*2.2. Establishing the Baseline*

Hydrogen is a very mobile gas that can leak towards the surface and accumulate in the groundwater and the soil. A complete monitoring protocol could interest the saturated zone, the unsaturated zone, the soil, and the surface because hydrogen can be detected in all these compartments as it can be seen in natural hydrogen emission areas [15]. However, the leakage simulation protocol (see under) is based on the injection of water saturated with dissolved gas (helium or hydrogen) directly into the aquifer. The water will be previously saturated at atmospheric conditions, i.e., at a pressure of 0.10 MPa, and then injected from 2 m to 11 m under the water table, where the hydrostatic pressure is between 0.12 and 0.21 MPa. Thus, the water will always remain undersaturated and nor helium nor hydrogen will degas. In these conditions, the only way for the dissolved gas to propagate is to follow the groundwater flow. In this study, the monitoring system has thus be designed for the saturated zone, with a light control of eventual weak degassing in the internal atmosphere of the piezometers but only for safety purposes.

Prior to setting up a monitoring system for the hydrogen injection test, a baseline of the initial piezometric, chemico-physical and hydrogeochemical values of the aquifer was established over 388 days starting on 27 October 2018. On all of the 7 main piezometers of the site (PZ1, PZ2, PZ2BIS, PZ3, PZ4, PZ5, and PZ6), the baseline corresponds to more than 200 measurements of each of the main chemico-physical parameters of the water: pH, temperature, electrical conductivity (EC), oxidation-reduction potential (ORP) and dissolved O<sup>2</sup> (Table 1). The water has a neutral pH (pH = 7.25), is moderately mineralized (EC <sup>=</sup> <sup>562</sup> <sup>µ</sup>S·cm−<sup>1</sup> ), and oxygenated (O<sup>2</sup> = 5.44 mg·L −1 ) due to its proximity to the soil surface, and is thus globally oxidative (ORP = +103 mV).

**Table 1.** Baseline of the chemico-physical parameters.


Legend: O<sup>2</sup> = dissolved oxygen; T = Temperature; EC = Electric Conductivity; ORP = Oxidation-Reduction Potential; SD = Standard Deviation.

These chemico-physical parameters are quite stable over space and time. Figure 3 represents the boxplots of the dissolved oxygen and the oxidation-reduction potential at each piezometer during the baseline. Figure 4 represents the evolution of these main chemico-physical parameters over time and seems to show a certain sensitivity to the depth of the aquifer, which varies from 13.06 m to 13.94 m (Figure 3). *Appl. Sci.* **2020**, *10*, x FOR PEER REVIEW 6 of 18 and seems to show a certain sensitivity to the depth of the aquifer, which varies from 13.06 m to 13.94 m (Figure 3).

**Figure 3.** Boxplots of dissolved oxygen concentration and oxidation-reduction potential along the experimental site during the baseline: (**a**) Dissolved oxygen concentration; (**b**) oxidation-reduction potential. The colored boxes represent the 1st and 3rd quartiles (respectively Q1 and Q3), the black line is the median line, the dots are the measured values, the whiskers are the upper and lower extreme limits calculated according to the Tukey's formula: Q1 + 1.5 (Q3 − Q1) and Q3 − 1.5 (Q3−Q1). **Figure 3.** Boxplots of dissolved oxygen concentration and oxidation-reduction potential along the experimental site during the baseline: (**a**) Dissolved oxygen concentration; (**b**) oxidation-reduction potential. The colored boxes represent the 1st and 3rd quartiles (respectively Q1 and Q3), the black line is the median line, the dots are the measured values, the whiskers are the upper and lower extreme limits calculated according to the Tukey's formula: Q1 + 1.5 (Q3 − Q1) and Q3 − 1.5 (Q3−Q1).

(**a**)

(**b**)

− Q1) where Q1 and Q3 are the 1st and 3rd quartiles.

**Figure 4.** Evolution of dissolved oxygen concentration and oxidation-reduction potential and their extreme limits during the establishment of the baseline: (**a**) Dissolved oxygen concentration; (**b**) Oxidation-reduction potential. The solid and dashed lines correspond respectively to the upper and lower extreme limits calculated according to the Tukey's formula: Q1 + 1.5 (Q3 − Q1) and Q3 − 1.5 (Q3 m (Figure 3).

and seems to show a certain sensitivity to the depth of the aquifer, which varies from 13.06 m to 13.94

(**a**) (**b**) **Figure 3.** Boxplots of dissolved oxygen concentration and oxidation-reduction potential along the experimental site during the baseline: (**a**) Dissolved oxygen concentration; (**b**) oxidation-reduction potential. The colored boxes represent the 1st and 3rd quartiles (respectively Q1 and Q3), the black

**Figure 4.** Evolution of dissolved oxygen concentration and oxidation-reduction potential and their extreme limits during the establishment of the baseline: (**a**) Dissolved oxygen concentration; (**b**) Oxidation-reduction potential. The solid and dashed lines correspond respectively to the upper and lower extreme limits calculated according to the Tukey's formula: Q1 + 1.5 (Q3 − Q1) and Q3 − 1.5 (Q3 − Q1) where Q1 and Q3 are the 1st and 3rd quartiles. **Figure 4.** Evolution of dissolved oxygen concentration and oxidation-reduction potential and their extreme limits during the establishment of the baseline: (**a**) Dissolved oxygen concentration; (**b**) Oxidation-reduction potential. The solid and dashed lines correspond respectively to the upper and lower extreme limits calculated according to the Tukey's formula: Q1 + 1.5 (Q3 − Q1) and Q3 − 1.5 (Q3 − Q1) where Q1 and Q3 are the 1st and 3rd quartiles.

During the acquisition of the baseline data, 94 water samples were taken to analyze the major ions (Ca2+, Mg2+, Na+, K+, HCO<sup>3</sup> <sup>−</sup>, Cl−, SO<sup>4</sup> <sup>2</sup>−, NO<sup>3</sup> −) and the main minor ions liable to be modified by a hydrogen-water-rock interaction (NO<sup>2</sup> <sup>−</sup>, NH<sup>4</sup> <sup>+</sup>, SO<sup>3</sup> <sup>2</sup>−, S<sup>2</sup> −, Fe, Mn). To comply with the storage conditions for all the elements, the analyses were carried out within 24 h of each sample being taken, using the methods presented in Table 2.

**Table 2.** Analytical methods and detection thresholds for the analyzed elements (mg·L −1 ).


Legend: ICP-MS = Inductively Coupled Plasma-Mass Spectrometry; IC = Ionic Chromatography; DL = Detection Limit.

Regarding the major elements, Table 3, Figures 5 and 6 show that their behavior is also very stable throughout the baselining. The water generally exhibits bicarbonate-calcium facies, characteristic of using the methods presented in Table 2.

using the methods presented in Table 2.

of nitrate ions from agricultural inputs.

of nitrate ions from agricultural inputs.

chalk waters. This dominant hydrochemical facies is slightly altered by the presence of nitrate ions from agricultural inputs. Regarding the major elements, Table 3, Figures 5 and 6 show that their behavior is also very stable throughout the baselining. The water generally exhibits bicarbonate-calcium facies, characteristic of chalk waters. This dominant hydrochemical facies is slightly altered by the presence Regarding the major elements, Table 3, Figures 5 and 6 show that their behavior is also very stable throughout the baselining. The water generally exhibits bicarbonate-calcium facies, characteristic of chalk waters. This dominant hydrochemical facies is slightly altered by the presence

*Appl. Sci.* **2020**, *10*, x FOR PEER REVIEW 7 of 18

**Table 2.** Analytical methods and detection thresholds for the analyzed elements (mg·L<sup>−</sup>1). **Parameter HCO3<sup>−</sup> Ca2+ Mg2+ Na+ K+ Cl− SO42<sup>−</sup> NO3<sup>−</sup> NO2<sup>−</sup> NH4+ SO32− S2<sup>−</sup> Fe Mn**  Method Titration Ionic Chromatography ICP-MS DL (mg·L<sup>−</sup>1) 0.04 0.05 0.01 0.02 0.02 0.01 0.001

**Table 2.** Analytical methods and detection thresholds for the analyzed elements (mg·L<sup>−</sup>1). **Parameter HCO3<sup>−</sup> Ca2+ Mg2+ Na+ K+ Cl− SO42<sup>−</sup> NO3<sup>−</sup> NO2<sup>−</sup> NH4+ SO32− S2<sup>−</sup> Fe Mn**  Method Titration Ionic Chromatography ICP-MS DL (mg·L<sup>−</sup>1) 0.04 0.05 0.01 0.02 0.02 0.01 0.001 Legend: ICP-MS = Inductively Coupled Plasma-Mass Spectrometry; IC = Ionic Chromatography; DL

During the acquisition of the baseline data, 94 water samples were taken to analyze the major ions (Ca2+, Mg2+, Na+, K+, HCO3−, Cl−, SO42−, NO3−) and the main minor ions liable to be modified by a hydrogen-water-rock interaction (NO2−, NH4+, SO32−, S2−, Fe, Mn). To comply with the storage conditions for all the elements, the analyses were carried out within 24 h of each sample being taken,

During the acquisition of the baseline data, 94 water samples were taken to analyze the major ions (Ca2+, Mg2+, Na+, K+, HCO3−, Cl−, SO42−, NO3−) and the main minor ions liable to be modified by a hydrogen-water-rock interaction (NO2−, NH4+, SO32−, S2−, Fe, Mn). To comply with the storage conditions for all the elements, the analyses were carried out within 24 h of each sample being taken,

*Appl. Sci.* **2020**, *10*, x FOR PEER REVIEW 7 of 18


Legend: SD = Standard Deviation.

SD 7.4 0.8 1.0 0.23 10 2.3 2.8 2.5 7.4 Legend: SD = Standard Deviation.

SD 7.4 0.8 1.0 0.23 10 2.3 2.8 2.5 7.4 Legend: SD = Standard Deviation.

−1 ).

**Figure 5.** Boxplots of major and minor ions concentration along the experimental site during the baseline: (**a**) major elements; (**b**) minor elements. The colored boxes represent the 1st and 3rd quartiles (respectively Q1 and Q3), the black line is the median line, the dots are the measured values, the whiskers are the upper and lower extreme limits calculated according to the Tukey's formula: Q1 + 1.5 (Q3 − Q1) and Q3 − 1.5 (Q3 − Q1). **Figure 5.** Boxplots of major and minor ions concentration along the experimental site during the baseline: (**a**) major elements; (**b**) minor elements. The colored boxes represent the 1st and 3rd quartiles (respectively Q1 and Q3), the black line is the median line, the dots are the measured values, the whiskers are the upper and lower extreme limits calculated according to the Tukey's formula: Q1 + 1.5 (Q3 − Q1) and Q3 − 1.5 (Q3 − Q1). **Figure 5.** Boxplots of major and minor ions concentration along the experimental site during the baseline: (**a**) major elements; (**b**) minor elements. The colored boxes represent the 1st and 3rd quartiles (respectively Q1 and Q3), the black line is the median line, the dots are the measured values, the whiskers are the upper and lower extreme limits calculated according to the Tukey's formula: Q1 + 1.5 (Q3 − Q1) and Q3 − 1.5 (Q3 − Q1).

**Figure 6.** *Cont.*

*Appl. Sci.* **2020**, *10*, x FOR PEER REVIEW 8 of 18

**Figure 6.** Evolution of the cumulative concentrations of major and minor ions and their extreme limits during the establishment of the baseline in 2019: (**a**) Sum of major ions; (**b**) Sum of minor ions. The solid and dashed lines correspond respectively to the upper and lower extreme limits calculated according to the Tukey's formula: Q1 − 1.5 (Q3 − Q1) and Q3 + 1.5 (Q3 − Q1) where Q1 and Q3 are the 1st and 3rd quartiles. **Figure 6.** Evolution of the cumulative concentrations of major and minor ions and their extreme limits during the establishment of the baseline in 2019: (**a**) Sum of major ions; (**b**) Sum of minor ions. The solid and dashed lines correspond respectively to the upper and lower extreme limits calculated according to the Tukey's formula: Q1 − 1.5 (Q3 − Q1) and Q3 + 1.5 (Q3 − Q1) where Q1 and Q3 are the 1st and 3rd quartiles.

Regarding the minor elements analyzed, Table 4 shows the absence of nitrite and sulfide ions above the detection thresholds, as well the absence of sulfide ions except in five samples taken at PZ2 in the first half of 2019 where the concentrations ranged from 0.03 to 0.11 mg.L−1. Ammonium ions are present in 74% of the samples, probably related to the application of nitrogenous fertilizers nearby, with a fairly fluctuating concentration with an average of 0.10 mg.L−1. The totals of dissolved iron and manganese were also analyzed, but they were only minutely present in the water due to the mineralogical composition of the aquifer rock, which is made up of more than 95% calcite [4]: their ionized forms were therefore not researched. Ultimately, the evolution in the total of these minor ions varied little during baseline monitoring, the fluctuations observed being mainly due to the ammonium ions (Table 4). Regarding the minor elements analyzed, Table 4 shows the absence of nitrite and sulfide ions above the detection thresholds, as well the absence of sulfide ions except in five samples taken at PZ2 in the first half of 2019 where the concentrations ranged from 0.03 to 0.11 mg·L −1 . Ammonium ions are present in 74% of the samples, probably related to the application of nitrogenous fertilizers nearby, with a fairly fluctuating concentration with an average of 0.10 mg·L −1 . The totals of dissolved iron and manganese were also analyzed, but they were only minutely present in the water due to the mineralogical composition of the aquifer rock, which is made up of more than 95% calcite [4]: their ionized forms were therefore not researched. Ultimately, the evolution in the total of these minor ions varied little during baseline monitoring, the fluctuations observed being mainly due to the ammonium ions (Table 4).


**Table 4.** Main characteristics of minor ions analyzed during baselining. **Table 4.** Main characteristics of minor ions analyzed during baselining.

Legend: DL = Detection Limit; SD = Standard Deviation; CV = Coefficient of Variation; Total (N+S) = Total of sulfur and nitrogen ions. Legend: DL = Detection Limit; SD = Standard Deviation; CV = Coefficient of Variation; Total (N+S) = Total of sulfur and nitrogen ions.

## *2.3. Preparing the Test 2.3. Preparing the Test*

The helium was injected with the aim of testing and optimizing a future hydrogen injection device using an inert gas, and to configure the monitoring protocol (types of measurement and time intervals) depending on the piezometer being monitored. The objective of this test is to create a plume of dissolved helium in the aquifer, comparable to the future plume of dissolved hydrogen, and to The helium was injected with the aim of testing and optimizing a future hydrogen injection device using an inert gas, and to configure the monitoring protocol (types of measurement and time intervals) depending on the piezometer being monitored. The objective of this test is to create a plume of dissolved helium in the aquifer, comparable to the future plume of dissolved hydrogen, and to monitor its propagation in the saturated zone.

monitor its propagation in the saturated zone. Before this test, the propagation of the dissolved He plume was modeled in 1D using PHREEQC. Modeling parameters were determined using the results of previous CO2 injection tests [4,5]. The result is shown in Figure 7 and shows a maximum dissolved helium concentration between 1.46 mg.L−1 and 8 × 10−21 mg.L−1 from PZ2BIS to PZ6, and a peak arrival time between 100 min and 23 days. Peak values at PZ5 and PZ6 are expected to be below 1 µg·L−1 and thus it will not be possible to detect helium in these two piezometers. Before this test, the propagation of the dissolved He plume was modeled in 1D using PHREEQC. Modeling parameters were determined using the results of previous CO<sup>2</sup> injection tests [4,5]. The result is shown in Figure 7 and shows a maximum dissolved helium concentration between 1.46 mg·L <sup>−</sup><sup>1</sup> and <sup>8</sup> <sup>×</sup> <sup>10</sup>−<sup>21</sup> mg·<sup>L</sup> −1 from PZ2BIS to PZ6, and a peak arrival time between 100 min and 23 days. Peak values at PZ5 and PZ6 are expected to be below 1 µg·L <sup>−</sup><sup>1</sup> and thus it will not be possible to detect helium in these two piezometers.

**Figure 7.** Model of the propagation of the dissolved He plume using PHREEQC. **Figure 7.** Model of the propagation of the dissolved He plume using PHREEQC. **Figure 7.** Model of the propagation of the dissolved He plume using PHREEQC.

Then, the water from the aquifer was extracted beforehand by pumping in the PZ2 piezometer (future injection well) to fill two HDPE tanks (Figure 8a): a first 1 m3 tank which contains the tracers to help determine the arrival of the plume of dissolved gas and precisely quantify its kinetics, and a second 5 m3 tank in which the water will be saturated with helium by bubbling it. It was decided not to incorporate the tracers in the tank of water saturated with helium to avoid the risk of reducing the solubility of this gas. Then, the water from the aquifer was extracted beforehand by pumping in the PZ2 piezometer (future injection well) to fill two HDPE tanks (Figure 8a): a first 1 m<sup>3</sup> tank which contains the tracers to help determine the arrival of the plume of dissolved gas and precisely quantify its kinetics, and a second 5 m<sup>3</sup> tank in which the water will be saturated with helium by bubbling it. It was decided not to incorporate the tracers in the tank of water saturated with helium to avoid the risk of reducing the solubility of this gas. Then, the water from the aquifer was extracted beforehand by pumping in the PZ2 piezometer (future injection well) to fill two HDPE tanks (Figure 8a): a first 1 m3 tank which contains the tracers to help determine the arrival of the plume of dissolved gas and precisely quantify its kinetics, and a second 5 m3 tank in which the water will be saturated with helium by bubbling it. It was decided not to incorporate the tracers in the tank of water saturated with helium to avoid the risk of reducing the solubility of this gas.

**Figure 8.** Views of the two tanks and the helium saturation device: (**a**) View of the 1 m3 tracers tank [A], the 5 m3 He saturated tank [B] and the compressed He cylinder [C] ; (**b**) View from the manhole of the bubbling device in the 5 m3 He tank (20 m of PVC pipe pierced with 200 needles of 0.5 mm **Figure 8.** Views of the two tanks and the helium saturation device: (**a**) View of the 1 m3 tracers tank [A], the 5 m3 He saturated tank [B] and the compressed He cylinder [C] ; (**b**) View from the manhole of the bubbling device in the 5 m3 He tank (20 m of PVC pipe pierced with 200 needles of 0.5 mm **Figure 8.** Views of the two tanks and the helium saturation device: (**a**) View of the 1 m<sup>3</sup> tracers tank [A], the 5 m<sup>3</sup> He saturated tank [B] and the compressed He cylinder [C]; (**b**) View from the manhole of the bubbling device in the 5 m<sup>3</sup> He tank (20 m of PVC pipe pierced with 200 needles of 0.5 mm diameter).

diameter). In the first 1 m<sup>3</sup> tank, two types of tracers were used:

diameter).

In the first 1 m3 tank, two types of tracers were used: • fluorescent organic tracers that exhibit no analytical interference between them, to allow in situ detection in real-time of the arrival of the injected plume thanks to the installation of a GGUN FL-30 field fluorimeter; these are uranine or fluorescein sodium (green dye), sulforhodamine B (red dye) and Amino G Acid (a colorless tracer emitting in blue); however, previous experiments In the first 1 m3 tank, two types of tracers were used: • fluorescent organic tracers that exhibit no analytical interference between them, to allow in situ detection in real-time of the arrival of the injected plume thanks to the installation of a GGUN FL-30 field fluorimeter; these are uranine or fluorescein sodium (green dye), sulforhodamine B (red dye) and Amino G Acid (a colorless tracer emitting in blue); however, previous experiments • fluorescent organic tracers that exhibit no analytical interference between them, to allow in situ detection in real-time of the arrival of the injected plume thanks to the installation of a GGUN FL-30 field fluorimeter; these are uranine or fluorescein sodium (green dye), sulforhodamine B (red dye) and Amino G Acid (a colorless tracer emitting in blue); however, previous experiments with these types of organic tracers with a long carbonaceous molecule (C<sup>20</sup> to C40) have shown

with these types of organic tracers with a long carbonaceous molecule (C20 to C40) have shown

with these types of organic tracers with a long carbonaceous molecule (C20 to C40) have shown

that not all of them were conservative when transferred to an aquifer composed of chalk with finely porous matrix permeability, as is the case in Catenoy [5]; that not all of them were conservative when transferred to an aquifer composed of chalk with

*Appl. Sci.* **2020**, *10*, x FOR PEER REVIEW 10 of 18

• inorganic ionic tracers, which are highly conservative but colorless; they are analyzed a posteriori in the laboratory, from a water sample to precisely quantify the kinetics of the plume; these are lithium (as lithium chloride, LiCl) and bromide (as potassium bromide, KBr) finely porous matrix permeability, as is the case in Catenoy [5]; • inorganic ionic tracers, which are highly conservative but colorless; they are analyzed a posteriori in the laboratory, from a water sample to precisely quantify the kinetics of the plume; these are lithium (as lithium chloride, LiCl) and bromide (as potassium bromide, KBr)

For the final hydrogen injection experiment, only the most efficient fluorescent tracer and ionic tracer in terms of their recovery will be used. The objective of this first test is, therefore, both to select these two tracers from the five tested, and to validate the principle of a prior injection of tracers to predict the arrival of the dissolved gas plume and, as a result, to improve the monitoring system. A quantity of 1 g of each tracer was diluted in the 1 m<sup>3</sup> tank: it will be noted that, for ionic tracers, this is 1 g of tracer ion (Li<sup>+</sup> and Br−), which corresponds to 6.14 g of LiCl and 1.49 g of KBr. For the final hydrogen injection experiment, only the most efficient fluorescent tracer and ionic tracer in terms of their recovery will be used. The objective of this first test is, therefore, both to select these two tracers from the five tested, and to validate the principle of a prior injection of tracers to predict the arrival of the dissolved gas plume and, as a result, to improve the monitoring system. A quantity of 1 g of each tracer was diluted in the 1 m3 tank: it will be noted that, for ionic tracers, this is 1 g of tracer ion (Li+ and Br−), which corresponds to 6.14 g of LiCl and 1.49 g of KBr.

To saturate the water with helium, a bubbling device was installed on the interior floor of the second tank to create a curtain of bubbles facilitating the dissolution of the gas. It is a PVC pipe of 20 m length, pierced with 200 holes (Figure 8b). This device is connected in a loop to a rotameter to regulate the flow of gas injected from a compressed gas cylinder (Figure 8a). To saturate the water with helium, a bubbling device was installed on the interior floor of the second tank to create a curtain of bubbles facilitating the dissolution of the gas. It is a PVC pipe of 20 m length, pierced with 200 holes (Figure 8b). This device is connected in a loop to a rotameter to

The two tanks were then emptied successively by gravity into the PZ2 borehole and the water then entered the aquifer. regulate the flow of gas injected from a compressed gas cylinder (Figure 8a). The two tanks were then emptied successively by gravity into the PZ2 borehole and the water then entered the aquifer.

The injection device consists of a reinforced PVC pipe 70 mm in diameter. This pipe is directly connected, by means of a tee fitting, to the outlet located at the base of the two tanks, each being isolated by a valve (Figure 9a). The injection pipe is plugged at its lower end and ballasted to ensure that it descends to the bottom of the well, at a depth of 25 m. To better distribute the injected fluid over the entire screened height in the injection well, the submerged part of the injection pipe is drilled with 46 holes that are distributed two by two from 12 to 23 m deep, and four by four from 23 to 25 m deep (Figure 9b). The shallowest holes are located 0.2 m below the water table to ensure that the dissolved helium is injected under a slight hydrostatic overpressure, and therefore cannot degas. The injection device consists of a reinforced PVC pipe 70 mm in diameter. This pipe is directly connected, by means of a tee fitting, to the outlet located at the base of the two tanks, each being isolated by a valve (Figure 9a). The injection pipe is plugged at its lower end and ballasted to ensure that it descends to the bottom of the well, at a depth of 25 m. To better distribute the injected fluid over the entire screened height in the injection well, the submerged part of the injection pipe is drilled with 46 holes that are distributed two by two from 12 to 23 m deep, and four by four from 23 to 25 m deep (Figure 9b). The shallowest holes are located 0.2 m below the water table to ensure that the dissolved helium is injected under a slight hydrostatic overpressure, and therefore cannot degas.

**Figure 9.** View of injection installation: (**a**) injection pipe and its connection to the tanks; (**b**) diagram of the arrangement of the holes along the injection pipe. **Figure 9.** View of injection installation: (**a**) injection pipe and its connection to the tanks; (**b**) diagram of the arrangement of the holes along the injection pipe.

### *2.4. Conducting the Test 2.4. Conducting the Test*

The tanks were filled with groundwater on the morning of 1 April 2019 using a submerged electric pump installed in PZ2 piezometer, the future injection well: The tanks were filled with groundwater on the morning of 1 April 2019 using a submerged electric pump installed in PZ2 piezometer, the future injection well:


The injection of the water and tracers from the first tank (1 m<sup>3</sup> ) was performed by gravity on the 2nd of April 2019 from 10:35 to 11:10, which represents an injection flowrate of 1.7 m<sup>3</sup> ·h −1 . The helium-saturated water from the second tank (5 m<sup>3</sup> ) was injected again by gravity immediately after, i.e., from 11:12 to 12:47. This injection lasted 1 h 35 min, which corresponds to a flow rate of 3.2 m<sup>3</sup> ·h −1 .
