Insulated glass filled with inert gases has gained widespread application and attention as an energy-efficient and environmentally friendly building material in recent years. Due to its high volume fraction in the air and low inflation cost, argon has become the preferred gas for ensuring the thermal insulation and soundproofing performance of glass. However, over time, argon leakage or air infiltration may occur in insulated glass, leading to a reduction in both thermal insulation and soundproofing performance. Existing detection methods, such as high-voltage spark discharge and gas chromatography, suffer from limitations such as low measurement accuracy, complex operation, or inability to conduct on-site rapid detection. Therefore, in order to meet the practical needs of high-precision, convenience, and on-site applicability for leak detection of argon filled insulated glass. this study aims to develop an argon volume fraction measurement method based on laser-induced breakdown spectroscopy (LIBS). By optimizing experimental design and data processing methods, the precision and prediction accuracy of the detection system are improved, providing technical support for gas leakage monitoring in the production, manufacturing, and use of insulated glass.

According to the detection requirements of argon volume fraction in argon filled insulated glass, the experimental research on argon detection based on LIBS is carried out. First, the LIBS system for gas analysis is built, and the experiments of argon samples with different volume fractions are carried out. The LIBS data of argon with nitrogen or air as the dilution gas are obtained, and three argon atomic emission spectral lines are selected for analysis. The precision of spectral line intensity is analyzed, and the stability of spectral line is improved by the multiple averaging method. Then, the sensitivity and linear dynamic range of spectral line intensity relative to argon volume fraction are analyzed, and the cross validation of volume fraction intervals is carried out according to the results. Finally, the calibration curves are constructed by linear fitting and cubic fitting, respectively, and the calibration analysis is carried out to analyze the accuracy of argon quantitative analysis.

The experimental results with nitrogen as the dilution gas show that the multiple averaging method can reduce the fluctuations in spectral line intensity to improve precision and resolution. The relative standard deviation of argon data with 85% volume fraction is reduced from 5.5% to 2.8%, and that of argon data with 90% volume fraction is reduced from 5.4% to 2.4% (Fig. 4). Additionally, the sensitivity and linear dynamic range for volume fraction detection are analyzed using the first and second derivatives of the spectral line intensity with respect to gas volume fraction . Among these three spectral lines, the sensitivity of the Ar I 738.40 nm line is higher than the other two, but it shows a significant decrease at higher volume fractions [Fig. 6(a)]. The second-order differential of the spectral line Ar I 727.29 nm is closer to zero [Fig. 6(b)], and its linear dynamic range is the largest. The sensitivity and linear dynamic range of the spectral line Ar I 696.54 nm are between the two. Further analysis using piecewise fitting methods reveals that cubic curve fitting provides more accurate prediction results than linear fitting. Moreover, the relative error of piecewise fitting is significantly smaller than that of overall fitting. The relationship between predicted volume fraction and true volume fraction is very close to the c=c? [Fig. 8(a)]. The absolute errors for all three fitting results are within the range of -2% to 2% [Fig. 8(b)], and the relative errors are generally less than 5%, with the largest relative error occurring at a volume fraction of 5% [Fig. 8(c)]. The experimental results with air as the dilution gas show that the overall fitting using the spectral line Ar I 696.54 nm has the relative error of less than 10% except for 5% volume fraction gas. The results of piecewise linear fitting using spectral line Ar I 727.29 nm and spectral line Ar I 738.40 nm show that the relative error of prediction is less than 5% except for 5% volume fraction gas (Fig. 9).

In response to the need for detecting argon volume fraction in argon filled insulated glass, this study establishes a Gas-LIBS system for gas analysis and employs LIBS technology to excite and measure three characteristic argon emission lines. Multiple repeated experiments are conducted under different argon volume fractions, with nitrogen or air as background gas. The experimental results show that the use of the multiple averaging method significantly improves the precision of the LIBS analysis. Furthermore, it is found that selecting spectral lines for piecewise fitting effectively enhances the accuracy of the prediction results. For different volume fractions of argon diluted with nitrogen, calibration curves are established using piecewise linear fitting. The absolute errors for the three fitting results are larger at the volume fraction extremes, and the absolute errors exhibit distinct cubic curve characteristics. Given the nonlinear relationship between LIBS spectral line intensity and argon volume fraction , cubic curve fitting is applied. Compared to linear fitting, cubic curve fitting provides more accurate predictions of argon volume fraction , with absolute errors for all three fitting results within the range of -2% to 2%. The cubic curve fitting method is also applied to argon diluted with air to simulate real-world conditions. The results indicate that the fitting results for air as diluent gas have slightly larger fluctuations compared to those for nitrogen as diluent gas, but the absolute errors are also within the range of -3% to 3%, further demonstrating the advantages of piecewise fitting. The investigation here confirms the feasibility of its application in the analysis of argon volume fraction in the production and application of argon filled insulated glass, and provides a new detection scheme for the industry of argon filled insulated glass.