The use of a semi-aqueous organic film stripper and residue remover that does not contain N-Methyl-2 pyrrolidone (NMP) is compared with current NMP-based chemistry.
By NIK MUSTAPHA and DR. GLENN WESTWOOD, Avantor Performance Materials, Inc. MARKUS TAN, JOACHIM NG, and YANG MING CHIEH, Philips Lumileds Singapore
Philips Lumileds collaborated with Avantor Performance Materials, a global manufacturer of high-performance chemistries, to evaluate one of Avantor’s post-etch residue remover and photoresist stripper products as a replacement for a current chemistry. Avantor’s J.T. Baker ALEGTM-368 organic film stripper and residue remover is an engineered blend of organic solvents and semi-aqueous compo-nents suitable for bulk photoresist removal and post-etch/ash residue and sidewall polymer removal. Designed to provide broad process latitude in terms of processing times and temperatures, ALEGTM-368 organic film stripper and residue remover is completely water soluble, requires no intermediate solvent rinse, and contains no hydroxylamine (HA), NMP, or fluoride elements.
The authors worked together to assess whether a change to Philips Lumileds’ process of record (POR), using this product, could be accomplished without impacting yield or device quality, and with the desired cost savings.
NMP replacement challenges
Pending changes in environmental, health, and safety regulations in key manufacturing locations around the world may prohibit the use of NMP-based post-etch residue and photoresist removal products in LED manufacturing. The shift can already be observed in Europe and in some parts of Asia and the United States, where companies are moving toward NMP-free manufacturing environments. In today’s competitive environment, it is vital for companies to find alternative chemistries that are not only effective and emphasize good performance, but also provide better cost of ownership. Philips Lumileds is taking a significant step to be part of this change.
Initial verification tests of NMP-free product
As part of the process verification, several wafers were used to check etch rate on critical substrates such as III/V Nitride, Al, Ag, and Au. These wafers were also used to verify the effectiveness of the ALEGTM-368 product to remove photoresist. Data were then compared with the current POR (TABLE 1).
It was important to confirm the effectiveness of the ALEGTM-368 product in stripping capability of negative photoresist. A wafer with 5μm thickness was used as an experiment. The wafer was dipped in the ALEGTM-368 product at 75°C followed by a water rinse step. To ensure uniformity of chemical performance, five locations were inspected by a scanning electron microscope (SEM) before and after treatment with the NMP-free product (FIGURE 1). Post-treatment images after dipping the wafer in the ALEGTM-368 product indicated that no resist remained on top of the metal surface (FIGURE 2). This supports the effectiveness of the ALEGTM-368 product; it is capable of stripping photoresist completely, without visible damage to the metal surface.
Resist stripping and residue remover verification test on pattern wafers
Further tests were conducted on pattern wafers comparing POR and the ALEGTM-368 product at 75 °C, for 30 minutes. Wafers were then cleaved and subjected to SEM inspection.
FIGURE 3. Post-treatment for POR material. No photoresist remained under high-magnification confocal microscope inspection. POR material showed good stripping capability on patterned wafers.
FIGURE 4. Post-treatment using the ALEGTM-368 product. No resist remained under high-magnification confocal microscope inspection. POR material showed good stripping capability on patterned wafers.
High-magnification images were obtained to verify cleaning performance and stripping capability of the ALEGTM-368 product and POR wafers. For top-view inspection, a high-magnification confocal microscope was used to verify complete removal of photoresist. Results are shown in FIGURES 3 and 4. Both the POR and the ALEGTM-368 product showed equal performance in terms of cleaning polymer residues and stripping photo resist on patterned wafers (FIGURES 5 and 6). The next critical step was to verify electrical performance for both the POR and the ALEGTM-368 product.
FIGURE 5. SEM images showing post-treatment for POR.
FIGURE 6. SEM images showing post-treatment for the ALEGTM-368 product.
Electrical performance for engineering lots
Wafers were sampled from several production lots before being split into two groups, one group using the baseline and the other using the ALEGTM-368 product. Both groups were processed in an automated tool following the recommended process condition at an operating temperature of 75°C and a processing time of 30 minutes. To achieve wafer uniformity, the tool was equipped with a mega-sonic function and recirculation to ensure effective cleaning of post-etch residues and stripping of negative photoresist.
After chemical treatment, the wafers were given an intermediate rinse using an IPA solvent to remove any remaining traces of the ALEGTM-368 product from the surface of the wafer. Without this step, chemical left on the surface of the wafer could cause corrosion, water marks, or other device defects. Wafers were then subjected to a QDR (quick dump rinse) to remove all remaining solvent on the wafers. This step normally takes five to ten minutes, with noticeable CO2 bubbling to serve as extra protection from corrosion of exposed metal. Finally, all wafers were subjected to a nitrogen dry for five minutes, a vital process since any remaining moisture could cause severe corrosion and impact electrical performance and final yield.
Once all process steps were performed, both groups were subjected to electrical tests to ensure the chips on the wafers were functioning well and within specifications. Results, as indicated in FIGURE 7, showed no significant differences in term of electrical performance for both the baseline and the ALEGTM-368 product. All wafers met specification and were subject to final yield probe.
Comparable Performance in Final Yield
The same production wafers which were processed using the ALEGTM-368 product at 75 °C were then subjected to final yield analysis and compared to current POR. There was slight improvement in the standard deviation for the ALEGTM-368 product when compared to baseline chemistry. Overall, both products showed comparable final yield at 98 percent (FIGURE 8).
Reduced Cost of Ownership
It is undeniable that operating cost is a major consideration in LED manufacturing. Prior to adopting the current POR chemistry, Philips Lumileds tried both HA-based and NMP-based chemistries. Using the HA-based chemistry, a pre-treatment process was needed to soften the photoresist prior to stripping, followed by a solvent intermediate rinse. A strip process with the ALEGTM-368 product eliminated this step and resulted in significant cost savings and increased throughput due to process simplification (TABLE 2).
The NMP-free ALEGTM-368 product was comparable to POR when tested in various steps of the LED manufacturing process, including: substrate compatibility on critical layers, electrical performance on actual device, and final yield. In terms of process simplification, use of the ALEGTM-368 product also showed similar technical benefits as POR, in which a significant reduction of the number of steps and chemicals used in the process leads to improved cost of ownership.
This collaboration demonstrates how a manufacturer can translate its commitment to environmental, health, and safety improvements and reduction of cost of ownership into the commercialization of a new cleaning process which can bolster its competitive position in the global LED manufacturing industry.
NIK MUSTAPHA is a Principal Applications Engineer, AvantorTM Performance Materials, Inc. MARKUS TAN is Chief Process Engineer, JOACHIM NG is Senior Manager-Process Engineering, and YANG MING CHIEH is a Process Engineer at Philips Lumileds Singapore. DR. GLENN WESTWOOD, Senior Research Scientist, AvantorTM Performance Materials, Inc.