Thick Film Resistors Ndt by Electro-ultrasonic Spectroscopy

The non-linear electro-ultrasonic spectroscopy was used as a non-destructive testing method for the polymer based and cermet thick film resistors evaluation. We carried out the correlation between this method and standard testing methods as noise spectroscopy and the third harmonic voltage measurements. The measuring set-up with the ultrasonic transducer working on the frequency range 10 kHz to 1 MHz was designed and realized. Proposed method is exploiting two different signal sources – ultrasonic wave and alternating electric current. Mechanical vibrations affect the defects in the sample structure and it influences the electric charge transport through the measured structure. Resulting information is measured on the differential frequency given by the superposition or subtraction of exciting signals frequencies. Low-pass filter with limit frequency 5 kHz was used for the suppression of exciting signals amplitudes. Measured sample together with the ultrasonic transducer creates the resonant system. For the thick film resistor samples prepared on the alumina substrate we achieved high sensitivity measuring on the resonant frequencies determined by the alumina substrate size. Four-point method was used for the elimination of the contact influence on the measured characteristics. The intermodulation signal was measured on the frequency in the range 2 to 5 kHz. The intermodulation signal amplitude increases linearly with the amplitude of alternating electric current, and with the square of increasing voltage on the ultrasonic transducer. The intermodulation signal amplitude is further influenced by the sample technology. The intermodulation voltage is influenced by the AC current value and on the resistance change of contacts among the conducting grains in the thick film resistor structure. Ultrasonic signal changes the area of the contact between the conducting grains in the resistor structure, hence the square-law between the voltage on the ultrasonic transducer and the intermodulation voltage is observed. The method sensitivity is influenced by the measuring set-up noise background. The intermodulation signal amplitude was measured to be about one order above the measuring set-up background noise voltage for the current 1 mA flowing through the structure. In this case no sample heating is observed even for through the long term measurements. The relative resistance change is of the order of 10 for the polymer based thick film resistors, and of the order of 10 for the cermet samples. There is a correlation between the resistance value and the intermodulation voltage. At the same time there exists the correlation between the amplitude of intermodulation voltage and the value of the noise spectral density or the value of the third harmonic signal amplitude, respectively. It happens in all cases that with the increasing number of contacts among the conducting grains in the sample there are decreasing the amplitude of intermodulation voltage, as well as the value of the noise spectral density, the value of the third harmonic signal amplitude, and the resistance for the samples with identical geometry. This enables to evaluate the sample technology.