Testing CAFS systems in the industrial sector

By Luca Parisi

Italian language version here.

Aim of the tests

The tests are aimed at verifying whether CAFS systems can be used in the industrial sector to mitigate the spill and/or fire effects of some products that are present at the Ferrara petrochemical plant. Since these substances are transported by road they could present a problem even outside the industrial area if involved in a road accident:

  • Pyrophoric substances;
  • Products which, upon contact with air and water, react violently and develop toxic and/or flammable gases;
  • Flammable solvents.

After searching the Internet and meeting some CAFS users at the symposium iCAFS (https://icafs.com/) it can be assumed that some of these chemicals have never been tested with CAFS.

Characteristics of the tested substances

Diethylaluminum Chloride (DEAC)

DEAC is a pyrophoric liquid. It is a highly dangerous organic compound (C4H10AlC) which develops vapours heavier than air (4.2). When dissolved in hexane, DEAC is a colourless liquid.

Since it is extremely flammable it must be kept away from sparks, flames or any source of ignition by storing it in a hermetically sealed steel container in a cool and dry place where temperatures and pressures do not undergo variations. In case of leakage, ignition sources must be eliminated and water contact must be avoided by covering the leak with sand or vermiculite. The substance is so water reactive that it might ignite upon exposure with humid air. Vapours can spread and ignite if there is an ignition source. The resulting fire releases toxic gases if inhaled.

In the event of a fire use foam, chemical powder or CO2 and bear in mind that the compound may re-ignite after extinction because of its high reactivity. Water should never be used to extinguish the flames because it causes a violent reaction.

Main physical and chemical properties:

  • physical state: liquid;
  • Water solubility: it reacts violently;
  • Colour: colourless;
  • Flash point: -22 ° C
  • Density: 0.711 g / mL
Video 1: DEAC reaction in contact with water (Youtube)

Titanium tetrachloride TiCl4

Titanium tetrachloride is an inorganic compound with the formula TiCl4. TiCl4 is a strong Lewis acid which is explosive upon contact with water, releasing HCl. It is used as a polymerization catalyst.

TiCl4 is a volatile liquid. Upon contact with humid air, it forms spectacular clouds of titanium dioxide (TiO2) and hydrated hydrogen chloride. Hazards posed by titanium tetrachloride are usually related to hydrogen chloride (HCl) being released. Titanium tetrachloride is highly irritating to the skin, eyes, and mucous membranes in humans. Acute (short-term) exposure may result in surface skin burns and marked congestion and constriction of various sections of the upper respiratory tract in humans. Acute exposure may also damage the eyes. The HCL has a TLV-C equal to 2 ppm. As an example, Carbon Monoxide has a TLV-C of 200 ppm.

Main physical and chemical properties:

•Physical state: liquid;

•Colour: light yellow;

•Odour: pungent;

•Odour threshold: no information available;

•Relative vapour density: 6.55

Isohexane C6H14

Isohexane is an alkane composed of 6 carbon atoms bonded to 14 hydrogen atoms. It is a very important solvent. It is used in reactions involving very strong bases. At room temperature and pressure, it is a colourless liquid with a smell of gasoline, of which it is an important constituent. It is almost immiscible with water. It is an extremely flammable, irritating and harmful compound. It is also dangerous for the environment and toxic to the reproductive system.

Main physical and chemical properties:

• Physical state: liquid;

• Colour: colourless;

• odour: gasoline;

• odour threshold; no information available;

• Density 0.66 g / cm³ at 20°C;

• Vapour density 2.79 (air = 1);

• Flammable temperature: -22°C (251 K);

• Autoignition temperature: 240°C (513 K);

• flammability range: 1.0 – 8.1 vol%.

List of tests

The set of tests to be carried out has jointly been decided. The tests were conducted on 27 September 2018 at the testing ground of the Ferrara petrochemical plant. The following list reports the substances and the quantities used in the tests:

  1. DEAC:

a) 1 test with 5 kg of DEAC (test No. 1)

2. Titanium tetrachloride TiCl4:

a) 2 tests, each with two litres of TiCL4 (test No. 2 and 3)

3. Isohexane C6H14:

a) 2 tests, each with 200 litres of hexane (test No. 4 and 5)

4. Protection of a tank with dry foam:

a) 1 dry test (test No. 6)

TestSubstanceQuantityExpansion ratioFoam: Oneseven B-ARLitres of waterLitres of foamCAFS type Type of hoseline
1DEAC5 kgSuper dry1 %2423Oneseven Ø 70 mm
2TiCl42 lSuper dry1 %871Oneseven Ø 70 mm
3TiCl42 lSuper dry1 %3063Rosenbauer Ø 45 mm
4Isohexane500Wet0,6 %4343Rosenbauer Ø 45 mm
5Isohexane500Wet0,6 %n.a.n.a.Rosenbauer Ø 45 mm
6Protection of a tankn.a.Dry0,6 %n.a.n.a.Rosenbauer Ø 45 mm

Table 1: List of tests

Features of CAFS and foam used

Two vehicles equipped with the CAFS system of the Trento Fire Brigade were used.

  1. Scania Rosenbauer (2003);
Figure 1: Scania Rosenbauer, Trento Fire Brigade (Trento FB)

2. ABP Man Gimaex (2013).

Figure 2: ABP Man Gimaex, Trento FB (Gimaex)



Oneseven AFFF B-AR

  • Induction Rate 0,6 %;
  • Expansion Rate (0,6% F.W.) >6,5;
  • Drainage Time >2´;
  • Expansion Rate (0,6%, S.W.) >7;
  • Drainage Time >3´;
  • Its shelf life is about 20-25 years if stored following the recommendations by Vs FOCUM.

Discharge mode and parameters of CAFS foam

In tests 1 to 3 (DEAC and TiCl4) the foam expansion ratio was super dry with foaming agent concentration at 1%. In test 4 and 5 the expansion ratio was wet foam with foaming agent concentration at 0.6%, while in test 6 the expansion ratio was dry foam at 0.6%.

In the super dry foam mode, the nozzle should be pointed at the external edge of the pool. In this way, the growing foam blanket is pushed forward by the foam itself. By applying the foam with this technique, you make sure that the foam slides over the liquid in a very gentle way.

Wet foam should be applied in direct mode. The only precaution is to avoid fuel splashes.

Dry foam should be applied with a gentle application. The operator must be far enough to make sure that the foam arrives in the descending phase and has already lost most of its energy before hitting the tank surface.


Disclaimer on the execution of tests.

Data collection and initial parameters entail critical issues:

  • The tests performed were not carried out with a scientific method;
  • The tests were not repeated to verify whether the results were the same;
  • The measurements of containers, extinguishing agent amounts and some products have a variable approximation margin;
  • The tests were carried out by two teams, one had no previous experience with CAFS systems;

On the other hand, it can be stated that:

Tests and data have been carried out, processed and interpreted with the utmost intellectual honesty.

Diethylaluminum Chloride (DEAC)

In test 1, DEAC is poured in a concrete tank open on one side with a metal container of approximately 1×1 mt in it. The DEAC (5 kg) is stored inside a pressurized vessel. The substance is pushed out by nitrogen. Once the container is opened, the liquid flows out through a thin metal pipe.

Figure 3: vessel containing DEAC used in test No. 1 (Trento FB)


DEAC spillage is usually addressed by covering the spill with sand and/or vermiculite. This should prevent the exothermic reaction (triggered by air humidity) from igniting other fuels. In doing so, however, the product remains still active and poses a potential danger. That’s because the exothermic reaction that consumes it is not underway;

The purpose of the test with dry CAFS is to verify if the low water content in the foam combined with the foam consistency is able to mitigate the reaction.

Video 2: Test No. 1. Coverage of a DEAC spill with CAFS (Trento FB)

Test No. 1

During the test, the cylinder containing the substance was closed and reopened twice. More precisely, it was opened at minute 00.01, closed at 00.58, reopened at 01.26 and reclosed at 02.10 and finally opened at 03.07 until exhaustion.

This explains why flames seemed to disappear at certain times.

The test showed that DEAC reacts with CAFS foam producing outbursts and flames.

The water content of CAFS foam, even if in super dry mode, reacts with the organic metal and triggers an exothermic reaction.

However, the projection of flames and smoke during the reaction is limited and no fireball was produced as is the case with water. The total amount used (5 kg) was completely consumed.

Titanium tetrachloride TiCl4

Tests No. 2 and 3 were performed in a circular cement tank with a diameter of about 2.5 meters. Inside the tank there was a metal container (0.4×0.25 meters, 0.12 meters deep). In test No. 2 the metal container was completely dry to minimize the evaporation of the product while in test No. 3 the same container had water on the bottom before the start of test.

The substance was contained in a glass bowl of 2 litres which was then broken exposing the product to air. As soon as the container was broken, the liquid began to emit a large amount of vapours into the atmosphere. The aim of these tests was to verify whether CAFS foam can cover Titanium tetrachloride spills and prevent vapours from spreading.

Some TiCL4 users cover spills with medium expansion foam. The foaming agent they use is called «Neutral Plus» produced by Bioex (http://www.bio-ex.com/images/Fiches_gamme/BioEx.Neutral_Anglais_Web .pdf). The manufacturer suggests a minimum critical flow of 20l/min∙mq.

Video 3: Test No. 2: covering a TiCl4 pool with CAFS (Trento FB)

Test No. 2

In test No. 2, the tank was covered by about 10/15 cm of super dry CAFS foam.

After applying the foam, we realized that the vapours pierced the foam and leaked out. After applying another layer of foam the vapours stopped leaking.

Video 4: Circular tank containing a metal container for TiCL4 (Trento FB)
Video 5: Test No. 3, rupture of the bowl with TiCL4 (Trento FB)

Test No. 3

In test No. 3 on TiCL4, the tank was covered by about 20/25 cm of super dry foam. After applying the foam, no vapours leaked out.

The layer of foam was not only thicker than the previous one but also more homogeneous. The foam blanket remained compact, preventing evaporation for a significant amount of time without need for more foam.


The test on hexane fire took place inside a rectangular pool of about 10 meters x 3 meters.

The hexane was manually transferred form a tank into the concrete tank. Ignition occurred via an open flame. Pre-burn time lasted approximately 30 seconds.

Video 6: Test No. 4, hexane pool fire (Trento FB)

Test No. 4

Test No. 4 consisted in a hexane pool fire. The fire was extinguished with a CAFS wet outlet Ø 45 mm, from a distance of about 10 meters;

90% of the flames were extinguished in 40 sec;

Complete shutdown happened in ca 120 sec;

Total amount of water: 434 litres;

Total amount of foam: 3 litres.

Test No. 5

Figure 4: test No. 5, hexane pool fire: second attempt (Trento FB)

Test 5 is the repetition of test 4 (pool fire hexane). The test could not be carried out because the previous foam coverage prevented re-ignition even though the foam layer was removed and new fuel was added.

The characteristics of the foam used indicate a good resistance to re-ignition according to the EN 1568-3 standard:

I: extinguishing capability with direct stream on hydrocarbons;

B: Resistance to re-ignition: class B in the ranking from A (the best) to D (the worst).

Covering a tank with CAFS foam

Test No. 6 consisted in covering a cylindrical metal tank. Using foam to protect structures from radiant heat is one of the most frequent situations in which CAFS is used.

The most suitable expansion ratio is dry foam. If the surface is rough or irregular, the dry foam adheres well even if on a vertical surface (plaster and wood). The dry CAFS foam was applied with a Ø 45 mm hose.

Video 7: test No 6, covering a tank with dry CAFS foam (Trento FB)

Test No. 6

The foam on the vertical part of the tank started to fall down after a few moments, while the foam on the top (less steep) remained in place for a few minutes.

The team holding the hoseline had never used a CAFS system before.

Final thoughts

Here are a few considerations on the tests carried out.

Diethylaluminum Chloride (DEAC)

The use of CAFS on DEAC triggers an exothermic reaction because of the presence (even if very low) of water in CAFS. The reaction has a much lower intensity than the one triggered by pure water or NAFS (Normal Air Foam System). The blanket of CAFS foam that covers the spill seems to partially contain and reduce the energy released by the reaction.

Operators work in safer conditions since they are at a greater distance than with other systems.

At the end of the test the DEAC is completely consumed.

Titanium tetrachloride TiCl4

Covering a pool of TiCL4 with CAFS foam prevents vapours from being released. Note that the covering layer must be at least 20/25 cm thick. Even in case of rapid evaporation due to a wet container, CAFS foam is able to contain the release of vapours.

Operators work in safer conditions since they are at a greater distance (10/15 meters against 2/4 needed for NAFS) than with other systems of vapour abatement and containment (Medium Expansion foam).


The short shutdown time of 90% of the flames reduces the possibility of a domino effect on other plants or buildings. The impossibility of re-ignition, despite adding new fuel, makes the area much safer for rescuers and reduces the amount of foaming stock need to keep the foam coat intact.

The fact that the foam blanket recaps quickly limits the release of flammable vapours, which are due to the temperature of the liquid far above its Flashpoint. Shorter shutdown time means less hot structures with less chance of reignition.

Covering a tank with CAFS foam

The dry CAFS foam doesn’t remain long enough on the vertical walls of a cylindrical metal tank. The gently applied dry foam remains on the top of the cylinder but as soon as it reaches the vertical walls it falls back on the ground.

Using foam instead of pure water allows to reduce the flow rate and makes it easier to find hot spots. If the surface of the structure to be protected were rough (e.g. the plaster coating of a building), the foam would remain on the vertical surface for a longer time.

Proposals for the future

Considering the potential of the tests and the results obtained, the tests should be repeated. In order to collect useful information which can help meet the needs of emergency teams, the tests should:

  • use a scientific method in data collection;
  • use larger quantities of substances;
  • consider different types of foam;


We wish to thank the following partners for working so hard at our joint project:

  • IFM Ferrara S.C.p.A;
  • Trento Fire Brigade;
  • Ferrara Fire Brigade;
  • Eni Versalis;
  • Basell Poliolefine Italia Spa.