Developing a resist specific to EUV
2018-07-10 10:21:21
Multi-Trigger chemistry, which is designed specifically for EUV, creates a high- chemical gradient at pattern boundaries, significantly reducing blurring and improving line-edge roughness to reduce the RLS trade off.
BY DAVID URE, ALEXANDRA MCCLELLAND and ALEX ROBINSON, Irresistible Materials, Wellesley, MA and Birmingham, U.K.
The semiconductor industry has invested billions of dollars to develop extreme ultraviolet (EUV) lithography and high-volume deployment of the technology is imminent. However, EUV lithography is not yet a complete solution. Most notably, new photoresist materials that enable the full benefits of EUV have yet to be developed.
While incremental modifications of incumbent �chemically amplified resists� will be used for the planned initial EUV introduction in 2019, there are presently no clear solutions that address the industry feature size targets, defectivity requirements, and sensitivity needs for 2020 and onwards. This is a significant concern and continues to cast a shadow over the industry�s long anticipated switch to EUV lithography. Indeed, the lack of a suitable resist for EUV lithography is now one of the biggest problems faced by the semiconductor industry.
What makes a good resist?
The critical performance parameters for any successful resist are: 1) Resolution (R): How narrow the lines on a microchip are, 2) Line-edge roughness LER (L): How �wobbly� the lines are; and 3) Sensitivity (S): How small a dose of radiation is required (how quickly the pattern can be formed). These performance metrics are known as the RLS targets, and they are set out in the ITRS. For a given material, these metrics have a conflicting relationship (one can only be improved at the cost of another): The �RLS tradeoff�. For a given material, improving one or two of the metrics leads to a loss in the third. To improve the RLS tradeoff, it is necessary to move to a new RLS graph. This can only be done by changing the resist material as illustrated in FIGURE 1.
In addition to the primary RLS targets, there are a series of critical secondary peformance metrics a commercially successful resist system needs to address, including the ability to pattern with extraordinarily low level of defects, high durability in the post processing steps, ultra-low contamination levels and wide process latitude.
The limitations with current state-of-art resist technology
Existing state-of-the-art photoresists are polymer- based platforms known as Chemically Amplified Resists (or CARs). The original CAR was based on a poly(hydroxystryene) chain with acid-labile tBOC protecting groups on the phenols, mixed with a photoacid generator. The photoacid released upon light exposure diffused through the polymer matrix catalytically removing the protecting groups, leading to a strong change in the solubility. While modern chemically amplified resists have increased in complexity, often using proprietary co-polymers with multiple functional units to address etch durability, adhesion and other properties, the core mechanisms of patterning have remained the same as the original CAR technology.
Such materials have demostrated significant design flexibility to address the evolving needs of the lithog- raphy industry. However, as feature sizes have continued to shrink, the diffuse nature of the acid � required for high senstitivity � has hampered resolution, and the acid quenchers, added to address this, have driven defects and roughness up. These limitations have risen to the fore as the industry prepares for the introduction of EUV lithography and the targeted feature sizes are increas- ingly incompatible with CAR technology.
Solving the EUV resist problem?
Given the limitations of polymer-platform photoresists originally developed for 193nm lithography, as the industry prepares for EUV introduction, the approach to photoresist development is being challenged. Indeed, device manufacturers and scanner suppliers have urged the photoresist suppliers to consider novel approaches to design photoresist systems specifically to meet the needs of EUV lithography.
One of the new photoresist platforms that has risen to prominence has been given the name �molecular resist� because it represents a departure from polymer- based photoresists to formulations based around �small molecules.� Originally developed to reduce the chemical��pixel� size of the resist, this platform has demonstrated promise in reducing line-edge roughness, but until recently has not fulfilled its early promise in EUV.
Another novel approach has been the development of metal-oxide resist platforms. These have demonstrated a compelling combination of high resolution, and low-line edge roughness, and sensitivities have improved recently. However, like other contenders, these materials currently demonstrate high defects and face a hurdle due to concerns over the use of metals in a cleanroom environment.
Another leading new �EUV specific� resist system is being developed by Irresistible Materials Ltd (IM), a company headquartered in Birmingham, England. IM has developed a new approach to achieve high-resolution, high sensitivity, and a low LER resist called the Multi-Trigger Resist platform(MTR). MTRs comprise a small proprietary resin molecule; an MTR process compatible cross-linker; and (like a chemically amplified resist) a photo-acid generator (PAG). However, the novel Multi-Trigger chemistry creates a high-chemical gradient at pattern boundaries, significantly reducing blurring and improving line-edge roughness to reduce the RLS trade off (FIGURE 2).
In a Multi-Trigger material, resist exposure proceeds via a catalytic process in a similar manner to a chemically amplified resist. However, instead of a single photoacid causing a single deprotection event and then being regen- erated, the Multi-Trigger resist uses multiple photoacids to activate multiple acid sensitive molecules, which then react with each other to cause a single resist event while also regenerating the photoacids. Importantly, it is only when two complimentary activated molecules react with each other that the resist is exposed � a single activated molecule, which is not near another will quench the acid, and remain unexposed.
In areas with a high number of activated photoacids (higher dose areas, for instance at the centre of a pattern feature), resist components are activated in close proximity and the multi-step resist exposure reaction proceeds, ending with photoacids regeneration and thus further reactions, ensuring high sensitivity. In areas with only a low number of activated photoacids (lower dose areas, for instance at the edge of a pattern feature), the activated resist components are too widely separated to react and the photoacids are thus removed, stopping the catalytic chain. The Multi- Trigger resist creates an increase in the chemical gradient at the edge of patterned features and reduces undesirable acid diffusion out of the patterned area. FIGURE 3 and 4 illustrate how the Multi-Trigger approach departs from the traditional approach used in existing state-of-the-art resist systems (CARs).
How good is the MTR system and where is it in its development cycle?
The MTR system is presently in an advanced development phase. Results have already shown this system can match and exceed the performance capabilities of state-of-the- art CARs. Furthermore, the specific formulation of the MTR system can be tailored by changing the ratio of the components within the resist. To date, IM has demonstrated that the sensitivity of the resist can be varied from 12 mJ/cm2 to over 50 mJ/cm2, with the patterned resolution ranging from 20nm half pitch to under 16nm half pitch respectively, to meet varying lithographic requirements.
Some example data from the ASML NXE 3300 scanner at IMEC in Belgium is included for reference below. ASML�s NXE platform is the industry�s first production platform for extreme ultraviolet lithography (EUVL), using 13.5 nm EUV light, generated by a tin-based plasma source.
FIGURE 5 shows results for 20nm half-pitch lines patterned on a pitch of 40nm. At a dose of 44.5 mJ/cm2, the LER is 2.6nm. FIGURE 6 shows 16nm half-pitch lines patterned on a pitch of 32nm. At a dose of 38.5 mJ/cm2, the LER is 3.7nm (unbiased values). These LER values compare very favorably with existing state-of-the-art CAR resists modified for EUV lithography. Importantly, the MTR technology is at the very beginning of its optimization cycle, with significant further performance enhancements expected as the technology matures. To this end, IM is in the process of scaling operations to accelerate the optimization of the MTR system in preparation for commercial launch.
The roadmap to commercial readiness
Prior to commercial integration into a Fab, it is also critical to address the 'secondary' performance metrics previously discussed. It is these tests that often prove a stumbling block to progressing from a promising new material. For an SME such as Irresistible Materials, passing this testing is a challenge as often new infrastructure and a specialist, custom tool set is required to pass stringent tests such as contamination. A resist that meets all lithography criteria could still fail to be adopted if, for example, the solubility of the components has not be synthesised with the required solubility in common fabrication solvents which will be present in the waste system.
For IM�s MTR, a precipitation test using waste drain solvents passed the precipitation test with no precipitate optically visible. These results indicate that the IM resist can be used within a fabrication facility with no precipitation issues. The resist also passes outgassing requirements so that it does not contaminate the lithog- raphy tool. Furthermore, because the resist is not metal based, there are no inherent track contamination issues. Metallic ion migration is a key concern for advanced device manufacturers and IM has implemented several protocols to address metal ion related concerns -- the current contaminant metal levels are below 15ppb for each individual metal and will reduce further as production system are optimized.
Another major step in the commercialization roadmap is the ability to produce material in a quality controlled, high-volume manufacturing process at commercially competitive costs. To address this requirement, IM has established a partnership with Nano-C for the high- volume supply of IM�s proprietary resin molecule. Nano-C, Inc. is a leading supplier of specialist small molecules and has recently doubled the footprint at its Massachusetts site as preparations are made to scale production of the IM materials.
Looking towards the future
IM is targeting launch of its initial MTR products in 2020 (to address the industry N5 node),and is presently engaged in a variety of tests/trials with potential end-user and distribution partners as the resist system is optimized, scaled and readied for commercial release. However, IM also recognizes the potential of this resist system to go beyond N5 and has a clear pathway for addressing future industry nodes, to N3 and potentially beyond. Notable upgrade pathways from the gen 1 MTR include optimizing the metastable nature of the proton quenching, increasing opacity, reducing the number of components in the resist to reduce the impact of stochastics, and optimizing the ancillary process.