-1

The effects of beating, web forming and sizing on the surface energy of

Eucalyptus globulus kraft fibres evaluated by inverse gas chromatography

M.G. Carvalho1,*, J.M.R.C.A. Santos2, A.A. Martins3 and M.M. Figueiredo11Department of Chemical Engineering, University of Coimbra, Pinhal de Marrocos, 3030-290 Coimbra,Portugal; 2Department of Chemical Technology, Polytechnic Institute of Braganca, Campus de SantaApolonia, 5301-857 Braganca, Portugal; 3RAIZ–Forest and Paper Research Institute, 3801-501 Eixo, Aveiro,Portugal; *Author for correspondence, e-mail: mgc@eq.uc.pt

Received 7 April 2004; accepted in revised form 13 December 2004

Key words: Eucalyptus globulus, IGC, Kraft pulps, Paper properties, Surface energy

Abstract

In this work attention has been focused on the effects of papermaking beating, web forming and sizingoperations on the physical/chemical surface properties of bleached Eucalyptus globulus kraft fibres. Inversegas chromatography (IGC) was used to determine the dispersive component of surface tension (cs

d) as wellas the acidic/basic character (according to the Lewis concept) of the solid surfaces (pulp fibres and pulphandsheets). The results have shown that the main effect of beating is to increase the fibre’s Lewis acidiccharacter. Web forming caused a strong decrease in cs

d and significant increments in the adhesion works ofboth acidic and basic probes, lowering the ratio between the two. Nevertheless, the surface of handsheetsstill exhibited a dominant acidic character. The sizing operation did not change the dispersive component ofthe surface tension significantly but decreased the difference between the adhesion works of the acidic andbasic probes, rendering the handsheet surface less Lewis acidic and more Lewis basic. Thus, althoughinternal sizing is expected to strongly influence liquid spreading at the paper surface and liquid penetrationof the fibre’s network, it is concluded that beating and web forming lead to important changes in the surfaceenergetic properties of the Eucalyptus globulus kraft fibres.

Abbreviations: A – molecular surface area of the probes used in the IGC study; Acet – acetone; AN* –Gutmann’s modified acceptor number; CHCl3 – trichloromethane; DCM – dichloromethane; DN – Gut-mann’s donor number; EtAcet – ethyl acetate; F – carrier gas flow rate (measured with a digital flow meter);GC – gas chromatography; IGC – inverse gas chromatography; J – correction factor for gas compress-ibility; Ka – surface Lewis acidity; Kb – surface Lewis basicity; N – Avogadro’s constant; R – ideal gas lawconstant; T – temperature in Kelvin; t0 – retention time of the non-interacting molecule; tr – retention timeof the probe molecule; THF – tetrahydrofuran; Vn – net retention volume; DG – free energy of adsorptionof the probe on the stationary phase surface; DGd – dispersive component of the free energy of adsorption;DGs – specific component of the free energy of adsorption; DHs – enthalpy of adsorption; DSs – entropy ofadsorption; cl

d – dispersive component of the surface free energy of the liquid; csd – dispersive component of

the surface free energy of the solid; csp – polar component of the surface free energy of the solid; Wa

s –specific component of the adhesion work between the polar probes and the stationary phase

Cellulose (2005) 12:371–383 � Springer 2005

DOI 10.1007/s10570-004-7738-0

Introduction

Eucalypt fibres are being used worldwide to pro-duce the highest quality printing and writing papergrades. This is due, firstly, to the characteristics ofthe eucalypt fibre itself. Compared to the mostcommonly used hardwood fibres, in general euca-lypt fibres exhibit lower fibre length and width butsimilar wall thickness, which leads to a highernumber of fibres per unit mass and a uniquecombination of stiffness and conformabilitycapacity upon beating. These features provide thepaper with high opacity, high bulk, excellent for-mation, excellent surface smoothness and enoughstrength, for good runability and superior endusage performance. However, it is well known thatthe specific conditions of pulp production (cookingand bleaching) affect the internal and the surfacechemical composition of the fibres, which is ofconsequence to the surface, structural, optical andstrength properties of the papersheet. Further-more, the papermaking operations of beating, webforming and sizing play an important role in thefinal quality of the paper. The beating operation (amechanical process in which fibres undergo highshear and compression forces in the presence ofwater) aims at improving bonding properties,which decreases the pulp drainability, the paperporosity and optical properties and increases thepaper mechanical resistance. Furthermore, in or-der to enhance paper resistance to liquid penetra-tion, chemical sizers like AKD (alkyl ketenedimer) are added to the stock.

Among other factors, differences in the surfaceenergy determine changes in the surface adhesiveproperties and spreading of liquids (like inks) aswell as changes in their penetration into the fibresnetwork. Since the fibre surface properties caninfluence the final product quality, it is not sur-prising that a large amount of research work isbeing carried out on the study of the surfaceenergetic characteristics of papermaking cellulosefibres and the interaction of these surfaces withother systems, such as printing inks (Banerjee1991; Santos et al. 2001). Adhesion between thesesystems is directly related to their capability tointeract by means of dispersive and acid–base in-termolecular interactions. Therefore, the charac-terisation of these phenomena is of majorimportance.

In what concerns eucalypt fibres, studies havebeen undertaken to investigate the effects ofcooking and bleaching conditions on the surfaceand strength properties of the corresponding kraftpulps (Shen and Parker 1999; Aquino et al. 2002).The study of Aquino et al. (2002) on Eucalyptusglobulus kraft fibres has revealed that these prop-erties are highly dependent on the history of thecooking and bleaching processes. For instance,high cooking temperatures strongly decrease thefibre surface free energy and the introduction of anoxygen delignification stage prior to bleachingenhances the Lewis acidic character of the fibres.However, the effect of subsequent papermakingoperations, like beating and web forming, on thefibre’s surface energy properties is not yet wellknown.With regard to the effects of AKD sizing,the work of Shen et al. (1998) and that of Lee andLuner (1989) are worth mentioning.

The purpose of the present study is to examinechanges in the surface energy of fibres due tobeating and sizing operations. Comparisons arealso made between fibre surface free energy of pulpand of handsheets in order to investigate the effectsof fibre bonding and collapse, and migration offines that occur at the web formation stage. Thepresent work is part of a more extensive workwhose main objective is to study the relationshipbetween chemical and physical properties of fibresand fibre networks and the functional properties ofthe paper, in order to assess the contribution ofthese relationships to the printing performanceof office papers.

The growing awareness of the importance ofsolid surfaces, interfaces and interphases in deter-mining the useful properties of polymeric systems,has led to the development of inverse gas chro-matography (IGC) as a useful technique in eval-uating the potential for interaction of differentcomponents of polymer blends, composites, andmulticomponent polymeric systems in general. Thefirst papers on IGC go back 32 years. Since then,and in particular since 1990, the number of pub-lications per year has increased exponentially.Data obtained from IGC experiments may, infavourable cases, correlate directly with observedperformance criteria, such as colour development,gloss, rheological properties, adhesion andmechanical properties (Lee 1991; Mukhopadhyayand Schreiber 1993; Schultz and Lavielle 1989).

372

Moreover, IGC has been increasingly used tocharacterize the surface thermodynamic propertiesof many inorganic and organic materials, likepolymers, pigments, coatings, wood meal, woodpulp, paper, etc. (Al Saigh 1994; Papirer and Ba-lard 1999; Santos et al. 2001; 2002a,b). Moreinformation on theoretical and practical aspects ofthis technique can be found in reviews such as thatof Papirer and Balard (1999) and that of Al Saigh(1994).

The IGC technique enables the evaluation ofseveral surface energy related parameters, namelythe dispersive component of the surface tension,arising from London and Van-der-Waals forces,and the acid/base character according to the Lewisconcept (i.e., ability to accept and donate elec-trons). IGC is very similar to conventional gaschromatography (GC), the main difference beingthe fact that the material under study is the sta-tionary phase introduced into the chromato-graphic column, whereas the compounds injected(probes) have well-known properties. As men-tioned above, in this work, the packed materialsunder investigation are pulp and pulp handsheets.

Materials and methods

Sample preparation

Eucalyptus globulus pulp was produced in a 7 llaboratory MK digester (model 409 MII from M/K Systems, Inc.) by the kraft process. The cookingconditions were 160 �C cooking temperature dur-ing 45 min, 1 �C/min heating rate, a liquor towood ratio of 4, 16% active alkali charge, asNa2O, and 30% sulphidity, resulting in a kappanumber of 14.8. The pulp was then bleached byapplying a DEDED sequence to achieve abrightness of 90% ISO (sample A). In order toanalyse the changes on the surface free energy dueto the beating process, a portion of the bleached

pulp was beaten in a laboratory PFI mill toachieve 30 Schopper Riegler (sample B). To studythe influence of sheet formation and sizing on thesurface characteristics of fibres, handsheets weremade (according to ISO 5269/1-1979) using thebeaten pulp with and without the sizing agentAKD, samples D and C, respectively. The sizingagent emulsion (with 13.5% active componentsupplied by RAISIO Chemicals Ltd.) was added tothe pulp suspension after addition of sodiumhydrogen carbonate (0.1%), which is the reactioncatalyst. After strong mixing, laboratory roundsheets were formed and subjected to drying (forAKD curing) during 10 min at 80 �C. The AKDamount was adjusted to obtain 180 s HST (Her-cules Sizing Test) level. Table 1 summarises thecharacteristics of the several samples tested.

Inverse gas chromatography

A Perkin Elmer 8410 gas chromatograph equippedwith a hydrogen flame ionization detector (FID)was used for IGC data collection. The instrumentwas connected to a Kompensograph Siemensintegrator. Stainless steel columns, 0.5 m long and0.4 mm ID, were degreased, washed and driedbefore packing. In order to reduce the particlessize, samples A and B were milled and samples Cand D were cut into small pieces (about2 mm·2 mm). On average, approximately 1.3 g ofmilled pulp and 2 g of handsheet pieces werepacked into the GC columns using a vacuumpump. The packed columns were conditionedovernight at 105 �C under a helium flow beforeany measurements were made. This procedureaims to remove any volatiles, including watermolecules, that may be adsorbed at the stationaryphase surface and that, consequently, could affectthe retention of the probe molecules. As all thecolumns are subject to the same pre-treatmentconditions, this procedure is not expected to

Table 1. Summary of sample characteristics.

Sample A Sample B Sample C Sample D

Sample type Pulp Pulp Handsheet Handsheet

PFI rpm to 30 SR 1200 1200 1200

Grammage (g/m2) 80 80

Sizing (mg AKD/10 g of pulp) 75

373

contribute to the observed differences in the sur-face energy of the samples analysed.

Experiments were carried out at column tem-peratures between 40 and 60 �C, at 5 �C intervals,in order to evaluate the variation of cs

d withtemperature as well as the enthalpies of adsorptionfor the polar probes. The injector and detectorwere kept at 180 and 200 �C, respectively. Heliumwas used as carrier gas and its flow was selected toensure that neither absorption nor diffusion of theprobes would occur inside the column stationaryphase. Minute quantities of probe vapour (<1ll)were injected into the carrier gas flow to ensurethat the experiments took place at infinite dilution.The probes were of chromatographic grade andused as received (Sigma–Aldrich Ltd.). Their re-levant properties are listed in Table 2. The resultsreported for the work of adhesion of acidic andbasic probes were estimated from the retentiontimes obtained for tetrahydrofuran (basic probe)and chloroform (acidic probe). Replicated mea-surements were performed with two columns pre-pared in the same way. Retention times were theaverage of at least five injections. Reproducibilitybetween runs was always better than 2%. Theretention times were determined graphicallyaccording to the Conder and Young method(Kamdem and Riedl 1992).

Natural gas with 83.7% methane was used todetermine the dead retention volume. It was en-sured that no extraneous peaks were obtained dueto the use of natural gas, the methane peak beingclearly identified, and peaks relating to othercomponents being not significant (as expected, asthe runs were carried out under infinite dilutionconditions).

IGC data analysis

IGC data processing was carried out according tocustomary methods, well described in the literature

(see for instance Chtourou et al. (1995)). In IGC, aninert carrier gas elutes a minute quantity of a probemolecule through a column packed with the mate-rial under study. Due to the interactions betweenthe two phases, the probemolecules are retained fora certain time, tr, which is used to calculate the netretention volume, Vn, according to Eq. (1):

Vn ¼ ðtr � t0ÞF � J: ð1Þ

Here t0 is the dead retention time of the markerprobe, F is the carrier gas flow rate (measured witha digital flow meter) and J is the correction factorfor gas compressibility (Belgacem et al. 1995).

Determination of the dispersive componentof surface tension

Assuming that experiments take place at infinitedilution, the free energy of adsorption of the probeon the stationary phase surface per mole, DG, canbe determined from the retention volume, Vn,according to

DG ¼ �RT lnðVnÞ þ C1: ð2Þ

Here R is the ideal gas constant, T is the absolutecolumn temperature and C1 is a constant thatdepends upon the chromatographic column and thereference state (Dorris and Gray 1980). Consider-ing that the dispersive and specific components,respectivelyDGd andDGs, are additive, as suggestedby Fowkes (1987), Eq. (2) can be rewritten as

DGd þ DGs ¼ �RT lnðVnÞ þ C1: ð3Þ

On the other hand, the free energy of adsorptioncan be related to adhesion work, Wa (Dorris andGray 1980), according to

�DG ¼ N � A �Wa: ð4Þ

Here N is Avogadro’s number and A the cross-sectional area of the probe to be tested (Table 2).If non-polar components (such as n-alkanes) are

Table 2. Characteristics of IGC probes (Kamdem et al. 1993; Santos et al. 2001).

Polar probes A (A2) cld (mJ/m2) ANa (kJ/mol) DN (kJ/mol) Alkanes A (A2) cl

d (mJ/m2)

CHCl3a 44.0 25.0 22.6 – C6H14 51.4 18.4

DCMa 31.5 27.6 16.3 – C7H16 57.0 20.3

THFa 45.0 22.5 2.1 84.2 C8H18 62.8 21.3

EtAceta 48.0 19.6 6.3 71.7 C9H20 68.9 22.7

C10H22 75.0 23.9

aPlease consult the list of abbreviations.

374

used, only dispersive interactions occur and theadhesion work is given by

Wa ¼ 2ðcds cdl Þ1=2: ð5Þ

Here csd and cl

d are the dispersive components ofthe surface tension of the solid (stationary phase)and of the liquid, respectively.

Replacing Eqs. (4) and (5) in Eq. (2) leads to

2Nðcds Þ1=2

Aðcdl Þ1=2 þ C ¼ RT lnðVnÞ: ð6Þ

The slope of the straight line, referred to as thereference line, obtained by plotting RTln(Vn) vs.2N(cl

d)1/2A, for a homologous n-alkane series(Figure 1(a)), leads to the determination of cs

d fora given temperature.

Determination of the surface acid–base properties

Acid–base characteristics of fibre surface aredetermined by analysing the interaction of thepolar probes with the solid surface and quantifyingthe deviation from the reference line, leading to theestimation of the specific free energy, DGs, as

�DGs ¼ RT lnðVnÞ � RT lnðVn;refÞ: ð7Þ

Here Vn,ref is the retention volume established bythe n-alkanes reference line (Eq. (1)), Vn being nowthe retention volume of the polar probes. Thiscalculation is also illustrated in Figure 1(a).

The adhesion work between the polar probestested and the fibres, Wa

s, can be evaluated fromthe specific free energy, given by Eq. (7), as

Wsa ¼

RT

N � A ln

�Vn

Vn;ref

�: ð8Þ

By carrying out experiments at different tem-peratures, it is possible to determine the enthalpyand entropy of adsorption, respectively DHs andDSs, from the plots of DGs=T vs. 1/T, according tothe following equation:

DGs

T¼ DHs

T� DSs: ð9Þ

The acidic and basic constants, Ka and Kb,respectively, are calculated using Eq. (10) from thelinear relationship DHs/AN* vs. DN/AN* for theseries of polar probes tested, characterized bydifferent AN* (modified acceptor) and DN (donor)numbers (Table 2). The procedure is illustrated inFigure 1(b).

ð�DHsÞAN�

¼ Ka

DN

AN�þ Kb: ð10Þ

The retention times obtained from the IGCexperiments are converted in net retention vol-umes, according to Eq. (1). For subsequent datareduction the net retention volume can be nor-malized by using the stationary phase surface areaor mass (the latter being the rule). In any case, andbearing in mind the data reduction mathematicalprocedures, it should be noted that the normali-zation of the net retention volume, using either thestationary phase mass or surface area, does notinfluence the values of cs

d or its variation withtemperature (dcs

d/dT), nor the values of Was, DHs,

Ka and Kb.The limitations of the IGC are well described

in the literature, and have actually been thesubject of discussion in a paper published by oneof the authors (Santos et al. 2002a). The best

Figure 1. Determination of (a) csd (Eq. (6)) and DGs (Eq. (7)) for the basic probe THF, and (b) acidic and basic constants (Eq. (10)) for

sample C.

375

approach to assess the surface energetic proper-ties (surface tension, surface Lewis acidic/basicproperties) of a material by IGC is the side-by-side analysis of the energy of adsorption,enthalpy of adsorption and surface Lewis acidityand basicity constants. This procedure allows fora coherent and systematic interpretation of theIGC results, leading to a more complete analysisthan the sole analysis of the semi-empirical val-ues of Ka and Kb. This was the procedureadopted in the present paper.

BET surface area determination

The BET specific surface area of the IGC columnstationary phases was determined by nitrogenadsorption in a Micromeritics ASAP 2000 unit.

SEM imaging

SEM images of the IGC column stationary phaseswere obtained in a JEOL JSM-5310 unit, using anaccelerating voltage of 10 kV.

FTIR analysis

FTIR spectra of the IGC column stationary pha-ses was carried out in a Nicolet Magma IR TM750unit (ATR MKII Golden Gate).

Contact angle measurement

An OCA20 Dataphysics unit was used for thedetermination of static contact angles relating tothe handsheets. The liquids used were water,formamide, diodomethane, ethylene glycol andpropylene glycol. The values of contact angle wereprocessed according to the Owens, Wendt, Rabeland Kaelble approach (OWRK method).

Results and discussion

BET surface area determination

In Table 3 are summarised the results concerningthe determination of the specific surface area of theIGC column stationary phases.

In Table 3 it can be observed that the beatingprocess applied to pulp A (production of sampleB) increases the specific surface area due to themore open structure and the production of fibrilsalong the fibre’s surface. Moreover, the specificsurface area of the handsheet stationary phases(samples C and D) is lower than that of the pulpsamples (A and B), as expected. The stationaryphase corresponding to the sized handsheet (sam-ple D) has a lower specific surface area than that ofthe non-sized handsheet (sample C). This may bedue to reduction in porosity caused by covalentbonding between the AKD molecules and thefibres and fines surface.

SEM imaging

SEM images of the IGC columns stationary pha-ses are presented in Figures 2–5, concerning sam-ples A, B, C and D, respectively.

As mentioned, samples A and B consist ofmilled pulp. The milling operation aimed atreducing the size of pulp flocks so that the IGCcolumn stationary phase could be packed. It canbe observed in Figures 2 and 3 that the effects ofthis procedure do not dominate over the effect ofthe PFI beating process (which results in a moreopen structure and the production of fibrils alongthe fibre’s surface). This is further supported bythe fact that even after the milling proceduresample B has a greater surface area than sampleA (Table 3). The consequences of the pulp mill-ing, if this operation had been extended, wouldbe increased exposure of the cross sectional sur-face and lumen surface in addition to thedetachment of fibre wall fragments. Conse-quently, the thermodynamic properties (cs

d, dcsd/

dT, Wa(THF), Wa(CHCl3), Ka and Kb) wouldreflect this varied composition and not that of theoriginal pulp because the milled pulp surface isexpected to be richer in cellulose and hemicellu-lose. Nevertheless, the above phenomena did not

Table 3. BET surface area (m2/g) of the IGC columns sta-

tionary phases.

Sample A Sample B Sample C Sample D

1.621±0.014 1.961±0.036 1.388±0.016 0.991±0.029

376

occur in the studied samples because of the gentlemilling.

In addition, morphological differences areobserved between sample B (beaten and milledpulp) and samples C and D (paper sheets madefrom beaten pulp), in particular the collapse of

the surface fibre’s and the presence of fines andcolloidal material at the fibre’s surface, due tothe web-forming operation. The effects ofthese phenomena on the surface energeticproperties will be discussed in detail further inthis article.

Figure 4. SEM images of the IGC stationary phase relating to sample C, at magnification levels of 750 · (on the left) and 3500 · (on

the right).

Figure 3. SEM images of the IGC stationary phase relating to sample B, at magnification levels of 750 · (on the left) and 3500 · (on

the right).

Figure 2. SEM images of the IGC stationary phase relating to sample A, at magnification levels of 750 · (on the left) and 3500 · (on

the right).

377

Inverse gas chromatography

The elution peaks obtained were generally sym-metrical. For probe molecules that adsorb stronglyon the stationary phase, such as THF, a tail wasdetected in the elution peak that did not influencethe symmetry of the peak significantly as themagnitude of the tail was prominently lower thanthat of the elution peak. The presence of anasymmetrical peak would be due to the existenceof dead volumes inside the IGC column or tosignificant surface energetic heterogeneity. In thatsituation, in order to account for the distributionof acidic/basic sites of various strengths, IGC un-der finite dilution conditions would be necessary.The Gaussian peaks observed indicate that theenergetic heterogeneity is not significant, for anyof the samples studied, under the operation con-ditions used for the IGC study. It should bementioned that IGC carried out under infinitedilution conditions is more sensitive to the higherenergy sites. IR spectroscopy was used in order todetect any contamination of the samples. Noextraneous peaks, related with contaminants, weredetected.

The retention times were not extrapolated topeaks of zero area as the volume of molecularprobe injected in the IGC columns was residualand the IGC unit was set to the limit of probe

detection by the FID detector. Typically, the syr-inge was filled with 1ll of gaseous probe, flushedwith air about 10 times, in order to ensure thecreation of a Henry’s infinite dilution region, andinjected manually.

Determination of the dispersive componentof the surface tension

The dispersive component of the surface tension,csd at 40 �C, determined according to Eq. (6), and

its variation with temperature (dcsd/dT), are pre-

sented in Table 4. Figure 6 illustrates the depen-dence of cs

d on temperature, for all the samplesanalysed.

Although the modifications to which pulp A wassubjected (beating, handsheet formation and siz-ing) greatly affect the sample surface energeticproperties, all the values obtained for cs

d areconsistent with those published in the literature forthis type of materials (Belgacem 2000). A Whil-elmy balance was used in an attempt to determinethe contact angles for the fibres, but, due to theirsmall length, it was not possible to prepare sam-ples such that reliable results could be obtained.With regard to the handsheets, the sessile dropletmethod was used but the size of the pores,roughness and hydrophilicity of the surface didnot allow for accurate results to be obtained in the

Figure 5. SEM images of the IGC stationary phase relating to sample D, at magnification levels of 750 · (on the left) and 3500 · (on

the right).

Table 4. Dispersive component of surface tension, csd, and its variation with temperature for all the samples tested.

Sample A Sample B Sample C Sample D

csd (mJ/m2), 40 �C 45.0±0.38, r2 = 0.995 48.2±0.48, r2 = 0.998 31.3±0.02, r2 = 1.000 33.4±0.04, r2 = 0.999

dcsd/dT (mJ/m2K) �1.06 �1.28 �0.61 �0.57

378

case of samples A, B and C. For sample D a valueof 33.9 mJ/m2 for cs

d and of 0.3 mJ/m2 for csp

were determined. Thus, the values of csd for sample

D, obtained using IGC and the sessile dropletmethod, are in good agreement.

The increased exposed surface area, due to theproduction of fibrils and to the more open struc-ture formed upon beating of the pulp (samples Aand B), is thought to result in increments in theamount of hydroxyl and carboxyl groups accessi-ble at the surface and, thus, of bonding sitesavailable for interaction with the IGC molecularprobes. Consequently, the surface ability toestablish dispersive interactions increases (highercsd). The dependence of cs

d on temperature(Figure 6) also increases for pulp B, denoting ahigher entropic contribution to the surface freeenergy, a consequence of the greater mobility ofmolecular segments, namely fibrils, in the case ofsample B.

The effect of web forming can be examined bycomparing sample B (milled pulp) and sample C(pieces of handsheet). The latter shows a muchlower value of the dispersive component of surfacetension and of its temperature dependence (abouthalf). This is most probably due to the enrichmentof fines and colloidal material at the surface of thehandsheets. Although the milling procedure,adopted in the case of samples A and B, couldinfluence the surface properties of the pulp, asdiscussed earlier (it would contribute, for example,to the higher values of cs

d observed for samples Aand B in comparison with those observed forsamples C and D), the web-forming operation is

thought to be the major cause of the differencesobserved upon this operation, based on the SEMimages (Figures 3–5) and on the IGC results(Figure 7 and Table 5). In Figures 4 and 5, it canbe clearly seen that in samples C and D the pres-ence of fines and colloidal material at the fibre’ssurface is very significant and, thus, these compo-nents are expected to influence very significantlythe handsheet surface thermodynamic properties.Conversely, at the surface of the beaten pulp(sample B, Figure 3) no colloidal material can beidentified (only the existence of fibrils originatedfrom the beating process).

The differences in fines and extractives concen-trations between the pulp and the handsheet fibressurface have been the subject of several studies.Three types of fines have been identified: primaryfines, associated with the virgin pulp, secondaryfines, produced during pulp processing prior to theforming operation, and tertiary fines, producedduring the flow of non-retained stock around thepaper machine white water system (Truong et al.2003). Shen et al. (1998) and Truong et al. (2003)state that residual lignin and extractives of thefines fraction are present in a higher concentration(at the surface as well as in the bulk) than they arein the fibres fraction. Furthermore, the concen-trations of lignin and residual extractives on thesurface of the fibres are not the same as those inthe bulk of the fibres, due to migration, resorptionand reprecipitation phenomena (Shen and Parker1999), during the cooking (Suurnakki et al. 1996)and the washing processes (Shen et al. 1998). Shenet al. (1998) have determined that the dispersive

Figure 7. Specific component of the adhesion works for the

basic (THF) and the acidic (CHCl3) probes on the surfaces of

samples A, B, C and D, at 40 �C.

Figure 6. Variation of the dispersive component of surface

tension with temperature for the various samples.

379

component of the surface tension of the eucalyptfines is lower than that of the whole pulp (34.7 and38.4 mJ/m2, respectively). In another study, Shenand Parker (1999) found the values of 56.6 and48.2 mJ/m2 for cellulose and lignin, respectively.Other researchers have also found that lignin has alower cs

d than cellulose (Belgacem 2000, Belgacemet al. 1995). Thus, it is concluded that the decreaseobserved in cs

d when the pulp is converted tohandsheets is due to reprecipitation/migration offines and colloidal material to the surface of thehandsheet. This phenomenon would be caused bywater flow during the web-forming process.

Regarding the influence of sizing, expressed bythe differences between the results obtained forsamples C and D, no significant change was de-tected for the surface free energy, despite otherauthors having reported a decrease in thisparameter with sizing (Shen et al. 1998; Belgacem2000). Possible explanations for this lack of liter-ature agreement are the differences in both thetype of sizing agent and procedure and the factthat, in this work, the effect of sizing was evaluatedin handsheets instead of pulp fibres. Thus, the ef-fect of sizing could be masked by the effect of webforming.

Acid–base characteristics

Although the discussion that follows is mainlybased on the results obtained for THF (as a modelbasic probe) and CHCl3 (as a model acidic probe),the acidic/basic character of the samples wasevaluated using all the polar probes referred to inTable 2, the results being consistent.

In Figure 7 are presented the values obtainedfor the specific component of the adhesion worksof THF and of CHCl3 on the surface of all thesamples analysed (derived from Eq. (8)). The val-ues obtained for the specific component of theenthalpy of adsorption of THF and of CHCl3(derived from Eq. (9)), and the values obtained for

Ka and Kb (calculated from Eq. (10)) are coherentwith the tendencies observed in Figure 7. The rel-ative values found for the specific component ofthe adhesion works and of the enthalpy ofadsorption of THF and of CHCl3, as well as therelative values found for the acidic and basicconstants (Ka and Kb, respectively) are presentedin Table 5.

From the results presented in Table 5, namelythe ratio Ka/Kb, it can be concluded that, excludingsample D, all the samples have a predominantacidic character which decreases in the followingorder:

sample B > sample A > sample C > sample D.

This tendency is corroborated by the values of theWa

s(THF)/Was(CHCl3) and DHs (THF)/

DHs(CHCl3) ratios, revealing a good qualitativecorrelation between the three approaches.

From IGC experiments Shen et al. (1998) haveshown that eucalypt pulp fibres are Lewisamphoteric with a strong acidic character. Ourresults are, thus, in agreement with expectationbearing in mind reports in literature and also thenature of cellulosic materials (extensive presence ofOH and COOH functional groups).

As far as the beating effect is concerned, theincreased OH and COOH groups concentration atthe fibre’s surface in the case of sample B (a con-sequence of the beating action as discussed above)results in increased works of adhesion with thepolar probes, namely with the basic probe (thusreflecting a significant increase in the Lewis acidiccharacter), Figure 7. The latter effect is also clearin Table 5. Here it can be observed that the pres-ence of acidic sites relative to that of basic sitesincreases upon the beating operation.

As mentioned previously, the differences insurface energy upon web forming are thought tobe due mainly to the increase in surface fines andcolloidal material concentration. The primary fineshave been shown to have a strong basic character,

Table 5. Ratios between the adhesion works (Was) and enthalpies of adsorption (DHs) of tetrahydrofuran and chloroform, and

between the acidic and basic constants (Ka and Kb, respectively), relating to samples A, B, C and D.

Sample A Sample B Sample C Sample D

Was(THF)/Wa

s(CHCl3), 40 �C 8.7 11.4 5.4 2.0

DHs(THF)/DHs(CHCl3) 17.5 21.6 4.1 2.7

Ka/Kb 2.0 3.1 1.8 1.0

380

in an order of magnitude similar to that of cellu-lose (Shen and Parker 1999; Shen et al. 1998). Onthe other hand, lignin has a lower acidity thancellulose. In addition, Shen and Parker (1999) andBelgacem et al. (1995), showed that extraction ofeucalypt fibres causes an increase of the works ofadhesion with both acidic and basic probes. Ourresults (Figure 7, samples B and C), however,show that web forming increases the works ofadhesion of both acidic and basic probes despitedecreasing the relative acidity of the surface (Ta-ble 5). The greater concentration of fines observedin sample C (when compared to sample B) con-tributes indeed to the increase of the work ofadhesion of the acidic probe (and to the decreaseof the surface relative acidity). The discrepancyrelated to the basic probe may be justified by (1)the fact that lignin and extractives are not the onlyconstituents of fines and colloidal material thatalso comprises carbohydrates, (2) the greaterdensity of COOH and OH groups in the case ofthe fines and (3) the preferential orientation ofthese groups at the handsheet surface due to theweb-forming operation conditions (towards thecellulosic surface used to absorb the water and toform the paper handsheet).

The last factor is important in many polymericsystems, as extensively illustrated by Schreiberet al. (1993). A good example is that of PMMAfilms prepared either by drying in air or by dryingbetween plaques of different materials. When thePMMA film is dried in air, the methacrylategroups are oriented towards the bulk of the film, inorder to minimize the PMMA film surface freeenergy (as the PMMA backbone is apolar). Whenthe PMMA film is formed between two plaques ofgreater surface free energy (in this case metallic),the methacrylate groups are preferentially orien-tated towards the PMMA film/metallic plaqueinterface, thus minimizing the interface surfacefree energy. The result is a greater interactioncapability of the PMMA film surface throughspecific (Lewis acid/base) forces. In the case of oursamples, the pulp, after being thoroughly washed,is dried in air. On the other hand, the web-formingoperation consists of placing wet handsheets be-tween two plaques of high water absorbency pa-per, followed by application of pressure.Afterwards, the sheets are dried in air. The factthat the pulp is dried in air and the handsheet isformed between two cellulosic substrates would be

expected to lead to differences in functional groups(hydroxyl and carboxyl groups) orientation at thesurface, and, in particular, to increased Lewisacidic and basic character in the case of thehandsheet. This reasoning must, nevertheless, bethe subject of further experimental evidence.

The fines have a much larger specific surfacearea than the fibres and have been shown to have amuch higher surface charge than the fibres, on amass basis (Truong et al. 2003). When fines areremoved, the eucalypt pulps studied by Truong etal. (2003) lose an average of 50% of their surfacecharge. The charge of the kraft pulp fibres andfines results from ionization mainly of carboxylicacid groups. The difference in COOH and OHfunctional groups per unit mass between the finesand the fibres can be explained by differences insize. Bearing in mind the typical dimensions of thefibres and of the primary fines, the surface chargeof fines (and, thus, the COOH and OH functionalgroups concentration), can be estimated to bearound 10 times that of the fibres. A practicalevidence is that a large amount (50–75%) of AKDsizing agent has been proven to be adsorbed by thefines in eucalypt pulps (Truong et al. 2003). Thus,the results obtained for the differences in the Lewisacid/base properties after web forming are thoughtto be due mainly to the higher concentration offines (with higher charge density) in the handsheetsurface.

It is clear from Figure 7 that the differences be-tween the adhesion works of the acidic and thebasic probes decreased from sample C to sample D,due to sizing (that is, the surface of the sizedhandsheet is less Lewis acidic and more Lewis basicthan the unsized sample). In addition, the ratio Ka/Kb not only decreased but is equal to unity forsample D, denoting that sizing renders the samplemore evenly bipolar. The loss of acidity and in-creased basicity with sizing is consistent with otherstudies reported elsewhere (Shen et al. 1998), and islikely to be related to the basic nature of AKD,which must have formed strong covalent bondswith the fibre’s surface, decreasing the presence ofacidic sites. It should be noted that the IGC resultsobtained in the present study do not indicate thatsizing promotes wetting as (1) the AKD moleculesform bonds with the fibres and, thus, decrease thesites for interaction with water or other molecules,and (2) the specific surface area (Table 3) decreasesupon sizing and so does porosity (SEM imaging,

381

Figures 4 and 5), thus, decreasing opportunity forsorption and spreading of liquids.

Conclusions

In this work, inverse gas chromatography wasused to study the influence of beating, web formingand sizing on surface free energy of bleachedEucalyptus globulus kraft fibres. Milled pulps orpieces of handsheets were used as the stationaryphase.

The results have shown that beating, althoughincreasing only slightly the dispersive componentof the surface tension, greatly enhances the acidiccharacter of the fibres surface, probably due to agreater number of hydroxyl and carboxyl func-tional groups accessible at the fibre’s surface. Thespecific surface area has increased as expected.

The process of web-forming decreases the spe-cific surface area and causes the retention of finesand colloidal material at the surface of the fibres.Consequently, this operation greatly decreases theability of forming dispersive intermolecular forcesbonds, resulting in significantly lower values forcsd. Nevertheless, the specific (acidic/basic) inter-

molecular interactions with the polar probesgreatly increase, probably due to increased hy-droxyl and carboxyl functional group concentra-tion and exposure at the papersheet surface (aconsequence of the greater concentration of finesand colloidal material).

Because of the basicity of AKD and also due tothe strong bonds formed between this sizing agentand the fibre surface, the availability of electron-accepting sites, relative to the electron donor sites,decreases and the surface of the sized handsheetbecomes more Lewis basic.

Although internal sizing plays an important rolein controlling liquid spreading at the paper surfaceand liquid penetration on the fibre’s network, it isconcluded that web forming is probably thepapermaking operation that mostly contributes tothe changes of the surface energetic properties ofthe cellulose fibres.

Acknowledgements

Thanks are due to Sapiens Program for the re-search grant (PE044-PE/PPQ/34159/99). The

skilful work of Enga A. Aquino in IGC analysis,Engo V. Redondo in BET surface area determi-nation, Engo J.P. Dias in SEM imaging and Enga

M.J. Travassos in FTIR analysis is also acknowl-edged.

References

Al Saigh Z.Y. 1994. Recent advances in the characterisation of

polymers and polymer blends using the inverse gas chroma-

tography method. Polym. News 19: 269–279.

Aquino A.I., Ferreira P.J., Carvalho M.G., Amaral J.L. and

Figueiredo M.M. 2002. Influence of cooking and bleaching

conditions on the surface and paper properties of Eucalyptus

globulus kraft pulps. Proceedings of the 2nd Iberoamerican

Congress on Pulp and Paper Research, CD-ROM, CIADI-

CYP, Sao Paulo, Brasil.

Banerjee R.K. 1991. Surface characterization of paper and

printability. IPPTA – Indian Pulp & Paper Technical Asso-

ciation Annual General Meeting & Seminar on Develop-

ments in Pulp mill and Paper Machine Clothings, New Delhi,

pp. 75–90.

Belgacem M.N. 2000. Characterization of polysaccharides,

lignin and other woody components by inverse gas chroma-

tography: a review. Cell. Chem. Technol. 34: 357–383.

Belgacem M.N., Czeremuszkin G. and Sapieha S. 1995. Surface

characterization of cellulose fibres by XPS and inverse gas

chromatography. Cellulose 2: 145–157.

Chtourou H., Riedl B. and Kokta B.J. 1995. Surface charac-

terization of modified polyethylene pulp and wood pulp

fibres using XPS and inverse gas chromatography. Adhes.

Sci. Technol. 9(5): 551–574.

Dorris C.M. and Gray D.G. 1980. Adsorption of n-alkanes at

zero surface coverage on cellulose paper and wood fibres.

J. Colloid Interface 77(35): 353–362.

Fowkes F.M. 1987. Role of acid–base interfacial bonding in

adhesion. J. Adhes. Sci. Technol. 1(1): 7–14.

Kamdem D.P. and Riedl B. 1992. Inverse gas chromatography

of lignicellulosic fibres coated with a thermosetting polymer:

use of peak maximum and Conder and Young methods.

J. Colloid Interface Sci. 150: 507–516.

Kamdem D.P., Bose S.K. and Luner P. 1993. Inverse gas

chromatography characterization of birch wood meal.

Langmuir 9(11): 3039–3044.

Lee L.H. 1991. Relevance of the density-functional theory to

acid-base interactions and adhesion in solids. In: Mittal K.L.

and Anderson J. (eds), Acid–Base Interactions: Relevance to

Adhesion Science and Technology. Utrecht, The Nether-

lands, pp. 25–46.

Lee H.L. and Luner P. 1989. Characterisation of AKD sized

papers by inverse gas chromatography. Nordic Pulp Paper

Res. J. 4: 164–172.

Mukhopadhyay P. and Schreiber H.P. 1993. Inverse gas chro-

matography for polymer surface characterization above and

below Tg. Macromolecules 26: 6391–6396.

Papirer E. and Balard H. 1999. Inverse gas chromatography: a

method for the evaluation of the interaction potential of solid

surfaces. Surfact. Sci. Ser. 80: 145–171.

382

Santos J.M.R.C.A., Gil M.H., Portugal A. and Guthrie J.T.

2001. Characterization of the surface of a cellulosic multi-

purpose office paper by inverse gas chromatography. Cellu-

lose 8: 217–224.

Santos J.M.R.C.A., Fagelman K. and Guthrie J.T. 2002a.

Characterisation of the surface Lewis acid/base properties of

the components of pigmented, impact-modified, PC/PBT

blends by inverse gas chromatography – phase separation

and phase preference. J. Chromatogr. A 969(1–2): 119–132.

Santos J.M.R.C.A., Fagelman K. and Guthrie J.T. 2002b.

Characterisation of the surface Lewis acid/base properties of

poly(butylene terephthalate) by inverse gas chromatography.

J. Chromatogr. A 969(1–2): 111–118.

Schreiber H.P. 1993. Aspects of component interactions in

polymer systems. In: Akovali G. (ed.), The interfacial inter-

actions in Polymeric composites. Kluwer Academic Publish-

ers, The Netherlands, pp. 21–59.

Schultz J. and Lavielle L. 1989. Interfacial properties of carbon

fiber-epoxy. In: Lloyd D.R., Ward T.C. and Schreiber H.P.

(eds), Inverse Gas Chromatography, Characterization of

Polymers and Other Materials. A.C.S, Washington, pp. 185–

202.

Schultz J. and Lavielle L. 1989. Interfacial properties of carbon

fiber-epoxy. In: Lloyd D.R., Ward T.C. and Schreiber H.P.

(eds), Inverse Gas Chromatography, Characterization of

Polymers and Other Materials. A.C.S, Washington, pp. 185–

202.Schultz J. and Lavielle L. 1989. Interfacial properties of

carbon fiber-epoxy. In: Lloyd D.R., Ward T.C. and Schreiber

H.P. (eds), Inverse Gas Chromatography, Characterization

of Polymers and Other Materials. A.C.S, Washington, pp.

185–202.

Shen W., Yao W., Li M. and Parker I.H. 1998. Characteriza-

tion of eucalypt fibre surface using inverse gas chromatog-

raphy and X-ray photoelectron spectroscopy. Appita J. 51(2):

47–51.

Shen W. and Parker I.H. 1999. Surface composition and surface

energetics of various eucalypt pulps. Cellulose 6: 41–55.

Suurnakki A., Heijnesson A., Buchert J., Viikari L. and Wes-

termark U. 1996. Chemical characterisation of the surface

layers of unbleached pine and birch kraft pulp fibers. J. Pulp

Paper Sci. 22(2): 43–47.

Truong B.Y., Shen W. and Parker 2003. I Effect of primary

fines and surface charge of hardwood pulps on AKD sizing.

Appita J. 56(1): 30–34.

383