Optics and Lasers in Engineering 49 (2011) 536–541
Contents lists available at ScienceDirect
Optics and Lasers in Engineering journal homepage: www.elsevier.com/locate/optlaseng
Laser plasma shockwave cleaning of SiO2 particles on gold film Yayun Ye n, Xiaodong Yuan, Xia Xiang, Wei Dai, Meng Chen, Xinxiang Miao, Haibing Lv, Haijun Wang, Wanguo Zheng Research Center of Laser Fusion, China Academy of Engineering Physics, Mianyang 621900, PR China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 5 June 2010 Received in revised form 25 November 2010 Accepted 13 December 2010 Available online 7 January 2011
A Nd:YAG laser (1064 nm) induces optical breakdown of the airborne above the gold-coated K9 glass surface and the created shockwave removes the SiO2 particles contaminated on the gold films. The laser cleaning efficiency has been characterized by optical microscopy, dark field imaging, ultraviolet–visible– near infrared spectroscopy, Fourier transform infrared spectroscopy, X-ray photoelectron spectroscopy and the Image-pro software. The relationships between removal ratio and particle position and laser gap distance have been studied in the case of single pulse laser cleaning. The results show that the 1064 nm laser induced plasma shockwave can effectively remove the SiO2 particles. The removal ratio can reach above 90%. The effects of particle position and laser gap distance on the cleaning efficiency are simulated for the single pulse laser cleaning. The simulated results are consistent with the experimental ones. & 2010 Elsevier Ltd. All rights reserved.
Keywords: Laser cleaning Laser plasma shockwave cleaning Gold-coated K9 glass SiO2 particles Removal ratio
1. Introduction Micron and sub-micron particles can be removed by laser cleaning without damage to the substrate [1–3]. Laser cleaning has widely been investigated and applied in many fields, e.g., microelectronics, rust removal, paint removal, restoration of cultural relics [3–7]. According to the different mechanisms of laser cleaning, particle removal mechanisms can be classified as (a) dry laser cleaning mechanism [8–11], (b) steam laser cleaning mechanism [12,13], (c) Matrix laser cleaning [14], and (d) plasma shockwave [15–18]. Dry laser cleaning is based on the direct interaction between the incident laser light and the substrate to be cleaned. It has been assumed that the acceleration forces produced by the thermal expansion of the substrate or the particle due to absorption of the applied irradiation are sufficient to remove the particles. However, the detailed analysis shows that the local field enhancement effects underneath the particles can result in the local ablation of the substrate. This effect is undesirable for the particle-removal process [8,14]. For the steam laser cleaning (SLC), a liquid, in most cases water or water/alcohol mixture, is applied to the sample prior to the laser pulse. The mechanism responsible for particle ejection in SLC is believed to be the explosive evaporation of the liquid film in the vicinity of the particle/substrate interface. Nevertheless liquids are not wanted in the semiconductor industry due to various reasons, e.g., the risk of watermark formation or possible damage of structures due to capillary forces [8,14]. Matrix
n
Corresponding author. E-mail address:
[email protected] (Y. Ye).
0143-8166/$ - see front matter & 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.optlaseng.2010.12.006
laser cleaning is based on the use of the matrix. In most cases, the solid CO2 matrix is applied to the sample surface and the particles are embedded in the solid matrix before the laser pulse. After the laser shot a small part of the matrix undergoes a phase transition forming an expanding vapor layer acting on the particles. In addition, the major part of the matrix as an intact layer leaves the surface, resulting in a drag force acting on the particles. In this way, the particles have been removed [14]. The disadvantage of this method is additional contaminations resulting from the matrix. Laser plasma shockwave cleaning have been presented to avoid the disadvantages. The laser irradiation mode is different from the other laser cleaning methods, where laser is utilized to induce optical breakdown of the airborne above the substrate surface to remove particles on the surface. So the plasma shockwave cleans the surface not only without damage to substrate but also without any additional contaminations. In the high-power laser facility, surface contaminations on the optics will worsen the laser beam quality and damage the optics. Some gold-coated optics, e.g., the gold-coated reflectors and the gold-coated gratings, are expensive and their delicate structure tends to be damaged by the conventional cleaning techniques. Hence it is necessary and urgent to find a cleaning technique to remove the pollutions without damage. As a non-contact technique, laser plasma shockwave cleaning is suitable to clean the delicate optics. At present, the laser cleaned objects in the most reports [8–18] are silicon wafer, glass, printed circuit board, and so on. To our knowledge, there is no report about laser cleaning of gold-coated optics. Due to the complex and delicate structures of the gold-coated grating, the gold-coated reflector is selected for investigation in this work. The results will be helpful for the
Y. Ye et al. / Optics and Lasers in Engineering 49 (2011) 536–541
cleaning of the gold-coated gratings. In this work, laser plasma shockwave cleaning is utilized to remove SiO2 particle contaminations on the gold-coated K9 glass. A Nd:YAG (1064 nm) laser was used to produce laser plasma shockwave. The laser cleaning effect in the scanning mode has been characterized by optical microscopy, dark field imaging, ultraviolet– visible-near infrared (UV–vis-NIR) spectroscopy, Fourier transform infrared (FT-IR) spectroscopy, X-ray photoelectron spectroscopy (XPS). The single pulse laser cleaning experiments have been performed to understand the relationships between removal ratio and particle position and laser gap distance, which has also been simulated by a simple theoretical model [17].
2. Experimental The gold-coated K9 glasses were prepared by the magnetic sputtering. The 2 mm-thick gold films were deposited on the 30 30 mm K9 glass. The preparing procedure of the contaminated samples is described as follows: the SiO2 particles (15 nm in diameter) were dispersed in alcohol and vibrated for 30 min in an ultrasonic cleaner. The solution was then agitated for 2 h by a magnetic stirrer to obtain a uniform suspension. After that, the gold-coated K9 glass samples were immersed in the solution to obtain the SiO2 contaminated samples by the Czochralski method with an average velocity 200 mm/min. The process was repeated for 5 times for each sample. The schematic diagram of the cleaning process is shown in Fig. 1. The cleaning experiments were performed using a Q-switched Nd:YAG laser operated at 1064 nm with a pulse width of 10 ns. The laser beam was parallel to the gold-coated K9 glass surface and focused above the area to be cleaned by a 200 mm focal length lens. The pulse energy used for laser cleaning was up to 600 mJ. When the airborne breakdown occurred due to the intense electric field induced by the focused laser pulse, the gas was ionized and rapidly heated, producing a shockwave at the focus of a laser beam. The airborne breakdown occurred only when the laser energy was above 70 mJ in this work. There was a big audible snapping sound during the sparking process. In the scanning mode, the contaminated sample was scanned twice and a region was cleaned by the laser with frequency of 1 Hz and energy of 120 mJ. The scanning speed was 1 mm/s and the laser gap distance was 1 mm. In the case of single pulse laser cleaning, some contaminated samples were cleaned at the laser gap distance of 0.5 mm and the other contaminated samples were cleaned at the different laser gap distances at the laser energy of 94 mJ. A Nikon optical microscope was used to observe and analyze the surface contaminations before and after cleaning. Dark field imaging was utilized to record the whole gold-coated samples. XPS, UV–vis–NIR and FT-IR spectrometers were used to obtain the elemental information, reflectivity or absorption spectra before and after cleaning, respectively. The Image-pro software was used to determine the number of particles on the surface to calculate the removal ratio.
Shock wave front Pulsed laser beam
Gap distance
SiO2 particle Substrate
Fig. 1. Cleaning schematic of SiO2 particles by laser plasma shockwave.
537
3. Results and discussions Fig. 2(a) shows the dark field image of the whole contaminated gold-coated K9 glass after local scanning. In the picture, there is an obvious boundary. The left of the sample is the laser cleaned area and the right is the un-cleaned area with contaminated particles on the surface. Fig. 2(b) and (c) are the 50 surface morphology images of the laser cleaned area and un-cleaned area, respectively. The contrast between two areas indicates that the cleaning effect is good. Fig. 2(b) shows that the particles are removed and the gold films are not damaged. Furthermore, the ablation of gold films is not observed in our experiments. Fig. 3(a) shows the UV–vis–NIR reflectivity spectra of the goldcoated K9 glass before and after cleaning. The three curves are the reflectivity spectra of the originally clean sample, contaminated sample and laser cleaned sample, respectively. From the figure it can be concluded that the reflectivity decreases after contamination. After cleaning, the reflectivity nearly returns to that of the originally clean sample. Fig. 3(b) shows the FT-IR spectra of the gold-coated K9 glass before and after cleaning. The three FT-IR spectra are attributed to the originally clean sample, contaminated sample and laser cleaned sample, respectively. The absorbance of the originally clean sample is near zero when wavenumber ranges from 500 to 4000 cm 1. There is an obvious absorption peak at 1093.6 cm 1 ascribed to Si–O–Si stretching vibration in the FT-IR spectrum of contaminated sample. However, the absorption peak disappears in the FT-IR spectrum after laser cleaning. Fig. 4(a) and (b) show the X-ray photoelectron spectra of the contaminated gold-coated K9 glass before and after cleaning, respectively. In the contaminated samples, there is an obvious peak ranging from 101 to 107 eV, which is due to Si2p3/2 energy level of SiO2 particles. The blue line shown in Fig. 4(a) is the background line of the spectrum. However, the XPS peak disappears in Fig. 4(b), indicating that the SiO2 particle contamination has been removed. Figs. 2, 3 and 4 confirm that the SiO2 particle contamination can be effectively removed from the gold-coated K9 glass surface by the laser plasma shockwave cleaning. The single pulse laser cleaning experiments was conducted and the curves of particle removal ratio versus particle position and laser gap distance are plotted in Fig. 5. The particle position is defined as the distance s between the projection point of laser focus and the center of the particle, as shown in Fig. 6. The laser gap distance is defined as the distance d between the laser focus and the substrate surface (Fig. 6). The Image-pro software was used to analyze and account the particle number in the micrograph. Fig. 7(a) and (b) are the 50 surface morphology of the contaminated sample and the analyzed picture by the Image-pro software. The red circles were labeled by the Image-pro software in Fig. 7(b). The Image-pro software can select the contaminated zones in the picture and count the total area of the selected zones. The removal ratio Z can be defined as Z ¼(1S / St) 100%, where St is the contaminated area before cleaning, and S is the contaminated area after cleaning. For reliability, the micrograph were measured and analyzed for several times to obtain the average particle numbers. Fig. 5(a) shows the experimental curve of removal ratio versus particle position at laser energy of 94 mJ and laser gap distance of 0.5 mm. The curve indicates that the removal ratio is above 90% when the particle position is less than 1.8 mm. The removal ratio begins to decrease when the position is 2.1 mm and comes down to 10% when particle position is 2.7 mm. Fig. 5(b) shows the experimental curve of removal ratio versus laser gap distance at laser energy of 94 mJ and particle position of 1 mm. The curve shows that the removal ratio reaches 91% when the laser gap distance is 0.5 mm and decreases to 10% when the laser gap distance is 3 mm. The removal ratio monotonously decreases with the increase in laser gap distance.
538
Y. Ye et al. / Optics and Lasers in Engineering 49 (2011) 536–541
50μm
50μm
Fig. 2. (a) A dark field picture of the gold-coated K9 glass after laser cleaning; (b) 50 surface morphology of the laser cleaned area; (c) 50 surface morphology of the un-cleaned area.
4. Theoretical model
100
0.008
On the base of the experimental results in the case of single pulse laser cleaning, the process of laser plasma shockwave cleaning is discussed by a simple model. Fig. 6 shows the force diagram representing the cleaning force acting on a partiallydeformed particle sitting on a solid surface. In the dry environment, the particle is attracted to the surface by Van der waals force Fvdw when the particle size is less than 50 mm in diameter [15]. The cleaning force Fs produced by plasma shockwave acts on the particle and is described in the literature [19]. If the pressure induced by the cleaning force is larger than the compression strength of substrate, the particles may get embedded in the substrate and the substrate may be damaged. In this work, the laser energy is selected slightly larger than the threshold of airborne breakdown to avoid damaging the substrate. Therefore, the condition that the particles are pushed into the substrate has been not considered in the theoretical model. According to the particle position on the substrate surface, there are two cleaning mechanisms, as discussed below. When the particles locate just below the projection point of laser focus, there are three forces acting on the particles [16], i.e., Van der waals force Fvdw, cleaning force Fs, and the elasticity Fe. The cleaning mechanism is the elasticity-removal model based on the elastic deformation of the particle induced by the cleaning force. Elasticity Fe and elastic potential energy Ee can be calculated by [16]
0.006
Fe ¼
4 qffiffiffiffiffiffiffiffiffiffiffiffi E rL3p ðtÞ, 3
ð1Þ
Ee ¼
8 qffiffiffiffiffiffiffiffiffiffiffiffi E rL5p ðtÞ, 15
ð2Þ
90
Reflectivity/%
80 originally clean sample contaminated sample laser cleaned sample
70 60 50 40 30 400
0.012
Absorbance/a.u.
0.010
500
800 600 700 Wavelength/nm
900
1000
1100
originally clean sample contaminated sample laser cleaned sample
0.004 0.002
Lp ðtÞ ¼ Lp0 þ f ðtÞ
0.000 2100
1800
1500
1200
900
600
Wavenumber/cm-1 Fig. 3. (a) UV–vis–NIR reflectivity and (b) FT-IR spectra of the gold-coated K9 glass before and after cleaning.
ð3Þ
where E* is Young modulus, r is the radius of the particle, Lp0 is the height of the particle deformation induced by Van der waals force, f (t) is the height of the particle deformation dependent on the time under the impact of cleaning force, Lp (t) is the sum of Lp0 and f (t). The particles can be removed when the elastic potential energy
Y. Ye et al. / Optics and Lasers in Engineering 49 (2011) 536–541
539
110 100 90 Removal ratio/%
80 70 60 50 40 30 experimental point fitting curve
20 10 0 0.0
0.5
1.0
1.5 s/mm
2.0
2.5
100 90
Removal ratio/%
80 70 60 50 40 30 20
experimental point fitting curve
10 0 0.0
0.5
1.0
1.5
2.0
2.5
3.0
d/mm Fig. 5. Experimental curves of removal ratio versus particle position (a) and laser gap distance (b).
Fig. 4. X-ray photoelectron spectra of the gold-coated K9 glass before cleaning (a) and after cleaning (b).
hy ¼ r sin a þ hx ¼ r cos a
induced by the elasticity is larger than the energy of adhesion induced by Van der waals force [16]. When the particles locate offset from the projection point of laser focus, the three forces also act on the particles, but the cleaning force is dominant and elasticity Fe is negligible. The cleaning mechanism is the rolling motion induced by the cleaning force. In this paper, the mechanism of rolling motion has been considered. In Fig. 6 the cleaning moment Mc and resisting moment Mr with respect to the contact point p can be calculated by [17] Mc ¼ Fx hy
ð4Þ
Mr ¼ Fy ðhx þ rc Þ þ Fvdw rc
ð5Þ
Fx ¼ Fs cos a,
ð6Þ
Fy ¼ Fs sin a,
ð7Þ
ð8Þ ð9Þ
The force Fx push the particle and the particle can be moved by a initial rolling motion. This happens when the cleaning moment of the shockwave is larger than the resisting moment associated with the adhesion forces, means Mc/Mr 41. d is the laser gap distance, s the particle position, rs the shockwave radius which is the distance between laser focus and the center of particle in the model, a the intersection between Fx and Fs, then Mc/Mr can be expressed pffiffiffiffiffiffiffiffiffiffiffiffiffi Fs cos a r sin a þ r 2 rc2 Mc ¼ ð10Þ Mr Fs sin aðr cos a þ rc Þ þFvdw rc tan a ¼
where the cleaning force Fs is resolved to x-axis force Fx and y-axis force Fy. hx and hy are the arms of the Fx and Fy forces, which is given by
qffiffiffiffiffiffiffiffiffiffiffiffiffi r 2 rc2 ,
d s
ð11Þ
It is possible to calculate the ratio of the two moments as a function of s and d. The environmental and material parameters are listed in Table 1 and Table 2, respectively. The particle radius is 7.5 nm and laser energy 94 mJ. According to the model, Fig. 8(a) shows the relationship between Mc/Mr and particle position when the laser gap distance is 0.5 mm. The inset is a magnified curve in the scale of 0–50 mm.
540
Y. Ye et al. / Optics and Lasers in Engineering 49 (2011) 536–541
Focal point
rs Shockwave
d Fs
Fs Fe
r
Fvdw
hy
Fx Fy α
r
hx
rc
s
p
Fvdw Fig. 6. Sketch showing the interaction between shockwave and one particle.
Table 1 Environmental parameters for calculations. Parameter
Value
Separation distance/nm Air density/(kg/m3) Specific heat ratio Dimensionless parameter Y (g)
0.4 1.237 1.4 1.03
Table 2 Material parameters for calculations. Parameter Hamaker constant/J Poission coefficients Young’s modulus/GPa
Fig. 7. (a) 50 surface morphology of the contaminated sample and (b) after analyzing by the Image-pro software.
It indicates Mc/Mr 41 when s ranges from 36 mm to 2.82 mm, suggesting that we cannot remove the particles located in the circular area with a radius of 36 mm, i.e., in the center of shockwave
Gold 40 10 0.42 79.5
SiO2 20
5 10 20 0.17 73
action region. Fig. 8(b) shows the curve of Mc/Mr versus laser gap distance when the particle position is 1 mm. It shows that the particle cannot be removed when d 41.9 mm and the cleaning efficiency increases with the decrease in laser gap distance. Fig. 5(a) indicates the radius of cleaning region is about 2.7 mm at laser energy of 94 mJ and laser gap distance of 0.5 mm. The simulated result shows Mc/Mr 41 when s ranges from 36 mm to 2.82 mm. It suggests that the radius 2.82 mm of cleaning region can be cleaned except the circular area with a radius of 36 mm. The simulated result is almost consistent with the experimental ones. The difference between the experimental and simulated results is that the center of shockwave action region with a radius of 36 mm cannot be cleaned. The reason is that the particles in the center region are removed by elasticity mechanism. Fig. 5(b) indicates that the removal ratio reaches 90% when the laser gap distance is 0.5 mm and decreases to 10% when the laser gap distance is 3 mm at laser energy of 94 mJ and particle position of 1 mm. It means that the particles cannot be removed when d is larger than 3 mm. The simulated results indicate that the particles cannot be removed when d is larger than 1.9 mm. The simulated and experimental data have similar tendency. The difference in the d-values probably because the material parameters used for calculation are the constants at room temperature. However, the parameters depend on temperature, which is not considered in the simulations.
Y. Ye et al. / Optics and Lasers in Engineering 49 (2011) 536–541
541
gold-coated optics, e.g., the gold-coated reflectors and gratings, which will be investigated in the future work.
Acknowledgment This study was supported financially by the Hi-Tech Research and Development Program of China (no. 2009AA8044005). References
Fig. 8. Relationships between Mc/Mr and particle position (a) and laser gap distance (b).
5. Conclusion The airborne breakdown induced plasma shockwave can effectively remove the SiO2 particles contaminated on the gold-coated K9 glass. Optical microscopy, dark field imaging, UV–vis–NIR and FT-IR spectroscopes, and XPS results have confirmed the cleaning effect. For the single pulse laser cleaning, the removal ratio can reach above 90%. The simulated results are consistent with the experimental ones. Based on the results, the laser plasma shockwave cleaning technique would be useful for the cleaning of other
[1] Imen K, Lee SJ, Allen SD. Laser-assisted micron scale particle removal. Applied Physics Letter 1991;58:203–305. [2] Zapka W, Ziemlich W, Tam AC. Efficient pulsed laser removal of 0.2 mm sized particles from a solid surface. Applied Physics Letter 1991;58:2217–9. [3] Tam AC, Leung WP, Zapka W, Ziemlich W. Lasercleaning techniques for removal of surface particulates. Journal of Applied Physics 1992;71:3515–23. [4] Pasquet P, Del Coso R, Boneberg J, Leiderer P, Oltra R, Boquillon JP. Laser cleaning of oxide iron layer: Efficiency enhancement due to electrochemical induced absorptivity change. Applied Physics A: Materials Science & Processing 1999;69:727–30. [5] Mosbacher M, Chaoui N, Siegel J, Dobler V, Solis J, Boneberg J, Afonso CN, Leiderer P. A comparison of ns and ps steam laser cleaning of Si surfaces. Applied Physics A: Materials Science & Processing 1999;69:331–4. [6] Daurelio Andriani SE, Catalano IM, Albanese A. Laser re-cleaning of a bronze age pre-historic dolmen. in: Proceedings of SPIE 2007;6346:634635. [7] Ostrowski Roman, Marczak Jan, Jach Karol, Sarzynski Antoni. Selection of radiation parameters of lasers used for artworks conservation. in: Proceedings of SPIE 2003;5146:99–107. [8] Chaoui N, Solis J, Afonso CN, Fourrier T, Muehlberger T, Schrems G, Mosbacher M, B¨auerle D, Bertsch M, Leiderer P. A high-sensitivity in situ optical diagnostic technique for laser cleaning of transparent substrates. Applied Physics A: Materials Science & Processing 2003;76:767–71. [9] Curran C, Lee JM, Watkins KG. Ultraviolet laser removal of small metallic particles from silicon wafers. Optics and Lasers in Engineering 2002;38: 405–15. [10] Lee JM, Watkins KG. In-Process monitoring techniques for laser cleaning. Optics and Lasers in Engineering 2000;34:429–42. [11] Song WD, Hong MH, Lu YF, Chong TC. Laser cleaning of printed circuit boards. Applied Surface Science 2003;208-209:463–7. [12] Frank P, Lang F, Mosbacher M, Boneberg J, Leiderer P. Infrared steam laser cleaning. Applied Physics A: Materials Science & Processing 2008;93:1–4. [13] Kruusing A. Underwater and water-assisted laser processing: Part 1—general features, steam cleaning and shock processing. Optics and Lasers in Engineering 2004;41:307–27. [14] Graf J, Luk’yanchuk BS, Mosbacher M, Hong MH, Chong CT, Boneberg J, Leiderer P. Matrix laser cleaning: a new technique for the removal of nanometer sized particles from semiconductors. Applied Physics A: Materials Science & Processing 2007;88:227–30. [15] Lee JM, Watkins KG. Removal of small particles on silicon wafer by laser-induced airborne plasma shock waves. Journal of Applied Physics 2001;89:6496. [16] Zhang P, Bian B, Li ZH. Ejecting removal of particles in laser-induced plasma shockwave cleaning. Chinese Journal of lasers 2007;34:1454–5. [17] Lim H, Jang D, Kim D, Lee JW, Lee JM. Correlation between particle removal and shock-wave dynamics in the laser shock cleaning process. Journal of Applied Physics 2005;97:054903. [18] Kim T, Lee JM, Cho SH, Kim TH. Acoustic emission monitoring during laser shock cleaning of silicon wafers. Optics and Lasers Engineering 2005;43: 1010–20. [19] Lammers N, Bleeker A. Laser shockwave cleaning of EUV reticles. in: Proceedings of SPIE 2007;6730:67304P.