To exploit the advantages of laser technology for material processing, e.g. micro drilling and cutting, versatile trepanning systems based on rotating optics have been designed and implemented. Depending on the development stage, the trepanning systems enable the controlled adjustment of beam displacement and inclination during fast rotation.

The typical specifications of the new versatile trepanning systems, developed by the LMTB and redesigned by 3D-Micormac, are the following:

• Free choice of laser wavelength (from the UV to the NIR, 10 µm is a future option)
• Free choice of laser pulse width (from sub-ps to ms, or even cw)
• Free choice of focusing optics with a focal length  ³ 60 mm  (e.g. doublet lens, gradient lenses, etc.)
• Drilling / cutting diameter: 50 to 1500 µm (larger diameters are optionally addressable)
• Laser beam inclination angle: -5..0°..+5° (0° = cylindrical, i.e. taper-less)
• Rotation speed: set at 18,000 rpm and 10,000 rpm
• Protection window disks and processing gas nozzle are included in all types
• Low weight and compact size for straightforward work-station integration: from <4 kg, <(200 x 120 x 120) mm³ to  <7 kg, <(200 x 250 x 120) mm³ , depending on the type (manual or motorized).

The trepanning systems are customized for very different laser parameters, e.g. for ultra-short laser pulses at wavelengths of 355, 515, 532, 1030, 1064 and 1550 nm.  The implementation of the newly developed trepanning systems can be identified already in several R&D laser laboratories and in 24/7 productions facilities around the world.

Fig. 1a presents an example of nanosecond laser drilling of aluminum nitride (AlN) ceramic. The laser system utilized for this application is an industrial q-switched, diode-pumped Nd:YVO₄ laser system with an amplifier unit manufactured by IB-Laser, Berlin, generating ca. 30 ns laser pulses at a maximum average power of 38 W @ 20 kHz at a wavelength of 1064 nm. Fig. 1b depicts a typical SEM image of the cylindrical bore wall at cross section. Note the smooth processed side-wall surface, demonstrating the exceptional processing quality using standard nanosecond laser pulse technology. The angular beam steering and circular pulse distribution in laser trepanning generally yields far improved processing conditions compared to conventional laser percussion with a fixed focusing optics alone.

The primary processing results with the novel trepanning systems utilizing nanosecond laser pulse technology are very encouraging. Further studies with different materials are conducted at the LMTB laser application laboratory to optimize the laser trepanning system for even more precise and reliable micro drilling. Present investigations include applications using picosecond laser pulses and other materials, such as metals.

Fraunhofer Institute for Laser Technology ILT, Aachen, Germany

In recent years the use of modern solid-state lasers has brought about a distinct increase in operational speed in laser materials processing. Whether with scanners or fixed optics, high speeds – as far as possible in various axes at the same time – have almost become the norm. But although the movement of the optic is precisely calculated, the position of the processing point can deviate from the planned contour. Help is at hand, thanks to a process monitoring system which precisely tracks the relative movement of workpiece and optic. It enables acceleration-related deviations from the set contour and speed to be measured exactly and the numerical control system to be adjusted accordingly.

Research scientists at the Fraunhofer ILT in Aachen have developed a camera-based system which analyzes the movements of the workpiece through the optical axis of the laser beam before or during processing. It does not matter whether a fixed or scanner optic is used – in both cases the system measures the movement of the processing point on the workpiece and documents deviations from the set contour during machine setup or operation.

The process monitoring system uses image sequence frequencies of up to 10 kHz. In various applications, contours have been measured with a processing speed of up to 10 m/min (fixed optic) and up to 15 m/s (scanner optic). The deviation from a reference system was less than 3 cm/min. At present the measured data are evaluated separately. Whilst the same technology does permit real-time measurement (there are no technical barriers to this), the accuracy class of this has not yet been completely specified.

The special design of the system means that it can be used in a very wide range of applications, including laser cutting and welding, soldering, drilling, ablation, microjoining, SLM and hardening. The various modes of operation are interesting both for system integrators and for end users. On the one hand, the system can track the processing point during machine setup, enabling the planned contour to be adjusted.

On the other hand, the system permits process control during actual operation. This means not only can the processing contour be adjusted, the laser output can also be controlled to ensure an even energy input at different laser spot speeds. That is a critical factor in particular when processing thin materials. As a result, existing processes can be optimized and new processes are made possible.

Bild1:  Measured feed rate and trajectory (Welding)

Bild2: Work piece surface and calculated velocity vectors

Fraunhofer IWS

Carbon fiber reinforced polymers (CFRP) are increasingly applied in the aircraft industry as well as the automobile industry. The main reason is the highly mechanical load on one hand and the low density on the other hand. Moreover, the corrosion resistance plus the damping behavior of the material can fully be utilized in highly stressed structures. However, the concept of manufacturing CFRP-parts near- net- shape does not substitute the need of cutting them. The different properties of fiber- and matrix-material constitute an ambitious challenge for the CFRP cutting process with a laser beam. Main influencing parameters are the different thermo-optical characteristics of fiber and matrix. The carbon fibers have a more than 20 times higher sublimation temperature than the decomposition temperature of the resin. Because of a good heat conduction of the fibers, the resin fitting at the fibers will quickly be decomposed, before the carbon fiber itself is cut (Fig.1). The theory to decrease the heat affected zone is to minimize the interaction time between laser radiation and material.

The benefit of the availability of high power and high brilliant laser sources enables ablation and/or cutting processes with high processing speeds. However, the relation between wavelength and absorption on non-isotropic materials on the one hand, the influence of the intensity and the processing speed on the other hand, require fundamental research.

Within the experiments a continuous-wave (cw)-single mode fiber laser was applied in comparison with a Q-switched pulsed laser system (50 W, 100 ns). Using the Remote technology (the laser beam will be deflected by two galvo – driven tilting mirrors) a minimized heat interaction time between material and laser beam can be reached. The setup of the cw-Remote processing is shown in Fig. 2.

The material to be processed was a 3 mm thick consolidated CFRP with epoxy resin. Its fiber volume percentage is about 50% with multidirectional orientation of the long fibers.

As a result two cross sections can be seen:

Fig.3

Cross section of CFRP using cw – laser & Remote processing

Pl = 3 kW, v_average = 9 m/min, E_T = 11 J/mm²

Fig.4

Cross section of CFRP using Q-switched – laser & Remote processing

Pl = 50 W, v_average = 0,04 m/min, E_T = 28 J/mm²

The Remote technology proved to be an excellent tool to cut consolidated CFRP with good cut qualities. Average processing speeds of more than 10 m/min can be reached using high brilliant and high power cw – lasers. The heat affected zone can be minimized by decreasing the interaction time. One possible way is to use Q-switched lasers with short pulse lengths. However, because of the limited overall laser energy, the average processing speed is quite low.

Fig. 1: REM image of CFRP, gas assisted laser cutting

Fig. 2: Experimental setup for Remote processing

Ultra-slim flexible glass substrates have many potential applications, spanning from photovoltaics to e-paper to touch sensors. Previously, these applications generally incorporated glass substrates in the thickness range of 0.3-1.0 mm and benefited from inherent glass properties including high optical transmission, low surface roughness, high thermal and dimensional stability, and low CTE. Recently, however, there is interest in reducing the thickness of the substrate to ≤200 mm. Glass substrates at this thickness still provide the inherent beneficial properties of glass, but they also enable substrate flexibility and end product devices that are thinner and lighter weight.

CO₂ laser cutting of flexible glass substrates possesses several advantages over mechanical and other laser-based cutting methods. It is non-contact, and it uses tensile stress to propagate a full-body crack along the direction of cutting.  The tensile stress is generated with CO₂ laser heating and a subsequent active or passive cooling process. A typical setup using CO₂ laser cutting technique is shown below.

We investigated CO₂ laser cutting of 100 mm and 200 mm flexible glass substrates suitable for display applications. The speed and quality of cutting 100 mm and 200 mm substrates were evaluated.  We observed that:

1. There is minimal dependence on the laser beam length for the laser cutting speed. A shorter laser beam is preferred since the cutting speed was observed to drop faster for longer laser beams as the jet-to-laser beam center distance increases.
2. The laser cutting speed dependence on width of the laser beam was minimal in all cases for 100 mm and 200 mm flexible glass. The cut edge straightness improves with a decreasing laser beam width. However, process stability generally decreases with decreasing beam width over large distances.
3. Cutting can be carried out without active cooling (water mist jet). In this case a rapid temperature drop due to air convection generates the tensile stress necessary for the propagation of the full-body crack. The cutting speed in this case was significantly lower than what can be achieved with mist jet cooling.
4. The cutting speed was affected by thermal contact between the glass sheet and the metal chuck. Adding a thermal insulation barrier such as polypropylene or polytetrafluoroethylene increased the cutting speed by as much as an additional 40 mm/s.

In summary, we investigated CO₂ laser cutting of flexible glass substrates using active cooling jets. We found that the laser cutting speed varies little with laser beam length and width. We attribute the weak dependence to a rapid thermal homogenizing process in flexible glass. We discussed the optimal laser cutting process and recommend that the laser cutting speed should be operated in a region which is defined by a lower boundary and an upper boundary. The lower boundary is defined by cutting action without active cooling. The upper boundary is defined by the maximum cutting speed. The distance of the water jet should be kept close to the laser beam to avoid reduction in speed due to thermal loss to the environment.



Today’s high speed production drills in PCB board manufacture operate at speeds in excess of 250,000rpm. The spindle shafts in these devices need to be individually balanced by very selective removal of material to ensure that, in operation, they run true. Any eccentricity in rotation can cause drill bit wobble, resulting in breakage and down time. Current manufacturing methods are manual and time consuming, requiring skilled labour.

Using DLS proprietary software in conjunction with an SPI 40W pulsed fiber laser, an automated laser balancing system has been created. The system identifies the amount of material for removal and then accurately removes through a laser milling process while the spindle is still rotating.

The spindles are loaded into a rotating jig that analyses the rotation and separates the complex motion into its static and couple components. The software then automatically calculates the amount and location of material that needs removal. The pulsed laser is then used in-conjunction with a scanner to locally machine the spindle while it’s still rotating. The material is removed from a small arc on the circumference of the spindle (photo 1).

The process has had a dramatic effect on throughput and efficiency, resulting in a 80% reduction in balancing time and a 20% increase in yield. Quality is also significantly improved, in that balance accuracy is down to 100’s of micro grams and shaft runout (TIR) can be reduced to less than 500nm.

Purdue University

Replacement costs for high value engineering components are very high due to long lead times and special tool required for processing high strength materials.  Laser direct deposition provides an attractive and cost effective means for repairing or remanufacturing high value engineering components.  Traditional repair processes are limited in applicability and bond strength.  With the emergence of Laser Direct Deposition (LDD) technologies the limitations of traditional repair processes are overcome.

Remanufacturing utilizes existing material resources to restore a component to “as new” condition or better.  Remanufacturing can produce a part better than the original by incorporating more advanced materials or by adapting to the improved design.  It also reduces waste and extends the life of the component. Laser-based remanufacturing provides a means to restore parts that were previously deemed “non-repairable”.

One difficulty in laser based remanufacturing is how to determine the missing geometry because of a void or defect.  A complete reverse engineering process was used to generate a parameterized geometric model required for LDD based repair.  The process starts by taking a laser scan of the defective are and building a mesh.  Then extract Prominent Cross Sections (PCS) from the non-defective region and reconstruct the repaired model in CATIA™ by interpolating the damaged region.  Finally, extract the Boolean difference to obtain the repair volume. The repaired blade matched the geometry of the original blade with a maximum deviation of 0.145 mm from the nominal model.

This study demonstrates the successful repair of defective voids in turbine airfoils based on a new semi-automated geometric algorithm and a direct laser deposition process. A Boolean difference between the original defective model and the final reconstructed model yields a parameterized geometric representation of the repair volume. The experimental results of this method demonstrate the effectiveness of laser direct deposition in remanufacturing and its ability to adapt to a wide range of part defects.

A Life Cycle Assessment (LCA) on the energy and environmental impacts showed that LDD only required 7.6% of Carbon footprint over replacing with a new part.

Re manufacturing saves energy and is better for the environment than producing a new investment casted component.  In addition, LDD remanufacturing can accurately restore the shape and strength of high value engineering components.

Ceramic particulate reinforcement enhances the already superior mechanical properties of nickel-based super alloys.   Reinforcement particles alter the matrix microstructure of the metal matrix composites (MMC) to improve wear resistance, hardness, etc.   Laser direct deposition allows localized addition of MMC with properties tailored to the application.   This research focuses on the effects of particle concentration on microstructural changes and mechanical property improvement.

In the present work, laser direct deposition was used to build multi-layer structures of nickel-based alloy Inconel 690 reinforced with titanium carbide particles at 18.4%, 27.2%, and 40.0% by volume.  Optical and electron microscopy of sectioned specimens revealed a uniform distribution of titanium carbide particles throughout the multi-layer depositions and a near absence of internal voids or incomplete melting.  The ratios of Inconel 690 to TiC in the depositions matched the levels in the initial powder mixtures.  The presence of carbide particles coincided with refinement of matrix microstructure and introduction of a finely dispersed crystalline phase when the titanium carbide percentage was increased to 40%.  High-temperature dissolution of titanium carbide was not detected.   The hardness test revealed that the hardness increased with increasing percentage of TiC particulates from 20 in the base metal to 40 HRC in 40% TiC reinforced metal matrix composite.  Wear tests results using pin on a disk apparatus also showed that wear scars were generally elliptical, nearly rectangular, and higher carbide content caused more wear of the steel ball used, thus indicating higher hardness of the surface.

Predictive numerical modeling results were also shown.   The multiphysics numerical model assisted in designing a suitable operating condition of the laser direct deposition and also predicted the even dispersion of the particulates during the laser direct deposition of metal matrix composites.

Figure:  SEM micrograph of the deposited Inconel690/TiC40% metal matrix composite, exhibiting a uniform distribution of the particulates.

Laser marking has penetrated a wide range of industries and is today used for traceability and product identification in markets as diverse as automotive, micro-electronics, semiconductor, photovoltaic, medical and packaging industries.

However, the pharmaceutical industry faces specific challenges which have until now limited the use of laser technology for anticounterfeiting purposes:

• The identification mark should be as close to the primary package as possible, e.g. the syringe vs. the carboard box it comes in, or the vial vs. the palette.
• The marking process should not alter the product.
• The mark should be tamper-proof and easy to read.
• The process speed should be compatible with the production environment.

Alternate technological solutions include inkjet printing and radio-frequency identification (RFID). Inkjet printing is widely used, still is vulnerable to tampering, adds additional consumables and leads to additional production steps. RFID, although well suited for remote checking of a batch, is not compatible with the primary package. Surface engraving by laser is also attractive, but may lead to physical alteration of the container and is also vulnerable to tampering by polishing the code.

Laser internal engraving is therefore today the only technological solution which fulfills all the criteria of the pharmaceutical industry. However, this approach brings serious constraints to the choice of the laser source. The laser intensity at the focus should be sufficient for marking, ruling out continuous wave lasers. The container should be transparent to the laser emission, ruling out UV lasers. Finally, the container integrity should be preserved, which prevents the use of nanosecond infrared lasers. Due to the thermal nature of the laser-matter interaction, micro-cracks are created, which can propagate over time and ultimately lead to glass fracture.

The recent development of industrial-class ultrafast lasers enable internal marking without any damage to the container. Because of the extremely high optical intensity and very short pulse duration delivered by ultrafast lasers, there is no heat dissipation during the interaction process, which means that there is no micro-crack formation. Since the individual spots can be made very small, it is also possible to achieve virtually invisible marking, and yet to guarantee reliable reading under proper lighting.

A first generation of internal engraving systems using diode-pumped Ytterbium ultrafast lasers was developed in 2004 by an industrial consortium, with the support of the European Union. In 2007, a company called TrackInside was formed to exploit the commercial potential of the technology in industrial environments Because of the moderate average power of the laser used, this first generation system was first implemented in the luxury industry, which has the same technical requirements as the pharmaceutical industry, but needs lower production speed.

The development of high power industrial ultrafast fiber lasers enabled the development of a second generation equipment, dedicated to the pharmaceutical industry. While maintaining the quality of the ultrafast laser internal engraving technology, the average power of the laser enables a marking speed compatible with the requirements of pharmaceutical industries. Additional developments were conducted related to the scanning speed, the software control and mechanical handling, to guarantee high quality marking at speeds up to 15 containers per second.

The TrackInside system has a very high accuracy and its flexible engraving process can create a 500 x 500 μm data matrix in less than 40 mm, as well as logos and text, on a field up to 60 x 60 mm. The readability of the marking is rated as grade A-AIM, i.e. the most stringent international regulation.

Ultrafast lasers are today at the nexus of several key technological advances. Strong innovation in the scientific community gives rise to an increasing number of industrial application. At the same time, as the average power of ultrafast lasers increases, more of these applications reaches the level where industrial deployment is economically competitive, further increasing the drive for further laser developments. Let us expect ultrafast lasers to play an ever increasing role in industrial processes, touching on many aspects of our daily life.

Ophir-Spiricon, LLC

It’s time to buy another car. Like everyone else these days,you’re a bit cost-conscience, so you’re looking at getting the most for your money. You decide to see what the local used car lothas to offer. Before you get out of your car, the slick used car salesman approaches you, shoves a card into your hand and is a little too happy to help you find a new automobile. You approach a couple of cars that you think might fit your budget. The closer you get, you notice something that seems odd. You’re standing in front of two identical cars – same make, same model, same year, same color, even the same warranty. But you see that one is $5,000 less than the other. Hmm. The first, more expensive car seems to be ingood order, looks nice, smells okay. You climb into the second car and it hits you why there is a difference in price – the second car has no dashboard instrumentation panel! No speedometer, no fuelgage, no warning lights.

Mr. Slick sees the concern on your face and quickly tries to thwart off any objections, “Don’t worry about not having a speedometer, you can just go along with traffic. Not a problem with not having a fuel gauge, it starts making a knocking noise when you need to fill up. And these are great cars, you shouldn’t worry about not having any warning lights.” Oddly enough, this all makes sense to you. You make the decision that these kinds of indications really aren’t that important and decide on the less-expensive car.

Of course you don’t decide on the car that doesn’t give you any indication on how it’s performing! That doesn’t make any sense to you at all! Yet, there are many laser operators and laser jocks out there that do just that when it comes to their lasers. They are content to be blissful in the ignorance of their own laser’s performance. They rely on simple sometimes time-consuming methods of beam diagnostics that are oftentimes subjective to the technician, don’t comply with industry standards and rely on a skill set that is difficult and time consuming to transfer to other employees.

Or, even worse, some laser technicians employ a “don’t fix what isn’t broken” approach to their laser performance monitoring, which can and have resulted in poor process, scrapped or failed parts, or even worse, product recalls. Several industry experts, from laser manufacturing engineers, research scientists, process engineers, to name just a few, agree that the best way to accomplish a comprehensive laser maintenance program is with a planned maintenance schedule that includes beam power/energy measurement along with a process called beam profiling.

About Beam Profiling
Laser beam profiling can be described as using an imaging device, such as a camera or a scanning slit profiler, to capture and display the spatial intensity of a laser’s energy. The software that is interfaced with this imaging device will then perform attribute measurements such as beam size, beam wandering, peak energy to centroid (or the geometrical center of the beam) location, as well as other beam characteristics and even using the latest insoftware developments to incorporate an average power or energy per pulse measurement to calibrate these measurements. There are even devices that will give you all of this information in one package. Bottom line is that it can be as simple or as complex as you want it to be, but the benefits of implementing laser beam profiling practices can be very beneficial at the end-user stage of a laser’s life.

Laser power/energy measurement is a quick and basic, yet certainly a vital practice to monitoring your laser’s performance and efficiency. Laser engineers will use such equipment for measuring the laser’s average power or energy per pulse over time to predict and plan for flash lamp changes or optics alignments or replacements. The equipment can also be used to characterize a laser for purposes of process validation or run off of new equipment. However, as important as this information is, it does not tell you all you need to know about your laser. For instance,the laser’s average power could be stable while the laser’s mode could be unstable or not optimized, ultimately causing undesired effects during your process.

Depending on the process that the laser is intended for, the laser’s mode, or structure of energy distribution, should be optimized for that process. For instance, a Gaussian, or “cone-shaped” mode, with a relatively high peak power near the centroid of the beam, should be achieved for marking, etching, microwelding and some cutting applications. A flat-top mode should be achieved for most welding applications. And a TEM01 or TEM01* (or “donut”) mode is common for high-powered cutting lasers. Several different factors come into play when trying to optimize your laser’s mode; process parameters, optical alignments, condition of laser components are just a few.

Driving an automobile without an instrument panel and not monitoring your laser’s performance on a regular basis can oftentimes have the same results. You can wait for an expensive and time-consuming laser failure the way you could use a policeman to provide you with information about your speed.You wouldn’t ever operate a motor vehicle without knowing how it’s performing because it’s just not safe to. With consequences being far more serious, why would you operate your laser without knowing how it’s performing?

Health Protection Agency, UK

The development of technology has resulted in fast penetration of LEDs into a wide range of consumer products, including toys. The significant increase of optical output and expansion of the emission wavelength range, from ultraviolet to infrared, raised a concern about optical radiation safety of LED use in toys and led to the development of a methodology to assess the safety of the LED. The simplified approach is based on LED photometric characteristics from datasheets and avoids the need for additional complex measurements.

The emission of an LED in a toy should not be greater than required for the intended purpose. The level of a child’s exposure to optical radiation from an LED during normal use and under foreseeable misuse should not exceed Exposure Limit values for the eyes and the skin recommended by the ICNIRP [1-3]. The Accessible Emission Limit (AEL) values in this document establish a correlation between personal exposures under worst case exposure scenarios (distance, duration of the exposure) and the emission of the LEDs.  The AEL values are maximum level of LED emission not expected to result in overexposure to optical radiation under worst case scenarios. Two optical radiation hazards are considered to derive the AELs: UVA hazard to the eyes and Blue Light photochemical hazard to the retina of the eye, with the more restrictive AELs applied. Another consideration is that UV transmittance of the crystalline lens is much higher in infants under the age of 2 than in older children and the Blue Light spectral weighting for the infant eye is significantly higher up to 440 nm [1].

In the UV spectral range, the risk of adverse health effects to the eyes and the skin increases significantly at shorter wavelengths, especially below 315 nm. At the same time, the practical usefulness of short wavelength LEDs in toys to create visual effects is questionable. Therefore, on a risk/use balance, the use of LEDs emitting below 315 nm isn’t justified for children’s toys.

UVA-emitting LEDs are used in toys as excitation sources for luminescent materials, for example, “secret writing.” The UVA AEL includes only the LED emission within the spectral band of 315-400 nm. Because of the narrow-band emission and very low visual stimulus, the aversion response to the exposure from a UVA LED may be compromised and exposure in close proximity should be considered as foreseeable. A child’s curiosity triggered by fluorescence of the eye lens could increase the risk of such an exposure.

The AELs of LEDs emitting in the visible spectral range (300-700 nm) are defined by the more restrictive of the two exposure scenarios, UVA (as discussed above) or Blue Light photochemical hazard. For visible light,  AEL values are given in watts, which are spectrally weighted with the Blue Light hazard function, whereas in datasheets the LED output is often expressed in derivatives of un-weighted watts or in photometric units, such as candela or lumen [4].

Due to the different wavelength dependence of Blue Light and luminous efficiency weighting, the visible light AEL is wavelength and emission-bandwidth dependent if expressed in candela. The visible light AEL should also take into account the ICNIRP recommendation for a luminance limit of 10⁴ cd/m² and depends on the LED emission angle.

For LEDs with a peak emission wavelength below ~500 nm, the visible light AELs defined by the Blue Light hazard are more restrictive than the ICNIRP luminance requirements; above ~550 nm the luminance limit is more restrictive even for broad-band LEDs. The spectral bandwidth of the LED emission has a smaller effect on the visible light AELs than the peak emission wavelength.

More than 280 LEDs intended for incorporation into toys and a range of toys were measured. The majority of the tested LEDs could be considered as eye-safe for use in toys (see Fig.1). Measurements showed good agreement with the predictions of the simplified safety screening technique.

• Miniature red, orange and yellow LEDs could be considered eye-safe for use in toys, as a single component or in arrays;
• Low power green LEDs are safe for use in toys; high power green LEDs may require more detailed analysis, especially if used in arrays;
• The Blue Light hazard level of some of the blue LEDs is above ICNIRP ELVs, even for low power single components;
• Low power white LEDs could be considered as eye-safe; the Blue Light hazard level of the high power white LEDs may exceed the ICNIRP ELVs;
• Low power UVA LEDs are eye-safe for use in toys for children 3 years of age and older

Fig.1. Comparison of datasheet information with visible light AEL

References

[1] ICNIRP Guidelines on Limits of Exposure to Ultraviolet Radiation of Wavelengths between 180 nm and 400 nm (Incoherent Optical Radiation). Health Physics, 87 (2): 171-186; 2004.

[2] ICNIRP Guidelines on Limits of Exposure to Broad-band Incoherent Optical Radiation (0.38 to 3 mm). Health Physics, 73 (3): 539-554; 1997.

[3] ICNIRP Statement on light-emitting diodes (LEDs) and laser diodes: implication for hazard assessment. Health Physics, 78 (6), 744-752, 2000.

[4] CIE S 010/E: 2004 Photometry – The CIE system of physical photometry.