How should humid tank air be sampled for H₂S and ammonia?
Take the water out at one controlled point, keep it from condensing anywhere else in the line, and prove with calibration gas that the target gas still reaches the analyzer. Air drawn from a covered wastewater tank is humid. If its water condenses in the sample line, it dissolves part of the H₂S and much more of the ammonia, and the analyzer reads low without showing a fault.
- Remove the water in a cooler with an automatic drain, or keep the line heated and dry the sample in a condenser. See how the water is removed.
- Make sure every filter in the path passes the target gas: one moisture filter took a 1,000 ppm ammonia reading to zero. See how to check the path.
- Apply calibration gas at the probe inlet, not only at the analyzer, to measure what the line removes.
- Allow for the time the gas needs to travel the line, on top of the sensor's own response time. See line length and response time.
Why do H₂S and ammonia readings drop in a wet sample line?
Because the gas dissolves in the water that condenses in the line. In extractive sampling, a pump draws air from the tank through a sample line to an analyzer outside the tank. The instrument stays out of the wet air space above the liquid (the headspace) and can be serviced without opening the tank, as in both Eranntex projects described below. [S4] [S5] The reading then depends on what happens to the gas inside the line.
Humid air and liquid water are different problems. Relative humidity (%RH) says how close air is to holding all the water vapor it can at its temperature. The GTYQ-MIC-600 fixed detector is specified for 10–95 %RH, so humid air alone is within its range. [S9] Liquid water forms when air cools below its dew point, the temperature at which that air would reach 100 %RH. Warm, humid tank air, like the air above the anaerobic tank in Example 1, can cool below its dew point in a line that runs through colder surroundings, and part of its vapor then turns into liquid water (condensate) on the tube walls. [S4]
Both H₂S and ammonia dissolve in that water, ammonia far more readily. At 25 °C, a liter of water in equilibrium with the sample air holds about 1,500 times as much ammonia as a liter of that air, but only about 2.5 times as much H₂S. [S1] One milliliter of condensate can therefore hold the ammonia from about 1.5 liters of sample. Gas that dissolves in the line never reaches the sensor, so the instrument reads low. A low reading is the dangerous error, because the system still looks as if it is working.
Some detectors are instead built to sit in a duct or a flow cell in condensing air. Which areas of a plant need personal, fixed or extracted monitoring is covered on the water and wastewater industry page.
How is the water removed without losing the gas?
At one controlled point, chosen from the sample, rather than wherever the line happens to be cold. Water has to come out before the sample reaches the detector: a sample near saturation is above the MIC-600's 95 %RH limit, and liquid carried with the sample can damage the analyzer. [S9] Ahead of that point, the sample is kept above its dew point. The US EPA reference method for instrumental gas analyzers requires this of every component except the sample conditioning, and asks for minimal contact between condensate and the sample gas. [S12]
- Cooling stage with automatic drain. For the warm, humid air above the anaerobic tank in Example 1, the TH-2000-C conditioning system cooled the sample in a thermoelectric (solid-state) cooler inside the cabinet so the water condensed there, drained it automatically, and passed the dried sample to a GTYQ-MIC-600 in the same cabinet. [S4] The TH-2000-C is documented for high-humidity, low-dust samples in this way. [S8]
- Heated line and two-stage condenser. Where low concentrations of water-soluble gases have to be measured accurately, the sample line is heat-traced, that is, wrapped with an electric heating cable, so water cannot condense in it. [S12] The sample is then dried in a two-stage condenser that holds its dew point at 4–5 °C. [S2] On the DOAS-3000 analyzer, this arrangement exists to limit losses of water-soluble components. [S3] Where an outdoor line can freeze, it is heat-traced for that reason as well. [S8]
- Simple pretreatment. In Eranntex's documented pretreatment options, simple uncooled pretreatment is matched to ambient-temperature, low-humidity, dust-free samples; humid samples are dried in a cooling stage. [S8] [S9]
Whichever stage is used, its filters have to pass the target gas. The whole-path gas check shows whether they do.
How do you know the sample path is not absorbing the gas?
Apply calibration gas, a cylinder gas of known concentration, at the probe inlet and compare the reading with the same gas applied directly at the analyzer. The difference is what the line, filters and conditioning remove. US EPA Method 7E, a US stack-testing method, calls this difference system bias; for the system check, the gas goes in at the probe, upstream of the filter and all sample conditioning. [S12]
Eranntex ran this check before shipment on an ammonia treatment project: the analyzer read 1,000 ppm ammonia calibration gas correctly on its own but read zero with the sample conditioning fitted. The moisture filter was taking up the ammonia. It was removed, and the calibration gas then reached the analyzer. [S6] Filter material matters: in a laboratory study, paper and fuel filters took up ammonia, most strongly when first exposed, while PTFE (Teflon) filters took up almost none. [S13] For any path that must deliver ammonia, test the assembled path with ammonia before relying on it.
After installation, repeat the check at the probe inlet. A bump test, a short exposure to test gas, shows that gas gets through and the alarm works, but not how much is lost, and a calibration done only at the analyzer port cannot show that loss either. OSHA's guidance on direct-reading monitors, a US advisory bulletin, states that a bump test does not measure accuracy and counts sample tubing among the links in the calibration chain. [S11] The differences between a bump test, a calibration check and a full calibration are explained in the personal gas monitoring guide. The water and wastewater page describes an HSE investigation in which a complete sampling arrangement responded far more slowly to a flammable vapor than the detector alone.
Example 1: alarms for H₂S, methane and three other gases above an anaerobic tank
Above a sealed anaerobic wastewater tank, the monitoring task was alarms for five gases. Anaerobic treatment warmed the water, so the air above it was hot and heavy with vapor. The operator wanted an alarm whenever H₂S, methane (CH₄), carbon monoxide (CO), carbon dioxide (CO₂) or oxygen (O₂) moved outside its limit. An alarm would send staff to check whether the ventilation had failed. [S4]
Eranntex supplied a TH-2000-C sample-conditioning system with a GTYQ-MIC-600 five-gas detector, in an outdoor rainproof cabinet. The specified ranges were:
| Gas | Range |
|---|---|
| H₂S | 0–100 ppm |
| CH₄ | 0–5 %Vol |
| CO | 0–1,000 ppm |
| CO₂ | 0–5,000 ppm |
| O₂ | 0–30 %Vol |
A brushless DC pump drew the tank air over a 20 m sampling distance, and a 10 m sample line was supplied with the system. Inside the cabinet, a thermoelectric cooler condensed the water out, a filter removed remaining moisture and dust, and a peristaltic pump (a pump that squeezes a tube to move liquid) and a solenoid valve drained the condensate automatically. The 7-inch touchscreen showed all five readings. [S4]
The site's requirements were not fully stated at the first enquiry. When the specification was reviewed, the site confirmed a hazardous-area requirement at the mounting location, and the cabinet specification was updated before shipment. The alarm points were set at the factory, an Eranntex engineer guided the installation and wiring by video, and the system went into service. [S4]
These ranges serve alarms above a process tank. They say nothing about exposure limits for people working at the tank, who still need personal monitors.
Example 2: ammonia and H₂S at 0.01 ppm to measure treatment performance
At a covered sewage storage tank, the measurement was for treatment performance, not alarms. The customer had built equipment to remove ammonia and H₂S from the air above the tank and needed data showing how well it worked, with readings resolved to 0.01 ppm. The air was at ambient temperature and atmospheric pressure, humid and free of dust. [S5]
The configuration Eranntex specified drew air from the tank's air duct through a 10 m hose to a pumped online monitor in an outdoor 304 stainless-steel cabinet rated IP67. Both gases were measured with electrochemical sensors on 0–25 ppm ranges at 0.01 ppm resolution, the ammonia sensor was chosen to resist interference from H₂S, and temperature was logged alongside. A pretreatment stage removed water vapor ahead of the sensors. [S5]
An alarm on a 0–100 ppm H₂S range needs reliable readings well above zero. A treatment-performance measurement at 0.01 ppm needs the path to pass very low concentrations without loss.
Measuring both sides of a treatment unit means two measuring points with different requirements. On an industrial ammonia treatment project, the inlet sample was hot (75 °C), under negative pressure (3.5 kPa) and specified on a 0–15,000 mg/Nm³ range (milligrams per normal cubic meter), so it was specified with cooling and a tunable diode laser absorption spectroscopy (TDLAS) analyzer, an optical method chosen here for the high concentration. The outlet sample was at ambient temperature on a 0–20 mg/Nm³ range, where an electrochemical sensor was specified. [S6]
Which conditions at the sampling point change the system?
High temperature, a near-saturated or wet sample, dust, pressure or suction, and interfering gases each add a component to the sample path. A standard detector configuration is meant for open air with a simple gas background, so when gas is drawn from a wastewater pipe or a closed tank, the Eranntex Technical Team first reviews these conditions. [S7]
| Condition at the sampling point | What it adds, and documented Eranntex options |
|---|---|
| High sample temperature | Adds: a cooling sampling handle or cooling stage. [S7] Documented: TH-2000-C sample temperature −40 to 600 °C standard, up to 1,300 °C optional; its −40 to 70 °C figure is the cabinet's operating range, not the sample's. [S8] |
| Sample near saturation or carrying liquid | Adds: moisture removal ahead of the detector. [S7] Documented: a cooling stage with automatic drain (TH-2000-C), or a heat-traced line with a two-stage condenser at a 4–5 °C dew point (DOAS-3000); the MIC-600 is specified for 10–95 %RH. [S2] [S8] [S9] |
| Dust | Adds: dust filtration so the path does not clog. [S7] Documented: TH-2000-C adds a gas-cleaning and backflush unit for dusty samples; MIC-600 pretreatment options come in three grades: ambient, low-humidity and dust-free; hot and humid with little dust; hot, humid and dusty. [S8] [S9] |
| Positive pressure or suction | Adds: a pressure-reducing valve and flowmeter for positive pressure, or a higher-capacity sample pump for suction. [S7] Documented: TH-2000-C is documented for sampling points under positive or negative pressure. [S8] |
| Interfering, corrosive or acid gases in the background | Adds: a sensing principle chosen for the gas mixture, such as high-accuracy infrared or a photoionization detector (PID). [S7] Documented: Example 2 used an ammonia sensor chosen to resist H₂S; the high-concentration ammonia inlet was specified with TDLAS. [S5] [S6] |
How long does the gas take to reach the analyzer?
Roughly the line's internal volume divided by the sample flow, plus the sensor's own response time. For example, a 4 mm bore line holds about 0.25 L per 20 m, so at 2 L/min the gas takes about 8 seconds to cover 20 m and about 26 seconds to cover 70 m. The sensor's response comes on top: Eranntex lists a T90 (the time to reach 90% of the final reading) of up to 30 s for its H₂S and ammonia sensors and up to 10 s for methane. [S10]
Conditioning stages add volume, and adsorption or condensate can slow the response much further, so the response that matters is the one measured with gas applied at the sample inlet after installation. US EPA Method 7E defines response time the same way, from a change in concentration at the sampling point. [S12] The TH-2000-C samples over 30–40 m as standard; an optional jet or vacuum pump extends this beyond 70 m. Sample flow is 2 L/min as standard, with 4 L/min optional. [S8]
Put the cabinet as close to the tank as the site allows. The measured sample also has to go somewhere afterwards; with H₂S in it, the outlet needs a safe discharge point.
What should you send with an enquiry?
The gases, the conditions at the sampling point and the route of the line; a model does not need to be chosen first.
- The gases and ranges, and whether the measurement is for an alarm or for process or treatment-performance data.
- The tank or duct, and whether the sampling point is under suction or pressure.
- The sample temperature, its moisture, and the coldest place the line will pass.
- Dust, grease, foam or liquid carry-over at the sampling point, and any other gases present.
- The line length and route, and where the measured sample can be discharged.
- Indoor or outdoor installation, available power, and the signal your control system expects. TH-2000-C provides 4–20 mA, RS485 and passive contacts. [S8]
- The country, and any hazardous-area classification at the cabinet location.
Certificate coverage for a hazardous-area installation depends on the model and configuration. Our sales engineers confirm it for the configuration you specify. The quotation checklist lists general project details.
Frequently asked questions
References
Bracketed numbers in the text link to these sources. Project configurations and the factory test record on this page come from Eranntex Technical Team project records. Product data comes from Eranntex product documentation for the TH-2000-C, GTYQ-MIC-600 and DOAS-3000.
| Ref. | Source | What it supports |
|---|---|---|
| S1 | R. Sander, Compilation of Henry's law constants (version 5.0.0) for water as solvent, Atmospheric Chemistry and Physics 23, 10901–12440 (2023); entries for ammonia and hydrogen sulfide, retrieved 3 November 2026 | Henry's law solubility at 25 °C: ammonia 0.59 and H₂S 1.0 × 10⁻³ mol/(m³·Pa). Multiplied by RT, these give the water-to-air ratios in the text, about 1,460 and 2.5. |
| S2 | Eranntex DOAS-3000 product documentation: sample conditioning description | Heat-traced sample line and two-stage condenser at a 4–5 °C dew point. |
| S3 | Eranntex DOAS-3000 product page | Heated transport and two-stage condensation to limit losses of water-soluble components. |
| S4 | Eranntex project record: five-gas monitoring system above an anaerobic wastewater tank | Example 1 conditions, configuration, ranges and commissioning; extractive layout. |
| S5 | Eranntex configuration record: ammonia and H₂S monitoring above a covered sewage storage tank | Example 2 conditions and configuration; extractive layout. |
| S6 | Eranntex project record: ammonia monitoring at the inlet and outlet of a treatment unit, including the pre-shipment gas test | Inlet and outlet specification; filter adsorption found in the pre-shipment test. |
| S7 | Eranntex Technical Team note on sampling from pipes and closed spaces | Conditions that add components to the sample path. |
| S8 | Eranntex product specification table, online monitoring systems; TH-2000-C product page; Oil & Gas industry page | TH-2000-C sample temperature, distance, flow, pressure, outputs, application range and dust option; heat tracing of lines that can freeze. |
| S9 | Eranntex product specification table, fixed detectors; GTYQ-MIC-600 product page | MIC-600 operating humidity, sampling options and pretreatment grades. |
| S10 | Eranntex product specification table, sensor options; GTYQ-MIC-600 product page | Sensor T90 for H₂S, ammonia and methane. |
| S11 | OSHA Safety and Health Information Bulletin: Calibrating and Testing Direct-Reading Monitors, SHIB 11-26-2024 (US advisory guidance) | A bump test does not measure accuracy; sample tubing is a link in the calibration chain. |
| S12 | US EPA Method 7E, 40 CFR Part 60, Appendix A-4, sections 3 and 6.1 (US reference method for stack testing with instrumental analyzers) | System calibration at the probe, system bias and response time definitions; components other than sample conditioning kept above the moisture dew point; minimal contact between condensate and sample gas. |
| S13 | Z. Zhu, H. Xin, H. Li, R. T. Burns, H. Dong, Assessment of in-line dust filter type and condition on ammonia adsorption, Transactions of the ASABE 50(5), 1823–1830 (2007) | Ammonia taken up by paper and fuel filters, highest at first exposure; negligible uptake by PTFE filters. |




