PHTN1432: Vacuum Systems and Thin Film Technology
Lab #2 - Optical Emission Spectroscopy (2019W)


In this lab you will use Optical Emission Spectroscopy (OES) to analyze the discharge spectrum of both pure and mixed compositions of gases to determine their composition as well as (for pure gases) the presence of impurities. Skills form this lab include: a basic understanding of how vacuum systems are used, a basic understanding of gas discharges, and the ability to apply spectroscopy techniques to analyze discharge spectra.

Pre-Lab (Do this before your assigned lab period)


This experiment is to be performed on the Alcatel system. The basic procedures to operate this system may be found in the SOP however be aware the system is designed to be reconfigured rapidly for various labs. In this case the system is configured as follows:

Alcatel System - Photo
A photo of the Alcatel system. The pump sits under the bench (the FORELINE valve is atop the mechanical pump).

Alcatel System - Valves
Valves of the Alcatel system. The yellow pipe is the main line from the turbomolecular pump, the orange pipe is the main high-vacuum manifold, and the red pipes the experiment (subject to change). Shown here are valves for this optical emission spectrum discharge tube.

Alcatel System - Valves
A photo of the valves in the system. The MFC is unused in this experiment.

Alcatel System - OES
The dicharge tube setup. The fiber which connects to the OOI mounts in front of the glass window of the tube while the opposite side consists of two electrodes connected to a high-voltage supply. At the top, the red tube connects the helium tank to the needle valve allowing helium to be injected into the discharge tube manifold in a controlled manner.

The Experiment

As with the majority of labs in this course, the vacuum pumps will altready likely have been started ahead of time and will be running as you enter the lab. In part, we do this to prolong the life of the equipment since start/stop cycles are harder on the equipment than simply allowing it to run between lab sections.

If the vacuum system is not already running:

Start the mechanical forepump (under the bench, the power switch is on the rear of the pump), open the foreline valve, MAIN valve, and MANIFOLD valve. In this state, the experiment manifold is being rough-pumped. Wait until the system pressure decreases to about 10 torr as indicated on the manifold gauge (which reads to 1000 torr ... the gauge on the pump manifold will not read at all until pressure is below 1 torr). Start the Alcatel turbomolecular pump. You should observe the manifold pressure decrease on the Baratron gauges. Ensure the type 246 MFC controller is set to OFF and the gas inject needle valve is closed while the system is evacuated to ensure gas does not enter the manifold. Continue pumping until the manifold gauge as well as the high-vacuum Baratron gauges both read zero.

With the vacuum system running:

With the pumps running and manifold valves open, wait a few minutes, (until both Baratron gauges should read zero) then turn the ion gauge ON to observe the base pressure (Filament ON) - it should be in the 10-5 torr range by now. Turn the ion gauge OFF (a 30 second ON cycle time will help preserve the life of the gauge). If the Baratron gauges do not read zero, they should be zeroed now (this happens, especially, if the gauge controllers were not left on ... one Baratron has a special heater and should be left on overnight during the term). To zero the gauges, use a small screwdriver to adjust the ZERO setting on each.

During the pumpdown process (and before beginning the rest of the lab), capture a spectrum from a pure mercury discharge tube to verify calibration. The OOI uses "Oceanview" software on a PC to capture spectra which can then be saved as a data file for later analysis. Aim the fiber optic at a mercury discharge tube and set the integration time so that largest peak observed is under the maximum value allowed and the output is not saturated (i.e. peaks do not have a "flat top"). Capture the spectrum and save it to a USB key.

Oceanview Software Screen

When capturing the spectrum from a gas discharge, you might need to store a dark spectrum (with the tube off) then subtract the spectrum of the discharge (with the tube on) from that spectrum to obtain an output clean from electrical and optical noise in which peaks will appear prominently. Ask for clarification in the lab if required.

When pumpdown is complete and the lowest vacuum possible is achieved, close the manifold valve to isolate the discharge tube completely from the vacuum system. Turn the high-voltage power supply for the discharge tube on: since the tube is at high vacuum no discharge will be observed. Now, open the needle valve SLOWLY to admit just enough air into the tube to start the glow discharge. Admit enough gas to cause a bright glow discharge - if the pressure is too high, simply open the manifold valve a little to reduce pressure within the tube. When a stable and bright glow is observed, adjust the integration time again (to ensure the largest peak is within the range of the OOI) then capture the spectrum of air, saving it to a USB key.

Open the manifold valve and evacuate the discharge tube again to the lowest level possible.

Connect the helium line to the needle valve, purging the line of air for purity. Close the manifold valve and admit helium into the discharge tube to start a bright glow (which will be a significantly different colour than that of air). Capture the helium spectrum.

You should hence have three spectra captured: mercury, air, and pure helium.

When complete, turn the high-voltage supply OFF and gas inlet valves OFF. If you are the last lab group close the MANIFOLD, MAIN, and FORELINE valves, then turn the turbo pump off and the mechanical pump off. If you are not the last group, leave the pumps running.


Uncertainty Calculations When comparing peaks, it is important to know the uncertainty in observed results so the range in which the observed peak falls can be determined. The basic minimum uncertainty for each observed peak (the "plus/minus" value) can be estimated simply by observing the wavelength difference between adjacent points to the observed wavelength in the observed data file. If, for example, a peak was observed at 349.84nm, by subtracting the adjacent wavelength reading from 349.84nm the uncertainty may be determined (let's say, for example, the next sampled point was at 350.22nm - the uncertainty is then plus/minus 0.38nm). For simplicity, use the largest difference (the "+" or "-") whichever is largest (although the difference between the two numbers will be small). If the peak was observed at 349.84nm the actual wavelength would be expressed as 349.84nm +/- 0.38nm. And note that it will NOT be constant across the spectrum so you will have a different uncertainty for each reading as determined from the data (it differs by wavelength due to the dispersion of the grating).

This represents the minimum uncertainty - the actual observed uncertainty is a combination of errors and may be considerably larger than this minimum. This can be determined, experimentally, by finding the actual difference between observed and known wavelength peaks of a known element such as mercury (which is one portion of this experiment).

Now that the uncertainty of measurements made with the spectrometer has been determined (so, for example, if a peak was found at 807.50nm and the uncertainty was determined by comparison with the mercury spectrum to be 0.61nm then the actual wavelength of that peak lies between 806.89nm to 808.11nm), a search can begin for unknown lines. With a pure gas such as helium, most lines can be attributed to known helium lines but some might well be from impurities in the discharge tube. In this case, if a line does not match those for helium, look for wavelengths in air gases and other possible contaminant gases (e.g. nitrogen, oxygen, argon, and possibly hydrogen from dissociated water vapour). Any possible wavelengths should fall within the range of uncertainties found above.

Lab Report

For this experiment, an abbreviated lab report is required (word processed, never hand-written) with the same format as PHTN1300. Answer each question in the form "4a., 4b., 4c. ..." with each new question (#4, #5, etc) beginning on a new page. Do NOT answer an entire question (e.g. question 4) as a single paragraph without identification of sub-parts ('a', 'b', etc). Submit the lab report in a bound folder NOT simply a pile of loose, stapled papers nor a thick binder!

Each student must submit a unique lab report - no portion other than the results must be shared between lab group members.

    Observations and Results

  1. Outline the uncertainty results of the OOI spectrometer as follows: Show actual observed data as a table (with columns for the observed wavelength from the OOI data, known wavelength from the NIST eBook, and difference) for at least four wavelengths of mercury across the spectrum (preferably covering both ends of the spectrum). From this data, determine an uncertainty based on these observations. Report the observed uncertainty (based on comparison of observed wavelengths in the mercury spectrum) which will be the larger of the difference shown here or that based on the wavelength resolution of the OOI, whichever is larger

  2. Calibration Data
    Above: An example of calibration data presented as a table (although not specifically the spectrometer used in this experiment which is likely more accurate). Observed wavelengths are as read from the OOI data text file and known wavelengths must be expressed in as many decimal places as possible (from the NIST tables or eBook).

  3. Determine the components in air showing a complete analysis of the air discharge and assignment of each observed spectral component as follows: For each major spectral line found in air (you must identify at least TEN, the ten largest peaks found), list these in a table showing (a) the observed wavelength shown WITH +/- TOLERANCES, (b) the known wavelength from the NIST eBook and (c) the gas component from which the line originates. When assigning observed lines to a specific source, be reasonable ... was Yttrium present anywhere in the tube? If not, it is not likely it would be seen. You'd expect common gases like oxygen, nitrogen, argon, etc. to be present in an air discharge, not SOLIDS like praseodymium!

    Observations Table With Uncertainties
    Above: An example of observations presented as a table. All observed wavelengths require a tolerance, known wavelengths must be expressed in as many decimal places as possible (from the NIST table), and each line must be assigned to a known species (otherwise, how could you possibly have a "known" wavelength if you don't know the atom is ???).

  4. For the discharge in pure helium, repeat the process above identifing at least ten spectral lines. Logically some will be attributed to helium and some to contaminants. Where a contaminant is found, identify the exact gas from which it originates (e.g. oxygen) by comparison with known lines in the NIST tables. The observations for each gas must be in the form of a table listing observed wavelengths WITH TOLERANCES, the known wavelengths from the NIST eBook, and the gas species of the line (helium, as expected, or the specific gas acting as a contaminant) ... similar to the example provided above.