Evaluation of a XVIII Century Steel Bar in the Cathedral of Mexico

Abstract

A cracked segment from a XVIII century steel bar was removed from the Cathedral of Mexico City for analysis in hopes of determining the source of the bar, its manufacturing process and the significance of several cracks. The historical and experimental analysis showed that the bar was manufactured by a puddling process typical of English steels and one crack was caused by lack of fusion when small bars were forge welded to form the bar used in the Cathedral and another crack was a fracture that also accompanied the forge welding process. The cracks developed during manufacture and were not the result of in-service degradation.[PUBLICATION ABSTRACT]

J Fail. Anal. and Preven. (2010) 10:8289 DOI 10.1007/s11668-010-9325-z

CASE HISTORYPEER-REVIEWED

Evaluation of a XVIII Century Steel Bar in the Cathedral of Mexico

M. E. Noguez D. Mojica T. Robert

Submitted: 1 October 2009 / in revised form: 20 January 2010 / Published online: 18 February 2010 ASM International 2010

Abstract A cracked segment from a XVIII century steel bar was removed from the Cathedral of Mexico City for analysis in hopes of determining the source of the bar, its manufacturing process and the signicance of several cracks. The historical and experimental analysis showed that the bar was manufactured by a puddling process typical of English steels and one crack was caused by lack of fusion when small bars were forge welded to form the bar used in the Cathedral and another crack was a fracture that also accompanied the forge welding process. The cracks developed during manufacture and were not the result of in-service degradation.

Keywords Steel microstructure Forging Ancient iron

Introduction

A segment from a XVIII century steel bar that had been installed inside the East Tower of the Cathedral of Mexico City was analyzed as requested by the Institute of Engineering of the Universidad Nacional Autonoma de Mxico (UNAM, National Autonomous University of Mexico). This analysis is part of the interdisciplinary studies involved in restoring and repairing this vice royal monument that had been constructed over a 187-year period, beginning in 1626. The Cathedral was part of the Spanish

construction on the site of a previously destroyed Aztec city. Tenochtitlan, the Aztec city, was made above a lake, on islets [1] and was destroyed by the Spaniards who built a city, the capital of New Spain, in the same place, in spite of the difculty of building above water. The construction of the second Cathedral of New Spain, now Mexico City began in 1626 and nished in 1813 [1, 2].

There are two ways to obtain information about the bar segment being studied: (1) by historical research and (2) by laboratory characterizations of the segment. The quantity of steel used in the experimental phase of the analysis was minimal because of the historical signicance of the bar. The procedure used in this investigation incorporated both information pathways and is outlined in Fig. 1.

The segment investigated was from a support rod used to hold a stone globe on one of the Cathedral towers and the primary purpose of the investigation was to characterize the bar. Characterization included determining the probable source of the bar, the manufacturing process and the cause of several cracks found in the bar.

History of the Bar in the Cathedral

By the middle of the XVIII Century, the interior of the Cathedral was almost nished, but the outside was not. There were missing towers and facades and nishing the Cathedral would require more years of construction. In the second part of the XVIII century, there was a contest for nishing the towers. The winner of the contest was the Architect Josef Damian Ortiz de Castro (17501793) and his design and construction scheme were accepted [3]. From 1787 to 1791, the completion of the towers was accomplished and each tower was topped with a stone globe and a cruise, as per the Ortiz de Castro proposal. Ortiz de Castro commented: the huge weight of the belts

M. E. Noguez (&) D. Mojica T. Robert

Facultad de Qumica, Departamento de Ingeniera Metalrgica, Universidad Nacional Autnoma de Mxico,Mexico, D.F. 04510, Mexicoe-mail: [email protected]

123

J Fail. Anal. and Preven. (2010) 10:8289 83

supporting the weight of the globes, especially in the collars of four inches thick, on which the inconsiderable weight of the globes and the cruises rest, do not have iron bolts or drills [2]. The concept for the towers can be seen in Fig. 2(a) and (b).

The stone globes were held in place by a steel rod. The rod was a metallic bar or tensor that was 56 varas (about 4.8 m long) and tied the center of the globes to a wood diaphragm (blades crosshead or cedar planks) inside the tower (Figs. 3 and 4). The metallic bar is shown in Fig. 5. The top of the bar has a at head (Fig. 6).

Historical Research of the Iron-Making Process

The bar could have originated from several places: New Spain (Mexico), Europe (Spain or England), and even Asia (maybe China or Philippines).

In New Spain (Mexico), there were no iron factories, as a way of keeping the Spaniard monopoly. In 18041805, Spain and France had a conict with England and, as a consequence, the famous Trafalgar battle took place, which England won. After the English victory, Spain stopped exporting iron and supplies to America [4]. New Spain

Fig. 1 Flow diagram for the analysis of a sample of the bar located in the Cathedral of Mexico

Fig. 2 (a) East tower with globe and cruise; (b) scheme of the tower

123

84 J Fail. Anal. and Preven. (2010) 10:8289

Fig. 3 Open stone globe showing the hole for the croisse

Fig. 4 Bar with cedar vanes anchorage

Fig. 5 Metallic bar

Fig. 6 Flat head of the bar

therefore decided to begin making steel (1807). The steel making was headed by Andres Manuel del Rio and used the puddling process [5].

In the archive of the Cathedral: Copia de las Cuentas de la Fbrica de la fachada y torres de Catedral, Tomo II [2], it is written that the steel erro platina used, arrived to the Veracruz port, in the Gulf of Mexico through the Atlantic Ocean, so the possibility of the bar coming from Asia is discarded (Fig. 7). The bibliographic research, however, does not tell the origin of the bar.

The source of the bar is probably either Spain or England. Two main primary iron-making processes used at that time were the Catalan forge in Spain and Puddled iron in England.

Catalan forging is a process employed in Spain and the south of France from the XVII to the XIX Century [5].

The operation began in a furnace with a good re of wood and charcoal. Then, the iron ore and charcoal were charged in

separated columns. The ore must be charges in small quantities in order to obtain its complete reduction. Workers removed the slag and took the reduced iron particles to the hottest part of the nozzle, where the iron particles were welded together forming iron balls. The balls were joined by strokes and any remaining slag, with lower melting point than the iron, was eliminated. Iron billets, bars, utensils, etc. were then obtained by successive forges [6]. Rolling is not mentioned in the descriptions of the steel production process. This process has three fundamental requirements: proximity to a stream with enough ow to operate the forge, the presence of a forest to provide wood (charcoal) as fuel, and easy transportation of raw materials and nished products [7].

In England, iron-making factories had a great development in low-temperature furnaces, which needed vegetal coal. Later, mineral carbon (anthracite) was employed, but its phosphorous and sulphur content contaminated and embrittled the steel. Darby, 1709, burnt coal to remove impurities and to transform the anthracite into coke, a

123

J Fail. Anal. and Preven. (2010) 10:8289 85

charcoal suitable for steelmaking [8, 9]. This change in fuel, from vegetal carbon to mineral carbon (coke), is one of the main factors that promoted the Industrial Revolution, which began about 1750 in England.

Puddling iron remained the common steel making process during the second half of the XVIII century, in Europe. The puddling furnace had two small compartments separated by a wall: one where the fuel was burned and the other, the hearth, where the metallic load was melted. Workers divided the metallic mass from a single charge into ve or six balls that were put in the maximum temperature region, for welding and separating the slag. After the separation, the balls were taken to the power hammer and hammered till the iron particles joined, eliminating a large part of the slag that was still in liquid or paste state. The hammered or forged balls were rolled to bars. Rolled bars were placed next to each other and hot rolled together, to produce one welded material. In this way, particles of slag were elongated and distributed in the steel [810].

Bearing in mind that Mexico, at that time, was under the rule of Spain, it is probable that the bar was from Spain but, with the enthusiasm of the new process, the bar could be brought from England. Vicua, in 1874, mentions that even in Spain, the English steel was preferred over the Spanish steel [11].

The possibility of forging the bar in Mexico, using imported iron must also be considered. Buchwald and Wivel [12] state that the larger objects had to be built up

by forge welding many small bar by the so-called piling processing. In the archives of the Cathedral [2], the amount paid to the blacksmith workshop of Mr. Juan Pedro Vallesteros, for forging the bars that would help to hold the two stone globes and the pedestals is recorded. This demonstrates that denitively there was a forge in New Spain (Mexico) with the capability of making iron bars from steel brought from abroad.

Metallurgical Analysis and Discussion

The purpose of this analysis was to identify the iron type, its structures, the probable making process and to estimate, if possible, the source of the crack in the bar.

Macroscopical Observation of the Bar and the Sample

Figures 5 and 6 show the 5 m bar. It is barely oxidized. The bar diameter is 55.3 cm while the head of the bar is enlarged in diameter and attened to about 11 cm diameter and 3 cm thickness. Due to the length and shape of the bar, it could have been fabricated rst by hot rolling the puddled steel and then by forging those small rolled bars to the required size.

A small segment, near the end of the bar, was cut for metallographic analysis. The amount of sample removed was minimal because of the historical signicance of the bar. The sample is rounded, 0.50.8 cm thick and 55.3 cm diameter. The surface of the bar is rough, typical of a struck surface or forged surface. The sample is from the terminal part of a long bar and because of this it may not have a proper nish appearance. A small part is missing in the center of the sample, as if an instrument, perhaps a striking object, had been used to remove a section and left a hole (Fig. 8). The opposite side of the sample, the cut surface,

Fig. 7 Photograph of the page indicating that the iron arrived to theVeracruz Port in Mexico Fig. 8 Outer extreme segment of the bar

123

86 J Fail. Anal. and Preven. (2010) 10:8289

transversal to the bar, exhibits a crack near the center and the crack seems to spread in several branches (Fig. 9).

Chemical Analysis

Chemical analysis was done by an Emission Spectroscopical apparatus; Spectro M8. The porosity in the surface caused some inaccuracy. The average analysis is shown in Table 1. A full chemical analysis, using drilled chips in different parts, was not allowed in this precious archeological sample.

For Buchwald and Wivel [12], it appears that puddled irons are essentially Si and Mn free (\0.04%) and have a P content around 0.204%. Moreno Revilla [13] report that puddled irons have, on average, up to 0.47% P and 0.20% Si and forged steels have lower contents 0.03% Si and0.08% P. Although Moreno Revilla and Butchwald indicate different P and Si content, it may be signicant that Mo-reno Revilla species a maximum rather than a typical analysis. The Si and Mn contents in the sample appear to be consistent with that expected in puddled iron. It must be remembered that the primary iron bars were forged and that the slag in the forging operation may have affected the original analysis.

Hardness Testing

Rockwell B hardness was determined in two different regions, due to the uneven surface (Fig. 10). The values obtained are consistent with a very heterogeneous microstructure with

regions of primarily ferrite and other regions of ferrite and pearlite. Inclusions were also present in the steel.

Metallographic, Stereographic, and SEM Analysis

The sample was prepared for optical and SEM observations in the etched and unetched conditions. Some inclusions were microanalyzed with SEM. Although difcult, it was possible to prepare a small longitudinal zone to observe the structure in this direction.

As expected for an ancient iron sample and as indicated by the hardness measurements, the structure is heterogeneous. There are low and medium carbon zones which form wide ferrite regions that coexist with pearliteferrite zones. In Fig. 11, there is an example of some of the different structures found. In Fig. 12, there are two steel structures: ferrite and ferritepearlite. These structures are separated by a crack and suggest that the crack is simply a lack of union. There is

Fig. 9 Cut surface of the segment

Table 1 Average chemical composition

%C %Si %Mn %P %S %Cr %Mo

0.229 0.126 0.027 0.002 0.006 0.001 0.039

%Ni %Al %Co %Cu %Nb %Ti %Fe

0.008 0.074 0.0038 0.206 \0 0.006 99.2

Fig. 10 Semipolished showing RB surface

Fig. 11 Ferritepearlite at left, ferrite at right

123

J Fail. Anal. and Preven. (2010) 10:8289 87

Fig. 14 Silicate inclusions in ferriticpearlitic structure

Fig. 15 Silicate aligned in the rolling direction, in ferritic and pearlitic matrix

Fig. 16 Inclusions and pores showed mainly in ferrite

Fig. 12 A lack of union between ferrite and ferritepearlite

Fig. 13 Widsmanstatten ferrite with pearlite

also pearliticwidmanstatten ferritic structure, mostly in the edge surface of the sample (outer surface of the bar; Fig. 13). This microstructure is consistent with the higher rate of cooling of the outside of the bar as well as a rather fast cooling during the forging process. In these pearlitic ferritic regions, silicate-type inclusions are observed (Fig. 14), and are elongated in the longitudinal direction (Fig. 15). The presence of elongated inclusions point toward a rolling direction. They provide a clear indication that a rolling process was used as part of the puddling. According to Buchwald and Wivel, in the forging process [12], the absence of iron oxide slag inclusions is consistent with a process that used a strong reducing gas such as a CO/ CO2 mixture. The metal phase becomes carburized and the slag is poor in FeO.

123

88 J Fail. Anal. and Preven. (2010) 10:8289

Fig. 17 Slag inclusion showing FeO wustite in glassy matrix

Fig. 18 Slag inclusion showing FeO wustite in points 1 and 5 mostly FeO; point 2: MgO 1.81%, Al2O3 4.62%, glassy matrix SiO2 34.84%,

K2O2 0.66%, CaO 8.61%, MnO 5.03%, FeO 42.23%

Fig. 19 Structures localized on the sample of the bar

The ferritic structure with inclusions and porosities is dominant in the sample. Ferrite varies in grain size (Fig. 16). Many of the inclusions are FeO wustite slag in glassy phase (Figs. 17 and 18), as determined by the SEM and microanalysis. Following Buchwald and Wivel [12], this observation suggests that the forgers were working at low temperatures where the reducing power of CO/CO2 gas is limited. The slag is FeO rich and these wustite slag inclusions are usually encountered in ferrite.

Figure 19 shows the location point of the optical microstructures in the sample. It is evident that there were many different masses joined with diverse heating and cooling rates. The ferritic and pearlitic microstructures have different grain sizes and different inclusions topographies.

In the puddled process as well as in the Catalan forge different iron balls were joined. In the puddled, at the end, there was a rolling. While in Catalan forge, rolling is never mentioned. The silicate inclusions are aligned, thus the primary iron bar could have been made by puddling and rolling.

The nal bar was made by forging the puddled rolled primary bars by the Vallesteros blacksmith shop.

However, some cracks or crack-like defects were found in the bar. The wider crack is viewed between ferrite microstructures with large grain sizes (Figs. 16 and 20). It is meaningful that the crack coincides with the location of the hole on the outer side, where an object probably struck the rod, possibly during the manufacturing process. The bar was probably struck with an object from the outside and the weakest part (the ferrite) separated; because, although being ductile, it has low strength and the many inclusions and porosities observed in Fig. 19 enhanced this lack of mechanical resistance. The other cracks were simply crack like defects, which are present because of the lack of complete welding during the forge welding of the various components used to produce the bar.

123

J Fail. Anal. and Preven. (2010) 10:8289 89

Fig. 20 Stereographic view, 50x

Conclusions

The iron bar, from the XVIII Century, located in the East tower of the Cathedral of Mexico, was made of ancient iron fabricated in Europe. Most of the laboratory characterizations of the bar point to puddling as the primary iron production process. However, the size and shape of the bar suggest that the puddled iron was forged by a skilled smith, probably Vallesteros as suggested in the archive found in Mexico [2]. The different steel microstructures, ferrite and pearlite, contain slag inclusions typical of those usually found in forged iron. The cracks found originated from two sources: a lack of union between at least two different iron bars that were rolled and forged welded together and as the result of a strong strike to the bar which fractured the weakest part.

Acknowledgments The authors thank Gonzalez G. and Puente I. for their microscopy support and Snchez R. for Figs. 2b, 4, 5, 6.

References

1. Catedral de Mxico. In: Encyclopedia de Mxico, vol. 3, pp. 14311434. Sabeca Edicin especial para Encyclopaedia Britannica de Mxico, S. A. de C. V. (1993) (in Spanish)

2. Copia de las cuentas de la Fbrica de la fachada y torres de Catedral, Tomo II. Archivo del Cabildo Metropolitano de la Arquidicesis de Mxico Memorias semanarias de lo erogado desde 29 enero 1787 en que dio principio, hasta n de diciembre 1791 (in Spanish)

3. Toussaint, Manuel, La Catedral de Mxico y el Sagrario Metro-politano: su historia, su tesoro, su arte, Editorial Porrua 3a (ed.), pp. 6265 (1992) (in Spanish)

4. Encyclopaedia Britannica, USA, 15th ed., Macropaedia, Trafalgar Battle, vol X, p. 84 (1981)

5. Uribe, J.A.: Labor de Andrs Manuel del Ro en Mxico: Profesor en el Real Seminario de Minera e innovador tecnolgico en Minas y Ferreras, Asclepio. Revista de Historia de la Medicina y la Ciencia, vol. LVIII, no. 2, Julio-Diciembre, pp. 231260 (2006). ISSN: 0210-4466 (in Spanish)

6. Apraiz, J.: Fabricacin de hierro, acero y fundiciones, Tomo I (ed.). Urmo, S. A. de Ediciones, pp. 6172 (1978) (in Spanish)

7. Diez de Salazar, L.M.: La Industria del hierro en Guipuzcoa (Siglos XIII XIV) aportacin al estudio de la industria urbana, En la Espaa Medieval, no. 6, p. 252 (1985). ISNN 0214-3038 (in Spanish)

8. Tylecote, R.F.: A History of Metallurgy, 2nd ed., p. 97. The Institute of Materials, London (1992)

9. Cottrell, A.: An Introduction to Metallurgy, 2nd ed., p. 2, 63, 130. Edward Arnold, London (1975)

10. Life of Henry Cort, http://henrycort.net/01life.htmp1

Web End =http://henrycort.net/01life.htmp1 , taken from the net on August 3rd 2008

11. Vicua, G.: El hierro en Vizcaya, Revista Europea, Num. 44, December 27th, pp. 265272 (1874) (in Spanish)

12. Buchwald, V.F., Wivel, H.: Slag analysis as a method for the characterization and provenancing of ancient iron objects. Mater. Charact. 40, 7396 (1998)

13. Moreno Revilla, J.: PhD thesis, Integridad estructural de vigas roblonadas de acero estructural antiguo, Departamento de Ciencia de los Materiales, Universidad Politcnica de Madrid (2005) (in Spanish)

123