Article pubs.acs.org/Langmuir
Foams Stabilized by Multilamellar Polyglycerol Ester Self-Assemblies Corina Curschellas,*,† Joachim Kohlbrecher,‡ Thomas Geue,‡ Peter Fischer,† Bertrand Schmitt,§ Martine Rouvet,§ Erich J. Windhab,† and Hans Jörg Limbach§ †
Laboratory of Food Process Engineering, ETH Zürich, Schmelzbergstrasse 9, 8092 Zürich, Switzerland Laboratory for Neutron Scattering, Paul Scherrer Institut, 5232 Villigen PSI, Switzerland § Nestec Ltd., Nestlé Research Center, Vers-chez-les-Blanc, 1000 Lausanne 26, Switzerland ‡
ABSTRACT: The importance of surfactant self-assemblies in foam stabilization is well-known. The aim of the current study was to investigate the self-assemblies of the nonionic surfactant polyglycerol ester (PGE) in bulk solutions, at the interface and within foams, using a combined approach of small-angle neutron scattering, neutron reflectivity, and electron microscopy. PGE bulk solutions contain vesicles as well as open lamellar structures. Upon heating of the solutions the lamellar spacing increases, with significant differences in the presence of NaCl or CaCl2 as compared to the standard solution. The adsorption of the multilamellar structures present in the bulk solutions lead to a multilayered film at the air−water interface. The ordering within this film was increased as a result of a 20% area compression mimicking a coalescence event. Finally, PGE foams were shown to be stabilized not only by strong interfacial films but also by agglomerated self-assemblies within the interstitial areas of the foams.
■
INTRODUCTION Foam stabilization in food products is frequently achieved by using lamellar phase forming low molecular weight surfactants.1−3 Many of those amphiphilic molecules are based on a chemical structure containing one or several hydrocarbon chains as the hydrophobic part of the molecule, with some examples being monoglycerides,4−7 various other glycerolbased units esterified with fatty acids,8,9 or phospholipids.10,11 Some of the most prominent systems of this category have been studied in great detail resulting in phase diagrams, either because of their importance in industrial applications or because they are used as a model for structures found in biological systems, such as membranes.12 In many cases the phase diagrams reveal complex, condition-dependent structuring such as lamellar phases, vesicles, and hexagonal or cubic phases. The lyotropic liquid crystals, which is the generic term for these structures, are typically investigated by a combination of different techniques such as small-angle X-ray or neutron scattering,13,14 differential scanning calorimetry,15 and nuclear magnetic resonance16,17 as well as different microscopy techniques such as optical9,18 or scanning and transmission electron microscopy.8,19−21 The phases formed by a specific surfactant depend on its molecular structure. Based on this structure, the so-called critical packing parameter can be determined, from which possible self-assemblies can be predicted.22 By tailoring the molecular structure of the surfactant and the conditions in which it is used, this can be taken advantage of when designing specific products. Regarding the resulting foam stability, fatty acid systems that form micelles in aqueous environments were shown to perform poorer as compared to fatty acid systems that form lamellar structures.23 Indeed, many of the remarkably stable low molecular weight surfactant foams described in the literature © 2012 American Chemical Society
are based on surface-active agents that form large selfassemblies, such as different types of rodlike systems24,25 and vesicles26 but also other liquid crystalline phases.27 In several of those studies they are discussed to not only stabilize the foams due to their surface activity but also through their presence in the interior of foam films and Plateau borders, where they provide steric stabilization and also reduce liquid drainage out of the film. In contrast to the investigation of surfactant self-assemblies in bulk solutions, studying their influence on the stability of foams, beyond simply recording the foam volume over time, has proven to be a challenging task. In recent years, different studies successfully applied small-angle neutron scattering (SANS) to investigate foams in their native state24,28−30 and provided information on thicknesses of foam films separating the bubbles, surfactant structuring within the foam, and bubble sizes. As mentioned in these studies, using this approach not only the surface-active material at the interface is sampled but also all the material contained within the interior of the foam films and the Plateau borders. This can of course also be advantageous, as any material entrapped in these areas, especially if it is organized in some higher order resulting in a signal visible by SANS, will also play a role in the stabilization of foams. The entrapped self-assemblies act by immobilizing water and thereby reducing further drainage of the foam, and in some very specific cases they can also remain entrapped between the films and thereby stabilize the foams against coalescence and Ostwald ripening.26 Received: July 19, 2012 Revised: October 23, 2012 Published: December 6, 2012 38
dx.doi.org/10.1021/la3029116 | Langmuir 2013, 29, 38−49
Langmuir
Article
maximum quantity possible for any given solution, with higher concentrations causing a destabilization of the given solutions. Small-Angle Neutron Scattering. The small-angle neutron scattering experiments were carried out at the SANS I beamline of the Paul Scherrer Institute (PSI, Villigen, Switzerland).40 The measurements were performed at a neutron wavelength of λ = 0.8 nm. The distances selected for the 3He two-dimensional detector were 2, 6, and 18 m for the solution measurements and 2, 5, and 16 m for the foam measurements, allowing for an investigation of a maximum qrange of 0.025−2.5 nm−1, translating into a real space of 2.5−250 nm. All measurements were performed at a collimation of 18 m (analogous to ref 35). The data reduction was performed using the BerSANS software,41 and the data were further treated and fitted using SASfit.42 Investigation of Solutions. For all measurements of solutions, custom-made quartz glass measuring cells with a thickness of 2 mm were used. For the temperature scans the measuring cells were placed into a custom-made sample holder, and vacuum was applied to the environment of the measuring cells in order to prevent condensation on the glass walls of the cells. The temperature was adjusted using a water bath. During the temperature scans, the first measurement at every new temperature was only started once the temperature sensor within a control cell had reached the desired temperature. Investigation of Foams. The measuring cell for the foam measurements consisted of two stainless steel frames each of them enclosing a neutron-transparent quartz glass slide (13.5 cm × 6 cm). A distance holder of 1 cm, which thereby also fixed the sample thickness, was placed in between the steel frames, and the whole setup was secured with multiple screws. The foams were produced using a Buchner funnel with a sintered borosilicate glass filter of porosity 5 (1.0−1.6 μm) from ROBU (Germany). Compressed oil-free air (1− 1.5 bar) was pushed through the membrane, and once the air pressure was constant, the 1 wt % PGE solution to be foamed was poured onto this membrane. Using this model process, foams were created and carefully transferred into the open measuring cell (one steel frame and the distance holder) using a spoon. These foams displayed a broad bubble size distribution, with maximum bubble sizes of several millimeters. The measuring cell was then closed with the second steel frame, the screws were tightened, and the whole cell was brought into an upright position and placed into the neutron beam. All measurements were performed at room temperature. Overall, a measurement including all three detector distances investigated lasted about 30 min. Neutron Reflectivity. Neutron reflectivity experiments were performed on the planar air−water interface of a Langmuir trough (Risø Research Center, Denmark) at the reflectometer AMOR43 (PSI, Villigen, Switzerland). Prior to experiments the trough was cleaned with 2-propanol (Merck, Germany) and chloroform (Merck, Germany). The trough had a surface area of 209 mm × 160 mm. In order to minimize the amount of deuterium oxide required, glass slabs were placed into the Langmuir trough to decrease its internal volume. About 250 mL of the solution to be investigated were then poured into the Langmuir trough. To reduce evaporation during the measurement, the Langmuir trough was enclosed within a sealed container. After aligning the beam with respect to the interface the measurement was started immediately. The measurement of all three angles investigated was always performed in the order of 0.4°, 1.6°, and finally 2.8°. To test for changes of the film over time, this measuring sequence, which had a total duration of about 6.5 h, was repeated once leading to a total experimental time of at least 13 h. After these 13 h of adsorption time, the film was compressed with a barrier speed of 0.15 mm/s until 80% of the initial area was reached. Following this surface compression the beam was realigned with respect to the new surface and another measuring cycle of the three angles mentioned above, with a total duration of 6.5 h, was started. Reflectivity curves have been calculated using Parratt’s recursive method44 applying the software program package Parratt32.45 Scanning Electron Microscopy (SEM). PGE solutions were frozen rapidly with a propane jet cooled at liquid nitrogen temperature, employing a JFD030 device (Baltec-Leica, Germany). Solution drops were deposited onto 300-mesh gold grids. These samples were then
To investigate solely the structures present at the air−water interface, neutron reflectivity is a successfully applied method.31−33 These experiments require a planar air−water interface and are usually performed using a Langmuir trough. This has the additional advantage of providing the means to vary the surface pressure, and thereby surfactant films can be investigated in different states of compression. Depending on the surfactant investigated, films can be formed through surfactant spreading or adsorption from the bulk phase. The aim of the current study was to improve the understanding of the mechanism behind the remarkable foam stability of a low molecular weight surfactant from the group of polyglycerol esters (PGE). Whereas PGE foams are not stable against drainage, they were previously shown to resist coalescence over a 48 h period of investigation.34 In this food-grade, nonionic surfactant the glycerol moieties are esterified with octa- and hexadecanoic acid. In aqueous solution it forms lamellar phases, such as multilamellar vesicles,35 and films transferred from model foams showed complex, multilayer structuring.36 Despite the nonionic nature of this surfactant, the pH has been shown to influence both the foamability of such solutions and the resulting foam properties, presumably as a result of the action of free fatty acids, which are residues from the production process.36−38 Therefore, the pH, as well as the addition of specific ionic species, was one major focus within the current investigations. By applying both neutron reflectivity and small-angle neutron scattering, the structures at air−water interfaces as well as those inside foam films and Plateau borders were investigated, yielding a more concise image of how these foams are stabilized. Small-angle neutron scattering and electron microscopy of the bulk solutions as well as confocal microscopy of the foams complemented these results.
■
EXPERIMENTAL SECTION
Materials. All glassware used to prepare solutions was cleaned with hot water, Milli-Q water (resistivity of 18.2 MΩ·cm, purified with a Millipore purification unit (Millipore)), 2-propanol (Merck, Germany), and chloroform (Merck, Germany) (analogous to ref 39). The polyglycerol ester investigated in the current study was PGE 55. It was purchased from Danisco (Braband, Denmark), and it was used without any further purification. For the neutron scattering and reflectivity experiments all the solutions were prepared using deuterium oxide, D2O (99.8 atom % D), purchased from Armar Chemicals (Döttingen, Switzerland). For the adjustment of the pH, 1 M HCl (Merck, Germany) was used; for the investigations of the effect of ions NaCl (purity ≥99%, Merck, Germany) and CaCl2 (calcium chloride dihydrate, purity ≥99%, Sigma-Aldrich, Switzerland) were added to the solutions. To test the influence of free fatty acids, some solutions were enriched in fatty acids through additions of hexadecanoic acid (purity 98%, Sigma-Aldrich, Switzerland). Methods. Solution Preparation. Solutions were prepared (adapted from ref 35) by weighing the corresponding amounts of PGE and if used additional dry ingredients, such as NaCl, hexadecanoic acid, or CaCl2 together with the deuterium oxide or Milli-Q water into clean and dry glass flasks. These mixtures were then heated to 80 °C in a water bath and kept at 80 °C for 10 min, before they were cooled quiescently in an ice−water bath. Prior to all experiments, the solutions were left to mature for a minimum of 12 h (analogous to ref 36) and a maximum of 5 days. Respecting a specific maturation time was shown to be important, as these solutions are in a nonequilibrium state.35 If required, pH adjustments were made immediately before the experiments. The concentration of PGE was chosen as to best fit to the experiment performed, such as 0.01 wt % for adsorption investigations, 1 wt % for foaming trials, and 10 wt % for small-angle neutron scattering experiments. The additional additives, such as NaCl and CaCl2, were always added in their 39
dx.doi.org/10.1021/la3029116 | Langmuir 2013, 29, 38−49
Langmuir
Article
Figure 1. Electron microscopy and Porod plot. Scanning electron micrographs of 10 wt % PGE (a), 10 wt % PGE, 0.01 M NaCl (b), and 10 wt % PGE, 0.001 M CaCl2 (c) and their corresponding scattering curves (d). The scale bars correspond to 10 μm. The profiles (symbols) and respective fits (solid lines) in (d) have been shifted by multiplication with arbitrary values in order to better visualize the data. pressed flat into a “sandwich” between two copper plates and were frozen through the application of a liquid propane jet, while being clamped in a sample holder. The frozen specimen were conserved in liquid nitrogen prior to their transfer into the cryo-preparation unit Alto 2500 (Gatan, Germany) at −170 °C. The “sandwich” was then opened and the internal structure of the sample made accessible. An alternate procedure was used for 10 wt % PGE solutions due to their high viscosity. Those samples were frozen in nitrogen slush using an Alto 2500 freezing station. Following their transfer into the cryopreparation unit at −170 °C, the samples were fractured through the impact of a razor blade. A slight etching of all specimen was performed on the cryostage of the microscope for 10 min at −105 °C, followed by a sample stabilization at −120 °C. This was followed by an application of a 5 nm gold layer onto the surface. For visualization of all samples a Quanta 200 FEG (FEI, Eindhoven) at 8 kV in highvacuum mode was used. Transmission Electron Microscopy (TEM). The vitrification of samples was performed using a controlled-environment vitrification system for cryo-transmission electron microscopy (cryo-TEM) built in house, with the chamber humidity set to 100%. 200-mesh copper grids (Quantifiol R2/2 or S7/2, Germany) covered with carbon films containing holes were used. 5 μL drops of solution were deposited onto the grids. Before being propelled into liquid nitrogen, the specimen were blotted for 2 s between two filter papers. Frozen grids were conserved in liquid nitrogen. For observation they were transferred into a cryo-holder (Gatan 626-DH), kept at −180 °C. The investigations were performed on a Philips CM12 TEM (Philips, The Netherlands) at 80 kV, with the images being recorded with a Quemesa camera (Olympus, Japan). Confocal Microscopy. The foams were prepared from 1 wt % solutions with a standard kitchen mixer (Hobart), using the whisk attachment. The mixer was set at its highest rotational speed for a total whipping time of 10 min. Immediately after the foam production, a small sample from below the top layer of the foam was transferred into the cavity (depth 2.5 mm, cavity dimensions 24 mm × 20 mm) of a
microscopy slide built in house. Six drops of a 0.1% nile red solution in ethanol (Sigma-Aldrich, Germany) were dissolved in 7 mL of Milli-Q water. A few drops of this stain solution were carefully dripped onto the surface of the foam in order to label the PGE molecules. The foam was then covered with a coverslip, avoiding strong compression of the sample. Confocal microscopy observations of PGE foams were performed using a Zeiss LSM710 microscope. Images were acquired with a 10× objective, and a 561 nm laser source was used for fluorochrome excitation. The emitted fluorescence was collected through a 564−735 nm filter.
■
RESULTS AND DISCUSSION PGE Self-Assemblies in Bulk Solutions. Recent studies suggested that PGE foams are stabilized by interfacial PGE structures as well as by PGE self-assemblies inside the foam films and Plateau borders.34,36 All these structures are either already present in the bulk solutions prior to the foaming process, or they are formed through reorganization processes from the bulk structures, as a result of their adsorption onto the newly created interfaces during the foaming process. Therefore, it can be assumed that primary PGE self-assemblies present in the bulk form the basis of all secondary self-assembled structures present in foams. A previous investigation addressed the influence of temperature and surfactant concentration on the self-assemblies in PGE bulk solutions using different methods including small-angle neutron scattering.35 It was shown that the Bragg peaks appearing at concentrations higher than 5 wt % are indicative of multilamellar structures, which improve in respect to their ordering at a concentration of 10 wt %. The q−2 slope in the low q-range was interpreted as an effect of lamellar objects that locally appear planar, such as the very large vesicles known to be present in PGE solutions.35,46 40
dx.doi.org/10.1021/la3029116 | Langmuir 2013, 29, 38−49
Langmuir
Article
Figure 2. Influence of temperature: 10 wt % PGE (a), 10 wt % PGE, 0.01 M NaCl (b), and 10 wt % PGE, 0.001 M CaCl2 (c) and resulting lattice parameters (d). The open symbols at 23 °C in (d) represent the measurement after the heating cycle. The data in graphs (a)−(c) are displayed by symbols and the fits by solid lines. The solutions displayed in (a) had already previously been investigated by SANS.35 They were remeasured for comparison with the solutions of adjusted ion content. The profiles and respective fits have been shifted by multiplication with arbitrary values in order to better visualize the data.
Influence of Monovalent and Divalent Ions. An addition of 0.015 wt % CaCl2 to a 1 wt % PGE solution (= 0.14 mM) was previously shown to lead to a partial agglomeration of the surfactant structures.35 This concentration allows for 1% of PGE molecules to form a complex with a Ca2+ ion, assuming a 1:1 complex. A similar destabilization phenomenon was also shown to take place for solutions at pH 3.37 The exact mechanism behind the separation of PGE agglomerates from the bulk solution through the addition of cations or the lowering of the pH is not fully understood. It is believed to be mainly based on the reduction of the repulsion between vesicles, due to charge screening, with the charge being introduced into the vesicles by free fatty acids.35,37 In order to investigate the influence of cations on the surfactant structuring within the current study, concentrations for NaCl and CaCl2 additions were chosen just below the concentration leading to agglomeration and flocculation. The choice for Na+ as a monovalent ion and Ca2+ as a divalent ion was based on their relevance for food products. Scanning electron micrographs and the corresponding scattering curves recorded at 23 °C of 10 wt % PGE solutions and 10 wt % PGE solutions containing 0.01 M NaCl and 0.001 M CaCl2 are displayed in Figure 1d. At the time of purchase, PGE contains ∼0.134 wt % sodium.a Assuming a molecular weight of PGE of Mw ≈ 770 g/mol,39 about 5% of PGE molecules can directly interact with a Na+ ion. The addition of 0.01 M NaCl to a 10 wt % PGE solution will allow for another 8% of PGE molecules to form a complex with a Na+ ion,
Furthermore, exceeding the Krafft temperature TK of PGE was shown to lead to a less ordered structure. This confirms the transition of the fatty acid chains into their liquid state at temperatures above TK.35 Based on these results, solutions of 10 wt % PGE were selected for the current investigation of the influence of cations and deliberately added free fatty acids on the self-assemblies at different temperatures. The SANS data were fitted with a form factor describing stacks of lamellar discs. The size of the vesicles is far outside the size range that can be investigated by SANS, which was previously concluded from the q−2 dependence of 10 wt % PGE solutions at low q values as well as from optical microscopy.35 Therefore, this model is considered suitable for the investigation of solutions containing large vesicles, as on the length scale investigated by SANS they appear like planar lamellar objects.35 Furthermore, this approach also allows to account for the planar lamellar structures forming the network that surrounds the vesicles at PGE concentrations of 10 wt % (see Figure 1a−c). The structure factor chosen is based on the “modified Caillé theory”. This theory includes a term for the bilayer bending rigidity, the stacking parameter, a term accounting for the scattering of uncorrelated bilayers and the mean number of bilayer stacks. This mean number of stacks is based on a Gaussian distribution to account for the polydispersity in stack size.47,48 Furthermore, as the thickness of the lamellae is not expected to be homogeneous a lognormal distribution of this thickness was introduced. 41
dx.doi.org/10.1021/la3029116 | Langmuir 2013, 29, 38−49
Langmuir
Article
resulting in a total of ∼13% of PGE molecules capable of interacting directly with Na+ ions. In the case of Ca2+, virtually none is present in the surfactant (
