The influence of shives on newsprint strength Øyvind Weiby Gregersen1, Åge Hansen1 and Torbjørn Helle2 1: Norwegian Pulp and Paper Research Institute 2: Norwegian University of Science and Technology ABSTRACT Late wood shives act as sheet rupture inducing weak spots in hard nip calendered TMP-based newsprint paper. The late wood shives raise local basis weight over a critical level where all voids are completely compressed. This high degree of compression cuts and damages the fibre material close to the shive and make it act like a short cut in the paper. Some 90% of the total amount of shives in the newsprint were found to consist of early wood fibres, however these shives did not seem to act as sheet rupture inducing weak spots in the same manner as did the late wood shives. INTRODUCTION The strength of paper sheets is most frequently assessed by standard tensile strength test e.g. SCAN-P 67:93. This does however not provide us much information about end use properties like “paper web runnability”. It is generally assumed that web breaks in printing and converting occur when a transient stress peak coincides with a paper area having less than average strength i.e., a so-called weak spot. Different origins for weak spots in paper have been identified as triggers to paper web breaks. Shives have been reported to be the most important factor trigging web breaks followed by scabs and hairs [1]. It may however be questioned how important shives are for web breaks today, as improved screening technology has reduced the shives content drastically. In the literature much consideration has been given to shives [1-7]. A clear correlation between the number of shives and the frequency of web breaks has been found [5,6]. Machine calendering has been shown to generally reduce paper strength [3,7,8], however this effect is more pronounced if the paper contains much shives [3] or when the shives are thick [6]. The strength is reduced if the calendering nip clearance is reduced below a certain minimum [8]. Such calendering may change the fracture line of the paper from the low basis weight areas to the high basis weight areas. From this it appears licely to conclude that calendering damages the high basis weight areas if these are compressed below a certain caliper in the nip. EXPERIMENTAL PROCEDURE To investigate the effect of shives in paper four newsprint sheets of 45 g/m2, 1800 mm long and 1000 mm wide were strained to fracture in a specially designed tensile apparatus [9]. One sheet was strained at 25% RH, one at 50% RH and two at 75% RH. A low strain rate (0.1%/min) was chosen to allow visual identification and marking of micro cracks during straining prior to the catastrophic fracture. The micro cracks are identified as small out of plane buckles caused by the stress anisotropy around the crack [3]. After the catastrophic fracture the position of each identified micro crack was noted together with the position of the crack actually initiating the catastrophic fracture. A 20 mm by 20 mm area sample surrounding the crack was cut. The length of the micro crack and the length of the shive causing the crack (if any) were measured using the Kontron image analysis program applied on light microscope images magnified 50x from a Leica LEITZ DMRXE microscope. 7KHVPDOOPLFURFUDFNDUHDVDPSOHVZHUHWKHQLPDJHGE\ UDGLRJUDSK\DW67),7KH UDGLRJUDPVZHUH digitized and analysed in a Kontron image analysis program to explore the local basis weight distribution in the areas containing the micro cracks. Each micro crack area was mounted in epoxy resin, ground perpendicular to the shives direction, polished, cleansed, carbon coated and imaged in the SEM, BEI-mode, to measure the cross sectional characteristics of the micro crack inducing shives [10]. For reference a raw specimen of the same newsprint paper were re-pulped and screened in a Somerville screen. The retained shives as well as the +48 mesh Bauer McNett long fibre fraction were parallellised, embedded in epoxy and the cross sections were analysed the same way as the micro crack inducing shives. RESULTS AND DISCUSSION The tensile index and fracture strain of the studied sheets are shown in Table 1. As can be seen the tensile index does not vary much, but the fracture strain is increasing with rising humidity in accordance with the general reduction of paper tensile stiffness with increasing humidity.

Table 1: Fracture load and strain for the different climates. 25% RH

50% RH

75% RH

75% RH

Tensile index (kNm/kg)

35.5

35.5

30.5

32.9

Fracture strain (%)

0.54

0.54

0.64

0.71

The shive length distribution for the crack inducing shives of the sheets tested in different climates is shown in Figure 1. 0.5

Fraction

0.4

0.3

0.2

0.1

0 0

1

2

3

4

5

Shive length (mm)

Figure 1: The length distribution of shives inducing cracks In 87% of the 69 cracks identified in the paper during streaching, a shive was observed. By subtracting the shive length from the crack length for each crack of the different climates three new crack growth length distributions were made. The cracks had grown significantly (95% level) more for paper strained at 75% RH compared to paper strained at 25% RH. The paper strained at 50% RH falls in between these values. The correlation between crack growth and humidity is given in Figure 2. The shive cross sections were imaged in the SEM, BEI-mode at 150x magnification. Some shives were also imaged at 500x resolution. Figure 4 shows the cross section of a typical crack inducing shive in the newsprint. 0.4

Crack growth (mm)

0.2

0

-0.2

-0.4

-0.6 0

25

50

75

100

RH (%) Figure 2: The average shive length subtracted from the crack length (i.e., crack growth) for the ruptured sheets in three different climates. The error bars indicate the 95% confidence limits.

Figure 3: SEM-BEI cross sectional image of a crack inducing late wood shive in newsprint. The cross sectional characteristics of crack inducing shives were acquired from images like Figure 3. Visual inspection of the shive cross sections indicated that the crack inducing shives usually were composed of late wood fibres. To investigate this more closely the cross sections of crack inducing shives, the shives screened from repulped paper as well as the long fibre fractions (+48 mesh) were imaged in the SEM. To distinguish between early and late wood the Zparameter was applied [11]. Z is defined as the cross-sectional fibre wall area divided by the area of a circle with the same outer perimeter (Eq. 1). Here, Aw denotes the cross-sectional fibre wall area and P denotes the outer perimeter. It has been showed that the distribution of Z for Norway Spruce (Picea Abies) could be very well fitted by two partly overlapping normal

Z=

4 Aw * π P

2

* 100%

Eq. 1

distributions [11]. The normal distribution corresponding to the higher Z-values represents the late wood fibres and the distribution corresponding to the lower numbers represents the early wood fibres. Figure 4 depicts the distribution of the Z-parameter for the long fibre fraction (+ 28 mesh) of the studied newsprint. Figure 5 shows the Z-parameter distribution for the crack inducing shives together with the Z-paramerter distribution of all the shives in the repulped paper. 0,6

0,5

Probability density

0,4

0,3

0,2

0,1

0,0 0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

0,9

1

Z-parameter

Figure 4: The distribution of the Z-parameter for the long fibre fraction of the newsprint has a peak in the early wood area, but even substantial amounts of late wood.

0,6 All shives

Crack shives

0,5

Probability density

0,4

0,3

0,2

0,1

0 0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

0,9

1

Z-parameter

Figure 5: The Z-parameter distribution of the fibres in the shives from the repulped paper together with the fibres in the crack inducing shives reveals that the crack inducing shives are mainly composed of late wood fibres. From Figure 5 it is evident that the crack inducing shives usually are composed of late wood fibres and that the crack inducing thick walled shives represent only some 18% of the total number fraction of shives. Fortunately the shives in TMP based newsprint have been found to mainly consist of early wood fibres [11]. The early wood shives do not seem to affect the newsprint strength much, it is the more rare late wood shives that cause cracks. This also gives the clue to the mechanism by which shives reduce the effective web strength. It seems as the conventional steel-steel machine calender applied to the studied newsprint causes short cuts in the paper where the local basis weight is too high. This calendering process tends to give the paper uniform thickness regardless local basis weight before calendering. As shown in Figure 6 this may cause a total collapse of all void volumes both within and between the fibres in the area covered by a shive.

Figure 6: A fully compressed shive that induced a web crack The local calender pressure will be very high in this area, causing high local compression and shear strain. This will partly cut, partly weaken the good fibres crossing the shive (Figure 7). The calendered late wood shive will thus HVVHQWLDOO\ DFW DV D FXWWLQJ HGJH LQ WKHSDSHU ZHE $VVHVVPHQWV RI ORFDO EDVLV ZHLJKW XVLQJ

UDGLRJUDPV UHYHDOHG WKH

average basis weight of the shive areas to be 64.5 g/m2. Assuming complete compression and a density of the fibrous material of 1.5 g/cm3 this corresponds to a FDOLSHU RI  P )LJXUH  VKRZV WKDW WKLV LV YHU\ FORVH WR WKH JHQHUDO caliper of the uncompressed newsprint sheet. The high degree of compression gives further evidence that the

calendering pressure in the shive areas will be very high. An early wood shive will usually not cause a sufficiently high local basis weight to establish a dangerously high pressure in the calender. When seen from above the crack inducing shives look totally flattened and smooth (Figure 8). In an optical light microscope the shives appear transparent and yellow indicating optical contact. These observations support the hypothesis of an extremely high local calender pressure causing damage to the fibres in the shive area For catastrophic fractures initiated by shives, one rarely observes fibres bridging across the shives after rupture.

Figure 7: A shive with crossing fibres that have been damaged in the calendering process.

Figure 8: SEM image of the area above the crack inducing shive that has been flattened in the calender.

CONCLUSIONS There are hundreds of shives per square metre of newsprint, however only 10-20 of them are causing micro cracks to develop before the catastrophic failure. Thus, it is easy to conclude that some of the shives reduce strength and others do not. From the results presented here it seems that the shives actually reducing paper strength are composed of thickwalled late wood fibres causing sufficiently high local basis weight to yield a total compression of all voids during calendering. This total compression will cause a very high local deformation. This local deformation of the fibre

material will reduce the strength of the fibre material around the shive to an extent that makes the shive act like a short cut in the paper web. This finding is in agreement with the earlier observations that: The strength reduction in calendering started when the calender clearing was below a certain minimum [8]. The strength reduction is more pronounced the thicker the shive [7]. From the new results presented here and the reviewed literature data it seems reasonable to conclude that a shive is reducing the web strength when the local basis weight of the shive area is a certain percentage above the average basis weight of the sheet. In the case studied here local basis weights in the crack inducing shive areas were on average 143% of the average basis weight of the paper. After calendering, this shive will act as a short cut in the paper. The length of the cut and the angle to the machine direction will of course depend on the length and orientation of the shive. References 1. Adams, R.J. and Westlund, K.B., "Off-line testing for newsprint runnability", 1982 Printing and Graphic Art Conference, 1, pp 13-18, (1982). 2. Hopkins, R.M., MacPerson, R. and Morin, L.J., "Analysis of the effects of centrifugal pulp cleaners and pressure screens on newsprint runnability", P. & P. Can., 63 no 12, pp T563-T570, (1962). 3. Sears, G.R., Tyler, R.F. and Denzer, C.W., "Shivers in newsprint: The role of shives in paper web breaks", P. & P. Can., 66 no 7, pp T351-T360, (1965). 4. MacMillan, F.A., Farrel, W.R. and Booth, K.G., "Shives in newsprint: Their detection, measurement and effects on paper quality", P. & P. Can., 66 no 7, pp T361-T369, (1965). 5. Stephens, J.R. and Pearson, A.J., "The cleaning of eucalypt groundwood by the use of the 623 Bauer hydrocyclone", Appita, 21 no 3, pp 79-93, (1967). 6. Laurila, P., Smook, G., Cutshall, K. and Mardon, J., "Shives -- how they affect paper machine runnability", P. & P. Can., 79 no 9, pp T285-T289, (1978). 7. Höglund, H., Johnsson, E. and Tistad, G., "Shives in mechanical pulp", Svensk Papperstidning, 79 no 11, pp 348353, (1978). 8. Moffat, J.M., Beath, L.R. and Mihelich, W.G., “Major factors governing newsprint strength”, 5th FRS: The fundamental properties of paper related to its uses, Cambridge, England, Vol. 1, pp 101-127, (1973). 9. Hansen, Å., "Ny papirbanestrekkmaskin utviklet ved PFI", Papirforskning, 26 no 1, pp 10-12, (1995). (in Norwegian) 10. Fjerdingen, H., Forseth, T.F., Gregersen, Ø.W., Helle, T., Johnsen, P.O., Kure, K.A. and Reme, P., “Some mechanical pulp fibre characteristics, their process relationship and papermaking significanse”, 11th FRS: The fundamentals of papermaking materials, Cambridge, England, Vol. 1, pp 547-605, (1997). 11. Reme, P.A., Tufa, L.D., Helle, T. and Johnsen, P.O., “The fibre characteristics of shives initiating web rupture”, Paper Physics Seminar, Vancover, Canada, 9-14 Aug. 1998.