Effects of a carboxymethyl fenugreek polysaccharide/sodium carboxymethyl cellulose composite binder on the physicochemical properties and sensory quality of reconstituted tobacco leaves Authors: Huayu Li, Lin Jiang, Hongtao Xu, Shaohui Yan, Ling Chang, Xiaoyu Li, Mengmeng Zu, Xiaopeng Yang, Xiaoming Ji Published: 2026 ABSTRACT Reconstituted tobacco leaves (RTLs) are widely used in heat-not-burn products, but single-component binders can limit their sheet integrity and sensory quality. Here, carboxymethyl fenugreek polysaccharide (CFP) was blended with sodium carboxymethyl cellulose (CMC) at CFP : CMC mass ratios of 0 : 10, 1 : 9, 3 : 7, 5 : 5 and 10 : 0 and evaluated in thick-slurry RTLs. The viscosity, FT-IR spectra, moisture behaviour, contact angle, SEM morphology, tensile and tear properties, thermal behaviour, pyrolysis products, smoke volume and sensory quality were analyzed. The 5 : 5 blend showed the highest adhesive viscosity (912 cP at 1 wt% and 25 °C). Relative to CMC alone, the 5 : 5 blend increased the tensile strength by 11.9%, tear index by 8.74% and bulk specific volume by 9.93%, while the softness value decreased by 26.45% and the sensory score increased from 71.0 to 77.2. Contact angle and moisture measurements indicated reduced hygroscopicity and improved moisture retention. TGA showed comparable thermal degradation profiles, while Py-GC/MS detected 85 products for the 5 : 5 blend compared with 48 for CMC alone. The smoke volume changed only modestly across the puff sequence. These results support CFP/CMC blending as a promising formulation strategy for improving the physical and sensory performance of thick-slurry RTLs. Introduction Heat-not-burn (HNB) tobacco products represent a novel category of tobacco goods. They release nicotine and flavor compounds from tobacco materials through low-temperature heating without combustion.1 Due to their lower heating temperature, reduced emission of harmful substances, such as tar and carbon monoxide, and the ability to satisfy consumers' physiological needs, they have gained widespread popularity among consumers.2–4 Currently, the primary component in HNB products is reconstituted tobacco leaf (RTL). Also known as tobacco flakes, RTL is a sheet or shredded recycled product manufactured by blending tobacco waste powder, binders, and atomizing agents.5,6 Its preparation methods include papermaking, thickening, rolling, and dry processing.7,8 Among RTLs, RTL prepared using the thick-slurry method offers advantages such as high atomizer loading capacity, excellent thickness uniformity, superior characteristic aroma release, simple processing, low energy consumption, and high production efficiency. These qualities make it the most widely used raw material for heated tobacco products both domestically and internationally.9 During the preparation of RTLs using the thick-slurry method, an adhesive is typically added to improve sheet integrity and handling. Sodium carboxymethyl cellulose (CMC) is widely used as an adhesive for RTLs. Due to the abundance of polar groups such as hydroxyl and carboxyl groups in its molecular chain, CMC can form strong van der Waals forces or hydrogen bonds with tobacco pulp and other components in the RTL, thereby improving its physical properties.10 However, sheets prepared with a single adhesive exhibit drawbacks such as low tensile strength and high breakage rates.11 The synergistic effect of blending CMC with other adhesives enhances the strength of the RTL while also improving its sensory properties.6 Existing reinforcement strategies include blending CMC with guar gum or introducing plant fibers, nanocellulose, and carboxymethylated cellulose fibers, which improve tensile performance through hydrogen bonding, fiber entanglement, and physical network formation.2,6,7 Nevertheless, fiber-based approaches require additional pretreatment and may cause aggregation or undesirable woody notes, while conventional guar gum formulations do not utilize chemically functionalized fenugreek-derived binders.2 Recent research has shown that carboxymethylation alters the solubility, molecular structure, moisture-retention capacity, and aroma-generation behavior of fenugreek polysaccharides in cigarettes.12 Fenugreek polysaccharides are highly branched galactomannans. Their molecular structure features an α-(1→4)-β-d-mannan backbone with α-d-galactose side units at the O-6 position. The galactose-to-mannose ratio is approximately 1 : 1, which contributes to high water solubility and viscosity.13,14 Carboxymethyl modification increases the number of polar groups in polysaccharides.15,16 We hypothesize that when CFP is blended with CMC, the mixture can form additional hydrogen bonds with the hydroxyl groups on the surface of RTL fibers, thereby enhancing the interfacial cohesion and mechanical properties of tobacco sheets. This study systematically evaluated CFP/CMC composite adhesives at five blending ratios in thick-slurry RTLs using CMC alone as the principal comparator. We assessed rheological, structural, mechanical, moisture, thermal, pyrolysis and sensory properties within the same formulation system. The objective was to determine whether partial CFP replacement of CMC could improve sheet performance without compromising thermal or sensory properties. The work extends conventional single-binder evaluation by integrating physicochemical performance with volatile-release and sensory measurements. Materials and methods Materials and reagents CMC (food-grade), glycerol, propylene glycol, a 200-mesh tobacco powder, and external fibers were all provided by Henan Cigarette Industrial Reconstituted Tobacco Sheet Co. Carboxymethyl fenugreek polysaccharide (substitution degree = 0.52) was prepared in-house in the laboratory. Preparation of blended adhesives CFP and CMC were mixed at CFP : CMC mass ratios of 0 : 10, 1 : 9, 3 : 7, 5 : 5, and 10 : 0 (w/w) using a magnetic stirrer. After homogenization, the mixtures were freeze-dried to obtain the blended polysaccharide adhesives. Viscosity measurement of blended adhesives The viscosity measurements of the blended adhesives (1 wt%) were performed at 25 °C using a Brookfield DV-2 Pro rotational viscometer. Fourier transform infrared (FT-IR) spectroscopic analysis of blended adhesives For structural characterization, the composite adhesives were mixed with a dry KBr powder (1 : 50, w/w) and pressed into a pellet. FT-IR spectra were then collected on a Nicolet iS10 spectrometer (Thermo Fisher Scientific, USA) over the range of 4000–400 cm−1. Preparation of RTLs using the thick slurry method External fibers (6.9 g) and water (240 g) were combined and mixed thoroughly. Glycerol (8.69 g) and propylene glycol (1.02 g) were then added, and the mixture was blended for 20 min. Subsequently, 1.5 g of CMC, CFP, or the composite adhesive was incorporated, followed by an additional 5 min of blending. The tobacco powder (51.09 g) was then added, and the entire mixture was stirred for 30 min. The MS-ZN320A coating machine was set to a gap of 0.7 mm, a stroke of 350 mm, and a speed of 200 cm min−1. The resulting slurry was cast, dried at 70 °C for 5–7 min, and then carefully detached from the coater using a plastic spatula. Sample preparation and dry-basis moisture content determination for RTL and shredded tobacco samples RTLs prepared with different adhesives were shredded using a shredder and then sieved through a 20-mesh screen to remove debris. Samples were conditioned at 22 °C ± 1 °C and 50% ± 2% relative humidity for 24 h. After equilibration, 5 g of the samples of RTL shreds were weighed and placed in desiccators containing a saturated KCl solution (22 °C ± 1 °C, relative humidity = 84% ± 2%) and a saturated magnesium chloride solution (22 °C ± 1 °C, relative humidity = 32% ± 2%) for moisture absorption and desorption experiments, respectively. Samples were retrieved at regular intervals, weighed and used to calculate the dry-basis moisture content. The initial moisture content of RTL shreds was determined using the oven method as per YC/T31—1996. Contact angle testing of RTLs Contact angle testing was performed according to the TOMMI method with slight modifications; specifically, RTLs with different adhesives added were flattened and fixed on a glass slide.17 The contact angle between deionized water and the RTL was measured using a contact angle tester 5 seconds after the droplet contacted the leaves. The droplet volume was 5 µL. Scanning electron microscopy (SEM) testing of RTLs The samples were frozen in liquid nitrogen for 30 min, and longitudinal cross-sections were prepared with a cutting tool. The cut RTL samples were adhered to conductive adhesive and sprayed with gold, and then, the cross-sectional morphology and structure of the samples were observed using a Zeiss Sigma 360 scanning electron microscope. Physical property testing of RTLs The prepared RTL samples were equilibrated for 24 h at 22 °C ± 1 °C and 50% ± 2% relative humidity. The basis weight and thickness were measured according to GB/T 451.2-2002 and GB/T 451.3-2002, respectively, using an SE051 thickness gauge (L&W, Sweden). The bulk density (V, cm3 g−1) was calculated according to formula (1) as follows:1V = T/W.In the formula, T represents the thickness of the sample, µm, and W represents the mass per unit area of the sample, g m−2. Tensile strength was measured with a KZW-300 tester according to GB/T 12914-2018 (ISO 1924—2:2008, MOD). Specimens were cut into strips with a width of 15 mm and a length of about 250 mm. The gauge length was set to 180 mm. Softness was measured with a TSA tester (Emtec, Germany), and tear resistance was measured with an NJ-08091 tester. Thermogravimetric analysis of RTLs Thermogravimetric analysis was performed using an STA 449 F3 analyser (NETZSCH, Germany). Spectroscopic-grade Al2O3 was used as the reference material. Argon was supplied at 60 mL min−1 and nitrogen at 20 mL min−1. Samples were heated in an Al2O3 crucible from 30 °C to 450 °C at 10 °C min−1. Pyrolysis analysis of RTLs Pyrolysis-GC/MS was performed using a Pyroprobe 5250T unit coupled to an Agilent 7890/5975 GC/MS system. The pyrolysis method was adapted from Gao's approach,18 as detailed below. (1) The pyrolysis probe's initial temperature was 40 °C, which was held for 0 s. It was then heated at a rate of 30 °C s−1 to 250 °C/300 °C/350 °C, holding at each temperature for 15 s. The pyrolysis atmosphere was nitrogen or air. Cleaning occurred at 1200 °C with a rate of 5.0 °C s−1. Drying occurred at 80 °C with a rate of 1.0 °C s−1. The pyrolysis atmosphere consisted of a nitrogen–oxygen mixture. Trap parameters: reaction gas heating rate = 0.63 °C min−1, standby temperature = 50 °C, initial temperature = 50 °C, desorption temperature = 250 °C, heating rate = 2.0 °C min−1, transfer line temperature = 280 °C. (2) Chromatography conditions: Agilent capillary column HP-5MS (30 m × 250 µm × 0.25 µm), injector temperature = 300 °C. Temperature program: 50 °C, equilibrate for 1 min, ramp at 6 °C min−1 until reaching 80 °C, then ramp at 4 °C min−1 until reaching 110 °C, hold for 2 min, finally ramp at 5 °C min−1 until reaching 280 °C, and hold for 2 min. High-purity helium (He, 99.999%) was used as the carrier gas at a 1 mL min−1 flow rate and split ratio 20 : 1. (3) Mass spectrometry conditions: electron ionization (EI) source, electron energy = 70 eV, transfer line temperature = 300 °C, ion source temperature = 230 °C, quadrupole temperature = 150 °C, scan range = 30–550 m/z, solvent delay = 3.9 min. Qualitative identification was performed using the NIST 14 standard library with a match score of ≥80%. Smoke volume test of RTLs RTLs with different formulations were rolled into heated tobacco products and conditioned at 22 °C ± 1 °C and 50% ± 2% relative humidity for 24 h. The smoke volume was measured using a VDM 100 light transmittance tester. Test parameters were set as follows: 5 puffs, a puff volume of 55 mL, puff at 30-s intervals, a diffusion time of 15 s, a diffusion factor of 1, and a bell curve mode for puffing. Sensory evaluation Heated tobacco products were rolled using RTLs prepared with blended adhesives in ratios of 0 : 10, 1 : 9, 3 : 7, 5 : 5, and 10 : 0. These samples were labelled as 1, 2, 3, 4, and 5, respectively. The technical center's evaluation panel at Henan Cigarette Industry Tobacco Sheet Co., Ltd conducted smoking evaluations and comprehensive assessments on these cigarette samples. The scoring criteria are provided in Table 1. Table 1Sensory quality evaluation scorecard of heated tobacco productsIndicator itemFragrance 25Flue gas 30Vigor 10Irritant 10Aftertaste 15Uniformity 10Single-item indicatorIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIHighest score in a single event252015302520108610861512101064IRich, delicate, and harmonious, with distinct characteristicsThe smoke is rich and full-bodied, with no noticeable burning sensation in the mouthModerate intensityNo significant irritationClean, comfortable, and not noticeably dryConsistent throughoutIIAdequate, slightly coarse, and fairly harmonious, with distinct characteristicsThe smoke is rich and full-bodied, with moderate intensity, leaving a slight warmth in the mouthSlightly more vigorous or slightly less vigorousSlightly irritatingCleaner and more comfortable, with a slight dry sensationThere are slight differences between the front and backIIIFaint, somewhat coarse, and lacking harmony, with indistinct featuresThe smoke lacks richness and fullness, tasting thin and watery, with a pronounced burning sensation in the mouthWith great or little enthusiasmExcessive stimulationSignificant dryness and noticeable residueThere is a noticeable difference between the front and back Sensory quality evaluation scorecard of heated tobacco products Results and discussion Viscosity determination of blended adhesives The viscosity of individual and blended adhesive solutions was measured to evaluate the effect of the composition on rheological behavior. The viscosity values for CFP : CMC ratios of 0 : 10, 1 : 9, 3 : 7, 5 : 5 and 10 : 0 were 760, 803, 860, 912 and 647 cP, respectively (Fig. 1). The viscosity increased up to the 5 : 5 blend and then decreased for CFP alone, showing a nonmonotonic composition response. The maximum at 5 : 5 is consistent with complementary intermolecular interactions between CFP and CMC. Viscosity of different adhesives. The enhanced viscosity relative to that of single-component systems can be attributed to the rheological properties of polysaccharides, which are influenced by polarity, hydrogen bonding, the presence of functional groups, molecular weight, and dispersion state.19 Specifically, both CFP and CMC contain polar carboxyl (–COOH) and hydroxyl (–OH) groups that facilitate noncovalent intermolecular interactions, such as hydrogen bonding and chain entanglement, thereby increasing the viscosity and adhesive strength of the blended system.20 Thus, blending CFP with CMC effectively improves the viscosity of the polysaccharide-based adhesive. Hydrogen bonding enhances cohesion in CFP-CMC adhesives according to FT-IR spectroscopy FT-IR spectroscopy is highly effective for identifying various functional groups, glycosidic bonds, and the configuration of sugar rings.21 The formation of the CFP-CMC composite adhesive was confirmed by FT-IR analysis. As shown in Fig. 2, the spectra of CFP, CMC, and the composite adhesives exhibit characteristic absorption bands at approximately 1600 cm−1, 1400 cm−1, and 1300 cm−1, respectively, which are assigned to carboxymethyl groups. Since both CMC and CFP are polysaccharides with similar chemical structures, they exhibit analogous characteristic peaks. Among these, the absorption peak near 3600 cm−1 corresponds to the stretching vibration of the –OH group on the monosaccharide unit.22 The absorption peak at 2900 cm−1 corresponds to the C–H stretching vibration, while the characteristic peak at 897 cm−1 indicates the presence of monosaccharide molecules, which is associated with the presence of galactomannan.23 In the composite polysaccharide system, the absorption peak of CMC at 1631.48 cm−1 shifted to approximately 1600 cm−1 (specifically to 1600.50, 1599.96, and 1600.32 cm−1 for the respective formulations), indicating the formation of intermolecular hydrogen bonds between CFP and the carboxyl groups of CMC. Hydrogen bonding increases intermolecular forces, directly enhancing the cohesion and bonding strength of the adhesive.24 FT-IR spectra of different blended adhesives. Restoring the moisture absorption and retention properties of RTLs To evaluate the moisture absorption and retention properties of RTLs, the samples prepared with different adhesives (Fig. 3) were equilibrated under various relative humidity conditions. As shown in Fig. 4a, in a high-humidity environment of 84% RH, the dry-basis moisture content of all samples increased progressively over 0–100 h. After 100 h, the equilibrium moisture contents were 23.45%, 21.11%, 22.05%, 22.32% and 20.97% for CFP : CMC ratios of 0 : 10, 1 : 9, 3 : 7, 5 : 5 and 10 : 0, respectively. This indicates that blending CFP with CMC reduces the hygroscopicity of RTLs. The dry-basis moisture content of RTLs decreased significantly within 0–12 h, with the rate of decrease gradually slowing after 12 h (Fig. 4b). This indicates that the addition of blended adhesives enhances the moisture retention capacity of RTLs. A possible reason is the increased intermolecular hydrogen bonds between carboxymethyl polysaccharides, which facilitate the binding of large amounts of water and its retention within an extensive network. This process converts free water into bound water that is less prone to evaporation, thereby achieving the goal of maintaining moisture in RTLs.25 The above experiments demonstrate that the addition of blended adhesives affects the moisture absorption and retention properties of RTLs. Among these, the CFP : CMC (5 : 5) blend exhibited both moisture retention and moisture-proofing properties, rendering it suitable for application in RTLs. This formulation was developed to prevent moisture loss in low-humidity environments and excessive moisture absorption under high-humidity conditions, thereby effectively mitigating the impact of moisture on RTLs. Photographs of thick-slurry RTLs prepared with different CFP : CMC mass ratios of (a) 0 : 10, (b) 1 : 9, (c) 3 : 7, (d) 5 : 5, and (e) 10 : 0. Moisture absorption (a) and retention curves (b) of RTLs prepared using different adhesives. Determination of the contact angle of RTLs The contact angle refers to the angle (θ) formed between the tangent line of the gas–liquid interface and the solid–liquid boundary line at the triple point where a gas, liquid, and solid meet. It serves as a measure of the degree of wetting.26 All samples were wettable by water, with contact angles below 90° (Fig. 5). The measured angles were 54.73°, 65.47°, 59.07°, 57.44° and 67.78° for CFP : CMC ratios of 0 : 10, 1 : 9, 3 : 7, 5 : 5 and 10 : 0, respectively. Introducing CFP at a ratio of 1 : 9 increased the angle from 54.73° to 65.47°. The angle then decreased to 59.07° and 57.44° at 3 : 7 and 5 : 5, respectively, before rising to 67.78° for CFP alone. Thus, the contact angle did not increase monotonically with the CFP content. All CFP-containing sheets had larger angles than the CMC control, indicating lower apparent surface wettability under the measurement conditions. This finding aligns with the moisture absorption and retention curve results for RTL. Contact angles of RTLs prepared using CFP/CMC blended adhesives with different mass ratios: (a) CFP : CMC = 0 : 10, (b) CFP : CMC = 1 : 9, (c) CFP : CMC = 3 : 7, (d) CFP : CMC = 5 : 5, (e) CFP : CMC = 10 : 0. SEM analysis of RTLs To investigate the effects of different adhesives on the RTL, SEM was employed to observe the microstructures of various RTL samples. SEM images showed that the longitudinal sections of RTLs prepared with CMC or CFP alone exhibited loose structures with numerous visible gaps (Fig. 6a–e). In contrast, the blended adhesive systems showed more continuous cross-sections and fewer large voids, with the most uniform morphology observed for the intermediate formulations. This aligns with the adhesive viscosity test results: higher viscosity correlates with tighter bonding between the tobacco powder and fibers. Overall, blending CFP with CMC enhances the physical structure of the RTL, mitigating the interlayer loosening and inhomogeneity typically associated with single-adhesive formulations. Longitudinal cross-sections of RTL slices prepared using CFP/CMC blended adhesives at different mass ratios. (a1–a4) CFP : CMC = 0 : 10, (b1–b4) CFP : CMC = 1 : 9, (c1–c4) CFP : CMC = 3 : 7, (d1–d4) CFP : CMC = 5 : 5, (e1–e4) CFP : CMC = 10 : 0. Physical property analysis of RTLs Tensile strength and tear resistance are critical indicators in the production process of the RTL, determining its mechanical strength during manufacturing.27,28 Tear strength is the force required to tear the RTL to a specified length. Relative to CMC alone, the three blended formulations increased the tensile strength by 3.0%, 7.8% and 11.9% and increased the tear index by 2.64%, 4.68% and 8.74%, respectively (Fig. 7). This increase continued as the blending ratio increased, with the maximum tensile strength and tear index achieved at a CFP : CMC ratio of 5 : 5. The tensile strength and tear resistance of the RTL are primarily influenced by inter-fiber bonding strength, and weaker inter-fiber bonding results in poorer cohesion of the RTL. This is likely attributed to the entanglement of adhesive polysaccharide macromolecules, forming a relatively dense physically cross-linked network structure. Concurrently, intra- and inter-molecular hydrogen bonding occurs between adhesive molecules and functional groups such as –COOH and –OH in the tobacco powder. The adhesive molecules act as intermediate linkers, bonding fibers with other components within the RTL to form a compact structure. These physical cross-linking and hydrogen bonding enhance interactions among tobacco flake components, thereby improving mechanical properties. Tensile strength (a) and tear index (b) profiles of RTLs prepared using different adhesives. Bulk density is one of the key physical indicators for RTLs. An optimal bulk density enhances combustion performance and reduces the harm index of RTL products. Softness is closely related to the filling capacity, tactile comfort, and packaging properties of RTL products. Good softness significantly impacts subsequent processes such as slicing and shredding. The lower the softness value, the higher the softness of the RTL. Compared to CMC alone, the three blended adhesives increased the bulk density by 1.65%, 5.94%, and 9.93%, respectively, while reducing softness by 3.53%, 22.17%, and 26.45%, respectively (Fig. 8). This indicates that the addition of blended adhesives improves the bulkiness and softness of RTLs produced by the thick-slurry method. Looseness and softness plots of RTLs prepared using different adhesives. Thermogravimetric analysis of regenerated tobacco leaves Derivative thermogravimetric (DTG) analysis can be used to evaluate the thermal stability and purity of materials, while thermogravimetric (TG) analysis is employed to characterize the residual mass of a sample. The DTG curve represents the first derivative of the TG curve and can be used to assess the rate of change of a sample with temperature.29 As shown in Fig. 9, all five formulations showed three principal mass-loss regions. Our findings are consistent with previous studies.7 The first mass loss stage occurred at approximately 30 °C–130 °C, with a mass loss of about 12–14%. This stage primarily involves the decomposition of thermally unstable and volatile components such as free water, monosaccharides, oligosaccharides, volatile acids, and free amino acids.30,31 The second stage occurred at 127–280 °C, accounting for about 30% mass loss. The peak at 310 °C is primarily attributed to the degradation of cell wall biopolymers such as cellulose, hemicellulose, pectin, and lignin.32 Above 350 °C, the mass-loss rate gradually decreased, indicating that most substances have been fully decomposed or released by this temperature. The minor mass loss observed may result from further cracking and carbonization of residual, non-pyrolyzed materials. Overall, in the 30–450 °C range, the blended and single adhesive samples showed comparable thermal degradation profiles, thereby minimizing potential sensory discrepancies during smoking that might otherwise arise from differences in adhesive decomposition temperatures. Thermogravimetric (a) and differential thermogravimetric (b) curves of RTLs prepared using different adhesives. Pyrolysis product analysis of RTLs Pyrolysis tests of RTLs with different adhesives were performed using Py-GC-MS at 350 °C to simulate the thermal decomposition conditions of heated tobacco products. As presented in Fig. 10, Py-GC/MS detected 48, 63, 79, 85 and 51 products for CFP : CMC ratios of 0 : 10, 1 : 9, 3 : 7, 5 : 5 and 10 : 0, respectively. The 5 : 5 blend produced the largest number of detected products. Based on quantitative analysis, the pyrolysis products were classified into the following categories of thermal decomposition products: (1) ketones, aldehydes, and esters, including hydroxyacetone, methylcyclopent-2,3-dienone, 2,3-pentanedione, 5-methylfuranaldehyde, phenylacetaldehyde, and methyl 2-furanoate; aldehydes, ketones, and esters typically impart sweet, aromatic, and fruity notes, often associated with fresh and caramel flavors.33,34 Alcohol compounds can enhance the floral and fruity aromas of tobacco leaves.35 (2) Acids and alcohols, including nonanoic acid, 3-methylpentanoic acid, propylene glycol, and glycerol. (3) Heterocyclic compounds, including acetylfuran, 2,3,5-trimethylpyrazine, and 3-methylindole. Furan compounds primarily contribute baked and caramelized sweet aromas.36 Pyrazine compounds enhance the nutty and roasted aromas in tobacco leaves. (4) Phenolic and olefinic compounds such as guaiacol, 4-ethylguaiacol, neophytadiene, and limonene. Neopentadiene serves as a key precursor for developing a fresh, clean aroma and also undergoes further conversion into smaller-molecule fragrance components.37,38 (5) Nicotine derivatives including nicotine, mesalamine, and nicotinamide. Additionally, Fig. 10 indicates that RTLs treated with blended adhesives released higher levels of ketone compounds (17.46–42.40%), including hydroxyacetone (8.72–19.23%) and 2,3-butanedione (0.51–10.32%) while simultaneously reducing nicotine release. This demonstrates a slight reduction in nicotine release. These results indicate that the combination of CFP and CMC can release more aromatic compounds and reduce the nicotine content in RTLs, thereby supplementing the overall aroma and improving the blandness caused by the addition of a single binder. Types and relative contents of the thermal pyrolysis products from tobacco flakes prepared using blended adhesives. Determination of the smoke volume from RTLs To evaluate the performance of different adhesive additives in heated tobacco products, the smoke volume was measured using a heated tobacco product smoke transmittance tester. The smoke volume gradually increased across the puff sequence and peaked at the fourth and fifth puffs (Fig. 11). Blended formulations showed transmittance values of approximately 80% during the first two puffs, with only small differences from the CMC control during the final puffs. CFP/CMC blending, therefore, had a limited effect on the smoke volume under the tested conditions. Smoke volume profiles of RTLs prepared using different adhesives. Sensory evaluation To further compare the effects of different adhesives on the sensory quality of RTLs, evaluators conducted sensory assessments of heated tobacco products rolled with five types of RTLs. The sensory scores were 71.0, 73.9, 75.2, 77.2 and 73.7 for samples 1–5, respectively (Table 2). Samples 2–4 formulated with blended adhesives achieved higher sensory scores, indicating that the CMC/CFP composite effectively enhances the sensory quality of RTLs as well as the overall sensory comfort of heated tobacco products. Notably, the sample with a CFP : CMC mass ratio of 5 : 5 exhibited optimal sensory performance, which is consistent with the pyrolysis results. The positive correlation between the yield of pyrolytic aroma compounds and sensory scores further highlights the practical potential of this formulation. Table 2Sensory quality evaluation score sheet of the heated tobacco productsCigarette numberFragrance 25Flue gas 30Vigor 10Irritant 10Aftertaste 15Uniformity 10TotalIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII252015302520108610861512101064120206811671221.621.26.2810.76.273.9321.522.46.4810.86.175.2423.422.76.57.810.56.377.2521.221.56.37.910.66.273.7 Sensory quality evaluation score sheet of the heated tobacco products Conclusions In summary, the composite adhesive prepared by blending CFP with CMC exhibits outstanding application potential in the production of thick slurry-processed RTLs. A CFP-to-CMC mass ratio of 5 : 5 provides the optimal balance of overall performance. Adhesive property tests demonstrate that blending effectively enhances both the cohesion and bond strength of the adhesive. Physicochemical tests indicate that the blended adhesive significantly regulates moisture sorption and retention and optimizes the sheet structure. It also simultaneously improves mechanical properties, including tensile strength, tear strength, bulk density, and flexibility. Meanwhile, the blended adhesive exerts no obvious adverse effect on the thermal stability of the RTL. It not only prevents sensory deterioration induced by thermal decomposition, but also facilitates the release of aroma-generating pyrolysis products and reduces the contents of nicotine-related components. Sensory evaluation results indicate that the blended adhesive greatly improves the overall sensory quality of the RTL. It also does not interfere with smoke delivery, which supports its use in manufacturing high-quality RTLs. Therefore, the CMC/CFP blended adhesive holds promising prospects for application in the slurry-processed RTL sector. Author contributions Huayu Li: conceptualization, visualization, writing – original draft, review, and editing. Lin Jiang: conceptualization, visualization, writing – original draft, review, and editing. Hongtao Xu: conceptualization, writing – original draft, review, and editing. Shaohui Yan: methodology, resources, supervision, and validation. Ling Chang: data curation, formal analysis, and visualization. Xiaoyu Li: investigation, methodology, and software. Mengmeng Zu: conceptualization, visualization, and supervision. Xiaopeng Yang: conceptualization, visualization, and supervision. Xiaoming Ji: conceptualization, visualization, and supervision. Conflicts of interest The authors affirm that they have no known financial or interpersonal conflicts that would have appeared to have an impact on the research presented in this study.