Journal of Petrology Advance Access originally published online on April 27, 2006
Journal of Petrology 2006 47(8):1637-1671; doi:10.1093/petrology/egl023
© The Author 2006. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: journals.permissions@oxfordjournals.org
Petrogenesis of Tertiary Mafic Alkaline Magmas in the Hocheifel, Germany
CAROLINE JUNG1,
STEFAN JUNG1,2,*,
EDGAR HOFFER1 and
JASPER BERNDT3
1 INSTITUT FÜR MINERALOGIE, PETROLOGIE UND KRISTALLOGRAPHIE, FACHBEREICH GEOWISSENSCHAFTEN, PHILIPPS UNIVERSITÄT MARBURG LAHNBERGE/HANS-MEERWEIN-STRASSE, 35032 MARBURG, GERMANY
2 MAX-PLANCK-INSTITUT FÜR CHEMIE, ABT. GEOCHEMIE POSTFACH 3060, 55020 MAINZ, GERMANY
3 INSTITUT FÜR MINERALOGIE, UNIVERSITÄT MÜNSTER CORRENSSTR. 24, 48149 MÜNSTER, GERMANY
RECEIVED
JANUARY 11, 2005;
ACCEPTED
MARCH 24, 2006
 |
ABSTRACT
|
|---|
Primitive nephelinites and basanites from the Tertiary Hocheifel
area of Germany (part of the Central European Volcanic Province;
CEVP) have high Mg-number (>0·64), high Cr and Ni
contents and strong light rare earth element enrichment but
systematic depletion in Rb, K and Ba relative to trace elements
of similar compatibility in anhydrous mantle. Alkali basalts
and more differentiated magmatic rocks have lower Mg-number
and lower abundances of Ni and Cr, and have undergone fractionation
of mainly olivine, clinopyroxene, Fe–Ti oxide, amphibole
and plagioclase. Some nephelinites and basanites approach the
Sr–Nd–Pb isotope compositions inferred for the EAR
(European Asthenospheric Reservoir) component. The Nd–Sr–Pb
isotope composition of the differentiated rocks indicates that
assimilation of lower crustal material has modified the composition
of the primary mantle-derived magmas. Rare earth element melting
models can explain the petrogenesis of the most primitive mafic
magmatic rocks in terms of mixing of melt fractions from an
amphibole-bearing garnet peridotite source with melt fractions
from an amphibole-bearing spinel peridotite source, both sources
containing residual amphibole. It is inferred that amphibole
was precipitated in the asthenospheric mantle beneath the Hocheifel,
close to the garnet peridotite–spinel peridotite boundary,
by metasomatic fluids or melts from a rising mantle diapir or
plume. Melt generation with amphibole present suggests relatively
low mantle potential temperatures (<1200°C); thus the
mantle plume is not thermally anomalous. A comparison of recently
published Ar/Ar ages for Hocheifel basanites with the geochemical
and isotopic composition of samples from this study collected
at the same sample sites indicates that eruption of earlier
lavas with an EM signature was followed by the eruption of later
lavas derived from a source with EAR or HIMU characteristics,
suggesting a contribution from the advancing plume. Thus, the
Hocheifel area represents an analogue for magmatism during continental
rift initiation, during which interaction of a mantle plume
with the overlying lithosphere may have led to the generation
of partial melts from both the lower lithosphere and the asthenosphere.
KEY WORDS: alkali basalts; continental volcanism; crustal contamination; partial melting; Eifel, Germany
|
INTRODUCTION
|
|---|
The geochemistry of primitive, alkaline mafic volcanic rocks
together with geochemical evidence from mantle-derived xenoliths
can potentially yield valuable information about the nature
of the inaccessible parts of the Earth's upper mantle. One problem
usually addressed in studies on volcanic rocks erupted in continental
areas is the identification of the source region of the alkaline
magmas—either the subcontinental lithospheric mantle (SCLM)
or a sublithospheric source, e.g. a mantle plume. In contrast
to the suboceanic mantle, the subcontinental lithospheric mantle
is likely to have remained isolated from the convecting upper
mantle after initial crustal extraction. This upper mantle can
have had a complex geological history, involving ancient depletion
events, followed by later re-enrichments, most probably from
metasomatizing fluids or melts (e.g. Hawkesworth
et al., 1990
).
The relatively large volumes of alkaline mafic magmas (nephelinites,
basanites, alkali basalts) that are generated during continental
rifting are often considered to be produced predominantly by
partial melting of asthenospheric mantle (McKenzie & Bickle,
1988
; White & McKenzie, 1989
; Wilson & Downes, 1991
;
Arndt & Christensen, 1992
). On the other hand, there is
growing evidence that, at least in some volcanic provinces,
magmas are generated by partial melting of metasomatically enriched
SCLM (Hawkesworth
et al., 1990
; Gallagher & Hawkesworth,
1992
; Bradshaw
et al., 1993
). In this respect, the SCLM can
contribute in several ways to the composition of continental
basalts including: (1) mixing of small-degree melt fractions
from the lithosphere with magmas from the asthenosphere (Ellam
& Cox, 1991
); (2) direct melting of the lithosphere (Bradshaw
et al., 1993
); (3) partial melting of detached fragments of
SCLM recycled into the asthenosphere during earlier tectonic
episodes (Zindler & Hart, 1986
).
In central Germany, several thousand km3 of mafic magma were generated during the Tertiary and constitute part of the Central European Volcanic Province (CEVP, Fig. 1). These volumes are probably too large to have been produced solely within the lithosphere (Wedepohl, 1985; Jung & Hoernes, 1998; Jung & Masberg, 2000; Bogaard & Wörner, 2003). Consequently, it has been suggested that the individual volcanic fields of the CEVP (Massif Central, Bohemian Massif, Eifel, Siebengebirge, Westerwald, Hessian Depression, Rhön, Vogelsberg, Urach, etc.) were fed from asthenospheric partial melts generated within upwelling upper mantle plumes or diapirs (Granet et al., 1995; Wilson & Patterson, 2001). The trace element and Sr–Nd–Pb isotope geochemistry of the most primitive alkaline mafic lavas from the CEPV suggest derivation from mantle sources similar to those of ocean island basalts (Wörner et al., 1986; Blusztajn & Hart, 1989; Wilson & Downes, 1991; Hegner et al., 1995; Wilson et al., 1995; Jung & Masberg, 1998; Wedepohl & Baumann, 1999; Jung & Hoernes, 2000). Moreover, with the advent of high-resolution mantle seismic tomography, showing slow velocity domains in the mantle at various depths, an upper mantle origin for the European volcanism linked to a series of diapiric upwellings has been suggested (Granet et al., 1995; Hoernle et al., 1995; Goes et al., 1999; Ritter et al., 2001; Wilson & Patterson, 2001; Keyser et al., 2002). For the Rhenish Massif, a columnar low P-wave velocity anomaly was detected beneath the Eifel (Ritter et al., 2001). This 100 km wide structure extends up to 400 km depth and could be interpreted to be equivalent to an excess mantle temperature of 150–200°C in the absence of volatiles or partial melts. The geochemical diversity of lithospheric mantle xenoliths entrained within the mafic magmas of the CEPV indicates that, locally, subduction zone processes during the Hercynian orogeny may have induced substantial trace element and isotopic enrichment of the lithospheric mantle beneath central Europe (e.g. Witt-Eickschen & Kramm, 1997). Partial melting of such zones of metasomatized lithospheric mantle, combined with interaction between asthenospheric melts and lithospheric melts, has been proposed to explain the geochemical characteristics of the most primitive mafic alkaline rocks of the CEPV (Wilson & Downes, 1991; Granet et al., 1995; Wilson & Patterson, 2001). In addition, crustal contamination of the mantle-derived magmas has been widely documented within the CEVP (Massif Central: Wilson et al., 1995; Vogelsberg: Jung & Masberg, 1998; Bogaard & Wörner, 2003; Rhön: Jung & Hoernes, 2000; Jung et al., 2006; Urach–Hegau: Blusztajn & Hegner, 2002; Westerwald: Haase et al., 2004). Elucidation of the details of interaction of asthenosphere-derived melts with the lithosphere (both crust and mantle) is often difficult because both crust and ancient subcontinental mantle can have similar geochemical and Sr–Nd–Pb isotope characteristics.
Given the well-characterized nature of the crust and mantle
lithosphere in the Eifel (Stosch & Lugmair, 1984
, 1986
;
Loock
et al., 1990
; Rudnick & Goldstein, 1990
; Stosch
et al., 1992
; Witt-Eickschen & Kramm, 1998
; Witt-Eickschen
et al., 1998
, 2003
), this area provides an ideal setting in
which to attempt to identify the sources of this specific type
of intra-plate volcanism. However, neither comprehensive whole-rock
geochemistry nor Sr–Nd–Pb isotopic data have been
published for the Tertiary Hocheifel area. Such data are essential
to constrain the role of fractional crystallization, crustal
contamination, and mantle source heterogeneities in the petrogenesis
of the magmas. In this study, major and trace element data and
Sr–Nd–Pb isotope data are reported for primitive
alkaline mafic magmas from the Hocheifel area; these data are
used to constrain the mantle source region of these basalts.
Major and trace element and Sr–Nd–Pb isotope data
for more differentiated lavas from the same area are used to
highlight the effects of fractional crystallization and crustal
contamination in the genesis of these alkaline lavas.
|
GEOLOGICAL SETTING
|
|---|
The Hocheifel volcanic field has an areal extent of
1400 km
2 and consists mainly of deeply eroded volcanic plugs and necks.
Locally, remants of lava flows appear. The Hocheifel volcanic
field forms part of an east–west-trending belt of Tertiary–Quaternary
volcanic fields in central Germany; these include the Eifel,
Westerwald, Vogelsberg, Hessian Depression, Rhön, Heldburg
and Oberpfalz (
Fig. 1a). The trend of these volcanic fields
is perpendicular to the main NNE–SSW-trending Rhine graben
rift system of Central Europe, which has been interpreted as
the result of Alpine tectonism further south (e.g. Ziegler,
1992
). In Germany and elsewhere in Central Europe, Tertiary
(mainly Miocene to Pliocene) basin development provides evidence
for lithospheric extension, although the huge volumes of basaltic
rocks in the Vogelsberg area (
500 km
3) and the Cantal (Massif
Central, France) are unlikely to be attributed to continental
extension alone. As noted by Wilson & Downes (1991)
, most
of the major volcanic fields sit on uplifted Variscan basement
massifs. However, basement uplift is not coeval with rift development,
typically starting some 20–40 Myr after the beginning
of rifting (Ziegler, 1992
). Within Central Germany, although
some of the Cenozoic volcanic fields are located mainly on Hercynian
fault blocks within the Rhenish Massif (e.g. Eifel, Westerwald,
Heldburg), Tertiary volcanic activity in the Rhön, the
Hessian Depression and the Vogelsberg occurs within graben-like
structures that transect the Rhenish Massif. In the Rhön
area and Hessian Depression magmatism is not obviously associated
with basement uplift, and dextral strike-slip movement of the
lithosphere probably caused passive asthenospheric upwelling
(Schreiber & Rotsch, 1998
).
Geophysical data indicate that the Cenozoic rifts of the CEVP are associated with a marked uplift of the Moho discontinuity. The maximum crustal thinning coincides with the trace of the northern Rhine graben, although this area has been shown to be largely non-magmatic (Wilson & Patterson, 2001). Crustal thickness beneath the Eifel is estimated to be between 28 and 32 km (Mengel et al., 1991; Prodehl et al., 1992). Babuska & Plomerová (1988) estimated a lithosphere thickness of 100–140 km prior to the Cenozoic rifting and suggested a present-day depth of less than 60 km for the asthenosphere–lithosphere boundary beneath the Rhenish Massif.
Volcanism within the CEVP spans the entire Cenozoic period (Wilson & Downes, 1991); in the Hocheifel area it appears to have ranged from middle Eocene to late Oligocene (45 Ma–24 Ma; Lippolt, 1982), although new Ar–Ar data suggest a smaller age range and two distinct periods of activity (44–40 Ma and 38–34 Ma; Fekiacova et al. 2003). The magmatic rocks are mainly basanites, nephelinites and alkali basalts plus rare hawaiites, mugearites, benmoreites and trachytes. Volcanism in the neighbouring East and West Eifel produced about 300 small-volume monogenetic centres between 700 and 10·8 ka BP (Schmincke et al., 1983; Mertes & Schminke, 1985; Wörner et al., 1985). Two geochemically, spatially and temporally distinct groups of sodic–potassic alkaline volcanic rocks were erupted in the East Eifel. In the NW nephelinites, leucitites and more differentiated rocks were erupted >400 kyr ago whereas in the SE basanites and more differentiated rocks erupted between 400 and 10 ka BP. The west Eifel volcanic field consists of leucitites, basanites and nephelinites, which cover an area of 600 km2 and erupted between 700 and 10 ka BP (Mertes & Schmincke, 1985, and references therein). Wilson & Downes (1991) suggested that the most primitive mafic alkaline volcanic rocks have major and trace element and Nd–Sr–Pb–O isotope systematics that suggest the involvement of both lithospheric and asthenospheric mantle source components in their petrogenesis. The sodic magma types (melilitites, nephelinites, basanites, alkali olivine basalts) originated by partial melting of a common asthenospheric mantle source, termed the EAR (European Asthenospheric Reservoir), whereas the potassic lavas (leucitites, leucite basanites) were derived from locally enriched portions of the mantle lithosphere. The Hercynian basement through which the magmas erupted consists mainly of greenschist- to amphibolite-facies metapelites, metabasites and orthogneisses of the Mid German Crystalline Rise and is overlain by Palaeozoic (Lower to Upper Devonian) limestones and sandstones and Cenozoic (Triassic) sandstones, carbonates and clays (Mengel et al., 1991, and references therein).
|
ANALYTICAL TECHNIQUES
|
|---|
Sixty-five samples were taken from the Hocheifel volcanic field
according to the distribution of deeply eroded volcanic edifices
given by Huckenholz & Büchel (1988)
. Based on the petrographic
descriptions and average chemical compositions given by Huckenholz
& Büchel (1988)
, a significant number of accessible
sample sites were re-investigated to cover the entire range
of lava compositions (
Fig. 1b and
Table 1). All samples were
taken from remnants of lava flows, plugs and necks that cover
the entire volcanic field. Whole-rock samples were prepared
by crushing in an agate shatterbox to obtain
250 g of the macroscopically
freshest material. Aliquots were analysed for major and trace
elements in fused lithium tetraborate glass beads using standard
X-ray fluorescence (XRF) techniques (Vogel & Kuipers, 1987
)
at the Mineralogical–Petrological Department at the University
of Bonn. Rare earth elements (REE) were analysed by inductively
coupled plasma atomic emission spectrometry (ICP-AES) following
separation of the matrix elements by ion exchange (Heinrichs
& Herrmann, 1990
) at the Department of Mineralogy, Petrology
and Crystallography at the University of Marburg. Loss on ignition
(LOI) was determined gravimetrically at 1050°C (Lechler
& Desilets, 1987
) and FeO was measured titrimetrically using
standard techniques. Accuracy was monitored by repeated measurements
of international and in-house standards; the results are in
good agreement with the recommended values for the international
rock standard JB 2 given by Govindaraju (1994)
(
Table 2).
Pb, Sr and Nd isotope analyses were carried out at the Max-Planck-Institut
für Chemie at Mainz by thermal ionization mass spectrometry
using a Finnigan MAT 261 multiple sample, multicollector mass
spectrometer operating in the static mode. Whole-rock chips
were leached in 6N HCl for at least 2 h on a hotplate. Subsequently,
the samples were washed three times with ultrapure H
2O. After
this treatment, the samples were dissolved in concentrated HF
and after evaporation redissolved in 2·5N HCl and 0·6N
HBr and loaded on Teflon
® columns filled with DOWEX
® AG 1
x8 anion exchange resin (100–200 mesh) in chloride
form (Mattinson, 1986
). The Pb was extracted using conventional
HBr–HCl techniques and was loaded on Re single filaments
following the H
3PO
4–silica gel method (Cameron
et al.,
1969
). Strontium and REE were separated by using standard cation
exchange columns with a DOWEX
® AG 50 W-X 12 resin using
2·5N HCl for Sr and 6N HCl for the REE. Nd was separated
from the other REE by using HDEHP-coated Teflon
® columns
and 0·12N HCl. Neodymium isotopes were normalized to
146Nd/
144Nd = 0·7219. Repeated measurements of the La
Jolla Nd standard gave
143Nd/
144Nd = 0·511848 ±
0·000021 (2
;
n = 28). The reproducibilitiy of the Sr
standard (NBS 987) is
87Sr/
86Sr = 0·710224 ± 0·000024
(2
;
n = 14) and the fractionation was corrected to
86Sr/
88Sr
= 0·1194. Lead analyses were corrected for mass fractionation
by a factor of 0·11% per a.m.u. The reproducibility of
the standard NBS 982 was estimated to be 0·068%, 0·064%
and 0·071% for the
206Pb/
204Pb,
207Pb/
204Pb and
208Pb/
204Pb
ratio, respectively. The total procedure blank is <60 pg
Pb during this study and is therefore considered negligible.
|
PETROGRAPHY
|
|---|
All samples are porphyritic and contain partly altered olivine
and clinopyroxene phenocrysts. Usually, the majority of the
olivine and clinopyroxene phenocrysts have grain sizes of
1–2
mm and 0·5–5 mm, respectively, and the samples
containing such phenocrysts do not appear to be accumulative.
Rare olivines in some basanites and nephelinites have a mosaic
texture and incipient undulose extinction. They probably represent
entrained material from disintegrated peridotite xenoliths.
Chemically, these olivines appear to be enriched in MgO and
depleted in FeO relative to the dominant euhedral olivines (
Table 2).
The dominant olivines within the nephelinites and basanites
are euhedral phenocrysts with sharply defined crystal edges.
Sometimes, the olivines are more skeletal with evidence of marginal
resorption, re-entrants and internal cavities. These olivines
have slightly lower MgO but higher FeO than the olivines, with
incipient undulose extinction (
Table 2). Both types of olivine
show a narrow rim with depletion of MgO and enrichment of FeO.
Zoned clinopyroxenes are generally composed of a subhedral to anhedral colourless to pale brown core and a darker brown, slightly pleochroic mantle. The core of these clinopyroxenes is usually more MgO-poor and FeO-rich than the rim (Table 2). The evolution of clinopyroxene in the sequence basanite–mugearite–tephrite shows increasing MgO but decreasing FeO, suggesting a complex pre-eruption history (Fig. 2a). In some samples, clinopyroxenes with an olive green to light green core, a colourless to pale brown mantle and a dark brown rim occur in addition to the clinopyroxenes described above. The green core of these clinopyroxene is enriched in FeO and Al2O3 and depleted in MgO. Whereas the core is unzoned, the pale brown mantle shows increasing MgO starting at a much higher MgO content. This evolution is followed by a decrease in MgO and a narrow zone of MgO enrichment when approaching the rim. In this sequence, FeO, Al2O3 and TiO first decrease and then increase through the mantle. Towards the outermost rim, the composition of the clinopyroxene shows a complex evolution of generally increasing Al2O3, FeO and TiO2, starting at lower values than the mantle values, whereas MgO decreases towards the outermost rim (Fig. 2a).
View larger version (71K):
[in this window]
[in a new window]
|
Fig. 2 (a) Representative clinopyroxene compositional profiles from samples HEJ 21 (basanite), HEJ 10 (mugearite), HEJ 12 (nephelinite) and HEJ 1 (tephrite); (b) representative amphibole compositional profiles from samples HEJ 21, HEJ 10 and HEJ 1, and a plagioclase compositional profile from sample HEJ 10.
|
|
The mostly euhedral form of some of the green cores suggests
that they crystallized from a melt. This melt must have been
more differentiated than the host basanite because of the lower
MgO but higher Al
2O
3, FeO and TiO
2 abundances. On the other
hand, the chemical characteristics of the mantles suggest that
they crystallized from a more mafic magma (higher MgO, lower
Al
2O
3, FeO and TiO
2). The outermost rims of the green-core cpx
vary according to normal low-pressure fractionation trends (e.g.
increasing Ti, Al and Fe, and decreasing Mg). These features
are characteristics of clinopyroxene crystallizing from an alkali
basaltic magma (e.g. Duda & Schmincke, 1985
). The development
of the more primitive mantles around the more evolved cores
can be attributed to mixtures of the host basanitic melt with
evolved (?tephritic) melts stored in the mantle, which resulted
from earlier episodes of magma generation and fractionation.
Mixing between evolved and primitive melts, believed to be associated
with magma replenishment (e.g. Huppert & Sparks, 1988
),
can account for the progression from evolved cores towards mafic
rims of the clinopyroxene phenocrysts. Therefore, the crystallization
scenario shown by the different clinopyroxene types suggests
discrete storage zones for the alkali basaltic magmas.
Some basanites and some of the more differentiated rocks contain optically homogeneous, unzoned brown amphibole (kaersutite) phenocrysts with a grain size between 1 mm and 2 cm (Table 2 and Fig. 2b). Plagioclase is common only in some alkali basalts and in the more differentiated rocks and is generally unzoned (Table 2). Rare large plagioclase crystals show the development of K-feldspar-rich rims (Fig. 2b).
|
GEOCHEMISTRY
|
|---|
Chemical composition of the Hocheifel lavas
The mafic magmatic rocks from the Hocheifel area are mostly
relatively primitive basanites, nephelinites and alkali basalts
with rare hawaiites, mugearites and benmoreites, according to
the total-alkali vs SiO
2 classification sheme of Le Bas
et al.
(1986)
(
Table 3;
Fig. 3a). According to the classification proposed
by Huckenholz & Büchel (1986) and Le Bas (1989)
, samples
with >10% normative nepheline, CaO + Na
2O + K
2O < 18 at
SiO
2 + Al
2O
3 < 55 and <41 wt % SiO
2 are classified as
nephelinites. For the nephelinites and basanites, TiO
2, MgO,
CaO and FeO
(total) decrease and K
2O and Al
2O
3 increase with
increasing SiO
2 (
Fig. 3). Na
2O (not shown) shows considerable
scatter among the nephelinites and basanites but generally increases
with increasing SiO
2. For the evolved rocks (hawaiites, mugearites
and benmoreites), TiO
2, MgO, CaO and FeO decrease with increasing
SiO
2, whereas K
2O, Al
2O
3 and Na
2O increase (
Fig. 3). P
2O
5 decreases
from nephelinites to alkali basalts but increases with increasing
SiO
2 in the more differentiated rocks. Basanites and nephelinites
have high CaO/Al
2O
3 ratios and this ratio decreases in the more
differentiated rocks with increasing SiO
2 (
Fig. 3b).
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 3 (a) Total alkali–silica diagram (Le Bas et al., 1986), (b) CaO/Al2O3 vs SiO2, (c) K2O vs SiO2, (d) FeO (total) vs SiO2, (e) TiO2 vs SiO2, (f) CaO vs SiO2, (g) MgO vs SiO2 and (h) Al2O3 vs SiO2 for the Hocheifel lavas.
|
|
Trace element data are reported in
Table 3 and
Figs 4 and
5.
Most nephelinites, basanites and the alkali basalts have Ni,
Cr and Co contents that approach the values commonly assumed
for primary magmas (e.g. Frey
et al., 1978
). Scandium contents
range from 10 to 40 ppm for the nephelinites, basanites and
the alkali basalts (
Fig. 4d). The most interesting feature is
the strong overlap in compatible trace element abundances for
nephelinites, basanites and alkali basalts. Similarly, incompatible
trace element abundances (Nb, Ba, Zr, La) show strong overlap
in nephelinites, basanites and the alkali basalts (
Fig. 5),
although the nephelinites tend to have higher La abundances
than the basanites and alkali basalts. Consequently, ratios
of incompatible trace elements (Zr/Y, La/Nb, Zr/Nb;
Fig. 6)
also show some overlap for nephelinites, basanites and alkali
basalts. On the other hand, ratios of Ba/La, Ba/Nb and K/Nb
tend to increase with increasing SiO
2 from nephelinites to basanites
and alkali basalts (
Fig. 6). Most of the differentiated rocks
have higher Ba/La, Ba/Nb, K/Nb, Rb/Nb and Zr/Nb ratios than
the nephelinites, basanites and alkali basalts, and Ba/La and
Ba/Nb decrease with increasing SiO
2 within this group (
Fig. 6).
The Rb/Nb and K/Nb ratios are positively correlated and the
nephelinites tend to have lower Rb/Nb and K/Nb ratios than the
basanites and alkali basalts. The differentiated rocks have
the highest Rb/Nb and K/Nb ratios. La/Nb ratios are remarkable
constant among the nephelinites, basanites and alkali basalts.
Nephelinites, basanites and alkali basalts have light REE (LREE)-enriched
REE patterns similar to those of many ocean island basalts (OIB)
and alkaline volcanic rocks from continental settings (
Fig. 7).
The differentiated rocks have similar REE patterns; two samples
have a pronounced depletion in middle REE (MREE) (
Fig. 7). Nephelinites,
basanites and alkali basalts show strong enrichment of highly
incompatible and moderately incompatible trace elements (
Fig. 8).
K and Rb are strongly depleted relative to elements with similar
incompatibility (
Fig. 8) whereas in some nephelinites, the basanites
and alkali basalts additional slight depletions of P and Ti
can be observed. Some nephelinites are enriched in P. Apart
from these anomalies, primitive mantle-normalized concentrations
increase with increasing incompatibility and show typical OIB
or intracontinental basalt patterns (
Fig. 8).
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 4 (a) Cr (in ppm) vs SiO2 (wt %), (b) Ni (in ppm) vs SiO2, (c) Co (in ppm) vs SiO2 and (d) Sc (in ppm) vs SiO2 for the Hocheifel lavas.
|
|
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 5 (a) Nb (in ppm) vs SiO2 (wt %), (b) Ba (in ppm) vs SiO2, (c) Zr (in ppm) vs SiO2 and (d) La (in ppm) vs SiO2 for the Hocheifel lavas.
|
|
View larger version (23K):
[in this window]
[in a new window]
|
Fig. 6 (a) Zr/Y vs SiO2, (b) La/Nb vs SiO2, (c) Zr/Nb vs SiO2, (d) Ba/La vs SiO2, (e) Ba/Nb vs SiO2 and (f) K/Nb vs SiO2 for the Hocheifel lavas.
|
|
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 7 Rare earth element abundances of (a) nephelinites, (b) basanites, (c) alkali basalts and (d) more differentiated rocks from the Hocheifel area. Normalization values are from Boynton (1984).
|
|
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 8 Primitive mantle-normalized incompatible element patterns of (a) nephelinites, (b) basanites, (c) alkali basalts and (d) more differentiated rocks from the Hocheifel area. Normalization values are from Sun & McDonough (1989).
|
|
Sr–Nd–Pb isotope chemistry
Sr–Nd–Pb isotope data are reported in
Table 4. The
87Sr/
86Sr ratios of the Hocheifel basalts are low and the
143Nd/
144Nd
ratios are high; thus, the samples plot in the ‘depleted
field’ relative to Bulk Earth in the Sr–Nd isotope
diagram (
Fig. 9). Sr and Nd isotope data for the primitive nephelinites,
basanites and alkali basalts form an elongated trend from more
depleted compositions similar to the European Asthenospheric
Reservoir (EAR; Cebriá & Wilson, 1995
) towards Bulk
Earth values. Generally, this trend is broadly similar to the
trends defined by Cenozoic mafic alkaline rocks from elsewhere
in Germany (Wörner
et al. 1986
; Wedepohl
et al., 1994
;
Hegner
et al., 1995
; Jung & Masberg, 1998
; Jung & Hoernes,
2000
; Bogaard & Wörner 2003
; Haase
et al., 2004
) and
also to other CEVP provinces; for example, the Massif Central,
Poland, and the Pannonian basin (Alibert
et al., 1987
; Blusztajn
& Hart, 1989
; Downes, 1984
; Wilson & Downes, 1991
, 2006
;
Embey-Isztin
et al., 1993
; Harangi, 1994
; Downes
et al., 1995
).
It is noteworthy that the trend defined by the Hocheifel lavas
is similar to that of the neighbouring East Eifel and West Eifel
volcanic fields (Wörner
et al., 1985), which is, however,
displaced to slightly higher
87Sr/
86Sr ratios at a given
143Nd/
144Nd
ratio. The more differentiated samples tend to have more radiogenic
87Sr/
86Sr and less radiogenic
143Nd/
144Nd than the mafic alkaline
lavas and overlap with the compositional fields of Eifel mantle
xenoliths and lower crustal granulite xenoliths (
Fig. 9).
The Pb isotope compositions of the nephelinites, basanites and
alkali basalts overlap and are variable, defining a linear array
subparallel to the Northern Hemisphere Reference Line (NHRL).
This trend ranges from high
206Pb/
204Pb ratios (
20) similar
to the EAR to more unradiogenic values (
19) similar to other
volcanic provinces from the CEVP (
Fig. 10). A subset of basanites
has distinctly lower
207Pb/
204Pb ratios than the other samples
(
Fig. 10). Published Pb isotope data for East Eifel and West
Eifel volcanic fields have slightly higher
207Pb/
204Pb and
208Pb/
204Pb
ratios at a given
206Pb/
204Pb ratio (Wörner
et al., 1986
).
|
DISCUSSION
|
|---|
Fractional crystallization
Most of the nephelinites, basanites and alkali basalts from
the Hocheifel volcanic field have MgO, Cr and Ni contents high
enough for these rocks to represent near-primary magmas (e.g.
Hart & Davies, 1978
). Some samples have lower concentrations
of MgO, Ni and Cr, and for these samples fractionation of olivine
and clinopyroxene is likely. For these samples, decreasing CaO
and increasing Al
2O
3 leds to decreasing CaO/Al
2O
3 ratios with
increasing SiO
2, which is also consistent with clinopyroxene
fractionation. Increasing Al
2O
3 contents and the lack of negative
Eu anomalies indicate that plagioclase was not a major fractionating
mineral phase at this stage, implying that fractionation took
place at pressures >5 kbar, equivalent to depths >15 km
within the lower crust. The more differentiated rocks have the
lowest Ni, Cr and V abundances, indicating that olivine, clinopyroxene
and Fe–Ti oxides were important fractionating mineral
phases in the petrogenesis of the more evolved Hocheifel magmas.
In accordance with previous studies on the evolution of alkaline
magma series, it is suggested that the alkali basalts represent
the parental magmas from which the more differentiated rocks
originated by fractional crystallization (Wilson
et al., 1995).
The three most differentiated samples with SiO
2 > 50 wt %
(HEJ 10, HEJ 11, HEJ 29) have the lowest Sr/Nd ratio and small
negative Eu anomalies, indicating removal of Sr by plagioclase
fractionation (
Fig. 7). Another fractionated sample (HEJ 37;
Fig. 7) shows a deficency of the MREE relative to the LREE and
heavy REE (HREE), which can be explained by significant fractionation
of amphibole and/or titanite. Some nephelinites and basanites,
and most of the alkali basalts, contain green-core clinopyroxenes;
in accordance with previous studies (e.g. Duda & Schmincke,
1985
; Jung & Hoernes, 2000
; Haase
et al., 2004
) the appearence
of green-core clinopyroxenes (with cores enriched in Al, Fe
and Na and with low Ti/Al ratios) is evidence for high-pressure
or, at least, polybaric fractionation of the host magmas. Polybaric
fractionation at deep crustal levels may also be associated
with crustal contamination, which will be evaluated below.
Crustal contamination
Major element and compatible trace element variations in the alkali basalts and the more differentiated rocks of the Hocheifel indicate that fractional crystallization processes affected the magmas during ascent. Fluid-inclusion barometric studies have shown that similar alkaline magmas from the Quaternary Eifel volcanic field stagnated in the lower crust at pressures of about 0·65 GPa, equivalent to 20 km depth (Duda & Schmincke, 1985; Sachs & Hansteen, 2000). In view of the occurrence of petrographically similar, green-core, clinopyroxenes in most of the alkali basalts, basanites and nephelinites (Table 1) from the Hocheifel, it is suggested that these magmas also stagnated in the lower crust at comparable depths.
The lower crust beneath the Eifel is composed of mafic and felsic granulites, in which mafic granulites, interpreted as basaltic cumulates, predominate over felsic granulites (Mengel et al., 1991; Sachs & Hansteen, 2000). The upper crust consists of Palaeozoic sedimentary and volcanic rocks as well as rare Mesozoic and Cenozoic sedimentary rocks. Rare metasedimentary granulites also occur. Some granulites show evidence of metasomatism and partial melting (formation of secondary hydrous phases, presence of glass) and it has been shown that this metasomatic event is most probably related to the Quaternary–Tertiary magmatism (Sachs & Hansteen, 2000). Lower crustal xenoliths from the Eifel have been extensively studied (Stosch & Lugmair, 1984; Stosch et al., 1986, 1992; Loock et al., 1990; Rudnick & Goldstein, 1990) and, therefore, major and trace element and Sr–Nd–Pb isotope data are available. These granulites have Sr–Nd isotope compositions that extend from Bulk Earth values towards more unradiogenic 143Nd/144Nd but more radiogenic 87Sr/86Sr isotope compositions (Fig. 9). Felsic granulites tend to have more radiogenic 87Sr/86Sr isotope compositions, although some mafic granulites are also fairly radiogenic. The nephelinites, basanites and alkali basalts have higher 143Nd/144Nd and lower 87Sr/86Sr than the lower crustal xenoliths, and only a few of the more differentiated rocks overlap with the Sr–Nd isotope composition of the xenoliths (Fig. 9). The Pb isotope compositions of the xenoliths plot above the NHRL in 207Pb/204Pb vs 206Pb/204Pb and 208Pb/204Pb vs 206Pb/204Pb space (Fig. 10). Metasedimentary granulitic xenoliths have higher 87Sr/86Sr and lower 143Nd/144Nd than the mafic and felsic granulites, but similar Pb isotope compositions to them.
The primitive nephelinites, basanites and alkali basalts have a considerable spread in K/Nb that ranges from 52 to 209 despite their limited variation in SiO2 (Fig. 6). For the CEVP as a whole the range in K/Nb ratios has been explained as a result of mixing of partial melts of two different mantle end-members (Wilson & Downes, 1991). On the other hand, the positive correlation between K/Nb and SiO2 (Fig. 6), and the highest K/Nb ratios in the most evolved samples may also reflect crustal contamination processes, as all crustal components (lower crust, bulk crust, upper crust) have high K/Nb ratios (>500; Taylor & McLennan, 1985).
Primitive alkaline volcanic rocks with OIB affinities commonly have low Zr/Nb ratios ranging from 2 to 4 (Weaver, 1991), whereas the continental crust has higher and more variable Zr/Nb ratios ranging from 8 to 14 (Taylor & McLennan, 1985; Rudnick & Fountain, 1995). The higher Zr/Nb ratios in most of the differentiated lavas and the observed correlations of Zr/Nb and Zr, Zr/Nb and K/Nb, Zr/Nb and 87Sr/86Sr, and Zr/Nb and 143Nd/144Nd suggest assimilation of lower crustal rocks with a composition similar to that of lower crustal xenoliths from the Eifel (Fig. 11). Assimilation of lower crustal rocks and fractional crystallization would have occurred simultaneously. However, thermal considerations suggest that bulk assimilation of lower crustal rocks is unlikely and that contamination of the fractionating alkali mafic magma with a partial melt of the lower crustal wall-rocks is more appropriate. The heat required for partial melting is released by the fractional crystallization process. Recent models indicate that this process is an energy-constrained assimilation–fractional crystallization process (EC-AFC; Spera & Bohrson, 2001). The Spera & Bohrson (2001) model was used to test the influence of concurrent crustal assimilation and fractional crystallization upon the composition of the differentiated lavas from the Hocheifel using the parameters given in Table 4. In contrast to the model parameters given by Spera & Bohrson (2001), we used a higher initial temperature for the lower crust of 900°C. This higher temperature is in agreement with recent estimates of lower crustal temperatures from the Eifel (>800°C, Sachs & Hansteen, 2000), and probably mirrors more closely the effects of rifting, uplift of the asthenosphere–lithosphere boundary and continuing magmatism in Tertiary–Quaternary times. Moreover, at this high inferred temperature, high rates of assimilation relative to fractional crystallization are likely (Reiners et al., 1995). In our model, we used sample HEJ 53 as the parental melt; this is one of the most unfractionated alkali basalts based on its moderately high Ni and Cr abundances and low 87Sr/86Sr and high 143Nd/144Nd isotope ratios (Tables 2 and 3). As the assimilant we used sample S 32 (Stosch & Lugmair, 1984; Loock et al., 1990), which is a mafic granulite xenolith with fairly high 87Sr/86Sr and low 143Nd/144Nd. Zr and Nb concentrations are not available for this granulite xenolith and for modelling purposes we used 70 ppm Zr and 6 ppm Nb (average of lower crust; Taylor & McLennan, 1985) for our lower crustal contaminant. It can be seen that the range in Sr and Nd isotope composition, Zr/Nb and K/Nb ratios and Zr concentrations of the differentiated lavas can be reproduced by an EC-AFC model (Fig. 11), implying that energy-constrained assimilation–fractional crystallization processes played an important role in the evolution of the differentiated lavas from the Hocheifel. Based on this model, the trace element and isotope composition of some of the differentiated lavas can be explained by 40–70% fractional crystallization and 10–50% assimilation of a granulite-facies lower crust with a trace element and isotope composition similar to S 32 (Figs 11 and 12). The degrees of assimilation are rather high and probably unrealistic; however, the composition of the lower crust beneath the Rhenish Massif is somewhat unconstrained with respect to its trace element and isotope composition and more suitable end-members may exist.
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 11 (a) Zr/Nb vs Zr, (b) Zr/Nb vs K/Nb, (c) Zr/Nb vs 143Nd/144Nd and (d) Zr/Nb vs 87Sr/86Sr for mafic alkaline lavas from the Hocheifel. Lines show the results of EC-AFC calculations with model parameters and end-member compositions from Table 4. Grey lines with regular-font numbers denote mass crystallized and black lines with italic numbers denote mass assimilated, both as wt %.
|
|
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 12 (a) 87Sr/86Sr vs 143Nd/144Nd, (b) 87Sr/86Sr vs 206Pb/204Pb and (c) 143Nd/144Nd vs 206Pb/204Pb for mafic alkaline lavas from the Hocheifel. Lines show the results of EC-AFC calculations with model parameters and end-member compositions from Table 4. Grey lines with regular-font numbers denote mass crystallized and black lines with italic numbers denote mass assimilated both as wt %.
|
|
Magma generation and partial melting processes
It is generally accepted that the upper mantle is composed predominantly
of peridotite with minor amounts of pyroxenite, both with or
without garnet. Primitive alkaline, silica-undersaturated melts
such as nephelinites and basanites can form at high pressure
from garnet peridotite sources (Kushiro, 1996
), but not from
garnet pyroxenite (Rapp
et al., 1991
). Therefore, the most likely
magma source for the primitive nephelinites and basanites from
the Hocheifel is a garnet peridotite. The strong fractionation
of HREE, with Dy/Yb >2 (
Fig. 13), further suggests that the
primitive lavas from the Hocheifel represent partial melts of
garnet peridotite. Seismic models indicate that the lithosphere–asthenosphere
boundary beneath the Rhenish Massif is strongly elevated and
is located at
50–60 km (Babuska & Plomerová,
1992; Goes
et al., 2000), whereas the transition from garnet
to spinel peridotite is estimated at 2·5–3·0
GPa, equivalent to 75–90 km depth (McKenzie & Bickle,
1988
; Robinson & Wood, 1998
). Previous xenolith-based studies
on the composition of the upper mantle beneath the Rhenish Massif
indicate that the upper mantle consists of metasomatized spinel
peridotite with amphibole and phlogopite (Witt-Eickschen &
Kramm, 1998
; Witt-Eickschen
et al., 1998
, 2003
).These xenoliths
are interpreted to represent fragments of the lithospheric mantle,
and it is reasonable to assume that partial melting must have
occurred at depths in excess of 60 km (i.e. below the base of
the lithosphere).
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 13 La/Yb vs Dy/Yb covariation for the Hocheifel basalts. Partial melting curves were calculated using a non-modal, fractional melting model (Shaw, 1970). Sources are grt–amph peridotite (cpx 0·07, opx 0·19, ol 0·55, grt 0·08, amph 0·11), which melts in the proportions cpx 0·25, opx 0·15, ol 0·05, grt 0·3, amph 0·25, and sp–amph peridotite (cpx 0·08, opx 0·25, ol 0·554, sp 0·033, amph 0·083), which melts in the proportions cpx 0·27, opx 0·25, ol 0·08, sp 0·13, amph 0·27. Source composition (La 2·1 ppm, Yb 0·17 ppm, Dy 0·31 ppm) represents average of 36 peridotite xenoliths from the Hessian Depression (Hartmann & Wedepohl, 1990). Mineral–melt distribution coefficients are taken from McKenzie & O'Nions (1991), Hart & Dunn (1993), Kelemen et al. (1993), Johnston (1994) and LaTourette et al. (1995). Numbers on model curves indicate the per cent melting. Points at 80grt/20sp, 60grt/40sp and 40grt/60sp indicate mixing proportions of melts from garnet peridotite (i.e. 80%) with melts from spinel peridotite (i.e. 20%).
|
|
A useful approach to model partial melting of common upper mantle
sources is based on REE systematics (e.g. a plot of La/Yb vs
Dy/Yb,
Fig. 13; Thirlwall
et al., 1994
; Baker
et al., 1997
);
such plots can easily distinguish between melting in the garnet
peridotite stability field and melting in the spinel peridotite
stability field because of the strong fractionation of HREE
by garnet. Additionally, mixing of melts from garnet and spinel
peridotite sources should produce linear arrays in such a diagram.
The nephelinites and most basanites and alkali basalts appear
to form a coherent group with higher La/Yb ratios than the rest
of the basanites and one alkali basalt, which form another group
with higher Dy/Yb ratios (
Fig. 13). Considering each group individually,
it appears that the samples plot on mixing lines between melts
from garnet peridotite and melts from spinel peridotite. Moreover,
this diagram suggests that simple partial melting exclusively
in the garnet peridotite stability field or spinel peridotite
stability field cannot account for the spread of data. Partial
melting of spinel peridotite should also produce a positive
correlation between Ce/Yb and Yb abundances. Nephelinites, basanites
and alkali basalts appear to show a series of negative correlations
between Ce/Yb and Yb abundances (
Fig. 14a). Additionally, nephelinites,
basanites and alkali basalts show a positive correlation between
Ce/Yb and Ce abundance (
Fig. 14b). These features indicate that:
(1) nephelinites, basanites and alkali basalts originate from
sources with similar LREE enrichment; (2) nephelinites represent
smaller melt fractions than most basanites and alkali basalts;
(3) at least the nephelinites originate from a garnet-bearing
source in which, during partial melting, garnet was progressively
eliminated from the source. High CaO/Al
2O
3 ratios in the nephelinites
and decreasing CaO/Al
2O
3 with increasing SiO
2 in the sequence
nephelinite–basanite–alkali basalt are also consistent
with increasing degrees of partial melting of a garnet-bearing
source. Therefore, the most plausible model that can account
for the REE variation involves initial partial melting in the
garnet stability field, followed by mixing of melts from garnet
peridotite with melts from spinel peridotite, both containing
amphibole. Most of these samples cluster at Dy/Yb ratios between
2 and 3; this is typical for the upper mantle beneath the CEVP;
the mantle xenoliths have flat to slightly LREE-enriched REE
patterns (Stosch & Lugmair, 1986
; Witt-Eickschen & Kramm,
1998
).
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 14 (a) Chondrite-normalized Ce/Yb vs Yb and (b) chondrite-normalized Ce/Yb(n) vs Ce for mafic alkaline lavas from the Hocheifel. Normalization values are from Boynton (1984).
|
|
It is noteworthy that the nephelinites, basanites and alkali
basalts have a large range in K/La ratios and that the nephelinites
have the highest Ce/Yb
(norm.) and La concentrations and the
lowest K/La ratios (
Fig. 15). The model curves shown in
Fig. 15 imply that the primitive Hocheifel lavas could result from 1–2%
partial melting in the garnet peridotite stability field, compatible
with experimental results that indicate melting degrees in excess
of 1% to generate basanites from peridotite sources (Kushiro,
1996
). Similar low degrees of melting have been inferred from
a number of volcanic provinces of the CEVP (Wilson & Downes,
2006
). In the Hocheifel, small-degree partial melts from garnet-bearing
peridotite were mixed with melts produced by a similar degree
of melting of a spinel peridotite source (
Fig. 13), suggesting
that partial melting and mixing of melts occurred close to the
spinel–garnet transition zone. The transition from garnet
peridotite to spinel peridotite occurs at 2·5–2·7
GPa (Robinson & Wood, 1998
), indicating that the Hocheifel
lavas formed at
80 km depth. The nephelinites, basanites and
alkali basalts display a negative correlation between K/La and
Ce/Yb
(norm.) (
Fig. 15a) implying that partial melting of garnet
or spinel peridotite alone is not likely because low-melt fractions
from such sources have both high Ce/Yb
(norm.) and K/La ratios
(Haase
et al., 2004
). It is, therefore, very likely that a residual
mineral phase that fractionates K from La was present. This
mineral phase was probably amphibole rather than phlogopite,
because phlogopite fractionates K/La even more efficiently than
amphibole and also fractionates Ba/La. Additionally, Ba concentrations
are high in the primitive Hocheifel lavas, which argues against
significant amounts of phlogopite in the melt residue. Because
the nephelinites with high Ce/Yb
(norm.) and low K/La ratios
formed in the stability field of garnet peridotite in the presence
of amphibole, it is suggested that the melting region is located
at 2·5–3·0 GPa and 1250–1300°C
(
Fig. 16). The amphibole may have formed by mantle metasomatism
caused by migration of small-degree melts from an upwelling
plume beneath the Rhenish Massif.
It is still a controversial issue whether the volcanism of the
CEVP is related to adiabatic decompression melting caused by
thinning of the lithosphere during rifting or due to raised
mantle temperatures (up to 200°C, Ritter
et al., 2001
) as
a consequence of mantle plume activity. Data from experimental
investigations can be used to constrain the conditions of formation
of the Hocheifel lavas. Basanites, and even more Si-undersaturated
melts such as nephelinites, can be generated by melting of amphibole-
an
TOP
ABSTRACT
INTRODUCTION
). Inset shows the location of the Hocheifel volcanic field relative to the east (Osteifel) and west Eifel volcanic fields.
, experimental results of phlogopite–garnet peridotite melting from Mengel & Green (1986)