Processing

Several important factors need to be taken into consideration in the design of food processing plants in order to assure the quality of the end products. One of them is the question of rheology, which concerns the ow behaviour of the products. A main issue is also the measurement, adaptation and application of viscosity data, which concerns the design calculations of processing equipment.

Viscosity & Rheology

Theoretical and practical considerations in liquid food processing
Dr Ulf Bolmstedt, Company Specialist in Heat Transfer & Fluid Mechanics,Tetra Pak AB

In the dairy industry there are cream and cultured milk products whose characteristics can be partially or completely spoiled if their ow behaviour is not understood. In the prepared food industry there are dessert products often containing starch based additives, soups and sauces containing particles and tomato products that require certain considerations regarding choice and design of equipment. What follows here is a brief guide to the ow behaviour of some typical food industry products, how to classify the products and how to measure and adapt viscosity data. Denition Rheology is dened as the science of deformation and ow of matter. The term itself originates from Greek rheos meaning to ow. Rheology is applicable to all types of materials, from gases to solids. The science of rheology is young,only about 70 years of age, but its history is very old. In the book of Judges in the Old Testament the prophetess Deborah declared The mountains owed before the Lord....Translated into rheological terms by Professor Marcus Reiner1, this expression means everything ows if you just wait long enough, a statement that is certainly applicable to rheology. It was also described by the Greek philosopher Heraclitus as panta rei - everything flows. Professor Reiner, together with Professor Eugene Bingham, was the founder of the science of rheology in the mid-20s. Rheology is used in food science to dene the consistency of different products. Rheologically the consistency is described by two components, the viscosity (thickness,lack of slipperiness) and the elasticity (stickiness, structure). In practice, therefore, rheology stands for viscosity measurements,characterisation of ow behaviour and determination of material structure. Basic knowledge of these subjects is essential in process design and product quality evaluation. Characterisation of materials One of the main issues of rheology is the denition and classification of materials. Normal glass, for

instance, is usually dened as a solid material, but if the thickness of an old church window is measured from top to bottom a difference will be noted. Glass does in fact ow like a liquid, albeit very slowly. Also, an air bubble captured in a glass object will rise, safely but slowly, due to the difference in density between air and glass. The process may, however, take centuries to observe and is of course not evident to human beings - the time of observation is too short. One way of characterising a material is by its relaxation time, i.e. the time required to reduce a stress in the material by ow. Typical magnitudes of relaxation times for materials are: seconds Gases < 10-6 Liquids 10-6 - 102 seconds Solids > 102 seconds The Deborah Number, D, named after the prophetess Deborah,is a way of characterising the ow behaviour of a material.The Deborah Number is the ratio between time of relaxation and the time of observation: D = time of relaxation time of observation Consequently, the Deborah Number is large for materials of high viscosity and low for materials of low viscosity. Another way of dening materials rheologically is by the terms viscous, elastic or viscoelastic. Gases and liquids are normally described as viscous uids.By denition an ideal viscous uid is unable to store any deformation energy. Hence it is irreversibly deformed when subjected to stress; it ows and the deformation energy is dissipated as heat, resulting in a rise of temperature. Solids, on the other hand, are normally described as elastic materials. An ideal elastic material stores all imposed deformation energy and will consequently recover totally upon release of stress. A viscous uid can therefore be described as a uid, which resists the act of deformation rather than the state of deformation, while an elastic material resists the act as well as the state of deformation. A number of materials show viscous as well as elastic properties, i.e. they store some of the deformation

Reprinted from New Food Volume 3 Issue 2, Summer 2000 For more information please contact Russell Publishing Ltd., Tel: +44 (0)20 7436 3642 Fax: +44 (0)20 7436 3610 Email: info@russellpublishing.com

Processing
energy in their structure while some is lost by ow. These materials are called viscoelastic and there are many examples among foodstuffs,such as starch-based puddings, mayonnaise, and tomato pures. dv yx = . __ =. dy The proportionality constant is thus equal to the viscosity of the material. The ow curve, which is a plot of shear stress versus shear rate, will therefore be a straight line with slope for a Newtonian uid. The viscosity curve, which is a plot of viscosity versus shear rate, will show a straight line at a constant value equal to . A Newtonian uid can therefore be dened by a single viscosity value at a specied temperature. Water, mineral and vegetable oils and pure sucrose solutions are examples of Newtonian uids. Low-concentration liquids in general, such as whole milk, skim milk and most single strength fruit juices, may for practical purposes be characterised as Newtonian uids. The viscosity temperature dependence of e.g. a 20OBx sugar solution is around 1.5% per degree at 20OC and around 3% per degree at 80OC. The gures increase with increased concentration.

Fig 1: Different types of shearing Shearing In rheology, shearing of a substance is the key to knowledge of ow behaviour and structure. A sheared ow is achieved through ow between parallel planes, rotational ow between coaxial cylinders where one cylinder is stationary and the other one is rotating, telescopic ow through capillaries and pipes, and torsional ow between parallel plates. To enable study of the viscosity of a material, the shearing must induce stationary ow of the material. The flow occurs through rearrangement and deformation of particles and through breaking of bonds in the structure of the material. Shear stress is dened as: F yx = __ A F = force, N A = area, m2 shear rate as: dv . d = __ = __ dt dy

Fig 2: Denition of shear stress and shear rate is based on shearing between planes (Isaac Newton, 1687). Non-Newtonian uids Materials, which cannot be dened by a single viscosity value at a specified temperature, are called non-Newtonian. The viscosity of these materials must always be stated together with a corresponding temperature and shear rate. If the shear rate is changed the viscosity will also change. Generally speaking, high concentration and low temperature induce or increase non-Newtonian behaviour. Apart from being shear rate dependent, the viscosity of non-Newtonian uids may also be time dependent, in which case the viscosity is a function not only of the magnitude of the shear rate but also of the duration and, in most cases, of the frequency of successive applications of shear. Non-Newtonian materials that are time independent are dened as shear thinning, shear thickening or plastic. Non-Newtonian materials that are time dependent are dened as thixotropic, rheopectic or anti-thixotropic. Shear thinning ow behaviour The viscosity of a shear thinning uid (sometimes also denoted pseudoplastic fluid) decreases with increasing shear rate. Most liquid food systems belong

and apparent viscosity of a uid as:

. / .

If we want to study the elasticity (structure) of a material, the shearing must be very gentle so as not to destroy the structure. One way to achieve this is to apply an oscillating shear to the material with amplitude low enough to allow an unbroken structure to be studied. Shearing between parallel planes is normally used for the basic denition of shear stress and shear rate, corresponding to how much deformation is applied to the material and how fast. Newtonian uids Newtonian uids are those having a constant viscosity dependent on temperature but independent of the applied shear rate. One can also say that Newtonian uids have direct proportionality between shear stress and shear rate in laminar ow:

Processing

Fig 3: Flow curves for Newtonian and non-Newtonian uids. to this category of uids, e.g. yoghurt, fruit juice concentrates, salad dressings and tomato ketchup. The shear rate dependency of the viscosity can differ substantially between different products, and also for a given liquid, depending on temperature and concentration.The reason for shear thinning ow behaviour is that an increased shear rate deforms and/or rearranges particles, resulting in lower ow resistance and consequently lower viscosity. It should be noted that although sucrose solutions show Newtonian behaviour independent of concentration,fruit juice concentrates are always signicantly non-Newtonian. Hence a non-Newtonian uid like yoghurt or fruit juice concentrate being pumped in a pipe shows decreased apparent viscosity if ow rate is increased. This means in practice that the pressure drop of a non-Newtonian uid in laminar ow is not directly proportional to the ow rate as for Newtonian uids in laminar ow. Shear thickening ow behaviour The viscosity of a shear thickening uid increases with increasing shear rate. This type of ow behaviour is generally found among suspensions of very high concentration. A shear thickening uid exhibits dilatant ow behaviour, i.e. the solvent acts as a lubricant between suspended particles at low shear rates but is squeezed out at higher shear rates, resulting in denser packing of the particles. Typical examples of shear thickening systems are wet sand and concentrated starch suspensions. Plastic ow behaviour A uid, which exhibits a yield stress, is called a plastic fluid. The practical result of this type of flow behaviour is that a signicant force must be applied before the material starts to ow like a liquid (often

Fig 4:Viscosity curves for Newtonian and non-Newtonian uids.

Fig 5: Flow curves for time-dependant non-Newtonian uids. referred to as the ketchup effect). If the force applied is smaller than the force corresponding to the yield stress, the material stores the deformation energy, i.e. shows elastic properties, and hence behaves as a solid. Once the yield stress is exceeded, the liquid can ow like a Newtonian liquid and be described as a Bingham plastic liquid, or it can ow like a shear thinning liquid and be described as a viscoplastic liquid. Typical plastic uids are quarg, high pectin pineapple juice concentrate,tomato paste and certain ketchups. Outside the liquid food world toothpaste, hand cream and greases are typical examples of plastic uids. A simple but still very effective way of checking a uids possible plastic properties is to just turn the

Processing
ty increases with increasing shear rate, but also with time at a constant shear rate. As with thixotropic uids, the ow behaviour is illustrated by a loop test. This type of ow behaviour is very uncommon among foodstuffs. Flow behaviour models For the adaptation of viscosity measurement data to process design calculations some kind of mathematical description of the ow behaviour is required. For that purpose several models are available, e.g. Ostwald, Herschel-Bulkley, Steiger-Ory, Bingham, Ellis and Eyring. These models relate the shear stress of a uid to the shear rate, thus enabling the apparent viscosity to be calculated, as always, as the ratio between shear stress and shear rate. By far the most general model is the HerschelBulkley model, also called the generalised power law equation, which in principle is an extended Ostwald model. The main benet of the generalised power law equation is its applicability to a great number of
Shear rates sedimentation chewing stirring pumping spraying rubbing 10-6 101 101 102 103 104 10-4 102 103 103 104 105 s-1 s-1 s-1 s-1 s-1 s-1

Fig. 6: Logarithmic ow and viscosity curves for a shear thinning power law uid jar upside down. If the uid will not ow by itself it probably has a signicant yield value. If it ows by itself, but very slowly, it probably has no yield value but a high viscosity. Information of this kind is of vital importance to process plant design regarding the dimensions and layout of storage and process tank outlets and pump connections. Time-dependent ow behaviour A thixotropic uid can be described as a shear thinning system where the viscosity decreases not only with increasing shear rate but also with time at a constant shear rate. Thixotropic ow behaviour is normally studied in a loop test. In this test the material is subjected to increasing shear rates followed by the same shear rates in decreasing order. The time-dependent thixotropic ow behaviour is seen from the difference between the ascending and descending viscosity and shear stress curves. To recover its structure, the material must rest for a certain period of time which is characteristic for the specic material. This type of ow behaviour is shown by all gel-forming systems. Typical examples of thixotropic fluids are yoghurt, mayonnaise,margarine and ice cream.Outside the liquid food world, perhaps brush paint is the most well-known product showing thixotropic behaviour. A rheopectic uid can be described as a thixotropic uid but with the important difference that the structure of the uid will only recover completely if subjected to a small shear rate. This means that a rheopectic uid will not rebuild its structure at rest. An anti-thixotropic fluid can be described as a shear thickening system, i.e. one where the viscosi-

Viscosities air water olive oil glycerol syrup molten glass glass 10-5 10-3 10-1 106 102 1012 1040 Pas Pas Pas Pas Pas Pas Pas

n and k values fruit concentrate molten chocolate sour milk quarg apple puree tomato paste grease

n = 0.7 n = 0.5 n = 0.3 n = 0.3 n = 0.3 n = 0.2 n = 0.1

K K K K K K K

= 2 Pasn = 50 Pasn = 3 Pasn = 4 Pasn = 10 Pasn = 70 Pasn = 1000 Pasn

Yield stress ketchup mustard mayonnaise 14 Pa 38 Pa 85 Pa

Fig. 7:Typical data on shear rates, viscosities and power law constants n and K.Valid around room temperature (1 Pas = 1000 mPas = 1000 centiPoise). N.b.: only given as examples and should not be used for design calculations.

Processing
non-Newtonian uids over a wide range of shear rates. Furthermore, the power law equation lends itself readily to mathematical treatment, for instance in pressure drop and heat transfer calculations. The generalised power law equation is applicable to plastic as well as shear thinning and shear thickening uids according to the following:

(0) = . n
where = 0 = K = . = n =

shear stress yield stress consistency shear rate ow behaviour index

Pa Pa Pasn s-1 -

Suitable modication of the generalised power law equation makes it possible to rewrite it to express each type of ow behaviour. For Newtonian uids the power law equation looks like this: (K = and n = 1):
n =. = .

For a plastic uid the power law equation is used in the fully generalised form, with n < 1 for viscoplastic behaviour and n = 1 for Bingham plastic behaviour. For a shear thinning or shear thickening uid the power law equation becomes: =.

.n
Fig 8: Operating principles of different types of viscometers. Instead, a special design of the capillary viscometer is the tubular viscometer, with a diameter of e.g. 25 or 38 mm compared to a few mm for the capillary type. The tubular viscometer is used for the determination of the power law constants and is especially suitable for particulate products. The drawback of the tubular viscometer is that it often requires large product volumes and that the measuring system can be quite bulky and expensive. Measurement of non-Newtonian uids requires instruments where the applied shear rate is accurately dened,i.e.where the shearing takes place in a narrow gap with a small shear rate gradient. This fundamental requirement excludes viscometers where the gap is too big or even undened, as it is in viscometers of spindle type. It must be strongly emphasised that viscosity measurements of non-Newtonian uids carried out at undened or out-of-range shear rates should not be used as a basis for quantitative analysis of viscosity gures or rheological parameters. Rotational viscometers are available as portable as well as stationary instruments. Portable types usually

with n < 1 and n > 1, respectively. For time-dependent uids, which in practice means thixotropic uids, the mathematical models required for description of rheological behaviour are generally far more complex than the models discussed so far. These uids are therefore often described by timeindependent process viscosities normally tted to the power law equation. Measuring equipment The main types of viscometers are rotational and capillary. Rotational viscometers are of spindle, coneplate, plate-plate or concentric cylinder type. The last-named may be of Searle (rotating bob) or Couette (rotating cup) type. Capillary viscometers may be of atmospheric or pressurised type. Generally speaking, rotational viscometers are easier to use and more exible than capillary viscometers. On the other hand, capillary viscometers are more accurate at low viscosities and at high shear rates. However, for practical use in liquid food viscometry they are less applicable due to their sensitivity to even small particles like fruit juice bres.

Processing
temperatures related to the process to be studied. The intended use of the measured data should therefore be considered before measuring takes place, for instance if the viscosity data are to be used in the design of a deep cooler or of the heating section of a steriliser. Due to practical limitations the maximum applicable temperature for most viscometers is around 900C. At higher temperatures the risk of evaporation from the surface of the test sample followed by skin formation leading to increased momentum and hence false readings is signicant. Hence a special type of pressurised measuring system has to be employed. With these systems temperatures up to 1500C are possible, i.e. a typical sterilisation process up to 1400C can be fully covered regarding viscosity data. It is also most important that the temperature is kept constant during the test period and, of course, that it is accurately measured. A temperature change of 30C can often cause a change in viscosity of 10 per cent. To increase the accuracy of data evaluation, measurements should be made at as many different shear rates and temperatures as possible. In addition, heating effects must be considered.In a substance containing warm-swelling starch, for example, the viscosities before and after heating above swelling temperature will differ signicantly.

Fig 9: Output from a rheometric measurement showing the elastic modulus G, the viscous modulus G and the phase angle . come in a shockproof case equipped with all necessary accessories. They are basically manually operated, although some manufacturers provide connections for use with personal computers. Today many of the portable instruments are equipped with processors capable of running the viscometer according to the desired scheme and also of storing all measuring data for later download to a printer or a PC. Stationary installations are normally computer controlled for automation of measuring sequences and data evaluation. The software usually includes possible tting to a number of rheological models, plotting of ow curves, etc. A rotational viscometer is normally insufcient for carrying out a complete rheological analysis, for instance determination of structure breakdown in yoghurt. This type of analysis requires a more sophisticated instrument, generally called a rheometer. With a rheometer, operating with torsional vibration or oscillation rather than rotation, the uid can be rheologically analysed without its structure being destroyed. Typical applications are viscoelastic uids, for which a rheometer can be used to determine the viscous and elastic properties of the uid separately. Ordinary viscometers and rheometers should not be used for measurement of substances with very high viscosities, such as butter, cheese and vegetable fats. Certain types of penetrometers are available instead, but these cannot be used to obtain scientific rheological results since a penetrometer gives only empirical information. A special type of consistometer is preferably used within the tomato industry. This type of instrument gives the result in so-called 0Bostwick, which is a unit applicable only to comparison of different products. Measuring techniques Viscosity measurements should always be carried out for a representative range of shear rates and

Fig 10: Database for storage of viscosity and rheology measurement data. Example showing data for 5% Colo 67 starch solution from 30 to 1400C. Furthermore, storage conditions and time factors must be taken into consideration. The rheological properties of many products, e.g. fermented dairy products, change with time, and if the purpose of the viscosity measurement is to supply data for process design, the measurements should preferably be made in as close connection as possible to the actual processing stage. When measurements are performed at a regular basis the results are preferably stored in a database in order to facilitate comparison of various products. In practice all varieties of liquid food products are unique regarding viscosity data, meaning that data measured on one type of vanilla pudding, one type of tomato pure or one type of yoghurt cannot be safely applied to another type or brand of a product with the same name or even with roughly the same

Processing
composition. However, with access to a database containing data on a substantial amount of products there is always a possibility to extract a range of viscosities for a certain type of product in case no other information is available. Pressure drop in pipes Some useful equations are given below for manual calculation of pressure drop and shear rates for laminar pipe ow. All equations are based on the power law expression, as most food systems in processing conditions can be described by this expression. The equations are applicable to Newtonian as well as non-Newtonian uids depending on the value of n used in the calculation: n<1 for shear thinning (pseudoplastic) uids, n=1 for Newtonian uids, and n>1 for shear thickening (dilatant) uids. The relationship between ow rate and pressure drop and between ow rate and wall shear rate in a circular channel is described as follows:

p=

( (

4.n+2 ________ n

)( )
.
Q ______ w . h2

2.L.K . ______ h

w =

2.n+1 ________ n

)( . )
.
Q ______ w h2 m m

The new parameters are: w = channel width h = channel height

Pressure drop in ttings For calculation of pressure drop in ttings,e.g.valves, bends, expansions and tees, the following equation can be employed: p = Kf .v . _____
2

Q=

n ________ 3.n+1

. .r .
3

r. p _______ 2.L.K

1/n
with the parameters Kf = friction loss coefcient = density of uid kg/m3 v = velocity of uid m/s Values of the friction loss coefcient can be found in ordinary chemical or food engineering textbooks as well as in specialised rheological textbooks. For laminar ow, however, the data found are scarce and hence accurate estimation of pressure drop for typical liquid food ow conditions is difcult to make. Since the actual pressure drop is dependent on the type of uid as well as on the type and shape of the restriction and the friction loss, coefcients should therefore preferably be determined from experimental data. Reference 1) Reiner, M., The Deborah Number, Physics Today, January 1964, p. 62

or

p=

( (

3.n+1 ________ n

)( )
. Q ______ . r3 . Q ______ r3

n .L.K .2 _______ r

and

w =

.n+1 3 _______ n

) ( . )
m3/s m Pa m s-1 Pasn

The parameters are: Q = ow rate r = channel radius p = pressure drop

w
n K

L .

= = = =

channel length wall shear rate ow behaviour index consistency

The corresponding equations for rectangular channels are as follows:

Q=

n ________ . 4 n+2

. w. h .
2

h. p _______ 2.L.K

1/n

Dr Ulf Bolmstedt joined Alfa Laval in 1975 to implement dissertation work on computer software for simulation of evaporators. He continued with Alfa Laval and Tetra Pak, to work with evaporators, heat exchangers and rheology, mainly within liquid food applications. He has published a number of articles within these areas and holds patents, mainly within evaporation equipment. He currently holds a position within Tetra Pak as Company Specialist in Heat Transfer & Fluid Mechanics and at the Lund Institute of Technology as External Associate Professor

Get More
from your processing supplier
Youll need a very large bag when you come shopping for liquid food processing systems at Tetra Pak. This is because we are the only one-stop-shop for processing, packaging and distrib ution solutions for the food and beverage industries. We supply everything from single components, through modular systems to complete lines. You get expert support all the way from initial project evaluation to packaged product and technical service. As a single-source supplier, we maximise communication and minimise decision-making to ensure tight project control and completion on time, to agreed terms and performance. Our competence is reected by thousands of plants installed worldwide. Our global reach gives you long-term production security at the local level. Make Tetra Pak your partner and get more from your processing supplier.

Contact your nearest Tetra Pak company or visit www.tetrapak.com

CODE 63551 EN1