Sustainability of LED production?

Question

Not all LEDs are the same. Standard LEDs use various combinations of indium, gallium, arsenic, silicon, phosphorus, zinc, selenium, and a couple other elements. Most of these need to be mined from the earth. Organic LEDs sound at first like they might be better, but they are only organic in the sense of "organic chemistry," not "grown without pesticides." However, in addition to organic compounds they also use indium, tin, barium, calcium, and aluminum so mining is still necessary.

Are there any LED technologies that avoid depleting limited mineral resources either by using only abundant materials or by being fully recyclable? Failing that, are there any LED technologies that are clearly more sustainable to produce than the others in terms of resources and environmental impact? Or do all LEDs effectively have equal impact and only differ in performance?

Answer 1

I think this is a great example of how sustainability really is multi-faceted. LED lighting may be great from an energy consumption standpoint, and not as good with respect to finite material usage.

Energy Use vs Material Supply

I think it's important, however, to apply some weighting to the different factors. I don't believe we're yet at a critical point with respect to rare earth material supply, but I believe we are already at a critical point with respect to climate change (primarily, that we have already passed the 350 ppm CO₂ threshold that many climate scientists believe is necessary to preserve the climate that humans have flourished in). So, in my opinion, it would be poor strategy to pass up energy-saving technologies in favor of technologies that save raw materials that we aren't running out of yet.

Light bulbs are unique devices, because they're small, but use lots of energy over their lifetimes. Their relative material footprint isn't that big, compared to their energy footprint. LED lights are expensive to produce, and use much more scarce material than incandescent bulbs, but they last for years (or decades).

LED lights are already the most energy efficient lighting solution available for most residential applications, and there is good reason to think they'll soon pass compact fluorescents by a significant margin.

Short Term Supply Constraints

Here is a report on GE's website about rare earth supply constraints. It's important to note, however, that current price spikes in these materials are largely an economic phenomenon. 95% of these materials are currently supplied by China, because China's policy is heavily influenced by engineers who saw this trend coming. China is also imposing export quotas. There are significant rare earth material reserves in the US (e.g. Nevada) and Australia, but the higher cost of labor here/there hasn't yet convinced those countries' companies or governments to ramp up production. It also takes years to bring new mines online. So, I would expect in the future, as climate problems worsen, and China's economy surpasses that of the US, this supply limitation will ease. Whether it will ease enough to satisfy increased demand requires a crystal ball (also a rare earth material).

Recycling

Another issue sometimes brought up is recylability. It's important not to assume that because a material isn't currently recycled, it's not recyclable. This assumption is made repeatedly with electric car batteries. Until products like LED lights are getting disposed of in large numbers, economic factors limit the opportunities for recycling them. That's not a long-term problem though, if LED lights (and other rare earth applications) catch on.

Long Term

Keep in mind that the theoretical potential for renewable energy is enormous. Just between solar and wind power, there is thousands of times more energy hitting the earth than we need (and even the harvestable potential is much more than we need). We're just not harvesting much of it now, in favor of consuming ten million year old sunlight stored in fossil fuels. But, in the long term future, we may be able to (once again) use technologies that use more energy, if the energy is renewable. At that point, we may decide to reduce, or reinvent LED lighting. But for now, I think the large energy saving potential of LED lighting makes it a good medium term (0 - 50 years) sustainability choice.

Answer 2

Its been many years since this question was asked, technology has developed, newer LEDs are far more efficient than LEDs from a decade ago. The control technology powering LEDs has also become far more efficient.

This introduces complications in estimating the sustainability of LED lighting. If you take a snapshot of LED technology in 2010, and asked how long can we go on like this, manufacturing the same LEDs in the same quantities, you might get an answer like fifty years.

But if you took a snapshot of 2019 LED technology and asked how long can we go on make the same quantities as 2010, you might get an answer like two hundred years.

But additionally as LEDs have become more efficient, they're also cheaper and so demand for them has increased, so you might only have the resources to build 2019 technology LEDs for a hundred years.

Since the efficiency, price and size of LEDs keeps improving and becoming broadly more sustainable, then the question is perhaps about limits.

In 1997 the UK reached 'Peak light bulb', since then the amount of energy required for lighting has reduced as consumers changed from incandescent lighting to more efficient technologies. In this respect lighting is now generally sustainable.

The old-fashioned incandescent light bulb was in production for over 150 years, the great leaps in efficiency were mostly in the first few decades of production, and by 1997, it couldn't be manufactured any more efficiently, it had reached its technological optimum.

LED lighting is still making leaps in efficiency and sustainability and will continue to for a few more decades. At each leap it is more sustainable than it ever has been, using smaller quantities of the limited mineral materials.

Answer 3

I think you are imagining a problem that isn't there.

It is true that LEDs use several materials, some of which are rare. But incredibly small amounts of these materials are used.

The most common type of LED today used is the blue gallium nitride LED. White light is produced from it using a phosphor material that converts blue wavelengths to longer wavelengths. Common phosphor is cerium-doped yttrium aluminum garnet (containing yttrium, aluminum, oxygen and very little (almost no) cerium).

For example, a common LED that can be run at 8 watts is Osram P9. Osram doesn't specify the dimensions of the chip, but based on the datasheet I can estimate it's at most 9 square millimeters because larger chip wouldn't fit to the final product.

So we can say LEDs need about 1 square millimeter per watt.

Gallium nitride isn't actually used for wafers because a good production method to make gallium nitride wafers doesn't exist. So actually the wafer (most of the material) is sapphire or aluminum oxide. Both aluminum and oxygen are very common materials. Then a gallium nitride buffer layer is grown on the sapphire wafer, and the active layers are grown on the buffer layer.

According to several searches on scientific papers, the buffer layer thickness of GaN is about 20 micrometers. So one watt has 20 µm x 1 mm x 1 mm of gallium nitride. That's about 0.123 milligrams of gallium nitride per watt. 83% of that is gallium rest being nitrogen, so we can estimate that about 0.1 milligram of gallium is needed for 1 watt of LED.

Gallium doesn't occur in rich deposits but at least 1 million tons of gallium are available in bauxite and zinc ores (according to Wikipedia).

This means that our gallium deposits can be used for 10¹⁶ watts or ten petawatts of LEDs.

I have around 700 watts of LED lights in my house, car, bikes, etc. And that's a lot -- I really enjoy lots of bright lights. If every person had the same amount of lights, and if world population stabilizes at 10 billion, we could make about 1400 iterations of these lights for everyone. If one iteration is good for 3.5 years, that means we are good for at least 5000 years.

There are far more limiting resources that become a problem quicker than in 5000 years, such as phosphorus reserves used for fertilizers running out.

I also searched for a typical thickness of YAG used in LEDs. One paper said 75 micrometers as an example thickness. At this thickness, 0.342 milligrams of YAG are needed per watt. 44% of that is yttrium so about 0.15 milligrams of yttrium is needed per watt of white LED. Yttrium oxide reserves are half million tons and 78% of that is yttrium so we have about 0.4 million tons of yttrium.

This is enough for 2.7 petawatts of white LEDs. Assuming every person on this planet would use 700 watts of white LEDs, that's enough for 386 iterations of lights or 1350 years if each iteration is 3.5 years.

The other materials are used in such small quantities that they don't matter. For example the cerium doping in YAG is only doping, so very small amounts of it are used.

Also both yttrium and gallium can and will be recycled, unlike fossil fuels which are used once and then they're gone.