We all know that ideal gases are a convenient fiction. But when it comes to working with gas analyzers, the minute deviations of real gases from PV = nRT are usually small enough to ignore. Some gases, however, have entirely different ways to thumb their noses at you and your attempts to measure them. Here we list a few from our own Most Wanted list, along with their unique nasty tricks
Arsine (arsenic(III) hydride) is an especially toxic gas that is used in modern semiconductor processing. Residents of Silicon Valley were terrified to learn, years ago, that truckloads of this Godzilla of Gases were being trucked down Main Street. Less well known, is that arsine is also a problem in the metal reclamation and finishing industries. Slags from reclaiming zinc or aluminum will often evolve arsine gas just from wetting. Some storage batteries will emit arsine under certain conditions. Arsine is generally controlled by absorption onto activated charcoal. Unfortunately, this common process does not necessarily destroy the gas. Under quite mild changes in temperature, pressure, or R.H., arsine may be re-emitted from charcoal absorbents. The same is true for rubber or elastomer fittings in gas handling equipment. You can destroy arsine by oxidation, however, with an agent like Purafil (tm).
Formaldehyde may be the most frustrating compound with which to make reliable standards. When left to stand, it polymerizes into trioxane. Just when you think it is a stable white solid, trioxane decomposes back into formaldehyde. But its pranks do not end here -- trioxane itself will sublime without decomposing, so merely driving it into the vapor phase does not guarantee a reliable source of formaldehyde.
Our solution to making formaldehyde standards was to use a fairly new bottle of 40% formalin (stored at room temperature, and not refrigerated!) and inject the appropriate amount of the water solution into a sample bag. Some complex, and certainly more accurate, methods were described by Gary O. Nelson in his excellent reference Gas Mixtures: Preparation and Control (Wiley, 1992).
Hydrogen Sulfide. People who work with electrochemical H2S sensors are aware that this gas will "activate" or "sensitize" the working electrode, so that prior exposure of a sensor to H2S causes the sensitivity to increase by as much as 25%. This is presumably due to formation of a layer of gold sulfide on the electrode, with different catalytic properties than gold itself. Worse, this activation phenomenon slowly fades if the sensor is not repeatedly exposed to the gas. It is therefore not possible to maintain calibration in an ordinary H2S sensor. To be fair, it is also true that an error of just 25% does not really affect the utility of an electrochemical sensor for detecting dangerous concentrations of the gas. The electrochemical sensor is still more convenient and sensitive than other portable methods.
Nitric Oxide is rapidly oxidized by oxygen to nitrogen dioxide. This matters when dilutions of NO are made and stored in sample bags, even for short times. Most sample bags are at least slightly permeable to oxygen. We have observed that a 50 ppm dilution of NO in nitrogen in a Tedlar bag would give a response equivalent to about 10 ppm in just four hours. Even NO standards made in nitrogen and stored in cylinders are only reliable for about three months.
Nitrogen Dioxide is in equilibrium with nitrogen tetroxide, N2O4, at elevated pressures such as those inside a standard gas cylinder. Even for NO2 standards as low as 20 ppm, most of the gas is in the tetroxide form while in the cylinder. N2O4 does not react on the gold electrochemical sensor, and it has different spectroscopic properties than the monomer. When drawing a sample from a cylinder of standard NO2, a sample bag should be used, and the gas should be allowed to dissociate for at least 15 minutes before use. Dynamic dilution systems are not a good idea for nitrogen dioxide.
Ozone is one of the most reactive of the common gases. When low concentrations are passed through new equipment or tubing, most of the ozone can be consumed in destroying the dirt, dust, and oils that contaminate all new materials. All systems that are used for low concentrations of ozone should be scrubbed for hours at an elevated ozone concentration before use. This includes sensors, too, unless the manufacturer indicates otherwise.
Ozone also reacts with many common plastics and elastomers. Ambient concentrations of ozone can be seriously depleted from the sample stream by a single rubber gasket or a few inches of vinyl or latex tubing. Pumps with Neoprene (tm) or Buna-N diaphragms are similarly lethal to ozone.
Volatile Organic Compounds (VOCs) are important workplace hazards, with TWA's ranging from 1000 PPM for ethanol and acetone, down to 1 PPM for benzene. They can also be tricky to control at the lowest concentrations. Absorption by rubber, NeopreneTM, or many other elastomers can be serious. A sample of VOCs should never be passed through a diaphragm-type pump before it reaches the sensor. They are not destroyed when absorbed, but may be gradually released back into the gas stream, perhaps at a time when you are setting the baseline and expecting a stream of clean air.
This list of bad-tempered gases is not exclusive by any means. It is not meant to be. It is just a reminder that, no matter how clever or sophisticated the software and hardware design of your instrument or experiment, you can never forget the chemistry.