Acetone sensor and method

Abstract

A sensor for measuring acetone, including: an array of nanowires extending from a substrate, each of the nanowires having a corresponding surface composed substantially of a compound semiconductor, and a diameter less than 100 nm; an electrical contact material disposed on a first portion of the surface of each nanowire to form a Schottky barrier contact therewith; an amine-functionalised polymer disposed on the nanowires to functionalise a second portion of the surface of each nanowire; and an electrical contact to the substrate to enable measurement of electric current flowing between the electrical contact and the Schottky barrier contacts on the nanowires when the nanowires are exposed to light or biased, such that, when the nanowires are exposed to a mixture of molecular species including acetone, the measured electric current is representative of a concentration of acetone in the mixture.

Description

[0001] ACETONE SENSOR AND METHOD

[0002] TECHNICAL FIELD

[0003] The present invention relates to sensing, and in particular to a sensor for measuring acetone, and a method of producing a sensor for measuring acetone.

[0004] BACKGROUND

[0005] Diabetes mellitus is a life-threatening condition that affects a significant portion of the population. Currently, people with diabetes are required to self-monitor their blood sugar levels at least twice per day by pricking one of their fingers to produce a drop of blood, and placing the drop on a test strip of a device such as an electronic glucometer to detect dangerously high levels of sugar in the blood. This procedure is inconvenient, invasive and unpleasant. However, low levels of insulin can cause a dangerous condition known as diabetic ketoacidosis ("DKA"), in which the body breaks down fat instead of glucose, producing ketones which can damage body organs at high levels. These ketones can be detected in urine, blood and breath. In particular, recent studies have shown that the level of ketones in blood is correlated with the level of acetone in breath, suggesting that measurements of breath acetone could be a viable alternative to daily blood tests.

[0006] For example, the presence of acetone in exhaled breath is a distinctive biomarker for people with diabetes (type 1 or type 2), or those undertaking a 'keto' diet. However, prior attempts to develop acetone breath sensors suffer from a number of difficulties, in particular poor performance such as poor sensitivity, selectivity, and dynamic range, and high operating temperatures.

[0007] It is desired to alleviate one or more difficulties of the prior art, or to at least provide a useful alternative. SUMMARY

[0008] In accordance with some embodiments of the present invention, there is provided a sensor for measuring acetone, including: an array of nanowires extending from a substrate, each of the nanowires having a corresponding surface composed substantially of a compound semiconductor, and a diameter less than 100 nm; an electrical contact material disposed on a first portion of the surface of each nanowire to form a Schottky barrier contact therewith; an amine-functionalised polymer disposed on the nanowires to functionalise a second portion of the surface of each nanowire; and an electrical contact to the substrate to enable measurement of electric current flowing between the electrical contact and the Schottky barrier contacts on the nanowires when the nanowires are exposed to light or biased, such that, when the nanowires are exposed to a mixture of molecular species including acetone, the measured electric current is representative of a concentration of acetone in the mixture.

[0009] In some embodiments, each of the nanowires has a diameter of about 50 to 60 nm.

[0010] In some embodiments, the amine-functionalised polymer is chitosan.

[0011] In some embodiments, the electrical contact material is composed substantially of a metal selected from the group consisting of Pt, Pd, Ni, and Au. In some embodiments, the electrical contact material is composed substantially of Pt.

[0012] In some embodiments, the sensor is self-powered when exposed to light.

[0013] In some embodiments, the compound semiconductor is a compound semiconductor selected from the group consisting of: InP, InAs, GaP, and GaAs. In some embodiments, the compound semiconductor is un-doped InP.

[0014] In some embodiments, the sensor has a response time < 20s, and a detection limit < 0.1 ppm. In some embodiments, the sensor is flexible and a component of a wearable sensing device. The mixture of molecular species may be in a gas phase or a liquid phase.

[0015] In accordance with some embodiments of the present invention, there is provided a handheld acetone measuring device, including any one of the above sensors. In some embodiments, the handheld acetone measuring device includes a light-emitting diode configured to emit light onto the sensor to cause the electric current to flow between the electrical contact and the Schottky barrier contacts. The device may be configured to diagnose diabetes and ketoacidosis from human breath.

[0016] In accordance with some embodiments of the present invention, there is provided a method of producing a sensor for measuring acetone, the method including the steps of: providing an array of nanowires extending from a substrate, each of the nanowires having a corresponding surface composed substantially of a compound semiconductor and a diameter less than 100 nm; depositing an electrical contact material on a first portion of the surface of each nanowire to form a Schottky barrier contact therewith; depositing an amine-functionalised polymer onto a second portion of the surface of each nanowire; and forming an electrical contact to the substrate to enable measurement of electric current flowing between the electrical contact and the Schottky barrier contacts on the nanowires when the nanowires are exposed to light or biased; wherein, when the nanowires are exposed to a mixture of molecular species including acetone, the measured electric current is representative of a concentration of acetone in the mixture.

[0017] In some embodiments, the step of providing the array of nanowires includes growing the array of nanowires from the substrate by an additive bottom-up growth process.

[0018] In some other embodiments, the step of providing the array of nanowires includes forming the array of nanowires by a top-down etching process. In some embodiments, the step of depositing an amine-functionalised polymer includes drop-casting chitosan onto the nanowires.

[0019] In some embodiments, the step of depositing an electrical contact material on a first portion of the surface of each nanowire includes tilt-angle deposition of the electrical contact material onto the nanowires so that the electrical contact material is deposited onto only one side of each nanowire.

[0020] In some embodiments, the electrical contact material is composed substantially of Pt.

[0021] In some embodiments, each of the nanowires has a diameter of about 50 to 60 nm.

[0022] In some embodiments, the compound semiconductor is a compound semiconductor selected from the group consisting of: InP, InAs, GaP, and GaAs. In some embodiments, the compound semiconductor is a undoped InP.

[0023] In some embodiments, the electrical contact material is composed substantially of a metal selected from the group consisting of Pt, Pd, Ni, and Au.

[0024] BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Some embodiments of the present invention are hereinafter described, by way of example only, with reference to the accompanying drawings, in which:

[0026] Figure 1 (a) includes a schematic illustration of an acetone sensor in accordance with an embodiment of the present invention, including an array of InP nanowires ("NWs") functionalised with chitosan and being used to measure acetone in the exhaled breath of a human subject by measuring electric current using a digital electrometer and displaying a graph of the measurements as a function of time on a computer display; scanning electron microscope ("SEM") images of the sensor are also shown;

[0027] Figures 1 (b) to (f) are a set of schematic diagrams and accompanying SEM images illustrating an acetone sensor production method in accordance with an embodiment of the present invention;

[0028] Figure 2 is a schematic diagram illustrating the bottom-up growth of InP nanowires through openings in a patterned SiO2 mask layer on an InP wafer substrate;

[0029] Figure 3 (a) is a schematic diagram illustrating testing of the acetone sensor; Figures 3 (b) to (h) are graphs showing measurements of the acetone sensor using the arrangement shown in (a);

[0030] Figure 4 includes two graphs of the response and recovery times of the acetone sensor as a function of acetone concentration, for acetone concentration ranges of 0 to 10 ppb, and 100 to 1000 ppb, respectively;

[0031] Figure 5 includes graphs illustrating the stability and reliability of the acetone sensor over time, and for acetone concentrations of 0.1, 2 and 10 ppm;

[0032] Figures 6 and 7 are sets of graphs illustrating the performance of acetone sensors under various conditions, including other ketone groups, and a range of relative humidities;

[0033] Figure 8 include schematic and SEM images of an acetone sensor during production by a top-down production method;

[0034] Figure 9 includes a schematic illustration of the acetone sensor produced by the top-down production method, and graphs demonstrating the performance of the resulting sensor;

[0035] Figure 10 is a circuit diagram showing the components of a portable handheld electronic acetone sensing device for sensing acetone in exhaled human breath, and including the acetone sensor; and

[0036] Figure 11 illustrates the use of the portable handheld electronic acetone sensing device to detect acetone in exhaled human breath, and performance measurements thereof.

[0037] DETAILED DESCRIPTION

[0038] In order to address shortcomings of existing acetone sensors, the inventors have developed miniaturised acetone sensors with high sensitivity and high selectivity sensing of acetone at room temperature. Effective breath detection for diabetes monitoring requires the detection of acetone concentrations in the range of 0.3-1.8 ppm. As described below, embodiments of the present invention are able to detect acetone in breath over an extremely wide concentration range, from < 1 ppb to > 105ppm, indicating the capability to detect acetone levels ranging from levels characteristic of low level diabetic on-set to levels characteristic of dangerous DKA, where the acetone concentration in breath can be higher than 75 ppm, and with high selectivity and a short response time of ~ 5s. The sensor can also be transferred to a flexible substrate to facilitate integration into a wearable device. In the described embodiments, the sensor is also self-powered, enabling continuous monitoring without requiring a power supply.

[0039] As shown in Figure 1, an acetone sensor includes an array of compound semiconductor nanowires extending from a substrate, each of the nanowires having a diameter less than 100 nm. An electrical contact material disposed on a first portion of the surface of each nanowire forms a Schottky barrier (or Schottky junction) contact, and an amine- containing polymer is deposited on the nanowires to functionalise a second portion of the surface of each nanowire. By forming an electrical contact to the common substrate, the processed nanowires effectively constitute a set of photovoltaic cells electrically connected in parallel, and electric current flows between the substrate contact and the Schottky barrier contacts on the nanowires when the nanowires are exposed to light, or when the Schottky junctions are biased. The electric current constitutes the sensor signal baseline under light illumination. Due to the amine-functionalised nanowire surfaces, the current changes upon exposure to acetone. In particular, when the nanowires are exposed to a mixture of molecular species including acetone, the measured variation of electric current is linearly correlated to the concentration of acetone in the mixture, and can therefore be calibrated to quantitatively measure that concentration.

[0040] In the described embodiments, the nanowires are composed substantially of the compound semiconductor indium phosphide (InP); however, in other embodiments the nanowires may be composed of other compound semiconductors, such as GaAs, InAs, or GaN, for example. The inventors found that the sensitivity and selectivity of the sensor to acetone are particularly high when the electrical contact material is platinum, the nanowires are composed of undoped InP, the amine-functionalised polymer is chitosan, and the nanowires are self-powered by exposing them to light (rather than being biased by an external power source). However, as described below, any or all of these parameters can be changed in other embodiments. Figure 1 is a composite figure illustrating an acetone sensor and a method for producing an acetone sensor, in accordance with respective embodiments of the present invention. Part (a) at the top of the Figure shows a schematic representation of the sensor being exposed to light and to acetone from the breath of a human subject. In the described embodiments, the amine-functionalised polymer is chitosan, whose chemical structure is shown to the left. However, it will be apparent to those skilled in the art that other amine-functionalised polymers may be used in other embodiments. The two scanning electron microscope (SEM) images to the right respectively show the entire nanowire array in plan view (and prior to the addition of chitosan), and a portion of the array at a higher magnification, and at a 30° tilt angle.

[0041] The acetone sensors described herein were and can be produced by a production method as described below, and as illustrated in Figures 1 (b) to (e), although it will be apparent to those skilled in the art that the acetone sensors described herein can also be produced by alternative production methods. In the described embodiments of the production method, an array of InP nanowires with a diameter less than 100 nm is fabricated. In the described embodiments, this is achieved by way of a bottom-up additive processing step, specifically involving selected-area epitaxial growth of InP nanowires on an InP substrate by way of metal-organic chemical vapor phase epitaxy ("MOVPE", also known as metalorganic chemical vapour deposition, or "MOCVD") on mutually spaced regions of an InP wafer exposed through respective openings in an electron-beam patterned SiO2 masking layer deposited on the InP wafer, as shown schematically in Figure 2.

[0042] In addition to forming correspondingly small openings in the masking layer (e.g., using electron-beam lithography), a key to growing thin (< 100 nm) high aspect ratio nanowires is reducing radial growth, and this can be achieved by growing nanowires having a Wurtzite ("WZ") crystal structure on a single-crystal substrate wafer, which results in predominantly axial growth and negligible radial growth, and defect and taper free nanowires. In the described embodiments, Si-doped InP wafers of (lll)A orientation are used.

[0043] In work leading up to the invention, the inventors investigated the effect of nanowire diameter on the sensitivity of the resulting sensors, and found that the sensitivity was poor for nanowire diameter greater than 100 nm. Accordingly, in the described embodiments, nanowire diameters of 50 to 60 nm are produced, as these provide high sensitivity, as described further below. The spatial dimensions of the nanowire arrays are 400 x 400 pm.

[0044] Once the array of nanowires has been produced, as shown in Figure 1 (a), a photoresist is deposited onto the InP substrate to mask it from a subsequent metal contact deposition step. The photoresist also has the benefit of providing mechanical stability to the nanowires. In the described embodiments, the photoresist is SU8-5. Because the photoresist also covers the nanowires, a controlled etching step is subsequently used to remove an upper portion of the photoresist layer and thereby expose a corresponding (and tuneable) portion of the nanowires, leaving the remaining portion of the nanowires and the substrate coated with the photoresist, as shown in Figure 1 (c). In the described embodiments, this is achieved by plasma etching. However, alternative etching methods may be used in other embodiments.

[0045] A metal contact is then formed on only a portion of each exposed nanowire portion to form a Schottky electrical contact on a portion of the nanowire, while leaving the remaining exposed portion of the nanowire for subsequent functionalisation. In the described embodiments, this electrical contact is achieved by angled shadow deposition of a suitable contact metal. That is, the contact metal is deposited at an angle relative to the substrate, with the angle selected such that the contact metal is deposited only on one side of each nanowire, the other side of the nanowire being shadowed from the metal during deposition. The contact metal is also deposited on the exposed planar surface of the remaining photoresist layer, interconnecting the nanowires, and resulting in the structure shown in Figure 1 (d). It can be seen from the SEM image that the deposition of metal on only one side of each nanowire causes the nanowires to slightly bend.

[0046] In work leading up to the invention, the inventors investigated different contact metals to form the Schottky barrier with the InP nanowires, including platinum (Pt), Aluminium (Al) and gold (Au). However, it was found that the choice of metal significantly affects the sensitivity of the resulting sensors, and that platinum provided the best performance. Consequently, platinum is used as the contact metal in the described embodiments.

[0047] An amine-functionalised polymer (chitosan in the described embodiments) is then deposited onto the nanowire array and Schottky contacts. The chitosan is diluted in an acetic acid aqueous solution with a concentration of 2.5% (diluted at 1 : 10 ratio from the saturated solution, concentration 28%). In the described embodiments, the deposition is achieved by drop-casting, followed by baking on a hotplate at 50 °C to remove the solvent, which the inventors found gave better results than spin coating. However, alternative deposition methods, dilute mixtures and concentrations may be used in other embodiments.

[0048] As shown in the SEM image and schematic illustration of Figure 1 (e), the drop-casting produces a film-like layer of chitosan over the nanowire array, with the chitosan layer covering or embedding an upper portion of the nanowires. In some embodiments, this porous hetero-structure is further configured to enhance the sensor performance for gas exposure; for example, by fine tuning the molecular weight of chitosan and its degree of acetylation, tuning the chitosan layer thickness via alternative deposition methods such as spin-coating, and / or using alternative chemicals containing a -NH2 functional group for functionalization, as a substitute for chitosan.

[0049] As shown in Figure 1(f), a second electrical contact (Ti / Au in the described embodiments) is formed on the InP wafer substrate, and when the Schottky barrier contacts on the nanowires are exposed to light, an electric current flows between the second contact and the Schottky barrier contacts on the nanowires in a circuit, as the nanowire array is a form of photovoltaic cell. Although this 'self-powered' mode of operation is particularly convenient and provides the best sensing performance, the sensor can also be operated in the dark by applying a bias voltage across the contacts, albeit at the expense of reduced sensitivity to acetone concentration, as described below.

[0050] Regardless of whether operated under illumination or in darkness, when the nanowires are exposed to a mixture of acetone and simulated air or breath, the interaction of acetone with the chitosan and the InP nanowires changes the electric current in dependence on the concentration of acetone in the gas mixture. Accordingly, measurements of the resulting electric current variations are representative of the concentrations of acetone, and the measurements can be calibrated to provide a direct quantitative measure of the concentration of acetone in a gas mixture. In the context of diabetes monitoring, the sensor can be used to detect different levels of acetone in breath, including dangerous levels of acetone in the breath of a diabetic patient. The production method described above is a type of method known in the art as a 'bottom-up' method because the nanowires are incrementally grown from the substrate. In an alternative embodiment, a top-down production method forms the nanowires by masked etching of an InP substrate (e.g., using an inductively-coupled plasma (ICP) etching tool) to remove selected portions of the substrate to form high aspect ratio InP nanowires from the remaining unetched regions. An advantage of the top-down (subtractive) production method is that it is likely to be lower cost than bottom-up (additive) production methods based on epitaxial growth.

[0051] Although embodiments and examples of the present invention are described herein primarily in the context of human breath sensing, the described sensors are equally applicable to detecting acetone in the breath of other animals, including pets such as dogs, cats and cows, for example, where diabetes is also a significant problem. Additionally, the described sensors can also be used to detect acetone in fluids, including detecting acetone in sweat.

[0052] EXAMPLE I

[0053] The following describes examples of the production method and resulting sensors, and measurements of acetone sensing performance and other characteristics of the sensors.

[0054] Substrate Preparation

[0055] In one example, a 30 nm SiO2 layer was deposited on an n-type (Si doped, 1018-1019cm-3, resistivity 0.6-6 x 10'3Q-cm) (lll)A InP substrate by plasma enhanced chemical vapor deposition (PECVD) at 300 °C; the thickness was measured by ellipsometry. A negative photoresist AR6200.09 was spin-coated (step 1 : 500 rpm 5 s; step 2: 2000 rpm for 60 s) on the SiO2 layer, and baked at 150°C for 1 min on a hotplate. The baked resist was patterned using a Raith 150 electron beam lithography (EBL) system. The pattern defining the desired nanowire layout was a 400 x 400 pm square array of hexagonal dots with a diameter of 40 nm and a pitch of 600 nm. The regions of the photoresist layer modified by the electron beam were removed by the corresponding photoresist developer, and an oxygen plasma (300 W, 2 min, 300 seem O2 flow) was used to remove any residues from those regions. The exposed regions of the SiO2 layer were then removed from the InP substrate by Reactive Ion Etching (RIE) (20 seem CHF3, RF power: 20W, 4.5 min) to pattern the SiO2 layer and thereby define a masking layer for the subsequent nanowire growth process. Na no wire Growth

[0056] To provide a clean and smooth exposed InP surface for nanowire growth, a 10% H2O2 solution was used to oxidize the newly exposed regions of the InP surface for 2 min, and the resulting oxide layer was stripped using a 10 % H3PO4 solution for 2 min. These two steps were repeated sequentially 5 times, and immediately after the final etch the sample was transferred into a MOVPE reactor to grow the nanowires epitaxially from the regions of the InP substrate exposed through the patterned SiO2 masking layer.

[0057] InP nanowires were grown using an AIXTRON 200 / 4 MOVPE reactor operating at a base pressure of 100 mbar, using H2 as a carrier gas at a total flow rate of 14.5 litres / min. Trimethyl-indium (TMIn) and phosphine (PH3) were used as precursors for the group III (In) and group V (P) elements, respectively. The molar fractions of TMIn and PH3 were set at 9.38xl0'6and 7.59xl0'4, respectively, corresponding to a V / III ratio of 80. The sample was hot baked at 750 °C for 10 min under a PH3 protective flow, and InP nanowires were grown for 4 min at 730 °C. Under such MOVPE growth conditions, even without the introduction with any dopants, nominally undoped InP nanowires ("un-InP") are known to be lightly n-type due to background doping.

[0058] Acetone sensor fabrication

[0059] As described above, a sensor was produced from the nanowire array by forming a Schottky barrier contact on a first portion of each nanowire and functionalising a second portion of the nanowire.

[0060] To achieve the above, SU8-5 photoresist was spin-coated onto the sample to fully cover the nanowire array. To protect the thin nanowires from damage during the fabrication process, a low spin speed of 1000 rpm was applied, followed by a two-step soft baking process at 65 °C and 95 °C for two minutes each to remove the solvent from the photoresist film.

[0061] The SU8-5 film was then partially removed by etching using a barrel-etcher (PVA Tepla Gigabatch 310 M) with an O2 flow rate of 300 seem and a power of 500 W to expose the top ~500 nm of the nanowires. Then the sample was flood exposed to UV illumination and baked at 150 °C to solidify the remaining photoresist. The remaining photoresist isolates the InP substrate from the subsequent metal deposition, and also provides mechanical support to the nanowires. A 60 nm Pt layer was then deposited on only one side of each nanowire by mounting the sample on a tilted sample holder (to provide a deposition angle of about ~45°) during evaporation in an electron beam evaporator. The Pt was also deposited over the top of the remaining photoresist layer, forming a continuous conductive film that electrically interconnects all the nanowires in parallel, thereby forming a single 'top' contact of the sensor. Because the contact layer is deposited on only one side of each nanowire, the other side of the nanowire remains exposed and available for sensing analytes.

[0062] The bottom contact (Ti / Au: 10 nm / lOOnm) was formed after the top contact, by removing the polymer from a portion of the sample substrate next to the nanowire array and depositing metal on the exposed substrate. Other conductive metals, such as Al, may alternatively be used.

[0063] With the bottom contact formed, the array was then functionalised by drop-casting a chitosan acetic acid aqueous solution onto the array, followed by baking on a hotplate at 50 °C to remove the solvent.

[0064] Electrical characterisation

[0065] Electrical properties of the sensor were measured using a Keysight B2900 precision source measure unit and a solar simulator, as shown schematically in Figure 3(a). When in operation, the solar simulator was used to illuminate the sensor with 1 Sun @AM1.5 ( / .e., the standard air mass coefficient of 1.5) illumination at 42.3 mW / cm2, as shown in Figure 3(b). Figure 3(c) is a graph showing the resulting current-voltage (I-V) curves of the sensor under dark and light conditions, respectively. Both curves demonstrate rectification behaviour typical of Schottky barrier electrical contacts, in this case resulting from the high work function of the Pt deposited on the nanowires, and the surface band bending of the InP under the deposited Pt. A photovoltaic effect is observed with light illumination from the solar simulator, with a short circuit current (Isc) of 136 nA and an open circuit voltage (Voc) of 80 mV, which enables the self-powered sensing operation. In particular, the short circuit current constitutes the sensing signal with constant light illumination. Acetone sensing measurements

[0066] The gas sensing performance of the sensor was measured using an FTIR600 environmentally sealed sample stage (from Linkam Scientific Instruments Ltd), with Au probes for contacting the sensor contacts / electrodes, mass flow controllers (MFCs) from Bronkhorst, the solar simulator, and gas cylinders, as shown schematically in Figure 3(a). For the gas sensing measurements, the carrier gas simulated air with an N2 to O2 volume ratio of 4 (VN2 / V02 = ~ 4, N2 and O2, BOC gas). The total gas flow rate was maintained at 1 litre / min for ppm level concentrations, and 0.5 litre / min for sub-ppm level concentration measurements. For the measurement of analyte gases, the target Volatile Organic Compounds (VOCs) (ethanol, 9.91 ppm in N2, Coregas; NO2, 10.1 ppm in N2, Coregas; methanol, 10 ppm in N2, BOC gas; acetone, 10 ppm in N2, BOC gas,) were diluted with the simulated air to a desired concentration before purging the sample stage chamber.

[0067] Figure 3(d) is a graph of the sensor response (shown as the percentage change in the short circuit current) as a function of time as the concentration of acetone is repeatedly cycled from zero up to a target concentration and returned to zero for successive target acetone concentrations of 2, 4, 6, 8 and 10 ppm, both under illumination (no bias voltage) and dark (bias voltage of 0.5 V) conditions. Figure 3 (f) similarly shows the illuminated sensor response for cycled acetone concentrations of 100, 200, 400, 600, 800 and 1000 ppb, and Figure 6 (b) for much higher concentrations from 4500 ppm to > 10%. It is apparent from the graphs that the illuminated sensor is effective at detecting the presence of acetone in the gas mixture, and provides a response signal that appears to be approximately linear with acetone concentration over the applicable range. These results confirm that the sensor provides a suitable response for detecting a wide range of acetone levels resulting from DKA, diabetes, and weight loss under a ketogenic diet.

[0068] The sensor performance at lower acetone concentrations of 0.4 - 10 ppb is shown in Figures 3 (g) and (h). The sensor still provides a meaningful response (0.7 %) at 0.4 ppb and provides a linear response in this concentration range from which the acetone limit of detection ("LOD") is estimated as 0.18 ppb. This is an excellent result, demonstrating that the sensor operating range extends down into the part-per-trillion range of acetone concentration. The apparently linear sensor response is confirmed by the data summarised in Figure 3 (e), which include graphs of the peak sensor response R as a function of acetone concentration for two ranges of acetone concentration: from 100 to 1000 ppb (main graph), and from 0.4 to 10 ppm (inset graph). Each graph also includes the linear regression results for the concentration range, where the slope defines the sensor sensitivity S = R / C( / ppb). For the higher concentration range (100 - 1000 ppb), the sensor sensitivity is 0.032 % / ppb, significantly smaller than the 0.48 % / ppb sensitivity for the 0.4 - 10 ppb low concentration range.

[0069] As shown in Figure 4, the sensor exhibits rapid response (Tres: ~25 s) and recovery (Tree: ~39 s) times for acetone levels of 0.1-1 ppm (although about 5 times slower than the sensor produced by a top-down production method, as described below). In practical applications, such a rapid response speed is important for quasi-real-time diagnosis. As summarised in Figure 5, the sensor provided a consistent response signal over 16 cycles of acetone sensing, demonstrating good stability and reproducibility.

[0070] In addition to high sensitivity, in the context of acetone breath sensing it is important that sensor has a high selectivity to acetone because human breath typically includes numerous other VOCs at ppt to ppm concentrations. Consequently, a selective sensor for diabetes monitoring is required to distinguish acetone from these interfering VOCs.

[0071] Figure 6 (a) is a chart showing the measured responses of the sensor to various VOCs and atmospheric gases at 1 ppm concentrations, including acetone (CH3COCH3), methyl nitrides (CH3NO2), ethanol (C2H5OH), propane (CeHe), and carbon dioxide (CO2). It is apparent that the sensor is highly selective to acetone relative to these other VOCs, all of which produce responses less than 5% in comparison to the 49 ± 2% response to acetone, at least an order of magnitude higher.

[0072] In a further measurement, a bubbler apparatus was separately used to evaporate acetone and 2-butanone solvents at higher concentrations. Figure 6 (b) is a graph of the sensor response to each of these solvent vapours as a function of time during cycling for solvent concentrations from 0.45% (4,500 ppm) to 13.1% (131,000 ppm), as shown. Clearly, the sensor maintains a positive sensing response correlation over at least this range of concentrations, which extend to concentrations that are more than 10,000 times higher than the sensing measurements of Figure 3, demonstrating consistent operation over an extremely wide concentration range. Even though both acetone and 2-butanone molecule are small volatile ketones, the sensor demonstrated a high sensitivity only to acetone. This property is of great significance for clinical breath analysis since 2-butanone is a smaller ketone (like acetone) and a biomarker that can be present in breath (e.g., in lung cancer patients), so the ability to distinguish it from acetone can greatly avoid false positives.

[0073] Finally, a major challenge for breath analysis is the presence of water vapour in breath, which could degrade sensing performance by saturating the sensor and consequently decreasing its sensitivity to target gases. The effect of humidity on acetone sensing was investigated using the same bubbler apparatus described above. As shown in Figure 6(c), relative to sensing in dry air, the sensor response to acetone decreased by 10 ± 2 %, 30 ± 6 % and 50 ± 12 % for relative humidity ("RH") values of 20%, 50%, and 65%, respectively. However, Figure 6 (d) shows that the sensor maintained a consistent sensitivity that is largely independent of humidity, indicating that the sensor is not overwhelmed by humid environments. The sensor response to potentially interfering molecules in exhaled breath, ethanol and CO2, was also measured at different relative humidities, and the results are summarised in Figure 6 (e) . It is apparent that the sensor maintained its high selectivity to acetone, with an acetone response that was more than 10 times higher than its responses to interfering gases, even at the high relative humidities.

[0074] The sensors described herein can also be used in darkness by applying a bias voltage across the sensor, but at the expense of sensitivity to acetone concentration, as shown in Figure 7(b). A similar result arises when a sensor is made from n-doped InP nanowires rather than undoped InP nanowires. Figure 7(a) compares the sensor response to cycled acetone to the response of a similar sensor with undoped rather than n-type InP nanowires. The undoped InP nanowire sensor provides a higher response value to acetone generally, because doping in nanowires only increases the electric current value instead of the electric current variation. EXAMPLE II

[0075] In another example, the top-down production method described above was used to produce an acetone sensor, as shown schematically in Figure 8 (a). A 200 nm SiO2 layer was deposited on an undoped InP wafer by PECVD. An ARP6200.09 photoresist layer was deposited over the SiO2 layer, and patterned using electron-beam lithography (EBL) to form openings in the photoresist generally corresponding to the desired nanowire locations and diameter. A 70 nm Ni layer was deposited over the patterned photoresist layer by e-beam evaporation, and the photoresist layer removed to transfer the pattern to the Ni layer by lift-off, forming the structure shown schematically in the leftmost schematic of Figure 8(a) and in the SEM image of Figure 8 (b).

[0076] The exposed regions of the SiO2 layer were then removed by ICP-RIE with fluorine gas, leaving a Ni / SiO2 bilayer as the mask, as shown schematically in the second image of Figure 8 (a) and in 30° tilted angle view in the SEM image of Figure 8 (c), for subsequent InP etching by ICP-RIE with chlorine gas to form the structure shown in the third image of Figure 8 (a), and, following removal of the Ni / SiO2 bilayer mask, the structure shown schematically in the fourth and rightmost image of Figure 8 (a), and in the SEM image of Figure 8 (d).

[0077] The nanowire diameter is ultimately determined by the mask and the ICP etching process, with the smallest achievable mask size being around 40-50 nm through EBL and Ni deposition. The SEM image of Figure 8 (d) shows that the resulting nanowires are tapered due to etching of the exposed sidewalls of the nanowires during the deep etching of the substrate, the resulting diameter of each nanowire increasing from ~50 nm at the top to ~130 nm at the base of the nanowire. Another limitation of the top- down production method is the smaller nanowire length due to the limited etching selectivity of the mask and the InP substrate, causing the mask to be slowly removed during the etching process. The average length of nanowires produced by the top-down production method is about 2 pm, considerably shorter than the nanowires grown by the bottom-up production method described above, which are longer than 4 pm.

[0078] Once the nanowires have been formed as described above, the steps forming the electrical contacts and chitosan drop-casting are the same as described above for the bottom-up production method. Figure 8 (e) is an SEM image showing the nanowires after the tilt angle deposition of the Pt contacts, and Figures 8 (f) and (g) both show the nanowires after the chitosan drop-casting. Figure 9(a) is a schematic image of the resulting sensor, and Figures 9 (b) to (h) show measurements of the sensor produced by the top-down production method. Figure 9(b) is a graph showing the I-V characteristics of the sensor under dark and light illumination conditions with the solar simulator @1.5AM.

[0079] Figures 9 (c) and (e) show the time-dependent responses of the illuminated sensor to pulsed acetone concentrations from 1 to 0.1 ppm, and from 10 to 2 ppm, respectively. Figure 9(d) is a graph showing the 1 ppm acetone sensing response at light intensities decreasing from 42.3 to 4.2 mW-crrr2. Significantly, the acetone response only decreases slightly from 6.5 % to 4.7 % over an order of magnitude decrease in light intensity, indicating suitability for use in real-world environmental settings where the ambient light intensity may vary significantly.

[0080] Figure 9 (g) confirms that the sensing signal has a linear relationship with acetone concentration. Figure 9 (h) is a chart comparing the responses of the sensor to a 10 ppm concentration of each of: acetone, 2-butanone, ethyl benzene, ethanol, propane, NO2, and CO2 of 10%, confirming that the top-down produced sensor also has excellent selectivity to acetone. Figure 9 (f) shows the time-dependent response of the sensor to an acetone concentration of 2-10 ppm at relative humidity (RH) levels of 0%, 20%, 50%, and 65%.

[0081] It will be apparent from the data shown in Figure 9 that acetone sensors produced by the top-down production method have generally similar characteristics to sensors produced by the bottom-up production method. However, as summarised in Table 1 below, there are significant differences in performance. For example, the acetone sensing response of the top-down sensor (6.5% @ 1 ppm) is significantly smaller than that of the bottom-up sensor (49% @ 1 ppm), and also the LOD of the top-down sensor (4.5 ppb) is larger than that of the bottom-up sensor. Conversely, the response and recovery times of the top-down sensor are about five times shorter than those of the bottom-up sensor. Regardless, the top-down sensor has a high sensitivity of 6.61 % / ppm to acetone over the relevant diabetic acetone concentration range of 0.1-10 ppm. Table 1: Performance comparison of sensors produced by the bottom-up and top-down production methods

[0082] Tres / Trec Production Method R (1 ppm) S (% / ppm) LOD (ppb)

[0083] (s) bottom-up 49 % 480 ~ 25 / 39 0.33 top-down 6.5 % 6.61 ~ 5 / 6 36

[0084] EXAMPLE III

[0085] The top-down acetone sensor described above was also used as the sensing component of a portable handheld electronic acetone sensing device, also referred to herein as the "ketone whistle", suitable for breath sensing in point-of-care applications. The top-down acetone sensor was chosen for this application owing to its shorter response time Tres(~5 s) and recovery time Tree (~7 s) and ease of fabrication, while exhibiting humidity tolerance, insensitivity to light intensity, and acetone selectivity, despite its slightly inferior sensitivity.

[0086] Figure 10 is a block diagram showing the electronic components of the device, in which the sensor 1002 is illuminated (and thus powered) by a red light-emitting diode (LED) 1004. The device also includes a commercial CO2 sensor 1006 (to ensure a breath deep enough to contain acetone), signal-processing circuitry 1008, and an OLED screen 1010 for displaying the acetone and CO2 measurement results. The red LED 1004 was located directly on top of the sensor to eliminate any interference from ambient light fluctuations.

[0087] Considering the wide acetone concentration range corresponding to a variety of physiological states, as shown in Figure 11(b), calibration of the device for breath acetone measurements is required. The device was calibrated by the laboratory gas system described above, with quantified acetone concentrations and relative humidity. The resulting calibration curves shown in Figures 11(c) show a sensitivity decreasing with acetone concentration from 6.11 mV / ppm (0.1 - 10 ppm) to 0.013 mV / ppm (50 - 1000 ppm), which is consistent with the laboratory measurements shown in Figure 3 (e). The approximately linear correlation of the output voltage signal with acetone concentration over several orders of magnitude confirms that the device is suitable for diagnosing diabetes, ketogenic diet assistance, and even ketoacidosis (DKA) monitoring which, if undetected, can lead to dehydration, confusion, loss of consciousness, and even loss of life if not timely treated.

[0088] For breath testing, Tedlar bags (PVDF 0.6 L, Sigma-Aldrich) were used for VOC sampling and collecting breath samples from healthy subjects and simulated diabetic breath (see Figure 11 (a)). Exhaled breath from diabetic patients typically contains acetone gases at concentrations of 1.8 - 20 ppm (as shown in Figure 11 (b)), so the concentration of simulated breath of diabetic patients was adjusted to ~ 5 ppm by filling the Tedlar bags with ~50 % of healthy exhaled breath and ~50 % of 10 ppm acetone in simulated air. Principal component analysis (PCA) was applied to analyze the output voltage dependent breath test measurements, and the results are shown in Figure 11 (e), where the distinction between the 10 simulated diabetic breath samples and the 10 healthy subjects was clearly categorized into two distinguishable clusters without any overlap. This confirms that the device is effective to diagnose diabetes in human subjects.

[0089] Many modifications will be apparent to those skilled in the art without departing from the scope of the present invention.

Claims

CLAIMS:

1. A sensor for measuring acetone, including: an array of nanowires extending from a substrate, each of the nanowires having a corresponding surface composed substantially of a compound semiconductor, and a diameter less than 100 nm; an electrical contact material disposed on a first portion of the surface of each nanowire to form a Schottky barrier contact therewith; an amine-functionalised polymer disposed on the nanowires to functionalise a second portion of the surface of each nanowire; and an electrical contact to the substrate to enable measurement of electric current flowing between the electrical contact and the Schottky barrier contacts on the nanowires when the nanowires are exposed to light or biased, such that, when the nanowires are exposed to a mixture of molecular species including acetone, the measured electric current is representative of a concentration of acetone in the mixture.

2. The sensor of claim 1, wherein each of the nanowires has a diameter of about 50 to 60 nm.

3. The sensor of claim 1 or 2, wherein the amine-functionalised polymer is chitosan.

4. The sensor of any one of claims 1 to 3, wherein the electrical contact material is composed substantially of a metal selected from the group consisting of Pt, Pd, Ni, and Au.

5. The sensor of claim 4, wherein the electrical contact material is composed substantially of Pt.

6. The sensor of any one of claims 1 to 5, wherein the sensor is self-powered when exposed to light.

7. The sensor of any one of claims 1 to 6, wherein the compound semiconductor is a compound semiconductor selected from the group consisting of: InP, InAs, GaP, and GaAs.

8. The sensor of any one of claims 1 to 7, wherein the compound semiconductor is un-doped InP.

9. The sensor of any one of claims 1 to 8, wherein the sensor has a response time < 20s, and a detection limit < 0.1 ppm.

10. The sensor of any one of claims 1 to 9, wherein the sensor is flexible and a component of a wearable sensing device.

11. The sensor of any one of claims 1 to 10, wherein the mixture of molecular species is in a gas phase.

12. A handheld acetone measuring device, including the sensor of any one of claims 1 to 11.

13. The handheld acetone measuring device of claim 12, including a light-emitting diode configured to emit light onto the sensor to cause the electric current to flow between the electrical contact and the Schottky barrier contacts.

14. The handheld acetone measuring device of claim 12 or 13, wherein the device is configured to diagnose diabetes and ketoacidosis from human breath.

15. A method of producing a sensor for measuring acetone, the method including the steps of: providing an array of nanowires extending from a substrate, each of the nanowires having a corresponding surface composed substantially of a compound semiconductor and a diameter less than 100 nm; depositing an electrical contact material on a first portion of the surface of each nanowire to form a Schottky barrier contact therewith; depositing an amine-functionalised polymer onto a second portion of the surface of each nanowire; and forming an electrical contact to the substrate to enable measurement of electric current flowing between the electrical contact and the Schottky barrier contacts on the nanowires when the nanowires are exposed to light or biased; wherein, when the nanowires are exposed to a mixture of molecular species including acetone, the measured electric current is representative of a concentration of acetone in the mixture.

16. The method of claim 15, wherein providing the array of nanowires includes growing the array of nanowires from the substrate by an additive bottom-up growth process.

17. The method of claim 15, wherein providing the array of nanowires includes forming the array of nanowires by a top-down etching process.

18. The method of any one of claims 15 to 17, wherein the step of depositing an amine-functionalised polymer includes drop-casting chitosan onto the nanowires.

19. The method of any one of claims 15 to 18, wherein the step of depositing an electrical contact material on a first portion of the surface of each nanowire includes tilt-angle deposition of the electrical contact material onto the nanowires so that the electrical contact material is deposited onto only one side of each nanowire.

20. The method of any one of claims 15 to 19, wherein the electrical contact material is composed substantially of Pt.

21. The method of any one of claims 15 to 20, wherein each of the nanowires has a diameter of about 50 to 60 nm.

22. The method of any one of claims 15 to 21, wherein the compound semiconductor is a compound semiconductor selected from the group consisting of: InP, InAs, GaP, and GaAs.

23. The method of any one of claims 15 to 22, wherein the compound semiconductor is a undoped InP.

24. The method of any one of claims 15 to 23, wherein the electrical contact material is composed substantially of a metal selected from the group consisting of Pt, Pd, Ni, and Au.

25. The sensor of any one of claims 1 to 11, or the device of any one of claims 12 to 14, or the method of any one of claims 15 to 24, wherein the mixture of molecular species including acetone is sweat or breath.

Images