How often OSHA requires exposure monitoring, by substance

This page explains what the rules and common practice call for. Every site is different, and what a survey or a filing needs at your facility is something we work out with you after we see it.

There is no single answer to how often air monitoring is required.

OSHA has no general schedule for exposure monitoring. Each substance-specific standard sets its own trigger for initial monitoring, its own schedule for periodic monitoring, and its own conditions for when you may stop. They differ in ways that matter:

• Benzene is once a year where silica is every 6 months.
• Cadmium has no quarterly schedule at all.
• The general industry asbestos standard has no action level.
• Lead records have to be kept 10 years longer than everything else.

Get one of those wrong and you either overspend on monitoring you were never required to do, or you monitor too little and hand an inspector a citation.

This page gathers the requirements into tables, and the tables cite the paragraph of each standard so you can look it up.

Monitoring frequency by substance

All values are for OSHA's general industry standards, which cover plants, shops and most other fixed workplaces. The construction and maritime standards differ, as the note at the end explains.

A few terms first:

• Permissible exposure limit (PEL). The enforceable limit on what a worker may breathe in, expressed here as an 8-hour time-weighted average (TWA).
• Action level. A lower trigger level, at or above which most of these standards require you to continue monitoring.
• Short-term exposure limit (STEL). A limit over a short period, such as 15 minutes. The asbestos standard calls its short-term value an excursion limit.
• Units. µg/m³ means micrograms per cubic meter of air, ppm means parts per million, and f/cc means fibers per cubic centimeter of air.

Asbestos is the outlier. For every other worker, the asbestos standard sets no interval at all and asks only for monitoring at a frequency and pattern that represents exposures "with reasonable accuracy."

When you may stop monitoring

This is where the standards differ most, and where mistakes cost the most. If you stop one sample too early, your decision to discontinue monitoring will not be defensible when someone reviews it.

The distinction that matters is initial monitoring versus periodic monitoring.

Almost every standard in this table applies its confirmation requirement to periodic monitoring only. An initial result below the action level lets you stop with no confirming round at all. That is true for:

• Silica
• Hexavalent chromium
• Beryllium (the initial result must also be at or below the STEL)
• Lead. The measurements need not be repeated, except where the lead standard's restart triggers apply (see the next section).
• Benzene
• Methylene chloride. Below the action level and at or below the STEL, no 8-hour TWA monitoring or STEL monitoring is required.

Cadmium is the exception. The cadmium standard requires a confirming round after the initial monitoring or the periodic monitoring, so cadmium never lets you stop on one sample.

A note on wording: the standards use two phrasings that look like different requirements but are not. Some say the result must be "confirmed by another monitoring taken at least seven days later" (hexavalent chromium, cadmium). Others say "two consecutive measurements, taken 7 or more days apart" (silica, beryllium, lead, benzene, methylene chloride).

Both come to the same thing: two results below the action level, at least 7 days apart. The standards do not differ in the arithmetic. They differ in whether the requirement reaches the initial round at all.

What restarts the clock

Every standard here requires additional monitoring when something changes. The wording varies slightly, but the trigger is consistent enough to state plainly. You must monitor again when either of these is true:

• A change in production, process, control equipment, personnel or work practices may reasonably be expected to result in new or additional exposures.
• You have any reason to believe such exposures have occurred.

Benzene and methylene chloride spell out a second trigger: spills, leaks, ruptures or equipment breakdowns.

Formaldehyde adds one that catches employers off guard. If you receive reports of signs or symptoms of respiratory or skin conditions associated with formaldehyde exposure, you must promptly monitor the affected worker's exposure. It does not matter what the last sample showed, and there is no waiting period. A worker complaint is a monitoring trigger.

In practice this is the requirement most often missed. A plant adds a shift, changes an abrasive, moves a weld cell, or loses the operator whose exposure the whole assessment was built on, and the exposure assessment quietly goes out of date. The schedule table above gives you the minimum, and a process change resets it.

Employee notification: same deadline, three different starting points

Every standard in this table gives you 15 working days. You may notify each worker in writing, or post the results in a location workers have access to.

But the standards do not all start the clock at the same moment. When the laboratory is slow, that difference is real:

Suppose your laboratory takes 10 business days and you are working on a silica exposure assessment. The clock that starts when the assessment is complete may already be running while you wait. If you treat every standard as if its clock starts "when the results land," you end up with a late notice.

Where the exposure was above the PEL, the written notice must say more: it must state that the PEL was exceeded and describe the corrective action being taken.

This is an easy citation to avoid and an easy one to collect. The sampling gets done, the report comes back, the results go in a file, and the workers are never told. Having the report does not satisfy the notification requirement.

How long to keep records, and the one that isn't 30 years

Most exposure records fall under OSHA's standard on access to employee exposure and medical records. That standard says at least 30 years.

But read the opening line of that standard's paragraph on record retention. The 30 years applies "unless a specific occupational safety and health standard provides a different period of time." That clause is not decoration. It is the hook that lets a substance-specific standard override the default.

Lead does exactly that. The lead standard requires lead exposure monitoring records to be kept at least 40 years, or for the duration of employment plus 20 years, whichever is longer.

If you run one 30-year policy across all exposure records, your lead records are short by at least 10 years.

(Separately: employee medical records under the same standard are kept for the duration of employment plus 30 years, which is a different clock from exposure records.)

Nine details that are easy to get wrong

These are the nine points where the text of the standard most often differs from the general rule people carry in their heads.

1. Above the PEL does not always mean quarterly monitoring. It does for silica, hexavalent chromium, beryllium, methylene chloride and lead. It does not for cadmium, benzene, formaldehyde or asbestos.

Cadmium has no quarterly schedule at all. The cadmium standard puts one floor under everything at or above the action level, but read that paragraph carefully before you decide 6 months is your schedule. It requires monitoring at a frequency and pattern that represents workers' exposure levels, and where exposures are above the PEL the monitoring must do more. It must also assure that the respirators selected are adequate and that the engineering controls and work practice controls are effective. The 6 months is an "at least."

So above the PEL, 6 months is a minimum, not a schedule. That broader obligation can require more, and an inspector will read it that way.

Benzene above the PEL is 6 months, not 3.

2. Asbestos has no action level. The definitions in the asbestos standard do not include the term. 0.1 f/cc is the PEL, and the two are easy to confuse.

3. Benzene does not follow the 6-month / 3-month pattern. Benzene is once a year at or above the action level and 6 months above the PEL, which makes it the slowest schedule in the table. Treating it like silica means monitoring twice as often as required.

4. Formaldehyde has no schedule above the PEL. Formaldehyde's periodic monitoring is tied to the action level and the STEL, not the PEL. The standard sets 6 months at or above the action level, and once a year, under worst-case conditions, at or above the STEL.

5. A confirming round is not always required before you can stop. On an initial result, most of these standards let you stop outright.

• Silica. With an initial result below the action level, you stop with no confirming round. The requirement for two measurements 7 days apart covers later rounds only.
• Hexavalent chromium. Same structure. You may stop on an initial result below the action level, and the single-confirmation requirement is expressly limited to periodic monitoring.
• Beryllium. Same again. You may stop on an initial result below the action level and at or below the STEL, and the two-consecutive-measurement requirement applies to later monitoring.
• Lead. With an initial result below the action level, "the measurements need not be repeated" except where the restart triggers apply. The two-consecutive-measurement requirement applies when results are at or above the action level. A separate one, which calls for results below the PEL but at or above the action level, moves you from quarterly back to every 6 months.
• Benzene. You may stop on an initial result below the action level, and the two-consecutive-measurement requirement covers periodic monitoring.
• Methylene chloride. Below the action level and at or below the STEL, no 8-hour TWA monitoring or STEL monitoring is required. The 7-day requirement covers periodic monitoring only.
• Cadmium is the exception. It requires a confirming round after the initial monitoring or the periodic monitoring.

This costs money in both directions. A confirming round taken after an initial lead, benzene or silica result was never required, while one skipped on cadmium was required.

6. Beryllium monitoring is not triggered at the action level. The beryllium standard requires an exposure assessment for each worker who is exposed to airborne beryllium, or may reasonably be expected to be, with no threshold attached. That is a much broader trigger than silica's, which expressly covers only exposure at or above the action level.

7. Existing data does not replace periodic monitoring for lead or cadmium. Some standards allow objective data in place of new monitoring. There are two points here, and the second is the useful one.

The lead standard has no objective data option at all. The term does not appear in its exposure monitoring paragraph. What lead has is earlier sampling data with a hard limit: the standard lets you rely on "measurements of airborne lead made in the preceding 12 months" and stops there. Data older than 12 months does not qualify, no matter how well it represents the job.

Cadmium does have both. One paragraph covers earlier monitoring results, and another covers objective data.

But neither one carries periodic monitoring, and the reason is subtler than a prohibition. By their own terms, the cadmium paragraphs cover initial monitoring only. One satisfies the initial monitoring requirement, and the other applies "instead of implementing initial monitoring." The paragraph on periodic monitoring allows no substitute, and the lead standard is built the same way.

The limit comes from what the paragraphs cover, not from a stated prohibition. Neither standard has a sentence that says "objective data shall not be used for periodic monitoring." The standards simply do not authorize it. The practical result is the same: objective data will not carry an ongoing lead or cadmium monitoring program.

Silica, hexavalent chromium, beryllium and formaldehyde are different. Their standards contain genuine performance or objective data options that can carry the exposure assessment.

8. Not all exposure records are kept 30 years. Lead is 40 years, or employment plus 20, whichever is longer, as described above.

9. The respiratory protection standard does not require air monitoring. It requires an assessment, not a sampling schedule. OSHA's respiratory protection standard says the employer must identify and evaluate respiratory hazards, and a "reasonable estimate of employee exposures" must be part of that evaluation. Just as important, where the employer cannot identify or reasonably estimate the exposure, the atmosphere must be treated as immediately dangerous to life or health.

No monitoring schedule is set anywhere in the respiratory protection standard, so it is not the source of a sampling schedule. But you cannot ignore it either: issuing respirators is no substitute for knowing what people are breathing.

Three things this table can't tell you

Construction is different. These are general industry standards. Construction has its own silica standard, its own lead standard and its own asbestos standard, so if your work is construction, the schedules above do not apply.

The construction silica standard has a Table 1. The general industry standard does not. Table 1 lets construction employers follow a table of specified controls and respiratory protection for listed tasks, in place of assessing exposures. The general industry silica standard has nothing like it. General industry employers get the performance option, which rests on data, not on a task list. Construction guidance is easy to land on first, but if your work is general industry, the Table 1 route is not open to you.

Substances without their own standard still have limits. Several hundred air contaminants have no standard of their own and sit instead on OSHA's general air contaminant tables, known as Table Z. Each has a PEL but no set monitoring schedule at all. The obligation there comes from the General Duty Clause of the Occupational Safety and Health Act and, if respirators are in use, from OSHA's respiratory protection standard. That is a judgment call, not a lookup. Note also that Table Z substances carry no employee notification requirement of the kind the substance-specific standards impose.

A note on what "compliant" means here

Everything above is the minimum. Meeting it keeps you off an inspector's citation list, but it does not necessarily mean your workers are protected.

Most OSHA PELs date from 1971 and were adopted from voluntary standards older than that. For a number of substances, the current threshold limit values (TLVs) published by ACGIH are meaningfully more protective, and in some cases the gap is a factor of ten.

Monitoring to the PEL and stopping there is legally defensible but medically incomplete. It also does not shield an employer from General Duty Clause liability, or from a lawsuit, where a more protective limit was well known and available.

It is worth knowing which of your substances that applies to before you build a monitoring program around this table.

Compiled from the current text of OSHA's general industry standards at osha.gov, cross-checked against eCFR, the electronic Code of Federal Regulations. Standards are amended, so check the current text before you rely on any value for a compliance decision.

Not sure which of these applies to your operation, or whether your current program is on the right schedule? Send us your list of substances and job classifications, and we will tell you what the standards require. → Contact us