The Novalak/DNQ material was such a great photoresist for IC processes that it had dominated the semiconductor industry for over a quarter of century. But guided by the Moore’s Law, the semiconductor industry had never stopped its pace to search for new light sources and new patterning materials. When deep UV photolithography technology came around at the end of the last century, demands for new photoresist material innovation became consensus for material development scientists/engineers. For this time, researchers from IBM made significant contributions for their outstanding work in developing the 3rd generation photoresist for 248nm DUV photolithography.
Novolak/DNQ has too much absorption at wavelength around 250nm so it was not a viable option for the DUV photolithography. In addition, at the time, the light intensity from mercury lamp for DUV lithography was too low that the traditional photo active compound (PAC) based photoresists can no longer meet the demand for the exposure step. In that regard, new photoresist materials with new patterning mechanism other than photo active compounds was in great need.
To circumvent the issue of low light intensity of the DUV light sources, chemical amplification from photo acid generator (PAG) became the winner for the new photoresist development. The concept of chemical amplification had then been studied for many years. Autocatalysis mechanism behind the chemical amplification was well understood in other fields but rarely used for photoresist formulations. For more details for the autocatalysis mechanism, the photo acid consumed during reaction is re-generated at the end of the reaction and therefore it could be used again for another reaction. So unless the catalyst is quenched, the acid generated could be used over and over. Therefore the efficiency for photo utilization is much higher for CA photoresists than that from traditional PAC-based photoresists. Chemical amplification has been an essential element for the success of almost all modern advanced photoresists. Several mechanisms including crosslink , polycondensation, de-protection (cleavage), and molecule rearrangement had been explored to formulate new photoresists using chemical amplification mechanism.
The pioneering work of Ito, Willson and coworkers from IBM led the search for new DUV photoresist material. They used poly (4-hydroxystyrene) or PHOST as the main resin and photo acid generators were chosen over photo active compounds after numerous studies. As shown in the following Figure 1, the acid from PAG decomposition was used to de-protect PHOST compounds. In more details, to introduce solubility contrast after exposure, the phenolic –OH group are protected by t-butoxycarbonyl (t-BOC). Upon exposure to acid, the protection group is uncapped, which introduces polarity change from a hydrophobic state to a hydrophilic state. As a result, the exposed area would dissolve in aqueous developer in a much faster rate than the film from the unexposed dark area. In this manner, the film in the dark area is left on the substrate after development.
Figure 1. De-Protection Reaction of t-BOC-PHOST.
In the early days, the t-BOC-PHOST photoresist suffered from skin (T-top) and post exposure bake (PEB) delay issues. Many material scientists/engineers had tried to understand the issues. Material improvement to address the issues had been the main tone for DUV photoresist development. A common cause was later identified for the issues by scientists from IBM. After careful design and execution of experiments, they found that an insoluble surface layer was formed after PEB from a trace amount of airborne basic substances, which led to film skin and PEB delay issues. After the causes were identified, activated carbon filter to purify surrounding atmosphere and protective coating had been explored. Both were demonstrated to be effective in solving the problem. Improvements in the process were thus made and enforced in every semiconductor fabs. But these improvements based exposure tool modification or extra processes by a protection layer are not preferred, at least by material development scientists.
Many material scientists still tried to solve the problem by improving material formulation. Eventually, Environmentally Stable Chemical Amplification Positive (ESCAP) resist was developed at IBM. The new ESCAP photoresist was based on poly-hydroxystyrene (PHOST) and tert-butyl methacrylate of high glass transition point. The basic concept was to use higher PEB temperature to reduce film free volume for this resin due to its excellent thermal stability. With less free volume, the diffusion of airborne bases into the film was limited and skin/PEB delay issues were therefore minimized.
Many different DUV 248nm photoresists have been marketed but the variation among them is usually not as big as manufacturers claimed. The poly-4-hydroxystyrene polymer backbone and de-protection mechanism from chemically amplification are the key for the success of DUV photoresists. Some product differential factors include capping agents, PAGs, and/or surfactants, etc.. The design concept using PAGs based chemical amplification have been carried over to the next generation of DUV photoresists (@193nm) development, which we will discuss in the next blog.