DEVICE AND METHOD FOR DETECTING GAS

Abstract

An analyzer for detection of a target compound in a complex gas mixture in the gas phase includes a detector for detecting the target compound and further has a sensing cavity. The detector is arranged in the sensing cavity. The analyzer also has a separate first membrane that is equipped to close the cavity and simplifies the composition of the complex gas mixture into a first gas mixture, wherein the first gas mixture includes the first target compound and wherein the first membrane is equipped to let the first gas mixture traverse through the first membrane into the sensing cavity. The detector is equipped to detect the first target compound.

Claims

1. A gas analyzer for detection of a target compound in a complex gas mixture in the gas phase, the analyzer comprising a detector for detecting the target compound and a sensing cavity, wherein the detector is arranged in the cavity; the analyzer further comprises a first membrane being separate, wherein the first membrane is equipped to close the sensing cavity, and wherein the first membrane is equipped for simplifying the composition of the complex gas mixture into a first gas mixture, wherein said first membrane is capable of performing a chemical and a physical separation of composition of the complex gas mixture, wherein the first gas mixture comprises the first target compound and wherein the first membrane is equipped to let the first gas mixture traverse through the first membrane into the sensing cavity; wherein the detector is equipped to detect the first target compound.

2. The analyzer according to claim 1, wherein the first membrane comprises a pore size with a diameter of maximally 10 nm.

3. The analyzer according to claim 1, comprising an inlet and an outlet, wherein the inlet is arranged such that the complex gas mixture is guided onto the membrane closing the sensing cavity, and the first membrane is arrangeable between the inlet and the outlet, and wherein the outlet is arranged in an axial and/or non-axial orientation with respect to the inlet.

4. The analyzer according to claim 1, wherein the membrane is equipped to block the direct flow of the complex gas mixture into the sensing cavity.

5. The analyzer according to claim 1, wherein the sensing cavity comprises a detector housing with an opening, wherein the first membrane is arrangeable in the opening and/or the sensing cavity comprises a base part, wherein the first membrane is fastenable to the base part and builds the sensing cavity together with the base part in a fastened state.

6. The analyzer according to claim 1, further comprising a separate second membrane, wherein the first membrane is interchangeable with the second membrane, wherein the second membrane is equipped to close the sensing cavity, and wherein the second membrane is equipped for simplifying the composition of the gas mixture into a second gas mixture, wherein the second gas mixture comprises a second target compound and wherein the second membrane is equipped to let the second gas mixture traverse through the second membrane into the sensing cavity; wherein the detector is equipped to detect the concentration of the second target compound.

7. The analyzer according to claim 6, further comprising at least a first sensing cavity and a second sensing cavity, wherein the first membrane is equipped to close first sensing cavity and wherein the second membrane is equipped to close the second sensing cavity, wherein in the first sensing cavity a first detector is arranged and the first detector is equipped for detecting a first target compound and wherein in the second sensing cavity a second detector is arranged, wherein the second detector is equipped for detecting a second target compound.

8. The analyzer according to claim 7, wherein the second sensing cavity is arranged in a parallel manner and/or in a serial manner with the first sensing cavity.

9. The analyzer according to claim 6, wherein the first membrane and the second membrane are equipped to close the sensing cavity.

10. The analyzer according to claim 1, wherein the analyzer is equipped for selective detection of the concentration of the target compound in the complex gas mixture and wherein the detector is equipped for detecting the concentration of the target compound.

11. The analyzer according to claim 1, wherein the analyzer is equipped for detection of target compound in the gas phase at a trace level.

12. The analyzer according to claim 1, wherein the first membrane and/or the second membrane is closing the sensing cavity, wherein a transition between the membrane and the sensing cavity is sealed.

13. The analyzer according to claim 1, wherein the first membrane respectively the second membrane comprises a microporous material, with a pore size comparable to the size of the target compound.

14. The analyzer according to claim 1, wherein the detector is at least one of a chemoresistive detector, a mass spectrometer, and an optical system, and wherein the detector is selected from a group comprising: a doped SnO2 detector, in particular doped with Pt, Pd, Si, Ti; a WO3 detector, in particular a Si-doped WO3 detector, in particular a epsilon phase of Si-doped WO3 detector; a MoO3 detector, in particular a Si-doped MoO3 detector, in particular an alpha phase of Si-doped MoO3 detector; a ZnO detector, in particular a Ti-doped ZnO; a CuBr detector and a CuOSnO2 heterojunction.

15. The analyzer according to claim 1, wherein the target compound is a sulfuric compound, a ketone, a hydrocarbon, an aldehyde, an pnictogen hydride, an acid, an alcohol, and/or hydrogen.

16. The analyzer according to claim 1, wherein said analyzer is equipped for use in medical and/or biological fluid analysis; for use in breath analysis, in particular in lung cancer detection from exhaled breath; and/or for use in analysis of skin emissions, and/or for use in headspace analysis of fluids; and or for use in air quality analysis, in particular air quality monitoring of target compound released from indoor sources; and/or for use in food quality assessment, food processing control and/or monitoring; and/or for use in monitoring and/or controlling agricultural and/or chemical processes and products; and/or for use in environmental analysis or monitoring and/or controlling exhaust emission; and/or for use in detection of explosives and other hazardous compounds.

17. Use of the analyzer according to claim 1 for medical and/or biological fluid analysis; for breath analysis; for analysis of skin emissions; for headspace analysis of fluids; for air quality analysis; for food processing control; for food quality assessment; for air quality analysis, for indoor air analysis; for monitoring and/or controlling agricultural and/or chemical processes and products; and/or for environmental analysis; monitoring and/or controlling exhaust emission; and/or for detection of explosives and other hazardous compounds.

18. A method for detecting a target compound in a complex gas mixture comprising the step of: providing an analyzer according to claim 1; letting a complex gas mixture interact with the first mixture, respectively, a second membrane, wherein the first gas mixture traverses through the membrane into the sensing cavity; and wherein the detector detects the target compound.

19. The analyzer according to claim 1, wherein the first membrane comprises a pore size with a maximum diameter of 2 nm.

20. The analyzer according to claim 1, wherein the first membrane comprises a pore size with a maximum diameter of 1 nm.

Description

[0072] The subject matter of the invention will be explained in more detail in the following text with reference to exemplary embodiments which are illustrated in the attached drawings, in which:

[0073] FIG. 1a Gas analyser 1, as membrane 2,21,22detector 3 system;

[0074] FIG. 1b pore size of the membrane 2,21,22 and a support element;

[0075] FIG. 2 target compounds 64 with their kinetic diameter in comparison to the determined MFI pore size (kd);

[0076] FIG. 3 Film resistance (R) and normalized response S.sub.n of a detector 3 with and without a membrane 2 exposure to 1 ppm of acetone, NH.sub.3, ethanol, isoprene and formaldehyde;

[0077] FIG. 4 detector 3 resistance of the membrane 2,21,22detector 3 system upon exposure to 100, 70, 60 and 30 ppb of formaldehyde at 50% RH

[0078] FIG. 5 detector 3 calibration curve for formaldehyde in the range of 0-1000 ppb at 400 C. and 50% RH.

[0079] FIG. 6 sketch of an analyser 1

[0080] The reference symbols used in the drawings, and their meanings, are listed in summary form in the list of reference symbols. In principle, identical parts are provided with the same reference symbols in the figures.

EXAMPLE

[0081] A MFI/alumina membrane enables a non-specific Pd-dopedSnO2 detector to selectively detect formaldehyde (FA).

[0082] Membrane Fabrication

[0083] A MFI precursor solution is prepared as follows: 1.4 g sodium hydroxide (97%, Sigma-Aldrich) are dissolved in 100 ml tetrapropylammonium hydroxide (1 M TPA(OH) in H.sub.2O, Sigma-Aldrich) in a closed Teflon flask at room temperature. 20 g of fumed silica (Aerosil 200, Evonik) is added at 85 C. and dissolved under vigorous stirring and reflux. 3.2 mL of deionized water is added to the clear solution followed by a subsequent heating step to 105 C. for 15 min. The solution is cooled down within 45 min and aged at room temperature for 135 min. MFI powders are obtained by sedimentation using a centrifuge (Rotina 35, Hettich Lab Technology), flushing with deionized water and subsequent drying of the sediment.

[0084] The membrane 2,21,22 is made by placing up to four 16.3 mm0.5 mm porous and polished alumina disks as support 43 with the polished surface upwards in a 250 mL Teflon beaker. Such support 43 is made from alumina powder (CT 3000 SG, Almatis) that is pelletized at 30 kN hydraulic pressure (GS15011, Specac) and sintered for 30 h at 1150 C. in a furnace (Type 48000, Barnstead Thermolyne). The MFI synthesis solution is added and the Teflon beaker sealed in a stainless steel autoclave and heated for 8 h at 185 C. After rapidly cooling the autoclave down with tap water, the membranes are removed, washed with deionized water and stored overnight in a water bath at 50 C. Drying of the membranes is carried out at 50 C. for at least 3 days in an oven (KB53, Binder). The TPA structuring template is removed by heating the membrane 2,21,22 and powders to 450 C. for 6 h with heating and cooling rates of 30 C. h.sup.1. All experiments are conducted with template-free membranes and powders.

[0085] Membrane Characterization

[0086] The micropore size distribution of the zeolite powder is assessed by nitrogen sorption with a 3Flex (Micrometrics Instrument Corporation) in the pressure range of p/p.sub.0=4.5.Math.10.sup.70.047, where p and p.sub.0 represent the partial vapour pressure and saturated vapour pressure of the adsorbate gas, respectively of the complex gas mixture. The data is analyzed by the Horwath-Kawazoe method, that assumes cylindrical pore shape, which is consistent with the shape of MFI zeolite membrane 2,21,22. Prior to the analysis, the membrane 2,21,22 was activated (degassed) overnight at 250 C. The macropore size distribution of the alumina support 43 is determined from full nitrogen adsorption and desorption isotherms with a TriStar 3000 (Micromeritics Instrument Corporation) in the pressure range of p/p.sub.0=0.05-0.99. The data is analyzed according to the Barrett-Joyner-Halenda method. The membrane 2,21,22 is degassed under vacuum for 1 h at 150 C. prior to analysis. The film morphology of the membrane and sensing film is investigated by scanning electron microscopy (S-4800, Hitachi) operated at 3 kV.

[0087] Gas Detection

[0088] The chemoresistive gas detector 3 comprises flame-made (1 mol %) Pd-doped SnO.sub.2 nanoparticles directly deposited onto silicon wafer-based microsubstrates. The detector 3 is either combined with a membrane 2,21,22 or installed alone (for reference tests without membrane 2,21,22) in a sensing cavity 41 comprising stainless steel.

[0089] The gas mixing set-up for producing a complex gas mixture 6 is described for example in Sens. Actuators B 2016, 223, 266-273. The complex gas mixture 6 flow is 600 ml min.sup.1 with synthetic air (PanGas 5.0, C.sub.nH.sub.m and NO.sub.x100 ppb) as carrier gas that is humidified with a water bubbler to achieve 50% relative humidity (RH) forming the sensor baseline. The target compound 64, also called analyte gas, can be formaldehyde (FA) (10 ppm in N.sub.2, PanGas 5.0), acetone (10 ppm in syn. air, PanGas 5.0), ethanol (10 ppm in syn. air, PanGas 5.0), ammonia (NH.sub.3) (10 ppm in syn. air, PanGas 5.0), isoprene (10 ppm in syn. air, PanGas 5.0). For tests with TIPB (1,3,5-Triisopropylbenzene 95%, Sigma-Aldrich), the complex gas mixture 6 gas is obtained as follows: 5 ml of liquid TIPB is poured into a 50 mL wide neck flask. TIPB vapor is formed in its headspace that is purged with 100 mL min.sup.1 of synthetic air. That way, 18 ppm of TIPB in synthetic air/complex gas mixture 6 is obtained, as measured with a proton transfer reaction time-of-flight mass spectrometer (PTR-TOF-MS1000, Ionicon).

[0090] The detector 3 response S is defined as:

[00001]$S=\frac{R_{\mathrm{air}}}{R_{\mathrm{analyte}}}-1$

where R.sub.air and R.sub.analyte represent the film resistances in absence and presence of the analyte, respectively. Response and recovery times are defined as the time needed to reach and recover 90% of the response resistance change, respectively. Signal-to-noise ratio (SNR) is defined as the ratio of the signal R=R.sub.airR.sub.analyte to the standard deviation of the noise.

[0091] Membrane 2,21,22Detector 3 System for Detection of a Target Compound 64 (Gas Analyser 1)

[0092] The gas analyser 1, as membrane 2,21,22detector 3 system, is illustrated schematically in FIG. 1a. The complex gas mixture 6, also called real gas mixture, (e.g. indoor air (Indoor Air 1994, 4 (2), 123-134) or breath (J. Breath Res. 2014, 8 (1), 014001)) comprises a large number of different molecules respectively compounds. These can be separated respectively simplified with a membrane 2,21,22, for example a zeolitic membrane 2 as described above. Ideally, only a single target compound 64 (medium sized ball) can permeate respectively traverse through the membrane 2,21,22 while other interfering compounds of the complex gas mixture 6 (small and big balls) are held back. Eventually, the target compound 64 is detected with a chemoresistive detector 3 placed after the membrane 2,21,22 in a sensing cavity 41 and the concentration of the target compound 64 is deduced from the detector's 3 electrical resistance change (i.e. response). The outer wall of the sensing cavity 41 is not shown.

[0093] The membrane 2,21,22 is coin-type and rather small (d=16.3 mm) and comprises a compact and coherently grown 3 m MFI membrane 2,21,22 that is supported on a porous Al.sub.2O.sub.3 support 43. The pore diameter distribution for the MFI membrane 2,21,22 (left side; d.sub.p<1 nm) and -Al.sub.2O.sub.3 support 43 (right side; d.sub.p>10 nm) are shown in FIG. 1b: MFI micropores were measured to be in a size range of 0.57 to 0.61 nm, slightly larger than silicalite (alumina-free MFINature 1978, 271, 512-516). In contrast, the -Al.sub.2O.sub.3 support pellet 43 has pores>40 nm, significantly larger than most target compounds 64. FIG. 1d depicts the pore diameter d.sub.p in nm (nano meter) as a function of the normalized pore volume V for the MFI membrane 2 and the Al.sub.2O.sub.3 support element 43.

[0094] The applied detector 3 is a chemo-resistive type and consists of a semiconductive metal-oxide film on Si wafer-based micodetector 3 substrates with a size smaller than a match head. In particular, the sensing film is formed by flame-made Pd-doped S.sub.nO2 nanoparticles that aggregate to a highly porous network providing high surface area available for target compound 64 interaction. Such flame-made and nanostructured Pd-doped SnO2 sensors (without membrane) are highly sensitive and can detect, for instance, formaldehyde FA down to 3 ppb (at 90% RH) with typical response times of 2 min and always complete recovery during continuous application (ACS Sens. 2016, 1 (5), 528-535).

[0095] The sensing cavity 41, membrane 2,21,22 and detector 3 are decoupled (mechanically and thermally) and thus can be combined flexibly and operated independently. Due to their compact and modular design, such membrane 2,21,22detector 3 systems (analyser 1) can be easily incorporated into analysers, e.g. compact indoor air monitors or portable breath analyzers, and they feature low sensor power consumption of 500 mW at 400 C.

[0096] MFI/Al2O3 Membrane 2,21,22 Turns Detector 3 Formaldehyde-Selective

[0097] To evaluate the membrane effect on the detector 3, various target compounds 64 are tested covering a wide range of physical and chemical properties. FIG. 2 lists the chosen target compounds 64 with their kinetic diameter (kd) in comparison to the determined MFI pore size (ps). 1,3,5-TIPB (TIPBdiagonal shaded) represent a symmetric molecule being larger than the pore size of the MFI-membrane 2 (shown as diamond outlined pattern) and that way, the size filtering effect of the membrane can be assessed. Formaldehyde (FAvertical lines), isoprene (Isoplarge squared), acetone (Acsmall squared), ethanol (EtOHdiagonal lines) and ammonia (NH.sub.3horizontal lines) are smaller than the membrane 2 pores and they are selected due to their different functional groups, introducing a diversity of chemical properties, and their relevance for indoor air monitoring (Indoor Air 1994, 4, 123-134) and breath analysis (Breath Res. 2014, 8, 014001).

[0098] FIG. 3a shows the change in detector 3 resistance (R) of the chemoresistive Pd-doped (1 mol %) SnO2 detector 3 without membrane upon exposure to 1 ppm of acetone (dotted lineAc), Ammonia (dash dotted lineNH.sub.3), ethanol (long dash lineEtOH), isoprene (dashed lineIsop) and formaldehyde (solid lineFA) at 400 C. and 50% RH. When injecting FA, the detector 3 resistance drops rapidly from 128 to 10.2 k, corresponding to a response of 11.3. Also the other target compounds 64 are detected clearly and the corresponding responses (normalized to maximum response of Formaldehyde (FA)S.sub.nnormalized response) are shown in FIG. 3b. This indicates the rather non-specific character of Pd-doped SnO.sub.2 and thus it cannot detect formaldehyde (FA) selectively in a complex gas mixture 6 as it is not possible to distinguish it from interfering gases (e.g. acetone).

[0099] This problem is solved when adding the membrane 2,21,22. Indeed, when exposing the detector 3 now to the different target compounds 64, only formaldehyde (FA) is detected (FIG. 3c, and FIG. 3d), so the membrane 2,21,22 turns the non-specific Pd-doped SnO2 detector 3 formaldehyde-selective. Thus the detector 3 does hardly respond to other gases. More specific, the FA selectivity to acetone is improved to >100 and the one to NH3, isoprene, ethanol and TIPB is even >1000, much higher than without membrane 2,21,22. So it seems that FA can permeate through the MFI/Al2O3, similar as observed before (J. Membr. Sci. 2004, 240 (1), 159-166), while other compounds are held back by the membrane. Actually, TIPB should be separated due to its larger molecular size compared to the distinct MFI pore diameter range (FIG. 2). This size cut-off is rather important for indoor air monitoring and breath analysis since both gas mixtures contain a myriad of such larger molecules potentially interfering with the sensor. The MFI layer should filter out all of them similarly efficient as TIPB. In case of other compounds smaller than the size cut-off (i.e. isoprene, NH.sub.3, acetone and ethanol), these might be separated due to their different sorption and diffusion properties.

[0100] Size cut-off for TIPB represents many other unknown gases with larger size than the pore size.

[0101] The calculated selectivities (S.sub.FA/S.sub.target) for the Pd-doped SnO2 detector 3 with and without membrane are shown in Table 1 along with other state-of-the-art FA gas detectors: while the Pd-doped SnO2 detector 3 features rather weak selectivity, this is dramatically improved with membrane 2,21,22. Actually, the selectivity to acetone is >100 while the one to NH3, isoprene and ethanol is even >1000. This is also superior to other metal-oxide sensors, such as Ag-doped LaFeO3 (J. Mater. Chem. C 2014, 2 (47), 10067-10072) or TiO2 nanotubes (Sens. Actuators B 2011, 156 (2), 505-509) that had been proposed as FA sensors.

[0102] In Table 1, the FA selectivity of the membrane-sensor system (analyser 1) is benchmarked with other state-of-the-art FA gas detectors. Various chemoresistive gas detectors, including metal-oxides (e.g. Ag-doped LaFeO.sub.3J. Mater. Chem. C 2014, 2 (47), TiO.sub.2 nanotubesSens. Actuators B 2011, 156 (2), 505-509) and metal-organic frameworks (e.g ZIF-67Inorg. Chem. 2014, 53 (11), 5411-5413) had been developed for high FA selectivity. However, all are outperformed clearly by the present membrane-sensor system 1. While this indicates the immediate impact of the novel membrane approach, it shows also the limitations of sensor material optimization.

TABLE-US-00001 TABLE 1 Selectivity of formaldehyde detectors Formaldehyde selectivity S.sub.Formaldehyde/S.sub.x [] Type Material Ammonia Ethanol Acetone Isoprene chemo-resistive MOx sensor + Pd:SnO.sub.2 + >1000 >1000 >100 >1000 membrane MFI/Al.sub.2O.sub.3 MOx sensor only Pd:SnO.sub.2 12 3 26 1.4 Ag:LaFeO.sub.3 35 27 50 TiO.sub.2 nanotubes 12 .sup.57.sup.a ZIF sensor ZIF-67 43 .sup.2.sup.b 2 Coated sensor ZIF-8 coated ZnO 5 7 11 optical Photoelectric Colorimetric sensor high 8 photometry Fiber-optic NADH based flow cell high high .sup.alinearly interpolated to same concentrations; .sup.bselectivity to methanol

[0103] In another study, zeolitic ZIF-8 had been applied directly as coating onto ZnO to pre-filter molecules (ACS Sens. 2016, 1 (3), 243-250), similar to the membrane here. While improved FA selectivity was observed, a distinct size cut-off was not obtained (ACS Sens. 2016, 1 (3), 243-250). In fact, larger molecules than the ZIF-8 pore size (0.34 nm), e.g. ethanol and acetone, were detected by the sensor (ACS Sens. 2016, 1 (3), 243-250) resulting in significantly lower FA selectivity compared to the membrane-sensor (analyser 1) configuration here (Table 1). This is probably associated to the high temperature of the coating (300 C.) that is thermally coupled to the ZnO and thus heated as well. This could lead to catalytic fragmentation of larger molecules on the ZIF-8 surface and that way, smaller product molecules can enter the ZIF-8 pores undermining the desired size cut-off. So the thermal decoupling of the present membrane-sensor approach is rather important allowing independent operation of the both to avoid such unwanted effects.

[0104] Finally, FA as a target compound 64 can be detected also by optical sensors and these achieve rather high selectivity with respect to ethanol (Table 1) while their performance for NH3 and isoprene is unknown (FP-30 RKI Instruments, Biosens. Bioelectron. 2010, 26 (2), 854-858). However, while the commercial FA detector FP-30 (RKI Instruments) features insufficient selectivtiy to acetone (Biosens. Bioelectron. 2010, 26 (2), 854-858), fiber-optic sensors that require enzymes for irreversible reaction with FA (Biosens. Bioelectron. 2010, 26 (2), 854-858) might have rather limited life-time. In case of the later, these enzymes are continuously consumed and depleted at some point, as observed for similar devices with a performance deterioration of 80% after 6 days (Anal. Chem. 1994, 66 (20), 3297-3302).

[0105] Response and Recovery Times

[0106] Fast response and recovery times are desirable properties for gas sensors. While the Pd-doped SnO2 detector 3 without membrane 2,21,22 possesses response and recovery times of 1 and 9 min, respectively, introducing the membrane delays them to 8 and 72 min. An increase is expected from gas diffusion theory, as the membrane 2,21,22 and, in particular, its dense and microporous MFI layer represents an additional permeation barrier. However, this could be minimized by reducing the MFI-membrane 2,21,22 layer thickness. In principle, maximal molecular diffusion would be obtained when the layer thicknesses approaches the dimensions of a zeolite's single unit cell. Previous studies had demonstrated the synthesis of 2 nm thick MFI nanosheets (Nature 2009, 461, 246-249), much thinner than the 3 m MFI layer applied here.

[0107] Lower Limit of Detection and Calibration Curve

[0108] The detection of formaldehyde FA levels below 100 ppb is crucial for indoor air monitoring to distinguish normal from hazardous conditions (Crit. Rev. Toxicol. 2011, 41 (8), 672-721) and also in medical diagnostics where typical breath formaldehyde (FA) concentrations in lung cancer patients and healthy humans are smaller (Int. J. Cancer 2010, 126 (11), 2663-2670). FIG. 5 shows the detector 3 resistance [R (k)] change over time [t (min)] to 100, 70, 60 and 30 ppb of formaldehyde FA with the membrane 2,21,22 at 50% RH and 400 C. When exposed to 100 ppb, the resistance R drops rapidly from 35.1 to 25 k corresponding to a response of 0.4 within a response time of <14 min. Most notably, formaldehyde FA levels down to 30 ppb can be detected with a high signal-to-noise ratio (SNR, >80), which is sufficient for breath analysis and indoor air monitoring. Actually, the extrapolated lower limit of detection (LOD) for formaldehyde FA is 0.2 ppb at a SNR=1. This is rather comparable to single Pd-doped SnO.sub.2 (0.1 ppb at 90% RHACS Sens. 2016, 1 (5), 528-535) indicating the little interference of the membrane 2,21,22, and superior to zeolite-coated ZnO (5.6 ppm 50-60% RHACS Sens. 2016, 1 (3), 243-250). Additionally, the membrane 2,21,22detector 3 system (analyser 1) has excellent formaldehyde (FA) resolution, in fact, the responses for 60 to 70 ppb are clearly distinguishable.

[0109] Another important feature for a gas detector 3 is repeatable usability. As shown in FIG. 5, the membrane 2,21,22detector 3 system (analyser 1) always fully recovers the initial baseline when flushed with air enabling repeatable exposure to formaldehyde (FA). This indicates reversible and fast FA permeation through the membrane 2,21,22 and interaction with the detecting structure without any observable deactivation. These detector responses are also stable and nicely reproducible. In fact, when exposing the detector 3 twice to 60 ppb of formaldehyde (FA), the same response resistance is achieved (dashed line, FIG. 5). Reproducible responses are also obtained during continuous operation for several days (results not shown here) without any observable degradation. Similar flame-made Pt-doped SnO.sub.2 had shown stable detector 3 performance during 20 days of continuous operation at 10% RH (J. Nanopart. Res. 2006, 8 (6), 783-796) and also MFI membranes 2,21,22 maintained constant selectivity and permeance for at least 5 days even with concentrated gas mixtures (Microporous Mesoporous Mater. 2014, 192, 76-81). While these first results are promising, long-term stability still needs to be investigated.

[0110] The detector 3 calibration curve (at 400 C.) for formaldehyde (FA64) in the range of 0-1000 ppb at 50% RH is shown in FIG. 4 (triangles). FIG. 4 depicts the detector 3 calibration curve for formaldehyde in the range of 0-1000 ppb at 400 C. and 50% RH (triangles). FIG. 4 illustrates the detector 3 response S versus the formaldehyde (FA) concentration c (ppb). Interestingly, this calibration curve does not change even when introducing 1 ppm of NH.sub.3, acetone, isoprene and EtOH, all at the same time (squares) highlighting the excellent formaldehyde selectivity of the membrane 2,21,22detector 3 system (gas analyser 1). That way, hazardous FA levels above the recommended indoor air limit (100 ppbACrit. Rev. Toxicol. 2011, 41 (8), 672-721) and eye irritation threshold (500 ppbBRegul. Toxicol. Pharmacol. 2008, 50 (1), 23-36) can be detected rapidly. The responses increase continuously with increasing FA concentration and this allows to distinguish them clearly. That way, and most importantly, FA levels exceeding the recommended indoor air limit of 100 ppb (Crit. Rev. Toxicol. 2011, 41 (8), 672-721) and the eye irritation threshold at 500 ppb (Regul. Toxicol. Pharmacol. 2008, 50 (1), 23-36) can be rapidly recognized to protect from potential cancer risks and sensory impairment.

[0111] Selectivity in Gas Mixtures

[0112] Indoor air and breath consist of complex gas mixtures 6 and some target compounds 64 may be present at even higher concentrations than formaldehyde (FA). FIG. 4 (squares) shows the detector 3 calibration curve to formaldehyde (FA) when additional 1 ppm of NH.sub.3, acetone, isoprene and ethanol are introduced at the same time. Remarkably, the calibration curve for FA (even at 30 ppb) does not change despite the four interfering compounds at substantially higher concentrations. This emphasizes the excellent separation properties of the membrane 2,21,22 being superior to E-noses that can trace FA in comparable gas mixtures as well, however, with an estimation error of 9 ppb (ACS Sens. 2016, 1, 528-535).

[0113] Further embodiments are evident from the dependent patent claims. Features of the method claims may be combined with features of the device claims and vice versa.

[0114] While the invention has been described in present preferred embodiments of the invention, it is distinctly understood that the invention is not limited thereto, but may be otherwise variously embodied and practised within the scope of the claims.

[0115] FIG. 6 shows a schematic sketch of an analyser 1 with a sensing cavity 41, a membrane 2,21,22 with support element 23 and a detector 3 within the sensing cavity 41. The analyser 1 further comprises an inlet 51 and an outlet 52. The separate membrane 2,21,22 is closing the cavity 41. A transition between the membrane 2,21,22 and the cavity 41 is sealed, by an O-ring,