1 Introduction

In the coming decades, the decommissioning activities in the Netherlands are forecasted to increase because many mature wells are reaching the end of their production phase (Nexstep 2020). Moreover, hundreds of new wells for geothermal energy are planned to be drilled in the future to meet the production goals of 50PJ in 2030 (EBN 2018). Finding economic and durable solutions for sealing and well plugging would contribute to significant cost savings for abandonment activities and improve the rate of return of renewable projects.

To this date, most practices adopted in P&A operation use an engineered cement-based material as a plug. Conventional Portland cement is economically affordable and has high strength and low permeability when set. Despite this, there are concerns related to the ability of cement to deliver a long-term seal, especially at certain downhole conditions that negatively impact the cement’s performance. For example, annular cement and cement plug can crack or debond due to hydration shrinkage and/or mechanical or thermal loading (Aas et al. 2016; Bois et al. 2011; Corina et al. 2020; Meng et al. 2021; 2022; Nagelhout et al. 2010; Torsæter et al. 2015; Vrålstad et al. 2019). In carbon capture storage wells, chemical reactions with CO2 rich brine can degrade the cement which can lead to leakage pathways, particularly when mechanical defects are present (Bachu and Bennion 2009; Duguid and Scherer 2010; Wolterbeek et al. 2013). In geothermal wells, the combination of cement deterioration due to high downhole temperature and cyclic thermal loading increases the risk of integrity failure (Kaldal et al. 2016; TerHeege et al. 2017). Integrity issues in cement plugs mean costly and complicated well remediation, and hence there has been significant interest in recent years to explore alternative sustainable plugging materials.

Bentonite is a natural clay-based material, consisting of mostly montmorillonite minerals, which can attract water molecules resulting in swelling and expansion. This is especially the case in Na-montmorillonite, which can swell to complete colloidal dissolution of the clay platelets (Pusch, Roland 1978). Bentonite has been identified as a solution to provide reliable well isolation due to its superior sealing capabilities. Under a constrained volume, hydrated bentonite can form a compacted solid and hydraulically tight plug (Corina et al. 2021; Mitchell and Soga 2005). Due to swelling, the pore space of the compacted bentonite shrinks, hence the transport of fluids within the pore space is inhibited considerably. The water- and gas permeability of hydrated bentonite ranges from 10–3 to 10–4 milliDarcy (Clark and Salsbury 2003; Idialu et al. 2004) and 10–3 to 10–6 milliDarcy (Gutiérrez-Rodrigo et al. 2021; Villar et al. 2014), respectively. There have been numerous applications of bentonite for plugging shallow boreholes such as radioactive waste disposal (Daemen and Ran 1996; Galamboš et al. 2011; Pusch, R. 1992; Pusch, Roland 2008), water and groundwater monitoring wells (Edil et al. 1992; Hodder 2016; Papp 1996), and sealing coal-bed methane wells (Towler et al. 2016). Trials in oil and gas wells with depths of up to 1300 m have been successfully performed (Clark and Salsbury 2003; Englehardt et al. 2001; Idialu et al. 2004; Towler et al. 2016). The bentonite placement only requires a simple operational job and potentially can be applied rig-less. Clark and Salsbury (2003) suggested that replacing the cement with bentonite as plugging material resulted in a substantial reduction in abandonment costs by 50%.

In terms of mechanical integrity, slipping or shearing at the bentonite interfaces due to high axial loads has been identified as the common failure mechanism of bentonite columns in a borehole or shaft (Akgün. et al. 2006; Akgün, Haluk 2010; Pusch, Roland 1978; Randolph et al. 1991; Towler et al. 2020; Towler et al. 2008). This is different from the typical tensile failure encountered in cement interfaces (Bois et al. 2019; Moghadam et al. (2022); Moghadam and Orlic 2021; Orlic et al. 2021). The mechanism of shear failure at the interface is also recognized as one of the potential failure modes in the well abandonment guidelines (Oil & Gas UK 2015). The capacity of a plug to oppose the displacement load is governed by the shear strength at the plug interfaces with casing/rock, also referred to as the shear bond strength or interfacial shear strength. In accordance with standards from the Dutch State Supervision of Mines (NOGEPA - OPCOM 2019), the emplaced sealing material must retain their original position and remain attached to the interfaces they have been placed against in the wellbore. This implies that the emplaced bentonite plug should be mechanically stable to withstand high axial loads, e.g. from reservoir fluid inflow, to act as an isolating barrier. Hence, the interfacial shear strength of hydrated bentonite is a critical parameter to assess the mechanical integrity of bentonite plugs.

Several studies have reported the interfacial shear strength of hydrated bentonite for annular- and full-bore sealing. Ogden and Ruff (1991) reported that the annular bentonite prepared in steel pipes (ID 2.67–4.83 cm x OD 10.19 cm) has an interfacial strength varying between 3.4 and 27.3 kPa. In a similar experiment by Tveit (2012), the annular bentonite prepared in plastic pipes (ID 5 cm x OD 8.97 cm) generated an interfacial shear strength of 1.02 kPa. Towler et al. (2020) performed an hydraulic dislodgment test on a full-bore plug prepared using compressed bullet-shape bentonite and cured in a 114 mm casing OD. It was shown that the interface shear strength of the bentonite plug increased after multiple dislodging/re-curing cycles ultimately reaching 337.15 kPa after six cycles. Holl (2019) conducted a similar experiment using bentonite bullet treated with polyvinylpyrrolidone (PVP). The ultimate shear strength at the interface of the bentonite plug after three dislodging/re-curing cycles ranged between 7.8 and 14.3 kPa for a 97.2 mm casing ID, and between 23.7 and 51.4 kPa for a full-bore plug prepared for a 139.8 mm casing ID.

The behavior of clay-based material is strongly dependent on the physico-chemical interactions between the clay particles, the pore fluid chemistry, and the surrounding conditions. Generating a compact and highly dense bentonite plug is important to provide the required mechanical stability (Dixon et al. 2011; Pusch, Roland 1978). During abandonment, the well candidates are commonly filled with a completion fluid containing inhibitors to prevent downhole corrosion (e.g. by adding scavengers that target corrosion source or by increasing the fluid pH). The completion fluid can be made from different fluid bases to achieve the required hydraulic pressure and ensure overbalance to the reservoir. Brine-based completion fluids are typically used for a high-pressure reservoir, whereas a freshwater-based completion fluid can be used for depleted reservoirs. In addition, higher temperatures and pressures in deeper wells can impact the performance of the bentonite plug. Some research has investigated the physico-chemical interactions of clay under various levels of pH, salinity, pressure, and temperature (Cui et al. 2019; Dixon A. 2000; Herbert et al. 2008a; Ye et al. 2013, 2014). However, research on the impact of these parameters on the sealing capacity of the bentonite plug is currently lacking.

This paper aims to assess the interfacial properties, shear strength, and mechanical–chemical stability of bentonite plugs for plugging deep subsurface wells at depths below 1500 m. The influence of the fluid composition (e.g. salinity, pH, and saturation), and in-situ conditions (e.g. diameters, pressures, and temperatures) was explored. This was done in a set of experiments including sliding test, push-out test, and hydraulic dislodgement. Sets of large-scale experimental setups were built in the Rijswijk Center for Sustainable Geo Energy (RCSG) to match in-situ downhole conditions and realistic full wellbore diameters. The operational placement of bentonite pellets in a large-scale setting was also investigated. The bentonite plug application for well decommissioning needs to meet the material requirements defined by the Dutch regulators. Hence, the results obtained from the experiments were used to predict the minimum bentonite column length required to abandon a deep well in compliance with national regulations.

2 Materials

In this study, we evaluated two types of bentonite pellets. Both pellets have a cylindrical shape and an average length of 5–20 mm and an average diameter of 6.5 mm. Pellet A has a high swelling capacity, whereas Pellet B has a medium swelling capacity. Both pellets have an initial dry density of approx. 2100 kg/m3 and are mostly composed of montmorillonite clays.

Fluids with different levels of salinity and pH were prepared and used to cure the bentonite pellets. Four saline solutions were prepared with different salt compositions and concentrations, as summarized in Table 1. The fluid pH investigated in this experiment varied from pH of 7 (neutral), 9, and 10.5. To achieve elevated fluid pH, NaOH was added to the solution.

Table 1 The description of saline fluid mixtures

3 Methodology

3.1 Sliding Test

Previous studies have provided measurements on the internal shear strength of compressed bentonite blocks or pellets (Holl and Scheuermann 2018; Sinnathamby et al. 2015). However, shear strength measurements at the interface of bentonite with adjacent material are still limited. A sliding test was performed to determine the friction angle and cohesion at the interface of the bentonite plug with various surfaces. The tests were performed by measuring the required shear force to slide the specimens while applying a normal load perpendicular to the interface. The correlation can be described following the Coulomb failure criterion:

$${\tau }_{s}=c+{\sigma }_{N}^{^{\prime}}\mathrm{tan}\left(\phi \right)$$
(1)

where \(\tau\) is the shear strength, \(c\) is the cohesion, \({\sigma }_{N}^{^{\prime}}\) is the total effective normal stress acting at the interface, and \(\phi\) is the friction angle.

A custom-built sliding test setup was built (Fig. 1) with a similar mechanism to a direct shear test. The test specimen consisted of a rectangle steel container (L 140 mm x W of 100 mm x H 30 mm) that was filled with a bentonite block. The blocks were made of pellet A that was previously hydrated in freshwater. The average bulk density of the bentonite blocks was 1.5 g/cm3. The bentonite sample had a distance of 5 mm from the steel container to prevent contact between the container and the test surface.

Fig. 1
figure 1

a Representation of the bentonite plug in the wellbore and b the schematic drawing of the interface sliding test

Before testing, the specimen was placed on the test surface, loaded with weighting blocks for 2 h, and was covered to prevent drying. The test was performed by pulling a cable connected to the container until the bentonite block slid on the test surface. The normal load applied in the tests varied between 3 and 12.3 kPa. The actual effective normal stress at the downhole conditions might be larger than the load applied in this experiment due to the setup limitation. By using the assumption of linear Coulomb failure criterion, the results would also be relevant for conditions with higher effective normal stress. The tests were conducted with different representative surface materials namely steel, high-porosity (sandstone), and tight (igneous) rocks. Two rock saturation conditions were evaluated: dry and saturated conditions. In the dry tests, the rock was undersaturated adding only a small amount of water to the surface to prevent the bentonite from drying. In the saturated tests, the rock was saturated with freshwater (FW) or seawater (SW) before testing. During the test, the saturated rock was submerged in the fluid to a depth of 1 mm.

3.2 Interfacial Shear Strength Measurements

The plug interfacial shear strength measurements were performed using two different types of tests, namely the push-out test and the hydraulic dislodgment test. In both tests, an axial load was applied to one end of the plug until dislodgement occurred. The push-out tests were used for a quick quantitative comparison of bond quality for a wide range of curing conditions by using a mechanical load. The hydraulic dislodgment tests were designed to mimic unwanted pressurized fluid inflow below the plug and to test the leak tightness of the plug. In both tests, the bentonite plug strength would be calculated following the axial force equilibrium, which can be written using Eq. (2):

$${F}_{max}={P}_{dis }A={\tau }_{s}{A}_{c}$$
(2)

where \({F}_{max}\) is the maximum force measured in the push-out test, \({P}_{dis}\) is the dislodgement pressure measured in the hydraulic dislodgment tests, \({\tau }_{s}\) is the interfacial shear strength, \({A}_{c}\) is the contact surface between the plug and casing, and \(A\) is the cross-section area of the plug.

The experiments were done using various pellet types, curing fluids, and under different curing pressure and temperature. The summary of all test combinations in each experimental work is listed in Table 2.

Table 2 Test combinations of experimental work of interfacial shear strength measurements

3.2.1 Push-Out Test

In these tests, small-scale bentonite plugs were built in a transparent tube (OD 50 mm) sealed with a rubber flange (Fig. 2). Each sample was prepared with 250 gr of bentonite pellet A, which was placed after the fluid. Gravel was added on top of the bentonite to prevent excessive vertical expansion. After curing, the gravel and the rubber flange were removed. The sample was placed under a load frame to apply a force on the bentonite plug. The force and displacement were recorded during each test. All tests were run with a similar displacement rate of 1.5 mm/s to ensure comparative results. The peak load observed from the load vs. displacement plot was used to calculate the interfacial shear strength.

Fig. 2
figure 2

(Left) Schematic of small-scale sample for push-out test and (right) push-out test mechanism

In the first experiment set using the push-out test, we measured the plug strength under various combinations of curing fluid salinities, pH (up to 10.5), and curing temperatures (up to 80 °C). The bentonite plugs were cured for 24 h before the push-out tests, and the density of the final plugs was measured. In the second experiment set, exposure tests were conducted to understand the impact of changes in fluid pH and salinity on the strength of the plug. The exposure tests were run for both short- and long-term periods. Bentonite plugs were initially cured in freshwater for 4 h and 1 month, respectively for short- and long-term tests. Afterward, the fluid on top of the plugs was replaced by a 1.03 NaCl solution with a pH of 9. The push-out tests were performed after an exposure period of up to 2 weeks for the short-term and up to 3 months for the long-term tests. Benchmark samples with no exposure were also prepared and tested.

3.3 Hydraulic Dislodgment Test

3.3.1 Small-Scale Experiment

The setup (Fig. 3) consisted of a dual-glass vessel with a total length of 800 mm. The annular space of the dual-glass vessel was used to circulate hot water to maintain a constant temperature. The inner- and outer glass diameters were 65 mm and 80 mm, respectively. Both sides of the vessel were connected to pressure control lines for providing curing pressure and to perform a dislodgement test.

Fig. 3
figure 3

Schematic of small-scale setup for dislodgement tests

The bentonite sample (approx. 500 mm) was placed in between gravel columns with a length of 60 mm. It was prepared by placing 1850 g of bentonite pellets, followed by approximately 1 L of fluid. The samples were hydrated for 4 days under selected temperatures and pressures. During the test, the top pressure was slowly decreased to create a differential pressure until dislodgment was observed. Dislodgement was confirmed visually and through a sharp pressure change at the top of the sample due to the movement of the plug. In this experiment, both pellets A and B were tested under various curing pressures (up to 7 bars), temperatures (up to 80 °C), and fluid salinities.

3.3.2 Large-Scale Experiment

These experiments were performed using the typical casings in deep wells to generate a large-scale bentonite plug with realistic diameters. Two different experiments were conducted using a separate configuration: the first experiment was set up for low-pressure and low-temperature (LPLT) curing conditions, whereas the second experiment was for high-pressure and high-temperature (HPHT) conditions.

The setup for the LPLT experiments (Fig. 4) was constructed with varied casing sizes of 7–5/8″ (76.2 kg/m or 51.2 lbs/ft), 9–5/8″ (79.6 kg/m or 53.5 lbs/ft), and 13–3/8″ (90.8 kg/m or 61 lbs/ft). The end of the casing sections was secured with flanges and connected to pressure control lines. The bentonite pellets were placed on the top of a gravel column (200 mm) to create an approx. 6 m long plug. In some samples, a sand column of 100 mm was added below the plug. In these experiments, the pellet placement, the curing fluid salinity, and the curing pressure were varied. There were two different techniques of pellet placement evaluated: (i) gravitation, i.e. dropping the pellets at once into the casing filled with fluid, and (ii) batch-wise, i.e. placing fluid and pellets in multiple batches.

Fig. 4
figure 4

The schematic drawing of large-scale LPLT setup and the setup realization

All plugs were made of pellet A and were initially cured for 3 days, except for one specimen that was cured for 1 month to understand the effect of long-term curing. After the plug was cured and the curing pressure was released, the first pressure (or dislodgement) test was performed by gradually increasing the fluid pressure below the sample. Plug dislodgment was indicated based on the response of the displacement sensor located on top of the plug. After the occurrence of dislodgement, the pressure below the sample was released, the plug was re-cured for 1 day, and a second pressure test was conducted.

The setup for the HPHT experiments (Fig. 5) was built with a 5.5″ (39.9 kg/m) steel casing. A heating vessel was placed enclosing the 5.5″ casing to provide an elevated temperature of 80 °C. The bentonite pellets A (approx. 6 m length) were placed on top of gravel (200 mm) and sand (100 mm) columns by using the batch technique. After bentonite pellets and water were placed, the casing was heated up and pressurized. The samples were cured with freshwater for 4 days. Two samples were prepared, each cured at a pressure of 4 and 180 bars, respectively. After curing, the pressure test was carried out by gradually reducing the top pressure to create a differential pressure. Plug dislodgment was indicated by a sharp drop in bottom pressure and an increase in the top pressure. The displacement of the plug was confirmed post the experiment. In these experiments, only one pressure test was performed.

Fig. 5
figure 5

The schematic drawing of large-scale 5.5″ HPHT setup and the setup realization

4 Results

4.1 Interfacial Properties of Bentonite

The tests using dry surface material were conducted on steel and sandstone. Results show that the cohesion and friction angle of the interface of bentonite/steel was 5.6 kPa and 10.20°, respectively. The bentonite/sandstone interface was observed to have a higher cohesion and friction angle of 7.8 kPa and 21.28°, respectively. The plot of the shear stress against the normal stress from the sliding tests under dry conditions is shown in Fig. 6a.

Fig. 6
figure 6

Shear and normal stress correlation from sliding tests performed in a dry- and b saturated condition

The interfacial properties of bentonite/saturated sandstone were measured under freshwater- and seawater saturating conditions. In the FW-saturated setting, the cohesion and friction angle of the bentonite/sandstone interface was 5.5 kPa and 14.03°, respectively. In the SW-saturated setting, the cohesion of the bentonite/sandstone interface was reduced to 2.7 kPa and the friction angle is similar at 14.57°. A sliding test was also conducted on an igneous rock saturated with freshwater. The interfacial constants of the bentonite/igneous rock were found lower than that of bentonite/FW-saturated sandstone, with a cohesion of 3.2 kPa and a friction angle of 6.84°. The relationship between shear stress and normal stress under saturated conditions is shown inFig. 6b. A summary of the measured interface cohesion and friction angles is provided in Table 3.

Table 3 Summary of the cohesion and friction angle at the interface of bentonite with various materials

4.2 Interfacial Shear Strength of Bentonite Plugs

4.2.1 Push-Out Test

A total of 10 samples were tested in the push-out test experiments with different combinations of curing conditions. The measured interfacial shear strength varied between 1.25 and 9.75 kPa. The density of the samples was within the range of 1.50–1.68 g/cm3. A summary of the plug strength and measured density is shown in Table 4. The interfacial shear strength of all tests is presented in Fig. 7.

Table 4 Summary of the results from push out tests with curing condition, measured interfacial shear strength, and density
Fig. 7
figure 7

Interfacial shear strength from the push-out test of bentonite plugs cured with various fluid salinity and pH grouped by the curing temperature (i.e. room temperature and 80 °C)

High interfacial shear strengths were observed by samples cured in freshwater, with strength values ranging between 4.5 and 9.75 kPa. In contrast, the samples cured in saline conditions showed strengths varying between 1.25 and 3.81 kPa. Sample 6, which was cured in freshwater with pH 10.5 and at 80 °C, produced an interfacial strength of 9.75 kPa. Sample 9, which was cured in 1.03 NaCl brine with similar pH and curing temperature, generated a lower strength of 2.89 kPa. Increasing the NaCl concentration to 1.06 SG further reduced the plug shear strength to 1.75 kPa, as observed by sample 10.

At room temperature, the interfacial shear strength of the freshwater samples increased from 4.5 kPa (sample 1) to 6.85 kPa (sample 2) with a pH increase from 7 to 10.5. A similar trend of strength gain with increasing pH was also indicated in samples cured with 1.03 NaCl brine at room temperature. The strength of these samples (3 and 4) increased from 1.25 kPa to 2.1 kPa with increasing pH. At 80 °C curing conditions, the gain in strength with increasing pH became less significant. The strength of freshwater-cured samples (5 and 6) at 80 °C increased from 8.17 to 9.75 kPa with increasing pH. Meanwhile, results from NaCl brine-cured samples (7, 8, and 9) at 80 °C, which were cured in pH 7, 9, and 10.5, respectively, had the least influence on fluid pH. These samples had a similar interfacial strength ranging between 2.9 and 3.8 kPa.

The interfacial shear strength of all plugs improved at a curing temperature of 80 °C. In the freshwater condition, it was observed that the strength of samples cured at pH 7 increased by 3.67 kPa (samples 1 and 5). The strength gain with increasing temperature was less observed in samples cured in saline and base conditions. The strength improvement of samples cured in freshwater pH 10.5 (samples 2 and 6) was observed to be reduced by 2.9 kPa. In the 1.03 NaCl brine curing condition, the strength of samples cured in fluid pH of 7 (samples 3 and 7) and 10.5 (samples 4 and 9) increased by 1.8 and 0.79 kPa, respectively. In general, the density of plugs cured in ambient temperature varied from 1.50 to 1.55 g/cm3. Under the same curing condition, the density of plugs cured at 80 °C was typically higher, ranging from 1.54 to 1.68 g/cm3.

Results from the exposure experiments are presented in Fig. 8 for both short- and long-term periods. Samples that were initially cured in the freshwater for 4 h show a decrease in strength with exposure time. The sample strength dropped from 11 to 9 kPa after 20 h of exposure to 1.03 NaCl brine with pH 9. The strength subsequently dropped to 7.1 and 4.0 kPa after 1 and 2 weeks of exposure, respectively. On the other hand, samples that were initially cured for 1 month maintained their strength more effectively when exposed to a higher salinity and base environment. The strength of the long-term sample dropped from 10.9 kPa to 10.1 after 2 months of exposure. In addition, the strength dropped from 8.7 to 6.3 kPa after 3 months of exposure.

Fig. 8
figure 8

Interfacial shear strength from the exposure experiment: a short-term and b long-term period. The indicated period is the cumulative period of curing and exposure

4.2.2 Small-Scale Hydraulic Dislodgment Test

A total of 15 tests were performed with different combinations of pellets, curing pressures and temperatures, and fluids. The results are presented in Fig. 9 for both pellets A and B.

Fig. 9
figure 9

Shear bond strength from small-scale hydraulic dislodgment tests of samples made of a Pellet A and b Pellet B

The shear strength of the freshwater-cured samples made from pellet A varied between 12.03 and 32.5 kPa under the given curing condition. The shear strength of freshwater-cured samples made from pellet B ranged between 10.7 and 32.5 kPa. At room temperature curing conditions, the plug shear strength of pellets A was slightly higher by around 1 kPa than those of pellet B. The strengths of plugs made from both pellets cured at 7 bars and 180 °C were similar at 32.5 kPa.

Results from pellets A and B that were cured at 7 bars and 80 °C show that the interfacial shear strength of samples cured in saline solutions reduced significantly compared to those cured in freshwater. In addition, the strength further dropped as the fluid salinity was increased from seawater to API brine. In tests using pellet A, the highest shear strength was produced by the freshwater-cured sample at 32.5 kPa. The strength was reduced to 12.03 kPa and 4.88, respectively for samples cured in seawater and API brine. Results from pellet B show that the shear strength of samples cured in freshwater, seawater, and API brine was 32.5 kPa, 7.2 kPa, and 5.9 kPa, respectively. At the other curing pressures and temperatures, pellet B that was cured in saline solutions also generated lower shear strength than those cured in freshwater. The shear strength of pellet B samples cured in seawater and API brine was ranging from 4.9 to 7.2 kPa and 3.3 to 7.2 kPa, respectively.

It was observed that the interfacial shear strength of the resulted plugs increased with curing pressure. The strength of pellet A- and B samples cured in freshwater slightly increased by 0.4 kPa and 0.7 kPa with a pressure increase of 7 bars, respectively. The same trend was observed in the seawater-cured samples made from pellet B. The strength of those samples increased from 3.3 to 4.9 kPa with a pressure increase of 7 bars. The interfacial shear strength of the plugs also improved with curing temperature. The strength of pellet A samples cured in freshwater and at 7 bars increased from 12.4 kPa to 32.5 kPa with a temperature increase from ambient to 80 °C. Similarly, the shear strength of pellet B samples cured in freshwater improved from 11.4 to 22.8 and to 32.5 kPa with increasing temperature from ambient to 60 °C and 80 °C, respectively. The strength of samples cured in a saline solution was less influenced by increasing temperature. The shear strength of pellet B samples cured in API brine at curing temperatures of ambient, 60 °C, and 80 °C was 4.2, 7.2, and 5.9 kPa, respectively.

4.2.3 Large-Scale Hydraulic Dislodgment Test

A total of nine tests were performed in the LPLT setup, and two tests in the HPHT setup. The test conditions and results are summarized in Table 5, and the measured interfacial shear strength is presented in Fig. 10. An example of the readings during the LPLT dislodgment test is presented in Fig. 11. In the first pressure test (Fig. 11a), it was observed that the displacement of the bentonite plug was initiated at a dislodgment pressure of 15 bars. Afterward, the pump was stopped and the pressure below the plug dropped instantaneously to a stabilization pressure of 11.3. During displacement, leakage was observed at the top of the plug for a short duration until the bottom pressure stabilized. The second pressure test (Fig. 11b), which took place after re-curing the plug, had a similar pressure, displacement, and flow profile. The general observation in LPLT tests shows that the dislodgment pressure from the second pressure test was higher than that from the first pressure test.

Table 5 Large-scale LPLT and HPHT experimental results
Fig. 10
figure 10

Calculated shear strengths from hydraulic dislodgement tests in LPLT and HPHT experiments

Fig. 11
figure 11

The response of bottom pressure and displacement during the a first and b second pressure test of LPLT-6

Samples LPLT-1, LPLT-2, and LPLT-3 were cured in freshwater and large casings of 7–5/8″, 9–5/8″, and 13–3/8″, respectively. Sample LPLT-1 had the highest initial shear strength at 6.6 kPa. The initial strength of plugs in larger diameters was smaller at 4.1 kPa (LPLT-2) and 3.0 kPa (LPLT-3). In the second pressure test, it was observed that the shear strength of these three samples increased to a similar value at approximately 13 kPa. This observation shows that the shear strength measured in the second pressure tests was less influenced by the casing size.

Four freshwater-based samples were prepared in a 7–5/8″ casing with gravitation placement. Sample LPLT-4, which was prepared by gravitation and cured at ambient pressure, had an initial strength of 1.9 kPa. The obtained strength was lower than that of sample LPLT-1, which was prepared with batch placement and cured in a similar setting. The gravitated sample LPLT-5, which was cured at a pressure of 4 bars, showed a significant increase of initial strength to 7.4 kPa. Other gravitated samples cured with additional sand layers (LPLT-6) and cured at a longer curing time (LPLT-7) of 30 days had a higher initial strength at approximately 11 kPa. The initial strength of the three gravitated samples (LPLT-5, 6, and 7) surpassed the initial strength of the batch-prepared sample of LPLT-1. All samples prepared by the gravitation placement method gained strength in the second pressure test, with an average increment of 4.5 kPa. The secondary strength of gravitated sample LPLT-6 at 13.5 kPa was comparable to those of batch-prepared samples (LPLT-1, 2, and 3).

Samples LPLT-8 and LPLT-9 were cured in seawater inside 9–5/8″ and 13–3/8″ casings and produced an initial shear strength of 8.3 and 6.1 kPa, respectively. Contrary to the freshwater-cured plugs, the gain in interfacial shear strength from the first- to second pressure test was less significant at 0.3 kPa in both samples LPLT-8 and LPLT-9. Assuming a negligible impact of casing size on the secondary strength, the difference of the secondary strength between the freshwater-cured sample (LPLT-6) and seawater-based samples of LPLT-8 and LPLT-9 was 4.8 and 3.2 kPa, respectively. Post-mortem observations indicated that samples cured in saline conditions have a soft texture, while the samples in freshwater are more solid (Fig. 12).

Fig. 12
figure 12

Post-mortem observations of the large-scale bentonite plug cured in (left) freshwater and (right) seawater

Two 5.5″ plugs tested in the HPHT experiment were cured at a temperature of 80 °C, one at a pressure of 4 bars (HPHT-1) and another at 180 bars (HPHT-2). The high pressure applied is within the range of the expected hydrostatic pressure at the deep wells. The results indicate that the initial strength of the plug increased nearly two-fold from 7.6 to 14.3 kPa with increasing curing pressure from 4 to 180 bar.

5 Discussion

5.1 Impact of Wellbore Fluid Properties on Bentonite Sealing Capacity

The results from both the push-out test and hydraulic dislodgment tests show that the interfacial shear strength of the bentonite plug is significantly reduced in a saline fluid curing condition. The measured hydraulic dislodgement pressures of pellet A from Fig. 9a and Table 5 are plotted against the ratio of sample height to diameter (H/D) in Fig. 13. The expected correlation follows the below equation by re-writing Eq. (2)

Fig. 13
figure 13

Relationship of dislodgement pressure with H/D for all small- and large-scale samples of pellet A in ambient conditions. The dislodgement pressure from the large-scale test was taken from the second pressure test

$${P}_{dis}=4\frac{H}{D}{\tau }_{s}$$
(3)

where H is the plug length and D is the inner diameter of the casing. A linear trendline was observed in both freshwater- and seawater-cured plugs, with R2 of 1.00 and 0.75, respectively. The average interfacial shear strength of freshwater- and seawater-cured samples was calculated from the curve slope using Eq. (3), which yields 13.32 and 9.12 kPa, respectively. The shear strength is related to the effective normal stress at the bentonite/steel interface following Eq. (1). Assuming that interfacial constants of bentonite/steel are \(c\) =5.6 kPa and \(\phi\) = 10.2° (Table 3), the normal effective stress of both the freshwater and seawater-cured bentonite can be calculated. The result (Table 6) shows that the \({\sigma }_{N}^{^{\prime}}\) from freshwater- and seawater- plugs are 42.5 and 19.6 kPa, respectively.

Table 6 Summary of the calculated shear strength and effective normal loads following results in Fig. 13

Existing research shows that the bentonite swelling pressure contributes to the effective normal stress acting at the interface (Akgün, H. et al. 2006; Pusch, R. 1992). Therefore, the reduced effective normal stress in the seawater-cured plug indicates that the presence of salt in swelling/contacting fluids suppresses the bentonite swelling. By mechanism, bentonite swelling upon wetting consists of two stages of crystalline swelling, which is driven by cation hydration, and diffuse double layer (DDL) swelling, which is the act of ion balancing due to osmotic pressure between the curing fluid and the fluid in the clay particles (Mitchell and Soga 2005; Mukherjee 2013). Previous research indicates a reduction of swelling pressure of various bentonite types in different saline solutions and concentrations (Castellanos et al. 2008; Herbert et al. 2008b). During clay hydration with saline fluid, Ye et al. (2015) reported that the DDL swelling is significantly inhibited compared to crystalline swelling. The higher ion concentration from salt in solution reduces the osmotic suction, and hence the DDL’s thickness. This might explain the plug strength reduction with increasing fluid salinity, which is observed in the present experimental results from the push-out test (1.03 vs. 1.06 SG NaCl brine) and small-scale hydraulic dislodgment test (seawater vs. API brine).

The interfacial properties of bentonite are influenced by the salinity of the saturating fluid at the interface between the bentonite and adjacent materials, as shown in the sliding test. This setting represents plug placement in an open hole, where pore fluid in formations is likely in direct contact with the plug at the interface. According to the results of the sliding test, the cohesion at the interface of bentonite and seawater-saturated sandstone is reduced by half compared to freshwater-saturated sandstone from 5.5 to 2.7 kPa (Table 3). However, the friction angle values did not materially change between the two tests.

The exposure experiments demonstrate the chemo-mechanical stability of hydrated bentonite exposed to an ionic fluid, which contained 4.22% wt. NaCl and has a pH of 9. In these experiments, the contact surface with the exposed fluid was limited at the top of the plug. The results show that both short-term (4 h) and long-term (1 month) freshwater-cured bentonite plugs were gradually reduced in shear strength after a certain period of fluid exposure. The long-term cured freshwater bentonite plug is found to be more stable and tolerate the exposure of the ionic fluid more effectively. The strength of the long-term cured plug is reduced by 0.9 and 2.5 kPa after exposure for 2 and 3 months, respectively ( Fig. 8b). On the other hand, the strength reduction of the short-term cured sample was more significant, up to 6 kPa after a short exposure period of 2 weeks ( Fig. 8a). The reduction of the shear strength might be due to the entry of the exposure fluid into the bentonite bulk, which changes the initial state of the clay–freshwater interaction. If ionic fluids penetrate the interface of the bentonite plug, the interface mechanical properties could be negatively altered, as shown by the results from the sliding test. The long-term cured sample might acquire low permeability and porosity that inhibit the transport rate of the exposed fluid into the bentonite more effectively. Previous studies report that the water permeability of a bentonite plug is very low, ranging from 10–3 to 10–4 milliDarcy (Clark and Salsbury 2003; Idialu et al. 2004; Towler et al. 2016). Existing studies also suggest that the low porosity and permeability of hydrated bentonite limit advective flow. Hence, the solute transport in hydrated bentonite is mainly through diffusion, which occurs at a slower rate than advective flow (Appelo 2013; Mitchell and Soga 2005).

The influence of fluid pH on the bentonite strength is captured in the push-out test using fluid pH variations of 7 (neutral pH), 9, and 10.5. For the given curing temperature and fluid type, the plug strength is observed to improve as the pH increases. Previous studies show that the swelling pressure of bentonites at the later stage of hydration reduces with increasing pH (Liu et al. 2018; Ramı́rez et al. 2002; Ye et al. 2014). During the early age of clay hydration in low (< 0.1 M) NaOH solution, the change in crystalline swelling is not significant and the role of OH to increase the DDL swelling is more prominent (Liu et al. 2018). In the present experiment of push-out tests, the mechanical stability of bentonite plugs was measured after a curing period of 24 h. Therefore, further investigations measuring the long-term mechanical stability of plugs cured at various elevated pH are still required.

5.2 Impact of Wellbore Operating Conditions on Bentonite

The plug interfacial shear strength changes with curing pressure. According to the results from small-scale hydraulic dislodgment tests, there was a slight increase in interfacial strength of 0.4–0.7 kPa with a pressure increase of 7 bars (Fig. 9). On the other hand, the gain in strength with increasing curing pressure is more substantial in large-scale tests. The strength of 7–5/8″ plugs cured at room temperature increased by 6.5 kPa with a pressure increment of 4 bars. At 80 °C, the 5.5″ bentonite plug gained strength by 7.9 kPa as the curing pressure was increased from 4 to 180 bars. This implies that the increase in plug strength with curing pressure is more significant for large-scale samples. This could be explained by the size effect and/or pellet compaction. In a large-scale setting, the placement of a large volume of pellets can lead to less uniform packing. Applying a higher curing pressure likely consolidates the pellets to form a more compressed plug. This behavior has also been observed in the experiments by Baille et al. (2010).

Experimental results show that the interfacial shear strength of bentonite plugs improves with temperature. In the small-scale hydraulic dislodgment test, increasing the curing temperature from ambient to 80 °C increases the strength by approximately 20 kPa, for both pellets of A and B cured in freshwater. This trend is confirmed by the results of the push-out tests. The pellets cured in freshwater with a pH of 7 show the highest gain in strength with increasing temperature. The increase in strength at higher temperatures is less significant for plugs cured in saline and base fluids. This implies that the high ionic strength present in the curing fluid counteracts the effect of curing temperature. In addition, the density of plugs cured at 80 °C is typically 0.1 g/cm3 higher than that cured in ambient conditions. Increasing curing pressure and temperature promote a higher interfacial shear strength and density in bentonite plugs. Consequently, the in-situ condition of deep wells should be beneficial to the mechanical stability of bentonite plugs.

5.3 Application of Bentonite for Deep Subsurface Wells

Bentonite plug is observed to be hydraulically tight to inhibit liquid flow at a differential pressure below the dislodgment pressure, as indicated by results of the LPLT hydraulic dislodgment tests. It was also observed that the dislodged bentonite resealed the casing cross-section hydraulically after being re-cured due to the nature of the self-healing properties of bentonite. A gain in interfacial shear strength is also observed after re-curing. The strength increase in freshwater-cured plugs is much more significant than in the seawater-cured plugs. Dislodgment experiments by Holl (2019) and Towler et al. (2020) indicate a similar finding that plugs cured and tested using freshwater generated higher dislodgement values after each cycle of failure/re-curing. Towler et al. (2020) also reported that the bentonite plug cured in highly saline conditions produces a consistent dislodgement pressure in each test cycle, similar to the findings presented in this work. According to Holl and Scheuermann (2018), during the dislodgement process, the applied axial load on the plug could induce compaction in the bentonite bulk. Subsequently, further compaction propels the fluid to penetrate the bulk and migrate through the defects, which could promote further expansion.

The influence of casing size and bentonite pellet placement on plug mechanical stability was evaluated in the LPLT hydraulic dislodgment test. Results from bentonite plugs cured in realistic casing sizes (7–5/8″, 9–5/8″, and 13–3/8″) show that the initial interface shear strength of plugs is inversely proportional to casing size. In the second pressure test, the strength of these plugs increased to a similar value at approximately 13 kPa. Two placement methods of batch and gravitation were performed to create a plug in the 7–5/8″ casing. The plug strength of the sample prepared with gravitation was 1.9 kPa, which is lower compared to the batch-prepared sample at 8.9 kPa (Table 5). The plug strength of gravitated samples increases substantially to 11.2 kPa after a 30-day curing period. Increasing the curing pressure by 4 bars and adding a 100 mm sand layer below the plug also promote a strength gain of the gravitated samples. This finding implies that the gravitational placement of bentonite pellet is a feasible technique to produce a mechanically stable plug in the field. Compared to the conventional cement plug placement job (e.g. balanced-plug method) that requires rig, the placement of bentonite pellets can be achieved in a rigless setting, which substantially reduces the abandonment cost. To address the size effect from upscaling the plug volume and other operational and placement aspects in a deeper well, we are currently conducting a field-scale test using a test well located in the RCSG laboratory.

National regulators and industry best-practice require that a permanent sealing material shall be verified after placement. The NOGEPA 45 standard by the Netherlands Oil and Gas Exploration and Production Association describes three verification methods and states that the plugs should be verified with at least one of these methods: 1) weight test (tagging), 2) pressure test, and 3) inflow tests (NOGEPA - OPCOM 2019). The pressure test verification is relevant to confirm the hydraulic sealing properties of the bentonite plug. The standard states that the pressure test validation should be performed at 50 bars above the value at which leakage via leak path would occur if the shut-off failed for a minimum test duration of 15 min. By extrapolating the averaged interfacial shear strength from the experiments, the required bentonite plug lengths in casings ranging from 7–5/8″ to 13–3/8″ to withstand 50 bars can be calculated. For bentonite application in freshwater curing fluid, a minimum plug length of 15, 21, and 30 m in 7–5/8″, 9–5/8″, and 13–3/8″, respectively, is sufficient to meet the verification requirement (Fig. 14a). For the application in seawater, a longer plug with a minimum length of 22, 30, and 43 m in 7–5/8″, 9–5/8″, and 13–3/8″, respectively, is sufficient to meet the standard (Fig. 14b).

Fig. 14
figure 14

The required length of bentonite plug cured in a freshwater (\(\tau\)-avg = 13.3 kPa) and b seawater (\(\tau\)-avg = 9.12 kPa) to meet the requirement of pressure test of 50 bars

6 Conclusions

In this work, a series of experiments were carried out to measure the interfacial properties and shear strength at the bentonite plug interface under various settings relevant to deep wells. Furthermore, the experimental results were compared to the Dutch regulatory requirements to show the outlook of bentonite applications as an alternative sealing material to plug deep wells. The following conclusions can be drawn from the present study:

  • The effect of salt concentration and fluid pH on the interfacial shear strength of bentonite plugs was studied. According to the results, the plug strength decreases with increasing salinity and increases with the fluid pH. The average interfacial shear strength of large-scale bentonite plugs was 13.3 kPa and 9.1 kPa for plugs cured in the freshwater and seawater fluids, respectively.

  • The interface mechanical properties of bentonite and steel, porous sandstone, and igneous rock were measured. The cohesion at the bentonite interface is reduced as the salinity of the pore fluid increases, while no significant change in the friction angle is expected.

  • The chemo-mechanical stability of freshwater-cured bentonite plug exposed to saline and base fluids was studied using the exposure tests. Results show that the interfacial plug strength is reduced after exposure. The long-term cured plug (1 month) is more stable and less influenced by the fluid change than the short-term cured plug (4 h).

  • The bentonite plug is hydraulically tight to prevent water flow at a differential pressure below the dislodgement pressure. After re-curing the dislodged plug, the hydraulic seal of the plug is restored due to the self-healing ability of bentonite, and the interfacial shear strength of the plug is improved.

  • The plug strength increases with curing temperature and pressure. The influence of curing temperature becomes less effective in a high ionic fluid environment, such as saline and alkaline fluids. A large-scale 5.5″ bentonite plug was successfully created in conditions of 180 bars and 80 °C, representing operating conditions of deep wells, with a shear strength of 14.3 kPa.

  • To simulate the downhole setting, large-scale bentonite plugs were prepared and tested in casing sizes of 7–5/8″ to 13–3/8″. The initial strength of the bentonite plug is inversely proportional to the casing diameter. The gravitational bentonite placement method was also presented to demonstrate a potential bentonite placement operation in the field, which could be applied rig-less and cost-efficient. The resulting plug is found to be mechanically stable.

  • The minimum length of a bentonite column in full-bore casing sizes from 7–5/8″ to 13–3/8″ was estimated using the current experimental data. A minimum plug length between 15 and 43 m would be sufficient in most cases to comply with the verification criteria of the Dutch regulators for permanently sealing deep geo-energy wells.