In a single generation, we have come to take for granted how the transistor and its offspring, the integrated circuit, have transformed our lives. From our laptop to our iPod, from X-Box to TIVO, solid-state devices have created new markets and liberated old ones. Vacuum tubes and CRTs have become electronic fossils in only 25 years. These tiny, solid-state devices seem to last forever, use virtually no energy and can be relied upon to work under even the most taxing conditions.

Now, solid-state UV emitters, such as UV LEDs, UV laser diodes or Semiconductor Light Matrix (SLM) technology promise to alter the discussion of how UV materials are cured in a manner similar to the way the microwave oven has transformed the way we talk about cooking. Just as a bowl of popcorn may look and taste identical when cooked either way, so will UV coatings look and perform identically whether cured with conventional medium-pressure mercury lamps or with solid-state devices. This will also require a new language when we describe the UV process, since it is just as inappropriate to describe cure in terms of watts/inch as it is to ask at what temperature the microwave oven needs to be set for popping.

A Photon is Just a Photon

UV curing reactions are driven by "light" that can be described using several parameters that affect the reaction: wavelength, peak irradiance and "dose" (or energy density) related to the time of exposure. In purely photochemical terms these are the only variables regarding the UV source that matter. The manner in which the light is produced is irrelevant to the chemistry. For example, exposing a coating to 365 nm UV light at 200 mW/cm2 peak irradiance for 30 seconds produces the same effect whether the UV is produced by an arc lamp, microwave lamp or an array of solid-state light sources. A photon is a photon, and from the perspective of the photo chemistry the source is a "black box".

This is not to say that factors such as heat, stability of the light over time and other factors related to photon production do not affect the process - they do. The costs to purchase and operate a UV source are also important - sometimes critical to a project's viability. But these factors are separate from the interaction of light with the photoinitiator. The differences between solid-state and traditional lamp-based UV sources are described below.

Solid-State Versus Traditional UV Sources Spectral Output

The most striking difference between solid-state UV light sources and their mercury lamp brethren is the narrow spectral output that characterizes the solid-state devices. Figure 1 shows the actual emission spectra of a commercial solid-state UV source (Phoseon RX 10) measured using a spectral radiometer at an overall intensity of 1.5 W/cm2. Most industrial solid-state UV sources operate in the 380-420 nm region because the emitters offer the best cost, power and curing performance in this region. Additionally, the 380-420 nm region offers advantages as that wavelength is especially safe and above the interference of many popular additives.

For reference, the output spectrum of a medium-pressure mercury lamp is shown in Figure 2. This well-recognized mercury, or "H" lamp footprint, produced by all medium-pressure lamps, contains several well-known peaks corresponding to the mercury spectrum. These UV bands occur at 253 nm, 312 nm and 365 nm (the so called mercury "i-line"). The radiant emission in other peaks in the visible and the infrared region (not shown) detract from the curing ability of the source, for example, heat-sensitive materials cannot be cured with mercury lamps unless the lamps have additional reflectors and filters installed.

Additives are often used to "dope" mercury lamps to produce spectral shifts at other wavelengths. The output of doped lamps depends on the specific additive used. The most common additives are iron (the "D" lamp) and gallium ("V" lamps), though others also exist.

A Language Barrier

At the top of the FAQ (frequently asked question) list for UV lamp makers is the popular question: "How much power does it have?" Our pre-occupation with a simple measure to describe lamp power is like asking about the horsepower of a car engine. The answer can lead to problems. For example, a Shelby Series 1 and a Mitsubishi 3000GT both have identical horsepower (320 hp) engines. But the Shelby accelerates from 0-60 mph in just 4.4 seconds while the Mitsubishi takes 5.8 seconds. Is horsepower a good measure of going fast? If it were, a locomotive would beat a Ferrari hands down on the drag strip.

The problem of answering the "power" question is exaggerated by the fact that the tools used for this purpose (typically UV radiometers) have been developed to measure conventional mercury lamp sources.

Measuring the energy produced by light-emitting semiconductor devices with radiometers that were designed to work well with mercury lamps is a bit like listening to a dog whistle with human ears. We just are not hearing what they are - so it appears there is no sound. But a dog whistle is loud and clear to a pair of canine ears. To speak meaningfully about the light output of semiconductor devices, we must create new tools and a new dialog that relate to narrow bandwidth energy.

Output Stability Over Time

Solid-state UV emitters share the same characteristic we experience with other common semiconductor devices. That is, they seem to last forever.

Quantifying the lifetime of a UV source is a bit of a moving target. Conventional sources are typically specified to last from 1,000 to 8,000 production hours (depending on a wide range of factors). But usually the lamp energy is diminishing over time so that while the lamp is still operating it is producing less UV output than at startup. This degradation is affected by the number of on/off cycles, power, doping, ballast design, etc.

Solid-state UV devices are not affected by these variables. The lifetime of such devices appears unaffected (in fact studies show that the lifetime may actually be enhanced) by turning the source on and off. Figure 3 shows the actual measured output data for a solid-state UV source running for over 14,000 hours without any noticeable sign of degradation as compared to the typical output of an arc lamp.

While tiny individual emitters within the solid-state UV source array may degrade or fail at different points in time, with proper design and construction the overall effect is negligible at the work surface. Solid-state devices appear to operate at, or near, their original output for their entire operating life, eliminating the need for frequent radiometric measurements or compensation measures.

The Optimal Form Factor: SLM

Most solid-state UV devices (and products) available today are based on individually packaged light-emitting diodes (LEDs). Each LED is typically composed of a single, minute silicon "chip," mounted in a discrete package that usually includes a lens and some metal leads for connectivity. This physical format has some significant limitations, among which is the difficulty of concentrating enough UV light in a small space to be effective in applications such as curing UV coatings.

A much more useful way of integrating the semiconductor light sources is via "macro packaging" technology. One such technology is Semiconductor Light Matrix (SLM), which involves mounting hundreds or thousands of "chips" in very close proximity on a substrate, with micro optics to collect and direct the UV light, and cooling via conductive packaging to manage the thermal issues associated with close integration of a large number of solid-state devices (Figure 4).

SLMs are not inherently limited in shape or size. In theory any shape or size array can be constructed. A meters-long light bar, a doughnut-shaped source or a source only millimeters across are all equally achievable. The limitation is the cost of manufacturing custom shapes compared to using standard, modular-shaped arrays.

These individual arrays may be positioned in any orientation and will produce uniform output. This flexibility and scalability of the technology allows SLM arrays to be constructed for virtually any physical curing task without waste or inefficiency.

SLM arrays may be literally butted up against each other with virtually no edge effects. Using modular arrays, large surfaces can be irradiated "seamlessly" by multiple sources (Figure 5).

Surface Uniformity

A characteristic related to the combination of form factor and output consistency is the general two-dimensional homogeneity of large SLM arrays. Precision radiometric measurements made over a wide 8" x 8" planar surface found less than 4% variation in UV intensity measured anywhere by probes monitoring peak intensity at a distance of 2.1" from the solid-state UV source.

The uniformity for UV SLM technology is superior to the variations observed with arc lamp sources, which range between 10-20% with a three sigma confidence interval of +/-30-60%, causing users to overdrive the light source to ensure reliable curing. This compounds the problem of inefficient coupling and long term stability.

Rapid On/Off Capability

There is a good deal of common misunderstanding concerning the true "on/off" capability for different UV sources. Microwave lamps require nearly 20 seconds to cycle from 100% to zero to 100% again. However most of these microwave systems employ "standby" circuitry, which allows the lamp to cycle from 100% to standby and back to 100% in just under 10 seconds.

If switched completely off, arc lamps could require over 15 minutes to restart and stabilize enough for use. The pressure buildup inside the lamp makes re-striking the arc more difficult until the internal atmosphere in the quartz tube returns to a low temperature (and pressure). Therefore, shutters are commonly employed to block the escape of UV light mechanically. These shutters can be triggered in a fraction of a second, making them quicker in practice than even microwave sources.

By comparison, a solid-state UV source cycles from zero to 100% in only 3 milliseconds. The near-instantaneous cycling allows for immediate start and stop of UV energy, and may offer other advantages by pulsing or generating curing recipes to control curing characteristics of the formulation. The performance of the cured product can be adjusted by tight control of irradiance and dose.

New Formulations

To take full advantage of solid-state UV curing, new formulations may be needed. Suppliers of photo-reactive chemicals have responded over the last few decades by developing photoinitiators that react well to the light sources popular for UV curing. Their commercial sales literature ties together their chemistry and the popular lamps.

While these chemistries usually work with solid-state sources, they were not developed, and are not ideal for them. By a more thoughtful selection much better results can be obtained.

The results of field trials show that materials optimized for UV SLM technology not only cure 6x faster, but cure with better results than using a conventional UV arc lamp (Figure 7).

Additional field trials in the wood market revealed that adding even a small percentage of a second photoinitiator (TPO) suited to 380-420 nm light to a commercial formulation containing Ciba Irgacure 819 (BAPO) had a significant and positive effect on the rate of cure and surface properties.

Work is now underway to categorize what chemistry works well with these new sources to aid formulators and to develop successful applications. Work is also underway to develop new cure mechanisms and photo-reactive chemicals that are optimized for these devices in the same way that the previous generation of lamps and chemistry were developed. c

Acknowledgements

The author wishes to thank the great support and assistance of Mark Owen, Tom Molamphy, Dr. Alex Schreiner and Jon Marson at Phoseon Technology. We are also indebted to the team at the University of Akron for their ongoing assistance with real-time IR studies; Dr. Mark Soucek, Dr. Hua Gu and Kosin Wutticharoenwong and to our colleagues at Kalcor Coatings, Tim Hlabse and Eileen Adams for the help in coating formulation.

For more information, visit www.phoseon.com.