Parametrically Engineered Laser-Induced Graphene with Tunable Electrical and Morphological Properties for Flexible Sensor Applications

Abstract

Laser-induced graphene (LIG) provides a scalable and maskless route for fabricating three-dimensional porous graphene on polymer substrates for flexible electronics. However, achieving reproducible control over its electrical and structural properties remains challenging due to the complex interplay of laser processing parameters. Here, we systematically investigate the effects of laser power, scan speed, dots per inch (DPI), energy per unit area (EPUA), and depth of focus (DOF) on the morphology, crystallinity, and conductivity of LIG formed on polyimide substrates. The optimized LIG exhibits excellent structural quality (I _D/I _G ∼ 0.25, L_a ∼ 76 nm) and low sheet resistance (∼10 Ω/□), within the lower range of reported values. The material is further validated through flexible piezoresistive sensors and electrothermal heaters. The pressure sensor demonstrates two distinct sensitivity regimes (1.25% N^1– and 32% N^1–), a low limit of detection (∼150 mN), and minimal hysteresis (∼5–6%). Piezoresistive behavior is governed by deformation-induced modulation of conductive pathways involving interflake contact resistance, tunneling effects, and microstructural rearrangement within the porous LIG network. The LIG-based microheater achieves a temperature of 201 °C within 6.2 s at a low operating voltage of 6.5 V, outperforming many reported flexible heaters in terms of heating efficiency. These results establish a transferable framework for engineering high-quality LIG with tunable properties and demonstrate its potential for scalable, low-power multifunctional flexible electronic applications.

1. Introduction

Laser-induced graphene (LIG) has emerged as a promising approach for the fabrication of three-dimensional (3D) graphene, offering a rapid, cost-effective, and scalable alternative to traditional complex techniques such as chemical vapor deposition (CVD) or mechanical exfoliation, which typically require multistep processing, high temperatures, or expensive precursors. LIG enables direct, maskless patterning of porous graphene on polymeric substrates under ambient conditions. In 2014, Tour et al. became the first to demonstrate the conversion of commercial polyimide (PI) tape into 3D porous graphene using a CO₂ laser under ambient conditions. Since then, significant research efforts have focused on tailoring LIG’s electrical and chemical properties through variations in laser processing parameters (laser scan speed, laser power, dots per inch (DPI), and depth of focus). For example, Minhas-Khan et al. showed that laser power and laser scan speed directly influence the graphite domain size and sheet resistance of the LIG. While Jo Y et al. reported that environmental atmospheres, such as inert gas or oxygen-rich conditions, can further modulate surface functionalities and conductivity, such variability introduces complexity and limits the reproducibility and scalability of LIG. Its porous nature with a high surface area has enabled applications in microsupercapacitors, , microfluidics, sensing, and a wearable heating system. This ease of tunable features has integrated LIG into multiple application domains. It should be noted that comparisons across different laser systems (e.g., CO₂, femtosecond, and visible lasers) are meaningful only when performed under comparable substrate and processing conditions; therefore, most prior studies remain system-specific. Despite these advancements, there remains a need for a more systematic understanding of LIG formation. Most studies optimize individual laser parameters in isolation rather than accounting for their cross-interactions, which limits reproducibility. Luka et al. reported significant variation in the LIG quality for different substrates under the identical conditions, highlighting substrate dependence. Similarly, Stanford et al. observed poor gas sensing performance due to inadequate substrate considerations, while Wang et al. demonstrated that improper combinations of DPI and scan speed can lead to highly resistive networks. These findings indicate that LIG optimization is often empirical and application-specific, lacking a unified framework to understand the combined influence of processing parameters.

In contrast to prior studies that primarily focus on application-specific or single-parameter optimization, the novelty of this work lies in developing a transferable and scalable framework for LIG fabrication by introducing energy per unit area (EPUA) as a physically meaningful parameter that integrates laser power, laser scan speed, and DPI. Rather than targeting only the lowest sheet resistance, this study focuses on controlled and reproducible tuning of LIG properties across a wide processing window. This approach acknowledges the broad variation in reported electrical performance in the literature. Using this framework, we systematically investigate the effects of EPUA, DPI, depth of focus, and substrate support on the morphology, crystallinity, and electrical properties of LIG. The material is comprehensively characterized using Raman spectroscopy, XRD, FTIR, XPS, SEM, and TEM, along with thermal and mechanical analyses to evaluate stability. It is important to note that this study is experimentally limited to a CO₂ laser system and polyimide substrate; however, the proposed EPUA-based framework is adaptable and can be extended to other laser systems and precursor materials through appropriate parameter tuning. The optimized LIG is further validated through proof-of-concept devices, including a piezoresistive pressure sensor and a microheater, demonstrating reliable performance and scalability.

2. Materials and Methods

2.1. Materials and Instruments

For this study, various materials and equipment were utilized to fabricate and characterize the LIG. A polyamide sheet (PI) with a thickness of 75 μm, Kapton tape with a thickness of 100 μm, and aluminum (Al) tape were procured from Nano Global Technologies in Uttar Pradesh, India. The optimization of LIG parameters was conducted using a 100W CO2-based laser system (model UT-9060) provided by Unique Technologies India. The laser processing parameters were systematically varied, including laser power (5% to 45%), laser scan speed (10 to 100 mm/s), dot per inch (DPI) 50 to 500, and depth of focus (h, 3 to 10 mm), to investigate the influence on the structural, electrical, and morphological properties of LIG. Electrical characteristics of the fabricated sensors were evaluated using a Keithley 4200A-SCS Parameter Analyzer, while mechanical testing was performed using a Mark-10 Series-5 force gauge. Thermal stability and compositional analysis were evaluated using thermogravimetric analysis (TGA), and thermoelectrical response was recorded using an infrared camera (FLIR ETS 320, Teledyne). Morphological characterization was conducted using scanning electron microscopy (SEM, JSM-7800F Prime, 8 kV). To further validate the structural and chemical properties of LIG, X-ray diffraction (XRD), Raman spectroscopy, Fourier transform infrared (FTIR) spectroscopy, X-ray photoelectron spectroscopy (XPS), transmission electron microscopy (TEM), and high-resolution TEM (HRTEM) were employed. For consistent and reliable measurements, SEM, EDX, and Raman analyses were performed on 2 × 2 mm LIG samples fabricated under varied conditions, while XPS and XRD analyses were conducted on fine LIG powder obtained by scraping the films. These systematic procedures and advanced characterization techniques ensured a comprehensive evaluation of the structural, thermal, electrical, and mechanical properties of the fabricated LIG.

3. Results and Discussion

3.1. Effect of Laser Scan Speed (%) and Laser Power (%)

In this work, laser scan speed, ranging from 100 to 180 mm/s with laser power varying between 5% to 50% was utilized. The experimental setup depicted in (Figures a and S1a) comprises an air-assisted 100W CO₂ laser having a 10.6 μm wavelength. However, the substrate (PI) was fixed to the base for better resolution, as depicted in (Figure S1b). PI exhibits high tensile strength (∼80 MPa), excellent thermal stability (up to 500 °C), and also offers strong chemical resistance to hydrocarbons, oils, and organic solvents. , As shown in (Figure b), the chemical structure of PI essentially comprises of two major groups, the Imide group and the aromatic moiety group. The Imide group contains two carbonyl (CO) groups bonded to a nitrogen (N) atom, while the aromatic moiety typically features a benzene ring connected to the Imide group. When exposed to CO₂ laser irradiation, PI efficiently absorbs infrared energy due to its strong vibrational absorption, which leads to photothermal pyrolysis rather than direct photochemical excitation. The localized heating causes bond scission of C–N, C–O, CO, and C–H groups, releasing volatile species such as H₂, CO, CO₂, and N₂. The remaining carbon atoms recombine and rearrange into sp²-bonded domains, resulting in a porous graphene-like network on the PI surface (Figures c,d and S1c). , This 3D porous structure provides a large surface area and excellent electrical properties, making it attractive for applications in energy storage, sensing, catalysis, and flexible electronics. , To determine the optimal parameters for achieving high receptivity and stable LIG formation, various combinations of laser power (%) and laser scan speed (mm/s) were tested. This experimentation process, illustrated in (Figures e and S1d), began with a laser power of 5% and gradually increased up to 50%, while the laser scan speed ranged from 100 to 180 mm/s. Darker patterns observed corresponded to stable LIG formation, while fainter patterns indicated insufficient Carbonization. Notably, the absence of LIG, depicted in (Figure e) for power beyond 40%, resulted in complete burnout of the substrate itself. Through this investigation, a threshold was identified suggesting that higher laser scan speeds (mm/s) and power (>15%) favor pattern quality. However, excessive power led to the incineration of the polyimide film, affecting the longevity of LIG. Similarly, as illustrated in (Figure f), the schematic presents a clear understanding of LIG pattern variation with varying laser power or scan speed. Low magnification SEM images, as shown in (Figure g,i-iv), supported the assumption that carbonization commenced only after exceeding 15% laser power, with (Figure g,(iv)) clearly illustrating the missing LIG due to overpower. An optical profilometry study was carried out to assess the LIG thickness and its dependence on laser scan speed at a specified laser power (Figure S1e). It was found that slower scan speeds led to longer exposure times, resulting in thicker LIG or complete burnout of the underlying PI. In the burnt regime, LIG exhibited a collapsed and discontinuous carbon framework, which disrupted conductive pathways and significantly increased resistance. Conversely, optimal LIG patterns, achieved by balancing laser power and scan speed, prevented burnout and formed conductive graphene (1k to 16 Ω), making it suitable for various sensors, electrodes, or active layers. To establish a relationship between laser parameters and LIG properties, unstrained resistance was measured from the slope of the I–V curve (voltage range: ±5 V) (Figure S1f). It shows that increasing power at a constant laser speed (100 mm/s) significantly reduced electrical resistance. However, surpassing a certain power threshold (30%) caused substrate melting, which increased the resistance to the MΩ range. These findings indicate that optimal laser power (approximately 20%) and higher scan speeds (>100 mm/s) result in lower resistance, whereas excessive laser power sharply increases resistance. After repeated measurements, it was observed that reaching a specific threshold of laser power and scan speeds resulted in the complete carbonization of PI. Consequently, after substrate-specific optimization (Figure h­(i-iv)), LIG could be shaped into various flexible sensor designs without the risk of burnout.

1.

Fabrication of laser-induced graphene (LIG). (a) CO₂ laser setup. (b,c) Photothermal pyrolysis of PI and release of gaseous byproducts forming porous LIG. (d) SEM image of porous LIG. (e,f) Effect of scan speed and power on LIG type and degree of carbonization. (g) Optical/SEM images of LIG at different powers (5–40%). (h) Examples of patterned LIG designs for various applications.

3.2. Effect of Dot Per Inch (DPI) on the Quality of LIG

To investigate the impact of Dots Per Inch (DPI) on the quality of LIG, the study focused on maintaining the optimal laser power (%) and laser scan speed (mm/s) as determined in the previous section while varying the DPI settings, as illustrated in Figure a. For this study LIG samples, each measuring dimension 18 × 14 mm, were fabricated with varying DPI values. It was observed for lower DPI values (such as 100) resulted in LIG patterns that appeared faint, accompanied by a higher electrical resistance measured in kilo ohms (kΩ). Conversely, increasing the DPI to 250 led to a notable improvement in resistance, with values around ∼269 Ω. However, beyond a DPI of 350, a significant increase in resistance was observed, coupled with the undesirable consequence of substrate burning. Deeper insights into the underlying mechanisms governing these phenomena were gained through an examination of the overlapping region of the two laser spots, as depicted in Figure b,c. This analysis highlighted the critical importance of optimizing the overlapping distance (g) between the adjacent laser spots to ensure the continuity of LIG formation and maintain controlled electrical resistance. These findings underscore the intricate interplay between the role of DPI and the resulting quality of LIG, shedding light on crucial factors that must be considered to enhance the fabrication process and improve the performance of LIG-based devices. To investigate further, a relation for the overlapping distance between two laser spots is given by:

$$\mathrm{O}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{l}\mathrm{a}\mathrm{p}\mathrm{D}\mathrm{i}\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{c}\mathrm{e}(g)=\left|\frac{25400}{\mathrm{D}\mathrm{P}\mathrm{I}}-s\right|$$ 1

where g represents the overlap distance between two laser spots and s is the laser spot diameter. The laser machine utilized in this work has a focal length of 50.8 mm with a spot size to overlap approximately 120 μm. Higher DPI values of >300 increase g, causing laser spot overlapping. This overlapping usually leads to burning or tears of the substrate and thus results in increased resistance, as depicted in (Figure d). Conversely, lower DPI values of <200 decreased g, leading to incomplete LIG network formation with a noncarbonized PI in between the spots. The overlapping area of the laser spots can be calculated as:

$$\mathrm{O}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{l}\mathrm{a}\mathrm{p}\mathrm{A}\mathrm{r}\mathrm{e}\mathrm{a}=4[\pi r^2{\sin }^{-1}\left(\frac{h}{2\pi r}\right)-\frac{1}{2}\mathrm{h}\mathrm{D}]$$ 2

2.

Schematic illustrating the role of DPI on the LIG stability and resistance. (a) Effect of DPI on the stability of LIG and resulting resistance (b) relation of overlapping distance with DPI (c) relation of DPI on the resistance of LIG. (d) Effective area of two laser spots with minimum desistance g maintained to keep resistance in check, and the optical micrographs showing LIG formation for various DPI values. (e) Raman analysis for different DPI values where the least defect is obtained for the sample created with DPI of 250.

If this overlap area exceeds the spot size area, disconnected regions are created which leads to high resistance. This assumption correlated with Raman spectroscopy results, which revealed that the I _D/I _G (intensity of D-band to G-band) ratio for a DPI value of 150 was around ∼0.98, for a DPI value of 250 was around ∼0.63 while the same for a DPI value of 350 was ∼0.98, as shown in (Figure e). A higher I _D/I _G ratio indicates lower graphitization, , confirming that a DPI around 250 is optimal for stable LIG at optimized laser Scan speed (mm/s) and laser power (%). Thus, for efficient and stable LIG growth, the DPI was fixed at 250 to ensure lower resistance and higher stability. Table S1 provides better insight with all possible combinations of varying DPI with LIG resistance, obtained based on optimized laser power (%) and laser scan speed (mm/s). The parameter optimization in this work followed a physics-guided sequential strategy rather than a full factorial design-of-experiments approach. Laser power and scan speed were first examined because they define the primary thermal dose and determine whether PI undergoes insufficient carbonization, stable LIG formation, or burnout. DPI was subsequently optimized to control laser spot/line overlap and electrical percolation of the LIG network. EPUA was then used as a unifying parameter to combine the effects of power, scan speed, and DPI, while depth of focus and substrate support were optimized afterward to account for local power density, spot size, and heat dissipation. This approach enabled exclusion of nonuseful destructive regimes and allowed the optimized condition to be validated through electrical, morphological, chemical, and device-level measurements. Although a full factorial design could further quantify higher-order parameter interactions, the present framework provides a practical and physically grounded route for reproducible LIG fabrication.

3.3. Effect of Energy per Unit Area (EPUA)

The energy per unit area (EPUA) is defined as the amount of energy delivered to a specific area during a laser-induced process, for formation of LIG. It is a critical parameter that influences the outcome of the process. It can also be used to define the threshold energy that is needed to attain for direct conversion of stable LIG on any LIG precursor. Importantly, EPUA integrates the combined effect of laser power, scan speed, and dots per inch (DPI), and can therefore serve as a unifying metric for reproducible LIG fabrication across different laser systems.

$$\mathrm{E}\mathrm{P}\mathrm{U}\mathrm{A}={P_{\mathrm{l}\mathrm{a}\mathrm{s}\mathrm{e}\mathrm{r}}}^{*}P{(\%)}^{*}\frac{\mathrm{D}\mathrm{P}\mathrm{I}}{\mathrm{l}\mathrm{a}\mathrm{s}\mathrm{e}\mathrm{r}\mathrm{s}\mathrm{c}\mathrm{a}\mathrm{n}\mathrm{s}\mathrm{p}\mathrm{e}\mathrm{e}\mathrm{d}\left(\frac{\mathrm{m}\mathrm{m}}{\mathrm{s}}\right)}$$ 3

where P _laser is the rated machine power (100 W in this study), P(%) is the percentage of laser power applied (0–50%), DPI is the spatial resolution. Laser scan speed (mm/s) is the speed at which the laser tip is scanned over the substrate. From our earlier analysis, laser speed of 120 mm/s and DPI of 250 were fixed while laser power (%) was varied. From (Figure a), it can be observed that only beyond the EPUA of 18 J/mm², laser energy was able to start the carbonization. This difference in EPUA leads to the formation of different zones for LIG as depicted in Figure b. For EPUA greater than 85 J/mm², burnout condition was reached, while the stable LIG was obtained for EPUA ranging between 35 J/mm² and 60 J/mm² and were verified using Raman analysis. The samples for all the succeeding analysis were prepared based on the procedure highlighted in Figure S1g. The analysis suggested that a lower I _D/I _G ratio for this range indicated lower defect or stable LIG formation. Raman spectra of various LIG patterns (Figure S2a,b) at varied EPUA values exhibit the characteristic carbon peaks including D band generated by the sp² with Raman frequency of ∼1350 cm^–1, G band produced by intraplane vibration at ∼1580 cm^–1 and 2D band at ∼2700 cm^–1 indicating no of graphene layers. , The ratio of I _D/I _G indicates the amount of defect in case of LIG or degree of the disorder. Having positive or higher ratio indicates higher defects. A higher EPUA or an EPUA lower than the cut off resulted in a higher I _D/I _G ratio with higher resistance. However, for low resistive LIG, EPUA values ranging between 35 J/mm² to 60 J/mm² would be beneficial for proposed applications. These findings correlate to our previous assumptions for LIG fabricated with a DPI of 250, shown in (Figure a). Further, LIG samples were investigated using SEM as depicted in (Figure c) where the morphology of the fabricated LIG could be controlled precisely. (Figure c,i-iv) shows different variants of LIG whose morphology was influenced by the change in the EPUA value. (Figure c,ii) shows that when the laser EPUA exceeded the threshold value, the high temperatures during the LIG process caused the rapid release of various components, leading to the formation of micropores. These micropores turn to fibrous form as depicted in (Figure c,iii) when EPUA was increased further to greater than 60 J/mm² but below 80 J/mm². The presence of these microporous nature increases the surface area of LIG, which can act as the localized site for other elements or provide the pathways for electrolytic ions that can open up the potential use case in supercapacitor electrodes. Upon increasing the EPUA >80 J/mm², the fibrous LIG translate to more complex structure as depicted in (Figure c­(iv)). However, due to the very high EPUA, the energy received by the substrate led to burnout, resulting in unstable and poor bonding of LIG, which could easily peel off and thus lose its LIG pattern. Such findings match well with the EDX results as depicted in (Figure d). Carbon weightage (%) increases with rise in power while the content of nitrogen and oxygen decreases. However, for the burnout condition, the oxygen weightage (%) increases due to pertinent oxidation that leads to reduced carbon content (%) for LIG. This reduction in carbon (%) is the major reason for inconsistent and unstable LIG formation. For pristine PI, the EDX analysis is depicted in (Figure S2c) wherein the carbon weightage equals 72.1% while with just 5% increase in laser power, the carbon weightage (%) rose by ∼3% without any morphological alteration. As the laser power is increased beyond the stable range, the LIG degradation is confirmed by observing the decrease in carbon weightage (%) while increase in the other elements. Figure S2d shows the intensity ratios I _D/I _G and I _2D/I _G of LIG prepared at different EPUA levels, indicating that defect density and carbon atom stacking in graphene vary accordingly. A high I _D/I _G ratio reflects higher defect density. With optimized EPUA, the I _D/I _G ratio initially decreases and then rises, consistent with previous findings. The I _2D/I _G ratio indicates the number of graphitic layers, with a larger I _2D/I _G ratio signifying fewer graphene layers. After surpassing the threshold EPUA value, a higher I _2D/I _G ratio was observed, indicating a porous and more stable LIG in the 35 J/mm² to 45 J/mm² range. However, when the EPUA exceeds 45 J/mm², the porous LIG becomes fibrous. This is due to the high EPUA causing bond recombination, making the structure tight and restricting movement, resulting in fragile and unstable LIG. Therefore, an optimal EPUA range not only enhances carbonization but also helps to create few-layer LIG. To reveal the effect of optimized LIG parameters on the grain size (L _a ) along the a-axis, different LIG samples were analyzed using Raman spectroscopy. The optimized LIG sample fabricated with EPUA in the range of 35 J/mm² to 60 J/mm² shows a sharper G band, which indicates a higher graphitization compared to other fabricated samples with fewer graphitic layers. For the LIG sample fabricated with laser power beyond the optimization parameters 2D band and G band peaks got wider while the intensity of L_a decreased (Figure e). The reason for this behavior can be attributed to excessive absorption of energy by the PI and deterioration of the LIG structure. The value of L _a from the ratio of D peak to G peak is calculated as.

$$L_a=2.4\times {10}^{-10}\times {\lambda }_l^4\times {\left(\frac{I_{\mathrm{D}}}{I_{\mathrm{G}}}\right)}^{-1}$$ 4

3.

Effect of EPUA on LIG quality: (a) resistance and I _D/I _G variation at optimized DPI and speed with varying laser power. (b) Effect of EPUA on nature of LIG grown. (c) Relation between laser power, EPUA, and LIG morphology with SEM images: (i) PI just above threshold (>20 J/mm²), (ii) porous LIG (>50 J/mm²), (iii) fibrous/grassy LIG (60–80 J/mm²), (iv) PI burnout (>80 J/mm²). (d) EDX analysis at different laser powers, showing the highest C % and lowest N/O % at 20% power. (e) I_D/I_G ratio and crystallite size (L _a ).

Here, λ _l (∼532 nm) represents the wavelength of the laser used in the Raman analysis of the LIG sample. As shown in (Figure e), a higher L _a (∼50 nm) corresponds to porous LIG, while a smaller L _a (∼15 nm) indicates an LIG sample processed with excessive laser power (%). EPUA plays a crucial role in achieving optimal laser processing conditions, independent of the specific laser machine used. Regardless of whether researchers are working with a lower-power or higher-power laser system, laser scan speed or the dots per inch (DPI) can be adjusted to maintain the optimal EPUA range. By controlling these parameters, the consistency and quality of laser-induced features can be ensured, highlighting the significance of EPUA in achieving reproducible results across different laser systems.

3.4. Effect of Depth of Focus (DOF) on LIG Spot Size and Quality

To comprehensively understand the impact of the depth of focus (DOF) or off-focus value (h) on the quality of LIG, a custom scale was designed to precisely manipulate h, as illustrated in Figure a. By systematically adjusting (h) to its optimal value while maintaining optimized laser parameters, we aimed to identify the ideal setting for achieving highly stable LIG with a minimal laser spot size (diameter). This investigation facilitated the fabrication of smaller, more precise LIG devices characterized by enhanced stability and quality. In (Figure b), the schematic depicts the working area of the laser beam, where L represents the distance between the Polyimide (PI) substrate and the focus lens, F_L denotes the focal length of the focus lens (50.8 mm in our case), D _B signifies the diameter of the unfocused beam, L _D indicates the lens diameter (20 mm in our case), and h = (L–F_L) represents the off-focus value. The off-focus value h can be positive or negative, depending on its position relative to the focal plane of the focusing lens. At the focal length F_L, where the distance between the laser tip and the PI substrate is h = 6 mm, L measures 56.8 mm. The diameter (D) of the laser spot on the PI surface is calculated as

$$D=D_{\mathrm{o}}\sqrt{1+{\left(\frac{h}{F_L}\right)}^2}$$ 5

4.

Effect of depth of focus (DOF) on LIG quality: (a) Custom scale for adjusting h (DOF) between laser tip and PI. (b) z-axis adjustment for spot size optimization. (c) Microscopic image of optimized spot at h = 6 mm (±1 mm defocus). (d) Raman spectra for varying h under optimized laser settings. (e) Top/bottom views of LIG with and without metallic support. (f) Resistance variation with h, showing the lowest resistance at h = 6 mm.

This equation considers the Gaussian beam characteristics of the laser beams and accurately predicts the diameter of the laser spot, which increases with h. (Figure c). confirms this assumption, further validated by optical investigation of LIG spots with SEM shown in Figure S3a–f. The variation of LIG spot with (h) also significantly affects the quality of LIG formed. Upon decreasing the value of (h), it was observed that the diameter of the laser spot would decrease, but the effect of EPUA became prominent. With a very small value of DOF (h < 3), higher EPUA is achieved, leading to complete melting of the PI substrate. From the results shown, it became evident that the most stable LIG was obtained at h = 6 mm, while unstable LIG were obtained for h < 6 mm or h > 6 mm. The Raman analysis depicted in (Figure d) confirmed the highest quality LIG formation at h = 6 mm, showing the lowest I _D/I _G ratio ∼0.44. For h < 4 mm, high defects were observed due to substrate burning, while for h > 7 mm, incomplete carbonization led to poor LIG quality and higher defect levels I _D/I _G ∼ 1.13. Thus, h = 6 mm was chosen for further optimization of LIG parameter with ±1 mm of defocus correction factor. SEM results in (Figure e) corroborate similar assumptions, demonstrating that at h = 6 mm, LIG with a laser spot diameter of ∼110 μm and thickness of ∼65 μm was obtained, with better I _D/I _G compared to other h values. To investigate the effect of (h) on the resistance of LIG, square samples measuring 3 × 3 mm were fabricated at varying (h) while maintaining other laser parameters consistent with the optimized settings. As depicted in (Figure f), high resistance was observed for h < 4 mm and h > 8 mm, like samples fabricated with high and low EPUA. Samples fabricated with h = 6 mm exhibited properties similar to porous LIG, further emphasizing the critical role of (h) in determining LIG stability and quality. Lower values of (h) resulted in higher power incidents, leading to unstable LIG or substrate burning. The correlation of varying (h) with optimized laser power (%), laser scan speed (%) and DPI has been shown in Table S2. The overall effect of varying (h) has been depicted in the high-resolution SEM image in Figure S3g,h with the side view for clear visualization, wherein the interconnected LIG network depicted in Figure S3i with a thickness of 0.86 μm is obtained at optimized h = 6 mm. This interconnected network allows the resistance of the optimized LIG to change with applied force and return to its original state when the force is removed, demonstrating high flexibility and mechanical robustness compared to LIG grown at other (h) values.

3.5. Effect of Nature of Support/Base on the Quality of LIG Quality and Stability

To investigate the effect of the nature of the base/support on the quality and stability of LIG. fabrications were performed using either aluminum or any other metal support, acrylic support, or no support, as depicted in (Figure a). To validate the obtained results, optimized LIG parameters were used separately to grow LIG on a metallic base. During LIG formation of PI, various gases such as water vapor, CO₂, and O₂, along with considerable heat, are released. Metals, due to their high heat dissipation ability, provide a superior pathway for heat transfer compared to acrylic or plastics. Using aluminum (Al) as the support for LIG on PI yielded higher quality, as shown in (Figure b), with an I _D/I _G ratio as low as ∼0.55. Similarly with optimized parameters, the I _D/I _G ratio improved further to ∼ 0.27, indicating a highly graphitized structure and stability. In contrast, acrylic support or no support led to an I _D/I _G ratio of ∼0.8, indicating lower LIG quality and reduced stability. For the sample without the Al base, as illustrated in Figure a, the top and bottom views confirmed that the LIG did not remain intact, contrary to observations for the sample with Al support. To confirm this, an array of LIG samples depicted (Figure c) with a width W = 5 mm and length L = 20 mm was fabricated with varying laser power (%). It is evident that sheet resistance (R _sh) decreases with increased power for all three cases, but samples with Al base and optimized settings show minimum R _sh (∼30 Ω/□ and ∼10 Ω/□) and higher stability, respectively. This trend corresponds with the decreasing I _D/I _G ratio, confirming the formation of higher quality of graphitic structures with fewer defects. The improved performance is attributed to the efficient heat dissipation provided by the metallic support, which prevents localized overheating and enhances LIG uniformity. Moreover, samples fabricated on Al exhibited lower initial resistance, indicating a more conductive LIG network formed via better thermal management during laser exposure. The dependence of resistance of LIG on laser power (%) was investigated experimentally using a digital source meter (Keithley 2450), while the R _sh is calculated as

$$R_{\mathrm{s}\mathrm{h}}=R_{\mathrm{m}}\times \frac{W}{L}$$ 6

where R _m represents the measured resistance, W represents the width, and L represents the length. According to the EPUA study of the formation of actual LIG, the laser processing energy density needs to surpass LIG thresholds. Below this threshold, the LIG sample shows very high resistance (MΩ). Beyond the optimal point (>20%), porous LIG turns into a fibrous nature with conductive paths, but it loses contact with PI, becoming unstable with high resistance compared to its counterpart porous LIG. So, for stability and porous LIG, a region of laser power ranging from 15% to 22% can be explored, as depicted in (Figure d). All three types of LIG samples performed well with an almost linear relation of R _sh with laser power within this range. Another combination of varying laser power (%) while keeping laser scan speed (120 mm/s), DPI (250), and DOF (h = 6 mm) constant for all three samples is given in Table S3. To evaluate the effect of optimized LIG parameters, all three samples were tested for bending performance, as shown in Figure e measured with the response expressed as ΔR/R %.

$$\frac{\mathrm{\Delta }R}{R}\%=\frac{R_{\mathrm{m}}-R_{\mathrm{b}}}{R_{\mathrm{b}}}$$ 7

where ΔR = R _m – R _b, change in resistance, R _m is the initial resistance without bending, and R _b represents the resistance with respect to bending. The fractional change during bending (b) is given by:

$$b(\mathrm{i}\mathrm{n}\%)=\frac{L_{\mathrm{l}}-k}{L_{\mathrm{l}}}$$ 8

where L _l is the length of the LIG surface and k is the clamped distance between the two ends of PI strip bent into an arc. It can be observed from that the LIG is on the inner side of the bending arc thus it experiences compressive strain when the sensor is bent. This leads to the formation of microcracks, causing microscopic deformations that increase the overall resistance. When b = 0%, the sensor remains straight, meaning, L _l = k while at b = 100%, the sensor is fully folded with k = 0. These observations are depicted in (Figure e); performance of LIG samples under applied bending (%) shows that the optimized LIG exhibits a higher response compared to the other two. The electrical conductivity of LIG relies on the quantity of contact points or electrically conductive pathways between neighboring carbon units. Increased laser power can generate denser and more interconnected carbon units, resembling fibrous structures, thereby reducing the sensitivity of LIG to bending strain. Also, from previous analysis, the effect of laser power on LIG quality in Figure c­(i-iv) clearly identified a strong correlation between laser power and the nature of LIG growth, which was higher compared to optimized LIG devices. Optimization significantly enhanced the consistency of LIG samples, as evidenced by the findings in (Figure f). An array of sensors, each with dimensions L = 5 mm and W = 2 mm, was fabricated as depicted in Figure f­(ii) under optimized parameters, resulting in sensor resistances ranging between 88 Ω and 95 Ω, demonstrating remarkable uniformity. In contrast, Figure f­(iii) showcases the growth of large-area LIG with dimensions L = 100 mm and W = 60 mm. Additionally, Figure f,iv illustrates the diverse benefits of LIG optimization, enabling tailored designs for various applications. This versatility underscores the significance of optimization in maximizing the potential of LIG for multifaceted uses. Power tends to produce more grassy/fibrous LIG with a weak binding force with the grown substrate, leading to high resistance and poor performance when used for bending applications.

5.

Effect of support on the quality of LIG. (a) Top and bottom views of LIG samples with and without metallic support. (b) PL Raman of LIG formation on acrylic base, on aluminum (Al) with and without optimized LIG recipe. (c) Effect of laser power (%) on sheet resistance (R _sh) keeping other LIG parameters as optimized values. (d) Almost linear relation of R _sh with laser power (%). The inset in (e) shows the way of bending characterization of the LIG sample fabricated at different laser Power (%). (f) Different LIG designs developed with optimized process parameters.

4. Physical and Mechanical Characterization of Optimized LIG

4.1. XPS and FTIR Characterization

The XPS and FTIR results from Figures and S4 confirm the successful fabrication of highly stable and high-quality LIG on PI. XPS analysis reveals the presence of primary C 1s, N 1s, and O 1s peaks at 284.4 eV, ∼400.1 eV, and 533.1 eV, respectively, indicating sp² hybridized Carbon atoms. − Additionally, the S 2p peak at ∼ 169 eV in the PI sample is attributed to the adhesive layer (Figure a). The experiments were conducted with a constant laser scan speed of 120 mm/s, DPI of 250, and DOF of 6 mm while varying the laser power from 0% to 30%. For the PI without laser treatment, the carbon (C) content was 72.1%, nitrogen (N) 7.7%, and oxygen (O) 20.1%, as shown in Figure S4a. As laser power increased, the carbon percentage rose, while the nitrogen and oxygen percentage decreased. At 20% laser power, the carbon content reached a maximum of 96.4%, and the nitrogen content decreased to ∼1.3%, compared to its initial value before laser processing of the PI. This high Carbon content suggests the formation of highly graphitic and conductive LIG, which is consistent with the previous literature. ,− Beyond this optimized laser power (20%), the Carbon percentage decreased, while the Oxygen percentage increased significantly, likely due to the burning or melting of PI, leading to oxidation rather than carbonization. The higher Carbon content at a laser power of 20% indicates the optimal conditions for LIG fabrication. At laser power (%) above (>30%), fibrous LIG formation was observed owing to excessive heat trapping in a smaller area, resulting in poor adhesion of LIG on PI and increased resistance. The deconvoluted XPS peaks for the pristine PI and optimized LIG are shown in Figure b,c. Both exhibit four major peaks: C–C at ∼ 284.6 eV, C–N at ∼285.4 eV, CO at 287.3 eV, and O–CO at 288.7 eV. , Figure d,e demonstrate that postlaser pyrolysis, the C–C intensity for optimized LIG increased compared to pristine PI, which is attributed to the photothermal reaction breaking C–O, CO, and N–C bonds, followed by the rearrangement of Carbon atoms into the graphene structure. Additionally, the increase in Carbon content can be primarily attributed to the removal of imide (OC–N–CO) bonds present in the PI. , High-resolution XPS of O is 1s for optimized LIG, shown in Figure S4b, exhibits two major peaks: CO (531.8 eV and C–O (533.1 eV). High-resolution XPS of N 1s for optimized LIG, depicted in Figure S4c, shows two major peaks: N–C–N at 398 eV and C–N at 400.1 eV, attributed to the decomposition of imide structures. For FTIR analysis, powdered LIG samples were obtained by scraping LIG from the PI. Figures f and S4d. present the FTIR spectra of the PI before and after laser pyrolysis. Initially, PI exhibited clear peaks ranging from 1090 to 1776 cm^–1, representing the characteristic stretching and bending vibrations of the C–O, C–N, and CC bonds. Following laser scribing, a broad absorption spanning from 1000 cm^–1 to 1700 cm^–1 indicates significant alterations in the chemical structure induced by the laser process. , Figure S4d displays the FTIR spectrum obtained from the finely powdered optimized LIG sample. It shows a vibrational peak at 3452.0 cm^–1, indicating the presence of hydroxyl groups (O–H) with strong intensity and a broad peak. The region between 3000 and 2800 cm^–1 corresponds to the stretching vibration of the methyl group (C–H bond). The peak at 1456.0 cm^–1 corresponds to the methylene group (CH₂). Additionally, the vibrations at 1629.6 cm^–1, 1029.8 cm^–1, and 599.8 cm^–1 indicate the presence of aromatic (CC) groups with strong intensity, an alkoxy group (C–O), and some halo compounds, respectively. These vibrational peaks confirm the presence of high-graphite Carbon materials such as graphene and graphene oxide (GO), indicating interactions with Oxygen and some impurities.

6.

Chemical analysis of PI and optimized LIG: (a) XPS survey of PI surface. High-resolution C 1s spectra of (b) pristine PI and (c) optimized LIG. Carbon intensity comparison in (d) pristine PI and (e) optimized LIG, confirming graphene formation. (f) FTIR spectra of PI before and after laser pyrolysis, showing bond modifications.

4.2. Atomic-Scale Morphology Study of Optimized LIG

Figures a and S5a present a low-resolution TEM image of an optimized LIG sample prepared with a 20% laser power, resembling a thin silk towel rolled. Figures b and S5b,c depict a selected area showing a few layers of the LIG sheet. Increasing laser power leads to thicker or multiple LIG layers with a higher number of defects, as shown in Figure S5d. Raman analysis (Figure c) confirms that the optimized LIG exhibits low defect density, with an I _D/I _G ratio of ∼0.25 and a low I _2D/I _G ratio of ∼0.74, indicating few-layer graphene, consistent with the TEM observations.

7.

Atomic-level structural investigation of optimized LIG using HRTEM, TEM, Raman, and XRD: (a) Low-magnification TEM (400 nm scale). (b) Enlarged TEM view showing few-layer regions (scale bar = 100 nm). (c) Raman spectra of PI and LIG with low I _D/I _G (∼0.25) and high I _2D/I _G (∼0.74), confirming few-layer structure. (d) XRD peak at 2θ ≈ 26.17° (d ≈ 3.4 Å). (e) HRTEM image (scale bar = 5 nm). (f) FFT of HRTEM showing lattice spacing ∼ 0.348 nm, consistent with XRD.

XRD results (Figure d) show a sharp peak at the (002) plane, suggesting a highly crystalline graphene structure, corroborating the Raman findings. , The interlayer spacing of LIG was calculated using Bragg’s law.

$$d_{\mathrm{o}}=\frac{n\lambda }{2\sin (\theta )}$$ 9

where d _o is the interplanar spacing (distance between adjacent atomic planes), θ is the X-ray incidence angle, n is the diffraction order, and λ is the X-ray wavelength. From the observed 2θ value of 26.13°, the interlayer spacing was determined as d ₀₀₂ ≈ 0.341 nm. These results agree with Raman analysis (Figure c), where the low I _D/I _G ratio and strong I _2D/I _G peak indicate well-ordered graphene domains and high structural quality of the LIG. Further confirmation of the interlayer spacing was obtained from HRTEM and FFT images (Figures e,f and S5e). FFT patterns were obtained from HRTEM images (acquired at 200 kV) and analyzed in ImageJ software; lattice spacings reported above were measured from calibrated FFT spots. Using ImageJ in FFT mode, the LIG was found to be few-layered with an interatomic spacing d ≈ 3.48 Å (0.348 nm), consistent with XRD results and confirming the formation of graphitic carbon under the optimized laser parameters. Detailed discussions of TEM morphology, HRTEM lattice analysis, and SAED patterns are provided in Figure S5f. These findings highlight the importance of optimizing laser parameters to achieve crystalline LIG structures. Table summarizes key material quality metrics, including I_D/I_G, I_2D/I_G, sheet resistance (R_sh), and crystallite size (L_a), reported in recent LIG studies. For consistency, comparison is primarily made for PI-based LIG, while a few entries involving different laser systems or modified substrates are included for completeness. As shown, the reported sheet resistance spans a wide range (∼6 to >500 Ω/□), even within similar CO₂–polyimide systems, reflecting the strong dependence on processing conditions. Although several studies report low sheet resistance values, structural parameters such as I_D/I_G, I_2D/I_G, and La are not consistently reported across the literature, making direct comparison challenging.

1. Comparison of Structural and Electrical Properties of LIG from This Work and Prior Reports .

laser type	substrate	I D/I G	I 2D/I G	R sh (Ω/□)	L a (nm)*	refs
CO2 (10.6 μm)	PI	0.33	1.05	-	∼ 60
CO2 (10.6 μm)	PI	1	0.64	∼35	∼19
CO2 (10.6 μm)	PI	0.26	0.45	∼12	∼73
CO2 (10.6 μm)	PI	0.94	0.66	∼6.5	∼20
CO2 (10.6 μm)	PI	1.40	0.66	13–572	∼13
CO2 (10.6 μm)	PI	0.31	0.50	∼15	∼62
CO2 (10.6 μm)	PI	1.2	0.50	20–100	∼16
CO2 (10.6 μm)	PI	0.83	0.78	20–60	∼23
CO2 (10.6 μm)	PI + K5V12 deposition	0.72	0.49	-	∼26
NIR (1064 nm)	PI	0.5	0.4	∼233	∼38
Femto (515 nm)	PI	1.08	-	∼32	∼17
Femto(1035 nm)	PI	1.04	0.72	∼20	∼18
CO2 (10.6 μm)	PI	0.25	0.74	∼10	∼76	This Work

^aPI = Polyimide, K₅V₁₂ = vanadium oxidenanocluster, NIR = Near-Infrared laser.

In this context, the LIG produced in this work exhibits a low I _D/I _G ratio (∼0.25), indicating reduced defect density. This is accompanied by a relatively high I _2D/I _G ratio (∼0.74) and a large crystallite size (L _a ∼ 76 nm), consistent with the formation of few-layer graphene with improved structural order. Moreover, the sheet resistance (R _sh ∼ 10 Ω/□) is within the lower range of reported values for LIG synthesized via CO₂ lasers, highlighting the effectiveness of our parametric tuning approach. Collectively, these results demonstrate our method’s capability to fabricate high-quality LIG suitable for flexible, scalable electronic applications.

5. Application of Optimized LIG

5.1. Flexible Piezoresistive Sensor

Following confirmation of the optimized laser parameters through comprehensive experimental testing, the next phase involved applying these findings to real-time applications. To achieve this, a flexible piezoresistive sensor (FPS) was meticulously crafted utilizing the optimized LIG parameters on a PI substrate. The fabricated sensor is illustrated in Figure . The silver contacts are fabricated on the two ends of the LIG sensor for electrical measurements, as depicted in Figure a­(i). The fabricated sensor is attached to the measurement setup and strained/curved by application of vertical forcing as shown in Figure a­(ii-iv). A schematic diagram of the FPS measurement configuration is presented in Figure b, where a controlled tensile force is applied to the sensor. The piezoresistive response of the device originates from deformation-induced changes in the conductive network of porous LIG. The LIG forms a three-dimensional interconnected structure composed of graphene flakes that establish percolative conduction pathways, as confirmed by the SEM images in Figures d and c. Under applied force, compression of the porous network increases interflake contact area and reduces tunneling distances between adjacent conductive domains, resulting in enhanced conductive pathways and a corresponding decrease in resistance. At higher forces, partial microcrack evolution and structural rearrangement further modify the conduction network, leading to nonlinear sensitivity. Upon unloading, the viscoelastic recovery of both the PI substrate and the porous LIG structure restores the original conductive configuration, producing reversible resistance changes. A schematic illustration of this mechanism is provided in Figure c, which highlights the evolution of conductive pathways under compressive deformation. A parameter analyzer was employed to measure the change in resistance in response to the applied force, while data acquisition was facilitated by a laptop. To apply a continuous cyclic force, a Mark-10 Series 3 Force gauge was utilized, with a force up to 7N applied. Figure d illustrates the time series response to periodic forcing of ∼7N and the resulting resistance change. The frequency of forcing was fixed at 0.1 Hz for the best response. The sensor exhibits clear and reproducible resistance modulation under applied load. The output characteristics for varying force amplitudes are presented in Figure e. The sensor demonstrates two distinct operational regimes with different sensitivities, a behavior commonly observed in strain-based piezoresistive devices. In the low-force region (0–4 N), the sensitivity is approximately 1.25% N^1–, while in the higher-force region, the sensitivity increases to approximately 32% N^1– before saturation beyond 6.9 N. The enhanced response at larger forces is attributed to increased deformation of the porous LIG network, leading to greater modification of conductive pathways.

8.

Optimized laser parameters and reliability of flexible piezoresistive sensor (FPS): (a) Schematic of measurement setup with data acquisition interface. (b) Actual FPS measurement system. (c) Schematic illustration of the piezoresistive mechanism in porous LIG, highlighting deformation-induced evolution of conductive pathways through increased interflake contact and structural rearrangement. (d) Periodic response under ∼ 7 N force. (e) Force sweep (0–7 N) with two sensitivity regions. (f) Mechanical stability under 7 N bending for 4000 s; insets highlight ΔR/R % at 0–50 s and 3950–4000 s, confirming no performance degradation.

To provide a comprehensive assessment of sensor performance, key parameters, including linearity, hysteresis, and limit of detection (LOD), were evaluated in addition to sensitivity. Linear fitting of the ΔR/R versus applied force curve reveals two distinct sensing regimes with coefficients of determination of R ² ≈ 0.89 (0–4 N) and R ² ≈ 0.98 (4–6.9 N), respectively. The limit of detection (LOD) is estimated as:

$$\mathrm{L}\mathrm{O}\mathrm{D}=\frac{3\sigma }{S}$$ 10

where σ represents the standard deviation of baseline noise and S is the sensitivity in the low-force regime, yielding an LOD of approximately 150 mN, which is primarily limited by the force-gauge resolution and the viscoelastic response of the polyimide substrate. Hysteresis was evaluated using cyclic loading–unloading experiments and quantified using the maximum normalized deviation between the resistance responses measured during loading and unloading at identical force levels.

$$\mathrm{H}\mathrm{y}\mathrm{s}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{i}\mathrm{s}(\%)=\frac{\max |{(\mathrm{\Delta }R/R)}_L-{(\mathrm{\Delta }R/R)}_U|}{{(\mathrm{\Delta }R/R)}_{\max }}\times 100$$ 11

where (ΔR/R)_L and (ΔR/R)_U denote the normalized resistance change during loading and unloading, respectively. Using this definition, the hysteresis of the flexible piezoresistive sensor was found to be approximately 5–6% over the full operating range. This hysteresis is primarily attributed to the viscoelastic relaxation of the polyimide substrate and the time-dependent mechanical recovery of the porous LIG network. Figure S6a depicts a single loading pulse applied to the FPS at 7 N, where the response time and recovery time are approximately 4.8 and 4.4 s, respectively. These times are governed by the viscoelastic behavior of the PI substrate and the mechanical relaxation of the porous LIG network rather than intrinsic electronic transport in graphene. The porous LIG layer is mechanically fragile and relies on the compliant polymer substrate for structural recovery after deformation. Consequently, time-dependent viscoelastic relaxation leads to delayed recovery and hysteresis under cyclic loading. In addition, the applied force was delivered using a mechanical force gauge operating at 0.1 Hz (≈10 s per sweep), which further influences the measured response time. These characteristics indicate that the sensor is most suitable for quasi-static pressure sensing and human-motion monitoring, rather than high-frequency dynamic applications. The real-time mechanical reliability and dynamic response of the fabricated FPS were also captured in Movie S1 where the optimized flexible piezoresistive sensor was subjected to cyclic loading using a Mark-10 Series 3 Force Gauge while resistance variations were simultaneously recorded by a Keithley 4200A-SCS Parameter Analyzer. The observations in Movie S1 corroborate the results presented in (Figure d–f), confirming that the FPS exhibits stable and reproducible performance under repeated mechanical stress. Each subfigure illustrated in Figure S7 gives a unique application, demonstrating the wide-ranging potential of optimized LIG parameters in creating highly functional, flexible, and sensitive sensors for various industries. Finally, at the micro level, the FPS, when examined after the application of this continuous force, revealed that there was not much change in morphology. The testing continued further to assess the reliability of the sensor for approximately 4000 s. Due to cyclic bending, the resistance fluctuated between approximately 85 Ω and 160 Ω (Figure f). This change in resistance, denoted as ΔR/R %, exhibited a strong correlation with the applied tensile force. Higher forces resulted in a larger change in ΔR/R %, while lower forces exhibited smaller variations. Comparing the initial resistance before bending to after bending, only a slight change was observed, as shown in the enlarged graph in (Figure f). This excellent mechanical performance indicates that the fabricated FPS is well-suited for flexible applications.

5.2. Flexible Pressure Sensor Based on PDMS-Transferred LIG

To further demonstrate the versatility of the optimized LIG parameters and broaden the application scope of this work, the LIG pattern was transferred onto a soft polydimethylsiloxane (PDMS) substrate to realize a flexible pressure sensor. In contrast to the PI-supported configuration described in Section , the compliant nature of PDMS enables efficient stress transfer under low-pressure stimuli and conformal contact with curved or irregular surfaces, making it highly suitable for wearable and tactile sensing applications. The fabrication process of the PDMS-supported device is illustrated in Figure a­(i–vi). A 125 μm thick nonadhesive PI sheet was mounted on a metallic base to maintain flatness and ensure uniform heat dissipation during laser processing. LIG was directly patterned using a 100 W CO₂ laser under optimized conditions (scan speed: 110–125 mm s^–1, laser power: 20–25%, DPI: 256–300, focal height: 6 mm, frequency: 4 kHz, scan direction along the x-axis). A Rectangle-shaped electrode consisting of two terminal pads (5 × 5 mm) connected by a central pathway (14 × 8 mm) was fabricated. For transfer, a PDMS precursor (base: curing agent = 10:1) was prepared and thoroughly mixed for 15–20 min, followed by degassing in a vacuum desiccator for 30 min to remove trapped air bubbles. The degassed PDMS was poured over the patterned LIG and thermally cured at 70 °C for 2 h. After curing, the PDMS film was carefully peeled off, yielding an LIG/PDMS composite layer as shown in Figure a­(iii). Electrical contacts were formed by applying silver paste at both ends of the device. Subsequently, an additional PDMS layer was cast to encapsulate the structure, resulting in a PDMS/LIG/PDMS sandwich configuration, as shown in Figure a­(v). Figure a­(vi) shows the optical image of the fabricated sensor along with the cross-sectional SEM view, highlighting the porous LIG layer embedded within the elastomeric matrix. SEM observations confirm that the transfer process preserves the three-dimensional porous morphology of LIG, which is essential for maintaining its conductive network. The pressure-sensing mechanism of the LIG/PDMS device is governed by deformation-induced modulation of conductive pathways within the porous graphene network. Upon application of compressive pressure, the compliant PDMS substrate transmits stress to the LIG layer, increasing the interflake contact area and reducing tunneling distances between adjacent conductive domains, thereby enhancing conductive pathways and decreasing resistance. When the pressure is released, the elastic recovery of PDMS restores the original structure, producing reversible resistance changes. The pressure-sensing performance of fabricated pressure sensors was evaluated in terms of sensitivity.

$$S_P=\frac{\mathrm{\Delta }R/R_0}{\mathrm{\Delta }P}$$ 12

where ΔR is the change in resistance relative to the initial resistance R ₀ and ΔP is the applied pressure (kPa). As shown in Figure b­(i), the sensor exhibits three distinct sensitivity regimes over the pressure range of 0–220 kPa, with sensitivities of 0.005 kPa^–1 (Region I: 0–80 kPa, R ₁ ≈ 0.98), 0.00673 kPa^–1 (Region II: 85–145 kPa, R ₂ ≈ 0.99), and 0.00123 kPa^–1 (Region III: 155–220 kPa, R ₃ ≈ 0.98). At low pressures, small deformations significantly alter conductive pathways, resulting in a strong and nearly linear response. Increasing pressure compresses the porous structure, limiting further pathway modification and reducing sensitivity.

9.

Fabrication and performance of the PDMS-transferred LIG pressure sensor: (a) stepwise fabrication process and device structure (PDMS/LIG/PDMS) with optical image and cross-sectional SEM confirming porous LIG embedding; (b) sensing performance including pressure-dependent response, dynamic behavior under varying frequency and pressure, response/recovery characteristics, cyclic stability, and wearable motion detection (finger, wrist, knee, and neck).

At high pressures, pore collapse leads to structural densification, where additional deformation minimally affects electron transport, resulting in low sensitivity and eventual saturation. Figure b­(ii) shows stable periodic resistance signals at a constant pressure of 100 kPa while the loading frequency varied from 1 to 10 Hz, demonstrating reliable dynamic tracking capability. Figure b­(iii) presents the corresponding response and recovery times as a function of frequency, both of which decrease from 0.38/0.42 s at 1 Hz to 0.15/0.22 s at 10 Hz, reflecting faster mechanical deformation and restoration of the PDMS/LIG structure at higher loading rates. The dynamic response was further evaluated at different pressures (50, 90, 100, and 140 kPa) at a fixed frequency of 7 Hz (Figure b­(iv)). Stable and repeatable ΔR/R ₀ signals were obtained under cyclic loading, with signal amplitude increasing with applied pressure. This trend is consistent with the static response shown in Figure b­(i), where higher loads produce larger resistance changes due to enhanced compression of the porous LIG network. Long-term stability was assessed by subjecting the sensor to 6000 tapping cycles at 220 kPa and 10 Hz. As shown in Figure b­(v), the response signals remain highly repeatable with minimal waveform distortion, indicating excellent mechanical reliability. The stable performance suggests that the porous LIG network remains structurally intact and firmly adhered to the PDMS substrate, demonstrating robustness under prolonged high-pressure operation. Owing to its high sensitivity, fast response, and excellent repeatability, the PDMS/LIG pressure sensor can be directly attached to human skin for monitoring physical movements such as finger, wrist, and knee bending. To demonstrate practical applicability, the LIG/PDMS sensor was attached to various parts of the human body to monitor motion-induced pressure variations. As shown in Figure b­(vi–viii), the device successfully detects finger bending, wrist movement, and knee motion, producing distinct electrical signals corresponding to different mechanical actions. The sensor can also distinguish motion intensity, indicating potential for gesture recognition and activity monitoring. Furthermore, the device is capable of detecting subtle pressure variations associated with neck motion during speech (Figure b­(ix)). Distinct resistance responses were observed when the volunteer pronounced the words “I,” “Love,” and “Laser,” demonstrating sensitivity to small mechanical vibrations. These results confirm that the fabricated LIG/PDMS pressure sensor is highly suitable for wearable electronics, human–machine interfaces, and health monitoring applications. Integration of the sensor output with machine learning algorithms could further enable advanced functions such as gait recognition and motion analysis.

5.3. Flexible Microheater

The optimized LIG parameters demonstrate strong tunability for tailoring the electrical and mechanical properties of LIG-based devices. Flexible heaters are gaining attention due to their diverse wearable applications much as they can be easily integrated into textiles for applications such as medical diagnostics, sports monitoring, and military systems requiring controlled thermal management. ,− Conventional flexible heaters are based on materials such as indium tin oxide and silver nanowires of require complex fabrication methods, including photolithography and 3D printing, pose challenges, − and often suffer from slow thermal responses, which limits their effectiveness in demanding applications. In contrast, carbon-based materials such as graphene offer a promising solution by leveraging their high thermal efficiency and responsiveness. , To evaluate the versatility of the optimized, LIG parameters, different LIG-based microheaters geometries were fabricated, as shown in Figures a­(i-iii) and S8a. Heaters were fabricated using optimized LIG parameters: a laser scan speed of 120 mm/s, laser power at 20%, a depth of focus (DOF) of 6 mm, and dots per inch (DPI) of 250. To record the electrothermal response of these fabricated LIG heaters, a FLIR camera setup, depicted in Figure a­(iv), was used. Considering the performance of these microheaters, electrical resistance plays a crucial role in Joule heating.

$$P_{\mathrm{h}\mathrm{e}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{r}}=\raisebox{1ex}{$V^2$}\!\left/ \!\raisebox{-1ex}{$R_{\mathrm{h}\mathrm{e}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{r}}$}\right.$$ 13

where P _heater is the power, V represents the voltage, and R _heater represents the resistance. Three heater geometries were investigated: a two-finger heater (∼535 Ω), a hexagonal heater (∼95 Ω), and a rectangular heater (∼35 Ω), depicted in Figure a­(i–iii). Other designs of finger heaters, such as the one shown in Figure S8a, were not considered due to their higher resistance. The lower resistance of the hexagonal and rectangular designs is attributed to shorter current paths and wider conductive channels. The results in Figure S8b–c show that resistance increases significantly when electrode width decreases below 500 μm, likely due to current crowding and narrowing of the conductive path. The voltage-dependent electrothermal response of the optimized LIG heater (2-Finger) is shown in Figure b, where the peak temperature increases monotonically with applied voltage (3–8 V), reaching ∼98 °C at 8 V. Figure c depicts that the surface temperature of the two-finger heater increases rapidly from ∼ 23 °C to different steady-state temperatures under applied voltage. A rapid rise occurs within ∼ 12 s, followed by saturation at approximately 98 °C under 8 V. The heater also maintains nearly identical thermal responses under flat and convex bending conditions (Figure S8d and Movie S2), indicating stable electrothermal performance under mechanical deformation. Similar responses were observed for the LIG heater (Rectangle type) depicted in Figure d, with much higher max temperature observed for much lower applied voltage. The fabricated heater exhibits exceptionally rapid heating and cooling characteristics, as shown in Figure e. Under an applied DC voltage of 6.5 V, the device demonstrates a heating time (H_t) of approximately 6.2 s and a cooling time (C_t) of approximately 12 s. The heating rate (H _r) is calculated as.

$$\mathrm{H}\mathrm{e}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{n}\mathrm{g}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}(H_{\mathrm{r}})=90\%*\raisebox{1ex}{$\mathrm{\Delta }T$}\!\left/ \!\raisebox{-1ex}{$H_{\mathrm{t}}$}\right.$$ 14

where ΔT is the temperature change during heating. As shown in Figure f, the heating rate increases from 6.32 to 28.38 °C/s as the applied voltage increases from 3.5 to 6.5 V. The device reaches ∼ 201 °C at 6.5 V within ∼ 6.2 s, corresponding to a ramp rate of ∼ 25.82 °C V^–1. It should be noted that 6.5 V represents the upper measurement limit, as higher temperatures exceeded the detection capacity of the FLIR thermal imaging system. To further evaluate the intrinsic thermal tolerance beyond this measurement limit, thermogravimetric analysis (TGA) was performed (Figure g). Only minor weight loss (∼2%) occurs at 100–120 °C due to moisture removal, and ∼ 3% between 230 and 300 °C from decomposition of residual oxygen groups. Importantly, negligible mass loss near 201 °C confirms that the optimized LIG structure remains thermally stable and can withstand temperatures beyond the FLIR detection limit without structural degradation. The operational stability of the rectangular heater was evaluated through repeated on/off cycling (Figure h). The device maintained nearly identical thermal responses during ∼2000 s of operation, indicating excellent electrothermal reliability. A comparison with recently reported flexible heaters is presented in (Figure i), highlighting the superior performance of the optimized LIG heater. − While many reported devices depicted voltages ≥10 V to achieve 100–150 °C, the heater developed in this work reaches 201 °C at only 6.5 V, demonstrating significantly improved heating efficiency. For a more detailed quantitative comparison of key electrothermal parameters including operating voltage, maximum temperature, and heating rate, the performance of the optimized heater is summarized in Table . To assess mechanical durability, the rectangular heater was transferred from PI to a flexible PDMS substrate for deformation and wearable testing (Figure ). An in-house bending and stretching setup was used. Figure a illustrates a heater that maintains a uniform thermal distribution even under different bending radii (15 mm to 7 mm), demonstrating excellent electrothermal stability under mechanical deformation. In addition, the fabricated heater was subjected to controlled tensile strain conditions ranging from 0% to 55% depicted in Figure b. The corresponding infrared images reveal nearly identical thermal profiles during stretching, indicating that the conductive LIG network maintains stable Joule heating behavior under mechanical strain. Figure c,d shows a minor reduction in steady-state temperature under increasing bending strains and tensile, likely resulting from the increase in electrical resistance induced by mechanical deformation of the device (Figure e,f). The practical applicability of the device was further demonstrated through wearable heating tests. As shown in Figure g, the flexible heater was attached to the human wrist and elbow, where consistent heating was observed in both relaxed and bending states. Infrared images confirm uniform temperature distribution across the device, highlighting its potential for wearable thermotherapy and personal thermal management applications. Importantly, the optimized LIG platform developed in this work supports both piezoresistive sensing and electrothermal heating within the same material system. Although these functionalities are demonstrated independently, the results establish a foundation for integrated multifunctional devices, such as self-heating pressure sensors. Such integration could enhance sensing reliability under varying environmental conditions and expand the applicability of LIG in wearable and smart electronic systems. To further evaluate the fabrication reproducibility of the optimized laser parameters, multiple sensors were fabricated under identical processing conditions. The statistical distribution of the electrical resistance values across eight independently fabricated devices is summarized in Figure S9 (Supporting Information) and discussed in Section S3. The devices exhibit very a small statistical variation with relative standard deviations below 2%, confirming excellent device-to-device reproducibility and fabrication reliability. A real-time batch fabrication demonstration using the optimized laser parameters is provided in Movie S3.

10.

(a) Optical images of LIG-based microheaters with different geometries (two-finger, hexagonal, and rectangular) and corresponding thermal imaging setup. (b) Voltage-dependent electrothermal response of the two-finger heater. (c) Time-dependent heating profile showing rapid temperature rise and saturation. (d) Electrothermal performance of the rectangular heater. (e) Heating and cooling behavior under applied voltage. (f) Heating rate as a function of voltage. (g) Thermogravimetric analysis (TGA) indicates thermal stability. (h) Cycling stability under repeated ON/OFF operation. (i) Comparison with reported work on flexible heaters.

2. Comparison of Electrothermal Performance of Flexible Graphene/LIG-Based Heaters Reported in the Literature .

material/heater type	substrate	voltage (V)	max temperature (°C)	heating time (s)	heating rate (°C s–1)	ref
LIG	PI/Nomex	5	∼160.2	∼30	∼4.5
LIG	Aramid Fabric	10	∼170	∼44	∼3.2
LIG	PI	10	∼195	∼45	∼3.7
Kirigami LIG	PI	8	∼120	∼20	∼4.8
Paper-based LIG	Cellulose Paper	10	∼96.6	∼20	∼3.6
LIG	PI/PDLLA PLA Film	10	∼75	∼30	∼1.7
MXene	Textile	12	∼106	∼40	∼2.1
Mxene/PLA	Electrospun PLA membrane	2	75	∼25	∼2.0
Ag NW	Transparent film	10	∼110	∼40 s	∼1.8
WDC	Carbon composite film	4	∼73	∼30	∼1.6
AgMM	Flexible polymer film	4.5	120	∼30	∼3.1
LIG This work	PI	6.5	201	6.5	∼28.38	This work

^aAbbreviations: PLA = Poly­(lactic acid), PDDLA = Poly­(d,l-lactic acid),Ag NW = Silver Nanowires, Ag MM = Silver Micro Mesh, WDC = Wood Derived Carbon.

11.

Mechanical stability and wearable demonstration of the LIG heater transferred from PI to PDMS. (a) Infrared images under different bending radii (15–7 mm). (b) Thermal response under tensile strains (0–60%). (c,d) Steady-state temperature variation under bending and stretching. (e,f) Corresponding resistance change during deformation. (g) Wearable heating on the wrist and elbow showing stable thermal performance.

6. Conclusion

In conclusion, this work establishes a systematic and reproducible framework for engineering laser-induced graphene by introducing the energy per unit area (EPUA) as a unifying laser parameter to capture the combined effect of laser processing conditions. Unlike prior studies that rely on trial-and-error approaches, this strategy enables controlled tuning of LIG morphology, crystallinity, and electrical properties, thereby improving process reliability and scalability. Beyond material optimization, this study provides key insights into the structure–property–performance relationship in porous LIG. The piezoresistive behavior is shown to arise from the coupled effects of interflake contact resistance, tunneling modulation, and microstructural rearrangement under mechanical deformation, offering a clearer understanding of sensing mechanisms in LIG-based devices. The PDMS-transferred flexible pressure sensor demonstrates a wide operating range (0–220 kPa) with distinct sensitivity regimes (∼0.005–0.0067 kPa^–1), fast dynamic response (0.38–0.15 s), and excellent mechanical stability over 6000 cycles. The optimized LIG also demonstrates efficient electrothermal performance, achieving rapid heating to 201 °C within 6.5 s at a low operating voltage of 6.5 V, corresponding to a high heating rate of ∼ 28.38 °C s^–1. These results highlight the capability of LIG for low-power thermal applications with fast response characteristics.

Overall, this work moves beyond empirical parameter tuning by providing a physically grounded and transferable methodology for LIG fabrication. The proposed framework provides a basis for adaptation to other laser systems and precursor materials through appropriate parameter tuning, including potentially low-cost visible diode laser systems. However, wavelength-dependent absorption and thermal conversion efficiency may influence the achievable graphitization quality and electrical performance. Future studies will investigate the applicability of this framework to alternative low-cost laser platforms for scalable multifunctional flexible electronic devices.

The data that support the findings of this study are available in the article and its Supporting Information.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.6c03731.

• Effects of DPI, depth of focus, and substrate on the electrical and structural properties of LIG; LIG fabrication process, effects of EPUA and DOF on morphology and quality, structural and chemical characterization, sensor response characteristics, application demonstrations, microheater performance, and device reproducibility (PDF)

• Mechanical reliability testing of an optimized flexible piezoresistive sensor (MP4)

• Electrothermal performance of an optimized LIG-based flexible heater (MP4)

• Real-time batch fabrication demonstration (MP4)

F.T.B: conceived and designed the experiments, performed laser fabrication, material characterization, and data analysis, and wrote the manuscript draft. N.T.B: Methodology, Investigation, Writing review and editing. D.M: supervised the research, contributed to result interpretation, funding acquisition, and revised the manuscript.

The authors declare no competing financial interest.