Morphological Variation and Recovery Mechanism of Residual Crude Oil by Biosurfactant from Indigenous Bacteria: Macro- and Pore-Scale Experimental Investigations

Abstract

Microbial enhanced oil recovery (MEOR) is being used more widely, and the biological contributions involved in MEOR need to be identified and quantified for the improvement of field applications. Owing to the excellent interfacial activity and the wide distribution of producing strains in oil reservoirs, lipopeptides have proved to be an essential part of the complex mechanisms in MEOR. In this study, crude lipopeptides were produced by a strain isolated from an indigenous community in an oil reservoir. It was found that crude lipopeptides can effectively reduce the IFT (interfacial tension) to 10^-1~10^-2 mN/m under high salinity without forming stable emulsions, and the wettability of natural sandstone can be enhanced (Amott index, from 0.36 to 0.48). The results of core flooding experiments indicate that an additional 5.2% of original oil in place can be recovered with a 9.5% reduction of injection pressure. After the shut-in period, the wettability of the core, the reduction of injection pressure, and the oil recovery can be improved to 0.63, 16.2% and 9.6%, respectively. In the microscopic flooding experiments, the crude oil in membrane, cluster, and throat states contribute nearly 90% in total of the additional oil recovery, and the recovery of membranestate oil was significantly enhanced by 93.3% after shut in. Based on the results in macro and pore scale, the IFT reduction and the wettability alteration are considered primary contributors to oil recovery, while the latter was more dominant after one shut-in period.

Introduction

After the primary and the secondary oil recoveries (the elastic recovery and the water-flooding development), one third to one half of the original oil remains in reservoirs [26]. The limited opportunities to find new oil reserves have drawn focus on the enhancement of the recovery of the remaining oil, which relies on the application of enhanced oil recovery (EOR), known as the tertiary oil recovery, to meet the increasing demand of fossil energy [36]. Among EOR methods, microbial enhanced oil recovery (MEOR) is considered as the most economically viable and environment-friendly technology, utilizing microbes (indigenous and exogenous microorganisms stimulated by injected nutrients) and/or their metabolites to release trapped oil. For decades, numerous studies have been carried out in laboratories and fields. Most lab-scale experiments of microbial flooding have exhibited positive results [2,3,10,35], and some field applications have shown comparable oil recovery and better economic efficiency than the chemical EORs commonly used at present [12,17,25,43,47,48]. However, owing to the various field conditions and unclear mechanisms, the effectiveness of MEOR is currently unpredictable, and it continues to restrain industrial applications [17,44].

A variety of mechanisms related to types of microorganisms and their metabolites, including biosurfactants, biogases, biopolymers, and organic acids, are involved in the complicated process of MEOR [7,17,22,39,40,45,46,49]. The complexity and uncertainty of the microbial processes on oil recovery have been considered as the major barrier for the mechanism research [26,50]. In order to clarify the dominant processes, a few studies have focused on the contributions of the compound mechanism involved in MEOR [26,36,38,50,51]. Among those, the biosurfactant is the most attractive one in recent years, including lipopeptides and rhamnolipids [35]. Lipopeptides can be produced by many species of indigenous bacteria in oil reservoirs, especially the reservoirs with high temperature [3,7,9,11,19,23,30,37,40]. However, the bacterial cells and the produced biosurfactant showed both independent and interactional contributions [5,6,26,50,51]. Therefore, there is a need for the function and the mechanisms of the biosurfactant itself to be identified.

The macro-scale performance of MEOR relies on microscale processes. Therefore, a pore-scale investigation is necessary to illustrate the various MEOR mechanisms. Research on the oil recovery efficacy and mechanisms of lipopeptide has been conducted by macro- and pore-scale experiments discretely. So far, the distinctions of macroand pore-scale physical simulation have not been contrastively analyzed in one single study, and the existing porous media of two- and three-dimensional pore-scale models have only been constructed by particles with a regular shape, such as spheres and hexagons [1,5,6], which differ from the realistic morphology of the pore space. Finally, refined n- or iso-alkanes, instead of crude oil, were used in these microscopic experiments to investigate the effects on hydrocarbons independently, avoiding the interference by non-hydrocarbons [1,5,6]. With the research development, crude oil should be used to mimic the practical processes, considering the significant differences in properties between crude oil and the refined hydrocarbons.

In our previous research, a microscopic flooding system with an artificial pore space was designed to operate under high pressure and temperature, which has demonstrated the effectiveness of the indigenous microbial community and the isolated strain producing emulsifier on oil recovery [51]. As an essential participant of the MEOR process, the mechanism and contribution of biosurfactants should be meticulously studied and determined. Thus, the present paper is focused on the mechanism of crude oil recovery by a biosurfactant. The extracted lipopeptide was used in one-dimensional macro-scale experiments (sandstone core flooding) and two-dimensional pore-scale experiments (microscopic flooding), and the lipopeptide was characterized by the IFT (interfacial tensions), the emulsification, and the core wettability indexes. The objectives of this study were to clarify the mechanisms on crude oil recovery taken place in pore space by the lipopeptide; to distinguish the contribution of each mechanism caused by the lipopeptide; and to quantify the effectiveness of the lipopeptide in the holistic MEOR process.

 

Materials and Methods

Material Preparation

The crude oil used in this study was sampled from two oil reservoirs of the Shengli Oil Field (China) with a viscosity of 598 mPa·s (Crude Oil I) and 1,422 mPa·s (Crude Oil II). Crude Oil I was used in all the experiments to keep consistency, and Crude Oil II, which was more viscous, was used only for IFT measurement as a comparison with Crude Oil I. The brine used in this study was composed of MgSO4 (22 mg/l), NaHCO3 (51 mg/l), KCl (610 mg/l), NaCl (2,335 mg/l), MgCl2 (18 mg/l), and CaCl2 (105 mg/l).

In order to illustrate the correlation between the core wettability, the oil recovery, and other metrics, the properties of experimental cores should be as consistent as possible. Thus, the cores were collected from the same low-permeable sandstone oil reservoir in the Shengli Oil Field, and their characteristics are listed in Table 1. The porosity (%) was calculated as the pore volume (PV) divided by the total volume of the core, where the PV was the volume of water required to saturate the vacuumized cores. The permeability was determined by Darcy’s law [19].

Table 1.Characteristics and applications of the natural cores utilized in this study.

The strain of Bacillus subtilis used in this study had been isolated from the produced water of an oil reservoir in the Shengli Oil Field [13], and it was cultured to produce lipopeptide. It was grown aerobically at 37°C for 48 h to accumulate lipopeptide in liquid medium. The liquid medium was composed of glucose (20 g/l), urea (3 g/l), yeast extract powder (1 g/l), K2HPO4 (3 g/l), and NaCl (5 g/l), and the pH was adjusted to 7.0 [42]. After being cultured, the crude lipopeptide solution (cell-free supernatant, about 0.93~1.16 g lipopeptide per liter solution) was obtained by centrifuging the prepared bacterial culture (10,000 rpm, 5 min). The cell concentration was counted in a Helber Bacteria Counting Chamber (Auvon, UK), and the lipopeptide biosurfactant was quantified using high-performance liquid chromatography [13]; the quantitative results are showed in Fig. 1.

Fig. 1.Time-course profiles of cell growth and lipopeptide production during culture.

Measurement and Analysis

Interfacial tension measurement. The IFT was measured at 65℃ between the crude oil and the crude lipopeptide solution diluted by the brine (1% (v/v)) with different salinity, using an automatic interfacial tensiometer (Powereach, TX-500, Shanghai, China). In order to guarantee reliability and repeatability, each measurement was performed four times.

Emulsification analysis. Emulsification activity of the supernatant was characterized at 65℃ by the emulsion index [18]. First, 5 ml of the hydrocarbon liquid (paraffin and crude oil, respectively) was added to 5 ml of lipopeptide solution. Thereafter, the mixture was shaken in a vortex for 2 min and then kept still. The heights of the emulsion (He) and the total mixture (Ht) were measured, and the emulsification index (E) was then calculated with the ratio of He to Ht.

Wettability analysis. The Amott-Harvey wettability test, which is the commonly used measurement of average wettability, was used to analyze the wettability variation of the entire core at 65℃ [4,38] and is described as follows:

where Iw is the Amott water index, Io is the Amott oil index, Vwsp is the displaced brine due to the spontaneous oil imbibition, and Vwt is the overall displaced brine due to the spontaneous and forced oil imbibition. Similarly, Vosp and Vot are the displaced oil due to the spontaneous brine imbibition and the overall displaced oil due to the spontaneous and forced brine imbibition, respectively. I is the overall Amott Harvey wettability index.

Flooding Experiments

Core flooding experiment. The core flooding apparatus used in this study was the same as other reports [36,38]. After the porosity and permeability measurements, the cores were placed in the core holder and then saturated under vacuum condition by injecting 2.0 PV (pore volume) of the brine (the stage of water saturation). Next, 2.0 PV of the crude oil was injected continuously to simulate the original oil in place (OOIP) (the stage of oil saturation). After that, the cores were flooded by the brine with 5.5 PV to form the residual oil saturation status (the stage of water flooding, capillary number of 3.5 × 10-8). Finally, 0.5 PV lipopeptide injections were carried out (the stage of lipopeptide flooding, capillary number of 2.8 × 10-6) and the brine was injected (the stage of post-water flooding) continuously until no oil was produced. During the stages of water, lipopeptide, and post-water flooding, the pressure at the inlet boundary of the core holder was recorded. The injection rate of the water phase (the brine and the lipopeptide) and the oil phase was 0.05 ml/min and 0.01 ml/min, respectively, and the temperature was held constant at 65℃. As a comparison, a contrast experiment with a 3-day shut-in period after the lipopeptide flooding was performed.

Microscopic flooding experiment. The two-dimensional microscopic flooding apparatus is shown schematically in Fig. 2. The experimental setup is composed of two main systems, the flooding and image capturing systems with some accessories. The flooding system consists of a syringe pump, three accumulators, a model holder, and a light source (white light); the image-capturing system includes a microscope, a camera, and a computer. The heart of the apparatus is the micro model, which is made of a two-dimensional pore structure etched onto the surface of a flat glass plate, and the pore structure is identical with the realistic cross-section image of a sandstone core from the Shengli Oil Field [50,51]. It was covered by another glass plate to create an enclosed pore space. This covering plate had an inlet and an outlet hole at two opposite corners, allowing fluids to flow through the pore network. The external size and pore volume of the model are 40 mm × 40 mm and 50 µl, respectively. The photograph of the micro-model used in this study is shown in Fig. 3. In the steel model holder (a hollow cylinder), the micro-model was clamped horizontally by thick glasses and filled with water to load the overpressure, which allowed light to penetrate through vertically. This special holder and a back-pressure valve were used to mimic the high pressure in oil reservoirs, and the accumulators were used for holding and injecting the liquids (crude oil, brine, and biosurfactant).

Fig. 2.Microscopic flooding experiment apparatus.

Fig. 3.Top view of the oil-saturated two-dimensional microscopic model.

The procedure for the microscopic flooding experiment is similar to the core flooding experiment process. After the water saturation, oil saturation, and water flooding stages, 0.8 PV lipopeptide was injected followed by 1.0 PV brine flooding. After that, the micromodel was cleaned by circulating petroleum ether and distilled water. For the second round, after the water saturation, oil saturation, and water flooding stages, the injection of 0.8 PV lipopeptide was followed by a shut-in period of 3 days, and continued with post-water flooding as a comparison. The rates of all injections in the microscopic flooding experiments were 0.003 ml/min (the capillary number of water-flooding, 5.3 × 10-7). During the process, the micro-model was kept at 65℃ and 5 MPa of overpressure. The morphology of residual oil was captured constantly.

Oil saturation determination from images is difficult and time consuming. To tackle the issue, the original images were sharpened to improve the contrast, and then analyzed with a program developed using Matlab. The program was able to distinguish the water, oil, and glass phases by given gray thresholds. Finally, the oil saturation was calculated by the area proportion of oil to pore space [50].

Considering the mobilization of crude oil in the micro-model with the artificial pore system, the morphology of remaining oil can be categorized and selected manually in images according to the interaction between brine, oil, and porous media [27,28], including cluster, blind-end, islet, throat, and membrane states (Fig. 4) [50,51]. Among them, the islet and membrane states refer to the isolated oil drops in the pore space and the residual oil attached along the pore wall, respectively; the blind-end state is the oil trapped in a blind throat with only one exit; the oil in throats that cannot be pushed out by flooding is defined as the throat state; and a series of throat states that are interconnected to each other are called the cluster state [27,28].

Fig. 4.Types of residual oil states in the image of the microscopic model (the lighter region: glass; the darker channel: pore space; Arrow A: membrane; B: blind-end; C: cluster; D: islet; E: throat).

 

Results

Effect of Biosurfactant on Interfacial Tension and Emulsification

The IFT is a major influence factor of oil recovery for the EOR and MEOR processes [26,30]. The results in Fig. 5 showed that the IFT between the brine and the crude oil was reduced to 100~10-2 mN/m and kept declining with an increase of salinity, which indicates that the lipopeptide can maintain activity under high salinity conditions, unlike most kinds of chemical surfactant.

Fig. 5.Effects of salinity on the interfacial tension between crude oil and saline water containing biosurfactant.

Biosurfactants can emulsify various hydrocarbons under differing conditions [14,22], and the emulsification is considered as another important factor for releasing trapped oil. The emulsification ability on hydrocarbons was characterized by the emulsification index (E) in this study. In order to demonstrate the abilities to trigger and stabilize the emulsion, the emulsification index was measured at 5 min (E5min) and 24 h (E24), respectively. Fig. 6 shows that the lipopeptide can emulsify the crude oil with greater stability than paraffin under mild conditions (salinity <10 g/l). An increase in salinity can inhibit the emulsification severely. Concerning the high salinity condition (salinity ≽10 g/l), the crude oil can be dispersed effectively by vortex, and the scattered oil drops tend to aggregate once they contact each other without forming a stable emulsion. Unlike the colorless paraffin, it was difficult to distinguish the nontransparent crude oil emulsion phase from the black crude oil, when the height of the emulsion was less than the oil phase. Thus, the emulsification index of crude oil was not reported under high salinity condition (see Fig. 6).

Fig. 6.Effects of salinity on the emulsification index of hydrocarbons by crude biosurfactant.

Wettability Alteration and Retention Time

In recent years, more research supports that the wettability alteration is one of the most important factors for oil recovery, and sufficient retention time is required to achieve an effective wettability alteration by the lipopeptide [32,38]. Thus, to clarify the effects of lipopeptide and retention time on wettability, the Amott-Harvey wettability index (I) of sandstone cores was measured with and without a shutin stage, respectively. The first measurement was executed immediately after the injection of lipopeptide, and the other one was carried out after a 3-day shut-in. The results in Table 2 indicate that the lipopeptide could increase the I value significantly from 0.36 to 0.63, which means the wettability of the sandstone core was increased with an increase of retention time and lipopeptide concentration. It is necessary to state that the effect of cells on wettability can be excluded, although the crude solution obtained by centrifugation had never been sterilized, because the mesophilic strain used in this study cannot grow effectively at 65°C [13].

Table 2.Amott-Harvey wettability index of the cores treated by the crude lipopeptide.

Core Flooding Experiment

The variations of injection pressure and accumulated oil recovery are shown in Fig. 7. During the 5.5 PV brine injection (from 0.0 PV to 5.5 PV of the injection volume), the pressure and recovery eventually stabilized at 3.65 MPa and 36.4% OOIP, respectively. With 0.5 PV of the lipopeptide and 4 PV of the post-water flooding, the pressure declined gradually and stabilized at 3.3 MPa, and the recovery was enhanced to 41.6% OOIP subsequently. The results of the shut-in period are shown in Fig. 8. During the injection of lipopeptide, and following the shut-in and the post-water flooding periods (from 5.5 to 10.0 PV of the injection volume), the injection pressure declined from 3.7 to 3.1 MPa, and the recovery increased from 38.5% to 48.1% OOIP. Additionally, the control experiments, in which the lipopeptide was replaced by brine, did not show detectable variations of pressure and recovery. Comparing the results in Figs. 7 and 8, the improvements of the pressure reduction (0.35 to 0.60 MPa) and the additional oil recovery (5.2% to 9.6% OOIP) can be attributed to the insertion of the shut-in stage.

Fig. 7.Injection pressure and accumulated recovery of core flooding without shut-in.

Fig. 8.Injection pressure and accumulated recovery of core flooding with shut-in.

the pressure and recovery eventually stabilized at 3.65 MPa and 36.4% OOIP, respectively. With 0.5 PV of the lipopeptide and 4 PV of the post-water flooding, the pressure declined gradually and stabilized at 3.3 MPa, and the recovery was enhanced to 41.6% OOIP subsequently. The results of the shut-in period are shown in Fig. 8. During the injection of lipopeptide, and following the shut-in and the post-water flooding periods (from 5.5 to 10.0 PV of the injection volume), the injection pressure declined from 3.7 to 3.1 MPa, and the recovery increased from 38.5% to 48.1% OOIP. Additionally, the control experiments, in which the lipopeptide was replaced by brine, did not show detectable variations of pressure and recovery. Comparing the results in Figs. 7 and 8, the improvements of the pressure reduction (0.35 to 0.60 MPa) and the additional oil recovery (5.2% to 9.6% OOIP) can be attributed to the insertion of the shut-in stage.

Microscopic Flooding Experiment

The various morphologies of remaining oil in porous media, caused by different flooding processes [1,5], can verify the mechanisms concluded from macro experiments and be helpful to reveal potential mechanisms [50]. The microscopic flooding experiments were conducted to illustrate the morphology variation of remaining oil during each flooding stage, and to analyze the processes taking place in the pore space. The quantified results of residual oil in different states without and with shut-in periods are listed in Tables 3 and 4, respectively. The residual oil saturation was reduced by the lipopeptide injection and the post-water flooding. Furthermore, between the stages of the lipopeptide injection and the post-water flooding, the saturation of islet-state oil was slightly higher than that before the lipopeptide injection. The reason for this unexpected phenomenon is that the oil droplets dispersed by the IFT reduction and the flooding shear were captured in the image analysis before the post-water flooding and counted as isle-state remaining oil.

Table 3.The variation of the remaining oil of different morphological types by surfactant flooding.

Table 4.The variation of the remaining oil of different morphological types by surfactant flooding with the 3-day shut in.

Concerning the contribution of different remaining-oil morphologies to the additional oil recovery, the states of membrane, throat, and cluster contributed the most (close to 90%). Compared with the results without shut-in stage, the total additional oil recovery raised from 13.0% to 19.1% OOIP after the shut-in. The additional oil recovery in the three largest contributors (membrane, cluster, and throat states) increased by 93.3%, 48.8%, and 15.6%, respectively.

Mechanisms involved in the lipopeptide injection and post-water flooding stages can be draw from the morphological variations of residual oil (from Figs. 9 to 12). After the lipopeptide injection, fractions of the residual oil were first dispersed as hollow spheres owing to the IFT reduction (Fig. 9A). Once the post-water flooding began, the scattered oil droplets tend to aggregate as a continuous phase. The aggregated oil band can easily adapt its shape to the narrow throats and successively go through without clogging (Fig. 9B). Finally, the oil bands travelled along the pore channels without contact with the porous medium (Fig. 9C).

Fig. 9.Microscopic images of dispersed crude oil droplets and balloons (A), mobilized oil through the throat (B, pointed by the arrow), and the mobilized continuous oil band (C, along the arrow).

Moreover, a middle phase, which is a mixture of the crude oil and the brine, was observed in this experiment (Fig. 10). The middle phase was transparent and yellow in appearance (darker than water and lighter than oil). The reason is that the solubility between oil and brine was improved by the biosurfactant. According to the observation, the middle phase could be generated gradually and increased in volume during the static shut-in period (Fig. 10A), and travelled along during the following flooding (Fig. 10B).

Fig. 10.Microscopic images of crude oil solubility enhancement by the lipopeptide biosurfactant (the light yellow sections pointed by arrows). (A) Snapshot of the spontaneous formation during a static shut-in period; (B) snapshot of the mobilizing state along pore spaces during post-water flooding.

In addition, there was another interesting phenomenon that occurred during the shut-in stage. A considerable fraction of the residual oil blobs mobilized spontaneously (Arrow A in Fig. 11), which is driven by the Marangoni effect due to the IFT gradient [50]. Some sections of solid surface were detached from the residual oil and no longer adhered by the oil blobs passing by, which indicated that the surface wettability had increased (Arrow B in Fig. 11) [5,6]. Moreover, water was found to penetrate the shallow blind end filled with crude oil with enough retention time (Fig. 12).

Fig. 11.Microscopic images of morphology variations of residual oil during the shut-in stage. (A) Before the shut-in stage; (B) after the shut-in stage. (Arrow A) Spontaneous mobilization of oil blobs; (Arrow B) wettability alteration.

Fig. 12.Microscopic images of morphological variations of residual oil during the shut-in and post-water flooding stages. (A) After biosurfactant injection and before the shut-in stage; (B) after the shut-in stage; (C) after the post-water flooding. The arrows indicate the oil displacement in a blind end.

 

Discussion

Plenty of research has indicated that surfactants, including biosurfactants, can enhance oil recovery by reducing the IFT between brine and petroleum hydrocarbons [2,3,10,35]. Under low salinity conditions, the lipopeptide solution used in this study can effectively reduce the IFT and form a stable emulsion [14,41]. Because of the stronger molecular polarity of the raw mixture compared with the refined hydrocarbon, the emulsion formed with crude oil (raw) is more stable than that with liquid paraffin (refined) [14]. As for the relation between the IFT and the emulsification, the lipopeptide could maintain activity under high salinity (Fig. 5), whereas the stability of emulsification decreased severely with increasing salinity (Fig. 6). Therefore, the ability of the lipopeptide to reduce surface tension was not sufficient to form stable emulsions, which is consistent with a report published previously [19]. The mechanism of the IFT reduction could be inferred to play a greater role than emulsification under high salinity, and the aggregation tendency of dispersed oil droplets was found to be a useful characteristic for oil transport in porous medium.

Other than measuring contact angle, porous media wettability is often quantified by macro-scale indices with the methods of Carter or Amott [6], which represent the integrated wettability, and they are convenient to ascertain the correlation with macro performance. Based on the results, the Amott-Harvey wettability index (I) of low permeable sandstone cores was effectively changed toward an increased wettability by the lipopeptide. However, the directions in which the lipopeptide change wettability have not been consistent. Both changes toward an increased wettability [1,5,6,31] and the opposite trend [34,38] have been reported.

According to the results of the core flooding experiments, the injection pressure declined significantly after the lipopeptide injection. Therefore, the plugging mechanism is apparently absent for additional recovery in this scenario. Overall, the IFT reduction, emulsification, wettability alteration, and pore-space plugging are proposed as the dominant mechanisms for additional oil recovery by biosurfactants [21,26,38]. Based on the results, the emulsification (unstable emulsification under high salinity) and the pore-space plugging (decline of injection pressure) have been excluded.

Then, the question remains of which one is the more primary contributor between the IFT reduction and the wettability alteration. The literature has reported that in carbonated formations, the wettability alteration is more important than the IFT reduction in oil recovery during chemical flooding processes [8]. As for the biosurfactant flooding, the wettability alteration, contributing 60~70% to the oil recovery after adequate retention time (shut-in stage in physical simulation), plays the more significant role than the IFT reduction in carbonated formations [32,38], while less dominant in sandstone [21,36].

In this study, the increase of the wettability index (I) of sandstone cores was enlarged from 0.12 to 0.27 after the 3-day shut in (Table 2), while the result also indicated that the wettability had already been effectively altered without shut-in, which is different from the report mentioned above [38]. Therefore, in the first core flooding without a shut-in period, the two mechanisms could not be distinguished clearly. In the second core flooding with the shut-in period, oil recovery improved from 5.2% to 9.1% OOIP (Figs. 7 and 8), and the wettability index (I) decreased without further IFT reduction, which means the additional 3.9% OOIP of oil recovery depends mostly on the wettability alteration. Comparing with the results using purified lipopeptide [21], the improved wettability performance in this study might have been caused by more types of metabolites in the crude solution compared with the purified biosurfactant, which improve the wettability alteration by additional effects [36].

Furthermore, the process and phenomenon of the microscopic experiment could validate the dominant mechanism mentioned above and aid in the discovery of the accompanying and potential mechanisms in pore scale caused by biosurfactants. First, unlike the living bacterial cells, the biosurfactant solution cannot release oil trapped in deep blind-end tunnels (Arrow B in Fig. 4) effectively, which is the most obvious difference between them [50,51]. The recovery of oil trapped in blind ends by MEOR depends mostly on the propagation, metabolism, dispersion, and chemotaxis of living cells, which are exclusive properties of microbial processes [16]. Our previous report showed 3.1% OOIP from blind ends by MEOR [51], whereas the crude lipopeptide solution recovered only 1.2%~1.6% OOIP in this study. However, the remaining oil filled in the shallow blind end (Fig. 11) can be penetrated by water after the lipopeptide injection with enough shut-in stage, which demonstrated that the wettability of the fractional pore had been increased. The ability of penetration might be due to the exclusive properties of the lipopeptide, including the flexibility, large volume, and strong polarity of the molecule. Based on the flexibility of the molecule, the hydrophilic residues of amino acids can be shielded from the external environment by the change of molecular conformation, which helps the lipopeptide molecule to go across a hydrophobic medium [33], which is the major biological property of lipopeptides to interact with cell membranes [29], and to initiate lipid phase transitions [20] and membrane destabilization [24]. In addition, because of the large volume and strong polarity of the hydrophilic head group, the normal translational coefficients of surfactants are much lower than the tangential translational coefficients, demonstrating the strong anisotropy of the translational motion at the interface [15]. These molecular properties effectively support the capability of penetration, diffusion, and stabilization of lipopeptides between the oil-water interfaces, and determine the necessity of the retention time for spontaneous mobilization of the residual oil.

Secondly, the residual oil was dispersed into drops as hollow spheres by the crude lipopeptide, due to the IFT reduction. Compared with the solid oil drops dispersed by other bioemulsifier [51], the sphere was covered with a thin film of crude oil and filled with water, which have not been observed yet in the experiments using refined alkanes. Once the post-water flooding began, the hollow-sphere oil droplets preferred to aggregate and flow as a continuous phase. The aggregated oil band can easily adapt its shape to narrow throats and successively penetrate, which improves the relative permeability of the oil phase. Furthermore, the oil bands flow nearly along the central axis of pore channels without contact with pore walls, indicating an increase in wettability. The considerable enhancement of the residual oil recovery in the membrane state further evidenced the wettability alteration as a primary contributor. Another observed phenomenon is the formation of a middle phase between oil and water, indicating the enhanced oil solubility by the biosurfactant. The microparticles in the middle phase were smaller than the pore holes in diameter, which assists the mobilization of residual oil and the reduction of injection pressure.

Lastly, the quantified morphology variations of residual oil showed that the greatest contributor to the additional oil recovery is the membrane-, throat-, and cluster-state oil, which indicates the synergy of the IFT reduction and the wettability alteration in this study. Concerning the effect of retention time, the increase of the additional oil recovery between membrane and throat states varied considerably (93.3% and 15.6%, respectively), which indicates that the decline of viscous force on the oil/solid interface played a more dominant role than the reduction of capillary pressure. The additional oil recovery was enhanced by 46.9% after the shut-in period, which was consistent with the results of the core flooding experiments. Therefore, the wettability alteration became more dominant than the IFT reduction after the shut-in period.

Overall, the hollow-sphere oil droplets and their aggregation tendency were proved by the unstable emulsification under high salinity at the macro scale; the spontaneous release and mobilization of the residual oil and the adaptation to narrow throats were caused by the IFT reduction; the spontaneous residual oil detachment and the flow trajectory show that the porous media had an increased wettability; and all the pore-scale effects above reduced the flow resistance and no clogging was observed in the microscopic experiment, which verified the decline of injection pressure in macro-scale core flooding. Based on these results, the phenomenon of the macro-flooding experiments was well supported by the results from the pore-scale experiments, and the IFT reduction and wettability alteration are proposed as the primary mechanisms on crude oil recovery in sandstone by the crude lipopeptide. In addition, the wettability alteration turns into the uppermost contributor during the flooding after the shut-in period. The mechanisms concluded from the macro- and pore-scale experiments need to be further quantified in future studies to clarify the contribution of each in the fields.