X-Ray Computed Tomography In Situ: An Opportunity for Museums and Restoration Laboratories

: X-ray Computed Tomography (X-ray CT) is a sophisticated non-destructive imaging technique to investigate structures and materials of complex objects, and its application can answer many conservation and restoration questions. However, for Cultural Heritage investigations, medical CT scanners are not optimized for many case-studies: These instruments are designed for the human body, are not ﬂexible and are di ﬃ cult to use in situ. To overcome these limitations and to safely investigate works of art on site—in a restoration laboratory or in a museum—the X-ray Tomography Laboratory of the University of Bologna designed several CT systems. Here we present two of these facilities and the results of important measurement campaigns performed in situ. The ﬁrst instrument, light and ﬂexible, is designed to investigate medium-size objects with a resolution of a few tens of microns and was used for the CT analysis of several Japanese theater masks belonging to the collection of the “L. Pigorini” Museum (Rome). The second is designed to analyze larger objects, up to 200 cm and was used to investigate the collection of the so-called “Statue Vestite” (devotional dressed statues) of the Diocesan Museum of Massa.


Introduction
X-ray Computed Tomography (X-ray CT) is a powerful non-invasive and non-destructive technique capable of revealing crucial information about the construction techniques, conservation state, and materials of works of art. Exploiting the high penetration power of X-rays, CT investigates the entire volume of the objects, revealing internal hidden details. Here we describe two of the transportable X-ray tomographic facilities especially designed and developed by our research group to investigate Cultural Heritage objects. Additionally, the capabilities of these instruments will be presented through the important results of two recent measurement campaigns: the investigation of a Japanese Theater mask collection of the L. Pigorini Museum in Rome (currently Museum of Civilities), and the analysis of the so-called Statue Vestite collection (devotional dressed statues) of the Diocesan Museum in Massa (Massa, Tuscany, Center Italy).
Nowadays, a large number of scientific investigations of works of art can be achieved in situ, avoiding risky transports. Examples are the imaging of superficial layers, being varnish, pigments or preparatory layers, using techniques, such as infrared reflectography [1], ultraviolet fluorescence [2] or The mask is then painted on the back with a dark protective layer of Urushi lacquer (made from the sap of Japanese Urushi tree) and on the front with many thin layers of Gofun, a fine white powder made from ground oyster shells, mixed with Nikawa, a natural glue. The finished painting is made with traditional colors: red, made from natural minerals, yellow from natural soil, and black, using Chinese ink [25]. Occasionally, some masks could have movable parts with metal inserts, such as eyes or lower jaw, to mimic and highlight facial expressions.
The MATEGIAPPI project aimed to conduct a comprehensive scientific investigation of the L. Pigorini Japanese mask collection: from the documentation to the material characterization, from the investigation of possible previous restorations to the study of the conservative state of the objects.
Within this framework and to investigate in detail the construction technique of the Pigorini's mask collection, an X-ray CT investigation has been performed in situ. In Figure 1a, the 3D CT rendering of all the acquired masks is shown. In the next paragraph, two of them will be analyzed in detail: a "Karasu Tengu" mask (Kagura rituals, XIX century) and a "Ran-Ryô" mask (Bugaku court dances, unknown date).
The second measurement campaign was performed on part of the collection of Statue Vestite (dressed statues) of the Diocesan Museum of Massa.
These Italian artifacts of medieval origin are usually located in churches or private chapels as venerated objects and are still nowadays carried in procession during the Christian holidays. As the name suggests, these wooden statues are dressed in rich and delicate costumes, donated by noble ladies as votive offerings. Being covered, the wooden bodies of the Statue Vestite are only outlined, often without any anatomical detail. Contrariwise, the uncovered parts, such as visage, chest, arms, and feet, are carefully modeled in plaster and finely painted. An interesting detail of this collection is the use of the so-called "Stucco di Lucca" to model these details. This particular stucco is a mixture of plaster, papier-mache, and other various materials [26][27][28][29][30].
Despite the wide dispersion of these artifacts, even outside of Tuscany, and their historical interest and the peculiar conservation and restoration problems related to material-mix used to mold them, these statues have only been investigated episodically and rarely with a scientific approach [27,30].
The STA.Ve project focused on a comprehensive examination of these artifacts: from the chemical analysis of the Stucco di Lucca to the investigation of its complex stratification and, more generally, to the evaluation of the conservation state of the statues.
Within this framework and to investigate the construction techniques of these peculiar objects, an X-ray CT investigation on a selection of five statues was performed in situ. Figure 1b shows the 3D tomographic rendering of the investigated statues. In the next paragraph, we will analyze in detail the X-ray CT results of one of them, the "Madonna del Rosario" statue.  The second measurement campaign was performed on part of the collection of Statue Vestite (dressed statues) of the Diocesan Museum of Massa.
These Italian artifacts of medieval origin are usually located in churches or private chapels as venerated objects and are still nowadays carried in procession during the Christian holidays. As the name suggests, these wooden statues are dressed in rich and delicate costumes, donated by noble ladies as votive offerings. Being covered, the wooden bodies of the Statue Vestite are only outlined, often without any anatomical detail. Contrariwise, the uncovered parts, such as visage, chest, arms, and feet, are carefully modeled in plaster and finely painted. An interesting detail of this collection is the use of the so-called "Stucco di Lucca" to model these details. This particular stucco is a mixture of plaster, papier-mache, and other various materials [26][27][28][29][30].
Despite the wide dispersion of these artifacts, even outside of Tuscany, and their historical interest and the peculiar conservation and restoration problems related to material-mix used to mold them, these statues have only been investigated episodically and rarely with a scientific approach [27,30].
The STA.Ve project focused on a comprehensive examination of these artifacts: from the chemical analysis of the Stucco di Lucca to the investigation of its complex stratification and, more generally, to the evaluation of the conservation state of the statues.
Within this framework and to investigate the construction techniques of these peculiar objects, an X-ray CT investigation on a selection of five statues was performed in situ. Figure 1b shows the 3D tomographic rendering of the investigated statues. In the next paragraph, we will analyze in detail the X-ray CT results of one of them, the "Madonna del Rosario" statue.

Materials and Methods
Considering the substantial difference between the two collections, the analysis was performed in situ with two different systems, illustrated in Figure 2.

Karasu Tengu-Kagura Rituals-XIX Century
In Japanese traditions, Tengu are spirits generally represented as bird-men with the characteristic prominent beak. Due to its ambiguous morality, they could be represented both as evil and as good spirits. Different kinds of Tengu exist, and the Karasu Tengu is one of the less powerful forms of them, and are generally ascetics or monks living in remote mountain regions and forests.
The Karasu Tengu mask here presented ( Figure 3a) is a traditional carved wooden mask, without metal inserts and with the conventional prominent beak. As shown in Figure 3b, the 3D tomographic rendering confirms the presence of a quite homogeneous preparatory layer. Nevertheless, the tomographic slices revealed several unexpected details. As shown in Figure 3c, the mask is shaped with two wooden blocks, apparently of the same wood type: one for the main part of the visage and one for the beak. The two blocks are positioned with the fiber direction perpendicular one to the other and joined using four wooden nails and a glue layer. Another interesting detail is the use of wooden blocks with defects or potentially fragile areas. Indeed, as shown in Figure 3d, the area of the forehead is characterized by a large wood-knot. Moreover, the wood block chosen to shape the beak is the central part of the trunk, the most fragile due to the presence of mallow. To prevent its degradation, as generally done in Western sculptures, the mallow has been removed, and the hole has been covered with a wooden stopper (Figure 3c).  Figure 2a shows the system used for the Japanese mask investigation, designed for medium-size objects. Thanks to the small dimensions of the system, it can be easily placed on a desk and thanks to its components, the system is optimal to perform a highly-resolved investigation of not too high-density materials, such as the wooden Japanese mask here described. Indeed, the facility enables high-resolution 3D imaging, up to 50 µm of voxel size, of objects up to 50 by 50 cm.
The system is equipped with a microfocus X-ray tube (Kevex mod. PXS10, maximum voltage of 130 KV and maximum current of 0.5 mA), a large-surface indirect conversion flat panel detector (Varian PS2520 D with a CsI:TI scintillator, 1536 × 1920 pixel and 127 µm pixel size) and a rotational stage (Physik Instrumente PI, mod. M-038.PD1). To perform the tile-scan and reach a 50 by 50 cm 2 field of view, the detector is placed on an X-Y stage system (linear stages PI, mod. M-413.3PD) and the X-ray tube on a Z-stage (PI, mod. M-521.DD). Considering the object dimensions and the resolution desired for these investigations, the source-object distance was fixed at 800 mm and the source-detector distance at 1030 mm. In this configuration, it was possible to perform the tile-scanning of the masks in a small number of frames: only two for the "Karasu Tengu" mask (dimensions: 22 × 18 × 17 cm) and four for the Ran-Ryô mask (dimensions: 33 × 29 × 17 cm). Both masks were analyzed using 100 kV, 160 µA, 2 fps (frame per second) and 900 projections over 360 • , obtaining a voxel size of 100 µm.
In Figure 2b, the system designed for large-size objects is shown. Thanks to its flexibility, combined with the two broad translational stages and the fast acquisition process-each tomographic frame requires only 3 minutes-the facility is optimal to perform CT of large objects, such as the Statue Vestite collection here illustrated. Indeed, the facility enables medium high-resolution 3D imaging, up to 100 µm voxel size, of objects up to 150 by 150 cm.
The system is equipped with an X-ray tube (Yxlon, Smart EVO 200D, 30-200 kV and maximum current of 6 mA), a fast flat panel detector (Hamamatsu C10900D with a CsI:TI scintillator, 1216 × 1232 pixel and 100 µm pixel size) and a heavy load rotational stage (stage PI, mod. PRS200). To perform the tile-scan acquisition, the detector is placed on a custom X-Y stage system and the X-ray tube on a custom Z-stage. Considering the large object dimensions and the resolution required, for these analyses, the source-object distance was set at 253 mm and the source-detector distance at 2650 mm. Due to the dimensions of the statue (39 × 38 × 141 cm) several frames were necessary to acquire the entire object. The acquisition was performed using 120 kV, 2 mA, 5 fps, and 900 projections over 360 • . To obtain a voxel size of 167 µm, a binned acquisition mode (2 × 2) was used.
For maximum flexibility and transportability, both systems are designed without shielding. The radiation safety is guaranteed by a no-go zone defined based on the criteria established by a qualified expert in radiation protection. The X-ray CT investigations are performed in rooms not accessible to the public and, during the X-ray activities, the access to the area is interdicted to everyone (including authorized operators).
The tomography reconstruction is performed using the cone-beam algorithm of Feldkamp, Davis, and Kress (FDK) developed in house [31]. The use of high-performance computing (HPC) techniques allows the fast processing of large dataset as with the 65 GB of the statue here analyzed. In the case of tile-scanning, the frames are stitched taking into account the opportune geometrical parameters. The 3D rendering and the image display are then performed using the two open-source software Slicer and ImageJ.

Karasu Tengu-Kagura Rituals-XIX Century
In Japanese traditions, Tengu are spirits generally represented as bird-men with the characteristic prominent beak. Due to its ambiguous morality, they could be represented both as evil and as good spirits. Different kinds of Tengu exist, and the Karasu Tengu is one of the less powerful forms of them, and are generally ascetics or monks living in remote mountain regions and forests.

Karasu Tengu-Kagura Rituals-XIX Century
In Japanese traditions, Tengu are spirits generally represented as bird-men with the characteristic prominent beak. Due to its ambiguous morality, they could be represented both as evil and as good spirits. Different kinds of Tengu exist, and the Karasu Tengu is one of the less powerful forms of them, and are generally ascetics or monks living in remote mountain regions and forests.
The Karasu Tengu mask here presented (Figure 3a) is a traditional carved wooden mask, without metal inserts and with the conventional prominent beak. As shown in Figure 3b, the 3D tomographic rendering confirms the presence of a quite homogeneous preparatory layer. Nevertheless, the tomographic slices revealed several unexpected details. As shown in Figure 3c, the mask is shaped with two wooden blocks, apparently of the same wood type: one for the main part of the visage and one for the beak. The two blocks are positioned with the fiber direction perpendicular one to the other and joined using four wooden nails and a glue layer. Another interesting detail is the use of wooden blocks with defects or potentially fragile areas. Indeed, as shown in Figure 3d, the area of the forehead is characterized by a large wood-knot. Moreover, the wood block chosen to shape the beak is the central part of the trunk, the most fragile due to the presence of mallow. To prevent its degradation, as generally done in Western sculptures, the mallow has been removed, and the hole has been covered with a wooden stopper (Figure 3c).

Ran Ryô-Bugaku Dances
The Ran-Ryô is one of the main characters of the traditional Japanese court dances called Bugaku. There are several interpretations of this dance. The most common explains it as a celebration of battle victories of the Chinese Prince Lanling (Ranryoo) [32]. According to the legend, he was an intelligent and brave warrior but so handsome that he could not inspire terror in his enemies. For this reason, he had to wear a scary mask to scare his enemies.
The Ran-Ryô mask is generally wooden carved with movable parts, such as lower jaw and, as in this case, movable and metallic eyes as shown in Figure 4a.
Analogously to the Karasu-Tengu mask, the CT investigation reveals a quite homogeneous preparatory layer. As shown in Figure 4b, exceptions are the red areas of the dragon wings, related to the use of absorbing pigments or to a higher thickness of the ground layer.
In addition, for the Ran-Ryô, the nose is realized with a wooden block positioned with the wood fibers perpendicular to those of the visage, as shown in Figure 4d.
Despite these similarities, the analysis of the tomography slices reveals a completely different internal structure. Indeed, contrary to the two main blocks of the Karasu-Tengu, the mask is The Karasu Tengu mask here presented (Figure 3a) is a traditional carved wooden mask, without metal inserts and with the conventional prominent beak. As shown in Figure 3b, the 3D tomographic rendering confirms the presence of a quite homogeneous preparatory layer. Nevertheless, the tomographic slices revealed several unexpected details. As shown in Figure 3c, the mask is shaped with two wooden blocks, apparently of the same wood type: one for the main part of the visage and one for the beak. The two blocks are positioned with the fiber direction perpendicular one to the other and joined using four wooden nails and a glue layer. Another interesting detail is the use of wooden blocks with defects or potentially fragile areas. Indeed, as shown in Figure 3d, the area of the forehead is characterized by a large wood-knot. Moreover, the wood block chosen to shape the beak is the central part of the trunk, the most fragile due to the presence of mallow. To prevent its degradation, as generally done in Western sculptures, the mallow has been removed, and the hole has been covered with a wooden stopper (Figure 3c).

Ran Ryô-Bugaku Dances
The Ran-Ryô is one of the main characters of the traditional Japanese court dances called Bugaku. There are several interpretations of this dance. The most common explains it as a celebration of battle victories of the Chinese Prince Lanling (Ranryoo) [32]. According to the legend, he was an intelligent and brave warrior but so handsome that he could not inspire terror in his enemies. For this reason, he had to wear a scary mask to scare his enemies.
The Ran-Ryô mask is generally wooden carved with movable parts, such as lower jaw and, as in this case, movable and metallic eyes as shown in Figure 4a.

Madonna del Rosario, XVII Century
The statue, generally located in the Church of S. Remigio (Fosdinovo, Massa, Italy), is a typical example of devotional dressed statues. The main body is carved in wood while the upper partshead, shoulders, and chest-are molded in Stucco di Lucca, apparently made with a mixture of materials, such as canvas, plaster, and papier-mache.
The incomplete documentation about the statue's origin [26], combined with the clear evidence of several interventions, give rise to a peculiar local tradition that identifies the Madonna del Rosario as a re-adaptation of a more ancient angel statue. In particular, an Annunciation's Angel that, together with a Madonna, constituted a XV century wooden pair of statues named Madonna dell'Annunziata. Several historical documents support this hypothesis, but while it is possible to find documentation Analogously to the Karasu-Tengu mask, the CT investigation reveals a quite homogeneous preparatory layer. As shown in Figure 4b, exceptions are the red areas of the dragon wings, related to the use of absorbing pigments or to a higher thickness of the ground layer.
In addition, for the Ran-Ryô, the nose is realized with a wooden block positioned with the wood fibers perpendicular to those of the visage, as shown in Figure 4d.
Despite these similarities, the analysis of the tomography slices reveals a completely different internal structure. Indeed, contrary to the two main blocks of the Karasu-Tengu, the mask is composed of a multitude of small wooden parts, held together thanks to the use of wooden joints and glues, as evident in the construction of the wings and of the Dragon (Figure 4c,d). Moreover, metal nails have been used to reinforce the junction of the Dragon with the main part of the mask (Figure 4d).

The Statue Vestite CT Analysis-Diocesan Museum of Massa
Madonna del Rosario, XVII Century The statue, generally located in the Church of S. Remigio (Fosdinovo, Massa, Italy), is a typical example of devotional dressed statues. The main body is carved in wood while the upper parts-head, shoulders, and chest-are molded in Stucco di Lucca, apparently made with a mixture of materials, such as canvas, plaster, and papier-mache.
The incomplete documentation about the statue's origin [26], combined with the clear evidence of several interventions, give rise to a peculiar local tradition that identifies the Madonna del Rosario as a re-adaptation of a more ancient angel statue. In particular, an Annunciation's Angel that, together with a Madonna, constituted a XV century wooden pair of statues named Madonna dell'Annunziata. Several historical documents support this hypothesis, but while it is possible to find documentation regarding the decades following for the Madonna's creation, all traces of the Angel were lost after the year 1628. Moreover, there is no information about The Madonna del Rosario's origin and the first indirect mention of the object was found on a receipt of rich and detailed clothes for a statue dated 1676. Stronger indications of the modification can be found in the structure of the statue itself, where several modifications appear evident, as described in detail in the following paragraphs.
In addition to that (supposed) main change, verified historical documentation reports several interventions on the statue. The first one, commissioned in 1692, consists of just in a new silver vest. The second and the third ones, both in 1693, were more invasive adducing wide and important alterations of the wooden parts of the statue, in particular on the hips area and probably adding the two wooden bars still present today. Presumably, the intervention was aimed to enlarge the hips to adapt the statue silhouette to what was in vogue at the end of the XVII century. Another important modification occurred in 1743, when the statue was completely repainted and, in the same year, three metal screws and two metal bars were added to the arms and to the wooden platform.
The CT investigation revealed the extensive use of plaster, primarily as a preparation layer and as filler, the latter most probably related to the several interventions mentioned above ( Figure 5b). As emphasized by restorers, the Stucco di Lucca parts, such as the head and hands, are removable and the tomography highlighted how these parts fit with the main wooden body (Figure 6a,b). The second and the third ones, both in 1693, were more invasive adducing wide and important alterations of the wooden parts of the statue, in particular on the hips area and probably adding the two wooden bars still present today. Presumably, the intervention was aimed to enlarge the hips to adapt the statue silhouette to what was in vogue at the end of the XVII century. Another important modification occurred in 1743, when the statue was completely repainted and, in the same year, three metal screws and two metal bars were added to the arms and to the wooden platform.
(a) (b) The CT investigation revealed the extensive use of plaster, primarily as a preparation layer and as filler, the latter most probably related to the several interventions mentioned above ( Figure 5b). As emphasized by restorers, the Stucco di Lucca parts, such as the head and hands, are removable and the tomography highlighted how these parts fit with the main wooden body (Figure 6a,b).
The X-ray investigation of the statue skirt (Figure 7a) clearly revealed the modifications and three main overlays are recognizable: the main, and presumably original, body (in red) has been cut and refined to create a flat base for a second layer (in green). The current statue shape has been achieved with the same technique, and a supplementary layer (in blue) has been added. The analysis at different heights revealed the use of massive metal nails to connect the several parts ( Figure 7b).
Moreover, X-rays reveal in each layer the presence of a finishing plaster coating, clearly intended to be visible, but currently covered.  The X-ray investigation of the statue skirt (Figure 7a) clearly revealed the modifications and three main overlays are recognizable: the main, and presumably original, body (in red) has been cut and refined to create a flat base for a second layer (in green). The current statue shape has been achieved with the same technique, and a supplementary layer (in blue) has been added. The analysis at different heights revealed the use of massive metal nails to connect the several parts (Figure 7b).

Conclusions
The investigation campaigns performed proved the possibility of totally safe in situ X-ray tomography investigations. Indeed, the design of flexible and transportable facilities, optimized for Cultural Heritage investigations, allow to perform studies on a wide range of objects, from medium size wooden Japanese masks of the Pigorini Museum in Rome to the (almost) natural-size Statue Vestite of the Diocesan Museum in Massa. Moreover, the joined competences of the restorers and of the scientists, allow to discover details about the construction techniques of less-known objects such as the Karasu Tengu mask or the peculiar restoration/modification history of the Madonna del Rosario.
Finally, these experiences could be great opportunities for outreach. Excellent examples are the two events organized by the Diocesan Museum of Massa for the general public: a temporary exposition titled "Statue vestite, lavori in corso: dalla devozione alla ricerca scientifica"(Dressed Statues, work in progress: from devotion to scientific research) and a scientific conference. For both, the response of the public was above expectations. Moreover, X-rays reveal in each layer the presence of a finishing plaster coating, clearly intended to be visible, but currently covered.

Conclusions
The investigation campaigns performed proved the possibility of totally safe in situ X-ray tomography investigations. Indeed, the design of flexible and transportable facilities, optimized for Cultural Heritage investigations, allow to perform studies on a wide range of objects, from medium size wooden Japanese masks of the Pigorini Museum in Rome to the (almost) natural-size Statue Vestite of the Diocesan Museum in Massa. Moreover, the joined competences of the restorers and of the scientists, allow to discover details about the construction techniques of less-known objects such as the Karasu Tengu mask or the peculiar restoration/modification history of the Madonna del Rosario.
Finally, these experiences could be great opportunities for outreach. Excellent examples are the two events organized by the Diocesan Museum of Massa for the general public: a temporary exposition titled "Statue vestite, lavori in corso: dalla devozione alla ricerca scientifica"(Dressed Statues, work in progress: from devotion to scientific research) and a scientific conference. For both, the response of the public was above expectations.