Cathodoluminescence
During the late 19th and early 20th centuries, partially evacuated tubes and bulbs filled with various and sundry substances were all the rage. Thanks to researchers, such as Geissler, Hittorf, Plücker, and Crookes (see Radioluminescence on this Web site for more about William Crookes), people in their parlors could experiment with liquids, gases, and solids (e.g., gems and minerals) to see how they reacted when excited by high voltage discharges. The luminescence of minerals, in particular, presaged the advent of the CRT tube that lies at the heart of television sets and oscilloscopes. Recently, LCDs and LEDs have replaced the phosphor-coated glass screens. Of course, those tubes with glowing gases were the forerunners of modern day neon lighting and fluorescent lamps.
The luminescence produced when electrons and negative ions bombard substances is called cathodoluminescence (CL). In a cold-cathode discharge, the type of discharge used in the early experimental apparatus, electrons were accompanied by both positive and negative ions. Electroluminescence would seem a better all-inclusive term, and it certainly rolls off the tongue easily, but it is firmly associated with other forms of lighting.
The mineral of interest is usually positioned at the anode, and a high negative voltage is applied to the cathode. Not having a Crookes Minerals Tube—as the device is often called—I rigged up a bell jar containing a couple of large, optical calcite crystals. (They are referred to as optical calcite due to their transparency.) The bell jar was connected to a two-stage mechanical pump that could evacuate the bell jar down to 2 millitorr. For the photos seen below, I applied high voltage when the vacuum gauge read around 30 millitorr. The photo at lower left shows the crystals glowing under the cold-cathode discharge. When I turned off the high voltage, the crystals continued to glow for some minutes. The 30-second exposure at lower right shows their dramatic phosphorescence. Had I placed a yellow filter over the camera lens, the blue ambient plasma glow seen in the left-hand photo would have been subtracted out. The crystals would then have appeared similar in color to the phosphorescence seen at right.
I checked the calcite crystals under both shortwave and long-wave UV. They displayed a rather weak fluorescence under SW and no discernible glow under LW. I have a third optical calcite crystal, but it exhibited neither UV fluorescence nor cathodoluminescence. Often, this is what you find when you experiment with CL and potentially luminescing substances. The crystals came from different sources, and they undoubtedly contain different trace elements that are responsible for the presence or apparent absence of luminescence.
UV fluorescence begins with the excitation of electrons in a molecule by UV photons. The excited electrons soon return to the ground state and may release their stored energy by emitting characteristic photons. It is also possible for the energy to be dissipated as heat or by the excited molecule entering the triplet state that leads to phosphorescence. The emitted wavelengths of light will usually be longer than the excitation wavelengths (Stokes Law).
For cathodoluminescence, concepts of semiconductors are introduced. When insulating and semiconducting materials are excited by bombarding electrons, electrons in the material's valence band are promoted across the band gap to the conduction band. Sometimes an exciton pair is formed. Other times, the free electron in the conduction band and the generated hole in the valence band are free to diffuse. When electron and hole rejoin, they may encounter recombination centers called charge traps. The particular recombination pathway followed will determine the observed cathodoluminescence. These charge traps are a function of trace ions and structural defects. In addition, when bombarding electrons strike a material and are stopped, x-rays are generated that produce fluorescence. Another pathway may lead to the production of phonons dissipating energy nonproductively as heat. Questions arise as to why some substances look much the same under UV and CL and why others differ so markedly. An example of this will be seen below in the discussion of aragonite.
For quite a while, I had wanted to construct an apparatus to view CL. There are a number of approaches to producing and observing CL in the laboratory, including the use of suitable light sensors and filters located within a scanning electron microscope (SEM). When a beam of electrons strikes a specimen within the SEM, in addition to secondary and back-scattered electrons that form the image, x-rays and light are often produced. Nice, but I ruled out that approach real fast. Two other approaches involve designing suitable high-voltage vacuum stages that permit either cold-cathode bombardment (relatively easy) or hot-filament electron bombardment (much more difficult because of the high vacuum needed). I opted for cold-cathode bombardment.
A major advantage of optical CL is the visibility of colors that can be readily photographed. SEMs, by their nature, do not produce colors. That is why the light sensors/ filters technique is necessary. For most of my work, low magnification suffices as entire specimens can be examined. This falls into the realm of macro CL (MCL). This is fortunate since the usual optical micro cathodoluminescence techniques require rather expensive equipment. (See PowerPoint presentation by Marshall in References.)
Macro CL
To produce a cold-cathode source for my macro CL device, I took the following approach:
A standard, KF-40 5" full nipple was modified by the manufacturer by welding a KF-16 half-nipple angled at 45°. This angled tube holds a KF-16 mounted high-voltage feedthrough. The top of the KF-40 tube has a clear glass viewport bracketed on. Over time, the window will cloud up with contaminants, and it is vital that it be easily removed for cleaning.
A metal sleeve is positioned within the tube to act as a specimen support. I machined the sleeve to produce a friction fit within the tube. The top of the sleeve has a wire mesh (square openings) serving to hold specimens and to allow nearly unfettered vacuum pumping. The top of the metal tube along with any specimen sits below the elbow of the KF-16 half-nipple. It occupies the (inside) lower third of the KF-40 fitting. The bottom of the KF-40 fitting is hooked up to a vacuum pump through a butterfly valve, etc.
Stable clamping is used to support the apparatus so that an integrating color CCD camera can be used to study and photograph materials undergoing CL. The camera is equipped with a macro-zoom lens functioning as a low-magnification microscope with an 8" working distance. Magnification in my unit with the Mintron CCD ranges from approximately 15X to 100X. The photo at right shows the complete unit. Not shown is a cone-shaped piece of black fabric positioned between the zoom lens and the glass viewport to block off room light.
The apparatus uses an empty forepump filter as a vacuum ballast. This enables quicker evacuation of the MCL chamber when the butterfly valve is opened. Also, the extra vacuum storage allows anchoring the MCL chamber when loading a new specimen without having to have the forepump operating. A quick flip of the butterfly valve, and the chamber is vacuum locked to the rest of the system. It can then be easily secured with a KF clamp.
The system is adequately energized by a 0-15 kVDC @ 1 milliamp power supply controlled by a Variac. There is also a high-voltage, current-limiting resistor in the circuit to protect the power supply. The high-voltage lead and power supply are shown close to the macroscope for ease in photographing the setup. Normally, that lead is kept well away from any possible user contact. The power supply's ground lead is connected to the macroscope. A plexiglass shield for high-voltage protection is advisable.
WARNING: IN ADDITION TO THE PRESENCE OF HAZARDOUS HIGH-VOLTAGES, X-RAYS ARE PRODUCED IN THIS APPARATUS. DUE TO THE VOLTAGES USED IN THEIR PRODUCTION, SIGNIFICANT ABSORPTION BY HUMAN TISSUES, SUCH AS SKIN AND EYES, SHOULD BE ANTICIPATED!
Some specifics
My vacuum pump is a Varian SD-90, two-stage rotary vane, 3.2 cfm unit. The manufacturer states that the ultimate pressure is 2x10-3 mbar. You'll notice that there is no vacuum gauge in my setup. It is important that you know your pump and what it can do. After starting the pump with the column butterfly valve closed, I wait 20 minutes before the first MCL experiment. This gives me time to change the specimen, turn on the associated video equipment, etc. With the pump running continuously, changing samples is fast with only 5 minutes needed to pull an adequate vacuum for the MCL. If you choose to gauge the system, look for 10-20 millitorr as a reasonable range. Too high a vacuum will prevent the electron beam from being generated. If you encounter this problem, you may need a controllable leak in your system. Too low a vacuum, and there will be a glow discharge but no CL. I just look for a decent image on the monitor and open or close the butterfly valve as needed. This requires a little experimentation, but soon you will become one with the macroscope.
The camera is a Mintron integrating CCD model MTV-6268. I leave off the automatic gain control (AGC) unless luminescence is so minimal that I can't get a recordable image. In the AGC mode, considerable image noise is present.
The KF-16 arm coming off the KF-40 has a pyrex sleeve running internally from the feedthrough just up to the oval hole in the main tube junction. I cut off a length of pyrex tubing and wrapped it with one or two turns of teflon tape to produce a friction fit. The pyrex isolates the high-voltage electron beam plasma from the stainless steel KF-40 housing. The electron beam is unfocussed, so there is no need for fancy electrodes in the electron source. That white collar seen wrapped around the KF-16 arm is a split delrin tube intended to support a solenoidal coil for magnetic experiments.
Finally, PC image capturing is performed by a Zarbeco Videolink USB interface and its associated software.
Studying minerals using MCL
Ruby was the first gem synthesized. Rubies can be produced as boules by oxy-hydrogen flame techniques, such as the Verneuil process. Made from corundum (crystallized aluminum oxide) along with small amounts of chromium oxide, synthetic ruby is used for lasers, masers, bearings, and gems. The fragment shown here was broken off a boule.
Sodalite is a sodium aluminum silicate chloride valued as an ornamental gem. Colors range from royal blue, green, yellow, white, to pink. Its CL is mottled.
Hydrozincite is a zinc carbonate hydroxide and appears blue under UV. The mineral is often associated with limonite, as was this specimen. Only certain areas luminesce strongly, and that is probably due to variations in composition.
Pectolite is a sodium calcium silicate hydroxide. It is often found with radiating masses, as seen here. It is fluorescent under UV.
Calcite is an extremely common form of calcium carbonate. It forms limestone that, in turn, can be metamorphosed into marble. It may exhibit an orange or red fluorescence under both UV and CL, usually due to the presence of manganese.
Fluorite, also called fluorspar, is calcium fluoride. It lends its name to the term fluorescence. Due to fluorite's high transmission of UV light, it can be made into superb lenses for cameras, telescopes, and microscope objectives.
Aragonite is a form of calcium carbonate. Aragonite's chemical properties are virtually identical to calcite, but its crystal structure is different. Accordingly, it's called a polymorph of calcite (and calcium carbonate). Although this specimen of aragonite exhibits a pervasive yellow-white UV fluorescence, only certain areas luminesce under CL. Those areas may be chemical modifications or just contaminants. The multicolored speckles in the background are tiny fragments from minerals that preceded the aragonite onto the mesh stage.
Spodumene is a lithium aluminum silicate. Its gemmy forms are kunzite and hiddenite. Spodumene often exhibits a yellow-orange fluorescence under UV and under CL. Finely ground spodumene is sold by dealers in clay for glazing and flameware. It seems likely that powdered spodumene would make a very useful screen for electron beam detection. The finely ground material I received glows strongly under CL but not at all under UV.
Franklinite/willemite/calcite is a favored association for UV fluorescent mineral collectors. Obtained from Franklin, New Jersey, under SW UV the trio fluoresce a beautiful green (willemite) and red (calcite). Franklinite is zinc iron manganese oxide and is an ore of zinc and manganese. The iron does a superb job of quenching any hope for that mineral fluorescing. Willemite is zinc silicate. Calcite with traces of manganese fluoresces orange to red. The non-luminous Franklinite nicely sets off the two fluorescent species.
Under CL, the calcite's red color in this specimen is very weak. To make it visible, I increased the electron energy (the voltage). As a result, some willemite regions were overexposed. The inset shows a lower energy bombardment recorded at a higher magnification of the butterfly-shaped, washed-out region adjacent to it. Zinc silicate is an excellent phosphor for CRT tubes.
Looking at gems with MCL
Pink ruby and pink sapphire are both composed of corundum mixed with a little chromium. They are considered to be the same gem but are not as red as a true ruby, which has more chromium. The specimens seen here are from two different locales and differ in their observed CL.
Brown zircon is a gemstone composed of zirconium silicate. Unlike the brown zircon seen in a later photo, two of the three gems seen here are quite blue.
Blue sapphire is corundum with small amounts of titanium and iron. As with the Franklinite discussed earlier, the iron quenches any CL. What is seen in the photo are the gems reflecting the bluish plasma produced in the MCL chamber. Note the graininess in the picture. I set the CCD camera into its AGC mode. The camera dutifully delivered a visible image complete with noise.
Due to the iron quenching, the three blue sapphire rough stones appear like chunks of rock devoid of both life and luminescence. The exception being some spots along the edge of the topmost sapphire and a few isolated speckles here and there.
Locating areas of interest with MCL
Identified in 1781 by the great Carl Wilhelm Scheele, Scheelite is an important ore of tungsten. The fluorescence of Scheelite under SW UV has long been key in helping locate the mineral. The glowing blue areas seen at left under the CL macroscope are mostly calcium tungstate. When larger amounts of molybdenum (as calcium molybdate) are mixed in with the calcium tungstate, the blue color shifts first to the green and then to the yellow.
The two photos are of the same specimen. After locating regions of calcium tungstate, I zoomed in for the lower photo to enhance detail. Both pictures were taken with ambient room light that was permitted to enter the chamber. This fill light illuminates the entire mineral specimen and serves to pinpoint those regions exhibiting the telltale luminescence.
I didn't run any spectra on the luminescent regions, but I suspect that in the absence of spurious ambient lighting any recorded spectra should stand out against a noiseless background. The addition of optical spectroscopy can only enhance the use and application of MCL. To do this, the stereo trinocular microscope arrangement mentioned earlier may be needed.
Scheelite is one of those minerals that appear much the same under UV as under MCL. Moreover, I only had to raise the voltage to ~ 3000 VDC in order to get very strong luminescence. Not surprisingly, artificial calcium tungstate-coated screens are used as image enhancers in x-ray film cassettes. The compound boasts a 5% efficiency in converting x-ray photons into the visible light photons capable of exposing sheet film. It is very efficient with electrons as well, which explains its use as a phosphor in electron beam tubes.
A zircon by any other name may also look different under MCL
Chemically, all zircons are zirconium silicate with varying amounts of other ions, including hafnium and yttrium. Zircon is mined in large part for its zirconium and hafnium content. Valuable gems in their own right, some zircons are used as diamond simulants. At left are samples of both brown and blue zircons. The specimens clearly luminesce in vastly different regions of the optical spectrum. The blue zircons look blue, and one of the brown zircons that I carefully selected to prove the point looks yellowish-beige.
Some zircons are radioactive and contain significant amounts of thorium and, possibly, uranium as well. These so-called metamictic zircons often develop a fuzzy internal appearance as the crystal structure breaks down, forming an amorphous material. Metamictic zircons, as well as those with unfavorable coloration, are often heat treated to make them marketable. The zircons seen here are not radioactive.
Summing up
Cathodoluminescence (micro and macro) can be of great value in the geological sciences for studying mineral provenance, structural differences, composition of crystals seen in thin section, differences due to meteoritic impact, etc. The technique is actively used in forensics as well. Numerous other fields, both in science and the arts, likely would benefit from applying CL.
In many applications, standard (i.e., micro) cathodoluminescence techniques are unquestionably required, and the expense is fully warranted. But for other work, especially where specimens are too large for chambers that fit on a microscope stage, macro CL may be sufficient. That has been the case for me—so far.
Researchers involved with materials science, fluorescence, phosphors, the study of small fossils, or evaluating semiconductors for LED use could benefit from MCL. Don't limit yourself to the visible spectrum. Be sure to examine the resulting cathodoluminescence in the UV and NIR too.
Concluding thoughts
It's quite straightforward to set up an MCL system geared specifically for your work. Those 19th century explorers back in their Victorian parlors would wholeheartedly have approved, while at the same time envying us our newer technologies. I hope that the experiments illustrated above demonstrate just how easy it is to use macro cathodoluminescence to find your own diamond in the rough!
Up next, we examine the use of sparks and lasers for spectroscopy.
