Posted on 10th Dec 2013 @ 1:35 AM

Sewage, Energy and Utilities work - and mineral production - especially in confined spaces - exposes people to a variety of toxic and explosive gases.

Fast, accurate measurement of dangerous gases is vital. Here I take a look at the methods of detection for flammable gases and some of their limitations.

Today, hand-held gas detectors use three, main sensor types for detecting explosive gases: galvanic sensors (limited to chemically reactive gases, like chlorine), catalytic bead sensors and infrared sensors. Laboratories will use other techniques, in larger instruments; such as Mass Spectrometry and gas Chromatography.

Galvanic sensors are electro-chemical systems, whereby an anode and a cathode, placed in a passive electrolyte cell, are connected to the sensor circuits of the detector. There is normally no potential difference between the electrodes but, when a ‘target’ gas is detected (by passage through a semi-permeable membrane, comprising the face surface of the sensor cell) the electrolyte becomes active and a potential difference is created, amplified and compared to that given by known concentrations of the target gas. Usually, the concentration of gas is proportional to the voltage produced and therefore, linear calibration is made.

Flammable gases detected by ‘galvanic’ detector cells include: Ammonia, Hydrogen Sulphide, Carbon Monoxide, Chlorine and most gaseous reducing agents and aerosols.

A problem with this technique: of using what, in effect, is a very tiny battery, just waiting for its electrolyte to make it ‘live’: is that the ‘battery’ will respond to a number of cross-sensitive chemicals. Therefore, if, say, I am using a Hydrogen Sulphide detector, I must be aware that it will respond to ANY Sulphurous gas, to Chlorine and to Carbon Monoxide, for example.

Manufacturers provide a table of cross-sensitivities, with their instruments; so that an approximation of other gas concentrations might thus be estimated.

The ‘Up-side’ of this problem is that no gas to which the cell responds, represents something that is healthy to breathe and therefore, the ‘false-alarm’ represents a reason why there should, in fact be an alarm or some concern about the surrounding atmosphere!

Both pellistor and infra red detector cells, are mainly used for alkane detection, such as Methane and LPG, but, in the case of the pellistor; this works by catalysing oxidation of the sample – in other words burning it – and then comparing the resistance of its ‘hot-wire’ with the known scale of temperature change, as the wire warms or cools.

Years ago, the first, portable detectors employed a large battery, a hot wire in a heat-sink cage (so we did not set fire to the entire atmosphere!) and a Wheat-stone bridge circuit, linked to a galvanometer.

Thus, we might determine the concentration of Methane, in a mine shaft, more accurately than by using canaries and observing the change in flame colour, from yellow to blue, in a DAVY LAMP, as the Methane became richer. So, then: no change in principle but a change in the techniques, reliability and efficiency of the systems we use today.

Catalytic Bead, or “pellistor”, sensors were developed forty years ago, as a less energy consumptive, more accurate means of FLAMMABLE gas detection than were Davy lamps or ‘hot-wired, Wheatstone Bridges’.

These sensors feature a Palladium catalyst, mounted in a bead of alumina, fixed to a tiny coil of Platinum wire (usually!). Flammable gas oxidises on contact with the catalyst, causing the temperature of the bead to soar, which in turn causes the resistance of the wire to increase.  The change in resistance directly with the partial pressure of combustible gas mixture present – using exactly the same ‘Wheatstone’ comparator circuit we used before – but, this time, full of cell-phone-technology-parts, rather than stuff we once used to repair with a soldering iron.

Such sensor heads respond differently to different types of combustible gases. This is for several reasons; the most significant being that longer chain hydrocarbon molecules are characterised by higher heats of combustion and therefore, differing temperature affects on the wire.

For example, it is harder to set fire to Butane than Methane – and even harder to light a candle (whose alkane chain length will be between seventy and a hundred). BUT, of the three, when burning, the candle offers the most heat per molecule and Methane the least. Have you ever noticed how ‘town gas’ takes much longer to boil a kettle than does Butane, when you’re camping?

These differences; of molecular chain length and reactivity with Oxygen also influence both the lower and upper explosive limits of gas mixtures, the liquid point and auto-ignition temperature of the gas.

Also rates of diffusion affects both the speed of detector response. Hydrogen, for example, diffuses quickly, ignites very easily and burns explosively but, in burning, it generates less heat than Butane and most other alkanes.

A further influence is; how readily the Palladium, or other catalyst, will accelerate oxidisation in a particular gas. A catalyst, remember, has no influence on a chemical reaction apart from speeding it up, or reducing the pressure and temperature at which the reaction  can take place.

Thus, there are several reasons why close relatives, Methane and Propane, will each cause a differing response in the detection cell and as to why different calibrations are needed, therefore.

There have been vast developments in the electronics side of gas detectors; so reducing power consumption, size and weight  - and improving reading accuracy and circuit response times but most of the improvements to Pellistors, have been focused on catalyst development, finer wires and reduction in bead size to further reduce power consumption.  

Pellistors, although not directly involved in the oxidisation/reaction, have limited life times. The catalyst can become ineffective, its surfaces becoming partly oxidised and chemically obscured – leading to consequent reduction in vigour and reduced sensitivity.

Typically, catalytic bead sensors have a useful lifespan of around five years, before calibration becomes unreliable or impossible. Pellistor sensors usually consume about two-hundred milliwatts, whereas galvanic cells used almost no current and little power, causing consumption only when a target gas is detected. Pellistors are therefore the limiting factor in battery times, for most, multi-gas detectors.

Pellistor cells require a case, made from sintered metal, as a heat sink, to prevent any chance of reaction heat igniting the surrounding atmosphere.

Used in an environment where smokes and aerosols are present – especially things like diesel exhaust – pellistors are easily ‘poisoned’ by aerosol condensate collecting on the bead.

In the case of condensing oil vapour the sensor will then forever, register a ‘flammable’ gas, refuse to be calibrated or otherwise prove unreliable. It will need to be expensively replaced and the machine recalibrated.  Cells so damaged cannot be cleaned or repaired.

Infra-red detectors have been made for many years, the first of them using infra-red spectroscopy to determine the presence of gas in a method not dissimilar from the Doppler principles, used in an echo-sounder but here, instead, measuring absorption and refraction, rather than reflection and interferrence.

The problem with using these was that lots of heavy, complex equipment was involved: and that, where theoretically, a very accurate measurement method, one had to know that the target gas was present because the ‘image’ showing the presence of gas and its density relied on qualitative certainty of the gas composition. Where this was unknown, other, additional techniques had to be used. But, for known gases, the method was almost as reliable as Mass Spectrometry and a fiftieth of the price.

More recently, small, Non-Dispersive, Infrared Sensors have been developed – an alternative to Pellistors – albeit, for the moment, more expensive and not reliable for very small partial pressures of most, flammable target gases.

There are three main types of non-dispersive infrared spectrometers: Total absorption; negative filter; and positive filter. Total absorption analysers have no selectivity towards any gas and measure only total absorbance of infrared radiation. Negative filter analysers offer selectivity by removal of a specific spectral region. In contrast positive filter analysers offer selectivity by using a detector containing the infrared absorbing target gas.

Most NDIR, personal gas detector instruments use two infrared sources and the difference in intensity of the two is measured and compared. There is a reference cell containing a non absorbing gas, a sample cell containing a sample of target gases and a detector.

Energy from each infrared source passes through the reference and sample cells to the detector. When the sample cell is either evacuated, or filled with an inert gas; the sample beam radiation reaching the detector is the same as the reference beam radiation.

When the sample cell contains a sample of the target gas, radiation is absorbed, reducing the radiation reaching the detector. The difference in signal received from the two beams is measured by the detector and is related to the amount of absorbing gas in the detector cell.

Infrared sensors are built in to a containment gas cell, through which infrared radiation passes. Sample gas molecules in the cell absorb known ranges of IR frequency, at atomic bonds between dissimilar atoms, such as Carbon and Hydrogen.

The telling difference in RECEIVED radiation, measured by the reference and active detectors, indicates absorption by the target gas and therefore, the density, or partial pressure, of target gas present. Each gas has a characteristic wave length absorption band and, as years ago, without using filtering techniques and specific emission wavelengths, it would be easy to confuse one gas with another.

More recent advances in NDIR sensors have dispensed with incandescent radiation sources, instead using less power-hungry, photodiodes - similar to LEDs. Such application potentially reduces power consumption 100mW to less than 5mW.   It also increases the useful life of the source. But it is very recent that these have been reliably used – let’s face it – how long will it be before everyone trusts LED kitchen lights to replace their power-station-wrecking ‘down-lighters’?

These days, the infra-red radiation source is a doped, Photo-diode cell. I do not mean that the little light has been on a bender; doping in this case, means adding a material that filters the wave band transmitted.

These cells still tend to be power consumptive and Non-Dispersive-Infra-Red detectors (NDIR) continue limited by battery capacity - especially as they are far more accurate in measurement when working at higher emission powers! There is reference to this aspect, published on the HSE web site, where the result of substantive research has been published.

Alike all, modern gas detectors, NDIR sensors are made intrinsically safe. Detected gases are not chemically changed and the cells are not so prone to ‘poisoning’ or aerosol contamination.

In effect; Pellistor cells measure the flammability and heat of combustion in a gas mixture – its reactivity, in other words: whereas, NDIR sensors reflect the partial pressure of an assumed target gas sample and may be reliably used in reduced-oxygen or even, completely anoxic environments.

The radiation sources and infra-red sensors are usually calibrated for measuring absorption in Methane hydrocarbon bonds at wavelengths around 3.4 microns, to which the gas is responsive; in an instrument designed generally sensitivity in a range between 2 and 15 microns.

NDIR instruments used for this type monitoring application are typically limited to the detection of higher concentrations of hydrocarbons and carbon monoxide (1% and above by volume).

Carbon dioxide absorbs infrared radiation very strongly: it would not otherwise be a ‘Greenhouse gas’: and many monitors are available to detect carbon dioxide in concentration ranges of 0.1% and higher.

NDIR detectors are not well able to differentiate between dioxides of Carbon, Nitrogen, and Sulphur, in low concentration. Also, the IR signal needs modulation, or ‘chopping’ so that the IR receivers are able to eliminate back-ground and incidental radiation, from heat sources.

In garnering accuracy and focus, it is imperative that a very narrow wave band is used to irradiate the gas molecules, so that absorption measured is certain to have resulted from target gas presence.

Although the technique may, theoretically, be used to assess many gases, remember that Atomic species and homo-nuclear diatomics do not meet the criteria for infrared-active molecules – so, the technique is no good for measuring air gases, rare gases and most inorganic gases, that have no dipolar function – to simplify this, for non-chemists – it is no good using a torch to find a sheet of glass under water!

Therefore, gases such as Hydrogen, are not be detected by this method. Likewise; NDIR sensors are no good for detecting Acetylene, Instead, Hydrogen and Acetylene may be detected by galvanic cells, designed for Carbon Monoxide (CO), with reference to a cross-sensitivity chart – or by splashing out on gas chromatography equipment!

Unless you live by an active volcano, on a star, or near a super-nova, Hydrogen does not exist naturally, being too reactive – so if you don’t work at a power station, in a battery factory, or a chemical works; such limitations are probably irrelevant to you.

NDIR sensors are more responsive to larger hydrocarbons than to smaller ones. Bigger molecules absorb more energy. For this, same reason, Pellistors are less sensitive to big molecules because the cell is a tiny ignition source and bigger alkane molecules need more heat of ignition and a higher temperature.
Try setting light to a steel Brillo pad with a match. It’s easy. Now try setting fire to a steel nail, with your match. I bet you can’t – especially if you set fire to the Brillo pad in the house, which you therefore burned down, before finding a nail!

There is much push and shove, by manufacturers to sell NDIR gas detectors – the latest generation – today’s technology and so forth. Take no notice. Any responsible scientist will spell out for you, the pros and cons of what is available and, given they are all imperfect methods, which of their imperfections will least endanger you!