Monitoring Methanol Exposure during Liquor Distillation with a Hand-Held Device

Abstract

Introduction Toxic methanol occurs naturally in most distillates. Yet, suitable detectors to check product adherence to legal limits and, most importantly, monitor the methanol concentration in situ during distillation are not available. The legal limits for methanol contamination in the E.U. are 1350 g methanol per hL ethanol for Williams pear spirits, 1200 g/hL for apple and plum spirits and 1000 g/hL for cherry spirits. Significantly lower ones apply for brandy (i.e. 200 g/hL), Vodka from agricultural alcohol (30 g/hL) and London gin (5 g/hL). However, these limits are exceeded frequently, as shown for 183 commercial Williams pear spirits with methanol contents up to 1865 g/hL [1]. Usually, distillers rely on error-prone human olfaction while “gold standard” but off-line, time-consuming and expensive liquid or gas chromatography (GC) is rarely used.Here, we monitor methanol concentrations during industrial distillation of cherry, apple, plum and herb liquor (196 samples) with a low-cost and handheld detector. Therein, it quantified individual methanol concentrations (0.1 – 1.25 vol% in liquid) with high coefficient of determination (R2 = 0.92) by headspace analysis, as confirmed by GC. Typical ethanol contents (5 – 90 vol%) and aromas did not interfere, whereas methanol levels above legal limits were recognized within two minutes. As a result, this device enables distillers to guide distillation and check product quality, to prevent deadly methanol exposure and occupational hazard. Method Mash of plum, apple, cherry as well as a mixture of agricultural alcohol and herbs (for herb spirit) were processed in a distillery located in Switzerland (S. Fassbind AG). Distillate samples were drawn every 1 to 10 min (as compatible with the still operation) for sensor and GC analysis. The methanol detector is shown in Figure 1a. In brief, it consists of a needle (Sterican, B. Braun AG) mounted on a 4 mm (inner diameter) Teflon tube, a packed bed column with 150 mg of Tenax TA particles (35 m2/g, Sigma Aldrich), a flame-made chemoresistive gas sensor of 1 mol% Pd-doped SnO2 [2] nanoparticles and a rotary vane pump (135 FZ 3V, Schwarz Precision). The pump and sensor are controlled by a microcontroller (Raspberry Pi Zero W) mounted on a tailor-made printed circuit board with wireless communication to a laptop or smartphone [3].For each measurement, the vial containing the sample is shaken for 30 s, the sampling needle is inserted through the septum and the headspace is extracted at 25 ml/min for 10 s. Thereafter, the sampling needle is removed and ambient air is drawn for 6 min to convey the sample through the separation column and to analyze it by the sensor. Finally, the column and sensor are recovered by flushing with ambient. Methanol is quantified from the sensor response at a retention time of 1.5 - 1.7 min [4], based on previous calibrations in ternary mixtures of 0.1, 0.3, 0.5, 0.70, 1, and 1.25 vol% methanol (> 99.9 %, Sigma Aldrich) in 10, 30, 50, 70, 90 vol% ethanol (> 99.8 %, Fisher Scientific) and water (Milli-Q Synthesis A10, Merck). Results and Conclusions Figure 1b shows the results of selected samples (for better visibility) for plum (triangles), apple (circles), cherry (squares) and herb (stars), Already low methanol concentrations of 0.1 (cherry), 0.13 (apple) and 0.21 vol% (plum) in the mash (first data points) of the fermented fruits are detected. Highest methanol levels occur in plum distillates (triangles) that vary between 1.2 and 0.85 vol% and only drop significantly after 80 min reaching 0.2 vol% at the end of distillation. Similar methanol dynamics were observed for apple (circles), though at slightly lower concentrations ranging from 1 to 0.7 for t < 80 min. Distinctly lower methanol levels (0.5 – 0.4 vol%) were measured for cherry (squares). Lastly, there is herb spirit (stars) where methanol concentrations stayed constantly below 0.22 vol%. So, the handheld detector is able to follow the individual methanol dynamics of fruit and herb during distillation. Figure 1c shows a scatter plot of methanol concentrations in all (196) samples of herb (stars), cherry (squares), apples (circles) and plum (triangles) distillates, as measured by the handheld device and GC. Close agreement (ideal line, dashed) is observed over the entire concentration range of 0.1 to 1.25 vol%, as quantified with high coefficient of determination (R2 = 0.92).As a result, this handheld device is capable to identify critical methanol levels and could support distillers who mostly rely on olfactory analysis that is error-prone (given the indistinct smell of methanol over ethanol). Since the detector is fully integrated, pocket-sized, battery operated, indicating results on a laptop or smartphone and consists mostly of low-cost components, it is ideal for on-site application and should be affordable even for small distilleries.