[Gasification] Gas Sensing Rig Update

[Bear Kaufmann]

Hi all,

I've been promising to post a message about development of the gas 
sensing rig, so here goes:

On a rainy day a few weeks ago, Jim and I did a good deal of internet 
research on gas sensing.
We came to the conclusion that we don't need to sense tar per se to 
optimize the system, as the concentrations of other gases will give a 
good indicator of the status of the system. When running the systems 
experimentally, we'll want to get as much data as we can, but when 
working towards an automated system, in all likelihood, one sensor 
should suffice to give a good indication of the status of the system.

There are numerous fully built gas sensing devices, ranging from car 
exhaust analyzers for smog checks, to industrial flue gas sensing, 
personal safety, and industrial process sensing. Most of these devices 
run into the thousands of dollars. Most of these devices won't allow 
direct integration into an automated unit anyway. So we decided to focus 
our research on individual sensing components. Most of these should be 
reasonably easy to interface with a microcontroller for data logging and 
eventual automation.

There are a few main types of gas sensing technologies. I'll list these 
below, and briefly describe my current understanding of their operating 
principles:
*NDIR* (non-dispersive infrared) - Uses the IR absorption spectrum of a 
gas to sense gas concentration. Probably won't shift like some other 
techniques will. Generally very selective (one gas only).
*Catalytic Bead* (Pellisistor)- Uses a heated reference bead and 
catalyzed bead, difference in temp between catalyzed/uncatalyzed is a 
function of gas concentration. Not very gas specific.
*Semiconductor* - Gas adsorbs on metal oxide (usually tin dioxide) and 
changes resistance of the sensor. Most likely cheaper then IR, but often 
not very selective, and output may shift over time. They generally sense 
HC gases, though there are some sensors with pretty high selectivity to H2.
*Solid electrolyte* - Gas generates voltage with voltaic cell. Generally 
pretty gas specific, limited lifetime.
There are additional technologies out there, like *Thermal 
Conductivity*, which requires binary mixtures either in a carrier gas or 
air, which works well with a gas chromatography column, but not a mixed 
real-time gas stream. *Tunable Laser Spectroscopy* and other 
spectroscopy techniques are out there. Often packaged for large 
industrial facilities.

I'm currently focusing on development using sensors from Figaro which 
are mostly Semiconductor and Solid Electrolyte for O2 and CO2. The 
semiconductor sensors handle H2, CO, Methane, and a variety of other HCs 
with cross sensitivities. They vary their resistance, so I'll be 
measuring changing voltage of 0-5V with an Arduino microcontroller board 
which will communicate with a computer via USB for data-logging. The SE 
sensors with require a little more circuitry to produce a 0-5V voltage. 
I'm still looking for NDIR components, most companies selling components 
are looking for an OEM manufacturer. I'm also looking at chips for 
interfacing thermocouples to the Arduino so all the sensing will be done 
by one unit.

Since most sensors will be poisoned by high levels of gas (excluding IR 
and some electrolyte) and most max out at sensing around 10,000 ppm 
(1%), I'm planning on diluting the gas 100x with air. Current dilution 
setup is two variable area rotameters with integrated valves. Gas will 
need to be cleaned of tars first, which I plan to do with a water column 
given the relatively small volume of gas needing cleaned, though I'm not 
an expert on gas cleanup. Diluted sample gas flow will be around 1-2 
L/min so as to avoid high velocities within the chamber containing the 
sensors (which can skew reading on some sensors), this flow rate may 
need to be increased by an order of magnitude.

So there's the current status. I welcome feedback, though I'll be quite 
busy the next few weeks, so communication and development of the rig 
will likely be delayed.

Cheers,
Bear