The purpose of this study was to investigate the applicability of different ways of calculating the average fibre length based on length measurements of individual particles of mechanical pulps. We have found that the commonly used average length-weighted fibre length, which is based on the assumption that coarseness is constant for all particles, as well as the arithmetic average, may lead to erroneous conclusions in real life as well as in simulations when used as a measure of the amount of long fibres. The average length-length-weighted fibre length or a weighting close to that, which to a larger extent suppresses the influence of shorter particles, is a relevant parameter of the “length” factor, i. e. amount of long fibres. Our findings are based on three studies: refining of different assortments of wood raw material in a mill; data from LC refining in mill of TMP, including Bauer McNett fractionation; mixing of pulps with different fibre length distributions. If the acceptable average fibre length for different products can be lowered, the possibility of reducing the specific energy input in refining will increase. Therefore, we need a reliable and appropriate way to assess the “length” factor.
Funding statement: We gratefully acknowledge the Swedish Energy Agency, the Knowledge Foundation, and Stora Enso for financial support.
We gratefully acknowledge partners in the e2mp-program for valuable discussions; employees at Stora Enso Skoghall and Stora Enso Kvarnsveden for stimulating trials and discussions; Maria Vornanen and Alf Gustafsson at Stora Enso for skilful experimental work.
Conflict of interest: There is no conflict of interest.
Kraft cook conditions
The chip samples (333 g BD) were cooked in a 30-litre digester with recirculation of the Kraft liquor at an effective alkali of 22 % as NaOH. Liquor to wood ratio was 3.8 l/kg. Temperature profile was 122–132 C (40 min), 132–155 C (5 min), 155–157 C (70 min) and 157–159 C (110 min). The yield was 49 %. The CTMP samples (100 g BD) were cooked in autoclaves with an effective alkali of 25 % as NaOH. The temperature was raised to 160 C for 90 min and remained at this level for 155 min. The yield was 47 %.
Average fibre length of digested CTMP
The CTMPs were also digested to allow a more strict comparison. Average fibre length (arithmetic, length-weighted, length-length-weighted) of digested CTMP is plotted versus the corresponding value for the original CTMP in Figures 12–14. This gave a similar picture of the higher amount of the longest fibres in the pulp made of the sawmill chips.
Modelling the influence of fibre length distribution
For deeper insights into why the ranking of pulps with respect to average fibre length can change, we have modelled the influence of the amount of fines.
Consider two model pulps, A and B, with different fibre length distribution, each consisting of two fibres, see Figure 15. Pulp A consists of two 3 mm long fibres while the two fibres in Pulp B are 2 and 4 mm long. Pulp B is also “refined” to peel off fines from the fibre wall in different amounts without shortening the fibres, resulting in Pulp B1, B2, B3, and B4. The fines are 0.2 mm long. Pulp B1 contains two fines particles, while Pulp B2 contains 20, Pulp B3 200, and B4 600. The value of 600 fines particles was derived by assuming that the pulp consists of 30 % fines where each particle has a width of 4 µm and a thickness of 1 µm. These values are within the range given by Marton (1964), but above 0.5 µm in width as reported by Gavelin et al. (1975). Smaller fines dimensions than we used here would have rendered more than 600 fines particles. Average fibre lengths of these model pulps are also included, calculated in four ways corresponding to a value of power of weight (p) ranging from 0 to 3.
Pulps B–B4 obviously contain the longest fibre compared to Pulp A regardless of the amount of fines. This is reflected in the higher averages when p = 3. However, using the length-weighted average (p = 1) the ranking of the pulps is strongly influenced by the amount of fines: Pulp B2, B3, and B4 exhibit a lower average compared to Pulp A. Also for the length-length-weighted average (p = 2), the higher fines content of B3 and B4 rank these pulps lower than A. The arithmetic average (p = 0) does not distinguish between Pulp A and B. This average is even more influenced by the amount of fines, Pulp B1–B4 may be interpreted as having the shortest fibres. In conclusion, the length-length-weighted average (p = 2) seems to be a more relevant way to assess the amounts of long fibres compared to length-weighted average (p = 1). More studies are needed to justify a higher value of p than 2, although the study with model pulps and the graph in Figure 16 indicates that p may be above 2. Studies by Carvalho et al indicates that coarseness increases faster than linear with fibre length (1997) meaning that p is slightly higher than 2.
The point at which the ranking of the model pulps shift has been studied more closely, i. e. the point where the average fibre length of B-pulps becomes higher than that of Pulp A. To do this, the average fibre length has been calculated as a function of power of weight ranging continuously from 0 to 3, Figure 16.
The figure illustrates that at higher fines contents it is necessary to increase the power of weight (p) in order to maintain the ranking of the pulps with respect to average fibre length. This explains why the ranking of the CTMPs became the opposite of what was found based on the right hand side of the whole fibre length distributions (c.f. Figure 4). If the length distribution had been narrower, the ranking of the pulps would have been less affected by how the weighting of the length-data was made. The higher amount of long fibres in the CTMP made of saw-mill chips was accompanied by a higher amount of fines, i. e. the length distribution was wider. This had a strong impact on the arithmetic and the length-weighted average fibre length. Other changes in raw material and process conditions may sometimes lead to similar changes in the relation between long fibres and fines. Therefore, it is important not to let an increased fines content be reflected in a too low reading of the average fibre length if we wish to reflect the amount of long fibres.
Modelling the influence of curl on the measurement of fiber length ()
A fibre model was constructed according to Figure 17. Each of the three sections of the fibre was 1 mm long which means that the fibre in total was 3 mm long (). The fibres were allowed to be bent in the two points A and B to different extent (cf. ϕA and ϕB in Figure 17) in order to get fibres with different level of curl. Furthermore, the section 1 and section 3 were also allowed to rotate randomly 0–360 degree (cf. θA and θB in Figure 17). For each of six levels of ϕ (ϕA = ϕB) about 2000 fibres were run and the projected fibre length in 2D (Lc) was recorded. Curl was calculated according to curl = (Lc-Lp)/Lp where is the straight distance between the end points of the fibre in 2D. A higher level of curl resulted as expected in a lower value of , Figure 18. A straight fibre has a curl of zero and will get a projected length of 3 mm. A “normal” level of curl for mechanical pulps are in the range of 10 to 20 %. At this level the value of the projected length was about 10 to 15 % lower than the real length in 3D with this simple model.
A conclusion from this study is that fibres with a lower level of projected fibre length in 2D () and a higher level of curl may have a higher real fibre length in 3D (). Hence the ranking with respect to fibre length may be shifted.
Bailey, I. W. (1920) “The Cambium and its Derivative tissues. II. Size variations of Cambial initials in Gymnosperms and Anginosperms”. Am. J. Bot. 7:11.10.1002/j.1537-2197.1920.tb05590.xSearch in Google Scholar
Bentley, R. G., Scudamore, P., Jack, J. S. W. (1994) “A comparison between fibre length measurement methods”. Pulp Pap. Can. 95(4):41–45.Search in Google Scholar
Brecht, W., Holl, M. (1939) “Schaffung eines Normalverfahrens zur Gütebewertung von Holzschliffen”. Der Papier-Fabrikant 24(10):74–86.Search in Google Scholar
Brecht, W., Klemm, K.-H. (1953) “The Mixture of Structures in a Mechanical Pulp as a Key to the Knowledge of its Technological Properties”. Pulp Pap. Can. 51(1):72–80.Search in Google Scholar
Carvalho, M. G., Ferreira, P. J., Martins, A. A., Figueiredo, M. M. (1997) “A comparative study of two automated techniques for measuring fiber length”. Tappi J. 80(2):137–142.Search in Google Scholar
Clark, J. d’A. (1942) “The Measurement and Influence of Fiber Length”. Paper Trade J. 115(26):36–42.Search in Google Scholar
Clark, J. d’A. (1962) “Weight Average Fiber Length- A Quick, Visual Method”. Tappi J. 45(1):38–45.Search in Google Scholar
Clark, J. d’A. (1978) “Pulp Technology and Treatment for Paper”. 0-87930-066-3, Miller Freeman Publications, San Francisco.Search in Google Scholar
Clark, J. d’A. (1985) “Pulp Technology and Treatment for Paper”. 0-87930-164-3, Miller Freeman Publications, San Francisco.Search in Google Scholar
Falk, B., Jackson, M., Danielsson, O. (1987) “The Use of Single and Double Disc Refiner Configurations for the Production of TMP for Filled Super-Calandered and Light-Weight Coated Grades”. In: Proceedings International Mechanical Pulping Conference. Vancouver. pp. 137–144.Search in Google Scholar
Forgacs, O. L. (1963) “The Characterization of Mechanical Pulps”. Pulp Pap. Mag. Can. 64:T89–T116.Search in Google Scholar
Gavelin, G., Kolmodin, K., Treiber, E. (1975) “Critical point drying of fines from mechanical pulps”. Sv Papperstid 78(17):603–608.Search in Google Scholar
Höglund, H. The Ljungberg textbook, Chapter 23 (Mechanical pulping), Fiber and Polymer Technology. KTH, Stockholm, Sweden, 2004.Search in Google Scholar
International Mechanical Pulping Conference (2005) Proceedings, Oslo.Search in Google Scholar
International Mechanical Pulping Conference (2009) Proceedings, Sundsvall.Search in Google Scholar
International Mechanical Pulping Conference (2011) Proceedings, Xian.Search in Google Scholar
International Mechanical Pulping Conference (2014) Proceedings, Helsinki.Search in Google Scholar
Kauppinen, M. (1998) “Prediction and Control of Paper Properties by Fiber Width and Cell Wall Thickness measurement with Fast Image Analysis”. In: PTS Symposium: Image Analysis for Quality and Enhanced Productivity, Munich.Search in Google Scholar
Klem, P. (1929) “On the Calculation of Weighted average Fiber Length in Paper”. Pulp Pap. Can. 28(3):173–177.Search in Google Scholar
Levlin, J.-E., Söderhjelm, L. “Paper Making Science and Technology. Pulp and Paper Testing”, ISBN 952-5216-17-9. Fapet Oy, Helsinki, 1999.Search in Google Scholar
Lidbrandt, O., Mohlin, U.-B. (1980) “Changes in fiber structure due to refining as revealed by SEM”. In: Int symposium on fundamental concepts of refining, IPC, Appleton, Wisconsin, USA, preprints p. 61–74.Search in Google Scholar
Marton, R. (1964) “Fiber structure and properties of different groundwoods”. Tappi J. 47(1):205A–208A.Search in Google Scholar
Mohlin, U.-B. (1980) “Properties of TMP fractions and their importance for the quality of printing papers”. Sv Papperstid 83(16):461–466.Search in Google Scholar
Paavilainen, L. (1990) “Importance of particle size, fibre length, and fines for the characterization of softwood kraft pulp”. Pap. Puu 72(5):516–526.Search in Google Scholar
Parham, R. A., Church, J. O. “On the Calculation of Weighted average Fiber Length in Paper”. Institute of Paper Chemistry Technical Paper Series, No. 44, pp. 546–554, 1977.Search in Google Scholar
Pulkkinen, I., Ala-Kaila, K., Aittamaa, J. (2006) “Characterization of wood fibers using fiber property distributions”. Chem. Eng. Process. 45:546–554.10.1016/j.cep.2005.12.003Search in Google Scholar
Reiyer Österling, S. (2015) Distributions of Fiber Characteristics as a Tool to Evaluate Mechanical Pulps, ISSN: 1652-893X, ISBN: 978-91-86694-66-3, PhD thesis, Mid Sweden University.Search in Google Scholar
Ring, G. J. F., Bacon, A. J. (1997) “Multiple Component Analysis of Fiber Length Distributions”. Tappi J. 80(1):224–231.Search in Google Scholar
Strand, B. C. (1987) “Factor Analysis as Applied to the Characterization of High Yield Pulps”. In: TAPPI Pulping Conference, Proceedings. pp. 61–66.Search in Google Scholar
Sundholm, J., Heikkurinen, A., Mannström, B. (1987) “The Use of Single and Double Disc Refiner Configurations for the Production of TMP for Filled Super-Calandered and Light-Weight Coated Grades”. In: Proceedings International Mechanical Pulping Conference. Vancouver. pp. 45–51.Search in Google Scholar
Turunen, M., Ny, C. L., Tienvieri, T., Niinimäki, J. (2005) Comparision of fiber morphology analysers. Appita J. 58(1):362–375.Search in Google Scholar
Tyrväinen, J. “Newsgrade TMP from three different Norway spruce (Picea Abies) wood assortments in mill-scale”. Pulp Pap. Can., 1997. 98(10):51–60.Search in Google Scholar
© 2018 Walter de Gruyter GmbH, Berlin/Boston