Material Marvels: Cattedrale di Santa Maria del Fiore

Materials World magazine
,
1 Sep 2017

Kathryn Allen looks at the mysteries of engineering behind the dome of the Cattedrale di Santa Maria del Fiore in Italy.

Construction began on the Cattedrale di Santa Maria del Fiore, Italy, – also known as Il Duomo di Firenze – in 1296, but by 1418 it was still incomplete. The cathedral’s original architect, Arnolfo di Cambio, died in 1310 before finishing his project, which was overseen by various architects until its completion – minus the dome – in 1418. 

Di Cambio left no indication of how the dome was to be constructed and no one knew how to build a dome 45.5m wide, sitting atop the cathedral walls at a height of around 55m on an imprecise octagonal base – the difference between the longest and shortest sides being 0.57m. An architecture competition was held in 1418 to find the ideal design. With a prize of 200 gold florins, architects attempted to come up with a design for what is now the largest masonry dome in the world.

Brunelleschi’s dome

In 1420, after much controversy, the design of goldsmith Filippo Brunelleschi was accepted. His design consisted of building two concentric ogival shells with a space in between for a walkway and stairs ascending to the top of the dome. Florentine artist Lorenzo Ghiberti, creator of the bronze doors of the Florence Baptistery – a project for which Brunelleschi competed – was appointed to the project, as well as Battista d’Antonio, alongside Brunelleschi. 

In an unprecedented move, forced by its sheer size, Brunelleschi planned to construct the dome without a freestanding supporting structure. It would have been a huge undertaking, in terms of both time and cost, to build a wooden structure large enough to support the emerging dome. However, Brunelleschi never revealed exactly how he constructed the dome without such a structure.

The first issue of construction was lifting the materials to the top of the cathedral, where the dome would sit, as existing mechanisms were not capable of carrying such weights at height and would require oxen to be unharnessed and re-harnessed to then lower materials. Brunelleschi engineered a three-speed hoist, driven by oxen turning a wooden tiller, which was capable of lifting and lowering materials without re-harnessing. Brunelleschi also engineered the castello – a crane, of just under 20m, used to move materials laterally once they had been lifted.

Four million bricks

The inner shell of Brunelleschi’s dome, more than two metres thick, is self-supporting. Made of brick in a herringbone pattern, this dome is sheltered by the outer dome, which is clad in terracotta tiles. An estimated four million bricks make up the dome, which sits on a sandstone base. The herringbone pattern, in which bricks placed vertically intersect horizontal rows of brick, forming a zig-zag pattern, strengthens the wall. This pattern avoids the planes of weakness caused by layers of mortar in horizontally laid brick and bonds new layers of brick to previous layers. The spines that form throughout the brickwork – lending it its name of spinapesce in Italian – travel round the dome, spreading the stress. It is this pattern that is generally credited, along with the use of rings of stone and wood that interlock with the dome’s arches, as providing support to the dome and allowing it to be constructed without scaffolding.

Intent on working out exactly how the dome was built, architect Massimo Ricci built a one-fifth-scale model of Brunelleschi’s dome in Parco dell’Anconella, Florence. Construction of Ricci’s model, which uses the herringbone pattern and measures 11m at its base, began in 1989. A supporting structure was not used in the model’s construction, however Ricci did use the partial centring pine templates he believed Brunelleschi used. These eight templates, one for each corner, covered only a minor portion of the dome’s curve and could be re-positioned as the dome grew higher.

The results of Ricci’s experimental model are explained in a paper he co-authored with Barry Jones and Andreas Sereni, Building Brunelleschi’s Dome: a practical methodology verified by experiment, published in Journal of the Society of Architectural Historians. The paper notes that the two shells of Brunelleschi’s dome are joined together by 24 ribs, eight of which rise from the corners of the octagonal base, with two minor ribs positioned between each, forming a series of gothic arches. The paper proposes that, since the dome did not have an exact centre, Brunelleschi likely used ropes to position the ribs and to work out the curvature of the brickwork so that it was uniform and met at the top of the dome. The curvature of the gothic arches was to be built to the quinto acuto, or the pointed fifth, which means that the radius of the arch is the same length as four-fifths of the dome’s base diameter.

The radial method

On his model, Ricci replicated the method he believed Brunelleschi used. Diagonal lines are drawn between opposite corners of the dome. The distance from opposite corners is divided by five, and the fourth point is used as a centre point to calculate the curvature of the dome on the opposite side. Between consecutive centre points an arc can be drawn – all of the points along which are equidistant to the corresponding point on the opposite side of the dome. This arc is then used, along with the two centre points it joins and the corner ribs of the dome, to position the bricks at the correct angle to form a uniform curvature. Ropes mark these lines.

Ricci, Jones and Sereni also propose that the herringbone brickwork does not serve only a supporting role, but that the vertical bricks, forming spines throughout the brickwork, offer guidance on where to position the next layer of bricks and at what angle. They argue that if the herringbone pattern were merely supportive, Brunelleschi would only have used it for the top of the dome, where the angle is severe. The paper claims Brunelleschi was already using the time-consuming herringbone pattern at the height of the second walkway, where the incline is not nearly great enough to require it, indicating it had another purpose. 

The level of the brickwork, which begins horizontal but becomes visibly concave nearing the top of the dome, is also cited by Ricci, Jones and Sereni as a supporting feature. However, they do not believe it was intentional, forming from Brunelleschi’s use of the radial method and ropes to position the bricks. This concave shape does, however, stabilise the wall, compressing the brick courses. Other theories suggest Brunelleschi knew this concave shape would be stronger, observing catenary curves in existing architecture. The paper’s authors claim that their theory of the dome’s radial construction may satisfy all the known facts, as previous theories have not. However, it notes that Brunelleschi was deliberately secretive about his methods, claiming that the brickwork now visible in the second and third walkways and the stairway between them is cosmetic, constructed after the dome was completed to confuse viewers. 

This paper offers one theory among many that attempt to explain the dome’s construction. However, many of these theories are based on the idea that Brunelleschi used guide ropes. Giacomo Tempesta, Associate Professor of Structural Design in the Department of Architecture at the University of Florence, Italy, told Materials World, ‘The use of strings or rope lines was essential to define the exact position of the bricks [in Brunelleschi's dome] and, at the same time, the slope of their laying, depending on the height of the rings.’

In the 1970s, Rowland Mainstone proposed that a series of cones, decreasing in size from the dome’s base to its top, indicated by guide ropes – with the tapered point facing down and the axis aligned with the dome’s centre – were used to indicate the curve of the shell.  This theory, and its assumption that Brunelleschi treated the octagonal dome as if it were circular has been disputed, not least by Ricci. In Mainstone’s paper, Brunelleschi’s Dome of S. Maria del Fiore and some related structures, he agrees with the theory that the herringbone brickwork allowed the dome to be constructed without supporting centring and that the centres used only outlined the specified curve that the bricks needed to follow. Mainstone claims the herringbone brickwork would only have needed temporary support while the mortar set. 

Robert Arnold, National Director of Education for the International Masonry Training and Education Foundation, the training arm of the International Union of Bricklayers and Allied Craftworkers, was part of a team of four masonry instructors who travelled to Florence to work on Ricci’s model in 2013. Arnold told Materials World, ‘In a normal brick dome pattern, the brick would be laid course-by-course creating a possible shear point in every course. By introducing the herringbone pattern the shear points are non-existent. The herringbone brick pattern also ties the brick to previous and future courses creating a bond that would not be created with normal brick coursing.’ 

Explaining the stress that the dome is under, Arnold said, ‘Hoop stress, especially at the lower portion of the dome, increases as the dome is being constructed. Tensile forces that are pushing out grow stronger as the dome is built. Some type of tension ring and/or a massive abutment at the base of the dome typically resolve these forces.’ It is known that the base of the dome is sandstone and that Brunelleschi planned to bind the walls with rings of wood and iron.

An engineering miracle

Pope Eugene IV consecrated the cathedral on 25 March 1436 and Brunelleschi’s dome was sealed later that year. The dome’s crowning lantern, although designed by Brunelleschi, was built following his death in 1446. The gilt copper sphere and cross on top of the lantern, designed by Andrea del Verrocchio, were positioned in the late 1460s. Further work continued on the dome, with Baccio d’Agnolo commissioned to build a marble gallery at its base, but d’Agnolo only completed one of the eight sides before work was stopped when Michelangelo supposedly compared it to a ‘cricket cage’. The gallery remains incomplete. The fresco of the Last Judgement, visible on the interior of the inner dome, was begun in 1572 by Giorgio Vasari and, following Vasari’s death in 1574, completed by Federico Zuccari.

Brunelleschi was buried in the crypt of Florence Cathedral and with him the definitive answer to how he constructed his dome. The cathedral remains a popular tourist attraction and its dome is highly acclaimed. Arnold said, ‘I would consider Florence Cathedral’s dome to be an engineering miracle. This dome was designed by a goldsmith in the early 1400s and was built before the understanding of calculating stress. Over four million bricks were installed without a centring system and the fact that the dome still stands today, almost 600 years later, in my opinion is a miracle.’

To read Building Brunelleschi’s Dome: a practical methodology verified by experiment, visit bit.ly/2tmFwVC

To read Brunelleschi’s Dome of S. Maria del Fiore and some related structures, visit bit.ly/2uquVbW