Diodes and low grade thermal energy harvesting

[ADesilets]

TL;DR

Be it IR photocurrent or dark current, there's always a use for current.

Infrared is ubiquitous and to do with dark current -> real meters have finite resistance, no real meter is a superconductor, if you can measure dark current: there's emf.

Since we can measure and quantify 0 bias dark current generation and we can see the measurements certified on manufacturers data sheets we should be looking at doing something with it.

For current to flow through a finite resistance it has to be propelled by EMF. In my experience, you can charge capacitors with dark current, in fact you have to if you want to have any chance of measuring the EMF behind it.

Main issue is that with diodes I have access to they are still tuned to SWIR and the real energy harvesting starts in MWIR and LWIR.

I'm at a loss, most of the MCT diodes are 1000's of $ a piece and InGaAs and PbSe diodes are cheaper but not exactly cheap either.

Ideally I'd like to solder together a series-parallel MWIR or LWIR harvesting panel but need to figure out a way to get access to 4-10 uM harvesting diodes for a reasonable or a suggestion on how to modify something else to get access to those bands.

 

[Dale]

Sounds like a financial problem. You should probably reach out to a venture capitalist.

 

[Baluncore]

Main issue is that with diodes I have access to they are still tuned to SWIR and the real energy harvesting starts in MWIR and LWIR.

There is little energy at those wavelengths because the voltage is low.

Electron energy can be specified in eV.
Longer wavelengths have lower energy, lower eV.
Given wavelength in μm; Energy; eV = 1.23984 / λ μm
Infrared energy ranges from 1.653 eV to 0.00124 eV

Near IR: 0.75 μm to 1.4 μm : 1.653 eV to 0.886 eV
SW IR: 1.4 μm to 3 μm : 0.886 eV to 0.413 eV
MW IR: 3 μm to 8 μm : 0.413 eV to 0.155 eV
LW IR: 8 μm to 15 μm : 0.155 eV to 0.083 eV
Far IR: 15 μm to 1,000μm : 0.083 eV to 0.00124 eV

From that you can see that MW IR and LW IR have low energy and low voltage. You would need very many PN junctions in series to produce a useful voltage.

 

[ADesilets]

It's fine that the energy levels are low. I've got no expectation off matching a solar panel at noon. My main thought is that the power available indoors is probably better in the IR spectrum than what indoor solar panels are tuned to.

 

[Baluncore]

My main thought is that the power available indoors is probably better in the IR spectrum than what indoor solar panels are tuned to.

Solar panels do not work in the dark. They are designed to operate at visual wavelength, in sunlight.

Indoors, there will be thermal radiation from the surrounding structure. You know it has a temperature of about 20°C, so you can work out the black body radiation curve; λpeak = 9.898 um, in the LW IR.
Using eV = 1.23984 / λ μm, we get photon energy, eV = 0.125 V, at the peak of the BB curve. You will need more than 8 PN junctions per volt.

Don't forget your panel will be at equilibrium with the surroundings, radiating as much energy to the room as the room does to the panel. Have you thought of how you will cool the panel to reduce re-radiation losses?

 

[ADesilets]

If/when I eventually get my hands on the right diodes I want to start by characterizing dark current generation so I wouldn't cool it. No interest in accurate sensor readings, from what I understand you have to cool these diodes to get clean signals, kind of the opposite of what I want to play with. I want to see what the limits of thermally generated carriers are.

I know it can be measured, so that means that dark current leaves the diode to flow externally through a measurement circuit. That means that there is some finite EMF there, which means that there is some finite power dissipation external to the diode as a result of thermally generating carriers.

I want to research and characterise that.

Then I want to research ir photocurrent harvesting.

 

[Baluncore]

I know it can be measured, so that means that dark current leaves the diode to flow externally through a measurement circuit.

Look for HgCdTe (MCT) photodiodes, or LW IR LEDs.

Multiply the room temperature dark current, (probably less than 10^-9A) by 0.25V to get the power, ≈ 0.25nW.

To evaluate if it is a cost-effective way of harvesting energy, compare the cost of manufacturing the diode, to the energy harvested during the Return-on-Investment period, or the diode's lifetime, whichever is shorter.

The sensors in a non-contact IR thermometers, generate a voltage proportional to temperature. That voltage is zero at room temperature.

 

[Dale]

I want to start by characterizing dark current generation

Or you could just look at the spec sheet. These things are usually pretty well characterized by the manufacturer.

https://www.vishay.com/docs/81503/bpv10nf.pdf

 

[ADesilets]

Hey Dale, yeah the dark current is on the sheets. But that's usually something like "dark current @ 0 bias @ 25 deg C" I'm interested in how it changes with temperature. MCT would be amazing because the dark current generation at room temperature is so much it'll completely wash out any measurement signal unless you cool it, but alas mct is like 4000 USD for a single diode with 1mm² of surface area.

As for power, yeah, probably very tiny power per diode, but the diodes generally have extremely tiny cross-sectional areas and 100 nA of dark current coming off an InGaAs diode like this one, https://www.digikey.ca/en/products/detail/marktech-optoelectronics/MTPD2601N-100/17126103, with 1mm² of surface area translates to 100 mA/m². I have one of these because this diode is only 26$, the dark current will charge a capacitor, but the emf is only 1 mV. The thing will put out 300 uA @ 300 mV pointed at a candle flame, which is super impressive to me, because thats like 300 amps/m² and makes me want to do a separate project of flame power generation.

I don't know about you guys but I find that the dark current impressive considering thats the work of thermal carriers alone without even catching any photocurrent. So then, how much current and emf do I get if I don't stick the diode in a blackbox and just measure the energy collection rate pointed at different surfaces in my house? I can answer that question with the MTPD2601N-100 diodes, but it would be really nice to see how this works out for diodes that respond to frequencies of light pushing closer to 10uM.

Just yesterday I took 4 BPW34 diodes in series into a pitch black room for shits and giggles to see what would happen to the emf and what would go on with dark current, that was more interesting than I was expecting and I was not expecting to see fluctuating bursts of emf on my meter.

Any ways, I just found out yesterday night by doing some more research that while MCT and InGaAs might be basically impossible to make at home. Lead sulfide or lead selenide might be plausible. PbS and PbSe might not have as ideal a band gap as MCT but it still has a pretty small band-gap and if it can be made to be the size of a glass slide per cell this could enable some fun energy harvesting research.

I'm going to have to do more reading but if it's actually possible to make PbS or PbSe cells at home. I might have a direction to go in.

Do you guys know anything about manufacturing PbS or PbSe or can point me

 

[ADesilets]

I also want to see the limit of what cheaper components can harvest is, so I bought 100 bpw34 photodiodes and I also bought 200 3mm swir diodes but I'm still waiting on shipping.

I only have 20 swir diodes on hand right now and those will charge a capacitor to 30 mV over night in the dark. So I'm curious if the effect will be linear and 200 in series charges the capacitor to 300 mV in the dark.

 

[Dale]

I'm interested in how it changes with temperature

That is there too.

 

[ADesilets]

I'm just looking for people to collaborate on experiments with, I might not be able to put an array of MCT diodes in series and see how fast it can charge a capacitor, but someone else here might have access to them and be interested in trying it out on their own bench and letting the rest of us know right?