Iron contamination: The bolts corrode again

Analyses of the Vasa's wood after the salvage showed up to 2 weight% iron (Arrhenius 1973). Our x-ray fluorescence line scans of core samples show large variations, with high iron concentration especially close to the surface, then decreasing to 0.1% or lower deep inside the wood. The source is Fe²⁺(aq) ions from corroding bolts and nails that  penetrated the submerged wood.

X-ray fluorescence line scan along 70 mm of the core sample No. 13, through a dunnage plank in the cabin, shows high iron content (> 10 mass%) close to the outer surface. The iron concentration increases at cracks and at the inner plank surface at about 52 mm depth.

Periodically, especially in the beginning, the oxygen supply was sufficient to corrode all articles of wrought iron in the hull of the submerged Vasa.  During and after the salvage in the 1960s, about 5500 new iron bolts (as counted in April 2002) coated with epoxy and/or zinc were inserted into the empty holes left by the rusted ones. Now, some forty years later these "new" bolts display severe corrosion damage. Some unforeseen factors are that the increasing acidity dissolves the zinc coating, and that corrosion of iron metal is accelerated in contact with PEG in humid wood (Giulminot 2000).

The sometimes metre-long epoxy and zinc treated iron bolts, replacing the old corroded bolts in the hull during and after the salvage, show corrosion damage. Subsidence when the hull dried will prevent new replacements of all bolts without damage to the wood.

The oxidation of sulfur to sulfuric acid described previously (see) is probably catalysed by the presence of iron ions. Another detrimental effect is that iron compounds catalyse oxidative degradation of cellulose (Johansson 2000). Tests have shown that wood in contact with rusting iron loses a considerable part of its tensile strength over time (Marian 1960). Therefore, it is desirable to remove as much iron as possible from the wood.

Blood red fertilizer dissolves rust

To dissolve and remove iron compounds from the wood and at the same time raise the pH is a chemical contradiction, since iron(III) oxyhydroxide (rust) precipitates already at low pH values (Figure). In order to dissolve rust in an alkaline solution, special agents must to be used that form very strong and soluble complexes with iron(III) ions. Especially strong complexes are formed when the complexing agent encloses and binds to the iron(III) ion with six surrounding atoms, arranged as a well-fitting cage. Such complexing agents are called chelates (chela = claw). A well-known chelate is EDTA, ethylenediaminetetraacetic acid, commonly used for capturing many metal ions. Chelates with particularly strong bonds to iron(III) ions can be created by adjusting the size of the cage (Ahrland 1990). Such special chelating agents derived from EDTA are ethylenediiminobis (2-hydroxy-4-methyl-phenyl) acetic acid and diethylenetriamine-pentaacetic acid, with acronyms EDMA (or EDDHMA) and DTPA, respectively ( see).

The [Fe(III)-EDMA]- complex with the iron(III) ion at the centre, enclosed by the six ligand atoms (oxygen O and nitrogen N) of the EDMA chelate. The phenol rings make the bonds in the complex stronger and also create the blood red colour. 3-D pictures of iron(III) complexes with DTPA and EDTA are found here.

It was found in the the 1960's that citrus trees on carbonate rich alkaline soils had certain diseases and low productivity due to lack of the essential elements iron and manganese. At high pH all natural sources of iron become unavailable to the plants, since insoluble iron(III) compounds (rust) form in the alkaline soil (Engelmark 2000). Because plants take up these essential elements through the roots the remedy is to provide iron as water-soluble complexes in aqueous solution, and the more stable the complexes are the less is their tendency to precipitate as rust (goethite). Chelates with phenolic rings were found to form especially strong bonds to iron(III). EDMA is such an EDTA-derivative, and has been thoroughly characterised (Ahrland 1990). The sodium salt of the iron complex [Fe(III)-EDMA]^- is produced by Akzo-Nobel Rexolin AB in Kvarntorp, Sweden, and is used in large quantities as iron micronutrient in citrus agriculture, primarily for orange groves in Spain but also when growing lemons and grapes in Italy, France, Greece, Israel, etc.

Ingmar Persson's idea was then to reverse the procedure, and use these types of chelating agents to extract the iron compounds from the Vasa's wood. The feasibility of using such chelates to dissolve iron compounds can be evaluated by means of calculated pH-diagrams (Puigdomenech 2002), which show that EDMA can keep iron(III) ions dissolved in alkaline solution up to pH = 12, before goethite, FeOOH(s), precipitates (Figure) DTPA and EDTA without phenol rings form weaker complexes with iron(III), and goethite precipitates then at pH > 9 and 8, respectively. 

The chelators [EDMA]4- and [DTPA]5-, form strong complexes with iron(III) ions and the fractions of the dominating species are displayed vs. pH in aqueous solution. The very stable deep red [Fe(EDMA)]- complex can keep iron(III) in solution up to pH = 11 before solid goethite FeOOH(s), precipitates (as rust). The yellow [Fe(DTPA)]2- complex is less stable and solid goethite,FeOOH(s), can precipitate for pH > 8. Also below pH 6 geothite or slightly less stable natrojarosite can form. The Chemical Equilibrium Software MEDUSA (Puigdomenech 2002) and stability constants from the HYDRA database and from Ahrland et al. 1990 have been used for the calculations.

Akzo Nobel Rexolin AB has generously produced EDMA without iron ions, and also DTPA, for our tests of extracting iron by using EDMA and DTPA to dissolve iron(III) compounds from planks of the Vasa. The extraction process, which can continue for months, is very clearly indicated by the dark red colour of the [Fe-EDMA]^- complex. The yellow iron(III) complex with DTPA is less strongly coloured, but DTPA is also less effective for dissolving rust, natrojarosite, and other iron(III) compounds.

Our preliminary experiments indicate that treatment with iron(III) chelating agents in alkaline solution can be developed into an efficient conservation method to remove iron and acid. For smaller wooden objects this could be performed in a series of baths, where successively the acid is neutralised, the iron compounds (not pyrite) dissolved along with the PEG, at least in the outermost parts of the wood structure, and the complexes extracted and washed away. As the last step a surface layer (with e.g. PEG 4000) should be restored. However, the long-term stability of the complexes must be evaluated before large-scale use is attempted.

It would be desirable to replace the iron bolts with bolts of inert material, e.g. epoxy coated carbon fibres. However, probably less than half of the 5500 bolts can be removed without unduly damaging the wood. Some bolts are difficult to access, and when the hull shrank and subsided during the drying process many may have become squeezed and bent.

When submerging a piece of black oak from the Vasa in alkaline sodium-EDMA solution the iron compounds in the wood form the soluble blood red complex, [Fe-EDMA]-.  Ingmar Persson brings EDMA solution for testing iron extraction at the Beckholmen laboratory of the Vasa Museum.