A characterization of laser cleaning painting layer from steel surface based on thermodynamic model

In this study, the environmentally friendly nanosecond ultraviolet (UV) laser is an innovatively employed laser cleaning to remove the painting layer from the AH36 steel substrate. The feasibility of UV laser cleaning the painting layer is innovatively proposed and it has been calculated by the model theoretically, followed by elaborating the prominent interaction mechanism of UV laser exactly. The initial cleaning threshold and completely cleaning threshold are 2 J/cm2 and 5 J/cm2, respectively. Afterwards, the UV laser-cleaned surface quality is evaluated by the scanning electron microscopy (SEM), energy dispersive spectroscopy (EDS), optical microscopy (OM), and optical profiler (OP), respectively. The mechanical properties have enhanced dramatically after laser cleaning and characterized by the Vickers hardness tester and universal testing machine. By varying laser fluences (2, 5, 7 J/cm2) during laser cleaning, microstructures registering various sizes of corrugated shaped, craters, and ring shaped could be acquired. In addition, the mechanical properties analysis including rapid melting, quenching, and dislocation density effects illustrates that laser cleaning could effectively increase surface microhardness, tensile strength, and bending strength. Thus, laser cleaning method has emerged as a favorable means to strip painting layer in lieu of traditional methods for marine industry as well as this study could promote the development of laser cleaning in the field of marine engineering.


Introduction
In the past a few decades, surface cleaning has been attracting considerable attention in various fields including aerospace, ocean engineering, microelectronics, and medicine [1][2][3][4]. With regard to the contaminants on the surface, they are mostly of oxides, paints, polymers, coatings, microorganisms, and particles [5][6][7][8][9][10]. Marine ships are a significant concern, which is mainly due to the corrosion as exposed to seawater. In this regard, ships appear to experience the paint removing alongside repainting for the maintenance of shipyard. Of particular note, the above issue has been reported to cost billions of dollars annually. Therefore, it is vital to strip off the painting layer from the substrate in order to repaint and extend the service life of the ship. Seen from prior studies, the traditional surface cleaning methods including the mechanical and chemical cleaning are recognized to be the most popular approaches. However, they have been reported to trigger the risks of polluting the atmosphere and producing secondary wastes. Standing on this view, the laser cleaning method, emerging as a promising technique, is recognized to be an alternative to those conventional methods as it is more ecofriendly and does not need to contact cleaned surface. Depending on a host of traits including but not limited to excellent plasticity and toughness, high mechanical strength alongside light density and fair stiffness, steel, e.g., type AH36, has been widely used in ocean engineering [11][12][13]. However, it should be emphasized here that the service life of devices made with steel is strongly affected by marine environment as the latter contains a large amount of erosive ions and microorganisms which may collaboratively trigger the spalling of painting layer coating the exterior of steel base.
What in follows, the internal steel base is degraded as the subsequent exposure continues. Thus, surface cleaning is extremely necessary as it allows to remove the old painting layer that has been damaged, together with the attaching erosive substances. There so far exist extensive studies to provide constructive hints for solving the above concern through laser cleaning technique. In 1974, J.A. Fox pioneered a study to conduct surface cleaning upon paint layer, which was essentially achieved by using a Q-switched neodymium laser [14]. It was reported that the paint layer can be removed effectively under the strong photo-induced stress wave generated by the associated laser. Besides, K. Liu and E. Garmire adopted different types of laser alongside varying pulse width to remove the lacquer [15]. Seen from that, the Nd: YAG laser along with Q modulation was reported to show the stronger removal effect in comparison to other examined lasers, i.e., CO 2 laser, excimer laser, and continuous wave laser. More importantly, Chen et al. pointed out that the laser cleaning technique may overcome the main flaw existing in conventional cleaning methods, viz., secondary pollution, which was confirmed through optical microscope observation [13].
As it is well reported, many factors including but not limited to the pulse frequency, the scanning line interval, the scanning speed, and the laser fluence are able to affect the efficiency of laser cleaning [16][17][18][19][20][21][22][23][24][25][26][27][28][29][30]. According to Zhao et al., the energy incorporated in each single pulse increased as the pulse frequency decreased, which would potentially trigger the plasma shielding effect near the painting layer surface once the pulse energy reached to a certain value. Of particular note, this resulting plasma shielding effect has a potential to cause the recession phenomenon through absorbing the energy involved in lasers and accordingly hinders the further cleaning effectiveness to the painting layer [17][18][19]. Besides, the larger scanning line interval was reported to decrease the corresponding cleaning efficiency. This may be attributed to the fact that the increase in the scanning interval could dilute the unit received energy or fluence at the scanned region on the one hand and sometimes even leads to the appearance of residual area on the second hand [20][21][22][23]. In addition, the scanning speed influences the cleaning efficiency essentially through changing the corresponding overlapping rate. In this regard, the larger scanning speed, the better cleaning performance, and vice versa. [24,25]. Apart from the above, laser fluence has been widely recognized as a predominant factor when investigating the laser cleaning effectiveness [2,4,11,[26][27][28][29][30]. In an early study conducted by Tian et al., two different laser fluences, namely 1.08 J/cm 2 and 4.14 J/cm 2 , were respectively set and the higher cleaning efficiency was noted in the case of larger laser fluence [4]. This has also been confirmed by a recent report [11], wherein the increase in laser fluence within a certain range, from 1.19 to 4.43 J/cm 2 , was once again noted to help improve the corresponding cleaning efficiency. The underlying reasons behind this may be that the increasing fluence empowers the thermodynamic behavior of molecules and accordingly makes the attached contaminants easier to be removed after a series of process including melting, vaporizing, and solidifying. As well, the larger fluence allows to produce the plasma in a quicker and more efficient way, which also favors the further cleaning [2,11]. However, it should be emphasized here that any increase in laser fluence beyond a threshold might trigger a risk of damaging the substrate. For instance, such an upper bound in terms of laser fluence was reported to be about 11.9 W/cm 2 , which induces the excessive ablation of the substrate [31][32][33]. Together with experimental investigations, numerical simulations have also been extensively carried out to understand the underlying mechanism involved in laser cleaning. W.F. Zou et al. established a one-dimensional heat conduction model based on the thermal stress mechanism alongside the Fourier transfer equation. Then, the process of laser cleaning was successfully simulated and the cleaning and damage thresholds were so numerically predicted [34]. On the basis of heat conduction and Hertz-Knudsen theory, V. Oliveira et al. developed a 2D finite element model to describe the TiC ablation effects led by pulsed laser irradiation with varying laser fluences, and the corresponding results agreed well with pertinent experimental values [35]. Aside from this, Vasantgadkar et al. also predicted the ablation depth and temperature distribution, on account of plasma shielding effect [36]. Yue et al. adopted the laser respectively shaped as rectangular and Gaussian to numerically investigate the associated effect upon the cleaning efficiency [37].
However, it should be emphasized here that the laser employed in most of the above studies was an infrared nanosecond laser that usually triggers thermal induced defects, manifesting as the recasting layer and the heat-affected zone (HAZ). On the other hand, the ultraviolet (UV) laser may avoid these disadvantages effectively as it registers a greater photon energy and accordingly is easier to break the intermolecular bonds. To the authors' knowledge, there exists very limited study that adopted UV laser to clean the painting layer and more importantly, the influence led by associated parameters upon cleaning efficiency is still not clear. In this study, the authors mark the first in using UV laser to clean the painting layer and varying laser fluence levels ranging from 2 to 7 J/cm 2 were examined. Further, the resulting samples were also evaluated through optical microscopy (OM), field emission scanning electron microscopy (SEM), optical profiler (OP), and mechanical measurements.

Materials
A commercially sourced AH36 steel, as widely used in marine engineering, was employed as the substrate in this study, and its elementary compositions as determined by XRF technique are now presented in Table 1. A locally supplied paint comprising chlorinated rubber was used to coat the substrate twice. Note here that the inter layer (white color) registered a thickness of 70 μm while the outer layer (red color) was as thick as 100 μm. The dimensions of specimens, as being used to experience laser cleaning and further characterization including SEM-EDS, OP, and OM, were set as 40 mm × 40 mm ×5 mm. In addition, another two batches of specimens, respectively sized as 40 mm × 5 mm ×5 mm and 30 mm × 10 mm ×2 mm, were prepared as well for tensile and bending tests, conforming to ASTM E8 and ASTM E290. The representative samples are illustrated in Fig. 1.

UV laser cleaning surface treatment
For starter, the laser cleaning process is operated in an ambient atmosphere, i.e., (20°C), using a nanosecond fiber laser (Huaray, China) coupled with a two-dimensional galvanometer scanning system alongside the progressive scanning mode. The parameters in terms of employed laser were set as follows: maximum average power = 20 W, wavelength = 355 nm, laser pulse width = 15 ns, and laser repetition rate =100 kHz. Furthermore, three different laser fluences were examined, respectively corresponding to 2 J/cm 2 , 5 J/cm 2 , and 7 J/cm 2 .

Surface characterization
The morphology of resulting specimens was characterized by a digital ultra-depth-of-field microscope (Keyence, VHE-1000, Japan) and a field emission scanning electron microscope (FE SEM) coupled with an energy dispersive X-ray spectroscope (EDS). The EDS was performed to analyze the chemical compositions of laser-cleaned surface and the original painting layer surface. An optical profiler (OP, Zygo Corporation, USA) was adopted to evaluate the surface roughness and 3D morphology. In addition, the mechanical performance, i.e., tensile and bending strength, was measured through the Instron 5569 universal testing machine, which offered the uniaxial tension and 3-point approach bending tests. The microhardness was determined using the Vickers durometer.

Results and discussion
In order to realize the UV laser cleaning the painting layer feasibility, the underlying analysis and discussion are crucial and essential factors to obtain the outstanding laser-cleaned effects theoretically. As for the mechanisms of UV laser cleaning, it is widely recognized that the photothermal and photochemical interactions and combination of them are underscoring effects, which has been expressed in various models detailedly [38][39][40][41][42][43][44][45]. It is worth pointing out that during the period of laser cleaning, the photochemical interactions, viz., the energy of UV laser emitted photons, are greater than the molecular bond energies within the material, which in turn further break the molecular bonds and manifest as the cold machining. One notes from Table 2 the detailed information as per the molecular bond energies of the painting layer. The photon energy is expressed as E=hγ=hc/λ, where E is photon energy, h is Planck constant, and γ and λ are frequency and laser wavelength respectively. According to the theoretical calculation, the energy of 355 nm UV laser is 6.2 eV, which is much greater than molecular bond energies of the painting layer, thereafter inducing molecular bonds broken chemically. It is well agreed with the prior mentioned theory and can be verified by theoretical calculation. From the table, it is clear that the energy of photon emitted by the UV laser (6.3 eV) is considerably greater than that most of molecular single bond, such as C-H bond registering 4.30 eV. Therefore, the adopted

Surface morphology
One notes from Fig. 2 that the representative images of lasercleaned surface are exhibited at varying laser fluences upon the painting layer adhered on the surface of AH36 steel substrate. It is noteworthy that there exists a strong smell of burning and a huge sound of vibration during the period of laser cleaning, which is mainly due to the interactions of the UV laser and the painting layer. Aside from this, it can also be inferred that both photothermal and photochemical reactions trigger on the surface of the painting layer. It is coincidence with the aforementioned theoretical analysis, namely there exists both the photochemical and photothermal interactions showing up in the laser cleaning process, along with breaking molecular bonds chemically. Typically, Fig. 2a, e, and i display the original painting layer surface before laser cleaning, while Fig. 2 (b, f, j), (c, g, k), and (d, h, l) indicate laser-cleaned surface at varying laser fluences of 2 J/cm 2 , 5 J/cm 2 , and 7 J/cm 2 respectively. Interestingly, seen from the Fig. 2b, it can be shown that the superficial color changes on the surface, viz., different from red painting layers, which is due to the fact that there exist two kinds of painting layer (red and white) adhering on the substrate surface. After the laser cleaning at laser fluence of 2 J/cm 2 , the red painting layer has almost been peeled off from the surface, whereas the white one left on the surface, as evident from Fig. 2a. This may attribute to the laser fluence that cannot reach the threshold of the painting layer and the energy is not strong enough to remove the painting layer directly. Notably, the laser fluence at 2 J/cm 2 is called initial cleaning threshold. Aside from that, it worth mentioning here that there are some uncleaned areas at the edge of specimen after laser cleaning, which is mainly due to the laser-satisfied Gaussian distribution in space, namely the energy in the middle is evidently higher than the edges. As expected, the laser fluence distributed at the edges is not strong enough to strip the painting layer, which coincides with the prior mentioned phenomena. One notes from Fig. 2c that it  can be presented as a superficial color variation from the zebra-like color to bar-shaped grey-black color in the macro at a laser fluence of 5 J/cm 2 . Specifically, UV laser cleaning induces a prominently complete removal and manifests as a metallic luster on the surface of the substrate. This may infer that improving the laser fluence could help enhance the laser cleaning effects and complete cleaning threshold of laser cleaning painting layer is 5 J/cm 2 . The resulting surface indicates that the cleaning effects have a positive correlation with laser fluence. Moreover, if the laser fluence increases to initial cleaning threshold, the painting layers begin to remove and there exists some traces on the surface. In contrast, if the laser fluence enhances to complete cleaning threshold, the corresponding painting layers are peeled off thoroughly. As noted earlier, the enhancement of laser fluence could promote the behavior of molecule thermodynamics and a host of process, the resultant surface is agreed well with previous study [11]. Compared with the laser-cleaned surface at fluence of 5 J/cm 2 , craters and tracks are clearly found in black color after the laser cleaning at a laser fluence of 7 J/cm 2 , which is mainly due to the existing excessive ablation upon laser-cleaned surface. This may be attributed to the photothermal effects playing a prominent role in the laser cleaning at a laser fluence of 7 J/cm 2 and the resultant surfaces may experience the evaporation, melting, re-solidification, and ablation in an iterative dynamic process, which in turn further lead to the excessive ablation. Figure 3 exhibits a low-magnification secondary electron SEM image of the micro-morphology of laser-cleaned painting layer at varying laser fluences, including 2 J/cm 2 , 5 J/cm 2 , and 7 J/cm 2 respectively. For the captured surfaces observed in Fig. 3a the employed laser fluence was 2 J/cm 2 and it can be found that there are some cracks and concaves exposed on the surface of laser-cleaned surface, which is mainly due to the local temperature increase, viz., the enhanced lattice vibrationinduced cracking result in the temperature enhancement, which coincides with the mechanism of the UV laser cleaning, namely the photochemical interaction reactions between the laser and the painting layer. As noted from Fig. 3d instead of concaves and cracks, relatively smooth and evenly surface, viz., some corrugated shape morphology, can be presented from the images, which indicates laser fluence reaching the cleaning threshold of the painting layer and without destroying the underlying substrate. The corrugated shape morphology is well agreed with a metallic luster surface of the substrate noted in Fig. 2c which is suggested that it is a thoroughly complete removal. As well, it is worth mentioning here that there are some ring-shaped microstructures showing  Fig. 3g. It is mainly due to the fact that the laser fluence exceeding the theoretical threshold considerably triggers the phase change of the substrate, followed by production of the ring-shaped microstructures. This is supported by the previous results [46]. As noted in Fig. 3, there are many craters in each laser-cleaned surface and the craters cannot be filled with the liquid metal immediately; therefore, it generates various crater microstructures on the surface of AH36 steel substrate.

Surface element distribution
For evaluating the laser cleaning painting layer effects, the energy dispersive X-ray spectroscopy (EDS) is widely recognized as a general method to examine the chemical composition. One notes from Fig. 4 that it exhibits the corresponding element content in the marked black box straightforwardly and the percentages of weight at varying laser fluences are expressed in Table 3. Together with the EDS analysis of the UV lasercleaned surface, the authors find that Fe element weight percentage is 53.04% and oxygen is 5.58% in weight percentage in area A. As noted from Fig. 4a and b, compared with area A, the area B in both the Fe and O elements has dramatically enhanced, which indicates area B has a relatively cleaned surface and exposed more substrate surface than area A. As seen therein, the process of area A belongs to incompletely cleaning period, whereas area B indicates it is a completely cleaning process, which is well agreed with the earlier mentioned Fig.  3a and d, respectively. As for area C, it can be seen that the  weight of C and O elements is higher than that of the areas A and B, which suggests the area experiences excessive ablation and the corresponding iron oxide registers the ring-shaped microstructure. This is also agreed well with the aforementioned Fig. 3g phenomena, including the ring-shaped microstructure and the excessive ablation effects.
To reveal the surface element distribution more clearly, the line EDS is performed to investigate changes of Fe, O, and C contents along with some trace amounts of Cr and Ti. Figure 5a, b, and c exhibit the line EDS report of element changes in the resulting laser-cleaned area at varying laser fluences of 2 J/cm 2 , 5 J/cm 2 , and 7 J/cm 2 respectively. From the curves, it can be observed that the Fe Kα1 appeared in (a) and (c) oscillates significantly more than (b), which is mainly due to the laser-cleaned surface (b) is relatively flat, viz., painting layer has been peeled off from the substrate completely. Interestingly, it can be found there are some periodic curves appearing in the line EDS-examined surface in Fig. 5a. This indicates that the produced ring-shaped microstructure, namely iron oxide, is approximately periodic, which is due to the UV laser satisfying the Gaussian distribution.
One notes from Fig. 6a that presents a secondary electron SEM image of UV laser-cleaned surface morphology and manifests as disparate contrast areas thoroughly. In Fig. 6bg, it can be observed that certain specific elements including Fe, O, C, Ni, Cr, and Mo are distributed on the surface with various colors. From the mapping of the AH36 steel substrate surface, it is presented the distribution of Fe, O, and C is the major elements on the surface, which is mainly due to Fe and C elements which are the prominent elements of substrate and there may exist a laser ablation during the laser cleaning period. Clearly, from the mapping, it can be seen that the distribution of the O and C is relatively homogeneous. As per the surface morphology of laser-cleaned surface, it is suggested that the convex exposed on the laser-cleaned surface is rich in the O and C elements, whereas they are almost absent in the center. This may be attributed to the UV laser-satisfied Gaussian distribution and the energy density at the center of the spot is greater than that at the edge, which results in the O element exposed in the center much more the edges. That is the reason why the microstructure could generate the ringshaped microstructure and corrugated-shaped morphology respectively. Herein, the present findings are very encouraging and it can also infer that the nanosecond UV laser could successfully strip off the painting layer thoroughly.

Hardness characterization
It is widely recognized that in the exact evaluation of the mechanical properties, the hardness characterization is an indispensable, essential evaluation index, which needs to be considered during the UV laser cleaning the painting layer. In this study, the Vickers microhardness test is performed to investigate the mechanical properties, the details including the load of 290 g and holding time duration of 15 s, which can be demonstrated in Fig. 7. As such, from the curves, it can be noticed that the UV laser-cleaned surface increases to 150 HV while the uncleaned surface is 92 HV. The dramatically microhardness enhancement can be attributed to the fact that the UV laser-cleaned surface shows up rapid melting and quenching, which is followed by generating the microstructure, including ring shaped and corrugated shaped. The other reason is due to the fact that the UV laser cleaning could induce the resultant surface producing dislocation density, while the traditional cleaning techniques are not available. This is supported by the previous results [17,47,48]. Specifically, the standard deviation of laser-cleaned surface and uncleaned surface microhardness is 2.875 and 0.525, respectively. Due to the uneven painting layer, the maximum Vickers microhardness UV laser-cleaned surface is 178 HV and the minimum is 150 HV. Thus, it is suggested that UV laser cleaning painting layers could enhance the surface microhardness considerably. Aside from the environmentally friendliness, this is another reason why authors take this prospective method to remove the painting layer.

Tensile and bending characterizations
For further investigation of the tensile and bending properties of laser-cleaned surface, there are batches of standard tensile and bending test specimen produced by UV laser cleaning method. The commensurately tensile stress and strain curves are illustrated in Fig. 8. From the curves, it exhibited five stages in these tensile curves, such as elastic deformation, yield deformation, plastic deformation, necking, and fracture respectively [49][50][51][52][53][54][55]. The curves are just like the parabolic shape and there is a rapid increase in the elastic and yield deformation stages. Followed by gradually enhanced plastic deformation stage, there exist maximum values, viz., ultimate strength, appearing before reaching necking period, along with the reduction to facture stage sharply. Notably, the tensile strength of laser-cleaned surface is much stronger than that of before cleaning, which is mainly due to the fact laser treatment is conducive to enhance the elastic and plastic deformation properties of the substrate.
As for bending properties, it is widely recognized that bending displacements and bending stress are essential factors, exhibited in Fig. 9. As noted in the curves, it can be observed that  The microhardness characterization after UV laser cleaning both of the resulting laser-cleaned surface and uncleaned surface experience complete elastic deformation and plastic deformation stage, whereas the fracture stage does not exist in the sample, in spite of the bending angles exceeding 90°and bending stress over 1000 MPa. In this regard, it indicates that both of the laser-cleaned surface and uncleaned surface have excellent bending strength and plasticity. Yet, laser-cleaned surface manifests as better plasticity in the plastic deformation stage. Thus, the laser-cleaned surface conduces to improve the bending strength and plasticity properties.

Roughness and profile characterizations
As seen in Fig. 10, it presents typical 3D morphologies along with corresponding line-scanned profiles of original painting layer and resulting laser-cleaned surface, which is examined by optical profiler and commensurately captured zone is 840 μm × 840 μm. As evident from Fig. 10a, it can be seen that the original painting layer is relatively evenly and smooth, along with the surface roughness 1.968 μm. Of particular note, white lines indicate maximum height difference is approximately 10 μm at original surfaces. With regard to the laser fluence 2 J/cm 2 , the maximum height difference is five times larger than that of the original painting layer surface and corresponding surface roughness is 12.751 μm. This may be contributed to the laser energy is less capable to remove the painting layer directly and there exists laser cleaning-induced cracks and concaves in this layer, which is well agreed with the prior mentioned Fig. 3a. Thus, the roughness of lasercleaned surface at laser fluence 2 J/cm 2 increases dramatically. Specifically, as for laser-cleaned surface at laser fluence of 5 J/cm 2 , the resultant surface is pretty smooth, viz., the surface roughness is 2.471 μm and relevant maximum height difference is approximately 6 μm, which indicates laser fluence at 5 J/cm 2 is the most suitable for UV laser cleaning painting layers. As such, it is suggested that laser-cleaned surface at this fluence without destroying the underlying substrate and the thermal ablation is minimal from the aforementioned surface morphology in Fig. 3d. Therefore, laser fluence at 5 J/cm 2 is regarded as the threshold of UV laser cleaning painting layer. To be noted here, the melted layer flowing and resolidification forms a corrugated shape morphology, which coincides well with Fig. 3d. In comparison, the laser-cleaned surface at laser fluence 7 J/cm 2 has a relatively rough surface (6.298 μm) and the maximum height difference is about 27 μm, which is suggested that the surface experiences the excessive ablation and it is consistent with the captured images in Fig. 3g. Thus, it is imperative to avoid the ablative conditions that occurred during UV laser cleaning painting layer as far as possible. Fig. 8 The relationship between the strain and tensile stress Fig. 9 The relationship between the bending displacements and bending stress

Theoretical model analysis
The experimental results illustrate initial cleaning threshold and complete cleaning threshold based on laser and painting layer interactions. In order to explain the phenomena more in detail, it is necessary to establish a thermodynamic model to describe the laser cleaning mechanism. Based on the study of Zhang et al. [56], the theoretical relationship between the temperature and energy are expressed as: Herein, T 1 , T 2 , n, C P , and ΔH stand for initial temperature, final temperature, amount of substance, heat capacity, and molar heats of phase transition.
where, A, B, and C are constants correlated with substance properties.
Based on the thermodynamic model, it is employed for theoretical calculation and the corresponding assumptions are under following conditions: (a) The painting layer and AH36 steel substrate are taken as heat insulators. (b) The distribution of laser energy is homogeneous.
One notes from Fig. 11 the schematic diagram of laser absorption and reflection. Wherein, E L , E R , δ 1 , and δ 2 indicate incident laser energy, reflection laser energy, thickness of painting layer, and substrate correspondingly. As well, E Fe , E A , and E B are energy absorption of Fe substrate, incident laser, and reflection through painting layer respectively. A a and A b stand for laser absorption coefficients of painting layer and Fe substrate respectively. The detailed equations are shown as follows: Herein, α 1 and α 2 are laser absorption coefficients of painting layer and Fe substrate respectively. In addition, it can be calculated As per Eq. (1), it can be obtained Eq. (6), as follows: Afterwards, the laser fluence can be deduced from Eq. (7) and the relationship is expressed as: Correspondingly, the laser fluence of Fe substrate can be illustrated as follows: Wherein, ρ 1 , ρ 2 , M 1 , and M 2 are densities of painting layer and Fe substrate as well as molar masses of painting layer and Fe substrate, respectively. Inspired from Eqs. (7) and (8) and substituted corresponding physical constants shown in Table 4, the calculated the theoretical laser fluence is 1.789 J/cm 2 if temperature approaches its melting points. As for the Fe substrate, it also  can be derived from functions that laser fluence is 3.216 J/cm 2 and 4.65 J/cm 2 with regard to the melting point and boiling point, respectively. It is noteworthy that the calculated theoretical threshold (4.65 J/cm 2 ) exceeded initial cleaning threshold, which is mainly due less consideration of the plasma shielding effects and thermal expansion effects as well as various thickness of painting layer. These points will be taken into account in future studies to modify this model effectively.
As for the UV laser cleaning the surface of painting layer, it experiences melting, evaporation, rapid solidification, and resolidification as well as the corresponding hydrogen precipitates from the surface, followed by the formation of hydrogen. Comprehensively, the laser cleaning is a facile, environmentally friendly, and promising method to strip off the painting layer from the marine engineering surface. Hopefully, this study would provide an experimental and theoretical analysis reference in the UV laser cleaning the painting layer and pave the way for any further potential applications in industrial field.

Concluding remarks
In this paper, a study based on the nanosecond UV laser cleaning method is innovatively proposed, which is successfully utilized to strip off the painting layer from the AH36 steel substrate. This study innovatively verifies the feasibility of the UV laser cleaning the painting layer on the surface of AH36 steel in theory and briefly elaborates the primary interaction mechanism of UV laser, such as the photothermal and photochemical interactions. The thermal dynamic model is established to describe the relationship between temperature and laser fluence, which indicates the experimental results (4.65 J/cm 2 ) are close to theoretical cleaning threshold (5 J/cm 2 ) and the differences between them are discussed exactly. Moreover, SEM, EDS, optical profiler, and mechanical tests are detailedly performed to study the morphologies, chemical compositions, and mechanical properties of original painting layer surface and resulting laser-cleaned surface at varying laser fluences respectively. From the SEM test, there are some typical corrugated shaped, craters, and ring-shaped microstructures exhibited on the surface of laser-cleaned surface at various laser fluences. With regard to the mechanical properties, it is worth mentioning here that the UV lasercleaned painting layer surface could enhance the surface microhardness, tensile strength, and bending strength dramatically. This can be attributed to the UV laser-cleaned surface experiencing rapid melting and quenching, followed by generating the ring-shaped and corrugated-shaped microstructure as well as the produced dislocation density. Therefore, this promising UV laser cleaning method is not only environmentally friendly, but also enhances the mechanical properties of laser-cleaned surface significantly. Hopefully, there is a great potential to utilize this promising method to large-scale cleaning the painting layer of the marine engineering surface and make some contributions to the marine and industrial fields in the future.