Subscribe to our newsletter to receive the latest news and events from TWI:

Subscribe >
Skip to content

Welding fume - Do you know your WEL? (July 2006)


Paper published in Health & Safety International, July 2006; (also in Welding & Cutting, No 4 2007).

Dr Dave McKeown, Manager, Corporate Projects at TWI Ltd, the Research and Technology Organisation focussed on joining, reviews the generation and control of hazards from welding and cutting.


Much has been written about the hazards associated with welding and cutting. Many are well understood, even if not always proficiently guarded against in the workplace; eg ultra-violet radiation, electrical hazards, burns, fire and explosion risk. But perhaps the most talked about and least understood is the hazard from fume. This article takes a look at the factors involved and attempts to demystify the technicalities. Much of the content is based on a series of frequently asked questions and job knowledge articles published on TWI's website. [1]

Welder with local fume extraction
Welder with local fume extraction

What is welding fume?

We really should start from a definition of what we are talking about. Fume is a mixture of particles generated by vaporisation, condensation and oxidation of substances transferred through the welding arc. The particles are very small and remain suspended in the air for long periods, where they may be breathed. Small particles are respirable which means that they may penetrate the innermost regions of the lung where they have the most potential to do harm. If inhaled, welding fume may be hazardous to health and must be controlled to limits laid down by regulations.

Toxic gases may also be generated during welding and cutting. Gases encountered in welding may be:

  • Fuel gases which, on combustion, form carbon dioxide and, if the flame is reducing, carbon monoxide
  • Shielding gases such as argon, helium and carbon dioxide, either alone or in mixtures with oxygen or hydrogen
  • Carbon dioxide and monoxide produced by the action of heat on the welding flux or slag
  • Nitric oxide, nitrogen dioxide and ozone produced by the action of heat or ultraviolet radiation on the atmosphere surrounding the welding arc
  • Gases from the degradation of solvent vapours or surface contaminants on the metal.

The degree of risk to the welder's health from fume/gases will depend on:

  • Composition
  • Concentration
  • The length of time the welder is exposed

We need to approach these variables in a way that reduces risk to a level that will not damage welders' health.

Risk assessment

To find out if our welders are at risk we need to consider the working conditions. The Management of Health and Safety at Work Regulations 1999 require that employers assess the risks to health of employees arising from their work. The actions arising from the risk assessment are dictated by other more detailed regulations, the Control of Substances Hazardous to Health (COSHH) Regulations 2004.

Hazards to health from welding fume arise from inhalation of the fumes. A risk assessment involves estimating the exposure of workers to the fume and considering the steps required to control that exposure, if prevention is not reasonably practicable. Further information on risk assessment and estimation of exposure is available in the document, 'COSHH assessments', available from the Health and Safety Executive.

It is essential to know the type of parent plate, together with any coating, and the composition of the fume generated. This is because different fume components vary in toxicity. The limits to which welding fume and its component parts must be controlled are provided in Guidance Note EH40 'Workplace Exposure Limits' available from the Health and Safety Executive (HSE). This is updated annually. The 2005 edition changed the way that exposure to fume is rated.

Previously, occupational exposure limits (OEL) were set as maximum exposure limits (MEL) and occupational exposure standards (OES). Both of these have been discontinued in favour of a single type of OEL known as the 'workplace exposure limit' or WEL.

MELs and OESs were introduced in 1989, when the first COSHH regulations came into force, but the WEL approach superseded them on 6 April 2005. The underlying principle was to make the system more easily understood by having a single criterion for exposure to airborne hazards. However, a further aspect of the new framework is that certain OESs have not been converted to WELs because of doubts that the limit was soundly based. Welding fume is one of those hazards that do not have a WEL. Individual components must therefore be addressed.

What's in the fume?

Exposure to fume may be measured according to the methodology defined in BS EN ISO 10882-2: 2000. Account must be taken of the exposure limits of the individual fume constituents. For example, iron oxide, limestone, titanium dioxide have WEL of 4 or 5 mg.m-3 and have no notes in EH40 of specific hazard. They may therefore be taken to be similarly hazardous to any dust - not specifically causing a medical condition but needing control to ensure proper lung function. Some components of fume have lower WEL, manganese, trivalent chromium and soluble barium are set at 0.5 mg.m-3, copper at 0.2 mg.m-3, though these are still without specific warning of health hazards. But hexavalent chromium compounds and nickel oxide are noted in EH40 as potential carcinogens and pose greater health risks at lower concentrations. Nickel and its water-insoluble compounds have WEL of 0.5 mg.m-3 and hexavalent chromium compounds only 0.05 mg.m-3. These exposures are over a time-weighted average reference period of 8 hours.

Clearly, welding stainless steel, likely to generate both nickel and chromium in the fume, poses a very different set of conditions than welding mild steel.

What about gases?

Exposure to gases may be measured according to the methodology defined in BS EN ISO 10882-2: 2000. Health and safety in welding and allied processes - Sampling of airborne particles and gases in the operator's breathing zone - Part 2: Sampling of gases. Exposure measurement may be used to verify compliance with regulations, identify a need for exposure control or to identify faults with existing control systems.

For gas shielded welding processes such as TIG, MIG/MAG, FCAW, shielding gases may be inert gases, such as argon, helium and nitrogen, or argon-based mixtures containing carbon dioxide, oxygen or both. Helium may be added to argon/carbon dioxide mixtures to improve productivity. Carbon dioxide (CO2) may be used, on its own, in MAG and FCAW. With the exception of CO2, these gases are not defined as hazardous to health under the COSHH Regulations but they are asphyxiants. None of the gases can be seen and none has a smell - so their presence in hazardous concentrations is difficult to detect without prior knowledge or measuring equipment.

The main hazard arising from exposure to shielding gases is accumulation in confined spaces. Argon is heavier than air, so tends to collect in low areas such as pits. Inhaling a gas, such as pure argon, which contains no oxygen can cause loss of consciousness in seconds. Workers should not enter an atmosphere that contains less than 18% oxygen.

Carbon monoxide (CO) and CO2 may be generated in fluxed welding processes by the action of heat on flux materials such as carbonates and cellulose. In MAG welding they can both originate from CO2 in the shielding gas, CO2 undergoing reaction in the vicinity of the arc to form CO. Flame processes also generate CO and CO2. The relative amounts depend on whether the flame is oxidising or reducing, with CO present in higher concentrations when the flame is reducing.

CO is by far the more hazardous of the two gases. It can cause a reduction in the oxygen carrying capacity of the blood that can be fatal. In lower concentrations it causes headache and dizziness, nausea and weakness. CO2 acts mainly as an asphyxiant, as indicated above. CO has a WEL of 30ppm and CO2 is listed at 5000ppm (8 hour time-weighted average).

The amounts of CO and CO2 generated by arc welding processes are small and, generally, they do not present an exposure problem. The amounts of CO and CO2 generated by flame processes are also small, so the risk of over-exposure is usually low. In special cases, such as high velocity oxy-fuel gas cutting, where large quantities of gas are consumed in a short period of time, the risk of over-exposure to CO may be a problem.

Nitric oxide (NO) and nitrogen dioxide (NO2) are known collectively as nitrous gases (NOx). Welding generates only small amounts of nitrous gases so exposure to NOx does not present a problem. Exposure problems may arise during cutting activities, particularly if the cutting is hand-held, as this places the operator closer to the emissions. Hotter flames generate higher concentrations of nitrous gases, so using acetylene generates more nitrous gases than using propane or natural gas. Plasma cutting with air or nitrogen generates higher levels of nitrous gases than oxy-fuel gas cutting and there is considerable risk of over-exposure.

Free-burning flames generate the highest concentrations of NO and NO2, and the risk of over-exposure is also highest. Caution should be exercised during activities such as flame heating, flame straightening, flame brazing, flame spraying, etc - particularly as emissions from these processes are difficult to control. The flame should be extinguished when not in use.

NO is a severe eye, skin and mucous membrane irritant. NO2 is a highly toxic, irritating gas. Yet neither feature in the WEL published in EH40/2005. This is presumably because doubt exists of the soundness of the limits. However, Chemical Hazard Alert Notices were issued in April 2003 for both gases (CHAN 28 and 29) and we must assume that these are still valid until further limits are published. These recommend exposure limits of 1ppm (8-hour reference period) for each gas. After inhalation, NOx act more on the deeper rather than the upper respiratory tract. The following symptoms are an indication of the primary stage of poisoning:

  • Irritation of the eyes, nose and trachea
  • Intensive cough
  • Narrowness in breathing
  • Dizziness and headache
  • Sickness and fatigue

The symptoms of over-exposure may not be apparent for several hours after the cutting or welding activity has ceased. Severe over-exposure may lead to an accumulation of water in the lungs impairing oxygen supply to the blood and possible death.

Ozone can be generated by reaction between UV light from the arc and oxygen in the air. It has a low WEL of 0.2ppm for a 15-minute reference period.

At the levels of exposure to ozone found in welding the main concern is irritation of the upper airways, characterised by coughing and tightness in the chest, but uncontrolled exposure may lead to more severe effects, including lung damage.

MIG welding of aluminium alloys with an aluminium/silicon filler wire generates by far the highest concentrations of ozone. Other process/material combinations that may generate hygienically significant concentrations of ozone are MIG with Al or Al-Mg wires, MAG/mild steel, MAG/stainless steel and TIG/stainless steel.

Ozone is only generated during arcing and decays quickly on arc extinction. Therefore, exposure to ozone is very dependent on the duty cycle employed. Although research in the laboratory has shown that ozone concentrations at points around a welding arc can exceed 0.2ppm, it is uncommon to find that average exposure to ozone, in a real work situation, exceeds the ozone exposure limit. An exception to this statement is exposure to ozone during MIG welding with an aluminium/silicon consumable.

Where is the welder's nose?

No, not the obvious answer: we need to consider the relationship of the person's breathing zone to the concentration of fume and gas generated during the process. A protocol for measuring exposure has been specified in BS EN 10882-1. HSE Guidance Note EH 54 also offers a matrix of working conditions to consider when contemplating exposure measurements. Carter[2] reports on this matrix but notes the impracticability of testing with a supposed fixed position of the welder's head to the plume of fume.

The HSE guide also calls for definition of welding current and equates increased current with increased fume. Furthermore, it does not differentiate between high fume conditions caused by head position or by welding current. Carter rather convincingly argues that this is fallacious and claims that an HSE assessment is difficult to carry out and results in very conservative conclusions. This leaves the employer and the welder in something of a quandary when trying to decide exposure conditions.

Good practice is to carry out exposure monitoring but care must be taken that this is representative and taken in a standardised manner. For welding when a helmet is worn, sampling must be performed behind the helmet, to provide a true indication of the quantity of fume breathed. Measurements may also be carried out at pre-selected points in the workshop. These can be useful in estimating the amount of fume breathed by other workers in the area.

A very significant factor is the duty cycle, ie how much arcing time there is in a standard working day. Thus, a welder exposed to even 25 mg/m3 for only infrequent, short periods may still be well inside the 5 mg/m3 WEL for iron oxide fume (averaged over the shift).

Is there a solution?

Fortunately there is - isolate the welder's breathing zone from the hazardous area. There are two approaches to achieving this: removing the hazards from the vicinity of the welder's face, or supplying clean air directly for him/her to breathe.

Removing the hazards from in front of the welder can be achieved by not generating them in the first place or by extracting what is there. Clearly if fume is generated in sufficiently small quantities, the situation for the welder will be improved. Processes where metal is transferred across an open arc produce the most fume, e.g. MMA, FCAW, MIG/MAG. On the other hand, TIG welding, in which filler is added as a non-current carrying wire, does not produce particulate fume in appreciable amounts.

The fume generation rate of any one process, e.g. MMA, depends on the parameters but the effect is relatively small. Generally higher currents and voltages equate to more fume but it is not practical to attempt to control fume generation by welding parameters alone. Some manufacturers state that the generation rate of their consumables is below that typically expected. Such control of the chemistry and physics of the process is possible but again may be only of second order. Certainly the position of the welder's head and the space confinement of the operation can have larger effects.

Welding position is an important variable as it affects the proximity of the fume to the welder's breathing zone and has a major effect on exposure. Welding vertically-up usually results in the welder's head being away from the path of the fume plume. Positions that place the welder closer to, or worst of all, above the plume of fume lead to highest exposures, so leaning over a flat position weld is more hazardous.

If the welding operation is in a confined workspace accumulation of fume may be expected to increase exposure. Similarly if the duty cycle is high the concentration of fume in the vicinity, and the time that the welder is exposed to it, will increase.

Provision of local extraction to suck away the fume from the welder's breathing zone is an obvious remedy. It is, indeed, quite efficacious, but only when used correctly. It is most useful for fixed welding stations where repetitive jobs are carried out. Here, the extraction nozzles can be placed close to the weld and need little re-positioning. Even for applications where the welder has considerable movement, positioning of extraction nozzles will provide adequate protection if used correctly. There may be reluctance on the part of the welder - 'one task too many' - but this is not really acceptable.

Respiratory protective equipment, known as RPE, must be provided and used when welding fume cannot be adequately controlled using ventilation techniques. RPE should always be regarded as a last resort solution to an exposure problem and should be used in addition to, rather than instead of, other control measures.

RPE is broadly grouped into two classes; respirators and air supplied equipment. Respirator equipment includes disposable respirators, powered respirators and face masks with filters. They take in contaminated air and filter or clean it before it is inhaled. Air supplied equipment includes devices such as air-fed helmets and self-contained breathing apparatus. They deliver air from a separate source to the welder.

If used, RPE should be incorporated into a formal management system so that effective controls are put in place and monitoring is undertaken regularly.

The employer's obligation

Records from monitoring should be readily retrievable for inspection and should be in an easily understood form. The employer must keep the results of personal sampling for at least 40 years and for at least five years in all other cases.

Employees or their representatives must be informed of the results of assessments and of any monitoring carried out, particularly the monitoring of carcinogens or asthmagens (e.g. nickel and hexavalent chromium compounds in fume) which has been exceeded. Collective results of any health surveillance must also be provided, but in a form which preserves the anonymity of individuals.

Employees (and other persons as appropriate) must, as described in the General COSHH Approved Code of Practice, be kept well informed on:

  • The nature and degree of risks to health arising as a consequence of exposure; including factors that may influence that risk (such as the substances involved) and factors that may increase that risk (e.g. smoking).
  • Control measures adopted, reasons for these, and how to use them properly.
  • The reasons for personal protective equipment and clothing, and the jobs where this is necessary.
  • The monitoring procedures, the arrangements for access to the results, and methods of notification if a WEL is exceeded.
  • The role of health surveillance, employees' duty to attend, the health surveillance procedures, arrangements for access to individual records and the collective results of health surveillance.

Instruction must be given to ensure that employees know:

  • What they must do, the precautions that must be taken and when they must take them.
  • What cleaning, storage and disposal procedures are in place, why they are required and when they are to be carried out.
  • The procedures to be followed in an emergency.

Training must be provided for the effective application and use of:

  • Methods of control
  • Personal protective equipment
  • Emergency measures

To keep such records and to inform and train a workforce may seem onerous, but it is the law and it is necessary to plan and implement these things effectively. Do things correctly and welding is a safe operation. Ignore the precautions and it can be very costly both for your company and your welders.


The Health and Safety Executive website[3] gives guidance on a responsible approach to the workplace environment. In particular the pages on welding have a COSHH section[4] that has a number of downloadable documents covering such aspects as 'Advice for Managers' (WL0); 'Exposure Measurement: Air Sampling' (G409); UK Standard Assigned Protection factor 20 (R3, a document covering the selection of respiratory protective equipment); and a series on engineering control for each of the welding process, eg 'Metal Inert Gas (MIG) and metal active gas (MAG) Welding' (WL10).

Whilst these documents refer to COSHH 2002 rather than the newer WEL system, they give sound, practical advice on creating and maintaining a safe environment for welders.


  1. TWI website:
  2. Risk assessment and control of exposure during arc welding of steel; G. J. Carter; International Conference: Health and Safety in Welding and Allied Processes 9-11 May 2005, Copenhagen, Denmark
  4. COSHH essentials for welding, cutting and allied tasks:

For more information please email: