An explosion occurred during hot work activities, killing three contract workers and injuring seven others.

The DeRidder mill produces container board that is used in products such as boxes and cardboard displays.

The container board is made from pulp produced at the facility.

During the process of creating pulp, vapors are generated.

These vapors, which contain water, turpentine and various sulfur compounds, are collected and separated.

To separate the vapors, they first enter a turpentine stripping column.

There, most of the turpentine is removed and sent to a condenser.

The remaining vapors in the stripper condense to a liquid, containing mostly water, but also residual amounts of sulfur compounds and turpentine.

This liquid is known as foul condensate.

The foul condensate is sent from the stripping column to an approximately 100,000-gallon capacity atmospheric storage tank, used to store the liquid at or close to ambient pressure.

The foul condensate tank is primarily used to regulate the flow of liquid between the turpentine stripping column upstream and a downstream unit that removes the remaining sulfur components from the water.

During the February 8th incident, the mill was undergoing its annual shutdown.

The shutdown required mill employees and contract workers to perform maintenance, inspection and upgrade tasks throughout the facility.

One of these tasks was the repair of water piping, located above and connected to the foul condensate tank that had shifted and cracked months earlier.

The repair required welding on the piping, commonly referred to as hot work.

In preparation for the hot work, valves were closed leading into and out of the foul condensate tank, while ten feet of liquid remained inside.

The liquid was left within the tank, partially because there were no plans to work directly on the tank during the outage.

The company also assumed that the tank contained mostly water, was sealed off from the atmosphere and did not pose a safety risk.

These assumptions were, however, incorrect.

Residual turpentine, normally present in foul condensate, collected on the top of the liquid in the tank due to its density.

The foul condensate tank was designed so that by changing the height of the liquid level inside the tank, the residual turpentine would be skimmed off and sent to a turpentine recovery system.

But in the months leading up to the incident, the turpentine was not removed, because there was confusion as to who at the mill was responsible for operation of the foul condensate tank.

Due to this confusion, a valve designed to direct skimmed turpentine to the mill's turpentine recovery system remained closed for months.

As a result, leading up to the incident, more flammable turpentine was present in the foul condensate tank than anyone expected.

In addition, there are normally vapors inside the foul condensate tank in the space above the liquid.

Although those vapors can become flammable, they are supposed to be kept at a concentration that is too rich to burn.

But due to the non-routine conditions experienced during the annual shutdown, the contents of the tank likely cooled, creating low pressure within the tank.

This most likely triggered a relief valve on the tank's roof to add more air, to avoid damaging the tank from the vacuum created by low pressure.

The vacuum relief valve was likely one of the few potential sources that allowed enough air into the tank to create an explosive atmosphere.

In preparation for the hot work, the water piping was isolated from the foul condensate tank and the rest of the process by closed valves and the piping was separated physically from the tank.

Around 8 a.m. on February 8th, a mill employee used a gas detector to check for a flammable atmosphere in and around the water piping and found none.

As a result, the company issued a hot work permit for the welding work.

But even though a flammable atmosphere was not present outside the tank, there was a flammable atmosphere inside the tank.

Without knowing that the tank posed a serious hazard, three contract workers began welding on the water piping located above it.

The CSB was unable to determine an exact ignition source, but it is likely that sparks or molten slag produced from the hot work landed on or near the foul condensate tank, heating up the tank wall or otherwise igniting the contents inside.

Or it is possible that the hot work was complete, but as the tools were lowered, a welding torch fell and created an electric arc on the tank or its vent piping.

Regardless, the CSB determined that hot work activities likely ignited the flammable vapors and liquid turpentine inside the foul condensate tank.

At 11.05 a.m., there was a large explosion.

The tank separated from its base and launched up and over a six-story structure, landing approximately 375 feet away.

Three people were killed and seven were injured.

All were contract employees working near the foul condensate tank.