Unlike SO2, SO3 possesses no warning smell, however its vapours are extremely corrosive and carcinogenic. They are fortunately visible - and this should be used as a cue to avoid exposure.

Theory

The SO3 / H2O system

One way to regard the sulphuric acid system is as a mixture of the components H2O and SO3

a H2SO4 + (b - a) H2O == a SO3 + b H2O.

Note that this is not a reaction equation. The precise structure of such a system, in most situations is much more complex than either left or right, rather its two alternative useful descriptions of the system. Mixtures where b >> a are called dilute sulphuric acid, b = a is refered to as 100% H2SO4, while b < a is called oleum, both a fuming liquid and solid depending on its composition. This property of the b < a oleum system is due to a peculiar non-monotonic phase diagram, which shows two maxima. Thus oleum is a solid at room temperature (20C) in the interval 30%-60% SO3, and again at >75% SO3. A consequence of this peculiarity is that one can not determine whether one has >75% purity SO3, or a lower concentration oleum by the appearance of a solid phase alone. A quantitative analysis needs be conducted. Due to the hysteresis in the system a melting point determination is very inaccurate, rather a titration of the homogeneous product, or evaporation of SO3, is required.



Returning to the equation above, b = 0, corresponds to pure SO3, a white deliquescent fuming solid with a melting point of 16C and boiling point of 44C. This very narrow range of temperatures where SO3 is a liquid, is the source of its condensation problems during its production by pyrolysis. SO3 easily polymerises (enhanced by H2O, inhibited by P2O5), with three modifications, with melting points 32C, 62C, and 95C, respectively.

If one tries distilling sulphuric acid with b >> a, initially a dilute acid distils, with the fraction of the SO3 progressively increasing until the maximum boiling point at 337C of the liquid/vapour azeotrope of 98%/2% H2SO4/H2O is reached. After this no further concentration in the SO3 component can be obtained. Hence distilation can not be used to prepare oleum/SO3, it is also quite wasteful for preparing conc. H2SO4 from the weak acid.

The SO3 / H2O / Na2SO4 system

A way to obtain oleum is provided by the acid salt NaHSO4. One way to regard the anhydrous salt is as an equimolar mixture of 100% H2SO4 and the basic salt Na2SO4 (it is basic being the salt of the weak acid NaHSO4). Certainly an aqueous solution of NaHSO4 is perfectly equivalent to such a mixture

2NaHSO4 == H2SO4 + Na2SO4

Again this is not a reaction, but an equivalence in the sense of the first equation. Since NaHSO4 has a relatively high pH (1mol/L has pH~2) we see from the equivalence above that Na2SO4 has a marked effect on the pH of H2SO4 solutions with an equimolar mixture raising the pH of 1mol/L H2SO4 from pH << 0 to pH ~ 2 (buffering). Na2SO4 thus being a weakly basic salt explains why the pyrosulphate can be decomposed at fairly low temperatures in what follows (compare to CaSO4 -> CaO + SO3 at 950C < T < 1200C, CaO being a much stronger base).

The formula above, which was presented as being merely an equivalent representation, turns out to be a realisable reaction in practice. Indeed H2SO4 can be distilled from the acid salt NaHSO4. Here we have found the distillation is 97% efficient. Collecting the distillate as a single fraction gives 100% H2SO4 in 97% yield. However, unlike the SO3/H2O system, in the Na2SO4/SO3 system, introducing a cut in the distillate fraction will yield us oleum, with the lower b.p. fraction being therefore a correspondingly weaker acid (since a combination of the fractions must give 100% H2SO4). We can say that the addition of Na2SO4 to H2SO4 breaks the azeotrope.

This distillation is generally 'idealised' as two reactions the first being evaporation of H2O

2NaHSO4 -> Na2S2O7 + H2O 200C < T < 650C

the second being evaporation of SO3 leaving behind a weakly basic salt

Na2S2O7 -> Na2SO4 + SO3 260C < T < 780C

The temperatures quoted are the ones I have measured here rather than those in the literature, which suggest almost no overlap of the reactions. In actual fact both reactions occur at the same time over a wide temperature range and the system consists of a complex mixture of NaHSO3/Na2S2O7/Na2SO4 at all but the bottom and top temperature ranges. Nevertheless substantial H2O evolution occurs at 200-260C while almost no SO3 is being evolved, and substantial SO3 doesnt get released until 780C, when no water is essentially left behind, and this explains the separating power of the NaHSO4 distillation.

Nonetheless substantial SO3 is released below the 'nominal' decomposition temperature of the pyrosulphate at 450C, about 30% in the present experiment, which indeed is the major component of the yield loss in the 69% SO3 yield quoted here. In addition some H2O is released with the SO3 at the low-end of the SO3 evolution range of 650C. This demonstrates that the pyrosulphate is a convenient idealisation for what in reality are complex mixtures of Na2SO4/H2O/SO3 existing at the various temperatures.

The efficiency with which the various components H2O/H2SO4/oleum/SO3 are separated is dependent on the homogeneity of the NaHSO4 composition (distilation of 100% H2SO4 added equimolarly to Na2SO4 without intimate mixing will just return the 100% acid) and the time - temperature regime.

Definition of yield and strength of oleum

One has to define carefully what one means by yield in such a system. The simplest definition is percentage yield H2SO4 based on sulphate in the NaHSO4 - with 100% yield corresponding to distillation of 50% of the total sulphate - with Na2SO4 forming the undistilled remainder. However generally one is interested in obtaining oleum. The yield for such a procedure, which requires one to take a cut on the distilation can be defined as so much percentage yield for oleum of quoted strength. Obviously the higher the strength of oleum desired, the higher the temperature at which the product starts to be collected, and the lower the yield.

Generally one defines the strength of oleum as a percentage of the weight of free SO3 to total weight of oleum. Free SO3 is defined as the weight of SO3 after the amount of SO3 bound as 100% H2SO4 has been subtracted out. In practice, for an SO3/H2O system with weight composition a SO3 + b H2O this prescription amounts to performing the arithmetic:

[a - b * (80/18)] / (a + b)

100% SO3 (b=0) then corresponds to 100% oleum, while 100% H2SO4 is zero percent oleum - with no free SO3 being able to be distilled off.

Within this formula lies a difficulty in determining the strength of oleum with any precision experimentally (as will be seen below). Because of the small molar mass of H2O its weight is multipluied by a factor of 4.4 in the formula, which, on performing the substraction, has the effect that small errors in H2O weight lead to large errors in oleum strength. Put another way, it only takes 18 grams of water to convert 80 grams of SO3 (100% oleum - a fuming solid evaporating at 44C with all SO3 available), into ordinary sulphuric acid boiling at around 300C (0% oleum with no free SO3 available). Hence glassware in which SO3 is generated and reacted should be oven dried to avoid loss of yield.


In the present article a 460C cut on the distilled fraction gave oleum of strength 80-100% (the large uncertainty for the reason explained above) with 69% yield based on free SO3 determined by titration, while the percentage yield H2SO4, based on total fraction with no cut-off temperature was 97%.

Acid-base sulphate and bisulphate equilibria

pH of HSO4- solutions

The bisulphate anion acts as a weak acid due to the low value of the following equilibrium constant

HSO4- -> SO4-- + H+, [SO4][H]/[HSO4] ~ 10^-2 -> pKa ~ 2

pKa = 2 is sufficienly large compared to pK = 7 in water that we may neglect protons contributed by water. Its also sufficiently small that in solutions of order 1 mol/L loss of HSO4- to dissociation can be negleted. If c is the latter concentration

[H] = sqrt(c 10^-2) -> pH = 1 - 1/2 log c

pH of SO4-- solutions

These solutions are weakly basic, and we have to use the proton equilibrium in water

[H][OH] ~ 10^-14

and the mass balance equation for the protons [H]

[OH] = [H] + [HSO4]

We neglect SO4 dissociation, whose concetration is c, again due to the small pKa above, hence

[H]^2 + [H][HSO4] = [H]^2 + [H]^2 c / 10^-2 = 10^-14

It follows

pH = 8 + 0.5 log c

A 1 mol/L Na2SO4 solution is expected to have pH ~ 8.

Method

All operations described below were carried out in a fume hood, wearing a gas mask, latex gloves, and with a drought established in the laboratory.

Ambient temperature in the lab 25C.

Although the NaHSO4 distillation can be carried out in a single step, taking a cut at the desired temperature, I did not wish to expose myself to excess SO3 fumes such as entailed in making connections when replacing receiver and condenser. Hence the process was carried out in two steps with the lower temperature part taking place outside in a shed.

There is a precedent for this in a procedure described in the Practicum on Inorganic Chemistry, Editor Yu. Tretyakov, p. 235

Step 1 - Distillation of low temperature fraction

SO3 released by the build up of pressure inside a warming container and sprayed in the eye will blind you instantly. Please wear glasses and be careful.

329 grams of NaHSO4 (pH-down for pools) were placed in a porcelain crucible and heated on a shielded sand bath. The exhaust fumes were not collected, so the procedure was carried out in a shed. The temperature of the bath was initially held at 250C until the salt melted - this occurred at about 190C as measured at the melt surface by an IR thermometer. When the melt reached 210-220C water vapour evolution commenced - but ceased in about 20mins with the melt at 220C. A picture of the melt at this stage is shown below.



The temperature of the bath was now raised to 350C - when the melt reached about 260C sulphuric acid vapour evolution commenced - as evidenced by a deep white smoke. Vapour evolution continued with the bath temperature rising to 285C in about an 90mins, the temperature being limited by the heat carried away by the evaporating H2SO4. At this time the bath temperature was set to 450C, whereupon the surface temperature rose to 337C where is stayed relatively constant for about an hour - due presumably to evaporation of the H2SO4 azeotrope. Distillation was stopped when the melt surface temperature reached 425C (about 450C in bulk), at which stage the reagent was fuming and had the form of a viscous liquid impregnated with tiny SO3/H2O bubbles. At no stage, from the nominal melting point of the NaHSO4 to the nominal melting point of the Na2S2O7 did the melt solidify. It is obvious, that quite a substantial amount (as turned out subsequently almost the entire yield loss) of SO3 was evaporated at this stage. It is possible that some of this could be avoided by shielding the melt from the atmosphere - as reaction of SO3 vapour with atmospheric moisture (in addition to that evaporating off the surface of the liquid), SO3 + H2O -> H2SO,4 decreases its partial pressure substantially and enhances evolution of further SO3 by Le Chattliers principle. It is also probable that yield could be enhanced by maintaining the melt in the 240-260C range for longer than the 2hrs here. However evaporation gets progressively slower, and as NaHSO4 is a very cheap reagent, experimental time was considered the more valuble comodity.

The remaining material now weighed 272 grams. The corresponding weight for a pure Na2S2O7 composition is 304 grams. Hence if one assumes the material consists of a Na2S2O7/Na2SO4 mixture (all H2O evaporated) 32 grams SO3, or 29% of total available SO3, co-evaporated with the H2O as H2SO4 in the previous time-temperature regime.

The sulphates shrink more than porcelain on cooling (as opposed to an NaCl/CaCl2 mix which cracks porcelain), and hence are easily separated by inversion and light tapping. The material was then broken up into small lumps and powder in a mortar as seen below.




Step 1 - Pyrolysis and high temperature cut


The material is now placed in our reaction vessel, which is a 50cent vinegar dispenser from Cheap-as-Chips. This container will easily withstand the 800C or so encountered in this experiment, as it was fired in a kiln at almost twice that temperature. It is however meant to operate at temperature gradients of the type encountered in everyday life, and hence will crack if its oulet is at temperatures of 600C or so below oven temperature, such as would occur in the present experiment. Hence, a piece of borosilicate tubing - which has one of the lowest expansion coefficients of any glass and a softening point of 821C - was designated to take this temperature gradient, and was fitted snugly over its outlet, with a glass-wool filler sealed by twisting in between. If one wished to eliminate carbon impurities due to subsequent carbonization of the pigment if the wool is coloured, the pigment should be burned-off at 500C. 97 grams of the mixture from the first step were placed in the reagent vessel which is seen below. Due to its retort shape bumping should be minimised, while tilting gives the melt maximum exposed surface. Despite the simplicity of the arrangement it will be seen subsequently that this step gives almost 100% yield.





Next a 120 degree bend was inserted into the quickfit end of the tubing. Since the temperature of this joint reaches 200C degrees, and I did not wish the joint to freeze - while grease can not be used with SO3 - I inserted some glass wool into the joint. This arrangement worked well - with the exception that the small amount of pigment in the wool was carbonized and added a tint of brown to the first oleum/SO3 distillate. A long tube with quickfit ends, to minimise losses, was used as the main (air condenser) leading to a U-tube maintained in water cooled below 10C, where the product collects. Since H2SO4 evaporation is notorious for generating hard to condense fogs, a reflux condenser was fitted to the other end of the U-tube, and cooled by a small water pump from the cooling container. The exhaust from the reflux condenser was vented through polyethelene tubing - this could be passed outside - but as it turned out, the cooling was so effective that it produced no fumes. Since a box oven rather than a crucible oven is used the oven door was removed and the exit sealed with an asbestos plank and porcelain plate cut to enable the borosilicate tube to pass. The entire arrangement is seen below.





The oven temperature was initially set to 300C and upon equilibration, raised about 200C every hour. Up to about 520C nothing could be seen, except 1-2 H2SO4 drops condensing and dripping down again in the lower half of the 120C bend - this thus formed a kind of fractional distillation apparatus. The same mixture at stage 1 produced a great deal of smoke at 450C, hence this was mostly due to air moisture, which was absent inside the reactor tube. At 550C a white fog filled the air condenser and intensified at 600C. The fog extended to the U-tube, but the reflux condenser remained clear, showing the effectiveness of sulphuric acid condensation. Effectively no product was collected in this temperature range. At 650C the insides of the U-tube frosted over with oleum snow and some product was collected - this temperature was maintained until the air condenser was clear again.



Raising the temperature to 700C resulted in a mass flux of white smoke and liquid streaks down the air condenser, with oleum drops dripping down the 120C bend. The smoke worked its way up the glass wool and carbonized the pigment. The U-tube was now completely coated with solid - the smoke extended to the first few coils of the condenser - but no further. The top of the condenser was clear. When the air condenser was clear again, the temperature was raised and the procedure repeated at 750C and 780C - the latter producing the last large efflux of SO3. After this raising the temperature to 800C and then 820C - considered the limit of this setup - produced no more efflux. Subsequent weighing in the next section shows almost 100% distillation efficiency.

Temperatures measured with an IR thermometer showed a temperature of about 220C on the borosilicate glass tube at a point 2cm from where it exited the oven, and 120C at the junction with the 120 degree bend, at the highest temperature setting of the oven - 820C. The air condenser heated to about 80-90 C when passing material.

The glass tube showed no-annealing indicating the temperature could have been taken higher if needed. Pictures of the oleum drops, streaks, and white smoke are shown below. Note the almost complete carbonization of the first part of the polyethelene tubing on the efflux from the reflux condenser. This was achieved with no visible fumes in the top half of the condenser at all - i.e. at the SO3 residual vapour pressure for 10C. In this way the tubing serves as a very useful indicator of the progress of SO3 vapour out of the system. Note that the radiation emanating from the oven appears violet (which would correspond to about 5000K) this is a frequently seen camera artefact - the actual colour at the maximum temperature was bright orange.





The oven was turned off and the U-tube removed. The condenser - which was clear throught the procedure - started to smoke profusely when its insides were exposed to air (see below).



The U-tube was disconnected and the contents weighted in at 26.9 grams. It was quickly wraped in cellophane and placed in a fridge. If this is not done the stoppers are likely to pop-out due to build up of pressure as the SO3 warms. The fumes this releases are extensive and dangerous. I must admit to being horrified at the danger this material obviously poses.

It follows that SO3 sprayed in the eye will blind instantly. Please wear glasses and be careful

Calculations, yield, and oleum/SO3 stregth

The material left in the retort after distillation weighed 67.2 grams. The amount expected from the reaction

2NaHSO4 -> Na2SO4 + SO3 + H2O

with a fraction of 97/272 of the initial weight of 329 grams of NaHSO3 used in the pyrolysis is 69.4 grams, the distillation efficiency is thus within experimental error 100% (the fact that its slightly over can be explained by the presence of moisture in the 'anhydrous' salt). This supports the earlier finding that at 780C the reaction is essentially complete.

The 67.2 grams Na2SO4 together with the 26.9 grams of the SO3/H2O oleum distilled add up to 94.1 gram - thus giving a 97% yield in the pyrolysis - this is the quoted yield on H2SO4 - essentially no SO3 or H2O was lost outside the system.

The available SO3 content of the initial 97 grams, corresponding to 97 * 329 / 272 = 117 grams starting NaHSO4 is 39.1 grams. If one assumes the entire product gathered is pure SO3, the yield based on available SO3 in the initial NaHSO4 is 26.9 / 39.1 = 69%. However, as pointed out in the theory section, the final product is actually oleum, whose strength is yet to be determined. The water presence in the latter will alter this yield (only slightly because of the low molar mass of water).

The final step is to find the H2O/SO3 ratio in the product and determine oleum strength for our 450C cut. This can be done in two ways. The first is to titrate the oleum. However this is very inconvenient as it exposes one to SO3 fumes. An alternative method is to use the fact that we have almost 100% distillation efficiency. This means the H2O/SO3 ratio of the reagent from step 1 is the same as of our product oleum - and so the reagent can be titrated either with BaCl2 to determine sulphate content, or a base to determine SO3 content in the form of Na2S2O7 and NaHSO4. We use the latter method.

11.2 grams of the product from step 1 were dissolved in about 100ml H2O and titrated against a 2% NaOH solution ([NaOH] known to an accuracy of 3%]. 3.62 grams NaOH raised the pH from an initial pH < 1 to 8.40 at the end point. It is interesting that the pyrosulphate was self-indicating - so a pH meter is not really needed. It changed from colourless to a deep yellow during the addition of the last 1/2 ml, and a small amount (about 0.1 grams) of a yellow-orange precipitate (presumably colloidal sulphur) formed.



This evidenced the presence of 3.62 grams SO3 in the 11.2 grams of step 1 product (coincidently NaOH neutralizes SO3 in a 1:1 weight ratio). The Na2SO4 content in the latter can be determined as follows: since distilling 97 grams of product gives a residual of 67.2 grams, the Na2SO4 content of 11.2 grams of product is ~ 7.7 grams. The H2O content is then

11.2 (+/- 1%) - 3.62 (+/-3%) - 7.7 (+/-2%) = (-0.12 +/- 0.3)grams H2O

The amount of H2O in the 11.2 grams of product then ranges from 0 to 0.28 grams (obviously we have to exclude negative values). Plugging this into the formula for oleum concentration gives oleum > 72%, but thats the lower 70% confidence limit.

Distillation of SO3 from Oleum

As mentioned in the introduction the amount of SO3 in oleum can be determined by evaporation. Since in this case we are directly determining the weight of SO3 which can be distilled from a given strength oleum, rather than the weight of H2O, this method is much more sensitive than the titration in the paragraph above. The distillation is hazardous, due to once again possible exposure to SO3 fumes, and this part must certainly must be conducted in a fume-hood.

As with any distillation a cut must be chosen, and this was taken as 100C for convenience, meaning a water bath can be employed. Obviously the temperature is a compromise between the amount of H2O coming over with the SO3 and the amount of free SO3 left undistilled. The water vapour pressure above 0% oleum (100% H2SO4) at 100C is minimal, while the oleum remaining at that temperature is 37% free SO3, from the following table:

% Oleum b.p.

0% 290C
10% 175C
20% 140C
37% 100C

The optimum cut for many applications is actually between 140-175C being less wasteful, while still keeping the vapour pressure of H2O negligible.

The U-tube filled with oleum from the previous part was connected as shown in the figure below, with the tube being immersed completely in water heated from below. Its initial temperature was about 30C, and a final temperature, established by adding boiling water and heating was 100C.




The oleum melted at about 55C bath temperature - it would indeed have been possible to pour it between the containers at that stage. This shows the SO3 was not in one of the two higher m.p. modifications (59C and 95C), which can be attained by remelting and solidifying, and enhanced by the presence of water. The SO3 started boiling at that temperature - and initially white fumes filled the chilled container subsequently being replaced by a flux of clear drops - this always happens when SO3 passes into a container with a hint of moisture present, once that is combined, pure SO3 is colourless. An air blower was used to warm the 90 degree still bend to ensure no condensation of SO3 there.

After as much SO3 as possible at 100C collected in the chilled flask, the still bend was disconnected and both the U-tube and flask stoppered. See picture below.




The small amount of organics discolouring the SO3, mentioned above, stayed with the 37% oleum in the U-tube. The interchange of stoppers, performed with the oleum at 100C is the only procedure to generate any SO3 smoke in the entire experiment. The bend smoked profusely, so the fume-hood could not cope, the laboratory had to be briefly evacuated.

The weiht gain of the flask was 22.4grams, the weight of 37% oleum was 4.2grams, meaning a loss of only 0.2gram or <1% oleum in the distillation. The 37% oleum contained 1.5gram SO3, with 2.7gram being 100% H2SO4. Together with the 22.4gram pure SO3 this means initial oleum strength prior to the distillation was 90% - in agreement with the titration figure. The distillation of 90% oleum to essentially 100% oleum, or pure SO3, was 94% efficient.

Being concerned to achieve maximum yield for this very simple distillation I waited until the last drops of SO3 had collected in the receiver (the flux for the last few drops was low and took about 5mins). These promptly froze in the narrow tube creating an icicle, this could not be removed by heating the upper part of the bend from the outside, part of it was shaken off when the bend was removed, nonetheless the remainder, while not contributing to the yield, was clearly < 0.2grams. It had to be taken outside (see picture below) as it threatened to fill the whole lab with a mist, and the procedure shows the importance of maintaining a reasonable flux of SO3 during the distillation and not cooling the flask too much.

Health effects

SO3 has no smell but fortunately minute amounts fume profusely in the atmosphere (it is used as a smoke agent in battle), and hence provide easy warning. It is extremely agressive on metals and anything organic as can be seen by the instant carbonization of polyethelene piping by minute amounts of SO3 due to its vapour pressure at T < 10C. It 'develops' as black dots the smallest traces of organics, such as left by the hands, on all glassware, especially particles trapped in quickfit joints even after several washes. It is also carcinogenic - I provide the following quote as a warning

Wow!! Once again len1: Beautiful work with both the experiment and writing!

Wow. Execellent write up and photos. That was awesome (and that PE tube is scary)

Oke a microwave oven is a bad idea.

I placed 0.5 gram of sodiumhydrogensulfate on a watchglas en placed it in a microwave oven at 100% power (800watts)
The timer was set to 30 seconds, nothing happend no melting or decomposition what so ever.
Another 30 seconds was run, after about 15 seconds an explosion occurred.
The watch glass was scattered into hundreds of pieces, the microwave survived.
Also there was a large could of vapour.
The vapour was probably consisting out of water (based on smell).

What I think happened was that there was water trapped in the cristals and in could escape causing super-heating and thus explosive ejection.

Another lesson learned today.
No one was hurt

A lot of substances dehydrate perfectly in a microwave, this one clearly not.
Even under adapted conditions i cant imagine how this could work.
I would have been nice to make pyrosulfate this way.

Quote: Originally posted by User

What I think happened was that there was water trapped in the cristals and in could escape causing super-heating and thus explosive ejection.

No, it was just the solid that melted, that's all. The only problem is in that ionic melts have several magnitudes higher microwave absorption compared to solids. When the MW source is 800W suddenly most of this energy is absorbed by the the small amount of the melt. This essentially means lots of broken glass and the owen full of decomposition products.
It would be nice if you would read the theory about microwaves before engaging in potentially dangerous games with it. The least you could do was to check the melting point of sodium bnsulfate (59°C monohydrate and 315°C anhydrous). You can not use such low melting materials in a MW oven without power control, since even though their MW absorption is poor, sooner or later they will reach the melting point, and at that point they accept too much MW energy in too short a time.
You can try this experiment with some other low melting material that does not decompose as easily, like for example KOH or NaOH. Put a pellet in a sacrificial test tube and put this in a ceramic crucible. Heat it in the MW owen and report back your findings.

thankyou for sharing your experience!

Quote: Originally posted by Hawkguy

Could Ammonium Persulfate be thermally decomposed to yield Sulfur Trioxide? I'm exploring the route with Persulfates as I have no Sodium Bisulfate.

Regeneration of bisulfate

Quote: Originally posted by chloric1

If you clean up and crystallize your sodium sulfate byproduct, would you get sodium bisulfate if you added dilute sulfuric acid?