Of course, the vast majority of LEDs are completely safe and do not represent any hazard to the human eye. However, many of us have been dazzled by LEDs at some point, possibly by staring at the brake lights of the car in front, or by looking too closely at that LED key ring picked up as a free gift from a conference.

At a recent LEDs workshop in the UK, Andrew Dennington of Carclo Technical Plastics, presenting a series of optical design tips, also included a word of caution; “The latest generation of LEDs is not safe, and someone will have their eyes damaged by a high-power LED product,” he warned. “Check your products to the relevant standards.”

One ongoing issue with LED safety is the problem of whether to classify an LED as a laser or a lamp – both have merits and both present problems, depending on how the LEDs are arranged and used.

Laser safety standard

So what are these standards currently in place that cover LEDs? The most important is IEC 60825-1, published by the International Electrotechnical Commission (IEC), which has been adopted in Europe as EN 60825-1. This standard is commonly known as the Laser Safety Standard but crucially also covers LEDs – at least for now – and treats them as lasers. Until quite recently, very few people realized there was a legal requirement (see footnote 2) for LED products sold in the EU to be tested according to 60825-1, but awareness has grown in line with LED performance.

The situation is different in the US, where laser products are covered by Federal Law 21CFR 1040.10, and compliance with 60825-1 is acceptable - but LED products are excluded. This is because the US, rightly or wrongly, consider all LEDs to be safe!

60825-1 sets accessible emission limits (AELs) for classes, which are used to give an indication of the risk from the beam of the laser or LED. Class 1 is by definition safe, and the majority of LED products fall either into this class or are exempt. LED and laser products are exempt from 60825-1 if the emission level does not exceed the AEL of Class 1 under all conditions of operation, maintenance, service and failure.

The most recent version of the 60825-1 standard indicates that LED manufacturers should provide more detailed data to help their customers decide if the end product is exempt or not. "Most LED manufacturers are not prepared, or not able, to provide the necessary information," says John O'Hagan, who leads the Laser and Optical Radiation Dosimetry Group at the UK's Health Protection Agency.

The 60825-1 standard refers to the product itself, rather than any packaged LED or module within the product. Robert Wells, a test engineer at Lasermet, a specialist test house, says that the addition of focusing or collimating optics can significantly alter the hazard posed by an LED and potentially can place the end product in a different class from the LED itself.

Products can be self-declared - i.e. the testing is performed in-house - or the appropriate measurements can be carried out by companies such as Lasermet or Laser Optical Engineering, another UK test house offering specialist expertise in LED testing. Although significant errors can be made by inexperienced testers, which could lead to a product being declared class 1 when it’s actually very dangerous, there is no legal compulsion to have a product tested by an external body, says Robert Wells.

It remains a bone of contention that LEDs are treated in the same way as coherent laser sources by IEC 60825-1, but this might not remain the case for long. John O'Hagan, who is part of the working group involved with redrafting 60825-1, says that it is "highly likely" that LEDs will be taken out of the laser safety standard. "However, it is the UK's position that this should not happen until some other documentation is in place to cover LEDs," says O'Hagan. "We feel that the properties of some LEDs are so close to those of lasers that they could be dangerous."

Treating LEDs as lamps

The International Commission on Illumination (CIE) treats LEDs as lamps, which are extended (rather than point) sources that generally have widely divergent beams. The CIE and IEC approaches (lamp or laser) generally don't produce the same results for the same product. Within CIE's division 6 (Photobiology and Photochemistry) is technical committee TC6-55, which is looking at the different methods of assessing photobiological safety of LEDs.

CIE has a Lamp Safety Standard, CIE S009/E:2002, which recommends that manufacturers place their LED products in one of several categories: exempt, low-risk and medium-risk. However, this has no legal basis in the EU. "CIE will typically produce lots of rationale, followed by a standard," says O'Hagan, "which might then be taken up by a body such as IEC and ISO, and eventually find its way into the EU legal framework." However, this hasn't happened in the case of CIE S009/E:2002.

Another wrinkle is that a Physical Agents Directive covering optical radiation is likely to be introduced in Europe in the next few years, but O'Hagan does not think this adds anything beyond existing general safety legislation.

What needs to be measured?

Clearly the optical output power of an LED source is important, but it’s also necessary to measure the apparent source size, rather than the beam diameter. “What’s important is the size of the image formed on the retina, since a larger image relates to a lower power density,” explains Robert Wells. “We have to measure the size of the apparent source, and work out exactly where it is.” This can be done in various ways including modeling the optical system, or taking measurements using a CCD imager.

The classification of the light source is based on the most hazardous condition that is reasonably foreseeable, even though this could represent an unlikely situation. “For most LED products, looking at them at close range through a magnifying glass is probably the most dangerous scenario,” says Wells.

However, this is not always the case, and in some circumstances the most hazardous viewing position may be some distance from the source. John O'Hagan cites the example of traffic-light LED arrays. When viewed close up, the radiation from only one LED is imaged in the eye, whereas from further away the radiation from all of the LEDs is imaged.

Footnote 1 - What damage can LEDs cause?
Lasers and LEDs cause a thermal heating effect in proportion to the power density of the radiation, which can result in tissue damage to the retina. Shorter wavelength radiation causes a photochemical effect in the retina, changing the chemistry of the cells, and there are dual limits (thermal and photochemical) in IEC 60825-1 between 400 and 600 nm. At shorter wavelengths below 400 nm, UV light is largely absorbed by, and can cause damage to, the cornea and/or the lens.

Footnote 2 – Legal requirements
In Europe, the Low Voltage Directive requires products operating between 50 and 1000 V to conform to the 60825-1 standard. However, there are moves to extend this down to 0 V, which is sensible since, as Lasermet’s Robert Wells points out, “You can get an awful lot of optical power [from laser or LED sources] at 48V.”

There is also a General Product Safety Directive which (you guessed it) requires that all consumer products should be safe; it doesn't reference 60825-1 but it is reasonable to use this standard if it helps define "safe". Furthermore, in different EU member countries there are various guidelines (rather than directives), for example that no laser pointer can be above Class 2, and that all LEDs in children’s toys must be Class 1.