Research Journal


Wednesday, August 11th, 2010 Last week ended well as I got some very good correlation between sideband coupling and mode-locking as expected. I've yet to put everything together in a paper. I'm somewhat comfortable (that is to say a lot more comfortable than two weeks ago) on explaining AM mode-locking. If some loss can be introduced into the cavity, these oscillations will have frequency components that will appear as sidbands on the spectrum analyzer. This must mean that they are beating at frequencies close to the mode spacing frequency. I still don't feel its enough to say that modulation magically results in sidebands without there being a physical change in the wave structure of the longitudinal modes. For intracavity AM locked lasers the math and the theory is not hard to follow, but for the extra-cavity method we are following I am finding little to no theoretical explanation of what is occuring. The most that I've found is papers with a lot of math and very little physical explanation. The key has to lie in how the injection light interacts with the longitudinal modes.

The following is a guess to how the injection of modulated light can produce for sidebands. The laser has say 10 longitudinal modes. Each successive mode is 160Mhz greater than the next, so thats why you see that the 160Mhz beating is the highest peak on the spectrum analyzer. Then you inject light from the outside, assume its another laser where each of its 10 modes (modes 1-10) is 160 Mhz greater than the modes (1-10) of the first laser. Now you look at the beating between those 20 modes inside the first cavity. Each one of those modes is still 160Mhz apart so you see even more beating at 160Mhz. If the spacing of the modes in the injection light is somewhat off 160Mhz than there will be beating at 160Mhz +/- that difference. When it is exactly 160Mhz then the modes will couple in phase, promoting mode-locking.

How does this qualify as amplitude modulation? The modes of the injection light are smaller in amplitude, and this amplitude varies as well. The electric fields of the cavity and injection modes will superimpose, modulating the amplitude of the former's modes. Wouldn't this cause the sidebands, or does what occurs in the previous paragraphs result in sidebands? Couldn't non-frequency shofted light with the same modes spacing be used to mode-lock the laser as well? There are still plenty of questions to be answered.

Friday, July 30, 2010

This week went from a high degree of uncertainty to some tangible results in the experiment. We were able to see clear pulses with the expected frequency spacing; but they were too borad and of low intensity to be considered completely consistent with the theory. Today Marty suggested that I find a way to mechanically stabilize the mirror mount because it is affecting the external cavity optical length as it is very sensitive to vibrations. I remember the interferometer set up in the lab and how a rubber band can be used to move the fringes which are on the order of nanometers. I placed at least a dozen rubber bands from the mirror mount to screws in the optics table. This has reduced the mirror sensitivity somewhat, and if I pull on the rubber bands I can see the pulses sweeping across the oscilloscope, so I think I can use that to carefully adjust the external cavity length. However to be more accurate I need a better way to stabilize the mount. Though I am glad that I do have tangible results to put into a paper this weekend.


Wednesday, July 28, 2010

I've had the mode-locking experiment set up for about a week with no results yet. Putting the RF components together was easy after becoming familiar with the various pieces and seeing that the AOM can be driven a number of ways. After a couple of days struggling with misbehaving oscilloscopes I'm getting better at seeing what's going on in the laser. Laser Sam's Fabry Perot is still very handy as it clearly shows the 8-10 modes oscillating in my laser (Spectra Physics model 127). Marty says that when the laser is mode locked, the modes will stand in place and that if I check the signal from a photodetector on an oscilloscope I should see pulses whose separation is that of the mode spacing (160 MHz). A couple of days ago I thought I saw pulses with this separation, but that was when the equipment was misbehaving, so I cannot take that to be accurate. Today I tried to look for pulses on the same scope and on a newer digital oscilloscope, but what I saw instead was a flat signal 5V DC on both scopes so I think I can conclude that the equipment is reliable.

Mostly I've been staring at the modes from the Fabry Perot, and I cannot see any significant effects from the feedback light. It could be an alignment issue, or it could just be that I haven't hit that precise external cavity length; I feel like its a hit or miss chance to achieve mode-locking, or like the proverbial "needle in a haystack." There are just so many variables to think about, and a couple of changes to the setup may be useful such as changing the distance between the AOM and the laser output to allow for greater separation between the zeroth and first order diffracted beams from the reflected light. This would decrease the separation of the diffracted beams from the first pass through the AOM and this separation is small as it is. I've also been working on learning the theory behind the setup so that I can at least make a quality presentation and paper at the end of the program.


Friday, July 16th, 2010

My interest in acousto-optic physics and applications is growing as I read more about the subject. I started by reading material on tensors since it is the strain that a crystal experiences that results in the change of its index of refraction. Yet the optical properties of the crystal under some external field can be understood without getting into the details of tensor mathematics (although I do find it interesting). The optical axis of a material can be modelled by the use of an indicatrix which is basically an ellipsoid whose radii in the x, y, and z directions are proportional to 1/n^2 in that direction. The changes in the ideces of refraction can be calculated with some matrix math, where each type of crystal has its own matrix describing how it responds to external forces from electric fields or strains from acoustic waves (these matrices I assume are derived from the tensor description of stress on the material).

I have yet to finish reading on this matter, as I don't understand all the math. At this point it is probable that I will be doing a project with an acousto-optic modulator which is driven by radio frequency components which I have no prior experience with, and finding web pages with information on how to put the different parts together for such an experiment seem to not exist. The concept of the potential project is simple. When a laser is incident with acoustic waves at the Bragg angle it experiences both frequency shifts and diffraction. The stress induced by the sound waves in the crystal results in a series of reflecting planes separated in space by the wavelenth of the sound. From my understanding, these planes are reflecting because as each wave strains the material the index of refraction changes, and from basic optics changes in index cause reflection of incident light. The various planes of reflection along the crystal act as a diffraction grating as the path differences between the various reflect waves will result in destructive and constructive intereference at different points in space. Since the acoustic waves are travelling, the light experiences a frequency shift from the Doppler effect.

The project will attempt to use the acousto-optic frequency shifts to mode lock a He-Ne laser. The laser will be shone through the AOM, hit a mirror, take a secon pass back through the AOM resulting in double the frequency shift, and this beam reflected back into the laser cavity. The setup has to be built so that double the frequency shift is equal to the laser's mode spacing (c/2L), and so that the beam comes back with the proper phase. I must think about these last couple of things a little more so that I can understand why this sort of feedback can result in mode-locking.

Right now I am trying to become familiarized with the RF components I will need. The AOM (consisting of the PZT and crystal) comes in one package. This has to be driven by a component that makes radio frequencies (RF driver), but its not as simple as putting a cable between these two boxes. I saw on one internet source that you can take a voltage controlled oscillator which will produce the desired frequencies (this project will need a frequency of ~70Mhz; a VCO for this is available), the VCO is attached to an attenuator - I don't know why, since I doubt much power is coming from a circuit a square inch in size- and then finally attached to an RF amplifier before going into the actual AOM. Many of the parts I have to look at are mismatched. The AOM's (with one exception) are fabricated for 800nm beams, though I think it is possible to get around it, I would just need to recalculate the Bragg angle for 632nm. Then, I am not sure what to do with the RF drivers, attenuators, and VCO's since I don't know what is compatible with what since the spec sheets don't say much on the power requirements for these components. I guess those type of specs are more appropriate for electronics, and RF components are rated differently. But this problem is a large pothole since I cannot really do anything until I know what to do with these pieces.


Monday, June 28th, 2010

Although its two weeks into the REU program, I haven't gotten around to writing anything in the journal because I wasn't sure what I would be doing this summer. But I saw an article in OPN (Optics and Photonics News) on integrated optics and how the industry is trying to replicate electronic devices with photonic ones. I've been reading different books on photonics and have become interested in doing a project on acousto-optics. This phenomenon can be used to diffract a laser beam, as well as be used in optical signal filters and frequency shifters.

From the small amount of material I've been able to read, electrodynamics is an important part of acousto-optics and integrated photonics because the light often must be treated on the wavelength scale (as opposed to represented as rays). This is due to the fact that integrated optics devices are microscopic in size so the propagation of light must be carefully described according to Maxwell's equations. One book I've been reading specifically on acousto-optics relies on tensor notation to describe the optical properties of different media. Yet when describing acousto-optic Bragg diffraction this book does not use tensors to describe this process. The tensor-based geometry and physics seems to be more important in the materials-science part of the Piezo transducers used in acousto-optics. It may be more necessary for me to brush up on wave properties, though learning about tensors and materials science is also interesting to me.


Thursday, April 22nd, 2010

Its done. I hope. Today I finished the poster which will be the final product of this project, and will be presented at URECA's celebration of undergraduate research. In the last month I've gotten some very nice results for the laser, locking is consistent, and I've made estimates of stability, all is found on the poster. For anyone else who attempts this in the future, here are a few tips:

1) Keep a good notebook, with good detail of what you are doing. You may need to look something up later.

2) Knowledge of the electronics isn't a must, but if you need to know about PID circuits, wikipedia isn't a bad place to start.

3) Make sure that the power supplies for the laser and the circuit are in good shape and behaving the way they are supposed to. Not knowing this held me back a few weeks.

4) When you place the PBS in back or in front of the laser, make sure the laser is aligned vertically or horizontally so that you get a good separation of the perpendicular modes. Check this by placing a polarizer sheet in front of the tube either vertically or horizontally, and you should see a minimum periodically.

5) When you are ready to start getting data with the USB data collector, read the manual beforehand. Its a bit technical but you should make sure that no voltage going into the device is more than 10V. Also, make sure that the voltage difference between ground on whatever voltage you are measuring and the ground input on the data collector is small. What you don't want is that this difference plus the voltage of what you are measuring add up to more than 10V, or less than -10V. This wasn't an issue in my setup, but its better to be safe than sorry.

6) If you are using a coil heater like I did, you have to wind it bifilar around the tube. To make things easier, I unwound all 200 feet of it in the hallway, walking 100 feet one way, placing a stand on the ground so that the wire could double back (and taping the midpoint of the wire to that stand so that I wouldn't lose it), and walking back 100 feet to where I started. Then what I should have done was take two spools, and spool each loose end back to the midpoint on its own spool. Then when you wrap it around the tube, you start from the midpoint of the wire, and you can easily wind both spools around the tube. This really helps avoid entanglements (I learned the hard way).


Friday, March 5th, 2010

So I've been able to continue locking the laser, so now the final steps of this project are getting some data, and writing the report. Today I made some graphs with the USB 1208 LS. The graphs display the voltage from a photodetector over time, and really show the difference between the laser when it is locked and unlocked. Here they are:

Stabilized Laser

Unlocked Laser


Friday, February 19th, 2010

!!! LOCK ACHIEVED !!!

Well, it turns out that the quality of the power supplies made all the difference. I got another power supply for the laser, and now the locks are lasting much longer. Today I observed a lock of atleast 15 minutes (with a minor hiccup around minute 10) before I started tinkering with the circuit again. What is great is that the supply powering the circuit has an ammeter in it that lets me see approximately how much current is being used by the heater, and it lets me see in real time how the circuit responds to the change in temperature.

As always, I've been tracking the behavior of the laser with the scanning Fabry Perot Interferometer, and as the laser threatens to change modes the current read by the ammeter drops as the circuit lets the tube cool, or jumps as the circuit tries to heat it. From this point forward, I'm going to try to get more detailed measurements using an A/D data acquisition module (the model the Dr. Noe has is the USB 1208 LS from Measurement Computing ) which I have to become more familiar with before I can start using it. Ultimately I would like to get side-by-side graphs of the behavior of the circuit over time, and output intensity of the laser over time, that way I can really get into what the setup is doing. Hopefully everything goes as smoothly as things have gone this week!


Wednesday, February 17th, 2010

I know it has been a long time since my last journal entry, but it does not mean I have not been working on my project. But today I am enocuraged by the latest developments, as I appear to be closing in on a lock. the power supply for the circuit I had been using seemed to have trouble supplying a constant current to it, so today i switched it with the more reliable supply that was powering the laser. Furthermore, over the last few times I've been in the LTC I've been probing various parts of the circuit with a voltmeter to see how the circuit responds to different inputs such as different currents from the photodiodes, and different resistances from the gain and offset pots. I'm still working on getting a clear picture of what is going on in the circuit, but I'm getting there.

Well, after all this the circuit appears to be getting closer to locking, as it managed to hold the Scanning FP Interferometer's output for 2 minutes! I think that by tinkering a little more with the hipots I can get it to improve. Also, since the unreliable power supply is the one powering the laser, it also is probably affecting the quality of the setup, so if I can get good power supplies for both it should improve my chances of getting a lock.


Thursday, January 28th, 2010

The semester has started, which brings many other burdens to weigh on my mind for the next few months. I think that this project is in its final stages as today Dr. Cohen and I got Sam's Scanning FP Interferometer running, which gave some very interesting patterns on the oscilloscope. We could distinctly see the modes oscillating in the Interferometer's cavity. It took me a while to understand what I saw, but what I could tell was that laser by itself (with no feedback circuit) is unstable, as the peaks sweep across the screen. With the circuit on, the sweeping slows, indicating that the circuit is having its intended effect. However, the~ laser does not "lock" in the sense that the peaks are stationary. To give you, the reader, and myself a reference for what I saw, here is a link to a website that shows examples of the same experiment we did today. This site is useful because it gives examples of some measurements to take and what they mean and how to use them. I will probably take the same measurements to test the performance of the controller circuit.

We also used the oscilloscope to see what is happening in Sam's circuit. Test point 2, by my understanding, is where I can see how the circuit is computing the error signal. I can change the range of this signal with the gain hi-pot, but changing the offset hi-pot has no noticeable effect at this test point. Hopefully with some more of Dr. Noe's instruction I can take more strides towards the completion of this project. As for now, its time to do some Quantum Mechanics for homework...


Friday, January 22nd, 2010

The spring semester begins Monday, and I can definitively say that this winter break was the most productive one since I've been in college. Unfortunately, the project is still not 100% finished, but that means that I can continue coming into the LTC and learn a few more things about optics. The basic setup of the thermally stabilized laser is finished. You can see some photos here. This past week was mainly spent making the heater (200 feet of 30AWG wire which had to be hand-spun into 100 feet of bifilar = lots of fun!) and soddering the last of the connections between the circuit and heater and photodiodes via the yellow ethernet chord as seen in the photos. After all of this had been done, Dr. Cohen and I spent a couple of hours on the oscilloscope exploring how the circuit worked (will get back to you later on that) and discussing how we would check that the controller was indeed stabilizing the laser. It looks like we will check that with a Fabry-Perot interferometer that Laser Sam has left in the LTC. Its another piece of equipment that I'm going to have to do some reading about before this project is finished.


Friday, January 15th, 2010

Today I made a third attempt in modeling the effects of a PBS on horizontally and vertically polarized light. After today's measurements, I think I can make a decision on which PBS is best for the stabilization project. Yesterday, Dr. Cohen suggested I change the method by which I did the experiment. Instead of polarizing the light by rotating the laser tube until the polarization sheet yielded a maximum, I rotated it until it was a minimum. This is an improvement because the human eye is better at seeing a minimum, though either way, this point was verified with the photodetector. Another improvement was fixing the photodetector to the work bench for all measurements.

The PBS provided by Prof. Weinacht edges out the 830nm cube based on today's testing. The new set of excel sheets are here:

1)Unknown PBS provided by Prof. Weinacht.

2)The 830nm cube.

Professor Weinacht's cube transmitted 95% of vertically polarized light through output 1, and although the 830nm cube transmitted 98% through output 1, it allowed a little more of it to go through output 2. It also outperformed the 830nm cube in directing horizontally polarized light through output 2 at 88% to 80%, and allowed less through output 1. In reality, either PBS would probably suffice, and these models serve to anticipate any extra measures that would need to be taken in order to accomodate the experiment for either PBS.

At this point, much in the project has been completed. With the help of Dr. Cohen, I've been able to understand Laser Sam's circuit a little better, and I'm beginning to setup the stabilization experiment on the work bench for initial testing. However, I am still missing the wire needed to make the heater and the potentiometer to tune the controller. I hope getting these pieces are not as hard as finding a good PBS. Still plenty of work to be done ...


Thursday, January 14th, 2010

I'm still working on choosing which PBS to use in the experiment. Yesterday's Excel models were apparently incorrect. Today I repeated the measurements with the 830nm cube. Updated excel sheets are here:

1)Unknown PBS provided by Prof. Weinacht.

2)The 830nm cube.


Wednesday, January 13th, 2010

The last couple of days we were able to solve a couple of problems. First, some solder in the cathode wire became undone; this imperfection had been the cause of some flickering when we first tested the laser. This event afforded me the experience of soldering for the first time, which was fun, and now the laser appears to be working better than before with no flickering. However, that distraction prevented me from doing some more technical work which actually was a good thing because this morning Dr. Noe sent me a Microsoft Excel file which models the behavior of a PBS. The PBS is a crucial element in the stabilization experiment because the signals coming from its outputs determine the behavior of the controller circuit. In my last journal entry I detailed some measurements I made with a PBS made for 830nm wavelengths, which is not an ideal setup. Fortunately, Prof. Weinacht provided us with another PBS, which after today's measurements, looks to be the PBS we need! I used Dr. Noe's Excel file to plot my results. If I have done both the experiment and entered the data correctly, then this PBS should work well in the experiment.

Another lingering concern is that the waste beam of the laser will not be intense enough to satisfy the controller circuit. The desired output power is 25μW for each polarization that comes from the PBS. Since the new PBS is quite large, 1", the power from the waste beam languishes at around 7μW. Laser Sam has offered some solace: The 25μW requirement applies to a similar controller with a gain hi-pot of just 500kΩ. Laser Sam's controller has a hi-pot of 1MΩ, which can be used to increase the power of the signal. Furthermore, he assured me that even a small voltage swing should be enough to do the trick! In summary, the experiment should be able to go through; now we must begin to costruct the laser mount and the heater, unless, as is sure to happen, Dr. Noe points out another important bug that needs to be fixed.


Monday, January 11th, 2010

I've been itching to get to work with a beam splitter, but as I found out a few days ago, beam splitters can be tricky things. Some are designed to work for specific wavelengths, while others are more broadband. Unfortunately, the LTC does not have any for 632nm, but we do have some for 830nm. I spent last friday and today exploring whether the 830nm polarized beam splitters (PBS) we have are appropriate for use in the laser stabilization project. The basic setup I used is seen in this diagram.

The laser I used is a commercial Melles Griot Laser with a measured output of about 15mW (measured using a ThorLabs photodetector with a responsivity of about 0.4 A/W near 632nm). The polarizing sheet at 45 degrees is used to attain equal contributions of both polarizations of the laser. Then I oriented the laser (turned the laser tube on its axis) so that the intensity transmitted by the polarizing sheet was at its maximum (11mW).

Then I proceeded to measure the light transmitted by both sides of the PBS. One side yielded 4.7mW (point C on the diagram) and the other 6.25mW (point D on the diagram). Then I placed polarizing sheets between the PBS and the screens, points C and D once again, and measured the intensity of the polarized light. These were 2.9mW and 2.5mW at points C and D, respectively. Whether one uses these polarized or unpolarized intensities, it seems that about one polarization composes 53% of the total intensity while the other 47%. So even though this PBS is not designed for this wavelength of light, it may be useful. And even if it were not, now that I have seen how much these optical elements can cut down on a beam's intensity, it may be possible to use a non-polarized BS with polarizing sheets in the laser stability experiment. According to Laser Sam, the photodiodes in the controller circuit needs only 25μW at each photodiode to function. The setup with the polarizers at points C and D provided about 16% of the laser's original intensity at screens 1 and 2. The laser I am going to use in the stabilization experiment has a raw output intensity of 1mW, so 16% is 160μW, more than enough for the experiment. However, since one of the objectives is to use the less-powerful waste beam, this could be a serious obstacle.

Tomorrow, I plan to run take some more measurements with this setup, and then repeat it with a non-polarizing BS with polarizing sheets.


Thursday, January 7th, 2010

Its been a long layoff since the last time I was in the LTC. There were finals and the holiday break and other distractions. But I'm glad to be back and eager to see this project to completion. After reading my previous journal entry, i'm glad to note that I have a better grasp of the project. Yesterday Dr.Noe and I were able to start the laser, which led to a discussion of the power supplies needed for both the laser and the controller circuit. The laser power supply needed between 21 and 31 volts which was supplied by a bench power supply set to 25.7V. The circuit needs 12VDC at a maximum of 1A regulated power according to Laser Sam's instructions so we will have to look further into this matter. Since the objective is to have all the parts of the project (laser tube, circuit, beam splitter, power supplies, etc.) contained in a single unit, we would need some sort of power cube for the circuit. However, most conventional power cubes have only 2 prongs (i.e. they don't provide a ground) so that the DC is "floating" which may not be appropriate for the circuit.

Yesterday was also productive because we began looking at designs for the unit that would support all of the parts. So far I am leaning towards an aluminum surface, about one inch in thickness around 15" in length by 10" or so in width (the actual dimensions to be worked out later). The laser tube with the electrodes is about 10" in length, and we will need extra length on the platform for the beam splitter and photodiodes. the laser power supply takes up a maximum space of 4" X 2". The circuit PCB is less than two square inches; however, Dr.Noe suggests buying a power cube for it and then making a box containing both it and the PCB. The box would also need space for a potentiometer to control the offset of the circuit (the offset helps "tune" the modes underneath the gain curve). With this added feature, the size of the platform may need to be changed. We also have to decide what type of power supply we are going to use to power the laser's power supply, so we may need room for that as well. One last consideration we had about the platform was adding jack screws to the corners in order to adjust the height of the whole platform in order to make it useful in experiments with common optical devices which can be at a height of several inches. Another concern is that the laser tube is only about 1/2" off the ground with the electrodes attached, which may not be enough clearance for the magnet wire and any insulation we place on it. The solution favored at this moment is milling into the surface of the platform to increase this clearance.


Thursday, December 1st, 2009

After waiting a couple of weeks for the experimental equipment to arrive from Laser Sam, it becomes apparent that I did not do enough preparation for the project. The circuit I was going to use was explained in a paper in the American Journal of Physics which consisted of a few transistors and op-amps, but the one that arrived was much more complex. Happily, Laser Sam provided one that was already constructed, but I had trouble grasping exactly how it was to be coupled to the HeNe laser tube.

First I should explain the experiment, or atleast how I understand it to date. Laser tubes can produce multiple longitudinal modes as long as they are within the gain bandwidth. These modes are separated in frequency space by the relation c/2L, where c is the speed of light and L is the length of the tube. As the laser heats or as ambient temperature fluctuates, the length of the tube can change, affecting mode spacing, thus resulting in mode sweeping. This means that the positioning of the modes underneath the gain curve changes, resulting in the chaging of the output intensity of the laser, which is undesireable.

There are many methods to fix this, and the project I will do is to explore the method of thermal feedback, which to my delight involves some electronics. The electronic setup basically compares the intensity of the laser to the desired intensity, takes the difference or "error" and gives this signal to the heater portion of the electronic system. The heater control circuit leads to a coil of wires wrapped around the HeNe laser tube and seeks to stabilize its length by stabilizing its temperature.

That is my qualitative understanding as of now. What I should strive to understand next is the different working components of the electric system (where is the heater and how does this connect to the coil around the HeNe tube?; where does the circuit receive the intensity inputs?; how do I interface it with a computer so that I get graphical data?).

As for the second question, I think that the circuit receives the intesity inputs via photodiodes (photodetectors would also work, but may be unnecessary). But there is a twist. Laser Sam provided two photodiodes. It turns out that the laser tube I have lases tow modes, each with a different polarization (orthogonal to each other). Dr. Noe and I were exploring this concept the other day by shining a laser through a small calcite rock. The calcite transmitted two beams with different polarizations. This was verified by placing a polarizer between the calcite and the screen. It was seen that when the polarizer was turned to only transmit one beam, the other's intensity was severely diminished. Since the HeNe produces two polarizations, a photodiode is needed for each one. So at the output of the laser, I think we will need a beam splitter and then polarizers oriented at the splitter's output to select which polarization enters which photodiode.

A list of supplies we need for the experiment: Magnet wire to wrap around the tube, beam splitter, polarizing sheets, power supply for the laser and other things that will surely pop up...


Ewuin Guatemala
January 2010
Back
Laser Teaching Center