University of Birmingham Whey protein fluid gels for the stabilisation of foams

The ability of whey protein fluid gels to produce very stable foams was demonstrated. These systems were prepared by heat induced gelation within the turbulent flow field of a pin stirrer at pH 5 and 8. The effect of pH and final protein concentration on the morphology of the particles, the bulk, interfacial and rheological properties and finally the foaming properties of their aqueous suspensions were investigated. Whey protein fluid gels, when produced close to the isoelectric point, consist of small spherical protein aggregates without significant functionality. Micrographs taken suggest that the protein aggregates created have the ability to adsorb at the air/water interface. Nevertheless, the lack of further increase in interfacial viscosity or elasticity indicates that either the adsorption is easily reversible or that it is only partial due to lack of material available to provide complete coverage. By increasing the pH of these systems the protein entities present acquire a negative charge, which causes an increase to both the bulk and interfacial viscoelasticity and increase of the stability of foams. The proposed mechanism is that during foaming, the smaller and mobile protein entities diffuse fast to the interface and provide the necessary interfacial tension reduction to facilitate foam formation. Subsequently, the larger protein particles fill the free space between the air bubbles and increase the local bulk viscosity, which improves foam stability mainly by preventing drainage. Whey protein fluid gels were able to create the same amount of foam as non-treated whey proteins but with substantially increased stability. analysis of variance (one-way ANOVA) in order to determine significant differences 188 between the samples. Data where checked for following normal distribution and equality 189 of variance prior to carrying out the ANOVA. Finally, Tukey’s test for pairwise 190 comparison of means was used in order to determine the trends of the samples that 191 exhibited different behaviour. The level of significance of p<0.05 was chosen. The 192 statistical analysis was performed using Minitab 17 statistical package (Minitab Inc., PA, 193 US). viscosity values of the control systems (native and pre-heated WPI) show a maximum at 336 3 wt% concentration indicating that at this level the packing at the interface is optimum 337 giving the highest film strength. The WPI fluid gels produced at pH 5 showed a 338 statistically significant decrease in both surface elasticity and viscosity compared to the 339 control systems indicating that the Gibbs criterion is not met.

Foams are stabilised by amphiphilic entities, which can adsorb at the air/water 24 interface and reduce the surface tension. Food foams are usually stabilised by high 25 molecular weight surfactants or particles. Milk and egg proteins are very common 26 ingredients present in foods that contain bubbles. Particle stabilised foams on the other 27 hand are only present in whipped cream and ice cream where the foam structure is 28 stabilised by a particulate matrix of either partially coalesced fat or ice crystals or both 29 (Dickinson & Murray, 2006). Particles have shown an increased potential in producing 30 ultra stable foams but in most of the cases the material used is non food (Eric Dickinson, 31 2010). Trying to fulfil this need for finding food grade materials than can produce 32 particles with foam stabilising properties, several studies have focused on plant 33 carbohydrate materials (starches and cellulose derivatives) in combination with 34 surfactants like proteins. These systems have demonstrated potential (Murray,Durga,35 Yusoff, & Stoyanov, 2011). Particles present in foams can provide stability whether they 36 adsorb on the interface or remain on the continuous phase. In case of adsorption, they 37 form rigid films that are stable against drainage and disproportionation, which is evident 38 from an increase of the interfacial elasticity and viscosity. When there is no adsorption 39 the particles either go through a percolation process and create a gel-like network or act 40 as corks reducing the drainage of the continuous phase due to gravity (Rullier, Novales, 41 & Axelos, 2008). 42 As the recent consumer trend demands foods to be 'natural' and 'wholesome' 43 accompanied by a 'clean label', there is a need of producing particles from readily 44 available food ingredients (Brockman & Beeren, 2011). A solution lies on the ability of 45 several proteins to produce aggregates and gels when a set of conditions are met. In 46 this study the heat denaturation of whey proteins is being explored as a mechanism to 47 create micron sized aggregates or gel particles that will entail the necessary surface 48 M A N U S C R I P T

A C C E P T E D ACCEPTED MANUSCRIPT
properties to adsorb on the air/water interface and produce very stable foams. Similar 49 systems containing discrete gel protein particles created by freeze drying quiescently 50 gelled whey protein isolate (WPI) suspensions prepared at a range of pH (5 to 8) has 51 shown prominent results (Lazidis et al., 2014). The whey protein gel particles when 52 rehydrated were able to produce foams with an increased stability by up to an order of 53 magnitude compared to native proteins at the same concentration. That method allowed 54 a significant proportion of whey proteins to remain in their native form and potentially in a 55 form of soluble aggregates. It has been demonstrated in the past that the foaming 56 properties of mixtures of soluble proteins with insoluble aggregates have enhanced 57 foaming properties (Zhu & Damodaran, 1994). 58 This study focuses on exploring a process of producing gel micro particles that has a 59 potential in future upscaling for manufacture through an industrially applicable process. 60 The ultimate goal is to produce micro particle suspensions that will be in a state which 61 will allow transport through a pump to a drying operation in order to achieve a powder 62 formulation. The route chosen was the implementation of the well established method of 63 producing fluid gels by subjecting a polymer solution to gelling conditions (pH, ionic 64 strength, concentration, temperature) while being under a shear field (Norton, Jarvis, & 65 Foster, 1999). In this case WPI solutions were heated at the critical gelation temperature 66 while subjecting them to the turbulent flow field of a pin stirrer device. In this context, the 67 effect of processing conditions (heating and mixing rate) along with the effect of 68 environmental conditions (pH) on the physicochemical characteristics and foaming 69 properties of the WPI gel particulates were assessed. left stirring overnight at 4 ºC to fully hydrate. The pH of the solutions was then adjusted 76 to the desired value using 5M NaOH or 5M HCl (Sigma Aldricht, Dorset, UK). For 77 creating the fluid gels the solutions were pre-heated to 40 ºC and then fed through a 78 peristaltic pump to a series of two jacketed pin stirrers of well defined geometries 79 (Gabriele, 2011). The first pin stirrer device was set to 80 ºC and 2000 rpm rotation 80 speed. The speed of the feed pump was adjusted in order to provide a retention time 81 inside the pin stirrer corresponding to a heating rate of 2 ºC min -1 . The outlet of the first 82 pin stirrer was connected to an identical second pin stirrer rotating at 2000 rpm and set 83 to a temperature of 5 ºC that was used for cooling and diluting the primary fluid gel. For 84 diluting the primary fluid gel RO water at room temperature (25 ºC) was fed through a 85 second inlet of the second pin stirrer at a rate that would allow the final dilution of the 86 initial 12 wt% to 5, 3 and 1 wt% which represented more relevant concentrations to the 87 study of foaming. The final fluid gel was stored in glass bottles at 4 ºC until further 88 characterisation. 89

Preparation of foams and determination of foam stability
90 Foams were prepared with two different methods, gas sparging and mechanical 91 whipping. When foam stability and rate of drainage was assessed, foams were prepared 92 by bubbling using a method adapted from literature (Waniska & Kinsella, 1979). Foam 93 was created inside a clear acrylic circular column (75 mm internal diameter and 500 mm 94 in height) by air sparging at a rate of 3 l min -1 and pressure of 2 bars through the bottom 95 of the column where a porosity 3 (15 -40 µm pores) glass sintered plate is located. A M A N U S C R I P T A C C E P T E D ACCEPTED MANUSCRIPT sample of 150 ml suspension was placed in the column and then air sparging was 97 initialised until the production of a foam head of 20 cm. The reduction of the height of the 98 foam head was then recorded with a CCD camera and the foam half-life was later 99 calculated. All measurements were carried out in four replicates. 100 For all other measurements, foams were produced by mechanical whipping using a 101 commercial milk frother with a spiral impeller of 10 mm diameter rotating at 102 approximately 2000 rpm. For the means of foam production, 20 ml of sample was placed 103 in a 100 ml glass beaker and whipped for 30 seconds. All measurements were 104 performed at 25 ºC unless otherwise stated. 105

Determination of foam overrun 106
The foaming ability of the systems studied was investigated by measuring the 107 amount of air that was able to be incorporated after foaming by mechanical whipping. (2) 123 Where ܸ is the volume of the liquid in the foam and ‫ܥܣ‬ ோெௌ is the voltage square mean 124 reading through the whole volume of the foam. 125 2.5 Measurement of particle size The ζ-potential was determined using a Zetasizer (Malvern Instruments, UK) 135 equipped with an automatic titration unit that allowed the determination of ζ-potential 136 over a wide range of pH (2-12). The titrator used NaOH and HCl of 0.5 and 1M to 137 increase and decrease the pH accordingly. The rheological properties of the bulk phase were determined by oscillatory rheology 144 by applying a strain controlled frequency sweep at a strain rate within the linear 145 viscoelastic region of the samples as defined by an amplitude sweep performed 146 beforehand using a similar sample. The oscillations were performed using the same 147 parallel plate geometry. All rheological measurements were carried out in three 148 replicates. 149

Surface tension measurements 150
The Wilhelmy plate method was used to measure the static surface tension of the 151 solution using a K100 Tensiometer from Kruss GmbH (Hamburg, Germany). The plate 152 was allowed to equilibrate for 1200 s while the surface tension was being recorded. All 153 measurements were carried out in three replicates..
where A is the area of the drop (mm 2 ), σ a/w the air/water interfacial tension (mN m -1 ), E' (predominately β-lactoglobulin and α-lactalbumin) unfold, exposing many of their active 199 groups. On a second step, these exposed groups form covalent bonds causing the 200 proteins to aggregate and form a 3-dimensional network (Mulvihill & Donovan, 1987). 201 When heat denaturation occurs at pH higher or lower than the isoelectric point the therefore, seems to be detrimental for the structure of these aggregates. In order for 207 whey proteins to form a gel, a set of conditions has to be met. For the given 208 environmental conditions (pH and ionic strength) there is a critical protein concentration 209 (C g ) and a critical temperature (T g ) needed in order to form a gel. Generally both C g and M A N U S C R I P T A C C E P T E D ACCEPTED MANUSCRIPT T g increase while moving away from the isoelectric point. At pH 8 the critical 211 concentration is approximately 12 wt% and the critical temperature is 80 ºC, when heat 212 is provided with a rate of 1 ºC min -1 (Stading & Hermansson, 1990). 213 During the heat gelation of whey proteins under shear, nuclei are formed as a 214 consequence of the disruption to the secondary structure, which leads to the formation 215 of oligomers. These nuclei grow through a series of weak interactions and covalent 216 bridging up to an aggregate size, which is limited by the magnitude of the shear applied. 217 Subsequently, the disruption of the shear intra particle interactions can cause bridging 218 mainly due to disulphide bonds, which lead to the formation of a continuous network. A 219 way of manipulating the size of the aggregates formed is by choosing the magnitude of 220 the applied shear and the heating rate. It has been shown that at large shear rates 221 (higher than 600 s -1 ) the size of the aggregates is limited by the shear whilst at lower 222 shear rates the size of the aggregates is generally larger but can be retained by 223 The main purpose of this study was to create WPI fluid gels in a continuous manner. 257 In terms with this, the production of fluid gels was done as described in Section 2.1 in a 258 pin stirrer device for both systems at pH 5 and pH 8. The production set up did not allow The shown differences in ζ-potential suggest also changes to the magnitude of the 295 electrostatic repulsions between particles in these systems, even at the same pH. This 296 dependence of the pH at which the fluid gels were made on the electric charge of the 297 system can be due to the difference in the amount of free non-aggregated proteins still 298 present in the system following the heating step. Recent work (unpublished) by the 299 authors has shown that the amount of non-aggregated proteins in the system is higher 300 when the gelation takes place at pH 5 compared to pH 8. 301

Interfacial properties of WPI fluid gels 302
The functional properties of proteins also depend on their ability to adsorb at 303 surfaces/interfaces and lower the surface/interfacial tension of the system stabilising the 304 newly formed interface. This behaviour is important when looking at colloidal systems 305 such as foams, where the surface tension (SFT) of the system is a good indicator of the 306 ability of the surfactants, in this case proteins, to adsorb on the interface during foaming. 307 When statistically comparing the SFT values (Table 1) of the samples it seems that the 308 samples made at pH 5 and then readjusted to 8 and the heat treated WPI have 309 significantly lower values which is also dependant on the concentration. Whilst the M A N U S C R I P T

A C C E P T E D ACCEPTED MANUSCRIPT
14 samples made at pH 8 and at pH 5 have higher values. It is worth mentioning that the 311 difference between the lowest (heated WPI 5 wt%) and highest (WPI FG pH 5 1 wt%) 312 mean is less than 4 mN/m 2 . Therefore there might be statistical differences between the 313 samples but these are in reality small. This indicates that the ability of the protein 314 systems to lower the surface tension remains unaffected by exposure to heat during fluid 315 gel production. This is probably due the presence of a significant amount of non-316 aggregated proteins, which rapidly adsorb and saturate the surface across all samples 317 studied, but also due to the ability of the formed protein aggregates to adsorb on the a/w 318 interface. The WPI fluid gels contain enough protein material that can diffuse fast and 319 adsorb on the interface lowering the surface tension and allowing the formation of a 320

foam. 321
Colloidal particles have shown to stabilise foams due to their ability to adsorb on the 322 air/water interface and form a strong film (Dickinson & Murray, 2006). The strength of 323 this film provides an energy barrier that prevents the diffusion of gas between different 324 sized bubbles (disproportionation). During disproportionation the bubbles tend to shrink 325 and in order for that to happen the bubbles have to work against the interfacial elasticity 326 and viscosity which can suppress shrinkage (Murray & Ettelaie, 2004). When particles 327 adsorb at the air/water interface the resulting increase in interfacial elasticity and 328 viscosity reduces significantly the diffusion of gas between bubbles of different size and 329 therefore disproportionation. The interfacial rheological properties of the systems studied 330 here were therefore investigated using pedant drop tensiometry. The interfacial elasticity 331 and viscosity values of native and heated WPI were also measured for means of 332 comparison as shown on Table 1. 333 The conventional condition that bubbles are stale against disproportionation is the 334 M A N U S C R I P T

A C C E P T E D ACCEPTED MANUSCRIPT
viscosity values of the control systems (native and pre-heated WPI) show a maximum at 336 3 wt% concentration indicating that at this level the packing at the interface is optimum 337 giving the highest film strength. The WPI fluid gels produced at pH 5 showed a 338 statistically significant decrease in both surface elasticity and viscosity compared to the 339 control systems indicating that the Gibbs criterion is not met. 340 When the pH of the same fluid gels was adjusted to 8 then both the surface elasticity 342 and viscosity increased significantly (statistically not different to the control systems for 343 the 5 wt% concentration). The fluid gels produced at pH 8 had higher values of 344 interfacial elasticity and viscosity than the ones at pH 5 and statistically similar to the 345 ones when the pH was readjusted from 5 to 8. This implies that both electrostatic 346 repulsions and size affect the interfacial rheological properties of these systems. More 347 specifically the interfacial elasticity increases enough to fulfil that ‫ܧ‬ ᇱ > ߪ /௪ /2, implying 348 that disproportionation can be significantly limited. The smaller gel particles produced at 349 pH 5 when provided with a charge by increasing the pH to 8 can produce films that have 350 high elasticity and viscosity but not higher than the ones of native or heat treated 351 proteins. This is either due to weak adsorption of these particles on the interface or lack 352 of complete interfacial coverage. All samples show a shear thinning behaviour that suggests the presence of 361 interactions and structural formation, which eventually disrupted by shear (Figure 6a). The functionality of WPI fluid gel systems in terms of producing and stabilising foams 387 is presented here. The ability of the fluid gels to incorporate sufficient amount of air and 388 produce foams was assessed in terms of overrun (Table 1). All fluid gel samples were 389 able to produce foams with very similar overrun to the ones made by native proteins. 390 The statistical evaluation has shown that the fluid gel samples made at pH 8 and pH 5 391 produced foams with the highest overrun values amongst the samples studied but still 392 with values close to the native proteins. The ability of the system to incorporate air was 393 not hindered as it is common with particle containing systems (Murray & Ettelaie, 2004). 394 The overrun results are in agreement with the surface tension data ( Table 1)  foaming properties which are dependent on both the pH at which these structures were 448 originally formed but also the pH at which they were aerated. 449 Fluid gels made at pH 5 whilst not having any specific foaming functionality at their 450 intrinsic pH, when adjusted and aerated at pH 8, demonstrated an enhanced foaming 451 capacity. This is due to what is proposed as an enhancement to the electrostatic inter-452 particle interactions promoted by the change to the pH environment. This is supported 453 by both the ζ-potential measurements but also by the presented interfacial elasticity and 454 viscosity data. The small aggregates formed at pH 5 seem to be able to fill effectively the Fluid gels formed at pH 8, while having stronger charge and higher bulk viscosity and 459 elasticity were able to produce less viscoelastic films probably due to the larger size of 460 the aggregates present, which limited their ability to pack which lead to exclusion from 461 the spaces around the air bubbles. This resulted in producing stable foams compared to 462 native proteins by reducing the rate of liquid drainage but not to the same extent as the 463 fluid gels made at pH 5 and aerated after adjustment to pH 8.  Table 1 Interfacial properties of fluid gels and native WPI samples.

M A N U S C R I P T
A C C E P T E D ACCEPTED MANUSCRIPT Table 1 Interfacial properties of fluid gels and native WPI samples.