Chemists in China have built a new analytical tool that can fish out trace amounts of flavor-related compounds from Baijiu, the Chinese liquor whose mellow character develops over years of storage. The resulting magnetic solid-phase extraction method, coupled with gas chromatography–mass spectrometry, achieved more than 90 percent of equilibrium adsorption within one minute and detected the targets at concentrations as low as 9.4 micrograms per liter.
The compounds in question matter because acetalization is one of the key chemical reactions that transforms freshly distilled liquor into a smoother product. In the ethanol-rich environment of Baijiu, aldehydes—which often contribute pungent or unpleasant notes to fresh spirit—react with ethanol through an acid-catalyzed equilibrium to form diethyl acetals. Compared with their aldehyde precursors, these acetals are chemically more stable and carry softer fruity or sweet aroma characteristics. Monitoring them may therefore offer chemical information about aldehyde conversion, aroma mellowing, and the emergence of mature flavor in light-aroma Baijiu, making them candidate aging markers.
Previous studies support that idea. A metabolomics-based investigation of Chinese Highland Qingke Baijiu identified 15 candidate aging markers, seven of which were acetals; among the nine markers that increased in the 6-to-11-month aging group, eight were acetals. In roasted-sesame-like Baijiu, 1,1-diethoxymethane has been reported as an age marker, while hexanal diethyl acetal was identified as a key age-related differential compound in Feng-flavored Baijiu samples of different storage ages. Acetals have also been examined as potential age-related marker groups in Madeira wine and other distilled spirits. Yet systematic, targeted quantification of structurally related diethyl acetals across different aging periods has remained scarce for light-aroma Baijiu, particularly Fenjiu.
Existing analytical approaches—liquid-liquid extraction, solid-phase extraction, headspace solid-phase microextraction, followed by GC–MS or comprehensive two-dimensional GC–MS—are powerful for broad volatile profiling but face limitations for trace acetals. Target analytes are present at low abundance, coexist with numerous esters, alcohols, and aldehydes, and broad-spectrum methods lack structural selectivity for a specific compound class. Multi-step extraction, long equilibration times, and solvent consumption further reduce efficiency when several trace acetals must be measured simultaneously. The researchers therefore sought a pretreatment strategy combining rapid separation, selective enrichment, and structural recognition.
Their solution began in the computer rather than the laboratory. The team selected phenylacetaldehyde diethyl acetal (PADA) as the template molecule because all four target analytes share the diethyl acetal moiety, providing a common structural basis for recognition. Ten candidate functional monomers were screened using density functional theory calculations at the B3LYP-D3(BJ)/6-31+G(d,p) level, with high-accuracy single-point energies at M06-2×/def2-TZVP and basis set superposition error correction. Allylglycine (AG) showed a strong calculated interaction energy of −33.63 kcal/mol with the template—only slightly weaker than the best performer, NAT, at −36.60 kcal/mol—but offered a more balanced distribution of hydrogen-bond donor and acceptor sites, with electron-rich carboxyl oxygen atoms and electropositive amino hydrogen atoms. The authors note that no direct experimental comparison between NAT- and AG-based polymers was performed, so the computational superiority of AG over NAT was not established experimentally.
The same computational framework guided the remaining components. Solvent-effect calculations using the solvation model based on density across eight solvents identified acetonitrile as the most favorable environment for maintaining template–monomer association, with an interaction energy of −21.86 kcal/mol. Cross-linker screening showed that TRIM bound AG too strongly (−17.28 kcal/mol), potentially competing with the template–monomer interaction, while DVB bound too weakly (−8.77 kcal/mol). EGDMA, at −13.54 kcal/mol—stronger than DVB but weaker than the 1:1 PADA–AG reference interaction of −15.79 kcal/mol—was selected. The template-to-monomer ratio was then optimized experimentally, with an imprinting-factor maximum at 1:5, and 1.0 mmol of Fe²⁺ proved optimal for magnetic functionalization.
Characterization confirmed the material’s design. Electron microscopy revealed rough, irregular particles with Fe₃O₄ nanoparticles of roughly 10–25 nm embedded in a polymer matrix. Vibrating-sample magnetometry showed a magnetization of 46.6 emu/g, sufficient for magnetic separation within 10 seconds. The polymers displayed a type IV nitrogen isotherm with mesopores mainly in the 2–10 nm range, and X-ray photoelectron spectroscopy detected iron, oxygen, carbon, and nitrogen, consistent with AG-derived nitrogen-containing functionalities. In adsorption tests, equilibrium capacities followed PADA > TEP > iVDA > HxDA, at approximately 9.60, 9.50, 9.21, and 8.87 mg/g, and the pseudo-second-order kinetic model fit the data far better than the pseudo-first-order model, with R² values of 0.99992–0.99998. Langmuir-estimated maximum capacities were similar across the four acetals, at roughly 46 mg/g.
Selectivity testing distinguished the imprinted material from its non-imprinted counterpart. The polymers adsorbed approximately 6.4–8.9 mg/g of the target acetals but only about 0.4–2.8 mg/g of four non-acetal ester references. Imprinting factors for the acetals ranged from 2.6 to 5.5, with the template PADA highest at 5.5, while non-acetal compounds scored between 0.13 and 0.95. Relative selectivity coefficients across all 16 target-versus-reference comparisons exceeded 1, ranging from 4.14 to 104.54. Wavefunction analyses—energy decomposition, independent gradient model, and atoms-in-molecules calculations on simplified AG–acetal complexes—indicated that dispersion (39.47–41.95 percent) and electrostatic interactions (38.33–40.08 percent) dominated the attractive energy, supplemented by localized weak hydrogen bonding; all identified intermolecular contacts fell into the weak, closed-shell category. The authors caution that these analyses of simplified complexes do not independently establish the specific binding forces within the polymerized recognition sites.
After optimizing extraction conditions—30 mg of adsorbent, 1 minute of adsorption, and 10 minutes of desorption with ethyl acetate containing 0.5 percent acetic acid—the method showed strong analytical performance. Linearity extended over 0.1–500 mg/L with R² values of 0.9995–0.9998; limits of detection ranged from 9.4 to 19.0 μg/L and limits of quantification from 32.7 to 48.2 μg/L. Recoveries at three spiking levels spanned 82.8–110.6 percent, with intraday and interday relative standard deviations of 3.2–9.9 percent and 3.7–13.3 percent. The adsorbent retained recoveries of 77.4–88.6 percent after five consecutive extraction cycles, demonstrating reusability in complex liquor matrices.
Applied to nine commercial Fenjiu samples with labeled production years from 1983 to 2022—corresponding to estimated aging periods of roughly 4 to 43 years—the method revealed distinct concentration patterns. Isovaleraldehyde diethyl acetal rose from 1.87 to 4.84 mg/L before a slight decline to 4.32 mg/L in the oldest sample, while TEP and hexanal diethyl acetal increased overall from 0.57 to 1.28 mg/L and 0.39 to 0.97 mg/L, respectively. PADA showed a non-monotonic pattern, rising from 2.37 to 4.28 mg/L in intermediate samples before falling to 1.64 mg/L. The authors emphasize that these trends are preliminary exploratory observations: only one commercial sample represented each labeled year, the samples came from labeled vintages rather than a controlled longitudinal aging experiment, and the differences cannot be distinguished from variation in raw materials, fermentation, or individual samples. Further validation through independently replicated controlled aging studies will be needed before any of these compounds can be regarded as validated aging markers.
Subject of Research: Development of a DFT-guided magnetic molecularly imprinted polymer method for selective enrichment and GC–MS determination of diethyl acetals in light-aroma Baijiu
Article Title: DFT-guided magnetic molecularly imprinted polymers for rapid and selective enrichment of diethyl acetals in Baijiu
Article References: Zhang, Z., Diao, Y., Guo, J., Li, W., Sun, B., Zhang, L., Dong, W., Sun, X., & Wang, S. (2026). DFT-guided magnetic molecularly imprinted polymers for rapid and selective enrichment of diethyl acetals in Baijiu. Food Chemistry: X, 40, Article 104641. https://doi.org/10.1016/j.fochx.2026.104641
Image Credits: AI Generated
DOI: 10.1016/j.fochx.2026.104641
Keywords: Baijiu, molecularly imprinted polymers, density functional theory, magnetic solid-phase extraction, diethyl acetals, GC-MS, aging markers, Fenjiu, food chemistry, acetalization, flavor analysis, adsorption selectivity
News Source: Beatrice Stafford. (October 29, 2026). Computational design yields magnetic polymers that rapidly capture aging markers in Baijiu. Scienmag.
