Partition Coefficients and Interfacial Activity for Polar Components in Oil/Water Model Systems

Journal of Colloid and Interface Science 212, 33– 41 (1999) Article ID jcis.1998.5988, available online at http://www.idealibrary.com on

Partition Coefficients and Interfacial Activity for Polar Components in Oil/Water Model Systems Sylvi Høiland Standal,* ,1 Anne Marit Blokhus,* Jostein Haavik,* Arne Skauge,† and Tanja Barth* *Department of Chemistry, University of Bergen, Alle`gaten 41, N-5007 Bergen, Norway; and †Norsk Hydro Research Center, Bergen, Norway Received May 4, 1998; accepted November 18, 1998

phase, or otherwise cross through the water film, eventually to be adsorbed onto the rock surfaces and cause wetting alteration of the reservoir rock (4). The wettability of reservoir rocks, that is, the tendency of one fluid to spread on or adhere to a solid surface in the presence of other immiscible fluids (5), is a determining factor with regard to distribution and permeability of crude oil in a reservoir and depends on the interfacial tensions among the three phases present. Although considerable amount of work has been done to determine the influence of different factors such as pH, salinity, crude oil composition, temperature, and pressure on the wettability of reservoir rocks, this is not fully understood. Several review papers are available on this topic (6 –9). In recent years several authors have investigated the partitioning of organic compounds between two immiscible liquids as a function of pH (10 –14). The distribution coefficient is found to be strongly pH dependent. The decrease in IFT between acidic oil solutions and water of increasing pH has also been investigated by several authors (12, 15–19). The effect is observed to approximately correspond to the dissociation of the acidic functional groups with increasing pH. In order to understand the processes acting in a reservoir it is crucial to explore the properties of the systems on a molecular level. In this work a description of how the chemical properties of molecules in the system govern their distribution between the phases is considered. The natural COBR consists of complex mixtures of fluid and rock, and investigating the effects of small compositional variations often gives results that are difficult to interpret. This work is therefore concentrated on simple model systems in order to develop an analogy to more realistic systems. The model system consists of isooctane modeling the oil phase, and molecules of moderate molecular weight to represent the polar organic components. The substances investigated are 1-naphtoic acid, 5-indanol, and quinoline, i.e., strong and weak acids, and base, all representative of the aspects that have been found to be important for the interactive properties in crude oil. The components chosen represent functions that are assumed to be active in high-molecular-weight oil components, such as asphaltenes, but still can be considered well-defined model substances.

Partition coefficients, surface tension, and interfacial tension for some polar organic components dissolved in oil/water model systems have been investigated. The systems consist of isooctane modeling the oil phase and of water solutions of NaCl and CaCl 2 modeling the water phase. The organic compounds examined were 1-naphtoic acid, 5-indanol, and quinoline, all well-defined molecules known to be representative of polar components in crude oil. The dependence on pH, salinity, and ionic strength in the water phase was investigated. The surface tension and interfacial tension were also examined as a function of component concentration. The results show a connection between the distribution of the polar components and the interfacial tension. Correspondence between the partition coefficient and the pK a value for the components is also reported. For 1-naphtoic acid none of the two ionization forms of the molecule are found to be surface active in aqueous solution. For 5-indanol both forms are surface active, and for quinoline only the nonionic form of the molecule is found to be surface active. The results indicate that the aqueous phase is the one that governs the interfacial tension. Increasing salinity increases the concentration of the component in the oil phase and decreases the interfacial tension between the oil phase and the aqueous phase. The results are explained due to the “salting-out” effect and to changes in the electrostatics for the various systems. © 1999 Academic Press Key Words: interfacial tension; surface tension; partition coefficient; pH effect; salinity effect.

INTRODUCTION

The presence of polar organic components in crude oil affects the complex rock/fluid and fluid/fluid interactions in crude oil-brine-rock (COBR) systems. These polar compounds are determining factors with regard to the solubility and stability of large molecules in the oil (such as asphaltenes), for example, and also with regard to the interfacial activity in the three-phase COBR system. The crude oil composition is thus found to be crucial for COBR interactions (1). Polar compounds in crude oil that are generally believed to be responsible for surface interactions include carboxylic, phenolic, and indolic acids and bases such as pyridine (2, 3). Oil components may be extracted from the oil into the water 1

To whom correspondence should be addressed. 33

0021-9797/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.

34

STANDAL ET AL.

To model the water phase we use simple solutions of NaCl and CaCl 2, varying pH, salinity, and ionic strength. The main objective of this present work is to demonstrate the importance of the transfer of different polar components between oil and water with regard to the interfacial activity between the two liquid phases. The connection between distribution and interfacial tension is reported. Surface activity is examined to support the findings, giving an indication of the affinity for the two liquid phases, and stating the importance of molecule dissociation.

TABLE 1 Specifications for the Different Polar Components Used in the Experiments

Organic compound

Supplier and purity

Solubility in water (w) and isooctane (o) a

1-Naphtoic acid

Fluka

4 3 10 24 m (w) b

.97%

;2 3 10 23 m (o)

Aldrich

2.3 3 10 22 m (w)

99%

;5 3 10 22 m (o)

Fluka

7 3 10 22 m (w) c

.97%

;3.7 m (o)

Structural formula and dissociation constant d

THEORETICAL BACKGROUND

Surface and Interfacial Tension

5-Indanol

In a system consisting of two phases, a and b, the molecules in the interface region between the two will differ energetically from the molecules in the bulk phases due to a different molecular environment. Intermolecular interactions will result in an intermolecular interaction energy of the molecules in the interface region that is different than that in either bulk phase. The relative hydrophilic character of a given polar organic compound, that is, the affinity for either the hydrophilic or the hydrophobic phase present, is governing its surface or interfacial activity in aqueous solution. This hydrophilicity is strongly connected to the dissociation constant of the compound. The dissociation constants for both the acids and the base examined in this work are given in Table 1 and are defined for the reaction written in the direction of decreasing protonation: HA 5 H 1 1 A 2 BH 1 5 B 1 H 1

Ka 5 Ka 5

[H 1][A 2] [HA]

[B][H 1] . [BH 1]

[1] [2]

Accumulation of surface-active molecules at the interface between a and b will cause a reduction in the surface tension. This reduction may quantitatively be represented by Gibbs adsorption equation (20), 2

1 RT

S

D

­g 5 G i, ­ ln C i

1 . Gi N

a

Approximate values from visual solubility experiments (no pH adjustment, 20°C). b Halford, J. O., Am. Soc. 53, 2939 (1931) (24°C, the solubility corrected for dissociation). c Jones, G., “Quinolines. Part 1” (G. Jones, Ed.), Wiley, London, 1977 (no pH adjustment reported, 23.5°C). d “CRC Handbook of Chemistry and Physics” (Lide, D. R., Ed.), 71st ed., CRC Press, 1990 –1991. e Kortu¨m, G., Vogel, W., and Andrussow, K., “Dissociation Constants of Organic Acids in Aqueous Solution,” Butterworths, London, 1961.

According to the literature one might expect a reduction in the IFT values between oil solutions of water-insoluble fatty acids and aqueous solutions with increasing pH due to dissociation. This has been reported by several authors (15, 17–19). Similarly, the dissociation of water-insoluble bases will result in low IFT values at low pH (19). At crude oil/brine interfaces both acids and bases are active, thus IFT is highest near neutral pH and decreases as pH is either increased or decreased (7, 16).

[3] Partition Coefficient

where C is the total bulk concentration of the component i, g is the surface tension, and G i is the surface excess of component i (relative to the solvent). Thus if (­g/­ ln C i ) is negative, G i is positive, and we have an actual excess of the component present at the surface. If G i is negative there is a surface deficiency of component i. The molecular area, A i, of component i at the surface may be calculated from Ai 5

Quinoline

[4]

When a polar organic compound dissolved in an organic solvent is in contact with an immiscible aqueous solvent, the component will distribute itself between the two phases depending on pH, salinity, and ionic strength in the water phase. The partition coefficient is defined as K;

C (oil) , C (water)

[5]

where C (mol/kg) equals the equilibrium total concentration of

O/W MODELS: DISTRIBUTION AND INTERFACIAL BEHAVIOR

every form of the molecules (e.g., HA, A 2, and dimers). K 5 1 thus corresponds to equal amounts of component in the oil and water phase. MATERIALS AND METHODS

Chemicals The properties of the organic compounds used are summarized in Table 1. Isooctane was supplied by Merck (.99%, for chromatography) and by Lab-Scan (GC 99.5%). Sodium chloride and calcium chloride-dihydrate were both supplied by Merck. Sodium hydroxide and hydrogen chloride were used for adjusting to the right pH value. The buffer solutions used for pH calibration were pH 4.00 6 0.02 (citrate-hydrochloric acid), pH 7.00 6 0.02 (phosphate), pH 9.00 6 0.02 (boric acid/potassium chloride-sodium hydroxide), and pH 10.00 6 0.05 (boric acid/potassium chloride-sodium hydroxide), all supplied by Merck. All the chemical compounds were used without any further purification, and the measurements were performed at room temperature, 20°C. A GC-MS chromatogram was made of quinoline as the presence of impurities would be crucial in interpreting the interfacial tension results. The chromatogram indicated a content of 99.6% quinoline, along with some isoquinoline and a pure hydrocarbon. Methods The organic compounds were initially dissolved in the oil phase before equilibration with the aqueous phase. The solutions were prepared from equal amounts of each phase by weight, shaken firmly, and left to settle for at least 24 h before measurements were performed. The partitioning of the components versus equilibrium time was checked by measurment of the concentrations in each phase after 24, 48, and 72 h. No changes in the concentrations were detected, and 24 h was thus used as sufficient equilibration time. pH Measurements The pH measurements were performed by a ROSS combination pH electrode (0 –14 pH, glass body, Model No. 8102) connected to an Orion pH meter (Model EA 940) supplied by Metric Analysis, Norway. The aqueous solutions were adjusted to the chosen pH value and stored for at least 8 h before final corrections. The pH was then measured in the aqueous phase after equilibrium was established in the two-liquid solutions. Distribution Measurements The distribution measurements were performed using a fixed concentration of the organic component, and varying pH, ionic strength, and salinity in the aqueous phase. The concentrations were selected based on the relative solubility of the components in each liquid phase (assuring that the experiments were

35

performed well below the solubility limit), and on avoiding the saturation limit for the UV detector. From this, the initial oil concentrations used in the experiments were as follows: C(1naphtoic acid) 5 2.7 3 10 24 mol/kg, C(5-indanol) 5 2.54 3 10 23 mol/kg, and C(quinoline) 5 1.00 3 10 23 mol/kg. All the polar organic components examined are UV active, and the quantification of the different components in solution was performed by a Lamda6/PECSS double-beam UV/visible spectrometer system from Perkin Elmer. The observed l(max) for each component was 300 nm for 1-naphtoic acid (21a), 280 nm for 5-indanol (21b), and 270, 300, and 313 nm (the latter used in these experiments) for quinoline (21c). The component concentration was measured in both the oil bulk phase and the water bulk phase, and the total amount was compared with the fixed original concentration in the oil phase. The results show good correspondence, though some portion of the component examined is assumed to accumulate at the interface. This agreement made it possible to ignore the fact that dissociated and undissociated forms of the acid molecules slightly differ in UV absorbtivity, which is especially true for 1-naphtoic acid. The two dissociation forms of quinoline differ quite substantially in UV absorbtivity. For the nonionic form at low pH values, only one peak (at 313 nm) can be observed, and the peak intensity is significantly increased compared to equal concentrations of molecules in the ionic form. This phenomenon is described in more detail by Jones (22). For quinoline the UV absorbtivity is thus pH dependent. However, in this work, we observed that for these low pH values the UV absorbtivity in the oil phase was minimal. We could then assume that approximately all of the molecules partitioned into the aqueous phase, giving a partition coefficient of zero. For all the systems investigated the effect of pH (and salinity) on the UV absorbtivity could then be neglected. Calibration curves were obtained for both isooctane and water as solvents. The reproducibility of the distribution experiments was very good, approximately 61%. Surface and Interfacial Tension Measurements Interfacial tension (IFT) measurements in the systems were performed by the ring method, using a Du Nou¨y ring connected to a KSV Sigma 70 tensiometer (supplied by KSV Chemicals, Finland). For the surface tension (ST) measurements a Wilhelmy plate connected to the same equipment was used. The methods are described in detail elsewhere (23). The IFT and ST for each solution were measured continuously until a stable and reproducible value was obtained. The given values for IFT and ST thus represent equilibrium values. Some uncertainties are connected to the results from the IFT measurements. Low concentrations and some degree of concentration changes from using the same solutions in distribution, pH, and IFT measurements can result in minor errors in the values. High salinity solutions (e.g., 1 M NaCl) were problematic, as the salt seemed to stick on the ring. The use of

36

STANDAL ET AL.

FIG. 1. Calculated concentrations of 1-naphtoic acid in different forms in the isooctane and water phase as a function of equilibrium pH in the aqueous phase.

CaCl 2 (aq) resulted in precipitation of Ca(OH)2 (s) for the more alkaline solutions and made IFT measurements impossible. Nevertheless, the trends observed are unambiguous, and the results from the interfacial tension measurements are reproducible to a reasonable degree. The IFT and ST were generally determined within a range of uncertainty of 60.5 mN/m. Uncertainties within a range of IFT 62 mN/m were observed for systems containing high salinities. RESULTS AND DISCUSSION

Partition Coefficients Knowing the total concentration of the component in each phase and the pK a value it is easy to calculate the concentration of each form of the molecule as a function of pH in the aqueous solution. This is illustrated for 1-naphtoic acid in Fig. 1, where the concentration of ionized and unionized acid in water and also the total concentration of the acid in both phases are given as a function of equilibrium pH in the aqueous phase. (The concentration of unionized acid in the dimer form in the oil phase could not be calculated, as the dimerization constant is not known.) It is thus possible to study the partitioning of, e.g., only the unionized acid in the two-liquid system, provided that the dimerization constant is known. However, in this present work only the total concentration of the component in each phase was studied, and the partition coefficient was defined according to Eq. [5]. Thus the intercept between the C total (aq) line and the C total (o) line in Fig. 1 corresponds to K 5 1, and the intercept between the C A2 (aq) line and the C HA (aq) line corresponds to the pK a value. Rudin and Wasan (10) have presented a detailed theoretical study of the surface chemistry in acidic oil/alkaline solution systems, looking at the partitioning of acid between the two phases. In an experimental study

by Rudin and Wasan (11) the intercept between the acid concentration in oil and the acid anion concentration in water is defined as the partition coefficient equal to unity, which in their oleic acid/oil/water system is at a pH value several units above the pK a value in water. The measured partition coefficients for 1-naphtoic acid, 5-indanol, and quinoline as a function of pH are given in Figs. 2A, 2B, and 2C, respectively. As can be seen, 1-naphtoic acid tends to concentrate in the oil phase for the lower pH values and to transfer to the water phase when pH is increased. In fact, the calculated K 5 2 indicates that the concentration in the oil phase is two or three times higher than the concentration in the water phase for the lower pH values, depending on the salinity of the water phase. 1-Naphtoic acid has a pK a value at 3.70 (24). The transfer of the component from oil phase to aqueous phase corresponds with the pK a value. In fact, the pK a value corresponds to K 5 1, that is, for equal amounts of the component in the two liquid phases (illustrated in Fig. 2A). Phenols are weak acids, and their pK a values are rather high (pK a 5 9.89 for phenol itself (24)). The transition of 5-indanol from the oil phase to the water phase, presented in Fig. 2B, is observed starting at pH ; 9 (depending on salinity). As for 1-naphtoic acid, K 5 1 corresponds to the pK a value (pK a 5 10.3 for 4-indanol (25)). Quinoline is weakly basic (the pK a value is 4.90 (24)) and is present in dissociated form at the lowest pH values. Figure 2C shows that the component tends to concentrate in the water phase at the lower pH range, and the transition to the oil phase corresponds with the pK a value, again at K 5 1. The partition coefficients are significantly higher for the undissociated molecules in this system than in the previous systems. The concentration in the oil phase is 10 to 14 times the concentration in the water phase (corresponding to over 90% of the total amount of quinoline molecules), thus considerably more of the component transfers to the oil phase compared to the acidic compounds examined. This is due to the difference in polarity for the molecules investigated. Quinoline in its undissociated form is less polar than the carboxylic acid and the phenol and has therefore a greater affinity for the oil phase. In all cases the partition coefficients are found to be closely related to the dissociation of the components. Salt Effects In addition to the pH dependence, Fig. 2 shows the effect on K from increasing salinity in the water phase, that is, when no salt is present compared to when 0.5 M NaCl is added to the water phase. The dependence of K with increasing salt concentration is presented in Table 2. It is clear that increasing concentrations of monovalent salt in the water phase increase the concentration of the components in the oil phase. This is consistent with the findings of Bennett and Larter (26), who observed increased phenol preference for the petroleum phase at higher brine salinity. The increase in partition coefficient

O/W MODELS: DISTRIBUTION AND INTERFACIAL BEHAVIOR

37

FIG. 2. Variations in interfacial tension, IFT, and distribution coefficient, K, for the various systems investigated as a function of equilibrium pH and salinity. (A) 1-Naphtoic acid, (B) 5-indanol, and (C) quinoline. Denotation: F, IFT; E, K without salt added; , IFT; ƒ, K with 0.5 M NaCl(aq).

with increasing concentrations of monovalent salt was also observed by Javfert et al. (13). The observations are due to the well-known salting-out effect of organic compounds in aqueous solution. The aqueous solubility of petroleum hydrocarbon species in brines of varying salinity was investigated by Price (27), who found that the aqueous solubility decreased with increasing salinity. For calcium chloride in the aqueous phase, the observed effect is even stronger than for the corresponding concentration of sodium chloride (Table 2). This is in agreement with the observations of Jafvert et al. (13), who found that the partition

coefficient is significantly higher with divalent salt in the aqueous phase compared to monovalent salt at equal concentrations. Surface Activity The surface tension for 1-naphtoic acid, 5-indanol, and quinoline was measured as a function of concentration in aqueous solutions, and of concentration in isooctane solutions. The concentration ranges investigated were based on the relative solubility of the compound in the solvent (given in Table 1).

38

STANDAL ET AL.

TABLE 2 Oil-Water Partition Coefficients for the Compounds Examined as a Function of Salinity and Ionic Strength Salt concentration

K (1-naphtoic acid) at pH 2.5

K (5-indanol) at pH 2.5

K (quinoline) at pH 12

0 M salt 0.1 M NaCl 0.5 M NaCl 1 M NaCl 0.5 M CaCl 2

1.76 2.21 2.49 2.90 2.81

1.96 2.01 2.35 2.75 2.56

10.80 11.30 13.20 — 13.70

For the aqueous solutions, two pH values, 2.5 and 12, were chosen to correspond to complete dissociated and undissociated forms of the molecules, so that the difference in surface activity for the two dissociation forms could be examined separately. For all the three components examined the surface tension in isooctane was observed to be constant, approximately equal to 18.7 mN/m, with increasing concentration. This indicates no accumulation of molecules at the surface. The affinity for the hydrophobic bulk phase thus keeps the molecules in solution. Possible dimer formation in this hydrophobic environment, which is especially significant for the carboxylic acid, will result in reduced polarity and, thereby, reduced surface activity. The surface tension of aqueous solutions of 1-naphtoic acid showed no effect from either pH or concentration changes and is therefore not presented. For both pH 2.5 and pH 12 the surface tension remained constant at approximately 71 mN/m throughout the whole concentration range, and no accumulation of molecules at the surface was observed.

In a study of an o-xylene/water/octanoic acid system, Lord et al. (12) found that the acid form (HA) of the octanoic acid molecule was more surface active than the octanoate form (A 2). In addition, Rudin and Wasan (11) have observed that unionized oleic acid by itself is surface active. However, in this work the poor solubility of 1-naphtoic acid in both the oil phase and the water phase makes the investigated concentration range very narrow. This may be the reason why no significant effects on surface tension from concentration variations are observed in the system. The surface tension for 5-indanol in aqueous solution is presented in Fig. 3A. It is clear that the surface tension of 5-indanol decreases with increasing concentration at both pH values, meaning that both forms of the molecule accumulate at the surface. Calculating the molecular area at the surface from Eq. [4], we find that A equals 39 Å 2/mol for the acid form of the molecule, and 67 Å 2/mol for the acid anion form. An explanation for this difference in molecular area may be electrostatic repulsion between acid anions, and also hydration. For quinoline in aqueous solution the surface tension results are presented in Fig. 3B. At pH 2.5 the surface tension remains constant, approximately equal to 71.2 mN/m, throughout the whole concentration range. At pH 12 the surface tension decreases with increasing concentration. From the surface excess G i , the surface molecular area is calculated, and a value of 41Å 2/mol is found for the nonionic form of the molecule. The high hydrophobic character of this nonionic form of the molecule thus causes an affinity for the surface rather than for the polar bulk phase. In summary, for 1-naphtoic acid none of the two ionization forms of the molecule are found to be surface active in aqueous solution, for 5-indanol both forms are surface active, and for quinoline only the nonionic form of the molecule is surface active.

FIG. 3. Interfacial tension as a function of pH and water bulk concentration (0 M salt) at equilibrium conditions. Also shown is the aqueous surface tension as a function of pH and component concentration. (A) 5-Indanol and (B) quinoline. Denotation: F, ST/IFT at pH 12; E, ST/IFT at pH 2.5.

O/W MODELS: DISTRIBUTION AND INTERFACIAL BEHAVIOR

Interfacial Activity as a Function of Concentration In a two-liquid system the different compounds examined will enhance affinity for both phases, depending on the relative hydrophobicity of the molecules, concentration, pH, and salinity. (For simplicity no inorganic salts were introduced at this stage.) The interfacial tension for the oil/water/1-naphtoic acid system remained constant throughout the whole concentration range for both pH 2.5 and pH 12 and is therefore not presented. The lack of change in the IFT from increasing concentration, and also from dissociation of the molecules, is most likely caused by the very narrow concentration range examined (due to the low solubility). The interfacial tension for 5-indanol as a function of concentration and pH is presented in Fig. 3A. The IFT at both pH 2.5 and pH 12 decreases with increasing concentration, implying that both dissociation forms of the molecule are interfacial active. Assuming that the partition coefficient is independent of concentration below the solubility limit for each phase, the concentration of 5-indanol in the water phase of the equilibrated system can be calculated. As long as the total amount of molecules distributed between the two liquid phases is unchanged, and the solutions are prepared in 1:1 by weight, the following applies: C~water)equilibrium 5 C~oil)initial/~K 1 1!. At pH 12, K 5 0 (Fig. 2B), and at pH 2.5, K 5 1.95, meaning that for the same initial oil concentration in the system, a larger amount of the molecules is accumulated in the water bulk phase at pH 12 than at pH 2.5. By comparing IFT values for an equal amount of component in the bulk water phase, rather than comparing equal total amounts in the two-liquid system, we find that approximately the same IFT values are observed for the same water bulk concentrations at the two pH values (Fig. 3A). This result implies that the interfacial tension is independent of the concentration in the oil phase, and that the two dissociation forms of the molecule seem to exhibit a nearly equal effect on the interfacial tension. Figure 3A further shows, by comparing the IFT values and the ST values, that the ST of 5-indanol in aqueous solution decreases in parallel with the IFT with increasing concentrations. As mentioned earlier, ST of 5-indanol in isooctane remained constant with increasing concentrations. This observation is in agreement with Lord et al. (12) who investigated the o-xylene/water/octanoic acid system. Together, these two observations imply that the aqueous concentration of the acid is the operative variable governing the interfacial tension in the isooctane/water/5-indanol system. The IFT results for quinoline are presented in Fig. 3B. As for 5-indanol the IFT is plotted as a function of the calculated component concentration in the aqueous phase. At pH 2.5 we find that the IFT is weakly decreasing with increasing concen-

39

tration in the aqueous phase. At pH 12 this decrease is substantially more significant. That is, by comparing equal water bulk concentrations, we find that from the aqueous phase the nonionic form of the molecule is significantly more interfacial active than the corresponding ionic form. The observations are in agreement with the results from the aqueous ST measurements (also presented in Fig. 3B), indicating that the aqueous phase is the one that governs the IFT, and that the IFT is mainly affected by the nonionic form of the molecules. Interfacial Activity as a Function of pH Keeping the component concentration fixed in the two-liquid system and varying the pH in the aqueous phase, it is important to consider the variation in concentration due to distribution of the component between the two phases. These experiments thus combine pH and concentration dependence of the systems. At this stage we also include the dependence of salinity and ionic strength to the systems. The concentrations used in these experiments are the same as the concentrations used in the distribution measurements. The interfacial tension of the isooctane/water system was fairly constant throughout the whole pH range, at approximately 44 mN/m. However, at high pH values (above pH 12 for the system without added salt, and above pH 11.5 for the system with 0.5 M NaCl added to the water phase) the IFT values decreased, probably caused by impurities in either the isooctane or the water phase. Therefore, IFT experiments above these pH values are not presented, as effects observed in this pH range are influenced by the shift in the IFT value for the pure isooctane/water system. Increasing salinity did not affect the IFT values of the isooctane/water system significantly, except for the system containing 1 M NaCl (discussed under Materials and Methods) where the IFT value was observed at about 30 mN/m. Lord et al. (12) have reported a slight lowering of the IFT for the o-xylene/water system with increasing NaCl concentration. These observations were explained from enhanced interfacial activity of impurities existing in the two-liquid system. Similar behavior is also reported by Gaonkar (28), who found that salt led to a considerable reduction in the IFT of a commercial soybean oil/water system. In our work we observe that only 1 M NaCl has a significant effect on the IFT values of the pure isooctane/water system (in addition to high pH values as already mentioned). For 1-naphtoic acid, presented in Fig. 2A, the IFT is constant with increasing pH throughout the whole pH range. It thus seems that neither dissociation effects nor concentration changes due to partitioning of the acid affect the IFT. As already mentioned, one would according to literature expect a decrease in the IFT as pH is increased. A behavior with a decrease in IFT several pH units above the bulk pK a value of the acid has been observed by Cratin (15) and by Joos (29) for stearic acid. This was explained by the difference in dissocia-

40

STANDAL ET AL.

tion of the acid in bulk and at the oil/water interface, a phenomenon first proven by Schulman and Hughes (30). Calculations by Cratin (15) and Joos (29) show that the interfacial pK a is about 3–5 pH units higher than the bulk pK a. However, in this work, there could not be detected any influence from 1-naphtoic acid on the IFT values as pH was increased, probably due to low concentrations. The IFT curves for 5-indanol (Fig. 2B) show that the IFT is constant as long as K is constant, that is, in the acidic and neutral pH range. As the component transfers from the oil phase to the water phase a simultaneous decrease in the IFT is observed, and at approximately the same pH value, the molecules start to dissociate (due to a pK a value of 10.3). From Fig. 3A it is clear that increasing concentration in the aqueous bulk phase at pH 12 reduces the IFT. The lowering of the IFT may then be attributed to increasing water bulk concentration. The addition of NaCl to the aqueous phase does not introduce any difference in the shape of the IFT curve as a function of pH. This is shown in Fig. 2B for 0.5 M NaCl, and the same behavior was also observed for 0.1 and 1 M NaCl. Generally the IFT curve was shifted to lower values as the NaCl concentration increased due to the screening effect of the ions. Using CaCl 2 in the aqueous phase (not presented in any figure), the observed interfacial activity curve for 5-indanol is slightly different in the alkaline pH range compared to the NaCl systems. The IFT remains rather constant for both the 5-indanol and the 1-naphtoic acid system (at approximately 42 mN/m for 1-naphtoic acid and 37 mN/m for 5-indanol) throughout the whole pH range, and no decrease for alkaline solutions was observed. Jennings et al. (31, 32) showed that only a small amount of calcium increased the IFT between caustic and crude substantially, whereas sodium chloride reduced the amount of caustic necessary to give maximum interfacial activity. Cooke et al. (33) have reported the reaction between organic acids and calcium ions in alkaline solution to form the corresponding soaps, which are much less surface active. However, in our system, the concentrations of acids are very low, and probably far below the solubility of the corresponding calcium salt. For comparison, the solubility of calcium benzoate (Ca(C 6H 4COOH) 2) in cold water is as high as 2.7 g/cc (24). A probable explanation for the rather stable IFT curve, in the CaCl 2 systems, may then be a complexation between Ca 21 ions and acid anions to give the monovalent complex in the alkaline pH range. These complexes will tend to concentrate in the water bulk phase rather than at the interface. The IFT curves obtained for quinoline (presented in Fig. 2C) show a behavior different than that of the acidic systems. The IFT is constant in the acidic pH range and then passes through a minimum in the neutral pH range. For the system without salt in the water phase, the IFT slightly increases after the minimum value and then remains constant in the alkaline pH range. For 0.5 M NaCl, the IFT increases continuously after the

minimum value. (0.1 M NaCl showed a behavior similar to that of 0.5 M NaCl.) Looking at the partition coefficient (Fig. 2C) we observe that the IFT lowering corresponds to the transition of molecules between the two liquid phases. The minimum in IFT is observed when the undissociated form of the molecule dominates the system, indicating that this form is more interfacial active than the protonated form. This observation is in agreement with the results from measuring IFT as a function of concentration, where the IFT values measured at pH 12 decreased whereas the IFT values measured at pH 2.5 remained approximately constant with increasing concentration in the aqueous phase (Fig. 3B). The partition coefficient from Fig. 2C shows that the “salting-out” effect is substantially more significant for 0.5 M NaCl compared to 0 M salt added to the aqueous phase. This may explain the difference between the two IFT curves beyond the IFT minimum. The increase in IFT for 0.5 M NaCl after the minimum value may be attributed to the large excess of Na ions in the aqueous phase, which increases the affinity for the nonpolar environment. This will result in an increase in IFT from the removal of molecules from the interface into the oil bulk phase, and also a slight increase in K with increasing pH. For 0 M salt in the aqueous phase, the “salting-out” effect is not significant, and the partitioning from aqueous to oil phase is mainly attributed to the change in relative hydrophilic character of the molecules due to dissociation. The molecules may then accumulate at the interface “undisturbed” after the partitioning into the oil phase. Lord et al. (12) have reported similar results in an octanoic acid/o-xylene/water system. However, in their work, the minimum IFT value was attributed to the presence of other surfaceactive species in addition to the neutral acid and anionic forms, such as ion pairs or multimer complexes (although this effect was neglected at an earlier stage in the work due to low association constants). In our opinion it is more likely that the IFT minimum is directly connected to the distribution of the acid from oil to aqueous phase, as indicated by their distribution results, which in turn is well in agreement with the findings in this work. Rudin and Wasan (11) have reported a similar IFT minimum with increasing pH in an oleic acid/decane/water system, attributing it to simultaneous adsorption of dissociated and undissociated molecules upon the interface. When pH was increased past the minimum, the increase in IFT was explained by increased dissociation of the acid, resulting in less contribution to the IFT. Like the 5-indanol and 1-naphtoic acid systems discussed, CaCl 2 does not contribute to changes in the IFT. A complexation between Ca 21 ions and the free electron pair on the nitrogen atom of quinoline in alkaline solutions may be an explanation for this effect.

O/W MODELS: DISTRIBUTION AND INTERFACIAL BEHAVIOR

SUMMARY

For the compounds examined in this work, the distribution between the two liquid phases is closely connected to the pK a value of the component. For all the systems K equals 1 at this value. Consequently, increasing salinity in the water phase increases the concentration of the component in the oil phase. For 1-naphtoic acid none of the two ionization forms of the molecule are found to be surface active in aqueous solution. For 5-indanol both forms are surface active, and for quinoline only the nonionic form of the molecule is found to be surface active. IFT for the oil/water/1-naphtoic acid system is found to be independent of both oil phase concentration and aqueous phase concentration (in the concentration range examined). IFT for the oil/water/5-indanol system is found to decrease with increasing pH in alkaline solution. We have shown a dependence of the aqueous bulk phase concentration, and that IFT is independent of the oil bulk phase concentration. For both the 5-indanol and the quinoline system we find that IFT is closely connected to the distribution of the components. For quinoline, this results in a minimum value on the IFT curve, corresponding to the transition of the component between the two liquid phases. For the quinoline system we have also observed that the unprotonated form of the molecule is the most interfacial active. The aqueous phase seems to be the operative variable governing the IFT in both the 5-indanol and the quinoline system. For all the systems investigated we find that increasing concentration of NaCl decreases IFT. The presence of CaCl 2 does not contribute significantly to IFT lowering. ACKNOWLEDGMENT We acknowledge the Norwegian Research Council for financially supporting this work.

REFERENCES 1. Jadhunandan, P. P., and Morrow, N. R., Soc. Petrol. Eng., paper 22597, presented at the 66th Annual Technical Conference and Exhibition of the SPE, Dallas, 111, October 1991.

41

2. Seifert, W. K., and Howells, W. G., Anal. Chem. 41, 554 (1969). 3. Dutta, P. K., and Holland, R. J., Fuels 63, 197 (1984). 4. Cuiec, L. E., paper presented at the 21st Intersociety Energy Conversion Engineering Conference, San Diego, August 1986. 5. Craig, F. F., Jr., “The Reservoir Engineering Aspects of Waterflooding,” monograph series, Vol. 3. Soc. Petrol. Eng., Dallas, 1971. 6. Cuiec, L. E., in “Interfacial Phenomena in Petroleum Recovery” (N. R. Morrow, Ed.), p. 319. Marcel Dekker, New York, 1991. 7. Buckley, J. S., Ph.D. Thesis, Dept. of Petroleum Engineering, Heriot-Watt University, Edinburgh, United Kingdom, 1996. 8. Morrow, N. R., J. Petrol. Technol. 1476 (Dec. 1990). 9. Anderson, W. G., J. Petrol. Technol. 1605 (Dec. 1987). 10. Rudin, J., and Wasan, D. T., Colloids Surf. 68, 81 (1992). 11. Rudin, J., and Wasan, D. T., Colloids Surf. 68, 67 (1992). 12. Lord, D. L., Hayes, K. F., Demond, A. H., and Salehzadeh, A., Environ. Sci. Technol. 31, 2045 (1997). 13. Jafvert, C. T., Westall, J. C., Grieder, E., and Schwarzenbach, R. P., Environ. Sci. Technol. 24, 1795 (1990). 14. Schwarzenbach, R. P., Gschwend, P. M., and Imboden, D. M., “Environmental Organic Chemistry.” Wiley, New York, 1993. 15. Cratin, P. D., J. Dispersion Sci. Technol. 14, 559 (1993). 16. Reisberg, J., and Doscher, T. M., Producers Monthly, 43 (Nov. 1956). 17. Danielli, J. F., Proc. R. Soc. London Ser. B 122, 155 (1937). 18. Hartridge, H., and Peters, R. A., Proc. R. Soc. London Ser. A 101, 348 (1922). 19. Peters, R. A., Proc. Roy. Soc. London Ser. A 133, 140 (1931). 20. Levine, I. N., “Physical Chemistry.” McGraw-Hill, Singapore, 1988. 21. Friedel, A. R., and Orchin, M., “Ultraviolet Spectra of Aromatic Compounds,” (a) No. 250; (b) No. 50; and (c) No. 270. New York, 1951. 22. Jones, G., “Quinolines,” Part 1. Wiley, London, 1977. 23. Adamson, A. W., “Physical Chemistry of Surfaces,” Wiley, 1976. 24. CRC Handbook of Chemistry and Physics, (D. R. Lide, Ed.), 71st ed. CRC Press, 1990 –1991. 25. Kortu¨m, G., Vogel, W., and Andrussow, K., “Dissociation Constants of Organic Acids in Aqueous Solution.” Butterworths, London, 1961. 26. Bennett, B., and Larter, S. R., Geochim. Cosmochim. Acta 61, 4393 (1997). 27. Price, L. C., Am. Assoc. Petrol. Geol. Bull. 60, 213 (1976). 28. Gaonkar, A. G., J. Colloid Interface Sci. 149, 256 (1992). 29. Joos, P., Bull. Soc. Chim. Belges 80, 277 (1971). 30. Schulman, J. H., and Hughes, A. H., Proc. Roy. Soc., London Ser. A 138, 401 (1932). 31. Jennings, H. Y. Jr., Soc. Petrol. Eng. J., 197 (June 1975). 32. Jennings, H. Y. Jr., Johnson, C. E. Jr., and McAuliffe, C. D., J. Petrol. Technol. 1344 (Dec. 1974). 33. Cooke, C. E., Jr., Williams, R. E., and Kolodzie, P. A., J. Petrol. Technol. 1365 (Dec. 1974).