Analytical methods

The determination of moisture is probably the most used analytical procedure in a milk powder laboratory. The conventional oven drying method is quite laborious and time consuming. Equipment, which is based on determination of loss of weight of a sample exposed to some source of heat (infrared lamp, microwave etc.) over a period of time, is now available. Nowadays the dominating method in large milk powder factories is based on infrared reflection. All these new instruments reduce labour and time, the latter to just less than 1 minute. Nevertheless, the conventional oven drying method still remains the reference method, against which all other methods have to be calibrated. Moisture content is of great importance from a commercial point of view. If no other method is agreed on between two parties then the standard oven drying method is decisive. However, one has to be aware that the ‘moisture’ determined by this method is just the ‘loss of weight’ and does not necessarily expresses the true water content. From a scientific point of view, it may be interesting to know in which form water appears in the product. In dry milk products water can exist as free moisture or water of crystallization. The content of water of crystallization is negligible in normal milk powders, but it is high in whey powders.

Determination of loss of weight of about 3 g sample, exposed to drying in an oven heated to 102 ± 2°C to constant weight. First check weight after 2 hours and then after each hour. The constant weight is achieved when the difference between two successive weightings is ≤ 0.5 mg. The weight loss is expressed in percentage. This method determines not only the free moisture but also some part of the water of crystallization. A non-standard modification of this method is the routine procedure under the same conditions, but drying in oven for 3 hours.

Determination of loss of weight of an about 3 g sample, exposed to drying in an oven heated to 87 ± 2°C for 6 hours. Fig. 11.1 shows the effect of various temperatures used for determination of ‘moisture’ for whey powder (W) and α-lactose-monohydrate (L). It can be seen clearly that the temperature 87°C is the most preferable for the determination of free moisture.

The total moisture, involving both free moisture and water of crystallization, is determined by Karl-Fisher titration. It is based on the reaction between iodine and sulphur dioxide in the presence of water. Apparatus and reagents for this method are commercially available with instruction manual for the procedure.

The water of crystallization is the difference between the total moisture (11.1.3) and free moisture (11.1.2).

Insolubility index according to the IDF Standard 129A:1988 [33] expresses the volume of insoluble material in the product under the conditions of the method. 

 

This method is primarily defined for skim and whole milk powder, but can be used also for other powders which require reconstitution without any sediment. 10 g of skim milk powder or 13 g of whole milk powder are added to 100 ml water of 24°C (50°C for drum dried product) with 3 drops of silicone antifoaming agent in the mixing jar (Cenco-mixer) and agitated at 3600 RPM for 90 seconds (Fig. 11.2.), then left for 15 minutes. The content is then gently stirred and filled into a centrifuge glass up to the 50 ml mark. The glass is centrifuged for 5 minutes at 160 G. Using the vacuum pump, the supernatant is removed to leave about 5 ml liquid above the sediment. The glass is refilled with water up to the 50 ml mark, the sediment is dispersed by means of a wire and the centrifugation is repeated. The volume of the sediment is the insolubility index.

 

In order to get good reproducibility, it is important to stick exactly to the prescribed conditions. One of the most important factors is temperature. Therefore, before the analysis, not only the water, but also the mixing jar must be tempered to 24°C. This high sensitivity to temperature can be utilized for increasing the classification ability of the method to distinguish powders with insolubility index less than 0.1 ml (which is according to the standard procedure the lowest obtainable reading). For instance when using 15°C instead of 24°C the sediment will increase about 10-times, i.e. the 0.1 ml/24°C result will be about 1 ml/15°C.

Bulk density of a powdered product is determined as the volume of 100 g after exposure to compaction by standardized tapping. The most utilized apparatus is the Engelsmann tapping machine (the German Stampfvolumeter – see Fig. 11.3.A). The IDFStandard 134-1986 [34] uses a 250 ml glass cylinder and 625 taps which is supposed to be almost tapped-to-extreme, but in practice also 0- (loose bulk density), 100- and 1250-times tapping are used. The GEA Niro-method uses the Engelsmann machine with a brass cylinder of 100 ml volume (Fig.11.3.B) with removable extension. After tapping, the extension is removed and the top of the powder levelled at the top edge. The advantage is that the bulk density is obtained by weighing directly in g/100ml. British Standard Method [35], used only seldom nowadays, applies manual tapping 10 times of a 250 ml glass cylinder on a folded towel. In Holland the Ledoux-method is still used in some factories. In this method a weight of 28.34g (1 ounce) is placed on top of 28.34g powder in the 100 ml measuring cylinder which is tapped 100 times. The results are expressed as ml/g.

The air pycnometer method is excellent for normal milk powders, which have an impermeable continuous phase formed by lactose glass. On the other hand it may give too high results for protein powders as the compressed air can penetrate into the vacuoles. Anyhow, usable results can be obtained by fast operation. An alternative method is the petroleum ether method. 25 g of powder are transferred into a 100 ml calibrated measuring cylinder (with glass stopper). By means of a 50 ml pipette petroleum ether is added and the contents are shaken until all powder is suspended. After flushing the walls with further 10 ml petroleum ether in order to bring all the particles down into the liquid, the reading of the total volume is taken. This volume minus 60 is equal to the total volume of the powder particles. The particle density, occluded air and interstitial air content for both methods are calculated as follows:

The density of the solids is calculated using the equation (3, 17) and densities of the components from the table 3.1. Both the content of occluded air and interstitial air can be expressed also in percentage of the total powder volume:

The determination of scorched particles is important not only for evaluation of product quality, but also from the point of view of safety, as this is a check of possible occurrence of some undesired combustion processes in the hot zone of the installation or heat generation in powder deposits due to Maillard reaction. The official ADMI method [36] uses 25 g of skim milk powder or 32.5 g whole milk powder mixed for 50 seconds with 250 ml water of 18 - 27°C and addition of antifoaming agent. Using compressed air or vacuum, the solution is filtered through a standard filter pad, 32 mm in diameter. The result is evaluated using the standard ADMI photographic scale. When used as a safety precaution, the frequency and quickness is most important and therefore volume measurement of amount of powder instead of weighing is recommendable.

Wettability is the essential requirement for instant products. Nowadays the IDF-Standard 87:1979 [37] is widely used where 10 g of skim milk powder or 13 g of whole milk powder is dropped into water of 25°C in a beaker of 400 ml. The weighed amount of powder is transferred into a glass cylinder placed on a glass plate over the beaker (see Fig. 11.5). The glass plate is then withdrawn and simultaneously the stop watch started. The wettability or wetting time is taken when the last particles of the powder penetrate the water surface. The Niro method uses a funnel made of antistatic plastic foil which is placed on the beaker edge with a glass pestle as a stopper above the water of 20°C. 10 g of powder is spread around the pestle. The stop watch is started simultaneously with lifting the pestle.

Dispersibility is often considered the most important characteristic to decide whether a product is instant or not on the basis of a single property. There are many methods for the determination of dispersibility. All of them are based on reconstitution of a powder under standard conditions and sifting the solution through a defined mesh. Most of the available methods evaluate the residue on the screen in comparison with a standard photographic scale.

 

The IDF Standard 87:1979 is based on the determination of the total solids of the solution and expressing the dissolved amount in percentage. 250 g of water at 25°C is weighed into a dry glass beaker. Using arrangement shown in Fig. 11.6 with glass tubing fixed by a clamp on a stand (the glass plate should remain free enough to be withdrawn) 26 g of skim milk or 34 g of whole milk powder are weighed and transferred to the glass plate inside the tubing.

The glass plate is withdrawn by gentle continuous movement and a stop watch started simultaneously. Remove the beaker and after 5 seconds insert the spatula along the wall until it touches the bottom. During the next 20 seconds stir the contents with the spatula touching continuously the bottom moving forth and back across the diameter making one complete movement per second. During the first 5 complete stirrings the spatula is slightly tilted to avoid that the upper part will touch the edge of the beaker, while during the next 15 movements it is held vertically. Simultaneously with the stirring, the beaker is slowly rotated along its axis to achieve a 360° turn at the end. Allow the contents to stand for 30 seconds (until the stop watch shows 55 seconds). Without disturbing any sediment, pour about 100 ml of the contents onto a test sieve 150 μm (diameter 200 mm) while collecting liquid in the Erlenmayer flask. Put on the stopper and mix the contents thoroughly. Carry out the determination of dry matter content of the filtrate in duplicate and use the mean value for the calculation:

The IDF-method is very laborious and time consuming. The dispersibility, being one of the most important instant properties, requires frequent checking in at least 2 hours intervals. In practice methods are preferred using similar procedures, but for the end-result a visual evaluation of the residue on the sieve is applied and compared with standard photos. The NZDB-method is such a method and uses 300 μm stainless steel screen cut into strips to fit into a sediment testing funnel. 26 g of powder is reconstituted in 200 ml of distilled water, the contents are stirred using a fork and while still swirling, the contents are poured into the sediment funnel/sieve assembly and sieved under vacuum within 5 seconds. 

 

The residue remaining on the funnel is rinsed with 100 ml of 45°C water; the sieve is removed, dried at 40°C and compared with the standard chart.

Originally the only official IDF methods for the determination of instant properties were the Wettability and Dispersibility, but later the methods for determination of White Flecks Number and Coffee Test have been accepted (see section 11.8.4 and 11.8.5.). These properties are important, but they are not sufficient to evaluate the complex instant performance. Therefore a number of methods have been developed by various producers with the aim of detecting all other possible reconstitution defects of the powder, which can be observed by the consumer. It has been emphasized several times that instant milk products and especially whole milk powder are used in many different ways, and ideally they must not exhibit any unpleasant performance that can be visually detected. It is impossible to discuss all these methods, because many instant milk powder producers have their own methods, which they consider as confidential. It is probably not surprising that an excellent collection of methods has been prepared in New Zealand. The properties determined by these methods have been discussed in section 10. Achieving product properties.

The sludge test is a kind of dispersibility determination detecting those elements of undispersed material, which cannot pass a mesh of 600 μm. This material is not so much oversized agglomerates, but rather lumps created by conglomeration of fines to form thick slurry at the bottom of the beaker.

 

The basic equipment for the determination of sludge is a 600 μm screen soldered to a brass ring of 75 mm diameter and 20 mm high. 12.5 g of powder is tipped on the surface of 100 ml water in a 250 ml beaker (water temperature see Table 11.1) and immediately stirred using a teaspoon. Stirring consists of 24 revolutions (6 clockwise, 6 anticlockwise, 6 clockwise, 6 anticlockwise) completed within 10 seconds. After standing (standing time see Table 11.1) and possibly removal of the skin by the teaspoon (see comment at Table 11.1), the content is resuspended gently with one circular and transverse movement and the content then poured onto the pre-weighed screen. The screen is then drained for 60 s on 4 layers of 2-ply tissue. After removing the liquid from the residues and the bottom and possibly walls of the screen, it is reweighed. The difference is expressed with two decimals as sludge.

The test for slowly dispersible particles (SDP) is conducted simultaneously with sludge determination using the liquid from the screen filtration. This filtrate, which is collected in a 400 ml beaker is filled into a test tube (150 x 25 mm) and immediately poured back into the beaker. After 2 minutes the appearance of the film on the wall of the test tube is compared with the SDP index standard photo. It is essential that the tubes are absolutely clean and dry. Similarly to sludge, the measurement distinguishes cold SDP which is SDP 25 or SDP 45 and hot SDP or SDP 85 depending on the same criteria as outlined for sludge.

Hot water sediment test follows in the same operation after conducting the sludge and SDP tests. The filtrate, poured back into the beaker from the SDP test tube is filled into two 50 ml centrifuge tubes (the same as for insolubility index) and centrifuged for 5 minutes at 164 G. The top liquid is then sucked off using a water jet vacuum pump down to 5 ml level. The tube is refilled with water to the 50 ml mark taking care not to disturb the sediment and then centrifuged again as before. The volume of the sediment is read on the nearest scale mark in each tube. The result is the sum of these two readings.

A very important factor for good reproducibility of the coffee test is the origin and quality of the instant coffee powder. Suitable powders will have a pH of about 4.9 in 1% solution. The precipitate after conducting the test consists of so-called floaters (usually a few quite large particles remaining on the surface), flakes (dispersed tiny particles in the whole body of the solution) and sediment at the bottom of the beaker. Some methods count the floaters and sediment and express the results subjectively. It is obvious that floaters are the most undesirable because they cause an unpleasant appearance of the beverage that can be seen immediately by the consumer.

 

The NZDB method expresses the result objectively with a number, however without distinguishing between the characters of the appearance of coffee and milk beverage. 100 ml of boiling water is added to 0.8 g instant coffee in a 250 ml beaker. After the black coffee temperature drops to 80 ± 0.5°C, 2 g of the milk powder is added and the stop watch started simultaneously. After 5 seconds, the contents are stirred with a teaspoon using a circular motion (6 complete revolutions clockwise followed by 6 complete revolutions anticlockwise). The total stirring time should be 5 seconds. After 10 minutes the sediment is re-suspended with a single gentle stir, and filled into two 50 ml centrifuge tubes (the same as for insolubility index). The tubes are centrifuged for 5 minutes at 164 G and the volume of sediment is read to the nearest scale mark in each tube. The result is the sum of both readings.

 

In 2005 an IFD standard method ISO 15322/IDF 203:2005 based on the above described NZDB method was published as ‘Dried milk and dried products – Determination of their behaviour in hot coffee (Coffee test).

White flecks are tiny flakes floating in the reconstituted solution. If it is allowed to stand for several minutes they rise to the surface forming a thin layer. The original method for the detection of white flecks was just a visual observation of the reconstituted milk in a thin layer with a teaspoon placed close to the wall of a beaker as background.

The method utilizes the fact that white flecks clog the mesh, and depending on their quantity, allow only a limited amount of liquid to pass through.

The standard method for the determination of total fat content in milk powders is the Röse-Gottlieb method as described in IDF-Standard 123A:1988 [38]. The amount of 1 g whole milk powder or 1.5 g skim milk powder is weighed into a graduated shaking cylinder with wellfitting stopper. Add 10 ml of water for dissolving and heat if necessary. Add 1.5 ml 25%-NH3-solution and heat in a water bath for 15 minutes at 60 - 70°C with occasional shaking. Cool down; add 10 ml 96%-ethanol and mix. Add 25 ml ethyl ether (b.p.34 - 35°C), close the cylinder tightly and mix by turning upside down for 1 minute. Add 25 ml petroleum ether (b.p. 40 - 60°C) and repeat mixing as above. Allow to stand for at least 1 hour to achieve an ether phase clearly separated from the water phase. By means of a siphon, transfer the ether phase to a pre-weighed 150 ml Erlenmeyer flask rinsing at the end the siphon with a little ether. Take care not to introduce any water phase into the flask. Repeat the addition of 25 ml ethyl ether and 25 ml petroleum ether keeping the same procedure as above collecting the ether phase into the same flask. Evaporate the ether and finally dry the flask for 1 hour in an oven at 102 ± 2°C. Cool in a desiccator and weigh. The result is expressed in percentage on powder.

 

The quick method for routine determination is the Gerber-Teichert method. Into a special butyrometer with scale 0 - 35 or 0 - 70% is added successively 10 ml sulphuric acid (90 - 91%, density 1.816 ± 0.003 g/ml), 8 ml distilled water (not to be mixed with the acid), exactly 2.5 g powder and 1 ml amyl alcohol (density 0.811 ± 0.002 g/ml). The butyrometer is closed with a rubber stopper, shaken vigorously for 5 minutes and turned several times upside-down to mix all the acid with the contents. The tube is then centrifuged for 15 minutes in a centrifuge heated to 65°C at 1200 RPM. The 5-minutes shaking and centrifuging is repeated once more. By means of the rubber stopper adjust the fat column to appear in the graduated part of the tube, spin again for 5 minutes and read the fat percentage directly.

The determination of free fat content of a milk powder is based on extraction by fat solvents. There are many alternatives especially as to extraction time and used solvent. Suitable solvents include petroleum ether and carbon tetrachloride, although for environmental reasons the latter should not be used anymore. Extraction time can vary between 15 minutes and 24 hours. Longer extraction times give higher values. In case of free fat determination, however, the most interesting is the surface free fat which is extracted very quickly and therefore there is no reason for long extraction time. 

 

The routine method is as follows: Weigh 10 g powder into 250 ml Erlenmeyer flask with ground glass stopper. Add 50 ml solvent, close the flask and agitate in a shaking device for 15 minutes. Filter the solution into a 100 ml Erlenmeyer flask. Pipette 25 ml of the filtrate into a pre-weighed 50 ml Erlenmeyer flask. Evaporate the solvent on a hot plate (or similar) and dry in an oven at 105°C for 1 hour. The content of the free fat is expressed as percentage of the powder:

where: A = evaporation residue from 25 ml of solvent B = amount of used powder in g 

 

Alternatively the free fat can be expressed also as a percentage of total fat:

The original methods for the determination of particle size distribution involved microscopic counting or sifting. Nowadays a number of sophisticated instruments are available, e.g. Malvern instruments, based on laser beam diffraction, Coulter counter etc. 

 

The microscopic counting method, used for non-agglomerated powders was very laborious, time consuming and the results were influenced by the subjective judgement of the analyst. It was necessary to count with simultaneous evaluation of the size of at least 1000 particles dispersed in toluene on the microscopic glass under the microscope. Nowadays this method has been virtually replaced by automatic modern methods.

 

The sieving test is suitable only for agglomerated powders and is based on sifting 100 g of powder through a number of sieves in a shaking apparatus, usually for 5 minutes. The used screens are 200 mm in diameter and recommended mesh sizes are for instance: 500, 315, 250, 212, 180, 125 and 90 μm. Fat containing powders require an addition of 2% of free flowing agent, mixed gently with the powder, before testing. The expression of results can be found in Table 3.3.

 

In the Malvern method, the powder is presented to the laser beam while airborne in a stream of air or kept in a suspension in isopropanol in a cuvette under magnetic agitation. The results are calculated by a computer and shown on a screen and print-out. 

 

Every method determining particle size distribution is influenced by the breaking down the agglomerates due to mechanical handling involved (friction on the sieve during shaking, agitation of the suspension etc.). A demonstration of this breaking-down effect can be seen in Fig. 11.8. An agglomerated sodium caseinate powder was kept under constant gentle agitation in the measuring cuvette for 600 s. A size print-out was taken every 100 s. During this time, the mean particle size dropped from 400 to 200 μm and the fines increased from roughly 2, 5 and 7 to 4, 22 and 30 μm respectively.

However, the Malvern method is gentler than the sieving procedure. When comparing both methods, the Malvern method indicates smaller mean particle than the sieving test. As mean particle size increases the difference between measurements decreases and at about 200 μm the results are close to each other. Above 200 μm the Malvern method gives higher values. The former effect is because Malvern detects the fines better. The latter effect is due to the gentler conditions of treatment. The large agglomerates are not broken down as much with the Malvern method.

The principle for determining mechanical stability of agglomerated powders is based on applying a defined mechanical treatment followed by determination of fines created. These are normally defined as the sifting fraction smaller than 150 μm. Before applying the mechanical treatment, this fraction must be removed from the original powder by gentle sifting. Mechanical treatment usually involves 10 minutes shaking. For a comparison of mechanical stability of various samples Malvern analysis can be used, as shown in Fig. 11.8.

The hygroscopicity is a property, which together with the method for determining degree of caking, is particularly suitable to classify whey powders as to their ability to pick up moisture from the surrounding air during storage. It is defined as the final moisture content of powder after exposure to humid air of 79.5% relative humidity at 20°C under the conditions of the method.

The apparatus for the determination of hygroscopicity is shown on Fig. 11.9. The washing bottle is filled with a saturated solution of NH4Cl with surplus of crystals at the bottom and the apparatus is connected by means of a three-way cock to a vacuum pump. The other passage of the cock is open to the atmosphere. The cock must always be in this position when starting or stopping the vacuum pump. The surrounding temperature must be 20 ± 2°C. After assembling the apparatus with empty Gooch filter and starting the vacuum pump, the cock is turned to suck the air through the apparatus and the flow rate is adjusted to 300-400 ml/ min. After 5 minutes the flow is stopped, the Gooch filter weighed first empty (a) and then with about 0.5 g of the test powder (b). The apparatus is assembled again and the air circulation started. Check the weight increase after 4 hours and then after each hour. The measurement is completed when the difference in weight between two successive weightings is negative and the result is then the second but last weighing.

where: a= weight of powder

           b= weight increase

           %FM= free moisture (see section 11.1.2.)

 

The powder component, which is mainly responsible for moisture uptake, is the amorphous lactose, but also proteins and minerals can pick up moisture. During the determination the powder starts to pick up moisture, but when the level of moisture is sufficiently high the lactose starts to crystallize. When this happens, the water activity of the powder decreases and the powder loses moisture. If the process is allowed to continue to a condition of final equilibrium, then all whey powders of normal composition would exhibit final moisture content of about 12%. This occurs when the lactose is completely crystallised and the moisture increase is exclusively caused by other-than-lactose components. Well pre-crystallized whey powders reach this point with continuous moisture increase and quickly, while non-pre-crystallized powders can reach up to 30% weight increase before the moisture starts to decrease. It might be difficult to determine the maximum weight increase before the moisture begins to decrease. Therefore a much better expression for hygroscopicity is the next method, i.e. degree of caking.

 

Thus the hygroscopicity of whey powders, expressed in terms of maximum weight increase, is classified as follows:

 

Non-hygroscopic max. 10%

slightly hygroscopic 10.1-15%

hygroscopic 15.1-20%

very hygroscopic 20.1-25%

extremely hygroscopic > 25%

The degree of caking, called also cakeness is the portion of powder remaining on a given mesh when sifting under prescribed conditions after the powder has been exposed to hygroscopicity test (section 11.13.) and then re-dried.

 

where: a = the amount of dried sample

           b = the amount remaining on the sieve.

 

The results are evaluated by following scale:

 

non-caking max 10%

slightly caking 10.1-20%

caking 20.1-50%

very caking > 50%

extremely caking 100%.

The determination of α-lactose content (of total lactose) is based on two refractometrical readings. For the first reading, all operations must be done at low temperature (below 5°C) to avoid mutarotation of the lactose. The second reading is then done after completing the mutarotation and achieving equilibrium. 

 

All the chemicals, measurement equipment and samples must be placed in the refrigerator overnight and the operation has to be done quickly to avoid a temperature increase above 5°C. Preferably, all operations should be done in a cold room. The procedure is as follows: The weighed amount of sample (a) corresponding to 1.0-1.5 g lactose is dissolved in about 10 ml cold distilled water using a mortar and pestle. The obtained paste is then diluted and transferred quantitatively into volumetric flask of 100 ml. Add 5 ml of 2.5% tannin solution, 10 ml of 10% lead acetate solution and fill up to the mark. Mix the contents and filter into a 200 ml Erlenmeyer flask. The first portions of the filtrate should be returned on the filter because it is sometimes not quite clear (if no cold room is available conduct the filtration inside the refrigerator). The solution is then filled into a pre-cooled polar metric tube. The polar metric reading (P1) must be done within the first minute. The rest of the filtrate is heated to 80°C and kept for 30 minutes. After cooling down to 5°C, the second polar metric reading (P2) is taken. Following equations apply:

where: %TL % total lactose content (as anhydride)

           % αLTL % α-lactose of total lactose

           % A % amorphous lactose

           % αLan % α-lactose monohydrate (as anhydride)

           % H2Ocryst % water of crystallization

           % αLM % α-lactose monohydrate (as monohydrate)

           % cryst. % crystallized lactose of total lactose (degree of crystallization)

 

The value y in the equation [11,15] is calculated from the value x which is the proportion of β-lactose to α-lactose at the temperature of the concentrate prior to drying. Both values are shown in Table 11.2.

Otherwise a good indication of the content of α-lactose monohydrate can be obtained from the difference between total and free powder moisture (methods 11.1.3. and 11.1.2.) and multiplying the result by 19 (compare with equations 10.16 and 10.17).

The determination of titratable acidity is a very common laboratory procedure in the dairy industry. The various expressions for titratable acidity have been explained in section 9.4.4. Nowadays the mostly used procedure expresses acidity as ‘lactic acid’ according to ADMI standard described in Bulletin 916 [36]. The IDF-standard is IDF 86:1978 [39]. 

 

10 g of skim milk or butter milk powder, 13 g of whole milk powder or 6 g of whey powder are dissolved in 100 ml distilled water using a mixer (as for Insolubility index, see section 11.2.). The amount of powder for other products should correspond to their natural concentration. After mixing, the solution should stand for 1 hour. Thereafter follows a gentle mixing and the transfer of 17.6 ml into a white glazed porcelain casserole using a pipette. 0.5 ml 1%-alcoholic phenolphtalein solution is added and titration is conducted using 0.1 N NaOH until a faint pink colour persists for 30 seconds. Titratable acidity is then the consumption of 0.1 N NaOH in ml divided by 20. The conversion Table 11.3. for various expressions for titratable acidity is below (%l.a.=% lactic acid, Th=Thörner, D=Dornick, SH=Soxhlet-Henkel).

The procedure for the determination of whey protein nitrogen index (WPNI) was originally introduced by ADMI, exclusively for classifying skim milk powders according to heat treatment. Nowadays it is used also for other products. For instance, knowledge of WPNI is very useful when investigating quality problems of instant whole milk powder. 

 

Reconstitute 2 g of skim milk powder in a test tube (25x150 mm) in 20 ml distilled water, add 8 g of NaCl, place in water bath at 37 ± 1°C for 30 minutes while shaking 10 times during the first 15 minutes. Without cooling, shake the mixture and filter through S&S 602 filter paper (or Munktell 20H+110 or S&S 605 or S&S Selecta folding filter 572½), re-filter the first portions if cloudy and collect 6-7 ml of filtrate. Pipette 5 ml into 100 ml Erlenmeyer flask, add 50 ml saturated solution of NaCl and mix slowly. Fill two photometer cuvettes with a known amount of filtrate and add 1 drop of HCl solution (23 ml of concentrated HCl-37% in 77 ml distilled water) per each 5.5 ml into the first cuvette. Close the cuvette and mix by inverting twice. The second cuvette is used for adjusting the spectrophotometer. Set the photometer to wavelength 420 nm and adjust the transmission by means of second cuvette to 100%. 5-10 minutes after adding HCl to the first cuvette, invert it once again and make a duplicate reading of transmission. If the difference between the duplicates is greater than 2% another pair should be analysed. The result is expressed as mg of un-denatured whey protein nitrogen per g powder and is found using the graph in Fig. 11.10.

The WPNI is defined for 1 g of powder without considering the moisture content. When analysing whole milk powder the amount of sample must contain the same amount of non-fatsolids as 2 g of skim milk powder (supposed to contain 4% moisture and 0.5% fat).

 

Thus the amount of whole milk powder for the analysis is calculated as follows:

where: %M= % moisture

           %F = % fat

The flowability is an important property of milk powders and it can be easily judged by the naked eye. However, it is not that easy to find a method which could cope with the whole range of products from easy flowable to so-called ‘dead’ powders. For instance the measurement of ‘angle of repose’ or ‘time of flow’ of a given amount of powder through a given funnel may be good enough for a particular powder. However, it may not work with a less flowable product.

 

The equipment for the determination of flowability is a stainless steel drum according to the drawing in Fig. 11.11, attached horizontally to a motor with gear so as to operate at 30 RPM. An amount of powder corresponding to 25 times the bulk density (tapped 100 times, expressed as g/cm3) is poured into the drum. The plastic lid is fastened and the rotation started simultaneously with the stop watch. When all powder has left the drum the time in seconds is recorded. The flowability is an average of three measurements.

The advantage of the method described here is that it is quite universal with a broad scale. Table 11.4 shows some examples of expected flowabilities of various powders.

Lecithin in cold water instant products is contained in the surface free fat layer. Thus the principle of the method is extraction of the surface free fat and determination of its phosphorus content gravimetrically. 60 g of sample is extracted by means of 300 ml solvent using the same procedure as described in section 11.10. The solvent is then evaporated to about 30 ml and transferred quantitatively into a pre-dried and pre-weighed quartz dish in which the evaporation is continued. The residue is then dried at 105 ± 2°C for 1 hour and the weight of the residue is recorded. The residue is covered with 2.5g MgO and heated on a Bunsen burner until ignition. The dish is then placed in an oven heated to 800°C and left for 2 hours or until white ash remains. The ash is transferred into 250 ml beaker with 10 ml distilled water. Using 15 ml HNO3-H2SO4-mixture (1 l of nitric acid D=1.2 with 30 ml concentrated sulphuric acid), rinse the dish 3 times to transfer the residue quantitatively, heating the dish each time. The solution is heated until everything is dissolved. Using a pipette, 50 ml of ammonium molybdate solution is added (50 ml ammonium sulphate is dissolved in 450 ml concentrated nitric acid d=1.4 and 150 g of ammonium molybdate is dissolved in 400 ml hot water, which is then cooled down, mixed with the previous solution and filled up to 1000 ml). After stirring with a glass rod, the solution is stored in darkness for 36 hours. A Jena A4 glass filter funnel is rinsed with acetone and dried in vacuum at 15 mm Hg for 30 minutes. Weigh the beaker, place it on a suction flask attached to vacuum pump and filter the contents of the 250 ml beaker transferring first the supernatant. Transfer the sediment quantitatively by means of NH4NO3 solution (1 l of 2% ammonium nitrate solution mixed with 5 ml of 20% nitric acid) repeating 6 times this procedure. Suck the A4 glass filter beaker dry, fill it with acetone stirring the sediment with glass spatula and suck it dry again. Repeat the same procedure once more. Dry the A4 beaker in vacuum at 15 mm Hg and record the weight. The calculation is the following:

where: a = evaporation residue from c ml filtrate

           b = weight of sample

           c = ml of filtrate used = weight of sediment

           e = weight of free fat

There are two methods available: the oven drying method, which is considered as an exact reference method, and the refractometrical method as a quick routine control procedure. With oven drying, milk concentrate is well mixed with sea sand. 20-30 g of sea sand is dried in a small weighing dish with well-fitting cover with a small spatula for 2 hours at 100°C. After cooling down with the lid on, the dish is weighed first empty (W1) and then again after adding of about 1.5 g of concentrate (W2). 5 ml of distilled water is added and the components are well mixed with the spatula. The excess of water is first evaporated on a water bath for about 20 minutes under occasional stirring and then dried in the oven heated to 100°C for 2 hours. After cooling, the dish is weighed again (W3). The result is:

The refractometrical method is based on the determination of the index of refraction of a drop of concentrate, and the measurement is conducted according to the instructions for the apparatus in question. It is an advantage to use a refractometer with scale in sugar degrees (°Brix); otherwise it is necessary to use a table for recalculation of the index of refraction to sugar degrees. It is important to keep the refractometer prism clean and grease-free by washing carefully with distilled water after each use and possibly occasionally with petroleum ether.

 

The total solids content of the concentrate is then obtained reading in °Brix multiplied by an empirical factor, which depends on the product composition:

 

for skim milk      0.9

for whey            0.97

for whole milk    1.0

 

The solubility problems of milk powder are often considered as a spray drying problem. However, when troubleshooting insolubility index problems, it is a useful practice to occasionally check the insolubility index of the concentrate to find out whether the deterioration of the solubility has taken place prior to spray drying.

 

When sampling the concentrate from the evaporator using a sampling valve, it is important to wash carefully the mouth of the valve to remove any possible dry residues of the concentrate. The method is in principle the same as described for powders in section 11.2. The amount of concentrate for the determination should be such to contain 10 g and 13 g of solids for skim milk and whole milk respectively. The concentrate is diluted with distilled water of 24°C to give 107 ml. The further procedure and precautions are the same as described previously (section 11.2).

For routine check of the viscosity of all milk and milk product concentrates the viscosity is the value obtained by Brookfield viscometer type LVT, equipped with spindle no. 2 at 60 RPM and 40°C. The reference temperature of 40°C was chosen in order to enable comparison of viscosities of concentrates obtained under various conditions. Furthermore it is easy to adjust being always lower than the temperature of the concentrate immediately after sampling and because the age thickening at such a temperature is not that fast. 

 

The concentrate is filled into a 250 ml beaker to reach the level of the recess on the spindle, and stirred, possibly in a cold water bath, until the correct temperature is reached. The apparatus is adjusted to the correct height of the spindle and started at speed 60 RPM. The result is an average of three readings after the previous readings have stopped decreasing. The value of dynamic viscosity in cP is the reading average multiplied by 5.

 

For concentrates which are supposed to be supplied cold to the spray dryer, as for instance whey concentrates, the measurement is taken at the actual temperature and using a spindle corresponding to expected viscosity. The total solids content and the temperature of the concentrate must be recorded together with the viscosity.

The degree of crystallization of whey concentrates is the amount of lactose in form of -lactose monohydrate present in the total lactose content expressed as percentage. The method is based on two refractometrical readings of the concentrate before and after crystallization. For the initial refractometrical reading before crystallization, it is necessary to take an average of readings taken at regular intervals (at least 10 readings for a batch) of the concentrate leaving

the evaporator. The second reading is taken on concentrate from the crystallization tank. 

 

For exact determination it is necessary to know the content of lactose and total solids content of the concentrate using methods described in sections 11.15 and 11.20.1, respectively. Then the degree of crystallization is:

where: S1 = first refract metrical reading

           S2 = second refract metrical reading

           L = total lactose content of the concentrate

           LTS = total lactose content of whey solids

           % TS = total solids content of the concentrate

 

For a routine fast determination, it is possible to use for total content of lactose LTS a value, which is usually known in a given factory and given season. For the total solids content of the concentrate the S1-reading is multiplied by 0.97 (see 11.20.1). See also Fig. 10.22.