Journal of Petrology Advance Access originally published online on January 28, 2005
Journal of Petrology 2005 46(5):945-972; doi:10.1093/petrology/egi006
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Regional Variations in the Mineralogy of Metasomatic Assemblages in Mantle Xenoliths from the West Eifel Volcanic Field, Germany
1 DEPARTMENT OF GEOLOGY, UNIVERSITY OF NEW BRUNSWICK, 2 BAILEY DRIVE, FREDERICTON, NEW BRUNSWICK, CANADA, E3B 5A3
2 DEPARTMENT OF EARTH SCIENCES, UNIVERSITY OF WESTERN ONTARIO, LONDON, ONTARIO, CANADA, N6A 5B7
3 DEPARTMENT OF EARTH SCIENCES AND THE GREAT LAKES INSTITUTE FOR ENVIRONMENTAL RESEARCH, UNIVERSITY OF WINDSOR, WINDSOR, ONTARIO, CANADA, N9B 3P4
4 DEPARTMENT OF EARTH SCIENCES AND THE GREAT LAKES INSTITUTE FOR ENVIRONMENTAL RESEARCH, UNIVERSITY OF WINDSOR, WINDSOR, ONTARIO, CANADA, N9B 3P4
RECEIVED JULY 7, 2003; ACCEPTED DECEMBER 3, 2004
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONGEOLOGY OF THE WEST...GOALSSAMPLE LOCATION AND PETROGRAPHYANALYTICAL TECHNIQUESMINERAL CHEMISTRYDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Xenoliths record two distinct events in the mantle below the Quarternary West Eifel Volcanic Field, Germany. The first, during the Hercynian Orogeny, led to widespread formation of secondary, Ti-poor amphibole, clinopyroxene and phlogopite. The signature of the second event, related to Quaternary volcanism, varies across the field. At Dreiser Weiher and Meerfelder Maar, this event is characterized by amphibole–phlogopite–clinopyroxene veins, hosted in lherzolite and harzburgite xenoliths brought to the surface by sodic olivine nephelinite–basanite suite lavas. These veins formed from crystallization of sodic magma that flowed along fractures in the mantle. At Rockeskyller Kopf, Gees and Baarley, the Quaternary event is characterized by wehrlite xenoliths, many of which have phlogopite–clinopyroxene veins, that were transported by potassic foid suite lavas. Wehrlite formed by reaction of lherzolite–harzburgite, with a large volume of potassic magma that flowed along grain boundaries rather than in fractures. During reaction, orthopyroxene was consumed and secondary clinopyroxene, olivine and phlogopite precipitated. Veins formed in wehrlites only during periodic over-pressure events. The composition of the magmas parental to the veins is similar to the lavas that carried the xenoliths to surface, indicating that the source of foid and olivine nephelinite–basanite suite magma is domainal, as was the flow regime and magma flux.
KEY WORDS: Eifel; mantle xenoliths; metasomatism; trace elements
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGY OF THE WEST...GOALSSAMPLE LOCATION AND PETROGRAPHYANALYTICAL TECHNIQUESMINERAL CHEMISTRYDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Metasomatic addition of new minerals to depleted mantle via interaction with melts or fluids is a well documented phenomenon in mantle xenoliths and massif peridotites (e.g. Roden & Murthy, 1985
; Bailey, 1987; Kelemen, 1990; Bedini et al., 1997; Edgar, 1997; Bodinier et al., 2004). Metasomatic events can be pervasive, if the metasomatizing agent infiltrates along an interconnected pathway of grain boundaries (Dawson, 1984; Harte et al., 1993), or localized in veins if the agent flows through fractures (Wilshire et al., 1980; Wilshire & Kirby, 1989; Fabries et al., 2000). Depending on the permeability of the rocks adjacent to the veins, there can be considerable infiltration metasomatism of the wall rocks (Gamble & Kyle, 1987; Kelemen, 1990; Xu et al., 1996), perhaps coupled with chromatographic separation of trace elements (e.g. Hofmann, 1972; Navon & Stolper, 1987). The nature of a single metasomatic event in the lithospheric mantle may, therefore, vary from place to place, depending on: the flow regime of the magma, the reactivity of the host rocks with respect to the magma, the total flux of magma and the duration of magma flow.
Witt-Eickschen et al. (1993, 1994, 1998, 2003a), Witt-Eickschen & Kramm (1998) and Kempton et al. (1988) presented clear evidence for multiple metasomatic events in the mantle below the Quaternary West Eifel Volcanic Field (Fig. 1; Table 1). The available data indicate that there was a pervasive metasomatic event, most likely during the Hercynian Orogeny, that is characterized by the presence of disseminated Ti-poor, LREE-enriched amphibole and clinopyroxene in previously depleted lherzolite and harzburgite. This was overprinted by a younger, shorter-lived event associated with Quaternary magmatism. At the two localities studied by Witt-Eickschen and co-workers (Meerfelder Maar and Dreiser Weiher; Fig. 1), this younger metasomatic event is characterized by the presence of anhydrous clinopyroxenite veins and clinopyroxene- and phlogopite-bearing hornblendite veins in lherzolite and harzburgite hosts. Our goal in this study is to determine the petrogenesis of veined and unveined wehrlite xenoliths at three localities in the West Eifel to determine whether they record spatial variations in vein mineralogy and petrology across the West Eifel Volcanic Field that can be related to variables such as magma flux and composition.
|
|
|
GEOLOGY OF THE WEST EIFEL VOLCANIC FIELD |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGY OF THE WEST...GOALSSAMPLE LOCATION AND PETROGRAPHYANALYTICAL TECHNIQUESMINERAL CHEMISTRYDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Volcanic rocks
The volcanic rocks in the West Eifel field (Fig. 1) can be divided into two groups, based on age and composition (Mertes & Schmincke, 1983
; Fuhrmann & Lippolt, 1987; Schnepp & Hradetzky, 1994). In general, feldspathoid-rich, potassic lavas, termed ‘F-suite’ by Mertes & Schmincke (1983, 1985), range in age from 
0·66 to 0·1 Ma and are found in the northwestern and central part of the field. The youngest eruptive centres (younger than 0·1 Ma) are concentrated in, though not limited to, the southeast, between Daun and Bad Bertrich. These younger lavas form a subordinate group of sodi-potassic olivine nephelinites and basanites called the ONB-suite (Mertes & Schmincke, 1983, 1985).
Mertes (1983) pointed out that the West Eifel field is cross-cut by the eastern edge of the Eifel North–South Zone depression (ENSZD). This is a large-scale graben structure that extends from southern France through Luxembourg into the Eifel region and has been active since the Devonian (Echle & Gussone, 1985). The eastern edge of the ENSZD is defined here as the NNE–SSW trending ‘zone’, marking the eastern extension of Middle–Upper Devonian and Triassic sediments (Fig. 1). To the east of the graben edge, these younger sedimentary rocks are absent and only the lower Devonian is exposed. The edge of the depression is marked by a zone of NNE–SSW trending faults (Büchel, 1994) that are particularly well developed just north of Meerfelder Maar (Fig. 1).
Mertes & Schmincke (1985) suggested that the F- and ONB suites were derived from two distinct sources: F-suite lavas originated in a source enriched in CO2 and K, whereas the ONB source was less enriched in alkalis and CO2. Both magma types formed within the garnet stability field, though Mertes & Schmincke (1985) suggested that the F-suite magmas had a deeper source than the more sodic ONB suite. This model is generally consistent with any of the three proposed models for the origin of volcanism in the Central European Volcanic Belt (see Haase et al., 2004, and references therein). Wilson & Downes (1991) suggested that the dominant magma tyre in the West and Central European Volcanic Province is sodic and that potassic lavas are derived from interaction of sodic melts with potassium-enriched lithospheric mantle; this disagrees with the interpretation of Mertes & Schmincke (1985) and the observation that potassic rather than sodic lavas dominate in the West Eifel field (Mertes, 1983).
Peridotite and clinopyroxenite xenoliths
The data of Mertes (1983), as well as our observations, indicate that the West Eifel volcanic field can be divided into two based on the peridotite–clinopyroxenite xenolith assemblage. To the east of the eastern edge of the ENSZD (Fig. 1), the lavas contain clinopyroxenite xenoliths only. Some of these clinopyroxenite xenoliths contain xenocrysts and rare polycrystalline xenoliths of olivine (Shaw, 2004). The clinopyroxenite xenoliths are generally interpreted to represent cumulates formed in subvolcanic magma chambers (Becker, 1977; Duda & Schmincke, 1985; Shaw, 2004). West of the eastern edge of the ENSZD, the xenolith suite comprises both mantle peridotite and clinopyroxenite cumulates.
Most of the detailed studies of metasomatism in the sub-Eifel mantle have been on samples from two localities (Dreiser Weiher and Meerfelder Maar), located near or within the zone that marks the eastern edge of the Eifel North–South Zone Depression (Fig. 1). The host lavas at these localities are characteristic of the ONB suite (Aoki & Kushiro, 1968; Mertes, 1983).
A number of studies have been carried out on mantle xenoliths hosted in F-suite lavas at volcanic centres within the ENSZD (Fig. 1), such as Gees, Baarley and Rockeskyller Kopf (Edgar et al., 1989; Lloyd et al., 1991; Zinngrebe & Foley, 1995; Pizzolato, 1997; Eyzaguirre, 1999; Shaw & Eyzaguirre, 2000; Shaw & Klügel, 2002). The nature of the metasomatic events that affected the mantle in this region is less well known. In contrast to Dreiser Weiher and Meerfelder Maar, the main xenolith type at localities between Hillesheim and Gerolstein, within the ENSZD, is wehrlite (Mertes, 1983, this study), and veins, where present, are thinner and of distinctly different mineralogy. It is worth noting that the region in which wehrlite dominates correlates with the maximum frequency of eruptive centres as defined by Mertes & Schmincke (1985). In the region west of Hillesheim, where the frequency of volcanic centres is lower, the main xenolith types are again lherzolite and harzburgite (Mertes, 1983, this study).
|
GOALS |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGY OF THE WEST...GOALSSAMPLE LOCATION AND PETROGRAPHYANALYTICAL TECHNIQUESMINERAL CHEMISTRYDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Our purpose in this study was to examine the mantle xenoliths from three localities within the ENSZD (Fig. 1) to determine:
- the petrogenesis of the veined and unveined wehrlite xenoliths and their relationship to the Quaternary vein-forming event identified by Witt-Eickschen and co-workers;
- how vein mineral assemblages and compositions differ from those described at Meerfelder Maar and Dreiser Weiher, and whether the vein assemblages correlate with the observed variations in the K/Na ratio of the host lavas.
|
SAMPLE LOCATION AND PETROGRAPHY |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGY OF THE WEST...GOALSSAMPLE LOCATION AND PETROGRAPHYANALYTICAL TECHNIQUESMINERAL CHEMISTRYDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The samples for this study were collected from scoria deposits at Rockeskyller Kopf, Baarley and Gees; a few samples were also collected from Dreiser Weiher and Meerfelder Maar (Fig. 1; Table 2). The host lavas are basanites and tephrites according to the IUGS classification (Le Bas et al., 1986
) or F-suite, group I/II (Rockeskyller Kopf, Baarley and Gees) and ONB suite, group IV (Dreiser Weiher and Meerfelder Maar) according to the classification of Mertes & Schmincke (1985).
|
As noted by a number of previous workers (e.g. Witt-Eickschen & Kramm, 1998
; Witt-Eickschen et al., 1998), the peridotitic xenolith suite can be divided into two groups—composite and single lithology. Composite xenoliths are those that are cross-cut by veins or have a variably continuous, lithologically distinct rim of material that we term the ‘selvedge’. At each locality, the selvedges and veins have the same mineralogy and texture; on this basis, we interpret selvedges to be fragments of veins.
Petrography of single lithology xenoliths
The localities shown in Fig. 1 are divided into two groups, based on the modal mineralogy of the single lithology xenoliths. At Dreiser Weiher and Meerfelder Maar, the xenolith assemblage is dominated by lherzolite and harzburgite with only minor wehrlite (Fig. 2a). These samples contain disseminated Ti-poor amphibole and phlogopite, as well as glass in veins and as inclusions (O'Connor et al., 1996; Witt-Eickschen et al., 2003a, 2003b).
|
At Gees, Rockeskyller Kopf and Baarley, the xenolith assemblage is dominated by wehrlite; harzburgite and lherzolite form only a minor (
10–15%) component of the peridotitic xenolith assemblage (Fig. 2a). Phlogopite is the most common hydrous phase at these localities, though amphibole is present at both Baarley and Rockeskyller Kopf. Lloyd et al. (1991) reported that amphibole is absent from the xenoliths at Gees. Although we have found no evidence for the occurrence of amphibole as an unaltered phase, many of the wehrlite samples contain olivine-, clinopyroxene-, spinel- and glass-bearing clusters similar to those found at Baarley and interpreted by Shaw & Klügel (2002) to be due to amphibole breakdown.
The former presence of orthopyroxene in the wehrlites is indicated by the presence of fine–medium-grained glass-bearing clusters of clinopyroxene and olivine that are similar in texture to reaction zones on orthopyroxene crystals (Tracy, 1980; Shaw et al., 1998).
Petrography of the composite xenoliths
Petrography of veins and selvedges
Selvedges and veins (hereafter referred to as veins) cut across the foliation in xenoliths (where present) and range in thickness from 1 to 15 mm (Table 2; Fig. 3a–c). The veins in peridotites from Meerfelder Maar and Dreiser Weiher are generally thicker than those found in peridotites at Baarley, Gees and Rockeskyller Kopf.
|
There is a distinct variation in the mineralogy of the veins across the field (Fig. 2b; Table 2). Veins in xenoliths hosted in F-suite lavas (Rockeskyller Kopf, Baarley and Gees) consist only of clinopyroxene and phlogopite, whereas veins hosted in ONB-suite lavas (Dreiser Weiher and Meerfelder Maar) are mainly amphibole, with lesser clinopyroxene and phlogopite.
The veins commonly show a well-developed zonation (Fig. 3d–f): with an inner zone that comprises 1–2 mm clinopyroxene ± phlogopite ± amphibole (Fig. 3d) and an outer zone, in contact with the host peridotite, that is finer-grained and has similar mineralogy to the inner zone, except that it also contains fragments of olivine (Fig. 3e). This type of zonation has also been noted laterally along an individual selvedge (Fig. 3f). All of the vein minerals are fresh and unaltered. There is no evidence of melting or reaction except adjacent to fractures filled by the host lava.
Vein–host peridotite transition
In all the samples examined, the vein–host (or selvedge–host) contact is sharp (Fig. 3a–c), although in most samples there are narrow, irregularly spaced stringers of vein minerals that project outwards (1–10 mm) into the host peridotite (e.g. Fig. 3c). In several samples, particularly those from Rockeskyller Kopf and Gees, the contact is marked by an abundance of spinel.
Orthopyroxene is absent from the transition zone adjacent to veins that cut lherzolite and harzburgite hosts. Instead, there is a significant proportion of unstrained, green, secondary diopside present. In most samples, this diopside is poikilitically intergrown with phlogopite. In most wehrlitic hosts, there is no obvious transition zone at the edge of the veins.
Host peridotite
Like the single lithology xenoliths, the peridotites that host veins vary in mineralogy with locality (Fig. 2c). At Dreiser Weiher and Meerfelder Maar, the host peridotite is mainly lherzolite or dunite (Fig. 2c; Witt-Eickschen et al., 1998), whereas at Gees, Baarley and Rockeskyller Kopf, the host is predominantly wehrlite. Disseminated phlogopite, as well as amphibole, or its breakdown products, are common in the wehrlites. On the basis of texture alone, there is no way to determine which of the metasomatic events (see Table 1) was responsible for the formation of these hydrous phases.
In addition to phlogopite and amphibole, the wehrlites typically contain abundant green poikilitic clinopyroxene similar to that found in transition zones and distinct from the colourless clinopyroxene in lherzolite and harzburgite. In a few samples, e.g. wehrlite 98KY04, it is clear that the proportion of this poikilitic clinopyroxene decreases with increasing distance from the vein and the colourless clinopyroxene increases in abundance.
|
ANALYTICAL TECHNIQUES |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGY OF THE WEST...GOALSSAMPLE LOCATION AND PETROGRAPHYANALYTICAL TECHNIQUESMINERAL CHEMISTRYDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Major element compositions of minerals were obtained at the University of Western Ontario using a JEOL 8600 Superprobe equipped with four wavelength dispersive spectrometers. Anhydrous minerals were analysed with an accelerating voltage of 15 kV, a beam current of 15 nA and a 1 micron spot size. Hydrous minerals were analysed with a reduced (10 nA) beam current. Counting times were 30 s on peaks and 15 s on each background position. Raw data were reduced using the ZAF correction built into the Tracor Northern automation system. Minimum detection limits for these analyses are 0·05 wt % oxide.
Laser ablation ICP–MS analysis of xenolith minerals was conducted at the Great Lakes Institute for Environmental Research at the University of Windsor. The facility is equipped with a non-homogenized, solid-state, 266 nm Nd-doped Y–Al garnet (Nd:YAG) laser that has 1 mJ energy/pulse and a pulse-width of 4–6 ns. The spot size of the beam at the sample is 10–15 microns. Sample ablation was conducted in an Ar gas-filled sampling cell, mounted to a modified polarizing microscope. The ablated material was transported from the ablation cell to the ICP–MS by Ar carrier gas. Analyses were performed on polished blocks and all samples were washed with ethanol prior to analysis. Trace-element concentrations in clinopyroxene and in two amphibole grains were determined in optically homogeneous, inclusion-free regions of the grains.
Elemental analyses were performed using a ThermoElemental X-7 ICP–MS, which has an average sensitivity (solution) of 4·5 x 108 counts/s/ppm for uranium. Calibration of the ICP–MS was accomplished using NIST glass standard 612. Instrument sensitivity was increased and mass bias reduced through the addition of ultra-pure N2 gas to the ICP–MS nebulizer Ar gas. The ratio of nebulizer Ar to N2 (approximately 26:1) was optimized by adjusting the gas-flow rates to obtain maximum sensitivity.
Conversion of the ICP–MS output data (integrated counts/s) to concentration units (µg/g) was accomplished using the LAMTRACE program developed by S. E. Jackson (Van Achterbergh et al., 2001). The abundance of Ca, determined using electron microprobe analysis (see below), was used as an internal standard for the calculation of the abundances of all other elements.
Independent analyses of Ca in the minerals analysed by LA–ICP–MS were obtained using the JEOL JXA-733 electron microprobe at the University of New Brunswick. The operating conditions for the analyses were the same as for the analyses performed at UWO, with the exception that raw data were reduced with the CITZAF routine.
|
MINERAL CHEMISTRY |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGY OF THE WEST...GOALSSAMPLE LOCATION AND PETROGRAPHYANALYTICAL TECHNIQUESMINERAL CHEMISTRYDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Clinopyroxene
Clinopyroxene occurs in three distinct textural settings. Colourless, variably strained type 1 clinopyroxene is the dominant type in lherzolite and harzburgite xenoliths from Dreiser Weiher and Meerfelder Maar, though it has also been found in a few wehrlite–dunite samples. Type 2 clinopyroxene is restricted to veins, is typically much coarser-grained than the other types and is intergrown with phlogopite and, where present, amphibole. The third type of clinopyroxene has a distinctive green colour in thin section and is typically poikilitic and unstrained. Type 3 clinopyroxene is the dominant type in wehrlite and dunite hosts and replaces type 1 clinopyroxene in the transition zones adjacent to veins in lherzolite and harzburgite.
Although all the grains analysed (Tables 3 and 4) fall in the diopside, Al-diopside or Al–Cr-diopside compositional field of Morimoto et al. (1988), three distinct compositional types of clinopyroxene have been identified on the basis of SiO2:TiO2 and Al2O3:Cr2O3 ratios (Fig. 4a and b). These three compositional groupings roughly correspond to the three textural types described above.
|
|
|
Type 1 clinopyroxene
Type 1 clinopyroxene is divided into two subtypes, based on small differences in major- and trace-element composition and distinct differences in textural setting. Type 1a clinopyroxene is Ti-poor, Cr–Al-diopside and diopside that occurs in the lherzolite and dunite host rocks at Meerfelder Maar and Dreiser Weiher. It is magnesian and Cr2O3-rich and is very similar in major-element composition (Fig. 4a) to Ti-poor, Cr-rich clinopyroxene found in unveined hydrous peridotite from Dreiser Weiher (Stosch & Seck, 1980
). Type 1b clinopyroxene has similar major-element characteristics to Type 1a but occurs in the cores of grains that have Ti-enriched rims in veins and their wehrlite hosts at Gees and Rockeskyller Kopf (Fig. 4a and b). These grains differ slightly from the Type 1a clinopyroxene, in that they are Na-depleted and Ca-enriched. In addition, relative to Type 1a, the Type 1b inclusions are depleted in LREE, Th, U, Nb and Sr, and slightly enriched in Ti (Table 4; Fig. 5a and b). In fact, these Ti-poor clinopyroxene inclusions in veins and wehrlite show a much closer affinity with the clinopyroxene found in wehrlitic transitional zones between lherzolite and harzburgite and hydrous veins in samples from Meerfelder Maar (Fig. 5a and b).
|
Type 2
Type 2 clinopyroxene, which is found only in veins, is Ti-rich with variable Cr2O3 (Fig. 4b), has Mg-number {atomic [Mg/(Mg + FeT)] x100} between 90 and 80·5 and Cr2O3 contents that vary between 1 wt % and the analytical detection limits with an average of around 0·2 wt %. The range in Mg-number is not a function of zonation within the veins (see below); however, individual samples show distinctly different Mg-numbers (Tables 3 and 4). Type 2 clinopyroxene is similar in REE and most incompatible trace elements to Type 1b (Table 4; Fig. 5c and d), the exception being Zr and Ti, which show a smaller negative anomaly or even a slight positive anomaly. There is no significant difference in the trace-element composition of these pyroxenes as a function of location, although clinopyroxene from one of the Gees samples (G96-1) is richer in REE than the others; however, these grains are from a 2 mm thick vein, whereas the other analyses are from selvedges that are more than 4 mm thick.
Type 2 clinopyroxene can be divided into two subtypes based on Na content (Fig. 4d). The first subtype is Na-rich and Ca-poor and is found in amphibole–phlogopite–clinopyroxene veins brought to the surface in ONB-suite lavas at Dreiser Weiher and Meerfelder Maar (as well as a single grain from Rockeskyller Kopf). The second subtype is Na-poor and Ca-rich; this subtype includes all but one analysis of vein clinopyroxene in phlogopite–clinopyroxene veins brought to the surface in F-suite lavas at Rockeskyller Kopf, Gees and Baarley.
Type 3
The third type of clinopyroxene is Ti- and Cr-rich and is found (a) in stringers and patches in the transition zone between veins and lherzolitic and harzburgitic hosts, and (b) in the wehrlite xenoliths (Tables 3 and 4). These grains are intermediate in composition between Type 1a/b and Type 2 clinopyroxene (Fig. 4) but show a much more restricted compositional range than the vein (Type 2) clinopyroxene. As was the case for Type 2 clinopyroxene, Type 3 can be broken down into two subtypes on the basis of Na and Ca content (Fig. 4c).
Type 3 clinopyroxene has similar trace-element composition to Type 2 (Fig. 5e and f), with the exception that most grains either lack a Zr-anomaly or have a positive Zr-anomaly. This contrasts with Type 2 clinopyroxene in which most grains show a negative Zr-anomaly.
Compositional traverses in a selvedge and wall rock
There is little variation in the composition of Type 2 clinopyroxene grains within a typical selvedge, from the contact with the host lava to the contact with the host peridotite. At the lava–selvedge contact, there is a distinct increase in Rb/Sr and Nb, whereas at the selvedge–host peridotite contact, there is a distinct increase in Cr2O3 (Fig. 6a–c).
|
There is also distinct compositional variation in Type 3 clinopyroxene grains between the wall rock and selvedge (Fig. 6d and e). The sample selected for this analysis has two distinct regions. The first region is close to the contact with the phlogopite–clinopyroxene selvedge. In this area, clinopyroxene is similar to that found in the selvedge, i.e. it is Cr- and Nb-poor (Fig. 6d and e) and Ti-rich. In the second region, approximately 10 mm from the selvedge–host contact, the clinopyroxene is Cr2O3- and Nb-rich and shows a pattern of decreasing Nb with increasing distance from the contact.
Witt-Eickschen et al. (1998) showed that there were distinct variations in the high-field-strength element and LREE contents of amphibole in the transition zones adjacent to veins in lherzolitic xenoliths. They noted that LREE and Nb were depleted in the transition zone, whereas Ti, Zr, and Hf were enriched. In the sample from Rockeskyller Kopf, we note a similar pattern of trace-element distribution.
Phlogopite
As was the case with clinopyroxene, three textural types of phlogopite have been identified. The first occurs as isolated tabular grains in lherzolite and harzburgite xenoliths from Meerfelder Maar. The second group is found in veins and in stringers projecting outward into the host peridotite. The third type is found mainly in wehrlite and in the wehrlitic transition zone, adjacent to a vein in sample 98DW05.
There are four distinct compositional populations of phlogopite (Fig. 7a and b; Table 5), three of which approximately correspond to the textural types described above. The fourth comprises all analysed grains from 96MF14; these are anomalously enriched in Ti. As with clinopyroxene, Type 1 phlogopite can be divided into two subtypes on the basis of textural setting. Type 1a phlogopite, found in a veined lherzolite from Meerfelder Maar, occurs as isolated grains in lherzolite, whereas Type 1b grains form cores to Type 3 phlogopite in a wehrlite from Rockeskyller Kopf. Regardless of textural setting, all Type 1 phlogopite is Ti-poor and Mg-rich (Fig. 7a and b) and is similar to disseminated phlogopite found in unveined lherzolite at Meerfelder Maar (O'Connor et al., 1996).
|
|
Type 2 phlogopite has an Mg-number of 88–80 and TiO2 contents greater than 3 wt % (Fig. 7a and b; Table 5). Phlogopite in individual veins can be grouped according to Mg-number and TiO2 content (Table 5). Three groups can be identified in the Meerfelder Maar samples: sample 96MF14 is the most magnesian, sample 97MF05 is intermediate in composition and samples 98MF12B, 96MF11 and 98MF09 are the most evolved. Similarly, the Gees and Rockeskyller Kopf veins define two groups: a high-Mg-number, low-TiO2 group (96G7, 98KY06, 98KY02A) and a lower-Mg-number, high-TiO2 group (98G03, 97G2, 97KY07, and 97KY04). The same grouping of samples can be seen for the Mg-number of clinopyroxene (Tables 3 and 4).
The third and final group have Mg-numbers from 92 to 88, with TiO2 contents from 2 to 3 wt % and are intermediate in composition between Type 2 (vein) phlogopite and Type 1 Ti-poor phlogopite (Fig. 7a and b; Table 5). These intermediate phlogopites are found in wehrlite hosts at Rockeskyller Kopf and Gees, and in the wehrlitic transition zone adjacent to a selvedge in a sample from Dreiser Weiher.
As was the case with the pyroxene data, two groups can be distinguished regardless of textural type on the basis of Na2O content. Phlogopite from Dreiser Weiher and Meerfelder Maar is invariably enriched in Na compared with that from Gees, Baarley and Rockeskyller Kopf for the same range of Mg-number (Fig. 7c). The Na2O-poor phlogopite from Baarley, Gees and Rockeskyller Kopf can be further subdivided (Fig. 7c). The Rockeskyller Kopf samples have a K2O/Na2O ratio of between 30 and 50, Gees samples have a ratio of 15–25 and the Baarley samples fall between these two end-members.
Amphibole
In common with clinopyroxene and phlogopite, there are three distinct textural types of amphibole. Type 1 amphibole is tabular and occurs as isolated grains in lherzolitic–harzburgitic host xenoliths. Type 2 amphibole is found only in veins in samples from Dreiser Weiher and Meerfelder Maar. The third type of amphibole is found in the wehrlite xenoliths at Baarley and Rockeskyller Kopf. This type also includes the amphibole found in wehrlitic transition zones adjacent to veins in lherzolitic hosts.
All of the grains analysed (Table 6; Fig. 8) range in composition from pargasitic hornblende through K–Ti pargasite to K–Ti magnesiohastingsite (according to Leake, 1978). Like clinopyroxene and phlogopite, the compositional types of amphibole roughly correspond to the textural types.
|
|
Ti-poor, Cr-rich Type 1 amphibole (Fig. 8a and b), like phlogopite and clinopyroxene, is divided into two subtypes on the basis of texture. Type 1a amphibole occurs in lherzolitic hosts to veins at Meerfelder Maar and is similar to the amphibole found as a disseminated phase in unveined xenoliths at this locality (Witt-Eickschen et al., 1998
). Type 1b amphibole occurs as Ti-poor cores to more Ti-rich Type 3 grains in a wehrlite from Baarley (Fig. 8a and b). In contrast to Type 1 clinopyroxene, which shows distinct differences in Na2O content, depending on whether the host lava is F-suite or ONB suite (Fig. 4c and d), disseminated amphibole shows no distinct difference in Na2O content with locality (Fig. 8c).
Type 2 amphibole occurs in veins and in stringers projecting from veins in samples from Meerfelder Maar and Dreiser Weiher. This type is enriched in Ti and is generally poorer in Cr than the first group, with an Mg-number that ranges down to 77 (Fig. 8a and b; Table 6). Where data from multiple samples are available, e.g. at Meerfelder Maar, the veins can be grouped by Mg-number (Table 6). Trace-element data for amphibole from one selvedge (Table 6) are within the range of compositions already reported by Witt-Eickschen et al. (1998) and are not considered further here.
Type 3 amphibole is intermediate in composition between Types 1 and 2, and has been found in transition zones adjacent to veins at Dreiser Weiher and Meerfelder Maar. This type has also been found in wehrlite xenoliths at Baarley and Rockeskyller Kopf that host amphibole-free veins (Fig. 8b).
Olivine
Olivine in lherzolite and harzburgite xenoliths is typically more magnesian than that in wehrlite, with average compositions being Fo89 and Fo86, respectively (see Shaw, 2004). Olivine in the wall rocks shows distinct variation in composition with proximity to veins. In sample 98DW05, there is a decrease in Fo content from the normal value of Fo90 at 25 mm from the selvedge–host contact to Fo86 at the contact. Similarly, for a vein hosted in a wehrlite (98G03), there is a decrease in Fo content from Fo88 at 15 mm from the vein–host contact to Fo82 at the contact. We interpret these compositional variations to be due to interdiffusion of Fe and Mg between the vein and wall rock over a period of 1400–2500 years [based on an Fe–Mg interdiffusion coefficient of 2·7 x 10–11 cm2/s at a temperature of 1200°C; calculated using data from Buening & Busek (1973) and Chakraborty (1997)]. This is of the same order of magnitude as estimates of the post-reaction annealing time for transition zone amphibole made by Witt-Eickschen et al. (1998), but is much greater than the 10–15 years that they estimated for the wall-rock–melt interaction.
Spinel
Spinel in samples that host phlogopite–clinopyroxene veins (Rockeskyller Kopf, Baarley and Gees) has higher Fe2O3 and Cr2O3 than that in samples that host amphibole–phlogopite–clinopyroxene veins (Dreiser Weiher and Meerfelder Maar). In all cases, spinel in harzburgite–lherzolite hosted veins shows a distinctly bimodal character, with significantly higher Fe2O3 in the vein spinels compared with those in the host (Table 7). Spinel in veined wehrlites shows pronounced zonation from Cr-rich, Fe- and Al-poor cores to Cr-depleted, Fe- and Al-enriched rims (Table 7).
|
Geothermometry
Olivine–spinel thermometry, using the calibration of Balhaus et al. (1991)
and an assumed equilibration pressure of 1·5 GPa, gives a bimodal distribution for the equilibration temperature of the peridotitic hosts to veins (Fig. 9). The lherzolitic hosts to amphibole–phlogopite–clinopyroxene veins yield equilibration temperatures of between 850 and 900°C (see also Witt & Seck, 1987, 1989), whereas the wehrlites and dunites that host both phlogopite–clinopyroxene and amphibole–phlogopite–clinopyroxene veins give equilibration temperatures of between 900 and 1100°C (Fig. 9).
|
The very wide range of temperatures calculated for wehrlites in part reflects differences between localities and rock types and suggests either large temperature gradients in the mantle or disequilibrium between olivine and spinel. The temperature difference between the lherzolite–harzburgite xenoliths and the wehrlites is also reflected in the Ca content of clinopyroxene. Clinopyroxene from high-temperature wehrlites typically has much higher Ca content than that from the lherzolites and harzburgites (Fig. 4c), even at Meerfelder Maar, where all the clinopyroxenes are Ca-depleted relative to those at Rockeskyller Kopf, Baarley and Gees. The same bimodal distribution of Ca contents in the vein clinopyroxene (Fig. 4d) may indicate either that the temperature of the vein-forming magma was lower at Meerfelder Maar and Dreiser Weiher than at the Rockeskyller Kopf, Gees and Baarley or the vein-forming magmas had distinctly different Ca contents. An alternative explanation is that high Ca content in wehrlite clinopyroxene may simply reflect the absence of orthopyroxene. In the presence of orthopyroxene, the Ca content of clinopyroxene is buffered to lower values because some of the orthopyroxene will dissolve in the clinopyroxene. Elimination of orthopyroxene by reaction with an infiltrating melt would have removed this buffering effect, resulting in Ca enrichment in clinopyroxene.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGY OF THE WEST...GOALSSAMPLE LOCATION AND PETROGRAPHYANALYTICAL TECHNIQUESMINERAL CHEMISTRYDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Witt-Eickschen and coworkers have developed a multistage model for the evolution of the sub-Eifel mantle, based on their studies of single lithology and composite peridotitic xenoliths from Dreiser Weiher and Meerfelder Maar (see Table 1). Our samples from Rockeskyller Kopf, Gees and Baarley are significantly different in their modal mineralogy and composition (Table 8) and cannot, at first glance, be fitted into the current model for the evolution of the Eifel mantle (Witt-Eickschen et al., 2003a
). The differences between the localities on the eastern edge of the ENSZD and those within it raise three major questions:
- Is there evidence for the Hercynian metasomatic event at Rockeskyller Kopf, Gees and Baarley?
- What was the composition of the magma responsible for vein formation and do differences in the mineralogy of veins and their host xenoliths correlate with differences in magma composition?
- Why is wehrlite the dominant peridotite type at Rockeskyller Kopf, Baarley and Gees and are wehrlite and vein petrogeneses linked?
|
Evidence of a Hercynian metasomatic event?
Witt-Eickschen et al. (2003a
and references therein) suggested that the Ti-poor, LREE- and MREE-enriched and Zr- and Hf-depleted amphibole and clinopyroxene in lherzolitc xenoliths formed during a pervasive metasomatic event. On the basis of isotopic data, they suggested that the mantle was affected by fluids carrying a crustal isotopic signature, probably during the Hercynian orogeny. This event occurred at relatively low temperatures (900–970°C) and, given its identification in xenoliths of the West and East Eifel Volcanic Fields, was interpreted to be regionally distributed.
What, then, is the evidence that this event also affected the mantle below Rockeskyller Kopf, Gees and Baarley? Although they are rare, lherzolites and harzburgites do occur at these localities and have similar modal mineralogy and mineral compositions to lherzolite and harzburgite at Meerfelder Maar (Lloyd et al., 1991; Shaw & Klügel, 2002). The wehrlite samples from Gees and Rockeskyller Kopf contain Ti-poor cpx; Ti-poor phlogopite occurs at Rockeskyller Kopf and Ti-poor amphibole is preserved in samples from Baarley. All these grains occur as cores to Ti-rich grains, indicating that they pre-date the formation of the rims. Although trace-element data are not available for all of these samples, the major-element compositions indicate a close similarity to clinopyroxene, phlogopite and amphibole associated with the Hercynian metasomatic event.
In detail, the Type 1b clinopyroxene at Gees and Rockeskyller Kopf has a transitional character in its trace-element content, indicating that these grains, like the pyroxenes in transition zones adjacent to veins in lherzolite and harzburgite (Witt-Eickschen et al., 1998), have been affected by a later overprinting event.
Although we do not have isotopic data for the samples from Gees, Rockeskyller Kopf and Baarley, we suggest that the similarity of major- and trace-element data for Type 1b minerals in these samples to the Ti-poor secondary mineral assemblages associated with the Hercynian metasomatism on the basis of isotopic data is too close to be coincidental and that the Hercynian metasomatic event affected the mantle below these localities.
Petrogenesis of veins
Witt-Eickschen et al. (1998) suggested that the hornblendite veins at Meerfelder Maar were a product of high-pressure crystallization of Quaternary Eifel magmas on the basis of close similarities in trace-element contents. Similarly, Shaw & Eyzaguirre (2000) suggested that many of the megacrysts found in the Eifel lavas were fragments of veins formed from high-pressure crystallization of alkaline magmas. Three lines of evidence suggest that the same interpretation can be applied to veins from Rockeskyller Kopf, Baarley and Gees. First, modelling of parent magma compositions using trace-element data (Fig. 10a–c) indicates that the calculated parent magmas to the veins are similar to Quaternary Eifel lavas. Secondly, there are distinct differences in vein mineralogy that correlate with the lava types erupted: phlogopite–clinopyroxene veins are found at Rockeskyller Kopf, Baarley and Gees, where the host lava belongs to the potassic F-suite, whereas amphibole–phlogopite–clinopyroxene veins are found only in the more sodic ONB lavas at Dreiser Weiher and Meerfelder Maar. Finally, there are distinct differences in the composition of vein minerals that also reflect differences in the host lava composition between localities. Clinopyroxene and phlogopite from Rockeskyller Kopf, Baarley and Gees are Na-poor, like the host lavas. Conversely, clinopyroxene and phlogopite from the sodic, ONB-hosted veined xenoliths from Dresier Weiher and Meerfelder Maar are Na-enriched.
|
Differences in the Ca content of clinopyroxene also appear to reflect differences in the bulk composition of the magmas. Clinopyroxene from veins in xenoliths from Rockeskyller Kopf, Baarley and Gees is Ca-rich, like the host F-suite lavas that contain an average of 14·6 wt % CaO. On the other hand, the clinopyroxene in veins hosted in ONB-suite lavas is Ca-poor, as is the host which contains
12–13 wt % CaO (Table 9).
|
The difference in the parent magma composition of ONB- and F-suite hosted veins is also reflected in small differences in the calculated trace-element compositions (Fig. 10d). Calculated magma compositions for veins from Gees and Rockeskyller Kopf are enriched in Rb, Ba, Sr and Nb and depleted in Th and U relative to the calculated parent magma to the Meerfelder Maar veins. These differences, particularly in terms of CaO, Na/K, Rb, Sr and Nb (Fig. 10; Table 9) are typical of the differences between the two lava suites (Mertes, 1983
; Mertes & Schmincke, 1985) and further support the interpretation that the phlogpite–clinopyroxene veins were formed from F-suite lavas, whereas the sodic amphibole–phlogopite–clinopyroxene veins formed from ONB-type lavas.
Origin of wehrlite
The domination of the peridotite xenolith assemblage by wehrlite at Rockeskyller Kopf, Gees and Baarley is anomalous within the West Eifel, where the dominant peridotite types are lherzolite, harzburgite and dunite. Lloyd et al. (1991) suggested that the original mantle below Gees was harzburgite that was transformed via reaction with a Ca- and alkali-rich infiltrating melt; this hypothesis was supported by the work of Zinngrebe & Foley (1995), who suggested that the melts responsible for wehrlite formation were of calc–alkaline andesite composition. However, our results suggest that wehrlite formation is directly linked to the vein-forming event in the mantle below Rockeskyller Kopf, Baarley and Gees, i.e. that wehrlite formed due to reaction between mantle and infiltrating silica-undersaturated, alkaline magma during the Quaternary magmatism in the West Eifel.
There are four main lines of evidence that support our interpretation. First, the secondary nature of the wehrlites is shown by the presence of relic grains of clinopyroxene and, in some cases, amphibole and phlogopite (Type 1b minerals) within the wehrlites that are similar in composition to minerals found in lherzolites and harzburgites that were only affected by the Hercynian metasomatic event (Type 1a). Secondly, there is abundant textural evidence for consumption of orthopyroxene and formation of secondary olivine and clinopyroxene in the wehrlites. This is particuarly clear at Gees (Zinngrebe & Foley, 1995) but can also be seen at Baarley (Shaw & Klügel, 2002) and Rockeskyller Kopf. Thirdly, the higher temperatures recorded by the wehrlites—50–200°C higher than those of lherzolites and harzburgites—indicate that wehrlite was accompanied by heating consistent with the infiltration of melt. Fourthly, the secondary mineral assemblage in both the discrete wehrlite xenoliths and the transition zones is identical to the vein mineral assemblage. This interpretation is strengthened by the observation that the major- and trace-element signatures of wehrlite minerals (Type 3) and those of the vein minerals (Type 2) are very similar, suggesting a direct link between the two (Figs 4, 5, 7 and 8). Minor differences between the veins and wehrlites can be attributed to chromatographic fractionation or buffering by the residual olivine in the wehrlites.
The reaction required to form wehrlite from lherzolite and harzburgite is consumption of orthopyroxene and formation of secondary olivine and orthopyroxene. Shaw (1999) showed that such a reaction, with a basanite as the solvent, was capable of producing the required mineral assemblage in a short period of time (hours). Orthopyroxene consumption also produces the wide range of secondary-melt compositions, now preserved as glass in the Eifel xenoliths (Edgar et al., 1989; Zinngrebe & Foley, 1995; Shaw & Klügel, 2002). In addition, this reaction produces olivine that is less magnesian than that in lherzolites and harzburgites (Shaw, 1999), which is a viable explanation for the lower Fo content of olivine in wehrlites. However, not all of the wehrlite olivine formed via orthopyroxene breakdown; at least some of it was originally present in the precursor and was enriched in iron via inter-diffusion with the infiltrating melt. As noted above, this process would have taken several thousand years to complete, although this does not necessarily mean that the magma–mantle reaction event was of this duration. The re-equilibration could have occurred in the absence of melt ([compare the annealing of amphibole reported by Witt-Eickschen et al. (1998)].
Clinopyroxene formed during orthopyroxene reaction is not a direct product of the breakdown reaction but rather a result of diffusion of Ca into the reaction zone that resulted in clinopyroxene saturation (Shaw et al., 1998; Shaw, 1999). During the formation of wehrlite, there may also have been some direct crystallization of clinopyroxene from the infiltrating melt. It is possible that clinopyroxene formed by orthopyroxene reaction and that due to direct crystallization was of different composition; however, over the annealing period that is required for olivine homogenization, clinopyroxene would likely have been at least partially homogenized via diffusion, though, for example, in G96-1 (Table 4), the range of clinopyroxene compositions suggest that homogenization was not complete. Similarly, pre-existing clinopyroxene, related to the Hercynian metasomatic event, would have been partially re-equilibrated and would have inherited at least some of the characteristics of the infiltrating magma. This explains the bimodal distribution of Na and Ca contents of Type 1b clinopyroxene: the original clinopyroxene at all localities would have been Ca-poor and Na-rich, i.e. similar in composition to that from Meerfelder Maar and Dreiser Weiher. During the wehrlitization event, inter-diffusion between this primary clinopyroxene and melt would have increased the Ca content and decreased the Na content.
Stosch & Lugmair (1986) suggested that the trace-element composition of clinopyroxene from a wehrlite in the Dreiser Weiher lavas indicates an origin involving a silicate melt. However, on the basis of Nd- and Sr-isotopic relations, they ruled out a direct origin from West Eifel magma, instead favouring a reaction mechanism similar to that proposed above. Stosch & Lugmair (1986) also noted that Sr-isotopic disequilibrium between phlogopite and amphibole in a Dreiser Weiher wehrlite was related to cooling of the wehrlitic mantle to a temperature where isotopic exchange ceased. In light of the arguments presented above, we suggest that the disequilbrium actually reflects two metasomatic events: an early one in which disseminated clinopyroxene, amphibole and phlogopite formed (Hercynian) and a later one where phlogopite, clinopyroxene (± amphibole) formed via reaction with infiltrating silicate melt. Overprinting of trace-element signatures and incomplete isotopic exchange between the minerals formed in these two events could have led to the observed disequilibrium. The details of the process would not have been resolvable from the data of Stosch & Lugmair (1986) because they collected their trace-element and isotopic data from mineral separates.
Timing and relationship between veins and wehrlite formation
The relative timing of vein and wehrlite formation can be interpreted from compositional profiles in the wall rock adjacent to veins. In the single sample for which we have data, the trace-element compositional gradients, measured in clinopyroxene at the vein margin (Fig. 6d and e), suggest that a pre-existing mineral assemblage in the wall rock was overprinted during the veining episode. The ‘unaltered’ clinopyroxene in the wall rock is not homogeneous in composition, but nevertheless is typical of Type 3. This relationship suggests that in this case, the formation of the wehrlite mineral assemblage pre-dated the formation of the vein. Similarly, in wehrlite sample 98G03, the systematic variation of olivine compositions in the wall rock to a vein suggests that the veining event overprinted the earlier olivine compositions.
A model for wehrlite and vein formation
Given the widespread occurrence of wehrlite at Rockeskyller Kopf, Gees and Baarley, we interpret that it formed by magma–mantle reaction during porous flow, whereas veins are the product of more focused flow. Yardley (1986) suggested that the pervasive permeability of a fluid is dependent upon the effective pressure and that high fluid pressures will increase the permeability. However, if the permeability rises further, the rock will fracture, leading to focused flow. At Gees, Baarley and Rockeskyller Kopf, it is likely that fluid pressure was sufficiently high that melt moved mainly along grain boundaries, leading to widespread formation of wehrlite. However, in general, the magma pressure was not high enough to induce widespread fracturing. Local increases in magma pressure would have caused focused flow, leading to the formation of veins. The variation in mineral Mg-number from vein to vein may reflect different episodes of over-pressure; alternatively, this variation may reflect fractionation of a single magma.
The remaining question is, why are wehrlites more common at Rockeskyller Kopf, Gees and Baarley compared with the two localities at the eastern edge of the ENSZD? There are several possibile reasons that are probably linked:
- Magma flow through the mantle was more voluminous at Rockeskyller Kopf, Gees and Baarley, as