1. Field of the Invention
The present invention relates to the production of an aerated frozen confectionery comprising of hydrophobin, a Secondary protein, and a Co-surfactant (CSF).
2. Background of the Art
It is desirable to make small size bubbles in aerated frozen confections that are stable to temperature abuse in order to improve texture, palatability, and reduce caloric content. It is well understood that frozen confections such as sorbets, sherbets and ice cream experience temperature fluctuations in the distribution chain and in consumers' home freezers. This results in bubble growth and a degradation of the product quality and the palatability on consumption. It is preferable to produce an ice cream with stable bubbles.
US Patent publication no. 2006024417A, published on Feb. 2, 2006 to Berry et al. discloses aerated products comprising hydrophobin where the hydrophobin is used to inhibit bubble coarsening.
US Patent publication no. 2008213453A published on Sep. 4, 2008 to Burmester et al. discloses aerated food products and methods producing them, where in the product comprises hydrophobin and a surfactant.
We have unexpectedly found that when the combination of hydrophobin, at least one Secondary protein and at least one Co-surfactant (as those terms are defined below) are used to make an aerated frozen confectionery product then:
In one aspect of the invention is a frozen aerated food composition including but not limited to
a. at least 0.01 wt. % of total hydrophobin(s) selected from the group consisting of HFBII, HFBI or Cerato Ulmin or blends thereof added in isolated form to the food composition;
b, one or more Co-surfactants in the total concentration range of about 0.001 to less than about 0.3 wt. % (preferably less than about 0.1 or 0.2 wt. %), and the Co-surfactant(s) to total hydrophobin wt, ratio is in the range of about 0.02 to less than 1.0. (Preferably the Co-surfactant to hydrophobin wt. ratio is at least about 0.05, more preferably at least about 0.3 and preferably at most about 0.75); and
c. one or more Secondary protein(s) that are different from hydrophobin(s): wherein said Secondary protein(s) are present in a total concentration range of about 0.25 to less than 6.0 wt. % (Preferably the total Secondary protein(s) are at least 0.5 or 1.0 wt. % and at most 4 or 5 wt. %).
In another aspect of the invention is a process of making an aerated food composition including but not limited to the steps of:
a. Blending the food composition ingredients together, mixing, and pasteurising;
b. Adding hydrophobin and a Co-surfactant or Co-surfactants to the chilled mix of step (a);
c. Aerating and freezing the mix of step (b) to produce the aerated and frozen product; and
d. Cooling the frozen product to a storage temperature of less than about −15 C.
In one aspect of the invention is a frozen aerated food composition including but not limited to:
a. at least 0.01 wt. % of total hydrophobin(s) selected from the group consisting of HFBII, HFBI or Cerato Ulmin or blends thereof added in isolated form to the food composition;
b. one or more Co-surfactants in the total concentration range of about 0.001 to less than about 0.3 wt. % (preferably less than about 0.1 or 0.2 wt. %), and the Co-surfactant(s) to total hydrophobin wt. ratio is in the range of about 0.02 to less than 1.0. (Preferably the Co-surfactant to hydrophobin wt. ratio is at least about 0.05, more preferably at least about 0.3 and preferably at most about 0.75); and
c. one or more Secondary protein(s) that are different from hydrophobin(s) wherein said Secondary protein(s) are present in a total concentration range of about 0.25 to less than 6.0 wt. % (Preferably the total Secondary protein(s) are at least 0.5 or 1.0 wt. % and at most 4 or 5 wt. %).
Advantageously the total hydrophobin(s) concentration is at most 1.5 wt. %. Preferably the Co-surfactants is or are water soluble non-ionic surfactant(s). More preferably the Co-surfactant(s) are selected from Polysorbates, polyglycerol esters of alkyl or alkenyl fatty acids, diacetyl tartartic acid esters of mono-/di-glycerides, sucrose esters with an HLB >about 8 or blends thereof. Most preferably the Co-surfactant(s) has a minimum effective HLB value of about 8. Effective HLB value is here defined as the arithmetic mean of the HLB values of a blend of Co-surfactants Advantageously the Co-surfactant(s) is selected from Tweens 20, 60 or 80, PGE-O-80; Panodan-Visco Lo 2000 and blends thereof.
Preferably the average bubble diameter (d3,2) is at least 10% smaller after the standard temperature abuse protocol described below than the same product prepared the same way but absent either hydrophobin(s), Co-surfactant(s) or if both are present then outside the total Co-surfactant(s) to hydrophobin(s) ratio range of about 0.02 to less than 1.0. Preferably the average bubble size is at least 20, 30, 40 or 50% smaller.
Preferably the average bubble diameter d(3,2) of the freshly prepared frozen product stored at below about −15° C. and pre-temperature abuse is at least 10% smaller, preferably 15% smaller, and more preferably 20% smaller than the same product prepared the same way but absent either hydrophobin(s). Co-surfactant(s) or if both are present then outside the total Co-surfactant(s) to hydrophobin(s) ratio range of about 0.02 to less than 1.0.
In another aspect of the invention is a process of making an aerated food composition including but not limited to the steps of:
a. Blending the food composition ingredients together mixing, and pasteurising;
b. Adding hydrophobin and a Co-surfactant or Co-surfactants to the chilled mix of step (a);
c. Aerating and freezing the mix of step (b) to produce the aerated and frozen product; and
d. Cooling the frozen product to a storage temperature of less than about −15 C.
Preferably the process further includes the step of extruding the frozen product from a freezer at a temperature of about −5 C or less.
All percentages, unless otherwise stated, refer to the percentage by weight, with the exception of percentages cited in relation to the overrun.
The term “frozen confection” means an edible confection made by freezing a mix of ingredients which includes water. Frozen confections typically contain fat, non-fat milk solids and sugars, together with other minor ingredients such as stabilisers, emulsifiers, colours and flavourings. Frozen confections include ice cream, water ice, frozen yoghurt, sherbets, sorbet and the like.
The term aeration means that gas has been incorporated into a product to form air cells. The gas can be any gas but is preferably, particularly in the context of food products, a food-grade gas such as air, nitrogen or carbon dioxide or a mixture of the aforementioned. The extent of the aeration can be measured in terms of the volume of the aerated product. The stability of the aeration can be assessed by monitoring the volume of the aerated product over time and or the bubble size change over time.
The microstructure of frozen confections is critical to their organoleptic properties. The air cells incorporated into frozen confections are preferably small in size which ensures that the frozen confections do not have a coarse texture and also ensures that they deliver a smooth creamy mouth-feel. In typical ice cream products, the air bubbles coarsen over time (through distribution and storage) leading to a degradation in quality.
Hydrophobins are a well-defined class of proteins (Wessels, 1997, Adv. Microb. Physio. 38: 1-45; Wosten, 2001, Annu Rev. Microbiol. 55: 625-646) capable of self-assembly at a hydrophobic/hydrophilic interface, and having a conserved sequence:
where X represents any amino acid, and n and m independently represent an integer. Typically, a hydrophobin has a length of up to 125 amino acids. The cysteine residues (C) in the conserved sequence are part of disulphide bridges. In the context of the present invention, the term hydrophobin has a wider meaning to include functionally equivalent proteins still displaying the characteristic of self-assembly at a hydrophobic-hydrophilic interface resulting in a protein film, such as proteins comprising the sequence:
or parts thereof still displaying the characteristic of self-assembly at a hydrophobic-hydrophilic interface resulting in a protein film. In accordance with the definition of the present invention, self-assembly can be detected by adsorbing the protein to Teflon and using Circular Dichroism to establish the presence of a secondary structure (in general, α-helix) (De Vocht et al., 1998, Biophys. J. 74:2059-68).
The formation of a film can be established by incubating a Teflon sheet in the protein solution followed by at least three washes with water or buffer Masten et al., 1994, Embo. J. 13: 5848-54). The protein film can be visualised by any suitable method, such as labeling with a fluorescent marker or by the use of fluorescent antibodies, as is well established in the art. m and n typically have values ranging from 0 to 2000, but more usually m+n<100 or 200. The definition of hydrophobin in the context of the present invention includes fusion proteins of a hydrophobin and another polypeptide as well as conjugates of hydrophobin and other molecules such as polysaccharides.
Hydrophobins identified to date are generally classed as either class I or class II Both types have been identified in fungi as secreted proteins that self-assemble at hydrophobilic interfaces into amphipathic films. Assemblages of class I hydrophobins are relatively insoluble whereas those of class II hydrophobins readily dissolve in a variety of solvents.
Hydrophobin-like proteins have also been identified in filamentous bacteria, such as Actinomycete and Steptomyces sp. (WO01/74864). These bacterial proteins, by contrast to fungal hydrophobins, form only up to one disulphide bridge since they have only two cysteine residues. Such proteins are an example of functional equivalents to hydrophobins having the conserved sequences shown in SEQ ID Nos. 1 and 2, and are within the scope of the present invention.
The hydrophobins can be obtained by extraction from native sources, such as filamentous fungi, by any suitable process. For example, hydrophobins can be obtained by culturing filamentous fungi that secrete the hydrophobin into the growth medium or by extraction from fungal mycelia with 60% ethanol. It is particularly preferred to isolate hydrophobins from host organisms that naturally secrete hydrophobins. Preferred hosts are hyphomycetes (e.g. Trichoderma), basidiomycetes and ascomycetes. Particularly preferred hosts are food grade organisms, such as Cryphonectria parasitica which secretes a hydrophobin termed cryparin (MacCabe and Van Alfen, 1999, App. Environ. Microbicl. 65: 5431-5435).
Alternatively, hydrophobins can be obtained by the use of recombinant technology. For example host cells, typically micro-organisms, may be modified to express hydrophobins and the hydrophobins can then be isolated and used in accordance with the present invention. Techniques for introducing nucleic acid constructs encoding hydrophobins into host cells are well known in the art. More than 34 genes coding for hydrophobins have been cloned, from over 16 fungal species (see for example WO96/41882 which gives the sequence of hydrophobins identified in Agaricus bisporus: and Wosten, 2001, Annu Rev. Microbiol. 55: 625-646). Recombinant technology can also be used to modify hydrophobin sequences or synthesise novel hydrophobins having desired/improved properties.
Typically, an appropriate host cell or organism is transformed by a nucleic acid construct that encodes the desired hydrophobin. The nucleotide sequence coding for the polypeptide can be inserted into a suitable expression vector encoding the necessary elements for transcription and translation and in such a manner that they will be expressed under appropriate conditions (e.g. in proper orientation and correct reading frame and with appropriate targeting and expression sequences). The methods required to construct these expression vectors are well known to those skilled in the art.
A number of expression systems may be used to express the polypeptide coding sequence. These include, but are not limited to, bacteria, fungi (including yeast), insect cell systems, plant cell culture systems and plants all transformed with the appropriate expression vectors. Preferred hosts are those that are considered food grade—‘generally regarded as safe’ (GRAS).
Suitable fungal species, include yeasts such as (but not limited to) those of the genera Saccharomyces, Kiuyveromyces, Pichia, Hansenula, Candida, Schizo saccharomyces and the like, and filamentous species such as (but not limited to) those of the genera Aspergillus, Trichoderma, Mucor, Neurospora, Fusarium and the like.
The sequences encoding the hydrophobins are preferably at least 80% identical at the amino acid level to a hydrophobin identified in nature, more preferably at least 95% or 100% identical. However, persons skilled in the art may make conservative substitutions or other amino acid changes that do not reduce the biological activity of the hydrophobin. For the purpose of the invention these hydrophobins possessing this high level of identity to a hydrophobin that naturally occurs are also embraced within the term “hydrophobins”.
Hydrophobins can be purified from culture media or cellular extracts by, for example, the procedure described in WO01/57076 which involves adsorbing the hydrophobin present in a hydrophobin-containing solution to surface and then contacting the surface with a surfactant, such as Tween 20, to elute the hydrophobin from the surface. See also Colien at al., 2002, Biochim Biophys Acta. 1569: 139-50; Calonje at al., 2002, Can. J. Microbial. 48: 1030-4; Askolin at a, 2001, Appl Microbiol Biotechnol, 57: 124-30; and De Vries at al. 1999, Eur J. Biochem. 262: 377-85.
The hydrophobin used in the present invention can be a Class I or a Class II hydrophobin. Preferably, the hydrophobin used is a Class II hydrophobin. Most preferably, the hydrophobin used is HFBI, HFBII, or CU (cerato ulmin). The hydrophobin used can also be a mixture of hydrophobins, e.g. Class II hydrophobins HFBI and HFBII.
The product should comprise at least 0.01 wt. % hydrophobin, more preferably at least 0.025 wt. % hydrophobin and most preferably at least 0.05 wt. %. Preferably the hydrophobin is present in an amount of 1.5 wt. % maximum and more preferably 0.5 wt. % maximum and most preferably 0.2 wt. % maximum.
The extent of aeration of a product is measured in terms of “overrun”, which is defined as:
Where the weights refer to a fixed volume of mix or product. Overrun is measured at atmospheric pressure.
Preferably the overrun of the product is between 10 and 400% overrun, more preferably between 10 and 300% overrun, and most preferably between 20 and 250% overrun. Preferably the measurements are taken immediately after aeration is ended.
Interfacial or surface rheology, defines the functional relationship between stress, deformation and rate of deformation at an interface in terms of coefficients of elasticity, and viscosity, arising from relaxation processes. The technique is referred to as dilationl interfacial rheology when the experimentally imposed interfacial deformation arises from variation of area at constant shape. The investigation of the dilational rheology of adsorbed layers is useful to access the macroscopic viscoelastic properties of interfaces and can be used to predict the stability of a foam once formed.
In the dilational deformation mode, the use of the interfacial tension (γ) response to relative area variation (ΔA/A) provides for the definition of the dilational viscoelasticity. Assuming harmonic area perturbations of small amplitude of frequency v, the dilational viscoelasticity can be written using a linear approximation approach, as
where the viscoelastic modulus E can be further split into its elastic and viscous components (see e.g. R. Miller, L. Liggieri, Interfacial Rheology, Brill, Leiden, 2009, Ch. 5, 138).
One criterion to reduce bubble coarsening (e.g. coalescence and/or disproportionation) is to confer adequate interfacial properties (particularly elasticity) to the air/water surface, i.e. the bubble surface. Experimental data show that high interfacial elasticity (for completely “elastic” interfaces) is able to slow clown the rate of disproportionation (see e.g. W. Kloek, T. van Vliet, M. Meinders, J Colioi Interf Sci, 2001, 237, 158), and consequently this bubble/foam coarsening.
This criterion has been used here to predict the stability of fully formulated aerated frozen confections such as ice creams stored under controlled thermal conditions, which includes a temperature abuse protocol. Verification of the stability of the ice cream was assessed using SEM (Scanning Electron Microscopy) and observations compared with predictions from the dilational interfacial rheology experiments.
A Co-surfactant (CSF) is defined as:
An ingredient which, when mixed in an aqueous solution containing:
0.001 wt. % hydrophobin
and between 0.0015 and 0.2 wt, % of at least one non-hydrophobin (Secondary) protein at a concentration effective to confer an air/water surface dilatational elasticity that is at least 30% of that of 0.001 wt. % pure hydrophobin (absent the Co-surfactant), more preferably at least 50%, more preferably at least 55%, more preferably at least 65% and most preferably at least 70%, measured between 600 and 4000 s at 5° C. for an air droplet in water subject to a continuous area change of between 2.5 and 3.5% oscillated at a frequency of 0.05 Hz using the procedure provided below. The effective concentration will depend on the identity of the Cc-surfactant. Preferably the effective concentration will be in the range of 0.001 to less than 0.2 wt. % based on the product; more preferably in the range of 0.005 to 0.2 wt. % and most preferably in the range of 0.01 to 0.1 wt.
Preferably the Co-surfactant is chosen as one of more of the following:
Secondary protein(s) are defined as non-hydrophobin proteins that when mixed with 0.001 wt. % hydrophobin in aqueous solution at a concentration of 0.04 wt. % results in an air/water surface dilatational elasticity that is at least 35% less than that of hydrophobin alone, more preferably at least 40% less, more preferably at least 50% less, where the dilatational measurement is made between 600 and 4000 s at 5° C. for an air droplet in water subject to a continuous area change of between 2.5 and 3.5% oscillated at a frequency of 0.05 Hz. Secondary proteins, are advantageously chosen from food proteins such as: skim milk protein (SMP), whey protein, soy protein, or mixtures thereof and the like.
It may be noted that many Secondary proteins such as skim milk powder are not typically received and used as pure proteins, since they consist of other ingredients such as lactose and other non-protein materials. Therefore the protein content must be taken into account during formulation.
Further additional ingredients are typically added to make the frozen confectionary product. These include and are not restricted to:
Sugars, e.g. sucrose, fructose, dextrose, corn syrups and sugar alcohols and the like. Fats, e.g. coconut oil, butter oil, palm oil and the like. Preferably the fat content of the product is less than 5 wt. %, more preferably less than 3 wt. %, more preferably less than 2 wt. %, most preferably less than 1.5 wt. %, 1.0, 0.5, 0.4, 0.3, 0.2, 0.1, 0.01, 0.001 wt. % or zero.
Emulsifiers, e.g. mono/di glycerides of fatty acids and the like other than Co-surfactants.
Stabilisers or thickeners, e.g. locust bean gum, guar gum, tara gum, carrageenans, alginates, pectins, citrus fibres, xanthan, gelatine and the like.
Flavours and colours. e.g. vanilla, fruit purees, chocolate, mint and the like.
The invention will now be described in greater detail by way of the following non-limiting examples. The examples are for illustrative purposes only and not intended to limit the invention in any way. Physical test methods are described below:
Except in the operating and comparative examples, or where otherwise explicitly indicated, all numbers in this description indicating amounts or ratios of materials or conditions or reaction, physical properties of materials and/or use are to be understood as modified by the word “about”.
Where used in the specification, the term “comprising” is intended to include the presence of stated features, integers, steps, components, but not to preclude the presence or addition of one or more features, integers, steps, components or groups thereof.
All percentages in the specification and examples are intended to be by weight unless stated otherwise
Surface dilatational rheology measurements were taken using formulations A, B, and C shown in Table 1.
The elastic modulus data are summarised in
This recovery is preferred since it will lead to a more stable foam structure in an aerated frozen confectionery where the ratio of HFBII:Co-surfactant:SMP is similar to that described in this example, although the absolute concentrations will be greater in order to make the food product.
Frozen aerated (ice cream) products (A), (B), and (C) were made using formulations summarised in Table 2. The relative amounts of hydrophobin, SMP, and Tween 20 Co-surfactant are the same as the equivalent examples measured in Example 1 (A, B, C).
The microstructures of the fresh and temperature abused products for A, B, and C, are shown in the SEM images depicted in
Inventive Product C with added Co-surfactant has the smallest air bubbles for the fresh sample (
It was found that the recovery of the elastic modulus for the inventive Co-surfactant containing solution (Example 1C) was associated with the improved microstructure and stability of the formed ice cream (Example 2C), as observed from SEM images shown in
Evaluating the extent of the recovery (if any) of the elastic modulus (Eel) in presence of a Co-surfactant over the comparative hydrophobin+SMP case (e.g. Example 1B) is a method that was found useful to predict the quality of the ice cream microstructure after the temperature abuse storage test.
The elasticity of the a/w (air/water) interface in presence or absence of a Co-surfactant can be listed at set time points (e.g. 1200 s, 1800 s, 2400 s, 3600 s) and compared with the elasticity (absolute and percentage) of reference hydrophobin example alone (measured at 3600 s).
Elasticity of HFBII (0.001 wt. %) measured after 3600 s is 212.9±19.4 mN/m (average of five repetitions)—referring to solution in Example 1A, and illustrated in
Aerated frozen products were produced using different examples of Co-surfactants. SEM images illustrated in
In all the comparative cases, after the temperature abuse protocol, the microstructure was observed to degrade (air bubbles visibly coarsened), as is also seen in the case when an inventive Co-surfactant is not present in the formulation (Example 2B and Example in
Evaluation of the optimized level of Co-surfactant (Panodan) for aerated frozen products formulated at a 0.1 wt. % concentration HFBII and 11.3 wt. % SMP (Table 4) was carried out using the interfacial rheology technique on model formulations A, B, C and D (concentration reduced 200 times) shown in Table 5.
Table 5. Model Interfacial rheology formulations used to determine optimum amount of Co-surfactant for use in the ice cream formulation stated in Table 4. Concentration for SMP (skim milk powder) is that stated for the total powder. The amount of protein in SMP was stated by the manufacturer as 35 wt. %.
The results from the interfacial rheology experiments are summarised in Table 6. The data predict that the concentration of Panodan (CSF) to be ideally used in the ice cream example (Table 4) should be used in the ratio of 0.0005 wt. % HFBII:0.0002 wt. % Panodan. The concentration used in the ice cream example was proposed as 0.1 wt. %. Therefore, the appropriate concentration of Panodan Co-surfactant should be preferably greater than 0.01% and preferably around 0.04 wt. %.
In the following section we show that the recovery of interfacial elastic modulus promoted by CSF is a common feature to different variants of class II hydrophobins. In this case, using either HFBI or CU, both of which are class II hydrophobins.
Surface dilational rheology experiments were carried out in the same manner as in the previous examples, except in these cases ΔA/A=2.5%.
The same pattern is seen in
These data demonstrate that the use of a Co-surfactant works with other class II hydrophobins.
Preparation of the solutions for the experimental work was done using the same ingredients as in the preparation of fully formulated ice creams except that tap water was used to prepare ice creams. All other solutions were prepared in de-ionised water (18.2 MΩ cm). Concentrations of ingredients are expressed as weight %.
Solutions were prepared using combinations of hydrophobin, Secondary protein e.g. a milk protein, and an added Co-surfactant.
In the interfacial rheology experiments the concentration of used ingredients was scaled down 200-times with respect to the levels used in ice cream manufacturing. The concentrations used for interfacial rheology experiments are (as an example and not restricted to these):
Reported values of the viscoelastic modulus (E) were measured using the Drop Shape tensiometer PAT-1 (Sinterface, Germany). The measuring configuration is that of a bubble emerging from a J-shaped capillary positioned inside the cell containing the solution. The PAT-1 tensiometer implements a feature allowing for an accurate control of the bubble interfacial area with the possibility of varying it during the measurement according to predetermined patterns. This feature is utilised for the measurement of the dilational viscoelasticity. Purely harmonic oscillations of the bubble interfacial area with small amplitude and frequency are imposed (immediately after the bubble formation) while the surface tension response, γ(t), is measured. From the amplitude of the two signals, A(t) and γ(t), and the phase shift between them, the elastic and viscous components of E are calculated. Amplitude and phase of the measured A(t) and γ(t) oscillatory signals are extracted by standard Fourier analysis techniques.
In the experiments reported here an air bubble of area A0=18 mm2 was formed at the tip of a J-shaped capillary in a glass cell containing about 27 ml of the solution. An area variation between 2.5 and 3.5% was imposed during oscillations at the frequency of 0.05 Hz and a temperature of 5° C., unless otherwise stated. A gentle nitrogen stream was directed onto the cell glass walls (front and back) to prevent air humidity condensation which obscures the cell field of view
The microstructure of each product was visualised using Low Temperature Scanning Electron Microscopy (LTSEM). The sample was cooled to −80° C. on dry ice and a sample section cut. This section, approximately 5 mm×5 mm×10 mm in size, was mounted on a sample holder using a Tissue Tek:OCT™ compound (PVA 11 wt. %, Carbowax 5 wt. % and 85 wt. % non-reactive components). The sample including the holder was plunged into liquid nitrogen slush and transferred to a low temperature preparation chamber: Oxford Instrument CT1500HF. The chamber is under vacuum, approximately 10−4 bar, and the sample is warmed up to −90° C. Ice is slowly etched to reveal surface details not caused by the ice itself, so water is removed at this temperature under constant vacuum for 60 to 90 seconds. Once etched, the sample is cooled to −110° C. ending the sublimation, and coated with gold using argon plasma. This process also takes place under vacuum with an applied pressure of 10−1 millibars and current of 6 milliamps for 45 seconds. The sample is then transferred to a conventional Scanning Electron Microscope (JSM 5600; JEOL LTD. Japan), fitted with an Oxford Instruments cold stage at a temperature of −160° C. The sample is examined and areas of interest captured via digital image acquisition software e.g. using the method described below.
Ice cream pre-mixes were prepared by adding the solid ingredients to hot water (>60° C.) with stirring to disperse. The pre-mix was then heated to 80° C. with a plate heat exchanger, then homogenised at 140 bar pressure and pasteurised at 82° C. for 25 seconds. The mix was then cooled via a plate heat exchanger to 5° C. and held at this temperature (aging) for at least 2 hours before further processing. After the aging step and prior to the freezing process (below), concentrated hydrophobin solution and Co-surfactant was added to the mix with gentle stirring to disperse. When used in combination, concentrated hydrophobin solution was mixed with the Co-surfactant before adding to the mix.
After aging the mixes were processed using an WCB MF75 (for small scale) or Hoyer KF 1000 (for large scale) freezer. All aerated products were produced at 100% overrun and extruded ca. −6° C. Frozen products were collected in 500 mL waxed paper cartons and hardened in a blast freezer at −35° C. for 2 hours before storage at −25° C.
Products stored at −25° C. are classed as “fresh” products since the microstructure is stable at this temperature and growth of bubbles and ice crystals will not be significant between the time of production and further analysis by SEM (<1 month).
“Temperature abused” products were transferred to a freezer wherein the temperature fluctuates between −20° C. and −10° C. over 24 hours as follows: 11 hours 30 mins at −20° C. followed by 30 mins at +10° C., then 11 hours 30 mins at −10° C. followed by 30 mins at +10° C. After 2 weeks abuse with daily temperature fluctuations as described above, the products were then transferred to a −25° C. freezer before further analysis.
The gas bubble size (diameter) distribution as used herein is defined as the size distribution obtained from the two dimensional representation of the three dimensional microstructure, as visualized in the SEM micrograph, determined using the following methodology.
Samples are imaged at 3 different magnifications (for reasons explained below), and the bubble size distribution of a sample is obtained from this set of micrographs in three steps:
1. Identification and sizing of the individual gas bubbles in the micrographs
2. Extraction of the size information from each micrograph
3. Combination of the data from the micrographs into a single size distribution
All of these steps, other than the initial identification of the gas bubbles, can conveniently be performed automatically on a computer, for example by using software such as MATLAB R2006a (MathWorks. Inc) software.
Firstly, a trained operator (i.e. one familiar with the microstructures of aerated systems) traces the outlines of the gas bubbles in the digital SEM images using a graphical user interface. The trained operator is able to distinguish gas bubbles from ice crystals (which are present in frozen aerated products and are the same order of magnitude in size) because the gas bubbles are approximately spherical objects of varying brightness 1 darkness whereas ice crystals are irregular-shaped objects of a uniform grey appearance.
Secondly, the size is calculated from the selected outline by measuring the maximum area as seen in the two dimensional cross-sectional view of the micrograph (A) as defined by the operator and multiplying this by a scaling factor defined by the microscope magnification. The bubble diameter is defined as the equivalent circular diameter d:
d=2√{square root over (A/π)}
This is an exact definition of the diameter of the two-dimensional cross-section through a perfect sphere. Since most of the gas bubbles are approximately spherical, this is a good measure of the size.
Extraction of the Size Information from Each Micrograph
Gas bubbles which touch the border of a micrograph are only partially visible. Since it is not therefore possible to determine their area, they must be excluded. However, in doing so, systematic errors are introduced: (i) the number of gas bubbles per unit area is underestimated: and (ii) large gas bubbles are rejected relatively more often since they are more likely to touch the border, thus skewing the size distribution. To avoid these errors, a guard frame is introduced (as described in John C. Russ, “The Image Processing Handbook”, second edition, CRC Press, 1995). The guard frame concept uses a virtual border to define an inner zone inside the micrograph. The inner zone forms the measurement area from which unbiased size information is obtained, as illustrated in
Bubbles are classified into 5 classes depending on their size and position in the micrograph. Bubbles that fall fully within the inner zone (labelled class 1) are included. Bubbles that touch the border of the virtual micrograph (class 2) are also included (since it is only a virtual border, there is fact full knowledge of these bubbles). Bubbles that touch the actual micrograph border (class 3) and/or fall within the outer zone (class 4) are excluded. The exclusion of the class 3 bubbles introduces a bias, but this is compensated for by including the bubbles in class 2, resulting in an unbiased estimate of the size distribution. Very large bubbles, i.e. those larger than the width of the outer zone (class 5), can straddle both the virtual (inner) border and the actual outer border and must therefore be excluded, again introducing bias. However, this bias only exists for bubbles that are wider than the outer zone, so it can be avoided by excluding all bubbles of at least this size (regardless of whether or not they cross the actual border). This effectively sets an upper limit to the gas bubble size that can be reliably measured in a particular micrograph. The width of the inner zone is chosen to be 10% of the vertical height of the micrograph as a trade-off between the largest bubble that can be sized (at the resolution of the particular micrograph) and the image area that is effectively thrown away (the outer zone).
There is also minimum size limit (at the resolution of the micrograph) below which the operator cannot reliably trace round gas bubbles. Therefore bubbles that are smaller than a diameter of 20 pixels are also ignored.
Combination of the Data from the Micrographs into a Single Size Distribution
As explained above, it is necessary to introduce maximum and minimum cut-off bubbles sizes. In order that these minimum and maximum sizes are sufficiently small and large respectively so as not to exclude a significant number of bubbles, some samples may need to be imaged at 3 different magnifications: e.g. 100×, 300× and 1000×. This occurs if there is a wide distribution in bubble sizes, and the skilled user can determine what magnifications are appropriate in order to capture the full size distribution: one magnification or more. As an example for the case of 3 different magnifications, each magnification yields size information in a different range, given in Table 7.
Thus bubbles as small as 2 μm and as large as 83 μm are counted. Visual inspection of the micrographs at high and low magnifications respectively confirmed that essentially all of the bubbles fell within this size range. The magnifications are chosen so that there is overlap between the size ranges of the different magnifications (e.g. gas bubbles with a size of 20-28 μm are covered by both the 100× and 300× micrographs) to ensure that there are no gaps between the size ranges. In order to obtain robust data, at least 500 bubbles are sized; this can typically be achieved by analysing one micrograph at 100×, one or two at ×300 and two to four at ×1000 for each sample.
The size information from the micrographs at different magnifications is finally combined into a single size distribution histogram. Bubbles with a diameter between 20 μm and 28 μm are obtained from both the 100× and 300× micrographs, whereas the bubbles with a diameter greater than 28 μm are extracted only from the 100× micrographs. Double counting of bubbles in the overlapping size ranges is avoided by taking account of the total area that was used to obtain the size information in each of the size ranges (which depends on the magnification), i.e. it is the number of bubbles of a certain size per unit area that is counted. This is expressed mathematically, using the following parameters:
N=total number of gas cells obtained in the micrographs
dk=the kth outlined gas cell with kε[l,N]
Ai=the area of the inner zone in the ith micrograph
Ri=the range of diameters covered by the ith micrograph (e.g. [20 μm, 83 μm])
B(j)=the jth bin covering the diameter range: [jW, (j+1)W)
The total area. S(d), used to count gas bubbles with diameter d is given by adding the areas of the inner zones (Ai) in the micrographs for which d is within their size range (Ri).
The final size distribution is obtained by constructing a histogram consisting of bins of width W μm. B(j) is the number of bubbles per unit area in the jth bin (i.e. in the diameter range j×W to (j+1)×W). B(j) is obtained by adding up all the individual contributions of the gas bubbles with a diameter in the diameter range j×W to (j+1)×W, with the appropriate weight, i.e. 1/S(d).
Magnifications used are chosen by the skilled user in order to extract bubble size through the analysis software.
The bubble size distributions are conveniently described in terms of the normalised cumulative frequency, i.e. the total number of bubbles with diameter up to a given size, expressed as a percentage of the total number of bubbles measured.
Alternative expressions of bubble size distribution can also be used, e.g. d3,2.