Joint effects of gas bubbles and solid particles on sonochemical inhibition in sonicated aqueous solutions Authors: Kunpeng Su, Binghui Li, Jianhua Wu, Pei Xin, Shangtuo Qian Published: 2023 ABSTRACT Highlights•Sonochemistry of gas–solid–liquid mixtures was experimentally investigated.•Gas bubbles and solid particles are jointly responsible for sonochemical inhibition.•Ultrasound attenuation caused by the gas and solid phases is the underlying cause.Keywords: Sonochemical effect, Solid particles, Gas bubbles, Ultrasound attenuation Introduction Since early research in sonochemistry, power ultrasound application has been a subject of growing interest in wastewater treatment and environmental remediation areas [1], [2], [3]. When cavitation bubbles generated by ultrasound oscillate and collapse, hydroxyl radicals (⋅OH) are produced owing to the ultrahigh temperature and pressure inside the bubbles [4]. These highly reactive species are crucial in the oxidative destruction of non-biodegradable organic pollutants in the aqueous phase. Ultrasound treatment is a simple technique, requiring only electrical energy; however, scaling up requires the knowledge of all possible factors that influence sonoreactions [5], [6], [7]. For example, besides organic compounds, wastewater also contains inorganic contaminants such as sediment and microplastics [8], as well as undissolved gas bubbles [9]. These rigid particles and deformable bubbles may influence the efficiency of the sonochemical reactors that operate prior to biological (secondary) treatment [10], [11], as primary treatment removes only ∼60% of suspended solids [12] and biological treatment requires aeration processes [13]. Therefore, particular attention must be given to the sonochemical effects in gas–solid–liquid mixtures. Several studies have investigated sonochemical activity in solid particle suspensions [14], [15], [16], [17], [18], [19], [20], [21], [22] and in bubbly liquids [23], [24], [25], [26] and are mostly concerned with how the quality and quantity of these solid/gas phases influence bubble characteristics and sonoreactions. Surface roughness, size, and shape are the most-studied indicators of particle quality. Irregularly shaped particles with rough surfaces may improve ⋅OH yields [27], according to the crevice model [28]. Large particles can induce asymmetric bubble collapse and form high-speed micro-jets that impact the surface, whereas small ones may be propelled by the shock waves from bubble collapse and collide with one another [29], [30], [31], [32]. In both cases, defects form on the surface and may act as nucleation sites. In the presence of the gaseous phase, the intrinsic nature of gases (e.g., polytropic index, solubility, and chemical reactivity) has been recognized as a major factor influencing sonochemistry. Monatomic gases can raise the temperature within the collapsing bubbles owing to high polytropic index [33], whereas diatomic gases such as oxygen and nitrogen can directly participate in reactions that promote or prevent radical production [34]. By cushioning bubble collapse with enhanced gas diffusion [35] and preventing bubble coalescence with decreased pH values [36], carbon dioxide inhibits radical production. Previous research has also demonstrated the sequence of radical production among different gases [37], [38] or gas mixtures [39], [40]. The quantity of the solid/gas phases exerts a decisive influence besides the physical and chemical qualities. Katekhaye and Gogate [41] studied the sonochemical activity in 0.2–0.6-vol% suspensions of titania powder and reported higher ⋅OH yields than in water; the yields increased with solids fraction. They ascribed this to additional nucleation sites, as well as titania’s catalytic activity [42]. Her et al. [43] found that ⋅OH yields in 3–12-vol% glass bead suspensions were lower than those in water, and decreased with increasing solids fraction, which tallied with the results of other studies [44], [45]. In more concentrated suspensions, gas/vapor diffusing through the bubble wall would be harder, which would in turn make it harder for bubbles to oscillate or collapse [46]. Tuziuti et al. [47] reported an improved sono-oxidation rate in the presence of micron-sized air bubbles. They suggested that further dissolved air was the probable cause, which led to additional cavitation bubbles [48]. Choi et al. [49] investigated the influence of the air bubbling rate on sonodegradation and discovered that ⋅OH yields first increased and subsequently decreased with increasing air bubbling rates (0–16 L⋅min−1). Gogate et al. [50], [51] and Xia et al. [52] also observed similar trends. Cushioned bubble collapse and increased compressibility were deemed responsible for reduced cavitation intensity with higher air concentrations [53], [54]. Although examining how the quantities of the solid and gas phases affect sonoreactions in gas–solid–liquid mixtures is important, the combined effects have rarely been studied. The consensus regarding the mechanism is that gas bubbles and solid particles can bring about points of weakness in the liquid continuum, lowering the cavitation threshold and leading to additional radical production. The only difference is that the cavitation nuclei in bubbly liquids are directly the free tiny bubbles [52], whereas the ones that solid particles cause are the undissolved gas adhering to the particle surface, e.g., in a cleft or crevice [28]. However, consensus regarding why the gas and solid phases inhibit sonoreactions in certain circumstances is lacking. Extensive evidence has demonstrated the attenuation of acoustic waves that propagate through solids suspensions [55], [56], [57] or bubbly liquids [58], [59], [60]. This could be a rational explanation since reduced ultrasound intensity makes it harder for cavitation inception to occur and bubbles to form. Although various attenuation regimes in the presence of gas or solid phases have been identified, limited studies touch upon the relation between sonochemical inhibition and ultrasound attenuation in gas–solid–liquid mixtures. This study aimed to examine the sonochemically inhibiting effects of the gas and solid phases and clarify the underlying mechanisms. Sonochemical experiments were performed in aqueous solutions containing argon gas bubbles and silica particles and the experimental results were analyzed using an integrated ultrasound attenuation model. Materials and methods Materials Analytical grade potassium iodide (KI), luminol (C8H7N3O2), and potassium hydroxide (KOH) were acquired from Beilian Fine Chemical Development Co., Ltd. (Tianjin, China), Haohong Biomedical Technology Co., Ltd. (Shanghai, China), and Jinshan Chemical Reagent Co., Ltd. (Chengdu, China), respectively. All chemicals were used as received. Argon, the cheapest noble gas, was selected as the gas phase owing to its ability to produce higher sonochemical activity than polyatomic gases [5]. Cylinders storing high-purity grade argon were supplied by Qiaopai Technology Development Co., Ltd. (Nanjing, China). Silica particles with a uniform diameter of 0.1 mm (density: 2.4 × 103 kg·m−3) were used as the solid phase, provided by Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan). The monodispersity in silica particle size was confirmed using a scanning electron microscope (Regulus 8100, Hitachi High-Tech Corp., Japan), as shown in Fig. 1. Image of monodisperse silica particles captured using a scanning electron microscope. Experimental setup A 20-kHz horn-type reactor equipped with a piezoelectric transducer (VCY−1500, Shanghai Y&Y Sonic) was employed (Fig. 2). Ultrasound was generated with an applied electric power of 250 W, and the absorbed acoustic power was approximately 135 W (measured using the thermal probe method [61]). The ultrasonic horn, made of Ti-6Al-4V titanium alloy, was placed erect at the center of a cylindrical PET vessel (inner radius: 6.0 cm). The horn tip surface was submerged 5.0 cm beneath the free surface of the test solution, transferring the acoustic energy from the transducer into the solution. The solution temperature was monitored using an alcohol thermometer and maintained at 25 ± 1 °C by positioning the vessel in a circulating water bath as horn heating was inevitable. Schematic of the experimental setup. Key: 1: ultrasonic generator, 2: piezoelectric transducer, 3: ultrasonic horn, 4: output tip, 5: fine bubble diffuser, 6: thermometer, 7: gas cylinder, 8: argon (inside), 9: pressure-reducing regulator, 10: rotameter, 11: cooling bath, 12: test solution containing argon gas bubbles and/or silica particles, and 13: supporting rod. A disc-shaped fine-pore ceramic diffuser was installed at the bottom of the vessel to continuously supply argon to the test solution, producing a plethora of rising gas bubbles (Fig. 3a–d). The volume of gas bubbles in the solution (Vg) at a specific point in time could be calculated based on the rise of the liquid surface caused by aeration (Δh), i.e., Vg = SΔh, where S is the vessel’s cross-sectional area. For example, Vg ≈ 11.3, 22.6, 33.9, 45.2, and 56.5 mL when Δh = 0.1, 0.2, 0.3, 0.4, and 0.5 cm, respectively. Volumetric gas flow rates (Qg) were regulated using a pressure-reducing valve (F60–0085, Foshan Feibao Tools) and measured using a rotameter (LZB–3WB, Hangzhou Darhor Tech.). Vg could be controlled during the tests by altering Qg owing to the linear correlation between Vg and Qg (Fig. 4). The average diameter of the gas bubbles was approximately 1.2 mm under typical operating conditions, determined from the images recorded by a high-speed camera (Phantom VEO 1010, Vision Research) at 10,000 fps. Owing to ultrasonic agitation, the gas bubbles were more evenly distributed in the sonicated solution (Fig. 3e) than in the solution without sonication (Fig. 3b). Images showing the rise of gas bubbles at different gas flow rates with no solids or ultrasound: (a) Qg = 0.1 L·min−1, (b) Qg = 0.4 L·min−1, and (c) Qg = 1.0 L·min−1 after being released from (d) a ceramic media diffuser (effective aeration diameter: 10.0 cm); and gas bubbling at Qg = 0.4 L·min−1 under sonication (e) without solids and (f) with 50.0-g suspended silica particles. Variation of gas bubble volume with the gas flow rate. The error bars represent one standard deviation. To avoid floc clogging of the gas-release orifices in the porous ceramics, routine removal of fouling materials was required, and the operation of the aeration system should be started before adding silica. The silica particles were observed to remain in suspension in sonicated aerated solutions (Fig. 3f) thanks to the downward forces caused by ultrasonic agitation and upward forces due to diffused aeration. The suspended particles moved vigorously in a random manner owing to the intense mixing. Thus, it seems reasonable to assume that the silica particles were well distributed throughout the vessel. The interaction between bubbles and particles, as well as their mutual influence in terms of size distribution and movement, was neglected, since the gas bubbles were supplied continuously and much larger in size. The volume of silica particles in the solution (Vs) was calculated as Vs = ms/ρs, in which ms and ρs are the mass and density of silica particles, respectively. An electronic balance was used to weigh the particles. All particles were unrecycled after every test completion, due to the inevitable reduction in the particle size that sonication might induce [32]. Measuring sonochemical yield The KI dosimetry method was used to quantify the amount of ⋅OH generated by acoustic cavitation. In this method, ⋅OH oxidized a fraction of iodide ions (I−) to generate molecular iodine (I2), and the remaining I− reacted with I2 to form the triiodide ion (I3−). The maximum absorption of I3− was at a wavelength of 355 nm [62], and the adsorption of I3− on the surfaces of the vessel, diffuser, and silica particles was neglected. Just before each test, aqueous KI solutions (0.1 mol·L−1) were freshly prepared using distilled water (air-saturated), with the basicity adjusted to a pH of 8 using KOH. Despite the initial presence of dissolved air in the solution, the effect of dissolved oxygen concentration on I3− production was neglected because of the continuous argon gas bubbling and the degassing effect induced by ultrasonic irradiation [63]. The duration of each test was 25 min. An aliquot of solution (6–7 mL) containing minimal bubbles and particles was sampled at regular intervals to establish ⋅OH yield versus time curves. It was centrifuged at 2500 rpm for 2 min and left undisturbed for 5 min. The supernatant was subsequently transferred into quartz cuvettes for spectrophotometric analysis, using an ultraviolet–visible spectrophotometer (UV754N, Shanghai Aucy Scientific Instrument). The I3− concentration generated, c(I3−), was determined according to the Beer–Lambert law: c(I3−) = A/(εw), where A is the measured absorbance, ε is the molar attenuation coefficient (ε = 23200 L∙mol−1∙cm−1), and w is the width of each cuvette (w = 1.0 cm). The procedure was repeated throughout the test at 5 min increments. Each test was repeated three times, and the mean values were used in the subsequent analyses. To visualize the sonochemically active zone under the argon gas bubbling conditions, the sonochemiluminescence (SCL) method was also utilized. SCL images were obtained using an aqueous solution of luminol (1 mmol·L−1) and KOH (0.1 mol·L−1), captured with a digital camera (D3400, Nikon) in total darkness. The shutter speed was fixed at 30 s (exposure time). Experimental cases For three-phase gas–solid–liquid mixtures, the gas volume fraction, φg, and the solids fraction, φs, are defined as follows: where Vg, Vs, and Vl are the respective volumes of the gas, solid, and liquid phases, respectively. In this study, Vl was unchanging (Vl = 1.1 L) as the distance from the initial surface of the additive-free solution to the ceramic diffuser was always set to 9.6 cm. The fraction of the total volume that is made up of gas bubbles and solid particles, φtotal, and the gas-to-solid ratio, λφ, i.e., the ratio of φg to φs, are the other two parameters of interest. Volume fraction of the liquid phase is denoted by φl = 1 − φtotal. Twenty-three experimental cases were performed, including ten three-phase cases in which gas bubbles and solid particles coexisted (Table 1), twelve two-phase cases with only gas bubbles (φg = 0.5, 1.0, 1.5, 2.0, 3.0, or 5.0 vol%) or only solid particles (φs = 0.5, 1.0, 1.5, 2.0, 3.0, or 5.0 vol%), and a single-phase case (pure water) for reference. Through the three-phase cases, we investigated the effects of the total volume fraction (Cases 1–4: φtotal = 2.0–6.0 vol%, λφ = 1:1) and the gas-to-solid ratio (Cases 2, 5, and 6: φtotal = 3.0 vol%, λφ = 1:2 to 2:1; Cases 3, 7, and 8: φtotal = 4.0 vol%, λφ = 1:3 to 3:1; and Cases 4, 9, and 10: φtotal = 6.0 vol%, λφ = 1:5 to 5:1). The Qg values for all the cases listed in Table 1 were determined by interpolation based on known data points in Fig. 4, owing to the strong positive correlation between Vg and Qg. Table 1Experimental cases (gas–solid–liquid mixtures).Case No.φga (vol%)φsb (vol%)φtotalc (vol%)λφdVge (mL)Qgf (L·min−1)Vsg (mL)msh (g)11.01.02.01:111.10.21211.126.621.51.53.01:116.80.31516.840.332.02.04.01:122.60.41922.654.343.03.06.01:134.70.63634.783.251.02.03.01:211.20.21422.453.762.01.03.02:122.40.41511.226.971.03.04.01:311.30.21633.981.483.01.04.03:133.90.62311.327.191.05.06.01:511.60.22057.8138.6105.01.06.05:157.81.05111.627.7aφg—volume fraction of the argon gas bubbles.bφs—volume fraction of the silica particles.cφtotal—total volume fraction of the argon gas bubbles and silica particles (sum of φg and φs).dλφ—gas-to-solid ratio (ratio of φg to φs).eVg—volume of the argon gas bubbles.fQg—volumetric gas flow rate.gVs—volume of the silica particles.hms—mass of the silica particles. Experimental cases (gas–solid–liquid mixtures). φg—volume fraction of the argon gas bubbles. φs—volume fraction of the silica particles. φtotal—total volume fraction of the argon gas bubbles and silica particles (sum of φg and φs). λφ—gas-to-solid ratio (ratio of φg to φs). Vg—volume of the argon gas bubbles. Qg—volumetric gas flow rate. Vs—volume of the silica particles. ms—mass of the silica particles. Results I3− production linearly increased with ultrasonic irradiation time in the test solution containing either silica particles or argon gas bubbles, as well as in the single-phase case (Fig. 5), which is well aligned with the results of past studies [45], [52]. This indicates that the formation rate of I3−, rf(I3−), defined as the increase in I3− concentration per unit time, was independent of the irradiation time (≤25 min). In both kinds of two-phase mixtures, the sonochemical yields after ultrasonic irradiation for 25 min attained the smallest value at the highest φs or φg (5.0 vol%) and the largest value at the lowest φs or φg (0.5 vol%), where more I3− was detected at every time point compared with the pure water scenario. A one-quarter reduction in I3− production was observed at 25 min with the solid or gas phase being added from 0.5 vol% to 5.0 vol%, reflecting the sonochemically inhibiting effects of these additives. I3− concentration plotted against sonication time for two-phase mixtures containing (a) 0.1-mm silica particles (volume fraction φs = 0.5–5.0 vol%), and (b) 1.2-mm argon gas bubbles (volume fraction φg = 0.5–5.0 vol%), as well as for the aqueous solution. Error bars are shown when they are larger than the data point markers. Despite the similarity in reaction inhibition at high φs or φg, considerable differences existed between the silica particles and argon gas bubbles in influencing sonochemistry. Adding argon gas bubbles at φg = 0.5 vol% generated 17% more I3− than that in the pure water case (Fig. 5b) and 8% more than that in the 0.5-vol% suspensions of the silica particles (Fig. 5a). This implies that the 1.2-mm gas bubbles had stronger enhancing capability than the 0.1-mm solid particles when φs = φg, even though the number of particles was considerably greater than that of the gas bubbles. The main reason may be the different nuclei-introducing abilities of the two phases. Additionally, silica particle introduction at φs = 1.5 vol% started to inhibit I3− formation, which was slightly less than that in the pure water case (Fig. 5a). Contrastingly, the argon gas bubbles had an overall enhancing effect on I3− formation when φg = 1.5 vol% (Fig. 5b). The inhibitory effect of the gas bubbles began to be dominant merely when φg was doubled to 3.0 vol%, demonstrating the strong capability of the gas bubbles in nucleus introduction and sonochemical enhancement. Although the sonochemically enhancing mechanisms of the gas bubbles and solid particles differ, their inhibitory effects at high volume fractions exhibited some similarities, implying similar mechanisms in reducing sonochemical yields. The inhibitory effect of gas bubbles was also visualized using the luminol method (Fig. 6). The SCL image under the non-aeration condition (Fig. 6a) was similar to the findings reported by Son et al. [64]. High-intensity SCL light was detected beneath the horn tip, with lower-intensity light shining above the vessel bottom and around the immersed horn. Interestingly, compared with the non-aeration condition, the region under the horn got darker when φg = 0.5 vol% (Fig. 6b). This may be because the large gas bubbles rising from the bottom induced mixing and modified the local sound energy field [25], inhibiting ultrasound transmission and resulting in limited active zones. The higher SCL intensity in the circular zone around the horn (Fig. 6b), however, might result from the introduction of additional nuclei. Although I3− production at φg = 0.5 vol% was higher than that in the pure water scenario (Fig. 5b), no significant increase in the total SCL intensity was observed, possibly due to the disturbance the rising gas bubbles caused to the capturing of emitted light [49]. As φg further increased, a decrease in the intensity of SCL light was noticed in the active zones (Fig. 6c and d), demonstrating the sonochemically inhibiting effect of the gas bubbles. Sonochemiluminescence images showing sonochemically active zones (a) without aeration, and at different gas volume fractions: (b) φg = 0.5 vol%, (c) φg = 2.0 vol%, and (d) φg = 5.0 vol%. The dashed line represents the location of the vibrating horn. We can first adopt an effective-medium approach to analyze the sonochemical effects in gas–solid–liquid mixtures. This implies assuming the solid–liquid mixture as a continuous medium, where gas bubbles are additives, or considering the gas bubbles and the liquid as a whole with solid particles as additives. In Fig. 7(a) and (b), the bubbly solution with φg = 1.0 vol% and the 1.0-vol% solid particle suspension were treated as effective media, respectively. Data in both figures follow the same “first increasing then decreasing” pattern of rf(I3−) against the volume fraction of the additives. This result is extremely similar to the results of the two-phase cases (Fig. 5), and the only difference is that the gas/solid phase in the effective medium altered the nucleation characteristics of the medium. For instance, adding argon gas bubbles with φg from 1.0 to 5.0 vol% into the silica suspension (φs = 1.0 vol%) caused a less sharp decrease in rf(I3−) compared to when silica particles were added into the bubbly solution (φg = 1.0 vol%) with φs from 1.0 to 5.0 vol%. This indicates that the gas phase in the effective medium performed better in introducing nuclei, enhancing sonochemical activity, and offsetting the inhibitory effect of the additives. Comparison of the I3− formation rates among the three-phase mixtures with (a) the same gas volume fraction (φg = 1.0 vol%) but different solids fractions (φs = 0–5.0 vol%); (b) the same φs (1.0 vol%) but different φg (0–5.0 vol%); (c) equal proportions of argon and silica (λφ = 1:1) but different total volume fractions (φtotal = 2.0–6.0 vol%); and the same φtotal but different λφ: (d) φtotal = 3.0 vol%, λφ = 1:2 to 2:1, (e) φtotal = 4.0 vol%, λφ = 1:3 to 3:1, and (f) φtotal = 6.0 vol%, λφ = 1:5 to 5:1. The dashed line represents the absence of silica particles and argon gas bubbles. Fig. 7(c)–(f) shows the experimental data for the three-phase cases in terms of different φtotal or λφ. Of all the three-phase cases in Table 1, the highest and lowest I3− formation rates occurred in Cases 1 and 4, respectively. When the volume ratio of the gas phase and the solid phase was equal (λφ = 1:1), values of rf(I3−) at φtotal = 2.0 and 3.0 vol% was larger than that in the case with only the liquid phase, whereas at high φtotal, I3− was produced at a low rate (Fig. 7c). The marked decline in rf(I3−) with increasing φtotal from 2.0 to 6.0 vol% demonstrates the combined effects of gas and solid phases in inhibiting sonoreactions at high φg and φs. A clear distinction existed between the values of rf(I3−) at φtotal = 3.0 vol% (Fig. 7d) and those at φtotal = 4.0 and 6.0 vol% (Fig. 7e and f) with various λφ. When φtotal = 3.0 vol%, the sonochemical effects were enhanced at all λφ considered, and a higher proportion of gas bubbles seems to have improved enhancing effects. Contrastingly, at φtotal = 4.0 and 6.0 vol%, I3− formation was inhibited in all cases, with the largest rf(I3−) values at λφ > 1 and smallest at λφ = 1:1 (Fig. 7f). The inhibition mechanism will be discussed in the following section. Discussion The inhibition in sonochemical activity caused by solid particles and gas bubbles can be attributed to several possible mechanisms: (a) the addition of solid particles increased the apparent viscosity of the mixture, which could in turn lead to increased bubble oscillation time [65] and the formation of larger bubbles due to coalescence [66], thereby retarding the cavitation rates. (b) When the non-condensable argon gas diffused into cavitation bubbles, it could strongly influence the bubble dynamics, e.g., by decelerating the bubble collapse, weakening the shockwaves emitted, and reducing the intensity of cavitation [67]. This cushioning effect also accounts for the alleviation of cavitation damage to dam spillways by installing aerators [53], [54]. (c) The presence of large amounts of solid particles and gas bubbles in the immediate vicinity of the ultrasonic horn (especially near the horn tip surface) could produce decoupling losses, minimizing the transfer of energy into the solution and decreasing the energy for the cavitation events [50]. (d) When ultrasound propagated in gas–solid–liquid mixtures, the gas and solid phases could also result in increased energy dissipation, attenuating the ultrasound and leaving even less energy available for acoustic cavitation [45], [52]. Among the possible mechanisms, the first one was neglected in this study as our φs values were far lower than 40–50 vol%, a range in which Stoian et al. [27] reported greatly enhanced slurry viscosity and decreased sonochemical yields with increasing φs. As regards the second and third ones, the cushioning and decoupling effects were somewhat limited, because the mass of argon gas that diffused into cavitation bubbles was not directly related to the quantity of rising gas bubbles, and the localized decoupling loss made up only a small proportion of the total energy loss. To understand the fourth mechanism, the acoustics of gas–solid–liquid mixtures should be examined at different concentrations of the gas and solid phases, from an energy perspective. The removal of acoustic energy in heterogeneous media is generally accepted to be caused by viscosity, scattering, and heat conductivity. Among the three main reasons, ultrasound attenuation in solid particle suspensions is primarily induced by scattering [55], [56], whereas thermal conduction plays the most important role in bubbly liquids [68], [69]. Given the different regimes, we consider the total attenuation coefficient of the gas–solid–liquid mixtures, αtotal, as the combination of the attenuation due to the solid particles, αs, and that due to the gas bubbles, αg, i.e., Based on the one-dimensional two-phase hydrodynamic equations and linear perturbation analysis, Atkinson and Kytömaa [57] modeled sound propagation in monodisperse sphere suspensions and obtained a seemingly complex formula for αs. With further algebraic simplification, we can reduce their formula to: in which f is the ultrasound frequency; ρs and ρl are the densities of the solid particles and the liquid, respectively; δs = [μl/(πfρl)]1/2 is the viscous boundary layer thickness of the solid particles, where μl is the liquid viscosity; Rs is the radius of solid particles; ζ = φl/2 is a simplified added mass coefficient; and the quantities k¯, ρ¯, and ρ* are given byk-1¯ = φsκs−1 + φlκl−1, ρ¯ = φsρs + φlρl, and ρ* = φlρs + φsρl, respectively, with κs and κl being the bulk moduli of the solid particles and the liquid, respectively. The model requires no viscous interactions between particles, which is satisfied in the limit that δ ≪ h/2, where h = [(φs/φm)−1/3 –1]Rs is the average interparticle spacing, and φm ≈ 64 vol% is the maximum packing fraction for monodisperse spheres [70]. As δs ≈ 0.1 μm and h/2 ranges from 33 μm to 101 μm in the current study, the viscous boundary layers that surround the adjacent particles cannot overlap, and the model can be used. In contrast to the limited variation of the sound speed in the solid particle suspensions, the speed of sound can be considerably lower in bubbly liquids than in either gas or liquid alone [71]. Brennen [72] simplified the phase velocity of ultrasound in the bubbly liquid, c, as in which γg is the heat capacity ratio for gas; pa is the ambient pressure; and ρg is the density of the gas in the bubble. Based on Eq. (7), we can rewrite Silberman’s formula for αg [58] as in which ρm = φgρg + φlρl is the mean density of the bubbly liquid; f* = f/fR, where fR is the bubble resonant frequency given by the Minnaert formula fR = (3γgpa/ρm)1/2/(2πRg) [73], with Rg being the bubble radius; and δ* = δf*2, where δ is the total damping constant of gas bubbles, which is the sum of the scattering term δr, the viscous damping term δv, and the thermal damping term δt. We assume that δ ≈ δt in Eq. (8) because the influence of thermal conduction is dominant. Eller [74] developed the following expression for δt based on the work by Devin [75]: in which cpg and Kg are the specific heat capacity (at constant pressure) and thermal conductivity of gas, respectively. As Domenico [59] opined, the model applies only to gas bubbles of Rg > 0.15 mm and far less than the sound wavelength. The gas bubble size in this study satisfies the requirements. The calculations of the total attenuation using Eq. (3) are based on the assumption that the mutual interaction between the gas bubbles and the solid particles is insignificant, and the bubbles and particles advect with the ambient liquid velocity (no slip). This implies that, when calculating αtotal, the volume fractions of separate phases in the expressions mentioned above should be revised by neglecting the presence of a third phase, i.e., replacing φs and φl in expressions for αs by φs/(1 – φg) and (1 − φtotal)/(1 – φg), and φg and φl in expressions for αg by φg/(1 – φs) and (1 − φtotal)/(1 – φs), respectively. Such revisions do not contradict the effective-medium approach when we neglect the bubble–particle interaction. It is reasonable to hypothesize, based on Fig. 3, that both the gas bubbles and the silica particles were uniformly sized and distributed in the solution in the calculation of αtotal. Eqs. (3), (4), (5), (6), (7), (8), (9), (10) show that αtotal is dependent on φg and φs. Fig. 8 illustrates the relationships between acoustic attenuation and additive volume fractions in a filled contour map. At a fixed λφ, the increase in φtotal results in larger values of αtotal, and consequently, a decline in rf(I3−) (Fig. 7c). αtotal increases with φs when the bubbly water is deemed as an effective medium with φg being constant. Similarly, αtotal increases with φg when φs is unchanged. However, the amount of increase in αtotal with the same increments of additive addition is different in the two circumstances. For example, the increase in αtotal caused by an increase in φs from 1.0 vol% to 5.0 vol% at φg = 1.0 vol% is 9% bigger than that caused by the same amount of increase in φg at φs = 1.0 vol%. This may explain why the decrease in rf(I3−) was sharper in Fig. 7(a) than in Fig. 7(b). When φtotal is kept constant (>3.0 vol%), the variation of αtotal with λφ is non-monotonic. As λφ increases, αtotal first increases and then decreases, and such a trend is more pronounced at high φtotal. Among Cases 4, 9, and 10 in which φtotal = 6.0 vol%, the lowest value of αtotal appears at λφ = 5:1, which can explain the highest I3− formation rate in Fig. 7(f). Total acoustic attenuation coefficient (αtotal) as a function of the total volume fraction and gas-to-solid ratio in three-phase mixtures, calculated using the following values (at 25 °C): (1) for/in water: ρl = 1.0 × 103 kg⋅m−3, μl = 0.89 mPa·s, κl = 2.2 GPa, and c = 1.5 × 103 m⋅s−1; (2) for argon gas (bubbles): Rg ≈ 0.6 mm, ρg ≈ 1.6 kg⋅m−3, γg = 5/3, cpg = 520.3 J⋅kg−1⋅K−1, Kg = 0.016 W·m−1·K−1, and pa ≈ 1.0 × 105 Pa; and (3) for silica particles: Rs = 50 μm, ρs = 2.4 × 103 kg·m−3, and κs = 42.5 GPa. The V-shaped dashed line represents acoustic attenuation at φg = 1.0 vol% (left) and φs = 1.0 vol% (right), and the markers correspond to the experimental cases in Table 1. Table 2 lists the calculation results of αg, αs, and αtotal for all three-phase cases in Table 1. To nondimensionalize the measured formation rates of I3− in these cases, we define a percentage change in rf(I3−), β, as follows: in which rfo(I3−) is the formation rate of I3− in a selected case for reference. The case for reference can be the single-phase case in which silica particles and argon gas bubbles (both treated as additives in the three-phase case) were absent. It can also be a two-phase case if the effective-medium approach is adopted, regarding the third phase as the additive. The parameter β can be either positive (sonochemically enhancing) or negative (sonochemically inhibiting), with a value of zero representing no difference in sonochemical effects in comparison with the case for reference. Table 2Calculated attenuation coefficients (gas–solid–liquid mixtures).Case No.φga (vol%)φsb (vol%)αgcαsdαtotale11.01.00.1270.0510.17821.51.50.1580.0770.23432.02.00.1840.1030.28743.03.00.2310.1570.38851.02.00.1280.1020.23062.01.00.1830.0510.23471.03.00.1290.1540.28383.01.00.2280.0520.28091.05.00.1300.2610.391105.01.00.3050.0530.358aφg—volume fraction of the argon gas bubbles.bφs—volume fraction of the silica particles.cαg—acoustic attenuation coefficient due to the argon gas bubbles.dαs—acoustic attenuation coefficient due to the silica particles.eαtotal—total acoustic attenuation coefficient (sum of αg and αs). Calculated attenuation coefficients (gas–solid–liquid mixtures). φg—volume fraction of the argon gas bubbles. φs—volume fraction of the silica particles. αg—acoustic attenuation coefficient due to the argon gas bubbles. αs—acoustic attenuation coefficient due to the silica particles. αtotal—total acoustic attenuation coefficient (sum of αg and αs). The strong negative relationships between β and the attenuation coefficients verify that ultrasound attenuation accounts for the inhibitory effects of gas bubbles and silica particles on sonoreactions, individually (Fig. 9a and b) and collectively (Fig. 9c). Although the gas and solid phases exhibited differences in the mechanisms of nucleus addition and acoustic attenuation, their sonochemically inhibiting effects were somewhat similar and combinative in three-phase mixtures at a relatively high φtotal. The sonochemical enhancement induced by introducing nuclei could be completely offset by inhibition when φtotal exceeded a critical value, which fell within 3.0–4.0 vol% under our test conditions (Fig. 7). Additional tests are required to analyze the critical φtotal in the future. Percentage change in I3− formation rate relative to (a) the bubbly solution, (b) the solids suspension, and (c) the pure water scenario, plotted against the attenuation coefficient due to the solid particles, αs, the one due to the gas bubbles, αg, and the total attenuation, αtotal, respectively. The dashed lines represent no change in the formation rate of I3−, and the markers symbolize data for the experimental cases with the same shapes as in Fig. 8. Conclusions Sonochemical activity in aqueous solutions containing millimeter-sized argon gas bubbles and submillimeter-sized silica particles was experimentally investigated using an ultrasonic horn reactor (20 kHz). The effects of the volume fractions of the gas and solid phases were carefully examined, and the mechanism that underlies the sonochemically inhibiting effects of the two phases was clarified. The key findings and conclusions are as follows: (a)The sonochemical yields decreased as the φtotal increased when the gas and solid phases had equal volume fractions. The sonoreactions were enhanced for all three cases at φtotal = 3.0 vol%, compared with the pure water case, and the yields increased with λφ. Contrastingly, the argon gas bubbles and silica particles had an inhibitory effect as the φtotal reached 4.0–6.0 vol%, with the highest and the lowest rates of radical production at λφ > 1 and λφ = 1:1, respectively. The sonochemical yields decreased as the φtotal increased when the gas and solid phases had equal volume fractions. The sonoreactions were enhanced for all three cases at φtotal = 3.0 vol%, compared with the pure water case, and the yields increased with λφ. Contrastingly, the argon gas bubbles and silica particles had an inhibitory effect as the φtotal reached 4.0–6.0 vol%, with the highest and the lowest rates of radical production at λφ > 1 and λφ = 1:1, respectively. (b)The model of the total acoustic attenuation is proposed (Eq. (3)), assuming that the bubble–particle interaction is insignificant. Two classic models, one considering scattering in solid particle suspensions (Eqs. (4), (5), (6)) and the other heat transfer in bubbly liquids (Eqs. (7), (8), (9), (10)), are integrated into this model. The calculated αtotal increases with φtotal for all λφ considered; when φtotal is kept constant (>3.0 vol%), αtotal first increases and then decreases with increasing λφ. The model of the total acoustic attenuation is proposed (Eq. (3)), assuming that the bubble–particle interaction is insignificant. Two classic models, one considering scattering in solid particle suspensions (Eqs. (4), (5), (6)) and the other heat transfer in bubbly liquids (Eqs. (7), (8), (9), (10)), are integrated into this model. The calculated αtotal increases with φtotal for all λφ considered; when φtotal is kept constant (>3.0 vol%), αtotal first increases and then decreases with increasing λφ. (c)The percentage change in radical production, β, is a direct indicator of enhancing or inhibiting sonochemistry compared with the case for reference. Based on the effective-medium approach, we found that ⋅OH yields decreased with increasing φs, if φg was fixed, and vice versa. The experimentally determined β values correlate well with αtotal that was theoretically calculated using our model, irrespective of whether the gas and solid phases were individually or collectively viewed. The percentage change in radical production, β, is a direct indicator of enhancing or inhibiting sonochemistry compared with the case for reference. Based on the effective-medium approach, we found that ⋅OH yields decreased with increasing φs, if φg was fixed, and vice versa. The experimentally determined β values correlate well with αtotal that was theoretically calculated using our model, irrespective of whether the gas and solid phases were individually or collectively viewed. Analyzing the sonochemical test results and ultrasound attenuation mechanisms in the gas–solid–liquid mixtures demonstrated that gas bubbles and solid particles in large quantities hold joint responsibility for inhibiting sonoreactions. An improved understanding of sonochemistry in three-phase mixtures would facilitate the scale-up of sonochemical reactors for practical wastewater treatment. The focus of future research should be on the size effects of the gas and solid phases. CRediT authorship contribution statement Kunpeng Su: Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Writing – original draft, Writing – review & editing. Binghui Li: Visualization, Data curation. Jianhua Wu: Supervision. Pei Xin: Supervision. Shangtuo Qian: Funding acquisition. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements This study was supported by the Natural Science Foundation of Jiangsu Province, China (Grant No. BK20230959), the National Natural Science Foundation of China (Grant No. 52279063), the Fundamental Research Funds for the Central Universities, China (Grant No. B220201043), and the Jiangsu Funding Program for Excellent Postdoctoral Talent (Grant No. 2022ZB163). References