# Gas station without pumps

## 2014 January 9

### Second day of freshman design seminar

Several students had looked for interesting ideas for projects on the web, but only one had sent me the URLs.  I asked the students to send me annotated links by Friday, so that I could put them into a web page over the weekend—they’ve now started to trickle in.

I started the class by going around the room getting a project idea from each student, which we discussed very briefly.  Several had come up with interesting projects, and a couple had come up with ones that seemed a bit too big for a 2-unit freshman course.  There was one that I thought was too ambitious even for a senior deign project, but I forget what it was now.  Two students came up with the same idea—a home-made centrifuge.

I then started a design exercise, which we’ll probably continue next week.  The exercise started with an idea—a spectrometer.  The exercise almost foundered right there, as almost no one in the class had ever heard of a spectrometer or a spectrum.  I ended up drawing a few colored lines on the board, which several students recognized as being spectral lines that they had seen in a chemistry or physics text book.  I then gave them the idea that one could plot the intensity of light as a function of wavelength.

Exercise 1: 2-minute writing, tell me what a spectrometer does.

Exercise 2: 2-minute writing, tell me what the inputs and outputs of a spectrometer are.

Exercise 3: 2-minute writing, tell me what a spectrometer might be used for—list uses cases.

A lot of the students were very frustrated by this assignment, as they had no idea what I was asking for or what a spectrometer might be used for.  That surprised me somewhat, as I thought that spectroscopy would have been covered either in chemistry or physics, which everyone had had in high school.

I then had students get into groups of 3 and share their results.  In some groups, there was active sharing and people seemed to be getting a clearer idea of what a spectrometer did.  Other groups seemed a bit dead, with no one having anything to say.  After a few minutes of this, I asked each group to report one thing from their discussion.  The first group reported something trivial (the input is light), the second reported something subtly incorrect (the output is the wavelength of light), and the third group echoed the first two.  I then picked on the third group, asking them to clarify the vague response.  [I had decided in advance to pick on the third group, whoever they were, unless they gave an awesome response.]  After several questions to the group, and increasingly vague answers, I finally got an answer that we could build on (that the output was an array).  But then they couldn’t clarify what the subscripts of the array were nor what the data in the array were.

I moved on to the third group and asked them to clarify what the outputs were.  After a fair amount of unclear description, they finally came out with the output being a graph, but once again they couldn’t say what the x and y axes were.   (I had a graph on the white board behind me, with the x-axis labeled with wavelength and the y-axis labeled with intensity, so I thought I was asking freebie questions.)  Eventually I gave up, as the frustration level of the students was rising, and not in a productive way.  I then explained the notion of the intensity of the light being a function of the wave length.

The fifth group surprised me in a good way, when I asked them for a statement.  They managed (in less clear wording) to come up with the idea that one may be measuring emission from a light source or absorbance of light.  This came up in the form of a use case—measuring UV in sunlight and determining how well sunblock blocks it.  I gave them the terminology.

Next I did a demo of a diffraction grating deflecting light, using 3 laser pointers (red, green, and violet).  I managed, by having several hands of students helping, to get all three lasers pointing through the 1000lines/mm diffraction grating at once at almost the same spot on the projection screen.  The diffracted spots then showed up nicely spaced on the screen, with the violet (405±10nm) deflected least, the red (650nm) deflected most, and the green 532±10nm in between.  Something I had not noticed before was that the green laser did not produce a single spot, but three distinct spots.  More on that below.

After the diffraction demo, I had the class break up into a different 5 groups of 3 and go to the white boards to start designing ways to convert what they now knew about diffraction into components for a spectrometer.  This exercise did not go well—I’ll need to scaffold it better next time we do it.  Students were stuck either on the output they wanted or on the physical phenomenon of diffraction, but no group came up with any connection between the two.  I think I tried to cram too much into the 70 minutes—I’ll revisit the spectrometer design problem after talking a bit about how one decomposes a design problem into subproblems.

In the meantime, I suggested that students look at Wikipedia articles about spectrometers, though the main article looks like it was lifted from a 30-year-old encyclopedia. The Spectrophotometry article looks better written.  I hope some of them think to Google the key words to find better explanations elsewhere.  Those taking physics may think to look in their physics texts, even if they are not taking the optics part of the physics course.

### Three dots

I took a photo at home of the three spots (a bit smeared, because I was holding the diffraction grating with one hand and operating the camera with another). I tried again with the camera on a tripod and the diffraction grating in my Panavise Junior, but it did not come out any better, even after fussing with the exposure settings.
The lowest, brightest dot is the least deflected.

I don’t yet have a good explanation for why there are 3 dots for the green laser. They are clearly different diffractions and not just distortions of the laser spot, since they rotate with the diffraction grating and not with the laser.

I looked up green lasers on Wikipedia, and found out that the 532nm green laser pointer is probably a diode-pumped solid-state laser consisting of an 808nm IR laser diode, whose output is converted to even longer-wavelength 1064nm IR light by a neodymium-doped yttrium orthovanadate crystal laser, and then frequency-doubled in a non-linear potassium titanyl phosphate (KTiOPO4 or KTP) crystal to get 532nm output. I don’t see any mechanism there for producing other wavelengths, unless the 1064nm laser is generating more than one wavelength.

I wonder if the multiple spots are not from multiple wavelengths but from reflections off the front and back of the diffraction grating, resulting in different amounts of diffraction. If that were the case one would get multiple dots for the red and violet also, though the dots may be dim enough that we only see the brightest one.

Nope, that explanation doesn’t work—as the green laser runs for a while the extra spots disappear, but blowing on the laser to cool it brings them back. The extra spots are temperature-dependent!

## 11 Comments »

1. I assume the 3 dots are from multiple diffraction orders, thought the blowing/cooling thing is weird. It’s possible the laser is getting dimmer (and the higher orders are getting really dim) and your logarithmic eye can’t see the dim ones any more but can’t tell that the bright one (the first order) is also dimmer.

Comment by Andy "SuperFly" Rundquist — 2014 January 9 @ 05:36

• No, I don’t think that these are the multiple diffraction orders, as the spots are quite close. The second-order diffraction spots are visible with the violet laser, at about twice the deflection of the first order spots, but those angles are huge. We do not see second-order spots with the green or red lasers, because the grid spacing is less than 2λ.

We are using a 532nm laser with a 1000nm diffraction grating, so λ/d = 0.532. According to http://www.calctool.org/CALC/phys/optics/grating, the first-order diffraction should be 32.1407° and it won’t calculate the second-order one. With $d \sin theta = m \lambda$ for an order-m diffraction we get $\theta = \sin^{-1} m \lambda/d$, solutions only exist for m=0 and m=1. If you want to show second-order diffraction, you want to use a 500-lines-per-mm grating rather than a 1000-lines-per-mm grating (or use a violet laser).

Comment by gasstationwithoutpumps — 2014 January 9 @ 08:22

• then I guess I like the notion of some multiple reflections going on, with my comment above about the other ones getting dim. I’m not sure where the multiple reflections are happening, though

Comment by Andy "SuperFly" Rundquist — 2014 January 9 @ 08:38

• I considered the possibility that there were multiple reflections off the front and back surface of the diffraction grating, but that seemed inconsistent with the changes in the spot intensity as the laser warmed up or cooled down. It isn’t heating or cooling of the diffraction grating, as moving to different locations on the grating does not change the spot pattern.

Given that there are 3 active optical elements in the green laser (the IR laser diode, the pumped IR laser, and the frequency doubler), I suspect that one of them has something else going on that is not covered in the simplified explanations I’ve found on the web.

Comment by gasstationwithoutpumps — 2014 January 9 @ 10:05

2. High power lasers can have multiple lasing modes very close together in wavelength, these are the longitudinal modes, and the spacing depends on the length of the cavity. http://olympus.magnet.fsu.edu/primer/java/lasers/gainbandwidth/laserfigure1.jpg

Temperature can cause the wavelength to drift a few tenths of a nm per degree, as the laser heats up you might be changing which of the lasing modes get the most gain (brightest).

Mode shape can also depend on the geometry of the waveguide. Broader, less confined cavities can also have multiple modes visible in the far field pattern in grad school we referred to these as lateral modes.

For sure the violet and blue laser pointers are direct emission from InGaN-based diodes in a Fabry-Perot geometry, which are likely single mode. The green is, as you said, frequency doubled with integrated optics and probably much higher power output so I’m not surprised you see multiple modes. I’m wondering if a red laser pointer (also higher power) would show the same thing.

This website has some useful info, not many images though.
http://www.rp-photonics.com/categories.html#lasers

Comment by KK PhD — 2014 January 9 @ 20:29

• I did not see any multiple dots with the red (650nm) or violet (405nm) lasers. Only with the green (532nm) laser. These were all <5mw laser pointers.

If I'm understanding you correctly, you are suggesting that the 1064nm IR laser is multimodal. (I'm assuming that it wouldn't matter if the pumping diode were multimodal.)
I played with the applet on http://olympus.magnet.fsu.edu/primer/java/lasers/gainbandwidth/index.html and saw that it is possible to get multiple modes that are not symmetrically spaced around the center, depending on how the cavity is tuned.

I'm seeing the most intense mode being the one deflected least, so the shortest wavelength. That is also the only one present when the laser is started completely cold or after the laser warms up.

I tried measuring the spots on my wall. The distances from the undeflected spot were 54, 54.8, and 55.6 cm. Assuming that the first spot was 532nm, then the wavelengths were 532nm, 540nm, and 548nm.

If the spots correspond to different modes and different apparent cavity lengths, we'd expect $\lambda_{N} = \frac{2L}{N}$ and
$\frac{\lambda_{N}-\lambda_{N+1}}{\lambda_{N}} = \frac{1}{N+1}$. For the measurements I made, the modes would be around N=68, which would imply that the cavity is only 36µm long. Given that the laser is a neodymium-doped yttrium orthovanadate crystal, this seems a little small to me. If the Wikipedia picture is to scale, I would expect the crystal to be about 10 times that size. (Maybe my calculations are off?)

Comment by gasstationwithoutpumps — 2014 January 9 @ 22:01

3. Very interesting post. I’m somewhat surprised also, if they hadn’t run across spectroscopy in high school, or at least emission spectra. Probably it doesn’t have much impact at the time and is quickly forgotten.

Comment by profbillanderson — 2014 January 9 @ 20:56

• Most of them had encountered emission spectra, as they recognized my line drawing of a red, green, and blue line as something they had seen. But the words “spectrum”, “spectrometer”, “photospectrometer”, and “spectroscope” all seemed to mystify them. They also had no idea what the axes of the plot should be. Well, they’re freshmen—they’ll learn. I’ll just have to adjust my pace and my expectations a little to match what they know—I’m doing this class to get them excited and knowledgeable about engineering design, not to judge what they were taught in high school.

When I stumble across results I didn’t expect and get to learn something new (like how green laser pointers work, and why there are multiple wavelengths at the output of the green laser), that’s just one of the perks of being a teacher.

Comment by gasstationwithoutpumps — 2014 January 9 @ 22:08

4. […] we continued the design exercise I started last week, designing a […]

Pingback by Fourth day of freshman design seminar | Gas station without pumps — 2014 January 16 @ 00:26

5. […] we continued the design exercise I started last week (see also 4th class) designing a […]

Pingback by Fifth day of freshman design seminar | Gas station without pumps — 2014 January 22 @ 20:17

6. […] Second day of freshman design seminar […]

Pingback by Twelfth day of freshman design seminar | Gas station without pumps — 2014 February 19 @ 23:12

Create a free website or blog at WordPress.com.