This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

Colloids and Surfaces A: Physicochem. Eng. Aspects 341 (2009) 127–133

Contents lists available at ScienceDirect

Colloids and Surfaces A: Physicochemical andEngineering Aspects

journa l homepage: www.e lsev ier .com/ locate /co lsur fa

Dynamics of colloidal pitch adsorption at the solid–liquid interface by surfaceplasmon resonance

Gerard Murraya, Karen Stackb, Douglas S. McLeanb, Wei Shena, Gil Garniera,∗

a Australian Pulp and Paper Institute, Department of Chemical Engineering, Monash University, Clayton, Victoria 3800, Australiab School of Chemistry, University of Tasmania, Hobart, Tasmania 7001, Australia

a r t i c l e i n f o

Article history:Received 31 July 2008Accepted 28 March 2009Available online 5 April 2009

Keywords:PitchAdsorptionSolid–liquid interfaceSurface plasmon resonanceSPR

a b s t r a c t

This study has two objectives. The first is to investigate the direct adsorption of pitch on pulp fibres asa papermaking strategy. The second objective is to evaluate the potential of surface plasmon resonance(SPR) to quantify the adsorption of polydisperse colloids at the solid–liquid interface. The affinity ofcolloidal pitch for carboxyl methyl dextran (CMD) surfaces was studied by SPR and optical microscopy.The dynamics of adsorption and desorption of concentration pulses of colloidal pitch on carboxy methyldextran were followed at the solid–liquid interface by SPR. The parameters investigated were temperature,pitch concentration, colloid size and pitch build-up.

The direct adsorption of pitch onto pulp fibres was found to be a poor strategy for pitch control in paper-making. This is because of the very low chemical affinity of pitch for polysaccharide surfaces. Pitch can becarried by fibres as the low desorption rate might not allow full desorption during papermaking. Pitch hasa stronger affinity for pitch covered surfaces than for polysaccharides and first adsorbs as colloids—not asindividual fatty/resin acid molecules. Adsorbed pitch particles then serve as nucleation centres for furtherpitch adsorption. SPR is a suitable analytical technique to quantify the adsorption dynamics of polydis-perse colloids at the solid–liquid interface. The low signal response measured for the larger colloid wasexplained with the bouncing ball mechanism. A fractionation of the polydisperse colloid was observed,with the small particles being eluded first, followed by the bigger one.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

Pitch deposits have been a worldwide problem for the pulp andpaper industry ever since mechanical fibres were used. It is esti-mated that 1–2% of total annual sales, amounting to 3–5 billion USdollars, is lost due to pitch problems [1–4]. Pitch deposits arise fromthe agglomeration of resinous material released from the wood inthe pulping process. These deposits adversely affect paper machineefficiency and product quality. Furthermore, if large amounts of theresinous material are discharged into effluent, they pose an envi-ronmental problem because of their natural toxicity [1–8]. Thereare two basic strategies for pitch control [9–15]. The first is to sta-bilise the colloidal pitch and segregate it into the process water forlater removal in the water treatment system; the second is to desta-bilise and remove pitch in the process. An attractive pitch removalstrategy is through direct adsorption onto the pulp fibres, whereinpaper acts as a sink and pitch is sold as paper. This study investi-gates the feasibility of adsorbing colloidal pitch directly onto fibres.

∗ Corresponding author.E-mail address: Gil.Garnier@eng.monash.edu.au (G. Garnier).

A critical piece of information is the natural affinity of pitch for thecellulosic fibres.

Surface plasmon resonance (SPR) has been a critical analyti-cal technique in Biotechnology, wherein the specific affinity of amolecule carried in a liquid stream for another one immobilized ona surface is optimized. This study investigates the potential of SPRto characterise the adsorption dynamics of polydisperse industrialcolloids at the solid–liquid interface. This topic is critical in manyindustrial processes with applications not only in papermaking butalso in paints, food, pharmaceutical, and more recently, bio-surfaceengineering. There are a few fundamental issues to consider beforeusing SPR as an analytical tool for colloids application. What is theeffect of the micro-fluidic system or liquid transport on colloidsof various stability? Is SPR detection, probing a region extending100–200 nm from the surface, relevant to detect colloids and par-ticles orders of magnitude larger? What is the effect of colloidspolydispersity on the adsorption process?

In this study, the adsorption dynamics of pitch at the solid–liquidinterface is quantified by surface plasmon resonance (SPR) spec-troscopy. Carboxy methyl dextran (CMD) was selected as a modelsurface for cellulose and hemicellulose, two main polymers whichform micro-fibrils; it allows to only measure the retention dueto the chemical interactions, without any contribution from the

0927-7757/$ – see front matter © 2009 Elsevier B.V. All rights reserved.doi:10.1016/j.colsurfa.2009.03.050

Author's personal copy

128 G. Murray et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 341 (2009) 127–133

Fig. 1. Schematic representation of a pitch colloid [6,8].

Table 1Particle size of softwood colloidal wood resins.

Colloidal particle size (�m) Literature source

0.35 McLean et al. [18]0.8 Allen [19]0.2 Wågberg and Ödberg [12]0.54 Swerin et al. [20]<2 Shetty et al. [21]0.15–0.4 (mean 0.26) Sundberg et al. [22]0.15–0.4 Nylund et al. [23]0.2–2 Allen [24]0.1–1.5 Thornton [25]

physical entrapment. The effect of water temperature, pitch con-centration, particle size and a sequential adsorption process onpitch adsorption is quantified by surface plasmon resonance usinga BIAcore-X® instrument. The suitability of the SPR technique toquantify the adsorption of industrially relevant polydisperse col-loids at the solid–liquid interface is evaluated.

2. Background

A model of a pitch aggregate has been proposed in which theresin acid and fatty acids interact via hydrogen bonding to forman outer region around a core of the hydrophobic triglycerides andsteryl esters [6,8,16] (Fig. 1). The carboxyl and hydroxyl groups ofthe resin and fatty acids extend into the water and stabilise thepitch aggregate. It is this outer region that controls the chemicalreactivity, charge, colloidal stability and deposition tendency of thepitch aggregate. The driving force for the deposition of the aggre-gates is due to a combination of factors largely governed by processconditions. At low pH, when the pitch particles are uncharged,hydrophobic and hydrogen bonding type interactions control theattraction between the pitch aggregate and surfaces. At higher pH’s,near pH 7, electrostatic interactions may play some role as the resin

acids and fatty acids are partially ionized with the resin acids havingpKa above 6.5 and the fatty acids having colloidal pKa above 7.0 [17].The pitch particle size measured in a series of studies is presentedin Table 1; five of the nine colloidal particle sizes are reported as asize distribution. The maximum reported particle size is 2 �m [24].

3. Experimental

3.1. Methods

3.1.1. Surface plasmon resonanceSurface plasmon resonance (SPR) spectroscopy allows to quan-

tify adsorption at the solid–liquid interface in real time withoutlabelling techniques. The SPR system typically consists of a flow sys-tem, a sensor surface, a light source and a SPR detector. A polarizedlight is reflected on the gold coated sensor surface and detected bythe diode array sensor [27]. Light is reflected by a thin film of gold,stimulating the generation of resonance as the energy and momen-tum of the incident photons are converted to surface plasmons.Plasmons are electromagnetic waves resulting from the disturbancefrom equilibrium of large numbers of electrons in matter with a reg-ular structure. For our experiments, carboxy methyl dextran chainsare grafted on a gold film deposited over a glass slide (sensor chipCM5). Changes in refractive index in the micro-fluidic cell on theopposite side of the gold foil to the incident light, attenuates thesurface plasmon resonance. SPR spectroscopy probes the change inrefractive index caused by the adsorption of material in the cell.The resulting sensogram is a plot of the variation in SPR incidenceangle as a function of time, which represents the evolution of massadsorbed (pitch) at the solid–liquid interface.

The SPR used was a BIAcore-X®. The eluent is delivered via acontrolled syringe pump to a micro-fluidic flow cell. The eluent isproportioned between two channels (Fig. 2). Both channels flowover the sensor chip surface where the change in refractive indexof one, both or the difference between the two can be determinedas a function of time.

CM5 Sensor chips, having medium molecular weight chainscarboxy methyl dextran chains (around 100 nm) grafted on goldfilms, were purchased from BIAcore and used as is. 90 �L aliquotsof the pitch aqueous suspension were acquired using 20–200 �Lautopipette and injected directly into the inlet port of the BIAcore-X®. The eluent employed was Milli-Q® water filtered and degassedthrough a cellulose filter membrane under vacuum. This water alsoserved as diluent for stock solutions. The eluent was kept at roomtemperature and delivered to the sensor chip surface by flow con-duits and a syringe pump at a flow rate of 5 �L/min. Sensogramswere obtained at 20 and 30 ◦C.

Fig. 2. Schematic of the flow path of the BIAcore-X® showing the two channel flow system.

Author's personal copy

G. Murray et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 341 (2009) 127–133 129

3.1.2. Sample preparation for optical microscopyA conventional optical microscope equipped with a stage

micrometer was used to estimate the size range of precipitatedwood extractives particulates. A drop of aqueous suspension wasdelivered to a clean, dry microscope slide using a Pasteur pipette. Aclean cover slip was placed over the drop. An Olympus BX-60 opticalmicroscope was used to image particulates at different resolutions.Bright field lighting from above illuminated the samples. Imageswere captured with the SPOT Advanced software. Feature dimen-sions were added by calibrating the system with a stage micrometerand stamping the micrographs electronically.

3.2. Materials

3.2.1. Precipitation of wood extractives (large particles)Pitch extracts in acetone were obtained by modifying the

method described by Sundberg [14]. Wood resin from Pinus radi-ata TMP pulp was obtained by soxhlet extraction with hexane fora minimum of twenty cycles. The hexane was removed by rotaryevaporation and the wood resins were then dispersed in acetoneand stored at −4 ◦C until needed. A 10.79 mg/mL solution of woodextractives from P. radiata TMP pulp in acetone resulted [3–7]; itwas used as is in our adsorption study. A colloidal dispersion of pitchwas prepared by dispersing an aliquot of the wood resin in acetonesolution in distilled water. The natural pH of the aqueous pitch sus-pension was constant at around 5.5 and left unchanged; pH of theMilli-Q® water eluent was 6.5. For each SPR study, a colloidal dis-persion of pitch particulates was obtained by mixing various 200 �Lpitch/acetone aliquots (1, 2, 3 or 4) into 50 mL of Milli-Q® waterunder constant stirring for 10 min.

The composition of the wood resins was determined using a Var-ian 3800 Gas chromatogram using a 15 m × 0.53 mm i.d. × 0.15 �mfilm thickness ZB-1 GC column as previously described [7].

3.2.2. Preparation of small pitch emulsions400–1600 �L of Milli-Q water was delivered by auto pipette to a

clean, dry 2 mL glass vial. 200 �L of analytical grade acetone bear-ing pitch (10.79 mg/mL) was added to the water, already underagitation in a sonic bath. Sonication was completed at ambienttemperature for no more than 1 min. Fairly monodispersed colloids1 �m in diameter resulted as observed by microscopy.

4. Results

This study has two objectives. The first is to analyse the feasi-bility of directly adsorbing pitch particles onto pulp fibres prior topapermaking. The second is to investigate the potential of SPR asan analytical technique to characterise the adsorption of polydis-perse colloids at the solid–liquid interface. Our aim is to quantifythe chemical affinity of colloidal pitch for cellulosic surfaces, and tobetter understand the mechanisms by which pitch particles candirectly interact with fibres through chemical interactions. Theassociation and dissociation dynamics of pitch suspensions ontocarboxyl methyl dextran (CMD) surfaces was studied by surfaceplasmon resonance (SPR) spectroscopy. SPR was selected as it isnot only a real time technique which does not require a labellingtechnique, but also because it is one of the most sensitive methodsto probe the solid–liquid interface. SPR has the potential to mea-sure the adsorption down to the individual molecule of fatty acidor resin acid forming pitch. CMD was investigated as a surface mod-elling the cellulose and hemicelluloses fractions of mechanical pulpfibres. Dextran has the same chemical composition as cellulose.However, the anhydroglucose units are linked by a combination of�-glycosidic bonds: � 1-6 for the straight segments, � 1-3 and � 1-2for the branches, instead of the unique � 1-4 bond found for cellu-lose; this results in a more soluble and a branched macromolecule.

Fig. 3. Photograph of the micro-fluidic cell pressed over the CM5 carboxy methyldextran surface. In our experiments, only one cell was used.

The carboxylic acid groups further contribute to increasing watersolubility of carboxy methyl dextran.

In our study, concentration pulses of two model colloidal pitchsuspensions (small and large particles) were injected sequentiallyinto the aqueous stream flowing over the same CMD surface.The short carboxyl methyl dextran chains (100 nm) are covalentlygrafted over the thin gold film uniformly covering glass.

4.1. Surface plasmon resonance study

The photograph of the BIAcore micro-fluidic cell pressed againstthe CMD surface is shown in Fig. 3. The inlet and outlet have a diam-eter of 300 �m, and are 2100 �m apart. An oval cross-section isexpected to form.

In our surface plasmon resonance (SPR) study, four successivepitch pulses of increasing concentration were injected over thesame surface. This is to reproduce the effect of slow build-up onpitch deposition typically encountered during papermaking. Thesensograms are shown superposed (offset) for clarity. The senso-gram indicates the mass of pitch in the surface vicinity (100–200 nmthick interphase) as a function of time. Only a horizontal translationwas applied to the sensograms; this means that the initial value ofthe baseline of the second sensogram corresponds to the final valueof the first injection.

A sensogram is typically made of three distinct sections: (1) anabsorption phase, (2) a plateau, and (3) a desorption phase. Theshape of the initial sensogram adsorption phase results from thebalance of material transport and its affinity for the surface as theinitial front of the concentration pulse flows over the area moni-tored. At a given flow rate and concentration, the initial slope is adirect indication of the adsorption constant. In the second phase,a maximum concentration is reached, once the CMD interphaseis saturated with pitch, or once the equilibrium surface-liquid isreached. The shape and slope of the onset-plateau are mainly dic-tated by the balance of the adsorption and desorption constants.The duration of this second phase is established by the amount ofmaterial injected and the flow rate. The last section of the senso-gram corresponds to the drop in concentration associated with theend of the pulse and represents the material removal or the disasso-ciation phase. This phase best describes the affinity of the substancefor the surface. The faster this decay, the weaker is the interac-tion between the pitch and the surface. At the end of the pulse,the baseline typically decays, as the pitch particles are sheared offby the continuous liquid flow, and levels-off. A substance adsorbedstrongly or irreversibly on the surface has a higher final baseline

Author's personal copy

130 G. Murray et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 341 (2009) 127–133

Fig. 4. Superimposed sensograms of four consecutive pitch pulses of increasing con-centration (40, 80, 120 and 160 mg/L) on CMD surface at 20 ◦C. The clearer the line,the higher the concentration (black = 40 mg/L, pale grey = 160 mg/L). Large particles.

than its initial baseline; the higher the final baseline, the moresubstance remained on the surface.

The effect of pitch concentration and build-up at 20 ◦C on aCMD surface was investigated by injecting four successive pulsesof increasing concentration on the same carboxy methyl dextransurface. The superposed sensograms are shown in Fig. 4. Four obser-vations are of interest. The first is that the signal responses for thepitch pulses are all very low. The signal response is typically of 1000relative units (RU) per nanogram of material adsorbed per mm2

of BIAcore chip surface (1000 RU = 1 ng/mm2) [26]. SPR responsesoffering good sensibility/accuracy ratio usually range from 250 to5000 RU [27]. The increase in relative units achieved with pitchis only of 130. The second observation is that the sensograms areindependent of the concentration of pitch injected and are all verysimilar, at the exception of the first. The first injection interactsthe strongest and passivates the fresh CMD surface. All other pitchpulses are interacting with a surface previously modified by pitchdeposition. The third observation is that colloidal pitch adsorp-tion on CMD is very weak, as indicated by the rapid drop in pitchconcentration corresponding to the end of the injection pulse. Themicro-fluidic cell provides sufficient shear to remove most depositsduring the time frame of the experiment. Last, there is an unusualramp (0–200 s) at the beginning of each sensogram preceding thenormal sharp rise associated with the introduction/detection ofpitch pulse in the cell.

The effect of temperature on pitch deposition was also studied.The same experiment was performed at 30 ◦C. Four pitch pulses ofincreasing concentration were successively injected over the samecarboxy methyl dextran surface and the sensograms were mea-sured at 30 ◦C (Fig. 5); a new CM5 chip was used. A few observationsare worthwhile. First, the sensograms are all different and concen-tration dependent. Second, the response signal remains low andsimilar to that at 20 ◦C. Third, pitch adsorbs at 30 ◦C on CMD surfacesand the amount increases non-linearly with the pitch concentrationor with the amount of pitch previously adsorbed. This can best beseen with the increase level of the final baseline. The shear createdby the eluent flow slowly desorbs pitch but never completely; thefinal baseline never drops to its initial level. The increase in heightof the final baseline for each sequential pitch injection suggests thatpitch has a low affinity for the CMD surface but a higher affinity forpitch modified surfaces.

4.2. Pitch suspension characterisation

The pitch suspension was characterised by optical microscopy.A wide distribution of pitch particle size and shape was observed(Fig. 6). Spheres and plates were the most common shapes. The crit-

Fig. 5. Superimposed sensograms of four consecutive pitch pulses of increasing con-centration (40, 80, 120 and 160 mg/L) on CMD surface at 30 ◦C. The clearer the line,the higher the concentration (black = 40 mg/L, pale grey = 160 mg/L). Large particles.

ical dimension of the particles ranged from 1 to 40 �m (Fig. 6A).Larger particles were also observed (Fig. 6B). There is an orderof magnitude difference between the size distribution of particlesin this study and the 0.1–2 �m colloids reported in the literature(Table 1). Plate-like particulates like the one in Fig. 6B would beexpected from the recrystallisation of fatty acids.

Fig. 6. (A) Micrograph showing large and small pitch particulates magnified (100×).(B) Micrograph of a large pitch particulate magnified at 100×.

Author's personal copy

G. Murray et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 341 (2009) 127–133 131

Table 2Composition of pitch in standards as measured by GC analysis.

Concentration Components Total (mg/L)

Fatty acids(mg/L)

Resin Acids(mg/L)

Triglycerides(mg/L)

X 10 23 10 432X 20 46 20 863X 30 69 30 1294X 40 192 40 172

The chemical composition of the acetone–pitch solution suspen-sions was analysed by gas chromatography (GC). The pitch acetonesolution contained 2.61 mg/mL total fatty acids, 5.83 mg/mL totalresin acids and 2.37 mg/mL triglycerides; no steryl esters weredetected. The calculated composition of the aqueous pitch suspen-sions is presented in Table 2. The pitch particles were made of fattyacid, resin acid and triglycerides in the ratio 1:2:1.

4.3. SPR study with model pitch particles

To better understand the effect of colloid size on adsorption,model small pitch emulsions of controlled size were prepared usinga novel technique by interface solubilisation with acetone as aco-solvent. Basic microscopic observation revealed fairly monodis-perse particles with a diameter of around 1 �m.

The same experiment was performed and four consecutivepulses of increasing pitch concentrations were injected in the BIA-core over the same surface, and the sensograms were recorded at20 ◦C and superimposed in Fig. 7. A new CM5 chip was used. Afew observations are of interest. First, a good response signal wasachieved with signal RU of 350. Second, the sensograms are all dif-ferent and concentration dependent. All sensograms have similarinitial slope and plateaux increasing in height non-linearly withconcentration. This differs from the big pitch particles reported inFig. 4. Third, the sensogram shape is as expected; there is no ramppreceding the initial rise at the beginning of the injection, as previ-ously reported in Fig. 4. Fourth, the small pitch particles have a lowaffinity for CMD surface, and a low level of adsorption results. Thisis indicated by the level of the final baseline slightly higher than theinitial.

The interaction of the small model pitch suspensions was alsorecorded on CMD surfaces at 30 ◦C. Four consecutive pitch pulsesof increasing concentration were injected over a new CM5 chip andthe offset sensograms are represented in Fig. 8. The response signalswere relatively low, at 100 relative unit (RU), and fairly unstable asindicated by the shifting baselines and the appearance of spikes. TheSPR instability might indicate some pitch phase change driven by

Fig. 7. Superimposed sensograms of sequential pulses of increasing concentration(1, 2, 3 and 4 mg/L) of the small model pitch–acetone suspension at 20 ◦C on CMD.

Fig. 8. Superimposed sensograms of pulses of increasing concentration (1, 2, 3 and4 mg/L) of the small model pitch–acetone suspensions at 30 ◦C on CMD.

temperature. The level of the plateaus increased with the concen-tration of the pitch suspension injected. No significant permanentadsorption resulted, as shown by most final baselines abruptlydecreasing to the pre-injection initial level.

5. Discussion

5.1. Adsorption of polydispersed colloids by SPR

This study investigated the dynamic adsorption of polydispersecolloids at the liquid solid interface using surface plasmon reso-nance (SPR). Pitch was used as a polydisperse colloid model ofindustrial interest. SPR spectroscopy has typically been restricted towater soluble macromolecules – not micron size colloids – mainlyby fear of obstruction of the micro-fluidic system. While no plug-ging occurred and low noise to signal ratio were achieved in thisstudy, the validity of the technique needs to be evaluated and a fewcritical questions deserve further analysis.

The SPR response signals were much lower than expected (1/10).A first explanation is that pitch never reached the detection cell,having either adsorbed or coagulated and precipitated earlier in theinjector or in the micro-fluidic cell of the BIAcore. The increase ininjection pressure observed with time and the existence of 50 �mpitch particles would support this hypothesis. However, the systemremained surprisingly clean considering the total amount of pitchinjected in the study (as observed during clean-up and replacementof micro-fluidic cell after the study). The bouncing ball mechanism,illustrated in Fig. 9, might provide another explanation.

The BIAcore micro-fluidic cell flowing over the SPR detectoris 2 mm long and has an oval cross-section around 300 �m wide(Fig. 3). With the flow of 5 �L/min and assuming the maximum cir-cular cross-section (300 �m), the average fluid velocity in the cellis of 70 mm/min for a retention time of 2 s. A Reynolds number of0.4 results, which defines a perfectly laminar flow. However, thepitch colloids affect the flow, and particle rotation results, furthercontributing to the radial particle diffusion and transport. The zoneanalysed by SPR decays exponentially with distance and extendsaround 100–200 nm from the surface; this represents 100–2% of

Fig. 9. Schematic representation of flow induced particle separation.

Author's personal copy

132 G. Murray et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 341 (2009) 127–133

Fig. 10. Schematic representation of the bouncing ball concept.

the diameter of the colloids probed. Under these conditions, itis reasonable to assume that the pitch particles behave as ballsbouncing in and out of the thin probed region (Fig. 9). On aver-age, each non-adsorbing particle spends a small fraction of itsretention time into the detection zone, resulting in a low responsesignal/sensogram, similar to those measured. The SPR detectionzone at the solid–liquid interface being smaller or equal to the pitchdiameter could explain the reduced effect of colloid concentration.

Microscopy observation of the large pitch suspension revealeda polydispersed particles size with diameter ranging from sub-micron to more than 50 �m. Furthermore, we also observed for thepitch suspension an unusual initial ramp preceding the expectedsharp sensogram rise normally associated with the arrival of thepitch plug flow in the detection zone. This initial ramp was non-existent for the small monodisperse model pitch suspensions. Flowinduced fractionation would be an explanation. Fig. 10 illustratesthe phenomenon following which the small particles are firsteluded, followed by the bigger one, therefore creating an initialramp.

5.2. Pitch adsorption in papermaking

For the first time, the adsorption dynamics of colloidal pitchonto a polysaccharide surface was measured in real time at theliquid–solid interface in conditions relevant to papermaking. A glasssurface fully grafted with short carboxy methyl dextran chains(around 100 nm) was used as model pulp fibre to represent thecellulose and hemicellulose fractions. This model allows to onlymeasure the effect of chemical interaction, without mechanicalentrapment contribution. The hydrophilicity of CMD is expectedto be very similar to that of hemicellulose but higher than that ofcellulose. Sequential pulses of pitch were delivered to the CMD sur-face to better understand the mechanism of pitch build-up. Surfaceplasmon resonance (SPR) spectroscopy was selected as an analyti-cal technique because of its unique sensitivity (250 D) [26] havingthe potential to detect individual molecules of the resin acid andfatty acids forming pitch.

Pitch particles showed very little affinity for the CMD graftedsurfaces at neutral pH and 20 ◦C at the solid–water interface. Lit-tle permanent adsorption, if any, resulted at equilibrium when theCMD surface was submitted to a pulse of pitch colloids. This followsexpectations as there is little driving force for the hydrophobic pitchparticles to adsorb onto the hydrophilic polysaccharide surface todecrease their energy. The direct extrapolation to mechanical pulpfibres is difficult. Mechanical pulp fibres contain around 40% lignin,

a more hydrophobic poly(phenol propyl) polymer. Measuring theaffinity of pitch particle for lignin is required. Chemical pulp fibres,made entirely of cellulose and hemicellulose, are not expected tobe good adsorbents for pitch. Furthermore, mechanical pulping notonly releases pitch but also oligomers of hemicellulose, well-knownto adsorb onto pitch colloids [6,14]. The affinity of hemicellulosestabilised pitch for polysaccharide surfaces is expected to be sig-nificantly higher than that of pure pitch; this deserves furtherinvestigation. Conditions of high ionic strength, destabilising pitchand low pH (below COOH pKa of 4.6), preventing the deprotonationof the fibre’s carboxylic groups, could allow more pitch adsorption.While thermodynamics do not favour pitch deposition onto fibresat equilibrium, the kinetic process deserves further attention. Uponadsorption, the pitch colloids desorbed steadily when submittedeven to a low shear. Pitch adsorption, driven by the high collisionfrequency with the surface, is a very fast process, while desorptionis a slow one. In the timeframe of papermaking, this means thatmany pitch particles adsorbed or entrapped by the fibres might nothave sufficient time to desorb while on the moving wire of the papermachine (1–10 s) [28].

While the affinity of colloidal pitch for CMD at 20 ◦C is low, thepitch affinity for the pitch covered surfaces is significantly higher.Any pitch particle adsorbed onto pulp fibres will act as a nucleationcentre for other pitch particles to deposit. Since adsorption is fastand desorption is a much slower process, a substantial amount ofpitch could still be weakly adsorbed onto pulp fibres at the endof the forming fabric. However, these pitch clusters, having a highorthogonal projection and a weak adhesion with the pulp surface,will be very prone to desorb once submitted to shear, dilutionor a pH shock which affects pitch surface charge and water. Thedesorbed pitch might remain as aggregates of particles instead ofindividual stable colloids.

The adsorption of pitch particles was also shown to be verysize dependant. This suggests that shear plays a critical part in thepitch adsorption and desorption process, which is well known forfiller and fines retention [29]. Very little permanent pitch retentionresulted – if any – for the big pitch particles onto CMD at 20 ◦C; per-manent adsorption was slightly higher at 30 ◦C. However, low levelsof pitch remained adsorbed on CMD surface for the small pitch col-loids. The question is whether this pitch first adsorbs as colloidsor as individual molecules of fatty acid and resin acids. The con-centration of free molecules of fatty acid and resin acid dissolvedin water must be independent of the pitch colloid size, but depen-dent upon temperature as thermodynamics dictates the partitionequilibrium between dissolved molecules and colloids for the fatty

Author's personal copy

G. Murray et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 341 (2009) 127–133 133

and resin acid forming pitch. Therefore, the different levels of pitchstrongly adsorbed onto CMD observed between the small and thelarge particles must be due to colloid retention, instead of moleculeadsorption, as the concentration of dissolved fatty/resin acid is con-stant. This suggests that pitch first adsorbs on a surface as colloidsand not as individual molecules of fatty acid and resin acid. Thisdeduction is also supported by the near water insolubility of thepitch components [30].

6. Conclusion

This study had two objectives. The first one was to investigatea pitch control strategy; the second was to evaluate the potentialof surface plasmon resonance to quantify the adsorption dynam-ics of polydisperse colloids at the solid–liquid interface. The directadsorption of pitch particles onto pulp fibres was investigated asa pitch control strategy. For the first time, the adsorption dynam-ics of colloidal pitch onto a polysaccharide surface was measuredin real time at the liquid–solid interface in conditions relevant topapermaking. The adsorption dynamics of colloidal pitch onto CMDsurfaces was analysed by surface plasmon resonance (SPR) and opti-cal microscopy. Surfaces fully grafted with short carboxy methyldextran chains (around 100 nm) were chosen as a model of thecellulose and hemicellulose fractions of the pulp fibres; this wasto dissociate the effect of retention by chemical interaction fromthe physical entrapment. The effect of pitch concentration, size,temperature and sequential adsorption cycle was quantified.

The direct adsorption of colloidal pitch onto pulp fibres is nota viable option for pitch control strategy in papermaking. This isbecause of the very low chemical affinity of colloidal pitch forpolysaccharide surfaces. Virtually no pitch remained adsorbed atequilibrium over CMD at 20; slight retention occurred at 30 ◦C,probably due to mechanical entrapment of the soften pitch. In spiteof its low affinity for polysaccharides surfaces, colloidal pitch can becarried by fibres as the low desorption rate might not allow full par-ticle desorption during the time frame of papermaking. Pitch has astronger affinity for pitch covered surfaces than for polysaccharidessuch as carboxy methyl dextran. The proposed mechanism is thatpitch first adsorbs as colloids – not as fatty/resin acid molecules –which then serve as nucleation centres for further colloidal pitchadsorption. It is therefore critical for the papermaker to preventequipment passivation by keeping all surfaces clean. Pitch adsorp-tion slightly increased with temperature. This can be caused eitherby diffusion/reorientation of the molecules changing the surfacecomposition of pitch, or by mechanical entrapment of the softeningparticles.

SPR is a suitable analytical technique to quantify the adsorp-tion of polydisperse colloids at the solid–liquid interface. A lowersignal response was observed for the larger colloids (up to 50 �m),which can be explained with the bouncing ball phenomenon. A frac-tionation of the polydisperse colloid was observed, with the smallparticles being eluded first, followed by the bigger ones.

References

[1] D.E. Richardson, T. Parsons, J. van den Bosch, P.E. Harden, Factors affectingthe formation and control of pitch deposits in newsprint manufacture frommechanical pulp and recycled fibre, in: 50th Appita Annual General ConferenceProceedings, 1996, pp. 499–506.

[2] D.E. Richardson, T. Parsons, S. Jenkins, P.E. Harden, The use of chemicals toremove pitch in a newsprint mills, in: 51st Appita Annual General ConferenceProceedings, 1997, pp. 51–57.

[3] K.R. Stack, E.A. Stevens, D.E. Richardson, T. Parsons, S. Jenkins, Factors affectingthe deposition of pitch in process waters and model dispersions, in: 52nd AppitaAnnual General Conference Proceedings in Brisbane, Australia, 1998, pp. 59–66.

[4] K.R. Stack, E.A. Stevens, D.E. Richardson, T. Parsons, S. Jenkins, Factors affect-ing the deposition of wood resins, in: 54th Appita Annual General ConferenceProceedings, Melbourne, Australia, 2000, pp. 567–570.

[5] D.S. McLean, The Effect of Wood Extractive Composition on Pitch Deposition,School of Chemistry, MSc Thesis, University of Tasmania, Hobart, 2003.

[6] D. Vercoe, Study of the Interactions Leading to Wood Resin Deposition, Schoolof Chemistry, PhD Thesis, University of Tasmania, Hobart, 2004.

[7] D.S. McLean, K.R. Stack, D.E. Richardson, The effect of wood extractives compo-sition, pH and temperature on pitch deposition, Appita J. 58 (1) (2005) 52–55,76.

[8] M. Qin, T. Hannuksela, B. Holmbom, Physio-chemical characterisation of TMPresin and related model mixtures, Colloids Surf. A: Physiochem. Eng. Aspects221 (1–3) (2003) 243–254.

[9] D. Richardson, N. Moore, A. Featherstone, T. Parsons, R. Lorz, A. Esser, P.Baumann, J. Adlam, The use of chemicals to fix pitch to fibre in newsprint man-ufacture, in: 56th Appita Annual General Conference in Rotorua, APPITA, NewZealand, 2002, pp. 265–272.

[10] D. Richardson, N. Waller, T. Parsons, J. Stallard, M. Young, T. Watkins, A.Dechandt, Optimisation of neutral papermaking wet end chemistry for pitchfree newsprint manufacture, in: 57th Appita Annual General Conference Pro-ceedings in Melbourne, Australia, 2003, pp. 219–226.

[11] D. Richardson, M. Grubb, Extractives removal from newsprint process waters bydissolved air flotation, in: 58th Appita Annual General Conference Proceedingsin Canberra, Australia, 2004, pp. 79–84.

[12] L. Wågberg, L. Ödberg, The action of cationic polyelectrolytes used for the fixa-tion of dissolved and colloidal substances, Nordic Pulp Paper Res. J. 6 (3) (1991)127–135.

[13] J. Kekkonen, H. Lattu, P. Stenius, Formation, adsorption and interactions of poly-DADMAC/pectic acid complexes, J. Pulp Paper Sci. 28 (1) (2002) 6–12.

[14] K. Sundberg, J. Thornton, B. Holmbom, R. Ekman, Effects of wood polysaccha-rides on the stability of colloidal wood resin, J. Pulp Paper Sci. 22 (7) (1996)J226–J230.

[15] L. Maher, K. Stack, D. McLean, D. Richardson, Adsorption behaviour of cationicfixatives and their effect on pitch deposition, 59th Appita Annual General Con-ference Proceedings in Auckland, New Zealand (2005), pp. 115–122.

[16] D. Vercoe, K. Stack, A Blackman, D. Richardson, A multicomponent insight intothe interactions leading to wood pitch deposition, Appita J. 58 (3) (2005) 208–213.

[17] D. McLean, D. Vercoe, K. Stack, D. Richardson, The colloidal pKa of lipophilicextractives commonly found in Pinus radiata, Appita J. 58 (5) (2005) 362–366.

[18] D.S. McLean, K.R. Stack, D.E. Richardson, P.R. Haddad, Wood pitch fixative selec-tion by laser particle size analysis, in: 60th Appita Annual General ConferenceProceedings, Melbourne, Australia, 2006, pp. 413–416.

[19] L.H. Allen, Pitch particle concentration: an important parameter in pitch prob-lems, Trans. CPPA 3 (2) (1977) TR32–TR40.

[20] A. Swerin, L. Ödberg, L. Wågberg, Preparation and some properties of the col-loidal pitch fraction from a thermomechanical pulp, Nordic Pulp Paper Res. J. 8(3) (1993) 298–301, 337.

[21] C.S. Shetty, C.S. Greer, G.D. Laubach, A likely mechanism for pitch depositioncontrol, TAPPI J. 77 (10) (1994) 91–96.

[22] K. Sundberg, C. Pettersson, C. Eckerman, B. Holmbom, Preparation and proper-ties of a model dispersion of colloidal wood resin from Norway Spruce, J. PulpPaper Sci. 22 (7) (1996) J248–J252.

[23] J. Nylund, H. Byman-Fagerholm, J.B. Rosenholm, Physico-chemical characteri-zation of colloidal material in mechanical pulp, Nordic Pulp Paper Res. J. 8 (2)(1993) 280–283.

[24] L.H. Allen, Characterization of colloidal wood resin in newsprint pulps, ColloidPolym. Sci. 257 (5) (1979) 533–538.

[25] J. Thornton, Dissolved and Colloidal Substances in the Production of Wood-Containing Paper. Department of Forest Products Chemistry, Faculty ofChemical Engineering, PhD thesis, Åbo Akademi University, Åbo, 1993,p. 33.

[26] Surface Plasmon Resonance – Technology Note - 1, http://www.biacore.com/lifesciences/index.html, date of access February 2007.

[27] M.-I. Aguilar, D.H. Small, Surface plasmon resonance for the analysis of �-amyloid interactions and fibril formation in Alzheimer’s disease research,Neurotoxicity Res. 7 (1,2) (2005) 17–27.

[28] G. Murray, K. Stack, D. McLean, W. Shen, G. Garnier, Mechanism of pitch adsorp-tion on carboxy methyl dextran surfaces, Appita J. 62 (1) (2009) 64–69.

[29] T.G.M. van de Ven, Physicochemical and hydrodynamic aspects of fines andfiller retention, in: Fundamentals of Papermaking, 1:471, Transactions of the9th Fundamental Research Symposium, Cambridge, 1989.

[30] The Merx Index, 11th edition, Merck &Co., Rahway, NJ, 1989.