Tubing and odors, a non-ideal combination

In olfactometry, one is constantly fighting to produce temporally precise and stable concentration outputs.  This is especially critical when you want to make detailed measurements of perception or neural activity.  Unfortunately, investigators are constantly confounded by artifacts that can distort the temporal properties and concentration profile of odor output.  On this blog, I hope to eventually cover several issues that typically come up in calibrating olfactometers.  In this first installment, I will deal with the issue of tubing after the jump.



There is general wisdom in the field that Teflon(R) tubing is the best choice of flexible tubing for olfactometry as it is least permeable to odors.  However, this tubing is rather incompatible with pinch-valve based olfactometers.  Here I will show how tubing choice influences odor output in a very simple system.

The setup is as follows:

1. Two output channels: one clean & one "dirty" 
2. Flow to the PID switched between the two channels by pinch valves 
3. Tubing is made dirty by running 30 sec of Amyl Acetate through it @ > 350 ml/min 
4. The dirty tubing is then placed into the path of the PID and clean air was flushed through.

The experiment is repeated with different types of tubing:  Silicone (C-flex), Tygon, Polyprophylene, and Teflon.  Unfortunately I did not yet do glass and stainless steel.  All tubing was 1/16" ID and ~16 cm in length.

The PID output from clean air being run through each type of tubing indicates how much odor is retained in the tubing and the decay in PID output indicates how fast the odor desorbes from the tubing.  The PID output is measured in Volts.  Output from a 100% AA bottle with Teflon tubing gives roughly 25 Volts.

Tubing sorption measures for different tubing types.  Voltage indicates how much odor is retained in tubing after a 30 second saturation period during which 100% saturated vapor is flushed through the tubing.  Odor is Amyl Acetate.   PID reads 25 Volts for 100% Amyl Acetate saturated vapor.  Odor flow = 350ml/min.  Time zero indicates switch from clean air to air delivered through the dirty tubing.

As we can readily see, by far the worst offender is C-Flex.  This tubing retains enough odor to give an output signal of 5 V, roughly 1/3 as strong as the fully saturated vapor (25 V).  Teflon is the best, giving a signal that is only a tiny fraction of the saturated vapor.  You can also see that the C-flex tubing shows the slowest desorption rate.  So much odor is retained in this tubing that the air passing through it is odorized for the full 5 seconds.  Other tubing shows a much faster odor release.

These results are easier to see in bar graph form:

Average amount of odor (Volts) retained by tubing.  For reference, 100% AA vapor reads 25 Volts.

Rate of odor release from tubing.

 So, as expected Teflon is the best, but polyethylene is pretty similar to Teflon.  Moral of the story: if you want to deliver odor, keep the amount of non-teflon tubing to a minimum!

PS.  C-Flex tubing is such an odor sponge that passing odor through it actually reduces concentration at the output.  Further, sorption/absorption and release are not symmetric at all!  Here is what happens when you run through four 5sec odor pulses through C-flex tubing and then run through four clean air pulses:

C-Flex tubing is an odor sink.  Red: saturated AA vapor passing through C-Flex tubing as a series of four 5 sec pulses.  Note distorted non-square waveform.  Purple: the same tubing immediately flushed with clean air.  Note decay in odor release from the tubing is much slower than the apparent rate of odor absorption/adsorption.

Compare this to what happens with Teflon tubing:

Teflon tubing is much better.  Red: saturated AA vapor passing through Teflon tubing as a series of four 5 sec pulses.  Note fairly square waveform.  Blue: the same tubing immediately flushed with clean air.  Note negligible odor retained in tubing relative to saturated vapor. NOTE: Voltage gain was reduced by a factor of 5 (5 V here is 25 V) on other plots.