Applied Science

In a previous video, I showed how to extract caffeinated water from green coffee beans with supercritical CO2. Here, I show how to extract dry caffeine crystals from the caffeinated water (ie coffee, in this case). I used a vacuum filtration setup to create very strong coffee, then mixed it with methylene chloride in a seperatory funnel. The methylene chloride sinks below the water, taking most of the caffeine with it. By opening the funnel carefully, only the methylene chloride and caffeine can be transferred to a flask where the methylene chloride is boiled away. I then mounted a test tube in a rubber stopped and filled the tube with ice. With the stopped in the flask, I evacuated the flask and applied heat. The caffeine will sublimate and collect on the cold test tube. I scraped off the caffeine and measured the mass.

In his recent post, Chris Gammell used astronaut ice cream as a fun example of how high tech innovation trickles down to consumer-level products. Seeing an image of astronaut ice cream in my G+ feed got me thinking about making some of my own. I managed to accomplish this with a harbor freight vacuum pump, some dry ice and various hoses and fittings. The process is known as freeze-drying and allows water to pass directly from ice to vapor, thus allowing the ice cream to maintain its physical structure while it is dried. What should I freeze-dry next? http://engineerblogs.org/2012/01/weekend-journal-the-trickle-down-techonomy/

In previous videos, I used supercritical CO2 to dry my homemade aerogels. This time, I soaked the aerogel in methanol, then raised the temperature and pressure of the methanol itself to make it supercritical. This allows the gels to be dried without an additional solvent exchange into CO2. The downside is that it requires a chamber full of methanol at 460*F and over 1200 psi, which is a much bigger hazard than using CO2. The aerogels dried with methanol shrank less than the ones dried with CO2, but there was still a lot of cracking, and I have yet to create a high-quality monolith.

These bakers' yeast cells might be reproducing in the video. It's hard to tell if its just movement from the water under the cover slip More photomicography: http://benkrasnow.blogspot.com/2010/03/microphotgraphy-with-lumix-gh1.html

I built an improved cold trap from an aluminum pan and a stainless steel bowl. It held more ice than my original copper pipe cold trap, but the 100ml of water from the gelatin still managed to almost plug up the trap.

I combined the High Intensity Discharge bulb and power supply from an old LCD with a desk lamp to create a 270W spotlight. I bypassed the power supply's interlocks and shorted an optocoupler to cause the power supply to always turn the lamp on when AC power is applied. I used an inline power switch to control the device. The desk lamp's existing wiring is able to withstand the high starting voltage of the HID lamp (probably a few KV). Once the arc strikes, the voltage is much lower and the current much higher.

A liquid lens works by changing the shape of a water drop by creating an electrostatic field that pulls on the water molecules. I show how water can be affected by a high voltage supply and an electrostatically charged comb. The process by which the droplet changes shape is known as electrowetting.

A brand-new power strip emitted a loud pop and left a soot mark on my friend's hand when he plugged it into the wall. No devices were attached to the power strip. I decided to take the thing apart to see what failed. The circuit breaker in the household electrical system did NOT trip during this event. My guess is that a very fine piece of wire bridged the neutral and hot terminals on the switch. When power was applied, the whisker vaporized.

Here I show how I made gassified candy (eg Pop Rocks) in my home shop. The basic procedure is to heat sugar water up to 280*F, carbonate it with 600 psi CO2 and vigorous mixing, then cool the mixture while it is still under pressure. After the mixture has solidified, rapid depressurization causes the candy to burst apart into small fragments. Each fragment has lots of tiny CO2 bubbles embedded in it. When placed on the tongue, the bubbles burst open as the candy dissolves and a popping sensation is felt. The main difficulty is building a high-pressure chamber that allows precise temperature control as well as thorough mixing of its contents while under pressure. I've built something that meets these requirements, and I've finally made a batch of candy that would pass as Pop Rocks, but they are pretty weak. I still need to adjust the temperature and mixing times to get a better product. Link to recipe: http://www.exploratorium.edu/cooking/candy/recipe-lollipops.html

I found some interesting materials at a local fleamarket (http://www.electronicsfleamarket.com): a canister of lithium sheet metal, and a bottle of fine zirconium powder (325 mesh). The zirconium is stored under water since the dry power can self-ignite, or may be ignited by static electricity or shock. Confusingly, the MSDS sheets say that dry zirconium powder must not be allowed near water, but if it is to be stored, it must be completely submerged in water. I carefully dried a very small amount, and sprinkled it over a flame. If you know of something interesting to do with these chemicals, please let me know.

Some photos here: http://benkrasnow.blogspot.com/2012/04/dissolving-lithium-in-anhydrous-ammonia.html Here's an interesting reaction where lithium metal is dissolved in anhydrous ammonia. I placed a couple small pieces of lithium in a test tube, then connected the tube to a larger flask of ammonium hydroxide via silicone tubing. Upon heating the flask, gaseous ammonia will be produced which is condensed in the test tube by cooling it with a bath of alcohol and dry ice. The resulting solution is brilliant bronze in color, but at lower lithium concentrations, the solution is deep blue -- a very beautiful reaction.

A cloud chamber shows alpha particles being emitted from americium sources. The chamber works by creating a layer of supersaturated alcohol vapor which visualize the radation. The alpha particles trigger condensation in trails through the vapor cloud and show up as tiny droplet tracks. In this chamber, I have two alcohol-soaked pads above an aluminum plate that is cooled by dry ice. The pads emit alcohol vapor, which is cooled by the air above the aluminum plate to the point where it is colder than would normally by necessary to condense. The tiny radioactive particle pushes it over the edge to cause condensation, which is visible droplets.

I made some whipped chocolate by melting standard semi-sweet morsels, and applying nitrous oxide at 250 psi. I then dispensed the melted, gassified chocolate into a chilled vacuum chamber, then applied a vacuum to create large bubbles within the chocolate. I maintained the vacuum level while the chocolate solidified with the bubbles still intact. This process yields a dessert that is very low-density, and has a very pleasant airy texture.

Chris Agerton (http://www.youtube.com/user/cagerton) suggested that I try adding nitrous oxide bubbles to an avocado -- a wonderful idea since avocados are fatty, and the gas should readily dissolve in the fat. I placed some avocado slices in my pressure chamber and dispensed two nitrous cartridges into the chamber. I let it sit for about 7 hours. I then took the slices out, and reduced the pressure with a vacuum pump. Unfortunately, the avocado was resistant to the nitrous infusion process. I will try again with guacamole and shaking/stirring to encourage gas absorption.

I built a liquid level controller that senses the level of water in my aquarium and powers a water pump when necessary to compensate for evaporated water. This is known as an "automatic top-off" system. The controller is based around a Keyence FS-V11, which sends light out a plastic fiberoptic, and measures the amount received via another fiberoptic. The optical probe makes use of total internal reflection to control how much light is returned to the Keyence based on how far the probe is immersed in the water. It is also completely passive, non-metallic, and has no moving parts for high reliability. The Keyence triggers a solid state relay that supplies power to a standard AC outlet. The pump is then plugged into the outlet, and moves water from a reservoir into the aquarium on command.

I took apart my microwave oven and measured the voltage and current supplied to the magnetron. The device appears to start conducting current at 4KV, and will allow lots of current to flow once this threshold voltage is reached. My current probe is pretty cheapo, so I wouldn't trust its measurement too much, but the final determination of 1300 W average seems pretty spot-on.

I recently bought some Fluorinert FC-40 on eBay since it is such an intriguing chemical. This fluid is 1.8 times as dense as water, but has almost the same viscosity. It is also very inert, has a boiling point of 160*C, is immiscible with water, and has exceptionally low electrical conductivity. The fluid can dissolve large amounts of gas and was shown in the movie The Abyss where it allowed a rat (and later a human via special effects!) to breath the oxygenated fluid, by submerging the rat in a container of Fluorinert and having the rat take the fluid into its lungs. I decided to fry a potato chip in the fluid to see if I could make a truly fat-free deep-fried chip. The experiment worked very well, however I am not convinced that eating the product would be a good idea. Your thoughts?

I forced an RCA 811A tube to produce some X-rays by operating the tube in cold-cathode mode at about 20KV and 150uA. The glass fluoresces nicely, but I didn't get any light from my X-ray intensifier cassette. MightyOhm's geiger counter kit: http://mightyohm.com/blog/products/geiger-counter/

A few years ago, a research group posted a video showing that X-rays could be created by unrolling scotch tape in a vacuum. The idea is that the static charge generated by the mechanical unrolling process would cause electrons to accelerate to 50KV (or so), and then produce x-rays when they slam into something. I tried to recreate this experiment, but I had pretty weak results. I may need a much larger vacuum pump to keep the pressure in the chamber very low while unrolling.

In this follow-up video, I show that I was able to get a P47 phosphor disc to illuminate when unrolling tape in the vacuum chamber. The blue x-ray intensifier screen did not illuminate although it was also in the chamber. Also, it seems the P47 was illuminated through a thin copper foil, though I will have to do some more tests to believe this.

I made a puree of maraschino cherries and added 2.5g of calcium lactate gluconate. I then dripped this mixture into a solution of 2.5g of sodium alginate in 500g of water. A gelatinous skin is formed at the interface between the two liquids, forming spheres of cherry puree. This process is known as reverse spherification. I bought the chemicals here: http://www.amazon.com/Experimental-Kit-Artistre-600-grams/dp/B0045KOOXU

Some people were curious about what my shop looks like, so here is a video tour. It seems that I was really rambling, since this video turned out much longer than expected.

I thought it would be interesting to try decapping some chips. This involves using fuming nitric acid, which also seemed fun, so I thought I would give it a go. The process starts by milling a precise pocket into the IC using my CNC machine. I used carbide tooling to cut the glass fiber/epoxy material. I then put a drop or two of the acid into the pocket, and raised the temperature to about 100*C. The acid dissolves the epoxy packaging as it sort of "dries". I added more acid to the pocket every few minutes. After about 10 minutes, I washed the IC in acetone, then reapplied acid if there was still material left on the die. Eventually, it was all cleared away, and I had a nice decapped IC.

This video shows my entry into the world of high-intensity flash tubes. I was inspired by the work done by the famous Doc Edgerton who built a strobe system for nighttime aerial photography. I've heard the biggest system he built was somewhere between 50 and 75 KJ per flash -- the power needed to illuminate the ground from 1000 feet in the air! The system in this video is a Speedotron 4803, which delivers up to 4800J per flash. Photographers use the units watt-seconds (Ws) to describe the energy in each flash, which are the same as Joules. I plan to add capacitance to the bank in order to increase the total discharge power, but the voltage may also need to be adjusted since the higher power flashes also tend to require higher voltages. I'm also planning to try different xenon flash tubes, which are capable of more intense flashes.

I found an X-ray tube on eBay and fired it up with my Spellman high voltage supply. I ran the tube at about 35-40KV at 180uA. This required a filament current of about 1.5A at 2.5V DC. With the room very dark, I could clearly see the phosphorescent screen glowing green. It wasn't bright enough for my eye to discern details in the X-ray images, but the camera with a 4 second exposure at f/5.6 and ISO400 was perfect. I took some pictures of various electronic components. I also X-rayed a dead, dried-up bumble bee, and it produced no image at all! The X-rays completely penetrated the insect without any visible absorption. I should have turned down the anode voltage, but I didn't think of this until after I threw away the bee. The tube is a Nago GFH2-0.3-85-60 Link to still images: http://benkrasnow.blogspot.com/2012/09/first-test-with-ebay-x-ray-tube.html

I used my DIY scanning electron microscope to view a 555 timer circuit while it was powered. The circuit is a simple oscillator with a very long time constant to make the changes easy to see. My plan was to view the silicon die itself, and hopefully discern changes in its internal circuitry as the oscillations occurred. As it turns out, I was only able to "see" the charge of the electrical wires going to the chip socket. There is likely a clear oxide layer that covers the silicon die, and needs to be removed with hydrofluoric acid in order to use the SEM to inspect the die itself while powered.

I built a small rocket engine for demonstration purposes. The engine is built from a 2" diameter acrylic rod through which I drilled a 0.5" hole. The oxygen at 80 psi or less is passed through the hole and then is forced through a convergent-divergent nozzle at the tail end. The nozzle's throat is about 0.25" and expands to 0.625". I lit the engine by inserting a burning cotton swab (with wooden stick) while a small amount of oxygen was flowing. The acrylic catches fire very easily in a pure oxygen environment. The engine can be throttled and shut off completely, which is a major benefit to hybrid engine designs. Solid-fuel rockets cannot be throttled or shut off, which makes them difficult to control.

I wired up a high-power photography strobe light to my oscilloscope and made some measurements of the input voltage, current and light output.

I've made a series of wooden vases and similar items by laminating contrasting boards, then turning them on the lathe. I like to use walnut, maple and either bloodwood or padauk to create a white/black/red theme. I almost never use stains, instead I just finish the wood with clear wipe-on polyurethane, using two coats for a fairly matte finish.

I bought a set of x-ray equipment on eBay, which included an x-ray head unit, 50KV supply, and filament power supply. The filament supply has a 25-pin connector, and no data or markings on the case or board. I spent some time reverse-engineering the circuit, and determined how it should be connected to make an emission-controlled x-ray system. I tested the whole thing, and it appears to be working as intended at 50KV / 1 mA. The head unit produces a beam of about 25 degrees total angle.

I built a large area detector that is sensitive to X-rays and will be used in an upcoming project. The detector is constructed from a "Lanex Regular" X-ray intensifying screen, and a Hamamatsu R6094 photomultiplier tube (PMT). The PMT amplifies the light from the intensifying screen which is emitted when an X-ray strikes the screen. The area of the screen allows it to catch many X-ray photons created by backscatter. The PMT is setup with a simple 100K resistor between the anode and ground. A 1V output corresponds to a 1 microamp anode current. I'll probably add a very simple opamp circuit to this, and then pipe it into an A/D converted for digitization.

It seems strange that a single piece of glass can compute the Fourier transform of an image, but it is true (sort of). I explore an optical arrangement known as a 4F correlator, and note the practical considerations that are rarely discussed in existing sources. http://www.imagemagick.org/Usage/fourier/

I'll be presenting this work at EHSM: http://ehsm.eu/ I built an X-ray backscatter imaging system from parts found on eBay. This system works by scanning a very thin beam of X-rays across the target, and measuring the amount of backscatter for a given beam position. The beam is scanned mechanically by a rotating chopper (collimator) wheel, and by tilting the rotating wheel on an orthogonal axis. The output image is generated on an oscilloscope by matching the horizontal scan speed to the rotating wheel, and using a potentiometer to measure the vertical axis position. The scope's brightness (z axis) is controlled by the amount of backscatter signal received by a large-area detector. Thus, the image is constructed bit by bit. I used a long-exposure shutter on my camera to see the image formed by the moving oscilloscope trace. http://www.google.com/patents/US5181234

I built a system to synchronize the two color wheel motors in a Wurlitzer Model 1015 jukebox. Originally, the jukebox used two synchronous clock motors, and the designers assumed that the motors would stay synchronous and keep the colors wheels at the same orientation after manually setting it. As it turns out, the torque required to spin the color wheels is enough to cause the motors to slip. Eventually, the color wheels get out of sync, and the left and right side of the jukebox do not match in color. My upgraded system uses two small DC gearmotors that interface with the original drive mechanics via a nylon gear from McMaster. I drive the two motors via a PN2222A transistor and sense the position of the color wheels with the optical sensors from an old computer mouse. An arduino controls the motors via PWM (20 or 30 KHz), and runs a phase-locked-loop routine with P-I control. This system could use a little tuning, but it's pretty close.