Effect of cooling and heating rates on polymorphic transformations and gelation of tripalmitin Solid Lipid Nanoparticle (SLN) suspensions (original) (raw)

Effect of Cooling and Heating Rates on Polymorphic Transformations and Gelation of Tripalmitin Solid Lipid Nanoparticle (SLN) Suspensions

Tarek S. Awad ⋅\cdot Thrandur Helgason ⋅\cdot
Kristberg Kristbergsson ⋅\cdot Eric A. Decker ⋅\cdot
Jochen Weiss ⋅\cdot D. Julian McClements

Received: 4 December 2007 / Accepted: 14 January 2008 /Published online: 19 February 2008
© Springer Science + Business Media, LLC 2008

Abstract

Solid lipid nanoparticle (SLN) suspensions undergo a phase transition from the α\alpha - to β\beta-polymorphic forms, which is accompanied by particle aggregation and gel formation. These processes are both time and temperature dependent, and so it is important to study the impact of cooling rates (CRs) and heating rates (HRs) on polymorphic transformations, particle aggregation, and gelation. Rheology measurements indicated that the temperature where gelation was first observed during cooling ( Tgel T_{\text {gel }} ) decreased with increasing CRs, with SLN suspensions remaining fluid at HR≥5∘C/min\mathrm{HR} \geq 5^{\circ} \mathrm{C} / \mathrm{min}. On the other hand, the temperature where gelation was first observed during heating of stable SLN suspensions increased with increasing HRs: 18, 24, 31, and 45∘C45^{\circ} \mathrm{C} at 2,5,102,5,10, and 20∘C/min20^{\circ} \mathrm{C} / \mathrm{min}, respectively. When the melted SLN suspensions were cooled again, two exothermic peaks were observed in the differential scanning calorimetry scans at 39 (which was attributed to coalesced oil) and 19∘C19^{\circ} \mathrm{C} (which was attributed to stable SLN). With increasing CR, the enthalpy of the coalescence peak (ΔHCO)\left(\Delta H_{\mathrm{CO}}\right) decreased, while that of the supercooled SLN⁡(ΔHSLN)\operatorname{SLN}\left(\Delta H_{\mathrm{SLN}}\right) increased. With increasing HR, ΔHCO\Delta H_{\mathrm{CO}} decreased and ΔHSLN\Delta H_{\mathrm{SLN}} increased, with no coalescence being observed at HR≥10∘C/min\mathrm{HR} \geq 10^{\circ} \mathrm{C} / \mathrm{min}. These results show that increasing the CRs or HRs retard the α→β\alpha \rightarrow \beta polymorphic transformation, which increased the stability of

[1]SLN against aggregation and retarded gelation. In addition, this study shows that the careful selection of HRs and CRs is required to examine polymorphic transformations and the stability of SLN suspensions.

Keywords Solid lipid nanoparticles ⋅\cdot Stability ⋅\cdot
Crystallization ⋅\cdot Polymorphism ⋅\cdot DSC ⋅\cdot Rheology ⋅\cdot Gelation

Introduction

In the food industry, there is a growing trend toward fortifying foods with lipophilic active ingredients such as oil-soluble vitamins, antioxidants, phytosterols, and polyunsaturated oils. The presence of such bioactive ingredients in the diet is essential for delaying the onset of many age-related diseases such as cancer, diabetes, and cardiovascular disease. 1−3{ }^{1-3} However, there are considerable practical challenges associated with successfully incorporating many of these lipophilic functional components into foods, such as low water solubility, susceptibility to chemical degradation, and crystalline form. Creation of solid lipid nanoparticle (SLN) suspensions has proved to be a powerful means of delivering poorly water-soluble drugs in the pharmaceutical industry. 4−6{ }^{4-6} SLNs offer many advantages over other colloidal carriers because of their high encapsulation capacity, high chemical and physical stability, improved oral bioavailability of drugs, large-scale production, and controlled release characteristics. 4,5,7−9{ }^{4,5,7-9} SLN may therefore have a great potential as carriers for the encapsulation and protection of lipophilic components in food applications. 10{ }^{10}

The use of solid lipids, such as high-melting triglycerides (TAGs), aids in restricting the mobility of encapsulated actives as compared to liquid oils. This property offers many

1. T. S. Awad ( ⊠)⋅\boxtimes) \cdot T. Helgason ⋅\cdot E. A. Decker ⋅\cdot J. Weiss ⋅\cdot D. J. McClements
   Department of Food Science, University of Massachusetts, 100 Holdsworth Way, Amherst, MA 01003, USA
   e-mail: tawadjp@gmail.com
   T. Helgason ⋅\cdot K. Kristbergsson
   Department of Food Science, University of Iceland, Hjardarhagi 2-6,
   Reykjavik 107, Iceland ↩︎

advantages for the use of SLN delivery systems in pharmaceutics for enhancing stability of sensitive drugs against decomposition and slowing their release until reaching the site of action. 11{ }^{11} The fatty acid chains in a TAG vary in their length (i.e., short, medium, and long), type (i.e., saturated or unsaturated), degree of unsaturation (i.e., mono-, di-, or polyunsaturated), and uniformity (i.e., mono-or mixed-acids). 12{ }^{12} These structural variations provide a wide spectrum of physical properties such as different crystallization and melting temperatures and polymorphic characteristics, which could be implemented to design stable bioactive carriers with a controlled release rate. For example, pure monoacid TAGs form highly dense crystalline structures, whereas mixed TAGs form less dense crystalline structures. In addition, the transition rate to the most stable polymorphic form is faster in short TAGs than in long TAGs. All of these factors will influence the physicochemical characteristics of SLN such as stability, loading capacity, release speed, and bioavailability of bioactive ingredients. 7,13{ }^{7,13}

SLN of triglycerides are usually prepared by controlled cooling of emulsions formed by high-pressure melt homogenization. 8{ }^{8} The particle size and distribution depends on the surfactant type and concentration, as well as on the homogenization conditions. 14−17{ }^{14-17} Previously, there have been a number of studies on the effects of emulsifiers and preparation conditions on the physical stability and functionality of SLN. 18−21{ }^{18-21} Many studies have focused on the role of emulsifiers on crystallization and polymorphic behavior of the lipid matrix. Westesen and Siekmann 22{ }^{22} observed gel formation in phospholipid-stabilized tripalmitate SLN suspensions, which was attributed to aggregation of platelet crystals not completely coated by the emulsifier. Freitas and Müller 23,24{ }^{23,24} investigated the effect of storage conditions, shear forces, and packing materials on the incidence of gelation. Recent research has shown that the particles formed in SLN suspensions remain spherical when the triglycerides stay in an α\alpha-modification, but they become elongated or platelet shaped when the triglycerides transform into the β\beta modification. 25−28{ }^{25-28} We recently investigated the mechanism of gelation in tripalmitin SLN suspensions and found that the incidence of gelation at different storage temperatures was associated with a polymorphic transition to the β\beta-form, which increased the overall particle surface area and promoted particle aggregation through hydrophobic attraction of lipids at surfactant-free patches. 29{ }^{29}

To make stable SLN suspensions for the encapsulation and protection of sensitive bioactive ingredients, it is necessary to either retard the polymorphic transition or to control the aggregation stability. 24{ }^{24} There have been many studies on modification of the crystallization and polymorphic behavior and stability of lipids in emulsions as well as SLN by varying surfactants, 19,20,30−35{ }^{19,20,30-35} lipid composition, 36{ }^{36} particle size and distribution, 30,37,38{ }^{30,37,38} aqueous phase composition, and
application of shear. 39−41{ }^{39-41} Because the gelation of SLN was observed shortly after crystallization of the lipid nanoparticles, 22−24{ }^{22-24} it is important to investigate the gelation phenomenon under different crystallization conditions. In this paper, we investigated the gelation and thermal behavior of tripalmitin SLN under different cooling (CRs) and heating rates (HRs). The results from this study provide important information about the relation between CRs and HRs and the phase transitions and stability of SLN. Ultimately, the information gained from this study will be useful for developing effective protocols to characterize SLN and to improve the stability of SLN suspensions.

Experimental Details

Materials

Tripalmitin was purchased from Fluka (Buchs, Switzerland). Sodium phosphate monobasic and sodium phosphate dibasic were purchased from Fisher Scientific (St. Clair Shores, MI). Sodium azide and polyethylene glycol sorbitan monolaurate (Tween 20) were purchased from Sigma-Aldrich Chemical (St Louis, MO). All chemicals were used as received.

Methods

Nanoparticle Preparation

Tripalmitin nanoparticles were prepared by high-pressure melt homogenization. 10%(w/w)10 \%(w / w) tripalmitin was melted and held at 75−80∘C75-80^{\circ} \mathrm{C} for more than 30 min to erase crystal history. The melt was then mixed with a hot buffered surfactant solution ( 1.5%[w/w]1.5 \%[w / w] Tween 20, 4 mM sodium phosphate monobasic, 6 mM sodium phosphate dibasic, 0.02%0.02 \% sodium azide, pH7,75∘C\mathrm{pH} 7,75^{\circ} \mathrm{C} ). The hot mixture was thoroughly blended for 1 min using a hand-held high speed blender (Model SDT-1810, EN Shaft, Tekmar, Cincinnati, OH ). Then, the resulting coarse premix was finely dispersed under pressure ( 9 kbar and ten times) using a microfluidizer (Microfluidics, Newton, MA). This produced an emulsion with a monomodal size distribution and a mean particle diameter of 145 nm . The emulsion was then stored in a plastic container at 37∘C37^{\circ} \mathrm{C} in a temperaturecontrolled room before use. No change in mean particle size was observed after 2 month of storage, and the lipid phase did not crystallize, indicating that this emulsion contained liquid droplets that were stable to aggregation.

Particle Size Determination

Particle size measurement was performed by photon correlation spectroscopy (PCS) using a Malvern Zetasizer NanoZS

(Malvern instruments, Malvern, UK). The samples were diluted 100 times in the same buffer solution at 37∘C37^{\circ} \mathrm{C}. PCS gives the mean diameter of the particle population and the polydispersity index (PI) ranging from 0 (monodisperse) to 0.50 (very broad distribution). To monitor the emulsion stability, the particle size distribution was measured at 37∘C37^{\circ} \mathrm{C} for the emulsion soon after preparation and during a period of 2 months.

Visual Observation and Microscopy

Sample gelation was detected in selected SLN suspensions by visual observation of samples stored in glass test tubes at fixed temperatures. The microstructure of SLN suspensions was also observed using an optical microscope (Nikon microscope Eclipse E400, Nikon, Japan). 29{ }^{29}

Rheology

The storage modulus (G′)\left(G^{\prime}\right) and loss modulus (G′′)\left(G^{\prime \prime}\right) of SLN suspensions undergoing a sol-to-gel transition were measured with an oscillatory rheometer (Physica MCR 300, Anton Paar, Ganz, Austria) using a cone and plate measurement system (diameter 49.94 mm , angle 2∘2^{\circ} ) thermostatted by a Peltier system. The cone was positioned 50μ m50 \mu \mathrm{~m} above the plate to avoid artifacts arising from the presence of any large particles in the samples. Initially, the linear viscoelastic range was determined by conducting a strain sweep at an oscillation frequency of 1 Hz . The strain sweep indicated that G′G^{\prime} and G′′G^{\prime \prime} of gelled samples did not decrease at strains less than 0.8%0.8 \%, and consequently, all gelation experiments were conducted at a constant strain of 0.1%0.1 \% and frequency of 1 Hz . The temperature of the loaded sample was equilibrated to 37∘C37^{\circ} \mathrm{C} using the Peltier system before any cooling/heating run. The effect of CR on rheology was studied by loading emulsion samples at 37∘C37^{\circ} \mathrm{C} then cooling them down to 5∘C5{ }^{\circ} \mathrm{C} at different CRs. The effect of HR was studied by loading emulsion samples at 37∘C37^{\circ} \mathrm{C}, cooling them down to 5∘C5^{\circ} \mathrm{C} at 10∘C/min10^{\circ} \mathrm{C} / \mathrm{min}, and then heating them up to 75∘C75{ }^{\circ} \mathrm{C} at different HRs. In both studies, the temperature where G′G^{\prime} and G′′G^{\prime \prime} were equal, i.e., where the phase angle was 45∘45^{\circ}, was defined as the gelation temperature Tgel T_{\text {gel }}.

Differential Scanning Calorimetry

A differential scanning calorimeter (DSC; Q1000, TA Instruments, New Castle, DE) was used to study polymorphic transitions at different CRs and HRs. An aliquot of the sample (8−10mg)(8-10 \mathrm{mg}) was placed in a hermetic aluminum pan and sealed. An empty pan was used as a reference. All the DSC pans were prepared at 37∘C37^{\circ} \mathrm{C} to avoid crystallization/gelation before measuring. Different heat-cool cycles
were used to study the effects of HRs and CRs on emulsion stability and lipid phase transitions:

• CRs: The effect of initial CR was examined by loading samples into a DSC pan at 37∘C37^{\circ} \mathrm{C} and then cooling to 5∘C5^{\circ} \mathrm{C} at 0.2 to 20∘C/min20^{\circ} \mathrm{C} / \mathrm{min}. After cooling, the samples were heated to 75∘C75^{\circ} \mathrm{C} at +10∘C/min+10^{\circ} \mathrm{C} / \mathrm{min} to melt the fat crystals and then cooled to 5∘C5^{\circ} \mathrm{C} again at −10∘C/min-10^{\circ} \mathrm{C} / \mathrm{min}.
• HRs: The effect of HR was examined by loading samples into a DSC pan at 37∘C37^{\circ} \mathrm{C} and then cooling to 5∘C5^{\circ} \mathrm{C} at −10∘C/min-10^{\circ} \mathrm{C} / \mathrm{min}. The samples were then heated to 75∘C75^{\circ} \mathrm{C} at 0.2 to 40∘C/min40^{\circ} \mathrm{C} / \mathrm{min} to melt the crystals and then cooled to 5∘C5^{\circ} \mathrm{C} at −10∘C/min-10^{\circ} \mathrm{C} / \mathrm{min}.

Results and Discussion

Preparation of Tripalmitin Oil-in-Water Emulsions

The mean particle diameter of freshly prepared emulsions cooled to 37∘C37^{\circ} \mathrm{C} was 146.9±0.6 nm146.9 \pm 0.6 \mathrm{~nm}, and the PI was 0.05 indicating a narrow particle size distribution. There were no significant changes in mean particle size and PI after 6 weeks storage at 37∘C37^{\circ} \mathrm{C} indicating that the emulsions were stable to droplet aggregation and Ostwald ripening. 15{ }^{15} When the emulsions were heated from 37 to 75∘C75^{\circ} \mathrm{C} in a DSC, there was no evidence of phase transitions (data not shown), which indicated that the tripalmitin remained in the liquid state during this time.

Polymorphic Behavior of Bulk and Emulsified Tripalmitin

Initially, the thermal behavior of bulk tripalmitin was characterized when it was cooled (from 80 to 5∘C5^{\circ} \mathrm{C} ) and then heated (from 5 to 80∘C80^{\circ} \mathrm{C} ) in a DSC (Figure 1a). During the cooling step, a single exothermic peak was observed around 39∘C39^{\circ} \mathrm{C}, which can be attributed to the crystallization of the tripalmitin into the metastable α\alpha-form. 12{ }^{12} Upon heating, a series of endothermic and exothermic peaks were observed in the thermogram, which can be assigned to particular lipid phase transitions. 12{ }^{12} The endothermic peak at 41∘C41^{\circ} \mathrm{C} corresponds to melting of the α\alpha-form, the exothermic peak at 47∘C47^{\circ} \mathrm{C} corresponds to formation of the β′\beta^{\prime}-form, the small endothermic peak at 53∘C53^{\circ} \mathrm{C} corresponds to melting of the β′\beta^{\prime}-form, the exothermic peak at 56∘C56^{\circ} \mathrm{C} corresponds to formation of the β\beta-form, and the large endothermic peak at 63∘C63^{\circ} \mathrm{C} corresponds to melting of the β\beta-form.

The thermal behavior of emulsified tripalmitin was characterized under the same conditions as for the bulk tripalmitin (Figure 1b). The emulsified tripalmitin was cooled from 37 to 5∘C5^{\circ} \mathrm{C} to crystallize the fat, then heated to 75∘C75^{\circ} \mathrm{C} to melt the fat, and then cooled to 5∘C5^{\circ} \mathrm{C} to crystallize the fat again. During the first cooling step, a

single exothermic peak was observed around 19∘C19^{\circ} \mathrm{C}, which can be attributed to the crystallization of the tripalmitin into the α\alpha-form. The crystallization temperature was about 20∘C20^{\circ} \mathrm{C} lower for emulsified tripalmitin (TC≈19∘C)\left(T_{\mathrm{C}} \approx 19{ }^{\circ} \mathrm{C}\right) than for bulk tripalmitin (TC≈39∘C)\left(T_{\mathrm{C}} \approx 39^{\circ} \mathrm{C}\right), which can be attributed to the effect of emulsification on fat nucleation. 42{ }^{42} In bulk oils, nucleation is initiated by the presence of catalytic impurities that are evenly distributed throughout the volume of the oil. Once heterogeneous nucleation promotes crystal formation in one part of the oil, then it rapidly spreads throughout the whole volume of the oil. On the other hand, the oil phase in emulsions is divided into a huge number of droplets, so that the probability of finding an impurity in a particular droplet can be extremely small. In addition, the droplets are effectively isolated from each other, so that crystallization in one droplet is not easily propagated to other droplets. 42{ }^{42}

Fig. 1 DSC thermograms of a bulk tripalmitin (80−5−80∘C)\left(80-5-80{ }^{\circ} \mathrm{C}\right) and b Tripalmitin nanoparticles (37−5−75−5∘C)\left(37-5-75-5{ }^{\circ} \mathrm{C}\right) both cooled or heated at the same rate (10∘C/min).TC\left(10^{\circ} \mathrm{C} / \mathrm{min}\right) . T_{\mathrm{C}} is the crystallization peak of SLN during the first cooling cycle, TCOT_{\mathrm{CO}} the coalescence peak, and TSLNT_{\mathrm{SLN}} the SLN crystallization peak after melting. Tcr,T//T_{\mathrm{cr}}, T_{/ /}, and T//T_{/ /}T//are the melting peaks of the three polymorphs

As a result, emulsified oil usually requires a lower temperature to promote nucleation and crystallization than bulk oil.

When the SLN suspension formed at low temperatures was heated, a number of exothermic and endothermic peaks were observed (Figure 1b), which corresponded somewhat to those observed when crystallized bulk tripalmitin was heated (Figure 1a). An endothermic peak was observed at 43∘C43^{\circ} \mathrm{C}, which can be attributed to melting of the α\alpha-form, an exothermic peak was observed at 47∘C47^{\circ} \mathrm{C}, which can be attributed to recrystallization of the tripalmitin, and a large endothermic peak was observed at 63∘C63^{\circ} \mathrm{C}, which can be attributed to melting of the β\beta-form. These results suggest that the α\alpha-polymorphic form rapidly transformed into the β\beta-form in the emulsified tripalmitin, which was not observed in the bulk tripalmitin.

When the tripalmitin emulsion was cooled for a second time, two exothermic peaks were observed at 19 and 39∘C39^{\circ} \mathrm{C} (Figure 1b). The peak at 19∘C19{ }^{\circ} \mathrm{C} can be assigned to the crystallization of tripalmitin contained within small (stable) droplets, whereas the peak at 39∘C39^{\circ} \mathrm{C} can be assigned to the crystallization of tripalmitin contained within large (coalesced) droplets or free oil. 43−45{ }^{43-45} These results indicate that the suspension is unstable to droplet aggregation during the fat crystallization/melting cycle. In addition, they show that DSC measurements are convenient means of monitoring droplet aggregation in SLN suspensions.

Effect of Cooling Rate on Gelation of SLN

Using different CRs, the gelation phenomenon of the tripalmitin SLN occurred at different temperatures. At greater than or equal to 5∘C/min5^{\circ} \mathrm{C} / \mathrm{min}, however, the sample did not gel and remained fluid during the cooling process. Figure 2 shows the change in G′G^{\prime} during cooling from 37 to 5∘C5^{\circ} \mathrm{C} using CRs of 0.2 and 2∘C/min2{ }^{\circ} \mathrm{C} / \mathrm{min}. The storage elastic modulus (G′)\left(G^{\prime}\right) of the colloidal suspension abruptly increased somewhat below the fat crystallization temperature indicating the formation of a semisolid network. The gelation temperature Tgel T_{\text {gel }} decreased

Fig. 2 Effect of two CRs on storage modulus ( G′G^{\prime} ) of tripalmitin SLN suspensions

from 24 to 12∘C12{ }^{\circ} \mathrm{C} when the CR was increased from 0.2 to 2∘C/min2{ }^{\circ} \mathrm{C} / \mathrm{min}.

DSC measurements were carried out on tripalmitin suspensions cooled at the same CRs as in the rheology measurements (Figure 3). The tripalmitin suspensions were cooled at different CRs, then heated at 10∘C/min10{ }^{\circ} \mathrm{C} / \mathrm{min}, then cooled again at −10∘C/min-10^{\circ} \mathrm{C} / \mathrm{min}. The transition temperatures of the peaks observed during heating of the SLN suspensions were fairly similar for samples cooled at different initial CR, although the areas under the peaks were somewhat different (Figure 3a). An endothermic peak was observed at 42∘C42{ }^{\circ} \mathrm{C}, which can be attributed to melting of the α\alpha polymorph, whose area increased with increasing CR. An exothermic peak was observed at 47∘C47{ }^{\circ} \mathrm{C} because of recrystallization, and an endothermic peak was observed at 62∘C62{ }^{\circ} \mathrm{C} because of the melting of the β\beta-polymorph (Figure 3a). The peaks obtained during a subsequent cooling of the melted SLN suspensions are shown in Figure 3b. Unlike the first cooling cycle, two exothermic peaks were observed: 39 (coalescence peak) and 19∘C19{ }^{\circ} \mathrm{C} (SLN peak). The coalescence peak was observed at the

Fig. 3 DSC thermograms of tripalmitin SLN suspensions supercooled at different CRs during a heating and b\mathbf{b} subsequent cooling. The samples were heated then cooled at the same rate (10∘C/min)\left(10^{\circ} \mathrm{C} / \mathrm{min}\right) after the initial crystallization

Fig. 4 Effect of CR on amounts of α\alpha structure (ΦB)\left(\Phi_{\mathrm{B}}\right) and stable SLN crystal (ΦC)\left(\Phi_{\mathrm{C}}\right). SLNs were heated and cooled at 10∘C/min10^{\circ} \mathrm{C} / \mathrm{min}. Data were extracted from Figure 3
same temperature as for the bulk tripalmitin crystallization, while the SLN peak was observed at the same temperature as for the emulsified tripalmitin crystallization (Figure 1). With increasing CR, the coalescence peak decreased, while

Fig. 5 Effect of HR of 2,5,102,5,10, and 20∘C/min20^{\circ} \mathrm{C} / \mathrm{min} on a value of the storage modulus (G′)\left(G^{\prime}\right) and b\mathbf{b} gelation temperature TGT_{\mathrm{G}} of SLN. The arrows indicate the gelation temperature ( TGT_{\mathrm{G}} )

Table 1 Melting and crystallization enthalpies of tripalmitin SLNs measured during a cooling process from 37 to 5∘C5^{\circ} \mathrm{C}, a subsequent heating to 75∘C75^{\circ} \mathrm{C} at different HRs, and a second cooling to 5∘C5^{\circ} \mathrm{C}

A rate of 10∘C/min10^{\circ} \mathrm{C} / \mathrm{min} was used for the cooling cycles
a { }^{\text {a }} The sample was cooled from 37 to 5∘C5^{\circ} \mathrm{C} at 10∘C/min10^{\circ} \mathrm{C} / \mathrm{min}
b{ }^{\mathrm{b}} Not detected (i.e., no peak)
the SLN peak increased (Figure 3b). With the highest CR (i.e., 20∘C/min20^{\circ} \mathrm{C} / \mathrm{min} ), the coalescence peak disappeared, while the SLN peak became large, indicating that the emulsions were stable to droplet coalescence during the cool/heat cycle. To obtain more insight into the relationship between the phase transitions of the tripalmitin and the gelation of the SLN suspensions, we calculated the fraction of tripalmitin present in the α\alpha-polymorph (Φα)\left(\Phi_{\alpha}\right) and the fraction of tripalmitin that was stable to coalescence during a cool/heat cycle (ΦC)\left(\Phi_{\mathrm{C}}\right) :
Φα=ΔHα/ΔHC\Phi_{\alpha}=\Delta \mathrm{H}_{\alpha} / \Delta \mathrm{H}_{\mathrm{C}}
ΦC=ΔHSLN/ΔHC\Phi_{\mathrm{C}}=\Delta \mathrm{H}_{\mathrm{SLN}} / \Delta \mathrm{H}_{\mathrm{C}}
where ΔHα,ΔHSLN\Delta \mathrm{H}_{\alpha}, \Delta H_{\mathrm{SLN}}, and ΔHC\Delta H_{\mathrm{C}} are the enthalpies associated with the melting of the α\alpha-polymorphic form, the SLN peak, and the total tripalmitin observed in the DSC scans. Both Φα\Phi_{\alpha} and ΦC\Phi_{\mathrm{C}} increased with increasing CR (Figure 4), which suggested that increasing the CR increased the fraction of tripalmitin that remained in the α\alpha-form and retarded the coalescence of SLN during a cool/heat cycle. These data support the hypothesis that it is the transition of emulsified tripalmitin from the α\alpha - to β\beta-polymorphic forms that promotes droplet aggregation and gel formation.

Effect of Heating Rate on Gelation of SLN

The effect of HRs on the properties of SLN suspensions containing solidified particles that were stable to aggregation was examined. A fast CR (10∘C)\left(10^{\circ} \mathrm{C}\right) was used to prepare the initial SLN suspensions so as to retard the α\alpha - to β\beta polymorphic transition and thereby retarding particle aggregation and gelation (Figures 1 and 3). The change in G′G^{\prime} during heating of SLN suspensions from 5 to 75∘C75^{\circ} \mathrm{C} at different HRs was measured (Figure 5a). The gelation
temperature increased from 18 to 46∘C46^{\circ} \mathrm{C} as HRs increased from 2 to 20∘C/min20^{\circ} \mathrm{C} / \mathrm{min} (Figure 5 b ).

The impact of HRs on the thermal behavior of SLN suspensions was measured using DSC using the following three-cycle process: a cooling cycle from 37 to 5∘C5^{\circ} \mathrm{C} (at −10∘C/min-10^{\circ} \mathrm{C} / \mathrm{min} ), a heating cycle from 5 to 75∘C75^{\circ} \mathrm{C} at different HRs, and a second cooling cycle from 75 to 5∘C5^{\circ} \mathrm{C} (at −10∘C/min-10^{\circ} \mathrm{C} / \mathrm{min} ). The α\alpha melting peak (Tα)\left(T_{\alpha}\right) was absent, and only a single β\beta melting peak (Tβ)\left(T_{\beta}\right) was observed at the lowest HR, i.e., 0.2∘C/min0.2^{\circ} \mathrm{C} / \mathrm{min} (Table 1). The α\alpha melting peak temperature (Tα)\left(T_{\alpha}\right) increased slightly, while its melting enthalpy (ΔHα)\left(\Delta H_{\alpha}\right) increased dramatically from 0 to 84 J/g84 \mathrm{~J} / \mathrm{g} with increasing HR from 0.2 to 10∘C/min10^{\circ} \mathrm{C} / \mathrm{min}. In contrast, the β\beta melting peak enthalpy decreased slightly from 225 to 203 J/g203 \mathrm{~J} / \mathrm{g} with increasing HR from 0.2 to 10∘C/min10^{\circ} \mathrm{C} / \mathrm{min}. Above 10∘C/min10^{\circ} \mathrm{C} / \mathrm{min}, both ΔHα\Delta H_{\alpha} and ΔHβ\Delta H_{\beta} decreased. On a subsequent cooling process, the crystallization enthalpy of the coalescence peak (ΔHCO)\left(\Delta H_{\mathrm{CO}}\right) decreased greatly from about 170(HR=0.2∘C/min)170\left(\mathrm{HR}=0.2^{\circ} \mathrm{C} / \mathrm{min}\right)

Fig. 6 Effect of HR on amounts of α\alpha structure (ΦM)\left(\Phi_{\mathrm{M}}\right) and stable SLN crystal (ΦC)\left(\Phi_{\mathrm{C}}\right). SLN were cooled at 10∘C/min10^{\circ} \mathrm{C} / \mathrm{min}

to 39.5 J/g(HR=5∘C/min)39.5 \mathrm{~J} / \mathrm{g}\left(\mathrm{HR}=5{ }^{\circ} \mathrm{C} / \mathrm{min}\right). At HR≥10∘C/min\mathrm{HR} \geq 10{ }^{\circ} \mathrm{C} / \mathrm{min}, no coalescence peak appeared, but a large peak that represents the supercooled SLN (TSLN)\left(T_{\mathrm{SLN}}\right) was observed at 19.5∘C19.5^{\circ} \mathrm{C}, and its enthalpy increased with increasing HR. Figure 6 summarizes the fraction of tripalmitin present in the α\alpha-polymorph (Φα)\left(\Phi_{\alpha}\right) and the fraction of tripalmitin that was stable to coalescence during a cool/heat cycle (ΦC)\left(\Phi_{\mathrm{C}}\right) after heating at different rates. These results were similar with those obtained at the different CRs, which indicated that the aggregation of the nanoparticles is retarded by cycling at higher HRs. This is reasonable because the faster the HR, the less is the time given for the metastable crystal structure to rearrange its molecules into a more stable form. The DSC and rheology results thus suggest that rapid cooling and heating of SLN suspensions lead to the formation of systems that are stable to particle aggregation and gelation. Nevertheless, one must be extremely careful when using the results of fairly rapid DSC and rheology tests to predict the long-term stability of SLN suspensions. One may also expect differences between the DSC and rheology measurements of some emulsion systems because of the effect of shear, which may increase the rate of particles collision and destabilization. 46{ }^{46} When we took SLN suspensions and cooled them rapidly by placing them in a 5∘C5^{\circ} \mathrm{C} water bath and then heated them rapidly by placing them in a 37∘C37^{\circ} \mathrm{C} water bath, we found that the samples gelled after about 5 min . Although both the CRs and HRs were fast, this did not prevent particle aggregation and gelation, if the sample was left long enough for the polymorphic transition to occur. It is therefore important when studying the stability of SLN suspensions using DSC and rheology measurements to choose HRs and CRs that give results that correspond to the long-term stability of the system. In particular, the HRs and CRs should be relatively slow compared to the rate of the α\alpha - to β\beta-polymorphic transition if DSC and rheology are going to be used to reliably study the stability of SLN suspensions.

Conclusions

We have shown that particle aggregation and gelation are retarded in SLN suspensions by increasing the CRs and HRs in DSC and rheology experiments. The stability of the SLN suspensions was mainly improved by retarding the transition of emulsified tripalmitin from the α\alpha - to β\beta polymorphic forms. The α\alpha-form produces spherical particles that are fully covered by the surfactant, whereas the β\beta-form produces nonspherical particles that are not fully covered by the surfactant and thereby aggregate through hydrophobic attraction. These results suggest that the preparation of stable SLN suspensions requires either retardation of the α\alpha-to- β\beta polymorphic transition (e.g., by changing fat type or adding inhibitors) or prevention of
particle aggregation (e.g., by adding sufficient surfactant to completely cover the nonspherical particle surfaces or by increasing the repulsive interactions between the particles). Finally, this study shows that DSC and rheology can be used to study particle aggregation and gelation in SLN suspensions but that one must be careful to select appropriate CRs and HRs to obtain results that correlate to the long-term behavior of the systems.

Acknowledgments This grant was supported by USDA CSREES Hatch grants (MAS 0911 and MAS 831) and grants by the USDA National Research Initiative Programs (Award number 2005-01357). Additional financial support was provided by the Leifur Eiriksson Foundation, Hrafnkellssjodur, and Rannsoknarnamsjodur, all located in Reykjavik, Iceland.

References

1. P.M. Kris-Etherton, W.S. Harris, L.J. Appel, Circulation 106(21), 2747-2757 (2002)

2. P.M. Kris-Etherton, K.D. Hecker, A. Bonanome et al., Am. J. Med. 113(Suppl 9B), 71S-88S (2002)

3. J.H. Weisburger, Eur. J. Cancer Prev. 11(S2), 1-7 (2002)

4. R.H. Müller, W. Mehnert, J.S. Lucks et al., Eur. J. Pharm. Biopharm. 41, 62-69 (1995)

5. A.J. Almeida, E. Souto, Adv. Drug Deliv. Rev. 59(6), 478-490 (2007)

6. S.A. Wissing, O. Kayser, R.H. Müller, Adv. Drug Deliv. Rev. 56 (9), 1257-1272 (2004)

7. P.M. Bummer, Crit. Rev. Ther. Drug Carrier Syst. 21(1), 1-20 (2004)

8. W. Mehnert, K. Mader, Adv. Drug Deliv. Rev. 47(2-3), 165-196 (2001)

9. R.H. Müller, M. Radtke, S.A. Wissing, Adv. Drug Deliv. Rev. 54 (Suppl 1), S131-S155 (2002)

10. J. Weiss, T. Helgason, D.J. McClements, K. Kristbergsson, E. Decker, T.S. Awad, Food Biophys. (2008 in press)

11. A.A. Attama, C.C. Muller-Goymann, Int. J. Pharm. 334(1-2), 179-189 (2007)

12. C. Himawan, V.M. Starov, A.G.F. Stapley, Adv. Colloid Interface Sci. 122(1-3), 3-33 (2006)

13. R.H. Müller, K. Mader, S. Gohla, Eur. J. Pharm. Biopharm. 50(1), 161−177161-177 (2000)

14. E. Dickinson, D.J. McClements, M.J.W. Povey, J. Colloid Interface Sci. 142(1), 103-110 (1991)

15. D.J. McClements, Food Emulsions: Principles, Practice, and Techniques, 2nd edn. (CRC, Boca Raton, 2005)

16. D.J. McClements, E. Dickinson, S.R. Dungan, J.E. Kinsella, J.G. Ma, M.J.W. Povey, J. Colloid Interface Sci. 160(2), 293-297 (1993)

17. J.N. Coupland, Curr. Opin. Colloid Interface Sci. 7(5-6), 445-450 (2002)

18. P. Ahlin, J. Kristl, J. Smid-Korbar, Acta Pharm. 48, 257-267 (1998)

19. H. Bunjes, M.H. Koch, J. Control. Release 107(2), 229-243 (2005)

20. H. Bunjes, M.H. Koch, K. Westesen, J. Pharm. Sci. 92(7), 15091520 (2003)

21. B. Siekmann, K. Westesen, Colloids Surf. B: Biointerfaces 3(3), 159−175159-175 (1994)

22. K. Westesen, B. Siekmann, Int. J. Pharm. 151(1), 35-45 (1997)

23. C. Freitas, R.H. Müller, Int. J. Pharm. 168(2), 221-229 (1998)

24. C. Freitas, R.H. Müller, Eur. J. Pharm. Biopharm. 47(2), 125-132 (1999)

25. H. Bunjes, F. Steiniger, W. Richter, Langmuir 23(7), 4005-4011 (2007)

26. S. Hatziantoniou, G. Deli, Y. Nikas, C. Demetzos, G.T. Papaioannou, Micron 38(8), 819-823 (2007)

27. M.A. Schubert, M. Harms, C.C. Müller-Goymann, Eur. J. Pharm. Sci. 27(2-3), 226-236 (2006)

28. M.A. Schubert, C.C. Müller-Goymann, Eur. J. Pharm. Biopharm. 61(1-2), 77-86 (2005)

29. T. Helgason, T.S. Awad, E. Decker, K. Kristbergsson, D.J. McClements, J. Weiss, J Am Oil Chem Soc (2008 in press)

30. M.J.W. Povey, T.S. Awad, R. Hue, Y. Ding, ed. by E. Dickinson, M. Leser. Food Colloids: Self-Assembly and Material Science (the Royal Society of Chemistry, Cambridge, UK, 2007), pp. 399-412

31. T.S. Awad, K. Sato, Colloids Surf. B: Biointerfaces 25(1), 45-53 (2002)

32. T.S. Awad, K. Sato, J. Am. Oil Chem. Soc. 78(8), 837-842 (2001)

33. T.S. Awad, K. Sato, ed. by A.G. Marangoni, S. Narine. Physical Properties of Lipids (Marcel Dekker, New York, 2002), pp. 37-62

34. T.S. Awad, Food Res. Int. 37(6), 579-586 (2004)

35. K. Golemanov, S. Tcholakova, N.D. Denkov, T. Gurkov, Langmuir 22(8), 3560-3569 (2006)

36. H. Bunjes, K. Westesen, M.H.J. Koch, Int. J. Pharm. 129, 159173 (1996)

37. H. Bunjes, M.H.J. Koch, K. Westesen, Langmuir 16, 5234-5241 (2000)

38. T. Unruh, H. Bunjes, K. Westesen, M.H.J. Koch, J. Phys. Chem. B 103, 10373-10377 (1999)

39. K. Boode, C. Bisperink, P. Walstra, Colloids Surf. 61, 55-74 (1991)

40. K. Boode, P. Walstra, Colloids Surf. A: Physicochem. Eng. Asp. 81, 121-137 (1993)

41. K. Boode, P. Walstra, A.E.A. Degrootmostert, Colloids Surf. A: Physicochem. Eng. Asp. 81, 139-151 (1993)

42. P. Walstra, Physical Chemistry of Foods (Marcel Decker, New York, NY., 2003)

43. P. Thanasukarn, R. Pongsawatmanit, D.J. McClements, Colloids Surf. A: Physicochem. Eng. Asp. 246(1-3), 49-59 (2004)

44. P. Thanasukarn, R. Pongsawatmanit, D.J. McClements, Food Hydrocoll. 18(6), 1033-1043 (2004)

45. S.A. Vanapalli, J. Palanuwech, J.N. Coupland, Colloids Surf. A: Physicochem. Eng. Asp. 204(1-3), 227-237 (2002)

46. S.A. Vanapalli, J.N. Coupland, Food Hydrocoll. 15(4-6), 507-512 (2001)