Laser cleaning of organic contaminants on a glass substrate using ns-pulsed Nd:YAG laser was studied. Laser cleaning is a new promising dry cleaning technology under developing as an alternative to the conventional wet cleaning technology in semiconductor and display industries. PR (photoresist) was...
Laser cleaning of organic contaminants on a glass substrate using ns-pulsed Nd:YAG laser was studied. Laser cleaning is a new promising dry cleaning technology under developing as an alternative to the conventional wet cleaning technology in semiconductor and display industries. PR (photoresist) was selected as a representative material of organic contaminants. The PR and the glass substrate are widely used in the manufacturing process of semiconductor and display products. Nd:YAG laser beam was irradiated to a PR coated glass wafer to investigate the performance and mechanism of the laser cleaning. The key experimental variables were laser fluence, number of irradiation pulses, irradiation directions and the average flow velocity in the cleaning chamber. For irradiation directions, the laser beam was irradiated to the front of the PR layer (forward irradiation, FWI) or through the back of the glass substrate (backward irradiation, BWI). The surface of the laser-irradiated specimen was examined by SEM and microscope to study the cleaning performance. The size and number characteristics of particles generated during the cleaning process were measured by OPC (optical particle counter), Aerosizer and SMPS (scanning mobility particle sizer). The particles were sampled by LPI (low pressure impactor) to investigate the morphology of them by SEM. The mechanism of cleaning was analyzed on the basis of the above-mentioned characteristics of particles. As a preliminary study, laser cleaning was conducted in an open environment (clean booth of class 10,000). The laser beam with the wavelength of 532 nm was ineffective for the cleaning of PR. After 2,500 pulses irradiation at 0.3 J/㎠, the PR was not removed from the glass substrate but only slightly melted because of the low absorptivity of PR at 532 nm. On the other hand, the PR layer was effectively removed by 355 nm laser beam due to its high absorption feature at 355 nm. So, all the experiments were conducted with the 355nm laser beam. The cleaning and damage threshold fluences with forward irradiation were 0.027 J/㎠ and 3.0 J/㎠. respectively. Those with backward irradiation were varied to 0.036 J/㎠ and 2.25 J/㎠, respectively. The cleaning results for various fluences ranging from 0.05 J/㎠ to 5.0 J/㎠ were studied. Particles from these experiments were measured by OPC. High sampling flow rate (1.0 cfm) of OPC was suited to the experiment in an open environment. The backward irradiation was much more effective for the PR cleaning than the forward irradiation. The concentration of large micron-sized particles with the BWI was higher than that with the FWI. These results suggested that the BWI had additional PR removal mechanism which was absent in the FWI. Experimental results and the 1-D temperature simulation results indicated that the most probable additional removal mechanism in the BWI was the blasting mechanism. Also, secondary contamination was observed on the nearby of the cleaned area along the suction flow. Secondary contamination means the re-deposition phenomenon of the vapor and particle which generated during the laser cleaning. To confirm the hypothesis of the blasting mechanism, the cleaning chamber was designed to eliminate the external disturbance on the characteristics of the particles and to evaluate the effect of the flow velocity. The concentration of the micro particles with BWI was higher than that with FWI. But the concentrations of nanoparticles with the FWI were much higher than that with BWI. Also the number concentration of both micro particles and nanoparticles with FWI varied significantly according to the flow velocity. Most particles generated by the FWI were aggregates which mainly consisted of primary particles of tens of nanometers. The tendency of number concentrations for the flow velocity and the morphology of particles in the FWI indicated that the most of PR were removed by evaporation. On the contrary, the variance of particle concentration with the BWI to the flow velocity was hardly notable. Particles of several micrometers with the BWI had irregular plate shapes which were not found in the samples by the FWI. These indicated that most micro particles in the BWI were the results of the direct interaction between the laser beam and the PR layer like the photomechanical ablation. Experimental results with the cleaning chamber could be logically explained only by the blasting hypothesis. It validated the existence of the blasting mechanism in the BWI and the absence of that in the FWI. Despite the clean gas flow, the secondary contamination was also found frequently in the cleaned area. To find the feasible prevention solution for the laser cleaning, numerical simulation was conducted. The results showed that the combination of thermophoresis and gravity could effectively prevent the deposition of debris in broad size range. When the temperature gradient of about 200 K/cm applied, particles larger than 10 nm were not re-deposited to the substrate in the numerical simulation. In experiment, the temperature gradient was limited to about 13 K/cm by constraints of the cooling system. Even with relatively low temperature gradient, the inspections by SEM and microscope showed that the deposition of debris prevented effectively. Also, both experiments in an open environment and the cleaning chamber, the size characteristics of particles generated by the laser irradiation showed the different tendency for the laser fluence from that in the laser ablation of bulk material. The amounts of the large particles increase with increasing laser intensity in that case. But in this study, the concentration of the large particles decreased with increasing laser intensity at the high fluence. This might be due to the thin layer feature of the target material. If the laser fluence was sufficiently high, the whole volume of the PR layer in the irradiated area could be heated above its boiling temperature, leading the enhancement of evaporation resulting in the increased generation of the small particle. The PR film on the glass wafer was successfully cleaned by the backward irradiation of nanoseconds pulsed Nd:YAG laser. The third harmonic beam (355 nm) was effective in removing the PR while the second harmonic beam (532 nm) was not because of the low absorption. The backward irradiation was much more efficient than the forward irradiation for the cleaning of the PR from the glass substrate. In search of dominant cleaning mechanisms, the particle characteristics and the PR temperature profiles were investigated. As a result, it was found that the PR layer irradiated by the BWI was mainly removed by blasting, an additional removal mechanism absent in the forward irradiation. However, the thermal evaporation played a significant role in the forward irradiation. In conclusion, the laser cleaning by the backward irradiation of 355 nm nanoseconds pulsed laser beam was outstanding in cleaning film-type organic contaminants on an optically transparent substrate.
Laser cleaning of organic contaminants on a glass substrate using ns-pulsed Nd:YAG laser was studied. Laser cleaning is a new promising dry cleaning technology under developing as an alternative to the conventional wet cleaning technology in semiconductor and display industries. PR (photoresist) was selected as a representative material of organic contaminants. The PR and the glass substrate are widely used in the manufacturing process of semiconductor and display products. Nd:YAG laser beam was irradiated to a PR coated glass wafer to investigate the performance and mechanism of the laser cleaning. The key experimental variables were laser fluence, number of irradiation pulses, irradiation directions and the average flow velocity in the cleaning chamber. For irradiation directions, the laser beam was irradiated to the front of the PR layer (forward irradiation, FWI) or through the back of the glass substrate (backward irradiation, BWI). The surface of the laser-irradiated specimen was examined by SEM and microscope to study the cleaning performance. The size and number characteristics of particles generated during the cleaning process were measured by OPC (optical particle counter), Aerosizer and SMPS (scanning mobility particle sizer). The particles were sampled by LPI (low pressure impactor) to investigate the morphology of them by SEM. The mechanism of cleaning was analyzed on the basis of the above-mentioned characteristics of particles. As a preliminary study, laser cleaning was conducted in an open environment (clean booth of class 10,000). The laser beam with the wavelength of 532 nm was ineffective for the cleaning of PR. After 2,500 pulses irradiation at 0.3 J/㎠, the PR was not removed from the glass substrate but only slightly melted because of the low absorptivity of PR at 532 nm. On the other hand, the PR layer was effectively removed by 355 nm laser beam due to its high absorption feature at 355 nm. So, all the experiments were conducted with the 355nm laser beam. The cleaning and damage threshold fluences with forward irradiation were 0.027 J/㎠ and 3.0 J/㎠. respectively. Those with backward irradiation were varied to 0.036 J/㎠ and 2.25 J/㎠, respectively. The cleaning results for various fluences ranging from 0.05 J/㎠ to 5.0 J/㎠ were studied. Particles from these experiments were measured by OPC. High sampling flow rate (1.0 cfm) of OPC was suited to the experiment in an open environment. The backward irradiation was much more effective for the PR cleaning than the forward irradiation. The concentration of large micron-sized particles with the BWI was higher than that with the FWI. These results suggested that the BWI had additional PR removal mechanism which was absent in the FWI. Experimental results and the 1-D temperature simulation results indicated that the most probable additional removal mechanism in the BWI was the blasting mechanism. Also, secondary contamination was observed on the nearby of the cleaned area along the suction flow. Secondary contamination means the re-deposition phenomenon of the vapor and particle which generated during the laser cleaning. To confirm the hypothesis of the blasting mechanism, the cleaning chamber was designed to eliminate the external disturbance on the characteristics of the particles and to evaluate the effect of the flow velocity. The concentration of the micro particles with BWI was higher than that with FWI. But the concentrations of nanoparticles with the FWI were much higher than that with BWI. Also the number concentration of both micro particles and nanoparticles with FWI varied significantly according to the flow velocity. Most particles generated by the FWI were aggregates which mainly consisted of primary particles of tens of nanometers. The tendency of number concentrations for the flow velocity and the morphology of particles in the FWI indicated that the most of PR were removed by evaporation. On the contrary, the variance of particle concentration with the BWI to the flow velocity was hardly notable. Particles of several micrometers with the BWI had irregular plate shapes which were not found in the samples by the FWI. These indicated that most micro particles in the BWI were the results of the direct interaction between the laser beam and the PR layer like the photomechanical ablation. Experimental results with the cleaning chamber could be logically explained only by the blasting hypothesis. It validated the existence of the blasting mechanism in the BWI and the absence of that in the FWI. Despite the clean gas flow, the secondary contamination was also found frequently in the cleaned area. To find the feasible prevention solution for the laser cleaning, numerical simulation was conducted. The results showed that the combination of thermophoresis and gravity could effectively prevent the deposition of debris in broad size range. When the temperature gradient of about 200 K/cm applied, particles larger than 10 nm were not re-deposited to the substrate in the numerical simulation. In experiment, the temperature gradient was limited to about 13 K/cm by constraints of the cooling system. Even with relatively low temperature gradient, the inspections by SEM and microscope showed that the deposition of debris prevented effectively. Also, both experiments in an open environment and the cleaning chamber, the size characteristics of particles generated by the laser irradiation showed the different tendency for the laser fluence from that in the laser ablation of bulk material. The amounts of the large particles increase with increasing laser intensity in that case. But in this study, the concentration of the large particles decreased with increasing laser intensity at the high fluence. This might be due to the thin layer feature of the target material. If the laser fluence was sufficiently high, the whole volume of the PR layer in the irradiated area could be heated above its boiling temperature, leading the enhancement of evaporation resulting in the increased generation of the small particle. The PR film on the glass wafer was successfully cleaned by the backward irradiation of nanoseconds pulsed Nd:YAG laser. The third harmonic beam (355 nm) was effective in removing the PR while the second harmonic beam (532 nm) was not because of the low absorption. The backward irradiation was much more efficient than the forward irradiation for the cleaning of the PR from the glass substrate. In search of dominant cleaning mechanisms, the particle characteristics and the PR temperature profiles were investigated. As a result, it was found that the PR layer irradiated by the BWI was mainly removed by blasting, an additional removal mechanism absent in the forward irradiation. However, the thermal evaporation played a significant role in the forward irradiation. In conclusion, the laser cleaning by the backward irradiation of 355 nm nanoseconds pulsed laser beam was outstanding in cleaning film-type organic contaminants on an optically transparent substrate.
주제어
#Nd:YAG 유리 Glass substrate
※ AI-Helper는 부적절한 답변을 할 수 있습니다.