Method development for rheological characterization of microfibrillated cellulose

(1)

Method development for

rheological characterization of

microfibrillated cellulose

Utveckling av en metod för reologisk karakterisering av mikrofibrillerad

cellulosa

Elin Wahlkrantz

Fakulteten för hälsa, natur- och teknikvetenskap Civilingenjörsprogrammet Kemiteknik

30 HP

Handledare: Björn Sjöstrand, Adrianna Svensson Examinator: Magnus Lestelius

(2)

i.

Abstract

This thesis contributes to a development of a method for rheological characterization of microfibrillated cellulose. The intended use of the method is to be able to distinguish between different grades of microfibrillated cellulose. The method that was developed had preparation procedure of suspensions, pH, dry content and conductivity as well as measuring geometry and measuring sequence in mind. The method resulted in using a propeller mixer for sample preparation and the most suitable properties of the samples for comparison of different qualities of microfibrillated cellulose was evaluated to be pH 8 with a dry content of 2.0 wt% and a conductivity of 110 µS/cm. The rheology of the microfibrillated cellulose suspensions was examined by using a dynamic rotational rheometer and a splined bob and cup (C25G/PC25G). The complex viscosity from amplitude sweeps is used as the parameter to distinguish between different grades of 2 wt% microfibrillated cellulose suspensions. At 1.0 wt% the pH of the suspensions appeared to have a very small impact on the results from rheological measurements while an increased conductivity of the suspensions resulted in an increased complex viscosity. The dry content dependency appeared to be exponential in the range of 0.5 to 3.0 wt% and it was thus easier to distinguish between different grades of microfibrillated cellulose when the dry content is 2.0 wt% compared to 1.0 or 1.5 wt%.

Keywords microfibrillated cellulose, MFC, rheology, dynamic rotational rheometer, complex

(3)

ii.

Executive summary

The aim of this thesis is to contribute to the development of a method for rheological characterization of suspensions of microfibrillated cellulose (MFC). The intended use of the method is to be able to distinguish between different grades, or qualities, of MFC. The method is developed with examination of the most suitable preparation procedure of the samples, measuring sequence, measuring geometry, dry content, pH, conductivity in mind. The rheological property that is most suitable to use to be able to distinguish between different grades of MFC is evaluated as well. To gain knowledge about new materials it is important to characterize their properties. For example, when developing new materials and when predicting how they behave in processing and/or handling of them it is important to rheologically characterize these materials. Rheology is the study of the flow of a material. The rheology of the materials can thus tell us a lot about how the material will flow and deform when exposed to shear forces. The rheological properties of a material will tell us how the material will behave in processes, during storage or during transport and what type of end products the material is suitable for.

MFC is a material with a lot of different possible applications. MFC is most often produced from a paper pulp which can be produced from hardwood or softwood with several different pulping processes such as kraft, sulfite or thermomechanical pulping. The cellulose fibers from the pulp are fibrillated millions of times and creating a three-dimensional network of fibrils. Due to the fibrillation the material has a great surface area. The material can be used in fiber material mixtures as strength enhancer, as a rheological modifier, as barrier material or as different types of specialty papers such as filter paper etc. MFC suspensions is behaving shear-thinning and thixotropic and the rheology of the suspension can be affected by parameters such as aspect ratio of the fibrils, the morphology of the surface of the fibrils and crowding number which in turn is affected by both dry content and aspect ratio, pH as well as surface charge and conductivity.

When examining the most suitable sample preparation procedure several mixers, containers, propellers, rotational speeds and mixing times was evaluated. The best suited mixer was chosen to be a propeller mixer, and to use a container with a diameter as close to the propeller diameter as possible. This is to ensure favorable mixing of the suspension and thus obtain a suspension that is as homogenous as possible. Two propellers were evaluated and the propeller with propeller blades with a smaller surface area and a larger angle from the horizontal plane appeared to be advantageous to use since the suspension did not cling to the propeller blades to the same extent when using such a propeller. The rotational speed that was chosen is 1400 rpm, which was fast enough to obtain a mixed, homogenous sample when mixing was performed for 10 minutes when samples up to 3.0 wt% were prepared. If using a higher rpm some of the suspension was splashing out of the container and more air appeared to enter the suspension. If using a lower rpm, the obtained suspensions was not as homogenous.

(4)

iii. to obtain reliable results. To make the method practical, the samples is thereby left to rest overnight, for 20 hours to be exact.

The best suited measuring geometry appears to be the splined bob and cup (C25G/PC25G). The loading of the sample is performed carefully after a slow mixing of the sample performed with a metal spoon. The mixing is performed by circulating the spoon slowly 5 times to the left and 5 times to the right. This mixing is performed to stir up eventual sedimented fibrils from the bottom of the container to make sure the samples are homogenous while still affecting the rheology of the samples as little as possible. Four full spoons are then loaded to the measuring cup which results in a sample of a total of 16-20 g in the measuring cup.

The effect of pH is examined for pH values between 2.00 and 12.00. Even though the span is large, the effect appears to be small. There seems to be almost no effect of the pH for 1.0 wt% samples which is unexpected if looking at other results found in the literature. A possible explanation to this is that the surface charge or conductivity “knocks out” the effect of pH. Another possible explanation could be that the process or raw material used when manufacturing the MFC samples differ between the studies found in the literature and the MFC samples examined in this thesis. It could thus be interesting to examine the effect of pH for MFC qualities prepared by using different raw materials and manufacturing processes.

The best suited dry content to use when performing rheological measurements of MFC by using a dynamic rotational rheometer appears to be 2.0 wt% compared to 1.0 or 1.5 wt%. At 2.0 wt%, it is still possible to use the splined bob and cup (C25G/PC25G) as measuring geometry. Moreover, the difference between results from the rheological measurements between the samples is higher when the dry content is higher and it is thus easier to distinguish the results from the different samples from each other.

(5)

iv.

Övergripande sammanfattning

Syftet med denna uppsats är att bidra till utvecklingen av en metod för reologisk karakterisering av suspensioner med mikrofibrillerad cellulosa (MFC). Tanken är att metoden skall användas för att kunna särskilja olika kvalitéer av MFC. Metoden utvecklades genom att undersöka och ta fram en lämplig provförberedelseprocedur med avseende på mätsekvens, mätgeometri, torrhalt, pH och konduktivitet. Vilken egenskap som är lämpligt att använda för att kunna särskilja mellan resultaten från mätningar av olika kvalitéer av MFC utvärderas också. För att få kunskap om nya material är det viktigt att karakterisera egenskaperna hos materialet. Till exempel när nya mterial utvecklas och framställs eller då man vill kunna förutspå hur ett material kokmmer att bete sig vid tillverkning eller hantering av materialet är det viktigt att reologiskt kunna karakterisera materialet. Reologi är läran om flödesegenskaper hos ett material. Ett materials reologiska egenskaper kan därför berätta mycket om hur materialet kommer att flöda och deformeras när det utsätts för skjuvkrafter. De reologiska egenskaperna kommer även att ge information om hur materialet beter sig i olika processer, vid lagring eller under transport samt vad som skulle vara passande slutprodukter.

MFC är ett material med många olika möjliga tillämpningar. MFC tillverkas oftast från pappersmassa vilket i sin tur kan vara tillverkad av barrved eller lövved med olika tillverkningsprocesser såsom sulfat, sulfit eller termomekanisk massaframställning. Cellulosafibrerna i massan fibrilleras miljontals gånger och ett tredimensionellt nätverk av fibriller skapas. Till följd av fibrilleringen får materialet en mycket stor ytarea. MFC kan användas i fiberblandningar för att ge materialet en ökad styrka, det kan även användas som reologisk modifierare, som barriärmaterial eller till olika typer av specialpapper såsom filterpapper etc. MFC suspensioner uppvisar ett skjuvtunnande och tixotropiskt beteende och reologin hos suspensionerna påverkas av parametrar såsom förhållandet mellan bredd och längd hos fibrillerna, fibrillernas morfologi samt hur troligt det är att fibrillerna krockar och på så sätt skapar kontakt mellan varandra (eng. ”crowding number”). Detta i sin tur påverkas av både torrhalt och fibrillernas dimensioner samt pH, ytladdning och konduktivitet.

Vid undersökningen av vilken förberedelseprocedur som skulle vara lämplig undersöktes olika omrörare, behållare, propellrar, rotationshastigheter samt tider för omrörning. Den omrörare som passade bäst var en propelleromrörare i kombination med en behållare som ger en så liten glipa mellan propeller och behållare som möjligt. Den här kombinationen gav en fördelaktig omrörning av suspensionerna och en så homogen suspension som möjligt kunde då erhållas. Två olika propellrar undersöktes och den propeller vars propellerblad hade en mindre ytarea och en större vinkel från det horisontella planet ansågs vara fördelaktig att använda då suspensionen inte fastnade på propellerbladen i lika stor usträckning. Rotationsastigheten som användes var 1400 rpm, vilket var nog snabbt för att erhålla ett homogent, slätt prov då omrörningen utfördes under 10 minuter när prover med en torrhalt upp till 3.0 wt% förbereddes. Om en högre rotationshastighet användes skvätte suspensionen mer och mer luft verkade komma in i suspensionen. En lägre rotationshastighet i sin tur gav inte lika homogena suspensioner.

(6)

v. uppvisas tydligt i resultaten som visar en tydlig regenerering av viskositeten hos proverna. Dessa resultat understryker att alla prover behöver vilka lika länge mellan förberedning och mätning för att erhålla tillförlitliga resultat. För att metoden ska vara praktiskt genomförbar och så enkel som möljigt att använda får proverna vila under natten, under 20 timmar mer exakt.

Den mätgeometri som passar bäst är den räfflade cylindergeometrin tillsammans med en räfflad mätkopp (C25P/PC25G). Provet laddas försiktigt efter en långsam omrörning som görs för hand med hjälp av en metallsked. Omrörningen genomförs genom att röra skeden 5 gånger åt vänster, och därefter 5 gånger åt höger. Omrörningen är viktigt för att röra upp eventuellt sedimenterade fibriller från botten av behållaren och för att på så sätt säkerställa att proverna fortfarande är homogena medans man försöker påverka provernas reologi i så liten usträckning som möjligt. Fyra hela skedar prov laddas därefter till mätkoppen vilket resulterar i ett prov med storleken 16-20 g.

Effekten av pH undersöks mellan 2.00 och 12.00. Trots att det är ett stort spann verkar effekten vara mycket liten. Det är i princip ingen förändring av den komplexa viskositeten för proverna med 1.0 wt% vilket är oväntat då studier i litteraturen påvisat andra resultat. En möjlig förklaring till detta skulle kunna vara den höga konduktivteten i proverna som i sin tur skulle kunna ”slå ut” effekten av pH. En annan möjlig förklaring kan vara att tillverkningsprocessen eller råmaterialet som använts vid tillverkningen av MFC:n kan skilja sig mellan studierna som kunde hittas i litteraturen och MFC:n som använts i den här studien. Det skulle därför kunna vara intressant att undersöka effekten av pH för MFC kvalitéer som har tillverkats med olika råmaterial och tillverkningsprocesser.

Då de reologiska egenskaperna hos MFC undersöks med hjälp av en dynamisk rotationsreometer verkar 2.0 wt% vara mer passande att använda jämfört med 1.0 eller 1.5 wt%. Då torrhalten är 2.0 wt% är det fortfarande möjligt att använda den räfflade cylindern och mätkoppen som mätgeometri, och skillnaden mellan resultaten från de reologiska mätningarna av de olika proverna är större. Det är därmed enklare att särskilja resultaten för de olika kvalitéerna från varandra.

(7)

vi.

List of symbols and abbreviations

Symbols

G Shear modulus (Pa) G’ Storage modulus (Pa) G’’ Loss modulus (Pa) N Crowding number (-)

Greek Letters

τ Shear stress (Pa) γ Shear strain (-) γ̇ Shear rate (s-1₎

η Shear viscosity (Pa s) η* Complex viscosity (Pa s) ζ Zeta-Potential (mV)

υ Kinematic viscosity ω Angular frequency (rad/s)

Abbreviations

AFM Atomic force microscopy CSS Controlled shear stress CSR Controlled shear rate CNC Cellulose nanocrystals CNF Cellulose nanofibrils MFC Microfibrillated cellulose NFC Nanofibrillated cellulose PD Pressure difference

(8)

vii.

1 Introduction

There is an increasing demand of products and materials, made in a sustainable way using renewable and sustainable raw materials and manufacturing processes. This demand is what can direct the economy towards a more sustainable and circular bioeconomy. Development of new biomaterials can easily be connected to several of the United Nations Sustainable

Development Goals such as decent work and economic growth, industry, innovation and

infrastructure, sustainable cities and communities, responsible consumption and production, climate action as well as life on land. Using biomaterials, instead of fossil-based materials, can reduce the emissions of carbon dioxide and thus slow down the global warming. A contribution to this solution, could be the forest industry. A growing forest will bind carbon during its whole lifecycle, this is the case for all forest industry products as well as the trees themselves [1,2].

Forestry industry products can be described as biodegradable, renewable, carbon neutral and safe for human and animals. Cellulose rich materials have been used for thousands of years but when the demand after products and materials for new applications and with new properties is increasing, the development in the industry needs to keep up. One material from the forestry industry, is microfibrillated cellulose (MFC). MFC is most often produced from paper pulp made from some kind of wood raw material (softwood or hardwood) and can be prepared from several pulping processes (kraft, sulfite or thermomechanical pulping) [3,4]. The cellulose fibres are fibrillated millions of times, creating a large three dimensional network with a very large surface area. MFC can be mixed with fiber based materials or used as water based suspensions with purposes such as increasing strength or lower the weight of a material, as rheological modifier or as barrier material [5].

To gain knowledge about materials, the properties of the materials needs to be characterized. The results from this thesis, will contribute to the development of a method for rheological characterizing of MFC. The aim of this thesis is to examine how to prepare the suspensions for rheological characterizations, the effect of pH, dry content as well as the effect of the conductivity of the samples. Another important aim is to examine which rheological property to look at to be able to distinguish between different grades of MFC. All this is described in this thesis.

1.1 Aim and research questions

The aim of this thesis is to develop a method for rheological characterization of MFC. Characterization of a material is important to be able to further interpret and predict the behavior of a material. A reproducible and easily performed method is thus valuable to when it comes to efficient work. The thesis will address the following research questions:

• What is a suitable preparation procedure for samples of MFC for rheological measurements?

• What type of spindle geometry is suitable when performing rheological measurements on a sample of MFC?

• If and how will pH, conductivity and dry content of the prepared sample of MFC affect the results from rheological measurements?

(11)

2.

2 Literature background

2.1 Rheology

Rheology is the study of flow and deformation of a material or matter. Examples of rheological properties is viscosity, yield stress and elastic behavior of materials. The rheological properties will be further described in this literature background. When characterizing properties of a material, rheology can be interesting to examine since knowledge about the rheological properties of a material facilitates the understanding and prediction of the materials behavior in different scenarios. The rheological properties are of high importance for the handling, processing and usage of the material. Many of the everyday materials we see are used due to their rheological behavior. This makes rheology an important field to study when developing new materials. It is important to be aware that the intermolecular forces of a material are affecting the rheology and thus the behavior of the material. In summary, the rheological properties possessed by a material can give insights about a materials behavior, limitations and capacity. It is thus very important to examine these properties when characterizing a material [6]. As mentioned, the rheological properties of a material are also critical in handling of the material in processes such as pumping, mixing, storage and extrusion etc. [7]. The study of rheology is not only applicable to pure materials but also to solutions and dispersions. The texture, flow properties or deformation of a material when exposed to stress or strain are all parts of the materials rheological properties [6,8]. The flow behavior of a material can be described by shear and extensional flows that are properties connected with the viscosity of the material. The elasticity is characterized by swell ratios or modulus of the material. To examine these properties to an fully extent a rheometer can be used [9]. The rheological properties of materials with known surface chemistry most often can be predicted accurately. Due to the complex nature, and lack of knowledge, of the surface chemistry of nanocellulose such as MFC, it is still hard to predict the rheological behavior completely [10].

2.1.1 Flow behavior and viscosity

2.1.1.1 Shear

Shear means that parallel, adjacent layers are moving relative to each other. For example, when an object is pushed over a surface a movement between parallel layers occurs which is illustrated in Figure 1. This is the same principle that occurs when a material is exposed to shear forces. The force which can be applied to a matter can be described as stress or strain. Depending on the material, it could be more suitable to control the stress or the strain load that is applied to the material. For example when the material is not a pure solid it could be preferred to control the rate of the strain, i.e. the shear rate [11].

(12)

3.

2.1.1.1.1 Shear stress

Stress is defined as the force applied to a surface divided by the area over which it is applied. One kind of stress is shear stress. Shear stress (τ), which is also denoted by σ in some literature, occurs when parallel layers of a matter successively move in the plane of shear due to applied stress, i.e. the force that needs to be applied to obtain a certain deformation. During rheological measurements an instrument will apply the force to the sample via a measuring geometry which will make the sample move [6]. How the shear stress can be calculated is described by Equation 1.

Equation 1 𝜏 =𝐹

𝐴

τ = shear stress (Pa) F = applied force (N) A = area (m2)

2.1.1.1.2 Shear rate

Shear strain (γ), is another type of shear force. Shear strain describes the relative in-plane movement of two layers which are parallel to each other. Shear strain can be calculated by the ratio of the deformation of a matter in one direction and the length of the perpendicular axis. If the material is placed between two parallel plates where one of the plates is moving the length of the perpendicular axis is simply the distance between the plates which is illustrated in Figure 2 and described by Equation 2.

Figure 2. Illustrates the deformation (u) and distance between the plates (h) which is used to calculate the shear strain and shear rate described by Equation 2 [8]. Equation 2 𝛾̇ = 𝑣 ℎ (s -1₎ 𝛾̇ = shear rate (s-1₎ v = velocity of u (m/s)

h = distance between plates (m)

(13)

4. time. The instrument used will apply a rotational speed to the sample via a measuring geometry [11].

2.1.1.2 Shear viscosity

Shear viscosity, η (Pa s), is one of the most commonly described rheological properties. When it comes to fluids, all molecules are moving relative to each other. This relative motion is combined with frictional internal forces which creates a flow resistance. This flow resistance is defined as the shear viscosity of the fluid or the soft solid [8].

2.1.1.2.1 Newtonian fluids and Newton’s law

When a fluid is showing an ideally viscous flow it is called a Newtonian liquid. The shear viscosity is defined as the ratio of the shear stress and shear rate, this relationship is called Newton’s law which is described by Equation 3. If the relationship between the shear stress and shear rate is linear the liquid is by definition Newtonian [8].

Equation 3 𝜂 = 𝜏

𝛾̇ (Pa s)

η = shear viscosity (Pa s) τ = shear stress (Pa)

𝛾̇ = shear rate (s-1₎

2.1.1.2.2 Non-Newtonian fluids

(14)

5.

Figure 3. Illustrates how the change in viscosity depending on shear rate for a Newtonian, a shear thickening and a shear thinning material [8].

2.1.1.3 Time-dependent flow behavior

Thixotropy and rheopexy are time-dependent rheological behaviors. Most dispersions, i.e. suspensions, foams and emulsions such as gels, ketchup, drilling fluids, pastes and creams are thixotropic [8]. Thixotropy is a reversible decrease in viscosity at a particular shear load. Due to shear load the structure of the material is broken down over time. Conversely, a material is rheopectic if the structural strength at first is increasing for a period of time when exposed to a higher shear load. When the fluid is at rest the structure is instead broken down [8,11].

To determine the thixotropy or rheopexy of a material both the decomposition and the regeneration of the structure needs to be taken into consideration, a typical viscosity curve for showing thixotropic behavior and rheopectic behavior is shown in Figure 4. During a measurement a constant shear load is used while the time is the variable factor [8].

(15)

6.

2.1.2 Elastic behavior

The definition of a material’s elastic behavior is the ability of deformation due to applied stress and then regaining the original shape and size [12]. The elastic behavior of a material can be characterized by parameters as yield point and the shear modulus of the material [8].

2.1.2.1 Yield point

The yield point can also be called the yield stress or the elastic limit, i.e. when a solid material goes from elastic behavior to plastic behavior. When a material changes from elastic to plastic behavior the material can no longer return to its previous form but have undergone a permanent change of shape or size [11]. In other words, if a material is by definition plastic the characteristic yield point needs to be present. When it comes to fluids the yield point is the point where the externally applied forces are greater than the internal forces. I.e. if the applied force exceeds this yield point a flow will occur. This is seen in everyday lives in for example toothpaste tubes or ketchup containers [8].

2.1.2.2 Ideally elastic behavior

The shear modulus of an elastic material can be described as the ratio of the shear stress and the shear strain of the material, which is known as Hooke’s law and described in Equation 4. If a material is showing a linear correlation between the shear stress and the shear rate the material is showing ideally elastic behavior. This ideally elastic behavior is also called Hookean [8].

Equation 4 𝐺 =𝜏

𝛾

G = shear modulus (Pa) τ = shear stress (Pa) γ = shear strain (-)

2.1.3 Viscoelastic behavior

(16)

7.

2.1.3.1 Storage modulus (elastic modulus)

The storage modulus or the elastic modulus which is denoted by G’ (Pa), represents the elastic part of the viscoelastic behavior of a material. The value of the storage modulus is a measure of the deformation energy stored by the sample during the shear process. The elastic portion of energy which is stored in the deformed material will, when the material is released, function as a driving force to regain the original shape of the material [8].

2.1.3.2 Loss modulus (viscous modulus)

The loss modulus or the viscous modulus which is denoted by G’’ (Pa), represents the viscous part of the viscoelastic behavior of a material. The value of the loss modulus is a measure of the deformation energy used up by the sample during the shear process. The viscous portion of energy is used up when the material is deformed. This energy will thus be lost from the material when the material is deformed [8].

2.1.3.3 Complex viscosity

The complex viscosity, η* (Pa s), of a material is measured using oscillatory tests compared to “common” shear viscosity, η (Pa s), which is measured using constant shear conditions. The complex viscosity consists of a real and an imaginary part. The real part (η’) of the complex viscosity describes the materials viscous behavior while the imaginary part (η’’) of the complex viscosity describes the materials elastic behavior. The complex viscosity can thus be described as shown in Figure 5 and by Equation 5 and Equation 6 [8].

(17)

8. Equation 5

𝜂

′

=

𝐺′′

𝜔

η' = the viscous part of the complex viscosity (Pa s) G’’ = the loss/viscous modulus (Pa)

ω = angular velocity of measuring geometry (s-1₎

Equation 6

𝜂′

′

=

𝐺′ 𝜔

η'’' = the elastic part of the complex viscosity (Pa s) G’ = the storage/elastic modulus (Pa)

ω = angular velocity of measuring geometry (s-1₎

2.2 Measuring systems

Rheological measurements can be performed to obtain information about rheological properties of different materials. A viscometer is most useful when the fluid is Newtonian and thus exhibits an ideal viscous behavior that is easier to anticipate and interpret. A rheometer will, compared to when using a viscometer, acquire a lot more information, even regarding non-Newtonian fluids. When using a rheometer the measurement can not only be speed- or shear rate-controlled but can also be torque- or shear stress-controlled. When using a rheometer additional information about rheological properties of a material can be obtained which is not the case when using a viscometer [8].

There are several different viscometers available and several different designs of rheometers exist as well. The modern equipment can be divided into two main groups of rheometers, i.e. dynamic rotational rheometers and capillary extrusion rheometers. The capillary rheometer is mostly used for relatively concentrated suspensions while a dynamic rotational rheometer is mostly used for less concentrated suspensions [9]. Another newer approach is to use a pipe-rheometer with ultrasound velocity profiling and pressure difference measurements (UVP-PD) [13].

2.2.1 Viscometers

2.2.1.1 Capillary viscometers

(18)

9.

2.2.1.2 The falling ball method

Another type of viscometer that is using gravity is the falling ball method. As the name implies a ball is falling through the fluid and viscosity values can be determined using the travel time and the falling distance This simple type of viscometers is not as accurate as more advanced instruments [8].

2.2.1.3 Rotational viscometer

A rotational viscometer will give information about the viscosity of the material, i.e. the resistance to deformation at a certain rate. There are several different manufacturers and brands providing rotational viscometers but the principle used is the same. A rotational viscometer uses a rotating spindle to exert a shear tension to the fluid in a container. By using the rotating spindle, the speed and shear rate exerted to the fluid can be controlled [8,14].

2.2.2 Rheometers

2.2.2.1 Dynamic rotational rheometer

A dynamic rotational rheometer consists of a rotating shaft which is driven by an integrated motor. Measurements can be performed using rotating spindles with different geometries. The geometries can be rod-shaped or a vane in a concentric cylinder (“bob in a cup” and “vane in a cup”). Another type of a measuring set-up can consist of two parallel plates, i.e. plate-plate, or a cone-plate, alternatively a cone-cone set-up which rotates or oscillates [8]. The surfaces can be smooth or have a roughened or serrated surfaces to get a better grip of the sample and thus avoid wall slip [15]. Depending on the viscosity of the sample different diameters of the plates can be used if plate-plate or plate-cone set-up is used. If a plate-plate set-up is used the gap between the plates can be modified to suit the specific sample. A general rule of thumb is to use a gap which is 10 times the size of the largest particle in the sample, but a gap which is 5 times the size of the largest particle in the sample could also work. However, this requires one to know the particle size in the sample. Another well approved gap when working with gels and dispersions is 0.5 to 1.0 mm which does not require one to know the size of the particles in the sample [8]. A dynamic rotational rheometer is preferably used when viscosity and viscoelastic properties is measured. The measurements is most commonly performed in low shear rate ranges, typically from 0.001-100 s-1 [16].

2.2.2.1.1 Wall depletion (wall slip effect)

(19)

10.

2.2.2.2 Capillary rheometer

In a capillary rheometer a sample is forced through a well-defined opening under high pressure. The opening can be a barrel or a die-opening [8]. When measuring the viscosity of both Newtonian and even some non-Newtonian fluids the capillary methods is considered to be the most accurate. Another advantage is that the set-up can be rather simple and not as expensive as a dynamic rotational rheometer. These kind of extrusion method is mostly used when the material is a rather soft solid [9].

2.2.2.3 Pipe rheometer using UVP-PD

Some opaque and flocculated suspensions with high viscosity even at low mass concentrations can be hard to examine using conventional rheometers. An option to the conventional rheometers is to use a pipe rheometer with ultrasound velocity profiling (UVP) and pressure difference measurements (PD). This method is non-invasive and there is no need of assumptions about the material, i.e. the method is solely based on directs measurements. The set-up of the rheometer can consist of two chambers connected by a measurement tube and a pump. The bottom chamber is filled with the fluid which is then pumped to the upper chamber. The measurement is then started by opening the valve in the measurement tube connecting the two chambers. The flow can be gravitational and controlled by an air valve. Another option for the flow is to be driven by overpressure in the upper chamber which can be controlled by a pressure regulator [13].

2.3 Cellulose

Cellulose is a common organic polymer on Earth and it is also described to be an almost infinite source of material to environmentally friendly products. Cellulose is most often found in trees and plants but can also be found in algae or tunicates or it can be produced by bacteria. The structure of cellulose is alternating between crystalline and amorphous regions [4,18]. Native cellulose consists of D-glucose units that are linked together by β-1,4-glycosidic bonds. Multiples of this glucose units are connected together in a total of 10’000-15’000 units. The β-1,4-glycosidic bond which is repeatedly occurring between the glucose units connects the C4 of one of one unit with the C1 of the next unit which is schematically shown in Figure 6.

Due to the β-1,4-glycosidic bonds, the cellulose chain is linear. Moreover, the cellulose chain is stabilized by hydrogen bonds between the hydroxyl groups and the oxygens of adjacent chains [4].

(20)

11.

Figure 6. Shows the structure of native cellulose and the β-1,4-glycosidic bonds between the C1 and C4 of the different glucose units [19].

2.3.1 Microfibrillated cellulose (MFC)

MFC is produced from cellulose rich materials such as paper pulp. There are several different names used for similar materials such as nanocellulose, nanofibrillated cellulose (NFC), cellulose nanocrystals (CNC), cellulose nanofibrils or cellulose nanofibers (CNF). These materials are similar to each other but the definitions are not perfectly clear. Generally the term “nano” implies a smaller size range than the term “micro”. In this thesis MFC will be used to describe the highly fibrillated cellulose material prepared by mechanical processing which can be accompanied by chemical or enzymatic treatments in accordance with the definition by Hubbe et al [10]. There are several different possible production processes for MFC and the closely resembling materials (NFC, CNC, CNF, nanocellulose etc.) which will be described shortly. Since the rheological properties of a material are important to understand the behavior and function of the material some of MFC’s behavior will be briefly described as well.

2.3.1.1 Production of microfibrillated cellulose

(21)

12. surfaces. The fibrils surface charge is mostly affected by the carboxylic groups on the hemicelluloses and using a kraft pulp as raw material when producing MFC will thereby affect the surface charge of the fibrils as well as the rheological behavior of the suspensions [3].

The pretreatment used to lower the energy consumption could be a chemical treatment, or enzymatic. A chemical treatment could be to hydrolyse the fibers using sulfuric (H2SO4) or

hydrocholoric (HCl) acid. Another pretreatment is oxidation using TEMPO (2,2,6,6-tetramethylpiperidine-1-oxyl) or to use enzymes as endoglucanase. When using different pretreatments, the functional groups on the surface of the fibers can be varied. The purpose of the pretreatment is to lower the time- and energy needed when the fibrillation and separation of the cellulose fibers are performed. This is done by weakening or breaking the connections between the fibers [22,23].

After the pretreatment of the fibers, it is time for the fibrillation and separation. This could be performed by mechanical grinding, cryocrushing, homogenization, microfluidization or ultrasonication [24]. The mechanical grinding would mean that the hydrogen bonds and the walls of the cellulose fibers is grinded between a rotating grindstone and a grindstone that doesn’t move. This repeated movement makes the fibers separate. Cryocrushing means that liquid nitrogen is used to freeze the cellulose suspension. Thereafter the frozen cellulose suspension is crushed to separate the fibers. This method is not suitable for large scale production [22]. Homogenization means that the cellulose suspension is forced through a container using very high pressure. In this container the material will be exposed to both impact and shear forces due to the high pressure [24]. Microfluidization uses a principle kind of similar to homogenization since a chamber is exposed to a high pressure caused by a pump. In this chamber the material is exposed to impact and shear forces which performs the fibrillation and separation of the fibers. Ultrasonication means that oscillating ultrasonic waves helps forming microscopic gasbubbles between the fibers which then expands. These bubbles will then collapse due to a high pressure and the fibrils can be separated from each other [20,22,24].

The possible uses of MFC are wide. It could be as a replacement of materials made from fossil-based raw materials, as a rheological modifier or strength enhancer. MFC could be used to produce different packaging, such as packaging for liquids, take away products, food packaging, or as barrier layer, special filter paper, coatings or even glue [25].

2.3.2 Rheology of MFC

(22)

13.

2.3.2.1 Morphology of MFC

The dimensions of MFC particles is estimated to be around 10-100 nm wide and 0.5-100 µm long, the aspect ratio it thus often between 50-100 [27]. The methods used to measure the diameter are most often microscopic methods such as atomic force microscopy (AFM), transmission electron microscopy (TEM) or scanning electron microscopy (SEM). Due to the high aspect ratio and entanglement of fibrils, it can be hard to determine the dimensions of the fibrils. It’s due to the fact that the microscopic images needs to be magnified to be able to determine the fibril diameter. When the images are greatly magnified, the whole length of the fibril will not fit in the image. If instead the whole length is fitted in the image, it is impossible to distinguish between the different fibrils due to the entanglement. The fibrils are thus too long, thin and possibly entangled which makes it hard to measure the length of the fibrils exactly. Even though it can be hard to measure the length of the fibrils exactly by using imaging methods, they are normally estimated to be several micrometers long which gives the fibrils a high aspect ratio. The diameter and length of the MFC fibrils differs depending on the raw material and manufacturing processes used. The surface of the fibrils will be irregular since the fibrils are partially fibrillated. The surface chemical composition of MFC will be affected by the choice of raw material as well as manufacturing processes and pretreatments [3].

2.3.2.2 Forming of networks and crowding number

The MFC fibrils will already be entangled after the manufacturing and if the fibrils are exposed to shear-flow there can be further entanglements and connections between the fibrils. When a solution consists of solid particles, like fibrils in a liquid, the rheological behavior of the liquid will change. One reason behind the change in rheological behavior is the collisions between the particles due to shear flow [28] but also due to hydrodynamic disturbance of the flow field [3]. When it comes to papermaking and conventional pulp there is a “crowding number”, N, which is used to estimate the number of collisions or contacts between the fibrils in a solution. How to calculate the crowding number is described by Equation 7. This is interesting to know since the number of contacts between the fibrils give information about the eventual flocculation behavior of the solution.

Equation 7 𝑁 =2 3𝐶𝑣( 𝐿 𝑑) 2 N = crowding number (m-3) Cv = volumetric concentration (m-3) L = fiber length (m) d = fiber diameter (m)

(23)

14. between the fibrils will occur when the concentration is high enough [3,28]. When the fluid has a low viscosity, the individual fibrils don’t align with the flow to the same extent as when the viscosity of the fluid is higher. This means that the fibrils won’t be aligned but instead will have a higher tendency to hook each other when the viscosity of the fluid is low. Due to the irregular surface of the fibrils, the hooking will also be further favored [3]. The viscosity of MFC is connected to concentration or dry content. The viscosity is also a time- and shear-dependent property of MFC.

2.3.2.3 Flocculation

If the fiber concentration is varying in the suspension there will be fiber networks forming alternately. These smaller networks are called flocs and the phenomenon is known as flocculation. The formed flocs are not permanent though, but there will be a constant formation and breakdown of flocs. Besides the mechanical forces described in 2.3.2.2 Forming of networks and crowding number there are also colloidal forces that affects the rheological behavior and the flocculation of suspensions [3]. The colloidal forces can be described by the Derjaguin Landau Verwey Overbeek (DLVO) theory. The theory describes interactions between particles. The interactions considered are two types of forces, van der Waals forces and electrostatic double layer forces. It is important to mention that the theory was developed for pair interactions and not multiparticle interactions which can occur if the solution is concentrated [29].

If flocculation of the suspension occurs due to mechanical entanglement the contact between the fibrils is an important factor [3,28]. The flocculation between fibers, at least in pulp suspensions, can to a large extent be explained by mechanical forces and not due to the weak Van der Waals forces when there are no additives added to the suspension. This is depending on the size of the surface area though. MFC which has a bigger surface area than cellulose fibers themselves will probably be more affected by the electrostatic forces and not only mechanical forces [28,30].

2.3.2.4 Gel formation

When MFC is suspended in water the suspension is usually not transparent but opaque. This is due to the network of fibrils formed when the fibrils are entangled and hooked to each other [3]. The cellulose fibers have several hydroxyl-groups on their surfaces. The MFC fibrils will, due to these hydroxyl-groups and their large surface area, be able to bind large amounts of water. Since the fibril network can bind large amounts of water, the suspension will be gel-like [20]. The MFC suspensions are often in this range of concentrations where a gel network is formed at rest. A gel suspension is elastically dominated which means that the storage modulus is bigger than the loss modulus [31]. The point which is when a solution turns into a gel is reached when the loss factor, 𝑡𝑎𝑛𝛿 is equal to 1. The loss factor is defined as the ratio between the loss modulus and the storage modulus as shown in Equation 8 [8].

Equation 8 𝑡𝑎𝑛𝛿 =𝐺′′

𝐺′

(24)

15. If 𝑡𝑎𝑛𝛿 > 1 the suspension is a liquid. If 𝑡𝑎𝑛𝛿 < 1 suspension is gel-like. When 𝑡𝑎𝑛𝛿 = 1 or 𝛿 = 45° the gel formation is taking place [8]. The gel point and the gel strength is dependent on the aspect ratio of the fibrils as well as the surface charge [3,31]. The concentration for gel formation of MFC can be rather low compared to other materials, in one study the concentration for gel formation is as low as 0.125 wt% which was measured by Pääkkö et al for a MFC suspension prepared from bleached sulfite softwood cellulose pulp and enzymatic treatment [32].

2.3.2.5 Thixotropic and shear-thinning behavior

MFC is exhibiting both thixotropic and shear-thinning behavior. This means that both a higher shear-load and a longer period of time of applied shear will result in a lowered viscosity of the sample. If the shear-load is removed, a material that is showing thixotropic properties, will regain its original viscosity (after resting for some time). This is due to a regeneration of the initial structure of the material, which due to the shear-load was broken down [31,33].

Shear-thinning behavior can occur when suspensions with elongated particles such as MFC are exposed to flow. The reason behind shear-thinning behavior can be that the elongated particles, or fibrils in this case, are aligning in the direction of flow when the shear force applied is large enough. When the particles are aligned in the direction of flow the entanglement of the particles can be diminished and the viscosity will thus be lowered. Due to the flocculation of fibrils, there can be other explanations but alignment of particles to the shear-thinning behavior of MFC. The flocs can due to the applied shear force be disintegrated which lowers the viscosity. This behavior can be seen in a typical viscosity curve of MFC where the viscosity is plotted against the shear rate. The viscosity is decreasing with an increasing shear rate (and shows shear thinning behavior) when the shear rate is in the lower region. In the middle region, the viscosity increases which is explained by flocculation. When the shear rate is increased even further, the viscosity decreases again, which can be explained by a disintegration of the flocs. This typical MFC viscosity curve can be seen in Figure 7 [34].

(25)

16.

2.3.2.6 Effect of pH

The pH of a solution will affect the dissociation of the negative groups on MFC surfaces. The carboxylic groups on the surface of cellulose-based materials will be in their protonated form when the pH of the solution is much lower than the pKa of the carboxylic groups. This leads to a strong network or clusters of MFC due to the lack of electrostatic forces between the fibrils. When the pH is higher than the pKa the carboxylic groups will be in their deprotonated form. This will lead to a negative charge of the surfaces and thus repulsion between the fibrils and a more dispersed material [10]. This dissociation of the carboxylic acid groups should thus lead to a decreasing viscosity of the MFC suspension. To sum up, a higher pH ought to give a lowered viscosity of a MFC solution and vice versa which have been observed in the range of pH 2-10. The dry content of the MFC was 0.25 wt% in this study [35]. Another study showed contradictory results, i.e. no effect of the viscosity for different pH-values was seen in 1.0 wt% MFC suspension [36]. A suggested explanation to this is that the lower concentration in the study when using 0.25 wt% makes the suspension more responsive to the addition of ions when adjusting the pH [35]. This will thus affect the viscosity of the suspension to a higher extent [10,36]. The pKa-value of the carboxylic acid groups of MFC is approximated in the literature to be 4.0 by some and 4.8 by others [10,37]. The pH could possibly affect the size of the formed flocs as well. At least this is the case when it comes to native cellulose fibers [30]. The floc size will affect the homogeneity of the sample and thus affect the results from rheological measurements. The floc size can affect the viscosity as well, since the number or form of interactions between the fibers are affected [34].

2.3.2.7 Effect of zeta-potential

The zeta-potential or the electrokinetic potential [38] is linked to the surface charge density and the surface potential. Measurements of the zeta-potential can give information about charged groups at the surface of the particles. The electrostatic repulsive forces between the particles, or fibrils, are often due to the presence of these charged groups. The zeta-potential can thus affect or at least give some information about the behavior of a particle suspension. The colloidal stability of cellulose based materials as MFC is connected to the repulsive forces between the particles. The colloidal stability is thus connected to the potential. The zeta-potential can also give some insight regarding the particles tendency to aggregate or form flocs [16].

Suspensions with a measured zeta-potential value greater +30 mV or less than -30 mV are generally considered stable. If the zeta-potential is big enough there is a sufficient mutual repulsion between the particles which results in colloidal stability. If the zeta-potential value is between -30 and +30 mV the solution is considered unstable and there is a risk for agglomeration [39,40]. If the zeta-potential is between -15 and +15 mV there is an impending risk of agglomeration, but even at these lower values the suspensions can be stable and thus not agglomerate. The stability of the suspensions will for these cases be due to steric stabilization or electrostatic stabilization at low ionic strength [41].

(26)

17.

2.3.2.8 Effect of conductivity

The conductivity of a solution is a measurement of the amount of electrolytes, or ions, in the solution. An increase of the conductivity, i.e. an increase of ions in the suspension, will decrease the repulsive forces between the fibrils. This decrease is a result of the compressed electrostatic double layer as well as decreasing the surface charge. When solutions containing salt such as NaCl are added to the suspensions, the fibrils could thus become more aggregated since they will not repel each other to the same extent. This will in turn affect the viscosity of the sample. When adding salt, the suspension could also become more gel-like, but this depends on the sample itself and other things such as shear history [10].

When the materials are prepared in different ways the response to the addition of salts will differ. The effect will also be different depending on what type of ions that are added to the suspension [10]. For example when a 0.1 wt% fibrillated cellulosic material prepared by using TEMPO-oxidation is exposed to monovalent electrolytes in form of NaCl in the concentrations of 10-400 mM, the viscosity of the suspension increases when the concentration is lower than or equal to 100 mM. When the concentration is increased even further, the viscosity is instead decreasing. This could be explained by the gel structure that was clearly noticeable up to 100 mM. When further addition of NaCl was performed, there was instead a clear phase separation in the suspension, thus indicating that agglomeration was occurring. When instead divalent electrolytes such as CaCl2 or MgCl2 was used, this aggregation occurred at

concentrations between 2-4 mM. This critical coagulation concentration of CaCl2 and MgCl2

differs from the critical coagulation concentration of NaCl but corresponds well to the critical coagulation concentration calculated by using the Schulze-Hardy rule that is supported by the DLVO theory [42]. Another study found in the literature is performed on a 0.5 wt% MFC suspensions after freeze drying. In this study NaCl or CaCl2 is added to the suspensions. The

results shows, as expected, an increase in viscosity due to the addition of ions. The effect was thus a bit larger when CaCl2 was added compared to when NaCl was added for the same final

concentrations of 0.05 and 0.1 M [36].

When looking at another cellulosic material, cellulose spheres prepared by using the viscose process and linter pulp of coniferous or broad-leaved trees, the forces between the cellulose spheres were purely repulsive when the particles were approaching each other. When instead the particles were separated there were no attractive forces between them either. This was the case for the cellulose spheres in both 0.1 and 10 mM KBr solutions. When the concentration of KBr was increasing, the magnitude of the repulsive forces was decreasing. This was explained by the interactions between the electrostatic double layers of the cellulose particles in the suspensions [43] which is the same behavior that can be observed for MFC particles in suspensions with added ions [10].

2.3.3 Wall depletion (wall slip effect)

(27)

18. content is estimated to have a thicker depletion layer and the thickness is also supposed to increase with an increasing shear rate. When the shear rate is around or lower than 1000 s-1_and

the dry content is 2.0 wt% or more the depletion layer seemed to have a rather low impact on the results [7].

3 Experimental part

The materials examined were 10 different grades of MFC labelled sample A-J. All samples were obtained from Stora Enso Oyj. The samples A-C were made of a different pulp type and with different equipment than the samples D-J. Each sample was pretreated or/and fibrillated in a slightly different way. All details regarding the manufacturing of the samples such as pulp type, equipment as well as pretreatments and manufacturing processes remained secret for this project.

3.1 Sample preparation

3.1.1 Mixing of suspension

When preparing the samples of MFC for rheological measurements different stirrers, propellers and containers were tested to examine which combination was best suited. The samples were dispersed to dry contents between 0.5 and 5.0 wt%. Different rotational-speeds were examined as well as different mixing times. A Janke & Kunkel IKA-WERK Ultra Turrax T-45 and a propeller mixer Janke & Kunkel IKA-WERK RW20 as well as a Kenwood ThermoResist™

KAH359GL blender were tested. When using the propeller stirrer several types of propeller

blades were examined in combination with different containers. The following containers were tested: glass beakers with different sizes (150, 100 and 200 ml) and a 250 ml plastic jug and a 200 ml plastic container. Different speeds of rotations were tested as well, ranging from 700 rpm to 1600 rpm and a mixing time between 3-35 minutes when using the Janke & Kunkel mixer. The Ultra Turrax mixer had only one setting which was 20’000 rpm but the mixing time was between 1-5 minutes. When using the food processor the maximum setting was used which corresponds to 202.5 rpm and the mixing time was between 5-10 minutes. The pH and conductivity were adjusted when the suspension was agitated using a radleys CarouselTM

Stirring hotplate. The magnetic stirrer was set to 500 rpm and the agitation was performed for

a total of 30 minutes. This was to make sure the pH and the conductivity could be adjusted and settled within the time. The pH was adjusted using NaOH with concentrations of 0.5, 5 and 50 % and HCl with concentrations of 0.5, 5 and 37 %. The conductivity was adjusted by using NaCl solutions with concentrations of 0.25, 1 and 5 M which was prepared from NaCl powder and deionized H2O.

3.1.2 Determination of dry content

(28)

19. moisture analyze to make sure it did not contain any moisture. The reason behind using the glass fiber filter was to avoid the dried MFC-film from rising to the top of the analyzer and thus giving a faulty result. This method was evaluated by performing it for 10 times on the same sample.

3.2 Spindle geometry

When using a dynamic rotational rheometer, different measuring geometries can be used. Samples of MFC with a dry content between 0.5 and 10.0 wt% were examined. Six different combinations of upper and lower geometries were used. The first three geometries were smooth plate-plate geometries with an upper diameter of 40 mm (PU40) and 20 mm (PU20) as well as a cone-plate geometry where the cone has an angle of 4° and a diameter of 40 mm (CP4/40). To all these upper geometries a smooth plate with a diameter of 61 mm (PLC61) was used as a lower geometry. A plate-plate geometry with serrated surfaces and an upper diameter of 40 mm and a lower diameter of 61 mm (PU40X/PLC61X) was tested as well. The two last geometries were a four bladed vane with a diameter 25 mm (4V25) and a splined bob with a diameter of 25 mm (C25G) combined with a splined cup with a diameter of 25 mm (PC25G) as lower geometry. The combination of upper and lower diameter is described by “upper geometry/lower geometry” and the abbreviation for all used geometries are presented in Table 1 to give an overview.

Table 1. Presentation and summarization of the geometries used and details regarding them as well as the names for each geometry.

Geometry

Upper/lower

geometry Diameter (mm) Surface Name

Plate upper 40 serrated PU40X

Plate upper 40 smooth PU40

Plate upper 20 smooth PU20

Cone (4° angle) upper 40 smooth CP40/4

Bob upper 25 splined C25G

4-bladed vane upper 25 n/a 4V25

Plate lower 61 serrated PLC61X

Plate lower 61 smooth PLC61

Cup lower 25 splined PC25G

3.3 Resting before measurement

(29)

20.

3.4 pH dependency study

The pH dependency was examined by preparing a large batch of Sample C as well as Sample F with a dry content of approximately 1.0 wt% to be able to compare to another study of pH dependency of MFC [44]. A large batch (800 ml) was prepared and fractioned into smaller containers instead of multiple small batches. This was to avoid differences in dry content and to make sure that the only difference between the samples was the pH. A spot-check was performed by preparing a small batch (80 ml) of each Sample C and Sample F. This was to compare the results from the large batch and the small batch with each other at one pH. The pH of the fractionated samples was then adjusted to 10 different values between 2.00 and 12.00 by using solutions of 50, 5.0 and 0.5 % NaOH as well as 37, 5.0 and 0.5 % HCl. The solutions with a concentration of 5.0 and 0.5 % was prepared by diluting 50 % NaOH and 37 % HCl with deionized H2O. The pH was measured using a Mettler Toledo™ Seven2Go™ pH meter. The pH

meter was always calibrated to an accuracy of at least 97.0 %

3.5 Dry content dependency study

Samples for the dry content dependency study were prepared from the Sample C and F. The pH of all the samples were adjusted to 8. The aim was to prepare samples with a dry content of 0.50, 0.80, 0.90, 1.00, 1.10, 1.20, 1.50, 2.00 and 3.00 wt%. The dry content dependency of Sample D was examined around 2.00 wt% by preparing samples of 1.50, 1.80, 1.90, 2.00, 2.10, 2.20 and 3.00 wt%.

3.6 Effect of conductivity study

The effect of conductivity was examined by measuring Sample D at 0.94 wt% with its original conductivity 36.6 µS/cm, and adjusting the conductivity to 77.6, 142.0 and 280.0 µS/cm by using NaCl solutions with concentrations of 0.25 and 1 M. The conductivity is measured using a Mettler Toledo™ FiveEasy™ Plus Benchtop Meter which has an accuracy of ± 0.5 %.

3.7 Rheological measurements

When performing all the above described studies a sequence of rheological measurements was performed. The measurements were performed using a rheometer Kinexus Pro+ which is shown in Figure 8 a) and the associated software rSpace. The loading of the sample to the measuring geometry is performed by using a metal spoon. The sample is stirred by hand slowly 5 times to the left and 5 times to the right by using the spoon. This is to make sure eventually sedimented fibrils is stirred up and to make sure the suspension is as homogenous as possible while affecting the rheology of the sample as little as possible. The spoon used holds 4-5 g of sample and 4 spoons (16-20 g sample) is loaded to the measuring geometry. A principle of the measuring geometry is shown in Figure 8 b).

(30)

21. measurement. The individual measurements from the sequence is described further under the associated heading.

a) b)

Figure 8. a) A dynamic rotational rheometer (Kinexus Pro+) and b) the principle of a serrated cup and bob measuring geometry used.

3.7.1 Amplitude sweep

An amplitude sweep is a measurement where the chosen measuring geometry is oscillating with a controlled increasing shear stress or shear strain. The reason that an amplitude sweep was performed was to find the linear viscoelastic region of the material (LVER). This region gives information about how large stress or strain can be applied to the material without destroying the sample. The maximum stress or strain calculated from this measurement will be used in the next measurement in the sequence. When performing a sequence of measurements it is important to stay within the LVER to avoid breaking down the sample in an early stage of the sequence [8].

To make sure the stress was well within the LVER the stress used in the following frequency sweep was 0.3 ∗ 𝜏_𝑚𝑎𝑥 (Pa). From this measurement the complex viscosity η* (Pa s) that the sample exhibits when it was exposed to different stress will be obtained as well. To examine approximately how large the LVER was the amplitude sweep was performed for several samples prepared from Sample C, D and F at a frequency of 1 Hz, with a shear stress between 0.01-500.00 Pa and collection of 10 data points per decade. The amplitude sweep had triggers that detected a decrease or increase of 5 % or more in G*, G’ or G’’ or a decrease or increase of 1̊ for δ together with a harmonic distortion within 5 % and thereafter skip to next action. This was to detect the LVER and to not measure beyond it even in the amplitude sweep. From these measurements the average LVER was calculated to an average of 0.78 Pa and the shear stress interval used was thereafter 0.01-0.70 Pa. The trigger that was supposed to detect the LVER was thereafter turned off and the amplitude sweep will thus result in the same number of points for each sample. This was to avoid the measurement from being stopped due to noise in the measurement.

3.7.2 Frequency sweep

(31)

22. sweep was used. From the frequency sweep the material can be characterized as a viscoelastic liquid, viscoelastic solid or a gel depending on the ratio between the viscous and the elastic modulus [8]. The frequency sweep will measure the complex viscosity as well since it is an oscillatory measurement. The frequency was ranging between 0.1-10 Hz with 7 data points collected per decade.

3.7.3 Controlled shear rate (CSR) viscosity measurement

The last measurement in the sequence was a viscosity measurement. Before the CSR measurement starts, the sample was conditioned using a pre shear, and was thereafter resting in the cup. The pre shear consists of a static shear stress of 1 s-1 that was applied for 1 minute, this was to condition the sample and to create a shear history that was as identical as possible for each sample. Thereafter the sample was resting for 2 minutes without any shear applied to it. The purpose of the resting period was to minimize an eventual shear thinning or thixotropic effect from the previous amplitude and frequency sweeps. The viscosity measurement is a rotational measurement which can be performed with a controlled shear rate (CSR) or shear stress (CSS). The CSR measurement is usually chosen when the material is showing self-leveling behavior while the CSS measurement is usually chosen when the material has a clear yield point and thus needs an applied force to flow. Since the measurements were performed with a dry content around 1.0 wt% and 2 wt% where the material exhibits self-leveling behavior CSR was chosen over CSS.

The reason that this measurement was placed last in the sequence was due to the expectation of a non-Newtonian shear thinning behavior of the material. A shear thinning behavior means that the viscosity of the sample will decrease when the sample is exposed to increasing shear forces. If this viscosity measurement would be placed first in the sequence the sample would need a long period of time to eventually recover from being exposed to these shear forces. It was thus more advantageous to perform this measurement last in the sequence. From the controlled shear rate viscosity measurement information regarding several rheological properties as shear thinning/thickening, some thixotropic/rheopectic behavior, eventual yield stress and reorganization of the fibrils will be obtained [8]. The CSR measurement was split into two regions where the shear rate of the first region was ranging between 0.008-0.100 s-1 and collects 20 data points per decade. The shear rate of the second region was ranging between 0.100-1000 s-1 and collects 60 data points per decade.

4 Results and discussion

4.1 Sample preparation

4.1.1 Mixing of suspension

(32)

23. blades which have a larger surface area compared to “propeller a” and thus leading to an inferior mixing of the sample.

The Janke & Kunkel mixer with the three rounded propeller blade geometry were examined using different sizes of glass-beakers, a plastic jug and a plastic container. A smaller gap between the propeller and the glass-beakers gave a better agitation of the suspension. The samples prepared had a volume of approximately 80 ml. Since the diameter of the propeller were approximately 50-55 mm, the best suited glass-beaker was a 150 ml with a diameter of 58 mm to make the gap as small as possible. But when using this at higher rpm resulted in the suspension splashing out of the container. The 200 ml plastic container was thus preferred since the gap between the wall and the propeller was tiny and the container was high enough to avoid spilling out the suspension. Compared to the 250 ml plastic jug it appeared to be easier to obtain the desired dry content. When using the 250 ml plastic jug the dry content often turned out to be lower than expected compared to when using the 200 ml container. The reason could possibly be that the suspension doesn’t need to be poured between several containers which means that a smaller amount of fibers will be loosed.

a) b)

Figure 9. Two different propellers used with the Janke & Kunkel RW 20 stirrer, a) a three bladed rounded propeller and b) a four bladed square propeller.

(33)

24. avoided by turning on the mixer and thereafter slowly turning the rpm to the right setting. A well dispersed and homogenous looking suspension is shown in Figure 10.

Figure 10. 2 wt% MFC suspension at pH 8 stirred with a Janke & Kunkel propeller mixer at 1400 rpm for 10 min.

The Ultra Turrax stirrer was too powerful to use when preparing the samples. It caused severe splashing of the suspension which can affect the dry content of the suspension since there is no knowledge about the amount of fibers splashing out of the container, especially when the suspension is not homogenous. When the Ultra Turrax was used for one minute the temperature of a 100 ml sample raised from 21.8 ⁰C to 27.8 ⁰C. When it was used for 5 minutes the temperature was raised to 43.0 ⁰C and the flocculation in the suspension became very obvious which is shown in Figure 11. Thus, the Ultra Turrax was not appropriate and excluded.

a) b) c)

(34)

25. When using the Kenwood mixer the samples did not appear to become homogenous at all, which can be seen in Figure 12.This could possibly depend on the low rotation speed (202.5 rpm). A longer mixing time (10 minutes compared to 5 minutes) did not appear to have any positive effect on the homogeneity of the sample which can be seen by the eye. Due to this, the Kenwood mixer was ruled out as well from being used in this study.

a) b)

Figure 12. Sample B 0.5 wt% stirred with the Kenwood mixer used for a) 5 and b) 10 min.

When adjusting the pH and conductivity of the samples, the suspensions were agitated using a magnet stirrer at 500 rpm. The agitation was performed for 30 minutes in total. This is to make sure the pH and the conductivity can be adjusted and settled to the desired values within this time.

Method development for rheological characterization of microfibrillated cellulose