Effective electron backscatter

Although the single-layer negative resists commonly utilized are simple, they typically do not resolve narrow (0.5 ftm) gaps between wide lines or pads due to proximity effects from backscatter from the substrate during electron beam (E-beam) exposure. Further, there are variations in linewidth which occur when images are written in a single-layer resist which overlies steps in the substrate. 

Thick or bulky specimens are commonly used in scanning electron microscopy. However, as shown in Fig. 4.4.2, there are two disadvantages in using such thick specimens (1) a low spatial resolution, and (2) interferences from the effects of backscattering, absorption and fluorescence (Russ 1974). The poor spatial resolution can be remedied by using thin specimens (Fig. 4.4.2). The use of cross-sections of 0.5thickness, for example, would permit analysis of the compound middle lamella in cell walls as a separate entity (Saka and Thomas 1982). The second problem can also be minimized or eliminated by using thin specimens (Russ 1974, Saka and Thomas 1982). 

In a typical direct write e-beam lithographic system, the resolution of a dense line-space array is often limited by the effect of electrons backscattered from the substrate. In these arrays, the backscattered electrons from one exposed line increase the net exposure density in an adjacent line. While this problem of non-uniform exposure can be corrected by varying the exposure dose within the pattern, this form of proximity correction requires sophisticated algorithms and extensive computer facilities. 

Another method for minimizing the proximity effect involves changing the electron beam acceleration potential from the 10 to 20 KV range that is typically used in direct write applications. L. D. Jackel and co-workers have proposed that increasing the acceleration potential to 100 KV will minimize the need for proximity corrections as the backscattered electrons will be averaged over a larger area (2). While this approach may reduce the effect of backscattered electrons, it will also significantly reduce the resist sensitivity since a smaller number of electrons will be deposited in the resist at 100 KV. 

Ema data can be quantitated to provide elemental concentrations, but several corrections are necessary to account for matrix effects adequately. One weU-known method for matrix correction is the 2af method (7,31). This approach is based on calculated corrections for major matrix-dependent effects which alter the intensity of x-rays observed at a particular energy after being emitted from the corresponding atoms. The 2af method corrects for differences between elements in electron stopping power and backscattering (the correction), self-absorption of x-rays by the matrix (the a correction), and the excitation of x-rays from one element by x-rays emitted from a different element, or in other words, secondary fluorescence (the f correction). 

Electrons can collide with atoms in the sample and be scattered back backscat-tering becomes more effective as the mass of the atom increases. If a region of the sample contains heavier atoms (e.g. Ft) than the surroundings, it can be distinguished due to a higher yield of backscattered electrons. 

Fig. 3.19 Schematic illustration of the measurement geometry for Mossbauer spectrometers. In transmission geometry, the absorber (sample) is between the nuclear source of 14.4 keV y-rays (normally Co/Rh) and the detector. The peaks are negative features and the absorber should be thin with respect to absorption of the y-rays to minimize nonlinear effects. In emission (backscatter) Mossbauer spectroscopy, the radiation source and detector are on the same side of the sample. The peaks are positive features, corresponding to recoilless emission of 14.4 keV y-rays and conversion X-rays and electrons. For both measurement geometries Mossbauer spectra are counts per channel as a function of the Doppler velocity (normally in units of mm s relative to the mid-point of the spectrum of a-Fe in the case of Fe Mossbauer spectroscopy). MIMOS II operates in backscattering geometry circle), but the internal reference channel works in transmission mode...

Quantitative analysis of Auger spectra is quite possible, but less straightforward than in XPS. The main reason is that the effective intensity of primary electrons in the sample may be higher than that of the primary beam, because of backscattering effects. For example, the Auger yield measured from a thin film of... 

M. L. Polignano and G. Queirolo, Studies of the Stripping Hall Effect in Ion-Implanted Silicon J. Stoemenos, Transmission Electron Microscopy Analyses R. Nipoti and M. Servidori, Rutherford Backscattering Studies of Ion Implanted Semiconductors... 

Figure 10. Contrast versus linewidth for 25 kV and 50 kV electrons exposing a 1 / thick resist layer on a silicon substrate. Improved immunity to proximity effect has been reported by Neill and Bull (64). Backscattering for 50 kV electrons was obtained by extrapolation from data given in references...

---