Content
Citation(s) & Footnote(s)
Book(s)Whiskey Knowledge Databases › Literature Notes
Complete analytical Literature Note produced from a full-source review. Claims below are bounded by the recorded evidence and limitations.
Scope and core summary
Complete synthesis of a four-page symposium paper comparing extraction methods for a 25-compound model spirit and applying an optimized HS-SPME method to gin, rum, and whisky.
Author argument
The authors use a deliberately simplified model matrix to compare LLE, SPE, SPME, and SBSE/HSSE, concluding that HS-SPME was the best operational compromise and could recover differentiating volatile patterns in three commercial spirits.
Researcher synthesis
The paper is a compact methodological precursor to the authors’ longer 2022 study. Its main value is procedural: controlled model mixtures can isolate sampling-method effects before the method is applied to real whiskey, reducing the risk that apparent product differences are artifacts of extraction.
Evidence assessment
Primary conference research; clear method comparison but preliminary and superseded in detail by the later journal article.
Limitations and open questions
Limitations: Short proceeding, only three commercial applications, limited validation, and no direct sensory-causation test. Retain without duplicating the stronger 2022 paper’s evidence.
Open questions: No source-completion question remains; cite the 2022 article when the same claim is available there.
Connected records
- Contributors: 4
- Verified Excerpts: 3
- Citations: 1
- Zettels: 2
Completion record
Full-source pass completed: 4/4 local PDF sheets and 2,570 extracted words reviewed. Abstract, method comparison, optimization, results, conclusion, and references checked. Evidence locators retained at local sheets 1 and 4. No completion hold remains.
Barnes et al. (2021) — Complete Symposium Paper Review
Quentin Barnes, Marie-Anne Contamin, Nicolas Papaiconomou and Xavier Fernandez. Developing methods for characterising flavour compounds in distilled spirits: Application to gin, rum and whisky. Proceedings of the 16th Weurman Flavour Research Symposium, edited by E. Guichard and J. L. Le Quéré. DOI 10.5281/zenodo.5617614. SRC-386 / LIT-310. Reviewed September 27, 2026.
Coverage and Source Identity
All four supplied pages read, including each of the 25 model-mixture rows, all five extraction-mode rows, every cell of the 14-class/three-product table, both panels of Figure 1, conclusion and six numbered references. Visual inspection covered pages 3–4. Table 1's separate van Gemert (2011) attribution was also read; it is not a fully specified entry among the six references. No appendix or supplement is mentioned. The existing Notion attachment is retained and the original local PDF is unchanged. Proton name/size matching remains provisional; browser rendering is not verified.
Contribution and Method
The authors develop a model spirit containing 25 compounds representing fifteen chemical classes to compare sample-preparation methods before testing commercial products. This is an earlier version of the same research program as their longer 2022 microdistillery article, not independent replication. Two authors are affiliated with Comte de Grasse and its 44°N gin was provided by the company. The rum is Diplomático Selección de Familia; the whisky is Glenfiddich 14-year Bourbon Barrel Reserve, a Scotch whisky, not bourbon. One sample per product does not characterize its category. [p. 1]
The model spans roughly a 20,000-fold concentration range, from sotolon at 0.000938 mg/L to furfural at 18.2 mg/L. The design tests extraction coverage at these selected concentrations; it is not a validated universal average spirit. Water odor thresholds in Table 1 are reference values, not experimental thresholds in the hydroalcoholic mixture. Superscript a/b/c markers for selected thresholds lack explanations in this supplied paper, so their exact matrix qualifications cannot be recovered here. [pp. 2–3]
Liquid-liquid extraction detected 23/25 compounds, SPE 17/25, SPME 21/25, and both stir-bar modes 22/25. LLE's missing sotolon and hexanoic acid reflect concentration/method response limitations, not absence from the prepared mixture. HS-SPME was chosen for operational balance and potential automation despite detecting fewer compounds than LLE. “Best compromise” must not become “most complete extraction.” [p. 3, Table 2]
The selected fiber was DVB/CAR/PDMS in headspace mode, with dilution from 40% to 10% ethanol and added salt. At least sixty minutes extraction was selected, and 35°C was the chosen compromise temperature. Figure 1 shows measured bars at room temperature, 30, 45 and 60°C; no 35°C bar is supplied. Its kinetic panel shows compound-specific behavior: beta-caryophyllene continues increasing, furfural levels off, alpha-pinene declines toward a plateau, and beta-damascenone rises slowly. Thus one hour does not establish equilibrium for every analyte. The prose only describes the room-to-45°C range despite the chart's 60°C series. Error bars are visible but not defined in the caption. [pp. 3–4]
The protocol also specifies 24 hours of headspace equilibration at 25°C before extraction. Table 2's 0.2-hour SPME handling time therefore is not total turnaround time. The LLE/SPE/SPME oven program totals 107.5 minutes, calculated from its holds and ramps, in addition to preparation and equilibration. Stir-bar methods use a different GC-MS platform, column and desorption system, introducing another comparison variable beyond extraction alone. [p. 2]
Commercial Results and Arithmetic Audit
All Table 3 class counts sum to the printed totals: gin 122, rum 61, whisky 74. The whisky column contains 9 alcohols, 43 esters, 2 lactones, 7 carbonyls, 2 acetals, 2 acids, 0 phenols, 2 furans, 4 aromatics, 1 monoterpene, 0 oxygenated monoterpenes, 0 sesquiterpenes, 1 oxygenated sesquiterpene and 1 other compound. Gin's four terpene classes total 92 of its 122 identifications; rum has 32 esters. These are method-dependent identification counts, not concentrations, odor activity, consumer preference or comprehensive flavor inventories. Zero phenols detected in the whisky does not establish that the bottle contains no phenols. No sensory panel or causal aroma reconstruction is performed. [p. 4]
The authors suggest a polar column could improve phenol/furan recovery and propose future purge-and-trap work. Those are proposed improvements, not validation results in this four-page report. Claims of general applicability are limited to these three examples and the model mixture.
Version Differences and Reproducibility Limits
This 2021 conference version uses 25 model compounds and reports SPE coverage of 17, unlike the later 2022 article's 24-compound model and SPE result of 22. Commercial identification totals also differ from the later report. Do not splice the later paper's count into this version or count the two papers as independent confirmation. The 2022 article offers more detail but also has its own count/label issues recorded in its separate review.
Several protocol details need resolution before copying this as an SOP: SPE conditioning lists ethanol/water 9:1 despite substantial subsequent dilution; the stir-bar sample and water volumes total 12.5 mL; tabulated rounded sample/solvent volumes differ from the procedural amounts (for example three 60 mL LLE solvent portions versus the table's 200 mL). These may be rounding or reporting choices, but the supplied text does not reconcile them. Quantitative recovery, calibration, matrix effects, detection limits, interday precision and independent product replication are not sufficiently established here to call the method universally validated. Some identities rely on library spectra/retention indices; only some are confirmed with standards.
Proposed Academy Uses and Links
Use this short paper beside the longer 2022 study to teach version-specific citation and why detection counts change with methods. A practical worksheet could compare breadth, hands-on time, total elapsed time, solvent consumption and automation as separate decisions. The 24-hour equilibration instruction provides a concrete reason not to advertise a twelve-minute analysis from handling time alone.
Link to the existing method-selection Zettels and to Kew's formula-space work: a chromatographic identification, an assigned elemental formula and a demonstrated aroma contribution are different evidence types. Pair with Poisson and Schieberle for the additional experiments needed to establish sensory importance. Keep the symposium paper as a distinct preserved source with a clear precursor relationship. Proposed Academy applications remain internal; no public course page was modified.
Related Journal Version
