Semiconductor dopant profiling by off-axis electron holography
Introduction
With the continual downsizing of device dimensions, two-dimensional (2-D) dopant profiling is an ongoing issue of great interest and importance to the semiconductor industry [1], [2]. Obtaining reliable information about the dopant distribution in real structures as a function of processing parameters remains a key step in validating process and device simulations used for developing and refining prototype device designs [3]. In practice, the spatial distribution and concentration of active dopant atoms play a major role in determining the electrostatic potential within the device, which in turn determines the device response. It would therefore be highly beneficial to develop an experimental method capable of characterizing the potential with high spatial resolution and sensitivity.
Off-axis electron holography in the transmission electron microscope (TEM) provides access to the phase of the electron wavefront that has traversed a sample, so that the technique should in principle enable the electrostatic potential within semiconductor devices to be quantified. Electron holography was used by McCartney et al. to observe the potential profile across Si/Si p–n junctions [4], and the technique was later used by Rau and colleagues to map the 2-D electrostatic potential in deep-submicron transistor structures [5], [6]. Electrostatic potential variations across an AlGaN/InGaN/AlGaN heterojunction diode have recently been quantified using electron holography, and also explained in terms of polarization effects and free polarization-induced interface charge [7].
In this present study, we have systematically explored the application of off-axis electron holography to a test structure of known dopant distribution. We demonstrate that accurate potential profiles can be extracted routinely from unbiased materials using a refined approach to hologram analysis and carefully controlled sample preparation conditions. Experimental factors that are liable to affect quantitative measurements are also discussed.
Section snippets
Experimental details
The specimens studied here were taken from a group of test structures fabricated at IBM for a dopant metrology round-robin specifically targeted at comparing and evaluating different methods for dopant profiling [1]. The substrate was 〈1 0 0〉 p-type silicon, boron-doped at 11–25 Ω cm−1. The test structures were fabricated using low-temperature epitaxial growth, and consisted of an abrupt p–n junction, doped with boron and phosphorus (1018 cm−3). Additional features included three narrow B-doped (1020
Basis for hologram interpretation
Off-axis electron holography provides a quantitative measure of the phase change experienced by the electron wave that has passed through the sample. For non-magnetic materials, and neglecting dynamical diffraction effects and external fields, the phase change is given by the expressionwhere x,y represent the plane of the sample, z is the incident beam direction and V is the electrostatic potential of the object. The interaction constant CE is given by
Dopant mapping
Fig. 4 shows the amplitude (a) and phase (b) images reconstructed from one set of holograms. The band of about 100 nm width at the (left) edge of the sample is the SiO2 protection layer. The amplitude image shows no visible differences in contrast within the sample but the phase image clearly reveals stripes of contrast as a result of the presence of the dopants. From a comparison with Fig. 3(b), it can be seen that the central broad band corresponds to the thick (∼100 nm) B-doped layer, and the
Discussion
Our observations have shown that charging can have a serious effect on electron holographic measurements of electrostatic potential, especially in specimen regions close to the vacuum edge which tend to become positively charged, due to the emission of secondary electrons. The number of secondary electrons ejected is related to the nature of the surface as well as the overall thickness of the sample. A greater fraction of secondary electrons is more likely to be emitted at the thinnest edge
Conclusion
Electron holography has recently attracted a great deal of attention in the semiconductor industry because it appears to provide a possible solution to the critical but unresolved issue of dopant profiling in deep-submicron devices [1]. The results reported here offer real promise that the problems of surface charging can be overcome and that accurate potential profiles should be obtainable on a routine basis provided that there is some prior knowledge of the likely junction depth so that a
Acknowledgements
The authors thank Professor Jian-Min Zuo for use of his CBED simulation program, and Dr. Rick Hervig for assistance with SIMS analysis. We acknowledge use of facilities in the Center for High Resolution Electron Microscopy at Arizona State University.
References (22)
- et al.
Ultramicroscopy
(1993) - et al.
Ultramicroscopy
(1993) - et al.
Ultramicroscopy
(1994) - et al.
Ultramicroscopy
(1993) - et al.
Ultramicroscopy
(1987) - et al.
J. Vac. Sci. Technol. B
(1996) - ITRS Metrology 2000 Update, Table 84a: 2000 Front End Processes Technology Requirements—Near Term, Semiconductor...
J. Vac. Sci. Technol. B
(1992)- et al.
Appl. Phys. Lett.
(1994) - et al.
Phys. Rev. Lett.
(1999)
Phys. Stat. Sol. B
Cited by (37)
Quantitative measurement of nanoscale electrostatic potentials and charges using off-axis electron holography: Developments and opportunities
2019, UltramicroscopyCitation Excerpt :Prior knowledge of the local value of the MIP, which is usually written V0, is often crucial for quantifying the electrostatic potential distribution. However, calculations based on neutral atom scattering factors tend to overestimate V0 [18], highlighting the need for experimental measurements of this key parameter [19]. Measurements of V0 usually require an independent estimate of the specimen thickness profile.
Dopant mapping in thin FIB prepared silicon samples by Off-Axis Electron Holography
2014, UltramicroscopyAdvanced split-illumination electron holography without Fresnel fringes
2014, Ultramicroscopy