As the R p of ion implants steadily decreases an ever-increasing percentage of the implant species lies in the oxide layer and is, therefore, not electrically active. For this reason, it is important to have analytical techniques capable of accurately measuring the thickness of ultrathin oxide layers. A round-robin study was performed on a series of SiO 2 films ranging from 0.3 to 20 nm in order to evaluate the advantages and disadvantages of five commonly used analytical techniques. High-resolution cross-section transmission electron microscopy ͑TEM͒ offers the only true measurement of oxide thickness because no density assumptions are made. In this study, TEM is used as the standard for all the other techniques. X-ray photoelectron spectroscopy and Auger electron spectroscopy offer precise measurements for ultrathin ͑Ͻ3 nm͒ films, but are limited for thicker films ͑Ͼ15 nm͒ due to the exponential decay functions that describe the sampling depth in both techniques. Secondary ion mass spectrometry ͑SIMS͒ has historically been used for characterizing relatively thick films ͑Ͼ10 nm͒ but not for thinner films because of atomic mixing effects. Encapsulating oxides with amorphous silicon prior to performing a SIMS experiment can negate these effects. A comparison of SIMS on encapsulated and as received films is made. Rutherford backscattering is an excellent technique for determining oxide thickness over a wide thickness range by channeling the Si signal from the crystalline substrate and analyzing oxygen from the amorphous oxide. Ellipsometry, being both rapid and low cost, is one of the most widely used techniques capable of accurate measurements on thick films ͑Ͼ10 nm͒.

Measurements on a variety of doped (magnesium and/or iron) and undoped lithium niobate crystals in the oxidized state demonstrate an Arrhenius dependence of dark conductivity on reciprocal temperature between 460 and 590 K. All of the crystals had roughly the same conductivity and activation energy (1.21 eV) over the temperature range, implying that all have about the same free-carrier concentration and mobility. The enhanced photoconductivity of magnesium-doped lithium niobate is attributed to a greatly reduced trapping cross section of Fe3+ for electrons, the smaller cross section being due to a changed substitutional site for Fe3+. The Fe3+ trapping cross section is calculated from photoconductivity data to be of order 10−18 m2 in undoped lithium niobate. This implies a photoelectron lifetime of order 6×10−11 s in a relatively pure (2-ppm Fe) oxidized crystal.

No abstract

No abstract

No abstract