Removal of water from solvent

[Mattias]

How would you do if you would like to remove absolutely all water from a solvent (like acetonitrile)?

I want to derivitise an alcohol with a flourescent anhydride to a flourescent ester. However, the sample contains a small amount of water (about 2%) which seems to consume my reagent. The reaction takes place in a mixture of acetonitrile and triethylamine.

Could metallic sodium be used for this? My knowledge in organic synthesis is limited I am afraid..

[chromatographer1]

Redistilling is the best way, but 2% water is a lot to remove by any other means.

When you get closer to dryness then use molecular sieve 4A to remove most of the remaining water.

Other means are to distill under certain chemical additives which remove the water first and leave you with dry acetonitrile. It really should only have a few tenths of a percent of water present. Take a look in the library for the best procedure. I don't have one readily available today.

best wishes,

Rod

[Bruce Hamilton]

The easiest solution is to purchase an anhydrous grade, but if you want to prepare and dry any solvent, then go to the library and collect the series of papers by D.R.Burfied published in the 1970s and 1980s on dessicant efficiency in solvent drying.

For example, he showed that using 3A molecular sieves is far more effective than 4A in acetonitrile - after 24 hours at 5% w/v loading, 3A MS had dried to 49 ppm, 4A to 450 ppm and, after 168 hours, 3A was 27 ppm, and 4A at 500 ppm, from an initial moisture content of 2800 ppm. From "Molecular Sieves - Dessicants of choice ", D.R.Burdfield and friends J. appl. Chem. Biotechnol, 1978, v28, p23-30

In general, for drying of solvents, 3A are usually preferred over 4A.

Bruce Hamilton

[chromatographer1]

Thanks for your valuable reference, Bruce.

best wishes,

Rod

[Victor]

It seems from Mattias' question that the source of the water is mainly from his sample rather than from the solvent acetonitrile.

But in the past I have been interested in removing water from acetonitrile to then use it as a mobile phase in normal phase chromatography. In this I considered molecular sieve but was concerned about how "dirty" it might be and how my nice HPLC grade ACN would be contaminated. So instead I used anhydrous sodium sulfate because this is available in high purity grade.

Bruce-thanks for your information. For those of us who do not have this reference to hand, is their any comment on the efficacy of sodium sulfate?

Bruce- I echo your point about dry solvents. I was surprised to see that my HPLC grade ACN quotes a water level of only 70 ppm anyway....another problem is that if you filter your sodium sulfate or molecular sieve dried ACN, I guess this is an excellent way to reintroduce atmospheric moisture. Or do you just hope that the stones on the mobile phase inlets will stop the muck getting into the pump and scratching the pistons.....any comments on this? Thanks.

[rhaefe]

Acetonitrile for Biosynthesis (oligonucleotide synthesis) has usually less than 10ppm water.

[Bruce Hamilton]

In another of their papers, they list the following residual water contents ( in ppm ) in acetonitrile after 1 day, from a starting water content of 2800 ppm. and a loading of 5% of dessicant.

P2O5 = 9
3A MS = 49
B2O3 = 59 ( distilled )
CaSO4 ( Drierite, carefully prepared ) = 180
K2CO3 = 250
4A MS = 450
Silica Gel = 1300
Al2O3 = 1600
CaH2 = 1900
KOH ( powdered ) =2200
CaSO4 ( Drierite, manufactuer's guide ) = 2200

It other experients, they found Na2SO4 similar to Drierite for drying other solvents, but there's no obvious data for CH3CN and Na2SO4.

I should note that some subsequent researchers have disputed some of the data for dessicants like CaSO4, claiming the humid environment in Malaysia ( where Burfield worked ) didn't allow proper oven conditioning of some of the salts.

I don't tend to dry solvents these days, but there are several competing issues, eg the finer the MS particles, the quicker the drying - however fine particles are harder to remove.

Some people have used columns of fine MS and pumped solvent through them to dry the solvent, and that has the advantage you can put a prefabricated, disposable 0.2um filter on the outlet line as it goes into the bottle.

I've used a simple pressure system to pump MS3A statically-dried solvent ( MS3a Beads tossed into solvent in a pressure-rated vessel ) through a 0.2 um filter into a storage container. worked OK, but I had to replace the filter after each few hundred mls of solvent.

As I mainly use 5 um particle columns, 0.2um filtration is OK, but smaller particle columns would be more sensitive, and use less solvent, so perhaps finer filters could be used after the 0.2 um, but...

My preference is to purchase HPLC grades that are appropriately filtered these days. Every filter step has the potential to contaminate the solvent.

Often the solvent line filters are several microns, as they are only removing large gunk, may have large flows, and have small hydraulic heads.

Please keep having fun,

Bruce Hamilton

[Mattias]

Thanks a lot for your good replies!

It is true that the main source of water is from my sample, which I dissolve in acetonitrile. The acteonitrile is of HPLC grade. If I dry the sample, I am certain that the alcohol will evaporate as well.

I have an HPLC method that easiliy separates the ester and the "hydrated anhydride" after the reaction, so I can just put in more reagent. The problem is that the anhydride costs 250$ per gram...

Molecular sieves sounds like the best idea, I will try that first!

[Consumer Products Guy]

If your alcohol is so volatile, why not consider GC???

[Mattias]

That is the obvious weapon of choice, but we don't have any GC in our lab!

(It is also quite fun to try new things also, could become a poster if it works well...)

[Victor]

Bruce-many thanks for your reply.

My questions still remain about using molecular sieve. Could it introduce contaminants into the mobile phase? And for to keep solvents dry they need to stand over the drying agents, which is practically impossible in HPLC since particulates in the mobile phase are not a great idea. I guess there are no simple answers to these questions.

[Peter Apps]

Hi Mattias

Rather than drying the solution of the sample in acetonitrile, dry the sample with sodium sulphate (easily available, chaep and clean) before you dissolve it in the acetonitrile.

Peter

[Bruce Hamilton]

My questions still remain about using molecular sieve. Could it introduce contaminants into the mobile phase? And for to keep solvents dry they need to stand over the drying agents, which is practically impossible in HPLC since particulates in the mobile phase are not a great idea. I guess there are no simple answers to these questions.

Sorry - I should have been more explicit. In my experience, using molecular sieves with a subsequent 0.2 um filter works fine for 5 um particle size columns. I almost always use guard columns - unless client specifies otherwise - and didn't notice quicker pressure increases.

I used to just put about 100g of the molecular sieves beads into the 2.5L solvent bottle and leave for a week or so with occasional very gentle swirling, filtering as required. Molecular sieves are very friable, so care is needed.

If you want to dry the solvent quickly, it would become more of an issue, as you need to purchase ( or grind, then rinse fines away ) the MS as a powder ( eg GC packed column material ), and then ensure all fines are removed.

I recycled some MS used on volatile solvents ( dried under flowing nitrogen at 265C for 24 hours - after removing the solvent by vacuum ), but only used the regenerated MS on the same solvents.

I've never had a problem with contamination from MS, and have assumed that the method of manufacture would preclude most organic contaminants. For solvents, I used laboratory grade molecular sieves, I would be a little more wary of the technical/refrigeration grades, but they should also be OK.

Please keep having fun,

Bruce Hamilton

[Mattias]

Just to let you know the outcome: We had some anhydrous MgSO4 in-house which I tested. It seems to work very well and my "water-peak" has almost disappeared.

It seems as one mol of MgSO4 can bind 7 moles of water, and for me it was sufficient to use 200 mg for 25 ml. No extra peaks form the salt what I can see (purity of salt 99.95%).

[Mattias]

A spin-off problem...

Since my sample is in almost 100% acetonitrile, I can obviously not inject so much on my RP column (max 5 µl). The molecule is quite hydrophobic since I need about 50% acetonitrile to elute it. I have tried to use HILIC but the molecule has no retention on HILIC even with 100% acetonitrile in the mobile phase. Any idea of how I can inject more, and improve my QL?? Thanks again.