Contact us
10.

Achieving product properties

About 40 years ago milk powder quality was evaluated using the same criteria as for liquid milk products. The aim of such evaluation was:


a) to ensure that the final product met the specified composition i.e. fat content, total solids content and possibly content of other ingredients, if any (sugar etc.).
b) To ensure that the product during processing was not affected by some undesirable microbiological or chemical processes.


The only properties related specifically to milk powder were solubility index and content of scorched particles. Later on it was found that it was possible to influence the propertiesof the final product by certain pre-treatment processes, by choosing certain conditions for evaporation and spray drying, by dividing the water removal process into spray drying and fluid bed drying and by applying various after-treatment processes. This resulted in the development of many new products with properties tailor-made for a special end-use, i.e. having special functional properties.

The properties of the final products are influenced by a number of factors involving quality and composition of the raw milk and operating conditions applied. As some of the factors are subjected to both seasonal and daily variations, it is necessary to frequently control those properties, which might be affected by those variations and to make the appropriate
correction to the operation parameters.

10.1.

Moisture content

The moisture content of the final product is a property, which is required by the product specification, defining the permissible maximum level (for instance max 3 %). From the point of view of functionality, too high a final moisture may result in inferior shelf life due to Maillard reaction, creation of lumps, and possibly bacteriological problems or growth of yeast and mould. Thus the moisture contents for individual products and that required by legal specification have been chosen with respect to the above.

The final moisture content is important from the point of view of final powder quality and achieving a standard product. From the economical point of view, it is important to operate as close as possible to the limit. In large spray drying installations each 0.1% of moisture can represent a great sum of money on a yearly basis. However, not less important is the intermediate moisture content of a product leaving the individual processing steps during two
stage or three stage drying. The intermediate moisture content has great influence on such properties as solubility index, bulk density, particle density, agglomeration (i.e. particle size distribution) and also on overall drying economy.


In a single-stage dryer the final moisture is influenced by combination of factors involving properties of the feed (concentration, temperature and viscosity), conditions of atomization (rotating wheel atomizer speed or atomization pressure with pressure nozzles) and conditions of the drying air (inlet and outlet temperature and absolute humidity). The magnitude of influence of some factors on the moisture levels ex-drying chamber is known. For instance,
as shown in Fig. 10.1, an increase of inlet temperature by 10°C, ambient air absolute humidity by 2.8 g/kg or total solids of the feed by 1% and reduction of the outlet temperature by 1°C will result in an increase of powder moisture by 0.2% with skim milk and by 0.16% with whole milk. The direction of change for other factors is indicated by the ±-symbols but not the exact magnitude.

Influence of various factors on moisture content
Fig.10.1. Influence of various factors on moisture content
Fig. 10.1 shows how the moisture content is influenced by variations of a number of other factors. The absolute value of the outlet temperature is determined also by a number of other factors, mainly the efficiency of the mixing of the cloud of droplets with the drying air. 
Fig. 10.2 shows the powder moisture (ex-chamber) as a function of outlet temperature for both skim and whole milk at the inlet temperature 180°C and feed concentration 48%.
The relation between powder moisture and outlet temperature
Fig.10.2. The relation between powder moisture and outlet temperature

In a two stage drying system the intermediate moisture should be kept reasonably constant because, as mentioned above, it influences several other properties. The importance of the second drying stage is the fine adjustment of the

final moisture below the rejection level from the standard quality point of view, but at the same time it should be as close as possible to that level from the economical processing point of view. Furthermore, it is also important during the second drying stage and cooling stage to ensure continuous reduction of moisture, as the powder passes through the drying/cooling fluid bed system. As mentioned in section 3.6, the milk powders, when cooled down to low final temperatures, can pick up moisture from the cooling air.

Depending on the type of product and humidity of the cooling air, this process may start already at a powder temperature of 34-40°C. Therefore it is advisable to check the powder moisture content both before and after the cooling section. The traditional way of in-process moisture control consists of sampling powder from the various stages of processing at regular intervals and checking the moisture by fast routine methods. The in-line infrared measurement is now more and more common. This can be combined with direct in-line moisture control and in computer controlled spray dryers even with feed-forward system adjusting the set-point of the outlet temperature according to variations of the factors shown on Fig. 10.1. Such a system is able to keep the standard deviation of the final moisture below 0.1%.
10.2.

Insolubility index

One of the first defects observed in milk powders was the presence of some insoluble material when centrifuging reconstituted powders. It was especially the case with whole milk powders. Waite and White [18] concluded that protein was carried into the fat layer and fat into the sediment layer in quantities directly related to the degree of insolubility. Furthermore they concluded that a well washed sediment consists mainly of calcium caseinate together with
calcium and phosphorus in the same proportion as in tri-calcium-phosphate Ca3(PO4)2, with most of the calcium-phosphate probably being in the form of a casein-phosphate complex. Howat [19] and Wright [20] concluded that the creation of an insoluble form of milk protein during the drying process is most probably identical in nature to the ordinary heat coagulation of a protein.
 
As mentioned above, the insolubility problem is much more severe with whole milk than with skim milk and is even more emphasized by homogenization. In more recent work Mol [21] came to the conclusion that whole milk concentrate is much more sensitive to drying temperatures, if the concentrate is standardized by “casein cream” (i.e. normal cream) than when using “casein-free cream” (based on whey proteins). Therefore he concluded that the impairment of the heat stability is due to casein micelles absorbed on the fat globule membrane.
 
There have been several methods elaborated for determining the insolubility of milk powders, some of them gravimetric. For routine purposes the most widely used procedure was developed in USA by American Dry Milk Institute and illogically called Solubility Index. This method was modified and defined with more precision by IDF (International Dairy Federation) in 1988. In order to distinguish this method from the original ADMI method and also to follow a more logical approach it was named Insolubility Index.
 
The main factor controlling the insolubility index is the particle temperature during the first drying stage from initial feed moisture content down to below about 10%. It is said that the most critical phase of first stage drying is when the powder moisture lies between 20 and 10%, whereby the critical factor is the powder temperature. The factors influencing the powder temperature are in fact all those shown in Fig. 10.1, i.e. all factors increasing the outlet temperature and thereby the whole profile of the particle temperature. Thus, when faced with insolubility index problems, attention must be given to all factors that increase the viscosity, droplet size and generally the outlet temperature. Means to reduce insolubility index include:
 
a) reduce pasteurization effect, i.e. lower the temperature and/or reduce the holding time,
b) increase the feed temperature by concentrate heating,
c) avoid long holding of the concentrate before drying and especially after reheating,
d) reduce the protein content by adding lactose,
e) reduce the homogenization to the lowest possible level and use two stage homogenization in case of fat containing powders,
f) reduce total solids content of the feed,
g) apply higher atomization pressure with nozzles or higher speeds with atomizer wheels,
h) reduce the air inlet temperature,
i) reduce the outlet temperature.
The influence of outlet temperature on insolubility index at constant inlet temperature of 200°C and 48% concentrate total solids
Fig.10.3. The influence of outlet temperature on insolubility index at constant inlet temperature of 200°C and 48% concentrate total solids
The above are just general rules and the decision which of them has to be applied depends on many other circumstances. However, in most cases just one of them is often sufficient to solve the problem. The strongest factor controlling the insolubility index is the powder temperature and this is mostly influenced by the outlet temperature.
 
The relationship of insolubility index and outlet temperature for both skim and whole milk is shown on Fig. 10.3. 
 
The rules above for troubleshooting insolubility index problems are valid generally. However, various types of dryers, even when operating with identical conditions, can produce powders of very different insolubility index. A very important factor is namely the efficiency of mixing the atomized droplets with the drying air, i.e. the design and the adjustment of the air disperser. Agglomeration also has a slight adverse effect on insolubility index as this affects the rate of evaporation by reduction of the evaporative surface and thereby increasing the outlet temperature.
10.3.

Bulk density, particle density, occluded air

Bulk density expresses the weight of a volume unit of a powder and in practice is expressed in g/cm3, kg/m3 and more seldom g/100ml. The reciprocal value of bulk density, often incorrectly called “bulk density” as well, is the bulk volume and is expressed as volume in ml of 100 g of powder. Bulk volume is also often used in the milk powder industry, as it is a value received directly from the analytical method.
 
Bulk density of milk powders is a very important property, both from the point of view of economy, functionality and market requirements. When shipping milk powder in bulk over long distances the producer is interested in high bulk density to reduce shipping costs, since in most cases transportation costs relate to volume. Also, high bulk density saves packaging material for a given weight shipment. 
 
Low bulk density may be interesting from the marketing point of view, so that larger amounts of powder per given weight are seen on the shelves of a supermarket than in a package of a competitive brand. Furthermore low bulk density as achieved by agglomeration is an important factor influencing other powder properties, mainly flowability and instant properties.
 
Currently the bulk density is determined by measuring the volume of 100 g of powder in a graduated glass cylinder of 250 ml after loose filling, followed by tapping manually or using apparatus designed solely for this purpose (10, 100, 600 or even 1250 times). See also 11.3. The bulk density of milk powders is a very complex property being the result of many other properties and being influenced by a number of factors shown in Fig. 10.4. The primary factors determining bulk density are:
 
a) the density of the solids (see 3.1.2),
b) the amount of the air entrapped in the particles (occluded air) or the particle density,
c) the sphericity of the particles, determining the amount of interstitial air, i.e. the air between the particles or agglomerates, inclusive of the air inside porous agglomerates.
 
Ideal spherical-shaped particles or agglomerates create low content of interstitial air resulting in higher bulk density powders, while irregularly shaped agglomerates with attached smaller particles lead to lower bulk density or bulky products. The content of occluded air together with the density of solids (given by the composition of the solids) determines the particle density. This together with the content of interstitial air results in bulk density of the final powder. For non-agglomerated powders the content of interstitial air is given exclusively by the particle size distribution. The broader the size distribution, the higher is the bulk density.
The factors influencing bulk density
Fig.10.4. The factors influencing bulk density
The mechanism for creation of occluded air is as follows: high protein content and especially non-denatured whey proteins support the feed foaming ability. Therefore low-heat products always have higher contents of occluded air than high-heat products. Aeration of the feed before spray drying, i.e. vigorous agitation or often just by filling the balance tank from the top with liquid falling onto the surface incorporates air into the feed. During atomization with rotating wheels, air is drawn into the feed, the extent depending upon wheel type (see 4.4.1). This effect is, however, negligible with pressure nozzles. It is difficult to predict the bulk density of agglomerated powders from particle size distribution. The content of interstitial air is a dominating factor here, overriding the influence of other factors. The final factor controlling the bulk density, especially of agglomerated powder and to a much lesser extent also of nonagglomerated powders, is attrition as a result of mechanical treatment of the powder during pneumatic transport, vigorous fluidization or even packaging in bag- or can-filling machines. Therefore mechanical treatment during production of high quality instant products must be minimised.
Influence of WPNI and % TS on various skim milk powder properties
Fig.10.5. Influence of WPNI and % TS on various skim milk powder properties
The series of graphs in Fig. 10.5 demonstrate interesting work conducted by Mol [22]. The full lines here show the viscosity of the skim milk concentrates and various powder properties when working with constant concentration and using heat treatment resulting in values of WPNI 2, 4 and 6 respectively.
 
The dotted lines show the same. However, here the viscosity was kept constant while the concentration was varied. Otherwise the drying conditions were identical. It can be seen that with constant total solids content from low heat to high heat:
 
a) viscosity increases,
b) final moisture increases as the denatured whey proteins formed by high heat treatment bind better the moisture,
c) occluded air content decreases as high heat treatment reduces the foaming properties,
d) therefore bulk density increases and,
e) solubility index decreases.
 
The latter is probably due to particle temperatures being lower due to higher moisture content in spite of an identical outlet temperature. With constant viscosity the following trends can be observed:
 
a) the total solids content of the concentrate is lower with high heat treatment,
b) the final moisture approaches a constant value and the same goes for occluded air content and bulk density,
c) the solubility index decreases due to lower feed concentration.
The occluded air content has also a considerable influence on bulk density, as it increases the particle volume, i.e. decreasing particle density. High bulk density is desirable, for instance, for skim milk powder used in the Far East for recombining purposes. The amount of occluded air, as illustrated in Fig. 10.4, depends on a number of factors. One is mode of atomization, and in this respect pressure nozzles are superior to wheels. Other factors include conditions of the feed, especially the degree of denaturation of whey proteins, concentration and temperature. Decades ago with old-fashioned recirculation evaporators yielding a concentrate of low solids content, it was known that cooling of the concentrate resulted in high bulk density. Nowadays when working with total solids content of skim milk concentrate of around 50%, it is advantageous to heat the concentrate up to 80°C in order to get a low occluded air content
and thereby a high bulk density. Fig. 10.6. explains this phenomenon.
Whipping ability of skim milk concentrate at various % TS and temperatures
Fig.10.6. Whipping ability of skim milk concentrate at various % TS and temperatures
Medium heat skim milk concentrates of 40, 44 and 50% total solids were whipped at various temperatures in a similar way as when testing whipping ability of egg albumen. Results indicate clearly that a concentrate of low solids content is less whippable (i.e. is able to incorporate less air) when cooled down, while a concentrate of high solids content exhibits an opposite effect. Concentrating to high solids and heating to high temperature is definitely favourable. The incorporation of air into the concentrate may take place either during transfer between the evaporator and spray dryer or during the atomization. 
 
Fig. 10.7. shows the contribution of the volume of occluded air to the total volume of the powder as a percentage. Under certain circumstances, the volume of occluded air can occupy up to about 10% of the total volume. Thus in order to obtain high bulk density powder, it is important to pay attention to the problem of occluded air as outlined above.
The contribution of occluded air (expressed as particle density) to the total powder volume of skim milk powder
Fig.10.7. The contribution of occluded air (expressed as particle density) to the total powder volume of skim milk powder
A similar relationship has been found by Westergaard [23] who expressed the bulk density of skim milk powder directly as a function of heat treatment (expressed as WPNI - see Fig. 10.8.).
Influence of heat treatment (as WPNI) on bulk density of skim milk powder
Fig.10.8. Influence of heat treatment (as WPNI) on bulk density of skim milk powder
The relationship seems to be close to linear, and as in a previous example, the lower bulk density at high WPNI (low heat treatment) is due to high foaming ability and therefore higher content of occluded air. In order to get high bulk density of low heat skim milk powder, it is therefore even more important to avoid incorporation of air into the concentrate prior to and during atomization, hence the preferred use of pressure nozzles.
 
The occluded air appearing as bubbles inside milk powder particle, often called vacuoles or entrapped air, is therefore one of the most important factors for controlling bulk density. Its volume is usually expressed in ml/100g powder or in percent of powder total volume.
High or low content of occluded air may be either an undesirable or wanted property depending on the powder bulk density specification. For high bulk density specification, high contents of occluded air are unwanted. The factors of creation of occluded air in connection with manufacture of high bulk density powders have been discussed previously.
 
If low bulk density is required, the powder volume can be increased by agglomeration (see next section) or by increase of the content of occluded air. With rotating wheel atomization, a small increase of the volume can be achieved by changing to a straight vane wheel from a curved vane wheel. Injection of carbon dioxide (CO2) into the feed before atomization is more effective and controllable. CO2 injection can be done with a wheel atomizer. However, much better control of the injected amount of gas is achieved with a pressure nozzle atomizer. For wheel atomizer operation, CO2 may be injected in slight excess prior to the feed tank. When dosing directly into the feed line, overdosing must be avoided, as it will cause pulsation to the feed flow to the atomizer. With a pressure nozzle atomizer gas is injected into the pipeline prior to or after the high pressure pump. Commercial units, such as the GEA Niro DENSISET™ for CO2 injection on the high pressure side, is available.
 
CO2 is soluble in water even under atmospheric conditions, but its solubility can be increased by increasing the pressure or lowering the temperature. After atomization the dissolved CO2 will be released from solution as a gas, expand and blow up the droplets. For obtaining the best results, the feed must be able to form voluminous and stable foam containing for instance non-denatured whey proteins and maltodextrins.
 
The solubility of CO2 in water at various temperatures and various pressures is shown in Fig. 10.9 and Fig. 10.10, respectively.
Solubility of CO2 in water at various temperatures
Fig.10.9. Solubility of CO2 in water at various temperatures
Solubility of CO2 in water at various pressures
Fig.10.10. Solubility of CO2 in water at various pressures
Instead of CO2 also N2O can be used, but this is much more expensive and has therefore no practical application. Nitrogen has also been used for so-called foam spray drying. Nitrogen is, however, hardly soluble in water, and therefore it can be used only with high pressure nozzle atomization requiring dosing at high pressure into the high pressure line (between high pressure pump and nozzles).
 
It is often believed that the higher the mean particle size, the higher the powder volume, i.e. the lower bulk density, but this is not always true. In a powder consisting of ideal spheres of the same size, packed in an ideal way in tetrahedron pattern, the solids would occupy 74% of the volume regardless particle size. The bulk density of such powder when considering air-free particles will be theoretically about 1.2 and 1.4 g/ml for whole and skim milk respectively, i.e. about 70% higher than experienced in practice. A distribution of different particle size when still maintaining the ideal spherical shape would contribute to even heavier powder. Thus the main factor controlling bulk density is not so much mean particle size, but particle shape.
 
From the above it is obvious that to obtain high bulk density, the powder must consist of airfree particles having as near as possible spherical shape with smooth surfaces and appropriate particle size distribution without agglomeration. This can be best achieved using preheated feed of high total solids content and a two stage drying method. The effect of the latter is more a compensation of the influence of high concentration on other powder properties (mainly insolubility index). Otherwise the influence of two stage drying on bulk density is to reduce the volume of occluded air, however on the account of deforming spherical shape due to shrinking of the particle. Therefore, if all other conditions are the same, the nozzle atomization is superior to wheel atomization where the particles are subjected to higher shrinkage due to higher contents of air in the atomized droplets, i.e. their shape deviates more from an ideal sphere. Under all circumstances, two stage dried powders are always slightly agglomerated and this has an adverse effect on bulk density. The agglomerate bindings are, however, in this case weak requiring only gentle mechanical treatment, like pneumatic or blow line transport to break them down.
 
The main tools for obtaining a low bulk density are agglomeration and increasing the occluded air content. An extremely low bulk density is achieved by the above discussed CO2 atomization. The agglomeration is discussed below.
10.4.

Agglomeration

Agglomeration is the formation of porous clusters of single particles with the aim to increase the volume of interstitial air of a powder, which is one of the main factors for obtaining easily dispersible and easily dissolving - so-called instant - powders. Generally the agglomeration is achieved by intentional collisions between wet particles and dry particles. It can be done either by the so-called straight-through or rewet method. Straight-through agglomeration means agglomeration during the drying process, while for rewet agglomeration the ingoing material is a dried non-agglomerated powder.
 
Primary and secondary agglomeration can be distinguished as follows:
Types of agglomeration
Fig.10.11. Types of agglomeration
1. Primary agglomeration takes place through coincidental collisions within the atomizer cloud between droplets of various sizes and in various stages of drying (see Fig. 10.11.).
 
a) spontaneous primary agglomeration appears in any atomizer cloud. The higher the amount of feed flowing through a single atomizing device, the higher is the probability of collisions.
 
b) forced primary agglomeration can be achieved by directing atomizer clouds of two or more atomizing nozzles (this type of agglomeration is only possible with nozzles) towards each other.
 
2. Secondary agglomeration takes place through coincidental collisions inside the atomizer cloud between droplets of various sizes and in various stages of drying and dry particles.
 
c) spontaneous secondary agglomeration means agglomeration by coincidental recycling of dry particles into the atomizer cloud. It takes place due to the turbulent movement of the drying air in the drying chamber. It is most evident in chambers where air dispersers cause air rotational flow, but it is weak in chambers with streamline air flow (plug flow).

d) forced secondary agglomeration is obtained by classifying the agglomerated powder, i.e. separating the non-agglomerated particles and re-introducing them back into the atomizing cloud.

Agglomeration is a complex process and its effect on mean particle size and powder bulk density depends on the equilibrium between several partial processes as shown schematically in Fig. 10.12. Agglomeration in this model means exclusively the formation of agglomerates in the atomizer cloud, and agglomeration efficiency means here the percentage of agglomerates of total powder after this stage. As mentioned in the above scheme, the efficiency of agglomeration depends on a number of factors, an important one being the amount of recycled fines.

Model of fines generation and recirculation
Fig.10.12. Model of fines generation and recirculation
The agglomeration stage is followed by separation, the aim of which is to separate the fines (or more correctly the non-agglomerated particles) from agglomerates. Separation efficiency means here the percentage of fines (of total fines) separated from the powder, leaving the drying chamber together with the exhaust air and passing to the cyclone for separation.
 
The fraction of agglomerated powder leaving the chamber has usually an excess of moisture and has to be dried to final required value. This process takes place in a fluid bed. This can be either stationary integrated within the drying chamber base or an external vibrating fluid bed. The fluid bed has several functions. Besides removing the excess of moisture it also classifies the powder, i.e. to remove the excess of fines. An unwanted side effect, which takes place during fluidization, is attrition. Some attrition is unavoidable yet acceptable, but excessive attrition has an adverse effect on particle size, and especially on the shape of agglomerates where grinding off the appended particles (which form the distance bridges important for high volume) and “polishing” their surface to round shape takes place. The result is an excessive loss of volume, i.e. too high bulk density. The extent of attrition depends on the conditions of fluidization and on the type of perforated plate. The main characteristics of perforated plates are the percentage of free area and the number of holes per given area determining the size of individual holes. Anticipating that the fluidizing velocity and the gas rate are given by the required duty (see section 4.7.2.) the total fluid bed area is also fixed. Too small a free area of the perforated plate will require too high a pressure in the plenum below the plate. This results in a high gas inlet velocity through the individual holes - so-called jet velocity. Another expression for the destructive effect of fluidization on agglomerates is momentum, which is the jet velocity multiplied by the mass of gas through one hole. These are the two factors most important for attrition, along with the mechanical stability of the treated powder, which also plays an important role.
 
The mechanical stability depends mainly on the composition of the powder and on the structure of the agglomerates. As to composition, stable agglomerates are obtained with powders containing carbohydrates creating a continuous phase of particle solids, which are sticky in wet state and relatively hard when dry. Lactose is an excellent binding agent for agglomeration and its natural content in milk powders is sufficient for stable agglomeration. The conditions of agglomeration, i.e. mainly the moisture content of droplets at the moment of collision have an influence on the type of agglomerates and on their mechanical stability and dispersibility in water. Too stable a powder can be difficult to disperse or even dissolve, and a too well dispersible powder may have poor mechanical stability. The types of agglomerates can be described as follows (see Fig. 10.13.):
Types of agglomerates
Fig.10.13. Types of agglomerates
a) Onion: These are created when droplets of very high moisture, i.e. before any, or only very little, drying has taken place, contact the recycled fines. These droplets, still very fluid, just cover the surface of the fines, gradually building up layers and increasing the size of the original particle, thus forming a structure that resembles an onion. This type of “agglomerate” is obtained intentionally by the process called spray-fluidization, in which liquid feed is sprayed onto fluidized powder in a fluid bed. Obviously, such particles are of relatively large size, have high mechanical stability, but are difficult to dissolve and are degraded by heat. Therefore this type of agglomeration is not desirable for dairy powders.
 
b) Raspberry: These are created if large droplets of high moisture collide with a high amount of fines. As with the onion type this type of agglomeration does not offer well dispersible, soluble or voluminous milk powders.
 
c) Grape: These represent the proper structure of agglomerates for milk powders and are created by mutual collisions of wet particles and fines. The grape structure can be either loose or compact depending on the moisture content at the moment of collision. The compact grape structure is characterized by agglomerates having small amounts of interstitial air, and therefore the powders have low volume but high mechanical stability, whereas the loose grape type agglomerates support high volume, but they are more sensitive to mechanical treatment. The latter exhibits the very best reconstitution properties as to dispersibility, absence of slowly dispersible particles, fast rate of hydration etc.
 
Generally, the higher the moisture content of the droplets at the point of collision with the returned fines, the larger and more stable are the agglomerates. Therefore two stage systems agglomerate better than single stage systems and three stage systems are superior to both of them. The distance from the atomizing device at which the fines are introduced into the cloud is also important. On the other hand, too large and too stable agglomerate can cause problems with reconstitution especially as to instant properties. Therefore the optimum lies between the loose and compact grape structure ensuring reasonably good mechanical stability without too much loss of reconstitution properties.
10.5.

Flowability

Good flowability is a property, which definitely increases the marketing value of a product, since it can be seen and compared visually with a competitive product. In some instances, good flowability is the necessary property of a product for its intended utilization, e.g. powders to be used in vending machines for various hot or cold drinks, for feeding calves by means of automatically working reconstituting and feeding apparatus (artificial cow machine) or for further industrial processing, equipped with mechanical handling and dosing devices.
 
On the other hand, a better flowability is associated with higher bulk density, especially the loose value. The tapped-to-constant bulk density value remains almost unchanged. Different types of milk powder exhibit very different degrees of free-flowing ability, and the factors influencing this property can be summarized as follows:
 
a) important factors in non-agglomerated powders are the particle size, shape and structure of the surface. Large mean particle size, narrow particle size distribution, spherical shape and smooth surfaces are the factors contributing to better flowability. Pressure nozzle powders are superior in this respect to rotary wheel powders due to their lower occluded air content.
 
b) agglomeration improves the free-flowing properties and again the same factors as mentioned above apply. The shape of agglomerates can exhibit even greater deviation from a spherical shape of single particles, and therefore long chain-type agglomerates should be avoided. Low amount of fines is also an important condition.
 
c) increasing fat content of the milk powder reduces flowability and it is well known that skim milk powder is more free-flowing than whole milk powder. On the other hand, it has been shown that with milk powders of very high fat content 65-80%, powders with fat content in the upper part of this range had better flowability than those in the lower.
 
d) a very detrimental effect on flowability is a high free fat content, especially if low melting fat is involved. Therefore lecithin treatment of fat-containing powders used to achieve improved wettablility adversely affects flowability.
 
e) addition of free-flowing agents improves flowability. However, there are limitations to the type of product to which they can be applied. Typical free flowing agents include sodiumaluminium silicates, calcium silicate, calcium phosphate, pre-crystallized whey powder or lactose as -lactose-monohydrate.
10.6.

Free fat content

The free fat of fat containing milk powders is (apart from exceptions discussed later) an undesirable property, jeopardizing the keeping quality due to fast development of oxidized flavour and tallowiness. This causes unpleasant appearance to reconstituted solutions with fat layers forming on the surface. It also deteriorates flowability. 
 
The free fat is sometimes defined as the fraction of fat, which is not protected by a protein film and is present in form of fat pools or patches rather than globules predominantly on the surface of fat containing milk powders. An excellent study of the appearance of free fat content has been presented by Buma [24, 25]. Based on his findings the free fat content can be defined as that fraction of fat, which is extractable by organic solvents under defined conditions (solvent type, time and temperature of extraction). He proved as well, that at least a part of free fat is located inside the powder particles and that the occurrence of free fat in whole milk powder can be related to the presence of micropores and cracks. There is a relationship between moisture and free fat content. With increasing moisture from 2 to 4-5% the free fat decreases, but with moisture content above 6-7%, the fat is 100% extractable. The former phenomenon can be ascribed to the swelling of particles and thus closing of the
microspores thus retarding the penetration of the solvent into the particle interior. The latter is attributed to the crystallization of lactose, proven by microscopic observation in polarized light, which provokes the creation of a network of fine interstices and cracks along the side and edges of the lactose crystals, which makes the particle permeable to gases and solvents. The changes of free fat content taking place during moisture absorption, resulting in swelling of particle mass and closing the microspores are fully reversible as long as the critical moisture level of 6-7% is not exceeded.
 
The continuous phase of milk powder particles is formed by amorphous lactose together with other milk serum constituents, in which fat globules and casein micelles are dispersed and which is impenetrable by organic solvents. During homogenization some casein particles adhere to fat globules covering them partly. The physical structure of a particle, mainly the particle size, distribution of fat and porosity plays a dominating role. High occluded air content results in high free fat and therefore powders made from concentrates with low solids content or foam spray dried have higher free fat contents.
 
In spite of free fat content not being quite proportional to surface area, small particles have much more free fat than large. Buma’s conclusion was that small particles have more and wider micropores. However, the small and large particles in his study were cyclone fractions and chamber fractions of the same powders. Thus a more probable explanation is that the cyclone fraction had been exposed to considerable friction during pneumatic transport and especially in the cyclone, attacking the surface and liberating the fat from fat globules located close to the particle surface.
 
Buma also explained why the amount of extractable fat is time dependent, whereby the major part of free fat is extracted during the first 10 minutes and the residual portion needs many hours. The difference between short and long extraction remains almost constant with increasing moisture up to a critical level. While long time extraction exhibits a critical moisture level between 6-7%, the short time extraction is reaching that point around 9-10%. Based on these findings, Buma presented a model of a whole milk particle with four forms of extractable fat. A free interpretation of Buma’s work dealing with the extraction of whole milk powder with 29% fat and various moisture content is presented in Fig. 10.15 and his model of a milk powder particle in Fig. 10.14.
Buma’s model of four forms of extractable fat in whole milk powder
Fig.10.14. Buma’s model of four forms of extractable fat in whole milk powder
These four forms of extractable fat are:
 
a) surface fat, present as pools or patches on the particle surface, particularly in the irregular folds or at contact points between particles.
 
b) outer layer fat, formed by fat globules almost touching the particle surface and thus easily accessible to solvents.
 
c) capillary fat, consisting of fat globules almost touching the surface of microspores and cracks and thus, similarly to b), can be reached by solvent.
d) dissolution fat, consisting of fat globules almost touching any hole remaining after dissolving the outer layer and capillary fat globules inclusive dissolution fat globules, forming a chain of fat globules.
Influence of moisture content on free fat (acc. To Buma)
Fig.10.15. Influence of moisture content on free fat (acc. To Buma)
In good quality whole milk powder with 28% fat, the free fat content is low, in the range less than 1% up to 1.5%. It is obvious that free fat is accessible not only to solvents, but also to gases and atmospheric oxygen developing oxidized flavour and tallowiness. Wewala [26] recommended an increase of moisture content of commercial whole milk powder from the usual 2.5-3.0% to 3.4%. More about the background of this recommendation is given in section 10.10., but a strongly contributing effect of such a step for extended shelf life is definitely the decrease of free fat content as a function of moisture described above. Moreover, according to Buma’s findings, the free fat decrease between 2 and 7% moisture is close to linear. Nevertheless, it exhibits a very significant drop, followed again by an increase, around 3.5% moisture. This phenomenon may be caused by onset of lactose crystallisation, which eventually may lead to severe product quality deterioration during storage.
The work of Snoeren [4 - 8] on viscosity of milk concentrates has already been mentioned. It can be seen in Fig. 9.5 that the viscosity of whole milk concentrate increases with both increasing concentration and increasing homogenization pressure. No doubt the homogenization effect on the size of fat globules has a dominating effect on decreasing free fat content. However, Snoeren proved that it is also a function of viscosity (which controls the particle size distribution and thereby many other properties). In other words low free fat content can also be achieved by a combination of high concentration with low homogenization pressure instead of low concentration with excessive homogenization, without any adverse effects on other properties. This finding can probably be utilized in practice to achieve more economical conditions. However, it will require a deep knowledge of the influence of various factors on the viscosity and standardization of protein content.
 
The factors controlling the level of free fat are:
a) total fat content of the powder. Below approximately 26% fat, the free fat fraction is low but above this level it increases rapidly.
b) type of fat, i.e. vegetable fats of low melting point tend to increase the free fat level.
c) product composition, i.e. if the composition of non-fat-solids is dominated by carbohydrates, especially lactose, the free fat level is low; with dominating proteins it is higher.
d) physical state of lactose, i.e. amorphous lactose protects the fat against extraction while crystallized lactose provokes free fat,
e) gentle drying conditions result in particles with smooth surface, giving lower free fat than with high temperature drying, which creates cracks and microspores,
f) gentle powder treatment, i.e. avoid pneumatic transport, use a dryer type giving a low cyclone fraction, operate with a low pressure drop over the cyclones and cool the powder in a fluid bed, avoid too high powder moisture from the first drying stage.
g) avoid standardization by buttermilk of fat of whole milk.
h) use high feed concentrations up to the level which will still ensure good solubility index.
i) avoid using lecithin as emulsifier for fat filled powders.
j) avoid over drying of the product to too low final moisture content.
k) use two stage homogenization with medium pressures.
l) using crystallized lactose as ingredient dissolves it completely.
10.7.

Instant properties

Reconstitution of a milk powder is a complicated process and many analytical methods have been developed to determine how successfully it has been completed and what the defects of the reconstituted solution are. The most important instant properties are wettability and Dispersibility. Originally sinkability was considered as a part of the reconstitution process also expressing how fast particles sink to the bottom of the glass, but later on it was found that this was of secondary importance and difficult to measure. Sometimes the expression ‘sinkability’ is used as a synonym of wettability. 
 
Ideally, the reconstitution of milk powder in water should result in a homogeneous solution and suspension having the appearance of pasteurized milk, but in practice there is almost always some un-dissolved or un-dispersed residue. This may consist either of inside un-wetted lumps or slurry at the bottom of a glass, agglomerates, single particles or fine flakes floating in the reconstituted solution or tiny flakes and possibly some particles or small lumps floating on the surface, so-called floaters. Apart from wettability all the other tests try to detect this insoluble residue. The biggest insoluble elements are determined by the Sludge and the Dispersibility tests. The flakes floating in the solution after removing the Sludge by filtration through a 600 μm mesh are detected visually in the milk film, remaining on the walls when emptying the test tube as slowly dispersible particles abbreviated to SDP. Even smaller flakes, designated as white flecks, are expressed by a White Fleck Number.
 
Instant milk powders were originally developed to be used as a fresh milk equivalent, i.e. consumed as cold drinks and therefore they should preferably be cold water instant. However, consumer demands have shown that they must also be reconstitutable in hot water or even  hot coffee or tea. In countries with a lack of good quality drinking water, this has to be boiled before use. This water is often used for reconstitution of the powder just after boiling. Therefore good instant powder must exhibit an absence of Sludge and slowly dispersible particles also under these conditions. Thus, we are talking about cold and hot Sludge and cold and hot SDP. Moreover, there must be good thermo-stability in hot beverages which is expressed by the Coffee Test and Hot Water test, also called hot sediment. For cold tests the temperature is 25°C while for hot tests 80 - 85°C. For agglomerated but non-lecithinated powder 45°C has been accepted for cold tests.
 
When the hot milk is allowed to stand for 15 minutes a skin is formed on the surface. Its thickness and colour is evaluated visually and expressed as skin index.
 
Generally, the essential conditions for good instant behaviour are good agglomeration and wettable surfaces. Milk powders with fat content less than 1% have a wettable surface. Fat containing powders, as instant whole milk powder, have due to the free fat a hydrophobic surface. Lecithin, being both hydrophilic and lipophilic and, being also a natural component of milk and therefore acceptable as an additive, is most suitable to be used for preparing the wetting agent, which is a solution of lecithin in oil. This is sprayed on the final powder when still warm and exposed to violent fluidization to ensure as complete distribution on the total surface of the agglomerates and particles as possible. Any fat with low melting point can be used as the solvent for the lecithin. Originally pure butter oil was used for ethical reasons to avoid any ‘foreign’ fat. For functional reasons, however, low melting vegetable oils or fractionated butter oil with melting point well below 18°C is superior. Practice has shown that there are no objections to using a vegetable oil for this purpose, and it is now legal in most countries.
10.7.1.

Wettability

The Wettability or wetting time determines the time necessary for a given amount of powder, dropped onto still water, to pass through the surface. Factors of importance for good Wettability of whole milk and other fat-containing powders are the following:
 
a) good agglomeration with an absolute minimum mean particle size of 180 μm (preferably 200 - 300 μm), having a size fraction below 125 μm less than 20% (but preferably lower than 15 or even 10%) and size fraction greater than 500 μm not higher than 10%.
b) efficient lecithin treatment - the details are explained separately below.
c) particle density at least 1.15 g/cm3 but preferably above 1.2 g/cm3.
d) conditioning of the powder after lecithin treatment to get a final temperature about 45°C.
e) gentle transport to silos or hoppers.
f) packaging in cans or transport boxes before the temperature drops below 40°C or alternatively after it has dropped below 25°C.
 
The philosophy and theoretical backgrounds behind lecithin treatment are:
 
The occurrence of free fat in fat-containing powders was discussed in previous section. As shown in Fig. 10.14, some part of this free fat appears on the surface. The fat is hydrophobic, i.e. water-repellent, which is the reason why such powders when dropped on the surface of cold water remains on the surface almost un-wetted. Lecithin is a component which is both hydrophilic and lipophilic and is fat-soluble. Thus incorporation of lecithin into the free fat is able to convert the hydrophobic surface to hydrophilic one. In whole milk powder the milk fat is a mixture of a number of triglycerides with melting points between 0 and 45°C. The powder should be instant at room temperature, i.e. around 20°C. At this temperature, part of the fat is in a solid state while a part is still liquid. These are the high melting and low melting fractions. The latter constitutes about 30-50% in a normal milk fat.
 
For achieving good Wettability by means of lecithin treatment, the active part is the low melting fraction, which remains liquid after the powder has been cooled down to the surrounding temperature. The high melting fraction will crystallize. The solution of lecithin in oil is sprayed on the treated powder and it is mixed with the original free fat. It is an advantage if the fat used as a solvent is a low melting fat, i.e. oil, which is liquid at normal temperature. However, using milk fat (in form of butter oil) is possible as well. The disadvantage of using butter oil is that the solution is more viscous and besides, it requires higher addition (because together with the active low melting component the high melting fraction is added). This so-called wetting agent has to be sprayed onto the powder in such a way as to ensure both good distribution and, during subsequent treatment (fluidization with hot air), good mixing with the original free fat to finally create a homogeneous solution of lecithin in melted free fat together with added oil. During cooling, the high melting fractions gradually crystallize. They are transformed to solid crystals, whereby the lecithin concentration in the low melting fractions increases. At the very end of this process (which in fact should terminate in cans or transport bins) the agglomerates and particles are covered with crystals of the high melting fat fraction and the whole surface is coated by the remaining low melting fraction with dissolved lecithin.
 
If the powder is cooled rapidly and exposed simultaneously to mechanical treatment, as for instance fluidization in a fluid bed cooler, the whole solution will become very viscous. Later on the high melting fraction will also very slowly crystallize. However, the crystals will be very small and spread throughout the whole coating layer, appearing partly also on the surface.
 
These crystals are hydrophobic and such powder will not be wettable. Keeping in mind the above theoretical considerations, the condition for achieving good wettable powders are:
 
a) lecithin concentration in the liquid fraction of the free fat must be 15-25% and the absolute amount of lecithin on powder 0.12-0.22%.
b) wetting agent consisting of 30-50% lecithin in oil solvent.
c) the total amount of free fat inclusive lecithin should be between 0.8-1.8% whereby the amount of low melting fraction should be 0.5-1.2% to ensure that the whole surface of the agglomerates and particles will be coated.
d) the oil used as solvent for lecithin should be without any foreign taste and odour and have a melting point preferably below 12°C. If butter oil is used, it must match the requirements for pure butter oil.
 
From the above it is obvious that the amount of low melting fraction and therefore also of the lecithin depends on the specific surface area of the powder, and that good agglomerated powders of large mean particle size require less lecithin than poorly agglomerated powders. Furthermore, using low melting point oil results in lower final free fat content than with butter oil. The low limits of the above ranges refer to powders having low specific surface area. A good Wettability can also be achieved when the amount of lecithin and free fat are higher than the above levels. However, high levels will adversely affect the other instant properties. Proper composition of the wetting agent, the kind of oil used, and dosing are in practice found empirically.
 
For the lecithination process as such, the requirements are as follows:
a) the wetting agent temperature should be 60-65°C to secure a reasonably low viscosity and good atomization,
b) the powder temperature must be min. 50°C,
c) after lecithination the powder should be fluidized at min. 45°C for at least 5 minutes.
10.7.2.

Dispersibility

Dispersibility determines how completely the product has been dissolved without leaving any un-dispersed residue, which can be detected visually as Sludge or lumps and separated from the solution by means of sifting. Sifting of the reconstituted solution and estimation of the amount of non-reconstituted residue retained on the sifter mesh are the operations used for the determination of the Dispersibility and Sludge. These two methods are using different sifter mesh sizes: 150 μm for IDF-Dispersibility, 300 μm for NZ-Dispersibility and 600 μm for Sludge. The part of retained residue is also determined by the Sludge method. Thus the Sludge test is in a way also a dispersibility test. Both the IDF Dispersibility and Sludge express quantitatively the un-dispersed residue, the NZ-Dispersibility just by comparing with the standard photo-scale.
 
Poor dispersibility is due to all kinds of un-dispersed material, which cannot pass the mesh. These are:
 
1. Large and compact agglomerates (raspberry or compact grape type), for which the conditions of the analytical methods are insufficient - as to the time and turbulence available - to complete the dispersion. Such agglomerates are created when droplet/droplet or droplet/particle collisions take place at high moisture content, i.e. close to the atomizing device. Agglomeration of poorly dispersed fines with wet droplets can also create compact agglomerates.
 
2. Large lumps and slurry settling at the bottom of the glass as a result of poor agglomeration. The reason for this is too small mean particle size and an excessive amount of fines smaller than 125 μm. A too high level or poor distribution of lecithin can result in creation of a slurry. 
 
The experiences with various types of spray dryer design as to dispersibility have shown that the Multi-Stage (MSD™) concept is in this respect superior to any other design, obtaining almost constantly the highest classifications. This is most probably due to the large mean particle size and low amount of fines together with spontaneous primary and secondary agglomeration taking place in this type of dryer.
10.7.3.

Sludge

The cold Sludge is determined at 25°C for instant powder and at 45°C for agglomerated non-lecithinated powder. 85°C is used for hot Sludge. These are currently called Sludge 25, Sludge 45 and Sludge 85. As commented above, the Sludge test is also a kind of Dispersibility test, determining a part of un-dispersed residue. The standard for Sludge is max. 0.1 g. As the weighed wet Sludge has about 50% moisture, expressing this amount in terms of IDFDispersibility would correspond to 99.6%. In practice, however, the Sludge of a good quality powder is much lower than 0.1 g, and therefore the contribution of Sludge to Dispersibility is minimal.
 
Nevertheless the cold Sludge is a useful and fast routine test especially for drying systems where achievement of reasonably high mean particle size and low amount of fines can be a problem. On the other hand, there can be expressed doubts about the usefulness of hot Sludge. It is seldom too different from cold Sludge.
10.7.4.

Heat stability

Coagulation of proteins leading to precipitation is the reason for many powder faults. The constituents of milk solids, which are the direct cause of precipitation, are mainly casein and -lactalbumin. The size of precipitated matter can vary widely, ranging from microscopic to macroscopic. The precipitation can take place in any processing step, during which the milk, the concentrate during the evaporation process, the droplet or wet powder are exposed to heat. The heat exposure can be indirect, in which the heat from the heating medium (steam or hot water) is transferred to the milk solids via a heating surface. In direct heating, the heating medium (steam) can be either injected into the heated liquid (Direct Steam Injection or DSI) or the heated liquid injected into the heating medium (Steam Infusion). In any case, there is a surface layer of heated material, which is primarily exposed to the first heat chock. Besides, precipitation can take place also during the analytical tests, in which the powder is exposed to the heat of the water or coffee used for dissolving.
 
Whether precipitation will occur under the various conditions of processing depends on following factors:
 
1. The temperature difference (T) between the heating medium and heated milk solids.
 
2. How fast, if at all, the material of the surface layer is replaced by new material. In case of liquid heating this depends on the product viscosity and on the flow velocity or turbulence or on the effectiveness of agitators. On the other hand during spray drying, where the surface layer of the droplet remains constant, the important factor is the air velocity. Direct heating by means of steam injection is usually considered gentler. However, even in this case a surface layer of the heated liquid on steam bubbles can occur, if the T is too high and mixing is unsatisfactory due to low flow velocities.
 
3. Heat stability of the milk solids can be defined as a relative resistance to precipitation during processing or testing. Generally speaking the heat stability of milk decreases during the evaporation process. Moreover, the danger of precipitation increases due to rising viscosity, influencing the flow characteristics. An important contributing factor in the evaporator is insufficient coverage of the tubes when liquid distribution is unequal. All properties based on dissolving powder in hot water or hot beverages are in a way expressing heat stability of the product. Good heat stability is thus a common precondition for achieving good SDP 85, Hot Water test, Coffee Test and also Sludge 85, although the latter is much less sensitive. 
 
The factors affecting the heat stability of milk powders can be divided into three groups:
 
1. Factors defining the basic properties of the processed milk such as chemical composition, including properties that express freshness.
2. Factors induced by the whole manufacturing process.
3. Factors given by the physical structure of the powder.
4. Factors given by the conditions of the test method.
 
The basic properties of the milk involve mainly the acidity, total protein concentration, α-lactalbumin/casein proportion and salt balance. The influence of acidity is shown in the next section on SDP 85. High protein concentration increases the concentrate viscosity and thereby the size of primary particles, i.e. affecting the physical structure of the powder.
 
The manufacturer is obviously mostly interested in the performance of the final powder in which the effect of precipitation, which took place during processing or during the analysis,
will be as little as possible. 
 
This means that the amount of the precipitate will be below the level specified for the individual tests or that the size of the precipitated elements will be undetectable by the method in question. 
 
The proteins are the constituent subjected to precipitation under certain circumstances and therefore a reduction of total protein content is a useful step to reduce the susceptibility of the milk to precipitation. However, in many countries, the protein content is much higher than the average of that from Frisian cows. A reduction to 38-39% total protein of non-fat solids by the addition of lactose is definitely one of the most useful steps for the control of hot properties. Such content is also still well above the Frisian milk average. This standardization of the protein content is sometimes called frisianisation. Since 1999 the FAO Codex 207-1999 has allowed for protein standardisation down to 34% protein of non-fat solids using either milk permeate or lactose.
 
The main protein components are casein and α-lactalbumin. The amount of α-lactalbumin can vary over a wide range and can be very high especially at the beginning of the season. Normally the milk powder manufacturer has no information about the actual content of α-lactalbumin. During heat treatment, it always coagulates creating, depending on the conditions of the heat treatment, coarse or fine precipitates. Pasteurization prior to evaporation is a necessary step not only for bacteriological reasons but also to ensure the required shelf life. Acceptable pasteurization conditions for both heat stability and shelf life control should result in WPNI 2.5-3.5. However, lower temperatures with long holding time have preference for the control of the heat stability. A useful step, which has been experienced in the production of UHT-milk for years, is holding the processed milk at 80°C for 6 minutes before the main pasteurization. This step is called stabilization and leads to an extremely fine precipitation not detectable by the naked eye.
 
The homogenization of the concentrate, used for the control of free fat, progressively decreases the heat stability with rising pressure. Therefore high homogenizing pressures, single-stage homogenization and recirculation of the concentrate over the homogenizer (when using constant speed homogenizer as a feed pump for the dryer) must be avoided. Experience has shown, however, that it is seldom necessary to homogenize with total pressure higher than 80 bar with 20 bars in the second stage.
 
One of the most important factors influencing the susceptibility to precipitation is the salt balance. In connection with the sterilization of evaporated milk, it has been found that casein has maximum heat stability when combined with an optimum amount of calcium. When the calcium content is higher or lower than this optimum, the casein/calcium complex is less stable. The calcium in normal milk is distributed between casein, phosphates and citrates. The optimum salt balance is achieved when the calcium (and also magnesium) cations, which are not bound in the casein complex, are in balance with citric and phosphoric acid anions.
 
Obviously, the optimum balance can be achieved by addition of the component, which is in deficiency or removing the surplus component. Normal milk usually has an excess of calcium, thus the increase of heat stability can be achieved by the addition of the acid anions in form of sodium phosphates or citrates. Citrates are more expensive than phosphates but are more effective. The tri-sodium citrate is much better soluble in water than the corresponding phosphate. Moreover the presence of the natural calcium phosphate content in milk is one of the reasons for precipitation and thus any increase should be avoided.
 
The addition of citrates for improving the heat stability is permitted, because citrates are a natural component of milk. Furthermore they are listed among the additives approved by WHO and FAO for baby food. Nevertheless, in some countries they can still be considered undesirable, probably because they have to be declared as additives, and this can spoil the product image.
 
The heating of the concentrate prior to atomization improves the so-called hot properties (tests where powder is reconstituted in hot liquid).
 
One effect of this heating is reduction of the viscosity. Another effect may be that due to the decrease of tri-calcium phosphate solubility with rising temperature, this treatment removes
the excess of calcium ions from the solution. 
 
The factors for increasing the heat stability and thereby improving all properties at higher reconstitution temperatures can be summarized as follows:
 
1. Standardization of total protein content below 39% (in non-fat-solids) or lower preferably by lactose. However, if problems should occur, it is recommendable to run a comparable production with lactose only.
 
2. The addition of ascorbic acid as an antioxidant, if required, should be done in cold milk giving good dilution and by adding where milk flow is turbulent. The higher the level of Vitamin C added the more careful the dosing. Addition in the form of a neutral salt of ascorbic acid or ascorbyl-palmitate eliminates dosing problems.
 
3. Addition of citrates or polyphosphates for milk in the off-peak milk season (where milk is unstable) is recommendable to achieve the desired heat stability.
 
4. Pasteurization resulting in WPNI 2.5-3.5. Low temperatures with long holding time are to be preferred. Pre-treatment with holding for 4-6 minutes at 80°C may be advantageous.
 
5. Concentrate leaving the evaporator to be free of any insoluble residue. Rapid fouling of the evaporator is usually a good indication of problems in this respect and appears especially in the early season where the milk stability is poor. It has been experienced that rapid fouling of the evaporator is accompanied by poor functionality results. Poor stability of milk is the main reason for both these problems, but undoubtedly fouling increases the functionality problems. Thus improvement of quality can be achieved by limiting any fouling through modification to evaporator components that foul up, conditions of pasteurization used, better mixing between milk and steam in case of direct steam injection (DSI), milk flow distribution and coverage of the tubes (possibly by addition of water).
 
6. Heating of the concentrate to min. 75° (preferably to 80°C).
 
7. Avoid excessive homogenization. Usually two stage homogenization is sufficient with 80 bar total pressure drop, 20 bar in the second stage.
10.7.5.

Slowly dispersible particles

Slowly dispersible particles, called SDP 25 and SDP 85 for cold and hot SDP, respectively, are the properties creating most problems. They are based both on powder structure as well as on the chemical nature of the product. There is often a significant difference between hot and cold SDP, the latter being most problematic.
 
As the name indicates, this property expresses also a kind of dispersibility evaluation. While the standard NZ-Dispersibility test retains all elements greater than 300 μm on the screen, SDP test is conducted with reconstituted milk after the Sludge test, i.e. from which all elements greater than 600 μm have been separated by sifting. Experience has also shown that there is no connection between Dispersibility and cold SDP. A powder can exhibit excellent or poor Dispersibility combined with either excellent or poor cold SDP. The appearance of the white spots observed in the film of milk on the wall of the test tube in this test does not indicate the presence of particles or agglomerates. Moreover, the fact that it remains at all in this film indicates that these are rather flakes than particles.
 
A comprehensive investigation conducted over 8 months at a large spray dryer of the MSD™ design has given the following findings:
Influence of titratable acidity on Hot Water Test and SDP 85
Fig.10.16. Influence of titratable acidity on Hot Water Test and SDP 85
1. No correlation between SDP 25 and any other property or operating parameter was found.
 
2. Strong relation between SDP 85 and many other properties and also to some operating
parameters was found. The correlation of SDP 85 to Hot Water sediment and titratable acidity isshown in Fig. 10.16. Table 10.1 shows the relation between SDP 85 and WPNI, titratable acidity, Hot Water test and Coffee Test. The operating parameters, which were of influence, were the fluidizing velocity in static fluid bed, the outlet drying air temperature, and atomization pressure.

Property or condition

Unit

Min.

Max.

Correl.

SDP 85 SDP 85 coef

WPNI

SDP 85

mg N/g

A=0

2.43

0.00

2.00

3.04


0.82

Titratable acidity

SDP 85

ml l.a.

A=0

0.10

0.51

0.12

3.92


0.75

Hot water sediment

SDP 85

ml

A=0

0.20

1.23

3.95

4.13


0.57

Coffee test

SDP 85

ml

A=0

0.39

0.84

1.26

3.70


0.36

SFB-Fluidizing vel.

SDP 85

m/s

A=0

0.97

0.00

0.88

3.23


0.73

Outlet temperature

SDP 85

°C

A=0

67.20

0.00

63.60

3.18


0.60

Atomization pressure

SDP 85

kPa

A=0

28.10

0.56

21.30

3.01


0.51

Table. 10.1. The Influence of various parameters on SDP 85
The other average properties of approximately 400 samples were:
 
- Sludge 25 and Sludge 85 was 0.02 and 0.03 with no single results higher than 0.05,
 
- Dispersibility was within the range 1-2 with average 1.22,
 
- SDP 25 was A-D with average 1.54, i.e. between B and C.
 
It can be concluded that there is no correlation between both SDPs and neither Dispersibility nor other properties, indicating a kind of dispersibility, like Sludge 25 and 85. However, there is a strong correlation between SDP 85 and all other hot properties, i.e. Hot Water sediment and Coffee Test. All depend mostly on WPNI and acidity. As to the latter, it is necessary to comment that none of the samples indicated really sour milk as all measured acidities were in the sweet range.
 
Another example is from a large spray dryer of CDI design with fines return system of FRADdesign where, apart from Dispersibility, very good results were achieved as shown in table 10.2. However, in contrast to the previous example, this powder had a very poor Dispersibility between 4 and 6. In the previous example the constantly excellent Dispersibility and Sludge values were achieved together with SDP 85 varying between A - E and SDP 25 A - D. In this case excellent Sludge, very good SDP 25 and reasonable SDP 25 appeared together with poor Dispersibility. As can be seen, the oversize fraction was quite high, i.e. 17.70% which was probably the reason for that Dispersibility, however without any significant influence on SDPs. One of the reasons for much better SDP 85 in this CDI-powder in comparison with previous MSDTM-powder is the WPNI, which is on the high side of the recommended range, while in previous example it was even outside the low side.
Property Result Property Result
Titr. acidity 0.11 Hot water test 0.30
Sludge 25 0.04 Coffee test 1.06
SDP 25 1.43 Free fat 0.72
Dispersibility 5.00 WPNI 3.37
Sludge 85 0.04 Fraction >500µm 17.70
SDP 85 0.29 Fraction <125µm 9.95
Table. 10.2. Powder quality results from CDI-plant with FRAD
From these findings it can be concluded that SDP 85 is controlled by the same factors as the other hot properties, i.e. Hot Water sediment and Coffee Test. The reasons for poor hot properties are most probably oversize primary particles due to high viscosity or low atomizing pressure and also particles originating from the jelly-like lumps originating in the evaporator and concentrate heaters.
 
The cold SDP is, as mentioned above, a more problematic property. It seems that it has more to do with white flecks than with dispersibility, but reliable evidence is not available. Experience indicates that feed temperatures around 72-74°C result in better cold SDP than 76-78°C. 
 
Conditions for improving heat stability as recommended in the previous section are important for control both SDP 25 and 85.
 
Guidelines are:
 
1. Effective atomization whereby the exact conditions have to be found individually for each plant. The general rules for non-protein standardised milks are however:
 
a) Nozzle atomization: - concentrate total solids 46-48%
- spray angle as close to 75° as possible
- pressure min. 220-300 bar
- frequent check of nozzle insert wear
 
b) Wheel atomization: - concentrate total solids 48-49%
- curved vane wheel
- peripheral wheel speed 135-165 m/s
- good dispersion of returned fines
 
2. Lecithin treatment: - avoid excessive amount of lecithin
- see precautions described in 10.7.1.
- avoid recycling of lecithinated fines
 
3. Ensure gentle treatment of final powder and pack when powder temperature still exceeds 40°C.
10.7.6.

Hot water test and coffee test

These two properties are closely related following the same trends, but the Hot Water test appears to be the most sensitive especially as to acidity. As described in the previous section, the main factor for both properties, apart from acidity, is WPNI which should be 2.5-3.5. In fact the higher the better, however, the upper limit is controlled by the requirements for shelf life. The relationship of both these tests to WPNI tested on a milk of generally good quality with 37-38% total proteins (on non-fat-solids) and without any additives is presented in Fig. 10.17.
Influence of whey protein nitrogen index on coffee and hot water test
Fig.10.17. Influence of whey protein nitrogen index on coffee and hot water test
The higher sensitivity, i.e. higher amount of sediment of NZ-Hot Water test in comparison with NZ-Coffee Test is caused mainly by the higher specified amount of powder used (12.5 g against 2 g). Besides, experience has shown that when poor, quite high fluctuations (sometimes almost ±100%) are exhibited. Both tests are controlled by the same factors as all other hot properties.
 
The NZ-Coffee Test is the only one of a number of coffee/ tea tests, whose results are expressed by a number.
On the other hand, it does not give the complete picture of the performance of a powder in hot coffee. The precipitation appears usually both as a sediment on the bottom and floaters on the top. The floaters escape evaluation in the NZ-Coffee Test. The CCF-Coffee Test evaluates both floaters and sediment, unfortunately only by a subjective estimate. The so-called Middle-East-Coffee Test is not at all a scientific test. The coffee and milk powder is stirred into almost boiling water in the same way as done by a busy consumer. This rough treatment creates good conditions for precipitation. The amount of sediment is evaluated visually after pouring out the liquid. Powders produced on spray dryers with rotating wheel atomizers are inferior to pressure nozzle powders with this test.
10.7.7.

White Flecks Number (WFN)

This is a fairly new test which has been elaborated by IDF and accepted as an IDF-standard in 1990. The white flecks are the minutest elements of a precipitate appearing in great numbers, however occupying a relatively small volume. When a reconstituted solution is allowed to stand in a glass for several minutes, the white flecks create a layer on the surface. When dipping a teaspoon and taking it out again, the white flecks stick to the spoon and can be seen by the naked eye. They remain also on the walls of the glass when it is emptied. The IDF-test is based on observations that these flecks can clog the holes of a fine mesh (the test is using 63μm mesh) and thus obstruct the flow.
 
It is probable that white flecks relate to SDP 25, but as this test is very new, no industrial experience is available. A study of factors promoting the appearance of white flecks has been  reported by Ruyck [27]. Unfortunately this work is of limited value, as it was conducted on agglomerated powder formed by rewetting powder from a single stage pilot plant spray dryer. None of the tested samples were truly instant. Following conclusions can be made from this work:
 
1. The White Flecks Number remained almost constant under all heat treatment conditionsresulting in WPNI 2.98 - 6.68 (WFN 0.65 - 0.75) and <0.1 - 7.22 (WFN 0.86 - 0.89) in spite of the viscosity varying between 30.5 - 93.5 and 39.2 - 440.0 mPa*s respectively. The WFN remained also constant while the insolubility index increased from 0.05 ml (at WPNI 3.35 - 7.22) to 3.50 - 7.2 (at WPNI 0.81 - <0.1).
 
2. Homogenization pressure had a negative effect on WFN. However the work was conducted under pressures of (I.stage/II.stage) 50/50 to 150/50 bar (i.e. pressures which are not suitable for instant whole milk powder).
 
3. Ruyck concluded also that increasing the inlet drying air temperature and/or outlet temperature had negative effect on white flecks, the latter being most important. However during these tests the concentrate solids varied over a broad range, thus it appears that this was a dominating effect.
Influence of outlet temperature and solids content on White Flecks Number (acc. to Ruyck)
Fig.10.18. Influence of outlet temperature and solids content on White Flecks Number (acc. to Ruyck)
The Ruyck’s results of the influence of the outlet drying air temperature on the White Flecks Number as a function of total solids content are presented in Fig. 10.18. It can be seen that there is a good relationship of both parameters to WFN. However, it can also be seen that during these trials, low outlet temperatures were combined with low solids and vice versa. Thus, it is difficult to decide, which of them is the dominating factor. Nevertheless results from other tests indicate that it is most likely the concentration.
 
There is no doubt that White Flecks Number is a useful test. However, more work is necessary to fully understand its significance.
10.8.

Hygroscopicity, sticking and caking properties

Hygroscopicity means the ability of a powder to absorb moisture from the surrounding air. All powdered milk products and their components, with the exception of fat, are hygroscopic. The relative tendency of how much moisture a given product will, under given conditions, pick up (or possibly loose) is given by the sorption isotherm of the product in question. Amorphous lactose, proteins and salts are strongly hygroscopic. However, while absorption of moisture by products containing high content of lactose is accompanied by increasing stickiness and finally caking, protein-rich products absorb moisture almost without these phenomena.
 
All three properties define the requirements on the packaging material and storage conditions of final products. However, they also set limits for the drying conditions as to inlet and outlet air temperatures and feed solids content. Probably all milk powder producers have experienced blocking of the ducts, cyclones etc. The reason for most of such incidents relate to these particular properties. Fig. 10.19 illustrates the meaning of stickiness and sticking point.
Explanation of sticking phenomena
Fig.10.19. Explanation of sticking phenomena
The sticking temperature line in Fig.10.19 defines the combination of moisture content and temperature (Tp) of a product above which it is inside the sticking zone. If, for instance, under normal air humidity, the outlet air temperature is 89°C (line To-1) then the product temperature (line Tp-1) is just below the sticking temperature line, i.e. no sticking is taking place. The corresponding 2.5% moisture is, in case 1, the upper limit for drying of that product. It is possible, of course, to use higher outlet temperatures resulting in lower moisture levels to increase the difference between sticking temperature and powder temperature. If suddenly the ambient air humidity increases, then the relationship of moisture content to outlet air temperature and powder temperature is given by lines To-2 and Tp-2. Continuation of the process with 90°C outlet temperature (case 2) will result in about 3.1% moisture and the corresponding powder temperature will be well inside the sticking zone. The solution is to increase the outlet air temperature to about 96°C (case 3) whereby the corresponding moisture content will drop below 2% and the product temperature will come just out of the sticking zone. This graph is illustrative and the figures must not be considered as absolute. However they are close to conditions for drying of non-pre-crystallized whey.
One of the components contributing mostly to hygroscopicity and stickiness is amorphous lactose, and therefore the mentioned problems occur when drying whey and whey products. The acidity due to presence of lactic acid is strongly contributing to stickiness as well. These are the reasons why acid whey is one of the most difficult products to dry. One of the possibilities to reduce the hygroscopicity and thereby stickiness of whey powders is precrystallization of lactose. Pallansch [28] in his work on drying acid whey developed a method for determination of the sticking temperature and using this method presented the influence of lactose crystallization and presence of lactic acid on the temperature of sticking. These relationships are given on Fig. 10.20 and 10.21.
Influence of pre-crystallisation on sticking temperature
Fig.10.20. Influence of pre-crystallisation on sticking temperature
As already mentioned, the hygroscopicity is controlled by water activity and this again depends on product composition. Hygroscopicity cannot be reduced without changing the composition or without changing the physical state of some components. 
 
The possibilities for reducing the hygroscopicity and disposition to sticking or caking are as follows:
 
1. Transformation of amorphous lactose to -lactose-mono-hydrate by precrystallization. The effect is shown in Fig. 10.20.
2. Denaturing of whey proteins by pasteurization. For acid whey, this pre-treatment was recommended as a must. However such treatment also increases the viscosity significantly, thus setting limits for maximum manageable concentration.
 
3. Neutralization of lactic acid by means of calcium hydroxide (lime). Use of sodium or potassium hydroxide is not advisable as the created lactic acid salts contribute to stickiness practically to the same extent as lactic acid.
 
4. If maltodextrin is a component, a maltodextrin in the lower dextrose equivalent range is recommended.
 
5. In order to delay the absorption of moisture and avoid caking, lecithin treatment can be applied in the same way as for instant whole milk.
 
6. Addition of a free flowing agent such as aluminium-sodium-silicate reduces the stickiness and tendency to caking.
Influence of added lactic acid on sticking temperature
Fig.10.21. Influence of added lactic acid on sticking temperature
As far as whey powders and whey based products are concerned, the most usual and effective way is pre-crystallization of lactose. The degree of crystallization expressed as percent of crystalline lactose (of total) for a whey concentrate with 72% lactose in total solids, based on initial and final refracto metric reading and final temperature, can be found using the graphs in Fig. 10.22. In the example in this graph, point 1 indicates that a concentrate exhibiting an initial refractometric reading 53°Bx and the second one at 20°C 36°Bx has achieved about 76% crystallization (point 2) and the total solids content before crystallization has been 52%.
The glass transition theory approach to better understanding the sticking mechanisms was discussed in 9.4.7.
Relation between initial and final refractometrical readings, total solids and degree of crystallization
Fig.10.22. Relation between initial and final refractometrical readings, total solids and degree of crystallization
10.9.

Whey Protein Nitrogen Index (WPNI)

The whey protein nitrogen index, WPNI, expresses the amount of un-denatured whey protein. It is a measure of the sum of heat treatments to which the milk has been subjected. The heat treatment of a concentrate has only a negligible effect on WPNI and thus the main operation to adjust the required value is the pasteurization processes, possibly at the milk reception and in the evaporator, i.e. time/temperature combinations. However, there are many other factors influencing the WPNI including the total amount of whey protein and the overall composition of processed milk as influenced by animal breed, seasonal variations, and possibly protein standardization. The individual design of the processing equipment, i.e. the pasteurizer and holding cells, has also great significance. Therefore it is difficult to predict the conditions of achieving the required WPNI on a general basis, or to transfer the experiences from one plant to another or even from one season to the next.
 
Obviously the primary purpose of heat treatment is to ensure that the product is free of quality jeopardizing bacteria, yeasts, moulds and milk enzymes. In milk powder production, the influence of heat treatment on denaturation of whey proteins for achieving the desired properties of the final product is just as important. Skim milk powder for cheese manufacture should have as much un-denatured proteins as possible, i.e. it should be low-heat (WPNI > 6), while for bakeries, high-heat powder with high denaturation is required (WPNI < 1.5). Denatured -lactalbumin can bind moisture about seven times its own weight and this is important for the structure of the dough and bread volume.
 
The importance of heat treatment resulting in WPNI 2.5 - 3.5 for instant whole milk powder, i.e. in the middle of medium-heat (WPNI >1.5 - <6.0) range, is explained above. 
 
Generally, direct contact heat treatment requires about 5 - 6°C higher temperature than indirect heating in order to get the same WPNI. The time/temperature relationship for achieving the desired WPNI is presented in Fig. 10.23. However, from reasons given above, this relationship shown has to be understood as a guideline only. In order to bring the WPNI under control every powder plant should collect the results from each plant at least twice a week.
Guideline for time/temperature relation of WPNI
Fig.10.23. Guideline for time/temperature relation of WPNI
10.10.

Shelf life

The shelf life is an important property especially of products for the retail market. Deterioration of product quality can be caused by moisture uptake with consequential chemical, physical and bacteriological changes and by oxidation of fat. As milk powders are hygroscopic, the only way to reduce the hygroscopicity is crystallization of lactose. However, such a step is applicable only for products where crystallization will not affect the functional or other desirable properties, and where there is a reasonably high content of lactose. In practice it is used only for whey powders. Therefore vapour-tight packaging is necessary to give the product the required shelf life.
 
For fat containing products and mainly whole milk powder for the retail market, there is the danger of deterioration by fat oxidation. Free sulphhydryl-groups (SH-groups) which are created in milk by heat treatment have anti-oxidative properties and thus a positive influence on prolonging shelf life. The higher the pasteurization temperature the faster is the creation of SH-groups. Kirchmeier et al. [29] studied the rate of creation of SH-groups in milk at various 
temperatures.
 
According to his results, the formation of free SH-groups commences at 72°C, reaching the maximum at 95°C and then declines at higher temperatures. This phenomenon is attributed to masking due to creation of a complex with casein. The results of Kirchmeier work are presented in Fig. 10.24 and 10.25.
Creation of SH-groups at various temperatures
Fig.10.24. Creation of SH-groups at various temperatures
Creation of SH-groups in milk at various temperatures and times
Fig.10.25. Creation of SH-groups in milk at various temperatures and times
According to Labuza’s [30] theory on stability of foods as a function of water activity, the reaction rate of lipid oxidation reaches the minimum at aw=0.24 (see Fig.10.26). This has inspired Wewala [26] to recommend to increase the moisture content of commercial whole milk powder from usual 2.5-3.0% to 3.4%, corresponding to water activity 0.24. This was confirmed by test work, which showed achievement of an extended shelf life of the powder.
Relative reaction rates of major deteriorative reactions in food (acc. to Labuza)
Fig.10.26. Relative reaction rates of major deteriorative reactions in food (acc. to Labuza)
Previously there have been tendencies in whole milk powder manufacture to keep the moisture content at a value sometimes as low as 2.3%. Thus the background of Wewala’s recommendation is theoretically correct and as mentioned in section 10.6. The strongly contributing effect of this step for extended shelf life is definitely the decrease of free fat content.
 
Nevertheless, a critical analysis of the consequences of this recommendation cannot approve of such practice, based upon practical experiences regarding actual shelf life of whole milk powder with such high moisture content. For example much instant whole milk powder is marketed in areas having very high ambient temperatures, even at above 50°C. The storage rooms are not always conditioned to temperatures recommended for milk powder storage. On the other hand, supermarkets in these areas are mostly well air-conditioned. The set of graphs in Fig. 10.27 shows the conditions of interstitial air in a tin of whole milk powder with moisture content 2.0 - 4.0% at temperatures 20 - 50°C. 
The conditions of interstitial air in powder cans under various temperature conditions
Fig.10.27. The conditions of interstitial air in powder cans under various temperature conditions
The figures are calculated for normal atmospheric air. The humidities for nitrogen would be slightly higher, but for carbon dioxide slightly lower. The usual mixture of N2 with CO2 will give similar values as for air. The dew point will be the same for all types of gas. It can be seen that the dew point of the gas in a can of whole milk powder with 3.4% moisture, heated to 50°C, is about 25°C. Therefore condensation can take place when this can is transferred from the
storage room to an air-conditioned room. 
 
Thomsen et al. [57] studied the phase transition of pure amorphous lactose and showed instability, i.e. lactose crystallisation, at aw ~ 0.4 at 25°C and aw ~ 0.25 at 38°C. 
 
Thomsen et al. [58] also studied the temperature effect on lactose crystallisation, Maillard reactions, and lipid oxidation in whole milk powder with an initial aw of 0.23. Powder samples were stored on closed vials at 37, 45, and 55°C. At 37°C only minor changes occurred during the 147 days of observation. At 45 and 55°C quite dramatic deteriorative processes took place, at 55°C after just about 6 days and at 45°C after about 2 months. It is suggested that lactose crystallisation, being initiated at storage temperatures above the glass transition temperature, Tg, is the main cause of these quite dramatic and deteriorative changes in important quality parameters. The aw and hence the final moisture content should therefore be selected, depending on the expected thermal stress the milk powder product may be exposed to during transportation and storage.

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.