Synthesis and crystal structure of an expanded square grid metal organic material, [Cu(L1)(DMF)]n· (2.64 DMF)

Article abstract of DOI:10.1007/s10870-011-0180-x

A new square grid metal organic material, [Cu(L1)(DMF)]_n· (2.64 DMF) (1; L1= N-(4-carboxyphenyl)-trimellitimide), has been prepared by solvothermal reaction of Cu(NO₃)₂·2.5H ₂O, H₂L1, and caffeine in

Full text of DOI:10.1007/s10870-011-0180-x

J Chem Crystallogr (2011) 41:1834–1838
DOI 10.1007/s10870-011-0180-x
ORIGINAL PAPER
Synthesis and Crystal Structure of an Expanded Square Grid
Metal Organic Material, [Cu(L1)(DMF)]nÁ(2.64 DMF)
•
Patrick Nugent Lukasz Wojtas
•
Michael J. Zaworotko
Received: 5 May 2011 / Accepted: 26 September 2011 / Published online: 5 October 2011
Ó Springer Science+Business Media, LLC 2011
Abstract A new square grid metal organic material,
[Cu(L1)(DMF)]nÁ(2.64 DMF) (1; L1 = N-(4-carboxy-
phenyl)-trimellitimide), has been prepared by solvothermal
reaction of Cu(NO3)2Á2.5H2O, H2L1, and caffeine in N,N-
dimethylformamide (DMF). The title compound crystal-
lizes in the monoclinic crystal system, space group C2/c,
organic linkers have afforded a wide variety of crystalline
polymeric materials that are net-like and can exhibit
porosity coupled with additional functionality attributable
to the nodes and/or linkers [1–8]. These metal organic
materials, MOMs, exhibit properties such as selective gas
sorption [9–13], catalysis [14–16], magnetism [17, 18] and
small molecule separations [14, 19, 20].
˚
with a = 18.5176 (6) A, b = 23.7473 (7) A, c = 18.1998
˚
3
˚
˚
(5) A, b = 107.527 (2)°, and V = 7631.7 (4) A . The
undulating sheets, formed by linking of Cu2(l2-CO2)4
square paddlewheel nodes by the angular L1, stack to form
A noteworthy subset of MOMs is exemplified by those
sustained by dimetal tetracarboxylate ‘‘square paddle-
wheel’’ nodes, a ubiquitous structure in carboxylate coor-
dination chemistry [21–25]. As its name suggests, the
square paddlewheel can be simplified to a square planar
node, which can then be linked to four neighboring nodes
via four dicarboxylate linkers. Square grid or two-dimen-
sional (4,4) networks constructed of square paddlewheel
nodes and dicarboxylate linkers are very well studied and
the dicarboxylate linkers may be linear or angular in nat-
ure. The partially rigid molecules 1,3-benzenedicarbox-
ylate (1,3-bdc) and 1,4-benzenedicarboxylate (1,4-bdc)
represent prototypal linkers that form the square grid motif.
However, in addition to the square grid motif, 1,3-bdc and
1,4-bdc may link paddlewheels to form a number of other
supramolecular isomers [26], all possessing 1:1 metal-to-
linker stoichiometry. Indeed, there have been eight supra-
molecular isomers observed involving 1,3-bdc and four
with 1,4-bdc [27]. The first square grid network sustained
by this variety of linker and node, MOF-2, was reported in
1998 by Yaghi and coworkers [28]. MOF-2 is composed of
divalent zinc paddlewheels and 1,4-bdc linkers and was
one of the first MOMs shown to exhibit permanent
porosity. The crystal structure of a copper-based analogue
with a Langmuir surface area significantly higher than that
of MOF-2 (752 vs. 270 m2/g) was reported in 2009 [29].
In this contribution we detail a new member of this class
of two-dimensional (4,4)-network, [Cu(L1)(DMF)]nÁ(2.64
2
˚
large square channels of approximately 11.9 9 11.9 A
running along the c axis. The geometry and conformation
of the linker result in slight distortion of the grids. Further,
the 147° bend angle between coordinating moieties of L1 is
amenable to forming a diverse array of paddlewheel-based
supramolecular isomers.
Keywords Paddlewheel Á MOM Á Coordination polymer Á
Supramolecular isomer Á Square grid
Introduction
Over the past two decades, judiciously selected combina-
tions of metal ion or metal cluster nodes and multitopic
Electronic supplementary material The online version of this
article (doi:10.1007/s10870-011-0180-x) contains supplementary
material, which is available to authorized users.
P. Nugent Á L. Wojtas Á M. J. Zaworotko (&)
Department of Chemistry, University of South Florida,
CHE 205, 4202 E. Fowler Avenue, Tampa, FL 33620-5205,
USA
e-mail: xtal@usf.edu
URL: http://chemistry.usf.edu/faculty/zaworotko/
123

J Chem Crystallogr (2011) 41:1834–1838
1835
DMF) (1; H2L1 = N-(4-carboxyphenyl)-trimellitimide
[30]). The bent shape and torsional angles present in L1
play an important role in determining the shape of the
resulting grid layers. H2L1 (Fig. 1) can be synthesized in a
solvent-free manner through condensation of the appro-
priate anhydride with the appropriate primary amine. Sol-
vent-free synthesis is an environmentally, financially, and
practically appealing alternative to methods that rely on
toxic, expensive and/or wasteful solvents and has been
used by us to prepare other imide based ligands [31–33].
powder (83.2% yield). Synthesis of H2L1 was supported by
IR spectroscopy, which exhibited absorptions consistent
with the conversion of anhydride to imide (Fig. S3).
[Cu(L1) (DMF)]nÁ(2.64 DMF), 1
15 mg of H2L1, 6.0 mg of Cu(NO3)2Á2.5H2O, and 6.0 mg
of caffeine were dissolved in 0.50 mL of N,N-dimethyl-
formamide and heated at 115 °C for 24 h, affording blue
plate-shaped crystals in ca. 98.1% yield (based on Cu).
Crystal Structure Determination
Experimental
X-ray diffraction data were collected on Bruker-AXS
Materials and Physical Measurements
SMART-APEXII CCD diffractometer using CuKa
(k = 1.54178 A) radiation. Indexing was performed using
˚
All starting materials are commercially available and were
used as received without further purification. The linker
H2L1 has been reported previously [30]. IR data was col-
lected using diffuse reflectance on a Nicolet Avatar Smart
MIRacle 320 FT-IR spectrometer. Powder x-ray diffraction
(PXRD) was conducted on a Bruker D8 ADVANCE, h/2h
APEX2 [34] (difference vectors method) whereas data
integration and reduction were performed using SaintPlus
6.01 [35]. An absorption correction was performed by
multi-scan method implemented in SADABS [36]. Space
groups were determined using XPREP implemented in
APEX2 [34]. The structure was solved using SHELXS-97
(direct methods) and refined using SHELXL-97 (full-
matrix least-squares on F2) contained in APEX2 [34] and
WinGX v1.70.01 [37–40] program packages. All non-
hydrogen, framework atoms were refined anisotropically.
Disordered molecules of DMF were refined using restraints
with isotropic displacement parameters. The contribution
of the remaining disordered, unidentifiable solvent (DMF
or water) was treated as diffuse using Squeeze procedure as
implemented by Platon [41, 42]. Hydrogen atoms attached
to the framework atoms were placed in geometrically
calculated positions and included in the refinement process
using a riding model with isotropic thermal parameters:
Uiso (H) = 1.2 Ueq (CH) and Uiso (H) = 1.5 Ueq (CH3).
Crystal data and refinement parameters are presented in
Table 1.
˚
diffractometer using Cu-Ka radiation (k = 1.54056 A). 2h
scans between 3 and 40° with a step size of 0.02° were
performed for a duration of 92 min and 33 s. Thermo-
gravimetric analysis (TGA) was carried out on a Perkin
Elmer STA 6000 instrument under nitrogen atmosphere.
The as-synthesized sample (14.91 mg) was heated from 28
to 600 °C at a rate of 5 °C/min. TGA (Fig. S1) indicates
that the initial weight loss for 1 corresponds to the removal
of coordinated and uncoordinated DMF (calculated 41.6%,
observed 40.0%). The PXRD pattern of the bulk sample
(Fig. S2) is in good agreement with the pattern calculated
from single-crystal data.
Synthesis
N-(4-carboxyphenyl)-trimellitimide, H2L1 [30]
4-aminobenzoic acid (5.00 mmol, 686 mg) and trimellitic
anhydride (5.00 mmol, 961 mg) were ground with a mortar
and pestle for several minutes to produce a powdered
mixture of the starting materials. This mixture was then
heated at 230 °C for 2.5 h, resulting in a light yellow
Results and Discussion
The title compound crystallizes in the monoclinic space
group C2/c and consists of two-dimensional undulating
square grid layers which stack in a slipped ABCDA fash-
ion. Each layer is composed of 4-connected paddlewheel
nodes that are linked by L1. The asymmetric unit of 1 is
presented in Fig. 2. Each copper atom in 1 assumes a
slightly distorted square pyramidal geometry, with the
carboxylate oxygen atoms of four linkers coordinated at
the basal positions and one DMF oxygen atom occupying
the apical position. The Cu–Cu distance in each paddle-
˚
wheel is 2.632 A. Adjacent paddlewheels are linked by L1
Fig. 1 N-(4-carboxyphenyl)-trimellitimide, H2L1 (created with
ChemDraw Ultra 8.0)
to form a distorted square grid motif with void dimensions
123

1836
J Chem Crystallogr (2011) 41:1834–1838
Table 1 Crystal data and structure refinement summary for 1
Empirical formula
Formula weight
Temperature (K)
C19H14CuN2O7Á2.64(DMF)
639.95
100 (2)
˚
Wavelength (A)
1.54178
Crystal system
Space group
Monoclinic
C2/c
˚
a (A)
18.5176 (6)
23.7473 (7)
18.1998 (5)
107.527 (2)
7631.7 (4)
8
˚
b (A)
˚
c (A)
b (°)
Volume (A )
3
˚
Z
Calculated density (g/cm3)
Absorption coefficient (mm-1
F(000)
1.114
)
1.222
Fig. 3 The distorted square grid motif in 1 (created with Materials
Studio (MS Modeling) version 4.0)
2666
Crystal size (mm3)
0.10 9 0.08 9 0.06
3.12–65.87
(-20/21, -27/27, -21/21)
25165/6503 [Rint = 0.1062]
97.9%
2
2
10.3 9 10.3 A and 10.8 9 10.8 A , respectively while an
h range for data collection (°)
Limiting indices (h, k, l)
Reflections collected/unique
Completeness to h = 65.87°
Absorption correction
˚
˚
analogous 1,3-bdc structure [43] contains alternating cav-
2
ities of 10.1 9 10.4 A (ignoring apical pyridine ligands)
˚
2
and 6.4 9 6.4 A .
˚
Semi-empirical from
equivalents
Caffeine was included in the synthesis of 1 because its
structure and molecular recognition features make it rep-
resentative of natural products and pharmaceutically rele-
vant compounds. In addition, caffeine is readily identifiable
by a variety of instrumental methods including IR and UV/
Vis spectroscopy, gas chromatography, and x-ray crystal-
lography. The study of caffeine sorption from solution by
MOMs therefore serves as a benchmark and will aid in
developing an understanding of specific host–guest inter-
actions for the design of new MOM sorbents.
Max and min transmission
Refinement method
0.9303 and 0.8876
Full-matrix least-squares on F2
Data/restraints/parameters
Goodness-of-fit on F2
Final R indices [I [ 2r (I)]
R indices (all data)
6503/34/371
1.023
R1 = 0.0775, wR2 = 0.2084
R1 = 0.1053, wR2 = 0.2261
Largest different peak and hole (e/ 0.750 and -0.552
3
˚
A )
The distorted hourglass shape of the cavities in the grids
and the undulating nature of the grids in 1 can be attributed
to the geometry of L1. If L1 is simplified to two vectors
that extend from each carboxylate carbon atom and meet at
the centroid of the phenyl ring which is fused to the imide
ring, the approximate angle subtended between the vectors
is 147°. The torsional angles present in L1 also play an
important role in defining the structure of 1. While the
torsional angle between the carboxylate group attached to
the N-phenyl ring and the phenyl ring itself is nearly zero,
that between the other carboxylate of L1 and the phenyl
ring of the bicyclic portion to which it is attached is 19.8°.
Also, the five-membered imide ring and the N-phenyl ring
of L1 are oriented at a torsional angle of 43.0–53.7° with
respect to one another. This angle cannot be precisely
measured because the imide nitrogen atom is not ideally
trigonal planar. The geometry and conformation of L1
result in adjacent paddlewheels being inclined at an angle
of 23.8° relative to one another.
Fig. 2 ORTEP view of the asymmetric unit of [Cu(L1)(DMF)]nÁ
(2.64 DMF), 1, with solvent omitted and ellipsoids at 30% probability
level (created with Ortep-3, version 2.2)
2
of ca. 18.8 9 12.3 A (excluding van der Waals radii) as
˚
depicted in Fig. 3. The void space within the squares of 1 is
substantially greater than that in prototypal square grids
containing 1,4- and 1,3-bdc. For comparison, MOF-2
and its copper analogue possess void dimensions of
As mentioned earlier, the sheets stack in an ABCDA
arrangement in the c direction. The interlayer separation, as
123

J Chem Crystallogr (2011) 41:1834–1838
1837
calculated by dividing the distance between paddlewheels
˚
in nearest ‘A’ layers by four, is 7.40 A. The distance
˚
between nearest copper atoms in adjacent layers is 7.59 A,
while the methyl groups of apically-coordinated DMF
molecules extend nearly into the square voids of the layers
above and below. Stacking of the layers results in square
channels (Fig. 4) with dimensions of approximately
2
˚
11.9 9 11.9 A (not taking into account van der Waals
2
radii) extending along the c axis (8.7 9 8.7 A with van
˚
der Waals). Free DMF molecules occupy the void space of
1, which exhibits a solvent-accessible volume of 60.1% as
determined by Platon [42].
An analysis of 11 square grid structures [44, 45] formed
by the square paddlewheel and 1,4-bdc linkers reveals that
the number of sheets before repetition of the stacking
arrangement varies from 2 to 8 (mean = 4.4), with inter-
Fig. 4 View of the square channels running along the c axis in 1
(created with Materials Studio (MS Modeling) version 4.0)
˚
˚
layer distances ranging from 5.1 to 9.4 A (mean = 6.6 A)
[46]. It is apparent upon examination of these structures
that, in general, an inverse relationship exists between
stacking number and interlayer distance. In other words, a
higher stacking number is consistent with more efficient
packing between layers. Meanwhile, examination of 14
undulating grids containing 1,3-bdc linkers [47] reveals
that the stacking number varies from 1 to 4 (mean = 2.8)
dicarboxylate linker, L1. The linker, which can be syn-
thesized in a solvent-free manner, creates a distortion of the
grid motif due to its distinct geometry and conformation.
The grids themselves have significantly expanded dimen-
sions in comparison with those of prototypal square grids
based upon 1,3- or 1,4-bdc. Stacking of the layers in 1
generates square channels which have the potential to
accommodate a wide variety of guest species. Finally, we
note that the intermediate bend angle of L1 may endow it
with the potential to form paddlewheel-based supramo-
lecular isomers that are accessible by either of the proto-
typal linkers.
˚
while interlayer distances range from 7.7 to 10.6 A
˚
(mean = 9.0 A). Judging by mean interlayer distance,
1,4-bdc grids pack more efficiently than their 1,3-bdc
counterparts. This observation is perhaps related to the
undulating nature of 1,3-bdc grids and/or the nature of the
axial ligands. The majority of known 1,3-bdc grids possess
bulkier pyridyl-type axial ligands while 1,4-bdc struc-
tures contain smaller axial ligands like DMF, water, and
methanol.
Supplementary Data
As mentioned previously, there are eight supramolecular
isomers (i.e. topologies) that have been observed when 1,3-
bdc links square paddlewheels and four when 1,4-bdc
serves as the linker. In the case of 1,3-bdc, possible
Supplementary crystallographic data for [Cu(L1)(DMF)]nÁ
(2.64 DMF) is contained in CCDC 823167. Copies of this
information may be obtained free of charge via http://www.
ccdc.cam.ac.uk/data_request/cif (or from the Cambridge
Crystallographic Data Centre (CCDC), 12 Union Road,
Cambridge CB2 1EZ, UK; fax: ?44(0)1233-336033;
e-mail: deposit@ccdc.cam.ac.uk).
´
topologies include kagome [48], square grid [43, 49–52],
nanoball-1 [53, 54], nanoball-2 [54], usf [55], CdSO4 [55],
mot [50], and NbO. In the case of 1,4-bdc, the known
´
topologies are kagome [56], square grid [28], CdSO4 [50],
and NbO [57]. Linker geometry in these isomers can be
quantified by three parameters: the bending angle between
carboxylate moieties (1,3-bdc = 120°, 1,4-bdc = 180°),
the twisting angle between carboxylate planes, and the
bending angle between carboxylate planes and the benzene
plane [50]. The 147° bend angle between carboxylates in
L1 lies roughly midway between that in 1,3- and 1,4-bdc,
meaning that L1 may have the potential to form isomers
that have been observed with either of the prototypal
linkers.
Acknowledgment This work was generously supported by the
Department of Defense—Defense Threat Reduction Agency
(DoD-DTRA) through HDTRA1-08-C-0035.
References
1. Kitagawa S, Kitaura R, Noro S (2004) Angew Chem Int Ed Engl
43:2334–2375
2. Janiak C (2003) Dalton Trans :2781–2804
3. Eddaoudi M, Moler DB, Li HL, Chen BL, Reineke TM, O’Keeffe
M, Yaghi OM (2001) Acc Chem Res 34:319–330
In summary, we have prepared a MOM, 1, which is
sustained by copper (II) paddlewheel nodes and an angular
123

1838
J Chem Crystallogr (2011) 41:1834–1838
4. Ferey G (2008) Chem Soc Rev 37:191–214
5. Yaghi OM, Li HL, Davis C, Richardson D, Groy TL (1998) Acc
Chem Res 31:474–484
6. Rowsell JLC, Yaghi OM (2004) Microporous Mesoporous Mater
73:3–14
31. Cheney ML, McManus GJ, Perman JA, Wang ZQ, Zaworotko MJ
(2007) Cryst Growth Des 7:616–617
32. Perman JA, Dubois K, Nouar F, Zoccali S, Wojtas L, Eddaoudi
M, Larsen RW, Zaworotko MJ (2009) Cryst Growth Des
9:5021–5023
7. Batten SR, Neville SM, Turner DR (2009) Coordination poly-
mers: design, analysis and application. Royal Society of Chem-
istry, Cambridge
8. MacGillivray LR (2010) Metal-organic frameworks: design and
application. John Wiley & Sons, Hoboken
33. Perman JA, Cairns AJ, Wojtas L, Eddaoudi M, Zaworotko MJ
(2011) CrystEngComm 13:3130–3133
34. Bruker (2010) APEX2. Bruker AXS Inc, Madison
35. Bruker (2009) SAINT data reduction software. Bruker AXS Inc,
Madison
9. Eddaoudi M, Kim J, Rosi N, Vodak D, Wachter J, O’Keeffe M,
Yaghi OM (2002) Science 295:469–472
36. Sheldrick G (2008) SADABS. Program for empirical absorption
correction. University of Gottingen, Germany
10. Rosi NL, Eckert J, Eddaoudi M, Vodak DT, Kim J, O’Keeffe M,
Yaghi OM (2003) Science 300:1127–1129
11. Rowsell JLC, Yaghi OM (2005) Angew Chem Int Ed Engl
44:4670–4679
12. Murray LJ, Dinca M, Long JR (2009) Chem Soc Rev 38:
1294–1314
13. Li JR, Kuppler RJ, Zhou HC (2009) Chem Soc Rev 38:
1477–1504
37. Sheldrick G (1997) SHELXL-97. Program for the refinement of
crystal structures. University of Gottingen, Germany
38. Sheldrick G (2008) Acta Crystallogr Sect A 64:112–122
39. Sheldrick G (1990) Acta Crystallogr Sect A 46:467–473
40. Farrugia L (1999) J Appl Crystallogr 32:837–838
41. Vandersluis P, Spek AL (1990) Acta Crystallogr Sect A 46:
194–201
42. Spek AL (1990) Acta Crystallogr Sect A 46:c34
43. Bourne SA, Lu J, Mondal A, Moulton B, Zaworotko MJ (2001)
Angew Chem Int Ed Engl 40:2111–2113
14. Seo JS, Whang D, Lee H, Jun SI, Oh J, Jeon YJ, Kim K (2000)
Nature 404:982–986
15. Wu CD, Hu A, Zhang L, Lin WB (2005) J Am Chem Soc
127:8940–8941
16. Cho SH, Ma BQ, Nguyen ST, Hupp JT, Albrecht-Schmitt TE
(2006) Chem Commun :2563–2565
17. Halder GJ, Kepert CJ, Moubaraki B, Murray KS, Cashion JD
(2002) Science 298:1762–1765
18. Batten SR, Murray KS (2003) Coord Chem Rev 246:103–130
19. Pan L, Olson DH, Ciemnolonski LR, Heddy R, Li J (2006)
Angew Chem Int Ed Engl 45:616–619
20. Chae HK, Siberio-Perez DY, Kim J, Go Y, Eddaoudi M, Matzger
AJ, O’Keeffe M, Yaghi OM (2004) Nature 427:523–527
21. Chui SSY, Lo SMF, Charmant JPH, Orpen AG, Williams ID
(1999) Science 283:1148–1150
44. Bruno IJ, Cole JC, Edgington PR, Kessler M, Macrae CF,
McCabe P, Pearson J, Taylor R (2002) Acta Crystallogr Sect B
58:389–397
45. CSD refcodes: ACASUT, AFOSIY, AGAWOU, AGAWUA,
DAXNOG, GECXUH, HIQVUZ, LOWPUI, POHPOS, SATFEZ
46. Layers were considered equivalent if they aligned with one
another when viewed along an axis perpendicular to the layer
plane
47. CSD refcodes: DIKFAF, EBUCAF, EBUCEJ, KAHHIL, KAH-
HOR, KAHHUX, ODUREL, ODURIP, PACVOF, PACWEW,
PACWIA, PACWOG, PACXAT, TOPKOZ
48. Perry JJ, McManus GJ, Zaworotko MJ (2004) Chem Commun
:2534–2535
22. Nouar F, Eubank JF, Bousquet T, Wojtas L, Zaworotko MJ,
Eddaoudi M (2008) J Am Chem Soc 130:1833–1835
23. Wong-Foy AG, Lebel O, Matzger AJ (2007) J Am Chem Soc
129:15740–15741
24. Xiang SC, Zhou W, Gallegos JM, Liu Y, Chen BL (2009) J Am
Chem Soc 131:12415–12419
49. Abourahma H, Bodwell GJ, Lu JJ, Moulton B, Pottie IR, Walsh
RB, Zaworotko MJ (2003) Cryst Growth Des 3:513–519
50. Eddaoudi M, Kim J, Vodak D, Sudik A, Wachter J, O’Keeffe M,
Yaghi OM (2002) Proc Natl Acad Sci USA 99:4900–4904
51. Gao L, Zhao BJ, Li GH, Shi Z, Feng SH (2003) Inorg Chem
Commun 6:1249–1251
25. Xiao B, Wheatley PS, Zhao XB, Fletcher AJ, Fox S, Rossi AG,
Megson IL, Bordiga S, Regli L, Thomas KM, Morris RE (2007)
J Am Chem Soc 129:1203–1209
26. Moulton B, Zaworotko MJ (2001) Chem Rev 101:1629–1658
27. Zhang Z, Wojtas L, Zaworotko MJ (2011) Cryst Growth Des
11:1441–1445
52. Xue D-X, Lin Y-Y, Cheng X-N, Chen X-M (2007) Cryst Growth
Des 7:1332–1336
53. Eddaoudi M, Kim J, Wachter JB, Chae HK, O’Keeffe M, Yaghi
OM (2001) J Am Chem Soc 123:4368–4369
54. Moulton B, Lu JJ, Mondal A, Zaworotko MJ (2001) Chem
Commun :863–864
28. Li H, Eddaoudi M, Groy TL, Yaghi OM (1998) J Am Chem Soc
120:8571–8572
29. Carson CG, Hardcastle K, Schwartz J, Liu X, Hoffmann C,
55. Moulton B, Abourahma H, Bradner MW, Lu JJ, McManus GJ,
Zaworotko MJ (2003) Chem Commun :1342–1343
56. Kondo M, Takashima Y, Seo J, Kitagawa S, Furukawa S (2010)
CrystEngComm 12:2350–2353
57. Eddaoudi M, Kim J, O’Keeffe M, Yaghi OM (2002) J Am Chem
Soc 124:376–377
Gerhardt RA, Tannenbaum
2009:2338–2343
30. Van Strien RE, Eiszner JR (1965) US Patent 3217014
R (2009) Eur J Inorg Chem
123