(August 24, 2010) — A method was presented three years ago for controlling ammonium sulfate (AS) haze by maintaining 193nm reticles in a low humidity environment. Since then, this approach has became an industry standard and is widely used in production fabs around the world. Based on analysis of practical applications in HVM fabs, this paper describes a successful approach to reticle haze control, outlines its critical elements and explains its limiting factors. The authors provide practical recommendations for lithography practitioners on haze control equipment selections and reticle management strategy development. Oleg Kishkovich, Tom Kielbaso, David Halbmaier, Entegris Inc.

The authors proposed a practical solution for critical in-fab reticle haze control solution three years ago [1,2]. This approach emphasizes that, in order to control haze, moisture (humidity) level in the reticle environment must be reduced, in addition to traditionally controlled molecular bases (ammonia, NH₃) and acids (sulfate). This approach has become an industry standard successfully implemented in fabs [3] and endorsed by both mask makers [4] and scanner manufacturers [5]. Though we are unable to discuss exact numbers, we can mention that overall improvement — measured in number of wafers exposed between consecutive reticle cleans (WBC) — is approaching in some cases two orders of magnitude [6], with memory fabs printing in some cases 200,000-300,000 wafers and logic fabs over 50,000 wafers without patterned side AS haze development.

With AS time-dependent reticle defects, organic haze is most commonly reported. It develops very fast (often after exposing less than 10 wafers), with an uncharacteristically large number (~10,000 or more) of small particles (<100nm) [5]. After exposing ~10,000 wafers, this type of haze permanently disappears. Arguably, but most often, it is attributed to the out-gassing of the pellicle and pellicle frame adhesives.

Mechanism of haze formation: Accumulation of chemical contamination on the surface of the mask

The presence of AS haze precursors (ammonium HN₄+ and sulfate SO₄– ions) or even ammonium sulfate molecules NH₄SO₄ on the surface does not mean that haze is present. These molecules need to move together and form crystals before haze is recognized.

Figure 1. Humidity, ammonia and SO2 entrain the reticle pod through diffusion at the door seal. Differences in pod design may account tenfold for differences in airborne molecular contamination (AMCs) accumulation rate; compare pods 1 and 2.

Reticle pods cannot be hermetically sealed and ambient impurities, including moisture, ammonia and SO₂, entrain the pod by diffusion around the pod’s door seal (Fig. 1). We measured that ambient ammonia and SO₂ diffuse into the commercial reticle pod at a rate equivalent to the ambient air inflow of 12cc/min and 5cc/min, respectively. At this AMC entraining rate, and considering realistic values for SO₂ concentration in a cleanroom at 0.3ppbv, it will take only ~1000 hours, or ~40 days of unprotected storage to bring enough SO₂ into the pod to meet the haze threshold concentration on the surface of the reticle, even if we assume a hypothetical case of the reticle coming from the mask shop without traces of molecular surface contamination.

While in transit between purge locations, the diffusion of ambient impurities into the pod around the door seal increases contamination levels inside the pod (Fig. 1). Accumulation of sulfate and ammonium from cleanroom ambient sources may be controlled by either reducing cleanroom ambient levels of AMC (achieved by using HVAC air chemical filters) and/or by improving the immediate microenvironment inside the pod while in transit. Placing chemical purifiers inside the reticle pod microenvironment (Fig. 2) effectively controls entrainment of cleanroom ambient AMC at much lower cost than the use of HVAC chemical air filters.

Figure 2. Purifier/getter installed on the door of the reticle pod effectively removes haze-causing impurities from the reticle microenvironment. The removal kinetics and efficiency are governed by the diffusion process, so it is important that purifier is installed in the closest proximity to reticle’s patterned side.

Interaction with UV light

The exact mechanism by which UV light initiates the AS haze growth has not been established yet. Shimada et al. [7],  reported that this process has a threshold characteristic, with the threshold value dependent on the level of the chemical contamination on the reticle surface. Rate of haze growth depends on the cumulative exposure to the 193nm light. It is one of the reasons why back side haze develops much faster than pattern side haze, and also explains why higher transmission masks are more prone to haze formation.

Gordon et al. [4], reported that AS haze does not develop in dark conditions without previous UV exposure, yet with certain levels of 193nm UV light exposure, haze growth will continue in dark conditions. The authors observed that AS haze develops on different surfaces both with 193nm and 248nm UV light exposures, as well as without light exposure at all  (the latter was detected at much higher levels of surface contamination).

Growth of haze particles

Unfortunately, there is not an exhaustive study available on the subject of the kinetics of the haze growth rate in the dark as a function of the reticle surface contamination and water vapor concentration (humidity). Gordon et al. [4], studied the increase of the haze particle counts as a function of the reticle’s environmental condition, varying SO₂, ammonia and humidity levels without quantifying or qualifying the surface contamination conditions. This impressively detailed study indicated that moisture exposure has a cumulative effect on reticle haze formation and confirmed that controlling water vapor concentration is a major factor in controlling ammonium sulfate growth. This study also revealed that even in polluted environments (5-10 times above ambient concentration), haze growth rate is substantially reduced by reducing humidity.

Models of the continuous growth of reticle haze particles at different rates depending on water vapor pressure (humidity) level allows us to understand how AS reticle haze severity progresses with advancement of technology nodes. Haze particle growth increases in mass, not size. Being proportional to particle volume, particle mass increases as the mass of the particle size cubed. Approximating the ratio of printable particle sizes between 65nm and 32nm lithography by a factor of two, the ratio of particle volumes will be the factor of eight. Under similar conditions, printable AS haze defects may develop eight times faster on a 32nm mask then on a 65nm mask. Further advancement of 193nm lithography may require more stringent control of reticle microenvironments, including more effective humidity management. 

Improving humidity control of the reticle environment

It takes a long time to purge the reticle pod to dry conditions (~10 hours) and even longer to dry-down the sub pellicle space. Humidity inside the pod decreases even after ~100 hours of purging. This process cannot be expedited by increasing the purge flow. Humidity inside the pod and the kinetics of pod microenvironment humidity is limited by the water out-gassing from the pod materials of construction, all of which adsorb considerable amounts of water (~1 gram water per pod). It takes a long time to dry the pod material down at room temperature.

The process of humidity recovery is much faster and its kinetics are controlled by diffusion of the ambient moisture around the seal between the door and the dome of the pod. This seal is not completely engaged; otherwise, existing automated reticle load-ports would have difficulty opening the pod. Minute details of the reticle pod design strongly affect the entraining rate of the ambient moisture inside the pod and the speed of the humidity recovery inside the pod (Fig. 1).

Reticle pods with chemical purifiers (Fig. 2) have a considerable advantage in humidity control. Purifiers act as a desiccant and slow down the moisture recovery in the pod when purge is interrupted (Fig. 3). Unlike traditional desiccants that require elevated temperatures to regenerate, getter/purifiers regenerate at room temperature and restore their capacity for moisture when the reticle pod is purged with XCDA.

Figure 3. Shown here is the recovery of humidity level inside the purgeable reticle pod after XCDA purge interruption. Maroon bars are with getter/purifier, blue bars are without getter/purifier.

Some important practical recommendations may be derived from this study:

• Continuous purge of the reticle pod with XCDA is required to keep reticles dry and clean. Intermittent purge schemes do not work;  
• Indirect purge reticle storage cabinets also provide poor humidity management;
• The time to move a reticle in the pod between purge locations should be minimized. The use of a properly designed getter/purifier will extend this time. It is important to implement proper reticle management to minimize the time when the reticle is in transit, moving between purged long term storage (reticle stocker) and scanner. 
• Use of bare reticle stockers may present a challenge from the standpoint of humidity management.

We identified areas of improvement in the reticle pod design as follows:

• Developing pods with low ambient entrainment rate (Fig. 1), achieving long-term storage humidity levels of <0.05% RH;
• Customized getters/purifiers to meet ambient AMC challenges of specific site, with target levels of contamination inside the pod of less than 0.05ppb of acids, bases and condensable organics without purge;
• Developing pod materials that are permeation resistant to water vapor and AMC; and
• Increasing the moisture-removing capacity of getter/purifier to provide a reticle microenvironment of <2% RH for two hours without purge.
   

Basic principles of haze control

At first, controlling haze can seem to be an arduous task given the limitations and restrictions that existing fabs are faced with today. Some of these limitations and restrictions include: established/fixed floor plans restricting the ability to add a new piece of equipment; increased cost control requiring further justification of expenditures; and implementation of new operating procedures. If we focus on the basics, it seems a bit more feasible to create a system that can adequately control haze. Vital elements to such a control system are listed below.

Control moisture: Keep moisture levels in the mask environment controlled and you can successfully keep haze from forming. 

Optimize mask routes and logistics: When a podded mask leaves the storage area, moisture begins to infiltrate the controlled environment quickly. Limiting the time a mask is absent from active purge is critical to preventing a haze occurrence. 

Figure 4. Example of successful of implementation of haze control equipment in the fab.

Placement of purge equipment: Cost effective and creatively designed pieces of equipment available today expand options available to fabs for controlling haze. See Fig. 4 for an example of an optimized reticle flow with proper environmental control stations.

Utilize performance-capable mask carriers: Mask carriers play an important role in haze control. Measurable differences in moisture levels can be seen from one carrier to the next (Fig. 1). The amount of time that moisture permeation is kept low within the pod equates to a longer period of time before contamination will occur when the mask is not actively purged.

Plan for the future: Contamination occurrences such as haze are likely to become more and more prevalent as technology progresses. For the future, plan now to incorporate purging of load-ports of process and metrology tools. Process and metrology tools with buffer capability should be outfitted with the capability to provide continuous dry gas flow.

Conclusion

Effectively controlling haze requires changes to current reticle handling practices. From the improved products that enable purging and purification of the reticle pod to the equipment used to purge the reticle pod when stored, effective haze control in the fab must follow some basic principles. 

Acknowledgments
Authors would like to acknowledge Yingkai Liu, Tony Tieben and Matt Reber (all from Entegris) for their valuable contributions.

References:
[1] O. Kishkovich, A. Grayfer, F. V. Belanger, "Contrarian approach to and ultimate solution for 193nm reticle haze, "Metrology, Inspection, and Process Control for Microlithography XXI, edited by C. N. Archie, Proc. of SPIE Vol. 6518 (SPIE, Bellingham, WA 2007) 65183R.

[2] O. Kishkovich, D. Halbmaier, X. Gabarre, B. Grenon, J. Lo, A. Lam, T. Chen, "A practical solution to the critical problem of 193nm reticle haze," Photomask Tech. 2007, edited by R. J. Naber, H. Kawahira, Proc. of SPIE Vol. 6730 (SPIE, Bellingham, WA 2007) 67301A.

[3] W. Goodwin, M. Welch, B. Laquidara, O. Kishkovich, A. Habecker, "More on practical solutions to eliminate reticle haze and extend reticle life in the production environment: specially designed RSPs, internal POD purifiers, and XCDA purged reticle stockers," Optical Microlithography XXI, edited by H. J. Levinson, Mircea V. Dusa, Proc. of SPIE Vol. 6924 (SPIE, Bellingham, WA 2008) 69244N.

[4] J. Gordon, D. Chan, D. L. Frisa, C. Weins, C. Chovino, J. Keagy, S. Mahoney, "Influence of environmental components on haze growth," Photomask and Next-Generation Lithography Mask Technology XIV, edited by H. Watanabe, Proc. of SPIE Vol. 6607, (SPIE, Bellingham, WA 2007) 660707.

[5] J. Gordon, M. Silova, B. Connolly, J. Huijbregtse, N. Maxim, L. Frisa, et al., "Contamination control for ArF photomasks," 25th European Mask and Lithography Conference," edited by U. F. W. Behringer, Proc. of SPIE Vol. 7470 (SPIE, Bellingham, WA 2007), 74700C.

[6] B. J. Grenon, O. Kishkovich, "Back side photomask haze revisited," Lithography Asia 2009, edited by Alek C. Chen, Woo-Sung Han, Burn J. Lin, Anthony Yen, Proc. of SPIE Vol. 7520 (SPIE, Bellingham, WA 2009) 752018.

[7] S. Shimada, N. Kanda, N. Takahashi, H. Nakajima, H. Tanaka, H. Ishii, et al., "A new model of haze generation and storage-life-time estimation for mask," Photomask and Next-Generation Lithography Mask Technology XIV, edited by H. Watanabe, Proc. of SPIE Vol. 6607 (SPIE, Bellingham, WA 2007), 660706.

Oleg Kishkovich received his MS in engineering physics from the Moscow Institute of Physics and Technology and holds a doctorate in chemical physics from the Moscow Institute of Chemical Physics. He is Director of Contamination and Defect Control Technology at Entegris Inc., 10 Forge Park, Franklin, MA 02038 USA; ph.: (1) 508 878-7492; email [email protected]; http://www.entegris.com

Tom Kielbaso received his BS degree in mechanical design and is Global Product Marketing Manager at Entegris Inc.

David Halbmaier received his BS degree in Industrial Technology from the U. of Wisconsin, Stout, and is Engineering Program Manager at Entegris Inc.