Sciencemadness Discussion Board

Sulfur trioxide from sodium hydrogen sulfate

garage chemist - 23-3-2008 at 08:54

With this preparation I think I have found the easiest, cheapest and highest-yielding method of sulfur trioxide production for the amateur chemist.

The idea of making sulfur trioxide by heating sodium hydrogen sulfate has been brought up since the beginning of discussion on oleum production on this board, more specifically here:
The old Oleum & SO3 thread
But as far as I know, I am the first one to have success with it.

For more information about other methods, look at the ferric sulfate method, the persulfate method and the vanadium(V)oxide catalysed method.

Sodium hydrogen sulfate gives off water when heated at 300- 500°C and turns into sodium pyrosulfate:

2 NaHSO4 -----> Na2S2O7 + H2O

This in turn gives off SO3 at 680- 880°C:

Na2S2O7 -----> Na2SO4 + SO3.

The temperatures stated are the ones I observed, they were measured at the outside of the test tube.
In literature you will find very much lower temperatures- I assume those are the temperatures at which the corresponding decomposition just starts, not the ones at which the decomposition goes at a useful rate.

The two reactions are not cleanly separated. Even when one heats at 480°C until absolutely no more water vapor is being produced, as I have done, some concentrated H2SO4 will come over from 500- 670°C, no matter how slowly the heating is carried out.
The yield of SO3 coming over at 680- 880°C is still extremely good, though.

But what makes the NaHSO4 method so much better than the ferric sulfate method, which operates at a similar temperature, is the fact that the SO3 does not decompose to SO2 and O2 at all.
The equilibrium for the thermal decomposition and formation of SO3 indicates that ca. 70% of the SO3 should be decomposed at 800°C- see the diagram that I posted in the ferric sulfate method.

I cite Ullmann's Encyclopedia:

Pure sulfur trioxide is in fact extremely resistant to thermal decomposition because of kinetic inhibition, even at elevated temperatures where thermodynamic equilibrium is shifted heavily toward SO2 + O2. However, certain catalytically active substances are able to increase the rate of equilibration substantially. In the presence of metals such as platinum or of metal oxides and sulfates (e.g., of iron, copper, and, of course, vanadium) the decomposition approaches equilibrium at temperatures above ca. 700 °C.

The ferric sulfate/oxide in my ferric sulfate method catalyzed the dissociation of SO3.
No iron or other heavy metal compounds are present in the NaHSO4 method, hence no dissociation of SO3 here.

This explains the unusually high yield of SO3 obtained from NaHSO4 and the complete absence of any SO2 smell.


The precursor is technical grade sodium hydrogen sulfate, purchased from the pool section of the home store.
It is used for lowering the pH of swimming pools, and comes in buckets of 1,5kg.

Apparatus and Procedure

100g NaHSO4 were filled into a quartz glass test tube with joint, the same one I had used for the ferric sulfate pyrolysis.

After melting, the melt filled only half of the test tube.

This was then slowly heated in the tube furnace. At 300-450°C there was a strong evolution of water vapor after melting.

It was held at 480°C for 45min, after which the steam evolution had ceased entirely.

Upon further heating from 500- 670°C, small amounts of concentrated H2SO4 slowly distill off. To capture and condense the vapors, I placed a round-bottom flask in front of the test tube:

A quartz joint with a bent tube would have been much better to use, but I didn't have such a thing.

At 680°C, SO3 evolution started. A quartz glass extension was put on the test tube, and a glass joint adapter carrying a bent glass tube was put on the extension. The glass tube led into a cylindrical receptacle cooled with water from the outside, so this was essentially the same apparatus I used for the ferric sulfate pyrolysis.
Since the SO3 is pure and not diluted with SO2, no ice is necessary for cooling.

The SO3 had a tendency to crystallize inside the glass tube, potentially leading to plugging. To remedy this, I increased the heating power on the tube furnace so that the SO3 came over faster and at a temperature above the melting point of solid SO3. Here is a picture of the SO3 in the tube melting:

Some crystals of solid SO3 deposited on the walls of the receptacle, while the bulk of the product collects below the cooling water surface:

The evolution of SO3 was complete at 880°C. I briefly heated to 930°C to melt the sodium sulfate into a single chunk:

And this is the product after solidification- an astounding 23,8g of SO3, from 100g NaHSO4, a 71% yield!

I think the yield of the NaHSO4 method speaks for itself. This is a much better method than the ferric sulfate pyrolysis.

Adding to that, the precursor can simply be bought OTC instead of having to prepare the ferric sulfate precursor beforehand.

Now we have a high-yielding and cheap method of SO3 production.
This should open up lots of possibilities- the synthesis of TNT, which requires oleum in the last step, springs to mind.

I have already explained in the other thread how useful SO3 and oleum are for the synthetic chemist. Thionyl chloride, dimethyl sulfate, chlorosulfonic acid, Oleum as dehydrating agent etc... are just a few applications of SO3.

If you choose to make SO3, please make yourself aware of the dangers of this substance first.
The warning "Extremely corrosive" does not really express the dangers appropriately. Oleum destroys many plastics, like PE and PP, that are completely resistant to 98% H2SO4- only glass and fluoropolymers like PTFE are really safe.

The reaction of SO3 with water is explosively violent. SO3 must only be diluted with concentrated H2SO4!

SO3 and Oleum also create an extremely dense fog when handled in open air, due to reaction of gaseous SO3 with aerial moisture forming droplets of H2SO4. If using a fume hood, be aware that all this thick fog will come out the exhaust! Most filters are incapable of precipitating H2SO4 fogs!

[Edited on 24-3-2008 by garage chemist]

DJF90 - 23-3-2008 at 09:28

Awesome write-up garage chemist. A couple of questions I'd like to ask though. Would a bunsen/other gas torch flame be sufficient, so long as the required temperatures can be achieved? What alternatives are there to the quartz test tube?

garage chemist - 23-3-2008 at 09:43

I have not been able to reach the required temperature with a bunsen burner in my many experiments with NaHSO4 years ago.
I only got to the stage where conc. H2SO4 slowly distills off, never to the SO3 evolution.
If you can reach the required temperature with a burner, why not. But you should definately do the pyrosulfate formation separately in an open crucible beforehand, as quite long heating is necessary for this.

I am afraid standard borosilicate glass becomes soft at 600-700°C. Can anyone confirm this?

There is a glass called Supremax that can take temperatures of up to 1000°C, this would be very useful here. You will have to look for it, it certainly isn't common.
Quartz glass is more common, and a glassblower can make custom parts like this test tube and the extension from it.

A fired and glazed retort made of clay or porcelain would be another good container for the SO3 production. If you know someone who does pottery as a hobby, you could ask him/her to make such a thing for you, or make it yourself and have him/her fire and glaze it.
This could potentially be cheaper than a quartz glass test tube.

Metals are unsuitable as a container material as far as I know, as SO3 is a powerful oxidiser especially at elevated temperatures and already at room temperature.

[Edited on 23-3-2008 by garage chemist]

DJF90 - 23-3-2008 at 10:22

I was at the hardware store the other day, and noticed the propane/butane blowtorches. On some of the packaging it listed the temperature that could be achieved by the torch. The ones I was looking at had temperatures of 1300C, 1370C and 1600C, and so should be more than sufficient for this purpose. The main problem for me would be the reaction vessel. I also believe that borosilicate glass starts to become soft in the range of 600-800C, rendering it unuseful for this particular use. I'm assuming any metal reaction vessel is also unusable due to the catalytic effect it may provide on the decompostion of the sulphur trioxide. Would "fire cement" (a refractory, probably good up to 1300C) be suitable for making a retort? Unfortunately I do not know the composition, but I would expect that it contains substances like alumina, silica and magnesia. These will react with the sulphur trioxide yes? What do you suggest to glaze the retort with? Sorry, I do not know much about pottery :(

garage chemist - 23-3-2008 at 10:46

The stated temperatures on the blowtorches don't mean anything. They are just the flame temperatures.
I can easily melt thin glass tubes with my bunsen burner, but a 100ml flask with sodium pyrosulfate I can't get to 700°C with it.
What matters is the total heat output and the size of your vessel.

As a general rule, high temperatures are much easier reached with electrical heating than with a burner.
You need to get some Nichrome resistance wire, coil it around your vessel and thermally insulate it with kaowool, just as I have done with the ceramic pipe of my tube furnace.
Nichrome wire is good for max. 1200°C surface temperature, meaning at least 1000°C in the vessel.

Yes, the catalytic effect of almost all heavy metals makes metal vessels unsuitable here, in addition to the corrosion issues.

I don't know if fire cement would be resistant. I think it would be very difficult to make a gastight retort from this.

There certainly are forums dealing with pottery, as it probably is a more popular hobby than chemistry. Just search for one and ask there how they would make such a retort and a gastight connection to a glass pipe for it.
I don't know much about pottery myself either.

microcosmicus - 23-3-2008 at 11:05

The issue here is not so much temperature as heat.
Pretty much any flame is going to achieve 900C.
However, with an open flame most of the heat is
going to be lost to surroundings, so little actually
stays in the tube to heat it. The best remedy for this
is to make some sort of oven and confine the flame
in there --- you can make yourself a nice little
furnace to accomodate your test tube or retort
with some firebrick and furnace cement.

Fire cement is likely porous so wouldn't work too
well. Use it for building a furnace instead. As for
glaze, that is simply a layer of glass deposited on
the ceramic surface to make a non-porous layer.
I would suggest clear glaze over colored ones
just in case the colorant might leach out and
react (after all, we usually make labware from clear
glass) but, other than that possible glitch, there
should be no problem --- the glaze is not going to
react with SO3 any more than with the glass used
in the set-up above.

As for how to manufacture such a retort, you could
pinch the clay into shape by hand, turn it on a
wheel, or make a plaster mold and cast in with
slip. Then you would have to dry it, fire it, glaze it,
and fire the glaze. Just as with glassblowing, the
question here is whether you are willing to learn
a new craft and set up a shop with tools and
equipment (notably the kiln) or rather find (and
perhaps pay, at least worth a bottle of beer) a
craftsman to do it for you. Personally, I do both
glassblowing and pottery as hobbies parallel to
my scientific hobby and can say that, if you are
good with your hands and enjoy that sort of thing,
then doing it yourself is a good idea and doubles the
fun (both make and use the apparatus) but if
ceramics does not suit your style, you would do
better to find a potter or look around for a ceramic
item that could be used as a retort. The price of
equipment and materials spent in learning is not
likely to come out to less than finished items if
you only need a few things.. Also, if you want
to try your hand at making it, ther might be a "make
your own pottery" place where you can go and
make the retort and they will fire it for you.

[Edited on 23-3-2008 by microcosmicus]

woelen - 23-3-2008 at 11:13

This is the best synth of SO3 I have ever seen on the net. I do not have such a nice heating setup, but could it be done with a normal test tube (16 mm wide, 16 cm length) made of supremax glass and heated with a big propane blow torch at the bottom? Although this does not work with 100 grams, it could work with 10 grams, making 2 grams of SO3, which I lead directly into some conc. H2SO4.

If I only could make a few tens of ml of 30% oleum, then I would be very happy with that. I can obtain oleum, but I have to pay EUR 50 for half a liter, and I have to pick it up in a place 250 km from my home. Not very good, so this cheap source of oleum would be very good.

garage chemist - 23-3-2008 at 11:23

If you only need such small amounts of SO3, the P2O5/H2SO4 method would be a much better choice, as this can actually be done in normal glass flasks. 100g P2O5 in 75ml conc. H2SO4 are a good ratio I hear- look at the synthesis of chlorosulfonic acid on

The_Davster - 23-3-2008 at 12:30

Originally posted by woelen
I can obtain oleum, but I have to pay EUR 50 for half a liter, and I have to pick it up in a place 250 km from my home.

Wow! Thats a pretty good price...we ordered some from aldrich at work, and it was several hundred dollars for 500mL.

I have the heat source and obviously the bisulfite, and I can get a quartz tube made.
The tube you used to lead the SO3 into the cooling vessel, you used quartz, I imagine a borosilicate tube would break or break the quartz tube due to the different thermal expansion coefficients?

I also wonder in terms of minimal quartz glass parts, if a small quartz beaker could be heated to the desired temperature, and a round bottom flask of cold water held a bit above the beaker would work to isolate small amounts of SO3

garage chemist - 23-3-2008 at 12:41

No, the tube was ordinary glass. Only the test tube and the extension are quartz.

woelen - 23-3-2008 at 13:31

Price has changed since last time I checked. Now the minimum offtake is 1 liter of 65% oleum for EUR 107 excluding VAT.
Probably in absolute terms this is not expensive, but for me, this is quite a lot of money, which can buy me many other interesting items.

Magpie - 23-3-2008 at 20:26

Congratulations GC for developing this elegant method for making SO3! You continue to prove the high utility of your newly developed tube furnace where you also made pioneering innovations. It is tantalizing to think that such powerful reagents as thionyl chloride and chlorosulfonic acid will soon be within the reach of the home chemist. ;)

len1 - 23-3-2008 at 23:27

This is very interesting. Essentially its like the persulphate method because after the low-T phase both give you sodium pyrosulphate which actually decomposes to SO3. The difference being that the bisulphate dehydrates in the initial phase, whereas the persulphate liberates O2. The difference in yields between this and what you posted for the persulphate is large - I presume its entirely due to the later not being heated to nearly the temperature of the present method.

The difficultly with the method is its hard to get the quartz tube. There are several alternatives - I was wondering if you considered using an all iron construction, steel is almost impervious to conc. H2SO4 where its used as a lubricant, while it'll be interesting to see what effect it actually does have on the yield. My suggestion - not very much - contact with the walls is far less efiicient than with the Fe2(SO4)3 mix the gases had to pass through last time.

[Edited on 24-3-2008 by len1]

garage chemist - 24-3-2008 at 04:33

Yes, you could simply use sodium persulfate instead of bisulfate and probably get more than 80% SO3 yield because the substance is anhydrous to begin with.
When heating sodium persulfate over the bunsen burner without H2SO4 addition I never got any SO3, simply because I did not get it hot enough. Now I know that persulfate would work extremely well without H2SO4 addition, it just needs MORE HEAT.
The H2SO4 addition seems to enable a small amount of the SO3 in pure sodium pyrosulfate to be released below 600°C- but only a small amount, not more.

The essence of this story is that you are going to miss a lot of interesting and useful things if the bunsen burner or a blowtorch is your strongest heat source in the lab.
Everyone who does experimentation on his own should make an electric heater capable of reaching 1000°C or more, it really pays off!

Im am absolutely confident that we will eventually achieve a valid and workable synthesis of Phosphorus if people would just start working on it again, and this time use electric resistance heating.
Just look at what BromicAcid has done in this field in the past, it reminds me a lot of my early experiments on SO3 production with NaHSO4! He -or someone else- needs to do those experiments again with an electric furnace.
And for Phosphorus, metal would be a suitable construction material for the retort.

For SO3, I cannot see iron or even stainless steel working though. Contrary to cold conc. H2SO4, hot concentrated H2SO4 attacks steel very strongly, by oxidising it, with evolution of SO2. SO3 would do the same.

[Edited on 24-3-2008 by garage chemist]

microcosmicus - 24-3-2008 at 08:13


he essence of this story is that you are going to miss a lot of interesting and useful things if the bunsen burner or a blowtorch is your strongest heat source in the lab.
Everyone who does experimentation on his own should make an electric heater capable of reaching 1000°C or more, it really pays off!

Or enclose the flame ---- a blowtorch can serve quite well as the burner
for a small furnace. Your typical plumber's torch is going to put out
a kilowatt or more of power, likely several when run at full blast, but
open flames are inefficient, with only a few percent of the heat produced
going towards heating the material in the retort. However, once you
enclose the flame in a furnace, the same blowtorch which could not
heat 100g of bisulfate all the sudden can easily melt 100g of brass.

As far as power output, flames are generally going to put out more
than electrical heat. Rather, the advantage of electricity is that it
can be controlled more precisely and that no chimney is required.

Fleaker - 24-3-2008 at 10:35

Excellent work!! This is a cheap and effective synthesis for SO3, one I will probably do at home as well. The only limitation this method has (shared in common with your other attempts GC) is that it is a batch process and is not a continuous means of producing SO3. However, this could easily be scaled up and run inside an improvised kiln.

I firmly believe that stainless steel would be suitable for contact with SO3, at least at the lower end.I used it before when working with my vanadium pentoxide method. Borosilicate will not work at all for this, it stresses* from 600C on and is unusable for this synthesis. Vycor would work, and it's cheaper than quartz.

As a correction to garage chemist: PTFE, PFA, and FEP and other fluoromers are not suitable for use with SO3 at temperatures greater than 170 C. In my personal experience, significant charring occurred to both FEP and PFA tubing. PTFE handles it the longest, but will discolour. I have pictures of what it did to FEP tubing in the ''SO3 by catalysis thread'' along with my stainless steel setup which held up flawlessly to ~650 C.

@microcosmicus--it is true that enclosing your heat source makes a massive difference. For us at home, it is easily done with kaowool and kiln brick: I build improvised small 'furnaces' with a kiln brick as a base, and a rolled up bit of kaowool blanket. I highly recommend that anyone interested in high temp. work or catalysis, or distillation (I find more uses for this material every day) buy some kaowool ceramic insulation! It is extremely useful for many things, I can not emphasize that enough!

Edit: correction from S.C. Wack noted--at 600*C, all Pyrex formulations are not yet plastic, but they do begin to strain, and upon cooling may very well break. They do begin to melt from 850 Centigrade on (according to Corning's site), so around orange heat.

[Edited on 24-3-2008 by Fleaker]

Twospoons - 24-3-2008 at 13:17

Seems to me the needed piece of kit is a small ( say 1 cu.foot?) electric ceramic kiln. With a top temp of around 1250C it would let you make a nice slip-cast porcelain retort, and with slight modification would become the heat source for the SO3 reaction as well.
Little kilns like this go for resonable money on ebay - usually being sold off by rich, middle aged women who have dabbled in ceramics, got bored, and are selling off all the expensive gear they've bought :D

DJF90 - 24-3-2008 at 14:03

gc, can I ask where you got that quartz tube from? I've been searching google for the last 20 minutes for "quartz furnace tube", "quartz test tube" and the like, yet the closest i could find was a tube with a glass joint on one end and the other end was open. Not to mention it was only half of the size of yours :(

garage chemist - 24-3-2008 at 14:07

I got this quartz test tube from german ebay. You have to search for the terms quarzglas reagenzglas on ebay to find the specific model I use- with joint 14/23, the one for EUR 18.50.

The_Davster - 24-3-2008 at 15:37

Most universities also have glassblowing shops, and for money may make you things. I have seen some that sell glass flowers and glass swans and the like on the side.

S.C. Wack - 24-3-2008 at 17:28

Originally posted by DJF90
I've been searching google for the last 20 minutes for "quartz furnace tube", "quartz test tube"

First hit on google (20 seconds of searching) - 2.5x20 cm. test tube for $30. Maybe it's a location thing...Not so sure about 600C for Pyrex softening...another google challenge
...maybe it weakens some to weight before that.

As said before in other threads, the metalcasters on the net are quite familiar with builing gas-fired furnaces. The refractory on my older Lindberg is 2 layers of different kinds of material, held together with some grey mortar. The inner bricks look for all the world like ground-up perlite held together with some sort of binder, like typical firebrick from any local hobby pottery or welding shop supplier. But the outer layer planks look exactly like ground-up vermiculite. So I wonder if there are different ways to skin that cat, in making your own refractory brick instead of cement (maybe I should google?). Seem to be drifting off topic here.

microcosmicus - 24-3-2008 at 18:07

Yes, perlite or vermiculite held together with furnace cement
makes for a good high-temperature insulating material. For
a formula and further information see Wasser's page:

A coffee-can furnace made of this stuff powered by a blowtorch
should be useful for heating retorts to generate SO3, P, etc.
Personally, I prefer to use this stuff as insulation around an
inner shell made of a harder refractory material.

IIRC, vermiculite, it is not so good for high temperature
applications as perlite because of steam issues.

The annealing point of Borosilicate glass is 600C, and its
strain point is 510C. Already at the lower temperature.
the glass has become soft enough that it will sag and
bend noticeably under pressure. By contrast, the strain
point of Vycor is at 890C and that of pure silica is 950C.

As for metalcasters, both 12AX7 and I happen to work
with molten metal. Maybe I should make a publication
here about furnace building one of these months.

Fleaker - 24-3-2008 at 18:34

@microcosmicus--I also do casting and metalwork. If you've not already joined, please join us at and There is ample material on that site on constructing many furnaces in far grander design than we've attempted here. I think it is a good idea to put up a post on furnace construction and proper materials.

I am thinking more and more that this reaction could be conducted in an improvised setup--a reaction vessel buried in charcoal, with a louver to control the amount of oxygen (thus preventing overheating). Charcoal without a ready supply of oxygen will barely melt aluminum, which melts about 650 C, so it should work well

MagicJigPipe - 24-3-2008 at 19:47

"Little kilns like this go for resonable money on ebay - usually being sold off by rich, middle aged women who have dabbled in ceramics, got bored, and are selling off all the expensive gear they've bought"

That's funny you should say that as my Mom is a semi-rich, middle aged woman that dabbles in glass-melting (to make jewelry). She has a small kiln. One of these days, she will quit using it and I shall harvest the nichrome out of it.

Anyway, what I wanted to ask is I have a piece of equipment that I bought for $25 from Ebay a while back. It is some sort of glass tube/probe (annealing?) oven. It works sort of like a toaster. The tube goes in and when the timer is done it pops it out. The max temp appears to be 750C. I wonder if I could modify this to make some sort of tube furnace? I know you need more information... I shall post a picture soon.

PS... Is this question in the wrong place? If so, I apologize and I will not object if it is moved.

len1 - 3-4-2008 at 17:20

It has occured to me that H2SO4 mist is a known carcinogen on a par with benzene. The SO3 which is hard to condense must be the worst fom of this H2SO4 mist as the particles are tiny. Can anyone confirm.

Strong inorganic acid mists containing sulfuric acid are known to be
human carcinogens based on sufficient evidence of carcinogenicity from
studies in humans that indicate a causal relationship between exposure
to strong inorganic acid mists containing sulfuric acid and human
cancer. Occupational exposures to strong inorganic acid mists
containing sulfuric acid are specifically associated with laryngeal and
lung cancer in humans.

Im beginning to lose some of my interest in this hobby, its seems I cant step left right or centre without fear of a carcinogen, a precursor, or a poison. In physics the scares were much fewer and far between.

PS I hope people dont think health questions are unrelated to the relevant chemistry. On a hobby site I think its paramount (and in professional literature its concetrated on far too little in my opinion).

[Edited on 4-4-2008 by len1]

microcosmicus - 3-4-2008 at 18:35


I'm beginning to lose some of my interest in this hobby,
its seems I cant step left right or centre without fear of a
carcinogen, a precursor, or a poison. In physics the scares
were much fewer and far between.

As someone else approaching a chemistry hobby from a physics
background, this sounds rather odd to me. I mean, in chemistry,
one is only fiddling around with the outer electrons of atoms at
relatively low energies and conditions not that far removed from
everyday life, not trying to rearrange the constituents of nuclei,
recreate the conditions of the early universe, or deal in realms
where matter can become energy, and the very structure of
spacetime can be tied in knots.

Some physics scares:

Almost any particle beam or radioisotope is a carcinogen.

Potentials in the millions of volts.

Intense RF fields which can fry anyone in the way.

Runaway nuclear chain reactions which can destroy a city and
leave it radioactive for years..

Temperatures so high that the thermal radiation is in the form of x-rays

Compared to this, I consider chemistry a pleasant stroll through the park.
For me, what makes this hobby so attractive is the opportunity to
play with the strange microscopic world of quantum mechanics in a
relatively tame setting. For instance, I can observe a singlet-triplet state
transition in oxygen which gives off a visible photon. By contrast, if I was
looking at an analogous transition in a nucleus, that would involve
an x-ray and its analogue in particle physics would involve a
gamma ray.

garage chemist - 3-4-2008 at 20:27

You would encounter the same type of fume if you just concentrate sulfuric acid by boiling, unless you use a distillation setup (which I always use to concentrate and denitrate waste nitrating acid and other dilute waste acids from processes where only the hygroscopic property of H2SO4 is needed. The recycled conc. H2SO4 finds use in washing bottles.).
I think a significant number of members here have at some time made conc. H2SO4 from battery acid.

Such carcinogens are only a big deal if you work with them pretty much every day for years of your life, as in an industrial setup.
If you are going to make some SO3 in your lab for occasional use the risk is negligible.
This stuff has to be made and handled under the fume hood anyway as you will quickly find out the first time you do it.

Although the smoke itself is pretty tame and only very slightly irritating, as chemoleo has pointed out in the old oleum and SO3 thread, you just don't want to fill your lab with it.

not_important - 4-4-2008 at 06:37

And if you are really worried about SO3 fumes, add a large bottle with some 5% aqueous ammonia in it to the end of the wash train, then vent that through a long tube through a plain water wash. The large bottle and long tube are to give plenty of opportunity for the SO3 to react with NH3(g) and H2O(g) to make ammonium sulfate.

Pixicious - 9-4-2008 at 04:21

As someone else approaching a chemistry hobby from a physics
background, this sounds rather odd to me. I mean, in chemistry, ...

Well said microcosmicus

A few years ago (Has this been achieved yet) they were attempting to make a few black holes in a giant machine..

Back to Chemistry..

[Edited on 9-4-2008 by Pixicious]

Polverone - 18-4-2008 at 11:06

EDIT: After doing some searching, I realize that BromicAcid and S.C. Wack have previously discussed the Wolters magnesium modification described below in the Oleum and SO3 thread. This may not be as interesting as I thought.

According to George Lunge's works on sulfuric acid (The Manufacture of Sulphuric Acid and Alkali, with the Collateral Branches: A Theoretical and Practical Treatise, various editions) there is a patented modification that permits lower working temperatures for the pyrosulphate decomposition. Wolters (given as Walters in some other books), in German patent 3110, March 5, 1878, describes using a 1:1 equivalent mixture of magnesium and sodium sulphates, converted to acid sulphate and subsequently pyrosulphate, ultimately liberating SO3.

Lunge says "Experiments made in my laboratory with Wolters's process have to a great extent confirmed these statements [given in the patent]." So it does not appear to belong to the class of misleading and/or irreproducible patents which are so easy to find.

The Principles of Chemistry By Dmitry Ivanovich Mendeleyev, A. J. Greenaway, George Kamensky, Thomas Atkinson Lawson also suggests the use of the magnesium/sodium salt mixture, though it is unclear whether the authors tested the procedure also or if they are just parroting the patent literature or Lunge.

Manual of Chemical Technology By Johannes Rudolf Wagner, William Crookes, Ferdinand Fischer makes reference to the mixed salt approach again and says that the decomposition takes place at 600 degrees Celsius. It is unclear if this is the start of decomposition or the peak temperature for the process, and it is again unclear if the process description comes from experience or from other literature.

If the working temperature is 600 degrees, or not much higher, it raises the interesting possibility of using much cheaper borosilicate apparatus. This is far above the suggested working temperature limits of borosilicate, but considerably below its softening point. I would not hesitate to attempt a 600 degree reaction in borosilicate, though I would not expect to reuse the apparatus. Given the relative cost and availability of borosilicate vs. vycor or quartz, it may not be a bad trade-off to use borosilicate vessels in a disposable manner. Of course, as len1 has shown, sometimes the right vinegar dispenser is all you need.

Even if it does not permit the use of cheaper vessels, the mixed magnesium/sodium process may be of interest simply for consuming less power, and producing less wear on the apparatus, due to lower temperatures.

All the books I've mentioned can be found on Google Books if you wish to read more.

[Edited on 4-18-2008 by Polverone]

len1 - 18-4-2008 at 15:56

I think the NaHSO4 decomposition can be done in borosilicate right now. I have done the pyrosulphate decomposition three times in the box oven - which has the great advantage over the tube oven of a much more uniform temperature (therefore max T ~ average T) each time the first SO3 appears at 650C peaks at 720C, and the last remnants come over at 780C.

I have deliberately placed some borosilicate glassware not dear to my heart (ie previously chipped or broken - like not _important I throw nothing away) at various places in the oven. There is no change in its appearance, except for a hint of fogging in some places. So if you dont mind a flask and a bend possibly freezing and losing their shine, its all go. Its even better than you say - they can be reused for this processes - though probably better keep them out of other reactions.

MgSO4 is more basic than Na2SO4 so im surprised at its purported effect (50-100 degree lowering of temperature). It might be worth a try as an experiment - though the hassle of having to make the Mg(HSO4)2 might not be worth the slight lowering in temperature for a preparative technique. Len

[Edited on 19-4-2008 by len1]

Magpie - 8-1-2011 at 19:52

@garage chemist

I have assembled an apparatus for making SO3 from NaHSO4 per your method. This quartz glass assembly has ground glass connectors. The one clip used is ptfe. The ID of the final tube is 8mm. The tube in the furnace is 400x22x25mm.

Let me know if you have comments or suggestions. Thanks.

SO3 apparatus2.JPG - 39kB

[Edited on 9-1-2011 by Magpie]

garage chemist - 9-1-2011 at 02:56

This looks good and should work without problems.
The PTFE clip is probably OK to use, but the temperature of the evolving H2SO4 vapor may be a bit too high for it. If you can replace it by steel wire it would be more reliable.
Do you plan to leave the joints completely unlubricated (as I have done in the past, but would not recommend) or are there PTFE liners in there? PTFE can't be used at the connection from retort to extension, it's way too hot.
I like to dust the quartz joints with graphite powder before assembly- this doesn't provide any real lubrication, but has always prevented them from freezing.

metalresearcher - 9-1-2011 at 08:19

Interesting. I have an electric (Kanthal) furnace which can heat till 1200oC.
Where did you buy the quartz glassware ?
I found this : Click ! but these are brutally expensive such as $30 for a simple test tube !

[Edited on 2011-1-9 by metalresearcher]

Rosco Bodine - 9-1-2011 at 09:04

There was an interesting bit of related topic discussion between 497 and myself
in another thread where the possible use of a basic ferric sulfate was being described. The following post and page describes the Monsel's salt which may fit the bill as another useful precursor for pyrolytic decomposition yielding SO3.
Related discussion began on the page preceding, so there are two pages there about this possible method which may be worth an experiment if anyone is curious to see if the basic ferric sulfate Monsel salt would actually work as contemplated in hope of producing a higher yield of SO3.

Magpie - 9-1-2011 at 09:34

@garage chemist

My plan was to leave the joints unlubricated. I have no ptfe liners. I do just happen to have acquired some graphite powder, however, so will use that as you suggest.


Quartz glassware can be expensive. Custom glassware (quartz or boro) can be even more expensive. I prepared drawings of what I wanted then sent them out for bid to 3 suppliers. The successful bidder was half the cost of the high bidder.

Bikemaster - 10-1-2011 at 08:03

@ Magpie
Have you ever try galiper slider grease(car brake) for high temperature joint sealer?? They says that the grease can withstand 850 C. I have no idea of the reactivity with sulfuric acid or sulfur trioxide, but i don't this that it is compose of any organic compound.

What happen with the ceramic test tubes??

Magpie - 10-1-2011 at 08:19

Quote: Originally posted by Bikemaster  
@ Magpie
Have you ever try galiper slider grease(car brake) for high temperature joint sealer?? They says that the grease can withstand 850 C. I have no idea of the reactivity with sulfuric acid or sulfur trioxide, but i don't this that it is compose of any organic compound.

What happen with the ceramic test tubes??

Hello Bikemaster. No, I don't know anything about that brake grease. Maybe it's MoS2? I'm going to look locally for some 325mesh graphite today as the graphite I have is only 80 mesh - way too coarse. And you can't grind it in a mortar. Maybe a ball mill would work, but my ball mill is not clean enough.

The ceramic tubes are resting comfortably in a storage closet. They are just waiting for me to resume my phosphorus experiments. ;)

hissingnoise - 10-1-2011 at 08:30

Simply scribbling graphite from a soft (8B) pencil onto joints might suffice.

metalresearcher - 10-1-2011 at 09:54

What about the 'Haldenwanger' porcelain crucibles (with a lid glued with the abovementioned kit or 'stove glue') to use as a retort ? In an electric (Kanthal) furnace of course.
These crucibles withstand 1050oC.

[Edited on 2011-1-10 by metalresearcher]

The WiZard is In - 10-1-2011 at 10:00

Quote: Originally posted by Magpie  
@garage chemist

My plan was to leave the joints unlubricated. I have no ptfe liners. I do just happen to have acquired some graphite powder, however, so will use that as you suggest.

How 'bout "Motor Mica" extremely fine ground mica commonly used
as a dry lubricant. I use it to lube cartridge cases when reloading

Humm Motor Mica seems to have disappeared from the market. Have found this —

Sedit - 10-1-2011 at 18:55

Woelen the fire cement more then likely would prove problematic since most of the same stuff I have is Sodium Silicates which on firing converts to Silica Dioxide meaning you will more then likely have issues with it holding form at first on the initial firing. Making a form may be helpful perhaps coating a tin can and firing then firing it slowly till it converts then touching up cracks and what not until you achive an air tight retort.

I personally would go with ceramics before attempting to use the fire cement and just using the cement to seal up the openings providing you with a removable seal(it cracks easy).

I have been doing pottery for years and have considered making ceramic retorts to sell but I don't know if I would be wasting my time and if they would sell or not.

Sedit - 11-1-2011 at 18:30

For the time being I will have to retract my statement on the Sodium Silicate forming SiO2 on heating because I can not find a reference for it and other sites seem to suggest otherwise. This was the indication I got from studing it a while back but I am unsure if this is correct.

However its stated HERE that at 210–220 °F the sodium silicate loses water molecules to form a very powerful sealant that will not re-melt below 1500 °F which is around 800 degress Celcius.

Magpie - 12-1-2011 at 10:49

re: my post of 8/1/11, this thread

Yesterday I made my first attempt at making SO3. Everything was going well until the furnace reached a temperature of ~860C. At this time the 19/38 joint (the lower one) began leaking. As it leaked considerable SO3 smoke was released. Since I waited too long before shutting down there was enough smoke to challenge my efficient fume hood to the max. When the system stopped smoking I removed the quartz tube from the furnace with kaowool padded tongs (thanks, gc) and placed it in a flower vase to finish cooling. This morning I placed this tube in a tray of water, where the Na2SO4 is quickly dissolving.

I pushed that lower joint together manually during operation but this did not stop the leakage. So it is not a problem with the ptfe clip, which performed admirably, suffering no damage. I believe the problem is due to the upward orientation of the male connector, which allows liquid SO3 to form a small pool then penetrate down through the joint. The orientation should be just opposite, with the male piece on top. This mis-orientation is an artifact of a previous design intent which was discarded.

The 24/40 connector, having proper orientation for refluxing liquid, did not leak.

So, it's back to dye making for me until I get this problem resolved.

[Edited on 12-1-2011 by Magpie]

[Edited on 12-1-2011 by Magpie]

[Edited on 12-1-2011 by Magpie]

garage chemist - 12-1-2011 at 13:49

I'm sorry to hear that it did not work well.
I always have a bit of leakage at the joint as well and it fumes there, but the losses are negligible.
Can you try wetting the joint with some conc. H2SO4 to help prevent SO3 from creeping through it?
Alternatively, a joint lube that's compatible with SO3 is syrupy phosphoric acid. It is convenient to let some P2O5 attract moisture until it has a suitable consistency. If you don't have that you could boil down some 85% phosphoric acid until it is viscous enough when cool.
The problem is that very hot H3PO4 will attack quartz (not very much, but in a joint, any attack is unacceptable). You will have to keep this joint from getting too hot with e.g. a computer cooling fan next to it. Also, you could do some preliminary experiments on the temperature at which H3PO4 starts to attack quartz.

Magpie - 12-1-2011 at 14:03

Yes, I was willing to tolerate a small leakage as long as my hood fan was handling it. But I waited way too long and I had a large leak. I strongly recommend not letting that happen!

I have my glassware all washed up now. Last night the offending joint seemed locked as I couldn't open it. This morning it came apart easily.

Thanks for the suggestions. I may try the con H2SO4 as a joint sealant. I would hate to have to go back to my glassblower with a new design.

Magpie - 19-1-2011 at 16:57

re: my 8/1/11 post

This is a progress report on my quest for SO3. After thinking about my recent failure for a few days I decided to give it another try. This time I prepared Na2S2O7 first and used it as precursor. I did this for the same reason Len did: I didn't want to have to manipulate my quartz extension tube during operation at 460C.

To prepare the Na2S2O7 I first dried NaHSO4 at 150C in a drying oven for 1 hour to drive off unbound moisture. I then loaded a 25x200mm borosilicate test tube to the 125mm mark with 71.6g of the dried NaHSO4, placed it in my tube furnace, and slowly (over 4 hours) heated it up to 460C. It was clear that H2SO4 and likely SO3 was being given off during this heating. In fact there was so much smoke coming off that I shut off the furnace promptly upon reaching 460C. The cooled salt was very dense having shrunk in length by half with a final weight of 66.8g. Coincidentally or otherwise this weight corresponded almost exactly to the theoretical weight of Na2S2O7. I had to break the test tube to recover the salt (I won't use that method again), crushed it with a hammer, then ground it to a powder in a mortar.

Today this Na2S2O7 was loaded into the quartz setup as shown in the picture of my 8/1/11 post above. 98% H2SO4 (reagent grade) was used to seal the two connectors. A slow heat-up was then begun in anticipation of making SO3. Everything was progressing well until the temperature reached about 600C. At this time a drop of liquid would periodically seep out of the lower joint. It did not fume so I assume it was H2SO4. The the upper joint started leaking smoke, presumably SO3. So I shutdown the furnace.

Other than sealing the joints, here are improvements I would make for future runs:

1. Insulate the quartz tube up to the bend. This will prevent refluxing liquid from dripping back into the furnace tube.

2. Cut off my extension tube 25-50mm so that it is never submerged in accumulating liquid in the receiver. This will eliminate back-pressure and eliminate any chance of suckback, a potentially disasterous event. In lieu of this the furnace could be raised 25-50mm with blocks.

To solve the leaking connectors issue I am first planning to go back to fine graphite for the upper connector. For the lower connector I am planning on fusing this to a solid connection with sodium silicate solution. I will be able to fire this to nearly 1300C in the tube furnace which I hope will provide a leakproof seal. If it doesn't I will be forced to go back to my glassblower with a corrected design.

Any questions, comments and/or suggestions are welcomed.

[Edited on 20-1-2011 by Magpie]

[Edited on 20-1-2011 by Magpie]

Magpie - 28-1-2011 at 11:23

This is another update on my adventures with SO3. I decided I needed to step back and get to know this treacherous creature a little better before further attempts at capture. I did some small crucible testing with a Mekker burner on MAPP gas where I was just able to drive off all the SO3 at red heat. I also attempted to drive SO3 out of some hydrated Al2(SO4)3 with absolutely no SO3 evolution. I experimented with making Na2S2O7 using a sand bath (my first one). But the product was so hard I wasn't sure I could get it out of the ceramic dish.

Then I gave the NaHSO4 another try in the quartz tube set-up, with the lower joint sealed with RTV silicone and the upper joint as bare glass. The upper joint leaked profusely so this run was aborted at about 570C.

On the next try (#3) I rubbed the upper glass joint, inner and outer surfaces, to saturation with extrafine dry graphite, which is available at some hardware stores. This proved, after a little initial leaking, to provide an excellent seal. So I was on my way to making SO3 as shown in the first picture. This arrangement was assembled with the intended use of the SO3 in mind, ie, gassing it with HCl to make chlorosulphonic acid. The condenser was a contingency that proved to not be of much use. At T=572C I sucked out the few mls of water in the receiver using a Pasteur pipet. At T=652 smoke was showing in the angle bend. At 682C no more smoke, just a very thin stream of clear liquid slowly running into the receiver, ie, liquid SO3. As heating continued more SO3 was accumulating as white crystals in the receiver. Then the RTV began to deteriorate, which discolored the liquid SO3 to a brown color. Also I noted that the end of the 8mm ID tube was plugging with solidified SO3. Although I could not reach this location with the heat gun I warmed the falling liquid in the tube just above the flask with the gun. This was successful in clearing out the plug. I continued the heat up to 900C. At this point there seemed to be very little product forming so I turned off the furnace.

I consider this synthesis mostly a success with the brown contamination of the SO3 the only detraction. It is obvious that good sealing of any joints is essential. I will next try removing the RTV and replacing it with a permanent seal using sodium silicate.

My charge of NaHSO4 was 100g. I'm estimating by eye that there may be about 10g of SO3 in the 250ml RBF receiver. I will be able to report this accurately only at a later date as I neglected to tare the receiver.

SO3 success.JPG - 46kBSO3 contaminated.JPG - 50kB

garage chemist - 28-1-2011 at 13:58

I would recommend against permanently sealing the joints with sodium silicate. The silicate has a much greater coefficient of expansion than the quartz glass, and such a sealed joint would very likely crack upon heating or cooling.
Try the polyphosphoric acid before resorting to such a desperate measure- or just order a 90° bent 8mm quartz glass tube with one male joint, as I have done. That way, there is only one joint in the setup, and it's facing upwards so that no liquid can leak through it.
Discoloration of the crude SO3 is not an indicator of low quality. I often had that happen due to residual vaseline on the joints of my glassware- it turns into black goop that runs down into the liquid SO3 and makes it deep brown or black. A simple redistillation renders it clear again.

Magpie - 28-1-2011 at 14:36

@garage chemist -

Do you think I could gas the discolored SO3 first with HCl? This would eliminate vessel transfers and reduce the number of distillations to one.

[Edited on 28-1-2011 by Magpie]

garage chemist - 28-1-2011 at 14:56

Yes, of course you can do that, chlorosulfonic acid is isolated by quick atmospheric distillation.
But I would recommend to make more SO3 first- pool a few batches, and convert them to ClSO3H together. It does not require more lab time to make 200g chlorosulfonic acid than it does to make 20g.

Magpie - 3-2-2011 at 14:36

Quote: Originally posted by Magpie  

My charge of NaHSO4 was 100g. I'm estimating by eye that there may be about 10g of SO3 in the 250ml RBF receiver. I will be able to report this accurately only at a later date as I neglected to tare the receiver.

I now have the weight of oleum from my 18-1-11 preparation from NaHSO4. I titrated 0.5mL with 1N NaOH to a phenolphthalein endpoint. This yielded a value of 81% oleum. The total weight was 25.9g. The pool grade NaHSO4 was 95% strength so the yield was 78.9% of theory.

My quartz glassware is being modified to eliminate the lower joint. When complete I will be back in production of this useful reagent. ;)

Magpie - 17-3-2011 at 13:42

My quartz glassware modification was finally completed so yesterday I made another SO3 run using NaHSO4 in my tube furnace. This time the quartz tubing performed flawlessly and I had no leakage. Net weight of oleum is 29.2g based on a charge of 100g of 95% pool grade NaHSO4.

As you can see in the picture below the oleum is a mixture of liquid and solids. This presents a problem for assay and usage. I withdrew 0.5mL of the liquid and titrated it. It is 50.5% oleum. But now I don't know the assay of the solids, and I don't know what % of the whole is solids. So I can't determine a yield.

More importantly, for my intended usage, I need to prepare 20% and 66% oleum. I could melt the solids by heating to 36C, but I estimate the vapor pressure would rise to 375mmHg (for 100% oleum) at that temperature. Anyone have any suggestions as to how to solve this problem?

modified glassware.JPG - 98kB oleum, liquid & solids.JPG - 88kB

[Edited on 17-3-2011 by Magpie]

garage chemist - 17-3-2011 at 14:31

From oleum with a SO3 concentration between 20 and 65%, solid disulfuric acid H2S2O7 crystallizes out.
This problem is solved by distilling off the SO3 (H2S2O7 also loses its SO3 then) and adding the calculated amount of conc. H2SO4 to form the desired oleum strength.

Magpie - 17-3-2011 at 14:54

Thank you GC. I'll post the results.

Magpie - 18-3-2011 at 15:13

I tried to distill my oleum today to obtain pure SO3 as distillate. I had to abort, however, as the amount of smoke coming out the tubulation on the vacuum adapter was threatening to overcome the capability of my fume hood. I had the heat set low (45%) on the heating mantle and had not as yet seen any distillate come over. Cooling water was cold tap water - no ice added.

The solids in the pot did melt so I took another 0.5mL sample for titration. The results indicated 13% oleum.

I feel that on future SO3 production runs with the furnace I should make sure that all water/H2SO4 has been withdrawn from the receiver before the SO3 starts coming over in earnest. I did this at T=660C using a pasteur pipet. Apparently that was not good enough.

Any other suggestions?

Rosco Bodine - 18-3-2011 at 15:46

Spin the vapor using a long graham condenser to see if that centrifuges and coalesces out the fine droplets you can maintain the condenser above the m.p. of the SO3 so it doesn't plug.

Magpie - 19-3-2011 at 19:24

Quote: Originally posted by Rosco Bodine  
Spin the vapor using a long graham condenser to see if that centrifuges and coalesces out the fine droplets you can maintain the condenser above the m.p. of the SO3 so it doesn't plug.

Rosco that is an interesting thought. I do have a Graham condenser but it is short and doesn't have ground glass connections.

Some lessons learned

Up until now I have been assaying my oleum by titration. I now feel that this really is not reliable for two reasons: 1) determining the amount of oleum added to the titration flask using a 1 mL pipet is not accurate due to the high viscosity of the oleum and the lack of accurate sp gr information, and 2) the oleum fumes badly when placed in the titration flask containing water. Any slight variance in the volume delivered or the sp gr used greatly affects the calculation of the assay.

So, today, taking a lesson from Len, I did an assay by evaporation of a weighed amount of oleum, ie, 1.58g. Firstly, I would like to say that it took forever for the SO3 to completely evaporate and I had to heat the 25mL RBF to get it to proceed to completion. All during this evaporation copious fumes of dense white smoke are coming off, at times challenging the fume hood. The result of this assay was 60.7% oelum. Much different than the 13% I reported yesterday using the titration method.

Using 60.7% oleum, the yield of SO3 for the furnace run is 85.6% based on the NaHSO4 charged and 29.2g of product.

SO3 and oleum are most interesting and useful. Handling the oleum is tricky and dangerous, and I'm still learning. My efforts so far would not have been possible without an efficient fume hood.


Here's a short video showing my dilution of 0.5mL of the 60.7% oleum for titration:

[Edited on 20-3-2011 by Magpie]

[Edited on 21-3-2011 by Magpie]

Rosco Bodine - 20-3-2011 at 08:15

An electrostatic precipitation chamber might do the job of marshaling SO3 vapor into a coalesced form. Something custom built for the purpose would probably be required. Graphite might hold up for electrodes .....I'm not sure.

Jon_Swars - 25-5-2011 at 06:33


You obviously havent a clue. Reading the board in general you appear to persist in
evacuating the contents of your bowels all over same.


hissingnoise - 25-5-2011 at 10:08

Reading the board in general you appear to persist in
evacuating the contents of your bowels all over same.

That's nice shite, spewing from a newb with three pretty useless fucking posts already.

Lambda-Eyde - 25-5-2011 at 10:29

Magpie: Then why don't you weigh it? Put an arbitrary amount of sulfur trioxide/oleum into a tared round bottom flask, weigh, dilute, titrate?

Magpie - 25-5-2011 at 11:50

Quote: Originally posted by Lambda-Eyde  
Magpie: Then why don't you weigh it? Put an arbitrary amount of sulfur trioxide/oleum into a tared round bottom flask, weigh, dilute, titrate?

That's a good approach and would probably give the most reliable results. Right now I don't have a scale that can measure an RBF + stopper + oleum then get the oleum weight (by difference) with adequate accuracy.