ABSTRACT

FIBRE COLLAPSE AND SHEET STRUCTURE

J . Gorres*, R . Amiri, M. Grondin andJ.R . Wood

Pulp and Paper Research Institute of Canada570 St . John's Blvd .Pointe Claire, Quebec

H9R 3J9

*University of Rhode IslandDepartment of Natural Resources

Kingston, Rhode Island02881U.S .A .

A stylus profilometer has been used toevaluate the effect of wet pressing on thecollapse of individual fibres from mechanical,chemimechanical and chemical pulps . Thechemimechanical and chemical pulp fibres beginto collapse at low pressures and approachcomplete collapse at high pressures, while themechanical pulp fibres do not exceed 80%collapse to 5000 kPa . The degree of collapseof southern pine TMP at a given pressure isabout the same as that of northeasternspruce/balsam TMP . Since the thick-walledSouthern pine fibres are less flexible, it isconcluded that transverse collapsibility andflexibility are two independent fibreproperties .

On the basis of modelling results and thedifference between fibre thickness-measurementfrom networks pressed in contact with smoothand with rough surfaces, it is suggested thatwet pressure transferred locally at fibrecontacts within a sheet leads to localcollapse forces higher - than expected fromnominal wet pressure vlues. Fibre contactsare initially present in the unpressedsheet.On pressing, free fibre segments will bedeflected into contact with other segmentsabove or below them, producing additionalfibre crossings as pressing progresses . Theeffective pressing pressure will be highest atthe initial fibre contacts, decreasing to zeroat contacts just formed at the end ofpressing . Because of the demonstrated effectof wet pressure on fibre collapse, it is thethickness at the initial fibre contacts or

Preferred citation: J. Görres, R. Amiri, M. Grondin and J.R. Wood. Fibre collapse and sheet structure. In Products of Papermaking, Trans. of the Xth Fund. Res. Symp. Oxford, 1993, (C.F. Baker, ed.), pp 285–310, FRC, Manchester, 2018. DOI: 10.15376/frc.1993.1.285.

those formed early in the pressing processthat is important in determining sheetdensity .

Use of the more appropriate fibre thicknesses,substantially improves the prediction of sheetdensity by the Interactive Multiplanar Modelof sheet structure for a range of pulp types .

INTRODUCTION

Fibre collapse and the thickness of mechanicalpulp fibres have been recognised as importantfactors in the development of pulp andultimately sheet quality . Kibblewhite [1,2]and Kibblewhite et al . [3] investigated theeffect of pulp treatment and fibre properties,including collapse, on sheet quality . Heitnerand Hattula [4] and Rioux and Silvy [5]suggested that fibre collapse increasedinterfibre bonding thus affecting sheetdensity and strength . Rahman et al [6] usingthe method developed by Page [7], presenteddata showing that both breaking length andsheet density increase with increasing fibrecollapsibility . Similar relationships betweensheet properties and fibre collapse orthickness were reported for chemical pulps byDinwoodie [8] .

The process by which fibre thicknessinfluences sheet structure is considered inthe Interactive Multiplanar Model (IMPM) [9] .The model predicts that interlayer fibrebonding increases as fibre thicknessdecreases . A sheet density model based on theIMPM [10,11] suggests an even strongerdependence of sheet density on fibrethickness . However, the model relies heavilyon accurate fibre thickness measurements, anddoes not account explicitly for relationshipsbetween fibre collapse and wet pressure .

While methods for measuring other fibreproperties such as coarseness [12], and length[13,14], are well known, measurement of fibrethickness is less common . Evaluation ofmicroscopic cross sections with the VAX ImageProcessing System (VIPS) algorithm [15], andconfocal microscopy [16], are attractivemethods but will probably remain exclusive toa few research centres . Measurement of fibrethickness by the vertical displacement of ascanning stylus, which is simple, rapid, andaccessible at reasonable cost was the methodused in this study .There were three objectives of this study . Thefirst was to quantify and compare the degreeof fibre collapse of different pulps inresponse to wet pressure . The second was toinvestigate the effect of the smoothness of

288the surface against which the sheet waspressed on the fibre thickness and to decideon the thickness measurement appropriate fordescribing sheet structure . The third was todetermine if this thickness measurement wouldimprove the correspondence between measuredsheet densities and those calculated with theIMPM.

During handsheet making, fibre collapse iscaused by couching and wet pressing [19] . Thefinal magnitude of collapse is stronglyinfluenced by chemical treatment [4,6] or byrefining [2] . In this report, the effect ofwet pressure on fibre collapse and therelationship between wet pressure, collapseand sheet structure is examined . It isrecognized that the conditions of drying couldalso affect fibre collapse but this effect isnot addressed here .

EXPERIMENTAL

Pulp Preparation and Sheet Making

In a previous study [11], the effect of fibreorientation on sheet density was studied usingthe R48 fraction of different pulps . Theunderestimation of sheet densities calculatedusing the IMPM was postulated to be due, inpart to the presence in the R48 fractions ofparticles small enough to act as "structuralfines" and fit in the interstices betweenfibres, and in part to inadequate measurementof fibre thickness . In this report the P14/R2 8fraction was used both for fibre thicknessmeasurements and for standard sheetmaking, toreduce the influence of such structural fines .In addition, since some difference in the meanthicknesses of the two fractions might beanticipated, fibre thicknesses of the R48fractions were also determined with theimproved method to calculate sheet densitieswith the IMPM for direct comparison with theprevious results .

Five pulps were used in this study : twothermomechanical pulps (TMP), from northeastern softwood and from southern pine, oneeastern softwood chemimechanical (CMP), onewestern ._.softwood chemithermomechanical pulp(CTMP) and an eastern kraft pulp . The pulps

were screened in a Bauer-McNett classifier toisolate the fibre fractions . For each pulp,four series of sheets were formed with theP14/R28 fraction, differing in the degree ofwet pressure that they received . Pressuresused were 350, 860, 2240 and 4830 kPa . Inaddition, sets of thin fibre networks, lessthan 0 .1 g/m2 , were prepared from this fractionon glass slides and pressed at the samepressures for measurement of individual fibrethicknesses . Fibre thicknesses of the R48fractions pressed at only one pressure, 350kPa, were also measured .

Fibre Thickness Measurement

As described previously (11], fibrethicknesses were determined on the thin fibrenetworks after transfer to the glass slides,pressing, and air drying . The fibres werescanned by the stylus of a Taylor-Hobson,Model 3, surface profilometer . The mean fibrethickness for each pulp was obtained byscanning 60 fibres, each at five locationsalong its length .

In the previous work the fibres had beentransferred to the slides under wet pressurewith blotters backing the network duringpressing . Fibre thickness was observed to varygreatly along the length of each fibresuggesting that prominent fibres at theblotter surface had led to an uneven transferof pressure and thus to non-uniform collapse .

In this study, to eliminate direct contactbetween blotter and network fibres, networkswere backed during transfer by a milliporefilter which was in turn backed by twohardened smooth surface paper filters (WhatmanNo . 50) . Figure 1 shows a diagrammaticrepresentation of the new transfer sandwich .Figure 2 shows Scanning Electron Microscopy(SEM) photomicrographs (100X) of the surfacesused in the transfer of fibre networks . Theblotter surface (Figure 2A) comprises manyprominent -fibres, whereas the surface of thehardened smooth surface filter andparticularly of the millipore filter (Figures2B and 2C) are much smoother .

Figure 1:Improved network transfer arrangementwhich gave a more even pressure distributionover the network area .

A B

CFigure 2 :Surfaces used in the transfer offibre networks (A) blotter surface with manyprominent fibres . (B) hardened smooth surfacefilter with a fibrous but smooth appearance .(c) millipore filter with a very smoothappearance . Magnification 100X .

RESULTS

Double wall thickness was computed using theequation of Mistry [17] and Jordan [18] :

t2w= CWPcel l

where Pcetl is the density of cellulose, C isthe fibre coarseness and w is the fibre width .For chemical fibres, Jordan [18] found goodcorrespondence between literature values ofcell wall thickness and values obtained withthe above formula . For this study, it isassumed that the density of the cell wall ofmechanical and chemical pulp fibres is thesame . It is recognized that equation (1) is anapproximation which will overestimate wallthickness slightly because a fibre which isnot completely collapsed will show a smallerfibre width than if it were completelycollapsed .

In this report we define the degree ofcollapse as

S=[1 _ t -t2w ] *100%

(2)to -t2w

where t is the uncollapsed fibre thickness,t w isthe double cell wall thickness and t istge fibre thickness after a treatment thatleads to fibre collapse . The initialuncollapsed fibre is assumed to becylindrical ; . it follows then that theuncollapsed fibre thickness, to , is equal toinitial' fibre width, wo . Equation (2) showsthat d = 0 when t = to and S = 10 0 when t =t2w , which implies complete fibre collapse . Theuncollapsed cylindrical fibre diameter wascalculated using the following equations :

Fibre Coarseness = Fibre Weight

(3)Fibre Length

.

or

Fibre Coarseness .=FibreVolume - pcel <

(4)

Fibre Length

Rearranging equation (4) yields

C+

(5)Net l

to -

~

. t

twc w

where tw is the cell wall thickness calculatedfrom equation (1) .

For the R48 fraction of the 5 pulp samples,mean fibre thicknesses and standard deviationsobtained from profilometry measurements onnetworks pressed in contact with blotters andwith millipore filters are listed in Table I .All values measured from networks pressed incontact with a blotter are significantlyhigher than those measured from the networkspressed more uniformly, with the milliporefilters . An explanation why the blotter methodgives higher mean thicknesses is given inAppendix 1 . For the three pulps for which acomparison of standard deviations with the F-statistic is valid, more even distributions ofthickness are recorded from networks pressedwith the millipore filter . For relevantstatistical information see Appendix 2 .

Table I : Mean thickness values (R48 and standarddeviation

fractions)of the thickness using the blotter and the

method to transfer fibremillipore

networks .

Fibre TypeMean Thickness

[Am)Standard Deviation

[Am ]Blotter Millipore Blotter Millipore

TMP-NE 18 .3 12 .9 * 5 .7 6 .2

TMP-S 20 .3 16 .2 * 7 .5 6 .8 *

CTMP 20 .4 13 .4 * 8 .5 5 .9 *

CMP 15 .0 9 .6 * 7 .4 4 .6 *

kraft 6 .9 5 .0 * 3 .3 1 .9

* Significantly different at the 95% level .

Fibre widths, and coarsenesses [12] determinedon the P14/R28 samples are given in Table IItogether with the cell wall thickness anduncollapsed diameters calculated usingequations 1 and 5 .

The effects of wet pressure on fibre thicknessand the degree of collapse are summarized inTable III . The table lists, for each pulp,thickness and percent collapse at the wetpressures used . At high pressures, CMP andkraft reach high levels of fibre collapse,whereas CTMP and TMP fibres only collapse to80% . Figure 3 shows the fibre thickness of thepulps as a function of wet pressure . Figure 4presents the same information for the degreeof collapse as defined by equation (2) . Thedegree of collapse is well represented by asimple logarithmic equation of the form :

6 =A+B (log P)

(6)

where A and B are constants for each pulp andP is the applied wet pressing pressure . Thedensities of handsheets made from the P14/R28

Figure 3 :Increasing wet pressure decreasesfibre thickness for all pulps . FractionP14/R28 .

Figure 4 :The degree of collapse of each fibretype in response to pressure is wellrepresented by a logarithmic function .

pulp fractions increase markedly with pressureat lower wet pressures and approach a plateauvalue at pressures higher than 2,000 kPa(Figure 5) . Bulk, the reciprocal of density isshown as a function of fibre thickness inFigure d . At high pressures the relationshipappears to be the same for all pulps .

Table II : Measured and calculated fibre properties(P14/R28 fraction) .

Pulps

i

Width

(microns)

Coarseness

(mg/m)

Cell wallthickness(microns)

Calculateduncollapseddiameter(microns) *

TMP-NE 31 .6 0 .25 2 .50 23 .0

TMP-S 31 .4 0 .48 4 .01 28 .1

CTMP 37 .0 0 .32 2.75 26 .4

CMP 33 .0 0 .27 2 .65 23 .9

KRAFT 38 .0 0.20 1 .65 26 .4

IF Calculated using equations (1) and (5) .

Table III : Fibre thickness Measurements and percentcollapse at different pressures . P14/R28 Fraction .

Pressure Fibre Thickness [Am] (upper) and PercentCollapse (lower) Figure

[kPa]TMP-NE CTMP CMP kraft

TMP-S

350 13 .3 16 .5 13 .7 10 .2 6 .554 58 61 74 86

860 11 .2. 17 .1 12 .2 9 .5 5 .566 55 68 77 91

2240 10 .9 13 .1 11 .5 7 .4 4 .267 75 71 89 96

4830 8.6 12 .3 9 .6 6 .6 3 .980 79 80 93 97

DISCUSSION

Comparison of the Degree of Collapse ofDifferent Pulps

The response of different types of fibres towet pressing is shown in Figure 4 . The degreeof collapse of all types of fibres increasesrapidly at low pressures, then much moreslowly at higher pressure . The kraft and CMPfibres, which have been extensively modifiedwith chemicals, collapse rapidly at lowpressures and approach Figure 4 . The degree ofcollapse of each fibre type in response topressure is well represented by a logarithmiccomplete collapse at 5000 kPa wet pressingpressure . The essentially mechanical pulpfibres, TMP and CTMP, which are untreated orlightly treated with chemicals, collapse lessrapidly at low pressures and only reach about80% collapse at 5000 kPa . CTMP is slightlymore collapsible at low pressures but similarto the other two at high pressures . Datareported by Miller [20] also suggest adependence of collapse on the pretreatment, Anparticular on beating and beating consistencyof Pulps .

Figure 5:Sheet density increases with 'wetpressure for all pulps . Fraction P14/R28 .

Figure 6:At high pressures bulk is a uniquelinear function of fibre thickness for allpulps . The solid line is a least squares fitthrough the two higher pressure for all, pulpsFraction P14/R28 .

The similar collapse behaviour of these threemechanical pulp fibres and in particular theabsence of any major difference between thenortheastern softwood TMP and the TMP fromsouthern - pine, -which have very differentcoarsenesses and flexibilities [11], indicatesthat fibre flexibility and transverse fibre

298collapsibility are distinct fibre propertieswhich are not necessarily correlated . Sincefibre collapsibility has a major effect onsheet thickness, and presumably smoothness,separation of these two fibre properties isessential to an understanding of how fibreproperties control sheet structure .

one may speculate that collapsibility isdetermined by the S1 layer which is of similarwidth in the different species and has afibril orientation at a high angle to thefibre axis, while flexibility is determined bythe S2 layer which is much wider in Southernpine and has a fibril angle close to the fibreaxis .

If this speculative explanation for thesimilar collapsibility of mechanical pulpfibres is correct it implies that mechanicalpulping procedures which strip off S1 materialfrom the fibre wall will make the fibre morecollapsible, though not necessarily much moreflexible .

Similarly the phenomenon of fibre roughening,which is essentially a reversal of fibrecollapse, maybe due to stresses "frozen" intoa relatively intact S1 layer which arereleased on application of water . Pulps inwhich the S1 layer has been extensivelydisrupted such as groundwood, and rejectsrefined at high specific energy, would beexpected to show less fibre roughening thanTMP prepared at lower specific energy .

The results shown in Figures 3 and 5 are inaccord with the inverse relationship betweensheet density, and fibre thickness found byDinwoodie -[8] and predicted by the IMPMapparent density model [10] . Figure 6 however,indicates that there are apparently twodifferent - mechanisms controlling thisrelationship . At lower pressures bulkdecreases rapidly with decreasing fibrethickness along a distinct line for each pulp .At higher wet pressing pressures, the decreaseof bulk with thickness is more gradual and allpulps fall on the same straight line .Presumably, in the former region, fibrecollapse and the deflection of free fibresegments described by the IMPM [11] are the

predominant effects, while in the latter, allspace within the sheet which can be filled bysuch deflections has been filled . Completionof collapse of the mechanical pulp fibres andperhaps other effects, then predominate . Asecond effect is necessary to explain thecontinuing decrease in bulk at high pressurewhere CMP and kraft fibres are totallycollapsed . A recent study of thecompressibility of high consistency pulp matshas also shown that kraft and CMP approach thesame density at high pressures [21] .

The observation that CMP gives higherdensities than TMP is in accord with thefindings of Corson and Kibblewhite [22] whoexplained' their results by betterconsolidation properties of the CMP . In lightof our findings on the collapsecharacteristics of CMP and TMP, most of thedensity difference between sheets from the twopulps may be explained by the greater collapseof the CMP fibres .

Effect of Backing Smoothness on FibreThickness

This work has shown that a smooth backingduring pressing will produce a lower mean, anda more uniform, fibre thickness . In additionsheet densities predicted with the IMPM usingthe measured fibre thicknesses, both in theprevious work [11] and, as shown in the nextsection, in this work, are often too low . Toexplain these two observations we mustconsider fibre collapse within the sheet atthe microscopic level.

When widely spaced fibres deposited on a slideare pressed with a smooth surface, pressure isapplied uniformly to the whole length of thefibre and to the slide surface between fibres .The pressure received by the fibres will becomparable to the applied pressure . on theother hand, when such a thin fibre network ispressed directly with the rough blottersurface shown in Figure 2A, pressure cannot betransferred evenly along the fibre length butwill be transferred at the crossings betweenthe network fibres and the blotter fibres . Thepressure at the crossings will besubstantially higher than the applied

300pressure . This leads to non-uniform collapsealong the length of the fibres and produceshigher mean thicknesses (see Appendix 1) .

Fibre collapse in a paper sheet, which may beconsidered to be a stack of single fibrenetworks, also occurs at fibre crossings .Fibre contacts are initially present in theunpressed sheet . On pressing, free fibresegments will be deflected into contact withother segments in layers above or below themproducing additional fibre crossings aspressing progresses . The effective pressingpressure will be highest at the initial fibrecontacts, decreasing to zero at contacts justformed at the end of pressing . It is thereforethe thickness at the initial fibre contacts orthose formed early in the pressing processthat is important in determining sheet.density .

Validity of Sheet Densities Predicted with theIMPM

Figure 7 shows diagrammatically the collapsepattern within a section of a sheet . Althoughfibres are only collapsed at fibre crossings,the distance between free fibre surfaces indifferent layers of the sheet is reduced tothat of the fibre thickness at fibrecrossings . The assumption made by the IMPMthat the thickness of a layer is equal to theaverage fibre thickness is not correct forfibres at pressures that give some localisedcollapse at fibre contacts in the sheet . Theimportant z-directional sheet constant forinput into the IMPM should be approximatelythe thickness at fibre-to-fibre contact .

As a result of the uniform distribution ofpressure application when measuring fibrethickness after pressing with a milliporefilter, thicknesses are higher than thethicknesses at locally collapsed crossings inthe sheet . However, under certain conditionsthe thickness values obtained by the milliporetransfer method- Inay be a good approximation ofthe thickness values of locally collapsedfibre segments . Figure 8 compares densitiescalculated with the IMPM for sheets from theP14/R28, used in this work, with measureddensities . At low and high pressures the

Figure 7 :Schematic showing the effect oflocalization of wet pressure at fibre contactson layer separation . The distance betweenlayers, tj , is less than the average fibrethickness, tf . Deflection of free fibresegments to produce additional contacts hasbeen omitted .

calculated and predicted values are inagreement . When wet pressure is so high thatthe fibres have approached complete collapseeven under homogenous wet pressure, fibrethickness at fibre contacts is similar to theaverage fibre thickness measured with themillipore transfer method . At 4830 kPapressure, CMP and kraft, which have beensignificantly softened by chemical treatment,are highly collapsed and for these two pulpspredictions at this pressure are close to themeasured values . At low pressures there islittle difference between the pressure atfibre to fibre contacts because the higherpressure at contacts must arise by transfer ofstress along daflecting free fibre segmentsand little deflection has taken place . Thepredicted densities for all but the kraft pulpare close to the measured densities at thelowest pressure, 350 kPa . At intermediatepressures IMPM predictions underestimate sheetdensity .

Figure 8 :Comparison of measured and predicteddensities forsheets made from the P14/R28fraction and pressed to pressures ranging fromabout 350 to 5000 kPa .

The thicknesses of the fibres from the R48fractions measured on networks transferred incontact with millipore filters were used torecompute apparent density values for theoriented sheets used previously [11] . As isevident in Figure 9, the IMPM densitypredictions with the mean fibre thicknessesobtained using a smooth backing were in betteragreement with the measured sheet densitiesthan the predictions reported previously [11],which were computed with the thickness valuesmeasured from networks transferred by theblotter method . However, the improvedpredictions still remain lower than measuredsheet densities .

Figure 9 : Correlation of measured and predictedsheet densities for sheets of reference[11]using fibre thicknesses obtained with a smoothand a blotter backing . R48 fraction .

As discussed earlier, the R48 fraction, incontrast to the P14/R28 fraction, contains"structural fines", material somewhat shorterthan the mean free fibre length which will fitinto the polygonal interstices between fibresin a layer, and increase the sheet densityabove that predicted by the IMPM . Thediscrepancy between measured and calculateddensities for sheets from the R48 fractions isattributed to the presence of this material .'

It is clear from the above discussion that thecollapse behaviour of pulps and the effectivefibre thickness at contact points within thesheet remain important issues both for usingthe IMPM, and more .generally, forunderstanding paper manufacturing, structureand properties .

304It is also apparent that it would be useful toextend the IMPM to include the effect of"structural fines" and- to calculate thedistribution of fibre thickness at crossingsby explicitly describing the balance betweenfibre flexibility and fibre compactibilitywhich controls the transfer of the pressingpressure from the pressing surface to thefibre crossings .

CONCLUSIONS

The collapse of wet pulp fibres depends onpulp type . Chemically treated fibres, CMP andkraft, are more collapsed at low pressure thanmechanical pulp fibres, TMP and CTMP . Athigher pressures the chemically treated fibresapproach complete collapse while themechanical pulp fibres do not exceed 80%collapse even at 5000 kPa .

There is little difference in the response topressing pressure of thick-walled, stiff,southern pine TMP and thin-walled moreflexible spruce/balsam TMP. Transversecollapsibility and flexibility are thereforeindependent fibre properties which must beevaluated separately .

The pressure dependence of collapse hasimportant consequences for our understandingof the development of sheet structure duringthe pressing process . The pattern of fibrecollapse within the sheet is focused at fibre-to-fibre contacts where wet pressure istransferred . The most important z-directionaldimension of partially collapsed fibre withinthe pressed sheet, which is of consequence indetermining sheet density, is the fibrethickness at fibre contacts .

ACKNOWLEDGEMENT

The technical contributions of J . Mathieu andG . Stepien in measuring fibre properties aregratefully acknowledged . The contribution ofP . Lego, McGill University,- in the initialdevelopment of the fibre thickness measurementmethod is appreciated . The authors thank Dr .A . Karnis for useful discussions and guidancethroughout this work .

REFERENCES

1 . Kibblewhite, R.P . "The Fibres ofMechanical Pulps - Drying and LatencyEffects ." Appita 35 (3) : 216 (1981) .

2 . Kibblewhite, R.P. "Fibres and Fines ofsome Radiata Pine Corewood and SlabwoodThermomechanical and Refiner MechanicalPulps ." Appita 37 (8) : 650 (1984) .

3 . Kibblewhite, R.P ., Corson, S.R . andGraham, K.L . "Chemimechanical andThermomechanical Pulps of Radiata PineCorewood and Slabwood . Part 3 . - FactorsDetermining Paper Quality ." Appita 40 (2)121 (1987) .

4 . Heitner, C . and Hattula, T . "Ultra-HighYield Pulping . Part VI : The Effect ofSulphonation on the Development of FibreProperties ." JPPS 14 (1) : J6 (1988) .

5 .

Rioux , P . and Silvy , J . "Some FundamentalCharacteristics of -High Yield Pulps andCorrelation with Paper Properties ."

Int.Sym . Wood & Pulping Chem (1) :3 (1987) .

6 . Rahman, L., Tay, C.H . and Nye, J.W ."Neutral Sulphite/SAQ Cooking as a Meansof Upgrading TMP and TCMP Rejects." TappiProceedings of the Int . Sulfite PulpingConference, 57 :(989) .

7 .

Page, D.H . "The Collapse Behaviour of PulpFibres ." Tappi 50 (9) : 449 (1967) .

8 . Dinwoodie, J .M. "The Influence ofAnatomical and Chemical Characteristics ofSoftwood Fibres on the Properties ofSulfate Pulp ." Tappi 49 (2) : 57 (1966) .

9 .

Gorres, J., Sinclair, C.S . and Tallentire,A. "An Interactive Multi-Planar Model ofPaper Structure." Paperi ia Puu 71 (1) : 54(1989) .

10 . Gorres, J . and Luner, P . "An ApparentDensity Model of Paper." JPPS 18 (4) : J127(1992) .

11 . Amiri, R., . Wood, J.R ., Karnis, A . and

Gdrres, J . "The Apparent Density ofPaper ." Transaction of the CPPA AnnualMeeting, B7 : (1992) .

12 . Tappi Standard Method No . T-234 .

13 . Bichard, W .

and Scudamore,P .

"AnEvaluation of the Comparative Performanceof the Kajaani FS-100 and FS-200 FibreLength Analyzers ." Tappi 71(12) :149(1988) .

14 . Tasman, J.E . "The Fibre Length of Bauer-McNett Screen Fractions." Tappi 55(1) :136(1972)

15 . Kibblewhite, R.P . and Bailey, D.G ."Measurement of Fibre Cross-SectionDimensions Using Image Processing ." Appita41 (4) :297 (1988) .

16 . Jang, H.F ., Robertson, A.G . and Seth, R.S ."Measuring Fibre Coarseness and WallThickness Distributions with ConfocalMicroscopy." Transaction of the CPPAAnnual Meeting B189 : (1992) .

17 . Mistry, J . M.Sc . Thesis, University ofManchester Institute of Science andTechnology, 1985 .

18 . Jordan, B .D . "A Simple Image AnalysisProcedure for Fibre Wall Thickness." JPPS14 (2) : J44 (1988) .

19 . Nanko, H . and Ohsawa, J . "Mechanisms ofFibre Bond Formation." FundamentalResearch Symposium Cambridge 783 :(1989) . .

20 . Miller, P.R . "Fibre and Sheet PropertiesResulting from Refining StockConcentration Variation ." Appita 42 (2) :125 (1989) .

21 . Amiri, R., May, W.D . and Miles, K.B ."Design and construction of an Apparatusfor Measuring the Dynamic CompressiveStress-Strain of Pulp Samples." PPR 962,September, 1992 .

'

22 . Corson, S .R . and Kibblewhite, R.P ."Chemimechanical and Thermomechanical

Pulps of Radiata Pine Slabwood andCorewood . Part 2 : Characteristics ofFibres and Fines ." Appita 39 (5) : 379(1986) .

307

APPENDIX 1

Network Statistics of Blotter and Fibrecrossings

The presented results of fibre thicknessmeasurements suggest that the surfacestructure of the material backing a thin fibrenetwork during wet pressing affects the degreeand the uniformity of fibre collapse . A roughsurface, like the blotter surface shown inFigure 2A, transfers wet pressure mostdirectly where pronounced surface featurescome into contact with fibres of the thinnetwork . In these places the pressure exceedsthe pressure experienced by the fibre sectionsin other locations of the network so thatfibres in the network are more collapsed wherethey are crossed by fibres at the surface ofthe blotter . These locations represent a smallnumber of fibre sections in the network, whilethe majority of the fibre sections in thenetwork are less collapsed .

In contrast, when the network is pressed incontact with a smooth surface, like thesurface of the millipore filter shown inFigure 2C, pressure is transferred more,uniformly across the entire area of thenetwork . The pattern of collapse is expectedto be more uniform resulting in a lowerstandard deviation of fibre thickness comparedto when the network is pressed in contact witha rough surface .

In order to understand the effect of surfaceroughness of a backing material on the meanvalue of thickness, one needs to consider thefrequency with which the fibres at the blottersurface cross the fibres of the network. Thisfrequency can be estimated with networkstatistics . If the blotter surface layer isregarded as a two- dimensional network ofweight g, as defined by the IMPM, the number ofcrossings per unit length of network fibre isgiven by9, 1 ,11)

N= 2g171C

where C is the coarseness of the blotterfibres . For a softwood blotter, for which we

assume C = 0 .0002 g/m and a width of 3 6 Am,the layer weight ` is 0 .8 g/m2 . Thus, N is 2546per meter of network fibre which yields aboutone crossing per 400 hum of network fibre . Theaverage area of a crossing is

A,,= "w22

The average crossing length is approximatelythe square root of A~, which is 45 Am . Thefraction of network fibre length affected bythe high wet pressure expected at crossingswith the fibres at the blotter surface is 45Am divided by 400 gum, which gives 0.12 . Thismeans that about 88 % of the fibre length ofthe network fibres receive lower pressures . Webelieve that this results in a higher meanfibre thickness than one would expect whenpressure is more uniformly applied . Thisexplains the higher mean values obtained forfibre thickness when using the old method ofslide preparation .

FIBRE COLLAPSE AND SHEET STRUCTURE

R Amiri, J Gorres, et al

Dr A F Nissan, Westvaco Corporation, USAIn the last slide, the last sentence, what do you mean by fibre

collapse as an intrinsic property of mechanical pulp?

R AmiriWhat we wanted to emphasise is as fibre collapses from a

mechanical pulp it is not really related to flexibility . It is anindependent property, and we should look at it . It is something notnecessarily related to flexibility of fibres . It depends on the type ofpulp too .

A NissanWould not that apply to other pulps?

R ArniriWell for a chemical pulp in general they are very highly collapsed .For example for a kraft pulp, even at 350 kPa the starting pressurewe had almost 88 or 90% collapse . So you cannot really use that forchemical pulp it is basically for mechanical pulp which becomesimportant . We have a wide range of collapse of fibres so it is notreally applied to chemical pulps and that is what I wanted todistinguish between chemical and mechanical .

Transcription of Discussion

J F Waterhouse, Institute of Paper Science & Technology, USA.Have you looked at the relationship between Relative Bonded Area{RBA} and apparent density and could you also comment on otherthings that might be motivating this study?

R AmiriThis is part of our work which will be published shortly .

J F WaterhouseWould you be able to generally sketch what the relationship wouldbe between RBA and density for the pulps you have discussed inthe modelling work?

R AmiriUnfortunately I cannot talk about these results but they will bepublished very soon. What I can say is that we did not look at RBAfor the moment in the density model but at shear bond strength .

Prof D Wahren, Stora Teknik AB, SwedenWhat was the method employed to measure sheet thickness?

R AmidThe standard TAPPI method not soft platen . We use that becausewe feel that this is more in line with the IPM because I feel that whatwe say is the superposition of layers . So each layer is important sowe use that so the surface roughness should take that intoconsideration .

D WahrenThat would imply that for the measured density, had you used amore modern method, it would be even higher than indicated in yourdiagrams .

R. AmidThe caliper would be reduced .

D WahrenThese sheets were airdried . Were they prevented from wrinkling togive what MacGregor calls micro-striations?

R AmidThis is the standard handsheet so we used the same method ofhandsheeting dried under restraint .