Mechanism Investigation and Standardized Equipment Development of Sequential Multi-Frequency Ultrasonic Extraction for Plasticizers in Polymer Materials

TL;DR

This paper introduces a new ultrasonic extraction method for plasticizers in polymer materials, offering higher efficiency and lower costs.

Contribution

A four-frequency ultrasonic extraction technology with improved recovery and a mechanistic understanding of sequential multi-frequency extraction.

Findings

The four-frequency composite ultrasonic extraction achieved 95.2% recovery of plasticizers.

Stepped-frequency ultrasound allows precise control of cavitation effects while protecting the polymer matrix.

The method provides a more efficient and environmentally friendly alternative to traditional extraction techniques.

Abstract

Phthalates (PAEs), commonly incorporated into materials such as polyvinyl chloride (PVC) and polyvinylidene chloride (PVDC), are easily to migrate readily into the surrounding environment, which have become a matter of increasing concern. Traditional PAEs extraction methods have been prevented by long extraction times and high costs, requiring substitute to accelerate the extraction speed while reducing extraction costs. Ultrasonic-assisted extraction facilitates the release and dissolution of target compounds through the combined effects of acoustic cavitation and molecular vibration acceleration, which could be an effective means to overcome the limitations of traditional extraction methods. Herein, we have developed a four-frequency composite ultrasonic extraction technology for PAEs, with a recovery of 95.2%, approximately 38.2% higher than mode MU 20 kHz. Besides, an in-depth study on the mechanism of ultrasound-assisted extraction with sequential multi-frequency was conducted and we confirm that stepped-frequency ultrasound could achieve precise control of cavitation effects by dynamically adjusting frequency distribution, ensuring high extraction efficiency while maximally protecting the PVC matrix structure, providing a new technical path for efficient and green recovery of plasticizers.

Figures

Full text

1. Introduction

Phthalates (PAEs) are among the most widely used plasticisers, commonly incorporated into materials such as polyvinyl chloride (PVC) and polyvinylidene chloride (PVDC) to enhance flexibility and ductility [1,2,3,4,5]. Yet, because they are bound to these polymers only through weak van der Waals forces and hydrogen bonding, PAEs are not chemically fixed within the matrix. This physical rather than covalent association allows them to migrate readily into the surrounding environment [6,7,8]. The consequences of such migration have become a matter of increasing concern. Numerous epidemiological studies have linked human exposure to PAEs with adverse health outcomes, including impaired reproductive development, metabolic disorders, and a higher incidence of tumours [9,10,11,12]. Beyond their impact on human health, PAEs also pose a serious ecological threat: they can persist in the environment, bioaccumulate in aquatic organisms, and become magnified through food chains, ultimately disrupting ecosystems and re-entering human systems via dietary exposure [13,14,15]. Given these risks, it is imperative to develop more sensitive and reliable methods for the detection and quantification of PAEs, to strengthen standardised monitoring frameworks, and to ensure that plastic products meet strict safety criteria before reaching the market. Such measures are not only essential for protecting environmental and public health but also for advancing the sustainable and responsible development of polymer-based materials.

According to the international standard ISO 8124-6:2014 [16], both Soxhlet extraction and accelerated solvent extraction (ASE) are recognised methods for detecting PAEs. Although Soxhlet extraction requires a relatively simple apparatus, it presents several drawbacks, including long extraction times, high solvent consumption, and low operational efficiency. By contrast, accelerated solvent extraction (ASE), which is performed under elevated temperature and pressure, markedly shortens processing time and reduces solvent usage, offering clear advantages in efficiency and environmental sustainability [17,18,19]. For example, Liu Chunhong et al. [20] successfully quantified six PAEs in plastic student supplies using ASE coupled with gas chromatography–mass spectrometry, achieving both accuracy and reliability. However, the wider adoption of ASE remains limited by the high capital and operational costs of the equipment.In a related development, Wang Yang et al. [21] established an efficient analytical approach to determine DEHP in plastic food packaging materials through microwave-assisted extraction. This method achieved average recovery rates of 88.9–98.4%, while providing rapid and energy-efficient processing, making it particularly well suited to heat-sensitive substances. Nevertheless, its applicability is restricted by the specific polarity requirements of the target compounds, reducing its effectiveness for non-polar substances. In addition, the relatively high cost of the equipment continues to limit its use in large-scale testing laboratories.

Ultrasonic-assisted extraction (UAE), incorporated into the international standard ISO 8124-6:2018, offers an effective means of overcoming the limitations of traditional extraction methods [22,23,24,25,26]. This technique facilitates the release and dissolution of target compounds through the combined effects of acoustic cavitation and molecular vibration acceleration. By employing UAE, the extraction time for PAEs can be greatly reduced, while degradation of heat-sensitive components is minimised through optimisation of the ultrasonic frequency. Moreover, the method imposes no particular restrictions on solvent polarity. Consequently, ultrasonic-assisted extraction has gained increasing recognition in recent years as a versatile and efficient approach for PAEs determination [27,28]. Nevertheless, most existing studies on the ultrasonic extraction of PAEs employ a single fixed frequency (typically 25 kHz or 40 kHz) [29,30], and research exploring the directional regulatory effects of varying ultrasonic frequencies on the phase morphology of polymer materials remains limited.

To enrich the theoretical understanding of ultrasonic extraction technology and also provide an innovative solution for the efficient detection of PAEs in practice, we create a equential multi-frequency ultrasonic-assisted method for PAEs extraction. By using this method, the technical standards within the testing and inspection industry could be advanced, which will promote the standardisation of ultrasonic extraction techniques, and enables green extraction processes that safeguard human health and environmental safety. More specifically, this study systematically investigated the effects of single-frequency ultrasound (MU), dual-frequency ultrasound (DSEU), triple-frequency ultrasound (TSEU), and quadruple-frequency ultrasound (FSEU) modes on the recovery efficiency of PAEs in PVC. The changes of PVC’s micro-morphology main chain structure after ultrasonic treatment was also observed. On this basis, we optimized the combination of various ultrasound frequencies and constructed an system with a low-medium-high frequency gradient ultrasound energy field. Combining the macroscopic erosion and membrane-breaking effects of low-frequency ultrasound with the local micro-cavitation of high-frequency ultrasound, this system significantly improving the extraction efficiency of PAEs while ensuring the integrity of the matrix material structure.

2. Materials and Methods

2.1. Materials and Chemicals

Dichloromethane (AR, Macklin, Shanghai, China), Ethanol (AR, Macklin), Polyvinyl chloride homogeneous powder (PVC, 99%, Aladdin, Shanghai, China), Dibutyl phthalate (DBP, 99%, Aladdin), Benzyl butyl phthalate (BBP, 99%, Aladdin), Bis(2-ethylhexyl) phthalate (DEHP, 99%, Aladdin), Di-isononyl phthalate (DINP, 99%, Aladdin), Di-n-octyl phthalate (DNOP, 99%, Aladdin), Di-isodecyl phthalate (DIDP, Aladdin), Di-iso-butyl phthalate (DIBP, 99%, Aladdin).

2.2. Preparation of Standard Samples

For sample preparation, a precise quantity of homogeneous PVC powder, plasticizer, and stabilizer (at a ratio of PVC-to-stabilizer of 20:1) was weighed and placed in a high-speed mixer. The mixture was processed until its temperature reached 40 °C, after which it was cooled and the pre-plasticized powder collected. The twin-roll mill was preheated to 160 °C, and the powder was milled for 10 min. After being cooled to 40 °C in a mixing unit, the material was transferred to the extruder hopper. Through screw rotation and barrel heating, the compound was plasticized and extruded to particles with an average diameter of 3 mm. The samples were allowed to stand at room temperature for 24 h before being sealed. The prepared samples are presented in Figure S1. The adding materials were listed in Tables S1 and S2.

2.3. The Extraction of PAEs by Single-Frequency Ultrasound

Before sampling, all the sampling tubes were washed 3 times by ethanol and then baked at 350 °C to remove all of the organic compounds. Weighed portions of 0.5000 ± 0.001 g of standardized PVC samples containing seven PAEs (including DBP, BBP, and DEHP) were placed in 25 mL sample tubes. In each tube, 15 mL of 99% pure dichloromethane was added. The tubes were then immersed in a water bath maintained at 60 ± 2 °C, with a liquid depth of 220 mm. The ultrasonic power density was set to 50 W/L, and the extraction instrument was operated at frequencies of 20, 40, 60, and 80 kHz for 10–30 min. After extraction, the samples were cooled, and the supernatant was collected and filtered through a 0.45 $[eqn]$ m membrane filter for subsequent quantitative analysis. The solid and liquid phases obtained after extraction were separated. The solid fraction was air-dried at room temperature to remove residual dichloromethane and then stored as a retained sample. The liquid extract was subjected to vacuum filtration, after which the filtrate was dried and preserved for subsequent structural analysis.

To better evaluate the impact of ultrasonication mode on PAEs extraction efficiency, we employed n-hexane as the solvent while maintaining all other experimental conditions constant. Ultrasonic treatment was performed at frequencies 20 kHz for a duration of 30 min. The corresponding sample was designated as MU 20 kHz (n-hexane).

2.4. The Extraction of PAEs by Sequential Multi-Frequency Ultrasound

Before sampling, all the sampling tubes were washed 3 times by ethanol and then baked at 350 °C to remove all of the organic compounds. Weighed portions of 0.5000 ± 0.001 g of standardised PVC samples containing seven PAEs (including DBP, BBP, and DEHP) were placed into 25 mL sample tubes. In each tube, 15 mL of 99% pure dichloromethane was added. The tubes were then immersed in a water bath maintained at 60 ± 2 °C, with a liquid depth of 220 mm. The ultrasonic power density was set to 50 W/L. Under the sequential multi-frequency operation mode, the ultrasound alternates between each frequency for a set duration, with each frequency proceeding in an orderly sequence. The frequency combinations include: 20/40 kHz, 20/60 kHz, 20/80 kHz, 40/60 kHz, 40/80 kHz, 60/80 kHz, 20/40/60 kHz, 20/40/80 kHz, 20/60/80 kHz, 40/60/80 kHz, and 20/40/60/80 kHz. The ultrasound time was set as 10–30 min. After extraction, the samples were cooled, and the supernatant was collected and filtered through a 0.45 $[eqn]$ m membrane filter for subsequent quantitative analysis. The solid and liquid phases obtained after extraction were separated. The solid fraction was air-dried at room temperature to remove residual dichloromethane and then stored as a retained sample. The liquid extract was subjected to vacuum filtration, after which the filtrate was dried and preserved for subsequent structural analysis.

To better evaluate the impact of ultrasonication mode on PAEs extraction efficiency, we employed n-hexane as the solvent while maintaining all other experimental conditions constant. Ultrasonic treatment was performed at frequencies 20/40/60/80 kHz for a duration of 30 min. The corresponding sample was designated as FSEU 20/40/60/80 kHz (n-hexane).

2.5. Comparative Extraction Analysis of Commercial PVC

To assess the applicability of the multi frequency ultrasound assisted extraction method, we undertook comparative evaluations using commercial PVC samples. Typically, commercial PVC toys were reduced to fragments 0.5 cm × 0.5 cm. 0.5000 ± 0.001 g of above fragments were placed into 25 mL sample tubes (Before sampling, all the sampling tubes were washed 3 times by ethanol and then baked at 350 °C to remove all of the organic compounds). In each tube, 15 mL of 99% pure dichloromethane was added. The tubes were then immersed in a water bath maintained at 60 ± 2 °C, with a liquid depth of 220 mm. The ultrasonic power density was set to 50 W/L. Under the sequential multi-frequency operation mode, the ultrasound alternates between each frequency for a set duration, with each frequency proceeding in an orderly sequence. The frequency combinations include: 20 kHz and 20/40/60/80 kHz. The ultrasound time was set as 30 min. The image of the testing toy is shown in Figure 1.

2.6. Materials Characterization

The crystalline structure of the samples was characterized by X-ray diffraction (XRD, Cu/K $[eqn]$ , Rigaku, Tokyo, Japan). The morphology of the samples were observed by emission scanning electron microscopy (SEM, ZEISS, GeminiSEM 300, Oberkochen, Germany). The particle size of fragments detached from the PVC matrix was measured by laser particle size and shape analyzer (LPSA, Microtrac SYNC, Montgomeryville, PA, USA). The pore size distribution of the PVC matrix was measured by BET method (Microtrac 3Flex, Norcross, GA, USA). The recovery rate of PAEs was measured by gas chromatography (GC, Agilent Technologies 8890-5977B, Santa Clara, CA, USA). The chemical structural information of samples were obtained through Fourier transform infrared spectroscopy (FTIR, Thermo Fisher Scientific Nicolet iS50, Madison, WI, USA).

2.7. Plotting of PAEs Calibration Curves

Standard stock solutions were serially diluted by $[eqn]$ $[eqn]$ to prepare a set of calibration solutions with concentrations of 1, 5, 10, 25, and 50 $[eqn]$ g/mL. Analyses were performed under optimized instrumental conditions. Calibration curves were constructed by plotting peak area against plasticizers concentration, yielding excellent linearity for all seven PAEs across the 1–50 $[eqn]$ g/mL concentration range, with correlation coefficients ( $[eqn]$ ) between 0.9991 and 0.9999. The results was recorded in Table 1.

2.8. Quality Assurance and Quality Control

In order to ensure the reliability of this method, quality control is demanded in data collection and analysis. The standard deviation (SD) and relative standard deviation (RSD) were measured by following approach: Standard stock solutions were serially diluted by dichloromethane to prepare a set of calibration solution with concentrations of 1 $[eqn]$ g/mL. The above solution was measured 9 times repeatedly, and the SD and RSD were calculated. The calculation formula of SD is as follows.

$$[eqn]$$

where N is the number of measurements; $[eqn]$ is the i-th measured value and $[eqn]$ is the arithmetic mean of all measurements.

The calculation formula of RSD is as follows.

$$[eqn]$$

where SD is the standard deviation and $[eqn]$ is the arithmetic mean of all measurements.

The sensitivity of the method was evaluated using the limit of detection (LOD) and limit of quantitation (LOQ). Following international conventions, the LOD was determined as 3 times of standard deviation (3 SD), while the LOQ was defined as 10 times of standard deviation (10 SD). The calculation formulas are as follows:

$$[eqn]$$ $$[eqn]$$

where SD is the standard deviation. The results was recorded in Table 2.

2.9. Calculation of the Diffusion Coefficients of PAEs in PVC

To quantify the migration process of PAEs in the PVC matrix, a mathematical model based on Fick’s second law was employed. This equation can be applied to describe the migration behavior of contaminants from materials. The shape of the sample was approximately considered as a spherical particle. The mass transfer of PAEs within the particle was regarded as a diffusion-controlled process, with negligible external mass transfer resistance. The corresponding initial and boundary conditions were as follows [31]:

$$[eqn]$$

where $[eqn]$ represents the extraction mass of PAEs at time t, $[eqn]$ is the total extraction mass of PAEs after sufficient ultrasound time, D symbolizes the effective diffusion coefficient and r is the radius of spherical particles. This formula can be transformed into

$$[eqn]$$

where $[eqn]$ represents the recovery rate of PAEs at time t, $[eqn]$ is the balance recovery rate of PAEs after sufficient ultrasound time.

Due to the potential degradation of PVC caused by excessively long ultrasonic exposure time, we employed the Weibull distribution function to extrapolate the recovery rate-time data, aiming to determine the theoretical $[eqn]$ . The Weibull function [32] is widely utilized in describing the release kinetics of analytes from solid matrices, and its form is as follows:

$$[eqn]$$

where $[eqn]$ represents the recovery rate of PAEs at time t, $[eqn]$ is the balance recovery rate of PAEs after sufficient ultrasound time, $[eqn]$ delegates shape parameter, $[eqn]$ is the scale parameter, which is related to extraction rate and t − $[eqn]$ reprsents the position parameter. $[eqn]$ could usually be regarded as 0, so Equation (7) can be simplified as

$$[eqn]$$

The experimental data was brought into Equation (8), and was fitted by using the Origin nonlinear least squares method. $[eqn]$ , $[eqn]$ , and $[eqn]$ were optimized by calcution. The fit goodness was evaluated by coefficient of determination ( $[eqn]$ ).

The greater the number of roots, the more reliable the results are. To make the estimation feasible, 5 roots were calculated:

$$[eqn]$$

where $[eqn]$ represents the recovery rate of PAEs at time t, $[eqn]$ is the balance recovery rate of PAEs after sufficient ultrasound time and D symbolizes the effective diffusion coefficient.

Input all experimental data into Formula (9), and optimize parameter D through nonlinear fitting in Origin.

3. Results and Discussions

3.1. Crystalline and Chemical Structure of PVC with Different Ultrasonic Treatment

The cavitation and mechanical effects generated by ultrasound may induce breakage or rearrangement of PVC molecular chains. To determine whether the crystalline structure of PVC was affected, X-ray diffraction (XRD) analyses were conducted on the residual PVC matrices following ultrasound-assisted extraction at different frequencies. The results are presented in Figure 2. The XRD images of samples extracted using different ultrasonic modes showed no significant differences. All samples exhibited a broad (100) diffraction peak centred at approximately 2 $[eqn]$ = 20°, indicating that ultrasonic treatment primarily influenced the amorphous regions of the PVC or disrupted secondary intermolecular bonds [33,34,35], without altering the inherent amorphous nature of the original material.

To investigate changes in the functional groups of the PVC matrix before and after ultrasonic treatment, FTIR analysis was performed, and the results are presented in Figure 3a,b. The spectra show a C=O stretching vibration peak near 1735 cm^−1^ [36,37,38], corresponding to the ester carbonyl group of PAEs. A C–C stretching vibration peak around 1260 cm^−1^ is also observed [39,40], primarily associated with the PVC backbone conformation and interchain interactions. The peak near 630 cm^−1^ corresponds to the C–Cl stretching vibration [41], a characteristic feature of PVC, reflecting the vibrational state of the C–Cl bond. Additionally, peaks near 2921 cm^−1^ are attributed to C–H stretching vibrations [42,43], while the broad peak near 3320 cm^−1^ corresponds to O–H stretching vibrations [44].

For the PVC1 samples treated with single-frequency ultrasound, the C=O peak near 1735 cm^−1^ shows no significant shift compared with the untreated sample (US) across all treatments (MU 20 to MU 80). This indicates that the chemical structure of PAEs remains intact during migration from the PVC matrix, suggesting that the extraction process primarily involves physical desorption and diffusion. Regarding the C=O peak intensity, the trend follows: US > MU 80 > MU 60 > MU 40 > MU 20. This demonstrates that, for the same ultrasonic treatment duration, low-frequency ultrasound (20 kHz) promotes more efficient PAEs migration, leaving fewer residual PAEs in the deeper regions of the PVC matrix. Near 1260 cm^−1^, the C–C peaks of MU 20 exhibit a blue shift relative to US, indicating increased effective C–C bond strength and reduced constraints on bond rotation and vibration. This suggests that ultrasonic energy enhances the motion of PVC chain segments, increasing free volume in the molecular chains and facilitating PAEs migration from amorphous regions. In contrast, no significant blue shift is observed in MU 60, 80, implying that at 60, 80 kHz, chain segment mobility is limited, restricting PAEs migration to superficial layers. Near 2921 cm^−1^, the C–H stretching vibration intensity follows the trend: US > MU 80 > MU 60 > MU 40 > MU 20. This reflects more extensive PAEs removal under low-frequency ultrasound, which reduces the number of C–H bonds. It also indicates that 20 kHz ultrasound generates stronger mechanical shear and localized high-temperature/pressure conditions, causing greater main-chain bond breakage, surface erosion, and disorder—randomizing C–H vibration orientation, reducing directional absorption intensity. As the frequency increases from 40 to 80 kHz, the C–H peak intensity partially recovers relative to MU 20 but remains below US levels. This suggests that high-frequency ultrasound produces shorter cavitation cycles with lower collapse energy, preserving the structural integrity of amorphous regions. However, this also results in higher residual PAEs in deeper layers, consistent with the absence of a C–C blue shift in MU 80. Under the same single-frequency ultrasound conditions, PVC2 samples display similar trends, further confirmed the above conclusions.

For the PVC1 sample subjected to dual-frequency ultrasonic treatment, comparison of the C–H stretching vibrations near 2900 cm^−1^ revealed the following trend in peak intensity: US > DSEU 60/80 > DSEU 40/80 ≈ DSEU 40/60 > DSEU 20/40, 60, 80. This trend is consistent with observations under single-frequency conditions, confirming that lower ultrasonic frequencies produce more pronounced damage to the PVC main chain structure. Moreover, the C–H stretching peak intensities for DSEU 20/40, DSEU 20/60, and DSEU 20/80 were found to be comparable. Despite employing high–low frequency composite modes, these configurations did not mitigate macroscopic fracture of the PVC matrix. For the PVC1 sample subjected to triple-frequency ultrasonic treatment, the trend in C–H stretching peak intensity near 2921 cm^−1^ was as follows: US > TSEU 20/60/80 > TSEU 20/40/80 > TSEU 20/40/60. This pattern again confirms that lower ultrasonic frequencies result in greater degradation of the PVC main chain.Under quadruple-frequency ultrasound, however, the FSEU 20/40/60 /80 sample exhibited a higher C–H peak intensity than triple-frequency modes (TSEU 40/60/80, TSEU 20/40/80, and TSEU 20/40/60). This observation indicates that the FSEU 20/40/60/80 kHz configuration induces less structural damage to the PVC backbone compared with these triple-frequency treatments. The FSEU 20/40/60/80 kHz system achieves this through a spatiotemporally complementary mechanism, comprising: Low-frequency dominance, generating strong cavitation and erosive effects; Mid-frequency modulation, maintaining a uniform micro-flow field; and High frequency induction, promoting relaxation of chain-segment vibrations. Together, these effects establish a selective disruption pathway that weakens secondary interactions between PAEs and the PVC matrix while avoiding covalent bond cleavage within the polymer backbone. This validates the “synergistic weakening effect” of frequency-switching ultrasound, which enhances the dissociation of weak intermolecular interactions in complex polymer plasticiser systems.

3.2. Morphology of PVC with Different Ultrasonic Treatment

To elucidate the effect of ultrasonic frequency on the structural and morphological characteristics of PVC, SEM analysis was performed and the results was in Figure 4. Under single-frequency ultrasound, the frequency and treatment duration showed a clear synergistic influence on the PVC microstructure and PAEs leaching behaviour. At the low frequency of 20 kHz, the high cavitation energy density produced micro-pores of approximately 0.4 $[eqn]$ m on the PVC surface within 10 min. However, extending the treatment to 30 min intensified the mechanical stress from collapsing cavitation bubbles, causing pronounced main-chain scission and severe surface damage caused by pore collapse, incomplete channels, and enlarged micro-pores of about 1.6 $[eqn]$ m (a 300% increase). These findings correspond with FTIR observations showing decreased C–H peak intensity and blue-shifted C–C vibration peaks, indicative of irreversible structural degradation at 20 kHz. Hence, this frequency proved ineffective in balancing enhanced extraction efficiency with structural preservation. In contrast, the mid-frequency ultrasound at 40 kHz exhibited milder cavitation effects. After 10 min of treatment, the MU 40 sample developed 0.3 $[eqn]$ m micro-pores, while 30 min resulted in partial fracture but retained semi-ordered channels with enlarged pores (1.0 $[eqn]$ m, a 233% increase). The surface roughness was lower than that of MU 20 after 30 min, consistent with FTIR results showing a weaker C=O peak, a slight blue shift in C–C vibrations, and higher C–H intensity than MU 20. The MU 60 sample displayed well-developed micro-channels after 30 min with improved surface integrity compared with low-frequency treatment. At the high frequency of 80 kHz, the MU 80 sample showed only small (0.05 $[eqn]$ m) micro-pores after 10 min, with incomplete channel formation, and these pores expanded to approximately 0.2 $[eqn]$ m after 30 min. FTIR spectra revealed no notable blue shift in the C–C peaks and C–H intensities comparable to the untreated sample, confirming that the main-chain structure remained intact. These results indicate that high-frequency ultrasound is preferable for applications where maintaining PVC structural integrity is critical.

Under tri-frequency ultrasonic mode, sample TSEU 20/40/60 kHz exhibited a large number of ordered pores with sizes ranging from 0.05 to 0.2 $[eqn]$ m, accompanied by partially closed structures Figure 5. The formation of these closed pores may suggest the structural retraction and collapse of ultrasound-induced micropores and channels. Compared with the single-frequency MU 20 kHz sample, after 30 min of ultrasonic treatment, the TSEU 20/40/60 kHz sample surface showed smaller pore radii and a higher proportion of closed pores, indicating that the acoustic field energy distribution in the combined frequency mode (20/40/60 kHz) is more uniform and gentler, resulting in less damage to the overall material structure. Further comparison among the three combinations—TSEU 20/40/60 kHz, TSEU 20/40/80 kHz, and TSEU 40/60/80 kHz—revealed that the TSEU 40/60/80 kHz sample surface had the highest number of closed pores. This phenomenon suggests that as the combined frequencies shift toward the higher range, the cumulative thermal effect generated by ultrasound becomes more uniform. Additionally, a microcrack network was observed on the surface of the TSEU 40/60/80 kHz sample, which is hypothesized to be related to the local resonance effect induced by high-frequency ultrasound. When the combined high frequencies match the natural vibration frequencies of specific molecular segments or microstructures in the PVC matrix, local resonance may be triggered, leading to stress concentration that exceeds the material’s tolerance threshold and resulting in the formation of surface microcracks. This phenomenon was more pronounced in the quad-frequency sample FSEU 20/40/60/80 kHz, where a large number of closed and unclosed pores were present, with highly uniform pore sizes of approximately 0.04 $[eqn]$ m. Based on all SEM characterization results, it can be concluded that the FSEU 20/40/60/80 kHz sequential multi-frequency ultrasound, through dynamic frequency switching, creates a gradient distribution of cavitation intensity, thereby inducing the formation of a progressive microcrack network on the material surface. This mechanism combines the macro-pitting and membrane-breaking effects of low-frequency ultrasound with the efficient extraction advantages of high-frequency localized micro-cavitation, significantly enhancing the dissolution efficiency of plasticizers. It is noteworthy that the spatiotemporally dispersed distribution of sequential multi-frequency energy effectively avoids non-directional pore collapse caused by over-focused energy in single frequency modes (20 kHz), as well as disordered coarsening of cracks due to cumulative mechanical stress from a single source.

To better analyze the changes in the PVC matrix caused by sequential multi-frequency ultrasound compared to single-frequency ultrasound, we plotted the pore size distribution of the PVC matrix after ultrasound treatment Figure 6. As shown in the figure, the average pore size of the FSEU 20/40/60/80 kHz samples is only 0.2 $[eqn]$ m, with 90% of the pore radii below 0.25 $[eqn]$ m. This result represents a 57% reduction in the average pore size compared to the samples treated with MU 20 kHz ultrasound.

Furthermore, by comparing the box plots of the pore size distribution, we found that the box plot of the FSEU 20/40/60/80 kHz samples is narrower, indicating a more uniform pore size distribution. The radial expansion of the pores on the sample surface after 30 min of MU 20 kHz ultrasound treatment is nearly 0.8 $[eqn]$ m, suggesting that the PVC structure collapsed to varying degrees under the low-frequency ultrasound mode.

In order to assess the damage of ultrasound to PVC structures, we conducted particle size analysis on the detached fragments. The experimental results (Figure 7a) indicate that the average particle size of PVC fragments is inversely correlated with the ultrasonic energy input. For sample MU 20 kHz, due to its high cavitation intensity, the brittle fracture of the PVC main chains were directly induced, resulting in submicron-sized fragments with an average size of 1257 nm after just 5 min ultrasonic treatment. In contrast, for 60 kHz and 80 kHz ultrasonic treatment, due to the high-density distribution characteristics of cavitation bubble clusters, primarily induce the initial expansion of microcracks on the substrate surface, with fragment sizes reaching 1392 nm. As the exposure time accumulates, ultrasonic energy facilitates secondary fragmentation of the fragments through a self-organizing shear flow field, gradually reducing the particle size of the detached fragments to 450–570 nm. This phenomenon aligns with the three-stage model of “microcrack initiation–interfacial delamination–fragment crushing” validated by the Weibull distribution function. Notably, in high plasticizer systems (PVC2), the initial particle size of detached fragments (1800–2000 nm) significantly increases. This is due to the formation of localized cavity weakness areas after a large amount of PAEs leaches out, making the matrix more prone to millimeter-scale peel fragments. The continuous ultrasonic energy input eventually reduces the fragment size closer to that of low-plasticizer systems, with a 77.5% decrease from the initial fragment size. For sequential multi-frequency ultrasonic extraction, different frequency combinations significantly affect the fragment size of polymer materials. Under the FSEU 20/40/60/80 kHz mode, the average size of detached PVC fragments is smaller and more concentrated, indicating less material structural damage and a milder extraction process. Specifically, the average particle sizes of PVC1 and PVC2 samples under FSEU are 730 nm and 830 nm, respectively, which are decreases of 42% and 40.4% compared with the single-frequency 20 kHz mode, demonstrating a significant particle size reduction effect. Additionally, under DSEU combined frequency, the fragment size of PVC1 is 816–700 nm, and PVC2 is 894–812 nm, showing a trend of gradually decreasing particle size with increasing complexity of frequency combinations. The synergistic effect of cavitation and mechanical effects brought by dynamic frequency switching allows the formation, growth, and collapse of cavitation bubbles more complex and uniform. The brittle fractures caused by localized cavitation concentration were prevented, which promoting more uniform crushing on PVC. This results indicates that dynamic multi-frequency regulation has significant advantages in weakening plasticizer bonding and promoting secondary fragmentation. In contrast, for DSEU and TSEU modes, due to the limited frequency combinations, the dynamic variation in cavitation effects is insufficient, resulting in larger and more dispersed fragment sizes. Therefore, by dynamically switching frequencies, FSEU 20/40/60/80 kHz mode achieves diversification and uniform distribution of cavitation intensity, promoting significant reduction and concentrated distribution of fragment sizes. In conclusion, FSEU 20/40/60/80 kHz mode demonstrating notable advantages in maintaining a balance between material structural stability and improving extraction efficiency, thus providing a theoretical basis for optimizing ultrasonic extraction processes.

To investigate the effect of ultrasonic energy input on the porosity of PVC material, we conducted BET tests on the PVC matrix, and the results are recorded in Figure 7b. For sample MU 20 kHz, after 30 min ultrasonication, the average pore size of the matrix expanded from 5.7 nm to 16.5 nm. This result indicates that 20 kHz ultrasonic frequency effectively disrupts the entangled network in the amorphous region of PVC, enabling rapid extraction of PAEs, followed by gradual pore closure. As the ultrasonic frequency increases, the time required for PAEs extraction prolongs, manifested by a delayed peak in maximum average pore size. For example, the maximum average pore size of MU 40 kHz and MU 80 kHz appear at 40 and 50 min, respectively, suggesting that high frequency ultrasonic modes require longer extraction times compared to low-frequency modes. In high frequency ultrasonic modes, the collective effect of cavitation bubbles forms a distributed energy field with lower transient energy, which is consistent with the phenomena observed in FTIR and SEM results. It is noteworthy that the samples under different ultrasonic frequencies exhibit the same trend in pore size variation with different ultrasonication time. As the exposure time extends, sample MU 20, MU 40, MU 60, and MU 80 kHz all show a reduction in average pore size. Combined with SEM images, the chain segment structure of the PVC matrix was damaged by 20 kHz ultrasonic treatment, leading to a pores collapse, which is reflected in the SEM images as numerous depressions. In contrast, the PVC chain segment structure remains relatively intact under 80 kHz ultrasonic treatment. This experiment demonstrates that there is a window of pore size regulation during the ultrasonic extraction process, which can achieve a balance between pore expansion, structural stability, and extraction efficiency. A similar trend in average pore size was observed in high-concentration PVC systems (PVC2). Under sequential multi-frequency ultrasonication, after 30 min of treatment, the average pore size of the PVC matrix decreases with the extension of the high-frequency segments, consistent with the results from single-frequency modes.

The average pore sizes of PVC1 and PVC2 samples under FSEU mode are 11 nm and 10 nm, respectively, reaching increases of 97.4% and 78.7% compared to the untreated samples (5.7 nm and 5.6 nm). However, these values are lower than the maximum pore size achieved under the single-frequency 20 kHz mode (16.5 nm), indicating that variable-frequency ultrasonication could promote pore formation while avoiding severe local structural damage. Notably, after 30 min ultrasonication, the FSEU 20/40/60/80 kHz and TSEU 40/60/80 kHz samples exhibit similar average pore sizes. Although the FSEU 20/40/60/80 kHz mode includes longer low-frequency segments, its structural impact is comparable to that of TSEU 40/60/80 kHz. This results suggests that quad-frequency composite ultrasonication may achieve an optimal balance between high extraction efficiency and minimal structural damage.

3.3. The Recovery Rate of PAEs with Different Ultrasonic Treatment Mode

To investigate the effect of ultrasonic frequency on PAEs extraction, the PAEs recovery rate was observed after carrying out 5–30 min single frequency (20–80 kHz) ultrasonic treatment (Figure 8a,b). Under different ultrasonic frequencies, the PAEs recovery rates of sample PVC1 and PVC2 showed consistent trends. Taking PVC1 as an example, within 30 min ultrasonic treatment, longer exposure times resulted in higher recovery rates. This finding aligns with the BET test results, which suggested a gradual opening of pore sizes within 30 min treatment. Furthermore, higher ultrasonic frequencies led to lower PAEs recovery rates. For instance, for 20 kHz mode, the PAEs recovery rate was 57% after 30 min ultrasonication, whereas decreased to 43% under 80 kHz treatment. Similarly, for PVC2 samples, the recovery rate decreased from 50% (under 20 kHz treatment) to 43% (under 80 kHz treatment). These results confirm that low-frequency ultrasound generates larger cavitation bubbles, which release stronger mechanical energy upon collapse, leading to higher extraction efficiency. Additionally, low-frequency ultrasound enhances solvent penetration and solute diffusion rates, thereby shortening the mass transfer equilibrium time. In terms of variable-frequency ultrasonic extraction, using the 30 min treated PVC1 as the benchmark, the overall extraction efficiency followed the order: FSEU > TSEU > DSEU. The optimal dual-frequency combination (DSEU) was 20/40 kHz, achieving a PAEs recovery rate of approximately 66.4%. The optimal tri-frequency combination was TSEU 20/40/60 kHz, with a PAEs yield of 78.4%.

Notably, FSEU 20/40/60/80 kHz achieved the best extraction performance, with a recovery rate of about 95.2%. This represents an increase of approximately 43.4% and 21.4% compared to the highest recovery rates under DSEU and TSEU modes, respectively (Figure 8c,d). This result indicating a significant influence of the ultrasonic operating mode on PAEs extraction efficiency. Similar trends were observed for PVC2 samples. Combined with FTIR and particle size analysis, the FSEU 20/40/60/80 kHz mode demonstrated better preservation of the C–H backbone integrity compared to the TSEU 40/60/80 kHz mode, while the size of detached fragments remained similar. For quad-frequency ultrasonic system, low frequency stage was leveraged to enhance cavitation erosion, high frequency stage was used to induce segmental vibrational relaxation, while medium frequency stage was introduced to maintain a uniform micro-flow field. These synergistic effects create a dynamically matched cavitation mechanism which could effectively promote the deep release of PAEs while minimizing excessive damage to the PVC matrix, thereby achieving an optimal balance between high extraction efficiency and low structural degradation.

To maintain the superiority of the FSEU mode, we conducted repeated experiments and statistical analyses comparing the FSEU and MU 20 kHz modes for PVC1. BY using an independent samples t-test (n = 5), we confirmed a significant difference in PAEs yield between these two modes (p < 0.0001). The average PAEs yield from these replicates are presented in Figure 9.

3.4. Extraction Kinetics Analysis

Table 3 and Figure 10 record the impact of different single-frequency ultrasonic modes on the extraction kinetics of PAEs. It can be observed that the diffusion coefficient (D) of PAEs decreases with increasing ultrasonic frequency, resulting in a slower diffusion rate during the extraction process. This confirms that low-frequency ultrasonic waves could generate larger cavitation bubbles compared to high-frequency ultrasonic waves, inducing the formation of large-diameter pores in the PVC matrix and accelerating the release of PAEs. High-frequency ultrasonic waves have a milder effect on the PVC matrix. Especially when the ultrasonic frequency is 80 kHz, the diffusion coefficient of PAEs is much lower than other single-frequency ultrasonic modes, indicating that most of the pores generated in the PVC matrix are micropores or cracks.

3.5. Other Comparative Experiments

To better assess the impact of ultrasonication mode on PAEs extraction efficiency, we employed n-hexane as the solvent while keeping all other conditions unchanged. N-Hexane is a non-expansive solvent which will not significantly dissolve or soften the matrix. The results was recorded in Figure 11. By using n-hexane as a comparative experiment, the solvent swelling effect could be isolated, thereby accurately reflects the actual contribution of different ultrasonic modes. The recovery rates of 7 kinds of PAEs under MU 20 kHz and FSEU 20/40/60/80 kHz modes were recorded in Figure 8. According to the results, the recovery rates of the seven PAEs under sequential multi-frequency ultrasound were significantly higher than those under the single-frequency 20 kHz ultrasonic mode. The average recovery rate in FSEU mode was 20.48%, representing a 2-fold increase compared to the MU 20 kHz ultrasonic mode. This confirms that the proposed sequential multi-frequency ultrasonic mode can effectively enhance the extraction efficiency of PAEs.

To validate the applicability of the sequential multi-frequency ultrasonic-assisted extraction method, we conducted comparative tests on commercial PVC samples using the MU 20 and FESU 20/40/60/80 extraction modes. The results was recorded in Figure 12. revealed that the FESU 20/40/60/80 modes yielded 345 mg/kg, whereas the MU 20 mode produced 309 mg/kg. This demonstrates that the FESU 20/40/60/80 modes exhibit an advantage over the MU 20 mode in extraction efficiency, attributed to their multi-frequency synergistic mechanism.The 11.6% higher extraction yield in FESU 20/40/60/80 confirms that this mode better protect PVC’s crystalline structure while improve PAEs’ recovery rate, making it more suitable for detecting PAEs in PVC matrices. The repeated experiments were conducted to verify this conclusion. The statistical analyses were conducted by comparing the PAEs yield of FSEU and MU 20 kHz modes. Using an independent samples t-test (n = 5), we confirmed a significant difference in PAEs yield between these two modes (p = 0.035).

4. Conclusions

This experiment developed a quad-frequency ultrasonic-assisted extraction technique for PAEs from PVC matrices. Through frequency adjustment, the final extraction rate reached approximately 95%. By optimizing the ultrasonic frequency combination parameters, high extraction rates and low material damage were successfully achieved simultaneously. Experimental results show that composite frequency modes can produce an uniformly distributed acoustic energy field, effectively reducing the overall structural damage of the materials. In particular, the FSEU 20/40/60/80 kHz mode dynamically switches frequencies to form a cavitation intensity field with a gradient distribution, inducing a progressive microcrack network on the material surface and significantly enhancing the integrity of the material structure. This study achieves high extraction efficiency while maximally protecting the PVC matrix structure, providing a new technical path for efficient and green recovery of PAEs.

It should be noted that, even the theoretical mechanism of the ultrasound-assisted extraction method suggests its capability to separate all chemical substances interacting with the matrix via van der Waals forces, this study specifically focused on extraction and migration experiments for 7 types of PAEs. Therefore, expanding the scope to include other chemical compounds represents a critical direction for future research.