|
|||||||||
|
|||||||||
This publication was processed (scanned, OCR'ed and manually corrected) by Andrew Wong. It has not yet been proofread. The original is in Solid State Ionics, 25 (1987), 295-300. SODIUM
ION CONDUCTION IN SINGLE CRYSTAL VERMICULITE M. Stanley WHITTINGHAM Schlumberger-Doll
Research, Old Quarry Road, Ridgefield, CT 06877-4108, USA Received
23 September 1987; accepted for publication 16 October 1987 The diffusion
of sodium in single crystals of Llano vermiculite has been studied using
conductivity measurements. The crystals were maintained in the bilayer
water state by using aqueous contacts. The diffusion coefficient at
25°C
was found to be 1.9 X 108 cm2/s, with an
enthalpy of motion of 11.7 kcals/mole. These values are in good agreement
with sodium tracer and proton NMR studies and indicate that the sodium
ions probably diffuse with their hydration spheres. The conductivity
in the range 10 to 90°C
is less than that in sodium beta alumina and much less than that of
either surface clay cations or of aqueous sodium chloride solutions. 1.
Introduction A number of studies
have been made of the diffusion of cations and water in aluminosilicate
materials [1-5], yet surprisingly there have been noreports of the conductivity
in single crystal samples.These layered clay/mica compounds have the
ideal composition Al3(Si4O10)(OH)2,
but substitution in the tetrahedral (Si) and octahedral (Al) sites lead
to incorporation of cations in the Van der Waals layer to maintain charge
balance. These cations are frequently hydrated, the degree of hydration
being a function of the water partial pressure as well as of the cation
and the charge density in the layers. Themontmorillonite class of clays
have the lowest charge density (i.e. lowest cation content) and are
found only as poorly crystalline fine powders; they swell continuously
as the salinity of their environment decreases [6]. At the other extreme
are the micas which have the compositions M+(Si3AlO10)(Mg3)(OH)2
and M2+(Si2Al2O10)(Mg3)(OH)2;
their very high charge densities preclude the incorporation of water
or other solvents except in a few special cases. In between reside the
vermiculite class of aluminosilicates with the idealized formula M+(Si3AlO10)(Mg3)(0H)2,
yH20 where y can vary from 0 to above 5; in
practice the M content is somewhat less and the aluminum content somewhat
greater than unity. A schematic of the three phases Fig.
1. Schematic of aluminosilicate structure. of
sodium vermiculite is shown in fig. 1. The vermiculites and micas are
readily available from natural sources as single crystals so are ideal
for studies of transport properties like conductivity. Thus in these
cases the external surface area is minimal in contrast to the montmorillonites
where in the presence of water the surface area can be in excess of
100 m2/gm. We therefore do not anticipate any problems with
surface conduction by native or foreign species such as protons. In
addition, there are no grain boundaries to hinder or enhance the transport
of cations. Whereas, Calvet and Mamy [1] suggested that the major charge
carriers in sodium montmorillonite Fig.
2. Arrangement of (Si,Al)04 tetrahedra showing the ditrigonal
cavities and approximate size of a hydrated ion. are
proton species, possibly due to the Bronsted acidity of the surface,
one might expect that in single crystals the dominant charge carrier
will in most cases be the native cation. The aluminosilicate tetrahedral
building blocks facing the interlayer spaces are arranged in such a
way that there are large ditrigonal cavities every 5.3 Ǻ, as indicated
in fig. 2. In the micas the tetrahedra are distorted so that e.g. a
potassium ion has six nearest neighbor oxygen atoms at 2.8 Ǻ and
six oxygens 0.5 Ǻ further away. This bonding and burying of the
cation in the structure precludes any significant ionic mobility in
the mica class of material. However, in the vermiculites with their
lower charge density cations such as sodium are readily hydrated separating
the layers by 5 Ǻ and thus permitting some mobility. Here we report
on a study of the most hydrated phase of a sodium vermiculite where
one might expect the highest ionic mobility. 2.
Experimental The vermiculite single crystals used
here were from Llano, Texas and were obtained from the Clays Minerals
Society. They were immersed in aqueous so dium chloride solution at
600C for two months to ensure complete exchange to the sodium form.
They have recently been extensively analyzed [7], and shown to have
the formula: Na0.93(Si2.86Al1.14)(Mg2.94Al0.05)O10(OH)2,yH2O (1) Fig.
3. Water isotherm for sodium vermiculite, after ref. [8]. where
y is a function of the water vapor pressure, and for this material
varies from 0 to »4.6 in a stepwise function as shown
in fig. 3. This isotherm was obtained using a Cahn microbalance at 28°C [8]. The number of water molecules
per sodium ion are 2.0 and 5.0 for the monolayer and bilayer hydrate
respectively. All the measurements reported here were carried out with
the crystals in contact with aqueous solutions where the bilayer water
structure is the stable phase with y around 4.6 and an aluminosilicate
basal spacing of 14.89 Ǻ (Philips diffractometer using Cu Ka radiation). The sodium ion
transference number and conductivity were measured in a simple cell
in which the single crystal of vermiculite was glued into a glass tube
using a silicone rubber cement (General Electric RTV 106). This tube
was then filled with a sodium chloride aqueous solution and placed in
another tube also containing the aqueous solution. By this means the
crystal could be maintained in the bilayer water state throughout the
temperature range of study, 5-90°C; even in saturated air, water loss
would be observed by 35°C. In addition a sodium chloride
concentration gradient could be established across the crystal, allowing
a determination of the transference number to be made. Electrical contact
to the solutions was made by Ag/AgCl electrodes whose surface area exceeded
that of the sample by more than a factor of a 100. This minimized potential
polarization problems at the electrodes, and in fact the conductivities
measured varied by less than 5-10% from 10 Hz to 100 kHz. The conductivities
were measured using a General Radio 1616 Bridge in a GR 1621 Precision
Capacitance System, generally with 0.02 molar solutions at a frequency
of 1 kHz with frequent scans over the above range. The EMF's of the
concentration cells were measured using a Hewlett Packard 3468A multimeter. 3.
Results and discussion The EMF of the concentration cell: Ag/AgCl/NaCl(a1)/membrane/NaCl(a2)/AgCl/Ag (2) is
given at 25°C by the expression: E=(2t+-1)59log(a1/a2)+59log(a1/a2),
(3) where
the first term is the membrane potential with t+ the
transference number of the membrane or liquid junction, and the second
term the electrode potential difference. For a pure ionic conductor
t+ is unity and for a simple liquid junction t+
is simply the transference number of Na+ in sodium chloride
which is around 0.38. The results for sodium vermiculite are shown in
fig. 4. All the data points fall on the unity transference number line
and clearly indicate that this material is a pure cationic conductor.
The values fell sharply from this line when any leakage occurred across
the cell, e.g. from minor cracks in the seal or in the crystal. Thus
the cell EMF was monitored before and after conductivity measurements
to ensure that the cell still had a tight seal. The liquid junction
potential line was measured using a clay free Fontainbleau sandstone,
with a porosity of 20%, as the membrane. The conductivity
of sodium vermiculite is shown in fig. 5 as a function of temperature.
At 25°C the conductivity is 3.0X10-4
S/cm, and in the expression: sT=s0exp(-H/RT) (4) s0 and H have the values 3.15X107
S.K/cm and 11.7 kcals/mole respectively. The linear behavior of the
conductivity over the temperature range studied indicates that there
is probably a single conductivity Fig.
4. EMF of NaCl concentration cell containing sodium ver- miculite as
the membrane. The reference solution is one molar NaCl. mechanism with an enthalpy of motion
of 11.7 kcals/mole (a plot of logs versus 1/T gives an activation
enthalpy of 11.1 kcals/mole). This enthalpy of motion is essentially
the same as that measured by sodium tracer diffusion, 11±1 kcals/mole [2], in World vermiculite
(of different source and composition). This is a strong indication that
the mobile cation is a sodium rather than protonic specie. The sodium ion diffusion coefficient
can be calculated from the Nernst-Einstein relation by making Fig.
5. Ionic conductivity of sodium vermiculite. two
assumptions. These are that the concentration of charge carriers equals
the concentration of sodium ions, 4.2 mmoles/cc, and that the correlation
coefficient (Haven ratio) is unity. Without definitive knowledge concerning
the diffusion mechanism these are the simplest assumptions to make.
With these assumptions, one finds the following values for the sodium
ion diffusion coefficient, D, and mobility, μ, at 25°C: D=1.9X10-8 cm2/s, (5) μ
=7.4X10-6 cm2/V/s. (6) In all probability the hydrated cations
jump from the neighborhood of one ditrigonal site to a neighboring one,
7% of which are vacant. If such a vacancy mechanism is operative then
the number of charge carriers is equal to the vacancy concentration,
so that the above values would be higher. However, these vacant sites
are only vacant in the sense that there are no sodium ions occupying
them; they are believed to be occupied by "free" water molecules. In
addition the generally accepted size of a hydrated sodium ion, 3.58
Ǻ radius [9], is too large for the cation to sit in a regular manner
in the ditrigonal cavities, and there are insufficient water molecules
per sodium ion, 5, to satisfy the normal coordination requirements of
the cation so it is difficult to define the in-plane structure and hence
the diffusion path. A study of conductivity in a series of vermiculites
of different sodium contents would help in elucidating the mechanism;
the Llano vermiculite has one of the highest sodium contents, hence
lowest vacancy concentrations of all the vermiculites. Proton NMR studies [3] of Llano vermiculite
have shown two relaxation processes, with the faster being associated
with the motion of free water molecules and the slower with those water
molecules associated with the sodium ion's hydration sphere. For the
latter it was assumed that the relaxation process involved the slower
motion of the sodium ion together with its hydration sphere; a diffusion
coefficient of 0.5X10-8 cm2/s was calculated [3]
which is in relatively close agreement with the value calculated here
considering the assumptions made about the mobile ion concentration
and the correlation coefficient. We can thus assume that the diffusion
mechanism operative in the bilayer hydrate of sodium vermiculite Fig.
6. Comparison of the conductivities of sodium Llano vermiculite and
the fast ion conductor sodium β-alumina. are
jumps of the sodium ion, with its hydration sphere, from site to site.
Thus the mechanism is not dissimilar to that in aqueous solutions but
is severely constrained by the geometry of the clay layers thus leading
to a conductivity several orders of magnitude lower than that for sodium
chloride solutions. In agreement with this mechanism is the high enthalpy
for motion of the sodium ions, 11.7 kcals/ mole, which is a factor of
3 higher than the temperature coefficient of resistivity, »3.8 kcals/ mole, for sodium chloride
solutions at 25°C It is interesting to compare the
values obtained here with those [10] of sodium beta alumina, one of
the highest known solid state ionic conductors. These are shown in fig.
6 and table 1. Although it would appear that the vermiculite might have
the higher conductivity at elevated temperatures if the water can be
retained inside the crystal by an external pressure, the presence of
water of hydration does not appear Table
1 Sodium
ion mobility in sodium vermiculite and (beta) alumina. Property
Vermiculite
β-alumina σ, S/cm 25°C
3.0X10-4
1.4X10-2 D, cm2/s
25°C 1.9X10-8
6.7X10-7 μ, cm2/V/s
25°C 7.4X10-6
6.8X10-5 H, kcals/mole 11.7
3.79 [Na], moles/cc
4.2X 10-3
5.5 X 10-3 Fig.
7. Conductivity of Bandera sandstone as a function of salinity for some
sodium chloride and tetraethyl ammonium chloride solutions, after ref.
[11]. to
assist in the diffusion of the sodium ions at ambient temperatures where
the conductivity is a factor of 50 lower than in the sodium beta alumina;
this is despite the sodium ion concentration being essentially the same
in the two structures. However, preliminary results in our laboratory
indicate that when the water content is reduced to give the monolayer
hydrate, y»1.9, the conductivity drops by about
two ordes of magnitude. Recent data [4] on the high surface area clay
montmorillonite also show very low conductivities when dry. Thus it
appears that there is very strong bonding between the sodium ions and
the aluminosilicate matrix. This may be associated either with binding
of the mobile cation in the large trigonal sites as in the mica class
of clays, or with ordering of the cations toward those oxygen coordinated
with the lower valent aluminum ions. This study will be extended to include
the effects of change of cation and of solvating molecule, and of the
large difference in mobility between surface and interlayer cations.
In the latter case we know from studies on sandstones containing aluminosilicates
that the mobility of surface sodium ions approaches that in bulk aqueous
solutions. An example of this behavior is shown in fig. 7, where the
conductivity of Bandera sandstone [11] is plotted against that of the
bulk chloride solution over a wide salinity range for two cations, sodium
and tetraethyl ammonium. The tetraethyl cations are essentially immobile,
so that a linear plot is observed in that case. However, for the sodium
ions the overall conductivity is significantly enhanced, and the data
is fitted after [12] with a sodium ion mobility corresponding to an
equivalent conductivity of up to 38 S.cm2/eq. This mobility
is 40% of those in 1 molar solution. This may be compared with the vermiculite
interlayer cations studied here which have a mobility of much less than
1% of those in solution. This may also explain why other workers [13]
have found higher conductivities for the very high surface area montmorillonites.
For these materials it may be conjectured that a significant part of
the conductivity is associated with surface rather than bulk hydrated
cations. 4.
Conclusions The sodium ion conductivity in the
layered structure of the aluminosilicate, Llano vermiculite, has been
measured from 5 to 90°C and found to have the value 3X10-4
S/cm at 25°C with an enthalpy of motion of 11.7
kcals/mole. This conductivity value is significantly less than that
of the fast ion conductor sodium beta alumina. Combining the conductivity
data with tracer sodium and proton NMR studies it seems likely that
in these single crystal specimens the dominant diffusing species are
sodium ions and that they jump together with their hydration spheres.
Acknowledgement This work was presented as a recent
news paper at the 6th International Conference on Solid State Ionics
in Garmisch-Partenkirchen, September 7, 1987. I would like to thank
D. Hines, N. Wada and P.-z. Wong for many helpful discussions. References [1] R. Calvet and J. Mamy, C.R. Acad
Sci. (Paris) 273D (1971) 1251. [2] J. Keay and A. Wild, Soil Science
92 (1961) 54. [3] J. Hougardy, W.E.E. Stone and J.J.
Fripiat, J. Chem. Phys. 64 (1976) 3840. [4] T. Kawasa, H. Yokokawa and M. Dokiya,
in: 6th Int. Conf. Solid State Ionics, 9th Sept. 1987, to be published. [5] D.J. Cebula, R.K. Thomas and J.W.
White, Clays Clay Miner. 29 (1981) 241. [6] K. Norrish, Disc. Faraday Soc. 18
(1954) 120. [7] P.G. Slade, C. Dean, P.K. Schultz
and P.G. Self, Clays Clay Miner. 35 (1987) 177. [8] M. Susuki, N. Wada, D.R. Hines and
M.S. Whittingham, Phys. Rev. B36 (1987) 2844. [9] E.R. Nightingale, J. Chem. Phys.
63 (1959) 1381. [10] M.S. Whittingham and R.A. Huggins, J. Chem.
Phys. 54 (1971) 414. [11] P.-z. Wong, in: Physics and chemistry of
porous media II, AIP Conf. Proc. # 154, ed. J.R. Banavar, J. Koplik
and K.W. Winkler (1987). [12] H.J. Vinegar and M.H. Waxman, Geophysics
49 (1984) 1267. [13] Y.-Q. Fan, in: 6th Int. Conf. Solid State
Ionics, 9th Sept. 1987, to be published. |
|||||||||
|