Index Home About Blog
From: (Bruce Hamilton)
Newsgroups: sci.chem
Subject: Re: Dry Gas
Date: Thu, 14 Aug 1997 18:51:54 GMT (Louis Hom) wrote:

>I was wondering if someone out there could tell me the mechanism behind
>that fuel additive that you buy to remove water from your fuel system.  I'm
>guessing there's some sort of hydrolysis (maybe make a fatty acid or
>something) or maybe addition to an alkyne or something. (I _am_ just a cell
>biologist though.)

No. it's purely to do with improving the miscibility of the
normally-immiscible fluids. Let' start with the Gasoline FAQ,
( available from the FAQ archives at in

[ begin extract ]

8.7  How can I remove water in the fuel tank?

If you only have a small quantity of water, then the addition of 500mls of
dry isopropanol (IPA) to a near-full 30-40 litre tank will absorb the water,
and will not significantly affect combustion. Once you have mopped up the
water with IPA, small, regular doses of any anhydrous alcohol will help
keep the tank dry. This technique will not work if you have very large
amounts of water, and the addition of greater amounts of IPA may result in
poor driveability.

Water in fuel tanks can be minimised by keeping the fuel tank near full, and
filling in the morning from a service station that allows storage tanks to
stand for several hours after refilling before using the fuel. Note that
oxygenated gasolines have greater water solubility, and should cope with
small quantities of water.

[ end extract ]

So your question may be " why IPA, why not methanol or ethanol,
which are usually cheaper ?". The answer is dependent on two
aspects, the actual ability to act as a cosolvent, and the
effect of temperature on the phase separation point of gasoline/
cosolvent mixtures. The aromatics content of the fuel also has
a profound effect on the amount of water that can be tolerated,
with 25% aromatics tolerating about 0.1% water and 40% aromatics
tolerating about 0.2% water at 15C.

Adding 15% by volume of anhydrous methanol to a typical
non-oxygenated gasoline will provide a water tolerance ( the
amount of water that can be added before phase separation occurs )
to around 0.16% at 15C, but at 0C the tolerance would be close to
0%, ie the water/methanol would be a separate phase without the
addition of water. Adding more that 15% of methanol can adversely
affect driveability on engines without management systems designed
for oxygenated fuels.

Adding about 3% by volume of anhydrous iso propyl alcohol
will provide a water tolerance of about 0.35% at 15C and about
0.2% at 0C, thus it's ability as a cosolvent is superior on
a volume basis, and far less temperature sensitive than

Ethanol is between the two, with a 15% blend tolerating about
1% water at 15C. The best additives are the isomeric propyl
and butyl alcohols, and above them, the tolerance decreases
as the alcohols become less water miscible.

The above numbers are from memory, and are approximate, but
you get the general idea. The experiment is easy to do if you
want real numbers for your local gasoline. Take 100 mls of
gasoline ( remembering that it's highly flammable ) in a
stoppered glass measuring cylinder and measure temperature.

Add one drop ( ideally a 1ml syringe, but any narrow tubing or
dropper can be used, and use 20 drops of water = 1 ml ), stopper,
and shake. If the solution goes hazy, warm slightly in a bucket
of warm water and measure the temperature that it clears.
If it stays clear, reseal and put it in a fridge or freezer ( in
a container that will contain the fuel should the glass break
- and remember that all vapour will be flammable and that the
controller and light use non-flameproof switches ).

When hazy, remove and record temperature that the haze clears,
and if it's still clear, add another drop and repeat.

Now repeat adding 5,10,15% anhydrous MeOH, EtOH, or IPA ( remembering
that many IPA products ( eg as an antistatic ) actually are not
anhydrous, and can contain up to 40% water - which somewhat
defeats the purpose :-) ). You can draw a graph of water tolerance
versus temperature that shows the different positions and gradients
of the alcohols.

          Bruce Hamilton

Index Home About Blog