This paper reports our progress on the temperature control of a room temperature Si3N4/SiO2 membrane-type electrical substitution radiometer (ESR), using two control loops and a chopping procedure. Sensing and heating elements were patterned in a platinum thin film, deposited on a 1560 mu m x 1560 mu m membrane made of a 280 nm thick Si3N4/SiO2 bilayer. The sample was fabricated in a 500 mu m thick silicon substrate by chemical anisotropic micromachining and then passivated with a 1 mu m thick SiO2 layer. To simplify operating conditions, the device was operated at room temperature in a primary vacuum chamber, with no coolant or no large heat sink other than the sample holder itself. The ESR technique is based on the periodic substitution of Joule heating power by the input power, whereas the thermometer delivers the error signal, kept as small as possible. Dynamic aspects of the loops are detailed. The input noise level was measured, leading to a signal-to-noise ratio of 80 for an incoming 1 mu W driving power in a 1 Hz bandwidth. Even if still slightly lower than our requirements for applications in metrology, these achievements are extremely promising in a wide range of applications.
A room temperature Si3N4/SiO2 membrane-type electrical substitution radiometer using thin film platinum thermometers
2007
Abstract
This paper reports our progress on the temperature control of a room temperature Si3N4/SiO2 membrane-type electrical substitution radiometer (ESR), using two control loops and a chopping procedure. Sensing and heating elements were patterned in a platinum thin film, deposited on a 1560 mu m x 1560 mu m membrane made of a 280 nm thick Si3N4/SiO2 bilayer. The sample was fabricated in a 500 mu m thick silicon substrate by chemical anisotropic micromachining and then passivated with a 1 mu m thick SiO2 layer. To simplify operating conditions, the device was operated at room temperature in a primary vacuum chamber, with no coolant or no large heat sink other than the sample holder itself. The ESR technique is based on the periodic substitution of Joule heating power by the input power, whereas the thermometer delivers the error signal, kept as small as possible. Dynamic aspects of the loops are detailed. The input noise level was measured, leading to a signal-to-noise ratio of 80 for an incoming 1 mu W driving power in a 1 Hz bandwidth. Even if still slightly lower than our requirements for applications in metrology, these achievements are extremely promising in a wide range of applications.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.