This guide summarises information published by INRS (France), OSHA (United States) and HSE (United Kingdom) on atmospheric testing in confined spaces and the use of portable gas detectors. It describes published regulations and practices. It does not determine whether a particular rule applies to your site, activity or equipment: this must be checked against the official source cited and, where necessary, with the competent authority. This text was prepared with the assistance of AI using the sources listed at the end of the page, consulted on 2 November 2026.
1What This Guide Covers and What It Does Not Cover
This guide covers atmospheric testing before and during an operation, the order of measurements, the response to an alarm and key considerations concerning gas detection instruments. It does not cover the complete organisation of a confined-space operation (permits, rescue arrangements, respiratory protection), or the legislation of other countries (Belgium, Switzerland, Canada, etc.), which is not described here.
A Review of Published Documents, Not a Certificate of Compliance
Each numerical value is attributed to its source document. These documents are not all of the same nature (recommendations from a prevention organisation, regulations, legal definitions): they cannot be compared directly (see the comparison of the three regulatory frameworks below), and no value is presented here as a single universal reference value.
Complementing the OSHA 29 CFR 1910.146 Reference Page
The content of the US regulation (definitions, testing sequence, equipment) is already described on the OSHA 29 CFR 1910.146 reference page. This guide does not reproduce that content: it follows its practical approach.
2Why the Atmosphere in a Confined Space Must Be Tested
Categories of Hazards
INRS (see also ED 6184, p. 4) describes a confined space as a fully or partially enclosed volume that was generally not designed for permanent occupancy and whose limited or insufficient openings restrict air exchange with the outside environment.
The atmosphere may become hazardous because of inadequate natural ventilation, materials or substances present, the design or location of the space, equipment used or the nature of the work being performed (ED 6184, p. 4).
The hazards identified include oxygen deficiency, toxic substances, explosion and fire, difficulties in carrying out rescue operations, as well as falls, drowning and engulfment.
For atmospheric hazards, NIOSH calls for monitoring three parameters: oxygen concentration, flammable gases and vapours, and potential toxic airborne contaminants. This NIOSH page was written for post-hurricane operations and was last revised on 10 March 2023.
INRS ED 6184 uses 19% oxygen as a threshold (p. 6) and indicates that below 6%, death can occur within minutes. It also specifies that an oxygen measurement alone does not establish whether the air is safe.
Why Human Senses Are Not Enough
According to NIOSH, workers cannot rely on their senses to assess the atmosphere; a calibrated direct-reading instrument is used before entry and for subsequent monitoring.
INRS identifies inadequate preparation and risk assessment, including insufficient atmospheric testing, among the main causes of accidents. It also specifies that detectors, even when equipped with alarms and properly checked, are not sufficient on their own.
3Before Entry: Assessment and Preparation
First Question: Is Entry Necessary?
Before organising entry, it is necessary to determine whether the work can be carried out from outside.
HSE places avoiding entry as the first duty, ahead of establishing a safe system of work and emergency arrangements. L101 (Regulation 4(1), paragraphs 70 to 72) gives examples such as sampling or cleaning from outside and remote visual inspection.
INRS ED 6184 requires confirmation that the operation is essential (p. 9) and provides for certain operations to be performed from outside the structure (p. 13).
OSHA 1910.146(c)(3) requires an employer who decides that employees will not enter a permit space to take effective measures to prevent them from entering.
Preliminary Risk Assessment
INRS describes an approach that includes risk assessment (DUERP, the French occupational risk assessment document), a prevention plan when an external contractor is involved, an operating procedure, a person in charge, an attendant, an entry permit, medical fitness and an emergency plan (Prévenir les risques).
Its Regulations page refers to Articles R. 4222-23 and R. 4222-24 of the French Labour Code.
Article R. 4222-23 covers wells, tanks, reservoirs, pits, galleries and locations where permanent compliance with the chapter's ventilation and air-quality requirements cannot be ensured. Work may only begin after checking that there is no risk to workers' health and safety and, where necessary, after cleaning or making the atmosphere safe and emptying the contents.
Article R. 4222-24 is separate: during work, ventilation must maintain a healthy atmosphere and provide continuous air circulation.
Neither article specifies a measurement method or instrument.
Texts consulted on code.travail.gouv.fr (displayed update date: 1 May 2008); the version in force on Légifrance at the time of publication remains to be confirmed.
Ventilation Before Testing: What the Sources Say
| Source | What It States |
|---|---|
| INRS, Prévenir les risques | Mechanical ventilation for at least 20 minutes before atmospheric testing |
| INRS, Ce qu'il faut retenir | Airflow velocity of at least 0.30 m/s |
| INRS ED 6184, p. 13 | At least 20 minutes and at least 0.3 m/s |
| CNAMTS/INRS R 447, §6.3 | At least 10 air changes per hour, followed by a waiting period of at least 20 minutes; target oxygen concentration between 19% and 21% (recommendation limited to certain sectors) |
| OSHA 1910.146(d)(3) | Isolation and ventilation of the space as necessary, without a specified minimum duration |
Two points should be noted regarding the sequence.
For construction, OSHA 1926.1204(e)(1) requires atmospheric conditions to be tested before any alteration of the space's natural ventilation and again before entry is authorised, with an exception for certain spaces that cannot be isolated. The INRS sequence of “ventilate, then test” therefore cannot be applied unchanged to operations governed by this US regulation.
Furthermore, R 447 specifies 20 air changes per hour during operations likely to release gases or vapours. A different value may be established in the prevention plan, provided written justification is given.
Where to Test and How Long to Sample
INRS describes measurements taken from outside at a minimum of three points, with at least one minute of sampling at each point, progressing gradually through the space.
R 447 requires testing at each access point, at three levels (top, middle and bottom), for at least one minute at each level.
HSE (L101) requires competent testers who can interpret the results and take appropriate action, as well as the retention of records (paragraph 97). The risk assessment determines whether testing is necessary before each subsequent entry (paragraph 90). When sampling from outside using a probe, sufficient time must be allowed for the sample to replace the air in the probe, and enough locations must be sampled to identify pockets of poor-quality air (paragraph 98).
One minute per point does not guarantee a valid measurement.
INRS ED 6184 (p. 13) requires consideration of the sample transit time through the sampling line, particularly when using a probe, and the detector's response time, with a sampling duration of at least one minute. Measurements must cover the entire volume in which workers will operate.
Testing may therefore take longer and require more sampling points: “three points for one minute” does not, on its own, establish that the atmosphere is safe.
Appendix B to OSHA 1910.146 requires each parameter to be measured for at least the minimum response time specified by the instrument manufacturer.
To measure without exposing the operator, INRS ED 6088 (p. 8) recommends a probe extending the sampling distance by several metres, while accounting for the resulting increase in response time.
4The Order of Atmospheric Testing
Oxygen First, Then Flammable Gases, Then Toxic Contaminants
For general industry, OSHA specifies testing in this order (29 CFR 1910.146(d)(5)(iii); details are provided on the reference page).
The construction regulation, 29 CFR 1926.1204(e)(3), follows the same sequence.
HSE (L101) and NIOSH also describe the sequence as oxygen, then flammable gases, then toxic contaminants.
The INRS pages reviewed for this guide do not specify this order.
Appendix B to OSHA 1910.146 explains the reasoning: most combustible-gas meters depend on oxygen and do not provide reliable readings in oxygen-deficient atmospheres, while fire or explosion hazards are generally more immediate than exposure to toxic substances.
It also states that testing equipment must have sufficient sensitivity and specificity for the contaminants being investigated.
For catalytic sensors, INRS ED 6088 (p. 13) specifies that at least 10% to 12% oxygen is required. The OSHA Technical Manual (2014) states that lower explosive limit (LEL) measurements are unreliable in oxygen-deficient or oxygen-enriched atmospheres.
Known Limitations of Each Measurement
According to the OSHA Technical Manual (2014), an LEL reading may appear to indicate a non-combustible atmosphere when the concentration is above the upper explosive limit.
An LEL measurement is also unsuitable for assessing a toxic substance against its occupational exposure limit.
INRS ED 6184 reiterates that measuring oxygen alone does not establish whether the air is safe.
5During the Operation: Atmospheric Monitoring
A Personal Detector Positioned Close to the Breathing Zone
INRS describes a personal portable detector that remains operational throughout the intervention and is positioned as close as possible to the breathing zone, allowing continuous atmospheric monitoring.
INRS ED 6184 (pp. 12 to 14) provides for a portable multi-gas detector for each worker.
R 447, within the sectors it covers, requires a personal detector to be worn continuously.
External Attendant and Evacuation Equipment
INRS describes an attendant stationed outside the confined space and an escape mask. It states that cartridge-type air-purifying respirators are prohibited where oxygen is deficient.
The organisation of rescue operations is not addressed in this guide.
Continuous or Periodic Monitoring
| Source | What It States |
|---|---|
| OSHA 1926.1204, construction, (e)(1)(ii) and (e)(2) | Continuous monitoring by default; periodic monitoring only when justified. Paragraph (c)(5) requires monitoring capable of detecting deteriorating conditions in time if ventilation fails |
| OSHA 1910.146, general industry, (d)(5) | Testing before entry, followed by testing or monitoring as necessary; continuous monitoring in the case of spaces that cannot be isolated (see reference page) |
| HSE L101, paragraphs 90 and 91 | Repeated testing according to the risk assessment; continuous monitoring when forced ventilation is used or work may alter the atmosphere |
| INRS | A personal detector operating throughout the intervention (continuous atmospheric monitoring); preliminary testing before entry |
OSHA also provides for entrants or their representatives to observe testing (1910.146(d)(5)(iv) and 1926.1204(e)(4)).
Test results must be communicated to them immediately under 1926.1204(e)(6) and recorded on the permit under 1910.146(f)(10).
HSE requires test results to be recorded.
6What to Do When a Gas Detector Alarm Sounds
Situation 1: An Alarm During Pre-Entry Testing
During atmospheric testing before entry, INRS specifies an additional 20 minutes of ventilation if an alarm sounds, followed by another test.
If the alarm sounds again, access must be secured against entry and management informed (Prévenir les risques).
INRS ED 6088 (pp. 7 and 10) requires each alarm to be associated with a simple, clear action known to all operators.
Situation 2: An Alarm While People Are Inside the Space
The additional 20-minute ventilation procedure applies only to the pre-entry testing phase. It must never be interpreted as permission to wait inside the confined space.
When people are already inside, the cited sources require immediate evacuation.
NIOSH instructs workers to leave immediately if a hazardous condition develops and to monitor the atmosphere continuously throughout entry.
INRS ED 6184 (p. 14) states that if the detector signals danger, workers must immediately put on the designated escape equipment and evacuate the hazardous area. The attendant orders evacuation if ventilation fails.
A self-rescue breathing device is intended only for evacuation, never for continuing work.
R 447, in the sectors concerned, requires workers to put on their masks, evacuate the space, close the access points and alert the employer.
The INRS Prévenir les risques page does not explicitly describe this response. It provides for an external attendant who orders evacuation when informed of a problem or when a ventilation failure is detected.
Before any subsequent entry, OSHA's alternative procedures under 1910.146(c)(5)(ii)(G) require workers to leave the space immediately, an assessment of how the hazardous atmosphere developed, and the implementation of protective measures.
Under the permit-required programme, the permit must be cancelled when an unauthorised condition arises (1910.146(e)(5)(ii)), and identified deficiencies must be corrected before another entry (1910.146(d)(13)).
Site-specific instructions, including operating procedures and emergency plans, must incorporate the requirements of the applicable regulatory framework.
The French sources describe securing access and informing management if another alarm occurs during pre-entry testing. They do not describe the conditions for resuming work after an evacuation.
Criteria and Recommended Alarm Settings Cited in the Sources
The values below have different purposes. Each is therefore presented separately in its own context.
Regulatory criteria (definitions)
| Source | Criterion |
|---|---|
| OSHA 1910.146(b) | Hazardous atmosphere: oxygen below 19.5% or above 23.5%; flammable gas above 10% of the lower flammable limit (LFL). These are criteria used in the US regulatory definition, not alarm settings |
Published recommendations for alarm settings
| Source | Recommendation |
|---|---|
| INRS ED 6088, p. 10 | Explosion: 10% of the LEL when people are working in the area. Oxygen: recommended minimum concentration of 19%, with the first alarm set at that value. Toxic substances: first alarm at the eight-hour occupational exposure limit (VME), second alarm at the 15-minute short-term exposure limit (VLCT) |
Alarm settings selected for a particular operation: These depend on the risk assessment and the manufacturer's instructions.
“First alarm” and “second alarm” do not, by themselves, describe a complete configuration. The manufacturer's instructions specify whether an alarm is based on an instantaneous concentration or a calculated time-weighted average, and over what period.
For toxic substances, VME (eight-hour occupational exposure limit) and VLCT (15-minute short-term exposure limit) are values associated with specified reference periods, according to the INRS page on chemical-agent regulations. An instantaneous concentration displayed by a detector is not equivalent to these values.
What an Alarm, or the Absence of an Alarm, Does Not Tell You
INRS emphasises that detectors are not sufficient on their own.
INRS ED 6088 (p. 10) notes that alarm settings must account for the instrument's accuracy, repeatability and reproducibility. The interferences described below can also distort readings.
7Understanding the Instrument
Selecting the Gases to Detect: Risk Assessment Comes First
The risk assessment determines which substances must be detected, followed by the required sensors, measurement ranges and performance characteristics.
INRS ED 6184 (pp. 11 and 13) requires a portable detector suited to the identified risks and equipped with detection cells corresponding to the gases likely to be encountered.
INRS ED 6088 (p. 6) requires the equipment specification to identify the substances to be detected and other substances potentially present. It recommends choosing the technology first, then the brand and model.
An instrument described as “multi-gas” therefore covers only the gases for which it has suitable sensors. It does not automatically cover every atmospheric hazard associated with an operation.
Sensor Technologies
INRS ED 6088 (second edition, April 2022) describes several technologies. The following table summarises some of them as general information from the publication.
The service life, response times and minimum oxygen concentrations relevant to selecting and operating a detector must be taken from the documentation for the specific instrument and sensor concerned.
| Sensor | Information From INRS ED 6088 |
|---|---|
| Catalytic (combustible-gas detector) | Measurement from 0 to 100% LEL; t90 (at 50% LEL) of 20 to 30 seconds; service life of 2 to 5 years; requires at least 10% to 12% oxygen (p. 13) |
| Infrared | t90 (at 50% LEL) of 10 to 20 seconds; does not detect hydrogen (pp. 14–15) |
| Electrochemical (toxic gases) | A sensor is designed for a particular gas but can generally respond to other gases; a minimum oxygen concentration is required (p. 15) |
| Photoionisation (PID) | Fast response, non-selective, measurement in ppm (pp. 15–16) |
| Oxygen | Liquid electrolyte; solid electrolytes are used more commonly in analysers and less frequently in gas detectors (p. 17) |
| Colorimetric tubes | Accuracy of 30% at best; should not be used for atmospheric checks (wording from ED 6088) (p. 17) |
The OSHA Technical Manual (updated February 2014) describes the same main technology families: electrochemical, catalytic, PID and infrared.
Interferences and Cross-Sensitivities
According to INRS ED 6088:
- Electrochemical CO sensors are generally sensitive to other gases, particularly H₂S and H₂ (note 1, p. 6).
- Temperature, humidity, oxygen concentration and electromagnetic fields can distort measurements (note 3, p. 6).
- Catalytic sensors can become unsuitable when exposed to high concentrations of organochlorine compounds, silicones, lead, H₂S or mercaptans (p. 13).
- PID sensors underestimate concentrations when humidity is high relative to the humidity conditions used during calibration (p. 15).
The OSHA Technical Manual (2014) likewise states that silicones and sulphur can desensitise catalytic sensors.
Response Testing, Calibration and Adjustment: What the Sources Say
INRS ED 6088 (p. 9) uses three terms that must be distinguished:
Response test: Exposure to a reference gas to observe the instrument's response and response time. INRS presents this as the only check that makes it possible to assess detection capability and recommends performing it very frequently, if possible before every use.
Calibrage (adjustment): Adjustment of zero and sensitivity using a reference gas, following an unsuccessful response test, maintenance or malfunction.
Étalonnage (metrological calibration): Establishment of a response curve without modifying the instrument.
Other sources provide additional guidance.
INRS ED 6184 (p. 11) requires instruments and the date of their last inspection to be checked before an operation.
HSE L101 (paragraph 93) refers to the manufacturer's instructions or a risk-based schedule, with calibration potentially forming part of daily response checks. Paragraph 184 provides for testing, inspection and, where necessary, calibration at manufacturer-specified intervals, if available, or at appropriate intervals, with records maintained.
The OSHA Technical Manual (2014) describes practices by sensor type: oxygen sensors checked in fresh air, combustible-gas detectors checked daily, and toxic-gas detectors checked before and after use. This is an older document and should be read with its publication date in mind.
The OSHA Safety and Health Information Bulletin dated 26 November 2024, Calibrating and Testing Direct-Reading Portable Gas Monitors, distinguishes between:
- Bump test: A qualitative functional check confirming that test gas reaches the sensors and alarms activate, without assessing measurement accuracy.
- Calibration check: Comparison of the instrument reading against a test gas, within the manufacturer's specified tolerances.
- Full calibration (adjustment): Adjustment using a certified calibration gas.
The bulletin repeats the ISEA recommendation to perform a bump test or calibration check before each day of use.
An instrument that fails the check must undergo full calibration before use and must be removed from service if calibration fails.
This bulletin provides recommendations. It does not create new regulatory obligations and is not a standard.
It refers to the manufacturer's instructions and states that the gas used for adjustment must be traceable.
Terminology. The documents do not all use the same terms. This guide uses:
- Verification: The response test described in ED 6088, or the bump test or calibration check described in the OSHA bulletin.
- Metrological calibration: In the sense used by ED 6088, establishing a response curve without modifying the instrument.
- Adjustment: Changing zero and sensitivity settings, called calibrage by ED 6088 and full calibration by OSHA.
The definition in the cited document and the manufacturer's instructions should always be consulted.
US regulations and calibration: OSHA 1926.1204 requires equipment provided, including testing and monitoring equipment, to be properly maintained ((d) and (d)(1)), without specifying a calibration frequency (see the SHIB bulletin above).
INRS ED 6184 does not provide detailed calibration requirements and refers to ED 6088.
The expression “calibrated direct-reading instrument” appears in OSHA 1910.146 at paragraph (c)(5)(ii)(C), concerning alternative procedures for certain spaces.
This guide therefore does not prescribe a specific calibration frequency.
8Three Regulatory Frameworks Compared
| France | United States | United Kingdom | |
|---|---|---|---|
| Documents reviewed | INRS ED 6184, INRS web pages, R 447, Labour Code articles cited by INRS | OSHA 1910.146 (general industry), OSHA 1926.1204 (construction) | HSE L101, third edition (Confined Spaces Regulations 1997, applicable in Great Britain according to L101; Northern Ireland has a separate regulatory framework, not described here) |
| Nature of the documents | INRS/CNAMTS publications and recommendations; French Labour Code | Federal regulations | Approved Code of Practice (ACOP), with a specific legal status |
| Oxygen threshold cited | 19% recommended in these publications (ED 6184, R 447, ED 6088) | Hazardous atmosphere below 19.5% or above 23.5% (1910.146(b)): regulatory definition criterion, not an alarm setting | No numerical threshold in paragraphs 89 to 98 |
| Monitoring | Personal detector operating throughout the intervention (continuous atmospheric monitoring) | Continuous by default in construction; as necessary in general industry | Based on risk assessment; continuous when forced ventilation is used |
| Calibration | ED 6088 (not detailed in ED 6184) | 1926.1204(d): equipment must be properly maintained, with no specified calibration frequency. Non-binding recommendations: SHIB bulletin of 26 November 2024 | Manufacturer's instructions or risk-based schedule (paragraphs 93 and 184) |
| Ventilation and initial testing | Ventilation for at least 20 minutes, followed by atmospheric testing (INRS) | Construction: testing before any alteration of natural ventilation, then before entry, except for certain spaces that cannot be isolated (1926.1204(e)(1)) | Testing according to the risk assessment (L101) |
These documents differ in nature (recommendation, regulation, legal definition), scope and date.
A value from one framework cannot automatically be transferred to another.
9Equipment Used in Explosive Atmospheres
According to INRS ED 6088 (p. 11), detectors used in explosive atmospheres must comply with Directive 2014/34/EU (ATEX), as indicated by the relevant marking, and the Electromagnetic Compatibility Directive 2014/30/EU.
The same document states that removing protective components or connecting accessories may compromise ATEX compliance (p. 8).
INRS ED 945 (2020) refers to continuous monitoring of the internal atmosphere during maintenance operations (box in section 4.3, p. 18) and to a detection system that is regularly checked and calibrated (section 4.1, p. 16).
The IEC 60079-29-0:2025 standard concerning gas detector requirements is presented on a separate Indu X reference page.
Equipment must be suitable for the relevant zone and atmosphere, rather than merely carrying an ATEX marking.
INRS ED 945 (Table 4, p. 17) relates hazardous zones to equipment categories:
- Zones 0 and 20: Category 1.
- Zones 1 and 21: Category 2 or 1.
- Zones 2 and 22: Category 3, 2 or 1.
Directives 2014/34/EU and 2014/30/EU are European Union legislation. This guide does not describe equivalent requirements in the United States or the United Kingdom.
Frequently Asked Questions
In What Order Should Atmospheric Testing Be Performed?
Oxygen first, then flammable gases, then toxic contaminants, according to OSHA (1910.146(d)(5)(iii) and 1926.1204(e)(3)), HSE and NIOSH.
The INRS pages reviewed do not specify this order.
Is a Personal Gas Detector Required?
INRS describes a personal portable detector positioned as close as possible to the breathing zone.
INRS ED 6184 provides for one portable multi-gas detector per worker.
What Should Be Done if an Alarm Sounds?
If people are already inside the confined space, they must evacuate immediately (NIOSH; INRS ED 6184, p. 14; R 447).
During pre-entry testing, INRS describes an additional 20 minutes of ventilation, followed by another test. If the alarm sounds again, access must be secured and management informed.
This 20-minute procedure must never be interpreted as permission to wait inside the confined space.
Is an Oxygen Detector Sufficient?
INRS ED 6184 states that measuring oxygen alone does not establish whether the air is safe.
INRS also emphasises that gas detectors are not sufficient on their own.
Further Reading
- OSHA 29 CFR 1910.146 reference page
- INRS ED 6088 reference page: Portable Gas Detectors
- IEC 60079-29-0:2025 reference page
- All Regulations and Standards
- Gas Detection in the Oil and Gas Industry
- Products & Prices: Dated Market Data (No Product Recommendations)
- NewsPulse: Gas Detection Industry News
Sources and Dates Consulted
Sources consulted on 2 November 2026. Document dates are shown in parentheses.
- INRS, Espaces confinés : ce qu'il faut retenir and Prévenir les risques (25 February 2025); Réglementation (25 February 2025).
- CNAMTS/INRS, Recommendation R 447 (adopted 25 June 2009, February 2010 edition).
- INRS, ED 6184 (2014, updated 2016); ED 6088 (April 2022); ED 945 (December 2020).
- OSHA, 29 CFR 1910.146; 29 CFR 1926.1204; Technical Manual, Section II, Chapter 3 (11 February 2014).
- HSE, L101 Safe Work in Confined Spaces (third edition, December 2014).
- NIOSH, Working Safely in Confined Spaces (archived page concerning post-hurricane operations, last revised 10 March 2023).
- French Digital Labour Code, Articles R. 4222-23 and R. 4222-24 (displayed update date: 1 May 2008).
- OSHA, 29 CFR 1910.146, Appendix B; SHIB dated 26 November 2024, Calibrating and Testing Direct-Reading Portable Gas Monitors.
- HSE, Confined Spaces: Main Duties.