Comparison of Pickering and Network Stabilization in Water-in-Oil

ARTICLE pubs.acs.org/Langmuir

Comparison of Pickering and Network Stabilization in Water-in-Oil Emulsions Supratim Ghosh, Tu Tran, and D
erick Rousseau* Department of Chemistry and Biology, Ryerson University, Toronto, Ontario, Canada ABSTRACT: We compared the efficacy of Pickering crystals, a continuous phase crystal network, and a combination thereof against sedimentation and dispersed phase coalescence in water-in-oil (W/O) emulsions. Using 20 wt % water-in-canola oil emulsions as our model, glycerol monostearate (GMS) permitted Pickering-type stabilization, whereas simultaneous usage of hydrogenated canola oil (HCO) and glycerol monooleate (GMO) primarily led to network-stabilized emulsions. A minimum of 4 wt % GMS or 10 wt % HCO was required for long-term sedimentation stability. Although there were no significant differences between the two in mean droplet size with time, the free water content of the network-stabilized emulsions was higher than Pickering-stabilized emulsions, suggesting higher instability. Microscopy revealed the presence of crystal shells around the dispersed phase in the GMSstabilized emulsions, whereas in the HCO-stabilized emulsion, spherulitic growth in the continuous phase and on the droplet surface occurred. The displacement energy (Edisp) to detach crystals from the oil
water interface was ∼104 kT, and was highest for GMS crystals. Thermal cycling to induce dispersed phase coalescence of the emulsions resulted in desorption of both GMS and GMO from the interface, which we ascribed to solute
solvent hydrogen bonding between the emulsifier molecules and the solvent oil, based on IR spectra. Overall, Pickering crystals were more effective than network crystals for emulsion stabilization. However, the thermal stability of all emulsions was hampered by the diffusion of the molten emulsifiers from the interface.

’ INTRODUCTION Many oil-continuous emulsions (e.g., in foods, pharmaceuticals, cosmetics, and crude oil) are kinetically stabilized by crystals [paraffins, triacylglycerols (TAGs), etc.]. Although progress has been made in establishing the fundamentals of such crystalstabilized emulsions, important questions regarding the stabilization capacity of continuous phase (“network”) crystals vis-a-vis interfacially adsorbed (“Pickering”) crystals in industrially relevant multiphasic systems remain.1
3 As monoacylglycerols (MAGs) are amphiphilic, they readily adsorb to oil
water interfaces, with their polar groups oriented toward the aqueous droplet surface and nonpolar groups residing in the continuous oil phase.4 Furthermore, if saturated, MAGs undergo a liquid
solid phase transition in the temperature range where emulsions are processed and stored, which suggests the possibility of interfacial crystallization. The effectiveness of such lipids as emulsion stabilizers depends on whether they remain at the interface, which is a function of size, shape, concentration, composition, and, importantly, hydrophobicity and wetting behavior.1 Particle wettability at an interface is described by the contact angle (θ) formed at the boundary between the oil, water, and solid phases. Wettability occurs when the three-phase contact angle (measured across the water phase) is greater than 0° and less than 180°. Highly hydrophilic (θ ∼ 0°) or hydrophobic particles (θ ∼ 180°) tend to diffuse away from the interface toward water or oil phase, respectively, and cannot be used as Pickering stabilizers.5 For particles of intermediate hydrophobicity, when θ is less than r 2011 American Chemical Society

90°, the particles stabilize oil-in-water (O/W) emulsions, whereas when θ is greater than 90°, water-in-oil (W/O) emulsions will be stabilized.1,6 At θ = 90°, the particles are equally wetted by the oil and aqueous phases. Fully saturated fats such as hydrogenated canola oil (HCO) are often used for network stabilization (tablespreads, baking, etc.).7
9 As these demonstrate little or no surface activity, they stabilize emulsions, not by interfacial adsorption, but by physically encasing the dispersed phase in a continuous crystalline matrix.10 This approach typically requires >5% solid fat for emulsion stabilization, yet by tailoring composition and processing, lower crystal volume fractions (e.g., j ∼ 0.02) can retard emulsion breakdown.11
13 Lucassen-Reynders14 explained that the presence of a fat crystal network could be deemed an energy barrier preventing the free diffusion of crystals away from the network. Once formed, such a network should also keep droplets separate from one another due to the presence of interstitial crystals, thus reducing flocculation and coalescence. If crystals were simultaneously adsorbed to the oil
water interface and present in the continuous oil phase as a fine particle network, both Pickering and network stabilization should further slow droplet
droplet coalescence and sedimentation, assuming a sufficiently high crystal concentration. Such W/O emulsion Received: January 6, 2011 Revised: March 16, 2011 Published: April 29, 2011 6589

dx.doi.org/10.1021/la200065y | Langmuir 2011, 27, 6589–6597

Langmuir

ARTICLE

Table 1. Solid Fat Composition of the Continuous Phase and SFC of Emulsions Stored at 25°C for 7 Daysa fat blend

a

GMS (wt %)

GMO (wt %)

HCO (wt %)

total solid fat (wt %)

SFC (%)

1% GMS

1

0

0

1

0.45

2% GMS

2

0

0

2

1.03

4% GMS

4

0

0

4

1.97

6% GMS

6

0

0

6

3.59

8% GMS

8

0

0

8

4.67

2% HCO
GMO

0

4

2

2

1.61

4% HCO
GMO

0

4

4

4

2.92

6% HCO
GMO 8% HCO
GMO

0 0

4 4

6 8

6 8

4.56 5.50

10% HCO
GMO

0

4

10

10

5.98

3%GMS
2.5%HCO

3

0

2.5

5.5

2.70

2%GMS
5%HCO

2

0

5

7

3.70

1%GMS
7.5%HCO

1

0

7.5

8.5

4.60

4%GMS
10%HCO

4

0

10

14

7.78

The SFC standard deviations are not shown in the table for clarity.

stabilization has been observed in numerous applications.15
21 Nevertheless, little is known regarding the interplay of these two means of stabilization, and which approach is more efficient on a solids basis. In the present study, we investigated the role of continuous and Pickering type crystals on the stability of W/O emulsions using two lipids: the saturated MAG glycerol monostearate (GMS) and HCO, which consists primarily of tristearin (3  C18:0). We demonstrate that Pickering stabilization is more efficient and that otherwise surface-inactive HCO can crystallize at the oil
water interface in the presence of glycerol monooleate (GMO). Findings from this study are equally applicable to food and other systems such as petroleum products, cosmetics, and pharmaceuticals.

’ MATERIALS AND METHODS Materials. Deionized water with a resistivity of >15 MΩ 3 cm (Barnstead E-Pure, Thermolyne, Ottawa, ON, Canada) was used for the aqueous phase. Canola oil (CO) [acid value ∼0.2%22] was purchased from a local grocery store and stored at room temperature (RT). Distilled GMS (Pationic 901, > 95% GMS) and GMO (Dimodan MO 90, >92% GMO) were kindly provided by Caravan Ingredients (Lenexa, KS, USA) and Danisco (New Century, KS, USA), respectively. HCO (fully hydrogenated canola oil) was purchased from Bunge (Oakville, ON, Canada). Pickering emulsions were made with 1
8 wt % GMS in the continuous oil phase, whereas network-stabilized emulsions consisted of 4 wt % GMO plus 2
10 wt % HCO in the continuous oil phase. In these emulsions, GMO was added to aid in emulsification only and alone, had no stabilizing effect. For combined Pickering and network stabilization, emulsions were made with blends of GMS (1
4 wt %) and HCO (2.5
10 wt %) in the oil phase. The solid fat composition of the emulsions’ continuous oil phase is shown in Table 1. Emulsion Preparation. W/O emulsions were prepared by premixing the oil phase (80 wt %) with water (20 wt %) in a rotor/stator (PT 10/35, Kinematica, Inc., Bohemia, NY, USA) for 1 min at 27 000 rpm. These coarse mixtures were emulsified in a high pressure valve homogenizer (APV-1000, APV, Albertslund, Denmark) at a pressure of 10 000 psi with 6 cycles. Homogenization conditions were based on preliminary experiments that showed similarly narrow droplet size distributions for the GMS and GMO emulsions. Emulsification was performed at >70 °C to ensure that all components were liquid. All emulsions were cooled with continuous stirring (500 rpm with a magnetic stirrer) to RT

(25.5 °C ( 0.5 °C), which allowed the GMS and HCO to crystallize and prevent sedimentation of the dispersed phase. Static cooling was ineffective as the dispersed droplets sedimented and coalesced prior to GMS shell or HCO network formation. A stirring speed of 500 rpm was the lowest speed that resulted in stable emulsion formation, where stability was defined as the lack of change in droplet size and sedimentation for 7 days. The effect of stirring speed was tested on emulsions containing either 4 wt % GMS or 4 wt % GMO and 10 wt % HCO. Emulsion Storage Stability. Emulsion samples were transferred to NMR tubes (ID = 0.8 cm, L = 20 cm) and glass vials (ID = 2.5 cm, L = 9.5 cm) and stored at 25 °C for 7 days. Stability was assessed via sedimentation, droplet size, and microscopy. Sedimentation. Emulsion sedimentation was recorded visually and with a digital camera. Digital images of 6 cm high emulsions in glass vials were taken on days 0, 1, 2, 4 and 7, and the height of the sedimented emulsion layer was calculated using image analysis (Fovea Pro v3.5, Reindeer Graphics, Asheville, NC). Emulsion destabilization was measured based on the percent emulsified layer (Femulsion) retained during storage: Femulsion ¼

hemulsion  100% htotal

ð1Þ

where hemulsionand htotal are the height of the emulsion layer and total height of the sample, respectively. Droplet Size Determination. The dispersed droplet size distribution of the emulsions was determined at 25 °C using a Bruker Minispec Mq pulsed field gradient nuclear magnetic resonance (pfg-NMR) unit (Bruker Canada, Milton, ON, Canada) that allows unimodal characterization of emulsion droplet size distributions via restricted diffusion measurement.23
26 The Minispec Mq NMR software version was 2.58 revision 12/NT/XP (Bruker Biospin GmbH, Rheinstetten, Germany), and the water droplet size application was v5.2 revision 4a. The pulsed gradient separation and number of pulse widths were 210 ms and 8, respectively. The oil suppression delay was 85 ms, and the magnet gradient strength was 2 T/m. The pfg-NMR field gradient strength was calibrated with CuSO4-doped water (diffusion coefficient =2.3  10
9 m s
1 at 25 °C). Emulsion samples (height = 1 cm) in glass tubes (ID = 0.8 cm, L = 20 cm) were placed in the NMR unit and their d3,3 (volume-weighted geometric mean diameter) and distribution breadth σ (geometric standard deviation) were acquired.27,28 Free water, used as an indicator of emulsion instability, was defined as dispersed water droplets sized above the instrumental limit (∼ 500 μm) that had not phase-separated from the emulsion as bulk water. As pulsed NMR relies on the molecular movement of water molecules within droplets, it detects size increases in 6590

dx.doi.org/10.1021/la200065y |Langmuir 2011, 27, 6589–6597

Langmuir the droplets themselves and not their clustering, thereby differentiating coalescence from flocculation/coagulation. Solid Fat Content (SFC). Emulsions were pipetted into NMR tubes to a height of 4 cm. Samples were prepared and stored at 25 °C. Samples were analyzed using the Bruker Minispec Mq pulsed NMR unit with the application sfc_lfc v2.51. X-ray Diffraction (XRD). A Rigaku Geigerflex (Danvers, MA, USA) XRD unit (λ = 1.79 Å) was used to determine powder diffractograms of the fat samples at RT. Scans from 5° to 90° 2θ at 5°/min were performed. Thermal Analysis. Aliquots of fresh emulsion (∼10 mg) were hermetically sealed in aluminum pans and temperature-cycled in a differential scanning calorimeter (DSC) (Pyris Diamond model, PerkinElmer Life and Analytical Sciences, Woodbridge, ON, Canada). Samples were cooled from 25 °C to
50 °C, reheated to 70 °C and then recrystallized to
50 °C at a rate of 1 °C/min. During the last cooling cycle, the appearance of the crystallization peak associated with free (i.e., demulsified) water was used to indicate emulsion instability. Controls based on the pure components and bulk fat mixtures (in the same proportion as in emulsion continuous phase) were also temperature-cycled in the DSC. Both crystallization and melting points were determined from the peak temperatures using the DSC data analysis software (Pyris v.7, PerkinElmer Life and Analytical Sciences, Woodbridge, ON, Canada). Microscopy. Polarized light microscopy was used to characterize emulsion morphology as a function of storage time. Samples were placed on viewing slides at RT (Fisher Scientific, Ottawa, ON, Canada), covered with a coverslip (Fisher Scientific, Nepean, ON, Canada), and analyzed with an inverted Zeiss Axiovert 200 M light microscope (Zeiss Inc., Toronto, ON, Canada) equipped with a temperature-controlled stage (model TS60 with STC200 controller, Instec, Boulder, CO, USA). Contact Angle. The fat phase associated with a given emulsion was melted at ∼75 °C with stirring, poured into an aluminum weigh boat at RT and allowed to crystallize statically for ∼24 h. Solidified fat disks were then cut into ∼1 cm3 cubes and placed in spectrophotometer cuvettes, with the smooth side facing upward. Cuvettes were then filled with either CO or 4 wt % GMO in CO. A small water droplet (D ∼ 1 mm) was injected from a needle (gauge 22, 0.394 mm nominal ID, Fisher Scientific, Nepean, ON, Canada) onto the surface of the solid fat using a 10 mL syringe and a programmable syringe pump (kd Scientific, Markham, ON, Canada) operated at 0.1 mL/min. Images of sessile water droplets were captured daily for 7 days with a Teli CCD camera with a macro lens assembly and IDS Falcon/Eagle framegrabber (DataPhysics Instrument GmbH, Filderstadt, Germany). Droplet contact angles against the solid surface were measured digitally (SCA 20 software, version 2.1.5 build 16, DataPhysics Instrument GmbH, Filderstadt, Germany). The CO/water and CO
GMO/water interfacial tensions and CMCs were determined using ASTM method D971,29 with the DuNouy ring method (Fisher surface tensiometer Model 21, Fisher Scientific, Nepean, ON, Canada). Statistical Analyses. Triplicate analyses were performed on all sedimentation and droplet size measurements. Analyses of variance were performed, and statistical differences were considered significant at p = 0.05.

’ RESULTS AND DISCUSSION Sedimentation. After 7 days, Pickering-stabilized emulsions with 1 and 2 wt % GMS showed clear oil separation, leaving a sedimented emulsion layer in the vial (Figure 1). The emulsion with 4 wt % GMS showed little oil separation (emulsion height ∼97%), while those stabilized with 6 and 8 wt % GMS did not visually phase-separate. Network-stabilized (GMO-HCO) emulsions with e6 wt % HCO showed 53% emulsion retention, whereas those with 8 wt % HCO showed 68% retention. With 10 wt % HCO, the emulsion was stable against sedimentation.

ARTICLE

Figure 1. Sedimentation of GMS and HCO-stabilized emulsions stored at 25 °C for 7 days.

Improved stability was seen when a portion of the HCO were replaced by GMS, though oil separation was still observed. The GMS-HCO emulsion made with 4 wt % GMS and 10 wt % HCO (combined Pickering and network stabilization) was stable. Overall, on a weight percent solid fat basis, sedimentation stability was highest in the GMS-only emulsions. Emulsion flowability was examined visually by tilting the sample glass vials. All formulations were flowable, except for the 10 wt % HCO emulsions, as their fat crystal networks were sufficiently rigid to prevent flow. SFC and XRD. Emulsion SFCs after 7 days at 25 °C ranged from 0.5% to 4.7% in the GMS-stabilized emulsions, 1.6% to 6.0% in the GMO
HCO emulsions, and 2.7% to 7.8% in the GMS
HCO emulsions (Figure 2). All experimental SFCs were lower than calculated, due to lipid solubilization in the oil phase.30 At similar added solid fats (wt%) (Table 1), GMO
HCO emulsions yielded higher SFCs compared to GMS-only and GMS-HCO emulsions. For example, the emulsion containing 8 wt % GMS had an SFC of 4.7%, whereas the emulsion with 4 wt % GMO and 8 wt % HCO (total solid fat 8 wt %) yielded an SFC of 5.5%. The SFC of the emulsion with 1 wt % GMS and 7.5 wt % HCO (total solid fat 8.5 wt %) was only 4.6%. Thus, crystallization of HCO was substantially suppressed in the presence of GMS and less so with GMO. Katsuragi et al.31 worked on stearic and oleic acid-based sucrose polyester surfactants and noted that the former suppressed equilibrium SFC of hexadecane in O/W emulsions, whereas the latter had no effect. Awad and Sato32 noted similar effects on the crystallization of palm kernel oil. Aronhime et al.33 proposed that GMS could be incorporated within tristearin crystal lattices leading to vacancies between TAG chains, which presumably would result in a lower SFC than with HCO alone. 6591

dx.doi.org/10.1021/la200065y |Langmuir 2011, 27, 6589–6597

Langmuir

ARTICLE

Figure 2. SFC of emulsions stored at 25 °C for 7 days (n g 4).

Figure 3. Emulsion sedimentation stability as a function of SFC and composition. Higher values indicate higher stability (n g 4).

Femulsion plotted versus SFC (Figure 3) was used to estimate emulsion separation. For GMS-stabilized emulsions, a small increase in SFC led to a sizable improvement in emulsion stability (∼ 44% Femulsion at 0.5% SFC to ∼97% Femulsion at 2.0% SFC). In contrast, GMO-HCO-stabilized emulsions showed a slow rise in stability up to a critical SFC (4.6%), above which interdroplet contact was prevented. Partial replacement of GMS with HCO negatively impacted emulsion stability [e.g., ∼ 97% Femulsion with 4 wt % GMS (SFC 2.0%) compared to ∼83% Femulsion with 3 wt % GMS and 2.5 wt % HCO (SFC 2.7%)], further demonstrating that Pickering crystals provided enhanced stabilization compared to network stabilization. On the basis of powder XRD, GMS
HCO and GMO
HCO present in the emulsion continuous phase was in the β-polymorph at RT. Droplet Size Distribution and Free Water Content. The d3,3 values for both Pickering (4 wt % GMS) and network-stabilized (4 wt % GMO
10 wt % HCO) emulsions were 20
23 μm and did not change with storage time (p > 0.05) (Figure 4A). The breadth of the droplet size distributions (σ) also remained constant (1.0
1.4) with storage (p > 0.05) (Figure 4B). In the mixed systems, the 1 wt % GMS and 7.5 wt % HCO blend showed the highest d3,3 and σ values throughout the 7 day period. Decreasing d3,3 values for the other mixed systems (2 wt % GMS
5 wt % HCO and 3 wt % GMS
2.5 wt % HCO) were

Figure 4. Evolution in (A) dispersed phase mean droplet size (d3,3), (B) breadth of droplet size distribution (σ), and (C) free water content of W/O emulsions as a function of continuous phase composition and storage in days. (--2--) 4 wt % GMS; (--(--) 4 wt % GMO-10 wt % HCO; (--9--) 4 wt % GMS-10 wt % HCO; (--0--) 1 wt % GMS-7.5 wt % HCO; (----) 2 wt % GMS-7.5 wt % HCO; (--O--) 3 wt % GMS-7.5 wt % HCO.

interpreted as the loss of larger droplets as bulk water, leading to an apparent decline in droplet size. These results were supported by a decrease in the breadth of the distribution (σ values) (Figure 4B) and an increase in the free water content (Figure 4C). Mixed systems with 4 wt % GMS and 10% HCO were stable against coalescence and did not show any change in d3,3, σ or free water content. Thermal Behavior. The DSC cooling and heating curves of the components used in the emulsions’ continuous phase are 6592

dx.doi.org/10.1021/la200065y |Langmuir 2011, 27, 6589–6597

Langmuir

ARTICLE

Figure 5. DSC heating (---) and cooling (—) thermograms of the oil phases used in the emulsions. GMS, GMO, and HCO were mixed in the same proportions as in the emulsions. Samples were temperature-cycled from 70 °C to
50 °C at a rate of 1 °C/min.

shown in Figure 5. The 4 wt % GMS in CO blend representative of the Pickering-stabilized emulsion showed a primary exothermic peak at 48 °C and a secondary peak at 37 °C, likely resulting from impurities (95% purity GMS was used) or perhaps a polymorphic transition.18 On heating, there was a small GMS endothermic peak at 61 °C, confirming its solid state at RT. The network stabilization fat (CO containing 4 wt % GMO and 10 wt % HCO) showed an HCO crystallization doublet at ∼32 °C and an endothermic peak at ∼55 °C, confirming network stabilization at RT. Crystallization of the mixed GMS
HCO blend (4 wt % GMS and 10 wt % HCO in CO) resulted in a small peak at 50 °C and a large peak at 37 °C, ascribed to GMS and HCO crystallization, respectively. On heating, the HCO and GMS melted as a single peak at 56
59 °C. HCO’s higher crystallization temperature when mixed with GMS likely resulted from GMS-mediated heterogeneous nucleation. A similar behavior was observed by Kalnin et al.29 who explored palm oil crystallization in the presence of propylene glycol monostearate. In all systems, CO crystallized at
38 to
41 °C and melted over a broad range, with two overlapping peaks at ca.
22 °C and ca.
17 °C. To discern the dispersed phase stabilization mechanisms, emulsions were temperature-cycled from 25 °C to
50 °C to crystallize water and CO, heated to 70 °C to melt all components and recooled to
50 °C (Figure 6). During the first cooling cycle, the GMS-stabilized emulsion showed several small irregular peaks from
20 °C to
30 °C and a large peak at
36 °C (Figure 6A), ascribed to the crystallization of polydispersed and narrowly sized water droplets, respectively.18 Embedded within this latter peak was also CO crystallization. Melting of CO and GMS in the emulsions was similar to the GMS
CO neat oil (Figure 5), along with an ice melting peak at 0 °C. During the second cooling cycle, the sharp peak at
20 °C resulted from demulsified water crystallization, as the now-melted GMS did not interfacially recrystallize.18 Cooling of the network-stabilized emulsion revealed a large exothermic peak at
38 °C that combined CO and water crystallization (Figure 6B). The small peak at 7 °C resulted from GMO crystallization. Upon heating, CO melted at ca.
20 °C, and ice melted at 0 °C followed by GMO at 12 °C and HCO at 60 °C. During the second cooling cycle, HCO crystallized at 32 °C, and CO crystallized at
38 °C. This emulsion also showed complete destabilization, based on bulk water crystallization at
19 °C. The inability of GMO to stabilize W/O emulsions was also observed by

Figure 6. Temperature-cycled thermograms of (A) the 4 wt % GMS emulsion and (B) the 4 wt % GMO
10 wt % HCO emulsion. Samples were cooled from 25 °C to
50 °C and heated to 70 °C followed by a second cooling cycle to
50 °C at a rate of 1 °C/min in a DSC. The inset shows a magnified view of the fat phase melting and crystallization region.

Johansson et al.9 who found that, without any stabilizing fat crystals, ∼70% of the dispersed phase in a 4 wt % GMO-stabilized 20 wt % water-in-soybean oil emulsion separated after one day. Microscopy. Pickering-stabilized emulsions consisted of GMS shells around the dispersed aqueous phase (Figure 7A). In the network-stabilized emulsions (Figure 7B), a continuous fat crystal network was apparent. Single crystals as well as larger HCO spherulites were visible in the continuous phase, and, to our surprise, some crystals were associated with the dispersed water droplets (arrows in Figure 7B). In the combined GMS
HCO system, both Pickering and continuous phase crystals existed (Figure 7C). Our initial hypothesis of exclusive network stabilization in the HCO-stabilized emulsion was not entirely correct as spherulites were present at the water droplet surface (Figure 8A). Similar effects were also seen in the GMS
HCO system where HCO spherulites grew on GMS Pickering shells (Figure 8B). Wetting Behavior and Interfacial Crystallization. The contact angles (θ) after immediate water droplet deposition onto the fat surfaces ranged from 148° to 165°, and remained constant for 7 days (Table 2). The smallest θ was obtained for water droplets in GMO-doped CO against HCO and was highest for water 6593

dx.doi.org/10.1021/la200065y |Langmuir 2011, 27, 6589–6597

Langmuir

ARTICLE

Figure 7. Polarized light micrographs of Pickering and network-stabilized W/O emulsions. (A) 4 wt % GMS emulsion (Pickering stabilization); (B) 4 wt % GMO
10 wt % HCO emulsion (network stabilization), and (C) 4% GMS
10% HCO emulsion (combined Pickering and network stabilization). The presence of HCO crystals at the GMO-stabilized water droplet surface is shown by arrows in panel B. Scale bar: 40 μm.

droplets in CO against HCO. The displacement energy (Edisp), which is intimately linked to θ, was used as a measure of the capacity of the interfacially adsorbed GMS and HCO crystals to promote emulsion stability:5,6 Edisp ¼ πr 2 γow ð1 ( cos θÞ2

ð2Þ

where γow is the interfacial tension at the oil
water interface and r is the radius of a fat “particle”, which is assumed to be much

smaller than the curvature of the interface. The sign inside the bracket is negative for displacement into the water phase, and positive for displacement into oil.1 With γow = 28.8 and 5.2 mN/m for the CO/water and CO
GMO/water interfacial tensions, respectively, and r ∼ 1 μm, Edisp ranged from 2.8 to 14.3  104 kT (Table 2). The Edisp to remove a GMS crystal from the water/CO interface was ∼5 times higher than for an HCO crystal of the same size, with GMO doubling the Edisp of HCO. Thus, GMS crystals were the most effectively “anchored” to the aqueous droplets.1 Other than its role in reducing interfacial tension, GMO promoted HCO interfacial adherence owing to the carbon chain length match between the kinked oleic acid (GMO) and saturated stearic acids (HCO). This chain
chain ordering also resulted in HCO heterogeneous nucleation at the interface (Figure 8A), as noted by Davies et al.34 in their work on GMO
tristearin systems. In the GMS
HCO blend, heterogeneous nucleation was further promoted as GMS crystallized at a higher temperature than GMO (48 vs 7 °C, Figure 5), and the stearic acids in the GMS and HCO were more apt to enmesh, given their matched chain length and level of saturation.35 Similarly, Arima et al.36 found that high-melting sucrose palmitic acid oligoesters acted as a template for the interfacial nucleation and growth of palm oil midfraction at the surface of oil droplets in O/W emulsions. On the basis of eq 2, the maximum Edisp will occur when θ values are close to 90°, i.e., when the fat crystals are equally wetted by the oil and water phases. Tailoring θ by using different emulsifiers to approach ∼90° is an easy way to increase displacement energy. Edisp also depends on the square of the particle radius, and will significantly increase for larger particles. As interfacial crystallization of HCO and GMS continued well after the initial liquid
solid phase transition, these crystals grew and sintered over time, thus improving their tendency to remain at the interface. Finally, surfactants that demonstrate molecular complementarity with, and nucleate at a higher temperature than, the crystallizing lipid will also promote interfacial crystallization. Mechanisms of Emulsion Destabilization. DSC of the emulsions’ freeze
thaw stability (Figure 6) led us to question the stabilization efficacy of both GMO and GMS. Our results suggested that these did not maintain the integrity of the dispersed droplets once the stabilizing fats had melted, and that perhaps more favorable interactions between the surfactants and continuous phase existed. As Bus et al. have shown that the hydroxyl groups in MAGs could hydrogen bond (H-bond) with the carbonyl moieties in vegetable oil fatty acids,37 we theorized that preferential solute
solvent H-bonding significantly reduced the availability of MAGs at the water/oil interface in fully molten emulsions. With their adsorption/desorption equilibrium heavily shifted toward desorption, the MAGs were no longer able to stabilize the dispersed aqueous phase, which led to emulsion coalescence and sedimentation. To test this hypothesis, IR absorbance spectra of the starting compounds and 4% GMO dissolved in either CO or light mineral oil (LMO) were compared (Figure 9). LMO was used as it only consisted of alkanes, with no functional groups to H-bond with the MAGs.37 The spectrum of GMO alone showed a strong peak at ∼3400 cm
1, corresponding to intermolecular H-bonding between
OH groups in neighboring GMOs.37 With this peak absent, the LMO spectrum did not show any indication of H-bonding. By contrast, CO showed a strong peak at 3476 cm
1 and several weaker peaks from 3510 to 3580 cm
1, which we ascribed to the intermolecular H-bonding between 6594

dx.doi.org/10.1021/la200065y |Langmuir 2011, 27, 6589–6597

Langmuir

ARTICLE

Figure 8. Polarized light micrographs demonstrating Pickering stabilization in the (A) GMO
HCO and (B) GMS
HCO emulsions. The presence of HCO crystal spherulites at the oil
water interface can be seen in both images (see arrows). Scale bar: 25 μm.

Table 2. Contact Angles of Sessile Water Drops on the Solid Fat Surface (GMS or HCO) against a Continuous Oil Phase and Corresponding Displacement Energy (Edisp) to Remove a Spherical Fat Particle (r = 1 μm) from the Water Droplet Surface solid phase

contact angle

Edisp (kT)a

CO

HCO

164.6 ( 1.0°

2.8  104

CO
GMO

HCO

152.7 ( 4.1°

4.9  104

CO

GMS

156.9 ( 0.6°

continuous phase


23

k is the Boltzmann constant (1.38  10 absolute temperature (K). a

2

m kg s


2

14.3  104


1

K ), and T is

Figure 10. Emulsion sedimentation: (A) GMO
CO emulsion after 1 h of preparation, (B) GMO
LMO emulsion after 24 h of preparation, and (C) GMS
CO emulsion temperature-cycled from 25 to 70 °C to 25 °C.

Figure 9. IR spectra of CO, LMO and GMO along with mixtures of 4 wt % GMO in CO or LMO.

TAG carbonyl groups and hydroxyl groups in neighboring partial acyglycerols and free fatty acids.37 A small peak at 3620 cm
1 was due to free
OH groups in partly hydrolyzed TAGs.38,39 Based on the peak at ∼3375 cm
1, there was only intermolecular H-bonding among GMO molecules in the GMO
LMO solution, demonstrating exclusive solute
solute interactions. In

the GMO
CO solution, intermolecular H-bonding between CO carbonyl groups and GMO hydroxyl groups occurred, based on the major peak at 3476 cm
1 and moderate peaks from ∼3510
3580 cm
1.37,40 The peak at 3375 cm
1, typical of H-bonding between
OH groups, was not observed. Rather, excess CO ester carbonyl groups H-bonded with the
OH groups in GMO, resulting in strong solute
solvent interactions. On the basis of IR spectra, we concluded that GMO was not readily available for emulsion stabilization, given its extensive H-bonding with CO. Further evidence was provided by GMO’s critical micelle concentration (CMC) in the two solvent oils. The CMC was determined from the plateau of the GMO
oil/water interfacial tension vs GMO concentration graph. While the CMC of GMO in LMO was only 0.15 wt %, in CO a significant increase was observed (CMC = 4.0 wt %). This was presumably due to the formation of solute
solvent H-bonds that prevented GMO’s diffusion toward the oil/water interface to reduce interfacial tension.37,41 Had GMO been used in LMO-based emulsions, lack of H-bonding between the two should have resulted in more stable 6595

dx.doi.org/10.1021/la200065y |Langmuir 2011, 27, 6589–6597

Langmuir

ARTICLE

Figure 11. Micrographs of the GMS emulsion as a function of temperature. Samples were heated and cooled on a microscope stage while micrographs of the same region were taken. Scale bar represents 50 μm.

emulsions. To test this, W/O emulsions were prepared with a 4% GMO in LMO solution as the continuous phase. Sedimentation results showed that, contrary to the GMO
CO emulsion that completely phase-separated after 1 h following preparation (Figure 10A), the GMO
LMO emulsion was stable with ∼57% of its initial emulsion height retained after 24 h (Figure 10B). This strongly suggested that the capacity of an emulsifier to stabilize W/O emulsions depends on how extensively it interacts with the continuous oil phase.37 GMS’ superior stabilization capacity over GMO was solely dependent on its rapid solid-state transition following emulsification. When GMS-stabilized emulsions were DSC temperaturecycled from 25 to 70 °C to 25 at 1 °C/min, the low cooling rate and quiescent conditions prevented GMS interfacial recrystallization, thereby leading to emulsion destabilization. Such breakdown was also observed with temperature-cycled emulsions (∼20 g) stored in glass vials (Figure 10C) and under the microscope (Figure 11). At temperatures lower than the melting point of GMS, the emulsions remained stable (Figure 11A, B). Upon melting of the GMS shells (60
70 °C), water droplets began to coalesce (Figure 11C, D). Cooling this destabilized emulsion led to GMS crystallization in the continuous phase and around large coalesced water droplets (Figure 11E, F). Binks et al.2 recently corroborated these results by observing the desorption of stabilizing wax microparticles from an oil
water interface upon their melting.

’ CONCLUSION The solids stabilization of W/O emulsions can be achieved with crystals located at the oil/water interface and/or in the continuous phase, with Pickering-type stabilization more effective than network stabilization. This study has also shown that fat crystals may possess a “switchable” polarity if liquid-state emulsifiers can absorb onto their surfaces. As per Johansson and  4,42 Bergenstahl, GMO may have conferred surface polarity to the HCO crystals, which promoted their presence at the oil/water interface. This was borne out by the contact angle measurements that indicated a higher Edisp for the HCO crystals in the presence of GMO. As a result, GMO promoted interfacial crystallization of the otherwise nonsurface active HCO. This led to a combination of Pickering and network-type stabilization with the use of a single, surface-inactive fat. Finally, this study suggests that the capacity of oil-tending emulsifiers such as GMO and GMS to stabilize W/O emulsions strongly depends on how extensively they H-bond with the continuous oil phase. If an emulsifier does not preferentially H-bond with the solvent oil, it does so with the dispersed aqueous phase at the water
oil interface (as in the case of mineral oil). Conversely, with vegetable oil, the surfactant adsorption/desorption equilibrium is heavily shifted toward desorption due to stronger H-bonding between the vegetable oil’s ester carbonyl groups and the surfactant’s hydroxyl groups 6596

dx.doi.org/10.1021/la200065y |Langmuir 2011, 27, 6589–6597

Langmuir (CdO 3 3 3 H) compared to the H
O 3 3 3 H bonds between the dispersed water molecules and the surfactant.43,44

’ AUTHOR INFORMATION Corresponding Author

*Address: Department of Chemistry and Biology Ryerson University Toronto, Ontario, Canada M5B 2K3. Telephone: (416) 9795000 ext. 2155. Fax: (416) 979-5044. E-mail: [email protected].

’ ACKNOWLEDGMENT The Natural Sciences and Engineering Research Council of Canada (NSERC) and the Ryerson Post-Doctoral Fellowship (RPDF) program are gratefully acknowledged for funding support. ’ REFERENCES (1) Binks, B. P. Curr. Opin. Colloid Interface Sci. 2002, 7, 21–41. (2) Binks, B. P.; Rocher, A. J. Colloid Interface Sci. 2009, 335, 94–104. (3) Macierzanka, A.; Szelag, H.; Szumala, P.; Pawlowicz, R.; Mackie, A. R.; Ridout, M. J. Colloids Surf. Physicochem. Eng. Aspects 2009, 334 (1
3), 40–52. (4) Johansson, D.; Bergenstahl, B.; Lundgren, E. J. Am. Oil Chem. Soc. 1995, 72 (8), 921–931. (5) Binks, B. P.; Lumsdon, S. O. Langmuir 2000, 16, 8622–8631. (6) Levine, S.; Bowen, B. D.; Partridge, S. J. Colloids Surf. 1989, 38, 325–343. (7) Ghotra, B. S.; Dyal, S. D.; Narine, S. S. Food Res. Int. 2002, 35 (10), 1015–1048. (8) Vaiseygenser, M.; Ylimaki, G. Cereal Foods World 1989, 34 (3), 246–258. (9) Johansson, D.; Bergenstahl, B.; Lundgren, E. J. Am. Oil Chem. Soc. 1995, 72 (8), 939–950. (10) Johansson, D.; Bergenstahl, B. J. Am. Oil Chem. Soc. 1995, 72 (8), 911–920. (11) Hodge, S. M.; Rousseau, D. Food Res. Int. 2003, 36 (7), 695–702. (12) Hodge, S. M.; Rousseau, D. J. Am. Chem. Soc. 2005, 82 (3), 159–164. (13) Garti, N.; Benyamin, H.; Aserin, A. J. Am. Oil Chem. Soc. 1998, 75 (12), 1825–1831. (14) Lucassen-Reynders, E. H.; van den Tempel, M. J. Phys. Chem. 1963, 67, 731–734. (15) Heertje, I. Food Struct. 1993, 12 (1), 77–94. (16) Heertje, I. Pol. J. Food Nutr. Sci. 1998, 7 (supplement), 2S–18S. (17) Juriaanse, A. C.; Heertje, I. Food Microstruct. 1988, 7 (2), 181–188. (18) Ghosh, S.; Rousseau, D. J. Colloid Interface Sci. 2009, 339, 91–102. (19) Rousseau, D.; Ghosh, S.; Park, H. J. Food Sci. 2009, 74 (1), E1–E7. (20) Binks, B. P.; Lumsdon, S. O. Phys. Chem. Chem. Phys. 1999, 1 (12), 3007–3016. (21) Abend, S.; Bonnke, N.; Gutschner, U.; Lagaly, G. Colloid Polym. Sci. 1998, 276 (8), 730–737. (22) Free Fatty Acids, method Ca 5a-40. In Official Methods and Recommended Practices of the AOCS; American Oil Chemists' Society: Champaign, IL, 1998. (23) Balinov, B.; S€oderman, O.; W€arnhelm, T. J. Am. Chem. Soc. 1994, 71, 513–518. (24) Lee, H. Y.; M., M. J.; Dungan, S. R. J. Am. Chem. Soc. 1998, 75, 463–475. (25) Li, X.; Cox, J. C.; Flumerfelt, R. W.; Am., J. Inst. Chem. Eng. 1991, 38, 1671–1674. (26) van den Enden, J. C.; Waddington, D.; van Aalst, H.; van Kralingen, C. G.; Packer, K. J. J. Colloid Interface Sci. 1990, 140, 105–113. (27) Alderliesten, M. Part. Part. Syst. Charact. 1991, 8 (3), 237–241. (28) Alderliesten, M. Part. Part. Syst. Charact. 1990, 7 (4), 233–241.

ARTICLE

(29) Kalnin, D.; Schafer, O.; Amenitsch, H.; Ollivon, M. Cryst. Growth Des. 2004, 4 (6), 1283–1293. (30) Hildebrand, J. H. J. Am. Chem. Soc. 1916, 38, 1452–1473. (31) Katsuragi, T.; Kaneko, N.; Sato, K. Colloid Surf., B: Biointerfaces 2001, 20 (3), 229–237. (32) Awad, T.; Sato, K. Colloid Surf., B: Biointerfaces 2002, 25 (1), 45–53. (33) Aronhime, J. S.; Sarig, S.; Garti, N. J. Am. Oil Chem. Soc. 1988, 65 (7), 1144–1150. (34) Davies, E.; Dickinson, E.; Bee, R. Food Hydrocolloids 2000, 14 (2), 145–153. (35) Davies, E.; Dickinson, E.; Bee, R. D. Int. Dairy J. 2001, 11 (10), 827–836. (36) Arima, S.; Ueno, S.; Ogawa, A.; Sato, K. Langmuir 2009, 25 (17), 9777–9784. (37) Bus, J.; Groeneweg, F.; van Voorst Vader, F. Prog. Colloid Polym. Sci. 1990, 82, 122–130. (38) Stuart, B. Infrared Spectroscopy: Fundamentals and Applications; John Wiely & Sons, Inc.: Hoboken, NJ, 2004. (39) Ksendzova, G. A.; Polozov, G. I.; Skornyakov, I. V.; Sorokin, V. L.; Tolstorozhev, G. B.; Shadyro, O. I.; Yakunin, A. A. Opt. Spectrosc. 2007, 102 (4), 551–556. (40) Brune, B. J.; Koehler, J. A.; Smith, P. J.; Payne, G. F. Langmuir 1999, 15, 3987–3992. (41) Debye, P.; Prins, W. J. Colloid Sci. 1958, 13, 86–98. (42) Johansson, D.; Bergenstahl, B. J. Am. Oil Chem. Soc. 1992, 69 (8), 705–717. (43) Fainerman, V. B.; Miller, R.; Ferri, J. K.; Watzke, H.; Leser, M. E.; Michel, M. Adv. Colloid Interface Sci. 2006, 123
126, 163–171. (44) Pacios, L. F. In Hydrogen Bonding - New Insights; Slawomir, G., Ed.; Springer: Dordrecht, The Netherlands, 2006; pp 109
148.

6597

dx.doi.org/10.1021/la200065y |Langmuir 2011, 27, 6589–6597