Creating 193nm immersion resists with embedded top barriers
09/01/2007
Immersion technology helps extend the capability of ArF (193nm wavelength) photoresist to 45nm and smaller nodes by increasing the refractive index (n) of the medium between the photoresist and the projection lens, effectively increasing the numerical aperture (NA) and resolution of the optical system. For the current immersion technology, the medium between the photoresist and the projection lens is deionized (DI) water (n = 1.44 at λ = 193nm). Because of the direct water contact, photoresist components such as photo-acid generator (PAG), quencher base, and additives can leach into the water.
As a direct result of this leaching, the photoresist will likely perform improperly. The leached photoresist components may also cause contamination of the projection lens, changes in the water’s refractive index, and contamination of the water recirculation mechanism. It is also very important for defectivity reasons that the top surface of the film meets certain critical water contact angle (hydrophobicity) requirements.
For these reasons, usage of a separate “impervious” immersion topcoat layer was the early technological approach for creating a water barrier on top of a dry ArF photoresist to convert it to an immersion photoresist. The extra spin-coating step and the additional cost for the topcoat materials however can make the immersion topcoat a less desirable approach than a topcoat-free photoresist. However, redesigning the entire photoresist, with the incorporation of new hydrophobic monomers into the resist polymer and immobilization of the photoacids and quencher bases to leaching might compromise the imaging performance of the resist, and also delay the availability of production-ready immersion resists [1].
Thus we were highly motivated to discover a novel approach to a production-worthy immersion 193nm resist. We took the approach of chemically designing an internal self-assembling barrier material that creates a thin embedded layer and functions as a topcoat. This invention creates all the advantages of a separate spin-on topcoat without needing a separate spin-on layer, saving the user extra process steps and cost. In addition, this invention makes it very easy to convert existing dry 193nm resist chemical approaches into an immersion 193nm resist by simply adding the self-assembling barrier material as a trace additive to the original photoresist formulation.
The first public disclosure of this discovery was presented in a paper at the IEEE Workshop on Lithography on August 2, 2006 [2]. We view the “embedded barrier layer” as a key innovation in the development of practical 193nm immersion resists [3].
The embedded top barrier process
A desired amount of a suitable barrier material is added to a photoresist formulation to form a uniform solution with the photoresist components. During spin coating, the barrier material migrates to the photoresist film’s surface to form a self-assembled water barrier whose thickness is determined by the amount of the barrier material added to the formulation. In general, the required barrier thickness on a photoresist film needs to be 20-30Å to reduce the PAG and quencher leaching to an insignificant level; the required thickness, of course, also depends on barrier composition.
For barrier material to migrate to the photoresist surface during spin coating, the material must satisfy two criteria at the same time:
- It has a lower surface energy than the photoresist polymer (or polymers).
- It does not mix well with the photoresist polymer (or polymers).
Criterion 1 ensures the barrier material stays on the top surface to form a water barrier between photoresist and the water, whereas criterion 2 is essential for the phase separation to occur during the spin coating. The phase separation of a barrier material from a photoresist matrix is an enthalpy-driven phenomenon since entropy always drives materials to mix. The enthalpy of the phase-segregated barrier-photoresist system needs to be much lower than that of the mixed barrier-photoresist system to overcome the unfavorable entropy contribution for the phase separation to occur naturally.
The enthalpy of a system depends on the intermolecular interactions of different species or functional groups of the materials in the system. Stronger barrier-barrier and resist-resist interactions and a weaker barrier-resist interaction will ensure the phase separation. Fluorinated materials are known to have much weaker interactions with non-fluorinated materials than with themselves and to have lower surface energy so they are among a group of candidate materials for investigation as embedded top barrier materials.
Besides forming a water barrier on the top of a photoresist, the embedded top barrier also serves as a means of adjusting photoresist surface properties such as the static and dynamic contact angles, and particularly the receding and sliding angles. These angles are correlated with defect modes. With an EBL, the surface receding angle can easily rise to >70° or even >80°, which is desirable for avoiding watermark defects at high scan speed [4, 5].
Because of the lower surface energy nature of the embedded top barrier materials, a photoresist with embedded top barrier, in general, has a higher static contact angle (and advancing angle) with water than that of the photoresist without the barrier. At high scan speeds, the scanning process can trap microbubbles if the advancing angle is too high, leading to “bubble defects” [5]. We have observed that at scan speeds >500 mm/sec, when the static contact angle is greater than 90° and the advancing angle greater than 93°, bubble defects are likely to appear. The materials we developed for the embedded top barrier offer a higher receding contact angle and a lower sliding angle (or a lower hysteresis) while maintaining relatively lower static and advancing contact angles.
Figure 1 shows the receding and sliding contact angles as a function of the loading of a barrier material in a test photoresist. The surface receding angle reached the onset of plateau at a loading of <1% relative to the total solid of the formulation. This material forms an excellent barrier on the photoresist at 1.5% loading, which is capable of reducing PAG leaching to 1.45 × 10-13 mole/cm2 for 60-sec leaching from 1.2 1× 10-11 mole/cm2, a nearly 100× reduction. The hysteresis, defined as the difference between advancing and receding angles, is as low as 11°.
 |
Moreover, the novel embedded top barrier materials that we designed and synthesized contain the functional groups-including photo-switchable and developer-switchable groups-to reduce development defects. The photo-switchable functional groups ensure the alkaline developer solubility of the embedded top barrier materials in the exposed areas upon PEB, and the developer-switchable functional groups are able to lower the developer contact angle in the dark field, enhancing development uniformity.
The performance of embedded top barriers
Patternable embedded barrier materials. When we first began this embedded top barrier program, the immediate concern was whether the added barrier material would negatively impact the performance of the dry resist. With this concern in mind, we designed our initial embedded top barrier materials as being photo-switchable by incorporating acid labile functional groups into the barrier polymer, which just like resist polymers, will become developer soluble upon photoacid attack during PEB.
The static and dynamic contact angles of the photo-switchable embedded top barrier materials are unaffected with exposure prior to PEB. This property ensures the exposed areas on the resist surface in the overlapped scan path behave identically to the unexposed resist surface. In certain applications, such as double exposure scanning, this property will be of great benefit. With the deprotection of the acidic functional groups during PEB, the barrier material becomes more hydrophilic and soluble in alkaline developer. Table 1 shows the static contact angles of water and alkaline developer measured under different conditions on the surface of a resist that contains an embedded top barrier material.
Embedded barrier materials with developer switchable functional groups. Although the photo-switchable functional groups take care of the developer solubility of the barrier material in the exposed areas, the dark regions still maintain relatively higher static contact angle for both water and developer. We considered that the higher surface contact angle for developer is a potential cause of the “missing contact hole” defect. Fall-on particles and blob defects may also be more likely to adhere to the dark areas on the resist surface due to its hydrophobic nature. To reduce the chance of causing these defects with embedded top barrier materials in a resist, additional developer-switchable functional groups are added to the barrier polymers. The technique of switching the surface of a film from hydrophobic to more hydrophilic on contact with developer (in order to reduce defectivity) has been previously shown by Rohm and Haas in the design of bottom antireflective films (BARCs) [6].
The switching functional groups on the embedded barrier layer polymer have higher affinity to alkaline developer, which can lower the developer contact angle. If the polymer is designed appropriately, such groups will not significantly affect the water dynamic contact angles, especially the receding and sliding angles. In addition, the changes in polymer structure can affect the ability of the embedded barrier to prevent leaching. With these performance challenges in mind, we have identified a novel barrier material that demonstrates a much lower alkaline developer contact angle than that of DI water, while maintaining excellent water dynamic contact angles as shown in Table 2. A difference of 18° is seen between the static contact angles of DI water and the alkaline developer.
Lithographic performance. With the success of forming a water barrier and a means of adjusting contact angle for high scan speed immersion, the lithographic performance of the resist containing the embedded barrier layer remains the primary concern.
 Figure 3. A comparison of resist profiles without (left) and with (right) an embedded top barrier material. (75nm , L/S 1:1 patterns obtained under 0.75NA) |
In a parallel study comparing resists with and without an embedded top barrier material, we found that a resist with a barrier material performs essentially the same as the resist without the barrier in terms of line-edge and linewidth roughness (LER & LWR), mask error factor (MEF), and process window. This result is well within our expectation, since these embedded barrier polymers are chemically similar to existing resist polymers. From a lithographic point of view, these barrier materials behave as a part of the resist polymers in the resist film; the only difference is that they are distributed primarily on the resist surface.
Figure 2 shows that essentially the same LER and LWR were observed with and without the embedded barrier. In addition, the feature profile remains unchanged with the addition of the barrier material in the resist formulations, as seen in Fig. 3. Immersion lithographic images with 50nm 1:1 line:space resolution obtained under a high NA condition are shown in Fig. 4, using the topcoat-free resist product EPIC EB 5000 TM. As seen, very low LER and LWR were obtained for this resist-containing embedded top barrier material.
Eliminating defects
Watermark and bubble defects are immersion-process added defects, which depend strongly on resist surface properties such as static and dynamic contact angles in DI water. Receding angles >70° are desirable for avoiding watermark defects at higher scan speed. However, a static or advancing contact angle that is too high would likely cause bubble defects.
 |
In a study with an ASML 1400i immersion tool using a scan speed of 500mm/sec, we found that the bubble defects become a significant issue when the static and advancing contact angles are larger than 90° and 93°, respectively. The bubble and watermark defect counts are summarized in Table 3 for the resists containing different embedded barrier layers along with their corresponding static and dynamic contact angles.
Moving forward
As inventors of this embedded top barrier technique, we are pleased to see that this technique has been widely accepted in this industry. This approach provides a reduced cost for immersion materials (separately coated topcoat) and increased throughput by eliminating the need for coating a topcoat on a resist surface [3].
References
- S. Kanna, H. Inabe, Kei Yamamoto, S. Tarutani, H. Kanada, K. Mizutani, et al., “Study and Control of the Interfacial Mass Transfer of Resist Components in 193nm Immersion Lithography,” SPIE, Vol. 5753, pp. 40-51, 2005.
- P. Trefonas, G. Barclay, S. Caporale, C. B. Xu, D. Wang, IEEE Workshop on Lithography, Aug. 2, 2006, Prince Edward Island, Canada.
- Rohm and Haas Electronic Materials, US patent application US20070087286 A1, European patent application EP1720072 A1, and Japan patent application JP2006309245 A. Please contact Rich Hemond for licensing details at [email protected].
- Ching-Yu Chang, Da-Ching Yu, John C.H. Lin, Burn J. Lin, “Watermark Defect Formation and Removal for Immersion Lithography,” SPIE, Vol. 6154, pp. 437-444, 2006.
- Yayi Wei, Stefan Brandl, Frank Goodwin, David Back, “193nm Immersion-related Defects and Strategies of Defect Reduction,” Future Fab Intl., Vol. 22, 2007.
- Ching Yu Chang, D.C. Yu, J.H. Chen, John C.H. Lin, Burn J. Lin, James W. Thackeray, et al., “A Novel Switchable BARC (SBARC) and Process to Improve Pattern Collapse and Defect Control,” SPIE, 2006.
Deyan Wang is a project team leader working on ArF immersion lithography at Rohm and Haas Electronic Materials, 455 Forest Street, Marlborough, MA 01752 United States; e-mail [email protected].
Chengbai Xu is a project leader in 193nm advanced material development at Rohm and Haas.
Stefan Caporale is part of the Embedded Barrier Materials team in Rohm and Haas’s immersion lithography program.
Peter Trefonas is the director of R&D for Rohm and Haas Microelectronics; ph 508/229-7350, e-mail [email protected].