Method for the determination of residual carbon dioxide saturation using reactive ester tracers

Abstract

The mechanisms for storage of CO₂ in rock formations include structural/stratigraphic, mineral, solubility and residual trapping. Residual trapping is very important in terms of both containment security and storage capacity. However, to date, the contribution from residual trapping (i.e. immobilisation of supercritical fluid via capillarity in pore spaces) is still relatively difficult to quantify accurately. Using a laboratory-based testing program, this study demonstrates the feasibility of using reactive ester tracers (i.e. triacetin, propylene glycol diacetate and tripropionin), which partition between a mobile water phase and a stationary supercritical CO₂ phase, to quantify the residual CO₂ saturation, S_gr, of a rock formation. The proposed single-well test involves injecting these tracers into the subsurface, followed by CO₂ saturated water, where the ester tracers slowly hydrolyse to form products with differing partition coefficients. After a suitable period of time, allowing for partial hydrolysis, water containing the tracer mixture is produced from the subsurface and analysed using gas chromatography mass spectrometry (GCMS). A numerical simulator of the tracer behaviour in a reservoir is used to explain the differential breakthrough of these tracer compounds during water production to estimate S_gr. Computer modelling suggests that the use of esters tracers to determine CO₂ residual saturation is a potentially robust method. The supercritical CO₂/water partition coefficients directly dictate the amount of time that each tracer spends in the CO₂ and water phases. As such for modelling of tracer behaviour and estimating S_gr, knowing the tracer partition coefficient is essential; in this paper, the first laboratory study to determine the partition coefficients of these reactive ester tracers is described.

Introduction

Carbon capture and storage (CCS) technologies are increasingly being refined and tested in pilot, field and commercial-scale demonstrations around the world. Examples include Weyburn, Canada (Wilson, 2004); the Frio Brine Project, USA (Hovorka et al., 2006); the Otway Project, Australia (Sharma et al., 2007); In Salah, Algeria (Ringrose, 2009); Cranfield, USA (Hovorka et al., 2009, Han et al., 2010) and Sleipner, Norway (Torp and Gale, 2004). One of the primary goals of these projects is to better understand and define the dynamics (i.e. plume evolution and changes over time in the various storage mechanisms) of CO₂ storage. The storage capacity of the target reservoir in these projects is limited by the trapping efficiency of CO₂ in the pore space of the rock matrix. Trapping efficiency and the contribution from each type of CO₂ trapping remain the biggest uncertainties in storage capacity assessments (Bachu et al., 2007, Michael et al., 2010). Therefore, a more comprehensive understanding of the trapping mechanisms of CO₂ in the rock matrix can produce accurate models with improved prediction of migration and storage potential. This allows for a calculation of the overall capacity of potential sites and reduces uncertainties related to containment security.

Proposed storage scenarios most commonly involve injecting CO₂ into rock formations at depths where the pressure and temperature exceeds its critical point of (7.38 MPa, 31.1 °C), typically greater than 800 m. There are four primary trapping mechanisms for long term storage of CO₂ in saline aquifers or depleted natural gas fields: structural/stratigraphic, mineral, solubility and residual trapping (Metz et al., 2005, Suekane et al., 2008). Over the lifetime of a CO₂ storage site, the contribution from each mechanism continually changes (see Fig. 1). Residual trapping is considered to be of greatest importance in the first several hundred years (Han et al., 2010, Metz et al., 2005); however, the contribution of residual trapping in current storage capacity estimates constitutes a large uncertainty. Quantifying the residual CO₂ saturation of a rock formation is, therefore, important in assessing the viability of large volume/long term storage reservoirs.

This study demonstrates the feasibility of using reactive ester tracers (i.e. triacetin, propylene glycol diacetate and tripropionin) to quantify the amount of residually trapped CO₂ through an integrated program of laboratory experiments and computer simulations. Accurately determining the tracer partition coefficients and understanding the hydrolysis kinetics of these tracers is essential for both determining the feasibility of the proposed tracers in field studies and any subsequent modelling of the tracer behaviour to estimate residual saturation. Laboratory high pressure/high temperature sampling experiments were implemented to determine the CO₂/water partition coefficients and understand the in situ hydrolysis kinetics. This tracer research program was designed to replicate the reservoir conditions at the Otway Project in Victoria, Australia to assess the potential for a field trial as a component of the CO2CRC Stage 2B Residual Saturation and Dissolution Test (Zhang et al., 2011).

This paper is focused on methods to quantify the mechanism of residual CO₂ saturation, also known as residual CO₂ trapping, using reactive tracers. Residual CO₂ trapping occurs when the supercritical fluid displaced by water is immobilised via capillarity in pore spaces. The non-wetting phase (i.e. supercritical CO₂) is trapped in pore spaces by the wetting phase (i.e. water) in the pore throats. Estimates of residual trapping range from 20% to 40% of overall trapped CO₂ (Bachu et al., 2007, Flett et al., 2007, Bennion and Bachu, 2008, Suekane et al., 2008). Laboratory studies that mimic or replicate subsurface conditions using core floods have been an invaluable tool in studying the various trapping mechanisms of supercritical CO₂ in rock formations (Muller, 2011). Suekane et al. (2008) attempted to quantify residual and solubility trapping using the well-characterised Berea Sandstone. They were able to estimate maximum trapped gas saturation of 25–28%, over a range of temperatures (38–50 °C) and pressures (7.6–10.0 MPa), representing depths from 750 to 1000 m.

To date, measurement or prediction of residual trapping beyond the core scale in a particular formation remains difficult due to abundant subsurface heterogeneities. According to Han et al. (2010), research in this area is hampered by three key factors: (i) not all trapping mechanisms are considered simultaneously in the models, (ii) simplification of the models is such that heterogeneities are effectively neglected and (iii) heterogeneity and associated data from field trials are either absent or not used. The primary limitation of core-flood tests to determine rock properties is that the core is only representative of the very near-well environment and that extrapolations of these results is unlikely to be accurate. A pulsed neutron test or reservoir saturation test (RST) can be used to estimate the amount of displaced formation water and from this to infer the residual CO₂ saturation; however, this test, like laboratory tests, is also only representative of the near well environment (Hovorka et al., 2006).

To better understand rock behaviour on a larger scale, several CCS projects have used chemical tracers. They have been used to distinguish native and injected CO₂ and to aid in the measurement of the first arrival of CO₂ in the formation (Freifeld et al., 2005, Boreham et al., 2011, Johnson et al., 2011). Inert perfluorocarbon, Kr and SF₆ were used in the Frio Brine Project, USA (Freifeld et al., 2005) and for In Salah, Algeria (Ringrose, 2009). These tests have yielded information which has improved the understanding of CO₂ migration and has also provided constraints for models. For the CO2CRC Otway Project in the Otway Basin in Victoria, Australia, SF₆, Kr, perdeuterated CH₄ and 1,1,1,2-tetrafluoroethane were used to observe the evolution in CO₂ saturation (Stalker et al., 2009, Boreham et al., 2011). Results to date have shown discrete changes throughout the reservoir with differences in CO₂ arrival times and gas/tracer compositions at varying sampling depths.

A single well “push–pull” test using Kr and Xe as tracers is currently being considered at the CO2CRC Stage 2B test (Zhang et al., 2009, Zhang et al., 2011) to characterise reservoir heterogeneity and quantify residual gas saturation. Single well “push–pull” inert gas tests consist of first injecting a suite of tracers dissolved in formation water (e.g. SF₆ or noble gases) into the target reservoir and then producing water samples from the same well used for injection (see Fig. 2). In addition, for “push–pull” tests, each inert tracer is exposed to approximately the same volume of rock during both injection and production. If the supercritical CO₂–water partition coefficients of each tracer are known, then the relative differences in the dispersion seen in the breakthrough profiles (see Fig. 4) between tracers can be correlated to the residual gas saturation, S_gr, using a simulation of tracer behaviour (Zhang et al., 2011). However, these models are susceptible to large errors due to several factors, including a lack of partitioning coefficient data for many species in different formation waters with variable concentrations of dissolved cations and anions and varying pH, temperature and pressure (Curren and Burk, 2000). Formation heterogeneities, flow channelling, diffusion and other phenomena also contribute to considerable errors in these models. Multiple tracers are typically injected for these experiments and the breakthrough curves for each tracer are compared; the risk is that the differences in these curves will be minor and the observed retention times might be nearly identical (see Fig. 4). Nevertheless, modelling of these differences is the basis for determining residual saturation, potentially leading to large errors in the estimates (Zhang et al., 2009, Zhang et al., 2011).

This paper proposes an alternative to this method using reactive ester tracers which, after injecting the primary or “parent” tracer compounds, undergo a subsurface hydrolysis reaction to generate “daughter” tracers (see Fig. 3). Amides, which are known to hydrolyse, are not typically used in this type of test due to the much harsher conditions (very low or very high pH, high temperature) necessary for the hydrolysis reaction. This technique using ester tracers is based on a patent by Deans (1971) where esters (e.g. ethyl acetate) were used to determine residual oil saturation in depleted oilfields (i.e. one that is at residual oil saturation). This patent details a field experiment where ethyl acetate is injected into a depleted oilfield, followed by water, and then allowed to ‘soak’ for several days while the ethyl acetate partially hydrolyses to yield ethanol and acetic acid. During water production, the unreacted and more hydrophobic ethyl acetate preferentially partitions into the oil phase, while the more hydrophilic ethanol preferentially partitions into the water phase. In this case, compared to the inert tracers, there is a significant difference in the retention time allowing for a more sensitive and accurate determination of the residual saturation (see Fig. 4). Chromatographic/reaction models of the concentration profiles, in which the movement of analytes is primarily dictated by phase partitioning, were shown to be effective in determining the residual oil saturation level in this reservoir (Deans, 1971, Tomich et al., 1973).

The use of multiple tracers combining both inert and reactive tracers is potentially a powerful tool to reduce the ambiguities caused by using only inert tracers (Tomich et al., 1973, Field and Pinsky, 2000, Geyer et al., 2007). The purpose of this study was to identify several reactive ester tracers and test their usefulness for determining the quantity of residually trapped CO₂ in geological formations under a wide range of conditions. The reactive ester tracers will have a higher partition coefficient into CO₂ compared to water and are referred to in this paper as CO₂ partitioning tracers. The corresponding hydrolysis products of the esters will, compared to the parent esters, be partitioned into the water more and are referred to in this paper as water partitioning tracers.

The laboratory characterisation of selected tracers was undertaken at pressure and temperature regimes appropriate for injection into the target reservoir at the CO2CRC field site. The critical parameters set for assessing the suitability of individual esters as tracers in field trials are as follows:

• Environmental safety.

• High solubility in water.

• Detectability of tracers at low concentrations using GCMS.

• Suitable hydrolysis kinetics.

• Differential partitioning of parent and daughter products between supercritical CO₂ and water.

• Minimal effects on the surface properties of the rock matrix.

Once a suite of suitable chemicals had been identified an experimental program to determine their supercritical CO₂/water partition coefficients was set-up. This required the design and commissioning of a high-pressure cylinder which allowed for the sampling of fluids at reservoir conditions. The selection criteria for the proposed tracers will be addressed in detail in the Discussion section of this manuscript.