Investigation on the Deviation and Thermal Damage Effects in Laser-Induced Lateral Crack Propagation of Soda–Lime Glass

Abstract

This study is based on the laser-induced thermal-crack propagation (LITP) technology, focusing on the issues of deviation and thermal damage during the transverse crack propagation process, with the aim of achieving high-purity, non-destructive, and high-precision cutting of glass. A 50 W, 1064 nm fiber laser is used for S-pattern scanning cutting of soda–lime glass. A moving heat source model is established and analyzed via MATLAB R2022a numerical simulation. Combined with the ABAQUS 2019 software, the relationships among temperature field, stress field, crack propagation, and deviation during laser-induced thermal crack cutting are deeply explored. Meanwhile, laser thermal fracture experiments are also carried out. A confocal microscope detects glass surface morphology, cross-sectional roughness and hardness under different heat flux densities (HFLs), determining the heat flux density threshold affecting the glass surface quality. Through a comprehensive study of theory, simulation, and experiments, it is found that with an increase in the HFL value of the material, the laser-induced thermal crack propagation can be divided into four stages. When the heat flux density value is in the range of 47.2 to 472 W/m², the glass substrate has good cross-sectional characteristics. There is no ablation phenomenon, and the surface roughness of the cross-section is lower than 0.15 mm. The hardness decreases by 9.19% compared with the reference value.

1. Introduction

Soda–lime glass has the advantages of high transparency, mechanical strength, uniform texture, smooth surface, and corrosion resistance [1] and is widely used as substrate for display panels and optical devices, such as the LCD and OLED displays on smart phones, televisions, computer, and electrical cars. In the manufacturing process of these display panels, the cutting of the glass is crucial [2]. The traditional glass-cutting method involves first scoring the glass surface with a diamond or hard metal wheel, and then breaking it by mechanical means [3,4]. However, this processing method is prone to issues such as the splashing of cutting powder, irregular cutting edges, the generation of micro-cracks, and even the breakage of the glass [5,6]. These problems seriously affect the cleanliness of the processing and reduce the quality of the glass cutting, thus influencing the quality of the display panel evaporation and the display effect. Therefore, a glass-cutting processing technology with high cleanliness, no pollution, and no defects has significant theoretical and practical demands.

As a non-contact micromachining method, laser cutting technology [7,8] has a small heat-affected zone, and the fracture surface is free of slag debris or microcracks, resulting in exceptionally high surface quality and mechanical strength. It has been widely used in various material processing fields.

Numerous scholars have conducted multi-dimensional studies on laser glass cutting. Some focused on optimizing process parameters (e.g., laser power, pulse frequency, scanning speed) to enhance cutting efficiency and precision while reducing defects like micro-cracks and chipping [9,10]. Liu et al. [11] used a Bessel beam picosecond laser for glass cutting and analyzed the influence of the focal position, speed, and power on the cutting roughness. A CO₂ laser with a wavelength of 10.6 μm and a maximum power of 100 W was used to separate the glass, and the glass cracked due to the internal stress generated by heating. By adjusting the speed, power, and focusing position, high-quality cutting with an edge breakage of less than 3 μm was achieved. Wlodarczyk et al. [12] explored the feasibility of picosecond laser cutting and drilling of 50 μm and 100 μm thick AF32^®Eco thin glass sheets at different wavelengths. The results show that the effective cutting speed is the highest at a wavelength of 1030 nm, the cutting quality is the best at a wavelength of 343 nm, and the cutting quality at a wavelength of 515 nm is good with a smaller heat-affected zone. The wavelengths of 343 nm and 515 nm can also drill micro-holes on a 100 μm thick TFG substrate. Dario et al. [13] used a near-infrared picosecond pulse laser with an average power of 200 W to cut soda–lime glass with a relatively large thickness (1 mm) while immersed in water. The influence of different pulse repetition frequencies, pulse energies, and scanning speeds on the cutting process and the quality of the cutting notch was studied. The results show that under the parameters of 400 KHz and 360 μJ, when the scanning speed changes from 0.5 m/s to 3 m/s, complete cutting of the 1 mm-thick glass is achieved after 300 to 1900 scans, respectively. Luo et al. [14] used an ultrafast laser-based composite cutting method for cutting alkali-free borosilicate glass. The effects of laser pulse duration, laser pulse energy, and pulse interval rate on the internal modification area of the ultra-thin glass by laser were studied. Considering both the cross-section roughness and edge strength, a composite laser cutting process with a punched hole-type groove structure was adopted. As a result, a maximum edge strength of 369 MPa and a minimum cross-section roughness of 0.148 μm were achieved, enabling high-quality cutting of ultra-thin glass.

Some scholars have utilized a 532 nm nanosecond laser to cut solar glass and explored the influence of the cutting path on the surface quality. Dong et al. [15,16] employed an X–Y spiral cutting method from the bottom upwards and carried out nine groups of experiments with different cutting angles ranging from 5° to 45°. The experimental results revealed that the error was distributed between −0.12° and 0.07°, exhibiting a random characteristic and being independent of the cutting angle. The roughness at the extreme corner positions was randomly distributed between 3.55 μm and 5.71 μm. Chen et al. [17] investigated the influence of the spiral geometric parameters (width and overlap rate) on the surface quality. Through experiments, three dropping modes, namely smooth dropping, serrated crack, and snowflake-like crack, were summarized, and the optimized range of the spiral parameters (0.45 mm ≤ w ≤ 0.55 mm, 65% ≤ r ≤ 80%) for ensuring the stable dropping of the glass was obtained. Li et al. [18,19] found that when the laser pulse repetition frequency was 55 kHz, the corresponding laser power was 22.6 W, the scanning speed was 300 mm/s, and the spiral trajectory parameters (spiral width and spiral overlap rate) were 0.45 mm and 70%, respectively, the minimum chipping width could reach 104.81 μm, and the cutting efficiency was relatively high. It was also found that in the concentric circle scanning mode, the minimum edge chipping area and the minimum surface roughness were reduced by 9.4% and 16.4%, respectively.

In particular, laser-induced thermal-crack propagation (LITP) technology utilizes the thermal stress generated by the laser’s thermal effect to induce internal cracks within the material. During the crack propagation process, only material separation occurs without material removal, which not only prevents material damage but also significantly improves material utilization [20,21]. It is widely recognized as the cutting method that provides the highest cross-sectional quality for display panels. Figure 1 shows the principle of laser cutting in LITP technology. The laser beam undergoes partial transmission through the glass substrate. As the glass has a uniform absorption coefficient, heat accumulates near its upper surface, creating stress and cracks. These cracks then spread along the path of the scanning laser beam. Figure 1a–c shows the diagrams of the thermal-cracking process.

The coefficient of thermal expansion has a profound influence on the crack expansion process in LITP. Yamamoto et al. [22,23,24,25] conducted in-depth research on the cutting of glass by CO₂ laser-induced thermal cracking from 2008 to 2010. By comparing the two-dimensional thermoelastic simulations and experimental results of three materials, namely soda–lime glass, aluminosilicate glass, and quartz glass, they studied the influence of the coefficient of thermal expansion of the materials on laser-induced thermal-cracking cutting. It was concluded that only when the coefficient of thermal expansion is greater than a given threshold, and the softening temperature and thermal stability temperature of the glass are higher than certain thresholds, can this material be processed by laser-induced thermal-cracking cutting. Mishchik et al. [26] employed a 1030 nm wavelength femtosecond laser to induce thermal cracking at a speed of 1 mm/s to cut a 1 mm thick soda–lime glass. They adapted the principle of filamentary laser cutting and enhanced it by generating high temperatures in a filamentary region of a considerable length outside the focal point of the beam. This resulted in thermal expansion of the material, enabling the induced thermal-cracking cutting process. Karube et al. [27] presented a summary of the technical difficulties associated with laser-induced thermal fracture monolithic cutting of glass and the proposed solutions. It was determined that the optimal light source for laser-induced thermal fracture monolithic cutting is the body absorption mode. Furthermore, the tensile stress induced by the expansion of the material under the constrained state of the glass is insufficient to cause the material to fracture. Consequently, the triggering crack must be prefabricated.

The dynamic change of laser parameters has a direct and significant effect on crack extension. Miyashita et al. [28] investigated the crack nucleation and extension behavior during laser cutting of soda–lime glass. The research revealed that manipulating the laser parameters enables control over the crack propagation path by regulating the stress distribution ahead of the crack tip. Furumoto et al. [29] employed a CO₂ laser to induce thermal fracture in 0.7 mm thick chemically strengthened glass, investigating the law of crack propagation. They utilized a 532 nm short-pulse laser to generate grooves on the surface of the strengthened glass, initiating crack propagation. The initial grooves produced by these methods often yield a significant quantity of minute chips.

In the process of laser heating, a significant temperature gradient will be formed inside the material, resulting in uneven thermal stress distribution. Cracks often start to sprout from the parts where the temperature gradient is large and the thermal stress is concentrated, and expand along the direction of thermal stress. Kondratenko et al. [30] used a CO₂ laser to irradiate a glass substrate in order to study the temperature field distribution after the laser beam stream had irradiated the float glass during the thermal-cracking method of cutting. The results demonstrated that the temperature gradient during the cutting process is the key factor affecting the quality of the float glass. Cai et al. [31,32] deeply analyzed the cutting mechanism of LITP. He proposed a three-dimensional model for calculating heat generation and applied it to finite element analysis, thus drawing conclusions regarding the temperature and stress distributions, as well as the crack morphology. Moreover, the high cutting quality was verified through experiments.

In conclusion, when domestic and foreign scholars study the use of laser-induced thermal processing (LITP) technology for glass cutting, they mainly focus on adjusting laser parameters, improving processes, and exploring the heat transfer and stress distribution inside the glass through the analysis of the temperature field and stress field. However, in practical applications, the heat flux density directly affects the direction and propagation speed of cracks, which in turn determine the quality of the cutting surface. This study is based on the laser-induced thermal-crack propagation (LITP) technology to investigate the deviation and thermal damage during the transverse propagation of cracks, aiming to achieve high-purity, non-destructive, and high-precision cutting of glass.

2. Moving Heat Source Model

In this study, the fiber laser is considered as a heat source that moves uniformly in the thickness direction of the panel. The absorption of laser energy by glass exhibits volume absorption characteristics, and its energy distribution conforms to the Gaussian distribution model. To get the volume absorption heat source model of a 1064 nm laser for soda–lime flat glass, a rectangular coordinate system is set up with the laser spot center on the material surface as the origin. The material surface is the x–y plane, and the optical axis is the z axis. Then, taking the rectangular coordinate system’s origin O (O, x, y, z) as the origin and the z axis as the h axis, a cylindrical coordinate system (O, r, u, h) is established, as shown in Figure 2.

According to Dickson’s theory [33], the heat flux density q(r, u, h) at any position (r, u, h) in the plane perpendicular to the direction of the beam propagation can be expressed as follows:

$$\begin{array}{l}q(r,u,h)=2q(0,0,h){\mathrm{e}}^{\left[-2{\displaystyle \frac{r^2}{{R_h}^2}}\right]}\\ R_h=R_0+\tan \beta \times h\end{array}$$

(1)

where q(0, 0, h) is the heat flux density at the center of the laser beam at the depth of h; R_h is the radius of the laser beam at the height of h; R₀ is the beam radius of the laser incident plane; and β is the angle between the optical axis of laser beam and the edge, due to the focus effect on the beam.

According to the Beer–Lambert law [34,35], when a laser passes through a substance, the laser intensity exponential decays along the propagation direction. The heat flux density q(0, 0, h) at the center of any cross section can be expressed as follows:

$$q(0,0,h)=q(0,0,0)e^{-\alpha h}$$

(2)

where α is the absorption coefficient of the glass at the laser wavelength.

After the total power of the laser is reflected by the glass, the energy incident into the glass will be completely absorbed within the absorption depth of the material and it is the integral of the heat flux density, that is, the integral holds:

$$(1-\eta )P={\displaystyle {\int }_0^{\frac{1}{\alpha }}{\displaystyle {\int }_0^{2\pi }{\displaystyle {\int }_0^{R_h}q(r,\theta ,h)}}}drd\theta dh$$

(3)

where η is the reflectivity of soda–lime glass to 1064 nm laser.

The relationship between the laser power P and the heat flow density q(0, 0, 0) at the center of glass surface is as follows:

$$(1-\eta )P=4\pi q(0,0,0){\displaystyle {\int }_0^{\frac{1}{\alpha }}{\displaystyle {\int }_0^{R_h}{e}^{-\alpha h}}}\times e^{\left[-2{\displaystyle \frac{r^2}{{R_h}^2}}\right]}\times rdrdh$$

(4)

Comprehensively considering the laser absorption rate of the glass, as well as the absorption of the lens during laser transmission and the loss coefficient of reflection at the panel, the expression for q(0,0,0) can be obtained as follows:

$$q(0,0,0)={\displaystyle \frac{3\eta P}{\pi h_0{R_0}^2}}$$

(5)

where R₀ is the radius of the laser spot and h₀ is the thickness of the glass.

By solving the heat flow density q(0, 0, 0) at the center of laser beam corresponding to different diameters and powers, the heat flow density q(x, y, z) at any position on the soda–lime glass under laser irradiation (in the form of a soda–lime body absorption) can be obtained.

$$q(x,y,z)=q(0,0,0)e^{\left(-{\displaystyle \frac{8\left\{x^2+{(u-vt)}^2\right\}}{h_z}}\right)}$$

(6)

The angle β between the laser axis and the beam edge is approximately 90°, and the vertical component (h_z) of the beam diameter is constant at the center of the incident surface of the material.

3. Simulation Modeling

Considering the steps of the basic heat transfer process of LITP during the cutting of soda–lime flat glass, this study uses the finite element simulation software ABAQUS to establish its physical model. The thermal conductivity, specific heat, Poisson’s ratio, and Young’s modulus of soda–lime glass vary with temperature. The physical and optical properties of soda–lime glass are shown in

Table 1. Physical and optical properties of soda–lime glass.

Material Properties	Values
Density (g/cm3)	2.53
Young’s modulus (GPa)	74
Poisson’s ratio	0.23
Softening point/TBDT (°C)	586
Heat transfer coefficient (W·K−1·m−2)	10
Laser absorption rate/η	3.4%
Specific heat [J/(kg·K)]	828–1160
Expansion coefficient (K−1)	8 × 10−6–10−5
Thermal conductivity [W/(m·K)]	1.06–1.38

[36]. The initial temperature is set at 20 °C, and the volumetric heat source is loaded along the laser scanning path.

The moving heat source model for laser-induced thermal-cracking cutting is coded into ABAQUS’s DFLUX subroutine using FORTRAN. This allows accurate calculation of the temperature field distribution. Then, the stress field is derived indirectly, ignoring the crack propagation effect during simulation. In the transient thermal analysis, it is assumed that the glass material is homogeneous and isotropic, and the laser energy density is uniformly distributed. In the thermal stress analysis, assumptions include: the glass is homogeneous and isotropic, the glass is homogeneous and isotropic, has no residual stress, and its stress-strain relationship is linearly elastic.

Figure 3 shows a finite element simulation system based on the volumetric absorption heat source model. The geometric model of the glass substrate has dimensions of 30 mm × 10 mm × 0.5 mm and is discretized using eight-node hexahedral elements.

In this study, through MATLAB numerical simulation, a level grid system (of quadratic geometric refinement) was constructed based on the reference glass thickness. Under the conditions of a power of 50 W and a spot diameter of 0.5 mm, the simulation shows (Figure 4) that the HFL reaches its peak in the central region and decays in a Gaussian distribution outward. The results of numerical integration (Figure 5) indicate that the HFL values gradually converge with the grid refinement. The difference between the calculation results of the h₀/4 (0.125 mm) and h₀/8 (0.0625 mm) grids is only 2.7%, but the calculation time is reduced by 75%. Considering both accuracy and efficiency comprehensively, the 0.125 mm grid meets the engineering requirements.

The element size is 0.125 mm, and finally, a convergent mesh system containing 97,605 nodes and 76,800 elements is generated. In the numerical model, the moving laser heat source is defined as a dynamic load scanned along the z-axis direction (thickness direction), and the diameter D of its effective action area has a strict corresponding relationship with the measured laser spot size. To achieve the regulation of heat flux density over a larger range, the simulation conditions are set as shown in

Table 2. Parameter setting in simulation.

Parameters	Value
Laser power, P/(W)	50
Feed speed, v/(mm/s)	0.2
Laser beam diameter, D/(mm)	0.5

.

4. Simulation Results and Discussions

4.1. Temperature Distribution

In Figure 6a, six monitoring positions for the laser cutting-in points are uniformly arranged along the beam trajectory, and the temperature distribution law of the laser spot is shown: the temperature at the center of the laser spot is the highest, and the temperature decreases as the distance from the laser spot increases. Figure 6b presents the temperature response curves at different positions on the laser path. The result analysis shows that in the initial stage of laser entry, the cumulative effect of incident energy is significant, and the peak temperature rapidly rises to 152 °C; after entering the steady-state processing stage, due to the dynamic balance of the heat conduction mechanism, the temperature stabilizes in the range of 130 ± 3 °C, and there is a temperature rise again in the stage of the laser exit point. The transition from the heat conduction dominant mode to the mode dominated by natural convection leads to a decrease in the heat dissipation efficiency. Finally, the monitoring point records a critical temperature of 181 °C, which is an increase of 20.4% compared with the temperature reference value in the initial stage. Figure 6c gives the simulated temperature distribution curves within the range of 0.01–0.2 mm/s. It can be seen that even when the laser cutting speed is 0.01 mm/s, the peak temperature only reaches about 200 °C (30% of the softening point temperature). This value has a significant safety margin compared with the softening point temperature of the display panel substrate (586 °C), avoiding its transition from the solid state to the viscous flow state under the action of the laser. The heat flux density (HFL) is calculated based on the laser power of 50 W and spot diameter of 0.5 mm, corresponding to the non-destructive cutting stage (472 W/m²).

4.2. Stress Field

Figure 7 shows the distribution characteristics and cross-sectional characteristics of the Sxx stress field on the surface of the glass in the direction perpendicular to the laser path at t = 70 s. As can be seen from the figure, a significant tensile stress field is formed at the laser entry point, providing the driving conditions for the transverse propagation of cracks. A compressive stress state is present at the center of the laser spot, and it shows a distribution pattern with higher values in the middle and lower values on both sides in the thickness direction. A uniformly distributed tensile stress state is observed at a certain distance in front of the laser spot. This uniform stress distribution can control the propagation direction of thermal cracks and ensure the flatness of the cross-section. According to the size of glass in the paper, 30 mm × 10 mm × 0.5 mm, 1 mm is selected as the appropriate scale.

Figure 8 shows the evolution law of the stress distribution along the glass cross-section at the center of the laser spot (point B-B) and 3 mm ahead of it (point A-A). As can be seen from the figure, during the process of the laser scanning along the preset path, the stress state at any measurement point undergoes a dynamic evolution process. Specifically, the initial tensile stress gradually accumulates and increases, then transitions to a compressive stress state, and finally forms a stable tensile stress distribution pattern again. From the curve of stress in each cross-section changing with time, at time t1, point B at the center of the laser spot reaches the maximum amplitude of compressive stress. At this time, point A in front of it simultaneously enters the peak state of tensile stress. When the laser spot moves to the position at time t2 and covers point A, point A changes from being dominated by compressive stress to being dominated by tensile stress. Meanwhile, point B shows the characteristics of tensile stress again, but its intensity is significantly lower than the historical peak.

Stress analysis shows that the synergistic effect of tensile stress and compressive stress forms a balanced effect of the dual-mode crack propagation of “driving-restraining”, ensuring the stable transverse propagation of cracks along the preset path. The characteristic of the stress distribution with compressive stress in the front and tensile stress in the rear precisely restricts the crack propagation direction within the range of the laser path, avoiding the damage of the substrate structure of the display panel caused by bifurcated cracks. The uniformity of the stress field distribution reduces the risk of local energy concentration. These three factors jointly construct a cutting regulation system of “controllable path-integral structure-balanced energy”, which is the prerequisite for achieving the straight propagation of thermal cracks.

5. Experiment Validation and Analysis of the Heat Flux Density Threshold

5.1. Experimental Setup

The experiment uses a PEDB-400F pulsed fiber laser (Wuhan Perfect Laser Co., Ltd., Wuhan, China), with parameters: 50 W average power, 1064 nm wavelength, 50–100 kHz repetition frequency, 360 mm focal length, and 0.1–0.3 mm numerical aperture. A soda–lime flat glass with dimensions of 30 mm × 10 mm × 0.5 mm was taken as the research object. To ensure the stability of the workpiece, it was fixed on the workbench with an aluminum plate and gaskets, and the temperature was monitored in real time by an infrared thermal imager, which was manufactured by FLIR Systems, with the model number FLIR A325sc, a wavelength range of 7.5–14 μm, and an accuracy of less than 1 °C in an indoor environment. The experimental setup is shown in Figure 9a. During thermal imaging, the glass is temporarily placed on a quartz substrate to maintain uniform heat distribution. This dual-setup ensures both thermal measurement accuracy and cutting stability. Due to the limitations of the equipment, after being focused, the laser formed a spot with a diameter of 0.036 mm and accurately irradiated the surface of the glass. In order to expand the area of thermal processing, an S-shaped path filling strategy was adopted to simulate the linear movement of the spots. The scanning line spacing was 1 mm, and the path diagram is shown in Figure 9b. After the cutting was completed, a laser confocal microscope was used to observe the cutting section, analyze the morphological features and ablation conditions, and evaluate the cutting quality.

5.2. Temperature Measurement

Figure 10a shows that the thermal-imaging system accurately captures the spatial evolution law of the thermal action area along the cutting path. Experimental observations indicate that the lateral expansion range of the heat-affected zone is approximately 10 mm, accounting for 33.3% of the total width of the substrate. The temperature evolution curve exhibits typical three-stage characteristics: the temperature gradient reaches 162 °C in the initial stage, the temperature remains around 140 °C in the stable stage, and the temperature rises again to 172.0 °C at the end stage when the laser cutting is finished. This dynamic characteristic is in good agreement with the results of the finite element analysis (the relative error is less than 5.88%). Figure 10b shows that there is a systematic deviation of 3.2–4.1 °C between the temperature measurement results using the contact-type thermocouple wire and the simulation results, and the relative error is strictly controlled within the range of 1.7%–2.3%. Figure 10c shows a comparison chart of measured temperature values and simulated temperature values under four groups of different process parameters. The relative errors are 0.1%, 0.8%, 3.1% and 5.4% respectively, and the error range is controlled within 6%, verifying the accuracy of the heat flux density model. This further proves the accuracy of the finite element simulation model. The discrepancy is attributed to minor variations in laser absorption rate during experiments. Heat flux density values are derived from the laser power and spot diameter.

5.3. Analysis of the Heat Flux Density Threshold for Crack Propagation

The flatness of the cutting line on the glass surface is characterized by measuring the maximum height difference (Sz), which is the sum of the maximum peak height value (Sp) and the maximum valley depth value (Sv) of the cutting cross-section profile. The distribution of Sz with respect to the heat flux density value is shown in Figure 11. The experimental results show that when the HFL value is increased to 787 W/m², the cutting cross-section deteriorates significantly. Specifically, when the HFL is lower than this value, the cutting cross-section is smooth and the edge is neat. The cutting line on the upper surface of the cut is thin and straight, without significant deformation, and the maximum height difference is about 1 mm. As the HFL value gradually increases and exceeds this value, the maximum height difference of the cutting cross-section gradually increases to more than 6 μm. At this time, an irregular wavy morphology appears on the cross-section, the surface melting marks are obvious, and the cutting line trajectory is irregular. This is because an excessively high HFL value will intensify the thermal accumulation effect inside the glass, and the thermal stress on the glass surface during the cutting process increases significantly, resulting in local melting and rapid solidification of the glass surface, thus forming irregular ripples and melting marks. Through comprehensive analysis of the cross-section quality, it is found that when the HFL value is in the range of 472– 787 W/m², the glass cutting process is mainly dominated by the local thermal effect mechanism. Although the laser energy input at this stage forms ablation spots on the glass cross-section and is accompanied by thermochromic phenomena, the maximum height difference Sz of the cross-section always remains stable at about 1 mm within this parameter range, indicating that the thermal damage to the cross-section at this time does not cause macroscopic morphological instability, and the cutting line is still in a straight state.

In order to more intuitively reflect the quality of the transverse propagation of the cutting cracks on the display panel substrate, Figure 12 shows the distribution of the cross-sectional roughness (Sa) under different HFL values. It can be seen from the figure that as the HFL increases, laser-induced thermal crack propagation evolves through four stages: below 47.2 W/m², the thermal stress field is insufficient to overcome initial crack tip stress, causing ineffective or incomplete crack propagation; when the HFL value is in the range of 47.2 W/m² to 472 W/m², the glass substrate exhibits good cross-sectional characteristics at this time, with no ablation phenomenon occurring, and the cross-sectional roughness is lower than 0.15 mm; when the HFL value is in the range of 472 W/m² to 1.18 × 10³ W/m², the glass cross-section is in a slight ablation stage at this time, with some ablation spots and thermochromic regions existing, and the cross-sectional roughness is still controlled within 0.25 mm; when the HFL value exceeds 1.18 × 10³ W/m², the glass cross-section shows a serious ablation phenomenon at this time, the ablation and thermochromic regions are further expanded, melting and ablation occur at the glass edge, the cross-sectional roughness surges to above 0.5 mm, increasing by 123% compared with the third stage, and the cross-sectional quality deteriorates seriously.

This study measured the surface and cross-sectional hardness of the display panel substrate after transverse crack propagation using a Vickers hardness tester, with test parameters of a load of 200 gf and a dwell time of 15 seconds. Each indentation was imaged at a magnification of 500× using a laser confocal microscope to evaluate surface integrity, ensuring no millimeter-scale cracks within the non-destructive cutting range. A blank control group was first set up. The reference hardness value of the display panel substrate was measured to be 710 N/mm². Secondly, along the horizontal direction of the thermal crack line, seven groups of sampling points were set at intervals of 1 mm (1–7 mm from the crack line), covering the heat-affected zone to the undamaged area, so as to analyze the hardness gradient distribution. Selection of the cross-section indentation area: Along the thickness direction of the crack cross-section, five groups of sampling points were set at intervals of 0.08 mm (from the surface layer to the center of the cross-section), with a focus on the depth of the thermal damage layer. Each sampling point was measured three times, and the average value was taken as the final result. The error range is controlled within 5%. The results of the hardness curve are shown in Figure 13.

It can be seen from Figure 13a that the hardness value of the surface of the display panel substrate is positively correlated with the distance from the thermal crack line, and the hardness value gradually approaches the reference value in the area far away from the thermal crack line. When the HFL is less than 472 W/m², the average hardness is 690 N/mm², which is only a 2.8% decrease compared with the reference value, indicating that the display panel substrate can maintain excellent structural integrity within the non-destructive cutting range. When the HFL is greater than 472 W/m², the hardness of the area close to the thermal crack line decreases significantly. The lowest hardness value is 603 N/mm², which is a 15.1% decrease compared with the reference value. This is because the generated thermal damage leads to structural defects such as micro-cracks and atomic bond fractures within the material, weakening its ability to resist local brittle fracture. As can be seen from Figure 13b, the cross-sectional hardness shows a distribution pattern of being higher in the middle and lower on both sides along the thickness direction. The average value in the non-destructive cutting range is 653 N/mm², which is a decrease of 9.19% compared with the reference value. When the HFL is greater than 472, the lowest hardness value is 551 N/mm², and the hardness of the crack cross-section is 22.3% lower than the reference value. A comprehensive analysis indicates that within the non-destructive cutting range, the W/m² decrease in the hardness of both the surface and the cross-section of the material is within 10%, having little impact on the material properties.

Figure 14 shows microscopic photographs of Vickers indentations with different HFL values. In Figure 14a, when HFL = 472 W/m², the indentation center exhibits a distinct rhombus shape with smooth edges and no burrs. The four corners of the indentation are sharp, and no crack propagation is observed. The glass surface is flat without obvious signs of ablation or melting, and the uniform texture of the glass matrix is clearly visible. This indicates no significant damage, which is consistent with the 9.19% decrease in hardness (Figure 13). In contrast, in Figure 14b, when HFL = 1181 W/m², the edges of the Vickers indentation are irregular, and the four corners show notches. Radial micro-cracks can be seen around the indentation, extending outward from the edges, along with molten bulges or depressions, indicating a 22.3% decrease in hardness.

6. Conclusions

In this study, through thermo-mechanical coupling finite element simulation and laser-cutting experiments, the distribution characteristics of the temperature field and stress field during the laser-induced thermal-cracking process are systematically analyzed, and the crack propagation mechanism under the synergistic action of tensile and compressive stresses is revealed to ensure the straight propagation of cracks. The influence of the heat flux density value on the maximum height difference of the crack cross-section profile (Sz), the cross-section roughness (Sa), and the degree of hardness reduction is systematically evaluated, and its significant influence and variation laws are revealed. The critical heat flux density for crack propagation and the heat flux density threshold without thermal damage are taken as the parameter ranges for precision cutting. The conclusions are as follows:

(1) Within the experimental laser cutting speed range of 0.01 to 0.2 mm/s, the peak temperature generated in the display panel substrate only reaches about 200 °C (30% of the softening point temperature). This value has a significant safety margin compared with the softening point temperature of the display panel substrate (586 °C), preventing it from transforming from the solid state to the viscous flow state under the action of the laser.

(2) A “driving-restraining” dual-mode balance is formed through the synergism of tensile/compressive stresses to control the crack propagation along the preset path. The stress distribution mode with compressive stress in the front and tensile stress in the rear restricts the crack direction and suppresses crack bifurcation damage. The uniform stress distribution reduces the risk of energy concentration.

(3) With an increase in the HFL value, the laser-induced thermal crack propagation can be divided into four stages: incomplete crack propagation, non-destructive cutting, slight ablation, and severe damage. In the non-destructive cutting stage (47.2 to 472 W/m²), the cutting cross-section is smooth, flat, and free of damage, and the surface roughness of the cross-section is lower than 0.15 mm. The hardness decreases by 9.19% compared with the reference value.

(4) For industrial applications, the non-destructive cutting stage (HFL: 47.2–472 W/m²) is recommended for high-precision display panel manufacturing, as it balances cutting quality and efficiency. Future research could explore the integration of LITP technology with real-time temperature monitoring systems to further optimize crack propagation control.