Some thoughts about high power diode laser for Holography

I had been interested in Holography since 1994 and making beautiful hologram is one of my dreams. However, even if it was invented in 1948, it took a long time for holography to reach ordinary people and making hologram is still an expensive and “laboratory” work. Finding cheap laser with long coherence length and cheap recording material is a pain.

I made my first hologram in September 2006 and didn’t continue due to my Ph.D research. After almost 4 years, I began using high power diode laser and DPSS (532nm) for holography recently. What I found is bad results for using 200mW 660nm diode laser and 200mW 405nm diode laser. Wonderful webpages at and give me the right answer.

My comments and suggestion:

Quan Says:
November 13, 2010 at 12:41 amReply

Thank you so much for the wonderful testing of all these laser diodes. I  made holograms 4 years ago using 5mW 650nm diode laser (wonderful result) and tried to use high power diode laser (150-200mW, 660nm, 532nm, 405nm) to make holograms this month. The problems I have are the “sliced bread” in transmission holograms and “dim, sliced bread, wrong color, out of focus” in Denisyuk holograms. I realize that this is caused by the short coherence length of the laser but wonder why 5mW diode laser is much better than 200mW diode laser ( I tried 2 different 200mW 660nm diode laser but all showed “sliced bread”, and the “space” between “sliced bread” is different for two diode lasers). Now I realized why the single longitude mode laser is so expensive and not very powerful. Before come here to read all of your testing graphs, I googled the wikipedia and find the relationship of frequency difference of different modes is c/(2nL),
while c is speed of light and L is the length of resonator (mirrors), since the L of diode is 0.02-0.3 mm and the delta wavelength (changes) between different modes are much higher than those of gas lasers (L~20cm, or meters). My calculation of the delta wavelength between muti-modes of 600nm diode for L=0.05mm is about 3.6nm.  (delta lamda=((lamda)^2)/(2*L) and that gives about the correct coherence length as the size of “sliced bread”. My result of 532nm is much better and I think that is due to the longer L of the YVO4 crystal for 1064nm laser. I failed to make Denisyuk hologram using 200mW 405nm diode laser for the length of 2 cm of object. It is easy to understand that higher current will lead to higher photon energy of the muti-modes. But it seems that the distribution of different modes should follow the boltzmann distribution. I was happy to see the growing market of high power diode laser but how can we keep the power and still get a single longitudinal mode? Do you think lowering the temperature of the diode will do the job? I mean using liquid nitrogen or dry ice to keep the diode in very cold temperature. Glad to see your web site and blog! Thanks a lot.

rxlaser Says:
November 13, 2010 at 9:55 amReply
Hi Quan, yes I agree with your statements and computations. Many low power diodes can run single mode, and most high power diodes too but not up to high powers. I would say that typical red DVD diodes can run single mode up to approx 30mW, sometimes higher. Some are completely unsuitable, though, see my tests. Blu ray diodes 405nm seem pretty bad in general, but the 445nm diodes much better again.

Thanks for your interest!

Quan Says:
November 13, 2010 at 2:38 amReply

Lowering the temperature should increase the ratio of distribution between different modes. According to Boltzmann distribution, n1/n2=exp(-E1/(kt))/exp(-E2/(kt))=exp(delta E/(kt)). So even if we lower the temperature a little bit, the ratio of amplitudes from different modes should increase a lot. And the modes that we don’t want will be significantly lowered. So by using lower temperature, if the depth of “sliced bread” increases, that means GOOD! I think it is easy and feasible for hobby holographer to lower the temperature of high power laser diode rather than buying expensive high power single longitudinal mode gas laser.

rxlaser Says:
November 13, 2010 at 10:03 amReply
Hi again, I don’t have experience with low temperatures, because sooner or later there is moisture condensation and ice appears on the diode and optics, so one needs to hermetically seal the whole thing and desiccate, professionals use nitrogen as ambient atmosphere. I have no idea how well diodes behave then, all what I can say that lowering the temp from like 25C to 15C, does already have a positive effect in most cases.

Quite generally, there seems a magical barrier at approx 100mW for almost all diodes, includung ECDL setups, above which multimode operation appears. I wonder whether this can be improved at low temperatures.

Quan Says:
November 13, 2010 at 6:59 pmReply

Hi, glad to see your quick reply. Yes, the moisture in the air (room temperature) will condense on the surface of cold objects but we still feel very dry in cold winter. The ambient relative humidity is about 30-50% but the pressure of water vapor in the air varies a lot from summer to winter. The reason of condensation is that the pressure of water vapor in ambient temperature is above the saturated water vapor at lower temperature. It’s already cold in north hemisphere at night now and I wonder if you can test the diodes at night in a room without heater or move your instruments out of the room. (Sorry for bad suggestions) Another way to get rid of ambient humidity in a closed small room is put a lot of ice (3-5 kg, -5 degree to -10 degree Celcuis, the colder the better) in the room until the humidity is condensed on the surface of ice and then you can test the profile of diodes in a little bit higher (5 degree C) temperature than the ice. I will try to make hologram with my 200mW 660nm diode laser in really cold winter night in a room without heater.

Quan Says:
November 14, 2010 at 1:56 amReply

BTW, I forget to mention that don’t use liquid nitrogen or dry ice in a closed room that will cause suffocation and can kill people. But ordinary ice from super cold freezer (-80 C) is best for decrease the humidity of the air in closed room. In another word, you just put big container with higher surface area full of cold ice for condensation to compete the unwanted condensation on diode.

qyuan0366 Says:
November 14, 2010 at 9:53 pmReply

Hi, I have something to tell you about my thoughts and what I did. Since I don’t have the instruments to measure the spectrum of multi modes, I use a simple way to see the mode hop. Put a clean thin glass (2mm) in front of the laser and see the interference pattern of the reflected light about 2 meters away or further. The rings of the interference pattern on the wall is slowly changing at the beginning when diode is turned on (100mW 660nm). But the interference pattern is sharp and clear. The slow changing of the rings is due to the temperature change of the diode. (wavelength change caused by temperature change). After a while, about 2 minutes, the rings are constant and don’t move, suggesting the thermo-equilibrium is reached at the current power. If increasing the power, the rings begin changing again. And will take a while to reach the heat equilibrium within diode. When the power is high, a sudden change of all interference pattern can be seen and the change might last for several seconds to minutes. During this change, the previous sharp and clear interference pattern become “evenly bright” or “less contrast” and some new patterns occur. I think that is due to mode hop. This is common for 660nm 200mW diode laser and also for 200mW 532nm DPSS laser. I carefully watched the interference pattern of 5mW diode for several minutes and no pattern change was observed. So here is the question, is the multi-mode at higher power caused by just high current or due to the high heat? Even if we soak the diode in liquid nitrogen, as long as the diode is on, there is a temperature gradient across the center of diode and the surface of the diode. I wonder the multi-mode of the diode at higher power is due to high heat generation inside of the diode and the actual temperature in the diode center is much higher than the ambient temperature even if the thermo-equilibrium is reached. For 5mW diode, the heat generated in diode is so low and can quickly dissipate. If the temperatures in the center of the diodes are same, there is no reason that 200mW diode runs multi-mode but 5mW diode runs single mode. I calculated the ratio of different modes using Boltzmann distribution. To my surprise, the ratio is not highly dependent on temperature. Here is my calculation:

For 660nm diode laser, assuming L(length of resonator)=0.1mm, so the wavelength difference between different modes is 2.18nm, the energy difference between two modes next to each other is
the ratio of different mode at 20C:
n1/n2=exp(delta E/(kt))=exp((9.945E-22)/((1.38E-23)*298))=exp(0.2417)=1.27
the ratio at 273k is 1.30 and 2.55 at 77k (liquid nitrogen).

So the reason that 5mW diode runs in single mode is completely due to the fast heat dissipation rather than the ratio of different mode at low power consumption. When you measure the properties of the diodes at 15C at high power, the actual temperature of diode might be as high as 50 or 60C even if the surface temperature of diode is 15C. So when power is above 100mW, it’s so hard to get rid of the heat fast enough and multi-mode occurs. I am pretty sure that diode can get single mode even at 200mW or higher as long as the heat is quickly dissipated. “Quickly” here I mean the temperature in the center of diode must be less than 20C or 25C.

rxlaser Says:
November 15, 2010 at 10:51 amReply
Hi Qyuan, thanks for your comments! I believe these are difficult questions and I am not really quaified to give a meaningful answer, as to what all the causes for multimode behavior at higher currents are and whether this can be helped by low temperatures. Obviously a higher gain will always give a tendency for more modes to lase, but why there seems to be a quite univeral border at around 100mW for practically all diodes incl ECDL setups escapes me. Probably thermal gradients and induced optical inhomogeneities, lensing effects in the material, etc, play a role and I just dont know what effect a very low temperature would have. This would be interesting to investigate.

But I guess for practically making holograms, dealing with ultracold substances would make things unnecessarily complicated, it is difficult enough to reach thermal stability and avoid drafts and thermal gradients over the holographic setup. SInce one can reach in the order of 100mW with ECDL, which is quite a lot, there is no strong reason to handle liquid nitrogen etc to reach a bit more power!

Quan Says:
November 15, 2010 at 7:42 pmReply

Hi, rxlaser, thanks. I agree with you that “there is no strong reason to handle liquid nitrogen etc to reach a bit more power”. I was thinking that if we can dissipate the high heat from the 1W 445nm diode using liquid nitrogen, we might get over 1W in single mode and it probably can be used for portrait or “non-still” objects without using the vibration-free platform. I am a biochemist and can get liquid nitrogen or dry ice easily in our lab. Also, there is a pulsed Nd:YAG laser with a pulse duration of 4ns and energy of 200mJ/pulse (532nm after SHG) in our lab for laser photolysis study and I am using it for research. I wish I can use it for portrait of people but obviously NOT because of the burning of the eyes. So I am interested in CW semiconductor laser. Let’s do a little bit calculation again here: For slavich emulsion (VRP-M, PFG-03C or PFG-M) the sensitivity at 445nm is about the same as logS=-3.4, which is about 2500uJ/cm^2 for the energy need for exposure. For a 1W laser, the time need for 2500uJ at 100cm^2 is only 0.25 sec. If the 660nm laser can run single mode at 250mW, the time need for 30uJ/cm^2 (sensitivity of red 660nm emulsion from Tianjin, China) at 100cm^2 is 0.012 sec. Such short time will allow the speed of movement of 0.01mm/sec. (for 1/4 wavelength tolerance). I know it might not as good as the prediction here but I just do some calculations for fun before doing the real work.

I observed the interference pattern again last night and want to tell you more. For 5mW diode, there is also mode hop (interference pattern change) but the rings are still sharp and clear. For my 200mW 532nm DPSS, if it is powered by 2.4V (before the circuit), the interference pattern is pretty stable for quite a long time. But when powered with 3.7 V (before the circuit, the normal working voltage), I see the interference changes often in about 5 minutes. For 200mW 660nm diode, I just saw very bad interference pattern. There was no high-contrast rings like those by 5mW 650nm diode or 532nm DPSS.

Quan Says:

December 2, 2010 at 8:36 pmReply

Hi, rxlaser, how are you recently?
Find a good post about single and multi mode for laser diodes:

Also, I just bought a 445nm 1 W Nicha laser diode from ebay. What I found is another problem: To make sure I can get single mode at low power, I monitored the current very carefully and it won’t excess 300mA. (Operating current is 1400mA). I use a standard 5.6mm diode housing heat sink.when the current was set as 200mA(monitored with digital multi-meters for voltage and current), 445nm laser is coming out. But the current is slowly going up to 300mA (keeping the voltage constant) in about 3 minutes and 445nm laser was replaced by blue light like LED. If I turn it off and wait for just 3 seconds, the laser is back again but will slowly change to ordinary blue light after a while. I don’t have the problem for my 405nm diode/660nm diode. Does that mean I was running just above the threshold limit or because of above the thermal limit?I find the laser modular is warm (40-50C degree) after 3 minutes even if powered by 300mA? Is this normal for 1W 445nm laser diode? It seems that 1W 445nm laser diode has a lower electric-optical conversion efficiency than 405nm/660nm laser diodes.

rxlaser Says:
December 3, 2010 at 5:37 pmReplyHi Quan,

you should always run a diode with constant current and not constant voltage (because the forward voltage is strongly temperature dependent). I guess the diode got very hot and then below the lasing threshold. Certainly 50C can be easily reached by 300mA. Better cool it properly, this is good for its life time and also for holography application better have a very stable temperature control.

Quan Says:
December 4, 2010 at 1:36 amReply

Hi, rxlaser, thanks for advices.
I removed the diode from the modular and made an big aluminium heat sink which fully contacts with diode case as much as possible. Yes, as long as the temperature is constant, I didn’t see the increase of the current. (Stays there at 210mA for about 8 minutes). I began using constant voltage rather than constant current when I found the fluctuation of the optical output of 660nm 200mW diode. For some reasons, (660nm 200mW diode, low heat generation than 445nm 1W diode) even if the current is constant, the photometer for optical output measurement is changing up and down after 10 minute stabilization. No optical power fluctuation was found in 5mW 650nm diode. So I tried to risk those in constant voltage.

qyuan0366 Says:
December 6, 2010 at 4:23 amReply

Hi, rxlaser. Sorry for bothering you again for another question. A laser diode behaviors like LED below thresh hold limit. Can we increase the current of a ordinary LED or high power LED (keeping the temperature low with dry ice or liquid nitrogen for fast heat dissipation) to get laser? I don’t know the structural difference between LED and laser diode and wonder how the population inversion can be obtained in laser diode while no population inversion in LED.

rxlaser Says:
December 6, 2010 at 8:16 amReply

Hi Qyuan,

no a laser diode needs a decent resonator to start with ;-( While a LD can work as expensive LED, sometimes unvoluntarily, it doesn’t work the other way around,

qyuan0366 Says:
December 16, 2010 at 6:11 amReply

Hi, Wolfgang. More questions for holography. I didn’t use the special polarized laser for holography just the 5mM diode laser. What’s the advantage of using the polarized laser? (except to remove the unwanted reflection between surfaces of the plate) If I split the random polarized laser to two beams of laser with perpendicular polarization planes as object beam and reference beam, using metal object or anything without Brewster’s effect, will I get the interference pattern/hologram? Or for Young’s double slits experiment, if I put two polarize filters on each slit (polarization planes are perpendicular to each other ), can I see the interference pattern on the screen?

此条目发表在Uncategorized分类目录,贴了, , , , , 标签。将固定链接加入收藏夹。


Fill in your details below or click an icon to log in: 徽标

You are commenting using your account. Log Out /  更改 )

Google+ photo

You are commenting using your Google+ account. Log Out /  更改 )

Twitter picture

You are commenting using your Twitter account. Log Out /  更改 )

Facebook photo

You are commenting using your Facebook account. Log Out /  更改 )


Connecting to %s