PHTN1432: Vacuum Systems and Thin Film Technology
Lab #1 - Gas Discharge Tubes (2017W)


Observing The Discharge via a Spectroscope, Photo by Andrew Klapatiuk at Niagara College
A student is seen here observing the output spectrum of a neon discharge. The direct-reading spectroscope used here allows the user to quickly read the wavelength of various emission lines and assess qualitative parameters such as emission line sharpness. A computerized Ocean Optics spectrometer is also used in the analysis of emissions. As well as expected lines from neon, impurity gases (residual gases such as oxygen primarily liberated from the electrodes as they heat) are also revealed by their emission lines. Note the flexible stainless-steel tube connecting the tube to the turbomolecular vacuum system.


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

Preparing the Vacuum System

Discharge Tube The discharge tube will already be connected to the Swagelok connector on the isolation valve for the system. The main valve is kept closed at all times when the system is not in use to keep the RGA under vacuum, leave it closed for now. Configure the system as per the instructions on the COURSE NOTES page and evacuate the discharge tube.

It is IMPERATIVE that you review the operation of the various valves and gauges in the system before attempting to use them!

When evacuating the tube and manifold, the evacuate valve must be opened very slowly to avoid overwhelming the pump (which, in the case of a turbomolecular pump causes the pump to slow with an accompanying "whine").

Pressure is monitored via the dual BARATRON gauges on the manifold. Pump the tube until this gauge 'bottoms-out' at which point the system is below 10-3 torr. Close the evacuate valve to isolate the tube from the pump. Pressure in the tube may now be monitored by the gauges on the manifold and may be controlled with two sets of valves: pressure in the manifold/tube may be increased by using the vent valve to allow small amounts of air into the manifold or the appropriate needle valve on the gas supply bottles to allow a specific gas into the tube. To decrease pressure in the manifold/tube the sampling valve may be used to allow small amounts of gas in the manifold to be leaked out to the turbo pump.

The Experiment

Do NOT touch the RGA nor open unused high-pressure gas bottles (only helium and neon are used in this lab) until required in the next lab. Also, do not open high-pressure (bottle) valves for any longer than required (i.e. to pressurize the regulator) in order to: (i) preserve contents and (ii) prevent contamination. These gases are research or UHP grade and are extremely expensive.

Also, be sure to cycle the tube on/off rather than leave it on continually to ensure the tube does not overheat. It is sealed with high-vacuum WAX which will MELT if overheated. Tube failure (in the form of violent cracking) may result from overheating so be sure the acrylic plastic safety shield is installed at all times.

As per the procedure from PHTN1300, the spectrometer must be calibrated. You will be provided (on the PC connected to the spectrometer or via download from this page) with a captured emission spectrum of a mercury emission with which to calibrate the offset of the spectrometer. The spectrum is optimized for emissions in the visible portion of the spectrum. When analyzing collected data from the experiment later, correct ALL observed wavelengths using this (Hg spectrum) data in the lab. Failure to compensate for the offset of the spectrometer will lead to incorrectly assigned lines (and a big loss of marks). More on calibration follows in this lab.

Evacuate the tube to as low a pressure as the vacuum system will allow (a few minutes of pumping after the gauge reads "zero" will suffice). Investigate the nature of the air discharge first. Close the evacuate valve and increase the manifold (and tube) pressure using the vent needle valve. Increase the pressure slowly and note the effect on the discharge (shape, intensity). Draw diagrams of the discharge in your lab book at various pressures (e.g. minimum, close to minimum, optimal, close to maximum, and maximum) so that this may be explained in the lab report. Note, also, the minimum pressure at which a discharge occurs and the maximum pressure at which discharge occurs (as well as how the discharge appears under each of these conditions e.g. does it fill the entire tube? Is it filamentary? Does the colour of the discharge vary across the length? Answering these questions will help you describe exactly HOW you know the tube pressure is within operational limits simply by observing the discharge). This part of the experiment may be done as many times as required - feel free to increase or decrease the pressure to repeat observations as required.

Now, pump the tube back down to a pressure of 2 torr where the air discharge is constant, bright, and fills the tube (i.e. approximately an "optimal" pressure). Observe the spectral emissions from the tube using the Ocean Optics (OO) Spectrometer so that the sources of individual lines may be identified in the lab report by the gas component from which they originate. Set the integration constant of the OO spectrometer to vary the intensity of observed peaks so that peaks in the visible portion of the spectrum are optimized (i.e. they are not so large as to 'saturate' the spectrometer nor are they too small to discern from noise). To eliminate "bouncing" of the displayed peaks, simply set the 'average' to a larger number. Be sure to capture the output of the spectrometer as a file (wavelength versus intensity) from the FILE menu for later analysis in the lab report. Optionally, a graphical output may be captured (this may be captured from the OO software or preferably generated later using a spreadsheet).. Remember to correct all observed wavelengths later.

For the air discharge, you will now be able to identify the spectral components (i.e. which lines originate form oxygen, which from nitrogen, etc) which demonstrates the basic technique of OES (Optical Emission Spectroscopy). For a discharge in helium, and then neon, gases we will now identify the source of emission lines and attempt to identify those originating from contaminants.

Now, evacuate the tube fully (close the vent valve first, of course, and open the evacuate valve slowly to avoid 'dumping' the pump) and admit a small amount of helium into the manifold to a pressure of 2.0 torr using the helium supply needle valve. Using the Ocean Optics spectrometer capture an emission spectrum to disk (be sure to set the integration time first) - later, you will analyze the data and identify the spectral emission lines from helium and identify any _NOT_ from helium. Since this is a first-fill one might well expect some residual gases to be present in the tube. Now evacuate once again. Overfill with helium to the maximum pressure which will sustain a discharge (note that pressure). Lower the pressure slowly (use the sampling valve to slowly pump-out excess helium) and observe the minimum pressure at which the discharge in helium can be sustained.

Open the vent valve and allow the tube to pressurize with air to atmospheric pressure (760 torr). Close the vent valve and evacuate the tube to the lowest pressure possible and repeat the above experiment with neon gas, observing and capturing the emission spectrum at 2 torr then pumping out the tube and determining the minimum and maximum pressure at which the discharge in neon can be sustained.

When done, evacuate the tube to low pressure. Close all valves, and shut down the vacuum system (the mechanical pump will automatically turn off when the turbomolecular pump has slowed - this takes up to ten minutes).

In Summary ...

You need observations of the following:

Calibrating the Spectrometer ...

Download the Latest Mercury Spectrum (2015) for the setup. Use neutral atom lines only to compare.

In order to calibrate the spectrometer properly, an offset must be determined. This offset is a function of wavelength. In order to determine the relationship between λTRUE and λOBSERVED, we will use linear regression (which you used in PHTN1300 on several labs). On a spreadsheet, make two columns of data: the "x" column represents observed wavelengths and the "y" column represents corresponding known wavelengths from the NIST tables (search for these peaks manually - the known and observed wavelengths will generally be only a few nanometers apart so that, for example, the known large mercury peak at 404.6565nm will be found in the OOI data file around 406.8nm). You must use at least five visible mercury lines for the calibration. GRAPH the points with the x-axis being the observed wavelength from the OOI spectrometer and the y-axis being the known wavelength from the NIST tables. Next, add a TRENDLINE and display the equation of this line on the graph as well as R-squared (make sure you display the equation with at least six digits of accuracy). If R-squared is not at least 0.999 the relationship is not 100% linear likely due to an incorrect choice of wavelength ... correct this by reviewing the data and try again.

Remember, ALL graphs have a title and axis labels (with units).

As well (and you'll need to do this anyway to determine tolerance) use linear regression (on Excel 2013, select the "Data Analysis" from the "Data" menu) to determine a linear relationship between the two columns, the result being the "Intercept coefficient" and the "X variable coefficient". The ACTUAL wavelength may now be determined using the standard line equation "y=mx+b" where "y" is the actual/corrected wavelength, "m" is the "X variable coefficient", and "b" is the "Intercept coefficient". Both the trendline equation and that found using the linear regression analysis tool must agree.

An equation will be formed transforming the observed wavelength into the true wavelength. To verify the relationship make a third column called "Actual wavelength" and calculate this using your new equation ... these calculated wavelengths should be very close, numerically, to the known wavelengths.

You will be required to outline the entire calibration procedure in the prelab as well as provide the aforementioned table of wavelengths.

Uncertainty Calculations Now, determine the uncertainty for each observed peak (the "plus/minus" value) 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 it would first be calibrated with respect to wavelength as per the above process. Next, 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). We're not done yet, though, since THIS uncertainty only describes part of the picture ...

The above uncertainty represents that of the acquisition system, it represents the possibility of an error between two adjacent sensor pixels in the OOI (much in the way a multimeter reading is plus or minus the last digit so that when displaying "12.00" the actual voltage is between 11.99 and 12.01), but we must also compensate for the uncertainty of the conversion equation in terms of how well the data from the mercury spectrum fits the equation we have determined. To do this, simply add the uncertainty determined above to two-times the standard error as determined using the linear regression analysis (this will be found in the SUMMARY OUTPUT under "Regression Statistics"). Statistically speaking, 95% of all points will fall within a range defined by a tolerance of +/- two standard error values. This standard error is the same for all points while the "pixel error" will be variable over a small range

Determination of uncertainty is important, so be sure to do it early (the night before it is due is too late) and ASK if you need help interpreting your data.

All observed wavelength peaks MUST BE REPORTED IN THIS MANNER, i.e. as a Calibrated wavelength and an uncertainty. This will help with identification of unknown peaks. For the calibration procedure, present the wavelength of all five (or more) peaks observed from the OOI spectrometer as a corrected wavelength with a tolerance ... these should be very close to the known NIST wavelengths and the NIST wavelengths should easily fall within the tolerance range calculated (if not, you likely chose the wavelength of one or more peaks incorrectly at the start - try again).

In the report (not the prelab), present peak identification data for each discharge (ten required in total) as a table. The table must have headings for wavelength (recalibrated, and WITH a +/- UNCERTAINTY figure) of each major spectral emission peak seen in the discharge as well as a probable source gas and the known wavelength of that known emission line (for example, if an emission was seen at 557.1nm, calibrated with respect to the offset for that specific wavelength, the table might list the observed line as '557.1nm +/-0.34nm', the probable gas as 'krypton', and the known wavelength of that emission line as '557.02nm'). One expects the known line to fall within the uncertainty range presented (assuming you calibrated the spectrometer wavelengths correctly and specifically as a function of wavelength) ... in the example given the range is 556.76nm to 557.44nm so the known line fall nicely into that range (Remember that 95% of the data will fall into the uncertainty range determined so a few points might be outside this range).

For the air discharge, one would expect all lines to originate from known gases in air (including some trace gases which comprise less than 1% of the total composition of air), for helium and neon, most lines would indeed originate from those gases (list the known wavelengths, of course) but some might be from impurity gases in the tube - determine what these are by their wavelengths.

Research (when writing the report)

Lab Report

This term there will be a special emphasis on PROFESSIONALISM in lab reports. You'll understand the tone of the report by reading below. AMATEUR HOUR is over ... cut-and-paste lab reports with irrelevant details and unedited text will be severly penalized.

For this lab, you are to prepare a test report and may assume the following:

You are working as an employee for the Polaris Spectra corporation and are in charge of processing discharge tubes for use as wavelength references. Specifically, the company endeavours to produce both helium and neon discharge tubes. Several questions now exist, namely (i) What pressure should be used and why not use something arbitrary, and (ii) How can OES be used to identify contaminants in the "pure" gas tubes with a demonstration of how this works.

Using an air discharge, you will demonstrate how the technique of OES works including details such as spectrometer calibration. Air components will be identified explicitly in the discharge. Next, using helium and neon, the OES technique will be used to identify contaminants in the tube showing how the technique could be used in the production environment for quality control.

Prepare a formal report (word processed, never hand-written) as follows and submit the lab report bound in a folder NOT simply a pile of loose, stapled papers not in a large binder. All work is to be done individually including interpretation & analysis of data and background & procedure sections. Any evidence of duplication (e.g. duplication of text in a section) will be submitted to the dean's office as suspected plagiarism. Plagiarized reports receive a mark of zero and all students involved will reprimanded in accordance with college policy.

The title of this test report, then, is "Feasibility of OES usage for contamination control in discharge sources" and has the following elements:


Similar to an abstract, this is one or two (at most) paragraphs outlining what the intent of the report is, what is trying to be accomplished by the technique, the very basics of the technique (without a huge explanation).

Remember your audience here: managers at the corporation who understand the basic process of making a discharge tube (since it it the company's line of business) but aren't completely aware of how OES can help them accomplish quality control nor how a newer spectrometer like an "OOI" can work for this!

What are we trying to accomplish? Discuss how OES can be used to determine contamination in a discharge tube and, better yet, WHAT those contaminant gases in a "pure" gas discharge tube usually are (since that will help them solve the problem).

Basics of the technique? Discuss what OES is, the proof-of-concept using an air discharge to identify the known air gas components in the tube, how it can be used with a "pure" gas discharge to identify contaminants (i.e. how do you tell a helium line from a rogue oxygen line).

READ THIS THROUGH after you write it and make sure it makes sense! Would a reader unfamiliar with the technique be able to follow what you did and how this technique could be applied to a production environment?

Be careful NOT to put too much detail in this section: no one reading a summary needs, or wants, to know that a SwagelokTM SS-4BK valve was used in the system !!! Keep it generic and refer to "a vacuum system employing a turbomolecular pump" instead of a verbose (and useless) description. The purpose of this section is NOT to outline everything, just to summarize the IMPORTANT points.


Outline how the vacuum system was configured, how it was used to process the discharge tubes, and how observations were acquired for this feasibility study.

Start with a page summarizing specific theory and principles used in this lab. In this case the (i) basic configuration of the vacuum system used for this feasibility study, and (ii) the operating principles of a turbomolecular pump

A diagram of the system components is required here as is a diagram of how a turbo pump operates (since a picture is worth 1000 words and no one wants to read a 1000 word description on how this pump works). Remember, though, that a picture without an explanation is useless. If you include a figure, give it a figure number and REFER to it in the text that follows (that is the acid-test as to whether or not it is a useful diagram). Usually one decent figure paired with one decent paragraph is enough to explain a concept as applied in a lab.

Now, outline the very basic steps needed to perform the task. This should include enough detail to duplicate the study and obtain your results. Here you would NEVER include a cut-and-paste of this outline, instead include a PROPERLY EDITED description of what you actually did. Avoid irrelevant details such as "I walked into the lab and took off my coat ...", "Our group consisted of three people, one turned the valves while the other picked his nose", or "We turned on the power switch as instructed by Mr. Csele" (I'm not kidding, I got that one year). Verbal diarrhea such as "We opened the needle valve by turning clockwise ..." is completely useless in this case - simply point out that "this valve was opened to fill the tube" or "this valve was opened to evacuate the tube" refering to the system diagram included already in this part of the report. Something like: "Neon was admitted into the evacuated manifold using a needle valve until the pressure in the manifold was increased to xx.x torr as read on the Baratron gauge" is sufficient and preferred. It is likely that anyone in the world trying to reproduce this experiment would have a similar but not identical system so details _too_ specific to this system are unnecessary. Another trap NOT to fall into is copying instructions from the text verbatim (e.g. pp.198) where they do NOT apply to this specific system. Discussions of diffusion pumps do not belong here since they aren't used in this lab.


Unlike a procedure, this is more like a practical example of how the technique works. In our case, as applied to the air discharge since the purpose of the air discharge was to show the feasibility of detecting gases in a discharge using spectrum analysis. It provides a detailed example of the technique but does not describe details such as which valves were turned when, etc.

Required explanations here include:
(i) an explanation in words of how calibration was accomplished and uncertainty was determined. Show actual observed data here used to calculate the offset (observed wavelength from the OOI data, known wavelength from the NIST eBook, and difference) for at least three wavelengths: use a table to present this data, similar to that used in lab #1 of PHTN1300 last term. Next, include a graph showing offset vs. wavelength and the calibration equation. EXPLAIN, in a paragraph, if the offset is constant or not, and finally summarize the offset used throughout the rest of the analysis (the equation or the constant).

Calibration Data
Above: An example of calibration data presented as a table. Observed wavelengths are as read from the OOI data and known wavelengths must be expressed in as many decimal places as possible (from the NIST table). The difference is shown as a negative constant here: adding thast constant to the observed wavelength will correct it to the actual (known) wavelength). To analyze this data further, a graph is required and finally a mathematical equation relating ACTUAL wavelength to OBSERVED wavelength.

(ii) application to the determination of the components in air showing a complete analysis of the air discharge and assignment of each spectral component. For each major spectral line found in air (you must identify at least TEN), list these in a table showing (a) the observed wavelength, corrected for the offset and shown WITH +/- TOLERANCES, (b) the known wavelength from the NIST eBook and (c) the gas component from which the line originates. As for 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!

Be sure to show ONE complete example calculation which describes how you found the ACTUAL/TRUE wavelength from the observed wavelength.

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 possible have a "known" wavelength if you don't know the wavelength ???).

(iii) In addition, discuss the required pressure for a discharge tube (both minimum, maximum, and optimal) - by observing the discharge, what is considered "too low" and what is considered "too high" and justify WHY (i.e does intensity decrease? stability poor?) in both cases. This must be explained in words in a PARAGRAPH, again detailing WHY you consider a pressure to be "too high" or "too low" and finally what the approximate "optimal" range is for an air discharge tube.


A detailed example of how the technique was applied to determination of contaminants in both a "pure" helium and neon discharge tube. Give specific examples and for each gas identify at least ten spectral lines attributing some to the gas 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 TRUE observed wavelengths (i.e. corrected), WITH TOLERANCES, the known wavelengths from the NIST eBook, and the gas species of the line (helium, or neon, as expected or the specific gas acting as a contaminant) ... similar to the example provided above.

REFERENCES must not be on a separate page but rather as a footnote on each page. If you used a diagram from a reference in the background section quote the source as a footnote right on that page. Similarly if you quoted a pressure like "discs are known to occur at xxx.x torr1" or that the red line of krypton gas has a known wavelength of 647.1nm, use a footnote to show the source of this figure. Marks will be deducted for numbers and figures without proper references.