COSHH 2002 and EH40 workplace exposure limits, UK wide
COSHH Assessments
Occupational lung disease and skin disease build slowly. By the time a worker reports symptoms the exposure has usually been running for years. AL23 Safety assesses the substances your work creates or uses, compares exposure against the current workplace exposure limits and sets out controls that hold up under inspection, for clients UK wide.
What is a COSHH assessment?
A judgement about exposure, not a folder of safety data sheets
Regulation 6 of the Control of Substances Hazardous to Health Regulations 2002 requires a suitable and sufficient assessment of the risk to health before the work starts. A safety data sheet is an input to that assessment. It is not the assessment. It describes the substance rather than your task, your quantity, your ventilation and your people.
The substances that cause most harm are often the ones nobody bought. Respirable crystalline silica, wood dust, welding fume and flour dust are generated by the work itself. They never arrive with a data sheet, which is why an assessment covering only labelled containers misses the exposures most likely to make someone ill.
COSHH deals with harm to health. Where the hazard is fire or explosion the duty sits elsewhere, covered on our DSEAR risk assessment page.
Who needs one
Anyone whose work creates a dust, fume, mist, vapour, gas or biological agent
- Machining, cutting, grinding, blasting or sanding stone, concrete, brick, tile or any other material that generates dust
- Soldering, welding, brazing or hot cutting of any metal, indoors and outdoors
- Woodworking, joinery, furniture manufacture and MDF machining
- Use of solvents, adhesives, resins, isocyanate paints or laboratory reagents
- Food production involving flour dust, enzymes or cleaning-in-place chemicals
- Any workplace where health surveillance has flagged dermatitis, asthma or a lung function change
Engineered stone fabrication deserves a separate mention. HSE published its first COSHH guidance sheet for engineered stone in May 2026. The practical effect is that dry cutting, grinding and polishing should not happen unless a business can demonstrate equally effective alternative controls. A nationwide campaign launched alongside it, with more than 1,000 inspections planned across 2026 and 2027. Four companies had received prohibition notices by the end of May 2026. HSE research found dry fabrication generates respirable crystalline silica exposure five to ten times higher than wet methods.
The regulations that apply
Your legal framework, in plain terms
The Control of Substances Hazardous to Health Regulations 2002 sit beneath the Health and Safety at Work etc. Act 1974. The Approved Code of Practice is L5, sixth edition, published in 2013. Exposure limits are published separately in EH40/2005 Workplace exposure limits.
Regulation 6
A suitable and sufficient assessment before work starts, reviewed regularly, when no longer valid, on significant change and when health surveillance results suggest it. There is no fixed statutory review interval.
Regulation 7
Prevent exposure first. Where prevention is not reasonably practicable, control it adequately by applying the eight principles of good control practice in Schedule 2A. Protective equipment comes last, never first.
Regulation 7(7)
Where a substance has a workplace exposure limit, exposure must be below it. For carcinogens, mutagens and asthmagens, exposure must be reduced as low as is reasonably practicable whatever the limit says.
Regulation 9
Control measures must be maintained, examined and tested. This is where the statutory examination of local exhaust ventilation comes from.
Regulations 10 to 12
Exposure monitoring where required, health surveillance where an identifiable disease is linked to the exposure and training for everyone affected.
The limits people ask about most
Respirable crystalline silica
0.1 mg/m³ as an 8-hour time-weighted average, classified as a carcinogen. The ALARP duty therefore applies on top of the limit.
Hardwood dust
3 mg/m³ inhalable as an 8-hour time-weighted average, classified as a carcinogen and a sensitiser.
Softwood dust
5 mg/m³, classified as a sensitiser. Where hardwood is mixed with any other wood dust, the 3 mg/m³ hardwood limit applies to the whole mixture.
Welding fume
There is no workplace exposure limit for welding fume. Following the reclassification of all welding fume, including mild steel fume, as a Group 1 human carcinogen, local exhaust ventilation is required for all indoor welding. Respiratory protection is needed in addition where extraction cannot achieve adequate control. General ventilation alone is not acceptable.
The current list of limits is EH40/2005, fourth edition, published in January 2020. It remains the current edition as at August 2026. Treat any article advertising a newer edition with caution.
What we assess
Every substance and the tasks that put it in the air
- Every substance used, stored or generated, including process dusts and fumes with no data sheet
- The task itself, covering quantity, duration, frequency, energy input, route of entry and how much becomes airborne
- Exposure against the relevant EH40 limit, including the mixture rule where more than one wood dust is present
- Whether any substance is a carcinogen, mutagen or asthmagen, which raises the duty beyond meeting a limit
- Substitution options and the engineering control in place, covering enclosure, on-tool extraction, water suppression and segregation
- Condition and adequacy of local exhaust ventilation and whether its statutory examination is current
- RPE selection, adequacy, face-fit testing records and clean-shaven policy
- Skin protection, glove breakthrough data, washing facilities and skin care regimes
- Health surveillance arrangements, including who is in scope and how results feed back into control
Our process
Build the inventory, judge exposure, test the controls
- 01
Substance and task inventory
We walk the process and build the inventory from what the work generates, not only from the chemical store.
- 02
Exposure evaluation
We judge likely exposure task by task against EH40 limits and advise where air monitoring is needed to settle a question rather than guessing.
- 03
Control assessment against Schedule 2A
We test existing controls against the eight principles of good control practice in order, identifying where protection is covering an engineering gap.
- 04
Assessment and action plan
We issue task-based assessments written for the people doing the work, with a prioritised action plan behind them.
- 05
Embedding and review
We brief supervisors, set the review triggers and agree what happens when a substance, supplier or process changes.
What you get
Assessments the operator can actually work from
- Task-based COSHH assessments naming the substance, exposure route and control, written to be used at the bench
- A substance inventory cross-referenced to safety data sheets and EH40 limits
- A prioritised control action plan following the Schedule 2A hierarchy, with owners and target dates
- An RPE specification with face-fit testing requirements and a maintenance and storage regime
- A health surveillance recommendation naming the roles in scope and the trigger for each
- A monitoring recommendation where exposure cannot be settled by inspection alone
What we need from you
Your substances, your processes and honest access
- A list of purchased substances with current safety data sheets
- Process descriptions, quantities used and typical task durations
- Any existing local exhaust ventilation examination reports and air monitoring results
- Details of RPE issued and the face-fit records behind it
- Access to the work as it normally runs, including cleaning and maintenance
Why AL23 Safety
Built around the dust and fume that cause the disease
Accountable
If controls rely on RPE where extraction should be doing the work, the report says so plainly and sets out the fix.
Built around the real hazard
We start with process-generated dust and fume, where the enforcement and the disease both sit.
Practical, not just compliant
Assessments are written for the operator and the supervisor. A document nobody reads controls nothing.
UK wide
We assess COSHH risk for clients across the UK from our Manchester base, in workshops, factories, laboratories and on site.
Talk to us about COSHH
A straight read on where your exposure control actually stands
If you are unsure whether your dust or fume control would survive an inspection, describe the process on a call and we will tell you what we expect to see. No obligation and no pressure.
Common questions
Answers, up front
Cannot see your question? Get in touch and we will answer it directly.
Contact usCost follows the number of distinct substances and, more importantly, the number of distinct tasks. Fifty products used in three simple ways is a smaller job than eight products used forty different ways. Air monitoring or a face-fit programme adds to it. Tell us the process and the headcount and we will scope it rather than price by the data sheet.
A single workshop is usually one site visit with assessments issued afterwards. Multi-process manufacturing takes longer because the inventory has to be built task by task. We agree scope and delivery date before starting.
COSHH sets no fixed interval. Regulation 6 requires review regularly, whenever there is reason to believe the assessment is no longer valid, on significant change and where health surveillance suggests the controls are not working. A change of supplier, substance, process or ventilation is a trigger. Many clients also set an annual sense check.
On engineered stone, inspectors are checking whether lower-silica material has been considered, whether on-tool water suppression is fitted and used, whether RPE has been face-fit tested, whether health surveillance is running, whether extraction is present and effective and whether machinery is guarded. Wet methods are the expectation. The same thinking is spreading across other high-dust trades, which is why we pair the assessment with health and safety training.
If you rely on local exhaust ventilation as a control, then yes. Regulation 9 requires control measures to be maintained, examined and tested. For most LEV that means a thorough examination and test at least once every 14 months. We cover the intervals and what an examination involves on our LEV testing page.
Related services