Copyright 2006, IADC/SPE Drilling Conference This paper was prepared for presentation at the IADC/SPE Drilling Conference held in Miami, Florida, U.S.A., 21–23 February 2006. This paper was selected for presentation by an IADC/SPE Program Committee following re-view of information contained in a proposal submitted by the author(s). Contents of the paper, as presented, have not been reviewed by the International Association of Drilling Contractors or Society of Petroleum Engineers and are subject to correction by the author(s). The material, as presented, does not necessarily reflect any position of the IADC, SPE, their officers, or members. Electronic reproduction, distribution, or storage of any part of this paper for com-mercial purposes without the written consent of the International Association of Drilling Con-tractors and Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The ab-stract must contain conspicuous acknowledgment of where and by whom the paper was pre-sented. Write Librarian, SPE, P.O. Box 833836, Richardson, TX 75083-3836, U.S.A., fax 1.972.952.9435.

Abstract Storing carbon dioxide (CO2) underground is considered the most effective way for long-term safe and low-cost CO2 sequestration. This recent application requires long-term wellbore integrity. A leaking wellbore annulus can be a pathway for CO2 migration into unplanned zones (other formations, adjacent reservoir zones, and other areas) leading to economic loss, reduction of CO2 storage efficiency, and potential compromise of the field for storage. This CO2 leakage through the annulus may occur much more rapidly than geologic leakage through the formation rock. The possibility of such leaks raises considerable concern about the long-term wellbore isolation and the durability of hydrated cement that is used to isolate the annulus across the producing/injection intervals in CO2-related wells. With the lack of industry standard practices dealing with wellbore isolation for the time scale of geological storage, a methodology to mitigate the associated risks is required. This requirement led to the need and development of a laboratory qualification of resistant cements and the long-term modeling of cement-sheath integrity. This article presents the results of a comprehensive study on the degradation of cement in simulating the interaction of the set cement with injected supercritical CO2 under downhole conditions. The methodology and the equipment are described for testing conventional Portland cement and measuring the evolution of its alteration process with time under CO2 conditions. Experimental details and analytical methods are discussed. Data relating cement-strength loss and CO2 penetration in Portland cement are presented. The evolution of cement chemistry and porosity with time is highlighted by scanning electron microscopy analyses, back-scattered electron images, and Hg-porosimetry measurements. A first fluid-flow-geochemistry modeling for Portland cement is proposed. The results are compared to equivalent studies on a new CO2-resistant material; the comparison shows significant promise for this new material. This CO2-resistant material will

enable the hydrocarbon production industry to store the burnt residue over the long term in a safer and more responsible manner.

Introduction Storing carbon dioxide (CO2) underground is considered the most effective way for long-term safe and low-cost CO2 se-questration1,2. There are three main types of geological reser-voirs3 with capacity sufficient to store captured CO2: depleted oil and gas reservoirs, deep saline aquifer reservoirs, unmi-nable coal beds. The reservoirs need to be at a depth greater than 800 m so that the CO2 is in a supercritical state at a tem-perature and a pressure above its critical point (31.6°C, 7.3 MPa). These pressure and temperature conditions allow stor-ing CO2 in a relatively small volume. The ideal storage site would involve high pressure but at the lowest possible tem-perature to be in the most dense properties of the supercritical CO2. For these reasons, regions with low geothermal gradients are preferable. Piping CO2 emissions for underground injection is not a novel concept and is already often used for the purposes of enhanced oil and gas recovery4,5,6. However, this application of CO2 in-jection is not intended for long-term storage, which is a more recent concept that needs a long-term wellbore integrity strat-egy to be developed. Indeed, the major risk associated by the public with CO2 injection is a well failure, which may result in escape of CO2 that will migrate upwards. The likelihood of a sudden escape of all CO2 stored in an underground reservoir is extremely small. The main risks are: CO2 and CH4 leakage, seismicity and ground movement (subsidence or uplift). Failure of the cement, in the injection interval or beyond it, may create preferential channels for CO2 migration back to the surface. This may occur on a much faster time scale than geological leakage. It is hence important to explain that wellbore integrity will en-sure that CO2 stays underground for several hundred years and beyond. Conventional materials used for well isolation in oil or gas production are Portland-based cement systems. These systems present the advantage to be low cost and efficient for well service applications when considering a 20 year-long well life. But cement is known to be thermodynamically unstable and not durable in CO2-rich environments. It tends to strongly degrade7, once exposed to such acid gases, by reacting with calcium hydroxide formed from hydrated calcium silicate phases. As carbonates are dissolved in low pH environment, the cement-carbonation process will not become a self-

IADC/SPE 98924

Mitigation Strategies for the Risk of CO2 Migration Through Wellbores V. Barlet-Gouédard and G. Rimmelé, Schlumberger; B. Goffé, CNRS/ENS*; and O. Porcherie, Schlumberger

2 IADC/SPE 98924

plugging effect in the cement sheath. Some data8 have been already published under this type of environment. Long-term isolation and integrity of CO2 injection wells clearly must be improved to ensure long-term environmental safety. A good simulation of wellbore conditions for CO2 storage application, as pressure, temperature but also CO2-rich fluids (wet supercritical CO2 fluid and CO2 dissolved in water fluid) must be demonstrated. With the lack of industry-standard practices dealing with well-bore isolation for the time scale of geological storage, a meth-odology to mitigate the associated risks is required. This re-quirement led to the need and development of a laboratory qualification of resistant cements and the long-term modeling of cement-sheath integrity. This article presents the results of a comprehensive study on the degradation of cement in simulat-ing the interaction of the set cement with injected supercritical CO2 under downhole conditions. The methodology and the equipment are described for testing conventional Portland ce-ment and measuring the evolution of its alteration process with time under CO2 conditions. The results are compared to equivalent studies on a new CO2-resistant material.

Experimental Sample preparation To study the effects of carbonic acid on hydraulic Portland cements, two cylindrical sample geometries have been used: 0.5-inch-diameter x 1-inch (1.27-cm OD x 2.54-cm) and 1-inch-diameter x 2-inch (2.54-cm OD x 5.08-cm). All cement slurries were prepared according to API Specifica-tion 10, section 59, and using fresh water. Before mixing the slurry, an antifoam agent, a dispersant and a retarder were added to the mix water to optimize the main slurry properties. The cement samples were cast by slowly pouring the degassed slurry down the cubic mold before launching the curing cham-ber. The samples were cured for 72 hours at 20.68 MPa [3000 psi] and 90°C [194°F]. The cubic samples were removed from the moulds and placed in water. Then the cubic samples were cored to obtain either 0.5-inch-diameter or 1-inch-diameter cylindrical samples. Finally, the core samples were cut to 1-in or 2-in length. Experimental Apparatus Cement carbonation was performed under static conditions using the following experimental set-up and procedure. The static conditions were considered as realistic simulation of the CO2-exposure conditions at the formation/cement sheath inter-face, except around the perforations where the exposure is un-der a dynamic state during CO2 injection. So static carbona-tion has been selected as representing downhole condi-tions10,11. The experimental set-up accommodates temperatures from 30°C [86°F] to 300°C [572°F] and pressure from 1 to 50 MPa [145 - 7252 psi] to cover the targeted applications ranges. Ex-periments with wet supercritical CO2 and with water saturated by dissolved CO2 were performed in a titanium annular vessel (TAV, Figure 1) and in a simple titanium cylinder vessel (TSCV, Figures 2 and 3). To avoid thermal gradient into the TAV where core samples are arranged along a crown, the temperature is controled by two resistive-heating coils: one is introduced in a titanium finger placed in the center of the ves-

sel while the other is wrapped around the vessel perimeter. In the TSCV, any vertical thermal gradient is elimi-nated/compensated using two independent resistive-heating coils wrapped at two levels around the vessel. The temperature is measured using a K-type thermocouple (four external ther-mocouples in the TAV and two internal ones in the TSCV) and monitored using a microprocessor-based controller. Pres-sure is monitored using a class 1 Bourdon manometer and electronic pressure gauge. Test conditions are fixed for all ex-periments to 90°C [194°F], 28 MPa [4061 psi]. The tests are performed at different durations: 44 hrs (~2 days), 88 hrs (~4 days), 188 hrs (~1 week), 523 hrs (~3 weeks), 1006 hrs (~6 weeks) and 2033 hrs (3 months). The volume content of water and CO2 are 10% and 90% respectively of the free volume of the vessel (volume of the vessel minus the volume of the core samples) at 0.1 MPa [14.50 psi] and room temperature. There is no fluid mixing in the vessel. At 90°C [194°F], 28 MPa [4061 psi], water is liquid while the CO2 is supercritical12,13,14 (Figure 4) with a specific volume of 0.65. The choice of the vessel depends on the number of core samples and their size (different types of systems are not tested in the same vessel). For any given type of core sample, no difference was observed between experiments made in the TAV or in the TSCV.

Safety considerations

The pressure test cells were designed for a maximum working pressure of 50 MPa. All the materials used in the cell construction were certified. Prior to delivery, each cell was hydraulically pressure tested at 150 MPa. Multiple pressure relief valves and pressure gauges were used on the system to minimize the risk of trapped pressure and/or overpressure. The cells were placed in a bunker room with no access permitted. All control functions and data recording were remotely performed using the Internet network possibility (remote control from the office desktop). After each test, the cell is inspected to prevent any damage to the cell during a subsequent CO2 test.

Test procedure The set-up and starting procedure included ten steps: 1- Cores samples were distinctly identified, measured,

weighed and placed into the vessel. When two cores are stacked, they are separated by a Teflon washer to prevent fluid capillarity (Figures 3 and 5).

2- A volume of water, measured using a graduated test tube, was poured into the vessel.

3- The vessel was closed and connected to the CO2 pressure line (CO2 bottle, manometer).

4- Vessel was cooled to 2-4°C [35.6 to 39.2°F] using a re-movable cooling blanket wrapped around the vessel (TAV) or in an ice box (TSCV) (Figure 3).

5- Valves connecting the CO2 bottle to the vessel were opened.

6- As a result of the temperature difference between the CO2 bottle (at room temperature) and the vessel 4°C [39.2°F], CO2 condensed in liquid phase inside the vessel. When the vessel is totally filled with liquid CO2 (after about 1/2 hour, detected by equalization of pressure between bottle and vessel), the cooling circuit is removed, the CO2 bottle is

IADC/SPE 98924 3

closed and the vessel is totally isolated, except through the manometer.

7- The heating process was started. With temperature, pres-sure increased very quickly along steep isochores corre-sponding to high specific volume (0.9).The leak valve pro-gressively adjusted the pressure to reach the target of 28 MPa [4061 psi] at 90°C [194°F], corresponding to the spe-cific volume of 0.65 (Figure 4).

8- Temperature was stabilized (after about one hour). 9- Internet cameras were installed/set up. Pressure was con-

tinuously monitored with the electronic pressure gauges. 10- Any pressure drop during the experiment was compensated

using the CO2 pump connected between the CO2 bottle and the vessel. Any such pressure decrease would be the result of CO2 dissolution in water during the 2 first days and mineral carbonation afterwards.

The experimental set-up in the TSCV was designed to get a quick overview of the behavior of cements with supercritical CO2. The set-up simultaneously obtained data for the three situations: CO2 dissolved in water (bottom of reactor), water dissolved in CO2 (top of reactor), and interface between the two media. For this experiment, core samples were disposed in three crowns (figure 5): the bottom samples were submerged in water, the upper crown is completely exposed to CO2. The intermediate crown simulates the conditions at the water-CO2 interface. In the TAV, the samples were placed in two crowns, the bottom crown submerged in water and the top crown in supercritical CO2 (Figure 5).

The stopping and breakdown procedure includes seven steps: 1- Electric power of the heater was switched off, and the thermal shield around the vessel was removed. 2- Pressure naturally decreased slowly with temperature (more than one hour), initially along the isochore 0.65 from 90°C [194°F] to the CO2 critical point 32°C [89°F] and then along the CO2 critical curve to 6 MPa [870 psi] at room temperature (Figure 4). In the experiment with the TAV, the water was re-moved from the vessel through the discharge valve and re-tained in a graduated test tube. The pH was measured for the first time at this stage. 3- CO2 pressure was then slowly decreased through the leak valve (about 15 min.) to prevent any possible spalling of the core sample and to avoid freezing of the leak pipe. 4- When the internal pressure reached 0.1 MPa [14.50 psi], the vessels were disconnected and opened. 5- Residual water (TAV) or total water (TSCV) was sampled and measured in a graduated test tube. The pH was then meas-ured with a pHmeter. 6- The core samples were removed from the vessel, photo-graphed and weighed. Their dimensions were also measured. 7- Finally, mechanical strength, chemical and microscopic composition were systematically analyzed. Alteration Measurements

Carbonic-acid corrosion on the core samples was indirectly measured: the pH of the sample water, the weight and volume change of the cores, the change in density and porosity, the

evolution of compressive strengths, the thickness of the car-bonation front, and microscopic measurements. The alteration front is determined in cutting each core sample to prepare a thin section in the axial plane. The front is meas-ured both directly and with a scanning electron microscope averaging several zooms. The microscopic evolution was determined by scanning elec-tron microscopy and back-scattered electron analysis through the entire core section. Additional SEM-EDS (Energy Disper-sion Spectroscopy) analyses and Xray imaging were per-formed in the same material with a quantitative EDS device (Quantex on Hitachi 2500). Compressive strengths were determined with an ADAMEL press. As length-to-diameter ratio of the cylindrical samples was always equal to two, the resulting strengths were directly used. The compressive strength of each core sample was measured before and after CO2 attack without additional treatment. The porosity variation was measured with a Mercury-porosimetry instrument. Mercury intrusion porosimetry (MIP) evaluated the pore size over a wide range (theoretically from 0.003 µm to 360 µm). MIP measurements were carried out by two different laboratories on 0.5 inch diameter by 1 inch length cylindrical samples, dried at 100°C at standard atmosphere during 60 hours. Two intrusion-extrusion cycles were performed to determine irreversible and reversible intrusion volumes. After an additional extrusion-intrusion cycle, the initial volume was always reached to within a small error, indicating that no significant damage occurred to the sample during the first intrusion. The maximum intrusion pressure is 60000 psi (413 MPa) in the best case, and the equilibrium time is 20 seconds. Results Comprehensive study on the degradation of Portland ce-ment Our kinetic study tested conventional Portland cement (class G, 1.89 g/cm3, water/cement ratio equal to 0.44) and measured the temporal evolution of its alteration process under CO2 ex-posure. The alteration of Portland cement during CO2 attack is a very effective process. A sharp alteration front is clearly observed at the rim of the samples already after 44 hrs of CO2 attack. For 0.5-inch-diameter x 1-inch (1.27-cm OD x 2.54-cm) cy-lindrical samples, the thickness of this alteration front (e) in-creases with time from about 1-2 after 44 hrs to 5-6 mm after 3 weeks of attack (Figures 6 and 7). After 6 weeks, the front reaches the sample core (7 mm deep). For 1-inch-diameter x 2-inch cylindrical samples (2.54-cm OD x 5.08-cm), after 3 months of attack, the front is 8 mm thick in CO2 dissolved in water and reaches the core of sample (12 mm deep) in wet su-percritical CO2 fluid (Figure 7). The alteration rate slightly de-creases with time (Figure 7). Furthermore, at each moment, the alteration front is always slightly thicker (e+1mm) for samples located in wet supercritical CO2 fluid (top of vessel) than for samples immerged in water fluid with dissolved CO2 (bottom of vessel). This observation clearly indicates that the alteration process is more efficient in the wet supercritical CO2 phase than in the CO2 dissolved water fluid (Figures 6 and 7). The alteration process follows a diffusion law according to

4 IADC/SPE 98924

equation 1 [e=0.26.sqrt(time)] for wet supercritical CO2 fluid and equation 2 [e=0.22.sqrt(time)] for CO2 dissolved in water (Figure 8a). The model predicts that after 20 years of CO2 at-tack, the alteration front reaches 90 mm in CO2-saturated wa-ter and 110 mm in wet supercritical CO2 respectively (Figure 8b). Portlandite [Ca(OH)2] and Calcium Silicate Hydrates (CSH) of cement are progressively consumed to produce carbonates (aragonite, vaterite and/or calcite), silica amorphous gel and water. The attacked cement cores systematically display a par-ticular geometry/concentric zones at different test durations on back-scattered electron images: from the core towards the rim of samples, the cement sample consists of (1) an uncarbonated zone, (2) a dissolution front, (3) a carbonation front, and (4) a dissolution back-front (Figure 9). The uncarbonated zone is the internal cement, surrounded by the alteration front. The dissolution front is a high-porosity zone where CSH progres-sively dissolves. The carbonation front is a 50 to 100 µm-large zone of very low porosity. The dissolution back-front is lo-cated at the rear of the carbonation front. In this zone, neo-formed carbonates are dissolved, increasing the porosity and resulting in a strong degradation of the cement. Figure 10 shows the compressive-strength results at each mo-ment with several 1.27-cm diameter x 2.54-cm cylindrical samples (1 inch OD x 2-inch length) under the same pressure-temperature conditions. After six weeks in wet, supercritical CO2 fluid, the relative strength decreases to 65%. In carbon dioxide dissolved water fluid, it decreases to 33% (Figure 10). However, compressive strength measurements are very dis-persed in both fluids. It may come from the heterogeneity of the samples due to the carbonation process (Figures 6 and 9). During the compressive strength measurements, the carbona-tion layer cracks. These results cannot be used as absolute compressive strength values but indicate clearly the weakness of the carbonate layer or the interface at this front. Figures 11 and 12 show the temporal weight and density variations of 1.27-cm diameter x 2.54-cm cylindrical samples (1 inch OD x 2-inch length) in wet, supercritical CO2 and in CO2 -saturated water at 90°C [194°F] and 28 MPa [4061 psi]. The weight of the cores rapidly increases by about 10% in both fluids during the first four days (88 hours). Then, it in-creases slightly from 10 to 12% in CO2-saturated water and from 10 to 15% in wet supercritical CO2 after 6 weeks. A similar trend is observed in both fluids, although the weight increases more in the wet, supercritical CO2 fluid (Figure 11). This observation results from a higher carbonation in samples in wet supercritical CO2 than those in CO2-saturated water. However, high value dispersion is observed for samples at-tacked by CO2-saturated water. The density evolution follows the same trend in the first week: it strongly increases from 16 to around 18 ppg in both fluids (Figure 12). The density continues to increase up to 18.5 ppg after 6 weeks of attack in wet supercritical CO2 while it seems to stabilize to about 18 ppg in CO2-saturated water (Figure 12). Supercritical CO2 that is injected very quickly becomes hy-drated in the reservoir. The resulting acidic water may cause a severe deterioration of the cement sheath. These experiments at downhole temperature, pressure and CO2 conditions accu-

rately display the potential interaction and degradation proc-esses between the rock formation and the cement sheath. The pH of water in equilibrium with Portland cement cores initially equals to 13. After CO2 attack, core samples are stored in water and the pH is measured at equilibrium. After each duration of CO2 attack, the new equilibrium pH is around 7 up to 3weeks and decreases again up to 6 after six weeks. Such CO2 attack-related decrease of pH has already been re-ported in the literature15,16. This decrease results from the reac-tion between CO2 and calcium from calcium silicate hydrates (CSH) or Ca(OH)2 coming from the Portland cement hydra-tion (Figure 13). Water-permeability measurements have been performed on the Portland-cement-core samples. No discrimination has been possible with this method before and after the CO2 attack because the permeability always stayed below the detection limit of 8 microdarcy. The initial Hg-porosity of 32.5% is reproducible and repeat-able. After CO2-attack, the porosity evolves during the test pe-riod with different behavior in each fluid (Figure 14). In wet supercritical CO2, the porosity continuously decreases, from initially 33% to 15% after 6 weeks. It rapidly decreases (from 33% to 27%) in the first two days, before the decrease slows down over the next 6 weeks (Figure 14a). In CO2 -saturated water, the porosity changes in two distinct phases. During the first phase of approximately three weeks, the porosity rapidly decreases from 33% to 9%. At this stage, the carbonation front has quite completely entered the cores. During the second phase after three weeks, the porosity in-creases from 9% to 21% up to 6 weeks (Figure 14 b).

Comparison with a new carbon dioxide resistant material The kinetic methodology presented above compares Portland cement to a new CO2-resistant material of 2 g/cm3 (17 ppg) density, which contains a reduced amount of Portland cement. The samples tested in both fluids do not show any visible car-bonation front after 3 months (Figure 15), in contrast to the Portland-cement-core samples. Even at the microscopic scale, a homogeneous pattern with a low calcium carbonate forma-tion is observed. The cement mass of 1.27-cm diameter x 2.54-cm length cylindrical samples keeps a good integrity and does not show microcracking. The compressive strength before and after carbon dioxide at-tack initially decreases, then stabilizes up to 3 months in both CO2 phases with very good repeatability, in contrast with the Portland cement behavior (Figure 16). Parallel to this kinetic study, compressive strength tests were performed on core samples with densities of 1.5 and 1.89 g/cm3 (12.5 ppg and 15.8 ppg). The results confirm the stable behavior of CO2-resistant material over a large density range. For 15.8 ppg and 12.5 ppg systems, the compressive strength decreases from 50 MPa to 30 MPa after three months of CO2 attack in both fluids and from 38 MPa to about 20 MPa after one month respectively. Figures 17 and 18 show the weight and density variations of 1.27-cm diameter x 2.54-cm cylindrical samples of 2g/cm3 with time in wet supercritical CO2 and in CO2-saturated water at 90°C [194°F] and 28 Mpa [4061 psi]. Again, a good stabil-

IADC/SPE 98924 5

ity is observed with time in either CO2 phase after an initial weight increase around 4%. The density variation stays very low with time with an initial increase around 6%. As for Portland cement, the pH data decrease from 11-12 ini-tially to 7 after CO2 attack (Figure 13). Water permeability measurements have been performed on the new CO2-resistant material. The measurement method did not permit to detect any change due to the CO2 attack. The initial Hg-porosity is low and equals to 23% (Figure 19). In wet supercritical CO2, the maximal porosity remains stable up to 3 months. The entry pore size slightly increases in the first two days andt stays stable up to 3 months (Figure 19a). In CO2-saturated water, the porosity increases from 23% to 25%. Concerning the entry pore size, a similar trend is ob-served as the one in wet supercritical CO2 (Figure 19b). These porosity results confirm the trend observed with other physical parameters (weight, density, compressive strength, microstructural characterisations), that CO2-resistant material remains stable from the first two days to a longer duration. Its behavior is comparable in both wet supercritical CO2 and CO2 -saturated water. The CO2 resistant material exhibits behavior drastically different from the one observed with Portland cement. Indeed, carbonation occurs very rapidly in the first days of CO2 attack and is then limited. Conclusions In the framework of carbon capture and storage application, this study leads to conclude that: • A new methodology accurately simulates downhole

conditions in pressure, temperature and CO2 fluid content, for both wet supercritical CO2 and CO2-saturated water. This procedure has been validated; it is reproductible and repeatable.

• Portland cement is not resistant enough to wet supercritical CO2 or to CO2-saturated water. Its alteration is characterized by a complex series of concentric fronts after CO2 attack in both fluids. This alteration is a very effective process following a diffusion law. The model predicts an alteration front of 100 mm after 20 years of CO2 attack possibly destroying zonal isolation and triggering casing corrosion since pH of interstitial fluid significantly decreases.

• An initial sealing by carbonation is followed by a dissolution stage, which starts earlier in CO2-saturated water than in wet supercritical CO2. Indeed, carbonation does not continuously plug Portland cement.

• In contrast, an homogeneous pattern with a limited car-bonation threshold is observed with the CO2-resistant ma-terial, which has a good mechanical behavior over a wide density range. This material remains comparably inert in both wet supercritical CO2 and CO2-saturated water.

• Weight, density, compressive strength, microstructural characterizations and Hg porosity measurements confirm the good stability of the CO2-resistant material.

Acknowledgments The authors would like to thank the management of Schlum-berger for permission to publish this paper and the laboratory

staff of Schlumberger Riboud Product Center and of CNRS/ENS for the experimental work. References

1. Gielen, D, “The future role of CO2 capture and storage. Results of the IEA-ETP model”, Report Number EET/2003/04, Paris November, 2003.

2. Sarmiento, J.L. and Gruber, N., “Sinks for anthropogenic carbon” 30-36, Physics Today, 2002.

3. Bachu. S, “Sequestration of CO2 in geological media: criteria and approach for site selection in response to climate change”, Energy conversion and management, 41, Issue 9, 953-970, 2000.

4. Mizenko, G.J., “North Cross (Devonian) Unit CO2 Flood: Status Report” Paper SPE/DOE 24210, 1992.

5. McDaniel Branting, J.K., Whitman, D.L., “The feasibility of using CO2 EOR Techniques in the powder river basin of Wyoming” Paper SPE 24337, Casper, Wyoming, 1992.

6. Power, M.T., Leicht, M.A., Barnett, K.L., “Converting Wells in a Mature West Texas Field for CO2 Injection” Paper SPE 20099, 1989.

7. Bruckdorfer R.A., “Carbon Dioxide Corrosion in Oilwell Cements”, paper SPE 15176 presented at the rocky Montain Regiona Meeting of the SPE held in Billings, MT, May 19-21, 1986.

8. Onan D.D., “Effects of Supercritical Carbon Dioxide on Well Cements” paper SPE 12593 presented at the 1984 Permian Basin Oil & gas recovery Conference, Midland, TX, March 8-9, 1984.

9. Recommended Practice for Testing Well cements, API Recommended Practice 10B, Twenty- Second Edition, December 1997, American Petroleum Institute.

10. V. Barlet-Gouédard, T. S. Ramakrishnan, K. Bennaceur, M. Supp, B. Goffé, G. Rimmelé, E. Nelson, “Testing of CO2-resistant material for well integrity under Carbon dioxide supercritical environment”, Wellbore integrity workshop, Houston-Texas, April 2005.

11. V. Barlet-Gouédard, T. S. Ramakrishnan, K. Bennaceur, M. Supp, B. Goffé, G. Rimmelé, E. Nelson, “Mitigation Strategies for CO2 Migration through Wellbores”, 4th annual conference on carbon capture and sequestration, Virginia, May 2005.

12. Hollister, L.S.: “Information intrinsically avalaible from fluid inclusions” in “Fluid inclusions: applications to petrology” Hollister and Crawford ed’s., Mineral association of Canada, Short course handbook, 6, 1-12, 1981.

13. Blencoe, J.G., Naney, M.T. and Anovitz, L.M.: “The CO2-H2O system: III. A new experimental method for determining liquid-vapor equilibria at high subcritical temperatures” American Mineralogist, 86, 1000-1111, 2001.

14. Blencoe, J.G: “The CO2-H2O system: IV. Empirical, isothermal equations for representing vapor-liquid equilibria at 110-350°C, P≤150 MPa” American Mineralogist, 89, 1447-1455, 2004.

15. T. Van Gerven, D. Van Baelen & al. “Influence of carbonation and carbonation methods on leaching of metals from mortars” Cement and Concrete Research, 34, 149-156, 2004.

16. T. Van Gerven, Johny Moors & al. “Effect of CO2 on leaching from a cement-stabilized MSWI fly ash” Cement and Concrete Research, 34, 1103-1109, 2004.

6 IADC/SPE 98924

Figure 1. (a) View on the two levels of 1-inch-diameter (2.54 cm) cylindrical core samples for CO2 experiments in the titanium annular vessel (TAV). TAV in open state before experiment (b), with the first level of core sample crown (c), and in close state before heating (d). Notice the titanium finger at the center of the device that allows heating the center of the vessel together with the external coil.

Figure 2. (a) View of the three levels of the half-inch-diameter (1.27 cm) cylindrical core samples for CO2 experiments in the titanium simple cylinder vessel (TSCV). Note that the three cores at the middle are not cut in order to study the interface between the supercritical CO2 fluid and CO2-saturated water fluid. These samples are only used for microscopic investigations, not for compressive strength measurements. (b) Introduction of a crown of samples into the TSCV.

IADC/SPE 98924 7

Figure 3. (a) Titanium simple cylinder vessel (TSCV) with screws, washer and Teflon slices. (b) TSCV connected to the CO2 line and cooled in an ice box to condense CO2 before heating at 90°C (note the pressure of about 6 MPa [870 psi] corresponding to the pressure of the CO2 bottle at room temperature).

8 IADC/SPE 98924

Figure 4. (a) Phase diagram12 for CO2 showing the experimental conditions (white circle, 90deg.C/28 Mpa [194deg.F/4061 psi], specific volume 0.65) and the pressure-temperature path (gray arrows) followed during the beginning of the experiments (see text). (b) Comparison of the CO2 and H20 phase diagrams12 showing the different states for water (liquid state on the critical curve) and CO2 (supercritical state on the specific volume 0.65 curve).

IADC/SPE 98924 9

Figure 5. Simplified scheme (above, lateral view; below, view in cross-section) showing the position of the different core samples weither in wet supercritical CO2 fluid or in CO2-saturated in water, in both titanium annular vessel (TAV) and titanium simple cylinder vessel (TSCV). [not to scale]

10 IADC/SPE 98924

Figure 6. Photos showing the alteration of Portland cement after CO2 attack during two days and three weeks for instance. The thickness of the alteration front (e) increases with time and is greater in wet supercritical CO2 than in CO2-saturated water fluid, at each test duration.

0

2

4

6

8

10

12

14

0 500 1000 1500 2000Time (hours)

Thic

knes

s of

the

alte

ratio

n fro

nt (m

m)

Wet supercritical CO2 fluidCO2-saturated water fluid

Figure 7. Evolution of alteration with time for a Portland cement in wet supercritical CO2 fluid and in CO2-saturated water fluid.

IADC/SPE 98924 11

Equation1: y = 0.2622xR2 = 0.9998

Equation2: y = 0.2182xR2 = 0.9995

0

2

4

6

8

10

12

14

0 5 10 15 20 25 30 35 40 45 50

Sqrt Time (hours)

Thic

knes

s of

the

alte

ratio

n fro

nt (m

m)

Wet supercritical CO2 fluidCO2 saturated water fluid

Equation1: y = 0.2622xR2 = 0.9998

Equation2: y = 0.2182xR2 = 0.9995

0

20

40

60

80

100

120

0 100 200 300 400Sqrt Time (hours)

Thic

knes

s of

the

alte

ratio

n fr

ont (

mm

)

Wet supercritical CO2 fluidCO2 saturated water fluid

Figure 8. (a) Evolution of alteration with sqrt(time) for a Portland cement in wet supercritical CO2 fluid and in CO2-saturated water fluid. (b) Model prediction of Portland cement alteration front thickness after 20 years under CO2 –rich environment: 90 to 110 mm in wet supercritical CO2 fluid and in CO2-saturated water fluid respectively.

(a)

(b)

12 IADC/SPE 98924

Figure 9. (A) Photo of a Portland cement half-core sample after a 88-hr-CO2 attack (about 4 days) in wet supercritical CO2 fluid. (B) Back-Scattered Electron images though the entire section of cement core (location of photos in Fig. A). Core is decomposed into different zones of variable porosity due to the carbonation and dissolution processes. (C) Detail of Fig. B focussing on the transition between the altered zone (dissolution and carbonation processes) and the uncarbonated internal part of cement core. (D) Detail of Fig. C showing the thin carbonation front between both dissolution fronts.

IADC/SPE 98924 13

0

5

10

15

20

25

30

35

40

45

50

0 400 800 1200

Time [h]

Com

pres

sive

stre

ngth

[MPa

]

Wet supercritical CO2

CO2-saturated water

Error bar : +/-10%

6 weeks3 weeks1 week

Figure 10. Compressive strength on half-inch diameter Portland cement cylinders with time in wet supercritical carbon dioxide fluid and in carbon dioxide saturated in water at 90°C [194°F] under 28 MPa [4061 psi].

1.00

1.02

1.04

1.06

1.08

1.10

1.12

1.14

1.16

1.18

0 400 800 1200Time [h]

Norm

aliz

ed w

eigh

t

Wet supercritical CO2

CO2-saturated w ater

Figure 11. Weight variation in half-inch diameter Portland cement cylinders with time in wet supercritical carbon dioxide fluid and in carbon dioxide saturated in water at 90°C [194°F] under 28 MPa [4061 psi].

14 IADC/SPE 98924

15.0

15.5

16.0

16.5

17.0

17.5

18.0

18.5

19.0

0 400 800 1200

Time [h]

Den

sity

[ppg

]

Wet supercritical CO2

CO2-saturated water

Figure 12. Density variation in half-inch diameter Portland cement cylinders with time in wet supercritical carbon dioxide fluid and in carbon dioxide saturated in water at 90°C [194°F] under 28 MPa [4061 psi].

0

2

4

6

8

10

12

14

Portland cementpH(i)

Portland cementpH(f)

CO2-resistantmaterial pH(i)

CO2-resistantmaterial pH(f)

pH o

f wat

er in

equ

ilibr

ium

with

cem

ent c

ores

44 hrs88 hrs188 hrs 523 hrs1006 hrs

Figure 13. Evolution of pH of water in equilibrium with cement cores, at each test duration, for a Portland cement and CO2-resistant material. pH(i)=pH before CO2 attack, pH(f)=pH after attack.

IADC/SPE 98924 15

Wet supercritical CO2

0

5

10

15

20

25

30

35

0.001 0.01 0.1 1 10Pore diameter [µm]

Cum

ulat

ive

poro

sity

[%] 0h

44h188h

523h1006h

CO2-saturated water

0

5

10

15

20

25

30

35

0.001 0.01 0.1 1 10Pore diameter [µm]

Cum

ulat

ive

poro

sity

[%] 0h

44h188h523h1006h

Figure 14. Cumulative porosity obtained during the first mercury intrusion in Portland cement core samples at several durations, in wet su-percritical CO2 fluids (a) and in CO2-saturated water fluid (b).

Figure 15. Photos showing the alteration of CO2-resistant material (density=2 sg) after CO2 attack during two days and three weeks, for in-stance. In contrast to Portland cement, no clear alteration front has been deciphered at each test duration, neither in wet supercritical CO2 nor in CO2-saturated water fluid.

(a) (b)

16 IADC/SPE 98924

0

10

20

30

40

50

60

0 500 1000 1500 2000 2500Time [h]

Com

pres

sive

str

engt

h [M

Pa]

Wet supercritical CO2

CO2 dissolved in water

3 weeks1 week 6 weeks 12 weeks

Figure 16. Compressive strength on half-inch diameter CO2 resistant material cylinders (density=2 sg) with time in wet supercritical carbon dioxide fluid and in CO2 saturated in water at 90°C [194°F] under 28 Mpa [4061 psi].

1.00

1.02

1.04

1.06

1.08

1.10

1.12

1.14

1.16

1.18

0 400 800 1200

Time [h]

Nor

mal

ized

wei

ght

Wet supercritical CO2

CO2-saturated water

Figure 17. Weight variation in half inch diameter CO2 resistant material cylinders (density=2 sg) with time in wet supercritical carbon dioxide fluid and in carbon dioxide saturated in water at 90°C [194°F] under 28 MPa [4061 psi].

IADC/SPE 98924 17

15.0

15.5

16.0

16.5

17.0

17.5

18.0

18.5

19.0

0 400 800 1200

Time [h]

Dens

ity [p

pg]

Wet supercritical CO2

CO2-saturated water

Figure 18. Density variation in half inch diameter CO2 resistant material cylinders (density=2 sg) with time in wet supercritical carbon dioxide fluid and in carbon dioxide saturated in water at 90°C [194°F] under 28 MPa [4061 psi].

Wet supercritical CO2

0

5

10

15

20

25

30

0.001 0.01 0.1 1 10

Pore diameter [µm]

Cum

ulat

ive

poro

sity

[%] 0h

44h188h523h2033h

CO2 saturated water

0

5

10

15

20

25

30

0.001 0.01 0.1 1 10

Pore diameter [µm]

Cum

ulat

ive

poro

sity

[%] 0h

44h188h523h2033h

(a) (b)

Figure 19. Cumulative porosity obtained during the first mercury intrusion in CO2-resistant material core samples (density=2 sg) at several durations, in wet supercritical CO2 fluids (a) and in CO2-saturated water fluid (b).