Petrophysical Characterisation and Suitability of Serpentinites from the Monteferrato Area (Tuscany, Italy) for Architectural Restoration

Abstract

This study investigates the mineralogical and physical properties of serpentinite from the Monteferrato area (Tuscany, Italy) to evaluate its potential use in Tuscany architectural restoration. The research addresses the need to identify replacement materials compatible with historic stones while preserving their original features. Representative specimens from the Bagnolo quarry were analysed through physical testing and a wide range of mineralogical and geochemical techniques, including polarised light microscopy, X-ray diffraction, electron probe micro-analysis, whole-rock chemistry, and fibre quantification. The results show a mineralogical composition dominated by serpentine-group minerals and magnetite, with physical properties generally consistent across samples. Measured capillary water absorption ranges from 3.27 to 5.27 g/m²·s^0.5, open porosity from 5.25% to 8.93%, apparent densities range from 2.49 to 2.56 g/cm³, and imbibition coefficient from 2.16% to 3.71%. Comparative analysis with serpentinite from historic sources (Figline di Prato quarry, Tuscany) and from monuments (Baptistery of San Giovanni, Florence) demonstrates close compositional and textural affinities, supporting the suitability of the rock from the studied quarry for restoration purposes in Tuscany monuments. However, chrysotile concentrations up to 14,153 mg/kg, exceeding Italian regulatory thresholds, represent a critical limitation. This not only requires the implementation of strict safety measures but also raises serious concerns regarding the practical feasibility of using this stone in conservation projects. More broadly, the presence of asbestiform minerals in serpentinites highlights a significant and often underestimated health risk associated with their extraction, processing, and use. Despite its importance, detailed fibre count data are rarely published or made publicly accessible, hindering both transparent risk assessment and informed decision-making. By integrating petrographic, mineralogical, and physical–mechanical characterisation with fibre quantification, this study not only assesses the technical suitability of Monteferrato serpentinites for restoration of Tuscan monuments but also contributes to a more responsible and evidence-based approach to their use, emphasising the urgent need for transparency and health protection in conservation practices.

1. Introduction

Natural stone has played a fundamental role in architectural heritage since antiquity, valued for its mechanical strength, durability, and aesthetic appeal. From the earliest civilisations, stone has been used to construct enduring monuments, temples, and infrastructure, establishing a legacy of resilience and permanence. Historically, builders relied on locally sourced stone due to the high cost and difficulty of transport, resulting in towns constructed primarily with a single material. This gave rise to distinctive cultural landscapes defined by regional geology. Today, these traditions are being revalued, as natural stone is increasingly recognised for its role in expressing local identity and supporting both community heritage and tourism [1,2]. Despite the widespread belief in the inherent durability of stone, all stones are ultimately subject to weathering and deterioration [2]. Once quarried and fashioned into building elements, natural stone becomes vulnerable to environmental conditions, construction techniques, and maintenance practices. Over time, the degradation of stone-built structures has become a critical concern in conservation and heritage management. Finding suitable replacement materials often requires searching for new quarries. Quarry landscapes are dynamic, often transitioning from active exploitation to abandonment. The cessation of quarrying activities may result from various factors: depletion of extractable material, hazardous properties of the stone, or its environmental impact. This is particularly relevant for older quarries located near built-up areas and, in some cases, even close to historic centres. These conditions have influenced territorial planning, leading to the cessation of ancient quarry exploitation and the redevelopment of abandoned quarry landscapes. Consequently, fostering a technological culture in the quarry industry that embraces ecological principles has become essential [3,4,5]. In addition, stricter environmental regulations in industrialised nations have led to the closure of many traditional quarries, limiting access to historically used materials. This shift has disrupted traditional restoration practices, contributed to the loss of regional architectural identity, and created a reliance on imported stones, often sourced from less regulated countries. While these materials are often more affordable, they rarely match the geological and aesthetic qualities of native stones, complicating conservation efforts. Understanding the long-term performance of stone in built environments requires an interdisciplinary approach. Contributions from geologists, materials scientists, architects, and conservation specialists are crucial in assessing stone durability and developing effective preservation strategies. The use of ornamental stones, such as serpentinite, granite, and marble, reflects both historical and cultural preferences, while also demanding scientific evaluation of their physical and mechanical properties. Non-destructive techniques have become vital in heritage conservation, enabling assessments that respect the integrity of historic structures. Within this broader context, this paper focuses on serpentinite sourced from the Bagnolo quarry (Monteferrato area, Tuscany), evaluating its intrinsic properties and its interaction with environmental factors. The search for serpentinite materials is driven by the need to find suitable replacements for deteriorated original stones. This is a widespread issue in many historic cities, for example Florence, where serpentinite was widely used in the cladding and decorative elements of numerous monuments (e.g., Cathedral of Santa Maria del Fiore and churches of Santa Croce, Santa Maria Novella, and San Miniato). This study analyses serpentinite from the Monteferrato area to evaluate its suitability for architectural restoration, with particular attention to its mineralogical features and chrysotile content. Despite its significant implications for public health, there is a notable lack of studies reporting detailed data on chrysotile in ornamental stones. Addressing this knowledge gap is crucial for both heritage conservation and the protection of public health.

2. Overview of Serpentinite

Petrologically, serpentinite is a fine-grained, dark green to black, metamorphic rock predominantly composed of serpentine-group minerals. It forms through low-temperature hydration of olivine-rich ultramafic rocks, resulting in the transformation of olivine, orthopyroxene, and clinopyroxene into hydrated magnesium silicates, magnetite, and occasionally brucite. This process reduces the density of the rock relative to its peridotitic protolith. Three main serpentine minerals, lizardite, chrysotile, and antigorite, are commonly present. Lizardite has platy crystals and frequently dominates the mesh textures, replacing olivine. Chrysotile is characterized by a cylindrical fiber morphology, which can reach lengths of several centimetres. These fibres, which may possess hollow or amorphous-filled cores, commonly fill secondary microfractures crosscutting the host serpentinite. While valued industrially for its insulating properties, chrysotile is also a recognised health hazard due to the inhalation risks associated with its fine, airborne fibres [6,7,8,9,10]. Antigorite, in contrast, forms at higher temperatures and is structurally distinct, with curved, modulated layers bonded by strong Si–O linkages. This gives it greater hardness, higher seismic velocities, and reduced cleavage, making it common in recrystallised and metamorphosed serpentinites [11,12].

Serpentinite has been used since ancient times, notably by the Egyptians, as both a building material and an ornamental stone [13,14]. Today, prominent commercial sources include Rajasthan Green and Emerald Green from India, often heavily altered to carbonates. In contrast, serpentinites from other regions, including Vermont Verde Antique (USA) and Verde Pirineo (Spain), preserve more of their primary mineralogy [15,16]. In Italy, serpentinite outcrops are widespread across both the Alps and the Apennines, with Tuscany representing a particularly important region. Occurrences are also found in the southern Apennines, especially in Calabria, although they were not traditionally employed as building stones.

In Tuscany, serpentinites associated with the Northern Apennine Ophiolitic Complex occur near Florence (Impruneta), Prato, Pistoia, Siena, and Volterra. The serpentinite examined in this study was sourced from a quarry (Bagnolo), located in the Monteferrato district, historically regarded as the primary source of the serpentinite employed in the Baptistery of San Giovanni, Florence [17].

3. Geological Overview of the Bagnolo Quarry

The Bagnolo quarry, located within the Monteferrato geological complex, near Montemurlo (Tuscany), forms part of the Northern Apennine orogenic system. This area is dominated by ultramafic and metamorphic rocks, with serpentinite as the prevailing lithology. These rocks originated through the hydration of peridotite within an ancient oceanic lithosphere and are associated with ophiolitic sequences emplaced during the Jurassic period [18,19]. Subsequent Alpine tectonic processes led to their deformation and uplift [20,21,22]. The Monteferrato serpentinites display massive to foliated textures. Their distinctive greenish color results from extensive serpentinization. The region represents a significant tectonic window, exposing fragments of the former Tethyan oceanic crust and offering valuable insights into the structure and evolution of ophiolite complexes in the Apennines.

Several historic quarries were exploited across the region, where serpentinite was locally known by names such as Verde di Prato, Nero di Prato, Marmo Nero, and Paragone [23,24,25,26,27,28,29]. This rock was widely used during the Romanesque and Renaissance periods in the construction of religious and civic buildings [17,29]. Among the earliest known extraction sites are the quarries at Piano di Gello, on the eastern slope of Monte Piccioli in the Monteferrato area. Active since at least the 11th century, these quarries are now abandoned but remain significant for understanding historic stoneworking practices.

The Bagnolo quarry (Figure 1), also known as Cava Guarino after its owner, is a historic serpentinite quarry of significant geological and cultural interest. The site is part of a broader area known for the presence of serpentinite, commonly referred to as “Prato green marble”, a metamorphic rock with distinctive green to dark bluish tones. The quarry has been exploited since the Middle Ages, supplying serpentinite blocks for use in notable architectural works throughout Tuscany. Today, the quarry is no longer active for industrial extraction, but it remains an important heritage site. Its surroundings, now partially reclaimed by vegetation, bear traces of historical quarrying activity and offer insight into traditional stoneworking techniques.

Sampling and Sample Preparation

The studied area is integrated into local trekking routes, such as the “Cave di Bagnolo trail”, which guides visitors through former extraction sites, natural amphitheatres, and panoramic viewpoints. These trails are part of a broader initiative to promote the sustainable use of geological heritage and to raise awareness of the region’s geo-cultural landscape. In addition to its geological relevance, Bagnolo quarry represents a valuable case study for the conservation and reuse of natural stone in architectural restoration, and serves as a living archive of Tuscany’s stone-working history. The stone material present in the quarry shows, when freshly cut, a general green colouration with varying textural characteristics, marked by the presence of granules and dark veins of various sizes. These features result in clear differences in macroscopic appearance. Therefore, the sampling was conducted in a way that would be representative of the different observed types. Five different rock portions were collected from different areas and were identified with the codes from CB1 to CB5. Before characterising the lithotypes, all rock samples were cleaned by removing degraded outer surfaces.

Two different sizes of test specimens (Figure 2) were prepared from each sample for physical testing. To assess reproducibility, two specimens per sample were tested. For mineralogical and geochemical characterisation, powder of each sample was obtained for X-ray diffraction (XRD) and whole rock analysis by inductively coupled plasma (ICP-MS). Selected samples were ground and sieved for fibre quantification. Polished thin sections of the rocks were also prepared for observation under a polarising microscope (PLM) and for mineral phase analysis using an electron probe micro-analyser (EPMA).

4. Analytical Methods

4.1. Physical Analyses

In studying the physical properties of stone materials used in cultural heritage, it is essential to understand how water moves within them. This requires determining open porosity, apparent density (also known as bulk density, calculated as the mass of the sample divided by its total volume, including both mineral grains and pore spaces), imbibition coefficient, and capillarity water absorption. These parameters were measured on Bagnolo samples in accordance with the European Standards [30,31,32]. For the capillary water absorption tests, two prisms measuring 4 cm × 4 cm × 4 cm were cut from each rock sample, while for the analyses of open porosity, apparent density, and imbibition coefficient, two smaller prisms of 2 × 2 × 2 cm were prepared (Figure 2).

4.2. Polarising Microscope Observations

Petrographic observations in polarised light were carried out using a Zeiss Axioscope A.1 microscope (Carl Zeiss, Jena, Germany), equipped with a 5 Megapixel resolution video camera and Zen lite 3.1 image analysis software. This technique enabled the characterisation of the texture and mineralogical composition of the samples [33].

4.3. X-Ray Powder Diffraction Analysis

Qualitative mineralogical analyses were performed by X-ray powder diffraction on powdered samples, using a Philips PW 1050/37 diffractometer (Philips, Almelo, The Netherlands), coupled with a Panalytical X’Pert PRO system and HighScore software version 2021 for data acquisition and interpretation. The detection limit of the method is 4%. The operating conditions were set at 40 kV and 20 mA, with a Cu anode, graphite monochromator, and a goniometer speed of 2°/min, over a scanning range of 5–70° 2θ [34,35,36].

4.4. Electron Probe Micro-Analysis

The chemical composition of the main mineral phases was obtained using a JEOL JXA-8600 electron microprobe (Jeol Ltd., Tokyo, Japan) equipped with four wavelength-dispersive spectrometers (WDS), under the following operating conditions: accelerating voltage of 15 kV, beam current of 10 nA, and beam diameter of 3–5 µm. Counting times were 10 s for Na, 15 s for the other major elements, and 40 s for minor elements (e.g., Mn, Sr, and Ba). Matrix effects were corrected using the PAP algorithm following the method of Pouchou and Pichoir [37]. Primary standards included natural and synthetic phases (e.g., Albite for Si/Na, Olivine for Mg, Diopside for Ca, Apatite for P, Ilmenite for Ti/Fe). Secondary standards from Smithsonian, MAC, and Astimex were analysed to verify accuracy. Long-term replicate measurements on secondary standards show precision <1% for Si, ≤2% for other major elements, and ≤5% for minor elements. Accuracy is within 1% for major and most minor elements. Detection limits are typically 100–300 ppm for minor elements, depending on the element and counting conditions.

4.5. ICP-MS Whole Rock Analysis

Whole-rock major and trace element composition was obtained at the Actlabs (Ancaster, ON, Canada), where samples were prepared using a batch system. Each batch included a method reagent blank, a certified reference material, and 17% replicates. Samples were mixed with a flux of lithium metaborate and lithium tetraborate and fused in an induction furnace. The resulting melt was immediately poured into a 5% nitric acid solution containing an internal standard and continuously stirred until completely dissolved (~30 min). Analyses were performed using a Thermo Jarrell-Ash ENVIRO II ICP in simultaneous/sequential mode. Calibration was carried out with seven certified reference materials (USGS and CANMET). One of the seven standards was analysed with every group of ten samples. The detection limit ranges between 0.01% and 0.001%.

4.6. Fibre Quantification

The fibre quantification was carried out at the laboratory of ECOLStudio/lifeanalytics (Vicenza, Italy). According to the Italian Ministerial Decree DM 06.09.1994, which sets regulatory limits and safety requirements for the use and handling of asbestos-containing materials, representative samples (0.5–1 g) were ground and sieved to <500 µm, then milled to 10–100 µm. An aliquot of the powdered material was suspended in a dispersant solution, filtered through a polycarbonate membrane (0.8 µm porosity, 25 mm diameter), and mounted on stubs for scanning electron microscopy (SEM) analysis. Fibres were counted only if they exhibited characteristic morphology and elemental composition (EDS) matching standard asbestos species (chrysotile, amosite, crocidolite, tremolite, actinolite, anthophyllite). The minimum count of 4 fibres defines the method detection limit (LOD ≈ 1.2 × 10⁴ fibres/mg, ~120 ppm), and the limit of quantification (LOQ) is 100 ppm. Fibre dimensions were measured and concentrations calculated assuming a density of 2.6 g/cm³ for chrysotile. Expanded measurement uncertainty (k = 2.95% confidence) is ~50%, considering contributions from sample weight, dilution, SEM magnification, fibre sampling, and repeatability. Method robustness was verified by varying milling times and assessing particle size distributions (10–100 µm), while repeatability was confirmed on samples of known concentrations (300, 1000, 3000 mg/kg) with ten replicates each.

5. Analytical Results

5.1. Physical Analyses

Table 1. Physical properties of serpentinite from Bagnolo (CB) and Figline di Prato (SF) quarries, and pavement of Florentine Baptistery (Bpt). SF and Bpt data are from Santo et al. [17].

Sample	Capillarity Water Absorption (g/m2·s0.5)	Sample	Open Porosity (%)	Apparent Density (g/cm3)	Imbibition Coefficient (%)
CB1_A	3.74	CB1_C	6.52	2.52	2.64
CB1_B	3.83	CB1_D	6.41	2.52	2.60
CB2_A	3.27	CB2_C	7.83	2.49	3.23
CB2_B	3.32	CB2_D	7.65	2.49	3.16
CB3_A	3.86	CB3_C	5.25	2.54	2.35
CB3_B	3.62	CB3_D	5.66	2.54	2.28
CB4_A	4.70	CB4_C	5.43	2.56	2.16
CB4_B	4.42	CB4_D	5.98	2.55	2.40
CB5_A	5.27	CB5_C	8.93	2.49	3.71
CB5_B	5.14	CB5_D	8.63	2.50	3.56
SF	3.46	SF	5.59	2.55	2.24
Bpt	3.18	Bpt	7.14	2.51	2.92

shows the results of the physical analyses obtained from replicate measurements for each block. The dataset highlights some differences in the physical behaviour of the Bagnolo serpentinite samples. Capillary water absorption varies from 3.27 g/m²·s^0.5 (CB2) to 5.27 g/m²·s^0.5 (CB5), with CB5 showing the fastest uptake, well above the mean (3.79 g/m²·s^0.5). Open porosity ranges between 5.25% (CB3) and 8.93% (CB5).

CB5 is the most porous variety, while CB3 displays lower open porosity, suggesting greater compactness. In general, these results indicate a relatively low to moderate capacity for water uptake, consistent with the petrophysical behaviour of this rock type. Apparent density is fairly uniform, oscillating between 2.49 and 2.56 g/cm³, reflecting a consistent mineralogical composition across the samples. Finally, the imbibition coefficient varies from 2.16% (CB4) to 3.71% (CB5). The lowest values (CB3 and CB4) are associated with the denser and less porous materials, while CB2 and especially CB5 show a greater capacity for water uptake. Overall, CB3 and CB4 emerge as the most compact and potentially more durable samples, whereas CB5 and, to a lesser extent, CB2, exhibit higher porosity and water absorption, features that may negatively affect long-term resistance and durability. The data obtained from the Bagnolo rock samples are consistent with values reported in the literature for serpentinite from various regions [38,39,40,41,42,43]. Table 1 also reports, for comparison, the results obtained on a serpentinite specimen (SF) from another historic quarry in the Monteferrato area (Figline di Prato quarry; Figure 1), as well as on the serpentinite (Bpt) from the pavement of the Baptistery of San Giovanni in Florence [17,29]. The physical properties of the SF and Bpt samples are consistent with those of the CB quarry samples. In fact, their values lie within the variability range observed for the CB series or are close to their mean values. The Bpt sample shows relatively high porosity and water absorption, comparable to the more porous CB samples (e.g., CB2 and CB5). Conversely, the SF sample displays lower porosity and imbibition, similar to the more compact CB samples (e.g., CB3 and CB4). This indicates that both SF and Bpt materials fall within the natural variability of serpentinite from the Bagnolo and Figline di Prato quarries, suggesting that the lithotypes used in the Baptistery pavement and those from the quarry sources share broadly comparable physical characteristics.

5.2. Minero-Petrographic Analyses

5.2.1. Polarising Microscope Observations

Under the polarising microscope (Figure 3a,d), the samples from the Bagnolo quarry exhibit overall similar textures, characterised by a medium-grained serpentine matrix crosscut by a dense network of thin, chrysotile-rich veinlets. These veinlets are irregularly oriented and intersect at various angles, forming a pervasive mesh texture typical of highly serpentinized ultramafic rocks. The fibrous habit of chrysotile is clearly visible under crossed nicols, with intense interference colours and well-developed alignment along fracture planes. Veins are evenly distributed throughout the rock, producing a fairly homogeneous microstructure. Primary silicate minerals (olivine, clinopyroxene, and orthopyroxene) are completely or almost completely replaced by serpentine-group minerals, magnetite, and spinel, although some pyroxene remnants are still visible. Serpentine minerals occur as fibres (100–250 μm), forming mesh textures typically associated with the replacement of olivine, and as pseudomorphs after pyroxene (commonly referred to as bastite). Samples from the Figline quarry (Figure 3e,f) display mineralogical and textural features broadly comparable to those of Bagnolo, but with a denser and more regularly oriented vein network, resulting in a slightly more anisotropic texture. Chrysotile fibres show similarly intense interference colours and strong alignment, indicating a comparable degree of serpentinisation. Pyroxene relics are rarer and more scattered than in the Bagnolo samples. In contrast, the samples from the baptistery pavement (Figure 3g,h) are characterised by a less pervasive vein network, with veins that are fewer, thicker, and more irregularly spaced. This gives the rock a more massive and less fibrous appearance. Chrysotile fibres are still present and display clear interference colours, but their distribution is more localised. Pyroxene relics are more abundant compared to the other two quarries, suggesting a slightly lower degree of serpentinization. Overall, the three groups of samples share a common mineralogical assemblage dominated by serpentine-group minerals, magnetite, and spinel, and similar textural features (mesh textures and bastite pseudomorphs), but differ mainly in the density and orientation of chrysotile-rich veinlets and the abundance of primary mineral relics.

5.2.2. X-Ray Powder Diffraction Analysis

The XRD patterns of the Bagnolo samples show very similar profiles (Figure 4). The identified phases are chrysotile, lizardite, and magnetite. Peak broadening of the main phases indicates a predominance of lizardite (although it is not readily visible in the thin section images), with chrysotile also present in significant amounts, and magnetite occurring in substantially lower amounts. For comparison, Figure 4 reports the XRD patterns of the Figline di Prato quarry sample (SF) and the Baptistery pavement sample (Bpt), which closely match those of the Bagnolo quarry specimens.

5.2.3. Electron Probe Micro-Analysis

The analysed Bagnolo samples include both mesh textures (olivine-derived) and bastite (pyroxene-derived) pseudomorphs (

Table 2. Representative microprobe analyses of serpentine-group minerals from Bagnolo (CB) and Figline di Prato (SF) quarries, and pavement of Florentine Baptistery (Bpt).

Sample	CB1-1	CB1-2	CB2-1	CB2-2	CB3-1	CB3-2	CB4-1	CB4-2	CB5-1	CB5-2	SF *	Bpt *
Texture	bastite	mesh	mesh	bastite	bastite	mesh	mesh	mesh	bastite	mesh	#10	#25
SiO2	38.3	40.7	40.9	39.2	40.2	41.3	36.4	41.9	40.5	38.8	39.7	39.1
TiO2	0.07	0.02	-	0.04	0.01	-	0.11	-	-	0.03	0.04	0.15
Al2O3	5.09	0.90	1.62	3.56	4.07	1.22	9.18	2.16	2.41	1.86	1.96	3.11
FeO	5.98	10.7	5.71	6.57	3.71	2.86	4.92	1.78	4.49	10.2	5.61	5.47
Cr2O3	1.36	bdl	0.03	0.96	1.63	0.04	0.44	0.23	0.92	bdl	0.70	1.00
MnO	0.15	0.14	0.08	0.11	0.07	0.05	0.10	0.01	0.09	0.15	0.11	0.13
MgO	28.8	30.9	34.5	35.6	38.2	37.9	34.4	31.7	27.2	35.8	37.5	36.2
CaO	0.10	0.05	0.05	bdl	bdl	0.08	0.08	0.14	0.06	bdl	0.07	0.06
Na2O	bdl	0.09	0.07	bdl	0.02	0.01	0.13	0.06	bdl	bdl	0.03	0.06
K2O	0.01	bdl	bdl	bdl	0.06	bdl	bdl	0.01	0.05	0.01	0.01	0.02
NiO	0.28	0.10	0.20	0.03	0.01	0.42	0.04	0.13	0.29	0.05	na	na
H2O **	11.6	11.6	12.1	12.4	12.8	12.3	12.5	11.7	11.2	12.2	12.4	12.3
Mg#	0.83	0.74	0.86	0.84	0.91	0.93	0.87	0.95	0.86	0.78	0.87	0.87
Cations per 14 O
Si	3.95	4.13	4.05	3.80	3.76	4.01	3.50	4.28	4.33	3.80	3.84	3.80
Ti	0.01	-	-	-	-	-	0.01	-	-	-	-	0.01
Al	0.62	0.11	0.19	0.41	0.45	0.14	1.04	0.26	0.30	0.21	0.22	0.36
Fe2+	0.52	0.92	0.47	0.53	0.29	0.23	0.40	0.15	0.40	0.83	0.45	0.44
Mn	0.01	0.01	0.01	0.01	0.01	-	0.01	-	0.01	0.01	0.01	0.01
Mg	4.43	4.62	5.10	5.14	5.33	5.49	4.93	4.83	4.33	5.22	5.41	5.25
Ca	0.01	0.01	0.01	-	-	0.01	0.01	0.02	0.01	-	0.01	0.01
Na	-	0.02	0.01	-	-	-	0.02	0.01	-	-	0.01	0.01
K	-	-	-	-	0.01	-	-	-	0.01	-	-	-
Cr	0.11	-	-	0.07	0.12	-	0.03	0.02	0.08	-	0.05	0.08
Ni	0.02	0.01	0.02	-	-	0.03	-	0.01	0.02	-	-	-
sum	9.68	9.82	9.86	9.96	9.96	9.92	9.96	9.58	9.49	10.09	10.02	9.98

* = mean values; #10 = ten analyses; #25 = 25 analyses; ** = calculated from stoichiometry (4 H per formula unit); bdl = below detection limit; na = not analysed; Mg# = MgO/(MgO + FeO).

). They exhibit wide ranges in FeO (1.8–11 wt.%) and Al₂O₃ (0.9–9.2 wt.%), including an anomalously Al-rich specimen (CB4-1), as well as locally elevated Ni contents, reflecting contributions from olivine and accessory phases such as plagioclase or spinel. When considering the mean compositions of the Bagnolo samples and comparing them with serpentinites from the Figline di Prato quarry and the Baptistery pavement, the three groups display broadly similar bulk composition, characterised by high MgO, moderate FeO, SiO₂ contents close to 40 wt.%, and very low concentrations of CaO, Na₂O and K₂O. All samples are strongly magnesian, with high Mg# values. Overall, the chemical compositions of serpentine phases are broadly comparable across CB, SF, and Bpt samples. The observed variations fall largely within analytical uncertainty and are interpreted to reflect minor fluctuations linked to precursor mineralogy and local alteration conditions rather than fundamental geochemical differences. All samples consistently plot within the compositional field of Mg-rich serpentines, highlighting a common serpentinization signature modified locally by texture and protolith variability.

5.2.4. ICP-MS Whole Rock Analysis

Table 3. Major (%) and trace (ppm) element composition of serpentinites.

Sample	CB	SF *	Bpt *
SiO2	38.9	37.7	37.8
TiO2	0.05	0.03	0.04
Al2O3	1.70	0.85	1.60
Fe2O3	8.14	8.30	7.60
MnO	0.13	0.07	0.12
MgO	37.5	39.1	38.6
CaO	0.06	0.02	0.03
Na2O	0.01	<0.01	0.02
K2O	<0.01	<0.01	<0.01
P2O5	<0.01	<0.01	<0.01
LOI	13.7	13.6	13.9
Sc	11	7	11
V	43	32	51
Cr	2440	2480	2380
Co	92	106	97
Ni	2120	2230	1970
Rb	<1	<1	<1
Sr	2	2	<2
Y	1.1	<0.5	1.1
Zr	4	1	<1
Nb	0.3	<0.2	<0.2
Cs	<0.1	<0.1	<0.1
Ba	8	<2	<2
La	<0.05	<0.05	<0.05
Ce	<0.06	0.08	<0.05
Pr	<0.01	0.02	<0.01
Nd	<0.05	0.11	<0.05
Sm	0.09	0.03	0.04
Eu	0.070	0.018	0.014
Gd	0.19	0.06	0.09
Tb	0.03	0.01	0.02
Dy	0.28	0.06	0.18
Ho	0.04	0.02	0.04
Er	0.11	0.04	0.12
Tm	<0.005	0.006	0.020
Yb	0.09	0.04	0.14
Lu	0.018	0.005	0.022
Hf	<0.1	<0.1	<0.1
Ta	0.04	0.04	<0.01
Th	<0.05	<0.05	<0.05
U	0.01	<0.01	<0.01

* Major elements from Santo et al. [17]; CB= mean of Cava Bagnolo samples; SF = Figline di Prato; Bpt = baptistery pavement; LOI = loss on ignition.

reports the average whole-rock composition of the serpentinite from Bagnolo quarry. The different samples analysed show very similar concentrations of both major and trace elements, and therefore only the average values are presented. For comparison, the table also includes the composition of serpentinite from Figline di Prato and that from the pavement of the Florentine Baptistery [17].

The whole-rock composition of the CB sample corresponds to a typical ophiolitic serpentinite. Major elements are dominated by SiO₂ (38.9 wt.%) and MgO (37.5 wt.%), consistent with serpentine-group minerals. The relatively high Fe₂O₃ content (8.1 wt.%) reflects the presence of magnetite, while Al₂O₃ (1.7 wt.%) indicates a slightly more aluminous character compared to most serpentinites. Calcium, sodium, and potassium oxides are nearly absent, in line with the highly depleted nature of the protolith. The loss on ignition (13.7 wt.%) corresponds to structurally bound water in serpentine minerals. Trace elements are characterised by high concentrations of Cr (2440 ppm) and Ni (2120 ppm), as expected from ultramafic precursors, together with significant Co (92 ppm) and V (43 ppm). The abundance of incompatible elements such as Sr, Zr, and REE is very low, confirming the depleted geochemical signature. When compared to the other samples, CB exhibits values very close to Bpt, especially in terms of major oxides and Cr–Ni contents, highlighting strong compositional homogeneity across different serpentinite sources. SF, however, displays slightly higher MgO (39.1 wt.%) and Ni (2230 ppm) but lower Al₂O₃ (0.85 wt.%), suggesting a more depleted protolith. Despite these minor variations, all three serpentinites share a comparable chemical fingerprint, reinforcing their common ophiolitic affinity and supporting the close compositional match between the Bagnolo serpentinite and materials historically used in Florentine monuments.

5.2.5. Fibre Quantification

Quantitative analyses were conducted on three serpentinite samples collected from different blocks within the Bagnolo quarry to assess their chrysotile content. The results revealed the presence of chrysotile fibres with concentrations of 14,153 mg/kg, 15,881 mg/kg, and 18,461 mg/kg, found in the samples CB1, CB3, and CB5, respectively. These values represent the chrysotile content quantified and, therefore, should not be directly compared to the total serpentine mineral content of typical serpentinites. The values indicate the amount of fibrous chrysotile present under the adopted analytical protocol, rather than the bulk mineralogical composition of the rock. Figure 5 shows an SEM image of one of the analysed samples.

6. Discussion

The study conducted on serpentinite from the Bagnolo quarry enabled its characterisation. Historical sources (e.g., [17,44,45]) report that serpentinite, already in use by the 11th century in Florentine architecture, was quarried in the areas of Monteferrato (Prato) and Impruneta (Florence). Among the oldest known quarries were those located at Pian di Gello, on the eastern slope of Monte Piccioli (Monteferrato). A comparison with samples from a historic quarry in Figline di Prato, considered the source of the so-called marmo verde used for the Baptistery of San Giovanni in Florence, and with a sample from the original flooring of the Baptistery itself [29], suggests that the boulders from Bagnolo quarry closely resemble the serpentinite historically employed in terms of appearance, texture, and mineralogical composition. This stone was locally known by various names, including Verde di Prato, Nero di Prato, Marmo Nero, and Paragone. The differences observed among samples from different boulders at the Bagnolo quarry are attributable to heterogeneities within the outcrop and to local variations in the agents and processes responsible for the serpentinization of peridotite. Despite these differences, the mineralogical composition of the samples is fairly homogeneous, with the same mineral phases present across all specimens. In terms of physical properties, samples CB2 and CB3 show the lowest capillary water absorption, while CB1, CB3, and CB4 exhibit the lowest water uptake imbibition coefficients. The variability of the capillary water absorption values is likely related to differences in microfracturing, degree of alteration, and textural heterogeneity within the serpentinite. Nevertheless, all values remain consistent with the range commonly reported for this rock type, indicating a generally low to moderate water uptake capacity. Overall, apparent density values are nearly identical across all boulders. Sample CB5, however, displays the highest chrysotile fibre content and the greatest capillary water absorption and imbibition values, making it the least suitable in terms of performance. The observable differences at the hand-sample scale already evident within the quarry, together with the characterisation of the various available lithotypes, make it possible to select appropriate boulders depending on specific needs. These criteria include block size, workability, and intended use location, whether indoor or outdoor. It should be noted that only five blocks (two specimens each) were analysed, which may not fully capture the intra-quarry variability. Therefore, the results should be interpreted with caution, and future studies should include a larger number of specimens to provide a more robust assessment. Our sampling strategy was designed to be representative of the different macroscopic types, rather than to provide a statistical characterisation of the entire deposit.

The serpentinite from the Bagnolo quarry exhibits properties that make it suitable as a replacement material for monument restoration in Tuscany, ensuring structural and chemical compatibility with the original stones. In particular, these rocks display physical properties in line with those of the baptistery pavement. Among them, CB3 and CB4 are the most suitable for restoration or substitution, as their apparent density, open porosity, and water absorption closely match the original material, ensuring greater compatibility and durability. Conversely, CB5, with higher open porosity and imbibition, appears less appropriate for direct substitution in exposed architectural contexts. However, attention must be given to their chrysotile content due to well-known health risks [7,8,9,10,46,47]. In this context, quantifying fibres in this type of serpentinite is a critical parameter, directly determining the safe feasibility of its use. The concentrations of chrysotile fibres in the serpentinite from Bagnolo quarry substantially exceed the regulatory threshold of 1000 mg/kg (0.1%) established by Italian and European legislation for classifying a material as asbestos-containing [48]. The obtained technical data confirm the high chrysotile content through detailed morphometric and gravimetric analyses of 26 individual fibres.

The calculated total asbestos mass in the analysed area corresponds to an estimated chrysotile concentration of at least 14,153 mg/kg, with an associated uncertainty of 7%, validating the consistency of the results across the three samples. Given these findings, the serpentinite from Bagnolo quarry must be considered non-compliant with current safety regulations for architectural, ornamental, or structural applications unless rigorous containment and abatement protocols are applied. It is worth noting that, in addition to the mass-based classification threshold, Italian legislation also sets an occupational exposure limit of 0.1 fibres/cm³ [49], compared to the previous threshold of 0.6 fibres/cm³ established by Law No. 257/1992 [48]. The recently approved EU Directive 2023/2668 further reduces this limit to 0.01 fibres/cm³, to be adopted by December 2025 [50]. The elevated asbestos content also raises concerns regarding environmental and occupational safety during extraction, handling, and potential reuse.

In addition to considering the bulk concentration of fibrous minerals within the serpentinite, it is important to evaluate the potential for fibre release during quarrying and processing. The degree of fibre liberation largely depends on the texture and fabric of the rock, the morphology and orientation of the fibrous minerals (e.g., chrysotile or fibrous antigorite), and the mechanical actions involved in extraction, cutting, and polishing. Even serpentinites containing only minor amounts of fibrous phases may generate airborne fibres when subjected to intense mechanical stress, particularly if the minerals occur along cleavage planes or microfractures. Therefore, assessing fibre release potential provides a more realistic evaluation of occupational and environmental risks associated with the material’s exploitation and use. In the studied samples, the compact structure and the massive, non-oriented distribution of fibrous phases suggest a low propensity for fibre release during standard quarrying and processing operations [51,52,53].

Our results highlight the need for careful material selection and pre-emptive characterisation in the context of the historical reuse or commercial exploitation of serpentinite outcrops, especially in geologically complex ophiolitic settings such as the Monteferrato complex. In this regard, fibre quantification in serpentinite becomes a critical parameter that cannot be overlooked, as it directly determines the safe feasibility of its use. Furthermore, potential exposure to airborne asbestos fibres should be considered during mining and polishing processes, as well as through the wear or handling of products such as floor tiles made from asbestos-bearing serpentinite. Proper risk assessment and mitigation measures are therefore essential to ensure safe use.

Several recent studies support the importance of assessing fibre release, not just bulk fibre concentration. For example, Marzini et al. [53] found that serpentinite with a massive, undeformed texture, even with chrysotile content up to approximately 20%–25%, released very few inhalable fibres in grinding tests, while foliated or cataclastic serpentinite with fibrous veins exhibited much higher release indices. Similarly, in Valmalenco quarries, Cattaneo et al. [51] and Cavallo and Rimoldi [52] reported that airborne chrysotile concentrations varied greatly depending on the presence of veins, surface exposures, and the extent of mechanical disturbance during quarrying or stone processing.

7. Concluding Remarks

The comparative study of serpentinite from Bagnolo quarry demonstrates a strong compositional and textural affinity with historic serpentinite used in Florentine monuments, particularly the Baptistery of San Giovanni. The physical and mechanical properties of the studied blocks indicate their potential suitability for restoration, with samples CB3 and CB4 exhibiting the highest compatibility and durability. However, chrysotile concentrations substantially exceed Italian and EU regulatory thresholds, posing serious health and environmental risks and limiting the material’s practical application. To resolve the apparent contradiction between geological suitability and regulatory non-compliance, a structured decision framework is proposed, considering the following factors: Compatibility (matching mineralogical and textural characteristics with historic stones); Durability (long-term performance under restoration conditions); and Safety/Regulatory compliance (asbestos content and occupational exposure limits). Within this framework, Bagnolo serpentinite scores highly for compatibility and durability, but fails on the safety criterion. Therefore, despite its strong match with historic serpentinite, the practical use of this material in conservation is constrained, and any application would require stringent containment and risk mitigation protocols.