Admittedly, there still a lot of room for solar panel research for the upcoming years which is pretty exciting!

The discrepancies is often due to fact we're defining "solar panels" quite different for each application and also the definition of input energy is different.

A great example is engines. The theoretical (Carnot) efficiency is limited by the heat difference between the source and the sink. You can burn fuel at a higher temperature and pressure which will give you great efficiency, but it's not realistically possibly due to physical constraints (usually the steel that houses the explosions). That's why steel engines are often said to have a maximum fuel efficiency of 37%.

This same idea applies to solar panel but it's even more convoluted. What is input energy? Is it all the energy from the sun or is it only the energy within the energy band that the panel can "see"? What is efficiency? The efficiency for different energy band is different for each solar panel as well. Worst of all, this efficiency changes for different temperature and is effected by hysteresis! By the time you store the energy, you also lose efficiency depending on the internal resistance of the energy storage (That's why MPPT exist and fun fact of the day, from what I know, there is still no reliable analytical method to account for this. Most MPPT trackers are done empirically).

Sorry for the long rant, but yeah, solar panels are confusing and there is a surprising about of research that still needs to be done!

These things are well defined in the research community and they don't have any issues with this. The input energy is all the energy from the sun that is incident on the panel. Efficiency is electrical power generated by the panel divided by the power from sun (~1000 W/m^2). For the NREL efficiency records, the measurements are all taken at the same temperature at the NREL facility. To find the max power of a device you just sweep the voltage applied to the device and measure the current output and determine which voltage produced the most power (P=V*I). Then operate your cell at that voltage.

I heard about carnot energy but didn't realize the calculation was this simple. Suppose we say we get a insulated material that is string enough to handle high temperatures. We can beat the 37% right?

It's a partial yes. You have to worry about pressure as well. You can burn fuel at a higher temperature by igniting the gas at a higher pressure. But there are certain constraints due to the fuel. If the pressure is too high, the fuel will spontaneous combust. That's why there are different octane rating of fuels and why higher octane fuel gives you better fuel efficiency (given your engine has some sort of valve timing like VTEC). This is also why Diesel engines tend to have better mileage than gasoline engines. Even though gasoline engines will perform better at the same pressure as Diesel, there just no way to bring the gasoline to the same pressure that the Diesel can handle.

When and how you combust the fuel can give you different fuel efficiencies as well. For instance, the Brayton cycle can achieve easily over 50% fuel efficiency. But in order to achieve this, you need a special housing (jet engines uses this cycle) to be able to compress the fuel to the specific pressure and also handle the pressure. Also, special fuel is required. These two things makes it not really practical for the standard consumer.

On a tangent, Internal combustion engines is probably one of the funnest and hardest class I took in graduate school...

You can, but it's worth pointing out that cars don't really run on the Carnot cycle because, unlike your air conditioner, they're not reversible. Usually they use something like the Otto cycle. To improve efficiency you can change your cycle or push the limits (temperature, compression ratio, etc.). The current generation Prius apparently uses an Atkinson cycle and achieves 40% efficiency. Some F1 engines notably have a thermal efficiency of 47%.

https://arstechnica.com/cars/2016/05/turbulent-times-for-for...

It's mostly about the max temperature. If combustion in a steel engine runs at 250 C and operates in a 50 C environment, the max efficiency is 1-325K/525K = 38%.

Ceramic, for example, can operate at far higher temperatures, so in theory a ceramic engine can be more efficient.

However, ceramic is harder to work with, and not as durable as steel, so it's not a practical alternative.

There's a good explanation and derivation of Carnot's results on the efficiency of heat engines in the Feynman lectures [1].

[1] http://www.feynmanlectures.caltech.edu/I_44.html

Yes; I believe this is why jet turbine engines have blades made of single crystals of exotic materials.

> What is input energy? Is it all the energy from the sun or is it only the energy within the energy band that the panel can "see"? What is efficiency? The efficiency for different energy band is different for each solar panel as well

Don't research/industries usually come up with unambiguous terms and units for these things so they can compare metrics? e.g. Our prototype showed 94% VSE. VSE: "Visual Spectrum Energy" is the input energy between 400-650nm as measured by ... "

It's even worst than that in practice. I remember an article from many years ago where the author went through the actual efficiency of the various components in a gasoline-powered car. The conclusion was that the entire system was somewhere around 18% efficiency.

I can't remember the name of the article. It may have been in one of my engine/automobile design books. Don't know. Sorry.

When talking about solar cells, the efficiency is defined as electrical power produced divided by the solar irradiance on the panel (~1000 W/m^s without concentration). In this situation, the theoretical limit for a single junction solar cell on Earth is ~34% as shown by Shockley-Queisser in their detailed balance calculations. Multijunction cells can be more efficient than single junction cells and can theoretically reach ~70-80% efficiency if >30 junctions are used and concentrated sunlight is used. In practice it is prohibitively difficult and expensive to make a device with more than 4 or 5 junctions, so you will not see one sun efficiency greater than 70% in real life. The 95% number is for a Carnot engine and doesn't apply to an operating solar cell. In order to extract power from a solar cell, you need to operate it in forward bias (typically at the maximum power point). When at the maximum power point, the solar cell emits radiation and therefore has irreversible losses. This is why the real limit for a single junction cell is 34% and not 95%.

It's not that.

The 95% is the limit. You can't extract solar power with a higher efficiency. (Funny thing that you can extract energy from Earth, radiating it away, with ~99% efficiency.)

86.8% and and 40% are theoretical limits to some current technologies. The others are people debating what is the best available on the market (with varying definitions of "available").

Could you explain why 95% is "the limit"? Are you saying 95.0% efficient solar panels are possible, but 95.1% efficient solar panels are impossible? What's the limiting mechanism here and why is it such a nice round integer?

This is a good question - And this is a partial answer (from a link above): https://news.ycombinator.com/item?id=13939399

Not sure why you were downvoted. Seems like this community is showing signs of academic elitism

> why is it such a nice round integer?

Significant Figures [1]

[1] http://www.bbc.co.uk/schools/gcsebitesize/maths/number/round...

Those ultimate efficiencies come from thermodynamics[1], since both the Sun light and Earth heat are thermal.

Both are rounded.

1 - https://en.wikipedia.org/wiki/Carnot_cycle

And yet, you can take a solar cell out to Pluto where the mean temperature is 40K rather than 300K and theoretically get far higher efficiency than 95%. So 95% isn't really a fundamental limit.

Yep, I'm assuming the solar cells are on Earth. Seemed like a good assumption to make.

If you just send them into space with good enough radiators, you'll be able to get 3 9's of efficiency (mid way to 4). But you don't get to reuse the heat waste.

You may be interested on that:

https://en.wikipedia.org/wiki/Matrioshka_brain

I suppose it's similar to why you cannot have a 100% efficient wind farm?

Wind is not thermal energy, so energy extraction from it can achieve 100% efficiency in theory.

Turbines do have some theoretical limitation, but it's a technology limitation, not an ultimate theoretical one.

I thought it was more fundamental than that; to capture energy, there needs to be energy, if your wind turbine transforms all of it, there's no current anymore and the system will change phase.

Yes, Betz's law [0]:

"Consider that if all of the energy coming from wind movement through a turbine was extracted as useful energy the wind speed afterwards would drop to zero. If the wind stopped moving at the exit of the turbine, then no more fresh wind could get in - it would be blocked."

[0] https://en.m.wikipedia.org/wiki/Betz's_law

Thanks, I didn't know the name.

This is a fundamental limitation of turbines. Thus, a technological limitation for wind-power. Other technologies could have other limits.

Besides, that's a limit on collecting efficiency (how much you can collect / how much is there), not usage efficiency (how much you can use / how much you collect). You don't "destroy" the wind when your turbines don't collect it, it's still there.

I see, I considered the turbines as the only mean. Are there other transformer types for fluids ?

A solar cell could never achieve the Carnot limit, so the 95% number is not relevant. In order to extract the maximum amount of power from a solar cell, must operate it at it's max power point. When you do this, the cell is in forward bias and must emit light (because of the physics of absorption/emission). Emitting light is a loss mechanism and thus people take this into account when calculating the theoretical efficiency for solar cells. It has nothing to do with currently available device on the market and is just accounting for how solar cells work.