WO2015109410A1

WO2015109410A1 - Chemical sensor

Info

Publication number: WO2015109410A1
Application number: PCT/CA2015/050049
Authority: WO
Inventors: Prashanthi Kovur; Arindam Phani; Thomas Thundat
Original assignee: University of Alberta
Current assignee: University of Alberta
Priority date: 2014-01-23
Filing date: 2015-01-23
Publication date: 2015-07-30
Anticipated expiration: 2016-07-23

Abstract

A nanowire is exposed to a chemical molecules of which are adsorbed on the nanowire. The nanowire is exposed to infrared light and variations of impedance parameters of the nanowire are measured to detect the chemical. A dissipation factor of the nanowire forms a spectrum with respect to the wavenumber of the infrared light, the spectrum being characteristic to the adsorbed chemical allowing identification of the chemical based on the spectrum.

Description

CHEMICAL SENSOR

TECHNICAL FIELD

[0001] Infrared chemical sensors.

BACKGROUND

[0002] Infrared (IR) spectroscopy is a prolific technique that allows chemical discrimination with high precision. Traditional Fourier transform infrared spectroscopy (FTIR) with resolution limited to a few micrograms ug) is an entirely optical method. Its extension, coupling nano-mechanical response to IR absorption, has also been introduced in recent years with tens of pictograms (pg) sensitivity.

[0003] Chemical identification of exceedingly small number of adsorbate molecules with high selectivity and sensitivity has applications in many areas ranging from health care to national security. Immense efforts are presently underway in developing miniature chemical sensors that can be coupled with smart phones for all kinds of practical applications. Nanosystems such as nanowires and nanomechanical devices have been envisioned as a sensor platform for the next generation of highly sensitive chemical and biological sensors. These nanosystem platforms utilize molecular adsorption- induced changes in the physical properties (thermal, optical, electrical and/or mechanical) for detection. Most of the approaches of imparting selectivity in nanowire-based sensors have been extensions of similar concepts used in macro and micro sensor platforms which utilize immobilized receptors or chemically selective coatings. As a result, selectivity in nanosystems still remains a challenge just as in the case of microsystems. Therefore, developing concepts for chemical selectivity that do not depend on immobilized chemically selective interfaces are of interest. For example, response modulation in receptor free nanomechanical systems (microcantilevers) as a function of IR absorption by adsorbed species has been demonstrated earlier. A detailed analysis of the dynamic thermo-mechanical response elucidates the dependence of thermal mass of such a micromechanical system in coupling of IR initiated thermal phonons to acoustic phonons and also its subsequent time scale response. An extension using the analogous concept in a mechanical resonant nanowire platform can be envisioned for achieving magnified effects at a lower time scale, subject to proper conditions of high vacuum (10^_5 to lO^-7 Torr) in order to eliminate damping in ambient conditions. However, tracking mechanical motion of a nanoscale device is still a challenge.

SUMMARY

[0004] There is provided a method of detection of a chemical, comprising: adsorbing molecules of the chemical on the surface of a nanowire, irradiating the nanowire with infrared radiation, measuring variations in impedance parameters of the nanowire, and detecting the chemical based on the variations in impedance parameters. In various embodiments, there may be included any one or more of the following features: the step of measuring variations in impedance parameters of the nanowire may include measuring dissipation; a dissipation infrared spectrum of the adsorbed chemical may be obtained by determining dissipation of the nanowire at multiple wavelengths of infrared light; and the adsorbed chemical may be identified by comparing all or part of the obtained dissipation infrared spectrum to a library of dissipation infrared spectra.

[0005] There is further provided a chemical sensor comprising a nanowire having a surface exposed to adsorption of molecules of a chemical, a source of infrared radiation positioned to illuminate at least a part of the exposed surface of the nanowire with infrared radiation, and an impedance analyzer connected to measure impedance parameters of the nanowire.

[0006] These and other aspects of the device and method are set out in the claims, which are incorporated here by reference.

BRIEF DESCRIPTION OF THE FIGURES

[0007] Embodiments will now be described with reference to the figures, in which like reference characters denote like elements, by way of example, and in which:

[0008] Fig. 1 is a schematic representation of an embodiment of a sensor set up;

[0009] Fig. 2 is a schematic representation nanowire resonator (NWR) with equivalent electrical circuit model;

[0010] Fig. 3 is a schematic representation of a sensor set up showing the infrared source;

[0011] Fig. 4 is a graph of D-factor of a nanowire v. frequency with (lighter line) or without

(darker line) adsorbed RDX molecules;

[0012] Fig. 5 is a graph of D-factor of a nanowire v. frequency with (lighter line) or without

(darker line) adsorbed TNT molecules;

[0013] Fig. 6 is a graph of D-factor IR spectrum v. wavenumber with adsorbed RDX;

[0014] Fig. 7 is a graph showing D-Factor spectrum for RDX superimposed with FTIR spectroscopy of RDX;

[0015] Fig. 8 is a graph of D-factor IR spectrum v. wavenumber with adsorbed TNT;

[0016] Fig. 9 is a graph showing D-Factor spectrum for TNT superimposed with FTIR spectroscopy of TNT;

[0017] Fig. 10 is a graph showing dynamic electrical resonance response of NWRs with various lengths of nanowires;

[0018] Fig. 11 is a graph showing D-factor and resonance frequency response of NWR obtained for various nanowire lengths;

[0019] Fig. 12 is a schematic diagram of an equivalent circuit model for the nanowire and electrodes;

[0020] Fig. 13 is a schematic diagram of a simplified equivalent circuit model for the nanowire at the self -resonating frequency (SRF); [0021] Fig. 14 is a graph showing variation of electrical series inductance of the nanowire as a function of frequency;

[0022] Fig. 15 is a graph showing variation of electrical series capacitance of the nanowire as a function of frequency;

[0023] Fig. 16 is a graph showing variation in D-factor with respect to frequency for the nanowire with (lighter line) and without (darker line) RDX molecules;

[0024] Fig. 17 is a graph showing variation in capacitance with respect to frequency for the nanowire with (lighter line) and without (darker line) RDX molecules;

[0025] Fig. 18 is a graph showing variation in D-factor with respect to frequency for the nanowire with (lighter line) and without (darker line) TNT molecules;

[0026] Fig. 19 is a graph showing variation in capacitance with respect to frequency for the nanowire with (lighter line) and without (darker line) TNT molecules;

[0027] Fig. 20 is a graph showing variation in the D-factor of NWR irradiated at different IR wavelengths without molecules;

[0028] Fig. 21 is a graph showing variation in the D-factor due to change of temperature of the nanowire.

[0029] Fig. 22 is a graph showing change of capacitance with respect to electrode spacing; and

[0030] Fig. 23 is a flow diagram showing a method of identifying adsorbed chemicals on the surface of a nanowire.

DETAILED DESCRIPTION

[0031] Though the disclosure discusses on a specific technique for fabrication of the nanowire resonator (NWR) based sensor, the sensing methodology discussed in this disclosure would be valid with any other NWR based electrical system where the fabrication of nanowire could be by any technique known to those skilled in the art and where the characteristic electrical resonance frequency of the fabricated resonator system is of the order of tens of MHz to GHz.

[0032] There is disclosed a method of sensitively detecting and also at the same time chemically specifying a target analyte from the dynamic impedance response of a NWR based sensor. The method involves adsorption of target analyte molecules on the surface of the NWR sensor, irradiating the NWR with infra-red (IR) radiation and measuring the variations in impedance parameters of the NWR sensor. Impedance parameters can include, for example, resistance, capacitance, inductance and dissipation. Based on the above mentioned method, a specific dissipation IR (D-IR) spectrum as a function of the IR wavelengths unique to the adsorbed chemical species is generated. In a depiction, the adsorbed chemical species may be chemically identified by comparing all or part of the obtained dissipation spectra to a library of D-IR spectra. At present, the obtained spectra are being matched with the available FTIR spectra library to compare the absorption peak positions alone and not the entire shape of the spectrum. The present invention as an example, introduces "Dissipation Spectra" as a completely new method of analyzing chemical species. The generation of a comprehensive dissipation spectra library is conceived as an embodiment which would allow various computer based correlation techniques (e.g., image correlation) to be used in the comparison step to identify the target analyte. A sensor is disclosed comprising a source of IR radiation, a nanowire fabricated on a substrate between electrodes and an impedance analysis based sensing methodology as an example using a commercial Agilent make impedance analyzer. Any other equivalent circuitry with a frequency bandwidth of the order of KHz to GHz designed by those skilled in the art, capable of measuring impedance parameters will also work. The operation of each such analysing instrument comes within the purview of each manufacturer's design and standard operating procedures in order to gather the response data (impedance parameters) as desired or mentioned in the document.

[0033] This technique works for a class of nanowire sensors whose response mechanism may depend on several physical phenomena such as thermal characteristics, electrical potential, electrical resistance or even mechanical changes. In case of semiconducting or metal oxide based nanowire sensors, a change in electrical resistance against adsorption of chemical species may change its dissipation. To give an example: reducing gases such as hydrogen, diffuses in to a metal nanowire and forms dissociated hydrogen dipoles at the metal-insulator interface. Therefore, dissipation based IR spectroscopy may provide further specific information about molecular interactions on the sensor surface. Similarly, a chemical vapor adsorbed on the surface of a conducting polymer sensor may change its physical property; for example, transfer of electron causes the changes in resistance and work function of the polymer. On the other hand, introducing vapor molecules into the polymer matrix may increase its inter-chain distance, which would further affects the electron hopping between different polymer chains. In a similar way, the adsorption of molecules on a polymer may also cause swelling of the polymer due to weak Van der Waals force causing change in conductivity of the polymer. In case of thermal response mechanism, the nanowire' s sensitivity to a temperature change is reflected through a change in the impedance parameters of the NWR. In case of metals, a positive temperature coefficient of resistivity may increase the resistance of the nanowire due to heating when exposed to IR (IR absorption by the adsorbed chemical species) and the heat dissipation gives the dissipation spectra of the adsorbed chemical species. Alternatively, in semiconductors, heating through IR absorption by adsorbed chemical species, may result in charge separation in the nanowire (generate additional charge) effecting an increase in capacitance which in turn gives the dissipation spectra of the adsorbed chemical species.

[0034] In principle, this technique should work for all chemicals that show a dissipation spectrum, which in principle should be all chemicals. RDX and TNT are examples demonstrated. NWR with adsorbed chemical species, irradiated by IR results in heating (due to IR absorption by the chemical analyte) which in turn reflects a dissipation factor change giving the unique dissipation IR (D-IR) spectra of the adsorbed chemical species which is the signature of that chemical.

[0035] Currently, FTIR is one of the well-known conventional techniques to identify chemical signatures. Therefore, we have compared D-spectrum with FTIR. However, D-spectrum also shows certain peaks which are not detected by FTIR. The method may use a library of D-factor spectra developed through routine experimentation. D-spectrum based chemical identification is novel. Typical FTIR detectors are limited by time response of the order of few milliseconds to sub-milliseconds for the most sensitive mercury cadmium telluride (MCT) based detectors in an FTIR instrument. However, the molecular vibrational modes excitation variations are much faster compared to responses of commercially available detectors. Since the time response of our developed sensor is of the order of sub- microseconds, it is more sensitive to IR induced changes. Our proposed detection technique being faster than the most sensitive detector known, and is capable of capturing more sensitive responses as additional peaks which are not detectable by FTIR.

[0036] Atomic force microscopy (AFM) is one of the techniques to know whether the molecules are physically adsorbed on the surface or not. However, FTIR spectra of the adsorbed surface would give chemical signatures of that molecule. In case of other IR spectroscopy detection techniques, the quantity of sample under investigation is usually in the order of milligrams to micrograms. As the mass of adsorbed molecules on nanowire surface is typically in the order of sub-picograms, it needs a selectivity signature to confirm whether the molecules are adsorbed or not. The dissipation signature or the D-spectra itself is the validation that the material is adsorbed on the surface. In a practical system embodiment, the measurement technique always compares the spectra with respect to base signal where you take a spectrum by passing dry air. The D-spectra while passing dry air acts as the reference signal and the obtained D-spectra on adsorption of a chemical changes because of its unique IR absorption which in turn changes the impedance parameters of NWR sensors as discussed in the present invention.

[0037] Selection of IR range depends on the chemical of interest. Most of the molecules have absorption in mid-IR (3-8 um) which is considered as the molecular fingerprint regime. In the present art, we have used an IR source (with multiple lasers) which can scan from 3-15 um. Selection of source wavelength depends solely on the type of chemical under examination and one may acquire this information from known FTIR database. However, D-spectra itself may act as the basis for selection of wavelengths, once we generate a D-spectrum library.

[0038] In the present invention, we have used two different IR sources with different peak power for IR spectroscopy and our method works for both. Regarding optical setup, we have tested our method with direct laser illumination as well as with focussed optics to get D-IR spectra. Fabricating identical NWR sensors by conventional nanofabrication techniques may serve as the reference devices for characterizing the power requirements of the IR source. But as a concept, it should work for any such developed NWR based sensor provided the working frequency is the regime of tens of MHz to GHz. In terms of beam optics, the device is not limited by beam polarization effects or wave front modulations. It would surely show a variable response in terms of peak amplitude depending on the intensity of IR beam and the particular device being used. The present technique should eventually define or push the minimum detectability limit of the detector and is envisaged as the future work. As regards the line width resolution of the source - it would also depend on the type of target molecules. Higher the resolution of the source better it is for the NWR since the typical thermal broadening of observed spectra by our technique is much smaller compared to that observed in conventional FTIR based techniques. This further allows interpreting the peaks with a resolution of 1 wavenumber (that is the maximum resolution of the currently used lasers). The low thermal mass of the NWR sensor along with the faster time response allows us to obtain such minimal line width broadening in the obtained D- IR spectra.

[0039] Fig. 1 shows an embodiment of a sensor set up. Fig. 3 shows a similar embodiment and shows an infrared source 7 which is not shown in Fig. 1, although present in the embodiment of Fig. 1. In Fig. 2 and in Fig. 3, the substrate, nanowire and electrodes are shown together as a single wire labeled as "sensor" and indicated by reference numeral 9. Electrodes 1 are positioned on a substrate 2, and a nanowire 3 is connected between the electrodes. Impedence analyzer 4 is connected to the electrodes via lead wires 5 to measure impedance parameters of the nanowire. Infrared radiation 6 impinges on nanowire 3 from an infrared source (not shown in Fig. 1, but indicated with reference numeral 7 in Fig. 3). In the embodiment shown, the infrared source 7 is a quantum cascade laser (QCL).

[0040] The metal electrodes 1 are Pt/Ti electrodes in the embodiment shown, but any other conducting electrodes would also do. Fabrication procedure of electrode pads in our case has been mentioned in the document. Fabrication of electrode pads could be by any methodology known to those skilled in the art.

[0041] Substrate 2 comprises Si02 on Si in the embodiment shown, but any other substrate not showing absorption in the working IR range of the target chemical species could be used.

[0042] Nanowire 3 is a Bismuth Ferrite (BFO) Nanowire in the embodiment shown - for this specific material the range is 30 - 200 run diameter and a maximum length of 30 um which gives the operating self resonating frequency (SRF) of tens of MHz. For any other type of material to keep the SRF in the order of tens of MHz or higher, the diameter and length constraints will vary depending on the material property.

[0043] This technique works for any nanoresonator with self -resonance frequency in the range of tens of MHz to 100's of GHz. The above mentioned self-resonance frequency depends on the material properties and its dimensions. The nanowires can be synthesized by different techniques; Electrospinning is one way to make them. This technique doesn't depends on how nanowires are made rather it depends on the dimensions of the structure and the material property. Fabrication of nanowires could be by any methodology known to those skilled in the art.

[0044] The Impedance Analyzer 4 in the embodiment shown in Fig. 1 is an Agilent 4294A.

Any other equivalent circuitry with a frequency bandwidth of the order of KhZ to GhZ capable of measuring impedance parameters will also work. The operation of each comes within the purview of each manufacturer's design and standard operating procedures to get the response data (Impedance parameters) as desired or mentioned in the document.

[0045] Lead wire 5 connect the Metal electrodes(l) to the Impedance Analyzer(4) or equivalent circuitry and should have an impedance matched to the impedance analyzer or equivalent circuitry.

[0046] For the infrared radiation 6, any IR source with wavelength range in the molecular fingerprint region ( 2 - 14 um) works. For specific chemicals IR LED sources can also be used with fixed wavelengths. Any optical source to stimulate vibrational modes of target chemical species would work.

[0047] Here, we introduce a method of detecting adsorbed chemicals, where suitably pulsed IR on a semiconductor nanowire shows magnified response at its electrical resonance. IR absorption by even minute quantities of adsorbates on the nanowire excites thermal phonons internally. In turn, excited carriers repartition into the high number of available surface states in the nanowire, predominantly effecting the internal dissipation at resonance. Monitoring this dissipation as a function of IR irradiation wavelengths, offers a simple and elegant new platform for discriminating femtogram (fg) level adsorbed chemical species. Bismuth Ferrite (BiFe0₃, BFO) nanowires have been used for technique demonstration. This approach offers a unique and universal method for receptor-free molecular recognition of extremely small amounts of adsorbates using an electrical route with wide-ranging applicability. It also opens up avenues for studying phonon - carrier coupling dynamics at the nanoscale for basic and applied research.

[0048] A semiconductor BiFe0₃ (BFO) nanowire is driven to electrical resonance externally.

The nanowire resonator (NWR) behaves analogous to an electrical series RLC resonant circuit with effective inductance L and a non-ideal capacitance C , typical order of magnitudes of which at the

1

nanoscale, exhibits resonance frequency in the MHz regime following Variations in the response characteristics, subject to changes on the surface dominated low thermal mass system, are being monitored electrically. The high inherent electrical resonant frequency at the nanoscale significantly enhances the time response sensitivity by an order of magnitude compared to a mechanical microcantilever resonator system (~KHz). The observed electrical resonance response of the employed NWR, promptly recalls the characteristics of a non-ideal capacitive element with an effective inductance in series (circuit model shown as an inset in Fig 2, and also in Fig. 12). At the self -resonating frequency (SRF), i∑> = 2π f_res , the capacitive reactance {X_c ) and inductive reactance {X_L ) exactly compensate each other Xr Xr = coL„ making the device predominately resistive or dissipative. In coC„

effect, the storage of energy in such a non-ideal capacitive element at SRF is through X_c , per alternating cycle, while the dissipation is through its effective frequency dependent series resistive component R_s . R_s and X_c are representations in terms of the impedance parameters obtained from the impedance analyzer used in our study. Essentially, the ratio of the energy lost to that stored per cycle becomes a critical and sensitive measure of the quality of resonance. This ratio, commonly referred as the dissipation factor or D-factor (dimensionless) is given by

[0049] D =— =— = *C_S

[0050] The variations of resonance characteristics of the NWR are sensitively reflected by D- factor responses when subjected to external parameter variations, in turn reflecting changes in its internal dissipation. The observed variations in SRF from the D-factor response as a function of molecular adsorption are significant. On the other hand, optical IR excitation of adsorbed species significantly changes the peak amplitude of D-factor response with only minute changes in its SRF. This may be accounted for by minute temperature changes of the nanowire predominately affecting its internal dissipation. The thermally generated carriers change the way energy gets dissipated and stored in the NWR per cycle of its oscillation when driven electrically (externally) at resonance. This resonance response variation, typical of the nanowire system employed, provides deeper insights on its thermal response characteristics through electrical resonance, which is exploited here as a basis for receptor free IR chemical discrimination.

[0051] The coupling of electrical resonance of NWR with optical resonance of adsorbed molecules results in an entirely new platform that combines the high response sensitivity of nanowires and the high selectivity of IR spectroscopy enabling chemical discrimination as demonstrated in this work. Only nanosystems with high surface-to-volume ratio allow such coupling due to dominant surface effects, inherently providing high detection sensitivity without the need of any complex vacuum apparatus. The importance of surface states in modulating responses of semiconductor materials was raised in the early 1970's by the pioneering work of Gatos and Lagowski. We add to their insights the dimension of selectivity by utilizing the distinct molecular vibrational characteristics of chemical species in the mid-IR regime. As evident, the thermal response characteristics of a semiconductor nanowire are controlled by its low thermal mass. The observed temperature modulated electrical response sensitivity stems from the presence of a large number of surface states which govern the phonon de-excitation processes internally. These processes are usually in the form of a cascading multi- phonon relaxation process reflected as internal dissipation. Internal dissipation is often attributed to phonon-phonon interactions or scattering where defects in bulk or on the surface enhances it significantly. Internal dissipation is often attributed to phonon-phonon interactions or scattering where defects in bulk or on the surface enhances or facilitates multi-phonon relaxations through added lattice vibration anharmonicities. Realistically, there is a higher probability of phonon relaxation coupled charge carrier promotion to the allowable excess surface states (higher surface charges - carrier separation), often termed charge capture or carrier trapping leading to relative higher internal dissipation. Such carrier repartitioning into different allowable surface states reflects as a dominant X_c change in electrical terms, as observed through dominant variations in D-factor amplitude at resonance. A systematic recording of the variations in D-factor as a function of incident IR wavelengths gives the dynamic dissipation IR spectrum (DDIRS) of the adsorbed species which resembles the IR absorption spectrum of the adsorbates. The D-factor has been used as a sensor signal for differential detection of volatile chemicals, but without any discriminating chemical selectivity. Our approach of coupling the electrical resonance to mulit-phonon relaxations on a surface state dominated semiconductor NWR brings selectivity through specific (IR) absorption based DDIRS.

The metal oxide semiconductor (BFO) NWR shows considerable promise for sensitive response and selectivity, especially in terms of line-width resolution of the obtained spectra. This high line-width resolution is primarily due to the low thermal mass and high SRF of the NWR. The schematic of overall concept employed for the IR spectroscopy of surface adsorbed molecules on NWR platform is shown in Fig. 2. The diameter of the nanowire is about 100 nm whereas the length varies from 3-20 urn for different NWRs employed. Electrical resonance frequency of the nanowire changes due to molecular adsorption, enabling detection of fg levels of adsorbed mass. Resonant IR excitation of adsorbed molecules produce large changes in the dissipation of nanowire resonator due to population- depopulation of surface states by thermally generated carriers. The electrical resonance frequencies for nanowires of various lengths were measured at room temperature (Fig. 10). The wavelength dependent dissipation variation due to IR absorption induced lattice phonon relaxations in the nanowire is reflected through a dynamic dissipation IR spectrum (Fig. 6, Fig. 8) which corresponds to the adsorbed species' unique (IR) absorption characteristics. This plot of measured dissipation (D-factor) as a function of IR excitation resembles IR spectra of the adsorbed molecules. The NWR serves as an extremely sensitive thermal sensor and an electrical resonator platform, enabling sensitive recognition and quantification of adsorbed molecules through variations in its dynamic impedance parameters.

[0052] For demonstrating the capability of the proposed sensor platform we have chosen commonly investigated explosives, trinitrotoluene (TNT) and cyclotrimethylene trinitramine (RDX) as model systems. Since these explosives molecules bind to surfaces very easily, they remain on the nanowires for longer periods for repeated measurements. The changes in the D-factor due to the adsorption of RDX and TNT are presented in Figs. 4 and 5, respectively. Insets of Fig. 4 and 5 show higher magnification in the region of interest. The observed changes in X_c of the NWR at SRF may be explained in relation to surface state properties. Previously reported observations by the authors on size effects and relevant studies by others shed light on plausible explanations for the observed induced capacitance effects. In agreement with previous reports, the authors also believe that domain wall motion and space charge contributions in the BFO nanowire contribute to effective complex permittivity variations, thereby predominantly causing X_c variation in its response due to the adsorption of analyte molecules.. The higher surface area of a nanowire coupled with higher number of surface states promotes unprecedented mass resolution of detection. Adsorbed mass on the NWR from the resonance response shift and corresponding capacitance change as a function of charge donation or transfer from the adsorbed chemical species of unknown mass is estimated to be of the order of \0fg . This estimated mass is also verified within the same order of magnitude (60 fg), assuming fractional two dimensional (2D) surface exposure of the nanowire to 0.4 μ∑ droplet. The factor of variation between the measured capacitance change and the surface area exposure estimation may be accounted for the non-uniform and different evaporation kinetics of solvent on the surface of the nanowire and the substrate. The authors believe that the residual target analyte adsorption on the nanowire cannot be in the form of a continuous layer, instead would be in the form of discrete islands on the nanowire surface. The technique thus also opens up a novel way of estimating adsorbed mass from electrical property variation through dynamic dissipation study at resonance. Variations in the D-factor of the NWR with RDX and TNT molecules adsorbed on their surface as a function of IR wavelength were found. The D-factors, when graphed with respect to frequencies, showed peaks of height varying with respect to wavenumber. The higher peak dissipation responses correspond to the absorption wavenumbers of RDX and TNT, 1578 cm^"1 and 1160 cm^"1 respectively. The adsorbate specific IR absorption excites thermal phonons which couple to the surface states of the nanowire through multi-phonon assisted (MP A) relaxation processes effectively modulating the internal dissipation of system as demonstrated. The selectivity in detection is through to the unique spectral absorption characteristics of the adsorbates in the mid-IR region. A variation in internal dissipation of the NWR is reflected by its dynamic impedance variation (D-factor) in proportion to the small temperature changes due to IR absorption by adsorbates. The D-factor response of NWR with adsorbed molecules (without IR irradiation) served as the reference signal. The observed DDIRS of RDX and TNT molecules adsorbed on different NWR devices are presented in Fig. 6 and 8 respectively. The IR spectra were obtained while measuring D-factor at their respective resonant frequencies. D-factor spectra as a function of IR wavelengths for the NWR without the adsorbed analyte molecules becomes the reference or background signal in our analysis and is used for background corrections (Fig. 20). The proposed dissipation signature as a function of IR wavelengths matches the FTIR spectra of adsorbed analyte molecules (Fig. 7 and 9).

[0053] The variation in the D-factor is predominantly due to variations in the effective capacitance of the NWR. Capacitive spectroscopy has been employed in measuring the capacitance of a p-n junction or a Schottky barrier to analyze variations in the charge state of deep levels in the space charge layer associated with such junctions. The variation in the capacitance affecting the NWR response likewise reflects the variations in the charge state of surface energy levels of the semiconductor nanowire. It is believed that the observed increase in capacitance is a result of an increase in surface charge carriers (electrons or holes) in the unoccupied surface state energy levels of the BFO nanowire by MPA processes as has been reported previously for MOS thin films. Also, as widely known, typically in a wide bandgap semiconductor material, surface states exist in the forbidden energy gap with a specific distribution up to the Fermi level. Photons with energy smaller than that of the bandgap energy get absorbed (exhibiting absorption peaks) which are believed to be promoted by available surface states effecting allowable electronic transitions. Nanowires having a high surface-to-volume ratio exhibit even higher density distribution of surface states which allow higher or more efficient sub- band transitions. It has also been mentioned that these surface states may even lead to highly inhomogeneous current distributions in nanowires, which may affect other physical properties such as local temperature and heat conduction directly in agreement with our observed response. The occupation of a vacant site at a particular energy level should correspond to the structural lattice energy relaxation. The presence of high density surface states (degenerate states) in the employed BFO-nanowire allows these transitions to take place thus modulating the surface state occupancy. Conclusively, higher carrier separation inherently increases the effective series capacitance (or decreases the capacitive reactance) at the SRF of the NWR device as observed exhibiting a higher D-factor change, which in turn becomes the IR specific dissipation signature. The efficiency or probability of such carrier trapping can be explained in terms of a detailed capture cross section analysis with respect to the nanowires' degenerate surface energy levels at room temperature. The variation in complex electrical parameters (e.g., D-factor) subjected to higher energy floor variations due to temperature changes have been conducted to see the trend (Fig. 21).

[0054] Another fundamental aspect revealed through the obtained spectra relates to the temperature dependence of the carrier capture rate due to the extremely low thermal mass of the nanowire. Low thermal mass of the nanowire results in a high rate of change of temperature with respect f dT Δ

to time — I and since the operational frequency (SRF) of the device is of the order of tens of MHz, dt J

the time response of D-factor variation due to IR absorption comes to an order of sub^s. As a result the line-width resolution of the spectra is significantly reduced due to modulation by the thermal response characteristics of the nanowire. Photothermal deflection spectroscopy (PCDS) using receptor free bi- material cantilevers with mass of tens of nanogram (ng) can produce spectral lines for pictograms (pg) of adsorbed chemical species. However, the time period of response of the technique is still of the order of ~ ms [l/(l^st eigen mode frequency)] limiting the line-width resolution of the measured spectra. Usually, the broadening of the peaks in conventional solid and liquid phase IR spectra is caused by the relaxation and dephasing of the vibrational excited states and indicates the complex fast dynamic interaction of the molecule with its environment. The high inherent SRF along with the low thermal mass of the NWR has the advantage of faster dynamic response as reflected in the obtained spectra.

[0055] In conclusion, the possibility of achieving highly selective IR chemical discrimination with fg levels of adsorbed mass on a surface state dominated nanowire platform using electrical resonance is introduced. The key lies in coupling electrical resonance to phonon induced surface state occupancy perturbations in such a low thermal mass system. This concept exploits the strengths of nanosystems reflected by enhanced effects at resonance. The achievable high mass resolution and IR response sensitivity as demonstrated here should be viewed in the light of measurable impedance parameter variation as a function of IR induced thermal modulations. Resonant excitation of adsorbed molecules varies nanowire electrical characteristics due to thermally-induced population-depopulation of surface electronic states on the nanowire and in turn changes the electrical resonance parameters of the nanowire. The BFO NWR system described here utilizes the internal dissipation due to IR absorption by the adsorbed species and opens new opportunities for detecting minute amounts of surface adsorbed molecules on similar nano-resonating platforms using dissipation as the parameter. It is also possible that the thermal response of nanostructures as a function of excitation frequency in the UV- Visible region can be a tool to map the surfaces states of the semiconductor material. The NWR-DDIRS shows great promise as a tool for obtaining IR spectra of extremely small amounts of adsorbed molecules. With optimization, this method provides exciting opportunities in developing a sensitive platform for potential realization of single-molecule detection with superior selectivity performance.

[0056] Materials and methods [0057] Preparation of BFO NWR: BFO NWRs with various electrode spacing are fabricated directly on pre-patterned substrates by electrospinning technique. A gas injection system (GIS) available with motorized flexible xyz-drive was used for in situ platinum metal contact to these nanowires (RAITH150™). The residues of individual explosive molecules of RDX and TNT were deposited on the NWR using the droplet evaporation method.

[0058] Chemicals: The standard explosive samples (RDX and TNT) were purchased from

AccuStandard™, Inc. (New Haven, CT) and used without further purification. The standard concentration of each explosive is 1000 μg/mL in MeOH:ACCN (1 : 1) as indicated by the manufacturer.

[0059] Dynamic impedance IR spectroscopy setup: The IR radiation (pulsed at 200 kHz) from the quantum cascade laser (QCL) (Daylight Solutions™ UT-8) was focussed on the NWR. The wavenumber of IR source was fixed at a specific value (range: 1630 cm^"1 to 1150 cm^"1) and the corresponding dissipation parameter was measured. The impedance parameters of the NWR were measured using an Agilent™ 4294A impedance analyzer having a frequency range of 40 Hz to 110 MHz with nominal impedance accuracy: +/-0.08 % at 100Hz. The excellent high quality factor (Q) or D accuracy enables reliable analysis of low-loss components. The inherent high dynamic range of the equipment allows evaluation under actual operating conditions. A fixed ac test signal level - V_rms ~ 50 mV was employed for all the impedance measurements.

[0060] FTIR spectroscopy: The explosive residues were characterized using a standard FTIR

Thermo Scientific™ Nicolet™ Continuum™ infrared microscope with a potassium bromide (KBr) beam splitter and a MCT-A (narrow band 650 cm^"1 cut-off) detector microscope in reflection mode. The number of registered scans was 200 with resolution of 4 cm^"1.

[0061] Nanowire resonators (NWRs) and their electrical resonance

[0062] Dynamic measurements have been performed on NWRs in a wide range of frequencies

(ranging from 20-50 MHz). The resonances for different nanowire designs measured at room temperature are presented in Fig. 10. The resonance frequencies of NWRs exhibit a decreasing trend as a function of increasing nanowire length (electrode spacing); i.e. the leftmost peak in Fig. 10 is for 20 μιη, the next is 15 μιη, etc. The D-factor response is just the reverse. The variation of electrical resonance frequency and D-factor of NWR as a function of the nanowire length is presented in Fig. 11. The observed variations can be attributed to the increasing capacitance (decreasing capacitive reactance (X_c)) as a function of decreasing electrode spacing which effectively is the length of the nanowire. In essence it is the predominant variation of electrical length at the nanoscale that governs the operational self-resonating frequency (SRF).

[0063] Equivalent circuit model for the NWR

[0064] The device behaves as a typical electrical series RLC circuit as shown in Fig. 12 where its typical component values drive it to a resonance frequency of the order of tens of MHz.

R_s , C_S & L_s are representations of equivalent components as measurable through the impedance analyzer. The nanowire suspends freely in air at a height of 100 nm from the substrate surface and hence the air capacitance C_AIR , as seen by the two electrodes, does not show a dominant effect in determining the resonance frequency of the equivalent circuitry. Also typical value of the substrate capacitance C _substmte along with the contact resistances R_c /2 at the electrodes, exhibits a resonance frequency in the order of few KHz only, with no nanowire drawn between the electrodes. This ensures that our device response at the MHz frequency regime is that of the nanowire and not dominated by changes on the substrate or the changes in medium around it. The effective dominant D-factor change used in our study can be explained in terms of the ratio of the lost to stored energy through its frequency dependant effective complex impedance parameters. The response of the NWR recalls to mind the response of a non-ideal capacitive element following Fig. 13 where, it is well known that at SRF, the capacitive and inductive reactance values become equal Xr X , = oL„ and the device becomes predominately resistive ( R_s ), o being the SRF. The inductance L_s variation, in one of the employed fabricated NWR, as a function of frequency (Fig. 14) is negligible as is evident from its response curve, conclusively implying that the reactance change of the resonator circuit tends to be dominantly capacitive. Following the same argument, the capacitance response of the NWR (Fig. 15), indeed also exhibits maximum variation at SRFjustifying the series equivalent model (Fig. 13). Evidently, in such a non-ideal capacitive element at SRF, the storage of energy per cycle is through its capacitive reactance X_c while the dissipation is through its effective frequency dependent series resistance component R_s . Typical RLC values for Fig. 12 are R ~ Ω,

L ~10 ⁷ /f and C ~ « )-

[0065] D-factor response due to the adsorption of molecules

[0066] As discussed earlier, the effective D-factor is the ratio of dissipated to stored energy per cycle x, TIC., T

1

T being the time period of a cycle.

[0067] A quick analysis of the obtained data (Fig. 17) shows an increase in effective capacitance on adsorption of RDX molecules by a factor of 1.038 following equation (i). Similar trend (a factor of 1.039) in increase in capacitance is also observed in a different NWR on absorption of TNT molecules (Fig. 19). Increase in capacitance corresponds to an effective decrease in capacitive reactance which leads to reduced storage of energy per cycle of response thus leading to a higher dissipation at its SRF as also clear from equation (i) above. The shift in SRF on adsorption may be attributed to the effective complex variation in resonator elements (dominant real part variation of complex permittivity), This observed response agrees to an earlier report where it was envisaged that the D-factor variations can be a useful indicator of the change in dipole-dipole interactions caused by the adsorption of a volatile chemical onto the polymer film employed in the study. They related the higher D-factor at the device's SRF to the maximum efficient transfer of energy at the natural resonance frequency of molecules binding to the polymer. Such efficient energy transfer was conclusively associated to lower energy storage at SRF in their developed sensor system.

[0068] Figs. 16 and 17 are graphs showing variation in D-factor and capacitance respectively with respect to frequency for the nanowire with (lighter line) and without (darker line) RDX molecules. The order of shift is in KHz for even minute quantities of adsorbates of the order of fg.

[0069] Figs. 18 and 19 are graphs showing variation in D-factor and capacitance respectively with respect to frequency for the nanowire with (lighter line) and without (darker line) TNT molecules. Same KHz order shift is observed as for RDX reflecting the high detection resolution.

[0070] Adsorbed RDX mass calculation

[0071] From electrical resonance of the NWR

[0072] From the dynamic impedance parameter variations using equation (i),

AC = 0.038 x 4.48/iF

AQ = AC x V = 8.5 \2 x \ 0 ^u C

[0073] Now charge of an electron is e^~ = 1.602 x 1(T¹⁹ C

[0074] Therefore, Number of electrons N_e =— = 5.3134 x 10⁷

e

[0075] Now charge contribution per atom from individual sites of ¾ H₂ and N₂ in RDX

= 0.84e^~ . There are 3 such sites in a RDX molecule . Hence the number of atoms taking part in charge change

N , _Rny = ^Ne = 2.1085 x 10⁷

3 x 0.84

[0076] Therefore, weight of RDX can be calculated as

N atoms, RDX

AvagNo.

1.23 x lO⁹

moles x 222. \2gm I mole

6.023 x lO²

= 7.78 x \0 '⁵ gm≡\0fg

[0077] From surface coverage

[0078] Typically, in monolayer coverage, it is considered that the total number of atoms in 1 cm² is about 10¹⁵ atoms/cm². However, for adsorbed mass calculations, authors consider that only fraction of RDX/TNT droplet is exposed to effectively half the nanowire surface. Therefore, the fraction of half the surface area of nanowire to the total area covered by 0.4 μΐ droplet on the substrate surface (as measured from the droplet stain) multiplied with concentration of RDX/TNT molecules in the solution used gives the estimated mass of adsorbed molecules.

[0079] Diameter of nanowire = 100 nm, Length of nanowire = 3 μιη

[0080] Concentration of RDX/TNT molecules: 1000

[0081] Volume of droplet (RDX/TNT) = 0.4/Λ

[0082] Hence effective mass of analyte in = x 1 ig / Λ = 0.4 ig

[0083] Droplet diameter on surface =2mm (measured)

Ttd¹ ;r x (2 x l 0^{~3 2}

[0084] Area of droplet on substrate: = - — = 3.14 x \0^'6m²

4 4

[0085] Area of nanowire: d^l = π x 100 x 10^~9 x 3 x I0^'6m²

[0086] Assuming only top surface of nanowire is covered with droplet, area exposed

^ _ ^ Χ ΐ00 χ 10-⁹ χ 3 χ 10^ _{= 4 712 χ 1 ()}_₁₃^₂

2 2

4.712x l0^~13

[0087] Hence the estimated adsorbed mass is — — x 0.4 10 = 60 fg

3.14 x l0^~6

[0088] The variation in D-factor due to IR absorption by molecules

[0089] The overall D-factor rise as a function of IR wavelengths on absorption by the adsorbed analyte, predominantly becomes a function of the effective series capacitance change (higher capacitance and lower capacitive reactance - dominant imaginary part variation of complex permittivity reflecting more loss). The capacitance variation is due to excited phonon relaxation induced carrier separation amplified by the higher density of surface states. Wavenumbers with higher peak D-factor also had larger spikes in capacitance below the resonant frequency, but also shallower troughs in capacitance above the resonant frequency.

[0090] Fig. 20 is a graph showing variation in the D-factor of NWR irradiated at different IR wavelengths without molecules, which is used as a background signal. The inset shows a zoomed version of the background noise.

[0091] The variation in D-factor of NWR due to thermal changes

[0092] Fig. 21 is a graph showing variation in the D-factor due to change of temperature of the nanowire. Variation in the D-factor due to change in temperature of NWR. The electrical impedance parameters of BFO nanowire are very sensitive to changes in its temperature. Small changes in the temperature of the nanowire results in a large variation in dissipation factor (D-factor) at resonance due to thermally generated carriers.

[0093] Fig. 22 is a graph showing change of capacitance with respect to electrode spacing.

[0094] Fig. 23 is a flow diagram showing a method of identifying adsorbed chemicals on the surface of a nanowire. In step 20, molecules of a chemical are adsorbed on the surface of a nanowire. In step 22, the nanowire is irradiated with infrared radiation. In step 24, variations in impedance parameters of the nanowire are measured. In step 26, the adsorbed chemical is detected based on the variations in impedance parameters. If detection is sufficient, the method can be stopped at this point. In step 28, a dissipation infrared spectrum of the adsorbed chemical is obtained by determining dissipation of the nanowire at multiple wavelengths of infrared light. In step 30, the adsorbed chemical is identified by comparing all or part of the obtained dissipation infrared spectrum to a library of dissipation infrared spectra.

[0095] Excitation of vibrational modes of adsorbed molecules when irradiated with IR energizes lattice phonon states of the nanowire (NW) with low thermal mass. Such multi-phonon assisted structural lattice energy relaxations promotes carrier trapping at unoccupied trap levels of the BFO-NWR effecting variation in dynamic impedance parameters of the NWR. The proposed sensing technique colligates the dissipation factor (D-factor) change (as a function of the irradiated wavelengths) with the IR absorption spectra of the adsorbed chemical species on the NWR through the predominant variation in capacitive reactance of the NWR.

[0096] The developed metal oxide semiconductor (BFO) nanowire resonator (NWR) sensor shows considerable promise on both aspects of sensitivity and selectivity primarily for their low thermal mass and high SRF. On grounds of sensitivity, advantages pertain to material properties arising due their high surface-to-volume ratios and Debye lengths comparable to the target molecule. With respect to the aspect of selectivity, variation in D-factor due to IR absorption by adsorbed molecular species in such a NWR sensor is proposed as an additional signature. Such IR absorption initiates the transitions between vibrational energy levels of adsorbed molecules which in turn excites phonon states in the employed semiconductor BFO NWs. Recently surface states were shown to play an important part in the optical properties of BFO NWs. Presence of excess surface charges on account of the excess surface states in the BFO NWs is believed to be the governing factor for exhibiting such higher sensitivity with respect to IR absorption specific selectivity signature. Thus the present work not only brings selectivity through specific (IR) absorption based dissipation (D-factor) signature but also proposes the enhancement of sensitivity by employing NWR platform.

[0097] NWRs and their electrical resonance

[0098] The diameter of the NW used in the experiments done is lOOnm with length varying from 3-20 urn for different NWRs.

[0099] Dynamic measurements have been performed on NWRs in a wide range of frequencies

(10-70 MHz). The resonances for different NW designs measured at room temperature are presented in Fig. 10. The resonance frequencies of NWRs exhibit a decreasing trend as a function of increasing NW length (electrode spacing) whereas the D-factor response is just the reverse. The variation of electrical resonance frequency and D-factor of NWRs as a function of length of NWR is presented in Figure 11. The observed variations can be attributed to the increasing capacitance or decreasing capacitive reactance (X_c) as a function of decreasing electrode spacing (shown in Fig. 22) which effectively is the length of the NW. [00100] D-f actor IR response in the NWR

[00101] The overall D-factor change as a function of IR wavelengths on absorption by an adsorbed analyte species predominantly becomes a function of the effective series capacitance change (higher capacitance and lower capacitive reactance) which may be attributed to the excitation of non- radiative phonon states in the NW. IR absorption by the adsorbed molecules excites its vibrational modes which in turn excites the non-radiative phonon states of the low thermal mass NW component. It is believed that the observed increase in capacitance is due to increase in surface charge carriers (electrons or holes) in the unoccupied surface state levels of the BFO NW by multiphonon-assisted (MP A) processes as has been reflected upon by various authors for MOS thin films . The occupation of a vacant site at a particular energy level should correspond to the structural lattice energy relaxation. It is an observation similar to that reported for sub-bandgap optical absorption spectra. It is widely known that typically, in a wide bandgap semiconductor material, surface states exist in the forbidden energy gap with a specific distribution up to the Fermi level. It has been observed that photons with energy smaller than the bandgap energy gets absorbed (exhibiting absorption peaks) which are believed to be promoted by available surface states effecting allowable electronic transitions. NWs having a high surface-to-volume ratio exhibit even higher density distribution of surface states allowing even higher or efficient sub-band transitions . Such high density distribution of surface states causes the Fermi energy level to be pinned resulting surface band bending showing enhanced surface state emissions. It has been indeed observed that surface-state emissions in such NWs compete with band-edge emissions, greatly reducing the luminescence efficiency near the band-edge. It has also been commented that these surface states may even lead to highly inhomogeneous current distributions in NWs, which might affect other physical properties such as local temperature and heat conduction. The higher density distribution of surface states in the developed BFO NWRs (high surface-to-volume ratio) is believed to promote enhanced charge carriers at the surface states on excitation of the phonon states due to IR absorption by the adsorbed molecules. Higher carrier trapping increases the effective series capacitance or decreases the capacitive reactance at the SRF of the NWR device as observed exhibiting a higher D-factor change following equation (i), which in turn becomes the IR specific dissipation signature.

[00102] The proposed dissipation signature as a function of IR wavelengths resembles FTIR spectra of adsorbed analyte molecules (in this case RDX and TNT) supporting our claims. High inherent SRF has the advantage of faster dynamic response which along with the low thermal mass of the NW proposes to set a new benchmark on the aspect of reduced thermal broadening in the obtained IR absorption spectra. D-factor spectra as a function of IR wavelengths for the NW without the adsorbed analyte molecules becomes the reference or background signal in our analysis (Fig. 20).

[00103] The change in D-factor in a NWR as a response signature is being corroborated to the excitation of vibrational modes of adsorbed molecules, when irradiated with IR leading to excitation of phonon states in the low thermal mass NW. The occupation of a vacant trap site at a particular energy by an energized carrier corresponds to the excited structural lattice energy relaxation causing a change in the effective capacitive reactance of the NWR system which in turn relates to the IR absorption spectra of the adsorbed species through a variation in the D-factor. The developed BFO NWR sensor system uses to its advantage the dissipation of its response as an IR absorption signature of the adsorbed species and opens up opportunities for sensing surface adsorbed molecules on other resonating platforms using dissipation as the parameter. It is also believed that the thermal response of nano-dimensioned material structures as a function of frequency can also be an active study with such a sensor resonator devised out of such a material.

[00104] Immaterial modifications may be made to the embodiments described here without departing from what is covered by the claims.

[00105] In the claims, the word "comprising" is used in its inclusive sense and does not exclude other elements being present. The indefinite articles "a" and "an" before a claim feature do not exclude more than one of the feature being present. Each one of the individual features described here may be used in one or more embodiments and is not, by virtue only of being described here, to be construed as essential to all embodiments as defined by the claims.

Claims

THE EMBODIMENTS OF THE INVENTION IN WHICH AN EXCLUSIVE PROPERTY OR PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:

1. A method of detection of a chemical, comprising:

adsorbing molecules of the chemical on the surface of a nanowire;

irradiating the nanowire with infrared radiation;

measuring variations in impedance parameters of the nanowire; and

detecting the chemical based on the variations in impedance parameters.

2. The method of claim 1 in which the step of measuring variations in impedance parameters of the nanowire includes measuring dissipation.

3. The method of claim 2 in which a dissipation infrared spectrum of the adsorbed chemical is obtained by determining dissipation of the nanowire at multiple wavelengths of infrared light.

4. The method of claim 3 further comprising identifying the adsorbed chemical by comparing all or part of the obtained dissipation infrared spectrum to a library of dissipation infrared spectra.

5. A chemical sensor comprising:

a nanowire having a surface exposed to adsorption of molecules of a chemical;

a source of infrared radiation positioned to illuminate at least a part of the exposed surface of the nanowire with infrared radiation; and

an impedance analyzer connected to measure impedance parameters of the nanowire.