Mineralogy, Geochemistry, and Genesis of Agates from Chihuahua, Northern Mexico

: The present study aimed to investigate the genesis and characteristics of some of the world-famous agate deposits in the state of Chihuahua, Mexico (Rancho Coyamito, Ojo Laguna, Moctezuma, Huevos del Diablo, Agua Nueva). Geochemical and textural studies of host rocks showed that all the studied deposits are related to the same rock type within the geological unit of Rancho el Agate andesite, a quartz-free latite that shows clear indications of magma mixing. As a result of their large-scale distribution and various differentiation processes, as well as transport separation, different textures and local chemical differences between rocks of different localities can be observed. These differences have also inﬂuenced the properties of SiO 2 mineralization in the rocks. The mixing of near-surface ﬂuids from rock alterations with magmatic hydrothermal solutions led to the accumulation of various elements in the SiO 2 matrix of the agates, which were, on the one hand, mobilized during secondary rock alteration (Fe, U, Ca, K, Al, Si) and, on the other hand, transported with magmatic ﬂuids (Zn, Sb, Si, Zr, Cr). Different generations of chalcedony indicate a multi-stage formation as well as multiple cycles of ﬁlling the cavities with ﬂuids. The hydrothermal ﬂuids are presumably related to the residual solutions of a rhyolitic volcanism, which followed the latitic extrusions in the area and probably caused the formation of polymetallic ore deposits in the Chihuahua area. The enrichment of highly immobile elements indicates the involvement of volatile ﬂuids in the agate formation. The vivid colors of the agates are almost exclusively due to various mineral inclusions, which consist mainly of iron compounds.


Introduction
Agates from the state of Chihuahua in northern Mexico are some of the most beautiful and most expensive agates worldwide and are, therefore, of great importance, especially for the collector market and the jewelry industry.They occur in the Chihuahuan Desert in spatially and often widely separated deposits, mainly within the voluminous, intermediate volcanic units of Sierra del Gallego [1].In these locations, they are mostly exploited commercially by private, small-scale miners or small companies and are of considerable economic importance because of their high prices in the world market.
Despite the high prominence of the deposits, only a few studies have been carried out on the exact characteristics and genesis of their silica mineralization.Therefore, the present study aimed to carry out a comprehensive geological and mineralogical investigation into some of the most important agate deposits in the Chihuahua region for the first time.The chemical and mineralogical compositions of the host rocks, as well as their Minerals 2023, 13, 687 2 of 31 textures and alterations, were used to assign the individual deposits to geological units and to check whether all of the agates in the region could be related to the same volcanic event.Through various mineralogical and geochemical analyses of the agate samples, conclusions can be drawn about the formation mechanism and the relative formation time of the mineralization in the area in general and in the individual deposits.The aim was to reconstruct the geological and mineralogical-geochemical processes from the actual data and to develop a model that clarifies the mechanisms of agate formation in northern Chihuahua in more detail.

Geological Background
From a physiographic point of view, the agate-rich deposits that were studied are located in the area of the flat ranges and plains of northern Mexico, between the Sierra Madre Occidental to the west and the Sierra Madre Oriental to the east [2].This area is primarily characterized by NNW-trending mountain ranges of Cretaceous sediments and Tertiary volcanic rocks, separated by broad, endorheic, intramontane basins, so-called 'bolsons' [3].This region is located at an elevation of 1200 m to 2400 m above sea level, with the terrain rising towards the south.Because of the predominantly arid desert climate, a typical cuesta landscape with extensive pediments and steep fault terraces formed, resulting in the formation of extensive basins between the escarpments during the predominant weathering of soft rock units [3].Drainage in the near-perennially dry area is directed primarily into the intramontane basins during the short monsoon period in summer, sometimes resulting in nondurable lakes in late summer [4].
The geology of the studied area is dominated by late Mesozoic to Cenozoic volcanic units of the El Sueco-Sierra del Gallego complex (Figure 1), which is located between the calc-alkaline volcanic provinces of the Sierra Madre Occidental and the alkaline volcanic province of Trans-Pecos [3,5].Because of its geographic location and specific geochemical composition, it represents the intermediate igneous province between these two major complexes [5].Within this area, the volcanic complex is about 1200 m thick and can be subdivided into four major units according to Bockoven [3]: (1) Tuffs and volcanoclastic sediments of the Libres Formation; (2) Voluminous intermediate-to-mafic lava flows of the Rancho el Agate andesite; (3) Rhyolitic units with lava flows, domes, and veins (Mesteño, Gallego, Agua Nueva, Carneros, and El Dos); (4) Basalt flows and flow tuffs of the Milagro Basalt (compare Figure 1).
The subalkaline volcanism was dated to between 45 and 29 million years by Keller [5], with no volcanic activity for an extended period between 36 and 29.5 million years ago.No directly determined age data are available for the Rancho el Agate andesite, which is particularly relevant for this work.However, based on the potassium argon ages (K-Ar) of the overlying and underlying units, an age between 39 and 38 million years can be assumed [5].
Field observations indicate that all deposits are related to the same volcanic unit, probably even to the same lava flow of the so-called Rancho el Agate andesite.The agates occur mainly in former vesicular cavities, where they are paragenetically associated with iron oxides/hydroxides and carbonates.Pseudomorphs are also not uncommon.The current investigations indicated direct, hydrothermal silicification processes as well as secondary silicification by SiO 2 , which was released during the host rock alteration.Figure 1.Stratigraphic profile of the volcanic complex in the Sierra del Gallego area.It shows the relative position and average thickness as well as the K-Ar ages of the rock units [in Ma]; (modified after [5]).

Materials and Methods
The samples were collected during a field trip in 2020.The studied deposits (Figure 2) are specifically located in Agua Nueva (AN), Rancho Coyamito (Coy), Moctezuma (Mo), Huevos del Diabolo (HdD), and Arcoiris claim near Ojo Laguna (OL), with Rancho Coyamito being sampled at three different claims (the Los Alamos claim (CLA), the Japanese deposit (CLJ), and the La Fortuna claim (CLF)) and the Agua Nueva deposit at two locations (AN, ANZ).For each deposit, the studied material consists of both the representative host rock and associated agate and chalcedony samples.The mineralization and sample points in the agates were always considered to be representative for the locality.If possible, all visibly distinguishable chalcedony or quartz generations of different samples were separated in order to be able to compare them within the individual agate samples and between the individual deposits.The color, shape, and type of banding as well as the preparability of the individual chalcedony bands were considered during the selection.
Eight different rock samples (Table 1) were crushed manually to the size of <400 µm using a steel mortar.In addition, individual macroscopically recognizable bands of the agates were selected primarily according to color and texture criteria and then separated (Table 2).After the rough crushing, a McCrone percussion mortar with ZrO2 grinding

Materials and Methods
The samples were collected during a field trip in 2020.The studied deposits (Figure 2) are specifically located in Agua Nueva (AN), Rancho Coyamito (Coy), Moctezuma (Mo), Huevos del Diabolo (HdD), and Arcoiris claim near Ojo Laguna (OL), with Rancho Coyamito being sampled at three different claims (the Los Alamos claim (CLA), the Japanese deposit (CLJ), and the La Fortuna claim (CLF)) and the Agua Nueva deposit at two locations (AN, ANZ).For each deposit, the studied material consists of both the representative host rock and associated agate and chalcedony samples.The mineralization and sample points in the agates were always considered to be representative for the locality.If possible, all visibly distinguishable chalcedony or quartz generations of different samples were separated in order to be able to compare them within the individual agate samples and between the individual deposits.The color, shape, and type of banding as well as the preparability of the individual chalcedony bands were considered during the selection.
Eight different rock samples (Table 1) were crushed manually to the size of <400 µm using a steel mortar.In addition, individual macroscopically recognizable bands of the agates were selected primarily according to color and texture criteria and then separated (Table 2).After the rough crushing, a McCrone percussion mortar with ZrO 2 grinding elements was used.The crushed agate samples were selected by hand under binoculars and the single-colored bands, i.e., sample points, were separated.Subsequently, the separated subsamples were milled by hand in an agate mortar.
Minerals 2023, 13, x FOR PEER REVIEW 4 of 32 elements was used.The crushed agate samples were selected by hand under binoculars and the single-colored bands, i.e., sample points, were separated.Subsequently, the separated subsamples were milled by hand in an agate mortar.
Figure 2. Map of agate deposits in the northern part of the state of Chihuahua.Marked are the deposits studied in this work and the Rancho el Agate Ranch, which, according to Bockoven [3], is to be seen as the main source area of Rancho el Agate andesite; (modified after [6]).The mineral composition of the agate host rocks was analyzed using the X-ray diffractometer Orion Comet P2 (XRD-Eigenmann GmbH, SEIFERT Analytical X-ray) with Co-Kα radiation and the semiconductor detector Meteor OD in the measuring range from 5 to 80 • 2-Theta.A total of 20 wt.% corundum was added to each of the powdered rock samples as an internal standard for the quantification of X-ray amorphous phases.For single mineral measurements, separate mineral grains were crushed in an agate mortar and prepared on a polished silicon carrier using technical ethanol (96%).Clay mineral investigations were carried out on special texture preparations using the sedimentation method on a glass carrier.Measurements were taken on a X-ray diffractometer HZG4 (Seifert/Freiberger Präzisionsmechanik, Freiberg, Germany) using Cu-Kα radiation and a scintillation counter as the detector in the measuring range from 2 to 13 • 2-theta.The evaluation of all measurements was performed using the Profex software [7].Polished thin sections (30 µm) of the rocks and of one agate per occurrence were prepared.These thin sections were examined using transmitted light microscopy on an Axio Imager.A1m polarization microscope, equipped with an AxioCam color 105 and the AxioVision software (ZEISS, Thornwood, NY, USA).Studies using scanning electron microscopes (SE, BSE, EDS) were also performed on the polished and carbon-coated thin sections.These investigations were performed using a JEOL 6400 SEM with EDX detector (JEOL Ltd., Akishima, Japan).
Cathodoluminescence investigations (CL) using a "hot cathode" cathodoluminescence microscope HC1-LM (LUMIC, Bochum, Germany) were performed on the polished and carbon-coated thin sections of agates and rock samples [8].This instrument was operated at a 14 kV accelerating voltage and a current of 0.2 mA (10 µA/mm 2 ).Images were acquired using a Peltier-cooled digital video camera (OLYMPUS DP72, OLYMPUS Deutschland GmbH, Hamburg, Germany).CL spectra in the range of 370 to 1000 nm were measured under standardized conditions using an Acton Research SP-2356 digital spectrograph equipped with a Princeton Spec-10 CCD detector and WinSpec software (OLYMPUS Deutschland GmbH, Hamburg, Germany).Wavelength calibration was performed using an Hg-halogen lamp, a spot width of 30 µm, and a 5 s measuring time.
The precrushed and divided rock samples were milled (<63 µm) in an agate grinding pot using a Retsch vibratory disk mill.The powders were dried overnight at 105 • C, then annealed for 2 h at 950 • C to determine the loss on ignition.Subsequently, homogeneous fused beads were prepared from 1 g of the annealed sample powder and 8 g of lithium tetraborate.Measurements were carried out using a wavelength dispersive PANalyticalAxios Minerals spectrometer and WROXI package (PANalytical, Almelo, The Netherlands).
Trace-element analyses were carried out using 100 mg of powdered rock and agate samples dissolved with a hydrofluoric acid digestion containing nitric acid and mixed with 1 mL rhenium solution (100 µg/L) as an internal standard.Measurements were taken with a Perkin Elmer Sciex Elan 5000 quadrupole mass spectrometer (ICP-MS) with a crossflow nebulizer and a rhyton spray chamber (Perkin Elmer Inc., Baesweiler, Germany) [9].
The paramagnetic centers of the powder agate samples were analyzed using electron paramagnetic resonance (EPR) spectroscopy with a Bruker EMX spectrometer (Saskatchewan Structural Science Centre, Saskatoon, Canada).Measurements were performed on ~100 mg powdered sample material at both RT (295 K) and 30 K. RT-EPR measurements were operated at a microwave frequency of ~9.86 GHz, using three sets of experimental conditions: (1) a wide scan in the magnetic field of 0-700 mT at a microwave power of ~2.0 mW, a modulation frequency of 100 KHz, a modulation amplitude of 0.5 mT, and a spectral resolution of 0.05 mT; (2) a narrow scan in the magnetic field of 335-360 mT at a microwave power of ~0.21 mW, a modulation frequency of 100 KHz, a modulation amplitude of 0.1 mT, and a spectral resolution of 0.01 mT; and (3) a narrow scan in the magnetic field of 320-360 mT using a microwave power of ~21.0 mW, a modulation frequency of 100 KHz, a modulation amplitude of 0.1 mT, and a spectral resolution of 0.01 mT.
Powder EPR measurements at 30 K were taken on the same spectrometer equipped with an Oxford cryostat using liquid helium at a microwave frequency of ~9.38 GHz with two sets of experimental conditions: (1) a scan range in the magnetic field of 320-370 mT at a microwave power of ~0.21 mW, a modulation frequency of 100 KHz, a modulation amplitude of 0.1 mT, and a spectral resolution of 0.01 mT; and (2) a scan range in the magnetic field of 320-370 mT at a microwave power of ~21.0 mW, a modulation frequency of 100 KHz, a modulation amplitude of 0.1 mT, and a spectral resolution of 0.01 mT.
Raman spectroscopy was performed using an XplorRa-Plus Raman spectrometer (Horiba France SAS, Palaiseau, France) operated with a 532 nm laser.The measurements were taken on polished thin sections at 100× objective magnification in the ranges of 0 cm −1 to 1500 cm −1 and 3020 cm −1 to 4100 cm −1 Raman shift.In both intervals, 200 spectra with a 1 s acquisition time each were averaged per measurement point.

Chemistry
The results of the X-ray fluorescence analyses are shown in Table 3.The loss on ignition of each sample ranged from 2.5 wt.% to 4.3 wt.%.All of the host rocks showed very similar chemical compositions, which were characterized by an intermediate SiO 2 content, a high iron content, and a constant alkali metal content.Only the sample from Ojo Laguna (OL) showed a significantly higher SiO 2 content.A clear correlation among the main elements was revealed exclusively in the negative correlation of the calciumpotassium ratio.The commonly used TAS classifications of Cox [10], LeBas [11], and Middlemost [12] classify almost all of the investigated samples as intermediate rocks in the trachyandesite field (Figure 3).Thus, the rocks plot near the boundary between alkaline and subalkaline volcanism, with the stronger subalkaline influence becoming particularly evident in the classification of LeBas [11].A similar classification of the rocks as trachyandesites is shown in the Zr/TiO 2 -SiO 2 plot [13] and in the Nb/Y-Zr/Ti plot [14], with the rocks in the latter being classified at the andesite/trachyandesite boundary.A more exact subdivision is possible based on Na-K ratios, in which, according to Shelley [15], all examined trachyandesites are to be classified as latites.This also indicates the R1-R2 classifications [16], in which the samples are classified among the trachyandesites predominantly as latites and quartz-latites.Ultimately, this classification can also be confirmed by the purely trace-element-based classifications, as in the Co-Th plot [17], in which the rocks plot in the field of strongly potassic, calc-alkaline latites.One exception in the common chemical classifications is the host rock of Ojo Laguna (OL).It can be classified as either dacite [11,12], rhyodacite [16], or very acidic andesite [10] according to the TAS and R1-R2 classifications, primarily because of its higher SiO 2 content and low alkali metal content.

Phase Composition
Both qualitative and quantitative X-ray phase analyses were performed on all powder samples of the rocks.Texture preparations were also made to better distinguish and identify the clay minerals in the samples.Qualitatively, the different rocks exhibited the same mineral composition and a significant amorphous content (Table 4).
The clay mineral analyses of the texture preparations showed a clear and symmetrical reflex in the range of 15.0 Å to 15.5 Å in the untreated (air-dried) samples.In addition to this dominant signal, there was a peak at 10.0 Å in almost all samples.After treatment with ethylene glycol, a change in the position of the peaks at about 1.5 Å was observed, which is due to the swelling of the sheet silicate layers.In addition, an indistinct, broad peak appeared in the 8.3 Å to 8.5 Å range.No changes could be seen at the 10 Å peaks.After treatment at 400 °C, peaks in the small angle region disappeared completely and intensities at 10 Å increased.In the samples in which no peak had previously been present

Phase Composition
Both qualitative and quantitative X-ray phase analyses were performed on all powder samples of the rocks.Texture preparations were also made to better distinguish and identify the clay minerals in the samples.Qualitatively, the different rocks exhibited the same mineral composition and a significant amorphous content (Table 4).
The clay mineral analyses of the texture preparations showed a clear and symmetrical reflex in the range of 15.0 Å to 15.5 Å in the untreated (air-dried) samples.In addition to this dominant signal, there was a peak at 10.0 Å in almost all samples.After treatment with ethylene glycol, a change in the position of the peaks at about 1.5 Å was observed, which is due to the swelling of the sheet silicate layers.In addition, an indistinct, broad peak appeared in the 8.3 Å to 8.5 Å range.No changes could be seen at the 10 Å peaks.After treatment at 400 • C, peaks in the small angle region disappeared completely and intensities at 10 Å increased.In the samples in which no peak had previously been present in this area, a new, broad peak appeared.Judging from the X-ray diffractograms of the HdD sample, which show the typical 006 peak at 1.509 Å [18], and the microscopic observations, the 10 Å mica is celadonite.This is also supported by the missing 002 peak, which is suppressed by the strong scattering of the octahedral iron in celadonite [18].The typical behaviors of swellable clay minerals indicate that they are smectites, which does not contain significant mixed-layer components [19].The 006 peak, which was dominant in the 1.49 Å to 1.50 Å range, shows that the samples contained predominantly dioctahedral smectite [18].Only in sample OL can a trioctahedral smectite be assumed from the distinct 006 peak in the range of 1.545 Å [18].Most of the rocks were rich in phenocrysts, while only two of the samples were completely or nearly aphyric.The majority displayed a rather uniform appearance with a porphyric-dominated texture.They comprised partly large phenocrysts of fresh plagioclase as well as single feldspar crystals, which showed a distinct resorption structure.The rock samples from the central sites of Rancho Coyamito exhibited a fine-grained matrix with squat plagioclase crystals, which were on average 60-70 µm in size.The matrix structure was predominantly intergranular with a strong tendency to be intersertal and transitioned in some areas to felsitic, especially in the CLA sample (Figure 4c).Swallowtail plagioclase was concentrated in some zones inside the matrix of sample CLJ, where it was also combined with skeletal crystals.In the sample CLA, neither the swallowtail plagioclase nor the skeletal crystals occur.The samples from Moctezuma (Mo) and Agua Nueva 2 (ANZ) had intergranular-to intersertal-dominated matrix that included feldspar crystals at an average size of 100 µm (Figure 4b).Swallowtail plagioclase was found only very subordinately in the matrix of these samples.
The matrix of sample OL was also predominantly intersertal, only subordinately intergranular, and of a mainly red appearance (Figure 4d).Many of the matrix feldspars were almost exclusively swallowtail plagioclase or skeletal crystals and were on average 100 µm Minerals 2023, 13, 687 9 of 31 in length.The matrix of Agua Nueva, as well as that of Ojo Laguna, was intergranular to strongly intersertal and red in appearance.
In general, plagioclase phenocrysts within the rocks of Sierra del Gallego were found mainly as fresh, single-crystal individuals of up to 10 mm in size (Ø 1.5 mm) (feldspar generation 1) (Figure 5a).In the central areas of Rancho Coyamito and Agua Nueva, glomeroporphyritic aggregates as well as synneutic intergrowth occurred predominantly.In marginal deposits, these aggregates either did not appear (Mo) or occurred only subordinately in smaller agglomerates (OL), and phenocrysts were usually smaller and isolated.Synneutic intergrowth occurred less frequently.

R PEER REVIEW 10 of 32
In general, plagioclase phenocrysts within the rocks of Sierra del Gallego were found mainly as fresh, single-crystal individuals of up to 10 mm in size (Ø 1.5 mm) (feldspar generation 1) (Figure 5a).In the central areas of Rancho Coyamito and Agua Nueva, glomeroporphyritic aggregates as well as synneutic intergrowth occurred predominantly.In marginal deposits, these aggregates either did not appear (Mo) or occurred only subordinately in smaller agglomerates (OL), and phenocrysts were usually smaller and isolated.Synneutic intergrowth occurred less frequently.Glomeroporphyritic aggregates, in particular, usually showed no or only very weak resorption phenomena on feldspar crystals and often contained a lot of pyroxene.In contrast, some plagioclase phenocrysts occurred subordinately with only a small fresh area in the core of the crystal.The outer areas were completely altered and showed a typical fine-grained sieve structure (feldspar generation 2) (Figure 5b).A few of these crystals also had large embayments.A third feldspar generation was characterized by similar intense resorption phenomena and was often simple twinned.The crystals formed mostly idiomorphic individuals, with their edges apparently formed by epitaxially regrown feldspar.Resorption phenomena subordinately caused the typical fine-grained sieve structures but appeared much more frequently as large embayments (Figure 5c).Within the embayments, mostly glassy-dominated areas or crystallized matrix could be found.Large phenocrysts, in particular, sometimes showed several generations of resorption and epitaxial growth.Many feldspar crystals also possessed pronounced oscillatory zoning, which was still evident in even the highly resorbed individuals.In addition to the feldspars, hypidiomorphic-to-idiomorphic clinopyroxenes with sizes of up to 1000 µm (Ø 250 µm-400 µm) commonly occurred in the central deposits.
The pyroxenes were mostly untwinned and commonly found as inclusions in the Glomeroporphyritic aggregates, in particular, usually showed no or only very weak resorption phenomena on feldspar crystals and often contained a lot of pyroxene.In contrast, some plagioclase phenocrysts occurred subordinately with only a small fresh area in the core of the crystal.The outer areas were completely altered and showed a typical fine-grained sieve structure (feldspar generation 2) (Figure 5b).A few of these crystals also had large embayments.A third feldspar generation was characterized by similar intense resorption phenomena and was often simple twinned.The crystals formed mostly idiomorphic individuals, with their edges apparently formed by epitaxially regrown feldspar.Resorption phenomena subordinately caused the typical fine-grained sieve structures but appeared much more frequently as large embayments (Figure 5c).Within the embayments, mostly glassy-dominated areas or crystallized matrix could be found.Large phenocrysts, in particular, sometimes showed several generations of resorption and epitaxial growth.Many feldspar crystals also possessed pronounced oscillatory zoning, which was still evident in even the highly resorbed individuals.In addition to the feldspars, hypidiomorphic-to-idiomorphic clinopyroxenes with sizes of up to 1000 µm (Ø 250 µm-400 µm) commonly occurred in the central deposits.
The pyroxenes were mostly untwinned and commonly found as inclusions in the resorbed feldspar phenocrysts.While only a few crystals were fresh, most showed clear signs of alteration at the rim or along the cleavage planes.Some larger individuals were already completely altered, while others showed an incipient uralitization.Apatite occurred predominantly in long prismatic-to-acicular idiomorphic crystals in all of the porphyritic samples.It was found in the matrix as well as poikilitic in the phenocrysts.Overall, apatite occurred much more frequently in the central deposits, where the crystals were also significantly larger in size than in the marginal localities.In the latter ones, apatite was mainly concentrated in small crystals within the matrix.Opaque minerals were mostly xenomorphic, rarely hypidiomorphic, and mainly confined to the matrix structure.However, they often also formed large individuals of up to 600 µm.The samples of the central areas did not contain microscopic amygdales or vesicles, while these were common in the marginal deposits.In most cases, they were circular (sometimes oval), of very variable size, and filled with either calcite, celadonite, zeolite, or chalcedony.In the southern deposit of OL, vesicles occurred mainly as irregular-flat or angular cavities of up to 7 mm in size.Most of them were filled with calcite, and some with chalcedony.
The matrix of aphyric rocks from Huevos del Diabolo (HdD) can be described as predominantly microcrystalline intergranular-to-vitreous intersertal.It consists, to a large extent, of elongated feldspars with an average size of 100-200 µm.To a large extent, these reveal a weak alignment (trachytic microstructure) (Figure 4a).Overall, the rock appeared predominantly greenish, both macroscopically and in thin sections.In the thin sections, amygdales of up to 4 mm were found filled with celadonite and also, sometimes, iron (oxy-hydr)oxides.

Agate 4.2.1. Macroscopic and Microscopic Appearance
The color spectrum of the individual bands in the examined agates ranged from colorless and white over yellow, brown, pink, or red in different nuances to a strong violet, whereby the color could change laterally within chalcedony bands.The different colors were not characteristic of single occurrences, but certain color preferences could be associated with the different localities [1,6].Thus, strong violet colors as well as dark red or yellow colors appeared predominantly in rocks from Rancho Coyamito, while Moctezuma, which is located to the north, was characterized primarily by rocks with pastel pink, yellow, or red tones (Figure 6) [1].Similar differences could be seen in the color distribution of individual bands.While agates from Ojo Laguna showed almost exclusively sharply delineated and finely banded colors, the Coyamito agates displayed a less sharp, or smooth, color delineation [1,6].These significant similarities and the fact that almost all colors and shapes occurred at all localities suggest that the deposits had similar formation processes.However, the varying predominance as well as subtle differences in the characteristics of color and banding indicate specific and slightly differing geochemical situations and conditions during agate formation among the individual deposits in Sierra del Gallego.
The agate samples appeared both macroscopically and in the polarization microscope as classical 'wall-lining' agates, in which chalcedony banding is parallel to the outer wall of the former rock cavity.Only this wall-parallel banding was found in the examined samples.The agates consisted exclusively of length-fast chalcedony.The quartzine (length-slow) variety was not observed in the investigated Chihuahua agates of the Rancho el Agate andesite.In addition, a typical crystallization series (from microcrystalline to spherulitic chalcedony to fibrous chalcedony and, finally, to macrocrystalline quartz, Figure 7) occurred several times in some cases in the same agate.The initial microcrystalline or spherulitic chalcedony layer was usually only very thin and did not occur in all agates.Only the agate samples that had a strongly irregular boundary with the host rock or contained host rock fragments and initial mineral precipitates (mainly ferruginous) had a very dominant generation of microcrystalline chalcedony.The majority of the chalcedony in the agates consisted of long, fibrous chalcedony.Some of the individual chalcedony layers showed a "Runzelbänderung," which is developed differently in the individual bands of the agates, while it is usually more distinct in the central areas of the agates.Some single chalcedony layers that were contrastive in color in the thin section showed a very strong, pronounced "Runzelbänderung."In most specimen, the chalcedony layers were followed by macrocrystalline quartz.The quartz varied from being thin crusts to dominant layers with idiomorphic crystals.In some cases, these crystals were overgrown by further generations of chalcedony.In one agate, a second generation of chalcedony could be detected that did not overgrow macrocrystalline quartz but directly followed the long, fibrous chalcedony.This was interrupted by microcrystalline chalcedony, which showed a completely new spherulitic growth at the beginning of the layer and ended in long, fibrous chalcedony.
The agates showed a strongly varying spectrum of macroscopic and microscopic inclusions.These differed in their type and color as well as in their size and distribution.
White mineral inclusions were predominantly found in the outer areas of the agates, their abundance decreasing significantly towards the center of the mineralization.
Overall, iron-rich, yellow-red mineral inclusions were dominant.The major macroscopically recognizable minerals were arranged wall-parallel along the crystallization fronts of the chalcedony bands.Smaller inclusions were distributed irregularly within the individual chalcedony bands in addition to having a wall-parallel arrangement.Additionally, a Zr-bearing phase as well as several spheroidal and semitransparent "bubbles" of up to 100 µm in size and of an orange-brown color occurred in the bands of some agates of Moctezuma and Huevos del Diabolo (Figure 8).The initial microcrystalline or spherulitic chalcedony layer was usually only very thin and did not occur in all agates.Only the agate samples that had a strongly irregular boundary with the host rock or contained host rock fragments and initial mineral precipitates (mainly ferruginous) had a very dominant generation of microcrystalline chalcedony.The majority of the chalcedony in the agates consisted of long, fibrous chalcedony.Some of the individual chalcedony layers showed a "Runzelbänderung," which is developed differently in the individual bands of the agates, while it is usually more distinct in the central areas of the agates.Some single chalcedony layers that were contrastive in color in the thin section showed a very strong, pronounced "Runzelbänderung."In most specimen, the chalcedony layers were followed by macrocrystalline quartz.The quartz varied from being thin crusts to dominant layers with idiomorphic crystals.In some cases, these crystals were overgrown by further generations of chalcedony.In one agate, a second generation of chalcedony could be detected that did not overgrow macrocrystalline quartz but directly followed the long, fibrous chalcedony.This was interrupted by microcrystalline chalcedony, which showed a completely new spherulitic growth at the beginning of the layer and ended in long, fibrous chalcedony.
The agates showed a strongly varying spectrum of macroscopic and microscopic inclusions.These differed in their type and color as well as in their size and distribution.
White mineral inclusions were predominantly found in the outer areas of the agates, their abundance decreasing significantly towards the center of the mineralization.
Overall, iron-rich, yellow-red mineral inclusions were dominant.The major macroscopically recognizable minerals were arranged wall-parallel along the crystallization fronts of the chalcedony bands.Smaller inclusions were distributed irregularly within the individual chalcedony bands in addition to having a wall-parallel arrangement.Additionally, a Zr-bearing phase as well as several spheroidal and semi-transparent "bubbles" of up to 100 µm in size and of an orange-brown color occurred in the bands of some agates of Moctezuma and Huevos del Diabolo (Figure 8).The "bubbles" were aligned with banding but occurred irregularly and seemed randomly distributed within these individual bands.They occurred only in individual chalcedony bands in the mentioned agates and were always anisotropic in their overall appearance under crossed polars.Some of these spheres additionally contained a darker colored core.A few of them were elliptically elongated along chalcedony fibers.
These inclusions occurred in one sample within a second chalcedony generation, which had grown on macrocrystalline quartz.There, they were associated with secondary fractures that also extended into the quartz.The fissures within the chalcedony were filled with identical reddish-brown inclusions.A similar optical effect and zoning of darker central areas and lighter transparent rims, as in the "bubbles," could be observed along the fissures.In the macrocrystalline quartz, the filling of these fissures with secondary minerals was not observed.

Trace Elements
The trace-element composition of the investigated agate samples is compiled in Table 5.Overall, the depletion of certain trace elements was detected in the agates compared to the related host rocks, in which the trace-element contents were mostly higher in chalcedony compared to the associated macrocrystalline quartz, which always had the lowest concentrations.In particular, the immobile trace elements Th and Ta could not be detected in any of the agate samples because the contents were below the detection limit.Furthermore, the elements Nb, Rb, Co, Sc, V, and Cs were detected in contents below 1 ppm in the agates.Other trace elements, such as Li, Cu, Sr, Y, and Ba, which were mostly present in concentrations below 5 ppm in the agates, have occurred in higher concentrations in some discrete bands of several agates.
Only a few trace elements were enriched in the silica matrix of the agates.Several hundreds of ppm were analyzed in the chalcedony matrix of the agates for Fe, Na, K and Ca (Table 5).Elevated concentrations were also detected for Al and Mg, which additionally showed a strong statistic scatter.A high aluminum content was especially found in white agate bands, while, in reddish agates, the iron content was usually more elevated as a result of their iron (oxy-hydr)oxide inclusions, which mainly occurred within these red areas of the agates.The "bubbles" were aligned with banding but occurred irregularly and seemed randomly distributed within these individual bands.They occurred only in individual chalcedony bands in the mentioned agates and were always anisotropic in their overall appearance under crossed polars.Some of these spheres additionally contained a darker colored core.A few of them were elliptically elongated along chalcedony fibers.
These inclusions occurred in one sample within a second chalcedony generation, which had grown on macrocrystalline quartz.There, they were associated with secondary fractures that also extended into the quartz.The fissures within the chalcedony were filled with identical reddish-brown inclusions.A similar optical effect and zoning of darker central areas and lighter transparent rims, as in the "bubbles," could be observed along the fissures.In the macrocrystalline quartz, the filling of these fissures with secondary minerals was not observed.

Trace Elements
The trace-element composition of the investigated agate samples is compiled in Table 5.Overall, the depletion of certain trace elements was detected in the agates compared to the related host rocks, in which the trace-element contents were mostly higher in chalcedony compared to the associated macrocrystalline quartz, which always had the lowest concentrations.In particular, the immobile trace elements Th and Ta could not be detected in any of the agate samples because the contents were below the detection limit.Furthermore, the elements Nb, Rb, Co, Sc, V, and Cs were detected in contents below 1 ppm in the agates.Other trace elements, such as Li, Cu, Sr, Y, and Ba, which were mostly present in concentrations below 5 ppm in the agates, have occurred in higher concentrations in some discrete bands of several agates.
Only a few trace elements were enriched in the silica matrix of the agates.Several hundreds of ppm were analyzed in the chalcedony matrix of the agates for Fe, Na, K and Ca (Table 5).Elevated concentrations were also detected for Al and Mg, which additionally showed a strong statistic scatter.A high aluminum content was especially found in white agate bands, while, in reddish agates, the iron content was usually more elevated as a result of their iron (oxy-hydr)oxide inclusions, which mainly occurred within these red areas of the agates.Elevated concentrations were also measured for the elements Sb and U. Whereas uranium was enriched in almost all agate samples (up to >30 ppm), Sb reached high concentrations only in the agate samples from Ojo Laguna (>300 ppm).Such anomalous concentrations of Sb and U (also Cs, Sr) were also detected in the volcanic host rocks from Ojo Laguna, where they were up to 200 times and 20 times higher, respectively, compared to the host rocks of the other localities.
Contents of manganese and zinc were generally rather low and homogeneously distributed in all samples.Remarkably, there were pronounced Zn, Ba, and Sr anomalies in individual agate bands.Chromium occurred in some of the agate samples in high concentrations (up to 50.3 ppm), sometimes even higher than in the associated host rocks.Moreover, Cr and Co contents exhibited a straight, statistical correlation (Figure 9a).Elevated concentrations were also measured for the elements Sb and U. Whereas uranium was enriched in almost all agate samples (up to >30 ppm), Sb reached high concentrations only in the agate samples from Ojo Laguna (>300 ppm).Such anomalous concentrations of Sb and U (also Cs, Sr) were also detected in the volcanic host rocks from Ojo Laguna, where they were up to 200 times and 20 times higher, respectively, compared to the host rocks of the other localities.
Contents of manganese and zinc were generally rather low and homogeneously distributed in all samples.Remarkably, there were pronounced Zn, Ba, and Sr anomalies in individual agate bands.Chromium occurred in some of the agate samples in high concentrations (up to 50.3 ppm), sometimes even higher than in the associated host rocks.Moreover, Cr and Co contents exhibited a straight, statistical correlation (Figure 9a).

Cathodoluminescence (CL)
The results of the CL investigations identified differing luminescence colors and associated chalcedony generations in the agates.The main CL emission bands were found at 450 nm (blue) and 650 nm (red) (Figure 10).The appearance of the ~450 nm (2.8 eV) emission band cou ld be related to two different defects in quartz.In Ti-poor quartz, the band is mainly associated with oxygen deficiency centers (ODC), but, in Ti-rich quartz, the intensity of the 450 nm band also shows a correlation with the concentration of trace Ti [20].The red emission band at ~650 nm (1.95 eV) is attributed to the so-called nonbridging oxygen hole center (NBOHC) with a number of different precursors [21].
At least one generation of blue luminescent chalcedony occurred in most of the agates.If present, this chalcedony generation always marked the beginning of amygdale filling and was, therefore, generally present in the outermost regions.The emission spectra of this chalcedony type exhibited the pronounced band, which remained stable under prolonged electron irradiation, at 450 nm.In addition, in all agates, the red emission band appeared at 650 nm which had almost no influence on the visual luminescence color because of its low intensity.In a few samples, an indistinct peak at about 690-700 nm could be found below this broad band.

Cathodoluminescence (CL)
The results of the CL investigations identified differing luminescence colors and associated chalcedony generations in the agates.The main CL emission bands were found at 450 nm (blue) and 650 nm (red) (Figure 10).The appearance of the ~450 nm (2.8 eV) emission band cou ld be related to two different defects in quartz.In Ti-poor quartz, the band is mainly associated with oxygen deficiency centers (ODC), but, in Ti-rich quartz, the intensity of the 450 nm band also shows a correlation with the concentration of trace Ti [20].The red emission band at ~650 nm (1.95 eV) is attributed to the so-called non-bridging oxygen hole center (NBOHC) with a number of different precursors [21].
At least one generation of blue luminescent chalcedony occurred in most of the agates.If present, this chalcedony generation always marked the beginning of amygdale filling and was, therefore, generally present in the outermost regions.The emission spectra of this chalcedony type exhibited the pronounced band, which remained stable under prolonged electron irradiation, at 450 nm.In addition, in all agates, the red emission band appeared at 650 nm which had almost no influence on the visual luminescence color because of its low intensity.In a few samples, an indistinct peak at about 690-700 nm could be found below this broad band.Furthermore, in some samples, green luminescence colors could be detected within some chalcedony bands.The spectra show several sharp lines between 500 nm and 600 nm (Figure 10), which can be associated with the luminescence of the uranyl ion (UO2 2+ ) [22].These lines mostly followed the blue chalcedony bands with dominant 450 nm luminescence emissions or formed alternating layers with them (Figure 10).Sometimes, the colors overlapped, in which cases both bands could be seen in different intensity ratios within the spectra.Green uranyl luminescence never appeared in the marginal areas of a mineralization.
Yellow luminescence with a pronounced broad band at about 570 nm was detected exclusively in the agate sample from Ojo Laguna (AOL), where it prevailed throughout chalcedony (Figure 10).The intensity of the band decreased continuously alongside prolonged electron bombardment.This CL emission was not detected in any of the other Furthermore, in some samples, green luminescence colors could be detected within some chalcedony bands.The spectra show several sharp lines between 500 nm and 600 nm (Figure 10), which can be associated with the luminescence of the uranyl ion (UO 2 2+ ) [22].These lines mostly followed the blue chalcedony bands with dominant 450 nm luminescence emissions or formed alternating layers with them (Figure 10).Sometimes, the colors overlapped, in which cases both bands could be seen in different intensity ratios within the spectra.Green uranyl luminescence never appeared in the marginal areas of a mineralization.
Yellow luminescence with a pronounced broad band at about 570 nm was detected exclusively in the agate sample from Ojo Laguna (AOL), where it prevailed throughout chalcedony (Figure 10).The intensity of the band decreased continuously alongside prolonged electron bombardment.This CL emission was not detected in any of the other agate samples.
In the macrocrystalline quartz parts of the agates, blue, yellowish-green, and red luminescence prevailed.A stable blue emission band at about 450 nm and broad bands in the region of 650 nm to around 720 nm were mainly dominant in the spectra.In the spectra of the greenish luminescent quartz, the typical sharp emission peaks of uranyl ion [22] occurred between 500 nm and 600 nm.Quartz crystals with yellow CL exhibited a transient broad band between 500 nm and 600 nm, which disappeared during electron bombardment, whereas the partially overlapping red and blue bands remained stable.The strongly pronounced sector zoning under CL within the quartz crystals of all samples, which otherwise appeared homogeneous in polarized light, as well as oscillatory growth zoning, is also remarkable.
A common feature in agates, especially those found near Rancho Coyamito, was the presence of pseudomorphs of chalcedony/agate after carbonates.The initial minerals were, in most cases, completely replaced by chalcedony or macrocrystalline quartz.Only small relics of the initial carbonates (mainly calcite) were detected during CL measurements (Figure 11).

R REVIEW 19 of 32
which otherwise appeared homogeneous in polarized light, as well as oscillatory growth zoning, is also remarkable.A common feature in agates, especially those found near Rancho Coyamito, was the presence of pseudomorphs of chalcedony/agate after carbonates.The initial minerals were, in most cases, completely replaced by chalcedony or macrocrystalline quartz.Only small relics of the initial carbonates (mainly calcite) were detected during CL measurements (Figure 11).

Electron Paramagnetic Resonance (EPR)
The EPR spectra of all examined chalcedony samples showed a weak signal of the rhombic Fe 3+ center at g = 4.28 as well as various radiation-induced lattice defects, including the pronounced Eʹ1 center at g = ca.2.00, in the wide scan measured at room temperature (Figure 12).From the narrow scans, it is obvious that the studied macrocrystalline quartz samples displayed lower intensities of these defects (Figure 12), showing that the defect density in chalcedony was much higher than in the associated macrocrystalline quartz.Furthermore, the "hole centers" in chalcedony also differed from those in quartz and include an unknown defect at g1 = 2.010, g2 = 2.006, and g3 = 2.003.Additional signals were present at the effective g values of ~2.001, 1.998, and 1.993, and could possibly be assigned to various Ge-related electron centers, such as [GeO4/M + ] 0 , where M = Li or Na [23,24] (Figure 12).

Electron Paramagnetic Resonance (EPR)
The EPR spectra of all examined chalcedony samples showed a weak signal of the rhombic Fe 3+ center at g = 4.28 as well as various radiation-induced lattice defects, including the pronounced E 1 center at g = ca.2.00, in the wide scan measured at room temperature (Figure 12).From the narrow scans, it is obvious that the studied macrocrystalline quartz samples displayed lower intensities of these defects (Figure 12), showing that the defect density in chalcedony was much higher than in the associated macrocrystalline quartz.Furthermore, the "hole centers" in chalcedony also differed from those in quartz and include an unknown defect at g 1 = 2.010, g 2 = 2.006, and g 3 = 2.003.Additional signals were present at the effective g values of ~2.001, 1.998, and 1.993, and could possibly be assigned to various Ge-related electron centers, such as [GeO 4 /M + ] 0 , where M = Li or Na [23,24] (Figure 12).(2) at RT. Powder EPR-spectra wide scans (c) showing the rhombic Fe 3+ center at g = 4.28 and radiation-induced defects, including the Eʹ1 center, at g = ~2.00;measuring condition: (1) at RT.

Characteristics and Formation of Volcanic Host-Rocks
The mineralogical and chemical compositions of the investigated volcanic rocks show that the studied agates are related to the same host rock within the unit of the Rancho el Agate andesite.In addition, similarities between the major-and trace-element data indicate that all deposits belong to the same lava flow.This assumption is strengthened by the common geochemical classifications.Based on the mineral compositions determined using XRD, all samples can be classified as quartz-poor to quartz-free latite according to the Streckeisen diagram.Small differences in their quantitative-phase compositions can be explained by the differing textures of the rocks and their varying amorphous contents.
The distint chemical composition of the Ojo Laguna deposit, which led to a different classification, can probably be explained by primary and secondary effects.Both an initial deviant magma composition and a secondary overprint of the volcanic host rocks may be responsible for its higher SiO2 content as well as for differences in its alkali metal ratios and elevated concentrations of some trace elements (e.g., Sb, U, Cs, Sr).In general, the Ojo

Characteristics and Formation of Volcanic Host-Rocks
The mineralogical and chemical compositions of the investigated volcanic rocks show that the studied agates are related to the same host rock within the unit of the Rancho el Agate andesite.In addition, similarities between the major-and trace-element data indicate that all deposits belong to the same lava flow.This assumption is strengthened by the common geochemical classifications.Based on the mineral compositions determined using XRD, all samples can be classified as quartz-poor to quartz-free latite according to the Streckeisen diagram.Small differences in their quantitative-phase compositions can be explained by the differing textures of the rocks and their varying amorphous contents.
The distint chemical composition of the Ojo Laguna deposit, which led to a different classification, can probably be explained by primary and secondary effects.Both an initial deviant magma composition and a secondary overprint of the volcanic host rocks may be responsible for its higher SiO 2 content as well as for differences in its alkali metal ratios and elevated concentrations of some trace elements (e.g., Sb, U, Cs, Sr).In general, the Ojo Laguna magma appears to have been slightly more acidic than the others, as indicated by the yellow CL with an emission band at 570 nm.Such characteristic yellow luminescence of chalcedony has previously been detected exclusively in agates from acidic volcanic host rocks and is therefore a strong indication concerning the geological and physico-chemical environment of the mineralization [25].Nevertheless, the strong equivalence between the major element compositions indicates that all of the studied volcanic host rocks belong to the intermediate series of lava flows from the Rancho el Agate andesite.
Bockoven [3] described a vesicular texture in the whole rock unit (at least twelve lava flows), which was most pronounced in the two youngest and thickest flows.Therefore, the occurrence of agate is theoretically possible in different lava units with slightly differing chemical compositions.This assumption was confirmed by Keller [5], who assumed several source channels, which are supposed to be mainly located on the eastern flank of Sierra del Gallego, for the Rancho el Agate andesite in their study area.The isolated location of the Ojo Laguna deposit in Sierra San Martin, which is both geographically and morphologically separated from Sierra del Gallego, reinforces the assumption of an independent source in this area.The chalcophile character of antimony additionally suggests the involvement of hydrothermal fluids in the agate formation [26], which may have also affected the composition of the host rock.
The various matrix textures in the different samples, from nearly completely intergranular to almost completely intersertal as well as single felsitic areas, indicate very differentiated cooling mechanisms.The whole unit can be classified as a near-surface formation, in which an extensive lava flow with high fluid lava cannot be excluded [3] but is rather implausible as a result of the lack of flow textures in the rocks of most localities.
The different contents and sizes of the phenocrysts additionally indicate a transport separation of phenocryst-rich and -poor magma fractions during lava flow in addition to differentiated cooling at the surface.The samples that were situated furthest from the main source area of the rock unit near Rancho el Agate, as suggested by Keller [5] and Bockoven [3], had the highest amorphous contents.Likewise, they had the highest tendency to contain skeletal crystals, swallowtail plagioclase, and a glassy matrix (Ojo Laguna, Huevos del Diabolo).In contrast, the central localities, especially on the eastern flank of Sierra del Gallego, were characterized by intergranular-to-felsitic textures with many phenocrysts and large apatite crystals.This is in agreement with the work of Bockoven [3], who described the occurrence of long prismatic apatite crystals within the Rancho el Agate andesite only in phenocryst-rich samples, which mainly occur in the vicinity of the possible source center of the magma near the Rancho el Agate.Therefore, the textures are primarily geographically specific properties of the rocks that were formed by differentiation and transport processes during the emplacement of the lava and depend significantly on their distance from the magma source.
As already assumed by Bockoven [3], the presence of resorbed feldspar phenocrysts indicates a magma mixing.Some feldspars were already crystallized during magma mixing and were partially resorbed by this process as a result of their original composition [15].These were predominantly simple or untwinned, idomorphic-to-hypidiomorphic individuals that exhibited strong resorption phenomena in the form of large embayment structures (feldspar generation 3).The former phenocrysts were no longer stable in the hybrid magma as a result of their chemical compositions and the temperature differences, which led to a strong resorption of the crystals [15].The subsequent epitaxial crystallization of intermediate feldspar formed the idiomorphic outer rims that are present today [15].Generation two feldspar, which contained unresorbed, calcium-and sodium-rich, polysynthetically twinned cores, exhibited a strong sieve texture in the external regions of the crystals.These phenocrysts either formed after magma mixing or were present before mixing and, as a result of their composition, were more stable to the hybrid magma than generation three.The sieve texture seems to be a decompression effect from the magma ascent [15,27,28].
Another indication of magma mixing is the occurrence of apatite accumulations along the dissolution structures of feldspar generation three (Figure 13), which were absent in the sieve-textured crystals of generation two.The mass crystallization of apatite was, in this case, due to the rapidly decreasing solubility of the mineral alongside its decreasing temperature and increasing SiO 2 content [29,30].Apatite mineralization accumulates mainly on already crystallized phenocrysts, which themselves are unable to incorporate phosphorus [31].In this particular case, this happened mainly along the edges of resorbed generation three feldspar crystals.This association also indicates that the present embayment structures are dissolution phenomena and not dendritic growth, as demonstrated in other deposits [32].
Minerals 2023, 13, x FOR PEER REVIEW 22 of 32 Another indication of magma mixing is the occurrence of apatite accumulations along the dissolution structures of feldspar generation three (Figure 13), which were absent in the sieve-textured crystals of generation two.The mass crystallization of apatite was, in this case, due to the rapidly decreasing solubility of the mineral alongside its decreasing temperature and increasing SiO2 content [29,30].Apatite mineralization accumulates mainly on already crystallized phenocrysts, which themselves are unable to incorporate phosphorus [31].In this particular case, this happened mainly along the edges of resorbed generation three feldspar crystals.This association also indicates that the present embayment structures are dissolution phenomena and not dendritic growth, as demonstrated in other deposits [32].The fresh, polysynthetically twinned, calcium-rich phenocrysts of feldspar generation one are relicts from one of the source magmas that, because of their chemical composition and the high melting point of anorthite, could not be dissolved by either magma mixing or rapid decompression [15].A later formation of these crystals during the outgoing crystallization of the magma can be ruled out, since no evidence of poikilitic texture was found in the individual crystals.Furthermore, the frequently observed synneusis of phenocrysts is exclusively related to the higher fluid stages of magma under turbulent conditions [31].Glomeroporphyritic aggregates, which consist mainly of fresh plagioclase crystals, pyroxene phenocrysts, and opaque minerals, occur exclusively in the porphyritic rocks with predominantly intergranular or felsitic matrix as well as large apatite crystals, as is the case in Rancho Coyamito and Agua Nueva.The lower mobility of these aggregates and their predominant origination in the deeper parts of the magma chamber indicate that these rocks solidified not far from the source vent [3].

Origin of Silica for Agate Formation
Agates form when silica fills formerly gas or liquid bubbles in lava.In this study, these vesicular cavities formed before the actual agate formation during the ascent and emplacement of the magma and were fixed in the rock during the cooling process.Most of these amygdales were irregularly distributed and concentrated in the upper part of the The fresh, polysynthetically twinned, calcium-rich phenocrysts of feldspar generation one are relicts from one of the source magmas that, because of their chemical composition and the high melting point of anorthite, could not be dissolved by either magma mixing or rapid decompression [15].A later formation of these crystals during the outgoing crystallization of the magma can be ruled out, since no evidence of poikilitic texture was found in the individual crystals.Furthermore, the frequently observed synneusis of phenocrysts is exclusively related to the higher fluid stages of magma under turbulent conditions [31].Glomeroporphyritic aggregates, which consist mainly of fresh plagioclase crystals, pyroxene phenocrysts, and opaque minerals, occur exclusively in the porphyritic rocks with predominantly intergranular or felsitic matrix as well as large apatite crystals, as is the case in Rancho Coyamito and Agua Nueva.The lower mobility of these aggregates and their predominant origination in the deeper parts of the magma chamber indicate that these rocks solidified not far from the source vent [3].

Origin of Silica for Agate Formation
Agates form when silica fills formerly gas or liquid bubbles in lava.In this study, these vesicular cavities formed before the actual agate formation during the ascent and emplacement of the magma and were fixed in the rock during the cooling process.Most of these amygdales were irregularly distributed and concentrated in the upper part of the lava flow (Figure 6a).Only a few are clearly flattened and follow a weak flow texture.This is consistent with the matrix samples examined, which also contain only minor or no evidence of a flow texture.
The yellow CL signal of the agate from Ojo Laguna provides a clear indication of the formation conditions of this agate deposit.Agates with the dominant 570 nm luminescence emission band form exclusively during the alteration of acidic volcanic host rocks [25].These formation conditions are associated with a rapid crystallization that is mostly below 250 • C and under an oxygen deficit [25].A relation between E 1 defect centers and the intensity of the 570 nm luminescence band, discussed in previous publications [33][34][35], could not be observed in the Mexican samples.On the other hand, the Laguna agates showed rather low concentrations of E 1 defect centers compared to the other studied samples.
The presence of extremely high antimony contents (>300 ppm) in the agates from Ojo Laguna is, among Mexican agates, a unique feature.High antimony contents were detected not only in the chalcedony itself, but also in the host rock in significant amounts (~90 ppm).The clearly lower contents in the host rock show that the element is enriched during the formation process of these agates.This is also confirmed by the other agate samples, in which Sb always shows elevated values compared to the host rock.However, only in the Ojo Laguna samples could antimony be detected locally (REM-EDS) within reddish iron oxide inclusions.Antimony (as a chalcophile element) is almost exclusively accumulated hydrothermally [26].Therefore, it can be assumed that a later hydrothermal influence occurred in the Ojo Laguna area.In general, high antimony contents in quartz have been detected in association with hydrothermal Sb-quartz mineralization [36,37].The Sb content of quartz, as a result of its lack of correlation with other trace elements, is apparently determined solely by the content of antimony in the fluid [36][37][38][39].No economically relevant antimony deposits are yet known within the Sierra del Gallego area [40].In the frequently occurring Pb-Zn-Ag-Cu deposits in the state of Chihuahua, however, high amounts of Sb, which occurs primarily in galena and replaces lead with coupled substitution together with silver [41], were detected.Thus, a connection between a hydrothermal formation of agates and possible polymetallic deposits seems plausible.
Also remarkable is the high chromium content of all analyzed agate samples.As an immobile element, chromium is only rarely accumulated in agate or in chalcedony [39].High contents (up to 9500 ppm [42]) have been thus far detected mainly in SiO 2 mineralization, which is spatially as well as genetically bound to chromium-bearing, ultramafic rocks or chromium deposits [43].No association between the high chromium content and chromiferous rocks or deposits was found in the Chihuahua area [5,40].However, this lack of association occurs in other localities as well.For example, Powolny [44] also described chromium-bearing agates (up to 140 ppm) without the typical green color of mtorolite from an alkaline and low-chromium series of the Borowno quarry in Poland.However, chromium can be mobilized via the hydrothermal alteration of volcanic rocks, amongst others, by the dissolution of chromite [45], resulting in a depletion of the element in these rocks [46].Chromium in the agates can therefore be interpreted as another indication of the involvement of hydrothermal solutions during their formation.It remains unclear from which rocks the chromium was mobilized, since the surrounding rocks are poor in chromium.While there was no evidence of chromium within the mineral inclusions of the agates, some of the CL spectra additionally showed an emission line at about 690 nm.At this position (678 nm-694 nm), the sharp R lines of Cr 3+ can be found [47,48].Therefore, the CL spectra indicate an incorporation of Cr 3+ within the chalcedony structure.

Microstructure and Color of Agates
Most of the analyzed trace elements are found within the numerous mineral inclusions occurring in the SiO 2 matrix of agate, while only a few of them can substitute small amounts of silicon within the structures of quartz or chalcedony (e.g., Ga, Ge, Ti, Al, Fe, B) [23,49].The substitution of Si ions with U in a quartz lattice is unlikely due to the different crystalchemical properties, instead, the data from Pan et al. [50] favor their incorporation as a uranyl-silicate complex in microcrystalline silicas.
White colors are common in some Agates from Chihuahua, while the frequency of white bands decreased clearly from the rim to the core.In the current samples, exclusively, a small part of the white bands was clearly colored, while the remaining part was almost completely transparent and did not contain any inclusions (Figure 14).White chalcedony bands had significantly lower iron contents compared to colored bands, as well as slightly elevated concentrations of Mg, Al, Sc, and Ca in comparison to colorless, transparent chalcedony.In general, it is assumed that the white color is related to chalcedony with large particles and low porosity [49].
Fe, B) [23,49].The substitution of Si ions with U in a quartz lattice is unlikely due to the different crystal-chemical properties, instead, the data from Pan et al. [50] favor their incorporation as a uranyl-silicate complex in microcrystalline silicas.
White colors are common in some Agates from Chihuahua, while the frequency of white bands decreased clearly from the rim to the core.In the current samples, exclusively, a small part of the white bands was clearly colored, while the remaining part was almost completely transparent and did not contain any inclusions (Figure 14).White chalcedony bands had significantly lower iron contents compared to colored bands, as well as slightly elevated concentrations of Mg, Al, Sc, and Ca in comparison to colorless, transparent chalcedony.In general, it is assumed that the white color is related to chalcedony with large particles and low porosity [49].However, an plate edge-like structure, which occurs in many white, especially, old agates worldwide [49,51,52], was not detected in any of the Mexican agates and can be excluded for the current samples based on their observed micro-structures and young ages.Since only minor differences in the trace-element contents, but no significant structural defects, could be detected in the CL or EPR spectra, it can be assumed that the coloration of white agate bands is mainly caused by different, not more closely specified, Al-Ca-Na-bearing mineral inclusions.
Within the red-colored agates, mineral inclusions were usually smaller and homogenously distributed inside chalcedony bands, which generally resulted in a more homogeneous color distribution within the bands.This distribution in the chalcedony bands, additionally occurred in the pink-, brown-, and yellow-colored agates, being most pronounced in the latter two.This is probably due to the significantly smaller particle size of the coloring components (Figure 15a,d,f) [53].In the trace element analyses, high iron contents, which correlate weakly with the intensity of the respective color shade, and somewhat lower Al contents were particularly noticeable.Agates in intense red or yellow shades had the highest iron contents.These high iron contents were also associated with high uranium contents, which, however, did not correlate exclusively with the iron contents.Uranium is mostly concentrated in the dark red-colored agates, while lower uranium contents were determined in the yellow and brown areas.The correlation between uranium and iron seems to be an effect of similar transportation mechanisms within the aqueous fluids either as a silicon complex [22,54] or as a result of the adsorption of uranium on colloidal iron particles [55][56][57].During the crystallization of agates, uranium is incorporated as a uranyl-silica complex within the structure of SiO2 [50], while iron forms different iron oxide/hydroxide inclusions within the SiO2 matrix.However, an plate edge-like structure, which occurs in many white, especially, old agates worldwide [49,51,52], was not detected in any of the Mexican agates and can be excluded for the current samples based on their observed micro-structures and young ages.Since only minor differences in the trace-element contents, but no significant structural defects, could be detected in the CL or EPR spectra, it can be assumed that the coloration of white agate bands is mainly caused by different, not more closely specified, Al-Ca-Nabearing mineral inclusions.
Within the red-colored agates, mineral inclusions were usually smaller and homogenously distributed inside chalcedony bands, which generally resulted in a more homogeneous color distribution within the bands.This distribution in the chalcedony bands, additionally occurred in the pink-, brown-, and yellow-colored agates, being most pronounced in the latter two.This is probably due to the significantly smaller particle size of the coloring components (Figure 15a,d,f) [53].In the trace element analyses, high iron contents, which correlate weakly with the intensity of the respective color shade, and somewhat lower Al contents were particularly noticeable.Agates in intense red or yellow shades had the highest iron contents.These high iron contents were also associated with high uranium contents, which, however, did not correlate exclusively with the iron contents.Uranium is mostly concentrated in the dark red-colored agates, while lower uranium contents were determined in the yellow and brown areas.The correlation between uranium and iron seems to be an effect of similar transportation mechanisms within the aqueous fluids either as a silicon complex [22,54] or as a result of the adsorption of uranium on colloidal iron particles [55][56][57].During the crystallization of agates, uranium is incorporated as a uranyl-silica complex within the structure of SiO 2 [50], while iron forms different iron oxide/hydroxide inclusions within the SiO 2 matrix.The mineral inclusions, themselves, were characterized using methods of REM-EDS and Raman spectroscopy.In different agate samples, mainly iron oxide/hydroxide, mineral inclusions were found.Most of them were purely iron bearing, while some samples from Ojo Laguna contained a significantly high amount of antimony.Iron mineral inclusions consisted of two different mineral phases, which existed in different transitional stages from each other.Raman spectroscopy of the more yellowlike inclusions The mineral inclusions, themselves, were characterized using methods of REM-EDS and Raman spectroscopy.In different agate samples, mainly iron oxide/hydroxide, mineral inclusions were found.Most of them were purely iron bearing, while some samples from Ojo Laguna contained a significantly high amount of antimony.Iron mineral inclusions consisted of two different mineral phases, which existed in different transitional stages from each other.Raman spectroscopy of the more yellowlike inclusions revealed mainly the modes of goethite, while the spectra of the transparent, mostly spherical and bubblelike inclusions (Figure 15c) additionally exhibit the chalcedony matrix.This was shown in the form of the typical Raman modes of the α-quartz as well as the additional mode at 502 cm −1 (Figure 16), which could be assigned to moganite [58].The more reddishappearing inclusions in the agates, as well as some central regions of the transparent "bubbles," contain only hematite.Especially in the latter one, but also in other inclusions, transitional states of goethite and hematite were observed.However, intense red colors were predominantly caused by the incorporation of hematite.Iron hydroxides, such as goethite, predominated in yellow and brown bands.Different compositions and mixtures of iron oxide, hydroxide, and oxyhydroxide phases cause the differing color shades.
Minerals 2023, 13, x FOR PEER REVIEW 26 of 32 revealed mainly the modes of goethite, while the spectra of the transparent, mostly spherical and bubblelike inclusions (Figure 15c) additionally exhibit the chalcedony matrix.This was shown in the form of the typical Raman modes of the α-quartz as well as the additional mode at 502 cm −1 (Figure 16), which could be assigned to moganite [58].
The more reddish-appearing inclusions in the agates, as well as some central regions of the transparent "bubbles," contain only hematite.Especially in the latter one, but also in other inclusions, transitional states of goethite and hematite were observed.However, intense red colors were predominantly caused by the incorporation of hematite.Iron hydroxides, such as goethite, predominated in yellow and brown bands.Different compositions and mixtures of iron oxide, hydroxide, and oxyhydroxide phases cause the differing color shades.Since manganese was present in relevantly high amounts only in brown agate areas, it can be assumed that manganese-rich compounds and minerals contributed to the color effect in these agate bands.However, neither manganese nor any manganese-bearing phase could be detected locally with REM-EDS or Raman spectroscopy.The particle size of the inclusions also had a decisive influence on the color impression, indicating that different color shades can also be caused by mineral inclusions of different sizes (especially iron minerals), which are chemically and structurally identical [49,53].This cannot be excluded as an explanation for the yellowish agates, which showed a very finely dispersed distribution of inclusions.Considering the increased Na content and only Since manganese was present in relevantly high amounts only in brown agate areas, it can be assumed that manganese-rich compounds and minerals contributed to the color effect in these agate bands.However, neither manganese nor any manganese-bearing phase could be detected locally with REM-EDS or Raman spectroscopy.The particle size of the inclusions also had a decisive influence on the color impression, indicating that different color shades can also be caused by mineral inclusions of different sizes (especially iron minerals), which are chemically and structurally identical [49,53].This cannot be excluded as an explanation for the yellowish agates, which showed a very finely dispersed distribution of inclusions.Considering the increased Na content and only slightly lower Fe content in the pink sample compared to the red samples, its color effect seems to be caused by a mixture of white and red mineral inclusions [1,59].

Agate Formation Process
The initial cavities of the andesite were subsequently filled with Si-rich fluids.Likely during a hiatus from volcanic activity or the fading rhyolitic volcanism, the exposure of emplaced units, primarily the Rancho el Agate andesite, led to an increased near-surface alteration of the rocks.During the alteration and extensive weathering of primary rockforming minerals, SiO 2 was mobilized in the voluminous rock units that were accumulated within aqueous solutions.Within trace-element data from the agate samples, there were indications (Al, Fe, K, Na, Ca, and U) of the formation of agate from such near-surface weathering solutions [49].Since thorium is strongly immobile under weathering conditions, unlike uranium, the measured U-Th ratio in the agates argues for mobilization under weathering conditions [46,60].The enrichments of chromium, antimony, zirconium, and zinc, among others, in the mineralization indicate the involvement of hydrothermal solutions during the agate formation process.In particular, the very high antimony contents in Ojo Laguna agate indicate intense hydrothermal activity resulting in an enrichment of the predominantly chalcophile element during agate formation.
The state of Chihuahua is one of the most deposit-and mineral-rich states in Mexico, and several (hydrothermal) ore deposits are known to exist in the area of Sierra del Gallego [61].In the vicinity of the studied deposits of agate, polymetallic Pb-Zn-Ag-Cu deposits are mainly found, but isolated uranium and manganese deposits also occur [40].Most of these deposits are related to Cretaceous sediments or directly to the volcanic rock units [40,61].The fluid origin of polymetallic deposits has not been conclusively determined.Megaw [62] assumed that hydrothermal polymetallic ore deposits were formed in association with felsitic intrusions.These intrusions can be related both geochemically and temporally to different felsitic/rhyolitic intrusions near various polymetallic ore deposits throughout all of northern Mexico, implying a similar genesis and magma source for these rocks [62].This also suggests the involvement of similar intrusions or extrusions in the formation of the agates and explains the high antimony contents and enrichments in the mineralization as a result of the participation of residual magmatic solutions during their formation.
The ages of the mentioned felsitic intrusions correlate approximately with the ages of rhyolites in the Sierra del Gallego area [61,62], suggesting that the rhyolites are a potential source of hydrothermal solutions.The involvement of Mesteño Rhyolite in the formation of Las Choyas Geodes, which occur regionally within the Libres Formation, has been assumed for a long time in this context [51].For the studied agates, various trace-element concentrations suggest that both near-surface weathering by meteoric waters and interaction with hydrothermal solutions resulted in the agate formation.
For the formation of the agates themselves, the migration of silica-enriched waters into vesicular cavities by diffusion is particularly of decisive importance [49], but suitable secondary pathways to the individual vesicles, such as fine cracks and veins, have also been found during field observations.The accumulation and condensation of monomeric silicic acid results first in the formation of dimers and oligomers, while changing physicochemical conditions (pH, Eh, temperature, changes in chemical composition, etc.) cause the precipitation of amorphous SiO 2 .The subsequent crystallization starts from the margins of cavities [49,63].As the SiO 2 concentration in the silica-bearing fluids decreases, macrocrystalline quartz finally crystallizes from the undersaturated residual solutions [49].Sector zoning, which was detected in the CL images of macrocrystalline quartz (Figure 17), suggests rapid crystallization under disequilibrium conditions [64].The chalcedony, itself, is normally strongly disordered, especially in young agates such as in the studied samples from Mexico [49,52].However, actual Raman spectroscopy investigations have shown a significant amount of moganite within agate samples, which additional to the young age of the mineralization suggests a lower crystallinity [65].Most of the micro-inclusions of coloring impurities were initially precipitated together with SiO2 and arranged during crystallization using a kind of "self-cleaning process" along the crystallization fronts in the later chalcedony [49].This process results in the typical wall-parallel arrangement of micro-inclusions in many agates.The deformation of this banding along so-called "escape tubes" shows that the mineral inclusions were already fixed before the final crystallization process (Figure 15a,b).Accordingly, a degassing through these escape tubes seems to have happened in the transition from amorphous SiO2-gel to crystalline chalcedony, since the chalcedony fibers along this structure show little to no deformation.However, in certain agate samples, there are also indications that some of the impurities entered the agate later by infiltration after the crystallization of chalcedony.This is shown by micro-inclusions of goethite and hematite in an agate sample from Huevos del Diablo, where an identical mineralization was detected both in the pore spaces (Figure 18) of selected chalcedony bands and in secondary cracks within the same agate (Figure 8).Because another generation of macrocrystalline quartz was precipitated after this infiltrative formation of inclusions, it can be assumed that not all agates were formed in a continuous process from a single silica supply.In the agates of Huevos del Diabolo, an additional Zr-bearing phase could be identified which seems to be zircon, according to the first analyses (Raman spectroscopy, EDS).The occurrence of this phase within secondary iron mineral accumulations indicates an earlier formation of the Zr-mineral (Figure 18).Most of the micro-inclusions of coloring impurities were initially precipitated together with SiO 2 and arranged during crystallization using a kind of "self-cleaning process" along the crystallization fronts in the later chalcedony [49].This process results in the typical wall-parallel arrangement of micro-inclusions in many agates.The deformation of this banding along so-called "escape tubes" shows that the mineral inclusions were already fixed before the final crystallization process (Figure 15a,b).Accordingly, a degassing through these escape tubes seems to have happened in the transition from amorphous SiO 2 -gel to crystalline chalcedony, since the chalcedony fibers along this structure show little to no deformation.However, in certain agate samples, there are also indications that some of the impurities entered the agate later by infiltration after the crystallization of chalcedony.This is shown by micro-inclusions of goethite and hematite in an agate sample from Huevos del Diablo, where an identical mineralization was detected both in the pore spaces (Figure 18) of selected chalcedony bands and in secondary cracks within the same agate (Figure 8).Because another generation of macrocrystalline quartz was precipitated after this infiltrative formation of inclusions, it can be assumed that not all agates were formed in a continuous process from a single silica supply.In the agates of Huevos del Diabolo, an additional Zr-bearing phase could be identified which seems to be zircon, according to the first analyses (Raman spectroscopy, EDS).The occurrence of this phase within secondary iron mineral accumulations indicates an earlier formation of the Zr-mineral (Figure 18).This multistage formation, which indicates multiple infillings of the cavities with SiO2-containing fluids or a crystallization interruption [66], is also emphasized by CL studies.Clearly differing chalcedony generations, each initiated by a separate spherulitic growth at its boundary with the next generation, could be detected (Figure 10).Another indication of multiple silica infill in some agates are secondary cracks, which cross-cut former chalcedony layers that have been subsequently filled with a new generation of chalcedony (Figure 15e).Different generations may have also be formed as a result of a rapid change in physicochemical conditions, e.g., a secondary supply of hydrothermal fluids.

Conclusions
The present study aimed to carry out a comprehensive mineralogical and geochemical investigation into some of the world-famous agate deposits in the Chihuahua region of Mexico.The volcanic host rocks of the agates are intermediate trachyandesites according to the classical TAS classifications and can be classified geochemically as quartzpoor latite, which belong to the intermediate unit of Rancho el Agate andesite.The agates from Ojo Laguna also belongs to the unit of the Rancho el Agate andesite but have slightly higher SiO2 content.In addition, this unit shows a much stronger rock alteration that was mainly caused by deep hydrothermal waters in addition to the general near-surface weathering that caused slight alterations in all host rocks.Geochemical data indicate that these hydrothermal fluids are probably closely related to the fluids that were involved in the formation of widespread polymetallic ore deposits in the Chihuahua region.
The micro-textures of the investigated rocks range from intersertal and intergranular over trachytic to felsitic and have differing phenocryst contents and sizes.They represent local-specific textures of an intermediate unit formed close to the surface.The different textures have been formed by various differentiation and transport processes during lava extrusion and the final emplacement of this unit.On the other hand, they are an expression of the different cooling mechanisms within sections of lava flow.Different characteristics of the rock-forming minerals indicate that the intermediate trachyandesitic magma was formed by the mixing of acidic magma with basic (basaltic) magma.
Overall, the volcanic rocks are strongly vesicular, especially in the upper parts of the unit, with partly large amygdaloid cavities.The later infill of silica-rich solutions resulted in the formation of agates.Mineralogical and geochemical data indicate that, in the investigated agate occurrences within the area of Sierra de Gallego, both secondary rock This multistage formation, which indicates multiple infillings of the cavities with SiO 2 -containing fluids or a crystallization interruption [66], is also emphasized by CL studies.Clearly differing chalcedony generations, each initiated by a separate spherulitic growth at its boundary with the next generation, could be detected (Figure 10).Another indication of multiple silica infill in some agates are secondary cracks, which cross-cut former chalcedony layers that have been subsequently filled with a new generation of chalcedony (Figure 15e).Different generations may have also be formed as a result of a rapid change in physicochemical conditions, e.g., a secondary supply of hydrothermal fluids.

Conclusions
The present study aimed to carry out a comprehensive mineralogical and geochemical investigation into some of the world-famous agate deposits in the Chihuahua region of Mexico.The volcanic host rocks of the agates are intermediate trachyandesites according to the classical TAS classifications and can be classified geochemically as quartz-poor latite, which belong to the intermediate unit of Rancho el Agate andesite.The agates from Ojo Laguna also belongs to the unit of the Rancho el Agate andesite but have slightly higher SiO 2 content.In addition, this unit shows a much stronger rock alteration that was mainly caused by deep hydrothermal waters in addition to the general near-surface weathering that caused slight alterations in all host rocks.Geochemical data indicate that these hydrothermal fluids are probably closely related to the fluids that were involved in the formation of widespread polymetallic ore deposits in the Chihuahua region.
The micro-textures of the investigated rocks range from intersertal and intergranular over trachytic to felsitic and have differing phenocryst contents and sizes.They represent local-specific textures of an intermediate unit formed close to the surface.The different textures have been formed by various differentiation and transport processes during lava extrusion and the final emplacement of this unit.On the other hand, they are an expression of the different cooling mechanisms within sections of lava flow.Different characteristics of the rock-forming minerals indicate that the intermediate trachyandesitic magma was formed by the mixing of acidic magma with basic (basaltic) magma.
Overall, the volcanic rocks are strongly vesicular, especially in the upper parts of the unit, with partly large amygdaloid cavities.The later infill of silica-rich solutions resulted in the formation of agates.Mineralogical and geochemical data indicate that, in the investigated agate occurrences within the area of Sierra de Gallego, both secondary rock alteration and hydrothermal solutions have led to agate formation.These conditions resulted in the extensive mobilization of Si and other trace elements (Fe, U, Al, Ca, K, etc.).The hydrothermal activity in the area, which was mainly caused by outgoing rhyolitic volcanism, provided additional element-enriched fluids (Si, Sb, Zn, etc.) and the necessary heat for the improved transportability of SiO 2 in aqueous solutions.Likewise, the migration of hydrothermal fluids may have led to mobilization of additional elements, such as Cr and Zr, from host rocks.Volatile fluids acted as transport media, enabling the transport of immobile trace elements and silicon.The migration of fluids occurred both on fractures and by diffusion on intergranular pore spaces, which caused an enrichment inside the vesicular cavities and later crystallization.Thus, the different chalcedony generations, and also the appearance of micro-inclusions, show that several cycles of crystallization and fluid replenishment took place in the cavity.
The predominant similarities between agates from spatially separated deposits prove a fundamental connection in the formation processes of all agates within the area.Depending on the local influences of host rocks and hydrothermal systems and the influence of different fluids, agates with slightly different characteristics were formed.For example, Ojo Laguna agates seem to have formed with significantly higher influences from hydrothermal fluids.This can be seen in the texture of the secondary rock as well as in the trace elements of the agates.Significant amounts of antimony can be associated with a metal-bearing hydrothermal fluid, seemingly related to the fluids that caused the polymetallic ore deposit formations in the area, while the yellow CL of quartz in this deposit normally indicates rapid crystallization in an oxygen-deficient environment under low temperatures.
The colors of the agates are mainly due to mineral micro-inclusions.These were previously formed during crystallization in the chalcedony via self-cleaning processes.In addition, there are also agates in which inclusions were formed by subsequent infiltration.These inclusions are different iron (oxy-hydr)oxides which cause differing yellow, red, and brown colors as a result of their different crystallinity and crystallite sizes.White colors are also exclusively caused by foreign mineral inclusions and are not related to the formation and texture of the chalcedony.The common violet color of agates from Rancho Coyamito is not caused as a result of mineral inclusions but seems to be a special defect in the agate's structure.This unique color, which can predominantly be found in Mexico, is the subject of further research.

Figure 1 .
Figure 1.Stratigraphic profile of the volcanic complex in the Sierra del Gallego area.It shows the relative position and average thickness as well as the K-Ar ages of the rock units [in Ma]; (modified after [5]).

Figure 2 .
Figure2.Map of agate deposits in the northern part of the state of Chihuahua.Marked are the deposits studied in this work and the Rancho el Agate Ranch, which, according to Bockoven[3], is to be seen as the main source area of Rancho el Agate andesite; (modified after[6]).

Figure 3 .
Figure 3. TAS-classification diagrams following Cox [10] (a) and Middlemost [12] (b) show that nearly all host rocks can be classified as trachyandesite.Only the sample from Ojo Laguna shows a different, more acidic composition.

Figure 3 .
Figure 3. TAS-classification diagrams following Cox [10] (a) and Middlemost [12] (b) show that nearly all host rocks can be classified as trachyandesite.Only the sample from Ojo Laguna shows a different, more acidic composition.

Figure 4 .
Figure 4. Micrographs in transmitted light (crossed polars) showing typical matrix textures in the studied samples.(a) Poorly trachytic matrix due to slight alignment of plagioclase in sample HdD; (b) intergranular, slightly glassy matrix of sample ANZ; (c) strong interlocking of plagioclase crystals in felsitic matrix of sample CLA; (d) strongly glassy, intersertal matrix with swallowtail plagioclase and skeletal crystals in sample OL.

Figure 4 .
Figure 4. Micrographs in transmitted light (crossed polars) showing typical matrix textures in the studied samples.(a) Poorly trachytic matrix due to slight alignment of plagioclase in sample HdD; (b) intergranular, slightly glassy matrix of sample ANZ; (c) strong interlocking of plagioclase crystals in felsitic matrix of sample CLA; (d) strongly glassy, intersertal matrix with swallowtail plagioclase and skeletal crystals in sample OL.

Figure 5 .
Figure 5. Micrographs in transmitted light (crossed polars) showing the different generations of feldspar in the host rock CLJ.(a) Fresh polysynthetically twinned plagioclase crystal of generation 1; (b) sieve textured plagioclase crystal with fresh core of generation 2; (c) idiomorphic feldspar crystal with large embayment-like resorption structures of generation 3; all images are from sample CLJ.

Figure 5 .
Figure 5. Micrographs in transmitted light (crossed polars) showing the different generations of feldspar in the host rock CLJ.(a) Fresh polysynthetically twinned plagioclase crystal of generation 1; (b) sieve textured plagioclase crystal with fresh core of generation 2; (c) idiomorphic feldspar crystal with large embayment-like resorption structures of generation 3; all images are from sample CLJ.

Figure 6 .
Figure 6.Investigated agate deposits and representative samples (scale bar for agates is 2 cm); (a) Upper part of the lava flows of the Rancho el Agate andesite at the Rancho Coyamito with empty vesicles and those filled with agate; (b) characteristic Coyamito agate; (c) strongly altered volcanic rocks and tuff layers at the Arcoiris claim, Ojo Laguna; (d) Laguna agate with typical fine banding; (e) vesicular andesite at the Moctezuma deposit; (f) Moctezuma agate; (g) weathering debris with loose agate geodes at Huevos del Diablo; (h) agate from Huevos del Diablo with characteristic brownish-yellow color banding; (i) reddish andesite from Agua Nueva containing agate geodes; (k) agate from Agua Nueva.

Figure 6 .
Figure 6.Investigated agate deposits and representative samples (scale bar for agates is 2 cm); (a) Upper part of the lava flows of the Rancho el Agate andesite at the Rancho Coyamito with empty vesicles and those filled with agate; (b) characteristic Coyamito agate; (c) strongly altered volcanic rocks and tuff layers at the Arcoiris claim, Ojo Laguna; (d) Laguna agate with typical fine banding; (e) vesicular andesite at the Moctezuma deposit; (f) Moctezuma agate; (g) weathering debris with loose agate geodes at Huevos del Diablo; (h) agate from Huevos del Diablo with characteristic brownish-yellow color banding; (i) reddish andesite from Agua Nueva containing agate geodes; (k) agate from Agua Nueva.

Figure 7 .
Figure 7.A complete agate thin section from Moctezuma depicted in transmitted light (crossed polars (a) with additional λ-compensator (b)).The typical sequence for wall-lining agates is clearly visible; from left to right: microcrystalline/spherulitic chalcedony to fibrous chalcedony to macrocrystalline quartz; in the lower image with λ-compensator, the continuous length-fast chalcedony is visible.

Figure 7 .
Figure 7.A complete agate thin section from Moctezuma depicted in transmitted light (crossed polars (a) with additional λ-compensator (b)).The typical sequence for wall-lining agates is clearly visible; from left to right: microcrystalline/spherulitic chalcedony to fibrous chalcedony to macrocrystalline quartz; in the lower image with λ-compensator, the continuous length-fast chalcedony is visible.

Figure 8 .
Figure 8. Micrographs in transmitted light showing orange-brown inclusions in an agate sample of Huevos del Diablo (HdD).In the overview image (a), the irregular distribution of the bubblelike inclusions is clearly visible, as well as the secondary fractures, which extend into quartz generation 1 (Qz 1) (red arrow) and end at generation 2 (Qz 2); In the detail image (b), a similarity between the bubblelike inclusions and those along the secondary fractures can be seen.

Figure 8 .
Figure 8. Micrographs in transmitted light showing orange-brown inclusions in an agate sample of Huevos del Diablo (HdD).In the overview image (a), the irregular distribution of the bubblelike inclusions is clearly visible, as well as the secondary fractures, which extend into quartz generation 1 (Qz 1) (red arrow) and end at generation 2 (Qz 2); In the detail image (b), a similarity between the bubblelike inclusions and those along the secondary fractures can be seen.

Figure 9 .
Figure 9. Correlation pattern of Cr, Co, Fe, and Ti of the agate samples (a).Note the perfect correlation of Cr and Co and Fe and U content in ppm (b).Marked are the differently colored agate samples: red agate (A), colorless/white agate and quartz (B), and brown/yellow agate (C).

Figure 9 .
Figure 9. Correlation pattern of Cr, Co, Fe, and Ti of the agate samples (a).Note the perfect correlation of Cr and Co and Fe and U content in ppm (b).Marked are the differently colored agate samples: red agate (A), colorless/white agate and quartz (B), and brown/yellow agate (C).

Figure 10 .
Figure 10.CL images of the blue (a), green (b), and yellow (c) luminescent chalcedony with associated spectra.The letters a-c represent measuring points and their associated spectra.Sample ACJ (left); sample AOL (right).Note the spherulitic growth in chalcedony at the transition from blue luminescent chalcedony to green luminescent chalcedony.

Figure 10 .
Figure 10.CL images of the blue (a), green (b), and yellow (c) luminescent chalcedony with associated spectra.The letters a-c represent measuring points and their associated spectra.Sample ACJ (left); sample AOL (right).Note the spherulitic growth in chalcedony at the transition from blue luminescent chalcedony to green luminescent chalcedony.

Figure 11 .
Figure 11.(a) Typical pseudomorphs in an agate from Rancho Coyamito; (b,c) micrographs of the pseudomorphic micro-structure in transmitted light (crossed polars (b)) and CL (c); the pre-existing minerals were almost completely replaced by chalcedony and macrocrystalline quartz; only small relics of former calcite were detectable by CL (orange dots, see arrows).

Figure 11 .
Figure 11.(a) Typical pseudomorphs in an agate from Rancho Coyamito; (b,c) micrographs of the pseudomorphic micro-structure in transmitted light (crossed polars (b)) and CL (c); the pre-existing minerals were almost completely replaced by chalcedony and macrocrystalline quartz; only small relics of former calcite were detectable by CL (orange dots, see arrows).

Figure 12 .
Figure 12.EPR spectra of some chalcedony samples (a) as well as a comparison of chalcedony and macrocrystalline quartz from Huevos del Diabolo (b).The Eʹ1 defect center, another signal at g = 1.993 (arrow), presumably related to various Ge-related electron centers, as well as the clearly lower intensity of these signals in the macrocrystalline quartz samples are notable; measuring condition: (2) at RT. Powder EPR-spectra wide scans (c) showing the rhombic Fe 3+ center at g = 4.28 and radiation-induced defects, including the Eʹ1 center, at g = ~2.00;measuring condition: (1) at RT.

Figure 12 .
Figure 12.EPR spectra of some chalcedony samples (a) as well as a comparison of chalcedony and macrocrystalline quartz from Huevos del Diabolo (b).The E 1 defect center, another signal at g = 1.993 (arrow), presumably related to various Ge-related electron centers, as well as the clearly lower intensity of these signals in the macrocrystalline quartz samples are notable; measuring condition: (2) at RT. Powder EPR-spectra wide scans (c) showing the rhombic Fe 3+ center at g = 4.28 and radiation-induced defects, including the E 1 center, at g = ~2.00;measuring condition: (1) at RT.

Figure 13 .
Figure 13.Micrographs in CL (a) and transmitted light (b) of generation three feldspar phenocryst with prominent precipitates of microcrystalline apatite along the resorption rims (yellow spots in CL) formed by magma mixing inside the idiomorphic crystal.

Figure 13 .
Figure 13.Micrographs in CL (a) and transmitted light (b) of generation three feldspar phenocryst with prominent precipitates of microcrystalline apatite along the resorption rims (yellow spots in CL) formed by magma mixing inside the idiomorphic crystal.

Figure 14 .
Figure 14.Micrograph in transmitted light shows different inclusions that are accumulated only within a small part of each chalcedony band in an agate sample of the La Fortuna Claim, Rancho Coyamito (a).The same distribution is also macroscopically visible in the polished section (b).

Figure 14 .
Figure 14.Micrograph in transmitted light shows different inclusions that are accumulated only within a small part of each chalcedony band in an agate sample of the La Fortuna Claim, Rancho Coyamito (a).The same distribution is also macroscopically visible in the polished section (b).

Figure 15 .
Figure 15.Typical micro-structures of the investigated agate samples ((a,c,d,f); transmitted light; (b,e) crossed polars).(a,b) Escape tube in a sample from Coyamito with obvious material displacement caused by the rapid escape of gas/liquid; in polarized light it is visible that the surrounding chalcedony fibers were only partially deformed by the process, and material of the inner bands was crystallized in the channel; (c) micrograph of a sample from Huevos del Diablo showing the wall-parallel arrangement of micro-inclusions, while the large roundish and semitransparent inclusions occurred mainly irregularly arranged within the individual bands; (d) distinctly colored wall-parallel banding in an agate from Moctezuma; in addition, similar semitransparent inclusions, as in agates from HdD, were present within the bands; (e) secondary chalcedony mineralization, crosscutting fibers arranged perpendicular to the crack surfaces in an agate from Ojo Laguna; (f) very tiny micro-inclusions concentrically arranged within thin bands of an agate from the Japanese deposit of the Rancho Coyamito; the inclusions were seemingly cut off by the banding and formed shapes reminiscent of "micro-escape tubes" in some places (uppermost band).

Figure 15 .
Figure 15.Typical micro-structures of the investigated agate samples ((a,c,d,f); transmitted light; (b,e) crossed polars).(a,b) Escape tube in a sample from Coyamito with obvious material displacement caused by the rapid escape of gas/liquid; in polarized light it is visible that the surrounding chalcedony fibers were only partially deformed by the process, and material of the inner bands was crystallized in the channel; (c) micrograph of a sample from Huevos del Diablo showing the wall-parallel arrangement of micro-inclusions, while the large roundish and semi-transparent inclusions occurred mainly irregularly arranged within the individual bands; (d) distinctly colored wall-parallel banding in an agate from Moctezuma; in addition, similar semi-transparent inclusions, as in agates from HdD, were present within the bands; (e) secondary chalcedony mineralization, crosscutting fibers arranged perpendicular to the crack surfaces in an agate from Ojo Laguna; (f) very tiny micro-inclusions concentrically arranged within thin bands of an agate from the Japanese deposit of the Rancho Coyamito; the inclusions were seemingly cut off by the banding and formed shapes reminiscent of "micro-escape tubes" in some places (uppermost band).

Figure 16 .
Figure 16.Raman spectra of analyzed micro-inclusions in sample AHdD.The spectra of light brown inclusions contain mainly goethite while the matrix of the chalcedony with the modes of quartz and moganite is also visible (a).The spectra of dark red inclusions contain modes of hematite and quartz (b).

Figure 16 .
Figure 16.Raman spectra of analyzed micro-inclusions in sample AHdD.The spectra of light brown inclusions contain mainly goethite while the matrix of the chalcedony with the modes of quartz and moganite is also visible (a).The spectra of dark red inclusions contain modes of hematite and quartz (b).

Figure 17 .
Figure 17.Micrographs of chalcedony and macrocrystalline quartz in an agate from Huevos del Diablo in CL (a) and transmitted light (crossed polars, (b)) Strong sector zoning between yellowgreenish and blue luminescent areas within the macrocrystalline quartz grains, which appear homogeneous in polarized light, is visible; a decrease in the greenish-yellow luminescence color as a result of longer electron bombardment can be seen in the lower right part of the CL image.

Figure 17 .
Figure 17.Micrographs of chalcedony and macrocrystalline quartz in an agate from Huevos del Diablo in CL (a) and transmitted light (crossed polars, (b)) Strong sector zoning between yellow-greenish and blue luminescent areas within the macrocrystalline quartz grains, which appear homogeneous in polarized light, is visible; a decrease in the greenish-yellow luminescence color as a result of longer electron bombardment can be seen in the lower right part of the CL image.

Figure 18 .
Figure 18.REM-Micrographs of the "bubble" mineral inclusions in an agate sample from Huevos del Diabolo.Goethite and hematite form big accumulations within the matrix of the chalcedony while small Zr-bearing inclusions are randomly arranged within the matrix (a).The detailed image shows the distribution within the pore spaces of the chalcedony for the Fe-minerals as well as the bright Zr-bearing inclusions (b).

Figure 18 .
Figure 18.REM-Micrographs of the "bubble" mineral inclusions in an agate sample from Huevos del Diabolo.Goethite and hematite form big accumulations within the matrix of the chalcedony while small Zr-bearing inclusions are randomly arranged within the matrix (a).The detailed image shows the distribution within the pore spaces of the chalcedony for the Fe-minerals as well as the bright Zr-bearing inclusions (b).

Table 1 .
Description of the host rock samples.

Table 1 .
Description of the host rock samples.

Table 2 .
Description of the separated agate, chalcedony, and quartz samples.

Table 4 .
Quantitative composition of the samples.Determined with XRD and the program Profex using 20 wt.% corundum as internal standard.[data in wt.%].

Table 5 .
Trace-element contents (in ppm) of the separated chalcedony and macrocrystalline quartz bands analyzed using ICP-MS.