Positron emission tomography dataset of [11C]carbon dioxide storage in coal for geo-sequestration application

Positron Emission Tomography (PET) imaging has demonstrated its capability in providing time-lapse fluid flow visualisation for improving the understanding of flow properties of geologic media. To investigate the process of CO2 geo-sequestration using PET imaging technology, [11C]CO2 is the most optimal and direct radiotracer. However, it has not been extensively used due to the short half-life of Carbon-11 (20.4 minutes). In this work, a novel laboratory protocol is developed to use [11C]CO2 as radiolabelled tracer to visualise and quantify in-situ CO2 adsorption, spreading, diffusion, and advection flow in coal. This protocol consists of generation and delivering of [11C]CO2, lab-based PET scanning, subsequent micro-CT scanning, and data processing. The lab-based PET scanning setup integrates in-situ core flooding tests with PET scanning. The real-time PET images are acquired under different storage conditions, including early gas production stage, depleted stage, and late storage stage. These datasets can be used to study across-scale theoretical and experimental study of CO2 flow behaviour in coal with the application to CO2 geo-sequestration.


Background & Summary
To achieve the net zero emission goal to mitigate the global warming issues, negative emission technologies must be deployed to remove carbon dioxide (CO 2 ) from air or point sources and store it.CO 2 geo-sequestration is one of the critical components in decarbonisation technologies for long-term and large-scale CO 2 storage 1 .Feasible CO 2 geological storage formations include depleted hydrocarbon reservoirs, saline aquifers, and coal seams.Compared with the former two options, CO 2 storage in coal seams has a unique adsorption trapping mechanism, as CO 2 can be permanently retained within coal seams in adsorbed state if the in-situ pressure is maintained.Therefore, CO 2 geo-sequestration in coal seams could be safer with high storage capacity 2 .Due to the complexity and heterogeneity of coal, CO 2 storage in coal seams is more complicated, which includes gas spreading, trapping, migrating, and interacting with coal matrix.These coupled physical phenomena are both spatially localised and time-dependant, which are yet thoroughly known.Thus, it is of importance to provide direct insights about the subsurface processes during CO 2 geo-sequestration.
Micro-Positron Emission Tomography (micro-PET) can create four-dimensional images of the radiotracer transport, which has been used in visualising real-time fluid flow in rock cores due to its high temporal (≈10 seconds) resolution [3][4][5][6] .[ 18 F]FDG (Fluorodeoxyglucose) is the most used aqueous radiotracer for solute transport study in porous media because of its availability and suitable radioisotope half-life (109 mins).However, for the application of CO 2 geo-sequestration, we need to directly visualise gaseous CO 2 flow dynamics in coal by using [ 11 C]CO 2 as the radiotracer.But radioisotope 11 C is rarely used for the study of fluid flow in porous media, as it is difficult to complete radiotracer delivery and injection within its short active lifetime (half-life = 20.4mins).The gaseous phase of [ 11 C]CO 2 further makes the radiotracer handling a challenge, only two studies are found from the literature 7,8 .Herein, for the first time, we use 11 C to trace CO 2 gas flow in coal using PET scanning, aiming to measure temporal and spatial evolution of CO 2 storage process in coal.The real-time PET images are acquired under different in-situ storage conditions, including early gas production stage, reservoir depleted stage, and late storage stage.In this data descriptor, the laboratory-based PET image acquisition method is described, including the radiotracer generation, delivery and handling, apparatus, and experimental design.All PET images and high-resolution micro-CT images of tested samples are publicly available on Figshare.Figure 1 shows a schematic overview of the workflow.This experimental protocol gives direct observation of CO 2 gas adsorption, spreading, diffusion, and advection flow in coal.These datasets could be reused by other researchers to validate modelling methods and to study in-situ CO 2 storage physics in coal for the application of CO 2 geo-sequestration.

Methods
[ 11 C]CO 2 generation, entrapment, and delivery.Carbon-11 ( 11 C), a radioisotope of carbon, has a half-life of 20.4 minutes.In this work, 11 C is prepared as [ 11 C]CO 2 using an on-site cyclotron ("Cyclone-18" fixed energy cyclotron, Ion Beam Applications, Belgium) at the Centre for Advanced Imaging, University of Queensland.Generated [ 11 C]CO 2 is automatically delivered at 500 mL/min via a mass flow controller (Synthra, GmbH) and trapped in a cryotrap at -195 °C within a shielded "hot-cell" (Tema Sinergie, Italy).The cryotrap is made of a coil of 1/8" OD stainless steel tubing with two ball valves at each end.Its internal volume is 3.64 mm 3 .A dose calibrator is installed within the hot-cell to measure the quantity of 11 C in the coil.After trapping [ 11 C] CO 2 from the cyclotron, the two ball values of the cryotrap are closed to secure the trapped [ 11 C]CO 2 .Next, the cryotrap coil, disconnected from the hot cell, is inserted into a shielded dewar containing liquid nitrogen, which is then delivered to the PET scanning lab via a specific sample lift (Fig. 2).

Laboratory-based PEt imaging setup. The experimental design integrates PET scanning with in-situ
core flooding setup which enables imaging and gas injecting at the same time (Fig. 2).The scanning instrument is Siemens Inveon micro-PET/CT Platform, where the ring has 120 cm internal diameter and is 12 cm long.The core holder is specifically designed for the Siemens Inveon PET scanner, such that it can be positioned inside the PET scanner ring.The upstream flowline connects the sample inlet with gas sources (CO 2 , He, CH 4 gas cylinders).Pressurised gas from the gas cylinders is pumped into the sample, measured at the outlet using a gas mass flowmeter, then collected at the end of the flowline in a shielded "Tedlar" gas capture bag (Grace Discovery Sciences, USA).A back-pressure regulator at the downstream is used to maintain a desired outlet pressure.
[ 11 C]CO 2 loading.A dual-position six-port injection valve is installed at the upstream to load [ 11 C]CO 2 into the injection flowline.Initially, the gas is injected into the core directly when the injection valve is under loading position.Meanwhile, the cryotrap coil that collects [ 11 C]CO 2 from the hot-cell is connected to the injection valve.For loading the radiotracer, the injection valve is switched to the loading position, such that gas from the cylinder will flow through the cryotrap coil to mix [ 11 C]CO 2 into the injection stream.
Samples and pre-treatments.Two coal samples and one Bentheimer sandstone core are studied in this work (D = 5.2 cm, L = 6 cm).One coal sample is relatively tight with fewer visible fractures, named "COAL1".The other coal sample is banded, where the bright band is highly fractured and percolates along the length, named "COAL2".Coal samples are pre-treated prior to CO 2 injection to mimic different underground conditions, including dry condition, dewatered condition, and gas-saturated condition.The different initial conditions of the tests could represent various storage stages.For instance, dewatered coal sample with residual CH 4 in place represents early-stage coal seam gas production condition, dewatered coal sample without residual CH 4 represents depleted reservoir condition, CO 2 -saturated coal sample represents the late-stage of CO 2 storage approaching storage capacity.The dewatered condition is achieved by injecting helium into a water-saturated sample to drain Fig. 1 Schematic overview of the workflow.This experimental protocol integrates PET scanning with in-situ core flooding, to visualise CO 2 storage process.[ 11 C]CO 2 is prepared at an on-site cyclotron and injected into three types of rocks as the radiotracer.PET image datasets and pressure data during core flooding tests are deposited publicly.In addition, micro-CT imaging technology is also applied for all three samples to obtain microscale pore/fracture features.the water phase until no water is produced from the outlet to reach the residual water saturation.To saturate the sample, CO 2 (or CH 4 ) is pumped into the dry sample under desired pressures for at least 12 hours.For the dewatered condition with residual CH 4 in place, we pump water into the CH 4 -saturated coal sample under the same gas pressure to keep the pore pressure constant, followed by releasing the outlet pressure to start dewatering and gas desorption.

Experiment procedures.
After the sample is pre-treated, carrier gas is injected under a constant gas pressure into the core under room temperature.When the pressures and outlet gas flowrate become stable, we load and inject [ 11 C]CO 2 and start PET scanning.In this work, there are 5 sets of core flooding tests, and 10 [ 11 C]CO 2 injections and corresponding PET scans.Table 1 lists details of each test, where gas flow rates, inlet and outlet pressures, confining pressure, and carrier gas are recorded.

PEt scanning and image construction.
PET scanning is performed simultaneously while injecting [ 11 C]CO 2 .The scanning instrument is a Siemens Inveon preclinical PET/CT scanner at The Centre for Advanced Imaging, University of Queensland.The effective spatial resolution for the scanner is calculated to be 1.6 mm, which is quantitatively assessed during the scanner calibration but using a Fludeoxyglucose ( 18 F) calibration source rather than a carbon-11 calibration source.The scanning data are reconstructed to create time lapse images, using OSEM2D (ordered subset expectation maximization 2D) method.Time frames can be less than 10 seconds, depending on the intensity of the radiotracer and scanner properties.In PET acquisition settings of this work, PET data are reconstructed at timeframes according to the gas breakthrough time, ranging from 2 s to 300 s.Table 2 lists the initial radio activity measured at the hot-cell and at the time of PET scanning.

technical Validation
All PET scans can visualise the CO 2 spreading, diffusing, adsorbing, flowing, and breaking through the core.Some selected PET images are shown in Fig. 3. Column (a -c) are time-lapse PET images of three different tests.The breakthrough time, CO 2 concentration and distribution differ between different tests, which is due to different initial sample conditions and sample types.Figure 3a shows results of the unsaturated and tight coal sample.In this case, CO 2 storage is a slower process, where CO 2 diffuses within the sample with a late gas breakthrough.While for unsaturated fractured coal sample (Fig. 3c), CO 2 breakthrough happens first, followed by gas adsorption and diffusion.On the other hand, if the fractured coal sample has been saturated initially, CO 2 gas quickly occupies fracture space and breaks through at the outlet (Fig. 3b).In our preliminary analysis, we estimate the  dispersion coefficients for selected scans by fitting the experimental PET data with modelling data derived from Advection-Dispersion Equation (ADE).Scan 8 and 9 are repeat scanning for CO 2 -saturated coal sample under the same conditions with 2 hours apart.They give similar estimated dispersion coefficients at the magnitude of 10 −9 cm 2 /min.More analysis and discussions could be found in our previous paper 13 , which could further validate the reliability of obtained image data.
for more details about data processing.The micro-CT images can be segmented and registered with corresponding PET images using software such as Thermo-Scientific Avizo, to obtain porosity and permeability map of each sample Table 4.

Fig. 2
Fig. 2 Schematic of the experimental setup.The cryotrap coil collects [ 11 C]CO 2 from the cyclotron.It is then transferred via a sample lift to the PET lab and is connected to the injection valve for loading (loading position).After loading, switch the valve to injecting position to start injecting [ 11 C]CO 2 into the core.The whole flowline is shielded.

Fig. 3
Fig. 3 Selected PET images of three tests.Column (a) is from unsaturated and tight coal sample, COAL1; (b) is for saturated and fractured coal sample, COAL2; (c) is for unsaturated and fracture coal sample, COAL2.Timelapsed PET images present CO 2 spreading, diffusing, adsorbing, flowing, and breaking through the core.

Table 1 .
Overview of the core flooding experiments conducted under different conditions.

Table 2 .
Initial radio activity and time framing of each PET scan.