Laser-Induced Spectroscopy of Graphene Ablation in Air

Carbon Swan spectra are observed following laser ablation of graphene in laboratory air. Previous experiments showed temperatures that ranged from 4500 to 7500 K for the Δv = 0 transition and 4200 to 4500 K for the Δv = −1 transition for time delays on the order of 1.6 μs to 70 μs. This experiment explored in greater detail time delays > 10 μs for both molecular bands. Temperatures were found to be similar, ranging from 4500 to 6700 K for the Δv = 1 transition and 3200 to 5500 K for the Δv = -1 transition. Investigation is also made into spatially resolving the plasma emissions along the slit height. In addition, efforts are made to investigate the applicability of the local thermodynamic equilibrium (LTE) assumption. Comparisons are discussed in view of previous work that utilized Stark broadening of the Hβ line, confirming LTE for delays < 10 μs, yet further research needed for later delays.


Introduction
Laser-induced breakdown spectroscopy (LIBS) is a technique utilized for analyzing a substance's composition [1]. It is a non-invasive and non-contact technique that only requires a small amount of the substance to be useful while still allowing for accurate determination of unknown material [2]. This becomes especially useful when the substance itself or its environment is hazardous in nature. However, in order to be of use as a diagnostic tool, known models are required. Previous investigations had been made [4,3,5] into C 2 , though limited. To that end, investigations are made into carbon Swan spectra emissions from laser-induced plasma.

Experimental approach
The experiment utilizes a Nd:YAG pulsed laser operating at 1064-nm with 13 nanosecond, 190 mJ/pulse to generate laser-induced breakdown. The emitted spectra were dispersed with an 1800 grooves/mm grating in a HR640 Jobin-Yvon 0.6 m spectrometer and then recorded with an Andor iSstar ICCD. Measurements were time-resolved with varying time delays (10 μs to 100 μs) and corresponding gate widths (1 μs to 20 μs). Spectra were sensitivity and wavelength calibrated with a standard tungsten and mercury lamp, respectively.

Computational approach
For diatomic molecules, spectral predictions require a temperature along with line-strengths [6], for the allowed spectral transitions. In order to fit the spectra, the Nelder-Mead algorithm is utilized [7] through the so-called Nelder-Mead-Temperature (NMT) program. The NMT program generates a single temperature spectral fit for the recorded data. This fit, within the accuracy and precision bounds, is also used to refine the calibration the data. This is repeated several times, with each run increasing from a linear, quadratic, to finally a cubic calibration. After that, the program is run three times in succession to ensure accuracy of the computed temperature. Figures 1 and 2 are representative of the results from the NMT program for carbon Swan spectra of the Δv = 0 and Δv = −1 bands. Seen in both are the clear vibrational peaks and underlying rotational structure. No atomic lines are present. Note, there is good matching between the experimental and theory spectra as the count difference between both is generally within the computed background line.

Conclusions
For our theory models, we have used a single temperature for fitting with assumption of LTE. For early time delays on the order of 1 μs with large laser energy per pulse, time-delayed spectra indicate LTE due to the presence of H β [3]. Past that, there is some indication for LTE at later delays: The experimental spectra having good agreement with computed theory spectra within background variations. However, there is no definite indication of LTE at later delays. As the electron density diminishes, H β emissions disappear in the measured spectra past 10 μs. There is some indication that LTE is absent since there appears to be different temperature ranges for the different wavelength bands, as seen in comparing the Δv = 0 and Δv = −1 bands.
From the recorded 2-D data, we conclude that the plasma shows a higher temperature towards the top of the recorded emission spectra near 7 mm but is cooler away from it. Just below that region, there appear to be significant variations in inferred temperature. These temperature differences could indicate non-LTE conditions. Consequently, exploration and utilization of Abel and Radon techniques [8] in better analyzing the plasma is called for.