Light against Viruses – 1:0

The Basis: 

the semi­con­ductor crystal

Alumi­nium gallium nitride (AlGaN) is the basis for UV-C light. This semi­con­ductor crystal is a compound of aluminum, gallium and nitrogen. This means that UV-C-LEDs are funda­men­tally diffe­rent from blue LEDs, on which the commer­ci­ally available white LEDs are based. The compound for blue LEDs is indium gallium nitride (InGaN). To achieve UV-C light the deve­lo­pers replaced the indium with aluminum. Sounds simple but it’s not, because the two compounds have diffe­rent mate­rial proper­ties. Which means that AlGaN behaves comple­tely differ­ently from InGaN. As a result, new deve­lo­p­ment and manu­fac­tu­ring processes are also needed.

Shine on you crazy crystal

The big chall­enge is to make the mate­rial emit light. That’s not some­thing it does by itself. The AlGaN semi­con­ductor is a very good insu­lator even at room tempe­ra­ture. This essen­ti­ally non-conduc­ting mate­rial only becomes a conduc­ting mate­rial through the selec­tive addi­tion of atoms. This process is called epitaxy. In the chemical depo­si­tion process, the raw mate­rials are trans­ferred to a reactor in gaseous form where they react with each other under extre­mely high tempe­ra­tures of well over 1,100 degrees Celsius on a suitable layer known as a substrate.

The conven­tional silicon used in the semi­con­ductor world is unsui­table as a substrate because the grid spacings of the two crys­tals differ too much and UV-C radia­tion is absorbed. In most cases, aluminum oxide (Al2O3), also known as sapphire, is ther­e­fore used for blue, green and white LEDs based on InGaN. Sapphire is also the substrate of choice for produ­cing UV-C light in the case of aluminum gallium nitride because its physical proper­ties match very well with those of the semi­con­ductor. It is also trans­pa­rent for UV-C radia­tion.

Wafer thin

The epitaxial process itself is largely the same as in the produc­tion of conven­tional LEDs. Well over one hundred wafer-thin layers are depo­sited one on top of the other. The most important ones, the so-called “quantum wells”, are thinner than ten atomic layers. However, a lot of theo­re­tical work is needed to ensure that the LED emits the short-wave UV-C light at the correct wave­length of 260 to 280 nano­me­ters. The layer struc­ture is complex, and the inter­ac­tions between the layers are also complex. How exactly they interact with each other is only known when the chip is ready, but their thic­k­ness and compo­si­tion are deter­mined in advance on a theo­re­tical basis in order to obtain the best possible result.

Finally, ever­y­thing comes toge­ther. The first step is to create a buffer layer, then the active LED layers and finally a contact layer is applied. At this point, rela­tively fine adjus­t­ments can be made to the wave­length of the light. The key is the ratio of aluminum and gallium in the active LED layers.

Trou­ble­shoo­ting with an atomic force micro­scope

Ques­tions and fresh chal­lenges arise again and again during the fabri­ca­tion process. Here’s a good example: As soon as the LED struc­ture is built up layer by layer as far as the contact layer it already emits some light, but not enough. The search for a means of adjus­t­ment begins. Is the selected struc­ture the best one? Where are the compo­si­tions wrong? Where do struc­tures need to be removed or added to make impro­ve­ments? Where are elec­trons being lost?

To search for the fault an atomic force micro­scope may be used to minu­tely scan the surface of the buffer layers and make the atomic struc­tures visible. Are they smooth or rough? Are there holes or defects caused by particles?

X-rays also show how the layers are composed and where the tensions between them lie. The LED layers are tested for their conduc­ti­vity. Depen­ding on how well the layers conduct, the deve­lo­pers can tell whether they are on the right track or whether they need to rethink the layers and their compo­si­tion. Again and again, this means taking one or more steps back or even starting again from scratch and rethin­king ever­y­thing.

Chips no bigger than a grain of sand

This also applies to the next step – chip proces­sing. The wafer produced in the epitaxy process is struc­tured, provided with metal cont­acts and then divided into small cubes, the chips. The tiny crys­tals, no bigger than a grain of sand, must be cut out of these wafer. They are the ones that will later emit UV-C light.

As things stand today, however, they do not yet do this effi­ci­ently enough. Whereas blue LEDs can achieve an effi­ci­ency of well over 60 percent, UV-C-LEDs curr­ently only manage about four percent. For deve­lo­pers, this is still the hardest nut to crack.

Well protected:

the package

A chip by itself is extre­mely deli­cate. It needs protec­tion, and that protec­tion is known as a package. The package has a number of func­tions. It provides the elec­trical connec­tion to deliver power to the LED. It also ensures that a place­ment machine can pick up the LED and place it on a pcb without dama­ging it. At the same time, it performs optical func­tions. With a small reflector in the package, it is possible to direct the light or radia­tion.

Hot stuff

Thermal manage­ment plays an important role in the design of the package. Because the hotter an LED gets, the faster it ages. Only four percent of the elec­trical energy that is fed into the LED is emitted as UV-C radia­tion, the rest is power loss – and that gene­rates heat. This means that the package has to remove as much heat from the chip as possible. The better it can do this, the cooler the chip stays and the longer it will last. This is a critical point for the deve­lo­p­ment process because a larger package, for example, will lead to better heat dissi­pa­tion, but larger often means more expen­sive.

Deve­lo­pers simu­late whether and how well thermal manage­ment will work before they build proto­types. They begin this process with very broad brush­strokes, explo­ring diffe­rent concepts. Finite element methods help them to assess the effects of mate­rial inter­ac­tions. The 3D element to be computed is broken down into smaller elements which are then analyzed for their physical beha­vior. This is important because the indi­vi­dual mate­rials that make up the package have diffe­rent coef­fi­ci­ents of expan­sion. This leads to internal tensions as the tempe­ra­ture rises. Simu­la­tions can deter­mine whether the connec­tion points will with­stand these tensions so that design solu­tions can be found if neces­sary. Small features such as angled or rounded edges often influence how well tensions can be miti­gated.

Shaping the light

Depen­ding on the appli­ca­tion in which the UV-C light is to be later used for disin­fec­tion, the emis­sion charac­te­ristics also have to be tail­ored. Among other things, the emitted light can either be fanned out widely or bundled tightly. If, for example, water or air passing over the LED is to be puri­fied, it is advi­sable not to closely cluster the radia­tion source but to distri­bute it along a water or venti­la­tion pipe so that the medium is uniformly irra­diated. In contrast, disin­fec­tion of water at a faucet or objects in small portable boxes requires signi­fi­cantly smaller and concen­trated flows of radia­tion. This can be achieved through the design of the package.

Rethin­king is also the order of the day when it comes to selec­ting the mate­rials for the package. Simply adop­ting the mate­rials used for conven­tional LEDs will not work because the plas­tics and organic encap­su­la­tion mate­rials used for those LEDs would decom­pose due to the high-energy radia­tion of the UV-C light. So deve­lo­pers look for mate­rials that will produce a stable, solid and durable package. Cera­mics, glass and meta­lized foils have proved their worth to date.

Test to failure

If a package variant seems promi­sing in the simu­la­tion it will be built as a proto­type and must survive destruc­tive testing. In this test, the package is subjected to incre­asing loads until it fails. It is exposed to signi­fi­cantly more sever opera­ting condi­tions and loads than would occur in the real envi­ron­ment. Endu­rance tests reveal whether joints are dete­rio­ra­ting and how the mate­rial is aging. This in turn allows conclu­sions to be drawn about the life of the LED.

Once all the requi­re­ments have been defined and the package has been desi­gned, the produc­tion depart­ment will check whether the package can be produced in large quan­ti­ties. If so, that signals the end of the deve­lo­p­ment project and the start of series produc­tion.

In use: 

UV-C-products

Products with conven­tional lamps as sources of UV-C are already available. OSRAM’s AirZing is one of them. First deve­loped for the Chinese market, the device was initi­ally used in early 2020 to disin­fect the air in the hospi­tals set up in Wuhan during the intense phase of the corona pandemic. AirZing is now being used throug­hout the world. If there are no people in a room, the device is swit­ched on and cleans the air and surfaces with its UV-C radia­tion. As soon as someone enters the room, sensors detect this move­ment and imme­dia­tely turn off the light to prevent damage to skin and eyes.

The issue of effi­ci­ency

But these devices are still equipped with conven­tional discharge lamps that produce UV-C light. The effi­ci­ency of the LEDs is not yet good enough to flood large rooms with UV-C light. For a conven­tional lamp with 30 W of elec­trical power, effi­ci­ency is between 30 and 40 percent, which equates to about twelve watts of UV-C output. If you want to replace a 30 W lamp for disin­fec­ting a room, you would need a large number of LEDs. That would not make economic sense. Repla­cing a lamp with an LED is just not possible yet.

The LED is ideal, however, for use in mobile appli­ca­tions. It is compact, shatter-proof, vibra­tion-resistant and does not contain mercury. For deve­lo­pers that’s reason enough to create devices in which the LED does a good job even with low effi­ci­ency. A compact device for disin­fec­ting small areas is a promi­sing concept.

Small areas at the start

The idea here is for UV-C light to disin­fect the air in a room by means of perma­nent circu­la­tion. The used air is drawn into the device where it is steri­lized and then directed back into the room. To disin­fect a defined volume of space, a certain UV-C output is required and a certain velo­city at which the air is chan­neled past the UV-C source.

With a normal LED unit this is not possible due to the low effi­ci­ency and the high cost. The solu­tion: a HEPA filter (high-effi­ci­ency air filter) installed upstream of the UV-C LEDs to trap the viruses. This germ trap is then cleaned by a small UV-C LED unit. Instead of disin­fec­ting all the air, the UV-C light shines only on the filter and steri­lizes it. LEDs could ther­e­fore soon be opera­ting in large areas.

Rese­arch is also being carried out on small devices, such as a UV-C LED box in which the light removes viruses and germs from the surfaces of cell phones and other objects. That would be a “germ killer to go”.