The High Cost of Doing Physics

Yesterday, I spent $52 (plus shipping) buying sand. Not a gret big sack of sand, either– just 200 grams of it. I count it as a bargain, too, because I was prepared to spend twice the amount for half as much.

Now, granted, the $1000/kg sand is extremely high purity silicon dioxide, designed to be used in putting high-quality coatings on optical elements, and I would’ve bought that if it hadn’t been back-ordered. The cheaper stuff is slightly lower purity– 99.9% instead of 99.95%– but it ought to work. And they had it in stock at Aldrich, so I decided to take a chance, and save $50 in the process.

But still, I’m paying $260/kg for sand.

This is a little more absurd than the usual scenario, but weirdly high costs are part of doing physics these days. When you’re looking for specialized devices to do very specialized tasks, you end up shelling out a lot of money for things.

Explaining the costs of things is one of the funnier parts of introducing research students to the lab. I had a student a couple of summers ago who shorted out a diode laser he was working with, and was really apologetic about it.

(Continued below the fold…)

“Don’t worry about it,” I said, “It was only about a hundred bucks.” His eyes just about came out of their sockets when he heard that, so I had to explain that that’s pretty cheap for a diode laser. I’ve got a bag full of diodes that were $10 each, but they’re basically useless to me. The ones that are useful cost between $100 and $500, and I’ve got two that were $675 apiece.

And from the outside, they all look exactly like the lasers in a cheap CD player. “But, they’re so small!” my student said when I explained the price structure.

The current controllers for those lasers range from about $300 (for a bare-bones model) to a bit over $1000 for a fancier model with a built-in display and a lot of safety features (money well spent, given the way the power has flickered on and off over the years). I went with the $300 temperature controllers, because they’re not as fragile. Miscellaneous other parts probably average out at $50 or so, putting the full grating-locked diode system at $2,000-2,500, not counting the cost of the labor to build the mounts. The higher-power laser I’m injection locking comes in at $1,500-$2,000

And this is the cheap route. A commercial system to do the same thing would be between $5,000 and $10,000, depending on the manufacturer and the laser wavelength you’re after. The deluxe solution would do away with the injection-locked laser, replacing the whole thing with a titanium-sapphire laser system, which will set you back something on the high side of $150,000 for the Ti:sapph and a doubled YAG to pump it. I’m not sure of the exact price of that sort of system, as it’s just too depressing to get a solid estimate…

And that’s just the lasers. If you want to do anything with that light, you need to move it around the lab by bouncing it off mirrors, and a research-quality dielectric mirror in an adjustable mount will set you back $110, when they’re on sale. I’ve got something like 75 of those in my lab at the moment, and I always need more. Lenses, beamsplitters, and waveplates start around a hundred bucks each, and go way up from there. The posts and clamps and holders to keep all this stuff in place are pretty cheap, but still, it adds up.

And that’s not even touching the vacuum hardware.

I’ve spent something in the neighborhood of $150,000 building up my lab, and I had the advantage of starting with a number of big items already in the room (three large optical tables being the most important part of that). The most expensive single item in the room is a turbopump system that set me back about $14,000 (thank you, NSF), which is at least impressive enough that the students don’t scoff when I tell them the price. The rest of that is made up of lots of little, expensive items.

And I’m doing this whole thing on the cheap. A colleague who’s at a large research university on the west coast told me that he figured out he needs to raise $500,000 in external grants to keep his lab running. Every. Year. And that’s still AMO physics, which is the cheap, table-top stuff– everybody’s seen the eye-popping price tags that the big accelerators carry.

So, really, it’s not surprising that a bunch of physicists managed to blow through a million dollars in a few short weeks. What would be surprising is if we hadn’t managed to find stuff to spend that money on.

After all, I’m paying $260/kg for sand.


  1. #1 NL
    August 15, 2006

    What do you need the sand for?

    I remember buying some to put in a basket and rest a diode laser on for vibration isolation…

  2. #2 Perry Rice
    August 15, 2006

    All that ‘spensive equipment. As a theorist I used to watch my friends in labs, and marvel at all the cool toys they had. Back in grad school days, a buddy worked for a fellow with a real tight budget. My friend went to him one day and said he needed a new beam splitter. Why, he was asked! Well he had used the first one up. The research advisor was a bit puzzled, but they were using Nd:YAG pulses, and my buddy had gotten ahold of a beam splitter with an antireflection coating, burned holes, then using the translator on the mount to move the beam splitter over a bit…..Yes indeed he used it up. 2nd place for weird lab stuff was the time another friend had to check the voltage across an Argon laser to see if it was underpressured. Being skittish around high voltage, he hooked up the voltmeter, and old black Simpson meter before he turned on the laser. Well the big ionization current pulse to make the Argon conduct fried that meter, and the two bid power diodes on the power supply were cracked. The door to the battery compartment on the Simpson meter was weled shut by the heat…..I’m glad I’m a theorist, I would have hurt myself!

  3. #3 SMC
    August 15, 2006

    Is it correct to think that the reason for these high costs is that in your field, all of the information that can be had from “less precise” research has essentially been done, so that in order to add to the state of knowledge in your field, it is necessary specifically to go after increasingly high levels of precision? (Presumably the reason for the extremely expensive versions of otherwise cheap-sounding components e.g. sand?)

    Or to turn it around the other way – the reason it is not possible to use less expensive substitutes is specifically because the less expensive equipment and materials are not precise enough?

  4. #4 Uncle Al
    August 15, 2006

    Good luck. Aldrich under professional management is not what Aldrich was under the good Dr. Bader. Look at Alfa-Aesar for inorganics. See if what you bought works, then trace down why if it doesn’t.

    Define “purity.” Is it volatiles, bulk stoichiometry, specific element impurities, surface configuration…? A chemist’s high purity silica is rarely a lens coater’s silica source. If you bought silica gel it is more than likely brimming with adsorbed water and carbon dioxide. Stick it in a half-filled fused silica ampoule, hook up to dynamic forepump vacuum (Carefully! It’s gonna outgas and jump), and pump overnight to 500 C max (large pores retain) or 800 C (collapses microstructure). Remember the liquid nitrogen trap or you’ll get oil vapor backstreaming. If you want the porous gel, 200 C under vacuum overnight. If you want surface activity, 120 C in *air* overnight – but though it will be drier it will retain surface crud. How clean is your oven? You’ll add way more crud than you’ll get out.

    If all you need is tiny chips as opposed to porous silica, crushed fused silica would have been better. Now you can argue ultrapure UV silica vs. IR silica. Different trace impurities.

    Did you want porous Vycor preform?

  5. #5 Chad Orzel
    August 15, 2006

    The sand is for putting an AR coating on the useless $10 diodes. Apparently SiO2 is a good material for that.

    Well, ok, eventually, I’d like to be able to coat one of the more expensive useful diodes, but for the moment, I have a student who’s going to test it on the cheap ones, since none of them are within 5 nm of the right wavelength, anyway.

    100g of the stuff ought to be plenty for our purposes.

    Or to turn it around the other way – the reason it is not possible to use less expensive substitutes is specifically because the less expensive equipment and materials are not precise enough?

    That’s a fair way of putting it, yes.

  6. #6 Shelley Batts
    August 15, 2006

    Yeah, we’re looking into buy a confocal microscope, all for the cheap cheap price of $300,000. Oh, and its used.

  7. #7 RPM
    August 15, 2006

    I remember seeing an MSDS or warning label for sand. I think it was in an Orgo lecture. Couldn’t find it on google. The prof presented all the warnings and safety info (abrasions to skin, dangerous if swallowed, etc) then told us it was for sand. Amusing. Along the same theme as this.

  8. #8 SteveG
    August 15, 2006

    The college president was complaining to the provost about the amount of money the physicists were requesting for research. “Look at the mathematicians,” he said, “they only ask for money for pencils, paper, and erasers. And the philosophers, they don’t even ask for erasers.”

  9. #9 iGollum
    August 15, 2006

    I won’t go into the prices we pay for materials/reagents in molecular biology, but suffice it to say that if you want to compare prices per kg, we’ve got you beat pretty much every time ;-)

  10. #10 Chad Orzel
    August 15, 2006

    I won’t go into the prices we pay for materials/reagents in molecular biology, but suffice it to say that if you want to compare prices per kg, we’ve got you beat pretty much every time ;-)

    I don’t know about that.
    I have a colleague who blew 3/4th of his start-up budget (something like $5-10K) on a microgram of isotopically enriched calcium. In the late 70’s, even.

    He said that when it arrived, he thought the vial was empty.

  11. #11 iGollum
    August 15, 2006

    Ouch 8-/

    Okay, you win! Heh. Though for the record I said ‘pretty much’, not absolutely every time ;-)

  12. #12 Jordin Kare
    August 15, 2006

    A long time ago (1977, as I recall) in a galaxy far, far away (Cambridge MA) I was an undergraduate working in a laser lab at MIT. There was a wheeled lab cart blocking the line-of-sight for an optical gadget I was testing, so I rolled it out of the way. Alas, I didn’t turn the lights up first, so I didn’t see that the power supply on the cart was connected to a large HeNe laser (one of those 4-foot-long SpectraPhysics ones) on the adjacent optical table. Crash!

    My advisor (and the lab’s owner) let me look up the replacement cost of the laser tube ($1800) and then stew about what I’d done overnight before explaining that I wouldn’t have to pay for it personally….

    (And RE MSDS for sand: I’ve seen those also. But my favorite such was when I was looking at setting up an outdoor test stand at LLNL and was told by the safety folks that I’d have to set up a double-walled tank to hold runoff from the roof, since rainwater on site was classified as potential hazardous waste and could not be discharged onto the ground.)

  13. #13 Alison Chaiken
    August 16, 2006

    The newish diode-pumped solid-state lasers seem like an incredible bargain: 15 mW at 532 nm for thousands of hours at 300 mA. See for example Digikey.

  14. #14 Leon Brooks
    August 16, 2006

    We’ve got as much sand as you could want — pure white stuff, yellow, red… all kinds, yours to haul away by up to many tens of thousands of tonnes per lot. Western Australia is reknowned for its sand.

    How much overweight room did you plan on leaving in your baggage? If there’s a couple of you, you might squeeze 50kg into a pair of lightly packed suitcases.

    As to purity, well… it’s pure sand. You can take a chunk from the middle of the sand-patch — below the surface but some way from the ground itself.

    Heck, choose the right place and you may well get a few diodes in with it (probably silicon rectifiers — 1N4004 or the like — but for hundreds of dollars a kg we can easily do specials like LEDs, tunnel diodes or zeners).

    Hey, wow, that leg didn’t come off! (-:

New comments have been temporarily disabled. Please check back soon.