Spectral Analysis App

Purpose: My classes are in need of a phone app that will analyze spectral images that are taken with their phone cameras so that unknown gases can be identified.

Calibration: The first step is to use a known spectrum and calibrate the imaging instrument. This is an example of Hydrogen spectrum taken with my iphone. Notice that there are emission lines on both sides of the gas tube (pink in the center). Only one side of is needed for analysis. The wavelengths for hydrogen are in the following table:

Red / 656.8 nm
Blue / 486.1 nm
Violet 1 / 434.5 nm
Violet 2 / 410.7 nm

These wavelengths would then be used to scale the image in nm/pix across a row of the image in a manner such that the origin of the scale is the center of the emission tube.

Gaussian fitting using

is an ideal method of determining the center of the observed emission line for accurate calculations. This is one possible method that can determine if the outer-wings of emission lines are overlapping and appear blurred in the image.

The relationship between the observed spectral lines and an intensity-vs-wavelength graph is demonstrated below.

There is a continuous background and the emission lines are spikes above this continuum. Therefore, background levels will need to be taken into account when determining the existence of an emission line and the center of the line.

Experiment: Using this calibrated scale, determine the wavelengths of the emission lines of unknown gases. The examples below are Krypton and Mercury. Notice that due to the resolution of the diffraction grating not all of the emission lines are observed by eye in this image. However, large groupings of lines and relative brightness levels are significant and can be used to identify the gases.

Krypton / Mercury
676.4 nm
568.2 nm
530.9 nm
520.8 mm
482.5 nm
476.3 nm
468.0 nm
415.4 nm
413.1 nm
406.4 nm / 578 nm
546 nm
436 nm
405 nm
365 nm
334 nm
312-313 nm
297-302 nm
254 nm

Output: The output of each image is an intensity-vs-wavelength plot with the emission lines labeled, similar to the following graph.