Assembly of Metal-Anion Arrays within a Perovskite Host. Low-Temperature Synthesis of New Layered Copper-Oxyhalides, (CuX)LaNb2O7, X = Cl, Br

Article abstract of DOI:10.1021/ja991566u

A low-temperature topotactic route is used to assemble metal-anion arrays within a perovskite host. Ion exchange between RbLaNb₂O₇ and CuX₂ (X = Cl, Br) results in a new set of layered copperoxyhalide perovskites, (CuX)LaNb₂O₇. Rietveld structural analysis of X-ray powder diffraction data confirms the formation of a two-dimensional copper-halide network in the double-layered perovskite interlayer. This new structure type contains unusual CuO₂X₄ octahedra that corner-share with NbO₆ octahedra from the perovskite slab and edge-share with each other along all four equatorial edges. Magnetic susceptibility measurements show that both products exhibit antiferromagnetic transitions below 40 K. Additionally, these materials are found to be low-temperature phases, decomposing completely by 700°C. The synthetic approach described in this work is significant in that it demonstrates how host structures can be used as templates in the directed low-temperature assembly of extended metal-anion arrays.

Full text of DOI:10.1021/ja991566u

J. Am. Chem. Soc. 1999, 121, 10743-10746
10743
Assembly of Metal-Anion Arrays within a Perovskite Host.
Low-Temperature Synthesis of New Layered Copper-Oxyhalides,
(CuX)LaNb2O7, X ) Cl, Br
Thomas A. Kodenkandath,†,§ John N. Lalena,†,§ Weilie L. Zhou,§ Everett E. Carpenter,†,§
Claudio Sangregorio,§ A. U. Falster,‡ W. B. Simmons, Jr.,‡ Charles J. O’Connor,†,§ and
John B. Wiley*,†,§
Contribution from the Departments of Chemistry and Geology and Geophysics and the AdVanced
Materials Research Institute, UniVersity of New Orleans, New Orleans, Louisiana 70148-2820
ReceiVed May 11, 1999
Abstract: A low-temperature topotactic route is used to assemble metal-anion arrays within a perovskite
host. Ion exchange between RbLaNb2O7 and CuX2 (X ) Cl, Br) results in a new set of layered copper-
oxyhalide perovskites, (CuX)LaNb2O7. Rietveld structural analysis of X-ray powder diffraction data confirms
the formation of a two-dimensional copper-halide network in the double-layered perovskite interlayer. This
new structure type contains unusual CuO2X4 octahedra that corner-share with NbO6 octahedra from the perovskite
slab and edge-share with each other along all four equatorial edges. Magnetic susceptibility measurements
show that both products exhibit antiferromagnetic transitions below 40 K. Additionally, these materials are
found to be low-temperature phases, decomposing completely by 700 °C. The synthetic approach described in
this work is significant in that it demonstrates how host structures can be used as templates in the directed
low-temperature assembly of extended metal-anion arrays.
Introduction
within the assembled layer itself. To our knowledge there are
no reports of low-temperature topotactic routes being utilized
to construct extended inorganic arrays in receptive host materi-
als. Herein we describe the synthesis and characterization of a
new set of perovskite-related layered copper-oxyhalides,
(CuX)LaNb2O7 (X ) Cl and Br), where a copper-halide
network was assembled within a double-layered perovskite host.
As part of our continuing efforts to develop new low-
temperature (e500 °C) synthetic strategies for the manipulation
of nonmolecular compounds,1 we have sought to construct
covalently bound metal-anion arrays within receptive host
materials. Our goal is to use the hosts as templates for the
rational assembly of specific structural motifs. Of particular
interest here are two-dimensional transition-metal systems in
that such layered compounds can demonstrate a range of
interesting properties including high-temperature superconduc-
tivity and colossal magnetoresistance (CMR).2 There are only
a few general examples of topochemical manipulation of layered
compounds that result in covalent cross-linkages between host
layers; bifunctional organic anions have been “grafted” to select
layered compounds while pillaring reactions have been reported
in several systems.3,4 We seek, however, to establish covalent
bonding not only between host layers but also continuously
Experimental Section
Synthesis. The n ) 2 Dion-Jacobson phase, RbLaNb2O7, was
prepared by a method similar to that of Gopalakrishnan et al.5 from
La2O3 (Alfa, 99.99%), Nb2O5 (Alfa, 99.9985%), and Rb2CO3 (Alfa,
99%). The La2O3 was heated at 1050 °C for 24 h prior to use to remove
any water or carbonates. Stoichiometric amounts of La2O3 and Nb2O5
along with a 25% molar excess of Rb2CO3 were ground together, heated
for 12 h at 850 °C, ground again, and then heated an additional 24 h
at 1050 °C. The product was washed thoroughly with distilled water
and then dried at 120 °C overnight. The unit cell parameters (a ) 3.896
(9) Å, c ) 11.027(2) Å) were in good agreement with literature values
(a ) 3.885(2) Å, c ) 10.989(3) Å).5
† Department of Chemistry.
‡ Department of Geology and Geophysics.
§ Advanced Materials Research Institute.
(CuCl)LaNb2O7 and (CuBr)LaNb2O7 were prepared by a single-step
ion-exchange reaction. Initially, CuCl2 (Alfa, 99%) and CuBr2 (J. T.
Baker, 99%) were dried overnight at 140 °C. RbLaNb2O7 was then
combined with a 2-fold molar excess of the anhydrous copper halide
in sealed, evacuated (<10-3 Torr) Pyrex tubes and heated for 7 days
at 325 °C. X-ray powder diffraction of the unwashed products showed
the presence of rubidium halide byproduct. Samples were washed with
copious amounts of water to remove excess copper halide and the
rubidium byproduct; the products were green and dark reddish-brown
for the chloride and bromide compounds, respectively. The exchange
of other related double-layered perovskites, ALaNb2O7 (A ) H, Li,
Na, K, NH4), was also screened.
(1) (a) McIntyre, R. A.; Falster, A. U.; Li, S.; Simmons, W. B.; O’Connor,
C. J.; Wiley, J. B. J. Am. Chem. Soc. 1998, 120, 217. (b) Lalena, J. N.;
Cushing, B. L.; Falster, A. U.; Simmons, W. B.; Seip, C. T.; Carpenter, E.
E.; O’Connor, C. J.; Wiley, J. B. Inorg. Chem. 1998, 37, 4484. (c) Cushing,
B. L.; Wiley, J. B. J. Solid State Chem. 1998, 141, 385. (d) Cushing, B. L.;
Wiley, J. B. Mater. Res. Bull. 1999, 34, 271.
(2) (a) Bednorz, J. G.; Mu¨ller, K. A. Z. Phys. B: Condens. Matter. 1986,
64, 189. (b) Moritomo, Y.; Asamitsu, A.; Kuwahara, H.; Tokura, Y. Nature
1996, 380, 141.
(3) (a) Kikkawa, S.; Kanamaru, F.; Koizumi, M. Inorg. Chem. 1980,
19, 259. (b) Rouxel, J.; Palvadeau, P.; Venien, J. P.; Villieras, J.; Janvier,
P.; Bujoli, B. Mater. Res. Bull. 1987, 22, 1217.
(4) Butruille, J.-R.; Pinanvaia, T. J. In ComprehensiVe Supramolecular
Chemistry; Alberti, G., Bein, T., Eds.; Elsevier Science Inc.: New York,
1996; pp 219-250. Alberti, G. In ComprehensiVe Supramolecular Chem-
istry; Elsevier Science Inc.: New York, 1996; pp 151-187 and the
references therein.
(5) Gopalakrishnan, J.; Bhat, V.; Raveau, B. Mater. Res. Bull. 1987, 22,
413.
10.1021/ja991566u CCC: $18.00 © 1999 American Chemical Society
Published on Web 11/04/1999

10744 J. Am. Chem. Soc., Vol. 121, No. 46, 1999
Kodenkandath et al.
RbLaNb2O7 + CuX2 f
(CuX)LaNb2O7 + RbX (X ) Cl, Br)
Typically exchange reactions in these perovskite systems only
involve cations. In RbLaNb2O7, however, we see the co-
exchange of both copper and halide ions. Though this behavior
is not common, some examples have been reported. Researchers
examining the exchange of zeolites for the preparation of
cracking catalysts noted co-exchange from aqueous copper-
chloride solutions.9 Others studying the absorption of copper(II)
on cation-exchange resins observed a similar effect.10 In both
these cases, reactions presumably involved the species [CuCl-
(H2O)3]+. In contrast, a very recent report on the synthesis of
CuLa2Ti3O10 from the solid-state reaction of CuCl2 and the
Ruddlesden-Popper triple-layered perovskite, Na2La2Ti3O10,
showed no anion incorporation.11 Here, however, it is expected
that the greater layer charge of the host, twice that of
RbLaNb2O7, is important: to maintain a charge balance, Na2-
La2Ti3O10 would have to exchange two “CuCl+” units, for which
there may not be enough room in the interlayer. In a similar
Figure 1. Experimental, calculated, and difference X-ray powder
patterns for the Rietveld refinement of (CuCl)LaNb2O7.
Characterization. The products would not dissolve in a variety of
acids and bases, even with the use of a microwave digestion bomb.
Compositional analysis was therefore based on energy dispersive
spectroscopy (EDS) of sets of individual crystallites. Analyses were
carried out on a JEOL JSM 5410 scanning electron microscope (SEM)
equipped with an EDAX DX PRIME microanalytical system. The
parent compound, RbLaNb2O7, CuCl (Aldrich, 99.999%), and CuBr
(Aldrich, 98%) were used as standards. The chloride product was found
to have copper, chlorine, lanthanum, and niobium in the ratios of
1.03(4):0.99(2):1.01(5):2.00, respectively. For the bromide product, the
copper, bromine, lanthanum, and niobium were in ratios of
1.04(5):1.2(1):1.03(4):2.00, respectively. Rubidium was not detected
in either of the products.
X-ray diffraction data from a Philips X’pert-MPD system (Cu KR
radiation, λ ) 1.5418 Å) equipped with a graphite monochromator were
collected in step-scanning mode between 10 and 110° 2θ with a 0.02°
step width and 10 s count time. Magnetic susceptibility measurements
were made on a Quantum Design MPMS-5S SQUID susceptometer
between 2 and 300 K at 1000 G (field cooled). The decomposition of
(CuCl)LaNb2O7 was examined by a combination of thermogravimetric
analysis (TGA) and by heating a series of (CuCl)LaNb2O7 samples at
various temperatures overnight in sealed, evacuated Pyrex or fused-
silica tubes. TGA were performed on a TA Instruments Thermal Analyst
2000 system over the temperature range 25-800 °C with a ramp of 10
deg/min under flowing nitrogen. Decomposition products were deter-
mined by X-ray powder diffraction.
result on the direct reaction of KLaNb2O7 and CuSO4, Cu0.48
-
LaNb2O7 formed with only a small amount of sulfur detected
in the product.12 The larger size, as well as the higher charge
of the sulfate ion, likely suppresses its exchange. Our initial
studies on the exchange of CuX2 (X ) Cl, Br) with the series
ALaNb2O7 (A ) H, Li, Na, K, Rb, NH4) also indicate that the
size of the anion may be important. The smaller chloride ion
(1.69 Å)13 readily co-exchanges into all members of this series,
while the bromide ion (1.82 Å) only reacts with the compounds
having the largest interlayer spacings, RbLaNb2O7 and NH4-
LaNb2O7. Further studies on RbLaNb2O7 and the other layered
perovskites are needed to develop a full understanding as to
when the co-exchange of cations and anions will occur.
Structure. On exchange, the layers of the perovskite host
were found to significantly expand (ca. 0.70 Å) in c relative to
the parent for both the chloride and the bromide (Table 1).
Table 1. Tetragonal Unit Cell Parameters for the Parent and the
Exchange Products
compd
unit cell (Å)
cell vol (Å3)
167.4
lit. cell (Å)
RbLaNb2O7
a ) 3.896(9)
c ) 11.027(2)
a ) 3.885(2)5
c ) 10.989(3)
this work
Rietveld refinements were carried out with the GSAS programs.6
Profile refinements utilized the psuedo-Voigt function. The background
was refined with six variables using a linear interpolation function.
The host has been well characterized by others,5,7 so the site occupancies
for the niobium, lanthanum, and oxygen were all fixed at 100% and
only the copper and halogen occupancies were varied in the refinement.
The R-factor (Rp) and the weighted R-factor (Rwp) are defined as Rp )
(CuCl)LaNb2O7 a ) 3.8792(1)
c ) 11.7282(3)
(CuBr)LaNb2O7 a ) 3.8995(1)
c ) 11.7060(3)
176.49
178.00
this work
Various structural models were considered that could account
for the expansion as well as accommodate the copper and halide
ions in the interlayer. The most reasonable one had copper
cations surrounded by four coplanar halide ions and bridging
between apical oxygens of the perovskite slabs. Rietveld
refinements on this model gave the best profile fits. The
observed, calculated, and difference plots for the Rietveld
refinement of (CuCl)LaNb2O7 based on this model are shown
in Figure 1. Table 2 presents the structural parameters for
(CuCl)LaNb2O7 and (CuBr)LaNb2O7. The thermal parameters
for both the chlorine (0.133 Å2) and bromine (0.087 Å2) atoms
∑|Io - Ic|/∑Ic and Rwp ) [∑w(Io - Ic)2/∑(wIo )]1/2, where Io and Ic are
2
the observed and calculated intensities, respectively.
Results and Discussion
Synthesis. (CuX)LaNb2O7 (X ) Cl, Br) form from an ion
exchange reaction between RbLaNb2O7 and the copper(II)
halides, CuX2.8
(6) Larson, A.; Von Dreele, R. B. GSAS: Generalized Structure Analysis
System; Los Alamos National Laboratory: Los Alamos, NM, 1994.
(7) See for example: (a) Sato, M.; Abo, J.; Jin, T.; Ohta, M. Solid State
Ionics 1992, 51, 85. (b) Sato, M.; Abo, J.; Jin, T. Solid State Ionics 1992,
57, 285. (c) Sato, M.; Abo, J.; Jin, T.; Ohta, M. J. Alloys Compd. 1993, 81.
(d) Sato, M.; Jin, T.; Ueda, H. Chem. Lett. 1994, 161.
(9) Tsutsumi, K.; Fuji, S.; Takahashi, H. J. Catal. 1972, 24, 146.
(10) Nardin, M.; Talbot-Besnard, S. C. R. Acad. Sci. Paris 1969, 269,
1608.
(8) These reactions are carried out below the decomposition/melting
points of the copper halides and of the rubidium-halide byproducts. It is
possible that a eutectic may form between the copper and rubidium halides
to assist in the diffusion process, otherwise these reactions likely proceed
by solid-state diffusion of the ions.
(11) Hyeon, K.-A.; Byeon, S.-H. Chem. Mater. 1999, 11, 352.
(12) Matsuda, T.; Fujita, T.; Miyamae, N.; Takeuchi, M.; Kunou, I. J.
Mater. Chem. 1994, 4, 955.
(13) Ionic radii are from the following: Shannon, R. D. Acta Crystallogr.
1976, A32, 751.

Synthesis of New Layered Copper-Oxyhalides
J. Am. Chem. Soc., Vol. 121, No. 46, 1999 10745
Figure 2. The structure of (CuX)LaNb2O7 relative to that of the parent,
RbLaNb2O7. The large dark spheres in the interlayer of RbLaNb2O7
are rubidiums while the small dark and the larger light-colored spheres
of (CuX)LaNb2O7 are the copper and halide ions, respectively.
Table 2. Crystallographic Data for (CuX)LaNb2O7 (X ) Cl, Br)
sample
atom site
x
y
z
g
Uiso(Å2)
a
(CuCl)LaNb2O7 Cu 1d 0.5 0.5 0.5
1.00(1) 0.011(1)
1.01(1) 0.133(6)
Cl 1b
La 1a
0
0
0
0
0.5
0
1
0.0017(6)
0.0022(6)
0.005(2)
0.036(6)
0.020(6)
Nb 2h 0.5 0.5 0.1908(1) 1
O1 4i 0.5 0.1595(6) 1
0
O2 2h 0.5 0.5 0.342(1)
O3 1c 0.5 0.5 0
1
1
b
(CuBr)LaNb2O7 Cu 1d 0.5 0.5 0.5
1.009(9) 0.024(2)
1.02(1) 0.087(3)
Br 1b
La 1a
0
0
0
0
0.5
0
1
0.011(7)
0.0070(7)
0.001(3)
0.024(7)
0.021(9)
Nb 2h 0.5 0.5 0.1904(2) 1
O1 4i 0.5 0.1583(8) 1
0
O2 2h 0.5 0.5 0.338(1)
O3 1c 0.5 0.5 0
1
1
a P4/m, Z ) 1, Rp ) 6.92%, Rwp ) 8.90%. b P4/m, Z ) 1, Rp )
6.39%, Rwp ) 8.31%.
Table 3. Selected Bond Distances (Å) for (CuX)LaNb2O7
compd
Cu-O2 Cu-Xa Nb-O1 Nb-O2 Nb-O3
(CuCl)LaNb2O7 1.97(1) 2.7430(1) 1.974(1) 1.66(1) 2.238(2)
(CuBr)LaNb2O7 1.99(1) 2.7573(1) 1.990(1) 1.63(1) 2.230(2)
Figure 3. (a) Magnetic susceptibility (ø) versus temperature for (CuCl)-
LaNb2O7 relative to the parent compound RbLaNb2O7. (b) Inverse
susceptibility (1/ø) versus temperature. The solid line indicates a fit to
the Curie-Weiss law above 50 K.
a See text.
were found to be large. Such large values are often attributed
to either nonstoichiometry or disorder. Analytical data were in
reasonable agreement with the copper-halide stoichiometries
we have clearly established the existence of copper halide arrays
within the host materials, refinements on higher quality data
would allow us to better address some of these structural details.
An idealized structure of (CuX)LaNb2O7 is shown in Figure
2 relative to the parent, RbLaNb2O7. As would be expected for
a topotactic reaction, the host layer itself is essentially un-
changed. In the interlayer of the product, distorted CuO2X4
octahedra are found to corner-share with the NbO6 octahedra
of the perovskite slabs while edge-sharing with other CuO2X4
octahedra in the a-b plane. This arrangement is similar to what
is seen for mercury in NH4HgCl3.16 Copper-anion bond
distances (Table 3) are consistent with known copper com-
pounds,17,18 though in light of the apparent disorder of the
halides, the Cu-X distances are best viewed as rough averages.
Other examples of perovskite-related layered copper oxyha-
lides are known. Mu¨ller-Buschbaum18 and others19 have studied
a number of compounds and recently there have been reports
determined from the refinement, so it was expected that the
1
halogens were disordered off the ideal 0, 0,
/ position.
2
Anisotropic refinement of the halogen atoms was consistent with
this in that both the chlorine and bromine thermal ellipsoids
were found to be oblate, extending significantly in the a-b
plane.14 It is known that copper(II), as a Jahn-Teller ion, tends
to have both short and long bonding interactions. If the halogens
are responding to the bonding preferences of copper, they would
disorder selectively in the a-b plane. Initial efforts to model
disorder in the bromide compound showed that bromine is
slightly displaced off the 0, 0, 1/2 site.15 It is expected that similar
behavior occurs in the chloride compound, though the larger
thermal parameter would indicate that chlorine moves much
further off the ideal position. Efforts to obtain synchrotron and
neutron data on these systems are currently underway; though
(14) Anisotropic refinement of the chloride and bromide atoms produced
U11, U22, and U33 values of 0.166, 0.166, and 0.063, and 0.101, 0.101, and
0.048, respectively.
(16) Harmsen, E. J. Z. Kristallog. 1938, 100, 208.
(17) Wells, A. F. Structural Inorganic Chemistry, 5th ed.; Oxford
University Press: New York, 1986; pp 1135-1141.
(18) Mu¨ller-Buschbaum, H. Angew. Chem., Int. Ed. Engl. 1977, 16, 674.
Mu¨ller-Buschbaum, H. Angew. Chem., Int. Ed. Engl. 1991, 30, 723 and
the references therein.
(19) (a) Kodenkandath, T. A.; Calestani, G.; Matacotta, F. C. J. Mater.
Chem. 1996, 6, 1575. (b) Ruck, K.; Eckert, D.; Krabbes, G.; Wolf, M.;
Mu¨ller, K.-H. J. Solid State Chem. 1998, 141, 378.
(15) In modeling disorder in the bromide compound, the bromine was
1
refined on an x, x, /2 site in the P4/mmm space group. This resulted in a
slight movement off the 0, 0, 1/2 position to x ) 0.061(2). Also, the bromine
thermal parameter dropped to 0.033(4), the site occupancy on this 4-fold
site refined to the expected value of 0.25 (0.249(2)), and the R-factors
decreased to 6.01% and 7.93% for Rp and Rwp, respectively.

10746 J. Am. Chem. Soc., Vol. 121, No. 46, 1999
Kodenkandath et al.
of copper-oxyhalide superconductors.20,21 Though most of these
materials have been prepared at temperatures in excess of 800
°C, some have used low-temperature methods in fluorine
intercalation.21 None of these copper oxyhalides, however, are
isostructural with those reported in this study.
to a minor impurity introduced in the exchange step, it is also
possible that there are some copper ions in the interlayer that
were not accompanied by halides on exchange. These species
could be magnetically isolated from adjacent coppers, resulting
in a paramagnetic response.
Magnetism. Figure 3a shows a plot of magnetization versus
temperature for (CuCl)LaNb2O7 relative to the diamagnetic
parent. The high-temperature data (T > 50 K) of the chloride
compound (Figure 3b) demonstrate Curie-Weiss behavior (C
) 0.404 emu‚K/mol, θ ) - 5.06 K) with an effective magnetic
moment (µeff) of 1.86 µB. Analogous behavior is observed in
the bromide (C ) 0.424 emu‚K/mol, θ ) -11.99, µeff ) 1.79
µB). The µeff values are consistent with the presence of Cu2+
where moments are typically found in the range 1.75 to 2.10.22
The low-temperature magnetic response of these materials gives
further support for the establishment of copper-halide arrays
within the perovskite layers. The compounds show antiferro-
magnetic (AFM) transitions at 16 and 37 K for the chloride
and bromide, respectively. Coupling necessitates the presence
of the halide in that it is required to mediate the magnetic
exchange process.23 Similar AFM transitions are seen in several
other copper-halide compounds.24 It should be noted that a
small amount of paramagnetism was observed below the
transition temperature in each sample. Though this may be due
Thermal Analysis. Thermal analysis experiments showed
that (CuCl)LaNb2O7 starts to decompose with loss of CuCl2 by
450 °C and Cu2OCl2 by 550 °C. Complete collapse of the host
structure then occurs by 700 °C. This indicates that these
compounds are low-temperature phases and are likely not
accessible by direct reaction since the parent compound must
be synthesized at temperatures well above 700 °C.
Conclusions
The assembly of metal-anion networks within receptive hosts
could prove to be an important advance in low-temperature
syntheses as applied to nonmolecular systems. The ability to
prepare hosts at high temperatures, where interatomic spacings
and topologies can be varied, and then to use them as templates
in the low-temperature construction of metal-anion arrays,
offers a significant route to new rationally designed materials.
We have already found that similar chemistry with the copper
halides can be performed in various n ) 2 and 3 members of
the Dion-Jacobson series.25 Should other effective ion-exchange
reagents also be identified, such as metal halides, chalcogenides,
pnictides, etc., then this unique approach to materials synthesis
could be further exploited, possibly leading to whole new classes
of low-temperature or even metastable compounds.
(20) Jin, C.-Q.; Wu, X.-J.; Laffez, P.; Tatsuki, T.; Tamura, T.; Adachi,
S.; Yamauchi, H.; Koshizuka, N.; Tanaka, S. Nature 1995, 375, 301.
(21) (a) Al-Mamouri, M.; Edwards, P. P.; Greaves, C.; Slaski, M. Nature
1994, 369, 382. (b) Francesconi, M. G.; Slater, P. R.; Hodges, J. P.; Greaves,
C.; Edwards, P. P.; Al-Mamouri, M.; Slaski, M. J. Solid State Chem. 1998
135, 17.
(22) O’Connor, C. J. Prog. Inorg. Chem. 1982, 29, 203.
(23) Goodenough, J. B. Magnetism and the Chemical Bond; Wiley and
Sons: New York, 1963; pp 165-185.
(24) See: Ko¨nig, E. In Magnetic Properties of Transition-Metal Com-
pounds; Landolt-Bo¨rnstein Series Vol. 2; Springer-Verlag: New York,
1966; pp 321-322. Escoffier, P.; Cauthier, J. Compt. Rend. 1961, 252,
271.
(25) We have successfully applied this chemistry to several members of
this series including RbLaTa2O7, RbBiNb2O7, RbCa2Nb3O10, and RbLa2-
TiNb2O10.
Acknowledgment. Financial support from the Louisiana
Education Quality Support Fund (LEQSF(1994-1997)-RD-A-
35), DOD (DARPA MDA972-97-1-0003), and the American
Chemical Society Petroleum Research Fund (ACS-PRF 30522-
G5) is gratefully acknowledged.
JA991566U