Re: How does an atomic emission spectrometer work?

Nick - Atomic emission spectroscopy is based on the phenomenon that when 
atome in a gaseous state (like in a flame) are heated the outer electrons 
are placed in a more energetic state.  At some point in time (very 
rapidly) that energy is lost as a photon of visible light and the electron 
goes back to its initial (ground) state.  A typical home example of that 
is the yellow flame you see when a bunen burner with a properly adjusted 
flame (transparent blue flame) is tapped and there is a burst of yellow 
seen in the flame.  That yellow is from sodium (wavelength 5890 Angstrom)
that is commonly found in the environment.  An atomic emission 
spectrometer has a heat source (gas burner, graphite furnace, carbon 
eletrodes, or induction coupled plasma) that excited the sample.  The 
sample is typically dissolved in liquid to form a solution.  The resulting 
solution is then transported into the heat source.  For my example, a gas 
burner with an aspirator that sucks the sample into the burner and then 
impacted the sample stream against a glass bead to form a mist of the 
sample. A small portion of the sample (5 to 15 percent)is transported into 
the flame and is analyzed.  The light is given off by the element of 
interest is focused using lenses onto the entrance slit of a 
monochromator.  A monochromator is a device that filters out all of the 
light except the color (Wavelength) that I wish to measure.  At the exit 
of the monochromator, a detector of some type is used such as a 
photomultiplier tube, photodiode, photodiode arrays, or a charged coupled 
device (CCD) that converts the light into electrical energy.  The amount 
of electrical energy is measured and recorded by a device like a 
voltmeter, a chart recorder, or a computer.  By keeping all variables 
constant like the temperature of the flame, the amount of sample going 
into the flame, if I vary the concentration of the material I want to 
measure, I will get a larger amount of electrical energy out of the 
detector if I have more of the analyte I'm trying to determine.  With five 
or so standard solutions with known concentrations, I can develop a linear 
relationship between the amount of light given off in the flame and the 
concentration of my standards.  I then can aspirate an unknown sample and 
calculated the concentration of the element I am interested in based on 
the amount of light given off.  One point-when I say light given off I am 
refering to light of a particular wavelength (color) that is 
characteristic for the element of interest.  I will attempt to forward 
more info and websites to you as my work schedule permits this week. Feel 
free to e-mail me directly if I can help you further.