Study on the effects of media-regulated ultrasonic pretreatment combined with elwectrohydrodynamic drying on the drying kinetics, active components, proteomic characteristics and volatile components of garlic

Abstract

This study investigated the effects of different pretreatment media-NaCl solution (NaCl aq.), sucrose solution (Suc aq.), and deionized water (DW)-on the combined ultrasonic (US)-electrohydrodynamic (EHD) drying of garlic. Comprehensive analysis revealed that the NaCl aq. + US pretreatment demonstrated optimal performance. It markedly improved the effective moisture diffusion coefficient (D_eff), thereby accelerating the overall drying kinetics, and enhanced the rehydration capacity, indicating better preservation of the garlic's microstructure. This pretreatment also promoted superior retention of allicin-derived compounds and total phenols compared to other methods. Furthermore, it helped preserve protein secondary structure, promoted the release of alliinase, and maximized the content of volatile sulfur compounds. These findings first reveal the synergistic mechanism of medium-regulated ultrasound, where osmotic pressure and cavitation effects jointly enhance bioactive compound retention. The results provide both theoretical and practical support for optimizing combined drying processes in garlic processing.

Graphical abstract

Highlights

• •Medium type critically affects ultrasound-assisted electrohydrodynamic drying of garlic.
• •NaCl + ultrasound yields higher bioactive content in garlic.
• •NaCl + ultrasound promotes alliinase release and sulfur volatiles.
• •Provides a theoretical-experimental basis for optimizing bulb vegetable drying.

1. Introduction

Garlic (Allium sativum L.) is an important economic crop and edible plant. As the world's largest producer, China has an annual output of 23–24 million tons in China, accounting for approximately 78% (Ammarellou, Yousefi, Heydari, Uberti, & Mastinu, 2022; Song et al., 2025). Modern research showed that garlic is rich in active ingredients such as organic sulfur compounds (like allicin), proteins, and polyphenols. However, the retention of these bioactive compounds during processing is challenging due to enzyme inactivation and structural destruction (Han et al., 2023). Allicin, as a core antibacterial ingredient, has a content level that directly determines the functional value of the product. Alliinase, a pivotal enzyme in garlic, is not only a key medium for enzymatic reactions but also a factor that affects texture characteristics and nutritional quality (L. Li et al., 2025). In terms of application, garlic is not only an important ingredient for condiments, but also widely used in health food and medicine, highlighting its significant economic value and health benefits (H. Liu, Mao, Wang, Wang, & Xie, 2015).

Fresh garlic contains more than 70% water, is prone to sprouting and mold, and is difficult to store for a long time. Scientific drying can effectively reduce moisture, inhibit quality deterioration, and extend shelf life. This feature has made processed products such as dehydrated garlic slices and garlic powder an important category in the market, ensuring product stability and safety by controlling water activity (Han et al., 2023; H. Liu et al., 2015). Conventional garlic drying methods each have distinct characteristics. Hot air drying has become the mainstream choice for industrial production due to its stable and efficient characteristics, but it has the disadvantages of high energy consumption and loss of heat-sensitive components (Wongsa, Bhuyar, Tongkoom, Spreer, & Müller, 2022). Freeze-drying can retain the active ingredients and flavor substances of garlic to the greatest extent, and the product quality is excellent (Thakur, Dhiman, Kumar, & Gautam, 2025), but the high equipment investment and operating costs limit its application range. Other drying methods have problems with energy efficiency or scale (Makarichian, Chayjan, Ahmadi, & Mohtasebi, 2021). This comparison establishes a clear technical benchmark and identifies research directions for advancing drying technologies in garlic processing.

EHD drying is a low-temperature technology that utilizes a high voltage (10–40 kV) applied between emitter and grounded collector electrodes, generating ionic wind (0z.1–10 m/s) to accelerate moisture evaporation and internal migration while minimizing loss of heat-sensitive nutrients (Onwude, Iranshahi, Martynenko, & Defraeye, 2021). It has been widely used for drying agricultural products such as fruits and vegetables, while retaining the nutritional content and sensory quality of products (Zhang et al., 2023). However, its drying rate drops significantly after the moisture content of the material decreases, and this efficiency bottleneck becomes a key problem restricting its application (Ni, Ding, Zhang, & Song, 2020). This decline originates from three interrelated mechanisms: first, residual moisture shifts to bound water (e.g., hydrogen-bonded water) with strong interaction with the material matrix (Duan et al., 2024); second, material shrinkage causes pore collapse, reducing the effective moisture diffusion coefficient (Deff) by over 50%; third, the thin and stable saturated air layer on the surface weakens the disturbance effect of corona wind, the core driving force of EHD drying (Zamani, Mehrabani-Zeinabad, & Sadeghi, 2019). Ultrasonic pretreatment (US) can specifically address this bottleneck through cavitation and mechanical effects: it forms microchannels in the material to prevent pore collapse, breaks hydrogen bonds to convert bound water into free water, and enhances corona wind's boundary layer disturbance(Duan et al., 2024), thus becoming a core optimization direction for EHD drying. Therefore, the optimization of the pretreatment process becomes the core direction to break through this bottleneck. Food pretreatment methods can be divided into two major categories: thermal pretreatment and non-thermal pretreatment. Thermal pretreatment methods such as hot water blanching, steam blanching, microwave blanching, and ohmic heating blanching can effectively inactivate enzyme activity and improve drying efficiency, but high temperatures may lead to nutrient loss or cause non-enzymatic browning. Non-thermal pretreatment methods such as osmotic dehydration, pulsed electric field, and cold plasma treatment mainly promote water migration by altering cell structure, but their application is limited by problems such as high equipment costs or complex processes (Iranshahi et al., 2023). Among them, US stands out, which uses cavitation and spongy effect to disrupt cell membranes and form microchannels, significantly increasing the drying rate, as demonstrated by Cakmak, Tekeoglu, Bozkir, Ergun, and Baysal (2016) with 32% reduction in drying time for mushrooms. Moreover, the low-temperature treatment better preserves heat-sensitive components such as polysaccharides and antioxidant substances (Tan et al., 2020). In the garlic drying experiment conducted by Guo et al. (2023), the US group retained the maximum amount of allicin. In addition, US can improve the texture of the product and can be combined with EHD technology to create a synergistic effect. Although it requires a liquid medium and overprocessing may affect cell structure, its high efficiency, energy efficiency, and quality optimization make it highly promising in the field of food drying.

Conventional US typically uses ordinary water as the medium, but has obvious limitations when dealing with special ingredients like garlic. In contrast, using deionized water, NaCl solution, and sucrose solution as the ultrasonic medium has significant advantages. The deionized water can effectively avoid impurity interference and ensure the purity of the pretreatment process (Ni et al., 2020). NaCl solution can significantly enhance ultrasonic cavitation through ionic effects and improve cell membrane permeability (Sleiman, Hallez, Pflieger, Nikitenko, & Hihn, 2022). Sucrose solution, on the other hand, promotes water excretion while better retaining nutrients through osmotic pressure regulation (Pei et al., 2019). These improved media can not only be optimized for the characteristics of garlic to enhance drying efficiency, but also better maintain the nutritional quality and sensory characteristics of the product, laying a good foundation for subsequent improvements in the drying process. However, previous studies mainly focused on single ultrasound or osmotic pretreatment, while the synergistic mechanism of NaCl solution and ultrasound on garlic components remains unclear.

Given the efficiency limitations of EHD drying at the low water content stage, and considering that existing studies have confirmed the individual effects of NaCl in enhancing ultrasonic cavitation and sucrose in regulating osmotic pressure (Yue, Lin, Zhang, Lu, & Jiang, 2025), there are few reports on the effects of different media types (deionized water, NaCl solution or sucrose solution) on the efficiency and active ingredients of US combined with EHD drying for the garlic. To explore the synergistic effects of NaCl solution and ultrasound pretreatment on the retention of allicin-derived stable sulfur compounds, total phenols, and protein structure of garlic, and elucidate the underlying chemical mechanisms, this paper selected four pretreatment methods: deionized water as the control group, US combined with deionized water (DW + US), US combined with NaCl solution (NaCl aq. + US), and US combined with sucrose solution (Suc aq. + US). The study aims to explore the effects of four pretreatment methods on the drying characteristics and quality of garlic in electrohydrodynamic systems, providing experimental evidence and theoretical guidance for the extended application of ultrasonic technology combined with EHD. At the same time, it provides a reference for the optimization of the US process.

2. Materials and methods

2.1. Materials and equipment

2.1.1. Material selection

Fresh garlic purchased from Hohhot farmers' Market in Inner Mongolia, transported to the laboratory within 24 h after harvest, stored at 4 °C to preserve the original quality characteristics for subsequent experimental preparations. Samples with plump bulbs and no mechanical damage were selected for the experiment.

2.1.2. Chemicals and laboratory equipment

The chemicals used in this article include analytical grade NaCl (Sinopharm Chemical Reagent Co., LTD.) and sucrose (Shanghai Maclin Biochemical Technology Co., LTD.). The experimental equipment mainly included: high-voltage power supply (YD (JZ)-5 KVA/50 KV, Wuhan Huate Power New Technology Co., Ltd., Wuhan, China), controller (KZX-1.5 KVA, Wuhan Kaikai Automation Control Equipment Co., Ltd., Wuhan, China) and the self-made needle-plate electrode system. The upper plate of the needle-plate electrode system is connected to a high-voltage power supply, using 720 mm × 360 mm stainless steel needle plate electrodes with a needle spacing of 40 mm × 40 mm. The lower plate is grounded using 900 mm × 630 mm stainless steel plates. The distance between the discharge electrode and the grounding electrode is 60 mm, with each needle measuring 30 mm in length and 1 mm in diameter. Auxiliary equipment: temperature and humidity meter (GJWS-A2, Jinlishi, Xingtai Qiaodong Jinlishi Watch & Glasses Store, Tianjin, China), ultrasonic cleaner (KQ-300DE, Shumei, Kunshan Ultrasonic Instrument Co., Ltd., Kunshan, China), electronic balance (BS124S, Shanghai, China), and digital display constant temperature water bath (HH-2, Changzhou Jintan Jingda Instrument Manufacturing Co., Ltd., Jiangsu, China, temperature control accuracy ±1 °C).

2.2. Pretreatment methods and drying methods

2.2.1. Pretreatment methods

The effects of four pretreatment methods on fresh garlic samples were systematically investigated. Samples in the deionized water (DW) group were immersed in the water bath maintained at 30 °C for 30 min. The US group used ultrasonic cleaners to treat in three media (Pei et al., 2019; Praharaj & Misra, 2018). Guided by literature, preliminary experiments evaluating both drying characteristics and component retention led to the selection of 2% sodium chloride and 13.5% sucrose to create distinct acoustic media: a representative electrolyte medium and a medium-viscosity medium, respectively(Asefi, Alirezaloo, Veisi, & Zafar, 2023; Chandra, Kumar, Kumar, & Nema, 2022). Based on preliminary experimental validation combined with relevant literature support, the ultrasonic power was set at 240 W to balance enhanced mass transfer and high retention of bioactive components without inducing thermal degradation, the pretreatment temperature at 30 °C, and the pretreatment duration at 30 min(W. Bai et al., 2025). To ensure the reliability of the experimental data, three parallel repeats were set for each pretreatment protocol. All treated samples were immediately transferred to the subsequent drying experimental process to avoid time-sensitive changes in the pretreatment effect. This experimental design examines the effects of different solution systems and assesses the synergistic effects of ultrasonic-assisted treatment.

2.2.2. Drying methods

After pretreatment, the garlic samples were dried using an EHD system with the operating voltage was set at 18 kV, identified as optimal via pre-experiments. This voltage balances drying efficiency and discharge stability: EHD drying efficiency correlates positively with voltage within a reasonable range, and 18 kV is the maximum voltage enabling stable, uniform corona discharge under the fixed 60 mm electrode distance. Exceeding this voltage triggers arc discharge, damaging garlic cells, causing active component loss, and compromising experimental safety and reproducibility, while 18 kV ensures efficient moisture diffusion and preserves product quality. The fresh garlic slices, after pretreatment, were made into standard sections with a diameter of 14 mm and a thickness of 3 mm. The initial wet base moisture content of the slices was determined to be 72.00 ± 0.85% by a rapid moisture meter (SH10A, Shanghai Precision Instrument Co., Ltd., Shanghai, China). All drying tests were conducted in a precisely controlled constant temperature and humidity environment, with the test conditions strictly controlled at a temperature of 25 ± 1 °C, a relative humidity of 20 ± 1%, and a wind speed of 0 m/s until the wet basis moisture content of the sample was reduced to 10% and the drying process was terminated. Through a systematic analysis of the dried slices, the drying kinetics characteristics, physical quality parameters, infrared spectral characteristics, proteomic characteristics and volatile flavor substances of garlic were investigated.

2.3. Moisture content determination

Calculate the dry basis moisture content and moisture ratio of the garlic sample during the drying process using the following Eqs. (1), (2) (Dorneles et al., 2019):

$$M_i=\frac{m_i-m_g}{m_g}$$ (1)

$$\mathit{MR}=\frac{M_i-M_e}{M_0-M_e}$$ (2)

The meanings of each parameter in Eqs.(1), (2) are as follows: M_i represents the dry basis moisture content of garlic at drying time t_i (unit: g water /g dry matter), which is calculated by the ratio of the total sample mass m_i (unit: g) minus the absolute dry mass m_g (unit: g) to m_g. MR is the moisture ratio, reflecting the relative relationship between the current moisture content state and the initial state, M_e represents the equilibrium dry basis moisture content of garlic, that is, the final stable moisture content achieved through drying, M₀ represents the dry basis moisture content of garlic at the initial time t₀, serving as a reference value for the drying process. In practical applications, when the equilibrium dry basis moisture content M_e value is small, it can often be ignored, at which point the moisture ratio can be simplified as shown in Eq. (3) (Y. Yue et al., 2023).

$$\mathit{MR}=\frac{M_i}{M_0}$$ (3)

2.4. Determination of drying rate

The drying rate of garlic slices is defined as Eq. (4) (Akbarbaglu et al., 2025):

$$\mathit{DR}=\frac{M_t-M_{t+∆t}}{∆t}$$ (4)

where DR is the drying rate, M_t is the dry basis moisture content of the garlic slices at time t, and M_t+Δt is the dry basis moisture content of the garlic slices at time t + Δt.

2.5. Determination of the effective moisture diffusion coefficient for drying

D_eff is calculated based on Fick's second diffusion law equation (C. Li et al., 2023), as shown in Eq.(5):

$$\mathit{MR}=\frac{8}{{\mathrm{\pi }}^2}\sum _{\mathrm{n}=0}^{\infty }\frac{1}{{\left(2\mathrm{n}+1\right)}^2}\exp \left(-\frac{{\left(2\mathrm{n}+1\right)}^2{\mathrm{\pi }}^2{D}_{\mathit{eff}}t}{L^2}\right)$$ (5)

Here, D_eff represents the effective moisture diffusion coefficient (unit: m²/s), L represents sample thickness (unit: m), t stands for time (unit: s), n is a positive integer. Eq. (5) can be simplified to Eq. (6), taking only the first term of the series, with no significant impact on the accuracy of the assumption:

$$\mathit{MR}=\frac{8}{{\pi }^2}\exp \left(-\frac{{\pi }^2{D}_{\mathit{eff}}t}{L^2}\right)$$ (6)

The effective moisture diffusion coefficient (D_eff) can also be calculated using the slope of Eq. (6). That is, plotting ln(MR) against time yields the slope k:

$$k=\frac{{\mathrm{\pi }}^2{D}_{\mathit{eff}}}{L^2}$$ (7)

The value of D_eff can be calculated from Eq. (7).

2.6. Determination of rehydration rate

Took dry samples from each group, placed them in 10 mL beakers, transferred them to a 30 °C water bath for soaking, and weighed them every 60 min. Each time before weighing, samples were removed, surface moisture was drained using filter paper, and then weighed. This continued until the sample weight stabilized (indicating no further water absorption). Rehydration rate (RR) as (J.-W. Bai et al., 2023):

$$\mathit{RR}=\frac{W_g-W_0}{W_0}$$ (8)

In the Eq.(8), W_g and W₀ are the masses of the sample after rehydration and before rehydration, respectively.

2.7. Fourier transform infrared spectroscopy (FTIR)

Based on the potassium bromide particle method and amide I band analysis method adopted by Kezban, Gunes, and Naşit (2020), the infrared spectroscopy study of meat products was improved. Weigh 2 mg of each treated sample powder, mix with 300 mg potassium bromide (1:150 mass ratio) in an agate mortar (Shanghai Sile Instrument Co., Ltd., B13-3Pro, China), and grind to <2 μm. Transfer the mixture to a pellet die (model SLP-10, matching the agate mortar, Shanghai Sile Instrument Co., Ltd., China), press into a transparent pellet at 8 tons pressure, and analyze immediately via Bruker Vertex 70 FTIR spectrometer. Acquire spectra (400–4000 cm⁻¹, 32 scans, 4 cm⁻¹ resolution) with H₂O/CO₂ atmospheric compensation, then output DPT file after baseline correction. Further analysis of the amide I band spectrum in the 1600–1700 cm⁻¹ range using PeakFit software (San Rafael, California, USA) yielded fine spectral information through baseline adjustment, Gaussian deconvolution, second derivative calculation, and curve fitting (Ao et al., 2024).

2.8. Proteomic analysis

Garlic tissue was subjected to protease digestion. Specifically, the target bands were placed in centrifuge tubes, washed twice with ultrapure water (5 min each time via vortex oscillation using Scilogex MX-S, USA), and decolorized by sonication for 15 min with a 1:1 mixture of 50 mM NH₄HCO₃ and acetonitrile, repeating until the solution and gel pieces were colorless. After a further wash with the same 50 mM NH₄HCO₃-acetonitrile mixture, the gel pieces were dehydrated with 100% acetonitrile by oscillation until white and vacuum-dried for 5 min using Thermo Savant, USA. Proteins were then reduced with 150 μL of 50 mM dithiothreitol (56 °C, 1 h) and alkylated with 150 μL of 100 mM iodoacetamide in the dark for 45 min. Following sequential washes with 25 mM NH₄HCO₃, 50 mM NH₄HCO₃-acetonitrile (1:1), and 100% acetonitrile, the gel pieces were re-dehydrated and vacuum-dried. Trypsin digestion was performed overnight at 37 °C with enzyme solution prepared using Thermo Fisher Scientific Trypsin Gold (Mass Spectrometry Grade, Cat. No. V5280) (diluted appropriately with 25 mM NH₄HCO₃ to submerge the swollen gel pieces), and peptides were extracted by sonication with 100 μL of 30% acetonitrile/0.1% trifluoroacetic acid and 100 μL of 60% acetonitrile/0.1% trifluoroacetic acid sequentially, followed by lyophilization. The extracted peptides were desalted using ZipTip C18: after reconstitution in 0.1% trifluoroacetic acid, the ZipTip was preconditioned with 50 μL of 60% acetonitrile/0.1% trifluoroacetic acid (10 washes) and equilibrated with 10 μL of 0.1% trifluoroacetic acid (10 washes); the sample was aspirated and dispensed 20 times through the tip, which was then washed 5 times with 10 μL of 0.1% trifluoroacetic acid before peptides were eluted with 10 μL of 60% acetonitrile/0.1% trifluoroacetic acid into a new centrifuge tube and vacuum-dried. For liquid chromatography-mass spectrometry analysis, peptides were reconstituted in 20 μL of 0.1% formic acid, vortexed thoroughly, centrifuged at 17,000 ×g for 20 min at 4 °C using Eppendorf 5418 R, Germany, and 3 μL of the supernatant was injected onto a Thermo PepMap RSLC C18 column (75 μm × 150 mm, 2 μm, 100 Å). The chromatographic separation was performed at a flow rate of 250 nL/min using a gradient of mobile phase A (0.1% formic acid in water) and mobile phase B (0.1% formic acid in acetonitrile): initial conditions (0 min) were 95% A/5% B, adjusted to 90% A/10% B at 8 min, 85% A/15% B at 33 min, 72% A/28% B at 43 min, 60% A/40% B at 50 min, 5% A/95% B at 60 min (held until 65 min), and reverted to 95% A/5% B at 66 min (held until 70 min). Mass spectrometry parameters were set to a resolution of 70,000 for primary mass spectrometry (automatic gain control target: 3e⁶; maximum ion time: 60 ms; scan range: 300–1400 m/z) and 17,500 for secondary mass spectrometry (automatic gain control target: 5e⁴; maximum ion time: 80 ms; TopN: 20; ormalized collision energy: 27). Data were searched against the UniProtKB database (uniprotkb_taxonomy_id_4678_2025_02_25.fasta) using PEAKS software with trypsin specificity, 10 ppm precursor mass tolerance, and 0.02 Da fragment mass tolerance, allowing for dynamic modifications (oxidation of methionine, N-terminal acetylation) and static modification (carbamidomethylation of cysteine).

2.9. Determination of major stable sulfur compounds (DADS and DATS)

Given the rapid decomposition of unstable allicin into stable organosulfur compounds during processing, this study followed the approach of Locatelli, Altamirano, Gonzalez, and Camargo (2015) and quantified the major end products, DADS and DATS, as reliable indicators of bioactive sulfur retention. All electrohydrodynamic-dried samples were first freeze-dried in a vacuum vacuum freeze-dryer (model LGJ-12H, Sihuan Freeze Drying (Beijing) Co., Ltd., China) at −50 °C and below 10 Pa for 48 h. The resulting material was ground, passed through an 80-mesh sieve, and stored at −20 °C until analysis. Freeze-dried powder was then extracted with dichloromethane using extracted with dichloromethane using Dionex ASE 350 accelerated solvent extraction system, Thermo Fisher Scientific, USA (60 °C, 1500 psi). The extract was concentrated, reconstituted in acetonitrile, and analyzed by UHPLC-UV on a Waters Acquity BEH C₁₈ column under isocratic conditions (acetonitrile/water, 60:40, v/v) at a flow rate of 0.4 mL/min with detection at 254 nm; no derivatization was performed. DADS and DATS were quantified by external calibration using commercial standards (Sigma-Aldrich, LKT Laboratories). Contents were calculated from their respective calibration curves and are reported individually and as a sum in grams per kilogram of dry weight (g/kg).

2.10. Determination of total phenol content

Based on the Folin-Ciocalteu method (Y. Yue et al., 2023), modifications were made to accommodate the garlic substrate. EHD dried garlic slices were ground into powder using a planetary ball mill (QM-3SP2, Nanjing University Instrument Factory, China) prior to extraction. Garlic powder (0.20 g) was heated in distilled water (100 °C, 30 min), filtered through a 0.45 μm organic phase filter membrane (Merck, Germany) using a Corning filtration funnel, USA, and diluted with 60% ethanol. The test solution (100 μL) was mixed with diluted FC reagent (0.75 mL) and 10% sodium carbonate (0.75 mL), incubated in the dark (90 min), and absorbance measured at 760 nm. Gallic acid standards (0.01–0.05 mg/mL) were used for calibration.

2.11. Analysis of volatile compounds by SPME-GC–MS

Based on established HS-SPME-GC–MS methodology (Molina-Calle, Priego-Capote, & Castro, 2017), volatile compounds were analyzed as follows: 1.5 g of garlic powder and 20 μL of 1% 2,4,6-trimethylpyridine (internal standard) pipetted with a BRAND Transferpette® pro single-channel pipette (10–100 μL volume range, Germany) were sealed in 20 mL vials and equilibrated for 10 min. Extraction proceeded for 40 min at 60 °C using a 50/30 μm DVB/CAR/PDMS fiber with 500 rpm magnetic stirring via IKA RCT Basic, Germany. GC–MS analysis employed an Agilent 7890 A-5975c system with an HP-5MS column (30 m × 0.25 mm × 0.25 μm) and splitless injection at 250 °C. The oven program was: 40 °C (5 min), ramp to 250 °C at 10 °C/min, hold 5 min. Helium carrier gas flowed at 1.0 mL/min. Mass detection used EI ionization (70 eV) scanning m/z 35–400 (solvent delay 2 min). Compounds were identified by retention indices (C6-C26 alkanes) and NIST/Wiley library matching, and quantified by internal standard calibration with peak area normalization. All analyses were performed in triplicate.

2.12. Statistical analysis

All experiments were performed independently in triplicate. The experimental results were expressed as mean ± standard deviation (SD). Statistical analyses were conducted using SPSS Statistics 27.0. Differences among the four pretreatment groups were evaluated by one-way analysis of variance (ANOVA), followed by Tukey's post-hoc test for multiple comparisons. A p-value <0.05 was considered statistically significant.

3. Results and discussion

3.1. Analysis of drying characteristics and rehydration capacity

Fig. 1(a) shows the variation of garlic moisture content with drying time under different pretreatments. In the early stage of deionized water pretreatment (DW) alone, the water content decreased relatively gently, while in ultrasound-assisted pretreatments (DW + US, NaCl aq. + US, Suc aq. + US), the water content decreased more significantly because US pretreatment formed ultrasonic microchannels inside the cells, thereby significantly accelerating the drying rate. Duan et al. (2024) also mentioned in their experiment on drying carrots that ultrasound can regulate the microstructure of materials through its cavitation effect. Among them, ultrasonic-assisted NaCl solution (NaCl aq. + US) and ultrasonic-assisted sucrose solution (Suc aq. + US) in the initial stage of drying, due to the synergistic effect of solution permeation and ultrasonic cavitation, facilitated a rapid decrease in moisture content. However, the later drying rate of the sucrose solution group dropped to the level of the DW group, which was associated with the formation of a viscous film layer of high concentration sucrose that hindered water diffusion. Rahaman et al. (2019) explicitly stated that sucrose molecules are more prone to adhere to the plum surface, forming a high-concentration layer that impedes water diffusion.

Fig. 1.

Drying properties and rehydration capacity. (a): Curve of water content ratio varying with drying time; (b): Graph of the relationship between average drying time and average drying rate for different pretreatment methods; (c): Effects of different pretreatment methods on the effective moisture diffusion coefficient of garlic; (d): Rehydration rate. Different letters indicate significant differences between sample means (p < 0.05).

Fig. 1(b) compares the average drying rate and average drying time of different pretreatments. In terms of average drying rate, the NaCl aq. + US group stood out, with a much higher peak than the other groups, while the Suc aq. + US group had the lowest rate. Regarding average drying time, the Suc aq. + US group took the longest time; ultrasound-assisted treatment significantly shortened the drying time, and the NaCl aq. + US group had the shortest drying time. Meanwhile, the rate of the Suc aq. + US group declined later, so its drying time was longer than that of the DW group. The effective moisture diffusion coefficient (D_eff) is a useful indicator for measuring dehydration efficiency (Lasik-Kurdys, Majcher, & Nowak, 2018). As can be seen from Fig. 1(c), D_eff is the largest in the NaCl aq. + US group, indicating that the pretreatment minimizes water diffusion resistance. The Suc aq. + US group had the smallest D_eff and significant resistance to water diffusion. This can be attributed to the ability of sucrose molecules to cross-link via hydrogen bonds with headgroups on the lipid membrane surface, forming a dense adsorbed or glassy sugar layer at the interface. This layer structure effectively blocks the transport of small molecules and establishes a stable diffusion barrier, which aligns with the findings of Bhatia, Christ, Steinkuehler, Dimova, and Lipowsky (2020). Although the DW + US group and the Suc aq. + US group increased D_eff due to structural destruction by ultrasound, the increase in diffusion coefficient in the sucrose solution group was weaker than that in the NaCl aq. + US group.

Rehydration defines the most important quality characteristics of dried food, highlighting the physicochemical changes brought about by pretreatment (Assad et al., 2025), sample composition and process conditions during the drying process. As shown in Fig. 1(d), the NaCl aq. + US group had the highest rehydration rate, suggesting that the pretreatment did not overly disrupt the cellular structure during accelerated drying, possibly forming a controllable porous structure. Water was easily permeated and filled during rehydration, maintaining good quality restorability. The DW group had the lowest rehydration rate. Long-term drying led to cell collapse, dense structure, difficulty in water penetration, and poor quality retention. The DW + US group had a better rehydration rate than the DW group because the structure was less damaged by ultrasound than by natural drying. However, in the Suc aq. + US group, affected by the sugar film effect, which is the film layer formed by sucrose during drying hindered rehydration (Agoda-Tandjawa, Mazoyer, Wallecan, & Langendorff, 2019), the rehydration rate was lower than that in the NaCl aq. + US group, indicating that different pretreatments had different effects on the structural integrity of the material after drying. The NaCl aq. + US pretreatment not only enhanced drying efficiency but also preserved rehydration capability.

3.2. Retention of active components: Allicin-derived DADS/DATS and total phenols

Allicin is garlic's core bioactive component, synthesized via the alliinase-catalyzed hydrolysis of alliin (Locatelli et al., 2015). Due to its instability during processing, it degrades into stable sulfides (DADS and DATS), which were used as reliable indicators of bioactive sulfur-containing compound retention in this study.

As shown in Fig. 2(a), the total content of DADS and DATS in the NaCl aq. + US group reached 3.69 ± 0.22 g/kg, significantly higher than the DW group (2.446 ± 0.15 g/kg), DW + US group (2.415 ± 0.095 g/kg), and Suc aq. + US group (2.93 ± 0.164 g/kg). Specifically, DADS and DATS in the NaCl aq. + US group were 60% and 45.6% higher than the DW group, respectively, attributed to the synergistic effect of ultrasound and NaCl. Ultrasonic cavitation disrupted garlic cell structures to promote alliin release (Yang, Gan, Ge, Zhang, & Corke, 2019), while NaCl's ionic effect plays a critical regulatory role in alliinase activity (Zhong et al., 2022). NaCl modulates ionic strength to screen electrostatic repulsion between alliinase surface residues, stabilizing a compact conformation that exposes the active site; Na⁺/Cl⁻ further interact with charged residues in the active site to optimize enzyme-substrate binding and catalytic efficiency. Additionally, ultrasound enhances enzymatic activity (Xu et al., 2019), and the controlled porous structure formed by the synergistic pretreatment reduces allicin oxidation, ultimately boosting DADS and DATS retention.

Fig. 2.

Garlic active substances. (a): Content of major allicin-derived stable derivative (DADS and DATS). (b): Total phenol content. Different letters indicate significant differences between sample means (p < 0.05).

In contrast, the DW and DW + US groups showed similar DADS and DATS contents, as ultrasound alone only disrupts cell structures without regulating alliinase conformation or catalytic kinetics. The Suc aq. + US group exhibited lower DADS and DATS retention due to the viscous film formed by sucrose, which hinders water diffusion and enzyme-substrate interaction.

Total phenols are one of the key indicators for measuring the antioxidant capacity of plants (Motta, Guaita, Cassino, & Bosso, 2019). For garlic, the level of phenols is closely related to the antioxidant effect of garlic. As shown in Fig. 2(b), the NaCl aq. + US group had the highest total phenol content, reaching 0.9 ± 0.03 g/kg, which was 32.4% higher than the DW group, 25.0% higher than the DW + US group (0.72 ± 0.02 g/kg), and 38.5% higher than the Suc aq. + US group (0.65 ± 0.02 g/kg). Goyeneche, Roura, and Scala (2014) reported similar results on the inhibition of polyphenol oxidase activity by NaCl. The total phenol content in the DW group was close to that in the DW + US group, differing by only 0.04 g/kg, suggesting that ultrasound alone did not significantly increase the total phenol content in garlic. After pretreatment with NaCl aq. + US, total phenol content increased due to the synergistic effect of NaCl solution and US. On the one hand, mechanical vibrations produced by the US disrupted the structure of garlic cells, prompting the release of total phenols within the cells. On the other hand, the NaCl solution regulates the ionic environment within the cells and may inhibit the activity of polyphenol oxidase, thereby reducing the oxidative decomposition of total phenols. When pretreated with Suc aq. + US, the total phenol content was lower than that in the NaCl aq. + US group, possibly due to the high osmotic pressure of the sucrose solution suppressing cellular metabolism and affecting the synthesis and accumulation of total phenols.

3.3. Proteomic characteristics and structural stability analysis

3.3.1. Protein secondary structure stability (FTIR analysis)

Fig. 3(a) shows the Fourier transform infrared spectra of garlic under different pretreatment conditions, used to investigate the effect (Yulizar, Ariyanta, & Abdurrachman, 2017). The spectra show that the wide-intensity absorption peaks in the 3000–4000 cm⁻¹ range correspond to the stretching vibrations of N—H, O—H and C—H. The absorption peaks at 2000–3000 cm⁻¹ are associated with the C C, C N stretching vibrations of groups such as methyl and methylene. 1200–1300 cm⁻¹ is the characteristic band of the protein, including C—C, C—O and C—N stretching vibrations of amide I, II and III groups. 500–900 cm⁻¹ mainly shows the characteristic absorption of polysaccharides. The spectral peaks of the different pretreatment groups were consistent but varied in intensity, indicating that pretreatment did not change the material composition of garlic, but significantly affected the component content. Specifically, the wide peaks near 3274 cm⁻¹ were reduced in width and height after pretreatment with DW + US, DW and Suc aq. + US, suggesting weakened absorption capacity of the corresponding functional groups. Transmittance analysis showed that the DW group was significantly lower than the DW + US group and the order of each group was DW + US > Suc aq. + US > DW > NaCl aq. + US. Absorbance at 1726 cm⁻¹ and 1625 cm⁻¹ from high to low was Suc aq. + US > NaCl aq. + US > DW + US > DW and absorption intensity near 1200 cm⁻¹ decreased significantly. Absorption peak intensities at 1000 cm⁻¹ were ranked as DW > Suc aq. + US > DW + US > NaCl aq. + US.

Fig. 3.

Infrared spectra and protein secondary structures of garlic slices under different pretreatment methods. (a): Fourier transform infrared spectra of garlic showing absorption peaks of functional groups (e.g., hydroxyl, amide, polysaccharide-related) under different pretreatments; (b): Proportion of protein secondary structure composition analyzed by Gaussian deconvolution of the amide I band, with β-turn content significantly increased in NaCl aq. + US and Suc aq. + US groups, indicating enhanced protein structural stability.

The amide I band spectra of 1600–1700 cm⁻¹ in Fig. 3(a) were analyzed using baseline correction, Gaussian deconvolution, second derivative and curve fitting to quantitatively analyze the proportion of protein secondary structure (Gabriel et al., 2022). The α-helix is located at 1651–1660 cm⁻¹, the β-sheet is located at 1657–1669 cm⁻¹ and 1680–1689 cm⁻¹ and the β-turn is located at 1664–1681 cm⁻¹. Random coils are located between 1649 and 1667 cm⁻¹. Fig. 3(b) visually shows the difference in content between random coil and β-sheet. Statistical analysis indicates that β-turn content is significantly increased in the Suc aq. + US and NaCl aq. + US groups, which is speculated to be related to the synergy of ultrasound and solution, that is, the mechanical vibration generated by ultrasonic cavitation enhances the stability (Cui et al., 2020). In addition, the four pretreatments had no significant effect on the content of random curls and α-helices.

3.3.2. Protein content and post-translational modifications

The total ion current diagram can visually show how the total intensity of all ions in the sample varies over time under different pretreatment methods, thereby evaluating the overall response and separation effect of compounds in the sample. It can be seen from Fig. 4(a) that there are significant differences in the total ion current intensity corresponding to different pretreatment methods. The total ion current intensity corresponding to the DW group was significantly higher than that of the other pretreatment methods, indicating that this pretreatment method had a better effect in extracting or retaining compounds in the sample.

Fig. 4.

Analysis of the effects of different pretreatment methods on garlic protein and alliinase. (a): Total ion current fingerprints of garlic proteins, reflecting differences in compound extraction/retention efficiency among pretreatment groups; (b): Proportion of post-translational modifications of garlic proteins, with N-terminal acetylation as the dominant modification and the highest acetylation level in NaCl aq. + US group; (c): Protein abundance (LFQ intensity) showing the highest value in DW group and moderate level in NaCl aq. + US group; (d): Alliinase abundance with the highest level in NaCl aq. + US group, which is key to allicin biosynthesis. Different letters indicate significant differences between sample means (p < 0.05).

The post-translational modification of proteins analysis is of great significance (Bobalova, Strouhalova, & Bobal, 2023). In this study, post-translational modification analysis was conducted on garlic samples treated with four different methods and it was found that the pretreatment methods significantly affected the modification characteristics of garlic proteins. Fig. 4(b) shows that in all pretreatment groups, N-terminal acetylation modification was dominant, ranging from 23.25% to 35.36%, with the highest acetylation level in the NaCl aq. + US pretreatment group at 35.36%. This high acetylation level may be a cellular adaptive response to the combined stress of ultrasonic cavitation (Qayum et al., 2021) and NaCl-induced ionic osmotic stress(Yu, Wang, Li, & Wang, 2022): N-terminal acetylation is well-documented to enhance protein structural stability by inhibiting proteolytic degradation, and this modification might specifically preserve the activity of key metabolic enzymes like alliinase —consistent with the highest alliinase abundance and the abundance of its stable end products (DADS and DATS) observed in the NaCl aq. + US group (Figs. 4d and 2a). Notably, NaCl-mediated electrostatic screening may further synergize with acetylation to stabilize protein conformation, while ultrasonic cavitation-induced mild structural disruption is compensated by this protective modification, avoiding excessive protein denaturation. The composite modification patterns of different treatment groups showed significant differences: the acetylation + carboxymethylation modification was the highest in the DW + US group at 20.98%, the acetylation + oxidation modification was the most significant in the DW group at 18.96% and the acetylation + carboxymethylation + oxidation triple modification was higher in the Suc aq. + US group at 20.08%. Particularly notable is the distribution of oxidative modifications, with the highest level of individual oxidative modifications in the Suc aq. + US group at 16.49% (which may impair the stability of functional proteins and explain the lower alliinase abundance in this group) and the lowest in the DW + US group at 10.22%, suggesting that US may have an antioxidant effect and that sucrose solution may induce oxidative stress through osmotic stress (Peran, Sabri, & Mittag, 2020). In addition, the proportion of individual carboxymethylation modification in the NaCl aq. + US group at 18.07% was significantly higher than that in the other pretreatment groups, suggesting that NaCl affected protein conformation and exposed more cysteine residues, potentially further optimizing enzyme stability by regulating protein folding and intermolecular interactions.

From the results of protein abundance in Fig. 4(c), the protein abundance in the DW pretreatment group was significantly higher than that in the other pretreatment groups and the LFQ intensity was close to 6, indicating that DW pretreatment alone had a unique advantage (Liu et al., 2017). Protein abundance was significantly reduced in the DW + US pretreatment group, with an LFQ intensity of only about 2.5, indicating that US had an impact on protein extraction or stability. The protein abundance in the NaCl aq. + US pretreatment group was at a moderate level, with an LFQ intensity of about 5, between the DW and DW + US pretreatment groups, indicating that the effect of the synergistic effect of NaCl solution and US on protein abundance was rather complex. Protein abundance was also relatively low in the Suc aq. + US pretreatment group, with an LFQ intensity of about 3.

Alliinase plays a significant role in the enzymatic reaction in which Allium plants release sulfur-containing compounds (Z. Wang et al., 2024).From Fig. 4(d), alliinase abundance, the NaCl aq. + US pretreatment group had the highest alliinase abundance, with an LFQ intensity of approximately 0.14, indicating that the pretreatment could significantly promote the retention or extraction of alliinase. The alliinase abundance in the DW pretreatment group was at a moderate level, with an LFQ intensity of approximately 0.11, indicating that the effect of DW alone on alliinase was relatively stable. The alliinase abundance was lower in the DW + US pretreatment group, with an LFQ intensity of about 0.08. This may be an adverse effect of US on the structure or activity of alliinase (Su & Cavaco-Paulo, 2021). The Suc aq. + US pretreatment group had the lowest alliinase abundance, with an LFQ intensity of only about 0.06. This phenomenon stems from two synergistic effects: first, the same ultrasound-induced structural damage to alliinase as observed in the DW + US group (Su et al., 2021); second, the high osmotic pressure of the 13.5% sucrose solution that impairs enzyme stability. As reported previously (Giannini et al., 2021; Sanchez-Fernandez et al., 2022), high sugar concentrations disrupt the protein hydration layer—osmotic pressure depletes bound water around alliinase, breaks intramolecular hydrogen bonds, and exposes hydrophobic groups, thereby inducing protein denaturation and aggregation. Specifically, sucrose-induced osmolarity promotes the unfolding of alliinase's secondary structure and formation of inactive oligomers, further reducing its detectable abundance. Thus, the combined effect of ultrasonic damage and sucrose-mediated denaturation collectively leads to the lowest alliinase abundance in the Suc aq. + US group.

3.3.3. Differential protein-protein interaction network

Protein-protein interactions play a crucial role in many fundamental biological processes, such as cellular signaling and immune responses (J. Wang & Miao, 2022). This paper constructs interaction networks for differential proteins in the DW and DW + US pretreatment groups. As can be seen from Fig. 5(a), nodes such as petD, psaA and rbcL are at the core and interact with multiple other differential proteins. PetD is involved in processes related to the electron transport chain in photosynthesis and psaA is an important component of photosystem I and plays a key role in light energy absorption and conversion (Króliczewski, Bartoszewski, & Króliczewska, 2017). Changes in its interaction reflect the regulation of photosynthetically associated proteins by US (Azarin et al., 2020). RbcL, as a large subunit of Rubisco enzyme, plays a significant role in carbon assimilation and the alteration of its interaction is associated with the impact of US on carbon metabolism. In addition, proteins such as accD, rpoA and rpoB also play a significant role in the network. accD is involved in fatty acid synthesis and rpoA and rpoB are associated with RNA polymerase function. The changes in their interactions suggest that DW + US pretreatment has an impact on fatty acid metabolism and gene transcriptional regulation in garlic.

Fig. 5.

Differential protein-protein interaction networks. Interaction networks of proteins with significantly altered abundance between (a): DW and DW + US groups, (b): DW and NaCl aq. + US groups, and (c): DW and Suc aq. + US groups. Key nodes (e.g., PsaA, RbcL, PetD) are highlighted, reflecting the impact of pretreatments on core biological processes such as photosynthesis, carbon metabolism, and gene transcription.

For the differential protein interaction network of the DW and NaCl aq. + US pretreatment groups, it can be observed in Fig. 5(b) that nodes such as ndhF, rbcL and psaA have more connections and are in key positions in the network. NdhF is involved in photosynthetic electron transfer and respiration-related electron transfer processes and its interaction with other differential proteins may reflect the regulation of energy metabolism-related proteins in garlic by NaCl solution combined with US. The key role of rbcL in carbon assimilation alters its interaction related to carbon metabolism (Iqbal, Lisitsa, & Kapralov, 2023). The changes in the psaA interaction network reflect the effect of the pretreatment on the photosynthetic system. Meanwhile, proteins such as matK, rpoB, and rps12 are also prominent in the network. MatK is associated with chloroplast gene transcription, rpoB is involved in RNA polymerase function and rps12 is related to ribosome assembly (Luan et al., 2022). The changes in their interactions suggest that the NaCl aq. + US pretreatment has an impact on garlic gene transcription, translation and ribosome function. In addition, proteins related to secondary metabolism, such as PDS, GAPC2 and MYB29 were also present in the network. PDS is involved in carotenoid synthesis, GAPC2 is related to glucose metabolism and MYB29 is a transcription factor. The changes in their interactions suggest that this pretreatment plays an important role in secondary metabolism and transcriptional regulation of garlic.

In Fig. 5(c), in the differential protein interaction network between the DW and Suc aq. + US pretreatment groups, nodes such as psaA, rbcL and ndhF are frequently connected and located at the core. The key roles of psaA and rbcL in the photosynthetic system and carbon assimilation made the changes in their interactions reflect the regulation of photosynthetic and carbon metabolism-related proteins in garlic by sucrose solution combined with US. NdhF is involved in the electron transport process and changes in its interaction network are associated with the effect of the pretreatment on energy metabolism. In addition, proteins such as rpoA, rpoB and matK also play a significant role in the network. RpoA and rpoB are associated with RNA polymerase function and matK is related to chloroplast gene transcription. The changes in their interactions suggest that the Suc aq. + US pretreatment has an impact on processes such as garlic gene transcription and expression regulation. There are also proteins in the network that are associated with signal transduction and the ubiquitin-proteasome pathway, such as PHYA, ADO1, SKP1A, etc. PHYA is involved in light signal transduction, ADO1 is associated with adenosine metabolism and SKP1A is part of the SCF ubiquitin ligase complex (Thompson, Rutherford, Lepage, & McManus, 2022). The changes in their interactions suggest that the processing plays a significant role in garlic signal transduction and protein degradation.

3.4. Volatile component analysis

Fingerprint profiling based on retention time and peak area parameters of volatile compounds can visually show the effects of different pretreatment methods on the extraction of volatile components from garlic. As shown in Fig. 6(a), there are significant differences in the fingerprint profiles of the four pretreatment groups: DW, DW + US, NaCl aq. + US and Suc aq. + US. The DW group had fewer peaks and most of them were less intense, possibly because the simple water immersion pretreatment did not destroy the garlic cell structure, resulting in limited release of volatile components and oxidative degradation of some oxidation-prone volatile components (e.g., sulfur-containing compounds) during the pretreatment process. The number of peaks in the DW + US group was higher than that in the DW group and the intensity of some peaks was significantly enhanced, which is attributed to the cavitation effect of ultrasound disrupting the garlic cell membrane to form microchannels, thereby promoting the release of intracellular volatile components—consistent with the mechanism by which ultrasound facilitates substance release through cell structure disruption. The NaCl aq. + US group had the largest number of peaks and generally high peak intensity; for instance, three key sulfur-containing compounds (diallyl disulfide, methyl 2-propenyl disulfide, methyl 2-propenyl trisulfide) showed peak areas of 76,365,120, 15,121,084, and 15,086,810, respectively, which were significantly higher than those of the other groups (Appendix Table S1). It is speculated that the addition of NaCl solution changed the ionic environment of the extraction system, inhibited the activity of some enzymes that led to the degradation of volatile components, and reduced the loss of volatile components. At the same time, the US further enhanced the penetration of the solvent into the garlic cells, promoting the release of volatile components. The combined effect of the two enhanced the extraction effect. The number and intensity of peaks in the Suc aq. + US group were between those in the DW group and the NaCl aq. + US group; for example, its diallyl disulfide peak area was 47,345,108, 37.9% lower than that of the NaCl aq. + US group. This is because the high osmotic pressure of the sucrose solution inhibited garlic cell metabolism to a certain extent, affecting the synthesis and accumulation of volatile components. Although ultrasound also promoted component release, the overall effect was partially offset by the inhibitory effect of sucrose, resulting in a less optimal outcome than the NaCl aq. + US group.

Fig. 6.

Effects of pretreatment on volatile compounds in garlic. (a): Total ion chromatograms (fingerprints) showing distinct volatile profiles across pretreatment groups. (b): Relative content of volatile compound classes; note the significantly higher sulfur-containing compound content in the NaCl aq. + US group. (c): Number of volatile compounds in each chemical class; NaCl aq. + US and Suc aq. + US groups contained the most sulfur-containing compound types. (d): Number of flavor compounds in each sensory category; the NaCl aq. + US group had the highest count of typical garlic sulfur aroma compounds.

Different pretreatment methods have specific effects on the content of volatile compounds in garlic, with the content of sulfur-containing compounds (the core substance of garlic characteristic flavor) being the most prominent (P. Liu et al., 2020). The content of sulfur-containing compounds in the NaCl aq. + US pretreatment group was significantly higher than that in the other groups: the relative content of diallyl disulfide was 96.37%, methyl 2-propenyl disulfide was 95.45%, and methyl 2-propenyl trisulfide was 96.33%, all of which were the highest among the four groups. This confirms that the synergistic effect of NaCl solution and ultrasound greatly promoted the generation or retention of sulfur-containing compounds. The sulfur-containing compounds in the DW pretreatment group were similar to those in the DW + US group and were at a lower level: the peak area of diallyl disulfide in the DW group was 58,012,738, and that in the DW + US group was 37,904,130. The DW group may have lacked conditions to promote the release or generation of sulfur-containing compounds, while the DW + US group, although ultrasound promoted release, may not have inhibited their degradation, so the content did not increase significantly. The content of sulfur-containing compounds in the Suc aq. + US group was between that in the NaCl aq. + US group and the DW group; for example, the peak area of methyl 2-propenyl trisulfide was 3,037,897. This is speculated to be due to the sucrose solution having a certain effect on the generation or retention of sulfur-containing compounds, but not as significant as that of the NaCl solution. In terms of other compounds, the content of nitrogen-containing compounds (e.g., 1,2-hydrazinedicarboxamide) in the DW group was higher than that in other pretreatment groups (relative content: 31.55%) (Appendix Table S1). This indicates that DW pretreatment alone was more conducive to the retention of nitrogen-containing compounds. The DW + US group had the lowest content of oxygen-containing acid compounds, suggesting that ultrasound might have had an adverse effect on the stability of oxygen-containing acid compounds. The Suc aq. + US group had a relatively high content of alcohols (e.g., 2-propen-1-ol, relative content: 97.66%), indicating that the sucrose solution had a certain promoting effect.

Different pretreatment methods also had a significant effect on the types of volatile compounds in garlic. The number of sulfur-containing compounds was the highest in the NaCl aq. + US and Suc aq. + US pretreatment groups (19 types each), while it was the lowest in the DW + US group (only 10 types) (Appendix Table S1). For example, (E)-1-allyl-2-(prop-1-en-1-yl) disulfide was not detected in the DW group but was present in the NaCl aq. + US group with a relative content of 82.44%. This suggests that NaCl solution and sucrose solution helped increase the variety of sulfur-containing compounds, while ultrasound alone may have reduced the variety—possibly because ultrasound alone caused structural damage to some sulfur-containing compounds. For nitrogen-containing compounds, there were three types in the DW group, far more than one in the other pretreatment groups, suggesting that DW alone was more conducive to preserving different types of nitrogen-containing compounds, while other pretreatments might have led to the loss or transformation of some nitrogen-containing compounds. There were no ester compounds in the DW + US, NaCl aq. + US and Suc aq. + US pretreatment groups, and only one type in the DW group, indicating that ultrasound and the addition of NaCl solution and sucrose solution were not conducive to the retention of ester compound types, and these pretreatments might have changed the environmental conditions required for ester compound synthesis.

Under different pretreatment methods, there were significant differences in the number of different flavor compounds. The typical garlic sulfur aroma was the most abundant in the NaCl aq. + US pretreatment group (15 types), which echoed the higher content of sulfur compounds in this group. Specifically, key sulfur-containing compounds such as diallyl disulfide, methyl 2-propenyl disulfide, and methyl 2-propenyl trisulfide—core contributors to garlic's characteristic sulfur aroma—were highly abundant in the NaCl aq. + US group. This indicates that this pretreatment enhanced the typical sulfur aroma of garlic and further demonstrated its advantages in preserving and promoting characteristic flavor substances. The number of pungent/irritating sulfur compounds was 0 in the Suc aq. + US group, in contrast to the other groups. It is speculated that the presence of sucrose solution may have inhibited the formation of such flavor compounds, which is of great significance for the development of low-pungent garlic products. In terms of fat/wax aroma, the DW + US group had the highest number (29 types), which may be related to the ultrasound-promoted decomposition and transformation of lipids in garlic (e.g., increased content of alkanes such as 2,4-dimethyldodecane). Other unscented compounds were the most abundant in the DW group (28 types), reflecting that DW pretreatment alone retained more unscented compounds, while other pretreatments caused some unscented compounds to transform or be lost.

Principal component analysis (PCA) is a multivariate statistical analysis technique that simplifies data and reveals interrelationships among different samples, used to describe unsupervised classification trends (Liao et al., 2022). Fig. 7(a) shows the PCA plots of garlic with different pretreatment methods based on the signal strength of volatile compounds. The cumulative variance contribution rates of principal component 1 and principal component 2 were close to 80%, indicating that these two principal components can better represent the differences in volatile components of garlic under different pretreatment methods. It can be clearly observed from the graph that garlic samples with different pretreatment methods are distributed in different regions on the PCA graph. Samples of the same pretreatment method are clustered together, while samples of different pretreatment methods are clearly separated—consistent with the quantitative differences in volatile compounds shown in Appendix Table S1. This confirms that different pretreatment methods have a significant impact on the volatile components of garlic. This is similar to the effect of different drying methods on the volatile components of iron rod yam (Zhang et al., 2023), that is, PCA can effectively distinguish samples treated with different methods.

Fig. 7.

Multivariate statistical analysis of volatile compounds. (a) Principal component analysis (PCA) score plot showing distinct clustering of samples by pretreatment method. (b) Partial least squares-discriminant analysis (PLS-DA) score plot demonstrating clear separation between groups. (c) Cross-validation plot of the PLS-DA model, confirming model validity (no overfitting). (d) Variable importance in projection (VIP) plot; compounds with VIP >1 (e.g., diallyl disulfide, methyl 2-propenyl trisulfide) are key contributors to group differentiation.

Partial least squares discriminant analysis (PLS-DA) is a supervised classification method (Kalogiouri et al., 2021). Fig. 7(b) shows that the garlic samples of different preprocessing methods are significantly separated in the PLS-DA score graph and the separation effect is more significant than that of PCA, indicating that PLS-DA can better reflect the differences in volatile components caused by different preprocessing methods. The fitting index of the independent variable (R²) and the fitting index of the dependent variable (Rᵧ²) in the model are important indicators for evaluating the model. When R² and Rᵧ² are close to 1, it indicates that the model fits well. Guan et al. (2024) also arrived at comparable findings.

Cross-validation is used to verify the reliability and stability of the PLS-DA model and to avoid overfitting (Allen, Williams, & Sigman, 2019). Fig. 7(c) shows the cross-validation results of the model. After 200 permutations, the intersection point of the Q² regression line with the vertical axis is less than 0, indicating that there is no overfitting and the model is valid. The cross-validation results show that the model has good stability and predictive ability, and can accurately reflect the relationship between different pretreatment methods and volatile components of garlic, providing a reliable basis for subsequent screening of differential volatile components. The variable importance in projection (VIP) values are used to assess the importance of volatile compounds in differentiating different pretreatment groups. Compounds with VIP > 1 were considered to be effective in distinguishing differential volatile components among different pretreatment groups. Fig. 7(d) shows the VIP values of volatile compounds in garlic; the higher the VIP value of a compound, the more important it is in differentiating garlic samples with different pretreatment methods. As summarized in Appendix Table S1, these high-VIP compounds include diallyl disulfide (VIP > 1.2), methyl 2-propenyl trisulfide (VIP > 1.1), and hexanal (VIP > 1.0)—their significant content differences across groups (e.g., 29.7% higher diallyl disulfide in NaCl aq. + US than in DW) directly drove the clear separation of samples in PCA and PLS-DA. Zhang et al. (2023) obtained similar results in their experiment on iron rod yam.

3.5. Statistical analysis

After normalizing data on drying characteristics (average drying rate, drying time, D_eff), quality indicators (rehydration rate), and functional components (total phenol content, allicin retention with DADS+DATS as indicators, protein abundance, alliinase abundance) under four pretreatments, correlation patterns were analyzed via heat maps (Fig. 8). The most biologically and technologically significant correlations included: positive correlation between alliinase abundance and allicin retention (reflecting garlic's core enzymatic synthesis pathway), between D_eff and average drying rate (key to optimizing efficiency and reducing energy consumption), and between rehydration rate and cell structural integrity (determining dried product quality restorability).

Fig. 8.

Heat map of the correlation between different pretreatment methods and drying indicators.

The NaCl aq. + US group had the highest D_eff (Fig. 1(c)), attributed to ultrasonic cavitation and NaCl ionic effects synergistically disrupting cell structure to form water migration channels. It also exhibited the highest rehydration rate (Fig. 1(d)), as the controlled porous structure avoided excessive collapse, facilitating water infiltration during rehydration. Notably, this group had the highest alliinase abundance (Fig. 4(d))—a key enzyme for alliin-to-garlicin conversion—directly promoting high accumulation of DADS and DATS (stable allicin derivatives, indicating effective allicin retention). These results demonstrated a synergistic effect of structural disruption-enzyme activity retention-functional component accumulation. In terms of drying efficiency, the NaCl aq. + US group was significantly positively correlated with drying rate and D_eff, and negatively correlated with drying time, confirming its advantages in shortening processing time and enhancing water diffusion. In contrast, the Suc aq. + US group showed opposite trends due to sucrose's viscous film hindering water diffusion, with weaker correlations than the NaCl aq. + US group.

The underlying synergistic mechanism is integrated as follows: NaCl enhances ultrasonic cavitation to disrupt garlic cell membranes, forming microchannels that increase D_eff and accelerate moisture migration; NaCl regulates the ionic environment to inhibit polyphenol oxidase (reducing total phenol oxidation) and promote alliinase retention/release; ultrasound further facilitates alliin release, driving enzymatic synthesis of allicin, which degrades into stable DADS and DATS during processing—ultimately achieving integrated optimization of drying efficiency, structural preservation, and functional component accumulation.

In terms of quality and functional components, the NaCl aq. + US group showed the most positive correlations, including with rehydration rate (reflecting structural integrity), total phenol content, allicin retention (DADS+DATS accumulation), and alliinase abundance. This aligns with the synergistic effect: ultrasound and NaCl form a porous structure for rehydration, promote alliinase-mediated garlicin synthesis (retention indicated by DADS+DATS), and reduce phenol oxidation. The DW group, despite positive correlation with total protein abundance, showed negative correlations with drying efficiency, active components, and rehydration rate due to cell collapse. The DW + US group had weaker correlations with drying efficiency and functional components than the NaCl aq. + US group, reflecting limitations of ultrasound alone. The Suc aq. + US group was positively correlated with protein β-turn content (enhancing structural stability) but weakly correlated with drying efficiency and functional components, indicating sucrose's limited promotion of functional accumulation.

4. Conclusions

To address the efficiency bottleneck of EHD drying at low moisture content, this study investigated four pretreatments for garlic drying. The NaCl aq. + US pretreatment emerged as optimal, benefiting from the synergy of ultrasonic cavitation and NaCl ionic regulation. Cavitation forms microchannels and NaCl enhances permeability, constructing a controllable porous structure that boosts drying efficiency and rehydration performance. Meanwhile, ultrasound promotes alliin release, and NaCl regulates ionic environment to enhance alliinase activity and inhibit polyphenol oxidase, achieving efficient accumulation of DADS+DATS (3.69 ± 0.22 g/kg) and total phenols (0.9 ± 0.03 g/kg), while preserving protein structure and volatile sulfur-containing compounds.

This study first clarifies the mechanism of medium-regulated ultrasound: osmotic pressure and cavitation effects synergistically drive “structural disruption - enzyme regulation - component accumulation”, filling research gaps. Future work can optimize NaCl concentration and ultrasonic power, extend the mechanism to similar bulbous crops, and develop high-value garlic products to meet market demands.

CRediT authorship contribution statement

Wei Zuo: Writing – original draft, Visualization, Validation, Methodology, Conceptualization. Changjiang Ding: Writing – review & editing, Supervision, Resources, Project administration, Conceptualization. Jingli Lu: Writing – review & editing, Supervision, Resources, Project administration, Conceptualization. Chuanqiang Che: Writing – review & editing, Supervision. Zhiqing Song: Methodology, Conceptualization. Chunxu Qin: Methodology, 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.

Appendix A. Supplementary data

Supplementary material

Data availability

Data will be made available on request.