A large amount of exposure and development experiments were carried out to obtain the graphics conforming to the process target. Graphics experimental results are generated by the application of SEM imaging to monitor the use of the equipment, Japan JEOL company JSM- 6401F, field emission, resolution 3 nm.
In the bulk silicon CMOS Fin FET device, the fine groove pattern printed substrate material is SiO2 (multilayer dielectric film), but in order to study the electron beam lithography technology itself, it is also studied in the experiment process
The exposure, development process and characteristics of UV3 on thick SiO2 and bulk silicon (mainly in the former) are investigated.
1. Influence of film thickness of resist
In the process of electron beam exposure, a higher precision figure can be obtained by the thinner thickness. This is because thinner resists only require a lower critical dose; and electrons with the same energy are more likely to penetrate the resist, so that the electron scattering effect is reduced and the resolution of the pattern is improved.
The specific process of coating is: on the automatic transfer homogenizing machine, firstly, 10 ml resist is dripped on the surface of 100 mm silicon wafer, and then spin evenly at the speed of 800 rpm. Finally, the glue is evenly distributed at a high speed for 30 s. The film thickness of the resist is determined by the rotational speed and the viscosity of the resist itself (i.e., the composition of the resist). The maximum speed allowed in the uniform plastic machine (7000 RPM), in order to get a thin film thickness, this paper selectively diluted UV 3, thinning agent for ethyl lactate.
Under the same exposure and development conditions (optimum conditions), the effect of different adhesive thickness on the size of groove pattern is compared, and the results are shown in table 1.
The experimental results show that the thinner the resist is, the smaller the print size is. The lower limit of the gel thickness is that the chemical properties of the gel are basically maintained during the dilution process and can be sufficiently masked in the subsequent etching process.
2. Influence of electron beam exposure parameters
Exposure to electron beam exposure process, the main parameters affecting the accuracy of the image exposure system exposure energy, beam current, scanning grid spacing, focused electron beam exposure UV3 positive resist process mirror 2003, they determine the exposure Spot size and area of the field.
There is an important correlation between the spot size and the pattern accuracy. Figure 1 shows the SEM groove pattern under three exposure conditions. The beam spot diameters for the three conditions (50 keV, 4th LO, beam currents 25 pA, 50 pA and 100 pA) are 30 nm, 50 nm, 100 nm and 50 keV, respectively.
For the 50 nm design linewidth, the results of its exposure are shown in Figure 1, respectively 160 nm, 180 nm, 230 nm. The last two graphics edges are smooth, while the first graphics edge has obvious jaggies.
This is because the proximity effect of the smaller diameter beam spot (30 nm) has less influence on the surrounding area at certain scanning pitches, so the transitional area between the scanning grids is conspicuous, causing the edge of the pattern to fluctuate.
The size of the field also affects the actual exposure time. For a groove grid pattern on a standard cell of 2. 8 mm × 2.1 mm, the exposure times for the above three conditions are 60 min, 30 min, and 15 min, respectively, and with the fifth objective, the time is increased by 50 %the above.
Another important factor affecting the accuracy of the pattern is the inherent proximity effect of electron beam exposure. Electron beam exposure process, the electron in the resist and the substrate multiple collisions, scattering, so that in the exposed area adjacent to the region to produce an unwanted exposure, resulting in the edge of the image blur, deformation, steepness decline, which is Proximity effect of electron beam exposure.
Proximity effects must be corrected for complex shapes, otherwise the accuracy of the shape will be dominated. The main methods to correct the proximity effect are dose modulation, pattern bias, GHO ST, software synthesis and so on.
Since the actual layout of the device layout is relatively simple, almost all straight-line graphics, and are far apart, so no proximity effect correction. In addition, a smaller accelerating electric field corresponds to a larger beam spot and proximity effect, and no changes have been investigated in this paper.
The final exposure optimization conditions were: an accelerating electric field of 50 keV, a beam of 50 pA, a scanning grid spacing of 12.5 nm, a focusing of the fourth lens with no correction of the proximity effect.
3. Effect of exposure dose
For positive resists, the residual glue thickness on the exposed pattern increases with decreasing exposure dose after development. UV3 electron beam exposure dose comparison curve shown in Figure 2 on the left axis.
At 50 keV, a beam current of 50 pA, a scanning grid spacing of 12.5 nm, electron beam exposure of the fourth lens, and CD-26 development for 1 min, the critical dose on the UV 3 topography was 18 μC / cm 2, The resist dose-sensitive contrast ratio of 2.84 (Definition 1 / (lo g10 Dc-log10 D0)).
On the other hand, exposure dose also affects the size of the pattern, which is caused by the electron beam proximity effect. The right axis of Figure 2 shows the relationship between the key figure size and exposure dose. Although smaller exposures are available at lower exposure doses, the edges of the pattern are poor and there may be some residual glue in the grooves; as the dose increases, the line width increases. Therefore, we need to optimize the exposure dose