Application Of Laser-Induced Breakdown Spectroscopy (Libs) In Scrap Recycling

What is LIBS?

LIBS stands for Laser-Induced Breakdown Spectroscopy. It is a cutting-edge analytical technique use to classify materials and determine their elemental composition. The LIBS analyzers work by using a highly focused laser pulse to ablate the surface of a sample. The excited surfaces form excited atoms and ions called plasma, under a decay time all the plasma redeposit on the surface and the electrons back into the round states, then emitting light at characteristic wavelengths or unique fingerprints for each kind of element, composition, or molecular are collected and converted by a spectrometer. The LIBS is an excellent tool for quantitive and qualitative measurements.

How LIBS help in sorting scrap?

LIBS technology offers rapid and multi-element analysis of materials in various physical states, including gas, liquid, and solid. In this application note, the LIBS is described as a powerful tool for the analysis of scrap for recycling purposes and aluminum consumption especially. Its unique capabilities include the analysis of light elements, in situ analysis outside the laboratory, fast spatially-resolved compositional imaging, chemical analysis in extreme environments, and remote and automated standoff analysis. This technology enables real-time, in situ analysis, providing data elements on the alloys. This method can not only set a single alloying element but all elements are possible together with their related, or any desired, concentration level (such as 0.4% < Mg <0.9%, 0.2% < Si <0.6%, <0.1% Mn, <0.1% Zn, <0.35% Fe, and <0.1% Cu, etc.). Furthermore, LIBS instrumentation is generally more cost-effective with lower day-to-day operating costs compared to many other analytical techniques.

Processing of LIBS

The LIBS steps include sorting targets, sampling and measuring procedures, signal pre-processing, and classification.

Firstly, the initial stage involves the utilization of magnetic separators to extract magnetizable ferrous scrap, thereby reducing the volume of ferrous materials present. Subsequently, eddy current separators are employed to isolate plastics from the input stream. Following this, a further refinement process entails the extraction of denser materials such as copper, brass, zinc, etc. from the aluminum scrap.

Secondly, the utilization of a laser pulse in conjunction with an analyzer results in the generation of plasma upon interaction with the sample surface. This plasma subsequently atomizes and excites the sample, leading to the emission of light. The emitted light is then conveyed through fiber optics into a spectrometer, thereby yielding spectral data called optical emission spectroscopy (OES). Every peak spectrum matches with the standard peak on the periodic table, and the intensity of peaks relates to their percentages in the substance.

Thirdly, the interpretation and analysis of this data are facilitated through the employment of deep learning algorithms. The spectra are pre-processed to enhance the performance of the classifying algorithms. This step is always applied to deal with a superposition of the emission signal of the analyte, plasma, and detector-associated signal. A second pre-processing step is the removal of recorded intensities that relate to saturated pixels and signal-to-noise.

Lastly, deep neural networks to model and predict continuous numerical values. It involves training neural networks with multiple layers to learn complex patterns and relationships within the data. The performance of these methods is expressed with four metrics: accuracy, weighted average precision, weighted average recall, and weighted average f1 score. Figure 1 illustrates the process of scrap sorting by using LIBS method.

Figure 1. Schematic diagram of the LIBS experimental processing.

Reference sorting result

To clarify the application, here is the referred work of S V Eynde, et al. shows the result of aluminum alloy sorting by using LIBS method.

There are three selected target fractions consisting of Premium, Desox, and Second classes. The first class has limits to the concentrations of the most common alloying elements. The second class is called "Desox" standing for "deoxidation aluminum" which shows a high concentration of metallic aluminum. The last target fraction is the "Secondary" class for the production of cast alloys.

Table 1 shows the composition of the three desired output fractions when all pieces in the dataset are assigned to the three classes according to the ground truth labels. The concentrations of the elements are expressed in weight percent (wt%). As a result of the chosen approach to assigning the ground truth labels, some less critical elements narrowly exceed the specified concentration limits for the desired output fractions (marked in yellow). The iron content is slightly higher than desired in each output fraction, as well as the magnesium content in the Desox class. However, all other considered alloying elements, including the most critical elements, meet the specified restrictions by a significant margin.

Table 1. Composition of desired output fractions when separated according to ground truth measurements and specified concentration tolerances (concentrations in wt%).

Recommended equipment

The LIBS method is divided into some different kinds of technologies that work on different kinds of laser energies showed in Table 2.

Table 2. Methods of LIBS signal enhancement.

Herein, Intins introduces to customers LIBS25000plus and SpeedSorter™ LIBS Sorting Sensor, two popular product lines from Ocean Optics. Ocean Optics’ LIBS functionality can be purchased as a complete system, including the spectrometers and related fibers, laser, and LIBS Imaging Module.

For LIBS25000plus, a high-intensity, 10 nanosecond-wide laser pulse beam is focused on the sample area. All elements have emission spectra in the 200-980 nm region. The detection system uses up to seven of our HR2000+ High-resolution Miniature Fiber Optic Spectrometers, each with a 2048-element linear CCD array. The camera is also useful for adjusting the laser focus above or below the sample surface. The USB-enabled color camera captures pre- and post-ablation images of the sample and provides up to 1280 x 1024 pixel resolution. You can also purchase LIBS fiber bundles in various wavelength ranges, suitable lasers such as electro-optic Q-switch, and a LIBS sample chamber.

Similarly with the SpeedSorter, separation of wrought from cast aluminum, aluminum from magnesium, and alloy class separations like 5xxx and 6xxx are all easily accomplished.

Figure 2. a) LIBS25000plus, b) SpeedSorter™ LIBS Sorting Sensor