BLOOD GLUCOSE LEVELS

TECHNICAL FIELD

[0001] The present application relates to the field of microelectronic sensors based on high-electron-mobility transistors and their use in detection and continuous monitoring of electrical signals generated by a human body in a sub-terahertz range. In particular, the present application relates to the open-gate pseudo-conducting high-electron-mobility transistors (PC-HEMTs) and their use in non-invasive monitoring of blood glucose levels.

BACKGROUND

[0002] The prevalence of diabetes is constantly on the rise. With over 420 million diabetic patients worldwide and the number of pre-diabetic people at almost three times this number, diabetes has become the largest global healthcare concern. People who suffer from diabetes can live fuller, longer, and healthier lives by carefully regulating their blood glucose levels. To do this effectively, blood glucose levels must be monitored on a regular basis, in some cases as often as four or five times a day and more. Diabetes Control and Complications Trial (DCCT) Research Group in "The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus ", N. Engl. J. Med. 1993, 329(14), 977-986, and UK Prospective Diabetes Study (UKPDS) Group in "Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in subjects with type 2 diabetes" in Lancet. 1998, 352(9131), 837-853, have shown conclusively that the blood glucose monitoring reduces the long-term complications of diabetes. Moreover, according to the American Diabetes Association in their Clinical Practice Recommendations, "Standards of medical care in diabetes", Diabetes Care, 2006, 29, S4-S42, self-monitoring of blood glucose has a positive impact on the outcome of therapy with insulin, oral agents, and medical nutrition. It is useful in generating knowledge about individual glucose profiles, as well as knowledge about the effects of one's habits, including exercise and food intake on that profile, thus helping to achieve specific glycaemic goals.

[0003] However, there are many situations whereby conventional blood glucose testing is challenging. The blood glucose level monitoring with invasive devices for many people is painful, expensive, inconvenient, and complexity involved in the self- monitoring of blood glucose lead to its underutilisation. The pain comes from a lancet pricking of the finger-tip to produce the required blood drop for testing. The expense comes from the consumable test strips being inserted into a reader, on which the blood drop is placed. Steady hands and a considerable level of dexterity are required when testing oneself. As a result of the pain, expense and inconvenience, many diabetics do not monitor their blood glucose levels as often as they should. Often people with diabetes will unnecessarily expose themselves to dangerous levels of glucose rather than prick their fingers. Availability of an accurate, painless, and easy-to-operate device will encourage more frequent testing, leading to tighter blood glucose control and a delaying or decreasing of long-term complications and their associated health care costs. Thus, people with diabetes have longed for a glucose monitor that is non-invasive, non- intrusive, affordable, and allows testing without pain.

Non-Invasive Blood Glucose Monitoring

[0004] Non-invasive glucose monitoring refers to the measurement of blood glucose levels without drawing blood, puncturing the skin, or causing pain or trauma. The search for a successful technique began about 40 year ago and has continued to the present without a clinically or commercially viable product. Until now, only one such blood glucometer had been approved for sale by the FDA, based on a technique for electrically pulling glucose through intact skin, which is described in Tamada et al. (1999), "Noninvasive glucose monitoring: comprehensive clinical results. Cygnus Research Team", JAMA, 1999, 282(19), 1839-44. This glucometer was however withdrawn after a short time owing to poor performance and occasional damage to the skin of users.

[0005] In general, non-invasive methods used for determination of glucose may fall into two major categories. The first is based on the measurement of glucose using one or more of its intrinsic molecular properties, such as NIR or mid-IR absorption coefficient, optical rotation, Raman shifts, and photoacoustic absorption, as well as others. These methods assume the ability to detect glucose in tissue or blood independently of other body components or physiological state. The second category measures the effects of glucose on the physical properties of blood and tissue. This category is based on an assumption that glucose is a dominant blood analyte (highly fluctuating) and, as such, contributes significantly to the change in the relevant physical parameters of the tissue. Khalil et al. (2004) in "Non-invasive glucose measurement technologies: an update from 1999 to the dawn of the new millennium", Diabetes Technol. Ther. 2004, 6(5), 660-697, demonstrated that measurement of such parameters can lead indirectly to evaluation of the blood glucose level. The measured parameters are evaluated relatively to calibration, performed through correlation of the non-invasive signal to a reference blood glucose value. Therefore, the relative change of glucose in blood or interstitial fluid plays the major role, as other blood analytes, which are less fluctuating, are fully or at least partially eliminated through calibration.

[0006] However, glucose determination in indirect and nonspecific non-invasive measurement faces several obstacles. Depending on the particular method used, readings may vary with changes in glucose level, but may also be affected by sensor-skin interface variations, changes in microcirculation and blood supply, medications that affect fluid distribution, comorbidities, a person's metabolic rate, and so on. The main concern, consequently, is to achieve high accuracy results and high sensitivity, despite the fact that no direct blood or interstitial fluid glucose measurement is performed.

[0007] Hundreds of millions of dollars have been invested in companies who have sought the solution to this long-standing problem. Approaches that have been tried include near infrared (NIR) spectroscopy (measuring glucose through the skin using light of slightly longer wavelengths than the visible region), transdermal measurement (attempting to pull glucose through the skin using either chemicals, electricity or ultrasound), measuring the amount that polarised light is rotated by glucose in the front chamber of the eye (containing the "aqueous humour"), and many others.

[0008] So et al. (2012), "Recent advances in non-invasive glucose monitoring" , Med. Devices, 2012, 5, 45-52, reviewed 10 technologies: bio impedance spectroscopy, electromagnetic sensing, fluorescence spectroscopy, Raman spectroscopy, NIR spectroscopy, mid-infrared spectroscopy, optical coherence tomography, optical polarimetry, reverse iontophoresis, and ultrasound technology, concluding with the observation that none of these had produced a commercially available, clinically reliable device and that therefore, much work remained to be done.

[0009] In 2014, disregarding some severe shortcomings mentioned above, the first non-invasive blood glucometer GlucoTrack ^® by Integrity Applications hit the market in a number of countries. Vashist (2013) however reported in his review "Continuous Glucose Monitoring Systems", Diagnostics, 2013, 3(4), 385-412, that the mean absolute deviation of this device was nearly 30% in clinical trials and 'further research efforts were desired to significantly improve the accuracy'.

[0010] Apple was the latest in a line of companies going back years that have tried to crack the challenge. The FDA approved a minimally invasive device, the GlucoWatch , back in 2001, only for the product to fail commercially. GlucoWatch ^® used the reverse iontophoresis to shuttle glucose across the skin for measurement. But, as the FDA noted, this meant its readings could 'differ substantially' and somewhat unpredictably from blood glucose levels. Studies linked the device to frequent false alarms. Calibration was time-consuming, and many users found the device uncomfortable. The size of the market meant the travails of GlucoWatch did little to deter interest in developing new noninvasive devices to measure glucose, but the field is still a long way from delivering the sort of sensors reportedly envisaged by Apple and discussed by many. In 2016, two startup companies planned to bring non-invasive products to market, but timelines seemed to have slipped.

Sub-THz Spectroscopy

[0011] Recently it has become clear that terahertz (THz) radiation could be extremely important for research related to the life sciences because of the unique capability of these low energy electromagnetic waves to interact with vibrations of atoms within biological molecules to produce specific molecular fingerprints (see for example, Globus et al. in "Terahertz Fourier transform characterization of biological materials in a liquid phase ", J. Physics D: Applied Physics, 39(15), 3405-3413). Sub-THz spectroscopy uses wavelengths beyond those traditionally used for chemical and biomolecular analysis. Biological materials have found to be active in the frequency range of 0.05-1 THz and above (the submillimetre-wavelength range, about 1.5 to 30 cm ^"1 ). These frequency and wavelength domains, the spectral range between the upper end of the radio frequencies and the lowest optical frequencies were named the 'Terahertz Gap', because so little was known about them and because of the absence of radiation sources and detectors.

[0012] Sub-THz vibrational spectroscopy is based on the interaction of THz radiation with internal molecular vibrations of low energy. A majority of the THz experimental data have recently been reported on frequencies above 1 THz and for relatively small biological molecules that are often prepared in crystalline form (for example, Heilweil et al. (2008), "Terahertz Spectroscopy of Biomolecules" , In Terahertz Spectroscopy, Taylor and Francis, London, 2008, Chapter 7, pp 269-297). Low energy THz radiation interacts with the low-frequency internal molecular motions (vibrations) involving the weakest hydrogen bonds (H-bonds) and other weak connections within molecules by exciting these vibrations. The width of individual spectral lines and the intensity of resonance features, which are observed in sub-THz spectroscopy, are very sensitive to the relaxation processes of atomic dynamics (displacements) within a molecule. Those relaxation processes determine the discriminative capabilities of sub-THz spectroscopy. Appropriate spectral resolution must be used in THz spectroscopy to be able to acquire qualitative as well as quantitative information used to identify the molecules that will, in turn, increase detection accuracy and selectivity.

[0013] Because of their small size and relatively low absorption coefficient, the waves of the THz radiation easily propagate through the entire biological object, such as cells and skin. Safrai et al. (2012) in "The remote sensing of mental stress from the electromagnetic reflection coefficient of human skin in the sub-THz range", Bioelectro- magnetics, 2012, 33(5), 375-82, and Safrai et al. (2014-1) in "Remote monitoring of phasic heart rate changes from the palm" in IEEE Transactions on Terahertz Science and Technology, 2014, 4, 618-624, reported that both physical and mental stress could be traced through the reflection coefficient of the hand, influenced by the activity of the sweat ducts, in the (75 GHz-110 GHz) and (110 GHz-170 GHz) frequency bands. The reflected signal was monitored from a distance of 72 cm using a Vector Network Analyser, while the patients' electrocardiograms (ECGs) were concurrently registered. Further, Safrai et al. (2014-2) in "The correlation of ECG parameters to the sub-THz reflection coefficient of human skin", IEEE Transactions on Terahertz Science and Technology, 2014, 4(5), 624-630, reported on a good correlation between the reflection coefficient in the same frequency bands and some of the parameters of the ECG, mainly to the ST elevation.

Sweat-Based Blood Glucose Sensing

[0014] Analysis of sweat using electrochemical devices on human skin provides a new route for non-invasive glucose monitoring without painful blood collection. Moyer et al. (2012) in "Correlation between sweat glucose and blood glucose in subjects with diabetes", Diabetes Technol Ther. 2012, 14(5), 398-402 showed that sweat glucose, when properly harvested to prevent contamination from other sources on the skin's surface, can accurately reflect blood glucose levels. Velmurugan et al. (2015) in "Sweat based blood glucose analysis", IJRASET 2015, 3(111), 550-555, reported that the determination of blood glucose from sweat is feasible and can deliver the correct blood glucose level, but includes certain constraints, such as processing of the microampere current obtained from the biosensor. In addition, they noted that only the sweat produced by the eccrine gland can be used for analysis. Velmurugan et al. (2015) demonstrated that the method can be used for continuous blood glucose monitoring where an overall average of the blood glucose is provided over a stretch of time. The described non-invasive method was proposed to implement in a wearable wrist band-like biosensor. It was concluded in the publication that if such technology could be brought into practice, it would save a great amount of pain for diabetic patients pricking their fingers 2-5 times a day, counting up to minimum of 1,820 lancets a year.

[0015] In light of the above, sweat patches are rapidly becoming a popular research development within diabetes. However, sweat-based glucose sensing still faces many challenges, such as difficulty in sweat collection, activity variation of glucose oxidase due to lactic acid secretion and ambient temperature changes and delamination of the enzyme from the glucose sensor when exposed to mechanical friction on soft human skin, deformation of the device and skin deformation causing mechanical fractures in the device. In addition to the difficulty of measuring glucose levels in sweat because of its much smaller concentration than that in blood, lactic acid in sweat, ambient temperature changes, and various medications can induce errors in enzyme -based glucose sensing.

Sweat Ducts as Helical Antennas

[0016] Hayut et al. (2013) in "The helical structure of sweat ducts: Their influence on the electromagnetic reflection spectrum of the skin", IEEE Transactions on Terahertz Science and Technology, 2013, Vol. 3, No. 2, pp. 2017-2015, suggested that the helical structure of human eccrine sweat ducts together with the dielectric properties of the human skin would result in their electromagnetic properties similar to those of an array of helical antennas. In order to test this assumption, they conducted numerical simulations in the sub-THz frequency range of 100-450 GHz, performed measurements of a reflection spectrum from the human skin and compared the obtained results to the simulation results. The spectral response obtained by their simulations coincided with the analytical prediction of antenna theory and supported the hypothesis that the sweat ducts can indeed be regarded as helical antennas. The magnitude of the spectral response was found to be dependent on the conductivity of sweat in these frequencies, but the analysis of the frequencies related to the antenna-like modes were found to be independent of this parameter. The performed simulations demonstrate that variations of the spectra are observed in the vicinity of the frequencies close to the predicted axial response mode of the sweat ducts (regarded as helical antennas) at approximately 380 GHz. The results clearly show that the structure of the sweat ducts plays a key role in the shaping of the reflected spectra. Thus, it has been established that the coiled nature of the human sweat duct can lead to electromagnetic behaviour reminiscent of a helical antenna.

SUMMARY

[0017] The present application describes embodiments of a microelectronic sensor based on a combination of an open-gate pseudo-conducting high-electron mobility transistor (PC-HEMT) and Vivaldi antenna. In some embodiments, the transistor comprises a substrate, on which a multilayer he teroj unction structure is deposited. This hetero-junction structure may comprise at least one buffer layer and at least one barrier layer grown from III-V single -crystalline or polycrystalline semiconductor materials and stacked alternately. In some embodiments, the ΙΠ-V single -crystalline or polycrystalline semiconductor materials are GaN/AlGaN.

[0018] A conducting channel comprising a two-dimensional electron gas (2DEG) or a two-dimensional hole gas (2DHG) is formed at the interface between the buffer and barrier layers and upon applying a bias to said transistor provides electron or hole current in the transistor between its source and drain electrodes. In a particular embodiment, the heterojunction structure may be a three-layer structure consisting of two GaN layers (one of which is a buffer layer) and one AlGaN layer squeezed between said GaN layers like in a sandwich. This may lead to formation of the two-dimensional electron gas (2DEG) or two-dimensional hole gas (2DHG) in the top layer dependent on the polarity of the structure.

[0019] An optional dielectric layer may be deposited on top of the heterojunction structure. The open gate area of the 2DEG/2DHG is formed between the source and drain areas as a result of recessing or growing of the top layer to a specific thickness.

[0020] The source and drain non-ohmic (i.e. capacitively-coupled) contacts are connected to the 2DEG/2DHG channel and to electrical metallizations, the latter are placed on top of the sensor and connect it to an electric circuit of the sensor. Since the source and drain contacts are non-ohmic, the DC readout cannot be carried out. Instead, in order to electrically contact the 2DEG/2DHG channel underneath, about 5-20 nm bellow the metallizations, the AC readout or impedance measurements of the electric current flowing through the 2DEG/2DHG-channel must be performed. In this case, the capacitive coupling of the non-ohmic metal contacts with the 2DEG/2DHG channel is normally induced at the frequency higher than 30 kHz.

[0021] In some embodiments, the PC-HEMT multilayer heterojunction structure of the embodiments is grown from any available ΙΠ-V single-crystalline or polycrystalline semiconductor materials, such as GaN/ AlGaN, GaN/AIN, GaN/InN, GaN/InAlGaN, GaAs/AlGaAs, GaN/InAIN, InN/InAIN or LaA10 ₃ /SrTi0 ₃ . In a particular case of the substrate grown from GaN/AlGaN, it has been experimentally and surprisingly found that the highest sensitivity of the sensor is achieved when thickness of the top recessed layer (GaN or AlGaN) in the open gate area between the source and drain contacts is 5-9 nm, preferably 6-7 nm, more preferably 6.2-6.4 nm. This recessed layer thickness corresponds to the pseudo -conducting current range between normally-on and normally-off operation mode of the 2DEG/2DHG conducting channel. In addition, surface roughness of the top recessed layer within the open gate area between the source and drain contacts has a roughness of about 0.2 nm or less, preferably 0.1 nm or less, more preferably 0.05 nm. Thus, the significant features of the PC-HEMT are that:

(i) thickness of the top recessed layer (GaN or AlGaN) of the multilayer heteroj unction structure in the open gate area between the source and drain contacts is 5-9 nm, preferably 6-7 nm, more preferably 6.2-6.4 nm, which corresponds to the pseudo- conducting current range between normally-on and normally-off operation mode of the 2DEG/2DHG conducting channel;

(ii) surface roughness of the top recessed layer in the open gate area between the source and drain contacts is about 0.2 nm or less, preferably 0.1 nm or less, more preferably 0.05 nm; and

(iii) the non-ohmic source and drain contacts for the capacitive coupling with the conductive 2DEG/2DHG channel replace the ohmic contacts.

[0022] In a further embodiment, the present application provides the sensing device suitable for use in blood glucose monitoring.

[0023] Various embodiments may allow various benefits, and may be used in conjunction with various applications. The details of one or more embodiments are set forth in the accompanying figures and the description below. Other features, objects and advantages of the described techniques will be apparent from the description and drawings and from the claims

BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Disclosed embodiments will be understood and appreciated more fully from the following detailed description taken in conjunction with the appended figures. The drawings included and described herein are schematic and are not limiting the scope of the disclosure. It is also noted that in the drawings, the size of some elements may be exaggerated and, therefore, not drawn to scale for illustrative purposes. The dimensions and the relative dimensions do not necessarily correspond to actual reductions to practice of the disclosure. Fig. 1 schematically shows the quantum well at three different biasing conditions:

Fig. la: positive gate potential (+VG) is much higher than threshold voltage (V _T ), Fig. lb: 0V gate potential, and

Fig. lc: negative gate potential (-V _G ) is below threshold voltage (V _T ).

Fig. 2 schematically shows a cross-sectional (XZ) view (A-A) of the PC -HEMT of an embodiment.

Fig. 3 schematically shows the top (XY) view and the basic topology of the sensor of an embodiment.

Fig. 4a schematically shows the dependence of the source-drain current (a charge carrier density) induced inside the 2DEG channel of a GaN/AlGaN HEMT on the thickness of the AlGaN layer recessed in the open gate area.

Fig. 4b illustrates a theory behind the 2DEG formation (charge neutrality combined with the lowest energy level) at the conduction band discontinuity.

Fig. 5a schematically shows the 2DEG area created in the step of the 2DEG-pattering via ion implantation during the manufacturing process. AZ 4533 is a positive thick resist. Fig. 5b shows the lithographic mask of the sensor layout of the present invention.

Fig. 5c shows the lithographic image of the 2DEG channel formed with AZ 4533 thick resist lithography over the mask shown in Fig. 5b.

Figs. 5d-5e show the mask and the corresponding lithographic image, respectively, of the sensor layout of the present invention.

Fig. 5f shows the ± 2-μπι alignment precision on 25 x 25 mm samples in the lithography of the sensor layout of the present invention.

Fig. 5g shows the lithographic images of the multichannel samples.

Fig. 5h shows the fixed sample on the Si-GaN/AlGaN wafer prepared for ion implantation and containing around 30-32 sensors with 4-8 channels on each sample. Fig. 5i shows the lithographic image of the sensor layout with the AZ4533 resist after development, prepared for ion implantation.

Fig. 5j shows the 2DEG channels (dark) patterned by ion-implantation after the resist removal.

Fig. 5k shows the visible non-implanted area containing the conductive 2DEG channel. Fig. 6a shows the AFM surface image of the top recessed layer of the PC-HEMT made by the manufacturing process of the present invention. The measured RMS value of the surface roughness is 0.674 nm in this case.

Fig. 6b shows the AFM surface image of the top recessed layer of the HEMT made by a conventional manufacturing process. The measured RMS value of the surface roughness is 1.211 nm in this case.

Fig. 6c shows the time -dependent plot of the drain-source electric current I _DS of the nitrogen oxide sensor measuring 100 ppb of the N0 ₂ gas in humid air, where the sensor is based on the PC-HEMT made by the manufacturing process of the present invention. Fig. 6d shows the time -dependent plot of the drain-source electric current I _DS of the nitrogen oxide sensor measuring 100 ppb of the N0 ₂ gas in humid air, where the sensor is based on the HEMT made by a conventional manufacturing process.

Fig. 7a schematically shows the formation of the 2DEG and 2DHG conducting channels in the Ga-face three-layer Ga/AlGaN/GaN PC-HEMT structure.

Fig. 7b schematically shows the formation of the 2DEG and 2DHG conducting channels in the N-face three-layer Ga/AlGaN/GaN PC-HEMT structure.

Fig. 8 schematically shows the formation of the 2DEG conducting channel in the N-face three-layer GaN/AlG aN/GaN PC-HEMT structure with an ultrathin Al(GaN)N layer for improved confinement.

Fig. 9 shows the model of a Vivaldi antenna realised on a thin dielectric substrate. An exponential function is used for the taper profile. The entire domain is bounded by a perfectly matched layer.

Fig. 10 shows the electric field distribution in the Vivaldi antenna plane at 480 GHz. Fig. 11 shows (a) the far- field directional radiation pattern of the Vivaldi antenna starting from 240 GHz till 780 GHz and (b) the corresponding 3D far-field pattern at 480 GHz. Fig. 12 shows the exemplary two-dimensional photonic crystals (a = 160 μπι, d = 60 μπι) deposited on top of the metal layer of the Vivaldi antenna.

Fig. 13 shows the position of the metal connector on the Vivaldi antenna for coupling with the PC-HEMT of an embodiment. Fig. 14 schematically shows a microelectronic sensor comprising a single PC-HEMT of the embodiments with the integrated Vivaldi antenna, for non-invasive monitoring of glucose levels in blood, with a remote readout.

Fig. 15 schematically shows a microelectronic sensor comprising an array of the PC- HEMTs of the embodiments with the integrated Vivaldi antennas, for non-invasive monitoring of glucose levels in blood, with a remote readout.

DETAILED DESCRIPTION

[0025] In the following description, various aspects of the present application will be described. For purposes of explanation, specific configurations and details are set forth in order to provide a thorough understanding of the present application. However, it will also be apparent to one skilled in the art that the present application may be practiced without the specific details presented herein. Furthermore, well-known features may be omitted or simplified in order not to obscure the present application.

[0026] The term "comprising", used in the claims, is "open ended" and means the elements recited, or their equivalent in structure or function, plus any other element or elements which are not recited. It should not be interpreted as being restricted to the means listed thereafter; it does not exclude other elements or steps. It needs to be interpreted as specifying the presence of the stated features, integers, steps or components as referred to, but does not preclude the presence or addition of one or more other features, integers, steps or components, or groups thereof. Thus, the scope of the expression "a device comprising x and z" should not be limited to devices consisting only of components x and z. Also, the scope of the expression "a method comprising the steps x and z" should not be limited to methods consisting only of these steps.

[0027] Unless specifically stated, as used herein, the term "about" is understood as within a range of normal tolerance in the art, for example within two standard deviations of the mean. In one embodiment, the term "about" means within 10% of the reported numerical value of the number with which it is being used, preferably within 5% of the reported numerical value. For example, the term "about" can be immediately understood as within 10%, 9%, 8%, 7%, 6%, 5%, 4%, 3%, 2%, 1%, 0.5%, 0.1%, 0.05%, or 0.01 % of the stated value. In other embodiments, the term "about" can mean a higher tolerance of variation depending on for instance the experimental technique used. Said variations of a specified value are understood by the skilled person and are within the context of the present invention. As an illustration, a numerical range of "about 1 to about 5" should be interpreted to include not only the explicitly recited values of about 1 to about 5, but also include individual values and sub-ranges within the indicated range. Thus, included in this numerical range are individual values such as 2, 3, and 4 and sub-ranges, for example from 1-3, from 2-4, and from 3-5, as well as 1, 2, 3, 4, 5, or 6, individually. This same principle applies to ranges reciting only one numerical value as a minimum or a maximum. Unless otherwise clear from context, all numerical values provided herein are modified by the term "about". Other similar terms, such as "substantially", "generally", "up to " and the like are to be construed as modifying a term or value such that it is not an absolute. Such terms will be defined by the circumstances and the terms that they modify as those terms are understood by those of skilled in the art. This includes, at very least, the degree of expected experimental error, technical error and instrumental error for a given experiment, technique or an instrument used to measure a value.

[0028] As used herein, the term "and/or" includes any and all combinations of one or more of the associated listed items. Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. It will be further understood that terms, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the specification and relevant art and should not be interpreted in an idealized or overly formal sense unless expressly so defined herein. Well-known functions or constructions may not be described in detail for brevity and/or clarity.

[0029] It will be understood that when an element is referred to as being "on", "attached to", "connected to", "coupled with", "contacting", etc., another element, it can be directly on, attached to, connected to, coupled with or contacting the other element or intervening elements may also be present. In contrast, when an element is referred to as being, for example, "directly on", "directly attached to", "directly connected to", "directly coupled" with or "directly contacting" another element, there are no intervening elements present. It will also be appreciated by those of skill in the art that references to a structure or feature that is disposed "adjacent" another feature may have portions that overlap or underlie the adjacent feature.

[0030] The polarization doped high-electron-mobility transistor (HEMT) is a field effect transistor (FET) in which two layers of different bandgap and polarisation field are grown upon each other forming the heterojunction structure. This transistor is essentially based on at least two layers of III-V semiconductor materials, such as gallium nitride (GaN) and aluminium gallium nitride (AlGaN). As a consequence of the discontinuity in the polarisation field, surface charges are created at the interface between the layers. If the induced surface charge is positive, electrons will tend to compensate the induced charge resulting in the formation of the channel. Since the channel electrons are confined in a quantum well in an infinitely narrow spatial region at the interface between the layers, these electrons are referred to as a two-dimensional electron gas (2DEG). This special confinement of the channel electrons in the quantum well actually grants them two-dimensional features, which strongly enhance their mobility surpassing the bulk mobility of the material in which the electrons are flowing.

[0031] Figs, la-lc schematically shows the quantum well at three different biasing conditions starting from the positive gate potential (V _G ), much higher than the threshold voltage (V _T ), and going down to the 0V gate potential and further to the negative values below the threshold voltage. The V _T is defined as a voltage, which is required to populate electrons at the interface between the GaN layer and the AlGaN layers, thereby creating conductivity of the 2DEG channel. Since the 2DEG channel electrons occupy energy levels below the Fermi level, the Fermi level in a quantum well is located above several energy levels when V _G >> V _T (Fig. la). This enables high population of channel electrons and consequently, high conductivity. The 2DEG channel is turned on in this case. However, when V _G decreases to 0V (Fig. lb), the Fermi level also drops with respect to the quantum well. As a result, much fewer electron energy levels are populated and the amount of the 2DEG channel electrons significantly decreases. When V _G much less than V _T (Fig. lc), all electron energy levels are above the Fermi level, and there is no the 2DEG electrons below the gate. This situation is called "channel depletion", and the channel is turned off. [0032] Many commercially available HEMTs based on the layers of III-V semiconductor materials have a negative value of VT, resulting in a "normally-on" operation mode at 0V gate potential. They are called "depletion-mode" semiconductor transistors and used in various power switching applications when the negative voltage must be applied on the gate in order to block the current. However, for safe operation at high voltage or high power density, in order to reduce the circuit complexity and eliminate standby power consumption, the transistors with "normally-off" characteristics are preferred. The high voltages and high switching speeds allow smaller, more efficient devices, such as home appliances, communications and automobiles to be manufactured. To control the density of electrons in the 2DEG channel and to switch the HEMT on and off, the voltage at the gate of the transistor is normally regulated.

[0033] Several techniques to manufacture the normally-off semiconductor structures have been reported. Burnham et al (2010) proposed normally-off structures of the recessed gate type. In this structure, the AlGaN barrier layer is etched and the gate is brought closer to the interface between the AlGaN barrier layer and the GaN buffer layer. As the gate approaches the interface between the layers, the VT increases. Thus, the normally-off operation of the 2DEG conducting channel is achieved once the depletion region reaches the interface and depletes the 2DEG channel at zero gate voltage. The major advantages of these structures are relatively lower power consumption, lower noise and simpler drive circuits. They are currently used, for example, in microwave and millimetre wave communications, imaging and radars.

[0034] Chang et al (2009) proposed instead of etching the relatively thick barrier layer to approach the AlGaN/GaN interface, to use a very thin AlGaN barrier. This structure also achieves the normally-off operation of the 2DEG channel by approaching the gate towards the AlGaN/GaN interface. Chen et al (2010) proposed to use the fluorine -based plasma treatment method. Although many publications have adopted various methods to achieve normally-off devices with minimum impact on the drain current, they unfortunately sacrificed device turn-on performance.

[0035] The present application describes embodiments of a microelectronic sensor based on a combination of an open-gate pseudo-conducting high-electron mobility transistor (PC-HEMT) and Vivaldi antenna installed in the open gate area of the transistor. As shown in Fig. 2, the PC-HEMT of the present application, capable of sensing sub-THz radiation produced by a body of the user, comprises:

(a) a multilayer heterojunction structure composed of III-V single-crystalline or poly- crystalline semiconductor materials, said structure comprising at least one buffer layer (11 ) and at least one barrier layer (12), said layers being stacked alternately, and said structure being deposited on a substrate layer (10);

(b) a conducting channel (13) comprising a two-dimensional electron gas (2DEG) or a two-dimensional hole gas (2DHG), formed at the interface between said buffer layer (1 1) and said barrier layer (12), and upon applying a bias to said transistor, capable of providing electron or hole current in the transistor between source and drain contact areas (15);

(c) electrical metallizations (14) capacitively-coupled to said 2DEG or 2DHG conducting channel (13) via the source and drain contact areas (15) for connecting said transistor to an electric circuit; and

(d) a Vivaldi antenna electrode (16) placed on a top layer between said source and drain contact areas (15) in an open gate area of the transistor and capable of detecting electrical signals in the sub-THz-frequency range of 200-800 GHz.

[0036] The functional basic topology of the microelectronic sensor of the present embodiments is schematically shown in Fig. 3. The sensor detection principle is based on the field effect of electric current modulation in a DC-mode within the 2DEG/2DHG conducting channel (13) achieved by Vivaldi-shaped gate antenna (16). In fact, the Vivaldi-shaped gate antenna (16) is capable of strongly concentrating the electric field strength of sub-THz radiation in a very small area above the ultra charge-sensitive 2DEG/2DHG channel, thereby affecting its conductivity in the DC-mode by accumulation effect. The S 11-S12 parameters of the PC-HEMT of the present embodiments however may also be measured at radio frequencies (RF) of 1 to 60 GHz using the 'beating' effect of sub-THz radiation.

[0037] The transistor shown in Figs. 2 and 3 may further comprise a dielectric layer of 1-10 nm thickness. This dielectric layer may be deposited on top of the barrier layer (12). The Vivaldi antenna gate electrode, made for example from gold, is then placed directly on the dielectric layer. This configuration prevents strong electrical leakage at the metal/top layer interface. The dielectric layer used for the device passivation, may be made, for example, of SiO-SiN-SiO ("ONO") stack of 100-100-100 nm thickness or SiN- SiO-SiN ("NON") stack having the same thicknesses. It may be deposited on top of the barrier layer by a method of plasma-enhanced chemical vapour deposition (PECVD), which is a stress-free deposition technique.

[0038] The electrical metallizations (14) connect the transistor to the electric circuit and allow the electric current to flow between non-ohmic contact areas (15) via the two- dimensional electron gas (2DEG) or two-dimensional hole gas (2DHG) channel (13). The metallizations (14) may be made of metal stacks, such as Cr/Au, Ti/Au, Ti/W, Cr/Al and Ti/Al. The Cr or Ti layers of the metal stack is, for example, of 5-10 nm thickness, while the second metal layer, such as Au, W and Al, is of 100-400 nm thickness. The metallizations (14) may be chosen according to the established technology and assembly line at a particular clean room fabrication facility.

[0039] In some embodiments, substrate layer (10) may be composed, for example, of sapphire, silicon, silicon carbide, gallium nitride or aluminium nitride. The hetero- junction structure (11 , 12) may be deposited on the substrate layer (10), for example, by a method of metalorganic chemical vapour deposition (MOCVD), thereby forming the pseudo-conducting 2DEG or 2DHG channel (13) in the close proximity to the interface between the buffer layer (11 ) and the barrier layer (12). The barrier layer (12) then may be either recessed or grown in the recessed area (17) as a very thin layer between the source and drain contact areas (15).

[0040] The 2DEG or 2DHG channel (13) formed near the interface between the buffer layer (11) and the barrier layer (12) serves as a main sensitive element of the transistor reacting to a surface charge and potential. The 2DEG or 2DHG channel (13) is configured to interact with very small variations in surface or proximal charge or changes of electrical field on the top layer/Vivaldi antenna gate interface interacting with the donor-like surface trap states of the barrier layer. This will be discussed below in detail.

[0041] The term "2DEG" mentioned in the present description and claims should not be understood or interpreted as being restricted to the two-dimensional electron gas. As stated above and will be explained later in this application, the two-dimensional hole gas may also be a possible current carrier in a specific heterojunction structure. Therefore, the term "2DEG" may be equally replaced with the term "2DHG" without reference to any particular PC-HEMT configuration.

[0042] In some embodiments, the PC-HEMT multilayer heterojunction structure of the embodiments is grown from any available III-V single-crystalline or polycrystalline semiconductor materials, such as GaN/AlGaN, GaN/AIN, GaN/InAlGaN, GaN/InAIN, GaN/InN, GaAs/AlGaAs, InN/InAIN or LaA10 ₃ /SrTi0 ₃ . In a specific case of the heterojunction structure grown from GaN/AlGaN, it has been experimentally found that the highest sensitivity of the sensor is achieved when thickness of the top recessed layer (GaN or AlGaN) in the open gate area between the source and drain contacts is 5-9 nm, preferably 6-7 nm, more preferably 6.2-6.4 nm. In addition, it was also found that the sensor exhibits its highest sensitivity when surface roughness of the top recessed layer is about 0.2 nm or less, preferably 0.1 nm or less, more preferably 0.05 nm.

[0043] Thus, the top layer recessed in the open gate area to 5-9 nm must be optimised for significantly enhancing sensitivity of the sensor. This specific thickness of the top layer was surprisingly found to correspond to the "pseudo-conducting" current range between normally-on and normally-off operation modes of the 2DEG channel and requires further explanation.

[0044] "Pseudo-contacting" (to distinguish from normally-conducting) current range of the 2DEG channel is defined as an operation range of the channel between its normally-on and normally-off operation modes. "Trap states" are states in the band-gap of a semiconductor which trap a carrier until it recombines. "Surface states" are states caused by surface reconstruction of the local crystal due to surface tension caused by some crystal defects, dislocations, or the presence of impurities. Such surface reconstruction often creates "surface trap states" corresponding to a surface recombination velocity.

[0045] Classification of the surface trap states depends on the relative position of their energy level inside the band gap. The surface trap states with energy above the Fermi level are acceptor-like, attaining negative charge when occupied. However, the surface trap states with energy below the Fermi level are donor-like, positively charged when empty and neutral when occupied. These donor-like surface trap states are considered to be the source of electrons in the formation of the 2DEG channel. They may possess a wide distribution of ionization energies within the band gap and are caused by redox reactions, dangling bonds and vacancies in the surface layer. A balance always exists between the 2DEG channel density and the number of ionised surface donors which is governed by charge neutrality and continuity of the electric field at the interfaces.

[0046] Thus, the donor-like surface traps at the surface of the top layer are one of the most important sources of the 2DEG in the channel. However, this only applies for a specific top layer thickness. In a relatively thin top layer, the surface trap state is below the Fermi level. However, as the top layer thickness increases, the energy of the surface trap state approaches the Fermi energy until it coincides with it. The thickness of the top layer corresponding to such situation is defined as "critical". At this point, electrons filling the surface trap state are pulled to the channel by the strong polarisation-induced electric field found in the top layer to form the 2DEG instantly.

[0047] If the surface trap states are completely depleted, further increase in the top layer thickness will not increase the 2DEG density. Actually, if the 2DEG channel layer fails to stretch the top layer, the later will simply relax. Upon relaxation of the top layer, many crystal defects are created at the interface between the top layer and the layer right underneath it, and the piezoelectric polarisation instantly disappears causing deterioration in the 2DEG density.

[0048] In order to illustrate the above phenomenon of the pseudo-conducting current, reference is now made to the following figures. Fig. 4a shows the dependence of the source -drain current (a charge carrier density) on the recessed AlGaN layer thickness. As seen from the plot, transistors that have a thickness of the top layer larger than 9 nm form normally-on 2DEG channels. In such transistors, due to the inherent polarisation effects present in the III-V materials, a thin sheet of charges is induced at the top and bottom of the interfaces of the top layer. As a result, a high electric field is induced in the top layer, and surface donor states at the top interface start donating electrons to form the 2DEG channel at the proximity of the hetero-junction interface without the application of a gate bias. These transistors therefore constitute normally-on devices. On the other hand, the transistors that have a thickness of the top layer lower than about 5 nm act constitute normally-off devices. Energy equilibrium between the donor surface trap states and AlGaN tunnel barrier leads to the 2DEG formation (charge neutrality combined with the lowest energy level) at the conduction band discontinuity. As explained above, decrease in the thickness of the AlGaN layer results in increase of the energy barrier. As a result, the ionisable donor-like surface trap states, which are responsible for electron tunnelling from the surface to 2DEG, drift bellow the Fermi level, thereby minimizing the electron supply to the 2DEG channel. This theoretical situation is further illustrated in Fig. 4b. Therefore, the recess of the AlGaN layer from 9 nm to 5 nm leads to huge drop in conductivity of the two-dimensional electron gas for six orders of magnitude.

[0049] In view of the above, it is clear that the mechanism of the 2DEG depletion based on recessing the top layer is strongly dependent on the donor-like surface trap states (or total surface charge). As the thickness of the top layer decreases, less additional external charge is needed to apply to the top layer surface in order to deplete the 2DEG channel. There is a critical (smallest) barrier thickness, when the 2DEG channel is mostly depleted but still highly conductive due to a combination of the energy barrier and the donor surface trap states energy. At this critical thickness, even the smallest energy shift at the surface via any external influence, for example an acoustic wave propagating along the surface, leads immediately to the very strong 2DEG depletion. As a result, the surface of the top layer at this critical thickness is extremely sensitive to any smallest change in the electrical field of the surroundings. Thus, the recess of the top layer from 9 nm down to 5 nm significantly reduced the 2DEG density, brought the sensor to the "near threshold" operation and resulted in highly increased surface charge sensitivity. The specific 5-9 nm thickness of the top layer is actually responsible for the pseudo- conducting behaviour of the 2DEG channel and gives the sensor an incredible sensitivity.

[0050] The top layer is recessed to this specific thickness after subjecting to short plasma activation by an ultra-low damage reactive-ion etching technique using inductively-coupled plasma (ICP) with a narrow plasma-ion energy distribution. Such short plasma treatment allows much lower roughness of the surface, which is a function of the semiconductor vertical damage depth during the plasma etching process. Such low surface roughness (about 0.2 nm and less) can be achieved only via this ICP-RIE ultra low damage etching process with a narrow plasma-ion energy distribution, and this inherently results in a very low vertical damage depth to the top layer, which allows the minimal surface scattering and minimal surface states-2DEG channel interaction with the maximum signal-to-noise ratio of the sensor. Thus, the depth effect of the vertical sub- nanometre damage to the top recessed layer, due to an ultra-low damage ICP-RIE etching process with a very narrow plasma-ion energy distribution, is the only way to optimally achieve the required sub-nanometre roughness of the semiconductor surface. This inherently results in an adjustable pseudo-conductive working point with the highest charge sensitivity ever possible. This depth effect is always inherent to the sub-nanometre roughness of the semiconductor surface, which was measured using AFM (atomic force microscope).

[0051] Thus, in addition to the recessed top layer thickness, roughness of the top layer surface is another very important parameter that has not been previously disclosed. It has been surprisingly found that the roughness of the top layer surface (in the open gate sensitive area) bellow 0.2 nm prevents scattering of the donor-like surface trap states. Thus, combination of these two features: 5-9 nm thickness of the top layer in the open gate area and strongly reduced roughness of its surface (bellow 0.2 nm) make the sensor incredibly sensitive.

[0052] In a certain aspect, the method for manufacturing of the PC-HEMTs of the present invention comprises the following steps:

Step 1: Plasma-enhanced atomic layer deposition (ALD) of alumina (AI2O ₃ ) on a pre- aligned masked Si-GaN/AlGaN wafer with nitrogen-plasma de-trapping for the thickness of the AI2O ₃ layer being 3-10 nm. The AI2O ₃ layer thickness was measured with an X-ray reflectometer.

Step 2: Plasma-enhanced atomic layer deposition (ALD) pattering of the wafer coated with the thin AI2O ₃ layer in Step 1 , with hydrogen fluoride (HF) or using the aforementioned reactive-ion etching (RIE) technique.

Step 3: Optionally creating the source and drain ohmic contacts (in case ohmic contacts are required) on the coated wafer obtained in Step 2 from metal stacks, for example Ti/Al/Mo/Au, Ti/Al/Ni/Au, Ti/Au and Ti/W, having 15-50 nm thickness, using spin-coating technique or e-beam physical vapour deposition (VPD) of the stack metals. The deposition rates using the e-VPD technique were determined for the ohmic-stack metals using the Dektak Profilometer with dummy lift-off samples.

Step 4: Two-dimensional electron gas (2DEG) channel-pattering of the wafer obtained in

Step 3 with argon- or nitrogen-ion implantation.

Step 5: Plasma-enhanced chemical vapour deposition (CVD) of the ONO stack over the wafer obtained in Step 4. This is the stress-free technique to deposit the layer of the SiO-SiN-SiO stack having an exemplary thickness of about 200-300 nm and structured by the ICP-RIE dry etching, which is the CF4-based etching method.

In this step, the pseudo-conducting channel areas and ohmic electrical contact pads of the transistor become available.

Step 6: Optional lift-off deposition of an Au or Ti/W-CMOS-gate electrode (in case a gate electrode is to be deposited on the top layer of the heterojunction structure for an integrated MMIC-HEMT -based amplifier manufacturing).

Step 7: Optional plasma-enhanced ALD pattering with RIE or HF above sensing area (in case the plasma-enhanced ALD layer deposited in Step 1 is removed separately to ONO stack).

Step 8: Atomic layer etching (ALE) of the wafer obtained in Steps 5-7. This sophisticated technique carried out in the clean manufacturing cluster of the applicant is the only technique allowing the removal of individual atomic layers (the top atomic layers of the wafer). ALE is a way better-controlled technique than RIE, though it has not been commercially used until now because very sophisticated gas handling is required, and removal rates of one atomic layer per second are the real state of the art. This step is the step of creating the pseudo- conducting working point of the transistor, because ALE allows achieving the specific thickness of 5-9 nm thickness of the top layer in the open gate area with the extremely low surface roughness of the top layer below 0.2 nm.

Step 9: Optional plasma-enhanced CVD or ALD of the dielectric layer used for device passivation and in some gas sensors.

Step 10: Optional deep reactive -ion etching (DRIE or Bosch process) of the Si-substrate under sensing areas (in case the substrate is on the free-standing membranes - used, for example, in RF-HEMTs, FBAR and SAW sensors). [0053] Reference is now made to Figs. 5a-5c showing the sensor, which is obtained in Step 4 of the 2DEG-channel pattering. The lithography of the sensor was performed with AZ 4533, which is a positive thick resist having optimised adhesion for common wet etching. The lithographic resist film thickness obtained at 7000-rpm spin speed and at 100° C for 1 min was 3 μπι. Thus, as seen in the lithographic image of Fig. 5c, the formed 2DEG channel (13) is approximately 2-3 μπι wide. The overall exposure time was 9 sec, followed by 5 -min development in MIF726 developer.

[0054] Fig. 5d-5e show the mask and corresponding lithographic image, respectively, of the sensor layout of the present invention. Fig. 5f demonstrates the high alignment precision of ± 2-μπι on 25 x 25 mm samples in the lithography of the sensor layout of the present invention. Fig. 5g shows the lithographic images of the multichannel samples. Fig. 5h shows the fixed sensor chip sample on the Si-GaN/AlGaN wafer, which contains approximately 30-32 sensors with 4-8 channels on each sample and prepared for ion implantation. Fig. 5i shows the obtained lithographic image of the present sensor layout with the AZ4533 resist after development, prepared for ion implantation. Fig. 5j shows the 2DEG channels (dark) patterned by ion -implantation after the resist removal. The argon-ion implantation was conducted with 20 keV and 30 keV energies and with an exemplary dose of 2.5e ¹³ /cm ² and a 7° tilt angle. AZ4533 was removed with oxygen plasma at 220 W for 10 min. Fig. 5k shows the visible non-implanted area containing the conductive 2DEG channel.

[0055] The atomic layer etching (ALE) performed in Step 8 of the manufacturing process is the most important stage in the process. As mentioned above, it allows the controlled recess of a top layer, removing a single atomic layer-by-layer, where the etch thickness is in the order of magnitude of a single atomic monolayer. As explained above, such ultra-low damage to the top layer of the heterogeneous structure, when the actual surface roughness is controlled by a single atomic monolayer, allows to achieve the sub- nanometre roughness (about 0.2 nm and less) of the top layer when its thickness is only few nanometres (5-9 nm).

[0056] The ALE process sequence consists of repeated cycling of process conditions. The total amount of material removed is determined by the number of repeated cycles. Each cycle is typically comprised of four steps: adsorption, first purge, desorption and second purge. During the adsorption step of the cycle, reactive species are generated in the reactor (for example, upon plasma excitation), adsorbed by, and react with material on the wafer. Due to the self-limiting process, and with the proper choice of reactants and process conditions, reaction takes place with only a thin layer of material, and the reaction by-products are formed. This step is followed by purging of the reactor to remove all traces of the reactant. Then the by-product desorption takes place due to bombardment of the wafer surface by noble gas ions with a tightly controlled energy. Again, by-products are purged from the reactor, and the wafer is ready for the last two (optional) steps of the manufacturing process.

[0057] Reference is now made to Fig. 6a showing the AFM image of the top recessed layer surface of the PC-HEMT produced by the manufacturing process of the present invention. The measured RMS value of the surface roughness is 0.674 nm in this case. Fig. 6b shows the AFM surface image of the top recessed layer of the HEMT made by a conventional manufacturing process. In this conventional process, the HEMT initially had a top ultrathin-grown AlGaN layer of the 6-7 nm thickness. This layer was recessed with inductively-coupled plasma (ICP) for 60 sec using a conventional reactive-ion etching (RIE) technique. The measured RMS value of the surface roughness is 1.211 nm in this case. Figs. 6c show the time-dependent plot of the drain-source electric current I _DS of the nitrogen oxide sensor measuring 100 ppb of the N0 ₂ gas in 80%-humid air, where the sensor incorporates the PC-HEMT made by the manufacturing process of the present invention. Figs. 6d show the time -dependent plot of the I _DS of the nitrogen oxide sensor measuring 100 ppb of the N0 ₂ gas in 80%-humid air, where the sensor incorporates and based on the HEMT made by the conventional manufacturing process. It is clear from these comparative examples that the manufacturing process of the present invention based on the ultra-low damaging RIE with a narrow plasma-ion energy distribution leads to much lower roughness of the semiconductor surface, which in turn leads to incredibly high sensitivity of the sensor.

[0058] In a further aspect, the hetero-junction structure may be a three-layer structure consisting of two GaN layers and one AlGaN layer squeezed between said buffer layers like in a sandwich, wherein the top layer is a buffer layer. This may lead to formation of the two-dimensional hole gas (2DHG) in the top GaN layer above the AlGaN layer which results in reversing polarity of the transistor compared to the two-layer structure discussed above.

[0059] In general, polarity of III-V nitride semiconductor materials strongly affects the performance of the transistors based on these semiconductors. The quality of the wurtzite GaN materials can be varied by their polarity, because both the incorporation of impurities and the formation of defects are related to the growth mechanism, which in turn depends on surface polarity. The occurrence of the 2DEG/2DHG and the optical properties of the hetero-junction structures of nitride-based materials are influenced by the internal field effects caused by spontaneous and piezo-electric polarizations. Devices in all of the III-V nitride materials are fabricated on polar {0001 } surfaces. Consequently, their characteristics depend on whether the GaN layers exhibit Ga-face positive polarity or N-face negative polarity. In other words, as a result of the wurtzite GaN materials polarity, any GaN layer has two surfaces with different polarities, a Ga-polar surface and an N-polar surface. A Ga-polar surface is defined herein as a surface terminating on a layer of Ga atoms, each of which has one unoccupied bond normal to the surface. Each surface Ga atom is bonded to three N atoms in the direction away from the surface. In contrast, an N-polar surface is defined as a surface terminating on a layer of N atoms, each of which has one unoccupied bond normal to the surface. Each surface N atom is also bonded to three Ga atoms in the direction away from the surface. Thus, the N-face polarity structures have the reverse polarity to the Ga-face polarity structures.

[0060] As described above for the two-layer heterojunction structure, the barrier layer is always placed on top of the buffer layer. The layer which is therefore recessed in the two-layer heterojunction structure is the barrier layer, specifically the AlGaN layer. As a result, since the 2DEG is used as the conducting channel and this conducting channel is located slightly below the barrier layer (in a thicker region of the GaN buffer layer), the hetero-junction structure is grown along the { 0001 } -direction or, in other words, with the Ga-face polarity. However, as explained above, the physical mechanism that leads to the formation of the 2DEG is a polarisation discontinuity at the AlGaN/GaN interface, reflected by the formation of the polarisation-induced fixed interface charges that attract free carriers to form a two-dimensional carrier gas. It is a positive polarisation charge at the AlGaN/GaN interface that attracts electrons to form 2DEG in the GaN layer slightly below this interface.

[0061] As noted above, polarity of the interface charges depends on the crystal lattice orientation of the hetero-junction structure, i.e. Ga-face versus N-face polarity, and the position of the respective AlGaN/GaN interface in the hetero-junction structure (above or below the interface). Therefore, different types of the accumulated carriers can be present in the hetero-junction structure of the embodiments.

[0062] In case of the three -layer hetero-junction structure, there are four possible configurations:

Ga-face polarity

1) The Ga-face polarity is characterised by the 2DEG formation in the GaN layer below the AlGaN barrier layer. This is actually the same two-layer configuration as described above, but with addition of the top GaN layer. In this configuration, the AlGaN barrier layer and two GaN layers must be nominally undoped or n-type doped.

2) In another Ga-face configuration shown in Fig. 7a, in order to form the conducting channel comprising a two-dimensional hole gas (2DHG) in the top GaN layer above the AlGaN barrier layer in the configuration, the AlGaN barrier layer should be p-type doped (for example, with Mg or Be as an acceptor) and the GaN buffer layer should be also p-type doped with Mg, Be or intrinsic.

N-face polarity

3) The N-face polarity is characterised by the 2DEG formation in the top GaN layer above the AlGaN barrier layer, as shown in Fig. 7b. In this case, the AlGaN barrier layer and two GaN buffer layers must be nominally undoped or n-type doped.

4) The last configuration assumes that the 2DHG conducting channel is formed in the buffer GaN layer below the AlGaN barrier layer. The top GaN layer may be present (three-layer structure) or not (two-layer structure) in this case. The AlGaN barrier layer must be p-type doped (for example, with Mg or Be as an acceptor) and the bottom GaN layer should be also p-type doped with Mg, Be or intrinsic.

[0063] Thus, there are four hetero-junction three -layer structures implemented in the transistor of the embodiments, based on the above configurations: A. Ga-Face GaN/AlGaN/GaN heterostructure with the 2DEG formed in the GaN buffer layer below the AlGaN barrier layer. In this case, the top GaN layer may be omitted to obtain the two-layer structure. For the three-layer structure, the top GaN layer must be recessed to 1-9 nm thickness in the open gate area or grown with this low thickness, with the roughness below 0.2 nm, and the thickness of the AlGaN barrier can be adjusted properly during growth

B. Ga-Face GaN/AlGaN/GaN heterostructure with the 2DHG conducting channel formed in the top GaN layer above the AlGaN barrier layer. The top GaN layer must be recessed to 5-9 nm thickness in the open gate area with the roughness below 0.2 nm, and the thickness of the AlGaN barrier layer can be adjusted properly. P-type doping concentrations of the GaN layer and AlGaN barrier have to be adjusted; the 2DHG has to be contacted (in the ideal case by ohmic contacts).

C. N-Face GaN/AlGaN/GaN heterostructure with the 2DEG in the top GaN layer above the AlGaN barrier layer. The top GaN layer must be recessed to 5-9 nm thickness in the open gate area with the roughness below 0.2 nm. Thickness of the AlGaN barrier can be adjusted during growth. N-type doping levels of the GaN buffer layer and the AlGaN barrier layer must be adjusted; the 2DEG has to be contacted (in the ideal case by ohmic contacts).

D. N-Face GaN/AlGaN/GaN heterostructure with the 2DHG in the GaN buffer layer below the AlGaN barrier layer. In this case, the top GaN layer may be omitted to obtain the two-layer structure. In both, the two-layer and three-layer configurations, the top GaN layer must be recessed to 1-9 nm thickness in the open gate area with the roughness below 0.2 nm, and the thickness of the AlGaN barrier can be adjusted properly.

[0064] In all the above structures, the deposition of a dielectric layer on top might be beneficial or even necessary to obtain a better confinement (as in case of the N-face structures). As shown in Fig. 8, for the above "C" structure, it may be even more beneficial to include an ultrathin (about 1 nm) AIN or AlGaN barrier layer with high Al- content on top of the 2DEG channel to improve the confinement.

[0065] The preferable structures of the embodiments are structures "B" and "C". In the structure "B ", the 2DHG conducting channel formed in the top GaN layer, which has a higher chemical stability (particularly towards surface oxidation) than the AlGaN layer. Concerning the structure "C", the 2DEG conducting channel might be closer to the surface. Therefore, the electron mobility might be lower than in the 2DEG structure with the Ga-face polarity. In general, the polarity of the heterostructure can be adjusted by the choice of the substrate (e.g. C-face SiC) or by the growth conditions.

[0066] Based on the above, one of the aspects of the present application is an open- gate pseudo-conductive high-electron mobility transistor (PC-HEMT) for amplifying signals produced by a body of a user in a sub-THz radiation range, comprising:

(1) a multilayer hetero-junction structure made of gallium nitride (GaN) and aluminium gallium nitride (AlGaN) single -crystalline or polycrystalline semi-conductor materials, deposited on a substrate layer, and characterised in that:

(a) said structure comprises (i) one top GaN layer recessed in an open gate area of the transistor to the thickness of 5-9 nm and having the surface roughness of 0.2 nm or less, (ii) one bottom GaN buffer layer, and (iii) one AlGaN barrier layer in between; said layers have Ga-face polarity, thus forming a two-dimensional hole gas (2DHG) conducting channel in the top GaN layer, close to the interface with said AlGaN barrier layer;

(b) said structure comprises (i) one top GaN layer recessed in an open gate area of the transistor to the thickness of 5-9 nm and having the surface roughness of 0.2 nm or less, (ii) one bottom GaN buffer layer, and (iii) one AlGaN barrier layer in between; said layers have N-face polarity, thus forming a two-dimensional electron gas (2DEG) conducting channel in the top GaN layer, close to the interface with said AlGaN barrier layer; or

(c) said structure comprises (i) one top AlGaN layer recessed in an open gate area of the transistor to the thickness of 5-9 nm and having the surface roughness of 0.2 nm or less, and (ii) one bottom GaN buffer layer; said layers have N-face polarity, thus forming a two-dimensional hole gas (2DHG) conducting channel in the GaN buffer layer, close to the interface with said AlGaN barrier layer;

(2) source and drain contacts connected to said 2DEG or 2DHG conducting channel and to electrical metallizations for connecting said transistor to an electric circuit; and (3) a Vivaldi antenna electrode placed on a top (GaN or AlGaN) layer between said source and drain contact areas in an open gate area of the transistor and capable of detecting electrical signals in the sub-THz-frequency range of 200-800 GHz.

[0067] Another important feature of the sensor of the present application is that an electrical connection of the heterojunction structure to the 2DEG or 2DHG channel is realised via capacitive coupling to the electrical metallizations through a Schottky barrier contact. "Capacitive coupling" is defined as an energy transfer within the same electric circuit or between different electric circuits by means of displacement currents induced by existing electric fields between circuit/s nodes. In general, ohmic contacts are the contacts that follow Ohm's law, meaning that the current flowing through them is directly proportional to the voltage. Non-ohmic contacts however do not follow the same linear relationship of the Ohm's law. In other words, electric current passing through non-ohmic contacts is not linearly proportional to voltage. Instead, it gives a steep curve with an increasing gradient, since the resistance in that case increases as the electric current increases, resulting in increase of the voltage across non-ohmic contacts. This is because electrons carry more energy, and when they collide with atoms in the conducting channel, they transfer more energy creating new high-energy vibrational states, thereby increasing resistance and temperature.

[0068] When electrical metallizations are placed over single-crystalline or poly- crystalline semiconductor material, the "Schottky contact" or "Schottky barrier contact" between the metal and the semiconductor occurs. Energy of this contact is covered by the Schottky-Mott rule predicting the energy barrier between a metal and a semiconductor to be proportional to the difference of the metal-vacuum work function and semiconductor- vacuum electron affinity. However, this is an ideal theoretical behaviour, while in reality most interfaces between a metal and a semiconductor follow this rule only to some degree. The boundary of a semiconductor crystal abrupt by a metal creates new electron states within its band gap. These new electron states induced by a metal and their occupation push the centre of the band gap to the Fermi level. This phenomenon of shifting the centre of the band gap to the Fermi level as a result of a metal-semiconductor contact is defined as "Fermi level pinning", which differs from one semiconductor to another. If the Fermi level is energetically far from the band edge, the Schottky contact would preferably be formed. However, if the Fermi level is close to the band edge, an ohmic contact would preferably be formed. The Schottky barrier contact is a rectifying non-ohmic contact, which in reality is almost independent of the semi-conductor or metal work functions.

[0069] Thus, a non-ohmic contact allows electric current to flow only in one direction with a non-linear current- voltage curve that looks like that of a diode. On the contrary, an ohmic contact allows electric current to flow in both directions roughly equally within normal device operation range, with an almost linear current-voltage relationship that comes close to that of a resistor (hence, "ohmic").

[0070] Since the source and drain contacts are non-ohmic (i.e. capacitively -coupled), the DC readout cannot be carried out. To electrically contact the 2DEG/2DHG channel underneath, about 5-20 nm bellow the metallizations, the AC-frequency regime must be used. In other words, the AC readout or impedance measurements of the electric current flowing through the 2DEG/2DHG-channel should be performed in this particular case. The capacitive coupling of the non-ohmic metal contacts with the 2DEG/2DHG channel becomes possible only if sufficiently high AC frequency, higher than 30 kHz, is applied to the metallizations. To sum up, the electrical metallizations, which are capacitively coupled to the 2DEG/2DHG channel utilise the known phenomenon of energy transfer by displacement currents. These displacement currents are induced by existing electrical fields between the electrical metallizations and the 2DEG/2DHG conducting channel operated in the AC frequency mode through the Schottky contact as explained above.

[0071] In one embodiment, a tapered slot antenna, also known as a Vivaldi antenna is placed in the open gate area of the transistor. The Vivaldi-antenna gate is capable of detecting various sub-THz-frequencies, i.e. 200-800 GHz. By applying additional plasmonic filters of any kind, the frequencies could be tuned precisely to a specific frequency of choice. In general, a Vivaldi antenna is a co-planar broadband-antenna, which is made from a dielectric plate metalized on both sides. Fig. 9 shows the model the model of the Vivaldi antenna realised on a thin dielectric substrate. An exponential function is used for the taper profile. The entire domain is bounded by a perfectly matched layer. The objective of this model is to compute the far-field pattern and to compute the impedance of the structure. Good matching is observed over a wide frequency band. In this model of the Vivaldi antenna, the tapered slot is patterned with a perfect electric conductor (PEC) ground plane on the top of the Si dielectric substrate. One end of the slot is open to air and the other end is finished with a circular slot.

[0072] On the bottom of the substrate, the shortened 50 Ohm microstrip feed line is modelled as the PEC surfaces. The entire modeling domain is bounded by a perfectly matched layer (PML) chamber absorbing all radiated energy. To excite the antenna, a lumped port is used. Fig. 10 shows the electric field distribution in the Vivaldi antenna plane at 480 GHz. The substrate size is 917x667x4 3 2

μπι with total area about 0.6 mm . Substrate material is a high-dielectric constant, low conductive semiconductor material with σ > 1000 Ohm*cm, such as for example, Si and GaAs. The following table provides the exemplary sub-THz Vivaldi antenna characteristics:

[0073] Fig. 11 shows (a) the far-field directional radiation pattern of the Vivaldi antenna starting from 240 GHz till 780 GHz and (b) the corresponding 3D far-field pattern at 480 GHz, with the maximum in the X direction (axis along the substrate), which means this is a unidirectional antenna. As seen in Figs. 10 and 11, the feeding line excites a circular space via a microstrip line, terminated with a sector-shaped area. From the circular resonant area the energy reaches an exponential pattern via a symmetrical slot line. In addition, the frequency response SWR of the Vivaldi antenna shows wide -band impedance matching better than 2: 1 in most of the simulated frequency range. Printed circuit technology makes this type of antenna cost effective at sub-THz frequencies.

[0074] Advantages of the Vivaldi antenna is its broadband characteristics suitable for ultra-wideband signals in the sub-THz frequency domain, its easy manufacturing process using common methods for PCB production, and its easy impedance matching to the feeding line using microstrip line modelling methods. Also, the Vivaldi antenna has been chosen because it permits to integrate a long meander delay without having undesired effects.

[0075] The Vivaldi antenna is a reciprocal device. It collects in a passive mode exactly the same frequencies that can be actively radiated. Due to its very broadband character, the Vivaldi antenna may receive signals outside the 240 GHz-780 GHz range. In order to limit the antenna sensitivity to the desired 0.3 THz-0.6 THz range, two- dimensional photonic crystals (18) may be superimposed on top of the antenna metal layer, as shown in Fig. 12. The exemplary dimensions of such Vivaldi antenna and lattice constant of the photonic crystal are summarised in the following table:

[0076] Fig. 13 shows the position of the metal connector (19) on the Vivaldi antenna for coupling with the PC-HEMT of the embodiment. Fig. 14 schematically shows a microelectronic sensor for non-invasive monitoring of glucose levels in blood, with a remote readout, comprises the following components (this is a single-transistor sub-THz solution):

(a) at least one PC-HEMT (100) of the present embodiments with an integrated Vivaldi antenna;

(b) an integrated circuit (101) for storing and processing a signal in a sub-THz frequency domain, and for modulating and demodulating a radio-frequency (RF) signals;

(c) an -pulse generator (102) for pulsed RF signal generation; (d) an integrated DC-RF current amplifier or lock-in amplifier (103) connected to said μ-pulse generator (102) for amplification of the signal obtained from said μ-pulse generator;

(e) an analogue-to-digital converter (ADC) (104) with in-built digital input/output card connected to the amplifier (103) for converting the received analogue signal to a digital signal and outputting said digital signal to a microcontroller unit;

(f) the microcontroller unit (MCU) (105) for processing and converting the received digital signal into data readable in a user interface or external memory;

(g) a wireless connection module (106) for wireless connection of said microelectronic sensor to said user interface or external memory.

[0077] Fig. 15 schematically shows a microelectronic sensor for non-invasive monitoring of glucose levels in blood, with a remote readout, comprises the following components (this is the DC/RF-based sub-THz antenna transistor-array solution for imaging):

(a) the array of the PC-HEMTs of the present embodiment (110), wherein each PC- HEMT in said array has an integrated Vivaldi antenna and connected to its dedicated electrical contact line;

(b) a row multiplexer (107) connected to said array for addressing a plurality of said transistors (PC-HEMTs) arranged in rows, selecting one of several analogue or digital input signals and forwarding the selected input into a single line;

(c) a column multiplexer (108) connected to said array for addressing a plurality of said transistors (PC-HEMTs) arranged in columns, selecting one of several analogue or digital input signals and forwarding the selected input into a single line;

(d) an integrated circuit for storing and processing said signals in a sub-THz frequency domain, and for modulating and demodulating a radio-frequency (RF) signals;

(e) an -pulse generator (102) for pulsed RF signal generation;

(f) an integrated DC-RF current amplifier or lock-in amplifier (103) connected to said μ-pulse generator (102) for amplification of the signal obtained from said μ-pulse generator; (g) an analogue-to-digital converter (ADC) (104) with in-built digital input/output card connected to the amplifier (103) for converting the received analogue signal to a digital signal and outputting said digital signal to a microcontroller unit;

(h) the microcontroller unit (MCU) (105) for processing and converting the received digital signal into data readable in a user interface or external memory;

(i) a wireless connection module (106) for wireless connection of said microelectronic sensor to said user interface or external memory.

[0078] The ADC card (104) may be any suitable analogue-to-digital converter data logger card that can be purchased, for example, from National Instruments ^® or LabJack ^® . Optionally, the current amplifier (103) can be operated directly with current flowing via the 2DEG/2DHG channel into the amplifier with small input resistance of 1ΜΩ at gain higher than 10 ⁴ and only 1Ω at gains lower than 200. This setup may directly amplify the electric current modulation in the 2DEG channel originated from an external body charges.

[0079] In a specific embodiment, the wireless connection module (106) may be a short-range Bluetooth ^® or NFC providing wireless communication between the wearable device or gadget and a smartphone for up to 20 m. If this module is Wi-Fi, the connection can be established with a network for up to 200 nm, while GSM allows the worldwide communication to a cloud. The external memory may be a mobile device (such as a smartphone), desktop computer, server, remote storage, internet storage or telemedicine diagnostics cloud.

[0080] In some embodiments, the sensors of the present application can be used for portable long-time-operation solution within remote cloud-based diagnostics. The portable sensor of an embodiment should have a very small power consumption saving the battery life for a prolong usage. In this case, the non-ohmic high-resistive contacts capacitively connecting the sensor to an electric circuit are preferable. The non-ohmic contacts actually limit an electric current flowing through the 2DEG/2DHG channel by having an electrical resistance 3-4 times higher than the resistance of the 2DEG/2DHG- channel, thereby reducing electrical power consumption without sacrificing sensitivity and functionality of the sensor. Thus, the use of non-ohmic contacts in some embodiments of the sensor of the present application is a hardware solution allowing to minimise the power consumption of the device. In another embodiment, the power consumption of the device can be minimised using a software algorithm managing the necessary recording time of the sensor and a battery saver mode, which limits the background data and switches the wireless connection only when it is needed.

[0081] In some embodiments, a method for non-invasive monitoring of glucose levels in blood comprises the following steps :

(1) Contacting a single sensing point on the user's body with, or remotely positioning in a space against the user's body, the microelectronic sensor of the embodiments;

(2) Recording electrical signals received from the user's body with the microelectronic sensor in a form of a source -drain electric current of the microelectronic sensor over time (I _DS dynamics) and/or measuring S11-S 12 parameters of the microelectronic sensor over time (S11-S12 dynamics);

(3) Transmitting the recorded signals from said microelectronic chip to an external memory for further processing; and

(4) Converting the transmitted signals to digital signals and processing the digital signals in the external memory, correlating said I _DS dynamics and/or SI 1-S12 dynamics with pre -calibrated waveforms for blood glucose levels stored in the external memory, and calculating the user's blood glucose levels from said signals in a form of a numerical or readable data, thereby monitoring the blood glucose levels of the user.

[0082] In conclusion, what makes the sensor of the present embodiments particularly useful and unique is the combination of the PC-HEMT and Vivaldi antenna in one single transistor. The helical structure of human eccrine sweat ducts, together with the dielectric properties of the human skin, suggested that their electromagnetic (EM) properties would resemble those of an array of helical antennas. In order to examine the implications of this assumption, numerical simulations in the frequency range of 100-450 GHz, were conducted. In addition, an initial set of measurements was made, and the reflection spectrum measured from the skin of human subjects was compared to the simulation results. The simulation model consisted of a three layer skin model (dermis, epidermis, and stratum cornea) with rough boundaries between the layers and helical sweat ducts embedded into the epidermis. The spectral response obtained by the simulations coincided with the analytical prediction of antenna theory and supported the hypothesis that the sweat ducts can be regarded as helical antennas.

[0083] The results of the spectrum measurements from the human skin were found to be in a good agreement with the simulation results in the vicinity of the axial mode. The magnitude of this response was found to be dependent on conductivity of sweat in these frequencies, but the analysis of the phenomena and the frequencies related to the antennalike modes were found to be independent of this parameter. Moreover, circular dichroism of the reflected electromagnetic field is a characteristic property of such helical antennas. The helical structure of the sweat ducts has the strongest effect on the reflected signal at frequencies above 200 GHz, where the wavelength and the dimensions of the ducts are comparable. In particular, the strongest spectral response (as calculated by the simulations and measured experimentally) was found to be around the predicted frequency (380 GHz) for the axial mode of the helical structure.

[0084] It was also experimentally proven, that besides the black body radiation between 1-20 THz, a human body actively irradiates sub-THz (0.1-1 THz) radiation from sweat ducts. Sweat containing glucose therefore may accurately reflect the blood glucose levels. Thus, the sweat-based electrolyte defining the emission behaviour of the sweat duct-based helix-shaped bionic antennas may change if the chemical content of this electrolyte -based antenna is changing. The radiation principle of the sweat-filled helix- shaped duct-antenna is based on the proton-proton sub-THz interaction changing with the sweat composition. The glucose content in sweat is around 5-10 ppm that correlates with the glucose content in blood, which is 100-400 ppm. Sweat ducts are emitting all the time independently of sweating process as they are always filled with a body electrolyte.

[0085] When the glucose level is changing, the emission spectra and intensity of the sub-THz radiation may shift, and this shift may be detected by the PC-HEMT -based sub- THz sensors of the present embodiments at room temperature. The Vivaldi antenna of the instantly claimed sensors is passively receiving these sub-THz signals from sweat ducts in approximately 200-800 GHz range. Sensing the glucose levels in blood is entirely noninvasive and may be contactless (remote). The patient may either touch the sensor or stay at a relatively short distance from it (few centimetres). The emitting sweat ducts may then change their emission spectra and intensity relative to glucose concentration inside the μ- spiral shaped antenna like glands.

[0086] While certain features of the present application have been illustrated and described herein, many modifications, substitutions, changes, and equivalents will be apparent to those of ordinary skill in the art. It is, therefore, to be understood that the appended claims are intended to cover all such modifications and changes as fall within the true spirit of the present application.