WO2025238111A1

WO2025238111A1 - Method and apparatus for detecting a characteristic of a target species within a liquid

Info

Publication number: WO2025238111A1
Application number: PCT/EP2025/063292
Authority: WO
Inventors: Thomas Joshua WADE; Diandian ZHANG; Thiyagarajan NATARAJAN; Sohini Kar-Narayan
Original assignee: Cambridge Enterprise Ltd
Current assignee: Cambridge Enterprise Ltd
Priority date: 2024-05-14
Filing date: 2025-05-14
Publication date: 2025-11-20
Anticipated expiration: 2026-11-14
Also published as: GB202406787D0

Abstract

A method of detecting a characteristic of a target species in a liquid, comprising the steps of: providing a capacitive element in contact with the liquid, determining a capacitance of the capacitive element across a range of measurement frequencies, determining turning point data representative of a turning point of the capacitance as a function of the measurement frequency, and determining a characteristic of the target species from the turning point data.

Description

METHOD AND APPARATUS FOR DETECTING A CHARACTERISTIC OF A TARGET

SPECIES WITHIN A LIQUID

BACKGROUND

[0001] Remote health monitoring (RHM) has many advantages including easier disease management, earlier and improved detection and diagnosis, reduced costs, and increased convenience for patients. RHM could facilitate a shift from reactive healthcare to predictive, preventative and personalised methods.

[0002] Whilst blood is commonly used to assess the health of a patient, it can only be measured at discrete times and requires invasive sampling in a clinical setting. These drawbacks may be avoided through the continuous monitoring of more accessible biofluids such as sweat, saliva and tears. For example, sweat has been shown to contain a wealth of information about the metabolic state, infections and diseases, and physical exertion of an individual. In particular, the concentration of salts such as NaCl and KC1 in sweat can be used to monitor hydration, amongst other physiological markers.

[0003] Existing monitoring devices rely on techniques such as colorimetry. Such devices may be single use and may require the patient to take a photograph of the device adding complexity and error to detection.

[0004] A common alternative to colorimetry is electrochemical detection, which has the potential to offer continuous monitoring. However, specific chemical compounds must be developed and used to measure different metabolites, and the need for high efficiency and specificity whilst maintaining a low cost limits the practicality of such detectors. In addition, the shelf-life and usage lifetimes of the specific chemical compounds used are often limited which prohibits long-term measurements. Evaluation over days, weeks, or even longer is an important advantage of RHM over traditional methods, but due to these limitations, electrochemical sensors are currently unable to achieve this.

[0005] Biosensors utilising electrical, mechanical and electromechanical detection techniques are also in development, but currently face technological limitations such as sensitivity and manufacturing cost.

[0006] Impedance spectroscopy has been used for analysis of fluids. However, this technology is yet to be suitably miniaturised to enable the production of a fully wearable device capable of accurately characterising liquids. [0007] It is an object of the present disclosure to overcome some or all of the aforementioned drawbacks associated with existing detectors.

BRIEF SUMMARY

[0008] An aspect provides a method of detecting a characteristic of a target species in a liquid, comprising the steps of: providing a capacitive element in contact with the liquid, determining a capacitance of the capacitive element across a range of measurement frequencies, determining turning point data representative of a turning point of the capacitance as a function of the measurement frequency, and determining a characteristic of the target species from the turning point data.

[0009] The turning point data may comprise the measurement frequency at the turning point. In particular, this may be useful for determining concentration and/or identity of an ionic target species in the liquid.

[0010] The method may include determining turning point data representative of a plurality of turning points of the capacitance as a function of the measurement frequency. The turning point data may comprise the measurement frequency of each of the plurality of turning points.

[0011] The turning point data may comprise a number of turning points within the range of measurement frequencies.

[0012] The turning point data may comprise the capacitance at the turning point.

[0013] The step of determining a characteristic of the target species from the turning point data may include comparing the turning point data to stored data representing turning points of the capacitance of the capacitive element as a function of measurement frequency for a plurality of known samples.

[0014] Determining the characteristic of the target species may involve detecting a concentration of the target species in the liquid from the turning point data.

[0015] Determining the characteristic of the target species may involve identifying the target species in the liquid from a set of known target species using the turning point data. [0016] The method may comprise the step of varying a temperature of the capacitive element within a range of measurement temperatures. The method may comprise the step of determining the capacitance of the capacitive element as a function of measurement frequency across the range of measurement temperatures.

[0017] The method may comprise the step of determining turning point data representative of turning points of the capacitance as a function of the measurement frequency across the range of measurement temperatures. The method may comprise the step of analysing the turning point data for each of the plurality of temperature measurements to determine an additional characteristic of the target species.

[0018] The step of determining the additional characteristic of the target species may involve calculating a gradient of the turning point data as a function of the measurement temperature and comparing the calculated gradient to stored gradients for a plurality of species to determine an identity or a concentration of the target species in the liquid. The

[0019] The method may comprise the step of measuring an impedance of the capacitive element across a range of measurement frequencies to obtain impedance measurements including a real component and an imaginary component. The method may include the step of calculating the capacitance of the capacitive element across a range of measurement frequencies from the impedance of the capacitive element across a range of measurement frequencies. The capacitance may be an effective capacitance. For example, the effective capacitance may be calculated according to the following equation. f lmiZ]

(Equation 1)

[0020] The method may comprise the step of determining a minimum of the imaginary component of the impedance measurements of the capacitive element across the measurement frequencies. The method may comprise the step of determining an additional characteristic of the target species by comparing the minimum of the imaginary component to stored minimum imaginary components for one or more species to determine an identity or a concentration of the target species in the liquid.

[0021] The step of determining the additional characteristic of the target species may involve calculating a gradient of the minimum of the imaginary component as a function of the measurement temperature and comparing the calculated gradient to stored gradients for a plurality of species to determine the identity or the concentration of the target species in the liquid.

[0022] The method may comprise the step of comparing the characteristic with the additional characteristic(s) to determine concentration and identity of the species of the target species in the liquid.

[0023] The method may comprise a step of measuring a temperature indicative of the environmental temperature. For example, the method may comprise measuring temperature of the liquid or of the apparatus used to carry out the method. The method may comprise a step of accessing stored data relating to the dependence of recorded turning point data on temperature. The method may involve accounting for the temperature at the time the impedance measurements were recorded when determining the characteristic of the target species from the turning point data.

[0024] The target species may comprise an electrolyte, such as a nitrate, phosphate, fluoride, or chloride (e.g., NaCl, KC1, MgCh, CaCh) or lactate. The target species may comprise a metabolite, such as lactate, lactose, glucose, urea, ammonia, cortisol, glycerol, pyruvate, tyrosine, or serine.

[0025] The liquid may comprise water. The liquid may comprise drinking water or potable water, or water stored in a reservoir or storage vessel or water distribution system. The liquid may comprise wastewater such as industrial wastewater or sewage.

[0026] A channel may be provided overlaying the capacitive element. The channel may comprise or consist essentially of a microfluidic channel.

[0027] The capacitive element may comprise a first electrode and a second electrode spaced from the first electrode. The first electrode and the second electrode may be interdigitated electrodes. The capacitive element may comprise any features discussed herein in relation to the capacitive element, sensor, sensor element or apparatus of any other aspect.

[0028] In some examples the liquid may comprise a complex solution comprising plural ionic species. Each or a selection of the plural ionic species may be target species.

[0029] The method may comprise detecting a characteristic of two or more target species in a liquid, wherein for each of the target species the method comprises the steps of: providing a capacitive element in contact with the liquid, determining a capacitance of the capacitive element across a range of measurement frequencies, determining turning point data representative of a turning point of the capacitance as a function of the measurement frequency, and determining a characteristic of the additional target species from the turning point data.

[0030] For example, where the method comprises detecting a characteristic of two, three or four target species, two, three or four different capacitive elements, respectively, may be used. [0031] Each capacitive element may be configured to preferentially detect a respective one of the target species. For example, each capacitive element may comprise an ion-selective membrane configured for a specific target species. In other examples, each capacitive element may comprise a molecularly imprinted polymer (MIP) configured for a specific target species. In some examples, one or more capacitive elements may be provided with an MIP and one or more other capacitive elements may be provided with an ion-selective membrane. In this way it is possible to determine the identity / concentration of both ionic species and metabolite species.

[0032] In some examples, a channel (such as a microfluidic channel) may be provided for each of the respective capacitive elements. Each channel may overlie electrode(s) of that respective capacitive element. The ion-selective membrane may cover an inlet of the channel. In some examples, the ion-selective membrane may form part of the capacitive element. The ion-selective membrane may cover the first electrode and the second electrode. For example, the ion-selective membrane may encapsulate the first and second electrodes. Each of the ion- selective membranes may cover the first and second electrode of that capacitive element. For example, each ion-selective membrane may encapsulate the first and second electrodes of that capacitive element. Each ion-selective membrane may be configured to select for a different ion (than each other ion-selective membrane is configured to select for). Each ion-selective membrane may be configured to select for any one of: nitrates, phosphates, fluorides or chlorides (e.g., NaCl, KC1, MgCE, CaCh ).

[0033] The apparatus may comprise a reservoir for receiving the liquid. The reservoir may be fluidically coupled to the inlet(s) of the channel(s). [0034] In some examples, each of the capacitive elements may comprise an ion-selective membrane. In other examples, at least one of the capacitive elements may not comprise an ion- selective membrane. In examples, there are a plurality of capacitive elements, a plurality of the capacitive elements comprise an ion-selective membrane, and (only) one of the capacitive elements does not comprise an ion-selective membrane. Capacitive elements comprising an ion-selective membrane may be termed ‘additional capacitive elements’.

[0035] The method may comprise determining turning point data of the complex solution from the capacitive element which does not comprise an ion-selective membrane.

[0036] In particular, the method may comprise determining turning point data (TPtotai) representative of a turning point of the capacitance as a function of the measurement frequency in dependence on the received signal indicative of the capacitance of the capacitive element.

[0037] TPtotai is indicative of a total turning point data for the complex solution.

[0038] For each of the additional capacitive elements which each comprise an ion-selective membrane configured to select for a target species, the method may comprise determining turning point data for that target species. In particular, determining the turning point data comprises back-calculating what the frequency turning point would be if there was only the target species in the liquid.

[0039] That is, for each of the one or more additional capacitive elements, the method may comprise determining turning point data (TP;_on) representative of a turning point of the capacitance as a function of the measurement frequency in dependence on the received signal indicative of the capacitance of that additional capacitive element.

[0040] TPion is indicative of turning point data specific to the ion which the ion-selective membrane of that additional capacitive element is configured to select for.

[0041] The sum of the individual contributions of each ionic species in the liquid to the turning point data equals the total (non-species specific) turning point data (TPtotai) measured. For an apparatus comprising N additional capacitive elements, by subtracting the contribution of each of the N ionic species in the liquid to the turning point data from the total turning point data measured at the capacitive element with no ion-selective membrane, turning point data representative of the individual contribution of an additional unknown ion can be calculated. I.e.:

This can in turn be used to determine a characteristic such as concentration of the additional unknown ion. In this way, for a solution comprising N+l target species, a characteristic (such as concentration) of each of the N+l target species may be calculated using only N ion- selective membranes. Beneficially, this may reduce the cost of manufacturing an apparatus for completing the method.

[0042] In some examples, the type of the unknown ion may be known, and the above method used to determine its concentration. In other examples, the type of the unknown ion may be unknown, and the method described above may be used to determine its concentration.

[0043] The method may comprise calculating turning point data (TPExtraion) representative of a turning point of the capacitance as a function of the measurement frequency for an extra ion, for which none of the ion-selective membranes of the additional capacitive elements are configured to select for, on the basis of TPtotai and TP;_on for each of the one or more additional capacitive elements.

[0044] For example, the method may comprise calculating TPExtraion by subtracting the sum of TPion for all of the one or more additional capacitive elements from TPtotai.

[0045] Another aspect provides an apparatus for detecting a characteristic of a target species within a liquid, comprising: a sensor, comprising a capacitive element for contacting the liquid and outputting a signal indicative of a capacitance of the capacitive element across a range of measurement frequencies; and a processor configured to: receive the signal indicative of the capacitance, in dependence on the received signal indicative of the capacitance, determine turning point data representative of a turning point of the capacitance as a function of the measurement frequency; and determine a characteristic of the target species from the turning point data. [0046] The processor may be configured to compare the turning point data to stored data representing turning points of the capacitance of the capacitive element as a function of measurement frequency for a plurality of known samples to determine an identity or a concentration of the target species in the liquid.

[0047] The sensor may comprise a channel for receiving the liquid, the channel overlaying the capacitive element. The channel may comprise or consist essentially of a microfluidic channel.

[0048] The capacitive element may comprise a first electrode and a second electrode spaced from the first electrode. The first electrode and the second electrode may be interdigitated electrodes.

[0049] The first electrode may comprise a first backbone from which a first set of protrusions project. The second electrode may comprise a second backbone from which a second set of protrusions project. The first set of protrusions and the second set of protrusions may interdigitate along a length of the capacitive element.

[0050] An electrode gap may be provided between each neighboring one of the first set of protrusions and the second set of protrusions.

[0051] The first set of protrusions and the second set of protrusions may be evenly spaced along the length of the capacitive element. A center-to-center electrode distance may be measured between the center of one of the first set of protrusions to a center of a neighboring one of the second set of protrusions. The center-to-center electrode distance may be less than or equal to 200pm, 150pm, 120pm, 100pm or 90pm. The center-to-center electrode distance may be greater than or equal to 140pm, 120pm, 100pm or 80pm. For example, the center-to-center electrode distance may be 80pm.

[0052] Each of the first set of protrusions may be parallel to each of the second set of protrusions such that the electrode gap is constant along a width of the capacitive element.

[0053] In some examples, the first electrode and/or the second electrode may be Aerosol Jet Printed (e.g., in silver ink and/or carbon ink).

[0054] The capacitive element may be Aerosol Jet Printed (e.g., in silver ink and/or carbon ink).

[0055] The sensor may comprise an ion-selective membrane. The ion-selective membrane may cover an inlet of the channel. In some examples, the ion-selective membrane may form part of the capacitive element. The ion-selective membrane may cover the first electrode and the second electrode. For example, the ion-selective membrane may encapsulate the first and second electrodes.

[0056] The apparatus may comprise a reservoir for receiving the liquid. The reservoir may be fluidically coupled to the inlet of the microfluidic channel.

[0057] The apparatus may comprise one or more additional capacitive elements. Each of the one or more additional capacitive elements may comprise an ion-selective membrane. Each of the ion-selective membranes may cover the first and second electrode of that capacitive element. For example, each ion-selective membrane may encapsulate the first and second electrodes of that capacitive element. Each ion-selective membrane may be configured to select for a different ion (than each other ion-selective membrane is configured to select for).

[0058] Wherein each additional capacitive element is for contacting the liquid and outputting a signal indicative of the capacitance of the capacitive element across a range of measurement frequencies.

[0059] The sensor may comprise a sensor element (e.g., for determining a capacitance of the capacitive element across a range of measurement frequencies). The sensor element may include the channel and the capacitive element. The apparatus may comprise one or more additional sensor elements (e.g., for determining a capacitance of a respective one of one or more additional capacitive elements) across a range of measurement frequencies.

[0060] Each of the additional sensor elements may have any of the features associated with the sensor element. Each of the one or more additional sensors may comprise an additional capacitive element for contacting the liquid.

[0061] Each of the one or more additional sensor elements may comprise an additional channel, comprising an inlet for receiving the liquid from the reservoir, overlaying the additional capacitive element. Each of the one or more additional sensor elements may further comprise an additional ion-selective membrane which covers the inlet of the additional channel.

[0062] The additional ion-selective membrane may be configured to select for a different ion than the ion-selective membrane of the channel. [0063] The sensor may include a plurality of additional sensor elements. Each additional ion- selective membrane may be configured to select for a different ion than the others of the additional ion-selective membranes.

[0064] In some examples, the sensor may comprise: a sensor element which does not comprise an ion-selective membrane; and one or more additional sensor elements each of which do comprise an ion selective membrane. Each additional ion-selective membrane may be configured to select for a different ion than the others of the additional ion-selective membranes.

[0065] In some examples the liquid may comprise a complex solution comprising plural ionic species. Each or a selection of the plural ionic species may be target species.

[0066] The apparatus (e.g., the processor of the apparatus) may be configured to, in dependence on the received signal indicative of the capacitance of the sensor element (which does not comprise an ion-selective membrane), determine turning point data (TPtotai) representative of a turning point of the capacitance as a function of the measurement frequency.

[0067] TPtotai is indicative of a total turning point data for the complex solution.

[0068] For each of the one or more additional sensor elements (which each comprise an ion- selective membrane), the apparatus (e.g., the processor of the apparatus) may be configured to, in dependence on the received signal indicative of the capacitance of that additional sensor element, determine turning point data (TP;_on) representative of a turning point of the capacitance as a function of the measurement frequency.

[0069] TPion is indicative of turning point data specific to the ion which the ion-selective membrane of that additional sensor element is configured to select for.

[0070] The sum of the individual contributions of each ionic species in the liquid to the turning point data may equal the total (non-species specific) turning point data (TPtotai) measured. For an apparatus comprising N additional sensor elements, by subtracting the contribution of each of the N ionic species in the liquid to the turning point data from the total turning point data measured at the sensor element with no ion-selective membrane, turning point data representative of the individual contribution of an additional unknown ion can be calculated. I.e. :

[0071] This can in turn be used to determine a characteristic such as concentration of the additional unknown ion. In this way, for a solution comprising N+l target species, a characteristic (such as concentration) of each of the N+l target species may be calculated using an apparatus comprising only N ion-selective membranes. Beneficially, this may reduce the cost of manufacturing the apparatus.

[0072] In some examples, the type of the unknown ion may be known, and the above method used to determine its concentration. In other examples, the type of the unknown ion may be unknown, and the method described above may be used to determine its concentration.

[0073] The apparatus may be configured to calculate turning point data (TPExtraion) representative of a turning point of the capacitance as a function of the measurement frequency for an extra ion for which none of the ion-selective membranes of the additional sensor elements are configured to select for on the basis of TPtotai and TPi_on for each of the one or more additional sensor elements.

[0074] For example, the apparatus may be configured to calculate TPExtraion by subtracting the sum of TPion for all of the one or more additional sensor elements from TPtotai.

[0075]

[0076] The sensor may comprise an impedance spectroscopy circuit for measuring an impedance of the capacitive element across a range of measurement frequencies. The processor may be configured to calculate the capacitance of the capacitive element across a range of measurement frequencies from the impedance of the capacitive element across a range of measurement frequencies.

[0077] The impedance spectroscopy circuit may be configured to measure a real component and an imaginary component of the impedance of the capacitive element across a range of measurement frequencies. The processor may be configured to determine a minimum of the imaginary component. The processor may be configured to compare the minimum of the imaginary component to stored minimum imaginary components for one or more species to determine an additional characteristic of the target species. [0078] The impedance spectroscopy circuit may be operably connected to each of the capacitive elements. The processor may be configured to analyse the characteristic data received from each capacitive element in combination with the identification data from each capacitive element to determine the concentration and the type of target species present in the liquid.

[0079] The apparatus may comprise a heating element configured to increase and/or decrease a temperature of the channel within a range of measurement temperatures. The apparatus may be configured to determine the capacitance of the capacitive element as a function of measurement frequency across the range of measurement temperatures. The processor may be configured to determine turning point data representative of turning points of the capacitance as a function of the measurement frequency across the range of measurement temperatures. The processor may be configured to analyse the turning point data for each of the plurality of temperature measurements to determine an additional characteristic of the target species.

[0080] The apparatus may comprise a water testing device. For example, the apparatus may comprise a drinking water or potable water testing device. The apparatus may be used on water in a reservoir, storage vessel, or water distribution system. In some examples the apparatus may comprise a wastewater testing device, in particular for testing industrial wastewater or sewage

[0081] The apparatus may comprise a wearable device. The wearable device may comprise the sensor and, optionally, the processor. Such a wearable device may be used to test sweat.

[0082] The sensor may comprise a cover layer which at least partially covers the substrate. The channel(s) may form part of the cover layer. The cover layer may comprise or essentially consist of a flexible plastic such as polydimethylsiloxane (PDMS) or Flexdym®. The cover layer may be attached to the substrate (e.g., by an adhesive).

[0083] The inlet, for receiving the liquid, may be provided at a first side of the channel. The channel may further include an outlet at a second side of the channel. The channel may be elongate and the first side of the channel may be substantially opposite the second side of the channel. The channel may be configured to overlie the capacitive element.

[0084] The apparatus may comprise a memory module which comprises stored turning point data representing turning points of the capacitance of the capacitive element as a function of measurement frequency for a plurality of known calibration samples. [0085] The processor may be configured to compare the stored turning point data to turning point data of the liquid being measured to determine the characteristic of the target species in the liquid.

[0086] The apparatus may comprise a results memory configured to store the characteristic data representing a characteristic of the liquid determined by the apparatus. For example, the characteristic data may be representative of the concentration and/or identity of the target species in the liquid. The processor may be configured to transmit the characteristic data to the results memory by wired or wireless means.

[0087] The processor may be configured to transmit sample identifiers to the results memory. The characteristic data in the results memory may be indexed by the sample identifiers. For example, the sample identifier may comprise a timestamp. Each entry of characteristic data may be associated with a timestamp representing a date and time at which the measurements used to calculate that characteristic data were recorded. The sample identifier may further include identification data representative of either the identity of the sensor, the sensor element or the identity of the user of the sensor at the time the measurements used to calculate that characteristic data were recorded.

[0088] The apparatus may comprise a user interface. The user interface may be configured to receive the characteristic data from the results memory by wired or wireless means. The user interface may comprise an interactive touch screen display.

[0089] The wearable device may comprise a strap for attaching to a user e.g., to a limb of the user.

[0090] The substrate, and optionally the cover layer, may be arcuate in shape. The sensor may be operable to bend to conform to the shape of a user's skin.

[0091] The apparatus may comprise a power source, such as a battery, operable to supply power to at least the sensor.

[0092] The apparatus may comprise an external module which is wirelessly connected to the wearable device. Each or any combination of: the processor; the memory module; the results memory; and the user interface may form part of the wearable device or alternately may form part of the external module. [0093] The apparatus may comprise a heating module. The heating module may comprise a heating element, such as a Peltier element, configured to increase and/or decrease a temperature of the channel within a range of measurement temperatures. The heating module may comprise a heat sink thermally coupled to the heating element opposite the substrate.

[0094] The apparatus may comprise a temperature sensor. For example, where present, the heating module may comprise the temperature sensor.

[0095] The processor may be configured to control the temperature of the heating element to vary the temperature of the channel of the sensor element.

[0096] The apparatus may comprise a temperature sensor for measuring an environmental temperature. For example, the temperature sensor may be configured to measure the temperature of the liquid and/or of the apparatus. The memory may store information relating to the dependence of recorded turning point data on temperature. The apparatus may be configured to account for the temperature at the time the impedance measurements were recorded when determining the characteristic of the target species from the turning point data.

[0097] The capacitive element may comprise a molecularly imprinted polymer (MIP). For example, the capacitive element may comprise a MIP layer. The MIP layer may be deposited on the first electrode and/or the second electrode such that the MIP layer contacts the liquid in the channel.

[0098] The first electrode and/or the second electrode may consist of a first material. A second material may overlie a selected area of the first electrode and/or the second electrode. The second material may have a higher affinity for the MIP than the first material.

[0099] The first electrode and the second electrode may be printed from carbon ink. Silver may then be deposited over a selected area of the first electrode and/or the second electrode.

[0100] In another example, the second material may underlie the selected area of the first electrode and/or the second electrode.

[0101] The second material and MIP layer may partially cover the first/second electrode such that some of the surface of the first/second electrode is exposed to the liquid in the channel in use. Alternately, the second material and MIP layer may completely cover the first/second electrode such that, in use, none of the surface of the first/second electrode is exposed to the liquid in the channel. [0102] The MIP layer may be functionalised for a metabolite. For example, the MIP layer may be functionalised as a lactate MIP, lactose MIP, glucose MIP, urea MIP, ammonia MIP, cortisol MIP, glycerol MIP, pyruvate MIP, tyrosine MIP, or serine MIP.

[0103] The sensor may comprise one or more further capacitive elements each comprising a molecularly imprinted polymer layer. For example, the sensor may comprise two, three, four, five, six or more capacitive elements each comprising a molecularly imprinted polymer layer. The molecularly imprinted layers of the various capacitive elements may each be different, or some may be the same. Each molecularly imprinted polymer layer may be functionalised for a different species such as a different metabolite (e.g., lactate, lactose, glucose, urea, ammonia, cortisol, glycerol, pyruvate, tyrosine, or serine).

[0104] For example, the sensor may comprise: a first capacitive element comprising a lactate MIP layer; and a second capacitive element comprising a glucose MIP layer. Such as sensor may further comprise a third capacitive element comprising an ion-selective membrane (e.g., an ion-selective configured to select for NaCl).

[0105] The apparatus may be operable to determine the identity and/or concentration of both ion(s) and metabolite(s) within the liquid.

[0106] For example, the apparatus may comprise a plurality of capacitive elements, one or more of which has a MIP layer, and one or more of which has an ion-selective membrane. In this way, the apparatus may be configured to determine the identity and/or concentration of both ion(s) and metabolite(s) within the liquid.

[0107] Another aspect provides a method of detecting a characteristic of a target species within a liquid using a wearable device, comprising: providing a wearable device comprising a capacitive element for contacting the liquid and a sensor for measuring a capacitance or an impedance of the capacitive element across a range of measurement frequencies, determining a capacitance of the capacitive element across a range of measurement frequencies using the sensor, and determining a characteristic of the target species from the capacitance of the capacitive element as a function of measurement frequency. [0108] Another aspect provides an apparatus for detecting a characteristic of a target species within a liquid, comprising: a wearable device comprising a sensor including a capacitive element for contacting the liquid, wherein the sensor is configured to output a signal indicative of a capacitance of the capacitive element across a range of measurement frequencies; and a processor configured to receive the signal indicative of the capacitance and determine a characteristic of the target species from the capacitance of the capacitive element as a function of measurement frequency.

[0109] The processor may form part of the wearable device.

[0110] A further aspect provides a method of detecting a characteristic of a target species in a liquid, comprising the steps of: providing a capacitive element in contact with the liquid; measuring an impedance of the capacitive element when it is in contact with the liquid; calculating a difference between the measured impedance of the liquid and a reference value indicative of the impedance of the capacitive element when it is in contact with a known fluid; and determining a characteristic of the target species from the calculated difference.

[0111] The impedance of the capacitive element may be recorded at a constant frequency. However, in some examples the impedance of the capacitive element may be recorded at more than one frequency, as described above.

[0112] Determining the characteristic of the target species may include calculating AZ/Z₀ according to the following equation:

AZ > Z i — Z_o

ZQ Z_O where Zi: magnitude of the measured impedance when the capacitive element is in contact with the liquid, Zo : magnitude of the impedance when the capacitive element is in contact with the known fluid. [0113] The characteristic of the target species may be determined by comparing the calculated AZ

AZ/Z₀ with data indicative of the relationship between — and concentration of the target zo species.

[0114] The method may comprise a step of measuring an impedance of the capacitive element when it is in contact with the known fluid. The known fluid may be a liquid such as de-ionised water. The known fluid may act as a blank.

[0115] The capacitive element may or may not comprise an ion selective membrane, as described previously. In some examples, the method may comprise providing a plurality of capacitive elements in the manner described above, and may comprise providing an ion- selective membrane for one or more of the capacitive elements, for example all of the capacitive elements or all but one of the capacitive elements.

[0116] Another aspect provides an apparatus for detecting a characteristic of a target species within a liquid, comprising: a sensor, comprising a capacitive element for contacting the liquid and outputting a signal indicative of an impedance of the capacitive element when it is in contact with the liquid; and a processor configured to: receive the signal indicative of the impedance of the capacitive element when it is in contact with the liquid, receive a signal indicative of the impedance of the capacitive element when it is in contact with a known fluid, in dependence on the received signals determine a characteristic of the target species.

[0117] The impedance of the capacitive element may be recorded at a constant frequency. However, in some examples the impedance of the capacitive element may be recorded at more than one frequency, as described above.

[0118] Determining the characteristic of the target species may include calculating AZ/Z₀ according to the following equation:

AZ > Z i — Z_o ZQ Z_O where Z;: magnitude of the measured impedance when the capacitive element is in contact with the liquid, Zo : magnitude of the impedance when the capacitive element is in contact with the known fluid.

[0119] The characteristic of the target species may be determined by comparing the calculated AZ

AZ/Z₀ with data indicative of the relationship between — z and concentration of the target o species.

[0120] The known fluid may be a liquid such as de-ionised water. The known fluid may act as a blank.

[0121] The capacitive element may or may not comprise an ion selective membrane, as described above. In some examples, the apparatus may comprise a plurality of capacitive elements in the manner described previously, and may comprise an ion-selective membrane for one or more of the capacitive elements, for example all of the capacitive elements or all but one of the capacitive elements.

[0122] Features discussed in relation to each aspect of the invention are to be understood to be applicable to any other aspect of the invention unless the context clearly indicates otherwise.

[0123] Other technical features may be readily apparent to one skilled in the art from the following figures, descriptions, and claims.

BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS

[0124] Examples of the invention are described with reference to accompanying drawings, in which:

[0125] FIG. 1 illustrates an apparatus for detecting a characteristic of a target species within a liquid.

[0126] FIG. 2 illustrates a capacitive element for the apparatus of FIG. 1.

[0127] FIG. 3 A illustrates a stage of assembly of a sensor for the apparatus of FIG. 1.

[0128] FIG. 3B illustrates another stage of assembly of the sensor of FIG. 3 A.

[0129] FIG. 3C illustrates the sensor of FIGS. 3A to 3B when assembled. [0130] FIG. 4A illustrates a real component of impedance measurements, recorded using the apparatus of FIG. 1, plotted against measurement frequency.

[0131] FIG. 4B illustrates an imaginary component of impedance measurements, recorded using the apparatus of FIG. 1 , plotted against measurement frequency.

[0132] FIG. 5A illustrates a plot of the calculated effective capacitance against the measurement frequency for different concentration solutions of NaCl.

[0133] FIG. 5B illustrates a plot of turning point frequency against ion concentration for different concentration solutions of NaCl.

[0134] FIG. 5C illustrates a plot of turning point frequency against ion concentration for different concentration solutions of sodium lactate.

[0135] FIG. 6 illustrates a plot of the turning point frequency as a function of electrode gap of the capacitive element for various concentration solutions of NaCl.

[0136] FIG. 7A illustrates a plot of a real component of impedance measurements against measurement frequency for various capacitive elements having different electrode gaps.

[0137] FIG. 7B illustrates a plot of an imaginary component of impedance measurements against measurement frequency for various capacitive elements having different electrode gaps.

[0138] FIG. 8A illustrates a plot of turning point frequency against ion concentration for liquids comprising one of various chloride ion species.

[0139] FIG. 8B illustrates a plot of a change of impedance magnitude against target species concentration for NaCl and KC1 solutions in the presence of an ion selective membrane for K⁺.

[0140] FIG. 8C illustrates an apparatus, comprising ion-selective membranes, for detecting a characteristic of a target species within a liquid.

[0141] FIG. 9 illustrates an apparatus, comprising a Peltier element, for detecting a characteristic of a target species within a liquid.

[0142] FIG. 10A illustrates a plot of turning point frequency against temperature data for various 2mM ionic solutions.

[0143] FIG. 10B illustrates a bar chart representing gradients of turning point frequency against temperature for various ionic solutions. [0144] FIG. 11 A illustrates a plot of the impedance of a minima of the imaginary component of the impedance against the recorded temperature for each of 2mM solutions of various ionic solutions.

[0145] FIG. 1 IB illustrates a bar chart representing gradients of the impedance of the minima of the imaginary component against the recorded temperature for 0.5 to 2 mM solutions of various ionic solutions. FIG. 12 illustrates a schematic exploded view of a capacitive element comprising a MIP layer.

[0146] FIG. 13 illustrates a method of manufacturing the capacitive element of FIG. 12.

[0147] FIG. 14 illustrates results of two metabolite solutions.

[0148] FIG. 15 illustrates a method of detecting a characteristic of a target species in a liquid.

[0149] FIG. 16 is a plot of turning point frequency against ionic concentration.

[0150] FIG. 17A is an updated version of FIG. 5 A showing further data collected.

[0151] FIG. 17B is an updated version of FIG. 5B showing further data collected.

[0152] FIG. 17C is an updated version of FIG. 5C showing further data collected.

[0153] FIG. 18 is an updated version of FIG. 6 showing further data collected.

[0154] FIG. 19 is an updated version of FIG. 8 A showing further data collected.

[0155] FIG. 20A is an updated version of FIG. 10A showing further data collected.

[0156] FIG. 20B is an updated version of FIG. 10B showing further data collected.

[0157] FIG. 20C is a plot of percentage change in turning point frequency with temperature vs concentration for various ionic solutions.

[0158] FIG. 21 is a plot of the turning point frequency against concentration for Ca, K and Na ions, using an ion-selective membrane.

[0159] FIG. 22A is a plot of the measured impedance relative to a known solution against concentration of Ca²⁺, using an ion-selective membrane.

[0160] FIG. 22B is a plot of the measured impedance relative to a known solution against concentration of K⁺, using an ion-selective membrane. DETAILED DESCRIPTION

[0161] FIG. 1 shows an apparatus 100 for detecting a characteristic of a target species within a liquid such as sweat. The apparatus 100 comprises a sensor 102 including one or more sensor elements 104. Each sensor element 104 includes a capacitive element 106 for contacting the liquid. One example of the capacitive element 106 will be described in greater detail later on with reference to FIG. 2.

[0162] As shown in FIG. 1, in this example the sensor element 104 includes a channel 112 for receiving the liquid.

[0163] The sensor 102 comprises a substrate 114 on which the capacitive element 106 lies. The sensor 102 further comprises a cover layer 116 which at least partially covers the substrate 114. The capacitive element 106 is disposed between the cover layer 116 and the substrate 114. The channel 112 is formed in the cover layer 116 such that the capacitive element 106 lies in the channel 112. In some examples, the cover layer 116 may comprise or essentially consist of a flexible plastic such as polydimethylsiloxane (PDMS) or Flexdym®. As will be described later on with reference to FIG. 3 A, the cover layer 116 may be attached to the substrate 114 (e.g., by an adhesive) to seal edges of the channel against the substrate 114. In other examples, the cover layer 116 may not be attached (e.g., adhered) to the substrate 114.

[0164] An inlet 118 is provided at a first side 122 of the channel 112 for receiving the liquid. In this example, the channel 112 further includes an outlet 120 at a second side 124 of the channel 112. The outlet 120 may allow air within the channel 112 to escape when the liquid enters via the inlet 118. In some examples, the sensor 102 may be adapted for continuous monitoring. The outlet 120 may enable overflow liquid from the channel 112 to leave the channel 112 ensuring a continuous flow of liquid through the channel from the first side 122 to the second side 124. In particular, the channel 112 may be a microfluidic channel. Such a microfluidic channel may only require a small volume of liquid to be filled, thereby completely covering the capacitive element 106 with the liquid. As such, the apparatus 100 may reliably determine the characteristic of the target species for a small sample of liquid (e.g., less than 10'⁵ liters).

[0165] As shown in FIG. 1, the capacitive element 106 lies directly beneath the channel 112. Liquid within the channel 112 directly contacts the capacitive element 106. In this example, the sensor 102 comprises an impedance spectroscopy circuit 108. The impedance spectroscopy circuit 108 is configured to measure an impedance of the capacitive element 106 across a range of measurement frequencies.

[0166] The apparatus 100 comprises a processor 110. The processor 110 is configured to calculate the capacitance of the capacitive element 106 across the range of measurement frequencies from a signal received from the sensor 102, in particular from an impedance signal received from the sensor 102.

[0167] Variation in ion concentration is known to vary the capacitance of fluids. Relationships between the measured capacitance and ion concentration are complex functions that depend both on the species of ion being measured and the measurement parameters used such as the specific capacitor arrangement and the measurement frequency sampled. These factors pose challenges to when attempting to use the measured capacitance to infer characteristics of the liquid.

[0168] Surprisingly, it has been discovered by the applicant that a turning point of the capacitance as a function of measurement frequency is strongly correlated with the ion concentration of the liquid. In particular, the frequency of the turning point has been found to be linearly correlated with ion concentration for various ions tested. As a result, the turning point data may be more easily interpreted than the capacitance measurements. In contrast, a measured capacitance at a given frequency could be produced by multiple different ion concentrations. Alternately, inferring the concentration of a solution from the capacitancefrequency curve would require complex curve fitting programs which are prone to error.

[0169] In the present application, the ion concentration can be inferred from the position of the turning point frequency. The linear relation between the turning point frequency and the ion concentration may also ensure that the accuracy of the results is more constant over the concentration range of interest. Systems and methods using this turning point data may provide a more reliable way to measure ion concentration or to deduce the type of ion present where the concentration is known.

[0170] The apparatus 100 comprises a processor 110 configured to determine turning point data representative of a turning point of the capacitance as a function of the measurement frequency. The processor 110 is also configured to determine a characteristic of the target species from the turning point data. For example, where the type of target species present in the 1 liquid is known, the characteristic may be indicative of the concentration of the target species. Where, instead, the concentration of the target species in the liquid is known, the characteristic may be indicative of the identity of the target species present in the liquid.

[0171] For example, the apparatus 100 may comprise a memory module 128 which comprises stored turning point data representing turning points of the capacitance of the capacitive element as a function of measurement frequency for a plurality of known calibration samples. The processor 110 may be configured to compare the stored turning point data to turning point data of the liquid being measured to determine the characteristic of the target species in the liquid.

[0172] The apparatus 100 may comprise a results memory 132 configured to store 'characteristic data' representing a characteristic of the liquid determined by the apparatus 100. For example, the characteristic data may represent the concentration and/or identity of the target species in the liquid. As shown in FIG. 1, the processor 110 may be configured to transmit the characteristic data to the results memory 132, for example by a wired or wireless connection.

[0173] The processor 110 may be configured to transmit sample identifiers to the results memory 132. The characteristic data in the results memory may be indexed by the sample identifiers. For example, the sample identifier may comprise a timestamp. Each entry of characteristic data may be associated with a timestamp representing a date and time at which the measurements used to calculate that characteristic data were recorded. The sample identifier may further include identification data representative of either the identity of the sensor 102, the sensor element 104 or the identity of the user of the sensor 102 at the time the measurements used to calculate that characteristic data were recorded.

[0174] The apparatus 100 may comprise a user interface 136. The user interface 136 may be configured to receive the characteristic data from the results memory 132 by wired or wireless connection. The user interface 136 may include a display for viewing the characteristic data. In some examples, the user interface 136 may be configured to allow a user to access and navigate characteristic data stored in the results memory 132. For example, the user interface 136 may comprise an interactive touch screen display.

[0175] In some examples, the apparatus 100 may comprise a wearable device 134. In examples where the wearable device 134 is present, the sensor 102 forms part of the wearable device 134. In some examples, the apparatus 100 may comprise an external module (not shown) which is wirelessly connected to the wearable device 134.

[0176] Each or any combination of: the processor 110; the memory module 128; the results memory 132; and the user interface 136 may form part of the wearable device 134 or alternately may form part of the external module. In the example shown in FIG. 1 , the processor 110, the sensor 102, the results memory 132 and the user interface 136 all form part of the wearable device 134. In another example, the processor 110 and the memory module 128 form part of the external module while the sensor 102, the results memory 132 and the user interface 136 form part of the wearable device 134. In a yet further example, the processor 110 and the memory module 128 may form part of the external module while the results memory 132, the sensor 102 and the user interface 136 may form part of the wearable device 134. In yet a further example, the memory module 128 may form part of the external module while the sensor 102, the processor 110, the results memory 132, the user interface 136 form part of the wearable device 134.

[0177] The wearable device 134, may comprise a strap (not shown) for attaching a user e.g., to a limb of the user. For example, the strap may comprise a watch strap for attaching to an arm of the user.

[0178] The sensor 102 may be shaped to conform to the user's limb. For example, the substrate 114, and optionally the cover layer 116, may be arcuate in shape. Alternately or in addition, the sensor 102 may be operable to bend to conform to the shape of a user's limb. For example, the substrate 114 may comprise or essentially consist of a polymer such as silicone. The cover layer 116 may comprise polymer or essentially consist of a polymer such as poly dimethylsiloxane (PDMS) or Flexdym®. The capacitive element 106 may be configured to bend when the sensor 102 is flexed. For example, the capacitive element 106 may comprise carbon nanotubes. For example, the capacitive element may be Aerosol Jet printed onto the substrate 114 using carbon ink. The carbon ink may comprise lwt% of multi- walled carbon nanotubes (MWCNT, Sigma- Aldrich) blended with the 0.5wt% of Polyvinylpyrrolidone (PVP, Sigma-Aldrich) and de-ionised water.

[0179] The sensor 102 may comprise a reservoir (not shown) for receiving the liquid. The reservoir may comprise a liquid collector for drawing liquid from a liquid source into the reservoir. Where present, the reservoir is fluidically coupled to the inlet 118 of the channel 112 such that liquid within the reservoir can be delivered to the channel 112 through the inlet 118.

[0180] In some examples, the cover layer 116, including the channel 112, may be omitted. Instead, the capacitive element 106 may form part of an external surface of the sensor 102. The capacitive element 106 may be exposed such that in use the capacitive element 106 contacts a surface of a skin of the user. Liquid, such as sweat, disposed on the skin may contact the capacitive element 106 so that the characteristic of a target species within the liquid can be detected.

[0181] The apparatus may comprise a power source, such as a battery, operable to supply power to at least the sensor 102. In the example shown in FIG. 1, the power source is a battery which forms part of the wearable device 134. The battery is configured to supply energy to the each of the components in the wearable device 134 including the processor 110 and the user interface 136.

[0182] In some examples, the apparatus 100 may be configured to function continuously. In other examples, apparatus 100 may comprise a means to switch the apparatus 100 between an OFF mode and an ON mode. When the apparatus 100 is in the ON mode, the apparatus 100 determines: the capacitance of the capacitive element across the range of measurement frequencies; the turning point data; and the characteristic of the target species from the turning point data. When the apparatus 100 is in the OFF mode, the apparatus 100 does not determine the capacitance of the capacitive element across the range of measurement frequencies. When the apparatus 100 is in the OFF mode, the apparatus may not determine the characteristic of the target species from the turning point data, and/or may not determine the characteristic of the target species from the turning point data.

[0183] Providing a means for switching the apparatus 100 between the OFF mode and the ON mode may reduce the energy requirements of the apparatus 100. Beneficially, this may reduce the environmental impact and running costs associated with using the apparatus 100. Where the battery is present in the wearable device 134, providing the means for switching the apparatus 100 between the OFF mode and the ON mode may reduce a size of the battery required to operate the apparatus over a given time period thereby reducing the weight and size of the wearable device 134. [0184] For example, the apparatus 100 may comprise a pressure sensor (not shown). The pressure sensor may be operable to detect whether or not the wearable device 134 is secured to the user. In some examples, the pressure sensor may be disposed on an interior surface of the wearable device 134 and configured to face the skin of the user. The pressure sensor may be configured to actuate the apparatus 100 between the OFF mode and the ON mode. For example, the pressure sensor may be configured to actuate the apparatus 100 from the OFF mode to the ON mode when the pressure sensor detects the presence of the user. The pressure sensor may be configured to actuate the apparatus from the ON mode to the OFF mode when no user is detected by the pressure sensor.

[0185] In some examples, the sensor 102 may comprise a liquid sensor (not shown) operable to detect the presence of the liquid (e.g., liquid at or near the capacitive element 106). The liquid sensor may be configured to actuate the apparatus 100. For example, the liquid sensor may be configured to actuate the apparatus 100 from an OFF mode to an ON mode when the liquid sensor detects the presence of liquid. The liquid sensor may be configured to actuate the apparatus from the ON mode to the OFF mode when no liquid is detected or when the liquid sensor detects less than a minimum threshold of liquid. For example, the minimum threshold may be representative of the quantity of liquid required achieve full contact of liquid over the capacitive element 106. Alternately, the minimum threshold may represent a minimum time period for which liquid has been continuously detected by the liquid sensor (e.g., 5 seconds). The liquid sensor may be operable to detect a sweat rate. For example, the liquid sensor may be configured to actuate the apparatus 100 from an OFF mode to an ON mode when the liquid sensor detects a sweat rate exceeding a minimum flow threshold. The liquid sensor may be configured to actuate the apparatus from the ON mode to the OFF mode when the detected sweat rate falls below the minimum flow threshold. In examples, the sweat rate may be displayed to a user e.g., via the user interface 136.

[0186] In an alternative, the liquid sensor may be omitted and the sensor 102 may be configured to actuate the apparatus 100 between the OFF mode and the ON mode (and vice versa). For example, the sensor 102 may be configured to switch the apparatus 100 from the OFF mode to the ON mode when the measured impedance decreases below an impedance threshold. The sensor 102 may be configured to switch the apparatus 100 from the ON mode to the OFF mode when the measured impedance exceeds the impedance threshold. The impedance threshold may be representative of the maximum impedance of the capacitive element 106 when the capacitive element 106 is fully covered by a liquid. The measured impedance exceeding the maximum impedance threshold may indicate that the capacitive element 106 is not completely covered by the liquid.

[0187] In some examples, the apparatus 100 may comprise an actuator (not shown), such as a push button, for user switching between the ON mode and the OFF mode (and visa versa). The actuator may form part of the wearable device 134. In some examples, the user interface 136 may function as the actuator, allowing the user to switch the apparatus between the OFF mode and the ON mode.

[0188] In some examples, the apparatus 100 may include a timer (not shown) for switching the apparatus between the OFF mode and the ON mode (and visa versa). The processor 110 may include the timer. The timer may be configured to periodically actuate the apparatus 100 according to a duty cycle. For example, the duty cycle may comprise continuous cycle between: a first period where the apparatus 100 is in the ON mode; and a second period where the apparatus is in the OFF mode. The first period may be shorter than the second period. For example, the first period may be less than 120 seconds. Preferably the first period may be less than 60 seconds, less than 10 seconds, or less than 5 seconds. The second period may be greater than 1 minute, 10 minutes or 30 minutes. Such an arrangement may allow the liquid to be periodically sampled by the apparatus 100. In some examples, the timer may be used in combination with the liquid sensor. The processor may determine the duty cycle according to the sweat rate.

[0189] In some examples, the duty cycle may be programmable via the user interface 136.

[0190] It will be understood by the skilled person that multiple different means for actuating the apparatus 100, as described above, may be used in combination. For example, the actuator may be used in combination with the timer such that when the actuator is switched ON (e.g., the push button engaged), the apparatus switched between the ON mode and the OFF mode periodically according to the timer. When the actuator is switched OFF (e.g., the push button released) the apparatus 100 may be in the OFF mode regardless of a status of the timer.

[0191] FIG. 2 illustrates the capacitive element 106 in isolation from the remainder of the sensor 102. As shown, the capacitive element 106 includes first electrode 201 and a second electrode 202. The first electrode 201 and the second electrode 202 are connected to the impedance spectroscopy circuit 108 at a first contact pad 203 and a second contact pad 204 respectively. The first electrode 201 is spaced from the second electrode 202. In this example, the first electrode 201 and the second electrode 202 are interdigitated electrodes. In particular, the first electrode 201 comprises a first backbone 205 from which a first set of protrusions 206 project. The second electrode 202 comprises a second backbone 207 from which a second set of protrusions 208 project. The first set of protrusions 206 comprises a plurality of linear protrusions which project substantially perpendicularly away from the first backbone 205 towards the second backbone 207. Similarly, the second set of protrusions 208 comprises a plurality of linear protrusions which project perpendicularly away from the second backbone 207 towards the first backbone 205. In other examples, the first set of protrusions 206 and the second set of protrusions 208 may be angled with respect to the first backbone 205 and the second backbone 207 respectively.

[0192] The first set of protrusions 206 and the second set of protrusions 208 interdigitate along a length of the capacitive element 106. The first set of protrusions 206 and the second set of protrusions 208 are evenly spaced along the length of the capacitive element 106. In this example, a center-to-center electrode distance 220 was 80pm. in other examples, the center-to- center electrode distance 220 may be less than or equal to 200pm, 150pm, 120pm, 100pm or 90pm. The center-to-center electrode distance 220 may be greater than or equal to 140pm, 120pm, 100pm or 80pm.. An electrode gap 209 is provided between each neighbouring one of the first set of protrusions 206 and the second set of protrusions 208. Each of the first set of protrusions 206 may be parallel to each of the second set of protrusions 208 such that the electrode gap 209 is constant along a width of the capacitive element 106.

[0193] Free ends of the first set of protrusions 206 are spaced from the second backbone 207 while free ends of the second set of protrusions 208 are spaced form the first backbone 205. The first electrode 201 and the second electrode 202 define a serpentine path therebetween.

[0194] The capacitive element 106 may be formed by any suitable means. For example, the exemplar capacitive elements 106 shown in FIG. 2 were formed by way of Aerosol jet printing (AJP). In particular, using an Optomec AJ200 (Optomec Inc., New Mexico, USA) and silver ink (produced by diluting Novacentrix Ag nanoparticles aerosol ink (JS-A221AE) 1:2 by volume with de-ionised water). The capacitive element 106 was AJP directly onto the substrate 114 (a glass slide in this example). The glass slide was cleaned using iso-propyl alcohol (Merck, Germany), prior to depositing the capacitive element 106. The ink and sheath flow rates were typically approximately 18 seem and 80 seem respectively, and were manually set and adjusted throughout printing to maintain print quality. The silver was cured in an oven (HeraTherm OGH60, Thermo Fisher Scientific) at 200 °C for 2 hours.

[0195] FIG. 3 A to FIG. 3C illustrate stages of assembly of the sensor 102. In these figures, the sensor 102 is shown to comprise a plurality of sensor elements 104. Each of the plurality of sensor elements 104 may be used to test a different sample of the liquid (e.g. To provide repeat measurements). In this example the sensor 102 comprises eight sensor elements 104 which alternate in orientation along a length of the substrate 114. It will be understood that the sensor shown in FIG. 3 A to FIG. 3C is merely an example and that the sensor 102 may comprise any number of sensor elements 104. For example, the sensor 102 may comprise one, two, five, ten or more sensor elements 104.

[0196] FIG. 3 A shows an exemplar cover layer 116, prior to assembly of the sensor 102. In this example, the cover layer 116 comprises PDMS. Specifically the example shown in FIG. 3A was formed from PDMS base (Sylgard-184, Dow Chemical Company, Michigan, USA) and curing agent were mixed 10: 1 (w/w), and degassed using a desiccator. The PDMS was poured into a re-useable mold and cured at 60 °C for a minimum of 3 hours, then removed from the reusable mold.

[0197] The cover layer 116 comprises: a basal surface 211 which faces the substrate 114; and a top surface 212 opposite the basal surface 211. As shown in FIG. 3A, the channels 112 of the sensor elements 104 are formed in the basal surface 211 of the cover layer 116.

[0198] In this example, the inlet 118 comprises a tunnel that projects substantially away from the basal surface 211. In use, liquid is inserted into the channel 112 using a needle such as a hypodermic needle. The needle penetrates the cover layer 116 at the inlet 118 and a syringe fluidically attached to the needle is depressed to inject the liquid into the channel 112. Similarly, the outlet 120 comprises a tunnel that projects substantially away from the basal surface 211. A needle, such as a hypodermic needle, may be used to penetrate the outlet 120 to open the channel 112 at the second side 124. As liquid is injected into the channel 112 at the inlet 118, existing air or liquid within the channel 112 may be forced out through the outlet 120.

[0199] In other examples, the inlet 118 may comprise a first hole which perforates the cover layer 116 fluidically connecting the top surface 212 with the first side 122 of the channel 112. Similarly, the outlet 120 may comprise a second hole which perforates the cover layer 116 fluidically connecting the top surface 212 with the second side 124 of the channel 112.

[0200] FIG. 3B illustrates another stage of assembly of the sensor 102. In this stage, the capacitive elements 106 are deposited onto the substrate 114. In this example, the first and second electrodes 201, 202 and contact pads 203, 204 of the capacitive elements 106 were printed directly onto the substrate 114 by AJP as described above in relation of FIG. 2. As shown in FIG. 3C, wires 126 were then connected to each of the contact pads 203, 204 for connecting to the impedance spectroscopy circuit (not shown).

[0201] As shown by comparison of FIG. 3 A and FIG. 3C, the sensor 102 comprises an adhesive layer 210 disposed between the substrate 114 and the cover layer 116. For example, the adhesive layer 210 may comprise adhesive or double-sided adhesive tape. During assembly, the cover layer 116 is aligned with respect to the substrate 114 to ensure the channel 112 of each sensor element 104 is aligned with the capacitive element 106 of that sensor element 104. The adhesive layer 210 is then sandwiched between the cover layer 116 and the substrate 114 to fixedly secure the cover layer 116 to the substrate 114. The adhesive layer 210 may further act to seal the cover layer 116 to the substrate 114 preventing liquid ingress therebetween.

[0202] In this example, the adhesive layer 210 consists of double-sided tape perforated by one or more slots 213. In particular, this example utilised double-sided tape (Tesa, Germany) of 90 pm thickness that was laser cut using an Epilog Zing 16 (30 Watt) laser cutter (Epilog Laser, Colorado, USA), with a laser frequency of 2500 Hz, and speed and power of 90% and 12% respectively. Each of the one or more slots 213 is configured to ensure contact between the capacitive element 106 and the channel 112 of that respective sensor element 104 is not obstructed by the adhesive layer 210. In this example, each slot 213 comprises a rectangle with dimensions 1.2 mm x 26 mm.

[0203] In this example, the channel 112 is a microfluidic channel having a width of 1mm, a height of 500pm and a length of 26mm. A width of the capacitive element 106 is sized to fit within the width of the channel 112. In this example, the width of the capacitive element 106 is less than 1mm.

[0204] The first electrode 201 and the second electrode 202 project away from the substrate 114 towards the impedance spectroscopy circuit 108 (not shown). The impedance spectroscopy circuit 108 is configured to measure the impedance of the capacitive element 106 across the range of measurement frequencies. In some examples, separate impedance spectroscopy circuits may be provided for each sensor element 104. In other examples, the impedance spectroscopy circuit 108 may be operably connected to each of the sensor elements 104. The impedance of each of the capacitive elements 106 of the sensor elements 104 may be measured simultaneously or sequentially (e.g. actuated by a manual or electronic switching means). In this example, only one capacitive element 106 is measured by the impedance spectroscopy circuit 108 at any time. The impedance spectroscopy circuit 108 may be configured to record a real component and an imaginary component of the impendence of the capacitive element across the range of frequencies.

[0205] FIG. 4A and FIG. 4B illustrate the impedance measurements recorded when different concentration NaCl solutions filled the channel 112. The impedance measurements were acquired using a Sciospec ISX3v2 Impedance Analyser with a SlideChipAdapter (Sciospec Scientific Instruments GmbH, Germany). The impedance analyser was operated in voltage- controlled mode, with an excitation amplitude of 250 mV and a current measurement range of ± 100 pA. Measurements were made in 2048 logarithmically spaced steps between 10 kHz and 25 MHz or 2 kHz and 25 MHz inclusive, at a precision setting of 1 (the precision setting being directly correlated to the relative bandwidth of the measurement). Three repeats of each measurement were carried out. Between measurements, the channel 112 was emptied by wicking and then refilled. Anomalous data (e.g. due to the presence of bubbles or incomplete filling of the channel) were manually identified and removed. Prior to taking measurements, the impedance analyser was calibrated against the effect of parasitic capacitances using an open circuit, closed circuit and known load (330 resistor).

[0206] Test solutions of aqueous NaCl used herein were produced according to the following procedure. NaCl (Merck, Germany) was added to de-ionised water (filtered in-house, Purite, UK, typical conductivity < 10 pS) to concentrations of 0.5 mM to 2.5 mM, in 0.5 mM steps.

[0207] FIG. 4A shows a first graph 401 illustrating the real component of the impedance measurements recorded by the impedance spectroscopy circuit 108. Reference numerals for each line of the first graph 401, representing the real component of the impedance signal for different concentration solutions of NaCl, are provided in Table 1 below.

Table 1: Reference numerals for lines in FIG. 4A representing the real component of the impedance signals for samples of different concentration ofNaCL

[0208] As illustrated in the first graph 401, the real component of the impedance is dependent on the ion concentration of the solution. However, the relationship between ion concentration and recorded real impedance signal is non-linear. For example, at 10⁴ Hz the difference between the recorded real impedance signal for 0.5-lmM NaCl (lines 411 and 412) is significantly greater than the difference between the recorded real impedance signal for 2- 2.5mM NaCl (lines 414 and 415).

[0209] FIG. 4B shows a second graph 402 illustrating the imaginary component of the impedance measurements recorded by the impedance spectroscopy circuit 108. Reference numerals for each line of the second graph 402, representing the imaginary component of the impedance signal for different concentration solutions of NaCl, are provided in Table 2 below.

Table 2: Reference numerals for lines in FIG. 4B representing the imaginary component of the impedance signals for samples of different concentration ofNaCL

[0210] As illustrated in the second graph 402, the imaginary component of the impedance is dependent on the ion concentration of the solution. However, the relationship between ion concentration and recorded imaginary impedance signal is non-linear. For example, at 10⁶ Hz the difference between the recorded real impedance signal for 0.5-lmM NaCl (lines 421 and 422) is significantly greater than the difference between the recorded real impedance signal for 2-2.5mM NaCl (lines 424 and 425).

[0211] The magnitude of both the real and imaginary signals decreases non-linearly with increasing NaCl concentration for most frequencies. This trend is observed at above ~10⁶ Hz for the real signal and below ~10⁷ Hz for the imaginary data.

[0212] Using a simplified circuit model consisting of a resistor and a capacitor in series, the effective capacitance (C ) of the circuit was calculated from the imaginary part of the impedance signal (Im(Z)), using the formula:

(Equation 1)

[0213] Where angular frequency co = 2nf, where f is the measurement frequency.

[0214] FIG. 5 A shows a third graph 510 illustrating a plot of the calculated effective capacitance against the measurement frequency for the different concentration solutions of NaCl. FIG. 17A shows an updated version of the graph of FIG. 5 A. As illustrated, the graph 510 plots the log of the calculated effective capacitance against the log of the measurement frequency. Reference numerals for each line of the third graph 510, representing the effective capacitance of the impedance signal for different concentration solutions of NaCl, are provided in Table 3 below.

Table 3: Reference numeral for lines in FIG. 5 A and FIG. 17A representing the effective capacitance, calculated according to Equation 1, for samples of different concentration of NaCl. [0215] The third graph 510 shows the effective capacitance has a strong dependence on NaCl concentration. Turning points of the capacitance with respect to the frequency are indicated by dots on the respective lines 511-515 in FIG. 5 A and FIG. 17A.

[0216] Surprisingly, it has been discovered by the applicant that the turning points are strongly correlated with the ion concentration of the liquid. In particular, the frequency of the turning points has been found to be linearly correlated with ion concentration for various ions tested. This is shown by FIG. 5B and FIG. 17B, which show a fourth graph 520 plotting the turning point frequency (TPF) against the ion concentration for five solutions of NaCl, illustrated by fitted line 522.

[0217] As shown by the fitted line 522, a linear correlation between the TPF and the concentration was observed. R² = 1.000, r = 0.9998, p = 2.1xl0'⁶.

[0218] Considering an NaCl solution of unknown concentration, this linear relationship may be used to determine the concentration of NaCl in such a solution. Table 4 below shows calculated concentrations of various NaCl test solutions determined using the above methodology on the apparatus 100.

Calculated Within error?

Tested cone. / mM cone. / mM

0.50 0.47 ± 0.04 Y

1.00 1.02 ± 0.04 Y

1.50 1.49 ± 0.05 Y

2.00 2.02 ± 0.04 Y

2.50 2.47 ± 0.08 Y

Table 4: Calculated concentrations of NaCl solutions, their associated uncertainties defined at the standard deviation over three refills of the same channel 112, the actual concentrations of NaCl in the solutions tested, and indication of whether the actual concentrations lie within the error of the calculated concentrations.

[0219] The calculated concentrations show good agreement with the known concentrations, and the errors are relatively small. The average relative uncertainty was 4.1%. In all cases the tested concentration falls within the error bounds of the calculated concentration. This demonstrates that the apparatus 100 and TPF based analysis described above is suitable for the characterisation of NaCl species. [0220] Due to the linear relationship between the TPF and the ion concentration, the concentration of the target ion in the liquid may be more easily interpreted from the TPF than from the impedance or the capacitance measurements (which are not linearly dependent on the ion concentration). Additionally, the TPF is less likely to be affected by sample-to-sample variation.

[0221] The linear relationship between the TPF and the ion concentration may also provide enhanced consistency in the sensitivity of the determined concentration over a set concentration range. As such, systems and methods using this TPF may provide a more reliable way to measure ion concentration of a liquid of unknown concentration than existing electrical detection methods. In particular, the systems and methods described herein may further be used to measure the concentration of a metabolite in a liquid.

[0222] This is shown in FIG. 5C and FIG. 17C, which illustrate a plot of the TPF against ion concentration for four different concentration solutions of sodium lactate. In FIG. 17C the turning point frequency (TPF) against the ion concentration for five solutions of NaCl is also illustrated (fitted line 522).

[0223] As shown by the fitted line 523, a linear correlation between the TPF and the concentration was observed. Considering a sodium lactate solution of unknown concentration, this linear relationship may be used to determine the concentration of sodium lactate in such a solution.

[0224] As the TPF is dependent on temperature, the apparatus 100 may comprise a temperature sensor (not shown) to enable temperature variation to be accounted for in post-processing. The temperature sensor may be configured to measure an environmental temperature or a temperature of the liquid in the channel 112. In some examples, the temperature sensor may be located on the substrate proximal to the channel. The processor may be configured to use measurements received from the temperature sensor to determine the characteristic of the target species. FIG. 20C illustrates a plot 1200 of the percentage change in turning point frequency (TPF) with temperature vs concentration for various ionic solutions. As shown, for a given change in temperature, the relative change in the TPF values is found to be independent of the ionic species and concentration. The apparatus 100 may therefore only require one calibration curve, which can be applied in the same manner to measurements, to account for the effect of temperature variation on TPF for all target species. [0225] The apparatus 100 may also have several advantages over existing chemical methods for recording target species concentration. Existing chemical methods of recording target species typically utilise one or more assays which are chemically altered during ion detection (e.g., to effect the colour change indicative of the presence of ions in colorimetric methods). Devices using such methods are either single use, or must be re-supplied with assay(s) after each use. In contrast, the apparatus 100 of the present disclosure is wholly re-usable. The characterisation utilising only purely electrical detection means and requiring no chemical reactions, no assay is required and so the running costs are significantly reduced when compared with existing chemical methods.

[0226] FIG. 6 shows a fifth graph 600 representing the TPF as a function of the center-to- center electrode distance 220 for various solutions of NaCl. FIG. 18 shows an updated version after collection of further data. Reference numerals for each line of the fifth graph 600, representing the TPF for different concentration solutions of NaCl, are provided in Table 5 below.

Table 5: Reference numerals for linear fit lines in FIG. 6 and FIG. 18 representing the turning point frequency as a function of the center-to-center electrode distance 220 for various solutions of NaCl.

[0227] As shown, for each of the concentrations of NaCl tested, the TPF was largely independent of the electrode spacing. In particular, a gradient of the TPF plotted against the center-to-center electrode distance 220, averaged over the five different NaCl concentrations tested, was -(1.7 ± 1.4) kHz pm ¹. [0228] Independence of the ion concentration from the center-to-center electrode distance 220 may be beneficial, limiting the dependence of the ion concentration on the capacitor element configuration. For example, such independence may limit the effect small manufacturing variations of the capacitive element 106 have on the determined ion concentration.

[0229] In contrast, it is apparent that an ion concentration deduced directly from the impedance (real or imaginary components) or the capacitance will be highly dependent on the configuration of the capacitive element 106. This is illustrated by FIG. 7A and FIG. 7B which show real and imaginary components of the impedance measurements recorded by the impedance spectroscopy circuit 108 for various capacitive elements 106 having different center-to-center electrode distances 220. Each of the capacitive elements 106 tested in FIG. 7A and FIG. 7B was substantially similar to the capacitive element 106 shown in FIG. 2. Each of the capacitive elements 106 tested in FIG. 7A and FIG. 7B had a length of 20 mm, center-to- center electrode distance 220 and the number of electrodes differed for each of the capacitive elements 106 tested.

Table 6: Reference numeral of lines representing the real and imaginary components of the impedance signal.

[0230] This linear relationship between TPF and ion concentration is not limited to NaCl specifically and has been reproduced for various chloride ion species.

[0231] FIG. 8A shows a sixth graph 800 representing plots of the TPF against the ion concentration for liquids comprising one of various chloride ion species. FIG. 19 is an updated version of the graph of FIG. 8A after further data was collected. In particular, the chloride ion species tested were: NaCl (line 801); KC1 (line 802); MgCh Oine 803)_; and CaCE (line 804). Test solutions of the aqueous ionic chloride solutions (NaCl, KC1, MgCE and CaCh) used herein were produced according to the following procedure. NaCl, KC1, MgCh or

CaCh (Merck, Germany) was added to de-ionised water (filtered in-house, Purite, UK, typical conductivity < 10 pS) to concentrations of 0.5 mM to 2.5 mM, in 0.5 mM steps.

[0232] As shown by FIG. 8A and FIG. 19, a linear relationship between the TPF and the ion concentration of the liquids was observed for each of the ion species tested.

[0233] Different species exhibited different gradients of the TPF over the ion concentration. These results are provided in Table 7 below.

Species TPF-conc gradient / kHz mM ¹

NaCl 797 ± 8

KC1 948 ± 13

MgCh 1612 ± 44

CaCh 2221 ± 22

Table 7: Gradients of the linear fits relating the turning point frequency (TPF) to the ion concentration (as shown in FIG. 8 A and FIG. 19).

[0234] As shown, gradients of the monovalent cation species are lower than those of the divalent cation species. The gradients of the species with cations with larger radii (K, Ca) have larger gradients than those with smaller radii (Na, Mg). These gradients are equal to the sensitivity of the device (which is inherently dependent on the resolution of the frequency sweep used).

[0235] As the gradients of the linear fits differ for different ion species, an unknown type of ion present in a solution of known concentration of that ion may be determined using the turning point frequency.

[0236] When using the apparatus 100 to determine the concentration of the target species, the identity of the target species must be known. Similarly, when using the apparatus 100 to determine the type of species present, the concentration of the unknown species must be known. This limitation stems from there being equivalent solutions to any given TPF result. Taking the species of FIG. 8A and FIG. 19 as an example, it can be seen that a given TPF 810 may correspond to multiple different equivalent solutions. For example, the given TPF 810 may correspond to a ImM CaCh solution or a 1.5mM MgCh solution. [0237] Where both the identity of the target species and the concentration of the target species are unknown, further information may be required to identify which of the equivalent solutions is correct.

[0238] One way of circumventing the problem of equivalent solutions is to make use of ion- selective membranes. Ion-selective membranes are known to select for a single type of ion by diffusion using ionophores. A wide range of ion-selective membranes are available, each tailored to allow diffusion of a specific ion and inhibit or prevent diffusion of other ions.

[0239] FIG. 8B illustrates impedance results from a sensor element comprising an ion- selective membrane. In this example, the ion-selective membrane was configured to select for KC1. The ion-selective membrane was deposited directly onto the electrodes using dropcasting. In some examples, the ion-selective membrane may be deposited using aerosol jet printing or spin coating.

[0240] FIG. 8B shows a plot of a change of impedance magnitude, relative to de-ionised water, against target species concentration for NaCl and KC1 solutions. As shown, there is a significantly larger response to a change in concentration of KC1 than for NaCl. This illustrates that the ion-selective membrane is selecting for KC1. In some examples, this plot of a change of impedance magnitude relative to a blank difference may be used to determine a characteristic (such as concentration) of the target species (without requiring a calculation of the turning point frequency).

[0241] FIG. 8C shows a further example apparatus 820 for detecting characteristics of one or more target species within a multi-ion liquid such as sweat. For conciseness, previously introduced reference numerals have been used in FIG. 8C to indicate corresponding components to the components of the apparatus 100.

[0242] The apparatus 820 comprises a sensor 102 including a plurality of sensor elements 104a-d. In this example, the sensor 102 comprises four sensor elements 104a-d. Each sensor element 104a-d includes a capacitive element 106, for contacting the liquid, and a channel 112 for receiving the liquid. The sensor 102 further comprises an impedance spectroscopy circuit 108 configured to measure an impedance of the capacitive elements 106 across a range of measurement frequencies. The apparatus 820 further comprises a processor (not shown) configured to calculate the capacitance of the capacitive elements 106 across the range of measurement frequencies from impedance signals received from the sensor 102. [0243] The sensor 102 comprises a substrate 114 on which the capacitive elements 106 lie. The sensor 102 further comprises a cover layer 116 which at least partially covers the substrate 114. The capacitive elements 106 are disposed between the cover layer 116 and the substrate 114. The channels 112 are formed in the cover layer 116 such that each capacitive element 106 lies in a channel 112.

[0244] An inlet 118 is provided at a first side 122 of each of the channels 112 for receiving the liquid.

[0245] In some examples, each sensor element 104a-d may comprise an ion-selective membrane 821a-d which covers the inlet 118 of the channel 112. Each sensor element 104a-d may be provided with a different ion-selective membrane 821a-d. I.e., each of the different ion- selective membranes being configured to select for a different ion. In this example, each of the four sensor elements 104a-d is provided with a different ion-selective membrane 821 a-d. In other examples, only some of the sensor elements 104a-d may be provided with an ion- selective membrane 821 a-d.

[0246] As shown in FIG. 8C, the sensor 102 may comprise a reservoir 822 for receiving the liquid. The reservoir 822 is fluidically coupled to the inlet 118 of each channel 112 so that liquid may be supplied from the reservoir 822 to the channels 112 through the inlets 118.

[0247] For each sensor element 104a-d, liquid entering the channel 112 via the inlet 118 is filtered through the ion-selective membrane 821 a-d. As such, the type of ion present in each channel 112 is known. This circumvents the equivalent solutions problem for solutions of unknown ion type and unknown concentration as there is no longer uncertainty on the type of ion present within the channel 112. The TPF recorded for a given sensor element 104a-d can therefore be directly attributed to the presence of an individual ion.

[0248] The sample identifier of each sensor element 104a-d may include identification data representative the type of ion-selective membrane 821 a-d comprised in that sensor element 104a-d. The processor 110 may be configured to analyse the characteristic data received from each sensor element 104a-d in combination with the identification data from each sensor element 104a-d to determine the concentration and the type of target species present in the liquid.

[0249] Use of ion-selective membranes circumvents equivalent TPF solutions for single electrolyte solutions allowing the concentration and type of such single electrolyte solutions to be correctly identified. However, ion-selective membranes may further enable the apparatus 100 to determine the individual concentrations of different ions within a multiple ion solution such as sweat. The processor 110 may be configured to analyse the characteristic data received from each sensor element 104a-d in combination with the identification data from each sensor element 104a-d to determine the concentration and the type of each target species present in the liquid.

[0250] The number of sensor elements 104a-d and type of ion-selective membranes used therein may be selected depending on the specific application for which the apparatus 100 is to be used. Where little or nothing is known about the constituent ions of the liquid, the sensor element 104a-d may comprise a large array of sensor elements 104a-d. For certain applications, the likely types of ions present in the liquid are known prior to characterisation using the apparatus 100. For example, sweat is known to comprise Na ions, K ions and Cl ions in varying proportions as well as trace amounts of other ions such as Mg ions and Ca ions.

[0251] In this example, the apparatus 820 is to be used to measure sweat, and the sensor 102 comprises: a first sensor element 104a comprising an ion-selective membrane 821a for selecting Na ions; a second sensor element 104b comprising an ion-selective membrane 821b for selecting K ions; a third sensor element 104c comprising an ion-selective membrane 821c for selecting Mg ions; and a fourth sensor element 104d comprising an ion-selective membrane 82 Id for selecting Ca ions.

[0252] Where each type of possible constituent ion of the liquid is known (prior to using the apparatus 100), the apparatus 100 may be operable to fully characterise the individual concentrations of each constituent ion. In some applications, full characterisation of the ion composition of the liquid is not required and the concentration of only certain ions are of interest. For such applications, the apparatus 100 may comprise one or more sensor elements 104 comprising ion-selective membranes for selecting a subset of one or more ions present in the liquid.

[0253] Ion-selective membranes represent one way of circumventing the problem of equivalent solutions. Another way of disambiguating equivalent TPF results is to measure how their electric properties vary in response to an external stimulus. As will be described with reference to FIG. 9 below, the applicant has discovered that the relationship between TPF and temperature is ion dependent. As such, data characterising this temperature-TPF relationship may be used to circumvent the problem of equivalent solutions.

[0254] FIG. 9 shows a modified apparatus 900 for detecting a characteristic of a target species within a liquid. The modified apparatus 900 is substantially similar to the apparatus 100 differing in that the modified apparatus 900 comprises a heating module 902. The modified apparatus 900 may comprise any combination of features associated with respect to the apparatus 100. For clarity, previously introduced reference numerals have been used in FIG. 9 to indicate corresponding components to the components of the apparatus 100.

[0255] As shown in FIG. 9, the modified apparatus 900 comprises a sensor 102 including one or more sensor elements 104. Each sensor element 104 includes a capacitive element 106 for contacting the liquid. Each sensor element 104 includes a channel 112 for receiving the liquid.

[0256] The sensor 102 comprises a substrate 114 on which the capacitive element 106 lies. The sensor 102 further comprises a cover layer 116 which at least partially covers the substrate 114. The capacitive element 106 is disposed between the cover layer 116 and the substrate 114.

[0257] The cover layer 116 comprises: a basal surface 211 which faces the substrate 114; and a top surface 212 opposite the basal surface 211. Channels 112 of the one or more sensor elements 104 are formed in the basal surface 211 of the cover layer 116.

[0258] An inlet 118 is provided at a first side 122 of the or each channel 112 for receiving the liquid. In this example, the inlet 118 comprises a first hole which perforates the cover layer 116 fluidically connecting the top surface 212 to the first side 122 of the channel 112. In this example, the channel 112 shown further includes an outlet 120 at a second side 124 of the channel 112 which comprises a second hole which perforates the cover layer 116 fluidically connecting the top surface 212 to the second side 124 of the channel 112.

[0259] The heating module 902 comprises a heating element 904 configured to increase and/or decrease a temperature of the channel 112 within a range of measurement temperatures. In this example, the heating element 904 comprises a Peltier element operable both to heat and to cool the channel 112. The heating element 904 is thermally coupled to the channel 112 of the sensor element 104. In some examples, the heating element 904 may be directly coupled to the substrate 114 opposite the cover layer 116. In this example, the sensor 102 comprises thermally conductive tape 906 which connects the heating element 904 to the substrate 114. [0260] The heating module 902 further comprises a heat sink 908 thermally coupled to the heating element 904 opposite the substrate 114. Thermally conductive tape 906 connects the heating element 904 to the heat sink 908 to ensure effective heat transfer therebetween.

[0261] In this example, the heating module 902 further comprises a temperature sensor 910. The temperature sensor 910 may be used to approximate a temperature of the liquid in the channel 112. As shown in FIG. 9, the temperature sensor 910 may be located on the substrate 114 proximal to the channel 112.

[0262] The temperature sensor 910 may comprise an impedance sensor. In this example, the temperature sensor comprised a PtlOO Resistance Temperature Detector.

[0263] The processor 110 (not shown) may be configured to control the temperature of the heating element 904 to vary the temperature of the channel 112 of the sensor element 104. For example, the processor 110 may be configured to increase and/or decrease the temperature of the channel 112 within a range of measurement temperatures. For example, the range of measurement temperatures may be 10°C to 30°C, 10°C to 20°C, 20°C to 30°C, or 15°C to 25°C. The processor 110, may be configured to incrementally increase or decrease the temperature of the channel 112 from one end point of the range of measurement temperatures to the other end point of the range of measurement temperatures, e.g., in steps of 1°C, 2°C or 5°C. Measurements of the impedance or capacitance of the capacitive element 106 may be recorded at each temperature step.

[0264] In an alternate embodiment, the heating element 904 may be configured to continuously increase or decrease the temperature of the channel 112 from one end point of the range of measurement temperatures to the other end point of the range of measurement temperatures. Measurements of the impedance or capacitance of the capacitive element 106 may be recorded at set time intervals (e.g., every 20s, 10s, 5s).

[0265] The processor 110 (not shown) may be configured to receive temperature data indicative of a temperature of the channel 112 of the sensor element 104. The sample identifiers of the characteristic data may comprise the temperature data. I.e., each entry of characteristic data may be associated with temperature data representing the temperature of the channel 112 at the time at which the measurements used to calculate that characteristic data were recorded. [0266] The processor 110 may be configured to determine the TPF of the capacitance of the capacitive element 106 across the range of measurement temperatures.

[0267] In the example modified apparatus 900 shown in FIG. 9, only one sensor element 104 is shown. Where the sensor 102 comprises multiple sensor elements 104 each sensor element 104 may comprise an individual heating element 904. Each heating element 904 may be thermally coupled to the channel 112 of a different sensor element 104. The processor 110 may be configured to control each heating element independently.

[0268] FIG. 10A shows a seventh graph 1000 illustrating a plot of the TPF against the temperature data from the temperature sensor 910 for various 2mM ionic solutions. FIG. 20A shows an updated version of the graph of FIG. 10A after further data was collected. As shown in FIG. 10A and FIG. 20A, the TPF is dependent both on the recorded temperature and on the type of electrolyte. Reference numerals for each line of the seventh graph 1000, representing the TPF- recorded temperature profiles for four different 2mM ionic solutions are provided in Table 8 below.

Table 8: Reference numerals for FIG. 10A, FIG. 20A and FIG. HA.

[0269] In some embodiments, the TPF-recorded temperature relationship may be used to disambiguate equivalent TPF results to determine the concentration and identity of the target species for single electrolyte solutions.

[0270] For example, certain electrolytes exhibit characteristic profiles. As shown in FIG. 10A, the gradient for 2mM CaCh increased with temperature above 25°C. Such a profile may be used to characterise specific electrolyte solutions by comparing the recorded TPF-recorded temperature profile to comparable plots of known single electrolyte solutions.

[0271] FIG. 10B shows a bar chart 1100 representing gradients of the TPF-recorded temperature plots for 0.5 to 2 mM solutions (left to right in FIG. 10B) of each of NaCl, KC1, MgCh and CaCh. FIG. 20B shows an updated version of the graph of FIG. 10B after further data was collected.

[0272] As shown, the gradients of the TPF -recorded temperature plots are dependent on the species of the electrolyte. The processor 110 may be configured to calculate a gradient of the turning point data as a function of the recorded temperature for the liquid. The calculated gradient of the liquid may be used in combination with the TPF (e.g., at room temperature: 20°C) to determine the concentration and the type of ion present in the liquid (i.e., by comparing the calculated gradient to comparable gradients of a plurality of species [stored in the memory module 128]). Specifically, the calculated gradient of the liquid may be used in combination with the TPF (e.g., at room temperature: 20°C) to determine the concentration and the type of ion present in the liquid. In this way, the problem of equivalent solutions may be circumvented for single electrolyte solutions where the type and concentration of the electrolyte is unknown. Additionally or alternatively, the gradients of the TPF-recorded temperature plots may be used in combination with the TPF results to determine the concentration or identity of the target species for single electrolyte solutions. For example, gradients of the TPF-recorded temperature plots may improve the confidence in results where TPF values for two electrolyte solutions are similar.

[0273] Characterisation using the heating module 902 may be used in combination with other characterisation methods discussed herein e.g., use of ion-selective membrane(s). In the example modified apparatus 900 shown in FIG. 9, the sensor 102 does not comprise any ion- selective membranes. However, it is to be understood that the heating module 902 may be used in combination with the ion-selective membranes. Use of ion-selective membranes in combination with the heating module 902 may allow further characterisation of the liquid. In particular, where the modified apparatus 900 comprises both the heating module 902 and at least one ion-selective membrane, the modified apparatus 900 may be operable to determine the individual concentrations of different ions within a multiple ion solution such as sweat.

[0274] An alternate way to circumvent the problem of equivalent solutions is to make use of the minima of the imaginary component of the impedance. As shown previously in FIG. 4B, plots of the imaginary component of the impedance against the measurement frequency exhibit a minima. The processor 110 may be configured to determine the minima of the imaginary component. [0275] FIG. 11 A shows an eighth graph 80 plotting an impedance of the minima of the imaginary component of the impedance against the recorded temperature for 2mM solutions of NaCl 1004, KC1 1003, CaCl₂ 1001, and MgCl₂ 1002.

[0276] As shown, the impedance of the minima is dependent both on the recorded temperature and on the type of electrolyte. In some embodiments, the minima may be used to disambiguate equivalent TPF results to determine the concentration and identity of the target species for single electrolyte solutions. For example, the processor 110 may be configured to compare the impedance of the minima of the imaginary component measured by the impedance spectroscopy circuit 108 to known minima impedance values of a plurality of species. Specifically, the calculated gradient of the liquid may be used in combination with the TPF (e.g., at room temperature: 20°C) to determine the concentration and the type of electrolyte present in the liquid.

[0277] FIG. 1 IB shows a second bar chart 1102 representing gradients of the impedance of the minima of the imaginary component against the recorded temperature for 0.5 to 2 mM solutions (left to right in FIG. 11B) of each of NaCl, KC1, MgCl₂, and CaCl₂.

[0278] As shown in the second bar chart 1102, the gradients of the impedance of the minima against the recorded temperature (termed the minima gradients) are dependent on the species of the electrolyte. The processor 110 may be configured to calculate the minima gradient for the liquid in the channel 112 of the sensor element 104. The calculated minima gradient of the liquid may be used in combination with the TPF (e.g., at room temperature: 20°C) to determine the concentration and the type of ion present in the liquid (i.e., by comparing the calculated minima gradient minima gradients of a plurality of species stored in the memory module 128). Specifically, the calculated minima gradient of the liquid may be used in combination with the TPF (e.g., at room temperature: 20°C) to determine the concentration and the type of ion present in the liquid. In this way, the problem of equivalent solutions may be circumvented for single electrolyte solutions where the type and concentration of the electrolyte is unknown. Additionally or alternatively, the calculated minima gradient may be used in combination with the TPF results to determine the concentration or identity of the target species for single electrolyte solutions. For example, the calculated minima gradient may improve the confidence in results where TPF values for two electrolyte solutions are similar. [0279] Characterisation of the liquid on the basis of the minima gradient may be used in combination with any of the other characterisation methods discussed herein (e.g., ion-selective membrane(s) and/or gradients of the TPF-recorded temperature plots and/or the impedance of the minima at room temperature (20°C)).

[0280] Another way of circumventing the problem of equivalent solutions is to modify the capacitive element 106 to be selective for the target species to be characterised. In particular, a molecularly imprinted polymer may be used to selectively bind with the target species. Molecularly imprinted polymers (MIPs) are synthetic polymers with an affinity for a given analyte, or group of structurally related compounds. MIPs achieve a high specificity for analytes through functional group cavities which are complementarily shaped to a specific analyte. The applicant has appreciated that a MIP with functional group cavities complementarily shaped to the target species may be used to selectively adhere the target species to the capacitive element 106. This allows the apparatus 100 to additionally or alternatively characterise non-ionic analytes such as metabolites. For example, the processor 110 may be configured to compare the turning point data to stored data representing turning points of the capacitance of the capacitive element as a function of measurement frequency for a plurality of known samples to determine the concentration of the target species in the liquid. Alternately, the processor 110 may be configured to determine the concentration of the target species by comparing the recorded capacitance-measurement frequency relationship to comparable curves for samples of known concentration of the target analyte.

[0281] MIPs may also be used to circumvent the equivalent solutions problem for solutions of unknown analyte type and unknown concentration as this significantly reduces the uncertainty of the type of ion producing the observed capacitance-measurement frequency relationship. As such, the features of the turning point frequency can be attributed to the target species for which the MIP is functionalised and the concentration of the target species in the liquid can be unambiguously determined.

[0282] In some examples, the capacitive element 106 may be modified to incorporate a MIP. Example methodology for determining the identity of the target species in the liquid, using the modified capacitive element 1202 is described below in relation to FIG. 14.

[0283] FIG. 12 shows a schematic exploded view of a modified capacitive element 1202. The modified capacitive element 1202 is substantially similar to the capacitive element 106 differing from the capacitive element 106 in that it comprises a MIP layer 1204. The MIP layer 1204 may be deposited on the first electrode 201 and/or the second electrode 202. In particular, the MIP layer 1204 may cover a surface of the first electrode 201 and/or the second electrode 202 such that the MIP layer 1204 contacts the liquid in the channel 112.

[0284] In this example, the MIP layer 1204 is deposited on the first electrode 201 only. Depositing the MIP layer 1204 on only one of the two interdigitated electrodes may minimise manufacturing time and costs associated with MIP deposition. The MIP can be deposited directly onto the electrodes using one of many suitable techniques such as drop-casting, aerosol jet printing or spin coating. The area onto which the MIP is deposited may be selected through varying the materials used in the first electrode 201 and the second electrode 202. For example, the first electrode 201 and the second electrode 202 may be printed from carbon ink and a separate material may then be deposited over a selected area of the first electrode 201 /second electrode 202. For example, silver may be printed onto a selected area. In this example a silver layer 1206 was printed by AJP onto the first electrode 201 and not the second electrode 202. The silver ink was produced by diluting Novacentrix Ag nanoparticles aerosol ink (JS- A221AE) 1:2 by volume with de-ionised water. After printing, the silver layer 1206 was cured at 150 °C for 2 hours to increase its conductivity and stability. The MIP used was synthesized by the functional monomer 3- -aminophenylboronic acid (3 -APB A). Alternately, the MIP may be synthesized by the functional monomer 4- -aminophenylboronic acid (4-ABPA), or any other suitable polymer, for example polyaniline or polypyrrole.

[0285] As shown in FIG. 12, the selected area follows the profile of the first electrode 201. In this example, the silver layer 1206 and overlying MIP layer 1204 partially covers the first electrode 201 such that some of the surface of the first electrode 201 is exposed to the liquid in the channel 112 in use. In other examples, the silver layer 1206 and overlying MIP layer 1204 completely cover the first electrode 201 such that, in use, none of the surface of the first electrode 201 is exposed to the liquid in the channel 112.

[0286] In this example, the MIP layer 1204 was functionalised as a lactate MIP. The lactate molecularly imprinted polymer (MIP) was lactate imprinted poly(3-APBA) synthesised by the functional monomer 3 -aminophenylboronic acid (3 -APB A) with template molecule of lactate in neutral PBS solution. The lactate MIP solution is composed of 0.012 g of lactate, 0.0363 g 3- APBA and 0.1095 g sodium chloride dissolved in 10 mL, 10 mM PBS solution (pH = 7.4). A method of manufacturing the modified capacitive element 1202 having the lactate MIP layer

1204 is illustrated in FIG. 13 below.

[0287] In other examples, the MIP layer 1204 may be functionalised for a different analyte such as a different metabolite (e.g., lactose, glucose, glycerol, pyruvate, urea, ammonia, cortisol, tyrosine, or serine).

[0288] Where the sensor 102 comprises multiple sensor elements 104, each sensor element 104 may be provided with a different MIP layer 1204. I.e., each of the different MIP layers 1204 being configured to select for a different analyte.

[0289] The sensor 102 may comprise a reservoir (not shown) for receiving the liquid. The reservoir may be fluidically coupled to the inlet 118 of each channel 112. Where the sensor 102 comprises multiple sensor elements 104, the channel 112 of each sensor element 104 may be fluidically connected to the reservoir so that liquid may be supplied from the reservoir to the channels 112 through the inlets 118. The turning point frequency(s) recorded for a given sensor element 104 can be directly attributed to the presence of an individual analyte.

[0290] Where multiple sensor elements 104 are present, the sample identifier of each sensor element 104 may include identification data representative the type of MIP layer 1204 comprised in that sensor element 104. The processor 110 may be configured to analyse the characteristic data received from each sensor element 104 in combination with the identification data from each sensor element 104 to determine the concentration and the type of target species present in the liquid.

[0291] Use of MIP layers 1204 may be used to circumvent equivalent TPF solutions for single analyte solutions allowing the concentration and type of such single analyte solutions to be correctly identified. However, MIP layers 1204 may further enable the apparatus 100 to determine the individual concentrations of different analytes within a multiple analyte solution such as sweat. The processor 110 may be configured to analyse the characteristic data received from each sensor element 104 in combination with the identification data from each sensor element 104 to determine the concentration and the type of each (of multiple) target species present in the liquid.

[0292] The number of sensor elements 104 and type of MIP layers 1204 used therein may be selected depending on the specific application for which the apparatus 100 is to be used. Where little or nothing is known about the constituent ions of the liquid, the sensor element 104 may comprise a large array of sensor elements 104. For certain applications, the likely types of analytes present in the liquid are known prior to characterisation using the apparatus 100. For example, sweat is known to comprise lactate, lactose, glucose, glycerol, pyruvate, and serine in varying proportions. Where the apparatus 100 is to be used to measure sweat, the sensor 102 may comprise one sensor element 104 comprising a MIP layer 1204 for selecting lactate. The sensor 102 may further comprise a further sensor element 104 comprising a MIP layer 1204 for selecting glucose and/or a further sensor element 104 comprising a MIP layer 1204 for selecting serine. Such a sensor 102 may comprise a further capacitive element 104 comprising a MIP layer 1204 for selecting glycerol and/or a further capacitive element 104 comprising a MIP layer 1204 for selecting pyruvate ions.

[0293] Where each type of possible constituent analyte of the liquid is known (prior to using the apparatus 100), the apparatus 100 may be operable to fully characterise the individual concentrations of each constituent analyte. In some applications, full characterisation of the analyte composition of the liquid is not required and the concentration of only certain analytes are of interest. For such applications, the apparatus 100 may comprise one or more sensor elements 104 comprising MIP layers 1204 for selecting a subset of one or more analytes present in the liquid.

[0294] FIG. 13 illustrates a method of manufacturing the modified capacitive element 1202 comprising a lactate MIP layer 1204. First the first electrode 201 and the second electrode 202 may be printed on the substrate 114. In this example, the first electrode 201 and the second electrode 202 were printed using carbon ink comprising lwt% of multi-walled carbon nanotubes (MWCNT, Sigma- Aldrich) was blended with the 0.5wt% of Polyvinylpyrrolidone (PVP, Sigma-Aldrich) and de-ionised water to produce the carbon ink. The suspension was magnetically stirred for 15 minutes and ultrasonicated for 1 hour prior to atomisation and printing using Aerosol Jet Printing. The parameters of which are provided in Table 9 below. The silver layer 1206 was then printed on top of the first electrode 201.

Table 9: AJP parameters. [0295] The printed electrodes were then cured at 200 °C for 15 minutes. In this example, the substrate 114 comprised a Kapton sheet. Following post- treatment, the first electrode 201 and the second electrode 202 were connected to a potentiostat (VersaSTAT4, Ametek Scientific Instruments, USA) and operated as working electrodes. An Ag/AgCl electrode filled with potassium chloride (KC1) saturated with silver chloride (AgCl) solution was used as the reference electrode, while Pt were utilized as the counter electrode. The lactate MIP solution was added into a cell in contact with the three electrode system. Polymerization was carried out by cyclic voltammetry (CV) scan at a scan rate of 0.05V/s in the potential range of -0.6 V and 0.6 V for 30 cycles. The lactate imprinted poly(3-APBA) was deposited onto the silver layer 1206. The lactate template was subsequently removed from the MIP layer 1204 by immersion in the 8wt% acetic acid in PBS solution, resulting in lactate complementary cavities. The resultant lactate MIP layer 1204 are operable to rebind with lactate due to complementary characteristics in functional groups, thereby achieving lactate selectivity. It will be understood that the above-described method an example method of manufacturing the modified capacitive element 1202 and is not intended as limiting. For example, other acids such as aqueous HC1 may be used instead of acetic acid to remove the lactate template from the MIP layer. The MIP- layer may be functionalised for a different metabolite. For example the MIP layer 1204 may comprise a lactose MIP, glucose MIP, urea MIP, ammonia MIP, cortisol MIP, glycerol MIP, pyruvate MIP, or serine MIP. The resultant MIP layer 1204 may be operable to rebind with the metabolite due to complementary characteristics in size, shape and/or functional groups, thereby achieving selectivity for that metabolite.

[0296] FIG. 14 shows results taken using the apparatus 100 with the modified capacitive element 1202 of FIG. 12. The left-hand side of FIG. 14 shows a first series of graphs 1300 showing results from when a lOmM lactate solution was tested as the liquid. The right-hand side of FIG. 14 shows a second series of graphs 1310 showing results from when a lOmM glucose solution was tested as the liquid. For the second series of graphs 1310, the MIP layer 1204 of the modified capacitive element 1202 was a glucose MIP layer 1204 instead of the lactate MIP layer 1204 shown in FIG. 12.

[0297] From top to bottom of FIG. 14, the first series of graphs 1300 incudes a real signal graph 1301, an imaginary signal graph 1302, a capacitance graph 1303, a first differential graph 1304 and a second differential graph 1305. Similarly, the second series of graphs 1310 incudes a real signal graph 1311, an imaginary signal graph 1312, a capacitance graph 1313, a first differential graph 1314 and a second differential graph 1315. As shown in FIG. 14, each of the first series of graphs 1300 and the second series of graphs 1310 are plotted against measurement frequency. The real signal graphs 1301, 1311 illustrate the real component of the impedance measurements recorded by the impedance spectroscopy circuit 108. The imaginary signal graphs 1302, 1312 illustrate the imaginary component of the impedance measurements recorded by the impedance spectroscopy circuit 108. The capacitance graphs 1303, 1313 illustrate plots of the effective capacitance calculated using Equation 1. The first differential graphs 1304, 1314 show the first differential of the capacitance graphs 1303, 1313. The second differential graphs 1305, 1315 show the second differential of the capacitance graphs 1303, 1313.

[0298] The plot of calculated capacitance against measurement frequency for the lactate solution (capacitance graph 1303) has four turning points. This is shown more clearly in the second differential graph 1305 where the turning points of the capacitance-frequency plot are illustrated by intersections with line 1306 which represents dC/df=0. In contrast, the plot of calculated capacitance against measurement frequency for the glucose solution (capacitance graph 1313) has two turning points. This is shown more clearly in the second differential graph 1315 where the turning points of the capacitance-frequency plot are illustrated by intersections with line 1316 which represents d²C/df²=0.

[0299] As shown by comparison of the second differential graphs 1305, 1315 for the lactate and glucose solutions, the number of turning points and the TPF of the turning points is dependent on the identity of the metabolite in the liquid. As such, the presence of specific metabolites in the liquid may be unambiguously determined through comparison of the recorded TPF and/or number of turning points with comparable data from known samples. As the TPF is dependent on the concentration of the target species within the liquid, the concentration of the target species in the liquid may be determined by comparison of the turning point data recorded with comparable data from various samples of known concentrations of the target species.

[0300] Additionally, where the sensor 102 comprises multiple sensor elements 104 having different MIP layers 1204, the apparatus 100 may determine the individual concentrations of different analytes within a multiple analyte solution such as sweat. The processor 110 may be configured to analyse the characteristic data received from each sensor element 104 in combination with the identification data from each sensor element 104 to determine the concentration and the type of each (of multiple) target species present in the liquid.

[0301] Characterisation of the liquid using MIP layers 1204 may be used in combination with any of the other characterisation methods discussed herein (e.g., ion-selective membrane(s) and/or gradients of the TPF-recorded temperature plots and/or the impedance of the minima at room temperature (20°C) and/or the minima gradient(s)). As such, the apparatus 100 maybe operable to determine the identity and/or concentration of both ion(s) and metabolite(s) within the liquid.

[0302] FIG. 15 illustrates a method 1500 of detecting a characteristic of a target species in a liquid such as sweat. The target species may comprise an electrolyte, such as NaCl, KC1, MgCh, CaCh, lactate, or a metabolite, such as lactose, lactate, glucose, urea, ammonia, cortisol, glycerol, pyruvate, tyrosine, or serine.

[0303] The method 1500 may comprise a first step 1501 of providing a capacitive element in contact with the liquid.

[0304] The method 1500 includes a second step 1502 of determining a capacitance of the capacitive element across a range of measurement frequencies.

[0305] In some examples, the second step 1502 may include measuring an impedance of the capacitive element across the range of measurement frequencies to obtain impedance measurements (e.g., using an impedance spectroscopy circuit). The impedance measurements may include a real component and an imaginary component. The second step 1502 may include calculating the capacitance of the capacitive element, across the range of measurement frequencies, from the impedance measurements. For example, the second step 1502 may involve calculating an effective capacitance (C ) using a simplified circuit model consisting of a resistor and a capacitor in series. The effective capacitance (C ) of the circuit model may be calculated from the imaginary component of the impedance signal (Im(Z) , using the formula:

[0306] Where angular frequency co = 2nf, where f is the measurement frequency.

[0307] The capacitance data as a function of frequency may be smoothed to reduce noise, for example using a rolling average or another suitable method. Logarithms of the capacitance and frequency data may be taken. This data may then be differentiated twice, and the turning point may be defined as the point at which the second derivative crosses 0. Linear interpolation between the two closest points to 0 may be used to determine the turning point to a greater accuracy. The turning point frequency may be determined as the measurement frequency corresponding to this point. Further smoothing (for example, using a rolling average) may be applied to the first and/or second derivative to reduce noise. This procedure may be carried out for each set of capacitance-frequency data obtained.

[0308] Prior to taking measurements, the impedance spectroscopy circuit may be calibrated against the effect of parasitic capacitances using an open circuit, closed circuit and known load (e.g., 330 resistor).

[0309] The method 1500 includes a third step 1503 of determining turning point data representative of a turning point of the capacitance as a function of the measurement frequency. For example, the turning point data may represent the measurement frequency at the turning point. In other examples, the turning point data may represent the capacitance, of the capacitive element, at the turning point.

[0310] The method 1500 includes a fourth step 1504 of determining a characteristic of the target species from the turning point data. For example, the characteristic may represent: the concentration of the target species in the liquid; and/or the identity of the target species (e.g., in a liquid of known concentration).

[0311] The fourth step 1504 may involve comparing the turning point data to stored data representing turning points of the capacitance of the capacitive element as a function of measurement frequency for a plurality of known samples.

[0312] In some examples, the method 1500 may involve varying a temperature of the capacitive element within a range of measurement temperatures. In such examples, the second step 1502 may involve determining the capacitance of the capacitive element as a function of measurement frequency across the range of measurement temperatures. The third step 1503 may involve determining turning point data representative of turning points of the capacitance as a function of the measurement frequency across the range of measurement temperatures. The method 1500 may include an optional fifth step 1505 of analysing the turning point data for each of the plurality of temperature measurements to determine an additional characteristic of the target species. The additional characteristic may represent: the concentration of the target species in the liquid; and/or the identity of the target species (e.g., in a liquid of known concentration). For example, the fifth step 1505 may involve calculating a gradient of the turning point data as a function of the measurement temperature and comparing the calculated gradient to stored gradients for a plurality of species to determine an identity or a concentration of the target species in the liquid.

[0313] Where the second step 1502 involves measuring an impedance of the capacitive element across the range of measurement frequencies to obtain impedance measurements having a real component and an imaginary component, the method may include a sixth step 1506. The sixth step 1506 may involve determining a minimum of the imaginary component of the impedance measurements of the capacitive element across the measurement frequencies and determining an additional characteristic of the target species by comparing the minimum of the imaginary component to stored minimum imaginary components for one or more species to determine an identity or a concentration of the target species in the liquid. The additional characteristic may represent: the concentration of the target species in the liquid; and/or the identity of the target species (e.g., in a liquid of known concentration). For example, the sixth step 1506 may involve calculating a gradient of the minimum of the imaginary component as a function of the measurement temperature and comparing the calculated gradient to stored gradients for a plurality of species to determine the identity or the concentration of the target species in the liquid.

[0314] Where the method 1500 includes the fifth step 1505 and/or the sixth step 1506, the method may further include an optional seventh step 1508 of comparing the characteristic with one or more additional characteristics to determine concentration and identity of the species of the target species in the liquid.

[0315] The steps of the method 1500 may be completed in any suitable order. FIG. 15 shows one example of order in which the method 1500 may be carried out. The steps of the method may be completed in ascending numerical order. The fifth step 1505 (where present) may be completed before, after or simultaneously with the sixth step 1506 of the method 1500. The fourth step 1504 may be completed before, after or simultaneously to the fifth step 1505 (where present) and/or the sixth step 1506 (where present). The method may be completed using an apparatus according to this disclosure (e.g., apparatus 100 or modified apparatus 900). Where the apparatus comprises one or more additional capacitive elements and/or one or more further capacitive elements. The method may include the step of analysing characteristic data received from each capacitive element in combination with identification data from each capacitive element to determine the concentration and the type of target species present in the liquid.

[0316] FIG. 16 shows a plot 1700 of the turning point frequency against ion concentration for various solutions comprising different concentrations of NaCl and KC1. Each of the capacitive elements 106 tested in FIG. 16 was substantially similar to the capacitive element 106 shown in FIG. 2. Test solutions of the aqueous ionic chloride solutions (NaCl, KC1) used herein were produced according to the following procedure. NaCl and/or KC1 (Merck, Germany) was added to de-ionised water (filtered in-house, Purite, UK, typical conductivity < 10 pS) to produce the required concentrations.

[0317] As shown, there is a linear relationship between the turning point frequency and the relative quantities of NaCl and KC1. This is evidence in support of the total calculated turning point frequency being equal to the sum of the individual contributions of each of the two target species.

[0318] As such, where the concentration or turning point frequency contribution of all but one ionic species in a complex solution are known (e.g., by use of capacitive elements with ion selective membranes) and the total turning point frequency of the complex solution is known (e.g., by use of an uncovered capacitive element), the turning point frequency and therefore the concentration of an additional ionic species in the complex solution can be determined. I.e., By subtracting the sum of the contributions of each of the known ionic species in the liquid to the turning point data from the total turning point data for the solution, turning point data representative of the individual contribution of an additional ionic species can be calculated. In this way, the concentrations of N ionic species may be measuring using N-l sensor elements with ion-selective membranes. This may beneficially reduce the cost of the measuring apparatus as few ion-selective membranes need to be manufactured.

[0319] This method of calculation is illustrated in FIG. 21 which shows a plot 901 of the turning point frequency against concentration for Ca, K and Na ions. In the figure, lines of best fit for data for Ca, K and Na solutions are represented by lines 911, 912 and 913 respectively. Here the apparatus comprises a sensor element with a K-specific ion-selective membrane, a sensor element with a Ca-specific ion-selective membrane and a sensor element with no ion- selective membrane. Calibration data (i.e., data indicative of the relationship between the concentration of each ion and the contribution of each ion to the turning point frequency) is known for Ca, K and Na. As shown, the turning point frequency of the capacitive element with the Ca-specific ISM (ion-selective membrane) is measured (TPFca, 914). The turning point frequency of the capacitive element with the K-specific ISM is measured (TPFK, 915). The turning point frequency of the capacitive element with no ISM is measured (TPFiotai, 916) As shown in FIG. 21, the turning point frequency of the capacitive element with the Ca-specific ISM is measured. In this example, the liquid comprises only three target species, Fiotai = TPFca + TPFK + TPF_Na . So the contribution of Na ions to the turning point frequency (TPF\_a) can be calculated. I.e., Fiotai - (TPFca + TPFK ) = TPF\_a. The concentration of Na can be determined from the calibration data for Na. In this way, the concentration of Na ions in the liquid can be determined without requiring an ion-selective membrane specific for Na.

[0320] FIG. 22A is a plot 1400 of measured impedance (relative to a known fluid) against concentration of Ca²⁺. In this example the complex solution comprised: 2 mM Na, 2 mM K, 1 mM Ca. As shown, the calculated Ca concentration was determined to be 0.76 (± 0.04) mM (i.e., a 24% error value).

[0321] FIG. 22B is a plot 1600 of measured impedance (relative to a known solution) against concentration of Ca²⁺. In this example the complex solution comprised: 2 mM Na, 2 mM K, 1 mM Ca. As shown, the calculated K concentration was determined to be 1.77 (± 0.03) mM. In the examples shown in FIG. 22A and FIG. 22B, the known fluid is de-ionized water. The method of calculating AZ/Zo is described in the paper entitled “Impedance-based sensor for potassium ions” by C. Day et al. Analytica Chimica Acta, Vol 1034 (2018) p.39-45. The calculation of AZ/Zo may be completed according to the following equation:

AZ _ Zj — Z_o ZQ Z₀

Claims

1. A method of detecting a characteristic of a target species in a liquid, comprising the steps of: providing a capacitive element in contact with the liquid, determining a capacitance of the capacitive element across a range of measurement frequencies, determining turning point data representative of a turning point of the capacitance as a function of the measurement frequency, and determining a characteristic of the target species from the turning point data.

2. The method of claim 1, wherein the turning point data comprises the measurement frequency at the turning point.

3. The method of claim 1 or 2, wherein the turning point data comprises a number of turning points within the range of measurement frequencies.

4. The method of claim 2, wherein the step of determining a characteristic of the target species from the turning point data includes comparing the turning point data to stored data representing turning points of the capacitance of the capacitive element as a function of measurement frequency for a plurality of known samples.

5. The method of any one of claims 1 to 4, wherein determining the characteristic of the target species involves detecting a concentration of the target species in the liquid from the turning point data.

6. The method of any one of claims 1 to 4, wherein determining the characteristic of the target species involves identifying the target species in the liquid from a set of known target species using the turning point data.

7. The method of any one of claims 1 to 6, further including the step of varying a temperature of the capacitive element within a range of measurement temperatures and determining the capacitance of the capacitive element as a function of measurement frequency across the range of measurement temperatures.

8. The method of claim 7, further including the steps of: determining turning point data representative of turning points of the capacitance as a function of the measurement frequency across the range of measurement temperatures; and analysing the turning point data for each of the plurality of temperature measurements to determine an additional characteristic of the target species.

9. The method of claim 8, wherein the step of determining the additional characteristic of the target species involves calculating a gradient of the turning point data as a function of the measurement temperature and comparing the calculated gradient to stored gradients for a plurality of species to determine an identity or a concentration of the target species in the liquid.

10. The method of claim 8 or 9, further including the steps of: measuring an impedance of the capacitive element across a range of measurement frequencies to obtain impedance measurements including a real component and an imaginary component; determining a minimum of the imaginary component of the impedance measurements of the capacitive element across the measurement frequencies; and determining an additional characteristic of the target species by comparing the minimum of the imaginary component to stored minimum imaginary components for one or more species to determine an identity or a concentration of the target species in the liquid.

11. The method of claim 10, wherein the step of determining the additional characteristic of the target species involves calculating a gradient of the minimum of the imaginary component as a function of the measurement temperature and comparing the calculated gradient to stored gradients for a plurality of species to determine the identity or the concentration of the target species in the liquid.

12. The method of claim 11, further comprising a step of comparing the characteristic with the additional characteristic(s)to determine concentration and identity of the species of the target species in the liquid.

13. The method of any one of claims 1 to 11, wherein the target species comprises an electrolyte, such as NaCl, KC1, MgCh, CaCh, lactate, or a metabolite, such as lactose, lactate, glucose, urea, ammonia, cortisol, glycerol, pyruvate, tyrosine, or serine.

14. An apparatus for detecting a characteristic of a target species within a liquid, comprising: a sensor, comprising a capacitive element for contacting the liquid and outputting a signal indicative of the capacitance of the capacitive element across a range of measurement frequencies; and a processor configured to: receive the signal indicative of the capacitance, in dependence on the received signal indicative of the capacitance, determine turning point data representative of a turning point of the capacitance as a function of the measurement frequency; and determine a characteristic of the target species from the turning point data.

15. The apparatus of claim 14, wherein the processor is configured to compare the turning point data to stored data representing turning points of the capacitance of the capacitive element as a function of measurement frequency for a plurality of known samples to determine an identity or a concentration of the target species in the liquid.

16. The apparatus of claim 14 or 15, wherein the sensor comprises a channel for receiving the liquid, the channel overlaying the capacitive element, optionally wherein the channel may comprise a microfluidic channel.

17. The apparatus of any one of claims 14 to 16 , wherein the sensor comprises an ion-selective membrane.

18. The apparatus of any one of claims 14 to 17, wherein the sensor comprises one or more additional capacitive elements each comprising an ion-selective membrane, optionally wherein each ion-selective membrane is configured to select for a different ion.

19. The apparatus of any one of claims 14 to 18, wherein the capacitive element comprises a molecularly imprinted polymer, optionally wherein the capacitive element comprises a molecularly imprinted polymer layer functionalised for a metabolite.

20. The apparatus of any one of claims 14 to 19, wherein the sensor comprises one or more further capacitive elements each comprising a molecularly imprinted polymer layer, optionally wherein each molecularly imprinted polymer layer is functionalised for a different species.

21. The apparatus of any one of claims 16 to 20, comprising a heating element configured to increase and/or decrease a temperature of the channel within a range of measurement temperatures, wherein the apparatus is configured to determine the capacitance of the capacitive element as a function of measurement frequency across the range of measurement temperatures and the processor is configured to determine turning point data representative of turning points of the capacitance as a function of the measurement frequency across the range of measurement temperatures; and analyse the turning point data for each of the plurality of temperature measurements to determine an additional characteristic of the target species.

22. The apparatus of any one of claims 14 to 21, wherein the sensor comprises an impedance spectroscopy circuit for measuring an impedance of the capacitive element across a range of measurement frequencies and wherein the processor is configured to calculate the capacitance of the capacitive element across a range of measurement frequencies from the impedance of the capacitive element across a range of measurement frequencies.

23. The apparatus of claim 22, wherein the impedance spectroscopy circuit is configured to measure a real component and an imaginary component of the impedance of the capacitive element across a range of measurement frequencies and wherein the processor is configured to: determine a minimum of the imaginary component; and compare the minimum of the imaginary component to stored minimum imaginary components for one or more species to determine an additional characteristic of the target species.

24. The apparatus of claim 22, wherein the apparatus comprises a wearable device which comprises the sensor and, optionally, the processor.

25. Method of detecting a characteristic of a target species within a liquid using a wearable device, comprising: providing a wearable device comprising a capacitive element for contacting the liquid and a sensor for measuring a capacitance or an impedance of the capacitive element across a range of measurement frequencies, determining a capacitance of the capacitive element across a range of measurement frequencies using the sensor, and determining a characteristic of the target species from the capacitance of the capacitive element as a function of measurement frequency.

26. An apparatus for detecting a characteristic of a target species within a liquid, comprising: a wearable device comprising a sensor including a capacitive element for contacting the liquid, wherein the sensor is configured to output a signal indicative of a capacitance of the capacitive element across a range of measurement frequencies; and a processor configured to receive the signal indicative of the capacitance and determine a characteristic of the target species from the capacitance of the capacitive element as a function of measurement frequency.

27. The apparatus of claim 26, wherein the processor forms part of the wearable device.