Application of Noble Gases to the Viability of CO₂ Storage

Abstract

Unequivocable evidence for warming of the climate system is a reality. An important factor for reducing this warming is mitigation of anthropogenic CO₂ in the atmosphere. This requires us to engineer technologies for capture of our carbon emissions and identify reservoirs for storing these captured emissions. This chapter reviews advances made in understanding multiphase interactions and processes operating in a variety of subsurface reservoirs using noble gases. We begin by discussing the types of reservoir available for carbon storage and the mechanisms of viable CO₂ storage, before summarising the physical chemistry involved in data interpretation and the sampling/sample storage techniques and requirements critical to successful sample collection. Theory of noble gas partitioning is interspersed with examples from a variety reservoirs to aid our knowledge of long term CO₂ storage in the subsurface. These include hydrocarbon reservoir and natural CO₂ reservoirs. In these examples we show how good progress has been made in using noble gases to explain the fate of CO₂ in the subsurface, to quantify the extent of groundwater interaction and to understand CO₂ behaviour after injection into oil fields for enhanced oil recovery. We also present recent work using noble gases for monitoring of subsurface CO₂ migration and leakage in CO₂ rich soils, CO₂ rich springs and groundwaters. Noble gases are chemically inert, persistent and environmentally safe and they have the potential to be extremely useful in tracing migration of CO₂. It is imperative that the many upcoming pilot CO₂ injection studies continue to investigate the behaviour of noble gases in the subsurface and develop suitable noble gas monitoring strategies.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

1.1 The Need for CO₂ Storage

Climate change and its effects are emerging as one of the most pressing environmental issues of the 21st century. It is generally accepted that increases in anthropogenic carbon dioxide (CO₂) emissions since the industrial revolution are a key factor in the climate warming observed to date (IPCC 2007). The period between 1995 and 2006 recorded eleven of the twelve warmest years in the instrumental record of global surface temperature, since 1850. The most recent assessment by the Intergovernmental Panel for Climate Change (IPCC) indicated that the global surface temperature has increased by 0.74 ± 0.18 °C between 1906 and 2005 (Fig. 1) (IPCC 2007). The IPCC report clearly stated “Warming of the climate system is unequivocal, as is now evident from observations of increases in global average air and ocean temperatures”.

Fig. 1

This time series is compiled jointly by the Climatic Research Unit and the UK Met. Office Hadley Centre showing mean global air temperature over the last ~150 years. The record is being continually up-dated and improved (Brohan et al. 2006)

This warming of the Earth’s climate has also resulted in a rise of sea levels due to melting of continental ice sheets and thermal expansion of the oceans. The 2007 IPCC report documented that global sea levels rose by 2.7 ± 0.7 mm between 1993 and 2003. The report also predicted a rise of 18–59 cm from 1990 to the 2090s, in addition to an unspecified amount that could come from partial melting of the large ice sheets covering Greenland and Antarctica. More recent studies have shown that this is probably a conservative estimate and rises of up to 1 m are not implausible (Rahmstorf 2010).

It is clear that reduction of CO₂ concentration in the atmosphere is paramount to limiting climate change in the near future and ameliorating the climate in the longer term. IPCC projections suggest a global surface temperature rise of a further 1.1–6.4 °C during the 21st century, as CO₂ concentrations rise from ~370 to ~700 ppm. This rise is largely due to global emissions of CO₂ from burning of fossil fuels at a rate of approximately 25 Gigatonnes of CO₂ equivalent per year (Marland et al. 2008). To reduce this value over the coming decades will require a combined approach of lower emissions from more fuel-efficient/alternative fuel technologies and removal of the on-going CO₂ emissions that produced as a by-product of power generation (Fig. 2).

Fig. 2

Diagram illustrating the life cycle chain of fossil fuel use with carbon capture and storage technology. CO₂ capture at the point source of emission allows the CO₂ to be transported to a suitable storage reservoir. Three different subsurface storage sites are being considered and tested; (1) Empty (depleted) oil and gas reservoirs, (2) Saline aquifers and (3) Unminable coal seams. Figure credit to Scottish Carbon Capture and Storage, School of Geosciences, University of Edinburgh

1.2 What is Geological Carbon Storage?

Geological carbon storage is the final aspect of the Carbon Capture and Storage (CCS) process. The CCS process involves capturing CO₂ from large point source emitters such as power stations, refineries or other industrial processes, compressing it to produce a supercritical fluid, then transporting, and then injecting it into geological formations that can act as natural underground storage sites. CO₂ becomes a supercritical fluid above the critical temperature of 31.1 °C and critical pressure of 7.39 MPa. Supercritical CO₂ has the properties of both a gas and a liquid, in that it expands to fill its container as a gas would whilst retaining the density of a liquid, making it easier to transport and inject than in gaseous form. The CO₂ injected into the geological carbon storage site must then remain isolated from the atmosphere for a significant period of time (1,000 to 10,000 years). In its simplest form, a geological carbon storage site will consist of a layer of permeable and porous reservoir rock situated beneath a layer of impermeable rock that will act as a seal preventing CO₂ migration to the surface.

There are three main types of proposed underground storage sites:

1.2.1 Depleted Oil and Gas Reservoirs

Oil and gas reservoirs consist of permeable and porous reservoir rocks overlain by impermeable cap rocks that have stored hydrocarbon liquids and gases over many millions of years. Those that have reached the end of their productive life are ideal for CO₂ storage. CO₂ can be pumped into the reservoirs, using the boreholes originally drilled to extract the oil and gas, to fill the empty pore space left by removal of hydrocarbons. The geology of these reservoirs is often well characterized to maximise oil recovery therefore fluid behaviour in these systems is well understood.

However, those reservoirs which have been subjected to the injection of water or gas to increase reservoir pressure and aid oil recovery may require extra production of the hydrocarbons in order to make space for the CO₂ (Haszeldine 2009). This technique is known as enhanced oil recovery (EOR) and can produce 10–15 % more oil. EOR is a mature technology that has been operating since the early 1970s. In addition, the natural gas industry has been reliant on underground natural gas reservoirs for many decades and enhanced gas recovery (EGR) is also a viable technique. These reservoirs could also be adapted for CO₂ injection and long-term storage. Depleted oil and gas reservoirs are expected to provide potential for storing 920 Gt CO₂ (IEA 2009)—45 % of the total CO₂ output from 2000 to 2050 assuming the IEA’s business as usual scenario for CO₂ emissions (i.e. no significant reduction in emissions).

1.2.2 Deep Saline Formations

Saline formations, also known as saline aquifers, are rock formations saturated with salt water that are unsuitable for supplying potable water. These exist worldwide, not just in hydrocarbon provinces and potentially could store large volumes of CO₂. CO₂ injected into these saline formations would partially dissolve into the saline water, storing the CO₂. Additionally, in some cases carbonate minerals could precipitate, locking up the CO₂ permanently. Recent estimates suggest the capacity for storage in geological reservoirs in North West Europe alone could be as much as 800 Gt CO₂ (with most of this in deep saline reservoirs), with other estimates suggesting that up to 10,000 Gt CO₂ could be stored globally (IEA 2009). However, these estimates are probably somewhat optimistic as they assume static injection efficiency of 5–10 %, when dynamic injection of CO₂ at the Utsira aquifer in the North Sea currently fills only 0.2 % of the pore space. Further study and the injection of CO₂ into other aquifers is required to fully determine the storage capacity and long term storage viability of saline aquifers.

1.2.3 Deep (Un-Minable) Coal Seams

Injection of CO₂ into deep un-mineable coal seams is also a possibility, where the CO₂ will be stored on the surface pores and fractures of the coal. This has the additional benefit of forcing the methane, which is already present within the coal beds, from the coal and so increasing recovery of the coal bed methane. This could offset the cost of CO₂ injection and even if the additional methane extracted is not of economic use, coal can adsorb twice as much CO₂ by volume as methane. Hence, conventional coal bed methane extraction and subsequent CO₂ injection will still result in net storage of CO₂.

1.2.4 Additional CO₂ Storage Options

There are several alternative options for CO₂ storage which include biological sequestration (Marland and Schlamadinger 1999), deep ocean sequestration (Brewer et al. 1999) and CO₂ injection into deep ocean carbonate sediments (House et al. 2006) or basalts (Goldberg et al. 2008) to form carbonate minerals. However, all of these are, with the exception of ocean sequestration, in their infancy and do not offer the necessary volume or security of storage required for large scale CO₂ storage at present (Baines and Worden 2004a).

Injection of large amounts of CO₂ into the marine environment may acidify large amounts of sea water (Haszeldine 2009) and ocean storage is not a permanent method of CO₂ storage. There is a potential risk that CO₂ will over time return to atmosphere, as there is not a distinct physical barrier to prevent the CO₂ from doing so. For these reasons this method of storage has been deemed unsuitable at the current time and is currently outlawed by international marine treaties.

In this chapter we focus on the depleted oil and gas reservoirs as this is the CO₂ storage setting that has received most attention from the scientific community to date and provides important lessons directly relevant to storage in other settings.

1.3 Mechanisms of Viable CO₂ Storage

The key requirement for successful geological CO₂ storage within the Earth is the safety and security of CO₂ retention over significant timescales. It is essential that protocols for monitoring and risk assessing CO₂ storage sites and potential leakage pathways are developed. The existence of CO₂ reservoirs that have been isolated from the atmosphere for millions of years (Ballentine et al. 2001; Gilfillan et al. 2008) demonstrates the feasibility of storing CO₂ for significant periods of time but understanding the exact behaviour of CO₂ in these systems is of paramount importance. Secure CO₂ storage can be achieved by ensuring that the stored CO₂ is immobile and trapped by one of the four mechanisms outlined below.

1.3.1 Physical Trapping: Stratigraphic and Structural Storage

Supercritical CO₂ has a lower density than water, so when CO₂ is injected into a reservoir it will ascend through the microscopic pore spaces between the grains in the reservoir rock to the top of the reservoir. Therefore, it is essential that a suitable impermeable cap rock such as a shale, clay or mudstone overlies the reservoir to act as a physical barrier to the CO₂ and prevent migration of the CO₂ from the reservoir. Cap rocks of oil and gas fields have trapped buoyant hydrocarbon fluids in this manner for millions of years.

1.3.2 Residual or Hydrodynamic Storage

Microscopic pore spaces which exist between grains in the reservoir rock are analogous to a foam sponge. CO₂ can enter ‘dead end’ pores in a similar fashion to how air is trapped within a foam sponge. This is known as residual trapping and is the reason why soaking a sponge in water cannot be achieved by submersing it: the sponge requires squeezing to expel the residually trapped air and replace it with water.

1.3.3 Dissolution Storage or Solubility Trapping

As CO₂ is readily soluble in water, it will dissolve in the formation water trapped in the pore spaces within a reservoir. The water containing dissolved CO₂ is denser than the formation water surrounding it. This dense water sinks to the bottom of the reservoir, trapping the CO₂. Provided that the formation water does not migrate from the reservoir this is a secure trapping mechanism.

1.3.4 Mineral Storage or Mineral Trapping

CO₂ dissolves in water to form a weak acid (carbonic acid) and this acid can dissolve and react with the mineral grains present in the reservoir rock. The reaction of the bicarbonate ions produced when the CO₂ dissolves into the formation water with calcium, magnesium and iron from silicate minerals such as micas, clays, feldspars and chlorites will form carbonate minerals. These will be produced in the interior of pores in the reservoir rock and have the potential to ‘lock away’ the CO₂ within the rock. This is the most permanent trapping mechanism. However, under normal reservoir conditions this process is very slow, typically taking thousands of years (IPCC 2005).

2 Sample Collection and Analytical Techniques

Within the Colorado Plateau and surrounding Rocky Mountain region there are at least nine producing or abandoned gas fields that contain up to 2,800 billion m³ of natural CO₂ (Allis et al. 2001). These fields act as a natural analogues of CO₂ repository sites and hence intensive studies have been undertaken to provide an understanding of CO₂ behaviour within them. Data included in this chapter are predominantly from five producing gas fields, namely McCallum Dome, (Jackson Co, CO); Sheep Mountain, (Huerfano County, CO); Bravo Dome, (Harding County, NM); McElmo Dome and the related field Doe Canyon, (Montezuma County, CO); St Johns Dome (Apache County, AZ) and eight natural CO₂ seeps at Green River, (Emery and Grand Counties, UT). All of these sites produce CO₂ in extremely high concentrations, averaging 95–99 % CO₂, 1–4 % N₂, 0.1–1 % He and other trace gases (Allis et al. 2001). The majority of these samples were taken directly from producing wellheads that tap the natural gas reservoirs, although samples from several ‘shut in’ wellheads with high gas pressures were also collected. Well sample sites were chosen on site to provide a wide range of depths and spatial distribution across the individual reservoirs. Two field-sampling techniques were used to collect these samples.

2.1 Well Head Collection Methods

The first method utilized ‘Swagelok^©’ 300 ml stainless steel sampling cylinders sealed at both ends with high-pressure valves. The cylinders were baked at 150 °C and evacuated prior to sampling use. The cylinders were attached directly to the wellheads before the gas had undergone any form of processing. A length of high pressure hosing was attached to the other end of the cylinder to act as an exhaust. The cylinders were then flushed with gas from the wellhead for 5 min and then purged 6 times. The purging process involves filling the cylinder with well gas, closing the valve to the wellhead, and then releasing the gas. This ensured that there is virtually no air contamination present in the sample. This method was used for wellheads that have high flow rates and typical wellhead pressures ranged from 1.38 to 3.45 MPa.

2.2 Water and Low Pressure Gas Collection Methods

The second collection technique used a 70 cm long piece of 3/8″ diameter refrigeration grade copper tubing as the sample container. This method was used for sampling of low-pressure wellheads with lower flow rates and the Green River seeps. The copper tubing is contained in an aluminium holder with clamping devices at each end and provides approximately 10 cm³ of gas for analysis (Fig. 3). In the field, high pressure hosing was secured to one end of the copper tubing, which was connected to a gas regulator attached directly to the wellhead by a suitable adapter. The remaining end of the tubing was connected to another section of tubing with a second length of high pressure hosing so a duplicate sample could be obtained. A third length of high pressure hosing was attached to the second length of copper tubing to act as an exhaust and prevent turbulent backflow of atmospheric gases into the sample. Gas from the wellhead was then pumped into both sections of tubing at a regulated pressure of 0.20 MPa (2 atm). The entire line was flushed with gas from the wellhead for 10 min. After flushing, the clamp furthest from the wellhead was closed sealing the end of the copper tube by cold welding. The other clamps were then shut in turn working back up the line to the wellhead. These clamps are then left closed until the samples are ready for analysis. Secure sample storage is achieved as copper is virtually impervious to noble gases (Hilton et al. 2002). Sample analysis was carried out by connecting one end of the copper tube to the mass spectrometer and releasing one of the clamps.

Fig. 3

Schematic of the apparatus used for noble gas sampling with copper tubing [from Becker (2005)]. The plastic hose is attached directly to the well head using a suitable regulator, CO₂ gas is allowed to flow through the copper tube until sample collection is complete. The tubing is then crimped at both ends

The Green River seeps (Sect. 7.1) were collected using a similar copper tube procedure. All of the seeps were actively releasing CO₂ within pools of water. To sample this gas a plastic funnel connected to a length of high pressure hosing was positioned over the bubbling gas, and submerged into the water. The water provides a seal from the atmosphere, preventing air contamination of the samples. The high pressure hose was directly connected to two copper tubes, allowing duplicate samples to be collected. Gas was allowed to flow freely into the copper tubes for a minimum of 10 min. The clamps were then shut sequentially working back from the clamp furthest from the seep. The water samples collected at St. Johns Dome (Sect. 7.3) and on the Kerr property (Sect. 7.4) were collected in the same manner using copper tubes after flowing water through the tubes for 5 min. The samples were either taken directly from producing well heads or from depth in groundwater pools or wells using a peristaltic pump and a weighed piece of hosing.

2.3 Issues of Long Term Sample Storage

Collection using the high pressure cylinder technique has been the convention in both the commercial and academic sectors. However, re-analysis of a suite of well gas samples after several years of lab storage has highlighted a serious issue, specifically that the double valved Swagelok cylinder approach is prone to air contamination. Over a period of two years, between summer 2005 and summer 2007, samples with ⁴⁰Ar/³⁶Ar ratio of ~22, 000 had been reduced by ingress of air to a ⁴⁰Ar/³⁶Ar ratio ~3,000. The mechanism by which this air contamination had occurred was not specifically identified, but because the partial pressure of noble gases in the atmosphere is significantly higher than in the well gases, any pathway for noble gas transfer will result in an influx of atmospheric noble gases to the cylinders. The high pressure valves were attached to the cylinders via steel screw threads wrapped in PTFE tape and the valve tips were also Teflon based. In a noble gas world where all metal systems are the norm, including linkages and valve tips, it is likely that one or both of these potential weak points are the source of the contamination. Samples are still collected in high pressure cylinders as required by health and safety issues at point of collection and due to commercial shipping regulations but in an attempt to alleviate this problem, the gases are rapidly transferred to an all metal storage device (currently metal spheres with a welded Conflat fitting and a standard copper tipped valve).

2.4 Analytical Techniques Specific to CO₂ Samples

At the University of Manchester, where some of the work reported in this chapter was undertaken, a specific vacuum line was constructed to handle the clean-up requirements of large volumes of well gas. This consisted of a calibrated volume containing a high pressure Baratron® capacitance manometer, a Ti sponge furnace, a Zr-Al SAES getter, cold finger, a quadrupole mass spectrometer and both an ion and turbo pump. The quadrupole was essential for ensuring that the samples had been cleaned up thoroughly prior to admission into the mass spectrometer clean-up/gas extraction line. The separate pumping system also ensured that no hydrocarbon or CO₂ was able to enter the main clean up line.

Well gas samples were extracted from either the copper sample tubes or from stainless steel cylinders into the volume containing a high pressure Baratron® capacitance manometer. The pressure was recorded, allowing calculation of the volume of CO₂ administered. The sample was then expanded onto a Ti sponge furnace to remove CO₂ and any other active gases. This furnace was preheated to 1,173 K (900 °C) with each sample held at this temperature for 10 min, and then cooled to room temperature over a period of 15 min. Approximately 10 g of Ti sponge material was capable of removing 1 L of CO₂, permitting multiple analyses of the same sample and high precision analyses of the rarer noble gas isotopes of Kr and Xe. These volumes of CO₂ would quickly saturate a Zr-Al SAES getter typically found on noble gas extraction lines and is a necessary step in preparation of CO₂ well gas samples for analysis. Subsequently, the sample was exposed for 10 min with a Zr-Al SAES getter heated to 523 K.

3 Noble gases in Geological Reservoirs

3.1 Introduction

Noble gases are chemically inert elements that exist as extremely volatile gases at standard temperature and pressure. Their lack of reactivity results in no change in their isotopic compositions via chemical processes but allows them to be used as a tracer of physical processes, which can alter both their isotopic and relative elemental compositions. In the context of storage in geological reservoirs, their relative abundances are sensitive to partitioning between different fluid phases such as oil, water and gas, enabling their use as tracers of physical processes even if they have subsequently been partitioned from their host fluids. The three key sources of noble gases in natural gas reservoir fluids are the crust, the atmosphere and the mantle. Each of these possesses different isotopic compositions permitting resolution of the different sources (see Fig. 4).

Fig. 4

Schematic diagram of a natural gas reservoir, highlighting the three distinct components and their corresponding isotopes which may occur in a natural gas reservoir. The radiogenic noble gases (e.g. ⁴He, ²¹Ne and ⁴⁰Ar) are produced within the Earth’s crust by the radioactive decay of U, Th and K. Atmospheric noble gases (e.g. ²⁰Ne and ³⁶Ar) are input into the gas phase as a result of equilibration with the groundwater at recharge. Within areas of continental extension or magmatic activity mantle derived noble gases (e.g. ³He) may also be present. The distinct elemental and isotopic composition of the noble gases from these three components permits the contribution from each to be quantitatively resolved from any crustal fluid (redrawn from Ballentine et al. 2002)

3.2 Sources of Noble Gases in a Subsurface Reservoir

Noble gases on the Earth are derived from two principle sources: those trapped during the planetary accretion process which are commonly known as ‘primitive’, ‘juvenile’ or ‘primordial’ noble gases (e.g. ³He) and radiogenic noble gases generated by radioactive decay processes within the Earth (e.g. ⁴He). The radiogenic noble gases are primarily produced by radioactive decay and nucleogenic reactions (Table 1). Minor amounts are also produced by cosmic ray spallation, although this is not relevant to study of CO₂ storage and is not discussed further. The following summary documents these main production mechanisms and focuses on the reactions that are directly relevant to the use of noble gases as tracers in natural gas and CO₂ reservoirs.

Table 1 Radiogenic nuclides and their half-lives for noble gases, from Porcelli et al. (2002) and references therein

3.3 Noble Gas Production

3.3.1 Radiogenic Helium: ⁴He

⁴He is produced by any α—decay radioactive pathway which emits α particles. The three most significant radioactive parent isotopes are ²³⁸U, ²³⁵U and ²³²Th. Both U and Th are incompatible elements which are concentrated in the continental crust. The majority of crustal rocks contain U and Th in an almost constant weight ratio of 3.3, meaning that ⁴He production from all three nuclides can be referenced to the abundance of U alone. The ⁴He production rate expressed in this way is 2.15 × 10⁻⁷ cm³ per gram of U per year (Ozima and Podosek 2001).

3.3.2 Radiogenic Argon: ⁴⁰Ar

Radiogenic decay of another incompatible element, ⁴⁰K, results in the production of ⁴⁰Ar. ⁴⁰K is strongly partitioned into the crust and therefore ⁴⁰Ar production is highly significant despite the low isotopic abundance (0.00012) of ⁴⁰K and only 11 % decaying to ⁴⁰Ar by electron capture (with the remainder decaying by β decay to ⁴⁰Ca). Over geologic time the average ⁴⁰Ar production rate is 3.89 × 10⁻¹² cm³ per gram of K per year and this has resulted in ⁴⁰Ar dominating the Ar isotope budget with an isotopic abundance of 99.6 % in the atmosphere.

3.3.3 Krypton and Xenon

Small amounts of fissionogenic Xe and Kr are produced by the fission of ²⁴⁴Pu and ²³⁸U. The most significant product of these reactions is ¹³⁶Xe which is produced at a rate of 4.98 × 10⁻¹⁶ cm³g⁻¹(U) a⁻¹ and is accompanied by lesser amounts of ¹³¹Xe, ¹³²Xe and ¹³⁴Xe. Minor amounts of Kr isotopes are also produced by U fission but are not relevant for these discussions so have been omitted.

3.3.4 Nucleogenic Helium and Neon

Alpha particles and neutrons emitted during the decay series of U and Th can bombard lighter elements and cause nuclear reactions producing noble gases. The production of Ne isotopes in the crust is almost entirely due to nucleogenic routes (Table 2). Their rate of production is directly related to radioelement and target-element concentrations and the distribution of the target element related to radioelement heterogeneity (Ballentine and Burnard 2002). Additionally, in regions with high U and Th or with unusual compositions, such as high Li contents, nucleogenic ³He production can increase the ³He/⁴He ratios.

Table 2 Nucleogenic reactions applicable to this study

3.4 Atmospheric Noble Gases

With the exception of He, which is lost to space, the Earth’s noble gases are concentrated in the atmosphere. This is a result of their highly volatile nature, meaning that the atmosphere consists of primordial noble gases that have been outgassed from the Earth’s mantle early in its history (e.g. Porcelli and Ballentine 2002) and possibly from external volatile sources such as comets (e.g. Holland et al. 2009).

The noble gas composition of the atmosphere is well constrained (Porcelli et al. 2002; Sarda et al. 1985) and is the standard used as a reference for discussion. With the exception of Ar which is dominated by radiogenic ⁴⁰Ar and accounts for almost 1 % of the atmosphere, the noble gases are present as trace constituents. The atmosphere contains almost the entire inventory of ³⁶Ar. The ⁴⁰Ar/³⁶Ar ratio of the atmosphere was originally equal to the solar ratio (<10⁻³) but this has been increased to its present value of 295.5 by radiogenic production of ⁴⁰Ar (Sarda et al. 1985). The atmospheric ³He/⁴He value is commonly referred to as R_A (where R_A is the ³He/⁴He ratio of air at 1.4 × 10⁻⁶). The He inventory of the atmosphere is a combination of material outgassed from the mantle (which is rich in ³He), material produced in and outgassed from the crust (which is rich in ⁴He), cosmogenic material (rich in ³He) and material lost to space (³He is preferentially lost over ⁴He as it is lighter and more volatile). The other significant atmospheric noble gases are ²⁰Ne, ⁸⁴Kr, ¹²⁹Xe and ¹³²Xe which have characteristic ratios relative to Ar of ²⁰Ne/³⁶Ar = 0.5185, ⁸⁴Kr/³⁶Ar = 0.0207 and ¹³⁰Xe/³⁶Ar = 0.7237 × 10⁻³ (Ozima and Podosek 2001).

Noble gases from the atmosphere are dissolved in water with the proportions dependent on their respective solubilities and subsequently migrate into basin aquifers, transported by groundwater (Mazor 1972). The fractionation of atmospheric noble gases present in the subsurface natural gases then provides information on the interaction between oil phases, if any, present in the fluid system (Ballentine et al. 1996; Pinti and Marty 1995) and the interactions between water and gas phases, as well as the extent of degassing of the groundwater.

3.5 Crustal Noble Gases

The noble gas inventory of the crust is extremely variable as it is dependent on age, geological history (e.g. degree of metamorphism) and composition of radioactive elements. U, Th and K are enriched in the crust as a result of their incompatible behaviour, and therefore so are ⁴He and ⁴⁰Ar, the products of their radioactive decay. Crustal ³He/⁴He ratios are typically within the range of 0.01–0.05 Ra (Ozima and Podosek 2001), but can be significantly higher in Li-rich regions, due to nucleogenic production of ³He (Tolstikhin et al. 1996). The present day ⁴He/⁴⁰Ar production ratio of the lower, middle and upper crust can be calculated as 3.09, 5.79 and 6.00 (Ballentine and Burnard 2002), using the average crustal composition values of the three separate crust portions. This gives a production weighted average value of 5.7. ³⁶Ar production in the crust is a minor component compared to the amount of atmosphere-derived ³⁶Ar that is introduced dissolved in the groundwater (Fontes et al. 1991). Fissionogenic Kr and Xe can be significant in regions rich in U, with fissionogenic Xe being characterised by a ¹³⁶Xe/¹³²Xe ratio of 1.733 ± 0.003 and ¹³⁴Xe/¹³²Xe of 1.435 ± 0.003 (Drescher et al. 1998).

3.6 Mantle Noble Gases

The presence of excess ¹²⁹Xe in the mantle, which is produced by the decay of the extinct ¹²⁹I with a half-life of 15.7 Ma, implies that the mantle outgassed very early in Earth history (<100 Ma). I–Pu–U systematics requires significant loss of noble gases over this period (Porcelli et al. 2001). Loss of almost all of the primordial ³⁶Ar to the atmosphere and radiogenic production of ⁴⁰Ar from the decay of ⁴⁰K has resulted in a high upper mantle ⁴⁰Ar/³⁶Ar ratio, with a upper limit of >35,000 (Burnard et al. 1997; Holland and Ballentine 2006; Moreira et al. 1998). The ³He/⁴He ratios of the mantle are also high and exhibit a distinct difference between the upper and primitive lower mantle sources. The ³He/⁴He signature of oceanic island basalts (OIBs) sourcing the primitive lower mantle is highly variable, typically ranging between 9–32 R_A (Farley and Craig 1994; Trieloff et al. 2000). This is in contrast to the narrow range of ~8 ± 1 R_A exhibited by mid ocean ridge basalts (MORB) (Graham 2002), which sample the upper portion of the convective mantle.

4 Understanding the Sources and Processing of Crustal Fluids

4.1 Introduction

As described in the previous section, noble gases are predominantly derived from the atmosphere, the crust and the mantle. We have also seen that the isotopic signatures of these three reservoirs are distinct from one another. These characteristic differences in isotopic composition allow us to distinguish between components.

4.2 Two Component Mixing

The atmosphere has relatively low concentrations of radiogenic/nucleogenic noble gases ²¹Ne, ⁴⁰Ar and ¹³⁶Xe compared to crustal sources. To distinguish between the two, the isotopic ratios can be compared with the atmospheric ratio to identify crustal excess (Ballentine et al. 2002).

In a two component crust/air mixture:

$$ \left[ {^{21} Ne} \right]_{crust} = \left[ {^{21} Ne} \right]_{tot} \times \left[ {1 - \left( {^{21} Ne/^{20} Ne} \right)_{air} /\left( {^{21} Ne/^{20} Ne} \right)_{s} } \right] $$

(1)

$$ \left[ {^{40} Ar} \right]_{crust} = \left[ {^{40} Ar} \right]_{tot} \times \left[ {1 - \left( {^{40} Ar/^{36} Ar} \right)_{air} /\left( {^{40} Ar/^{36} Ar} \right)_{s} } \right] $$

(2)

$$ \left[ {^{136} Xe} \right]_{crust} = \left[ {^{136} Xe} \right]_{tot} \times \left[ {1 - \left( {^{136} Xe/^{130} Xe} \right)_{air} /\left( {^{136} Xe/^{130} Xe} \right)_{s} } \right] $$

(3)

Where subscripts crust and total refer to concentrations and subscripts air and s refer to the isotopic compositions of the atmosphere and sample, respectively.

⁴He is also produced in significant quantities in the crust but the He system is complicated by the presence of ‘primitive’ ³He, sourced from the mantle. In order to resolve the mantle/crustal components, ³He can be used. Although He has a low abundance in the atmosphere due to thermal loss, excess air corrections must still be made by using the observed air-derived ²⁰Ne concentrations (Ballentine et al. 2002).

$$ \left( {\frac{{{}^{3}He}}{{{}^{4}He}}} \right)_{c} = \frac{{\left( {\left( {\frac{{{}^{3}He}}{{{}^{4}He}}} \right)_{s} \times \left( {\frac{{{}^{4}He}}{{{}^{20}Ne}}} \right)_{s} } \right) \div \left( {\left( {\frac{{{}^{4}He}}{{{}^{20}Ne}}} \right)_{air} \times \left( {\frac{{{}^{3}He}}{{{}^{4}He}}} \right)_{air} } \right)}}{{\left( {\left( {\frac{{{}^{4}He}}{{{}^{20}Ne}}} \right)_{s} - \left( {\frac{{{}^{4}He}}{{{}^{20}Ne}}} \right)_{air} } \right) - 1}} $$

(4)

Subscripts c, s and air refer to the corrected, measured and air-derived ratios, respectively. Once corrected for excess air, the ³He/⁴He ratio represents the sum of the crust and the mantle, with crustal ⁴He given by:

$$ \left[ {{}^{4}He} \right]_{c} = \frac{{\left[ {{}^{4}He} \right]_{tot} \times \left( {\left( {\frac{{{}^{3}He}}{{{}^{4}He}}} \right)_{mantle} - \left( {\frac{{{}^{3}He}}{{{}^{4}He}}} \right)_{c} } \right)}}{{\left( {\left( {\frac{{{}^{3}He}}{{{}^{4}He}}} \right)_{mantle} - \left( {\frac{{{}^{3}He}}{{{}^{4}He}}} \right)_{crust} } \right)}} $$

(5)

Where subscripts mantle, crust and c refer to the mantle, crust and air corrected values.

4.3 Three Component Mixing

In many crustal fluid reservoirs, a significant magmatic component is present in the fluids, in addition to the air and radiogenic components. Typically, all three components, including that of the atmosphere can be observed in the noble gases. In principle, it is possible to correct out the air component using an unambiguously air derived isotope. In practice, this is problematic because this correction depends on no (or at least well quantified) elemental fractionation in an individual sample which is often not the case. However, due to the large differences in isotopic composition of the mantle, crust and atmosphere end-members of Ne, this isotopic system alone can be used to make the correction for the atmospheric component (Fig. 5). This is because any sample that contains a mix of these components must plot within the envelope defined by the well-known ²⁰Ne/²²Ne and ²¹Ne/²²Ne air and crust end members and the resolved mantle end member values of 12.5 for ²⁰Ne/²²Ne and 0.06 for ²¹Ne/²²Ne (Ballentine et al. 2005). Hence, the Ne isotope system is a particularly useful tracer within natural gas reservoirs as the contribution from the mantle, crust and atmosphere to any sample can be resolved. In the following sections we shall see that Ne isotopic composition provides the most striking evidence of the different proportions of atmosphere, crust and mantle inputs to subsurface gas systems.

Fig. 5

Plot of ²⁰Ne/²²Ne against ²¹Ne/²²Ne illustrating the three distinct sources of Ne in natural gas reservoirs (redrawn from Ballentine 1997). As the three end member ratios are known it is simple to correct any sample for atmosphere, crust or mantle contributions, taking due account of the error propagation. Example shows correction for the atmospheric contribution. The same procedure can be used for correction of crust and mantle contributions

4.4 Role of Stable Isotopes

In many natural reservoirs, the dominant CO₂ source and basin scale processes that act on it are poorly constrained. This is primarily a reflection of the multiple origins of CO₂ in the subsurface which include methanogenesis, oil field biodegradation, kerogen decarboxyilation, hydrocarbon oxidation, decarbonation of marine carbonates and degassing of magmatic bodies (Jenden et al. 1993; Wycherley et al. 1999). Carbon isotopes can be measured within natural CO₂ gas and precision is usually better than 0.1 % (Javoy et al. 1986), the δ¹³C(CO₂) signature having been used previously to distinguish between different CO₂ sources. However, δ¹³CO₂ values alone do not provide an unambiguous means to distinguish between mantle-derived CO₂ and CO₂ derived from crustal sources such as the metamorphism of carbonate rocks. This is due to the average mantle carbon and bulk crustal carbon exhibiting an overlapping range of δ¹³CO₂ of −3 to −8 % (Fig. 6) (Sherwood Lollar et al. 1997). These are the only sources that can produce gas fields containing greater than 99 % CO₂ by volume, which are particularly pertinent as natural CO₂ sequestration analogues.

Fig. 6

Plot of CO₂ concentration against the range of δ¹³C (%) observed in natural systems. The range of δ¹³C values corresponding to a variety of natural processes are also labelled as is the clear overlap in the range of values that exists between magmatic CO₂ and the alteration of marine carbonates. Significantly these are the only sources that can account for reservoirs with high concentrations of CO₂

A much more unambiguous measure of CO₂ origin is provided by comparing the absolute amount of the magmatic noble gas, ³He, to the CO₂ concentration within the gas reservoir, because the mantle CO₂/³He ratio measured in mid-ocean ridge basalts (MORB) has a small range, from 1.0 × 10⁹ to 1.0 × 10¹⁰ (Burnard et al. 2002; Marty and Jambon 1987; Trull et al. 1993), compared with the range observed in natural gases. Figure 7 shows CO₂/³He plotted against CO₂ concentration. Any natural CO₂ sample that plots above this MORB range irrespective of CO₂ content can only be attributed to a CO₂ source containing minimal ³He, providing an unambiguous identification of a crustal CO₂ source. Samples within or below the MORB range contain a magmatic component but have been subjected to possible CO₂ loss, dilution (by addition of other gas species such as CH₄ or N₂) and/or crustal CO₂ addition. Careful study of both the δ¹³C (CO₂) isotopes and CO₂/³He ratios can distinguish between these processes and allow CO₂ origins to be determined (Ballentine et al. 2001, 2002; Sherwood Lollar et al. 1994, 1997).

Fig. 7

Plot of CO₂ concentration against CO₂/³He ratio. The range of CO₂/³He values exhibited by mid-ocean ridge (MORB) popping rocks is highlighted by the shaded region. Any CO₂ that plot within or below this range, must contain a significant quantity of magmatic ³He, implying a mantle origin for the CO₂. As the crust contains no ³He, gases sourced from the crust have CO₂/³He values that plot above the shaded region (Ballentine et al. 2001, 2002; Sherwood Lollar et al. 1999; Zhou et al. 2005)

5 Use of Noble Gases in Natural CO₂ Reservoirs: Colorado Plateau, USA

The study of noble gases in the understanding of water—CO₂—oil interactions is still in the early stages but in principle, noble gases and the combination of noble gases and stable isotope data can provide a quantitative tool for tracing the processes occurring in the hydrogeological system that have direct relevance for long term carbon sequestration.

5.1 Determining the Source and Residence Time of CO₂

Noble gases have been particularly effective in determining the source of CO₂ within natural CO₂ reservoirs. Comparing the concentration of ³He with the concentration of ⁴He (i.e. the ³He/⁴He ratio) can identify magmatic contributions to subsurface reservoirs (Ballentine et al. 2002). In a similar fashion, the concentration of ³He can be compared to the concentration of CO₂ within a natural CO₂ reservoir to determine the origin of the CO₂ (Marty et al. 1992; Oxburgh et al. 1986). Magmatic CO₂/³He ratios exhibit a small range when compared to other crustal fluids (1 × 10⁹ to 1 × 10¹⁰) (Ballentine et al. 2001) and hence can be used to identify and quantify the amount of mantle CO₂ in a particular reservoir.

Using this technique it has been established that five reservoirs (McCallum Dome, Sheep Mountain, McElmo Dome, Bravo Dome, St Johns Dome) surrounding the Colorado Plateau uplift region of the southwest U.S. primarily contain mantle-derived CO₂ (Fig. 8). Radiometric dating of surface volcanics surrounding these natural CO₂ reservoirs gives an indication of the approximate timing of CO₂ injection. The most recent magmatic event related to the natural CO₂ reservoirs investigated occurred at the Raton-Clayton volcanic field close to Bravo Dome between 10 and 8 ka (Broadhead 1998). The next youngest event is associated with St. Johns Dome, where the last phase of volcanic activity from the nearby Springerville Volcanic field dates from 0.3 to 2.1 Ma (Rauzi 1999). Magmatic activity associated with both the Sheep Mountain and McCallum Dome CO₂ reservoirs occurred in the Late Tertiary (Maughan 1988; Woodward 1983). The oldest intrusive igneous rocks close to a natural CO₂ reservoir are those associated with McElmo Dome, from the nearby Ute Mountain and La Plata Mountain laccoliths which have been dated at 40–72 Ma (Stevens et al. 2006). Whilst these ages were determined on surface intrusive igneous rocks, and it is probable that deep igneous activity will have continued for a significant period after these surface rocks were formed, it is clear that several of the natural reservoirs appear to have stored CO₂ for at least a million years. Indeed in two reservoirs CO₂ has been stored for considerably longer, namely McElmo Dome where CO₂ was injected ~40 million years ago (Gilfillan et al. 2008) and in the JM Brown Basset Field which is believed to have contained CO₂ for some 300 million years (Ballentine et al. 2001). In all these fields, CO₂/³He data show a magmatic origin for the CO₂ gases with evidence of CO₂ depletion relative to ³He (Fig. 8).

Fig. 8

Plot of CO₂/³He against CO₂ concentration for all of the reservoirs studied. All error bars are smaller than printed symbols. The shaded region highlights the range of CO₂/³He values measured in pure magmatic samples (MORB). All of the samples from the deep gas fields plot within or below this range, indicating the presence of a significant quantity of magmatic ³He, implying a mantle origin for the CO₂ within these fields (Gilfillan et al. 2008). Doe Canyon is a small natural CO₂ reservoir related to McElmo Dome

5.2 Determining the Long Term Fate of CO₂

In addition to identifying the source of CO₂, noble gases can also be used to determine the long term fate of natural CO₂. Measured CO₂/³He ratios from nine CO₂ reservoirs from around the world (specifically the U.S., Europe and China) show a marked decrease in CO₂/³He ratios with increasing ⁴He, (Fig. 9a) and where measured, ²⁰Ne (Gilfillan et al. 2009) (Fig. 9b). As ³He is almost solely derived from the mantle, there is no source in the shallow crust that could increase its concentration after the premixed CO₂ and ³He has entered the reservoir. Hence, the reduction in CO₂/³He can only be explained if the CO₂ component is being reduced relative to ³He.

Fig. 9

a CO₂-rich gas fields from around the world also show a strong correlation between decreasing CO₂/³He and increasing ⁴He concentrations. ⁴He accumulates in the formation water over time and further indicates the importance of formation water in controlling the mechanism of subsurface CO₂ removal (Gilfillan et al. 2009). b CO₂/³He variation plotted against ²⁰Ne in samples from natural CO₂ reservoirs from around the world. There is a general trend in this data set of decreasing CO₂/³He with increasing ²⁰Ne. ³He is conservative within the gas phase. Variations in CO₂/³He therefore represent subsurface removal of CO₂ from the emplaced CO₂ phase. As the only subsurface source of the ²⁰Ne which is now in the CO₂ phase is the formation water, the sink that is reducing CO₂ must be linked to the formation water which is in contact with the CO₂ (Gilfillan et al. 2009)

Due to its atmospheric origin, ²⁰Ne is a tracer of formation water interaction in crustal fluids. ⁴He is a tracer of radiogenic input to a crustal fluid system, and accumulates in the formation water over time (hence its use for formation water dating). Therefore, the correlation between the CO₂/³He reduction and increasing ²⁰Ne and ⁴He concentrations illustrates that contact with the formation water is directly controlling the removal of the CO₂ component in the natural CO₂ reservoirs.

5.2.1 Using Noble Gases and Carbon Isotopes to Identify Long Term Fate of CO₂

In each individual field, the reduction in CO₂/³He ratios allows the amount of CO₂ lost relative to ³He to be calculated. Several of the reservoirs show an excess of 90 % loss of CO₂ compared to the highest CO₂/³He ratio measured (Gilfillan et al. 2009). Using stable carbon isotopes in conjunction with noble gas data it is possible to determine how much of this CO₂ component is being lost by dissolution into the formation water or by precipitation of new carbonate minerals. This is because precipitation of new carbonate minerals is a chemical process which strongly alters the δ¹³C (CO₂) isotope values as a result of the preferential incorporation of ¹³C in the precipitate. As both ¹²C and ¹³C have similar solubility, dissolution of CO₂ into the formation water has a lesser effect on the δ¹³C (CO₂) isotope values, although this is also pH dependent.

All of the natural CO₂ reservoirs investigated exhibited a small range of carbon isotope values, despite large reductions in the CO₂ component relative to ³He. This shows that the majority of the CO₂ is being removed by dissolution into the formation water. Individual reservoir lithologies exhibited a clear control on the variation of δ¹³C (CO₂) isotopes both within and between the individual natural reservoirs. Within the carbonate dominated reservoirs very little variation in δ¹³C (CO₂) values is observed compared to the much wider range of values which were exhibited in the silicilastic dominated reservoirs (Fig. 10).

Fig. 10

Plot of CO₂/³He against δ¹³C (CO₂) for all of the reservoirs in the study. The carbonate dominated reservoirs exhibit very little variation in δ¹³C (CO₂) value, both within the individual reservoirs and between different reservoirs (left). The reservoirs which had a predominantly siliciclastic lithologies reservoirs exhibited a wider range of δ¹³C (CO₂) values within and between the reservoirs (right) (Gilfillan et al. 2009)

Evidence to suggest that precipitation of new carbonate minerals is responsible for some reduction of the CO₂ component is found only in sandstone-dominated reservoirs. This was particularly the case in the Bravo Dome CO₂ reservoir, where the measured values for one sample indicated that up to 18 % of the CO₂ component could have been precipitated (Fig. 11). The dataset from Sheep Mountain indicates that a maximum of 10 % of the CO₂ could have also have been removed by precipitation (Fig. 12a). However, the remaining four siliciclastic and all three of the carbonate dominated reservoirs, such as McElmo Dome (Fig. 12b) showed very little evidence of precipitation and it is clear that CO₂ dissolution into the formation water is the only significant CO₂ sink over geological timescales (Gilfillan et al. 2009).

Fig. 11

a Model of evolution of gas phase CO₂/³He ratio and δ¹³C of the CO₂ under Bravo Dome reservoir conditions. Using a combination of noble gas and δ¹³C(CO₂) stable isotopes determination and quantification of the different mechanisms of CO₂ removal from natural CO₂ reservoirs is possible. This is because δ¹³C(CO₂), the stable carbon isotope value of the residual CO₂ phase, is changed, or ‘fractionated’, by different amounts depending on the loss mechanism. Specifically, precipitation of CO₂ as carbonate minerals has a distinct fractionation from CO₂ dissolving to into the formation water in the reservoir. b Plot of δ¹³C(CO₂) against CO₂/³He for the siliciclastic Bravo Dome field as CO₂ gas is either dissolved under different pH (dashed lines) or precipitated in an open system by Rayleigh fractionation (solid line). The Bravo Dome samples exhibit up to 18 % CO₂ loss by carbonate precipitation and up to 60 % dissolved into the formation water (Gilfillan et al. 2009)

Fig. 12

a Plot of δ¹³C (CO₂) against CO₂/³He for the siliciclastic Sheep Mountain field. Whilst the majority of the dataset is consistent with dissolution of CO₂(g) at pH = 5, a number of data points require approximately 10 % of the CO₂(g) precipitating as carbonate, together with dissolution of CO₂(g) into the formation water at a pH of = 5 in a two process model similar to that outlined for Bravo Dome. b Plot of δ¹³C(CO₂) against CO₂/³He for the carbonate McElmo Dome field. This shows the relationship between CO₂/³He and δ¹³C (CO₂) observed in the majority of the CO₂ reservoirs. This relationship indicates that the dominant CO₂ sink in these natural systems is dissolution into the formation water. Error bars represent analytical uncertainty on the measured data points

5.2.2 Using Noble Gases and Carbon Isotopes to Determine Reservoir pH Conditions

Combining the CO₂/³He ratio reduction with the corresponding fractionation of the δ¹³C (CO₂) isotopes also permits determination of the reservoir pH conditions under which the CO₂ dissolution into the formation water is occurring (Fig. 11). This is the result of the different carbon stable isotope fractionation factors which act on CO₂ gas dissolving either to bicarbonate (HCO₃⁻) or carbonic acid (H₂CO₃) (Deines et al. 1974). The proportion of CO₂ dissolving to each species is directly controlled by pH, meaning that the δ¹³C (CO₂) fractionation trend can be used to determine the pH conditions under which dissolution is occurring (Fig. 11). This is a significant breakthrough in the understanding of natural CO₂ reservoir conditions, as subsurface pH measurements are notoriously difficult to obtain. As with δ¹³C (CO₂) values, the reservoir lithology exhibits a clear control on the pH range observed in the different CO₂ reservoirs. A pH range of 5.0–5.3 was observed in the six siliciclastic reservoirs, compared to the higher values of 5.4 to 5.8 determined for the three carbonate reservoirs. This can be accounted for by the buffering capacity provided by additional bicarbonate ions which originate from the dissolution of carbonate minerals in the carbonate-dominated reservoirs. The pH ranges recorded in the natural reservoirs are also consistent with the subsurface pH value of 5.7 observed in the recent CO₂ injection experiment in the Frio formation in Texas (Kharaka et al. 2006).

5.3 Physical Chemistry of Dissolved Noble Gases

The different noble gas isotope systems can tell us about the different proportions of air or crustal derived noble gases within a reservoir but they cannot shed light on the physical processes that occur within these reservoirs. This is because processes such as degassing and phase interactions do not alter the isotopic compositions unless the timescales over which they occur are short than the noble gas diffusion rates. Therefore one must identify the relative solubilities of the noble gases to determine their complex histories within a reservoir, acknowledging that each reservoir will have unique physical conditions and therefore unique ‘starting’ compositions upon which physical processes leave their fractionating fingerprints. To begin with we must quantify noble gas solubilities in different phases under reservoir conditions to establish these initial noble gas compositions within a reservoir.

Fundamentally, the solubility of noble gases in solution is governed by Henry’s Law:

$$ p_{i}\; = \;K_{i}x_{i} $$

(6)

where p_i is the partial pressure of gas i in equilibrium with a fluid containing x_i mole fraction of i in solution and K_i is Henry’s constant.

Henry’s coefficient varies with pressure, temperature and salinity of the fluid. However at the temperatures and pressures found in crustal fluid systems, empirical data are not available. Consequently the equation must be modified to include non-ideality in both the gas and liquid phase. The modified equation then becomes:

$$ \phi_{\rm i} {\text{p}}_{\rm i} = \, \gamma_{\rm i} {\text{K}}_{\rm i}x_{\rm i} $$

(7)

where Φ is the gas phase fugacity coefficient and γ is the liquid phase activity coefficient.

In the gas phase, the real molar volume can be calculated from:

$$ {\text{PVm}}/{\text{RT}} = 1 + {\text{B}}\left( {\text{T}} \right)/{\text{V}}_{{\text{m}}} + {\text{C}}\left( {\text{T}} \right)/{\text{V}}_{{{\text{m}}^{2} }} $$

(8)

From which the fugacity coefficient can be derived via:

$$ \Upphi\left( {{\text{P}},{\text{T}}} \right) = \exp [{\text{B}}\left( {\text{T}} \right)/{\text{V}}_{{\text{m}}} + ({\text{C}}\left( {\text{T}} \right) + {\text{B}}\left( {\text{T}} \right)^{2} /2{\text{V}}_{{{\text{m}}^{{\text{2}}} }} $$

(9)

Where P is the total pressure, R is the gas constant, T is the temperature and B(T) and C(T) the temperature dependant first and second order virial coefficients.

Non ideality in the liquid phase cannot be determined using equations of state because of the complexity of solute/solute and solvent/solute interactions. γ is considered to be independent of pressure therefore deviation from ideality caused by temperature and electrolyte concentration is assessed from empirically derived data. For a more extensive review of physical chemistry in liquid and gas phase see Ballentine and Burnard (2002).

5.4 Noble Gas Fractionation and Partitioning Between Phases

Of more direct relevance to assessment of subsurface CO₂ storage using noble gases, is the modelling of solubility fractionation between subsurface phases in open and closed systems (Ballentine et al. 1996; Hiyagon and Kennedy 1992; Pinti and Marty 1998; Torgersen and Kennedy 1999; Zartman et al. 1961). Noble gases alone can also be used to identify interactions between different fluid phases. In the simplest scenario, equilibration of a large gas/water volume ratio in the subsurface would be expected to cause near quantitative degassing of the formation water and result in a gas phase that inherits the formation water noble gas isotopic and elemental composition. This composition would be expected to be similar to the calculated air saturated water (ASW) noble gas composition, determined using appropriate recharge conditions (for the Bravo Dome field used as the example here an ASW ²⁰Ne/³⁶Ar of 0.152 was determined assuming air equilibration at 10 °C and a small excess air component of 10 % Ne (Kipfer et al. 2002). As the subsurface gas/water volume ratio decreases the least soluble gases will tend to be enriched in the gas phase until V_gas/V_water → 0 when the limit of the enrichment is controlled by the relative solubility of the respective gases. For example, the maximum enrichment of the Ne relative to Ar in the gas phase though phase equilibrium degassing is K_Ne/K_Ar where K_Ne and K_Ar are the respective Henry’s constants for Ne and Ar (e.g. Ballentine et al. 1991, 2002). K_Ne and K_Ar for average field conditions predict a maximum enrichment of Ne relative to Ar from ASW values by a factor of ~2 corresponding to a maximum ²⁰Ne/³⁶Ar of 0.304, which is significantly lower than the range of ²⁰Ne/³⁶Ar values (0.314–0.617) observed in the Bravo Dome dataset (Fig. 13). The Sheep Mountain data exhibits even more extreme ²⁰Ne/³⁶Ar values which range from 0.530 to 2.496. Clearly both datasets cannot be explained by simple equilibration between the CO₂ and formation water phases and must be explained by a more complex equilibration history.

Fig. 13

Plot of ²⁰Ne/³⁶Ar_(atm) against ⁸⁴Kr/³⁶Ar_(atm) for the Bravo Dome field. Also plotted are the Air and ASW values and the calculated Rayleigh and Batch fractionation curves for a gas phase exsolving from a groundwater that has an initial ASW composition (Eq. 6) (dashed lines). The tick marks on the dashed lines represent the proportion of Ar remaining in the groundwater phase. Also plotted is the Rayleigh fractionation line (Eq. 15) which would be produced by a gas phase dissolving into a water phase initially containing no noble gases (solid line). This is the re-dissolution model discussed in the text (Eq. 17). The bold tick marks on the solid fractionation line represent the proportion of Ar lost from the gas phase. The fractionation pattern observed in the air derived ratios can be explained by the re-dissolution model (Gilfillan et al. 2008)

5.4.1 Solubility Controlled Phase Fractionation

Henry’s Law (Eq. 1) defines the fractionation between the gas phase and the liquid phase with which it is in equilibrium. Again assuming ideal gas behaviour, this can be rewritten to relate the concentration in the gas phase, ${\text{C}_{\text{g}}}^{\text{i}} ,$ to the concentration in the fluid phase, ${\text{C}_{\text{l}}}^{\text{i}} ,$ via Henry’s constant ${\text{K}_{\text{l}}}^{\text{d}}$:

$$ {\text{C}_{\text{g}}^{\text{i}}} = {\text{ K}}_{\text{i}}^{\text{d}} {\text{C}}_{\text{l}}^{\text{i}} $$

(10)

The equilibrium concentrations are also dependent on the gas/liquid volume ratio (V_l/V_g), related by

$$ \left[ {\text{i}} \right]_{\text{g}} = \, \left[ {\text{i}} \right]_{\text{T}} \left( {{\text{V}}_{\text{l}} /{\text{V}}_{\text{g}} {\text{K}_{\text{i}}^{\text{d}}} + 1} \right)^{ - 1} $$

(11)

Where [i]_g is the number of moles in the gas phase and [i]_T is the total number of moles. As V_l/V_g → 0, [i]_g → [i]_T. Giving Henry’s constant units of molality leads to:

$$ A_{g} = A_{r} \left[ {\frac{{22400T\rho_{1} V_{1} }}{{1,000\, \times \,273\frac{{\gamma_{A} }}{{\phi_{A} }}K_{i}^{m} V_{g} }} + 1} \right]^{ - 1} $$

(12)

where:

A:

number of moles

ρl:

the density of the liquid l (g/cm³)

T:

Temperature

Vl, g:

Volume of liquid l, or gas g in cm³

Ki^m:

Henry’s constant (Kg atm/mol)

This determines the partitioning between the gas and liquid phase and makes it possible to calculate the volume of gas with which a liquid has equilibrated by using only one noble gas concentration in the liquid phase. This method assumes that during each stage of degassing, V_gas/V_water is small and by iteratively equilibrating small volumes of gas with progressively degassed groundwater, one can calculate the number of single stage equilibration episodes that must have occurred to produce an observed loss of the noble gas in question. Summing the volumes of gas defines the total of volume of gas exsolved from the groundwater, hence the gas/groundwater volume ratio (e.g. Zhou et al. 2005).

In practice it is often difficult to assess the initial concentrations of noble gases observed in the liquid phase. In this case, the relative fractionation between two different noble gases can be employed to determine the extent of fractionation processing.

$$ \left( \frac{A}{B} \right)_{gas} = \left( \frac{A}{B} \right)_{groundwater} \frac{{\left( {\frac{Vg}{Vl} + \frac{1}{{K_{B}^{{}} }}} \right)}}{{\left( {\frac{Vg}{Vl} + \frac{1}{{K_{A}^{{}} }}} \right)}} $$

(13)

as

$$ \text as\,\frac{Vg}{Vl} \to 0,\,\frac{{\left( \frac{A}{B} \right)_{gas} }}{{\left( \frac{A}{B} \right)_{groundwater} }} \to \frac{{\frac{rA}{\phi A}K_{A}^{{}} }}{{\frac{rA}{\phi B}K_{B}^{{}} }} = \alpha $$

(14)

where:

$\left( \frac{A}{B} \right)_{groundwater}$:

$\frac{A}{B}$ original groundwater ratio and A & B are different noble gases

$\left( \frac{A}{B} \right)_{gas}$:

$\frac{A}{B}$ ratio in the gas phase

α:

Fractionation coefficient calculated for the gas/groundwater system

K_A, K_B:

Henry’s constants for gases A and B

r_A, r_B:

Groundwater phase activity coefficients

$\phi_{A} \,\phi_{B}$:

Gas phase fugacity coefficients

5.4.2 Rayleigh Fractionation (Open System)

Noble gas partitioning from a water phase into a gas phase (degassing) can also follow the Rayleigh fractionation law, provided that the gas phase is continually removed. Rayleigh fractionation can be expressed as:

$$ \left( \frac{A}{B} \right)_{groundwater} = \left( \frac{A}{B} \right)_{0} f^{\alpha - 1} $$

(15)

$$ \alpha = \frac{{\frac{rA}{\phi A}K_{A}^{groundwater} }}{{\frac{rA}{\phi B}K_{B}^{groundwater} }} $$

(16)

Where:

$\left( \frac{A}{B} \right)_{groundwater}$:

$\frac{A}{B}$ ratio in the groundwater and A and B are different noble gases

$\left( \frac{A}{B} \right)_{0}$:

The original groundwater phase $\frac{A}{B}$ ratio

f:

Fraction of B remaining in the groundwater phase

Equally, one may also consider the opposite situation, in which noble gases in the gas phase are re-dissolved into groundwater that has been previously stripped of noble gases, as might occur during contact with a CO₂ phase. In this case, the ratio A/B will evolve in the opposite direction, given by:

$$ \left( {\frac{A}{B}} \right)_{{groundwater}} = \left( {\frac{A}{B}} \right)_{0} \,f^{{\left( {\frac{1}{\alpha } - 1} \right)}} $$

(17)

In principle, multiple stages of exsolution and dissolution could increase the magnitude of fractionation in a gas phase over that predicted by the phase equilibrium model limit (Zartman et al. 1961). In this case a gas phase exsolved from the water with a low gas/water volume ratio will have an elevated ²⁰Ne/³⁶Ar ratio. This could be re-dissolved into a smaller volume of water by a change in physical conditions such as a pressure increase, a decrease in temperature and salinity, or mixing with unsaturated (with respect to the gas phase) water. This would create a local increase in both the formation water noble gas concentration and the magnitude of noble gas fractionation in solution. Subsequent formation of a gas phase in equilibrium with the now modified formation water could result in a fractionation significantly in excess of that predicted by the single stage equilibrium model. For this process to occur the pressure/temperature conditions must fluctuate, and the resulting fractionated gas phase volume must be very small compared to the fluid/water volume required to produce significant fractionation (Ballentine et al. 2002). This is unlikely to have occurred on a gas field scale so it can be discounted as a viable mechanism to account for the extreme values observed in the Bravo Dome and Sheep Mountain datasets.

A second possible mechanism considered is non-equilibrium diffusion. Torgersen et al. (2004) have proposed that variable enrichments of noble gases can be accounted for by incomplete emptying followed by partial filling of lithic grains and half spaces, similar to the mechanism proposed by Pinti et al. (1999) for air diffusion into fresh pumice. They further demonstrate that both elemental and isotopic compositions of Ne and Ar show clear mass-dependent fractionation (Pinti et al. 1999). Whilst the significant mantle and crustal Ne contributions to the Colorado Plateau reservoirs makes subtle isotopic fractionation in Ne hard to identify, no significant mass-dependent fractionation in either the ³⁸Ar/³⁶Ar ratios or elemental abundance patterns was observed meaning that non-equilibrium diffusion could be ruled out (Holland and Ballentine 2006).

Another mechanism which could account for the extreme ²⁰Ne/³⁶Ar fractionation is interaction of the gas with an oil phase. This is because Ar (and also Kr and Xe), relative to Ne, is more soluble in oil than water meaning that equilibration of the groundwater with an oil phase could increase the ²⁰Ne/³⁶Ar of the water phase (Bosch and Mazor 1988). Subsequent degassing of the groundwater into the gas phase could then result in higher ²⁰Ne/³⁶Ar ratios in the gas. This has been suggested as an explanation for elevated ²⁰Ne/³⁶Ar ratios measured in the Indus Basin, Pakistan (Battani et al. 2000) and Mexico gas fields (Sarda et al. 2000), where the term double-distillation is used. However, there is no evidence of significant hydrocarbon components in the Bravo Dome CO₂ field and hence, oil interaction can be discounted as a valid explanation of the high ²⁰Ne/³⁶Ar ratios. Additionally, in the Indus study there is a correlation between the ²⁰Ne/³⁶Ar and 1/³⁶Ar ratios but no such correlation exists in the Bravo Dome or Sheep Mountain datasets.

Having ruled out the above mechanisms, a more appropriate model must be devised to explain the high ²⁰Ne/³⁶Ar ratios observed in the Colorado Plateau CO₂ reservoirs. One such model is a two stage partial re-dissolution model (Gilfillan et al. 2008). The preservation in some samples near ASW ratios suggests that some degree of ‘gas stripping’ of the formation water noble gas composition has occurred with complete removal of noble gases into the gas phase. Stage one involves such a complete degassing to provide an ASW starting composition for stage II, which consists of Rayleigh fractionation (open system) of the noble gases from the gas phase into the noble gas stripped formation water (potentially as a result of stage 1).

Neither closed system batch fractionation nor open system Rayleigh degassing models of gas exsolving from an ASW groundwater phase can account for the observed maximum fractionation of ²⁰Ne/³⁶Ar_(atm) and ⁸⁴Kr/³⁶Ar_(atm) in the Bravo Dome reservoir (Fig. 13). The values can however, be explained by the re-dissolution model. It is proposed that gas stripping of the reservoir formation water occurs during the injection and movement of magmatic CO₂ through the formation water during the initial reservoir filling stage. This is corroborated by the orders of magnitude lower ²⁰Ne concentration in the gas phase than that predicted by groundwater degassing (Fig. 14). The CO₂ migrates as a separate phase and as it is an excellent solvent, would have had the capacity to quantitatively partition any gases from the water phase into the CO₂, effectively ‘stripping’ the formation water of its dissolved gas content. This will produce a CO₂ reservoir with an ASW-like composition which will also have incorporated any radiogenic-crustal noble gases that had previously accumulated in the formation water. At this stage the formation water will be saturated with respect to CO₂, but completely depleted with respect to the atmosphere-derived noble gases.

Fig. 14

Plot of ²⁰Ne/³⁶Ar_(atm) against ²⁰Ne_(atm) for the samples from the Bravo Dome field. Also plotted are the predicted ²⁰Ne concentration values assuming Rayleigh Fractionation of a gas phase exsolving from the groundwater. The predicted concentrations are several orders of magnitude greater than the observed concentrations, implying that the atmospheric noble gases within the reservoir have been significantly diluted, most probably by injection of magmatic CO₂

In the Bravo Dome system, coherent fractionation of the crustal radiogenic and formation water-derived noble gases indicates that the formation water and CO₂ were well mixed prior to the fractionation occurring, hence the fractionation process must have occurred after the ASW gas stripping. Subsequent charges of magmatic CO₂ would then pass through the gas stripped formation water but will only dilute the ASW-derived and crustal-radiogenic noble gases in the CO₂ trapping structure. Bravo Dome preserves a clear ²⁰Ne concentration gradient, increasing from west to east in the field. It is argued this both corroborates the model and indicates the reservoir filling direction, which may still be ongoing, to be from the west (Baines and Worden 2004b).

Following reservoir filling the noble gases now in the CO₂ phase will then re-equilibrate with the degassed formation water. This re-equilibration will occur under either open system or closed system conditions depending on the available volume of water for re-equilibration and if there is significant water migration or recharge. For open system conditions, the re-dissolution process will follow a Rayleigh fractionation path. Under closed system conditions, the trend will be the exact reverse of that calculated for batch fractionation of a gas phase exsolving from water. Under the Bravo Dome reservoir conditions, the open system Rayleigh fractionation re-dissolution model accounts for all of the mantle corrected ²⁰Ne/³⁶Ar_(atm) and ⁸⁴Kr/³⁶Ar_(atm) ratios that are observed in the field (Fig. 13). The conceptual model is also presented as a cartoon (Fig. 15) from Gilfillan et al. (2008).

Fig. 15

Cartoon illustrating the two stage degassed formation water and re-dissolution model. Stage 1: Magmatic CO₂ injection into the formation water causes the water to become fully saturated in CO_2, with excess gas forming a free CO₂ phase. The free CO₂ phase migrates upwards, stripping dissolved air-derived noble gases and accumulated crustal/radiogenic noble gases as it does so. This results in formation water which is fully saturated in CO₂ and completely degassed with respect to ASW and crustal derived noble gases (and N₂). When this CO₂ containing the noble gases stripped from the formation water encounters a trapping structure, it provides the first reservoir fluid charge. Subsequent charges of CO₂ provide no more ASW and crustal noble gases, and serve only to dilute the original ASW and crustal noble gas rich CO₂. Reservoir scale preservation of concentration gradients in ASW-derived noble gases therefore provide an indication of CO₂ filling direction (cf Bravo Dome, St Johns). Stage 2: When recharge of the CO₂ cap stops, the noble gases begin to re-dissolve into the gas stripped groundwater. This results in a fractionation of the ASW-derived noble gases. The magnitude of this fractionation depends on the volume of water in contact with the gas phase and whether open or closed system fractionation conditions are present. Hence, only limited CO₂ re-dissolution can, and only as result of a change in reservoir conditions or CO₂ precipitation as new carbonate minerals causing a reduction of the CO₂ saturation level of the formation water. Figure reproduced from Gilfillan et al. (2008)

5.4.3 Quantifying the Extent of Water Interaction

Using the open system re-dissolution model outlined above, the fractionation of the measured gas phase ²⁰Ne/³⁶Ar ratios from the initial ASW formation water value can be used to define the proportion of ³⁶Ar lost from the gas phase into solution. Using this value it is then straightforward to calculate the initial concentration of ³⁶Ar, ³⁶Ar_init, within the gas phase prior to the re-dissolution process (Zhou et al. 2005). Using the concentration of ³⁶Ar in ASW it is then possible to calculate, for each sample, the volume of water required to have been gas ‘stripped’ to supply the ³⁶Ar_init to provide the Stage 1 Water/Gas volume ratio. For Bravo Dome system, Stage 1 Water/Gas volume ratio ranges from 0.0005 to 0.071 cm³ of water per cm³ gas (STP) (Table 3). Under reservoir conditions 1 cm³ of groundwater can hold 4.66 cm³ of CO₂ (STP) in solution (Spycher and Pruess 2005). This results in a water/gas ratio of 0.215 and is far higher than the largest estimated Stage 1 V_Water/V_Gas Bravo Dome value. This is consistent with the CO₂ gas charging the reservoir as a separate and distinct phase, and hence is consistent with the proposed CO₂ ‘stripping’ of the formation model (Fig. 15).

Table 3 Calculated formation water noble gas concentrations (cm³STPcm⁻³) and ratios (Gilfillan et al. 2008)

One can then calculate the volume of water, under reservoir conditions (Table 3), that the noble gases of the CO₂ phase must re-dissolve into, in order to fractionate the ²⁰Ne/³⁶Ar values from ASW (0.152) to the values observed. For the Bravo Dome reservoir these vary from 0 to 104 cm³ water per cm³ gas (RTP) (RTP = Reservoir Temperature and Pressure), which corresponds to 0–1.51 cm³ water per cm³ gas (STP). The volume of degassed water required to fractionate the ²⁰Ne/³⁶Ar from ASW to observed values is far larger than the volume of water ‘stripped’ to supply the gases, by up to a factor of 50. Little, if any, CO₂ will dissolve in this water while it remains CO₂ saturated. This analytical technique is able to quantify on a sample by sample basis the volume of water the CO₂ in this storage analogue has interacted with both (1) on reservoir filling; and (2) subsequent to reservoir filling.

5.4.4 Adsorptive Fractionation

Several studies have suggested that physical fractionation processes such as adsorption can result in a relative gas abundance pattern that is highly enriched in heavy noble gas isotopes (Kr and Xe). ¹³⁰Xe/³⁶Ar enrichment factors up to ~600 relative to air and ~200 relative to water in equilibrium with air having been observed. (Fanale and Cannon 1971; Kennedy et al. 1990, 2002; Torgersen et al. 2004). These components cannot be directly determined from ASW by distillation or re-equilibration with oil, gas or oil–gas systems and thus require desorption of these gases from organic matter into fluids (Kennedy et al. 2002; Torgersen and Kennedy 1999). Evidence for this hypothesis is provided by the fact that large quantities of Xe can be adsorbed onto carbon-rich materials at room temperature (Fanale and Cannon 1971; Podosek 1980). Recent work by Torgersen et al. (2004) has reported that these enrichments of noble gases could be accounted for by incomplete emptying followed by incomplete filling of lithic grains and half spaces.

5.4.5 Diffusive Fractionation

Maximum diffusive fractionation will occur under non-equilibrium conditions when the different diffusion properties of the noble gases and CO₂ control the partitioning rate from water into the gas phase. The fractionation or diffusion coefficient of a ratio A/B can be calculated using the Graham’s law of diffusion:

$$ \alpha_{\frac{A}{B}} = \sqrt {\frac{{M_{B} }}{{M_{A} }}} $$

(18)

where:

A and B:

Isotopes

M_A, M_B:

Mass of A and B

Hence, for the ²⁰Ne/³⁶Ar ratio α = 1.34, and for the ⁸⁴Kr/³⁶Ar system α = 0.655. As these values depend on the relative mass of the noble gas species, they are independent of the individual reservoir conditions. While some natural gas reservoirs show evidence of isotopic fractionation by diffusion (Zhou et al. 2005) others do not (Holland and Ballentine 2006; Holland et al. 2009).

6 Use of Noble Gases in Hydrocarbon Fields Relevant to CO₂ Storage

6.1 Tracing of CO₂ Injected for EOR, Mabee Oil Field Texas

Noble gas tracing has been used successfully for large-volume groundwater tracing in several different locations (Castro and Goblet 2003; Castro et al. 1998). Nimz and Hudson (2005) undertook an investigation to determine whether noble gases could also be used for CO₂ tracing. This study involved an analysis of noble gas tracer costs, availability, detection limits, and a comparison with other potential tracers.

At the time of the study no CO₂ storage operations were being conducted in which a noble gas tracer could be added therefore the work was undertaken at an active EOR location, the Mabee Oil field in Texas. Fortunately, the CO₂ being injected into the subsurface in the Permian Basin of West Texas contains unique and readily identifiable noble gas isotopic characteristics as the CO₂ originates from the extensive CO₂ reservoirs from the Colorado Plateau region previously discussed in this chapter. This CO₂ is transported via pipelines to West Texas. Therefore, Nimz and Hudson did not have to artificially add noble gases to the CO₂ stream and wait months or years for them to arrive at monitoring wells, they simply investigated the noble gas composition of the producing wells at the field.

Nimz and Hudson collected CO₂ from the pipeline prior to it being injected in the oil field and produced gases from 13 producing oil wells in the Mabee field. The gases extracted from all the Mabee field wells is collected and combined into a single pipeline and which is then mixed with incoming ‘fresh’ Colorado Plateau CO₂ and re-injected into the field. This ‘blend’ of CO₂ was also collected for noble gas analysis. This allowed comparison of the isotopic composition of the injected CO₂ with the CO₂ produced from wells a significant distance from the original injection point. This provided an excellent analogue of tracing in an engineered storage site where noble gases could be injected along with CO₂ and then their distribution could be monitored through outlying wells. It should be noted that the concentrations of noble gases in the injected natural CO₂ from the Colorado Plateau reservoirs is significantly lower than those that would arise from the addition of a spike of artificial noble gases, meaning the study represents a much more difficult monitoring situation than would arise in an engineered storage site.

The wells sampled are shown in Fig. 16, along with the wells that were injecting CO₂ during the period of sampling. As injection had been taking place for many years, the entire field has been subject to injected CO₂ flood.

Fig. 16

Map of the southern portion of the Mabee EOR field. The production wells that Nimz and Hudson sampled are shown with adjacent sample numbers. The injection wells shown are those that were known to be active during the sampling period. There are numerous other production wells which are not shown, these occur on approximately 1,000 ft spacings (Nimz and Hudson 2005)

Nimz and Hudson found that the incoming CO₂ (KMCO₂) was isotopically similar to the CO₂ found in McElmo Dome, indicating that the majority of the CO₂ injected into the Mabee field at the time of the study was from McElmo Dome (Fig. 17a). This was somewhat surprising for two reasons; firstly as the operators of the field thought they were receiving CO₂ from Bravo Dome and secondly as there are several CO₂ fields that feed the CO₂ pipeline and hence a more mixed noble gas composition was expected.

Fig. 17

a Plot of ²⁰Ne/²²Ne against ²¹Ne/²²Ne for the individual Mabee oil field samples (in grey), KMCO₂, which is the composition measured from the incoming CO₂ supply pipeline prior to injection and Blend CO₂ which is from the return pipeline and contains a mix of gases from all the active wells the incoming supply pipeline CO₂. Also plotted is the average Bravo Dome and McElmo Dome ratios from Gilfillan et al. 2008. It is clear to see that the KMCO₂ did not originate from Bravo Dome as the operators of Mabee had believed. b Plot of the excess ⁴⁰Ar against excess ²¹Ne compared to the concentration of air for the Mabee samples including the incoming pipeline CO₂ (KMCO₂) and the Blend CO₂. This shows clear mixing between the excess radiogenic ²¹Ne and ⁴⁰Ar from the injected KMCO₂ and atmospheric air. However the trend does not intersect the origin indicating that an additional crustal component, enriched in ⁴⁰Ar over ²¹Ne, must be present. c ²⁰Ne concentration plotted against ³⁶Ar concentration. This illustrates that the atmospheric air component in the Mabee samples did not originate from contamination during sampling as the samples do not plot near the fractionation trend expected if the injected KMCO₂ was in contact with atmospheric air. They plot on a similar trajectory to what would be expected if the samples were mixing with ASW but with elevated initial ²⁰Ne and ³⁶Ar concentrations. c ²⁰Ne concentration plotted against ³⁶Ar concentration. This illustrates that the atmospheric air component in the Mabee samples did not originate from contamination during sampling as the samples do not plot near the fractionation trend expected if the injected KMCO₂ was in contact with atmospheric air. They plot on a similar trajectory to what would be expected if the samples were mixing with ASW but with elevated initial ²⁰Ne and ³⁶Ar concentrations. d Plot of ¹³²Xe concentration plotted against ³⁶Ar concentration. This plot shows ¹³²Xe concentrations in the Mabee samples are enhanced relative to both the ASW and air fractionation trends starting from an initial KMCO₂ composition

Compared to KMCO₂ the Mabee production wells all exhibited more ‘air like’ noble gas composition. This was thought to be the result of either the baseline oil reservoir composition (present prior to CO₂ injection commencing) or from the large amount of water that is also injected into the field with CO₂. (This is because EOR operation in Mabee is a water alternating gas/CO₂ type (WAG), where water and CO₂ are injected into the field at separate times). It was possible to rule out atmospheric contamination during sampling as a cause of this component as the ²⁰Ne/³⁶Ar trend recorded from the samples was not close to the fractionation trend expected if the injected KMCO₂ was in contact with the atmosphere (Fig. 17c). Instead the samples plotted close to the expected fractionation which would occur if the KMCO₂ was in contact with air saturated water (ASW), indicating that the injected water in the field had a strong influence on the noble gas composition of the samples.

From Fig. 17 it is apparent that an additional component must be present in the ASW component because a clear excess in crustal radiogenic ⁴⁰Ar and nucleogenic ²¹Ne relative to the predominantly air-derived ³⁶Ar and ²⁰Ne was recorded. This excess component can be clearly seen in Fig. 17b as the trend exhibited by the Mabee samples toward a position with elevated ⁴⁰Ar relative to excess ²¹Ne. The radiogenic origin of ²¹Ne and ⁴⁰Ar requires this additional component must be crustal in origin.

A similar crustal component was observed in the Xe isotope composition. The samples were found to be enriched in crustal radiogenic ^132−136Xe relative to atmosphere derived ¹³⁰Xe. Figure 18d illustrates this excess ¹³²Xe relative to the ASW and air fractionation trends. Unfortunately it was not possible to collect a sample of the injected WAG water in the oil field and hence it is unclear if the crustal radiogenic component was present in this water. However, it is clear from the results of the study that if the groundwaters are the source of crustal radiogenic noble gases, they must have been isolated from atmospheric input for a significant period of time. It is believed that the radiogenic component may originate from the oil reservoir itself as the noble gas compositions measured in the samples from the oil field can be explained by a simple two-component mixture of KMCO₂ and a native subsurface component which had air-like neon isotope ratios along with a slightly enhanced ²⁰Ne concentration relative to ³⁶Ar, excess ⁴⁰Ar and enhanced ^132−136Xe relative to ¹³⁰Xe.

Fig. 18

a Water-derived and crustal-derived noble gas show evidence of coherent fractionation trend, which can be modelled as Rayleigh fractionation of a gas phase progressively exsolving from ASW water with a radiogenic component (recalculated and replotted from Zhou et al. 2005). b Ne concentrations predicted from the Rayleigh fractionation model are significantly higher than observed, requiring dilution of the groundwater-derived noble gases

The key conclusions from this study were: (a) the noble gas tracer was not overprinted in the oil reservoir because the noble gas composition of all of the production gas samples can only be explained by the presence of the KMCO₂ component. This is despite the presence of a crustal radiogenic component and a possible WAG component, showing that even if the injected CO₂ moved significantly from the active injection area, the KMCO₂ component would still be identifiable, proving that noble gases can be used as tracers of CO₂ migration in a deep reservoir; (b) the sampled KMCO₂ had a ¹³²Xe concentration of 1.62 × 10⁻⁷ ppmv, which is two orders of magnitude above what would be required for groundsurface monitoring indicating that ¹³²Xe could be used as a tracer of injected CO₂ within the subsurface and probably as a tracer of the migration of deep CO₂.

The noble gas compositions observed in the Mabee Field were very systematic and formed linear trends with only small deviations that could be accounted for by the presence of other known subsurface components, such as a crustal component. This coherent and systematic behaviour of the noble gas system bodes well for using noble gases as tracers of CO₂ in the subsurface. The systematic results obtained from the Mabee study clearly show that even very small isotope variations can be detected using existing noble gas analysis techniques. The isotope variations that would result from injection of artificial noble gas tracers would be much easier to detect and hence Nimz and Hudson demonstrated that noble gases could be used as reliable and robust tracers of CO₂ migration within an engineered storage site.

6.2 Noble Gas Tracing of Groundwater/Coalbed Methane Interaction, San Juan Basin, USA

Coal beds that are not of economic interest for coal extraction are becoming increasingly important as a possible sink for anthropogenic CO₂ emissions (Klara et al. 2003; Pashin and McIntyre 2003). Groundwater plays an important role in CO₂ sequestration in coalbeds in a similar fashion to the controls it exerts on CO₂ storage in CO₂ reservoirs. Consequently noble gases can be used to develop a model of groundwater interaction with coal beds that can inform on the CO₂ sequestering process.

The San Juan Basin natural gas field, located in northwestern New Mexico and southwestern Colorado in the USA, is a case-type coalbed methane system which was investigated by Zhou et al. (2005). Data from the San Juan field show ²⁰Ne/³⁶Ar ratios significantly lower than air but ⁸⁴Kr/³⁶Ar and ¹³⁰Xe/³⁶Ar mostly higher than air which suggests, like the CO₂ fields in the previous section, an open system Rayleigh fractionation process but one in which gas is continuously exsolved and then lost from the system. This can be modelled and, in the case of San Juan data, displayed as the variation in radiogenic ⁴He/⁴⁰Ar with atmospheric ²⁰Ne/³⁶Ar. In contrast to the Bravo Dome data, the data converge at the origin when V_gas → ∞ and noble gases have fully exsolved from the fluid phase (Fig. 18a).

However, noble gas data show us this is not the complete story. Using ASW as an initial fluid composition that plausibly fits the San Juan data allows us to predict the ²⁰Ne concentration at any stage of the degassing process (i.e. at any ²⁰Ne/³⁶Ar). The observed ²⁰Ne concentrations are orders of magnitude lower than predicted (Fig. 18b), as also observed in the CO₂ gas fields of SW USA. In the Bravo Dome CO₂ system, this was explained by CO₂ injection stripping the noble gases from the groundwater, owing to the higher solubilities of noble gases in CO₂ than in water (i.e. the re-dissolution model). However, large scale CO₂ migration is not present in the San Juan basin and hence another explanation is required. The model suggested by Zhou et al. (2005) is that a CH₄ component devoid of any noble gases, and therefore isolated from any groundwater since formation, has diluted the CH₄ component that does contain noble gases from the Rayleigh fractionation process as shown in Fig. 19 (adapted from Zhou et al. 2005).

Fig. 19

Cartoon of Rayleigh fraction and dilution model adapted from Zhou et al. 2005. Gas exsolved from the ASW is heavily fractionated and diluted with gas desorbed from the coal bed that has not previously interacted with the groundwater system and is therefore devoid of noble gases

Further evidence of phase interaction can be identified using the heavy noble gases. In a plot of ²⁰Ne/³⁶Ar v ¹³⁰Xe/³⁶Ar (Fig. 13 in Zhou et al. 2005), San Juan data show significant but variable excesses of Xe above that predicted from Rayleigh fractionation. When these Xe data are corrected for the dilution effect of the CH₄ desorbed from the coal bed (the ratio of [²⁰Ne] observed divided by [²⁰Ne] modelled), the Xe excesses disappear and the model fits the data excellently. This independent evidence of a dilution effect from Xe enriched coal corroborates the model presented in Fig. 19. This in turn shows that groundwater is not the only control on coal biodegradation and therefore groundwater alone does not determine the viability of coal beds as a CO₂ sequestration reservoir: other important factors include overpressure that may help maintain the integrity/isolation of coal bed seams.

As described in the Sect. 5, step-wise equilibrium degassing can approximate Rayleigh fractionation if at each stage of degassing, V_gas/V_water is small. In this case study, the differences between the modelled and observed Ne concentration can be quantified (Fig. 18b). Using this dilution ratio together with calculations of gas production history as performed on the CO₂ data in Sect. 5 (for details of this case study see Table 3 Zhou et al. 2005) one can quantify the volumes of water associated with a gas producing well. Assuming coal bed thickness and specific porosity, one can calculate the volume of water involved in gas production of specific wells. Expressed as radii from the well, in the San Juan field these volumes vary from 16 to 242 m (Zhou et al. 2005). This is significantly smaller than the distance between wells. This in turn indicates there is no need for dynamic groundwater flow between wells on the timescale of well gas production and, in this case, suggest that more CH₄ wells can be drilled without adversely affecting existing well gas production.

In the context of the carbon sequestration, this information is important for understanding reservoir connectivity and the presence of absence of continuous water flow through a basin in light of the evidence that the majority of CO₂ sequestration analogues contain CO₂ dissolved in groundwater (see Sect. 5). In exhausted CO₂ reservoirs or basins undergoing enhanced oil recovery, this type of analysis may also provide information on the extent to which new drilling may increase the storage capacity of a potential subsurface CO₂ storage site.

7 Use of Noble Gases for Monitoring of Subsurface CO₂ Migration and Leakage

7.1 Noble Gases as Tracers of CO₂ Leakage at Green River Springs, Utah, USA

Wilkinson et al. (2010 ) examined naturally occurring noble gas tracers and isotopic ratios from surface gas samples from a suite of natural CO₂ springs. These springs provide an analogue for some time in the future, when a very severe failure of a containment site has been detected after injection operations have ceased. In the imagined scenario, there was a reservoir of artificially injected CO₂ which existed as a discrete fluid together with shallower aquifers which also contain natural CO₂. It was imagined that a baseline survey had not been undertaken prior to CO₂ storage commencing, so there was only proxy information available from regional archives to identify the chemistry and gas contents of the diverse shallower CO₂ sources and aquifers. The key questions that the study aimed to address were; could CO₂ migration from an injection zone, through the overburden, to the surface, be quantified? Could different sources of CO₂ be discriminated?

The group of CO₂-rich springs sampled by Wilkinson et al. (2010) were located 10 km south of the town of Green River, southwest Utah. They are CO₂-charged springs associated with the Little Grand Wash and Salt Wash faults lying at the northern end of the Paradox Basin in south west Utah. These control the present day flow of gas and groundwater to the surface (Heath 2004; Shipton et al. 2004) with both being part of a series of steeply dipping (70–80º) WNW trending normal faults that are found in the region. Springs of CO₂-rich water and free gas, an active CO₂ charged geyser, and both active and fossil travertine deposits are associated with the Little Grand Wash fault zone. The most notable spring is the cold-water CO₂-charged Crystal Geyser, which erupts to heights of up to 25 m at 4–12 h intervals (see Shipton et al. 2004 for further details). In addition to Crystal Geyser, numerous CO₂-charged springs are associated with the Salt Wash fault zone (see Heath (2004) for further details).

The gases emitted by the Green River Springs are predominantly CO₂ but also include nitrogen and trace noble gases (Heath 2004). Prior to the study of Wilkinson et al. CO₂ and other gases in the springs were considered to originate from a combination of sources including mantle-degassing, clay-carbonate reactions, the thermal metamorphism of carbonates or atmospheric and soil gas recharge combined with diagenetic reactions such as the maturation of organic material.

In order to calculate crustal and mantle inputs to the CO₂, Wilkinson et al. first quantified the amount of air present in each sample. This was calculated using the measured ⁴⁰Ar/³⁶Ar ratio for each spring using the following formula:

$$ \left[ {^{{{\text{4}}0}} {\text{Ar}}} \right]_{{{\text{air}}}} = {\text{ }}\left[ {^{{{\text{4}}0}} {\text{Ar}}} \right]_{{{\text{meas}}}} {\text{ }} \times {\text{ }}\left( {^{{{\text{4}}0}} {\text{Ar}}/^{{{\text{36}}}} {\text{Ar}}} \right)_{{{\text{air}}}} /{\text{ }}\left( {^{{{\text{4}}0}} {\text{Ar}}/^{{{\text{36}}}} {\text{Ar}}} \right)_{{{\text{meas}}}} $$

(19)

The proportion of air-derived Ar was found to be high (33–91 %) as a result of the high ⁴⁰Ar content of the atmosphere, compared to the much lower crust and mantle contributions. Figure 20 clearly shows how the amount of dissolved air in the samples was controlling both the ⁴⁰Ar/³⁶Ar and ⁴He/²⁰Ne ratios. Using the air-derived ⁴⁰Ar concentration, it was straightforward to calculate the air contribution to the total erupted CO₂ using the known CO₂/⁴⁰Ar for air of 0.0398 (Porcelli et al. 2002). The air contribution to the CO₂ was found to be insignificant (0.00003 to 0.00018 volume  % CO₂).

Fig. 20

Plot of ⁴He/²⁰Ne against ⁴⁰Ar/³⁶Ar for the Green River Springs. It is clear that the quantity of dissolved air in the samples is controlling both the ⁴⁰Ar/³⁶Ar and ⁴He/²⁰Ne ratios, with the greatest amount of air in the samples with the lowest ratios

The measured ³He/⁴He of the gases erupting from the springs showed a small variation from 0.224 to 0.265 R_a, with the exception of Pseudotenmile geyser which exhibited a higher ratio of 0.386 R_a. This spring was also noted to have a very slow gas discharge rate, possibly enabling atmospheric noble gases to become dissolved in the spring water to form a component of the measured gas, possibly explaining the higher ³He/⁴He ratio observed.

Using the established technique (Craig et al. 1978) the authors attempted to correct the measured ³He/⁴He ratios for the air component, leaving just the crust and mantle input. However, the correction was very small and significantly less than the precision of the data. This was unsurprising given the measured range of ⁴He/²⁰Ne of 313 to 13,002, significantly above the air value of 0.288 (Ballentine et al. 2002). All of the ³He/⁴He ratios were significantly higher than the crustal ³He/⁴He ratio of 0.02 R_a but also lower than the sub-continental mantle ratio of 6 R_a (Fitton et al. 1991). Using these end-member values, it was possible to calculate the proportion of the He from both crustal and mantle sources using the following formula (from Ballentine et al. (2002)).

$$ ^{{\text{4}}} {\text{He}}_{{{\text{crust}}}} {\text{ }} = {\text{ }}\left( {^{{\text{4}}} {\text{He}}_{{{\text{measured}}}} \, \times \,{\text{ }}^{{\text{3}}} {\text{He}}/^{{\text{4}}} {\text{He}}_{{{\text{mantle }}}} - \,^{{\text{3}}} {\text{He}}/^{{\text{4}}} {\text{He}}_{{{\text{measured}}}} } \right)/\left( {^{{\text{3}}} {\text{He}}/^{{\text{4}}} {\text{He}}_{{{\text{mantle}}}} - {\text{ }}^{{\text{3}}} {\text{He}}/^{{\text{4}}} {\text{He}}_{{{\text{crust}}}} } \right) $$

(20)

The crustal contribution to ⁴He in the samples ranged from 94 to 97 % with the remainder originating from the mantle. It was also possible to calculate the crustal contribution to the measured ³He concentration using the known continental crust ³He/⁴He ratio of 0.02 R_a. Only 5–8 % of the total ³He originated from the crust with the remainder from the mantle.

The known range of mantle CO₂/³He ratios was used (1 × 10⁹ to 1 × 10¹⁰) to calculate the proportion of CO₂ which had originated from both the mantle and crust, knowing the atmosphere contribution to the CO₂ was negligible. An unknown is the crustal CO₂/³He ratio, which is variable (O’Nions and Oxburgh 1988). However, a simple mass balance approach allowed the crustal CO₂/³He ratio to be calculated.

$$ ^{3} {\text{He}}_{{{\text{mantle}}}} \times {\text{ CO}}_{2} /^{3} {\text{He}}_{{{\text{mantle}}}} + \,^{3} {\text{He}}_{{{\text{crust}}}} \times {\text{ CO}}_{2} /^{3} {\text{He}}_{{{\text{crust}}}} = 1 $$

(21)

The results of these calculations are shown in Table 4. Using a CO₂/³He ratio of 1 × 10 ⁹ the crustal contribution of CO₂ is uniformly between 84 and 99 % of the total CO₂. Use of a CO₂/³He ratio of 1 × 10¹⁰ gives a wider range of crustal compositions ranging from 0 % for Tenmile Geyser to 98 % for Big Bubbling Spring. Unfortunately from these results, it is not possible to determine if the crustal component was derived from deep crustal reactions, or from aquifer water that had interacted with crustal minerals. The calculated CO₂/³He ratios for the crust ranged from 5.81 × 10¹⁰ to 6.53 × 10¹², which are within the range considered by O’Nions and Oxburgh (1988) to be typical of the crust.

Table 4 Calculated origins of CO₂ erupted at Green River Springs

In this study noble gas tracers provided substantial constraints on the origins of CO₂ erupted at the Green River Springs, and hence were much more useful than stable isotope data alone. However, due to the lack of measured baseline data on the noble gas composition and CO₂ contents of the different aquifers beneath the springs, noble gases could not provide an unequivocal determination of the origin and transport mechanisms of the CO₂ erupted (Fig. 21).

Fig. 21

Cartoon summary of the imagined CO₂ sources beneath the Green River Springs. There is no evidence in the noble gas data for a large subsurface accumulation of free CO₂ (Wilkinson et al. 2010)

It was possible to determine that the erupted CO₂ is largely derived from crustal sources, with a 0.2–31 % contribution from the mantle, with the exception of a single spring (Tenmile Geyser) that could be erupting mostly mantle CO₂. This origin for the erupted CO₂ is in contrast to the large, commercially exploited CO₂ accumulations that surround the Colorado Plateau, which are of purely mantle origin. Hence, Wilkinson et al. (2010) were able to determine that the Green River Springs are not derived from these large accumulations. Additionally, they also found that the Green River Springs’ noble gases are quite variable in composition, arguing against a homogeneous reservoir of free phase CO₂ within the shallow Navajo Aquifer. Nevertheless, they also argue that some ascent of free CO₂ gas must have occurred because the measured surface volumes contained too much CO₂ to have been dissolved in the volume of water erupted at the springs. It is clear from this study that the CO₂ is not derived from a single reaction or source, as had been previously assumed, but is predominantly from crustal diagenetic reactions within the shallow Navajo Aquifer, except for the 1–31 % mantle contribution thought to have originated from the much deeper Paleozoic aquifer.

Whilst the noble gas tracers were not entirely diagnostic of the processes forming the CO₂, with uncertainty still surrounding CO₂ origin and transport mechanisms, the detection of leaks in engineered storage scenarios should be easier as the CO₂ will have a distinctive isotopic composition. Additionally a comprehensive baseline survey will be required at any proposed storage site and with better baseline geochemical data, more would be possible.

7.2 He and Rn in Soil Gases and CO₂ Seeps from the Massif Central, France

An extensive geochemical study of natural CO₂ release at the surface was undertaken by Battani et al. (2010) in Sainte-Marguerite, in the Massif Central region of France. This examined soil gas fluxes and concentrations as well as gas from CO₂ rich bubbling springs. These seeps provide a natural laboratory to study CO₂/fluid/rock interactions and the migration mechanisms of CO₂ migration to the surface. The Sainte-Margurite region is located in the southern part of the Limagne graben and contains many CO₂-rich springs caused extensive natural emissions of CO₂. The springs sampled were associated with two major fault systems dating from the tardi-Hercyian period. The area also contains travertine deposits associated with the CO₂-rich springs.

Soil gas measurements were made over a small area (50 × 100 m) surrounding an ancient thermal bath over two field campaigns in 2006 and 2007. ⁴He and ²²²Rn abundance measurements of the soil gases were made in the field and gas samples were also collected for analysis in the laboratory for noble gas abundances, isotopes and δ¹³C (CO₂) isotopes. All samples for laboratory analysis were collected using a micro gas chromatograph as a pump to collect gases from depth of 1 m. N₂, Ar and O₂ peaks were monitored during collection to ensure atmospheric gases had been purged from the collection system. Samples of gas were also collected from numerous CO₂ rich bubbling water springs, inferred to be deep gas (Battani et al. 2010).

Measured helium isotopic ratios in the CO₂ exsolving from the bubbling springs ranged from 0.76 to 6.62 R_a. The upper value indicates some samples contain a pure mantle component, given it is above the European Sub Continental Lithopsheric Mantle (SCLM) value of 6 R_a (Gautheron and Moreira 2002) and close to the upper mantle value of 8 R_a (Kurz and Jenkins 1981). Additional evidence of a predominantly mantle origin for the CO₂ came from the measured δ¹³C (CO₂) values which were close to −5 %, within the range of mantle derived CO₂ (Marty and Zimmermann 1999). Measured ⁴He concentrations ranged from 0.12 to 39.12 ppm and somewhat surprisingly the majority of the samples exhibited lower ⁴He concentrations than the atmospheric value of 5.24 ppm. The low ⁴He concentrations were particularly unexpected given that the springs emerge from granitic basement which is typically rich in U and Th and hence, as a result of their radioactive decay, also ⁴He. Battani et al. (2010) argue that this result can only be explained if the fluids migrate rapidly from depth to the atmosphere, and do not interact with crustal fluids. The rapid migration of the CO₂ gas is further confirmed by the observed Ar isotopic fractionation, which is consistent with a Rayleigh type fractionation process. This is illustrated in Fig. 22, which shows that both Ar isotopic ratios (³⁸Ar/³⁶Ar and ⁴⁰Ar/³⁶Ar) follow the Mass Fractionation Line (MLF). Such mass fractionated values are a strong indication of a fast migration process which does not allow time for re-equilibration with other fluids. The gas was therefore assumed to travel fast along the deep-rooted faults of the granitic basement.

Fig. 22

Plot of ⁴⁰Ar/³⁶Ar against ³⁸Ar/³⁶Ar from the CO₂ exsolving from the bubbling springs. This illustrates that the majority of the samples plot along the black mass fractionation black line. Battani et al. (2010) argued that the increase of some ⁴⁰Ar/³⁶Ar ratios could either be explained by a mantle derived input or the addition of a small component of radiogenic ⁴⁰Ar*. This type of kinetic fractionation is indicative of rapid migration of the fluid

This is a significantly different scenario to that observed at the Green River Springs in the Colorado Plateau (Sect. 7.1), where the noble gas and δ¹³C (CO₂) data indicated that most of the CO₂ was dissolved in water at depth prior to degassing at the surface.

The δ¹³C (CO₂) of the soil gas samples ranged between −3 and −5.1 %, slighter heavier than those measured in the CO₂ degassing from the Bubbling Springs, indicating either a crustal or mantle origin of the gas. Battani et al. argued that this was the result of interaction of the gas with the local travertines, which have a δ¹³C signature of between 4.4 and 7.4 % (Casanova et al. 1999). Thermal breakdown of travertine would also explain the higher than mantle CO₂/³He ratios which were measured in the gases from some of the bubbling springs. This interaction between soil gas and travertine was consistent with the measured ²²²Rn activity, as the soil gas samples from above the travertine deposits exhibited higher activities, reaching up to 2 × 10⁶ Bq/m³. Travertines are between 300 to 10,000 times more enriched in uranium and thorium than typical surface sediments (Casanova et al. 1999). Comparing the soil gases that contained over 50 % by volume of CO₂ with radon activities in excess of 50,000 Bq/m³ indicated a strong correlation between the two gases. The enriched Rn activity allowed the authors to identify preferential pathways for migration of gas from depth in the soil gas samples which trended NE-SW aligned to the Allier-river axis, a structure of related to the former opening of the Limagne graben.

The soil gas He concentrations were less diagnostic as both low (less than 5 ppm) or high (over 6 ppm) helium values were exhibited by the samples with high measured high CO₂ concentrations. However, significantly lower than atmosphere values of Ar were measured and this was attributed to the soil gases containing a significant amount of the deep gas flux also found in the bubbling springs. Despite the lack of a clear correlation of high CO₂ concentrations with He, this study highlighted the role that multiple noble gases (in this case He, Ar and Rn) and C isotopes can play in identifying the migration of CO₂ from depth. It also reinforced the importance of a detailed background geochemical study prior to a storage site being utilised, in order to determine the natural noble gas and C isotope geochemistry which might not be as predicted (i.e lack of ⁴He in the deep CO₂ and high natural ³He/⁴He as observed in this work).

7.3 He and Ne as Tracers of Natural CO₂ Migration at St. Johns Dome, Arizona, USA

An important aspect of engineered CO₂ storage is the ability to monitor the storage site, and to trace the CO₂ after injection. Future engineered CO₂ storage sites will consist of both the reservoir or saline formation into which the CO₂ is injected, and the overlying suite of rocks above the reservoir or saline formation up to the ground surface. Hence, the overburden will become legally defined as part of the ‘storage complex’ (EU 2009), and its containment performance needs to be evaluated to ensure that CO₂ can be safely stored over geological timescales. However, it is extremely difficult to unequivocally detect the small releases of anthropogenic CO₂ that could arise from a diffuse leakage of CO₂ from a storage site. This is because there are many natural sources of CO₂ within the crust with overlapping signatures, including breakdown of carbonate minerals or cements, biological activity or hydrocarbon oxidation. Therefore, a more specific and unambiguous method is needed to fingerprint deep-derived CO₂ migration, but it has not yet been understood if fingerprints remain intact during hundreds of metres of migration and diverse interacting geochemical and petrophysical processes.

As an analogue for post-emplacement seepage, Gilfillan et al. (2011) have recently examined natural CO₂ rich springs and groundwater wells in the vicinity of the St. Johns Dome CO₂ reservoir located on the border of Mid-Arizona/New Mexico, USA. Extensive travertine deposits surrounding the CO₂ reservoir highlight a past history of migration of CO₂ rich fluids to the surface. However, there is no evidence that travertine formation is occurring at present and no geysers with gaseous CO₂ are currently found in the area (Moore et al. 2005). High concentrations of HCO₃⁻ are common in surface springs, shallow groundwater wells used for irrigation and deeper wells used to obtain cooling water for a coal fired power plant in the region. Previous chemical analysis of these waters implied a possible connection between the formation water within the CO₂ reservoir and the HCO₃⁻ rich water. This link was not found to be conclusive, due to significant differences in water types (Moore et al. 2005) and the fact that a soil gas survey surrounding these springs and wells was unable to differentiate additional CO₂ flux from that of background biological activity (Allis et al. 2005).

δ¹³C measurements have proven to be extremely valuable in tracing CO₂ injected into early CO₂ storage test sites, such as Weyburn (Raistrick et al. 2006). However, in the St. Johns Dome waters there was no clear relationship between the HCO₃⁻ content of and the δ¹³C_DIC ratio or sample depth (Fig. 23a). This was in stark contrast to the relationship observed between HCO₃⁻ concentration and ³He/⁴He, which showed a distinction between the above air ratios measured in three spring water samples (Willowbank Spring, 24 Bar Ranch and New Mexico Spring—hereafter termed the anomalous springs) and the below air ³He/⁴He measured in all of the other samples (Fig. 20b). The above air values in the three springs were attributed to the presence of a small amount of excess ³He produced by tritium decay, indicating a contribution of recent tritium bearing water. There was also a marked distinction in both the ³He and ⁴He concentrations measured in these three anomalous samples, which were an order of magnitude lower than other water samples (Fig. 23). With the exception of Colorado River Spring, all of the remaining water samples exhibited higher HCO₃⁻ concentrations which corresponded to lower ³He/⁴He ratios (Fig. 23b).

Fig. 23

a δ¹³C_DIC variation plotted against HCO₃⁻ concentrations for the surface springs, groundwater and CO₂ well sampled at St. Johns. There is no clear relationship between the HCO₃⁻ concentration of the water and δ¹³C or the sample type. b ³He/⁴He ratios plotted against HCO₃⁻ concentrations for the surface springs, groundwater and CO₂ well sampled at St. Johns. Values obtained from groundwater wells cluster above ³He/⁴He = 1, equating to air. A clear correlation exists between increasing HCO₃⁻ concentrations and decreasing ³He/⁴He, trending toward the values measured in the deep CO₂ reservoir. A clear distinction existed between the above air ³He/⁴He ratios measured in three anomalous spring water samples; Willowbank Spring, 24 Bar Ranch and New Mexico Spring and the other water samples

Significantly higher than atmosphere concentrations of ⁴He in all of the groundwater wells and majority of surface springs implies the addition of a crustal component that had accumulated radiogenic ⁴He. Thus, a portion of the water originated at depth, having circulated in the crust for a significant period of time (Torgersen and Clarke 1985). The ⁴He excess was further illustrated by the significantly above air ⁴He/²⁰Ne measured in all water samples, excluding the similar to air values measured from the three anomalous springs. Figure 24a shows that the low ³He/⁴He and high ⁴He within the groundwater wells and majority of surface springs (bar the three anomalous springs) could be explained by simple mixing between low ³He/⁴He and high ⁴He values measured in Wells 10–22 and 22–1X and varying percentages of shallow groundwater (high ³He/⁴He with low ⁴He concentrations). This clearly illustrated that the excess ⁴He fingerprint entrained at depth was retained after migration of the waters to the surface.

Fig. 24

a ⁴He plotted against ³He/⁴He for the St. Johns waters. Mixing lines are plotted from the different compositions of deep groundwater measured from the end member CO₂ wells to 100 % air. The low ³He/⁴He and high ⁴He within the groundwater wells and majority of surface springs can be explained by simple mixing between the ³He/⁴He and ⁴He values measured in Wells 10–22 and 22–1X and varying amounts of air (with the per cent air indicated on tick marks). This proved that the excess ⁴He component entrained at depth was retained and not lost on migration of the waters to the surface. b ³He plotted against ³He/⁴He for the St. Johns water samples. As with ⁴He, high ³He concentrations and low ³He/⁴He ratios were observed in groundwater wells and majority of the surface springs. These can also be accounted for by the mixing of waters from wells 10–22 and 22–1X and varying amounts of air. This shows that the whilst the high magmatic ³He signature exhibited by the deep CO₂ wells, although diluted, was retained and not lost as a result of transport water to the surface

As with ⁴He, the low ³He/⁴He and corresponding high ³He values within the deep groundwater wells and most surface springs, could be explained by simple mixing between water from Wells 10–22 and 22–1X and air (Fig. 24b). This clearly indicated that all of the waters rich in ³He must have interacted to some degree with the magmatic CO₂ in the deep reservoir. Furthermore, the fingerprint of high ³He from magmatic origin, exhibited by the deep CO₂ wells, although diluted, was also retained during transport of water and CO₂ to the surface.

All of the sampled waters, bar the three anomalous air-like springs, exhibited CO₂/³He ratios which were extremely similar to the range observed in the three deep CO₂ wells (Fig. 25a and b). These results further reinforce the observation that all of the water samples, except for the anomalous springs, contained CO₂ of magmatic origin, which had originated from the CO₂ reservoir. The higher CO₂/³He ratios observed in the three anomalous springs indicated that they contained CO₂ with less ³He content compared to that of the deep reservoir, providing further evidence that the anomalous springs contain additional CO₂ from a different source. Several samples exhibited ratios below the lowest observed CO₂ well value of 9.75 × 10⁷, which implied that some of the CO₂ component had been lost relative to ³He.

Fig. 25

a CO₂/³He variation plotted against ²⁰Ne for the St. Johns water samples. All of the groundwater springs and the majority of the surface springs exhibit CO₂/³He ratios which are extremely similar to those measured in the deep CO₂ wells, so implying that they contain dissolved CO₂ which migrated from the CO₂ reservoir. A clear trend of decreasing CO₂/³He with increasing ²⁰Ne existed in all of the data. As ³He is a conservative tracer, reduction of CO₂/³He can only be explained by a reduction of the CO₂ component. As the groundwater is the main source of ²⁰Ne, this CO₂ reduction must be linked to contact with the groundwater. b CO₂/³He variation plotted against ⁴He for the St. Johns samples. A clear reduction in the CO₂/³He ratio corresponds to an increase in ⁴He. This can be explained by the mixing of ⁴He, which has accumulated in the deep formation water, with the younger shallow groundwater and then migrating from the reservoir, along with dissolved CO₂. This trend is identical to that observed in natural CO₂ reservoirs from around the world (see Fig. 9)

Gilfillan et al. (2011) also documented a clear relationship between CO₂/³He reduction and increases in both ²⁰Ne and especially ⁴He (Fig. 25) within the St. Johns water samples. As previously outlined in this chapter, similar relationships have been observed in numerous CO₂ reservoirs around the world (Gilfillan et al. 2009; Sherwood Lollar and Ballentine 2009). Whilst there are means by which crustal CO₂ (CO₂/³He > 10¹⁰) could be added to an initial mantle rich CO₂ accumulation (Bradshaw et al. 2004; Cathles and Schoell 2007), there is no plausible mechanism that would allow crustal CO₂ to be added whilst preserving the correlation between CO₂/³He reduction and increases in noble gases derived from the groundwater. Hence, the changes in CO₂/³He must be due to CO₂ loss in the subsurface by an amount directly related to the quantity of groundwater that the CO₂ has contacted. As CO₂ is soluble and reactive, the most probable mechanisms of subsurface CO₂ fluid phase removal are dissolution of the CO₂ into the groundwater and/or mineral trapping (Baines and Worden 2004b; Bradshaw et al. 2004). As previously highlighted earlier in this chapter it has been possible in some natural CO₂ reservoirs to distinguish between proportions of CO₂ that were dissolving into the formation water and those which were precipitating as carbonate minerals (Gilfillan et al. 2009). However, due to the lack of any correlation between the δ¹³C and CO₂/³He this was not possible within the St Johns water samples.

The results of this study illustrate that the He and Ne concentrations and ³He/⁴He and CO₂/³He ratios provide characteristic fingerprints, which could be identified in all of the groundwater wells and all but three of the surface springs. The simplest explanation is that the groundwater wells sampled shallow well and spring waters containing noble gases together with magmatic CO₂ derived from the deep reservoir. In this model, the high concentrations of HCO₃⁻ present in the sampled waters are then the direct result of the migration of dissolved CO₂ from the deep reservoir, illustrating for the first time that CO₂ can be fingerprinted from source to surface using noble gases, particularly He. This paves the way for the use of similar techniques to identify dissolved CO₂ migration from future engineered storage sites.

7.4 Noble Gas Investigation of Alleged CO₂ Leakage at Weyburn

In January 2011 it was extensively reported that the Kerr family had been forced to move from their property located above the Weyburn-Midale Monitoring and Storage Project in Saskatchewan, Canada. A geochemical consultant from Petro-Find GeoChem Ltd., who was hired on behalf of the Kerr’s, reported measurements of δ¹³C (CO₂) isotope values in soil gases rich in CO₂ which were similar to those of the CO₂ injected into the deep oil reservoir (Lafleur 2010). The Petroleum Technology Research Centre (PTRC), who are responsible for the environmental monitoring of the Weyburn CO₂-EOR and storage operation, published a detailed response correctly stating that Petrofind had not taken the similar baseline δ¹³C (CO₂) isotope measurements conducted prior to the injection into account (Petroleum Technology Research Centre 2011). Also as δ¹³C (CO₂) is not a unique tracer, there were several other natural sources that could account for the measured values. Whilst this response went some way to addressing the public perception fears raised by the alleged CO₂ leakage claims, it was clear that more research was required to re-establish confidence in the safety and security of CO₂ stored at Weyburn. This was imperative for both the project itself and the future acceptance of the safety and security of CO₂ storage technologies. The International Performance Centre for Geologic Storage of CO₂ (IPAC-CO₂) undertook a detailed independent incident response protocol focused on the near surface soil gases, the noble gas composition of the shallow groundwaters and a hydrogeological analysis. In the following section we summarise the findings of the noble gas investigation.

In order to determine if migration of dissolved CO₂ from the Weyburn oil field was responsible for the alleged CO₂ anomaly, Gilfillan and Haszeldine (2011) completed noble gas measurements on water samples collected from four groundwater wells surrounding the Kerr quarter, near Goodwater in Saskatchewan. These were compared to the noble gas composition measured from fluids obtained from a production well and CO₂ and water obtained from two separate injection wells on the Weyburn-Midale oil field, located near to the Kerr quarter. The aim of the work was to test the hypothesis that migration of dissolved CO₂, originating from either the free phase CO₂ or water injected into the Weyburn-Midale oil field or from CO₂ contained in the produced fluids, was responsible for the CO₂ anomaly reported at the surface. This focused on dissolved CO₂ in the groundwaters as no free phase CO₂ was present in the sampled well waters and as recent research has shown that the sensitivity of He as a tracer is two orders of magnitude greater in groundwaters than in soil gases (Mackintosh and Ballentine 2012).

Gilfillan and Haszeldine (2011) found that the measured ⁴He/²⁰Ne ratios in the injected CO₂ and produced fluid samples were significantly above the calculated air saturated water range of 0.250 to 0.296 (Fig. 26). This indicated that atmospheric He input to these samples was minimal. The injected water ⁴He/²⁰Ne ratio of 1.42 implied that there was a small amount of atmospheric He input in this sample. However, this was considerably above the range of 0.248 to 0.403 observed in the groundwater samples. This overlapped with the air saturated water range (assuming a 0–50 % excess air component and a re-equilibration temperature range on 10 to 25 °C) and indicated that almost the entire of the ⁴He dissolved in the groundwaters was of atmospheric origin.

Fig. 26

³He/⁴He ratios plotted against ⁴He/²⁰Ne ratios. The produced CO₂ and injected CO₂ exhibited ratios well in excess of the air saturated water values. The groundwater wells were extremely close to the air saturated water (ASW) ratio range indicating that almost all of the ⁴He they contain is of atmospheric origin (Gilfillan and Haszeldine 2011)

Similarly, there was a distinct difference in ³He/⁴He ratios in the different sample types, with the lowest values of 0.173 and 0.179 R_a measured in the produced CO₂ from the Weyburn field (Fig. 26). The injected CO₂ had a similar, albeit slightly higher ratio of 0.193 R_a, with the injected water sample ratio being higher again at 0.295 R_a. The range observed in the groundwater samples of 0.897 to 1.103 R_a was significantly above the ratios observed in the produced fluids, injected water and CO₂ samples, consistent with the ⁴He/²⁰Ne ratios in indicating that almost the entire of the ⁴He in the groundwater samples originated from the atmosphere.

⁴He and ⁴⁰Ar concentrations also exhibited marked distinctions depending on sample type (Fig. 27). The lowest ⁴He concentrations were those calculated for water equilibrated with the injected CO₂. The groundwater samples exhibited a measured range of ⁴He which was similar to the calculated air saturated water concentration range. This clearly showed that there was no significant addition of ⁴He to the groundwater samples. Measured ⁴He concentrations in the injected water were an order of magnitude higher, with the calculated ⁴He concentration of water in equilibrium with the produced fluid samples being over another order of magnitude higher still. As with ⁴He the lowest ⁴⁰Ar concentration was that calculated for water in equilibrium with the injected CO₂ under reservoir conditions (Fig. 27). The next highest were those calculated for water in contact with the produced fluids from the Weyburn field with the injected water having an order of magnitude higher ⁴⁰Ar concentration. The highest concentrations were those observed in the groundwater wells which were within the calculated air saturated water range.

Fig. 27

⁴He concentration plotted against ⁴⁰Ar concentration. All of the groundwater samples plot within the air saturated water (ASW) concentrations indicating that there is no input of crustal ⁴He and the ⁴⁰Ar is entirely derived from the atmosphere. The lower ASW concentrations are those with no excess air component and the higher values are those with a 50 % excess air component

Both ²⁰Ne and ³⁶Ar concentrations exhibited extremely similar concentration groups to ⁴⁰Ar, with the lowest observed in the calculated water concentration equilibrated with the injected CO₂ and highest being those observed in the groundwater wells (Fig. 28). As ²⁰Ne and ³⁶Ar are primarily derived from the atmosphere, this reinforced that the groundwater samples contained only atmospheric noble gases.

Fig. 28

²⁰Ne concentrations plotted against ³⁶Ar concentrations. There is a pronounced difference in concentrations between the produced fluids and injected CO_2, the injected water and the groundwater samples. All but one of the groundwater samples can be explained by the calculated air saturated water concentration range, indicating that all of the ²⁰Ne and ³⁶Ar in the waters is atmosphere derived

Gilfillan and Haszeldine (2011) observed no relationship between the ³He/⁴He ratio and HCO₃⁻ concentrations measured in the groundwaters from the Kerr site (Fig. 29). This is in stark contrast to the systematic relationship observed between HCO₃⁻ concentration and ³He/⁴He ratios in the St. Johns surface springs and groundwater wells (Sect. 7.3 and Fig. 23b). In the St. Johns Dome dataset there was a clear link between shallow samples and deep sources of CO₂ and He (Gilfillan et al. 2011). The majority of the St. Johns water samples exhibited higher HCO₃⁻ concentrations which corresponded to lower ³He/⁴He ratios and showed a clear trend towards the lower ³He/⁴He and higher HCO₃⁻ concentrations recorded in the deep CO₂ well water. In the Kerr groundwaters increased HCO₃⁻, indicating increasing concentrations of dissolved CO₂ did not result in any variation in the air like ³He/⁴He ratios.

Fig. 29

³He/⁴He ratios plotted against HCO₃⁻ concentrations in the produced reservoir fluids, groundwater wells and water injected into the reservoir surrounding the Kerr quarter. The black dash line indicates the air ³He/⁴He value of 1R_a. All of the groundwater wells have ³He/⁴He ratios which are close to the air ratio and exhibit no relationship between HCO₃⁻ concentrations and the ³He/⁴He ratio

There was also no relationship observed between ³He/⁴He ratios and the concentration of ⁴He in the Kerr groundwaters. This strongly implied that there is no addition of deep crustal ⁴He to the sampled waters surrounding the Kerr quarter and indicates that there was no migration of fluids from depth into the sampled wells. This was again in contrast to the clear relationship between decreasing ³He/⁴He and increasing ⁴He concentrations observed in the St. Johns groundwaters (Fig. 24a).

In summary, all of the noble gas concentrations and the majority of noble gas ratios measured in the groundwaters from wells surrounding the Kerr quarter were within or very close to the air saturated values. Gilfillan and Haszeldine (2011) found no evidence in any of the noble gas data that there was any presence of deep crustal derived noble gases within the groundwaters surrounding the Kerr quarter. Hence, they found no evidence of the migration of deep CO₂ from the Weyburn oil field into the groundwater on the Kerr quarter or the surrounding area (Fig. 30).

Fig. 30

³He/⁴He plotted against ⁴He concentrations in the produced reservoir fluids, groundwater wells and water injected into the reservoir surrounding the Kerr quarter. It is readily apparent that all of the groundwater samples exhibit ⁴He concentrations which are within the air saturated water concentrations and there is no correlation to the either the injected water, injected CO₂ or produced fluids from the Weyburn oil field

8 Summary and Future Work

Studies of noble gases in natural CO₂ storage analogues have proven to be extremely useful in establishing the viability of geological CO₂ storage. Noble gas data has shown that it is possible to store CO₂ for significant periods of time (up to 40 million years in the case of one reservoir). Measurements of noble gases and carbon stable isotopes have also permitted the identification of the main CO₂ storage processes that act in natural CO₂ reservoirs over geological timescales providing important lessons for engineered storage. These data indicate that the dominant mechanism of CO₂ loss is through dissolution of CO₂ into the formation water within a narrow pH window (pH = 5–5.8). Precipitation of carbonate minerals is only a minor sink, accounting for a maximum of 18 % of CO₂ loss in one of the CO₂ reservoirs.

For CO₂ storage technology to be universally deployed, it is essential that a robust, reliable and inexpensive means to trace the migration of CO₂ within the subsurface exists. Monitoring during injection will increase confidence that the site characteristics were correctly determined and met. Furthermore, should migration and subsequent surface leakage occur, the ability to track origin and ownership of CO₂ at near and ground surface will be critical for remediation purposes. However, one of the key necessities of any monitoring program will be the detection of any migration, however small, of CO₂ from the storage site. The main problem in detecting migration is the large variation in CO₂ concentrations within the Earth’s crust. This makes unequivocally determining if detected small quantities of CO₂ are the result of migration of CO₂ from a storage site extremely difficult. The use of tracers is the main technique that could be used to accurately determine if migration of CO₂ from a storage site has occurred.

Noble gases are chemically inert, persistent and environmentally safe—and possess significant potential as tracers of CO₂ migration within future storage sites. Their ability to trace CO₂ migration within a reservoir is illustrated by the study undertaken at the Mabee EOR field (Sect. 6.1). They have the potential to be extremely useful in tracing migration of CO₂ once it has migrated outside a reservoir as highlighted by the St. Johns Springs and Green River studies (Sects. 7.2 and 7.3), and also can play an important role in investigations of alleged leakage, as outlined by the Weyburn investigation (Sect. 7.4). Hence, there is enormous promise in the use of noble gases to monitor for any leakage of CO₂ or CO₂ rich groundwater from an engineered storage site and so provide a means of reassurance of the security of the storage site. This promise will only be fulfilled if more studies similar to those highlighted in this chapter are undertaken. It is imperative that the many upcoming pilot CO₂ injection studies continue to investigate the behaviour of noble gases in the subsurface and develop suitable noble gas monitoring strategies.

Whilst the aspirations for future CO₂ emissions reduction are in increasing energy efficiency and in developing alternative and renewable energy technologies, these cannot fulfil our current demands for energy. The amount of time that it will take to develop these technologies, and the fact that there are still plentiful supplies of fossil fuels in developing countries, means their combustion will remain the main source of global energy for a considerable time to come. Capturing and storing the CO₂ produced by fossil fuel combustion can drastically reduce CO₂ emissions and this technology can be operational very soon, given the right support. The lessons learned from noble gas studies show that CO₂ can be safely stored on geological timescales and that any migration from a storage site could be detected. CCS has the potential to act as a ‘bridge technology’ significantly reducing emissions whilst other low carbon energy technologies are developed and deployed. The fossil fuel power plants will be built anyway, as the world needs a reliable economic source of energy. Therefore, the only question to be answered is whether we are willing to build the bridge to capture and store CO₂ emissions from these plants, or prepare to face the climate change consequences.