Abstract: Introduction Chemical recognition with sensors offers tremendous potential in emerging applications including the next generation of non-invasive medical diagnostics [1], smart indoor air control [2] , exhaust gas analysis and food quality assessment. Nowadays, chemical gas sensors offer low price, compact size, easy applicability and they typically can detect target analytes at low part-per-billion (ppb) concentration. A major obstacle, however, in the realization of such sensors is their lack of selectivity [3]. They often fail in the application due to the complexity of the gas matrix (ambient air or breath) containing hundreds of interferants. Strategies to overcome this limitation include sensor material design and combining different selective sensors to arrays [4] . Often, however, the achieved selectivity is insufficient for the stringent requirements in the applications. Here, we present the use of compact separation columns upstream of sensors to achieve unprecedented selectivity to target analytes. Method The separation column consists of a packed bed of Tenax TA particles (60–80 mesh, ~35 m 2 g -1 ) inside a Teflon tube (4 mm inner diameter) at loading of 150 mg, secured on both ends with silanized glass wool and tension springs. Columns were flushed over night with 100 mL/min synthetic air at 50% relative humidity (RH) before use to desorb impurities that might be present on Tenax. Sensors were prepared by depositing flame-made Pd-doped SnO 2 (1 mol% Pd) nanoparticles directly onto micromachined sensor substrates. These microsensors were bonded onto leadless chip carriers and placed inside a Teflon sensor chamber. The separation column was secured upstream of the sensor and connected to it with Teflon connections. The performance of the column–sensor system was characterized with a dynamic gas mixing setup, where it was continuously flushed with synthetic air at varying relative humidity (RH). Analyte gases, supplied from calibrated gas standards, were admixed to the synthetic gas air stream to generate well-defined analyte exposures. Analyte retention times ( t R ) were defined as the time elapsed between the analyte exposure and sensor’s maximum response, analogous to gas chromatography. Results and Conclusions When exposing the non-specific sensor without separation column to analytes typically contained in ambient air [5] and breath [6] such as ethanol, methanol or acetone, it quickly responds to them (response time 〈 5 s) with high responses at low ppb concentrations [7]. However, it cannot accurately measure single compounds in gas mixtures, especially when interferants are present at elevated concentrations. This can be solved by a simple and compact separation column. In Figure 1a, the sensor resistance with separation column to 10 s pulses of methanol, ethanol and acetone is shown. The analytes are retained depending on their specific interaction with the Tenax surface. As Tenax is non-polar and adsorbs molecules primarily due to nonspecific van-der-Waals forces, the analyte retention times typically increase with higher molecular weight. As a result, methanol is detected first ( t R = 1.7 min), followed by ethanol (8.7 min) and acetone (33 min). Analytes that are retained longer are thereby released over a longer time period, leading to lower maximum concentration and lower sensor response. Similar responses are thus obtained for all compounds although concentrations vary greatly (1, 5, and 20 ppm). When exposing the column–sensor system now to the gas mixture (Figure 1b), all analytes are detected individually at their corresponding retention time, i.e., with very high selectivity. This concept can be utilized for selective detection of individual target analytes, as recently demonstrated for selective detection of toxic methanol over ethanol in breath and liquor [7] , or to detect multiple analytes with a single sensor. By flushing with air, the separation column regenerates, which can be easily sped up by slight heating of the column and increasing the flow rate (e.g., acetone from ~60 min to 20 s when increasing column temperate and flow rate briefly to 80 °C and 100 mL min -1 ). As a result, separation columns demonstrate how to possibly address chemical sensors’ long-standing challenge of selectivity. They give comparable performance to classical gas chromatographic columns, however, are much simpler in design (packed particle bed), inexpensive ( 〈 10 $) and compact (4 cm length). A broad variety of different commercial sorbent materials is readily available to design columns for specific applications and combine them freely with different sensor technologies typically suffering from low selectivity (e.g., metal-oxide, optical or electrochemical sensors). Based on their low price and compact size, they could enable the next generation of highly selective portable breath analyzers and environmental monitors. References [1] A.T. Güntner, S. Abegg, K. Königstein, P.A. Gerber, A. Schmidt-Trucksäss, S.E. Pratsinis, Breath sensors for health monitoring , ACS Sensors. 4 (2019) 268-280. [2] M. Mayer, A.J. Baeumner, A Megatrend Challenging Analytical Chemistry: Biosensor and Chemosensor Concepts Ready for the Internet of Things , Chemical Reviews. 119 (2019) 7996-8027. [3] A. Lewis, P. Edwards, Validate personal air-pollution sensors , Nature. 535 (2016) 29-31. [4] K. Persaud, G. Dodd, Analysis of discrimination mechanisms in the mammalian olfactory system using a model nose , Nature. 299 (1982) 352. [5] T. Salthammer, Very volatile organic compounds: an understudied class of indoor air pollutants , Indoor Air. 26 (2016) 25-38. [6] B. de Lacy Costello, A. Amann, H. Al-Kateb, C. Flynn, W. Filipiak, T. Khalid, D. Osborne, N.M. Ratcliffe, A review of the volatiles from the healthy human body , Journal of breath research. 8 (2014) 014001. [7] J. van den Broek, S. Abegg, S.E. Pratsinis, A.T. Güntner, Highly selective detection of methanol over ethanol by a handheld gas sensor , Nature Communications. 10 (2019) 4220. Figure 1

Abstract: Introduction In July 2019, more than 20 people died in Costa Rica due to methanol-tainted alcohol of registered and unregistered brands [1]. Similar outbreaks have been reported in India [2] (95 deaths, February 2019), Iran [3] (768 cases, 2018) or Cambodia [4] (237 cases, 2018). In fact, 5400 intoxications and 1500 deaths were reported due to methanol-tainted alcohol within the last two years [5]. Apparently, cheap methanol is intentionally added to alcoholic beverages to increase their potency and profit. In the human body, methanol is converted to formaldehyde and formic acid, which can interfere with the human metabolism leading to blindness or even death [6] . Therefore, handheld analyzers or sensors to screen beverages are urgently needed to prevent such intoxication. These could be applied by tourist or even first responders. However, compact gas sensors either suffer from insufficient detection limit (legal limit of 0.09 vol% in fruit distilled spirits [7]), cannot distinguish methanol from the chemically similar ethanol present at much higher concentration or have not been validated with real beverages. Here, we present a stand-alone, inexpensive and miniaturized analyzer for screening of potential safety hazards in beverages [8] that can be applied even for hand sanitizers. Method The handheld device consists of a printed circuit board with integrated microcontroller and circuitry to operate metal-oxide gas sensors (Figure 1a). A miniaturized pump (12 g) is used to guide the sample through a sorption filter for analyte pre-separation to a highly sensitive and flame-made Pd-doped SnO 2 sensor [9]. The sample was obtained from the headspace above liquors. The device was calibrated with laboratory mixtures containing methanol and ethanol in ultra-pure water at concentrations ranging from 0.01 – 10 vol% and 5 – 80 vol%, respectively. Subsequently, the device was tested with real beverages including beer, sake, Baileys, Arrack, pear spirit and Stroh rum. The drinks were chosen to cover a wide range of ethanol concentrations (5 – 80 vol%) and flavors that might interfere with the detector. The methanol concentration of the pure drinks was determined by liquid chromatography. To generate higher methanol concentrations, the pure beverages were spiked with methanol up to 10 vol%. Results and Conclusions In laboratory mixtures, methanol (1 vol%) and ethanol (5 vol%) passed through the separation column with retention times of 1.6 and 5.3 min, respectively, that can be detected subsequently (thus selectively) by the Pd-doped SnO 2 sensor. Thereby, the effective separation of methanol and ethanol is crucial. This was evaluated by comparing the retention time of methanol to the breakthrough time of ethanol. For methanol, the retention time ranged from 1.25 – 1.5 min depending on concentration, from 0.01 to 10 vol%. Importantly, the breakthrough time of ethanol was always larger than 2.2 min. This enabled reliable and simultaneous detection of methanol and ethanol, even at their lowest and highest concentrations, respectively. For all alcoholic beverages, no interference originating from the different flavors was observed on the methanol and ethanol quantification. Most importantly, the device predicted methanol and ethanol with average error of 21.9 and 13.4% (with respect to the actual concentration), when tested in 43 pure and methanol spiked beverages (Figure 1b). Also, the performance was reproducible and stable for at least 107 days. As a result, the validated device could readily be applied by laymen (e.g. tourists) or first responders to screen beverages for toxic methanol and prevent severe health damage. References [1] Knowles, H., Tainted alcohol has led to 20 deaths in Costa Rica, authorities say. The Washington Post July 24, 2019. [2] Kai Schulz; Kumar, H., Over 90 killed in India by toxic homemade liquor. The New York Times 23. Feb, 2019. [3] Aghababaeian, H.; Araghi Ahvazi, L.; Ostadtaghizadeh, A., The methanol poisoning outbreaks in Iran 2018. Alcohol Alcohol. , 54 (2019) 128-130. [4] Sen, D. More than 100 villagers return home after methanol poisoning. https://www.khmertimeskh.com/487671/more-than-100-villagers-return-home-after-methanol-poisoning/ (accessed June 6th 2019). [5] Oslo University Hospital; Doctors without borders The methanol poisoning initiative - suspected methanol poisoning incidents. https://legerutengrenser.no/mpi/ (accessed August 19, 2019). [6] Kruse, J., Methanol poisoning. Intensive Care Med. , 18 (1992) 391-397. [7] The european parliament and of the council, Regulation (EU) No 2019/787. Official Journal of the European Union, Ed. 2019; Vol. L130/1. [8] Abegg, S.; Magro, L.; van den Broek, J.; Pratsinis, S. E.; Güntner, A. T., A pocket-sized device enables detection of methanol adulteration in alcoholic beverages. Nature Food , 1 (2020) 351-354. [9] van den Broek, J.; Abegg, S.; Pratsinis, S. E.; Güntner, A. T., Highly selective detection of methanol over ethanol by a handheld gas sensor. Nat. Commun. , 10 (2019) 4220. Figure 1

Abstract: Introduction Over 800 people died this year in Iran alone as a result of methanol poisoning [1]. In fact, such outbreaks are quite common with thousands of victims each year [2] . However, no inexpensive and fast point-of-care method exists for diagnosis of methanol intoxication to rapidly respond to disasters. Currently, methanol poisoning is detected directly through blood analysis by gas chromatography or indirectly through blood gas analysis [3]. Both require trained personnel, are expensive and rarely available in developing countries where most outbreaks occur. Blood methanol levels can also be determined non-invasively in exhaled breath (Figure 1a), analogous to ethanol as widely applied by law enforcement [4] . However, current chemical sensors cannot distinguish methanol from the usually much higher ethanol background in the breath of poisoned victims. Here, we present an inexpensive and handheld detector for rapid and highly selective methanol detection. Method The handheld detector is shown in Figure 1b. It consists of a capillary inlet for sampling from Tedlar bags, a separation column consisting of a packed bed of Tenax TA polymer sorbent to separate the breath mixture [5], a chemoresistive Pd-doped SnO 2 microsensor to quantify the methanol and ethanol concentrations, and a rotary vane pump (SP 135 FZ 3 V, Schwarzer Precision, Germany) drawing the sample through the column to the sensor. A microcontroller (Raspberry Pi Zero W, Great Britain) with integrated circuits on a custom-designed printed circuit board (PCB) is used for autonomous sensor heating, film resistance readout, pump flow control as well as wireless communication with a computer or smartphone [6]. The detector was validated with methanol-spiked breath of drunken volunteers (0.1% blood ethanol) [7]. Late expiratory breath was sampled into Tedlar bags and subsequently spiked with 10–1000 ppm methanol on a dynamic gas mixing setup to simulate poisoning without intoxicating volunteers. Samples were then measured by the detector and a high resolution proton-transfer-reaction time-of-flight mass spectrometer (PTR-TOF-MS). Results and Conclusions The main challenge for chemical sensors is the selective detection of breath methanol over high ethanol concentrations present after consumption of contaminated beverages or during methanol poisoning treatment where ethanol is even used as an antidote. The present detector (Figure 1b) achieves this with a compact separation column, where ethanol absorbs stronger (and thus is retained longer) than methanol, analogous to GC. A downstream chemoresistive microsensor based on Pd-doped SnO 2 nanoparticles quantifies the methanol and ethanol sequentially with high sensitivity. Figure 1c shows the sensor response to breath (green) after consumption of an alcoholic beverage and when spiked with 23 (blue), 66 (purple) and 148 ppm (red) methanol. These methanol levels correspond to endogenous (0–10 ppm), harmless exogenous (10–52 ppm) and toxic concentrations ( 〉 52 ppm), respectively. Most importantly, the sensor detects no significant methanol concentration in the original breath (PTR-TOF-MS, 〈 1 ppm) with sensor response below the LOD, as expected from physiological breath methanol concentrations (median 0.26 ppm), while the spiked concentrations are recognized distinctly at 1.8 min. In total, 105 methanol-spiked breath samples from 20 volunteers after consumption of water, beer, liquor or wine were evaluated and the measured methanol concentrations of the detector and PTR-TOF-MS are shown in Figure 2a. Indicated also are the concentration ranges where antidote ( 〉 52 ppm, grey shaded) and hemodialysis ( 〉 131 ppm, red shaded) treatments are recommended from the corresponding blood methanol concentrations. The detector shows excellent agreement with PTR-TOF-MS (R 2 = 0.966) over the entire concentration range (14–1079 ppm) and in the presence of 0–316 ppm ethanol. As a result, this device is promising to screen methanol poisoning and classify severity. This detector can be equipped with a disposable mouthpiece, as for commercial breath alcohol testers, and readily applied as a point-of-care diagnostic tool for fast screening of methanol poisoning by first responders and clinicians. References [1] H. Hassanian-Moghaddam, N. Zamani, A.-A. Kolahi, R. McDonald, K.E. Hovda, Double trouble: methanol outbreak in the wake of the COVID-19 pandemic in Iran—a cross-sectional assessment , Critical Care. 24 (2020) 402. [2] The American Academy of Clinical Toxicology Ad Hoc Committee on the Treatment Guidelines for Methanol Poisoning:, D.G. Barceloux, G. Randall Bond, E.P. Krenzelok, H. Cooper, J. Allister Vale, American Academy of Clinical Toxicology practice guidelines on the treatment of methanol poisoning , Journal of Toxicology: Clinical Toxicology. 40 (2002) 415-446. [3] J.A. Kraut, Diagnosis of toxic alcohols: limitations of present methods , Clinical Toxicology. 53 (2015) 589-595. [4] A.T. Güntner, S. Abegg, K. Königstein, P.A. Gerber, A. Schmidt-Trucksäss, S.E. Pratsinis, Breath Sensors for Health Monitoring , ACS Sensors. 4 (2019) 268-280. [5] J. van den Broek, S. Abegg, S.E. Pratsinis, A.T. Güntner, Highly selective detection of methanol over ethanol by a handheld gas sensor , Nat. Commun. 10 (2019) 4220. [6] S. Abegg, L. Magro, J. van den Broek, S.E. Pratsinis, A.T. Güntner, A pocket-sized device enables detection of methanol adulteration in alcoholic beverages , Nature Food. 1 (2020) 351-354. [7] J. van den Broek, D. Bischof, N. Derron, S. Abegg, P.A. Gerber, A.T. Güntner, S.E. Pratsinis, Screening Methanol Poisoning with a Portable Breath Detector , Analytical Chemistry. (Just Accepted) (2020) Figure 1

Abstract: Introduction “Over 90 Killed in India by Toxic Homemade Liquor” read the headline in The New York Times on Feb. 23, 2019 [1]. In fact, such outbreaks are quite common with thousands of victims each year [2] . However, no inexpensive and fast point-of-care method exists for diagnosis of methanol intoxication or screening of laced alcoholic beverages to rapidly respond or even prevent such disasters. Currently, methanol poisoning is detected directly through blood analysis by gas chromatography or indirectly through blood gas analysis [3]. However, both require trained personnel, are expensive and rarely available in developing countries where most outbreaks occur. Blood methanol levels can also be determined non-invasively in exhaled breath, analogous to ethanol as widely applied by law enforcement [4] . However, current chemical sensors cannot distinguish methanol from the much higher ethanol background. Here, we present an inexpensive and handheld detector for rapid and highly selective methanol detection. It consists of a separation column (Tenax) separating methanol from interferants like ethanol, acetone or hydrogen and a chemoresistive gas sensor (Pd-doped SnO 2 nanoparticles) to quantify the methanol concentration. This way, methanol is precisely quantified in vivo in the breath of an ethanol-intoxicated subject and in liquor headspace without interference of much higher ethanol levels. Method The methanol detector consists of a separation column placed upstream of a chemoresistive microsensor. The separation column is a packed bed of 150 mg Tenax TA (60–80 mesh, ~35 m 2 g -1 ) inside a Teflon tube (4 mm inner diameter). The sensor consists of Pd-doped SnO 2 (1 mol% Pd) nanoparticles produced by flame spray pyrolysis (FSP) and directly deposited by thermophoresis onto micromachined free-standing membrane type sensor substrates (1.9 x 1.7 mm 2 ). The sensing film is heated to 350 °C through the back heater of the sensor substrate, while a miniature vane pump draws the sample through the column to the sensor at 25 mL min -1 . The detector is characterized with synthetic gas mixtures in a mixing setup adapted from [5] and tested directly on methanol-spiked liquor (Indonesian Arrack) and breath of a drunken volunteer. Results and Conclusions Figure 1 shows the separation column and the micromachined metal-oxide gas sensor of the handheld methanol detector. Breath or the headspace of a beverage is drawn by a pump through the separation column to the sensor. The separation column separates analytes based on their interaction with the Tenax particles, as in chromatography, although much more compact (4.5 cm long and 4 mm inner diameter) and simpler in design. The sensor is micromachined, offering small size and minimal power requirement (76 mW at 350 °C) readily suitable for integration into a battery-driven device. In combination, the resulting detector senses methanol within 2 min in a wide concentration range of 1–918 ppm and can differentiate normal from toxic breath methanol levels (Figure 2). Most importantly, due to separation of analytes in the column, it is not interfered irrespective of much higher ethanol concentrations, as well as other possible interferants (e.g., hydrogen, acetone). Figure 3a shows the response of the detector when sampling the headspace of methanol-laced liquor (Arrack). Thereby, concentrations down to 0.3 vol% are detected and distinguished from the pure liquor. Even small differences between 0.3, 0.4 and 0.5 vol% (i.e., close to the legal limit [6]) are clearly resolved with high signal-to-noise ratio 〉 100. As first proof-of-concept, the detector was also tested on breath of a volunteer after consumption of ethanol. Figure 3b shows the response of the detector to normal and methanol-spiked breath with a methanol concentration of 135 ppm, indicative for a serious intoxication ( 〉 133 ppm [2]). Thereby, the added methanol is clearly detected separated from the ethanol background at its specific retention time with high signal-to-noise ratio 〉 1000. In both applications (liquor headspace and breath) the detector fully recovers within 15 min by flushing with ambient air. As a result, this detector shows great promise for quick and non-invasive screening of methanol poisoning from breath and laced alcoholic beverages. Based on its compact size and low price, it could be used by locals, tourists and first responders in developing countries, where most outbreaks occur.[7] References [1] K. Schultz, H. Kumar. The New York Times: Over 90 killed in India by toxic homemade liquor . 2019 March 6, 2019; Available from: https://www.nytimes.com/2019/02/23/world/asia/india-poison-alcohol.html . [2] The American Academy of Clinical Toxicology Ad Hoc Committee on the Treatment Guidelines for Methanol Poisoning:, D.G. Barceloux, G. Randall Bond, E.P. Krenzelok, H. Cooper, J. Allister Vale, American Academy of Clinical Toxicology practice guidelines on the treatment of methanol poisoning , Journal of Toxicology: Clinical Toxicology. 40 (2002) 415-446. [3] J.A. Kraut, Diagnosis of toxic alcohols: limitations of present methods , Clinical Toxicology. 53 (2015) 589-595. [4] A.T. Güntner, S. Abegg, K. Königstein, P.A. Gerber, A. Schmidt-Trucksäss, S.E. Pratsinis, Breath sensors for health monitoring , ACS Sensors. 4 (2019) 268-280. [5] J. van den Broek, A.T. Güntner, S.E. Pratsinis, Highly selective and rapid breath isoprene sensing enabled by activated alumina filter , ACS Sensors. 3 (2018) 677-683. [6] A.J. Paine, A.D. Dayan, Defining a tolerable concentration of methanol in alcoholic drinks , Human & Experimental Toxicology. 20 (2001) 563-568. [7] J. van den Broek, S. Abegg, S.E. Pratsinis, A.T. Güntner, Highly selective detection of methanol over ethanol by a handheld gas sensor , Nature Communications. 10 (2019) 4220. Figure 1