Effect of different preparation methods on the quality of carrot preserves: Chemical composition, textural properties, and volatile compounds

TL;DR

This study compares different methods of making carrot preserves, finding that non-enzymatic browning-treated preserves are more nutritious and better tasting.

Contribution

The novel contribution is the development of low-sugar, low-sodium carrot preserves with enhanced polyphenol content and improved sensory qualities.

Findings

NEBCP preserved more polyphenols (5.24-fold higher than fresh carrots) and had better protein and fiber retention.

NEBCP showed increased volatile compounds from Maillard reactions, improving aroma and sensory acceptability.

NEBCP had lower hardness and adhesiveness, higher elasticity, and was most acceptable in sensory tests.

Abstract

Systematic comparisons of quality impacts of various carrot preserve preparation techniques remain scarce. We compared candied carrot preserves (CCP), salted sugar crystallised carrot preserves (SSCP), and non-enzymatic browning-treated carrot preserves (NEBCP) for chemical composition, texture, and aroma. CCP better preserved carotenoids. NEBCP exhibited the highest total polyphenol content, 5.24-fold that of fresh carrots (FC), and retained more dietary fibre and protein. Its total carbohydrate content was 0.41-fold that of CCP, and sodium content was 0.043-fold that of SSCP. Headspace solid-phase microextraction coupled with gas chromatography–mass spectrometry (HS-SPME/GC–MS) identified 115 volatile compounds. Relative to FC, alcohols, ketones, esters, phenols, and hydrocarbons decreased in CCP and SSCP, while NEBCP showed increased levels of acids, aldehydes, lactones, and heterocyclic compounds from Maillard reactions and carotenoid degradation. Sensory evaluation indicated that NEBCP scored higher than CCP and SSCP. These findings provide insights for developing low-sugar, low-salt nutritious carrot preserves.

Unlabelled Image

•Non-enzymatic browning–treated carrot preserves (NEBCP) were developed with low sugar and low sodium contents.•NEBCP increased total polyphenols 5.24-fold and improved protein and dietary-fibre retention vs. conventional preserves.•Maillard products and carotenoid-derived volatiles increased aldehydes, ketones and heterocycles, enhancing aroma.•NEBCP showed lower hardness and adhesiveness, higher elasticity, and the highest overall acceptability among samples.

Non-enzymatic browning–treated carrot preserves (NEBCP) were developed with low sugar and low sodium contents.

NEBCP increased total polyphenols 5.24-fold and improved protein and dietary-fibre retention vs. conventional preserves.

Maillard products and carotenoid-derived volatiles increased aldehydes, ketones and heterocycles, enhancing aroma.

NEBCP showed lower hardness and adhesiveness, higher elasticity, and the highest overall acceptability among samples.

Figures

Macroscopic morphology and chemical composition of carrot preserves processed by three methods, compared with fresh carrot.Macroscopic morphology of four samples (A); Moisture content (B); Total saccharide content (C); Sodium content (D); Dietary fibre content (E); Protein content (F); Fat content (G); Total phenolic content (H); Carotenoid content (I). Different letters indicate significant differences (*p* < 0.05).Fig. 1

Texture properties and SEM microstructures of carrot preserves processed by three processing methods and fresh carrot.Texture parameters: hardness (A); Chewiness (B); Springiness (C); Cohesiveness (D); Adhesiveness (E). Different letters indicate significant differences (p < 0.05).SEM micrographs: CCP (A1, A2), SSCP (B1, B2), NEBCP (C1, C2), and FC (D1, D2).Fig. 2

Comparative characterisation of VOCs in carrot preserves processed by different methods and in fresh carrots.Bar chart of VOC category quantities (A); Venn diagram (B); Radar plot (C); Distribution of absolute concentrations (D); Distribution of relative contents (E); Dual-axis plot showing absolute contents (bars, left axis) and percentage of total VOCs (line, right axis) for top 10 compounds across 4 samples (F).Fig. 3

Comparative analysis of VOCs in fresh carrots and candied carrots processed using three different methods.Clustered heatmap of volatile components across the four sample groups (A), with VOC categories indicated by colored outer rings, samples arranged radially, compounds arranged circumferentially, and a central clustering tree.; PCA score plot (B); HCA dendrogram (C); loading plot (D).Fig. 4

Correlation network between key aroma compounds and sensory attributes in the 4 samples (A); Relative composition of aroma compounds in four samples (B–E).Fig. 5

Differential analysis of 36 aromatic compounds with OAV ≥ 1. Clustered heatmap (A), with the outer ring representing chemical classifications, the second ring indicating aroma types, samples arranged radially, and a central clustering tree.; OPLS-DA score plot (R^2^Y = 0.984, Q^2^ = 0.933) (B); Permutation testing with 200 iterations (R^2^ = 0.327, Q^2^ = −0.914) (C); VIP value distribution (D).Fig. 6

Sankey Diagram of Aroma Profile Intensity Distribution in Fresh Carrot and Carrot Preserves Processed by Three Different Methods.Fig. 7

Sensory evaluation scores of candied carrots prepared using three different processing methods.Radar charts showing the sensory scores of CCP, SSCP, and NEBCP samples, respectively, across the attributes of colour, appearance, texture, flavour, aftertaste, and overall acceptability (A–C).Stacked bar chart illustrating the total sensory scores composed of the individual attribute scores for the three samples (D).Fig. 8

Full text

Introduction

1

Carrots (Daucus carota L.) constitute one of the world's 10 major vegetable crops and have attracted attention owing to their dietary fibre, organic compounds, and other bioactive constituents (Fartoosi et al., 2025). According to Food and Agriculture Organization of the United Nations (FAO) statistics, the global carrot production reached 59.86 million tons in 2023, with China alone producing 18.46 million tons (approximately 30.85% of the global yield). Despite their popularity, fresh carrots (FC) are highly perishable due to their high moisture content, posing challenges for storage and distribution (Mondal et al., 2025). Processing carrots into preserved products extends shelf life and increases value. Currently, carrot preserves are mainly produced by sugar- or salt-curing, which involves blanching and osmotic treatment. However, these processes cause water migration, leading to the degradation of thermosensitive components and losses of polyphenols, dietary fibre, and volatile compounds (Nunes, Coimbra, et al., 2008; Xie et al., 2025; Tang et al., 2022). Prolonged osmotic pressure disrupts cell wall integrity, affecting texture (Xu et al., 2015). Conventional preservation methods result in excessive sugar or sodium content. Following China's ‘three-reduction’ dietary initiative promoting reduced salt, oil, and sugar consumption (Bai et al., 2022), there is growing demand for nutritious and healthy foods with reduced sugar and salt content. Developing preserved carrot products with diverse flavours, high nutritional value, and low sugar and sodium levels aligns with this health-conscious trend and presents significant market opportunities.

Non-enzymatic browning technology has been applied to products such as black garlic (Ríos-Ríos et al., 2019). In this study, we applied an improved non-enzymatic browning technique to carrot preserves (Li et al., 2025). This process involves chemical reactions, with the Maillard reaction playing a central role. This reaction transforms volatile organic components (VOCs) by generating flavour heterocyclic compounds through interactions between reducing sugars and amino acids (Ríos-Ríos et al., 2019; Yang et al., 2019). The Strecker degradation pathway produces aldehydes via amino acid decarboxylation and deamination (Kebede et al., 2017), while carotenoid breakdown contributes characteristic VOCs (Meléndez-Martínez et al., 2023). High-temperature and high-humidity conditions during blackening weaken intercellular connections and promote cell-wall polysaccharides hydrolysis and protein denaturation, softening texture (Ríos-Ríos et al., 2019), and giving NEBCP its dark brown appearance, soft texture, and sweet-sour flavour (Li et al., 2025).

Sugaring, salting, and non-enzymatic browning yield distinct nutritional, textural, and VOC profiles. While studies have focused on single parameters (Nunes, Coimbra, et al., 2008; Wang et al., 2023), systematic, process-oriented comparisons of their overall quality impacts remain scarce. The texture and flavour characteristics of non-enzymatically browned carrot preserves are particularly undercharacterized. Because these processing routes differ substantially in operating conditions such as processing time, osmotic medium, and drying method, applying uniform control of intermediate variables would drive products away from their optimal quality states and reduce technological relevance. Therefore, this study adopts processing route as the primary factor. Carrots from a single batch were used to prepare laboratory-scale products under representative sugaring, salting, and non-enzymatic browning conditions, each optimized to its preferred consumption state (Waseem et al., 2024; Wu et al., 2025). Chemical composition analysis, texture profile analysis (TPA), and scanning electron microscopy (SEM) were employed. VOCs were analysed with relative odour activity values (ROAVs) calculated to identify key aroma compounds. This study examined the processing effects on carrot preserve quality, providing a basis for industrialisation and product innovation.

Materials and methods

2

Materials and equipment

2.1

Carrots (Daucus carota L. cv. ‘Lowucun’) were sourced at commercial maturity from a wholesale market in District 9, Daqing, China; they originated from the Shandong Province, China. The following reagents were used: carbinol, acetone, sodium carbonate (Sinopharm Chemical Reagents Co., Ltd., Shanghai, China), forinol (Coolaber, Beijing, China), deionised water (Wahaha, Hangzhou, China), gallic acid, phenol, ethanol disodium 1,2-cyclohexanediaminetetraacetate (CDTA), carbazole, galacturonic acid (Macklin, Shanghai, China), deuterium n-hexanol-d_13_ (purity 98.5%; C/D/N Isotopes, Quebec, Canada), n-alkanes (SIGMA, USA), and n-hexane, GR grade (Yonghua, Shanghai, China). The following instruments were employed: a microplate reader (SpectraMax ABS Plus, Molecular Devices, San Jose, USA), a UV–visible spectrophotometre (U-T181s, Qipu Instruments, Shanghai, China), a rotary evaporator (VRT-20, Shanghai Xiande Experimental Instrument Co., Ltd., Shanghai, China), and a flame photometre (640 A, Yidian Keyi, Shanghai, China).

Sample processing

2.2

Candied carrot preserves (CCP), salted sugar crystallised carrot preserves (SSCP), and non-enzymatic browning–treated carrot preserves (NEBCP) were prepared following published methods with minor modifications (Li et al., 2025; Nunes, Coimbra, et al., 2008; Tang et al., 2022), as detailed in Supplementary Material S1.1. Thereafter, 500 g of CCP, SSCP, NEBCP, and FC was mixed with 500 mL deionised water, sealed, and soaked at 4 °C for 24 h and homogenised (30,000 rpm, 2.5 min; B15, Joyoung, Shandong, China). The resulting slurries were freeze-dried (SCIENTZ 50FG/A, Ningbo Xinzhi Biotechnology Co., Ltd., Zhejiang, China; −40 °C, 5 h; −20 °C, 50 Pa, 20 h; 10 °C, 50 Pa, 10 h), ground, and passed through a 100-mesh sieve to obtain CCP, SSCP, NEBCP, and FC powders, which were packed in aluminium foil bags and stored at −20 °C until analysis.

Determination of chemical components

2.3

Detailed analytical procedures are provided in Supplementary Table S1.

Textural profile analysis

2.4

Three carrot preserve samples and FC were randomly selected. Phloem tissue was separated and cut into uniform cubes (5 × 5 × 5 mm) Each cube was centrally positioned on a texture analyser (EZ-X, Shimadzu, Japan). Double-cycle compression tests were conducted with the following parameters: a 36-mm-diameter disk probe; trigger force, 5 g; compression, test, and return speeds were all set to 1 mm s^−1^, respectively; compression ratio, 60%. Hardness, elasticity, cohesiveness, adhesiveness, and chewability were recorded.

SEM analysis

2.5

Three carrot preserve types and FC were randomly selected, with three specimens per sample. Phloem tissue was excised, and the central radial region was cut into cubes (≈3 mm edge length), then dehydrated at 65 °C for 8 h. The samples were sputter-coated with gold (10 nm) in an ion sputter coater (JFC-1600, JEOL, Japan) and examined using SEM (Gemini 300; Zeiss, Germany) under high vacuum at 15 kV, 8.5-nm working distance, and magnifications of 100× and 500×.

HS-SPME/GC–MS analysis

2.6

HS-SPME method

2.6.1

Following the method of Wu et al. (2025) with minor modifications, 5.0 g of freeze-dried sample powder was placed into a 20-mL headspace vial. Deuterium n-hexanol-d_13_ was added to the vial, which was immediately sealed. Extraction fibres (DVB/CAR/PDMS, 50/30 μm × 1 cm, Bellefonte, USA) were preconditioned at 270 °C for 10 min, then inserted into the vial. Samples were equilibrated at 60 °C for 10 min, followed by adsorption for 15 min. Once adsorption was complete, the SPME fibre was transferred to the GC injection port and desorbed at 250 °C for 5 min. Following desorption, the SPME fibre was aged at 270 °C for 10 min. Finally, we introduced 10 μL of n-alkanes into a 20-mL headspace vial under identical incubation and extraction conditions.

GC–MS detection

2.6.2

Volatile compounds were analysed on an 8890 A mass spectrometre (Agilent Technologies, Palo Alto, CA, USA) coupled with a LECO Pegasus BT 4D time-of-flight mass spectrometry (TOFMS) detector (LECO, St. Joseph, MI, USA), using a DB-HeavyWax capillary column (30 m × 250 μm × 0.5 μm, Agilent, CA, USA). Helium (purity >99.99%) served as the carrier gas at 1.0 mL/min. The temperature programme was as follows: the initial column temperature was held at 50 °C for 2 min, then increased at a rate of 5 °C/min to 85 °C and was held for 5 min. Both the transmission line and ion source temperatures were held at 250 °C. Mass spectrometry settings included an electron ionisation (EI) source energy of 70 eV, a detector voltage of 2035 V, an acquisition frequency of 10 spectra/s, and a scanning mass range of m/z 35–550.

Characterisation and quantification of volatile compounds

2.7

Compound identification was performed by simulated research against the NIST 2017 library. The retention indices (RI) of volatile substances were calculated based on the retention times of n-alkanes using Eq. (1).

$$[eqn]$$

where n and n + 1 represent the carbon numbers of normal alkanes, ti is the retention time of component i (between C_n_ and C_n+1_)/min, tn is the retention time of normal alkane Cn/min, and tn+1 is the retention time of normal alkane Cn+1/min.

A semi-quantitative method was used to calculate the relative contents of volatile substances. C_i_ was calculated using Eq. (2).

$$[eqn]$$

where Ci and Cs represent the content (μg/kg) and peak area of VOCs respectively; Ai and As represent the content (μg/kg) and peak area of the internal standard substance, respectively.

The aroma thresholds were sourced from Van Gemert (2011). The aroma descriptions were obtained by querying the (http://thegoodscentscompany.com/) database. The odour activity values (OAV) of volatile components were calculated following the method described by Wu et al. (2025) using Eq. (3). The component with the highest OAV in each group was set as ROAVmax = 100, and the ROAV was calculated using Eq. (4).

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

where Ci represents the content of volatile components (μg/kg), Ti is the aroma threshold of this volatile substance in water, and OAV_max_ is the maximum OAV value among the samples.

Sensory evaluation

2.8

A total of 20 screened panelists (10 males and 10 females, 27–50 years old) received systematic sensory training for 3 weeks in accordance with GB/T 46555–2025. A 100-point scale was adopted, with appearance as the primary attribute (30 points), texture and flavour as key attributes (25 points each), and colour, aftertaste, and overall acceptability as secondary attributes (10 points each), thereby reflecting the relative contributions of each sensory dimension to candied carrot quality (Table S2).

Data processing

2.9

Numerical analyses were conducted in Excel 2019, with data expressed as mean ± standard deviation (n ≥ 3). Normality was tested using the Shapiro-Wilk test, performed in SPSS 27.0 (IBM, Armonk, NY, USA). For normally distributed data, one-way ANOVA was applied. Duncan's post-hoc test was used for homogeneous variances, and Games-Howell post-hoc test was used for heterogeneous variances. Statistical significance was set at p < 0.05. Origin 2021 (Origin Lab, Inc., Northampton, MA, USA) for bar charts, box plots, column charts, radar charts, stacked charts, Venn diagrams and Sankey diagrams. SIMCA 14.1 (Umetrics, Umeå, Sweden) was used for principal component analysis (PCA), orthogonal partial least squares-discriminant analysis (OPLS-DA), and hierarchical cluster analysis (HCA), and visualisation. Rstudio (Posit, Boston, USA) with the ‘scale()’ function was used for data centring and heat map generation. Cytoscape v3.9.1 (Cytoscape Consortium, San Francisco, USA) was used for correlation networks. Figures were finalised in Adobe Illustrator 2023.

Results and analysis

3

Chemical component detection and analysis

3.1

Processing techniques markedly influence the chemical component of carrot preserves (Fig. 1, Tatel S3). Fig. 1A presents a comparison of the colour and morphology of products from the three processes with FC. FC exhibited a bright orange-red hue owing to intact cell structures and retained carotenoids, serving as the colour benchmark. CCP displayed a deep orange-red colour, primarily due to pigment concentration during dehydration. Maillard and caramelisation reactions during post-sugar infiltration drying generated dark chromophores (Nunes, Coimbra, et al., 2008; Yang et al., 2019). By contrast, SSCP appeared lighter because carotenoids undergo oxidative degradation during salting and air-drying, leading to pigment loss (de Oliveira et al., 2025; Meléndez-Martínez et al., 2023). The surface sugar crystallised layer further brightened the appearance via light scattering. NEBCP subjected to high temperature and humidity displayed alternating dark brown and orange hues. This process intensified the Maillard reaction, generating substantial brown substances such as melanoidins that partially mask the carrots' natural orange-red colour (Yang et al., 2019).Fig. 1. Macroscopic morphology and chemical composition of carrot preserves processed by three methods, compared with fresh carrot.Macroscopic morphology of four samples (A); Moisture content (B); Total saccharide content (C); Sodium content (D); Dietary fibre content (E); Protein content (F); Fat content (G); Total phenolic content (H); Carotenoid content (I). Different letters indicate significant differences (p < 0.05).Fig. 1

Fig. 1(B–I) illustrate significant differences in the chemical composition of the four samples. Fig. 1B indicates that FC had the highest water content (89.33 ± 0.86%), with the water content of the carrot preserves ranked as NEBCP (29.94 ± 1.75%), CCP (26.73 ± 1.22%), and SSCP (22.14 ± 0.40%). In Fig. 1(C–I), the blue area on the left denotes the fresh-weight content of the chemical components, while the pink area on the right indicates the dry-weight content. The fresh-weight values closely reflect nutritional and energy intake, thus providing a precise basis for dietary assessment. Additionally, the contents of key substances, such as total saccharides and sodium, are directly linked to sensory quality, serving as scientific indicators for predicting taste characteristics and consumer acceptance (Meijer et al., 2021). The saccharide content directly influences sweetness (Fig. 1C). As expected, CCP exhibited the highest total saccharide content (66.05 ± 2.49 g/100 g FW), attributed to the sugar infiltration process. Sodium contents are shown in Fig. 1D. Notably, SSCP underwent salting during processing, significantly elevating their sodium content to 4425.8 ± 166.55 mg/100 g FW. Using a dry-weight basis effectively captures absolute changes in nutritional components during processing by eliminating the influence of moisture variation. This method is particularly suitable for assessing the retention of heat-sensitive components and evaluating processing technologies, providing a scientific foundation for optimising food processing (Nielsen, 2017). Fig. 1(E and F) demonstrate that hydrothermal treatment and osmotic dehydration significantly affect the migration of nutritional components. The dietary fibre (CCP: 8.67 ± 0.61 g/100 g DW; SSCP: 14.35 ± 0.43 g/100 g DW) and protein contents (CCP: 2.01 ± 0.10 g/100 g DW; SSCP: 2.51 ± 0.08 g/100 g DW) in the CCP and SSCP groups were both significantly lower than those in the FC and NEBCP groups (p < 0.05), primarily because of cell-structure disruption from blanching in boiling water, which led to the dissolution and loss of soluble dietary fibre and certain hydrophobic proteins. Osmotic dehydration removes water and water-soluble components (Sun et al., 2025). During NEBCP processing under high-temperature and high-humidity conditions, some dietary fibres and proteins are lost and degraded, while other proteins undergo the Maillard reaction (Yang et al., 2019), Consequently, the dietary fibre (22.79 ± 0.52 g/100 g DW) and protein (4.78 ± 0.18 g/100 g DW) contents were both significantly lower than those in the FC group (p < 0.05).

Polyphenols are known for their anti-inflammatory and antioxidant properties. As illustrated in Fig. 1H, the polyphenol content in NEBCP (558.95 ± 65.41 mg GAE/100 g DW) was significantly higher than that in the other three samples (p < 0.05), and was 5.24-fold greater than that of FC. This is closely related to its processing conditions; under sustained high-temperature and high-humidity conditions, the Maillard reaction is accelerated and causes changes in system acidity. This microenvironment may promote the hydrolysis of polyphenols bound via ester or glycosidic linkages, releasing them into more easily extractable free forms, thereby increasing the measured polyphenol content (Li et al., 2025; Ríos-Ríos et al., 2019), significantly increasing the polyphenol content in NEBCP than the other three samples (p < 0.05). Carotenoids (Fig. 1I), the key active compounds in carrots, are known for their cancer-preventive properties and role in protecting cardiovascular, cerebrovascular, and eye health. Despite their benefits, carotenoids are thermally unstable and can degrade into volatile compounds at elevated temperatures (Meléndez-Martínez et al., 2023). Therefore, the carotenoid contents in all three types of carrot preserves were significantly lower than that of FC (34.61 ± 1.97 mg/100 g DW). Notably, carotenoid retention in CCP (25.90 ± 1.05 mg/ 100 g DW) was significantly greater than that in SSCP (9.98 ± 0.72 mg/ 100 g DW) and NEBCP (16.20 ± 2.13 mg/ 100 g DW) (p < 0.05). This increased retention may be because of sugar osmosis within the tissue, which partially inhibits carotenoid oxidation during the drying process (Wang et al., 2023).

Analysis of textural characteristics and micromorphology

3.2

The textural attributes of food play a vital role in its sensory quality and consumer acceptance (Nishinari et al., 2024). TPA was performed on carrot preserves prepared by three different processing methods, as well as on FC samples. Five key parameters—hardness, chewiness, springiness, cohesiveness, and adhesiveness—were selected for comprehensive characterisation (Table S4), revealing significant differences among groups (p < 0.05). In addition, the microstructure of each sample was examined by SEM (Figs. 2 and S1).Fig. 2. Texture properties and SEM microstructures of carrot preserves processed by three processing methods and fresh carrot.Texture parameters: hardness (A); Chewiness (B); Springiness (C); Cohesiveness (D); Adhesiveness (E). Different letters indicate significant differences (p < 0.05).SEM micrographs: CCP (A1, A2), SSCP (B1, B2), NEBCP (C1, C2), and FC (D1, D2).Fig. 2

Panels D1 and D2 in Fig. 2 show that FC exhibits a compact, honeycomb-like network structure. The surface wrinkles may result from dehydration or fracture of the primary cell wall and middle lamella pectin (Mondal et al., 2025). In the hydrated state, the intact cell wall and strong intercellular adhesion maintain turgor pressure and form a three-dimensional load-bearing network, directly conferring high hardness (164.22 ± 4.47 N) and relatively high springiness (0.62 ± 0.01) to the samples (Fig. 2A, C). In contrast, thermal processing, particularly blanching, induces tissue water loss, cell wall rupture, and pectin solubilisation, which severely disrupt the native network structure. This structural degradation is the fundamental reason for the marked reduction in hardness (p < 0.05) observed in all processed samples compared with FC (Mondal et al., 2025; Xu et al., 2015).

Building on the substantial hardness loss induced by blanching, CCP was further subjected to sucrose impregnation and drying, resulting in the formation of abundant polyhedral sucrose crystal clusters on the surface and within tissue pores (Fig. 2A1, A2 and S1). These crystals act as rigid fillers and structural reinforcements, partially compensating for the hardness loss caused by the disruption of the native network, thereby conferring significantly higher hardness (12.67 ± 1.66 N) than that in NEBCP (p < 0.05). The dense packing of the crystal clusters and the binding effect of the residual syrup (Nunes, Santos, et al., 2008) led to significantly higher springiness (0.91 ± 0.10), cohesiveness (0.60 ± 0.06), and adhesiveness (1.96 ± 0.44 N·mm), compared with the other three samples (p < 0.05; Fig. 2A, C**–**E). Within this texture profile, high hardness and springiness directly increase the biting force required at the onset of mastication and prolong chewing duration. High cohesiveness promotes bolus aggregation and formation during mid-mastication; however, the high adhesiveness causes extensive residue on oral surfaces, thereby increasing the subsequent oral clearance effort, which may in turn delay swallowing and diminish overall sensory comfort (Chen & Lolivret, 2011).

The microstructure of SSCP was markedly reconstructed after blanching, salting, and press dehydration. SEM observations (Fig. 2B1, B2 and S1) showed that the originally continuous honeycomb like cell lumen structure had almost disappeared, and replaced by numerous detached cell wall fragments and larger, dense aggregates. Gel like and granular deposits were present among these structures, and the lamellae were in close contact. This morphology suggests that the osmotic pressure generated during salting promoted the leakage of intracellular contents (Llorca et al., 2001; Tang et al., 2022), which were subsequently redistributed, filled, and compacted with fibrous debris during pressing. Consequently, porosity decreased and the solid–solid contact area increased, thereby enhancing compressive resistance and producing a hardness value (52.98 ± 4.21 N, p < 0.05) significantly higher than that of CCP and NEBCP (Fig. 2A). Meanwhile, processing disrupted the native cell wall network and the elasticity mechanism associated with cell turgor, resulting in springiness (0.40 ± 0.04, p < 0.05) (Fig. 2C). Once the cell wall network was disintegrated into fragments and re aggregated into stacked structures, the overall integrity depended mainly on physical compaction and weak interfacial adhesion. Microcracks or slip planes readily formed at the interfaces between aggregates and lamellae, compromising structural integrity and leading to cohesiveness (0.39 ± 0.03) that was significantly lower than that of the other samples (p < 0.05) (Fig. 2D). This ‘high hardness/low springiness’ texture combination indicated insufficient textural harmony: oral processing required greater initial chewing force, while the weak rebound, limited buffering, and poor force feedback could increase masticatory load and diminish sensory pleasure (Nishinari et al., 2024; Volkow et al., 2011).

The hardness of NEBCP (2.85 ± 0.32 N) was significantly lower than that of the other samples (p < 0.05; Fig. 2A). Compared with that in FC, its springiness (0.51 ± 0.09) decreased (Fig. 2C), whereas its adhesiveness (0.32 ± 0.11 N mm) increased significantly (p < 0.05) (Fig. 2E), and its cohesiveness (0.48 ± 0.04) was maintained or showed an increasing trend (Fig. 2D). SEM images of NEBCP (Fig. 2C1, C2) revealed penetrating crack-like separation bands accompanied by locally flat, continuous membrane-like coverage, suggesting that intercellular interfaces were more susceptible to cracking and separation, thereby weakening structural continuity. The processing mechanism of NEBCP differed from that of the other two samples and was characterised by pronounced softening coupled with increased adhesiveness, and the structural basis for these texture changes has not been systematically clarified; therefore, NEBCP and the control FC were further selected for pectin fractionation and FTIR analysis (Fig. S2) to investigate how alterations in cell wall pectin state and carboxyl-related chemical structures relate to the corresponding textural responses.

Pectin fractionation (Fig. S2A) showed that compared with those of FC, the contents of sodium carbonate–soluble pectin (SSP; 3.77 ± 0.90 mg/g) and chelator-soluble pectin (CSP; 1.93 ± 0.31 mg/g) in NEBCP decreased significantly, by approximately 13.02- and 9.08-fold, respectively, whereas the Water-Soluble Pectin (WSP) content (65.90 ± 1.32 mg/g) increased significantly (p < 0.01), by approximately 2.57-fold. This indicates that cell wall-bound pectin at the intercellular interface migrated toward water-soluble pectin, which may weaken middle lamella cementation and cell wall network constraints, thereby reducing resistance to compressive deformation and leading to decreased hardness and chewiness (De Roeck et al., 2007). The increase in WSP implies a higher proportion of mobile soluble macromolecules, which, during slow dehydration and setting, may enrich at the surface or in intercellular spaces and form a continuous adhesive phase, giving rise to the membrane-like coverage observed in the microstructure and enhancing adhesiveness. The maintenance or increase in cohesiveness may be related to the fact that the internal framework retains a certain degree of continuity and that the soluble phase may provide secondary bridging across cracks (Moelants et al., 2014). The FTIR results (Fig. S2B; Table S5) showed that both samples retained the typical absorption bands of pectin, indicating that the differences mainly arose from the chemical state of carboxyl groups rather than the loss of the pectin backbone. Compared with FC, NEBCP exhibited a strengthened esterified carbonyl peak at 1738 cm^−1^ and weakened –COO^−^ absorptions at 1619 and 1415 cm^−1^, suggesting an increased proportion of esterified/non-ionised carboxyl groups in pectin and a decreased proportion of –COO^−^ groups available for ionic binding (Kyomugasho et al., 2015). As Ca^2+^ bridges mediated by –COO^−^ groups are typically more closely associated with middle lamella adhesion and cell wall network constraints (Obomighie et al., 2025), these changes are consistent with the interfacial cracking and separation observed by SEM, as well as the reduced load-bearing capacity reflected in texture measurements. Furthermore, the enrichment of WSP, providing an increased mobile soluble phase, may also offer the material basis for enhanced adhesiveness (Moelants et al., 2014). The ‘softer and stickier’ texture of NEBCP mainly stems from the transformation of bound pectin (SSP, CSP) into soluble pectin (WSP), accompanied by a reduction in –COO^−^ groups, which weakens middle lamella cementation and cell wall constraints and diminishes structural load-bearing capacity. Concurrently, soluble pectin becomes enriched at surfaces and in intercellular spaces, forming a membrane-like continuous phase that enhances interfacial adhesiveness. These combined changes in pectin fraction distribution and carboxyl group chemistry reconstruct the mechanical coupling between intercellular interfaces and the cell wall network, mapping microstructural alterations onto macroscopic texture parameters. Overall, NEBCP exhibits a structure that is easily deformable, resistant to disintegration, and moderately sticky, thereby reducing chewing effort, improving swallowing smoothness, and decreasing oral residue (Volkow et al., 2011).

Detection of VOC compounds in FC and carrot preserves prepared by the three methods

3.3

VOC identification and class profiles

3.3.1

To compare FC with preserves produced by the three methods, we profiled VOCs by HS-SPME/GC–MS (Table S6). A total of 115 VOCs were identified across the samples, comprising 12 acids, 16 alcohols, 16 ketones, 15 aldehydes, 12 esters, 3 lactones, 27 hydrocarbons (5 alkanes, 4 alkenes, 8 aromatic hydrocarbons, and 10 terpene hydrocarbons), 2 phenols, 10 heterocyclics (5 furans, 2 pyrazines, and 3 other heterocyclic compounds), and 2 additional VOC types. Fig. 3A shows that 58, 58, 72, and 80 VOCs were detected in the CCP, SSCP, NEBCP, and FC samples, respectively. The Venn diagram in Fig. 3B highlights these compositional differences. Compared with FC, CCP and SSCP lost 38 VOCs each, while NEBCP lost 13. Meanwhile, 24 new VOCs appeared in CCP and SSCP, and 21 in NEBCP. Notably, 25 VOCs were consistently found across all 4 sample groups. Significant differences were observed in the compositional diversity of VOCs across the four sample groups (Fig. 3C). CCP and SSCP showed a broader array of acids, while NEBCP and FC exhibited a richer variety of chemical classes, including alcohols, ketones, aldehydes, and esters. Furthermore, NEBCP contained a more diverse profile of heterocyclic compounds than the other three samples. Quantitatively (Table S6; Fig. 3D and E), phenols were absent from CCP and SSCP, and alcohols, ketones, esters, and hydrocarbons were markedly reduced. Specifically, in the CCP group the following values were recorded: alcohols (612.16 ± 23.13 μg/kg, 6.16%), ketones (1300.27 ± 88.46 μg/kg, 13.09%), esters (368.82 ± 17.36 μg/kg, 3.71%), and hydrocarbons (1998.04 ± 116.46 μg/kg, 20.12%). In the SSCP group the following values were recorded: alcohols (734.20 ± 45.17 μg/kg, 6.88%), ketones (1921.50 ± 230.86 μg/kg, 18.00%), hydrocarbons (2955.30 ± 196.13 μg/kg, 27.68%), and esters (353.46 ± 17.51 μg/kg, 3.31%). These losses are attributable to processing. In CCP and SSCP processing, high-temperature blanching disintegrated the cell membrane and loosened the cell wall, causing cell contents to exude ((Wu et al., 2025; Xie et al., 2025). This initiates Strecker degradation, resulting in the loss of VOCs (Kebede et al., 2017). During sugar infiltration and salting, high osmotic pressure swiftly precipitates intracellular water. Under the combined influence of water and heat, thermally sensitive and highly hydrophilic hydrocarbons, phenols, alcohols, ketones, and esters undergo accelerated thermal degradation, facilitated by water migration, significantly reducing their content. The hot-air drying process was employed to solidify and shape CCP and SSCP, inducing Maillard, degradation, and redox reactions. These reactions facilitated the formation of new flavour compounds, including acids, furans, and aldehydes (Xie et al., 2025). During CCP hot-air drying, Maillard, thermal degradation, and redox reactions promote the formation of acids, furans, and aldehydes (Wang et al., 2023). These reactions likely facilitate the formation and accumulation of acids and heterocyclic compounds in CCP, reaching 1755.89 ± 165.81 μg/kg (17.68%) and 1679.89 ± 55.01 μg/kg (16.92%), respectively. In SSCP, possible fermentation during salting (Tang et al., 2022) leads to acid accumulation, while high-temperature sugar melting during sugar crystallisation, similar to CCP hot-air drying, further promotes the formation of acids and heterocyclic compounds. Consequently, acids and heterocyclics in SSCP reached 1032.35 ± 114.35 μg/kg (9.67%) and 2021.27 ± 70.60 μg/kg (18.93%), respectively.Fig. 3. Comparative characterisation of VOCs in carrot preserves processed by different methods and in fresh carrots.Bar chart of VOC category quantities (A); Venn diagram (B); Radar plot (C); Distribution of absolute concentrations (D); Distribution of relative contents (E); Dual-axis plot showing absolute contents (bars, left axis) and percentage of total VOCs (line, right axis) for top 10 compounds across 4 samples (F).Fig. 3

The preparation of NEBCP fundamentally differs from that of CCP and SSCP, resembling the process employed for black garlic. During processing, NEBCP undergoes the Maillard reaction alongside lipid oxidation and degradation (Yang et al., 2019), which substantially increases the levels of acids (958.4 ± 7.74 μg/kg, 5.60%), aldehydes (2648.04 ± 34.47 μg/kg, 15.46%), lactones (605.41 ± 35.73 μg/kg, 3.54%), and heterocyclic compounds (2655.03 ± 106.89 μg/kg, 15.50%). Research has shown that black garlic contains significantly higher levels of acids, aldehydes, and furans than fresh garlic (Yang et al., 2019). Similarly, the HS-SPME/GC–MS analysis indicated a marked increase in acids, aldehydes, and heterocyclic compounds during black garlic processing, contributing to its sour, caramel-sweet, and roasted aroma, and altering the sulphur-dominated flavour profile of fresh garlic (Yu et al., 2024). These findings align closely with the results of the present study. Furthermore, the absence of blanching and pickling renders this process ‘mild’, preventing VOC loss and preserving alcohols (3516.99 ± 18.15 μg/kg, 20.54%), ketones (3419.08 ± 84.27 μg/kg, 19.96%), esters (893.69 ± 47.02 μg/kg, 5.22%), hydrocarbons (2499.22 ± 194.03 μg/kg, 14.26%), and phenols (510.54 ± 13.08 μg/kg, 2.91%).

The top 10 VOCs in the 4 samples are shown in Fig. 3F, with group differences shown in Fig. S3. Among all samples, 6-methyl-5-hepten-2-one and β-caryophyllene were consistently detected. The former, a carotenoid-derived compound, contributes citrus, green, and lemongrass aromas (Meléndez-Martínez et al., 2023), whereas β-caryophyllene, a common sesquiterpene in carrots, imparts spicy, sweet, and woody aromas (Kjeldsen et al., 2003). Maillard-derived volatiles showed distinct enrichment patterns. Furfural and 5-hydroxymethyl-2-furaldehyde (5-HMF) were elevated in CCP and SSCP, with SSCP also containing abundant 2-acetylfuran. Sugar osmosis combined with hot-air drying promotes furfural formation, producing toasted, sweet, and almond aromas (Nunes, Coimbra, et al., 2008). 5-HMF formed during intermediate Maillard stages, mainly contributes caramel notes (Yang et al., 2019). 2-acetylfuran, prevalent in dried and candied fruits, enhances sweet and almond (van Boekel, 2006). In NEBCP, dihydroactinidiolide and hydroxyacetone were particularly abundant. Dihydroactinidiolide, a thermal degradation product of β-carotene, imparts plum, berry, and grape notes (Meléndez-Martínez et al., 2023), while hydroxyacetone, generated via the Maillard reaction or sugar degradation, contributes caramel and sweet flavours (van Boekel, 2006).

Differential analysis of VOCs

3.3.2

A heatmap (Fig. 4A) displays VOC content differences and clustering patterns among samples, highlighting compositional variations. PCA was employed for dimensionality reduction, focusing on features with maximum variance (Xie et al., 2025). In the PCA score plot (Fig. 4B), the first two principal components account for 80.9% of the variance, with PC1 and PC2 contributing 48.5% and 32.4%, respectively. The score plot shows separation among CCP, SSCP, NEBCP, and FC, although CCP and SSCP are closely positioned, suggesting similar profiles (Liang et al., 2024). The HCA results show that the CCP and SSCP samples cluster together, due to similar VOC profiles from analogous preparation methods (Fig. 4C). The NEBCP and FC samples exhibit closer clustering as NEBCP preparation better preserves original VOCs. Fig. 4D identifies acids, alcohols, ketones and aldehydes as key VOCs differentiating the principal components. The NEBCP and FC groups are characterised by alcohols and ketones, while the CCP and SSCP groups are distinguished by acids. However, differences in VOC content do not directly indicate flavour variations. Despite variations in aroma thresholds, compounds with high content but elevated thresholds may minimally impact overall flavour. For instance, 6-methyl-5-hepten-2-one and butyric acid, with thresholds of 0.1 ppm and 3 ppm, respectively (Van Gemert, 2011), may be key differentiators, yet have limited flavour impact. Thus, OAV evaluation is essential to determine true flavour differences among the groups.Fig. 4. Comparative analysis of VOCs in fresh carrots and candied carrots processed using three different methods.Clustered heatmap of volatile components across the four sample groups (A), with VOC categories indicated by colored outer rings, samples arranged radially, compounds arranged circumferentially, and a central clustering tree.; PCA score plot (B); HCA dendrogram (C); loading plot (D).Fig. 4

OAV analysis of key aroma compounds

3.3.3

The contribution of VOCs to food aroma is determined by their concentrations and odour perception thresholds. We utilised an OAV analysis with odour thresholds of each VOC (Table S7), to assess aroma impact. Compounds with 0.1 < OAV < 1 were identified as aroma-active, totalling 13 substances. Those with OAV ≥ 1 were key aromatic contributors (Liang et al., 2024), with 36 compounds identified: 1 acid, 4 alcohols, 9 ketones, 5 hydrocarbons, 9 aldehydes, 3 esters, 1 lactone, 1 phenol, and 3 heterocyclic compounds. The relationship between these compounds and aroma attributes is illustrated via network analysis (Fig. 5A). Among them, 11 VOCs had OAVs exceeding 100, including 1-octen-3-ol, 2-methylbutanal, carvacrol, dihydroactinidiolide, ethyl cinnamate, ethyl hexanoate, heptanal, hexanal, octanal, α-ionone, and β-ionone, forming the samples' fundamental aroma profiles. Dihydroactinidiolide, α-ionone, and β-ionone were all degradation products of β-carotene, impart floral, berry, and tropical fruit aromas. Thermal processing of carrots induced carotenoid cleavage and oxidative degradation of unsaturated fatty acids, producing numerous low-molecular-weight volatile aldehydes (C₆–C₈), notably heptanal, hexanal, and octanal (Meléndez-Martínez et al., 2023; Puganen et al., 2024) that confer green/leafy notes (Wu et al., 2025); hexanal may also present woody/citrus notes, and octanal herbal/fresh scents. Compound 1-octen-3-ol naturally occurs in carrots, and its levels rise significantly upon mechanical damage (Miyamoto et al., 2014), contributing earthy and mushroom aromas (Wu et al., 2025). Ethyl cinnamate and ethyl hexanoate contribute sweet, fruity, and pineapple aromas, while carvacrol, a terpenophenol, imparts pungent, herbal, and thyme-like aromas and exhibits antibacterial, antioxidant, and anti-inflammatory properties (Javed et al., 2023). During carrot thermal processing, reducing sugars produce α-dicarbonyl compounds through the Maillard reaction and undergo Strecker degradation with leucine, resulting in compound 2-methylbutanal, which adds cocoa, nutty, and caramel aromas (Kebede et al., 2017). Collectively, Fig. 5(B–E), shows that these 11 compounds define the samples' fundamental aroma.Fig. 5. Correlation network between key aroma compounds and sensory attributes in the 4 samples (A); Relative composition of aroma compounds in four samples (B–E).Fig. 5

Fig. 6A shows the clustering of 36 aromatic compounds with OAV > 1 across samples. Significant differences exist among the 4 samples. Orthogonal Partial Least Squares Discriminant Analysis (OPLS-DA), a supervised analysis method utilising orthogonal partial least squares regression for discriminant analysis and visualisation (Xie et al., 2025), was constructed using the 36 compounds (Table S7). R^2^X and R^2^Y denote the model's explanatory power for the X matrix (aromatic compound data) and Y matrix (grouping information), respectively, while Q^2^ indicates predictive capability. In Fig. 6B, R^2^Y = 0.984 and Q^2^ = 0.933; these values approximate 1, demonstrating a good model fit. CCP and SSCP appear in the first and fourth quadrants, while FC and NEBCP are in the second and third quadrants, respectively, indicating distinct separation among the samples. Model reliability was confirmed through 200 cross-permutation tests. In Fig. 6C, the abscissa reflects sample permutation status, with points at 1 corresponding to the original model's R^2^ and Q^2^. The model's R^2^ (0.327) and Q^2^ (−0.914) values are below 1.0, with Q^2^ regression line's intercept being negative, indicating no overfitting. In the OPLS-DA analysis, significant aroma components were identified using variable importance in projection (VIP) values, with VIP ≥ 1 indicating key differences. Fig. 6D highlights five aromatic compounds with VIP ≥ 1: β-ionone, α-ionone, octanal, carvacrol, and 2-methylbutanal, which distinguish the aroma profiles of carrot preserves processed differently. The loading plot (Fig. S4) reveals that 2-nonanone, 2-octanone, 6-methyl-5-hepten-2-one, 1-hexanol, 1-heptanol, ethyl acetate, and myrcene are predominant in FC. Furfural, benzothiazole, and naphthalene are key in CCP. SSCP are characterised mainly by piperonal, benzaldehyde, p-cymene, and hexanoic acid, while NEBCP are dominated by 3-methyl-1,2-cyclopentanedione, hydroxyacetone, acetoin, and phenylacetaldehyde.Fig. 6. Differential analysis of 36 aromatic compounds with OAV ≥ 1. Clustered heatmap (A), with the outer ring representing chemical classifications, the second ring indicating aroma types, samples arranged radially, and a central clustering tree.; OPLS-DA score plot (R^2^Y = 0.984, Q^2^ = 0.933) (B); Permutation testing with 200 iterations (R^2^ = 0.327, Q^2^ = −0.914) (C); VIP value distribution (D).Fig. 6

ROAV analysis

3.3.4

ROAV analysis uses OAV results, with the highest OAV compound as the benchmark (ROAV = 100); other compounds' ROAV values are calculated proportionally (Table S8). Components with ROAV ≥1 are key aroma contributors, while ROAV between 0.1 and 1 indicates modifying effects. ROAV analysis clarifies aroma profiles and quantifies flavour contributions (Wu et al., 2025).

Based on the ROAV values and aroma attributions, a Sankey diagram was constructed (Fig. 7). The aroma profiles are shaped by aldehydes, ketones, lactones, and esters. Octanal, hexanal, and 2-methylbutanal contribute green/aldehydic notes with roasted nuances; α-ionone and β-ionone impart floral notes (Meléndez-Martínez et al., 2023); dihydroactinidiolide provides fruity notes; ethyl hexanoate and ethyl cinnamate enhance fruity, slightly fatty notes (Ragavan et al., 2022). Processing alters volatile combinations, creating distinct aroma styles.Fig. 7. Sankey Diagram of Aroma Profile Intensity Distribution in Fresh Carrot and Carrot Preserves Processed by Three Different Methods.Fig. 7

FC exhibits a typical carrot flavour with floral, sweet, fruity, and green notes, with slight spicy and earthy nuances (Kjeldsen et al., 2003). CCP and SSCP show reduced volatiles and convergent profiles, with floral and aldehydic notes. NEBCP retains the floral and green notes but shows weakened earthy nuances and more pronounced sweet, fruity, aldehydic, fatty, and roasted characteristics, resulting in a more complex overall aroma.

Compared with FC, CCP, and SSCP, NEBCP showed greater contributions from heat-reaction volatiles and certain aldehydes, particularly those serving as characteristic and markedly enriched aroma contributors in NEBCP, such as 2-amylfuran and 2-methylbutanal (nutty and roasted notes), diacetyl and hydroxyacetone (buttery and caramel-like sweetness), and decanal (citrus- and waxy-like aldehydic notes). Their synergistic effects enhanced the overall aroma complexity of NEBCP, whereas the reduced contributions of these key volatiles in CCP and SSCP led to a more monotonous aroma profile.

Sensory evaluation of the three carrot preserves

3.4

As shown in Fig. 8, notable sensory differences were observed among the three carrot preserves. CCP achieved the highest external colour score (8.35 ± 1.14), owing to the suppression of enzymatic browning and the formation of a glossy, vivid surface under high osmotic pressure (Wang et al., 2023). SSCP ranked second (6.90 ± 1.25), while NEBCP scored lowest (4.35 ± 1.18). For appearance, although all samples maintained their segmented shapes, NEBCP obtained the highest score (14.95 ± 1.23), followed closely by SSCP (14.75 ± 1.33). CCP received the lowest score (13.55 ± 1.32), primarily due to syrup exudation that increased surface stickiness. NEBCP also showed the most favorable texture (19.45 ± 2.50), characterised by a soft, elastic, and slightly glutinous structure with good chewability and no tooth-sticking. CCP exhibited noticeable springiness (16.45 ± 1.57), but residual surface syrup compromised palatability. SSCP was harder and insufficiently elastic (15.25 ± 2.99), leading to poorer chewing tolerance (Chen & Lolivret, 2011). Regarding flavour, NEBCP again performed best (17.85 ± 2.06), displaying a well-balanced sweet–sour profile and a strong Maillard-derived aroma (Ríos-Ríos et al., 2019). CCP ranked second (16.15 ± 2.13), while SSCP scored lowest (12.55 ± 3.05) because of its unbalanced taste and limited aroma complexity. NEBCP also achieved the highest aftertaste score (7.50 ± 1.10). Consequently, overall acceptability was highest for NEBCP (7.10 ± 1.21), moderate for CCP (6.55 ± 1.00), and lowest for SSCP (4.95 ± 1.28). Collectively, NEBCP demonstrated notable advantages in texture, flavour, and aftertaste, with enhanced flavour complexity and coherent texture–flavour interactions, indicating strong sensory competitiveness as a novel preserve. CCP retained traditional strengths in colour and sweetness but required reduced stickiness, whereas SSCP's lower scores across attributes suggest that both its flavour profile and textural design need further optimization.Fig. 8. Sensory evaluation scores of candied carrots prepared using three different processing methods.Radar charts showing the sensory scores of CCP, SSCP, and NEBCP samples, respectively, across the attributes of colour, appearance, texture, flavour, aftertaste, and overall acceptability (A–C).Stacked bar chart illustrating the total sensory scores composed of the individual attribute scores for the three samples (D).Fig. 8

Discussion

4

This study systematically compared the effects of three processes—CCP, SSCP, and NEBCP—on the nutritional components, textural properties, and volatile flavours of carrot preserves. The findings indicate that each processing method significantly impacts the final quality attributes through distinct physical and chemical pathways, resulting in notable differences in nutrient retention, texture formation, and flavour development. From a nutritional standpoint, the CCP process caused significant losses of polyphenols and dietary fibre during blanching and sugar infiltration, although it effectively preserved carotenoids. However, its high sugar content conflicts with current healthy diet trends, necessitating a consideration of nutritional benefits versus potential risks. The SSCP process demonstrated generally low nutrient retention and posed health risks because of its high sodium content. The NEBCP process notably enhances polyphenol content and excels in preserving functional components such as dietary fibre, protein, and carotenoids. Crucially, it avoids the addition of high sugar or salt, aligning with health guidelines to ‘reduce sugar, salt, and fat’ (Bai et al., 2022).

In terms of texture, NEBCP undergoes relaxation and partial disruption of the cell wall and middle lamella under high-temperature and high-humidity processing, which weakens the supporting structure and significantly reduces hardness (Xu et al., 2015). Concurrently, partial degradation and solubilisation of cell wall polysaccharides facilitate the release of pectin originally confined within the wall network into the soluble phase, leading to its accumulation in the system (De Roeck et al., 2007). Under these conditions, soluble pectin absorbs water, swells, and forms a continuous viscoelastic phase through intermolecular interactions, thereby enhancing interfacial adhesion and reshaping the product texture (Moelants et al., 2014). In parallel, the released and enriched reducing sugars and free amino acids undergo Maillard reactions, generating brown polymeric pigments that impart a characteristic dark-brown appearance (Ríos-Ríos et al., 2019). Consistent with these structural changes, NEBCP showed decreased hardness and chewiness, increased adhesiveness, and slightly higher cohesiveness, resulting in a structure that was more deformable yet less prone to disintegration. This textural profile helps to reduce chewing effort, improves swallowing smoothness, and diminishes oral residue. By contrast, in CCP, infiltration of concentrated sugar solution promoted solid enrichment and the formation of a candied matrix, increasing system viscosity and network density. Consequently, hardness and adhesiveness increased concomitantly, while elasticity and cohesiveness were markedly enhanced, indicating stronger structural continuity and deformation recovery. In SSCP, more of the cell wall supporting structure was retained, resulting in relatively high hardness; however, weakened intercellular adhesion led to lower cohesiveness, causing structural loosening during mastication. Its limited elasticity further contributes to a relatively hard and insufficiently elastic texture (Chen & Lolivret, 2011; Nishinari et al., 2024).

Differences in flavour profiles primarily arise from distinct reaction pathways during processing. In NEBCP, the Maillard reaction and its associated Strecker degradation constitute the key routes responsible for generating characteristic flavours that distinguish this product from the other samples. These reactions produce heterocyclic compounds, mainly furans and pyrazines (van Boekel, 2006): 2-ethylfuran, 2-amylfuran, 2-acetylfuran, 5-hydroxymethyl-2-furaldehyde, pyrazine, and 2-methylpyrazine, which impart typical roasted and baked notes. In parallel, Strecker degradation yields aldehydes such as 2-methylbutanal and phenylacetaldehyde, contributing nutty and floral nuances (Kebede et al., 2017). In addition, thermal degradation of carotenoids gives rise to α-ionone, β-ionone, and β-cyclocitral, further reinforcing floral undertones (Meléndez-Martínez et al., 2023). The superposition of these multiple formation pathways results in a richer and more clearly layered aroma profile in NEBCP. By contrast, in SSCP and CCP, blanching and osmotic dehydration favour the leaching and thermal loss of aroma related small molecules into the aqueous phase. During blanching in SSCP and CCP, aroma active compounds readily diffuse into the blanching medium and volatilise upon heating, thereby reducing the retention of volatiles in the matrix (Xie et al., 2025; Xu et al., 2015). In CCP, osmotic dehydration during candying continuously drove free water outward, and part of the volatile flavour compounds migrated with the exuding water phase. In SSCP, salting promoted the loss of water-soluble small molecules via migration with the brine, while sun drying further increased the risk of volatilisation and oxidative degradation, weakening aroma retention. Overall, these aroma losses mainly originated from the coupled effects of hydrothermal treatment and osmotic dehydration, leading to decreased levels of volatile flavour compounds and a gradual simplification of the aroma profile (Nunes, Coimbra, et al., 2008; Tang et al., 2022). Of note, the present evaluation of aroma contribution is based on odour activity values calculated from literature reported thresholds in aqueous solutions, which may not fully capture binding, partitioning, and masking effects in a concentrated, browned, high sugar matrix. Nonetheless, cross-validation with sensory evaluation partially compensates for this limitation (Wang et al., 2025).

The sensory evaluation results are consistent with the physicochemical and volatile analyses. In terms of colour, CCP showed better colour retention because the high osmotic pressure and surface encapsulation by sugar syrup partly inhibit oxidative browning (Wang et al., 2023). NEBCP was darker due to non-enzymatic browning, leading to lower colour preference. Regarding appearance, NEBCP and SSCP maintained their shape well, whereas CCP was limited by high surface stickiness caused by sugar crystallisation. For mouthfeel and flavour, NEBCP performed the best, owing to the lower chewing effort required and more complex reaction derived aroma. In contrast, the palatability of CCP was reduced by stickiness (Chen & Lolivret, 2011), and SSCP received lower scores because of its harder, less elastic texture and weaker taste harmony (Volkow et al., 2011). Overall, NEBCP achieved a better balance between ease of eating and flavour complexity.

This study followed the principle of ‘replicating best commercial status’ by preparing candied, salted sugar-crystallised, and non-enzymatic browning–based carrot preserves under process-optimal, industry-relevant conditions, and comparing their overall quality. Accordingly, inherent differences in raw material geometry and final moisture content were allowed (Waseem et al., 2024; Wu et al., 2025). This increases the practical relevance of the results but also introduces variation in water activity and heat/mass transfer, which may alter apparent solute concentration and reaction kinetics and thus limit attribution to single factors. Future work should perform parameterised studies within each process to clarify how slice thickness/geometry, final moisture or water activity, and heating regime jointly affect flavour and texture, and to define key processing windows for pilot scale application. For NEBCP texture, time resolved sampling during thermal treatment combined with kinetic modelling, TPA, and CT-based microstructural analysis could quantify pore formation and tissue densification. By targeted adjustments of Ca^2+^ level, pH, and amino nitrogen, subsequent studies may track pectin solubilisation, de-esterification, and molecular-weight changes and their coupling with Maillard reactions, thereby elucidating their combined contribution to texture (Masztalerz et al., 2024). Regarding aroma, OAVs based on aqueous thresholds may not fully reflect binding, partitioning, and masking in a high-sugar browned matrix, but cross-validation with sensory data in this study improves confidence in the identified key odourants. Matrix-matched thresholds and aroma recombination/omission tests are needed to further refine these predictions (Wang et al., 2025). Long-time heating of NEBCP may also cause non-monotonic carotenoid changes—initially increased extractability followed by isomerisation and oxidative degradation (Fartoosi et al., 2025). Planned HPLC quantification combined with GC-IMS profiling will clarify the balance between carotenoid retention and characteristic flavour formation. Overall, NEBCP offers a ‘reduced-sugar without loss of aroma’ route for carrot preserves. Future consumer tests in different target groups, linked with instrumental indices, are required to relate physicochemical markers to actual acceptance and guide process Optimization and product development.

Conclusion

5

This study compared CCP, SSCP, and NEBCP in terms of composition, texture, and volatile flavour. Processing dependent differences in hydrothermal treatment, osmotic dehydration, and non-enzymatic browning led to distinct overall quality. CCP showed the highest carotenoid and carbohydrate contents; SSCP had the highest sodium; NEBCP significantly increased total polyphenols and best retained dietary fibre and protein. Texturally, CCP and SSCP formed denser matrices, with SSCP showing the greatest hardness, whereas NEBCP was softer, more elastic, and easier to chew and swallow. Among 115 identified volatiles, NEBCP exhibited the richest profile, driven mainly by Maillard reactions and carotenoid degradation, with β ionone and α ionone as key floral contributors and multiple aldehydes, furans, and ketones providing complementary notes. NEBCP achieved the highest sensory acceptance despite no flavour additives; CCP matched traditional candied products in colour and sweetness but was less preferred; SSCP showed the lowest acceptance due to its hard texture, weaker flavour, and elevated sodium. Overall, NEBCP offers superior nutritional retention, texture, and flavour compared with CCP and SSCP, supports sugar and salt reduction goals, and shows good potential for industrial application and further development of innovative candied fruit products.

CRediT authorship contribution statement

Guowei Li: Writing – review & editing, Writing – original draft, Software, Formal analysis, Data curation. Yu Shi: Writing – review & editing, Methodology, Data curation. Jie Shi: Writing – review & editing, Funding acquisition, Formal analysis. Hongwei Zhang: Project administration. Yan Dong: Investigation. Tianliang Wang: Investigation. Xuesong Ma: Investigation. Yanru Ji: Validation. Lianhui Wei: Validation. Zhenghai Zhang: Validation. Qingli Yang: Visualization. Yueming Wang: Software.

Ethical approval

All participants provided informed consent prior to participating in the sensory evaluation. Measures were taken to protect participant privacy and adhere to institutional guidelines, and all samples were confirmed to be safe for consumption. The study was approved by the Institutional Ethics Committee of the Daqing Branch of the Heilongjiang Academy of Sciences for sensory testing with trained assessors.

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.