the fractal dimension for individual fat systems. Through this methodology, a particular crystallized fat system is serially diluted determining the corresponding solid content and G′ for each dilution. The fractal dimension for the particular crystallized fat system is then calculated from the slope of the log(G′) − log(SMF). Therefore, the physical meaning of the fractal dimension calculated using the G′ and solid mass fractions of all type of organogels studied is not clear. As previously shown (Figures 3 and 4), the gelators investigated formed crystals with different shape and size that should develop 3D crystal structures with a different fractal dimension. Probably, this is

Wang et al.17 showed that the proportion between transient (i.e., entanglement of fibers) and permanent (i.e., branching of fibers) junction zones in a gel network determine the rheological properties of the self-assembled crystal network. Unfortunately, with the magnifications available on our optical microscopes, it was not possible to observe differences in fiber branching or microtopology of the crystal networks associated with the cooling rate effects (Figures 3 and 4). Overall, cooling rate did not affect G′ or σ* of CW, HSA, SA, and OA organogels. At both cooling rates, gels of SA (Figure 3, panels A and B, and Figure 1, panels A and B, of the Supporting Information) and OA (Figure 3, panels E and F, and Figure 1, panels E and F, of the Supporting Information) exhibited large platelets, the lowest solid contents, and the weakest viscoelastic properties (Tables 3 and 4). Gelators with a 12-hydroxyl and a primary or secondary amide group, HOA, PHOA, and OHOA, crystallized as spherulites and were attended by a significant cooling rate effect on elasticity (P < 0.01). In contrast to SA and OA (both lacking a 12-hydroxyl group), the gel networks of HSA (Figure 3, panels C and D) and HOA (Figure 4, panels A and B) in HOSFO were fibers and fibrillar spherulites, respectively. With the use of racemic hydroxy stearic acids in which the hydroxyl group was positioned at different points along the alkyl chain, Abraham et al.24 established that isomers with the hydroxyl beyond position 6 promoted crystal growth along the secondary axis of the molecule, leading to HSA crystallization as fibers. These HSA molecules can form effectively both carboxylic cyclic dimers and H bonds along the axis of the fibers via the secondary hydroxyl groups.24 The higher rheological properties of HSA, and particularly of HOA, resulted from the higher solid content and its higher degree of fiber interweaving. For example, the microstructure of HOA gels formed at 20 °C/min showed significant fiber interpenetration between vicinal spherulites (Figure 4B), which resulted in organogels with lower JMAX (higher G′) and higher % Rt than those achieved by the gels formed at 1 °C/min (Table 3). The substitution of an amide group in place of the carboxyl of HSA altered its crystallization kinetics and mechanism of crystal growth. Independent of the cooling rate, HOA, PHOA, and OHOA crystallized as spherulites (Figure 4, panels A−F, and Figure 1, panels G−L, of the Supporting Information), developing gels with higher G′ and σ* and, therefore, lower JMAX than the HSA gels (Tables 3 and 4). This is despite the fact that the HSA gels have a higher crystal mass than the HOA, PHOA, and OHOA gels, particularly in gels formed at 1 °C/ min (Table 1). At 20 °C/min, HOA, PHOA, and OHOA crystallized as smaller spherulites (Figure 4, panels B, D, and F) than those obtained at 1 °C/min (Figure 4, panels A, C, and E, and Figure 1, panels G, I, and K, of the Supporting Information). However, independent of the cooling rate, the spherulites formed by PHOA and OHOA were smaller than those formed by HOA. In polymer systems, such as poly(trimethylene terephalate)39 and poly(ethylene succinate),40 spherulite size decreases with increasing molecular weight because growth rate decreases as molecular weight increases. The overall result is that HOA gels formed at 20 °C/ min showed higher fiber interpenetration between vicinal spherulites (Figure 4A) than the gels formed at 1 °C/min (Figure 4B). Given the smaller size of the spherulites the opposite occurred with the PHOA (Figure 4, panels C and D) and OHOA (Figure 4, panels E and F) gels. As a result, HOA gels formed at 20 °C/min (and having higher solid contents) J

dx.doi.org/10.1021/la400809a | Langmuir XXXX, XXX, XXX−XXX

Langmuir

Article

why the determination coefficients (R2) of the log(G′) − log(SMF) relationships are relatively low. Several studies15,17,30,37,38 have established relationships between the microtopology and the rheology of organogels. Nevertheless, the combined effect of microstructure and solid content on organogels rheology has not been addressed. Independent of the cooling rate or as a function of the cooling rate, we could not find a particular relationship between G′ or σ* and the solid content of the organogels. This observation applies even when the SA and OA data, the gelators that formed the gels with the lowest crystal mass and weakest rheological properties, were excluded from the analysis. The same behavior was observed even when we excluded the CW data (data not shown). In an attempt to establish possible relationships that describe the viscoelastic properties of the organogels as a function of the solid content and parameters associated with gelator molecular structure, we used the forward stepwise multiple regression analysis available in the multiple regression module of STATISTICA version 9.0 (StatSoft Inc.). We used as predictor variables of the organogels rheological behavior, the parameters ΔH M , ΔH Cr , the SMF, and the corresponding linear interactions of the SMF with the thermal parameters (i.e., ΔHM*SMF, ΔHCr*SMF, etc.). The forward stepwise regression analysis involves starting with no variables in the regression model and testing the addition of each variable based on statistical criteria (i.e., % R2, % of variation in G′, or σ* fitted by the model), adding the variable (if any) that improves the model the most, and repeating this process until no further addition of a predictor variable improves the model.41 Essentially, the statistical procedure establishes which independent variable is the best predictor, the second best predictor, etc. of a particular property of a system. With the use of this approach, models that explain 53% (P < 0.02) and 39% (P < 0.07) of the variation in G′ and σ*, respectively, were obtained by including ΔHM and the solid content as significant variables. We performed further analyses excluding the SA and OA data from the model building procedure. The models obtained fitted 93% (P < 0.001) of the variation in organogel elasticity (eq 5) and 53% (P < 0.02) of the variation in organogel resistance for plastic deformation (eq 6). Because SA and OA were excluded from the model development procedure, eqs 5 and 6 apply only to organogels with SMF > 0.006 and G′ > 10 Pa. Note that the fits are independent of the cooling rate and shape of the crystals that form the 3D network in the organogels (i.e., microplatelets, fibers, fibrillar spherulites). The model for σ* requires the inclusion of additional variables to describe better the variation of this rheological parameter. Thus, using the same statistical procedure, we tried to develop models for σ* that provided higher % R2 at each cooling rate. However, the equations obtained did not provide increases in % R2 (data not shown). Evidently, the solid content did not have a significant effect on σ*; microstructural factors must be more relevant in determining σ*. Additional microstructural parameters of the crystal network (e.g., correlation lengths) must be investigated to evaluate their effect in determining σ*.

As previously stated, the type of functional groups in the gelator structure, their position in the molecule, and the extent of involvement of the weaker London dispersion forces in the molecular assemblies determine the amount of heat released during crystallization (i.e., ΔHCr) or absorbed during melting (i.e., ΔHM) of the gelator network in an organogel. Additionally, as established in studies involving the influence of the position of the hydroxyl group on the self-assembly of racemic HSA,24 these factors also have a direct effect on the packing arrangement of the molecules, the shape of the crystals and, therefore, the microstuctural organization of the crystal network in the gel phase. The eqs 5 and 6 are predictive models and thus they do not have a direct correspondence to the underlying mechanisms that establish the organogel rheology. Nevertheless, the presence of ΔHM in these equations might be associated with the effect of a functional group (e.g., inclusion of a primary amide group in place of a carboxyl group in 12-HSA or a n-alkyl chain in HOA) in the gelator microstructure and, subsequently, on the rheological properties of the organogels. Additionally, because the ΔHM of HOA, PHOA, and OHOA showed significant cooling rate dependence (Table 1), the crystallization conditions modify their microstructure. The inclusion of ΔHM in the equations accounts for these effects on the behavior of G′ and σ*. The presence of the variable (SMF)(ΔHM) in eq 5 indicates that the effect of the crystal mass on G′ depends on the microstructure developed by the gelator. The interaction effect between these variables on the viscoelastic parameters of the organogels was observed in a three-dimensional plot that describes the change of G′ as a function of the SMF and ΔHM (see Figure 2 of the Supporting Information). Furthermore, the variable SMF in eq 5 accounts for effect of the cooling rate on organogel elasticity.

4. CONCLUSIONS LMOGs without a 12-hydroxyl group, SA, and OA, crystallized in the vegetable oil as platelets, developing gels with the lowest solid content and G′ and σ* values; these parameters were independent of the cooling rate. The results obtained in this investigation and those previously reported by Rogers and Marangoni31 showed that the presence of a 12-hydroxyl group in LMOG makes its crystallization kinetics cooling ratedependent. Thus, at 20 °C/min HSA and its amide derivatives, HOA, PHOA, and OHOA, formed gels with higher solid content. Additionally, the 12-hydroxyl group increases the molecular polarity and the number of potential hydrogenbonding sites of the LMOG. Thus, in contrast to SA, the HSA crystallizes as fibers forming organogels with higher solid content than SA, particularly at 20 °C/min. Amides can form hydrogen bonds on both the oxygen and the nitrogen of the amide functional group. Therefore, the addition of a primary or a secondary amide functional group to HSA resulted in LMOGs with higher numbers of potential hydrogen-bonding sites. Thus, in contrast to HSA, primary and secondary amides crystallized as fibrillar spherulites at both cooling rates. The bonding of an alkyl group to the amide group of HOA decreases the potential number of hydrogenbonding sites but increases the influence of the London forces for intermolecular interaction as the length of the alkyl group increases. Consequently, the cooling rate effect on the thermal parameters, solid content, microstructure, and therefore, in the organogels, the rheology of OHOA and PHOA is different from that exhibited by HOA. Thus, at 1 °C/min, HOA

G′ = −95313.3 + 83612.3(SMF)(ΔHM) + 484.8(ΔHM) (5)

σ* = 18.13 + 1.03(ΔHM)

(6) K

dx.doi.org/10.1021/la400809a | Langmuir XXXX, XXX, XXX−XXX

Langmuir



developed gels formed by larger spherulites than at 20 °C/min. However, at the higher cooling rate the spherulites exhibited higher fiber interpenetration among vicinal spherulites, and therefore higher rheological properties than gels developed at 1 °C/min. In contrast, PHOA and OHOA developed smaller spherulites than HOA and thus the opposite behavior occurred. At 1 °C/min, the microstructure of PHOA and OHOA gels showed a higher extent of fiber interpenetration among vicinal spherulites and, therefore, G′ and σ* were higher than at 20 °C/min. This, despite the gels being developed at 1 °C/min, had lower solid content than those formed at 20 °C/min. Independent of the cooling rate applied, HSA and the primary and secondary amide gels showed thixotropic behavior with different extents of instantaneous (∼20% to ∼43%) and extended recovery (∼72% to ∼100%) after applying a constant strain. Of all these LMOGs, OHOA formed the gels with the highest G′ and σ* and the highest instantaneous recovery also. These observations were particularly evident at the lower cooling rate. Then, the gel microstructure has a more important role determining the organogels’ rheology than the solid content. The fitting models developed for G′ and σ* supported this argument. The elasticity of organogels depended on microstructural factors, which in turn depends on the LMOG structure and the cooling rate. Because the solid content did not have a significant effect describing the variability in σ* among the organogels, microstructural factors must be even more relevant in determining σ* of the gels. The results obtained for HOA, PHOA, and OHOA at both cooling rates suggests that the rheological properties of amide’s organogels depend on a balance between hydrogen-bonding sites and the alkyl chain length. This evidencing the importance of the London dispersion forces in establishing the mechanical strength of the organogels as the alkyl chain length of the LMOG increases. However, this might be a compound effect additionally associated with the inverse effect that molecular weight of LMOGs has on the crystal growth rate and its consequence on fiber interpenetration among vicinal spherulites. We found that independent of the cooling rates or as a function of the cooling rate, the log(G′) of the organogels had a significant linear relationship with their corresponding log(SMF). The corresponding slopes provided a relative fractal dimension between 2.65 and 2.82. However, the model used to calculate the fractal dimension was developed to calculate the fractal dimension for individual fat systems.33 Thus, we must take the physical meaning of this fractal dimension with caution. As a final note, CW organogels had higher elasticity than HSA gels but lower than the gels formed by the primary and secondary amides tested. Despite its lower σ*, the CW gels had a similar or even higher thixotropic behavior than HSA, HOA, PHOA, and OHOA gels. These remarkable rheological properties resulted from the microstructural organization of CW organogels, where the vegetable oil was structured by a high number of small microplatelets with a high extent of crystal−crystal interaction. This microstructural organization, stabilized mainly through London dispersion forces, ought to provide high oil-binding capacity to these organogels. We must undertake this study to establish the functionality of this edible LMOG as an alternative to reduce or even eliminate the use of saturated and trans fatty acids. Reducing the levels of such fatty acids will result in the development of healthier food products for the consumer.

Article

ASSOCIATED CONTENT

S Supporting Information *

Additional microphotographs for the 2% organogels, graphs showing the fitting of the organogels elasticity as a function of solid content and heat of melting, and organogels creep and recovery profiles. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

́ *Address: Facultad de Ciencias Quimicas-CIEP, Av. Dr. Manuel Nava 6 Zona Universitaria San Luis Potosi,́ SLP 78210, México. E-mail: [email protected]. Tel: 52-444-8262460, ext 101. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The investigation was supported by Grant CB-2012-01/177335 from CONACYT. The technical support from Concepcion Maza-Moheno and Elizabeth Garcia-Leos is greatly appreciated. V.A.M. and R.G.W. thank the U.S. National Science Foundation for Grants CHE-1147353 and CHE-0911089 that supported the portion of this research performed at Georgetown.



REFERENCES

(1) Abdallah, D. J.; Weiss, R. G. n-Alkanes gel n-alkanes (and many other organic liquids). Langmuir 2000, 16, 352−355. (2) Terech, P.; Weiss, R. G. How molecular mass gelators of organic liquids and the properties of their gels. Chem. Rev 1997, 97, 3133− 3159. (3) Bot, A.; Agterof, G. M. Structuring of edible oils by mixtures of γoryzanol with β-sitosterol or related phytosterols. J. Am. Oil Chem. Soc. 2006, 83, 513−521. (4) Toro-Vazquez, J. F.; Morales-Rueda, J. A.; Mallia, V. A.; Weiss, R. G. Relationship between molecular structure and thermo-mechanical properties of candelilla wax and amides derived from (R)-12hydroxystearic acid as gelators of safflower oil. Food Biophysics 2010, 5, 193−202. (5) Toro-Vazquez, J. F.; Morales-Rueda, J. A.; Dibildox-Alvarado, E.; Charó-Alonso, M. A.; Alonzo-Macías, M.; González-Chávez, M. M. Development of organogels with candelilla wax and safflower oil with high triolein content. J. Am. Oil Chem. Soc. 2007, 84, 989−1000. (6) Morales-Rueda, J. A.; Dibildox-Alvarado, E.; Charó-Alonso, M.; Weiss, R. G.; Toro-Vazquez, J. F. Thermo-mechanical properties of candelilla wax and dotriacontane organogels in safflower oil. Eur. J. Lipid Sci. Technol. 2009, 111, 207−215. (7) Morales, M. E.; Gallardo, V.; Clarés, B.; García, M. B.; Ruiz, M. A. Study and description of hydrogels and organogels as vehicles for cosmetic active ingredients. J. Cosmet. Sci. 2009, 60, 627−636. (8) Jibry, N.; Sarwar, T.; Murdan, S. Amphiphilogels as drug carriers: Effects of drug incorporation on the gel and on the active drug. J. Pharm. Pharmacol. 2006, 58, 187−194. (9) Vintiloiu, A.; Leroux, J. C. Organogels and their use in drug delivery: A review. J. Controlled Release 2008, 125, 179−192. (10) George, M.; Weiss, R. G. In Molecular Gels: Materials with SelfAssembled Fibrillar Networks; Weiss, R. G., Terech, P., Eds. Springer: Dordrecht, The Netherlands, 2006; Chapter 14. (11) Mallia, V. A.; George, M.; Blair, D. L.; Weiss, R. G. Robust organogels from nitrogen-containing derivatives of (R)-12-hydroxystearic acid as gelators: Comparisons with gels from stearic acid derivatives. Langmuir 2009, 25, 8615−8625. (12) Mallia, V. A.; Terech, P.; Weiss, R. G. Correlations of properties and structures at different length scales of hydro- and organo-gels L

dx.doi.org/10.1021/la400809a | Langmuir XXXX, XXX, XXX−XXX

Langmuir

Article

based on N-alkyl-(R)-12-hydroxyoctadecylammonium chlorides. J. Phys.Chem. B. 2011, 115, 12401−12414. (13) Baddeley, C.; Yan, Z.; King, G.; Woodward, P. M.; Badjié, J. D. Structure-function studies of modular aromatics that form molecular organogels. J. Org. Chem. 2007, 72, 7270−7278. (14) Tang, S.; Liu, X. Y.; Strom, C. S. Producing supramolecular functional materials based on fiber network reconstruction. Adv. Funct. Mater. 2009, 19, 2252−2259. (15) Li, J. L.; Liu, X. Y. Architecture of supramolecular soft functional materials: From understanding to mico-/nanoscale engineering. Adv. Funct. Mater. 2010, 20, 3196−3216. (16) Tu, W.; Li, Y.; Wang, T.; Liu, M.; Li, Z. Effect of solvent and molecular structure on the enhanced fluorescence and supramolecular chirality of Schiff bases in organogels. Acta Phys.-Chim. Sin. 2008, 24, 1535−1539. (17) Wang, R.; Liu, X. Y.; Xiong, J.; Li, J. Real-time observation of fiber network formation in molecular organogel: Supersaturation dependent microstructure and its related rheological property. J. Phys. Chem. B 2006, 110, 7275−7280. (18) Nippon Fine Chemicals. http://www.nipponseika.co.jp/eng/ industry/amides.htm (accessed August 28, 2012). (19) Nippon Kasei Chemical Company Ltd. http://www.nkchemical. co.jp/ver_03/english/product_e/fattyacidamid_e.html (accessed August 28, 2012). (20) Rudolf Group. http://www.rudolf.de/products/en/coproducer/fatty-acid-amide-softeners.htm (accessed August 28, 2012). (21) Alvarez-Mitre, F. M.; Morales-Rueda, J. A.; Dibildox-Alvarado, E.; Charó-Alonso, M. A.; Toro-Vazquez, J. F. Shearing as a variable to engineer the rheology of candelilla wax organogels. Food Res. Int. 2012, 49 (1), 580−587. (22) AOCS Official Method Cd 16b-93. Solid Fat Content (SFC) by Low-Resolution Nuclear Magnetic Resonance-The Direct Method. AOCS Methods 2009. http://search.aocs.org/methods/search_ methods_view_method.cfm?method=cd16b_93.pdf (accessed August 28, 2012). (23) Dolz, M.; Hernández, M. J.; Delegido, J. Creep and recovery experimental investigation of low oil content food emulsions. Food Hydrocolloids 2008, 22, 421−427. (24) Abraham, S.; Lan, Y.; Lam, R. S. H.; Grahame, D. A. S.; Kim, J. J. H.; Weiss, R. G.; Rogers, M. A. Influence of positional isomers on the macroscale and nanoscale architectures of aggregates of racemic hydroxyoctadecanoic acids in their molecular gel, dispersion, and solid states. Langmuir 2012, 28, 4955−4964. (25) Suzuki, M.; Nakajima, Y.; Yumoto, M.; Kimura, M.; Shirai, H.; Hanabusa, K. Effects of hydrogen bonding and van der waals interactions on organogelation using designed low-molecular-weight gelators and gel formation at room temperature. Langmuir 2003, 19, 8622−8624. (26) Sato, K.; Yoshimoto, N.; Suzuki, M.; Kobayashi, M.; Kaneko, F. Structure and transformation in polymorphism of petroselinic acid (cis-.omega.-12-octadecenoic acid). J. Phys. Chem. 1990, 94, 3180− 3185. (27) Kobayashi, M. In Polymorphism of Fats and Fatty Acids Crystallization; Garti, N., Sato, K., Eds., Marcel Dekker: New York, 1998; Chapter 4. (28) Beckmann, W.; Boistelle, R.; Sato, K. Solubility of the A, B, and C polymorphs of stearic acid in decane, methanol, and butanone. J. Chem. Eng. Data 1984, 29, 211−214. (29) Suzuki, M.; Ogaki, T.; Sato, K. Crystallization and transformation mechanisms of α, β and γ-Polymorphs of Ultra-Pure Oleic Acid. J. Am. Oil Chem. Soc. 1985, 62, 1600−1604. (30) Chopin-Doroteo, M.; Morales-Rueda, J. A.; Dibildox-Alvarado, E.; Charó-Alonso, M. A.; Toro-Vazquez, J. F. The shearing effect on the thermo-mechanical properties of candelilla wax and candelilla waxtripalmitin organogels. Food Biophysics 2011, 6, 359−376. (31) Rogers, M. A.; Marangoni, A. G. Non-isothermal nucleation and crystallization of 12-hydroxystearic acid in vegetable oils. Cryst. Growth Des. 2008, 8, 4596−4601.

(32) Steffe, J. F. Rheological Methods in Food Process Engineering, 2nd ed.; Freeman: East Lansing, 1996. (33) Narine, S. S.; Marangoni, A. G. Fractal nature of fat crystal networks. Phys. Rev. E 1999, 59, 1908−1920. (34) Litwinenko, J. W.; Rojas, A. M.; Gerschenson, L. N.; Marangoni, A. G. Relationship between crystallization behavior, microstructure, and mechanical properties in a palm oil-based shortening. J. Am. Oil Chem. Soc. 2002, 79, 647−654. (35) Pérez-Martínez, D.; Á lvarez-Salas, C.; Charó-Alonso, M.; Dibildox-Alvarado, E.; Toro-Vazquez, J. F. The cooling rate effect on the microstructure and rheological properties of blends of cocoa butter with vegetable oils. Food Res. Int. 2007, 40, 47−62. (36) Campos, R.; Narine, S. S.; Marangoni, A. G. Effect of cooling rate on the structure and mechanical properties of milk fat and lard. Food Res. Int. 2002, 35, 971−981. (37) Moniruzzaman, M.; Sundararajan, P. R. Low molecular weight organogels based on long-chain carbamates. Langmuir 2005, 21, 3802−3807. (38) Lam, R.; Pederson, T.; Quaroni, L.; Rogers, M. A. A molecular insight into the nature of crystallographic mismatches in selfassembled fibrillar networks under non-isothermal crystallization conditions. Soft Matter 2010, 6, 404−408. (39) Chen, X.; Hou, G.; Chen, Y.; Yang, K.; Dong, Y.; Zhou, H. Effect of molecular weight on crystallization, melting behavior and morphology of poly(trimethylene terephalate). Polym. Test. 2007, 26, 144−153. (40) Papageorgiou, G. Z.; Bikiaris, D. N.; Achilias, D. S. Effect of molecular weight on the cold-crystallization of biodegradable poly(ethylene succinate). Thermochim. Acta 2007, 457, 41−54. (41) Draper, N.; Smith, H. Applied Regression Analysis, 2nd ed.; New York: John Wiley & Sons, Inc. 1981.

M

dx.doi.org/10.1021/la400809a | Langmuir XXXX, XXX, XXX−XXX