Projects/IGMAA

Industrial Grade Material Analysis and Applications is an HEI applied research project to analyze different parameters (speed, resolution, sensitivity) of current Laser-induced breakdown spectroscopy (LIBS) equipment, tools and methodologies to find their limits and propose ways to increase them.

Our refernce spectrometers are the

• Ocean Optics Maya2000 Pro
• Ocean Optics USB2000

State of the Art

The whole test setup can be divided into 3 parts:

• laser controlling
• spectrometer and controlling
• data analysis

Laser

We have a TEEM Photonic PowerChip laser that sends laser pulses at a constant rate of 1 kHz with a wavelenth of 355 nm. The duration of the laser pulse is 0.3 ns. Apparently the laser can be turned on/off with a digital signal on its serial connector, but it has a preheat time of several seconds until it's fully operational.

Spectrometer

Software

With the Ocean Optics SpectraSuite Software the minimal integration time which can be set is 7200 us. It has a mode where it saves every scan into a separate file in a tab-separated mode (see official screenshot). With the minimal integration time set, we get approximately 40 (±10%) scans per second. During this operation the processor [TODO: insert spec] on the test laptop was occupied by less than 10%.

Data Analysis

Some data analysis has been done with LabView, but it seems to be very slow [TODO: add better measures and ref]

There are different use cases:

• Detecting the existence of one or more specific element(s)
• Finding the elements present in the sample by comparison to a database of known elements
• Specify the purity of a sample (amount of undesired elements compared to desired elements)

Peak Finding

A common task in all these use cases is to find peaks in the spectrum. With these libraries, this is quickly done.

Element Detection

Assigning the found peaks to a one or more given sample spectra proofs to be rather difficult. There are two distinct tasks: find the presence of a given element and estimate its proportion.

A simple approach is to cross-correlate the two spectra. If the sampled spectrum is cross-correlated to a list of element spectra, we get a list of potential proportions of the elements in the sample. This works well, if the sample only contains few elements.

Another idea is to calculate the distance of all the peaks in a sampled spectrum to all the significant peaks in an element. The results would be a good indicator of the potential presence of the element.

Outline

In this project we will set up an acquisition chain with the laser and spectroscope given above. We will use the Ocean Optics Software to take samples and try to analyze them in real-time with a Python script.

Improvement Ideas

Sensitivity

Nokia LCD

We have transparent Nokia LCDs that can be rendered black when current is applied. These could be used to block out the emission of continuum radiation and only integrate the characteristic radiation on the spectrometer.

[TODO] Measure noise/distortion on signal added by Nokia LCD in transparent mode.