Monday, February 8, 2010

Compact Flourescent Lamps: A Study in Complexity

In 2005, my father and I became obsessed with CFLs. I do not remember where we first heard about them, but after reading a few of the bold claims made we were infatuated with the idea of reducing power consumption while greatly increasing the lifespan of our lights. Though they were several times as expensive as incandescent bulbs, they promised to last about ten times as long, while saving 75% of power consumption. This, we told each other, was innovation at its finest: a technology that benefits consumers in multiple ways without requiring so much as a lifestyle change. He set out on a quest to replace every incandescent bulb in our house with a CFL.

Shortly after installing the first CFL, we noticed that the quality of light was significantly different. A room lit by a single CFL took on a sickly greenish or yellowish tint, depending on the specific lamp. A little disappointed, but not daunted, we continued replacing lights until, about 6 months after we installed the first CFLs in the house, one of them failed. It burned out. It died. Weren't these lamps supposed to last for several years? Maybe it was a fluke, we told ourselves. But it wasn't. Several more lamps failed within a year of installation.

If their lifespan wasn't significantly better than that of an incandescent lamp, suddenly the high installation cost looks a little more off-putting. With one of the lamps costing about $3, replacing them every year (or more frequently) could get quite expensive. My dad was upset, and basically gave up on the technology, switching back to incandescent lamps. I couldn't stop asking the question, "why?"

I knew that our experience must be somewhat atypical. If these lamps really had such a short lifespan, it would be common knowledge and the Australian government would think twice before passing legislation banning incandescent bulbs. No one seemed to be talking about this lifespan problem, however. I did a little bit of digging, and found out a few simple facts about the lifespan of flourescent lamps. First, their lifespan is greatly reduced if they are cycled on and off frequently. In our domestic setting, I expect that our lights went through a lot more power cycles than those of most businesses or even most houses. We had several young children living at home at the time who were constantly moving from room to room and had been taught to turn off the lights when they left a room - good advice regarding incandescent lamps. Flourescent lamps are also sensitive to temperature and humidity. Living in south Louisiana, our temperature and humidity were certainly above average. In addition, nearly all of our lights were in a base-up configuration common to overhead fixtures. This configuration has been demonstrated to decrease the lifespan of the lamps by increasing the operating temperature of the ballast. The conclusion I came to was that the combination of frequent power cycles, high temperature, and high humidity combined to reduce the average lifespan of our lamps just enough to cause a few outlying failures after a year. As far as I know, some of the CFLs we installed over 4 years ago are still fully functional, but the handful of early failures were enough to turn my dad off of the technology. He has since moved on to an infatuation with LEDs, which is probably a story for another post.

The experience of the CFLs taught me a number of valuable lessons:

First, it taught me the fundamentals of technology marketing. People want a better product, which usually means a cheaper product. More importantly, they don't want to put a lot of effort into getting their new product; they've been getting along just fine without it and can keeping going that way if adopting it is too difficult. We wouldn't have dreamed of replacing all of our light fixtures with tubular flourescent lamps, but since we could install CFLs as easily as incandescent lamps, we were willing to try them.

Second, it taught me about statistics, and human interpretation of statistics. The handful of early-failing lamps, while representing a significant deviation from our expectations, did not prove that the lamps were not economically feasible. They did destroy my dad's faith in them. If you're going to make promises, you'd better be able to live up to them.

Third, it taught me about innate complexity. George Whitesides (who I got to see at last year's ACS National Meeting) defined simplicity as "it's impossible to f**k it up" (at the end of this talk) which sounds to me like a good working definition. Applying that criterion to CFLs and incandescent lamps, what do we find? In an incandescent lamp, a current is passed through a filament, where ohmic resistance causes it to heat up and emit radiation. Bring in a materials scientist (or Thomas Edison with a lot of time on his hands) to choose a material that emits the right frequencies of radiation at an appropriate voltage. Seal the whole thing in a vacuum, and it's ready for use. Flourescent lamps are a bit more complicated. Not only does the radiation have to be converted from ultraviolet to visible by flourescent phosphor, the flow of electricity through the mercury vapor requires a ballast to regulate the current; not the most complex circuit ever designed, but a lot more complicated than the ohmic contacts needed for an incandescent lamp. In the final analysis, it is this complexity that stops flourescent lamps from pushing their incandescent competitors from the market. Their complexity increases their price, and though their lifespan is officially longer they have more failure modes, which makes them easier to f**k up.