Environmentalists are sort-of "torn" regarding what to think about fluorescent light bulbs. They are much more energy-efficient than old-fashioned incandescent light bulbs, but they also contain toxic mercury. Dead fluorescent bulbs often get broken outside of an appropriate processing facility, which
releases mercury into the environment.
It should be obvious, though, that if the bulbs were built to last longer, then fewer would be discarded each year for the reason of having become poorly-functional (a flickering bulb is often considered useless even if it isn't actually "dead"). Let us now examine the typical reasons why they fail, and what might be done about extending their lifespan.
Most commonly, fluorescent bulbs exist in two main varieties. Smaller bulbs tend to have 4 electrical connector pins; larger bulbs tend to have 2. We'll start with the 4-pin bulbs.
The 4-pin bulbs have 2 pins at each end of the glass tube; and inside each end is a coiled filament, much like the coiled filament of an ordinary incandescent bulb. If either of the two filaments break, in this type of fluorescent bulb, then the bulb fails to work properly. It should be noted that these filaments tend to last longer than incandescent filaments not so much because they are built tougher, but because they are operated at a much lower temperature (the higher the temperature, the faster that metal atoms boil off the filament; the more that boil off, the weaker it gets, until it breaks).
The 2-pin fluorescent bulbs don't have any filaments inside them. They simply use a higher "discharge voltage" than the 4-pin fluorescents, which causes an electric arc to travel inside the glass tube from one end to the other (the 4-pin bulbs can use a lower discharge voltage because the hot filaments make it easier for an electric arc to come into existence). Nevertheless, electric arcs can do damage to materials, and so metal atoms are caused to boil off of the specially shaped electrodes inside a 2-pin fluorescent bulb, until not enough metal is left, of the right shape, for an arc to form. (The blackened end of a dead fluorescent bulb is due to those boiled-off metal atoms condensing/plating-out onto the inside of the glass.)
For this Idea, then, the goal is to consider ways to make the bulbs last longer. One obvious notion is simply to make the electrodes heavier-duty; the more material physically exists in an electrode, after all, the longer it will take to boil away.
What I'd rather do, however, is figure out a way to make the boiled-off metal recondense onto the electrode! Then the bulb might last even longer than a bulb with even the heaviest-duty practical electrodes....
Before getting to that, though, let's examine the role of mercury in a fluorescent light bulb. It has 3 physical properties that make it almost ideal for the purpose, and so, despite improvements over the years that have reduced the amount of mercury per bulb, it is likely that as long as fluorescent bulbs get made, they will continue to contain some mercury.
The first relevant physical property of mercury is that it is a metal and conducts electricity fairly easily. The second property is that it has a high "vapor pressure" --it evaporates fairly easily. Taken together, inside a glass tube, these two properties mean that there is some gaseous electrical conductor present, making it easier for an electric arc to traverse the length of the tube.
The third relevant physical property of mercuy relates to what happens when its atoms interact with an electric arc --they produce significant amounts of ultraviolet light. Fluorescent lights, by definition, work by converting UV light into ordinary light (the coating on the inside of the glass does that), a process called "fluorescence".
Mercury has one other physical property that is relevant, but in a negative way (no, not its toxicity!), regarding the construction of fluorescent bulbs. It is a good solvent for other metals. Gold, silver, copper, and quite a few other metals can be dissolved in mercury; this means it is necessary to construct the electrodes of some metal that WON'T dissolve. Obviously such metals exist, or fluorescent bulbs wouldn't be using mercury.
It happens that Iron is one of the metals that does not dissolve in mercury. It also happens that Iron atoms can be influenced by a magnetic field. And now for the Idea (which mostly is about the 2-pin bulbs):
Start by making the ends of the glass tube somewhat differently than is now done. We want the whole inside area of the tube-ends to be very rough glass--spiky, call it. We use fairly standard Vapor Deposition technology to coat the interior spiky glass with a suitable layer of iron. This iron layer connects to the standard electrode metal that passes through the glass to the outside (that metal is special not only because it doesn't dissolve in mercury, it also expands and contracts with temperature at the same rate as glass). Note that while iron is not an excellent electrical conductor, neither is it so poor as to be unsuited for this purpose; the amount of current flowing through a fluorescent bulb is typically less than 1 Ampere.
All this work has to be done in an oxygen-free environment, of course, to keep the iron from rusting.
The special glass tube-ends are now attached to the main length of fluorescent tubing --with some mercury added, of course-- and the bulb is installed in the usual way, along with the final details: magnets (or, perhaps, electromagnets). One is placed at each end of the fluorescent tube.
Now, when the lamp is turned on, every single spiky place on the inside end of the tube, where the metal coating is of course also making spikes, will naturally be a place where an electric arc can jump from one end of the glass to the other, and fluorescent light will be produced as is normal.
Also, as normal, metal atoms will be boiled off of each spike, as the arcs do their thing. But there are plenty of spikes, remember, and they offer a lot of surface area where vaporized atoms can condense back onto the glass. The magnets exist to ensure the vaporized iron atoms don't travel too far away from the ends of the main glass tube; the atoms will prefer to condense back onto the spiky glass. For anyone objecting, regarding the magnetic properties of iron vapor, keep in mind that the Curie Point is about bulk magnetism; individual iron atoms are still magnetic; heat just causes their individual fields to be oriented randomly instead of uniformly.
(It is to be noted that fluorescent bulbs do get warm, and heat can degrade the performance of even the best permanent magnets. For this reason it may be necessary to use electromagnets instead.)
SO, the Idea here is that if iron boils off of one glass spike, making it unusable for an electric arc, then the arc will start to use some other metal-coated spike, boiling off atoms that can recondense onto the first glass spike! With enough spikes, it can take a LONG time before all of them become useless! (condensing vapor fills up spaces in-between spikes).
Which means these bulbs can last a very long time, and the environmentalists can be less concerned about the mercury used.