Persistence of persulfate in karst groundwater and its influencing factors

doi:10.21203/rs.3.rs-5335152/v1

Persulfate (S₂O₈^2-, PS) is a new type of oxidant used for in situ chemical oxidation (ISCO) during the remediation of contaminated groundwater. However, PS may be consumed by nontarget matters in aquifers, decreasing its persistence and remediation effect. To better understand the persistence of PS in widely distributed karst aquifers, microcosm, column, and conduit experiments were carried out in this study to simulate karst caves, fracture zones, and conduit environments under static or flow water conditions. Karst aquifer matters, including limestone and lime soil, and a novel carbonate rock conduit model were employed. PS decomposition at different concentrations, influencing factors, and hydro-chemical responses were discussed. The results of the study indicate that (1) In static water, the half-lives of 1, 8, and 20 g/L of PS in limestone media were 102, 185, and 202 d, respectively, and 19, 34, and 51 d in lime soil media, respectively. As the injection concentration increased, the persistence of PS also increased. The half-life range of PS in limestone column and conduit was 0.05 ~ 0.13 d and 0.36 ~ 1.70 d, respectively, indicating that PS exhibited poor persistence under flowing karst water conditions. (2) The pH remained at neutral to slightly alkaline levels in limestone media, which buffered the acidizing effect of high PS concentration. When PS concentrations were 8 and 20 g/L, the organic matter content in lime soils decreased from an initial 45.57 g/Kg to 35.07 and 24.63 g/Kg, respectively. The rich organic matters in lime soils greatly consumed PS with a rapid degradation rate constants of 20.31 and 13.47d^-1, respectively. (3) The decomposition of PS led to obvious hydro-chemical responses under static groundwater conditions. The pH values dropped to a minimum of 1.4 and the dissolved oxygen concentration increased from 8.5 to 17.3 mg/L in the absence of solid media. When the limestone media were present, PS injection at the concentration of 20 g/L stimulated the dissolution of carbonate, producing Ca²⁺ concentration 8–10 times higher than the background value. However, the hydro-chemical changes remained relatively stable under flowing karst water conditions.

Persulfate

Decomposition

Persistence

In situ chemical oxidation

Karst groundwater

Persulfate (S₂O₈^2-, PS) has been widely used as a highly efficient oxidant for in situ chemical oxidation (ISCO) in recent years (Devi et al., 2016; Liu et al., 2022; Nguyen et al., 2023) because it is more persistent than peroxide in the presence of aquifer materials and can be used to treat a wider range of organic compounds than permanganate (Sra Kanwartej S. et al., 2013). In field-scale applications, persulfates have been used to remove contaminants such as chlorinated solvents (Head et al., 2020), Polycyclic Aromatic Hydrocarbons (PAHs) (Wang et al., 2024), and organic matter (Eberle et al., 2017; Ding et al., 2024) from soil or groundwater. Sra Kanwartej S et al. (2013) injected persulfate into a gasoline source zone at the Canadian Forces Base (CFB) to monitor the degradation of 9 gasoline compounds. However, in aquifers, PS might be consumed by nontarget sediments in addition to undergoing reactions with contaminants (Teel et al., 2011; Teel et al., 2016; Li et al., 2017; Mustapha et al., 2021). For example, subsurface sediments that are rich in organic matter can consume PS (Ahmad et al., 2013; Watts et al., 2018; Zawadzki, 2022). Minerals containing iron or manganese oxides can activate PS and lead to the rapid decomposition of PS (Haizhou et al., 2014; Liu et al., 2023; Jiang et al., 2024; Liang et al., 2024). These factors allow for PS to be readily consumed before it comes into contact with target pollutants. In addition, PS is likely to be rapidly transported under fast water flow conditions, reducing the possibility to contact with pollutants. Therefore, to make ISCO technology more cost-effective, a better understanding of the persistence of PS in aquifers is needed.

Many researchers have suggested that the degradation of PS can cause a decrease in pH in porous media (Wu et al., 2019; Zheng et al., 2022). In contrast, in the presence of matter containing calcium carbonate, a decrease in pH can be buffered (Wei et al., 2020; Jiang et al., 2021; Wang et al., 2022), which may be beneficial to the persistence of PS in aquifers. Karst aquifers characterized by soluble carbonate rocks, karst caves, and karst conduit, exhibit significant differences from other aquifers, such as complex hydraulic connectivity, pronounced anisotropy, and rapid dynamics (Wu et al., 2009; Hao et al., 2020), which may complicate the persistence of persulfate. Such karst aquifers are easily polluted by organic contaminants due to their openness and vulnerability (Zhu et al., 2019). ISCO with PS is an available strategy for remediating such karst groundwater contamination. However, there is a lack of studies on the decomposition behaviour of PS in karst environments.

To understand the persistence of PS in karst environments, a set of experiments, including microcosm, column, and conduit experiments, were carried out with under static or flowing karst groundwater conditions. Limestone particles and lime soil were used as the experimental materials. A novel conduit model made of carbonate rock was used to simulate karst conduit environments. The aim of this study was to investigate the persistence of PS, the influencing factors, and hydro-chemical responses in karst aquifers. The findings of the study are expected to serve as a technical foundation for the implementation of ISCO in karst regions.

1.1 Experimental materials

Experiments in the study consisted of three parts, namely, microcosm, column, and conduit experiments, simulating water environments of karst caves, karst fracture zones, and karst conduits, respectively (Fig. 1). Glass bottles each with a volume of 1000 ml (Fig. 2a), stainless-steel columns each with a length of 120 cm and inner diameter of 4.5 cm (Fig. 2b), and a carbonate rock conduit model (Fig. 2c) were employed for these experiments. Limestone particles (1~5 mm) and lime soil collected from a karst aquifer located in the Guilin area, China, were used as media for the microcosm and column experiments, respectively. Carbonate rock was collected to construct the conduit model. It consisted of some carbonate rock pipes with different inner diameters ranging from 1 to 5 cm and a carbonate rock sphere with an inner diameter of 10 cm, linked by three stainless-steel tubes. The conduit model had a total length of 367.5 cm, an inner surface area of 5602 cm², and a volume of 8672 cm³. Karst groundwater was injected at the inlet K1 and discharged at the outlet K8, severally controlled by two peristaltic pumps for the same flow rate. The ports of K2~K7 on the conduit top were set for sampling water. Mineralogical compositions of the limestone, lime soil, and carbonate rock used were determined by X-ray fluorescence spectrometry (ZSX Primus Ⅱ, Japan), as given in Table 1.

Table 1 Chemical components of the aquifer materials used in the experiments

Material	Mass ratio	Al₂O₃		CaO	Fe₂O₃	K₂O	MgO	MnO	Na₂O	P₂O₅	SiO₂	TiO₂	LOI
Limestone	%	0.46	53.87		0.13	1.81	—	0.01	0.17	0.29	1.19	0.01	42.90
Lime soil	%	15.61	5.39		0.92	1.33	1.59	0.26	0.08	0.25	41.29	1.33	31.95
Carbonate rock	%	1.54	44.38		0.36	—	0.67	—	—	—	12.22	—	40.84

Note: LOI is the loss on ignition, and “—” means not detected.

Karst groundwater used in the experiments was collected from the Yanshan Campus of the Guilin University of Technology. Its initial physicochemical characteristics were as follows: DO, 7.3~9.1 mg/L; SO₄^2-, 2.36 mg/L; Ca²⁺, 62 mg/L; HCO₃^-, 252.6 mg/L; pH, 8.60; EC, 290 μs/cm; and the oxidation‒reduction potential (ORP), -70.5 mV. The experimental reagents including sodium persulfate (Na₂S₂O₈) , potassium bromate (KBr), and potassium iodide (KI) were analytically pure purchased from Xilong Science Co. in China.

1.2 Experimental methods

1.2.1 Microcosm experiment

The microcosms were divided into three groups: the control group (A1~A3) with no aquifer material, the limestone group (B1~B4) with 400 g of limestone added, and the lime soil group (C1~C4) with 400 g of lime soil added (as shown in Table 2). Each microcosm had a total groundwater volume of 800 mL. PS at concentrations of 1, 8, and 20 g/L was initially prepared for each group, corresponding to A1~A3, B1~B3, and C1~C3, while B4 and C4 were prepared with no PS. Each microcosm was prepared in triplicate and placed indoors at a temperature of 25±3°C.

1.2.2 Column experiment

The experimental column was fully filled with 2.16 kg of limestone particles and placed in a horizontal position. The groundwater was driven from Z1 to Z2 using a peristaltic pump at a rate of 56.4 ml/h. Before PS injection, the hydraulic parameters of the column were estimated using a bromide tracer test, which showed an average migration velocity of 1.09 m/d and a dispersion of 0.206 m. Three instantaneous injections of PS (10 mL each time) were conducted at concentrations of 5, 10, and 20 g/L (Table 2). The average water temperature in the column was 18.3°C during the experiment. For the column experiments, only limestone was used because aquifer materials in the fracture zones are mainly composed of broken limestone.

Table 2 Types and setups of the experimen

Experiment type	Groups	Solid medium	PS addition		Hydraulic conditions
Microcosm experiment	A1	No solid medium	1	Initial concentration, g/L	Static state, liquid volume 800 mL
	A2		8
	A3		20
	B1	Limestone, 400 g	1	Initial concentration, g/L
	B2		8
	B3		20
	B4		No added
	C1	Lime soil, 400 g	1	Initial concentration, g/L
	C2		8
	C3		20
	C4		No added
Column experiment		Limestone, 2.16 kg	5	Instantaneous injections three times, 10 ml each time	Flow rate was 1.09 m/d
			10
			20
Conduit experiment		No solid medium	5	Instantaneous injections three times, 10 ml each time	Flow rate was 1.08 m/d
			10
			20

1.2.3 Conduit experiment

The conduit experiment included three groups, in which the PS concentrations in the 10 mL injected solutions were 5, 10, and 20 g/L (Table 2). The inlet point was K1, and the outlet point was K8 for PS. Prior to the injection of PS, a sodium fluorescein tracer test was introduced in the system to estimate the hydraulic parameters of groundwater flow at a rate of 120 mL/h. According to the Qtracer2 software (EPA, 2002), the average migration velocity was 1.08 m/d, and the dispersion coefficient was 0.312 m. The average water temperature was 23.1°C during the conduit experiment.

1.3 Sample analysis

Water samples were collected at regular intervals for the three different experiments. A syringe was used to collect 15 mL of water sample for the microcosm experiment, 15 mL for the column experiment at Z2, and 20 mL for the conduit experiment at K8. The water samples were then tested for different parameters using various instruments. DO and pH levels were analysed using a portable water multiparameter meter (HQ30d, USA). The SO₄^2- concentration was measured using ion chromatography (Dionex ICS 2100, USA), while the S₂O₈^2- concentration was analysed by a UV‒visible spectrophotometer (UV-6000 PC, China). The concentrations of Ca²⁺ and HCO₃^- were determined using a reagent kit (Merck, Germany). The dissolved organic carbon (DOC) content was measured using a total organic carbon analyser (multi-N/C 3100, Germany), and the organic matter content (OMC) of the solid materials was determined using a potassium dichromate volumetric method. The total organic carbon (TOC) content of the solid particles was determined using an oscillatory centrifugation method.

2.1 Decomposition of PS

2.1.1 Under static water conditions

Figure 3 illustrates the normalized concentration of PS with small sample errors in the microcosms. The data indicate that PS decomposed to varying degrees when the PS concentration was 1, 8, or 20 g/L. PS decomposition reached 19.3–44.1% in A1 ~ A3, 28.8–48.9% in B1 ~ B3, and 56.8–89.8% in C1 ~ C3 (as shown in Table 3). The results indicated that PS decomposition in B1 ~ B3 was similar to that in A1 ~ A3 with no solid material. However, PS decomposition in C1 ~ C3, which had lime soil media, was significant, almost double that in B1 ~ B3, which had limestone media.

Table 3

Decomposition parameters of PS in the microcosm experiment
Parameters	Unit	Control			Limestone			Lime soil
Parameters	Unit	A1	A2	A3	B1	B2	B3	C1	C2	C3
PS^*	g/L	0.78	4.16	11.76	0.49	5.18	13.83	0.10	2.28	8.29
Decomposition percentage	%	19.3	44.1	38. 5	48.9	30.9	28.8	89.8	71.6	56.8
Decay rate constant k	10^− 3 • d^− 1	2.17	5.88	4.91	6.78	3.73	3.43	35.72	20.31	13.47
half-life	d	319.64	117.85	141.24	102.21	185.90	202.12	19. 41	34.13	51.56
R²	—	0.94	0.87	0.88	0.90	0.85	0.82	0.96	0.97	0.95
Note: * Residual concentration of PS at the end of the experiment.

The first-order decay rate constant (k) and half-life of PS were determined by fitting the PS normalized concentration curve using the exponential equation ln (C_t/C₀) = -kt, as presented in Table 3. The determination coefficient (R²) ranged from 0.82 to 0.97, indicating the high credibility of the fitting results. In the control groups A1 ~ A3, the rate constant k ranged from 2.2 to 5.9 × 10^− 3 /d, which was higher than that in a previous study k was estimated to be 0.37 × 10^− 3 /d under similar experimental conditions with deionized water at 20°C (Sra et al., 2010). This difference could be attributed to the high salinity of groundwater and high water temperature in our experiment, which can enhance the decomposition of PS (Tsitonaki et al., 2008). Sra et al. (2010) reported a wider range of PS decay rate constants (1.33 to 42.38 × 10^− 3/d) in sandy to silty-sand calcite-rich aquifer materials. In our study, the rate constants ranged from 3.4 ~ 6.8 × 10^− 3 /d for limestone groups B1 ~ B3 and 13.5 ~ 35.7 × 10^− 3 /d for lime soil groups C1 ~ C3, indicating the significant influence of aquifer materials on PS decomposition, compared with groups A1 ~ A3.

The aim of the microcosm experiment was to understand the persistence of PS in karst caves under relatively hydrostatic conditions. The results showed that the presence of lime soil more significantly decreased the persistence of PS than the presence of limestone. Generally, shallow karst caves are often filled with lime soil, which is the weathering product of limestone. Therefore, ISCO with PS in karst caves should investigate the effect of filled matters on the decomposition of PS.

2.1.2 Under flowing water conditions

(1) Column experiment

A column experiment using a limestone-filled stainless-steel column was conducted to simulate a fractured karst aquifer. The concentrations of PS at Z2 were monitored during three instantaneous injections of PS from Z1, as shown in Fig. 4a. Similar to the bromide nonreactive tracer, the PS concentration increased and decreased with similar peak times. The peak concentrations of PS observed at Z2 were 121.3, 207.5, and 300.0 mg/L, corresponding to injection concentrations of 5, 10, and 20 g/L, respectively, at Z1. Based on the fitting of the exponential equation (R² > 0.90), the peak concentrations followed a first-order decay model. The first-order decay rate constants of PS ranged from 5.272 to 14.895 d^− 1 with half-lives of 0.047 ~ 0.131 d, while bromide had a rate constant of 2.183 d^− 1 and a half-life of 0.317 d. PS had a greater decay rate and a shorter half-life than bromide, indicating that it can be consumed by chemical reactions. Some hydraulic parameters for the column were estimated based on the method of US EPA (2002). Compared with bromide, PS also had a shorter mean residence time, a greater average migration speed, and a smaller longitudinal diffusion coefficient, as shown in Table 4. With the increasing of PS concentrations injected, mean residence time and half-life of PS increased, and its persistence became better.

Table 4

Estimated parameters of tracers and PS in the column and conduit experiments
Group	Component	Concentration^*	Recovery rate (%)	Mean residence time (d)	Average migration speed (m/d)	Longitudinal diffusion coefficient (m²/s)	Decay rate constant (d^− 1)	Half-life (d)	R²
Column	Br⁻	256	96.9	0.85	1.35	0.26×10^− 5	2.183	0.317	0.94
	PS	5	53.8	0.55	2.08	0.44×10^− 6	14.895	0.047	0.96
	PS	10	74.2	0.58	1.99	0.71×10^− 6	6.904	0.100	0.99
	PS	20	67.0	0.66	1.75	0.73×10^− 6	5.272	0.131	0.97
Conduit	Fluorescein	44	91.9	3.48	1.08	0.39×10^− 5	0.59	1.174	0.98
	PS	5	60.4	4.14	0.89	0.51×10^− 5	0.408	1.698	0.93
	PS	10	68.2	3.92	0.94	0.75×10^− 5	0.436	1.591	0.90
	PS	20	59.5	3.20	1.15	0.98×10^− 5	0.511	1.356	0.94
Note: *Br⁻, mg/L; PS, g/L; sodium fluorescein (C₂₀H₁₀Na₂O₅), µg/L.

(2) Karst conduit experiment

Sodium fluorescein as a nonreactive tracer was used to carry out a tracer test prior to three instantaneous injections of PS at concentrations of 5, 10, and 20 g/L. The breakthrough curves of fluorescein and PS concentrations at K8 are shown in Fig. 4b. When PS was injected at a concentration of 5 or 10 g/L, the peak times were similar. However, an earlier peak occurred when the injected concentration of PS was 20 g/L. PS exhibited a similar trailing curve to fluorescein, with extremely low concentrations at the end of the experiment.

The mean residence times and the average migration speeds of PS ranged from 3.20 to 4.14 d and from 0.89 to 1.15 m/d, respectively, while those of fluorescein were 3.48 d and 1.08 m/d, respectively. PS and fluorescein had similar residence times, migration speeds, and hydrodynamic coefficients (as shown in Table 4). The first-order decay rate constants of PS ranged from 0.408 to 0.511 d^− 1, with half-lives of 1.356 to 1.698 d, which were close to those of fluorescein. These characteristic parameters suggest that PS strongly persisted in the karst conduit model.

Because bromide and fluorescein are two common nonreactive tracers in hydrogeologic filed investigation, they can act as reliable references for the decomposition of PS in the experiments. Compared to the column experiment, the conduit experiment showed lower decay rate constants, longer half-lives, and better persistence of PS under the similar flow rate conditions. Compared with the microcosm experiment under static water conditions, the persistence of PS in the column and conduit experiments decreased greatly, due to the processes of flow water. The static water condition is more favourable for ISCO-PS technology in the remediation of groundwater contamination.

2.2 Influencing factors of PS persistence

2.2.1 Advection and dispersion

Advection and dispersion are two main hydrodynamic factors affecting the fate and transport of a solute in aquifers (Sakr et al., 2023), which can lead to a mass loss of PS within an aquifer volume. The faster the velocity of flow water is, the greater the PS mass loss, and the shorter the PS mean residence time. For example, at an injected concentration of 5 g/L, the migration speeds of PS in the column and conduit were 2.08 and 0.89 m/d, respectively, and the mean residence times were 0.55 and 4.14 d, respectively (Table 4). In the limestone column, filled matters decreased the cross-sectional area and accelerated the migration speed. As shown in Table 3, the half-lives of PS in the column and conduit experiments were less than 2 d. However, the half-lives of PS in the microcosm experiment ranged from 19 to 320 d. This suggests that the persistence of PS in aquifers is subject to advection and dispersion, and that static water conditions greatly help to improve the persistence of PS.

2.2.2 Organic matter content (OMC)

The impact of OMC on the persistence of PS was examined in the microcosm experiment. PS can be consumed and activated by high OMC in lime soil. For example, Ahmad et al. (2013) suggested that phenol salts in organic matter can activate PS to produce $\:{\text{SO}}_{\text{4}}^{-·}$and $\:{\text{HOO}}^{\text{·}}$under neutral to alkaline conditions. As shown in Table 5, microcosms B4 and C4, which did not contain PS, had stable OMC, DOC and TOC contents without significant change. However, microcosms B1 ~ B3 and C1 ~ C3, which contained PS, showed significant changes in OMC, TOC, and DOC. As the initial concentration of PS increased, the residual contents of OMC and TOC decreased, and the mean concentration of DOC increased. In particular, microcosms C1 ~ C3 showed more significant changes than microcosms B1 ~ B3 (Fig. 4) due to the large differences in OMC. The initial OMC in the limestone and lime soil groups were 8.95 and 45.57 g/kg, respectively. More PS was consumed at higher OMC and more DOC was produced. The maximum concentrations of DOC reached 12.24 mg/L in B1 ~ B3 and 150.93 mg/L in C1 ~ C3. According to Fang et al. (2018), the reaction between OMC and PS can produce some intermediates, such as alkyl RH, as shown in the following reaction equation.

$$\:{\text{S}}_{\text{2}}{\text{O}}_{\text{8}}^{\text{2-}}\text{+SOM=S}{\text{O}}_{\text{4}}^{\text{·-}}\text{+RH}$$

1

$$\:\text{S}{\text{O}}_{\text{4}}^{\text{·-}}\text{+RH=S}{\text{O}}_{\text{4}}^{\text{2-}}\text{+R}$$

2

Therefore, OMC in aquifers can partially consume the injected PS. The higher the OMC is, the greater the consumption of PS, and consequently, the poorer the persistence of PS. When karst aquifers contain lime soil rich in OMC, the persistence of PS decreases.

Table 5

Variations in OMC, TOC, and DOC in microcosms
Group	Number	Initial concentration (g/L)	OMC (g/kg)		TOC (g/kg)		DOC (mg/L)
Group	Number	Initial concentration (g/L)	Initial	Residual	Initial	Residual	Range	Mean
Limestone	B1	1	8.95	7.92	0.49	0.44	4.42 ~ 10.06	6.8
	B2	8		4.02		0.24	4.42 ~ 14.27	11.01
	B3	20		2.88		0.18	4.42 ~ 17.12	12.24
	B4	–		8.89		0.48	5.68 ~ 12.95	9.43
Lime soil	C1	1	45.57	43.46	2.17	1.96	22.70 ~ 29.26	25.33
	C2	8		35.07		1.67	27.14 ~ 139.06	56.64
	C3	20		24.63		1.17	22.03 ~ 389.56	150.93
	C4	–		45.32		2.15	10.96 ~ 20.34	16.63

2.2.3 Type of aquifer materials

In carbonate rock areas, limestone and lime soil are the main aquifer materials. As shown in Table 1, limestone mainly contained calcium oxide with mass ratio of 53.87%. Lime soil was composed of silica oxide, calcium oxide, iron oxide, and manganese oxides with mass ratios of 41.29%, 5.39%, 0.92%, and 1.59%, respectively. Carbonates in limestone can adjust the pH and absorb hydrogen ions produced by PS decomposition, being favourable for inhibit the environmental acidification. In microcosms, PS at concentrations of 8 and 20 g/L exhibited better persistence in limestone material, primarily attributed to the pH regulation capability of limestone. However, the main component of limestone, CaO, undergoes lime hydration to form hydrated lime (Ca(OH)₂) when in contact with water, which can release heat and create an alkaline environment. Recent studies have shown that PS decomposition accelerates under increased temperature and alkaline conditions. Alkaline conditions can induce PS decomposition (Furman et al., 2010), generating $\:{\text{SO}}_{\text{4}}^{\text{-·}}$, $\:\text{HO}\text{·}$, and $\:{\text{O}}_{\text{2}}^{\text{-·}}$ radicals, which efficiently degrade polycyclic aromatic hydrocarbons (Liang and Guo, 2012). In this study, the pH range of the limestone group was 5.7–7.9, primarily neutral to slightly alkaline, and therefore the effect of alkaline catalysis on PS was not significant. Furthermore, carbonate ions (HCO₃⁻/CO₃²⁻) possibly consume the sulphate radicals ($\:{\text{SO}}_{\text{4}}^{\text{-·}}$) produced by PS, as shown in the Eq. 4–5 (Liang et al., 2006; Bennedsen et al., 2012). This may be detrimental to the efficiency of persulfate oxidation.

$$\:\text{S}{\text{O}}_{\text{4}}^{\text{•-}}\text{+}{\text{HCO}}_{\text{3}}^{\text{-}}\text{=}{\text{SO}}_{\text{4}}^{\text{2-}}\text{+}{\text{HCO}}_{\text{3}}^{\text{•}}$$

4

$$\:\text{S}{\text{O}}_{\text{4}}^{\text{•-}}\text{+}{\text{CO}}_{\text{3}}^{\text{2-}}\text{=}{\text{SO}}_{\text{4}}^{\text{2-}}\text{+}{\text{CO}}_{\text{3}}^{\text{•-}}$$

5

The composition of lime soil is complex, containing not only high levels of organic matter but also indigenous microorganisms (Zhu et al., 2018). Studies have indicated that PS can inhibit the activity of certain microbial species (Sutton et al., 2013; Chen, 2016). However, in this study, PS reacts with the organic matter in lime soil, resulting in partial consumption, which attenuates its inhibitory effect on microorganisms. In some contaminated soils, this attenuation facilitates the combined remediation by PS and biodegradation. For instance, Tsitonaki et al. (2008) found that persulfate concentrations up to 10 g/L did not impact the indigenous microorganisms. When iron or manganese oxides are present in aquifers, PS may react with them to produce $\:{\text{S}}_{\text{2}}{\text{O}}_{\text{8}}^{\text{-·}}$, which subsequently undergoes a free-radical chain reaction, resulting in the production of $\:{\text{SO}}_{\text{4}}^{\text{-·}}$ (Liu et al., 2016). Therefore, different aquifer materials may lead to different persistence of PS. Limestone aquifers are suitable for ISCO technology, while lime soil aquifers are more suitable for combined ISCO and biodegradation remediation.

2.2.4 PS concentration

In static water without added materials, the decomposition of PS accelerated as the concentration of PS increased. However, when aquifer materials were added, a higher PS concentration led to a longer half-life and improved persistence. For instance, with increasing initial PS concentration, PS decomposition increased in A1 ~ A3 but decreased in B1 ~ B3 and C1 ~ C4. This suggested that PS with higher initial concentration had better persistence in limestone and lime soil material. In flowing water with aquifer material, the persistence of PS also increased with increasing concentration, with half-lives of 0.047, 0.100, and 0.131 h for PS concentrations of 5, 10, and 20 g/L, respectively. However, in flowing water without aquifer material such as conduit flow, the persistence of PS decreased with increasing concentration, with half-lives of 1.698, 1.591, and 1.356 d for PS concentrations of 5, 10, and 20 g/L, respectively. Overall, with the increasing of the initial PS concentrations, the persistence of PS decreased in the absence of aquifer material but increased in the presence of aquifer material.

2.3 Hydro-chemical responses

2.3.1 pH

The decomposition of PS can generate H⁺, leading to a decrease in pH, and excessively low pH promotes the production of $\:{\text{SO}}_{\text{4}}^{\text{·-}}$ (Liang et al., 2007), as shown in Eqs. 6 ~ 8.

$$\:\text{2}{\text{S}}_{\text{2}}{\text{O}}_{\text{8}}^{\text{2-}}\text{+2}{\text{H}}_{\text{2}}\text{O=4S}{\text{O}}_{\text{4}}^{\text{2-}}\text{+}{\text{O}}_{\text{2}}\text{+4}{\text{H}}^{\text{+}}$$

6

$$\:{\text{S}}_{\text{2}}{\text{O}}_{\text{8}}^{\text{2-}}\text{+}{\text{H}}^{\text{+}}\text{=H}{\text{S}}_{\text{2}}{\text{O}}_{\text{8}}^{\text{-}}$$

7

$$\:\text{H}{\text{S}}_{\text{2}}{\text{O}}_{\text{8}}^{\text{-}}\text{=S}{\text{O}}_{\text{4}}^{\text{2-}}\text{+S}{\text{O}}_{\text{4}}^{\text{•-}}\text{+}{\text{H}}^{\text{+}}$$

8

The results of the microcosm experiments shown in Table 6 indicate that higher initial concentrations of PS lead to greater decreases in pH. When the PS concentration was 20 g/L, the average pH values of A3, B3, and C3 decreased to 2.0, 6.0, and 3.7, respectively. The limestone contained in microcosm B3 acted as a buffer for pH changes. Similarly, in the column and conduit experiments involving limestone, the average pH of the aqueous solution was above 7.5, indicating weak alkalinity. The results suggest that higher concentrations of PS had a more significant impact on pH under static water conditions, but the presence of limestone, especially under flowing water, mitigated this effect of PS on pH.

Table 6

Variation in pH during the experiments
Microcosms				Column			Conduit
Group	Condition	Range	Mean	Group	Range	Mean	Group	Condition	Range	Mean
Control	A1	7.0 ~ 7.8	7.3		8.17 ~ 8.64	8.35		K1	7.8 ~ 8.0	7.9
	A2	1.8 ~ 6.7	3	5 g/L			5 g/L
	A3	1.4 ~ 3.5	2					K8	7.5 ~ 8.0	7.8
Limestone	B1	7.2 ~ 7.8	7.5
	B2	6.4 ~ 7.0	6.6		8.13 ~ 8.36	8.21		K1	8.0 ~ 8.2	8.1
	B3	5.7 ~ 6.5	6	10 g/L			10 g/L
	B4	7.3 ~ 7.9	7.5					K8	7.3 ~ 8.2	7.8
Lime soil	C1	6.4 ~ 7.5	7
	C2	3.6 ~ 7.0	4.7		8.06 ~ 8.28	8.19		K1	8.0 ~ 8.4	8.2
	C3	3.2 ~ 6.6	3.7	20 g/L			20 g/L
	C4	6.7 ~ 7.7	7					K8	7.7 ~ 8.2	8

2.3.2 DO

In the control and limestone groups of microcosms, the DO concentration significantly increased as the initial concentration of PS increased. When the initial concentration of PS reached 20 g/L, the DO concentrations in A3 and B3 increased, with peak values of 17.3 and 16.6 mg/L, respectively (Fig. 6), while the DO concentration in B4, without PS added, remained stable. The results confirmed that the decomposition of PS generated oxygen according to the chemical equation shown in Eq. 7(Liang et al., 2007). In microcosms C1 ~ C4 of the lime soil group, the oxygen demand by high OMC caused DO concentrations decreased to 2.9, 1.8, 3.1, and 2.5 mg/L at the end of the experiment, respectively.

Compared with those under static water conditions, there was no significant accumulated in the DO concentration under flowing water. DO concentrations ranged from 9.51 to 9.73 mg/L in the column experiment and from 7.21 to 8.12 mg/L in the conduit experiment, close to the background groundwater value.

Oxygen is the strongest electron acceptor in the aerobic biodegradation. The generation of oxygen during the decomposition of PS can enhance bioprocesses for the removal of organic contaminants (Guo et al., 2020).

2.3.4 Calcium and bicarbonate ions

The presence of PS caused significant changes in the concentrations of Ca²⁺ and HCO₃⁻ in microcosms B1 ~ B3 and C1 ~ C3 (Fig. 7). Large amounts of Ca²⁺ were produced due to acid accumulation, which enhanced the dissolution of CaO. For instance, the Ca²⁺ concentration in microcosm C3 reached 550.7 mg/L on day 1, which was ten times the background value in groundwater. The concentration of HCO₃⁻ also increased in microcosms B1 ~ B3 but decreased notably in microcosms C2 and C3. In cases where pH buffering was poor, such as in C1 ~ C3, carbonic acid mainly existed in its molecular form at low pH values, leading to a significant decrease in the HCO₃⁻ concentration. For example, the concentrations of HCO₃⁻ in C2 and C3 were close to 0 mg/L on days 7 and 3, respectively, reflecting the carbonic acid balance (Lin et al., 2023).

In contrast, Ca²⁺ and HCO₃⁻ remained relatively stable in the column and conduit experiments and had high pH buffering capacities under flowing water conditions.

2.3.5 Sulphate ions

The decomposition of PS can result in an increase in the SO₄²⁻ concentration, as demonstrated by chemical equations (1) to (6). The rate of increase is greater in water environments with greater OMC. For example, at a PS concentration of 20 g/L, the SO₄²⁻ concentration varied from 886.1 to 6805.7 mg/L for the limestone group and from 871.4 to 7871.7 mg/L for the lime soil group under static water conditions. OM in the lime soil group consumed more PS and generated more sulphate. However, the detectable SO₄²⁻ concentration in the column and conduit experiments was lower than 25 mg/L due to SO₄²⁻ loss caused by convection under flowing water.

In this study, the persistence of PS in karst aquifers, such as hydrostatic caves, karst fracture zones, and karst conduits, was investigated using microcosm, column, and conduit experiments, respectively. The conclusions were as follows:

(1) In static water environments, such as a hydrostatic karst cave filled with limestone or lime soils, PS was more persistent at higher concentrations. Compared with limestone, lime soil with a high OMC significantly decreased PS persistence. In flowing water, such as in karst fracture zones and karst conduits, PS persistence decreased greatly compared due to mass loss with flowing water.

(2) The factors influencing PS persistence in karst aquifers included advection and dispersion, OMC, type of aquifer material, and concentration of PS injected. Advection and high OMC in aquifers obviously decreased PS persistence. The presence of limestone increased PS persistence due to its pH buffering capacity.

(3) The decomposition of PS resulted in pH decrease, oxygen production, and salinity increase mainly related to SO₄²⁻, Ca²⁺, even and HCO₃⁻. The processes of pH buffering and advection inhibited the hydro-chemical responses.

Author contributions

All the authors significantly contributed to the study. Yudao Chen carried out the conception and a thorough critical review throughout the process. Xue Yan took charge of drafting the manuscript and analysed experiment data. Dongbo Tang and Liu Du helped to perform the column experiment. Weixuan Li and Wei Yang were indispensable in reviewing and checking the paper.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This study was supported by the Project of the National Natural Science Foundation of China (Grant Nos. 42367011).

Ahmad M, Teel AL, Watts RJ. (2013). Mechanism of persulfate activation by phenols. Environmental science & technology, 47 11, 5864-5871.
Bennedsen LR, Muff J, Søgaard EG. (2012). Influence of chloride and carbonates on the reactivity of activated persulfate. Chemosphere, 86(11), 1092-1097.
Chen WT. (2016). Remediation of diesel-contaminated soil using in situ chemical oxidation (ISCO) and the effects of common oxidants on the indigenous microbial community: a comparison study. Journal of Chemical Technology & Biotechnology, 91(6).
Devi P, Das U, Dalai AK. (2016). In-situ chemical oxidation: Principle and applications of peroxide and persulfate treatments in wastewater systems. Science of The Total Environment, 571, 643-657.
Ding X, Wei C, Wei Y, Liu P, Wang D, Wang Q, Chen X, Song X. (2024). Field test of thermally activated persulfate for remediation of PFASs co-contaminated with chlorinated aliphatic hydrocarbons in groundwater. Water Research, 249, 120993.
Eberle D, Ball R, Boving TB. (2017). Impact of ISCO Treatment on PFAA Co-Contaminants at a Former Fire Training Area. Environmental Science & Technology, 51(9), 5127-5136.
EPA US. (2002). The Qtracer2 Program For Tracer-Breakthrough Curve Analysis For Tracer Tests In Karstic Aquifers And Other Hydrologic Systems (2002).
Fang G, Chen X, Wu W, Liu C, Dionysiou DD, Fan T, Wang Y, Zhu C, Zhou D. (2018). Mechanisms of Interaction between Persulfate and Soil Constituents: Activation, Free Radical Formation, Conversion, and Identification. Environmental Science & Technology, 52(24), 14352-14361.
Furman OS, Teel AL, Watts RJ. (2010). Mechanism of Base Activation of Persulfate. Environmental Science & Technology, 44(16), 6423-6428.
Guo Y, Wen Z, Zhang C, Jakada H. (2020). Contamination and natural attenuation characteristics of petroleum hydrocarbons in a fractured karst aquifer, North China. Environmental Science and Pollution Research, 27(1).
Haizhou, Liu, Thomas, A., Bruton, Fiona, M., Doyle, David, L. (2014). In Situ Chemical Oxidation of Contaminated Groundwater by Persulfate: Decomposition by Fe(III)- and Mn(IV)-Containing Oxides and Aquifer Materials. Environmental Science & Technology, 48, 10330−10336.
Hao Y, Zhang J, Guo Y. (2020). The Characteristics of Karst Aquifers and the Variation of Karst Hydrological Processes in Northern China. Eurokarst 2018, Besançon, Cham.
Head NA, Gerhard JI, Inglis AM, Nunez Garcia A, Chowdhury AIA, Reynolds DA, de Boer CV, Sidebottom A, Austrins LM, Eimers J, O’Carroll DM. (2020). Field test of electrokinetically-delivered thermally activated persulfate for remediation of chlorinated solvents in clay. Water Research, 183, 116061.
Jiang X, Jiang Y, Chen Y, Xu K, Lu X, Zheng G. (2021). Stability of persulfate in karst conduit groundwater and its removal effect of aromatics. Chinese Journal of Environmental Engineering, 15(4), 1395-1402.
Jiang Y, Lu R, Chen Y, Deng R, Deng X. (2024). Effect of Fe²⁺-activated persulfate combined with biodegradation in removing gasoline BTX from karst groundwater: A box-column experimental study. Environmental Science and Pollution Research, 31(38), 50733-50745.
Li W, Orozco RG, Camargos NG, Liu H. (2017). Mechanisms on the Impacts of Alkalinity, pH, and Chloride on Persulfate-Based Groundwater Remediation. Environmental Science & Technology, 51(7), 3948-3959.
Liang C, Guo Y-Y. (2012). Remediation of Diesel-Contaminated Soils Using Persulfate Under Alkaline Condition. Water, Air, & Soil Pollution, 223, 4605-4614.
Liang C, Wang Z-S, Bruell CJ. (2007). Influence of pH on persulfate oxidation of TCE at ambient temperatures. Chemosphere, 66(1), 106-113.
Liang C, Wang Z-S, Mohanty NR. (2006). Influences of carbonate and chloride ions on persulfate oxidation of trichloroethylene at 20 °C. Science of The Total Environment, 370, 271-277.
Liang C, Yin S, Huang P, Yang S, Wang Z, Zheng S, Li C, Sun Z. (2024). The critical role of minerals in persulfate-based advanced oxidation process: Catalytic properties, mechanism, and prospects. Chemical Engineering Journal, 482, 148969.
Lin K, Yu T, Ji W, Li B, Wu Z, Liu X, Li C, Yang Z. (2023). Carbonate rocks as natural buffers: Exploring their environmental impact on heavy metals in sulfide deposits. Environmental Pollution, 336, 122506.
Liu B, Huang B, Wang Z, Tang L, Ji C, Zhao C, Feng L, Feng Y. (2023). Homogeneous/heterogeneous metal-catalyzed persulfate oxidation technology for organic pollutants elimination: A review. Journal of Environmental Chemical Engineering, 11(3), 109586.
Liu H, Bruton TA, Li W, Buren Jv, Prasse C, Doyle FM, Sedlak DL. (2016). Oxidation of Benzene by Persulfate in the Presence of Fe(III)- and Mn(IV)-Containing Oxides: Stoichiometric Efficiency and Transformation Products. Environmental Science & Technology, 50(2), 890-898.
Liu T, Yao B, Luo Z, Li W, Li C, Ye Z, Gong X, Yang J, Zhou Y. (2022). Applications and influencing factors of the biochar-persulfate based advanced oxidation processes for the remediation of groundwater and soil contaminated with organic compounds. Science of The Total Environment, 836, 155421.
Mustapha NA, Liu H, Ibrahim AO, Huang Y, Liu S. (2021). Degradation of aniline in groundwater by persulfate with natural subsurface sediment as the activator. Chemical Engineering Journal, 417, 128078.
Nguyen T-M, Chen H-H, Chang Y-C, Ning T-C, Chen K-F. (2023). Remediation of groundwater contaminated with trichloroethylene (TCE) using a long-lasting persulfate/biochar barrier. Chemosphere, 333, 138954.
Sakr M, El Agamawi H, Klammler H, Mohamed MM. (2023). A review on the use of permeable reactive barriers as an effective technique for groundwater remediation. Groundwater for Sustainable Development, 21, 100914.
Sra KS, Thomson NR, Barker JF. (2010). Persistence of Persulfate in Uncontaminated Aquifer Materials. Environmental Science & Technology, 44(8), 3098.
Sra KS, Thomson NR, Barker JF. (2013). Persulfate injection into a gasoline source zone. Journal of Contaminant Hydrology, 150C(5), 35-44.
Sra KS, Thomson NR, Barker JF. (2013). Persulfate Treatment of Dissolved Gasoline Compounds. Journal of Hazardous, Toxic, and Radioactive Waste, 17(1), 9-15.
Sutton N, Langenhoff A, Lasso DH, Zaan BMVD, Gaans PVV, Maphosa F, Smidt H, Grotenhuis T, Rijnaarts H. (2013). Recovery of microbial diversity and activity during bioremediation following chemical oxidation of diesel contaminated soils. Applied Microbiology and Biotechnology, 98, 2751 - 2764.
Teel AL, Ahmad M, Watts RJ. (2011). Persulfate activation by naturally occurring trace minerals. Journal of Hazardous Materials, 196, 153-159.
Teel AL, Elloy FC, Watts RJ. (2016). Persulfate activation during exertion of total oxidant demand. Chemosphere, 158, 184-192.
Tsitonaki A, Smets BF, Bjerg PL. (2008). Effects of heat-activated persulfate oxidation on soil microorganisms. Water Research, 42(4), 1013-1022.
Wang H, Chen YD, He LW, Jiang YP, Xia Y, Yang PF. (2022). Feasibility of nitrate reduction combined with persulfate oxidation in the remediation of groundwater contaminated by gasoline (Nov, 10.1007/s10040-021-02417-w, 2021). HYDROGEOLOGY JOURNAL, 30(1), 327-327.
Wang Y, Tian H, Zhang C, Xu J, Liu X, Ma F, Wei X, Sun Y. (2024). Degradation and mechanism of PAHs by Fe-based activated persulfate: Effect of temperature and noble metal. Science of The Total Environment, 931, 172768.
Watts RJ, Ahmad M, Hohner AK, Teel AL. (2018). Persulfate activation by glucose for in situ chemical oxidation. Water Research, 133, 247-254.
Wei J, Jiang Y, Chen Y, Jian L. (2020). Study on removal of ethanol in groundwater by persulfate. Environmental Pollution & Control, 42(8), 953-958.
Wu P, Tang C, Zhu L, Liu C, Cha X, Tao X. (2009). Hydrogeochemical characteristics of surface water and groundwater in the karst basin, southwest China. Hydrological Processes, 23(14), 2012-2022.
Wu S, Chen Y, Wei Y, Meng W. (2019). Effects of Persulfate to Oxidize BTEX in Groundwater and Hydrochemical Responses. Environmental Science & Technology, 42(6), 76-81.
Zawadzki P. (2022). Persulfate activation by organic compounds: advancements and challenges. Current Opinion in Chemical Engineering, 37, 100837.
Zheng G, Jiang Y, Chen Y, Zhou Z, Qin J, Yang P. (2022). Effects of degradation of gasoline BTEX and influences of ethanol by persulfate in calcareous soil medium. Acta Scientiae Circumstantiae, 42(7), 217-224.
Zhu M, Liang M, Liu C, Jin Z, Li Q. (2018). Influential Factors and Indicative Significance of Microbial Abundance in Karst Limestone Soil. Ecology and Environmental Sciences, 27(03), 484-490.
Zhu Z, Wang J, Hu M, Jia L. (2019). Geographical detection of groundwater pollution vulnerability and hazard in karst areas of Guangxi Province, China. Environmental Pollution, 245, 627-633.

No competing interests reported.

Persistence of persulfate in karst groundwater and its influencing factors

Status:

Version 1

Abstract

Figures

Introduction

1 Materials and methods

1.1 Experimental materials

1.2 Experimental methods

1.3 Sample analysis

2 Results and discussion

2.1 Decomposition of PS

2.2 Influencing factors of PS persistence

2.3 Hydro-chemical responses

2.3.2 DO

3 Conclusions

Declarations

References

Additional Declarations

Status:

Version 1