Miocene sepiolite deposit near Calatayud accumu- lated by chemical precipitation in a playa-lake environment, and Inglds & Anad6n (1991) described Eocene ...
Clay Minerals (1996)31, 3344
MINERALOGICAL AND GEOCHEMICAL C H A R A C T E R I Z A T I O N OF P A L Y G O R S K I T E F R O M G A B A S A ( N E S P A I N ) . E V I D E N C E OF A D E T R I T A L P R E C U R S O R
A. LOPEZ-GALINDO,
A. BEN ABOUD, P. F E N O L L J. C A S A S R U I Z *
HACH-ALI
AND
lrLvtituto A ndaluz de Ciencias de la Tie rra. CS1C-Univ. Granada, A vda. Fuentenueva, s/n. 18002 Granada, Spain, and *TOLSA SA. Crta. Vallecas-Mejorada del Campo. 28080 Madrid, Spain (Received 15 February 1995; revised 30 June 1995)
A B S T R A C T : A mineralogical and geochemical study of the Gabasa outcrop (Huesca, NE Spain) was undertaken. It consists of Early Oligocene marly and clayey fluvial and lacustrine (playa-lake) sediments. The phases detected were quartz, amorphous silica, calcite, dolomite, palygorskite, illite, interstratified illite-smectite, Al-smectite and Mg-smectite. The palygorskite expands with ethyleneglycol. Statistical analysis of the geochemical data shows that the rare earth elements and transition trace elements are basically associated with the detrital phyllosilicates, although a considerable amount of the latter is contained in the palygorskite (Y~REE = 60-70 ppm, Cr+Co+Ni+V+Zn+Cu = 120-150 ppm), in contrast to the normally low values for neoformed minerals. This fact, together with the significant presence of AI and Fe in the palygorskite, suggest genesis involving alteration by dissolution of the 2:1 structure of the illite and/or Al-smectite, followed by re-ordering in a fibrous structure.
In the framework of a research project covering the geochemical and micromorphological study of fibrous clay deposits in Spain, the authors present here the first data on a new palygorskite deposit located near the city of Huesca. Our samples were taken from the continental Tertiary sediments of the Ebro Depression (Fig. 1), consisting of thick sequences of lacustrine and fluvial sediments with assorted lithology and frequent lateral facies changes. They are located between the Barbastro and Peraltilla Formations (Riba et al., 1980) and consist of limestones, calcareous conglomerates, sandstones and marls. The Early Oligocene sediments form a belt striking NW-SE, from Zurita to Calasanz and running through Gabasa. The exposure is ~ 2 2 m thick in the sampling zone (Fig. 2). It begins with a bed of white limestone, followed by 2 m of marly clays and green clays, 2.5 m of reddish marls with more lighter-coloured carbonate intercalations, 1.5 m of white and beige hollow carbonates and 15 m of green, red and white clays and marls containing the richest palygorskite levels. The sequence ends with
a bed of limestones containing calcareous conglomerates at its base. The exposure reaches over 35 m in the vicinity of Zurita and contains intercalated beds of saccharoid gypsum and carbonate crests. The presence of palygorskite in Tertiary Spanish deposits has been broadly reported (i.e. MartfnPozas et al., 1981, and Gal~in & Castillo, 1984, in the Tajo Basin; Gal~in & Ferrero, 1982, in the Guadalquivir Basin; Su4rez et al., 1989, in the Duero Basin; Gonz41ez-L6pez et al., 1994, in the Almaz~in Basin). Previous occurrences of fibrous clay minerals in the Ebro Basin have been cited by Huertas et al. (1974) and by Gonzzilez & Galfin (1984). In recent years, Arauzo et al. (1989, 1991) recognized a Miocene sepiolite deposit near Calatayud accumulated by chemical precipitation in a playa-lake environment, and Inglds & Anad6n (1991) described Eocene sequences in the NE margin of the Ebro basin, near lgualada, which contain a considerable amount of palygorskite in the sediments from playa-lakes, mudflats and poorlydrained flood plains.
9 1996 The Mineralogical Society
34
A. L6pez-Galindo et al.
GABASA 9 OUTCR~
BLower
Miocene
~ J
Oligocene
?
Eocene
FIG. 1. Geological sketch of the Tertiary materials in the Ebro Depression (from Riba et al., 1980) and location of the Gabasa outcrop.
BULK
CLAY
MINERALOGY
50% C.~15I'~ ~:l
100% r,f J f f f / / / J f ,
GA-I3I'~ \ I
MINERALOGY
i
(////////////A
GA-12r'~ ~: 9 9 9 9 9 9 9 9. ' f / / / / / / / / / / / / / , f J
3dH[IIIIIIIIIBIIIIIIIIIIIIIIIIIIIIIIIIIIfllIHHIIflIIHHWIHIIIIflUI'~ \
.'.'.'-'.'-'-'.'-'."
.:/H///.,/,~
:'"'-'-'-'-'-'.'.'.'-'-'-'.'.'-'.'.'.'.'.'.-#',f,#,#i GA.7 IX,,. I : - ' , ' , ' , ' - " f / / / / / / / / / / / / / / / / / J C,k6 I'~ I Ir/'////////////r/r////////]
flHIlllllllllllllUIIIllillllllllllllltllllllr\
:I/////////A
GA-2I~ I ~ 1 PI
[/////////.,Jl l
Quartz
f//////////.~
Amorp. silica
Calcite
gll]lllillllllllllllllillflllgllllr~
~
HIIIIllllIItlIIIlIIIIT~ [-~ f
\
J
Dolomite Phyllosil.
I
/
i jr
/
/
/
/
f
I
J
/
.f
f
/
illOlNllllllf |l]llllflfl~ f f gllllllllllllllU~ ~ gllllllllllllllll~ ] /
1
I
.#.-|
~
~
I
J
i
Jl
i
/
f
f
/
/
/
f/1
I /
/ /
/
i /
,Or /
i
i
/
Illite
Inter. I-S
/
/
AI-Smec
J
i ]
.,./I
/
I/ /
I
l l l J l i J
i
~illl[ll]llllllll[lllllllllllllllllllllllll~ \ | gl[flllllllllllllKHllllHllflllllfl~ ~ f
&
.f
d e" 1 /
IIHHr'4IIIIllilIliJ glllilllllhH /
:-'-'-'-'.'-'.'.'.'.'.'-'.'.'.'.'.'.'.'.'. ".'.."E/,~I GA-4 I ~ : ' . ' . ' . ' . ' . ' . ' . ' , ' , ' , ' , ' , ' , ' GA-3IX:. ". ". ". ". ". '. '. '. '. ". ". '. ". ". - " : / / / / / / / . / / J
I
gllllllilllllHHHIlilllllllgllllllllllllHgflBUflllHIr\ "~ "4 i ~1 OllllllilllilllllilgllllllilflHIIliflflililflllr~ \ I ~ ] f / f lillllll]lllilllllllllIIIl[lllllllll~lP,,. " ~
GA-10r~ : - ' . ' - - - -
\1
]
/
I
,/" 1 /
Paly
FIG. 2. Lithology, bulk and clay mineralogy of the samples studied in the Gabasa deposit.
/
1 f
f
Mg-Smec
Characterization of palygorskite from Spain MATERIALS
AND
METHODS
In all, 15 clayey, marly and marly-limestone samples were studied. Figure 2 shows their position in the lithological column. The samples were ground to 53 lam mesh and decarbonated for the diffractometric study of the fine fraction. Total and clay mineralogy were determined using a Philips PW1770 diffractometer, with automatic slit, Cu-Ket radiation and 2~ scanning rate from 2 - 6 0 ~ 20. The oriented aggregates were treated with ethylene glycol and dimethyl sulphoxide and heated to 550~ Quantitative analysis was carried out by computer on the basis of the chemical analyses of major elements and diffraction data. The values obtained by the traditional method of measuring peak areas and reflecting power (cf. Schultz, 1964; Barahona, 1974) were considerably improved by taking into account: (a) the chemical composition of the monomineralic samples and data taken from EDX analysis (the latter allowing us to distinguish between A1- and Mgsmectites); (b) assignation of K to illite, interstratified illite-smectite and K-feldspar; and (c) assignation of Ca to the carbonates and distribution between calcite and dolomite, according to the intensity relation between the 3.03 ,~ and 2.88 ,~ peaks (L6pezGalindo et al., 1994). The chemical analyses of major elements, except Si, were carried out using ICP Thermo Jarrel AshPlasma 300 equipment. Silica was determined by atomic absorption using IL-257 equipment. Trace
35
and rare earth elements were measured in an ICP-MS Perkin-Elmer SCIEX Elan-5000 equipment. The detection limit for all the trace elements was 0.8). (3) A1203 correlates strongly with Fe203 and TiO2 (r = 0.98), Na20 and K20 ( r = 0.8) and quartz ( r = 0.82). Likewise, these major elements correlate with the two previously mentioned groups of trace elements Variables Quartz
F1
F2
Pb
0.608 -0.449 -0.283 0.327 0.880 -0.508 0.683 -0.708 0.066 0.527 0.469 0.034 -0.189 0.364 0.884 0.382 0.064 0.829 -0.301 0.948 0.416 0.498 0.840 0.876 0.739 0.635 -0.246 0.115
0.567 0.396 -0.479 -0309 0261 0.769 0.175 0.327 0.905 0.761 0.755 0.140 -0.726 0.508 0.298 0853 0.391 0.524 -0.563 0.043 0.683 0.820 0.394 0221 0.628 0.705 -0.609 0.750
% Expi. v a r i a n c e
44.5
34.1
Arnorph silica Calcite Dolomite
I + I-S Palygorsk AI S m e c
Mg Smec SiO2 A]203 Fa2Oa
MgO CaO Na20 K20 TiO2 Mn2Oa TRTE
Sr Ba Cu Li REE
Y SUM Ga Ta
(PCA) allows us to differentiate three factors (F1, F2, F3) which together explain 87% of the whole variance (Fig. 5). Factor 1 (44.5%) separates the clearly detrital minerals and associated chemical elements from those components of presumably different origin, i.e. palygorskite, amorphous silica, Mg-smectite, calcite, Sr and Ta. Factor 2 (34%) distinguishes between silicates and carbonates, and factor 3 (9%) separates calcite from dolomite. Rare Earth Elements. Figure 6 shows the distribution of REE normalized to NASC. In general, all the samples presented a horizontal curve, with slight impoverishment in heavy REE (1.37 < LaJLu < 1.76). The REE content varies from 0.3 to 0.9 times the NASC content. There are no significant anomalies, except for the samples richest in palygorskite, that have a positive Ce anomaly (GA-6) and a slight negative Eu anomaly (GA-6 and GA-7). These same samples also have lower Tb/Yb ratios and higher A1203/Yb ratios. In order to determine the REE distribution in the phyllosilicates, particularly in detrital clays and palygorskite, and given the good correlation found, we established straight-line regressions based on: (1) quartz accumulates practically no REE; (2) calcite and dolomite contain hardly any REE
F3 0.194 0.598 -0.808 0.840 0.268 0156 -0.183 -0.338 0.386 0.356 0.363 0.925 -0.649 0.569 0.236 0.301 0.599 0.156 -0.439 -0.040 -0.187 -0.022 0.220 0.181 0.235 Q226 -0.361 0.027
8.6
1
SiO2
Fe 03 T~O2
Palygolskffe
Z
9C~w w AI203 9
9
Mn203 9
Arn0rph silioa
M~Smec
9
QUalIz 9
Na20
0.5
9
FRYE
~EE K20
,I. 9 . . . .
~m~
0 LI,
Oo,;m.~ ca~ite
-0.5
S r 9 OTa CaO
-1
-"
-0.5
0
0.5
Factor 1 REE = Rare earth elements
SUM = Rb+Zr+Nb+Cs+Hf+Th+U+Be+Sc+Sn+Mo T R T E = V+Cr+Ni+Co+Zn
FIG. 5. Factor analysis of principal components of main mineralogical and chemical variables.
17-
......
Characterization of palygorskite from Spain
39
10 GA-I E2]
GA-3 - JIF'-
-
GA-5
GA--6 -.4P--
GA-7 C~
E r
Palygorskite-rich samples
GA--9
0.1 GA-I 1
m
Calcite-rich sample
GA- 12 ~r
GA-I 3 0.01
La
I Ce
I Pr
I Nd
I Sm
t Eu
I Gd
I Tb
I Dy
I Ho
I Er
I Tm
t Yb
t Lu
FIG. 6. NASC-normalized REE patterns of the analysed samples.
(EREE form ~ 10 ppm in calcite and 20 ppm in dolomite, cf. Ronov et al., 1974, and sample GA-5), and (3) R E E p a l y g o r s k i t e = REErock - - R E E d e t r i t a l phyllosilicates
--
REEcarbonates
Figure 7 shows this calculation. So, for a sample containing carbonates, detrital phyllosilicates and palygorskite, and a ratio 25:75 among the two last
phases taken as an example, the quantity of REE contained in palygorskite (C) is equal to the REE of the sample (D) minus the REE provided by detrital phyllosilicates (B) minus REE of the carbonates (A). We can, therefore, infer that, in ideal conditions for pure palygorskite, the amount of R E E in this phase would be the intersection of the 'REE in
t W
o
t.tea
FIG. 7. Theoretical calculation of the REE content in palygorskite. A: REE in carbonates; B: REE in detrital phyllosilicates; C: REE in palygorskite; D: REE in sample.
40
A. L6pez-Galindo et al.
sample' line with the abscissae axis. As this seems to be true for every rare earth given, the very good correlation observed between them, the theoretical content (in ppm) of each element is: La: 11.14; Ce: 28.53; Pr: 2.67; Nd: 9.89; Sm: 1.93; Eu: 0.33; Gd: 2.41; Tb: 0.27; Dy: 1.34; Ho: 0.27; Er: 0.79; Tm: 0.11; Yb: 0.67; Lu: 0.10. When these values are normalized to NASC, we obtain a curve very similar to that of our GA-6 sample but below it and presenting a positive Ce anomaly and a negative Eu anomaly, and a slightly higher impoverishment in HREE. Other trace elements. The argument mentioned in the preceding paragraph is also valid for the trace elements strongly correlated with the detrital phyllosilicates. Thus, trace element contents in the Gabasa palygorskite would be (in ppm): Ni: 9.2; Co: 2.2; V: 17.6; Cr: 29; Zn : 58.7; Rb: 41.8; Zr: 42.5; Th: 5; Mo: 0.04; Sn: 1.9. Microanalysis (TEM). We analysed a large number of particles from the palygorskite-rich samples. The mean structural formula, on the basis of 20 oxygens, is: (Si7.79Alo.21)O2o(All.64Feo.42Mgl.9o)(OH)2(HzO)4"4(HzO)(Cao.o3Ko. 12) The variation observed in the different cation proportions is: Si: 7.54-7.91, total AI: 1.68-2.10, Fe: 0.34-0.49, Mg: 1.70-2.04, K: 0-0.23 and Ca: 0-0.07. Electron microscopy also reveals the presence of A1- and Mg-smectites, although they normally do not coexist in the same sample. Their mean structural formulae, on the basis of 10 oxygens, are: Mg-smectites: (Si3.95Alo.os)Olo(Alo.xoFeo.oaMg2.72)(OH)z(Cao.05Ko.o7) Al-smectites: (Si3.71Alo.z9)Olo(All.3zFeo.35Mgo.31)(OH)z(Cao.13Ko.39) Stable isotopes. Samples rich in calcite (GA-5, GA-8 and GA-10) and dolomite (GA-1, GA-2 and GA-15) were analysed. (see Table 2). Calcite presents ~513C(pi~m of - 7 . 2 3 to -7.45 %o and ~180(pD m of - 4 . 5 2 to - 4 . 8 6 %0. The values for dolomite oscillate between 813C(pDm = --4.95 and -5.29%0, and 8180(pDm = - 0 . 2 5 and -0.44%0. DISCUSSION
AND
CONCLUSIONS
The literature on palygorskite genesis in continental deposits is extensive (cf. Velde, 1985; Jones & Galen, 1988 and references therein). It appears mainly in association with soils, calcretes and
alluvial deposits (Singer, 1984) and lacustrine deposits (Galen & Castillo, 1984). Arid or semiarid climates, alkaline conditions and high Si and Mg activity and low A1 activity are invariably necessary for formation of palygorskite. Unlike sepiolite, which normally forms by chemical precipitation from natural surface water, the most usual mechanism of formation for palygorskite seems to be the dissolution of a previous silicate phase, particularly smectites, followed by palygorskite precipitation. Numerous examples of this mechanism can be found in Spanish deposits. Thus, by studying the geochemical characteristics of groundwater in the southwestern part of the Duero Basin, Sgnchez San Roman & Blanco (1986) showed that the Si and Mg concentrations decrease from recharge areas to discharge areas. They also observed automorphic palygorskite crystals in sediment pores, concluding that the origin was linked with fluid circulation and late diagenetic alteration of smectites. In a study of palygorskite developed in palaeosoils located on terraces along the Tagus River, Martfn de Vidales et al. (1987) considered that a TEM study of the fine fraction proved the origin of this phase by alteration of the structure of pre-existing montmorillonite. This conclusion was reached as they had observed small-sized (100-1000 A) globules intimately associated with smectite relics whose polymerization with Si and Mg in solution would have led to palygorskite formation. For lacustrine-palustrine sediments in the Madrid basin, Martfn de Vidales et al. (1988) concluded that palygorskite formation at the expense of dioctahedral smectite was the result of carbonatation processes involving calcrete formation, as well as diagenetic dolomitization processes. Using analytical electron microscopy, Su~rez et aI. (1994) found evidence of a precursor in the Bercimuel (Segovia) deposit. They suggested mechanisms of dissolution of detrital mica and opening of the structure, with K and AI loss and Si and Mg gain, so that, by way of an intermediate phase of disordered interstratified itlite-smectite, palygorskite finally formed. On the other hand, the stability of fibrous clay minerals in soils was treated in depth by Velde (1985). This author mentioned examples where palygorskite appears in caliches (carbonate soils in arid climates). Genesis is closely related to carbonate precipitation by evaporation of subsurface water rising by capillary action under highly
Characterization of palygorskite from Spain
evaporitic conditions. During the initial stages, the water would dissolve salts and silicates, mainly detrital aluminosilicates, characteristic of most sediments and soils. Velde concluded that it was highly probable that palygorskite was the result of equilibrium between the solution and the silicates, and that it recrystallized from a layer structure to a fibrous structure. The samples of Gabasa studied here are from lacustrine and fluvial deposits (playa lake) accumulated during flooding episodes. The particle size is very small, calcite and/or dolomite are abundant, the pH is alkaline and salinity is high. In this environment, the original detrital sediments, basically aluminosilicates and quartz, would have undergone post-depositional weathering processes that altered the structure of the phyllosilicates, as seems to be shown by the geochemical characteristics of the trace elements and REE. This fact is in accordance with the observations by Inglts & Anad6n (1991) in Eocene outcrops located to the NE of the Ebro Basin. Based on SEM and TEM studies, these authors suggested that in mudflats and flood plains, palygorskite is formed mainly by transformation of precursor clays in a Mg-rich environment, although it may be a neoformed phase in the playa lake and sabkha facies. In another paper on Spanish deposits of sepiolite and palygorskite, Torres-Rufz et al. (1994) showed that the REE, transition trace elements and fluorine contents differed considerably between the detrital phyllosilicates and the neoformed phases, taking the latter to be sepiolite, stevensite and carbonates. Thus, the amount of REE in the detrital minerals was approximately 250 ppm, whereas in the neoformed phases it was
