The advent of the transistor and the printed circuit board brought about a revolution in analytical chemistry of virtually unprecedented proportions. It could be argued that no period in scientific history witnessed so many changes in so short a time. The hour’s long Winkler method for determining the dissolved oxygen in a water sample has been reduced to mere seconds and now consists of nothing more than inserting a probe into the sample. The size of the sample necessary for an electronic analysis has also been reduced dramatically to the point where, in some cases, a few drops on a sheet of filter paper are all that is required. This fact has not been lost on those afflicted with diabetes who prick their fingers to test their blood sugar several times a day. Not so long ago, a “portable field laboratory” meant a van or trailer complete with electricity and running water. The wait time for some analysis was measured in hours or, in some cases, days. Now, a good sized suitcase can contain a complete analytical set up for a variety of applications and turn out results in a matter of minutes with a concomitant reduction in costs; all due to the versatility of that little metal or plastic thing with a couple of wires sticking out that is smaller than a grain of corn and is the very heart of electronic instrumentation. The advent of the transistor and the printed circuit has not only made analysis easier and quicker but they have made it cheaper as well.
It’s easy to believe that the electronic age solved all things for all people and that a complete analysis of any specimen is just a button press away. Not so fast my friends, the advent of “black box chemistry” is a Pandora’s Box and it brings up about as many questions as it answers. To get reliable data from an electronic instrument one must understand both the chemistry and the electronics behind the analysis.
Back in “the good ole days” when we did chemistry with funnels, filter paper, test tubes and a class S chain-o-matic balance, the analyst understood these things and took them into account when writing his report. The analyst holding the black box may not have any real idea of what the read-out on the screen actually means in terms of actual conditions at the test site.
The first question that the analyst should answer is; what are we actually measuring? Now at first blush this might seem a bit ridiculous but it really is significant. A case in point involves the venerable old explosimeter made by MSA. Now this workhorse, one of the first electric (not really “electronic”) instruments on the market, is intended to measure the concentration of flammable gases or vapors in the atmosphere and it has been around since Noah wore short pants. It is simple to use, easy to service and a glutton for the punishment inherent to field use. This “black box” consists of a platinum filament connected as part of a Wheatstone bridge and powered by electric current from 6 “D” cells. As a safety precaution, the test chamber is, covered by a wire gauze to prevent any possible propagation of flame outside the instrument. The filament is heated by the electric current to initiate a catalytic reaction with any flammable organic substance in the sample. As the sample is drawn into the instrument by the aspirator bulb it comes in contact with the hot platinum. The reaction (burning) of the flammable portion causes the temperature of the filament, and therefore its electrical resistance, to change. This causes a unbalanced condition in the Wheatstone bridge which is reflected in the meter reading.
The instrument is calibrated by means of a calibration gas which is usually a mixture of methane and air.
But this is an indirect measurement. We are not really measuring methane but the change in electrical resistance caused by the heating effect of the methane being burned on the surface of the platinum element. Since the amount of heat produced by a given amount of methane is a stoichiometric property it can be equated to the amount of methane present in the sampling stream.
While this instrument is widely known for its reliability there are some things of which the average operator should be aware. In the first place measurements are based on the temperature of the platinum filament. Extreme differences in the temperature at the time of calibration and the time of use (as when an instrument calibrated in July is stored until December) could cause erroneous readings since anything that effects the temperature of the filament will have an effect on the readings produced.
I once encountered a welder who swore that his explosimeter was sensitive to argon. He claimed that he used his unit to determine when a pipe or vessel was fully inerted with argon prior to his starting to weld. Since it is well known that argon is an inert gas, I doubted his story and he offered to prove me wrong. We set up a pipe and took an explosimeter reading then we introduced argon at one end of the pipe and when it reached the explosimeter probe the instrument indicated a sharp dip in the reading. Why? The answer lies in the difference between the thermal conductivity of argon and the thermal conductivity of air. Any gas with a different thermal conductivity could have done the same thing. Thus, while this is a useful way to indicate the filling of a space with a gas other than air, it may not be totally reliable as a means of verifying that the space is filled with argon and is therefore inerted.
Another thing that the operator should verify is that the instrument is functioning. Yes, it reads “zero” before we go in, but will it give a defining readout in the presence of the designated analyte? Old timers used to check this with un-lighted butane lighter, it wasn’t quantitative but it did let the operator know that the batteries were good and the platinum element was functional. It is to be noted also that in the past some operators engaged in the practice of checking their explosimeters against the fumes at the opening of their vehicle’s fuel tank. The lead in the gasoline “poisoned” the filament which would render the element useless. Now, with the advent of unleaded gasoline with 10 percent ethanol this is not as critical but it is still not a good idea. It is also important to use, if possible, a standard that contains the actual analyte in question. If we are checking for LPG then we should calibrate with LPG. If we are checking for natural gas (methane) then calibrate with methane and so on. It is also worth remembering that not all “explosimeters” (here used as a generic term) are equal. Some instruments use the platinum filament but others use a ceramic bead. There are several of these available. These distinct types of instruments behave very differently especially when challenged with vapors or fumes from slightly heavier petroleum derivatives such as gasoline and diesel fuel, to which some types are “blind.” In some instances the addition of ethanol to motor fuel has a major effect on analytical detectors since ethanol is water soluble and it is definitely hygroscopic. It is a good idea to check it out before you go into a dangerous situation.
Most electronic monitoring instruments are designed to operate in an atmosphere composed of air and one or more contaminates or analytes. This means that air is as important to the test as the analyte. When an explosimeter probe is inserted into an atmosphere that is above the flammable limit i.e. too “rich” to burn, the meter will first read 100 percent and then, as the air inside the instrument is consumed, drop, sometimes all the way to zero. This gives rise to the convention of reporting readings of “one hundred and xx percent.” This does not mean a reading of 1xx percent; it means that the meter read 100 percent and then dropped down to xx percent. There is actually a fire (of microscopic proportions) on the surface of the platinum and a fire will go out from a lack of air (oxygen) just as surely as it will from lack of fuel. Don’t forget this.
Many years ago we had a training film showing the operation of the Davey safety lamp (which, by the way is still legal in some states and in the US Navy). In this particular film, the not so bright deck hand was trying to check the atmosphere of a shipboard tank by lowering the safety lamp into the hold. The atmosphere contained petroleum vapors greatly in excess of the flammable range. The light went out and when it was withdrawn the deckhand sent his partner to “supply” to get “a good one”. This was repeated again before a rating stepped in and explained what was happening. Take-home lesson, know your instrument and how it can be expected to respond to extreme conditions. The fact that the lamp went out should have alerted the deckhand to the fact that the tank was filled with vapors to the point that there was not enough oxygen present to sustain a fire or a life. The Davey safety lamp was not defective, the operator was.
Once we have acquired an instrument we need to know what it is really telling us when it reads “XY percent.” If the “black box’ says that it is safe to enter a space can we depend on it? To answer this question we need to do all of several things. The first is standardization or “calibration”. To do this we challenge the instrument with a defined concentration of the analyte. In the case of an explosimiter or other atmospheric contaminate meter this would be a calibration gas with a “defined” (known) concentration of a flammable gas or vapor, commonly methane. For a pH meter it would be a buffer solution of a known pH. In some cases it may even be a “synthetic” standard which is actually a different substance from the analyte for which we are assaying. A case in point is the automated analysis for ethyl alcohol; in at least one protocol the standard is actually sodium bicarbonate (baking soda). This is due to the fact that alcohol is soluble, hygroscopic and very volatile. Therefore a standardized solution is almost impossible to maintain. In any case a “Standard” is defined by the concentration of the desired analyte it contains or represents. For example: we have a calibration gas that is known to contain 61percent LEL (lower explosive limit) of methane. We challenge our instrument with this gas and adjust it until it reads 61percent. When this happens the instrument is said to be “standardized”. Now we need to check the validity of our calibration. We do this by presenting a second gas of a known concentration, say 89 percent LEL, and see if the instrument reads 89 percent (within reasonable tolerances and for a direct reading instrument). If it does then the instrument is said to be “calibrated” within the range between 61 percent and 89 percent of the LEL. If it does not, adjustment or maintenance is required. Now, it is a fact of life that the response of electronic (and chemical) detectors and indicators is linear over only a relatively small segment of the response range. In the scenario presented above we assume that the instrument responds in a linear manner between 61 and 89 percent LEL. If we need to assay samples that are higher than these limits then we can dilute the sample and multiply the result by the appropriate dilution factor or we can calibrate the instrument to a higher level, say 95 percent LEL.
Linearity of response can be verified by assaying a number of samples of known concentration (for example: 61 percent, 66 percent, 71 percent, 81 percent and 86 percent) and graphing the results. Even if the instrument is not responding in a linear manner results can be obtained from this graph. It should also be remembered that the more points used to establish the graph the more accurate the readings from unknowns will be.
Any incident response of significance will almost certainly generate litigation. Having said this, the purchaser of any instrumentation will be well advised to spend the few extra dollars required to acquire instrumentation that is self documenting. This can either be via a printout, a computer or a “smart phone.” In any case it obviates the need of trying to record field notes while dressed in protective gear and also the task of transcribing them after the incident. It also obviates the possibility of errors in readings, a point that will make a session on the witness stand much less stressful, for the witness at any rate.
From the foregoing it should be obvious that the procurement and deployment of electronic test equipment is not as simple as the manufacturer’s ads would have one to believe. There is far more to it than simply sending the office boy down to the nearest supply house to pick one up. Support for the instrument is vital. Quality fresh reagents, standards and controls need to be readily available. Repair parts and service are a must and a provision for a back up in case an instrument is out of service needs to be implemented. Periodic maintenance is also required especially in the case of instruments which employ a wet cell chemical reaction as a detector. In the desert southwest these wet cells often dry out and become useless long before their expiration dates. If this type of instrument is to be employed then there needs to be some arrangement for recharge or replacement of the wet cells. This is a point all too often neglected until an instrument fails to work at the site on an incident. “An ounce of prevention is worth a pound of cure” is a particularly valid comment here.
It is also imperative that those using the instrumentation as well as those evaluating the reports be trained in what the instrument can and cannot do. They need to understand exactly what a reading of 25 percent LEL or a pH of 3.0 actually means in terms of the response protocols underway.
All this means training not only for the operating personnel but the evaluators as well. This is not cheap but it is not nearly as expensive as the loss of a litigated case or a death or injury due to failure to appreciate the warnings emanating from the “black box”.
Electronic instrumentation is a valuable tool both from the technical and economic viewpoints but they often speak in unknown tongues. To realize their benefits those who use them must learn their language.
The author can be scheduled for contract training to help assure your instruments are prepared to give you accurate information and that personnel understand readings.