They presented at the SPIE Advanced Lithography Conference this year and I also interviewed them just before the conference. During my interview, I spoke to Mark Shepherd (CEO), Alex Robinson (CTO) and Warren Montgomery (VP of Strategy).
Currently there are two main approaches to EUV Photoresists:
- Metal based photoresists utilize small molecules to try to give good edge roughness and resolution. Metal based photoresists are generally not as sensitive as other approaches because a single photon only initiates a single reaction (quantum efficiency ~1).
- Chemically amplified photoresists (CAR) are the current workhorse photoresists for optical lithography and represent an incremental approach for EUV. CARs are very sensitive because a single photon generates photo acids that can catalyze hundreds to thousands of reactions (quantum efficiency >>1). CARs require a post exposure bake adding processing time.
Successful photoresists need to achieve good performance on the RLS triangle. The R is resolution, you must be able to print the required pattern. L is Line Width Roughness; a measure of how smooth lines are. LWR has been a major sticking point in EUV to-date where large exposure doses are required to meet acceptable LWR specifications. S is sensitivity, measured as the dose required to achieve R and L and once again to-date S is much lower than desired for acceptable R and L values.
One of the key issues with EUV has been exposure source power and even as source power has improved it is still necessary to achieve high sensitivity in EUV photoresists to enable high throughput on the exposure tools. With EUV exposure tool prices in the range of $100 million dollars throughout is particularly important. The high sensitivity also needs to be achieved at the same time as good resolution and line width roughness - the RLS triangle discussed above. The typical goal for EUV photoresists is a dose of <20 mJ/cm2 but to-date most acceptable patterns have required higher doses.
For CAR at small linewidths you must limit acid diffusion to achieve the desired resolution and this reduces the sensitivity making CAR similar to metal containing resists for sensitivity.
In a CAR, exposure creates photoacid. During a subsequent post exposure bake the acid diffuses and a catalytic reaction occurs where the photoresist polymer is deprotected against developer by the acid, see figure 1.
Figure 1. Chemically amplified photoresist mechanism.
IM has developed a novel photoresist that is polymer based and therefore avoids any contamination concerns that can come up with metal based photoresist. The photoresist is like other carbon based photoresists and won't need a lot of etch process redesign. The photoresist is what IM calls multi-trigger and has chemically amplified behavior (quantum efficiency >1) without being chemically amplified. In an exposed area the following occurs, see also figure 2:
- An acid reacts with a molecule making it activated but the molecule is still not developer soluble.
- A second acid reacts with a second molecule activating it.
- If the two activated molecules are in close proximity they interact and become developer soluble.
Figure 2. Multi trigger photoresist behavior in exposed area.
An interesting property of this reaction is it is self-quenching at the edge of a feature because the reaction is only complete if two nearby molecules are activated, see figure 3. The probability of two random photons occurring in close proximity is low and this could possibly make this photoresist less sensitive to Shot noise.
Figure 3. Multi trigger photoresist behavior at the edge of an exposed area.
The result is a very sensitive photoresist that has demonstrated doses in the tens of millijoules with very high contrast. High contrast should provide better LWR than lower contrast photoresists.
At this point the evaluation of this photoresist is in the early stages and has not yet been run on full field production EUV systems. IM is seeking partners to help complete the evaluation and commercialization of the photoresist.
In conclusion the IM photoresist represents an interesting option for EUV photoresist that may offer a better set of RLS trade-offs than current solutions. IM is claiming their photoresist is 30% to 100% more sensitive than current EUV photoresists. If the promise of lower doses with acceptable LWR and resolution is met the cost savings due to improved system throughput could be substantial.