I've posted about this once in the past, I think in an Instructables comment, though I
can't find it again with Google. But I've never posted about it here, and it's been on my
list for two years now.
I shall begin by describing a conventional fluorescent lamp , and its failure modes.
consists of a glass tube filled with mixture of mercury vapor and argon gas (or other
noble gas) at low pressure, with the inside surface coated with phosphors, and a
filament at each end. The filaments are coated with an emission mix to promote
thermionic emission. After heating them, an electric arc is struck between the
filaments, through the mercury vapor. This causes the mercury vapor to emit
ultraviolet light, primarily at 253.7 and 185 nm. This ultraviolet light is absorbed by the
phosphors, which then re-emit that energy as visible light (usually at a mix of
wavelengths designed to give the appearance of white light).
This conventional fluorescent lamp can fail in a number of ways. The filaments can
sputter their emission mix coating away. They can also burn out like those of
incandescent bulbs. The phosphor and glass can adsorb the mercury vapor (which both
reduces the available mercury, the main problem, and blocks some UV from getting to
the phosphor). The phosphor can wear out through any of several phosphor
degradation processes . Also, the ballast can fail, but I am considering that out of
scope for this idea.
The first two problems, those of the filaments, have already been solved by the
electrodeless fluorescent lamp . This uses a sealed induction coil to couple energy
into the mercury vapor. Therefore, it has no electric parts exposed to the arc, and no
resistive heating is used either. Wikipedia also claims uncitedly that it has another
advantage, namely that you can use higher-efficiency "light-generating substances"
(vapor? phosphors? IDK) because they don't have to be chemically compatible with the
My invention is an improved version of the electrodeless fluorescent lamp. It solves the
remaining problems: adsorption of the mercury onto the glass and phosphor, and
First, the phosphor is encased inside the glass. This keeps the mercury from adsorbing
onto it. Also, with proper chemical design, it should be able to prevent phosphor
degradation. The oxidation process is blocked by the phosphor being encased in glass.
(The outer glass must be a non-oxygen-permeable one, obviously.) The crystal lattice
degradation and activator migration processes could be blocked by a close
encapsulation by the glassmultiple alternating molecular-thickness phosphor layers
and thicker glass layers might work. This will depend on the crystal lattice of the glass,
and different phosphors have different degradation processes, so the phosphor mix and
glass chemistry must all be designed holistically.
Obviously, the layer of glass between the phosphor and the mercury vapor must be UV-
transparent. There is a well-known UV-transparent glass, used in blacklights, called
Wood's glass.  However, Wood's glass gets less transparent to UV the more UV you
expose it to, meaning it would make the lamp mortal. Therefore, some candidate
materials are fused quartz (silica) , sapphire , alumina , and aluminium
oxynitride . With these, there should be enough flexibility to engineer a way to
incorporate phosphors into their crystal lattices. As well, these materials are quite hard
and strong, resulting in durability of the envelope. (And no, I have no idea how to
make a bulb or tube out of sapphire.)
The one remaining failure mode (as far as I know yet) is adsorption of the mercury
vapor onto the glass. This is solved by a simple bakeout process, performed whenever
the ballast detects low mercury vapor level (difficult to start) or at every startup. This
consists of heating the glass, using a radiant heater at the center of the tube/bulb
(sharing space with the induction coil if it's in the center). The radiant heater is
ceramic rather than filament-based, to avoid burning out.
In the case of fused quartz, you can get both UV-grade and IR-grade, and each one is not
very transparent to the other band. Therefore, using UV-grade fused quartz for the
inner glass would result in the inner glass getting heated efficiently. The glass covering
the radiant heater and induction coil should probably be IR-grade. It will absorb more
IR anyway due to being closer to the heater, and will also be heated by conduction. The
mechanical properties of the two grades are pretty much identical, so they should be
easily weldable together.
(A resistive heater embedded throughout the envelope (similar to the rear window
defroster of a car) could be used instead, but oxygen could get in through where it
entered the glass, oxidize the metal, and, even if it didn't crack the glass, result in the
Please point out any failure modes I've missed. :)