ARTICLE pubs.acs.org/Langmuir
Quantitative Analysis of Cationic Poly(vinyl alcohol) Diffusion into the Hairy Structure of Cellulose Fiber Pores: Charge Density Effect Pedram Fatehi,*,†,‡ Huining Xiao,† and Theo G. M. van de Ven§ †
Department of Chemical Engineering, Limerick Pulp and Paper Centre, University of New Brunswick, Fredericton, NB, Canada E3B 5A3 ‡ Deaprtment of Chemical Engineering, Faculty of Engineering, Lakehead University, Thunder Bay, ON, Canada P7B 1B8 § Department of Chemistry, Pulp and Paper Research Centre, McGill University, Montreal, QC, Canada H3A 2A7 ABSTRACT: The diffusion of charged polymers into the pores of cellulose fibers has not yet been fully understood due to the complexity of the interaction between polymers and fibers. In this paper, the diffusion of cationic-modified poly(vinyl alcohol) (CPVA) with tailored charge densities and a relatively high molecular weight into the pores of bleached aspen high-yield pulp (via a chemi-thermomechanical pulping process) was quantitatively investigated via an adsorption analysis, charge density analysis, and solute exclusion technique (SET). The results showed that the adsorption of the low-charged CPVA was substantially higher than that of the highcharged CPVA on fibers. The surface charge density analysis confirmed that approximately 17 mg/g of the high-charged CPVA adsorbed on the outer surface and on the macropores of fibers and the remaining (23 mg/g) diffused into the pores. The SET analysis confirmed that the pore size of fibers was more significantly reduced by applying the low-charged CPVA than the high-charged one. The influencing factors for the diffusion of CPVA into the large and small pores were related to the repulsion force developed between the adsorbed polymers and approaching polymers, entropy increase, and the polymer flexibility. The BrunauerEmmettTeller surface area analysis showed an increase in the surface area of fibers upon CPVA adsorption. It was proposed that the diffused CPVA prevented complete fiber pore collapse during drying, which eventually increased the surface area of fibers.
1. INTRODUCTION The adsorption of charged polymers on cellulose fibers has been extensively investigated in the literature.16 The characteristics of charged polymers significantly affect the amount of polymers adsorbed on cellulose fiber.18 Among those characteristics, the charge density of polymers was reported to affect the hydrodynamic size of polymers in solutions,9,10 the interaction of charged polymers with charged surfaces (e.g., cellulose fibers),11,12 and the configuration of polymers upon adsorbing on a charged surface.1315 However, the diffusion and adsorption of charged polymers into the pores of cellulose fibers have not been fully understood or characterized. The diffusion of charged polymers occurs if the electrostatic driving force developed between the opposite charges of polymers and fibers can compensate for the entropic loss associated with the confinement of polymers in the pores.1619 It was recently reported that the polymer flexibility affected the diffusion of cationic charged polymers into the fiber pores.9,10,19 It is well-known that the electrostatic interaction affects the adsorption of charged polymers on fibers. This interaction may also contribute to the diffusion of charged polymers into the fiber pores. The labeling technique has also been applied to investigate the diffusion of polymers into the fibers.9,10,19 Although it directly reveals the diffusion of polymers into the fiber pores, it cannot determine the changes in the porosity of the fiber wall upon polymer diffusion. r 2011 American Chemical Society
The diffusion can be determined by quantifying the amount of charged polymers adsorbed on, or diffused into, fibers via a surface charge analysis. In this method, adsorption analysis indirectly characterizes the location of polymers on fibers or inside the pores.1,20 One objective of this study was to assess how the diffusion of cationic-modified poly(vinyl alcohol) (CPVA) having two different charge densities could be quantified by employing the adsorption analysis. Another method to investigate the porosity of fibers is the solute exclusion technique (SET). In this method, the porous structure of the fiber wall is probed with macromolecules, e.g., dextrans, that do not adsorb onto the fibers. The idea is that the molecules of a given size will not penetrate into the pores smaller than the size of the molecules. By using a series of dextran molecules of different sizes, the pore size distribution can be probed.2123 However, this method is invalid unless the interaction between the dextran polymers and fiber surface is negligible. In our previous work, we demonstrated that the interaction of dextran molecules with CPVA was marginal,24 which facilitated the application of dextran molecules to investigate the changes in the pore size of fibers after CPVA modification. The SET analysis can be conducted if the polymers have a fixed configuration on the fibers or inside the pores. Therefore, since Received: April 26, 2011 Revised: September 29, 2011 Published: September 30, 2011 13489
dx.doi.org/10.1021/la203364x | Langmuir 2011, 27, 13489–13496
Langmuir polymers may reconfigure or migrate from the fiber surface to the fiber pores after adsorption, the SET analysis should be conducted at the equilibrium adsorption state. Another objective of this study was to employ and modify the concept of the SET method to quantitatively determine changes in the pores size of fibers upon CPVA diffusion. As is well-known, the lignin and hemicelluloses form a matrix in the cellulose fiber wall. The removal of lignin during pulping and bleaching results in pores with irregular shapes and hairy structures originating from the dissociation of hemicelluloses from the wall of the pores. The adsorption and diffusion of charged polymers will eventually affect the characteristics of these hairy pores. Another objective of the present study was to quantitatively relate the diffusion of CPVAs to the changes in the structure of these hairy pores and to the surface area of fibers in wet and dry states.
2. MATERIALS AND METHODS 2.1. Materials. Poly(vinyl alcohol) (PVA, with MW = 146186 kDa, 99% hydrolyzed, polydispersity index = 1.42), glycidyltrimethylammonium chloride (GTMAC; 75% in water), α-glucose and dextran polymers (with various MWs, i.e., 911, 3545, 6474, 100200, 400500, and 2000 kDa), and poly(diallyldimethylammonium chloride) (PDADMAC; with MW = 200400 kDa) were all obtained from Aldrich Co. and used as received. Anionic poly(vinyl sulfate) (PVSK) with MW = 100200 kDa (97.7% esterified) was provided by Wako Pure Chem. Ltd., Tokyo, Japan. Poly(ethylene oxide) (PEO) with MW = 22 kDa was obtained from Viscotek, Houston, TX. Aspen high-yield pulp (HYP), which was produced via a chemithermomechanical pulping process (CTMP) followed by a two-stage peroxide bleaching, was obtained from Tembec Inc., Temiscaming, QC, Canada, and used as received. The pulp was first soaked under acidic condition (pH 2) for 30 min via adding HCl (40%) to the fiber suspension (3% consistency) and then washed thoroughly with deionized and distilled water twice. The conductivity of the pulp suspension after washing was below 8 μS/cm. Then the fibers were filtered with a 200 mesh screen to remove fines. The moisture content was measured according to TAPPI T 412. The pulp freeness was measured by using a Canadian Standard Freeness (CSF) according to TAPPI T 227. The pulp characteristics were determined using a fiber quality analyzer (FQA), OpTest Equipment Inc., Hawkesbury, ON, Canada. Originally, the HYP contained fibers with a fiber length (length-weighted average) of 0.69 mm, fines content of 45 wt %, and a CSF of 620 mL. The total charges of fibers were determined via a conductometric titration.25 In this set of experiments, unmodified HYP fibers were titrated by a 836 Titrando titrator (Metrohm, Zofingen, Switzerland) with a 0.005 M NaOH solution at a rate of 0.05 mL/min until the mixture had reached pH 11. Then, the total carboxylic and sulfonate groups of fibers were determined from the conductivity curves.25 2.2. CPVA Preparation. The cationic PVAs with tailored charge densities were prepared according to our procedure established earlier.3,26,27 At first, PVA (7.25 g) was dissolved in water (60 mL) at 80 °C and stirred for 1 h. Then, 5 mL of NaOH (5 N) was added to the solutions. Afterward, GTMAC was added to PVA solution with the molar ratio of 0.1:1 or 0.5:1. The mixtures were stirred for 60 min at 75 or 98 °C, respectively, to produce low- or high-charged CPVAs. Unreacted GTMAC was separated by dialysis tubes with a MW cutoff of 1000, while changing water every 2 h for the first 6 h and then once a day for 2 days. To measure the degree of substitution (DS) of the hydroxyl group by GTMAC along the PVA backbone, CPVAs were first dried at 105 °C overnight. Then, 1H NMR spectra were recorded in D2O at 25 °C on an Oxford 300 MHz spectrometer operating at 300.13 MHz
ARTICLE
for nuclei. The DS was quantified on the basis of the proton peak area contributed from the CH3 of quaternary ammonium groups attached on a PVA chain and that from CH of the PVA backbone.3,27 The charge density analysis and the DS measurement were conducted 3 times, and the average results were reported in this work. 2.3. GPC Characterization. The hydrodynamic radius and the molecular weights of CPVAs and dextran polymers in aqueous solutions were measured using a gel permeation chromatography (GPC), Viscotek GPCmax VE2001 (Houston, TX, USA) coupled with an online VE 3580 refractive index (RI) detector and Viscotek 270 dual laser light scattering detector. The columns were poly[analytic], AquaGel series, PAA 202, 204, and 206 in series. The degassed NaNO3 solution in water (0.05 M) was used as an eluent.9,10 The polymer solutions, 4 mg/mL in concentration, were filtered using 0.2 μm pore size syringe filters prior to analysis. The flow rate and the operating temperature were 1 mL/min and 35 °C, respectively. Standard PEO samples were used for the GPC calibration. The hydrodynamic radius of polymers is estimated according to 3 ½ηM 1=3 ð1Þ Rh ¼ 4π 0:025 where [η] is the intrinsic viscosity (in units of liters per gram) and M is the viscosity-average molecular weight of the polymer sample. The detail of such measurements has been well-documented in the literature.2830 2.4. Adsorption and Charge Density. It is well-known that the ionic strength of solution affects the flexibility and chain length of polymers, both of which affect the adsorption characteristics. The ions can screen a part of the electrostatic repulsion force developed between the charged groups of polymers in polyelectrolyte solutions, thus enhancing the polymer flexibility and decreasing the polymer length. However, to simplify our experimental setup, we conducted our analysis in a salt-free solution. To investigate the adsorption of PVA on fibers, 40 mg of the unmodified PVA was mixed with 1 g (oven-dried (o.d.)) of fibers in the suspensions at 3% consistency and neutral pH in 125 mL Erlenmeyer flasks and shaken in a water bath shaker (Innova 3100, New Brunswick Scientific) at 30 °C for 24 h in one set of experiments. The supernatant was collected via filtering and the adsorption of the PVA on fibers was determined via using iodine calorimetric technique by a UV spectrophotometer.31 To investigate the adsorption isotherms of CPVAs on cellulose fibers, various amounts of CPVAs were mixed with approximately 1 g (o.d.) of fibers in the suspensions at 3% consistency and neutral pH in 125 mL Erlenmeyer flasks.3,26,27 The mixtures were then shaken in the water bath shaker at 30 °C for 24 h. To investigate the adsorption kinetics of CPVAs on cellulose fibers, 90 mg/g of CPVAs was added to the fiber suspension under the abovementioned conditions at various time intervals. Then, the samples were filtered and the supernatants were collected for adsorption analysis. Control samples without fibers were prepared under the same conditions. The adsorption amounts were calculated on the basis of the concentration differences of CPVA in filtrates and in their corresponding control samples according to3,26,27 C¼
ðV1 V2 ÞmH V1 m f
ð2Þ
where C is the amount of CPVA adsorbed (mg/g), V1 and V2 are the volume (mL) of PVSK used for the titration of the blanks and the samples. The mH and mf are the mass (oven-dried) of CPVA and fibers used in the experiment. The concentration of CPVA in supernatants was determined by titrating with the PVSK solution (0.5 mM) by using a particle charge detector, M€utek PCD 03 (Herrsching, Germany).9 The surface charge densities of fibers were determined via a backtitration method.3,32,33 In this method, PDADMAC or PVSK was used 13490
dx.doi.org/10.1021/la203364x |Langmuir 2011, 27, 13489–13496
Langmuir to neutralize the surface charge density of unmodified or CPVAmodified fibers. In one set of experiments, 50 mL of PDADMAC solution (0.2 mM) was mixed with 0.2 g (o.d) of unmodified and CPVA-modified fibers under the conditions of 30 °C and 150 rpm for different time intervals. Then, the fibers were filtered and the concentrations of PDADMAC in the solutions were determined using the PCD titrator. This analysis helped us determine the adsorption behavior of PDADMAC on unmodified and CPVA-modified fibers. In another set of experiments, an equivalent to 0.2 g (o.d.) of unmodified fibers or fibers modified with CPVA was added to 50 mL of PDADMAC or PVSK solution (0.5 mM) and mixed for 2 h. After filtering, the supernatants were collected and titrated by the abovementioned PCD. Then, the surface charge density of fibers was calculated on the basis of the concentration difference of PDADMAC or PVSK in supernatants and in their corresponding control samples. In the literature, it was reported that the interaction of high-charged polymers, e.g., PDADMAC or PVSK, with the cellulose fiber is stoichiometrically controlled.14,34 Three repetitions were conducted to get an average value for each sample. 2.5. SET Analysis. In our previous work, we reported that CPVA had no detectable interaction with dextran polymers, which facilitated the application of dextran polymers to investigate the changes in the pore size of fibers modified with CPVAs.24 The inaccessible water (σ) of fibers for dextran polymers was assessed by the SET method:20,22,23 At first, a dextran polymer was diluted to a concentration of 30 g/L. Then, 30 mL of the solution was added to the unmodified fibers or fibers modified with CPVAs for 24 h to create the fiber suspension of 3% consistency. The suspensions were stirred at 30 °C and 200 rpm for 3 h. Afterward, the fibers were filtered, and the concentration of dextran in the supernatant was measured and compared with the original concentration. The amount of inaccessible water (σ) was calculated according to the procedure comprehensively described in the literature.22 This procedure was conducted for several dextran polymers having different sizes to find a relationship between σ of fibers for dextran polymers and the hydrodynamic radius (Rh) of dextran polymers. 2.6. BET Analysis. Nitrogen adsorption/desorption isotherms were measured for unmodified and CPVA-modified fibers using a BELSORP-max, BEL Japan, Inc. (Osaka, Japan) instrument.35,36 At first, approximately 1 g (o.d.) of the samples was dried overnight. They were then pretreated for contamination removal at 70 °C and 107 Torr overnight. Afterward, the measurement was carried out using nitrogen, as a probe, at 77 K overnight. The isotherm data were recorded in a relative pressure of P/P0 in the range of 107 to 0.99999. The BET specific surface area was then calculated from the adsorption isotherms. Three repetitions were conducted in this experiment, and the average value for each sample was reported.
3. RESULTS AND DISCUSSION 3.1. Characteristics of Cationic PVAs. In this study, the cationic modification of PVA was carried out using the same PVA. To produce CPVAs with different charge densities, different numbers of quaternary ammonium groups should be associated with the PVA backbones, but the resulting CPVAs would have similar structures. The cationization of PVA and characteristics of CPVA was demonstrated in our previous work.3,26 CPVA with the charge densities of 0.2 ( 0.05 and 1.7 ( 0.08 mequiv/g, which corresponded to the degree of substitution of 0.9 ( 0.1 and 9.4 ( 0.3%, respectively, were selected for our analysis. The CPVA with the charge density of 1.7 mequiv/g had the highest charge density obtained for the cationic modification according to our previous work.26 The average hydrodynamic radius of the
ARTICLE
Figure 1. Adsorption isotherms of CPVA (with the MW = 124186 kDa) having various charge densities (mequiv/g) on cellulose fibers (conducted at 3% fiber consistency and 30 °C for 24 h).
low- and high-charged CPVAs were 11.7 ( 0.2 and 14.5 ( 0.3 nm, respectively. The MW and polydispersity of the lowcharged CPVA were 111188 kDa and 1.52, while those of the high-charged CPVA were 117201 kDa and 1.60, respectively. 3.2. Adsorption Isotherms of CPVAs. Generally, the structure, hydrodynamic radius, entropy increase, and the flexibility and charge density of polymers affect the adsorption of polymers on fibers.18 In previous studies, different PVAs possessed different adsorption characteristics on fibers.37 The adsorption of unmodified PVA was analyzed on the fibers in the current study, and no detectable adsorption was observed. This result implies that the adsorption of CPVA was mainly due to the electrostatic charge interaction. The adsorption characteristics of CPVAs on bleached sulfite cellulose fibers under different process conditions were discussed in our previous work.3 However, HYP was used instead of bleached sulfite fibers in this research. Because the characteristics of pulp fibers significantly affect the adsorption performance, and hence the diffusion performance, we initially evaluated the adsorption performance of CPVAs on HYP fibers. Figure 1 shows the adsorption isotherms of CPVAs on cellulose fibers. By increasing the amount of CPVAs in the fiber suspensions, the low-charged CPVA adsorbed more than the high-charged one. We previously reported that, if the adsorbing polymers did not have a narrow molecular weight distribution, the low and high MW portions of the polymers would have different adsorption isotherms, leading to an overall bimodal adsorption isotherm.4 Therefore, the adsorption of CPVA would have followed the bimodal adsorption model, as the CPVAs had relatively wide MW distributions (see section 3.1).4 A bimodal adsorption model was observed for polymers with various charge densities on cellulose fibers, on dissipative quartz-crystal wafers, and on latex.1,7,20,38 3.3. Adsorption Kinetics of CPVAs. Figure 2 shows the adsorption of CPVAs versus the time of adsorption. Although the adsorption of polymers on fibers is a quick process, polymers reconfigure and may migrate from the fiber surface to the pores, which takes a longer time. Results in Figure 2 show that the adsorption of CPVAs on fibers almost reached the saturation level after 200 min. One factor driving the adsorption of charged polymers on the fiber surface is the electrostatic attraction.9,10,19 The adsorption is mainly an ion-exchange process, i.e., cations inside the pores being replaced with cations of the polyelectrolyte. The released 13491
dx.doi.org/10.1021/la203364x |Langmuir 2011, 27, 13489–13496
Langmuir
Figure 2. Kinetic adsorption of CPVA having different charge densities (mequiv/g; 3% fiber consistency; 90 mg/g CPVA addition dosage).
cations increase the overall entropy and thus contribute to the driving force. When charged polymers adsorb on the fiber surface, a repulsion force is developed between the approaching polymers and preadsorbed polymers on a charged surface.9,39 The repulsion force arises from the entropy loss of dissociated counterions, which are pushed back into the surface when the polymers approach the surface covered partly by the charged polymers.37,39 The higher the charge density of CPVA, the higher such repulsion force is developed. This repulsion force decreases the overall driving force for adsorption (a lower adsorption rate) and decreases the total amount of CPVA adsorbed on the fiber surface. 3.4. Adsorption of PDADMAC on Unmodified and CPVAModified Fibers. Since PDADMAC is used for determining the surface charge density of unmodified and CPVA-modified fibers, the performance of PDADMAC on these fiber suspensions should be first evaluated. Figure 3 shows the concentration of PDADMAC that remained in the supernatants after treatment with unmodified and CPVA-modified fibers versus the time of the treatment. As can be seen, the PDADMAC concentration decreased significantly within 10 min of the experiment and then reached a plateau, regardless of the fiber modification. The concentration of PDADMAC was reduced more significantly for the unmodified fibers compared with the CPVA-modified fibers, because CPVA neutralized a part of the surface charge of fibers. These results are very interesting, because (1) the adsorption of PDADMAC was mainly on the fiber surface and macropores, since the concentrations of PDADMAC reduced to their final values in various suspensions in 10 min and no significant changes were observed afterward (no diffusion probably occurred during the first 10 min of experiment); and (2) as illustrated earlier, the adsorption of PDADMAC after treatment with CPVA-modified fibers was measured for determining the surface charge of fibers via the back-titration method. If the CPVAs of the CPVA-modified fibers had been displaced by PDADMAC during the back-titration, the concentration of PDADMAC would have been changed in the supernatant. Therefore, the total charges in the supernatants would have been different, because the charge densities of the low- and highcharged CPVA and PDADMAC were 0.2, 1.7, and 6.1 mequiv/g, respectively. Therefore, the titration of PDADMAC solutions should have shown different PDADMAC concentrations for different time intervals in Figure 3. However, Figure 3 shows
ARTICLE
Figure 3. Concentration of PDADMAC (mM) in the supernatants after treatning with 1 g of unmodified fibers and fibers modfied with 60 mg/g of 0.2 mequiv/g CPVA or with 10 mg/g of 1.7 mequiv/g CPVA (3% fiber consistency, 30 °C, and 150 rpm).
no significant changes over the time intervals evaluated, even for the HYP fiber modified with 10 mg of 1.7 mequiv/g CPVA. In fact, the trend of concentration changes was similar for all HYP fibers, regardless of the CPVA modification. These results underscore the possibilities of PDADMAC displacement with CPVA for the CPVA-modified fibers under the conditions investigated. Because no detectable diffusion and CPVA displacement occurred via PDADMAC titration, the PDADMAC could indeed be used for determining the surface charge of unmodified and CPVA-modified fibers. 3.5. Adsorption on the Outer Surface and Macropores of Fibers. As mentioned earlier, the polydispersity of the low- and high-charged CPVAs were 1.52 and 1.60, respectively, which would imply that the MW distribution of CPVA was relatively broad. Therefore, a part of CPVA with short chain lengths might have diffused to a larger extent than the other chain lengths.4 However, the CPVAs had very similar polydispersities. Since the purpose of this analysis was to compare the diffusion of CPVAs having different charge densities, but very similar polydispersities, into the fiber pores, the polydispersity was likely not an influencing factor in our studies. Figure 4 shows the surface charge density of fibers measured by the back-titration method versus the amount of CPVAs added to fibers. Evidently, the charge density of fibers was decreased from 26 to 22 μequiv/g by adding 80 mg/g of the lowcharged CPVA. The marginal change was due to the low charge of the low-charged CPVA (0.2 mequiv/g). The charge density of fibers was increased from 26 to +11 μequiv/g by adding 80 mg/g of the high-charged CPVA to fibers.The back-titration analysis showed that the surface charge density of fibers was 27 μequiv/g. The total charge density of the fibers used by a conductometric titration was 310 ( 10 μequiv/g. Therefore, about 9% of the total charges of fibers were located on the fiber surface, which is in agreement with the literature results.4042 The results in Figure 1 showed that, by adding 20 mg/g highcharged CPVA, 17 mg/g was adsorbed, which corresponded to a total charge adsorption of 28.9 μequiv/g (17 1.7 μequiv/g) on fibers. Therefore, the surface charge density of fibers will be 1.9 μequiv/g via adsorbing 17 mg/g high-charged CPVA (adding 20 mg/g high-charged CPVA). The experimental analysis in Figure 4 shows almost a zero charge density upon adding 20 mg/g high-charged CPVA, which reveals a marginal error 13492
dx.doi.org/10.1021/la203364x |Langmuir 2011, 27, 13489–13496
Langmuir
Figure 4. Fiber charge density versus the amount of CPVA added.
Figure 5. Inaccessible water of fibers, modified with various dosages of low-charged CPVA (0.2 mequiv/g), for dextran polymers as a function of dextran hydrodynamic radius.
compared with the theoretical estimate. Therefore, one can conclude that, among the 40 mg/g overall adsorption of the high-charged CPVA (Figure 2), approximately 17 mg/g was stoichiometrically adsorbed on the outer surface and macropores, and the remaining adsorbed in the micro- and nanopores of fibers, which is in agreement with the stoichiometrical adsorption of high-charged polymers on charged surfaces.14 The results in Figure 4 also show that the total charges adsorbed via the high-charged CPVA were 40 1.7 = 68 μequiv/g. Therefore, about 21% of the total charges of the fibers was accessible to the high-charged CPVA. Furthermore, as the charge density of fibers was marginally affected by the adsorption of the low-charged CPVA, the low-charged CPVA did not stoichimetrically adsorb on the fiber surface, even at the maximum adsorption level. It was reported in the literature that the lowcharged polymers do not always adsorb stoichiometrically on fibers, and a part of their charges is compensated for by the charges of counterions.14 3.6. Adsorption onto Micropores. There are always some pores in the structure of cellulose fibers that are not accessible to the dextran polymers. As the adsorption experiment was conducted in water, these pores contain water. The amount of water in these pores is called inaccessible water, σ (expressed as grams of water/(grams of fiber)), to the dextran polymers. Generally, σ in the SET method will increase as the radius of the probe molecule (dextran) increases, reaching a plateau for large molecules that
ARTICLE
Figure 6. Inaccessible water of fibers, modified with various dosages of high-charged CPVA (1.7 mequiv/g), for dextran polymers as a function of dextran hydrodynamic radius.
cannot penetrate into any of the pores.5 Figures 5 and 6 show the amount of inaccessible water for the unmodified fibers and fibers modified with various CPVAs as a function of dextran hydrodynamic radius. Clearly, the σ of water was increased by increasing the CPVA addition (mg/g) on fibers, regardless of the Rh of dextrans. For the dextrans with relatively large radius (Rh > 8 nm), σ reached a plateau, which indicated the fiber saturation point (FSP). The results in Figures 5 and 6 showed that the fibers seemed to be more swollen as the FPS of fibers was increased by adsorbing CPVA. It was described in the Introduction that the pores of fibers most probably have a hairy structure. The charged anionic groups associated with the hemicelluloses induce repulsion force between the hemicellulose hairy layers, which probably stretches these hairy structures. Therefore, the hairy layer is probably sufficiently rigid that it can impair the diffusion of dextran during solute exclusion. Thus, the SET analysis determines the size of pores excluding the thickness of this hairy structure. In other words, the dextran polymers can readily diffuse into the place that has no hairy barriers. If the thickness of this hairy layer for a pore is 15 nm for example, the dextran with the Rh of 8 nm (for example) can readily enter the pores with the original size of 38 nm, while the SET analysis will show 8 nm pore radius. The hairy structure plays an important role in the diffusion of dextran into fiber pores and hence in the SET analysis. We consider two pores with the same pore opening (excluding the hairy layers). If the wall of one pore has no or a limited hairy structure, the actual pore opening of this pore is larger than the pore having a considerable hairy wall, as this pore has less physical constrains for the diffusion of dextran in the SET analysis. If this is the situation for any pore with the actual size of 38 nm, the dextran with Rh = 14 or 27 nm could diffuse. Since this pore is accessible for these dextrans, it does not contribute to σ as FSP. Now, we consider the case of CPVA diffusion into the pores with partial hairy structure in the CPVA/fiber treatment system: the CPVA will form complexes with the hemicellulose hairy layers of pores upon diffusing and reduce the pore opening. In this scenario, the diffusion of the dextrans with Rh = 14 or 27 nm into the pore (with the opening of 38 nm, for example) of CPVA-modified fibers will be reduced in the SET analysis and σ for these dextrans will be increased (as observed for dextran with Rh = 14 and 27 nm in Figures 5 and 6). Another reason for the increase in the FSP (Figures 5 and 6) may be attributed to the enhancement in the hydrophilicity of the 13493
dx.doi.org/10.1021/la203364x |Langmuir 2011, 27, 13489–13496
Langmuir
ARTICLE
Table 1. Pore Radius of Fibers Estimated According to Equation 6 via Analyzing the Results Presented in Figures 5 and 6 20 mg/g of
60 mg/g of
20 mg/g of
60 mg/g of 1.7
fiber modification
Specification of unmodified
0.2 mequiv/g CPVA
0.2 mequiv/g CPVA
1.7 mequiv/g CPVA
mequiv/g CPVA
cylindrical pore radius, nm
14.4
12.6
11.7
13.1
12.8
slit-like pore size, nm
7.2
6.3
5.85
6.55
6.4
fiber surface by adsorbing CPVA, because CPVA is very hydrophilic.23 Some water molecules can also be trapped between CPVAs adsorbed on fibers and the fiber surface. These water molecules are taken into account as inaccessible water (σ) for dextran polymers in the SET analysis.24 For the dextrans with a small radius (Rh < 8 nm), the reason for the increase in σ is that, by the diffusion of CPVAs into the microfiber pores, the volume of the pores was reduced, which prevented the small dextrans from accessing the pores and resulted in an increase in σ for the dextrans. CPVAs might have also blocked the entrance of fiber pores upon adsorbing. However, since the amount of inaccessible water for small dextrans was less than that for large dextrans (Figures 5 and 6), it can be inferred that the small dextrans could diffuse into the pores more than the large ones did. Therefore, not all of the fiber pores were blocked by applying CPVAs.20 Also, by considering the Rh of CPVA (11.714.5 nm) and of various dextrans in Figures 5 and 6, it can be concluded that the CPVA might have also diffused into the pores that were smaller than its radius. 3.7. Pore Size Analysis via SET. In the SET analysis, if the probe molecule (dextran) cannot enter a pore, the polymer concentration in the pore is zero. On the other hand, even if a probe molecule enters the pore, its concentration in the pore will be less than its concentration in the bulk. This is due to the fact that the polymer concentration near the pore wall is less than that in the bulk on average, because the concentration of nonadsorbing molecules is zero at the surface.5 Thus, in a thin layer adjacent to the pore wall, i.e., depletion layer, the polymer concentration is lower than in the bulk concentration. The extension of depletion layer is typically the size of the polymer molecule in solution.5 This depletion layer causes the average polymer concentration in a pore to always be less than the concentration in the bulk. If we define K(a,r) to be the fraction of water in a pore of radius r accessible to a probe of radius a, then σðaÞ ¼ 1 FSP
Z ∞ 0
Kða, rÞ pðrÞ dr
ð4Þ
R where p(r) is the normalized pore size distribution (i.e., ∞ 0 p(r) dr = 1). To obtain the pore size distribution p(r) from the measurements of the amount of inaccessible water σ(a) obtained for various values of the probe size a, we need to invert eq 4. In this equation, K(a,r) is not precisely known, as it depends on the details of the depletion layers and the geometry of the pores. Thus, K is different for the slit-like and cylindrical pores. Depletion is often neglected, in which case K = 1 for a < r and K = 0 for a > r. Assuming cylindrical or slit-like pores and approximating the depletion layer with a step function (i.e., the polymer concentration changes from zero at the pore opening to bulk value at a distance a from the opening), we obtain K = (1 a/r)2 and K = 1 a/r for a < r for cylindrical and slit-like pores, respectively. Assuming that the pore size distribution is uniform, i.e., the eq 4 is reduced to σ=FSP ¼ 1 Kða, rÞ
ð5Þ
For small probes (a , r), eq 4 reduces further to σ=FSP ¼ 2a=r
or
σ=FSP ¼ a=r
ð6Þ
for cylindrical and slit-like pores, respectively. Thus, the initial slope of inaccessible water (σ) in the SET analysis can be related to the pore size, r, of fibers. The sizes of assumed cylindrical and slit-like pores were determined using the results presented in Figures 5 and 6, and listed in Table 1. Clearly, the pore radius of unmodified fibers was 14.4 nm, which is close to the pore size reported by others.22 For the cylindrical pores, the pore radius was reduced to 11.7 and 12.8 nm by modifying the fibers with 60 mg/g of the low- and high-charged CPVA, respectively. For the slit-like pores, the pore size was reduced from 7.2 to 5.85 nm and 6.4 nm for the fibers modified with 60 mg/g low- and high-charged CPVA, respectively. This analysis revealed that the pore size of fibers was reduced more significantly by applying the low-charged CPVA than high-charged one, implying a higher diffusion of lowcharged CPVA into the fiber pores. 3.8. Diffusion Analysis. Since no mechanical treatment, e.g., beating, was carried out on fibers during the experiments, the amount of CPVAs adsorbed inside the fiber pores was directly influenced by the diffusion of CPVAs into the fiber pores. Generally, the results in Figures 26 showed a higher diffusion of the low-charged CPVA than the high-charged CPVA into the fiber pores. Therefore, the lower adsorption rate of the highcharged CPVA than the low-charged CPVA in Figure 2 is ascribed to its lower diffusion rate into the fiber pores. Regardless of the hairy structure, the pores affected by the CPVA diffusion can be divided into two parts: (1) the pores that are sufficiently large that CPVA does not require a reconfiguration upon diffusing (micropores); (2) the pores that are smaller than the size of CPVA. The diffusion principles of CPVA into these pores are different as discussed in the following: 3.8.1. Diffusion into Large (Micro)Pores. The mechanism for the diffusion of CPVA into the micropores of fibers is elucidated as follows: when the CPVA monolayer has not been developed on the fiber surface (beginning of the adsorption), the approaching CPVA will face with the uncoated fiber surface (or uncoated hairy structure of pore opening). However, when a monolayer is developed, electrostatics repulsion force is developed between the approaching CPVA and the adsorbed CPVA close to the pore opening. As described earlier, the repulsion force is more substantial for the high-charged CPVA than for the low-charged CPVA. This higher repulsion force reduces the concentration of CPVA at the pore opening. We consider the pore depicted in Figure 7 as a fiber micropore representative. Due to the repulsion force, the concentration of polymers at the pore opening, Co, will be lower than that in bulk, Cb. In this case, the higher the repulsion force, the lower the Co. Consequently, if the charge density of polymer is significant, the repulsion force developed between the already adsorbed polymers and the approaching polymers will be significant, which reduces the concentration of polymers at the pore opening and eventually reduces the 13494
dx.doi.org/10.1021/la203364x |Langmuir 2011, 27, 13489–13496
Langmuir
ARTICLE
Figure 7. Approaching of CPVA to the fiber micropore for diffusion when a layer of adsorbed CPVA has been already developed on fibers. (Dashed line shows the electrostatic repulsion force created between the approaching CPVA and the adsorbed CPVA.)
Table 2. BET Surface Area of Unmodified Fibers and Modified Fibers with Different Dosages of the Low- and HighCharged CPVA Specification of
BET surface
Specification of
BET surface
fiber modification
area, m2/g
fiber modification
area, m2/g
unmodified
0.56 ( 0.02
20 mg/g of
0.58 ( 0.03
which was in agreement with the findings of others,4345 and implied that the pores were collapsed during the pretreatment of the BET analysis. Thus, the BET analysis represents the outer surface area. Interestingly, the BET surface area was increased to 1.17 and 2.38 m2/g upon adding 20 and 60 mg/g of the lowcharged CPVA to fibers. The BrunauerEmmettTeller (BET) surface area increased to 0.58 and 1.21 m2/g via adding 20 and 60 mg/g of the high-charged CPVA, respectively. Therefore, the BET surface area increased substantially via modifying the fibers with CPVA. There are two possibilities for this increase: (1) the CPVA/hemicelluloses complexes on the fiber surface increased the roughness of fiber surface after drying; (2) the complexes formed via the interaction of the diffused CPVA and hemicelluloses in the fiber pores prevented a complete collapse of fiber pores in drying. As confirmed in the above analyses, CPVA diffused into the pores of fibers and as a consequence, the pores were somehow filled with CPVA. These pores could not be completely closed upon drying. This is because the filled pores would require more considerable fiber wall deformation than unfilled pores do, which is unlikely to occur upon drying. The incomplete closure of the pores for CPVA-modified fibers induced a larger surface area in the BET analysis (Table 2). Since the diffusion of the low-charged CPVA was higher than that of the high-charged CPVA into the fiber pores, the BET surface area increased more substantially for the fibers modified with the low-charged CPVA.
1.7 mequiv/g CPVA 20 mg/g of
1.17 ( 0.04
0.2 mequiv/g CPVA 60 mg/g of
60 mg/g of
1.21 ( 0.04
1.7 mequiv/g CPVA 2.38 ( 0.06
0.2 mequiv/g CPVA
diffusion of polymers into the fiber pores. Therefore, the concentration of CPVA close to the pore (Co) would be lower for the high-charged CPVA than for the low-charged CPVA,20,34 which resulted in a lower diffusion and a lower pore size reduction (Table 1). Consequently, the hairy structure of large pores has limited influence on the diffusion of CPVAs, but the charge density of the CPVA has a great impact on the diffusion. 3.8.2. Diffusion into Small (Nano)Pores. As described earlier, CPVA may also diffuse into the pores that are smaller than its own size or have complex hairy structures. In this case, the CPVA should have been uncoiled and reconfigured to facilitate the diffusion. In addition to the repulsion force and CPVA charge effect described in section 3.8.1, the polymer flexibility may also play an important role in the diffusion of CPVA into the nanopores. The persistence length, which is an indicator of polymer flexibility, is affected by the electrostatic repulsion developed between the charged groups on the polymer backbones.9,10 Since the high-charged CPVA contains more charged group on its backbone than the low-charged CPVA does, the persistence length of the high-charged CPVA is larger than that of low-charged CPVA. Therefore, the low-charged CPVA is more flexible than the high-charged CPVA. Horvath et al. reported that the diffusion of polymers into the fiber pores was lowered as a result of the decrease in the flexibility of polymers.19 Therefore, the less diffusion of the high-charged CPVA than the low-charged CPVA into the nanopores was attributed to its lower flexibility. 3.9. Surface Area Modification. The BET surface areas of unmodified and CPVA-modified fibers are listed in Table 2. As can be seen, the surface area of unmodified fibers was 0.56 m2/g,
4. CONCLUSIONS The adsorption of the low-charged CPVA was substantially higher than that of the high-charged CPVA on fibers. The surface charge density analysis confirmed that approximately 17 mg/g of the high-charged CPVA adsorbed on the outer surface and macropores of fibers, and the remaining (23 mg/g) diffused into the pores. However, the high-charged CPVA could not access all the charges of fibers. The SET analysis confirmed that the pore size of fibers was more significantly reduced by applying the lowcharged CPVA than by the high-charged one. The adsorption, charge density, and SET analyses quantitatively confirmed that the diffusion of the low-charged CPVA into the fiber pores was indeed greater than that of the high-charged one. Besides the entropy increase, the influencing factor for the diffusion of CPVA into the large (micro) pores was claimed to be the repulsion force developed between the already adsorbed CPVA close to the fiber pore opening and the approaching CPVA from the bulk. The influencing factors for the diffusion of CPVA into the small (nano) pores were the repulsion force and the flexibility of CPVA segments as the CPVA had to reconfigure to facilitate the diffusion. Also, the BET surface area of fibers was increased upon the CPVA adsorption, and the increase was more significant for the low-charged CPVA. It was proposed that the diffused CPVA prevented the complete fiber pore collapse during drying, which eventually increased the surface area of fibers. ’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected]. Tel.: 807-343-8697. Fax: 807-346-8928.
’ ACKNOWLEDGMENT NSERC Canada and Atlantic Innovation Fund (AIF) are gratefully acknowledged for funding this research. 13495
dx.doi.org/10.1021/la203364x |Langmuir 2011, 27, 13489–13496
Langmuir
’ REFERENCES
€ (1) Tanaka, H.; Odberg, L.; Wagberg, L.; Lindstr€om, T. J. Colloid Interface Sci. 1990, 34, 219. (2) Van de Ven, T. G. M. J. Colloid Interface Sci. 1994, 48, 121. (3) Fatehi, P.; Xiao, H. Colloids Surf., A 2008, 327 (13), 127. (4) Alince, B.; Vanerek, A.; van de Ven, T. G. M. Ber. Bunsen-Ges. Phys. Chem. 1996, 100 (6), 954. (5) Alince, B.; van de Ven, T. G. M. In Fundamentals of Papermaking Materials Edition, Transactions of the 11th Fundamental Research Symposium; Baker, C. F., Ed.; Pira International: Surrey, U.K., 1997; Vol. 2, p 771. (6) Goddard, E. D.; Faucher, J. A.; Scott, R. J.; Turney, M. E. J. Soc. Cosmet. Chem. 1975, 26, 539. (7) Ondaral, S.; Wagberg, L.; Enarsson, L. E. J. Colloid Interface Sci. 2006, 30, 132. (8) Shirazi, M.; van de Ven, T. G. M.; Garbier, G. Langmuir 2003, 19, 10835. (9) Horvath, A. T.; Horvath, A. E.; Lindstrom, T.; Wagberg, L. Langmuir 2008, 24, 6585. (10) Horvath, A. T.; Horvath, A. E.; Lindstrom, T.; Wagberg, L. Langmuir 2008, 24, 7857. (11) Netz, R. R.; Joanny, J.-F. Macromolecules 1999, 32, 9013. (12) Svensson, A. V.; Huang, L.; Johnson, E. S.; Nylander, T.; Piculell, L. Appl. Mater. Interfaces 2009, 1 (11), 2431. (13) Dobrynin, A. V.; Ekaterina, B. Z.; Rubinstein, M. Macromolecules 2001, 34, 627. (14) Horvath, A. E.; Lindstrom, T.; Laine, J. Langmuir 2006, 22, 824. (15) Zhang, J.; Pelton, R. Langmuir 1999, 15, 5662. (16) Chakraborty, A.; Sain, M.; Kortschot, M. Holzforschung 2005, 59, 102. (17) Gurak, E.; Dupont-Gillain, C.; Booth, J.; Roberts, C. J.; Rouxhet, P. G. Langmuir 2005, 21, 10684. (18) Guzman, E.; Ritacco, H.; Rubio, J. E. F.; Rubio, R. G.; Ortega, F. Soft Matter 2009, 5, 2130. (19) Horvath, A. T.; Horvath, A. E.; Lindstrom, T.; Wagberg, L. Langmuir 2008, 24, 10797. (20) Wagberg, L.; H€agglund, R. Langmuir 2001, 17, 1096. (21) Forsstr€om, J.; Andreasson, B.; Wagberg, L. Nord. Pulp Pap. Res. J. 2005, 20 (2), 176. (22) Stone, J. E.; Scallan, A. M. Cellulose Chem. Technol. 1968, 2, 321. (23) Topgaard, D.; Soderman, O. Cellulose 2002, 9, 139. (24) Fatehi, P.; MacMillan, B.; Ziaee, Z.; Xiao, H. Colloids Surf., A 2009, 348 (13), 59. (25) Katz, S.; Beatson, R. P.; Scallan, A. M. Sven. Papperstidn. 1984, 87 (6), 48. (26) Fatehi, P.; Xiao, H. Nord. Pulp Pap. Res. J. 2008, 23 (3), 285. (27) Fatehi, P.; Xiao, H. J. Pulp Pap. Sci. 2009, 35 (1), 1. (28) Fox, T. G.; Flory, P. J. J. Am. Chem. Soc. 1951, 73, 1904. (29) Nagy, D. J. J. Liq. Chromatogr. 1993, 16 (14), 3041. (30) Coelho, J. F. J.; Goncalves, P. M. F. O.; Miranda, D. Eur. Polym. J. 2006, 42 (4), 751. (31) Williams, P. C.; Kuzina, F. D.; Hlynka, I. Cereal Chem. 1970, 47, 411. (32) G€ardlund, L.; Wagberg, L.; Gernandt, R. Colloids Surf., A 2008, 218, 137. (33) Meiczinger, M.; Dencs, J.; Marton, G.; Dencs, B. Ind. Eng. Chem. Res. 2005, 44, 9581. (34) Wagberg, L.; Odberg, L.; Glad-Nordmark, G. Nord. Pulp Pap. Res. J. 1987, 2, 71. (35) Buchhammer, H. M.; Mende, M.; Oelmann, M. Colloids Surf., A 2003, 218, 151. (36) Dabros, T.; van de Ven, T. G. M. Colloids Surf., A 1993, 75, 95. (37) Ishimaru, Y.; Lindstrom, T. J. Appl. Polym. Sci. 1984, 29, 1675. (38) Guzmana, E.; Ortegaa, F.; Baghdadlib, N.; Luengob, G. S.; Rubioa, R. G. Colloids Surf., A 2011, 375, 209. (39) Amsden, B.; Grotheer, K.; Angl, D. Macromolecules 2002, 35, 3179.
ARTICLE
(40) Drnvsek, T.; Perdih, A.; Perdih, M. J. Wood Sci. 2005, 51 (5), 507. (41) Horvath, A.; Lindstrom, T. Nord. Pulp Pap. Res. J. 2007, 22 (1), 87. (42) Zhang, Y.; Sjogren, B.; Engstrand, P.; Htun, M. J. Wood Chem. Technol. 1994, 14 (1), 83. (43) Jacob, P. N.; Berg, J. C. Langmuir 1994, 10, 3086. (44) Ouajai, S.; Shanks, R. A. Cellulose 2006, 13, 31. (45) Bismarck, A.; Aranberri-Askargorta, I.; Springer, J.; Lampke, T.; Wielage, B.; Stamboulis, A.; Shenderovich, I.; Limbach, H. H. Polym. Compos. 2002, 23 (5), 872.
13496
dx.doi.org/10.1021/la203364x |Langmuir 2011, 27, 13489–13496
