LIBS Analysis of Glued Wood Composite
Goal:
To assess the discrimination capabilities of LIBS for different layers in a glued wood composite
Method:
LIBS
Introduction:
Three features of interest can be distinguished visually in the glued wood composite. The very thin dark lines are a phenol formaldehyde resin bondline (sodium catalyst). The light tan (earlywood) and darker brown (latewood) lines represent different growth rates for the wood.
Experimental Conditions:
For the initial analysis to assess the discriminating capabilities of LIBS, a series of LIBS spectra were acquired at random locations on the sample. The goal was to obtain data that could be subjected to chemometrics analysis to determine if the different layers could be distinguished using LIBS and PCA and PLS analysis. Five different spots were measured at random locations on the sample with 7 spectra acquired at each spot to provide a depth profile. Each spectrum was single laser shot. To provide more specific data for chemometrics analysis, additional analysis was carried out on the specific features of the composite.
Hardware Used:
LIBS2000+ broadband, high-resolution spectrometer
50 and 200 mJ Nd:YAG Big Sky lasers
LIBS-SC sampling chamber with a 38 mm focal length lens (to provide a small spot size since spatial resolution is important)
Experimental Parameters:
-2 microsecond Q-switch setting
Measurement Mode:
LIBS
Results:
Based on the chemometrics analysis of the original set of data acquired with the 200 mJ laser at random locations on the glued wood composite (PLS and PCA analysis carried out by the customer – proprietary data), PCA and PLS analysis of the LIBS spectra were not sufficient for discrimination of the three layers of interest. PCA analysis showed greater differences between spectra acquired at the same spot than for the different areas sampled. PLS analysis showed a limited correlation between the depth and spectra suggesting that the actual laser pulse may have created a limited “memory” effect as it changed the wood (i.e. crater formation).
Since the first set of measurements were made at random locations on the composite, it was difficult to determine whether or not each of composite layers was sampled. To provide a more specific set of data to assess the discrimination capabilities of LIBS, an additional set of data was acquired for each of the specific layers in the glued wood composite. Analysis was carried out with the 50 mJ laser to minimize the spot size and allow sampling of the individuals layers. The average of 3 single shot spectra are shown in Figure 1 for each of the glued wood composite layers. Additional data for 6 consecutive LIBS spectra acquired at the same spot on the light tan layer are shown in Figure 2 to provide a pseudo-depth profile of this layer.
Conclusions:
A quick assessment of the spectral data showed a unique spectral line around 586 nm for the thin black bondline. This line is most likely a sodium line indicative of the sodium used to cure the phenol formaldehyde resin. Compared to the early and latewood layers, the black bondline had more sodium lines (as detected by the Elemental Identification software) and more intense sodium lines around 589 nm.
Unique spectral lines were observed for the early and latewood layers around 821, 822 and 857 nm. The identity of these lines is not known but all three of them occur near reported lines for nitrogen. The early and latewood layers also had a larger hydrogen alpha line around 656 nm related to the moisture content of the layer.
All of these unique spectral features, along with others that were not observed in the quick data assessment done, provide potential discriminating features for chemometrics analysis.
Figure 1: LIBS Spectra for Different Layers of Glued Wood Composite: Average of 3 Single Shot Spectra
Figure 2: LIBS Depth Profile for Light Tan Layer of Glued Wood Composite
