Technical Insight
Violet VCSELs poised for action?
Blue laser diode inventor, Shuji Nakamura, recently unveiled a non-polar violet VCSEL, claiming 'one of the biggest breakthroughs in laser diode technology'. Compound Semiconductor talks to Nakmura and industry colleagues, to find out more.
Researchers from the University of California, Santa Barbara, have developed a non-polar, violet VCSEL, which they say opens the door to a host of new applications including pico-projectors for smartphones, mobile cinema and automotive lighting.
Developed by Shuji Nakamura and colleagues Daniel Feezell, Casey Holder and Stephen DenBaars, the single-mode VCSEL operates at room temperature, and is the latest of several optoelectronic devices the team has grown, using MOCVD, on m-plane GaN substrates.
To date, most GaN-based devices have been grown on substrates sliced from the c-plane of a GaN crystal. Nakamura demonstrated the first c-plane (AlGaIn)N laser diode as early as 1996 and today c-plane devices can be found in myriad commercial products for high density optical storage, laser printing and display applications.
But despite success, c-plane GaN substrates have a major drawback. Large polarised electric fields induce charge separation in quantum wells, restricting the recombination efficiency of electrons and holes, and ultimately, limiting a device's performance.
As Nakamura points out: “Non-polar nitride devices have the potential to outperform their polar counterparts, for example, you can get a much higher optical gain using m-plane substrates.”
With this in mind, the UCSB researcher and colleagues have been developing devices, based on substrates cut from the m-plane within GaN crystals. In 2005, blue GaN LEDs were unveiled, followed by GaN violet laser diodes in 2007. Come 2009, Nakamura and colleagues had developed blue-green laser diodes, but now the team has upped the ante, by developing the GaN VCSEL.
While VCSELs tend to have a lower power output than edge-emitting lasers, the devices are cheaper and more efficient to manufacture. Edge-emitters cannot be tested until the end of production, while VCSEL arrays can be tested on the wafer, during fabrication. And because the VCSEL beam is emitted perpendicular to the laser's active region – as opposed to parallel with an edge-emitter – as many as ten thousand devices can be processed simultaneously on a single three inch wafer.
Given this potential to cut costs, researchers worldwide have been earnestly developing GaN VCSELs that emit at room temperature. In 2007, Switzerland-based Ecole Polytechnique Fédérale de Lausanne (EPFL) researchers demonstrated blue lasing in an optically pumped AlInN/GaN VCSEL, while a year later, Nichia revealed the first electrically-pumped GaN VCSEL to emit continuous-wave violet light. Last year, the Japan-based optoelectronics developer also unveiled blue and green VCSELs, but not one of these VCSELs was grown on an m-plane substrate. Why?
“The easiest orientation is the c-plane, as it's the most common substrate you can find, and also the best orientation if you consider the price,” says Eric Feltin, researcher at EPFL spin-off, Novagan. “UCSB has been working with m-plane substrates, and have fabricated good devices. But right now, high quality, c-plane substrates are more widely available.”
Nakamura disagrees, highlighting how his group has, for some years now, used low defect density, m-plane substrates, manufactured by Mitsubishi Chemical. Indeed, the Japan-based corporation started working on the novel substrate several years ago in an effort to produce next-generation white LEDs. And as Feltin concedes, if the demand for m-plane substrates increases, c-plane substrates could lose out.
.jpg)
Shuji Nakamura and his research group at UCSB demonstrate the first violet, non-polar, m-plane VCSEL based on GaN.
So where next for the Nakamura's violet VCSEL? While some researchers have poured their energies into developing novel distributed Bragg reflectors to ease VCSEL fabrication and avoid the lattice mis-match strains that can crack epi-wafers, Nakamura claims dielectric DBRs are adequate. “We don't have to worry about lattice mis-matches and these are easy to make,” he says.
And as he highlights, his group has developed a new processing technology to control the length of the VCSEL cavity. “While other groups use lapping and polishing, we have developed selective etching to control the VCSEL cavity length,” he asserts. “By doing this we have demonstrated single mode [lasing] for the first time.”
Still, research is clearly ongoing. At the time of writing, the violet VCSEL had a 70mA threshold current, using a 10 micron aperture, and an output power of 19 microW, under pulsed operation.
Nakamura is determined to build on this. “Yes it's a small power, but this is just a demonstration. We're going to minimise the optical losses and soon we will increase the power,” he says. “Right now we are working very hard.”
Developed by Shuji Nakamura and colleagues Daniel Feezell, Casey Holder and Stephen DenBaars, the single-mode VCSEL operates at room temperature, and is the latest of several optoelectronic devices the team has grown, using MOCVD, on m-plane GaN substrates.
To date, most GaN-based devices have been grown on substrates sliced from the c-plane of a GaN crystal. Nakamura demonstrated the first c-plane (AlGaIn)N laser diode as early as 1996 and today c-plane devices can be found in myriad commercial products for high density optical storage, laser printing and display applications.
But despite success, c-plane GaN substrates have a major drawback. Large polarised electric fields induce charge separation in quantum wells, restricting the recombination efficiency of electrons and holes, and ultimately, limiting a device's performance.
As Nakamura points out: “Non-polar nitride devices have the potential to outperform their polar counterparts, for example, you can get a much higher optical gain using m-plane substrates.”
With this in mind, the UCSB researcher and colleagues have been developing devices, based on substrates cut from the m-plane within GaN crystals. In 2005, blue GaN LEDs were unveiled, followed by GaN violet laser diodes in 2007. Come 2009, Nakamura and colleagues had developed blue-green laser diodes, but now the team has upped the ante, by developing the GaN VCSEL.
While VCSELs tend to have a lower power output than edge-emitting lasers, the devices are cheaper and more efficient to manufacture. Edge-emitters cannot be tested until the end of production, while VCSEL arrays can be tested on the wafer, during fabrication. And because the VCSEL beam is emitted perpendicular to the laser's active region – as opposed to parallel with an edge-emitter – as many as ten thousand devices can be processed simultaneously on a single three inch wafer.
Given this potential to cut costs, researchers worldwide have been earnestly developing GaN VCSELs that emit at room temperature. In 2007, Switzerland-based Ecole Polytechnique Fédérale de Lausanne (EPFL) researchers demonstrated blue lasing in an optically pumped AlInN/GaN VCSEL, while a year later, Nichia revealed the first electrically-pumped GaN VCSEL to emit continuous-wave violet light. Last year, the Japan-based optoelectronics developer also unveiled blue and green VCSELs, but not one of these VCSELs was grown on an m-plane substrate. Why?
“The easiest orientation is the c-plane, as it's the most common substrate you can find, and also the best orientation if you consider the price,” says Eric Feltin, researcher at EPFL spin-off, Novagan. “UCSB has been working with m-plane substrates, and have fabricated good devices. But right now, high quality, c-plane substrates are more widely available.”
Nakamura disagrees, highlighting how his group has, for some years now, used low defect density, m-plane substrates, manufactured by Mitsubishi Chemical. Indeed, the Japan-based corporation started working on the novel substrate several years ago in an effort to produce next-generation white LEDs. And as Feltin concedes, if the demand for m-plane substrates increases, c-plane substrates could lose out.
Shuji Nakamura and his research group at UCSB demonstrate the first violet, non-polar, m-plane VCSEL based on GaN.
So where next for the Nakamura's violet VCSEL? While some researchers have poured their energies into developing novel distributed Bragg reflectors to ease VCSEL fabrication and avoid the lattice mis-match strains that can crack epi-wafers, Nakamura claims dielectric DBRs are adequate. “We don't have to worry about lattice mis-matches and these are easy to make,” he says.
And as he highlights, his group has developed a new processing technology to control the length of the VCSEL cavity. “While other groups use lapping and polishing, we have developed selective etching to control the VCSEL cavity length,” he asserts. “By doing this we have demonstrated single mode [lasing] for the first time.”
Still, research is clearly ongoing. At the time of writing, the violet VCSEL had a 70mA threshold current, using a 10 micron aperture, and an output power of 19 microW, under pulsed operation.
Nakamura is determined to build on this. “Yes it's a small power, but this is just a demonstration. We're going to minimise the optical losses and soon we will increase the power,” he says. “Right now we are working very hard.”
