A high-temperature batch-spray process for implanted resist stripping
06/01/2006
Photoresist stripping in IC manufacturing has become more challenging as the number of photoresist levels has increased, while at the same time allowable material loss and surface damage has decreased. Heavily implanted photoresist is especially challenging due to the dehydrogentated, amorphous carbon layer that forms on the surface. Improvements to the standard sulfuric acid/hydrogen peroxide (“piranha”) resist stripping process have been made in a batch spray system to achieve on-wafer temperatures above 200°C. This has enabled the wet stripping of implanted photoresist exposed to doses greater than 1x1015 ions/cm2. This capability can eliminate the need for ashing on a large number of implanted photoresist stripping steps, reducing surface damage and material loss, and also improving overall IC manufacturing cycle time.
Clean and efficient stripping of photoresist masking material is an important step in integrated circuit (IC) manufacturing that is repeated numerous times. A 90nm high-performance logic device may use over 30 photoresist masking steps [1]. The nature of the photoresist and its process history greatly affect its ability to be removed cleanly from the surface of the wafer. Patterned photoresist that has been used as a mask during ion implantation steps is particularly difficult to remove. High-dosage ion implantation (>1x1014 ions/cm2) causes the surface of the photoresist to become dehydrogenated and highly cross-linked, similar in nature to amorphous carbon [2]. The thickness of this layer varies with the dosage, energy, and species of the ion implantation.
A liquid mixture of sulfuric acid (H2SO4) and hydrogen peroxide (H2O2), also known as “piranha” or sulfuric-peroxide mixture (SPM), can be used to remove photoresist that is unimplanted or only lightly implanted. Photoresist is also commonly removed by ashing, whereby the resist is exposed to an oxygen plasma, or ozone, while the substrate is heated.
However, both standard piranha chemistry and ashing processes have limited ability to remove the amorphous carbon layer that forms on heavily implanted photoresist. For the ashing process, energetic ion bombardment is added at the beginning to remove the amorphous film layer [2]. But even if ashing is used to remove the entire photoresist layer, a post-ash clean (PAC) process with piranha is also needed to remove any remaining organic and metallic contamination. This PAC step often includes a final treatment with an ammonium hydroxide (NH4OH) peroxide (H2O2) mixture (APM), also known as standard clean 1 (SC1), which helps remove residual sulfur and provide low particle levels.
Recently, there has been an effort to reduce the amount of silicon and silicon oxide loss on the surface of the wafer during the manufacturing process. Loss of silicon during processing can lead to reduced electrical performance. Initial efforts have focused on reducing the amount of silicon and silicon oxide consumption during PAC processes [3]. This has been accomplished largely by reducing the concentration and temperature of the APM step.
It is recognized that ashing steps cause oxidation and disorder to the wafer surface and lead to increased material loss during the PAC process [2, 4]. In addition to reducing APM temperature, replacing dry ashing processes with all-wet piranha ones to remove ion-implanted photoresist can decrease the amount of material consumption in the IC manufacturing process. However, a roadblock to implementing the all-wet strategy is the difficulty in removing heavily implanted photoresist using current piranha processes.
When H2SO4 is mixed with H2O2, monopersulfuric acid (H2SO5) is formed. H2SO5 is commonly known as Caro’s acid, which reacts with H2O2 to form the radicals OH and HSO4. These radicals react with the carbon polymer chain in photoresist to eventually form CO and CO2 reaction products [5]. Current piranha processes are heated as high as 150°C to achieve sufficient reactivity and stripping rates.
The dehydrogenated, amorphous carbon layer that forms on photoresist under ion bombardment is much less reactive with OH and HSO4 radicals. This layer can be broken up by physical processes such as ion bombardment or the swelling of the underlying photoresist material. Another approach is to use the piranha mixture at higher temperatures in order to improve reaction with the amorphous carbon layer. If we assume Arrhenius behavior and a radical formation activation energy of 200kJ/mol, then the radical formation rate will increase by over 500 times with a temperature increase from 150°C to 200°C.
It is difficult to use piranha temperatures higher than 150°C in an immersion-type cleaning system, due to chemical decomposition and evaporation. However, using a batch spray system, on-wafer temperatures of over 200°C have been achieved in this work. Spray systems can achieve higher on-wafer temperature by taking advantage of the exothermic mixing reaction between H2SO4 and H2O2. Depending on concentrations and flow rates, temperature increases of up to 100°C can be achieved. It is not possible to take advantage of this mixing exotherm in a batch immersion system because the chemical bath must be used over a long period (8 hrs or more), so the temperature of the bath is set and maintained by the chemical recirculation system. Because a spray system dispenses chemicals very efficiently, fresh chemicals can be used without excessive chemical waste. With higher on-wafer chemical temperature, efficient stripping of heavily implanted photoresist has been achieved, and the productivity of stripping non-implanted photoresist has been significantly improved.
Equipment and procedures
The equipment used was a commercially available FSI batch spray system (Fig. 1) for processing 300mm wafers. The wafers are held horizontally in PFA process cassettes. The 300mm system can hold 2 cassettes with 25 product wafers each. The cassettes are placed on a PFA turntable that can be rotated at speeds up to 300 rpm. The cassettes and turntable are held in a sealed nitrogen purged chamber.
 Figure 1. The equipment in this work was a commercially available FSI batch spray system used to process 300mm wafers. |
Chemicals, rinse water, and nitrogen are dispensed from a central spray post that extends from the chamber lid. Fluids can also be dispensed from a side spray post, mounted in the wall of the chamber. Chemicals are mixed and diluted using a mixing manifold. Mixture ratios are set by the process recipe and are carefully controlled by precision flow controllers.
An infrared (IR) heater can be used to heat the chemicals before dispensing them into the chamber. To prevent degradation of H2O2 and to achieve maximum mixing temperature, H2O2 can be mixed with H2SO4 after the H2SO4 is heated by the IR heater. In this work, the IR heater was set at 95°C. The temperature of the solution as it flows from the wafer surface in the process chamber is monitored with a sidebowl temperature probe mounted in the chamber wall, as shown in Fig. 1.
Modifications to the hardware and chemical sequences were made to achieve higher on-wafer temperatures. In addition to sidewall chemical temperature measurements, wafer surface temperature was indicated by attaching a set of color-changing, heat sensitive indicators (Omega Engineering, TL-10) to a monitor wafer’s surface during processing.
To evaluate the effectiveness of the new stripping process, various photoresist layers were processed and evaluated. Both blanket and patterned Shipley UV6 photoresist were used at a thickness of 7000Å on 200mm wafers and were subjected to a variety of ion implant conditions. For initial screening and process optimization studies, small pieces of the wafers were attached to whole wafers during processing. Final optimization was carried out with 300mm wafers. Table 1 lists the different implant levels used. The chemicals used were 98% by weight ULSI grade H2SO4 and 30% by weight ULSI grade H2O2. Resist stripping effectiveness was evaluated by wafer inspection with an optical microscope using brightfield and darkfield illumination.
 |
null
Results and discussion
Figure 2 shows the chemical temperature differences, measured by the sidebowl temperature probe, between a standard process and the new process with hardware and sequence modifications. In the case of the standard process in which pre-heated H2SO4 is mixed with room temperature H2O2 immediately before dispensing onto the wafers, the chemical temperature measured at the chamber wall is in the range of 60-70°C during most of the process. Temperature indicators attached to the wafer surface indicate a maximum wafer surface temperature of about 90°C for this standard process. It is clearly seen in Fig. 2 that the new process can achieve much higher sidebowl chemical temperatures, reaching over 130°C. In addition, the wafer surface temperature indicators give a reading of over 200°C. This much higher wafer surface temperature during exposure to the piranha mixture enables removal of implanted photoresist.
The results of the high-temperature strip process on various implanted samples is shown in Table 1. Samples that were judged “clear” showed no residual PR in any area by brightfield or darkfield optical microscopy. Samples judged to be “99% clear” had no residue in the patterned area, but had a small ring of residue at the inner radius of the edge-bead removal region. All of the resist on the “90% clear” blanket wafer had been undercut and removed, but small fragments of the amorphous carbon layer had redeposited on the cleaned surface.
Ash-free, all-wet stripping can be implemented in IC manufacturing in a cost-effective manner, as illustrated in Table 2. Replacing the ash-wet sequence, which requires two process tools and an intervening storage queue, with a single wet process, can eliminate queue time, and overall manufacturing cycle time can be reduced from 80 minutes to 35 minutes. Considering the large number of implanted resist stripping steps, the overall IC manufacturing cycle-time improvement can be as much as 12 hrs.
 |
null
Conclusion
Ash-free, all wet stripping of 1x1015 ions/cm2 implanted DUV photoresist has been achieved in a commercially available batch spray system. This was accomplished through a combination of hardware and chemical sequence modifications that result in on-wafer chemical temperatures of over 200°C during chemical exposure. This results in implanted photoresist stripping with less damage to the wafer surface and less material loss than the standard combination of ashing and wet cleaning. By removing the need for ashing implanted photoresist, this new process can be implemented in IC manufacturing in a cost-effective manner and can improve manufacturing cycle time.
Acknowledgment
This article is based on a presentation and manuscript originally presented at the SEMICON Korea 2006 Technical Seminar.
References
1. 2003 International Technology Roadmap for Semiconductors, http://public.itrs.net, Executive Summary, Table 5, p. 55.
2. T. Bausum, M. DeSarno, G. Dahrooge, “Stripping High-dose Implanted Resist for 300mm Production,” Semicondictor International, Web Exclusive, June 2003.
3. F. Arnaud, H. Bernard, A. Beverina, R. El-Farhane, B. Duriez, K. Barla, D. Levy, “Advanced Surface Cleaning Strategy for 65nm CMOS Device Performance Enhancement,” Ultra Clean Processing of Silicon Surfaces VII, Solid State Phenomena, Vols. 103-104, pp. 37-40, 2005.
4. Lee M. Lowenstein, Srikanth Krishnan, George A. Brown, “Damage During Ashing: A Characterization of Several Modern Ashers,” Electrochemical Society Proceedings, Vol. 95-20, pp. 225-234, 1995.
5. Steven Verhaverbeke, Kurt Christenson, “Organic Contamination Removal,” in Contamination-Free Manufacturing for Semiconductors and Other Precision Products, ed. Robert P. Donovan, pp. 320-322, 2001.
Jeffery W. Butterbaugh is chief technologist for FSI International, 3455 Lyman Blvd., Chaska, MN 55318; e-mail [email protected].
Kurt Christenson is a senior member of the technical staff at FSI International.
Nam Pyo Lee is a member of the technical staff at FSI International.