Ostwald reactor from platinized silica wool

I started a thread recently on making platinized silica wool from e-cigarette wick and platinum metal. The goal was to use the catalyst in a reactor based on the Ostwald process to make nitric acid (or nitrate/nitrites). Since the original thread was about the catalyst and specifically the silica support I used, I'm creating this new thread to report my progress in making the reactor.

The main idea behind the reactor is to pass a mixture of ammonia and excess air in a tube containing a finely dividied platinum catalyst preheated to red heat. If the ammonia air mixuture is right andthe flowrate is appropriate and a bunch of other things work out, you should obtain a mixture of nitrogen oxides and air which can be dissolved in water to make nitric acid.

I just did a pilot run with a source of 2.5 moles of ammonia gas to gauge how well the catalyst work, and get a feel for the different parameters that are important.

Experimental
The setup I used can be seen in the attachments w/ close ups of the different parts. The first part in the setup is the vacuum pump, the output of which supplies the air for the reaction. The input has a needle valve to control the air flowrate. The air goes into the three neck flask containing NH4NO3 (235 g of 85 % ammonium nitrate fertilizer in this run). The addition funnel contained 100 g of NaOH in water (solution density 1.53 g/ml).

To start the reactor, NaOH solution was dripped at a rate of about 1 drop/second and the catalyst tube was heated to red heat with a blowtorch. The air flowrate was the adjusted to maintain continuous oxidation without additional heat. About 10 large bubbles per second were coming through in the absorption flask.

Immediately a brown gas was seen forming inside the tube and white smoke appeared in the suckback trap. A tube at the end of the setup leading outside to vent excess gases was also spewing white smoke for the whole run. After the reaction is started, no much else to besides enjoying the show. At one point the plastic tube coming out of the reactor tube melted from the hot exhaust gases. The air flowrate was decreased to mitigate the reaction.

Results
At the end of the run, I tested the pH of the solution in the absorption flask: strongly basic! Obviously a lot of ammonia passed through the reactor without being oxidized, even though enough heat was generated to keep the catalyst red hot.

Discussion
A major problem with this run was that too much NH3 passed through the catalyst without being oxidized. This in turn was due to a too low air flowrate (not enough to oxidize all the NH3) or too high NH3 flowrate. Unfortunately high flowrates tended to push the catalyst towards the end of the tube, I'll have to somehow hold it in place next time. And at high flowrates the hot gases didn't have to time to tool down enough before reaching the plastic tubes, which melted it releasing NOx and NH3 everywhere.

The generator worked relatively well, stirring was surprisingly not an issue and the "waste" leftover is actually mostly sodium nitrate which will be useful. But I think too much ammonia is produced too quickly even at the slowest addition rate my funnel allows. (Quick calculation: adding 1 drop NaOH = 0.05 ml <=> ca. 1 mmol <=> 25 ml pure NH3 released.)

I haven't yet decided how to cool the exhaust gases from the catalyst tube. A fan would be practical but might not be sufficient.

This post is getting long, I might add other remarks later.

Anyway if you have ideas for improvement or experiences with those kind of reactors...

EDIT: pictures and legends: the setup, the red hot catalyst and the plastic tube that melted from the hot gasses.



[Edited on 15-7-2018 by Heptylene]

Quote: Originally posted by VSEPR_VOID

you should pack the excess space in the reaction tube with glass wool or calcium chloride. [Edited on 15-7-2018 by VSEPR_VOID]

Quote: Originally posted by unionised

Quote: Originally posted by VSEPR_VOID   you should pack the excess space in the reaction tube with glass wool or calcium chloride. [Edited on 15-7-2018 by VSEPR_VOID] Why?

Quote: Originally posted by unionised

Wet CaCl2 will react with nitric acid to form Cl2 and NOCl/ NO2Cl etc. Silica/ glass wool would be an idea to keep the catalyst in place. I'm not sure, but I think it might help with cooling the gas after the reaction.

Quote: Originally posted by Heptylene

What do you mean by halogenated plastic? PTFE? The tubes are soft PVC, which is not really chemically resistant but its all I can find easily. I guess PTFE would actually be a significant improvement over PVC, especially in terms of temperature resistance.

Quote: Originally posted by VSEPR_VOID

Is there a better material for coating in platinum than glass wool? What about activated carbon as a substrate?

Quote: Originally posted by WGTR

The trick is also to pick a substrate that will effect the decomposition of ammonia and/or nitric oxide much more slowly than the platinum catalyzes the oxidation of ammonia to nitric oxide. Even platinum catalyzes the decomposition of nitric oxide, but more slowly, so if the reaction zone is thin and the reactant flow rates and temperature are optimal, then 90-100% efficiency can be obtained.

This is currently unfinished, and I'm kind of in a hurry right at the moment, but I've uploaded a picture of the condensor. It is intended to be mounted vertically on a board. The left side is the bottom, and the right side faces upward.



Here's the parts list:

From Home Depot:

1. (1x) 3/4" PVC pipe, 2 ft. Part # 22075 Price $1.55 ea.
2. (2x) 3/4" slip x 1/2" FIP x 3/4" slip PVC tee. Part # C402-101 Price $1.14 ea.
3. (1x) 3/8" all-thread, 2 ft. Part # 802167 Price$1.97 ea
4. (1x) 3/4" slip x 3/4" FIP PVC adaptor. Part # C435-007 Price $0.74 ea.
5. (1x) 1/2" barb x 3/4" MIP adaptor, nylon. Part # 800229 Price $2.24 ea.

From Lowe's:

1. (2x) 3/8" barb x 1/2" MIP, nylon. Part # 877100 Price $1.89 ea.
2. (2x) 3/4" split ring pipe hanger. Part # 302049 Price $1.67 ea.
3. (1x) 1/2" split ring pipe hanger. Part # 302044 Price $1.68 ea.
3. (3x) 3/8" ceiling bracket. Part # 301689 Price $2.24 ea.
4. (1x) 5/8" OD x 1/2" ID clear tubing 10 ft. Part # 814319 Price $4.59 ea.
5. (1x) 1/2" OD x 3/8" ID c;ear tubing 10 ft. Part # 879263 Price $4.18 ea.
6. (1x) 3/4" OD x 5/8" ID clear tubing 10 ft. Part # 814320 Price $7.99 ea.
7. (1x) 3-pack 0.956" rubber diverter. Part # 891223 Price $1.49 ea.

From Napa Auto Parts:

1. 5/8" -> 9/16" fuel injection hose clamp. Part # 7051227 Price $0.99 ea.
2. 1/2" -> 7/16" fuel injection hose clamp. Part # 7051225 Price $0.99 ea.

I'm still waiting on the 8mm x 6mm x 4' quartz tube, but it will slide all the way through the PVC pipe and out either end. The bottom end has a plastic hose barb connector that the quartz tube will slide through. The short piece of clear tubing on the end of the barb is intended to provide a water-tight seal. If you look closely you can see that there are two pieces of clear tubing, a smaller piece inside of the larger one. A 5/8" fuel injection clamp is there to provide a seal between the clear tubing and the quartz tube.

Cool water will flow in through the bottom barb fitting, and out the top one. The condensor is intended to be completely open on the top. The water flow rate will have to be adjusted such that gravity is sufficient to keep water from spilling out the top of the condensor. I discovered that the hardware store also sells a fitting that adapts from a barb fitting to a standard garden hose fitting. Interesting.

The condensor will be positioned into place using a pair of 3/4" split ring pipe hangers. These things have a threaded 3/8" hole in the bottom that accepts a length of 3/8" all-thread. The other end of the all-thread screws into a ceiling bracket, that I am going to mount on a 2 X 4. If one drills out a hole through the wood under the bracket, then it is possible to screw a standard length of all-thread through the board (maybe a 6" long piece?). This means that you can adjust the precise stand-off of the pipe hangers without cutting each piece of all-thread to an exact length. I'll show this later.

The 2' piece of all-thread shown at the bottom of the picture shows how the hanger and the bracket fits together, although I'll be using shorter pieces of all-thread.

The "hot end" of the quartz tube will fit into the tube furnace that I referred to earlier. It's about 18" long. Right now the condensor is about 2' long. I may cut it and shorten it a bit later, so that a 4' long piece of quartz is long enough.



This is a rubber diverter that I picked up in the store. It's made to fit inside of a garden hose fitting to pick up debris and stuff. I think the screen is stainless steel, but I'm not sure yet. I'm planning to cut it carefully to squeeze tightly inside of the quartz tube, and provide a support for the catalyst. Once up to temperature the screen may anneal and fall out of the tube, meaning that I'll have to think about this a bit. Maybe a thin quartz rod placed into the tube would support the screen.

I didn't read the OPs last couple posts, so IDK if he figured out a way to keep the catalyst from being blown to the back of the tube. I was thinking that you might want to try finding a gravel quarry/pit, gravel dump or gravel driveway (if you have a limestone area) where there is often solid white rocks incorperated into the grey limestone. These are quartz rocks, some are beautiful and you can see the crystal structure in them and I've found ones 3"x4"x2" in size. They break nicely and could be filed down with something like a dremel (IDK if a metal file would work). They may be heavy enough to hold the catalyst in place, you could put a few in line and it would allow air flow around them as well as not contaminate your product with anything that isn't already there.

I know I used to be able to pick up a bucket full of solid quartz rock when they would dump a new load on the driveway. I have some samples I'll post pics of if you aren't sure what I'm talking about.



[Edited on 7-24-2018 by RogueRose]

Quote: Originally posted by WGTR

Thanks for the idea RogueRose. I'm familiar with the type of rock that you're referring to, and have seen them in driveways, etc. If one is choosy, it's possible to find them with large facets and pretty transparent. It's a possible option. It would be best to get the purest crystalline samples that you can scrape up, so that there aren't other minerals present that could decompose the product. It would probably be best to leach them in nitric acid beforehand to make sure they are chemically "clean". [Edited on 7-24-2018 by WGTR]

Quote: Originally posted by Chemetix

Smashed up brick or tiles have a great surface area and chemically neutral, It worked with cobalt, platinum oxide would also work.

I inserted the quartz tube into the furnace and ran some brief tests today.

I calibrated a flow of nitrogen to be 1L/min by measuring water displacement from an inverted beaker vs. time. The 9mm quartz tube was inserted into the furnace just enough so that the hot end barely protruded from the brick. A thermocouple junction was inserted about 2" into the tube to measure nitrogen temperature as it exited the furnace. The nitrogen flow was connected to the inlet of the quartz tube, and the connection secured with a fuel injection hose clamp.

After 30 minutes of running the furnace at 200-250W, a temperature of 520°C was measured by the thermocouple. This temperature seemed to be stable.



Power was removed from the furnace, and the tube allowed to cool down. Then, the tube was inserted fully through the furnace so that the condenser could be installed. Deionized water of about 2" of water pressure (measured at the top nylon barb fitting) was used to provide cooling. Water came in from the right rear end of the condenser, and left through the copper drain on the bottom left. The waste water was just run to a floor drain for this test.



With a flow of 1L/min of nitrogen and the furnace operating at 200W, the temperature of the quartz tube after the condenser never deviated from 23.8°C, during the 10 minutes that measurements were taken. With this type of gas flow rate the condenser is more than adequate for the job.

Here's a photo of the cover removed from the tube furnace:

Thanks for the compliments Heptylene. I don't think there is any problem with that temperature differential on the tubing. The temperature doesn't drop from 500°C to 25°C all at once; there is a transition that occurs once the tubing leaves the furnace. I discovered that I can measure temperature at various points in the tubing by merely inserting the temperature probe after everything is already installed and operational, like this:



The probe is inserted from the condenser end, to protect the wire insulation. For some reason the thermocouple insulation seems to hold up to the output temperature of the furnace, at least for the few seconds needed to take a measurement. I noticed that there is a considerable temperature gradient from the furnace output to the condenser input.

After thinking about it for a bit, I'd like to go back and measure the temperature at regular intervals inside the tubing, so that I can get an idea of how quickly the furnace heats up the air, and how quickly it cools down again in the condenser. I could plot tube temperature vs. position. I'll have to strip some more insulation off the thermocouple wire for that, so that I can push the junction further into the furnace. Do you think that kind of data would be useful?

Here's another data point after further testing:

Flow rate, air: 3.3L/min
Input power, furnace: 250W
Temperature at furnace output: 526.5°C
Temperature at condenser output: 31.8°C
Condenser water temperature, input: 25.7°C
Water pressure: 4"



I think that I have some quartz wool in another lab, so I'll borrow a page from your book, and try soaking some of it in my hexachloroplatinic acid. Some quick back-of-the-envelope calculations seem to show that a 3-4% ammonia solution should give a 3% concentration in air if air is bubbled through it at room temperature. I should get some results out of that, I would think.

I added graduated marks to the side of the condenser and furnace for reference, and photographed the setup with a tape measure for clarity.



My thermocouple could only reach three feet into the tube, and the tube is four feet long. This means that I was able to reach only halfway into the furnace with the thermocouple junction. I could go back later and gather some more data, but I'll have to remove the condenser and slide the quartz tube back through the furnace a foot or so to be able to measure temperatures in the first half of the furnace. Anyway, you can still see the temperature at the halfway point in the furnace, and that it is still rising until about 2/3 of the way through the furnace, then it starts to drop.

With 150W to furnace:


With 250W to furnace:


I figure that with 200W input, the temperature would max out around 500C.

Right at the 26 inch mark the temperature drops noticeably. This is because there is a cloth ring that is stuffed around the reaction tube at this location. Its purpose is to divert water into the drain that is flowing along the reaction tube, and it remains wet constantly. The temperature is going to vary a bit within the tube, and the gas flow may or may not be a different temperature than the walls of the tubing. The thermocouple junction is generally touching the tubing, not directly sampling the air temperature. This also explains why the tube seems to be getting hotter as it exists the bottom end of the condenser; the gas flow is about 40C as it exits the condenser, and there is no water flowing on the tubing right at that point to cool down the walls of the tubing. The temperature drops off near the end of the furnace. I think this can be attributed to air leakage from the furnace.

I set the air flow on this test to 4L/min, as this seems to approximate the conditions recorded in industry, considering the size of the tubing. It's a starting point anyway.

If a 10% ammonia gas feed is utilized, then a 300C preheat will probably be all that is needed. If ammonia feeds drop to 1-3%, then it may be necessary to raise the preheat temperature to 500-600C.

[Edited on 8-11-2018 by WGTR]

OK, I'm going to go back and make another set of measurements with a slower flow rate. In the references uploaded in this post for some reason I had understood that the flow rates mentioned there were for ammonia only, not the total ammonia/air mixture. Not sure why I thought that.

I was re-reading the references again as a sanity check, because I'm getting ready to order a flow meter. It's time consuming to calculate flow rate iteratively by measuring air displacement over water. If it's sized correctly it will be more accurate, as these meters are most accurate the closer you are to their maximum capacity. Anyway, the measurements that I took at 4 lpm were for a flow rate that is around 3 times too much under the most optimistic conditions.

A 7mm ID tube (9mm OD) should correspond to an internal area of 0.385 cm². For a flow rate of 0.636 lpm/cm², this corresponds to a total flow rate of 0.385*0.636 = 0.245 lpm.

For the fastest flow rate of 3.51 lpm/cm², this corresponds to a total flow rate of 0.385*3.51 = 1.35 lpm.

Realistically, I think that I'll keep the flow rate around 0.5 to 1.0 lpm. The condenser seems to handle a 1 lpm flow rate easily, and the air exiting the condenser under that condition was practically room temperature.

I'm considering bubbling air though an ammonia solution in water to formulate the correct gas feed. If I'm understanding this correctly, a roughly 2.5-3.0% solution of ammonia in water will give a partial pressure of ammonia that is equal to that of water at 20C. The idea is that as air is bubbled through the ammonia solution, the ammonia concentration in the solution will hopefully remain the same. This would provide a consistent gas composition over several hours if the solution temperature can be maintained.

I pulled some data from the Wikipedia pages (not exactly an authoritative source, I know) to create a chart. I looked at the curve for ammonia, and extrapolated the 2.5% data point.






The idea is that air would be dried with a desiccant and measured with a flow meter, then bubbled through the ammonia solution. If desired, the ammonia/air mixture could be dried with calcium oxide before sending it to the furnace.

Assuming 24.5 liters of gas per mole at 25C and 1 atm, that corresponds to 0.0408 moles per liter. With a 2.25% ammonia concentration in air, this gives 0.000918 moles per liter of gas feed. If the yield is 100%, this gives 3.47g of nitric acid per hour at 1 lpm gas flow. This also would consume 0.980g of ammonia during the same time. Feel free to check my math.

[Edited on 8-11-2018 by WGTR]

I managed to get a full temperature profile of the furnace and the condenser together, but not at the exact same time. The first set of readings was taken of the furnace itself without the condenser installed, after the furnace operated for 1.5 hours. It was necessary to remove the condenser for these measurements, as the temperature probe was not long enough to fit all the way through the furnace with the condenser installed. The next set of measurements was taken after 2.5 hours, after re-installing the condenser. Part of the furnace was re-profiled, in addition to the full length of the condenser.



These measurements were rather difficult and time consuming to do correctly, but as can be seen, the temperatures along the two curves still do not match up exactly. This is because even after 1.5 hours the furnace continued to heat up during the following hour.

From looking at all three charts, I think it is apparent where the hottest zone is located in the furnace, and also that the condenser parameters are adequate. After completing this one full temperature profile I don't think it is necessary to record other ones, even when other parameters such as air flow or furnace wattage change. It should be sufficient to measure the temperature at one location in the middle of the hot zone.

Unfortunately, I think it is necessary to add a PID controller to the furnace. I already have one so this is no problem for me. It is an added expense for others, though. At the same time it simplifies control of the furnace. I am currently providing power via an adjustable DC power supply, and most people do not have one of these high power units. In contrast a PID controller with an SSR can run directly from the AC outlet.

I currently have a flow meter on order for the gas supply. I am getting one of these for development purposes, but I don't think it is necessary for everyone to get one of these. It is possible to measure air flow by measuring how fast the air can push water out of an inverted beaker. If it pushes out 1 liter in one minute, then that is 1 lpm air flow.

One additional point is that I managed to center the thermocouple junction within the tube, so it's measuring actual air temperature, and there isn't that characteristic "hump" right there at the end of the condenser that was seen in previous charts.

That is all for now, probably until next weekend.

[Edited on 8-13-2018 by WGTR]

Quote: Originally posted by Heptylene

Thank you for the measurements, I might get away with using a 10-20 cm furnace then, although my tube is 18 mm outside diameter so extrapolating like is probably not quite correct.

Quote: Originally posted by Chemetix

Just a heads up with boro, it goes soggy very quickly as an oxidiser chamber. The catalyst will get to 700C easily. Boro starts getting a bit mushy at this stage. Quartz, aluminosilicate, or ceramic are going to be the materials of choice. I'm appreciating your ever precise building style. I am anticipating your first couple of runs. [Edited on 15-8-2018 by Chemetix]

Thanks for the compliments there, you and VSEPR_VOID. It’s much easier for me to pursue something like this when I know that others are interested in it as well.

I think that your words of caution are well-placed about the use of borosilicate tubing for this type of application, and they are probably prophetic. I don’t plan on trying out the boro right away since I have other points of investigation that are more pressing, and already have a good tube made of quartz to play with. However, I have a number of different ideas floating around back there in my head (too many to try and list), and it seems slightly possible with the right combination of variables that boro might actually work. If it did work it would be marginal for sure, but it would be cool, because I can buy 4’ borosilicate tubing on the way into work every morning if I wanted to, with no shipping charges. The same store doesn’t sell quartz.

Here’s my thought process on this: I’m not thinking of using a 10% ammonia feed. Referring to a chart from this reference https://www.sciencemadness.org/whisper/files.php?pid=525953&...



selectivity to nitric oxide over nitrogen gas begins to be noticeable around 400C with a ceria catalyst. At 500C it’s at 70-80% already. This is with only a 0.05% (500ppm) ammonia feed, so in this case I don’t think the temperature rise is very measurable during that oxidation. Using such a diluted feed would be impractical in our case, since with a 3% feed I think that yields would be around 3g per hour of acid if everything goes more or less correctly. It would be interesting to see how high I can get the ammonia feed % without getting much temperature rise from the oxidation.

Diluted ammonia feeds also give diluted nitric oxide products. These oxidize much more slowly to nitrogen dioxide as the dilution increases. This might seem to be a bug, but really I think it’s a feature. With a 7mm tube ID and, let’s say, a 600ml/min flow rate, that should translate to a flow rate of 10” per second through the tube. I’m already flowing more air than that through the condenser, and the condenser is still overkill. It cools down the air almost immediately. Anyway, the dilute nitric oxide would simply not have enough time to oxidize to a noticeable degree in the condenser due to the matter of seconds spent in the tube.

The output gasses could then be bubbled through a very cold (below freezing) nitrate salt solution in order to remove most moisture from the gasses. Some nitric oxide may be lost as nitrate in this solution, but the idea is that the losses would be very minimal. Silica gel can be used to both remove moisture, and also catalyze the oxidation of nitric oxide to nitrogen dioxide. The former reaction is much faster than the latter, however, and the overall effect is a matter of residence time. Also, moist silica gel does not catalyze nitric oxide oxidation. The first stage of silca gel would be located in a short tube to remove any residual moisture not captured by the chilled salt solution. Gasses would pass very quickly through this. Following a void space for oxidation, the following stage of silica gel would serve to catalyze the oxidation of the dry nitric oxide, and also to capture the resulting nitrogen dioxide. Gasses would pass very slowly through this catalyst bed. Practically I plan to use indicating silica gel for moisture adsorption, and non-indicating for nitrogen dioxide adsorption (it is it’s own indicator).

To measure yield, before and after weights of the second silica gel stage will be taken, and yields will be based on the weight of dry nitrogen dioxide.

Platinum Loaded Ceramic Beads

I made some ceramic beads and infused them with platinum.

20ml of deionized water was warmed gently in a 25ml glass beaker on a hot plate, and stirred vigorously using a magnetic stir bar. A glass cover was placed over the beaker to keep evaporation to a minimum. 2.00g of bentonite clay was carefully added to the beaker under heavy stirring. Eventually, all of the clumps dissolved into a brown suspension, and the solution was allowed to stir for an additional 30 minutes. Then, 18.00g of alumina was stirred in as well, and 30 more minutes were allowed for the bentonite and alumina to thoroughly mix together, and for the bentonite to fully hydrate.

The resulting suspension had low viscosity, and much extra water. The clay was de-watered by carefully pouring the contents of the beaker onto absorbent paper towels. After standing for several minutes excess water floated to the top of the clay and ran off, or was soaked into the towels. A metal spatula was then used to scrape the clay together from the towels.

With much care, small pieces of the clay were rolled by hand into round beads and pills. The clay seemed plastic enough to form, so long as the moisture remained sufficient. A bit too much moisture caused the clay to become very gooey and soft, and too little caused the clay to become stiff and non-cohesive. The right amount gave a firm clay that could still be gently shaped.

The resulting clay bodies seemed porous enough to dry rapidly, and they seemed dry under warm air after only an hour. Still, they were allowed to dry in ambient room air overnight before firing. In their green state the clay seemed fairly strong and could be easily handled, but no attempt was made to try to test the strength by crushing samples.

The beads were placed in a small kiln (linked in signature) and ramped slowly up to 1200°C over a period of about 3 hours. One half hour was spent at 100°C to make sure that everything was dry, then the kiln was ramped to 1200°C. The kiln spent 20 minutes soaking between 1200°C and 1210°C. The kiln was allowed to cool down naturally to 1000°C, and then the lid was opened partially. Cool-down was rapid at this point.

Once the ceramic bead was heated to the decomposition point of the platinum salt, the bead was cooled down and reintroduced back into the ethanol solution, and this cycle was completed four times in totality. The following video shows the fourth cycle in progress. Note how dark the bead is by this point. Practically none of the ethanol solution has been used, and it’s probably less than a few cents of platinum for each bead.

Attachment: Pt_loaded_ceramic_support_4th_cycle.mpg (4.9MB)
This file has been downloaded 678 times

Can you see the difference between coated and uncoated samples?



If you look closely you can see that there are some small pinholes that are not covered well with platinum. Looking under high mag with an optical microscope, it appears that those small areas are relatively non-porous. The ethanol solution was simply not able to soak into those areas very well.

Video file updated


[Edited on 8-25-2018 by WGTR]

Quote: Originally posted by Heptylene

Incredible! That's a nice catalyst support, I'm guessing the beads fit in the reaction tube right? They seem quite porous from their bubbling in water. How did you get access to a microbalance? I can't imagine those are common outside very specialized research labs. The best I've seen was at my uni had 10 micrograms resolution. The coating with platinum by thermal decomposition seems to be a better method than that by reduction with hydrazine which I've been using. The problem with hydrazine reduction is that most of the platinum just forms a suspension in the solution (which I recycle of course) and doesn't stick to the substrate. I'll try thermal decomposition if I need to make more catalyst. Btw I haven't been able to reply to your U2U, I'm having problems sending messages. I hope this will be fixed soon.

Tonight, I joined the Ostwald club...

For now, some pictures...

Furnace glowing:




Catalyst beads after furnace cool-down:




Jar full of NO₂:



[Edited on 8-30-2018 by WGTR]

PVC Tubing and 800°C Ceramic Beads have Compatibility Issues

I have a couple of hours of video from the last batch of experiments, but have to edit it down before posting anything. Also, there is some more video that I'd like to shoot using what I've learned from this latest go-around.

I made another catalyst support, but instead of making beads this time, I made a ceramic screen. Using the same alumina clay composition as last time, I pressed some clay into a 0.25" hole drilled through aluminum plate. The plate was used as a mold. The clay dried in the mold will practically no noticeable shrinkage. When dry, small 0.031" holes were drilled in the clay using a diamond burr bit. I used a small high speed drill press made for small work like this, but it was still very hard. Some of the hole walls are only 7 mils thick, but they did not break when drilling or handling. A previous one that I tried to make using a handheld Dremel tool was a mess. I simply was unable to hold the tool steady enough.

After successfully (!) drilling all the holes, I had to find a way to remove the fragile piece of dried clay from the mold without breaking it. It was too tight a fit to be pressed out without destroying the piece. Next time around I could make a 2 piece mold and simply take it out that way, but this time, I decided to dissolve the mold with HCl. The clay would also break apart in this solution, so I first made it waterproof by melting microcrystalline wax into the clay body. The wax soaked in like a sponge.

After this wax treatment the mold was dissolved away, and the clay piece survived with minimal cosmetic damage. The clay was then fired to 1200C, which at the same time burned away all the wax. After cooling down the ceramic screen, I accidentally gave it a drop test, and it simply bounced on the tabletop. I thought for sure it would break, but it didn't.





I used the same firing schedule that I did for the beads, and due to the very thin wall features, I think I overfired the piece a little bit. It wasn't quite as porous as the beads, but still soaked up enough platinum solution to give a good coating.

Quote: Originally posted by Heptylene

So my next step will be building a heating element for the tube to replace the blowtorch. Using a cheap lights dimmer seems to be the most cost effective way to regulate the power going through the element. Those can be bought on ebay for $2 apparently.

Quote: Originally posted by Heptylene

I find it impressive how dark the catalyst is colored given the small amount of platinum that is present. After your success with the catalyst beads, I'm sure this new catalyst will work great!

Quote: Originally posted by Heptylene

I ran a quick test this afternoon at a much reduced flowrate compared to my previous attempt to see if air cooling would be sufficient to cool the tube. I bubbled air through about 10 % commercial ammonia solution at an air flowrate of 0.6 liters per minute. The resulting gas mixture, which should consist of about 10 % ammonia by volume, was passed through the tube (18 mm OD, 16 mm ID) where the catalyst is located. I used a blowtorch to heat the catalyst to red heat. The gasses coming out of the tube were brown from nitrogen oxides and after a few minutes the tube was still cool to touch about 15 cm downstream from the catalyst. The gasses in the tube move at about 5 cm per second. A sample of the gas was collected in flask with some water and was found to be acidic, showing that a large portion of the ammonia was converted. Btw this 0.6 LPM flowrate is just what my cheap aquarium air pump outputs. Assuming a 100 % conversion, this should consume a few moles of ammonia per day.

Wow

The Blue Color of Dinitrogen Trioxide

No, the solution isn't gravity-defying, for some reason the picture just loaded sideways.





One thing that I’ve noticed when running my Ostwald reactor, is that a lot of moisture is produced:

4 NH₃ + 5 O₂ → 4 NO + 6 H₂O [1]

That’s to be expected with so much hydrogen being present in the ammonia. This may seem desirable; after all, we need water to produce the acid. One problem is that more water is produced in the reaction than we really need for this overall reaction:

4 NO₂ + O₂ + 2 H₂O → 4 HNO₃ [2]

Basically we have three times more water than needed.

A second problem is that the overall reaction to nitric acid doesn’t occur in one step. First nitric oxide is oxidized to nitrogen dioxide in air:

2 NO + O₂ → 2 NO₂ [3]

This is a relatively slow reaction, taking seconds, minutes, or days depending on concentration. At extremely dilute concentrations this reaction may effectively never complete. The reaction with water is somewhat complicated, but is

3 NO₂ + H₂O → 2 HNO₃ + NO [4]

overall under the conditions of bubbling the dilute gases through warm water or concentrated nitric acid. This reaction is reversible, and by bubbling nitric oxide through concentrated acid, nitrogen dioxide can be reformed and stripped away, diluting the acid with water. Thus it is important to ensure that ample time is given for [2] to occur in order to obtain the maximum concentration of nitric acid. Nitrous acid is an intermediate that forms during the overall reaction of [4]:

2 NO₂ + H₂O → HNO₃ + HNO₂ [5]

Nitrous acid can either disproportionate and form nitric oxide and more nitric acid:

3 HNO₂ → HNO₃ + 2 NO + H₂O [6]

or under very cold conditions can form the anhydrous dinitrogen trioxide:

2HNO₂→ N₂O₃ + H₂O [7]

The same compound can be formed directly from partially oxidized nitric oxide below 0°C:

NO + NO₂ ⇌ N₂O₃ [8]

The chemistry of nitrogen oxides and water is quite complicated, and I’m not doing it justice. The book “Absorption of Nitrous Gases” is in the Sciencemadness library, and dedicates a generous number of pages to the topic.

Anyway, armed with the previous information, I plan to make some changes to the reactor design, and borrow some ideas from Heptylene in the process. Recently I ran some clear PVC tubing from the condenser output to a container filled with ice, with the intention of lowering the vapor pressure of water enough to remove most of it from the product gases. This kind of thing has to be done quickly after the ammonia oxidation reaction takes place, so that most of the NO does not have enough time to oxidize to NO₂. NO does not react with water by itself, while NO₂ does. Without removing moisture with this ice bath condenser, moisture gradually reacts in [5] downstream. One byproduct of this reaction is the generation of heat, which in turn causes the acid product to vaporize. This mist gets carried along, where it deposits as a thin film on pretty much everything downstream.

Even with my overly-long condenser, I was still able to remove almost all of the moisture just in a couple loops of PVC tubing submerged in ice, without losing much NO. The gases in the downstream NO oxidation chamber remained clear and brown, with no cloudiness or traces of moisture showing up on the glass walls. Any NO that was already oxidized before hitting the ice bath was simply trapped in the tubing as dilute nitric acid.

The resultant NO₂ was then led into a “cold finger”, basically just a test tube that was partially submerged in liquid nitrogen. I happen to have large quantities of liquid nitrogen just sitting around so it’s handy to me, but dry ice, a freezing salt mixture, or running a tube through a cold freezer might work just as well. The result was a frozen mixture of N₂O₄ and N₂O₃, a rather pale greenish solid. Letting it warm up somewhere below 0°C allowed a rather pale-green liquid to gather at the bottom of the test tube. Slowly adding a few drops of water (this is exothermic, so this has to be done gradually, and with cooling, to keep the mixture around 0°C), caused the formation of two layers, one clear, and another one deep blue. The deep blue layer is quite dense and is primarily N₂O₃, whereas the upper layer is concentrated nitric, up to 60-70%, with possibly some N₂O₄ still dissolved in it. It is not possible to obtain greater than 60-70% nitric by only adding water to cold N₂O₄, because of the equilibrium that is established between nitric acid and N₂O₃. If 100% acid is desired, oxygen gas can be introduced to the mixture, and the blue N₂O₃ layer will slowly disappear as it oxidizes and forms more nitric acid. If red fuming acid is desired, less water is added than theoretically required, while adding oxygen.

Getting back to my planned design changes, I think that Heptylene has a good idea, and that a water condenser isn’t needed on the quartz reactor tube itself. The tube cools down below 300°C in air before even 15-20 cm or so. Kapton tape is usable up to 300°C, and I think that it’s possible to join the quartz tube to a length of borosilicate or soda-lime tubing while the gases are still fairly hot, by wrapping the joint with Kapton. The advantage to this is that even borosilicate tubing is pretty easy to shape over a propane torch, and can be shaped into a “U” or something that can be submerged in ice. An additional advantage is that I would feel more comfortable turning my back to something like this while it’s running, because I wouldn’t have to worry about the drain backing up in the condenser and running water out under the door into the hallway, if you know what I mean. Finally, it would allow a pretty short tubing length after the reactor, allowing minimal loss of NO to oxidation in the ice bath. Further design details need to be worked out in the lab.

Quote: Originally posted by Chemetix

This was what I found to be the limit of the concentration available from a simple NH3 oxidation, I was getting a crude value of 8M from the first condensation. The answer I suspect is a long length of black irrigation tube 30 or 60 meters long in a coil, and about 25mm dia for long residence times for the oxidation to NO2 and the subsequent reaction with water to re-emit NO in a iterative sequence towards full absorption. As for concentrating the acid once made, does a distillation from calcium nitrate seem a reasonable way to break the azeotrope with water? Seems a little easier than working with semi cryogenic oxidations with oxygen and N2O3 with water. Although the chemistry has always fascinated me since reading about N2O4 as a rocket oxidant, The image of blue N2O3 is really amazing.

Quote: Originally posted by Heptylene

@WGTR Kapton tape! I knew that some gas chromatography columns (which are heated in an oven) were made of polyimide, but I had no idea that there existed polyimide tape! Since I have 3 pieces of quartz tube (18x500 mm) I could always join them if needed in the future. Or join a borosilicate condenser for that matter. I should take a look at that book on nitrous gasses, looks interesting and directly relevant to what we're doing. @Morgan I had always thought of the catalyst as something that would not deteriorate. Let's hope this loss of platinum will not be too significant for an amateur reactor. Depositing additional platinum won't be expensive given the small quantities needed, but might be time-consuming if it has to be done too often.

Yes, Kapton is like some kind of miracle tape! Here, I made some changes to the reactor design this evening. The furnace was cut in half (9" long now), and the water condenser was removed and replaced with some borosilicate tubing in an ice bath. The quartz reactor tube is joined to the borosilicate tubing with Kapton tape. The boro was wound into a spiral condenser by hand (easy with a propane torch). Power input to the furnace was cut from 300W to 150W, saving energy. It could probably be less than half of this if I insulated things better. I keep the furnace running at bright red. Right now the catalyst is the honeycomb type, with a 300ml/min air flow rate, and an ammonia solution of 6% concentration. The tubing has 7mm ID, the same as before. I believe it is possible for someone to use a 12" long quartz tube in this application, meaning that it should not be so expensive for people to obtain this.




I looked at my earlier temperature profile measurements to determine how much quartz tubing length I needed outside of the furnace before the Kapton joint. Right now I have it at about 4" of spacing, and the joint stays barely warm to the touch. The Kapton seems to be holding up OK.

Unfortunately, I didn't realize that I've been working on this project for a month and a half now. I'm getting very behind on some other stuff, and need to move on after the end of this weekend. Are there any questions about my current setup, requests for better pictures, etc? I'm hoping that there is enough information in our posts at this point, between you, Chemetix, others, and myself, for someone else to put something together and get it working.

Here's some home-made copper nitrate:

Yeah, the ammonium salt has the advantage that there's no leftover residue other than platinum itself. It's worth the extra effort to make it, and a 100ml bottle of the weak solution should last you forever. I used ethyl alcohol as a solvent. I don't know how much effect that has on how easy the salt is to reduce later since I didn't try using water as a solvent, and my catalyst support was also quite absorbent.

I did find some high silica content wool in the lab, but I personally wasn't too happy with it. It felt like fiberglass, and after being heated with a propane torch became very brittle and friable. The stuff just broke into pieces just the right size to make hands and lungs itchy. If you don't mind me asking, what country are you located in?

If you can't get some insulative firebrick, then perhaps some ceramic wool would work. You could ask Chemetix what he was using. I think you'll need some kind of insulation, although you could compensate by using lots of power in the heater.

Chemetix, I'm not sure that I understand your idea either, but it's probably because I'm either making or not making certain assumptions about it. Could you describe the theory behind how it should work?

For the benefit of people reading, be careful handling platinum salts. The cautions about platinum salts being strongly sensitizing for people are no joke. Gloves are recommended. This stuff can sensitize people to various skin ailments at any time in their lives. Certain people are genetically predisposed towards skin sensitivities. I've known people who lived a good part of their lives with no problems, and then got some cleaning chemicals under a ring one day, developed a rash on the entire hand, and now randomly battle severe skin rashes than can only be mitigated with strong topical steroids, years later. Just practice good lab hygiene and one should be fine.

The ammonia scrubbing tower idea needed a bit of a diagram it seemed. So, have a look at the schematic and tell me what you think.

Quote: Originally posted by WGTR

Yeah, the ammonium salt has the advantage that there's no leftover residue other than platinum itself. It's worth the extra effort to make it, and a 100ml bottle of the weak solution should last you forever. I used ethyl alcohol as a solvent. I don't know how much effect that has on how easy the salt is to reduce later since I didn't try using water as a solvent, and my catalyst support was also quite absorbent. I did find some high silica content wool in the lab, but I personally wasn't too happy with it. It felt like fiberglass, and after being heated with a propane torch became very brittle and friable. The stuff just broke into pieces just the right size to make hands and lungs itchy. If you don't mind me asking, what country are you located in? If you can't get some insulative firebrick, then perhaps some ceramic wool would work. You could ask Chemetix what he was using. I think you'll need some kind of insulation, although you could compensate by using lots of power in the heater. Chemetix, I'm not sure that I understand your idea either, but it's probably because I'm either making or not making certain assumptions about it. Could you describe the theory behind how it should work? For the benefit of people reading, be careful handling platinum salts. The cautions about platinum salts being strongly sensitizing for people are no joke. Gloves are recommended. This stuff can sensitize people to various skin ailments at any time in their lives. Certain people are genetically predisposed towards skin sensitivities. I've known people who lived a good part of their lives with no problems, and then got some cleaning chemicals under a ring one day, developed a rash on the entire hand, and now randomly battle severe skin rashes than can only be mitigated with strong topical steroids, years later. Just practice good lab hygiene and one should be fine.

Quote: Originally posted by Heptylene

The resulting gas mixture, which should consist of about 10 % ammonia by volume, was passed through the tube (18 mm OD, 16 mm ID) where the catalyst is located.

Quote: Originally posted by Heptylene

I live in Switzerland. A bit of a nanny state chemistry-wise, but a nice place nevertheless. WGTR I assume you are in the US?

@Chemetix: I like the idea of an hydrometer, I'll keep that in mind when building the ammonia generator.

@WGTR: I don't have a caliper to measure the ID of the tube, but using a ruler I can say its at least close to 16 mm. Not 15 nor 17 mm. Why do you need a precise measurement of the ID?

The setup you propose would be useful to regulate the composition of the gas feed. One thing I don't like about it is the use of calcium oxide, which is consummed and has to be bought, or regenerated by heating Ca(OH)2 in a furnace. This could get expensive in the long run. But for testing purposes this should be a pretty stable source of gas.

The same idea could be used if, for instance, the gas generator is a 25 % ammonia solution.
I had originally intended to dilute 20 liters of 25 % ammonia to 50 liters of (approx.) 10 % solution and use that directly. But this would be wasteful as the solution would become too dilute at some point, wasting the remaining ammonia unless there was a way to concentrate it. If "too dilute" is a concentration below say 5%, then half the ammonia would go to waste. Fractional distillation could be used,but this would be time-consuming.

Instead ammonia gas could be stripped from the 25 % solution (call it "A") with air and absorbed through 10 % solution (call it "B") and fed into the reactor. If solutions A and B have about the same volumes, solution A should be progressively depleted and solution B should be enriched in ammonia and then be depelted too.

I made a drawing of the concentration curves vs time for both solutions (A and B). This is scenario 1. The drawing isn't pretty but should get the point across. The curves are only intended to show the general trend.

If instead one uses a 10 % solution only, the depletion should be much more rapid (Scenario 2 , solution B*).

I hope this makes sense, its late and my mind a bit foggy.

Quick update:
I got my nichrome wire in the mail yesterday (30 AWG/0.25 mm diam.). My original idea was wrapping the wire directly onto the quartz tube, but the wire doesn't stay in place on its own. In fact when released, the diameter of the wound coil approximately doubles. I might use a larger diameter wire (0.5-1 mm) that holds its shape more easily. This would also allow the use of lower voltages across the coil. Annealing the wire in place on a smaller tube and sliding the formed coil on the larger tube could work.

Regarding the catalyst, I recently had access to a stereo microscope and took a picture of silica thread with and without platinum deposit at 30x magnification. I thought this could be interesting to share.

The catalyst was prepared by dipping some silica thread in aq. H2PtCl6 (0.49 g Pt/100 ml) and heating the thread to 500 °C with a heat gun. This was repeated 3-4 times.

The two red marks are 4 mm apart to give a sense of scale.

Left: without platinum, right: with platinum.


You can see that the coating seems uniform. In fact I've observed some wicking of the Pt solution to the edges of the strand which we can't see here. When heated to dryness, the Pt solution refluxes and deposits more solid at the ends and surface of the strand than in the middle. Also to note is how fine the silica fibers are. Much finer than a human hair.

I also bought some silica sleeve to insulate the heating element, but it turns out to be too tight to slide over the coil, so I have yet another source of silica for the catalyst. The strands can be more easily separated than those of silica wick.

.

I finished the heating element today. I wound about 5 meters of 0.25 mm diam. nichrome wire into a 4 mm coil, stretched it a bit and wound the coil around the silica reaction tube (about 8 turns I think).

The nichrome wire on each end was secured to the tube using some silica sleeve. The wire goes under the sleeves and the sleeves themselves are tightened on the tube with some wire. Then silica thread (6 mm diam.) was wound around the silica tube between each turn of nichrome coil to provide electrical insulation. To provide some thermal insulation and rigidity, mica sheet (obtained from an old heat gun) was wrapped around the heating element and secured with 2 knots of nichrome wire.

I plan to wrap some rockwool/kaowool around the whole thing to provide additional thermal insulation.

I made a side-view diagram of the heating element:


The nichrome 4 mm coil:


The finished element (but still lacking sufficient insulation):


I briefly tested the heating element by connecting it directly to the mains (240 V). I'll be getting a multimeter shortly, but from what I've calculated it should have a resistance on the order of 100-150 Ω, giving a dissipated power of about 400-550 W. The heating element was glowing orange, although it was a bit too hot to be run for any length of time. A dimmer will be used to reduce the power slightly. I'm hoping the direct contact between the heating element and the quartz will not be an issue. Otherwise I will have to build a proper tube furnace like yours WGTR.

I might end up doing that anyway since a tube furnace would be handy to have for other projects (acetaldehyde from ethanol, anyhdrous element chlorides such as AlCl3, SiCl4, etc.)

I also followed up on the catalyst support I propsed in my previous post. I placed some silica fabric at the bottom of an evaporating dish and soaked it with 0.8 g of aq. H2PtCl6 solution (about 4 mg of Pt). The dish was heated on the hotplate to 500 °C for 20 minutes. The Pt deposited uniformly on the surface and was capable of decomposing 12 % hydrogen peroxide almost instantly. I'll upload a video on youtube at some point. I think this silica fabric will be the final choice for catalyst support. Its easy to work with and easy to coat with platinum. Much more practical than silica wick in fact.

Silica fabric soaked with Pt solution:


Silica fabric after heating:


[Edited on 9-10-2018 by Heptylene]

Update: Tube furnace finished

I recently completed my tube furnace for the Ostwald reactor, so here are some picture:




When the core of the furnace is at 800 °C, the tube is at about 50 °C on at the outlet, with 0.6 LPM of air running through it.

The two metal angles serve to hold the connectors for the nichrome wire. The conductors are in open air, which makes this thing very unsafe. I'll probably build a better (non-deathtrap) furnace at some point. The metal box is a 3$ dimmer from ebay rated for 4000 W. I would trust that thing with 4000 W though. The nichrome element resistance is about 130 ohms. At 240 V that should be about 450 W of power. On the picture it was running with the potentiometer about halfway, whatever that tells you. I have no way of measuring the duty cycle of the dimmer so I don't know the power consumption. I know full power is too much for the element to handle.

The temperature inside reached about 800 °C. Then the glass sleeve of my thermocouple started to melt and fuse to the silica tube.

Anyway that's it for now.

[Edited on 31-12-2018 by Heptylene]

[Edited on 31-12-2018 by Heptylene]

Thank you WGTR!

The tube is held at both ends using plumbing pipe clamps. The clamps initially had a rubber pad which I removed and replaced with a few layers of regular paper. This is to allow the tube to slip when it expands due to heating up. The clamps are not tightened much, to avoid breaking the tube. The clamps themselves have a threaded base that screws on threaded rod (M8). Overall I'm satisfied with what the furnace can do, but there is a major design flaw: if the silica tube breaks for whatever reason, I cannot remove it, I'd have to break the heating element apart.

When I finished the furnace I did a quick test and tried to pass some dilute (3% ± ish) ammonia/air mixture over the heated Pt catalyst and got some brown nitrogen oxides, with no visible ammonium nitrate residue. I still have to figure out what to do with the gases. Maybe dissolution in NaOH, as I don't really need nitric acid.

I'll also make some acetaldehyde eventually, by passing ethanol vapor over hot copper wool.

But I've been busy recently with college and other projects (namely a 150 kV power supply for x-ray experiments). It's been months since I did any proper chemistry!

P.S. There is a typo in my last post regarding the dimmer rated for 4000W: I meant to say "I wouldn't trust that thing with 4000 W."

Counter-current gas absorption tower!

I've been working on and off on my Ostwald reactor over the past year, and have improved the gas absorption system. Previously I used a simple glass tube leading into a (100 ml) graduated cylinder filled with water. This hardly worked at all and a lot of nitrogen oxides escaped into my workspace, which is both dangerous and wasteful. Indeed I haven't produced any usable quantity of nitric acid using this method.

The new system is a counter-current gas absorption tower, inspired by systems used at the pilot or industrial scale. See for instance this system by chemglass and a general illustration of the process:



I am currently testing this new system. So far, absorption seems to work well, and only a little NOx can be detected by smell at the outlet, but a little still goes through so I've added a simple NaOH scrubber at the outlet (inverted glass funnel in a beaker, classic type).

The heart of the system is a 500 mm long column, 30 mm O.D. filled with 10 mm PTFE raschig rings. Water is continuously passed over the packing from top to bottom. The gasses (nitrogen oxides here) are introduced at the bottom and escape at the top. This ensures intimate contact between gas and liquid. The liquid is pumped using a polypropylene centrifugal pump (similar to Iwaki-type pumps, 50 $ on aliexpress) at a rate of a few liters per minute. The liquid is sprayed on top of the packing using a homemade nozzle with 1 mm holes. The liquid at the bottom of the column is collected in a chromatography reservoir and withdrawn from the bottom of said reservoir by the pump. The pump is placed below the reservoir as it is not self-priming: Liquid must be present for it to start pumping, it produces no air suction when dry. In fact, the pump loses a lot of head pressures when even bubbles are present in the pumped liquid. An advantage of placing the pump at the very bottom is the ability to purge virtually all the liquid simply by turning a stopcock ("product outlet valve" in the pictures). This will make testing easy, no need to disassemble anything to get a small sample or recover all the product. The sep funnel is unrelated to this. It is simply a way to collect the condensate that comes out of the furnace so it doesn't stay in the hoses.

For the packing I used 10 mm PTFE raschig rings because the 6 mm raschig rings I already had (for distillation) did not let enough liquid pass through, which caused column flooding. I had some PTFE tubing on hand so I chopped it into pieces (10 mm long) so I didn't have to buy a whole kilo of glass packing from my supplier. I think 8 mm raschig rings would pack better, the 10 mm one tends to make channels, which isn't good for liquid-gas exchange.

A quick note about the connections from the hose to the quartz tube furnace: I replaced the rubber stoppers with GL-32 to ground glass adapters. The GL-32 thread can be used as a compression fitting with PTFE/silicone joints and a cap (red) and form an air tight seal with the quartz tube. There is no problem with thermal expansion and chemical resistance either. This provides my tube furnace with two male ground glass joints, to which I simply attach classic hose barb adapters, or in this case a distillation receiver and a sep funnel.
I have also used this system for liquid connection at the bottom of the reservoir, GL-18 thread this time.

The next step is to test the system for an extended period of time and titrate samples of the liquid over time. I actually got a burette for this specific purpose. I also want to play with the temperature of the furnace and the ammonia/air feed composition. It seems at high temperatures almost no NOx are produced, which I attribute to over-oxidation of ammonia to dinitrogen. I don't have a probe to measure the furnace temperature yet, so I'll need to get that first. So far, I've been running with air and measured the temperature with the furnace open (around 800 °C). I can only assume the temperature is close when running on ammonia/air, but it's likely to be higher.

The column could definitely be larger in diameter and longer. I think 50 mm by 1 meter would be good but I couldn't find a column this large. Ideally I would like a 100 mm diameter column with duran flanges, but apparatus at this scale is getting very expensive. Possibly a heat exchanger for the liquid, as it's getting warm under operation (maybe 30 °C?). Time will tell if this is an issue.

Below are some picture of the apparatus (with legends). Zoom in to see the text.

Quote: Originally posted by WGTR

The idea is to keep stacking reservoirs and condensers until this becomes impractical.