Introduction
In an earlier post, I used a sample of caesium chloride (CsCl) and a small Tesla coil to generate F centres. F centres are defects in the crystal lattice where, for caesium chloride, chlorine ions are ejected from the lattice and replaced by electrons. The electrons are able to absorb light in the visible spectrum, meaning that the crystals become coloured. More details of this effect are described in a previous video demonstration.
I went on to record a spectrum of the blue-white luminescent glow that immediately appeared when the Tesla coil’s electric field was applied. The spectrum of the blue glow is reproduced here for convenience:
I admitted at the time the post was published that I was having difficulty assigning these emission lines to the likely chemical species present in this glow discharge. So I decided to do some more research and digging, and below are the results of this search:
Emission Line Identification
From their relatively narrow profiles, the lines are almost certainly atomic emission lines, mostly from caesium but also chlorine. They are not due to any luminescence from the F-centres themselves. F-centre emission is typically broad and smooth, not sharply peaked as in the above spectrum.
A provisional identification, which is based on NIST atomic data and plasma emission literature, is provided in the following table with appropriate comments on uncertainty and any line shifts:
| λ (nm) | Probable Assignment | Comments |
|---|---|---|
|
390 |
Weak Cl(I) line at 390.3 nm |
Could also be N2. Common in discharges and possible if tube was not fully evacuated |
|
411.4 |
Cl (II) |
Singly excited Cl atoms Cl (II) centred at 413.25 nm |
|
449.2 |
Cs (II) |
Singly excited Cs atoms Cs (II) officially at 450.16 and 452.67 nm. So within experimental uncertainty. |
|
481.6 |
Cs (I) |
Several neutral Cs lines in this region. But could also be Cl (II) which has a line at 481.01 nm |
|
518.5 |
Cs (II) |
Fairly strong line at 520.9 nm. Spectrometer calibration may be poor in this region |
|
560.2 |
Cs (II) |
Another strong Cs line at 556.3 nm. Wavelength calibration may be poor here. |
|
606.2 |
Cs (II) & Cs (I) |
Several Cs (I) and Cs (II) lines in this region from 593 to 612 nm. Likely to be Cs (II) ; strong in discharge tubes |
|
651.0 |
Cs (II) |
Strong line at 653.6 nm |
|
674.0 |
Cs (II) |
Strong line at 672.4 nm |
To summarize, the dominant atomic emission lines in the discharge originate principally from singly excited caesium Cs(II), some neutral caesium Cs(I) and a possible small contribution from chlorine Cl(I).
None of this emission comes from F-centres. Any F-centre luminescence would appear as a broad, featureless band across several tens of nm. F-centres are clearly formed in the crystal lattice from the visible colour change, but detecting any luminescence from the F-centres themselves is not possible under these excitation conditions.
The most common research method for reproducibly generating F-centres in metal halide lattices is by exciting a sample of the salt with 10-100 keV X-rays. Photons in this energy range will eject halide ions from lattice sites in a controllable way and produce F-centres. These could then be examined by UV irradiation to look for any direct photoluminescence. Needless to say, specialized X-ray equipment of this nature is not available to me in my modest home lab!
Let’s now take a look at other examples of metal salts excited in a strong electric field.
Potassium Iodide
In the video below is a sample of potassium iodide, again dried and sealed under vacuum. A search was conducted for F-centre using both a tesla coil and a small high-voltage power supply, but in this case no F-centres appeared and no colour change.
However, the high voltage did immediately generate light emission above the crystals as shown in this very short video:
As before, I recorded the visible spectrum of the glow discharge which is reproduced here:
The strong emission line at 657.4 nm is suspiciously close to the hydrogen H-alpha line at 656.3 nm. This line can, and will, appear if trace amounts of water (hence hydrogen) are present in the sample. So the sample may not have been completely dry when it was sealed in the tube.
The remaining lines are due to neutral iodine atoms I(I) or singly ionized iodine I(II).
Unlike the example of caesium chloride above, where we saw a clear colour change indicative of F-centres, there is no observed colour change with this sample and F-centres are absent under these excitation conditions. Other forms of excitation (X- and gamma-ray irradiation) are presumably needed for this sample.
Tin (IV) Iodide
I wasn’t really looking at F-centres with this sample; I just wanted to see if there was any iodine-based glow discharge similar to the case with potassium iodide.
Tin (IV) iodide, also known as stannic iodide, is a red-orange crystalline solid as seen from the small sample in the tube here.
The supplier stated that the sample was pure tin iodide containing no “starter” gases, which are sometimes added to initiate gas discharge tubes of this type. So the tube was immersed in boiling water for a few minutes to increase the vapour pressure in the tube and then the same high voltage power supply as used above was applied. Here is a video (without audio this time) of the demonstration:
The purple/blue glow discharge is very similar to that seen with the potassium iodide example earlier and the spectrum of the discharge is almost the same:
There is still the H-alpha line strongly present, again indicating that the sample was not completely dry. All remaining emission lines originate from atomic iodine in different ionization states as before. The grouping from 580 – 640 nm is characteristic of molecular iodine I2 vibronic structure, similar to the absorption spectrum of I2 and its laser induced fluorescence spectrum reported in other posts elsewhere on this site. And the multiplet structure around 480 nm is where iodine has several transitions.
The strongest emission lines for tin (Sn) are in the UV and near UV/blue. Emission from Sn(I) and Sn(II) in the visible are relatively weak. If they are present in the above spectrum, they are likely hidden in the noise or masked by I atomic emission lines.
Superposing the discharge spectra from both potassium and tin samples shows virtually identical line profiles, confirming the emission lines to be largely due to I(I) and I(II).
Final Words...
A future article may look at other techniques to generate F-centres in metal halide and other inorganic salts.
Steve
