Foam forming leads to sheet structures with exceptional volume of large pores. The link between fibre network structure and foam properties is investigated by comparing pore structure with measured bubble-size distribution. In foams produced by mechanical mixing, higher rotor speed leads to smaller average bubble size, whereas the effects coming from air content and surfactant are less pronounced nor systematic. A significant drop in the average bubble size is seen when mixing fibres to foam. In sheets made with foam forming, there are more large pores compared to the water formed sheets. The size of these pores is affected by the sizes of the bubbles in the foam. Overall, pore size distribution is more strongly affected by the fibre type than by small changes in bubble size distribution.
Keywords: foam, bubble, surfactant, fibre network, pore.
ADDRESSES OF THE AUTHORS: Ahmad Al-Qararah ([email protected]), Tuomo Hjelt ([email protected]), Karita Kinnunen ([email protected]), Nikolai Beletski ([email protected]), Jukka Ketoja ([email protected]): VTT Technical Research Centre of Finland, P.O. Box 1000, FI-02044 VTT, Finland.
Corresponding author: Ahmad Al-Qararah.
Introduction
Pore size distribution is a key paper-based property in many processes such as drying, coating and printing, etc. With normal water forming operations, it is very difficult to change the characteristic shape of the distribution. The distribution usually has log-normal form, which means that the amount of large pores is relatively small (Niskanen 2008; Bliesner 1964).
On the other hand, quite different pore size distribution can be obtained by using foam as the carrier phase instead of water (Hjelt et al. 2011). In foam forming (Punton 1975a; Punton 1975b; Radvan, Gatward 1972; Skelton 1987; Smith 1974; Smith, Punton 1975; Wiggins Teap 1973), the stock is mixed with foam, and this mixture is transferred through the headbox to the wire, where foam is removed using a vacuum. This kind of forming leads to many positive effects. The fibres inside the foam are locked between the foam bubbles, and thus do not flock during transportation (Punton 1975a). This provides better formation and allows the usage of higher consistency stock. Also, the forces affecting fibres during dewatering are smaller compared to water forming. When combining foam forming with air drying, it is possible to make very bulky structures. The foam forming process and the resulting properties have been discussed by (Radvan, Gatward 1972).
Foam is a dispersal system consisting of gas bubbles separated by liquid (Weaire, Hutzler 1999). The relative proportion of liquid and gas determines whether the foam is "wet" or "dry". There is no exact boundary between these two regimes, but generally speaking, the flow of foam resembles that of liquid and the foam can be called "wet" in the case of gas content below 80% (Weaire, Hutzler 1999). In this paper, we consider wet foams with 60-77% air where the bubble shape is quite spherical and there is no direct contact between the bubbles. In order to have metastable foams, a foaming agent such as a surfactant is needed. At low concentration, the surfactant adsorbs onto the air-water interfaces, altering the free energies of the surfaces (Milton 2004). In this research, high surfactant concentrations from 2 up to 6 g/l were used due to the small size of our equipment.
Our target is to investigate the origin of the exceptional pore size distribution in foam forming. In particular, we determine the bubble size distribution produced by mechanical mixing and compare this to the pore size distribution in foam-formed paper samples. The effect of several parameters such as rotor speed and foam density together with the type and concentration of surfactant are investigated in order to determine how to control bubble size. The bubble size distribution is compared to the amount and size of bigger pores in paper samples prepared with various kinds of foams and fibres.
Material and Methods
Foam can be made in many different ways (Weaire, Hutzler 1999). In this study, foam was generated mechanically by mixing water with gas. The water contained surface active agent, and the gas in our case was air. Two types of ionic surfactant, the sodium dodecyl sulphate SDS (anionic surfactant) and hexadecyltrimethylammonium bromide HTAB (cationic surfactant) were studied. SDS is used in many different applications. It is commonly used in, e.g. shampoos, toothpastes and shaving creams. No articles are available concerning foam forming where various surfactants have been compared. However, a couple of results exist from related foam technologies. In the article concerning foam-assisted dewatering (FAD) three commercial surfactants (anionic, protein based and rosin based) and SDS (Skelton 1987) were mutually compared. According to these results, SDS offered the best performance. In one thesis studying foam coating, casein, PVA and SDS were mutually compared (Sievänen 2010). Also with reference to these studies, SDS was found to be most suitable surfactant. HTAB is the cationic surfactant we have used in foam-coating applications (Kinnunen 2011): it works well in this context and we have therefore included it in this study.
The experiment was divided into three parts: Firstly, bubble size distribution was determined for several cases by varying parameters, such as type of surfactant, concentration of surfactant, air content of the foam, and the rotational speed of the mixer. Rotational speed was varied between 1250 and 2000 RPM. Our foam-forming device has been constructed so that it always produces the same volume of foam. This means that if the starting volume of liquid is smaller, the amount of the air in the foam is bigger. The bubbles were detected by using a CCD camera with 980 x 1280 pixels of resolution. In order to do this, a few layers of bubbles were collected between two glass plates (Rodrigues 2003). In order to characterize the bubble size distribution and get good statistics, many images were taken from various positions of the foam. From the recorded images, we manually measured the radii of all clearly visible bubbles. A typical number of bubbles measured for a trial point was approximately 250.
In the second part of the experiment, unrefined pine kraft pulp was added to the water containing a surfactant. After mixing this suspension, we studied the effect of fibres on bubble-size distribution.
Finally, paper samples were made by applying foam forming in a laboratory sheet former. We used two different types of pulps - spruce CTMP and unrefined pine kraft - and two different kinds of surfactant mixtures - SDS and SDS+HTAB. The bubble-size distributions of the foam used in the sheet formed was measured in the same way as in other laboratory tests. The idea was to compare average bubble size with the pore-size distribution of the paper samples made.
Pore-size distributions in paper samples were measured using SEM cross cuts and image analysis. For the calculation of pore-size distribution, the image was enhanced and binarized so that object pixels were marked with value one and background pixels with zero. We then used the distance transformation (Rosenfeld 1966) of the binary image which is a grey-scale image, in which each object pixel was marked with the value of distance to its closest background pixel.
In our case, we transposed object pixels and background pixels, i.e. for each background pixel P, the mark of corresponding pixel in the distance transform image can be interpreted as a radius of a circle enclosed in the background and centred on P. Our algorithm applies sequential fulfilment of the image background, with the largest circles enclosed in the background. Figure 1 shows an example of the algorithm results realised for the SEM image of the foam-formed sample.
Figure 1. Example of of the realised foam-formed sample SEM image with the largest circles enclosed in the background.
Results and discussion
The histograms in Figure 2 describe the bubble size distributions for various cases. In particular, we studied the effect of air content, bubble size and surfactant concentration. In all the distributions, the variation in the bubble size is very large. In particular, some very large bubbles can be found, even though the majority of bubbles have a radius smaller than 100 microns.
Figure 2. Bubble-size distribution as a function of a) air content (0.1 g SDS surfactant, rotational speed 1250 RPM), b) SDS concentration (60% air content and 2000 RPM), c) rotational speed (2 g/l SDS and 2 g/l HTAB, 77% air content), and d) 77% air content and 0.5% fibre consistency.
Figure 3 shows the average bubble size as a function of several parameters, such as air content, type and concentration of surfactant, and rotational speed. In Figure 3a, the average bubble size is plotted as a function of air content. The other parameters are fixed. The surfactant type used is SDS with a fixed total mass of 0.1 g. Due to constant final foam volume, this means that surfactant concentration decreases when air content gets smaller (i.e. water volume increases). The resulting behaviour is shown for four varying rotation speeds. In general, the average bubble size increases with decreasing air content. However, we note an exception from this trend for low speed and large water volume (i.e. low surfactant concentration). In Figure 3b, it is shown that there is no clear relationship between the concentration and the bubble size. On the other hand, the average bubble radius decreases clearly with the increasing rotational speed of the mixer (see Figure 3c). This indicates that the energy dissipation within the mixer is an important factor affecting bubble size: higher speed means higher mixing energy. Figure 3d shows results for mixtures of two types of ionic surfactant (SDS and HTAB) both with and without fibres. Here the comparison is carried out at equal air content so that the surfactant concentrations are 2+2 g/l for SDS+HTAB and 2 g/l for the SDS surfactant. Despite the type of surfactant, the addition of fibres causes a significant decrease in the average bubble size. It is interesting that the relative drop in bubble size increases with mixing speed. This suggests that the drop has rather physical than chemical origin. A possible physical mechanism is the increase of dissipated energy in mixing due to higher viscosity caused by fibres.
Figure 3. Influence of a) air content, b) SDS surfactant concentration, c) rotation speed, and d) adding of fibres on average bubble size.
Figure 4. Bubble-size distribution for foams applied in sheet-making with two separate fibre types: a) with SDS surfactant, b) SDS+HTAB.
In laboratory sheet-making, we could affect the bubble size only with the type and concentration of surfactant and fibre type. This resulted from limitations of the larger size foam mixer, in which the mixing speed is fixed to 3000 rpm. In Figure 4, we show results of the bubble size distribution for CTMP and kraft pulps. In Figure 4a, SDS 0.24 g/l concentration was used with a fixed rotational speed, whilst Figure 4b corresponds to SDS and HTAB, 0.6 g/l and 0.36 g/l, respectively. In both cases, we obtain quite similar bubble-size distribution for both pulps. When comparing the average bubble size (see Figure 5), it appears the average bubble size is slightly smaller with CTMP, but the difference remains within the error bars. On the other hand, the addition of HTAB causes a significant drop in bubble size. Nevertheless, one must remember that the two cases correspond to varying surfactant concentrations. HTAB appears to cause, in particular, the increase of very small bubbles with a radius approx. 5 µm.
Figure 5. Average bubble size with two varying pulp types (CTMP and kraft). There is a huge difference in the average bubble radius attributable to the presence of HTAB.
We also measured the pore-size distributions of foam-formed samples. First, we made samples using CTMP fibre and SDS as surfactant. The pore size distributions were measured from the SEM cross-sections, as explained previously. The results are shown in Figure 6. They show that pore-size distributions of these variously formed samples are clearly different. In the case of the water-formed sample, there are more small pores than expected, whereas in the foam-formed samples, there are more large pores. We would like to emphasize that the effect seen in distribution becomes very strong when considering the total volume of the large pores. One possible reason for the large pores is the supportive effect of foam bubbles. The foam bubbles may act as "ghost" particles during forming. In a previous study (Hjelt 2011), where analysis was made using the µ-tomography method (Bliesner 1964), even more pronounced population of large pores could be observed.
Figure 6. Pore size distrtibution of water and foam formed samples. They were made using CTMP fibres and SDS as a surfactant.
To study the effect of the average bubble size on the paper pore size distribution, we made laboratory samples using CTMP and kraft fibre. Various bubble sizes were obtained using SDS and SDS+HTAB as a surfactant. The mixing speed in our laboratory sheet mould was fixed at 3000 rpm. The results are shown in Figure 7.
Figure 7. The pore size distributions of samples made using CTMP or kraft fibres and SDS+HTAB or SDS as surfactant.
The large difference in pore-size distributions clearly derives from different fibre types, as observed earlier (Niskanen 2008). The stiffer CTMP fibres are able to better retain large pores in the structure. However, the effect of the different average bubble size is also evident. When samples are made with SDS+HTAB as the surfactant, there are more small pores. Thus, foam which has smaller bubbles produces samples having more small pores. These results are in accordance of our assumption of foam bubbles acting as "ghost particles", as mentioned above.
Conclusions
Foam forming is an interesting new technology to make paper-like materials. However, relations between foam properties and the structure of the fibre network formed are still rather unclear. We have shown that in mechanical mixing, bubble size distribution can be affected by several parameters: for example, bubble size decreases with increasing rotational speed when more energy is dissipated in the foam to increase the total interface area between the bubbles. The type of surfactant is also an important parameter. When adding an unstable HTAB surfactant to a stable surfactant such as SDS, we can increase the number of small bubbles considerably. This effect may be caused by micelles formed in mixing of the surfactants. Our study shows an important interaction between added fibres and foam. With added fibres, we obtain a much smaller bubble size than the one in pure foam in comparison to equal water volume.
The largest pores in foam-formed paper samples have a similar size as the largest bubbles in the corresponding foam. In general, these pores are much larger than what is found in water-formed samples made with similar pulp. The interaction of fibres and foam introduces deviations from pore-size distribution obtained from water-formed samples. In the case of stiff fibres (CTMP), there are more large pores than in the case of flexible fibres (kraft). In both cases, the samples made with smaller bubbles in the foam had more small pores and fewer large pores, compared to samples made with bigger bubbles in the foam. This result clearly indicates that pore size distribution of paper made with foam forming depends on properties of foam. This enables the more flexible tuning of pore-size distribution than in the case of water forming.
Acknowledgments
The authors would like to thank the Graduate School at Technical Research Centre of Finland (VTT) for supporting this work. In particular, we would like to thank prof. Ali Harlin for valuable discussions. We would also like to extend our thanks for the financial support provided through the EffNet program of Forestcluster Ltd.
Literature
Bliesner W.C. (1964): A study of the porous structure of fibrous sheets using permeability techniques. Tappi J. 47 (7), 400.
Hjelt T., Kinnunen K., Lehmonen J., Beletski N., Hellén E., Liljeström V., Serimaa R., Miettinen A., and Kataja M. (2011): Intriguing structural and strength behaviour in foam forming, Progress in Paper Physics Seminar, Graz, 135.
Kinnunen, K., Hjelt T., and Kenttä E. (2011): Thin nanoparticle layers applied on the paper by foam coating technique, 25th PTS Coating Symposium, Munich.
Milton J. Rosen (2004): Surfactant and Interfacial Phenomena, John Wiley & Sons, Inc., Hoboken, New Jersey, pp. 1-30.
Niskanen K. (2008): Paper Physics, Finnish Paper Engineers' Association, Finland, pp.12-53.
Punton V.W. (1975a): Fibre Distribution in Foam and Foam-laid paper, Int. paper physics conf., 135-139.
Punton V.W. (1975b): The use of an aqueous foam as a fibre-suspending medium in quality papermaking. ed. Akers R.J. Foams. Proceedings of a symposium organized by the Society of the Chemical Industry. Colloid and surface chemistry Group, and held at Brunel University. September 8-10.
Radvan B., and Gatward A.P,J (1972): The Formation of Wet-Laid Webs by a Foaming Process, Tappi J. 55(5), 748.
Rodrigues R.T., Rubio J. (2003): New basis for measuring the size distribution of bubbles, Minerals Engineering 16, 757-765.
Rosenfeld, A. and Pfaltz, J.L. (1966): Sequential operations in digital picture processing, J. Assos. Comput. Mach., 13(4), pp. 471-494.
Sievänen H., Suitability of foam coating on application of thin liquid films, Master's Thesis, Lappeenranta University of Technology, 2010.
Skelton J., Paper Technology and Industry, March 1987, p. 431-436.
Smith M.K., Punton V.W., and Rixson A.G. (1974): The structure and properties of paper formed by a foaming process. Tappi. Vol 57, No 1. 107-111.
Smith M.K. and Punton V.W. (1975): Foam can improve formation. Pulp & Paper Canada. Vol. 76, No 1. 55-58.
Weaire, D., and Hutzler W. (1999): The Physics of Foams, Clarendon Press, Oxford, pp. 1-55.
Wiggins Teape, (1973): Make paper with foam not without it. A report from Wiggins Teape Research and Development Ltd. Pulp Trade Journal. Vol 16. 28 - 31.
