Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Symmetric wetting heterogeneity suppresses fluid displacement hysteresis in granular piles

Symmetric wetting heterogeneity suppresses fluid displacement hysteresis in granular piles Symmetric wetting heterogeneity suppresses uid displacement hysteresis in granular piles R. Moosavi, M. Schr oter, S. Herminghaus Max-Planck-Institute for Dynamics and Self-Organization, Bunsenstr. 10, 37073 Gottingen, Germany (Dated: January 17, 2018) We investigate experimentally the impact of heterogeneity on the capillary pressure hysteresis in uid invasion of model porous media. We focus on `symmetric' heterogeneity, where the contact angles the uid interface makes with the oil-wet ( ) and the water-wet ( ) beads add up to . 1 2 While enhanced heterogeneity is usually known to increase hysteresis phenomena, we nd that hysteresis is greatly reduced when heterogeneities in wettability are introduced. On the contrary, geometric heterogeneity (like bi-disperse particle size) does not lead to such e ect. We provide a qualitative explanation of this surprising result, resting on rather general geometric arguments. PACS numbers: 68.05.-n; 68.08.-p; 05.40.-a; 64.75.-g I. INTRODUCTION 45 minutes in a solution prepared from 50 ml bicylo- hexyl (C H ), 0:240 ml carbon tetrachloride CCl and 12 22 4 0:214 ml of the OTS [21]. This was followed by a chlo- Immiscible displacement of one uid by another within roform rinse to remove unbound OTS molecules. Subse- a permeable medium, such as soil [1, 2], a porous rock quently, the sample was rinsed with ethanol and acetone formation[3{6], or a lter cake [7], is of tremendous and then dried in an oven for 30 minutes. importance in environmental, geological and industrial The procedure was repeated three times in order to settings. A deeper understanding of the fundamental enhance the quality of the chemisorbed silane layer. On mechanisms governing the morphology of the propagat- standard glass slides, this was found to yield an advanc- ing uid front would be of immediate impact on, e.g., ing contact angle for water against air of 114 , which irrigation, oil recovery [8{10] and storage in geological underscores the high grafting density achieved [21]. formations [3, 4, 11, 12]. However, progress in this direc- Our experimental setup is sketched in Fig. 1a. The tion has been strongly impeded by the enormous com- bottom and top of the sample cell consist of a hydrophilic plexity inherent to these systems owing to their random and oleophilic porous membrane, respectively, in order geometry. This pertains not only to natural samples. to prevent the uid front from leaving the cell, and to Even in piles of perfectly spherical particles of equal size, thus keep the two liquids strictly separated outside the which represent the most frequently studied model of a cell. The hydrophobic (oleophilic) membrane at the top permeable solid, the random nature of the packing geom- is made of 135 m thick porous Te on, the one at the etry renders the morphology of the invading uid front bottom is made of Polyamid (both supplied by Sartorius poorly predictable. Nevertheless, it has been recently AG). As immiscible uid phases, we chose hexadecane shown that important aspects of the morphology of the (Alfa Aesar, density  = 0:77 g=cm ), which was ltered invading uid front can be understood on the basis of prior to use through an alumina column (height 20 cm) merely the wetting geometry at the pore scale [13]. to remove surfactants, and milipore water as the aque- In the present paper, we apply this paradigm to the ous phase. The interfacial tension between the two liq- impact of sample heterogeneity. Following established uids was measured by means of the pendant drop method schemes [14{19], we study uid displacement in sam- and found to be 52 1 mN/m. The particularity of this ples consisting of piles of spherical glass beads with sub- choice of liquids lies in the fact that the contact angles millimeter diameter, which are chemically functionalized the uid interface makes with the bead surfaces are lying in order to control their wettability. Furthermore, we symmetrically around =2. Experimental evidence and study the interplay between geometric heterogeneity and physical consequences will be discussed further below. heterogeneous wettability. Indications that correlations Samples are prepared by rst pouring dry mixtures of between geometry and wettability may play a role date beads into the oil phase, then thoroughly degassing. Sub- back to 1959 [20] and should therefore be considered. sequently the oil saturated with beads is lled into the sample cell. After sealing the cell, its top outlet is con- nected to a tube which ended in a beaker with the same II. EXPERIMENTS uid. The latter was positioned on a precision balance which allows to measure the amount of expelled oil. The Soda lime glass beads were purchased from MoSci Cor- bottom of the sample container is connected, with a ex- poration and cleaned with piranha solution before fur- ible tube, to the bottom of a beaker containing water, ther treatment. Oil-wet beads were created by liquid- which can be shifted vertically by means of a translation phase silanization with octadecyl-trichlorosilane (OTS). stage. This allows for precise adjustment of the pressure The procedure consisted of immersing the beads for of the invading uid (i.e., water). arXiv:1801.04990v1 [physics.flu-dyn] 15 Jan 2018 2 (solid triangles in Fig. 1b). This CPS curve was repro- ducible within experimental scattering. In a recent study [13] with the same system, where the uid interface was monitored by means of x-ray tomography, we could es- tablish that gravity has no noticeable in uence on the interface morphology and the behaviour of the system. FIG. 2. Matric suction measured with samples of mixed wet beads of equal size. The scaled quantity p R= is plotted on the ordinate, the abscissa is  =  = (  )=2 + . oil 2 2 1 Closed circles: R = 1169m. Open circles: R = 19516m. Inset: log-log plot of p vs. the bead size, measured with mono-disperse samples of water-wet beads. The straight line has unity slope. FIG. 1. (a) Sketch of the experimental apparatus. The sam- ple cell is sealed at the top and the bottom with an oleophilic (om) and a hydrophilic (hm) membrane, respectively. This III. RESULTS prevents the uid interface from leaving the cell. The hydro- static pressure di erence between water and oil is controlled by the height of the water reservoir using the translation When the interface between two uids is in contact stage. The mass of the displaced oil (and thereby the wa- with a solid surface, the equilibrium interface morphol- ter saturation) is measured by a high precision balance. (b) ogy is characterized by a certain contact angle which es- An example saturation curve. Open triangles: rst run. Solid tablishes itself between the uid interface and the solid triangles: all consecutive runs. surface. If and are the interfacial tensions between 1 2 uid 1 or uid 2 with the solid, respectively, the equilib- rium contact angle is given by Capillary pressure saturation (CPS) curves were ob- cos  = (1) tained with a home-built experimental setup which has 1 2 been described in detail before [14]. Each CPS measure- where is the tension of the interface between the two ment begins with the sample fully oil-saturated (S = 0 (assumed immiscible) liquids. If a uid is being com- in Fig. 1b). The water is made to invade the cell by verti- pletely expelled by a second, immiscible uid, the corre- cally translating the water reservoir in steps equivalent to sponding capillary pressure, p , is given by the excess in- 1 mbar (upward pointing open triangles in Fig. 1b). At terfacial free energy gain upon exchanging one uid with each incremental step, the water saturation S is cal- the other. It is directly proportional to the di erence culated from the mass of displaced oil from the sam- in the interfacial tensions, , and hence to cos , and is ple. After oil saturation has been reached for the rst max commonly called the matric suction. time ( S ), further increase of pressure does not pro- For a sample of equally sized spherical beads with con- duce more oil, because the uid front cannot pass the tact angle  and radius R, one readily obtains oleophilic membrane at the top of the cell. The capil- lary pressure is then decreased again stepwise back to 3 cos p = ; (2) its original value, thereby allowing the oil phase to re- 1  R invade the sample cell (downward pointing open trian- gles in Fig. 1b). Subsequently, this procedure is cycled where  is the lling fraction of the beads (fraction of several times in order to establish a steady CPS curve total sample volume) and R is their radius. That the 3 matric suction indeed scales with the inverse radius of angle hysteresis (CAH),   . [23, 24]. Intuitively, adv rec the beads in our experiments is shown in the inset of one expects a similar phenomenology for wetting of in- Fig. 2. homogeneous porous media. If we add heterogeneity to For a sample containing two types of beads with di er- the porous medium, we might expect the hysteresis, p ent wettability (i.e., di erent contact angles  and  ) (cf. Fig. 1b), to increase. 1 2 and di erent size, we obtain Quite surprisingly, however, we observe the opposite. While heterogeneity in geometry, as introduced by mix- 3  cos   cos 1 1 2 2 ing beads of di erent size, is found to have no discernible p = + ; (3) 1  R R 1 2 impact on imbibition hysteresis, heterogeneity in wetta- bility even dramatically reduces the hysteresis of uid where  +  = . This suggests that in a sample 1 2 invasion. This is rather counterintuitive, and we will containing a mixture of water-wet and oil-wet beads, we provide here not only the evidence but also a tentative should nd a linear variation of p with the relative con- explanation based on simple geometric arguments. tent of the bead surface area. Accordingly, we plotted in the main panel of Fig. 2 the matric suction as a function of the fraction of the oil wet beads in mixed wet sam- IV. DISCUSSION ples. The data of the lled and open symbols have been obtained with beads of di erent sizes and scaled with re- Let us rst discuss p for binary mixtures of beads spect to =R, according to eq. (2). Not only do we nd of di erent size but equal wettability (oil wet or water a reasonable data collapse, but also we indeed nd a lin- wet). There are two main contributions to this hystere- ear relationship within experimental scattering. That p sis. One is the amplitude of variations in the interfacial goes to zero at equal fractions of both bead types indi- free energy as the uid front advances within the pile, cates that cos   cos  . This is important to note 1 2 and has to accommodate varying geometries of the inter- because it suggests that the geometry of the invading and stitial space. We will call this contribution the geometry receding fronts should be identical. This is what we re- induced hysteresis. These pressure variations should scale fer to as symmetric wetting heterogeneity. We then can strictly with the variations of the mean curvature of the write  =  , or  +  = , and hence 1=2 1 2 uid front, and hence with the size of throats between ad- jacent beads. The throat size between beads of di erent 3 sin 2 1 p = : (4) s radii R can neither be expressed in terms of hR i nor in i i 1  R R 2 1 terms of h1=R i . Nevertheless, hR i certainly provides i i a reasonable measure for the intrinsic length scale of the From the slope of the straight line in the main panel of sample. Hence we will, in the present paper, generally Fig. 2 and from our system parameters ( = 52mN/m, use hRi as the characteristic length scale of the system. = 0:60) we obtain sin  = 0:31 and hence  = 18 deg. The other contribution is the hysteresis inherent to the Let us now turn to the hysteresis in the drainage (heterogeneous) solid surfaces of the beads themselves. curves, p = p p , where p and p are the crit- + + As mentioned above, this gives rise to CAH, which is ex- ical pressures for invasion and for drainage, respectively. pected to contribute additively to the hysteresis observed They correspond to the positions of the steep slopes of in the drainage curves. the saturation curves as displayed in Fig. 1. We de ne In fact, we observe that the hysteresis scales in a char- (j 2 f+;g) acteristic manner with the bead diameter for both water- max wet and oil-wet beads, as shown in Fig. 3a, where larger j min p = p S S (p) S dp; (5) beads lead to smaller pressure hysteresis and vice versa. j max w w However, we also observe a substantial additive di erence min of 2:03 hPa on average between both data sets. If this where p and p are chosen suciently far outside is subtracted from the data obtained for oil wet samples, min max one nds a reasonable data collapse with hysteresis data the range displayed in Fig. 1b to assure convergence of the branches of the saturation curves. S (p) and S (p) obtained for monodisperse water wet samples of variable w w are the water saturations for ascending (invasion) and bead size, as shown in Fig. 3b. The unity slope of the max min dotted line indicates that all pressures scale inversely pro- descending (drainage) pressure, and S = S S w w is the amount of reversibly expelled uid. portional to the average size of the beads, which strongly suggests that this is entirely due to geometric e ects, and If a uid interface moves with respect to a solid surface it makes contact with, one frequently observes a certain not noticeably a ected by CAH at the three-phase con- tact lines. hysteresis in the contact angle, i.e., a di erence between the contact angle measured with advancing front ( ) We conclude that we have appreciable CAH on the adv and the contact angle when measured with the receding surfaces of the oil wet beads, which gives rise to the ex- front ( ). The reason for this hysteresis is the hetero- tra 2:03 hPa of hysteresis apparent in Fig. 3a. This is rec geneity of the solid surface, be it topographical or chem- well conceivable, as the oil wet beads had to be chem- ical [22]. If heterogeneity increases, so does the contact ically functionalized, which tends to add some chemi- 4 ometric heterogeneity, which is maximal at equal frac- tions of large and small beads (center of Fig. 3a), has no discernible e ect on the hysteresis. Instead, the latter linearly interpolates between the extreme values corre- sponding to the mono-disperse samples, as supported by the data collapse in Fig. 3b. FIG. 3. (a) Hysteresis measured with bi-disperse piles of oil wet (open symbols) and water wet (full symbols) beads, as FIG. 4. Hysteresis measured with bi-disperse piles of beads, a function of the relative fraction  of large beads. Within as a function of the relative fraction . (a) Mixtures of beads scattering, the data sets are almost in parallel, the decreas- of di erent size (R = 116 9 m, R = 195 16 m). Closed 1 2 ing trend stemming from the change in average length scale symbols: beads of equal wettability. Open symbols: beads of (small beads and high pressures, left hand side, to large beads di erent wettability. Squares: large beads water wet, small and low pressures, right hand side). (b) Hysteresis as a func- beads oil wet. Diamonds: large beads oil wet, small beads tion of the capillary pressure scale, =hRi. Open symbols: water wet. (b) Mixtures of beads of di erent wettability. monodisperse beads of variable size. Full symbols: bidisperse Squares and diamonds, data as above. Upward pointing tri- mixtures of beads. Circles: R = 116m, R = 195m. Tri- 1 2 angles, large beads only. Downward pointing triangles, small angles: R = 58m, R = 231m. The dotted line with unity 1 2 beads only. Dashed curve, see text. slope represents what would be expected if the bead size were uniformly varying between the largest (left) and the smallest (right) size. Let us now include heterogeneity of wettability. Fig. 4 contains all data of geometry induced hysteresis for vari- able situations studied. For all data points involving oil- cal heterogeneity, and hence CAH. This contribution will wet beads, the intrinsic hysteresis has been subtracted henceforth be called the intrinsic hysteresis. Although it according to their fraction of sample surface. Further- is somewhat surprising that the CAH on the water-wet more, data have been scaled the same way as in Fig. 2, beads is too small to show up in our data, we have to with R having been replaced by the average radius, hRi, accept this as an experimental result and will henceforth as suggested by Fig. 3b. The main observation is that treat CAH on the water-wet beads as negligible for the while nd no noticeable in uence of geometric hetero- e ects to be considered here. geneity on the width of the hysteresis, we observe an The main conclusion we draw from Fig. 3 is that ge- almost complete suppression of the hysteresis in the case 5 of mixed wet piles, irrespective of the geometric hetero- geneity. To arrive at a tentative explanation of this counter- intuitive result, we consider the geometry of the uid interface close to a contact between two adjacent beads. If these have the same wettability, as sketched on the left hand side of Fig. 5a, the uid which wets the beads better ( uid 1 in this case) will form a capillary bridge [25], with the other uid as the surrounding phase. If the beads have di erent wettability, a capillary bridge of this kind is not likely to form. In particular, if the two contact angles are symmetric around =2 as in our case ( =   with  > 0), an axisymmetric solution like the one shown in the left sketch does not exist. Hence we have to discuss which uid interface morphology is likely to emerge at a contact point between two spherical beads with contact angles  and  , where  +   . 1 2 1 2 It is straightforward to appreciate that the arrange- ment sketched on the right side of Fig. 5a provides a solution. It consists of a plane uid interface which con- tains the point of contact and makes an angle  = j j with the line connecting the centers of the two beads (dash-dotted). One readily sees that this plane makes a contact angle of +  at its line of intersection with FIG. 5. (a) Interface (solid) between two uids (1 and 2) one of the beads, and a contact angle of  on its near a contact between two beads. Left: Contact between line of intersection with the other. As the at uid inter- two beads wettable by uid 1. The latter is likely to form face corresponds to zero Laplace pressure, this solution a capillary bridge between both beads, with uid 2 as the is expected to form easily, and not to add to the Laplace surrounding phase. Right: Contact between two beads of dif- pressure constituting the geometric hysteresis of the CPS ferent wettability. A capillary bridge is not likely to form. For curve. It is expected to be pinned to the contact point by the `symmetric' case  = =2 (as for our system), there is a mild minimum around zero Laplace pressure. Further- not even an axisymmetric solution. Instead, the interface goes more, it will provide a natural boundary between regions at through the contact point, making an angle  with the in which the liquid phase wetting the majority of beads line connecting the centers of the beads (dash-dotted). (b) represents the majority phase liquid. Excursions from The thresholds for site percolation ( lled symbols) and bond zero applied pressure (cf. Fig. 1b), in both positive and percolation (open symbols) for a range of ordered (squares) and random (circles) lattices ([26] and references therein). negative direction, should then easily displace uid on the low pressure side. On the other hand, the wetting phase will form capil- In the mixed wet pile, however, the situation is dif- lary bridges at the contact points between like wet beads, ferent since both uids have the same (or very similar) as sketched in the left part of Fig. 5. From studies of wet topology. It can be qualitatively discussed by consid- granular piles [27{29], it is well known that neighboring ering the statistics of capillary bridges (or equivalent capillary bridges start to grow into clusters already at a structures) occurring in the sample. In a random pile liquid content of about 2,5%. For piles of beads of equal of spheres, it is well known that there are about six con- wettability, there is a bicontinuous phase in a wide range tacts on each sphere [25, 26, 29]. Hence each contact has of liquid content, where both the wetting and the non- four nearest neighbor contacts on the same sphere. Con- wetting phase are percolated [29]. The former as merged sequently, each contact, and hence each capillary bridge, capillary bridges, the latter lling the interstitial space has z = 8 nearest neighbors in total on the two spheres in between. If both uids form percolated networks, between which it is formed. they can in principle be depleted or in ated without any substantial movement of contact lines or any other re- In order to discuss our results in the context of per- arrangements of uid interface geometry. However, for colation theory, we show in Fig. 5b the known percola- piles with a single wettability, this is only possible in the tion thresholds as a function of coordination number z late stage of expelling the wetting (i.e., capillary bridge) for several di erent lattices, including ordered and dis- uid [13]. For most of the range of saturation, expelling ordered ones. We see that aside from the data grouping one liquid with the other is accompanied by frequent re- into the two subgroups corresponding to bond percola- arrangements and concomitant changes in the topology tion (open symbols) and site percolation ( lled symbols), of the uid interface [27, 28], hence with considerable the percolation threshold is quite independent of the pre- hysteresis in the CPS curve. cise structure of the lattices, and can be well related to 6 the coordination number as the only relevant descriptor. V. CONCLUSIONS AND OUTLOOK We realize that at z = 8 about 25% of occupancy is sucient for site percolation. Hence for each of the liq- Clearly, the details of the complex topology of the uid uid phases a percolated cluster of capillary bridges will interface still leaves a lot of room for investigations, but persist as long as not more than 75% of the spheres are qualitatively the observed phenomenon seems to be cap- wet for the other liquid. This predicts that the e ect tured by the above considerations. One may think of a of hysteresis suppression should be present around equal wide range of potential applications here. For instance, fractions of beads of both wettabilities, and extend from it seems possible to provide a novel design for oil-water about 25% to about 75%. This is roughly what we ob- separation columns, where the interstitial geometry of serve in the experiments. the ne granulate provides a particularly large area of interaction between both phases. Furthermore, it will be interesting to investigate how universal the behaviour re- ported here turns out to be, in view of the rather simple mechanism behind it. In particular, it will be of great in- terest to investigate until how far away from symmetric heterogeneity this e ect persists. In other words: how A complementary approach which leads to a similar large has  +   to be for this e ect to break down? result may be gained from the number of oppositely wet 1 2 Moreover, many hysteresis e ects in uid displacement in contacts, as sketched on the right hand side of Fig. 5a. porous media should be revisited in view of the present If ' is the number fraction of beads wet by uid i, results. we have ' + ' = 1, and the fraction of like-wet con- i 2 2 2 2 tacts is ' = ' + ' = 2' 2' + 1. The fraction l 1 1 2 1 of oppositely-wet contacts is ' = 1 ' . If we plot o l ACKNOWLEDGEMENTS ' ' = 4' 4' + 1 into Fig. 4 as a proxy for the l o 1 necessity of topologocal rearrangements (dashed curve), we see that the tendency observed in the CPS curve hys- The authors appreciate generous nancial support teresis is quite well reproduced. from BP plc. within the ExploRe program. [1] R. A. Wooding, Scott W. Tyler, and Ian "White, \Con- [10] Tue Hassenkam, C. S. Pedersen, K. Dalby, T. Austad, vection in groundwater below an evaporating salt lake," and S. L. S. Stipp, \Pore scale observation of low salin- Water Resources Research 33, 1199{1217 (1997). ity e ects on outcrop and oil reservoir sandstone," Col- [2] Bert van Bocxlaer, Dirk Verschuren, Georg Schettler, loids and Surfaces A: Physicochemical and Engineering and Stefan Kr opelin, \Modern and eary Holocene mollusc Aspects 390, 179{188 (2011). fauna of the Ounianga lakes: implications for the paleo- [11] F. M. Orr, \Onshore geologic storage of CO ," Science hydrology of the central Sahara," J. Quaternary Sci. 26, 325, 1656{1658 (2009). 433{447 (2011). [12] Mike J. Bickle, \Geological carbon storage," Nature Geo- [3] W. G. Anderson, \Wettability literature survey- part 4: science 2, 815{818 (2009). E ects of wettability on capillary pressure," Journal of [13] Kamal Singh, Hagen Scholl, Martin Brinkmann, Marco Petroleum Technology 39, 1283{1300 (1987). DiMichiel, Mario Scheel, Stephan Herminghaus, and [4] Norman R. Morrow, \Wettability and its e ect on oil Ralf Seemann, \The role of local instabilities in uid in- recovery," Journal of Petroleum Technology 42, 1477{ vasion into permeable media," Scienti c Reports 7, 444 1484 (1990). (2017). [5] F. A. L. Dullien, Porous Media: Fluid Transport and [14] Julie Murison, Benot Semin, Jean-Christophe Baret, Pore Structure (Academic Press, 1991). Stephan Herminghaus, Matthias Schr oter, and Martin [6] M. Sahimi, Flow and Transport in Porous Media and Brinkmann, \Wetting heterogeneities in porous media Fractured Rock: From Classical Methods to Modern Ap- control ow dissipation," Phys. Rev. Applied 2, 034002 proaches (John Wiley and Sons, 2011). (2014). [7] Po-Zen Wong, Methods in the Physics of Porous Media, [15] Martin Blunt and Peter King, \Relative permeabili- Experimental Methods in the Physical Sciences, Vol. 35 ties from two- and three-dimensional pore-scale net- (Academic Press, San Diego, USA, 2013). work modelling," Tran. Porous media Journal 6, 407{433 [8] Arnaud Lager, Kevin John Webb, Ian Ralph Collins, and (1991). Diane Marie Richmond, \Losal enhanced oil recovery: [16] Martin Blunt, \Pore level modeling of the e ects of wet- Evidence of enhanced oil recovery at the reservoir scale," tability," SPE Journal 2, 494{510 (1997). SPE Journal (2008). [17] A. B. Dixit, J. S. Buckley, S. R. McDougall, and K. S. [9] T. Austad, S. F. Shariatpanahi, S. Strand, C. J. J. Sorbie, \Empirical measures of wettability in porous me- Black, and K. J. Webb, \Conditions for a Low-Salinity dia and the relationship between them derived from pore- Enhanced Oil Recovery (EOR) E ect in Carbonate Oil scale modelling," Tran. Porous media Journal 40, 27{54 Reservoirs," Energy & Fuels 26, 569{575 (2012). (2000). 7 [18] P al-Eric ren and Stig Bakke, \Reconstruction of Berea [24] C. Wylock, M. Pradas, B. Haut, P. Colinet, and S. Kalli- sandstone and pore-scale modelling of wettability ef- adasis, \Disorder-induced hysteresis and nonlocality of fects," SPE Journal 39, 177{199 (2003). contact line motion in chemically heterogeneous mi- [19] Xiucai Zhao, Martin J. Blunt, and Jun Yao, \Pore-scale crochannels," Physics of Fluids 24, 032108 (2012). modeling: E ects of wettability on water ood oil recov- [25] Mika M. Kohonen, Dimitrios Geromichalos, Mario ery," Journal of Petroleum Science and Engineering 71, Scheel, Christoph Schier, and Stephan Herminghaus, 169{178 (2010). \On capillary bridges in wet granular materials," Physica [20] I. Fatt, A. Waldemar, and Jr. Kliko , \E ect of frac- A 339, 7{15 (2004). tional wettability on multiphase ow through porous me- [26] M. J. Powell, \Site percolation in randomly packed dia," SPE Journal 11, 71{75 (1959). spheres," Phys. Rev. B 20, 4194 (1979). [21] M. Lessel, O. B aumchen, M. Klos, H. H ahl, R. Fetzer, [27] Mario Scheel, Ralf Seemann, Martin Brinkmann, Marco M. Paulus, R. Seemann, and K. Jacobs, \Self-assembled DiMichiel, Adrian Sheppard, Boris Breidenbach, and silane monolayers: an ecient step-by-step recipe for Stephan Herminghaus, \Morphological clues to wet gran- high-quality, low energy surfaces," Surface and Interface ular pile stability," Nature Materials 7, 189 (2008). Analysis 47, 557564 (2014). [28] Mario Scheel, Ralf Seemann, Martin Brinkmann, Marco [22] H. Perrin, R. Lhermerout, K. Davitt, E. Rolley, and DiMichiel, Adrian Sheppard, and Stephan Herminghaus, B. Andreotti, \Defects at the nanoscale impact contact \Liquid distribution and cohesion in wet granular as- line motion at all scales," Phys. Rev. Lett. 116, 184502 semblies beyond the capillary bridge regime," J. Phys.: (2016). Cond. Mat. 20, 494236 (2008). [23] R. Dufour, C. Semprebon, and S. Herminghaus, \Filling [29] Stephan Herminghaus, Wet Granular Matter: a Truly transitions on rough surfaces: inadequacy of Gaussian Complex Fluid, Series in Soft Condensed Matter, Vol. 6 surface models," Phys. Rev. E 93, 032802 (2016). (World Scienti c, 2013). http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Condensed Matter arXiv (Cornell University)

Symmetric wetting heterogeneity suppresses fluid displacement hysteresis in granular piles

Download PDF
7 pages

Loading next page...
 
/lp/arxiv-cornell-university/symmetric-wetting-heterogeneity-suppresses-fluid-displacement-kplpiuRnuP

References (31)

ISSN
2469-990X
eISSN
ARCH-3331
DOI
10.1103/PhysRevFluids.3.024304
Publisher site
See Article on Publisher Site

Abstract

Symmetric wetting heterogeneity suppresses uid displacement hysteresis in granular piles R. Moosavi, M. Schr oter, S. Herminghaus Max-Planck-Institute for Dynamics and Self-Organization, Bunsenstr. 10, 37073 Gottingen, Germany (Dated: January 17, 2018) We investigate experimentally the impact of heterogeneity on the capillary pressure hysteresis in uid invasion of model porous media. We focus on `symmetric' heterogeneity, where the contact angles the uid interface makes with the oil-wet ( ) and the water-wet ( ) beads add up to . 1 2 While enhanced heterogeneity is usually known to increase hysteresis phenomena, we nd that hysteresis is greatly reduced when heterogeneities in wettability are introduced. On the contrary, geometric heterogeneity (like bi-disperse particle size) does not lead to such e ect. We provide a qualitative explanation of this surprising result, resting on rather general geometric arguments. PACS numbers: 68.05.-n; 68.08.-p; 05.40.-a; 64.75.-g I. INTRODUCTION 45 minutes in a solution prepared from 50 ml bicylo- hexyl (C H ), 0:240 ml carbon tetrachloride CCl and 12 22 4 0:214 ml of the OTS [21]. This was followed by a chlo- Immiscible displacement of one uid by another within roform rinse to remove unbound OTS molecules. Subse- a permeable medium, such as soil [1, 2], a porous rock quently, the sample was rinsed with ethanol and acetone formation[3{6], or a lter cake [7], is of tremendous and then dried in an oven for 30 minutes. importance in environmental, geological and industrial The procedure was repeated three times in order to settings. A deeper understanding of the fundamental enhance the quality of the chemisorbed silane layer. On mechanisms governing the morphology of the propagat- standard glass slides, this was found to yield an advanc- ing uid front would be of immediate impact on, e.g., ing contact angle for water against air of 114 , which irrigation, oil recovery [8{10] and storage in geological underscores the high grafting density achieved [21]. formations [3, 4, 11, 12]. However, progress in this direc- Our experimental setup is sketched in Fig. 1a. The tion has been strongly impeded by the enormous com- bottom and top of the sample cell consist of a hydrophilic plexity inherent to these systems owing to their random and oleophilic porous membrane, respectively, in order geometry. This pertains not only to natural samples. to prevent the uid front from leaving the cell, and to Even in piles of perfectly spherical particles of equal size, thus keep the two liquids strictly separated outside the which represent the most frequently studied model of a cell. The hydrophobic (oleophilic) membrane at the top permeable solid, the random nature of the packing geom- is made of 135 m thick porous Te on, the one at the etry renders the morphology of the invading uid front bottom is made of Polyamid (both supplied by Sartorius poorly predictable. Nevertheless, it has been recently AG). As immiscible uid phases, we chose hexadecane shown that important aspects of the morphology of the (Alfa Aesar, density  = 0:77 g=cm ), which was ltered invading uid front can be understood on the basis of prior to use through an alumina column (height 20 cm) merely the wetting geometry at the pore scale [13]. to remove surfactants, and milipore water as the aque- In the present paper, we apply this paradigm to the ous phase. The interfacial tension between the two liq- impact of sample heterogeneity. Following established uids was measured by means of the pendant drop method schemes [14{19], we study uid displacement in sam- and found to be 52 1 mN/m. The particularity of this ples consisting of piles of spherical glass beads with sub- choice of liquids lies in the fact that the contact angles millimeter diameter, which are chemically functionalized the uid interface makes with the bead surfaces are lying in order to control their wettability. Furthermore, we symmetrically around =2. Experimental evidence and study the interplay between geometric heterogeneity and physical consequences will be discussed further below. heterogeneous wettability. Indications that correlations Samples are prepared by rst pouring dry mixtures of between geometry and wettability may play a role date beads into the oil phase, then thoroughly degassing. Sub- back to 1959 [20] and should therefore be considered. sequently the oil saturated with beads is lled into the sample cell. After sealing the cell, its top outlet is con- nected to a tube which ended in a beaker with the same II. EXPERIMENTS uid. The latter was positioned on a precision balance which allows to measure the amount of expelled oil. The Soda lime glass beads were purchased from MoSci Cor- bottom of the sample container is connected, with a ex- poration and cleaned with piranha solution before fur- ible tube, to the bottom of a beaker containing water, ther treatment. Oil-wet beads were created by liquid- which can be shifted vertically by means of a translation phase silanization with octadecyl-trichlorosilane (OTS). stage. This allows for precise adjustment of the pressure The procedure consisted of immersing the beads for of the invading uid (i.e., water). arXiv:1801.04990v1 [physics.flu-dyn] 15 Jan 2018 2 (solid triangles in Fig. 1b). This CPS curve was repro- ducible within experimental scattering. In a recent study [13] with the same system, where the uid interface was monitored by means of x-ray tomography, we could es- tablish that gravity has no noticeable in uence on the interface morphology and the behaviour of the system. FIG. 2. Matric suction measured with samples of mixed wet beads of equal size. The scaled quantity p R= is plotted on the ordinate, the abscissa is  =  = (  )=2 + . oil 2 2 1 Closed circles: R = 1169m. Open circles: R = 19516m. Inset: log-log plot of p vs. the bead size, measured with mono-disperse samples of water-wet beads. The straight line has unity slope. FIG. 1. (a) Sketch of the experimental apparatus. The sam- ple cell is sealed at the top and the bottom with an oleophilic (om) and a hydrophilic (hm) membrane, respectively. This III. RESULTS prevents the uid interface from leaving the cell. The hydro- static pressure di erence between water and oil is controlled by the height of the water reservoir using the translation When the interface between two uids is in contact stage. The mass of the displaced oil (and thereby the wa- with a solid surface, the equilibrium interface morphol- ter saturation) is measured by a high precision balance. (b) ogy is characterized by a certain contact angle which es- An example saturation curve. Open triangles: rst run. Solid tablishes itself between the uid interface and the solid triangles: all consecutive runs. surface. If and are the interfacial tensions between 1 2 uid 1 or uid 2 with the solid, respectively, the equilib- rium contact angle is given by Capillary pressure saturation (CPS) curves were ob- cos  = (1) tained with a home-built experimental setup which has 1 2 been described in detail before [14]. Each CPS measure- where is the tension of the interface between the two ment begins with the sample fully oil-saturated (S = 0 (assumed immiscible) liquids. If a uid is being com- in Fig. 1b). The water is made to invade the cell by verti- pletely expelled by a second, immiscible uid, the corre- cally translating the water reservoir in steps equivalent to sponding capillary pressure, p , is given by the excess in- 1 mbar (upward pointing open triangles in Fig. 1b). At terfacial free energy gain upon exchanging one uid with each incremental step, the water saturation S is cal- the other. It is directly proportional to the di erence culated from the mass of displaced oil from the sam- in the interfacial tensions, , and hence to cos , and is ple. After oil saturation has been reached for the rst max commonly called the matric suction. time ( S ), further increase of pressure does not pro- For a sample of equally sized spherical beads with con- duce more oil, because the uid front cannot pass the tact angle  and radius R, one readily obtains oleophilic membrane at the top of the cell. The capil- lary pressure is then decreased again stepwise back to 3 cos p = ; (2) its original value, thereby allowing the oil phase to re- 1  R invade the sample cell (downward pointing open trian- gles in Fig. 1b). Subsequently, this procedure is cycled where  is the lling fraction of the beads (fraction of several times in order to establish a steady CPS curve total sample volume) and R is their radius. That the 3 matric suction indeed scales with the inverse radius of angle hysteresis (CAH),   . [23, 24]. Intuitively, adv rec the beads in our experiments is shown in the inset of one expects a similar phenomenology for wetting of in- Fig. 2. homogeneous porous media. If we add heterogeneity to For a sample containing two types of beads with di er- the porous medium, we might expect the hysteresis, p ent wettability (i.e., di erent contact angles  and  ) (cf. Fig. 1b), to increase. 1 2 and di erent size, we obtain Quite surprisingly, however, we observe the opposite. While heterogeneity in geometry, as introduced by mix- 3  cos   cos 1 1 2 2 ing beads of di erent size, is found to have no discernible p = + ; (3) 1  R R 1 2 impact on imbibition hysteresis, heterogeneity in wetta- bility even dramatically reduces the hysteresis of uid where  +  = . This suggests that in a sample 1 2 invasion. This is rather counterintuitive, and we will containing a mixture of water-wet and oil-wet beads, we provide here not only the evidence but also a tentative should nd a linear variation of p with the relative con- explanation based on simple geometric arguments. tent of the bead surface area. Accordingly, we plotted in the main panel of Fig. 2 the matric suction as a function of the fraction of the oil wet beads in mixed wet sam- IV. DISCUSSION ples. The data of the lled and open symbols have been obtained with beads of di erent sizes and scaled with re- Let us rst discuss p for binary mixtures of beads spect to =R, according to eq. (2). Not only do we nd of di erent size but equal wettability (oil wet or water a reasonable data collapse, but also we indeed nd a lin- wet). There are two main contributions to this hystere- ear relationship within experimental scattering. That p sis. One is the amplitude of variations in the interfacial goes to zero at equal fractions of both bead types indi- free energy as the uid front advances within the pile, cates that cos   cos  . This is important to note 1 2 and has to accommodate varying geometries of the inter- because it suggests that the geometry of the invading and stitial space. We will call this contribution the geometry receding fronts should be identical. This is what we re- induced hysteresis. These pressure variations should scale fer to as symmetric wetting heterogeneity. We then can strictly with the variations of the mean curvature of the write  =  , or  +  = , and hence 1=2 1 2 uid front, and hence with the size of throats between ad- jacent beads. The throat size between beads of di erent 3 sin 2 1 p = : (4) s radii R can neither be expressed in terms of hR i nor in i i 1  R R 2 1 terms of h1=R i . Nevertheless, hR i certainly provides i i a reasonable measure for the intrinsic length scale of the From the slope of the straight line in the main panel of sample. Hence we will, in the present paper, generally Fig. 2 and from our system parameters ( = 52mN/m, use hRi as the characteristic length scale of the system. = 0:60) we obtain sin  = 0:31 and hence  = 18 deg. The other contribution is the hysteresis inherent to the Let us now turn to the hysteresis in the drainage (heterogeneous) solid surfaces of the beads themselves. curves, p = p p , where p and p are the crit- + + As mentioned above, this gives rise to CAH, which is ex- ical pressures for invasion and for drainage, respectively. pected to contribute additively to the hysteresis observed They correspond to the positions of the steep slopes of in the drainage curves. the saturation curves as displayed in Fig. 1. We de ne In fact, we observe that the hysteresis scales in a char- (j 2 f+;g) acteristic manner with the bead diameter for both water- max wet and oil-wet beads, as shown in Fig. 3a, where larger j min p = p S S (p) S dp; (5) beads lead to smaller pressure hysteresis and vice versa. j max w w However, we also observe a substantial additive di erence min of 2:03 hPa on average between both data sets. If this where p and p are chosen suciently far outside is subtracted from the data obtained for oil wet samples, min max one nds a reasonable data collapse with hysteresis data the range displayed in Fig. 1b to assure convergence of the branches of the saturation curves. S (p) and S (p) obtained for monodisperse water wet samples of variable w w are the water saturations for ascending (invasion) and bead size, as shown in Fig. 3b. The unity slope of the max min dotted line indicates that all pressures scale inversely pro- descending (drainage) pressure, and S = S S w w is the amount of reversibly expelled uid. portional to the average size of the beads, which strongly suggests that this is entirely due to geometric e ects, and If a uid interface moves with respect to a solid surface it makes contact with, one frequently observes a certain not noticeably a ected by CAH at the three-phase con- tact lines. hysteresis in the contact angle, i.e., a di erence between the contact angle measured with advancing front ( ) We conclude that we have appreciable CAH on the adv and the contact angle when measured with the receding surfaces of the oil wet beads, which gives rise to the ex- front ( ). The reason for this hysteresis is the hetero- tra 2:03 hPa of hysteresis apparent in Fig. 3a. This is rec geneity of the solid surface, be it topographical or chem- well conceivable, as the oil wet beads had to be chem- ical [22]. If heterogeneity increases, so does the contact ically functionalized, which tends to add some chemi- 4 ometric heterogeneity, which is maximal at equal frac- tions of large and small beads (center of Fig. 3a), has no discernible e ect on the hysteresis. Instead, the latter linearly interpolates between the extreme values corre- sponding to the mono-disperse samples, as supported by the data collapse in Fig. 3b. FIG. 3. (a) Hysteresis measured with bi-disperse piles of oil wet (open symbols) and water wet (full symbols) beads, as FIG. 4. Hysteresis measured with bi-disperse piles of beads, a function of the relative fraction  of large beads. Within as a function of the relative fraction . (a) Mixtures of beads scattering, the data sets are almost in parallel, the decreas- of di erent size (R = 116 9 m, R = 195 16 m). Closed 1 2 ing trend stemming from the change in average length scale symbols: beads of equal wettability. Open symbols: beads of (small beads and high pressures, left hand side, to large beads di erent wettability. Squares: large beads water wet, small and low pressures, right hand side). (b) Hysteresis as a func- beads oil wet. Diamonds: large beads oil wet, small beads tion of the capillary pressure scale, =hRi. Open symbols: water wet. (b) Mixtures of beads of di erent wettability. monodisperse beads of variable size. Full symbols: bidisperse Squares and diamonds, data as above. Upward pointing tri- mixtures of beads. Circles: R = 116m, R = 195m. Tri- 1 2 angles, large beads only. Downward pointing triangles, small angles: R = 58m, R = 231m. The dotted line with unity 1 2 beads only. Dashed curve, see text. slope represents what would be expected if the bead size were uniformly varying between the largest (left) and the smallest (right) size. Let us now include heterogeneity of wettability. Fig. 4 contains all data of geometry induced hysteresis for vari- able situations studied. For all data points involving oil- cal heterogeneity, and hence CAH. This contribution will wet beads, the intrinsic hysteresis has been subtracted henceforth be called the intrinsic hysteresis. Although it according to their fraction of sample surface. Further- is somewhat surprising that the CAH on the water-wet more, data have been scaled the same way as in Fig. 2, beads is too small to show up in our data, we have to with R having been replaced by the average radius, hRi, accept this as an experimental result and will henceforth as suggested by Fig. 3b. The main observation is that treat CAH on the water-wet beads as negligible for the while nd no noticeable in uence of geometric hetero- e ects to be considered here. geneity on the width of the hysteresis, we observe an The main conclusion we draw from Fig. 3 is that ge- almost complete suppression of the hysteresis in the case 5 of mixed wet piles, irrespective of the geometric hetero- geneity. To arrive at a tentative explanation of this counter- intuitive result, we consider the geometry of the uid interface close to a contact between two adjacent beads. If these have the same wettability, as sketched on the left hand side of Fig. 5a, the uid which wets the beads better ( uid 1 in this case) will form a capillary bridge [25], with the other uid as the surrounding phase. If the beads have di erent wettability, a capillary bridge of this kind is not likely to form. In particular, if the two contact angles are symmetric around =2 as in our case ( =   with  > 0), an axisymmetric solution like the one shown in the left sketch does not exist. Hence we have to discuss which uid interface morphology is likely to emerge at a contact point between two spherical beads with contact angles  and  , where  +   . 1 2 1 2 It is straightforward to appreciate that the arrange- ment sketched on the right side of Fig. 5a provides a solution. It consists of a plane uid interface which con- tains the point of contact and makes an angle  = j j with the line connecting the centers of the two beads (dash-dotted). One readily sees that this plane makes a contact angle of +  at its line of intersection with FIG. 5. (a) Interface (solid) between two uids (1 and 2) one of the beads, and a contact angle of  on its near a contact between two beads. Left: Contact between line of intersection with the other. As the at uid inter- two beads wettable by uid 1. The latter is likely to form face corresponds to zero Laplace pressure, this solution a capillary bridge between both beads, with uid 2 as the is expected to form easily, and not to add to the Laplace surrounding phase. Right: Contact between two beads of dif- pressure constituting the geometric hysteresis of the CPS ferent wettability. A capillary bridge is not likely to form. For curve. It is expected to be pinned to the contact point by the `symmetric' case  = =2 (as for our system), there is a mild minimum around zero Laplace pressure. Further- not even an axisymmetric solution. Instead, the interface goes more, it will provide a natural boundary between regions at through the contact point, making an angle  with the in which the liquid phase wetting the majority of beads line connecting the centers of the beads (dash-dotted). (b) represents the majority phase liquid. Excursions from The thresholds for site percolation ( lled symbols) and bond zero applied pressure (cf. Fig. 1b), in both positive and percolation (open symbols) for a range of ordered (squares) and random (circles) lattices ([26] and references therein). negative direction, should then easily displace uid on the low pressure side. On the other hand, the wetting phase will form capil- In the mixed wet pile, however, the situation is dif- lary bridges at the contact points between like wet beads, ferent since both uids have the same (or very similar) as sketched in the left part of Fig. 5. From studies of wet topology. It can be qualitatively discussed by consid- granular piles [27{29], it is well known that neighboring ering the statistics of capillary bridges (or equivalent capillary bridges start to grow into clusters already at a structures) occurring in the sample. In a random pile liquid content of about 2,5%. For piles of beads of equal of spheres, it is well known that there are about six con- wettability, there is a bicontinuous phase in a wide range tacts on each sphere [25, 26, 29]. Hence each contact has of liquid content, where both the wetting and the non- four nearest neighbor contacts on the same sphere. Con- wetting phase are percolated [29]. The former as merged sequently, each contact, and hence each capillary bridge, capillary bridges, the latter lling the interstitial space has z = 8 nearest neighbors in total on the two spheres in between. If both uids form percolated networks, between which it is formed. they can in principle be depleted or in ated without any substantial movement of contact lines or any other re- In order to discuss our results in the context of per- arrangements of uid interface geometry. However, for colation theory, we show in Fig. 5b the known percola- piles with a single wettability, this is only possible in the tion thresholds as a function of coordination number z late stage of expelling the wetting (i.e., capillary bridge) for several di erent lattices, including ordered and dis- uid [13]. For most of the range of saturation, expelling ordered ones. We see that aside from the data grouping one liquid with the other is accompanied by frequent re- into the two subgroups corresponding to bond percola- arrangements and concomitant changes in the topology tion (open symbols) and site percolation ( lled symbols), of the uid interface [27, 28], hence with considerable the percolation threshold is quite independent of the pre- hysteresis in the CPS curve. cise structure of the lattices, and can be well related to 6 the coordination number as the only relevant descriptor. V. CONCLUSIONS AND OUTLOOK We realize that at z = 8 about 25% of occupancy is sucient for site percolation. Hence for each of the liq- Clearly, the details of the complex topology of the uid uid phases a percolated cluster of capillary bridges will interface still leaves a lot of room for investigations, but persist as long as not more than 75% of the spheres are qualitatively the observed phenomenon seems to be cap- wet for the other liquid. This predicts that the e ect tured by the above considerations. One may think of a of hysteresis suppression should be present around equal wide range of potential applications here. For instance, fractions of beads of both wettabilities, and extend from it seems possible to provide a novel design for oil-water about 25% to about 75%. This is roughly what we ob- separation columns, where the interstitial geometry of serve in the experiments. the ne granulate provides a particularly large area of interaction between both phases. Furthermore, it will be interesting to investigate how universal the behaviour re- ported here turns out to be, in view of the rather simple mechanism behind it. In particular, it will be of great in- terest to investigate until how far away from symmetric heterogeneity this e ect persists. In other words: how A complementary approach which leads to a similar large has  +   to be for this e ect to break down? result may be gained from the number of oppositely wet 1 2 Moreover, many hysteresis e ects in uid displacement in contacts, as sketched on the right hand side of Fig. 5a. porous media should be revisited in view of the present If ' is the number fraction of beads wet by uid i, results. we have ' + ' = 1, and the fraction of like-wet con- i 2 2 2 2 tacts is ' = ' + ' = 2' 2' + 1. The fraction l 1 1 2 1 of oppositely-wet contacts is ' = 1 ' . If we plot o l ACKNOWLEDGEMENTS ' ' = 4' 4' + 1 into Fig. 4 as a proxy for the l o 1 necessity of topologocal rearrangements (dashed curve), we see that the tendency observed in the CPS curve hys- The authors appreciate generous nancial support teresis is quite well reproduced. from BP plc. within the ExploRe program. [1] R. A. Wooding, Scott W. Tyler, and Ian "White, \Con- [10] Tue Hassenkam, C. S. Pedersen, K. Dalby, T. Austad, vection in groundwater below an evaporating salt lake," and S. L. S. Stipp, \Pore scale observation of low salin- Water Resources Research 33, 1199{1217 (1997). ity e ects on outcrop and oil reservoir sandstone," Col- [2] Bert van Bocxlaer, Dirk Verschuren, Georg Schettler, loids and Surfaces A: Physicochemical and Engineering and Stefan Kr opelin, \Modern and eary Holocene mollusc Aspects 390, 179{188 (2011). fauna of the Ounianga lakes: implications for the paleo- [11] F. M. Orr, \Onshore geologic storage of CO ," Science hydrology of the central Sahara," J. Quaternary Sci. 26, 325, 1656{1658 (2009). 433{447 (2011). [12] Mike J. Bickle, \Geological carbon storage," Nature Geo- [3] W. G. Anderson, \Wettability literature survey- part 4: science 2, 815{818 (2009). E ects of wettability on capillary pressure," Journal of [13] Kamal Singh, Hagen Scholl, Martin Brinkmann, Marco Petroleum Technology 39, 1283{1300 (1987). DiMichiel, Mario Scheel, Stephan Herminghaus, and [4] Norman R. Morrow, \Wettability and its e ect on oil Ralf Seemann, \The role of local instabilities in uid in- recovery," Journal of Petroleum Technology 42, 1477{ vasion into permeable media," Scienti c Reports 7, 444 1484 (1990). (2017). [5] F. A. L. Dullien, Porous Media: Fluid Transport and [14] Julie Murison, Benot Semin, Jean-Christophe Baret, Pore Structure (Academic Press, 1991). Stephan Herminghaus, Matthias Schr oter, and Martin [6] M. Sahimi, Flow and Transport in Porous Media and Brinkmann, \Wetting heterogeneities in porous media Fractured Rock: From Classical Methods to Modern Ap- control ow dissipation," Phys. Rev. Applied 2, 034002 proaches (John Wiley and Sons, 2011). (2014). [7] Po-Zen Wong, Methods in the Physics of Porous Media, [15] Martin Blunt and Peter King, \Relative permeabili- Experimental Methods in the Physical Sciences, Vol. 35 ties from two- and three-dimensional pore-scale net- (Academic Press, San Diego, USA, 2013). work modelling," Tran. Porous media Journal 6, 407{433 [8] Arnaud Lager, Kevin John Webb, Ian Ralph Collins, and (1991). Diane Marie Richmond, \Losal enhanced oil recovery: [16] Martin Blunt, \Pore level modeling of the e ects of wet- Evidence of enhanced oil recovery at the reservoir scale," tability," SPE Journal 2, 494{510 (1997). SPE Journal (2008). [17] A. B. Dixit, J. S. Buckley, S. R. McDougall, and K. S. [9] T. Austad, S. F. Shariatpanahi, S. Strand, C. J. J. Sorbie, \Empirical measures of wettability in porous me- Black, and K. J. Webb, \Conditions for a Low-Salinity dia and the relationship between them derived from pore- Enhanced Oil Recovery (EOR) E ect in Carbonate Oil scale modelling," Tran. Porous media Journal 40, 27{54 Reservoirs," Energy & Fuels 26, 569{575 (2012). (2000). 7 [18] P al-Eric ren and Stig Bakke, \Reconstruction of Berea [24] C. Wylock, M. Pradas, B. Haut, P. Colinet, and S. Kalli- sandstone and pore-scale modelling of wettability ef- adasis, \Disorder-induced hysteresis and nonlocality of fects," SPE Journal 39, 177{199 (2003). contact line motion in chemically heterogeneous mi- [19] Xiucai Zhao, Martin J. Blunt, and Jun Yao, \Pore-scale crochannels," Physics of Fluids 24, 032108 (2012). modeling: E ects of wettability on water ood oil recov- [25] Mika M. Kohonen, Dimitrios Geromichalos, Mario ery," Journal of Petroleum Science and Engineering 71, Scheel, Christoph Schier, and Stephan Herminghaus, 169{178 (2010). \On capillary bridges in wet granular materials," Physica [20] I. Fatt, A. Waldemar, and Jr. Kliko , \E ect of frac- A 339, 7{15 (2004). tional wettability on multiphase ow through porous me- [26] M. J. Powell, \Site percolation in randomly packed dia," SPE Journal 11, 71{75 (1959). spheres," Phys. Rev. B 20, 4194 (1979). [21] M. Lessel, O. B aumchen, M. Klos, H. H ahl, R. Fetzer, [27] Mario Scheel, Ralf Seemann, Martin Brinkmann, Marco M. Paulus, R. Seemann, and K. Jacobs, \Self-assembled DiMichiel, Adrian Sheppard, Boris Breidenbach, and silane monolayers: an ecient step-by-step recipe for Stephan Herminghaus, \Morphological clues to wet gran- high-quality, low energy surfaces," Surface and Interface ular pile stability," Nature Materials 7, 189 (2008). Analysis 47, 557564 (2014). [28] Mario Scheel, Ralf Seemann, Martin Brinkmann, Marco [22] H. Perrin, R. Lhermerout, K. Davitt, E. Rolley, and DiMichiel, Adrian Sheppard, and Stephan Herminghaus, B. Andreotti, \Defects at the nanoscale impact contact \Liquid distribution and cohesion in wet granular as- line motion at all scales," Phys. Rev. Lett. 116, 184502 semblies beyond the capillary bridge regime," J. Phys.: (2016). Cond. Mat. 20, 494236 (2008). [23] R. Dufour, C. Semprebon, and S. Herminghaus, \Filling [29] Stephan Herminghaus, Wet Granular Matter: a Truly transitions on rough surfaces: inadequacy of Gaussian Complex Fluid, Series in Soft Condensed Matter, Vol. 6 surface models," Phys. Rev. E 93, 032802 (2016). (World Scienti c, 2013).

Journal

Condensed Matter – arXiv (Cornell University)

Published: Jan 15, 2018

There are no references for this article.