A Characterization of Laser Cleaning Painting Layer From Steel Surface Based on Thermodynamic Model

8 In this study, the environmentally friendly nanosecond ultraviolet (UV) laser is innovatively 9 employed laser cleaning to remove the painting layer from the AH36 steel substrate. The feasibility 10 of UV laser cleaning the painting layer is innovatively proposed and it has been calculated by the 11 model theoretically, followed by elaborating the prominent interaction mechanism of UV laser 12 exactly. The initial cleaning threshold and completely cleaning threshold are 2 J/cm 2 and 5 J/cm 2 , 13 respectively. Afterwards, the UV laser cleaned surface quality is evaluated by the scanning electron 14 microscopy (SEM), energy dispersive spectroscopy (EDS), optical microscopy (OM) and optical 15 profiler (OP), respectively. The mechanical properties have enhanced dramatically after laser 16 cleaning and characterized by the Vickers hardness tester and universal testing machine. By 17 varying laser fluences (2, 5, 7 J/cm 2 ) during laser cleaning, microstructures registering various 18 sizes of corrugated shaped, craters and ring-shaped could be acquired. In addition, the mechanical 19 properties analysis including rapid melting, quenching and dislocation density effects illustrates 20 that laser cleaning could effectively increase surface microhardness, tensile strength and bending 21 strength. Thus, laser cleaning method has emerged as a favourable means to strip painting layer in 22 lieu of traditional methods for marine industry as well as this study could promote the development 23 of laser cleaning in the field of marine engineering.


INTRODUCTION 1
In the past a few decades, surface cleaning has been attracting considerable attention in 2 various fields including aerospace, ocean engineering, microelectronics and medicine [1][2][3][4]. With 3 regard to the contaminants on the surface, they are mostly of oxides, paints, polymers, coatings, 4 microorganisms and particles [5][6][7][8][9][10]. Marine ships defouling the painting layer is a significant 5 concern, which is mainly due to the corrosion as exposed to seawater. In this regard, ships appear 6 to experience the paint removing alongside repainting for the maintenance of shipyard. Of 7 particular note, the above issue has been reported to cost billions of dollars annually. Therefore, it 8 is vital to strip off the painting layer from the substrate in order to repainting and extending the 9 service life of the ship. Seen from prior studies, the traditional surface cleaning methods including 10 the mechanical and chemical cleaning are recognized to be the most popular approaches. However, 11 they have been reported to trigger the risks of polluting atmosphere and producing secondary 12 wastes. Standing on this view, the laser cleaning method, emerging as a promising technique, is 13 recognized to be an alternative to those conventional methods as it is more eco-friendly and does 14 not need to contact cleaned surface. Depending on a host of traits including but not limited to 15 excellent plasticity and toughness, high mechanical strength alongside light density and fair 16 stiffness, steel, e.g. type AH36, has been widely used in ocean engineering [11][12][13]. However, it 17 should be emphasized here that the service life of devices made with steel is strongly affected by 18 marine environment as the latter contains a large amount of erosive ions and microorganisms 19 which may collaboratively trigger the spalling of painting layer coating the exterior of steel base. 20 What in follows, the internal steel base is degraded as the subsequent exposure continues. Thus, 21 layer, thereafter inducing molecular bonds broken chemically. It is well agreed with the prior 1 mentioned theory and can be verified by theoretical calculation. From the table, it is clear that the 2 energy of photon emitted by the UV laser (6.3 eV) is considerably greater than that most of 3 molecular single bond, such as C-H bond registering 4.30 eV. Therefore, the adopted UV laser 4 cleaning could remove the painting layer in theory by calculation of photon energy. 5

Surface morphology 6
One notes from Fig. 2 that the representative images of laser cleaned surface are exhibited at 7 varying laser fluence upon the painting layer adhered on the surface of AH36 steel substrate. It is 8 noteworthy that there exists a strong smell of burning and a huge sound of vibration during the 9 period of laser cleaning, which is mainly due to the interactions of the UV laser and the painting 10 layer. Aside from this, it can also be inferred that both photothermal and photochemical reactions 11 trigger on the surface of the painting layer. It is coincidence with the aforementioned theoretical 12 analysis, namely, there exists both the photochemical and photothermal interactions showing up 13 in the laser cleaning process, along with breaking molecular bonds chemically. Typically, Fig. 2  14 (a, e, i) display the original painting layer surface before laser cleaning, while Fig. 2 (b, f, j), (c, g, 15 k) and (d, h, l) indicate laser cleaned surface at varying laser fluence of 2 J/cm 2 , 5 J/cm 2 , and 7 16 J/cm 2 respectively. Interestingly, seen from the Fig. 2 (b), it can be shown that the superficial color 17 changes on the surface, viz., different from red painting layers, which is due to the fact that there 18 exist two kinds of painting layer (red and white) adhering on the substrate surface. After the laser 19 cleaning at laser fluence of 2 J/cm 2 , the red painting layer has almost been peeled off from the 20 surface, whereas the white one left on the surface, as evident from Fig. 2 (a). This may attribute 21 to the laser fluence cannot reach the threshold of the painting layer and the energy is not strong 1 enough to remove the painting layer directly. Notably, the laser fluence at 2 J/cm 2 is called initial 2 cleaning threshold. Aside from that, it worth mentioned here that there are some uncleaned areas 3 at the edge of specimen after laser cleaning, which is mainly due to the laser satisfied Gaussian 4 distribution in space, namely, the energy in the middle is evidently higher than the edges. As 5 expected, the laser fluence distributed at the edges is not strong enough to strip the painting layer, 6 which is coincide with the prior mentioned phenomena. One notes from Fig. 2 (c) that it can be 7 presented a superficial color variation from the zebra like color to bar shaped grey-black color in 8 the macro at a laser fluence of 5 J/cm 2 . Specifically, UV laser cleaning induces a prominently 9 complete removal and manifests as a metallic luster on the surface of the substrate. This may infer 10 that improving the laser fluence could help enhance the laser cleaning effects and complete 11 cleaning threshold of laser cleaning painting layer is 5 J/cm 2 . The resulting surface indicates that 12 the cleaning effects have positive correlation with laser fluence. Moreover, if the laser fluence 13 increases to initial cleaning threshold, the painting layers begin to remove and there exists some 14 traces on the surface. In contrast, if the laser fluence enhances to complete cleaning threshold, the 15 corresponding painting layers are peeled off thoroughly. As noted earlier, the enhancement of laser 16 fluence could promote the behavior of molecules thermodynamic and a host of process, the 17 resultant surface is agreed well with previous study [11]. Compared with the laser cleaned surface 18 at fluence of 5 J/cm 2 , it can be found craters and tracks clearly in black color after the laser cleaning 19 at a laser fluence of 7 J/cm 2 , which is mainly due to existing the excessive ablation upon laser 20 cleaned surface. This may be attributed to the photothermal effects play a prominent role in the 21 laser cleaning at a laser fluence of 7 J/cm 2 and the resultant surfaces may experience the 1 evaporation, melting, re-solidification and ablation in an iterative dynamic process, which in turn 2 further lead to the excessive ablation. of laser cleaned painting layer at varying laser fluence, including 2 J/cm 2 , 5 J/cm 2 and 7 J/cm 2 5 respectively. For the captured surfaces observed in Fig. 3 (a), the employed laser fluence was 2 6 J/cm 2 and it can be found there are some cracks and concaves exposed on the surface of laser 7 cleaned surface, which is mainly due to the local temperature increase, viz. the enhanced lattice 8 vibration induced cracking result in the temperature enhancement, which is coincide with the 9 mechanism of the UV laser cleaning, namely, the photochemical interaction reactions between the 10 laser and the painting layer. As noted from Fig. 3 (d), instead of concaves and cracks, relatively 11 smooth and evenly surface, viz. some corrugated shape morphology can be presented from the 12 images, which indicates laser fluence reaching the cleaning threshold of the painting layer and 13 without destroying the underlying substrate. The corrugated shape morphology is well agreed with 14 a metallic luster surface of the substrate noted in Fig. 2 (c), which is suggested that it is a 15 thoroughly completely removal. As well, it is worth mentioning here that there are some ring-16 shaped microstructures showing up in Fig. 3 (g). It is mainly due to the fact that the laser fluence 17 exceeding the theoretical threshold considerably, which triggers the phase change of the substrate, 18 followed by producing the ring-shaped microstructures. This is supported by the previous results 19 [46]. Notably, there are many craters in each laser cleaned surface, as noted in Fig. 3, which may 20 explain through the following: the AH36 steel substrate contains hydrogen. As for the UV laser 21 cleaning the surface of painting layer, it experiences melting, evaporation, rapid solidification and 1 re-solidification as well as the corresponding hydrogen precipitates from the surface, followed by 2 the formation of hydrogen. This is also supported by prior research results [5,47]. In what follows, 3 the craters cannot be filled with the liquid metal immediately, therefore it generates various craters 4 microstructure on the surface of AH36 steel substrate. 5

Surface element distribution 6
For evaluating the laser cleaning painting layer effects, the energy dispersive X-ray 7 spectroscopy (EDS) is widely recognized as a general method to examine the chemical 8 composition. One notes from Fig. 4 that it exhibits the corresponding elements content in the 9 marked black box straightforwardly and the percentages of weight at varying laser fluence are 10 expressed in Table 3. Together with the EDS analysis of the UV laser cleaned surface, the authors 11 find that Fe element weight percentage is 53.04 % and oxygen is 5.58 % in weight percentage in 12 area A. As noted from Figs. 4 (a) and (b), compared with area A, the area B in the Fe and O 13 element both have dramatically enhancement, which indicates area B has a relatively cleaned 14 surface and exposed more substrate surface than area A. As seen therein, the process of area A 15 belongs to incompletely cleaning period, whereas area B indicates it is a completely cleaning 16 process, which is well agreed with the earlier mentioned Fig. 3 (a) and (d), respectively. As for 17 area C, it can be seen that the weight of C and O element higher than that of the area A and B, 18 which suggests the area experiences excessive ablation and the corresponding iron oxide registers 19 the ring-shaped microstructure. This is also agreed well with the aforementioned Fig. 3 (g) 20 phenomena, including the ring-shaped microstructure and the excessive ablation effects. 21 To reveal the surface element distribution more clearly, the line EDS is performed to 1 investigate changes of Fe, O, C content along with some trace amounts of Cr and Ti. and (c) exhibit the line EDS report of element changes in the resulting laser cleaned area at varying 3 laser fluence of 2 J/cm 2 , 5 J/cm 2 and 7 J/cm 2 respectively. From the curves, it can be observed that 4 the Fe Kα1 appeared in (a) and (c) oscillates significantly more than (b), which is mainly due to 5 the laser cleaned surface (b) is relatively flat, viz. painting layer has been peeled off from the 6 substrate completely. Interestingly, it can be found there are some periodic curves appearing in the 7 line EDS examined surface in Fig. 5 (a). This indicates that the produced the ring-shaped 8 microstructure, namely, iron oxide is approximately periodic, which is due to the UV laser satisfied 9 the Gaussian distribution. 10 One notes from Fig. 6 (a) that presents a secondary electron SEM image of UV laser cleaned 11 surface morphology and manifests as disparate contrasts areas thoroughly. In Fig. 6 (b-g), it can 12 be observed certain specific elements including Fe, O, C, Ni, Cr, and Mo are distributed on the 13 surface with various colors. From the mapping of the AH36 steel substrate surface, it is presented 14 the distribution of Fe, O and C is the major elements on the surface, which is mainly due to Fe and 15 C elements are the prominent elements of substrate and there may exist the laser ablation during 16 the laser cleaning period. Clearly, from the mapping, it can be seen the distribution of the O and C 17 is relatively homogeneous. As per the surface morphology of laser cleaned surface, it is suggested 18 that the convex exposed on the laser cleaned surface is rich in the O and C elements, whereas they 19 are almost absent in the center. This may be attributed to the UV laser satisfied Gaussian 20 distribution and the energy density at the center of the spot is greater than that at the edge, which 21 13 results in the O element exposed in the center much more the edges. That is the reason why the 1 microstructure could generate the ring-shaped microstructure and corrugated shape morphology 2 respectively. Herein, the present findings are very encouraging and it can also infer that the 3 nanosecond UV laser could successfully strip off the painting layer thoroughly. 4

Hardness characterization 6
It is widely recognized that evaluate the mechanical properties exactly, the hardness 7 characterization is an indispensable, essential evaluation index, which needs to be considered 8 during the UV laser cleaning the painting layer. In this study, the Vickers microhardness test is 9 performed to investigate the mechanical properties, the details including the load of 290 g and 10 holding time duration of 15 s, which can be demonstrated in Fig. 7. As such, from the curves, it 11 can be noticed that the UV laser cleaned surface increases to 150 HV while the uncleaned surface 12 is 92 HV. As for the dramatically microhardness enhancement can be attributed to the fact that the 13 UV laser cleaned surface shows up rapid melting and quenching, which is followed by generating 14 the microstructure, including ring-shaped and corrugated shaped. The other reason is due to the 15 fact that the UV laser cleaning could induce the resultant surface producing dislocation density, 16 while the traditional cleaning techniques are not available. This is supported by the previous results 17 Vickers microhardness UV laser cleaned surface is 178 HV and the minimum is 150 HV. Thus, it 20 is suggested that UV laser cleaning painting layers could enhance the surface microhardness 21 considerably. Aside from the environmentally friendliness, this is another reason why authors take 1 this prospective method to remove the painting layer. commensurately tensile stress and strain curves are illustrated in Fig. 8. From the curves, it can be 6 exhibited five stages in these tensile curves, such as elastic deformation, yield deformation, plastic 7 deformation, necking and fracture respectively [50-56]. The curves are just like the parabolic shape 8 and there is a rapid increase in the elastic and yield deformation stages. Followed by gradually 9 enhanced to the plastic deformation stage, there exist maximum values, viz., ultimate strength 10 appearing before reaching necking period, along with the reduction to facture stage sharply. 11 Notably, the tensile strength of laser cleaned surface is much stronger than that of before cleaning, 12 which is mainly due to the laser treatment is conducive to enhance the elastic and plastic 13 deformation properties of the substrate. 14 As for bending properties, it is widely recognized bending displacements and bending stress 15 are essential factors, which can be exhibited in Fig. 9. As noted in the curves, it can be observed 16 that both of the resulting laser cleaned surface and uncleaned surface experience complete elastic 17 deformation and plastic deformation stage, whereas the fracture stage does not exist in the sample, 18 in spite of the bending angles exceeding 90° and bending stress over 1000 MPa. In this regard, it 19 indicates that both of the laser cleaned surface and uncleaned surface have excellent bending 20 strength and plasticity. Yet, laser cleaned surface manifests as better plasticity in the plastic 21 deformation stage. Thus, the laser cleaned surface conduces to improve the bending strength and 1 plasticity properties. 2

Roughness and profile characterizations 3
As seen in Fig. 10, it presents typical 3D morphologies along with corresponding line scanned 4 profiles of original painting layer and resulting laser cleaned surface, which is examined by optical 5 profiler and commensurately captured zone is 840 μm × 840 μm. As evident from Fig. 10 (a), it 6 can be seen that the original painting layer is relatively evenly and smooth, along with the surface 7 roughness 1.968 μm. Of particular note, white lines indicate maximum height difference is 8 approximately 10 μm at original surfaces. With regard to the laser fluence 2 J/cm 2 , the maximum 9 height difference is five times larger than that of the original painting layer surface and 10 corresponding surface roughness is 12.751 μm. This may be contributed to the laser energy is less 11 capable to remove the painting layer directly and there exists laser cleaning induced cracks and 12 concaves in this layer, which is well agreed with the prior mentioned Fig. 3 (a). Thus, the roughness 13 of laser cleaned surface at laser fluence 2 J/cm 2 increases dramatically. Specifically, as for laser 14 cleaned surface at laser fluence of 5 J/cm 2 , the resultant surface is pretty smooth, viz. the surface 15 roughness is 2.471 μm and relevant maximum height difference is approximately 6 μm, which 16 indicates laser fluence at 5 J/cm 2 is the most suitable for UV laser cleaning painting layers. As 17 such, it is suggested that laser cleaned surface at this fluence without destroying the underlying 18 substrate and the thermal ablation is minimal from the aforementioned surface morphology in Fig.  19 16 morphology, which coincides well with Fig. 3 (d). In comparison, the laser cleaned surface at laser 1 fluence 7 J/cm 2 has a relatively rough surface (6.298 μm) and the maximum height difference is 2 about 27 μm, which is suggested that the surface experiences the excessive ablation and it is 3 consistent with the captured images in Fig. 3 (g). Thus, it is imperative to avoid the ablative 4 conditions occurred during UV laser cleaning painting layer as far as possible. 5

Theoretical model analysis 6
The experimental results illustrate initial cleaning threshold and complete cleaning threshold 7 based on laser and painting layer interactions. In order to explain the phenomena more detailly, it 8 is necessary to establish a thermodynamic model to describe the laser cleaning mechanism. Based 9 on the study of Zhang et al. [57], the theoretical relationship between the temperature and energy Afterwards, the laser fluence can be deduced from Eq. (7) and the relationship is expressed 13 as: 14 Correspondingly, the laser fluence of Fe substrate can be illustrated as follows: 16 Wherein, 1 , 2 , 1 , 2 are densities of painting layer and Fe substrate as well as molar 18 masses of painting layer and Fe substrate, respectively. 19 Inspired from Eqs. (7), (8) and substituted the corresponding physical constants shown in 20 Table 4, it can be calculated the theoretical laser fluence is 1.789 J/cm 2 if temperature approaches 1 its melting points. As for the Fe substrate, it also can be derived from functions that laser fluence 2 are 3.216 J/cm 2 and 4.65 J/cm 2 with regard to the melting point and boiling point, respectively. It 3 is noteworthy that the calculated theoretically threshold (4.65 J/cm 2 ) is exceeded initial cleaning 4 threshold, which is mainly due to the fact that less consideration of the plasma shielding effects 5 and thermal expansion effects as well as various thickness of painting layer. These points will be 6 taken into account in future studies to modify this model effectively. 7 Comprehensively, the laser cleaning is a facile, environmentally friendly and promising 8 method to strip off the painting layer from the marine engineering surface. Hopefully, this study 9 would provide an experimental and theoretically analysis reference in the UV laser cleaning the 10 painting layer and pave the way for any further potential applications in industrial field. 11

CONCLUDING REMARKS 12
In this paper, a study based on the nanosecond UV laser cleaning method is innovatively 13 proposed, which is successfully utilized to strip off the painting layer from the AH36 steel substrate. 14 This study innovatively verifies the feasibility of the UV laser cleaning the painting layer on the 15 surface of AH36 steel in theory and briefly elaborates the primarily interaction mechanism of UV 16 laser, such as the photothermal and photochemical interactions. The thermal dynamic model is 17 bending strength dramatically. This can be attributed to the UV laser cleaned surface experiencing 7 rapid melting and quenching, followed by generating the ring-shaped and corrugated shaped 8 microstructure as well as the produced dislocation density. Therefore, this promising UV laser 9 cleaning method is not only environmentally friendly, but also enhances the mechanical properties 10 of laser cleaned surface significantly. Hopefully, there is a great potential to utilize this promising 11 method to large-scale cleaning the painting layer of the marine engineering surface and make some 12 contributions to the marine and industrial fields in the future. and Yang Wang edit this paper. All authors approve this paper. 3 Declarations 4

Ethical approval
Not applicable 5

Consent to participate
Not applicable 6

Consent to publication
All presentations of case reports have consent for publication 7

Declaration of Competing Interest 8
The authors declare that they have no known competing financial interests or personal 9 relationships that could have appeared to influence the work reported in this paper.         Macroscopic three-and two-dimensional optical images of original surface (a, e, i) and laser cleaned surface at scanning speed of 1000 mm/s with different uence, among them (b, f, j) at laser uence of 2 J/cm2, (c, g, k) at laser uence of 5 J/cm2, (d, h, l) at laser uence of 7 J/cm2 respectively.     The microhardness characterization after UV laser cleaning.

Figure 8
The relationship between the strain and tensile stress.

Figure 9
The relationship between the bending displacements and bending stress.

Figure 11
Schematic diagram of laser cleaning painting layer model.