XPS imaging of surface diffusion of alkylketene dimer on paper surfaces

XPS imaging of surface diffusion of alkylketene dimer on paper surfaces

Colloids and Surfaces A: Physicochem. Eng. Aspects 219 (2003) 35 /43 www.elsevier.com/locate/colsurfa XPS imaging of surface diffusion of alkylketen...

626KB Sizes 0 Downloads 11 Views

Colloids and Surfaces A: Physicochem. Eng. Aspects 219 (2003) 35 /43 www.elsevier.com/locate/colsurfa

XPS imaging of surface diffusion of alkylketene dimer on paper surfaces ¨ dberg b A.V. Shchukarev a,*, R. Mattsson b, L. O a

b

Department of Chemistry, Inorganic Chemistry, Umea˚ University, S-901 87 Umea, Sweden Department of Chemical Technology, Lulea˚ University of Technology, S-971 87 Lulea, Sweden Received 5 November 2002; accepted 19 December 2002

Abstract The alkylketene dimer (AKD) wax dispersion is introduced in the wet-end of a paper machine, and during drying it spreads and anchors to the fiber surface. Direct observation of the surface diffusion of the AKD wax at paper surfaces was obtained using X-ray photoelectron spectroscopy imaging technique. An AKD diffusion of approximately 400 mm was observed at a paper surface that was stored for 3 h at 80 8C. This gives a diffusion coefficient of approximately 10 11 m2 s 1. Storage for 1 week at room temperature also showed a remarkable spreading of AKD. # 2003 Elsevier Science B.V. All rights reserved. Keywords: X-ray photoelectron spectroscopy; X-ray photoelectron spectroscopy imaging; Alkylketene dimer; Surface diffusion; Paper

1. Introduction Dispersions of alkylketene dimer (AKD) are widely used in the paper industry, along with rosin size and alkenyl succinic anhydride to produce papers with a certain degree of hydrophobicity. AKD (Fig. 1) is a waxy substance with a melting point varying from room temperature up to 65 8C depending on the length and saturation of the substituted alkyl chains, R? and Rƒ. Commonly occurring is a combination of C16 and C18 chains. * Corresponding author. Tel.: /46-90-786-6328; fax: /4690-786-9195. E-mail address: [email protected] (A.V. Shchukarev).

A usual way to introduce AKD to the paper is by internal sizing, where a water dispersion of the AKD wax is introduced into the fiber furnish prior to the headbox of the paper machine. The AKD dispersions are most commonly cationic and consist of AKD particles in the size range of 0.5 /2 mm in diameter. These dispersions are electrostatically and most often also sterically stabilized with cationic polyelectrolytes. The solids content in dispersions that are commercially available is in the range of 10 / 30%. A normal dosage level in a paper machine is between 0.1 and 0.2% based on dry pulp. The mechanism of AKD sizing is well described in the Refs. [1 /4]. The sizing mechanism of AKD is often divided into the following sub processes:

0927-7757/03/$ - see front matter # 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0927-7757(03)00009-8

36

A.V. Shchukarev et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 219 (2003) 35 /43

Fig. 1. Structure of AKD and its believed interaction with cellulose.

retention , where part of the cationic AKD particles are deposited on the negatively charged fiber surface, preferably uniformly distributed effective spreading of AKD on the fiber surface in the dryer section of the paper machine chemical reaction , where the AKD binds via a b-ketoester to the cellulose. If the spreading of AKD is effective, it might be possible to use larger AKD particles or agglomerates of AKD particles for sizing. Larger agglomerates would be easier to retain in the paper web during papermaking than micron-sized particles which would most probably promote the sizing efficiency [5]. The spreading of AKD has only been studied in detail on model surfaces. There have been different proposals about the spreading mechanism of AKD. It was suggested by Hodgson [6] that AKD spreads via surface diffusion to a monomolecular layer and Stro¨m et al. [7] proposed that AKD also could spread to a thin multimolecular film. On the contrary, Garnier and Yu [8] concluded that AKD could not completely wet model surfaces of glass or cellulose. Shen et al. [9]

suggested that AKD migration by surface diffusion is limited and that migration over a few micrometers occurs via the vapor phase and redeposition. Seppa¨nen et al. [10] suggested that AKD spreads via surface diffusion with an autophobic precursor and they estimated the diffusion coefficient to 1011 m2 s 1 when studying the migration of an AKD wax with a melting point of 51 8C on silica model surfaces at a temperature of 50 8C. The influence of extractives on AKD migration in paper has recently been studied by Mattsson et al. [11]. Direct observation of sizing reagents at real paper surfaces, by using modern surface science techniques like TOF-SIMS and X-ray photoelectron spectroscopy (XPS), is limited to the works of Brinen et al. [12 /14]. They used SIMS and SIMS imaging to detect the presence and also the distribution of inorganic particles as well as sizing and de-sizing agents on paper surfaces. Papers presenting XPS imaging applications for studying the sizing of paper has not been found. The objectives of this study are to investigate the surface spreading of AKD wax on paper surfaces and also the spreading of an unflocculated AKD dispersion and a dispersion flocculated with car-

A.V. Shchukarev et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 219 (2003) 35 /43

boxymethylcellulose by using XPS imaging technique.

2. Experimental 2.1. Materials The paper used in this study was a Munktell Analytical Filter Papers Quality OOH, 80 g m 2 from Munktell Filter AB. The AKD wax used had a melting point of 51 8C. The structure of AKD and the believed reaction with cellulose are shown at Fig. 1. The AKD dispersion used in the study KeyDime E (KDE, EKA Chemicals AB, Sweden) consisted of the AKD emulsified in water and electrostatically stabilized with polyamideamine. The average volume weighted particle size was 0.8 mm in diameter. To study the spreading of aggregates of AKD particles, the AKD dispersion was flocculated with carboxymethylcellulose [15]. The average volume weighted size of the AKD agglomerates was 20 mm. Both the unflocculated and the flocculated AKD dispersions were diluted in deionized water to 0.1% w/w and applied to the paper by immersing the paper into the solution. Note that the difference in size of AKD particles may cause difference in the amount of wax deposited to the paper. Samples for studying the migration of the AKD wax were prepared by exposing one edge of the paper sample (:/1/15 mm2) to a solution of 1% w/w AKD in heptane (JT Baker, The Netherlands, min 88.0%). The initial AKD coverage, about 2 mm2, was confirmed with XPS before they were stored at different conditions. 2.2. Measurements All XPS experiments were performed with a Axis Ultra electron spectrometer using a monochromated Al /Ka source operated at 225 W (spectroscopy mode) or Mg/Ka source operated at 75 W (imaging mode). To compensate the surface charging a low energy electron gun was used. Processing of the spectra and images were KRATOS

37

accomplished with the KRATOS software. The binding energy (BE) scale was referenced to the C 1s line of aliphatic carbon set at 285.0 eV. Wide spectra (pass energy 160 eV) and spectra of individual photoelectron lines C 1s and O 1s (pass energy 20 eV) were acquired. In the case of XPS imaging pass energy of 40 eV and low magnification mode were used. Three images were acquired for each point of the samples: aliphatic carbon at 285.0 eV, cellulose C /OH group at 286.5 eV and background signal at 280.0 eV. The resulting chemical mapping were created by subtraction of the ‘background’ image from the images of the functional groups. It allows to reduce the influence of roughness on the images. Several overlapping areas of the same papers were imaged with a shift of approximately 200/300 mm each. Topographical heterogeneity of the paper surface makes it possible to reconstruct the whole chemical map with a size of few mm2 starting from the corner of the paper as the initial point. 2.3. X-ray degradation of paper samples It is well known that organic materials can degrade under X-rays. For wood samples changes in C 1s spectrum were detected after 20 min of irradiation using monochromated Al/Ka source operated at 225 W [16]. Reducing the power of Xray source allows to increase the analysis time necessary for XPS imaging. Control C 1s spectra, taken from the same points of the samples before and after the imaging, demonstrated no detectable changes for the AKD covered parts of the paper and not more than a 10% intensity increase of the aliphatic carbon component for ‘pure’ paper in 1 h.

3. Results and discussion 3.1. Storage effect in XPS spectra To study the spreading of AKD in the samples immersed in unflocculated and flocculated AKD dispersions, the carbon 1s spectra were recorded the first day and after 3 months of storage exposed

38

A.V. Shchukarev et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 219 (2003) 35 /43

to air and at room temperature. All XPS spectra are given in Fig. 2. A reference sample (without any treatment) was analyzed to confirm that neither degradation of cellulose nor contamination of the sample occurred.

The untreated paper (Fig. 2(a) and (d)) exhibits 3 main components in C 1s spectrum corresponding to aliphatic carbon contamination (285.0 eV), C /OH group (286.5 eV) and O /C /O bond (287.9 eV) of cellulose. The intensity of aliphatic con-

Fig. 2. C 1s spectra of paper samples: (a) initial untreated; (b) initial treated with unflocculated AKD solution; (c) initial treated with flocculated AKD solution; (d) untreated after 3 month of storage; (e) treated with unflocculated AKD solution after 3 month of storage; (f) treated with flocculated AKD solution after 3 month of storage.

A.V. Shchukarev et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 219 (2003) 35 /43

tamination is very low and does not change during the storage. Introducing the AKD molecules with a long aliphatic tails into the paper surface causes significant changes in C 1s spectra. Intensity of the hydrocarbon component at 285.0 eV increases for both treated papers but differently. In the case of unflocculated AKD dispersion, this intensity is much higher than for the flocculated one (Fig. 2(b) and (c)). Taking into account that the depth of analysis in XPS is about 10 nm and assuming that AKD particles applied to the paper maintain their shape during the first day, we can conclude that the initial surface coverage induced by the /0.8 mm wax dispersion is several times higher than for the /20 mm one. We can also conclude that a significant part of the paper surface remains uncovered in both cases. Note, that C 1s spectra of AKD treated papers contain an additional low intensity component with BE at 289.2 eV corre-

39

sponding to carbon in the /O /C /O functional group of AKD and possible AKD reaction products (Fig. 1). After 3 months of storage (Fig. 2 (f)) the relative intensity of the C /(C, H) component observed in the C 1s spectrum for the papers immersed in the flocculated AKD has increased significantly, an increase in the intensity of COOH is observed as well. The paper immersed in unflocculated AKD dispersion ( /0.8 mm particles) shows a negligible decrease (Fig. 2(b) and (e)). This observation indicates a process where the AKD wax has diffused from the large flocs ( /20 mm) over the paper surface. From the relative intensities of C /(C, H) and C /OH components we can conclude that the surface coverage after 3 months is approximately 2 times higher for the flocculated dispersions than the surface coverage achieved for the unflocculated AKD dispersion.

Fig. 3. XPS images of initial sample immersed into AKD solution in heptane: (a) C /(C, H); (b) C /OH.

40

A.V. Shchukarev et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 219 (2003) 35 /43

Fig. 4. Overlaying of C /(C, H) and C /OH images and corresponding intensities profile through the AKD/pure paper interface for ‘as immersed’ paper sample.

Even if 3 months is an extremely long diffusion time this observation agrees with the observation [5] that flocculated dispersions can be efficient for sizing. 3.2. XPS imaging of AKD surface diffusion The chemical ‘marks’ of cellulose (C /OH) and AKD (C /(C, H)) are well separated in BE and have dominant intensities in the corresponding C 1s spectra [17]. This makes XPS imaging (or chemical mapping) a very effective tool to visualize the wax behavior on the paper surface. Two paper

samples were studied. The samples were immersed in 1.0% AKD-heptane solution and the AKD coverage was confirmed using XPS imaging. One of the samples was then placed in a heating cabinet for 3 h at 80 8C, and during this time the sample was exposed to air. The other sample was kept exposed to air during 1 week at room temperature. The images in C /(C, H) (285.0 eV, AKD) and C /OH (286.5 eV, paper) obtained for one of the immersed papers are given at Fig. 3. Because it was impossible to place the sample in one plane out of the sample holder, the images are not uniform within the frames. To follow the wax behavior in one direction, the spectrometer analyzer was focussed at the paper edge for each image. In that case the maximum intensity of the image signal is achieved near the edge. Starting from the corner of the paper and by following the paper edge, we were able to observe three regions at the paper surface: an AKD covered part where C /(C, H) intensity is prevailing; an uncovered (‘pure’) paper corresponding to the predominance of C / OH intensity; a thin ‘interface’ in between. Overlaying the images taken from the last region (Fig. 4) allows to determine the interface as a field where C /(C, H) and C /OH intensities are equal. Doing this, the width of the interface is estimated to 200 mm. Noise in the profile is due to the roughness (morphology) of the paper surface. Because of the surface roughness it is not possible to perform simple mathematical fitting of the intensity curves. The chemical mapping was repeated after storage to observe changes in the AKD distribution at the paper surfaces. The heated sample shows remarkable changes in the XPS images (Fig. 5, only C /(C, H) images are shown). An increase of the C /(C, H) (AKD) component is clearly detected as far as 400 mm from the initial interface. This indicates a spreading of 400 mm in 3 h at 80 8C which gives a diffusion coefficient of approximately 1011 m2 s 1. This value agrees with earlier studies [10,11]. The figure is a rough estimate, considering that the topography of the paper surface could imply a longer distance of spreading. The C /OH intensity (the images are not shown) also increases at the surface covered by AKD. This can be explained by at least two effects: an

A.V. Shchukarev et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 219 (2003) 35 /43

Fig. 5. XPS images in C/(C, H) (AKD): (a) initial sample; (b) the same sample after 3 h at 80 8C.

Fig. 6. XPS images in C /(C, H) (AKD): (a) initial sample; (b) the same sample after 1 week of storage at room temperature.

41

42

A.V. Shchukarev et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 219 (2003) 35 /43

evaporation of the wax during heating and a diffusion of AKD to the uncovered part of the paper resulting in a decrease of thickness of the AKD layer. The effect of diffusion was proved by intensity profile of the C 1s components through the overlaying images assessed near the same point where the initial AKD/uncovered paper interface was found. The region of equal C /(C, H) and C /OH intensities is much more wide compared with the initial profile. In addition, the equal intensities of both C 1s components at the interface allow us to estimate the thickness of the AKD layer. The intensity ratio method used is well known in XPS [18]. The absolute intensities of C /(C, H) and C /OH components were measured with two samples: pure AKD wax and clean initial paper, respectively. A corresponding C 1s photoelectron attenuation length was calculated using NIST database [19]. The determined thickness of AKD layer at the interface is 1.6 /2.0 nm. This value is very close to the length of the wax molecule (C18 chain). This strongly indicates that AKD spreads as a monolayer. Similar results were found for the paper stored at ambient conditions during 1 week. The storage leads to an significant AKD spreading from the initial AKD/uncovered paper interface (Fig. 6, only C /(C, H) images are shown). The migration also was proved by the corresponding intensity profile over the interface region. The increase for the C /OH intensity of the 1 week sample was observed at the initially AKD covered paper area. The increase was not so prominent as in the case of the heated sample. It seems that evaporation of the wax occurs under the elevated temperature and should be taken into account. On the other hand, all XPS experiments were performed in ultra-high vacuum (106 /107 Pa) and a decrease in C /(C, H) (AKD) intensity was not detected measured at the same points of the samples. Moreover, we were also able to observe the melting process of large AKD particle ( /500 mm) at a paper surface inside the vacuum chamber during heating of the manipulator up to 130 8C. These experiments do thus not support the gas phase transfer mechanism of AKD proposed by Shen et al. [9].

4. Conclusions The migration of AKD at paper surfaces during storage was confirmed for papers treated with flocculated AKD dispersion. The XPS imaging technique allowed to visualize the AKD spreading directly. The migration of AKD was detected not only at temperature above the melting point of the wax but also during storage in room temperature. Moreover, our data support the AKD spreading as a monomolecular layer and do not confirm the gas transfer mechanism. Additional XPS experiments and mathematical processing of the intensities profiles are planned in order to calculate the diffusion coefficient more exactly.

Acknowledgements AssiDoma¨n AB and Eka Chemicals AB are acknowledged for their financial support. The paper was presented at International Symposium on Practical Surface Analysis PSA-01, November 2001, Nara, Japan.

References ¨ dberg, T. Lindstro¨m, B. Liedberg, J. Gustavsson, [1] L. .O Tappi J. 70 (4) (1987) 135. [2] T. Lindstro¨m, Fundamentals of papermaking, in: Baker, Punton (Eds.), Transactions of the Nineth Fundamental Research Symposium, vol. 1, Cambridge, UK, 1989, p. 311. [3] K.J. Bottorff, M.J. Sullivan, Nord. Pulp Pap. Res. J. 8 (1) (1993) 86. [4] J.C. Roberts, Fundamentals of Papermaking Materials, in: Baker (Ed.), Transactions of the 11th Fundamental Research Symposium, vol. 1, Cambridge, UK, 1997, p. 209. ¨ dberg, Nord. Pulp Pap. Res. J. [5] R. Mattsson, J. Sterte, L. .O 17 (3) (2002) 240. [6] K.T. Hodgson, Apita J. 47 (5) (1994) 402. [7] G. Stro¨m, G. Carlsson, M. Kiear, Wochenbl. Papierfabr. 120 (15) (1992) 606. [8] G. Garnier, L. Yu, J. Pulp Pap. Sci. 25 (7) (1999) 235. [9] W. Shen, N. Brack, H. Ly, I.H. Parker, P.J. Pigram, J. Liesegang, Colloids Surf. A: Physicochem. Eng. Aspects 176 (2001) 129. [10] R. Seppa¨nen, F. Tiberg, M.-P. Valigant, Nord. Pulp Pap. Res. J. 15 (5) (2001) 452.

A.V. Shchukarev et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 219 (2003) 35 /43 ¨ dberg, J. Pulp [11] R. Mattsson, D. Lindstro¨m, J. Sterte, L. O Pap. Sci., accepted for publication. [12] J.S. Brinen, R. Proverb, Nord. Pulp Pap. Res. J. 6 (1991) 177. [13] J.S. Brinen, R.J. Kulick, Int. J. Mass Spectr. Ion Proc. 143 (1995) 177. [14] J.S. Brinen, Nord. Pulp Pap. Res. J. 8 (1993) 123. ¨ dberg, The science of [15] R. Mattsson, J. Sterte, L. O papermaking. Transactions of the 12th Fundamental Research Symposium (Oxford), vol. 1, 2001, p. 393.

43

[16] A.V. Shchukarev, B. Sundberg, E. Mellerowicz, P. Persson, Surf. Interface Anal. 34 (2002) 284. [17] W. Shen, Y. Filonanko, Y. Truong, I.H. Parker, N. Brack, P.J. Pigram, J. Liesegang, Colloids Surf. A: Physicochem. Eng. Aspects 173 (2000) 117. [18] D.F. Mitchell, K.B. Clark, J.A. Bardwell, W.N. Lennard, G.R. Massoumi, I.V. Mitchell, Surf. Interface Anal. 21 (1994) 44. [19] NIST Standard Reference Database 82. NIST Electron Effective-Attenuation-Length database, version 1.0, 2001.