Texas City, Texas: Nov. 3, 2003

High school chemistry teachers forever warn students to add acid to water, not water to acid. The wisdom of this became treacherously clear when, in November 2003, a tank barge laden with 235,000 gallons of sulfuric acid capsized at a refinery dock in Texas City, Texas.

Sulfuric acid is transported for commercial purposes at 99 percent plus concentration. At this concentration, the acid is not corrosive, i.e., it will not eat the mild steel tanks and hulls of the barges used to transport it. When water began to enter the capsized barge two things happened.

First, the acid began to dilute, which is a reaction that generates heat. Second, the diluted acid began to corrode steel inside the barge, weakening its structural integrity while generating large amounts of highly flammable hydrogen gas.

For 10 days straight Coast Guard Cmdr. Paul Thomas, federal on-scene coordinator, gently tended a potentially explosive situation.

"Throughout the entire event incident command operated in emergency phase," Thomas said. "Every single day the thing pulled another trick on us. We never got to where the last plan we wrote could actually be executed."

But just as the old teacher's admonition defined the danger, it would eventually provide a safe resolution.

"It was a very hazardous situation for a long period of time but nobody got hurt," Thomas said. "At the height of the emergency more than 100 responders were on scene. To me, that says things went well."

Day One

On Nov. 3, tugs maneuvered tank barge NMS 1477 into its berth at a Texas City refinery situated on the 25-mile long Port of Houston. Video tape from security cameras show that the barge, measuring 195 feet long, 35 feet wide and 12 feet deep, had begun taking on a "pretty scary list," Thomas said.

"On the dock the guys said, 'We're not off-loading this thing,'" Thomas said. "The guys on the barge said, 'No, let's get it off-loaded as quick as we can.' The guys on the barge were trying to hook it up when it flipped over." Two people aboard escaped death narrowly.

Solid information about the contents of the barge was hard to come by in the initial phase of the emergency, Thomas said.

"The guys on the tug pushing the barge were not the guys who loaded it, so it was really hard for us to find out exactly what was on board, where it was supposed to go and how it was loaded," Thomas said. "That slowed down some of our decision making."

Authorities established incident command with a 500-yard hot zone enforced around the barge. Equipment to empty or lighter the barge was ordered to the scene.

Top priority was to drain the barge without the acid being released into the environment, Thomas said. However, there were signs that this might prove difficult, if not impossible. It was apparent early on that acid was leaking from the cargo tanks and into the voids between them. Structural components of the barge were being attacked by the diluted acid.

"We knew the acid was leaking because we were taking pH readings in the water," Thomas said. "We also knew water was going into the barge because we could see the draft increasing. We knew we had a reaction ongoing in the barge."

Day Three

At 7 a.m. on Nov. 5, with specialized equipment to lighter the barge still en route, a second dramatic event occurred. Because of cargo migration out of the barge and water's ingress, it shifted onto its port side. Surface bubbling was seen and water monitoring indicated that acid was leaking from the barge. It was calculated that the barge's displacement had decreased by approximately 800 tons.

"As soon as it came up we could see bubbling from the hatches that were still submerged, and we could obviously hear it coming out of the hatches above water," Thomas said. Unfortunately, the only hatches above water led into voids between the three main tanks, not the tanks themselves.

Thomas ordered that the hatches be closed, a decision he would later regret.

"Air was coming out and I knew that meant water was coming in," Thomas said. "The last thing I wanted this barge to do is sink at the pier. In retrospect, I wish I had sunk it at the pier, but at the time our response objective was still to minimize environmental impact by getting the acid out of the barge. The only way to do that was to keep it afloat."

Booms were deployed around the barge to contain diesel fuel lost from the cargo pump. No means existed for containing the acid leaking from the barge. Because the water was now highly corrosive, dispatching divers to more closely survey the damage was out of the question. Even remotely operated vehicles were unlikely to survive, Thomas said. Rigging the barge to be lifted would be nearly impossible under these conditions.

Personnel from Garner Environmental and the EPA continued to closely monitor both air and water.

"The water monitoring was to assess environmental impact but also to protect some critical infrastructure," Thomas said. "There was at one point an oil tanker nearby. The last thing we needed was for oil to start showing up because acid had eaten through the hull of that tanker."

The United States is divided among 13 regional response teams (RRT), each consisting of 16 federal agencies, including the Coast Guard, with interests and expertise in various aspects of emergency response to pollution incidents. Thomas immediately consulted with the Region 6 RRT, keeping them apprised of the situation.

"It was really sort of a courtesy notification at that point, although we did plant the idea that we might have to have a metered release," Thomas said.

Some of the other expertise available during the emergency was not so welcome, he said.

"We had technical experts coming out of the woodwork," Thomas said. "We had cargo experts, salvage experts, even alternative technology experts. Some of them were truly expert and some of them were not."

Newspapers quoted a University of Houston chemist as saying there was no way that hydrogen gas could be produced by the reaction the Coast Guard was describing, Thomas said.

"We figured we could get him to walk over and drill a hole in the side of the barge for us," he said. "One thing I learned is we probably need to have a list of who is truly expert in what areas and have them vetted ahead of time."

Day Four

Had luck been with the responders, this would have been the last day of the emergency. The equipment needed to lift the barge was in place and ready to go. Specialized pumps were on hand to transfer the cargo to a second stainless steel barge.

"Unfortunately, as the salvage personnel approached the barge they could hear the metal inside buckling and welds popping," Thomas said. "When we saw this happening we knew that we had lost the battle and the reaction had beaten us."

Because the hatches had been closed to keep the barge afloat and minimize environmental impact, an estimated 10 pounds per square inch of hydrogen gas had built up inside. Thermal imaging cameras revealed that the temperature was continuing to rise. Remote controlled cameras transmitted photos of bulging deck plates on the starboard side. The concern now was that a spark could be produced as welds broke, setting off the hydrogen gas.

The decision was made to increase the hot zone to 800 yards. Texas City harbor was closed to marine traffic. Two neighboring refineries were evacuated for fear that flying schrapnel from any explosion might puncture tanks containing hazardous chemicals. Determining the overall risk involved proved to be difficult.

"Everybody within each facility knows what they have, but there was no good consolidated listing of what might be affected in the Texas City harbor area in a worst case scenario," Thomas said. "Those secondary hazards were far more dangerous than the barge itself."

Texas City officials issued a Level II alert for its citizens indicating an emergency that was not under control but was confined to a small area.

"A barge full of hydrogen gas was bad enough but then put it in the middle of Texas City," Thomas said. "And the media did not miss that point." Film footage from the famous 1947 cargo ship explosions that devastated the city became standard viewing during television news reports.

"Immediately, our priority shifted from getting the acid out of that barge to getting this high-pressure hydrogen bomb defused," Thomas said.

Conventional methods for cutting through the metal hull such as an acetylene torch were immediately ruled out. Underwater operations using explosives to open one of the submerged hatch covers was likewise discounted. Drills using ultra high pressure water or lasers were considered but carried the risk that once penetrated the pressurized hull might tear apart. Escaping atmosphere could heat the jagged edges enough to cause a catastrophic spark.

Wild Well Control, experts on cutting into explosive atmospheres, were consulted. The Houston-based company is one of the world's leading suppliers of fire fighting, well control and related engineering to the oil and gas industry. Wild Well recommended applying the long established technique of cold tapping. Tapping is the precise process of drilling a hole in an on-stream system such as a tank or pipeline without spilling or igniting its contents.

A 15,000 psi rated high pressure tapping machine used for sub-sea drilling into oil wells was chosen, said Scott Powell, vice president of marine services for Wild Well. Spot modifications to adapt the machine to this specific job were made in short order.

"Normally this machine has to couple up to a flange," Powell said. "There were no flanges or anything to grab hold of on the barge. So we built a base plate for it. It had to be able to be set in place and maintain its own position just through gravity." Three-inch legs were added to the base plate so that cooling water could flow beneath it, further preventing any heat or spark buildup. The water carried away the tiny metal chips created as the tapping machine cut through the hull.

The added clearance also compensated for the noticeable bulge or crowning of the barge's deck plates, owing to the intense pressure building inside.

Day Five

Environmental monitoring indicated that the leaking sulfuric acid, heavier than water, was now affecting water within a 50-foot radius of the barge at a depth of 20 feet. Supported by a crane, the tapping machine was lowered into place on the barge. Wild Well technicians operated the machine by remote control, although it was nowhere as remote as Powell would have preferred.

"We were only 75 feet away from the machine," Powell said. "That doesn't give you a very big safety margin."

Using hydraulics, 35 to 40 pounds per square inch was applied to the bit as it turned at between 17 and 25 rpm. Setting up to cut each one-inch hole took one hour, as did actually cutting the hole. Water was kept running across the surface during the cutting and for at least a half hour afterward. Readings were checked for the presence of hydrogen coming from the hole before the equipment could be moved.

"The whole time we were drilling that barge was coming apart," Powell said.

All told, drilling five holes took nearly half a day.

"You could actually see the pressure relieved because the plates that were bulging out would fall back into place," Thomas said.

With hydrostatic balance again restored, cargo was no longer leaking from the barge. A diving crew from T&T Marine Salvage could now go into the water to wrap large slings under it for lifting.

"The divers were going in with monitoring equipment right on their shoulders so that we could pull them out if acid was detected," Thomas said. "It was still a high risk operation, but we did an extensive amount of monitoring."

Garner Environmental personnel wearing Level A hazmat gear handled all the work done above the water. The situation was still regarded as dangerous. Despite nitrogen being pumped into the barge for 12 hours, meter readings at the hatches still detected a 100 percent LEL (lower explosive limit) inside. Everything possible was done to reduce the chance of a spark.

Day Nine

The first attempt to lift the barge was made. Using a crane and derrick, the goal was to lift the barge high enough to gain access to the cargo hatches. Working in the hydrogen-filled voids between the tanks was judged too hazardous.

Even if the hatches could be reached, Thomas, working with the RRT, had come to the conclusion that lightering was no longer an option.

"First of all, nobody is going to volunteer their barge," Thomas said. "There just wouldn't be enough trucks to do the job. So the plan was now to do a metered release. We are going to put it into the water but in a very controlled fashion."

Once access to the tanks was gained, a pump operating at 600 to 1,200 gallons per minute with a special diffuser was used to release the acid 25 feet below the surface where dilution could safely occur. As with the drilling, the pumping was conducted under a water curtain with personnel wearing Level A hazmat gear.

"The plan was that when we hit pre-approved pH trigger points we would stop the pump until dilution occurred, then pump again," Thomas said.

Slowly, the barge started to lift as the amount of acid was reduced. However, this moment of success came to an abrupt halt. When the pump was temporarily shut down to allow dilution, a sudden violent reaction inside the barge sent personnel fleeing for the designated rally point about half a mile away.

"We had everybody out of the hot zone and at the rally point within seven minutes of our evacuation order," Thomas said.

Because the discharge for the pump was 25 feet down, water was siphoned back through the pump when it was shut down. The pump was not a salvage pump and did not have a check valve to prevent this.

"We ended up pouring a lot of fresh water on top of this acid and having a violent reaction," Thomas said. "The only way to stop the reaction was to start pumping again. So we suspended the triggers that we had agreed to with the RRT and got rid of the cargo as fast as we could."

With pumping resumed, the adverse reaction ceased. It would prove to be the last major upset in a problem-plagued operation. As pumping continued, the barge lightered enough that it righted itself in the water without incident. By day 10 of the emergency, the reaction inside had been brought under control.

Now able to examine the interior, responders were shocked to learn that flooring and other structural members outside the tanks had been eaten away.

"Based on the damage inside, the fact that it stayed afloat and didn't break in half when we lifted it is amazing," Thomas said.

Conclusion

Much environmental scrutiny had been invested in the Texas City harbor even before the acid barge emergency. According to Thomas, this data provided a good baseline to assess any damage to the environment. Monitoring in the months after the acid release showed that the environmental impact had been minimal.

"In fact, the ecosystem there is more healthy and robust now than it was during that study that set the baseline a year earlier," Thomas said.

Still, the decision to release the acid was not taken lightly. Only extreme circumstances made the environment a secondary concern.

"Even after the reaction that sent everyone running I got resistance from the regional response team and national response team," Thomas said. "They asked, 'Can't you come up with a better solution?' and 'Isn't there something else we can do?'"

However, seeing people run for their lives had a definite effect on his decision making process, Thomas said. Weighing the immediate danger against long term effects on plants and wildlife left Thomas with no choice. Releasing the acid involved the fewest risks for responders dealing with the emergency.

"I think that in certain hazmat situations the best solution is just to make sure nobody gets hurt," Thomas said.

Lessons Learned

• Some emergencies defy pre-planning. The acid barge incident in Texas City was a constantly evolving prelude to disaster. The worst case scenario was not a far fetched extreme but simply one wrong decision away. No set procedures could anticipate the degree of complication involved. This situation demanded thinking outside the box.
• Which expert is right? In this case the experts were often in conflict. This put an incredible amount of pressure on the incident commander. His job became a process of evaluating the conflicting opinions to choose one workable decision, knowing it may have to be further modified as the situation evolved. Some experts talk the talk, others walk the walk. Make sure you are dealing with the best people to resolve your particular problem.
• With no perfect option, the incident commander was forced to pick the least harmful of the negative alternatives open to him. Trying to protect the environment had to be weighed against the potential harm to responders and neighboring facilities. In the end, the environment had to take the hit and the consequences were not as negative as expected.

Editor's note: This after-action review appeared in "Disasters Man-Made" by David White and Anton Riecher that was published in 2011.