A Scanning Near-field Photon Emission Microscope (SNPEM) for monitoring photon emission sites with a spatial resolution of between 50 to 200 nm is described. A protrusion type probe with a base diameter larger than a wavelength is proposed as a good compromise between resolution and sensitivity. Photon emissions from silicon pn junction and n-MOSFET have been detected with resolution clearly better than the far-field PEM (FFPEM). Features in photon emission distribution smaller than 200 nm were revealed in spite the fact that metal lines prevented the SNPEM probe to reach near-field condition with an actual emission source.

No abstract

We report gain properties of the multi-wavelength pulsed Raman pump. Based on fiber dispersion property, we obtain the group velocity difference between pump pulses which imposes the upper limit of pulse repetition rate beyond which pump-to-pump interaction occurs. The pump-to-pump interaction can be severe when pump instantaneous power is high due to small duty cycle. However, bigger duty cycle imposes much lower limit to the pulse repetition rate. Therefore, an optimized level of duty cycle is required. In addition, shorter pump wavelength separation helps to increase the upper limit of pulse repetition rate. Ultimately this will improve the noise performance of TDM Raman amplifier.

In this paper, the application of scanning near-field photon emission microscopy for imaging photon emission sites is demonstrated. Photon emissions generated by a Fin-FET test structure with one metallization layer are imaged with spatial resolution of 50 nm using scattering dialectic probe. The potential applications and limitations of the technique are discussed. I. IntroductionCurrent fault localization techniques have limited resolution and sensitivity parameters, which are inadequate for the current technology node whereby substantial searching is still required for post-localization [1]. One of these techniques is Photon Emission Microscopy (PEM), which is an important tool for the localization of hot carrier sites and faults in silicon based integrated circuits [2]. The spatial resolution of current PEM tools, which are operated in the optical far-field, is limited by optical diffraction. The theoretical spatial resolution according to Rayleigh Criterion is about half the wavelength used but the practical spatial resolution is closer to one wavelength due to instrumentation constraints. For silicon devices, the significant contribution of photon emission occurs in the near-infrared (NIR) and the practical spatial resolution of far-field NIR PEM systems is limited to about 111m. Recent developments in using solid immersion lenses in PEM have improved the spatial resolution to around 230 nm [3]. However, for a critical defect size of 45 nm at the 90 nm technology node [4], the desired spatial resolution should be better than 50 nm.In our earlier work, a Scanning Near-field Photon Emission Microscope (SNPEM) was introduced [5]. It was used for imaging photon emissions from silicon devices in the visible spectral range from 400 to 900 nm [5]. SNPEM is based on a collection mode of a Scanning Near-field Optical Microscope (SNOM). In SNOM, a sub-wavelength aperture is used for near-field light detection and the dimension of the aperture determines the resolution of the SNOM [6]. The method of collecting photons in the near-field has been proposed [7], but the technique has not been developed. A probable reason could be the low sensitivity due to extreme attenuation of light intensity [8]. Our SNPEM system avoids this limitation and demonstrates a much higher sensitivity at a similar resolution. In our previous work, the improvement was achieved with a protrusion type probe (PTP), which had a

for their kind guidance, great help and continuous stimulation through the work. Working together with Dr Binh during my overseas attachment in Australia has given me so much knowledge, experience and more importantly deep passion in craving creative idea and working on the solution. On top of that, I have also gained valuable experience in transferring the knowledge to students under my mentorship. I'm also grateful to all NTU Research Scholars in Network Technology Research Centre (NTRC) for their fruitful discussion and help. I also acknowledge all the great assistance and continuous support from the research staff and technicians in the Network Technology Research Centre through my study. Last but not least, my sincere thank to my wife and my beloved family for all wonderful love and support they have given to me. Without them, this thesis would not have been completed successfully.