Ultrasonic Sensor for Photoresist Process Monitoring
In older I-line and deep-UV lithography, in situ monitoring of photoresist processing was not required to meet design objectives. However, as feature size shrank below sub-quarter-micron, the resist materials became more sensitive to processing conditions during preexposure bake, post-exposure bake, and development. After photoresist has been spin-coated to form a submicron film on a silicon wafer, it is necessary to prebake the freshly spun layer of resist prior to lithographic exposure. The purpose of this step is to evaporate the solvent from the film as well as to relax the resin polymer chains into an ordered array. In order to accomplish this, the resist must be heated beyond its glass transition or softening temperature and held above that temperature during evaporation. The time of prebake is crucial because if the solvent isn’t fully evaporated, the resist contrast can be compromised. Along with this, the temperature of prebake is also important since the resist can decompose if it is heated past its decomposition temperature. In either case, the feature size and therefore the device size will not be as expected. A sensor to detect resist cure state across the wafer would be necessary to optimize the process for each new wafer. After the prebaked resist has been exposed, there is a postexposure bake (PEB) step designed to produce a uniform image. Heating the resist to the post-exposure bake temperature enables diffusion of the exposed photoactive compound (PAC) within the resist, smoothing the standing waves that result from interference effects. In the case of chemically-amplified resists, the post-exposure bake is especially important since the deprotection reaction and the diffusion of photoactive compound continue throughout the PEB. A measure of this diffusion and the large changes in film properties that accompany it would be useful in characterizing the bake and the effect of its parameters on the final critical dimension of the devices. Once the resist has been through the PEB, it is developed to etch the desired circuit pattern into the resist. It is crucial that all of the soluble resist be removed by developer while the insoluble resist remains to sustain the exposed pattern. Thus, prediction and control of the development rate and time to clearance are necessary for repeatable feature size and device operation.
We developed an ultrasonic sensor to monitor photoresist processing in situ during semiconductor manufacturing. This sensor is based in principle on the measurement of the acoustic reflection coefficient for a wave incident on the interface between silicon and a thin film. This reflection coefficient is dependent upon the elastic properties of the thin film. As with optical reflection, there can be constructive and destructive interference in a thin film that will change as the film thickness and properties change. If the film thickness is of the same order as the wavelength of the incident wave, then these interference changes manifest themselves as measurable changes in the amplitude and phase of the reflection coefficient. With this sensor, we monitored photoresist development, pre-exposure bake, and postexposure bake for the Shipley 1800 series I-line resists, and the pre-exposure bake of Shipley APEX-E deep-uv (DUV) resist was monitored as well. Development monitoring was achieved by measuring thickness changes in the resist as it was removed. Data regarding dependence of development rate on exposure dose was obtained for the I-line resist with exposure doses varying from 20 to 68 mJ/cm2: Measurements showed an increase in average development rate from 0.04 to 0.155 μm/s, with the rate leveling off at around 55 mJ/cm2: Pre-exposure bake monitoring results demonstrated the ability of the sensor to measure the glass transition temperature of the resist film during prebake as well as the ability to invert out the elastic constants of the film using reflection theory. The glass transition temperature (Tg) is an important parameter in both the pre- and post-exposure bakes and therefore could be useful in monitoring these processes. Results of pre-exposure bake Tg measurements are presented for both I-line and DUV resists. The glass transition temperature during prebake was found to be higher for the DUV resist than for the I-line series. The I-line resist post-exposure bake measurement of glass transition temperature confirmed the reported Tg of 118 C for the I-line novolac resin. The multiple uses of this sensor make it suitable for integration into a manufacturing setting. More detail on this work can be found in [1].
FIGURE 1. Measurement setup.
FIGURE 2. Illustration of the measurement technique.
FIGURE 3. Phase change versus temperature during prebake, 0.6 _m Shipley 1805 film, 750 MHz.
References
[1] S. L. Morton, F. L. Degertekin, S. Calmes, B. T. Khuri-Yakub, “Ultrasonic Sensor for Photoresist Process Monitoring,” IEEE Trans. on Semiconductor Manufacturing, vol. 12, pp. 332 - 339, Aug. 1999.
Acknowledgement
This work was supported by the Defense Advanced Research Projects Agency of the Department of Defense and monitored by the Air Force Office of Scientific Research under Grant F49620-95-1-0525.
