Surfactant Foam Selection for Enhanced Light Non-Aqueous Phase Liquids (LNAPL) Recovery in Contaminated Aquifers
- 135 Downloads
Foam is promising for the remediation of non-aqueous phase liquids (NAPL) source zones; however, the production of foam and its behavior in porous media are poorly understood. A methodology for the selection of surfactants suitable for foam production applied to NAPL remediation was developed. Two criteria were initially used for surfactant selection: foamability as evaluated by the Ross–Miles test and interfacial tension reduction measured with the pendant drop method. Three promising surfactants were identified and used in sand column tests: Genapol LRO because it produced the highest foam height in the Ross–Miles test, Ammonyx Lo which exhibited the lowest interfacial tension with p-xylene and had the second highest foam height, and Tomadol 900 because it showed intermediate results in both tests. Viscosity was found to be proportional to foamability. Genapol LRO produced a foam so viscous that it destabilized by the end of the experiment. Ammonyx Lo produced a less viscous foam but with a stable front. Tomadol 900 produced an unstable foam with poor viscosity. Results from column tests gave indications of optimal conditions needed to produce a stable and viscous foam front.
KeywordsFoam Surfactant NAPL recovery In situ remediation Ross–Miles test
Many studies on surfactant foams have been carried out in the petroleum industry for their application to enhanced oil recovery (EOR) in deep oil reservoirs (Li et al. 2012a; Farajzadeh et al. 2012; Shallcross et al. 1990; Simjoo et al. 2012; Nguyen et al. 2007). However, environmental applications for non-aqueous phase liquid (NAPL) source zone remediation under field conditions are quite different from EOR: NAPL-contaminated shallow aquifers are mostly unconsolidated porous materials, a low injection pressure must be used to prevent soil heaving, and the natural groundwater present in the pores is not saline. The selection of foam for environmental remediation (ER) differs from the one for enhanced oil recovery (EOR) for two main reasons: (1) the low toxicity and high biodegradability criteria used to screen surfactants selected for ER and (2) the fact that the addition of salt for ER applications is not always desirable, whereas in EOR it is common practice to optimize the foam formulation with sodium salt to reach the lowest interfacial tension (IFT) (Lake 1989). One issue with salt in ER is related to the difficulty of maintaining the optimal salinity required for low IFT in (fresh) groundwater of aquifers during surfactant injection. Another issue is that salt has to be removed from the wastewater produced during ER, contributing to an increase in effluent treatment costs. In EOR, saline water present in the petroleum reservoir helps to maintain the optimized conditions to maintain low IFT.
A few laboratory studies have been carried out to assess foam behavior in porous media for its application to the ER of shallow contaminated soils (Mulligan and Eftekhari 2003; Couto et al. 2009; Tanzil et al. 2002). The application of foams to the ER of light non-aqueous phase liquids (LNAPL) is studied here, whereas some of the previous works considered dense non-aqueous phase liquids (DNAPL), such as trichloroethylene (TCE), as the type of contaminant to be removed (Jeong et al. 2000, Jeong and Corapcioglu 2003, 2005; Rothmel et al. 1998; Pennell et al. 1996). For a DNAPL, mobilization has to be avoided to prevent the sinking which would worsen site conditions. Some field-scale tests of foam injection in shallow soils were also conducted (Hirasaki et al. 2000).
The work reported in this paper was undertaken in the context of an application of foam to remediate a LNAPL-contaminated aquifer. The objective of the study was to develop a methodology for: (1) the selection of surfactants suitable for foam production applied to p-xylene remediation, (2) the ex situ production of foam and (3) the characterization of the behavior of foam injected in a water-saturated sand column. So, the surfactant selection method proposed here does not use salinity scan with the surfactant and the contaminant in test tubes to find the optimal salinity as proposed by Hirasaki et al. 2000.
1.1 Enhanced NAPL Recovery Mechanisms with Surfactants and Foams
Foam flow through porous media is defined by Hirasaki (1989) as a “dispersion of gas in liquid such that the liquid phase is continuous (i.e., connected) and at least some part of the gas phase is made discontinuous by thin liquid films called lamellae.” Foam has non-Newtonian and shear-thinning properties (Hirasaki and Lawson 1985; Falls et al. 1989), which make foam an alternative to the use of polymers for mobility control (Martel et al. 1998, 2004; Robert et al. 2006). Also, the presence of air in foam reduces significantly the amount of surfactant needed to sweep a given aquifer volume, compared to the injection of an aqueous surfactant solution. Thus, when compared to surfactant solutions, the use of foam can lower the mass of surfactant needed and even replace the use of polymers. Thus, if the efficiency of foam for NAPL source zone remediation can be demonstrated, foam could represent an economic alternative to other NAPL source zone flushing technologies.
The increase in viscous forces is achieved by increasing the injected fluid viscosity \((\mu )\) and its flux (or velocity) (q) in the porous media, whereas a decrease in capillary forces is possible via a reduction in interfacial tension \((\sigma )\) by the presence of a surfactant. The wettability is represented by the contact angle (θ). Therefore, maximizing Nc increases the possibility to displace a contaminant with foam. Replacing water by foam as the displacing fluid in porous media favors NAPL mobilization because, compared to water, it increases viscosity and decreases interfacial tension by at least one order of magnitude (Hirasaki and Lawson 1985).
1.2 Foam Properties
Two properties were considered in this study to characterize foam-forming surfactants: foamability and foam quality. Foamability corresponds to the ability of the surfactant to create bubble and its ability to maintain this state over time. Foam quality corresponds to gas fractional flow used to generate foam. Foam has a dynamic nature; its properties evolve significantly through time. Furthermore, the same surfactant can produce foam with different properties depending on the production method. Consequently, in order to compare different surfactant candidates for foam production, the same foam production method and the same measurement conditions need to be used.
2.1 Surfactant Selection Methodology
First, p-xylene was selected as the model LNAPL representative of gasoline: density of 0.8611 kg/m3(lighter than water), viscosity of 0.644 mPa s, vapor pressure of 0.0086 atm (all these properties are at 20 °C; Johnson et al. 1990; Mercer and Cohen 1990). Also, p-xylene is an aromatic hydrocarbon that has 8 carbons (C8) and is a good surrogate for gasoline that is a mixture of tens of hydrocarbons (alkanes linear or branched, alkenes and aromatics) having a number of carbons ranging from C5 to C10. Finally, on a ternary phase diagram, looking at the miscibility of p-xylene or gasoline in water in the presence of surfactant, both show the same miscibility zone.
Three criteria were considered to select the most suitable surfactant for foam production: biodegradability, foamability and interfacial tension. Foamability was measured with the Ross–Miles test (ASTM D1173). Interfacial tensions between different concentrations of surfactant aqueous solutions (0.01%, 0.1% and 1% mass/mass) and p-xylene were measured by the pendant drop method (Woodward 2000), and the best candidates with the lowest interfacial tension with the contaminant were considered for further investigation.
List of tested surfactants
Active matter (% w/w)b
No. of CMC at
Sodium alkyl ether sulfate
3.3 × 10−2
1.1 × 10−3
Sodium laureth sulfate
4.5 × 10−2
6.2 × 10−3
96% (Mouton et al. 2009)
2.3 × 10−3 c
1.5 × 10−2
1.0 × 10−2
2.5 × 10−2
1.7 × 10−3
Octylphenol ethylene oxide
1.9 × 10−2
2.7 × 10−2
Polyoxyethylene sorbitan monooleate
1.4 × 10−3
For the purposes of this study, foamability is considered as the height of foam produced during a Ross–Miles test, as described by Li et al. (2012b). This test involves the use of a fixed volume of surfactant solution that falls in a receiving tube of a certain height into a constant volume of the same surfactant solution. This method had the following advantages: Each test was done in a short period of time, it was repeatable, and the effect of humidity can be neglected due to the sufficient apparatus height. As reported by Li et al. (2012b), the distance between the top of the foam and the receiver top opening was crucial for the humidity gradient over the foam, which had an effect on foam collapse. When the humidity gradient is low, the evaporation rate is also low and the foam was more stable. Considering that the best foam tested reached a height of 18.6 cm and the receiver total height of 90 cm, a distance of 71.4 cm was considered sufficient to have negligible evaporation. However, this technique did not consider the change in foam quality through time, only the foam height.
2.1.2 Foam Quality
Foam quality was measured at the outlet of the foam production column and at the outlet of the sand column. Foam was recovered in a graduated cylinder of 1L to estimate its total volume and the liquid produced after foam breakage. The gas volume is the difference between the total volume minus the liquid volume and the foam quality calculated according to equation above.
2.1.3 Foam Viscosity
2.1.4 Interfacial Tension
IFTs between surfactant aqueous solutions and p-xylene were measured with a FTA 200 Dynamic Contact Angle Analyzer (First Ten Angstroms, Inc., Portsmouth, Virginia), which uses the pendant drop technique (Woodward 2000). Images were taken of a p-xylene drop injected upward with a “U”-shaped flat-ended needle (0.356 mm diameter) in a glass cell filled with the tested surfactant at concentrations of 0.01%, 0.1% and 1%. A computer-activated pump controlled p-xylene injection. A p-xylene drop needs to be large enough to be distorted by gravity as interfacial tension tries to balance this distortion. IFT was assessed by fitting the shape of the drop to the Young–Laplace equation (Lake 1989). The captured video image was calibrated with phase densities and with the size of the needle tip. FTA35 software was used to treat drop images and compute IFT values (First Ten Angstroms, Inc., Portsmouth, Virginia).
2.2 Foam Production and Injection System
The foam was produced by alternating injection of surfactant and air at the same pressure in a foam production column. With this method a constant flow of foam was easily produced and foam quality was easily measurable. Foam production at a constant pressure has been reported to be more efficient than production at a constant rate (Li 2011). Other methods were tested but not used, including the injection at a constant rate with syringe pumps and injection with a porous stone in a foam production column. Those methods did not perform well because no foam was produced and only alternate slugs of air and surfactant solution were generated.
Summary of column tests including the characteristics of each test; a surfactant concentration of 0.1% w/w was used for all sand column tests
Foam production column
Foam quality outlet of production column (% airb)
Injection pressure (cm H2O)
Foam quality outlet of sand column (% airb)
2.3 Sand Column Experiments
Sand column experiments are used to assess the suitability of foam for LNAPL remediation. Sand column tests were designed to help understand the effects of different experimental conditions on foam flow through porous medium, especially the use of the foam production column, injection pressure, surfactant concentration, pre-flush liquid prior to foam injection and surfactant types. Sand column experiments were performed with the three most suitable surfactants selected following foamability tests and interfacial tension measurements.
A coarse Temisca 20 (Opta Minerals Inc.) silica sand (99.9% quartz) having a d10 and a d50 (grain sizes larger than 10% and 50% of sand mass) of 0.75 and 1.3 mm, respectively, was used for all experiments to minimize fines (clay and silt) and organic matter contents of soils and the interaction of surfactant with other minerals.
Sand filling and compaction in columns were done following the method described by Martel and Gélinas (1996). Columns are made up of a transparent acrylic cylinder with 3.5 cm internal diameter and 14.5 cm length. Acrylic was selected for its resistance to compression and its transparency, which allowed the filming of foam flow. Both ends are sealed with a perforated Teflon cap combined with a reservoir that uniformly distributes fluids before entry into the sand column. The seal between the acrylic cylinder and the Teflon caps was provided by a Viton O-Ring. A nylon screen (125-μm mesh) in each Teflon cap prevents the loss of sand. Two holes were drilled on the side of the column to connect pressure transducers, and the same nylon screen was placed on each hole to prevent sand loss.
Compaction of each 5-mm sand layer was done by dropping a 500 g weight 12 times from a height of 8 cm. The top surfaces of compacted layers were lightly scarified to minimize preferential flow paths between subsequent layers. The sand column had a dry bulk density of 1.74 g/cm3 to prevent channeling (Ripple et al. 1973). Trapped air in the column was eliminated by circulating at least 30 pore volumes of CO2 through the column. Then, the column was saturated from the bottom up with degassed distilled water. At least 3 pore volumes (PV) of water were circulated through the sand column in order to flush or solubilize all CO2. The same column was used for all tests to produce comparable experiments. After each test, the column was rinsed with water and dried with compressed air. It was then purged with CO2 and saturated again with degassed distilled water. Doing so, the column did not have been repacked after each test. With this protocol the sand column used has a hydraulic conductivity to water of 4 × 10−4 m/s (permeability of 4 × 10−11 m2), a porosity of 0.34 and a corresponding pore volume of 47 mL.
The impact of the foam production column prior to foam injection in the sand column was evaluated with two column tests: one with a foam production column (Test 1) and one without (Test 2). Otherwise, both tests were done under the same conditions (Table 2) and results are shown in Fig. 5. For the test with the foam production column, a pressurized surfactant solution was first injected in the foam production column, followed by air. The resulting foam was then directed into the sand column connected downstream of the foam production column. In the test without the foam production column, the pressurized surfactant solution was injected directly in the sand column, followed by air. These tests were done to verify if the use of a foam production column before the injection of foam in the sand column would increase the foam apparent viscosity. Table 2 summarizes the sand column test conditions. The foam quality at the outlet of the foam production column was about the same (~ 86.5%) for all tested surfactants and was modified following its injection into the sand column [as measured at the outlet (Table 2)]. Note that in this type of injection (SAG) foam quality could vary among other test conditions depending on slug size and foam formation mechanisms.
In order to be representative of saturated conditions as found in the field, the sand column must be liquid saturated prior to foam injection. The effect of the pre-flush liquid in the sand column prior to foam injection was tested by comparing two alternatives: one using a column pre-flushed with water (Test 4) and one using a column pre-flushed with surfactant (Test 3), prior to foam injection. Both tests were done under the same surfactant at a concentration of 0.1% and foam injection pressure of 350 cm H2O. Results are shown in Fig. 6. The effect of surfactant type was evaluated by making three tests, each with a different surfactant solution (Test 1, Test 3 and Test 5) (Table 2). These three tests were done under the same foam injection conditions, and the results are presented in Fig. 8. Finally, the effect of injection pressure was evaluated by making two tests: one at a lower injection pressure (Test 3) and another at a higher injection pressure (Test 6). Other than injection pressure, both tests were done with the same surfactant and under the same foam injection conditions as shown in Fig. 11.
Foam apparent viscosity measurements were done for each column test to evaluate the effect of each distinct experimental condition. Viscosities were estimated via foam flow rate and pressure measurements. Injection pressure was maintained constant during each test. Foam flow rates were calculated by two indirect methods: the velocity of the foam front (front velocity method) and the foam flow rate at the system outlet (output method). The front velocity method involved the measurement of the foam front velocity when it passed the second pressure transducer in the column. Apparent viscosity was calculated using the pressure gradient between the two pressure transducers (PT-3 and PT-4) in the column at that time. This method provides an evaluation of foam viscosity at the beginning of the column test, before the column is completely swept by foam. The output method used foam outflow rate measured at the column outlet after a certain time of foam injection when the pressures measured in the column had stabilized. For this method, apparent viscosity was calculated using the stabilized pressure gradient between the two pressure transducers (PT-3 and PT-4) in the column and the measured flow rate at the column outlet. This method provides a stabilized measurement of foam viscosity just before the experiment ended.
3.1 Surfactant Selection
3.1.2 Interfacial Tension
3.2 Sand Column Tests to Study the Effect of Various Conditions
3.2.1 Effect of a Foam Production Column on Foam Stability
Foam stability during column tests is assessed through visualization of the foam front. A stable front shows a sharp demarcation between the foam- and water-invaded regions in the column. An unstable front shows an indistinct demarcation between the two regions, suggesting the front has collapsed. As foam was injected at the same pressure (350 cm water) in both cases, a more viscous foam would result in a lower flow velocity in the column. Relative flow velocity is indicated through the photographs of the advancing foam front as shown in Fig. 5.
Figure 5 indicates that the front in the test without a foam production column (top) moved significantly faster than in the test with a foam production column (bottom). This indicates a greater foam viscosity when a foam production column is used. Both experiments showed the formation of a foam with a stable displacement front even if the foam quality is different [a (weak) foam quality of 59.1% (top) and a strong foam at 93.5% quality (bottom) at the column outlet (Table 2)]. Evaluation of viscosity with the front velocity method indicated a value of 8 mPa·s without a foam production column, compared with 338 mPa s when using one. When the foam front passed the second pressure transducer, it was a lot more viscous in Test 1 than in Test 2. With the output method, viscosity estimates were 204 mPa s without foam production column and 254 mPa s with the foam production column. These results are of the same order of magnitude, thus indicating that when foam has stabilized and pressure is constant, both foam production methods tend to show a similar apparent viscosity.
3.2.2 Effect of Water or Surfactant Pre-flush
Figure 6 shows that, with a surfactant pre-flush, a straight uniform front is formed. For the water pre-flush experiment, a blurred foam front (made of air) was first formed, resulting in poor sweep, followed by a second front with more viscous foam advancing more slowly with better sweep efficiency. However, both this second, more stable front and the front observed in the surfactant pre-flush experiment took the same time to sweep the column, indicating that foam viscosity was actually similar. For the water pre-flush tests, the apparent viscosity of the foam was evaluated with the front velocity method using the second foam front that was more clearly defined. In fact, the apparent foam viscosities estimated, respectively, by the front velocity or the output methods were 40 or 44 mPa s when water was used as pre-flush fluid and 54 or 59 mPa s when surfactant B solution 0.1% was used as pre-flush fluid.
3.2.3 Effect of Surfactant Type
3.2.4 Effect of Injection Pressure
The front velocity method viscosities of 54 and 3 mP·s were recorded under shear rates of 109 and 565 s−1, respectively (Eq. 3). This confirms that at these injection pressures, the foam exhibits a shear-thinning non-Newtonian behavior.
For surfactant selection, the Ross–Miles test and the pendant drop technique are two simple and fast methods for the evaluation of surfactant foamability and interfacial tension, respectively. Column test results were in agreement with indications provided by the Ross–Miles test for the three surfactants tested, as the surfactant solution with the best foamability had the most viscous foam front, whereas the one with the worst foamability showed the least viscous foam front. Hence, the Ross–Miles test results can be used as a screening tool for foam selection. The other foam selection criterion based on the lowering of interfacial tension is not indicative of the behavior of foam when in contact with p-xylene. However, it can be used to maximize the capillary number achieved during the foam flood. Viscous foam provides NAPL displacement and a better sweep efficiency of the porous media via favorable mobility ratio (stable displacement), while interfacial tension decrease allows the contaminant to be drained, moved and recovered easier.
A third foam selection criterion evaluating the direct effect of p-xylene on foam would have been nice to have (not performed in this paper) as a selection criterion. This criterion would measure the production and collapse of foam without contact with p-xylene and when p-xylene is mixed with surfactant solution prior to foam production, therefore quantifying the effect of p-xylene on foamability as presented by Simjoo et al. (2012).
Summary of column test results and conclusions
Effect of foam production column
Without a foam production column, foam front is blurred and viscosity is low
With foam production column, foam front is sharp and foam viscosity is high
Foam production column is necessary
Water pre-flush creates an air front with poor sweep efficiency before the foam front
Surfactant pre-flush creates a sharp and stable foam front
Surfactant pre-flush instead of water is recommended
Surfactant A created a highly viscous foam that was destabilized at the end of the test, and surfactant I created a front with very low foam viscosity
Surfactant B created a sharp and stable foam front
Surfactant B was selected
An injection pressure of 210 cm of water created an unstable foam front with low foam viscosity
An injection pressure of 350 cm of water created a sharp and stable foam front
High injection pressure is recommended
The output method for foam viscosity measurements required a minimum of 5 mL of foam, and it took on average several minutes to collect this volume. A small flow rate combined with the fact that foam collapses over time made it difficult to obtain accurate measurement of foam flow. When the foam viscosity was very high, with surfactant A, the foam flow rate was very low and the measurement of foam flow rate was difficult because not enough foam exited the column. In this case, foam flow was underestimated, which increased calculated viscosity. Also, when foam viscosity was very low, with surfactant I, foam tended to collapse faster than with other surfactants. This often produced slugs of air mixed with foam at the column outlet, which could not be monitored or included in foam flow rate measurements. However, those slugs provided clues that this foam had a poor stability. Those considerations indicate that all measurements made with the Output Method underestimated the volume of foam coming out of the column and therefore overestimated foam viscosity.
In this study, an injection pressure as high as 350 cm of water (0.34 bar) was required to flood a 14.5-cm-long homogeneous coarse sand column (corresponding to 2.3 bar/m) according to the minimal pressure gradient models for strong foam generation considering a porous medium of 40 µm2 (Ransohoff and Radke 1988; Gauglitz et al. 2002; Rossen and Gauglitz 1990; Del Campo Estrada et al. 2015). More work needs to be done to minimize foam injection pressure and to produce a stable foam front in the porous media, related to the pressure gradient constraint in ER (0.9 bar/m) (Maire and Fatin-Rouge 2017). Also, it means that a higher foam injection pressure would be expected at the field scale in heterogeneous sand having a lower permeability. The use of other gases such as CO2 or N2 instead of air to generate foam, the use of other surfactants or the injection of particles to stabilize the foam are options that may be part of the solution.
The general objective of this study was to develop a methodology for the selection of a suitable surfactant to make foam for the remediation of LNAPL-contaminated aquifers. Surfactant solutions were tested to compare their foaming properties with the Ross–Miles test and their interfacial tension with p-xylene using the pendant drop method. Three surfactant solutions were chosen and tested in a homogeneous sand column. Column tests were also done to evaluate injection conditions needed to obtain a stable and viscous foam front: the use of a foam production column, sand column pre-flush with either water or surfactant solution prior to foam injection and injection pressure.
The Ross–Miles test proved to be suitable for the identification of surfactant solutions having good foaming properties. Column tests indicated that the surfactant solution with the best foamability in Ross–Miles test (Genapol LRO) formed a more viscous foam front during column injection compared to the surfactant solution with less foamability (Ammonyx Lo). Also, the surfactant solution with the worst foamability of the three (Tomadol 900) presented an unstable front with very low viscosity. The Ross–Miles test being simple and fast, it is an interesting tool to rank surfactant solutions in terms of foaming ability. The other column tests gave indications on conditions to consider for foam injection in further studies: a high injection pressure, the use of a foam production column and the column pre-flush with surfactant solution prior to foam injection.
The tests reported in this study for surfactant selection and foam behavior observed in column are only the first steps toward the field application of surfactant foam injection to remediate LNAPL-contaminated aquifers. Other tests are needed to: (a) evaluate the impact of horizontal injection through soil layers on foam front behavior and (b) verify the effect of LNAPL on foamability.
Funding for this study was provided by a NSERC-CRD Grant (412730-11) in partnership with TechnoRem and a NSERC Discovery Grant (326975-2011) held by Richard Martel. The authors would like to gratefully acknowledge the donation of surfactant samples by Stepan, the help of Mr. Luc Trépanier from INRS in designing and installing the foam injection system, of Mr. Marco Boutin from INRS for figures and tables improvements, and of Dr. Clément Portois for internal review. We acknowledge reviewers (M. Babs Oyeneyin from Robert Gordon University and anonymous) for comments and suggested corrections.
- Bear, J.: Dynamics of Fluids in Porous Media. Dover, New York (1972)Google Scholar
- Chauveteau, G., Zaitoun, A.: Basic rheological behavior of xanthan polysaccharide solutions in porous media: effects of pore size and polymer concentration. In: Foyers, F.J. (ed.) First European Symposium on Enhanced Oil Recovery, Bournemouth, England, pp. 197–214. Elsevier, Amsterdam (1981)Google Scholar
- de Marsily, G.: Quantitative Hydrogeology—Groundwater Hydrology for Engineers. Academic Press, San Diego (1986)Google Scholar
- Hirasaki, G.J., Jackson, R.E., Jin, M., Lawson, J.B., Londergan, J., Meinardus, H., Miller, C.A., et al.: Field demonstration of the surfactant foam process for remediation of a heterogeneous aquifer contaminated with DNAPL. In: Fiorenza, S., Miller, C.A., Oubre, C.L., Ward, C.H. (eds.) NAPL Removal: Surfactants, Foams, and Microemulsions, Part 1. AATDF Monograph Series, pp. 3–163. Lewis Publishers, CRC Press, Boca Raton (2000)Google Scholar
- Johnson, P.C., Stanley, C.C., Kemblowski, M.W., Byers, D.L., Colthart, J.D.: A practical approach to the design, operation, and monitoring of in situ soil-venting systems. Groundw. Monit. Remed. 10(2), 158–178 (1990)Google Scholar
- Lake, L.W.: Enhanced Oil Recovery. Prentice Hall, Englewood Cliffs (1989)Google Scholar
- Li, R.F.: Study of foam mobility control in surfactant enhanced oil recovery processes in 1-D, heterogeneous 2-D, and micro model systems. PhD Thesis, Rice University, Houston, TX (2011)Google Scholar
- Portois, C.: Comportement de la mousse en milieu poreux pour confiner une source de pollution: potentialités, contraintes et demonstration en site reel. PhD thesis, Université Bordeaux Montaigne (2018)Google Scholar
- Tanzil, D., Hirasaki, G.J., Miller, C.A.: Conditions for Foam Generation In Homogeneous Porous Media. SPE 75176 (2002)Google Scholar
- Woodward, R.P.: Surface Tension Measurements Using the Drop Shape Method. First Ten Angstroms (2000)Google Scholar