Thin Film Based Gas Sensor for Diabetes Mellitus Applications

Abstract

Diabetes or diabetes mellitus is a group of metabolic diseases in which the level of glucose rises in the human body. This disease results from imperfections in insulin secretion or insulin action in the body. Diabetes can be very dangerous and may even lead to long term damage, non-functioning of organs like heart, kidney and blood vessels. Detection of such a disease is very essential for early treatment. People have been monitoring the glucose level through blood by means of invasive methods. On the contrary, breath analysis has gained recent interest for detection of diabetes. Acetone has been stated to be the biomarker for diabetes mellitus. Breath analysis is a non-invasive method which can be used to detect acetone levels in human breath. For a healthy individual, the level of acetone in breath is less than 0.8 ppm, while it ranges from 1.7 ppm to 3.7 ppm for a diabetic individual. Acetone in breath is found due to the excess formation of ketone bodies in the human body. For detection of diabetes through breath, a very low concentration of 1.7 ppm had to be detected in breath. For this purpose, metal oxide thin films had been considered. A metal oxide gas sensor can be of high importance for the same, since it has been proven that the resistance/electrical conductivity of metal oxides varies when the concentration of surrounding gas changes. MEMS technologies had been used to develop a miniaturized thin film based gas sensor. Four metal oxide thin films, viz., tungsten dioxide (WO2), tungsten trioxide (WO3), tin dioxide (SnO2) and tin-doped tungsten oxide (Sn-doped WO3) were investigated to find the best for detection of acetone gas. After standard cleaning procedure of n-type silicon wafers, a 1 m thick layer of silicon dioxide (SiO2) was deposited over them by means of thermal oxidation. 100 nm thin layer of each WO2, WO3, SnO2 and Sn-doped WO3 were deposited over four different oxidized silicon wafers by means of reactive ion sputtering technique. For deposition of WO2, the ratio of argon gas to oxygen gas was kept to be 80:20. In case of WO3 deposition, the ratio was set as 70:30. Both WO2 and WO3 were annealed at an optimum annealing temperature of 500oC for one hour each. For SnO2 deposition, the ratio was set to be 60:40. This SnO2 film was annealed at 300oC for one hour. For deposition of Sn-doped WO3, a thin layer (1-2 nm) of tin was sputtered over WO3 and this was then diffused in WO3 by annealing at an elevated temperature of 400oC for one hour. The optimum annealing temperatures for each film were analyzed at which the film gave maximum response percentage towards acetone gas. Each wafer was then diced into multiple 5 mm x 5 mm chips. These chips were wire bonded using conducting epoxy at two points for carrying out the measurements. Gas sensing procedure employing the setup elements, viz., computerized gas mixture to mix acetone and dry air to achieve the desired concentration, gas sensing chamber, external heater, picoammeter, voltmeter, etc. was used to find the best film out of four, i.e., SnO2 was found and this was then employed in development of complete micro gas sensor using the fabrication steps. The micro gas sensor of miniature size 2.5 mm x 2.56 mm consisted of three primary elements, viz., the micro heater, the interdigitated electrodes (IDEs) and the sensing film (SnO2). The n-type silicon wafer was first cleaned using inorganic and organic cleaning techniques, followed by thermal growth of 1 m thick SnO2. Backside etching of silicon was done after the photolithography of cavity to be opened using TMAH etching. A cavity of size 1100 m x 1100 m which ended up to around 700 m x 700 m was developed to reduce power losses. Photolithography for micro heater was done on the front side of the wafer and DC sputtering technique was used to develop a platinum heater of thickness 200 nm, followed by lifting-off of excess platinum from the wafer. A passivation layer of silicon nitride (Si3N4) of thickness 0.6 m was deposited using PECVD technique to provide isolation between the micro heater and IDEs. Platinum IDEs of thickness 1300 angstroms were patterned using photolithography and were sputtered using DC sputtering technique over the passivation layer. Lastly, photography for sensing film was done and reactive ion sputtering was used to deposit 100 nm of SnO2 as sensing film. This device was then wire bonded to a TO header. Instrumentation of the developed sensor was done and voltage divider circuitry, current pump circuitry and read out circuitry were formed for displaying the initial and final resistance values (before and after exposure to acetone gas) on the display screen. All four sensing films, viz., WO2, WO3, SnO2 and Sn-doped WO3 were characterized using various techniques like XRD, SEM and AFM. The XRD patterns of tungsten oxide thin films revealed the formation of orthorhombic phase of WO2 and hexagonal phase of WO3. The other films depicted reflections of orthorhombic phase of tin (IV) oxide (SnO2) and Sn-doped WO3 resembling reflections of Sn and WO3. The topography of sensing film was characterized using AFM and SEM. The surface morphology of the thin film was seen by this method. The grain sizes and roughness had also been computed. Each sensing film was exposed to different concentrations of acetone gas and the record of resistance changes was maintained. Response percentages corresponding to various concentrations of acetone gas were computed. Optimum operational temperature for each sensing film was computed as 260oC for WO2, 220oC for WO3, 360oC for SnO2 and 300oC for Sn-doped WO3. Very low concentration of acetone gas, i.e., 1.5 ppm was detected by SnO2 with good percentage response of 30 %. Sn-WO3 could detect a low acetone concentration of 3 ppm but with a low response percentage of 0.6 %. WO2 and WO3 could detect a minimum acetone gas concentration of 10 and 15 ppm with response percentages 15 % and 17 %, respectively. Hence, SnO¬2 was chosen to be the best of all for detection of acetone gas with a minimum concentration of 1.5 ppm. Interfering gases like ethanol, carbon dioxide and benzene were also tested on SnO2 and it was observed that at an optimum operational temperature of 360oC, SnO2 responds the best to acetone gas.