Tuesday, April 25, 2017

On the nature of emitted light, Part 3

I said in a previous blog post that I wanted to talk about fluorescent bulbs. I do. Really. And I will... in this very blog post. But before we get to that dessert, we need  to eat our peas and carrots. Let's talk about fluorescence.

The soon-to-be-famous sunburn analogy

I am of Northern European stock. I sunburn easily. Naturally, I wound up in a place where the Sun doesn't shine. Milwaukee. On those rare occasions when the Sun does shine, I absorb ultraviolet light. Later, my skin emits red light.

Lobster, anyone?

That's fluorescence.

Well, not really. I do absorb UV, and my skin does turn red. But that red is a reflective red, rather than a emissive red. My skin doesn't actually give off light. Factoid: sunburnt skin is red due to the increased concentration of hemoglobin at the surface. Hemoglobin absorbs bucketloads of light in the OYGBIV part of the spectrum, and reflects some at the R end. The reflected light is thus comprised chiefly of red light so skin looks red when we burn. (Interested in more about the color of human skin?)

Just in case you were wondering, my normally pasty-white Anglo-Swedish skin matches 2R04 in the Pantone Skintone guide.

If I recall correctly, though, I was talking about fluorescence. My explanation about sunburn shares a lot of the features of fluorescence. Light is absorbed at one wavelength, and is emitted at another wavelength. It is always emitted at a wavelength with less energy, which is to say, at the more relaxed higher wavelengths. For some molecules, the absorbed light is in the UV, and the emitted light could be at the red region of the spectrum. 

My understanding of the fizzicks involved

This will thankfully be a short section. I dunno nothin' about the fizzicks behind fluorescence. I mean, a molecule absorbs a photon, and that photon "kicks it up into a higher energy state". I have no clue what that means. I just know that I don't want to be around when my wife gets kicked up into a higher energy state.

Happy little benzine molecule

Later, the excited molecule gives up that energy, but not all at once. For some reason, it only gives it up a parcel at a time. Hence each fluorescent emission is at a lower energy (higher wavelength) than the excitation.

Note that I said molecule, and not atom. In the last post, kicking an atom up into a higher energy state was all about the orbits of electrons. Now it's about molecules. Surely that's a clue about what is happening when something fluorescences. But I am pretty ignorant when it comes to all that chemistry stuff. I'm the guy who once looked for a quantum mechanic to fix my compact car.

If I really understood any of this stuff, I would explain that
in this diagram from Kurt Nassau's book,
the wavy lines represent fluorescence

But in the spirit of pretending I know something...

There is a closely linked phenomenon called phosphoresence. Actually, it's the same phenomenon with a different name. Light is absorbed and is later emitted at a higher wavelength. The only difference is in how much later the emission happens. If it happens on a time scale where we don't notice (like nanoseconds or milliseconds), it's called fluorescence. If the delay happens on a time scale that we notice, for example if the fluorescent emission continues for seconds or hours after the excitation goes away, then we call it phosphorescence.

The distinction between fluorescence and phosphorescence is thus strictly anthropocentric. Just like the distinction between electromagnetic radiation and light (described in a previous blog post), the distinction is along a continuum and is not based on anything physical other than our meager, pitiful senses.

Examples of phosphorescence

Back in the olden days, engineers made a lot of use of phosphorescence. Cathode ray tubes (CRTs) in electron microscopes and in radar systems had long-persistence phosphors so that the image stayed latent on the tube long enough for us to notice. Quick show of hands... how many in the audience have used one of these devices?

A vintage scanning electron microscope (left) and a vintage radar tube (right)

Ok... let's try to open this up a bit. Show of hands again. How many in the audience have been to a historical museum (or my basement, same thing) and saw a TV that was two feett deep and weighed more than a pregnant and cross-eyed mule? That, my friend, was a cathode ray tube display, with a phosphor on the inside of the display end of the tube. Then again... maybe that was more accurately called a fluorescor, since we definitely didn't want it to persist for more than a 30th of a second.

(By the way, a scanning electron microscope is nearly identical in structure to the cathode ray tube in a television set. In fact, the cathode ray tube display was invented along side the scanning electron microscope. Someday I will blog on that topic.)

Certain lichens and mushrooms will glow in the dark long after the sun has gone down. Perhaps there is an evolutionary advantage to being seen as a part of the night life of the forest? I dunno. When it comes down to it, being a visible part of the night life has never given me much of an evolutionary advantage. It usually kicks my wife up into a higher energy state.

Some minerals fluoresce like an Anglo-Swedish color scientist with sunburn. Party-loving minerals like fluorite come to mind. I wonder where it got that cool name? Come to think of it, where did phosphorous get it's cool name?

Welcome back to the 60's

You can also see phosphorescence if you look at a fluorescent bulb in the dark, just after it has been turned off. Here we see the vague distinction between fluorescent and phosphorescent. Some of the stuff inside the tube is fluorescing, and some of it is phosphorescing. But more on that when I finally get around to discussing fluorescent bulbs.

But by far the most useful application is the little rubber duckie that was sitting on my wife's desk, at least until I absconded with it for a photo shoot. The rubber duckie is impreganted with some phosphor with excitation in the violet to blue part of the rainbow, and emission in the green to yellow part.

You can't claim to be uber-cool until you have one of thes on your desk

Examples of fluorescence

Certain versions of the Pantone guide had a few cards with the ever-popular 800 series inks. These inks all have fluorescent properties.

Picture of my 2005 Pantone guide

One of the more hip of these colors is Pantone 804, which is the orange ink. I almost called this dayglo orange, but that would be a misuse of the word, since Dayglo is a company. They make Dayglo pigments.

To demonstrate the phenomenal fluorescent properties of Pantone 804, I set up my spectrometer, my camera, and dug out my red, green, and blue laser pointers. (Note the repetitive use of the first-person pronoun my. It's all about me. Even when it's not, it's still about me.)

Here is what happens when I point the green laser pointer at Pantone 804. There is a strong peak in the green, at 546 nm. This is the reflection of the light from the laser pointer. But note the broader spectral stuff that appears from 560 nm to 700 nm. Lasers only put out a very narrow range of wavelengths. The broader peak must be fluorescence.

Now have a look at the spectrum emitted when the blue laser pointer is swapped in. The laser wavelength appears way far to the left, tucked away nicely at 390 nm. Then there's a broad peak that looks a lot like the broad peak in the previous spectrum. The excitation wavelength has changed, but the emission spectrum has not. 

Or at least the fluorescent emission spectrum hasn't changed a lot. Have a close look at the region from 460 to 510 nm. We see another bump. Not a big one, but a bump all the same. Why didn't this show up in the experiment with the green laser?

The explanation can be found above. I don't mean in the Heavens, but earlier in this blog post. I wisely said: "each fluorescent emission is at a lower energy than the excitation." The green laser just didn't have the gumption to excite emission in the blue part of the spectrum.

This should help us explain the frankly quite boring results with the red laser pointer that are shown below. We see a red peak, which is way up at 688 nm. Ho-hum. Any fluorescence would have to be above that, so we get bupkis in the way of fluorescence.

Fluorescent whitening agents

How about another example? I have previously touched on the topic of additives that make paper whiter. I mentioned it in a blog about color-related standards in the print industry. I also blogged about how different spectros deal with the problems caused by the whitening stuff. And I blogged about a conference with a sub-conference on the little buggers.

The image below shows the reflectance spectrum of one paper stock. You might notice something a bit peculiar about it, especially around 430 nm. Go ahead. Have a look. And take note of the scale on the left-hand side. The observant reader will have noticed that over 120% of the light that hits the surface is reflected back. For the mathophobes in the crowd, 120% is more than 100%. So... this paper is creating light?

Note the attractive little bump at the blue end of the spectrum

So, here's the scam. Paper normally looks like a brown paper bag. You can make it whiter by various means, including bleaching it, but that's expensive. Not horribly expensive, but there is cost involved. And people like their paper to be white. In fact, studies have shown that people prefer paper that is just a tad on the blue side of true white.

A cheaper way to get white (and the only way to get blue) is to add fluorescent whitening agents to the paper. There is a family of compounds known under the name of stilbenes. Below is the excitation / emission spectra of stilbene stolen from a TAGA paper by Dr. David Wyble and some Anglo-Swedish guy who likes to think of himself as a color scientist. The blue line shows the amount of energy that the stilbene absorbs, as a function of wavelength. Note that this is in the UV region, mostly all between 300 nm and 400 nm. The red line is the wavelengths where that energy is fluorescently emitted. Pretty much what we would call the blue region of the spectrum, from 400 nm to 500 nm. 

Yest'day I's fluorescin', and today, I still-been fluorescin'

Adding stilbene to a paper stock will boost the blue. Since drab, dull, yellowish paper is blue-deficient, this will make it look whiter. Well, provided there is some UV light to get it excited. Paper is not creating light, it's redistributing the energy from the UV to higher wavelengths.

The image below illustrates that. There are three sheets of paper here. I wrote on them, annotating the amount of FWAs. On the right side, I took a picture of the three sheets under regular old garden-variety light. The three look similar. On the left we have a picture of those same three sheets under a UV flashlight. OMG! It is pretty obvious that there is some sorta difference going on! 

Three sheets to the fluorescent wind

BTW, FWA AKA OBA. Someone got the bright idea to call these brighteners OBAs. This stands for Optical Brightening Agents. I agree, the term fits. Stilbene brightens paper optically. But so does bleach, calcium carbonate, and titanium dioxide, and a good coat of white paint. These four will all increase the reflectance of paper in the blue region. But only stilbene does it with a fluorescent flair. So, if you hear someone call stilbene an OBA, wag your finger at them and tell 'em John the Math Guy says that they are using the term improperly.

Remember back when I took note of the little bump in the spectrum when I used the blue laser pointer? You may have guessed by now. It was stilbene. The paper that the Pantone book is printed on has quite a bit of FWAs. It's kinda hard to find paper today that doesn't.

Well. Look at the time! It's about time to wrap up this blog post on the nature of emitted light. Today I taught you everything I know (and a little bit more) about things that fluoresce in the night. There was something else I wanted to say about fluorescent light... Can't remember what it was. I guess it can wait until the next blog post. That one will be about fluorescent bulbs. I promise.

Wednesday, April 12, 2017

On the nature of emitted light, Part 2

I'd like to talk about one of the most ubiquitous light sources, the fluorescent bulb. I mean, not only are they ubiquitous, they're all over the place. And at least until the recent new wave of LED illumination, they were the number two light source that tried harder. (They try harder than the #1 light source, incandescent, which was featured in my last blog post.) And for those of you who are excited by viewing booths (and quote frankly, who isn't?) I'm sure you have been just chomping at the bit, waiting for a blog post about fluorescent lights, since almost all light booths use fluorescent bulbs.

As I said, I'd like to talk about fluorescent bulbs. But I have to talk about a different sort of light emitting thingie first. You see, florescent bulbs are kinda complicated. There is a combination of two physics thingies going on: gas excitation and fluorescence. Today's blog will be about gas excitation. If that phrase caused you to snicker, then ... well, so be it.

Neon bulbs

The simplest gas excitation bulb is the neon bulb. You start with a couple of electrodes close together, but not touching. You form a glass bulb around them, and squirt in a tiny amount of neon just before you seal it. Maybe you add a tiny tiny amount of argon as well. Now, you put a high voltage across the electrodes (at least 50V, but likely 110V). Lo and behold, a faint orange glow appears.

Neon bulbs have gained popularity as indicator lights. A recent Rasmussen poll put their popularity somewhere just above that of Mel Gibson. Why are they so popular? First off, they're cheap. You can buy a handful of these little puppies for about a dime apiece on Amazon. Second, they are very simple to hook into a device that plugs into household current (110V AC). All you need is a current limiting resistor, which is included in your investment of one thin dime on Amazon. Third, they draw a tiny amount of power. You would need about 1500 of them to draw the power of a 60W bulb. Fourth, they put out a pleasing warm glow that is very effective at telling someone that the power strip is live, that the soldering iron is on, or that the circuit is live.

A collection of indicator lights chosen to subliminally convey my machismo

Getting your knee on

I have no idea what is mean by the title of this section, but it has some sort of cool vibe. As does neon. I mean, it is one of the noble gases! This prestigious group of elements includes helium, neon, argon, krypton, xenon, and radon. Helium, of course, is the party gas, since it makes us talk funny. Krypton is so cool that it has a fictitious planet named after it, and it is so powerful that it makes Superman cower. And radon? What safety conscious household doesn't have a radon detector in its basement? Truly this is a noble group to belong to.

The group is characterized as those elements which have a full outer shell of electrons. (As you know, you don't want to be that guy who is one electron short of a full outer shell!) This means that they are inert, very reluctant to react. As a result, they don't get invited to many pep rallies or often get selected as game show contestants. But they do get selected for applications where engineers are trying to avoid chemical reactions. Such as light bulbs that are hot and that we want to last a long time. Argon's senior picture has the caption: most likely to be selected to make an appearance inside an incandescent bulb.

The shell game

I'm gonna start with a quote from Wikibooks, under the heading "General Chemistry, Shells and Orbitals": "Each shell is subdivided into subshells, which are made up of orbitals, each of which has electrons with different angular momentum." As I was going to Saint Ives... I sure wish I could talk purdy like that. Honestly, I have no idea what this means, but nonetheless, I will give my explanation.

Imagine a guitar string. It has a certain resonant frequency. Like, the G string will vibrate easily at around 200 Hz.(I am tempted to throw in a joke about how I frequently resonate with G strings, but that would be totally inappropriate. So I won't say anything.) This is a natural vibration mode for the string, where the whole string is moving back and forth the same way.

The G string will also vibrate at one octave above 200 Hz, around 400 Hz. If you were to watch a high speed video of the string at 400 Hz, you would see that the center of the string is not moving, and that the right and left side of the string are moving opposite from each other. Similarly, the G string has an affinity for vibrating at 600 Hz, where there are two points on the string that are immobile. This third mode of vibration is shown below. The astute reader will recognize this concept from a blog post of mine from almost exactly three years ago on the vibration of piano wires.

G string vibrating at 600 Hz

Atoms are like guitar strings. (I just googled that sentence, in quotes. Google is not aware of that sentence ever having been typed before. High fives all around! Just wait until next week!!) Just like a G string doesn't take kindly to vibrating at 260 Hz, the electrons that orbit an atom only exist in certain energy states. (Oh yeah. I forgot to mention that each frequency has a different energy level associated with it. It takes more energy to get something to vibrate quickly, so the higher the frequency, the higher the energy level. Each energy state corresponds to a specific frequency/wavelength.)

So, you got this atom. Let's get just for example that an electron in this particular type of atom can be at energy states of 13 banana units, 15 banana units, and 20 banana units. An electromagnetic field induces the electrons way up to the 20 banana unit state. Eventually, the electrons will grow tired of hanging around up there, and they will drop down to another state.

If they drop down to the 15 banana unit state, they will lose 5 banana units of energy. Since energy is conserved, a little packet with 5 banana units of energy needs to be spit out. It gets spit out as a photon with 5 banana units of energy. Since energy and wavelength are related, this photon proudly moves to its proper place in the rainbow - the location that corresponds to 5 banana units of energy.

If an electron drops all the way down from the 20 to the 13 banana unit state, it will lose 7 banana units of energy. Now we have photons that are at the 5 and 7 banana unit locations of the rainbow.

There is one other possibility - an electron that dropped to the 15 banana unit state could drop a second time and wind up at the 13 banana unit state. Hence we also see some photons in the 2 banana unit state. A third position on the rainbow.

Going through the possibilities, we can expect there to be photons at three discrete positions (that is to say, wavelengths), corresponding to 2, 5, and 7 banana units of energy, as illustrated below.

Monkeys falling from tree branch to tree branch
The size of the yellow circle represents how loud of an uf-da the monkey makes

Emission lines

If you were looking for the section on transmission lines, I suggest you might want to check out a different blog. On the assumption that you are actually interested in how all this orbital decay stuff ties into neon bulbs, then read on.

Based on this business about discrete energy levels leading to discrete energy levels for the emitted photons, we kinda expect that the spectral output of a neon bulb to be equally discrete. Here is my expectation, based on some website somewhere that looks like it's reliable. They use big words, anyway.

I got put my ultra-sophisticated spectrometer, for which I paid about two years' salary, and put one of my neon light sources in front of it. The spectra below shows what I saw. Strong peaks, but not really the very narrow lines that we might expect. I am going to blame that on my spectrometer. Although it reports every nanometer, the spectral resolution is around twelve nanometers. 

My spectrometer looks at neon

Tech note: There is a spectral blur in any spectrometer that has to do with a design trade-off. Most spectrometers require collimated light, which is accomplished by focusing light on a slit aperture. The narrower the slit, the finer the spectral resolution, but also the smaller the amount of available light. Less light means either longer integration time or more noise.

Actually, a neon bulb can be used to measure the spectral resolution of a spectrometer. I looked through the data to find the wavelengths on either side of the peak where you reach 50% of the max: 579 nm and 592 nm. The difference between these is the FWHM resolution. FWHM stands for "Full Width at Half Max".

Do my peaks line up with the advertised values?

Mine Theirs
585 585.2
612 609.6
637 640.2
702 703.2

Actually, I am rather impressed. The two peaks in the official-looking plot which are most isolated (585.2 nm and 703.2 nm) are almost right on the money. 

But why are the others off? The key is that we need isolated peaks to test for correct placement of emission peaks. Because the resolution of my spectrometer blurs the spectrum, several peaks got averaged together, and so the center got shifted.

Another tech note: This is the technique used to calibrate spectrometers. Typically, the factory calibration lab will have a set of  gas discharge lamps such as neon, but also maybe krypton, xenon, argon, and/or mercury.

How about doing color matching under neon bulbs?

Neon bulbs are very efficient and inexpensive. Individually, they don't emit a whole lot of light, but they're small and cheap. Presumably, I could wire up a gazillion or so of these to make a really groovy light booth for evaluating color. And since we know the spectrum so accurately, it should make for really accurate evaluation of critical color, right?

The short answer is no. And the long answer is "good golly gosh, no!" Take another gander at the spectral emission plot of the neon bulb. Note in particular what we see happening below 570 nm. Nothin'. Virtually no light at all.

Consider a yellow ink. Above 550 nm, it looks almost indistinguishable from the paper it's printed on. So, I would argue that a neon bulb color booth is about as useful as Braille on the keypad of a drive-through ATM.

Yellow ink 

They have a measure that is an index of how good a light source is at properly rendering color. Ironically, it's called the Color Rendering Index, or CRI for short. The color rendering index of a neon bulb is zero. That's on a scale from 0 to 100. So, kinda not so good.

A few similar bulbs

You know those orangey-yellow lights that are used for street lights? High-pressure sodium vapor lights, also known as HPS by the cool people. Not to be confused with high-pressure sodium vapor light salesmen, who tell you how great the bulbs are cuz they are very efficient.

But have you ever tried to find your car at night in a parking garage with these kind of lights? I bet that high-pressure sodium vapor light salesman never told you that the HPS lights have a CRI of 20. Color is greatly distorted.

You'll never guess what gas is used in these puppies!

How about those really bright bluish-white lights that are used as security lights, as overhead lights in high-bay factories and stores, and as floodlights in a stadium? Those are likely to be metal halide bulbs. As with all the other gas excitation bulbs, these have a gas and a high voltage which causes electrons to jump around to different energy levels, giving off light at specific discrete wavelengths. Theses bulbs come in at a whopping 54 CRI.

Rock concert? Make sure you get the heavy metal halide floodlights!
By the way, just to make sure I am not misunder-terpretted, getting a score of 54 on a 100 point test is not so good. Of the bulbs in this blog. the metal halide bulb does the best job of making colors look right, but note that even metal halide is kinda short of energy on the red end. That's where it lost a lot of points on the CRI test. But, I should point out that the spikes don't help a lot either. 

Next time, I'll talk about fluorescent bulbs. I promise.