Study on the Impact of Air Pressure on the Laser-Induced Breakdown Spectroscopy of Intumescent Fireproof Coatings

Abstract

Intumescent fireproof coatings protect steel structures and cables by forming a thick, fire-resistant layer under high temperatures. These coatings can deteriorate over time, impacting their fire resistance. Current testing methods are largely lab-based, lacking in-service evaluation platforms. Laser-Induced Breakdown Spectroscopy (LIBS) is emerging as a promising in situ detection technology but is influenced by low air pressure in high-altitude areas. This study investigates how air pressure affects LIBS signals in intumescent coatings on galvanized steel. Using pressures between 35 and 101 kPa, a linear model was developed to correlate laser pulses to ablation depth for characterizing coating thickness. Results show that spectral intensity decreases with lower air pressure. However, a strong linear relationship persists between laser pulses and ablation depth, with a fitting accuracy above 0.9. The coating thickness is identified by the number of laser pulses required to detect the Zn spectral line from the underlying galvanized steel. As air pressure decreases, the ablation depth increases. The study effectively models and corrects for air pressure effects on LIBS data, enabling its application for field detection of fireproof coatings. This advancement enhances the reliability of LIBS technology in assessing the fire performance of these materials, providing a reference for their in situ evaluation and ensuring better fire safety standards for building steel structures and cables.

1. Introduction

Intumescent fireproof coatings are widely used in the electrical sector for the fire protection of structural steel and cables due to their excellent fire resistance. These coatings work by forming a highly insulating carbon foam layer, which can expand to dozens or even hundreds of times the thickness of the original coating when exposed to flames or high temperatures. This layer is non-flammable and acts as a barrier to oxygen and external heat sources, effectively protecting the underlying material [1,2]. The thickness of the coating plays a crucial role in its fire protection performance [3,4,5]. Standards such as GB 50205-2001 [6] and GB 14907-2018 [7] specify the required thickness of fireproof coatings upon inspection.

As the years go by, fire-resistant coatings on steel structures and cables can start to peel, blister, fall off, and crack, significantly compromising their fire protection capabilities. Currently, the fire performance testing of these materials primarily relies on laboratory methods such as bundle burning tests and large-scale fire resistance tests [8]. These methods are highly specialized, require large sample sizes, have long testing durations, and incur high costs, making them more suitable for factory inspections and type testing of fire-resistant materials. However, for fire-resistant materials in use, specific application scenarios limit the sample shape, size, and quantity, making it difficult to conduct bundle burning tests and other performance evaluations. There is a lack of testing platforms and evaluation methods for assessing the fire performance of materials during their service life. Figure 1 illustrates the application scenarios of fire-resistant coatings.

Laser-induced breakdown spectroscopy (LIBS) is an advanced detection technology composed of a laser, optical components, fiber optics, a spectrometer, and a computer. The laser emits high-energy pulses, which are focused on the sample surface by optical components like mirrors and focusing lenses. This ablates the sample and forms plasma. During the cooling process, the plasma emits spectral information at specific wavelengths, which is collected by the fiber optics and transmitted to the spectrometer. The spectrometer, connected to a computer, transfers these spectral data to the computer [9]. Through this spectral information, a detailed analysis of the sample’s elemental composition, content, and surface conditions can be conducted.

LIBS has developed rapidly in recent decades, finding extensive applications in fields such as food safety [10,11], biomedicine [12], steel smelting [13,14], space exploration [15], environmental monitoring [16,17], archaeological excavation [18,19], and coal mining [20]. In the field of electrical engineering, materials such as fireproof coatings, ceramics, and silicone rubber are widely used [21,22,23]; however, the application of laser-induced breakdown spectroscopy technology in the detection of electrical materials is still in the exploratory stage. Chen and colleagues [24] used LIBS to test silicone rubber composite insulator materials with varying hardness (different filler content). They developed a hardness-spectral data calibration model by combining principal component analysis and neural network algorithms, achieving accurate measurement of the silicone rubber samples’ hardness. Wang and colleagues [25] used LIBS to analyze contaminants on the surfaces of insulators with varying compositions and concentrations. They developed a calibration model using partial least squares regression to correlate the spectral data with the contamination levels on the insulator surfaces, enabling the rapid detection of insulator contamination components. Li and colleagues [26] conducted a quantitative analysis of coal samples with different calorific values using LIBS. They established a quantitative model using the partial least squares method and compared the effects of various spectral data preprocessing algorithms, such as smoothing and the Savitzky–Golay filter, on the model. The results showed that second-order smoothing preprocessing could enhance the accuracy of the partial least squares model, reducing the root mean square error of the prediction model from 2.36 × 10⁹ mJ/kg to 2.76 × 10⁸ mJ/kg.

There has been considerable research on the impact of environmental factors on LIBS technology. Zou and colleagues [27] studied the effect of sample temperature on LIBS signals. The experimental results indicated that increasing the sample temperature significantly enhanced the intensity and reproducibility of the LIBS signals. In experiments with artificial samples, raising the sample temperature to 973 K led to a 280% increase in signal intensity, a 20% improvement in the signal-to-noise ratio (SNR), and a 260% reduction in the limit of detection (LOD). Lei and colleagues [28] investigated the influence of relative humidity on air plasma, finding that humidity has a significant impact on the emitted spectra. Scott, J.R. and colleagues [29] examined the LIBS spectra under various atmospheric conditions and discovered that altering pressure and the composition of the background gas can significantly affect the spectral results; lowering the pressure and using argon and helium gases can enhance the spectral intensity, SNR, and resolution.

Research on the use of LIBS for detecting coating thickness is relatively rare. Yuan and colleagues [30] characterized solid samples with LIBS, investigating the emission characteristics of aluminum (Al) and zinc (Zn) spectral lines in laser-induced plasmas, and exploring the relationships between element concentration and film thickness. Liu and colleagues [31] combined LIBS technology with a laser profile and interface roughness model (LPIR model) to study how the ablation rates of different materials affect the accuracy of thickness measurements of deposited layers. Their results indicated that the thickness correction method they proposed significantly improved the measurement accuracy of multilayer samples. Yang and colleagues [32] introduced a method for the online monitoring of the laser coating removal thickness of double-layer coatings on aluminum alloy plates using a high-frequency nanosecond infrared pulsed laser combined with LIBS technology. Through the standard curve method and principal component analysis-support vector regression (PCA-SVR) algorithm, they achieved high-precision online monitoring.

Around the world, many countries are actively researching high-voltage transmission technology, which relies on fireproof coatings to enhance safety. For example, Brazil’s Belo Monte Hydroelectric Plant project utilizes ±800 kV direct current transmission technology to transport electricity from the northern Amazon region to the southeast. India is implementing multiple ultra-high voltage transmission projects to connect its resource-rich northeastern region with other load centers. In the United States, projects like the Pacific DC Intertie are used to transmit renewable energy from resource-rich areas to load centers. China’s ultra-high voltage (UHV) and extra-high voltage (EHV) transmission lines cover extensive areas, including high-altitude regions such as Qinghai and Tibet. The Qinghai–Tibet Grid Interconnection Project was put into operation in 2011; in November 2014, the 500 kV Sichuan–Tibet Interconnection Transmission and Transformation Project was launched; and in December 2020, the ultra-high voltage transmission line known as the “Sky Road” connecting Ali and Central Tibet was officially put into operation. Unlike other countries, these transmission projects in China pass through high-altitude areas such as Lhasa in Tibet (with an atmospheric pressure of about 56 kPa) and Xining in Qinghai (with an atmospheric pressure of about 70 kPa). The low-pressure environment can affect the interaction processes between lasers and samples, as well as the light collection process in optical fibers, which can interfere with experimental data and impact the accuracy of LIBS detection. Therefore, studying the impact of atmospheric pressure on LIBS data and its mechanisms is particularly important.

This research utilizes LIBS to collect spectral data from fireproof coating samples under different pressures (35–101 kPa) and analyzes how the spectral data vary with pressure. A confocal microscope is used to observe the ablation morphology of the samples, and a linear model is established to relate the number of laser pulses to the ablation depth of the samples. The thickness of the fireproof coating is characterized based on the number of laser pulses required to penetrate the coating. Additionally, by establishing a linear model under different pressures, the impact of pressure on the measurement of fireproof coating thickness is mitigated. This study facilitates the accurate characterization of fireproof coating thickness under varying pressure conditions, providing a reference for assessing the fireproof performance of these coatings.

2. Materials and Methods

2.1. Equipment

Combining a pressure control device with a LIBS system allows for experiments under varying pressures. The pressure control device comprises components like a pump and a pressure gauge, while the LIBS system includes a laser, optical elements, a spectrometer, and a delay generator. The schematic of the setup is shown in Figure 2.

The laser emits a high-energy pulsed laser light, which enters the chamber through a window in the vacuum box. The laser is reflected by a medium film mirror to a long focal length focusing lens, and is focused onto the surface of the sample, ablating the sample to generate plasma. The plasma emits characteristic spectra as it cools, and part of the spectral signal is collected by a fiber optic probe after being focused by a short focal length lens, ultimately transmitted to a computer for processing via a spectrometer. A timer is used to adjust the triggering sequence between the laser and the spectrometer, thereby avoiding interference from background signals. A vacuum pump is used to extract gas from the chamber, and a pressure gauge is used to monitor the gas pressure inside the chamber.

The laser selected is the Nd:YAG laser from Lei Bao Optoelectronics Company (Beijing, China), model Nimma-900, which operates at a wavelength of 1064 nm. This wavelength of laser exhibits good energy stability and a low air attenuation rate, with a maximum single pulse energy of 900 mJ, pulse width of ≤ 9 ns, a laser beam diameter of 9 mm, and a working frequency of 1–10 Hz, utilizing an external trigger mode. The spectrometer selected is the AVS-RACKMOUNT-USB2 from Beijing Avantes Co., Ltd. (Beijing, China), with a wavelength range of 196 nm to 641 nm and a resolution of 0.01–0.15 nm. The timer chosen is the eight-channel digital delay generator DG645 form Stanford Research Systems (CA, USA), with a delay time set to 2.5 μs. The delay resolution for all channels of the signal delay generator DG645 is 5 ps, with jitter between channels being less than 25 ps, and a delay control accuracy of 1 ns. This signal generator is used to set the delays for the laser and spectrometer, emitting the Q-switch signal for the laser and the external trigger signal for the spectrometer to control the timing sequence.

The pressure vessel shell is made of stainless steel to prevent deformation of the device due to the pressure difference between the inside and outside of the chamber at low air pressure. The windows used are optical lenses, allowing lasers to enter the pressure control device through the optical lens window. The configured devices, such as air pumps and pressure gauges, can regulate and measure the air pressure within the pressure control device, which can be as low as 45 kPa. Additionally, the confocal microscope used is the VT6100 laser confocal microscope from Zhongtu Instruments Co., Ltd. (Shenzhen, China), with a height measurement resolution of 0.5 nm and a width measurement resolution of 1 nm. The light source is a white LED, and the travel range in the xyz directions is 100 mm.

2.2. Materials

The test sample is an expansion-type fireproof coating for steel structures, produced by Langfang Dahao Fireproof Materials Co., Ltd (Langfang, China). The applicable standard is GB 14907-2018 [7]. It was evenly applied with a brush on the surface of a galvanized steel sample measuring 5 cm × 5 cm × 0.2 cm, as shown in Figure 3. The galvanized steel was purchased from Foshan Guangdu Steel Co., Ltd (Foshan, China).

2.3. Data Preprocessing

First, the raw spectra were normalized using area normalization in order to mitigate the impacts of pulse energy fluctuations and matrix effects. Subsequently, discrete wavelet transform was employed for baseline correction and noise reduction, thereby weakening environmental noise and accentuating useful signals. Moreover, continuous wavelet transform was utilized for effective peak identification, acquiring the wavelength information of peaks for element recognition. Finally, based on the peak-seeking results, the central wavelengths of the test spectral peaks were compared with standard wavelengths from the Atomic Spectra Database (ASD) [33] of the National Institute of Standards and Technology (NIST) to determine the elemental attribution of the spectral lines.

3. Results and Discussion

3.1. Microstructure and Composition of Fireproof Coatings

Fireproof coatings generally consist of a binder, flame retardants, and fillers. The binder forms the foundation of the coating and is the main film-forming substance. It can be divided into organic and inorganic film formers. Organic film formers come in various types and are typically flame-resistant synthetic resins, such as phenolic and polyester resins, while inorganic film formers include silicates and phosphates. Flame retardants are the key components that enable fireproof coatings to perform their protective function. Common flame retardants include halogenated, inorganic, and phosphorus-based flame retardants. Fillers can enhance the mechanical and physical properties of fireproof coatings, such as color, acid and alkali resistance, and flame resistance. Common fillers include titanium dioxide and antimony trioxide.

Scanning electron microscopy (SEM) was used to test the fireproof coating applied to the surface of galvanized steel, revealing the microscopic structure of the coating as shown in Figure 4. The surface of the coating appears rough and porous, with uneven particle sizes and shapes. There are noticeable gaps and voids between the particles, which may facilitate the expansion of the coating at high temperatures, thereby forming an insulating layer that enhances fire resistance.

The Energy Dispersive Spectroscopy (EDS) test was conducted on the fireproof coating to study its composition, and the results are shown in

Table 1. Fireproof coating EDS test results.

Element	Quality Score (%)
C	40.57
O	37.29
Na	0.49
Al	1.07
Si	1.73
P	11.54
Ca	1.20
Ti	4.49
Mo	1.38
Cl	0.22

. According to Table 1, the main elements of the fireproof coating are C, O, Ti, P, Al, Si, and Ca. The right side of the table indicates that the highest contents are C and O, suggesting that the base material is an organic film-forming substance. The presence of Al, Ca, and Ti indicates that titanium dioxide, calcium carbonate, and aluminum hydroxide are added as fillers in the fireproof coating. The presence of P indicates that the flame retardant used is a phosphorus-based flame retardant.

3.2. Trends in Spectral Line Intensity Changes

To determine the characteristic spectrum of fireproof samples, we selected a galvanized steel sample coated with fireproof paint for the experiment. Initial tests were conducted using LIBS, with the laser energy set at 85.5 mJ, a frequency of 1 Hz, a delay time of 2.5 μs, atmospheric pressure at 101 kPa, and an ambient temperature of 19 °C. Since the fireproof paint was applied to the surface of the galvanized steel, we also performed LIBS on the galvanized steel sample to distinguish between the two materials. The spectrograms for the galvanized steel and fireproof paint are shown in Figure 3.

Based on the NIST database for line selection, as shown in Figure 5, the coating on galvanized steel is a zinc layer. Therefore, the Zn spectral lines were chosen as the characteristic spectral lines for galvanized steel materials, which include the elemental spectral lines Zn I 334.56 nm and Zn I 472.22 nm, among others. The spectral graph of the fireproof coating presents characteristic elemental spectral lines such as Ti II 334.94 nm, Ca II 393.37 nm, and Ti I 453.48 nm.

The fireproof coatings Ti II 334.94 nm and Ti I 453.48 nm were selected to study the effect of air pressure on the intensity of spectral lines. The pressure range was set at a range of 35 kPa to 100 kPa, with a laser energy of 75.9 mJ, a delay time of 2.5 μs, and the temperature during testing was the ambient temperature of 19 °C.

As shown in Figure 6, the intensity of the Ti II 334.94 nm and Ti I 453.48 nm spectral lines generally decreases as the air pressure drops. The decline is more gradual from 100 to 70 kPa, and more rapid from 70 to 35 kPa. This trend can be attributed to the weakening of the environmental gas confinement effect with decreasing pressure, which accelerates the expansion rate of the plasma. As a result, the number of particles per unit volume in the plasma decreases, reducing the likelihood of collisions between particles and weakening avalanche ionization, ultimately leading to a decrease in spectral line intensity.

3.3. Coating Thickness Measurement

Research has shown that there is a strong linear relationship between the number of laser applications and the ablation depth of the sample [34]. It is possible to attempt to establish a linear model relating the number of laser applications to the ablation depth of fireproof coatings (applied to galvanized steel surfaces) in order to characterize their coating thickness. When creating the linear model for the thickness of the fireproof coating with respect to the number of laser applications, it is essential to control the number of laser applications to prevent the laser from penetrating through the fireproof coating.

The air pressure was set at 101 kPa, 60 kPa, and 50 kPa, the laser energy was set at 85.5 mJ, the laser frequency at 1 Hz, and the number of laser applications was set at 5, 10, 15, 20, 25, and 30. The 3D morphology of the ablation pits on the sample surface was observed using a confocal microscope, with results shown in Figure 7.

As shown in Figure 7, the confocal microscope reveals the 3D ablation morphology of the fireproof coating sample’s surface. The surface’s 3D ablation morphology roughly resembles an inverted cone, but it is not uniform. This unevenness is due to the high temperature of the plasma generated during the laser ablation process, which causes the melted material to become fluid and splatter, resulting in inconsistencies in the ablation morphology of the sample.

For a more detailed quantitative analysis, the built-in analysis software of the confocal instrument was used to statistically evaluate the specific information regarding the depth of the ablation pits of the samples under different air pressures and laser pulse counts, as illustrated in Figure 8.

Figure 8 shows that the depth of laser ablation pits increases with the number of laser pulses at gas pressures of 101 kPa, 60 kPa, and 50 kPa. When the number of laser pulses is 5, the depth of the ablation pits is approximately 150 μm; at 30 laser pulses, the depth reaches about 400 μm. Furthermore, there is a strong linear relationship between the number of laser pulses and the ablation depth across all pressure levels, with coefficients of determination of 0.963, 0.997, and 0.981, respectively. The fitting results are satisfactory, indicating that the ablation depth of the sample can be determined by the number of laser pulses.

As the atmospheric pressure decreases, the linear fitting curve of the number of laser pulses versus the depth of the ablation pit gradually rises. This means that with the same number of laser pulses, the depth of the ablation pit increases as the pressure drops. The reason for this is that, as the pressure decreases, the expansion speed of the plasma increases, the plasma shielding effect weakens, and the interaction between the laser and the sample becomes more effective, leading to deeper ablation. This indicates that atmospheric pressure influences the measurement of coating thickness, therefore, it is necessary to establish a relationship curve between ablation depth and laser pulses at different pressures, in order to correct for the effects of pressure on measurements.

The characteristic spectral line of the Zn element (Zn I 334.56 nm) was chosen as the basis for determining whether the fire-retardant coating had been penetrated by the laser. Laser penetration tests on the fire-retardant coating were conducted under conditions of 101 kPa and 50 kPa, with the laser energy set at 85.5 mJ and the number of laser shots set at 60. The intensity of the Zn I 344.56 nm characteristic spectral line was observed regarding its variation with the number of laser shots, as shown in Figure 7.

From Figure 9, it can be noted that under conditions of 101 kPa and 50 kPa, the intensity of the Zn I 334.56 nm spectral line initially starts at 0 and suddenly appears during a specific laser pulse. The number of laser pulses required for the Zn element spectral line to become observable varies under different conditions.

Comparing the number of laser pulses corresponding to the intensity of the Zn element spectrum under the conditions of 50 kPa and 101 kPa, the number of laser pulses at 50 kPa is 16, while at 101 kPa it is 23. This indicates that the coating thickness at 50 kPa is between the thicknesses corresponding to 15 and 16 laser pulses, and at 101 kPa the coating thickness is between the thicknesses corresponding to 22 and 23 laser pulses. Based on the fitting curve of ablation depth versus the number of laser pulses, the coating thickness at 50 kPa is between 219.3 µm and 234.4 µm, while at 101 kPa, the coating thickness is between 337.8 µm and 350.2 µm. The main reason for the difference between these two measurements is attributed to the non-uniformity of the coating thickness. Theoretically, the thickness values obtained at the same point under different pressures should be the same, with only the number of laser pulses differing. The LIBS technology used in this article, combined with the coating thickness testing method established by confocal microscopy, can theoretically effectively avoid interference from air pressure in coating tests.

4. Conclusions

This paper explored the impact of air pressure on the laser spectral data of intumescent fire-retardant coatings using LIBS, and delved into the mechanisms driving these influences. Moreover, it examined how LIBS technology can be utilized to measure the thickness of fire-retardant coatings under varying air pressures, concluding with the following insights:

(1)

As atmospheric pressure decreases, the intensity of spectral lines for titanium elements in the fireproof coating similarly diminishes. This trend is attributed to the reduced constraining effects of environmental gases at lower pressures.

(2)

There is a strong linear relationship between the number of laser pulses in fire-retardant coating samples and their ablation depth, with a goodness-of-fit of 0.963 for the linear model at 101 kPa. The Zn spectral line, characteristic of galvanized steel materials, can be used to determine the thickness of the fire-retardant coating by identifying the specific number of laser pulses required for the Zn spectral line to appear.

(3)

A decrease in atmospheric pressure leads to an increase in the ablation depth of fireproof coating samples due to the weakening of the plasma shielding effect. By establishing a linear model of ablation depth versus laser pulses at different pressure levels, it is possible to mitigate the influence of pressure on the thickness measurement of fireproof coatings.

This study offers preliminary insights into how air pressure affects the laser spectral data of intumescent fire-retardant coatings and the underlying mechanisms involved. It also explores the application of LIBS for measuring coating thickness under different pressure conditions, providing a reference for assessing the fire performance of these coatings. Future research could advance the on-site characterization of the fire performance of fire-retardant coatings by incorporating technological innovations such as equipment miniaturization and integration with drone technology.