Contact us
9.

The composition and properties of milk

A great number of operational parameters and conditions influence the performance of a spray drying plant. One essential factor is the quality of the processed milk, evaluated not only according to general quality standards, but also composition and physical-chemical properties as influenced by the composition itself.

9.1.

Raw milk quality

The quality of the processed milk, both physical-chemical, bacteriological and organoleptic, must comply with the appropriate quality standard, which specifies the requirements to raw milk quality, taking into account the required quality of the produced product. Therefore this chapter deals only with the quality variations, which do not downgrade general quality.
9.2.

Milk composition

As with any other material of biological origin, milk composition is subject to geographical and seasonal variations as influenced by the breed of animal, lactation period, climate, region, feeding etc. All can have, under certain circumstances, great influence on the quality of the final product, especially if peculiarities of that milk supply have not been recognized and allowed for when choosing the processing technology. The gross composition of milk is given in Table 9.1.
Component % in liquid milk % in total solids
range average range average
Water 88.5-85.3 87.1 - -
Fat 3.4-5.0 3.9 27.4-34.7 29.8
Protein 3.3-3.9 3.5 26.0-27.0 26.7
Lactose 4.8-5.0 4.9 33.9-39.7 38.0
Ash 0.68-0.74 0.7 5.0-5.8 5.5
Non-Fat-Solids 8.8-9.6 9.1 65.2-72.2 70.2
Table. 9.1. Gross composition of milk
For a given region the main differences occur between spring and late autumn milk. From the technological point of view, the attention has to be paid to the quantitative variations of individual components especially those of non-fat-solids (NFS):
 
  1. fat/NFS ratio
  2. content of total protein in NFS
  3. content of lactose in NFS
  4. total protein/lactose ratio
  5. casein/albumin ratio
  6. mineral salts/protein ratio.
Some of these variations are given in Table 9.2.
Total protein in non-fat-solids 36.2-48.0 1)
Lactose in non-fat-solids 51.7-54.9
Protein/lactose ratio 0.63-1.06 1)
Casein/non-fat-solids-ratio 27.6-34.0 1)
Whey protein/non-fat solids ratio 7.4-8.3 1)
Whey protein/casein ratio 0.19-0.22 1)
1) The high range limit applies for Channel Islands breeds and end of season.
Table. 9.2. Variations of some components
The above tables do not express any abnormalities, but just the variations of an average milk of various origins. These variations are an important factor in milk powder production, since extremes can influence the sensitivity of milk to heat treatment, heat stability, tendency to lactose crystallization in concentrates etc. The seasonal variations of the components in non-fat-solids in an area of North Island of New Zealand are shown on Fig. 9.1.
An example of seasonal variations of skim milk components.
Fig.9.1. An example of seasonal variations of skim milk components.
9.3.

Components of milk solids

In this chapter only properties of individual components, which have direct connection to milk powder technology, will be highlighted.
9.3.1.

Milk proteins

Milk proteins consist of a complex of caseins (α -, β-, κ- and minor-caseins), which can be precipitated by acids or rennet, and serum proteins. Both groups exhibit different chemical and physical properties.
 
Casein is the most important and the most characteristic protein of milk. In milk, it is in a form of a fine dispersion of particles (colloidal system) of deformed globule shape (casein micelles). It is sensitive to acids and rennet enzymes, which cause aggregation of individual particles forming flake-like or continuous gel precipitates. In the production of milk powders, it is important to avoid any casein flocculation and to retain its fine dispersion.
 
Among serum proteins belong α-lactalbumin and β-lactoglobulin. They are water soluble, and on casein precipitation they remain in the whey. Whey proteins in cow’s milk, in comparison with milk of other mammals and also with human breast milk, represent a relatively small part (about 22%) of total protein. Therefore in the production of humanized baby food powders, cow’s milk is enriched with whey proteins up to about 60% whey protein.
9.3.2.

Milk fat

Milk fat is present as a dispersed emulsion of fat globules and their conglomerates. The size of fat globules varies considerably depending on various factors. Chemically, fat consists of triglycerides of fatty acids, both saturated and unsaturated. The proportion of individual fatty acids determines the chemical and physical properties of milk fat.
 
Under the influence of lipolytic enzymes, milk fat decomposes into glycerol and fatty acids, resulting in a characteristic, unpleasant taste and smell (mainly of butyric acid). Oxidative decomposition induced or accelerated by day light or the presence of certain metal ions manifests itself by rancid taste.
 
Milk fat globules are covered by a phospholipids/protein membrane. Under the influence of heat above 80°C, free sulfhydryl groups are created. They act as antioxidants protecting fat against oxidation. This is utilized during the production of whole milk powders. During homogenization, especially of concentrated milk, the size of fat globules is considerably reduced. This contributes to digestibility and reduces the amount of so-called free fat in milk powders.
9.3.3.

Milk sugar

Configuration of α- and β-lactose.
Fig.9.2. Configuration of α- and β-lactose.

Milk sugar or lactose is a carbohydrate, existing only in milk, in true water solution.

It is a disaccharide C12H22O11 consisting of glucose and galactose and occurs in two isomeric forms α- and β-lactose (see Fig. 9.2). They have different physical properties, especially solubility in water and polarized light rotation. Both forms can crystallize, but for the production of milk and whey powders the most important is α-lactose, which crystallizes as α-lactose-monohydrate from the supersaturated solution of lactose below the temperature of 93.5°C. During the production of normal milk powders, the water evaporation during spray drying is so fast that despite supersaturation, the lactose cannot crystallize but remains in the powder as amorphous lactose, also called lactose glass. Amorphous lactose is very hygroscopic. This can cause caking problems with powders having high content of lactose, as for instance whey powders. To avoid caking the lactose has to be crystallized as α-lactosemonohydrate, which is non-hygroscopic. This is done by pre-crystallization of the concentrates. The rate of crystallization in solutions of lactose is controlled by the rate of mutarotation, i.e. transformation of β-lactose into α-lactose. The rate of mutarotation decreases with falling temperature, being fairly high in the range of 40-20°C, but practically zero at temperatures below 10°C. The solubility of lactose is shown on Fig.9.3.

Solubility of lactose.
Fig.9.3. Solubility of lactose.

The specific optical rotation of α-lactose in water is [α]D/20°C = +89.4° and melting point 201.6°C under disintegration. The corresponding values for α-lactose are [α] D/20°C = +33.5° and 252.2°C. The equilibrium specific rotation is [α] D/20°C = +55.3°.

α-lactose-monohydrate crystallizes in prism shapes often called tomahawk. These crystals in milk products are detectable on one’s teeth. With crystals larger than 10 μm there is a “flourish” and over 15 μm a “sandy” feel on the teeth.

The density of pure α-lactose-monohydrate is 1.54 g/ml. The sweetness of lactose is only 30% of that of sucrose. For whey powder manufacturing technology a most important physical property is the heat of crystallization, which is 10.63 kcal/kg. This must be taken into consideration when calculating the consumption of cooling water for the crystallization tanks. The relationships between various forms of lactose acc. to. King [3] are shown in Fig. 9.4.

The relationships between different forms of lactose.
Fig.9.4. The relationships between different forms of lactose.
9.3.4.

Minerals of milk

From the point of view of the technology involved, the most important cations are the calcium and magnesium ions. These are bound to casein and to the phosphoric and citric acid anions. Casein has the maximum thermostability when bound to the optimum quantity of calcium. This is referred to as the salt balance. Low thermostability is mostly caused by a surplus of calcium. Thus addition of citric acid or phosphoric acid anions will improve the thermostability. The stabilizing salts used, also called sequestering agents, are secondary di-sodium phosphate or tri-sodium citrate.
9.4.

Physical properties of milk

Milk is a very complex polydisperse system. The individual components have various influences on its physical properties. The lactose and salts appear in true water solution. Fat exists in emulsion and partially in suspension, especially in deep chilled milk. The colloidal system is stabilized mainly by phosphoric and citric acid salts. Water is the main component of the colloidal medium. The fat appears as microscopic globules 0.1 - 20 μm in diameter. Casein occurs as submicroscopic particles 10 - 300 nm, albumin and globulin 5 - 15 nm. Lactose and minerals (both as molecules and dissociated salts) have the size 0.4 - 0.5 nm. Milk exhibits properties of both colloidal and true solution’s nature, which explains the various characters of its physical behaviour. Casein, which occurs as a calcium phosphoric complex, is hydrophobic and tends to cluster. It can be precipitated by acids at the isoelectric point, pH 4.6, or enzymatically by rennet. The degree of casein dispersion can be increased by dilution or reduced by intensive agitation, heating or boiling. The phosphoric and citric acid salts act as stabilizers of the colloidal system.
 
The most important physical properties of milk in the production of dried milk powders are viscosity, density and heat stability. Variations in milk composition, mainly the protein/lactose ratio and the salt balance have greatest influence on these properties. Lactose is the main component of milk solids. Therefore the physical properties of lactose are reflected to a great extent in the behaviour of milk.
9.4.1.

Viscosity

Shear stress in a flowing fluid occurs due to friction between adjacent layers, moving at different velocities. The unit of dynamic viscosity is Newton second per square meter, Ns/m², but in industrial practice, the unit of centipoises, cP is used. The viscosity of liquid, untreated milk is influenced mainly by composition, temperature and age. The viscosity of milk concentrates during the evaporation process reflects the viscosity of the incoming milk. However, the viscosity is also influenced by heat treatment (i.e. by the degree of denaturation of whey proteins as caused by the sum of all heat treatment effects), temperature and concentration. Agitation and especially homogenization also increase viscosity.

Milk and milk concentrates are non-newtonian fluids and exhibit thixotropic behaviour. There are two viscosity components: basic viscosity and structural viscosity. The structural component develops during storage (i.e. without movement) progressively with temperature. This viscosity increase in milk concentrates is known as age thickening and can be eliminated almost completely by high shear stress treatment. The viscosity measurement of milk concentrates has to be done at high shear rate or extrapolated to infinite shear rate to get the basic viscosity.
 
Viscosity plays an important role during atomization, since it is one of the decisive factors for the droplets size distribution and thus for the whole drying process. The structural component, unless it is excessively high, does not have too much influence during atomization, which proceeds under high shear stress. On the other hand it can have a strong influence where the milk flow is relatively slow as in the last stages of the evaporator, concentrate heat exchangers etc.
 
High viscosity prior to these treatments may result in further increase of viscosity due to overheating of the boundary layer on the heating surface.
 
The most important factor for increase of basic viscosity is the denaturation of the whey proteins by heat treatment, as expressed by Whey Protein Nitrogen Index (usually called WPNI). This increase in viscosity is caused by the ability of denatured whey proteins to bind an amount
of water up to seven times their own weight. Therefore high heat milk concentrates require drying at higher outlet air temperature than low heat concentrates, if all other conditions are the same, to get powder of the same moisture content.
 
The concentrate viscosity can be reduced by heating. This produces smaller droplets during atomization and thereby lower outlet drying temperature. This effect, however, is more noticeable at concentrations above 46% total solids. It is almost negligible below 44% TS.
The influence of heat treatment on viscosity
Fig.9.5. The influence of heat treatment on viscosity
Extensive studies on viscosities of milk concentrates have been done by Snoeren et al. [4- 8] who concluded that the viscosity depends on the volume fraction of the macromolecular material of milk and the viscosity of milk serum. They used Eiler’s [9] relationship between viscosity and the volume fraction occupied by the proteins. The latter depends on the content of casein, denatured whey protein and native whey protein (and thus on the heat treatment) of the liquid milk, plus concentration and temperature. Fig. 9.5-9.8 show Snoeren’s relationships of the viscosity vs. solids content as influenced by heat treatment (% denaturation of whey proteins), total protein content in non-fat-solids, temperature and homogenization. All these calculations are based on whey protein content of total protein 22.5%, liquid skim milk total solids content 9% with 0.5% fat in total solids, and liquid whole milk total solids content 12.5% with fat content 28% in total solids. Apart from the variable properties, the calculations relate to a total protein in non-fat-solids content 38%, denaturation of whey proteins 50% (medium heat) and concentrate temperature 50°C.
The influence of temperature on viscosity.
Fig.9.6. The influence of temperature on viscosity.
The influence of protein content on viscosity.
Fig.9.7. The influence of protein content on viscosity.
The influence of homogenization on viscosity.
Fig.9.8. The influence of homogenization on viscosity.
Regarding the computation of Snoeren’s relationships, it must be emphasized that it expresses the viscosity as measured with infinite shear rate, i.e. it is the true viscosity without the structural component. The calculations are very sensitive and their accuracy or agreement with measured viscosities is to a great extent dependent on many factors such as calculation of concentrate and milk densities, water density and viscosity, viscosity of lactose solution, volume fractions of various components etc. Nevertheless these calculations are very useful in practice to predict or explain the influence of the various factors involved.
 
These calculations are based on following equations:
[9,1]
[9,2]
[9,3]
[9,4]
[9,5]
[9,6]
[9,7]
[9,8]
[9,9]
[9,10]
[9,11]
[9,12]
The viscosity of cream with various fat contents according to Free [11] is shown in Fig. 9.10.
Viscosity of cream (acc. to free).
Fig.9.9. Viscosity of cream (acc. to free).
9.4.2.

Density

Density is the mass of a certain quantity of a material divided by its volume. It is expressed in kg/m³ or g/cm³ or g/ml. For milk concentrates of known composition, the approximate calculation of the density follows the Hunziker’s [12] equation:
[9,13]

where: %F = percent of fat

%NFS = percent of non-fat-solids

%W = percent water

ρfat = specific gravity of fat (acc. Hunziker 0.93)

ρNFS = specific gravity of non-fat-solids (acc. Hunziker 1.608)

ρwater = specific gravity of water (Hunziker used 1 for t=15°C)

Concentrate density at a temperature, t, can be calculated with a reasonable degree of approximation by:

[9,14]
Gosselin [13] modified Hunziker’s equation, so that it can be used directly over the entire temperature range using for specific gravities of fat, non-fat-solids and water equations:
[9,15]
[9,16]
[9,17]
[9,18]
The equation [9,13] can be used also for calculating the theoretical density of milk powder solids at laboratory temperature. The specific gravities of individual constituents use constant values: 0.94 for fat, 1.52 for non-fat-solids and 1 for water. For non-fat solids of whey powder 1.58 is used.
 
The densities of skim and whole milk concentrates calculated using the modified Gosselin’s relationships are shown on Fig. 9.10 and 9.11 respectively.
Density of skim milk concentrate.
Fig.9.10. Density of skim milk concentrate.
Density of whole milk concentrate with 28% fat in TS.
Fig.9.11. Density of whole milk concentrate with 28% fat in TS.
Density of whey concentrate
Fig.9.12. Density of whey concentrate
The density of whey concentrates can be calculated using the empirical equation:
[9,19]
in which %TS = concentration and t = concentrate temperature. The densities of whey concentrates are given in Fig. 9.12.
9.4.3.

Boiling point

The boiling point of milk under normal atmospheric pressure is about 100.6°C. It increases with higher concentrations. The increase is proportional to the molar concentration of the dissolved components. In practice it can be calculated by the equation:
 
Boiling point increase in
[9,20]
9.4.4.

Acidity

Fresh milk is a complex buffering system of proteins, phosphates, citrates, carbon-dioxide and other minor components. During bacterial activity, lactic acid and other organic acids are created. The acidity of milk is normally determined by titration using NaOH solution and phenolphthalein as indicator (corresponding to pH 8.3). In various countries various expressions of titratable acidity are used. It is usually expressed in ml of NaOH solution, the strength of which and the comparison between various methods are given in Table 9.3. and Table 11.4.
 
In spite of titratable acidity often being expressed “as lactic acid” the acidity of good fresh milk has only negligible amounts of true lactic acid. True lactic acid can be determined by an enzymatic method and gives the real expression of acidity increase due to bacterial activity.
Active acidity is expressed by pH. The pH of normal milk is 6.5 - 6.65 and decreases with increasing temperature. During concentration, the pH value decreases.
A B
°Th °SH °D % 1.a.
°Th - 0.4 0.9 0.0009
°SH 2.5 - 2.25 0.0225
°D 1.11
% lactic acid 111.1 44.44 100 -
Strenght NaOH 0.1 N 0.25 N 0.11 N 0.1 N
°Th=Thörner, °SH=Soxhlet-Henkel, °D=Dornick, %1.a.=% as lactic acid
Table. 9.3. Conversion factors of titratable acidities. (multiply A by factor to get B)
9.4.5.

Redox potential

Redox potential expresses the energy with which a system can oxidize or reduce present or added components. It depends on the presence of oxygen, ascorbic acid, free sulfhydryl groups, trace elements and products of bacterial activity. It is also influenced by acidity.
9.4.6.

Crystallization of lactose

Lactose is quantitatively the major component of milk, as shown in Fig. 9.3. The solubility of lactose in water is relatively low and concentration and drying thus yield a supersaturated solution. This happens already when milk is concentrated to 45 - 50% total solids and has temperature below 45°C. Under normal circumstances during the production of whole or skim milk powders, however, the lactose will not crystallize and will appear in the final product as amorphous lactose or lactose glass, which can be considered a very high viscous or solidified solution. Also the ratio of α- to β-lactose remains unchanged i.e. same as in the concentrate. If, however, some crystal nuclei are present and the concentrate is kept some time in the supersaturated state, crystallization occurs. The same process will take place in moist, warm powder having moisture content higher than about 6 - 7%.
 
From a supersaturated solution at temperatures below 93°C α-lactose will crystallize as monohydrate since it is less soluble than β-lactose. The rate of crystallization is governed by mutarotation, i.e. transformation of β-form into α-lactose. The rate of mutarotation is temperature dependent and is shown on Fig. 9.13.
Mutarotation rate of lactose.
Fig.9.13. Mutarotation rate of lactose.
For the production of normal milk powders the crystallization of lactose is undesirable and has to be avoided. It is very undesirable in the production of whole milk powder. The consequence of lactose crystallization in whole milk powder is an increase of the free fat content because the amorphous lactose, being a continuous phase of milk powder particles, creates a very tight membrane (encapsulation) through which the solvent cannot penetrate thus protecting the fat against extraction (and also against oxidation). Crystallization transfers the lactose from the continuous phase to a discontinuous phase, thereby creating craters and channels, enabling penetration of solvent. This can happen if some nuclei of lactose are present already in the concentrate due to high concentration and low temperature, and possibly also with long holding times before drying and/or if a semi-dried product is held for longer periods with a moisture content above 6 - 7%. On the other hand, high free fat content is an advantage for milk powder for chocolate industry. Establishing intentionally the above conditions makes possible the production of whole milk powder with over 90% free fat content (of total fat content).
 
The presence of crystallized lactose in milk powders can be traced by scanning electron microscopy (SEM) and microphoto techniques in polarized light. An example of lactose crystals in whole milk powder is shown on the polarized light-microphoto Fig. 9.14.
 
The reason for the presence of lactose in whole milk powder was too high moisture content from the first drying stage in an SDI dryer. In baby food powder the reason was not fully dissolved lactose in the mixture prior to drying. In both cases the presence of lactose crystals has resulted in high free fat content. Amorphous lactose is very hygroscopic and is the main reason for hygroscopicity and caking of powders with high content of lactose such as whey or permeate powders. In order to reduce hygroscopicity and caking tendencies, the lactose has to be transformed into α-lactose-monohydrate, which is non-hygroscopic. This is done by pre-crystallization of the concentrate prior to spray drying. It is possible to achieve above 80% pre-crystallization. Even higher crystallization in the final product is possible using a special drying process in which the crystallization, so-called after-crystallization, continues in the moist powder leaving the spray dryer chamber before final drying. Final powder can have over 90% of lactose as α-lactose-monohydrate and such product is non-caking.
Polarized light photos of whole milk powder particles with lactose crystals
Fig.9.14. Polarized light photos of whole milk powder particles with lactose crystals
Polarized light photos of whole milk powder particles with lactose crystals
Fig.9.15. Polarized light photos of whole milk powder particles with lactose crystals
9.4.7.

Water activity

The water activity (aw) of dried milk products is largely a function of moisture content and temperature. The composition and state of the individual components, as influenced by various processing techniques, also play an important role. The composition of the solids is given more or less by the contents of proteins. At low moisture content characterized by aw <0.2 (corresponding for instance to skim milk at <4% moisture) it is the casein, which is the main water absorber. Within the intermediate range of aw >0.2 and <0.6 (about 4 - 16% moisture on skim milk) the sorption behaviour is dominated by the transformation of the physical state of lactose. Above this level the salts have a marked influence.
 
As to the influence of temperature at low moisture levels up to about 12-18%, water activity increases with rising temperature. Above 12% moisture temperature has only minor influence, as reported by Warburton and Pixton [54] and above 18% moisture, i.e. in the area where the salts start to contribute significantly to the water activity, the influence of temperature is opposite.
 
The water activity of milk powders consisting of milk non-fat-solids and milk fat is predominantly controlled by the moisture content expressed on non-fat solids since the fat has no influence. Thus differences in water activities are due mostly to the state of proteins and physical state of lactose.
 
The main interest in water activity relationships concerns the connection between predicting and controlling the shelf life of foods. It has been recognized that the growth of most bacteria is inhibited at water activities lower than 0.9 and for mould and yeast strains between 0.88-0.80. Furthermore, many physicochemical changes as enzymatic reactions, Maillard reaction, lipid oxidation, textural changes, crystallization of carbohydrates, aroma retention etc. are, to a great extent, controlled by water activity.
 
As water activity plays important roles during the dehydration process, knowledge of desorption isotherms can provide useful guidelines for the design, engineering and the control of the drying process. Numerous publications are available in the literature presenting sorption isotherms of various dried dairy products. Attention has been paid to milk powders and other dehydrated products, such as whey powders, whey protein concentrates, caseinates, baby food powders etc.
 
It has been recognized that water activity, beside the moisture content and overall composition, also depends on the pre-treatment to which the material has been subjected prior to and during dehydration.
 
Practically all the isotherms of milk powders, which can be found in the literature, were obtained using the final product as the starting material. This starting material is exposed to air of well-defined and known relative humidity and brought to equilibrium, after which the moisture content is determined. Thus the published isotherms are designated as adsorption, desorption and re-adsorption (second adsorption) isotherms. Establishing equilibrium can very often take weeks and especially during adsorption many changes can take place when the moisture is rising. As to milk powders it is mainly crystallization of lactose which takes place when the moisture level rises to around 7%. The amorphous lactose forms about half of the content of non-fat-solids of milk and about three quarters of whey solids. Transformation to α-lactose-monohydrate, which is almost inert, has a quite dramatic influence on the shape of isotherms. They exhibit a sudden “break” at which the moisture drops at constant water activity. This drop corresponds often to the theoretical amount of water consumed by the crystallization of lactose. However, the test material has been irreversibly changed and any further data during the continuation of absorption, desorption and reabsorption are of little value because it has been conducted on a product different from the starting material.
 
Many mathematical equations, both theoretical and empirical, have been reported in the literature for expressing water sorption isotherms of milk powders. Iglesias and Chirife [14] have compiled from the literature experimental data on water sorption isotherms of hundreds of food products, among them many powdered milk based products. Referring to several mathematical two parameters equations for expressing water sorption isotherms, they applied regression programmes to the published experimental data. 
 
For instance as regards skim milk powder, they applied Halsey’s [15] equation,
[9,21]
expressing the moisture content x on a per cent dry basis, i.e. kg H2O/kg dry matter*100 on the data of Berlin et al. [16] measured at 34°C and found the constants:
 
a = 2.0544 and b = 54.3870 for 1.cycle desorption and,
a = 1.7764 and b = 24.8439 for 2.cycle adsorption (reabsorption).
 
Obviously no attempts were done to express first cycle adsorption due to the mentioned irregularities, i.e. curve break.
 
Two other sorption models often used are the BET (Braunauer, Emmett & Teller) and the GAB (Guggenheim, Anderson & de Boer) models:
 
BET model:
[9,22]

GAB model:

[9,23]
where: aw = water activity
           m = moisture content in g/g total solids
           mo = monolayer moisture content in g/g total solids
           c & k = constants
 
To avoid these problems connected with crystallization of lactose during prolonged exposure to humid atmosphere and to obtain water activity values under the primary desorption. Písecký [17] conducted series of measurements placing the sensor for water activity and temperature directly in the drying chamber. The measurements were conducted on skim milk powder, whole milk powder with 27% fat and two fat filled milk powders with 42 and 46% fat respectively. Based upon regression analysis of the measured values, the following empirical equations have been developed for calculation of water activity during primary desorption, i.e. during the dehydration:
[9,24]
[9,25]
[9,26]
[9,27]
[9,28]
where the constants are:     a = -1.2988 and b = -1.208*103 for [9,24]
                                          c = 1.4626 and d = -3.7668*105 for [9,25]
 
and:                                     %M = percent moisture on wet basis
                                            %F = percent fat on powder
                                            xNFS = moisture content kg/kg non-fat-solids
 
Agreement between measured and calculated values for 144 measurements of samples in the fat content range 0 - 46%, moisture range xNFS 0.0345 - 0.1337 and temperature range 20 - 60°C was expressed by mean relative deviation E = 2.59% and variance v = 0.0758 [17].
 
The empirical equations are for general use in predicting moisture content of milk powders on discharge from the spray dryer from measured values of water activity, temperature and fat content, and vice versa. The equations have been shown applicable over a wide range of all variables. Fig. 9.15 illustrates the moisture content xNFS plotted against water activities at 20, 34 and 80°C. For comparison are shown also published data by Berlin for skim milk powder at 34°C on first cycle adsorption, exhibiting the typical break at aw 0.45, due to crystallization of lactose.
Water activity of milk powder acc. to Písecký with data by Berlin for first cycle absorption.
Fig.9.16. Water activity of milk powder acc. to Písecký with data by Berlin for first cycle absorption.
The water activity is one of the main factors governing many of the phenomena occurring during thermal dehydration, mainly:
 
  1. ease of which water is evaporated from a liquid droplet,
  2. particle temperature history during the whole water removal process (see also section 3.2.3. Droplet temperature and rate of drying),
  3. the equilibrium moisture content which can be achieved under given conditions atinfinite residence time,
  4. the stickiness of the product (sticking temperature) and the outlet conditions (air temperature and moisture content) that can be used to dry without sticking problems occurring.
9.4.8.

Stickiness and glass transition

Powder build-up on drying chamber walls is a well-known phenomenon, caused by certain product properties described with terms as Stickiness or Thermoplasticity.
 
Thermoplasticity is a very descriptive term that implies that the product plasticizes at elevated temperatures. It is well known that increased powder moisture increases the stickiness of the products. The relation between powder moisture and T-out, influenced by other drying parameters as well, is shown later in Fig. 10.1. It shows that increased T-in or % TS - changes that increase the plant capacity - will result in increased powder moisture content and potentially in powder build-up if not compensated for by T-out. Increased ambient air humidity will have the same effect. However, T-out can only be increased within limits, partly for product quality reasons, but also because the powder particles become sticky at a certain T-out due to thermoplastic behaviour, even though the moisture content may be quite low. This phenomenon is shown in Fig. 9.16. It should be emphasized that the sticking curve is a product property depending on product composition only, whereas the T-out – moisture relation (and hence T-particle – moisture relation) as well is dependent on product properties (composition) as drying conditions (T-in, % TS, etc).
 
The sticking curve approach described above is a rather empirical one, but the use of the glass transition concept in the last decade or more has formed a more theoretical basis of understanding the phenomena.
Empirical sticking curve. Relationship between the moisture content, outlet air temperature, particle temperature and the sticking point temperature.
Fig.9.17. Empirical sticking curve. Relationship between the moisture content, outlet air temperature, particle temperature and the sticking point temperature.
A glass is an amorphous, high-viscous liquid in a non-equilibrium state, exhibiting mechanical properties of a solid, but with structural characteristics of a liquid, i.e. contrary to the crystalline state a glass is without any ordered molecular arrangement.
 
Due to the non-equilibrium state, a glass is thermodynamically unstable and can undergo phase transitions. The glass transition in amorphous systems is a reversible change in the physical state from a mechanically solid glass to a visco-elastic, rubbery state, which takes place at the characteristic glass transition temperature, Tg. Tg of a product depends on the product composition and in particular on the presence of plasticizers, of which water is very potent, that is the Tg declines drastically at increased moisture content.
 
In spray dried milk powders the lactose is usually in an amorphous state because the drying takes place so fast that there is no time for the molecular ordering required for crystallization of the lactose. Hence normal milk powders exhibit glass transition. Although the glass transition temperature of a milk powder is not identical to the ‘sticking temperature’, defined as the temperature at which powder build-up in dryers may occur, there is still a relation between Tand the sticking temperature. Indications are that the ‘sticking’ temperature is about 20 - 25°C above the Tg.
 
Glass transition is accompanied by measurable physical changes in viscosity, heat capacity and others. The change in heat capacity can be measured Differential Scanning Calorimetry (DSC).
 
Models describing Tg in binary system (one component and a plasticizer) are available. The Gordon – Taylor model [9,28] has been used extensively, but also the more elaborate Couchman – Karasz equation [9,29] is used.
[9,29]
[9,30]
where: wi = weight fraction of component i,
Tgi = glass transition temperature of component i,
ΔCpi = specific heat change of component i.
 
Table 9.4 shows the glass transition temperature of some food components. It can be seen that lower molecular weight carbohydrates exhibit lower Tg (mono-saccharides < di-saccharides and the higher the DE of maltodextrines, the lower the Tg). The high Tg of starch and low DE maltodextrines explains why these components are good ‘carriers’ for more difficult components in spray drying. The extremely low Tg of lactic acid also explains the difficulties of drying acid whey.
Component Tg(oC)
Fructose 5
Glucose 31
Galactose 32
Sucrose 62
Maltose 87
Lactose 101
Maltodextrin DE 36 (MW ~ 550) 100
Maltodextrin DE 25 (MW ~ 720) 121
Maltodextrin DE 20 (MW ~ 900) 141
Maltodextrin DE 10 (MW ~ 1800) 160
Maltodextrin DE 5 (MW ~ 3600) 188
Starch 243
Lactic acid -60
Water (amorphous) -135
Table. 9.4. Glass transition temperature of different food components
Although the relation between aw of milk powders and Tg is basically sigmoid it has been found to be almost linear in the aw range from 0.2 – 0.65,
[9,31]
Knowledge of the parameters in [9,28] or [9,29] for a given product and drying process together with knowledge of the corresponding desorption isotherm and the glass transition temperature as a function of aw can be very helpful in defining optimized, but still safe drying conditions.

Table of contents

  1. 1.Introduction
  2. 2.Evaporation
    1. 2.1. Basic principles
    2. 2.2. Main components of the evaporator
    3. 2.2.1. Heat exchanger for preheating
    4. 2.2.1.1. Spiral-tube preheaters
    5. 2.2.1.2. Straight-tube preheaters
    6. 2.2.1.3. Preheaters to prevent growth of spore forming bacteria
    7. 2.2.1.3.1. Direct contact regenerative preheaters
    8. 2.2.1.3.2. Duplex preheating system
    9. 2.2.1.3.3. Preheating by direct steam injection
    10. 2.2.1.4. Other means to solve presence of spore forming bacteria
    11. 2.2.1.4.1. Mid-run cleaning
    12. 2.2.1.4.2. UHT treatment
    13. 2.2.2. Pasteurizing system including holding
    14. 2.2.2.1. Indirect pasteurization
    15. 2.2.2.2. Direct pasteurization
    16. 2.2.2.3. Holding tubes
    17. 2.2.3. Product distribution system
    18. 2.2.3.1. Dynamic distribution system
    19. 2.2.3.2. Static distribution system
    20. 2.2.4. Calandria(s) with boiling tubes
    21. 2.2.5. Separator
    22. 2.2.5.1. Separators with tangential vapour inlet
    23. 2.2.5.2. Wrap-around separator
    24. 2.2.6. Vapour recompression systems
    25. 2.2.6.1. Thermal Vapour Recompression – TVR
    26. 2.2.6.2. Mechanical Vapour Recompression - MVR
    27. 2.2.7. Condensation equipment
    28. 2.2.7.1. Mixing condenser
    29. 2.2.7.2. Surface condenser
    30. 2.2.8. Vacuum equipment
    31. 2.2.8.1. Vacuum pump
    32. 2.2.8.2. Steam jet vacuum unit
    33. 2.2.9. Flash coolers
    34. 2.2.10. Sealing water equipment
    35. 2.2.11. Cooling towers
    36. 2.3. Evaporator design parameters
    37. 2.3.1. Determination of heating surface
    38. 2.3.2. Heat transfer coefficient
    39. 2.3.3. Coverage coefficient
    40. 2.3.4. Boiling temperature
    41. 2.4. Evaporation parameters and its influrence on powder properties
    42. 2.4.1. Effect of pasteurization
    43. 2.4.1.1. Bacteriological requirements
    44. 2.4.1.2. Functional properties of dried products
    45. 2.4.1.2.1. Heat classified skim milk powders
    46. 2.4.1.2.2. High-Heat Heat-Stable milk powders
    47. 2.4.1.2.3. Keeping quality of whole milk powders
    48. 2.4.1.2.4. Coffee stability of whole milk powders
    49. 2.4.2. Concentrate properties
  3. 3.Fundamentals of spray drying
    1. 3.1. Principle and terms
    2. 3.1.1. Drying air characteristics
    3. 3.1.2. Terms and definitions
    4. 3.1.3. Psychrometric chart
    5. 3.2. Drying of milk droplets
    6. 3.2.1. Particle size distribution
    7. 3.2.2. Mean particle size
    8. 3.2.3. Droplet temperature and rate of drying
    9. 3.2.4. Particle volume and incorporation of air
    10. 3.3. Single-stage drying
    11. 3.4. Two-stage drying
    12. 3.5. Expansion of air bubbles during drying
    13. 3.6. Extended Two-stage drying
    14. 3.7. Fluid bed drying
  4. 4.Components of a spray drying installation
    1. 4.1. Drying chamber
    2. 4.2. Hot air supply system
    3. 4.2.1. Air supply fan
    4. 4.2.2. Air filters
    5. 4.2.3. Air heater
    6. 4.2.3.1. Indirect: Gas / Electricity
    7. 4.2.3.2. Direct heater
    8. 4.2.4. Air dispersers
    9. 4.3. Feed supply system
    10. 4.3.1. Feed tank
    11. 4.3.2. Feed pump
    12. 4.4. Concentrate heater
    13. 4.4.1. Filter
    14. 4.4.2. Homogenizer/High-pressure pump
    15. 4.4.3. Feed line
    16. 4.5. Atomizing device
    17. 4.5.1. Rotary wheel atomizer
    18. 4.5.2. Pressure nozzle atomizer
    19. 4.5.3. Two-fluid nozzle atomizer
    20. 4.6. Powder recovery system
    21. 4.6.1. Cyclone separator
    22. 4.6.2. Bag filter
    23. 4.6.3. Wet scrubber
    24. 4.6.4. Combinations
    25. 4.7. Fines return system
    26. 4.7.1. For wheel atomizer
    27. 4.7.2. For pressure nozzles
    28. 4.8. Powder after-treatment system
    29. 4.8.1. Pneumatic conveying system
    30. 4.8.2. Fluid bed system
    31. 4.8.3. Lecithin treatment system
    32. 4.8.4. Powder sieve
    33. 4.9. Final product conveying, storage and bagging-off system
    34. 4.10. Instrumentation and automation
  5. 5.Types of spray drying installations
    1. 5.1. Single stage systems
    2. 5.1.1. Spray dryers without any after-treatment system
    3. 5.1.2. Spray dryers with pneumatic conveying system
    4. 5.1.3. Spray dryers with cooling bed system
    5. 5.2. Two stage drying systems
    6. 5.2.1. Spray dryers with fluid bed after-drying systems
    7. 5.2.2. TALL FORM DRYER™
    8. 5.2.3. Spray dryers with Integrated Fluid Bed
    9. 5.3. Three stage drying systems
    10. 5.3.1. COMPACT DRYER™ type CDI (GEA Niro)
    11. 5.3.2. Multi Stage Dryer MSD™ type
    12. 5.3.3. Spray drying plant with Integrated Filters and Fluid Beds - IFD™
    13. 5.3.4. Multi Stage Dryer MSD™-PF
    14. 5.3.5. FILTERMAT™ (FMD) integrated belt dryer
    15. 5.4. Spray dryer with after-crystallization belt
    16. 5.5. TIXOTHERM™
    17. 5.6. Choosing a spray drying installation
  6. 6.Technical calculations
    1. 6.1. Evaporation and product output
    2. 6.2. Heating of atmospheric air
    3. 6.3. Mixing of two air stream
    4. 6.4. Dry air rate, water vapour rate and air density
    5. 6.5. Air velocity in ducts
    6. 6.6. Air flow measurements
    7. 6.7. Barometric distribution law
    8. 6.8. The heat balance of a spray dryer
  7. 7.Principles of industrial production
    1. 7.1. Commissioning of a new plant
    2. 7.2. Causes for trouble-shooting
    3. 7.3. Production documentation
    4. 7.3.1. Production log sheets
    5. 7.3.2. General maintenance log book
    6. 7.3.3. Product quality specification
    7. 7.3.4. Operational parameter specification
    8. 7.4. Product quality control
    9. 7.4.1. Process quality control
    10. 7.4.2. Final quality control
  8. 8.Dried milk products
    1. 8.1. Regular milk powders
    2. 8.1.1. Regular skim milk powder
    3. 8.1.2. Regular whole milk powder
    4. 8.1.3. Whole milk powder with high free fat content
    5. 8.1.4. Butter milk powder
    6. 8.1.4.1. Sweet butter milk powder
    7. 8.1.4.2. Acid butter milk powder
    8. 8.1.5. Fat filled milk powder
    9. 8.2. Agglomerated milk powders
    10. 8.2.1. Agglomerated skim milk powder
    11. 8.2.2. Agglomerated whole milk powder
    12. 8.2.3. Instant whole milk powder
    13. 8.2.4. Agglomerated fat filled milk powder
    14. 8.2.5. Instant fat filled milk powder
    15. 8.3. Whey and whey related products
    16. 8.3.1. Ordinary sweet whey powder
    17. 8.3.2. Ordinary acid whey powder
    18. 8.3.3. Non-caking sweet whey powder
    19. 8.3.4. Non-caking acid whey powder
    20. 8.3.5. Fat filled whey powder
    21. 8.3.6. Hydrolysed whey powder
    22. 8.3.7. Whey protein powder
    23. 8.3.8. Permeate powders
    24. 8.3.9. Mother liquor
    25. 8.4. Other Dried Milk Products
    26. 8.5. Baby food
    27. 8.6. Caseinate powder
    28. 8.6.1. Coffee whitener
    29. 8.6.2. Cocoa-milk-sugar powder
    30. 8.6.3. Cheese powder
    31. 8.6.4. Butter powder
  9. 9.The composition and properties of milk
    1. 9.1. Raw milk quality
    2. 9.2. Milk composition
    3. 9.3. Components of milk solids
    4. 9.3.1. Milk proteins
    5. 9.3.2. Milk fat
    6. 9.3.3. Milk sugar
    7. 9.3.4. Minerals of milk
    8. 9.4. Physical properties of milk
    9. 9.4.1. Viscosity
    10. 9.4.2. Density
    11. 9.4.3. Boiling point
    12. 9.4.4. Acidity
    13. 9.4.5. Redox potential
    14. 9.4.6. Crystallization of lactose
    15. 9.4.7. Water activity
    16. 9.4.8. Stickiness and glass transition
  10. 10.Achieving product properties
    1. 10.1. Moisture content
    2. 10.2. Insolubility index
    3. 10.3. Bulk density, particle density, occluded air
    4. 10.4. Agglomeration
    5. 10.5. Flowability
    6. 10.6. Free fat content
    7. 10.7. Instant properties
    8. 10.7.1. Wettability
    9. 10.7.2. Dispersibility
    10. 10.7.3. Sludge
    11. 10.7.4. Heat stability
    12. 10.7.5. Slowly dispersible particles
    13. 10.7.6. Hot water test and coffee test
    14. 10.7.7. White Flecks Number (WFN)
    15. 10.8. Hygroscopicity, sticking and caking properties
    16. 10.9. Whey Protein Nitrogen Index (WPNI)
    17. 10.10. Shelf life
  11. 11.Analytical methods
    1. 11.1. Moisture content
    2. 11.1.1. Standard oven drying method (IDF Standard No.26-1964 [32])
    3. 11.1.2. Free moisture
    4. 11.1.3. Total moisture
    5. 11.1.4. Water of crystallization
    6. 11.2. Insolubility index
    7. 11.3. Bulk density
    8. 11.4. Particle density
    9. 11.5. Scorched particles
    10. 11.6. Wettability
    11. 11.7. Dispersibility
    12. 11.8. Other methods for determination of instant properties
    13. 11.8.1. Sludge
    14. 11.8.2. Slowly dispersible particles
    15. 11.8.3. Hot water sediment
    16. 11.8.4. Coffee test
    17. 11.8.5. White flecks number
    18. 11.9. Total fat content
    19. 11.10. Free fat content
    20. 11.11. Particle size distribution
    21. 11.12. Mechanical stability
    22. 11.13. Hygroscopicity
    23. 11.14. Degree of caking
    24. 11.15. Total lactose and α-lactose content
    25. 11.16. Titratable acidity
    26. 11.17. Whey Protein Nitrogen Index (WPNI)
    27. 11.18. Flowability (GEA Niro [31])
    28. 11.19. Lecithin content
    29. 11.20. Analytical methods for milk concentrates
    30. 11.20.1. Total solids
    31. 11.20.2. Insolubility index
    32. 11.20.3. Viscosity
    33. 11.20.4. Degree of crystallization
  12. 12.Troubleshooting operations
    1. 12.1. Lack of capacity
    2. 12.2. Product quality
    3. 12.3. Deposits in the system
    4. 12.4. Fire precaution
    5. 12.5. Principles of good manufacturing practice
    6. 12.6. The use of computer for quality control and trouble-shooting
  13. References
Reference: Schlünder,E.U.:Dissertation Techn.Hochschule Darmstadt D 17, 1962.