CO2 sensor design

deepwrecksc

This post is prompted by the sobering report posted in the "Rebreather Accidents" forum by Dave Sutton, detailing his own experience with hypercapnia at depth. I was astonished at how many other people reported similar experiences in that thread. It occurs to me that a reliable electronic "canary in a mine" for CO2 should be top priority for EVERY rebreather manufacturer right now.

I am an analytical chemistry professor by trade and so I know that there is absolutely no reason for absence of such a sensor in the marketplace. People who say it can't be done reliably simply don't understand the science behind these measurements. I know that there are several groups trying to get one of these to market right now. Laguna Research comes to mind.

I think that the average Rebreather diver could do with some knowledge on how these measurements might be made. Also, many of us here are technically-oriented professionals (electrical and mechanical engineers stand up). Maybe this can be a starting point at which we can exchange information on this topic to get something accomplished. I will go first. Below are three chemistries that can be used to build sensors for carbon dioxide. There are likely others.

1). The simplest method is to use infrared spectroscopy. CO2 has several unique absorbance bands in the IR spectrum. It turns out that water vapor also absorbs in the IR, and there is some spectral overlap (see spectra below). It is straightforward to use either multi-wavelength or full-spectral IR absorbance analysis to quantify low concentrations of CO2 in a humid atmosphere - this is done by atmospheric chemists on a routine basis.

Now there are several problems with using IR spectroscopy in a rebreather. First is that the optics are very different than those used in "visible" spectroscopy. You can't just use normal glass. Glass is opaque to IR radiation below a certain energy. Chemists use cells made of potassium bromide or in some cases cesium iodide. Unfortunately these are soluble salts and don't last too long in contact with humid air (or water for that matter). There are other specialty optical materials (such as sapphires) that have been developed to get around these problems so this can be solved. Another problem is that the IR energy source is difficult to make in a low-power configuration. An IR emitter is often just a heated piece of metal (or some other material), and such a resistive heating element would quickly run down a battery. One solution may be to periodically pulse the source to conserve power. Detectors are another challenge. IR radiation is not detected with high sensitivity by inexpensive silicon photodiodes. Instead, you need germanium or InGaAs detectors or thermopiles. These are expensive, slow, and often require peltier cooling (also consumes a lot of energy). It is not surprising that IR-based methods for CO2 analysis in rebreathers has not caught on in a big way, but it is certainly POSSIBLE to build a sensor if one had the inclination (and budget).

2). CO2, when it dissolves in water undergoes the following equilibrium:

CO2 + H2O <--> H2CO3

In other words, CO2 makes carbonic acid when it dissolves in water - this causes a decrease in pH for weakly buffered solutions. A second method for analyzing CO2 is to take advantage of this behavior by placing a pH electrode (electrochemical sensor that responds with a change in voltage as the concentration of acid changes in a solution) in a small amount of electrolyte behind a gas-permeable membrane in contact with the gas to be measured. CO2 from the gas phase permeates into the electrolyte, dropping the pH and causing a response in the electrode. These sensors can be made quite small. The one I use in my laboratory is from Microelectrodes Inc. (Dip-type Carbon Dioxide, Microelectrodes,Inc.). It is easily calibrated, inexpensive, and works very well under 100% relative humidity conditions (actually it is completely submersible).

Challenges for electrochemical CO2 measurement include sensitivity (this method is not as sensitive as an IR-based technique would be) and compensation for temperature drift and pressure effects (like a galvanic O2 sensor these electrodes measure gas partial pressure). Also, there is a measurement lag as it takes a bit of time for the CO2 to diffuse across the membrane. However, these sensors require ZERO power (simple mV output) and could be integrated fairly easily into a breather. It would be worthwhile simply trying the Microelectrodes Inc. sensor to see what would happen on a couch dive or in a pressure pot with calibrated gases.

3) The most sophisticated and complex way to make a CO2 sensor is with a technique called time-resolved fluorescence resonance energy transfer (TR-FRET) or dual-luminophore referencing phase fluorometry. This is an active area of research right now and is very exciting. I've attached a few papers that the TRUE techno-geeks might want to look at! The way it works is that you take two dyes - one with a fast fluorescence "decay" that is sensitive to pH (and as above CO2), and one that is inert but has a long "decay" or lifetime, and you immobilize these two in a polymer or sol-gel matrix together so that they may be excited together by a light source and interact with each other. When the CO2 concentration changes, it causes the pH-sensitive dye to "glow" less intensely (be quenched) when excited by a pulsed blue diode (~10-15 kHz). The other dye doesn't change.

By pulsing the excitation light and then measuring the frequency-domain phase shift of the light emitted by the combined dyes, you can measure the CO2 concentration without any interference from dye bleaching, light intensity fluctuation, etc. In other words, the signal is very stable (which is of course what you want). This clever modulation method (phase fluorometry) solves a lot of the problems that has made fluorescence "optodes" like this unattractive in the past. It turns out that the most common "inert" dye is a ruthenium complex that is highly sensitive to oxygen. You can make an oxygen sensor using this method too (see paper), but honestly I wouldn't trust my life to the numbers I got from the currently-available instruments. An "oh sh*t" CO2 alarm is a different story though. The exact numbers aren't important - just tell me if I have CO2 over a certain threshold! It is easy to make the ruthenium dye insensitive to O2. Simply encase it in a gas-impermeable nanoparticle. You can buy these commercially.

Check out Ocean Optics' "Redeye" patches (RedEye Oxygen Sensor Patch) for an example of what a company is currently marketing for O2 measurement using this technology. CO2 measurement is simply a matter of changing the sensor disc chemistry - the optics are the same. Incidentally, a very clever group at SUNY Buffalo has made a microminiature phase fluorometer that will fit (entirely) on a single CMOS-chip (including light source and detector) - check out the attached paper if you want to see how. Again, it was designed for O2 but CO2 would be straightforward.


OK - well I hope I've given you some food for thought and maybe some fodder for discussion. Please weigh in on this topic with your own expertise and maybe we can actually get something accomplished. Too bad there isn't a government agency that would be interested in this - the ideas are ripe for a grant proposal!

Lee

Attached Images

	DLR CO2.pdf (138.1 KB, 15 views)
	CO2 phase fluorescence.pdf (133.9 KB, 12 views)
	cmos O2 phase fluorescence.pdf (530.8 KB, 11 views)
	CO2.gif (4.7 KB, 664 views)
	water.gif (7.5 KB, 657 views)

PaulTG2

Hello,

Great summary. Thank you very much. Have some green.

So let me ask a dumb question:

If we took a few ounces of Sorb and put it in a miniature scrubber located after the main scrubber. If CO2 passed through the main scrubber this mini-scrubber would pick it up and have a chemical reaction to it. Is there a way to measure that reaction and use that signal to drive an alarm?

In this situation you could reload the mini-scrubber each time you reloaded the main scrubber.

Just thinking out loud. I expect there's a lot of reasons this doesn't work... or should i build a test bed.

hmmm.... Perhaps a small, thin 1 ounce scrubber just before the inhalation hose with the sorb between two metal screens coated in.... hmmm....a small PLA to measure the resistance/PH/??? of the sorb and alert if..... doubles as a flood detector.... hmm....

-p

__________________
Paul's first law states that the safety of an activity is determined by how forgiving of mistakes the activity is.

Paul's second law states that the difference between an adventurer and an explorer is whether the doing or the learning comes first.

cramerdn

So all we need is a clear tube packed with color change sorb with a couple of fittings to inert it into the loop next to the DSV. When the kitty litter turns purple its time to go.

__________________
"Its better to live one day as a tiger than an entire life as a worm."

"But who's ever heard of a worm skin rug?"

deepwrecksc

Actually a very good question. A color change on a "breakthrough" column like that is a very straightforward (and easy to see) diagnostic. You could also use a thermocouple to detect temperature change from the reaction (like APD does on the inspiration/evolution scrubbers). In fact this method (visual) is used very often on things like gas purifiers upstream of analytical instruments so that we know when to change out the cartridges.

A few challenges:

1) Will the reaction in the "canary cartridge" give you sufficient warning before an incapacitating hit? I have no idea how much CO2 is required before the color changes. It would have to be tested.

2) A small column of sorb like that might be a problem in terms of breathing resistance if incorporated into the loop directly.

Definitely an interesting idea though. Maybe instead of a full column of sorb, just a "pH indicator" dye coated on the inside of a transparent tube within the view of the diver? Obviously wouldn't be easy to see in the dark... but hey who knows?

jasondrake

I'm sure that paramedics regularly use a chemical colour indicator to show CO2 presence in open/closed breathing loops when providing gas to patients.
Maybe one of the medics can confirm/provide details? I proposed this as an interesting appraoch months/years ago but can't find my own posts!

It's designed to work once, reacts very fast and is designed to work in a breathing loop so is moisture/temperature tolerant within expected ranges. Also, as compromised/weak patients can breathe through it without significant WOB problems sounds ideal.

I thought placing it upstream of the BOV with some sort of window which was visible to the diver might work. Not a high tech/high cost retrofit probably. Question is...would it already be too late by the time it reacted?

PaulTG2

Hello,

I think there needs to be an alarm. Color change is great if you happen to be looking at it but seems worthless if you're not. I don't think I'd spend a lot of time looking at it while I dive. Checking if you think you have a problem may be a bit too late based on Dave's post.

I would expect a sensor detecting the change in a small amount of sorb located post-scrubber would be a better idea. The chemical reaction is quick and result in changes which should be readily measurable. It could trigger a visual and auditory alarm. Change it when you change the sorb keeps it fresh and easy because you already have the sorb at the ready.

Sincerely,

-p.


-p

__________________
Paul's first law states that the safety of an activity is determined by how forgiving of mistakes the activity is.

Paul's second law states that the difference between an adventurer and an explorer is whether the doing or the learning comes first.

PaulTG2

Hello,

So, is there a simple way to electrically measure/detect the chemical reaction of sorb in the presence of CO2? At what CO2 concentrations and with what delay?

Sincerely,

-p

__________________
Paul's first law states that the safety of an activity is determined by how forgiving of mistakes the activity is.

Paul's second law states that the difference between an adventurer and an explorer is whether the doing or the learning comes first.

deepwrecksc

Well it would be possible using a thermocouple probe - since the reaction is exothermic and releases heat. Pretty much it is just this reaction:

Ca(OH)2 + CO2 --> CaCO3 + H2O

The pH *does* change due to the reaction but by the time you sensed a pH change in the sorb stack you'd be toast due to the CO2. You have to be careful measuring a pH change in such an alkaline environment.

I don't know what the color indicator used in sorb is (on a chemical basis anyway), so I don't know how much CO2 "saturation" is required before the color change is noticeable.

Best bet would be to use a specifically-designed sensor instead of monitoring the sorb reaction. That would ensure the highest sensitivity and quickest response time.

Jasondrake: The color-change indicator you mentioned does sound interesting. Any idea what the manufacturer is? The challenge would be integrating this to some kind of automated monitoring transducer.

Lee

PaulTG2

Hello,

Thank you for the input. The reason I thought to use sorb is because you can change the material easily when you change the scrubber hence life of sensor isn't an issue and we know that sorb works in the environment very well.

I guess one temp sensor outside the sorb-sensor and one inside and then compare the two would give you a pretty fast read that the sorb was activated. Perhaps a small pipe made with screen filled with sorb and a temperature sensor running down the middle. Placed in the general flow of the post-scrubber gas. Any CO2 that activates the sorb would be bad so use that to generate an alarm.

Just thinking out loud.

-p

__________________
Paul's first law states that the safety of an activity is determined by how forgiving of mistakes the activity is.

Paul's second law states that the difference between an adventurer and an explorer is whether the doing or the learning comes first.

swampdiver

The older end-tidal CO2 monitors were simply a colourimetric device which changed colour at a particular threshold likely well above what would be safe for rebreather divers (i.e 5 percent CO2).

I see there are some new devices being used for the paramedic market described here.
Paramedic Blog: New Developments in End-Tidal CO2 Monitoring

The link to the Capnocheck doesn't seem to work but this one does. The very small portable unit seems to rely on IR spectroscopy.
http://www.smiths-medical.com/upload...sell-sheet.pdf

One would have to look at the relationship between end-tidal CO2 in mmHg at various inspired percentages of CO2.