☎ (888) 264-7050
Colorado Class VI Primacy Starts November 12. Your CO2 Compressor Building Still Answers to Federal OSHA.
← Back to Blog Industrial Hygiene

Colorado Class VI Primacy Starts November 12. Your CO2 Compressor Building Still Answers to Federal OSHA.

Colorado's ECMC takes over Class VI carbon storage well permits on November 12, 2026. Here is what OSHA still requires at your CO2 capture and compression skid.

by iSi Environmental

On November 12, 2026, the agency that issues carbon storage well permits in Colorado changes. EPA has approved Colorado’s application for primary enforcement responsibility over Class VI underground injection wells, the well class used for geologic carbon sequestration. From that date the Colorado Energy and Carbon Management Commission (ECMC) issues and enforces those permits. The final rule, 91 FR 64780 (FR Doc. 2026-20846, docket EPA-HQ-OW-2025-2829), published in the Federal Register on October 13, 2026 and amends 40 CFR Part 147.

If you run an ethanol plant, a cement kiln, or a gas processing train that is about to have a capture and compression island bolted onto it, that handoff changes who you call about the well. It changes nothing about what you owe on the plant side. A Class VI permit covers the injection zone, the well bore and the monitoring program. It has nothing to say about the compressor building where a colorless, odorless gas heavier than air collects at floor level while a millwright works a valve.

Carbon dioxide kills by displacing oxygen, and it does it without the warning symptoms people expect from a toxic gas. At 70,000 to 100,000 ppm it produces unconsciousness within a few minutes, per NIOSH’s IDLH documentation for carbon dioxide. That hazard shows up with the project. The permit that approved the project does not address it, and neither does the process safety standard most plant managers assume would.

Does a Class VI permit cover safety at the carbon capture plant?

No. A Class VI permit governs the injection well: site characterization, area of review, well construction, mechanical integrity, monitoring and post-injection site care. Aboveground capture, dehydration and compression hazards fall under OSHA, at 29 CFR 1910.1000, 1910.146, 1910.134 and 1910.38. The two are separately permitted and separately enforced.

Read the equipment list on a typical capture project and the split becomes obvious. The amine or physical-solvent contactor, the regenerator, the dehydration skid, the multi-stage compression train that takes CO2 to supercritical pressure, the pipeline tie-in, the analyzer house, the sumps and the pits. None of that is in the Class VI permit file, because none of it is the well. It is a process unit on your property, operated by your people, inspected by a federal OSHA compliance officer who will never read your plume model.

That division has a practical consequence worth building into your document control. Keep the plant-side exposure and confined-space record separate from the Class VI permit file. Two agencies will ask for them, at different times, for different reasons, and a single binder that mixes injection-zone data with atmospheric monitoring records serves neither request well.

Does OSHA Process Safety Management apply to a carbon capture unit?

Not on carbon dioxide inventory alone. CO2 is not listed in Appendix A to 29 CFR 1910.119 and is not a Category 1 flammable gas, so PSM is not triggered by CO2 quantity. PSM may still apply where the process co-locates 10,000 pounds or more of a Category 1 flammable gas or a flashpoint-under-100-degree liquid in one location, per 1910.119(a)(1)(ii).

Facilities read that as relief. It is the opposite. PSM is the management system that forces process hazard analysis, written operating procedures, mechanical integrity programs and pre-startup safety review on the most dangerous material in a plant. Carbon dioxide skips all of it on a technicality of list membership. The most hazardous new substance arriving at the site shows up without the wrapper that would normally catch an uncontrolled release path.

EPA’s Risk Management Program at 40 CFR Part 68 is not triggered on CO2 alone either. So the facility that assumes an existing PSM or RMP program will absorb the new hazard gets nothing, and the facility that concludes the hazard must therefore be minor has drawn the wrong inference. The General Duty Clause applies. So do 1910.146, 1910.1000 and 1910.38. The obligation is there; the familiar program structure that usually carries it is not.

Who issues Class VI carbon storage well permits in Colorado?

From November 12, 2026, the Colorado Energy and Carbon Management Commission issues and enforces Class VI permits statewide, under EPA’s primacy approval at 91 FR 64780. EPA Region 8 remains the permitting authority for Class VI wells on Indian lands within Colorado and continues to oversee the state program.

EPA’s oversight under the rule is prospective and specific: quarterly ECMC reports on permittee noncompliance, plus annual UIC program performance reports. The same rule codifies previously approved non-substantial changes to Colorado’s Class II program. Colorado becomes the fifth Class VI primacy state, behind North Dakota, Wyoming, Louisiana and West Virginia, with Texas and Arizona in process.

There is no ECMC Class VI enforcement history to cite, and any article claiming to describe ECMC enforcement trends today is making it up. The authority starts on November 12.

When does Colorado’s Class VI primacy take effect?

November 12, 2026. The final rule approving Colorado’s Class VI primacy application, 91 FR 64780, published October 13, 2026 and amends 40 CFR Part 147. Until the effective date, EPA remains Colorado’s Class VI permitting authority. Confirm which agency holds your file before you submit anything in the intervening weeks.

If your project sits on or adjacent to Indian lands, that confirmation is not a formality. The jurisdictional line runs through the middle of the primacy grant, and the answer determines who reviews the application, who receives the monitoring reports, and who shows up for the inspection.

What is the OSHA exposure limit for carbon dioxide?

5,000 ppm (9,000 mg/m3) as an 8-hour time-weighted average, listed for CAS 124-38-9 in Table Z-1 of 29 CFR 1910.1000. NIOSH sets the same 5,000 ppm TWA plus a 30,000 ppm short-term exposure limit (54,000 mg/m3), and a revised IDLH of 40,000 ppm.

Those three numbers are the design basis for everything else. Personal exposure monitoring for operators and maintenance crews working the capture, dehydration and compression units gets measured against the 5,000 ppm TWA, with the sampling basis documented so the result means something twelve months later. Fixed CO2 and O2 detection in enclosed compression and analyzer buildings and at below-grade locations gets alarm setpoints tiered against the TWA, the STEL and the IDLH, with audible and visual output.

A single sampling round does not characterize an exposure distribution. If the project changes the work, the monitoring has to cover the shifts and the tasks where exposure actually concentrates, which on a compression skid usually means filter changes, instrument work and anything that involves breaking containment at grade.

At what concentration does carbon dioxide become immediately dangerous?

NIOSH sets the IDLH for carbon dioxide at 40,000 ppm (4 percent), revised down from 50,000 ppm on human acute inhalation data. NIOSH documentation reports signs of intoxication after about 30 minutes at 50,000 ppm, and unconsciousness within a few minutes at 70,000 to 100,000 ppm.

Four percent sounds like a wide margin until you consider the volume of a compressor building and the mass flow through a failed seal. The gap between a nuisance alarm and an IDLH atmosphere in an enclosed space is minutes, not hours, and nobody inside will smell it or taste it on the way there. Set the alarm tiers so the first one means leave, and train people that the first one means leave.

Is a carbon dioxide compressor building a permit-required confined space?

It can be. Under 29 CFR 1910.146(b) a space is permit-required if it “contains or has a potential to contain a hazardous atmosphere,” and an atmosphere below 19.5 percent oxygen by volume qualifies. CO2 is roughly 1.5 times the density of air, so it settles into enclosed buildings, pits, sumps, trenches and below-grade vaults instead of dispersing upward.

Section 1910.146(c)(1) puts the obligation on the employer to evaluate the workplace and determine whether any space is a permit-required confined space. The timing matters more than the paperwork. Do the re-inventory when the capture project is scoped, not after it is commissioned, because new compression and dehydration equipment creates spaces that did not exist and changes the atmosphere in spaces that did. An analyzer shelter that was a non-permit space on Tuesday can be a permit space after a tie-in on Friday.

Ethanol and biofuel plants deserve a specific note here. Fermentation CO2 is already a confined-space hazard at those sites before any capture equipment arrives, which means the existing program has usually identified some CO2 spaces and has almost never identified the ones a compression island adds. The same standard, 1910.146(b), also flags oxygen-enriched atmospheres above 23.5 percent, which is relevant where a project brings bottled oxygen or an oxy-fuel stream onto the same pad.

How do you test a confined space for carbon dioxide before entry?

29 CFR 1910.146(d)(5)(iii) requires testing in a fixed order: oxygen first, then combustible gases and vapors, then toxic gases and vapors. Instruments have to be specified and calibrated for CO2 as well as oxygen. A catalytic-bead LEL sensor reads nothing on CO2, and an oxygen-only meter alarms after significant displacement has already happened.

This is where most programs fail an audit after a retrofit. The written procedure is correct, the entry permits are filled out, and the four-gas meter in the gang box has no CO2 channel. It was bought for a plant that did not have a carbon dioxide inventory. Check the sensor configuration against the hazard list for each space, check the calibration records, and check that the people doing pre-entry testing know why the order is oxygen first.

Will respirators protect workers from a carbon dioxide release?

Air-purifying respirators will not. Cartridges and filters do not remove carbon dioxide and they do not add oxygen, so they offer no protection in a CO2-enriched or oxygen-deficient atmosphere. 29 CFR 1910.134 requires supplied-air respirators or SCBA for oxygen-deficient and IDLH atmospheres, inside a written program that includes fit testing.

The rescue side of this is where people die. Under 1910.146(k) the employer has to have rescue capability in place, and in asphyxiant atmospheres the fatality pattern is consistent: the people who die are overwhelmingly would-be rescuers who went in without air. Build a non-entry rescue capability with retrieval systems, equip responders with SCBA, and drill it on the actual spaces at the actual site.

Coordinate with the local fire district before startup. Most rural Colorado departments have not run a dense-phase CO2 scenario, and the first time they see your compression building should not be the day they are asked to pull someone out of it.

Where should workers evacuate during a CO2 release?

Uphill and upwind. Because carbon dioxide is denser than air it flows downslope and pools in low ground, so a conventional low-lying muster point can concentrate people in the release instead of protecting them. 29 CFR 1910.38 requires an emergency action plan with evacuation routing and alarm procedures.

Walk your site with a topographic map and find out where your muster point sits relative to the capture island. Plenty of plants put it in the parking lot because the parking lot is paved, open and easy to count heads in. If the parking lot is downhill of a compression skid, it is a design error that costs nothing to fix and can be fixed this week.

The alarm itself needs the same attention. A CO2 alarm has to be distinguishable from a fire alarm, because the two call for opposite movements. A fire alarm sends people out and down a stairwell. A CO2 alarm has to send them up and crosswind. One tone for both means the first real release gets answered with the wrong reflex.

What are the OSHA penalties for confined space and exposure violations at a capture site?

OSHA’s current maximums are $16,550 per serious or other-than-serious violation, $16,550 per day beyond the abatement date for failure to abate, and $165,514 per willful or repeated violation, per OSHA’s penalties page. Confined-space cases typically generate multiple instance-by-instance citations rather than one.

The citation pattern in asphyxiant atmospheres is rarely a single missing permit. It is the chain: no reclassification of a space after a process change, atmospheric testing in the wrong order or with the wrong sensor, no rescue capability beyond calling 911, and an emergency action plan nobody updated for the new hazard. Those are separate paragraphs of the standard, so one incident produces several serious citations at $16,550 each. A second occurrence at the same site converts them to repeat at $165,514.

One honest note on the trajectory. The 2026 federal civil-penalty inflation adjustment was cancelled, so these figures are flat year over year rather than climbing. The reason to act is the incident, not a rising fine schedule.

Is Colorado a state-OSHA state for carbon capture projects?

No. Colorado has no OSHA-approved State Plan, so private-sector employers are under federal OSHA jurisdiction, Region 8 in Denver, per OSHA’s State Plans page. Wyoming, Colorado’s partner under the carbon storage memorandum of understanding signed May 4, 2026, is a State Plan state covering private employers. One project can span two enforcement regimes.

There is no Colorado-specific confined-space or carbon dioxide standard to layer on top of the federal requirements, and anyone describing a “Colorado OSHA” program to you has it wrong. The federal standards are the whole of the occupational-safety obligation in Colorado.

Wyoming is the live complication, and the MOU is what makes it live. Wyoming holds Class VI primacy and runs its own State Plan covering private employers. A storage project that crosses the state line can face ECMC and Wyoming’s UIC authority on the well side, and federal OSHA plus Wyoming OSHA on the worker-safety side. Multi-state operators who standardize one written program across both sides of the line should check it against the state-plan requirements rather than assume federal parity.

For the rest of the region iSi covers, no state currently holds Class VI primacy. Kansas, Oklahoma, Missouri and Nebraska remain under EPA for Class VI; Texas is in process. Kansas Class I underground injection is a separate program with separate requirements and should not be read across to Class VI. If that is the well class you are dealing with, the Class I injection well permitting rules by state are a different animal entirely.

What the permitting headline leaves out

The clean version of this story is that Colorado gets primacy, carbon capture accelerates, and everyone should prepare for a wave of projects. The counter-signals are worth carrying.

Class VI permit applications are slowing in several 2026 industry analyses even as approval throughput improves, which means primacy relieves a bottleneck that is no longer the binding constraint for every developer. A national backlog of pending Class VI applications persists; EPA does not publish a current count on its Class VI pages and directs users to the Class VI Data Repository, with data as of May 2026, so treat any specific backlog figure or state permit-timeline commitment you see quoted as contested until it comes from the repository or the agency directly.

Primacy has also been contested elsewhere. Louisiana’s Class VI primacy drew litigation and survived judicial review, and state-level pauses have occurred during 2026. Colorado’s grant is a final rule publishing October 13, 2026 and sits inside the ordinary judicial-review window, so it is not beyond challenge. And ECMC’s own program material is explicit that transferring regulatory authority approves no individual storage project. The number of permitted Class VI wells in Colorado on November 12 does not change because of this rule.

Here is the part that holds no matter which way the permitting story breaks. Every one of those counter-signals is about the well. None of them changes the plant-side hazard by one part per million. Whether Colorado issues two Class VI permits or twenty, the ethanol plant that adds a compression skid has a new oxygen-displacement hazard inside an enclosed building, governed by a standard that has not changed and enforced by an agency that is not ECMC.

Four questions that tell you whether your confined-space program survived the retrofit

Run these against your own site. They take an afternoon and they find most of what an inspector would.

  1. Were your confined spaces re-evaluated when the capture project was scoped? Not after commissioning, and not as a line item in the construction package. If the space inventory predates the project drawings, it is out of date by definition under 1910.146(c)(1).
  2. Do the meters your entrants carry actually read CO2? Pull one out of the gang box and look at the sensor configuration and the calibration sticker. An oxygen channel alone is not enough.
  3. Does your rescue plan depend on somebody volunteering to go in? If the written plan is “call the fire department,” you do not have rescue capability under 1910.146(k). Retrieval systems and SCBA-equipped responders are the control.
  4. Does your emergency action plan move people uphill? Find your muster point on a contour map. If it is below the capture island, relocate it.

If the answer to the first question is no and the answer to the second is “I would have to check,” the gap is on the plant side and no Class VI permit is going to close it.

The work that closes it is ordinary industrial hygiene and safety consulting: a space-by-space re-evaluation after the project scope is fixed, personal exposure monitoring against the 5,000 ppm TWA with a documented sampling basis, instrument and alarm specification tied to the TWA, STEL and IDLH tiers, a confined space and emergency response program rewritten for an asphyxiant that gives no warning, and a drill with the people who would actually respond. iSi does this work at industrial facilities in 40 states, and we are not well permitters. The well is ECMC’s business after November 12. The compressor building is yours.

If your site is adding capture and compression and you want the plant-side program looked at before startup rather than after an incident, contact iSi Environmental and ask for an industrial hygiene and confined-space evaluation of the new scope.

Sources

Tags: