Appl. Phys. A 63, 435—439 (1996)

Complete removal of paint from metal surface by ablation with a TEA CO laser 2 Akira Tsunemi1, Koji Hagiwara1, Norihito Saito2, Keigo Nagasaka2, Yasuaki Miyamoto3, Osamu Suto3, Hideo Tashiro1 1 The Institute of Physical and Chemical Research (RIKEN), 2-1, Hirosawa, Wakoshi, Saitama 351-01, Japan (Fax: #81!48!462!4682, E-mail: tsunemi@postman riken.go.jp) 2 SUT, Science University of Tokyo, 1-3 Kagurazaka, Shinjukuku, Tokyo 162, Japan 3 Power Reactor and Nuclear Fuel Development Corporation (PNC), 4-33 Muramatsu, Tokaimura, Nakagun, Ibaraki 319-11, Japan Received: 1 April 1996/Accepted: 7 June 1996)

Abstract. For high-speed metal surface cleaning, we applied TEA CO laser pulses to ablate painted materials on 2 metal surfaces and examined the efficiency of removal under different surface and irradiation conditions. Surfaces treated with a line-focused laser beam were analyzed with an energy dispersive X-ray analyzer and inspected with optical microscopic observation. Although paints were selectively ablated from the metal surface, the cleaning efficiency was found to depend on surface conditions of substrates. An application of a small amount of dimethyl formamide was effective for completely removing of resin without scorching the surface. PACS: 42.55E, 82.65

Material processing using laser ablation has been widely studied and applied for etching, marking and thin-film deposition. Recently, the laser ablation surface cleaning has attracted much attention for several reasons. First, it is a dry process in air. Second, ablated substances are effectively evacuated and collected by suction and filtering. Third, remote operation is feasible by optical power delivery with mirrors or optical fibers. Fourth, there are no secondary contaminated materials. Fifthly, treatment rates can be scaled up by increasing the applied laser power. Such characteristics have wide applications ranging from paint removal of airplanes or ships to systematic decontamination in atomic energy plants. Most surface cleaning treatment papers have reported surface cleaning with excimer lasers [1—4], which are suitable for removing thin layers because of their low penetration depth. For paint removal, however, TEA CO 2 lasers have several advantages. In general, the absorption coefficient of paint materials of organic or inorganic compounds for CO laser (10.6 lm) is reasonably high com2 pared with that for the metal substrate. The different absorption coefficients for paints and substrate provide a critical diffrernce in laser ablation threshold fluences. In consequence, surface substances are ablated without

damaging to the exposed metal surface. Though the ablation threshold for a CO laser is usually higher than that 2 for excimer lasers, higher removal rates per pulse will be obtained by CO lasers because of their deeper penetraton 2 depths. Recently, high-power TEA CO lasers capable of high 2 repetition rates and long lifetime operation have been developed in combination with all-solid state exciters [5, 6] and laser gas recycling systems [7]. Since TEA CO 2 lasers with an average power of more than 1 kW have become available [8, 9], we report here on a study of surface cleaning using a 500 W average power TEA CO 2 laser. In a previous paper [10], we demonstrated laser surface removal of paint and rust from metal substrates. Most of the paint and rust was easily removed, but the removal was not perfect. The amount of material left on the surface is small but could be serious if it contains poison or radioactivity. In this paper, we investigated the dependence of residue on the surface roughness and laser fluences and demonstrated a complete removal of the residue using a slight amount of liquid assistance. 1 Experimental apparatus and procedure The source laser for surface cleaning is a high-repetitionrate TEA CO laser developed in RIken’s Molecular 2 Laser Isotope Separation (RIMLIS) program [11, 12] as a master oscillator for Master Oscillator and Power Amplifiers (MOPA) chain. The CO laser medium is excited 2 by an all-solid-state exciter incorporated with a magnetic pulse compressor. The maximum output energy is 500 ml/pulse and the average power reaches 500 W at a repetition rate of 1 kHz. The discharge cross section is 16]16 mm. The electrodes are placed in A wind tunnel with a flow velocity of 40 m/s, coupled with a gas recycling system of 1.5 m3/min regeneration rate. The experimental setup is the same as the one reported previously [10]. A sample placed on a movable stage was irradiated with a cylindrically focused beam. The focal length was 20 cm. At the focal point, the beam width was

436

Fig. 1a–d. Photographs of the exposed substrate surfaces of SUS and iron taken with optical microscope. (a1) and (b1) are the mirror-like polished and dull substrate surfaces of SUS-316L before painting, respectively. (a2) is the two-pulse irradiated surface painted with synthetic resin of PbCrO as a pigment. 4 (b2) is the surface after the eight-pulse irradiation. (c) and (d) are the surfaces of sandblasted and roll formed iron substrates. (c1) and (d1) are the photographs of the surface before painting. For the sandblasted plate, five pulses were needed to remove the paint as shown in (c2). For roll-formed plate, more than 15 pulses were requirede to expose the substrate surface

300 lm. A nozzle was instaled near the irradiated region to suck the plume at a pump rate of 1.5 m/min. For substrates, iron and SUS plates were prepared with polished and dull surfaces for comparison. A synthetic resin containing pigments of PbCrO was used by spray4 ing a 25 lm thick layer on the substrate. 2 Results and discussion Figure 1 shows optical microscopic photographs of SUS and iron substrate surfaces exposed by laser ablation. Figure 1 (a1) shows mirror-like polished surfaces of SUS before painting and Fig. 1 (b1), dull surfaces. The paint on the smooth surface was almost removed by two-pulse irradiation, although a thin layer of paint remained judging from the interference color in Fig. 1 (a2). In contrast, the dull surface required more than eight shots to remove most of the paint to the level of the smooth surface as shown in (b2). Laser paint removal rates thus depended on the substrate surface condition. c

Fig. 2a–d. Energy dispersive X-ray spectra of painted and exposed surfaces. Synthetic resin with pigments of PbCrO was sprayed on 4 the substrate of a rolled steel plate. The signal intensities of the PbMa transition correspond to the amounts of paint remaining on the substrate survace. a the substrate surface before painting. b painted surface before laser irradiation. c surface after 5 pulse irradiation. d surface after 15 pulse irradiation. The peak appeared at the left edge in b and c shows signals from CKa originated from the main component of synthetic resin

437

The same type of experiment was conducted for an iron sample surface. Because iron has a higher thermal conductivity than SUS, more pulses should be needed to remove paint materials from an iron surface. It is well known that thermal conductivity influences the ablation

Fig. 3. Ratios of the PbMa intensity to the FeKa intensity are shown as a function of number of irradiated pulses. Two kinds of substrates were used, a rolled steel and b polished SUS. Results after application of DFM are also presented in c

efficiency. Sandblasted and roll-formed iron plates were used as smooth and rough surface substrates. Figures 1 (c1) and (d1) show substrate surfaces before painting. Two pulse irradiation, which had removed paints from the polished SUS, was ineffective when applied to both iron plates. For the sandblasted plate, at least five pulses numbers required to remove paint residue as shown in (c2). For roll-formed plate, more than 15 pulses were required to attain the exposure of the bulk surface as shown in (d2). In order to quantitatively evaluate the removal efficiency of paint, an electron dispersive X-ray analyzer was used to measure paint residue. Figure 2 presents changes of X-ray spectra before and after laser exposure. As a marker, the PbMa transition signals, which came from PbCrO , were analyzed in relation to the FeKa signal 4 from the bulk surface. Figure 2(a) shows signals before painting, (b) signals of a painted surface before irradiation, (c) signals from the surface after 5 pulse irradiation, and (d) signals after 15 pulse irradiation. Before laser irradiation, no detectable FeKa signals are found because the electron probe beam in the EDX system penetrates the material surface with a depth of only 2 lm, which is one tenth the thickness of the paint layer. As the number of irradiation pulses increased, the PbMa intensity decreased and the FeKa intensity increased.

Fig. 4a–d. Photographs of laser-irradiated surface as a function of the laser fluences. Substrates were polished SUS, a 5.1 J/cm2; b 3.1 J/cm2; c 1.8 J/cm2; d 1.6 J/cm2

438

Fig. 5a–c. EDX spectra of painted iron surfaces. a before laser irradiation, b after 10-pulse laser irradiation, c the same as b but a slight amount of DFM was dropped on the surface after the fifth laser pulse

Removal rates are indicated in Fig. 3, in which the ratio of PbMa to FeKa intensities is shown as a function of number of irradiation pulses. Two surfaces, (a) rolled steel and (b) polished SUS, were compared. As shown in Fig. 2, the ratio for (b) reached 0.01 while (a) needed 15 shots. Further irradiation up to 50 pulses, however, did not reduce the ratio below 0.01. One primary reason for the saturation of the Pb/Fe ratio may be the fast heat transfer to the substrate from a thin layer. After the major part of the paint layer was removed, the laser energy absorbed in the residual paint layer was directly transmitted to the bulk substrate. Since the substrate has a higher thermal conductivity than the paint, the transmitted energy easily diffused in the substrate material, and the energy accumulated in the thin paint layer could not reach the threshold energy for ablation. Since paints with resins have larger penetration depths for CO lasers than for the shorter wavelength lasers, 2

high-speed surface cleaning seems possible. An area removal rate of about 17 m2/hour was reported in the previous paper [10]. In order to remove the last layer of paint on the surface, we increased the fluence. Higher irradiation fluences increase removal rates of paint. However, the surface was scorched as a result of heating remnants after the laser beam reached the bulk surface. Figure 4 presents photographs of laser irradiated surfaces with different laser fluences (5.1 J/cm2 in (a), 3.1 in (b), 1.8 in (c) and 1.6 in (d)). It is obvious that the amount of residue decreased with higher fluences, but scorching occurred. In order to achieve complete paint removal without residue or scorching, we examined several chemicals to assist laser ablation cleaning. A small drop of water, ethyl alcohol, acetone and dimethyl-formamide (DFM) was applied on the surface with residue just after the laser irradiation removed the bulk layer of paint. Only DFM helped remove residual paint with no evidence of scorching. Acetone and ethyl alcohol enabled wiping off the residue with a cloth or swab, but did not help this laser ablation method as assisting liquid. Figure 5 shows the EDX spectra of (a) surface before laser irradiation, (b) surface irradiated with 10 laser pulses without the application of DFM, and (c) surface irradiated with 10 pulses with the application of DFM. In (c), DFM was applied after the fifth laser pulse and then another five pulses were irradiated. As a result, the CKa and PbMa signals were no longer detected. This complete removal of paints was also confirmed by microscopic observation in Fig. 2. The decontamination coefficient without the DFM application was about 160, which is calculated from signal values shown in Fig. 3. The corresponding coefficient with DFM exceeded 500, the detection limit of our EDX system. The Pb/Fe signal ratio obtained with the DFM assistance is plotted as (c) in Fig. 3 to show the advantge of DFM application. The effect of DFM is attributed to the decreased ablation threshold of the paint, which may be due to the reduced adhesion force of paint at the bulk surface. This liquid-assist method looks similar to the method used in the particle removal by excimer lasers, in which a very small amount of assist liquid was applied to the contaminated surfaces in a jet steam [1]. The latter’s principle, however, differs from ours because it utilizes the surface vibration excited with the absorption of laser pulse energy at the surface. The sprayed liquid only facilitates the absorption of laser energy and enhances vibration of the surface where particles are attached. The method using a small amount of liquid assistance will be classified as a semi-dry method, not a wet method since it evaporates the paint rather than dissolving it. Systems for large-scale paint removal and surface cleaning of metal are urgently needed not only for air planes and ship factories but also for atomic power plant. Planes and ships require repainting periodically every few years, and the old paint must be thoroughly removed before repainting. In nuclear power plants, the radioactive contaminated surface residue must be removed in situ, without generating of secondary activated materials. In general, chemical solutions such as sulfuric acid are effective, but not easy to operate in situ. Particle-blast-irradiation

439

can also be used for surface decontamination. Though this method provides large-scale, high-speed treatment, low decontamination rates and damage to exposed metal surfaces are the serious problems. Neither of these chemical and physical methods is completely satisfactory because secondary activated materials are unavoidable. Laser ablation surface cleaning with high-repetition-rated, long life-time TEA CO lasers is a possible solution of these 2 problems.

3 Conclusions We have applied a high-power TEA CO laser for high2 speed metal surface cleaning. Paint with synthesized resin on iron and SUS substrates was removed by a linefocused laser beam. The surface cleaning efficiency depend on surface roughness of substrates, so that the number of laser shots required to remove most of the paint increased although the number of pulses needed to reach the substrate surface showed little difference. After laser beams removed the paint layer and reached the substrate surface, a thin residue layer remained on the substrate surface, and the surface was scorched when an excess of pulses or fluence was applied. The paint layer

was completely removed with the help of dimethylformamide applied to the residual layer to assist laser ablation. References 1. A.C. Tam, W.P. Leung, W. Zapka and W. Ziemlich: J. Appl. Phys. 71 (1992) 3515 2. S.J. Lee, K. Imen and S.D. Allen: Appl. Phys. Lett. 61 (1992) 2314 3. Y.F. Lu and Y. Aoyagi: Jpn. J. Appl. Phys. 33 (1994) L430 4. Y.F. Lu, M. Takai, S. Komoro, T. Shiokawa and Y. Aoyagi: Appl. Phys. A59 (1994) 281 5. R. Dumanchin, M. Michon, J.C. Francy, G. Boudient and J. Rocca-Serra: IEEE J. Quantum Electron. QE-8 (1992) 163 6. H. Tanaka, H. Hatanaka, M. Obara, K. Midorikawa and H. Tashiro: Rev. Sci. Inst. 61 (1990) 2092 7. H. Tashiro, A. Suda, M. Maki, M. Kurachi, S. Shibata, M. Toshikuni, M. Maki and K. Midorikawa: J. Appl. Phys. 71 (1992) 2025 8. K. Midorikawa, H. Hatanaka, M. Obara and H. Tashiro: Meas. Sci. Technol. 4 (1993) 388 9. H. Hatanaka, K. Midorikawa, M. Obara and H. Tashiro: Rev. Sci. Inst. 64 (1993) 3061 10. A. Tsunemi, R. Hirai, K. Hagiwara K. Nagasaka and H. Tashiro: Rev. Laser Eng. 22 (1994) 566 (in Japanese) 11. H. Tashiro: Oyo Butsuri. 57 (1988) 1485 (in Japanese) 12. K. Takeuchi, H. Tashiro, S. Kato, K. Midorikawa, T. Oyama, S. Satooka and S. Namba: J. Nucl. Sci. Technol. 26 (1989) 301

.