2.1.3 Extended-Measurement-Stress-Measurement Setup

To save on time and devices when performing NBTI experiments the extended-MSM (eMSM) measurement routine was established [18]. Choosing each stress sequence tstr,i+1   to be significantly longer than the previous stress sequence tstr,i  ensures that the amount of degradation lost during the recovery within t
rel,i  is nearly completely restored within t
 str,i+1   . Consequently, regardless if the stress is interrupted or not, more or less the same amount of degradation is obtained after the total stress time, i.e.                    ∑
ΔVTH  (tstr) ≈ ΔVTH (  itstr,i)  . This is schematically depicted in Fig. 2.3, where the top dotted black line of the continuous degradation is always met by the individual sub-sequences (red dotted lines) of the eMSM-sequence after sufficiently long stresses. When the stress sequences are recorded via the on-the-fly method, which will be explained in Chapter 2.3, both stress and recovery can be monitored with the eMSM routine.


PIC


Figure 2.3: Schematic view of the last three out of N  = 9  stress/relaxation cycles building up an eMSM-sequence like performed by [1711]. The stress (dashed red) is interrupted N − 1  times to record N  − 1  short and one long final relaxation sequence on the relative time scales trel = t− tstr,i  . After the measurement delay tM   marked by the dashed blue lines the monitorable relaxation (solid blue) sets in. A permanent or slowly relaxing component P  is indicated for the last two cycles and will be explained in Chapter 4.