6362

Ind. Eng. Chem. Res. 2008, 47, 63626367

Designing Multiscale Structures for Desired Properties of Ice Cream

Ice cream is a complex multiphase structure consisting of ice, air, and fat as dispersed phases at a range of different length-scales, all embedded in a continuous phase consisting of unfrozen sugar solution known as the matrix or serum. The entire structure is the result of both the ingredients and all the processes used in ice cream manufacture including emulsication, freezing, and aeration. It is thermodynamically unstable and delivered quality can only be assured at low and stable temperatures. Physicochemical processes during storage can lead to loss of quality by coarsening of the ice particles, disproportionation of the air, and the loss of water from the matrix. Product design for specic sensory, stability, shape, and increasingly, nutritional properties, is a challenging task and must take account of all these aspects of the structure. Almost all properties are sensitive to the size, density, and morphology of the dispersed phases as droplets, cells, crystals, or even micelles. Finer structures, in general, result in more desirable organoleptic properties such as creaminess and smoothness but the interfacial dynamics are more rapid, leading to less stability. Even small changes in the relative densities of the dispersed phases such as in the case of low-fat or fat-free products can dramatically change key properties such as taste perception, mouth-feel, and rate of melt. Conventional formulation and processing techniques complemented by the use of specic additives such as emulsiers and stabilizers enable some control, albeit limited, over the interfacial dynamics and stability. New ingredients and new technologies (such as low temperature extrusion) have been developed to enable higher levels of control on the interfacial behavior either through direct molecular intervention on an interface or new structuring processes wherein interfaces are created in a new or different way. Examples of new ways of inuencing the ice, fat, and air interfaces will be discussed such as ice structuring protein and hydrophobins. Challenges that remain highlight the need for new types of molecular and microstructural interventions to achieve the next levels in design capability for the ice creams of tomorrow.
1. Introduction There is a widely held impression that the design and development of ice cream involves only issues relating to recipe and chefmanship borne out of kitchen skills. Although chefmanship is important, ice cream per se is a deeply technological product in all of its development, manufacture, distribution. and selling. Whereas product delivery to the consumer is aimed to achieve high appeal through various sensorial attributes, it is deep inside the structure where the science and technology is embedded. Designing ice cream to meet specic technical performance targets is a major challenge.1 This paper aims to show that, with deeper scientic understanding and some very exciting new technologies both in use and under development, some of the major barriers to better ice cream in terms of quality, stability, nutrition, and innovation can be overcome. 2. Design Challenge Ice cream is a product that truly operates on a range of spatial scales. On a macroscale, the sensory properties of the texture are perceived; these are determined by the microscopic details of the structure. In turn, microstructure is determined by complex molecular interactions. The main aim of ice cream manufacturers is to generate the correct microstructure in the ice cream to achieve the desired organoleptic characteristics such that the product can breakdown and melt away in the mouth, thus delivering the consumers preferences for taste. In addition, the structure needs to be sufciently robust to withstand transportation and storage, so there is quite a balancing act to perform to
* To whom correspondence should be addressed. E-mail: deryck.cebula@unilever.com. Tel.: 44-1234-222748. Fax: 44-1234222007.

reconcile these simultaneous and often conicting requirements. Therefore, in achieving the optimum microstructure, there are trade-offs between the formulation (levels and types of ingredients and actives such as process aids and stabilizers) and the processing regimes (heat transfer rate, temperature of freezing, etc). A general description of the science of ice cream is given by Clarke.2 Increasingly, as consumers demand healthier products, nutritional aspects of formulation become vastly more important. Whereas, for example, reduction of both saturated fat and sugar are desirable, they may not be immediately possible because these are crucial components for both the process conditions and the microstructure per se. A typical microstructure is one that consists of ice crystals and air bubbles in the size range 20 m to about 100 m, and fat droplets in the size range of 1 m to 0.1 mm (Figure 1). These ne entities are embedded throughout a viscous solution of sugars, polysaccharides, and milk proteins known as the matrix. At another order of magnitude lower in scale, it is possible to identify the location of the fat. Fat droplets of size <1 m can be seen that exist as clusters located on the surface of the air cells as well as distributed throughout the matrix. Not shown here, milk protein is also partially located on the air interface and together fat and protein both help to stabilize the air. Fat has an incredibly important role in the microstructure that relates directly to the sensory properties like mouth-feel, creaminess, and avor delivery but it is also critical to the stability properties of products such as rate of melt. It can be appreciated that reducing the fat by 50%, or more, to enable healthier products may not only compromise the sensory quality but may put the stability, specically of the air phase, at considerable risk.

10.1021/ie701773z CCC: $40.75 2008 American Chemical Society Published on Web 08/06/2008

Ind. Eng. Chem. Res., Vol. 47, No. 17, 2008 6363

Figure 1. Ice cream microstructure depicted through scanning electron microscopy. The overall structure is determined largely by freezing and aeration process conditions. Complex interactions exist between the structural phases and between fat, protein, and emulsier. Shown in the inset, air is stabilized by a coating of fat droplets (Pickering stabilization) and because the matrix is highly viscous.

As previously stated ice cream is thermodynamically unstable and even under ideal storage conditions the structure, specically the ice and the air phases, will coarsen over time, resulting in loss of quality and loss of stability. This situation is exacerbated by upward temperature uctuations and by pressure changes which affect the air phase. In addition, stability becomes a real problem when distributing products across different altitudes when the ice cream expands in response to lower pressures then shrinks to much lower volume as normal pressure is restored. Low-fat or reduced-calorie products are particularly susceptible to variations in ambient conditions. To fulll the requirements of the complete design challenge without unacceptable trade-offs of quality and stability, better control is needed of the starting microstructure and of the unstable dispersed phases: ice and air. These require solutions to overcome some major technical problems. The remainder of the paper will address new technical developments in process and ingredient technology that can help with the design challenge. 3. Technical Developments 3.1. Microstructure Control by Process. One of the most direct approaches to better control microstructure is via processing. Low-temperature extrusion is a new technology involving a single screw extruder3 that replaces conventional hardening and takes fresh ice cream straight from the manufacturing freezer (at about -5 C) and reduces the temperature rapidly below -10 C. The process continues to work the product to capture a ner microstructure. The technology consists of an extruder closely coupled with a conventional scraped surface freezer and operated at high torque to slowly churn and extrude the ice cream at much lower draw temperatures. These lower draw temperatures are possible because the mechanical energy dissipation experienced by the product in the screw extruder is much lower than in the conventional scraped surface freezer.4 This energy dissipation becomes more important when the ice cream viscosity increases rapidly with declining temperature. Low-temperature extrusion of ice cream has also been achieved using a twin screw extruder5,6 with draw temperatures as low as -18 C reported.

Low-temperature extrusion leads to a much ner microstructure. This is a direct consequence of the lower draw temperature, which slows the coarsening of the newly formed air and ice phases, which begins immediately upon creation of the ice cream structure. Detailed microstructural analysis of low-temperature extruded ice cream shows that both the air cells and the ice crystals are much smaller, but the presence of smaller air cells is the dominant factor for enhanced sensory properties, specifically creaminess. One additional benet of low-temperature extrusion is that because of the optimized microstructure, it is possible to reduce fat level substantially, in fact by ca. 50%, while delivering the same level of creaminess and better meltdown properties. This has been a major step forward in our drive to improve the nutritional prole of ice cream while retaining high quality and stability. A number of new higher-quality low-fat, or light, products showing major improvements have been launched across the world that exploit the low-temperature extrusion capability and these have been very much appreciated by the consumers in all markets. However, although low-temperature extrusion gives us the benets of a better initial microstructure, it does not solve the problem of instability due to temperature abuse on storage. For that, specic technical solutions for the ice phase and for the air phases per se are needed. 3.2. Controlling Ice Structure. Ice crystals in ice cream coarsen over time even under isothermal storage conditions,7 and the situation is exacerbated or accelerated by elevated temperatures or uctuations. This is a phenomenon called recrystallization. Figure 2 shows the consequence of this processsmany smalls crystals reduce to fewer bigger crystals and the air bubbles increase in size and may coalesce over time. These effects detract from the creaminess and smoothness, resulting in a product that is cold eating and has poorer resistance to meltdown (that is to say it melts and loses serum faster). There are two distinct processes involved in recrystallization (Figure 3) more fully described by Sutton et al.8 The rst is referred to as migratory recrystallization and is a manifestation of Ostwald ripening. Here, the tendency to minimize surface free energy in ice crystals of different sizes in a multiparticle distribution favors the survival of crystals with a large radius

6364 Ind. Eng. Chem. Res., Vol. 47, No. 17, 2008

Figure 2. Ice cream microstructure shown by scanning electron microscopy. Freshly made ice cream shows a much ner microstructure with smaller air cells and ice crystals, but after storage in which some temperature uctuations occur, there is a coarsening of all phases. Stability in ice cream can not be fully controlled. Both air and ice phases are intrinsically unstable, leading to reduced creaminess/smoothness, colder eating, poorer visual appeal, and rapid melting.

Figure 3. Ice recrystallization, as observed by low-temperature light microscopy (crystals are approximately 50 m across), occurs by two mechanisms: migratory and accretive crystallization. These occur simultaneously driven by reduction in curvature of crystal faces and necking.

of curvature. Storage of products at low temperatures can retard this process but cannot arrest it. The second process is accretive recrystallization and occurs when two crystals make contact. The negative curvature of the area of contact drives the development of a neck, leading to the coalescence of the two crystals. Repeated accretion leads to agglomerates of crystals of quite irregular shapes although the process of rounding during inevitable temperature uctuations in storage largely offsets irregularity. Using gums and other thickeners (called stabilizers) retards contact between crystals, thereby reducing accretion. Growth and accretion take place at the same time. Under highly abusive conditions (sometimes seen in storage or selling cabinets) the ice cream structure can completely breakdown. Clearly, with conventional ingredients, the basic laws of thermodynamics prevail and a new approach to ice control is required. 3.2.1. Ice Structuring Proteins. It seemed that nature already had the solution to ice control. A special type of protein in the blood of Antarctic sh held out the potential for a new mechanism in ice control.9 These proteins bind to ice crystals and protect the organisms from damage in freezing water.10 Initially referred to as antifreeze proteins (AFP), these have now been identied in more than 200 organisms, all of which survive in cold habitats. These molecules inuence the size and structure of the ice crystals and, subsequently, the overall properties of the ice phase. The potential for the use of antifreeze proteins in foods has been considered by Grifths and Ewart11 and, more recently and particularly for ice cream, by Ragand and Goff.12 These proteins are referred to as antifreeze because of their protective function in biology, but as an ice cream ingredient, it is their structuring function that is most important; this is why they have been renamed ice structuring proteins (ISP) by Clarke et al.13

The current body of research literature reports a variety of molecules, distinguished by their structure, either helical, globular, or other forms derived from different species. Each has been designed by nature most effectively to fulll its purpose of protecting the organism from ice damage. All ISPs have the common property of binding to ice but some lock onto specic crystal planes giving anisotropic ice structures. The globular structure (type III) obtained from Ocean Poutssh from the North Atlanticshas the most interesting properties in an ice control context. ISPs are associated with three phenomena: thermal hysteresis, recrystallization inhibition, and crystal habit modication. The latter two properties are of most interest. First, nearly all ISPs studied can limit the growth of ice crystals. A typical example of difference in size and morphology is shown in micrographs of ice dispersions with and without the proteins present (Figure 4). They almost completely arrest the isothermal recrystalliztion process. Second, because of the crystal face specicity of individual molecules, ISPs can change the morphology of the growing crystal as shown in Figure 5. These really are the properties being sought and the ISP type III is the molecule that has been selected by Unilever to develop new ice control technology. 3.2.2. New Ice Structure. Small crystals with different morphologies can result in some new and surprising ice structures in multiple particle dispersions. ISP type III gives very spicular crystals with high aspect ratios, as much as 5-6:1, resulting in high connectivities at ice contents of 60-70%. The micrographs in Figure 6 show the difference in ice structures in a water ice with and without ISP. ISP causes the ice phase, instead of the matrix phase, to become continuous, thereby locking in or encapsulating the matrix phase containing colors and avors in pockets within the ice. Whereas in an open ice structure, the solution-borne components such as the color and avor can leach out, with ISP

Ind. Eng. Chem. Res., Vol. 47, No. 17, 2008 6365

Figure 4. Effect of ISP type III is shown using light microscopy. ISP, even at concentrations as low as 1 10-4mg/mL (1 10-5 wt %), limits the size of ice crystals as they grow in sucrose solution.

Figure 5. Light microscopy showing ice crystal habit modication achieved by the ISP adsorbing to specic faces of the crystal and thus altering its shape as it initially grows.

Figure 6. With planar ice crystal faces comes a greater degree of aggregation. Scanning electron microscopy shows that at high ice content, the ice phase becomes continuous, increasing the mechanical strength. The hard line in the inset picture shows a continuous line, or path, that can be drawn between adjacent crystals. With an ISP modulated structure, no such line can be drawn, indicating that the crystals have become fully networked.

present, this cannot happen and this behavior has been exceptionally useful for stabilizing innovative products with higher levels of solids like fruit purees. Since 2003, more than 40 products containing ISP have reached the market, some new and some major improvements of current products, all exploiting the unique properties of this

protein to give new benets. These illustrate the level of ice control for stability with healthier products even under abusive conditions. We believe this is a breakthrough for better ice cream, which consumers continue to demand.14 However, arguably, the biggest challenge remains the effective control of the air phase.

6366 Ind. Eng. Chem. Res., Vol. 47, No. 17, 2008

3.3. Controlling Air: The Next Challenge. The air phase in ice cream is stabilized through a complex combination of fat droplets and milk proteins that adsorb onto the surface of the air cells during the shearing and aeration processes in the freezer.15 The semisolid fat droplets under go a process called partial coalescence in the high shear freezing process. This leads to a proportion of the fat particles adsorbing to the air bubbles, which helps stabilize the dispersed air phase by impeding coalescence of adjacent air cells when in contact. The presence of clusters of partially coalesced fat particles in the continuous phase also leads to the retention of bulk product shape on melting and provides creamier in-mouth characteristics. Deemulsiers (for example, mono/diglycerides) play an important role in increasing the level of partially coalesced, or destabilized, fat droplets. They facilitate the process of destabilization through reducing the stability of the fat particles to shear. This is achieved by displacing some of the protein from the surface of the particles during the aging step, prior to freeze-aeration.16,17 This process of controlling the stability of the air phase via adsorption of fat and protein is very susceptible to small changes in the protein/de-emulsier ratio, or to changes in the fat content, fat particle size, and also to changes in process conditions. However, although stabilization by fat adsorption to the air bubbles is a key process, long-term stability of the air cells (i.e., through shelf life of the product) is achieved principally by the presence of the very viscous matrix phase. At ideal storage temperatures (-18 C or below) the matrix is almost solid, yet still pseudoplastic, and even though the structural components of the ice cream are thermodynamically unstable, they are effectively held in place and kinetically trapped because of the solid nature of the product. Ultimately, air bubble stability through these mechanisms can be achieved under ideal storage conditions with constant temperature and pressure. Nevertheless, relatively small changes in these conditions readily leads to air bubble coarsening, product degradation, and loss of quality on consumption. 3.3.1. HydrophobinssNatures Solution. With conventional technology, proteins, fat, and de-emulsiers can help achieve a ner air phase microstructure (post extrusion) of the ice cream from the freezer. With temperature cycling (i.e., poor storage or distribution conditions), the coarsening of the air cells is irreversible and will lead eventually to channelling, air loss, and shrinkage of the product.18 Even with the optimum use of de-emulsiers and ideal processing, the air stability is nely balanced and precarious against ambient variations. Thus, a better and more direct solution is required to make a step change in enhanced air control, an ability that has been long sought after by ice cream manufacturers. In investigations for more powerful emulsiers, Unilever Research teams focused on a set of natural proteins referred to as hydrophobins (originally found in arboreal fungi) to help solve the problem. As in the case of ISP, it seemed that nature may have already developed a solution to the problem of air phase control. Currently, two classes of hydrophobins, I and II, distinguished by their specic aqueous solubility, have been identied.19 Class II hydrophobins, which readily dissolve in aqueous solutions, currently hold out the best potential as air control agents. The Class II hydrophobin, HFBII, from the Trichoderma reesei fungus exhibits a stable globular amphiphilic structure, where the majority of the surface is hydrophilic and a smaller area is distinctly hydrophobic.20,21 A distinguishing property of this protein is its very high surface activity at air/ water interfaces (see Figure 7). Measured surface tensions of

Figure 7. Data for the equilibrium surface tension for hydrophobin at an air/water surface (pH 7). This compares to typical values for skim milk powder (50 mN m-1) or sodium caseinate (45 mN m-1).

Figure 8. Data for the surface shear rheology comparing hydrophobin and skimmed milk powder (SMP) at an air/water surface using an ARG2 rheometer with a De Nouy ring. This shows that surface elasticity of hydrophobin is at least an order of magnitude greater than for a typical food protein.

this, and other hydrophobins, make this class of protein one of the most surface active known. However, the most important property of HFBII in relation to air phase stabilization is the ability for it to adsorb to the air/water surface and form exceptionally elastic and cohesive lms. This is demonstrated by the measurement of surface shear rheology, shown in Figure 8. The surface shear elastic and viscous moduli (Gs and Gs, respectively) are measures of the mechanical properties of the surface layer.22 The values of Gs for HFBII are at least an order of magnitude greater than that of other food proteins. Such mechanical properties are known to be directly related to the stability of foams.23 Effectively, each air cell has its own elastic membrane, giving it the ability to respond to and resist pressure and temperature changes without signicant coalescence. From the preliminary work done,24 the HFBII layers appear to provide some of the properties required to stabilize air cells individually so that they are robust even against temperature cycling and abuse. 4. Conclusions Designing structures in ice cream to meet all product performance targets is a very challenging task with many technical hurdles. Whereas the aim is to control kinetically

Ind. Eng. Chem. Res., Vol. 47, No. 17, 2008 6367

stabilized microstructures by processing methods, signicant results can also be achieved by addressing certain molecular features critical for good structure. Air control remains the biggest technical challenge and the hydrophobins development holds out great promise and has real potential for the future. With new developments in science and technology, much progress has been made in making better ice cream both healthier and better quality to meet modern consumers demands. Literature Cited
(1) van der Graaf, K.; Crilly, J. F. New Horizons for Ice Cream. Food Technology International, EFFoST 2004, 1315. (2) Clarke, C. J. The Science of Ice Cream; The Royal Society of Chemistry: London, 2004. (3) Burmester, S.; Russell, A. B.; Cebula, D. J. The Evolution of Ice Cream Technology. New Food 2005, 2, 4245. (4) Russell, A. Process innovation from research and development to production in a large companysdevelopment and commercialisation of a low temperature extrusion process. In Case Studies in Food Product DeVelopment; Earle, M., Earle, R., Eds.; Woodhead Publishing: Cambridge, U.K., 2008; Chapter 11, pp 202-222. (5) Bolliger, S.; Kornbrust, B.; Goff, H. D. Inuence of emulsiers on ice cream produced by conventional freezing and low-temperature extrusion processing. Int. Dairy J. 2000, 10 (7), 497504. (6) Eisner, M. D.; Wildmoser, H.; Windhab, E. J. Air cell microstructuring in a high viscous ice cream matrix. Colloids Surf., A 2005, 263 (13), 390399. (7) Donhowe, D. P.; Hartel, R. W.; Bradley, R. L. Determination of Ice Crystal Size Distributions in Frozen Desserts. J. Dairy Sci. 1991, 74, 3334 3344. (8) Sutton, R. L.; Lips, A.; Piccirillo, G.; Sztehlo, A. Kinetics of Ice Recrystallisation in Aqueous Fructose Solutions. J. Dairy Sci. 1996, 89, 741745. (9) deVries, A. L.; Wohlschlag, D. E. Freezing Resistance in Some Antarctic Fishes. Science 1969, 163, 10731075. (10) Harding, M. M.; Ward, L. G.; Haymet, A. D. J. Type I antifreeze proteins structure-activity studies and mechanism of ice growth inhibition. Eur. J. Biochem. 1999, 264, 653665.

(11) Grifths, M.; Ewart, V. Antifreeze Proteins and Their Potential Use in Foods. Biotech. AdV. 1995, 13, 375402. (12) Regand, A.; Goff, H. D. Ice Recrystallization Inhibition in Ice Cream as Affected by Ice Structuring Proteins from Winter Wheat Grass. J. Dairy Sci. 2006, 89, 4957. (13) Clarke, C. J.; Buckley, S. L.; Lindner, N. M. Ice Structuring proteins: a new name for anti-freeze proteins. Cryo-Lett. 2002, 23, 89 92. (14) Crilly, J. F. ISP: A Breakthrough for Better Ice Cream. New Food 2007, 3, 4044. (15) Goff, H. D. Colloid aspects of ice creamsA review. Int. Dairy J. 1997, 7, 363373. (16) Leser, M. E.; Michel, M. Aerated milk protein emulsions-new microstructural aspects. Curr. Opin. Colloid Interface Sci. 1999, 4, 239 244. (17) Wilde, P.; Mackie, A.; Husband, F.; Gunning, P.; Morris, V. Proteins and emulsiers at liquid interfaces. AdV. Colloid Interface Sci. 2004, 108-109, 6371. (18) Chang, Y.; Hartel, R. W. Stability of air cells in ice cream during hardening and storage. J. Food Eng. 2002, 55, 5970. (19) Wessels, J. G. H. Developmental regulation of fungal cell-wall formation. Annu. ReV. Phytopathol. 1994, 32, 413437. (20) Hakanpaa, J.; Linder, M.; Popov, A.; Schmidt, A.; Rouvinen, J. Hydrophobin HFBII in detail: ultrahigh-resolution structure at 0.75. Acta Crystallogr., Sect. D 2006, D62 (4), 356367. (21) Linder, M. B.; Szilvay, G. R.; Nakari-Setala, T.; Penttila, M. E. Hydrophobins: the protein-amphiphiles of lamentous fungi. FEMS Microbiol. ReV. 2006, 29 (5), 877896. (22) Bos, M. A.; van Vliet, T. Interfacial rheological properties of adsorbed protein layers and surfactants: a review. AdV. Colloid Interface Sci. 2001, 91, 437471. (23) Murray, B. S. Stabilisation of bubbles and foams. Curr. Opin. Colloid Interface Sci. 2007, 12, 232241. (24) Cox, A. R.; Cagnol, F.; Russell, A. B.; Izzard, M. J. Surface properties of class II hydrophobins from Trichoderma reesei and inuence on bubble stability. Langmuir 2007, 23, 79958002.

ReceiVed for reView December 27, 2007 ReVised manuscript receiVed July 4, 2008 Accepted July 9, 2008 IE701773Z