Saving Shipwrecks

The oceans are the world's biggest electrolyte solution. They are dangerous, corrosive places for ships that have sunk. When the ship is of historical significance—say, the U.S.S. Monitor, a Civil War-era ironclad, or the C.S.S. Hunley, a Civil War submarine—historians and archaeologists want to save it. And there are scientists who want to help them.

The electrochemistry of oxygen—in the form of oxidation of metals—is a main cause of damage to shipwrecks. It's only fitting, therefore, that electrochemical reduction should be one of conservators' main tools to reverse that damage. The standard way to treat marine artifacts is to soak them or to perform electrolytic reduction on them in high-pH solutions.

There are three main objectives in using electrochemical methods for marine artifact conservation, according to Donny L. Hamilton, director of the Center for Maritime Archaeology & Conservation at Texas A&M University. The first of these is to convert corrosion products back to a metallic state. "Reduction is critical for preserving something as close to the original surface as possible," he says. But electrochemical methods cannot reverse corrosion of iron or other reactive metals.

A second objective is to remove chlorides, especially from objects containing copper or iron. Chlorides from seawater make metals more prone to attack by dissolved oxygen through formation of complexes and changes in reaction mechanisms.

The third use of electrochemistry is for mechanical cleaning. Often, marine artifacts are encased in a "concretion" layer that is a mixture of corrosion products and biological and mineral deposits. Turning up the current density causes a vigorous evolution of hydrogen bubbles, which can knock off the encrustation. This approach works for wrought, but not cast, iron because the engraving in the delicate graphitized layer on cast iron would be lost in the process.

In one sense, these ocean deposits are a sign of decay, but in another sense, they're the best protection a shipwreck has. "It's the concretion that separates the corroding metal from the immediate physical environment and basically gives marine iron its inherent long-term stability," says Ian MacLeod, the executive director for collections management and conservation at Western Australian Museum.

Often the concretions conceal smaller artifacts. "You have to treat each concretion as if it were an excavation square, keeping track of where everything is and what all the associations are. Probably more archaeology is done in the conservation lab than in the field," Hamilton says.

Before electrochemical methods can be applied, conservators must first assess the condition of a marine artifact. Understanding the condition of a shipwreck can help site managers make better decisions about how to treat a site. Should they treat artifacts on-site or excavate them to bring back to the lab? Or, is an object stable enough that the decision can be postponed while more pressing needs are addressed at other sites? MacLeod sees himself as someone "who's learned the language of corroding metals on shipwrecks." He wants to use that language to provide site managers with enough information so that they can determine the best course of action.

MacLeod's team has been able to show that an object's profile on the seafloor ultimately determines its corrosion rate. Something located in a depression in a reef will decay much more slowly than something that is elevated on top of other artifacts. That means that on a single shipwreck, different sections can decay at different rates. These differences are the results of differences in the flow of water, and therefore dissolved oxygen, around the objects. "On a Japanese World War II ship, the gun corrodes at a much higher rate than the windlass, which corrodes at a higher rate than the bollard on the deck," he says. "Even though they may be at similar water depths, they have very different profiles."

MacLeod's goal, he says, is to "put myself out of a job" by determining equations that describe the conditions in all possible shipwreck environments. These equations relate the logarithm of the annualized rate of corrosion to the electrochemical corrosion potential. Someone using the equations would need to know the dissolved oxygen concentration, the water depth, and the corrosion potential. For most shipwrecks, the pH of the metal interface under the concretion can be used as a measure of the localized corrosion potential.

By working with a variety of shipwrecks in the open ocean and in lagoons around Micronesia, MacLeod has shown that the logarithm of the corrosion rate decreases linearly with water depth. "We've been able to develop equations that allow a deck-chair manager to predict how fast an iron wreck will be corroding," he says. "While there's still plenty of work to do, we're in a much better state of understanding deterioration on wrecks to be able to make informed decisions."

Shipwrecks behave differently depending on their age, and their behavior can change drastically in a short period of time. MacLeod saw a startling example in the Fujikawa Maru, a ship that sank in February 1944 when the U.S. Navy targeted Japanese ships moored at the Truk Lagoon during Operation Hailstone.

When he first measured the pH and voltage at the site in 2002, MacLeod was surprised to find that the relationship behaved as if the entire structure was governed by a single thermodynamically controlled electrochemical process. In older wrecks, corrosion disrupts the electrical continuity. Thus different parts of the ship—the engine, the hull, the boiler—behave like isolated electrochemical cells. The Fujikawa Maru, however, was showing signs of "unzipping," in which corrosion was eating into it but was not yet advanced enough to alter the electrochemistry. MacLeod predicted that the corrosion mechanism would change in four or five years. When he returned in 2006, the corrosion mechanism had indeed changed to that of a classic historic iron shipwreck.

"We've got a better idea of how long it takes before something turns," MacLeod says. "In another few years, we'll be able to give people a series of relationships between young, medium-age, and old shipwrecks. If a ship sinks in a national park, the government needs to know how fast it is going to decay, what's going to happen in years one to five, and then five to 50, and 100 years."

Once they have assessed the condition of a shipwreck, conservators can start saving it. Conservation can begin while an object is still on the seabed by attaching a "sacrificial anode" and reversing the polarity of the electrochemical reaction that is causing the metal artifact to corrode. In this way, conservators force the sacrificial anode to corrode instead of the object they're trying to conserve. MacLeod has shown that pretreating an artifact such as a cannon while it is still in the ocean can reduce the time required to conserve the object by years.

These anodes needn't look like your typical electrodes. "Because our museum operates on a very tight budget and because we've worked on remote islands where they had no anodes available, we use scrap aluminum alloy engine blocks from old wrecks and vehicles as anodes," MacLeod says. "They work perfectly well, and the price is normally free or you exchange them for a few beers."

MacLeod and his team simply bolt one end of a cable to the anode and another to a metal clamp. They drill a hole through the marine growth and attach the clamp to the object being conserved. The current from the anode draws the chloride ions out of the metal. With zinc or aluminum anodes, the reaction is gentle enough that it doesn't harm the marine growth layer. Magnesium would react too quickly and blow off the concretion.

Electrolytic reduction may be the standard treatment method for marine artifacts, but it's a painfully slow process. Back in the lab, cast-iron artifacts can take three years on average to conserve and wrought iron about half that time, according to Jean-Bernard Memet, the former head of the maritime department at Arc'Antique, a conservation lab in Nantes, France.

Researchers on the Hunley project at the Warren Lasch Conservation Center in Charleston, S.C., are working on a new treatment method that could slash the time required to treat an object. The Hunley is a Confederate submarine.

The Hunley researchers are using water at elevated temperatures and pressures to extract the chlorides from iron artifacts. Because the temperature and pressure used are not high enough to generate a supercritical fluid, the method is known as subcritical extraction.

The catalyst for these studies was Michael Drews, professor emeritus in the School of Materials Science & Engineering at Clemson University, who has done research in supercritical and subcritical fluids since the late 1980s. His initial involvement with the Hunley project during a sabbatical in 2002, however, was because of his expertise in textiles. He first worked on fabrics from crew members' clothing, but he quickly switched to working on corrosion and the conservation of the submarine itself.

From his earlier research, "I knew how to make metal go away in super- and subcritical water," Drews says. "Listening to Paul Mardikian [the conservator in charge of the Hunley project] talk about the traditional ways to stabilize archaeological iron, in the back of my mind a little light bulb went off. I know how to make [metal] go away; I think I might know how to stabilize it."

Drews spent two years at the Lasch Center figuring out how to use subcritical water extraction to stabilize iron artifacts. His team now uses water at temperatures of 180 °C and pressures of 600 to 800 psi. The temperature is a compromise to keep costs down when they move to larger objects. "If you were going to treat something very large, the higher the temperature, the more expensive that's going to be," Drews says. The pressure is actually much higher than necessary, he notes. In fact, the necessary temperature and pressure are relatively easy to attain and don't require any super-high-pressure apparatus, he adds.

The researchers have plans to build a 30-L reactor, but right now their largest reactor is only 600 mL. Using these smaller reactors, they've treated wrought-iron rivet heads from the Hunley and sections of cast-iron Civil War artillery shells as test samples. They've compared the results of subcritical extraction with those of soaking electrolysis and electrolytic reduction, the standard methods used for removing chlorides. They find that subcritical extraction is significantly faster than the standard methods. In head-to-head tests, they find that the chloride level reaches equilibrium in one to three days with subcritical extraction, whereas soaking or electrolytic reduction can take as long as 150 days.

They don't, however, know exactly why subcritical fluid extraction works so much faster. Drews suspects that it is related to the improved transport properties of the subcritical fluid. The subcritical water has a higher diffusion rate and lower viscosity than water under ambient conditions. "The corrosion goes deep into the pores. Things like surface tension and viscosity and transport properties become limiting factors," Drews says.

He and his team still have much to learn about the process. They don't know whether subcritical extraction is causing significant changes in corrosion products that may be accelerating the removal of chlorides and protecting the artifacts against further corrosion, Drews says.

Drews doesn't yet know whether subcritical extraction can be used with more complex objects, such as those with multiple types of metals or complex topographies. "Typically what people have done in the past with more complex artifacts is take them apart and treat each of the parts separately," he says.

Even though the results with subcritical extraction look promising, it won't be used as a main treatment method for the Hunley. There isn't a reactor large enough for the submarine, and the researchers are trying to minimize disassembly. "For the Hunley itself, we're recommending a traditional soaking process with no electrolysis and that the Hunley be displayed in a very controlled environment," Drews says.

Spectroscopic methods are also helping conservators understand the corrosion on historic shipwrecks. Desmond C. Cook, a physics professor at Old Dominion University in Norfolk, Va., is working with the conservators of the U.S.S. Monitor, a Civil War ironclad.

Cook is helping the conservators figure out what minerals are in the corrosion layer between the concretion and the metal on the ship's turret. Ultrasonic measurements of the corrosion indicate that it is 0.8 inches thick on the outside and 0.9 inches thick on the inside, which works out to a corrosion rate of about 0.037 mm per year.

Using X-ray diffraction and Mössbauer spectroscopy, Cook has shown that the ocean deposits on the turret are composed of FeCO₃, CaCO₃, quartz, and the iron oxide minerals magnetite and goethite. "Everything involved with the Monitor is rust related," Cook says. Given that the ironclad ship has been in the salty ocean environment for many years, such rust is not surprising, Cook says.

The real question, Cook says, is the location of the chlorides that promote corrosion. To answer that question, he has been studying regions of impurities in the wrought iron. Electron microprobe measurements indicate that these regions, known as inclusions, have variable chemistry, primarily iron silicates and iron phosphates. The inclusions that reach the metal surface provide conduits for chlorides to penetrate the metal, where they accelerate corrosion. Cook's research shows that the inclusions also trap chlorides once they are inside the metal, hindering their removal and leading to the possibility of further corrosion, even after treatment. He has shown that the subcritical extraction technique developed at Clemson is able to remove trapped chlorides.

But what if conservators can't or don't want to bring artifacts back to a laboratory? Memet devised a way to remotely monitor electrochemical treatment of artifacts while he was still working at Arc'Antique. Now, any site can become a makeshift conservation laboratory.

The impetus for developing a remote monitoring method was two-fold. First, most of the artifacts restored at the lab are large enough that transporting them is expensive. Second, Memet and his colleagues had calculated that a three-year conservation process for a cast-iron cannon usually involves only two weeks of hands-on work. Now, someone still must be on-site for the two weeks of manual intervention, but the process can be monitored from anywhere in the world for the rest of the time.

To monitor the process, Memet and his colleagues replaced a standard power supply with a programmable one that can be controlled remotely. Previously, Memet would measure the pertinent parameters—the current, cell voltage, and the object's potential—on a weekly basis. Now, measurements are taken every minute, and the hourly average is used to make any adjustments necessary to keep the artifact at optimal conditions. Memet believes that this will result in shorter overall treatment times.

Memet has remotely monitored the conservation of cannons recovered from sunken ships near Saint-Malo, France. His current location in Arles is more than 1,000 km from the site, and he has even checked on the site while traveling in the U.S. "We can have connections all around the world," he says. "It depends on the Internet connection."

Memet has started a company called A-CORROS, which will commercialize the system. Currently, the software is only in French, but the company is working on English and Spanish versions.

The challenge for the future, Hamilton believes, will be dealing with ships and airplanes from World War I and II. "I have never conserved an aluminum artifact," he says. His lab, however, is preparing to conserve an airplane that someone wants to retrieve from its watery grave. "Then you're dealing with all kinds of funny alloys and aluminum-clad composites they were using."

MacLeod also monitors the surroundings of new shipwrecks, so he can offer advice about components on modern ships. "Our data isn't just of use to the oddball managing the heritage of shipwreck sites," he says. "It actually has fundamental use for environmental management and disaster mitigation."