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 an 11-page protocol for producing approximately 500 mL of new-make bourbon from about 2.5 kg of grain with discrete sampling points.
Author argument
The authors present a repeatable lab-scale cooking, mashing, fermentation, and double-distillation method designed to make grain, process, and terroir experiments feasible with limited material.
Researcher synthesis
This is a bridge between commercial bourbon practice and controlled research. It can inform educational process diagrams and experimental design, but it should not be copied into consumer distillation instructions or treated as proof that lab spirit predicts matured whiskey.
Evidence assessment
Primary methods communication from the University of Kentucky and James B. Beam Institute, with batch-reproducibility and metabolite sampling built into the design.
Limitations and open questions
Limitations: Produces new make rather than matured bourbon; small equipment changes may affect heat and separation; the paper’s compressed origin narrative is not controlling history; legal and safety constraints bar casual replication.
Open questions: Which controlled maturation platform can extend this method from new make to comparable aged samples?
Connected records
- Contributors: 4
- Verified Excerpts: 3
- Citations: 1
- Zettels: 4
Completion record
Full-source pass completed: 11/11 local PDF sheets and 6,425 extracted words reviewed, including process steps, analytical checkpoints, conclusions, and references. Evidence locators retained at local sheets 1 and 9. No completion hold remains.
Independent full-source audit — September 27, 2026
Scope and contribution
All11supplied pages,38references,Figures1–4andTables1–2 read and visually inspected. The original supplementary workbook foods-2098078-supplementary.xlsx was recovered through Europe PMC's open-access service and preserved internally. Its only worksheet, Table S1. Preferm_data, has33rows×4columns: all headers and32data rows were read. It contains no formulas, charts or images. Original files remain unchanged.
Verges, Gollihue, Joyce and DeBolt, Foods2023,12,457,DOI10.3390/foods12030457, is a primary methods communication. It offers a manageable corn-based research workflow with stage-specific sampling. It does not demonstrate equivalence to commercial equipment, matured flavor, consumer preference, or independent interlaboratory reproducibility. The abstract's confident extrapolation claim exceeds the experiment shown.
Design and sequence
Pages1–2 frame grain research and a small-crew laboratory process. Eight corn samples came from a distillery reference and several Kentucky farm/hybrid/irrigation combinations. This is not a balanced factorial experiment isolating terroir or genotype. Two300gNIR subsamples per corn source characterize composition; these are not replicated fermentations of each treatment.
Pages3–4describe corn cleaning/milling, added commercial enzymes, yeast/nutrient, controlled fermentation, a copper first distillation and stainless air-still second distillation. Onlycorn is specified as the cereal input; the method does not reproduce the customary rye/wheat/malt combinations of every bourbon mash bill. It aims at standardized research comparison. Yeast “neutral flavor” is an asserted selection rationale, not a sensory result measured here.
Sampling spans pre-fermentation mash,96-hourbeer,lowwines andhearts. HPLC tracks sugars,ethanol,glycerol andorganicacids. Lowwines are standardized to40%ABV before the second distillation, a useful control because ethanol concentration affects separation. No actual sensory panel or tasting results appear, even though the output is intended for later sensory work.
Pages5–7report corncomposition,sugars,acids andethanol. Table1has small differences in starch across these samples, but the authors appropriately decline to infer a starch-yield relationship. Reported consistency must be assessed against the actual variability and replication structure. Grain-source variation and method variation are conflated in the eight batches. An approximately12.5%glucoseCV does not by itself establish precision across operators, laboratories or repeated batches of the same grain.
Pages8–9show four equipment photographs and discuss changes needed for other grains and for more precise cuts. The pictures document this setup, not commercial equivalence. The final conclusions propose future hybrid/terroir research. Noexternalfunding is declared; donatedcorn,enzymes/yeast andtechnicalsupport are acknowledged. Authors report no conflict. Other raw data are available on request; no external request was sent.
Supplement reconciled with the paper
TableS1pre-fermentation glucose mean83.7425g/L,SD10.5008,SE3.7126; maltosemean58.2075,SD6.2457,SE2.2082. At96hours glucose mean16.1375,SD5.6477,SE1.9968; maltosemean2.19,SD0.4651,SE0.1644. These independently calculated means agree with the rounded narrative andFigure2.
Residualglucose spans4.18–22.30g/L andmaltose1.30–2.74g/L. “Fermentation accomplished” therefore must not be paraphrased as all sugars exhausted. Mean residualglucose isabout19.3%of the initialmean; maltoseabout3.76%,rather than exactly3%. These ratios compare groupmeans,not individualbatch conversionefficiencies or a full starch/ethanolmassbalance. Continued starchhydrolysis and samplingbasis also matter.
The sum of pre-fermentation glucose andmaltose bysample is127.56,121.56,157.18,146.48,140.94,123.58,160.20and158.10g/L. The range supports the rounded text121.6–160g/L. Figure1's stacked bars appear appreciably higher than these totals for several samples, particularly3and7; use the workbookvalues for any derived graphic and flag the original chart discrepancy rather than silently reproducing it.
Statistical and procedural discrepancies
Table2is labeled mean±SE, but calculation from its eight printed observations reproduces the displayed errors as sample standard deviations: glycerolSD0.3800,aceticacidSD0.2689,lacticacidSD1.1636g/L. Corresponding SEs would be0.1343,0.0951and0.4114. Do not copy its error labels into Academy evidence. Figure2's errors approximately match the independently calculated SEs; Table2's confirmed labeling issue does not automatically prove every figure is wrong.
The abstract/aim report500–700mLfinaldistillate and the conclusion approximately500mL, while the stated second-pass method collects300mLhearts. The discrepancy is unresolved. Output volume must be described as an author claim with conflicting methods, not a verified500mLproduction result. The discussion's two passes do not explain the difference clearly.
The stated7.8Linitialwater plus grain followed by transfer into a7.6Lfermenter lacks a finalvolume,evaporation or split-batch explanation. The paper says water starts to boil at the85°C stage; temperaturelocation and intended meaning are unclear. Burner settings and measured still temperatures are different quantities. Those details prevent copying this text as a ready-to-run standard operating procedure.
Hydrometer/refractometer readings in an alcoholic,solids-containing matrix do not by themselves equal residual fermentable sugar. The stated below1°Brix endpoint needs matrix/calibration interpretation alongside the measured residuals. HPLC's two-stage acid-hydrolysis wording, biomass/lignin references and unrelated solvent/catalyst list need methodological clarification: hydrolyzed total sugars and native soluble sugars are not interchangeable. Calibration,recovery andconversion details are insufficient for independent reproduction of all reported concentrations.
“Scarification” and “liquidation” onpage8appear to be terminology errors for saccharification/liquefaction. Grainbasis,enzymepotency,coolingtime,evaporation,temperaturecorrection andsampling details need controlled documentation in a future implementation.
Critical assessment and Academy use
This paper is strongest as a research-design example: bounded graincharges,explicit samplingpoints,standardization before comparison,and measurement of several metabolites rather than alcoholstrength alone. It is weaker as a validation paper. There are no operator/laboratory reproducibility trials,validated commercialmatched controls,agingarms or sensorycomparisons. FinalABV is concentration,not alcoholyield; a yield claim needs volumes,masses andrecovery on a specifiedgrainbasis.
The opening Scotch-Irish origin narrative and old economicnumber are contextual assertions,not findings of this experiment and not controlling historical evidence. Avoid carrying the uncomplicated single-origin account into the Academy. Likewise, historical legalcitations and industrialgeneralizations need their own current sources when used in course content.
Proposed Academy applications: an annotated process-and-sampling diagram; a lesson distinguishing ABV,yield,residualsugar,SDandSE; and an experiment-design worksheet identifying controlledvariables,confounds andrequiredvalidation. The recoveredTableS1provides a compact source-grounded arithmetic exercise. These are proposed educational uses,not changes to publiclessons or an implemented lab.
Cross-source synthesis: Daute's wortpretreatment work shows how handling can alter the experiment; its sensorymethodcomparison separates aroma,similarity andcomposition; this cornmethod supplies a relevant Americanwhiskey researchplatform. Together they support a staged validation plan,not a claim that labnewmake predicts maturedbourbon. Preserve newmake versus matured endpoints and distinguish repeatable measurements from demonstrated external validity.
Verification record
Supplement recovered from https://www.ebi.ac.uk/europepmc/webservices/rest/PMC9914533/supplementaryFiles ; nested ZIP contained the single originalXLSX,nowattached to the existingSource. All publishedmaintext and suppliedsupplement examined. Unpublishedrawrecords have not been obtained. Existing evidence500mL/reproducibility wording must retain the conflicts above. OriginalNotionidentities andrelations preserved. Native readback,Protonbyteidentity andbrowserverification are separately recorded.
