Test structure design considerations for RF-CV measurements on leaky dielectrics

Application and Evaluation of the RF Charge-Pumping Technique

“…The approximation made in (7) is valid as long as Z line Z G . For properly designed test structures (see, e.g., [8]), this assumption is valid for the frequencies used in this paper.…”

“…The test structures are designed in a two-port ground-signalground configuration similar to [8], optimized for accurate RF measurements. The structures consist of transistors with the source and drain connection shorted; the gate is connected to one of the signal pads while the source/drain is connected to the other signal pad.…”

Reliability engineering in RF CMOS

“…This expression shows the dependence of Q on the measurement frequency and the magnitude of the model parameters of the three-element model of figure 5.1. For a 10 x 10 µm 2 MOS capacitor structure typical model parameters of the threeelement model are [100]: C ox = 2 pF, g ox ≈ 50•10 −6 Ω −1 and R s = 10-100 Ω. In figure 5.2 Q is plotted against frequency for such a typical example with R s set to 10 Ω, similar to what is done in [100].…”

“…For a 10 x 10 µm 2 MOS capacitor structure typical model parameters of the threeelement model are [100]: C ox = 2 pF, g ox ≈ 50•10 −6 Ω −1 and R s = 10-100 Ω. In figure 5.2 Q is plotted against frequency for such a typical example with R s set to 10 Ω, similar to what is done in [100]. In this example the lowest frequency for which Q > 1 is 4 MHz and only for frequencies higher than 42 MHz Q exceeds 10.…”

“…As explained in chapter 2 measurements at such high frequencies require the use of RF measurement techniques. In [97,100] it was shown that accurate C-V curves can be obtained on leaky dielectrics using two-port s-parameter measurements at radio frequencies using a VNA. These measurements were accompanied with SOLT calibration and OPEN-SHORT de-embedding.…”