Theory and Praxis of Capillary Viscometry

- An Introduction – Authors:

Prof. Dr.-Ing. habil. Jürgen Wilke Hochschule Anhalt Food and biotechnology (Process and environmental Technology Faculty)

Dr.-Ing. Holger Kryk Magdeburg

Dr.-Ing. Jutta Hartmann Rheinfelden

Dieter Wagner SCHOTT-GERÄTE GmbH Viscometry development dept. Table of contents Page

1 Viscosity – Rheology ...... 2

2 Basics of capillary viscometry ...... 5 2.1 Measurement principle ...... 5 2.2 Designs of glass capillary viscometers ...... 5

3 Measurement of flow time ...... 7 3.1 Manual time measurement ...... 7 3.2 Automatic time measurement ...... 7 3.2.1 Tasks and particularities ...... 7 3.2.2 Detection of the meniscus passage ...... 7

4 Working equation of glass capillary viscometers ...... 9 4.1 Procedure for viscosity determination ...... 9 4.1.1 Neglect of HC correction ...... 9 4.1.2 Calculation of HC correction resp. use of given table values ...... 10 4.1.3 Experimental determination of the individual HC correction ...... 12

5 Calibration ...... 14

6 Handling of capillary viscometers ...... 15 6.1 General guidelines for the selection of the measurement system ...... 15 6.2 Cleaning of capillary viscometers ...... 16 6.3 Preparation of the measurement ...... 17 6.4 Performing the measurement ...... 19

7 Causes of errors and special corrections ...... 23 7.1 Correctable errors and corrections ...... 23 7.2 Uncorrectable errors ...... 24 7.3 Frequently occurring error symptoms, possible causes of errors, and ways of elimination ...... 26

8 Special applications ...... 28 8.1 Testing of plastics ...... 28 8.2 Determination of the viscosity of oils and additives .... 30 8.3 Testing of food ...... 31

9 Formula signs and units used ...... 33

10 Bibliography ...... 35

11 Standards used in capillary viscometry ...... 37

1 Viscosity - Rheology

Viscosity characterises the flow properties, the inher- The relationship between dynamic viscosity h and ent friction of liquids and gases. density r is referred to as kinematic viscosity n:

If a fluid is trapped between two plane-parallel plates, h n = = [m2 /]s (1.5) it will require some amount of force to displace the r upper plate.

The fluid particles which are directly adjacent to the For reasons of convenience, the unit of mm2/s is plates are firmly bonded to the surface by adhesion used which then numerically corresponds to the for- forces. In this process the fluid layer neighbouring the mer cSt (Centistoke) unit. plate being displaced adopts the velocity of the plate. All neighbouring layers stay more and more behind In case of Newtonian liquids h will remain invariant with the increasing distance to the plate being if the shear rate changes with all other test conditions moved. The cause for this phenomenon can be remaining unchanged. found in cohesion forces which counter-act the reciprocal dislocation of the individual layers. Moving a liquid molecule requires a potential hill to

y be surmounted which will lead to the following rela-

F tionship if Maxwellian Boltzmann velocity distribution is being applied: v

Evisk D = k × e RT (1.6) x

Figure 1 Basic model of the shearing operation in k Potentiality factor the case of laminar, stationary layer flow Evisk Measure of the height of the energy maximum (activation energy of viscous flow) The fluid starts to flow inside the gap. A layered flow R Gas constant builds up (please ref. to Figure 1). T absolute temperature t s The shear strain (also referred to as 1,2) refers As a consequence of the differences in size, shape, the quotient of force F and the boundary surface A of and interaction between the molecules, h may the liquid: change within very wide limits in the case of pure liquids. F J = (1.1) A Examples: The speed drop, i.e. the shear rate D, is the differ- n-pentane 0.230 mPa • s (20 °C) ential quotient: Water 1.002 mPa • s (20 °C) Propane triol 1480 mPa • s (20 °C) dv D = (1.2) (Glycerine) dy According to Newton's Viscosity Law there is pro- In the case of liquids, and in contrast to gases, h will portionality between the shear strain t and the shear decrease in a strongly exponential manner with rate D. rising temperatures. As a rule, the decrease will be the higher, the higher the absolute values of viscosity t h = • D (1.3) are and the lower the temperature is, since the inter- The proportionality factor h is referred to as dynamic molecular interactions are decreasing with the mag- viscosity coefficient or, in short, as dynamic vis- nifying thermal movement of the molecules. cosity.

The unit of measurement is Pa • s, with the indication This effect indicates the major practical significance being made in mPa • s i.e. in numerical conformity of viscosity, for instance, with regard to lubrication with the former unit cP (Centipoise): technology, as will be shown below.

J 2 D = = [Ns / m] = [Pa • s] (1.4) D

In the case of liquids a complex molecule structure Shear-rate dependent flow behaviour: and an increasing pressure lead to an increase in viscosity. Dilatancy The shear viscosity increases with rising shear rate (for As regards water, an anomaly occurs owing to the work hardening, please refer to Figure 2, curve b). particular structure. If pressure increases, viscosity will pass through a minimum, since molecule aggre- D gates are being formed the reciprocal friction of b which is lower.

In the case of liquid miscible phases h is in general a not made up by the addition of h-values of the pure components. c The viscosity of the miscible phase may be greater or smaller than h of the isolated components, or may be D in between. Figure 2 Viscosity curves of fluids The viscosity of the solutions of solid matters is a - Newtonian fluid frequently greater than the one of the pure solvent. b - Fluid with dilatant flow behaviour The indication is mostly given in terms of relative or c - Intrinsic viscous fluid specific viscosity (please refer to chapter 8).

A particular behaviour can be observed with the con- Plasticity centration-dependability of viscosity of electrolyte so- The flow of the liquids begins only from a minimum lutions. shear strain. Below this yielding point the substance behaves like a solid matter. If the liquid layers are moving at different velocities, the deformation of the ion cloud will cause the occur- Examples: rence of additional inter-ionic interacting forces which - Paints, varnish/lacquer will affect friction between the individual layers. - Food (mayonnaise) - Toothpaste H. Falkenhagen used the theory of inter-ionic interac- - Vaseline tions, applicable to highly diluted electrolyte, solutions to derive the Limit Law of Viscosity: BINGHAM substances: t = f (D) is linear above the yielding point. DD = + K c (1.7) C 0 CASSON substances: D t = f (D) is non-linear above the yielding point. c Viscosity at ion concentration c D 0 Viscosity of the pure solvent at same temperature Pseudo-plasticity (intrinsic viscosity) K Constant depending on These substances are characterised by Newtonian the following influencing variables: behaviour at low shear rates. - Temperature At high shear rates h will increase with the shear rate - Relative permittivity (please refer to Figure 2, curve c). - Ionic valence - Ionic mobility Examples: - Lacquer/varnish Non-Newtonian flow behaviour - Thermoplastics Disperse systems, concentrated polymer solutions, - Lubricating oils (multigrade oils) and melts of macro molecules show a marked non- - Glues Newtonian behaviour with increasing shear rates. - Additives In their case there is a non-linear dependency between shear strain and shear rate.

In addition to these effects a shear-time dependent The complex nature of this field of work has lead to flow behaviour can be observed with some non- the crystallisation of an original term, i.e. rheology Newtonian matters: (science of flow behaviour). t = f (D, t) Rheometry deals with the specific methods and pro- h = f (D, t) cedures of determining rheological characteristics. This means that shear viscosity is influenced by the Within this nomenclature viscometry is a partial dis- duration of the shearing action (please refer to Figure cipline of rheometry. 3). Principles of viscosity measurement h Rheological measurement procedures are mainly based on mechanical methods, since tension and b elongation are mechanical values which are determined on the basis of a defined deformation of the a sample.

The simultaneous measurement of the electrical, c magnetic, and optical properties which may change during the deformation or flow process of the fluids is becoming more and more interesting.

ts Figure 4 shows the major manners of realising the Figure 3 Dependency of shear viscosity deformation of the sample. on the shear time 2 a = shear-time independent flow behaviour 4 b = Rheopexy M1 c = Thixotropy 2 3 5 The following distinction is made: 1 6 2 Thixotropy M2 Shear viscosity decreases at constant shear rate with v increasing shear time (typical for sol/gel transforma- v tion). a bc Rheopexy Figure 4 Measurement principles of viscometers Shear viscosity increases at constant shear rate with a = Capillary viscometer increasing shear time. b = Rotational viscometer c = Falling-ball viscometer Rheopexy can, for instance, be seen with PVC plas- 1 = Capillary 5 = Measurement ball tisols. They are used for corrosion protection on met- 2 = Sample 6 = Glass cylinder als. If the coating rate is increased the material be- 3 = Coaxial cylinder M1, M2 = Measurement marks comes more thick-flowing. Rheopex liquids are char- 4 = Torque sensor acterised by a gradual structure formation under The present brochure covers the methodological and shearing strain. metrological particularities of low-pressure capillary viscometers, the most important of which, in turn, In addition to these viscous properties one can ob- are the glass capillary viscometers. serve the occurrence of elasticities (1st and 2nd normal-stress difference) acting perpendicularly to They are in particular suited for viscosity measure- the flow direction. ments with Newtonian liquids with a kinematic viscosity of more than 0.3 mm2/s. The combination of viscous and elastic behaviour leads to the description of viscoelastic fluids. Poly- Perfection in the manufacture and the sophisticated mer solutions, and recently also biopolymers exhibit- quality-assurance methods form the basis of stan- ing molecular-structure dependent viscoelastic prop- dardised measurement systems which are meeting erties of this kind meet with more and more techno- today highest accuracy requirements as to reproduc- logical interest, e.g. in the production of paints and tion incertainties and absolute measurement incer- coatings, food, cosmetics, and pharmaceutics. tainty.

2 Basics of capillary viscometry

2.1 Measurement principle The first measurement principle can be used for the design of continuos viscometers the measurement Inside the capillary viscometers, the velocity drop re- accuracy of which is depending on the achievable quired for viscosity measurement is built up in the measurement incertainty in differential-pressure form of a laminar tube flow within a measurement measurement and the stabilisation of a defined vol- capillary. ume flow. Under idealised conditions This issue is approached in a satisfactory manner the laminar, isothermal flow design of device in the form of comparison meas- stationary flow condition urement methods. Newtonian flow behaviour of the liquid pressure-independence of viscosity An application of this can be found in solution vis- incompressibility of the liquid cometry where the viscosity of the pure solvent is wall adherence of the liquid used as a reference liquid. The measurement itself is neglect of the flow influences at the entry and made, inter alia, on the basis of a ”pneumatic Wheat- exit of capillary of sufficient length stone bridge”. the liquid flows in coaxial layers towards the pressure drop through the capillary. A parabolic velocity flow Another application of the first measurement principle occurs (please refer to Figure 5). is viscosity measurement on plastics melts. This process involves short capillaries, frequently gaps of a predefined geometry (high-pressure capillary viscometry). v R r max v 2.2 Designs of capillary viscometers

In the case of low-pressure capillary viscometers the imaging signal used for viscosity is the time required by a defined liquid volume to flow through a v = 0 measurement capillary.

Figure 5 Velocity profile with laminar tube flow The driving force is the hydrostatic pressure of the The Hagen-Poiseuille Law is the physical basis of liquid column. To achieve higher shear rates, it is viscometers working according to the capillary princi- possible to use over-pressure. ple /1, 2, 3, 4/: Irrespective of the specific design, the mostly V FR4 ,p U-shaped glass bodies have ball-shaped extensions = (2.1) t 8LD the volume of which determines the quantity of the sample. With regard to viscosity measurement, this results in two different fundamental measurement principles: Measurement marks on the glass body, or accurately Measurement of the differential pressure at a con- defined fixed sensors, allow the measurement of the stant volume flow of the sample through the cap- passage time of the boundary layer between the illary sample and the air (meniscus), a process which enables the passage time of a product volume re- Measurement of the volume flow through the cap- stricted in such a manner to be measured with illary at a given differential pressure. measurement incertainties < 1/10 s.

Figure 6 shows the two fundamentally different vis- In this way the hydrostatic pressure of the liquid col- cometer types after OSTWALD and UBBELOHDE. umn is independent of the sample quantity being filled in. 2 3 In addition, owing to the geometrical shaping of the levelling bulb (6), the influence of surface tension on the measurement result is almost eliminated.

In the case of the UBBELOHDE Viscometer, too, the measurement is aimed at the time required by the liquid meniscus to sink from the annular measure- M 1 ment mark M down to the annular measurement 8 1 mark M2. M2 7 In the case of very strongly tinted, opaque liquids, it L can be possible that a visual detection of the menis- hm cus passage through the measurement marks is im- 10 possible owing to the wetting of the tube. For manual 6 L operation, the Reverse-Flow Viscometer (please refer to Figure 7) is used in such cases. 7 4 a) b)

Figure 6 Glass capillary viscometers after a) UBBELOHDE and b) OSTWALD With both viscometers the liquid being examined is filled through the filling tube (3) into the storage con- tainer (4).

Considering that the mean pressure height in the h 3 = M1 m 2 case of the OSTWALD Viscometers depends on the 5 = M2 filling height, the prescribed measurement volumes h 7 = M3 m 1 L have to be observed under any circumstances. For this reason filling is done using a pipette. To perform the measurement, the sample is sucked into the tube (2). The measurement aims at the time the meniscus requires to sink from measurement mark M1 to measurement mark M2 (annular measurement marks). Figure 7 CANNON-FENSKE In the case of the UBBELOHDE viscometers the Reverse-Flow Viscometer transition point from the capillary (7) to the levelling bulb (6) has the shape of a ball joint being the end The sample is filled into the spherical extension of point of an additional venting tube (1) /32, 33/. After the capillary tube (2). The tube (1) is closed during filling the sample through the tube (3) into the con- thermostatisation and opened at the beginning of the tainer (4), the venting tube is closed. measurement. The imaging signal used for viscosity Depending on the operational mode, i.e. pressing or is the time required by the meniscus to flow through sucking action, the sample is filled by over-pressure the measurement marks M1, M2 and M3 at the re- applied to tube (3) or by suction via the tube (2) into verse-flow (1). the reference level vessel (6), the capillaries (6), the The standard viscometer introduced was the measuring sphere (8), and at least up to half of the CANNON-Master instrument with a capillary di- pre-run sphere (9). ameter of 0.45 mm and a capillary length of 400 mm. After venting tube (1), the liquid column in the level- With the determination of the viscosity of water ling bulb breaks off. At the exit of the capillary the so- h = 1.0019 [cP] ± 0.0003 [cP] 1) (20 °C), called suspended level develops (also refer to Fig- it was possible to define a viscosity scale. ure 22). For this reason only a limited sample quantity - max., min. filling marks (10) - may be filled in. After The capillaries of viscometers used for industrial ap- ventilating tube (2) the sample flowing out of the cap- plications are usually shorter (70 - 250 mm). illary will flow off along the inner wall of the levelling bulb (6) in the form of a film.

6 ______1) National Bureau of Standards, USA, 1953

3 Measurement of the flow time

3.1 Manual timing 3.2.2 Detection of the meniscus passage This task requires the use of sensors responding to In the most simple case the flow time is taken by an the difference between the material properties of the operator using a stop watch. Glass viscometers air and the product being analysed during the pas- manufactured for this purpose have annular meas- sage of the meniscus through the measurement urement marks burnt in above and below the meas- marks. urement sphere (please refer to Figures 6, 7). Optical sensors The disadvantages of this method are obvious: During the meniscus passage the optical conditions Subjective observation errors or differences in the such as refraction and reflection within the detection reaction time of the operator at the beginning and plane are changing. This leads to a change n the ra- end of the timing lead to increasing reproducibility diation intensity of the light arriving from the transmit- incertainties and, under certain circumstances, to ter at the receiver (please refer to Figure 8). For the systematic errors. measurement of time, for instance, the analogous In the case of opaque substances the meniscus signal provided by a photo diode is transformed into cannot be seen. One has resort to Reverse-Flow a pulse used for the start and stop of the time meas- Viscometers with their more awkward handling and urement. Specific threshold values of the analogous reduced accuracy. signal may be defined for the "filled" or "empty" status. 3.2 Automatic timing Advantage: Versatile application, simple set-up 3.2.1 Tasks and particularities Disadvantage: Highly tinted or opaque liquids, espe- In the case of automatic capillary viscometers an cially those which adhere strongly to electric signal has to be generated during the pas- the wall, cannot me measured. sage of the air/sample or sample/air boundary layer, On the viscometers from SCHOTT-GERÄTE all opti- respectively, through the measurement marks. This cal sensors are accommodated in a measurement electrical signal is required as tripod made of metal or plastic. Within the tripod the a start and stop signal for the timing process fixation rack and the glass viscometer are fastened and as using a clamping mechanism. Figure 8 shows the ar- a status signal for the automatic operation rangement of the optical sensors within the meas- (filling, emptying of the capillaries). urement tripod on the viscometer. The light is guided out of the tripod head through fi- The detection and transformation of a time signal bre optics into the tripod legs up to the upper and does not pose any metrological problems. In practical lower measurement plane. The watertight sealing viscosity measurement the measurement incertain- enables the measurement tripods to be placed in liq- ties are determined by the fluid-dynamic circum- uid thermostats. stances and the detection of the meniscus passage Owing to high precision in the glass-technological through the measurement marks. and mechanical production as well as through meas- The manufacturer of the measurement device has to ures of quality assurance it is ensured that the glass ensure by design and production measures that the bodies and tripods are freely interchangeable, with viscometer constant will not change even if the the certified viscometer constants remaining valid. measurement conditions should deviate from the calibration conditions (e.g. measurement and calibration temperature). As a result, there would be incidental errors which would have to be determined and identified for each device separately. Otherwise the user himself would 1 2 have to perform calibration. And this is the point where low-pressure capillary viscometry has a decisive advantage over other viscosity measurement procedures.

The well-adapted selection of materials, the engi- Figure 8 Arrangement of the optical sensors neering-technological mastery of the production on the viscometer processes, and the sophisticated methods of quality 1 = Optical fibre input assurance enable a calibration of the viscometers to 2 = Optical fibre output be made.

______7 1) National Bureau of Standards, USA, 1953

Conductivity sensors Electrolytically conductive measurement liquids (solu- U[V] tions of salts, acids, bases) can be detected using 14 b small-sized electrodes melted into the measurement 12 plane in the glass wall. For signal generation the electrical resistance is measured. 10 Advantage: Simple set-up; detection of tinted 8 and opaque liquids 6 Disadvantage: The sample must be electrocon- ductive; the supply lines to the sen- 4 a sors are to be protected against wa- 2 ter penetration if liquid thermostats 0 1 23 4 are being used. s t [s]

Thermal-conductivity sensors Figure 10 TC sensor signal Small-sized thermistors (NTC resistors), melted in on a during filling and b during emptying the level of the measurement plane, are heated up. S - switch point of the timer device As a result to the improved thermal conductivity of the liquid the thermistor will cool down at the Ultrasonic sensors air/sample transition, and its electrical resistance will The propagation of sound waves in the frequency diminish. range > 20 kHz is different in gases and liquids, and Advantage: Measurement-signal generation is inde- owing to the changing sound impedance (product of pendent of the tint, transparency, and conductivity of sonic speed and specific weight) the waves are re- the product being analysed. flected from boundary layers. Disadvantage: More demanding production owing In the case of the echo process (reflection) a sound to the required melting-in of the sensors; incrustation head, attached to one side of the measurement mark and contamination hazard in the case of thermally and acting both as emitter and receiver, detects decomposable samples. whether gas or liquid is present in the measurement Figure 9 shows a TC Viscometer from SCHOTT- plane. GERÄTE. In the tube axis the thermistors with a di- The radiation process uses separate emitting and re- ameter of < 1 mm in the sealed-in head portion are ceiving modulators located at opposite tube posi- clearly visible. tions.

Advantage: The signal formation is independent of other sample properties, i.e. the application of the process is versa- upper NTC sensor tile; no sealing in the glass required Disadvantage: Coupling of the sound heads bears production-technological difficulties, especially in the case of an application in liquid thermostats; greater lower NTC sensor signal-processing efforts required; higher price

Figure 9 TC Viscometer from SCHOTT-GERÄTE Gas-ionisation spark-discharge detection The electrodes melted in on the level of the detection The essential factor for safe operation is a good dy- planes are connected to a high-voltage generator. If namic behaviour. Figure 10 shows the signal course the liquid, acting as an electrical insulator, uncovers resulting developing during filling and run-off (meas- the electrodes a spark discharge will occur in the gas urement process) through the changing thermal con- chamber if a sufficiently high breakdown voltage is ductivity in the surrounding of the sensor. selected. The electrical pulse is used as a control To compensate the influence of the sample on dy- signal. namics, the SCHOTT-GERÄTE viscosity measure- Advantage: Detection is possible in dull, opaque ment devices perform an automatic calibration. The liquids working point of the start/stop timing is adaptively set Disadvantage: The process cannot be used in the by the device software during the filling process of presence of least traces of water in the capillaries on the basis of a respectively deter- the product being analysed (water mined dynamic ID value. contents > 0.5 %); high-voltage requires extensive insulation.

4 Working equation of glass capillary viscometers

In the metrological sense, the working equation The basic hydrodynamic process was first examined represents the statistical characteristic of capillary by Hagenbach /5/ and Couette /6/. viscometers. The user uses them for the determina- The difference between the measured and theoreti- tion of viscosity on the basis of the flow time. cal flow time tH is therefore referred to as Hagen- The starting point is formed by the flow model in the bach-Couette Correction Time (or, in short, HC form of the Hagen-Poiseuille Law (equation 2.1). The correction or Hagenbach correction): driving force is the hydrostatic pressure of the liquid t = t - t (4.3) column in the form of the mean pressure height hm H g (please refer to Figures 6, 7). This results in the following corrected working equa- Considering that the volume flow V is recorded via tion for glass capillary viscometers: the measurement of the flow time t, the following n equation results for kinematic viscosity n: = K • (tg - tH) (4.4)

h p R4 g h The smaller the flow time is, the greater becomes the n = = m t (4.1) r 8 L V Hagenbach-Couette Correction Time. Curve b in Figure 11 shows the real course of the characteristic. In addition to the flow time, equation (4.1) contains only constants and geometric details. In practical viscosity measurement there are in prin- For a given viscometer they can be summarised into ciple three ways to take into account the one characteristic magnitude, the so-called viscome- Hagenbach-Couette Correction and thus to deter- ter constant K: mine the kinematic viscosity of the product being analysed. n = K • t (4.2) In order to take into account the tolerances which are inevitable in the manufacture of the devices, K is de- 4.1 Methods of viscosity determination termined for each individual viscometer by way of a 4.1.1. Neglect of HC correction calibration (please refer to Chapter 5). The selection of a capillary with a small diameter, According to equation (4.2) there is a linear correla- adapted to the viscosity of the product being ana- tion between kinematic viscosity and flow time. Fig- lysed, involves long flow times. In this case HC cor- ure 11 shows this correlation in the form of a charac- rection takes such a small value that a correction teristic (curve a). may be omitted within the framework of the required accuracy. n The flow times to be observed if HC correction is neglected in order not to exceed a relative error e can be calculated according to equation (4.5) or equation (4.6), respectively:

a 1 b n æ m V ö 2 t ³ 19.95 ç ÷ (4.5) g è Î L Kø

1 t tg t 1 1 æ Vö 2 - - ³ ç ÷ e 2 3 Figure 11 a ideal and tg 4.9 ( L) R (4.6) è Kø b real viscometer characteristic

When applying the flow model in the form of the m = empirical coefficient of HC correction Hagen-Poiseuille Law, additional pressure losses oc- m = 1.12 (Re > 100) /N10/ curring at the capillary ends are not taken into account. Owing to the finite capillary length, however, Equation (4.5) is applicable to viscometers with the pressure losses occurring at the in- and outflow sharp-edged capillary ends. When using viscometers affect measurement accuracy. As a consequence of with funnel-shaped capillary ends, equation (4.6) these additional pressure losses the measured flow should be used /N10/. time tg is greater than the time t resulting from Hagen-Poiseuille Law.

4.1.2. Calculation of HC correction resp. use of given table values L l The manufacturer calculates HC correction times on e the basis of the geometrical dimensions as a function of the flow time and states them in the device de- scriptions. Understanding the calculation algorithm requires first an explanation of the theoretical basics of Hagen- ,p bach-Couette Correction. Figure 12 shows the true march of pressure in the capillary /7/.

The deviations from the ideal march result from hy- l drodynamic processes in the in- and outflow zone of the capillary. They are taken into account in the flow Figure 12 Axial march of pressure in the capillary model (please refer to Figure 13) in the form of additional terms.

Hagen-Poiseuille Law Hagenbach-Couette Correction

Viscous portion Pressure loss owing to the Pressure loss owing to Pressure loss owing increase in the kinematic the formation of the para- to the increased wall energy of the liquid when bolic velocity profile in the friction inside the flowing into the capillary flow path Ie flow-in path Ie

D r 2 H D 8 V L V D p = + + + pC (4.7) F R 4 2 2

Figure 13 Flow model with correction terms

The mean flow rate v in the capillary results from: This correlation was confirmed by Kerstin, Solokov, and Wakeham /8/ by a numerical solution of the Na- V vier Stokes' equations. v = (4.8) R 2 F , p F R 4 m H V D = - (4.10) 8 V (L + n R) 8 F (L + n R) In this way the following corrected Hagen-Poiseuille Law for the determination of viscosity results from An explicit determination of the Couette correction equation (4.7), Figure 13: poses problems in terms of metrology. However, since the viscometer constant K is determined by , p F R 4 H V , p F R4 D = - - c (4.9) way of calibration, Couette correction is implicitly 88 V L 8 F L V L taken into account in the form of a mean value. Cou- ette correction within the Hagenbach correction term Couette did already take into account the pressure of equation (4.10) is considered concurrently in the loss Dpc by way of adding a fictitious length n · R to form of the empirically determined parameter m. the capillary length L in equation (4.10). Therefore this correction is often briefly referred to as Hagenbach correction in literature /12, N6 ... N10/.

For a given glass capillary viscometer For Re > 100 the value as calculated was confirmed in experiments. In the case of Re numbers below 100 D r p = g hm (4.11) m will drop sharply and retain only approx. 30 - 40 % of its initial value at Re = 25/12/. With Re < 10, m is and so small that it can be neglected /13/.

V If the capillary ends are funnel-shaped, m will be a V = , (4.12) t g function of the Reynolds number all across the me- trologically utilised flow-time range. with equation (4.13) being derived as working equation. Cannon, Manning, and Bell /14/ arrived at the following functional correlation:

p R4 g h m V n = m t - (4.13) m = 0.037 Re (4.15) g p 8 V L 8 L t g Equation (4.15) forms the basis of the calculation of Parameter m mainly depends on the shape of the HC correction according to the applicable standards capillary ends and the Reynolds number (Re). /N6, N8, N10/.

The Reynolds number is an important non- In this way the following working equations result for dimensional similitude characteristic for fluidic de- viscometers with sharp-edged or funnel-shaped cap- scription of incompressible fluids: illary ends:

V v r 2 V sharp-edged capillary ends Re = = (4.14) hpn R t g B n = K × t - It characterises the flow shape, i.e. laminar or turbu- tg lent, conditioned by inertia and friction (viscosity). (4.16) 1.12 V Depending on the production technology the capillary B = ends of viscometers may be sharp-edged or funnel- 8 F L shaped (please refer to Figure 14). UBBELOHDE Viscometer: B = 2.5 /N6/

B tH = (4.17) K t g

funnel-shaped capillary ends If Hagenbach-Couette correction in the form of a time correction according to equation (4.4) is used, the HC correction time is calculated as follows:

E n = K × t - 2 tg a b 3 (4.18) 1.66 V 2 E = Figure 14 Capillary ends of viscometers 1 a - sharp-edged b - funnel-shaped L (2 K R)2

With regard to sharp-edged capillary ends a constant (4.19) value of m = 1.12 was calculated on a theoretical ba- E tH = sis /9, 10, 11/. This value is also contained as a 2 K tg maximum guidance value in /N10/. For reasons of production technology, however, ideally sharply cut capillary ends are not realisable.

The E / K correction terms for UBBELOHDE and 2. Determination of the Hagenbach correction tH Micro UBBELOHDE Viscometers can also be taken for the flow time tg by way of linear interpolation from the relevant DIN standards /N6, N7/. between the values t und t : H1 H 2 For reasons of production technology, capillary viscometers from SCHOTT-GERÄTE have funnel- æ 1 1 ö t = t - K ç - ÷ (4.21) shaped capillary ends. The correction times tH are H 2 H 1 12 ç ÷ è tg tg ø given in the operation instructions. 2 1

tH - tH K = 1 2 (4.22) 4.1.3. Experimental determination of the 12 1 1 individual HC correction - t t In the case of small flow times HC correction will g 1 g 2 have a increased influence on the measurement result. In addition, owing to after-flow effects of the liq- Figure 15 illustrates the correction procedure: uid and the beginning of the deformation of the suspended level, the viscometer characteristic of t UBBELOHDE Viscometers is affected. H c

If falling short of the measurement range as recom- t H 2 mended in the operating instructions is inevitable, an individual HC correction for the respective viscometer t H has to be determined in experiments. t H 1 a To do so, two standard liquids of a known viscosity are to be used, with the viscosity of the product being analysed lying between the viscosities of the stan- b dard liquids. The smaller the difference between the viscosities, the more accurate the result of the correction procedure.

Realisation of the correction procedure: 1. Determination of individual values for the 1/t

Hagenbach correction with the standard liquids: Figure 15 individual Hagenbach correction /N9/ n a Hagenbach curve according to equation (4.19) i b real course of the individual Hagenbach- tH = tg - i = {1;2} (4.20) i i K correction c interpolation straight line

Examples of viscosity determination Viscosity measurement of n-decane at J = 23 °C (n » 1.21 mm2/s) with UBBELOHDE Viscometers

1. Case 3. Case Selection of a viscometer with capillary 0 Selection of a viscometer with capillary Ic K = 0.00098 mm2/s2 K = 0.0303 mm2/s2 L = 90 mm V = 5.7 ml L = 90 mm V = 5.7 ml hm = 130 mm D = 0.36 mm hm = 130 mm D = 0.84 mm

Measurement range according to Measurement range according to operating instructions: operating instructions: 0.2 ... 1.2 mm2/s 3 ... 30 mm2/s mean measured flow time: tg = 1234.57 s mean measured flow time: tg = 39.95 s

HC correction time is approx. 0.3 s. This corresponds Calculated HC correction to approx. 0.024% of the flow time. This means that according to equation (4.19): neglecting the HC correction time would not cause tH = 1.03 s any significant change of the measurement result. The measurement range of the viscometer was fallen Calculation of viscosity: short of. Furthermore, the HC correction time is n = K · tg above the max. correction time of tH = 0.66 s as indi- n = 0.00098 mm2/s2 × 1234.57 s cated in /16/ for precision measurements. = 1.21 mm2/s In this case a viscometer with a smaller capillary diameter should be resorted to. If this is impossible, 2. Case the individual HC correction time for precision measurements has to be determined in an experimental Selection of a viscometer with capillary I manner. K = 0.0105 mm2/s2 L = 90 mm V = 5.7 ml hm = 130 mm D = 0.84 mm

Measurement range according to operating instructions: 1.2 ... 10 mm2/s mean measured flow time: tg = 116.05 s

Calculated HC correction time according to equation (4.19): tH = 0.69 s

Calculation of viscosity: n = K · (tg - tH) n = 0.0105 mm2/s2 × (116.05 - 0.69) s = 1.211 mm2/s

5 Calibration

The viscometer constant K is determined individually The constants of the test specimens are determined for each glass capillary viscometer by way of calibra- on the basis of kinematic viscosity of the test liquid tion. and the flow time (please refer to Figure 16).

By careful calibration in combination with the use of To ensure a high statistical certainty, two measure- high-quality measurement and testing means and ment cycles involving seven flow-time measurements close-tolerance reference standard sources the each are run, with the first measurement of the re- manufacturer guarantees a reproducible calibration spective measurement cycle being considered as of highest precision. Measurement and reproducibil- preliminary test. ity incertainties of calibration have a direct influence The measurement temperature is 23 °C ± 0.01 K. It on the measurement incertainty of the viscometers. is verified using at least two officially gauged mercury capillary-column thermometers with a resolution of Measurement principle 0.01 K. The determination of the constants is done by a simultaneous flow-time measurement in the viscome- Each calibration can guarantee the metrological cor- ters to be calibrated (test specimens) and in the ref- rectness of the viscometer constants only for a lim- erence standard sources the constants of which were ited period of time. It is therefore recommended to determined by the "Physikalisch-Technische Bunde- check the constants on a regular basis or to have sanstalt (PTB)" (Federal Physico-Technical Institute) them checked by the manufacturer, respectively. The in Brunswick. check may be done either by comparison measurements using reference standard sources (please see Realisation above) or with calibrating oils from the ”Deutsche Ka- In a thermostat bath with a constant temperature of librierdienst (DFD)” (German Calibration Service). ± 0.01 K the flow time of a test liquid through a multi- However, if regular oils are being used, the limitation tude of glass capillary viscometers is measured. of the accuracy of the test procedure caused by the incertainty of the regular-oil viscosity indication Test liquids are no reference standard sources. Their should be noted. Considering that this incertainty is in viscosity is only known within a tolerance range of general above the measurement incertainties stated ± 10 % around a guidance value. The test liquids for glass capillary viscometers, this calibration used are mono-substances or mineral-oil products method is not recommended for precision measure- with narrow boiling profiles. ments.

Two of the viscometers are reference standard Please refer also to DIN 51 561 - 4, Part 4: Viscome- sources the flow times of which is used to calculate ter calibration and determination of measurement in- the kinematic viscosity of the test liquid. Owing to the certainty, taking into account the user note /N9/. use of two Reference Viscometers a functional test is carried out automatically during calibration.

Test specimens Reference Viscometer

P1 P2 P3 R1 R2

n n 1 = K R 1 (t g R1 - tH R1 ) 2 = KR2 (tgR2 - tHR2 )

n + n n n = 1 2 KP1 = 2 t gP1 - tHP1

Figure 16 Calibration of glass capillary viscometers

6 Handling of glass capillary viscometers

6.1 General guidelines on the selection of the measurement system

Selection of the viscometer type The following viscometers from SCHOTT-GERÄTE For automatic viscosity determination, to be used in can be used for viscosity measurement with trans- particular with opaque oils and emulsions, parent liquids: TC-UBBELOHDE Viscometers are the choice. Ow- UBBELOHDE Viscometer ing to the fact that the thermistor sensors are glass- OSTWALD Viscometer sealed and melted hermetically tight in the viscome- CANNON-FENSKE-Routine Viscometer ters, it is, for instance, also possible to measure conductive and highly aggressive liquids. These include devices for both manual or automatic measurements involving optoelectronic detection of the meniscus passage. In addition it is possible to Capillary selection use TC-UBBELOHDE Viscometers equipped with The measurement range of the viscometers is de- thermistor sensors. Owing to the advantages re- termined by the capillary diameter (0.25 ... 10 mm). ferred to in chapter 2, UBBELOHDE Viscometers Each capillary diameter has a capillary number and a should be preferred over the other types in most ap- viscometer type number assigned which is indicated plications. on a test certificate.

In the case of measurements involving low-foaming To select a viscometer, the viscosity of the substance or bubbling liquids one should use OSTWALD or TC- to be analysed has to be estimated. UBBELOHDE Viscometers, since foam of bubbles affect the functioning of the photoelectric barriers. In The selection as such is based on a rough calcula- the case of highly foaming liquids, however, TC- tion of the flow time exclusive of the HC correction UBBELOHDE Viscometers should not be used, since according to equation (4.2). the thermistors' function may be affected by adhering foam particles. In addition, no clear detection of the In accordance with the DIN standard, the min. flow meniscus passage is possible in the presence of in- time to be sought after should be 200 s /N10/ for tense formation of foam. most viscometers. However, trials have shown that it is also possible to realise shorter flow times without For determining the viscosity of mixed substances impairing the measurement accuracy. containing highly volatile components and matters reacting with the ambient air, the use of OSTWALD When using micro viscometers the flow time can be or CANNON-FENSKE Routine Viscometer is rec- reduced to 30 s. According to the most recent re- ommended. search results, even flow times down to approx. 10 s /15, 34, 35/ are possible if individual Hagenbach- If only sample or solvent small quantities are avail- Couette Correction with automatic flow-time meas- able, the use of Micro UBBELOHDE or Micro urement is applied. OSTWALD Viscometers is favourable. In the operating instructions of the viscometers the For reason of thermally caused volume changes of min. flow times are stated as a function of the capil- the product being analysed, high- or low-temperature laries. measurements should always be performed using UBBELOHDE Viscometers. Table 1 shows as an example the measurement ranges as a function of the capillary diameter for CANNON-FENSKE Reverse-Flow Viscometers for UBBELOHDE Viscometers. manual measurement of the viscosity of opaque liquids, or BS/IP/RF U-Tube Reverse Flow Viscometers (from approx. 6000 mm2/s) for highly viscous substances are available.

Table 1 Measurement ranges of UBBELOHDE Viscometers /16/

Capillary Capillary diameter K (guidance value) Measurement range no. [mm] [mm2/s2] [mm2/s] 0 0.36 0.001 0.2 ... 1.2 0c 0.46 0.003 0.5 ... 3 0a 0.53 0.005 0.8 ... 5 I 0.63 0.01 1.2 ... 10 Ic 0.84 0.03 3 ... 30 Ia 0.95 0.05 5 ... 50 II 1.13 0.1 10 ... 100 In addition, SCHOTT-GERÄTE IIc 1.50 0.3 30 ... 300 offers a KPG utility pipette for determining the optimally suited IIa 1.69 0.5 50 ... 500 capillary number for the respec- III 2.01 1 100 ... 1000 tive measurement task. IIIc 2.65 3 300 ... 3000 IIIa 3.00 5 500 ... 5000 IV 3.60 10 1000 ... 10000 IVc 4.70 30 3000 ... 30000 IVa 5.34 50 6000 ... 30000 V 6.40 100 > 10000

6.2 Cleaning of capillary viscometers

Careful cleaning of viscometers is an essential pre- Initial cleaning requisite for an exact and reproducible measurement Especially as a result of transportation and storage, value. Practical experience has shown that increased severe contamination may occur so that a thorough scattering of the flow times is in most cases caused initial cleaning is inevitable. by contamination. In this context even smallest quantities of microscopically small particles of dust within the viscometer may lead to standard deviations of up The following cleaning agents have proven to several per cent. to be suitable: Particles which adhere firmly to the capillary wall and concentrated sulphuric acid with an addition of po- are frequently almost invisible are often the cause of tassium dichromate (chromic-sulphuric acid mix- systematic measurement errors. Errors of this type, ture); when working with chromic-sulphuric acid leading to an increase of the flow times, can hardly mixture, extreme care has to be taken; chromium- be told from the individual values of a measurement (VI) compounds are toxic series. The larger the capillary diameter, the smaller a solution consisting of 15 % hydrochloric acid and is the danger of contamination. 15 % hydrogen peroxide In addition to solid particles, oil or fat films adhering to the internal wall of the viscometer may affect the Cleaning methods: flow times. In particular when measuring substances 1. Fill the viscometer completely with one of the with a higher surface tension (e.g. aqueous media) above cleaning substances droplets, adhering to the wall and affecting the 2. Let the cleaning substance act for at least measurement result, may occur during the start-up 12 hours process. This is why it is recommendable to measure 3. Rinse the viscometer using distilled water only substances with similar properties in one and 4. Rinse with a filtered, miscible, highly volatile the same viscometer. If this is impossible, a particu- solvent, e.g. with acetone larly careful cleaning process has to be carried out. 5. Dry by way of purging with dry, dust-free air or in a As a principle, all cleaning agents should be filtered drying cabinet prior to use using glass frits with a corresponding pore width. Paper filters have a tendency of losing fi- bres and are thus not recommendable. They use of highly alkaline solvents may lead to irre- versible leaching in the glasses which may even cause a change in the viscometer constant.

Initial cleaning Automatic cleaning Immediately after each measurement, the viscometer Especially for examinations of mineral oils in has to be cleaned using suitable solvents. The use of UBBELOHDE or CANNON-FENSKE Routine Vis- a vacuum pump has proven suitable for this purpose. cometers, SCHOTT-GERÄTE is offering the AVS 26 Viscometer Cleaner. Using this device it is possible Cleaning method when using a vacuum pump: to clean viscometers without having to take them out 1. Connect the vacuum pump via a liquid trap to the of the thermostat baths. This process requires spe- capillary tube cial viscometers with an attached rinsing tube. 2. Fill the cleaning liquid into the filling tube and the venting tube The AVS 26 Viscometer Cleaner works in combina- (in the case of UBBELOHDE Viscometer) tion with the automatic viscosity-measurement de- 3. Periodically close the filling and the venting tube vices of the AVS series. Several rinsing programs while the liquid is being sucked off are available. During the rinsing process the vis- A pulsating liquid column will occur, dissolving cometer cleaner pumps solvent alternately through even set-in contamination all tubes of the viscometer. The device is intended for 4. Repeat the cleaning process two or three times use with up to two solvents. The rinsing process may 5. Rinse with a highly volatile solvent be followed by a drying cycle. For automatic clean- 6. Dry by way of sucking dry, dust-free air ing, the maximum viscosity limit of the product being through the assembly analysed is approx. 8000 mm2/s at 25 °C. Cleaning method without a vacuum pump: 1. Fill the cleaning liquid into the filling tube The use of an automatic rinser, however, does not 2. Suck the liquid several times into release the user from a periodical, careful manual the measurements sphere cleaning. 3. Clean the remaining viscometer parts by shaking the viscometer 4. Empty the viscometer 5. Repeat the cleaning process two to three times 6.3 Preparation of the measurement 6. Rinse using a filtered, highly volatile solvent 7. Dry by purging with a dry, dust-free air Preparation of the sample or in the drier Solid particles contained in the sample to be exam- In particular when cleaning without a vacuum pump, ined have a similar effect on the measurement result it is furthermore recommended to wait for an addi- as contamination in the viscometer tional 20 to 30 minutes prior to the beginning of the cleaning cycle. If measurements are not made im- For this reason, you should immediately prior to per- mediately subsequent one to another, the cleaned forming the measurement: viscometers are to be stored in a dust-free environ- Carefully clean and dry all parts coming in contact ment. Immediately prior to the next measurements, with the substance to be measured, the glass body is to be rinsed and dried once again. filter the samples - low-viscosity samples: If the viscometer was not in use for several weeks, glass filter, porosity 2 to 4 (10 - 100 mm) cleaning should be done using one of the substances - highly viscous samples: suitable for initial cleaning after an action time of at sieve, mesh width 0.3 mm. least one hour. The same cleaning process should also be performed if scattering of the measurements values above the repeatability limit specified for the Paraffin or resin-containing products as well as sub- viscometer, or systematic measurement errors, occur stances with a pour-point of less than 30°C below the during operation, with such errors not being elimi- testing temperature are to be treated thermally prior nated by cleaning using one of the correspondent to performing the measurement. The measurement solvents. temperature must be at least 20°C higher than the pour-point. In order to minimise the likelihood of the occurrence of such errors from the onset, regular cleaning of the viscometers using the liquid specified for initial cleaning is also recommended at larger timely intervals.

Filling UBBELOHDE and OSTWALD Filling CANNON-FENSKE Routine Viscometers Viscometers CANNON-FENSKE Routine Viscometers (please re- The substance to be examined is to be filled into the fer to Figure 17) are held upside down for filling. The liquid reservoir via the filling tube. capillary tube (1) immerses into the liquid to be measured, while suction is upheld at the other tube Considering that the average pressure height of the until the liquid has reached the annular mark M2. Af- OSTWALD Viscometer depends on the filling quan- ter filling, the viscometer is restored to normal meas- tity, the sample volumes for OSTWALD and Micro uring position. OSTWALD Viscometers indicated in table 2 are to be adhered to in any case. For this reason, a pipette is Considering that filling a Reverse Flow Viscometer is to be used for filling. somewhat more complex, reference is made at this point to the standards /N5, N28, N47/ as well as to UBBELOHDE Viscometers have two division marks the operating instructions. on the reservoir vessel showing the minimum and maximum filling quantity. In case of Micro UBBELOHDE Viscometers there is only one mark which is to be adhered to within a tolerance range of about ± 1 mm. This means that more accurate dos- ing is not required. It should only be ensured that the opening of the venting pipe on the reference level vessel is above the liquid level.

Considering that air bubbles occurring during the measurement process may lead to scattering of the measurements values, it has to be ensured that no bubbles occur during the filling of the viscometers. For this purpose, the viscometer is held in a slightly inclined position, and the liquid is filled in such a manner that it will float down into the reservoir vessel along the filling tube without any bubbles occurring.

Best results when filling UBBELOHDE Viscometers were achieved using throw-away syringes with an attached glass-tip filter. When using syringe filters, prior filtration is not necessary.

Especially when filling in substances of a higher viscosity into OSTWALD Viscometers the pipette should be immersed deeply into the filling tube in order to prevent after-flow errors.

Table 2 Filling quantities for Figure 17 CANNON-FENSKE Routine Viscometer various viscometer types 1 tube with capillary

2 venting tube Viscometer type Filling quantity 3 reservoir [ml] 4 lower timing mark M2 OSTWALD 3 5 upper timing mark M1 Micro OSTWALD 2 6 pre-run sphere UBBELOHDE 15 - 22 7 capillary 8 measuring sphere Micro UBBELOHDE 3 - 4 9 tube extension CANNON-FENSKE Routine 5 - 12 CANNON-FENSKE approx. 12 Reverse Flow BS/IP/RF U-Tube approx. 20 Reverse Flow

Suspending the viscometers in the racks The test temperature should be kept constant both SCHOTT-GERÄTE offers for all viscometer types locally as well as timely in a range between + 15 °C fixation racks or holders, respectively, which ensure a up to + 100 °C at an accuracy level of ± 0.01 K. Out- stable, vertical suspension of the viscometers in the side the indicated temperature range major fluctua- thermostat bath. In addition, they protect the vis- tion cannot be avoided in each case, but these fluc- cometers from breaking. tuations should still not exceed ± 0.05 K either. If, in particular cases, extreme precision is required, it is Prior to the measurement, UBBELOHDE Viscome- recommended to keep the test temperature timely ters are to be suspended in the racks provided for constant within a range of + 15 °C to + 100 °C at an this purpose, and fixed in position by pressing the accuracy level of 0.01 K, and outside this range at spring downwards. and a accuracy level of ± 0.03 K. The temperature should be checked using gauged mercury glass thermometers with a resolution of 0.01 K.

The liquid bath and in particular the thermometer are to be protected from direct exposure to light sources. There recommended bath liquids are. below 0 °C: antifreezers, e.g. glycerine + water 0 ... 80 °C: distilled water + tap water 80 ... 105 °C: water + glycol 105 ... 200 °C: propylene glycol, silicone oil, paraffin oil

The viewing thermostats of the CT series, developed by a SCHOTT-GERÄTE especially for capillary viscometry, meet the requirements with regard to the timely and local constancy of the temperature of the bath liquid. They are equipped with openings or in- serts, respectively, for two (CT 52/2, CT 1650/2) or four capillary viscometers (CT 1650/4). Once filled and placed in the fixation rack or the holder, respectively, the capillary viscometers are hung into the thermostat bath the temperature of which was pre-adjusted. When using viewing thermostats of the CT series, special viscometer-rack in- serts for manual measurement are available. Subsequently, the sample is exposed to thermostat treatment in the viscometer. When performing measurements using UBBELOHDE, OSTWALD or CANNON-FENSKE Routine Viscometers it is recommended to suck the

liquids at least three times into the measurement Figure 18 UBBELOHDE Viscometer sphere in order to speed up the heat transfer. This with fixation rack procedure is not possible the case of Reverse-Flow Viscometers. Their temperature adjustment should therefore be correspondingly longer. 6.4 Performing the measurement Thermostat treatment The following temperature-adjustment times are recommended: Viscosity is highly depending on the temperature /24/. For this reason, the viscometers have to be 10 min: low-viscosity substances; treated in a thermostat during each measurement. 20 min: high-viscosity substances, The thermostats used are automatically controlled low-viscosity substances in the case of liquid viewing thermostats. The viscometer has to be Reverse-Flow Viscometers; immersed until the bath liquid is at least 2 cm higher 30 min: high-viscosity substances in the case of than the liquid meniscus in the viscometer in its high- Reverse-Flow Viscometers. est position.

Manual measurement In order to make the measurement values available for For the measurement of the flow times, the liquid is statistical evaluation the measurement process should sucked into the measurement sphere by applying a be repeated several times. Especially in the case of vacuum to the capillary tube. When using viscome- UBBELOHDE Viscometers, on order to avoid any ters with a feeder sphere, the latter should be filled at formation of bubbles, it should be noted that a re- least up to its half. newed sucking or pressing up of the measurement substance must only begin when the drainage of the Viscometers without a pre-run sphere are filled until liquid from the capillary is completed. the liquid meniscus is approximately 20 mm above When using Reverse-Flow Viscometers, sucking the the upper annular mark. If UBBELOHDE Viscome- liquid into the measurements sphere is not applica- ters are used, the venting tube should be closed with ble. To perform the measurement, the tube which a finger tip prior to starting sucking in. Upon comple- was closed after filling, is opened on the side of the tion of the filling process the suction hose is removed measurement sphere, and subsequently one meas- from the capillary tube and, in the case of the ures the time over which the liquid rises from the UBBELOHDE Viscometer, the venting tube is re- lower to the upper annular mark. The CANNON- leased. FENSKE Reverse-Flow Viscometer is equipped with When measuring highly viscous samples it is rec- two measurement spheres one on top of the other, ommendable to keep the capillary tube closed after i.e. two measurement values are available after just releasing the venting tube until the levelling bulb has one liquid passage. To repeat a measurement when run empty and the suspended level has built up. using Reverse-Flow Viscometers, they have to be emptied, cleaned, and refilled after each measure- When examining highly volatile substances it is rec- ment. ommended to perform the filling of the measurement If the repeatability limit of a measurement series sphere by applying an over-pressure to the filling (2.8 times the standard deviation) exceeds the re- tube, if no bubbles occur in the liquid. Closing and producibility limit indicated for the specific vis- opening the venting tube in the case of cometer, one has to assume the presence of exter- UBBELOHDE Viscometers should be done analo- nal influences. In this case the measurements have gously. to be repeated on a new part of the filtered sample The measurement involves the period of time over after the viscometer has been cleaned. If only a which the lower for vertex of the meniscus sinks from ”maverick” is present it may be deleted or, as a better the upper edge of the upper annular mark down to alternative, be replaced by an additional measure- the upper edge of the lower annular mark. The stop ment value. If necessary, a check for runaway values watch used for timing should have a dissolution of at of this kind is to be performed /17/. least 0.1 s. When the meniscus passage is detected, The calculation of viscosity is done on the basis of it has to be made sure that the annual mark is at eye the mean value of the flow times. level. Automatic measurement Figure 19 (c) shows the proper detection of For automatic viscosity measurement using the meniscus passage. UBBELOHDE, OSTWALD, and CANNON-FENSKE Routine Viscometers, SCHOTT-GERÄTE offers the automatic viscosity measurement devices of the AVS series. Table 3 will give you an overview of the device pro- gram. The selection of the AVS/S, AVS-SK, and AVS/S-CF measurement tripods for automatic viscosity measurement with optical detection is determined by: Viscometer type Bath liquid of the thermostats (metal tripod for non-aqueous media,

PVDF measurement tripod as a corrosion-free Figure 19 Detection of the meniscus passage option) in the case if manual measurement For measurement using the TC-UBBELOHDE Vis- (a), (b) - wrong (c) – correct cometer no measurement tripod is required. The viscometer is clamped into a special fixation rack and suspended in the thermostat bath. The connection with the control unit is made using a cable which is plugged into a socket on the viscometer head.

Table 3 Automatic viscometer measurement devices from SCHOTT-GERÄTE

Type of device AVS 310 AVS 350 AVS 360 AVS 450 AVSPro

Meniscus detection Photoelectric Photoelectric Photoelectric Photoelectric Photoelectric barrier barrier barrier barrier barrier NTC sensor NTC sensor NTC sensor NTC sensor

Working mode pressing pressing sucking sucking sucking nmax 10 10 10 999 10 Interface RS-232-C RS-232-C RS-232-C RS-232-C RS-232-C Special features software control built-in printer fully automatic viscosity (optimal) measurement system

nmax = max. number of programmable single measurements in repetitive operation

The viscometers are pneumatically connected to the The displacement of the measurement liquid within AVS device via silicone or PTFE hoses. the viscometer is done via an internal diaphragm pump with pressing or sucking action. The pump is All automatic devices are microprocessor-controlled. controlled in such a manner that an optimal pumping An interface enables an external printer or computer pressure for a reproducible filling of the measure- to be connected. ment systems is built up as a function of the viscosity of the sample being analysed. Parameterisation and the triggering of the automatic measurement cycle are done on the control unit. For the Viscosity Measuring Units AVS 350 and AVS 360 a PC software is available which enables the control of up to 8 measurement devices and an evaluation of the measurement data.

Figure 20 shows a viscosity measurement station with 4 AVS 350 Viscosity Measuring Units.

Figure 20 AVS 350 Viscosity Measuring Units with UBBELOHDE and TC Viscometer and Viewing Thermostat

The AVSPro Automatic Viscosity Sampler is a fully The maximum viscosity limit of the product being automatic viscosity measurement system for routine analysed for use in the AVSPro Automatic Viscosity measurements. Sampler is approx. 800 mm2/s at 25 °C. This device performs measurements of kinematic The AVSPro Automatic Viscosity Sampler may be and relative viscosity up to calculation and documen- operated with a maximum of eight Micro UBBE- tation work in an independent manner. LOHDE Viscometers equipped with TC Sensors in two thermostat baths at two different measurement Filling, emptying, and rinsing of the viscometers are temperatures simultaneously. integrated in the automatic course of the measurement. Sample carriers for up to 16 sample bottles of 100 ml Operating guidance is via menu-control using a each, or sample carriers for a 56 sample bottles of monitor, a mouse, and a computer keyboard. 20 ml each can be used.

Figure 21 shows a viscosity measurement station with the AVSPro Automatic Viscosity Sampler.

Figure 21 AVSPro Automatic Viscosity Sampler equipped with 8 Micro UBBELOHDE Viscometers with TC Sensors in 2 Viewing Thermostats CT 53

7 Causes of errors and special corrections 7.1 Correctable errors and corrections

Rising-height error Thermal expansion of the capillaries and the measurement vessel Surface tension causes the liquid which is wetting the During high- and low-temperature measurements the tube wall to climb by a distance of Dh. radius and the length of the capillaries, the volume of the measurement sphere, and the average pressure height of the viscometer will change owing to the The size of the relative rising-height error e in terms large difference between the measurement and the of % can be calculated on the basis of the following calibration temperature. For this reason the viscome- formula: ter constant has to be corrected in the case of precision measurements. I 2 1 1 I 0 A = ( - ) ( - ) · 100 % (7.1) The corrected device constant according to the equa- g h r r H H m 12 0 tion (7.2):

K´ = K (1 + a (J - J0)) (7.2) h - mean pressure height m Viscometers from SCHOTT-GERÄTE are calibrated g - acceleration due to gravity at a temperature of J0 = 23 °C. The coefficient a of r - radius of the upper reservoir vessel at â 1 longitudinal expansion of the DURAN -glass2) used the liquid meniscus for production is 3.3 • 10-6 K-1. r2 - radius of the lower reservoir vessel at the liquid meniscus Thermal expansion of the s - surface tension of the measurement substance measurement substance s - surface tension of the calibration substance 0 In the case of UBBELOHDE Viscometers no correc- r - density of the measurements substance tion is required, since the measurement result is r 0 - density of the calibration substance largely independent of the substance quantity being filled in. If, in the case of viscometers without suspended level, the substance temperature should deviate from the measurement temperature during the In the case of precision measurements the influence process of filling the viscometer, a volume change of of the rising-height error is to be noted with the fol- the measurement substance leading to a change of lowing viscometer types, if the relation between sur- the viscometer constants will occur during the tem- face tension and density of the measurement liquid perature adaptation. deviates considerably from that of the substance used for a calibration: In this case the constants are to be corrected according to equation (7.3) for OSTWALD and CANNON- a) in the case of viscometers with a small FENSKE Routine Viscometers, or according to equa- pressure height, where the liquid flows tion (7.4) for Reverse-Flow Viscometers. from the upper vessel into another vessel the diameter of which is considerably 4 V (p2 - p1 ) K¢ = K (1 + ) (7.3) different from the one of the upper vessel, 2 e. g. CANNON-FENSKE Viscometer, F D m h m p2 OSTWALD Viscometer; b) in the case of all of pipette viscometers. 4 V (p2 - p1 ) K¢ = K (1 - ) (7.4) 2 F D m h m p2 In the case of UBBELOHDE Viscometers the correction will in general be no more than 0.1 to 0.2 % and can thus be neglected in most cases. Dm - mean diameter of the liquid meniscus in the reservoir vessel r1 - density of the measurement substance at filling temperature r2 - density of the measurement substance at measurement temperature

2) Registered trademark of SCHOTT GLAS, Mainz, Germany

Inclination error Viscometers have to be used in the position in which In the case of a temperature error of ½s½ £ 1K, the they were calibrated. If the connection line between relative error in the viscosity measurement is: the centre points of the reference level vessels deviates from normal position, the mean pressure height A of the viscometer will change. If, instead of the initial = Un • s • 100 % (7.7) angle f0 the connection line compared to perpendicular is at an angle of f the corrected device constant is to be calculated according to: The temperature coefficient Un is determined according to the corresponding DIN standard /N4/. cos f K¢ = K (7.5) cos f The temperature measurement should only be made 0 fully in mass in gauged thermal metres with a resolu-

tion of 0.01 K. Their requirements imposed on the fair The fixation racks or holders offered by SCHOTT- most bats to be used are it described in Chapter 6.4. GERÄTE ensure a perpendicular suspension of the viscometer with a deviation < 1°. This corresponds to a max. relative constant error of 0.02 %. This means 7.2 Uncorrectable errors that the inclination error can be neglected if these racks are being used. Turbulence Laminar flow is the basic requirement for viscosity Local independence of the acceleration measurement according to the capillary principle. of the fall Laminar flow is present if the Reynolds number Re is A correction is required if the acceleration of the fall < 2300. Owing to the sensitivity to disturbance of the at the calibration place g0 and the acceleration of the flow, it is useful to remain far below this value when fall at the measurement place g are significantly dif- performing measurements. ferent. For a given viscometer the Reynolds number can be Equation (7.6) is to be used to calculate the corrected calculated according to the following numeric-value device constant. equation:

¢ g V K = K (7.6) Re = 63.7 × (7.8) g0 2 R × K × t g

Accuracy of the watches V = [cm3] R = [cm] K = [mm2/s2] tg = [s] If mechanical stop watches are being used, these have to be adjusted in such a manner that their max. Considering that Hagenbach correction will increase accuracy error is less than 2 s per hour. In this case with an increase in the Reynolds number, one should the occurring error is less than 0.05 %. Prior to begin- work with a Reynolds number below 200 if this is ning the measurements the watch should be wound possible in practice /N10/. up to exclude variations of the spring force. It is recommended to check the accurate march of the watches regularly using a time standard. Disturbance of the suspending level If the time is measured electronically using a corre- in the case of UBBELOHDE Viscometers sponding frequency standard, the frequency being If viscosity measurements are performed with short used has to be constant and to correspond at least to flow times, a deformation of the suspended level may -4 10 of the set value. occur. This will lead to systematic measurement errors, since the average pressure height of the vis- Inaccurate adjustment and cometer will change. In addition, one has to reckon measurement of temperature with an increased scattering of the measurement Errors caused by inaccurate temperature adjust- values within the limit ranges between the disturbed ment or insufficiencies in the temperature stability or and the undisturbed suspended level, and the influ- temperature measurement are frequently very ence of surface tension on the measurement result large, since the viscosity of most of the liquids var- will increase. ies largely as a function of temperature /24/.

Figure 22 shows various stages of level disturbance. Drainage errors Drainage errors are caused by the fact that a small Table 4 will give you an overview of the limit values liquid volume DV is adhering to the wall of the vis- of the Reynolds numbers and the flow times up to cometer above the sinking liquid meniscus. DV will which in general no disturbances of the suspended increase with the viscosity and the sinking velocity of level will occur for UBBELOHDE Viscometers (nor- the meniscus. The magnitude of the error is also in- mal design). Considering further that the limits are fluenced by the wettability of the wall, the surface also depending on the surface tension of the liquid tension of the liquid, and the geometry of the vis- and the shape of the capillary outflow, disturbances cometer. Depending on the constructional shape of of this kind may even occur in the case of somewhat the device a shortening or extension of the flow times longer flow times. may occur.

Table 4 Limit values of tg and Re Radiation heat (UBBELOHDE Viscometer) /N9/ To avoid an uncontrolled heating up of the liquid to be tested by heat radiation, the liquid bath is to be Capillary no. 0c 0a I Ic protected from direct exposure to the sun or light sources. Cold lights or light sources with a pre- tg [s] 100 75 60 60 mounted infra-red filter should be used for illumina- tion preferably. Re 500 500 300 100

Start-up length One of the preconditions are for capillary viscometry is a parabolic velocity profile. For this reason the flow time has to be selected in such a manner that the start-length le for the formation of the profile is considerably smaller than the capillary length. According to Schiller /10/ the start-up length can be calculated as follows:

V l = 0,115 (7.9) e pn tg

Figure 22 Stages of distribution of the suspended level in the case of UBBELOHDE Viscometers /N9/ (a) no disturbance - measurement can be used (b), (c), (d) disturbance - measurement cannot be used

7.3 Frequently occurring error symptoms, possible causes of errors and ways of error elimination

Table 5 gives a summary of some of the major error occurrences occurring during viscosity measurements using glass capillary viscometers, including their possible causes and ways of elimination. Errors which can be attributed to device defects as well as improper use of the automatic viscosity measurement devices are not listed in the table below.

Table 5 Frequently occurring errors when using glass capillary viscometers

Error symptom Error causes Possible error elimination systematic measurement error: after-flow error, experimental determination of the Hagenbach flow time too large with Hagenbach correction correction using substances having a similar short flow times too small viscosity and a surface tension as the measurement product (please refer to Chapter 4) systematic measurement error: after-flow error, as above, better: Viscometer with flow time too small with Hagenbach correction a smaller capillary diameter short flow times too large (please refer to Chapter 6.1) systematic measurement error: substance quantity filled in was empty, clean and refill viscometer flow time too small with too small (please refer to Chapter 6.2 / 6.3) (Ostwald, CANNON-FENSKE or BS/IPRF U-Tube Reverse Flow Viscometer) systematic measurement error: substance quantity filled in was as above flow time too large too great (Ostwald, CANNON-FENSKE or BS/IP/RF-U-Tube Reverse Flow Viscometer) systematic measurement error: disturbance of the suspended select viscometer with flow time too small with level a smaller capillary diameter, short flow times (please refer to Chapter 6.1 / 7.2) (UBBELOHDE Viscometer) systematic measurement error: temperature of the bath liquid check temperature; flow time too small too high if necessary, readjust thermostat systematic measurement error: contamination in the capillaries empty and clean viscometer flow time too large (please refer to Chapter 6.2), repeat measurement temperature of the bath liquid check temperature, too low if necessary, readjust thermostat drift of the flow times drift of the bath temperature protect the thermostat from direct radiation exposure (please refer to Chapter 7.2), if necessary, replace thermostat temperature-adjustment of the continue temperature adjustment until measurement substance not the time values are stable completed (please refer to Chapter 6.4) evaporation of a highly volatile apply pressing operating mode component; reaction of the product being analysed with the air

Continuation of table 5

Error symptom Error causes Possible error elimination increased stochastic scattering of contamination in the empty and clean viscometer (please refer to Chap- the measurement values viscometer ter 6.2); repeat measurement contamination in the product empty and clean viscometer; repeat the measure- being analysed ment with a filtered sample; if necessary, use a filter with a smaller pore width (please refer to Chapter 6.2/6.3) air bubbles in the viscometer in the case calf pure matters with chemical and physical heat resistance, drive out bubbles by a shorter time increase of temperature clean and empty viscometer (please refer to Chap- ter 6.2); during refilling, ensure absence of bubbles (please refer to Chapter 6.3) excessive stochastic scattering oc- contamination of the remove the viscometer tripod from the thermostat curring during automatic meas- optical sensors bath; clean optical system using non-denatured al- urements using optoelectric cohol on a soft cloth barriers (Baron possibility of total errors triggered by the opto- use a TC-UBBELOHDE, OSTWALD, or CANNON- malfunction) electric barriers as a result of FENSKE Routine Viscometer the formation of bubbles, foam, (please refer to Chapter 6.1) or liquid lamellae excessive stochastic scattering oc- Incrustation of the sensors (in transparent media: curring during automatic meas- the case of thermally instable use optical flow-time measurement urements using TC Viscometers media) opaque media: (possibility of total malfunction) use Reverse Flow Viscometer wear and tear of the sensors replace viscometer increased stochastic scattering in beginning deformation of the select a viscometer with a smaller the case of short flow times suspended level capillary diameter (UBBELOHDE Viscometers) (please refer to Chapter 6.1/7.2) periodically fluctuating flow times heating-up or cooling-down set the heating and cooling of the thermostat phases of the thermostats in such a manner that at least two complete too long temperature cycles are completed during one viscosity measurement cycle no timely stability of the bath- replace the thermostat liquid temperature (please refer to Chapter 6.4) (defective thermostat) malfunction caused by air bubbles substance quantity filled in was UBBELOHDE Viscometer: during the sucking-in process of too small fill up the measurement substance; the liquid into the delivery vessel others: empty and clean viscometer; repeat measurement

8 Special application 8.1 Testing of plastics

Measurement problem Solution One of the major quality features of synthetic materi- That determination of the chain length, processing als is the mean molecular weight of the polymer properties, and quality of a synthetic material is done molecules. The molecular weight characterises the in the form of viscosity measurements on solutions of chain length of the polymer molecules which has a the plastic in suitable solvents using capillary vis- decisive influence on the processing properties of a cometers (solution viscometry). Table 6 will inform synthetic material. you about the solvents, viscometers, and the application of the relevant standards. The strain exerted on the plastic by the processing process may lead to changes in the polymer changes The viscosity number (for a definition, please refer (as a rule a decay of the chains). Under certain cir- to Table 7) gives information about the processibility cumstances the properties of the finished part might of plastic material. It plays a decisive role within the be changed to such an extent that it is no longer suit- framework of quality control of the granules. In addi- able for its intended purpose. tion, it is important to verify the viscosity number of the finished plastic part. In the research and development of polymers new polymers are being developed and produced. In this In most cases the indication of the viscosity number process, too, the chain length of the polymer mole- or of the relative viscosity (please refer to Table 7) cules is of essential importance as to the characteri- is sufficient as a quality criterion of established sation of the finished product. plants. This requires the determination of the viscosity of the solvent and of the plastic solution (concen- This results in the following measurement tasks: tration is mostly 0.5 g/100 ml). Polymer research and development Instead of the viscosity number the determination Determination of the mean chain length of mean frequently involves the K value after Fikentscher polymerisation degree of the polymer molecules /N21/. Objectives: - Characterisation of the finished product The determination of the mean molecular weight of - Optimisation of its chemical and the polymer molecules is done via the limiting vis- physical properties cosity number (please refer to Table 7). It is of par- Rating of polymerisation installations ticular importance in the range of research and de- Determination of process parameters velopment of polymers and of procedures and installations for their production and processing. In addi- Polymer chemistry (polymer production) tion, it is an important feature as regards quality as- Determination of the mean chain length or mean surance in the case of special applications, such as degree of polymerisation of the finished product plastic recycling and the processing of recycled plas- (raw granules) tics. Objectives: - Characterisation of the finished product To determine the limiting viscosity number, polymer - Quality assurance solutions of different concentrations are produced - Optimisation of the process parameters (so-called dilute series /36/). The limiting viscosity - Prevention of the production of spoiled batches number results from the extrapolation of the viscosity numbers to a concentration = 0. Polymer processing Characterisation of the properties and the capabilities of the starting material (raw granules) Objectives: - Rating of plants for polymer processing - Determination of optimum process parameters Determination of the chemical and physical properties of the finished part Objectives: - Quality assurance - Optimisation of the process parameters

Table 6 Choice of applications for solution viscometry (as a rule, concentrations of 0.5 g/100 ml are weighted in and the measurements of viscosity are performed at 25 °C).

Polymer Abbreviation Solvent Capillaries DIN Polyamide PA Formic acid (90%) I, Micro Ic 53 727 Sulphuric acid (96%) II, Micro IIc 53 727 m-cresol II 53 727 Polycarbonat PC Dichloromethane 0c 7744, Part 2 Polyethyleneterephtalate PET Phenol / 1.2- dichlorobenzene Ic 53 728, Part 3 Polybutyleneterephthalate PBT (1:1 parts by weight) 2- chlorophenol Ic m-cresol II Dichloroacetic acid II, Micro IIc Polyvinyl chloride PVC Cyclohexanone I 53 726 Tetrahydrofurane Ic Polyethylene 1) PE Decahydronaphthalene I 53 728, Polypropylene 1) PP (Decalin) Sheet 4 Polystyrol 2) PS Toluol I 7741, Part 2 o-Xylol I 1.2-Dichlorbenzol I Polymethycrylate 3) PMMA Chloroform 0c 7745, Part 2 Acetophenone I Cellulose acetate CA Dichloromethane / Methanol 0c 53 728, (9:1 party vy volume) Sheet 1

1) For concentration, please refer to DIN 53 728, Sheet 4 2) Measurement temperature: 135 °C 3) c = 0,26 g/l

Table 7 Definition of the terms used in dilute viscometry /18, N2/

Calculated value Description h dynamic viscosity n = h / r kinematic viscosity

hr = h / hS relative viscosity, viscosity ratio

(h - hS) / hS = hr - 1 relative viscosity change, specific viscosity

Jv = 1 / c · (h - hS) / hS Staudinger function, viscosity number

Ln (h / hS) / c inherent viscosity

Jg = lim [1 / c × (h - hS) / hS] Staudinger index, limiting viscosity number, intrinsic viscosity c®0

8.2 Viscosity determination of oils and additives

Measurement problem At high temperatures (for instance, in summer, when Petroleum is a mixture of hydrocarbons. By way of driving at full throttle; under extreme loads, e.g. when vacuum distillation it is split up into different fractions driving up a mountain) oil temperatures can raise up (fuels and lubricants, please refer to Table 8). above 100 °C. In this case, too, a still sufficient formation of lubricating film has to be ensured, so that Table 8 Typical for fractions of the the lubricating film will not break as a result of the in- distillation of crude oil /19/ sufficient viscosity in the places being subjected to friction. Fraction Boiling range [°C ] The life of an engine oil is limited, since in operation Natural gas below 20 ageing and external matters are building up on the one hand (e.g. caused by oxidation of the basic oil, Distillation of crude oil 30 ... 60 metal abrasion, formation of soot), and the additives Ligroin or white spirit 60 ... 90 are pooring down on the other (e.g. caused by the decay of the polymers owing to shearing action, oxi- Gasoline 85 ... 200 dation, and thermal strain) /20, 21, 22/. Kerosene 200 ... 300 The determination of viscosity is playing a major role Fuel oils 300 ... 400 in the production and development of doped oils (ba- Lubricating oil and grease, above 400 sic oil / additive mixtures). In the course of produc- paraffin, wax, asphalt tion, regular viscosity measurement ensures ade- quate quality control. As regards development on the

other side, the focus is on the examination of the viscosity-temperature behaviour of new oil/additive mix- Viscosity is a decisive characteristic for the flowing tures. and lubricating capabilities of an oil. The mixture of various raffinates leads to basic oils, with different In the case of used engine oils the determination of viscosities. Their properties can be considerably im- viscosity can be used to determine whether the for- proved by chemical additions or additives, such as mation of the lubricating film will still be sufficient viscosity-index improvers (VI improvers), detergents, even at higher temperatures. dispersing agents, wear-and-tear reduction agents, and oxidation or corrosion inhibitors. Solution One of the frequently used characteristics of viscosity-temperature behaviour (VT behaviour) of a lubri- So it is, for instance, that lubricating oils form a lubricating oil is the VI viscosity index. The VI of an oil cating film between the rubbing parts inside the engine can be calculated on the basis of the viscosities at which prevents a direct contact of the solid surfaces, 40°C and 100°C by using tables /N11/. The magni- and thus wear and tear. The thickness of this lubricat- tude of the viscosity drop occurring with increasing ing film depends on the viscosity of the oil. temperature depends on the chemical composition of the oil under concern. A minor temperature- dependence of the viscosity will lead to a higher vis- The viscosity of a mineral oil changes considerably cosity index. Multigrade, engine, and gear oils are with temperature. At low temperatures (for instance, characterised by a high VI /23/. in winter, during the cold start of the engine) the viscosity of the oil must still be low enough to enable the The classification of an engine lubricating oil in so- oil being pumped to the rubbing parts inside the en- called SAE viscosity classes is based on dynamic gine. viscosity at -17.8 °C (0 °F) and kinematic viscosity at 98.9 °C (210 °F) /N12, N13/.

Table 9 contains a list of examples of viscometers and accessories from SCHOTT-GERÄTE. Table 9 Measurement stations for viscosity measurements on oil and additives

automatic measurement manual measurement Viscometer UBBELOHDE UBBELOHDE CANNON-FENSKE Routine CANNON-FENSKE Routine TC Viscometer (for dark oils) CANNON-FENSKE Reverse Flow Viscometer (for dark oils) BS/IP/RF U-Tube Reverse Flow Viscometer (for viscid and / or dark oils) Viscosity AVS 350, AVS 360, AVS 450 Stop watch measurement AVSPro device (up to n » 1200 mm2/at room temperature) Accessories Thermostat and cooler Thermostat and cooler Viscometer Cleaner AVS 26 (optional)

8.3 Testing of food

Measurement problem Objectives of viscosity measurement: optimisation of the mashing properties The raw materials, semi-finishings, and finished selection of filtration strategy products to be processed in food industry are charac- quality evaluation of malt, terised by very much different rheologic properties. wort, and beer Provenience, temperature, water percentage, intensity of the mechanical processing, storage time, and b) Determination of viscosity of conditions of transportation are some of the factors fruit and vegetables juices influencing the mainly non-Newtonian flowing behaviour of food masses. Raw-pressed juices with a high viscosity are hard to But a number of fluids involved in food production clarify. Viscosity is mainly affected by the pectin per- also presents Newtonian behaviour. centage which, in the case of concentrated fruit juices, may rise so high in the course of production In food technology and machine engineering, knowl- that there is a danger of jellying of the contents of the edge of viscosity is of importance in multiple ways, tanks. Owing to the food-physiological importance, a e.g. for: complete decay of the pectin is not desired. controlling technological processes evaluating the products’ quality By way of an aimed pectinological decaying process designing food dispensers and in the course of the technological section of the fining conveying apparatus and clarification process of the juices one tries to ad- selecting and operating packaging installations. just an optimum pectin percentage /27/.

Within the framework of food-technological research Objectives of viscosity measurement: and development the measured viscosities can be gathering of control parameters for the used to draw valuable information about pectinological process of optimising of molecular structure optimising clarification and fining chemical composition quality surveillance efficiency of enzymes characterisation of the jellying capabilities percentage of viscosity-influencing of pectins, inter alia by the determination of the constituents and additives limiting viscosity number /28/. Examples of measurement tasks in c) Viscosity determination in sugar industry food industry a) Determination of viscosity of In the extraction and technical processing of saccha- beer wort and beer /26/ rarose solutions, information about viscosity is essen- Beers with a viscosity > 1.7 mPas are hard to filter, tial /29/. It increases in the form of an exponential and this leads to a reduction of the production output. curve with rising concentration and has thus a sub- On the other hand a high viscosity has a positive ef- stantial influence on the crystallisation readiness of fect on the full-bodiedness and the stability of the sugar solutions. foam.

So it is that the crystallisation of sacchararose solu- On the other side it was possible to show that the tions is favoured with increasing concentration (state homogenisation effects can be improved through vis- of over-saturation), but will decrease with the rise in cosity control. the percentage of more than saccharide. Viscosity increases with the rise in the molecular mass of the An addition of hydrocolloids (thickening, binding, and solution components (mono- and disaccharides, glu- jellying agents) and stabilisers has a highly viscosity- cose syrup /30/) and can be calculated as follows as raising effect. Viscosity measurement provides valu- regards saccharide solutions: able information required to reveal their chemical structure and their effect in combination with components of the milk. h = wA log hA + wB log hB (8.1) Objectives of viscosity measurement: w - Mass portions in the total mixture Technology surveillance A, B - Components Quality evaluation Development of recipes

Glucose syrups are characterised by different sac- Solution charide fractions with one and the same saccharifica- After examining the question of knowing whether the tion degree, a fact which results in diverging viscous food liquid to be analysed can be reasonably treated behaviour patterns. Considering that they are used as a Newtonian fluid, all types of capillary viscome- as crystallisation inhibitors in the production of con- ters can be used in principle. fectionery, viscosity is a major technological parameter. There may be some difficulties in the detection of the liquid meniscus. Objectives of viscosity measurement: Owing to their low degree of transparency and the af- gathering of control parameters for ter-flow effects, an optical detection of dairy products processing sugar solutions is problematic. quality surveillance development of recipes The use of TC Viscometers requires frequent, thor- provision of information for the rating of ough cleaning, since the thermistors tend to soil as a appliances and apparatus for sugar industry result of incrustation. d) Viscosity determination in milk industry There are less problems in the viscosity measurement on beer, fruit juices and the like. Owing to the Owing to the differences in the provenience and fact that these fluids have a tendency of foam forma- composition of milk and dairy products, the rheologic tion, OSTWALD Viscometers and even Micro behaviour of dairy products differs greatly /31/. OSTWALD Viscometers have proven their suitability The viscosity of milk, cream, condensed milk etc. is for use. According to /26/ the standard deviation with influenced by the fat contents, the concentration of viscosity measurements performed on wort using the dry matters, and, to a high degree, by the proc- OSTWALD Viscometers was 0.004 mPas compared essing conditions, e.g. by homogenisation. to 0.02 in the case of measurements made using the HÖPPLER Viscometer. Similarly good experience was gathered on viscometer measurements made on fruit juices.

9 Formula signs and units used

A Surface being parallel with flow direction m2

B Constant of Hagenbach correction with mm2 s sharp-edged capillary ends c Concentration g/cm3

D Shear rate 1/s

E Constant of Hagenbach correction with mm2 s funnel-shaped capillary ends

F Force acting in flow direction N g Acceleration of the fall at the place of the measurement m/s2

2 g0 Acceleration of the fall at the place of determination of the constant m/s hm Mean hydrostatic pressure height cm

Dh Capillary rising height of the liquid cm

3 Jg Limiting viscosity number according to STAUDINGER cm /g

3 Jv Viscosity number according to STAUDINGER cm /g

K Viscometer device constant mm2/s2

2 2 KP Viscometer device constant (device being tested) mm /s

2 2 KR Viscometer device constant (reference viscometer) mm /s

K´ Corrected viscometer device constant mm2/s2

L Length of the capillaries cm le Inflow length cm m Coefficient of Hagenbach correction - n Coefficient of Couette correction -

Dp Acting pressure mbar

DpC Pressure loss resulting from Couette correction mbar

R Radius of the capillaries cm

Re Reynolds number - r1 Radius of the upper reservoir vessel on the liquid meniscus cm r2 Radius of the lower reservoir vessel on the liquid meniscus cm s Temperature error K t Flow time s tg Measured flow time s tgP Measured flow time (device being tested) s

33 tgR Measured flow time (reference viscometer) s tH Hagenbach-Couette correction s tHP Hagenbach-Couette correction (device being tested) s tHR Hagenbach-Couette correction (reference viscometer) s ts Shear time s

T Measurement temperature K

T0 Calibration temperature K

U Voltage V

Un Temperature coefficient of kinematic viscosity 1/K

V Flow through volume cm3

V Flow volume cm3/s

DV Liquid volume adhering to the inner wall surfaces of the viscometer cm3 v Mean flow velocity m/s x Co-ordinate in flow direction m y Co-ordinate perpendicular to flow direction m a Longitudinal expansion coefficient of the glass grade 1/K used for the production of glass capillary viscometers e Relative error of the measurement value % h Dynamic viscosity mPa s hr Relative viscosity - hS Dynamic viscosity of the solvent mPa s n Kinematic viscosity mm2/s r Density of the liquid to be measured g/cm3

3 r0 Density of the normal liquid g/cm s Surface tension of the liquid to be measured mN/m s0 Surface tension of the normal liquid mN/m

J Temperature °C t Shearing strain Pa f Angle between the perpendicular and the connection line o of the upper and lower central point of the reference level vessel during measurement

o f0 Angle between the perpendicular and the connection line of the upper and lower central point of the reference level vessel during calibration

10 Bibliography

/1/ Hagen, G. Poggendorffs Annalen der Physik 46 (1839), 423

/2/ Poiseuille, J. L. Comptes rendus 11 (1840), 961; Mémoires des Savants Etrangers 9 (1846), 433

/3/ Prandtl, L. Strömungslehre Friedrich Vieweg u. Sohn, Braunschweig 1960

/4/ Eck, B. Technische Strömungslehre B.I; 75; Springer-Verlag, Berlin 1978

/5/ Hagenbach, E. Poggendorffs Annalen der Physik 109 (1860), 385

/6/ Couette, M. Annales de Chimie et Physique 21 (1890), 433

/7/ Hengstenberg, J.; Sturm, B.; Winkler, O. Messen, Steuern und Regeln in der chemischen Technik B. II, 432 - 499; Springer-Verlag, Berlin 1980

/8/ Kestin, J.; Sokolov, M.; Wakeham, W. Applied scientific research 27 (1973), 241

/9/ Boussinesq, V. S. Comptes rendus 110 (1890), 1160, 1238

/10/ Schiller, L. Forschung auf dem Gebiete des Ingenieurwesens (1922), H. 248

/11/ Riemann, W. Journal of the American Chemical Society 50 (1928), 46

/12/ Weber, W.; Fritz, W. Rheologica Acta 3 (1963), 34

/13/ Dorsay, N. E. The physical review 28 (1926), 833

/14/ Cannon, M. R.; Manning, R. E.; Bell, J. D. Analytical Chemistry 32 (1960), 355

/15/ Kryk, H.; Wilke, J. Möglichkeiten zur Meßbereichserweiterung von Mikro-UBBELOHDE-Viskosimetern Vortrag anläßlich der Jahrestagung der Deutschen Rheologischen Gesellschaft, Karlsruhe 1993

/16/ Gebrauchsanleitung UBBELOHDE-Viskosimeter mit hängendem Kugelniveau SCHOTT-GERÄTE GmbH, Hofheim am Taunus

/17/ Doerffel, K. Statistik in der analytischen Chemie Deutscher Verlag für Grundstoffindustrie, Leipzig 1982

/18/ Brown, R. P. Taschenbuch der Kunststofftechnik Carl Hanser Verlag, München 1984

/19/ Streitwieser, A.; Heathcock, C. H. Organische Chemie Verlag Chemie, Weinheim 1980

/20/ Jentsch, C. Chemie in unserer Zeit 12 (1978), 57

/21/ Klein, J.; Müller, H. G. Erdöl und Kohle, Erdgas, Petrochemie 32 (1979), 394

/22/ Klein, J.; Müller, H. G. Erdöl und Kohle, Erdgas, Petrochemie 35 (1982), 187

/23/ Stepina, V.; Vesely; V.; Trebicky, Vl. Tribologie und Schmierungstechnik 34 (1987), 113

/24/ Werner, S. Zur Temperaturabhängigkeit der Viskosität niedrigviskoser Newtonscher Fluide Diplomarbeit; TH Köthen 1993

/25/ Wilke, J.; Kryk, H. Temperaturkonstanz und Temperaturverteilung in Flüssigkeitsthermostaten VDI/VDE-GMA-Tagung TEMPERATUR ´92 VDI Berichte 982 (1992), 265 - 268

/26/ Greif, P. Monatsschrift für Brauerei 32 (1979), 356

/27/ Köller, M.; Grandke, I. Lebensmittelindustrie 24 (1977), 162

/28/ Kunzele, H.; Kleiber, U.; Bergemann, U. Lebensmittelindustrie 36 (1989), 78

/29/ Hoffmann, H.; Mauch, W.; Untze, W. Zucker und Zuckerwaren Verlag Paul Parey, Berlin 1985

/30/ Schiweck, H.; Kolber, A. Gordian 72 (1972), 41

/31/ Kessler, H. G. Lebensmittel- und Bioverfahrenstechnik/ Molkereitechnologie Verlag A. Kessler, Freising 1988

/32/ UBBELOHDE, L. Zur Viskosimetrie S. Hirzel Verlag, Stuttgart 1965

/33/ UBBELOHDE, L. Öl und Kohle vereinigt mit Erdöl und Teer 12 (1936), 949

/34/ Kryk, H; Wilke J. GIT Fachzeitschrift für das Labor 5 (1994), 463

/35/ Kryk, H; Wilke J. Erdöl und Kohle, Erdgas, Petrochemie 47 (1994), 467

/36/ Schurz, J. Viskositätsmessungen an Hochpolymeren Verlag Berliner Union GmbH, Stuttgart 1972

11 Standards used in capillary viscometry

1. National Standards

Basics

/N1/ DIN 1342 Teil 1 Rheologische Begriffe

/N2/ DIN 1342 Teil 2 Newtonsche Flüssigkeiten

/N3/ DIN 51 550 Bestimmung der Viskosität Allgemeine Grundlagen

/N4/ DIN 53 017 Bestimmung des Temperaturkoeffizienten der Viskosität

Measuring techniques

/N5/ DIN 51 366 Messung der kinematischen Viskosität mit dem CANNON-FENSKE-Viskosimeter für undurchsichtige Flüssigkeiten

/N6/ DIN 51 562 - 1 Messung der kinematischen Viskosität mit dem UBBELOHDE-Viskosimeter Normal-Ausführung

/N7/ DIN 51 562 - 2 Messung der kinematischen Viskosität mit dem UBBELOHDE-Viskosimeter Mikro-UBBELOHDE-Viskosimeter

/N8/ DIN 51 562 - 3 Messung der kinematischen Viskosität mit dem UBBELOHDE-Viskosimeter Relative Viskositätsänderungen bei kurzen Durchflußzeiten

/N9/ DIN 51 562 - 4 Messung der kinematischen Viskosität mit dem UBBELOHDE-Viskosimeter Teil 4: Viskosimeter-Kalibrierung und Ermittlung der Meßunsicherheit

/N10/ DIN 53 012 Kapillarviskosimetrie Newtonscher Flüssigkeiten, Fehlerquellen und Korrektionen

Test of mineral oils and related products

/N11/ DIN / ISO 2909 Berechnung des Viskositätsindex aus der kinematischen Viskosität

/N12/ DIN 51 511 SAE-Viskositätsklassen für Motoren-Schmieröle

/N13/ DIN 51 512 SAE-Viskositätsklassen für Kraftfahrzeug-Getriebeöle

/N14/ DIN 51 519 ISO-Viskositätsklassifikation für flüssige Industrie-Schmierstoffe

/N15/ DIN 51 563 Bestimmung des Viskositäts-Temperatur-Verhaltens Richtungskonstante m

/N16/ DIN 51 564 Berechnung des Viskositätsindex aus der kinematischen Viskosität

Test of polymers

/N17/ DIN/ISO 1628 Teil 1 Richtlinien für die Normung von Verfahren zur Bestimmung der Viskositätszahl und der Grenzviskositätszahl in verdünnter Lösung Teil 1: Allgemeine Grundlagen

/N18/ DIN 7741 Teil 2 Polystyrol (PS)-Formmassen Herstellung von Probekörpern und Bestimmung von Eigenschaften

/N19/ DIN 7744 Teil 2 Polycarbonat (PC)-Formmassen Bestimmung von Eigenschaften

/N20/ DIN 7745 Teil 2 Polymethylmethacrylat (PMMA)-Formmassen Herstellung von Probekörpern und Bestimmung von Eigenschaften

/N21/ DIN 53726 Bestimmung der Viskositätszahl und des K-Wertes von Vinylchlorid (VC)-Polymerisaten

/N22/ DIN 53 727 Bestimmung der Viskositätszahl von Thermoplasten in verdünnter Lösung Polyamide (PA)

/N23/ DIN 53 728 Blatt 1 Bestimmung der Viskosität von Lösungen Celluloseacetat in verdünnter Lösung

/N24/ DIN 53 728 Blatt 2 Bestimmung der Viskosität von Lösungen Polyamid (PA) in konzentrierter Lösung

/N25/ DIN 53 728 Blatt 3 Bestimmung der Viskositätszahl von Polyethylenterephthalat (PETP) oder Polybutylen- terephthalat (PBTP) in verdünnter Lösung

/N26/ DIN 53 728 Blatt 4 Bestimmung der Viskosität von Polyethylen (PE) und Polypropylen (PP) in verdünnter Lösung

/N27/ DIN 54 270 Bestimmung der Grenzviskosität von Cellulosen

2. International Organization for Standardization (ISO)

Viscometers

/N28/ ISO 3105 Glass capillary kinematic viscometers Specification and operating Instructions

Petroleum Products

/N29/ ISO 3104 Transparent and opaque liquids Determination of kinematic viscosity and calculation of dynamic viscosity

/N30/ ISO 3448 Industrial Liquid Lubricants ISO Viscosity Classification

Plastics

/N31/ ISO/R 175 Determination of Viscosity Number of Polyvinylchloride Resin in Solution

/N32/ ISO/R 307 Polyamides Determination of Viscosity Numbers

/N33/ ISO/R 600 Determination of Viscosity Ratio of Polyamides in concentrated Solution

/N34/ ISO/R 1157 Cellulose Acetate in dilute Solution Determination of Viscosity Number and Viscosity Ratio

/N35/ ISO/R 1191 Determination of Viscosity Number and Limiting Viscosity Number of Polyethylenes and Polypropylenes in dilute Solution

/N36/ ISO/DIS 1228 Determination of Viscosity Number of Alkylene Terephthalate Polymers and Copolymers in dilute Solution

/N37/ ISO/R 1336

Determination of Viscosity Number of Methylmethacrylate Polymers and Copolymers in Solution

/N38/ ISO/R 1599 Determination of Viscosity Loss on Moulding of Cellulose Acetate

/N39/ ISO 1628 / 1 Guidelines for the Standardization of Methods for the Determination of Viscosity Number and Limiting Viscosity Number of Polymers in Dilute Solution - General Conditions

/N40/ ISO 1628 / 2 Plastics - Determination of Viscosity Number and Limiting Viscosity Number Part 2: Poly (Vinyl Chloride) Resins

/N41/ ISO 1628 / 3 Plastics - Determination of Viscosity Number and Limiting Viscosity Number Part 3: Polyethylenes and Poly Propylenes

/N42/ ISO 1628 / 4 Plastics - Determination of Viscosity Number and Limiting Viscosity Number Part 4: Polycarbonate (PC) Moulding and Extrusion Material

/N43/ ISO 1628 / 5 Determination of Viscosity of Polymers in Dilute Solution using Capillary Viscometers Part 5: Thermoplastic Polyester (TP) Homopolymers and Copolymers

/N44/ ISO 1628 / 6 Determination of Viscosity Number and Limiting Viscosity Number Part 6: Methyl methacrylate polymers

3. American National Standard (ASTM)

Basics

/N45/ D 445 - 88 Standard Test Method for Kinematic Viscosity of Transparent and Opaque Liquids (and the Calculation of Dynamic Viscosity)

Viscometers

/N46/ D 446 - 89 a Standard Specifications and Operating Instructions for Glass Capillary Viscometers

/N47/ D 2515 - 86 Standard Specifications and Operating Instructions for Glass Capillary Kinematic Viscometers

Plastics

/N48/ D 789 Standard Test Methods for Determination of Relative Viscosity, Melting Point, and Moisture Content of Polyamide (PA)

/N49/ D 1243 Standard Test Method for Dilute Solution Viscosity of Vinyl Chloride Polymers

/N50/ D 1601 Standard Test Method for Dilute Solution Viscosity of Ethylene Polymers

/N51/ D 2393 Standard Test Method for Viscosity of Epoxy Resins and Related Components

/N52/ D 4603 Standard Test Method for determining inherent Viscosity of Poly (ethylene terephthalate) (PET)

/N53/ D 4878 Standard Test Method for Polyurethane Raw Materials: Determination of Viscosity of Polyols

4. British Standard (BS)

/N54/ BS 188 Methods for Determination of the Viscosity of Liquids

5. Norme Française Enregistrée

/N55/ NF T 60 - 100 Mesure de la viscosité cinématique

6. Normen zu Flüssigkeits-Thermostaten

/N56/ DIN 58 966, Teil 1, Thermostate, Allgemeine Begriffe

/N57/ DIN 58 966, Teil 2, Thermostate, Flüssigkeitsthermostate, Begriffe und Bestimmung der Kenndaten

These standards are valid as they are at the time of printing; newer standards are in process.

Manufacturer:

Schott-Geräte GmbH 55034 e 12990.2 Printed in Germany D-65701 Hofheim

SCHOTT GLAS Postfach 2480 Geschäftsbereich Industrieglas D-55014 Mainz Laborgeschäft Hattenbergstraße 10 D-55122 Mainz Telefon +49 (0)6131 / 66-4907 www.schott-in-the-lab.com Telefax +49 (0)6131 / 66-4051