No abstract

A technique (the capacitance-ratio test) is described for assessing the water content of hermetic packages containing an integrated circuit die. It is based on a theoretical analysis which-shows that the frequency dependence of-the capacitance between a selected pair of IC metallisation tracks can be used tc derive -the moisture induced-component of this capacitance. A measurement on an encapsulated integrat circuit die, held at.a constant temperature, is used to determine the ratio of the moisture inducedcapacitance and the -intrinsic capacitance betweenmetallisation tracks. By-comparing this ratio with that obtained from-a sample of integrated circuits exposed to a known relative humidity, the test-can indicate whether the relative humidity in the package is greater or less than a specified value.iThe accuracy of the test is comparable with that of other moisture sensing methods for encapsulations, and its sensitivity is adequate for the control of known moisture-related -failure mechanisms in integrated circuits. -IntroductionMoisture has been found to be a particularly deleterious hazard to monolithic and hybrid microcircuits-, and moisture induced failure mechanisms, such as electrochemical corrosion and dendritic growths have been widely reported.1'2 Moisture in hermeticpackages can,arise from imperfect encapsulation procedures that can give rise to leaks, or to water being trapped in crevices. Another source is the evolution of water from materials used in fabrication such as the sealing glasses of ceramic packages3 and *.the epoxy resins of hybrid microcircuits.4Much effort has been expended on the detection and accurate measurement of the small quantities of moisture involved, typically 4 x lo-4 mbar.l in a ceramic dual-in-line package. The three techniques which have received most attention are based on (i) masE spectrometers,5 (ii) the sealing of miniature A1203 hygrometers into packages6 and (iii) the temperature dependence of a moisture-sensitive parameter of the encapsulated die (or special test structure ) -a technique commonly referred to as the dew-point test.7 Because none of these techniques are ideally suited to production line control they have not been widely adopted. An operationally simpler test is considered here.-Recently, a development of the dew-point test has been reported in which it was shown that the capacitance between a pair of metallisation tracks of the die of any integrated circuit has a component which depends on the amount of water adsorbed on the surface of the die. By slQw, localised, cooling of the die, the relative humidity at its surface was raised until the amount of adsorbed water was sufficient to cause a detectable increase in ca;pacitance. The temperature at which such an increase occurred was taken as a measure of the water content of the package. The technique described in this paper is a derivative of the capacitance dew-point test with, however, two important differences. Firstly, the measurement of the capacitance is made at a constant temperature and, secondl...

Plasmons have been investigated in highly doped GaAs/A!o,2Ga,&i multiple quantum well structures using the far-infrared techniques of dispersive Fourier transform spectroscopy and polarized oblique-incidence reflection spectroscopy. The plasma-effective masses were obtained from comparison with van der P u w measurePmen!s, znd found to be considerably enhznced O L ~ !hs/r bulk values. This enhancement has been compared with that obtained using non-parabolicity theory. Good agreement is obtained, but the results suggest that the theory slightly underestimates the effect.

No abstract