Understanding Complex Electron Radiolysis in Saline Solution by Big Data Analysis.

In this article, we developed a new method to analyze the complex chemical reactions induced by electron beam radiolysis based on big data analysis. At first, we built an element transport network to show the chemical reactions. Furthermore, the linearity between the species was quantified by Pearson correlation coefficient analysis. Based on the analysis, the mechanism of the high linearity between the special species pairs was interpreted by the element transport roadmap and chemical equations. The time variation of the pH of the solution and bubble formation in the solution were analyzed by simulation and data analysis. The simulation indicates that O2 and H2 can easily oversaturate and form bubbles. Finally, the radiolysis of high-energy electrons in pure water was analyzed as a reference for the radiolysis of high-energy electrons in saline solution. This work provides a new method for investigating a high-energy electron radiolysis process and for simplifying a complex chemical reaction based on quantitative analysis of the species variation in the reaction.

Introduction

Radiolysis is a complicated phenomenon induced by ion beams, electron beams, and other radioactive particles on condensed materials. It is important to analyze the radiolysis of an aqueous solution, such as saline, in various applications. For instance, the efficiency of radiotherapy is dominated by the radiolysis of the external radioactive beam, radiotherapy implants, and injections on tumors and living cells.[1,2] Living cells exposed to beta rays and other radioactive sources are damaged by direct radiation hazards and radiation chemical reactions.[3−11] In addition to health science, radiolysis has been investigated in other fields. In radioactive waste disposal work, the service life of metal containers for high-radioactivity liquid storage is reduced because the corrosion of the metals is accelerated by the products of water radiolysis. In some chemical experiments, radiolysis products actuate the experiments for nanoparticle formation and evolution.[12−15] Nanostructures printed by electron beams are also controlled by free radicals induced by the radiolysis of high-energy electrons in water and other solutions.[13,16−22] Thus, the exploration of radiolysis is important and instructive to engineering and technologies.

Multiple water radiolysis product yield rates by pulsed electron beams have been measured since 1962.[23,24] Le Caër defined the water radiolysis process into three stages.[25] In the first stage, water molecules undergo relaxation processes after energy is deposited and provide excited molecules, ionized molecules, and subexcitation electrons. In the second stage, molecules undergo complex physical reactions such as ion–molecule reactions and dissociative relaxation. In the last stage, species undergo chemical reactions and diffuse in water. Schneider revealed the relationship between the water radiolysis product concentration and electron beam setting data by mathematical models and experiments.[26]

Based on the radiolysis of water, the relationship between the yield rate of the radiolysis products and the concentration of the aqueous saline solution was investigated.[27−32] Molecular decomposition and chemical reactions are used to explain the complex species generated during radiolysis. Energy absorption-induced molecular decomposition in solution yields free radicals and other species. Chemical reactions rebuild species chemical bonds and produce other species. These complex species irreversibly change solutions. This complex process is difficult to understand through the related more than one hundred chemical equations. Normal saline is a basic component of human cells, and normal saline radiolysis process research is important to understand radioactive damage to cells. Thus, we analyzed the radiolysis of high-energy electrons in normal saline solution by a big data method. The chemical reaction induced by radiolysis was clarified based on big data analysis.

Model Development

Kinetic Model

A kinetic model of the radiolysis process was established based on the chemical reactions in normal saline solution. In this model, we analyzed the formation of species and the chemical reaction among them. According to Le Caër’s three-stage theory, the radiolysis process can be divided into three stages. However, the theory of how the electron beams affect water during the radiolysis process has great development.[33] In Taylor’s words, the radiolysis effect is the reactions by free radicals produced by water and electron. Thus, the radiolysis process can be grouped into two stages by varied features in the species transformation. In the first stage, the electron beam transmits energy to the water molecules and yields free radicals. This process stops with the removal of the electron beam. In the first stage, new species yielded by the electron beam are listed in eqs and . In eq , the species yield hydrogen and oxygen atoms with new bonds. Moreover, in eq , chloride ions become chloride atoms with electrons lost, and sodium ions are always stable. In radiation dosimetry, the G value is used to define the rate of the new species’ yield or disappearance in the radiolysis process. In the second stage, all species react with others based on the chemical equations shown in the SI (Supporting Information).

The temporal evolution of the species induced by electron beam irradiation was analyzed in this paper; thus, we assumed that the cross area of the solution was exposed homogeneously. A simplified kinetic model was established to describe the temporal evolution of the species concentrations.[26] In addition, the mass of the analyzed solution is a constant in the model. It is suggested that the heat effect induced by laser pulse irradiation increases the temperature of the sample by more than 10 °C.[34] We find that the dose rate in Liu’s work[34] is about 2.5 × 107(Gy/S), this dose rate is the same as the electron beam with 300 keV voltage and 350 pA current. However, living cell research will not use such huge dose rates; on the one hand, the high current will kill the living cell quickly while, on the other hand, the high voltage cannot provide image information clearly. The settings of dose rate in living cell research[35,36] usually are under 2.5 × 106(Gy/s); thus, the heat effect is limited and it can be ignored in this model. The diffusion calculation was neglected because space influence was excluded from consideration in this paper.

The concentration variation rate of all species in the saline solution was calculated by eq with an improved Euler method, where R was calculated by eq , the dose rate of radiolysis was 7.5 × 107 (Gy/s), and the G values of the radiolysis are listed in Table . The detailed symbol description is listed in Table . In this work, we calculated the species concentration with time from 0 to 0.1 s with a step of 10–10 s; thus, we had 109 data points for each species.

Table 1

G Value for Nine Species

name	G value (100 eV)
eh–	3.58
H3O+	4.09
OH–	0.95
H2O2	2.83
H·	1
OH·	3.32
HO2·	0.08
H2	0.27
Cl	0.6175

Table 3

Symbol Description for eqs to 9

symbol	explanation
Ci	concentration of species i
T	exposed time
I	target species
J	species that can react with i and yield other species
L	species that can yield i in the chemical equations
K	species that can react with l and yield i in the chemical equations
ri,j	rate constant for the equation that uses i and j as reactants
Ri	yield rate of species i due to irradiation
ρ	solution density
Ψ	absorbed dose rate
Gi	G value for species i
F	Faraday constant
Pccl,i	PCC value for species l and i
N	length of the data list
V	data serial number in the data list
Vie(t)	convention rate from species i to e at time t
E	product species in chemical equations that use i as a reactant
X	possible product species in chemical reactions by species i
Pie(t)	convention possibility from species i to e
degi+	species i indegreea
degi–	species i outdegreeb
degi	sum of indegree and outdegree for species i

Indegree: the number of product types from the target species by chemical reactions.

Outdegree: the number of reactant types that can yield the target species by chemical reactions.

Pearson Correlation Coefficient (PCC) Calculation

PCC analysis is an effective method for displaying two database relationships in machine learning technique studies.[37] In this paper, the linearity of the concentration of the species was indicated by the Pearson correlation coefficient (PCC). After that, highly correlated species pairs were set according to linearity. In this work, the PCC of species was calculated with eqs –7. We chose 1.1 × 105 data points from 109 data points. The data picking rule was as follows: the complete data from 10–10 to 10–5 s were chosen; one data point for each 10–5 s from 10–5 to 10–1 s was chosen. We performed logarithmic calculations for previously selected data in the PCC calculation.

Normalized Conventional Rate Calculation

In eq , variable V(t) is the conventional rate, which stands for the transform rate from original species i to product e. r is the rate constant in the chemical reaction about species i and j to yield species e. C and C is the concentration of species i and j, respectively. In eq , variable P(t) is the normalized conventional rate, which stands for the form percent for species e, x in variable V(t) is the species that can be yielded by species i, for instance, species O3 can yield O2, HO2, and O3– ; thus, x stands for O2, HO2, and O3–. Species transform path effect can be qualified by P(t): high P(t) means species i transformation to e with high percent while low P(t) means species i transformation to e with low percent.

C and C are functions of time; thus, P(t) is dependent on time. According to P(t) development, transform paths can be classified into three groups as Table shows. The first group is the void path, which has P(t) < 0.03 at all times. In this group, the conversion of species from i to e is negligible even with a theoretical equation to interpret the reaction. In the second group, the path with P(t) > 0.03 and the disturbance of P(t) < 0.01 were defined as stable paths. This definition means that species e was convened from species i without time dependence. In the third group, paths with time-dependent P(t) were considered time-variant paths, and most transform paths belong to group 3. Group 2 and group 3 are shown in Figure with different colors.

Table 2

Path Classification

path type	Pie feature
void path	Pie < 0.03
stable path	Pie > 0.01 and ΔPie < 0.01
time-variant path	Pie depends on time

Figure 2

(a1) Initial H ETR (element transport roadmap). (a2) Complete H ETR. (b1) Initial O ETR. (b2) Complete O ETR. (c1) Initial Cl ETR. (c2) Complete Cl ETR. Arrows are transport paths, arrow colors from blue to red represent path transformation percentages from 99 to 3% for time-varying paths, and gray arrows are stable transformation percentage paths.

Results

Element Transport Roadmap (ETR)

The element transport roadmap (ETR) denotes the efficient element transport paths in the chemical reactions. The ETR was drawn from the analysis of 32 species based on big data on time-scale species concentrations and the corresponding chemical equations. The possible transport paths were provided based on the chemical equations. Then, those paths were classified into three groups according to different P features, the stable and time-variant paths were retained, and the void paths were removed. Table and Figure are instances of the calculation of the ClOH– transport path efficiency by V and P.

Table 4

ClOH– Transformation Path Efficiency Calculation

equation number	equation	rate constant	product	Vie	Pie
87	H + ClOH– = Cl– + H2O	8 × 109	Cl–	VClOH–,Cl–	PClOH–,Cl–
79	eaq– + ClOH– = Cl– + OH–	1 × 1010
113	ClOH– = OH + Cl–	6.1 × 109
110	Cl– + ClOH– = Cl2– + OH–	9 × 104	Cl2–	VClOH–,Cl2–	PClOH–,Cl2–
102	H+ + ClOH– = Cl + H2O	2.1 × 1010	Cl	VClOH–,Cl	PClOH–,Cl

Figure 1

ClOH– transformation path efficiency calculation in the Cl ETR.

First, all chemical equations that use ClOH– as a reactant were listed.

Second, V for each product was calculated. In Table ,

Third, P was calculated. In Table , the denominator for P.is

Last, P for the complete simulation time was plotted and those paths were classified into different groups. The paths from ClOH– to Cl– or Cl are time-varying paths, and the path from ClOH– to Cl2– is a void path because of the low PClOH–,Cl2–.

Hydrogen, oxygen, and chlorine reactions were analyzed with the corresponding ETRs. In the H ETR, time-variant paths are the overwhelming majority, and these complex time-variant paths display the fixability of the H element transformation network. In the O ETR, species have directional close relationships by effective paths that have high P. In the Cl ETR, species have a clear feature with transformation paths, and special species ClOH− and other Cl formed only species are linked by time-variant paths and the remaining species are linked by stable paths. Oxychloride species use stable paths to contract themselves in the O and Cl ETRs.

Pearson Correlation Coefficient (PCC) Calculation

Figure a shows the PCC for all possible special pairs in the chemical reaction induced by electron radiolysis. The species can be classified into different groups according to the PCC value. More details about the classification can be seen in the Discussion section. Species pairs with high correlation were arranged in the same group. Figure b1 shows that the high PCC species pair has a similar shape. Meanwhile, the low PCC species pairs have significantly different shapes, as shown in Figure b2.

Figure 3

(a) PCC (Pearson correlation coefficient) for each species pair. The x axis represents the first species, the y axis represents the second species, and the color and height represent the PCC value. (b1). Time-varying Cl2– concentration (solid line, left label) and HO2– concentration (dotted line, right label). (b2). Time-varying HO2 concentration (solid line, left label) and ClOH– concentration (dotted line, right label). (b3). Time-varying H+ concentration (solid line, left label) and OH– concentration (dotted line, right label). It is easy to see the relationship between different shapes and PCCs.

Discussion

Complexity Analysis for Element Transport Roadmap (ETR)

The complexity of the ETR is qualified by the number of paths connecting special species in the ETR. The indegree is the number of species that can transform into destined species. The outdegree is the number of destined species transformed from a special species. The degree is the sum of the indegree and outdegree in an ETR. The species degree for the H ETR was calculated from Figure a2. It can be seen from the species degrees of H2O (15) and H2O2 (10) that these two species are the dominating transformation stations in the H element transport system. This is because H2O is the original species in solution. Moreover, H2O2 has high activity, can be the reactant in bountiful reactions, and is the product of multiple species. The indegree of HO3 (0) and outdegree of HO3 (1) means that no species produces HO3 in the H ETR. However, the P(t) of the reaction to produce HO3 is small, inducing an insignificant transformation path. Therefore, the HO3 production paths are not shown in Figure a2.

The H element transport network is maneuverable, and most species have multiple removal paths. Figure a2 shows that H2O is the core transport station, is the largest source, and saves the most H atoms.

Figure b2 is the O ETR. The species degree of OH– (10) suggests that OH– is a dominant transport station for the O transport network. The species indegree of O2 (7) indicates that complex reactions produce O2. Moreover, species outdegrees of O2 (2) suggest that O2 only has two removal paths in the reaction. Thus, bubbles easily form in the solution because the species O2 generation rate is higher than the destruction rate. The indegree for species O3 and Cl2O is zero, meaning that the forming paths for these two species are removed in the O ETR because of low P(t).

Species in Figure b2 were classified into two groups. The first group includes O4, Cl2O4, ClO2, ClO3–, ClO, ClO2–, and Cl2O2, and the second group includes the remaining species. Group 1 is mainly formed by oxychloride and O4. Transformation paths for group 1 species except for ClO2– are almost stable. These paths are stable because reactions between group 1 species are sampled, and there are absolute disparities in the concentration of group 1 species. Thus, the effective transformation paths between species of group 1 are few. Group 2 species build a complex and flexible transport network with time-variant transformation paths based on chemical reactions.

Figure c2 shows the Cl ETR with all possible transformation paths. Cl2– has the largest degree, which suggests that Cl2– is an important species. The species indegree of Cl2O (0) means that the Cl2O yield paths are too small to be considered. Similar to the previous discussion, we classified the species into two groups: the first group included Cl2O4, ClO2, ClO3–, ClO, ClO2–, and Cl2O2, and the second group included the remaining species. Species transformation paths in group 2 are time-varying. HClO and HCl are the bridge that links the two groups. The first group species can transform to the second group species, but not vice versa.

ETR uses species transformation paths to exhibit the contact for species based on chemical reaction and species concentration data. The dominant species and important species were discovered by ETR. Species were classified into several groups for the Cl ETR and O ETR according to the transformation path features between them, which will simplify the complex species relationship.

PCC Result Analysis Based on the ETR

According to Figure a, species were classified into three groups based on the PCC analysis. The PCC of the species pair in the identical group was high. Moreover, the PCC of the species pair in different groups was low. The classification is shown in Table .

Table 5

Species Classification

group	species
I	H2O↔ Cl–, H↔ eh–, OH–↔ O–, ClOH–↔ OH
II	H+, Cl, HO2
III	HClO, H2O2, HO2–, O3, O3–, HO3, O2–, ClO, ClO2–, ClO3–, O2, H2, Cl2, Cl2–, Cl3–, Cl2O2, Cl2O, HCl, ClO2, Cl2O4, O4

The PCCs between the species pairs in group I, such as H2O-Cl–, H-eh–, OH–-O–, and ClOH–-OH are high. This result indicates that correlation bandings are only formed between the special species pairs. In group II and group III, the PCCs between every species in the identical group are high, as shown in Figure a. Here, species pairs with PCCs > 0.99 are listed in Table .

Table 6

PCC > 0.99 Species Pairs

	indirect pathb
direct patha	stable path	time-variant path
O3– ↔ O3	ClO2 ↔ O4	Cl2– ↔ ClO	O3 ↔ ClO
O3 ↔ HO3	O2 ↔ HCl	Cl3– ↔ HClO	HO3 ↔ Cl2–
O2 ↔ O4	HCl ↔ O4	Cl3– ↔ ClO	HO3 ↔ Cl2–
ClO2 ↔ Cl2O4	Cl2O ↔ O4	Cl3– ↔ ClO2-	HO3 ↔ HClO
ClO2– ↔ Cl2O2	HClO ↔ ClO	Cl3– ↔ ClO3-	HO3 ↔ ClO
ClO2 ↔ ClO3–	ClO2 ↔ ClO2–	Cl3– ↔ Cl2O2	O3– ↔ HO3
HCl ↔ Cl2O4	ClO2 ↔ HCl	ClO2 ↔ Cl2	O3– ↔ ClO
Cl2O4 ↔ O4	ClO2 ↔ Cl2O2	Cl2 ↔ HCl	O3 ↔ Cl3–
	Cl2O2 ↔ Cl2O	Cl2 ↔ Cl2O2	O3 ↔ HClO
	ClO2– ↔ Cl2O	Cl2 ↔ Cl2O4	HO2– ↔ HO3
	ClO3– ↔ Cl2O2	Cl2 ↔ Cl2O	H2 ↔ ClO2
	ClO3– ↔ Cl2O	Cl2 ↔ O4	H2 ↔ HCl
	ClO2 ↔ Cl2O	HO2– ↔ O3–	H2 ↔ Cl2O4
	Cl2O4 ↔ Cl2O	HO2– ↔ O3	H2 ↔ O4
	ClO2– ↔ ClO3–

A direct path between species pairs, as shown above, means that the transformation between the species pairs can be completed in one reaction.

Species pairs in the indirect path group need multiple reactions to complete the species transformation in ETRs.

The relationship between the high PCC species pairs was interpreted by ETR, as shown in Figures and 4. As shown in Figure , the transformation paths between species pairs with direct paths are single steps. For instance, the PCC value between O4 and O2 is more than 0.99, which can be interpreted as the transformation path between O4 and O2 being a single-step reaction, as shown in equation K108 (SI). Moreover, the PCC value between ClO2 and Cl2O is also higher than 0.99. However, the single-step reaction between ClO2 and Cl2O is absent from the reaction equation list in the SI. Only a multiple-step reaction (Cl2O - HClO - Cl2O2 - ClO - ClO2– - Cl2O4 - ClO2), as shown in Figure a, links the ClO2-Cl2O pairs with a high PCC value. In contrast to the species pairs with time-variant paths, as shown in Figure b, the transformation paths between the species list in Figure a are stable.

Figure 4

Species transformation network for indirect paths. This figure shows the intermediates for species pairs in the Table indirect path group, which includes stable path parts (a) and time-variant path parts (b).

The indirect transformation path group species pairs in the ETR lack the direct transformation path. However, species in one pair in the indirect path group both have the same strong linear correlation species. This intermediate species could be the bridge to contact the species pairs and induce high linearity. Figure a shows that the species pair of Cl2O4 and O4 is an important bridge that links the Cl ETR to the O ETR. Moreover, these species pairs have a high correlation and stable transformation path. Similar to Cl2O4 and O4, the oxychloride species (ClO, ClO2–, ClO3–, Cl2O2, Cl2O, ClO2, Cl2O4) are in contact with each other by stable transformation paths. Another species, HCl and oxychloride, showed a strong relationship because of the stable transformation paths from HCl to Cl2O4.

No direct stable transformation path or effective intermediate species contacts O2 and H2. Thus, instead of the ETR, chemical reactions were analyzed to determine the relationship between O2 and H2. The most important reactions of the O2-H2 pair were selected according to the reaction rate. Clearly, H2 is mainly formed by H2O, which reacts with eh– or H atoms, and H2 mainly reacts with OH to consume itself. O2 is yielded from OH, HO2, and O2–, and O2 mainly reacts with eh– or H atoms. Thus, OH, H, and eh– could be the intermediate species to provide a unique relationship for the O2-H2 pair.

The relationship between each species was analyzed by PCC calculation, which provided an efficient approach to analyze the radiolysis and reaction by a big data method. However, PCC data are defective in indicating the complete connection for all species because PCC results are not based on the complete transformation paths, as shown in the ETR, but are only calculated by the concentration data of two species. The transformation paths were ignored in the PCC calculation, inducing the low linearity of core species with other species. On the other hand, PCC is based on calculated concentration data that include complete chemical reaction information, although the ETR only analyzes the important reactants and overlooks other reactant influences on the species. Thus, the ETR analysis method can help to identify the dominant species, which is a protagonist in chemical reactions, and PCC is an effective tool for providing highly correlated species pair information that we cannot find in the ETR.

pH, Oversaturated Gas, and the Difference between Saline and Pure Water

Solution pH is well known as a time-varying property during the radiolysis process because chemical reactions change the H+ concentration. Moreover, when the radiolysis process is sufficiently long, the solution cannot dissolve all the O2 and H2 yielded by chemical reactions, and these gases quickly form bubbles and keep the concentrations of O2 and H2 constant in the solution.

In Figure a1, the pH decreases at a high rate before 10–6 s, and then the pH increases. Finally, the pH in a low stage suggests that the solution became acidic. The unusual increase in pH during 10–6–10–4 s can be explained in terms of chemical reactions. Several reaction equations that include H+ as a reactant or product and with a high reaction rate were analyzed as key equations. The rate of the chemical reaction that yields H+ always increases, while the H+ consumption chemical reaction rate increases and decreases. Thus, a possible reason is that the H+ decrease rate grew faster than the H+ increase rate and induced an increase in pH. The concentrations of HO2– and O2– increased, and the two species can react with H+. Thus, the increase in HO2– and O2– concentrations accelerated the H+ consumption rate and finally induced an increase in pH during 10–6–10–4 s.

Figure 5

(a1) pH and pOH of solution versus time. pH undergoes a drastic change in 0.1 s. (a2) Time variation of pH + pOH. (b) Time variation of O2 concentration (solid line) and H2 concentration (dotted line); the straight line is the gas saturation moment. (c) Escaped concentration for H2 (solid line) and O2 (dotted line). O2 escaped before H2, but both species have the same curve track.

Figure c shows that the shape of the escaped O2 concentration is similar to that of H2, and the O2 dissolution saturation time is shorter than that of H2. However, O2 is more soluble than H2 in solution, suggesting that O2-related reactions are stronger than H2-related reactions.

To find the differences in radiolysis inference between pure water and saline solution, species concentration databases of pure water and saline solution were plotted against time, as shown in Figure .

Figure 6

Species concentration ratio between the salt solution and pure water. x axis is time, and y axis is Csalt/Cpurewater. (a) Species whose concentrations are lower in salt solution rather than pure water. (b) Species whose concentrations are higher in salt solution rather than pure water. (c) Species whose concentrations are first lower and then higher in salt solution rather than pure water. (d) O2 and H2 concentration ratio in salt solution and pure water.

Here, we only consider those species that appeared in pure water under electron beam exposure.

The Schneider pure water model[26] was used to calculate the species concentration variation in pure water. The G value is identical for the same species in pure water and saline solution, and the saline solution has a G value for Cl– in addition. Other settings, including the calculation step, the initial value, and the simulation completion time, were identical for the two types of solutions. In Figure , the salt–water model species concentrations approach pure water before 10–6 s, and then the two models’ concentrations become different for most species.

After 10–5 s, the concentration ratios differ. Species OH–, HO2–, O–, O2–, and O3– are always in the decreasing stage (Figure a), which indicates that these species are less abundant in saltwater than in pure water. Moreover, species H+, H2O2, OH, HO2, and H2O are in the opposite state (Figure b). The concentration ratios of species eh–, O3, HO3, and H initially decrease and then increase (Figure c). O2 and H2 have multiple trend changes and finally obtain ratios of 1 (Figure d) because both species remain at saturation concentrations.

From Table , the yields of OH–, H2O2, HO2, O–, O3–, O3, HO3, and H2O and the applied rate in pure water are higher than those in salt solution. The yield and applied rate of species H, OH, HO2, O2–, H2, and H+ are lower in pure water than in salt solution. O2 has approximate data for yield and applied rate in the two solutions. The yield rate for eh– is much better in pure water than in saline solution, and the applied rate is similar in the two solutions.

Table 7

Variation Rate of 16 Species in 10–3 s by Two Modelsa,b

	reaction rate (μmol L–1 s–1)
species	yield by water	yield by salt	applied in water	applied in salt	Csalt/Cwater ratio
eh–	4.8 × 103	15	2.7 × 107	2.7 × 107	7.8 × 10–1
H+	1.1 × 106	5.7 × 107	3.2 × 107	8.9 × 107	6.2 × 102
OH–	5.2 × 1010	1.7 × 108	5.2 × 1010	1.8 × 108	1.5 × 10–3
H2O2	5.2 × 1010	1.6 × 108	5.2 × 1010	1.8 × 108	2.0
HO2–	5.2 × 1010	1.5 × 108	5.2 × 1010	1.5 × 108	2.8 × 10–3
H	9.6 × 104	1.8 × 106	7.8 × 106	9.6 × 106	2.0
OH	5.0 × 107	2.8 × 1010	7.6 × 107	2.8 × 1010	1.9 × 101
O–	4.0 × 107	1.0 × 106	4.0 × 107	1.1 × 106	2.5 × 10–2
HO2	2.5 × 107	7.0 × 107	2.6 × 107	7.1 × 107	1.0 × 102
O2–	3.7 × 107	6.3 × 107	3.6 × 107	6.3 × 107	1.4 × 10–2
O2	2.2 × 107	2.3 × 107	1.5 × 107	1.5 × 107	1
H2	2.5 × 103	3.8 × 103	9.4 × 104	1.3 × 106	1
O3–	1.9 × 106	4.1 × 104	1.9 × 106	4.2 × 104	1.4 × 10–4
O3	2.6 × 105	664	2.6 × 105	666	6.2 × 10–2
HO3	963	302	975	302	3.1 × 10–1
H2O	5.2 × 1010	2.0 × 108	5.2 × 1010	1.5 × 108	1

This table uses the pure water model and salt solution model results to calculate the rate for listed species. Only species that appeared in the two models were considered. This table uses species concentration data at 10–3 s.

“Yield by water” is the yield speed of species concentrations in the pure water model, and “applied in salt” is the consumption speed of species concentrations in the salt solution model.

For species OH–, HO2, O–, O3–, O3, HO3, and H2O, their reaction rates are higher in pure water than in saline solution, and the concentrations in pure water are higher than those in salt solution, suggesting that the addition of Cl decreases their reaction rate and provides a negative environment to accumulate these species. In contrast, the saline solution provides a positive environment for the accumulation of species H, OH, HO2, O2–, and H+ with a high reaction rate. In contrast to other species, the accumulation rate of H2O2 species is high in saline solution, but the reaction rate is high in a pure water environment. This is because the difference between the provision rate and depletion rate of H2O2 in saline solution is lower than that in pure water. O2 and H2 are already oversaturated in both solutions; thus, their concentrations in solution are constant. H2O has a larger basic concentration, and the yield and consumed concentration can hardly infer the H2O data. Thus, the H2O data in the two types of environments are similar.

Conclusions

In this article, we built the ETR of the chemical reaction induced by high-energy electron radiolysis (HEER) based on the chemical reaction equation and big data analysis. Based on the simulation and PCC analysis, the highly linear species pairs were selected and interpreted by the ETR. The ETR provides essential information on the chemical reaction, such as the element transport, reaction rate, and reaction direction. Combining ETR and PCC analysis, we developed an effective and reliable method for analyzing the complex chemical reaction induced by high-energy electron radiolysis in saline solution. The time variation of pH and bubble formation induced by high-energy electron radiolysis were analyzed based on this method.