1.1. Background

In its conventional context, fractionation means the separation of pulp into grades of different average fibre lengths. This approach is commonly used for recycled fibres, to split the paper pulp into a low-grade short fibre fraction and a high-grade long fibre fraction (Scott & Abubakr 1992). Fractionation is increasingly being used for virgin pulps nowadays. Other fibre properties are also of interest as well as fibre length, and the term — fractionation — is generally extended to cover the separation of a fibre suspension into fractions according to any average property, including cell wall thickness, specific surface, bonding potential etc.

Wood pulps, especially mechanically defibrated wood pulps, are highly heterogeneous. The fibres are widely distributed in their morphology, mechanical properties (Lammi & Heikkurinen 1997) and chemical composition (Chang et al. 1979, Koljonen et al. 1997), some of them having a high papermaking potential that can be exploited as such for the production of high quality papers, while others have no potential at all unless extensively processed. Appropriate classification of pulp into grades having different properties is considered an attractive possibility for utilizing its full potential. Savings in energy, chemicals and raw material costs can be envisaged, and further mechanical treatment (e.g. refining) can be focused selectively on those fibres that need it, while the chemical treatment of each fraction (e.g. bleaching) can be performed with the proper agents and dosages. Pulp quality targets can be achieved with lower energy or chemical consumption and without damaging the fibres through excessively severe processing. In addition, the use of different fibre properties in tailored paper products can reduce raw material costs, e.g. due to a lower need for reinforcement pulp. The economic potential of fractionation is obvious, but it is largely unclear how one can take full advantage of it in practice. Selective fractionation processes for different fibre properties are limited in number and their full fractionation potential has not yet been determined.

The practical fractionation choices available today are still centrifugal cleaning and pressure screening (Wakelin et al. 1999a, Duffy 1999). Centrifugal cleaners separate the fibres according to their specific surface (Wood & Karnis 1979) and cell wall thickness

(Kure et al. 1999), whereas it has been suggested that pressure screens fractionate pulp primarily by fibre length and secondarily by fibre flexibility (Karnis 1997). Screening of mechanical virgin pulps can be regarded as an example of fractionation in use today. Apart from shive separation, the aim is to separate out the underdeveloped long, coarse fibres for further treatment in a reject refiner to improve their papermaking properties. In fact, fractionation, i.e. removal of particles with poor papermaking properties (coarse fibres), and cleaning, i.e. removal of particles with no papermaking potential at all (shives and fibre bundles), are interconnected in this process.

Pressure screening is a straightforward process, but the phenomena affecting its results are complex. Basically, the separation of fibres is thought to take place by one of two mechanisms, probability screening and barrier screening. Probability separation implies that at least one dimension of a particle is smaller than the aperture size of the screen plate and thus the passing probability, P, of the particle is in the range 0 < P ≤ 1. In barrier screening, all the dimensions of a particle to be rejected must be greater than the aperture size of the screen plate, whereupon the passing probability of the particle will be zero (P = 0). The effects of an individual operating parameter or screen plate design parameter on the output are quite well known and predictable, at least qualitatively, but divergent views still exist on the effects of the parameters on the screening results and their mutual importance in this respect. Simultaneous changes in several parameters make prediction of the output very difficult, because the detailed fibre separation mechanisms that apply under different screening conditions are not fully understood.

Numerous articles concerned with pressure screening have been published, but their aim has usually been to a greater or lesser extent commercial and they have not always met the demands of scientific research. Their value as information sources is often low because essential details related to screen design and screening conditions have been reported incompletely. The fundamental phenomena occurring within the pressure screen unit were studied by Niinimäki (1998) in his doctoral thesis, and the present study is a continuation of his work. Other academic research has been published by Kumar (1991) and Olson (1996), who investigated the passing of fibres through a single aperture in the screen plate. A few attempts have also been made to model the pressure screening process, but no unique, universally applicable description has yet been developed. In their simplest forms, the models for screening efficiency are based on the mass reject rate and screening index (Kubat & Steenberg 1955, Nelson 1981). Nevalainen (1969) developed a group of complex equations to express the screening result, and Gooding & Kerekes (1989, 1992) demonstrated mathematically the relationship between reject thickening and the volumetric reject ratio. Later, based on the work of Gooding & Kerekes and of Kumar (1991), Olsson & Wherret (1998) used the fibre passage ratio to model fibre length fractionation by slotted screen apertures. Niinimäki et al. (1997) chose another approach for predicting the production rate and pulp properties and identified a model for pressure screening which was based on process experiments, theory and expert knowledge. An identical principle, but with much more extensive data, has been used by Hietanen et al. (1999). Three-dimensional curvilinear Navier-Stokes code has been also used to simulate the simplified passage of a single fibre through a screen slot (Dong et al. 2000), but the implementation of a simulation that will predict a practical screening situation computationally seems to be far away in the future.

A comprehensive understanding of the phenomena occurring within the pressure screen is still lacking, and the full potential of fibre fractionation with pressure screens has not yet become apparent. Pressure screening has been traditionally investigated as a process for removing debris (i.e. detrimental particles like shives, stickies etc.), but the aim of this thesis was to study it with a view to the fractionation of pulps, examining systematically and exhaustively the effects of screening parameters on fractionation under actual working conditions in order to provide an insight into its possibilities and limitations as a fractionation method. A further aim was to obtain new information relating to the modelling of pulp quality in pressure screening.