- Email: [email protected]
A simple optical device for measuring free surface deformations of nontransparent liquids
Journal of Colloid and Interface Science 288 (2005) 508–512 www.elsevier.com/locate/jcis
A simple optical device for measuring free surface deformations of nontransparent liquids E.G. Megalios, N. Kapsalis, J. Paschalidis, A.G. Papathanasiou, A.G. Boudouvis ∗ School of Chemical Engineering, National Technical University of Athens, Athens 15780, Greece Received 16 December 2004; accepted 12 March 2005 Available online 21 April 2005
Abstract In this paper, a novel optical device for measuring the deformation of liquid free surfaces is presented. The device employs a laser beam, which can be focused on any chosen location on the free surface. The key measurement is of the intensity of the beam reflected from a location on the free surface where the deformation exhibits a local extremum. The optics of the device is so designed as to measure a maximum intensity when the distance between the focusing lens and the selected point on the free surface is equal to the focal length, thus enabling a height measurement. The device is tested in ferrofluid pools where the height of the spikes of the normal field instability is measured. The simplicity of the suggested technique enables the fabrication of a quite cheap device for measuring surface deformation of nontransparent liquids, which provides good accuracy and reproducibility. 2005 Elsevier Inc. All rights reserved. Keywords: Nontransparent liquid; Free surface; Optical device; Ferrofluid; Laser beam
1. Introduction and motivation This work is motivated by the need to measure free surface static deformations of nontransparent liquids such as magnetic liquids (commonly called ferrofluids). Ferrofluids are stable synthetic colloidal suspensions of magnetite particles in carrier liquids, typically synthetic hydrocarbons or water [1]. Due to their strong polarization when subjected to magnetic fields they can easily be guided or deformed. Combining the magnetic properties of a solid with the rheological properties of a common liquid, magnetic liquids have many commercial applications. Moreover, many simple ferrofluid systems are studied as realistic prototypes of spontaneous pattern-forming systems ranging from physics and chemistry to developmental biology [2,3]. A case in point is the so-called normal field instability or Rosensweig instability [4]. The instability is caused by a uniform magnetic field applied normally to the (almost) flat free surface of a * Corresponding author. Fax: +30 210 772 3155.
E-mail address: [email protected] (A.G. Boudouvis). 0021-9797/$ – see front matter 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2005.03.033
ferrofluid pool. When the field strength exceeds a threshold, the flat surface turns unstable and it gives way to an appreciably deformed one, with pointed peaks, or “spikes” [5,6]; the peaks are arranged, usually, in a hexagonal lattice or, in some cases, a square lattice [4,6]. Measuring the wavelength of the lattice (order of 1 cm) is an easy task, since a top view snapshot gives the required information with the required detail. The measurement of the height and the profile of each peak was an issue under study for many years, since the needlelike tip of the peak hinders standard methods of measuring surface deformation. Shadowgraphy utilizes the deformed surface as a (de)focusing mirror for a parallel beam and can be applied successfully to measure the surface deformations as long as the free surface is not highly deformed [7]. However, the fully developed spike is too sharp, so that no image is given by the method. The simplest optical technique, the side view snapshot, fails to capture the development of the surface deformation, since the nontransparent liquid climbs at the sidewalls of the container and thus shadows the inner peaks. Thus, the application range of methods based on light absorption [8] is restricted. Recently, Richter and Blas-
E.G. Megalios et al. / Journal of Colloid and Interface Science 288 (2005) 508–512
ing [9] have measured the reduction of the intensity of an X-ray beam, when it passes through a film of magnetic liquid over a thin layer of Teflon, providing a 3-D image of the free surface. The method is accurate enough and gives the full height profile; however, the overall cost is prohibitive. A simple optical technique is presented here for the measurement of the height of a single peak. The proposed device utilizes the intensity of the reflected laser beam, which is focused on a single spot of the free surface with the assumption that the free surface is locally flat there, i.e., the spot is located at a local extremum of the surface. Measurements show adequate repeatability and accuracy. The simplicity of the suggested technique makes the measurement device cheap and easy to build. Its application is general and is not restricted to measurements in ferrofluid pools but is also useful wherever precise level monitoring in liquid vessels is required, especially where sensor–liquid contact is not allowed. In Section 2, the principle of operation of the device is explained. Section 3 describes the basic parts of the fabricated apparatus, and in Section 4, measurements performed are discussed, while advantages and drawbacks are outlined.
2. Principle of operation A sensitive optical method is utilized to provide highprecision measurements of liquid free surface deformations. A light beam is directed vertically and focused on a single spot on a surface. When the surface is locally flat, i.e., at
(a)
509
local minima or maxima of the deformation, the beam reflects back to the focal lens. The reflecting beam intensity is measured and reaches a maximum when the distance, d (see Fig. 1), between the lens and the free surface coincides with the focal length, f . The lens can be translated along the three axes. When the distance is kept fixed and equal to the focal length, any vertical displacement of the free surface results in an equal vertical displacement of the lens. Monitoring this movement yields the surface elevation at the selected point. This technique requires the use of a narrow light beam, which is generated by a laser diode module (λ = 670 nm). The laser beam is focused by a lens (DCX, doubly convex lens) directly on the liquid free surface. In Fig. 1a the path of the incident beam is shown. A fraction of the beam is reflected back to the focal lens, where it is defocused, and through a transparent mirror is directed to the receptor, which measures its intensity (see Fig. 1b). An important feature of the suggested method is that the intensity of the reflected beam is affected only by the distance, d, between the lens and the liquid free surface, and the local free surface slope. Fig. 2 shows the reflected beam intensity as a function of the distance between the lens and the free surface. The maximum value is obtained only when the distance becomes equal to the focal length, assuming that all the fraction of the reflected beam returns back to the lens (see Fig. 3a). Any vertical displacement of the lens from the focal distance results in a decrease of the corresponding intensity, since the fraction of the reflected beam that finally returns to the receptor decreases (see Figs. 3b and 3c). The intensity
(b)
Fig. 1. Illustration of the main parts of the device (not in scale). (a) The incident beam and (b) the reflected beam are shown.
510
E.G. Megalios et al. / Journal of Colloid and Interface Science 288 (2005) 508–512
surface is not locally flat and the reflected beam is directed away.
3. Experimental setup
Fig. 2. Intensity of the reflected beam as a function of the distance, d, between the lens and the free surface.
A laser lead is placed on the topside of the vertical and nonreflective tube, which is mounted in a mechanism that allows x–y–z movement in space; the x and y axes are on a horizontal plane and the z axis is vertical. The focal lens with a focal distance f = 6 mm is installed on the other side. A transparent mirror is placed so that the incident and the reflected beam are driven to the optical fiber and the receptor, respectively (see Figs. 1a and 1b); the operation of the device is indicated through the optical fiber. The laser receptor is connected to a three-scale signal amplifier: ×1, ×10, ×100. The amplified signal, corresponding to the intensity of the reflected beam, can be read out through an analog indicator. The x–y–z table is equipped with three micrometric slides enabling the precise movement of the lens in space, with a resolution of 1 mm per turn of the micrometric screw. The vertical displacement of the lens, that indicates the vertical displacement of the free surface at a point, is indicated by a digital micrometric dial, the precision of which is 0.01 mm. The x–y location of the point can be read out from the micrometric slides, giving a variation of the indicated position of ∼0.03 mm. The parts of the apparatus are illustrated in Fig. 4. All the apparatus components are made of nonmagnetic materials in order to avoid any field distortion due to magnetization. Since the operational principle and the device reported are simple, the total construction cost is kept very low, not exceeding 5000 EUR. The experimental setup is illustrated in Fig. 5.
4. Measurements Fig. 3. Instrument probe placed at in different positions relative to the liquid surface extremum: (a) d = f , (b) d < f , (c) d > f , (d) d = f , but the probe is not placed exactly above the local maximum of the deformation. The focal point is denoted by P.
becomes negligible for a displacement greater than ∼1.5 mm (see Fig. 2). When different samples, water and ferrofluid (APG-J10), are tested, it comes out that the intensity of the signal strongly depends on the extent of the laser light penetration into the main mass of the liquid. The less transparent the liquid is (the ferrofluid, in our case), the stronger the signal produced for the same range of displacement (see Fig. 2); thus the measurements in the less transparent liquid are more accurate. In Fig. 3d is depicted the case where the lens is not placed exactly above the local extremum of the deformation. Regardless of the distance between the lens and the free surface, the detected signal becomes weak or zero, since the
Measurements are reported for small-diameter pools where the free surface deformation is always axisymmetric, thus having a local extremum at the axis of symmetry of any applied field [10]. The height of the surface at that point is measured. Each measurement starts with a calibration where a reference zero level is set. This level corresponds to the zero applied field state. The probe (see Fig. 4) is placed above the selected point and its distance, d, from the free surface is adjusted so that a maximum of reflected intensity is obtained when d = f . The same adjustment is performed every time the surface elevation changes due to the applied field change. We have made measurements with different focal length lenses. Using a lens with a smaller focal length, the signal gets stronger, but the probe (tube) has to get closer to the liquid surface, thus making the measurement risky to perform, since touching the liquid is undesirable. Although a bigger focal length facilitates the measurement procedure,
E.G. Megalios et al. / Journal of Colloid and Interface Science 288 (2005) 508–512
511
Fig. 4. Drawing of the apparatus.
Fig. 5. Illustration of the experimental setup.
the acquired signal gets weaker. Thus, one has to compromise between accuracy and usability. The technique is tested for several hydrocarbon ferrofluid samples. Here measurements are presented for APG-J10 from FerroTec with the following properties: magnetization
at saturation Ms = 250 G, surface tension σ = 0.0258 N/m, initial susceptibility χj = 1.1, density ρ = 1.16 g/ml, viscosity µ = 70 cP. The dependence of the deformation height on the applied field is shown in Fig. 6. The free surface falls at the axis of symmetry while it climbs at the pool sidewalls as long as the field strength does not exceed a critical value, Bcr . When Bcr is exceeded a small spike is formed at the axis of symmetry. Its tip is initially lower than the zero level; however, it sharply increases as the applied field gets stronger, attaining a height of some millimeters for field strength roughly twice Bcr . The accuracy of the proposed method depends on the spike sharpness. Repeated measurements showed good reproducibility with a maximum absolute variation less than 0.25 mm observed for spikes 3 mm in height. In fact, the precision of the present method is similar to the one that was achieved with shadowgraphy or the radioscopy method. Another advantage of the proposed method is the precise determination of the onset of the normal field instability—a measurement that is hard to determine with shadowgraphy [7]. The instability is accom-
512
E.G. Megalios et al. / Journal of Colloid and Interface Science 288 (2005) 508–512
surface profile, since it precisely measures only the height of local extrema, it has the serious advantages of low cost and flexible operation. Measurements are reported for static deformations of ferrofluid pools in magnetic fields and the accuracy and repeatability of the method are shown to be comparable to those of other more complicated and expensive techniques based on shadowgraphy or radioscopy.
Acknowledgments
Fig. 6. Measurement of free surface deformation (peak height) versus applied field strength. The pool average depth is 18 mm. Bcr = 129 G.
panied by an abrupt change of the surface curvature and this change is read by an immediate signal loss. The precision is thus strongly dependent on the field strength increment resolution. The proposed device is the simplest possible and the x– y–z movement of the probe is performed by means of manually adjusted micrometer dials. A fully automated version utilizing computer-controlled motors driven in closed loop with the signal measurement would dramatically improve the measurement speed and flexibility.
5. Summary and conclusions The presented work suggests a simple technique for measuring static free surface deformations of nontransparent liquids. Although the device is not able to measure the full
The authors are indebted to M. Bistekos for his excellent machine work in the device construction. The ferrofluid used was kindly provided by the FerroTec Corporation, USA. Partial support was provided by the General Secretariat for Research and Technology through the ENTEP program.
References [1] R.E. Rosensweig, Ferrohydrodynamics, Dover, New York, 1997. [2] M. Seul, D. Andelman, Science 267 (1995) 476. [3] D’Arcy W. Thomson, On Growth and Form, Cambridge Univ. Press, Cambridge, UK, 1972. [4] M.D. Cowley, R.E. Rosensweig, J. Fluid Mech. 30 (1967) 671. [5] A. Lange, H. Langer, A. Engel, Physica D 111 (2000) 294. [6] A.G. Boudouvis, J.L. Puchalla, L.E. Scriven, R.E. Rosensweig, J. Magn. Magn. Mater. 65 (1987) 307. [7] J. Browaeys, J.-C. Flament, S. Neveu, R. Perzynski, Eur. Phys. J. B 9 (1999) 335. [8] W.B. Wright, R. Budakian, S.J. Putterman, Phys. Rev. Lett. 76 (1996) 4528. [9] R. Richter, J. Blasing, Rev. Sci. Instrum. 72 (3) (2001) 1729. [10] A.G. Boudouvis, J.L. Puchalla, L.E. Scriven, J. Colloid Interface Sci. 124 (1988) 677.
A simple optical device for measuring free surface deformations of nontransparent liquids E.G. Megalios, N. Kapsalis, J. Paschalidis, A.G. Papathanasiou, A.G. Boudouvis ∗ School of Chemical Engineering, National Technical University of Athens, Athens 15780, Greece Received 16 December 2004; accepted 12 March 2005 Available online 21 April 2005
Abstract In this paper, a novel optical device for measuring the deformation of liquid free surfaces is presented. The device employs a laser beam, which can be focused on any chosen location on the free surface. The key measurement is of the intensity of the beam reflected from a location on the free surface where the deformation exhibits a local extremum. The optics of the device is so designed as to measure a maximum intensity when the distance between the focusing lens and the selected point on the free surface is equal to the focal length, thus enabling a height measurement. The device is tested in ferrofluid pools where the height of the spikes of the normal field instability is measured. The simplicity of the suggested technique enables the fabrication of a quite cheap device for measuring surface deformation of nontransparent liquids, which provides good accuracy and reproducibility. 2005 Elsevier Inc. All rights reserved. Keywords: Nontransparent liquid; Free surface; Optical device; Ferrofluid; Laser beam
1. Introduction and motivation This work is motivated by the need to measure free surface static deformations of nontransparent liquids such as magnetic liquids (commonly called ferrofluids). Ferrofluids are stable synthetic colloidal suspensions of magnetite particles in carrier liquids, typically synthetic hydrocarbons or water [1]. Due to their strong polarization when subjected to magnetic fields they can easily be guided or deformed. Combining the magnetic properties of a solid with the rheological properties of a common liquid, magnetic liquids have many commercial applications. Moreover, many simple ferrofluid systems are studied as realistic prototypes of spontaneous pattern-forming systems ranging from physics and chemistry to developmental biology [2,3]. A case in point is the so-called normal field instability or Rosensweig instability [4]. The instability is caused by a uniform magnetic field applied normally to the (almost) flat free surface of a * Corresponding author. Fax: +30 210 772 3155.
E-mail address: [email protected] (A.G. Boudouvis). 0021-9797/$ – see front matter 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2005.03.033
ferrofluid pool. When the field strength exceeds a threshold, the flat surface turns unstable and it gives way to an appreciably deformed one, with pointed peaks, or “spikes” [5,6]; the peaks are arranged, usually, in a hexagonal lattice or, in some cases, a square lattice [4,6]. Measuring the wavelength of the lattice (order of 1 cm) is an easy task, since a top view snapshot gives the required information with the required detail. The measurement of the height and the profile of each peak was an issue under study for many years, since the needlelike tip of the peak hinders standard methods of measuring surface deformation. Shadowgraphy utilizes the deformed surface as a (de)focusing mirror for a parallel beam and can be applied successfully to measure the surface deformations as long as the free surface is not highly deformed [7]. However, the fully developed spike is too sharp, so that no image is given by the method. The simplest optical technique, the side view snapshot, fails to capture the development of the surface deformation, since the nontransparent liquid climbs at the sidewalls of the container and thus shadows the inner peaks. Thus, the application range of methods based on light absorption [8] is restricted. Recently, Richter and Blas-
E.G. Megalios et al. / Journal of Colloid and Interface Science 288 (2005) 508–512
ing [9] have measured the reduction of the intensity of an X-ray beam, when it passes through a film of magnetic liquid over a thin layer of Teflon, providing a 3-D image of the free surface. The method is accurate enough and gives the full height profile; however, the overall cost is prohibitive. A simple optical technique is presented here for the measurement of the height of a single peak. The proposed device utilizes the intensity of the reflected laser beam, which is focused on a single spot of the free surface with the assumption that the free surface is locally flat there, i.e., the spot is located at a local extremum of the surface. Measurements show adequate repeatability and accuracy. The simplicity of the suggested technique makes the measurement device cheap and easy to build. Its application is general and is not restricted to measurements in ferrofluid pools but is also useful wherever precise level monitoring in liquid vessels is required, especially where sensor–liquid contact is not allowed. In Section 2, the principle of operation of the device is explained. Section 3 describes the basic parts of the fabricated apparatus, and in Section 4, measurements performed are discussed, while advantages and drawbacks are outlined.
2. Principle of operation A sensitive optical method is utilized to provide highprecision measurements of liquid free surface deformations. A light beam is directed vertically and focused on a single spot on a surface. When the surface is locally flat, i.e., at
(a)
509
local minima or maxima of the deformation, the beam reflects back to the focal lens. The reflecting beam intensity is measured and reaches a maximum when the distance, d (see Fig. 1), between the lens and the free surface coincides with the focal length, f . The lens can be translated along the three axes. When the distance is kept fixed and equal to the focal length, any vertical displacement of the free surface results in an equal vertical displacement of the lens. Monitoring this movement yields the surface elevation at the selected point. This technique requires the use of a narrow light beam, which is generated by a laser diode module (λ = 670 nm). The laser beam is focused by a lens (DCX, doubly convex lens) directly on the liquid free surface. In Fig. 1a the path of the incident beam is shown. A fraction of the beam is reflected back to the focal lens, where it is defocused, and through a transparent mirror is directed to the receptor, which measures its intensity (see Fig. 1b). An important feature of the suggested method is that the intensity of the reflected beam is affected only by the distance, d, between the lens and the liquid free surface, and the local free surface slope. Fig. 2 shows the reflected beam intensity as a function of the distance between the lens and the free surface. The maximum value is obtained only when the distance becomes equal to the focal length, assuming that all the fraction of the reflected beam returns back to the lens (see Fig. 3a). Any vertical displacement of the lens from the focal distance results in a decrease of the corresponding intensity, since the fraction of the reflected beam that finally returns to the receptor decreases (see Figs. 3b and 3c). The intensity
(b)
Fig. 1. Illustration of the main parts of the device (not in scale). (a) The incident beam and (b) the reflected beam are shown.
510
E.G. Megalios et al. / Journal of Colloid and Interface Science 288 (2005) 508–512
surface is not locally flat and the reflected beam is directed away.
3. Experimental setup
Fig. 2. Intensity of the reflected beam as a function of the distance, d, between the lens and the free surface.
A laser lead is placed on the topside of the vertical and nonreflective tube, which is mounted in a mechanism that allows x–y–z movement in space; the x and y axes are on a horizontal plane and the z axis is vertical. The focal lens with a focal distance f = 6 mm is installed on the other side. A transparent mirror is placed so that the incident and the reflected beam are driven to the optical fiber and the receptor, respectively (see Figs. 1a and 1b); the operation of the device is indicated through the optical fiber. The laser receptor is connected to a three-scale signal amplifier: ×1, ×10, ×100. The amplified signal, corresponding to the intensity of the reflected beam, can be read out through an analog indicator. The x–y–z table is equipped with three micrometric slides enabling the precise movement of the lens in space, with a resolution of 1 mm per turn of the micrometric screw. The vertical displacement of the lens, that indicates the vertical displacement of the free surface at a point, is indicated by a digital micrometric dial, the precision of which is 0.01 mm. The x–y location of the point can be read out from the micrometric slides, giving a variation of the indicated position of ∼0.03 mm. The parts of the apparatus are illustrated in Fig. 4. All the apparatus components are made of nonmagnetic materials in order to avoid any field distortion due to magnetization. Since the operational principle and the device reported are simple, the total construction cost is kept very low, not exceeding 5000 EUR. The experimental setup is illustrated in Fig. 5.
4. Measurements Fig. 3. Instrument probe placed at in different positions relative to the liquid surface extremum: (a) d = f , (b) d < f , (c) d > f , (d) d = f , but the probe is not placed exactly above the local maximum of the deformation. The focal point is denoted by P.
becomes negligible for a displacement greater than ∼1.5 mm (see Fig. 2). When different samples, water and ferrofluid (APG-J10), are tested, it comes out that the intensity of the signal strongly depends on the extent of the laser light penetration into the main mass of the liquid. The less transparent the liquid is (the ferrofluid, in our case), the stronger the signal produced for the same range of displacement (see Fig. 2); thus the measurements in the less transparent liquid are more accurate. In Fig. 3d is depicted the case where the lens is not placed exactly above the local extremum of the deformation. Regardless of the distance between the lens and the free surface, the detected signal becomes weak or zero, since the
Measurements are reported for small-diameter pools where the free surface deformation is always axisymmetric, thus having a local extremum at the axis of symmetry of any applied field [10]. The height of the surface at that point is measured. Each measurement starts with a calibration where a reference zero level is set. This level corresponds to the zero applied field state. The probe (see Fig. 4) is placed above the selected point and its distance, d, from the free surface is adjusted so that a maximum of reflected intensity is obtained when d = f . The same adjustment is performed every time the surface elevation changes due to the applied field change. We have made measurements with different focal length lenses. Using a lens with a smaller focal length, the signal gets stronger, but the probe (tube) has to get closer to the liquid surface, thus making the measurement risky to perform, since touching the liquid is undesirable. Although a bigger focal length facilitates the measurement procedure,
E.G. Megalios et al. / Journal of Colloid and Interface Science 288 (2005) 508–512
511
Fig. 4. Drawing of the apparatus.
Fig. 5. Illustration of the experimental setup.
the acquired signal gets weaker. Thus, one has to compromise between accuracy and usability. The technique is tested for several hydrocarbon ferrofluid samples. Here measurements are presented for APG-J10 from FerroTec with the following properties: magnetization
at saturation Ms = 250 G, surface tension σ = 0.0258 N/m, initial susceptibility χj = 1.1, density ρ = 1.16 g/ml, viscosity µ = 70 cP. The dependence of the deformation height on the applied field is shown in Fig. 6. The free surface falls at the axis of symmetry while it climbs at the pool sidewalls as long as the field strength does not exceed a critical value, Bcr . When Bcr is exceeded a small spike is formed at the axis of symmetry. Its tip is initially lower than the zero level; however, it sharply increases as the applied field gets stronger, attaining a height of some millimeters for field strength roughly twice Bcr . The accuracy of the proposed method depends on the spike sharpness. Repeated measurements showed good reproducibility with a maximum absolute variation less than 0.25 mm observed for spikes 3 mm in height. In fact, the precision of the present method is similar to the one that was achieved with shadowgraphy or the radioscopy method. Another advantage of the proposed method is the precise determination of the onset of the normal field instability—a measurement that is hard to determine with shadowgraphy [7]. The instability is accom-
512
E.G. Megalios et al. / Journal of Colloid and Interface Science 288 (2005) 508–512
surface profile, since it precisely measures only the height of local extrema, it has the serious advantages of low cost and flexible operation. Measurements are reported for static deformations of ferrofluid pools in magnetic fields and the accuracy and repeatability of the method are shown to be comparable to those of other more complicated and expensive techniques based on shadowgraphy or radioscopy.
Acknowledgments
Fig. 6. Measurement of free surface deformation (peak height) versus applied field strength. The pool average depth is 18 mm. Bcr = 129 G.
panied by an abrupt change of the surface curvature and this change is read by an immediate signal loss. The precision is thus strongly dependent on the field strength increment resolution. The proposed device is the simplest possible and the x– y–z movement of the probe is performed by means of manually adjusted micrometer dials. A fully automated version utilizing computer-controlled motors driven in closed loop with the signal measurement would dramatically improve the measurement speed and flexibility.
5. Summary and conclusions The presented work suggests a simple technique for measuring static free surface deformations of nontransparent liquids. Although the device is not able to measure the full
The authors are indebted to M. Bistekos for his excellent machine work in the device construction. The ferrofluid used was kindly provided by the FerroTec Corporation, USA. Partial support was provided by the General Secretariat for Research and Technology through the ENTEP program.
References [1] R.E. Rosensweig, Ferrohydrodynamics, Dover, New York, 1997. [2] M. Seul, D. Andelman, Science 267 (1995) 476. [3] D’Arcy W. Thomson, On Growth and Form, Cambridge Univ. Press, Cambridge, UK, 1972. [4] M.D. Cowley, R.E. Rosensweig, J. Fluid Mech. 30 (1967) 671. [5] A. Lange, H. Langer, A. Engel, Physica D 111 (2000) 294. [6] A.G. Boudouvis, J.L. Puchalla, L.E. Scriven, R.E. Rosensweig, J. Magn. Magn. Mater. 65 (1987) 307. [7] J. Browaeys, J.-C. Flament, S. Neveu, R. Perzynski, Eur. Phys. J. B 9 (1999) 335. [8] W.B. Wright, R. Budakian, S.J. Putterman, Phys. Rev. Lett. 76 (1996) 4528. [9] R. Richter, J. Blasing, Rev. Sci. Instrum. 72 (3) (2001) 1729. [10] A.G. Boudouvis, J.L. Puchalla, L.E. Scriven, J. Colloid Interface Sci. 124 (1988) 677.