Comprehensive quality evaluation of ultrasound-assisted fermented Yuluxiang pear juice: Physicochemical characteristics, antioxidant activity, and multi-dimensional flavor analysis Authors: Huiling Duan, Danrong Yu, Chuan Xie, Chuanwei Zhang, Yaping Shen, Ling Ma, Liyan Jia, Shicong Zhang, Xiaoyu Zhang, Zhihui Yu Published: 2026 ABSTRACT Graphical abstractKeywords: Yuluxiang pear, Ultrasound-assisted fermentation, Electronic nose, Electronic tongue, Gas chromatography-ion mobility spectrometry, Flavor profile Introduction Yuluxiang pear (Pyrus bretschneideri Rehd. ‘Yuluxiang’), a premium mid‑to‑late‑season cultivar, was developed through hybridization using the ‘Korla Fragrant Pear’ as the female parent and the ‘Snowflake Pear’ as the male parent. Renowned for its thin skin, fine texture, low stone cell content, abundant juiciness, and sweet flavor, it has become one of the major cultivars promoted in Northern China [1]. In recent years, evaluations of its fruit quality have primarily focused on cultivation physiology, postharvest storage, and regional quality characteristics. For instance, previous studies confirmed that Yuluxiang pears grown in regions with significant diurnal temperature variation (such as Xi County, Shanxi) exhibit superior soluble sugar content [2]. Postharvest storage studies have demonstrated that ice-temperature storage can effectively delay fruit senescence and exerts specific effects on the composition of modulate volatile organic compounds (VOCs), including esters, aldehydes, and alcohols [3]. However, compared with studies on fresh consumption and primary storage, a comprehensive deep processing utilization system for Yuluxiang pear has not yet been established. Current product forms remain limited, and the added value is relatively low, which constrains the vertical development of its industrial economy. Therefore, developing new high value-added Yuluxiang pear processing products is of great significance for extending the industrial chain and enhancing industrial benefits. Yuluxiang pear juice processing has long faced critical technical challenges, including enzymatic browning, nutrient degradation during thermal processing, and flavor volatilization. While traditional thermal pasteurization can ensure safety, it often leads to the degradation of aromatic compounds and vitamins [4]. Therefore, non‑thermal processing and biotransformation techniques have attracted increasing attention. During traditional fermentation, fruit and vegetable juices are subject to the combined effects of enzymatic and non-enzymatic browning reactions, leading to reduced juice clarity and undesirable color deterioration. Meanwhile, natural characteristic flavor compounds (i.e., esters, aldehydes, and alcohols) are susceptible to volatilization loss during fermentation, while bitter and astringent substances such as bound phenols are difficult to degrade effectively, resulting in poor product palatability [5]. Furthermore, the physical barrier composed of plant cell walls and intercellular spaces limits the release of intracellular nutrients (i.e., bound phenols and fermentable sugars), leading to low substrate utilization efficiency by microorganisms and limited fermentation efficiency [6], [7]. Among them, microbial fermentation, particularly using lactic acid bacteria (LAB), is widely recognized as an effective strategy for enhancing pear juice quality. During fermentation, sugars and organic acids convert to lactic and acetic acids, which impart a mild acidity and inhibit harmful bacteria. Additionally, microbial enzymes such as β‑glucosidase can hydrolyze glycosidic bonds in phenolics, releasing bound bioactives and significantly enhancing the antioxidant activity of pear juice [6]. Studies have demonstrated that fermentation with LAB such as Lactiplantibacillus plantarum(L. plantarum) can effectively increase the total phenolic content of pear juice and enhance its DPPH and ABTS+ radical scavenging activities. After pear juice was fermented by L. plantarum for 36 h, an increase in total phenolic content was observed [7]. Additionally, Yuluxiang pear fruit is naturally endowed with characteristic differential volatile compounds identified by GC-IMS fingerprinting, including esters, aldehydes, and alcohols, whose profiles are transformed during fermentation through microbial metabolism. Specifically, LAB exhibit esterase activity to hydrolyze ester precursors, while concurrently generating new alcohols, acids, esters and ketones via amino acid and fatty acid metabolism [8]. Ultrasound (US), a non-thermal physical processing technology, offers considerable potential for enhancing the quality, extraction efficiency, and microbial safety of fruit and vegetable juices. The mechanical forces and microjets produced by its cavitation effect disrupt cell wall structures, which promotes the release of intracellular compounds such as polysaccharides and polyphenols, thereby increasing the nutritional density of the juice [9]. In recent years, the synergistic application of US with fermentation technology (US-assisted fermentation) has emerged as a promising research direction. Studies indicate that US treatment enhances fermentation by exerting micro-scale agitation on plant cell wall structures through mechanical shear and microjets, thereby promoting the release of intracellular nutrients and improving substrate accessibility. At an appropriate power levels, US can activate the cellular stress response of lactic acid bacteria, promoting probiotic proliferation and the synthesis of functional metabolites [10]. Furthermore, US activates key metabolic enzymes such as esterases and β-glucosidases, facilitating the synthesis of characteristic flavor compounds while promoting the conversion of bound phenols into free forms, thereby enhancing antioxidant activity and reducing bitterness and astringency [11]. Additionally, US significantly increases the volumetric mass transfer coefficient of the fermentation system by enlarging the gas-liquid interfacial area and reducing bubble size, thereby accelerating flavor formation processes such as esterification reactions [12]. However, current research, both domestically and internationally, has predominantly focused on conventional LAB fermentation of common pear varieties, such as 'Dangshan Suli' and 'Nanguo' pear. In particular, the application of US pretreatment combined with L. plantarum fermentation in Yuluxiang pear juice remains unexplored. The mechanisms by which this system influences fermentation kinetics, nutrient composition, and volatile profiles warrant systematic investigation. This study combines US treatment with L. plantarum fermentation to produce fermented Yuluxiang pear juice and systematically evaluates its effects on pH, clarity, browning index, soluble solids, total phenolic content, vitamin C, viable LAB counts, key enzyme activities (polyphenol oxidase (PPO), peroxidase (POD), superoxide dismutase (SOD)), sensory evaluation, color attributes, and flavor profile (GC-IMS), as well as and electronic-tongue (E-tongue) responses and overall fermentation quality. This study investigated US-assisted L. plantarum fermentation as a feasible strategy for improving the processing quality of Yuluxiang pear juice. By combining electronic-nose (E-nose), E-tongue, and GC-IMS, it established a multidimensional approach to evaluate flavor formation and quality changes during fermentation. The results are expected to deepen understanding of the power-dependent effects of US on fermentation behavior and quality attributes in pear juice, while providing a practical basis for reducing browning, enhancing product quality, and promoting the high-value utilization of Yuluxiang pear. Materials and methods Materials The Yuluxiang pears were procured from orchards in Xi County, a renowned Yuluxiang pear production region in Shanxi Province, China. Lactiplantibacillus plantarum (strain CICC-20265) was provided by the Dairy Laboratory, College of Food Science and Engineering, Shanxi Agricultural University. Pectinase was purchased from Tianjin Enzyme Preparation Factory (Tianjin, China). All chemicals were of analytical grade. Sodium chloride (NaCl), ascorbic acid, sucrose, methanol, oxalic acid, sodium carbonate (Na2CO3), potassium persulfate (K2S2O8), ferrous chloride (FeCl2), acetone, hydrogen peroxide pyrogallol, and phosphate buffer were supplied by Shanxi Tongjie Chemical Reagent Co., Ltd. (Yuncheng, Shanxi, China). Primary biochemical reagents, standard compounds, and culture media, including Folin-Ciocalteu reagent, gallic acid, DPPH (2,2-diphenyl-1-picrylhydrazyl), ABTS (2,2′‑azino‑bis(3-ethylbenzothiazoline-6-sulfonic acid)), ferrozine, catechol, guaiacol, pyrogallol and MRS agar powder, were procured from Aladdin Reagent (Shanghai) Co., Ltd. (Shanghai, China). Preparation of fermented pear juice Ripe and undamaged pears (soluble solids ≥ 13%) were selected, stored at 4 ℃, and processed within 24 h. After washing, peeling, and coring, the pears were cut into pieces and soaked in 0.1% (w/v) ascorbic acid solution for 15 min to prevent browning. To eliminate the influence of raw material variability (e.g., maturity, firmness, cell wall density) on fermentation performance, all pear samples used for fermentation were first subjected to a standardized US pretreatment (300 W, 15 min). This initial US step disrupts plant cell walls via cavitation effects, promoting the release of intracellular nutrients and ensuring a homogeneous substrate foundation across all fermentation groups. For fermentation, sucrose was added to the juice to a final concentration of 3% (w/v). The juice was then inoculated with 3% (v/v) L. plantarum seed culture (approximately 1×108 CFU/mL). The strain was stored at -80℃ in MRS broth supplemented with 20% (v/v) glycerol. Before use, it was subcultured three times on MRS agar at 37°C for 24 h. For seed preparation, a single colony was inoculated into MRS broth and incubated at 37℃ or 18 h. Cells were harvested by centrifugation and resuspended in sterile saline to approximately 1×108 CFU/mL. Fermentation was conducted at 37 ℃ in the dark. To evaluate US-assisted fermentation, the system received an US treatment (25 ℃, 15 min) at 2 h after inoculation. US power was set to 0 W (no-US fermentation control), 200 W, 300 W, 400 W, or 500 W. An uninoculated, non-US-treated group served as the blank control. In total, seven experimental groups were included. All inoculated groups were fermented at 37℃ for 24 h, rapidly cooled, and held at 4℃ for an additional 24 h to stabilize the fermented juice. This treatment ensured the stability of volatile components, achieved gas-liquid partition equilibrium, and reduced matrix interference. Moreover, it halted the metabolism of viable LAB and facilitated the precipitation of microbial cells and macromolecules. After a total of 48 h (24 h fermentation + 24 h cold storage), the samples were analyzed using GC–IMS, E-nose, E-tongue, descriptive sensory analysis. Color Difference Measurement Pear juice samples were loaded into cuvettes and equilibrated to room temperature. A colorimeter (CR-400, Konica Minolta, Tokyo, Japan) was calibrated using a standard white board. Each sample was measured in triplicate. The L*, a*, and b* values were recorded, and their means were calculated [13]. pH measurement The pH of fermented pear juice was dynamically monitored using a pH meter (Mettler-Toledo, Greifensee, Switzerland). Measurements spanned the entire fermentation and maturation process, beginning at 0 h. Samples were collected every 4 h until 24 h, covering seven time points (0, 4, 8, 12, 16, 20 and 24 h). Browning index The fermented pear juice treated with ascorbic acid for enzymatic browning inhibition served as the blank reference for instrument zeroing. The absorbance of each sample was measured at 420 nm using a spectrophotometer (Model WFJT200, Shanghai Mapada Instruments Co., Ltd., Shanghai, China). Unfermented raw pear juice was analyzed at the start of the experiment (pre‑inoculation) as the baseline [5]. Where Asample represents the absorbance of the sample at 420 nm, and Ainitial sample refers to the absorbance of the unfermented pear juice. Clarity The clarity of fermented pear juice was determined by spectrophotometry. The supernatant of the post‑fermentation sample was measured at 680 nm using distilled water as the reference. Transmittance T (%) was recorded, and the clarification rate (%) was calculated as follows: Where Ttreated represents the transmittance of the fermented sample, and Tinitial denotes the transmittance of the unfermented pear juice. Soluble solids Soluble solids content was measured using a refractometer (PAL-3, Atago Inc., Tokyo, Japan). The instrument was calibrated with distilled water prior to analysis. Clarified supernatant was applied to the prism, and °Brix was recorded at ambient temperature [14]. The prism was rinsed and dried between measurements. Total phenolic content Total phenolic content (TPC) was determined using the Folin-Ciocalteu method [15]. A gallic acid standard curve was generated by measuring absorbance at 765 nm. Appropriately diluted samples were analyzed identically. TPC was expressed as gallic acid equivalents (GAE) using: Where C is the determined concentration, V is the final volume, D is the dilution factor, and m is the sample mass. Vitamin C A standard stock solution of ascorbic acid (100 μg/mL) was prepared by dissolving the reference standard in 2% oxalic acid. The stock was diluted to a 10 μg/mL working solution and further diluted to generate a concentration gradient. Absorbance was measured at 265 nm to construct a standard curve. Samples were diluted with 2% oxalic acid and analyzed similarly. Vitamin C content was calculated from the standard curve and expressed as mg/100 g [16]. LAB count A 10 mL sample was serially diluted with sterile saline (10−1 to 10−6) under aseptic conditions. Based on preliminary tests, three consecutive dilutions yielding countable plates were selected. Aliquots (0.1 mL) were spread on MRS agar plates. Plates were incubated anaerobically at 37 ℃ for 48 h. Plates containing 30–300 colonies were counted, with three replicates per dilution [17]. Antioxidant activity Methanolic extracts were prepared by mixing samples with methanol and centrifuging at 5000 × g for 10 min. For DPPH scavenging, extract was mixed with 60 μmol/L DPPH methanolic solution, incubated for 30 min in the dark, and absorbance was measured at 515 nm [18]. For ABTS scavenging, ABTS+ stock solution was prepared by mixing 7 mmol/L ABTS with 2.45 mmol/L potassium persulfate (1:1, v/v) and incubating in the dark for 12–16 h. The solution was diluted with phosphate buffer to an absorbance of 0.70 ± 0.02 at 734 nm, mixed with extract, reacted for 6 min, and measured at 734 nm [19]. For ferrous‑ion chelating ability, sample, 2 mM FeCl2, and 5 mM ferrozine were added sequentially in a 96‑well plate, incubated at 27 °C for 10 min, and measured at 550 nm [19]. Determination of enzyme activities Fresh pear samples and fermented juice samples collected at twelve and twenty-four hours were used for enzyme activity analysis. For enzyme extraction, ten grams of sample were homogenized with ice-cold phosphate buffer solution at pH 6.5, and the homogenate was centrifuged at ten thousand times gravity for fifteen minutes at four degrees Celsius. The resulting supernatant was collected as the crude enzyme extract. PPO and POD activities were determined directly from this extract, whereas SOD activity was measured after acetone precipitation and re-dissolution in phosphate buffer at pH 7.8. PPO activity was assayed spectrophotometrically using catechol as the substrate, and the increase in absorbance at four hundred and twenty nanometers was recorded over one minute [5], [13]. POD activity was determined using guaiacol and hydrogen peroxide as substrates by monitoring the increase in absorbance at four hundred and seventy nanometers over one minute [5], [13]. SOD activity was measured by the pyrogallol autoxidation method at four hundred and twenty nanometers, and one unit was defined as the amount of enzyme required to inhibit fifty percent of pyrogallol autoxidation under the assay conditions [20]. PPO and POD activities were expressed as units per gram per minute based on the change in absorbance per minute per gram of fresh weight, whereas SOD activity was expressed as units per milliliter. All measurements were performed in triplicate, and the results were reported as mean values with standard deviations. Gas chromatography-ion mobility spectrometry (GC-IMS) A 1 mL aliquot of sample was transferred into a 20 mL headspace vial. Sodium chloride was added to facilitate volatile release, and the mixture was incubated at 60℃ for 20 min under continuous agitation at 500 rpm. For GC-IMS analysis, the injection needle temperature was set at 80℃, the autosampler was preheated to 80℃, the GC column oven temperature was maintained at 60℃, and the IMS detector temperature was kept at 45℃. High-purity nitrogen (99.99%) served as the carrier gas. The initial flow rate of 2 mL/min was held for 2 min, linearly increased to 100 mL/min over 18 min, and then maintained constantly for another 10 min. The drift tube temperature was 45℃, the drift gas flow rate was 150 mL/min, and a constant voltage of 5 kV was applied throughout the analysis. A series of n-ketones (C4-C9) were used for retention index (RI) calibration. Volatile compounds were identified by matching experimental drift time (DT) and RI values against the GC-IMS NEST library and NIST database using the GC-IMS Library Search software. Quantitative results were expressed as peak areas. The internal standard method was adopted to calculate relative response factors, thereby enabling reliable semi-quantitative analysis of volatile profiles. The GC-IMS analytical parameters were established based on previously reported methods for fermented fruit products [21]. E-nose analysis Volatiles were analyzed using an E-nose (PEN3, Airsense Analytics, Germany) with a metal‑oxide sensor array. Samples were equilibrated at 25 ℃ for 20 min. Sensors were purged with clean air at 300 mL/min for 300 s before measurement. Headspace was sampled for 60 s, followed by 120 s of sensor response recording. Stable signals were used for subsequent analysis [22]. E-tongue analysis After centrifugation and membrane filtration, samples were analyzed using a portable E-tongue (PEN2, Airsense, Germany) equipped with a multichannel electrochemical sensor array. After calibration, each sample was measured in triplicate to obtain stable responses. Signals were amplified, filtered, digitized, and analyzed by chemometrics [23]. Descriptive sensory analysis The study followed strict ethical standards. All participating students provided written informed consent and were informed of their right to withdraw unconditionally at any time. As the evaluation involved only commercially available edible products and no invasive procedures, it was exempt from ethical review approval according to the institutional guidelines of Shanxi Agricultural University. The design and implementation of the sensory evaluation methodology strictly adhered to relevant guidelines issued by the International Organization for Standardization to ensure procedural standardization and result reliability. This study recruited 15 food science students as panelists (8 females and 7 males, aged 22-26 years). All participants underwent preliminary screening and received two weeks of systematic training. The training covered basic taste recognition, intensity scaling, and familiarization with product characteristics. The final panel was established after passing repeatability tests and consistency checks. A structured scale was used for the evaluation. Descriptive sensory analysis was performed by ten trained assessors using a structured scoring sheet covering clarity, aroma, taste, and mouthfeel. Each attribute included graded descriptors with corresponding scores; the total score was 100. (e.g., clarity was categorized as “juice clear without sediment,” “slightly cloudy with no obvious sediment,” “some sediment, relatively cloudy,” or “severely cloudy with obvious sediment,” with scores ranging from 18-20 to 0-9). Similar graded standards were applied to aroma, taste, and mouthfeel. Panelists evaluated samples independently, and scores were averaged. Data processing All experiments were performed in three biological replicates, and data are presented as mean ± standard deviation. One-way ANOVA followed by Duncan's multiple range test was conducted using SPSS 26.0 (IBM Corp., Armonk, NY, USA), with p < 0.05 considered statistically significant. Multivariate statistical analyses, including principal component analysis (PCA) and partial least squares-discriminant analysis (PLS-DA), were performed using SIMCA 14.1 (Umetrics, Sweden). Prior to modeling, the data were mean-centered and auto-scaled to minimize scale-related bias and to enhance model robustness. Hierarchical cluster analysis (HCA) and heatmaps of the key volatile compounds were generated using OriginPro 2021 (OriginLab Corp., Northampton, MA, USA). Results and Discussion Effect of US treatment on color variation during fermentation of pear juice As shown in Fig. 1A, pear juice gradually darkened during fermentation, transitioning from a pale yellow-milky white to a deeper hue. In the early fermentation stage (0-4 h), the color parameters (a*, b*, L*) of all treatment groups increase significantly, indicating a shift towards a more reddish, yellowish, and brighter appearance. Notably, the 400 W group reached a peak L* value of 21.25 at 4 h, significantly higher than other groups (p < 0.05), suggesting that appropriate US power can enhance brightness in the initial phase. In the middle-to-late stage (4-24 h), the groups exhibited divergent color trajectories. The a* value fluctuates and shows varying degrees of recovery at the endpoint (24 h), with the 200 W group rising to +0.14, significantly higher than others (p < 0.05), indicating that lower US power favored redness in the later stage (Fig. 1B). Regarding yellowness, the b* values of the US-treated groups (200-500 W) remain high at 12 h and are significantly higher than the unfermented control (-2.36) (p < 0.05) indicating that US helps maintain the yellow tone (p < 0.05) (Fig. 1C). L* increases early on and generally declines from 4-12 h, possibly due to non-enzymatic browning (Fig. 1D). However, the L* value of the 200 W group rises significantly after 16 h (p < 0.05), staying high (19.36) at 24 h, along with the 500 W group (18.54), indicating the highest endpoint brightness. Mechanistically, medium-to-low power US (200-300 W) physically disrupts plant cell walls and microbial membranes via cavitation [23]. This not only promotes the release of pigment precursors such as polyphenols and flavonoids from pear cells [24], but also enhances microbial membrane permeability, thereby improving nutrient uptake and metabolic efficiency, ultimately influencing the fermentation process and the synthesis of related pigment metabolites [25]. Effect of US treatment on the physical photo (A), a*(B), b* (C) and L* (D) of fermented pear juice. Asterisks indicate significant differences between the untreated control (0 W) and each US-treated group (independent-samples t-test; *p < 0.05, **p < 0.01, ***p < 0.001). Effect of US treatment on the pH of fermented pear juice As shown in Fig. 2A, the pH of all fermentation groups decreased significantly within the first 24 h, confirming sustained acid production by microorganisms such as LAB. At 0 h, initial pH values were comparable among treated groups (4.60-4.90), whereas the unfermented control showed the highest pH (4.90) with minimal fluctuation. During fermentation, the pH decline rate in US‑treated groups generally exceeded that of the untreated fermentation group (0 W). At 24 h, the 400 W group reached the lowest pH (3.65), significantly lower than the 0 W group (3.85) and the 200 W group (3.90) (p < 0.05), while differences versus the 200 W and 500 W groups were not significant. The accelerated acidification is closely associated with cavitation, which disrupts pear tissue structures and promotes the release of fermentable substrates (sugars), thereby enhancing carbon availability and accelerating acid production [26]. The optimal effect at 400 W likely reflects a balance between cavitation intensity and microbial viability; higher power (500 W) may induce localized overheating and free‑radical formation that partially suppress microbial activity. Conversely, lower power (200 W) or no US results in insufficient disruption and limited substrate release, slowing acidification [27]. This finding aligns with the work of Piecko et al., who demonstrated that US treatment significantly enhances the soluble sugar content in juice, consequently promoting acid production during fermentation [28]. Overall, 400 W achieved the most rapid pH decline under the tested conditions. Effect of US treatment on the pH (A), browning rate(B), percentage transmittance (C) and total soluble solids content (D) of fermented pear juice. Different letters represent significant differences at p < 0.05. Effect of US treatment on browning rate and clarity of fermented pear juice As shown in Fig. 2B, the browning rate increased continuously in all groups with fermentation time. Differences were minor at 2-6 h but became pronounced after 12 h. At 24 h, the control group exhibited the highest browning rate (∼285%). In contrast, the 300 W group maintained the lowest browning rate throughout, indicating the most effective browning inhibition. When power increased to 400 W and 500 W, inhibition did not improve; instead, browning increased in later stages. This may be attributed to excessive cavitation at high power, which over‑disrupts cell structures and promotes the release of phenolics and their contact with polyphenol oxidase (PPO), thereby intensifying enzymatic browning. Appropriately powered US (300 W) has been reported to alter PPO conformation and reduce activity, delaying browning, whereas overly strong sonication can rupture membranes and expose more phenolics to oxygen and PPO, accelerating browning [29]. As shown in Fig. 2C, transmittance increased significantly over time across all fermentation groups, indicating an intrinsic clarification effect of fermentation. US significantly modulated clarification kinetics. At 18 h, the 300 W group achieved 60.78% transmittance, significantly higher than the control (48.85%) and other treatments (p < 0.05). This advantage peaked at 24 h (61.20%). Other power groups displayed an inverted “U-shaped” trend in final clarification performance, with the 400 W group performing the worst. US enhances clarification via cavitation‑induced disruption of suspended particles and colloidal structures and by facilitating enzymatic degradation and flocculation of soluble solids. The 300 W treatment appears to balance energy input and cavitation intensity, effectively disrupting tissues while avoiding excessive thermal/sonochemical stress that may impair microbial activity and destabilize the fermentation environment. In contrast, excessive power (400-500 W) may cause localized overheating and strong micro‑jets, leading to protein denaturation, partial microbial inactivation, and excessive colloid disruption, thereby impairing clarification [30]. Moreover, the control and 0 W fermentation groups consistently exhibited lower transmittance, confirming that microbial metabolism-driven degradation of pectin and related substances underpins the clarification foundation. However, this metabolic pathway operates significantly less efficiently than the synergistic effect observed in US-assisted fermentation [31]. Collectively, a 300 W US treatment optimally balances cavitation and enzymatic reactions, achieving the best clarification with the lowest browning within 24 h. Effect of US treatment on total soluble solids in fermented pear juice As shown in Fig. 2D, total soluble solids (TSS) decreased continuously during 24 h fermentation, reflecting carbohydrate consumption. The non‑fermented control remained stable (5.50–6.00 °Brix), whereas all fermentation groups showed significant decreases (p < 0.05). TSS reduction was enhanced in sonicated groups relative to 0 W. The 300 W group showed the lowest final TSS (∼4.40 °Brix), significantly lower than other groups (p < 0.05). The 500 W group exhibited a rapid initial decrease followed by a slower decline later, likely due to high‑power‑induced stress. This effect is primarily due to cavitation and mechanical shear forces induced by US, which disrupt cell wall integrity, enhancing substrate release and microbial access, thus accelerating sugar utilization. These findings suggest that moderate US treatment (300 W) effectively enhances polyphenol and soluble sugar release, improving fermentation efficiency. The high energy input of the 500 W treatment likely caused localized heating, which presumably inhibited LAB activity and slowed the later fermentation rate [27]. Effect of US treatment on polyphenol and ascorbic acid content in fermented pear juice As shown in Fig. 3A, fermentation (0 W) increased polyphenol content from 1.27 ± 0.04 mg/100 mL (control) to 1.36 ± 0.03 mg/100 mL, indicating that fermentation promotes phenolic release and biotransformation [24]. Polyphenol content showed a non‑linear response to US power, peaking at 1.44 ± 0.03 mg/100 mL at 300 W. In contrast, 200 W and 500 W produced significantly lower values than the control and 0 W groups (p < 0.05). These results indicate that moderate US treatment (300 W) effectively enhances polyphenol content. The underlying mechanism likely involves optimal cavitation effects that disrupt cellular structures, thereby promoting polyphenol release from cellular matrices and enhancing microbial metabolic activity. Conversely, excessive power input (500 W) may generate intense mechanical shear forces, which could lead to oxidative degradation of polyphenols and/or inhibit microbial activity, ultimately reducing the polyphenol content [32]. The lower polyphenol content observed at 200 W may result from insufficient US treated intensity, which was inadequate for releasing intracellular polyphenols [33]. Effect of US treatment on the total phenolic content (A), ascorbic acid content (B) and LAB count (C) of fermented pear juice. Different letters represent significant differences at p < 0.05. Different lowercase letters indicate significant differences among US power groups at 12 h, whereas different uppercase letters indicate significant differences among US power groups at 24 h (one-way ANOVA followed by Duncan's multiple range test, p < 0.05). Asterisks indicate significant differences between 12 h and 24 h within the same US power group (independent-samples t-test; *p < 0.05, **p < 0.01, ***p < 0.001). As shown in Fig. 3B, the fermentation process itself caused significant degradation of ascorbic acid. In the non–US treated group (0 W), the content decreased from 26.53 ± 0.73mg/100 g (control) to 11.20 ± 0.41 mg/100 g. This reduction is associated with enhanced microbial metabolism and oxidative stress [34]. Ascorbic acid responded nonlinearly to US power: 15.18 ± 0.26 mg/100 g (200 W), 20.17 ± 0.53 mg/100 g (300 W), a peak of 26.97 ± 0.57 mg/100 g (400 W; comparable to control), followed by 22.01 ± 0.28 mg/100 g (500 W). The protective effect of 400 W US on vitamin C likely arose from the combined action of several mechanisms. First, cavitation-induced shear force disrupted cellular structures and promoted the release of bound vitamin C and phenolic antioxidants into the juice matrix, which was consistent with the continuous increase in FRAP reducing power from 43.30% to 85.07% [37]. Second, the released phenolics, including chlorogenic acid and catechins, could preferentially scavenge hydroxyl radicals generated during cavitation, thereby alleviating oxidative consumption of vitamin C. Third, the suppression of PPO activity under moderate US further limited polyphenol oxidation and indirectly contributed to vitamin C retention [35]. This protective effect was no longer maintained at 500 W. Under excessive US intensity, over-cavitation likely accelerated hydroxyl radical generation beyond the scavenging capacity of the endogenous antioxidant system, resulting in direct oxidation and subsequent degradation of vitamin C. Meanwhile, transient localized heating produced during intense cavitation may have further promoted vitamin C loss. Excessive mechanical shear could also disrupt phenolic structures and weaken their radical-scavenging ability, thereby reducing their protective role in the juice system [34]. Collectively, these results suggest that vitamin C retention depended on a dynamic balance between antioxidant release and US-induced oxidative stress, and that this balance was most favorable under moderate US treatment. Effect of US treatment on LAB counts in fermented pear juice As shown in Fig. 3C, LAB counts at 12 h were low and comparable among groups (1.29–1.94 × 107 CFU/mL). A substantial increase in LAB populations was observed across all US-treated samples by 24 h. Among them, the 300 W treatment yielded the highest count (approximately 1.32 × 109 CFU/mL), which was significantly greater than that of all other groups (p < 0.05). In contrast, LAB counts in the 400 W and 500 W groups were significantly lower, reaching only about 3.91 × 108 and 6.00 × 108 CFU/mL, respectively. Notably, the 500 W group exhibited the lowest count at 12 h (∼1.27 × 107 CFU/mL), indicating an early inhibitory effect of high-power US on bacterial growth. The enhanced LAB proliferation under moderate US (300 W) is primarily attributed to improved substrate availability. The cavitation and mechanical effects of US disrupt macromolecular structures, such as polysaccharides and proteins in pear juice, thereby releasing additional fermentable sugars [35]. Concurrently, US increases cell membrane permeability, which facilitates nutrient uptake by the bacteria [36]. Conversely, excessively high power levels (400–500 W) generate intense cavitation and mechanical stress that can directly damage cellular integrity, leading to content leakage and cell death. This detrimental effect is particularly pronounced during the early stages of fermentation [37]. Furthermore, the localized temperature rise induced by high-power treatment may exceed the optimal growth temperature range for most LAB strains (30–45°C), thereby further suppressing their metabolic activity [38]. Effect of US treatment on the antioxidant capacity of fermented pear juice The effect of US power on the antioxidant activity of fermented pear juice was evaluated using FRAP, DPPH, and ABTS assays (Fig. 4A-C). The non‑US treated fermented group (0 W) exhibited significant increases in FRAP, DPPH, and ABTS values compared to the non-fermented control (p < 0.05). This enhancement is attributable to microbial metabolites such as phenolics, organic acids, and polysaccharides produced during fermentation [39]. The FRAP value increased continuously with US power, rising from 43.30% to 85.07%. The 500 W treatment yielded a six-fold increase over the non-fermented control, indicating power-dependent enhancement of reducing capacity. However, both DPPH and ABTS values peaked at intermediate power levels (300-400 W), with ABTS reaching its maximum (75.75%) at 400 W. At 500 W, the DPPH scavenging decreased significantly to 64.78% (p < 0.05), while the ABTS value declined from its peak. This bell-shaped response could be attributed to the cavitation and mechanical shear generated between 200 W and 400 W, which effectively disrupt cell walls and facilitate polyphenol and flavonoid release [26]. US also intensified microbial metabolic activity and promoted biotransformation of active components [33]. Furthermore, moderate US treatment may induce oxidative stress, thereby activating the internal antioxidant defense system of plant cells [40] However, when the power was increased to 500 W, excessive mechanical force and localized overheating could partially degrade heat-labile antioxidants (polyphenols) and potentially disrupt microbial consortia or inhibit key enzyme activity, ultimately weakening radical scavenging capacity [41], [42]. Overall, 400 W maximized free‑radical scavenging, whereas higher power compromised this benefit. Effect of US treatment on antioxidant capacity and enzyme activities in fermented pear juice. Ferric ion reducing antioxidant power (FRAP) (A), DPPH radical scavenging activity (B), ABTS radical scavenging activity (C), polyphenol oxidase (PPO) activity (D), peroxidase (POD) activity (E), and superoxide dismutase (SOD) activity (F). Different lowercase letters indicate significant differences among US power groups at 12 h, whereas different uppercase letters indicate significant differences among US power groups at 24 h (one-way ANOVA followed by Duncan's multiple range test, p < 0.05). Asterisks indicate significant differences between 12 h and 24 h within the same US power group (independent-samples t-test; *p < 0.05, **p < 0.01, ***p < 0.001). Effect of US treatment on the activities of key endogenous enzymes US markedly affected PPO, POD, and SOD activities in a power-dependent manner. Overall, moderate US was more favorable for suppressing browning-related reactions and enhancing antioxidant-associated enzyme activities, whereas excessive power attenuated these effects. PPO activity was strongly reduced by moderate US (Fig. 4D), particularly at 300 W. At 12 h, PPO activity decreased from 462.0 U/(g·min) in the 0 W to 136.0 U/(g·min) at 300 W, and remained significantly lower than the 0 W at 24 h (132.5 vs. 215.0 U/(g·min), p < 0.05). This pattern was consistent with the lower browning index observed in the 300 W group, suggesting that moderate US effectively suppressed enzymatic browning. By contrast, PPO activity partially recovered at 500 W, indicating that excessive cavitation may have promoted substrate exposure and weakened the inhibitory effect of US [29]. POD activity (Fig. 4E) showed the opposite trend and reached the highest level at 300 W. Compared with the 0 W, the 300 W treatment increased POD activity from 390.0 to 826.7 U/(g·min) at 12 h and from 253.3 to 586.7 U/(g·min) at 24 h (p < 0.05). The enhanced POD activity was in accordance with the higher total phenolic content and volatile accumulation in this group, indicating that moderate US favored the release or transformation of bound compounds [11]. However, POD activity declined at higher power levels, decreasing to 450.0 and 337.0 U/(g·min) in the 400 and 500 W groups at 24 h, respectively, which suggests partial enzyme inactivation under excessive ultrasound intensity. SOD activity (Fig. 4F) was also significantly enhanced by US and was highest at 300-400 W. At 12 h, SOD activity increased from 3.9 U/mL in the 0 W to 14.4 and 13.9 U/mL in the 300 and 400 W groups, respectively. At 24 h, the maximum value was detected in the 400 W group (19.4 U/mL), followed by the 300 W group (16.4 U/mL), both significantly higher than the 0 W (p < 0.05). The elevated SOD activity agreed well with the improved antioxidant capacity of the corresponding samples and may also reflect enhanced oxidative stress tolerance during fermentation [33], [40]. In contrast, SOD activity decreased to 10.0 U/mL at 500 W, coinciding with reduced viable LAB counts and slower acidification [37]. Taken together, US induced a clear non-linear enzymatic response. Among the tested treatments, 300 W provided the most balanced regulation of endogenous enzymes, as evidenced by pronounced PPO inhibition, high POD activity, and sustained SOD enhancement, which collectively contributed to reduced browning and improved functional quality of fermented pear juice. Effect of US treatment on volatile flavor compounds in fermented pear juice Gas chromatography-ion mobility spectrometry (GC-IMS) was employed to analyze the impact of US power on the volatile metabolome of pear juice after 48 h fermentation. The non-fermented control displayed weak signals confined to low drift times (< 6 ms), indicating a limited volatile repertoire (Fig. 5A&B). With increasing US power, the signal intensity and number of peaks rose markedly. Treatment groups ≥ 300 W produced pronounced high-signal zones at 6-8 ms, suggesting sonication-driven formation of key aroma compounds. The fingerprint map revealed region-specific responses to US treatment (Fig. 5C). In Region A, ethyl butanoate, 3-methylbutanal, and propionic acid, which were abundant in the control and possess fruity and malty aromas, diminished or disappeared in all fermented samples. This observation confirms that native volatile components of pear juice were metabolized or bioconverted during fermentation [43], [44]. Region B showed that 3-methylbutanol, (E)-3-hexen-1-ol, and isobutyl formate, which have fusel-like, fruity, and grassy aromas, peaked in the 0 W group but declined progressively as power increased (200-500 W). This decline likely results from oxidative degradation of these oxidation-sensitive alcohols by intense cavitation, local hotspots, and hydroxyl radicals (·OH) [45]. In Region C, 2-propanol and several acids/esters were significantly higher in the 400 W group. This suggests that elevated power intensified microbial metabolic pathways, probably by enhancing mass transfer efficiency [46]. US increases the gas-liquid interfacial area and reduces bubble size, elevating the volumetric mass-transfer coefficient (KLa) and optimizing metabolic flux [47]. Region D exhibited a stepwise increase in linalool, (Z)-4-heptenal, and 2-methylbutyl acetate, which have floral and fruity aromas, from 0 W to 500 W. The accumulation of linalool likely stems from sonication-induced cell disruption, which liberates terpenoid precursors and/or activates relevant enzyme systems [48]. Effect of US treatment on the 3D topographic plots of GC-IMS (A), 2D topographic plots of GC-IMS (B), and fingerprints of GC-IMS (C) of fermented pear juice. Multivariate statistical methods were applied to dissect the US-induced changes in volatile flavor compounds in fermented pear juice. Principal component analysis (PCA) explained 90.35% of total variance (PC1 + PC2, R2 = 0.9484, p = 0.0010), demonstrating strong model explanatory power. The score plot (Fig. 6A) showed clear separation between the control group and all sonicated groups along PC2. Furthermore, the sonicated samples shifted rightward with increasing power, evidencing dose-dependent modulation of flavor characteristics. The loading plot (Fig. 6B) revealed that linalool, (Z)-4-heptenal, and 2-methylbutyl acetate dominated the positive PC1 axis, correlating closely with high-power treatments (500 W). In contrast, 3-pentanone and pentyl acetate aligned with the negative PC2 axis, linking them to the low-power groups. The heatmap (Fig. 6C) visualized the relative enrichment of esters (e.g., 2-methylbutyl acetate) and aldehydes ((Z)-4-heptenal) under high-power sonication (400 W, 500 W). The control and low-power samples were characterized by ketones such as 3-pentanone. Concurrently, the enrichment of terpenes like linalool and 3-methyl-3-buten-1-ol in fermented groups suggested that microbial metabolism activated the conversion of endogenous precursors. Variable importance in projection (VIP) analysis (Fig. 6D) highlighted key differentiating factors with VIP > 1. Among these, 2-methylbutyl acetate and linalool exhibited the highest VIP values in high-power groups, qualifying them as signature aroma markers of US-assisted fermentation. The specific changes of the top-eight VIP differential flavor compounds are shown in Fig. 6E. These changes are closely tied to sonication-induced modulation of microbial metabolism. Studies have shown that US-induced cavitation effects improve cell membrane permeability and mass transfer, boosting metabolic efficiency [49]. High-power US releases polyphenols, sugars, and amino acids in pear juice, supplying abundant precursors that accelerate alcohol formation and subsequent esterification. This explains the observed surge in pleasant esters and aldehydes [50]. Concurrently, US may selectively reshape the microbial community structure or enzyme profile, suppressing the accumulation of unpleasant odor substances such as ketones [51]. Effect of US treatment on the PCA score plot of GC-IMS (A), bioplot score plot of GC-IMS (B), heatmap of GC-IMS (C), VIP > 1 of GC-IMS (D) and top 8 VIP value differential flavor compounds of GC-IMS (E) of fermented pear juice. Different letters represent significant differences at p < 0.05. Effect of US treatment on E-nose results of fermented pear juice As shown in Fig. 7A, the cumulative contribution rate for PCA of E-nose data was 92.61%. The PC1 and PC2 accounted for 87.73% and 4.88% of total variation, respectively. The non-fermented control and 0 W fermented group clustered on the positive side of PC1, whereas all sonicated fermented samples shifted to the negative side. This indicates that fermentation itself is the primary driver of fundamental transformation in volatile composition. Within fermented samples, a progressive displacement along PC1 accompanied the increase in US power, with the 500 W group deviating most strongly. This evidences a dose-dependent relationship between sonication intensity and flavor profile. The loading plot (Fig. 7B) revealed that sensors W5S, W1S, and W2S contributed to the positive PC1 region, while W6S and W2W aligned with the negative region and thus with sonicated samples. The radar plots (Fig. 7C) showed higher sensor responses for all sonicated groups, especially for W1S (sensitive to esters) and W2S (sensitive to alcohols and aldehydes). The 400 W and 500 W treatments produced the highest responses (p < 0.05), confirming that high-power US effectively promotes the formation of specific volatile flavor substances. This phenomenon is attributed to the mechanical forces and microjets generated by US treated cavitation, which disrupt pear cell structures and release more substrates for esterification (e.g., free fatty acids and alcohols). The local extreme conditions generated during cavitation bubble collapse accelerate mass transfer, promoting the synthesis of esters such as 2-methylbutyl acetate and ethyl 2-methylpropionate [52], [53]. Previous studies have demonstrated that US pretreatment can enhance the esterification activity of lipase by altering its conformation [54], [55]. Similarly, the elevated responses of W2S and W5S reflect sonication-driven sugar degradation and amino acid metabolism, yielding more aromatic precursor substances such as alcohols and aldehydes. Effect of US treatment on the PCA score plot of electronic nose analysis (A), bioplot score plot of electronic nose analysis (B), radar chart of electronic nose components (C), PCA score plot of electronic tongue analysis (D), bioplot score plot of electronic tongue analysis (E), radar chart of electronic tongue components (F) and radar chart of sensory evaluation (G) of fermented pear juice. The effect of US treatment on the E-tongue results of fermented pear juice The PCA showed that the unfermented pear juice control (Control) clustered in the positive region of PC1, clearly separated from all fermented groups (Fig. 7D). As the US power increased from 200 W to 500 W, the sample points shifted progressively along the negative direction of PC1 and toward the negative side of PC2, forming a power–dependent gradient. The loading plot (Fig. 7E) indicated that sourness was positioned opposite to richness and umami along PC1, suggesting a negative correlation, whereas bitterness primarily contributed to PC2. Fermentation combined with US moved samples along negative PC1, suggesting reduced sourness and increased body/umami, with stronger effects at higher power. Radaranalysis (Fig. 7F) further showed that fermented groups had higher sourness and aftertaste than the control, peaking at 400 W. Meanwhile, bitterness and astringency decreased across treatments, reaching minima at 400 W. Moderate US (400 W) may promote microbial metabolism and membrane transport, increasing organic acid production and ester synthesis [56]. Moreover, the mechanical and sonochemical effects of US efficiently degrade polyphenolic compounds, leading to a substantial reduction in bitterness and astringency [57]. However, excessive power (500 W) may cause over-mechanical damage or thermal effects that inhibit microbial activity, resulting in a decline in certain sensory attributes. Overall, 400 W provided the best balance among sourness, bitterness/astringency suppression, and aftertaste enhancement. Effect of US treatment on sensory evaluation scores of fermented pear juice As shown in Fig. 7G, all fermented groups scored significantly higher than the non‑fermented control in color, aroma, flavor, and overall acceptability, confirming the key role of fermentation in sensory improvement. Among fermented samples, overall scores showed a non‑linear dependence on US power, peaking at 300 W, where each attribute achieved the highest score. When the power increased to 500 W, a decline in sensory scores was observed. This decline may be attributed to excessive cavitation effects induced by high power, which could cause over-destruction of cellular structures, leading to the release of off-flavor compounds and degradation of aromatic components [58]. Additionally, all fermented groups displayed significantly superior aroma and flavor dimensions compared to the control group, indicating that the fermentation process effectively generated aromatic substances such as esters and alcohols, thereby enhancing flavor quality [59]. US treatment further optimized fermentation efficiency and product distribution by enhancing enzyme activity, promoting substrate contact, and accelerating the generation of metabolic products [12], with the optimal balance point identified at 300 W. Correlation analysis and involved mechanism Pearson correlation analysis combined with Mantel test revealed the complex association network among sensory attributes, volatile flavor compounds, and physicochemical parameters of fermented pear juice subjected to US treatment (Fig. 8A). Mantel test revealed that sourness and umami exhibited significant positive correlations with ABTS and browning rate (r ≥ 0.5, p < 0.05). The coordinated alterations of these sensory attributes may have been attributed to US-induced organic acid liberation and polyphenol biotransformation. Total polyphenol content, PPO activity and POP activity demonstrated highly significant positive correlations with color, including L* (r ≥ 0.75, p < 0.05), a* (r ≥ 0.75, p < 0.05) and b* (r ≥ 0.75, p < 0.05), confirming its predominant contribution to the overall color of the fermented juice. Acetic acid propyl ester demonstrated highly significant positive correlations with browning rate, pH and total soluble solida contents (r ≥ 0.75, p < 0.05). This suggests that the coordinated alterations of these quality parameters may have been attributed to US induced esterification reactions and the release of soluble sugars. Furthermore, total phenolic content exhibited a significant negative correlation with butyl butanote and 2-methyl-1-propyl acetate (r ≤ -0.75, p < 0.05), suggests that the coordinated alterations of these volatile compounds may have been attributed to phenolic oxidation and subsequent binding or degradation processes. PPO activity and POP activity were significantly negatively correlated with Acetic acid propyl ester, 2-methyl-1-propyl acetate, and Butyl butanoate (r ≤ -0.75, p < 0.05). This indicates that during hair or processing, when the activity of phenoloxidase increases, there will be a significant browning reaction on the surface, leading to a darkening of color and a significant decrease in the content of acid esters, ultimately resulting in a weakening of the fruity and ester aromas on the surface. The characteristic differential flavor compounds identified in this study were validated using semi-quantitative GC-IMS data and multiple chemometric approaches. The inter-group trends in their signal intensities have been corroborated through electronic nose responses, physicochemical correlation analyses, and statistical testing. Collectively, this correlation network elucidates the mechanistic pathways underlying US-assisted fermentation: acoustic cavitation effects modulate the physicochemical microenvironment (pH, optical transmittance, and cellular permeability), thereby regulating LAB metabolic activity and facilitating polyphenol transformation, which synergistically contribute to the enhanced flavor characteristics and antioxidant quality of the final product. Effect of US treatment on the Pearson correlation analysis (A) and involved mechanism (B) of fermented pear juice. The mechanism of US‑assisted fermentation in Yuluxiang pear juice is illustrated in Fig. 8B. US treatment significantly accelerated the proliferation of LAB and showed potential to improve juice brightness at the early fermentation stage. As fermentation proceeded, non‑enzymatic browning gradually intensified, leading to a darker juice color. Fermentation itself enhanced the antioxidant capacity of the system, and US further amplified this effect, primarily by intensifying microbial metabolic activity during fermentation and promoting the conversion of precursors into more active antioxidant compounds. In addition, US stimulated the formation of flavor compounds such as aldehydes and terpenes. The cavitation effect facilitated the release of substrates-including polyphenols, sugars, and amino acids-from the pear juice, thereby supplying richer precursors for LAB. This process accelerated alcohol formation and subsequent esterification reactions, which explains the significant increase in pleasant aromatic compounds, including esters and aldehydes, observed in this study. Conclusion This study revealed the non-linear, power-dependent effects of US treatment on Lactiplantibacillus plantarum fermentation of Yuluxiang pear juice. Moderate US (300 W) appeared to be optimal, yielding the lowest browning index and the highest LAB viability, whereas excessive power (500 W) compromised fermentation performance, likely due to over-cavitation. Enzymatic assays further suggested that US-induced cavitation may suppress PPO activity while enhancing antioxidant capacity. By integrating E-nose, E-tongue, and GC-IMS profiling with Pearson and Mantel correlation analyses, a mechanistic framework linking flavor formation to processing variables was established, in which linalool, 2-methylbutyl acetate, and (Z)-4-heptenal were identified as potential quality markers. Nevertheless, the present work is limited by the use of a single-strain system and the semi-quantitative nature of GC-IMS; targeted HS-SPME-GC–MS validation and systematic investigation of US frequency/pulse modes are therefore warranted. Future studies incorporating multi-strain co-fermentation and multi-omics approaches are expected to further clarify the molecular basis of flavor biosynthesis under US-assisted fermentation. CRediT authorship contribution statement Huiling Duan: Writing – original draft, Data curation, Conceptualization. Danrong Yu: Writing – original draft, Methodology. Chuan Xie: Writing – original draft, Data curation. Chuanwei Zhang: Writing – original draft, Data curation. Yaping Shen: Writing – original draft, Data curation. Ling Ma: Writing – original draft, Data curation. Liyan Jia: Writing – original draft, Data curation. Shicong Zhang: Writing – review & editing. Xiaoyu Zhang: Data curation, Conceptualization. Zhihui Yu: Writing – review & editing, Supervision, Formal analysis, Data curation, Conceptualization. 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. Acknowledgment This work was supported by the Doctoral Research Initiation Project of Shanxi Agricultural University (Grant No. 2021BQ29) and the Quwo County National Modern Agricultural Industrial Park Project (Doctoral Workstation Construction Project) (Grant No. SXRTFWZB2506-26-2). Contributor Information Xiaoyu Zhang, Email: xiaoyuzhang2005@163.com. Zhihui Yu, Email: zhihuisxau@163.com. References