Stable Isotope ( δ 18 O, δ D) Composition of Magmatic Fluids Exsolved from an Active Alkaline Magma Chamber—The Case of the AD 79 Magma Chamber of Vesuvius

: This work documents, for the ﬁrst time, the calculated oxygen and hydrogen isotope composition of the brines exsolved from the peripheral margin of the active magma alkaline chamber that fed the AD 79 (“Pompei”) eruption of Vesuvius. The exsolved hydrosaline ﬂuids had a constant δ 18 O composition and a variable δ D composition, showing a general lowering of δ D at nearly constant δ 18 O content. We argue that the progressive ﬂuid exsolution at the upper peripheral parts of the AD 79 magma chamber may explain this isotopic path. The modeling of the evolution of the hydrogen isotope composition of water remaining dissolved in the melt, and of the exsolved ﬂuid as a consequence of progressive degassing, would favor multiple ﬂuid-release events from the magmatic reservoir. A schematic model on the pulsed accumulation and release of ﬂuids at the top of the magmatic reservoir prior to the eruption is thus proposed. The assessment of the stable isotope composition of the ﬂuids exsolved from the AD 79 magma chamber of Vesuvius may be particularly relevant for the geochemical surveillance of the volcano as it may help to interpret the isotopic composition of fumarolic gases and its variations.


Introduction
Most magmas become saturated in volatiles and exsolve a free fluid phase(s) during their evolution in the shallow crust.This commonly results in the generation of an aqueous magmatic fluid (Edmonds and Wood [1] and reference therein).The exsolution of magmatic fluids from crystallizing silicate melt is a fundamental process, which strongly influences the physical and chemical properties of magmas in the subsurface environment, resulting in large-scale geologic phenomena that have major societal and economic implications.These phenomena include: the release of climate-altering gases into the atmosphere [2], the control on the explosivity of volcanic eruptions [3] and the generation of economically important magmatic-hydrothermal ore deposits [4,5].The exsolution of magmatic fluids is in fact pivotal in the partitioning of ore metals from melt to a hydrothermal system and the subsequent precipitation and concentration of metals [5][6][7].
The evolution of an immiscible aqueous fluid phase from magmas typically takes place in the poorly characterized interval between the orthomagmatic and hydrothermal regimes, which has proved extremely difficult to document due to the transient reactive nature of the exsolved volatile phases [4,8].As a consequence, although of crucial importance for many geological processes, as reported above, direct measurements of the chemical and isotopic composition of magma-derived fluids are not easy to obtain, and their knowledge mostly derives from the study of fumarolic emissions of active volcanoes (Fischer and Chiodini [9] and references therein).In that case, however, the composition of fluids can be mediated by fluid-rock interaction processes and by possible contamination with Minerals 2023, 13, 913 2 of 12 shallow waters.The rare opportunity where we can obtain direct measurements on the original composition of magmatic fluids, is offered by the investigation of fluid inclusions entrapped within crystals (from the juvenile fraction and/or cognate xenoliths) ejected during volcanic eruptions [10][11][12][13][14][15][16][17][18].Fluid inclusions therefore, represent in many cases the only direct witness of the exsolved fluid, and their investigation is pivotal to determine the chemical and isotopic composition of the fluids released by an active magma chamber [17].
The present study was undertaken to document, for the first time, the oxygen and hydrogen isotope composition of the brines exsolved from the peripheral margin of an active magma chamber, as deduced from the investigation of minerals and decrepitated fluid inclusions in cognate xenoliths (foid-bearing syenites) representing the upper portions of the AD 79 (Pompei) magma chamber of Vesuvius.Furthermore, we used δ18O and δD evolution of magmatic fluids to obtain inferences on the dynamic of magmatic fluid release from the AD 79 magma chamber.
Our study focuses on foid-bearing syenite cognate xenoliths representing the solidification front [23] of the upper parts of the AD 79 magma chamber of Vesuvius.These are nearly hypoyaline, nonequigranular rocks consisting of a network of euhedral crystals of K-feldspar; other constituents are sodalite, biotite, K-pargasitic amphibole, garnet, leucite and scarce nepheline.Many accessories are also present: sphene, apatite and monazite are among the most abundant.The foamy appearance of these rocks can be interpreted to result from the crystallization of K-feldspars in the presence of exsolved fluids accumulating under the roof of the magma chamber.On the other hand, K-feldspars host abundant multiphase fluid inclusions testifying the direct exsolution of a hydrosaline aqueous fluid phase from the crystallizing upper peripheral portions of the magma chamber [14,17].These fluid inclusions are generally primary and consist of a deformed vapor bubble + scarce liquid phase + several clear daughter minerals of cubic shape (halite and sylvite) + two or more rounded and tabular daughter minerals ± opaque crystals (Figure 1).Microthermometric experiments revealed homogenization temperatures ranging from 760 • to 830 • C, with a mode at about 810 • C, and salinity in the range of 62-70 wt.% NaCl equiv.[17].With the aim to characterize the isotopic composition of the fluids exsolved from the peripheral parts of the AD 79 magma chamber of Vesuvius prior to the eruption, oxygen and hydrogen isotope compositions of minerals, assumed to be in equilibrium with the exsolved magmatic fluid, and multiphase fluid inclusions (only δD) from foid-bearing syenite xenoliths were determined.

Analytical Methods
Oxygen isotope analyses on K-feldspar and biotite were carried out by laser fluorination of the total sample (typically between 1 and 2 mg) with excess ClF 3 using a CO 2 laser as a heat source (T in excess of 1500 °C; [24]).All fluorinations resulted in 100% release of O 2 from the silica lattice.This O 2 was converted to CO 2 by reaction with hot graphite, then analyzed on-line by a VG PRISM III spectrometer.Reproducibility is better than ±0.3‰ (1σ) based on repeated analyses of internal lab standard quartz SES-1 (+9.6‰).For H isotope analysis of fluid inclusion water, pure K-feldspar samples (mostly between 0.5 and 2 gr), hosting hypersaline fluid inclusions, were heated to 200 °C overnight under high vacuum to release labile volatiles.In the case of biotite, outgassing was carried out on several tens of mg of pure mineral separate at 150 °C.In both cases, samples were then placed in an outgassed Pt crucible, and gradually heated by radiofrequency induction in an evacuated quartz tube to temperatures in excess of 1500 °C.The released water was then reduced to H 2 in a chromium furnace at 800 °C [25], with the evolved gas measured quantitatively in a Hg manometer, then collected using a Toepler pump.The gas was subsequently analyzed on a VG 602D mass spectrometer with a manual Hg, high gas compression inlet system.Replicate analyses of water standards were used to calibrate raw data from the mass spectrometer (international std GISP, and internal standards Lt Std, LSW) and gave a standard error reproducibility better than ±3‰.Replicate analyses of international mineral standard NBS-30 (biotite) also gave reproducibility around ±2‰.

Results
The measured oxygen isotope compositions of K-feldspars (δ 18 O values between 9.8‰ and 10.1‰) and biotite (8.6‰) of the juvenile fraction (Table 1) were compatible with the values reported for Vesuvian products [26,27].The δ 18 O values of K-feldspars and biotite of foid-bearing syenites were consistently higher than those of the juvenile fraction and ranged from 10.6‰ to 12.5‰ (K-feldspars) and 10.2‰ (biotite), respectively (Table 1).The oxygen isotopic composition of the fluid in equilibrium with K-feldspars and biotite of foid-bearing syenites was calculated using the fractionation data of Zheng [28] and Zheng [29], respectively.The equilibrium temperature was assumed to be 810 °C

Analytical Methods
Oxygen isotope analyses on K-feldspar and biotite were carried out by laser fluorination of the total sample (typically between 1 and 2 mg) with excess ClF3 using a CO 2 laser as a heat source (T in excess of 1500 • C; [24]).All fluorinations resulted in 100% release of O 2 from the silica lattice.This O 2 was converted to CO 2 by reaction with hot graphite, then analyzed on-line by a VG PRISM III spectrometer.Reproducibility is better than ±0.3‰ (1σ) based on repeated analyses of internal lab standard quartz SES-1 (+9.6‰).For H isotope analysis of fluid inclusion water, pure K-feldspar samples (mostly between 0.5 and 2 gr), hosting hypersaline fluid inclusions, were heated to 200 • C overnight under high vacuum to release labile volatiles.In the case of biotite, outgassing was carried out on several tens of mg of pure mineral separate at 150 • C. In both cases, samples were then placed in an outgassed Pt crucible, and gradually heated by radiofrequency induction in an evacuated quartz tube to temperatures in excess of 1500 • C. The released water was then reduced to H2 in a chromium furnace at 800 • C [25], with the evolved gas measured quantitatively in a Hg manometer, then collected using a Toepler pump.The gas was subsequently analyzed on a VG 602D mass spectrometer with a manual Hg, high gas compression inlet system.Replicate analyses of water standards were used to calibrate raw data from the mass spectrometer (international std GISP, and internal standards Lt Std, LSW) and gave a standard error reproducibility better than ±3‰.Replicate analyses of international mineral standard NBS-30 (biotite) also gave reproducibility around ±2‰.

Results
The measured oxygen isotope compositions of K-feldspars (δ 18 O values between 9.8‰ and 10.1‰) and biotite (8.6‰) of the juvenile fraction (Table 1) were compatible with the values reported for Vesuvian products [26,27].The δ 18 O values of K-feldspars and biotite of foid-bearing syenites were consistently higher than those of the juvenile fraction and ranged from 10.6‰ to 12.5‰ (K-feldspars) and 10.2‰ (biotite), respectively (Table 1).The oxygen isotopic composition of the fluid in equilibrium with K-feldspars and biotite of foidbearing syenites was calculated using the fractionation data of Zheng [28] and Zheng [29], respectively.The equilibrium temperature was assumed to be 810 • C on the basis of fluid inclusion microthermometric results [17].The K-feldspar-water fractionation factor was very low (about 0.1‰) considering this high equilibrium temperature, and the δ 18 O calculated composition of the fluid in equilibrium with these K-feldspars was in the range of 10.7-12.6‰(Figure 2).The δ 18 O calculated composition of fluid in equilibrium with biotite was 12.5‰ and falls in the range obtained for the fluid in equilibrium with K-feldspars.
Minerals 2023, 13, x FOR PEER REVIEW 4 of 12 on the basis of fluid inclusion microthermometric results [17].The K-feldspar-water fractionation factor was very low (about 0.1‰) considering this high equilibrium temperature, and the δ 18 O calculated composition of the fluid in equilibrium with these K-feldspars was in the range of 10.7-12.6‰(Figure 2).The δ 18 O calculated composition of fluid in equilibrium with biotite was 12.5‰ and falls in the range obtained for the fluid in equilibrium with K-feldspars.[28] (oxygen, K-feldspar), Zheng [29] (oxygen, biotite) and assuming a temperature of 810 • C as deduced from fluid inclusion study [17].** = measured values from decrepitated multiphase fluid inclusions.*** = δD value for igneous biotite is taken as being representative of the water dissolved in the melt since there is little fractionation between OH minerals and melt water [30].Red = K-feldspar-biotite pair representative of the juvenile fraction reported in Figure 3. Blue = K-feldspar-biotite pair representative of the foid-bearing syenite reported in Figure 3.The δD values of water trapped in hypersaline fluid inclusions, hosted within K-feldspar from different samples of foid-bearing syenite xenoliths, are rather variable ranging from −55‰ to −111‰ (Table 1).The δD values of biotite of AD 79 and AD 472 juvenile fractions (for comparison) are −52‰ and −58‰, respectively, whereas the measured δD of biotite of foid-bearing syenite xenoliths is −143‰ (Table 1).

Discussion
The collected stable isotope data on minerals and fluid inclusions allow for constraining the δ 18 O and δD composition of the fluid exsolved from the peripheral portions of the AD 79 Vesuvius magma chamber.The δ 18 O values of the fluids in equilibrium with K-feldspar and biotite of foid-bearing syenites are in the range of 10.7-12.6‰,but most of the data are between 11.0‰ and 11.5‰ (Figure 2).These values are higher than the typical "magmatic water box" [31,32] but almost all fall close to the 18 O-enriched limit of the convergent margins magmatic waters (CMW) proposed by [33,34].However, these high δ 18 O values, and in particular the highest (12.5‰ and 12.6‰), may reflect exchange effects between magma and wall rocks at the margins of the magma chamber prior to fluid exsolution.This process is highlighted in several margins of plutons and seems to take place without any significant mineralogical or chemical change in the magmatic rocks.It appears to involve just 18 O exchange directly between magmatic melt and the wall rocks The δD values of water trapped in hypersaline fluid inclusions, hosted within Kfeldspar from different samples of foid-bearing syenite xenoliths, are rather variable ranging from −55‰ to −111‰ (Table 1).The δD values of biotite of AD 79 and AD 472 juvenile fractions (for comparison) are −52‰ and −58‰, respectively, whereas the measured δD of biotite of foid-bearing syenite xenoliths is −143‰ (Table 1).

Discussion
The collected stable isotope data on minerals and fluid inclusions allow for constraining the δ 18 O and δD composition of the fluid exsolved from the peripheral portions of the AD 79 Vesuvius magma chamber.The δ 18 O values of the fluids in equilibrium with K-feldspar and biotite of foid-bearing syenites are in the range of 10.7-12.6‰,but most of the data are between 11.0‰ and 11.5‰ (Figure 2).These values are higher than the typical "magmatic water box" [31,32] but almost all fall close to the 18 O-enriched limit of the convergent margins magmatic waters (CMW) proposed by [33,34].However, these high δ 18 O values, and in particular the highest (12.5‰ and 12.6‰), may reflect exchange effects between magma and wall rocks at the margins of the magma chamber prior to fluid exsolution.This process is highlighted in several margins of plutons and seems to take place without any significant mineralogical or chemical change in the magmatic rocks.It appears to involve just 18 O exchange directly between magmatic melt and the wall rocks or by the local assimilation of the wall rocks (Taylor and Sheppard [35] and references therein).
The δ 18 O composition of K-feldspar and biotite were also used to give an estimation of the temperature of equilibration.The fractionation equation for the K-feldspar-biotite pair (∆ feldsp-biot = 2.5(10 6 T −2 )−0.31) has been calculated by using the fractionation equations of O'Neil and Taylor [36] and Bottinga and Javoy [37,38] for K-feldspar and biotite, respectively.The oxygen isotope composition of K-feldspar and biotite of the juvenile fraction gave a temperature of equilibrium of 850-900 • C (Figure 3) that agrees very well with the temperatures estimated from melt inclusion microthermometries [20].The δ 18 O value of K-feldspar and biotite in foid-bearing syenites shows a per mil difference lower than that of the juvenile fraction, this would indicate higher temperatures of equilibrium that are reasonably unlikely, so δ 18 O disequilibrium is implied instead.
The hydrogen isotopic composition of the exsolved fluids was directly measured in decrepitated multiphase fluid inclusions, which are hosted in several samples of foid-bearing syenite xenoliths.Despite the homogeneous petrographic and microthermometric characteristics found in the various samples, fluid inclusions show a large δD range from −55‰ to −111‰ (Table 1).On a δ 18 O-δD diagram, our data depict a trend of δD impoverishment at nearly constant δ 18 O content for exsolved fluids (Figure 4).
than that of the juvenile fraction, this would indicate higher temperatures of equilibr that are reasonably unlikely, so δ 18 O disequilibrium is implied instead.
The hydrogen isotopic composition of the exsolved fluids was directly measure decrepitated multiphase fluid inclusions, which are hosted in several samples foid-bearing syenite xenoliths.Despite the homogeneous petrographic and microt mometric characteristics found in the various samples, fluid inclusions show a large range from −55‰ to −111‰ (Table 1).On a δ 18 O-δD diagram, our data depict a tren δD impoverishment at nearly constant δ 18 O content for exsolved fluids (Figure 4).  1) showing the oxy and hydrogen isotopic composition of magmatic fluids exsolved at the crystallizing upper per eral portions of the AD 79 Vesuvius magma chamber.PMW = primary magmatic water [39]; F = felsic magmatic water [40]; CMW = convergent margins magmatic waters [33,34].This is very similar to what was reported in several stable isotope studies on m matic fluids involved in potassic and phyllic alterations of porphyry copper deposits in skarn deposits [41][42][43][44].The highest δD values found in multiphase fluid inclusions close to that of the juvenile biotite of AD 79 and also AD 472 Vesuvius erupt (−52/−58‰, Table 1).These values are compatible with the δD composition of "magm ic" water proposed by Taylor [32] and Sheppard [39] and also with the δD compositio the water dissolved in felsic magmas of Taylor [40] (considering that Vesuvian mag emplaced into continental crust), and can be assumed to represent the hydrogen isoto composition of the water dissolved in Vesuvian magma before exsolution begins.C versely, several δD data are lower than those assumed to be representative of the typ magmatic waters (Figure 4).Excluding the contribution of external fluids (i.e., mete water), one explanation for such a large δD variation is the effect of progressive de  1) showing the oxygen and hydrogen isotopic composition of magmatic fluids exsolved at the crystallizing upper peripheral portions of the AD 79 Vesuvius magma chamber.PMW = primary magmatic water [39]; FMW = felsic magmatic water [40]; CMW = convergent margins magmatic waters [33,34].This is very similar to what was reported in several stable isotope studies on magmatic fluids involved in potassic and phyllic alterations of porphyry copper deposits and in skarn deposits [41][42][43][44].The highest δD values found in multiphase fluid inclusions are close to that of the juvenile biotite of AD 79 and also AD 472 Vesuvius eruptions (−52/−58‰, Table 1).These values are compatible with the δD composition of "magmatic" water proposed by Taylor [32] and Sheppard [39] and also with the δD composition of the water dissolved in felsic magmas of Taylor [40] (considering that Vesuvian magmas emplaced into continental crust), and can be assumed to represent the hydrogen isotopic composition of the water dissolved in Vesuvian magma before exsolution begins.Conversely, several δD data are lower than those assumed to be representative of the typical magmatic waters (Figure 4).Excluding the contribution of external fluids (i.e., meteoric water), one explanation for such a large δD variation is the effect of progressive degassing of the solidification front of the AD 79 magma chamber.The trend of δD impoverishment at nearly constant δ 18 O content can in fact be explained by considering the significant hydrogen isotopic fractionation as water in a melt exsolves to the aqueous fluid phase [45].In contrast, the δ 18 O fractionation between water and melt is negligible at magmatic temperature [46].Fluid inclusion evidences the direct exsolution of a hypersaline fluid phase from magma [17], and excludes that δD fractionation could be influenced by the phase separation of hypersaline liquid from vapor [47,48].According to Dobson et al. [49] and Taylor [44,50,51], the fractionation of hydrogen isotopes between magma and an exsolved aqueous fluid phase varies primarily as a function of the hydrogen speciation (H 2 O and OH) in the melt, which varies with water content [52] and can be calculated, according to Taylor [51], as: where X and 1 − X are, respectively, the atom fraction of hydrogen present in the melt as H 2 O and OH and α is the isotopic fractionation factor.The values of αH 2 O−(H 2 O melt ) and αH 2 O−(OH melt ) are 1.000 and 1.0415, respectively, as estimated by Dobson et al. [49].
The bulk fractionation factor between exsolved water and magma can therefore vary from about 0‰ to about 40‰.Thus, as water exsolves from the saturated melt, the residual melt is depleted in deuterium.The extent and rate of decrease in deuterium depend upon whether degassing is an open-or closed-system (open-system degassing leads to the largest variations in δD) process and on the ratio of H 2 O and OH species in the melt [44,[49][50][51]53].As a consequence, Hedenquist et al. [39] argued that magmatic water, as defined by Taylor [32], is more appropriately termed residual magmatic water because it represents the water left in magma upon complete crystallization, and thus, after the majority of the water has exsolved.The progressively lighter δD water that remains in the melt, as a result of fluid exsolution, may be incorporated by hydroxyl-bearing mineral and/or degas at later stage, and become depleted in deuterium from that of the initial exsolved fluid [41].The trend of δD impoverishment shown by Vesuvius magmatic fluids can be envisaged to be generated by this process.The fluids entrapped within inclusions found in foid-bearing syenite cognate xenoliths would record the progressive release of a magmatic aqueous fluid phase, which is progressively depleted in deuterium, from the upper peripheral crystallizing margin of the AD 79 magma chamber.The low δD values reached by these fluids (up to −111‰) sugget that the exsolution occurred (almost in part) in open-system conditions (Rayleigh fractionation).We might hypothesize that an initial closed-system degassing was followed by an open-system degassing, reflecting the shift from ductile and largely sealed conditions to brittle permeable conditions ("Closed-Open").More probably, we envisage a more complex behavior in which the exsolved magmatic fluids (closed-system degassing) accumulated at lithostatic pressure in the roof zone [1] of the magma chamber (Figure 5a).This accumulation of fluids caused a pressure increase near the top of the zone of accumulation, resulting in hydraulic fracturing or opening of pre-existing fractures, with the consequent release of accumulated fluids (Figure 5b) from the normally plastic region into the brittle, and depressurization [54][55][56].The released fluids flowed into the carbonate wall rocks producing exoskarn [14][15][16].After the decrease in pressure of the accumulated fluids, the fractures were closed by mineral deposition and plastic flow (Figure 5c), and magmatic fluids again restarted to accumulate at the top of the magma chamber (new closed-system degassing stage).This process probably recurred more times before the AD 79 Plinian eruption that disrupted the magma chamber, leading to the intermittent release of fluids (as already suggested for some porphyry systems [57,58]) that were even more depleted in deuterium.For this reason, it can be viewed as a quasi-open degassing process consisting of several closed or nearly closed stages ("Multi-Step Closed/Open").The evolution of the hydrogen isotope composition of water remaining dissolved in the melt and of the exsolved fluid as a function of the two proposed processes ("Closed-Open" and "Multi-Step Closed/Open") is modeled in Figure 6 (see Appendix A for calculations).The results show that both processes can well explain the δD evolution of the exsolved fluids recorded in fluid inclusions of the foid-bearing syenite xenoliths (Figure 6).
The very low δD value found in biotite from the foid-bearing syenites (−143‰) may represent the isotopic composition of the residual water, after the large majority of the water has exsolved.This process probably recurred more times before the AD 79 Plinian eruption that disrupted the magma chamber, leading to the intermittent release of fluids (as already suggested for some porphyry systems [57,58]) that were even more depleted in deuterium.For this reason, it can be viewed as a quasi-open degassing process consisting of  The very low δD value found in biotite from the foid-bearing syenites (−143‰) may represent the isotopic composition of the residual water, after the large majority of the water has exsolved.

Conclusions
This work reports for the first time the calculated stable isotope composition (oxygen and hydrogen) of the fluids directly exsolved from the peripheral margin of the active alkaline magma chamber that fed the AD 79 eruption of Vesuvius.These results show that the exsolved hydrosaline fluids had a constant δ 18 O composition and a variable δD composition, depicting a trend of δD impoverishment at nearly constant δ 18 O content.We argued that progressive fluid exsolution at the upper peripheral parts of the AD 79 magma chamber may explain this isotopic trend.This allowed us to propose a schematic model on the pulsed accumulation and release of fluids at the top of the magmatic reservoir prior to the eruption (Figure 5).The observed trend of progressive δD depletion of magmatic fluids released from an active magma chamber, as a consequence of the degassing process, agrees well with the conclusions reported in several previous works on "fossil systems" (e.g., porphyry systems, [39,41,42]) and suggests that this process may be more common than supposed before and does not need very long time to occur.The assessment of the stable isotope composition of the fluids exsolved from the AD 79 magma chamber of Vesuvius can also be particularly relevant for the geochemical surveillance of the volcano, as it may help to interpret the isotopic composition of fumarolic gases and its

Conclusions
This work reports for the first time the calculated stable isotope composition (oxygen and hydrogen) of the fluids directly exsolved from the peripheral margin of the active alkaline magma chamber that fed the AD 79 eruption of Vesuvius.These results show that the exsolved hydrosaline fluids had a constant δ 18 O composition and a variable δD composition, depicting a trend of δD impoverishment at nearly constant δ 18 O content.We argued that progressive fluid exsolution at the upper peripheral parts of the AD 79 magma chamber may explain this isotopic trend.This allowed us to propose a schematic model on the pulsed accumulation and release of fluids at the top of the magmatic reservoir prior to the eruption (Figure 5).The observed trend of progressive δD depletion of magmatic fluids released from an active magma chamber, as a consequence of the degassing process, agrees well with the conclusions reported in several previous works on "fossil systems" (e.g., porphyry systems, [39,41,42]) and suggests that this process may be more common than supposed before and does not need very long time to occur.The assessment of the stable isotope composition of the fluids exsolved from the AD 79 magma chamber of Vesuvius can also be particularly relevant for the geochemical surveillance of the volcano, as it may help to interpret the isotopic composition of fumarolic gases and its variations.These are critical parameters in detecting how Vesuvius works and the transitions in its volcanic activity.

Figure 2 .
Figure 2. Frequency histogram of δ 18 O composition of fluids in equilibrium with K-feldspar and biotite.
values from fractionation factors of: Zheng[28] (oxygen, K-feldspar), Zheng[29] (oxygen, biotite) and assuming a temperature of 810 °C as deduced from fluid inclusion study[17].** = measured values from decrepitated multiphase fluid inclusions.*** = δD value for igneous biotite is taken as being representative of the water dissolved in the melt since there is little fractionation between OH minerals and melt water[30].Red = K-feldspar-biotite pair representative of the juvenile fraction reported in Figure3.Blue = K-feldspar-biotite pair representative of the foid-bearing syenite reported in Figure3.

Figure 3 .
Figure 3. Diagram of δ 18 O values of coexisting K-feldspar and biotite.White dot K-feldspar-biotite pair in juvenile fraction, black dot K-feldspar-biotite pair in foid-bearing syenite.

Figure 3 .
Figure 3. Diagram of δ 18 O values of coexisting K-feldspar and biotite.White dot K-feldspar-biotite pair in juvenile fraction, black dot K-feldspar-biotite pair in foid-bearing syenite.

Figure 5 .
Figure 5. (a-c) Schematic sketch of the proposed Multi-Step Closed/Open degassing model at the roof of the AD 79 magma chamber of Vesuvius.

Figure 5 .
Figure 5. (a-c) Schematic sketch of the proposed Multi-Step Closed/Open degassing model at the roof of the AD 79 magma chamber of Vesuvius.

Figure 6 .
Figure 6.Plot of δD versus fraction of water remaining in melt, illustrating hydrogen isotope fractionation and variation of water content in the melt during exsolution of fluids at the roof of the AD 79 magma chamber of Vesuvius under Closed/Open system conditions and under Multi−Step Closed/Open system conditions (see Appendix A for calculation details).

Figure 6 .
Figure 6.Plot of δD versus fraction of water remaining in melt, illustrating hydrogen isotope fractionation and variation of water content in the melt during exsolution of fluids at the roof of the AD 79 magma chamber of Vesuvius under Closed/Open system conditions and under Multi−Step Closed/Open system conditions (see Appendix A for calculation details).

Table 1 .
Oxygen and hydrogen isotopic data of minerals and fluid inclusions, and calculated fluid compositions for waters in equilibrium with minerals at temperatures determined from fluid inclusion data.

Table 1 .
Oxygen and hydrogen isotopic data of minerals and fluid inclusions, and calculated fluid compositions for waters in equilibrium with minerals at temperatures determined from fluid inclusion data.