STABILIZATION OF A SUBMERGED BMW 801 D2 ENGINE
FROM A FOCKE-WULF 190
By Christian Degrigny - (paper published originally in French in the conservation journal Studies in
Conservation, 40 (1995) 10-18 and translated by the CCI-Ottawa)

• Back to site page

STABILIZATION OF SUBMERGED AIRCRAFT ENGINES
(translated by the CCI – Ottawa)

Abstract

In June 1944, allied forces shot down a Luftwaffe Focke-Wulf 190 fighter over France.
The aircraft crashed in Le Loiret River and remained there for 45 years, until it was
recovered in August 1990 by the Club Subaquatique d'Orléans. It was then deposited
in the Caen Memorial, Museum for Peace. This report presents a typical case of
stabilization of an aircraft engine following a rolonged period underwater. Here we
follow the story of the conservation of a BMW 801 D2 twin row radial engine as carried
out by the Valectra Laboratory, which developed a satisfactory treatment protocol.


A Focke-Wulf 190

Introduction

Objects made of aluminum alloys are often found in the reserve collections of technical museums; although these are relatively new metals (their industrial production dates from the late nineteenth century), they are an intrinsic part of the history of the technologies and industries that emerged with the twentieth century, such as aeronautics.

They are of particular interest to the public because they are closer to us and are still being used and developed today. Such objects include a number of aeronautical artifacts recovered from rivers and oceans, most of them dating from the last two world wars. The history of civil aviation also includes some tragically well-known wrecks, including those of Nungesser and Coli [1], Mermoz and Earhart [2]. A number of dedicated organizations are interested in this little-known heritage: TIGHAR (The International Group for Historic Aircraft Recovery, 2812 Fawkes Drive, Wilmington, DE 19808, USA) not only carries out regular exploration campaigns but also supports research programs on the conservation of light alloys, such as those conducted by Hallam [3] at the Australian War Memorial (AWM) in Canberra.

When a submerged artifact of this kind is located, there is an enormous desire to recover it, but a number of unfortunate recent experiences (the Arado 96 [4] and the Dornier 24s [5] from the lakes of Landes) have demonstrated the extreme instability of this type of object when exposed to the atmosphere; the corrosion already present is rapidly exacerbated, leading in more or less short time to the destruction of the object. The first treatments on submerged aeronautical artifacts were performed by MacLeod [6], who proposed immersing an aluminum-copper alloy seaplane float in an ammonia solution. Since 1987, numerous additional studies have been performed in France, particularly by the Valectra Laboratory of Electricité de France, and have led to the development of a less time-consuming method of treatment using electrolysis [7-9], which can be used for associated composite systems, since aluminum alloys are rarely found in isolation [10].

They are, in fact, often associated with iron alloys (engines and fuselages), with copper alloys (aircraft equipment), less commonly with magnesium alloys (which have almost totally disappeared) and, finally, with non-metallic materials such as wood, leather, rubbers or plastics.

Following a brief summary of the corrosion problems encountered on submerged aeronautic artifacts and the state of knowledge in the area of treatment, we will look at the typical case of the BMW 801 D2 twin row radial engine from the Focke-Wulf 190 that was shot down in June 1944 by Allied forces and crashed in Le Loiret River, where it remained for 45 years until it was recovered in August 1990 by the Club Subaquatique d’Orléans (France). Before it was deposited in the Caen Memorial, Museum for Peace, the Valectra Laboratory stabilized it, developing a satisfactory treatment protocol in the process. The work was carried out with the assistance of two specialists from the AWM [11].

State of knowledge on the treatment of corroded aluminum alloys -Top

While some aluminum alloys are known for their high resistance to saline atmospheres, those that are found on submerged aeronautical wrecks show significant corrosion as a result not only of the aggressive nature of the surrounding environment but also of the composition of the alloys. Despite a number of instructive counter-examples, such as the Loch Ness Wellington [12] found partially buried in the bottom of the lake at a depth of 80 m, the immersion medium is rarely harmless to these materials; the chloride ions present, even in small quantities, cause extensive damage, which is localized at first (pitting and intergranular corrosion) but eventually becomes generalized. In addition, the noble metals (steel and copper alloys) in contact with them create local galvanic cells which can lead to the virtually complete disappearance of the aluminum alloy parts on certain artifacts.

With the exception of seaplanes, which are specially protected, most aircraft are not made to withstand the damage caused by prolonged immersion in a marine environment. The alloys used in the first half of the twentieth century invariably contained copper, which considerably enhances their mechanical resistance properties but increases their sensitivity to the different forms of corrosion. After long immersion, a double layer of corrosion is formed, but it offers very little protection: the oxidized layer adhering to the metal surface is thin (0.2µm), while the layer covering it is pulverulent and contains all the polluting species, both metallic and non-metallic. Other common alloys, such as those containing manganese or a magnesium-silicon combination, are covered by hydrated white pustules over hemispheric pits. All these states are unstable in the atmosphere, because of the presence of chloride ions at the bottom of the active zones [7].

To stabilize these materials, we have for a long time opted for an electrolytic dechlorination of the alloys by cathodic polarization, which can be performed only within a very limited range of potentials and pH values to avoid the risk of pitting corrosion due to the chloride ions already extracted and present in solution, and of cathodic corrosion caused by the deterioration of the insulating properties of the surface oxide film as a result of cathodic polarization. In practice, the chloride ions are eliminated at an imposed potential, in a slightly acid, buffered (pH = 5.4) and stirred sodium citrate solution [7-9]. Because of the heavy pollution of the surface corrosion products by iron (due to the corrosion of steel objects in the vicinity) and possibly by copper (in the case of aluminum-copper alloys), these unstable compounds are dissolved before hand without polarization in a similar solution which has useful chelating properties. This approach avoids the risks of metal deposition in the case of direct polarization and, in addition, limits the thickness of the corrosion layer through which the chloride ions must diffuse. On completion of these steps, the object is rinsed in demineralized water or in tap water (under cathodic protection to avoid any fixation of additional chloride ions) and then dried.

While this treatment is satisfactory for objects containing only a single alloy, it is not appropriate when an alloy is associated with noble metals such as steel or copper alloys (bronzes, brasses and coppers). In this case, contact between the different metals has caused galvanic phenomena during prolonged submersion, resulting in accelerated corrosion of the light alloy to the benefit of the noble materials, which are well preserved but covered with concretions of varying thickness contaminated with chloride ions that must be removed during treatment. Those distributed on the metal surfaces can be eliminated by immersion in the buffered sodium citrate chelating solution. The mechanical action of the hydrogen bubbles formed at the interface between the metal surface and the concretions, provoked by cathodic polarization of the ferrous parts during a second step (a good metal core is required for this), permits the removal of the harder concretions. This cathodic treatment for cleaning iron alloys is performed in a basic solution to prevent corrosion in the event of improper polarization (the alloys would then become covered by a passivating oxide film). Caustic soda and potassium solutions are too corrosive on aluminum alloys; another electrolyte is therefore proposed for composite objects.


Weight loss (%)



Silicate concentration (M) -Top
Figure 1. Weight loss measurements for an aluminum-copper alloy versus the silicate ion concentration, after three weeks of immersion.

Figure 2. View of the FW 190 BMW 801 D2 engine from Le Loiret River before treatment.

During similar treatments performed on a Pratt & Whitney engine at the AWM, we found that sodium metasilicate Na2SiO3 was suitable because of its corrosion-inhibiting properties on aluminum alloys [10]; however, it was determined that, beyond a concentration of 0.08 M, significant corrosion occurred, with the formation of fibrous white aluminum silicate compounds. We therefore repeated these studies, establishing, in accordance with ANSI (American National Standards Institute) standard G-172, mass loss curves for aluminum-copper alloys (Figure 1) and aluminum-manganese or aluminum-magnesium-silicon alloys (materials commonly found in aeronautical wrecks) that are non-corroded and immersed in metasilicate solutions at various concentrations.

We found that the usable range of inhibitor concentrations (no evident corrosion) is broader for alloys without copper than for aluminum-copper alloys.

0.05 M < Cinhibitor for Al-Mn < 0.45 M approximately;
0.05 M < Cinhibitor for Al-Mg-Si < 0.3 M approximately and
0.05 M < Cinhibitor for Al-Cu < 0.25 M.

These tests were then repeated using identical materials that had been corroded as a result of immersion for 40 years in fresh water (Lake Biscarrosse, Landes). This time, we found that the usable concentration ranges were significantly lower. For instance, for aluminum-copper alloys, which are the most sensitive to any form of corrosion, the safe range is below 0.05 M; for clean materials, in contrast, only surface tarnishing occurs at this level. These low inhibitor concentrations are explained by the fact that it is difficult to form a protective layer when the material is already covered by corrosion products. If the concentration increases, the inhibitor combines with these products and protection becomes impossible, as was observed indeed on the AWM’s Pratt & Whitney engine. In order to retain a conducting solution, a concentration of 0.04 M was finally selected, with a pH of 12.4.

Table I. Composition of the aluminum alloys found on the BMW 801 D2 engine

The procedure selected for the treatment of aluminum alloy/iron alloy composites is summarized below:
_ pretreatment of the entire object by immersion in a chelating buffered sodium citrate solution (pH = 5.4);
_ removal of concretion (if necessary) from the ferrous parts by cathodic polarization in a stirred sodium metasilicate solution (0.04 M);
_ dechlorination of all metal parts of the object by cathodic polarization in a stirred and buffered sodium citrate solution (pH = 5.4);
_ final rinse of all metal parts of the object in tap water under cathodic protection to eliminate citrate species.
Following these operations, the item must be completely dried before any protection is applied.

Description of the BMW 801 D2 engine and diagnosis before treatment -Top



Figure 3. Radiographic image of a cylinder head from the BMW 801 D2 engine.

Figure 4. Fibroscopic image of the interior of an intake pipe, showing one of the valves.

This engine came from the most advanced German fighter aircraft of the Second World War, the Focke-Wulf 190. The original design was based on a Pratt & Whitney licence and consists of 14 cylinders in a twin row with direct injection (Figure 2). The lower part of the engine still reflects the violence of the impact: six cylinders have been ripped off. In addition to aluminum alloys (Table I) and iron alloys, it contains traces of magnesium alloys as well as rubber and various polymers. The apparent good state of preservation of the metal parts has been confirmed by radiographic and fibroscopic imaging. X-rays taken of one of the cylinders (Figure 3) and of the propellor speed reduction unit, with the assistance of a miniaturized linear accelerator called MINAC, show no cracks in the objects. A fibroscopic probe, inserted into the cylinders and intake and exhaust pipes, showed that the degree of corrosion inside the engine is very similar to that found on the exterior pitting on the aluminum alloys and thin concretions on the iron alloys (Figure 4).

Table 2.
Analysis of corrosion products present on certain metal surfaces of the BMW 801 D2 engine

Analysis of the corrosion products present on the various materials (Table 2), obtained by X-ray diffraction and X-r
ay fluorescence, revealed the presence of calcite (CaCO3) and quartz (SiO2) over the entire engine (derived from the environment in which it was submerged), ferrous hydroxide and carbonate on the iron-based alloys, and extensive white agglomerates of magnesium carbonate and hydroxide on the remaining magnesium alloy parts. No aluminum-based compounds, however, were found. It should be noted, finally, that the aluminum-silicon alloy pistons are covered with lead oxide under the layer of calcium carbonate, probably from the lead contained in the lubricating oil or fuel used. No chloride compounds were identified, possibly because of the non-aggressive environment (fresh water) in which the engine was submerged.

The aluminum alloys were not seriously corroded despite direct contact with ferrous parts. The use in this engine of aluminum-magnesium-silicon alloys, which are more resistant than aluminum-copper alloys, is of interest, since it is rare to find them in such proportions on artifacts from the Second World War.

Treatment of this engine thus amounted primarily to surface cleaning rather than stabilization by the extraction of chloride ions. Nonetheless, we applied the complete procedure for an iron alloy/aluminum alloy composite, in order to test it under standardized conditions for the treatment of a large object.

Description of the treatments -Top

The metal surfaces were covered with weakly adherent agglomerates, mixtures of corrosion products and debris from the environment in which the aircraft was buried. Since these agglomerates are likely to produce rapid saturation of the chelating sodium citrate solution used for the first phase of the treatment, a preliminary mechanical cleaning was performed. To avoid direct handling of the engine, a specially designed steel support was developed to permit safe and easy movement. Since one third of the engine’s cylinders were missing, the most complete (and therefore the heaviest) part was tipped down. The system thus obtained was not only more stable but also more compact (Figure 5).


Figure 5. Immersion of the engine in the tank for
dissolution of the corrosion products in a buffered
sodium citrate solution (pH = 5.4).

Dissolution of surface corrosion products -Top

The buffered sodium citrate solution is prepared in intermediate vats because of the large quantities required: 42.5 kg of anhydrous citric acid and 21.25 kg of caustic soda to 4.43 m3 of solution to produce a pH of 5.4. Because of the solution’s strong chelating capacity, the steel tank required cathodic protection. Preliminary tests showed that cathodic polarization of the steel at a potential of -1.2 V/SSE (SSE = saturated mercury sulphate electrode) would prevent any corrosion (no discolouration of the solution and no change in pH: high consumption of Fe2+ ions would, in fact, result in saturation of the solution, which would turn yellow as the pH increased). An Acore RGT 444 generator was used; the necessary potential was obtained with a cell voltage of 2V and a 5A current. Electrical contacts were provided by heavy cables fastened to the walls of the tank, and to the expanded 304 stainless steel anode cage placed inside it, by means of bolts, also made of stainless steel.

Any interference with the treatment due to the corrosion of the tank being prevented, the engine was submerged (Figure 6) and positioned on the bottom of the tank, insulated by a PVC (polyvinyl chloride) plate. The lifting system was left in place so that the engine could be raised at any time. The pH of the solution was monitored throughout this step (seven days). Figure 6 shows its evolution: there is a progressive rise, probably due to slight corrosion of the iron in the support and the iron alloy parts of the motor (crankcase, gears and cylinders), followed by stabilization around pH = 6.

pH

Time (hours)
Figure 6. Changes in the pH of the sodium citrate solution during pre-treatment.

Following this operation, the engine was removed from the tank and rinsed with tap water under strong pressure. It was noted at this time that the concretions had largely disappeared. Further mechanical cleaning was performed to eliminate the magnesium-based agglomerates. On conclusion of this step, some concretions were still present on the iron alloy parts; we attempted to eliminate them by cathodic polarization.

Cleaning of the ferrous parts under cathodic polarization.

The tank, still cathodically protected, was refilled, this time with 0.04M sodium metasilicate (Na2SiO3). Cathodic polarization of the ferrous parts of the engine requires electrical contact with all these areas. The aluminum alloy parts are polarized only indirectly (by contact with the ferrous parts); since they are very sensitive to the cathodic corrosion phenomena [7, 8] induced by hydrogen bubbles formed on the surface of the metal parts during treatment, the solution was vigorously agitated in their vicinity to eliminate these bubbles. This was done by wrapping the engine with a flexible rubber hose, which was then pierced in several locations in proximity to the aluminum alloys and connected to the discharge from a closed-circuit pump. The engine was then ready for immersion in the solution.

Table 3. Cathodic potentials of the various metal parts of the BMW 801 D2 engine during cathodic polarization
of the ferrous parts in a 0.04M Na2SiO3 solution

A second stabilized power supply was used for cathodic polarization of the engine. Table 3 indicates the operating conditions used for this phase of treatment, which lasted approximately ten days, after which the engine was removed for examination. It should be noted that the stabilized power supply used (Sodilec 12V/30A) made it possible to obtain average potentials compromised between -1 and -1.4 V/SSE on the steel parts, and values were lower on those situated near the anode cage (slight hydrogen bubbling visible).

While this step was not as long as our previous tests at the AWM [10], the solution used caused no observable corrosion of the aluminum alloy parts. The inhibitor concentration used here thus appears to be more appropriate. Following this operation, the engine was once again cleaned with tap water under strong pressure. It was found that all the concretions covering the ferrous parts had been eliminated. Before continuing with the cleaning process, the solution was neutralized prior to draining; this was done by adding anhydrous citric acid (21.5 kg) to reduce the pH from 12.4 to a value close to 6.

Dechlorination of all metal parts -Top

Despite the non-aggressive nature of the medium in which the Focke-Wulf 190 was submerged at the time of its discovery, chloride ions were nonetheless present in the materials. To extract them, we cathodically polarized all the metal parts of the engine, using the same device described in the second step above. In view of the low concentrations of chloride ions anticipated, this treatment was performed in a sodium citrate solution prepared with demineralized water.

Chloride concentration (mg/l)

Time (hours)
Figure 7. Dechlorination curve obtained during cathodic polarization of the entire engine in a sodium citrate solution.

The potential of the ferrous parts was maintained at -1.1 V/SSE, while that of the aluminum alloy parts was closer to -1.5 V/SSE (average values) because of their greater proximity to the anode. The pH of the solution increased slightly (from pH 5.4 to 5.9), as in the preceding phase (see ‘Dissolution of surface corrosion products’). After approximately ten days of polarization, the engine was again removed from the tank. This is the length of time normally used in many treatments of aluminum alloy-based systems [7-10]; in addition, chloride assays performed using the selective electrode on samples taken regularly throughout the polarization process revealed that, by the end of this period, extraction had ceased (Figure 7). It should be noted that these analyses must be performed immediately after the samples are taken since the chloride ions are largely consumed as a result of organic (bacterial) development in an unstirred sodium citrate solution. We could have continued with additional dechlorination phases to ensure total extraction of the chloride ions; however, by the conclusion of this step we were already approaching our initial deadline.

Rinsing under cathodic protection and finishing -Top

The engine was immersed in tap water one last time to eliminate all aggressive species produced by the different electrolytes used in the treatments. This brief (24h) phase took place under cathodic protection, with all the metal parts of the engine connected to the cathode. The cathodic potentials measured were slightly lower than those obtained in the preceding step. We found no change in the pH of the solution. Following this rinsing, a final mechanical cleaning was performed to remove the last corrosion layers. The engine was then rotated on its horizontal axis to drain it of any remaining water and dried with a stream of hot air.

The protective coating selected for long-term conservation of the engine was the wax Dinitrol 4010, which is widely used by aeronautical companies for storing engines prior to installation. This product, which consists essentially of metal-organic compounds and polymers in a white spirit-type solvent, is very similar to the Dinitrol AV 5B used in 1991 by the specialists at the AWM to protect the Pratt & Whitney engine of the Hudson I A16-19 [10]. It is applied in spray form (average thickness per coat: 40µm); it is transparent and has no effect on steel, aluminum, rubbers and plastics.

The engine was finally delivered to the Caen Memorial in September 1992. Since that time, it has been displayed in a number of temporary exhibits without showing any notable change in the various materials treated. At the present time, it is in a storage box in the museum’s reserve collection.

Discussion -Top

The treatment of the BMW 801 D2 engine required only one month because of its good general state of preservation. Additional dechlorination phases would have been desirable, although on a composite item like this, it is difficult to determine the stability of the materials involved; the chloride ions extracted are derived, in fact, from both the steel and the aluminum alloys.

In view of the success of this operation, the proposed procedure must now be applied to other similar objects showing far more advanced corrosion. For instance, in the case of the AWM’s Hudson I A16-19 engine, which was covered with very hard concretions over the iron alloys, we found that three weeks of cathodic polarization in the Na2SiO3 solution were barely sufficient to remove some of these concretions [10, 11]. It would thus be interesting to study the behaviour of the indirectly polarized aluminum alloys in this solution over a longer period. In addition, the electrolytic parameters during the final cleaning in tap water under cathodic protection are difficult to control because of the low conductivity of the solution. Another electrolyte which is safe for all materials could be considered [11].


Figure 8. The BMW 801 D2 engine following treatment and protection with the wax Dinitrol 4010.

The rubbers and plastics do not appear to have suffered as a result of the various immersions. It might even be argued that their surface appearance has improved (elimination of soiling and reduced opacity in the case of the plastics). In fact, when this engine was exposed to the outside atmosphere for one year following its removal from Le Loiret in 1990 and prior to the beginning of treatment in 1992, these materials deteriorated as a result of the effect of humidity, ultraviolet radiation and variations in temperature. Since they could not be removed from the engine, we decided to immerse and then protect them in the same way as the metal parts; regular monitoring of their general characteristics during storage will enable us to determine whether the solutions used are in fact harmful [11].

Conclusion -Top

Treatment of a composite article is always difficult when the electrochemical behaviour of the different materials is incompatible. The decision is often made to separate these materials in order to stabilize them separately. Here, we propose to treat the whole of a large object containing both aluminum alloys and iron alloys, using chemical and electrolytic methods. The proposed procedure involves four steps: pre-immersion in a chelating solution to dissolve the surface corrosion products, electrolytic cleaning of the iron alloys by cathodic polarization in a solution that inhibits corrosion of the aluminum alloys, dechlorination of the entire object by cathodic polarization and a final rinse under cathodic protection. This treatment appears to have produced satisfactory results on a BMW 801 D2 engine from a FW 190 removed from Le Loiret River in 1990, but requires further testing on other similar objects showing greater surface corrosion.

Acknowledgments -Top

I would like to thank Christopher Adams, a chemist at the Australian War Memorial, and Carolyn MacLennan, a restorer, for their invaluable assistance in the course of this project, as well as the Valectra Laboratory of Electricité de France for allowing me to perform the work.

Suppliers
Sodium hydroxide: Prolabo, 12 rue Pelée, Paris, France.
Citric acid and sodium metasilicate: Lambert Rivière, 17 av. Louison Bobet, Fontenay/Bois, France.
Expanded stainless steel (10 x 42 mesh, 1600 x 1000 mm sheets): EMT, 10-12 Bd des Martyrs de Chateaubriand, Argenteuil, France.
Electrical cable (2 x 6 mm2, ref RO2V): Société Franco-Belge d’Electricité, 61 rue du Lendit, Bt B. Aubervilliers, France.
Dinitrol 4010: Bonnot S.A., 183 rue Raymond Poincaré, Saulxures sur Mosclotte, France.

Bibliography
1 GARREAU, G., Nungesser et Coli, premiers vainqueurs de l’Atlantique. Acropole (1990).
2 GILLESPIE, R.E., ‘The Earhart project - an historical investigation’, TIGHAR 6 (1991).
3 HALLAM, D., and ADAMS, C., ‘Finishes on aluminium - a conservation perspective’, in Saving the Twentieth Century: The Conservation of Modern Materials. Ottawa (1993) 273-286.
4 LACOUDRE, N., Fana de l’Aviation 184 (1985) 11-12.
5 BOUSQUET, G., ‘Les Dornier oubliés du lac de Biscarrosse’, Fana de l’Aviation 144 (1980) 43-47.
6 MacLEOD, I.D., ‘Stabilization of corroded aluminium’, Studies in Conservation 28 (1983) 1-7.
7 DEGRIGNY, C., ‘Mise au point d’un traitement cathodique de stabilisation de pièces en alliage d’aluminium dégradées par corrosion en soution aqueuse’, Ph.D. (Engineering) thesis, Paris VI (1990).
8 DEGRIGNY, C., ‘Conservation et stabilisation d’alliages d’aluminium prélevés sur des épaves aéronautiques immergées en eau douce’, in Conservation des Biens Culturels 3. ARAAFU (1991) 27-39.
9 DEBRIGNY, C., ‘Mise au point d’un traitement cathodique de stabilisation de vestiges aéronautiques immergés en alliages d’aluminium’, in Saving the Twentieth Century: The Conservation of Modern Materials. Ottawa (1993) 373-380.
10 DEGRIGNY, C., ‘Traitements électrolytiques de vestiges subaquatiques de grandes dimensions à base d’alliages d’aluminium’, Archéologie en Yvelines 5 (1992) 30-49.
11 ADAMS, C., ‘The treatment of the BMW 801 D2 radial aero engine rescued from the Loiret river’, RA 93 1034, EDF-GDL-Valectra (1993).
12 FLOWER, S., ‘The Loch Ness Wellington’ in After the Battle (1985) 42-47.Author
CHRISTIAN DEGRIGNY, born 1961. Engineer, Ecole Nationale Supérieure d’Electrochimie et d’Electrométallurgie de Grenoble, 1985; Ph.D., Université de Paris VI, 1990; researcher, Valectra Laboratory, Electricité de France, 1991 and 1992; assignment to the Australian War Memorial in Canberra, 1991; currently with ART Métal (O. Morel Lab), a laboratory specializing in the conservation of historic metals.
Address: ART Métal (Labo O. Morel), 9 rue Mégevand, 25000 Besançon, France.

Back to site page