- Email: [email protected]
Nanobubble enhanced agglomeration of hydrophobic powders
Colloids and Surfaces A 530 (2017) 117–123
Contents lists available at ScienceDirect
Colloids and Surfaces A journal homepage: www.elsevier.com/locate/colsurfa
Nanobubble enhanced agglomeration of hydrophobic powders ⁎
MARK
Paul Knüpfer , Lisa Ditscherlein, Urs Alexander Peuker Institute of Mechanical Process Engineering and Mineral Processing, Agricolastraße 1, 09599 Freiberg, Germany
G RA P H I C A L AB S T R A C T
A R T I C L E I N F O
A B S T R A C T
Keywords: Agglomerate stability Agglomeration Hydrophobic particles Nanobubbles
Agglomeration of solid particles in aqueous liquid flows is generally influenced by the appropriate interacting forces. Between hydrophobic particles, these forces are dramatically increased compared to van der Waals forces due to the formation of capillary bridges, caused by nanobubbles located at the particle surface. In this study, both hydrophobic and hydrophilic agglomerates were generated in water with an agglomeration setup based on a stirred tank reactor. Variating the shear rate leads to different agglomeration sizes and provides information about the agglomerate strength. In principle, both stability and size of the hydrophobic agglomerates are larger than the hydrophilic ones. The presence of the nanobubbles is proved by applying atomic force microscopy to the particle surface.
1. Introduction Agglomeration of fine particles plays a role in different industries. One major representative of hydrophobic (non-wetting) agglomeration is metallurgical processing, especially liquid metal cleaning by filtration and flotation. Non-metallic inclusions in metal melts are undesired for a high stability of the cast products. To increase the fatigue strength of these products, a filtration is applied to the metal melt during the casting process. Therefore porous ceramic foam filters, consisting of
alumina, are used to separate the melt from inclusions. A depth filtration mechanism is assumed, because the pore diameter, where the inclusions enter the filter, are much larger than the inclusions themselves. If the inclusion attaches to the inner filter surface, the adhesion is provided due to van der Waals interactions, sintering and particularly capillary forces. Neither the inclusions nor the ceramic foam filter are wetted by the metal melt due to its high surface energy, which exceeds the surface energy of water about several magnitudes [1]. Gas reservoirs on the particulate inclusions as well as the filter surface are
Abbreviations: AFM, atomic force microscopy; ASD, agglomerate size distribution; IEP, isoelectric point; PSD, particle size distributions; rms, root mean square ⁎ Corresponding author. E-mail addresses: [email protected] (P. Knüpfer), [email protected] (L. Ditscherlein), [email protected] (U.A. Peuker). http://dx.doi.org/10.1016/j.colsurfa.2017.07.056 Received 6 June 2017; Received in revised form 17 July 2017; Accepted 18 July 2017 Available online 19 July 2017 0927-7757/ © 2017 Elsevier B.V. All rights reserved.
Colloids and Surfaces A 530 (2017) 117–123
P. Knüpfer et al.
a range of a few nm while decreasing exponentially [13]. The most common explanation of the long range forces is due to the formation of nanobubbles onto the hydrophobic surface, which are leading to strong capillary interactions [14–16]. The behavior and the generation of nanobubbles at smooth surfaces is summarized already in literature [17,18]. Nanobubbles on solid surfaces can be produced via gas supersaturation in liquids that are achieved by electrolysis [19,20], perturbation [2] or heating the substrate or the liquid [4]. A widely used procedure to generate nanobubbles onto hydrophobic surfaces is the solvent exchange method [21–24]. This procedure was developed by Lou et al. [25] and Ishida et al. [26] to generate nanobubbles in a reproducible way. Here, the hydrophobic surface is firstly immersed in a wetting liquid, which is a good gas solvent. The next step is to exchange the solvent, e.g. ethanol, with a second poorly wetting liquid (e.g. water) which contains less dissolved air. In addition, the enthalpy of the ethanol-water-mixing causes local temperature gradients. Thus, a supersaturation of gas occurs in the mixing zone during the exchange. Nanobubbles are forming spontaneously at the surface and grow to a height of several hundred nm. Another important property of nanobubbles is their high stability. Actual, the bubbles should dissolve in microseconds, but they stay stable even for hours and days [22,27,28]. Although it is not totally clear, the most common explanations for the high stability is due to line pinning [29,30] and a dynamic equilibrium of gas flux at the liquid-gas interface [31]. The detection of nanobubbles is commonly possible via atomic force microscopy (AFM), whereby a soft cantilever is used at intermittent mode [18]. Furthermore, due to the colloidal probe technique, it is possible to measure adhesive forces between a single particle and a surface via AFM. Therefore, a spherical particle is glued onto a tipless cantilever [32] and after recording force-distance curves between this particle and a surface, the maximum adhesive force can be determined. Because real surfaces are rough, it is more practical to quantify the detachment forces by an adhesive force distribution, as described in [33]. The relative mobility of the particles is strongly influenced by the micro turbulence of a liquid movement. It is necessary to transport the particles to each other, but can also lead to break up induced by larger shear forces [34]. If these two processes are equilibrated, the size of agglomerates is reaching a constant value after a certain time t. Additional stress to the agglomerates could lead to a breakage and disabling this stress provides a regrowth ending up to the previous size of agglomerates [35,36]. The stability of agglomerates towards shear stress can be determined by observation of the size of the agglomerates at different shear conditions [37]. According to the theory of turbulence, the eddy size is distributed. Starting from the macroscopic eddies the dissipation of energy finally occurs in the smallest micro eddies. While stirring the liquid in the vessel, the energy is dissipating at a rate of the integral energy dissipationε . The latter can be evaluated from the difference in energy consumption between stirring with and without the liquid relative to its mass. By this, the character of turbulence can be estimated. The microscale η of Kolmogorov (Eq. (1)) gives the length scale of the smallest eddies, where the Reynolds number Re equals 1 [38].
caused by dissolved gas in the metal melt or during immersion. The reservoirs can also develop at the moment of particle-surface contact [2]. A capillary is formed and leads to a strong binding force of the inclusions resulting in the deposition within the foam filter. As it is known from the theory of depth filtration, inclusion particles are able to cross the filter without establishing contact to the filter surface, thus a pre-agglomeration of these particles is necessary to increase the contact probability and thus to purify the metal melt in the most efficient way. Metal melts are opaque and solely treatable at high temperatures, that impedes investigations of inclusion agglomeration at the real system. Therefore, a room temperature model system is used, that mimics the poor wetting behavior of metal melts on nonmetallic inclusions and filter materials. In this model system alumina particles with a hydrophobic coating represent the inclusions while water is used as a non-wetting liquid. In the study, the agglomeration of hydrophobic powders is realized in a stirred vessel. To stimulate and control the hydrophobic agglomeration a novel approach is applied, which introduces defined nanobubbles in the system, to the particle surface respectively to increase the adhesion force between the particles dramatically due to capillary bridging [3,4]. In the recent past, similar investigations in this room temperature model system were done to understand and to improve the cleaning of metal melts. Saint-Raymond et al. investigated the agglomeration of fine alumina powders at different pH-values in water and in n-heptane [5]. Cournil et al. observed a higher agglomeration rate of hydrophobic particles than hydrophilic ones both theoretically [6] and experimentally [7]. Therefore, in situ measurements of turbidity in a stirred vessel at different agitation speeds were done [8]. Gruy et al. assumed the preexistence of nanobubbles on the particles and accomplished the agglomeration experiments at the pH-value of isoelectric point (IEP) and with dispersing the particles in a 15 vol.% ethanol-water mixture before the agglomeration starts [8]. This study has the aim to nucleate defined nanobubbles onto hydrophobic particles through a solvent exchange. A subsequent agglomeration of these particles in a non-wetting liquid is done. The emerged nanobubbles can be detected and qualified via AFM on comparable hydrophobic surfaces at corresponding physicochemical conditions. The hydrophobic agglomerates are compared with hydrophilic ones in size and stability, which were obtained in the same setup at the IEP. Also, high electrolyte concentrations were produced to reveal the enhancement of adhesion forces based on poor wetting property of the system. 2. Theoretical approach The agglomeration of hydrophobic powders can be described with established techniques of flocculation, as used e.g. in waste water treatment. Agglomeration occurs if primary particles are both transported and attached to each other [9]. The transport requires a relative displacement (e.g. sedimentation, flow field, Brownian motion) and the attachment an attractive force between the particles. Two components of interaction forces between particles immersed in polar liquids like water are known according to the DLVO-theory, which was established in the 1940s by Derjaguin, Landau [10], Verwey and Overbeek [11]. The van der Waals force represents the attractive component. This force is based on molecular dipoles of the different phases and scales with the distance to the negative power of three in the case of two colliding particles. The repulsive interactions is formed through the presence of the electrochemical double layer and decreases approximately exponentially relating to the distance. A further force occurs between hydrophobic surfaces in aqueous solutions, which cannot be explained by the traditional DLVO theory [12]. This attractive hydrophobic force is commonly divided into two parts as well, the so called short range and long range hydrophobic force. The short range force is commonly described as structural formation of the water molecules at the solid-liquid interface and exhibits
3
ν η=⎛ ⎞ ⎝ε ⎠ ⎜
1 4
⎟
(1)
Agglomerates with a size below the Kolmogorov scale are stressed by the dominating viscous shear forces and are consequently submitted to erosion. In this case, small parts in the size range of the primary particles detach. If the agglomerate size is significantly above the Kolmogorov scale, large scale fragmentation occurs due to dynamic pressures and the agglomerates decompose into several smaller agglomerates [37]. Camp and Stein [39] defined the global root mean shear rates G in a turbulent flow, which is related to the energy dissipation rate as well as 118
Colloids and Surfaces A 530 (2017) 117–123
P. Knüpfer et al.
the kinematic viscosity of the fluid, see Eq. (2). Thereby, this value is directly connected to the rotational speed of the propeller.
3.2. Particles and characterization For the measurements α−alumina particles T60/T64 (Almatis, Germany) classified to a size of 2-25 μm, CT9FG (Almatis, Germany) sized to 2–10 μm, respectively, are used. To hydrophobize the surface, the particles were coated with Tridecafluorotriethoxysilane (Dynasilan F8261 from Evonik, Germany) as described in [40]. The hydrophobic part of the silane molecules is comparable with the structure of PTFE [33]. Thereby the contact angle of a water drop on a F8261 coated glass plate amounts 105°. No differences between the particle size distributions before and after silanization, via analysis by a laser diffraction device (HELOS, Sympatec, Germany), were observed. Therefore, the PSD of the hydrophilic particles were measured a second time directly in the agglomeration setup using the dynamic image analysis. Treatment with a 3.76 mmol/L sodium pyrophosphate-water solution ensures a complete dispersion of the particles, where no agglomeration occurs and the primary particle size could be determined in the experimental setup. The IEP was determined by measuring the streaming potential which leads to a pH value of 8.5 for all powders.
1
ε 2 G =⎛ ⎞ ⎝ν⎠
(2)
Agglomeration at different shear rates leads to appropriate equilibrium agglomeration sizes x. The higher the shear rate, the lower the agglomerate sizes became, because break up is promoted. This behavior can be characterized with an allometric fit, see Eq. (3) [37].
x = C⋅G −γ
(3)
The constants C and γ are product-specific and referred to as agglomerate strength constant and agglomerate strength exponent, respectively. The higher the decay of equilibrium agglomerate sizes while increasing shear stress, the less stable the agglomerates are. Consequently, the lower the agglomerate strength exponent γ of the product, the higher the resistance against shear forces. The agglomerate strength constant C depends on the measuring setup as well as the method of particle size analysis. The agglomerate strength exponent is constant for certain materials and constant for different floc sizes (mean, median, maximum value of the agglomerate size distribution) [37].
3.3. Experimental procedure In this paragraph the preparation and measuring procedure for the agglomeration tests are listed.
• At first, the tank was filled with 4 l of deionized water or the ap-
3. Materials and methods 3.1. Measuring setup The measurements have been performed in a stirred vessel with a volume of 4 l. The vessel is closed with a lid that is equipped with three baffles to create a consistent turbulence. As turbulence inducing unit, a six blade propeller with a diameter of 0.1 m was used. The power characteristics of the stirred tank were determined, so the integral energy dissipation rate relating to the variable rotational speed is known. Within the tank lid there are two holes for sampling the liquid continuously for analyses. The liquid flows into the liquid cell of the dynamic image analysis device “QICPIC” (Sympatec, Germany). To ensure a constant level of liquid in the tank, the water is pumped back into the tank using a peristaltic pump. The tank is located about 1 m above the QICPIC, because pumping the liquid directly into the measuring cell through a pump may give an unintentional stress to the agglomerates and could falsify the measured size. With a frequency of 25 Hz, images are taken from the suspended particles and the agglomerate size distribution (ASD) can be analyzed precisely using the appropriate software from Sympatec. In addition, the agglomerates can be stressed by an ultrasonic horn (Sonopuls HD 200 with VS200-TT25 horn, Bandelin, Germany), located at the top of the tank. This is used to destroy preexisting agglomerates prior the measurement. A sketch of the experimental setup is shown in Fig. 1.
• •
•
propriate aqueous electrolyte solution. The liquid was bottled 24 h prior to the experiment to ensure an equilibrium of temperature and gas saturation. Otherwise, gas bubbles can be built on the wall of the vessel and might disturb the measured images if they detach. The lid containing baffles and stirrer are positioned on the vessel to reduce the surface between liquid and air. The stirrer has been immersed into the liquid and adjusted to the appropriate rotational speed. 2 g (in the case of 2–25 μm sized powder) or 1 g respectively (in the case of 2–10 μm sized powder) of the hydrophobic particles are firstly filled into a 50 ml beaker, subsequently 15 g of ethanol are added. To ensure the sufficient dispersion of the particles this ethanol suspension has been treated in the beaker with ultrasonic stress for 30 s. For pre dispersing the hydrophilic powder instead of ethanol the appropriate electrolyte solution is used. In the next step, the suspension is filled into a syringe and added completely into the stirred tank. After this, the dynamic image analysis is started. For 2 min the built-on ultrasonic horn is engaged to destroy preexisting agglomerates, which may be formed due to funneling the suspension and furthermore to create the same initial conditions at all experiments. According to Brotchi et al. [24] nanobubbles are even highly stable at presence of ultrasonic fields, so a negative effect due to the ultrasonic treatment is not considered. Moreover, the experiments were also performed without initial treatment of ultrasound that yields to no significant differences in agglomeration kinetics.
3.4. Atomic force microscopy Imaging of alumina particles and surfaces was done by applying the atomic force microscope XE-100 from Park Systems (South Korea) with a liquid cell in intermittent mode (Asp/A = 0.88–0.90) using a Cont-Al G cantilever from budget sensors (Germany) at a scan rate of 0.35–0.45 Hz. Scan sizes were 12 × 12 μm2 in intermittent and 5 × 5 μm2 in contact mode to prove that the detected caps are nanobubbles and no residuals. Scanning in contact mode was done with the same cantilever, whereby a contact force of 20 nN was used. Spring constant calibration was done using geometrical dimensions and the determined resonant frequency of the cantilever as it is described in [41]. The hydrophobic particles were carefully placed on a rough
Fig. 1. experimental setup: elevated tank with baffles, stirrer and ultrasonic (US), grounded the dynamic image analyzer QICPIC (Sympatec).
119
Colloids and Surfaces A 530 (2017) 117–123
P. Knüpfer et al.
a function with the agglomeration time in Fig. 3 (right). Furthermore in this diagram the equilibrium agglomerate size is visible. To vary the wetting behavior of the solid particles, the same fraction of uncoated, hydrophilic alumina particles were agglomerated in an electrolyte solution of 0.5 mol/l NaCl at the IEP. By this, only attractive van der Waals forces are acting between the particles and repulsive electrochemical double layer forces are depressed. The resulting agglomerate sizes of the hydrophilic powder are lower than the size of the hydrophobic ones. In an additional experiment the hydrophobic particles were processed in the agglomeration tank without solvent exchange. To create a comparable initial condition and to overcome the dispersion problems, the particles were dispersed in deionized water with an ultrasonic horn. After this, the suspension was funneled into the agglomeration tank in the known way. The particles agglomerate with a similar process behavior as the hydrophobic ones of Fig. 3 whereby a smaller equilibrium size distribution is reached. Due to the poor wetting of the particles it is possible that entrapped gas remains in the cavities and asperities onto the rough particle. This can lead to capillary forces between particles at contact. Fig. 4 left shows the equilibrium ASD after a stirring time of 20 min in the case of hydrophilic, hydrophobic (poor wetting, without solvent exchange) and hydrophobic (with solvent exchange) particles. For all shown systems different ASD are visible. The agglomerate sizes of the hydrophobic particles are significantly larger than the hydrophilic ones, whereby the solvent exchange leads to the highest values of agglomerate sizes. This is attributable to the formation of nanobubbles onto the hydrophobic surface, which induces capillary forces between particles. Assuming that the solvent exchange method is more controllable than immersion of the hydrophobic particles via ultrasonic treatment, the stability tests of hydrophobic agglomerates were performed only for agglomerates of the solvent exchange method. Fig. 4 right shows adhesive force distributions between hydrophilic alumina and between hydrophobic alumina as well. In addition, the force distribution between the hydrophobic surfaces are shown for both: the poor wetting state and after applying a solvent exchange to increase the number of nanobubbles. The difference between the measuring conditions are clearly visible and corresponds to the different ASD in Fig. 4 left. The forces between hydrophilic surfaces are quite low compared to the hydrophobic ones, furthermore, the solvent exchange leads to the largest adhesive forces. The forces of the poor wetting state are located between hydrophilic and nanobubbles enhanced forces after solvent exchange. The reason of this behavior is that without solvent exchange the gas remains in asperities and cavities and the nucleation of nanobubbles only occurs during particle-surface contact. Thus less nanobubbles are present and not all contacts show capillary bridges, as it can be seen in Fig. 4 right and [33]. If the solvent exchange is applied, additional local temperature gradients are present, subsequently a higher gas supersaturation is produced and more
alumina substrate (root mean square parameter of roughness, rms = 0.4 μm) and deposited in a petri dish filled with ethanol. The liquid cell was then filled with pure water, which had the same temperature as the alcohol. Subsequently the sample with alumina and some ethanol was carefully moved into the liquid cell. The particles remained inside the asperities of the rough alumina substrate [33]. Additionally, force measurements on rough alumina substrates were done via AFM with the same experimental procedure as described by Fritzsche et. al [33]. Thereby 8 × 8 force-distance curves at six random sample positions were recorded between a rough alumina particle (r = 13.5 μm, rms = 0.036 μm, 15 × 15 μm2 area) and a rough alumina surface (rms = 0.6 μm, 20 × 20 μm2 area). The rms values were determined via contact mode AFM. For generating of the force distribution, the maximum adhesive forces were extracted. The solvent exchange in the liquid cell of the AFM was done similar to the described method during detection of nanobubbles. 4. Results and discussion During agglomeration tests the dynamic image analysis allows to detect structures and shapes of the agglomerates in liquid environment. The formed agglomerates might have compact, fractal or porous shapes, which vary from round to elongate. Sample images of the agglomerates made by QICPIC or SEM respectively are shown in Fig. 2. For SEM analysis, the agglomerates where sedimented and dried, whereby rearrangement of particles and shrinking during this process is assumed. Fig. 2 shows a compact dried agglomerate but the previous shape seems to have been more fractal. For all measurements the particle size is defined as the circle diameter of equal projection area. 4.1. Agglomeration kinetics An example for an agglomeration process of hydrophobic alumina powder is presented in Fig. 3, where the initial size distribution of the primary particles and the ASD at different measuring times are shown. The primary particle size x90,3 for the fraction 2–25 μm exhibits a value of 24 μm. Furthermore the size distribution at initial state of agglomeration is shown. This measurement was done within 1 min while ultrasonic treatment was activated. During this step, a part of the preexisting agglomerates are destroyed but several ones are able to stay in the system. This might be the reason of the wider PSD while ultrasonic treatment is activated. After switching off the ultrasonic treatment, the free agglomeration of the particle starts and after 15–20 min, an equilibrium state of agglomeration was set in. The ASD becomes narrow compared to the distribution of primary particle sizes. Considering the fines of the distributions, a significant decrease of the amount of particles and agglomerates with a size lower than 10 μm is visible. Additionally, the 10%, 50% and 90% quantiles of the ASD are shown in as
Fig. 2. left: sample QICPIC pictures of several hydrophobic and hydrophilic agglomerates; right: SEM image of a single dried agglomerate consisting of 2–10 μm hydrophobic alumina particles.
120
Colloids and Surfaces A 530 (2017) 117–123
P. Knüpfer et al.
Fig. 3. Agglomeration kinetics of 2 g of hydrophobic alumina powder of a size of 2–25 μm in 4 l of deionized water, added via solvent exchange, G = 186 1/s; left: size distributions of the primary particles, agglomerates at ultrasonic treatment and agglomerates after 1, 5, 10 and 20 min of free agglomeration, right: 10%, 50% and 90% quantiles of the ASD within 20 min, the error bars represent the 95% confidence interval.
Fig. 4. left: Equilibrium ASD of alumina particles (2–25 μm primary particles, black line) after 20 min of stirring for the cases: hydrophilic particles at IEP and 0.5 mol/l NaCl (grey line), hydrophobic particles immersed in DI-water due to ultrasonic treatment (dotted line) and hydrophobic particles immersed via ethanol/water exchange (dashed line), right: AFM measurement of adhesive force distribution between a colloidal probe alumina particle and a rough alumina surface with the same wetting properties and measuring conditions as at the left side of the figure, the hydrophilic data is taken from [33], all forces are normalized on the appropriate particle radius.
had observed here and in previous studies that the bubbles remain stable for hours [4]. Interestingly, the bubbles did not detach from the particle surface. That means agglomerates which are exposed to higher shear forces might be destroyed, but a re-agglomeration due to stable nanobubbles is conceivable. Bubble sizes dependent on the shape and roughness of the particle surface as well as on the amount of dissolved gas and quantity of ethanol. Sizes of nanobubbles on the particles of Fig. 5 lie between 50 and 400 nm (height) and 400–2000 nm (diameter).
nanobubbles are formed. Thus the probability that a contact point between two particles shows capillary bridges increases. 4.2. Surface nanobubble detection Topographical and phase contrast images of a hydrophobic alumina particle surface before (a,b) and after (c,d) scanning in contact mode are shown in Fig. 5. Due to the ethanol-water exchange, gas supersaturation occurs. The cap shaped hills onto the particle surface were interpreted as nanobubbles because they are moved with low contact forces, were able to coalesce and the phase shift between alumina and bubble amounts 20°. The bubbles cover the entire surface but are also sitting in asperities. With contact forces of 20 nN, we were able to move the bubbles away from smooth parts and small asperities, too. It is assumed that the bubbles stay more stable in pores or larger asperities due to pinning effects caused by a larger surface energy. Furthermore we
4.3. Agglomerate stability analysis Because of the increased adhesion forces between the particles due to the nanobubble bridging, a higher agglomerate strength for hydrophobic powder is conceivable. To proof this assumption, the agglomerates were subjected to ultrasonic and shear stress. The behavior of the 121
Colloids and Surfaces A 530 (2017) 117–123
P. Knüpfer et al.
Fig. 5. nanobubbles detected via AFM after solvent exchange; hydrophobic alumina particle with nanobubbles (a) and corresponding phase contrast image (b) and line profile (1); hydrophobic alumina particle with coalesced nanobubbles after using contact mode (parallelogram, c) with corresponding phase contrast image (d) and line profile (2).
Fig. 7. agglomerate strength test of hydrophobic (solvent exchange) and hydrophilic agglomerates at different shear conditions. Fig. 6. quantiles of ASD of hydrophobic 2–25 μm (solvent exchange) alumina powder at G = 186 1/s; hatched areas indicate ultrasonic treatment.
Table 1 allometric fit parameters of the stability test of the used materials and size fractions.
agglomerates under ultrasonic stress can be seen in Fig. 6. After disabling the initial ultrasonic treatment, the agglomerates grow up to the equilibrium size. Subsequently ultrasonic treatment is reactivated for 3 min, resulting in a total breakage of agglomerates in the system. Disabling the stress leads to a re-agglomeration up to the previous size. This behavior substantiates on the one hand the absent influence of ultrasonic to the nanobubbles, because they are strongly fixed on the hydrophobic surface. On the other hand due to the ruptured gas bridge between two particles during breakage, nanobubbles remain on both particle surfaces. The stability of hydrophobic and hydrophilic agglomerates can be taken from Fig. 7. The agglomeration process was performed at different shear rates and agitation speeds, respectively. The quantiles of the ASD decrease at higher shear rates, because the equilibrium between agglomeration and breakage is moved toward breakage. For each
R2
alumina powder
hydrophobic 2–10 μm hydrophobic 2–25 μm hydrophilic 2–10 μm
log C
γ
2.24
0.10
0.539
2.37
0.26
0.813
2.47
0.44
0.973
material the test was done and eq. 3 was fitted to the data points. The agglomerate strength exponents γi and the agglomerate strength constants Ci for the appropriate material are shown in Table 1. Although the coefficients of determination R2 of the fit of hydrophobic material during this test are partially quite low, the influence of 122
Colloids and Surfaces A 530 (2017) 117–123
P. Knüpfer et al.
284 (2005) 548–559. [9] D.N. Thomas, S.J. Judd, N. Fawcett, Flocculation modelling: a review, Water Res. 33 (1999) 1579–1592. [10] B. Derjaguin, L. Landau, Theory of the stability of strongly charged lyophobic sols and of the adhesion of strongly charged particles in solutions of electrolytes, Prog. Surf. Sci. 43 (1993) 30–59. [11] E.J.W. Verwey, J.T.G. Overbeek, K. van Nes, Theory of the stability of lyophobic colloids, Dover Mineola, 1999. [12] H.K. Christenson, P.M. Claesson, Cavitation and the interaction between macroscopic hydrophobic surfaces, Science (New York, N.Y.) 239 (1988) 390–392. [13] H.-J. Butt, M. Kappl, Surface and Interfacial Forces, Wiley-VCH, Weinheim, 2010. [14] J.L. Parker, P.M. Claesson, P. Attard, Bubbles cavities, and the long-ranged attraction between hydrophobic surfaces, J. Phys. Chem. 98 (1994) 8468–8480. [15] M.A. Hampton, A.V. Nguyen, Nanobubbles and the nanobubble bridging capillary force, Adv. Colloid Interface Sci. 154 (2010) 30–55. [16] E.E. Meyer, K.J. Rosenberg, J. Israelachvili, Recent progress in understanding hydrophobic interactions, Proc. Natl. Acad. Sci. U. S. A. 103 (2006) 15739–15746. [17] X.H. Zhang, N. Maeda, V.S.J. Craig, Physical properties of nanobubbles on hydrophobic surfaces in water and aqueous solutions, Langmuir ACS J. Surf. Colloids 22 (2006) 5025–5035. [18] D. Lohse, X. Zhang, Surface nanobubbles and nanodroplets, Rev. Mod. Phys. 87 (2015) 981–1035. [19] L. Zhang, Y. Zhang, X. Zhang, Z. Li, G. Shen, M. Ye, C. Fan, H. Fang, J. Hu, Electrochemically controlled formation and growth of hydrogen nanobubbles, Langmuir 22 (2006) 8109–8113. [20] S. Yang, P. Tsai, E.S. Kooij, A. Prosperetti, H.J.W. Zandvliet, D. Lohse, Electrolytically generated nanobubbles on highly orientated pyrolytic graphite surfaces, Langmuir 25 (2009) 1466–1474. [21] M.A. Hampton, B.C. Donose, A.V. Nguyen, Effect of alcohol–water exchange and surface scanning on nanobubbles and the attraction between hydrophobic surfaces, J. Colloid Interface Sci. 325 (2008) 267–274. [22] X.H. Zhang, A. Quinn, W.A. Ducker, Nanobubbles at the interface between water and a hydrophobic solid, Langmuir 24 (2008) 4756–4764. [23] B. Zhao, Y. Song, S. Wang, B. Dai, L. Zhang, Y. Dong, J. Lü, J. Hu, Mechanical mapping of nanobubbles by PeakForce atomic force microscopy, Soft Matter 9 (2013) 8837. [24] A. Brotchie, X.H. Zhang, Response of interfacial nanobubbles to ultrasound irradiation, Soft Matter 7 (2011) 265–269. [25] S.-T. Lou, Z.-Q. Ouyang, Y. Zhang, X.-J. Li, J. Hu, M.-Q. Li, F.-J. Yang, Nanobubbles on solid surface imaged by atomic force microscopy, J. Vac. Sci. Technol. B 18 (2000) 2573. [26] N. Ishida, T. Inoue, M. Miyahara, K. Higashitani, Nano bubbles on a hydrophobic surface in water observed by tapping-mode atomic force microscopy, Langmuir 16 (2000) 6377–6380. [27] X.H. Zhang, A. Khan, W.A. Ducker, A nanoscale gas state, Phys. Rev. Lett. 98 (2007) 136101. [28] J. Weijs, D. Lohse, Why surface nanobubbles live for hours, Phys. Rev. Lett. 110 (2013). [29] X. Zhang, DerekY.C. Chan, D. Wang, N. Maeda, Stability of interfacial nanobubbles, Langmuir 29 (2013) 1017–1023. [30] Y. Liu, J. Wang, X. Zhang, W. Wang, Contact line pinning and the relationship between nanobubbles and substrates, J. Chem. Phys. 140 (2014) 54705. [31] M.P. Brenner, D. Lohse, Dynamic equilibrium mechanism for surface nanobubble stabilization, Phys. Rev. Lett. 101 (2008). [32] L. Ditscherlein, U.A. Peuker, Note Production of stable colloidal probes for hightemperature atomic force microscopy applications, The Review of scientific instruments 88 (2017) 46107. [33] J. Fritzsche, U.A. Peuker, Modeling adhesive force distributions on highly rough surfaces, Powder Technol. 289 (2016) 88–94. [34] D.P. Patil, J. Andrews, P. Uhlherr, Shear flocculation—kinetics of floc coalescence and breakage, Int. J. Miner. Process. 61 (2001) 171–188. [35] W. Yu, J. Gregory, L.C. Campos, Breakage and re-growth of flocs formed by charge neutralization using alum and polyDADMAC, Water Res. 44 (2010) 3959–3965. [36] M.A. Yukselen, J. Gregory, The reversibility of floc breakage, Int. J. Miner. Process. 73 (2004) 251–259. [37] P. Jarvis, B. Jefferson, J. Gregory, S.A. Parsons, A review of floc strength and breakage, Water Res. 39 (2005) 3121–3137. [38] A.V. Nguyen, D.-A. An-Vo, T. Tran-Cong, G.M. Evans, A review of stochastic description of the turbulence effect on bubble-particle interactions in flotation, Int. J. Miner. Process. 156 (2016) 75–86. [39] T.R. Camp, P.C. Stein, Velocity gradients and internal work in fluid motion, J. Boston Soc. Civil Eng. 85 (1943) 219–237. [40] P. Knüpfer, J. Fritzsche, T. Leistner, M. Rudolph, U.A. Peuker, Investigating the removal of particles from the air/water-interface −Modelling detachment forces using an energetic approach, Coll. Surf. A 513 (2017) 215–222. [41] H.-J. Butt, B. Cappella, M. Kappl, Force measurements with the atomic force microscope: technique, interpretation and applications, Surf. Sci. Rep. 59 (2005) 1–152.
wetting is clearly visible. The hydrophobic powder creates much larger agglomerates than the hydrophilic one even at higher shear rates. Furthermore, the stability of agglomerates which is quantized by the agglomerate strength exponent γ, is higher for the hydrophobic material. Considering the size of primary particles, for the smaller fraction of hydrophobic powder a higher agglomerate strength and in addition larger agglomerate sizes were detected. This might be attributed to the increased number of particles per agglomerate and thus, more individual bondings between the primary particles. 5. Conclusion In this study, the agglomeration of hydrophobic and hydrophilic alumina particles in aqueous solutions has been compared regarding to the equilibrium agglomeration size and the stability of agglomerates. It could be shown, that hydrophobic particles merge to larger agglomerates, which are additionally more stable than hydrophilic ones. This behavior is attributed to the nucleation of surface nanobubbles onto the hydrophobic particles due to the required solvent exchange. Thus, capillary forces between the particles are developed, that are larger in magnitude than van der Waals forces, acting between hydrophilic particles, respectively. This behavior was verified through adhesive force measurements via AFM between a single colloidal probe particle and the appropriate alumina surface. Additionally, the occurring nanobubbles were detected via AFM on representative hydrophobic alumina surfaces. Furthermore, the agglomerates were treated via ultrasonic and shear stress. Although ultrasonic stress leads to a total breakage of the agglomerates, the agglomeration is reversible. Different shear conditions lead to certain equilibrium sizes, from which the stability of agglomerates for every used material can be estimated. In general, the stability of hydrophobic agglomerates in aqueous flows is larger than the stability of hydrophilic ones. Due to the high surface energies of metal melts, similar effects like increased adhesion forces through their non-wetting behavior are conceivable in the real filtration process of metal melts. Acknowledgement The authors gratefully thank the German Research Foundation (DFG) for supporting the Collaborative Research Center CRC 920, subprojects B01 and B04. References [1] F. Delannay, L. Froyen, A. Deruyttere, The wetting of solids by molten metals and its relation to the preparation of metal-matrix composites composites, J. Mater. Sci. 22 (1987) 1–16. [2] O.I. Vinogradova, G.E. Yakubov, H.-J. Butt, Forces between polystyrene surfaces in water–electrolyte solutions: long-range attraction of two types? J. Chem. Phys. 114 (2001) 8124–8131. [3] J. Fritzsche, U.A. Peuker, Modeling adhesive forces caused by nanobubble capillary bridging, Coll. Surf. A 509 (2016) 457–466. [4] L. Ditscherlein, J. Fritzsche, U.A. Peuker, Study of nanobubbles on hydrophilic and hydrophobic alumina surfaces, Colloids Surf. A 497 (2016) 242–250. [5] H. Saint-Raymond, et al., Turbulent aggregation of alumina in water and nHeptane, J. Colloid Interface Sci. 202 (2) (1998) 238–250. [6] M. Cournil, F. Gruy, P. Gardin, H. Saint-Raymond, Modelling of solid particle aggregation dynamics in non-wetting liquid medium, Chem. Eng. Process. Process Intensif. 45 (2006) 586–597. [7] M. Cournil, F. Gruy, P. Gardin, H. Saint-Raymond, Experimental study and modeling of inclusion aggregation in turbulent flows to improve steel cleanliness, phys. stat. sol. (a) 189 (2002) 159–168. [8] F. Gruy, M. Cournil, P. Cugniet, Influence of nonwetting on the aggregation dynamics of micronic solid particles in a turbulent medium, J. Colloid Interface Sci.
123
Contents lists available at ScienceDirect
Colloids and Surfaces A journal homepage: www.elsevier.com/locate/colsurfa
Nanobubble enhanced agglomeration of hydrophobic powders ⁎
MARK
Paul Knüpfer , Lisa Ditscherlein, Urs Alexander Peuker Institute of Mechanical Process Engineering and Mineral Processing, Agricolastraße 1, 09599 Freiberg, Germany
G RA P H I C A L AB S T R A C T
A R T I C L E I N F O
A B S T R A C T
Keywords: Agglomerate stability Agglomeration Hydrophobic particles Nanobubbles
Agglomeration of solid particles in aqueous liquid flows is generally influenced by the appropriate interacting forces. Between hydrophobic particles, these forces are dramatically increased compared to van der Waals forces due to the formation of capillary bridges, caused by nanobubbles located at the particle surface. In this study, both hydrophobic and hydrophilic agglomerates were generated in water with an agglomeration setup based on a stirred tank reactor. Variating the shear rate leads to different agglomeration sizes and provides information about the agglomerate strength. In principle, both stability and size of the hydrophobic agglomerates are larger than the hydrophilic ones. The presence of the nanobubbles is proved by applying atomic force microscopy to the particle surface.
1. Introduction Agglomeration of fine particles plays a role in different industries. One major representative of hydrophobic (non-wetting) agglomeration is metallurgical processing, especially liquid metal cleaning by filtration and flotation. Non-metallic inclusions in metal melts are undesired for a high stability of the cast products. To increase the fatigue strength of these products, a filtration is applied to the metal melt during the casting process. Therefore porous ceramic foam filters, consisting of
alumina, are used to separate the melt from inclusions. A depth filtration mechanism is assumed, because the pore diameter, where the inclusions enter the filter, are much larger than the inclusions themselves. If the inclusion attaches to the inner filter surface, the adhesion is provided due to van der Waals interactions, sintering and particularly capillary forces. Neither the inclusions nor the ceramic foam filter are wetted by the metal melt due to its high surface energy, which exceeds the surface energy of water about several magnitudes [1]. Gas reservoirs on the particulate inclusions as well as the filter surface are
Abbreviations: AFM, atomic force microscopy; ASD, agglomerate size distribution; IEP, isoelectric point; PSD, particle size distributions; rms, root mean square ⁎ Corresponding author. E-mail addresses: [email protected] (P. Knüpfer), [email protected] (L. Ditscherlein), [email protected] (U.A. Peuker). http://dx.doi.org/10.1016/j.colsurfa.2017.07.056 Received 6 June 2017; Received in revised form 17 July 2017; Accepted 18 July 2017 Available online 19 July 2017 0927-7757/ © 2017 Elsevier B.V. All rights reserved.
Colloids and Surfaces A 530 (2017) 117–123
P. Knüpfer et al.
a range of a few nm while decreasing exponentially [13]. The most common explanation of the long range forces is due to the formation of nanobubbles onto the hydrophobic surface, which are leading to strong capillary interactions [14–16]. The behavior and the generation of nanobubbles at smooth surfaces is summarized already in literature [17,18]. Nanobubbles on solid surfaces can be produced via gas supersaturation in liquids that are achieved by electrolysis [19,20], perturbation [2] or heating the substrate or the liquid [4]. A widely used procedure to generate nanobubbles onto hydrophobic surfaces is the solvent exchange method [21–24]. This procedure was developed by Lou et al. [25] and Ishida et al. [26] to generate nanobubbles in a reproducible way. Here, the hydrophobic surface is firstly immersed in a wetting liquid, which is a good gas solvent. The next step is to exchange the solvent, e.g. ethanol, with a second poorly wetting liquid (e.g. water) which contains less dissolved air. In addition, the enthalpy of the ethanol-water-mixing causes local temperature gradients. Thus, a supersaturation of gas occurs in the mixing zone during the exchange. Nanobubbles are forming spontaneously at the surface and grow to a height of several hundred nm. Another important property of nanobubbles is their high stability. Actual, the bubbles should dissolve in microseconds, but they stay stable even for hours and days [22,27,28]. Although it is not totally clear, the most common explanations for the high stability is due to line pinning [29,30] and a dynamic equilibrium of gas flux at the liquid-gas interface [31]. The detection of nanobubbles is commonly possible via atomic force microscopy (AFM), whereby a soft cantilever is used at intermittent mode [18]. Furthermore, due to the colloidal probe technique, it is possible to measure adhesive forces between a single particle and a surface via AFM. Therefore, a spherical particle is glued onto a tipless cantilever [32] and after recording force-distance curves between this particle and a surface, the maximum adhesive force can be determined. Because real surfaces are rough, it is more practical to quantify the detachment forces by an adhesive force distribution, as described in [33]. The relative mobility of the particles is strongly influenced by the micro turbulence of a liquid movement. It is necessary to transport the particles to each other, but can also lead to break up induced by larger shear forces [34]. If these two processes are equilibrated, the size of agglomerates is reaching a constant value after a certain time t. Additional stress to the agglomerates could lead to a breakage and disabling this stress provides a regrowth ending up to the previous size of agglomerates [35,36]. The stability of agglomerates towards shear stress can be determined by observation of the size of the agglomerates at different shear conditions [37]. According to the theory of turbulence, the eddy size is distributed. Starting from the macroscopic eddies the dissipation of energy finally occurs in the smallest micro eddies. While stirring the liquid in the vessel, the energy is dissipating at a rate of the integral energy dissipationε . The latter can be evaluated from the difference in energy consumption between stirring with and without the liquid relative to its mass. By this, the character of turbulence can be estimated. The microscale η of Kolmogorov (Eq. (1)) gives the length scale of the smallest eddies, where the Reynolds number Re equals 1 [38].
caused by dissolved gas in the metal melt or during immersion. The reservoirs can also develop at the moment of particle-surface contact [2]. A capillary is formed and leads to a strong binding force of the inclusions resulting in the deposition within the foam filter. As it is known from the theory of depth filtration, inclusion particles are able to cross the filter without establishing contact to the filter surface, thus a pre-agglomeration of these particles is necessary to increase the contact probability and thus to purify the metal melt in the most efficient way. Metal melts are opaque and solely treatable at high temperatures, that impedes investigations of inclusion agglomeration at the real system. Therefore, a room temperature model system is used, that mimics the poor wetting behavior of metal melts on nonmetallic inclusions and filter materials. In this model system alumina particles with a hydrophobic coating represent the inclusions while water is used as a non-wetting liquid. In the study, the agglomeration of hydrophobic powders is realized in a stirred vessel. To stimulate and control the hydrophobic agglomeration a novel approach is applied, which introduces defined nanobubbles in the system, to the particle surface respectively to increase the adhesion force between the particles dramatically due to capillary bridging [3,4]. In the recent past, similar investigations in this room temperature model system were done to understand and to improve the cleaning of metal melts. Saint-Raymond et al. investigated the agglomeration of fine alumina powders at different pH-values in water and in n-heptane [5]. Cournil et al. observed a higher agglomeration rate of hydrophobic particles than hydrophilic ones both theoretically [6] and experimentally [7]. Therefore, in situ measurements of turbidity in a stirred vessel at different agitation speeds were done [8]. Gruy et al. assumed the preexistence of nanobubbles on the particles and accomplished the agglomeration experiments at the pH-value of isoelectric point (IEP) and with dispersing the particles in a 15 vol.% ethanol-water mixture before the agglomeration starts [8]. This study has the aim to nucleate defined nanobubbles onto hydrophobic particles through a solvent exchange. A subsequent agglomeration of these particles in a non-wetting liquid is done. The emerged nanobubbles can be detected and qualified via AFM on comparable hydrophobic surfaces at corresponding physicochemical conditions. The hydrophobic agglomerates are compared with hydrophilic ones in size and stability, which were obtained in the same setup at the IEP. Also, high electrolyte concentrations were produced to reveal the enhancement of adhesion forces based on poor wetting property of the system. 2. Theoretical approach The agglomeration of hydrophobic powders can be described with established techniques of flocculation, as used e.g. in waste water treatment. Agglomeration occurs if primary particles are both transported and attached to each other [9]. The transport requires a relative displacement (e.g. sedimentation, flow field, Brownian motion) and the attachment an attractive force between the particles. Two components of interaction forces between particles immersed in polar liquids like water are known according to the DLVO-theory, which was established in the 1940s by Derjaguin, Landau [10], Verwey and Overbeek [11]. The van der Waals force represents the attractive component. This force is based on molecular dipoles of the different phases and scales with the distance to the negative power of three in the case of two colliding particles. The repulsive interactions is formed through the presence of the electrochemical double layer and decreases approximately exponentially relating to the distance. A further force occurs between hydrophobic surfaces in aqueous solutions, which cannot be explained by the traditional DLVO theory [12]. This attractive hydrophobic force is commonly divided into two parts as well, the so called short range and long range hydrophobic force. The short range force is commonly described as structural formation of the water molecules at the solid-liquid interface and exhibits
3
ν η=⎛ ⎞ ⎝ε ⎠ ⎜
1 4
⎟
(1)
Agglomerates with a size below the Kolmogorov scale are stressed by the dominating viscous shear forces and are consequently submitted to erosion. In this case, small parts in the size range of the primary particles detach. If the agglomerate size is significantly above the Kolmogorov scale, large scale fragmentation occurs due to dynamic pressures and the agglomerates decompose into several smaller agglomerates [37]. Camp and Stein [39] defined the global root mean shear rates G in a turbulent flow, which is related to the energy dissipation rate as well as 118
Colloids and Surfaces A 530 (2017) 117–123
P. Knüpfer et al.
the kinematic viscosity of the fluid, see Eq. (2). Thereby, this value is directly connected to the rotational speed of the propeller.
3.2. Particles and characterization For the measurements α−alumina particles T60/T64 (Almatis, Germany) classified to a size of 2-25 μm, CT9FG (Almatis, Germany) sized to 2–10 μm, respectively, are used. To hydrophobize the surface, the particles were coated with Tridecafluorotriethoxysilane (Dynasilan F8261 from Evonik, Germany) as described in [40]. The hydrophobic part of the silane molecules is comparable with the structure of PTFE [33]. Thereby the contact angle of a water drop on a F8261 coated glass plate amounts 105°. No differences between the particle size distributions before and after silanization, via analysis by a laser diffraction device (HELOS, Sympatec, Germany), were observed. Therefore, the PSD of the hydrophilic particles were measured a second time directly in the agglomeration setup using the dynamic image analysis. Treatment with a 3.76 mmol/L sodium pyrophosphate-water solution ensures a complete dispersion of the particles, where no agglomeration occurs and the primary particle size could be determined in the experimental setup. The IEP was determined by measuring the streaming potential which leads to a pH value of 8.5 for all powders.
1
ε 2 G =⎛ ⎞ ⎝ν⎠
(2)
Agglomeration at different shear rates leads to appropriate equilibrium agglomeration sizes x. The higher the shear rate, the lower the agglomerate sizes became, because break up is promoted. This behavior can be characterized with an allometric fit, see Eq. (3) [37].
x = C⋅G −γ
(3)
The constants C and γ are product-specific and referred to as agglomerate strength constant and agglomerate strength exponent, respectively. The higher the decay of equilibrium agglomerate sizes while increasing shear stress, the less stable the agglomerates are. Consequently, the lower the agglomerate strength exponent γ of the product, the higher the resistance against shear forces. The agglomerate strength constant C depends on the measuring setup as well as the method of particle size analysis. The agglomerate strength exponent is constant for certain materials and constant for different floc sizes (mean, median, maximum value of the agglomerate size distribution) [37].
3.3. Experimental procedure In this paragraph the preparation and measuring procedure for the agglomeration tests are listed.
• At first, the tank was filled with 4 l of deionized water or the ap-
3. Materials and methods 3.1. Measuring setup The measurements have been performed in a stirred vessel with a volume of 4 l. The vessel is closed with a lid that is equipped with three baffles to create a consistent turbulence. As turbulence inducing unit, a six blade propeller with a diameter of 0.1 m was used. The power characteristics of the stirred tank were determined, so the integral energy dissipation rate relating to the variable rotational speed is known. Within the tank lid there are two holes for sampling the liquid continuously for analyses. The liquid flows into the liquid cell of the dynamic image analysis device “QICPIC” (Sympatec, Germany). To ensure a constant level of liquid in the tank, the water is pumped back into the tank using a peristaltic pump. The tank is located about 1 m above the QICPIC, because pumping the liquid directly into the measuring cell through a pump may give an unintentional stress to the agglomerates and could falsify the measured size. With a frequency of 25 Hz, images are taken from the suspended particles and the agglomerate size distribution (ASD) can be analyzed precisely using the appropriate software from Sympatec. In addition, the agglomerates can be stressed by an ultrasonic horn (Sonopuls HD 200 with VS200-TT25 horn, Bandelin, Germany), located at the top of the tank. This is used to destroy preexisting agglomerates prior the measurement. A sketch of the experimental setup is shown in Fig. 1.
• •
•
propriate aqueous electrolyte solution. The liquid was bottled 24 h prior to the experiment to ensure an equilibrium of temperature and gas saturation. Otherwise, gas bubbles can be built on the wall of the vessel and might disturb the measured images if they detach. The lid containing baffles and stirrer are positioned on the vessel to reduce the surface between liquid and air. The stirrer has been immersed into the liquid and adjusted to the appropriate rotational speed. 2 g (in the case of 2–25 μm sized powder) or 1 g respectively (in the case of 2–10 μm sized powder) of the hydrophobic particles are firstly filled into a 50 ml beaker, subsequently 15 g of ethanol are added. To ensure the sufficient dispersion of the particles this ethanol suspension has been treated in the beaker with ultrasonic stress for 30 s. For pre dispersing the hydrophilic powder instead of ethanol the appropriate electrolyte solution is used. In the next step, the suspension is filled into a syringe and added completely into the stirred tank. After this, the dynamic image analysis is started. For 2 min the built-on ultrasonic horn is engaged to destroy preexisting agglomerates, which may be formed due to funneling the suspension and furthermore to create the same initial conditions at all experiments. According to Brotchi et al. [24] nanobubbles are even highly stable at presence of ultrasonic fields, so a negative effect due to the ultrasonic treatment is not considered. Moreover, the experiments were also performed without initial treatment of ultrasound that yields to no significant differences in agglomeration kinetics.
3.4. Atomic force microscopy Imaging of alumina particles and surfaces was done by applying the atomic force microscope XE-100 from Park Systems (South Korea) with a liquid cell in intermittent mode (Asp/A = 0.88–0.90) using a Cont-Al G cantilever from budget sensors (Germany) at a scan rate of 0.35–0.45 Hz. Scan sizes were 12 × 12 μm2 in intermittent and 5 × 5 μm2 in contact mode to prove that the detected caps are nanobubbles and no residuals. Scanning in contact mode was done with the same cantilever, whereby a contact force of 20 nN was used. Spring constant calibration was done using geometrical dimensions and the determined resonant frequency of the cantilever as it is described in [41]. The hydrophobic particles were carefully placed on a rough
Fig. 1. experimental setup: elevated tank with baffles, stirrer and ultrasonic (US), grounded the dynamic image analyzer QICPIC (Sympatec).
119
Colloids and Surfaces A 530 (2017) 117–123
P. Knüpfer et al.
a function with the agglomeration time in Fig. 3 (right). Furthermore in this diagram the equilibrium agglomerate size is visible. To vary the wetting behavior of the solid particles, the same fraction of uncoated, hydrophilic alumina particles were agglomerated in an electrolyte solution of 0.5 mol/l NaCl at the IEP. By this, only attractive van der Waals forces are acting between the particles and repulsive electrochemical double layer forces are depressed. The resulting agglomerate sizes of the hydrophilic powder are lower than the size of the hydrophobic ones. In an additional experiment the hydrophobic particles were processed in the agglomeration tank without solvent exchange. To create a comparable initial condition and to overcome the dispersion problems, the particles were dispersed in deionized water with an ultrasonic horn. After this, the suspension was funneled into the agglomeration tank in the known way. The particles agglomerate with a similar process behavior as the hydrophobic ones of Fig. 3 whereby a smaller equilibrium size distribution is reached. Due to the poor wetting of the particles it is possible that entrapped gas remains in the cavities and asperities onto the rough particle. This can lead to capillary forces between particles at contact. Fig. 4 left shows the equilibrium ASD after a stirring time of 20 min in the case of hydrophilic, hydrophobic (poor wetting, without solvent exchange) and hydrophobic (with solvent exchange) particles. For all shown systems different ASD are visible. The agglomerate sizes of the hydrophobic particles are significantly larger than the hydrophilic ones, whereby the solvent exchange leads to the highest values of agglomerate sizes. This is attributable to the formation of nanobubbles onto the hydrophobic surface, which induces capillary forces between particles. Assuming that the solvent exchange method is more controllable than immersion of the hydrophobic particles via ultrasonic treatment, the stability tests of hydrophobic agglomerates were performed only for agglomerates of the solvent exchange method. Fig. 4 right shows adhesive force distributions between hydrophilic alumina and between hydrophobic alumina as well. In addition, the force distribution between the hydrophobic surfaces are shown for both: the poor wetting state and after applying a solvent exchange to increase the number of nanobubbles. The difference between the measuring conditions are clearly visible and corresponds to the different ASD in Fig. 4 left. The forces between hydrophilic surfaces are quite low compared to the hydrophobic ones, furthermore, the solvent exchange leads to the largest adhesive forces. The forces of the poor wetting state are located between hydrophilic and nanobubbles enhanced forces after solvent exchange. The reason of this behavior is that without solvent exchange the gas remains in asperities and cavities and the nucleation of nanobubbles only occurs during particle-surface contact. Thus less nanobubbles are present and not all contacts show capillary bridges, as it can be seen in Fig. 4 right and [33]. If the solvent exchange is applied, additional local temperature gradients are present, subsequently a higher gas supersaturation is produced and more
alumina substrate (root mean square parameter of roughness, rms = 0.4 μm) and deposited in a petri dish filled with ethanol. The liquid cell was then filled with pure water, which had the same temperature as the alcohol. Subsequently the sample with alumina and some ethanol was carefully moved into the liquid cell. The particles remained inside the asperities of the rough alumina substrate [33]. Additionally, force measurements on rough alumina substrates were done via AFM with the same experimental procedure as described by Fritzsche et. al [33]. Thereby 8 × 8 force-distance curves at six random sample positions were recorded between a rough alumina particle (r = 13.5 μm, rms = 0.036 μm, 15 × 15 μm2 area) and a rough alumina surface (rms = 0.6 μm, 20 × 20 μm2 area). The rms values were determined via contact mode AFM. For generating of the force distribution, the maximum adhesive forces were extracted. The solvent exchange in the liquid cell of the AFM was done similar to the described method during detection of nanobubbles. 4. Results and discussion During agglomeration tests the dynamic image analysis allows to detect structures and shapes of the agglomerates in liquid environment. The formed agglomerates might have compact, fractal or porous shapes, which vary from round to elongate. Sample images of the agglomerates made by QICPIC or SEM respectively are shown in Fig. 2. For SEM analysis, the agglomerates where sedimented and dried, whereby rearrangement of particles and shrinking during this process is assumed. Fig. 2 shows a compact dried agglomerate but the previous shape seems to have been more fractal. For all measurements the particle size is defined as the circle diameter of equal projection area. 4.1. Agglomeration kinetics An example for an agglomeration process of hydrophobic alumina powder is presented in Fig. 3, where the initial size distribution of the primary particles and the ASD at different measuring times are shown. The primary particle size x90,3 for the fraction 2–25 μm exhibits a value of 24 μm. Furthermore the size distribution at initial state of agglomeration is shown. This measurement was done within 1 min while ultrasonic treatment was activated. During this step, a part of the preexisting agglomerates are destroyed but several ones are able to stay in the system. This might be the reason of the wider PSD while ultrasonic treatment is activated. After switching off the ultrasonic treatment, the free agglomeration of the particle starts and after 15–20 min, an equilibrium state of agglomeration was set in. The ASD becomes narrow compared to the distribution of primary particle sizes. Considering the fines of the distributions, a significant decrease of the amount of particles and agglomerates with a size lower than 10 μm is visible. Additionally, the 10%, 50% and 90% quantiles of the ASD are shown in as
Fig. 2. left: sample QICPIC pictures of several hydrophobic and hydrophilic agglomerates; right: SEM image of a single dried agglomerate consisting of 2–10 μm hydrophobic alumina particles.
120
Colloids and Surfaces A 530 (2017) 117–123
P. Knüpfer et al.
Fig. 3. Agglomeration kinetics of 2 g of hydrophobic alumina powder of a size of 2–25 μm in 4 l of deionized water, added via solvent exchange, G = 186 1/s; left: size distributions of the primary particles, agglomerates at ultrasonic treatment and agglomerates after 1, 5, 10 and 20 min of free agglomeration, right: 10%, 50% and 90% quantiles of the ASD within 20 min, the error bars represent the 95% confidence interval.
Fig. 4. left: Equilibrium ASD of alumina particles (2–25 μm primary particles, black line) after 20 min of stirring for the cases: hydrophilic particles at IEP and 0.5 mol/l NaCl (grey line), hydrophobic particles immersed in DI-water due to ultrasonic treatment (dotted line) and hydrophobic particles immersed via ethanol/water exchange (dashed line), right: AFM measurement of adhesive force distribution between a colloidal probe alumina particle and a rough alumina surface with the same wetting properties and measuring conditions as at the left side of the figure, the hydrophilic data is taken from [33], all forces are normalized on the appropriate particle radius.
had observed here and in previous studies that the bubbles remain stable for hours [4]. Interestingly, the bubbles did not detach from the particle surface. That means agglomerates which are exposed to higher shear forces might be destroyed, but a re-agglomeration due to stable nanobubbles is conceivable. Bubble sizes dependent on the shape and roughness of the particle surface as well as on the amount of dissolved gas and quantity of ethanol. Sizes of nanobubbles on the particles of Fig. 5 lie between 50 and 400 nm (height) and 400–2000 nm (diameter).
nanobubbles are formed. Thus the probability that a contact point between two particles shows capillary bridges increases. 4.2. Surface nanobubble detection Topographical and phase contrast images of a hydrophobic alumina particle surface before (a,b) and after (c,d) scanning in contact mode are shown in Fig. 5. Due to the ethanol-water exchange, gas supersaturation occurs. The cap shaped hills onto the particle surface were interpreted as nanobubbles because they are moved with low contact forces, were able to coalesce and the phase shift between alumina and bubble amounts 20°. The bubbles cover the entire surface but are also sitting in asperities. With contact forces of 20 nN, we were able to move the bubbles away from smooth parts and small asperities, too. It is assumed that the bubbles stay more stable in pores or larger asperities due to pinning effects caused by a larger surface energy. Furthermore we
4.3. Agglomerate stability analysis Because of the increased adhesion forces between the particles due to the nanobubble bridging, a higher agglomerate strength for hydrophobic powder is conceivable. To proof this assumption, the agglomerates were subjected to ultrasonic and shear stress. The behavior of the 121
Colloids and Surfaces A 530 (2017) 117–123
P. Knüpfer et al.
Fig. 5. nanobubbles detected via AFM after solvent exchange; hydrophobic alumina particle with nanobubbles (a) and corresponding phase contrast image (b) and line profile (1); hydrophobic alumina particle with coalesced nanobubbles after using contact mode (parallelogram, c) with corresponding phase contrast image (d) and line profile (2).
Fig. 7. agglomerate strength test of hydrophobic (solvent exchange) and hydrophilic agglomerates at different shear conditions. Fig. 6. quantiles of ASD of hydrophobic 2–25 μm (solvent exchange) alumina powder at G = 186 1/s; hatched areas indicate ultrasonic treatment.
Table 1 allometric fit parameters of the stability test of the used materials and size fractions.
agglomerates under ultrasonic stress can be seen in Fig. 6. After disabling the initial ultrasonic treatment, the agglomerates grow up to the equilibrium size. Subsequently ultrasonic treatment is reactivated for 3 min, resulting in a total breakage of agglomerates in the system. Disabling the stress leads to a re-agglomeration up to the previous size. This behavior substantiates on the one hand the absent influence of ultrasonic to the nanobubbles, because they are strongly fixed on the hydrophobic surface. On the other hand due to the ruptured gas bridge between two particles during breakage, nanobubbles remain on both particle surfaces. The stability of hydrophobic and hydrophilic agglomerates can be taken from Fig. 7. The agglomeration process was performed at different shear rates and agitation speeds, respectively. The quantiles of the ASD decrease at higher shear rates, because the equilibrium between agglomeration and breakage is moved toward breakage. For each
R2
alumina powder
hydrophobic 2–10 μm hydrophobic 2–25 μm hydrophilic 2–10 μm
log C
γ
2.24
0.10
0.539
2.37
0.26
0.813
2.47
0.44
0.973
material the test was done and eq. 3 was fitted to the data points. The agglomerate strength exponents γi and the agglomerate strength constants Ci for the appropriate material are shown in Table 1. Although the coefficients of determination R2 of the fit of hydrophobic material during this test are partially quite low, the influence of 122
Colloids and Surfaces A 530 (2017) 117–123
P. Knüpfer et al.
284 (2005) 548–559. [9] D.N. Thomas, S.J. Judd, N. Fawcett, Flocculation modelling: a review, Water Res. 33 (1999) 1579–1592. [10] B. Derjaguin, L. Landau, Theory of the stability of strongly charged lyophobic sols and of the adhesion of strongly charged particles in solutions of electrolytes, Prog. Surf. Sci. 43 (1993) 30–59. [11] E.J.W. Verwey, J.T.G. Overbeek, K. van Nes, Theory of the stability of lyophobic colloids, Dover Mineola, 1999. [12] H.K. Christenson, P.M. Claesson, Cavitation and the interaction between macroscopic hydrophobic surfaces, Science (New York, N.Y.) 239 (1988) 390–392. [13] H.-J. Butt, M. Kappl, Surface and Interfacial Forces, Wiley-VCH, Weinheim, 2010. [14] J.L. Parker, P.M. Claesson, P. Attard, Bubbles cavities, and the long-ranged attraction between hydrophobic surfaces, J. Phys. Chem. 98 (1994) 8468–8480. [15] M.A. Hampton, A.V. Nguyen, Nanobubbles and the nanobubble bridging capillary force, Adv. Colloid Interface Sci. 154 (2010) 30–55. [16] E.E. Meyer, K.J. Rosenberg, J. Israelachvili, Recent progress in understanding hydrophobic interactions, Proc. Natl. Acad. Sci. U. S. A. 103 (2006) 15739–15746. [17] X.H. Zhang, N. Maeda, V.S.J. Craig, Physical properties of nanobubbles on hydrophobic surfaces in water and aqueous solutions, Langmuir ACS J. Surf. Colloids 22 (2006) 5025–5035. [18] D. Lohse, X. Zhang, Surface nanobubbles and nanodroplets, Rev. Mod. Phys. 87 (2015) 981–1035. [19] L. Zhang, Y. Zhang, X. Zhang, Z. Li, G. Shen, M. Ye, C. Fan, H. Fang, J. Hu, Electrochemically controlled formation and growth of hydrogen nanobubbles, Langmuir 22 (2006) 8109–8113. [20] S. Yang, P. Tsai, E.S. Kooij, A. Prosperetti, H.J.W. Zandvliet, D. Lohse, Electrolytically generated nanobubbles on highly orientated pyrolytic graphite surfaces, Langmuir 25 (2009) 1466–1474. [21] M.A. Hampton, B.C. Donose, A.V. Nguyen, Effect of alcohol–water exchange and surface scanning on nanobubbles and the attraction between hydrophobic surfaces, J. Colloid Interface Sci. 325 (2008) 267–274. [22] X.H. Zhang, A. Quinn, W.A. Ducker, Nanobubbles at the interface between water and a hydrophobic solid, Langmuir 24 (2008) 4756–4764. [23] B. Zhao, Y. Song, S. Wang, B. Dai, L. Zhang, Y. Dong, J. Lü, J. Hu, Mechanical mapping of nanobubbles by PeakForce atomic force microscopy, Soft Matter 9 (2013) 8837. [24] A. Brotchie, X.H. Zhang, Response of interfacial nanobubbles to ultrasound irradiation, Soft Matter 7 (2011) 265–269. [25] S.-T. Lou, Z.-Q. Ouyang, Y. Zhang, X.-J. Li, J. Hu, M.-Q. Li, F.-J. Yang, Nanobubbles on solid surface imaged by atomic force microscopy, J. Vac. Sci. Technol. B 18 (2000) 2573. [26] N. Ishida, T. Inoue, M. Miyahara, K. Higashitani, Nano bubbles on a hydrophobic surface in water observed by tapping-mode atomic force microscopy, Langmuir 16 (2000) 6377–6380. [27] X.H. Zhang, A. Khan, W.A. Ducker, A nanoscale gas state, Phys. Rev. Lett. 98 (2007) 136101. [28] J. Weijs, D. Lohse, Why surface nanobubbles live for hours, Phys. Rev. Lett. 110 (2013). [29] X. Zhang, DerekY.C. Chan, D. Wang, N. Maeda, Stability of interfacial nanobubbles, Langmuir 29 (2013) 1017–1023. [30] Y. Liu, J. Wang, X. Zhang, W. Wang, Contact line pinning and the relationship between nanobubbles and substrates, J. Chem. Phys. 140 (2014) 54705. [31] M.P. Brenner, D. Lohse, Dynamic equilibrium mechanism for surface nanobubble stabilization, Phys. Rev. Lett. 101 (2008). [32] L. Ditscherlein, U.A. Peuker, Note Production of stable colloidal probes for hightemperature atomic force microscopy applications, The Review of scientific instruments 88 (2017) 46107. [33] J. Fritzsche, U.A. Peuker, Modeling adhesive force distributions on highly rough surfaces, Powder Technol. 289 (2016) 88–94. [34] D.P. Patil, J. Andrews, P. Uhlherr, Shear flocculation—kinetics of floc coalescence and breakage, Int. J. Miner. Process. 61 (2001) 171–188. [35] W. Yu, J. Gregory, L.C. Campos, Breakage and re-growth of flocs formed by charge neutralization using alum and polyDADMAC, Water Res. 44 (2010) 3959–3965. [36] M.A. Yukselen, J. Gregory, The reversibility of floc breakage, Int. J. Miner. Process. 73 (2004) 251–259. [37] P. Jarvis, B. Jefferson, J. Gregory, S.A. Parsons, A review of floc strength and breakage, Water Res. 39 (2005) 3121–3137. [38] A.V. Nguyen, D.-A. An-Vo, T. Tran-Cong, G.M. Evans, A review of stochastic description of the turbulence effect on bubble-particle interactions in flotation, Int. J. Miner. Process. 156 (2016) 75–86. [39] T.R. Camp, P.C. Stein, Velocity gradients and internal work in fluid motion, J. Boston Soc. Civil Eng. 85 (1943) 219–237. [40] P. Knüpfer, J. Fritzsche, T. Leistner, M. Rudolph, U.A. Peuker, Investigating the removal of particles from the air/water-interface −Modelling detachment forces using an energetic approach, Coll. Surf. A 513 (2017) 215–222. [41] H.-J. Butt, B. Cappella, M. Kappl, Force measurements with the atomic force microscope: technique, interpretation and applications, Surf. Sci. Rep. 59 (2005) 1–152.
wetting is clearly visible. The hydrophobic powder creates much larger agglomerates than the hydrophilic one even at higher shear rates. Furthermore, the stability of agglomerates which is quantized by the agglomerate strength exponent γ, is higher for the hydrophobic material. Considering the size of primary particles, for the smaller fraction of hydrophobic powder a higher agglomerate strength and in addition larger agglomerate sizes were detected. This might be attributed to the increased number of particles per agglomerate and thus, more individual bondings between the primary particles. 5. Conclusion In this study, the agglomeration of hydrophobic and hydrophilic alumina particles in aqueous solutions has been compared regarding to the equilibrium agglomeration size and the stability of agglomerates. It could be shown, that hydrophobic particles merge to larger agglomerates, which are additionally more stable than hydrophilic ones. This behavior is attributed to the nucleation of surface nanobubbles onto the hydrophobic particles due to the required solvent exchange. Thus, capillary forces between the particles are developed, that are larger in magnitude than van der Waals forces, acting between hydrophilic particles, respectively. This behavior was verified through adhesive force measurements via AFM between a single colloidal probe particle and the appropriate alumina surface. Additionally, the occurring nanobubbles were detected via AFM on representative hydrophobic alumina surfaces. Furthermore, the agglomerates were treated via ultrasonic and shear stress. Although ultrasonic stress leads to a total breakage of the agglomerates, the agglomeration is reversible. Different shear conditions lead to certain equilibrium sizes, from which the stability of agglomerates for every used material can be estimated. In general, the stability of hydrophobic agglomerates in aqueous flows is larger than the stability of hydrophilic ones. Due to the high surface energies of metal melts, similar effects like increased adhesion forces through their non-wetting behavior are conceivable in the real filtration process of metal melts. Acknowledgement The authors gratefully thank the German Research Foundation (DFG) for supporting the Collaborative Research Center CRC 920, subprojects B01 and B04. References [1] F. Delannay, L. Froyen, A. Deruyttere, The wetting of solids by molten metals and its relation to the preparation of metal-matrix composites composites, J. Mater. Sci. 22 (1987) 1–16. [2] O.I. Vinogradova, G.E. Yakubov, H.-J. Butt, Forces between polystyrene surfaces in water–electrolyte solutions: long-range attraction of two types? J. Chem. Phys. 114 (2001) 8124–8131. [3] J. Fritzsche, U.A. Peuker, Modeling adhesive forces caused by nanobubble capillary bridging, Coll. Surf. A 509 (2016) 457–466. [4] L. Ditscherlein, J. Fritzsche, U.A. Peuker, Study of nanobubbles on hydrophilic and hydrophobic alumina surfaces, Colloids Surf. A 497 (2016) 242–250. [5] H. Saint-Raymond, et al., Turbulent aggregation of alumina in water and nHeptane, J. Colloid Interface Sci. 202 (2) (1998) 238–250. [6] M. Cournil, F. Gruy, P. Gardin, H. Saint-Raymond, Modelling of solid particle aggregation dynamics in non-wetting liquid medium, Chem. Eng. Process. Process Intensif. 45 (2006) 586–597. [7] M. Cournil, F. Gruy, P. Gardin, H. Saint-Raymond, Experimental study and modeling of inclusion aggregation in turbulent flows to improve steel cleanliness, phys. stat. sol. (a) 189 (2002) 159–168. [8] F. Gruy, M. Cournil, P. Cugniet, Influence of nonwetting on the aggregation dynamics of micronic solid particles in a turbulent medium, J. Colloid Interface Sci.
123