Figure from:The effect of bias-temperature stress on Na+ incorporation into thin insulating films

See full PDF downloadDownload figure

Related Figures (7)

Table 1 Comparison of detected Na’ doses with calculated values from cyaotr (PNa+, calc) AS Obtained by ToF-SIMS (#ya+, sms) and potentiostatic (Pya+, pc) sample processing. The given @-ratio DC/ * @at, calc for CNaoTe=0 ppm was obtained via F-AAS compared to ®ya+, sims and subsequently added as intrinsic Na’ dose (30 ppm) to ®na+, calc for Cnaotr=300, 600, and 900 ppm, respectively SIMS represents the excess of response current from potentiostatic experiments with respect to spectrometric data

Fig. 3 Nyqusit plot of electrical impedance spectroscopy measure- ments emphasizing the characteristics of the low frequency range of the spectrum (f=110'-1x10-' Hz), most likely related to processes in the SiO, film. Spectra were acquired on a sample with dgio2= 100 nm, cnaore= 300 ppm at T= 150 °C with representative features also observable on samples with thicker SiO, films. In the high frequency region (f=1*10°-110° Hz), related to processes in the PMMA film, all spectra exhibit identical characteristics (not shown here)

Table 2 Calculated £, for Na’ transport within SiO. without appliec Upias aS obtained by experimental data (Fig. 4) and data processing according to Eq. 4

Fig. 4 Arrhenius plot of data in Fig. 2. Linear regression has been applied for the different data sets (representing a respective applied Upias). The resulting slopes indicate reduced values for E,, which decreases with increasing U,jia;. The value for the unaltered £, within Si0,, i.e., without applied U,;,, could be evaluated using Eq. 4

Fig. 5 ToF-SIMS depth profile of insulating films (SiOz, SiO,N,, and Si;3N4; d=400 nm) after BT stress using E=4 MV/cm. The PMMA layer was removed, and Au was again deposited prior to depth profiling. The Na’ permeability of the insulating layer was suppressed with increasing nitride content In contrast to SiOz, the drop out rate was 0% for nitride- containing layers; thus, no dielectric breakdown was observed during BT stressing. In Fig. 5, the respective ToF-SIMS depth profiles after BT stress are illustrated and compared to a 400 nm SiO, film. Due to the enormous bias voltage for the 400 nm SiO, film, these samples could not be BT stressed with an internal electrical field of E=4 MV/cm using the methodology described above. For that reason, the 400 nm SiO, film was BT stressed with E=3 MV/cm, qualitatively resulting in the same profile characteristics. These more gentle conditions prevented dielectric breakdown, and a proper comparison could be given in Fig. 5. Note again that the PMMA layer was removed from the samples after BT stressing and prior to ToF-SIMS depth profiling. It is obvious that in contrast to SiOz, Si;N, did not exhibit any permeability to Na’ ions during BT stressing. Thus, no incorporation took place in SizN4, even at such rigorous bias voltage conditions. The SiO,N, film however showed a slight tailing towards the insulators bulk. In contrast to SiOp, the detected magnitude of Na in SiO,Ny was very low, and no Na’ reached the SiO,N,/Si interface in the chosen time interval, although rigorous BT conditions were applied. A lower Usa, or a lower sample temperature decreased the depth in SiO,Ny where Na’ still could be detected (In Fig. 5, ~200 nm) until for 80 V bias voltage no difference to a blank signal could be observed anymore (not shown in Fig. 5). This behavior could also be observed if one acquires ToF- SIMS depth profiles in the described sample before BT stressing and PMMA removal. Caused by migration, insulating materials are prone to be contaminated by Na