Development of Pt-based Sensitive Layer for Carbon Monoxide Work Function Change Based Sensor in Air at Room Temperature
Hansch, Walter, Prof. Dr.-Ing.
Hansch, Walter, Prof. Dr.-Ing. ; Eisele, Ignaz, Prof. Dr. rer. nat.
Tag der mündlichen Prüfung:
carbon monoxide, sensor, work function change, Pt-based, air, room temperature
Carbon monoxide (CO) gas is toxic to human health above a concentration of 30 ppm. Therefore a sensor is needed to detect the existence of CO. Nowadays, CO sensor based on resistance can be found easily in the market. There are some disadvantages of resistance based sensor. Instead of facing the obstacles of resistance based sensor, the second subtype of solid state sensor can be an alternative, i.e., work function change based sensor. A work function sensor like a Floating Gate Field Effect Transistor (FGFET) uses the gas induced surface work function change (called Contact Potential Difference / CPD) as sensing signal. Platinum (Pt) is a sensing material already known as detection layer, e.g., for ozone or hydrogen gas sensor. Additionally, there are some manners to involve Pt as a work function based sensor for carbon monoxide (CO) such as employing it as a catalyst on a metal oxide like Ga2O3 . However, pure Pt lacks the sensitivity to CO. In this research, Pt is used as a base material mixed with Au because the adsorption of CO on Au was studied since 1925. In other research, PtAu was also reported that it can detect CO in the hydrogen atmosphere. A model in some simulations also investigates the laboratory results of this research.
Experimentally, PtAu samples which consist of Pt as a majority element than Au cannot give good CPD signals as CO sensor started from 30 ppm in air at room temperature. On the other hand, PtAuTi samples which contain a small amount of Ti can always give good CPD signals. The simulation confirms it with/without adjustment factors. It proves that small amounts of Ti as metal oxides with less noble metal plays a role in stabilizing the CPD signals on PtAuTi samples. PtAuFe samples were also produced by an evaporating process, presenting CPD signals as predicted, although the signals are not as good as PtAuTi samples. Experimentally and in simulation, it is also proven that Pt, Au, and Ti cannot give good signal compared to PtAuTi.
By simulation, it is found that adjustment factors fadj2 = 0.296, fadj3 = 0.6 and fadj4 = 0.125 are needed for adsorption energy of CO on Pt (Ef1Pt), desorption energy of oxygen on Au (Er45Au) and dissociation energy of CO2(s) on Ti (Er4Ti) respectively.
There are no CPD signals from PtAuTi samples after exposed 24 hours in the air because of oxygen (O) occupation. O can diffuse into Pt (also applied for Au and Ti in simulations) with diffusion constant 10^(−19) cm2/s forming subsurface oxygen. Process from chemisorbed oxygen to be subsurface oxygen attenuates the sticking coefficient of every gas (O2 , CO, CO2 , H2O, OH, and H2) with simulated factor fadj1 = 1−0.00585exp(5.76664Osub). Samples can be refreshed using dry etching treatment or annealing treatment at 170 oC for 10 minutes in the air. Simulation confirms that annealing treatment can refresh the sample after exposure in the air for 24 hours.
Experimentally, annealing at 170 oC is really the optimum temperature for refreshing the sample. There is no significant difference of the CPD signals in temperature and time annealing variation experimentally and in the simulation. Self-heating chips of Pt80Au14Ti6 have been produced and show that it is not needed high electrical power for the annealing at 170 oC.
Simulation is able to predicts the oxygen atom coverage (ΘO), CO coverage (ΘCO) of 30 - 120 ppm CO, CO2 coverage (ΘCO2), H2O coverage (ΘH2O), OH coverage (ΘOH), H2 coverage (ΘH2), H coverage (ΘH), HCOO coverage (ΘHCOO) and C coverage (ΘC) are ca. 0.8 ML, 10^10 molecules/m2 , 10^9 molecules/m2 , 10^11 molecules/m2 , 10^12 molecules/m2 , 10 molecules/m 2 , 10^9 molecules/m2 , 0 molecules/m 2 and 0 molecules/m 2 respectively. Simulation also proves the relationship between the work function change and electric dipole moment. It is already demonstrated and its approach can give the calculated CPD signal is close to the experiment value i.e. 14.4 mV for 30 ppm CO on Pt80Au14Ti6.
Regarding the cross sensitivity properties of PtAuTi samples, there are cross sensitivities for PtAuTi samples ca. (−100 ± 20) mV, (50 ± 10) mV, (−297 ± 9) mV and 0 mV because of 50 ppm NH3, 5 ppm NO2, 15000 ppm H2 and 5000 ppm CO2 respectively. In application, these cross senstivities could be handled by using suitable reference layers for these gases, together with this PtAuTi layer on a FGFET.
Finally, since the CPD signals of PtAuTi are not high as other sensitive layers which use FGFET, the first differential of CPD signals can be an alternative for determining the CO concentration, instead of spending effort to saturate the CPD signals. «
Carbon monoxide (CO) gas is toxic to human health above a concentration of 30 ppm. Therefore a sensor is needed to detect the existence of CO. Nowadays, CO sensor based on resistance can be found easily in the market. There are some disadvantages of resistance based sensor. Instead of facing the obstacles of resistance based sensor, the second subtype of solid state sensor can be an alternative, i.e., work function change based sensor. A work function sensor like a Floating Gate Field Effect Tra... »
Fakultät für Elektrotechnik und Informationstechnik