Food Research International 67 (2015) 25–34

Contents lists available at ScienceDirect

Food Research International

j ourna l homepage: www.e lsev ie r .com/ locate / foodres

Crystallization behavior of monoacylglycerols in a hydrophobic and ahydrophilic solvent

Stefanie Verstringe ⁎, Kim Moens, Nathalie De Clercq, Koen DewettinckLaboratory of Food Technology and Engineering, Ghent University, Coupure Links 653, 9000 Ghent, Belgium

⁎ Corresponding author. Tel.: +32 92 64 61 68; fax: +E-mail address: Stefanie.Verstringe@UGent.be (S. Vers

http://dx.doi.org/10.1016/j.foodres.2014.10.0270963-9969/© 2014 Elsevier Ltd. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 29 July 2014Accepted 31 October 2014Available online 6 November 2014

Keywords:MonoacylglycerolCrystallizationMicrostructureSynchrotron radiation X-ray diffractionDifferential scanning calorimetryCryo-SEM

Systems containing monoacylglycerols (MAGs) in both a hydrophobic solvent (liquid oil) and a hydrophilicsolvent (water) can be used for the development of calorie-reduced food products. In this study, the crystalliza-tion behavior ofMAGs in a hydrophobic solvent (rapeseed oil) and a hydrophilic solvent (water)was studied andcompared. Pure monopalmitin (MP) and a commercial MAG containing MPwere used for this study. Differentialscanning calorimetry (DSC) data were coupled with X-ray diffraction (XRD) data obtained using synchrotronradiation and cryo-scanning electronmicroscopy (cryo-SEM) images were recorded to illustrate themicrostruc-tural characteristics of the systems. Although the polymorphic behavior of theMAGswas found to be the same inboth solvents, the crystallization onset temperature was found to be concentration-dependent in the systemswith liquid oil as solvent, in contrast to the systems with water as solvent. On the other hand, the temperatureof the polymorphic transition from the α to the sub-α polymorph was constant in both systems. Differences inmicrostructure could be attributed to the inherent properties of the hydrophobic or hydrophilic solvent.

© 2014 Elsevier Ltd. All rights reserved.

1. Introduction

The addition ofMAGs to liquid oil is interesting froma practical pointof view as they represent a strategy to structure oil without using solidfat, which implicates health advantages (Pernetti, vanMalssen, Flöter, &Bot, 2007; Rogers, 2009). However, the behavior of MAGs in liquid oil isscarcely documented in literature. In fact, most authors assume that thephase behavior of MAGs in liquid oil is similar to that in water.

A characteristic feature of MAGs in water, which is abundantlydescribed in literature, is their ability to form lyotropic mesophases orliquid crystals. MAGs crystallize in bilayers and are normally insolublein water. However, when a mixture of MAGs and water is heated, thehydrocarbon chains become liquid at a certain temperature while atthe same time water penetrates between the bilayers along the planeof the glycerol head groups. This results in the formation of liquid crys-talline phases. The introduction of the liquid chain concept by Chapman(1958) was an important step towards the elucidation of these phases.He demonstrated the phenomenon of the melting of hydrocarbonchains. This occurs because the van derWaals forces between hydrocar-bon chains are weaker than the hydrogen bonding between the polarhead groups (Nawar, 1996). Only a few years later, Luzzati, Mustacchi,Skoulios, and Husson (1960) revealed the structure of the most com-mon liquid crystalline phases.

Several types of liquid crystalline phases are possible dependent onthe MAG/water ratio and the temperature. Just above the melting

32 92 64 62 18.tringe).

point of the MAG, a lamellar mesophase is formed. This rather fluidmesophase consists of lipid bilayers alternated by water layers. This isschematically shown in Fig. 1A. If the water content is raised abovethe swelling limit of the lamellar mesophase, a transformation intospherical multilamellar vesicles (liposomes) occurs, also called a lamel-lar dispersion (Stauffer, 2005). The formation of closed vesicles withpreserved lamellar structure avoids direct contact between bulk watermolecules and water molecules associated with the lipid bilayers(Krog, 1997). At higher temperatures, a cubic mesophase and a hexago-nal mesophase are formed.

The interlayer spacing in the lamellar phase of non-ionic MAGs (dWin Fig. 1A) is determined by the balance between the long-range vanderWaals forces and the osmotic repulsion between the lipid bilayers. Inthis way, a water layer thickness of about 20 Å is reached in various sys-tems (Larsson&Krog, 1973). As investigated byVan deWalle, Goossens,and Dewettinck (2008), introducing charged groups on the surface ofthe lipid bilayers leads to an increased swelling (thicker water layers),a higher temperature stability of the lamellar phase (the transition tothe cubic phase is shifted to higher temperatures) and an expansion ofthe lamellar phase into the higher water content region. This is due tothe creation of electrical repulsive forces (Van de Walle et al., 2008).

If the lamellar phase or dispersion is cooled, a gel phase is formedwith an ointment-like consistency (Heertje, Roijers, & Hendrickx,1998). The gel phase is called the α-gel as it is the hydrated form oftheα polymorph. Theα-gel also has a lamellar structure but the hydro-carbon chains are now in a crystalline state, as schematically shown inFig. 1B. The gel consistency of this phase is due to the possibility of thelipid bilayers to slide in relation to one another (Larsson, 1967). The

Fig. 1. Schematic representation of (A) the lamellar mesophase and (B) the α-gel phase of a MAG/water mixture.

Fig. 2. Schematic representation of the bilayer structure of MAGs blended with (A) liquidoil and (B) water.

26 S. Verstringe et al. / Food Research International 67 (2015) 25–34

transition temperature from lamellar to gel phase is called the Kraffttemperature.

Phases of MAG/water systems can find use as main components inlow-calorie food products (Heertje et al., 1998). Similarly, MAGsadded to liquid oil can aid in calorie reduction of food products(Pernetti et al., 2007; Rogers, 2009). Most authors assume that thephase behavior of MAGs in liquid oil is similar to that in water. Surpris-ingly, only a limited number of authors have investigated MAG/liquidoil systems in detail. Ojijo, Neeman, Eger, and Shimoni (2004) reportedthe possible use of MAG/liquid oil systems as a healthy substitute formargarine and butter due to the formation of a gel network uponcooling. This gel networkwas assumed to be anα-crystalline gel, similarto aqueous systems. Moreover, a liquid crystalline lamellar phase wasassumed to exist before gel formation during cooling. A coagelmesophase was found to form during storage (Ojijo, Kesselman, et al.,2004). Chen, Van Damme, and Terentjev (2009) conducted a moredetailed investigation of the phase behavior of a commercial C18 MAGin hazelnut oil. Their proposed phase diagram displays three phases.On cooling the well-mixed liquid isotropic phase, an inverse lamellarphase is formed with the hydrophobic head groups now situated inthe middle of the bilayer. This is schematically shown in Fig. 2 whichshows the bilayer structure in both liquid oil and water. This phase rhe-ologically behaves like an elastic gel, as opposed to the fluid behavior ofthe aqueous lamellar phase but similar to the aqueous α-gel known atlower temperatures. Based on WAXD measurements, it was deducedthat the glycerol heads are packed in a hexagonal manner in the inverselamellar bilayer. The second phase transition from inverse-lamellar to a

crystalline phase includes the crystallization of the aliphatic chains in asub-α form with an orthorhombic packing. This phase retains themechanical properties of the gel phase.

MAG/water systems andMAG/liquid oil systems are both promisingblends for the development of calorie-reduced food products. Althoughthe behavior of both systems is often said to be comparable, a lot of un-certainty still remains in this matter. This paper seeks to redress the

Fig. 3.WAXDdiffraction patterns of ROduring cooling from80 °C to−40 °C at−10 °C/min.

Fig. 4. DSC crystallization curves of RO blended with 8% MP and 8% Myverol recordedduring cooling from 80 °C at−10 °C/min.

27S. Verstringe et al. / Food Research International 67 (2015) 25–34

situation by elucidating differences and similarities between the behav-ior of a saturatedMAG in liquid oil versuswater as solvent. To the best ofour knowledge, no study exists that compares the behavior of the sameMAG in both liquid oil and water under the same conditions.

2. Experimental section

2.1. Materials

Pure monopalmitin (N99%, MP) was obtained from Nu-Chek Prep(Elysian, USA). Myverol™ 18 04-PK (referred to as Myverol throughoutthe paper) was obtained from Kerry Group (Ireland). This commercialemulsifier contained 60.17 ± 0.22% palmitic acid and 37.29 ± 0.50%stearic acid as determined by gas chromatography. According to thesupplier, Myverol contained a minimum of 95% MAGs. Rapeseed oil(RO) was purchased at a local supermarket.

2.2. Preparation of the blends

One concentration of MP (8% w/w) and different concentrations ofMyverol (8, 25, 50 and 75% w/w) were dispersed in RO and stirredwith a magnetic stirrer at 70 °C until a homogeneous sample wasobtained. When the blend was visibly free of dispersed material, itwas further mixed for at least 30 min. 8% MP and different concentra-tions of Myverol (8, 25, 50 and 75% w/w) were added to distilledwater and held for 30 min in a waterbath at 60 °C while vortexing reg-ularly. In this way, a well-mixed lamellar dispersion was formed. Theblends were stored at −24 °C until analysis.

2.3. Differential scanning calorimetry (DSC)

The DSC experiments were performed with a Q1000 DSC with a re-frigerated cooling system and an autosampler system (TA Instruments,NewCastle, USA). The DSCwas calibratedwith indium (TA Instruments,New Castle, USA), azobenzene (Sigma-Aldrich, Bornem, Belgium) andundecane (Acros organics, Geel, Belgium) prior to analysis. Nitrogenwas used to purge the system. Between 5 and 10mg of samplewas her-metically sealed in an aluminum pan and an empty pan was used as areference. Non-isothermal crystallization curves were recorded bycooling the samples at −10 °C/min to −40 °C after a holding time of10 min at 80 °C. The DSC profiles were analyzed with the UniversalAnalysis software version 4.7A (TA Instruments, New Castle, USA).Each analysis was executed in triplicate. Crystallization onset tempera-tureswere determined by the intersection of the baselinewith the abso-lute highest tangent of the crystallization curve and the average onsettemperature was calculated.

2.4. Time-resolved synchrotron X-ray diffraction (XRD)

Polymorphic behavior of the blends during cooling was investigatedby XRD using synchrotron radiation. Small-angle X-ray scattering(SAXS) and wide-angle X-ray diffraction (WAXD) measurements wereperformed on the Dutch–Belgian (DUBBLE) beamline BM26B at theEuropean Synchrotron Radiation Facility (ESRF) in Grenoble (France).The experiments were performed at a fixed wavelength λ of 1.24 Å. Alinear 300 K photon counting Pilatus detector was used for WAXD,whereas a large area high sensitive photon counting 2D Pilatus detectorwas used to collect the SAXS images. The sampleswere enclosed in glasscapillaries and the temperature was controlled by a Linkam hot stage.The RO blends were cooled at −10 °C/min to −20 °C after holding at80 °C for 10 min. This holding temperature ensures complete meltingand erases the crystal memory. The water blends were cooled at−10 °C/min to −20 °C starting from a lamellar dispersion by holdingthe samples at 60 °C for 10 min. Scattering patterns were taken every2 °C during cooling. Known reflections of standard silver-behenateand alpha aluminum samples were used to calibrate the SAXS and

WAXD scattering angles, 2θ. The patterns are presented as a functionof d (Å), with d = λ/2sin(θ). All scattering patterns were corrected forthe detector response, normalized to the intensity of the primarybeam and corrected by the sample absorption before performing thebackground subtraction.

2.5. Cryo-scanning electron microscopy (cryo-SEM)

The microstructure of the blends at room temperature (20 °C) wasvisualized using scanning electron microscopy. After preparation ofthe blends (Section 2.2), a small amount of samplewas cooled in a ther-mostatic cabinet at 20 °C. A small piece of the samplewas thenmountedon an aluminum stub, quickly frozen in slushed nitrogen (−210 °C),fractured, sputter-coated with platinum and observed with a Jeol JSM-7100 F scanning electron microscope (Jeol (Europe) B.V., Zaventem,Belgium).

Fig. 5. (A) WAXD and (B) SAXS diffraction patterns of RO containing 8% MP during cooling from 80 °C at−10 °C/min.

28 S. Verstringe et al. / Food Research International 67 (2015) 25–34

2.6. Statistical analysis

To detect statistically significant differences (p b 0.05), one-wayANOVA was performed using SPSS Statistics 21. To identify the differ-ences, Tukey's test was used in the case of equal variances whileDunnett's T3 test was used when the variances were not equal.

3. Results and discussion

3.1. Crystallization behavior of monoacylglycerols in a hydrophobic solvent

MP and Myverol were added to RO in order to gain insight into thecrystallization behavior ofMAGs in a hydrophobic solventwithout crys-tallizing triacylglycerols (TAGs) present. XRDmeasurements confirmedthat RO only started to crystallize below −20 °C. This is illustrated in

Fig. 6. (A) WAXD and (B) SAXS diffraction patterns of RO contai

Fig. 3 showing WAXD diffraction patterns recorded during cooling ofRO. The first crystallization peak at 4.16 Å appeared when the tempera-ture reached −22 °C. Fig. 4 shows the DSC crystallization profilesrecorded during cooling of RO blended with 8% MP and 8% Myverol.The Myverol blend showed a significantly earlier crystallization onset(52.62 ± 0.19 °C compared to 49.12 ± 0.20 °C for the MP blend,p b 0.05), probably due to the presence of high-melting monostearin(melting point of the α polymorph of pure monostearin is 76.7 ±0.4 °C compared to 71.3 ± 0.4 °C for monopalmitin) in the commercialemulsifier. When the onset temperatures of the MAGs in RO are com-pared with those in palm oil (PO) determined in previous studies(Verstringe, Danthine, Blecker, Depypere, & Dewettinck, 2013;Verstringe, Danthine, Blecker, & Dewettinck, 2014), i.e. 43.4 ± 0.3 °Cfor MP in PO and 45.49± 0.26 °C for Myverol in PO, it can be concludedthat the MAGs started to crystallize earlier in RO than in PO. The highersolubility in PO can be attributed to the higher degree of similarity

ning 8% Myverol during cooling from 80 °C at−10 °C/min.

Fig. 7. Cryo-SEM images of RO containing 8% MP taken at 20 °C after cooling from 70 °C in a thermostatic cabinet at 20 °C. Crystal structures are embedded in amorphous RO (aRO).

29S. Verstringe et al. / Food Research International 67 (2015) 25–34

between the monoacylglycerols, both containing monopalmitin, andthe fatty acids present in PO.

Similar to the pureMAG substances (Verstringe et al., 2014), the firstcrystallization peak corresponds with the crystallization in the α poly-morph while the second crystallization peak corresponds with thetransformation of theα crystals to the sub-α polymorph. This transfor-mation temperature is much lower in the case of the commercial MAGas compared to the pure monopalmitin. This is probably caused bymo-lecular incompatibility of the different MAGs present in a commercialMAG.

The polymorphic behavior was confirmedwith XRDmeasurements.Figs. 5 and 6 showWAXDand SAXSdiffraction patterns recorded duringcooling to 0 °C. When the samples were in the molten state, a small,

Fig. 8. DSC crystallization curves of 8% MP in RO and 8% MP in water recorded duringcooling from 80 °C at−10 °C/min.

broad SAXS peak around d = 25 Å was present. This is indicative of acertain degree of liquid ordering. The 8% MP blend started to crystallizein αwhen the temperature reached 50 °C, as evidenced by the appear-ance of a WAXD peak at d = 4.17 Å and a SAXS peak at d = 46.2 Å. At38 °C, the transformation to sub-α started, as seen by the appearanceof typical WAXD peaks (d = 4.21 Å, d = 3.53 Å, d = 3.70 Å and d =3.86 Å) and the decrease of the long spacing value to d = 45.5 Å(Lutton & Jackson, 1948). In the case of Myverol, the WAXD patternsshowed two peaks for the α polymorph (4.19 and 4.22 Å), probably at-tributable to the presence of two main MAGs (monostearin andmonopalmitin) in Myverol. The crystallization of the α polymorphstarted at 52 °C while the transformation to sub-α occurred at 12 °C,as clear from the appearance of typical WAXD peaks at d = 4.21 Å,

Fig. 9. DSC crystallization curves of 8% Myverol in RO and 8% Myverol in water recordedduring cooling from 80 °C at−10 °C/min.

Fig. 10. (A) WAXD and (B) SAXS diffraction patterns of water containing 8% MP during cooling from 60 °C at −10 °C/min.

30 S. Verstringe et al. / Food Research International 67 (2015) 25–34

d=3.67Å, d=3.81 Å and d=3.95Å and the decrease of the long spac-ing value from d = 51.2 Å to d = 49.9 Å.

The microstructure of the RO blend containing 8% MP at roomtemperature (20 °C) was visualized using cryo-SEM. At ambient tem-peratures, the sample contains both solid and liquid phases. Duringthe cryo-treatment, the sample is plunged into slushed nitrogen(−210 °C) which effectuates a very rapid cooling. This rapid coolingpreserves the solid crystal structure and converts the liquid into a non-crystalline, vitreous or amorphous phase (Jewell & Meara, 1970). Afterfracturing, the crystal structure can be visualized using SEM. Fig. 7shows SEM images of the 8% MP blend crystallized at 20 °C. It can beseen that the MP crystals, embedded in amorphous RO, are composedof layers which seem to form needle-like lamellar structures. The thick-ness of the layers is of the order of 10 nm.

Fig. 11. (A) WAXD and (B) SAXS diffraction patterns of water cont

3.2. Crystallization behavior of monoacylglycerols in a hydrophilic solvent

Only a limited number of authors have investigated MAG/liquid oilsystems in detail (Chen et al., 2009). Most authors assume that thephase behavior of MAGs in liquid oil is similar to that in water (Ojijo,Kesselman, et al., 2004; Ojijo, Neeman, et al., 2004). In order to investi-gate similarities and dissimilarities in the phase behavior of MAGs inboth liquid oil and water, 8% MP and 8% Myverol were added to waterand their crystallization behavior was compared with that of the ROblends, as discussed in the previous section. Figs. 8 and 9 show anoverlay of the DSC crystallization curves of 8% of theMAGs in both liquidsystems. In RO, the MAGs crystallize in α after which they transform tosub-α, giving rise to two exothermic peaks in the DSC crystallizationprofile. The α crystallization onset temperature is 49.12 ± 0.20 °C for

aining 8% Myverol during cooling from 60 °C at−10 °C/min.

Fig. 12. Cryo-SEM images of water containing 8% MP taken at 20 °C after cooling from60 °C in a thermostatic cabinet at 20 °C. Crystal structures are embedded in amorphouswater (aW).

31S. Verstringe et al. / Food Research International 67 (2015) 25–34

MP and 52.62± 0.19 °C for Myverol. The transformation onset temper-ature is 37.32± 0.07 °C forMPwhile it is 12.2±0.4 °C forMyverol. Thesystems in water also show two exothermic peaks. The first peak has anonset of 48.7 ± 0.4 °C for MP and 52.71 ± 0.23 °C for Myverol. Thesecond exothermic peak has an onset of 14.2 ± 0.5 °C for MP whilethe onset is 2.82 ± 0.11 °C in the case of Myverol.

WAXD and SAXS measurements during cooling confirmed thatthese two peaks can be attributed to the transition to α and sub-αrespectively, as illustrated in Figs. 10 and 11. At 60 °C, the WAXD pat-terns showed a weak and relatively diffuse peak at about 3.3 Å, similarto a pure water sample (result not shown). Although no diffractionpeaks were present in the WAXD patterns, a small SAXS peak at52.1 Å for MP and 53 Å for Myverol was present which shifted to highervalues as the temperature decreased. In this temperature range, thistype of MAG/water blends is known to form a lamellar mesophaseconsisting of lipid bilayers alternated by water layers. The detectedlong spacing coincideswith the thickness (d) including theMAGbilayer(dMAG) and the water layer separating one lipid layer from another(dW). This is schematically shown in Fig. 1A. Typically, a water layerthickness of about 20 Å is reached in various systems (Larsson & Krog,1973). This means that the thickness of the MAG bilayer probably isabout 30 Å. This value is in the same range as the one found in themol-ten RO blends containing 8% MP and 8% Myverol (25 Å). This indicatesthat these lipid systems probably have a bilayer ordering of the MPmolecules in the molten state, which is comparable to the mesophases

observed in MAG/water systems. The shift of the long spacing peak ofthe lamellar mesophase to higher values with decreasing temperaturecan be attributed to a decreased flexibility and rearrangement of thelipid tails (Mezzenga et al., 2005).

When the temperature of the MP blend decreased to 48 °C, a SAXSpeak at d=58.6 Å and aWAXD peak at d=4.24 Å appeared, indicatingthe formation of a hexagonally ordered α polymorph. Similarly, a SAXSpeak at d = 63 Å and a WAXD peak at d = 4.24 Å appeared when thetemperature of the Myverol blend reached 54 °C. These temperaturesare comparable to the onset temperatures as determined by DSC(48.7±0.4 °C forMP and 52.71±0.23 forMyverol). This phase is calledtheα-gel phase andhas a lamellar structure but the hydrocarbon chainsare now in a crystalline state, as schematically shown in Fig. 1B. Only 8%of MAGs is present in these blends, so more water is present than theamount needed to form a water layer of about 20 Å between the differ-ent lipid bilayers. When this is the case, closed particles with a lamellarstructure (liposomes) are formed and the excesswater forms a separatephase. In this way, direct contact between bulk water molecules andwater molecules associated with lipid bilayers is avoided (Krog, 1997).Cryo-SEM enabled to visualize these liposomes, as shown in the imagesin Fig. 12 for the 8% MP blend. It should be noted that, although theformation of liposomes is a generally accepted given, visualization ofsuch structures in systems containing saturated MAGs is rarely foundin literature. Heertje et al. (1998) visualized onion-shaped multi-lamellar vesicles with a diameter of approximately 2.5 μm in the α-gelphase of a commercial saturated MAG. In a system containingpolyglycerol fatty acid esters (mainly palmitic and stearic acid), closed,spherical to ellipsoidal shells with a diameter of approximately 5 μmwere visualized (Duerr-Auster et al., 2007). As also remarked byLarsson and Krog (1973), the formation of closed vesicles in the α-gelphase is quite remarkable, as there should be a considerable strainagainst curved layers with ordered hydrocarbon chains.

As temperature decreased, the SAXS peak showed a small shift tolower values, indicating a shrinking of the structure (Figs. 10 and 11).When the temperature reached 12 °C in the MP blend and 4 °C in theMyverol blend, the SAXS peak shifted to a lower value (45.5 Å for MPand 54Å forMyverol) and the appearance of extraWAXDpeaks indicat-ed the transformation of theα crystals to sub-α. This coincides with theappearance of the second peak in the DSC crystallization profile (Figs. 8and 9). The onset temperature for this transformation as determined byDSCwas 14.2± 0.5 °C forMP and 2.82± 0.11 for Myverol. For bothMPand Myverol, the transformation to sub-α occurs later in water than inRO. This indicates a lower mobility of the crystallized MAGs in water,probably caused by the presence of hydrogen bridges between thehydrophilic heads of the MAGs and the water in between the lipidbilayers.

The behavior of theMAGs thus shows similarities and dissimilaritiesin liquid oil and in water. In themolten state as well as in the crystallinestate, theMAGmolecules are ordered in bilayers in both systems. How-ever, the bilayers are reversed in liquid oil. This is schematically shownin Fig. 2 which shows the bilayer structure in both systems. Moreover,different microstructures are formed: the layered structures in liquidoil form needle-like structures, while the layers are arranged in closedparticles in the water systems. With regard to polymorphism, initialcrystallization in α is followed by a transformation to sub-α in bothsystems.

3.3. Influence of concentration on the crystallization behavior of Myverol

In previous studies, it was shown that a higher concentration ofMAGs crystallized earlier in PO (Verstringe et al., 2013, 2014). Followingthe Hildebrand equation, the melting point of a high-melting compo-nent is reduced when blended with lower-melting components (Zhou& Hartel, 2006). In order to investigate concentration-dependence ofthe crystallization temperature in liquid oil and water systems, threeconcentrations of Myverol (25, 50 and 75%) were added to both

32 S. Verstringe et al. / Food Research International 67 (2015) 25–34

Fig. 14. Cryo-SEM images of RO containing (A) 25%Myverol, (B) 50%Myverol and (C) 75%Myverol and water containing (D) 25%Myverol, (E) 50%Myverol and (F) 75%Myverol taken at20 °C after cooling from 70 °C for the ROblends and 60 °C for thewater blends in a thermostatic cabinet at 20 °C. Crystal structures are embedded in amorphous RO (aRO) and amorphouswater (aW).

33S. Verstringe et al. / Food Research International 67 (2015) 25–34

solvents and the blends were investigated with XRD. Fig. 13 shows theSAXS patterns of the six blends. It can be noticed that the d-spacingsdo not change when the concentration changes. Similar to PO systems,the α crystallization onset temperature in RO increased with Myverolconcentration, from 58 °C for 25% Myverol to 60 °C for 50% Myveroland 64 °C for 75% Myverol. This can be attributed to the solubility ofthe high-melting component in the liquid oil, as a higher concentrationstarts to crystallize earlier. In contrast, α crystallization started around50 °C in all water blends. This transition temperature is called the Kraffttemperature and is known to be independent on concentration. This isprobably attributable to the incompatibility of MAGs with water.

With respect to theα to sub-α transition temperature, it can be no-ticed that this temperature remained constant at 12 °C in the RO blends

Fig. 13. SAXS diffraction patterns of RO containing (A) 25% Myverol, (B) 50% Myverol and (CMyverol, (E) 50% Myverol and (F) 75% Myverol during cooling from 60 °C at −10 °C/min.

while it remained constant at 2 °C in the water blends. Note that thesevalueswere also found for the blends containing 8%Myverol. The trans-formation to sub-α thus occurred later when the MAGwas dissolved inwater compared to liquid oil. It can be concluded that the crystallizationonset temperature of MAGs depends on concentration when it isdissolved in liquid oil, while it is concentration-independent whenwater is the solvent. Once crystallized, the transformation to sub-α isconcentration-independent in both systems. These findings confirmthose of Chen et al. (2009) who conducted a detailed study of thephase behavior of a commercial C18 MAG in hazelnut oil. They foundthat an inverse lamellar phase with a hexagonal ordering was formedon cooling the well-mixed liquid isotropic phase. When the tempera-ture was further decreased, the inverse lamellar phase transformed

) 75% Myverol during cooling from 80 °C at −10 °C/min and water containing (D) 25%

34 S. Verstringe et al. / Food Research International 67 (2015) 25–34

into a sub-α crystalline phase. They reported no signs of ordering in themolten state, aswas observed in the SAXS patterns in this study. Similarto our observations, α crystallization was followed by a transformationto sub-αwith a decreasing onset temperature ofα crystallizationwhenthe concentration decreased and a constant temperature of sub-αtransformation.

The microstructure of the six blends as visualized by cryo-SEM isillustrated in Fig. 14. All three concentrations of Myverol formed openlamellar structures in RO. 25%Myverol formed closed lamellar particlesinwater, similar to the liposomes observed in a blend of 8%MP inwater.In contrast, in the former blend the closed particles were not perfectlyspherical but seemed compressed due to the high concentration ofMyverol present and thus the high amount of closed particles. When50% Myverol was present in water, both closed particles and openlamellar structures were present. At a concentration of 75% in water,predominantly open lamellar structures could be observed. In thiscase, all the water molecules present are associatedwith the lipid bilay-ers and no separation between two different water phases is necessary.

4. Conclusion

The comparison between liquid oil and water as solvent for theMAGs revealed similarities and dissimilarities. In both systems, α crys-tallization was followed by a transformation to the sub-α polymorph.However, the crystallization onset temperature was concentration-independent in water, as opposed to the liquid oil systems, probablydue to the incompatibility of the MAGs with water. Once crystallized,the temperature of transformation to sub-α was concentration-independent in both systems. Cryo-SEM revealed differences andsimilarities in microstructures formed in the two solvents. The MAGscrystallized in bilayers forming lamellar structures in both systems.However, when a low MAG concentration was present in water, directcontact between bulkwater andwater associatedwith the lipid bilayerswas avoided by the formation of closed lamellar particles. The behaviorofMAGs in liquid oil and inwater is thus largely similarwith differencesattributable to the incompatibility of MAGs with water.

Acknowledgment

Stefanie Verstringe is a research assistant of the Fund for ScientificResearch — Flanders (F.W.O. — Vlaanderen). The authors furtheracknowledge the ESRF (Grenoble, France) for the use of the synchrotronfacilities. The Dutch–Belgian Beamline (DUBBLE) research group at theESRF and the Dutch organization for scientific research (N.W.O.) areacknowledged for their help and continuous support of the DUBBLEproject (ESRF, Grenoble, France). The Hercules foundation is acknowl-edged for its financial support in the acquisition of the scanning electronmicroscope JEOL JSM-7100F equipped with cryo-transfer system

Quorum PP3000T and Oxford Instruments Aztec EDS (grant numberAUGE-09-029). Benny Lewille is greatly acknowledged for his assistancewith the experiments.

References

Chapman, D. (1958). An infrared spectroscopic examination of some anhydrous sodiumsoaps. Journal of the Chemical Society, 33, 784–789.

Chen, C.H., Van Damme, I., & Terentjev, E.M. (2009). Phase behavior of C18 monoglyceridein hydrophobic solutions. Soft Matter, 5(2), 432–439.

Duerr-Auster, N., Kohlbrecher, J., Zuercher, T., Gunde, R., Fischer, P., &Windhab, E. (2007).Microstructure and stability of a lamellar liquid crystalline and gel phase formed bya polyglycerol ester mixture in dilute aqueous solution. Langmuir, 23(26),12827–12834.

Heertje, I., Roijers, E.C., & Hendrickx, H. (1998). Liquid crystalline phases in the structuringof food products. LWT—Food Science and Technology, 31(4), 387–396.

Jewell, G.G., & Meara, M.L. (1970). A new and rapid method for the electron microscopicexamination of fats. Journal of the American Oil Chemists Society, 47(12), 535–538.

Krog, N.J. (1997). Food emulsifiers and their chemical and physical properties. In S.E.Friberg, & K. Larsson (Eds.), Food emulsions (pp. 112–188). New York: Marcel DekkerInc.

Larsson, K. (1967). The structure of mesomorphic phases andmicelles in aqueous glyceridesystems. Zeitschrift für Physikalische Chemie, 56(3_4), 173–198.

Larsson, K., & Krog, N. (1973). Structural properties of the lipid—Water gel phase.Chemistry and Physics of Lipids, 10(2), 177–180.

Lutton, E.S., & Jackson, F.L. (1948). The polymorphism of 1-monostearin and 1-monopalmitin. Journal of the American Chemical Society, 70(7), 2445–2449.

Luzzati, V., Mustacchi, H., Skoulios, A., & Husson, F. (1960). La structure des colloidesd'association. I. Les phases liquide-cristallines des systemes amphiphile-eau. ActaCrystallographica, 13(8), 660–667.

Mezzenga, R., Meyer, C., Servais, C., Romoscanu, A.I., Sagalowicz, L., & Hayward, R.C.(2005). Shear rheology of lyotropic liquid crystals: A case study. Langmuir, 21(8),3322–3333.

Nawar, W.W. (1996). Lipids. In O.R. Fennema (Ed.), Food chemistry (pp. 239–319). NewYork: Marcel Dekker Inc.

Ojijo, N.K.O., Kesselman, E., Shuster, V., Eichler, S., Eger, S., Neeman, I., et al. (2004). Chang-es in microstructural, thermal, and rheological properties of olive oil/monoglyceridenetworks during storage. Food Research International, 37(4), 385–393.

Ojijo, N.K.O., Neeman, I., Eger, S., & Shimoni, E. (2004). Effects of monoglyceride content,cooling rate and shear on the rheological properties of olive oil/monoglyceride gelnetworks. Journal of the Science of Food and Agriculture, 84(12), 1585–1593.

Pernetti, M., van Malssen, K.F., Flöter, E., & Bot, A. (2007). Structuring of edible oils byalternatives to crystalline fat. Current Opinion in Colloid & Interface Science, 12(4),221–231.

Rogers, M.A. (2009). Novel structuring strategies for unsaturated fats—meeting the zero-trans, zero-saturated fat challenge: A review. Food Research International, 42(7),747–753.

Stauffer, C.E. (2005). Emulsifiers for the food industry. In F. Shahidi (Ed.), Bailey's industrialoil and fat products (pp. 229–267). USA: John Wiley & Sons.

Van de Walle, D., Goossens, P., & Dewettinck, K. (2008). Influence of sodium soap andionic strength on the mesomorphic behavior and the α-gel stability of a commercialdistilled monoglyceride. Food Research International, 41(3), 247–254.

Verstringe, S., Danthine, S., Blecker, C., Depypere, F., & Dewettinck, K. (2013). Influence ofmonopalmitin on the isothermal crystallizationmechanism of palm oil. Food ResearchInternational, 51(1), 344–353.

Verstringe, S., Danthine, S., Blecker, C., & Dewettinck, K. (2014). Influence of a commercialmonoacylglycerol on the crystallization mechanism of palm oil as compared to itspure constituents. Food Research International, 62, 694–700.

Zhou, Y., & Hartel, R.W. (2006). Phase behavior of model lipid systems: Solubility of high-melting fats in low-melting fats. Journal of the American Oil Chemists' Society, 83,505–511.