November_EDFA_Digital 1 1 ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 22 NO. 4 times before each electrical measurement. DHE also suffers from contact making-and-breaking noise which in turn reduces the repeatability and stability of the data. The accuracy and depth resolution of DHE has been gen- erally limited because of its manual nature. In contrast, the DHEM technique is a fully automated, fast, and high resolution (≤ 1 nm depth resolution) approach. In a DHEMmeasurement, first a test-pattern such as a Van der Pauw cross is prepared on the coupon sample to be characterized. The cross has amesa structure isolating the top film to be characterized from its surroundings. It is also important that there is an insulating barrier electrically separating the top film from the substrate below, otherwise the electrical measurements would be compromised by substrate effects (see Figs. 1a and b). The insulating barriermay be an oxide layer, a semi-insulating substrate, or a p-n junction (i.e., the film to be character- izedmay be an n-type film formed over a p-type substrate, or vice versa). Four electrical contacts are applied to the ends of the four arms of the test-pattern and the nozzle of a small process chamber with ability to deliver chemi- cal solutions, deionized water or nitrogen gas is sealed against the central area (test region) of the cross-shaped test-pattern as shown schematically in Fig. 1c. The size of the test region can typically be adjusted from 0.5 mm x 0.5 mm to 2 mm x 2 mm so that the data collected is representative of that area. Further miniaturization may also be possible if needed. After the nozzle of the process chamber is sealed against the test region and chemicals delivered, the electrically active thickness of the layer at the test region is reduced in a stepwise manner using chemical and/or electrochemical means. Figure 1c shows two approaches. Oneway is to etch awaymaterial forming a trench. The other way is to convert a portion of themate- rial into an insulating oxide. In either case, measurements of sheet resistance and mobility are carried out after each thickness reduction step using Hall effect and Van der Pauw techniques. Data collected is then processed to yield depth profiles of resistivity, mobility, and carrier concentration. Depth resolution of the data obtained from DHEM depends on the smallest thickness of the semiconductor layer that can be controllably and repeatably removed from the test region. Electrochemical etching and oxida- tion are well suited for this purpose provided that a solu- tion composition and ameasurement recipe is developed for each class of material to be characterized. In general, formaterials that formstable oxides, such as Si, the oxida- tion approach is more attractive. For materials that form unstable or poor oxide layers (such as Ge) an electro- chemical etching approach may be utilized. In any case, the thickness reduction process needs to be calibrated so that the thickness of material removed can be monitored and controlled. For Si and SiGe, using the anodization voltage as the control, it is possible to remove portions of the film from the measurement circuit in steps as thin as 0.3 nm. For Ge, using charge passed through the anodiza- tion circuit as the process monitor, one can control mate- rial removal to the tune of 0.5 nm/step, making sub-nm profiling of both these materials possible. Fig. 1 InDHEM, the electrically active thickness of the semiconductor layer at the test region canbe reducedby electrochemical etching or electrochemical oxidation. (a) and (b) show the top and cross-sectional view of the Van der Pauw structure, (c) schematically shows the thinning and measurement process. (a) (b) (c)