Note: Most of the information on Diode Pumped Solid State (DPSS) Lasers has moved to the chapter: Solid State Lasers.
Diode lasers use nearly microscopic chips of Gallium-Arsenide or other exotic semiconductors to generate coherent light in a very small package. The energy level differences between the conduction and valence band electrons in these semiconductors are what provide the mechanism for laser action. This is not the sort of laser you can build from scratch in your basement as the required fabrication technology costs megabucks or more to set up. You will have to be content with powering a commercial laser diode from a home-made driver circuit or using a pre-packaged module like a laser pointer. Fortunately, laser diodes are now quite inexpensive (with prices dropping as you read this) and widely available.
The active element is a solid state device not all that different from an LED. The first of these were developed quite early in the history of lasers but it wasn't until the early 1980s that they became widely available - and their price dropped accordingly. Now, there are a wide variety - some emitting many *watts* of optical power. The most common types found in popular devices like CD players and laser pointers have a maximum output in the 3 to 5 mW range.
A typical configuration for a common low power edge emitting laser diode is shown below:
+ +
o o
______________|______________ _______|_______
Laser | P type semiconductor | Laser | P type |
beam | | beam | |
<=======|:::::::::::::::::::::::::::::|=======> |ooooooooooooooo|
| Junction---^ | | |
End ->| N type semiconductor |<- End | N type |
facet |_____________________________| facet |_______________|
| |
o o
- -
(Side view) (End view)
|<----------------------- 1 mm ------------------------>|
This configuration above is called a 'homojunction' since there is only one
P-N junction. It turns out there are benefits to using several closely spaced
junctions formed by the use of layers of P and N type materials. These are
called 'heterojunction' laser diodes. There are many many more advanced
structures in use today and new ones are being developed as you read this!
For example, see the section: Vertical Cavity
Surface Emitting Laser Diodes (VCSELs) for a description of one type that
has the potential to have a dramatic impact in many areas of technology.The 'end facets' are the mirrors that form the diode laser's resonant cavity. These may just be the cleaved surfaces of the semiconductor crystal or may be optically ground, polished, and coated.
For these types of integrated laser diodes, everything takes place inside the chip. Therefore, the output wavelength is fixed and determined by the properties of the semiconductor material and the device's physical structure. Or, at least that's the way it is supposed to work though with some, reflection of the laser light back into the chip can cause stability problems or even be used to advantage to frequency stabilize the output. There are also tunable diode lasers using external cavity optics to provide a continuous and in some cases, quite wide range of wavelengths without mode hopping. See EOSI's Tunable Laser Diode Systemsfor an example of one commercially available product. You really don't want to ask about the price! :-)
There are also pulsed laser diodes requiring many amps to to reach threshold and providing watts of output power but only for a short time - microseconds or less. Average power is perhaps a few mW. These are gallium arsenide (GaAs) heterojunction laser diodes. They are not that common today but some surplus places are selling diodes like these as part of the Chieftain tank rangefinder assembly. They mention the high peak power output but not the low average power. :( Modern devices with similar specifications are also available. One manufacturer is Infineon Technologies AG.
Electrical input to the laser diode may be provided by a special current controlled DC power supply or from a driver which may modulate or pulse it at potentially very high data rates for use in fiber optic or free-space communications. Multi-GHz transmission bandwidth is possible using readily available integrated driver chips.
However, unlike LEDs, laser diodes require much greater care in their drive electronics or else they *will* die - instantly. There is a maximum current which must not be exceeded for even a microsecond - and this depends on the particular device as well as junction temperature. In other words, it is not sufficient in most cases to look up the specifications in a databook and just use a constant current power supply. This sensitivity to overcurrent is due to the very large amount of positive feedback which is present when the laser diode is lasing. Damage to the end facets (mirrors) can occur very nearly instantaneously from the concentrated E/M fields in the laser beam. Closed loop regulation using optical feedback to stabilize beam power is usually implemented to compensate for device and temperature variations. See the sections on CD and visible laser diodes later in this document before attempting to power or even handle them. Not all devices appear to be equally sensitive to minor abuse but it pays to err on the side of caution (from the points of view of both your pocketbook and ego!).
In their favor, laser diodes are very compact - the active element is about the size of a grain of sand, low power (and low voltage), relatively efficient (especially compared to the gas lasers they replaced), rugged, and long lived if treated properly.
They do have some disadvantages in addition to the critical drive requirements. Optical performance is usually not equal to that of other laser types. In particular, the coherence length and monochromicity of some types are likely to be inferior. This is not surprising considering that the laser cavity is a fraction of a mm in length formed by the junction of the III-V semiconductor between cleaved faces. Compare this to even the smallest common HeNe laser tubes with about a 10 cm cavity. Thus, these laser diodes would not be suitable light sources for high quality holography or long baseline interferometry. But, apparently, even a $8.95 laser pointer may work well enough to experiment in these areas and some results can be surprisingly good despite the general opinion of laser diode performance.
Even if not as good as a helium-neon laser in the areas of coherence and stability, for many applications, laser diodes are perfectly adequate and their advantages - especially small size, low power, and low cost - far outweigh any faults. In fact, these 'faults' can prove to be advantageous where the laser diode is being used simply as an illumination source as unwanted speckle and interference effects are greatly reduced.
As noted, not all laser diodes have the same performance. See the section: Interferometers Using Inexpensive Laser Diodes for comments that suggest some common types may indeed have beam characteristics comparable to typical HeNe lasers. And, for short range applications, see: Can I Use the Pickup from a CD Player or CDROM Drive for Interferometry?. Also see the section: Holography Using Cheap Diode Lasers.
The following sites provide some relatively easy to follow discussions of the principles of operation, construction, characteristics, and other aspects of laser diode technology:
li>Lumex Tech Notes. Articles on a variety of topics including laser diode construction.
The closeups below were scanned at 600 dpi - laser diodes (at least the small ones we are dealing with) are really not this HUGE! These two laser diodes can also be found in the group photo, above.
The Closeup of laser diode from the Sony KSS361A Optical Pickup shows a type that is found in many CD players and CDROM drives manufactured by Sony. The actual laser diode is inside the brass barrel shown in the photo of the optical pickup. The front of the package is angled so that the exit window (anti-reflection coated) is also mounted at what may be the Brewster angle, probably to further prevent stray reflections from the window's surfaces from feeding back into the laser diode's cavity or interfering with the detected signal. (At the Brewster angle, light polarized parallel to the window is totally reflected and light polarized perpendicular to it is totally transmitted. The output of these edge emitting laser diodes is polarized. See the section: What is a Brewster window?.)
The Closeup of Typical Laser Diode shows one that is from a laser printer. It was mounted in a massive module (relative to the size of this laser diode, at least) which included the objective lens and provided the very important heat sink. In some high performance laser printers, a solid state Peltier cooler is used to stabilize the temperature of the laser diode. The low power laser diodes in CD and LD players, and CDROM and other optical drives (at least read-only types) get away with at most, the heat sink provided by the casting of the optical block - and many don't even need this being of all plastic construction.
One can think of an LED as a laser without a feedback cavity. The LED emits photons from recombining electrons. It has a very broad spectrum.
When we add a high Q cavity to it, the feedback can be high enough to trigger true laser action. Most laser diodes have the cavity built right into the device but there are such things as external cavity diode lasers.
The addition of the high Q cavity cuts down drastically the number of modes operating (in fact, it is almost improper to speak of mode structure with an LED. The result is that the emission line narrows drastically (more monochromatic) and the beam narrows somewhat spatially. One can still not easily get true single mode lasing with normal diode lasers, however, so the line will not be as sharp as a gas laser, nor the beam as narrow.
For more info, see the section: How LEDs Compare to Laser Diodes - Wavelengths, Power, Focus, Safety.
(From: Don Stauffer (stauffer@htc.honeywell.com).)
Yes indeed, a diode laser is a true laser. That being said, looking at matters quantitatively, it is harder to make a diode laser with a very narrow line emission than a gas laser or large crystal laser. Adding cavity length to a laser in general acts to narrow the line (in spectral space, though a higher Q cavity does tend to narrow beam in space also). It is possible to use a larger, high Q external cavity with a laser diode to increase its coherence.
(From: David Schaafsma (drdave@jnpcs.com) and Rajiv Agarwal (agarca@giascl01.vsnl.net.in).)
A couple of minor points:
High Q cavities narrow the spatial profile only if they are confocal - planar high Q cavities (as in diode lasers, and especially vertical-cavity diode lasers) are prone to problems with walk-off and the mode must be confined physically.
In a gas laser, you also start with a much narrower fluorescence line and thus the gain spectrum is limited spectrally. Diode lasers (being band-to-band or excitonic semiconductor transitions) have much broader fluorescence spectra.
The typical edge-emitting diode laser actually lases in quite a few fundamental modes (especially when operated using its own facets as the cavity) and though each lasing mode is "monochromatic", the overall spectrum really isn't. External cavities are really the only way to obtain approximately single mode operation from an edge-emitting diode laser.
VCSELs are usually true single mode devices. The reason you can get away with lengthening the cavity in a gas laser is that you don't need to worry about lowering the free spectral range because the gain bandwidth is small.
DFB or DBR lasers achieve very similar results and have Side mode suppression ratios better than 30 db. These lasers have been the mainstay of Optical fiber base telecom for a while now.
DFB Lasers are use for long haul telecommunications network - the kind used by say Sprint (>1GB for up to 25 miles) for their phone networks between cities. These have been for Trans-Atlantic cables (TAT) between US and Europe. LEDs are used more for FDDI type application between computers (~100Mb and less than 1 mile).
(From: Vishwa Narayan (vishwa.narayan@ericsson.com).)
While LEDs are quite popular in Datacom applications (read short distances), Telecom applications typically use DFBs, either directly modulated for low speeds (e.g., OC-3 155 Mb/sec) or externally modulated for high speeds (e.g., OC-48 2.5 Gb/sec). Distances can typically range over tens of kilometers, to hundreds of kilometers with optical amplification, sans repeaters.
One should never look into the beam of any laser - especially if it is collimated. Use an indirect means of determining proper operation such as projecting the beam onto a white card, using an IR detector card or tester (where needed), or laser power meter.
Currently, green laser pointers are not simple diode lasers but are Diode Pumped Solid State Frequency Doubled (DPSSFD) lasers (this will probably change in the not too distant future, however). For a given power, green appears substantially brighter than red wavelengths but are also limited a maximum power of 5 mW.
See: NRPB Is1-98 for a discussion of some of the specific safety risks, myths, and concerns with respect to laser pointers.
(From: Gregory Makhov (lsdi@gate.net).)
According to a recent report by Dr. David Sliney, who is one of the leading "gurus" of laser safety, there are no confirmed accidents or injuries caused by laser pointer of 5 milliwatts radiant power or less. There is an awful lot of nonsense and false claims about this. Pointers are extremely bright, can cause visual distraction, afterimages, and other effects, such as headaches, but under most any typical usage condition, DO NOT cause eye injury. Dr. Sliney works for US Army, and has published papers and books on laser safety for over 20 years.
For IR laser diodes in particular, especially if you are considering selling a product:
(Portions from: Steve Roberts (osteven@akrobiz.com).)
You need to take a close look at the CDRH rules, because there is no blink reflex in the IR. IR diode lasers are considered much more dangerous and therefore are in a higher class. CDRH has a curve of power versus wavelength that is used for determining safety classes. The only way a IR laser gets less then a IIIb rating (read: dangerous) is if the beam is totally enclosed or of very low power. Go to CDRH, call them and request a manufacturers' packet by mail. It's huge and confusing, but covers the requirements for products using IR laser diodes such as 3-D scanners, perimeter sensors, and so forth.
Visible laser diodes have replaced helium-neon lasers in supermarket checkout UPC scanners and other bar code scanners, laser pointers, patient positioning devices in medicine (i.e., CT and MRI scanners, radiation treatment planning systems), and many other applications. The first visible laser diodes emitted at a wavelength of around 670 nm in the deep red part of the spectrum. More recently, 650 nm and 635 nm red laser diodes have dropped in price.
Due to the nonuniformity of the human eye's response, light at 635 nm appears more than 4 times brighter than the same power at 670 nm. Thus, the newest laser pointers and other devices benefitting from visibility are using these newer technology devices. Currently, they are substantially more expensive than those emitting at 670 nm but that will change as DVDs become popular:
Laser diodes in the 635 to 650 nm range will be used in the much hyped DVD (Digital Video - or Versatile - Disc) technology, destined to replace CDs and CDROMs in the next few years. The shorter wavelength compared to 780 nm is one of several improvements that enable DVDs to store about 8 times (or more - 4 to 5 GB per layer, the specifications allow up to 2 layers on each side of a CD-size disc!) the amount of information or video/audio as CDs (650 MB). A side benefit is that dead DVD players and DVDROM drives (I cannot wait) will yield very nice visible laser diodes for the experimenter. :-)
Like their IR cousins, the typical maximum power from these devices is around 3 to 5 mW. Cost is in the $10 to $50 for the basic laser diode device - more with optics and drive electronics. Higher power types (10s of mW) are also available but expect to spend several hundred dollars for something like a 20 mW module. Very high power diode lasers using arrays of laser diodes or laser diode bars with power output of WATTs or greater may cost 10s of thousands of dollars!
___
| | Metal case
| |_______________________________
| \
| _____________________________ |
| | | |
LD -------:===:------------------+ | |
| |__ | |__|
| | |___ ______|______ : :
| | | | | | : :
PD -------:===:----+ |<---|:::::::::::::|============> Main beam
| | |___|____|_____________|_ : : (divergent)
| | Photodiode Laser diode | :__:
| |\__________________________| | | Protective window
Com -------+ | Heat sink | |
| |_____________________________| |
| |
| _______________________________/
| |
|___|
The main beam as it emerges from the laser diode is wedge shaped and highly
divergent (unlike a helium-neon laser) with a typical spread of 10 by 30
degrees. External optics are required to produce anything approaching a
parallel (collimated) beam. A simple (spherical) short focal length convex
lens will work reasonably well for this purpose but diode laser modules and
laser pointers might use a lens where at least one surface is aspheric (not
ground to a spherical shape as are with most common lenses).In the case of a sample I removed from a dead diode laser module, the surface facing the laser diode was slightly curved and aspheric while the other surface was highly curved and spherical. The effective focal length of the lens was about 5 mm. It appeared similar to the objective lens of a CD player - which was perhaps its original intended application and thus a low cost source for such optics.
Due to the nature of the emitting junction which results in a wedge shaped beam and unequal divergence (10 x 30 degrees typical), a laser diode is somewhat astigmatic. In effect, the focal length required to collimate the beam in X and Y differs very slightly. Thus, an additional cylindrical lens or a single lens with an astigmatic curvature is required to fully compensate for this characteristic. However, the amount of astigmatism is usually small and can often be ignored. The general beam shape is elliptical or rectangular but this can be circularized by a pair of prisms.
The light from these edge emitting laser diodes is generally linearly polarized. You can easily confirm this even with a simple laser pointer by reflecting at about a 45 degree angle from a piece of glass (not a metal coated mirror). Rotate the pointer and watch the reflection - there will be a very distinct minimum and maximum with the elongated shape of the beam at close range being aligned with the glass and perpendicular, respectively. For the advanced course, determine the Brewster angle. :)
For addition information, see the section: Beam Characteristics of Laser Diodes.
The beam from the back end of the laser diode chip hits a built-in photodiode which is normally used in an opto-electronic feedback loop to regulate current and thus beam power. Note that the photodiode is likely mounted at an angle (not possible to show in ASCII) so that the reflection does not interfere with the operation of the laser diode.
CAUTION: Some complete modules may use the reflection from external optics along with an external photodiode for power stabilization as it is more accurate since the actual output beam is sampled. For these, one should never attempt to clean or even focus the lens when operating near full power as this may disturb the feedback loop and damage the laser diode.
Note: Some of the symbols below are not exactly what is found in the datasheet so they can be represented in ASCII. However, the meaning should be obvious.
Parameter Symbol Conditions Min Typ. Max Unit
------------------------------------------------------------------------------
Threshold current Ith 30 40 mA
Operating current Iop Po = 5mW 35 45 mA
Operating voltage Vop Po = 5mW 2.2 2.4 V
Wavelength lambdap Po = 5mW 650 660 nm
Radiation angle
Perpendicular theta_|_ Po = 5mW 22 30 40 Deg.
Parallel theta|| Po = 5mW 5 7 12 Deg.
Positional accuracy dx,dy,dz Po = 5mW +/-150 um
Angular accuracy
Perpendicular phi_|_ Po = 5mW +/-3 Deg.
Parallel phi|| Po = 5mW +/-3 Deg.
Differential eff. nD Po = 5mW 0.3 0.6 0.9 mW/mA
Astigmatism As Po = 5mW 7 15 um
Monitor PD current Imon Po = 5mW, Vr = 5V 0.05 0.1 0.25 mA
Descriptions of the parameters are provided below:
"I was just browsing Meredith Instrument's site, and noticed that they have 635 nm diodes rated at 500 mW. Has anyone ever dealt with these things? Looking around on the site, it appears I could put together a half watt red diode laser for under $600, or a 250 mW one for under $400. Is there some catch to using these? The whole setup would be cheaper than a 25 mW HeNe laser".Yes. Aside from the ease with which one of those pricey diodes can be blown out due to improper drive, the beam quality is no where near that of even a cheap HeNe laser. It is multimode and very non-circular and astigmatic. The latter can probably be dealt with using some (expensive) optics. However, multimode operation means that these are unsuitable for applications like holograpy or interferometry.
(From: Frank DeFreitas (director@holoworld.com).)
I have a 500 mW laser diode from Polaroid. 660nm I believe. It needs the heftier driver that Meredith offers - the one that can put out 1000 mA or so. The laser diode is gain guided/multi-mode, rather than index guided/single (mono) mode -- so you can pretty much forget any application that would call for any type of coherency or high contrast fringes.
The output beam profile is basically a line. It is very similar to taking a standard HeNe beam and sending it through a cylindrical lens. (However, on the other hand, I'm wondering if a cylindrical lens would actually help it when used in the other dimension. Or at least bring it to a spot which could be collimated utilizing secondary optics in the path.)
I'd also like to point out that it's not a diode to play around with. The optical output at 500 mW is not going to knock any missles out of the sky, but will certainly warrant caution when working with the beam. The beam is much more powerful than it appears at 660 nm due to the eye's reduced sensitivity at that wavelength compared to HeNe 632.8 nm.
About those laser diode bars:
(From: Walter Skrlac (Walter.Skrlac@t-online.de).)
"Bars are a 10 mm wide chip with typically 16 to 24 emitters, each emitter being about 150 microns wide and emitting up to 2 watts of power per emitter. The highest power for solid state laser pumping is 40 watts from a 19 emitter bar. Almost all bars are a single chip, multiple emitter device. I do know that in the beginning days of bars, Siemens produced a 5 watt device consisting of 5 separate 1 watt laser diodes mounted in a row 10 mm long. The individual laser diodes are connected in parallel so you can't switch them individually."
The good news is that this technology is developing very rapidly.
The bad news from our perspective is that there are no really low cost sources, new or surplus, for these diode lasers as far as I know at the present time.
For example, a 1 W 808 nm laser diode is currently (Spring 2000) $300 in the Lasershop catalog.
Actually, it isn't necessarily the diode itself that is so expensive. A 1.5 W 800 nm diode chip goes for about $10 when they are purchased in reasonably large quantities. However, these are only about 0.5 mm on a side and maybe 0.1 mm thick. Mounting means using low temperature solder and flux to bond the chip to a large heat sink and copper strip (for the two connections - no monitor photodiode, that function must be performed externally). The soldering is best done on a hot plate (to raise the temperature of the heat sink and chip to almost the melting point of the solder), with a fine tip iron for the last few degrees. They have an HR and OC side, and a top and bottom, and thus orientation matters. So, if you have access to a surface mount rework station with a stereo microscope, a steady hand, infinite patience, and don't sneeze much (which will blow your chips away to never be found again), you could try your hand at the mounting. I have a couple of these diode chips so once I get up the nerve to try this, I will report on success or failure.
The better way to deal with these laser diodes is to have them already mounted on a heat sink. But now we're talking about $100s for a single unit. But, for a number of reasons, the best type of high power laser diode to get is probably a fiber coupled module. Then you don't have to mess with beam shape issues, the diode is safely tucked away out of harm, and the fiber output can easily be adapted to your favorite crystal shape. Some power is lost in the coupling but it appears as though the specs I've seen are similar for the bare diode assembly and fiber coupled module. Of course, the cost for such a module now appoaches that of a nicely equipped PC. :) For more info, see the section: Anatomy of a Fiber Coupled Laser Diode.
Apparently, laser diode bars/assemblies of much higher power are now available at roughly equivalent prices if you multiply $10/W by the number of watts (but I don't know for what quantities these prices apply). Check out Industrial Microphotonics Company as one possible supplier. The IMC Products Page lists a wide variety of really interesting items but unfortunately doesn't have any prices. Bars can be connected in series to ease the power supply requirements enabling them to be driven with lower current at higher voltage (e.g., a 4 bar configuration would use 8 V at 50 A instead of 2 V at 200 A). With individual chips on a common heat sink, this really isn't an option.
Note that most high power diode lasers are near IR - often around 800 nm for pumping DPSS lasers and optical communications. High power visible laser diodes are much less common and usually limited to less than a W at 670 nm. Not that this is terrible. :)
If you have your heart set on one of these for your birthday, all I can suggest at the present time is to keep track of what is available surplus and to check with the manufacturers listed in the chapter: Laser and Parts Sources. If this is for some sort of academic project with a legitimate research objective, you may be able to obtain a cosmetic reject or one that doesn't quite meet specs by persistent pleading with one of the laser diode manufacturers. Or, if you can deal with the bare chips, it may be possible to beg a few from one of the companies that produces DPSS laser systems since they use them by the carload, and when purchased by the carload, the cost goes way down.
Keep in mind that obtaining the diode is only a small part of the problem. These devices are exceeding fussy about drive and cooling - even much more so than the wimpy little laser diodes found in CD players and laser pointers! However, if reasonable precautions are taken and they aren't run near their maximum ratings, actually blowing them out totally isn't that likely.
And, needless to say, at these power levels, your eyes (and flammable objects) don't get a second chance - laser safety must be at the top of your list of priorities.
VCSELs, on the other hand, emit their beam from their top surface (and potentially bottom surface as well). The cavity is formed of a hundred or more layers consisting of mirrors and active laser semiconductor all formed epitaxially on a bulk (inactive) substrate.
This approach provides several very significant technical advantages:
The beam from a typical VCSEL exits from a circular region 5 to 25 um in diameter. Since this is much larger than for the FP laser diode, the divergence of the resulting beam is much lower. And, because it is also circular, no corrections for asymmetry and astigmatism are required - a simple lens should be able to provide excellent collimation.
On the other hand, an entire wafer of VCSELs can be tested as a unit with each device evaluated for lasing threshold and power, and beam shape, quality, and stability, It is possible to form millions of VCSELs on a single wafer as a batch process and then test and evaluate the performance of each one automatically. The entire wafer can be burned in to eliminate infant mortalities and assure higher reliability of the final product. Each device can then be packaged or thrown away based on these findings.
For a general review article, see: "The Ideal Light source for Datanets", K.S. Giboney, L.B. Aronson, B.E. Lemoff, IEEE Spectrum V.35 (2) p. 43, Feb 1998.
In particular:
See the section: On-Line Introduction to Lasers for additional CORD LEOT modules and other courses relevant to the theory, construction, and power supplies for these and other types of lasers.
The following modules would be of particular interest for diode lasers (all in .pdf format):
EXP01 Emission and Absorption EXP03 Fabry Perot Resonator EXP04 Diode Laser EXP14 Erbium doped Fibre Amplifier EXP20 Laser safety EXP27 Bar Code Reader
To find typical laser diode characteristics and identify laser diode companies, check out the Thorlabs Laser Diode Scout Search Facility. The NASA Langley Photonics Group maintains a Laser Diode Manufacturers Database.
The divergence angle (half of total), Theta, (in degrees) is given by:
Wavelength * 720
Theta = -------------------------
pi * pi * Beam Diameter
At a wavelength of 670 nm, this works out to about 48 x 16 degrees for a
1 um x 3 um emitter and 48 x .48 degrees for a 1 x 100 um emitter (compared
to around .05 degrees for a 1 mm diameter beam from a 632.8 um helium-neon
laser).Note that since at least one of the dimensions of the end-facet is close to the wavelength that the laser diode emits - it may even be smaller - this simple equation is not very precise but typical low power laser diodes do produce beams with a divergence of around 10 x 30 degrees.
The divergence specification for laser diodes is measured to the half power points. T full width at the 10% level may be more like 70 or 80 degrees than the 30 degrees in the specifications.
There are ways of correcting for all of these artifacts with a single special lens close to the laser diode itself. For example, Blue Sky Research offers combined laser diodes and microlenses which they claim perform as well as larger more expensive diode laser modules using various discrete lenses and prisms to implement the beam correction.
Note that VCSEL (Vertical Cavity Surface Emitting Laser diodes) need not suffer from astigmatism and/or an elliptical beam profile since their emitting aperture can be made to be perfectly symmetrical. I would also expect them not to need to be polarized for this reason as well. See the section: Vertical Cavity Surface Emitting Laser Diodes (VCSELs).
At Philips we used three difference techniques to measure astigmatism in laser diodes:
An alternative technique, apparently used in many optical pickups, is to pass the beam through a thick optical plate having parallel sides at an angle (actually combined with the 45 degree beam splitter mirror when used for this application). This component has a very significant astigmatic effect whose magnitude is easily controlled by selecting the thickness or adjusting the angle of the plate. In the optical pickup, it is used to add astigmatism for the focusing servo but can just as easily be used to eliminate it. See the document: Notes on the Troubleshooting and Repair of Compact Disc Players and CDROM Drives for more information on optical pickup characteristics.
Whatever type of external optics are added, take care that significant power isn't reflected back into the laser diode itself. This can destabilize the lasing process as well as fooling the built-in photodiode into thinking the output power is higher than it really is causing the optical feedback circuit to reduce it.
Some additional comments are provided below:
(Portions from: Mark W. Lund (lundm@physc1.byu.edu).)
A simple short focal length lens will collimate the beam. However, laser diodes tend to be astigmatic which means that you will have one axis collimated at a different focus than the other. A typical value for this astigmatism is 40 microns. A cylindrical lens in addition to the spherical collimating lens or a special lens designed for this purpose can correct this but may not be needed for non-critical applications.
Any camera lens will be able to produce a reasonably well collimated beam (subject to the astigmatism mentioned above). Put the laser diode at the focal point of the lens. If you want the type of narrow beam produced by a HeNe laser, you need a short focal length lens, such as a microscope objective. A good compromise between cheap and short focal length would be an old disk camera lens. These cameras can be found at thrift shops, garage or yard sales, and flea markets for a couple dollars or less.
The longer the focal length the larger your beam will be, but the less effect the astigmatism will have. The diameter of the beam will be the size of the aperture of the lens (in which case you are throwing away light) or the size of the beam at the distance of one focal length, whichever is less.
(From: Steve Nosko (q10706@email.mot.com).)
One thing to note is that the laser diode actually has two apparent point sources. One for the wide axis of the beam and another for the narrow axis. This means that the lens must be more like two crossed cylindrical lenses with different focal lengths. There are different types of laser diodes with varying degrees of this so that some are easier to to design lenses for. There probably are types, by now, where there aren't two.
I think of it like this (right or wrong). The astigmatism has two components. One is the difference in divergence between the two axes. I think this can be even if there is ONLY one apparent point source. It is just a point source with an oval aperture letting light through. The other is the different apparent point sources for the two axes.
I have tested a PS106 which is a 650 nm, 7 mW Circulaser(tm). The beam is indeed nearly perfectly circular with a divergence of about 8 degrees FWHM - less than that of the lower divergence axis of the typical bare laser diode.
However, this general rule appears not to apply for all laser diodes including those in many common diode laser modules and even cheap ($9.95) laser pointers. These are now being used routinely for experiments in interferometry and even holography. While their stability over time (e.g., wavelength drift and susceptibility to mode hopping) - is probably less than stellar, over the short term, coherence lengths of 20 cm or more are not unusual. This is similar to that of a typical helium-neon laser.
For more on applications that may benefit from long coherence length diode lasers, see the sections: Interferometers Using Inexpensive Laser DiodesCan I Use the Pickup from a CD Player or CDROM Drive for Interferometry?. Also see the section: Holography Using Cheap Diode Lasers.
It depends on the laser diode, the power supply that is used, and the external optical feedback into the diode laser. With a single longitudinal mode diode, without external optical feedback, and a current noise of less than 1 uA RMS in a 1 MHz bandwidth, you can get linewidths of 10 MHz for a coherence time of nanoseconds. With optical feedback the linewidth can collapse to a few Hz or explode to several terahertz, depending on its intensity and the delay time between the light leaving the diode and returning to it.
The wavelength shift for 808 nm diodes is generally around 2.5 nm (+/- 0.2 or 0.3 nm) per 10 °C, with the wavelength shift to the red with increasing temperature.
(From: Lynn Strickland (stricks760@earthlink.net).)
It really depends on the laser (i.e., manufacturer) and temperature range you are talking about. A good rule of thumb is 0.3 nm per °C over the operating temperature range of the device (About 30 GHz per °C). That's the average slope of the curve though - it includes mode hops. If you're operating at a mode hop, you can get a lot more change than 30 GHz with a 1 °C temperature change. If you are between mode hops, it can be much less.
Mode hops can be a moving target too. Optical feedback can cause them (even minute amounts). Or, you can operate at a specific temperature where there are no mode hops today, but next week it might mode hop at that temperature.
Note that you can only go so far if you want to use temperature to reduce the wavelength. Even if you got the electronics to work under frigid conditions, there is a minimum laser wavelength you can get from a particular diode laser chip. I'm not a physicist, but it has to do with the bandgap of the materials used. What you would get, as you cooled the thing, is lower and lower threshold current, lower operating current, and longer lifetime.
(From: Richard Alexander (pooua@aol.com).)
Back in the old days, about 15 years ago, the only way to get visible light from a laser diode was by using cryogenic cooling. My textbooks from my laser degree program only knows of this type of visible laser diode (they were written in the early '80s). The first room temperature visible laser diode was invented about 1991; I still have a "Radio-Electronics" issue mentioning it.
(From: Flavio Spedalieri (flavios@ihug.com.au).)
All laser diodes have a tolerance when it comes to wavelength, these tolerances can be as high as +/- 10 nm.
The wavelength tolerances are due to thermal effects, and current. As the diode heats up, the wavelength will change 0.3 nm/°C. and results in mode-hopping.
However, neither of these devices is designed to be modulated at any more than a couple of Hz (if that) due to the heavy internal filtering to protect the laser diode from power spikes. Therefore, they are generally unsuitable for laser communications applications (though some laser pointers are so cheaply designed that such protection may be absent entirely). See the section: The Benefits of Cheap Laser Pointers for Modulation.
Common visible laser diodes have a maximum optical output power of 3 to 5 mW. Due to the sensitivity curve of the human eye, a wavelength of 635 nm appears at least 4 times brighter than an equivalent power level at 670 nm. Thus, shorter wavelength laser diodes will be best where maximum visibility is important. However, these are currently much more expensive - but this will change as DVD technology takes off.
Where the use of a diode laser module or laser pointer is suitable for your application, I would highly recommend this over attempting to cobble together something from a bare laser diode and homemade power supply - or even a commercial driver if it isn't explicitly designed for your particular laser diode. It really is all too easy to fry expensive laser diodes through improper drive or handling. Once blown, laser diodes don't even work very well as visible LEDs!
See the chapter: Laser Parts Sources for a number of suppliers of both diode laser modules and laser pointers. In additiona, Don's Klipstein (don@misty.com) maintains a Web page with a List of Suppliers of Inexpensive Lasers. While not exhaustive, it does include some popular distributors and he does strive to keep it reasonably up to date. Some of these companies now sell laser pointers for under $6! Pretty soon, you will be able to find free laser pointers in cereal boxes. :)
However, there is no way to know how reliable or robust an inexpensive laser pointer will be - or if the beam quality is acceptable before purchase. Diode laser modules are generally more expensive and of higher quality (though not always) so they may be a better bet for serious applications. Also consider a helium-neon laser since even the cheapest type is likely to generate a beam with better beam quality than the typical diode laser module or laser pointer. While any Tom, Dick, or Harry, can put together a laser pointer of questionable design from readily available parts and sell it on the Internet, only a handful of companies manufacturer HeNe tubes and their quality is all very high. With a HeNe laser, the tube alone determines most of its characteristics requiring at most a simple lens to collimate or focus the beam. See the chapter: Helium-Neon Lasers for more information.
On those that do have decent regulators, modulation frequency may be limited to a few Hz to a few hundred Hz depending on design and the actual output power may be more of a triangular wave shape due to the soft start (ramp up, ramp down) turn on, turn off behavior.
(From: John, K3PGP (k3pgp@alltel.net).)
The speed issue was true of many early (and pricey!) laser pointers which used a feedback power regulator. The capacitors and the feedback tended to reduce the speed at which the laser could be turned on and off.
Now that the price has fallen everyone is competing to make them even cheaper. What this means is that most laser pointers today have NO power regulator at all. What I've been finding is a laser diode, resistor, switch, and two 1.5 volt batteries in series. Laser pointers like these can be modulated up into the hundreds of Mhz as there is nothing to interfere with the speed at which the laser can be turned on and off.
Of course you stand the risk of easily damaging the diode in laser pointers like these with an overvoltage, spike, or static electricity if you don't use some common sense and are not careful when bringing wires out and hooking the laser pen to external circuitry.
Since we are dealing with a wide variety of styles and manufacturers, there will be some differences. For instance I've seen a few that have no power regulator, just a resistor to the 3 volt battery supply, BUT have an electrolytic capacitor across the diode. It was necessary to remove the capacitor to allow the laser to be switched at high speed.
As far as I know, CDRH approval will not be granted for any device of this type over 5 mW actual beam power since their classification would then need to be IIIb. So, don't expect to find a laser diode with an actual output power of 30 mW in anything like a laser pointer! Frankly, I don't understand how laser pointers with an output above 1 mW gain approval in any case. The 670 nm pointers especially (since they APPEAR less bright) represent a definite hazard to vision at close range. Do not underestimate the stupidity of some people who totally ignore all the safety warnings - "Wow, look at these cool afterimages." - and then wonder why their vision never quite returns to normal.
Another popular 'specification' is how far away the laser pointer is visible. What the seller is probably actually referring to is the distance that their Marketing department *thinks* the beam should be visible so long as this value is greater than that of their competition. :-)
Seriously, who knows? There is no standards organization overseeing these ratings. It could be the maximum distance to the screen that the beam is visible:
Laser pointer marketers don't appear to have discovered (3) as yet since the number would be extremely impressive - being in the many miles range!
A common red laser pointer contains the following components:
See the section: Basic Characteristics, Structure, Safety, Common Types
It may be possible to determine if your laser pointer uses optical feedback for power regulation: Shine the pointer at a mirror so that the beam returns precisely to its source. Where feedback is used, the output may dim a bit due to the reflected light hitting the monitor photodiode. This is fairly safe in that the laser diode can only be reduced in brightness unless there happens to be destructive interference right at the photodiode - but this is not likely unless you really work at it (with additional optics).
See the sections: Power Regulators in Laser Pointers and: Laser Diode Driver from Cheap Laser Pointer (LP-LD1).
See the section: Beam Characteristics, Correction, Comparison with Other Lasers
See the section: Laser Pointers that Produce Multiple Patterns
This type can be easily recognized because there will be a teeny-tiny replica of its pattern visible by looking closely at the beam aperture.
Also see the section: Pattern Generation Using Conventional Optics.
HOEs can be recognized by looking at them in normal lighting. What you will see is: Absolutely Nothing. Or, at most, a dirty smudge, but no resemblance to what results when used with the laser pointer.
For more info and suppliers, see the sections starting with: Diffractive Pattern Generating Optics.
Currently, nearly all green laser pointers are based on Diode Pumped Solid State Frequency Doubled (DPSSFD) laser technology. They are not just red laser pointers with a different laser diode or green lens! (See the section: Diode Pumped Solid State Lasers.)
The exceptions are older models using green helium-neon (HeNe) lasers. I bet you didn't know HeNe lasers came in green, huh? :) These had power outputs of less than 1 mW and are now about as common as raw dinosaur eggs. (See the section: HeNe Tubes of a Different Color if you are curious.)
The wavelength of these DPSSFD lasers is 532 nm based on the intracavity frequency doubling of a Nd:YAG or Nd:YVO4 chip using a KTP crystal inside the laser cavity. Their output may either be CW or quasi-CW. (You can tell which you have by moving the spot rapidly across a screen - the trace from the quasi-CW type will break into discrete spots.) There is no functional advantage to the pulsed system (it's actually less desirable) but it allows the manufacturer to use a cheaper laser diode without thermal management (e.g., active cooling) and is more efficient (desirable for battery life for these current hogs).
Visibility of these green pointers is about 4 to 5 times that of 635 nm diode lasers or 632.8 nm red HeNe lasers. These are in turn appear 6 or 7 times brighter than the older 670 nm laser diode based laser pointers. The maximum legal green laser pointer power is still only 5 mW but this would be equivalent in brightness to something like a 150 mW 670 nm device! And, sellers of these things don't let you forget it! :)
Battery life of any green pointer is likely to be much worse than that of the simpler red variety though for actual uses as a *pointer* (what a concept!), it probably doesn't matter all that much. The quasi-CW variety may somewhat better in this regard. This is because by its pulsed nature, the peak output can be greater (although the average may be the same) and at certain repetition rates (just under the flicker fusion frequency of human vision), they will appear slightly brighter for a given input power. Or, conversely, the power consumption can be lower for the same perceived brightness.
Note that since there is no real control of temperature, power output may change significantly (up or down or both) if the pointer is kept on for an extended period of time. Usually, since pointers are really intended to be used for brief periods of time for pointing at something, if any optimization was done, the manufacturer would attempt to select the laser diode wavelength to match the Nd:YVO4 absorption band when the components are cool. As the laser diode heats up, its wavelength increases (about 0.3 nm/°C) and drifts away from the optimal value. However, if the wavelength was low to begin with, the power would increase as the wavelength moved toward the peak absorption for the crystal and then decrease if it went far enough.
With the much higher price (currently about $300) for green pointers, make sure you get a decent written warranty. The additional complexity and more delicate nature of the individual components means that reliability and robustness may not be as good as for their red cousins (to the extent that these are reliable and robust!). Even a failed switch just out of warranty (assuming there is a warranty that will be honored in the first place!), can render a $300 pointer useless since there is often no non-destructive way of getting inside to repair it. (And, I've heard that the switches they use on these things are often not adequately rated for the much higher current green laser pointers use compared to red ones.)
For information on DPSSFD laser construction, see the section: Anatomy of a 3 mW Green DPSSFD Laser.
And, what about those other colors? As a practical matter, there isn't much need for anything beyond green since its wavelength (532 nm) is near the peak (555 nm) of the human eye's response curve. However, to impress those high flying corporate executives, blue might be cool - but expect to spend a $2,000 for one using DPSSFD technology that isn't as bright as a $5 red pointer. Violet (really hard to see) laser pointers could be built using the (currently very expensive) Nichia violet laser diodes but they would be even less useful.
The simple answer is: it all depends. There can be variability in any type of product. While the desired output of a laser pointer and collimated diode laser module is similar, how fussy the end-user is and how one gets there may not be:
There is also another difference between the two which relates to output power:
These aren't likely to be in the same league as the $300 diode laser modules from Edmund Scientific or even $100 units from other sources which will meet or exceed all specifications and have protection against all reasonable abuse, for the price, they can't be beat!
With respect to specifications:
CAUTION: Some diode laser modules are current controlled using optical feedback but expect a regulated DC power supply input. With these, the output will continue to increase more or less linearly as the input is cranked up until the point at which the smoke comes out :-(.
I know that in your fantasies, you have dreamed about the possibility of creating a burning laser or Star Wars style light saber from a laser pointer. Unfortunately, neither of these is even possible theoretically. The best you could ever hope for would be to obtain at most 5 mW from a device currently outputting 2 or 3 mW.
While it might be feasible to increase the current to the laser diode, unless you know its specifications AND have an accurate laser power meter (mucho $$$), there is no way of knowing when to quit. Above their rated maximum optical power, laser diodes turn into DELDs (Dark Emitting Laser Diodes) or expensive LEDs. Exceed this rating for even a microsecond and your whimpy 3 mW output may be boosted to precisely 0.0 mW. This is called Catastrophic Optical Damage (COD) to the microscopic end-facets of the laser diode. There can be also be thermal runaway problems or a combination of both of these depending on design - or lack thereof. However, if you have a bag of these gadgets and are willing to blow a few, here are some guidelines:
Where there is an internal regulator and adjustment pot, turning it may increases the brightness initially. However, as the laser diode heats up over a few seconds or minutes, its output with respect to current decreases and the regulator will keep increasing the current to compensate - a runaway condition which can also result in damage or death to the laser diode. A large heat sink, active (e.g., Peltier or heat-pipe) cooling, or dunking in liquid nitrogen may help if you are really determined to get every last photon from your laser pointer or diode laser module! :)
However, another risk, is that after having painstakingly set the current or resistance or whatever for a brighter output, the next time you turn it on, the laser diode may blow! The reason is that when cold, as noted above, the optical output of a laser diode is greater for a given current and may exceed what the laser diode facets can tolerate even if it was well within safe limits with the laser diode warmed up. With no optical feedback, there is no protection against this possibility
This really IS like playing Russian Roulette and my serious recommendation would be to leave well enough alone. Save for a more powerful unit or even just a 635 nm laser pointer if your current model is 670 nm (which will appear at least 5 times brighter for the same output power).
If you do insist on modifying the circuitry, use an antistatic wrist strap, grounded temperature controlled soldering iron, and the proper desoldering equipment (if needed). At least then, you'll know that it was more likely the changes to the circuit that blew out the laser diode, not your rework technique. :)
Also see the section: Determining Characteristics and Testing of Laser Diodes and those starting with: Laser Diode Life, Damage Mechanisms, COD and ASE, Drive, Cooling.
The problem with using NiCd or NiMH cells to replace Alkaline types is that since the voltage is lower (1.2 V/cell versus 1.5 V/cell when fresh), the output may not be as bright if the pointer doesn't include decent regulation or its compliance range is inadequate. Thus, it will be necessary to adjust or change whatever is used for current control in your pointer so it provides the proper current to the laser diode at the lower operating voltage of the rechargeable batteries. Note, however, that since the A-hr capacity of rechargeables is less than that of Alkalines, lasing time will be reduced if they are used. (This is somewhat compensated by the flatter discharge curve of NiCds and NiMH cells and your mileage may vary.) Of course, you risk blowing the circuitry and/or laser diode should you then install Alkalines, so you may not be able to easily go back to them. As with the other comments on modifications to laser pointers, this is quite risky both in terms of possible damage to the laser diode as well as being able to make any modifications to the teeny tiny circuit board if needed.
I've have heard of people (apparently with money to burn), successfully doing this with a green ($$$) laser pointer. They changed the value of the resistor used to set the laser diode current and were able to get slightly more power at the same time (expected life unknown). (Interestingly, at the original power, the beam was TEM00; with increased power, it became multimode.)
One way to tell which effect is causing the change in output power is to measure the laser diode current: If it drops with the reflection, the cause is likely the simple optical feedback mechanism. If on the other hand it increases, then laser instability is likely. Also see the section: Causes of Laser Pointer Output Power Changing When Directed at a Mirror.
Even if the photodiode sensitivity is the cause, several factors conspire against this being a viable technique in general (though it may work with specific devices):
And, if it is actually a lasing interference effect, good luck succeeding in getting anything to be repeatable or stable unless you have a granite block or sand-box holography setup. :)
If you still insist on experimenting, be aware that while this appears to be safe for the laser diode, there is no way of knowing for sure without tests. There could be funny resonances in the driver that will blow your laser diode at certain frequencies! And, if the effect is due to lasing instability, the regulator may attempt to boost the current to compensate resulting in possible overheating of the laser diode, driver, or both.
However, having just tried the reflect-off-the-mirror experiment with no results whatsoever, I'm inclined to write this off as something that might be useful for optical modulation or any other purpose except with certain select pointers or modules. Oh well, it was a good idea while it lasted...
One way to tell which effect is causing the change in output power is to measure the laser diode current: If it drops with the reflection, the cause is likely the simple optical feedback mechanism. If on the other hand it increases, then laser instability is likely.
It does seem that relatively low reflected power back to the laser diode can affect lasing. This has been used to advantage in narrowing the line width of common laser diodes with an external cavity. See, for example, Patent #4,907,237: Optical Feedback Locking of Semiconductor Lasers.
CAUTION: While I've never heard of any damage resulting from these sorts of experiments with common laser pointers or diode laser modules, anything that affects the lasing stability and monitor photodiode feedback could result in excessive current and damage to the driver, laser diode, or both.
(From: John, K3PGP (k3pgpalltel.net).)
This is pretty much my findings here also.
However, since laser pens seem to be built as cheaply as possible there are NO standards! What works with one may not work with another. This has caused me untold grief when trying to discuss most anything about laser pens!
I have a few laser pens here that go nuts when you aim them at a mirror. With some pointers the mirror has to be precisely aligned much the same as the mirrors at the ends of the laser cavity itself. With others the alignment isn't as critical. These same pens seem to be unaffected by other light sources shining back into the laser including light from another laser pen with the same approximate wavelength.
I think the important fact to those those units that were affected is whether or not the incoming radiation was precisely the same frequency as the oscillation in the laser cavity. When this experiment is set up with a pen that is sensitive to this effect, EVERYTHING affects the setup, even the slightest vibration which makes sense (to me anyway!). It kind of reminds me of the Michelson Interferometer or a holographic setup. I assume this interference effect is the same effect noticed with many HeNe lasers where no power sensing diode is involved.
(From: Sam.)
That would seem to confirm the hypothesis that interference with the lasing process is taking place, at least for those cases. I'm surprised they would be so sensitive.
(From: John.)
These pens seem to be somewhat rare though as most of the laser pens that I have don't seem to care what you shine back at them. Since laser pens differ so widely from one manufacturer to the next and even between identical model numbers from the same manufacturer I'm not sure if the differences are being caused by the use of different laser diodes or perhaps this effect is somewhat critical as to the amount of current passing through the laser diode or something else?
(From: Sam.)
Conceivably, the sensitive laser diodes are being operated on the verge of mode hopping or something like that but I'm more inclined to believe it is just a sample to sample variation or laser diode model dependent.
(From: John.)
When trying this experiment with several different HeNe lasers I've also noticed that some are effected to a much larger extent than others. I'm not sure why this is. Maybe it has something to due with the gas mixture, the pressure, the current passing through the tube, or what else?
(From: Sam.)
Also mirror reflectivity and curvature. The gas mixture, pressure, and current are probably less of an issue as long as it is running somewhere around the correct conditions.
When you reflect a beam back into a HeNe laser, it's only .5 to 2 percent of the strength of the output beam and order of .01 percent of the strength of the circulating photon flux inside the tube unless the external mirror is very close to being parallel to the output mirror. Then, there will be multiple bounces and much of the light makes it back to the cavity... Hmmm. The distance also matters due to interference effects and the curvature of the mirrors affect the shape of the wavefront. Possibly HeNe lasers with close to planar mirrors are more sensitive to this. However, just the light bouncing back and forth and interfering with itself outside the cavity can confuse the observations. What a mess. :)
For red laser pointers, note that some/many/most of the newest and cheapest imports may not even use a packaged laser diode - the bare chip is attached directly to a metal header next to the lens. I wouldn't be too optimistic about repair or reuse of one of those.
The deconstruction process for a typical green (DPSSFD) laser pointer - a much more complex device than the red variety - is shown in the Laser Equipment Gallery (Version 1.47 or higher) under "Dissection of Green Laser Pointer".
The quick answer is a definite maybe IFF the module or pointer can be opened for examination or repair. If it is a potted block, forget it.
The chances of success are much greater for a diode laser module since it is likely to have a proper laser diode driver with current regulation and optical feedback. These are typically so over-designed that while applying excessive voltage (well, within reason, not 120 VAC to a 5 VDC module!) or incorrect polarity may blow some components, chances are that the laser diode itself won't feel a thing and will survive unharmed.
Assuming you can get inside, repair should be possible. And, even if you end up having to replace a 5 mW laser diode (for, perhaps $10), you have made out well. High quality diode laser modules go for anywhere from $50 to $300.
However, depending on design, a laser pointer could be totally destroyed by even modest overvoltage (say 5 V instead of 3 V from 2 AAA batteries) or reverse polarity. Some of these don't have anything more than a resistor for current limiting. So the laser diode could very well have been damaged or turned into a DELD (Dark Emitting Laser Diode) or expensive LED. All you may end up with is a nice (or not so nice) case. :-( Of course that in itself may come in handy to package your own laser diode and driver - ignoring what was originally there.
There are at least 3 surfaces that can collect dirt - the two sides of the lens (it is probably a single element) and the exterior of the laser diode window. However, in all likelihood, only the exposed surface of the lens will need cleaning.
First, gently blow out any dust or dirt which may have collected inside the lens assembly. A photographic type of air bulb is fine but be extremely careful using any kind of compressed air source. Next, clean the lens itself. It may be made of plastic, so don't use strong solvents. There are special cleaners, but isopropyl alcohol usually is all that is needed. 91% medicinal should be fine, pure isopropyl is better. Avoid rubbing alcohol especially if it contains any additives.
Lens tissue is best, Q-tips (cotton swabs) will work. They should be wet but not dripping. Be gentle - the plastic (probably) or glass and particularly the anti-reflection coating on lens is soft. Wipe in one direction only - do not rub. Also, do not dip the tissue or swab back into the bottle of alcohol after cleaning the optics as this may contaminate it.
The alcohol should be all you need in most cases but some types of dirt (e.g., sugar) will respond better to just plain water.
The inside surface of the lens, any other optics, and the window of the laser diode can be cleaned in a similar manner should this be necessary. Usually, it is not.
Do NOT use strong solvents (which may attack plastic lenses) or anything with abrasives - you will destroy the optics surfaces.
CAUTION: Lenses or other optical components may be bonded or mounted using adhesives that are soluble in alcohol or acetone (but probably not water). Don't make the mistake I made and use too much solvent. I still have not found the tiny collimating lens that popped out of a laser diode module and is now likely lost forever to the basement floor. Crunch :-(.
If the camera is focused at infinity, a collimated laser beam will be focused to a tiny spot on the image sensor. Whether damage will occur depends on many factors including the type of image sensor, quality and focus of the optics, and how long the beam is held in one place. A 1 mW beam (much less than what some laser pointers produce) is roughly equivalent to the brightness of the noonday Sun at the equator on a clear day and when focused to a 10 um spot (the approximate size of one pixel on a typical video camera) it becomes 10,000 times more intense! Needless to say, pointing a camera at the Sun is generally not recommended.
WARNING: Class IV laser products - the output from the fiber will destroy vision and set things on fire!
Fiber coupled laser diodes are much easier to use than bare laser diodes even though they still need an external high current driver. (Of course, they are also much more expensive.) Aside from the physical protection provided by the packaging, the output of the fiber is a nice circular beam with modest divergence (about 16 degrees full angle) which doesn't require correction for astigmatism or asymmetry. Thus, simple lenses can be used for collimation and focusing. I've used a good sample of the 808 nm version of this laser to pump the guts from a green (DPSS) laser pointer just by holding the end of the fiber next to the Nd:YVO4 crystal. After adding a coupling with a GRIN lens for focusing, I can get a few mW of green light from it though I suspect the diameter of the pump beam is still larger than optimal.
The unit I dissected is typical of 0.5 to 1.5 W fiber coupled diode lasers. Refer to Typical 1 Watt Fiber Coupled Diode Laser Showing Interior Construction while reading the following description.
The overall package is 1.5"(L) x 0.75"(W) x 0.5"(H) and is made of a block of gold plated brass with a milled cavity. There are red and black wires for power and a single-mode fiber with SMA 905 connector for beam delivery.
After prying off the Epoxied lid, the following can be seen:
This is one reason why most applications of laser diodes include optical sensing to regulate beam power. The third lead on the laser diode package is connected to an internal optical sensing photodiode used to regulate power output when used in a feedback circuit which controls your current. This is very important to achieve any sort of stable long term operation.
You can easily destroy a laser diode by exceeding the safe current even for an instant. It is critical to the life of the laser diode that under no circumstances do you exceed the safe current limit even for a microsecond!
In addition, as the temperature of the laser diode changes (heats while powered), the current requirement to produce a given optical output increases as well. Without optical feedback if you set the current to be correct once the temperature of the laser diode stabilizes, it will likely blow out instantly the next you turn it on from a cold start!
Laser diodes are also extremely static sensitive, so take appropriate precautions when handling and soldering. Also, do not try to test them with an analog VOM which could on the low ohms scale supply too much current.
It is possible to drive laser diodes with a DC supply and resistor, but unless you know the precise value needed or have a laser power meter at your disposal, you can easily exceed the ratings before you realize it.
You might hear someone bragging "I have driven thousands of laser diodes by just connecting them to a battery and resistor and never have blown any". Sure, right. While it is quite possible that the susceptibility to instant damage due to overcurrent varies with the type of laser diode, unless you know the precise behavior, you must err on the side of caution. Some designers have gone to extremes, however. See the section: Laser Diode Power Supply 2 (RE-LD2) for a design with 5 levels of protection!
For testing, see the section: Testing of Low Power Laser Diodes.
For an actual application, you should use the optical feedback to regulate beam power. You should also use a heat sink if you do not already have the laser diode mounted on one. See the chapter: Laser Diode Power Supplies.
The raw beam from a laser diode is generally wedge shaped - 10 x 30 degrees is a typical divergence. You will need a short focal length convex lens to produce anything approaching a collimated beam. The optics from a dead CD player (even though CD players and CDROM drives use infra-red laser diodes, the optics can likely still be used with visible laser diodes), a low to medium power microscope objective, or even an old disc camera can provide a lens that may be entirely suitable for your needs.
Thus, these devices make truly lousy laser pointers or laser light shows as the emission is just barely visible in subdued light. If you hoped for a Star Wars type laser beam, better go hunting for a 25 W argon laser. :-)
However, for data or voice communications, various kinds of scanning or sensing, and electro-optic applications where visibility is not needed or not desirable, such low cost sources of coherent light are ideal.
Similar types are found in CDROM drives and newer LD (LaserDisc) players. CD-R recorders, Minidisc equipment, magneto-optical, and other writable optical drives including WORM drives, use devices that are similar in appearance and drive requirements but may be capable of somewhat higher maximum power output - as much as 30 mW or more.
Modern laser printers use laser diodes producing anywhere from 5 mW to 50 mW and beyond depending on their resolution and speed (pages per minute). High resolution laser imagers, typesetters, and plotters, may use laser diodes producing 150 mW or more. (However, equipment built before 1985 or so may use helium-neon or even argon lasers rather than diode lasers.)
The laser diode in a laser printer is located inside the scanner unit which is probably a black plastic case about 6 or 8 inches on a side and a couple of inches thick with a motor protruding from the bottom. The laser diode is mounted (along with its driver board, collimating optics, and even possibly a Peltier solid state cooler on some) either near one corner or inside. There should be a laser safety sticker on it as well - but these fall off sometimes!
It is essential that additional precautions are taken if you have a higher power laser diode from equipment of this sort (or don't really know where yours spent its earlier life).
There are now laser diodes (or possibly laser diode arrays) with optical output measured in 10s, even 100s of watts though these will not be what you would call tiny and will probably require buss bars for electrical power and plumbing for cooling!
This Laser Printer Diode Laser Module is from an older unidentified laser printer, laser scanner/duplicator, or similar device. It shows an example of a typical assembly consisting of an IR laser diode, collimating optics, and electronics driver board.
CD player laser diodes are infrared (IR) emitters, usually 780 nm, with a maximum power output of around 5 mW. Their emission will appear very slightly visible and deep red. This is the eye's response to the near-IR radiation but appearing about 10,000 times weaker than the actual beam would be it it's wavelength were centered in the visible part of the spectrum. Despite what the EM spectrum charts show, the eye's response does not drop off to zero at exactly 700 nm - there is decreasing sensitivity which may extend out beyond 820 nm depending on the individual (though some people can't even see the 780 nm). Just realize that the main beam is IR and almost totally invisible. Take care. A collimated 5 mW beam is potentially hazardous to your eyes. Don't be misled into thinking the laser is weak due to the dim appearance of the beam. It is not supposed to be visible at all!
If you don't want to take even the minimal risk of looking into the lens at all, project the beam onto a piece of paper held close to the lens. In a dark room, it should be possible to detect a red spot on the paper when the laser is powered. For any laser more powerful than this or where the beam may be even approximately collimated, viewing the spot on a diffuse surface is the only safe method for checking the beam.
Typical CD laser optics put out about 0.3 to 1 mW at the objective lens though the diodes themselves may be capable of up to 4 or 5 mW depending on type. If you saved the optical components, these may be useful in generating a collimated or focused beam. The aspheric objective lens will be optimized for producing a diffraction limited spot about 1 to 3 mm from its front surface when the optical system is used intact.
The optics may include a collimating lens, diffraction grating (to produce the three beams in a three beam pickup), beam splitter prism or mirror, turning mirror (for horizontally mounted optics), and focusing (objective) lens. Older pickups tend to have larger and more complex sets of optics. Despite the fact that they are mass produced at low cost, these are all very high quality optical assemblies.
However, depending on design, some of the parts may be missing or combined into one component. For example, many Sony pickups do not appear to use a collimating lens. For pickups with a collimating lens, if the objective lens is removed, you should get a more or less parallel main beam and two weaker side beams. Many newer designs have a combined laser diode/photodiode array rather than individual components. Mix and match parts for your needs (if you can get it apart non-destructively). Where there is no collimating lens, the objective lens may be used for this purpose if positioned closer to the laser diode.
For examples of typical optical pickup/optical block designs, see:
The coils around the pickup are used for servo control of focus and tracking by positioning the objective lens to within less than a um (1/25,400 of an inch) of optimal based on the return beam reflected from the CD. See the document: Notes on the Troubleshooting and Repair of Compact Disc Players and CDROM Drives for more information on optical pickup organization and operation.
Typical drive currents are in the 30 to 100 mA range at 1.7 to 2.5 V. However, the power curve is quite non-linear (though perhaps not as extreme as the typical visible laser diode). There is a lasing threshold below which there will be no coherent output (just IR LED emission). For a diode rated at a nominal current of 50 mA (typical for Sony pickups, for example), the threshold current may be 30 mA. This is one reason why most applications of laser diodes include optical sensing (there is a built in photodiode in the same case as the laser emitter) to regulate beam power. You can easily destroy a laser diode by exceeding the safe current even for an instant. It is critical to the life of the laser diode that under no circumstances do you exceed the safe current limit even for a microsecond!
Laser diodes are also supposed to be extremely static sensitive, so use appropriate precautions. Also, do not try to test them with an analog VOM which in particular could on the low ohms scale supply too much current.
It is possible to drive laser diodes with a DC supply and resistor, but unless you know the precise value needed, you can easily exceed the ratings.
For testing, see the section: Testing of Low Power Laser Diodes.
For an actual application, you should use the optical feedback to regulate beam power. You should also use a heat sink if you do not already have the laser diode mounted on one. CD laser diodes are designed for continuous operation. See the chapter: Laser Diode Power Supplies.
For salvaged LDs, poke their legs in anti-static (black) foam as soon as they are free and store in anti-static bags or boxes.
For salvaged LDs, add a shorting wire prior to unsoldering or removal from the circuit board if possible.
Caution: Removing the laser diode from the optical assembly may affect critical optical alignment since it will not be possible to replace it in precisely the same position. This probably doesn't matter for most purposes but is something to keep in mind if you intend to use the device in a manner similar to its original applications. See the section: Reasons to Leave the CD Laser Diode in the Optical Block.
Note that if you have a device from a CD player, CDROM, or other optical drive with 8 or 10 pins, it is a combined laser diode and photodiode array in a single package. You will first have to identify the three connections to the laser diode itself. You should be able to determine this by tracing the wiring - there may even be markings on the circuit board. In many cases, the laser diode is driven by discrete components whereas everything else goes to a preamp IC. Once the pinout of the laser diode is determined, it can be treated in exactly the same way as the more common 3 pin type.
The first step is to identify which pair of terminals are the laser diode and photodiode. Your laser diode package will be configured like one of the following:
LD LD LD LD
+--|>|--o LDC +--|>|--o LDC +--|<|--o LDA +--|<|--o LDA
| | | |
COM o--+ COM o--+ COM o--+ COM o--+
| PD | PD | PD | PD
+--|>|--o PDC +--|<|--o PDA +--|>|--o PDC +--|<|--o PDA
(1) (2) (3) (4)
The most common polarities for low power laser diodes seems to be (2). The COM
terminal will then be connected to a positive supply (+V) relative to LDC and
PDA.
Where you can see both the pins and the inside of the laser diode package, it is easy to identify which pins goes where:
* The connection to the photodiode (PD) will attach via a fine wire to the photodiode chip mounted (probably at a slight angle) deep inside the package.
The following assumes you did not have this luxury:
The photodiode's forward voltage drop will be in the approximately .7 V range compared to 1.7-2.5 V for the laser diode. So, for the test below if you get a forward voltage drop of under a volt, you are on the photodiode leads. If your voltage goes above 3 V, you have the polarity backwards.
CAUTION: Some laser diodes have very low reverse voltage ratings (e.g., 2 V) and will be destroyed by modest reverse voltage. Check your spec sheet. However, the laser diodes found in CD players seem to be happy with 4 or 5 volts applied in reverse. Of course, a shorted or open reading could indicate a defective laser diode or photodiode.
If the laser diode is still connected to its circuitry (probably a printed flex cable), it is likely that the laser diode will have a small capacitor directly across its terminals and the optical sensing photodiode will be connected to a resistor or potentiometer. In particular, this is true of Sony pickups and may help to identify the correct hookup.
R1 100 ohms 1 W
+ o--------/\/\--------+-----------+--------+
| | |
Power supply C2 + _|_ C2 _|_ __|__ LD1
0 to 10 VDC 10uF --- .01uF --- _\_/_ Laser diode
(No overshoot!) - | | |
| | |
- o--------------------+-----------+--------+
If your power supply has a current limiter, set it at 20 or 25 mA to start.
You can always increase it later. If a suitable bench power supply isn't
available, one which can be built for a few dollars and has the needed bells
and whistles is described in the section:
Sam's Laser Diode Test Supply 1.
R2 100 1W
+ o-----------+ +----/\/\------+-----------+--------+
| | | | |
10VDC / ^ | C1 +_|_ C2 _|_ __|__ LD1
Power supply \<----+ R1 10uF --- .01uF --- _\_/_ Laser diode
(No overshoot!) / 100 ohms - | | |
| 2W | | |
- o-----------+--------------------+-----------+--------+
R2 limits the maximum current. If you know the specs for your diode, this
is a good idea (and to protect your power supply as well). You can always
reduce its value if your laser diode requires more than about 85 mA (with
R2 = 100 ohms).
Before attempting to obtain lasing action with either of these circuits, monitor the voltage across what you think is the laser diode as you slowly increase the power supply or potentiometer.
Wavelength Operating Current
---------------------------------------
808 nm 60 - 70 mA
780 nm 45 - 55 mA
670 nm 30 - 35 mA
660 nm 55 - 65 mA
650 nm 65 - 85 mA
640 nm 70 - 90 mA
However, some laser diodes may have an operating current as low as 20 mA and
VCSELs tend to be much lower (but you probably don't have any of those
to play with!).Of course, if you inherited a bag of identical laser diodes and can afford to blow one: (1) I could use a few before you do this :-) and (2) you probably could fairly accurately characterize them by testing one to destruction.
For a current below the lasing threshold for your laser diode, there will be some emission due to simple LED action. As you slowly increase the current, at some point (if the laser diode is good) as you exceed the threshold current, the character of the emission will change dramatically and a very slight increase in laser diode current will result in a significant increase in intensity. Congratulations! The laser diode is lasing.
CAUTION: unless you have a laser power meter, don't push your luck. The maximum safe current may be as little as 5% above the lasing threshold. Go over by 6% and your diode may be history. The exponential power curve seems to be steeper with visible laser diodes but there is no way to be sure without specifications. It is all too easy to convert laser diodes into extremely useless DELDs (Dark Emitting Laser Diodes) or very expensive LEDs.
I have used this approach with laser diodes from dead CD players without difficulty. In the case of many of these, the operating current is printed on a sticker on the optical block, often as a 3 digit number representing the current in 10ths of mAs. Typical values are 35 to 60 mA (350 to 600). Sony pickups typically average around 50 mA. Without this information, the best you can do is to estimate when it is lasing at the proper intensity by comparing the brightness of the 'red dot' one sees by looking into the lens from a safe distance at an oblique angle. However, this is not very reliable as the optical power at the objective lens depends on the particular CD player.
WARNING: With multiple WATTs of output power, particularly for high power IR laser diodes, both eye safety and even possible heat/fire damage to materials must be taken seriously. NEVER look directly toward the output end of the laser diode unless there is no chance of any power being applied to it (even from residual capacitor charge). Direct the output in such a way that it isn't possible to for any eyeballs to intercept the beam or specular reflections. If the beam isn't focused, the heat/fire damage risk is minimal but something to take into consideration. Near-IR laser diode output may look weak and whimpy but realize that the actual intensity is 10s of thousands of times stronger than it appears!.
Also, even an output power as low as 10 mW is enough to affect dark materials when focused with even a simple lens. The beam from a 30 mW laser diode will easily melt black electrical tape and put tiny holes in paper and wood surfaces.
The same general testing approach can be followed as with low power devices. If no high quality adjustable laser diode driver is available, I would suggest a very simple rectified transformer with very large filter capacitor bank to minimize ripple. Control this from a Variac. Use a current limiting power resistor of several ohms between the caps and the diode. Depending on the size of the laser diode, anywhere between 1 and 10 A may be required. Put a modest load across its output to discharge the filter caps after power off. For up to 2 A, I've used a 16 VAC, 5 A power transformer, bridge rectifier, about 20,000 uF filter capacitor, an 8 ohm, 50 W power resistor, and a 100 ohm, 5 W load. The reason I suggest using such a simple power supply is that it is inherently free of overshoot on power cycling (which can't be said in general for active regulators unless specifically addressed in the design).
Note that these high power laser diodes usually don't have monitor photodiodes for optical feedback - output is determined via current and temperature control. For the purposes of testing, if you have a TEC (Thermo-Electric Cooler), set it for around 20 °C. If you don't have a TEC, mount the laser diode package on a large heat sink (with forced air cooling if necessary) to minimize temperature rise. As long as the laser diode package itself remains cool or just warm to the touch, it will be fine.
CAUTION: Change connections - including any meters - only with power OFF and the filter caps of the power supply fully discharged. Make sure the output of the laser diode is pointed safely away from you but don't put anything right up against the output facet or window - at these power levels, it may get toasted, especially if a dark color.
If you aren't sure, the best thing to do is locate the specs (!!) or trace the circuitry of the driver/controller if available. Else, it is possible to determine the pinout experimentally with little risk:
On laser diode packages with multiple pins (e.g., TO3), there are many more combinations to check but each can be tested in a similar manner. If you have the driver/controller, tracing its circuitry can greatly narrow the possibilities.
The safest way to monitor output power is with a proper laser power meter. An alternative is the IR Detector Circuit. Position its photodiode sensor an inch or so away from the laser diode's output. The beam shape is highly astigmatic - 5 to 10 degrees horizontally but perhaps 40 to 60 degrees vertically. Given the output power of these laser diodes, even with the sensor intercepting only a small part of the beam, the detector circuit may be overwhelmed (or literally smoked) quite quickly.
A very simple way of detecting optical output is to place a piece of black electrical tape close to (but not touching - a millimeter or so) the front of the LD; at power levels of a few tens of mW, spots will be melted in the black absorbing material quite quickly. At higher power levels, white paper will be charred. CAUTION: Don't let either of these touch the facet of the LD; at the very least it will be coated with burnt stuff (the power density is highest there); it may also be permanently damaged.
Size of LD Chip Threshold Current Max Current Max Output Power
-----------------------------------------------------------------------
0.5 x 0.5 mm 0.25 A 1 A 0.5 W
0.5 x 1.0 mm 0.5 A 2 A 1 W
1.0 x 1.0 mm 1.0 A 4 A 3 W
1.0 x 1.5 mm 1.3 A 6 A 6 W
Well, the answer is: maybe if you are willing to sacrifice one.
(From: Bob.)
As a GENERAL rule of thumb and barring infant mortality, ESD, or any other manufacturing defects in the laser diode, proper heat sinking:
So, yes, you can test a diode to failure by slowly increasing the current until failure occurs and take the current level that destroys the diode almost instantly and divide by 3. As far as whether this is an acceptable way to determine the rated current of the diode, the normally acceptable way is to have the manufacturer spec a current. :) Keep in mind that these numbers apply to diode bars and C mounted diodes. Can packages are a little less efficient in coupling heat away from the diode normally, so they may die a little quicker than normal. In that case you may be running at a bit lower than rated current if you divide by 3.
(Portions from: Pyroguy (pyromaniac_guy@hotmail.com).)
Actually you CAN use any old laboratory supply for your diodes if you want. :) It's just very inconvenient. If you are using a lab supply, make sure you adjust the voltage so you do not get too much voltage across the diode or too much current through it, make sure you connect a fast recovery rectifier diode between the anode and cathode of the diode to protect against voltage reversal, and most importantly, ALWAYS do things in the following order:
Having analyzed the circuit in the section: Laser Diode Power Supply 4 (RE-LD4), I then proceeded to try out a variety of typical visible laser diodes. For all the undamaged laser diodes that I tested, leaving SBT open resulted in safe feedback regulated operation at Vcc1 = Vcc2 = 7 V. But, depending on the particular sample's photodiode sensitivity, optical output power varied widely.
While testing, I used a regulated power supply with adjustable current limit. The voltage was set at 7 V and the current limit knob was used to ramp up the input to the driver while monitoring laser diode current and/or feedback voltage from the photodiode. This approach may have prevented damage to a laser diode on more than one occasion.
Sample SBT LD Current LD Power Output
----------------------------------------------------
1 (49) Open 79 mA .3 mW
39K 80 mA .5 mW
12K 82 mA 1.2 mW
2 (H81) Open 104 mA 1.5 mW
3 (H74) Open 80 mA 2.0 mW
4 (21)* Open >150 mA .3 mW
5 (696) Open 67 mA .2 mW
39K 69 mA .4 mW
12K 70 mA 1.0 mW
5.6K 72 mA 2.0 mW
3.3K 74 mA 3.0 mW
2.2K 89 mA 4.0 mW
6 (H32) Open 51 mA .2 mW
39K 52 mA .4 mW
12K 56 mA 1.0 mW
5.6K 60 mA 2.0 mW
3.3K 70 mA 3.0 mW
7 (D) Open 40 mA .6 mW
39K 43 mA 1.0 mW
12K 47 mA 2.0 mW
8.2K 50 mA 3.0 mW
8 (K)* Open 61 mA .1 mW
39K 66 mA .2 mW
12K 83 mA .5 mW
9 (E)* Open >150 mA 0.0 mW
The numbers in () do not mean anything - they were found marked on each sample
and are only used to identify them uniquely.Laser output power was estimated to seven significant digits based on the perceived brightness using my Mark-I eyeballs (with AutoCal(tm) option). :-)
The resistance of SBT (R7) is listed. However, the actual photodiode load is R7||R6 (33.2K) and thus the photodiode current is (Vcc1/2) = 3.5/(R7||R6) when optical feedback is successful in maintaining regulation. Since the photodiode current should be proportional to optical power, you will probably find that my high mileage eyeballs suffer from some slight non-linearity as well. ;-)
I do not have specifications for any of these laser diodes. However, they are typical of the 660 to 670 nm types capable of 3 to 5 mW maximum output power found in readily available diode laser modules and laser pointers.
Samples 1 through 6 were all in a large (9 mm diameter) package while samples 7 through 9 were in a small (6 mm diameter) package. As you will note, for these types of laser diodes, power output does not really correlate with package size. Each was mounted along with a collimating lens (adjustable in some cases) in an aluminum block or cylinder (variety of styles) which also acts as a heat sink.
I suspect that samples 2 and 3 were of similar construction but that this differed from that of samples 1 and 4. Note how sensitive sample 1 is to slight increases in current - dramatic evidence of the risks involved in running these without optical feedback. Samples 7 through 9 also appeared to be similar but I only had one fully operational unit of this type to test so no detailed comparison could be made.
I do not know whether the higher current for sample 2 is due to prior damage or just a normal variation in laser diode power sensitivity.
Samples 4, 8, and 9 (*) had been damaged to varying degrees previously due to running with excessive current. These disasters occurred prior to analyzing the behavior of this laser driver circuit. Sample 9 was absolutely positively beyond a shadow of a doubt totally dead laser-wise behaving like a poor excuse for a visible LED in a cool-looking fancy package. :-)
In the case of samples 5 and 6, I continued to decrease SBT until a distinct jump in laser diode current was required to maintain the voltage across SBT (and thus beam power). For example, with sample 5, the jump from 74 mA to 89 mA may have indicated that losses were building and damage or total failure would have resulted if pushed any further. However, at that point, no changes in laser diode behavior had occurred and all lower power levels ran at the same drive current as before. Note: I do not know if this is a valid approach for checking the limits of a laser diode but it may work for some types.
All of the other (undamaged) laser diodes tested could probably have been pushed to higher output power but without knowing their precise specifications and only using my Mark-I eyeballs for a laser power meter, I chickened out. However, there was definitely headroom above the power levels listed above.
Sorting by noticeable differences is almost useless - later model 40 milliwatt diodes come in the 5 mm package now. You can't tell much by looking at the packages!
My experience has been that lasing threshold current can vary by a factor of 2 (with temperature and this is verified by the Sharp catalog). Threshold current is NOT any sort of reliable indicator - that's why the drive electronics senses actual optical power output!
That's NOT to say that knowing the threshold isn't useful.
Here's my take on it:
I think that once the threshold has been reached, you can push the diode to about 10 percent past that current safely. For bigger diodes, you probably have 20 percent + of cushion.
Let's say I have a diode that snaps to laser mode at 50 mA. I'd drive it to 55 mA and measure the output quickly. I would set my APC to maintain that power level output and go on to the next diode.
For larger diodes, it's common to not even use a feedback photodiode for power sensing. Thats because these diodes have MUCH wider margins between the threshold and the smoke valve release ratings. Let's say I find a 2 lead LD that starts lasing at 400 mA. This diode can probably be pushed an additional 20 to 25 percent and driven with a constant current source.
With no name/unspecified diodes, in my opinion I'd stick with making them lase and holding them at that power output rather than squeezing every last milliwatt from them.
I might loose a few in testing, but I surely would not loose many.
Use a large area PD mounted right on the face of the LD under test. You can use a bias supply and a series resistor. Put your voltmeter across the resistor. As you slowly ramp up the LD current, you will see all hell break loose when observing the power output meter. Above threshold, the LD is fairly efficient and fairly linear (power out versus current above threshold).
As a ball park figure, you can assume that the threshold current is about 10 to 15 percent of maximum power out for the diode although it varies a lot for bigger and for IR diodes. So, trying to operate a LD to maintain 5 percent of it rated output is damn close to impossible because of the nature of the beast.
Again, all figures and numbers quoted widely variable. Don't take them too seriously.
PS: Make sure your LD testing supply is smooth (ramp up) and test it with an LED first!
(See Toshiba Visible Laser Diodes for a specification summary which includes the TOLD9421 and their other devices.)
According to the spec sheet for the TOLD9421 the monitor photodiode (PD) current can vary from .25 to 1.7 mA (at 5 mW) depending on the particular device sample. I started with RSET - the resistor that determines feedback sensitivity - of 50 K ohms and with the function generator disconnected (so that RMOD wouldn't matter). Based on the transfer function of PD current to RSET current, this would result in about 72 uA for the actual PD current - well below the worst case minimum value (at 5 mW) for any sample of the TOLD9421. Using my variable power supply, I ramped the voltage up gradually to assure that the device was going to regulate properly - it leveled off at a fixed but relatively weak output, above threshold but not very bright. After some trials with lower values of RSET, 15K resulted in an estimated output power of about 1 mW.
The next step was to try some modulation. Just attaching the function generator (powered off with its output control all the way down) doubled LD output since the output impedance of 50 ohms cut the value of RSET nearly in half (to 7.5K). Then, powering the function generator and cranking up it's output level allowed me to easily modulate the LD's output between near no light output (way below threshold) and perhaps 4 mW (still all estimated). I only tried frequencies I could see with my very accurately calibrated eyeballs waving from side-to-side - from .1 Hz to a 1000 Hz or so for these initial experiments.
Modulation works by varying the voltage on the input to RMOD and thus the current through it from the ISET pin which is maintained at a constant voltage (about 1.22 V nominal). The PD current is maintained at about 3 times (nominal) of this value.
I could detect no changes in the TOLD9421's behavior (either optical or electrical) so at least so far none of this has resulted in any detectable damage to the laser diode. There has been no increase in threshold or operating current and no measles (spots) in the device's output beam pattern. (For a couple of minutes I thought there had been damage but the spots turned out to be dirt on the LD window.)
CAUTION: For experiments like this with a signal or function generator, make sure that no power or output glitches (as when changing modes) could result in an excessively negative spike or offset which may force too much current through the LD and damage or destroy it. The addition of a reverse biased diode across the modulation input is recommended to prevent excessive negative voltage from appearing there.
Later, I popped in a Blue Sky Research PS106 which is a 7 mW Circulaser(tm) - a 650 nm laser diode with a built-in microlens to correct for beam asymmetry and reduce divergence. Since this device had a less sensitive monitor photodiode, I used an RSET of 39K which would run it at about 2.5 mW (I have a printout of this specific sample's complete electrical and optical characteristics). That worked fine as well though I didn't puch my luck any further (e.g., boosting power or modulation).
The toughest part about testing these was soldering the power supply leads to the NS102. I totally destroyed the first sample attempting to solder to what looked like a pad for the positive power supply input but despite its appearance, solder just wouldn't stick. And in the process, I managed to lift another pad clear of the device. After a total kludge soldering job that looked like it should have worked, there must still have been a problem because upon powering up using my variable voltage power supply with adjustable current limit, while the regulator appeared to be doing something based on the brightness of the LD output, power supply current kept going higher and higher as the input voltage was gradually increased. Eventually, the laser diode developed those dreaded spots and while still lasing, must have lost approximately half of its mirror facet(s) as there is also a large dead area in the beam pattern.
The second attempt was much smoother. Rather than trying to solder to that pad, the positive connection simply went to the common pin of the laser diode. So, wiring is as follows:
I later tested that damaged LD using the iC-Haus WJB driver (see the section: Testing the Toshiba TOLD9421 with the iC-Haus WJB Driver, above). It would still operate stably with an output of a milliwatt or so using optical feedback but about twice the normal current (50 mA) for 5 mW output. Of course, the unsightly blemishes in the beam pattern were still there. :( Interestingly, while determining a resistor value that would work, the current repeatedly spiked to more than 5 times its specified nominal value (pegging my 100 mA meter) for a good fraction of second. However, no further damage to the laser diode appears to have occurred. In fact, output power could still be pushed much higher - perhaps up to 3 mW or more - but then the current was way off scale and I didn't hang around to see what would happen next. :) This is in sharp contrast to the behavior of a laser diode I blew a while back where at a current only slightly above the rated maximum, the conversion to an expensive LED was quite rapid.
This combination is designed to fit entirely inside NVG's machined brass Laser Diode Module Housing which provides the much needed heat sink (the laser diode current would begin to creep up almost immediately due to the small thermal mass of the 5.6 mm laser diode package) and an adjustable collimating/focusing lens. Once assembled, the commercial units are potted in Epoxy and the laser safety sticker is wrapped around the outside. :)
While designed for CW applications, modulation of these drivers may also be possible (but I have not done any testing). See the section: Comments on Some Commercial Drivers and Detectors.
If you are trying to use a video camera or camcorder as an IR detector, confirm its sensitivity to near IR by looking at an active IR remote control through its viewfinder. It may have a built in IR blocking filter which will prevent it from being sensitive to IR. This may be removable.
Component values are not critical. Purchase photodiode sensitive to near IR (750-900 um) or salvage from opto-coupler or photosensor. Dead computer mice, not the furry kind, usually contain IR sensitive photodiodes. For convenience, use a 9V battery for power. Even a weak one will work fine. Construct so that LED does not illuminate the photodiode!
The detected signal may be monitored across the transistor with an oscilloscope.
Vcc (+9 V) o-------+---------+
| |
| \
/ / R3
\ R1 \ 500
/ 3.3K /
\ __|__
| _\_/_ LED1 Visible LED
__|__ |
IR ----> _/_\_ PD1 +--------o Scope monitor point
Sensor | |
Photodiode | B |/ C
+-------| Q1 2N3904
| |\ E
\ |
/ R2 +--------o GND
\ 27K |
/ |
| |
GND o--------+---------+
_|_
-
Note: Some designs combine the laser diode and photodiode into a single package which is then mounted in the optical block. This can still be used for either or both functions as long as you can identify the proper pins.
In some higher performance printers, there may be a Peltier cooler attacted to the back plate of the laser diode. Pretty cool :-) (no pun....).
Note: There are often a pair of adjacent solder pads connected to the laser diode circuitry on the flex cable or circuit board associated with the optical block. When handling the assembly but not actually attempting to power the laser diode, it is a good idea to short these together with a drop of solder using a grounded soldering iron. This will prevent the possibility of ESD damaging the laser diode.
Where the laser diode is to be used as part of a precise optical apparatus for close range sensing or scanning, for example, the entire optical deck (including the stable mounting and sled drive mechanism) may be useful intact. For the typical three-beam pickup (most common), this will provide precise control of beam position: Y (focus), X-coarse (sled drive), X-fine (tracking).
There are several good reasons to leave your CD laser diode installed in the optical block assembly even if you are not going to use it with the objective lens and focus and tracking actuators which were part of the pickup:
Remove the objective (front) lens and its associated coils unless you require them for a short range application. They will likely come off as a unit without too much effort. However, try not to destroy this assembly as you never can tell what might be needed in the future.
Here is the connection diagram for a typical Sony pickup:
_
R1 +---|<|----o A | +----o F+
+-/\/\---o VR | PDA | (
PD1 | ^ +---|<|----o B | ( Focus
+---|<|--+---+----o PD (sense) | PDB > Focus/ ( coil
| +---|<|----o C | data (
| LD1 | PDC | +----o F-
+---|<|--+--------o LD (drive) +---|<|----o D _|
| _|_ | PDD _ +----o T+
| --- C1 +---|<|----o E | (
| | | PDE > Tracking ( Tracking
+--------+--------o G (common) +---|<|----o F _| ( coil
| PDF (
Laser diode assembly | +----o T-
+----------o K (Bias+)
(includes LD/PD and Focus/tracking
flex cable with C, R). Photodiode chip actuators
The laser diode assembly and photodiode chip connections are typically all on
a single flex cable with 10 to 12 conductors. The actuator connections may
also be included or on a separate 4 conductor flex cable. The signals may
be identified on the circuit board to which they attach with designations
similar to those shown above. The signals A,C and B,D are usually shorted
together near the connector as they are always used in pairs. The laser
current test point, if present, will be near the connections for the laser
diode assembly.It is usually possible to identify most of these connections with a strong light and magnifying glass - an patience - by tracing back from the components on the optical block. The locations of the laser diode assembly and photodiode array chip are usually easily identified. Some regulation and/or protection components may also be present.
Note: There are often a pair of solder pads on two adjacent traces. These can be shorted with a glob of solder (use a grounded soldering iron!) which will protect the laser diode from ESD or other damage during handling and testing. This added precaution probably isn't needed but will not hurt. If these pads are shorted, then there is little risk of damaging the laser diode and a multimeter (but do not use a VOM on the X1 ohms range if it has one) can be safely used to identify other component connections and polarity.
See the document: Notes on the Troubleshooting and Repair of Compact Disc Players and CDROM Drives for additional information on construction and testing of optical pickup assemblies and photos of typical optical decks.
If you were to just pop in an IR laser diode in place of a visible one, either it will not work at anywhere near maximum output and/or it may blow instantly.
Some datasheets list expected lifetimes for laser diodes exceeding 100,000 hours - over 12 years of continuous operation. Of course, I trust these about as much as the latest disk drive MTBFs of 1 million hours. :-)
Laser diodes that fail prematurely were either defective to begin with or, their driver circuitry was inadequate, or they experience some 'event' resulting in momentary (greater than a few microseconds) overcurrent. What this means is that with cheap driver electronics such as found in many laser pointers, leaving the thing on continuously may result in much longer life than repeatedly pulsing it.
As noted elsewhere, a weak laser diode is well down on the list of likely causes for CD, LD, MD, and DVD player, as well as laser printer problems.
High power laser diodes may have considerably shorter life expectancies than the 5 mW variety - 10,000 hours or less.
And, high temperature operation can reduce life expectancy, possibly by as much as a factor of 2 for each 10 °C rise above the temperature quoted in the device's specifications. Thus, a laser diode with a quoted life of 10,000 hours at 25 °C, might only last 125 hours at 55 °C. Not that it will actually fail a 125 hours and 1 second, but its maximum output power will be reduced by 50 percent. I expect that there is a wide variation on the extent to which this applies depending on device type, how close it is operated to its specified maximum power, and all sorts of other factors.
Of course, in the grand scheme of things, even LEDs gradually lose brightness with use.
(From: Gregory J. Whaley (gwhaley@tiny.net).)
There is one thing to keep in mind about laser diode lifetimes. The time to failure probability distribution is quite wide, meaning that some laser's lifetime will be significantly less than the 5,000 hour mean, and some will be much, much longer than the mean. Lasers are not like light bulbs where they "wear out" and have a predictable lifetime. The main life limiting factors in a laser diode are related to how many crystal defects are present in the device when it is made. If you are lucky to have a diode with very few defects, then your laser may last nearly forever. If you are not so lucky, it may only last a few hours.
If you don't know the life story of your laser diode, see the section: Testing of Low Power Laser Diodes before you contribute to its demise!
Assuming the device was operating above its threshold current with a nice bright output beam prior to the 'event', some or all of the following may be in evidence:
If you return a damaged laser diode to a driver that uses optical feedback to stabilize output power, the laser diode will likely be destroyed if the circuit increases the drive current to its maximum limit in a futile attempt to achieve the expected output power.
Another tip-off that there is no laser action is that the beam intensity will not increase dramatically as the current is raised (as it would with the positive feedback of an intact laser cavity) and there will be no distinct threshold; output will be pretty much linear with respect to current.
For high power laser diodes such as the type used to pump solid state lasers, the location of facet damage can be very clearly seen in the beam pattern. Since the emitting aperture of these may be 100 um or more, projecting the output onto a white screen using a short focal length lens (e.g., one from a CD player) will yield the distribution of lasing along the aperture. Set up the distance between the lens and screen to be about 40 mm. This will require an LD-to-lens distance of a few mm (for the CD player lens of 4 mm focal length). The projection will then be a line 2 or 3 mm in length. A new/good LD will show a smooth variation in brightness (if visible or through an IR viewer) but a damaged one will have significant variations in brightness as well as places where there is no light at all. A common failure characteristic is to just have the side lobes with nothing in the middle. However, this terminal disease would also be obvious in the unfocussed beam pattern. Such serious damage may even be readily apparent as different color/rough areas on the end facet using a magnifier or low power microscope.
A way to determine if a laser diode is damaged is by shining the uncollimated beam on a white screen and looking at the spread of light intensity - the beam profile.
This method works with all laser diodes where the light is visible (up to a wavelength of about 800 nm), or with a CCD camera or other sensor array, further into the IR - or UV (wishful thinking).
A working laser diode, will produce an elliptical beam, that is brightest in the longitudinal axis, and tapers off in brightness towards the edges. Some may have slight bumps or dips or hints of an interference pattern but their location will usually be relatively symmetric - if one of these features occurs on one side, there will be a similar one on the other.
If you drive a diode at even very slightly above its maximum limit, you will cause permanent damage to the diode over time.
If you take a diode, then drive it with the correct current, the above beam profile will be produced. If you begin to slowly increase current, up to a certain point, the optical output will increase. Continuing to increase the current beyond this upper limit, the appearance of the beam will begin to change, the output will start to decrease, then the beam will have light and dark bands through it - the diode junction and/or mirror facets have now been damaged.
At this point, the diode is still producing coherent radiation, with slightly reduced output power. If you try and collimated this beam, you will end up with a spot that has light and dark areas.
This type of damage is caused by exceeding the limits of the structure of the semiconductor material and is irreversible.
Also see the section: Laser Diode Damage Mechanisms.
"I just connected a bare laser diode to an automobile battery without any other components and it is working just fine. I have never used any ESD precautions. In fact, I have a wool sweater on at this moment and can draw some really juicy sparks from everything I touch."through:
"I have blown several hundred laser diodes and I have been following all the manufacturer's guidelines with respect to ESD protection and drive. I am even using their recommended circuit layout and $4,000 power supplies. Nothing seems to help."Not all laser diodes are created equal and their susceptibility to damage through improper handling or improper drive likely varies widely. Here is a discussion of some of the issues:
(From: Eric Rechner (eric_r@3dm.com).)
"Does anyone have any experience with Hitachi laser diode HL7843MG 5 mW 780nm? I find this diode to be possibly extremely sensitive (ESD??), more so than any other 780nm laser diode. Does anyone know if there are problems with Hitachi MQW type diodes? Are MQW diodes more sensitive to ESD than Double Heterojunction diodes? Does anyone have info on possibly 'bad' or defective lasers out there?"
(From: Jon Elson (jmelson@artsci.wustl.edu).)
Strange. I think I've used some of these.
I hear everybody babbling about extreme static sensitivity on these devices, yet I've never had a failure, and I've been using just the usual minimum precautions with any semiconductor device. I suspect that people may be exceeding the optical power MAXIMUMS on the devices. I've been very conservative on that, since the devices only carry an optical maximum, and don't have that correlated to forward diode current (difficult, because it varies strongly with temperature). I try to run them at a good bit less than rated power, maybe 2-3 mW optical output. I'm using a diode sold by Digi-Key for $19.00, just because it is cheaper than the Panasonic in the 5.4 mm case. I think the manufacturer is NVG or something like that. I've got 10 of them I am working with, designing a closed-loop driver for a photoplotter, which pulses the lasers on and off as fast as 10 uS on, 10 uS off. It is working pretty well now. I included a series resistor (as well as the control transistor), so that if the loop becomes unstable or the sensing diode gets disconnected, it won't fry the laser diode.
(From: Dr. Mark W. Lund (lundm@xray.byu.edu).)
The babbling starts here: You don't have to be a total idiot to blow these things, in fact I have blown a few myself. Identifying the source of the trouble is extremely costly and difficult because it only takes a spike of a few nS to to the damage. I would say that 99.9999% of the time it is the power supply. Either it spikes on turn-on, turn-off, or at random. We used to toast lasers with a $5,000 laser diode power supply that would spike every time you sent certain signals on the IEEE 488 control line. This was a tough one to figure out, I can tell you. In the process we tried to damage one using static to try to get a handle on the sensitivity, but were not able to get a catastrophic failure this way (we may have induced some latent failures, however). Other laser diodes may vary.
(From: Jon Elson (jmelson@artsci.wustl.edu).)
Ah! This is good anecdotal evidence! I've often suspected that there might be more of this going on, and instead of examining the drivers, people just attribute problems to an invisible gremlin! I sure can see how a closed circuit driver can oscillate or overshoot on transients, and there could be a situation where some percentage of drivers will be less stable due to component tolerances. Unless you rigorously test a good batch of your drivers, you could have this sort of thing and not know it. (Of course, any time you put a computer in the loop, especially one that is canned inside an instrument, then the probability of unanticipated gremlins increases dramatically!).
Of course, I was designing a fixed-purpose driver to be used in a specific application, inside an instrument, so I had it easier than the guys designing a lab-quality pulser for who knows what application. So, I could put in a resistor, which will limit current to some 'safe' level, even if the loop is unstable, which it certainly was when I was tuning up my driver.
I DO use generally sound anti-static precautions, almost subconsciously, to protect all semiconductor devices. But, I am aware that I have occasionally, by accident, touched a cable going to the laser diode before I was grounded, and I have never noted a catastrophic failure.
I will have to go through some rigorous life-testing to make sure I'm not causing latent failures, but I've run these diodes for quite a few hours while testing things, and nothing of note has turned up yet.
By babbling, I meant some items in print media, as well as a lot on this and other newsgroups, indicating that if you even touch one lead of a diode laser, it is ABSOLUTELY destroyed, with a probability of 1.000! Obviously not true! Your comments are well reasoned, and indicate real experience. Others have also written that only a huge corporation, with millions in test equipment, could ever make their own laser diode driver. Now, clearly, the nanosecond multi-watt pulsers ARE much more difficult to do right, fast risetimes without overshoot is tricky. But, I did it in my basement with just over $1,000 in test equipment, mostly a decent oscilloscope. I also had the confidence that if I DID blow a few diodes, it wasn't so painful at $19 each.
So, now, I'm babbling!
(From: Eric Rechner (eric_r@3dm.com).)
Just an update on the outcome of my question about Hitachi laser diodes, above. At that time, large numbers of the diodes in question were dying prematurely (we were running at about 80% full power at a temperature between 20 and 30 °C, CW for several weeks in triangulation sensors). Our diode module supplier had the facilities to inspect the laser chips using electron microscopy and apparently found that new diodes exhibited oxidation on the facet. They believed this to be a process problem (contamination) at the manufacturer end. The last I heard, the diode module supplier credited us with replacement lasers - there were about 1000 pieces, but this took a great deal of 'fighting'....
With the active area of the end-facets of some laser diode being as small as 1 x 3 um, it isn't surprising that a little too much power will kill it. The power density of 5 mW through that aperture is 1,666,666,666 W/square meter or 167 kW/cm2! Apparently some types of optical materials when properly processed and undamaged can handle more than this without a problem but GaAlAs or whatever of the laser diode's mirrors isn't one of them. (Some manufacturers specify the emitting aperture of their laser diodes to be much larger - 10 x 60 mm being a typical value. However, these dimensions are inconsistent with their beam divergence which is similar to that of the much smaller aperture. If the actual emitter were that large, power density would drop by a factor of 200 and it would seem that COD would not be a major concern at the same power level.)
However, overall thermal damage is also possible even - or especially - with a laser diode driver using optical feedback. When you turn up the power control, there may initially be higher output. But as the laser diode heats up over a few seconds or minutes, its output with respect to current decreases and the regulator will keep increasing the current to compensate - potentially a runaway condition which can also result in damage or death to the laser diode. A large heat sink, active (e.g., Peltier or heat-pipe) cooling, or dunking in liquid nitrogen may help if you are really determined to get every last photon from your laser diode! :)
Or, where the laser diode is powered from a constant current source and set for a higher output when warmed up, it may blow instantly the next time it is turned on after having been off for a while. The reason: For the same current, the laser diode's optical output is greater when cold and may exceed the COD limits of the its end facets.
In other words, there are many interesting and creative ways to convert a laser diode into a DELD or expensive LED!
(From: Gregory J. Whaley (gwhaley@tiny.net).)
I will assume the effect is Catastrophic Optical Damage (COD) of the facet. This is an interaction between the temperature of the facet and its optical absorption. When the temperature of the facet grows, the absorption can also grow which feeds back positively to the temperature and the temperature "runs away" until it is physically damaged. My understanding is that this is extremely fast, certainly less than a microsecond, probably less than a nanosecond. COD is often cited as the mechanism which makes laser diodes extremely ESD sensitive and the ESD discharges can be quite brief.
Optical damage in a laser diode is a fairly complex phenomenon so it is hard to give time and/or power to damage. But based on my experience I'll give some numbers.
Typical 5 mW telecom laser diodes (1300 or 1550 nm) are really underated as far as optical power goes and they in general can be driven at 2 to 3 times their rated power without any immediate damage though their lifetime may be months instead of tens of years. High power diodes (e.g., 1 W) on the other hand are rated near their maximum optical power. How much higher they can be driven is a function of pulse width and duty cycle. To give some typical numbers at a pulse width of 1 ms and duty cycles of a few percent: A diode may be driven at up to 50 percent higher and at pulse width of about 50 ns; at a duty cycle of 0.1% it may driven at up to 5 - 10 times the rated power.
A diode that has suffered COD is already dead so its ESD sensitivity is a moot point. On the other hand a diode that has been overstressed optically is more ESD sensitive. This effect works in reverse too, i.e., a diode that has undergone an ESD discharge may only be able to handle lower optical power.
I don't think a time for optical damage can be stated without knowing the stress conditions and the type of diodes. A diode stressed at 20 to 50% may not suffer any catastrophic damage at all but just die out gradually - just much faster than normal lifetimes. At about 100% overstress, degradation can be catastrophic, and fairly fast. Even then the diode can generally be operated at the higher powers for quite a while (seconds) before the onset of COD. Once the COD starts it probably is quite brief. I'm not sure about the numbers and figures mentioned (nano - microseconds) may be correct for actual COD to occur.
ASE usually stands for Amplified Spontaneous Emission. It is part of any lasing process, and is just what it sounds like - spontaneous emission (not in the lasing mode) that gets amplified by the gain medium in the cavity. I find it easiest to think of this in terms of phase: The lasing mode will have one well-defined phase, while all the noise (ASE) modes will have some phase shift relative to the lasing mode. ASE is mostly a concern when you are trying to send modulated signals (e.g. bits) with your laser diode. In that case, ASE is essentially a noise source which degrades the signal (or S/N). In most electrically-pumped diodes, ASE is not so much a problem as RIN (Relative Intensity Noise), which can raise the bit error rate by changing the relative levels of the "on" bits.
L-I characteristic for ASE is going to follow the lasing mode for the low part of the current range, but at some point (depending on cavity Q and carrier lifetime), you're going to get spontaneous emission clamping, where the ASE will stop increasing superlinearly. I'm not sure that this is the same as COD, where you should see a sharp decrease in optical power output.
There are a number of good laser physics books which may discuss this - try Sargent, Scully and Lamb ("Laser Physics") or Yariv ("Quantum Electronics").
If you intend to use the laser without the feedback, one has to realize that there are a number of problems. One is that as the temperature goes down, the laser efficiency goes up. This tends to cause the laser diode to destroy itself at lower temperatures while running that same current that was OK at some higher temperature. Generally, if the temperature doesn't vary to much, one can use something as simple as a limiting resistor and not run the laser at its highest output. I once made a burn-in driver for some power lasers that used constant current sources that had no feed back but I had to preheat the diodes to 100 °C before using that high a level of current. The level of current used would have wiped the diodes out at room temperatures.
The hardest part of the whole thing was making the circuit to have controlled levels of current during power on and power off. Most things like op-amps are not specified under these conditions. My first attempt wiped out 10 diodes :-( when I turned the power on.
To run the diodes at there maximum light out safely, requires using the feedback photo diode.
Note that the photodiode is NOT part of the laser diode structure - it sits behind the laser diode in the typical package. So, you can actually test its frequency response with an external modulated light source (like an LED or another laser diode driven by a high speed pulse generator) independent of the laser diode itself. The light doesn't have to pass through the laser diode. Although not terribly clear, the photodiode can be see in the Closeup of a Typical Laser Diode.
(From: Richard Schmitz (optima-prec@postoffice.worldnet.att.net).)
The frequency response of the photo diode (PIN diode) is usually shown in the back of the manufacturers laser diode data book. In the case of Toshiba's visible diodes, the freq. response is shown as flat out to about 10 MHz and it rolls off to -3dB at about 175 MHz. With the newer diodes used in the DVD products, the freq. response seems to be a little better, curves for the TOLD9441 show the response out to 1 GHz, down -3dB. If you need exact details, contact a distributor and get the latest Toshiba data sheets.
K3PGP's Cryogenic Cooling of Laser Diodes Page suggests an experiment to demonstrate these effects. The information is taken from: [1] ("The Laser Cookbook, 88 Practical Projects", Gordon McComb, Tab Books Inc., Blue Ridge Summit, PA., 1988., ISBN: 0-8306-9390-4 (paperback)). However, I seriously question the magnitude of the results (an 850-fold increase in output power). :)
"I have read that cooling semiconductor laser diodes shortens wavelength and greatly increases efficiency some. Does this apply to the 635 nm diodes and what would be the result of super cooling one of these diodes?"
(From: Fred Kung (kung@ccf.nrl.navy.mil).)
One thing you will need to be careful about is that in super cooling a compound semiconductor diode laser, you will eventually take it out of its range of lasing operation (due to dispersion shifting). Dropping the temperature to -50C or so is OK, but don't expect them to work in LN2 or anything very cold unless they're designed for that.
The 0.3 nm/°C figure is good for GaAs quantum well lasers with AlGaAs cladding (which covers most of the commercially available ones), but only around room temperature.
One other thing that may happen if you cool the diode too far is that the thermal mismatch with the epoxy will cause it to physically come loose from its mount. Again, a TE cooler is fine, but don't dump cryogens on the thing.
The use of laser diodes in all sorts of mass produced products (CD, LD, MD, DVD, laser printers, bar code scanners, telecommunications, etc.) has driven down prices for lower power devices, at least.
However, shorter wavelength laser diodes had eluded researchers for many years. (The current crop of green laser pointers are DPSSFD lasers. See the section: Diode Pumped Solid State Lasers. Relatively recently, Nichia Chemical has started sampling and is about to begin commercial production of violet (400 nm, they actually call them blue) laser diodes based on gallium nitride. Other companies including Xerox Corporation have their own blue laser diodes near commercialization. Also see the section: Availability of Green, Blue, and Violet Laser Diodes.
Mid-IR (3 to 25 um) types are also available. These typically use lead salts for the active material, but may require a frigid operating environment while producing only around 100 uW output power. You won't find such devices in consumer electronics - their applications are more likely to be in spectroscopy research. (check out: Laser Components GmbH).
(Portions from: Anthony Cook (a.l.cook@larc.nasa.gov).)
The latest development in far-IR (greater than 3 um) laser diodes is the Quantum Cascade Laser which can produce 100s of mW of light at room temperature and up to a watt or more when cooled to about -100 °F (-73 °C). These operate in the range of 3 to 13 um. They are not commercially available yet (I don't think) but several research groups are doing work in this area:
Some nominally IR wavelengths are indeed very slightly visible. In favorable conditions (mainly isolating from more visible wavelengths) I have seen with my own eyes:
CAUTION: there is no advance warning of having exceeded eye-safe exposure to slightly visible wavelengths normally considered IR. You may permanently toast part of your retinas duplicating the above unless you verify retinal exposure below the Class I laser exposure limit.
I recently got a laser pointer with a wavelength of 660-661 nm or so and (guesstimated) 2 mW of output power.
I discovered that if I shine the beam through one of those dielectric interference bandpass filters, I got some weak beam output at other wavelengths. So, I investigated further.
About (very roughly estimated from standard issue eyeballs) .2 percent of the beam is spurious radiation with a continuous spectrum. I don't yet know well what it does at longer wavelengths, but a majority of the short wavelength side of this is in the few tens of nm below 660 nm. Slight traces exist down to 540 nm. With two 532 nm filters, I could stare into the beam and see a dim point of light. With a 570 nm filter, it was slightly bright to stare into and I could see the beam VERY DIMLY on a wall in a dark room. With a filter around 630 nm, I could easily see the beam on a wall in a dark room. I used my diffraction grating to verify that most of this was continuous spectrum in the passband of the filter.
The spurious radiation takes the same path that the laser radiation does.
With no filter, I could not see any continuous spectrum with my diffraction grating. The laser line was so much stronger.
As for IR lasers? If the spectrum is just a long-shifted version of what my visible laser does, the most visible part of the laser output would be the laser line. Having a wavelength 100 nm closer to visible increases its visibility only by about a factor of 1,000 and the total spurious output was (roughly) 1/1,000 of the laser line output. The wavelength of the bulk of this was nowhere near 100 nm shorter.
Although I can't be sure this would always be the case, the only spectrum components I could see using a diffraction grating with my CD player laser was the laser line at about 800 nm.
I suspect different IR laser diodes may have greatly different ratios of laser and LED output. If the LED output is only a fraction of a percent of the laser output, the visible output would be mainly the slightly visible laser line. If the LED output is equal to a few percent or more of the laser output, then it may be more visible than the laser line.
(From: Kjell Kraakenes (kkraaken@telepost.no).)
I once used 780 nm laser diodes similar to the types used in CD players, and something that puzzled me was that I was able to see some red radiation from these diodes. I used a microscope objective to focus the light on a wall a few meters away, and when properly focused, a red spot was visible to the naked eye. I had a piece of black card board on the wall, and there was no specular reflection. I used an IR viewer of the type sold by Edmund Scientific (Find-R-Scope), and if I looked at the spot with this IR viewer the beam appeared defocused. By adjusting the distance between the laser diode and the microscope objective, the spot (as it appeared through the IR viewer) could be brought to a better focus. The red, visible light was then so much defocused that it was no longer visible to the naked eye. From these observations, I assumed that the spot I saw through the IR viewer was the laser emission at 780 nm, and that the visible light was some weak emission at a shorter wavelength. Because of the chromatic aberrations in the microscope objective these two wavelength could not be expected to be in focus simultaneously. I did not notice whether the distance between the laser diode and the microscope objective was increased or decreased when shifting between the focus of the visible and the IR light, but since I did not know the chromatic aberrations of the microscope objective this information would not help me.
I damaged a few of these laser diodes. Probably by burning one of the facets such that the lasing threshold was increased. Electrically they were OK, and the visible output appeared as intense as before, but the total output was only a few microwatts.
I therefore believe that the light people see from NIR laser diodes is spurious emission within the visible band, and not intense NIR radiation.
(From: Don Klipstein (Don@Misty.com).)
According to the official 'standard observer' photopic response of the human eye, the long wave cutoff is a gradual one. Sensitivity roughly halves for each 10 nm further into the infrared. This trend holds close to true enough 'officially' from 700 to at least 780 nm.
It seems as if a small spot is usually (maybe only barely) visible to dark-adapted eyes in a dark room with eye-safe levels of any wavelength up to around 880-900 nm, maybe 950 nm for brief viewing. (If your eye's long wave sensitivity is not below average!)
But you may not want to push your luck. A milliwatt of IR can permanently cook a spot of your retina, maybe within a couple seconds, and with no pain or warning. Prolonged focusing of any quantity of light over .4 microwatt onto a single point on the retina is potentially damaging, although several microwatts won't do damage in only seconds.
Be careful if the main beam of the IR laser diode is collimated or not known to not be collimated. Some IR laser diodes have visible spurious emission, which may detract you from the main beam. In some other IR laser diodes and depending on your eyes, most of what you find visible is the main IR wavelength and you may be exposing your eyes to plenty of it if you find it visible.
"The spectrum of this laser diode (Sanyo) is supposed to be quite narrow (about 3 or 4 nm) in the range 635 to 645nm. But when I have tested that diode, I have found that it emits light from 635 nm up to 660 nm!!! So the width of its spectrum is more than 20 nm!"(From: Mark Summerfield (m.summerfield@ieee.org).)
Could you give some more details of your measurement?
(From: Harvey Rutt (h.rutt@ecs.soton.ac.uk).)
Most laser diodes emit a broad background of spontaneous emission as well as the laser output.
A student of mine made another error a while back. He simply had the gain on the detection system turned up too high; the very narrow laser line was heavily saturating the system, and he saw those big broad wings.
Which incidentally can extend extraordinary distances and have all sorts of structure. One of our 810nm diodes puts out a load of broad band mess out near 2,000 nm (yes, 2 um!) but virtually nothing in the 1 to 1.8 um region.
Ordinary LEDs have peak wavelengths and dominant wavelengths:
The dominant wavelength is the wavelength (mixed with white if necessary) that matches the color of the light source in question. The white, if not specified, is usually C.I.E. Standard Illuminant C which is approx. 6500 Kelvin. C.I.E. Illuminant E, which has chromaticity of (.3333, .3333) and is very slightly purpler than approx. 5500 Kelvin, may also be used. Most LEDs are either close enough to matching a spectral color or on a blue-yellow line that most whites are close to that it is not really necessary to specify the white.
But here are the peak wavelengths, dominant wavelengths, and approximate limunous efficacies (lumens in each watt out, not lumens per watt in that I mention in The Brightest and Most Efficient LEDs and Where to Get Them! for various LEDs. The luminous efficacy of 555 nm is approx. 681 lumens per watt.
Please note that I have misplaced some Hewlett Packard LED datasheets which contain most of the luminous efficacy data that I had on hand. I may be able to recover some from Hewlett Packard's web sites and refine this later.
Type Peak (nm) Dominant (nm) Efficacy (lm/W)
----------------------------------------------------------------------------
GaAsP on GaAs substrate red 660 650 ~55
GaP/ZnO (low current red, 697 (nom)
varies with current) 660-697 600-640 ~10-30
GaAsP on GaP substrate red 630 615 ? 180-200+
GaAsP on GaP substate yellow 590 588 ? 400
GaAlAsP (ultrabright red) 660 645 typ. ? 80
have seen 635-650
"T.S." AlGaAs (HP) 646-655 637-644 ? 80-95
InGaAsP (bright red-orange) 620-625 608-615 ~200
InGaAsP bright yellow 590 588 400
GaP green 565 upper 560s-570 ? 620
(Brighter greens are similar)
"Pure green" GaP near 550 near 555 ? 670
(There is an InGaP with similar color)
Nichia InGaN green 522 (?) 525 very roughly 450
Toyoda Gosei InGaN green 516 520 very roughly 425
InGaN blue 466 470 very roughly 95
(Nichia and Toyoda Gosie)
GaN blue (Panasonic 450 nm) 450 470 ? very roughly 100
(This is a broader band blue)
SiC ("Cree type") blue 466-470 around 480 ? very roughly 130
GaN on SiC substrate blue 430 around 450 ? maybe 50
(Radio Shack 276-311)
Most visible LEDs have their characteristics specified at 20 mA.
Here are approximate characteristics (at 20 mA unless otherwise specified) for some brighter LEDs. Output power is total of the main beam and all stray output. ALL FIGURES ARE APPROXIMATE and are based on some crude measurements.
Output Forward
LED Type Power (mW) Voltage (V)
---------------------------------------------------------------------------
The best GaAlAsP ultrabright reds 4 to 4.5 1.85
Hewlett Packard red (630 nm) "AlInGaP-II" 5 to 5.4 2.00
Better orange InGaAlP, non-Hewlett-Packard 2.5 1.95 to 2.00
Hewlett Packard red-orange "AlInGaP-II" 3.1 to 3.5 2.05
Better amber/yellow InGaAlP types .9 to 1 2.02 to 2.15
Toshiba TLGA183P 0.26 2.1
Green InGaN/GaN 520-525 nm 2 3.4
Blue-green InGaN/GaN 480-500 nm 2.5 3.4
Blue InGaN/GaN 466-470 nm 3 3.4 to 3.5
Gallium Arsenide 950 nm infrared, at 50 mA 7 to 8 1.4
Gallium Aluminum Arsenide 880 nm IR, at 50 mA 12 to 14 1.6
NOTE: The InGaN/GaN types are nonlinear with decreasing efficiency at higher
currents and are most efficient at currents of just a few mA. All other types
mentioned above have maximum efficiency generally around 20 to 30 mA and
sometimes higher. Efficiency = Po/(V * I).
(From: Lou Boyd (boyd@apt2.sao.arizona.edu).)
Opto Diode Corporation offers the OD-100, TO-39 single diode which is rated is rated at 80 mW minimum, 100 mW typical continuous output at 500 mA. Their OD-669 T066 is an array of 9 diodes and produces 390 mW minimum, 1/2 watt typical at 300 mA (13.5 volts). Both are 880 nm. They cost about $9 and $60 respectively but there's a minimum order.
Get the black "satellite grade" solar cell, such as Radio Shack 276-124 (version at least A). There are soldering surfaces on the top and bottom along one of the long edges. There may be oxide on those surfaces, so clean these surfaces very gently with very fine sandpaper or preferably fine steel wool. Maybe place the cell on a flat surface while scrubbing it, since it is about as fragile as a piece of glass of the same dimensions. Then solder a pair of wires onto it, preferably 26-30 gauge. Solder quickly to avoid cooking the silicon.
Connect the solar cell to a milliammeter. Shine the LED on it, with known current flowing through the LED. Make a small white paper cone to get the side and rear light (although there is usually not much from most LEDs) on to the cell. Do this in a reasonably dark room but with some illumination on the solar cell meter. Move the LED and the white paper cone around for the highest reading. Note that the solar cell can scratch LEDs which will mar clear ones.
Multiply the solar cell current by 1.04 to approximately correct for the little silvery strips on the solar cell blocking light from it.
Divide the (multiplied by 1.04) solar cell current by the LED current - this will be the photon/electron ratio. Give or take a few to maybe several percent. :)
Note that the solar cell will read low with blue LEDs, probably a little low with blue-green ones, and definitely a little low with white ones. The blue solar cells are not as good as the black ones - the blue ones are only reasonably accurate from yellow to near IR, while the black ones are good from mid-green through near-IR.
Amorphous silicon solar cells, selenium solar cells and flexible solar cells will not work as well as single crystal silicon ones.
I have gotten figures that don't run low compared to figures claimed by LED manufacturers, so I know this method is reasonably accurate.
With many LEDs, the photon/electron ratio is close enough to the efficiency. To refine this figure, multiply it by the "electron volt" photon energy, which is 1240 divided by the wavelength in nanometers. The peak wavelength is close enough to valid for most LEDs. Then divide this by the voltage drop of the LED in volts.
If you want a luminous efficacy figure in lumens per watt, many LED manufacturers publish such figures for the emitted light. If your LED does not have a figure for the efficacy of the emitted light, it is normally close enough to that of other LEDs of the same peak wavelength, bandwidth, and basic chemistry. Multiply the manufacturer's figure by the LED's conversion efficiency to get the overall luminous efficacy, in lumens of light per watt of electricity.
CAUTION: While not generally on par with laser diodes in the danger area, these super bright LEDs must still be treated with respect especially if collimated or assembled into multiple LED arrays.
For more information and suppliers, see: Don Klipstein's The Brightest and most Efficient LEDs and Where to get Them!.
The quick answer is that an LED does not appear as a point source and has as effective emitting area which is huge compared to a laser diode. Even though the emitting area of a laser diode is not a point, due to the way the laser beam is generated - collimation wise - it appears as a point source.
And, a point source can be focused to another point.
The effective emitting area of an LED is perhaps .25 x .25 mm. To focus an incoherent source like this to a 2 um spot with imaging optics would require a ratio of distances of roughly 125:1 for the LED-to-lens compared to the lens-to-image plane.
With any kind of real world optics, you will get a vanishingly small amount of power at the image plane. Similarly, an LED beam cannot be cleaned up with a spatial filter (pinhole) as very little of the beam will make it through.
The laser diode is coherent and monochromatic (enough) that relatively simple optics can be used to focus it to a spot smaller than 2 um. While the dimensions of the laser diode chip are not all that much different from the LED, the characteristics of the laser emission makes such focusing a relatively easy task.
Consider that the beam from a HeNe or ruby laser doesn't come from point source either. The beam can be sharply focussed because it is very well collimated.
The availability of relatively cheap laser diodes really was the enabling technology for the CD revolution.
One interesting side note: burnt out laser diodes - i.e., those that still work as LEDs but do not lase - can be focused or collimated nicely. Not quite like a true laser diode, but much better than an LED since the emitting area is still very small - typically 1 um by a few um for a low power laser diode. Of course, the maximum optical power output of these blown devices is also quite small. :-(
(From: Steve Nosko (q10706@email.mot.com).)
If a beam of light has nothing but *precisely* parallel rays, it can be focused to a point. Also, if the beam originated from a point, a lens will focus it to a point.
An LED has neither of these. First, it is an area source and light coming from that surface is not parallel. It would also be called a diffuse source, meaning light from all places on the surface travels in many directions. This kind of source can not be focused to anything but a smaller image of itself. The shorter the focal length of the lens, the smaller the image - but it is still an image of the source, not a spot. It is because of these rays, traveling in different directions, that a lens can't focus them all to the same point. If you draw the side view of a lens and trace rays this all should be obvious.
The gas laser, on the other hand, has rays which are much much closer to being parallel. The diode laser has rays which appear to come from an apparent point inside the diode.
There are two more subtle effects. One effect is the relatively wide range of wavelengths in the LED versus the narrow range of a laser. Simple optics don't focus all wavelengths at the same focal length. So the wide bandwidth of the LED causes a little trouble. There is another effect having to do with the size of the lens (diffraction limit) and the wavelength, but this is also secondary to an understanding of the *primary* reason why an LED can't be focused.
The intensity of a light source can be loosely defined as the optical power divided by the area of the source divided by the solid angle of emission. An LED and a laser diode with the same power output which both happen to emit light over the same angle still differ in emitter area - by a huge amount.
Note that for wavelengths that pass through the cornea, lens, and vitreous of the eye to be focused on the retina, it is optical power that matters regardless of how bright any given wavelength may appear. Thus, 5 mW of 555 nm green light (to which the eye is most sensitive) has about the same damage potential as 5 mW of 780 nm IR (which is nearly invisible) in terms of how much heat will be delivered to a spot on the retina. However, the 780 nm IR, being nearly invisible, will not trigger blink reflexes and aversion responses so it is in fact much more dangerous. For more info, see the section: Laser Safety and Diode Laser Safety.)
A typical laser diode is diffraction limited with an emitting area of about 1 um by 3 um. An LED will emit from the surface of its chip over a full hemisphere. In order to radiate over only a limited angle, a lens is added but with idea optics, the emitting area is still effectively the .3 mm x .3 mm or so area of the LED. This is about 30,000 times that of the laser diode. With the typical 5 mm molded lens, it is more like 1,000,000 times the area. That means the intensity and optical density of the laser diode can be over a million times that of the LED.
When the light from either device hits your eye you will not likely have any problems because your (at most) 7 mm diameter pupil will intercept very little of the light if you are several inches from the source.
The light from the lensed LED starts out 5 mm diameter at the lens and expands rapidly until it hits the eye. The amount of light entering your eye a couple of inches away is very small.
Even if you push the LED up against your cornea you cannot focus on anything that close to your eye even if you are extremely nearsighted. So the image of the very close LED that forms on the retina is blurred. That means it is very large and the power density is low. Adding another lens in front of the LED doesn't help. It is not possible to collimate it very well due to that large emitting area (and the plastic lens that is likely present).
But, with a simple lens you could collimate the laser diode to a 5 mm (or smaller) diameter beam with a very small fraction of a degree of divergence. All of the light from this collimated laser could enter the eye easily and be focused to a small spot nearly instantly burning a small pit in your retina. A collimated beam appears to be at an infinite distance even if its source is up close. The eye can focus it to a very small (diffraction limited) spot with a high power density.
There will likely be safety warnings on the packaging for high power LEDs though you won't find little tiny stickers on the LEDs themselves!
Also see the section: Why Can an LED Not be Focused Like a Laser Diode?.
In my experience, LEDs that shift to longer wavelengths when they overheat have one distinct emission band. That band shifts when they heat up. What I think happens is the conductance band and valence band widen from thermal agitation and the "bandgap" between these bands gets narrower. The result is longer wavelength at higher temperatures.
Most of the usual common green LEDs have a noticeable color shift when they overheat. In my experience, they degrade slightly when they get hot enough to shift to orange and stay that hot for a second or two. At the maximum temperature that they can withstand long term, there is hardly any visible color shift.
Most yellow ones will shift to orange or reddish orange when overheated.
Now for another color shift phenomenon that is quite nondestructive: "Low Current Red" or "697 nM red" which is GaP doped with zinc oxide often has color noticeably varying with current. At a couple milliamps or less, these are red and efficient. At higher currents, they operate less efficiently and turn orange. These are sometimes even yellowish or whitish-yellowish orange at 30 mA. The spectrum is very broad and the amount of content at longer wavelengths in the broad band does not increase proportionately with current while shorter wavelength content increases more linearly. In addition, there is a weak secondary band in the bluish green.
Other LEDs of types whose efficiency is maximized at low currents often have color visible shifting to shorter wavelengths at higher current and longer wavelengths at lower current. Toyoda Gosei and Panasonic bright blue LEDs are cyan-aqua at 100 uA. Most of the ultrabright 520 to 525 nm greens are yellowish at 50 to 100 uA, a beautiful green at 5 mA, and slightly blue-whitish at 30 mA.
Preliminary specifications (Source: EE Times, January 18, 1999):
There is also a 30 mW version in the works or already available (Novermber, 2000). I don't even want to think about its price. However, its MTTF is quoted by Nichia to be only 500 hours at 25 °C. This means half will fail by 500 hours. Ouch! :)
Ironically, it seems that it may be easier to produce reliable violet laser diodes rather than blue or green (despite possible previous reports of demonstrations of blue ones at least). This would be good news for next generation optical storage (beyond DVD) and high resolution laser printers but those wanting highly visible wavelengths (e.g., 555 nm green and full color displays) may have to wait a bit longer. The actual luminous efficiency (relative visibility) at 400 nm is only about .28 percent of that at 555 nm. This corresponds to about .2 lumen/watt compared to 16 to 20 lumens/watt for a 100 W incandescent light bulb! Nonetheless, this could be the start of something spectacular. :)
The availability of cheap, long lived, shorter wavelength (than the 635 to 650 nm types that are now used in better laser pointers and for DVD players and drives) laser diodes could usher in yet another quantum leap in solid state electro-optics technology. (Yes, I know, taken literally, 'quantum leap' may not make sense but you get the idea.)
When economical, these shorter wavelength laser diodes will represent the enabling technology for yet another revolution in the storage capacity of optical drives (at least a factor of two better than even DVD). Compared to the 4.5 GB capacity of one surface, one side of a DVD, a DUD (Digital Ultra Disc/k) drive would hold about 13 GB based on the wavelength difference alone (635/400 squared). As if we need yet another new standard. :)
Shorter wavelength laser diodes should also find applications in higher resolution laser printers and similar devices. Blue wavelengths (not violet though) would be ideal for underwater communications. With the addition of green laser diodes, compact full color displays and many other products would quickly follow. However, at the current time, only the violet laser diodes at around 400 nm are commercially available - blue and green may still be a few years away (as of January, 2000).
Blue and green has been widely demonstrated by SHG (second harmonic generation also known as frequency doubling) in nonlinear crystals. This approach is widely used now for lasers of all sizes. However, such technology is quite complex and currently very expensive. For example, a typical low power green (532 nm) device such as found in a *green* laser pointer includes a high power IR laser diode (emitting at around 800 nm) exciting a tiny Nd:YAG chip (which lases at 1064 nm) coupled to another chip of KTP which doubles its output to 532 nm - plus a whole bunch of needed optics to form a cavity, collimate the beam, and prevent stray IR from escaping, all mounted in precise alignment. No wonder they are cost several hundred dollars! A green laser diode would eventually cost no more than the common red ones resulting similarly priced green laser pointers (as if we need more of those!). See the section: Diode Pumped Solid State Lasers.
The direct emission from a semiconductor has been the Holy Grail for several of laser engineering years. The semiconductor materials available with a sufficiently wide band-gap are notoriously difficult to deposit and cleave. Many companies around the world have been working on this problem but until relatively recently, power output, operating temperature range, and/or laser diode life have been unacceptable. However, in late 1997, there was strong evidence that all this was about to change:
"Nichia Chemical Industries, Tokushima, Japan, has reported passing a major milestone in the development of blue laser diodes with the demonstration of a InGaN/GaN/AlGaN device with an estimated lifetime of more than 10,000 hours under CW operation at 20 °C. The announcement was made by Shuji Nakamura of Nichia on October 30, 1997, at the 2nd International Conference on Nitride Semiconductors, held in Tokushima, Japan. Working devices have been demonstrated (even a laser pointer!) and there is reason to believe that they may be commercialized in the near future. The same technology can also produce highly efficient laser diodes of other colors ranging from red through yellow and green."Since this news release over a year ago, there have been various hints that such devices were moving closer to commercial production but until the News Flash, above, in early 1999, no sample devices were available. Now, it would seem that the age of laser diodes of all colors of the spectrum is about to begin. [Hype mode off.]
I've been working in the nitrides for a couple of years and it is the case that the lasers lase easiest and best right around 400 nm, from ~395 to 420 nm. Going further either way is tough, but Nichia may be able to do it. Nakamura keeps astonishing us all. They do have amber nitride-based LEDs, which is another amazing accomplishment that no one else has repeated."
(From: P. Meyer (meyer@lps.u-psud.fr).)
Nakamura demonstrated a hand-held near-UV LD system some two years ago (if I remember right) at the Strasbourg EMRS meeting. He told, that visible (blue) laser operation was not yet possible (2 years ago - 1997). So, now the announcement of a 400 nm LD is good news - although this seems rather the limit for visible.
(From: Michael J. Bergmann (mjb@phy.duke.edu).)
I think most of the reports should be properly labeled violet, not blue. The ones I've seen have all been violet. The longest wavelength I've seen in the literature were Xerox's and Cree's lasers: ~430 nm.
Nichia started out around 420 to 411 nm under pulsed operation in 1996 and have been getting shorter in wavelength as they have gone CW and long lived. For instance, lifetime and wavelength went as follows:
One source for additional technical information on this work is: "Present status and future of blue LEDs and LDs", Review of Laser Engineering, vol. 25, no. 12, p. 850-4.
Xerox Corporation has just announced successful testing of a blue laser diode for use in high performance laser printers, phototypesetters, and similar equipment. Little information is currently available so life, cost, and detailed specifications are unknown. See the Xerox Press Release for some information (mostly marketing hype).
For some more technical info about the semiconductor physics of short wavelength laser diodes and other guaranteed cures for insomnia try these links:
(From: Gregory J. Whaley gwhaley@tiny.net).)