Synchysite-(Ce) from Cinquevalli (Trento, Italy): Stacking Disorder and the Polytypism of (Ca,REE)-Fluorcarbonates

: Synchysite-(Ce) at Cinquevalli occurs as ﬁne needles intergrown with quartz in quartz-dikes and in association with altered K-feldspar and oxidized chalcopyrite as major constituents. Synchysite-(Ce) [Ca 1.00 (Ce 0.43 La 0.26 Nd 0.17 Y 0.07 Pr 0.04 Sm 0.02 Gd 0.01 ) Σ = 1.00 (CO 3 ) 2 (F 0.58 (OH) 0.42 )], shows an overgrowth rim of bastnäsite-(Ce) [(Ce 0.34 La 0.25 Nd 0.17 Pb 0.07 C a 0.06 Y 0.06 Pr 0.04 S m 0.02 Gd 0.01 ) Σ = 1.00 C O 3 (F 0.75 (OH) 0.25 )]. Unit cell reﬁnement of synchysite ( C 2 / c ) and bastnäsite ( P 62 c ) led to a = 12.272(4), b = 7.100(2), c = 18.640(5) Å, β = 102.71(5) ◦ , and a = 7.085(1), c = 9.746(2) Å, respectively. Polysomatic faults are sporadic, but polytypic disorder is widespread. High resolution transmission electron microscopy images taken along [100] or (cid:104) 130 (cid:105) show an apparent order and the related di ﬀ raction patterns are streak-free. Conversely, along [010] or (cid:104) 110 (cid:105) , a high density of stacking faults is observed and the related di ﬀ raction patterns show hhl rows with h (cid:44) 3 n a ﬀ ected by streaks. No ordered domain larger than a few unit cells was detected. The stacking sequence of (Ca,REE)-ﬂuorcarbonates can be compared with subfamily-B mica polytypes (2 M 2 , 2 O and 6 H ), which are characterized by n · 60 ◦ ( n = odd) rotations. Subfamily-A polytypes (1 M , 2 M 1 and 3 T ), characterized by n · 60 ◦ ( n = even) rotations, should not be possible. Synchysite, characterized by ± 60 ◦ rotations, can be likened to the 2 M 2 polytype.

Diverse layer definitions in terms of composition, thickness, position (within the sequence), and symmetry have been employed so far, making the situation quite confusing (for a summary of all these  [19], synchysite-(Ce) [17], parisite-(Ce) [18] and vaterite [20], all as seen along [010]. Yellow = Ce, blue = Ca, green = F, black = C. Oxygen atoms at the vertices of triangular CO3 groups are omitted for clarity. Analogous layers in the different structures are connected by arrows and indicated with their relative codes.
In this study, we undertook a HRTEM study of synchysite-(Ce) from Cinquevalli (Trento, Italy) that have been never studied before, with the aim to contribute to the assessment of the BmSn polysomatic series. Structural defects are thoroughly characterized, a widespread polytypic disorder detected, and a comparison with the polytypism of mica is provided.

Materials and Methods
The studied (Ca,REE)-fluorcarbonates are hosted in quartziferous veins extracted from the gangue material of the Cinquevalli Mine, sited near Roncegno, Trento, Italy ( Figure 2). The mining area is located on the southern slope of the Mount Fravort, along the valley carved by the Argenta stream, at altitude around 1500 m. The ore body is represented by quartziferous dikes and veins, from a few centimeters to tens of meters in thickness (not reported in the map either because too small or cropping out only in mine tunnels), cross cutting Paleozoic phyllites and micaschists of the crystalline basement and the Roncegno granodiorite-monzogranite pluton. The beginning of the mining activity dates back to the 12th century, as documented by the finding of mining inlets, melting furnaces, and abundant melting slags. In the Middle Ages, silver and especially copper were mined at Cinquevalli. In the last century, until the end of the activity that took place in 1940, the area was exploited for sphalerite, galena, and copper. In the last years of activity and for a short period, fluorite was extracted [21]. Fluorcarbonates are scarce and dispersed and, therefore, have never represented an exploitable ore.  [19], synchysite-(Ce) [17], parisite-(Ce) [18] and vaterite [20], all as seen along [010]. Yellow = Ce, blue = Ca, green = F, black = C. Oxygen atoms at the vertices of triangular CO 3 groups are omitted for clarity. Analogous layers in the different structures are connected by arrows and indicated with their relative codes.
In this study, we undertook a HRTEM study of synchysite-(Ce) from Cinquevalli (Trento, Italy) that have been never studied before, with the aim to contribute to the assessment of the B m S n polysomatic series. Structural defects are thoroughly characterized, a widespread polytypic disorder detected, and a comparison with the polytypism of mica is provided.

Materials and Methods
The studied (Ca,REE)-fluorcarbonates are hosted in quartziferous veins extracted from the gangue material of the Cinquevalli Mine, sited near Roncegno, Trento, Italy ( Figure 2). The mining area is located on the southern slope of the Mount Fravort, along the valley carved by the Argenta stream, at altitude around 1500 m. The ore body is represented by quartziferous dikes and veins, from a few centimeters to tens of meters in thickness (not reported in the map either because too small or cropping out only in mine tunnels), cross cutting Paleozoic phyllites and micaschists of the crystalline basement and the Roncegno granodiorite-monzogranite pluton. The beginning of the mining activity dates back to the 12th century, as documented by the finding of mining inlets, melting furnaces, and abundant melting slags. In the Middle Ages, silver and especially copper were mined at Cinquevalli. In the last century, until the end of the activity that took place in 1940, the area was exploited for sphalerite, galena, and copper. In the last years of activity and for a short period, fluorite was extracted [21]. Fluorcarbonates are scarce and dispersed and, therefore, have never represented an exploitable ore.
Representative hand specimens were sliced with a diamond wheel saw and prepared as polished thin sections suitable for both optical and scanning electron microscopy (SEM) observations. Representative aliquots of the samples were powdered and used for energy dispersive X-ray fluorescence (EDXRF) and X-ray powder diffraction (XRPD) bulk analyses. For the TEM sample preparation, the green parts recognized on the hand specimen were scratched with a needle and left to settle directly Minerals 2020, 10, 77 3 of 19 in an agate mortar, where the grain size was reduced with a pestle. The resulting powder was ultrasonicated in isopropanol and a few microliters of solution were pipetted in holey-carbon Cu-grids. Representative hand specimens were sliced with a diamond wheel saw and prepared as polished thin sections suitable for both optical and scanning electron microscopy (SEM) observations. Representative aliquots of the samples were powdered and used for energy dispersive X-ray fluorescence (EDXRF) and X-ray powder diffraction (XRPD) bulk analyses. For the TEM sample preparation, the green parts recognized on the hand specimen were scratched with a needle and left to settle directly in an agate mortar, where the grain size was reduced with a pestle. The resulting powder was ultrasonicated in isopropanol and a few microliters of solution were pipetted in holeycarbon Cu-grids.
XRPD analyses were performed with a PANanalytical X'Pert-Pro PW3060 diffractometer, operating in Bragg-Brentano specular (θ-θ) geometry and equipped with a X'Celerator position sensitive detector. Diffractometer scans were recorded in the 3-70° 2θ range with step size of 0.02° and 30 s counting time, at 40 mA and 40 kV (CuKα radiation). A Ni filter along the diffracted beam path was used to filter out the CuKβ radiation. The sample holder was allowed to spin horizontally during measurements to improve particle statistics.
The identification of major and minor phases was done with the X'Pert High Score software (PANanalytical) using the ICSD PDF2-2004 database. Quantitative phase analyses (QPA) were performed with the Rietveld method [23,24] as implemented in the GSAS/EXPEGUI program [25].
Semi-quantitative bulk chemical analyses were obtained with a Pananalytical Epsilon 3 X EDXRF instrument using a standardless method. Volatile components (H2O and CO2) were determined through the weight loss on ignition (LOI).
SEM investigations were performed with a Tescan VEGA TS 5136XM instrument operating at 20 keV and equipped with an EDAX GENESIS 4000XMS energy dispersive system (EDS), and with a Zeiss Gemini 500 field emission gun (FEG) instrument, equipped with a Bruker XFlash ® EDS detector. The standardless method and the ZAF correction method were used for semi-quantitative analyses.
TEM observations were performed at the Department of Physical Sciences, Earth and Environment of the University of Siena with a Jeol JEM 2010 instrument operating at 200 keV and XRPD analyses were performed with a PANanalytical X'Pert-Pro PW3060 diffractometer, operating in Bragg-Brentano specular (θ-θ) geometry and equipped with a X'Celerator position sensitive detector. Diffractometer scans were recorded in the 3-70 • 2θ range with step size of 0.02 • and 30 s counting time, at 40 mA and 40 kV (CuK α radiation). A Ni filter along the diffracted beam path was used to filter out the CuK β radiation. The sample holder was allowed to spin horizontally during measurements to improve particle statistics.
The identification of major and minor phases was done with the X'Pert High Score software (PANanalytical) using the ICSD PDF2-2004 database. Quantitative phase analyses (QPA) were performed with the Rietveld method [23,24] as implemented in the GSAS/EXPEGUI program [25].
Semi-quantitative bulk chemical analyses were obtained with a Pananalytical Epsilon 3 X EDXRF instrument using a standardless method. Volatile components (H 2 O and CO 2 ) were determined through the weight loss on ignition (LOI).
SEM investigations were performed with a Tescan VEGA TS 5136XM instrument operating at 20 keV and equipped with an EDAX GENESIS 4000XMS energy dispersive system (EDS), and with a Zeiss Gemini 500 field emission gun (FEG) instrument, equipped with a Bruker XFlash ® EDS detector. The standardless method and the ZAF correction method were used for semi-quantitative analyses.
TEM observations were performed at the Department of Physical Sciences, Earth and Environment of the University of Siena with a Jeol JEM 2010 instrument operating at 200 keV and equipped with an Oxford Link EDS and an Olympus Tengra 2.3k × 2.3k × 14-bit slow-scan CCD camera. To remove noise contrast due to amorphous materials, HRTEM images were rotationally filtered [26] with the HRTEM filter [27], as implemented in the Gatan Digital Micrograph version 3.9. High resolution image Oxidized chalcopyrite (gold metal luster), quartz (white) and K-feldspar (pink) can also be recognized. The latter, however, is overestimated by visual inspection, because most of the reddish color is due to oxides and hydroxides that alter sulphides. (b) Photomicrograph (plain polarized light; 50X) showing bands of needle-like, randomly oriented crystals of (Ca,REE)-fluorcarbonates (dark) embedded in quartz. (c) Related cross polarized light photomicrograph. (d) Photomicrograph (cross polarized light) of coarse, euhedral K-feldspar (center) altered by white mica and bordered by ribbons of fibrous quartz (chalcedony). Photomicrograph side ∼9 mm. In thin section, under polarized light, the fluorcarbonates appear as opaque needles intergrown with granular quartz. The needle morphology may also result from cross sectioning thin lamellae of fluorcarbonates at high angle to the thin section plane (Figure 3b,c). Coarse, euhedral K-feldspar is sometimes altered to white mica (Figure 3d). In addition to granular quartz, fibrous quartz recalling chalcedony is present, in the latter case not intergrown with fluorcarbonates ( Figure 3d). Therefore, two generations of quartz seem present. In the thin section, chalcopyrite shows a reddish rim, suggesting later alteration in hematite/goethite. XRPD Rietveld refinement of a representative aliquot of the sample (R p = 8.18%; R wp = 11.22%; X 2 = 6.703) gave the following composition (wt%): quartz 79.4(1)%, muscovite 9.5(4)%, K-feldspar 6.9(3)%, chalcopyrite 1.4(1)%, bastnäsite 1.5(1)% and synchysite 1.2(1)%. These results are in keeping with the optical observations, although muscovite, possibly as an alteration of K-feldspar, is probably overestimated because of (001) preferential orientation. Unit cell refinement of bastnäsite (P62c) and synchysite (C2/c) gave the following results: a = 7.085(1), c = 9.746(2) Å, and a = 12.272(4), b = 7.100(2), c = 18.640(5) Å, β = 102.71(5) • , respectively, which are consistent with literature data, although should be taken with caution because of their low concentration.
EDXRF data of a representative aliquot of the studied samples are close to a typical granitic aplite composition, except the Al 2 O 3 content reduced by half, absence of NaO and higher abundance of exotic components, such as REE-oxides ( Table 1). The low Al 2 O 3 content and the absence of NaO determines the strong prevalence of quartz over feldspars evidenced by the Rietveld refinement. Among the other major components, K 2 O is present in K-feldspar and in its white mica alteration; S, Cu and Fe are hosted in chalcopyrite (of course not as oxides); Fe 2 O 3 is present in the chalcopyrite alteration halo; REEs (and CaO) are hosted in fluorcarbonates. On the basis of the texture, mineral assemblage, and chemical composition, the fluorcarbonate-hosting rock can be defined as a quartz dike.

Microstructure and Mineral Composition
In order to complement optical observations as well as XRPD and EDXRF results, a SEM-EDS investigation was undertaken. As argued from the optical examination, the fluorcarbonates form interwoven needle-like crystals embedded in quartz and K-feldspar ( Figure 4). An additional phase not detected with the previous techniques is metallic bismuth, which accounts for the Bi 2 O 3 component detected with EDXRF. The presence of metallic bismuth and chalcopyrite are consistent with a reducing environment. However, the pinite alteration of K-feldspar and the Fe-Cu-oxide halo wrapping chalcopyrite ( Figure 5) suggest a late alteration stage under high oxygen fugacity in the presence of water.
interwoven needle-like crystals embedded in quartz and K-feldspar ( Figure 4). An additional phase not detected with the previous techniques is metallic bismuth, which accounts for the Bi2O3 component detected with EDXRF. The presence of metallic bismuth and chalcopyrite are consistent with a reducing environment. However, the pinite alteration of K-feldspar and the Fe-Cu-oxide halo wrapping chalcopyrite ( Figure 5) suggest a late alteration stage under high oxygen fugacity in the presence of water.   Table 2, and are consistent with synchysite-(Ce), ideally CaCe(CO3)2F, and bastnäsite-(Ce), ideally Ce(CO3)F, respectively. In fact, Ce is the dominant REE in both minerals (0.43 and 0.69 a.p.f.u. in average in synchysite and bastnäsite, respectively) and the Ca/(Ca + REE) ratio is close to 0.50 for synchysite and very low in bastnäsite (0.03-0.09). Few analyses show Ca/(Ca + REE) lower than 0.50 (e.g., the n. 4 in Table 2), suggesting compositional (polysomatic) faults towards the not detected with the previous techniques is metallic bismuth, which accounts for the Bi2O3 component detected with EDXRF. The presence of metallic bismuth and chalcopyrite are consistent with a reducing environment. However, the pinite alteration of K-feldspar and the Fe-Cu-oxide halo wrapping chalcopyrite ( Figure 5) suggest a late alteration stage under high oxygen fugacity in the presence of water.   Table 2, and are consistent with synchysite-(Ce), ideally CaCe(CO3)2F, and bastnäsite-(Ce), ideally Ce(CO3)F, respectively. In fact, Ce is the dominant REE in both minerals (0.43 and 0.69 a.p.f.u. in average in synchysite and bastnäsite, respectively) and the Ca/(Ca + REE) ratio is close to 0.50 for synchysite and very low in bastnäsite (0.03-0.09). Few analyses show Ca/(Ca + REE) lower than 0.50 (e.g., the n. 4 in Table 2), suggesting compositional (polysomatic) faults towards the  Table 2, and are consistent with synchysite-(Ce), ideally CaCe(CO 3 ) 2 F, and bastnäsite-(Ce), ideally Ce(CO 3 )F, respectively. In fact, Ce is the dominant REE in both minerals (~0.43 and~0.69 a.p.f.u. in average in synchysite and bastnäsite, respectively) and the Ca/(Ca + REE) ratio is close to 0.50 for synchysite and very low in bastnäsite (0.03-0.09). Few analyses show Ca/(Ca + REE) lower than 0.50 (e.g., the n. 4 in Table 2), suggesting compositional (polysomatic) faults towards the bastnäsite composition, as commonly observed in (Ca,REE)-fluorcarbonates. Lanthanum (~0.26 and~0.50) and Nd (~0.17 and~0.33) are the other abundant REEs, whereas Y (~0.07 and~0.11) and Pr (~0.04 and 0.08) are less abundant and Sm (~0.02 and~0.04) and Gd (<0.02) are definitely scarce, especially the latter, which is sometimes below the detection limit.
Detectable amounts of Pb (~0.13 a.p.f.u.) are present in bastnäsite, whereas Pb is always below the detection limit in synchysite. Of course, CO 2 and any possible presence of H 2 O could not be detected (ideally, synchysite-(Ce) contains 27.6 wt% of CO 2 component and bastnäsite-(Ce) 20.08%). Additionally, the data show that F is lower than expected. In part this deficiency may be due to the presence of OH substituting for F in the structure, but mostly is due to diffusion of F under the highly focused electron beam, a problem further exaggerated in TEM-EDS analyses, where F is not detected at all (see ahead). In summary, taking into account that fluorine can be underestimated and the (OH) content, as calculated by difference, overestimated, the average composition of synchysite-(Ce) and bastnäsite-(Ce) read: Ca 1.00 (Ce 0. 43  .04) and Gd (<0.02) are definitely scarce, especially the latter, which is sometimes below the detection limit. Because the spatial resolution of the EDS is larger than the thickness of the lamellae, the EDS spectra of the latter contain always contribution of the surrounding matrix richer in Ca. The thinner the lamella, the higher the matrix contribution, which may explain the higher Ca/REE ratio observed in the thinner lamella.

EDS Analysis
Fluorcarbonate grains dispersed on the grid were first discerned from other more abundant grains (quartz, K-feldspar) on the basis of a fast EDS spectrum, then checked for their orientation (see ahead) and in suitable cases studied in HR mode. The final EDS analysis was taken at the end, in order to preserve the crystal for HR imaging.
TEM-EDS analyses are reported in Table 3. Overall, they are consistent with SEM-EDS analyses, although the low sensitivity of TEM-EDS did not allow for the detection of Pr, Sm, and Gd. Fluorine is also not detected because of diffusion during the analysis. In fact, in the TEM the electron beam is focused on a thin foil, the irradiated volume is therefore much lower than in the SEM and the electron dose per atom much higher. Moreover, the electron energy is an order of magnitude higher than for the SEM. Both factors determined the diffusion of the total amount of F.

[100] HRTEM Observations
HRTEM images of the 100 orientation type (hereafter [100] HR images), mostly show a zig-zag contrast which corresponds to a perfectly ordered sequence of synchysite half-cells (repeat unit~9.1 Å). The zig-zag contrast is caused by the alternating positions of fluorine atoms with respect to Ce atoms in CeF-layers ( Figure 8). Two adjoining half-cells determine the actual~18.3 Å (001) periodicity of synchysite. These observations suggest that: i) most samples are free of polysomatic disorder, and ii) any polytypic disorder, if present, can only have a component along the observation direction. The related SAED pattern, which is an average view of a larger area (in this case~0.5 µm in diameter), is streak-free (Figure 8b), suggesting that this apparent order is maintained over larger distances than those reproduced in HR images (~70 nm). The HR image above and the related simulation (Figure 8c), reveal that under optimal experimental conditions, i.e., close to Scherzer defocus (35-43 nm) and thin enough crystals (few nm), heavy atoms appear as dark fringes, which are thicker (darker) for CeF-layers (d) than for Ca-layers (f ), separated by rows of white dots, corresponding to C-layers (g). Average carbon positions appear offset along b* across the Ca-layers, whereas no shift is seen across the CeF-layers. Unfortunately, this detailed structural information could not be always achieved, because fluorcarbonates have been revealed to be very beam sensitive, and in some cases the samples could not be perfectly oriented.  (Figure 8b), suggesting that this apparent order is maintained over larger distances than those reproduced in HR images (70 nm). The HR image above and the related simulation ( Figure  8c), reveal that under optimal experimental conditions, i.e., close to Scherzer defocus (35-43 nm) and thin enough crystals (few nm), heavy atoms appear as dark fringes, which are thicker (darker) for CeF-layers (d) than for Ca-layers (f), separated by rows of white dots, corresponding to C-layers (g). The only exception to the situation described so far is the presence of sporadic polysomatic faults detected in few [100] HR images. They are revealed by the presence of a thicker dark fringe every four dark fringes, whereas in normal synchysite a thicker dark fringe occurs every five dark fringes ( Figure 9). The faulted lamellae measure~14.1 Å, which is consistent with the lack of a vaterite layer (~4.2 Å) within synchysite (~18.3 Å), therefore transforming the VBVB synchysite sequence into VBB, i.e., a single parisite half-cell. It should be noted, however, that if a bastnäsite layer (~4.9 Å) was missing, instead of a vaterite one, a slightly thinner lamella would have resulted (~13.4 Å), with sequence VVB. The difference involved is probably close to the measuring error. Unfortunately, whereas we can reasonably assume that darker fringes correspond to heavy atoms layers, we cannot confidently distinguish between Ca-layers and CeF-layers, because of the slight misalignment affecting this HR image. Therefore, the actual layer sequence cannot be demonstrated, although on the basis of the measurements, the VBB sequence seems more probable and in line with the belief that adjoining Ca-layers are metastable [18].
Minerals 2020, 10, x FOR PEER REVIEW 10 of 19 because fluorcarbonates have been revealed to be very beam sensitive, and in some cases the samples could not be perfectly oriented. The only exception to the situation described so far is the presence of sporadic polysomatic faults detected in few [100] HR images. They are revealed by the presence of a thicker dark fringe every four dark fringes, whereas in normal synchysite a thicker dark fringe occurs every five dark fringes ( Figure 9). The faulted lamellae measure 14.1 Å , which is consistent with the lack of a vaterite layer (4.2 Å ) within synchysite (18.3 Å ), therefore transforming the VBVB synchysite sequence into VBB, i.e., a single parisite half-cell. It should be noted, however, that if a bastnäsite layer (4.9 Å ) was missing, instead of a vaterite one, a slightly thinner lamella would have resulted (13.4 Å ), with sequence VVB. The difference involved is probably close to the measuring error. Unfortunately, whereas we can reasonably assume that darker fringes correspond to heavy atoms layers, we cannot confidently distinguish between Ca-layers and CeF-layers, because of the slight misalignment affecting this HR image. Therefore, the actual layer sequence cannot be demonstrated, although on the basis of the measurements, the VBB sequence seems more probable and in line with the belief that adjoining Ca-layers are metastable [18]. In the studied TEM samples, free bastnäsite grains could not be found. Bastnäsite usually forms fine lamellae at synchysite borders with coherent boundaries (Figure 10). Taking into account the poor to impossible distinguishability between certain diffraction patterns reported above, the following crystallographic relationships could be established: [100] Syn //[110] Bas ; (001) Syn //(001) Bas ; (010) Syn //(110) Bas . A higher frequency of compositional faults consistent with VBB parisite half-cells within VBVB synchysite were observed close to the contact with bastnäsite. The contrast in the HRTEM image shown in Figure 10 is characterized by rows of bright spots~9.1 Å apart, corresponding to the width of a synchysite half-cell. In several places, the distances between these rows are different from the 9.1 Å periodicity, indicating the presence of compositional faults. The fault in the center, closer to the boundary with bastnäsite, is characterized by a larger distance between neighboring bright fringes, i.e.,~14.1 Å, corresponding to an additional B-layer introduced in the regular VBVB sequence of synchysite, i.e., corresponding to a parasite half-cell. The bright fringes characterizing the faults in the upper part of the image are only~4.9 Å apart, which corresponds to the width of a B-layer. Despite the difference in contrast, also these faults correspond to additional B-layer in the regular VBVB sequence of synchysite. Indeed, as inferred from the SAED, the crystal is slightly misaligned, causing the two VB synchysite half-cells to be clearly distinguished. Therefore, a possible explanation of the different appearance of the VBB lamellae is that the extra B-layer enters the VBVB sequence in two distinguished half-cells, i.e., VBBVB vs. VBVBB. In the studied TEM samples, free bastnäsite grains could not be found. Bastnäsite usually forms fine lamellae at synchysite borders with coherent boundaries (Figure 10). Taking into account the poor to impossible distinguishability between certain diffraction patterns reported above, the following crystallographic relationships could be established:  Figure 10 is characterized by rows of bright spots 9.1 Å apart, corresponding to the width of a synchysite half-cell. In several places, the distances between these rows are different from the 9.1 Å periodicity, indicating the presence of compositional faults. The fault in the center, closer to the boundary with bastnäsite, is characterized by a larger distance between neighboring bright fringes, i.e., 14.1 Å , corresponding to an additional B-layer introduced in the regular VBVB sequence of synchysite, i.e., corresponding to a parasite half-cell. The bright fringes characterizing the faults in the upper part of the image are only 4.9 Å apart, which corresponds to the width of a B-layer. Despite the difference in contrast, also these faults correspond to additional B-layer in the regular VBVB sequence of synchysite. Indeed, as inferred from the SAED, the crystal is slightly misaligned, causing the two VB synchysite half-cells to be clearly distinguished. Therefore, a possible explanation of the different appearance of the VBB lamellae is that the extra Blayer enters the VBVB sequence in two distinguished half-cells, i.e., VBBVB vs. VBVBB.

[110] HRTEM Observations
Similarly to what has already been reported for parisite-(Ce) [7], synchysite-(Ce) imaged down 110 shows alternating bright and dark fringes that in the thinnest part of the sample and under conditions close to Scherzer defocus can be assigned to Vand B-layers, respectively (

[110] HRTEM Observations
Similarly to what has already been reported for parisite-(Ce) [7], synchysite-(Ce) imaged down 110 shows alternating bright and dark fringes that in the thinnest part of the sample and under conditions close to Scherzer defocus can be assigned to V-and B-layers, respectively (  HRTEM images of the 110 orientation type (hereafter [110] HR images), reveal that the order observed along 100 is only apparent. Whereas the (001) lattice periodicity of~18.3 Å is constant, the slant of the synchysite half-cells is highly variable (Figure 11). Basically, no ordered sequence larger than few unit cells was observed in these samples. In addition, SAED patterns along this zone are affected by streaks on hhl layers with h 3n, whereas hhl layers with h = 3n are streak-free (Figure 11b). These observations, in agreement with previous studies [7], indicate diffuse polytypic disorder and absence of polysomatic disorder. observed along 100 is only apparent. Whereas the (001) lattice periodicity of 18.3 Å is constant, the slant of the synchysite half-cells is highly variable (Figure 11). Basically, no ordered sequence larger than few unit cells was observed in these samples. In addition, SAED patterns along this zone are affected by streaks on hhl layers with h  3n, whereas hhl layers with h = 3n are streak-free ( Figure  11b). These observations, in agreement with previous studies [7], indicate diffuse polytypic disorder and absence of polysomatic disorder. Consistently, the HR image shows (002) half-cells with varying slants, but constant interplanar distance (lower left). Therefore, the VBVB sequence is maintained throughout the crystal, as well as the composition, but the relative orientation of the VB-layers within the sequence may be different from cell to cell. The white boxes in (a) include simulations for two different thickness and defocus conditions, which are in part reproduced and magnified in (c) (thickness = 15 nm, defocus = 43 nm, atomic vibrations = 0.05 nm) and (d) (thickness = 3.5 nm, defocus = 45 nm, atomic vibrations = 0.09 nm), respectively. Although disordered and affected by beam damage, the HR images allow recognition of vaterite (V) and bastnäsite (B) layers as alternating bright and dark fringes. Note, however, as dark and bright fringes invert contrast with increasing thickness of the thin foil (top and bottom arrows in (a) are aligned vertically). Consistently, the HR image shows (002) half-cells with varying slants, but constant interplanar distance (lower left). Therefore, the VBVB sequence is maintained throughout the crystal, as well as the composition, but the relative orientation of the VB-layers within the sequence may be different from cell to cell. The white boxes in (a) include simulations for two different thickness and defocus conditions, which are in part reproduced and magnified in (c) (thickness = 15 nm, defocus = 43 nm, atomic vibrations = 0.05 nm) and (d) (thickness = 3.5 nm, defocus = 45 nm, atomic vibrations = 0.09 nm), respectively. Although disordered and affected by beam damage, the HR images allow recognition of vaterite (V) and bastnäsite (B) layers as alternating bright and dark fringes. Note, however, as dark and bright fringes invert contrast with increasing thickness of the thin foil (top and bottom arrows in (a) are aligned vertically).

Rock Mineral Association
(Ca,REE)-fluorcarbonates from Cinquevalli are hosted in quartz-dikes with granitic texture made of quartz (79.4%), K-feldspar (6.9%), and chalcopyrite (1.4%) as major, primary, constituents. Quartz is present as: i) equant grains intergrown with (Ca,REE)-fluorcarbonates and ii) as in fibrous crystals recalling chalcedony and unrelated to fluorcarbonates. Muscovite (9.5%), possibly overestimated by Rietveld refinement because of (001) preferential orientation, has not been observed by optical or electron microscopy, but in the alteration of K-feldspar as white mica (pinite). Alteration also affects chalcopyrite, which shows an Fe-Cu-oxide rim. Metallic bismuth, detected by SEM-EDS, completes the mineral association.
These observations suggest that the crystallization of the quartz-dikes was a multistage event, at least judging from the two generations of quartz. Chalcedony possibly formed in a later stage under higher undercooling conditions. The graphic texture of the intergrowth between primary quartz and (Ca,REE)-fluorcarbonates suggest a simultaneous crystallization of quartz and fluorcarbonates. The presence of metallic bismuth and chalcopyrite are consistent with a reducing environment during the first stage of crystallization, but the pinite alteration of K-feldspar and the oxide rim wrapping chalcopyrite suggest a late alteration stage under higher oxygen fugacity and in the presence of water.
SEM-EDS analyses reveal that synchysite and bastnäsite are Ce-rich, with significant amounts of La and Nd, lower amounts of Y and Pr, and minor amounts of Sm and Gd. In total, in the REE-oxides amount to~2.8 wt% of the rock and are confined in fluorcarbonates finely intergrown with quartz. The small concentration and the dispersion within a resilient silicate matrix make this occurrence scarcely interesting for the ore mineral industry.

Polytypism of Synchysite
The crystal structure of synchysite-(Ce) has been refined and described in the C2/c space group [17]. The synchysite structure consists of Ca-layers at z = 0 and z = 1 2 , alternating with CeF-layers at z = 1 4 and z = 3 4 , all separated by intercalated CO 3 -layers. The heavy atoms form columns of Ca and Ce alternating along c* in a hexagonal stacking, as formerly predicted [16]. To accomplish chemical bonding, the CO 3 -layers above z = 1 2 are shifted by~2.37 Å along [110] and those above z = 1 by another 2.37 Å along [110] with respect to the position that they should have in bastnäsite, therefore breaking the hexagonal symmetry and leading to monoclinic symmetry. The overall shift of the CO 3 -layers along a is~4.11 Å, leading to a monoclinic angle of~102.7 • .
It can be easily realized that the layered structure of the (Ca,REE)-fluorcarbonates is at the origin of syntaxial intergrowths with a theoretically unlimited number of polysomatic and polytypic stacking variants. In the studied samples, polysomatic faults are sporadic, but polytypic disorder is widespread. Basically, no ordered domain, i.e., free of polytypic faults, larger than few unit cells was found. Polytypism results from the insertion of a Ca-layer between bastnäsite portions and the different types of linkages it can form with the bastnäsite portions above and below it [18]. As discussed above, in ordered synchysite-(Ce), the operating stacking vectors between CO 3 -layers in alternating bastnäsite portions are [110] and [110], respectively, which are related by ±60 • rotations ( Figure 12). A consistent bonding pattern, however, is also obtained for shifts in the second bastnäsite portion along [010] and [110], which are rotated by -60 • and 180 • , respectively, from the former shift (i.e., [110]). In contrast, shift directions rotated either 0 • or ±120 • relative to the first stacking vector (n·60 • , n = even) would put the CO 3 groups aligned with the heavy atoms columns along c*, impeding a consistent bonding pattern. Therefore, all polytypes and polytypic faults in synchysite-(Ce) are based on n·60 • (n = odd) shifts.
The complete set of all possible shifts and how they are recognized in HR images has been derived for parisite [7] and here reproduced for synchysite (Table 4). Such shifts are common to all (Ca,REE)-fluorcarbonates, once a consistent monoclinic unit cell is chosen.
When the structure is seen down [100], the [110] and [110] shifts would appear as -b/6 and +b/6 translations in HR images (Table 4), leading to a zig-zag sequence, as actually observed in Figure 8. Any polytypic departure from the ordered sequence will move the structure towards the allowed [010] shift, if occurring after the [110] one, or [010], if occurring after [110]. Both faults involve translation of b/3. Because of the half-cell periodicity due to the C-centring of the structure, a shift of +b/3 will appear as a -b/6 shift on [100] HR images, and a shift of -b/3 as a +b/6 shift, therefore undistinguishable from the structural changes expected in an ordered sequence. A similar analysis can be done for any of the 100 -type indistinguishable directions.
As can be seen in [110] HRTEM images, synchysite-(Ce) from Cinquevalli is extensively faulted. In [100] HR images, however, these faults cannot be seen, due to the equality between the projected shifts in regularly stacked sequences and in faulted sequences (and not because it has component along the observation direction only as assumed at first glance). When synchysite is observed along  Table 4).
The complete set of all possible shifts and how they are recognized in HR images has been derived for parisite [7] and here reproduced for synchysite (Table 4). Such shifts are common to all (Ca,REE)-fluorcarbonates, once a consistent monoclinic unit cell is chosen. [110] -a/6 0 a/6 a/6 0 -a/6 [100] [110] 0 a/6 a/6 0 -a/6 -a/6 [010] -a/6 -a/6 0 a/6 a/6 0 When the structure is seen down [100], the [110] and [110] shifts would appear as -b/6 and +b/6 translations in HR images (Table 4), leading to a zig-zag sequence, as actually observed in Figure 8. Any polytypic departure from the ordered sequence will move the structure towards the allowed [010] shift, if occurring after the [110] one, or [010], if occurring after [110]. Both faults involve translation of b/3. Because of the half-cell periodicity due to the C-centring of the structure, a shift of +b/3 will appear as a -b/6 shift on [100] HR images, and a shift of -b/3 as a +b/6 shift, therefore undistinguishable from the structural changes expected in an ordered sequence. A similar analysis can be done for any of the 100-type indistinguishable directions.
As can be seen in [110] HRTEM images, synchysite-(Ce) from Cinquevalli is extensively faulted. In [100] HR images, however, these faults cannot be seen, due to the equality between the projected shifts in regularly stacked sequences and in faulted sequences (and not because it has component along the observation direction only as assumed at first glance). When synchysite is observed along   Table 4).

A Brief Comparison with the Polytypism of Mica
The stacking sequence of synchysite-(Ce) is analogous to that of 2:1 mica, such as muscovite, if the CO 3 -layer is equated to the (Si,Al)O 4 -layer, the CeF-layer to the AlO 5 (OH)-layer, and the Ca-layer to the K-layer (interlayer) [17]. Mica polytypes result from the translation of the (Si,Al)O 4 -layers relative to each other, whereas synchysite polytypes, from the translation of CO 3 -layers above the Ca-layer.

A Brief Comparison with the Polytypism of Mica
The stacking sequence of synchysite-(Ce) is analogous to that of 2:1 mica, such as muscovite, if the CO3-layer is equated to the (Si,Al)O4-layer, the CeF-layer to the AlO5(OH)-layer, and the Ca-layer to the K-layer (interlayer) [17]. Mica polytypes result from the translation of the (Si,Al)O4-layers relative to each other, whereas synchysite polytypes, from the translation of CO3-layers above the Calayer.

6H +60°
Possible, but never observed as ordered single crystal.
* For a comprehensive review of mica-polytypes see Reference [57].

A Brief Comparison with the Polytypism of Mica
The stacking sequence of synchysite-(Ce) is analogous to that of 2:1 mica, such as muscovite, if the CO3-layer is equated to the (Si,Al)O4-layer, the CeF-layer to the AlO5(OH)-layer, and the Ca-layer to the K-layer (interlayer) [17]. Mica polytypes result from the translation of the (Si,Al)O4-layers relative to each other, whereas synchysite polytypes, from the translation of CO3-layers above the Calayer.

6H +60°
Possible, but never observed as ordered single crystal.
* For a comprehensive review of mica-polytypes see Reference [57].

A Brief Comparison with the Polytypism of Mica
The stacking sequence of synchysite-(Ce) is analogous to that of 2:1 mica, such as muscovite, if the CO3-layer is equated to the (Si,Al)O4-layer, the CeF-layer to the AlO5(OH)-layer, and the Ca-layer to the K-layer (interlayer) [17]. Mica polytypes result from the translation of the (Si,Al)O4-layers relative to each other, whereas synchysite polytypes, from the translation of CO3-layers above the Calayer.

6H +60°
Possible, but never observed as ordered single crystal.
* For a comprehensive review of mica-polytypes see Reference [57].
Synchysite-(Ce) is characterized by 60° rotations of subsequent bastnäsite portions above the Ca-layers, therefore can be likened to the 2M2 subfamily-B mica polytype, such as lepidolite, for instance. Hypothetical subfamily-A (Ca,REE)-fluorcarbonates, since entailing rotations of 0 or 120° that would align the vertical edges of the CO3 triangles above the heavy atoms columns have, as Paragonite [46] Polylithionite [47] Muscovite [48] Should not be possible

A Brief Comparison with the Polytypism of Mica
The stacking sequence of synchysite-(Ce) is analogous to that of 2:1 mica, such as muscovite, if the CO3-layer is equated to the (Si,Al)O4-layer, the CeF-layer to the AlO5(OH)-layer, and the Ca-layer to the K-layer (interlayer) [17]. Mica polytypes result from the translation of the (Si,Al)O4-layers relative to each other, whereas synchysite polytypes, from the translation of CO3-layers above the Calayer.

6H +60°
Possible, but never observed as ordered single crystal.
* For a comprehensive review of mica-polytypes see Reference [57].

A Brief Comparison with the Polytypism of Mica
The stacking sequence of synchysite-(Ce) is analogous to that of 2:1 mica, such as muscovite, if the CO3-layer is equated to the (Si,Al)O4-layer, the CeF-layer to the AlO5(OH)-layer, and the Ca-layer to the K-layer (interlayer) [17]. Mica polytypes result from the translation of the (Si,Al)O4-layers relative to each other, whereas synchysite polytypes, from the translation of CO3-layers above the Calayer.

6H +60°
Possible, but never observed as ordered single crystal.
* For a comprehensive review of mica-polytypes see Reference [57].
Synchysite-(Ce) is characterized by 60° rotations of subsequent bastnäsite portions above the Ca-layers, therefore can be likened to the 2M2 subfamily-B mica polytype, such as lepidolite, for instance. Hypothetical subfamily-A (Ca,REE)-fluorcarbonates, since entailing rotations of 0 or 120° that would align the vertical edges of the CO3 triangles above the heavy atoms columns have, as Anandite [55] Phlogopite [56] Possible, few unit cells observed at the TEM scale (this work, Figure 11).

A Brief Comparison with the Polytypism of Mica
The stacking sequence of synchysite-(Ce) is analogous to that of 2:1 mica, such as muscovite, if the CO3-layer is equated to the (Si,Al)O4-layer, the CeF-layer to the AlO5(OH)-layer, and the Ca-layer to the K-layer (interlayer) [17]. Mica polytypes result from the translation of the (Si,Al)O4-layers relative to each other, whereas synchysite polytypes, from the translation of CO3-layers above the Calayer.

6H +60°
Possible, but never observed as ordered single crystal.
* For a comprehensive review of mica-polytypes see Reference [57].
Synchysite-(Ce) is characterized by 60° rotations of subsequent bastnäsite portions above the Ca-layers, therefore can be likened to the 2M2 subfamily-B mica polytype, such as lepidolite, for instance. Hypothetical subfamily-A (Ca,REE)-fluorcarbonates, since entailing rotations of 0 or 120° that would align the vertical edges of the CO3 triangles above the heavy atoms columns have, as Possible, but never observed as ordered single crystal.
Synchysite-(Ce) is characterized by ±60 • rotations of subsequent bastnäsite portions above the Ca-layers, therefore can be likened to the 2M 2 subfamily-B mica polytype, such as lepidolite, for instance. Hypothetical subfamily-A (Ca,REE)-fluorcarbonates, since entailing rotations of 0 or ±120 • that would align the vertical edges of the CO 3 triangles above the heavy atoms columns have, as expected, never been observed. In contrast, subfamily-A mica polytypes are even more abundant in nature than subfamily-B mica polytypes. This can be ascribed to the different atomic arrangement in the interlayer. In subfamily-B polytypes, basal tetrahedral oxygens on the adjacent layers are exactly superimposed along the c* direction. In subfamily-A polytypes, these basal oxygens are laterally displaced from each other due to ditrigonal distortion of the tetrahedral rings, thereby reducing the repulsive forces between the oxygens facing across the interlayer and making the structural configuration energetically more favorable [58].
Parisite-(Ce), the other (Ca,REE)-fluorcarbonate whose structure has been refined [18], should be considered again as belonging to the 2M 2 polytype, since it is characterized by the same stacking vectors as synchysite-(Ce). Other possible short range (Ca,REE)-fluorcarbonate polytypes should be the 2O and the 6H. The 2O polytype has been observed as a few unit cells repeats at the TEM scale ( Figure 11), whereas the 6H has not yet been clearly characterized, not even at the TEM scale.