Abstract:
Objective To clarify the material basis of sourness in heated tobacco and the "tobacco raw material-chemical components-sensory" correlation mechanism.
Method 57 representative tobacco leaves of heated tobacco products were selected. 39 chemical compositions of the raw tobacco leaves was determined using methods such as continuous flow analysis, UPLC, and GC-MS; heated tobacco samples were prepared using a roller-pressing method, and 54 chemical composition of the smoke was qualitatively and quantitatively analyzed using central-cut two-dimensional gas chromatography-mass spectrometry (GC-GC/MS). The sourness was evaluated using a 0-5 point scale. Kruskal-Wallis tests, Spearman's correlation analysis, Random forest model, and cluster analysis were employed to identify the key chemical components associated with sourness.
Result 57 heated tobacco product samples were classified into three groups based on sensory evaluation of sourness: slightly sour (38), moderately sour (10), and sour (9). Analysis of differences indicated that 47 of the 93 detected components showed significant differences between groups (P < 0.05). Correlation analysis indicated that components such as chlorogenic acid (Rho = 0.84), sucrose (Rho = 0.82), rutin (Rho = 0.81), and acetic acid (Rho = 0.56) showed a highly significant positive correlation with the sourness score; components such as, isovaleramide (Rho = -0.77), oxalic acid (Rho = -0.72), and nicotine (Rho = -0.50) showed a significant negative correlation with the sourness score. Cluster analysis results indicate that the 57 samples can be divided into two clusters. The right-hand cluster (37 samples), consisting primarily of flue-cured tobacco, is enriched with components positively correlated with sourness, such as chlorogenic acid and sucrose. These samples exhibit a sensory profile of "sour" or "slightly sour", forming the material basis for the strong sourness profile of heated tobacco products; the left cluster (20 samples), composed of air-cured tobaccos such as Burley and cigar tobacco, was highly enriched in negatively correlated components such as nitrogen-containing compounds and was perceived as "slightly sour". The classification results objectively reflect how typical metabolic differences shape sensory profiles: flue-cured tobacco dominated by carbon metabolism extensively accumulates sugars and polyphenols, laying the material basis for strong sourness, whereas air-cured tobacco dominated by nitrogen metabolism is enriched in alkaline substances such as alkaloids, significantly suppressing sourness expression.
Conclusion The sourness of heated tobacco products is highly correlated with the chemical components of both raw materials and smoke. This study identified polyphenols and sugars represented by chlorogenic acid, sucrose, and acetic acid as the key positive acid-contributing components driving the expression of sourness in heated tobacco products, while nitrogen-containing compounds and specific dicarboxylic acids represented by isovaleramide, oxalic acid, and nicotine were identified as the negative regulatory components that inhibit sourness. The synergistic accumulation of carbon metabolites and the appropriate reduction of nitrogenous alkaline substances constitute the shared material basis for forming the strong sourness profile in heated tobacco products, providing a scientific basis for raw material screening and sensory regulation.