Abstract
Introduction
Heated cigarettes are a revolutionary tobacco product that uses a specialized heating device to heat treated tobacco material to temperatures below 300–500 °C, releasing aerosols for smoking (Petrache and de Boer, 2021; Ruprecht et al., 2017; Yi et al., 2001). The development of heated cigarettes dates back to the 1980s, but it is only in recent years that they have begun to gain popularity worldwide. At present, there are several major brands in the market, such as IQOS, Glo, Ploom Tech, etc. Among them, IQOS is the earliest brand to enter the Japanese market and occupy a large share, which occupies 15.8% of the Japanese tobacco market and 80% of the heated cigarette market. This method only heats the tobacco material, rather than burning it, significantly reducing the harmful components produced during high-temperature combustion of tobacco and decreasing the amount of sidestream and ambient smoke released (Borgerding et al., 1998; Cobb et al., 2021; Farsalinos et al., 2018; Gasparyan et al., 2018; Kim and An, 2020; Szparaga et al., 2021). Cigarettes are heated instead of burned, leading to a substantial decrease in the release of harmful components (Lu et al., 2022; Petrache and de Boer, 2021). Despite the high temperature of the smoke before it reaches the filter, caused by the heating method and product structure, the user experience is still affected (Tabuchi et al., 2016). Therefore, the development of adsorption cooling materials to reduce the temperature of heating cigarette smoke has gradually become a research priority.
Reducing the temperature of flue gas is crucial to improving the comfort of using heated cigarettes (Crenshaw et al., 2016; Horinouchi and Miwa, 2021). There are two commonly used cooling methods: designing a cooling structure (Lu et al., 2022) and adding cooling materials (Berthet et al., 2022). The cooling structure increases the contact area between flue gas and ambient air, promoting heat exchange (Goniewicz, 2019). Examples of cooling structures include hollow mouth sticks, double-layer filter sticks, cavity cooling structures, venturi cooling structures, and shunt cooling structures (Laking, 2019). For example, a multi-segment cavity structure can be used in cigarette construction in conjunction with highly permeable or perforated molding paper to increase the contact between ambient air and high temperature smoke (Martell, 1974). Containing multiple longitudinal straight passages can be regarded as a multiple cavity cooling structure, and the multiple cavities can be used in conjunction with the cavities (Muggli et al., 2008). At the same time, airflow baffles can be added to the cavity to enhance the collision and heat loss of the flue gas, so as to achieve the purpose of reducing the flue gas temperature (Malone, 2002). Or in the baffle plate set throttle flow holes, high temperature flue gas through the cavity, the formation of buffer diffusion and throttling pressure drop in each region, can realize the purpose of segmented throttling cooling. Cooling the flue gas through structural design or perforation of the filter nozzle has less impact on the concentration of the flue gas (Chen et al., 2021). While some cooling structures may be complex and restricted by patents, they are still a viable option to consider.
On the other hand, cooling materials absorb heat from the flue gas, causing its temperature to drop (Ma et al., 2022). This can be achieved through the use of phase change materials (Yang et al., 2020; Zervas et al., 2018), water, glycerin, or other temperature-reducing materials or structures (Bero, 2003). For phase change materials, PLA is the most common adsorption cooling material in the tobacco field, but its phase change temperature is high (60–100 °C), so some researchers have prepared polyethylene glycol/polylactic acid (PEG/PLA) composite films by using electrospinning technology to reduce the glass transition temperature of composite films (Jia et al., 2016; Ke et al., 2021). Crystalline hydrated salts such as sodium sulfate decahydrate can also be used to cool down the smoke of heated cigarettes by impregnating them with expanded graphite, drying and crystallizing them, and then adding them to the end of the filament of the filter rods and/or near the middle of the filter rods, or mixing them into the entire fiber bundle of the filter rods, or spraying or soaking them directly on the filter rods of the cigarette filters after they are heated up and melted down, which can reduce the temperature of the cigarette gas by 10 °C. However, phase change materials are greatly affected by their phase change temperature and environment (Li, Zhang & Monteiro, 2020). In humid or high temperature environments, the cooling materials will gradually fail (Biswas et al., 2022), resulting in poor cooling effect (Prueitt et al., 2009). Liquid water is a kind of natural and efficient heat storage material, which mainly utilizes the evaporation of water to absorb heat to achieve the purpose of cooling, liquid water can be added to cigarettes in a simple way, for example, by adding water-containing cooling segments to filters, or by means of applying a layer of water film to the outer layer of the cigarettes (Chen et al., 2022; Olonoff et al., 2019), which can be encapsulated and then be used as a cooling material in the heated cigarettes (Li, 2018; Yang et al., 2025). Liquid water as a cooling material for cigarettes also has some disadvantages, liquid water will increase the humidity of cigarettes, leading to softening, deterioration, and easy breakage of cigarettes. Meanwhile, liquid water evaporates as the cigarette is heated, producing water vapor (Boissiere et al., 2020; Chen et al., 2023), which increases the amount of gas inhaled by the smoker and may cause irritation or injury to the respiratory system.
Embossed rod powders are an excellent choice for reducing the cost of heated cigarettes because of their wide availability, as well as the fact that they do not easily deteriorate after preparation. This work researches new selective adsorption cooling materials that use (Jia et al., 2022) as the matrix. The study investigates the cooling effect, cooling mechanism, and application in heating cigarette products based on the materials’ selective adsorption characteristics. The findings provide technical support for enriching the cooling mechanism of smoke and expanding the application of cooling materials.
Experiment
Chemicals and materials
Embossed rod powder was purchased from Zongrun Mining Co., Ltd (Hebei, China). Glass powder was purchased from Lianyungang Huifu Nano New Material Co., Ltd (Jiangsu, China). Activated carbon powder, deionized water, and polyvinyl alcohol were purchased from Sinopharm Chemical Reagent Co., Ltd (Shanghai, China). The wood flour was purchased from Urumqi Xinlan Wood Industry Co., Ltd (Xinjiang, China). The alumina fibers were purchased from Zhejiang Huayi New Material Co., Ltd (Zhejiang, China).
Sample preparation
Selective adsorption cooling materials were prepared by mixed roasting method. Embossed rod powder, glass powder, activated carbon powder, wood powder, alumina fiber and polyvinyl alcohol solution were mixed according to the proportions shown in Table 1. Then deionized water of equal mass of the solid powder was added to stir the mixture, which was shade-dried at ambient temperature. At last, it was placed in a high temperature furnace for roasting, and the temperature and time of roasting were also as shown in Table 1. The six sets of samples prepared were named according to Table 1.
The dosing and experimental conditions for samples.
Performance test of selective adsorption cooling materials
The test sample was made of Soya cigarettes (Yoga) as a control sample. Preparation method of the test sample: Cut the vinegar fiber section of the cigarettes and the firmware section of the cigarettes of SU cigarettes (Yoga), take out the sieve of the end of the firmware with tweezers, fill the test particles into the firmware, and then reseal the vinegar fiber section of the cigarettes and the firmware section of the cigarettes with adhesive paper.
Vaping method for heating cigarette using IROD vaping machine: the heated cigarette sample was equilibrated at temperature (22 ± 1) °C and relative humidity (50 ± 3)% for 48 h. Cigarette samples were added into the smoking device for smoking, with a smoking volume of 55 mL, a smoking time of 2 s, and a smoking interval of 30 s. After the cigarette was smoked, the filter obtained was placed in a 100 mL triangular vial, and 1/4 of the blank filter was wiped to wipe the trap, and the wiped blank filter was put into the triangular vial. The filter was extracted with 50 mL of methanol solution containing the internal standard for 30 min, and the extract was filtered through a membrane for the detection of glycerol and nicotine release. The heated cigarette was disassembled, and the particles of the cooled fixtures were removed, and the particles of each 5 cigarettes were combined into a triangular vial, and the particles were extracted by ultrasonic extraction with 25 mL of methanol solution containing the quinoline internal standard for 30 min, and then continued to be extracted by vibration for another 30 min, and then the extract was filtered through a membrane to be detected for the glycerol and nicotine release of the particle samples.
The detection of glycerol and nicotine content was performed by GC-MS: column: HP-Innowax (30 m*0.25 mm, 0.25 μm); carrier gas: high purity helium; carrier gas flow rate: 1.0 mL/min; injection volume: 1 μL; inlet temperature: 240 °C; shunt ratio: 20:1; program temperature increased from 80 °C to 230 °C at 8 °C/min. Moisture detection method: GB/T 23203.1-2013 (Determination of Moisture in Cigarette Total Particulate Phase Part 1: Gas Chromatography) was used to detect moisture in the test samples.
Characterization of selective adsorption cooling materials
The materials under investigation was characterized by many techniques as detailed below. X-ray diffraction (XRD) patterns were collected in a 2θ range of 10–85° on an X-ray diffractometer (Smartlab TM 3Kw, Rigaku, Japan) at 40 kV and 40 mA. The morphology of the catalysts was investigated using a scanning electron microscope (SEM, JEOL, JEM-2010UHR) and high resolution transmission electron microscope (HR-TEM, JEM/2010). The specific surface area and porous parameters of the samples were measured by N2 physisorption at −163 °C using a surface-area analyzer (Micromeritics, ASAP 2020M). All the samples were degassed at 350 °C under vacuum for 3 h prior to the analysis.
Results and discussion
Cooling performance analysis
Figure 1 showed the maximum flue gas temperature of different samples during the suction process and the variation of the flue gas temperature profile of a typical sample, respectively. As shown in the figure, with the increase in the number of suction ports, the flue gas temperature showed an increase and then a decrease. The maximum flue gas temperature was generally reached during the second bite. Compared with the Yoga control sample (whose maximum flue gas temperature was 50.95 °C), the best adsorption cooling effect was sample 03 (temperature drop of 5.79 °C), followed by 04, 05, and 06 (temperature drop in the range of 2.03–2.92 °C). The cooling effect of samples 01 and 02 was close to that of the Yoga control sample.

(a) the maximum flue gas temperature of different samples and (b) flue gas temperature profile for sample 03.
Figures 2 and 3 showed the contents of each component of the flue gas adsorbed by the adsorption cooling materials and the contents of the components in the flue gas after cooling, respectively. It was found that for nicotine and moisture, the adsorption amount of sample 01–06 was much lower than that of the control (Yoga) particles, and the adsorption amounts of nicotine and moisture were reduced by 55%∼73% and 85%∼90%, respectively. The reduction of nicotine adsorption could promote the amount of nicotine that enters into the human oral cavity after cooling down of the smoke, which was beneficial to the enhancement of taste. For glycerin, the smaller amount of glycerin adsorbed by the samples, the more favorable the effect of cigarette smoking. Except for sample 03, which was at the same level with the control sample particles, the adsorption amount of glycerol in the smoke by the sample particles was larger than that of the control sample particles. This may be due to the rich hydroxyl groups on the surface of the embossed rod powder, while the control (Yoga) was a plant particle with a lower number of hydroxyl groups on the surface than that of the embossed rod powder, and the glycerol and nicotine were hydrophilic and lipophilic, respectively. As a result, the sample particles showed increased adsorption of glycerol but decreased adsorption of nicotine compared to the control (Yoga). In addition, it was also found that the inner wall of the cooling chamber had obvious water droplets when using the embossed rod powder material as the adsorption cooling material compared to the control group (Yoga), while the control group (Yoga) was relatively dry, which could be attributed to the fact that the water molecules in the flue gas condensed rapidly to form water droplets due to the high thermal conductivity of embossed rod powder material and were not adsorbed on the surface of the adsorption cooling material.

Adsorption capacity of glycerol, nicotine and moisture of different samples.

Glycerol, nicotine and moisture content in flue gas for different samples.
Corresponding to the trend of the amount of adsorbed nicotine in the samples, the amount of nicotine in the smoke of samples 01–06 were significantly higher than the control (Yoga). In addition, the amount of glycerol in the smoke of samples 03 and 06 was significantly higher than slightly higher than the control (Yoga), while the rest of the samples were lower than the control (Yoga). The moisture in the smoke of samples 01, 04 and 06 was significantly higher than that of the control group (Yoga), sample 03 was close to that of the control group (Yoga), and samples 02 and 05 were lower than that of the control group (Yoga). Considering the cooling effect and the effects of three factors, glycerol, nicotine and moisture, sample 03 was preferred as the best sample.
Physical property analysis
Figure 4 displayed XRD data plots of samples 01∼06. The peaks at 20.84°, 26.62°, 36.52°, and 50.11° corresponded to the (100), (101), (110), and (−212) crystal planes of α-SiO2 (PDF 77–1060), respectively. Additionally, the peaks at 21.76°, 28.21°, 30.96°, and 35.91° corresponded to the (101), (111), (102), and (112) of grainstone-like SiO2 (PDF82–0512), respectively (Borouni et al., 2018). The alpha-SiO2 samples from 01 to 06 exhibited the highest percentage of (101), (111), (102), and (112) crystalline facets. The sample from 02 had the highest percentage of alpha-SiO2 in the material. Sample 03 produced the highest percentage of grainstone-like SiO2 and α-SiO2 compared to the other samples (Chen & Liu, 2022). Additionally, all samples (03, 04, 05, and 06) roasted at 1200 °C showed a larger percentage of α-SiO2 in their XRD plots compared to samples 01 and 02 (Narváez et al., 2003). However, the anatase SiO2 content in the material produced by roasting at 1100 °C for samples 01 and 02 was significantly higher than that of the material produced at 1200 °C. Therefore, the high temperature condition caused the transformation of the crystalline form of silica from graptolite SiO2 to α-SiO2.

XRD patterns of different samples.
Table 2 showed the porous properties data for six sample groups.The pore diameter of the samples decreases with the addition of activated charcoal powder and wood powder, while the specific surface area increases. This is because the activated charcoal powder and wood powder are micron-sized, and their positive effect on the nanopore channels is not observed. Comparison of samples 02 and 03 reveals that increasing the roasting temperature from 1100 °C to 1200 °C results in an increase in specific surface area from 0.4m²/g to 10.5m²/g, as well as an increase in pore diameter from 3.43 nm to 3.83 nm. The reason for the 0.83 nm size is likely due to the roasting temperature of 1100 °C, which causes the gradual formation of ceramics in the samples (Yang et al., 2024). Higher roasting temperatures result in a more complete development of ceramic grains, leading to a decrease in the number of micron and submicron pores and an increase in the number of nanopores in the samples. The comparison between samples 03 and 06 indicates that a decrease in the mass of glass powder and alumina fibres results in a decrease in the pore surface area of the samples, while the pore diameter slightly increases. Furthermore, the comparison of samples 03–06 demonstrates that the amount of glass powder, alumina fibre, and polyvinyl alcohol can be adjusted to regulate the pore surface area, pore volume, and pore diameter of the adsorption cooling materials.
Porous properties of different samples.
Figure 5 shows N2 adsorption desorption and pore size distribution curves of different cooling material. The adsorption-desorption curves of samples 03 and 06 were separated at medium pressure. Based on the classification of physical adsorption isotherms, both samples belonged to Langmuir type IV physical adsorption, which is typical of mesoporous materials. The pore size distribution curves of both samples belonged to the H3-type hysteresis loops, which further confirmed their classification as mesoporous materials.

(a) N2 adsorption desorption and (b) pore size distribution curves of different cooling material.
Comparison of Figure 6(a) and (c), it could be observed that the surface of the 03 sample had a large number of block structure. The loose structure and a large number of blocks between the hollow structure caused a larger specific surface area (Algamdi et al., 2024; Sulu et al., 2024). Meanwhile, it could make the flue gas enter the sample pore inside the release of heat, to achieve the cooling effect. On the other hand, the 06 sample was formed with micron-sized lumpy particles of varying sizes due to the lack of glass powder in the preparation process, and the number of pores was relatively small, which was not conducive to the entry of flue gases into the interior of the sample, and thus the cooling effect was poor (Xing et al., 2023).

SEM patterns of different samples: (a) and (b) 03, (c) and (d) 06.
Figure 7 showed the high-resolution transmission electron micrographs of samples 03 and 06, respectively. It could be found that there was no obvious difference in the nanoscale morphology of the two groups of samples, which were both composed of irregular particles of about 30–100 nm. It is mainly due to the fact that the main raw material components of the two groups of samples were all embossed rod powder and they were all made by roasting at 1200 °C. However, the lattice stripes of sample 06 were more clearly spaced. This may be due to the fact that sample 06 had fewer types of components. In contrast to the XRD data, the XRD characteristic diffraction peaks of sample 03 had lower intensities, which similarly suggested that sample 03 was less crystallized, resulting in a larger specific surface area.

HR-TEM patterns of different samples: (a)–(c) 03, (d)–(f) 06.
Molecular simulation analysis
The geometric model of the smoke branch flow field was established, as shown in Figure 8. In the modeling process, the hindering effect of tobacco on the flow of smoke was not considered, and only the cooling effect of the particle stacking section on the heated smoke was considered. Therefore, the cigarette tobacco segment was simplified into a straight pipe, and the particle stacking layer was a porous structure (Yang et al., 2022). Although the cooling particles were irregular particles, considering that the particles were all ground and pass through a 14–19 mesh sieve, and in order to simplify the flow field model, the cooling particles were therefore simplified into spherical particles. It was divided into 12 layers, and arranged in accordance with a homogeneous disaggregation.

Geometric modeling of cigarettes.
The model was set to unsteady state calculations. The turbulence model was adopted as the RNG k-epsilong model, with both the viscous diffusion model and eddy model turned on. The near-wall region was modeled with the enhanced wall function model, with the pressure gradient effect option and the thermal effect option turned on. The wall surface of the smoke branch adopts the wood material property that comes with FLUENT, the pressure and velocity coupling adopts SIMPLE algorithm, the discretization format adopts Least Squares Cell Based, the momentum equation, the turbulent kinetic energy equation, and the turbulent kinetic energy dissipation rate equation adopts the first-order windward format, and the energy equation adopts the second-order windward format.
The temperature distribution at different moments was shown in Figure 9. As can be seen from Figure 9, the high-temperature flue gas spread along the filament section to one end of the filter. When the high-temperature flue gas flowed to the particle layer, the flue gas and the particles undergo a strong heat exchange, resulting in a rapid decrease in the flue gas temperature. The flue gas was closer to the branch wall, the temperature was lower. It indicated that the nanoparticles had a high cooling effect on the flue gas. As well known, the dissipation of heat in nature was mainly by convection, heat conduction and heat radiation. For the heat dissipation between the heated flue gas and the particles, the main ways were heat conduction and heat convection. Because the particles had high specific surface area and rich micro-pore structure, the flue gas passed through the particles in contact with their micro-pore, thus increasing the heat conduction area of the high temperature flue gas, and ultimately enhancing the heat conduction effect of the high temperature flue gas. In addition, in the nanoparticle stacked layer, a microchannel with special topological configuration was enclosed between particles and particles. When the high-temperature flue gas flowed through these microchannels, strong turbulence occurs, thus increasing the convective heat transfer effect. Therefore, the particles played an active role in the dissipation of flue gas heat.

Flue gas temperature distribution at different moments (a) 0.4 s. (b) 1.0 s. (c) 1.5 s.
The velocity distribution cloud was shown in Figure 10. As could be seen from Figure 10, the center velocity in the filament section was significantly higher than that in the wall region, which was because the fluid was affected by viscous forces in the wall region and a boundary layer was formed. The existence of the boundary layer reduced the flow velocity, which reduced the convective heat transfer effect between the fluid and the outside world. In the particle layer section, influenced by the microchannels between the particles, a high velocity region appears and a velocity pulsation with periodic characteristics was formed. This velocity pulsation destroyed the boundary layer, thus increasing the convective heat transfer between the fluid and the particles.

Characterization of the velocity distribution of flue gas in the smoke branch.
Through the numerical simulation analysis of the flow characteristics of the flue gas inside the cigarette containing the particle layer, it could be seen that increasing the filling amount of the particle layer could increase the heat conduction effect and heat convection effect between the flue gas and the particles, which can help to reduce the temperature of the flue gas. Meanwhile, the increase in the specific surface area of the particle layer was conducive to the enhancement of the heat conduction effect. It was conducive to the deposition and decomposition of hazardous substances in the flue gas, reducing the harmful effects of hazardous substances on the human body. Therefore, in order to improve the cooling effect of the particles, the preparation conditions could be optimized to improve the specific surface area and expand the number of microscopic pores, thereby improving the contact and heat transfer effect between the flue gas in the central region of the smoke branch and the cooling particles.
Conclusions
In summary, a embossed rod powder based adsorption cooling material was developed to address the problem of overheating of heated cigarette smoke affecting the taste, and the optimal preparation conditions and adsorption cooling effect were investigated. The mechanism of the adsorption cooling material was verified using molecular simulation technology. It was found that the optimal preparation conditions of the adsorption cooling material were: 3 g of embossed rod powder, 1.5 g of glass powder, 1.5 g of alumina fiber, and 1.20 g of polyvinyl alcohol mixed and roasted at 1200 °C for 2 h. Compared with the commercial Yoga cooling material, the embossed rod powder based adsorption cooling material was able to lower the maximum flue gas temperature to 45.16 °C. Meanwhile, the adsorption of nicotine by the cooling material was reduced, and the content of glycerol and nicotine in the cooled smoke was increased, which effectively promoted the taste of the smoke. In addition, the molecular simulation results found that the higher specific surface area and rich microscopic pore structure could increase the heat conduction area of high temperature flue gas, thus enhancing the heat transfer effect of flue gas. Finally, the periodic characteristic velocity pulsation destroyed the boundary layer formed by the fluid in the wall region, which increased the convective heat transfer effect between the fluid and the particles. The optimum sample 03 had the largest specific surface area and the lowest crystallinity, and thus had the best cooling effect. Future cooling adsorption materials in the field of heated cigarettes are mainly developed in the direction of their production cost, lowering the temperature of the smoke and reducing the adsorption of nicotine, and this work can provide a certain reference significance for the development of cooling materials for heated cigarettes.
