Question:

We manufacture an ozone based air and water purification system.  We need to determine the concentration of any bacteria that remain after the purification process.  We have been using a luminometer to determine the bacterial count, and verify the effectiveness of our system on solid surfaces, as well as in water.  However, it is very difficult to determine the concentration of bacteria present in the air.

Bacterial respiration produces a mixture of methane, CO2 and volatile organic gases. Is it possible to measure the waste products produced by bacterial respiration, and use this as an indicator?

We were wondering whether we might be able to use a photoionization detector (PID) for this purpose.  When we spoke with one of the other leading PID manufacturers we were told this probably would not work.  What do you think?

Answer:

I agree with the answer you received from that “other” PID manufacturer.  Direct reading instruments can be used to measure by-products of microbial respiration, but it takes a lot of microbes to produce a meaningful change in concentration.  In order to produce contaminants microbes have to be actively metabolizing.  Most of the bacteria present in the air are in the form of dormant spores, and are not actively respiring.  Spores are smaller, lighter, and remain suspended in the air for a much longer period.  You can easily culture any viable spores that remain after sterilization, but measuring atmospheric contaminants produced by the microbes while they are still in the air is next to impossible. Even when they are present on solid surfaces or in water, the bacteria have to be actively respiring in order to produce detectable metabolic by-products.

In a confined space, where there is no mixing with fresh air, microbial decomposition can easily create hazardous atmospheric conditions.

There are many different types of bacteria and microbes involved in this process.  Some types of “aerobic” microbes use oxygen, and produce carbon dioxide.  Other types of “anaerobic” bacteria that do not use oxygen produce methane and hydrogen sulfide.  Which types of bacteria are active at any moment depends on the type of organic material that is present in the confined space, the oxygen concentration in the space at that time, and other environmental conditions such as humidity and temperature.

The effects of microbial decomposition on the atmosphere in the space often (but not always) follow the same sequence.  Aerobic respiration, which utilizes oxygen, is the most efficient way to convert organic material into energy.  That’s why human beings are aerobic organisms that require oxygen.  When not actively metabolizing, bacteria and microbes are present in the form of dormant spores.  The still atmosphere in a confined space initially contains plenty of oxygen.  These early conditions are good for aerobic decomposition.  Oxygen using bacteria and microbes become active, and begin to proliferate.  Aerobic bacteria deplete the oxygen, and generate CO₂.  Being much heavier than fresh air, the CO₂ tends to accumulate in the bottom of the space, creating locally anaerobic conditions.  Anaerobic bacteria remain in the form of inactive spores until conditions become agreeable for their metabolism.  As the atmosphere becomes increasingly oxygen deficient, anaerobic microbes germinate and begin to metabolize.

Anaerobic microbes do not require oxygen.  Anaerobic decomposition is less efficient, and proceeds more slowly than aerobic decomposition.  The metabolic byproducts of anaerobic respiration include methane (CH₄) and hydrogen sulfide (if the organic material in the space includes sulfur).  The more sulfur the organic material in the space contains, the greater the concentration of H₂S that is likely to be produced by anaerobic bacterial action.  Being heavier than air, the H₂S also tends to accumulate near the bottom of the space.  Methane, being lighter than air, tends to rise, and accumulates near the top of the space, or escapes from the space, if there are any openings.

I would suggest continuing to assess the bacterial count on solid surfaces and in the water by means of the luminometer.  Unless you leave the sterilized area alone for a lengthy period of time, you are unlikely to see anything going on in the air, even if the purified armosphere includes viable spores.  The spores have to germinate to have an effect on the atmosphere.

On the other hand, a prime application for the G450 and G460 is to monitor the atmosphere where microbial action can cause dangerous conditions.  Anaerobic fermentation is used to produce alcohol, wine, distilled spirits and beer.  It is highly associated with the presence of dangerous levels of CO₂, as well as oxygen deficiency.  Hydrogen sulfide is highly associated with sewage and wastewater treatment, oil production and refining, commercial fish and meat processing, and many other industrial applications.  Methane produced by microbial action is highly associated with many types of confined spaces, including sewers, manholes, digesters, vaults and tunnels.

So, while a few dormant spores may not cause a measurable change in the atmosphere, large numbers of actively metabolizing bacteria can rapidly produce deadly conditions!

Thank you for the question.

Question: I’ve spoken with a number of other gas detection suppliers. They always seem to run up against the same problem. We have both CO₂ and HF that is produced in our process. The other suppliers say their SO₂ sensor has a relative response to HF of 1:1. Is there anything we can do? Are your sensors any different than those of the other manufacturers?

Answer:

The answer is to use multiple sensors to rule in or rule out potential interfering gases.

CO₂ does not affect SO₂ or HF sensors.  We use an IR sensor to measure CO₂.  The IR CO₂ sensor is not affected by exposure to SO₂ or HF.

The relative response of the HF sensor to SO₂ is about 0.8.  This is the only situation where you have a high relative response between one and another of the gases on the customer’s list.

The relative response of the SO₂ sensor to HF is very low.  For short exposures to HF gas, the SO₂ sensor relative response should be close to zero.  If the internal filters become saturated due to high exposure, the HF should theoretically produce a slight negative interfering effect on the SO₂ sensor.  However, I did not get a response even when I exposed an SO₂ sensor to 10 ppm HCl gas (which we use as a surrogate for HF) for over 10 minutes.

Putting this information into a logic chart allows you to determine which gas is likely to be present:

Thus, in the event of simultaneous exposure to both SO₂ and HF, the true concentration of SO₂ is determined by the SO₂ reading.  The true concentration of HF is determined by subtracting the 80% of the value of the reading of the SO₂ sensor from the reading from the HF sensor.

Let’s say you have a reading of 20 ppm from the HF sensor, and a reading of 10 ppm from the SO₂ sensor.  The true concentration of SO2 = 10 ppm.  The true concentration of HF = (20 ppm) – (10 ppm X 0.8) = 12.0 ppm.

Relative responses vary between sensors, and can change a little over the life of the sensor.  When you use relative response factors to calculate the actual concentration remember that relative response factors are approximate values!

HF sensors require a little extra effort when it comes to calibration.  Also they usually need to be replaced after one year, when the warranty expires.

We have a Technical Note which discusses HF sensor calibration issues that is posted on our website www.goodforgas.com at the following link:

TN2014 Calibrating GfG instruments equipped with HF sensors 1-14-13 (pdf)

The note discusses HF sensor calibration from the standpoint of the Micro IV, but the issues are exactly the same when the HF sensor is installed in a G460 multi-sensor instrument.  It is likely that at least some of the customer’s prior negative experience with HF sensors had as much to do with calibration as with issues related to relative response.  Also, HF sensors are very prone to exhibiting short term fluctuations due to abrupt changes in humidity.  Breathing on the sensor, or putting your hand over the HF sensor causes the reading to fluctuate.  The reading stabilizes as soon as the humidity stabilizes.  Customers need to be careful to avoid abrupt changes in humidity, and to give the sensor time to stabilize when exposed to new conditions in the ambient atmosphere.

I hope this information was helpful.

Question: I have a customer who is using an electrochemical sensor for CO to comply with NIOSH Method 6604, however they are using Nitrogen instead of air as Zero Gas; they have Nitrogen for Gas Chromatography purposes. This Nitrogen has 99.9% purity. An ISO 17025 auditor told them that this could affect the calibration process for this sensor, so the customer is asking me if this is true.

Do you think that the Nitrogen could affect the sensitivity of CO for this electrochemical sensor? Or is this procedure correct?

Thanks in advance for your help!

Answer:

The first step in the calibration process is the “fresh air calibration” or “fresh air zero adjustment”.  In this step the instrument looks at the electrical output of the CO sensor while it is located in contaminant free fresh air, and adjusts the instrument reading to 0 ppm.  This provides a point of comparison when the sensor is exposed to atmosphere that contains carbon monoxide.  The electrical output of the sensor when it is exposed to CO is higher than the electrical output of the sensor in fresh air.  The difference in electrical output is proportional to the concentration.  Calibration gas that contains CO is used to adjust the reading of the sensor while it is exposed to gas.  This is referred to as the “span calibration” or “gas calibration” step.

The electrochemical reaction used to detect CO requires oxygen.  However, the sensor does not require or use O₂ during the fresh air zero adjustment step.

Given the tolerances and accuracy of the electrochemical CO sensors used in health and safety instruments, using nitrogen instead of fresh air to adjust the CO sensor should not materially affect the accuracy of the fresh air zero adjustment. However, it is always better to use zero contaminant fresh air, not pure nitrogen for this calibration step.

All electrochemical CO sensors use the same two part reaction to detect gas.  The first half-cell reaction occurs at the sensing electrode.  The second half-cell reaction occurs at the counter electrode.

Carbon monoxide is oxidized at the sensing electrode.  Water from the sensor electrolyte is used up during this first step. The electrolyte in the sensor is an aqueous solution of sulfuric acid.

2CO  +  2H₂O    →    2CO₂  +  4H⁺  +  4e-

The counter electrode acts to balance out the reaction at the sensing electrode by reducing oxygen present in the air to water.  The water that is consumed in the first step is regenerated in the second step.  The oxygen used in this step comes from O₂ dissolved in the electrolyte.  The oxygen dissolved in the electrolyte comes from the atmosphere in which the sensor is located.

O₂  +  4H⁺  +  4e-    →   4H₂O

When you add the two half-cell reactions together you get the overall reaction:

2CO + O₂      →    2CO₂

For every 2 molecules of CO you detect you use up one molecule of O₂.

Note that unless the sensor is in the presence of a high concentration of CO, very little O₂ is actually used by the sensor.  Even when the atmosphere is oxygen deficient there is usually more than enough oxygen in the air to replace O₂ consumed in the detection reaction.

Long term exposure to pure nitrogen eventually depletes the oxygen in the electrolyte.  If the oxygen is depleted, the sensor is unable to detect gas.  However, this takes quite a while to occur.  The electrochemical sensors used in GfG portable instruments are full sized, “4 Series” sensors.  Span calibrating or zero-adjusting a GfG electrochemical CO sensor with pure nitrogen, or using CO in nitrogen to span adjust the sensor will not materially affect the accuracy of readings.

The time it takes for the electrolyte to become oxygen depleted is affected by the volume of electrolyte in the sensor.  Miniaturized or low volume (flat) CO sensors that contain less electrolyte, (and less dissolved oxygen) can become O₂ depleted more quickly.

So, as a general rule, it is better when possible to use contaminant free fresh air to zero-adjust the sensor, and CO calibration gas that includes oxygen when a span calibration adjustment is performed.

Question: Can hydrogen have an effect on CO readings, and can downloading the data help determine what’s going on?

Answer:

The attached .txt file is a very cool data download from a G460 instrument (the file is posted at the following link: 06262015.txt ). The readings are from an instrument that belongs to Robert Lindley, the Eastern Zone Manager for GfG.  Robert was in his garage, getting ready to calibrate his instruments prior to leaving on a business trip for GfG.  When he turned his G460 on he was shocked to see high readings for carbon monoxide.  He left the garage, opened the garage door from the outside, and did not re-enter until the readings returned to fresh air values.  The response of the instrument provides a good example of conditions that can affect readings, and how to interpret the results.  The data-logging interval is set to 1.0 minute, which means our “G450 / G460 Data Viewer Software” is perfect for looking at the results.  If you don’t yet have a copy, the software is posted in a .zip file on our website at the following link:  https://goodforgas.com/wp-content/uploads/2015/05/DDS_G450-G460_v4395.zip

What caused the readings?

Robert has a golf cart with a faulty battery charger.  Instead of the charger shutting down or going into trickle when the battery finished charging, it was continuing to pump power into the battery at the full charging rate.

We have a blog entry on www.goodforgas.com that talks about what gases to look for in areas where lead acid batteries are being charged:  https://goodforgas.com/hazardous-gases-associated-lead-acid-battery-charging-stations/

During charging, (especially in the event of overcharging), lead acid batteries produce oxygen and hydrogen.  These gases are produced by the electrolysis of water from the aqueous solution of sulfuric acid.  Since the water is lost, the electrolyte can be depleted.  This is why you need to add water to non-sealed lead acid batteries.  When a lead acid battery cell “blows” or becomes incapable of being charged properly, the amount of hydrogen produced can increase catastrophically.

Hydrogen is not toxic, but at high concentrations is a highly explosive gas.  The 100% LEL concentration for hydrogen is 4.0% by volume.  At this concentration, all it takes is a source of ignition to cause an explosion.  Sparking from a battery terminal as it is connected or disconnected from the charging system is more than adequate as a source of ignition energy.  That’s why lead acid batteries should only be charged in well ventilated areas.

Why did the hydrogen have an effect on the CO readings?

This is the chemical reaction that is used in electrochemical CO sensors to detect gas:

CO Sensing Electrode Reaction:    2CO + 2H₂O   →    2CO₂ + 4H⁺ + 4e^–

CO Counter Electrode Reaction:    O₂ + 4H⁺ + 4e^–   →   2H₂O

In the first part of the detection reaction, the sensor consumes water from the electrolyte (the solution in the sensor) and generates hydrogen (H⁺).  In the second part of the reaction, the hydrogen reacts with oxygen in the electrolyte to produce water.  Because the CO detection reaction includes hydrogen, it’s easy for hydrogen in the ambient air to have an effect on readings.

At GfG we use CO sensors that are deliberately designed to limit the response to hydrogen.  However, they still show at least some response.

The following chart shows the response of a substance-specific, electrochemical hydrogen sensor to 1,000 ppm hydrogen calibration gas (the green colored response curve).  The readings of the H₂ sensor are very accurate when the sensor is exposed to gas.

The blue colored line shows the response of the CO channel of a COSH sensor to the same 1,000 ppm hydrogen gas.  COSH sensors simultaneously measure both CO and H₂S.  We use the “two-in-one” COSH sensor whenever we need to save space in the G460.  Using the COSH sensor means we have another sensor position available to measure a different gas.

COSH sensors have one sensing electrode for H₂S, and a separate sensing electrode for CO.  The sensor is split into two chambers.  H₂S is measured in the outer chamber, and CO is measured in the inner chamber.  The sensor is designed for the rapid diffusion of atmosphere that contains CO into the inner chamber.  The CO channel of the COSH sensor has a relative response to hydrogen of about 33%.  As you can see, when we exposed the COSH sensor to 1,000 ppm hydrogen, it produced a reading of 330 ppm on the CO channel.

We only use the COSH sensor when we need to save space.  When we don’t need to save space, we use a single-channel CO sensor that is specifically designed to have the lowest possible response to hydrogen.  The single-channel “hydrogen nulled” CO sensor is the standard sensor we use in all 4-gas G450 instruments.  Even though the sensor is designed to show the lowest possible response, it still shows a relative response to hydrogen of 56 ppm (= 5.6 %).

Robert’s results:

“Session 5” started when Robert turned the instrument at 8:28 AM on June 26, 2015.  The session ended when he turned the instrument back off at 10:18 PM on the same date.  In the following set of graphs, the actual CO readings are shown in blue, the 15-minute STEL calculation is shown in black, and the 8 –hour TWA calculation is shown in green.  The software lets you view or hide each of the graphs by clicking on the legend for the graph at the bottom of the screen.

Let’s start by hiding the STEL graph.

When you look at the real-time CO readings, the concentration was well above 0 ppm from the moment Robert turned the instrument on.  But at 10:50 AM something really interesting happened.  Robert had been moving the instrument around the garage, trying to find the source of the CO readings.  You can see the concentration at any given moment by moving the cursor over the graph to the point of interest.

At 10:50 AM Robert positioned the instrument close to the lead acid battery in the golf cart, and the readings immediately climbed to 127 ppm.  Given that the CO channel of the COSH sensor has a relative response to hydrogen of about 33%, the true concentration of hydrogen in Robert’s garage where the instrument was located at 10:50 AM was about 100/0.33 ≈ 300 ppm H₂.

The readings rapidly went up and down when Robert moved the instrument back and forth near the battery.  Because hydrogen is actually part of the electrochemical CO sensor detection reaction, the sensor responds and recovers very quickly.  This is different than the response of the sensor to other gases (such as acetylene) that can interfere with CO sensor readings.  It can take hours for the sensor to recover after exposure to acetylene.  In the case of hydrogen the response and recovery is very rapid, which is exactly what you see in the graph.

When you look at the table (instead of the graph) of the logged results for Session 5, you see something else that is very interesting.  In the “Table View” of the session, any intervals where the concentration exceeds the alarm settings are flagged in red.  The low (A1) alarm for CO was set at 35 ppm.  The STEL alarm was set at 50 ppm, and the TWA alarm was set at 35.  There are three intervals in the table (flagged in red) where the A1 alarm was exceeded.  There were no intervals where the STEL or TWA alarms were exceeded.

Now for the interesting part.  At the exact moment the CO reached 127 ppm, the H₂S reading started to rise, reaching a concentration of 0.1 ppm.

You can also see this when you look at the H₂S graph for Session 5:

This makes perfect sense, and helps confirm that the CO readings were actually due to the presence of hydrogen.  The hydrogen sulfide (H₂S) channel of the COSH sensor has a very small relative response to hydrogen of about 0.03%.  If the true concentration of hydrogen in the area of the instrument at 10:50 AM was about 300 ppm, you would expect the H₂S channel of the COSH sensor to read about 300 x .0003 ≈ 0.09 ppm H₂S.   The actual reading was 0.1 ppm, which is extremely close to the expected value.  Sweet!

Why didn’t the LEL sensor detect the presence of hydrogen?

Hydrogen is a flammable gas.  The lower explosive limit (LEL) is the minimum concentration of gas that can explode if a source of ignition is present.   The LEL concentration for hydrogen is 4.0% volume, (which is equal to 40,000 ppm).

The LEL alarm is normally set at 10.0% LEL.  Assuming the instrument is properly calibrated, it would take about 4,000 ppm hydrogen to activate the LEL alarm.  However, Robert’s instrument was calibrated to methane, which has a slightly different response.

Assuming the highest true concentration of hydrogen in Robert’s garage was about 300 ppm, it’s easy to see why the LEL sensor didn’t show a reading.  Even if the instrument had been correctly calibrated for hydrogen, the concentration in Robert’s garage was still less than 1.0% LEL.  The concentration was just too low to register on the LEL sensor.

What if I wanted to directly measure ppm hydrogen?

No worries!  All you need to do is install a 0 – 2,000 ppm range, (= 0 to 5% LEL) substance-specific hydrogen sensor in your G460.

Robert’s downloaded results illustrate how useful it can be to actually look at the graphs and tabular response of the sensors over time.   With a little bit of experience the shape of the curves can provide all sorts of additional critical information.

Thanks again to Robert for sending the download to my attention.

Bob Henderson

President

GfG Instrumentation, Inc.

Question: One of our Canadian friends in the Industrial Hygiene community sent us the following question:

A colleague of mine was talking the other day about the presence of H₂S in crude oil.  In July, 2013 we had a terrible rail accident involving crude oil that occurred in the town of Lac-Mégantic in Quebec.   I have heard no mention of H₂S in official reports about the accident.   Nevertheless, some unofficial data is indicating that H₂S may have been present in very high concentrations, perhaps even reaching the flammable limit of 4%.  You have mentioned that H₂S is a poison to catalytic combustible LEL sensors.  Am I correct in guessing the LEL sensor would not be dependable in an H₂S LEL environment?   Also, if H₂S is damaging to the sensor, what concentration of H₂S can the LEL sensor handle, and for how long?

Answer:

This is a great question.  I think the answer will make a good post for our gas detection blog on the company website ( www.goodforgas.com ).

It is true that H₂S is a potentially flammable gas.  The LEL concentration for H₂S is only 4.0%.  H₂S can easily reach 100% LEL in many confined space and oil industry settings, especially when the situation involves “sour gas” or “sour crude” that contains high concentrations of sulfur.

Exposure to high concentrations of H₂S can inhibit or poison catalytic LEL sensors.   To keep this from happening, most LEL sensors, especially the ones used in portable instruments, have an internal filter to remove the H₂S before it can reach and damage the active bead in the sensor.  This is why you can use multi-component calibration gas that includes both H₂S and LEL gas to calibrate the combustible gas sensor without causing damage to the sensor.

Catalytic LEL sensors detect gas by catalytically oxidizing or “burning” the gas on an active bead or “pellistor” located within the sensor.  The sensor contains two coils of fine platinum wire which are coated with a ceramic or porous alumina material to form beads. The beads are wired into opposing arms of a balanced Wheatstone Bridge electrical circuit. The “active” bead is treated with a platinum or palladium-based catalyst that facilitates the oxidation of combustible gas on the bead. A “reference” bead in the circuit that has not been treated with catalyst provides a comparison value.  As oxidation occurs the active bead is heated to a higher temperature. Since heating due to oxidation of the combustible gas only occurs on the active bead, the difference in temperature between the two beads is proportional to the concentration of gas in the area where the sensor is located.  Because the two beads are strung on opposite arms of the Wheatstone Bridge circuit, the difference in temperature between the beads is registered by the instrument as a change in electrical resistance.

The LEL sensor depends on the activity of the catalyst to detect gas.  If the catalyst is harmed by exposure to sensor poisons or inhibitors, the sensor’s ability to detect gas is affected.  In the case of sensor inhibitors the effects may be reversible, and the sensor may recover over time (at least to an extent).  In the case of virulent LEL sensor poisons like tetraethyl lead and vapors that contain silicone, the damage is rapid and irreversible.  H₂S functions both as a poison and as an inhibitor.  The sensor may recover some of its lost sensitivity over time, but once the bead has been exposed to H₂S, there will be permanent harm.

The internal filter used to remove H₂S has a capacity on the order of 20,000 ppm minutes.  Each time you expose the LEL sensor to gas that contains H₂S you use up a little of the remaining capacity of the filter.  For most instrument users the capacity  of the filter is sufficient to protect the LEL sensor for the entire expected life of the sensor.  Bump testing the instrument before each day’s use has very little effect on shortening the life of the LEL sensor.

During the bump test the sensors are exposed to known concentration test gas.  The concentrations used in the test gas need to be high enough to activate the alarms.  In Canada, CSA takes the requirements a step further.  Because LEL sensors can be poisoned or damaged by exposure to sensor poisons like H₂S, it’s not enough simply to activate the LEL alarms.  To pass the test CSA requires that the reading of the LEL sensor is between  “minus 0% and plus 20%” of the concentration of gas applied.  In other words, if you use 50% LEL gas to test the combustible sensor, to pass the CSA version test the readings must stabilize between 50% LEL and 60% LEL.

Typical “four gas” multi-gas atmospheric monitors include sensors used to measure O₂, LEL, CO and H₂S.  The gas typically used to perform bump checks on these meters is a multi-component mixture that is designed to test all of the sensors at the same time.  The gas (depending on the manufacturer) usually includes 20 ppm or 25 ppm H₂S.

It typically takes less than 30 seconds to perform a manual bump test.  Thirty-seconds of exposure to 20 ppm H₂S is the same thing as “10 ppm minutes” of exposure.  Since the capacity of the filter is 20,000 ppm minutes, it takes around 2,000 bump tests to saturate the filter.   Even if you bump test the instrument every day, five days a week, 50 weeks per year, it will still take over 8 years before you begin to cause damage to the LEL sensor due to the H₂S in the calibration gas.

Docking stations are quite a bit faster, and expose the sensors to less gas, which is one of the reasons gas detector manufacturers encourage their use.   They save money as well as time, improve the accuracy of the test results, and simplify documentation.

By definition, “sour” natural gas contains at least 4 ppm H₂S.  However, the concentration can be much higher.  The gas from one well in Canada is known to contain 90% hydrogen sulfide, while other wells are documented have H₂S in the tens of percent range.  Some oil industry customers have specialized procedures that can expose workers (while wearing SCBA and other protective equipment) to H₂S concentrations that can reach several thousand ppm.  Even a short period of use in high concentration H₂S can rapidly saturate the filter, and damage the LEL sensor.  Twenty minutes of continuous exposure to 1000 ppm H₂S would be enough to saturate the filter.  Once the filter is saturated, the active bead can be damaged very rapidly by further exposure.  This is one of the reasons it is so important to perform a bump test to verify the performance of the LEL sensor before each day’s use.

We normally expect (and warrant) GfG catalytic LEL sensors to last for at least three years.  However, some of our sour gas customers need to replace LEL sensors on a yearly basis.

It’s hard to predict in advance when (or if) the internal filter is close to saturation.  While damage to the bead is electronically detectable, saturation of the filter is not.

The important thing is to verify that the readings of the LEL sensor are accurate by testing or calibrating the sensor before each day’s use!  Sensors that fail a daily bump test need to be calibrated or replaced before further use.

Thanks again for a great question!

Question: I work at a fire department.  We recently ran into a problem in the area where we charge lead acid batteries that we use in our engines and boats.  One of the batteries became overheated while being recharged.  What gases or hazardous atmospheric conditions are associated with lead acid battery charging stations?  Can batteries release flammable gas?  If the batteries contain sulfuric acid, does that mean they can produce and release toxic H₂S?  We thought we smelled something.  What should we measure?

Answer:

Thanks for your very good question.

Lead acid batteries are used to power forklifts, carts and many other types of machinery in many industrial settings.  Many facilities have charging areas where multiple heavy duty lead acid batteries are recharged at the same time.  In some cases facilities maintain large banks of lead acid batteries that are used to provide backup power to critical systems during an emergency.  Fire engines, HAZMAT and emergency response vehicles frequently include banks of lead acid batteries for the same purpose.  Gases produced or released by the batteries while they are being charged can be a significant safety concern, especially when the batteries are located or charged in an enclosed or poorly ventilated area, or on the truck.

Sulfuric acid contains sulfur, and hydrogen sulfide (H₂S) is a possible by-product of over-charging and battery decomposition. If you smell the rotten egg odor of H₂S in the charging area, you should assume that this very dangerous gas is a possibility.  You should leave the area, and use a gas detecting instrument with an H₂S sensor to confirm whether the gas is present before returning.

However, H₂S is not the most common gas associated with charging or discharging lead acid batteries that contain sulfuric acid.  Given the over-heating and other problems you mentioned, you may not find H₂S, but you probably will find the presence of other atmospheric hazards when you test the atmosphere in the area with your gas detector.

The most common reaction byproducts associated with sulfuric acid (H₂SO₄) are hydrogen and sulfur dioxide.  Overcharging, or lead acid battery malfunctions can produce hydrogen.  In fact, if you look, there is almost always at least a little H₂ around in areas where lead batteries are being charged.

Overcharging, especially if the battery is old, heavily corroded or damaged can produce H₂S. Deteriorated, old or damaged lead acid batteries should be removed from service, as damaged batteries are much more likely to be associated with production of H₂S.

Sulfuric acid reacts with a number of metals and substances to produce SO₂ as well as other “sulfur oxides” (SOx) such as SO₃, SO₄, S₂O, etc.  Many sulfur oxides have a pungent odor, but they are NOT H₂S.  H₂S is a reduced sulfide, not an oxide.  When you have a spill, SO₂ is generally the most common gaseous sulfur reaction by-product.

During discharge of a lead acid battery you have the following two half-cell reactions.  Neither SO₂ or H₂S are normally produced, even by catastrophic discharge!

Negative plate reaction:

Pb (solid) + HSO₄^– (aqueous) → PbSO₄ (solid) + H⁺ (aqueous) + 2e⁻

Positive plate reaction:

PbO₂ (solid) + HSO₄^– (aqueous) + 3H⁺ (aqueous) + 2e⁻ → PbSO₄ (solid) + 2H₂O

The total reaction can be written as:

Pb (solid) + PbO₂ (solid) + 2H₂SO₄ (aqueous) → 2PbSO₄ (solid) + 2H₂O

During charging, (especially in the event of overcharging), lead acid batteries produce oxygen and hydrogen.  These gases are produced by the electrolysis of water from the aqueous solution of sulfuric acid.  Since the water is lost, the electrolyte can be depleted.  This is why you need to add water to “wet” (flooded type) non-sealed lead acid batteries.  When a lead acid battery cell “blows” or becomes incapable of being charged properly, the amount of hydrogen produced can increase catastrophically:

Water is oxidized at the negative anode:  2 H₂O (liquid) → O₂ (gas) + 4 H⁺ (aqueous) + 4e⁻

The protons (H⁺) produced at the anode are reduced at the positive cathode:  2 H⁺ (aqueous) + 2e⁻ → H₂

So, in an area where lead acid batteries are being charged, the first gas to measure is H₂.

Hydrogen is not toxic, but at high concentrations is a highly explosive gas.  The 100% LEL concentration for hydrogen is 4.0% by volume.  At this concentration, all it takes is a source of ignition to cause an explosion.  Sparking from a battery terminal as it is connected or disconnected from the charging system is more than adequate as a source of ignition energy.  That’s why lead acid batteries should only be charged in well ventilated areas.

The best way to measure hydrogen in an area where you are charging batteries is with a permanently installed monitoring system.  You can use a standard catalytic LEL sensor, or you can measure the hydrogen by means of a substance specific electrochemical sensor.  The sensor and housing need to be designed and certified for installation and use in hazardous locations characterized by the potential presence of combustible gas.  Since hydrogen is lighter than air, H₂ sensors are usually mounted to the wall or ceiling at a height at least slightly above the source of gas.

Readings from the H₂ sensors can displayed right where the sensor is located, or on a remotely located controller or monitor.   Readings from the sensors can be used to activate relays, fans or alarms, or the information can be transmitted and integrated into the facility’s overall environmental health and safety and fire detection systems.

For LEL range measurement, using a standard catalytic combustible gas (CC) sensor with a range of 0 – 100% LEL is a good approach.  For situations where you need to take action at a lower concentration, using an electrochemical (EC) toxic gas sensor to measure the hydrogen may be a better approach.  The typical range for an EC hydrogen sensor is 0 – 2,000 ppm.  (This is equivalent to a range of 0 – 5.0% LEL.)

In the event of a sulfuric acid spill, or where the sulfuric acid is coming into contact with metals and / or other materials, you may need to measure SO₂ as well.  Although this is normally not necessary in charging areas where the acid is fully contained in the batteries.

If you are concerned with aerosol droplets of sulfuric acid, you can directly measure H₂SO₄ as well.  Once again, this is not normally a concern in battery charging areas.

Deteriorated, old or damaged lead acid batteries should be removed from service, as damaged batteries are much more likely to be associated with leakage leading to the production of SO₂, or charging malfunction which could lead to the production of H₂S

I hope this info is helpful!

Question: Why is the ammonia in the calibration cylinder odorless? and is the concentration of gas used to calibrate gas dangerous?

Answer:

Some people can smell ammonia down to 0.04 ppm.  However, according to OSHA the odor threshold for most people is between 5 and 50 ppm.  We normally use 100 ppm NH₃ to calibrate Micro IV and G460 ammonia sensors.  One hundred ppm is above the odor threshold for most people, so you would think you could smell the gas.  But there are other factors to consider.

Ammonia has a desensitizing effect on the sense of smell for people who work around it for prolonged periods of time.  There is a noticeable odor of ammonia throughout the facility where the NH₃ Micro IV detectors are being used.  Because of this on-going exposure, many people in the facility may be partly, and some may be almost completely desensitized to the odor of the gas.  For these people it can take a concentration higher than 100 ppm for the odor to be detectable.

However, a more important factor is the dilution effect when the ammonia is released into the air.

A cylinder of 100 ppm ammonia calibration gas doesn’t contain very much ammonia.  Remember that it takes 10,000 ppm to equal 1%.  Thus 100 ppm = 0.01% by volume.  When you release this already low concentration into the air it is rapidly diluted to a level well below the odor threshold.

One of the instructional GfG webinars we offered last year, “Gas detection myths and misconceptions”, included the following slide.  You can see in my example that even when you release the NH₃ calibration gas in a small enclosed space like the interior of a passenger vehicle, because of the dilution effect, the concentration is below the odor threshold.

Thank you for asking this question.

Question: I have a customer who uses silicon tetrachloride and germanium chloride in his production process. If the gases are released in air they form a dangerous chloride gas. He has an exhaust chamber where the gases could potentially be present, and needs a portable instrument he can use to sample the chamber before opening the door. Are we able to detect or measure these gases?

Answer:

Both silicon trichloride (SiCl4) and germanium chlorides (GeCl4 and GeCl2) can be detected by using a hydrochloric acid (HCl) sensor.

When these highly reactive gases are released into the atmosphere, they react with the humidity in the air to produce hydrogen chloride (HCl) gas.

SiCl4 + 2 H2O → SiO2 + 4 HCl

GeCl4 + 2 H2O → GeO2 + 4 HCl

Besides being able to detect the original gas, we are also able to detect the primary HCl reaction byproduct when it hits the air.

HCl is an extremely toxic and corrosive gas, with a TLV® (Ceiling) of 2.0 ppm. The OSHA PEL and NIOSH REL exposure limit (Ceiling) is 5.0 ppm.

The other byproducts, silicon dioxide (SiO2) and germanium oxide (GeO2), are normally stable, and in moderate concentrations, not very toxic. It’s the HCl that is the big concern.

GfG technical note TN 2015, “Electrochemical (EC) sensors: gases measured, ranges and resolution,” provides a complete list of the EC sensors that are available for use with GfG products, as well as optional ranges and resolutions (Table 1). The note also includes a list of additional gases that are reliably detectable by means of our standard available sensors based on relative response (Table 2). Both germanium chloride and silicon tetrachloride are on the list.

You can use the HCl sensor to detect a number of other common semi-conductor gases. The list in Table 2 is actually pretty extensive. For the chloride gases we generally use the HCl sensor. For the fluoride gases we use the HF sensor.

GfG has the ability to support many different sensors. In addition, we are often able to customize the range and resolution in order to optimize the sensor for a specific application. Not all of the available sensors, ranges and resolutions are listed in our data sheets, nor even (in some cases) in our price list. We just don’t have the space, and we don’t want to add additional complexity to the product ordering structure.

Technical note 2015 provides a more complete listing of our EC sensor options. The note is posted on www.goodforgas.com at the following link:

https://goodforgas.com/wp-content/uploads/2013/12/TN2015_EC_sensor_types_gases_ranges_and_resolution_08_07_13.pdf

The GfG Customer Service Department can provide a special part number, and add the additional configuration details when you place the order.

One word of caution about these particular gases. Silicon and germanium are both catalytic LEL sensor poisons. If you need to include an LEL sensor in the same instrument as the HCl sensor used to measure these gases, you should consider using our infrared (IR) LEL sensor. IR LEL sensors are not harmed by exposure to sensor poisons. Also, since hydrogen (H2) is almost always a concern in semi-conductor applications, and hydrogen cannot be measured with an IR LEL sensor, you should also include an electrochemical hydrogen sensor in the same instrument.

Thanks again for asking the question!

Question: Why do the readings for H₂S and NO₂ go up when I breathe on the sensors in my G460?  I notice they settle back down pretty quickly, but why do they go up in the first place?

Answer:

Thanks very much for sending us your question.

Electrochemical sensors respond to sudden changes in humidity.  They rapidly stabilize at the new humidity, but it can take a few seconds.  This is true even for H₂S sensors.  The air in your exhaled breath is very humid (close to 100% RH) and very warm (98.6°F / 37°C).  If you select the “PEAK” hold mode, then breathe on the H₂S sensor in a G460, the H₂S reading can climb as high as one or two ppm before it settles back down. However, this happens very quickly!  If you don’t use the “PEAK” hold button to see the highest reading that is momentarily reached, the sensor stabilizes so quickly you probably won’t notice there was a change in readings.

Reducing gas sensors, (like H₂S and CO sensors),  use oxygen dissolved in the electrolyte to detect the gas being measured.  The oxygen in the electrolyte comes from the atmosphere.  Water is used and regenerated during the detection reaction, but the overall reaction does not consume water.  The only thing that is used up is oxygen, and the gas that is detected.  The overall detection reaction used in an H₂S sensor converts hydrogen sulfide into sulfuric acid:

H₂S + 2O₂ → H₂ SO₄

H₂S sensors exhibit only small transients when the concentration of water in the atmosphere changes, and recover very rapidly.

Oxidizing gas sensors, (like NO₂, NO, Cl₂, ClO₂ and O₃), use water from the electrolyte to detect gas.  The water in the electrolyte that is used up in the detection reaction is replaced with water from the atmosphere.  The overall detection reaction used in an NO₂ sensor converts nitrogen dioxide into nitric acid:

NO₂ + H₂O → H₂NO₃

Because water is more directly involved in the detection reaction, breathing on oxidizing gas sensors (like NO₂) can cause a bigger spike in readings.

Both reducing gas and oxidizing gas sensors stabilize in the new humidity as soon as it stops changing. However, it can take oxidizing gas sensors a slightly longer time to stabilize fully in the new humidity.

GfG Instrumentation mining industry customers use a lot of NO₂ sensors.  The sensors do very well in underground and high humidity situations.  However, please make sure that you don’t hold a sweaty palm over the sensor compartment, and try not deliberately breathe on the sensors when the instrument is in normal operation.  Also, if you move from an air conditioned low humidity area to a hot, high humidity area, give the sensors a minute to stabilize in the new humidity before you perform a fresh air zero.

Thanks again for the question.

Question: We recently had a hydrogen leak which affected one of the production units at our refinery. However, the hydrogen leak mostly affected the CO readings, and not the readings from the LEL sensors in the instruments we are currently using. We would pick up 0-5% LEL from time to time, but pretty constantly stay at about 50-250 ppm CO even though we knew that could not be right since CO can only be present as a byproduct of combustion from a heater stack or vehicle exhaust. The wind direction and proximity to either of those things that day would not have allowed such a CO reading to occur.

I guess my questions are, what causes cross sensitivity in a CO sensor, what is it sensitive to, and are there any data I could take a look at on the subject?

Answer:

Thanks for the great question! Sorry you are having problems due to the response of the CO sensors in your instruments to hydrogen.

Substance-specific electrochemical sensors are available for many of the most common toxic gases including hydrogen sulfide, carbon monoxide, sulfur dioxide, chlorine, chlorine dioxide, ammonia, phosphine, ethylene oxide, nitrogen dioxide, ozone and others. “EC” sensors are compact, require very little power, exhibit excellent linearity and repeatability, and generally have a long life span. The detection technique is very straightforward in concept. Gas that enters the sensor undergoes an electrochemical reaction that causes a change in the electrical output of the sensor. The difference in the electrical output is proportional to the amount of gas present. EC sensors are usually designed to minimize the effects of interfering contaminants, making the readings as specific as possible for the gas being measured.

Effects of interfering gases on electrochemical sensors

One of the chief limitations of electrochemical sensors is the effect of interfering gases – the ones that you are not trying to measure with the sensor – on the sensor readings. Substance-specific sensors are ideally supposed to respond only to the gases they are supposed to measure. The higher the specificity of the sensor, the less likely the sensor will be affected by other gases. The composition of the electrodes and type of electrolyte, as well as the use of selective filters for the removal of interfering gases are all ways to increase the specificity of the sensor. For instance, on the inside, a CO sensor is very similar to a sensor used to measure H2S. The trick is to keep the H2S from reaching the CO sensing electrode. Most substance-specific CO sensors include an internal activated carbon filter designed to remove the H2S and other acid gas interferents before they reach the sensing electrode. Thus, the reading of the sensor is not affected by the presence of H2S in the atmosphere being monitored. While inclusion of a filter is frequently able to increase specificity, removal of a filter may be used to broaden response to a wider variety of gases. For instance, carbon monoxide sensors that do not include a filter are sometimes marketed as “dual purpose” sensors for the simultaneous detection of both CO and H2S. This type of sensor responds to both CO and H2S, but cannot tell them apart. The sensor produces a single signal, which is up to the instrument user to interpret. Even though care has been taken to reduce cross-sensitivity in substance-specific designs, interferences still exist. In some cases, the interfering effect is positive and results in readings that are higher than actual. In other cases, the interference is negative and produces readings that are lower than actual. It’s important to understand clearly the effects of potential interferents on the output of the sensors installed. Users should consult the owner’s manual or contact the manufacturer of the instrument they will be using to verify the correct values to use when making decisions based on interfering contaminants.

The reason that CO sensors are potentially susceptible to hydrogen (H2) interference is the reaction that is used to detect gas. Hydrogen is actually part of the detection reaction. The relative response to hydrogen depends on the brand and model of sensor. Some commonly used CO sensors show a relative response to hydrogen as high as 60%. (Not GfG sensors, of course!) Sometimes the cross sensitivity is presented as an advantage to customers. In some cases the response is so high the manufacturer tells customers the sensors should not be used at all in the presence of hydrogen.

GfG instruments can be equipped with a number of different models and types of single and dual channel CO sensors. For applications where hydrogen may be present we use specially designed CO sensors with a very low relative response to H2. For refinery applications we would normally specify a two electrode “hydrogen nulled” 2CF sensor with a catalyst system designed to limit the response of the sensor to hydrogen. The 2CF sensor has a relative response to hydrogen of about 5%. You can’t completely eliminate the relative response, but you can certainly reduce it far below what you are currently seeing!

How electrochemical CO sensors detect gas

Gas that enters the sensor undergoes an electrochemical reaction that causes a change in the electrical output of the sensor. The difference in the electrical output is proportional to the amount of gas present. Gas enters the sensor through an external diffusion barrier that is porous to gas but nonporous to liquid. An internal organic vapor filter is usually included in the CO sensor to remove or at least reduce the interfering effects of solvents, alcohols and other unsaturated hydrocarbon vapors. (The 2CF sensor has a robust internal organic vapor filter which further improves its fitness for use in refinery applications.)

Carbon monoxide that enters the sensor is oxidized at the surface of the sensing electrode, causing the potential of the sensing electrode to rise relative to that of the counter electrode. Current collecting filaments connect the electrodes with the external pins of the sensor. The instrument supplies power to the sensor, and interprets the output of the sensor by readings obtained through the external pins. In two-electrode sensor designs, the potential of the sensing electrode is compared directly to that of the counter electrode. In three electrode designs, what actually is measured is the difference between the sensing electrode and reference electrode.

The electrodes include materials necessary to catalyze the detection reaction. Without the presence of the catalyst the reaction would not occur. Changing the type or relative proportions of the metals in the catalyst system can profoundly affect both the speed of response as well as the relative response of the sensor to interfering contaminants.

The CO detection reaction is a two-step process. The electrolyte in which the reaction occurs is a weak solution of sulfuric acid. In the first step the carbon monoxide is oxidized at the sensing electrode to produce CO2. The reaction generates two electrons of electricity (2e-) for each molecule of CO detected, (which is how the concentration of CO is measured). The reaction uses one molecule of water from the electrolyte, and produces two hydrogen protons (2H+) for each molecule of CO reacted.

CO + H2O → CO2 + 2H+ + 2e-

The second step occurs at the counter electrode. In the second step the hydrogen protons (2H+) produced in the first step react with oxygen in the electrolyte to produce water. The oxygen in the electrolyte comes from the air in which the sensor is located. Electricity from the instrument’s power supply is used to provide the necessary electrons (2e-). Since the molecules of water consumed in the first step are regenerated in the second step, the reaction is said to be “balanced” by this second step.

½ O2 + 2H+ + 2e- → H2O

When the two “half-cell” steps are added together, the overall (or net) detection reaction becomes:

CO + ½O2 → CO2

The working efficiency of the sensing electrode is very high. This means the sensor is usually easily able to oxidize the incoming CO as fast as it reaches the sensing electrode. If the concentration of incoming gas exceeds the ability of the sensing electrode to oxidize the gas, the sensor becomes saturated, in which case the output reaches a maximum value and can’t rise any higher. However, as soon as the concentration of gas in the atmosphere drops below this critical concentration, the sensor rapidly recovers with no damage done to the sensor.

Water from the electrolyte is used, but is regenerated during the course of the reaction. The CO2 produced in the reaction accumulates in the acid electrolyte as carbonic acid. The only materials consumed during the detection reaction are the molecules of carbon monoxide, power from the battery of the instrument and oxygen. As long as the sensor is located in an atmosphere containing even small amounts of oxygen, the sensor is able to replenish itself directly from the atmosphere. This is the reason that non-consuming electrochemical sensors have such long life spans. The lifespan of the sensor is not affected by exposure to the contaminant that it measures. No part of the sensor is consumed during the detection reaction. You can expose the sensor to CO calibration gas every single day without shortening or affecting the lifespan of the sensor.

Effect of hydrogen on CO sensors

The first step in the CO detection reaction generates hydrogen protons. The second step uses hydrogen protons to regenerate the water used in the first step of the reaction. The materials in the sensor are designed to facilitate this process. Even though the sensor is designed for the detection of CO, hydrogen gas can react (and be detected) to at least some extent at the sensing electrode. The reaction for H2 at the sensing electrode is:

H2 → 2H+ + 2e-

The second step at the counter electrode is the same reaction as the second step in the CO detection reaction:

½ O2 + 2H+ + 2e- → H2O

The relative response of the CO sensor to hydrogen is determined by how well the catalyst and electrode system facilitates the hydrogen detection reaction compared to the CO detection reaction. It is possible to reduce the relative response of the sensing electrode to hydrogen by using a slightly different catalyst, and/or using a two electrode rather than three electrode sensor design. (It is very important, of course, to choose a catalyst system that is still a good choice for CO detection!)

The “hydrogen nulled” GfG 2CF sensor is a two electrode design, in a standard size 0.4 inch radius (4 Series) formatted housing. Three electrode single channel CO sensors, as well as four electrode “dual channel” CO / H2S sensors generally have a higher relative response to hydrogen. However, while the 2CF sensor has a much lower relative response to H2, it takes a little longer to clear back down to zero once the exposure is completed.

Most instrument users find that the performance of the 2CF hydrogen nulled sensor is an outstanding improvement over the performance of the standard CO sensors they have used in the past.

The following chart compares the relative response of the hydrogen nulled 2CF sensor and our standard 4CM CO sensor when the sensors are exposed to 1000 ppm hydrogen. The relative response of the 2CF sensor is less than 5%!

The second chart compares the performance of the 2CF hydrogen nulled sensor, and a standard 4 electrode dual channel CO / H2S “COSH” sensor exposed to 1000 ppm hydrogen calibration gas.

The third chart compares the performance of the GfG 2CF hydrogen nulled sensor, and a standard GfG 4CM CO sensor exposed to 200 ppm CO calibration gas. While the hydrogen nulled sensor takes a little more time to recover after exposure, the time to alarm is virtually the same for both sensors.

What other gases and vapors can have an interfering effect on CO sensor readings?

Although the internal organic vapor filter included in most CO sensors reduces the effects of interfering contaminants, once the filter is saturated, breakthrough can occur. Once breakthrough occurs the CO sensing electrode responds to a wide variety of interfering contaminants including alcohols, (such as methanol and isopropyl alcohol), solvent vapors, (such as toluene and MEK), combustible liquids, (such as kerosene and jet fuel), and unsaturated hydrocarbon gases (such as ethylene, propylene and isobutylene). Many CO sensors also show a significant response to nitric oxide, (NO). CO sensors also respond strongly to acetylene gas. Since the internal filter does not absorb acetylene, the sensor shows a very strong and immediate response to this gas. GfG Technical Note 2019, “G460 and Micro IV electrochemical (EC) toxic sensor relative response matrix” includes a chart listing some of the most common documented interferences for the electrochemical sensors used in GfG products. The note is posted on our www.goodforgas.com website at the following link: TN2019

Unfortunately, although CO sensors recover rapidly from exposure to hydrogen, it can take hours or even days for the sensor to recover completely after exposure to solvent vapor or acetylene. Although CO sensors are not usually harmed by exposure to low concentrations, exposure to very high concentrations of solvent or alcohol vapor can permanently damage the sensor. Never use alcohol or solvents to clean the instrument housing or sensor compartment area! Sample draw tubing or filters that are contaminated by solvent or exposure to heavy fuels (such as diesel vapor) should be discarded and replaced.

The following chart shows the response of an LEL sensor and a 2CF “hydrogen nulled” CO sensor to 10% LEL methanol (= 6,000 ppm). Note that this is a very high concentration of methanol! In the following chart you can see when the methanol starts to break through the internal filter in the CO sensor. The CO reading does not begin to rise until almost 4.5 minutes after the methanol exposure is ended. It took several hours for the CO sensor to finish stabilizing back on zero after the exposure to methanol. It should be noted that the final stable reading of 13 ppm for the 2CF CO sensor is extremely low compared to the response of most other CO sensors.

The following chart shows the response of the CO channel of a dual channel “COSH” type sensor for the measurement of CO and H2S to 5% volume (= 50,000 ppm) ethanol. Note that this is a very high concentration of ethanol! The red colored line on the chart shows the response of the LEL sensor. Because the COSH type sensor has a less robust internal filter, the relative response to the alcohol is quite a bit higher (450 ppm) than the response of the 2CF CO sensor in the previous example. It also takes a lot longer for the CO channel of the COSH sensor to recover after the exposure ends.

GfG supplies a substantial number of instruments equipped with the hydrogen nulled 2CF CO sensors for use in oil refinery applications. We receive very positive reviews for instruments equipped with this sensor. Please let us know if you would like to evaluate instruments that have been equipped with this type of sensor.

Thanks again for the question!