Phase selection and energetics in chiral alkaline earth tartrates and their racemic and Meso analogues: Synthetic, structural, computational, and calorimetric studies

Article abstract of DOI:10.1021/ja905690t

The hydrothermal reactions of calcium, strontium, and barium with L-, meso-, and D,L-tartaric acid were examined from room temperature to 220°C. We report the synthesis of 13 new phases and crystal structures of 11 alkaline earth tartrates, including an u

Full text of DOI:10.1021/ja905690t

Published on Web 10/01/2009
Phase Selection and Energetics in Chiral Alkaline Earth
Tartrates and Their Racemic and Meso Analogues: Synthetic,
Structural, Computational, and Calorimetric Studies
Leah N. Appelhans,† Monica Kosa,‡ A. V. Radha,§ Petra Simoncic,§
Alexandra Navrotsky,*,§ Michele Parrinello,*,‡ and Anthony K. Cheetham*,|
Solid State Lighting and Energy Center, UniVersity of California, Santa Barbara, California
93106, Computational Science, Department of Chemistry, and Applied Biosciences, ETH Zurich,
c/o USI Campus, Via Giuseppe Buffi 13, CH-6900 Lugano, Switzerland, Peter A. Rock
Thermochemistry Laboratory and NEAT-ORU, UniVersity of California, One Shield AVenue,
DaVis, California 95616, and Department of Materials Science and Metallurgy, UniVersity of
Cambridge, Pembroke Street, Cambridge CB2 3QZ, U.K.
Received July 15, 2009; E-mail: anavrotsky@ucdavis.edu; parrinello@phys.chem.ethz.ch;
akc30@cam.ac.uk
Abstract: The hydrothermal reactions of calcium, strontium, and barium with L-, meso-, and D,L-tartaric
acid were examined from room temperature to 220 °C. We report the synthesis of 13 new phases and
crystal structures of 11 alkaline earth tartrates, including an unusual I3O0 framework, [Ba(D,L-Tar)] (Tar )
C4H4O62-), with 3-D inorganic connectivity. Each alkaline earth exhibits different phase behavior in the
reactions with the three forms of tartaric acid. Calcium forms unique L-, meso-, and D,L-tartrate phases
which persist to 220 °C. Strontium forms three unique phases at lower temperatures, but above 180 °C
reactions with L- and D,L-tartaric acid yield the meso phase. Likewise, Ba forms three unique low-temperature
phases, but above 200 °C reactions with L- and meso-tartaric acid yield the D,L phase. Computational and
calorimetric studies of the anhydrous calcium phases, [Ca(L-Tar)] and [Ca(meso-Tar)], strontium phases,
[Sr(L-Tar)] and [Sr(meso-Tar)], and barium phases, [Ba(L-Tar)] and [Ba(D,L-Tar)], were performed to
determine relative phase stabilities and elucidate the role of thermodynamic and kinetic factors in controlling
phase behavior. The computational and calorimetric results were in excellent agreement. The [Ca(meso-
Tar)] phase was found to be 9.1 kJ/mol more stable than the [Ca(L-Tar)] phase by computation (total
electronic energies) and 2.9 ( 1.6 kJ/mol more stable by calorimetry (enthalpies of solution). The [Sr(meso-
Tar)] phase was found to be 13.4 and 8.1 ( 1.4 kJ/mol more stable than [Sr(L-Tar)] by computation and
calorimetry, respectively. Finally, the [Ba(L-Tar)] phase was found to be 6.4 and 7.0 ( 1.0 kJ/mol more
stable than the [Ba(D,L-Tar)] phase. Our results suggest that the calcium and strontium meso phases are
the most thermodynamically stable phases in their systems over the temperature range studied. The phase
transitions are controlled by relative thermodynamic stabilities but also by a kinetic factor, likely the barrier
to isomerization/racemization of the tartaric acid, which is hypothesized to preclude phase transformations
at lower temperatures. In the barium system we find the [Ba(L-Tar)] phase to be the most thermodynamically
stable phase at low temperatures, while the [Ba(D,L-Tar)] phase becomes the thermodynamic product at
high temperatures, due to a larger entropic contribution.
stable to the reaction conditions can be used,4,5 amino acids
and other naturally derived carboxylic acids have the distinct
advantage of being available as the pure enantiomers, relatively
cheap, and easy to obtain. A number of recent studies have used
these types of ligands to reliably synthesize chiral frameworks.6-12
Although obtaining a chiral framework from chiral building
Introduction
Interest in the synthesis of chiral metal-organic frameworks1-3
has led to the increasing use of naturally derived chiral building
blocks, particularly carboxylic acids, for chiral framework
synthesis. The use of a chiral, enantiopure ligand is the most
fail-safe method for synthesizing chiral extended solids since,
unless the ligand racemizes, the resulting metal-organic
framework must also be chiral. While any chiral ligand that is
(4) Cho, S. H.; Ma, B. Q.; Nguyen, S. T.; Hupp, J. T.; Albrecht-Schmitt,
T. E. Chem. Commun. 2006, 2563.
† University of California, Santa Barbara.
‡ ETH Zurich.
(5) Cui, Y.; Evans, O. R.; Ngo, H. L.; White, P. S.; Lin, W. B. Angew.
Chem., Int. Ed. 2002, 41, 1159.
§ University of California, Davis.
(6) Anokhina, E. V.; Go, Y. B.; Lee, Y.; Vogt, T.; Jacobson, A. J. J. Am.
Chem. Soc. 2006, 128, 9957.
| University of Cambridge.
(1) Bradshaw, D.; Claridge, J. B.; Cuseen, E. J.; Prior, T. J.; Rosseinsky,
M. J. Acc. Chem. Res. 2005, 38, 273.
(7) Barja, B.; Baggio, R.; Calvo, R.; Garland, M. T.; Perec, M.; Rizzi, A.
Inorg. Chim. Acta 2006, 359, 3921.
(2) Lin, W. B. J. Solid State Chem. 2005, 178, 2486.
(3) Ma, L.; Abney, C.; Lin, W. Chem. Soc. ReV. 2009, 38, 1248.
(8) Thuery, P. CrystEngComm 2007, 9, 460.
(9) Zhang, J.; Yao, Y. G.; Bu, X. H. Chem. Mater. 2007, 19, 5083.
9
10.1021/ja905690t CCC: $40.75  2009 American Chemical Society
J. AM. CHEM. SOC. 2009, 131, 15375–15386 15375

A R T I C L E S
Appelhans et al.
blocks is relatively straightforward, understanding and control-
ling the factors that affect the structures and properties of both
chiral and achiral metal-organic frameworks remains a fun-
damental challenge in the field. Structural control is best
exercised in systems that combine a geometrically rigid ligand
and a metal with a well-defined coordination geometry. Flexible,
polydentate ligands, like many of the natural carboxylic acids,
have less predictable structural behavior. Composition and
structure, as would be expected, also depend on an array of
synthetic variables, and some general trends in hydration and
connectivity have been identified.13 The prediction and control
of structure is further complicated because reactions can occur
under either thermodynamic or kinetic control.14,15 In chiral
systems, the additional complications of ligand racemization16
and conglomerate formation are also important. Finally, the
existence (or nonexistence) of a predictable structural relation-
ship between chiral and racemic frameworks is also of consider-
able interest.9,12,14,17 Series of readily accessible frameworks
based on available chiral ligands are ideal systems to advance
the understanding of these areas.
Figure 1. The three forms of tartaric acid.
new phases, with unique phase behavior for each of the metals.
By the combination of synthetic, calorimetric, and computational
studies, we have been able to identify and begin to distinguish
the kinetic and thermodynamic factors controlling phase be-
havior in this complex system. We focus in particular on the
anhydrous phases of Ca, Sr, and Ba, allowing the determination
of relative stabilities by computational methods without the
complications arising from different levels of hydration.
Results and Discussion
1. Hydrothermal Synthesis of Alkaline Earth Tartrates. The
reactions of tartaric acid with calcium, strontium, and barium
were studied from room temperature to 220 °C. Tartaric acid
exists in three crystalline forms (Figure 1), excluding consid-
eration of polymorphs and hydrates: the chiral D- or L-tartaric
acid (S,S- and R,R- respectively), the racemic D,L-tartaric acid
(racemic mixture of the enantiomers), and the achiral diastere-
omer, meso-tartaric acid (R,S-tartaric acid). Reactions were
carried out with each of the three forms. The reactions yield 18
phases between room temperature and 220 °C, including the
five previously known phases. Eleven new crystal structures
are reported (Table 1). Structural descriptions of the new
anhydrous phases 2, 7, 11, 13, and 17, which were used in the
computational and calorimetry studies, are included in the text.
Complete structural descriptions of all new hydrate phases are
included in the Supporting Information.
1.1. Calcium Tartrates. The reactions of calcium acetate with
tartaric acid yield seven different phases between room tem-
perature and 220 °C (Table 2). Throughout this paper, con-
nectivity is described using the InOm notation.13 In is the
dimensionality of inorganic (metal-oxygen-metal) connectiv-
ity. Om is the dimensionality of organic connectivity, excluding
organic connectivity that is redundant with inorganic connectiv-
ity. The sum, n + m, is the overall connectivity and must be
less than or equal to 3. For example, a compound with both
3-D inorganic and 3-D organic connectivity would be described
as I3O0 because the organic connectivity does not increase the
overall connectivity.
Subsequent to our previous work18 on magnesium tartrates,
in which we found that the magnesium tartrate system exhibited
unexpected complexity, we became interested in studying the
other alkaline earth tartrates in order to better understand the
importance of ligand chirality, cation identity, and thermody-
namic and kinetic factors in controlling phase diversity and
behavior. The study of tartaric acid and its salts has a long
history, and as early as 1935 the space group of the enantiopure
room-temperature phase, calcium tartrate tetrahydrate, was
determined.19 Complete structures of calcium tartrate tetrahy-
drate and strontium tartrate trihydrate were subsequently
reported in 1968.20 Despite this early start, the investigation of
new alkaline earth tartrates has been limited and continues to
focus on room-temperature phases. The room-temperature
barium tartrate phase has been synthesized and its structure
determined21 and the structure of the room-temperature phase,
strontium tartrate tetrahydrate, solved.22 However, even with
this recent work, only five unique structures of calcium,
strontium, and barium tartrates were collected in the CCDC prior
to this work. In this work we describe a comprehensive study
of the synthesis of a series of calcium, strontium, and barium
tartrate frameworks from room temperature to 220 °C. The
results reveal a very rich and complex system, generating 13
(10) Barrio, J. P.; Rebilly, J. N.; Carter, B.; Bradshaw, D.; Bacsa, J.; Ganin,
A. Y.; Park, H.; Trewin, A.; Vaidhyanathan, R.; Cooper, A. I.; Warren,
J. E.; Rosseinsky, M. J. Chem.sEur. J. 2008, 14, 4521.
(11) Zingiryan, A.; Zhang, J.; Bu, X. H. Inorg. Chem. 2008, 47, 8607.
(12) Anokhina, E. V.; Jacobson, A. J. J. Am. Chem. Soc. 2004, 126, 3044.
(13) Cheetham, A. K.; Rao, C. N. R.; Feller, R. K. Chem. Commun. 2006,
4780.
Phase Behavior. In the calcium tartrate system, no isomer-
ization of the tartaric acid ligand is observed, even at high
temperatures. Calcium forms a unique phase from each of the
(14) Bailey, A. J.; Lee, C.; Feller, R. K.; Orton, J. B.; Mellot-Draznieks,
C.; Slater, B.; Harrison, W. T. A.; Simoncic, P.; Navrotsky, A.; Grossel,
M. C.; Cheetham, A. K. Angew. Chem., Int. Ed. 2008, 47, 8634.
(15) Lee, C.; Mellot-Draznieks, C.; Slater, B.; Wu, G.; Harrison, W. T. A.;
Rao, C. N. R.; Cheetham, A. K. Chem. Commun. 2006, 2687.
(16) Thushari, S.; Cha, J. A. K.; Sung, H. H. Y.; Chui, S. S. Y.; Leung,
A. L. F.; Yen, Y. F.; Williams, I. D. Chem. Commun. 2005, 5515.
(17) Fleck, M.; Tillmanns, E.; Bohaty, L.; Held, P. Z. Kristallogr. 2004,
219, 101.
(23) Sheldrick, G. M. SHELXTL-97 A Program for Crystal Structure and
Refinement, Ver. 5.1; University of Gottingen: Gottingen, Germany,
1997.
(24) Sheldrick, G. M. SADABS User Guide; University of Gottingen:
Gottingen, Germany, 1995.
(25) Spek, A. L. PLATON, A Multipurpose Crystallographic Tool; Uni-
versity of Utrecht: Utrecht, The Netherlands, 2005.
(26) Flack, H. D. Acta Crystallogr., Sect. A 1983, 39, 876.
(27) Boese, R.; Heinemann, O. Z. Kristallogr. 1993, 205, 348.
(28) Hawthorne, F. C.; Borys, I.; Ferguson, R. B. Acta Crystallogr., Sect.
N: Struct. Sci. 1982, 38, 2461.
(29) Devries, A. J.; Kroon, J. Acta Crystallogr., Sect. C: Cryst. Struct.
Commun. 1984, 40, 1542.
(30) Dranka, I.; Lupashku, T.; Shafranskii, V.; Ropot, V.; Kharitonov, Y. Y.
Zh. Neorg. Khim. 1994, 39, 834.
(31) Mastai, Y.; Sedlak, M.; Colfen, H.; Antonietti, M. Chem.sEur. J.
2002, 8, 2430.
(18) Kam, K. C.; Young, K. L. M.; Cheetham, A. K. Crys. Growth Des.
2007, 7, 1522.
(19) Evans, R. C. Z. Kristallogr. 1935, 92, 154.
(20) Ambady, G. K. Acta Crystallogr., Sect. B: Struct. Cryst. Cryst. Chem.
1968, 24, 1548.
(21) Gonzalez-Silgo, C.; Gonzalez-Platas, J.; Ruiz-Perez, C.; Lopez, T.;
Torres, M. Acta Crystallogr, Sect. C: Cryst. Struct. Commun. 1999,
55, 740.
(22) Starynowicz, P.; Meyer, G. Z. Anorg. Allg. Chem. 2000, 626, 2441.
9
15376 J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009

Chiral Alkaline Earth Tartrates and Their Analogues
A R T I C L E S
9
J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009 15377

A R T I C L E S
Appelhans et al.
Table 2. Reactions between Ca(OAc)2 and Tartaric Acida
modate two adjacent metal ions. The shortest M-O bond, 2.384
Å, is to the relatively unhindered hydroxy oxygen. M-O bond
lengths are similar to those for other Ca tartrates, as are the
bond lengths and angles within the tartrate ligand itself. No
potential hydrogen bonds were identified.
Ca
temp
(°C)
L-Tar (L)
D,L-Tar (DL)
meso-Tar (m)
rt Ca(L)(H2O)2 ·2H2O (1)b Ca(DL)(H2O)4 (3)c Ca(m)(H2O)2 ·H2O (5)b
I0O3 2/2 234.6 I0O1 4/0 243.0 I0O2 2/1 226.9
Structure of [Ca(meso-Tar)] (7). The achiral meso-tartaric acid
(R,S-tartaric acid) forms a 3-D framework with Ca, crystallizing
in the space group C2/c. Like Ca(L-Tar), the meso phase is an
anhydrous framework with I1O2 connectivity. The asymmetric
unit contains one metal and one tartrate ligand. Each tartrate
ligand is µ5,κ6, bound to five metal centers through all six of
the oxygens (Figure 3a), and each metal is eight-coordinate.
The M-O-M polyhedra form 1-D edge-sharing chains that
extend along the [101] axis (Figure 3b). Each chain is connected
by tartrate linkages to the six surrounding chains (Figure 3c).
The average Ca-O bond lengths are typical, 2.476 Å. The
shortest and longest Ca-O bonds are 2.384 and 2.684 Å to
carboxylate oxygens, very similar to the bond lengths found in
Ca(L-Tar) (2). Two crystallographically unique hydrogen bonds
exist, between the hydroxy groups and carboxylate oxygens
(Figure 3d). The bond lengths and angles within the tartrate
ligand are unexceptional. The only major difference in tartrate
geometry between the meso and L-tartrate phases is the
C-C-C-C dihedral angle, which is 63.4° in Ca(m-Tar) and
33.8° in Ca(L-Tar). However, dihedral angles vary widely in
known alkaline earth tartrate structures, independent of the
tartrate diastereomer. The dihedral angles in Ca(L-Tar) and
Ca(m-Tar) are both well within this range.
Structural Trends. In terms of hydration and connectivity,
the overall phase behavior is fairly typical of hybrid frameworks
(see Table 2). Increasing reaction temperatures yield less
hydrated phases and generally higher connectivity, as expected.13
In the L-tartrate series, the hydration decreases from four waters
per metal atom in the low-temperature phase (1) to an anhydrous
phase (2) at high temperatures. Although the overall connectivity
is 3-D in both phases, the inorganic connectivity increases from
0-D to 1-D with the increase in temperature. The low-
temperature D,L-tartrate phase, 3, is a tetrahydrate, while the
high-temperature phase, 4, is a hemihydrate. In the meso system
the trend in hydration is maintained. The lowest temperature
phase, 5, contains three waters; the intermediate phase, 6,
contains two waters; and the high-temperature phase, 7, is
anhydrous. In the meso system, both overall and inorganic
connectivity increase from the low- to high-temperature phase.
Overall connectivity increases from 2-D in 5 to 3-D in 7, and
inorganic connectivity increases from 0-D to 1-D.
60 Ca(L)(H2O)2 ·2H2O (1) Ca(DL)(H2O)4 (3) Ca(m)(H2O)2 ·H2O (5)
I0O3 2/2 234.6
I0O1 4/0 243.0
I0O2 2/1 226.9
100 Ca(L)(H2O)2 ·2H2O (1) Ca(DL)(H2O)4 (3)
Ca(m)(H2O)2 (6)
- 2 -
Ca(L) (2)
I0O1 4/0 243.0
125
150
180
200
220
Ca(L) (2)
Ca(DL)(H2O)4 (3)
Ca(m) (7)
I1O2 0/0 137.8
I0O1 4/0 243.0
I1O2 0/0 140.9
Ca(L) (2)
Ca(DL)(H2O)4 (3)
[Ca(DL)]2 ·H2O (4)
Ca(m) (7)
I1O2 0/0 137.8
I1O2 0/0 140.9
Ca(L) (2)
Ca(DL)(H2O)4 (3)
[Ca(DL)]2 ·H2O (4)
Ca(m) (7)
I1O2 0/0 137.8
I1O2 0/0 140.9
Ca(L) (2)
[Ca(DL)]2 ·H2O (4)
- 0.5 -
Ca(m) (7)
I1O2 0/0 137.8
I1O2 0/0 140.9
Ca(L) (2)
[Ca(DL)]2 ·H2O (4)
- 0.5 -
Ca(m) (7)
I1O2 0/0 137.8
I1O2 0/0 140.9
a All products are phase pure except where two phases are listed. The
table lists composition (Tar
C4H4O62-), connectivity, bound/free
)
waters, and vol (Å3)/mol of metal. We were not able to obtain
single-crystal structures of phases 4 and 6 and thus do not know if the
waters are bound or free. b Phases 127,28 and 529 have been previously
described in the literature. c Compound 3 has been previously cited in
the literature30,31 and is commercially available (TCI), but no
characterization has been reported.
three forms of tartaric acid (Table 2). Although we were unable
to obtain a single crystal of the high-temperature phase, [Ca(D,L-
Tar)]2 ·H2O (4), we were able to confirm by powder X-ray
diffraction (PXRD) that it does not match any of the other
calcium phases or mixtures thereof. The reaction with L-tartaric
acid yields pure Ca(L-Tar) (2), even at 220 °C. Circular
dichroism spectroscopy confirms that there is no formation of
a racemic conglomerate, that is, a mechanical mixture of
enantiopure phases, e.g., Ca(L-Tar)/Ca(D-Tar). Neither is there
significant racemization of the tartrate ligand (see Supporting
Information).
Calcium Tartrate Structures. Only the structures of the two
anhydrous phases, Ca(L-Tar) (2) and Ca(m-Tar) (7), that are
used in the computational and calorimetric studies are described.
The structure of Ca(D,L-Tar)(H2O)4 (3), experimental procedures,
and characterization of all new phases are fully described in
the Supporting Information.
1.2. Strontium Tartrates. The reactions between strontium
acetate and tartaric acid yield six new phases from room
temperature to 200 °C (Table 3).
Structure of [Ca(L-Tar)] (2). Anhydrous Ca(L-Tar) (2)
crystallizes in the chiral space group C2221 as a 3-D framework,
with I1O2 connectivity. The asymmetric unit consists of the metal
center, which lies on a special position on the two-fold rotation
axis, and half a tartrate ion. The tartrate ligand overall is µ5,κ6,
bridging five metal ions and bound through each of its six
oxygen atoms (Figure 2a). The metal center is eight-coordinate
bound by three bis-chelating and two monodentate tartrates. The
extended structure is 3-D, connected along the c-axis by 1-D
zigzag chains of edge-sharing M-O-M polyhedra (Figure
2b,d). Tartrate linkages to six adjacent chains along the a and
b directions complete the 3-D network (Figure 2c). The average
M-O bond distance is 2.465 Å. The longest M-O bond, 2.597
Å, is to the bridging carboxylate oxygen, which must accom-
In the strontium system, we were surprised to observe
“disappearing polymorphs” 33 of two of the low-temperature
phases. The observed “disappearing” phases are actually
pseudopolymorphs; that is, the essential composition, Sr(L-Tar),
is the same, but the phases differ in level of hydration. Initial
experiments at room temperature and 60 °C yielded Sr(L-
Tar)(H2O)2 ·2H2O (8) and Sr(L-Tar)(H2O) (10), respectively. The
initial room-temperature phase, 8, which is isostructural to the
known room-temperature calcium phase, 1, has been reported
in the literature and the structure determined.22,32 The mono-
(32) Bohandy, J.; Murphy, J. C. Acta Crystallogr., Sect.B: Struct. Cryst.
Cryst. Chem. 1968, 24, 286.
(33) Dunitz, J. D.; Bernstein, J. Acc. Chem. Res. 1995, 28, 193.
9
15378 J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009

Chiral Alkaline Earth Tartrates and Their Analogues
A R T I C L E S
Figure 2. Views of Ca(L-Tar) (2). (a) Metal-tartrate coordination. (b) View down the b-axis, showing the 1-D edge-sharing chains of M-O-M polyhedra.
(c) View down the c-axis, showing (yellow bonds) the tartrate ligands linking each M-O-M chain to six others. (d) Alternate view of the M-O-M chains,
looking at the (-19 32 0) plane.
Figure 3. Views of Ca(m-Tar) (7). (a) Metal-tartrate coordination. (b) View down the b axis, showing the 1-D edge-sharing M-O-M chains. (c) View
down the [101] axis, showing the tartrate ligands (yellow bonds) linking each M-O-M chain to six others. (d) Hydrogen bonding showing two
crystallographically unique hydrogen bonds (yellow dashed bonds). Dashed bonds in blue are symmetry equivalent to yellow bonds.
hydrate Sr(L-Tar)(H2O) (10) is a new phase and was fully
were synthesized numerous times over a period of months.
characterized, including a single-crystal structure. Both phases
However, at the end of this period, a reaction at 60 °C, under
9
J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009 15379

A R T I C L E S
Appelhans et al.
Table 3. Reactions between Sr(OAc)2 and Tartaric Acida
temperatures, because in order to form the meso phase from L-
or D,L-tartaric acid, the isomerization of tartaric acid is required.
Tartaric acid does not racemize rapidly at low temperatures, so
at lower temperatures it may be the kinetic barrier to isomer-
ization, rather than thermodynamic control, that determines
phase formation. Alternatively, the meso phase may only be
the thermodynamically stable phase at higher temperatures, due
to entropic contributions that only become significant as the
temperature increases. These two scenarios cannot be easily
differentiated by synthetic experiments. For this reason, and to
better understand the different high-temperature phase behaviors
of the isostructural calcium and strontium systems, we undertook
both computational and calorimetric studies on the anhydrous
calcium and strontium phases, which will be described later.
Sr
temp
(°C)
L-Tar (L)
D,L-Tar (DL)
meso-Tar (m)
,
rt
Sr(L)(H2O)2 ·2H2O (8)b c Sr(DL)(H2O)4 (12)
Sr(m) (13)
I0O3 2/2 249.3
I0O1 4/0 255.1
I1O2 0/0 152.1
60
Sr(L)(H2O) (10)c
Sr(DL)(H2O)4 (12)
Sr(m) (13)
I1O2 1/0 182.7
I0O1 4/0 255.1
I1O2 0/0 152.2
100
125
150
180
200
Sr(L)(H2O) (10)
Sr(L) (11)
Sr(DL)(H2O)4 (12)
Sr(m) (13)
I0O1 4/0 255.1
I1O2 0/0 152.3
Sr(L) (11)
Sr(DL)(H2O)4 (12)
Sr(m) (13)
I1O2 0/0 148.9
I0O1 4/0 255.1
I1O2 0/0 152.4
Sr(L) (11)
Sr(DL)(H2O)4 (12)
Sr(m) (13)
Sr(m) (13)
I1O2 0/0 148.9
I1O2 0/0 152.5
Strontium Tartrate Structures. Only the anhydrous phases,
Sr(L-Tar) (11) and Sr(m-Tar) (13), that are used in the
computational and calorimetry studies are described. Full
structural descriptions of the new phases 10 and 12, as well as
experimental procedures and characterization of all new phases,
are included in the Supporting Information.
Sr(L) (11)
Sr(m) (13)
Sr(DL)(H2O)4 (12)
Sr(m) (13)
Sr(m) (13)
I1O2 0/0 152.6
Sr(m) (13)
Sr(m) (13)
Sr(m) (13)
I1O2 0/0 152.5
I1O2 0/0 152.6
I1O2 0/0 152.7
a All products are phase pure except where two phases are listed.
The table lists composition (Tar ) C4H4O62-), connectivity, bound/
free waters, and vol (Å3)/mol of metal. b Phase 8 has been previously
reported in the literature.22,32 c Disappearing pseudopolymorphs.
Later experiments yielded the known phase20 Sr(L-Tar)(H2O)3 (9).
Structure of Sr(L-Tar) (11). The anhydrous phase Sr(L-Tar)
is isostructural to Ca(L-Tar) (2). The only major differences are
in the M-O bond lengths, which are uniformly longer in the
strontium phase, as expected. The average M-O bond distance
is 2.566 Å in Sr(L-Tar). The longest and shortest M-O bonds
are 2.637 and 2.514 Å, respectively. M-O bond lengths are
similar to those in other strontium tartrates. The bond lengths
and angles within the tartrate ligand itself are normal. No
potential hydrogen bonds were identified.
identical experimental conditions to the original syntheses but
using a lower purity sample of D-tartaric acid, rather than
L-tartaric acid, was performed. Instead of yielding the expected
product, Sr(D-Tar)(H2O) (10), the reaction yielded the pseudopoly-
morph, Sr(D-Tar)(H2O)3 (9). Subsequent to this synthesis of 9,
the original phases, 8 and 10, could no longer be obtained under
the original conditions. The pseudopolymorph Sr(D-Tar)(H2O)3
(9) has been previously reported in the literature as the sole
product at room temperature, and the structure has been
solved.20,34 We suspect that impurities present in the D-tartaric
acid reagent were key in initiating the formation of the
“dominant” pseudopolymorph, 9, which subsequently represses
formation of the metastable pseudopolymorphs, 8 and 10.33 For
more details, see the Supporting Information.
Four of the calcium and strontium phases are isostructural:
the room-temperature phases M(L-Tar)(H2O)2 ·2H2O (1, Ca and
8, Sr), the D,L phases M(D,L-Tar)(H2O)4 (3, 12), and both the L
and meso anhydrous phases, M(L-Tar) (2, 11) and M(m-Tar)
(7, 13).
Phase Behavior. Although the anhydrous phases of calcium
and strontium are isostructural, the high-temperature phase
behavior of the strontium tartrates is strikingly different than
that of calcium. At a synthesis temperature of 200 °C, reaction
with any of the three forms of tartaric acid yields a single
product, the anhydrous meso phase, Sr(m-Tar) (13). Furthermore,
both the L phase, Sr(L-Tar) (11), and the D,L phase, Sr(D,L-
Tar)(H2O)4 (12), can be converted to Sr(m-Tar) by hydrothermal
treatment of a suspension at 200 °C over several days. The
reaction of strontium and tartaric acid at high temperatures is
clearly under thermodynamic control, and the meso phase, 13,
is the most thermodynamically stable phase.
Structure of Sr(meso-Tar) (13). The anhydrous phase Sr(m-
Tar) is also isostructural to the calcium phase, Ca(m-Tar). Again
only the M-O bond lengths are different. The average M-O
bond length is 2.587 Å, and the longest and shortest bonds are
2.747 and 2.516 Å, respectively. Two crystallographically
unique hydrogen bonds exist, between the hydroxy groups and
carboxylate oxygens. The bond lengths and angles of the tartrate
ligand are standard. As in the calcium tartrates, the geometry
of the tartrate ligand in Sr(L-Tar) versus Sr(m-Tar) only differs
significantly in the C-C-C-C dihedral angles, which are 35.2°
and 62.4°, respectively. The dihedral angles in each are
substantially similar to those in the respective isostructural
calcium phases.
Structural Trends. The strontium tartrates also follow the
expected trends in terms of hydration and connectivity (see Table
3). The hydration in the L-tartrate phases decreases from 4 to 3
to 1 to 0 water molecules per Sr with increasing temperature.
The overall connectivity for all the L-tartrate phases is 3-D, but
inorganic connectivity increases from 0-D to 1-D with an
increase in synthesis temperature.
Only Sr(L-Tar)(H2O) (10) maintained crystallinity with loss
of water, forming the high-temperature phase Sr(L-Tar) (11) on
dehydration, as shown by variable-temperature PXRD (see
Supporting Information). The transformation of 10 to 11 requires
loss of a coordinated water and a major rearrangement of the
metal-tartrate coordination (Figure 4). All other strontium
hydrate phases became amorphous on dehydration.
From the experimental work it is not possible to determine
if the meso phase is the most thermodynamically stable at all
Solid Solutions: Strontium and Calcium. The strontium and
calcium L-Tar phases, Sr(L-Tar) (11) and Ca(L-Tar) (2), form
infinite solid solutions at 150 °C with the lattice parameters
(34) Arora, S. K.; Patel, V.; Kothari, A.; Amin, B. Cryst. Growth Des.
2004, 4, 343.
9
15380 J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009

Chiral Alkaline Earth Tartrates and Their Analogues
A R T I C L E S
Figure 4. Coordination of tartrate ligand in (a) Sr(L-Tar) (11) and (b) Sr(L-Tar)(H2O) (10), showing the rearrangement that occurs on dehydration of 10 to
yield 11.
Figure 5. Powder X-ray diffraction patterns of Ca(1-x)Sr(x)(L-Tar) solid
Figure 6. Plot of volume (Å3) versus composition for Ca(1-x)Sr(x)(L-Tar)
solutions. The largest peak, at ∼25° 2θ, has been truncated in the bottom
solid solutions. Cell parameters were determined by Reitveld refinement
using GSAS/EXPGUI,35,36 and composition was determined by ICP.
three plots for clarity. The full diffraction patterns (0-60° 2θ) are included
in the Supporting Information.
(18) can be converted to Ba(D,L-Tar) by hydrothermal treatment
at 220 °C over several days. Thus, Ba(D,L-Tar) is clearly the
thermodynamic product relative to Ba(m-Tar)(H2O) above 200
°C.
following Vegard’s law (Figures 5 and 6). Solid solution
formation at other temperatures was not studied.
1.3. Barium Tartrates. The reactions between barium acetate
and tartaric acid from room temperature to 220 °C (Table 4)
form five unique phases. The barium tartrates are less diverse
than the calcium and strontium systems. In fact, only D,L-tartaric
acid forms more than one phase with barium over the temper-
ature range studied.
Phase Behavior. All three D,L phases form at room temper-
ature, and the composition of the final mixture is highly sensitive
to small variations in reaction conditions. Most often the room-
temperature reaction yields mixtures of all three phases, but
occasionally mixtures of only two phases are observed. Above
60 °C the only phase formed is the anhydrous phase, Ba(D,L-
Tar) (17).
The reaction with L-tartaric acid also yields mixtures of Ba(L-
Tar) (14) and Ba(D,L-Tar) at 220 °C. Pure Ba(D,L-Tar) can be
obtained starting from L-tartaric acid at 220 °C at low
concentrations (0.5 mM) and over long reaction times (5 days).
These observations suggest that, at 200 °C and above, Ba(D,L-
Tar) is the thermodynamically most stable phase. The formation
of the Ba(D,L-Tar) phase as the product of the reaction with
D,L-tartaric acid, rather than formation of a racemic conglomerate
of the Ba(L-Tar)/Ba(D-Tar) phase, supports this conclusion.
However, at our normally used concentrations (1 mM), pure
Ba(D,L-Tar) is never obtained from the reaction of barium and
L-tartaric acid, even with long reaction times. Furthermore,
treatment of a suspension of Ba(L-Tar) at 220 °C for days leaves
the sample unchanged. These results suggest that, at these
concentrations, the Ba(L-Tar) phase is the kinetic product of
the reaction and must have a sufficiently high barrier to
conversion to Ba(D,L-Tar) that the latter can not form under
the conditions studied. One possible scenario is that rapid
formation and precipitation of a highly insoluble kinetic product,
e.g., Ba(L-Tar), precludes formation of the thermodynamic
The high-temperature phase behavior of the barium tartrates
is more complex than that of calcium or strontium. At 200 °C
and above the reaction with meso-tartaric acid yields Ba(D,L-
Tar) (17), and a suspension of the meso phase Ba(m-Tar)(H2O)
(35) Larson, A. C.; Von Dreele, R. B. General Struture Analysis System
(GSAS) Manual; Los Alamos National Laboratory Report LAUR 86-
748; Los Alamos National Laboratory: Los Alamos, NM, 2000.
(36) Toby, B. H. J. Appl. Crystallogr. 2001, 34.
9
J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009 15381

A R T I C L E S
Appelhans et al.
Table 4. Reactions between Ba(OAc)2 and Tartaric Acida
bonds are reported in the previous work, one of the reported
H-bonds has a very small D-H-A angle (128°) and, therefore,
in this work is not included. We consider the structure to contain
a single crystallographically unique hydrogen bond, between a
hydroxy and carboxylate oxygen (Figure 7d).
Ba
temp
(°C)
L-Tar (L)
D,L-Tar (DL)
meso-Tar (m)
rt
Ba(L) (14)b
Ba2(DL)2(H2O)5 (15)
I1O1 2.5/0 216.0
Ba(DL)(H2O)2 (16)
I1O1 2/0 213.9
Ba(m)(H2O) (18)
I1O2 0/0 155.1
I1O2 1/0 185.1
Structure of [Ba(D,L-Tar)] (17). Compound 17 crystallizes
in P21/c, and the asymmetric unit is made up of one barium
atom and one tartrate ligand. The tartrate ligand is µ6,κ6, bound
through each oxygen to six barium atoms (Figure 8a). The
barium atom is 10-coordinate, and the structure is a rare example
of 3-D M-O-M connectivity (I3O0). Each barium atom is
connected by two bridging oxygen atoms to four adjacent barium
atoms in a pseudotetrahedral arrangement (Figure 9a). The
resulting 3-D structure of edge-sharing M-O-M polyhedra is
further interconnected by tartrate ligands, leading to a very
complex overall structure (Figure 8b). The average Ba-O bond
distance is 2.889 Å, and the shortest and longest bonds are 2.647
and 3.078 Å. Interestingly, the overall arrangement of the barium
atoms is diamondoid (Figure 9). The L- and D-enantiomers of
the tartrate ligand segregate within the crystal, forming alternat-
ing corrugated layers in the ab plane (Figure 8c). One crystal-
lographically unique hydrogen bond exists, between a hydroxy
group and carboxylate oxygen (Figure 8d).
Structural Trends. The meso and L-tartrate barium com-
pounds form only a single phase, but the multiple phases of
the barium D,L-tartartes follow the expected trends in connectiv-
ity and hydration (see Table 4). The hydration decreases from
2-2.5 in the low-temperature phases to 0 in the high-
temperature phase, and both inorganic and overall connectivity
increase with increasing temperature. The overall connectivity
increases from 2-D to 3-D, and the inorganic connectivity
increases from 1-D to 3-D.
Solid Solutions: Barium and Strontium. Although the pure
phases are not isostructural, barium and strontium form partial
solid solutions in reaction with meso-tartaric acid at 60 °C
(Figure 10). The structure of the Ba(m-Tar)(H2O) (18) phase is
described in detail in the Supporting Information. Over the range
of x ) 0 to x ≈ 0.1, a solid solution of the anhydrous strontium
phase, Sr(1-x)Ba(x)(m-Tar), exists. Between x ≈ 0.1 and x ≈ 0.2
there is a two-phase region, and from x ) 0.2 to x ) 1.0 there
isasecondsolidsolutionregionofthebariumphase,Sr(1-x)Ba(x)(m-
Tar)(H2O). It is interesting that the strontium forms an extensive
solid solution with the barium hydrate, Ba(m-Tar)(H2O), even
though the analogous phase Sr(m-Tar)(H2O) does not form under
the conditions studied. Likewise, barium forms a solid solution
with the anhydrous strontium meso phase, although over a much
smaller range of compositions, even though no anhydrous
barium meso phase forms below 220 °C. The hydrate Ba(m-
Tar)(H2O) does convert to a new high-temperature phase on
dehydration (see Supporting Information), but this phase is not
isostructural to the anhydrous strontium phase.
2. Computational and Calorimetric Studies. In order to better
understand the differences in high-temperature phase behavior
between the Ca, Sr, and Ba systems, we determined the relative
stabilities of the anhydrous high-temperature phases by com-
putational methods37-39 and acid solution calorimetry (Table
5) (see Supporting Information for additional details). The
hydrated phases, Ca- and Sr(D,L-Tar)(H2O)4 (3 and 12) and
Ba(m-Tar)(H2O) (18), were not considered in this part of the
Ba(DL) (17)
60
Ba(L) (14)
Ba(DL) (17)
Ba(m)(H2O) (18)
I1O2 0/0 155.1
I3O0 0/0 152.6
I1O2 1/0 185.1
100
125
150
180
200
220
Ba(L) (14)
Ba(DL) (17)
Ba(m)(H2O) (18)
I1O2 0/0 155.1
I3O0 0/0 152.6
I1O2 1/0 185.1
Ba(L) (14)
Ba(DL) (17)
Ba(m)(H2O) (18)
I1O2 0/0 155.1
I3O0 0/0 152.6
I1O2 1/0 185.1
Ba(L) (14)
Ba(DL) (17)
Ba(m)(H2O) (18)
I1O2 0/0 155.1
I3O0 0/0 152.6
I1O2 1/0 185.1
Ba(L) (14)
Ba(DL) (17)
Ba(m)(H2O) (18)
I1O2 0/0 155.1
I3O0 0/0 152.6
I1O2 1/0 185.1
Ba(L) (14)
Ba(DL) (17)
Ba(m)(H2O) (18)
Ba(DL) (17)
I1O2 0/0 155.1
I3O0 0/0 152.6
Ba(L) (14)
Ba(DL) (17)
Ba(DL) (17)
Ba(m)(H2O) (18)
Ba(DL) (17)
I3O0 0/0 152.6
a All products are phase pure except where two phases are listed. The
table lists composition (Tar
)
C4H4O62-), connectivity, bound/free
waters, and vol (Å3)/mol of metal. b Phase 14 has been previously
described in the literature.21
product, e.g., Ba(D,L-Tar), except at very low concentrations
where precipitation of the kinetic product is no longer favorable.
This scenario would explain our experimental observations at
high and low concentrations.
The high-temperature barium reactions appear to be controlled
by both thermodynamic and kinetic factors. The reactions with
L-tartaric acid at higher concentrations are under kinetic control,
and Ba(L-Tar) is formed. When concentration disfavors initial
formation of Ba(L-Tar), the thermodynamically stable product,
Ba(D,L-Tar), is observed. At low temperatures we cannot
determine, from the synthetic results, if the phase behavior is
controlled by the same factors. The relative phase stabilities at
low temperature may be different than at high temperature, or
the isomerization/racemization of tartaric acid may play a key
kinetic role. These questions will be addressed in the discussion
of the computational and calorimetric studies.
Barium Tartrate Structures. Only the anhydrous phases,
Ba(L-Tar) (14) and Ba(D,L-Tar) (17), used in the computational
and calorimetric studies are described. Full structural descrip-
tions of 15, 16, and 18, and experimental procedures and
characterization of all new phases are included in the Supporting
Information.
Structure of [Ba(L-Tar)] (14). The structure of 14 was solved
by Torres,21 but we will briefly describe it here in the terms we
have used for the previous compounds. Compound 14 crystal-
lizes in the chiral space group P212121. The tartrate ligand is
µ6,κ6 (Figure 7a). The barium is nine-coordinate, and the overall
connectivity is 3-D. The M-O-M connectivity is only 1-D,
with the Ba-O polyhedra forming chains of face-sharing
polyhedra along the c-axis (Figure 7b,c). The average Ba-O
distance is 2.816 Å, and the shortest and longest bonds are 2.769
and 2.875 Å. Although two crystallographically unique hydrogen
(37) CP2K, http://cp2k.berlios.de.
(38) VandeVondele, J.; Krack, M.; Mohamed, F.; Parrinello, M.; Chassaing,
T.; Hutter, J. Comput. Phys. Commun. 2005, 167, 103.
(39) Krack, M.; Parrinello, M. Forschungs. Ju¨lich, NIC Ser. 2004, 25, 29.
9
15382 J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009

Chiral Alkaline Earth Tartrates and Their Analogues
A R T I C L E S
Figure 7. Views of Ba(L-Tar) (14). (a) Metal-tartrate coordination. (b) View down the c axis showing an end view of the 1-D M-O-M chains. (c) View
down the [101] axis showing the M-O-M chains. (d) Hydrogen bonding showing one crystallographically unique bond (dashed yellow bonds). Dashed
bonds in blue are symmetry equivalents.
Figure 8. Views of Ba(D,L-Tar) (17). (a) Metal-tartrate coordination. (b) View of M-O polyhedra with two of the six-membered M-O-M rings tinted
as a visual aid. (c) View of the structure down the a axis showing segregation of the L- (yellow) and D-tartrates (green) into separate layers. (d) Hydrogen
bonding showing one crystallographically unique hydrogen bond (dashed yellow bond) from a hydroxy group to a carboxylate oxygen. Dashed blue bonds
are symmetry-equivalent hydrogen bonds.
study because of the difficulty in comparing the energetics of
phases with different stoichiometries by computational methods.
In the calcium system the computational work predicted that
the Ca(m-Tar) phase, 7, would be more stable in total electronic
9
J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009 15383

A R T I C L E S
Appelhans et al.
Figure 9. Diamondoid arrangement of the barium atoms in Ba(D,L-Tar) (a) with and (b) without bridging oxygens, and (c) the diamond unit cell. Nonbridging
oxygens in (a) are excluded for clarity. Each metal center in the diamondoid structure is bridged by two oxygens to each of four adjacent bariums, in a
pseudotetrahedral geometry.
are quite similar, and in all three cases the computational work
correctly identified the most stable phase. This is especially
notable in light of the very small differences in stabilities relative
to the total energies. These results are an excellent demonstration
of the effectiveness of advanced DFT methods37-40 as applied
to complex extended structures, such as metal-organic frame-
works, as well as the accuracy of modern acid solution
calorimetry.
Note that the computational work refers to total electronic
energy,41 while the calorimetric studies determine enthalpies.
Because the volume changes of the transformations are small
and the pressure is low, the difference between total electronic
energy and enthalpy can be neglected. Furthermore, the
calorimetric data refer to room temperature, while the computa-
tions determine total electronic energy at 0 K. Differences in
heat capacities, related to differences in lattice vibrations, will
make these values differ slightly, but the effect is not expected
to be significant. Thermodynamic stability is, of course,
determined by the free energy, which requires knowledge of
the entropy. Therefore, the experimentally observed phase
stabilities may not match the computationally and calori-
metrically determined phase sequence of energies, especially
at high temperatures where the entropic contribution to the free
energy may be significant.
The energetics provide insight into the different phase
behavior of the isostructural Ca and Sr systems. Experimentally
we observe that Sr(L-Tar) (11) converts to Sr(m-Tar) (13) at
200 °C, but Ca(L-Tar) (2) never converts to Ca(m-Tar) (7), even
up to 220 °C. The isostructural meso phases Ca(m-Tar) and
Sr(m-Tar) are both found to be lower in enthalpy/total electronic
energy than the respective isostructural L phases, Ca(L-Tar) and
Sr(L-Tar), by computational and calorimetric methods. Signifi-
cantly, both methods also determine that the difference in
enthalpy/total electronic energy of the meso phase compared
to the L phase is larger for the strontium phases than the calcium
phases. The greater difference in relative stability, providing a
greater thermodynamic driving force for transformation, may
explain why the meso phase is observed as the single thermo-
dynamic product in the strontium system, but not in the calcium
system. The thermodynamic data suggest that in both systems
the meso phases, which are lower in enthalpy/total electronic
energy than the L phases, should be the thermodynamic product
at low temperatures. Furthermore, if the meso phases are higher
in entropy than the L phases, then the data imply that the L
phases are metastable under all conditions. However, conversion
Figure 10. Powder X-ray diffraction patterns of solid solutions of
Sr(1-x)Ba(x)(m-Tar)(H2O) and Sr(1-x)Ba(x)(m-Tar). In the biphasic region, the
+ symbol indicates characteristic peaks of the Ba(m-Tar)(H2O) phase; the
* symbol indicates characteristic peaks of the Sr(m-Tar) phase.
Table 5. Total Electronic Energies, Enthalpies of Solution, and
Relative Stabilities of Phases As Determined by Computational
Methods (CP2K) and Acid Solution Calorimetry
relative stabilities
total
enthalpy of
electronic
solution,
calorimetry
(kJ/mol)
calculated
(kJ/mol)
calorimetry
(kJ/mol)
phase
energy,
calculated
(kJ/mol)
Ca(m-Tar) (7)
Ca(L-Tar) (2)
Sr(m-Tar) (13)
Sr(L-Tar) (11)
Ba(L-Tar) (14)
-415478.0 -2.40 ( 0.11
-415487.1 -5.34 ( 1.63
-399738.2 -1.79 ( 0.16
-399751.5 -9.89 ( 1.40
0
9.1
0
13.4
0
0
2.9 ( 1.6
0
8.1 ( 1.4
0
-385962.1
8.89 ( 0.78
1.84 ( 0.57
Ba(D,L-Tar) (17) -385968.6
6.4
7.0 ( 1.0
energy than the Ca(L-Tar) phase, 2, by 9.1 kJ/mol. The
calorimetric study determined the Ca(m-Tar) phase to be 2.9 (
1.6 kJ/mol more stable in enthalpy than the Ca(L-Tar) phase.
The Sr(m-Tar) phase, 13, was also determined to be more stable
than the Sr(L-Tar) phase, 11, by both methods. Computation
predicted Sr(m-Tar) to be more stable by 13.4 kJ/mol, while
by acid calorimetry the difference was determined to be 8.1 (
1.4 kJ/mol. In the barium system the computational and
calorimetric results also agreed. Both methods determined the
chiral Ba(L-Tar) phase, 14, to be more stable, by 6.4 and 7.0 (
1.0 kJ/mol for computational and calorimetric methods,
respectively.
(40) VandeVondele, J.; Hutter, J. J. Chem. Phys. 2007, 127, 114105.
(41) The zero-point vibrational energies (ZPVEs) are not included in the
calculations. Addition of the ZPVE is not expected to change the
relative stabilities of the calculated phases. See: Rivera, S. A.; Allis,
D. G.; Hudson, B. S. Cryst. Growth Des. 2008, 8, 3905.
A notable result of this study is the excellent agreement
between computational and calorimetric results. Although the
numerical values of the relative stabilities determined by
computational and calorimetric studies are not identical, they
9
15384 J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009

Chiral Alkaline Earth Tartrates and Their Analogues
A R T I C L E S
to the meso phase appears to be controlled by kinetics, perhaps
involving a barrier to isomerization of the tartaric acid. The
kinetic barrier precludes isomerization of the L to meso phases
at low temperatures. In the Sr system the kinetic barrier is
overcome above 180 °C and conversion of Sr(L-Tar) to Sr(m-
Tar) is observed, whereas in the Ca system there is no
conversion below 220 °C. The rate of isomerization may be
affected by many factors, including aqueous solubility and the
nature of the metal ion.
was observed in reaction with L-tartaric acid at temperatures
below 200 °C, even at low concentrations and long reaction
times. Racemization and formation of a Ba(L-Tar)/Ba(D-Tar)
conglomerate was ruled out by circular dichroism (see Sup-
porting Information). Previous work16 has shown that a solution
of sodium tartrate loses ∼50% of optical activity after 24 h of
hydrothermal treatment at 200 °C and suggests that hydrothermal
synthesis with tartaric acid should be limited to 3 days at 160
°C or 1 day at 180 °C. However, our results show that
racemization (and isomerization to or from the meso diastere-
omer) are not consistently temperature dependent. Only in the
strontium system is tartrate isomerization observed below 180
°C. There is no evidence in the barium reactions of isomerization
below 200 °C, even over 5 days, and there is no evidence of
any isomerization in the calcium system up to 220 °C. Likewise,
in previous work on magnesium18 no isomerization of the
tartrate ligand was observed up to 200 °C. Tartrate isomerization
clearly depends highly on other experimental conditions, includ-
ing pH, the identity of the cation, and, possibly, the relative
phase stability of the chiral and racemic or achiral phases. Thus,
while the possibility of isomerization of chiral ligands certainly
must be considered in the analysis of the products, high reaction
temperatures and long reaction times do not necessarily result
in isomerization and loss of chirality.
The interplay of kinetic and thermodynamic factors is evident
in the barium tartrate system. Both the computational and
calorimetric work determined that the Ba(L-Tar) phase, 14, is
more stable in internal energy/enthalpy than the Ba(D,L-Tar)
phase, 17. However, experimentally the Ba(D,L-Tar) phase, 17,
is formed and appears to be the thermodynamically stable phase
at high temperatures. This implies that the Ba(L-Tar) phase is
thermodynamically stable below some transition temperature
and the Ba(D,L-Tar) phase above. The calorimetrically deter-
mined enthalpy difference of 7 kJ/mol implies a transition
temperature of 1215 K. To obtain a transition temperature near
200 °C (473 K) as observed experimentally, one would need
an entropy change of 14.8 J mol-1 K-1
.
Ba(D,L-Tar) also forms at lower temperatures, including room
temperature. On the basis of the calorimetrically determined
stabilities, we would expect the reaction with D,L-tartaric acid
at room temperature to form a racemic conglomerate, a 50:50
mixture of Ba(L-Tar)/Ba(D-Tar). For the Ba(D,L-Tar) phase to
be the thermodynamically stable phase at room temperature,
Conclusions
The synthesis of chiral and achiral calcium, strontium, and
barium tartrates from room temperature to 220 °C was
studied, yielding 18 phases over the temperature range,
including 13 new phases. The high-temperature phase
behavior of each metal system was different. Calcium formed
a unique phase with L-, meso-, and D,L-tartaric acid undergo-
ing no isomerization or phase transitions up to 220 °C. In
contrast, in the reaction of Sr at temperatures above 180 °C,
the Sr(m-Tar) phase was the sole product of the reaction with
all of the tartaric acid starting materials. Likewise, in
reactions with barium, the Ba(D,L-Tar) phase was the only
product at synthesis temperatures above 200 °C. Determina-
tions of the relative energetic stabilities of several phases
by computational and calorimetric methods were in excellent
agreement and led to a more complete understanding of the
phase behavior. Furthermore, there was excellent agreement
between the relative energetic stabilities obtained by com-
putational and calorimetric means. Both the relative ther-
modynamic stabilities of the metal tartrate phases and kinetic
factors, such as ligand racemization and phase crystallization,
were determined to be key factors in controlling the phase
behavior. The occurrence of ligand racemization/isomeriza-
tion at high temperatures and the overall phase stability were
unique for each metal tartrate system, even for the isostruc-
tural calcium and strontium tartrates. The intricate relation-
ships between phase behavior, thermodynamic stabilities, and
kinetic factors emphasize the difficulty in predicting synthetic
outcomes in such systems. Nonetheless, the combination of
synthetic studies, calorimetry, and computation provides a
deeper understanding of these complex systems.
the entropy of transition would have to be 23.5 J mol-1 K-1
,
which is rather large. One possible explanation is that the
reaction with D,L-tartaric acid at room temperature is under
kinetic rather than thermodynamic control. Indeed, Brock and
Dunitz42 suggest that, from a solution of the racemate, the
crystallization of a racemic phase, such as Ba(D,L-Tar), may be
kinetically favored over the crystallization of the conglomerate
of the enantiopure phase. The barium system aptly demonstrates
the importance of energetic, entropic, and kinetic factors in
controlling phase behavior.
Despite the wealth of insight that has been gained by this
multifaceted approach to studying phase behavior, it is worth
noting that it remains difficult to directly relate the structural
details of the phases to the energetics. For example, in the
Ca and Sr phases the C-C-C-C dihedral angle of the
tartrate is quite different between the M(L-Tar) and M(m-
Tar) structures, ∼35° in the M(L-Tar) structures and ∼65°
in the M(m-Tar) structures. However, it is not clear what
role, if any, the difference in C-C-C-C dihedral angles
plays in determining the relative phase stabilities. Likewise,
the characteristics of the metal ion clearly influence the
relative stabilities, as demonstrated by the difference in
relative stabilities in the isostructural Ca and Sr systems.
Whether the differences are due simply to the different ionic
radii or to more subtle structural differences is difficult to
determine. Similarly, though there are major structural
differences between the Ba(L-Tar) and Ba(D,L-Tar) phases,
the energetics of the two phases are similar enough that the
relative stabilities invert over a 200 °C temperature range.
In all three systems (Ca, Sr, Ba) the isomerization of tartaric
acid is likely to influence the phase behavior. In the barium
system no racemization or formation of the Ba(D,L-Tar) phase
Acknowledgment. The authors thank J. VandeVondele (Uni-
versity of Zurich, Switzerland) and M. Krack (Paul Scherrer
Institute, Switzerland) for useful discussions and J. VandeVondele
for providing the basis sets. The computational work was performed
on the supercomputer Mare Nostrum at the Barcelona Supercom-
puting Center-Centro Nacional de Supercomputacio´n. The calo-
rimetric work was supported by NSF grant DMR-0601892. L.A.
(42) Brock, C. P.; Schweizer, W. B.; Dunitz, J. D. J. Am. Chem. Soc. 1991,
113, 9811.
9
J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009 15385

A R T I C L E S
Appelhans et al.
was supported by the Solid State Lighting and Energy Center,
University of California, Santa Barbara. A.K.C. thanks the ERC
for an Advanced Investigator Award. This work made use of MRL
Central Facilities supported by the MRSEC Program of the NSF
under award no. DMR05-20415.
Detailed description of calorimetry methods and additional
calorimetry data on tartaric acid. Additional information on
“disappearing polymorphs” as referenced in the text and on Ca/
Sr and Ba/Sr solid solutions. This material is available free of
charge via the Internet at http://pubs.acs.org. The supplementary
crystallographic data for this paper are also deposited as CCDC
738953-738963. These data can be obtained free of charge from
The Cambridge Crystallographic Data Centre via www.ccdc.
cam.ac.uk/data_request/cif.
Supporting Information Available: Experimental procedures
and characterization of all new compounds (elemental analysis,
TGA, PXRD). Crystallographic data for all new structures,
including CIF files, tables of bond distances and angles, and
structural descriptions of phases not detailed in the text (3, 10,
12, 15, 16, 18). Detailed description of computational methods.
JA905690T
9
15386 J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009