Introduction

For a long time, experts in earthquake forecasting have sought to deeply understand how various anomalies observed before earthquakes, such as underground fluids, geoelectricity, radon, atmospheric electrical field, ionosphere and others, interact with each other1,2,3,4,5,6,7,8,9,10. Therefore, Pulinets and Hayakawa respectively proposed the Lithosphere-Atmosphere-Ionosphere Coupling (LAIC) model, which suggests that the porosity, microfissures, and pressure of rocks near the future epicenter will change. These changes will mainly couple with the atmosphere and ionosphere through three channels: geochemistry channel, acoustic gravity wave channel, and electromagnetic channel11,12,13,14,15,16.

At present, there is a significant amount of observational data supporting the LAIC model, particularly from the MVP-LAI (monitoring vibrations and perturbations in the lithosphere, atmosphere and ionosphere) system in Sichuan, China17,18,19. Through various observation methods, this station observed the entire process of the propagation of earthquake precursor signals from the lithosphere to the ionosphere for the first time globally, validating the rationality of the LAIC model. Many scholars have also supplemented and improved the model based on observational data20.

We mainly focus on the geochemistry channels. In the geochemistry channel, the common view is that there will be a large release of the radioactive gas, radon, before the earthquake12. The decay of radon will cause a large amount of positive charges to accumulate on the surface, which will cause the vertical downward atmospheric electrical field to become vertical upward, even reaching 12 kV/m21,23. The decay of radon and its progeny will also change atmospheric electric conductivity (AEC). Changes in the atmospheric electrical field and AEC together lead to the generation of abnormal currents, affecting the ionosphere24. Our previous simulation results show that changes in the atmospheric electrical field can affect the ionospheric electrical field, ionospheric current, and plasma density, but these changes are still approximately three orders of magnitude smaller than the observed data4. Following Ohm’s law:

$$J=\sigma E$$
(1)

Here, J is the atmospheric current, \(\sigma\) is AEC, E is the atmospheric electrical field. Therefore, we believe that changes in AEC will also cause changes in atmospheric currents, thereby affecting the ionosphere. Our simulations have verified this well25. In-depth research on AEC and its changing factors will help to better carry out monitoring of earthquake precursors, understand the impact of earthquake precursors on LAIC, and thus contribute to disaster prevention and mitigation.

Although AEC is important for the study of earthquakes and LAIC, for a long time, AEC has not received enough attention in the field of earthquakes, except for some observations by meteorologists. Since Gerdien26 invented the atmospheric electric conductivity meter (AECM) in 1905, some scholars have conducted simple observations at heights of 0–80 km through hot air balloons or rocket-borne at various locations, obtaining vertical profiles of AEC27,28,29,30,31. Others have conducted route surveys of AEC above the ocean surface by steamer, preliminarily obtaining differences in AEC between ocean and land32,33,34,35. There are even fewer records of continuous observations at fixed points. Some observations were conducted in India, and the preliminary results of these observations suggest that AEC may be positively correlated with humidity and negatively correlated with air temperature, possibly due to the thermal motion of molecules36,37,38.

All of these studies were conducted by meteorologists, who primarily focused on meteorological processes. Their observations were either profile or route measurements, or short-term observations. Seismological observations, on the other hand, require long-term (several years) observations at fixed locations. Clearly, meteorological data are insufficient for earthquake research.

In addition, some scholars have also conducted comparative observations on radon and AEC, and they believe that their results support the view that there is a positive correlation between radon and AEC36,37,39. However, these observations were conducted in natural environments. The measurement results were greatly affected by the environment, and the radon concentration was very low (only about 20 Bq/m3), which was not significantly different from the environmental value. According to the report from the United Nations Scientific Committee on the Effects of Atomic Radiation, the environmental value is about 10 Bq/m340. Therefore, this is not sufficient to support their conclusion.

Seran28 conducted an experimental verification of the relationship between radon and AEC, their experiment was based on dry air and dry carbon dioxide, which is significantly different from the natural environment before the earthquake. In addition, their instrument has a limited range (10−13 ~ 10−9 S/m) and it is uncertain whether it can cover the range of conductivity changes caused by earthquake gases. Seran’s experiment inspired us to simulate the natural environment before an earthquake in the laboratory to study the changes in AEC.

In summary, the observation and research on the relationship between radon released before an earthquake and AEC are still insufficient, which affects our understanding of the propagation and coupling process of seismic disturbance signals in the lithosphere-atmosphere-ionosphere, especially the impact of radon and seismic gases (abnormally increased CO2, CH4, and other gases, before earthquakes) on AEC. To study the impact of radon and seismic gases on AEC, we developed this instrument based on the Gerdien sensor and conducted a preliminary experiment on radon and AEC.

This article is structured in a five section sequence. The first section briefly introduces the research history of AEC. The second section introduces the AECM we developed and the instruments used for the experiment. The third section outlines our experimental project. The fourth section presents the results and analysis. Finally, we discuss our plan for future work.

Instrument and experimental environment

In this section, we introduce the instrumentations used in the experiment. For more detailed design basis, process and indicators of the atmospheric electric conductivity meter we developed, please refer to our other article41.

Introduction of atmospheric electric conductivity meter

Our AECM adopts an integrated design, with all components installed in the shielding case to ensure that the measurement results are not affected by changes in the external environment. The AECM is powered by 220 V, 50 Hz AC. It can work immediately when powered on and needs to be reliably grounded. Table 1 shows some basic parameters of the AECM:

Table 1 Some parameters of the AECM.
Full size table

The AECM consists of three main parts. The first part is the Gerdien capacitor, which consists of two coaxial electrodes and a fan. The electrodes are referred to as the inner electrode and outer electrode, respectively. The fan’s function is to facilitate air containing the charged particles to flow through the Gerdien capacitor.

The second part is the weak current detection module, whose main function is to amplify the current output from the Gerdien capacitor to a measurable range. When a voltage difference is applied between the inner and outer electrodes, an electrical field is generated, deflecting charged particles in the air toward the electrodes, generating a weak current. This current is converted into a detectable voltage through a 10 GΩ resistor, and the conductivity is measured by measuring the voltage across the resistor.

The third part is the control module, which is responsible for controlling and storing the bias voltage, as well as collecting, calculating, storing, and transmitting the output from the weak current detection module. Currently, data transmission is done using WiFi and 4G, with a reserved network cable interface. Figure 1 is a schematic of the AECM’s principle and composition.

Fig. 1
figure 1

Schematic diagram of the AECM’s composition. The red and green circles represent positive and negative charges, respectively. The instrument inputs 220 V, 50 Hz AC power, which is converted into DC power (green line) through the power converter to power each module. The yellow circles represent the current detection module (A) and the bias voltage module (U). The red line provides direct current for the outer electrode, which can be converted from − 5 V to + 5 V, but during the testing phase we set it to 5 V. The GND of the current detection module and inner electrode are connected to the ground through the blue line. The control module controls and transmits data to the bias voltage module and weak current acquisition module through the communication line (purple line).

Full size image

The main principle of measuring AEC is that when charged particles are drawn into the capacitor, under the action of the bias voltage, the charged particles are deflected onto the electrodes, causing a change in the voltage between the inner and outer electrodes, which in turn leads to a change in the weak current26. Since the current output from the Gerdien capacitor is linearly related to AEC, the change in AEC can be determined by measuring the change in current. Since the output current is too weak (pA), we connected a 10 GΩ resistor in series and measured the voltage instead of the current.

The conversion relationship between the measurement results and AEC is41:

$$\sigma =\frac{{\ln \left( {{{{r_o}} \mathord{\left/ {\vphantom {{{r_o}} {{r_i}}}} \right. \kern-0pt} {{r_i}}}} \right)I}}{{2\pi l{V_i}}}=\frac{{\ln \left( {{{{r_o}} \mathord{\left/ {\vphantom {{{r_o}} {{r_i}}}} \right. \kern-0pt} {{r_i}}}} \right){V_o}}}{{2\pi l{V_i}R}}$$
(2)

Here, σ is AEC. ro is the outer electrode radius, 0.10 m. ri is the inner electrode radius, 0.01 m. l is the inner electrode length, 1 m. I is the current. Vi is the voltage between the inner and outer electrodes, 5 V. Vo is the measured voltage. R is the resistance, 10 GΩ41.

After substituting the above parameters into formula (2), the measured value of AEC is obtained41:

$$\sigma =\frac{{0.3665}}{5} \times {10^{ - 10}}{V_o}$$
(3)

The output of the AECM is voltage, which we convert to conductivity by formula (3).

Introduction of radon detector

The radon detector used in the experiment is AlphaGuard PQ2000Pro, which is widely used in the field of radon monitoring due to its wide measurement range, high detection efficiency for radon, fast response to concentration gradients, strong environmental adaptability, and maintenance-free long-term operation. Additionally, in extreme air humidity and temperature, it delivers reliable measuring values and is insensitive to vibrations and shock42. This instrument is also commonly used in China’s earthquake radon anomaly monitoring and is a standard instrument in the Radon Laboratory (more details about the laboratory are in 2.3).

AlphaGuard PQ2000Pro is based on the proven principle of pulse ionization chamber (alpha spectroscopy). As radon gas decays in the pulse ionization chamber, alpha particles are emitted and drift towards the collection electrode under the influence of an electrical field, creating a pulse current. The number of pulse currents is proportional to the number of alpha particles, and therefore, proportional to the radon concentration.

Some parameters of AlphaGuard PQ2000Pro are shown in Table 2:

Table 2 Some parameters of AlphaGuard PQ2000Pro.
Full size table

Introduction of radon sources and radon chambers

Our experiments were conducted at the Observatory for Geodynamic of the East Yangtze Block in Jiujiang, Jiangxi Province. The observatory is an important earthquake monitoring point managed by the China Earthquake Administration, close to Mount Lu. There are many active faults near the observatory, and the spring water in the station contains high concentrations of radon. To obtain gaseous 222Rn, spring water is introduced into the radon chamber system and degassed. The radon chamber system is equipped with a temperature and humidity control system that can ensure the temperature and humidity in the radon chamber are stable43.

Due to the large volume of the radon chamber system (200 m3), it takes a long time to reach a stable radon concentration and adjust the concentration. To address this issue, a small mobile radon chamber was specially designed. The small radon chamber (SRC) is cylindrical with a volume of only 1 m3 and is made of transparent organic plastic, allowing easy observation of the conditions inside from the outside. The SRC is equipped with an air pressure valve connected to the radon chamber system, enabling quick adjustment and balancing of the radon concentration. Additionally, air or other gases can be injected into the SRC through other valves to achieve various experimental purposes. However, since the temperature and humidity of the SRC cannot be controlled after the flow from the large radon chamber to the SRC, some fluctuations occurred. Fortunately, since the SRC is completely isolated from the outside, the absolute humidity remains constant. Furthermore, the laboratory’s air conditioning system remains on at all times, limiting fluctuations in temperature and humidity within the SRC.

Figure 2 is a photo from our experiment. We placed the AECM, AlphaGuard PQ2000Pro, and a small oscillating fan (used to accelerate 222Rn diffusion and mixing with air) together in the SRC. After confirming that the AECM was working properly, we closed the door and all valves of the SRC to isolate it from the outside.

Fig. 2
figure 2

Experimental scene and instrument placement.

Full size image

Experimental schemes of atmospheric electric conductivity meter

Since this is the first time that the AECM is being tested in a radon environment, we have designed three main steps to obtain as much experimental data as possible. Figure 3 is a flowchart of the experiment.

Fig. 3
figure 3

The Flowchart of the Experiment.

Full size image

The following is a detailed description of the experimental steps:

  1. (i)

    Since some radioactive substances will inevitably remain in the radon chamber, it is necessary to first measure the background value of AEC in the SRC. To obtain the background value, we placed the AECM in the SRC and measured it continuously for about 4 h (from 10:40 to 14:30, September 5, 2023) without injecting radon gas. The average value is 1.38 × 10–13 S/m, and the standard deviation is 1.81 × 10–14 S/m.

  2. (ii)

    To test whether the AECM responded to radon, we first conducted a preliminary experiment. In this experiment, 222Rn was injected into the SRC at a rate of 2 L/min. Once the concentration reached 50,000 Bq/m3, we stopped injecting 222Rn and started injecting air, causing the radon concentration to quickly drop to the background level.

The radon concentration in the SRC dropped quickly from 50,000 Bq/m3 to 1,000 Bq/m3. However, due to radioactive material residues, measurement errors of the AlphaGuard PQ2000Pro, gas unevenness, and other factors, the time required significantly increases for the radon concentration to drop from 1,000 Bq/m3 to 0 Bq/m3. Since the safe range of radon concentration is below 1,000 Bq/m3, we opened the SRC at 700 Bq/m3 and extracted the results of the preliminary experiment.

In the preliminary experiment, when the radon concentration increased from about 0 Bq/m3 to 50,000 Bq/m3 and then dropped back to about 0 Bq/m3, AEC increased from a background value of 1.38 × 10−13 S/m to about 90 × 10−13 S/m, and then dropped back to about 1.38 × 10−13 S/m. As shown in Fig. 4.a, there is a clear correlation between AEC and 222Rn. The Pearson’s correlation coefficient is 0.9925, more details are in Table 3. Subsequently, we began the formal experiment.

Table 3 Bivariate correlation analysis.
Full size table
Fig. 4
figure 4

Preliminary Experiment on AEC and 222Rn.In subfigure a, the left axis represents AEC (blue line), and the right axis represents radon concentration (green line). In subfigure b, the left axis represents temperature (red line) and relative humidity (blue line), while the right axis represents air pressure (green line).

Full size image

It should be noted that due to our setup errors, the time of the AECM and the radon detector deviated from the standard time. Fortunately, their deviations were fixed, and we corrected the discrepancies based on the experimental records and pictures.

  1. (iii)

    The formal experiment lasted for 30 days and was divided into three major stages. In the first stage, 222Rn was injected into the SRC at a speed of 2 L/min to quickly raise the radon concentration to 50,000 Bq/m3. Since the formal experiment required a stable radon concentration, we injected air and 222Rn into the SRC together; and ensured that the radon concentration was in dynamic stabilization by adjusting the injection rates of air and 222Rn.

After maintaining the radon concentration at around 50,000 Bq/m3 for 7 days, we dropped it to 20,000 Bq/m3 and maintained it for 7 days by increasing the injection rate of air and decreasing the injection rate of 222Rn. Finally, we dropped it to 15,000 Bq/m3 and maintained it for 7 days. After completing the three concentration experiments, we dropped it below 700 Bq/m3, and stopped the experiment.

The above concentrations are based on the concentration index and concentration gradient used by the observatory laboratory to calibrate other radon detectors. When designing this index, they took into account the needs of earthquake monitoring and instrument calibration43.

Through long-term comparison during three different concentration stages and adjustment stages, we obtained the performance parameters of the AECM, which provide a basis for field observations and other experiments.

Data analysis

Discrete point analysis and processing.

Figure 5 shows the comparison between 222Rn and AEC in the formal experiment. Obviously, there is remarkable consistency between them. However, we also noticed some discrete points in the data after September 10th. To determine the reason for the appearance of these discrete points, we conducted an auxiliary experiment.

Fig. 5
figure 5

Formal Experiment on AEC and 222Rn. In subfigure a, the left axis represents AEC (blue line), and the right axis represents radon concentration (green line). In subfigure b, the left axis represents temperature (red line) and relative humidity (blue line), while the right axis represents air pressure (green line).

Full size image

The auxiliary experiment was mainly conducted to allow the AECM to run continuously for 5 days to check for the appearance of discrete points. The experiment was conducted in the JinDiXiAnGuShi neighborhood, HongShan District, Wuhan, Hubei Province. To isolate the system from the external environment, the experiment was carried out in a separate room with closed doors and windows. It began at 22:00 on February 29, 2024, and finished at 12:00 on March 6, 2024. No discrete points appeared during the experiment period, and the results are shown in Fig. 6.

Fig. 6
figure 6

Discrete Point Experiment.

Full size image

However, we noticed a sudden rise in the data around 10:00 on March 3, which lasted for about 12 h and then suddenly dropped again. After inspection, we found that the weather had just turned sunny on that day, and combined with the SRC’s data, we suspected that the sudden change and the discrete points may be related to temperature.

To verify this speculation, we conducted a simple temperature experiment using the air conditioner in a bedroom from 12:30 on March 6 to 9:30 on March 7. First, it was in the heating mode, with the temperature set to 30 °C. After running for a few hours, we pushed the door open a small crack and used the remote control to set the air conditioner to the cooling mode, with the temperature set to 17 °C. Finally, we turned off the air conditioner at 20:40 on March 6 in the same way.

The results are shown in Fig. 7. It can be clearly seen that during the heating period, the measurements appeared as discrete points, whereas during the cooling period, they did not. We suspected this phenomenon may be related to molecular thermal motion. As the air in the room is heated, molecular thermal motion intensifies, leading to increased collisions and friction between molecules. This in turn causes charge to transfer between molecules, turning neutral molecules into ions. At the same time, we do not suggest that this phenomenon is caused by wind. This is because during the preliminary experiment and formal experiment, a fan was placed in the SRC. During the temperature experiment, wind also blew during the cooling process. If wind is the factor, then the measurement results of the preliminary experiment, the formal experiment, and the cool experiment should all be discrete, but they are not. Therefore, we do not suspect that wind is the factor, but rather molecular thermal motion. Of course, we cannot exclude the sensor as a factor, but this requires further testing.

Fig. 7
figure 7

Temperature experiment. The red line indicates that the air conditioner is set to heating mode, 30℃; the green line indicates cooling mode, 17℃; the black line indicates that the air conditioner is turned off.

Full size image

Since the bedroom could not be completely airtight and the temperature control of the air conditioner was not precise, the temperature experiment was rough. We are designing a more detailed and rigorous temperature experiment project, and will consider air pressure, humidity and other factors.

To avoid the influence of discrete points on the calculation results, we used the method of sliding average to remove them. The measurement period of the radon detector is 10 min, and the sampling rate of the AECM is 1 s. To match the radon detector’s measurement period, we set the size of the sliding window to 600 points for forward sliding. Based on the AECM’s characteristics, values exceeding 1 × 10−13 S/m of the sliding average were considered as discrete points and replaced with null values.

Correlation and relative error analysis

After removing the discrete points, we used the same parameters to perform a sliding average of AEC again. Following that, we selected the data corresponding to the radon detector’s measurement time from the above calculation results for analysis.

According to the experimental process and the trend of the data, we divided the data into several stages to calculate Pearson’s correlation with significance level is 0.05. For the smooth stage, we also calculated the relative error, mean value, and standard deviation. The calculation results are shown in Tables 3 and 4.

Table 4 Normalized relative error.
Full size table

These results show that there is indeed a correspondence between AEC and 222Rn, and also demonstrate that the AECM we developed has a fast and accurate response to changes in radon concentration.

Relationship between radon and atmospheric electric conductivity

Since the radon detector measures α particles produced by radon decay and does not measure α particles produced by the disintegration of radon daughters, this may result in a low calculated AEC.

However, through the analysis of radioactive balance, we found that since the half-life of 222Rn is hundreds to thousands of times longer than radon daughter elements, and the half-life of 210Pb is as long as 22.3 years. Therefore, it can be considered that when 222Rn decays to 210Pb, 210Pb no longer decays during the experimental period. Based on this, we conclude that when the 222Rn concentration is stable, 222Rn will reach equilibrium with radon daughter elements in 1 h, and the total concentration of radon daughter elements is only a few hundred (Bq/m3), which is much smaller than the concentration of 222Rn. Therefore, we ignore the effect of radon daughter elements on AEC and only consider the effect of radon decay. In other words, calculating AEC from radon concentration will not cause significant errors.

Liu et al.44 studied the relationship between the concentration of 222Rn and its daughters and AEC:

$${\sigma _{{\text{Rn}}}}=3.84 \times {10^{ - 17}} \times \sqrt {\frac{{0.22Q}}{{{\alpha _s}}}} \times \frac{{{p_0}}}{p}{\left( {\frac{{{T_0}}}{T}} \right)^{0.75}}$$
(4)

Here, σRn is AEC calculated from 222Rn. Q is the ions production rate of air ionization caused by 222Rn. αs is the ion recombination coefficient under standard conditions, 1.6 × 10−6 cm3/s. p0 is the standard atmospheric pressure, and its value is 1.01 × 105 Pa, p is the actual atmospheric pressure. T0 is the average atmospheric temperature, and its value is 288.15 K, T is the actual atmospheric temperature.

We did not measure the atmospheric pressure and temperature, but the AlphaGuard PQ2000Pro is capable of measuring atmospheric pressure and temperature. Therefore, we used the data from the AlphaGuard PQ2000Pro. However, since the changes in temperature and pressure are very small, the impact on the conversion is negligible and can be ignored.

It should be noted that since we set the bias voltage to 5 V in the experiment and only considered the case where a single particle was collected, the measured conductivity needs to be multiplied by 2. Then, we found that AEC calculated from the 222Rn concentration was approximately twice that obtained from the AECM measurement, as shown in the Fig. 8.

Fig. 8
figure 8

Comparison between theoretical conductivity and measured conductivity. Blue represents the original measured conductivity, red represents twice the measured conductivity, green represents the theoretical conductivity calculated from 222Rn with 1.6 × 10−6 cm3/s, and orange represents the theoretical conductivity calculated from 222Rn with 3.2 × 10−6 cm3/s.

Full size image

To analyze the cause of the deviation, we first considered the possibility that high-speed alpha particles escaped from the Gerdien sensor. We know that:

$$1{\text{eV=1}}{\text{.6}} \times {\text{1}}{{\text{0}}^{ - 19}}{\text{J=}}\frac{1}{2}m{v^2}$$
(5)

We calculated that the speed of alpha particles produced by radon decay is only 1.62 × 107 m/s, and the relativistic effect is not significant. Since the Gerdien sensor has a uniform electrical field of 5 V/m inside, alpha particles perform two kinds of motion inside the sensor. One is uniform linear motion, which will cause them to fly out of the sensor (when considering their ionization of the air, their flight distance is shorter). The other is the deflection caused by the electrical field. If the alpha particles can fly out of the sensor before reaching the inner electrode, it will not cause a change in the sensor current, which will cause the measured AEC to be low. We calculated that almost all alpha particles are captured by the electrical field inside the conductivity meter, and there are no escaped particles.

Another possibility is related to the ion-ion recombination coefficient. Since the work of Liu et al.44 was based on the natural atmosphere, and the radon concentration in the natural atmosphere is very low, there are fewer ion pairs ionized by radon in the atmosphere. However, the radon concentration used in our experiment is very high, which has a stronger ionization effect on the air and creates more ion pairs.

The increase in ion pairs leads to a higher collision probability, and thus an increase in the recombination coefficient. Differences in the chemical composition of the gas or ions also affect the recombination coefficient. The presence of neutral gas significantly influences the recombination process by causing the recombining ions to approach one another at grazing angles, in comparison to traditional theories which largely predict “head-on” collisions. This increases the recombination coefficient several times45,46.

Therefore, due to the large increase in ion pairs caused by radon, we suspected that the ion-ion recombination effect will also be enhanced. We doubled the ion-ion recombination coefficient from 1.6 × 10−6 cm3/s to 3.2 × 10−6 cm3/s, and the measured AEC is basically consistent with the conductivity calculated from the radon concentration.

In summary, as radon concentration increases, the number of alpha particles produced during radon decay also increases linearly. The ionization of the atmosphere by alpha particles, in turn, leads to a linear increase in AEC. The very high radon concentration used in the experiment resulted in a high number of ion pairs, which may increase the recombination coefficient. This may be the reason why AEC calculated from radon concentration is higher than the observed AEC. Because radon is a precursor gas to earthquakes, the experimental results help us understand the complex processes and mechanisms by which pre-earthquake radon anomalies affect atmospheric electric conductivity.

Figure 8 shows a comparison between the theoretical conductivity after correcting Formula (4) and the measured conductivity.

Conclusion

In order to verify whether atmospheric electric conductivity changes before an earthquake, we developed an atmospheric electric conductivity meter based on the Gerdien sensor. We tested the atmospheric electric conductivity meter with high concentrations of radon, and our results showed that:

  1. (1)

    There is a strong positive correlation between atmospheric electric conductivity and radon concentration.

  2. (2)

    Heating the air causes dispersion in the measurements of atmospheric electric conductivity, whereas cooling it does not. This may be due to the increase in triboelectric charging caused by the intensified molecular thermal motion.

  3. (3)

    Atmospheric electric conductivity calculated from the radon concentration is larger than the measured result, which may be related to the ion recombination coefficient. When the radon concentration is very high, the atmosphere is more easily ionized, and the ion pairs increase, which may lead to a larger ion recombination coefficient.

 

Because radon and other gases are released before earthquakes, the alpha rays released by radon during its decay process can ionize the atmosphere, causing changes in atmospheric electric conductivity. Our experiments have verified that an increase in radon leads to a linear increase in atmospheric electric conductivity, providing a basis for conducting atmospheric electric conductivity monitoring in seismic zones. This will help us understand the earthquake development process and promote disaster prevention and mitigation.

Future work

Overall, the first experiment of the AECM met expectations. Preliminary experiment results show that AEC is indeed correlated with radon concentration, and the AECM we developed can measure this change.

Next, we will test CH4, CO2, H2O, and other gases that may change abnormally before an earthquake. The purpose is to check what kind of change in AEC they would cause and what kind of change in AEC they would cause with the action of radon. The problems and phenomena occurred in the first test will also be tested. We will also add a series of sensors to monitor temperature, humidity, pressure, and concentrations of various gases, and add control circuits to ensure that temperature, humidity, and pressure fluctuate within 1%, reducing their impact on experimental results. Based on our plan and the experiments already underway, we expect to complete all experiments by December 31, 2025. The experiment draft is as follows, and Fig. 9 shows the flow chart of the experimental draft:

Fig. 9
figure 9

Flowchart of Subsequent Experiments.

Full size image
  1. (a)

    Let the AECM work continuously in a vacuum environment for 24 h. During this time, no charged particles will be collected on the electrode. This will allow us to obtain the baseline value of the AECM’s output and compare it with the theoretical output value that was designed for the AECM.

  2. (b)

    Control the air pressure to vary ± 10% to test how the air pressure affects AEC and the sensitivity of the AECM to air pressure.

  3. (c)

    Control the humidity to vary from 0% to 100% to test how the humidity affects AEC and the sensitivity of the AECM to humidity.

  4. (d)

    Control the temperature to vary from − 20℃ to 50℃ to test how the temperature affects AEC and the sensitivity of the AECM to temperature.

  5. (e)

    Let the AECM work continuously in the SRC for 24 h to obtain the baseline value of the SRC’s AEC.

  6. (f)

    Inject N2 into the radon chamber to ensure that only N2 is present in it to provide safety protection for the subsequent experiments involving flammable and explosive gases. Based on the observation reports of gas anomalies before earthquakes and for safety reasons, it is planned to control the volume of flammable and explosive gases to half of the lower explosion limit, and conduct 100% volume testing of other non-hazardous gases.

  7. (g)

    First, inject H2 into the radon chamber and observe AEC’s change with H2 in the absence of 222Rn.

  8. (h)

    After completing the observation, inject N2 again to remove any remaining H2. Once H2 is completely removed, inject 222Rn.

  9. (i)

    Once the radon concentration reaches a stable level, inject H2 again to study the impact on AEC when only H2 is ionized by 222Rn.

  10. (j)

    Finally, inject N2 again to restore the radon chamber to a safe state.

  11. (k)

    After completing the combined test of radon and H2, follow the testing process for H2 (steps (g)-(j)) to test moisture, CH4, CO2 and other gases.

  12. (l)

    Determine whether to conduct a combined test of multiple gases based on the actual situation.

While conducting experiments, we are also investigating field observation sites and hope to apply the AECM to seismic observation as soon as possible.