building a spectrophotometer from off-the-shelf parts

You owe me a nickel. Give it a charity in my name. The light source was a 4W blacklight lamp.

Yes, please, if you could, send me a scan. snipped-for-privacy@gmail.com

Michael

I'll probably try that next. I've been putting distillate into a graduated cylinder, weighing that, then using a density chart to get ethanol fractions. Crude, but it (kinda) works.

The link is confusing. One pic shows a white LED; the title implies use of a 100W light bulb.

I like Marvin's idea of using a small blacklight. Now, to get those wavelengths one at a time...

Thanks,

Michael

I just realized google groups (if you're using it) might mangle my email address. It's

m r d a r r e t t {at-symbol} g m a i l {dot} c o m

Thanks,

Michael

One problem I've had with using ordinary light bulbs in optics is that there is usually some writing on the bulb and it goes through the optics along with the light.

one?

My suggestion is that instead of finding ways to design a costly spectrophotometer, which can scan wavelengths and is of course delicate thing in terms of optics. You can design a colorimeter which has a cheap source and optical filters. Both should be easily available. BTW, does ethanol absorb in the UV? If it does, it would lie towards quite short UV which is near 200 nm. First of all you need to know the absorption spectrum of your analyte of interest and then start looking for the source which emits near the wavelength of maximum absorption and a optical filter. Tho, optical filters might be available, a UV source might be more difficult to hand and obtain. The rest of the work is for the electronics man!

Suggestions for IR or NIR, are perhaps impractical at home.

Read Dr. Marghoses reference on Ultraviolet Photometer, tho it is old (~ 1950s) but the idea is still useful.

M. Farooq

Isn't the accepted way to measure the SG before and after fermentation to determine how much sugar was converted to alcohol?

I make beer but never bother to measure the SG. Too much hassle, not much you can do about it (beer happens) and there's risk of contaminating the batch. There are some folks that claim it is necessary to tell when fermentation is complete, but I find it easier to control temperature and give it two weeks.

I did measure SG when I first started in 1990 - since a hydrometer/thermometer came with the brewing kit I bought.

The only real reason to do it (that I can see) is to meet some government requirement for labeling.

Single beam, single wavelength, crude, spectrophotometry is probably not the answer for determining alcohol content. In the lab gas chromatography was the first choice (another "easy" instrument to build) then HPLC. Most HPLC can also be preformed with thin layer chromatography (from what I understand) but TLC is better at finding out what is in a compound than quantifying the amounts.

BTW we had an old Beckman spectro that used an auto light bulb for the source and you adjusted the voltage (brightness) of the bulb to zero the instrument then put the full cuvette in and measured the sample. It was only useful for checking gross concentrations of substances with relatively broad wavelength absorption. I'm pretty sure it would be useless for alcohol.

Polarimetry may be a choice. Sugars twist light polarization - optical rotation. It is a common instrument for checking sucrose solutions. The instrument is a device with a light source, polarizing filter, long 2-10" sample tube, with flat, optically clear, ends, and a second - rotating polarizing filter before the detector. You measure how much the second filter has to rotate to stop light from reaching the detector between a blank (nothing in the cell or cuvette) and the sugar solution.

Like SG it also has to be corrected for temperature - density.

Presumably you use maltose, I assume it would work with all sugars but don't know for a fact.

Is this a hobby or business? Why do you care about alcohol and how precise do you need to be? I have a few friends who are actual chemists and do develop assays.

It's a hobby for now; it will maybe eventually become a business if I get enough time and funding. It's for fuel ethanol, not for drinkable ethanol. I'm actually starting with starch solutions (analogous to making vodka from potatoes, or making sake from rice followed by distillation). The requirement of hydrolyzing starch complicates things a bit. I have to calculate how much alpha-amylase I need to add to the starch solution, and how long I get them to react, before I reach a set glucose concentration for the yeast to much on (I'd like to keep it at no more than 200 g dextrose/L). Accuracy to 10% would be great. Research as to how it's done in industry revealed that some factories use on-line spectrophotometers to measure the ethanol concentration within the pipes. That's what inspired this whole post.

I'd imagine dextrose assays would be more accurate than using hydrometers...?

Thanks,

Michael

Your original post did say "beer." You lied.

I dabbled a bit with distillation. The usual technique is fractional distillation in fractionating packed columns - I did nothing so sophisticated.

Anything that makes it past the fractionating column(s) or thumper box is alcohol and water and not sugar or starch. Different ball game.

Think about the spectro for a bit. If you had sugar water, pure alcohol and plain water in glass test tubes it is likely you may not be able to distinguish which is which (save for things like refractive index, viscosity, meniscus or wetting properties). So it stands to reason that any resolution a spectro could offer wouldn't be in the visible wavelengths.

Then it is a matter of wavelength, so you'd need a monochromator or filter to let a narrow band of light through the cell - quartz so as not to absorb UV (probably rules out 100 watt light bulbs and LEDs as sources since incandescent bulbs are poor UV sources and LEDs are already pretty narrow band). see:

If you are really lucky the led would either absorb or transmit the exact color of the UV LED and the rest of the compound wouldn't. If that's the case, invest in a state run lottery, its bound to be more lucrative.

Reason why early spectros and lots of chemical assays are for 318 nano meters and 354 - they have lamps that are narrow band for those wavelengths.

Today they use quartz xenon lamps ($100-200) and power supply, for wide band light and a monochromator to select a 1-2 nanometer wide slice of that to pass through the cell, or a Deuterium lamp ($100-200 and constant current power supply) with monochromator.

My forte isn't chemistry but I did spend a lot of time in the production facilities and an assay lab for the pharmaceuticals.

For a pharmaceutical assay you generally start out with a known compound. You know what should have gone into it. The chemical precursors to the compound are either made on site or purchased. In any case, an assay is either performed on the precursors (and you know what you started with) or accepted from the manufacturer (how a lot of children died in South America when the company making cough syrup bought ethylene glycol that was certified to be glycerine by the Chinese company supplying it).

Anyhow, very few pharmaceutical assays are simple. Before it gets to the spectrophotometer, the chemical compounds are separated out using a packed column of silica gel particles, the usual procedure for most stuff. Each compound will have a specific affinity for the silica gel and larger molecules take longer - all the stuff is injected onto the head of the column in a lump (after the tablet is dissolved and a solution of liquid only is made). A high pressure low volume pump pushes the material through along with a "mobile phase" (the packed column is the "stationary phase" but just called a column). As the compound percolates through the tightly packed micron sized particles the molecules are slowed to varying degrees by their own affinity for the stationary phase. When they get to the end of the line, 2-15 minutes later, (columns being about 1/4" diameter and 2-12" long for a lot of HPLC) they go into a quartz flow cell which shines a narrow wavelength UV light and measures the absorption. To know what it is you measure the time it comes off the column and to know the quantity you measure the absorption with the spectrophotometer (just called a "detector" by the HPLC chemists)

With a lot of HPLC you only need one wavelength of light - but they can use scanning or dual wavelength detectors as well (costs lots more).

There are lots of variations on this same theme using gas chromatography and measuring the flame ionization potential, or ionization from a nuclear isotope, capillary chromatography using long coated capillary tubes and electrical conductivity or spectroscopy for a detector, and thin layer chromatography (as seen on TV when showing DNA analysis), better for substance ID but can be put into a box and illuminated with light and reflection or fluorescence used to quantify it.

PhD Chemists and especially professors, seem to think they know a lot about chemical instrumentation and electronics. They generally have no clue as to how incredibly sophisticated the instruments they use have become from a technology viewpoint.

I'm not surprised to see the DIY "spectrophotometer" given the source of the article - but I can say with a reasonable certainty that it probably won't work for what you want. This guy found a substance that absorbs in the near UV range with some source he had could supply. He demonstrated the principles of how a spectrophotometer works, not how to build a practical laboratory instrument. The fact that the "peak" is so sloppy would seem to attest to the lack of a good narrow wavelength source of light - and he does mention varying the wavelength so he's using a diffraction grating as a monochromator (or prism).

I couldn't find a representative spectrograph of KMnO4 on line, but some references to detecting it in the IR range. That would have been a good way to show how well his spectro works - comparing the spectral plot of a real instrument to the DIY one.

If he had a better source, that sample would probably look like a jagged series of sharp peaks. And You have no idea if one or more compounds of what You want to resolve is at same wavelength - one peak can cover and obscure another - one reason chromatography is used to separate out the compounds before they are analyzed with a spectro.

In beer and traditional whiskey we take grains and let them sprout or just germinate for a day or two, dry and crush them then let the enzymes work on the mash while the solution/suspension of malt and water is held at 150 F. The germination converts a lot of starch to sugar and the enzymes develop and react on the remaining starch while the drying stops the development of a plant and chlorophyll and other yucky tasting things.

You're seeming to imply you have pure starch with no natural enzymes. That doesn't sound like it will produce fuel economically - since to arrive at pure (dead) starch, you inevitably start with some natural substance and remove everything else that might convert it to sugar naturally. Vodka was made long before you could just go out and buy amylase enzyme . . .

Starch can always be used as food - so even if it were an unwanted by-product there is a market for it.

An oversimplification, perhaps. Some experiments I did with alpha- amylase and the kids' leftovers (closer to vodka, I guess); other experiments I did with malted barley and the kids' leftovers (closer to "beer" without hops), and some experiments I did with Aspergillus oryzae (koji) and moldy bread (closer to sake). The A. oryzae's need to be in a hot, aerobic environment made things... complicated.

It certainly doesn't make very good dinner conversation. "So, what have you been up to?" "Oh, I've been collecting garbage and trying to make fuel from it. Wanna hear about how my koji mold stank up my garage?"

I hadn't expected people to actually take an interest in my process, so I oversimplified. Sorry to disappoint you.

Yep... the methods I learned in school involved trays and structured packing. I remember stressing over q-lines while studying for the Chemical PE exam several years back.

~snip~

My feedstock is pretty much starchy food waste - kids' leftovers, rice, bread, etc. I'll have to look up the history of vodka making...

Thanks,

Michael

On Fri, 05 Oct 2007 20:06:47 +0000, mrdarrett wrote: ...

Have you tried getting methane from their poo? They have a virtually unlimited supply of it, you know. ;-)

Cheers! Rich

That reaction takes way too long - requiring really immense reactors as a consequence. I estimate about three gallons of poo reactor to get a single watt of electrical power. One single kWe would require about 3000 gallons of poo reactor.

Reference:

Then you have corrosion problems from the hydrogen sulfide...

Scale-up is impractical (at least on a home scale).

Michael