Anomalous non-prussian blue structures and magnetic ordering of K 2MnII[MnII(CN)6] and Rb 2MnII[MnII(CN)6]

Article abstract of DOI:10.1021/ic901903f

The reaction of Mn^II and KCN in aqueous and non-aqueous media leads to the isolation of three-dimensional (3-D) Prussian blue analogues, K₂Mn[Mn(CN)₆] (1a-d, 1e, respectively). Use of RbCN forms Rb₂Mn[Mn(CN)₆] (2). 1 and 2 are isomorphic {monoclinic, P2₁/n: 1 [a = 10.1786(1) A, b = 7.4124(1) A, c = 6.9758(1) A, β = 90.206(1)°]; 2 [a = 10.4101(1) A, b = 7.4492(1) A, c = 7.2132(1) A, β = 90.072(1)°]}, with a small monoclinic distortion from the face centered cubic (fcc) structure that is typical of Prussian blue structured materials that was previously reported for K2Mn[Mn(CN)₆]. Most notably the average Mn-N-C angles are 148.8° and 153.3° for 1 and 2, respectively, which are significantly reduced from linearity. This is attributed to the ionic nature of high spin Mn^II accommodating a reduced M-CN-M0 angle and minimizing void space. Compounds 1a, b have a sharp, strong.OH band at 3628 cm^-1, while 1e lacks a v _OH absorption. The v_OH absorption in 1a, b is attributed to surface water, as use of D₂O shifts the.OH absorption to 2677 cm^-1, and that 1a-e are isostructural. Also, fcc Prussian blue-structured Cs₂Mn[Mn(CN)₆] (3) has been structurally [Fm3hm: a = 10.6061(1) A] and magnetically characterized. The magnetic ordering temperature, Tc, increases as K⁺ (41 K) < Rb⁺ (34.6 K) < Cs⁺ (21 K) for A₂Mn[Mn(CN)₆] in accord with the increasing deviation for linearity of the Mn-N-C linkages [148.8 (K⁺) < 153.3 (Rb+) < 180° (Cs⁺)], decreasing Mn(II) · · · Mn(II) separations [5.09 (K+) > 5.19 (Rb+) > 5.30 A° (Cs+)], and decreasing size of the cation (increasing electrostatic interactions). Hence, the bent cyanide bridges play a crucial role in the superexchange mechanism by increasing the coupling via shorter Mn(II) 3 3 3 Mn(II) separations, and perhaps enhanced overlap. In addition, the temperature dependent magnetic behavior of K₄[Mn^II(CN) ₆] · 3H₂O is reported.

Full text of DOI:10.1021/ic901903f

1524 Inorg. Chem. 2010, 49, 1524–1534
DOI: 10.1021/ic901903f
Anomalous Non-Prussian Blue Structures and Magnetic Ordering of
K2MnII[MnII(CN)6] and Rb2MnII[MnII(CN)6]
Jae-Hyuk Her,‡,§ Peter W. Stephens,*,‡ Christopher M. Kareis,† Joshua G. Moore,† Kil Sik Min,†, Jong-Won Park,†
Garima Bali,† Bretni S. Kennon,† and Joel S. Miller*,†
†Department of Chemistry, 315 S. 1400 E. RM 2124, University of Utah, Salt Lake City, Utah 84112-0850 and
‡Department of Physics & Astronomy, Stony Brook University, Stony Brook, New York 11794-3800. §Current
address: GE Global Research, 1 Research Circle, Niskayuna, NY 12309. Current address: Department of
Chemistry Education, Kyungpook National University, Daegu, South Korea.
Received September 29, 2009
The reaction of MnII and KCN in aqueous and non-aqueous media leads to the isolation of three-dimensional (3-D)
Prussian blue analogues, K2Mn[Mn(CN)6] (1a-d, 1e, respectively). Use of RbCN forms Rb2Mn[Mn(CN)6] (2). 1 and
o
˚
˚
˚
2 are isomorphic {monoclinic, P21/n: 1 [a = 10.1786(1) A, b = 7.4124(1) A, c = 6.9758(1) A, β = 90.206(1) ]; 2
o
˚
˚
˚
[a = 10.4101(1) A, b = 7.4492(1) A, c = 7.2132(1) A, β = 90.072(1) ]}, with a small monoclinic distortion from the face
centered cubic (fcc) structure that is typical of Prussian blue structured materials that was previously reported for
K2Mn[Mn(CN)6]. Most notably the average Mn-N-C angles are 148.8° and 153.3° for 1 and 2, respectively, which
are significantly reduced from linearity. This is attributed to the ionic nature of high spin MnII accommodating a reduced
M-CN-M0 angle and minimizing void space. Compounds 1a,b have a sharp, strong νOH band at 3628 cm-1, while 1e
lacks a νOH absorption. The νOH absorption in 1a,b is attributed to surface water, as use of D2O shifts the νOH
absorption to 2677 cm-1, and that 1a-e are isostructural. Also, fcc Prussian blue-structured Cs2Mn[Mn(CN)6] (3) has
˚
been structurally [Fm3hm: a = 10.6061(1) A] and magnetically characterized. The magnetic ordering temperature, Tc,
increases as Kþ (41 K) > Rbþ (34.6 K) > Csþ (21 K) for A2Mn[Mn(CN)6] in accord with the increasing deviation for
linearity of the Mn-N-C linkages [148.8 (Kþ) > 153.3 (Rbþ) > 180° (Csþ)], decreasing Mn(II) Mn(II)
3 3 3
þ
þ
þ
˚
separations [5.09 (K ) < 5.19 (Rb ) < 5.30 A (Cs )], and decreasing size of the cation (increasing electrostatic
interactions). Hence, the bent cyanide bridges play a crucial role in the superexchange mechanism by increasing the
coupling via shorter Mn(II) Mn(II) separations, and perhaps enhanced overlap. In addition, the temperature
3 3 3
dependent magnetic behavior of K4[MnII(CN)6] 3H2O is reported.
3
Introduction
Prussian blue structure.1-3 Critical temperatures, Tc, as
great as ∼100 °C have been reported for KVII[CrIII-
Ferri- and to a lesser extent ferromagnetic ordering has
been reported for several hexacyanometalates possessing the
III
(CN)6] 2H2O 0.1(KO3SCF3)3a and K0.058VII
V
-
0.43
0.57
3
3
[CrIII(CN)6]0.79 (SO4)0.058 0.93H2O3b with the anticipation
3
3
that they may find applications in areas such as magnetic
shielding4 and spintronic memory storage devices.5 Magnetic
ordering occurs in these materials because of the strong spin
coupling between adjacent metal sites via the superexchange
interaction6 of the cyanide ligand. This family includes
*To whom correspondence should be addressed. E-mail: jsmiller@
chem.utah.edu (J.S.M.), peter.stephens@sunysb.edu (P.W.S.).
(1) Verdaguer, M.; Girolami, G. S. In Magnetism - Molecules to Materials;
Miller, J. S., Drillon, M., Eds.; Wiley-VCH: Weinheim, 2005, Vol. 5, p 283.
Hashimoto, K.; Ohkoshi, S. Phil. Trans. R. Soc. Lond. A 1999, 357, 2977.
Verdaguer, M.; Bleuzen, A.; Marvaud, V.; Vaissermann, J.; Seuleiman, M.;
Desplanches, C.; Scuiller, A.; Train, C.; Garde, R.; Gelly, G.; Lomenech, C.;
Rosenman, I.; Veillet, P.; Cartier, C.; Villain, F. Coord. Chem. Rev. 1999,
190-192, 1023.
MnIII[MnIII(CN)6] (Tc = 31 K),7 MnIII3[MnII(CN)6]2 zH2O
3
(4) (a) Morin, B. G.; Hahm, C.; Epstein, A. J.; Miller, J. S. J. Appl. Phys.
1994, 75, 5782. (b) Miller, J. S. Adv. Mater. 1994, 6, 322. (c) Landee, C. P.;
Melville, D.; Miller, J. S. In NATO ARW Molecular Magnetic Materials;
Kahn, O., Gatteschi, D., Miller, J. S., Palacio, F., Eds.; Kluwer Academic
Publishers: London, 1991, Vol. E198, p 395.
(2) (a) Ferlay, S.; Mallah, T.; Ouahes, R.; Veillet, P.; Verdaguer, M.
Nature 1995, 378, 701. (b) Dujardin, E.; Ferlay, S.; Phan, X.; Desplanches, C.;
Moulin, C. C. D.; Sainctavit, P.; Baudelet, F.; Dartyge, E.; Veillet, P.; Verdaguer,
M. J. Am. Chem. Soc. 1998, 120, 11347. (c) Ferlay, S.; Mallah, T.; Ouahes, R.;
Veillet, P.; Verdaguer, M. Inorg. Chem. 1999, 38, 229. (d) Verdaguer, M.;
Bleuzen, A.; Train, C.; Garde, R.; Fabrizi de Biani, F.; Desplanches, C. Phil.
Trans. R. Soc. Lond. A 1999, 357, 2959.
(5) Prigodin, V. N.; Raju, N. P.; Pokhodnya, K. I.; Miller, J. S.; Epstein,
A. J. Adv. Mater. 2002, 14, 1230.
(6) (a) Anderson, P. W. Phys. Rev. 1950, 79, 350. (b) Goodenough, J. B.
J. Phys. Chem. Solids 1958, 6, 287. (c) Kanamori, J. J. Phys. Chem. Solids 1959,
10, 87.
(3) (a) Holmes, S. M.; Girolami, G. S. J. Am. Chem. Soc. 1999, 121, 5593.
(b) Hatlevik, O.; Buschmann, W. E.; Zhang, J.; Manson, J. L.; Miller, J. S. Adv.
e
Mater. 1999, 11, 914.
(7) Buschmann, W. E.; Miller, J. S. Inorg. Chem. 2000, 39, 2411.
r
pubs.acs.org/IC
Published on Web 01/19/2010
2010 American Chemical Society

Article
Inorganic Chemistry, Vol. 49, No. 4, 2010 1525
(12 e z e 16; Tc = 35.5 K),8,9 CsMnII[MnIII(CN)6] 1/
received.20 K4[Mn(CN)6]21a and CsCN21b were prepared via
literature routes. Diethyl ether was purified through an
activated alumina dual-column purification system under a
positive pressure of N2, while H2O was purified through a
Barnstead “E-pure” water purification system and deoxyge-
nated through distillation under N2. All other solvents were
distilled from the appropriate drying agents under nitrogen
before use. All syntheses were performed in an oxygen-free
(<1.0 ppm O2) wet or drybox.
3
2(H2O) (Tc = 31 K),8 and MnII[MnIV(CN)6] 1.14H2O
3
(Tc = 49 K).7,10 In the case of MnII3[MnIII(CN)6]2 12H2O,
3
the C-bound low spin (ls) MnIII (t2g4, S = 1) interacts with
2
the N-bound high spin (hs) MnII (t2g3eg , S = 5/2) ferrimag-
netically.8,9 Additionally, face-centered cubic (fcc) K2MnII-
[MnII(CN)6]8,11a and NaMn[Cr(CN)6]12 were reported to be
ferrimagnets with Tcs of 41 and 60 K, respectively.
Recently, MII[MnIV(CN)6] (M = Co, Ni) and MIII[MnIII-
(CN)6] (M = V, Cr, Mn) made in non-aqueous media showed
unexpected magnetic behaviors.7 The formation of the water-
free lattice led to spin-glass behavior that is atypical of
Prussian blue analogues (PBAs).1-3 Also, the study of
anhydrous Fe[Mn(CN)6] and its hydrate revealed that they
have uncommon mixed valency and magnetic properties.13
Keggin and Miles originally reported the structure of
AnFe[Fe(CN)6] (A = alkali cation) as being fcc, and this
has been recognized as the general structure for AnM[M0-
Physical Methods. Infrared spectra were recorded from 400 to
4000 cm-1 on a Bruker Tensor 37 spectrometer ((1 cm-1) as
KBr pellets. Thermogravimetric analyses were performed at a
scan rate of 5 °C/min using a TGA 2050 TA Instruments located
in a Vacuum Atmospheres DriLab under nitrogen to protect air-
and moisture-sensitive samples. Samples were placed in
an aluminum pan and heated at 5 °C/min under a continuous
10 mL/min nitrogen flow. Elemental analyses were performed
by GCL & Chemisar Laboratories.
Magnetic susceptibilities were measured in 1000 (1a-d and 2)
and 500 Oe (1e and 3) applied fields between 5 and 300 K on a
Quantum Design MPMS superconducting quantum interfer-
ence device (SQUID) equipped with a reciprocating sample
measurement system, low field option, and continuous low
temperature control with enhanced thermometry features, as
previously described.22 Powder samples for magnetic measure-
ments were loaded in gelatin capsules. The direct current (DC)
magnetization temperature dependence was obtained by cool-
ing in zero-field and collecting the data on warming. The
remanent magnetization was taken in zero applied field upon
warming after cooling in a 5 Oe field. Alternating current (AC)
susceptibilities were measured at 10, 100, and 1000 Hz. In
addition to correcting for the diamagnetic contribution from
the sample holder, the core diamagnetic corrections of -136 (1),
-146 (2), and -176 ꢀ 10-6 emu/mol (3) were used.
(CN)6] zH2O PBAs.14 PBAs have two distinct metal sites
3
consisting of C bonded to strong field metal sites, M0C6, and
N bonded to weak field metal sites, MN6, and charge
balancing cations and/or water molecules are located within
the void spaces of each lattice. Often PBAs have incomplete
lattices with water coordinated to M rather than nitrogen.15,16
Herein, we investigate the magnetic properties of K2MnII-
[MnII(CN)6] (1a-e) and Rb2MnII[MnII(CN)6] (2), and report
the first non-face centered cubic structure of A2Mn[Mn-
(CN)6] (A = K, Rb) composition with bent Mn-CN-Mn
linkages. For comparison the magnetic behavior of K4[MnII-
(CN)6] is revisited.17 We also extend the fcc Prussian blue
structure type with the preparation and structural and
magnetic characterization of Cs2MnII[MnII(CN)6] (3). The
genesis of this study was to identify a route to make
[NEt4]2[MnII(CN)4],18 as well as to prepare [NEt4]4[MnII-
(CN)6], and these results will be reported later.19
Powder X-ray diffraction (XRPD) measurements for Rietveld
structure analysis were performed at Beamline X16C of the
National Synchrotron Light Source at Brookhaven National
Laboratory. The powdered samples were held in a 1.0 mm
diameter thin-wall quartz capillary. X-rays of wavelength
˚
0.70025(1), 0.69731(1), and 0.70051(2) A were selected by a
Experimental Section
Si(111) channel cut for K2Mn[Mn(CN)6], Rb2Mn[Mn(CN)6],
and Cs2Mn[Mn(CN)6], respectively. Diffracted X-rays were
selected by a Ge(111) analyzer and detected by a scintillation
counter. The incident intensity was monitored by an ion cham-
ber and used to normalize the measured signal. The TOPAS-
Academic program was used to index, solve, and refine the
crystal structures.23-25 Additional XRPD patterns were col-
lected on all samples with a Bruker D8 Diffractometer (Cu KR)
using Mica (NIST Standard Reference Material 675) as an
internal standard.
KCN, Mn(O2CMe)2 4H2O, and MnCl2 4H2O were used
3
3
as purchased. Mn(O2CMe)2 4H2O was dehydrated by heat-
3
ing in a vacuum oven at 100 °C over P2O5 for 24 h, and after
putting Mn(O2CMe)2 xH2O into the DryBox, x was deter-
mined to be 2 from a thermogravimetric analysis (TGA).
RbCNwasobtainedasagiftfromM.C. DeLong, andusedas
3
(8) (a) Entley, W. R.; Girolami, S. G. Inorg. Chem. 1994, 33, 5165.
(b) Entley, W. R.; Girolami, S. G. Inorg. Chem. 1995, 34, 2262.
(9) Franz, P.; Ambrus, C.; Hauser, A.; Chernyshov, D.; Hostettler, M.;
K2MnII[MnII(CN)6]. Method A (1a).11 To a ∼2 mL aqueous
€
€
Hauser, J.; Keller, L.; Kramer, K.; Stoeckli-Evans, H.; Pattison, P.; Burgi,
H.-B.; Decurtins, S. J. Am. Chem. Soc. 2004, 126, 16472.
(10) Klenze, R.; Kanellakopoulos, B.; Trageser, G.; Eysel, H. H. J. Chem.
Phys. 1980, 72, 5819.
solution of MnCl2 4H2O (100 mg, 0.505 mmol) was added a ∼2mL
3
aqueous solution of KCN (100 mg, 1.54 mmol). A gray pre-
cipitate immediately formed turning yellow then green within
5 min. After 3 h of stirring, the green powder was collected by
filtration, washed with water, ethanol, acetone, and diethyl
ether, and dried under vacuum at room temperature for 12 h
(Yield: 70 mg, 80%). IR (KBr), υOH 3628 (m), υCN 2057 (s), 2023
(sh) cm-1. Calcd for C6K2Mn2N6: C, 20.94; H, 0.00; N, 24.42;
obs C, 20.61; H, <0.20; N, 24.30.
(11) Qureshi, A. M.; Sharpe, A. G. J. Inorg. Nucl. Chem. 1968, 30, 2269.
(12) Dong, W.; Zhu, L.-N.; Song, H.-B.; Liao, D.-Z.; Jiang, Z.-H.; Yan,
S.-P.; Cheng, P.; Gao, S. Inorg. Chem. 2004, 43, 2465.
€
(13) Buschmann, W. E.; Ensling, J.; Gutlich, P.; Miller, J. S. Chem.;Eur.
J. 1999, 5, 3019.
(14) (a) Keggin, J. F.; Miles, F. D. Nature 1936, 137, 577. (b) Itaya, K.;
Uchida, I.; Neff, V. D. Acc. Chem. Res. 1986, 19, 162.
(15) Beall, G. W.; Milligan, W. O.; Korp, J.; Bernal, I. Inorg. Chem. 1977,
16, 2715. Mullica, D. F.; Zielke, J. T.; Sappenfield, E. L. J. Sol. State. Chem.
1994, 112, 92.
(21) (a) Figgis, B. N. Trans. Faraday Soc. 1964, 57, 298. (b) Markley, T. J.;
Toby, B. H.; Pearlstein, R. M.; Ramprasad, D. Inorg. Chem. 1997, 36, 3376.
(22) Brandon, E. J.; Rittenberg, D. K.; Arif, A. M.; Miller, J. S. Inorg.
Chem. 1998, 37, 3376.
€
(16) Herren, F.; Fisher, P.; Ludi, A.; Halg, W. Inorg. Chem. 1980, 19, 956.
(17) Kotani, M. J. Phys. Soc. Jpn. 1949, 4, 293.
(18) Buschmann, W. E.; Arif, A. M.; Miller, J. S. Angew. Chem., Int. Ed.
1998, 37, 781. Angew. Chem. 1998, 110, 813.
(23) TOPAS V3: General profile and structure analysis software for powder
diffraction data. - User’s Manual; Bruker AXS: Karlsruhe, Germany, 2005.
(24) Coelho, A. A. J. Appl. Crystallogr. 2000, 33, 899.
(19) Her, J.-H.; Stephens. P. W.; Kareis, C. M.; Moore, J. G.; Min, K. S.;
Park, J.-W.; Miller, J. S., in preparation.
(20) DeLong, M. C.; Rosenberger, F. J. Cryst. Growth 1986, 75, 164.
(25) TOPAS-Academic is available at http://pws.prserv.net/Alan.Coelho.

1526 Inorganic Chemistry, Vol. 49, No. 4, 2010
Her et al.
Method B (1b). To a ∼2 mL, colorless aqueous solution of
Mn(O2CMe)2 4H2O (100 mg, 0.408 mmol) was added a ∼2 mL
3
aqueous solution of KCN (80 mg, 1.22 mmol). The color change
of the precipitate was the same as observed for 1a. The slurry
was stirred for 2 h, and the green powder was collected by
filtration, washed with water, ethanol, acetone, and diethyl
ether, and dried under vacuum at room temperature for 12 h
(Yield: 48 mg, 69%). IR (KBr), υOH 3628 (m), υCN 2057 (s), 2023
(sh) cm-1. Calcd for C6K2Mn2N6: C, 20.94; H, 0.00; N, 24.42;
obs C, 20.79; H, <0.20; N, 23.96.
Method C (1c). To a 5 mL, colorless aqueous solution of
MnCl2 4H2O (250 mg, 1.26 mmol) was added a 5 mL aqueous
3
solution of KCN (418 mg, 6.42 mmol). The color change of the
precipitate was the same as observed for 1a. After stirring for 3 h,
the green powder was collected by filtration and washed with
2 ꢀ 10 mL water and 2 ꢀ 10 mL diethyl ether, and dried under
vacuum at room temperature for 3 h (Yield: 211 mg, 97%). IR
(KBr), υCN 2057 (s), 2023 (sh) cm-1
.
Method D (1d). To a 5 mL, colorless aqueous solution of
Mn(O2CMe)2 2H2O (250 mg, 1.20 mmol) was added a 5 mL
3
aqueous solution of KCN (389 mg, 5.97 mmol). The change in
color of the precipitate is the same as observed for 1a. After
stirring for 3 h, the product was collected by filtration and
washed with 2 ꢀ 10 mL water and 2 ꢀ 10 mL diethyl ether,
and dried under vacuum at room temperature for 3 h (Yield:
199 mg, 97%). IR (KBr), υCN 2057 (s), 2023 (sh) cm-1. Calcd
for C6K2Mn2N6: C, 20.94; H, 0.00; N, 24.42; obs C, 20.79;
H, <0.20; N, 23.96.
Figure 1. υOH and υCN IR spectra of 1a (blue), 1b-d (red), and 1e
(black).
Method E (1e). A 5 mL MeOH solution of Mn(O2CMe)2 (250mg,
1.44 mmol) was added to ∼15 mL MeOH solution KCN (329 mg,
5.05 mmol). The color change of the precipitate was the same as
observed for 1a. After stirring for 3 h, the green solid was collected by
filtration, washed with 3 ꢀ 10 mL MeOH, and dried under vacuum
for 2 h (yield: 160 mg, 65%). IR (KBr), υCN 2059 (s), 2023 (m) cm-1
.
Calcd for C6K2Mn2N6: C, 20.94; H, 0.00; N, 24.42; obs C, 20.61;
H, <0.20; N, 23.83.
Rb2MnII[MnII(CN)6] (2). To a 5 mL aqueous solution of
Mn(OAc)2 2H2O (172 mg, 0.701 mmol) was added a 5 mL
3
aqueous solution containing RbCN (352 mg, 3.16 mmol). A
gray precipitate immediately formed, and turned green within
5 min, and the mixture was stirred for an additional 2 h. The
green powder was collected by filtration and washed with 3 ꢀ 10 mL
of water, and 3 ꢀ 10 mL of diethyl ether, and dried under
vacuum over P2O5 for 12 h (Yield: 163 mg, 90%). IR (KBr), υCN
2064 (s) and 2028 (m) cm-1. Calcd for C6Mn2N6Rb2: C, 16.49; H,
0.00; N, 19.23; obs C, 16.34; H, <0.10; N, 18.89.
Cs2MnII[MnII(CN)6] (3). To a 7 mL aqueous solution of
Figure 2. TGA traces for 1a (blue) and 1b (red).
MnCl2 4H2O (112 mg, 0.566 mmol) was added a 7 mL aqueous
3
peak.8,11 When the reaction was carried out with an excess
of KCN in water with MnCl2 (1c) or with Mn(OAc)2 (1d) in
a ∼5:1 ratio this absorption was minimal. In contrast, when
prepared with Mn(OAc)2 in methanol (∼3:1) (1e), this
absorption does not appear. Use of D2O in the preparation
of 1b, or addition of D2O to solid 1a, leads to the shifting of
this absorption to 2677 cm-1, as expected for deuterium
substitution; thus, this peak is assigned to water. Also, the
weak absorption at 1675 cm-1 for 1a, taken as a Nujol mull,
is assigned as an OH bending mode. Hence, surface hydra-
tion occurs for samples prepared in water.
1a to 1e have identical υCN IR stretching bands [and X-ray
powder diffraction patterns (vide infra)]. The υCN absorption
and shoulder of K2Mn[Mn(CN)6] (1a-e) occur at 2058 and
2023 cm-1 (Figure 1) and occur at 2064 and 2028 cm-1 for
Rb2Mn[Mn(CN)6] (2). These are consistent with the υCN of
[MnII(CN)6]4- in PBAs (<2100 cm-1).6-8 The former value
is in accord with 2055 ( 4 cm-1 previously reported for
K2Mn[Mn(CN)6].8,11 These frequencies are lower than those
solution of CsCN (270 mg, 1.699 mmol). A precipitate immedi-
ately formed that turned blue-green within 1 h, and was col-
lected by filtration and washed with water and ethanol. The
solid was dried under vacuum over P2O5 at room temperature
for 1 h (Yield: 145 mg, 96%). IR (KBr), υCN 2073 (s) cm-1
.
Calcd for C6Cs2Mn2N6: C, 13.55; H, 0.00; N, 15.80; obs C,
13.14; H, <0.20; N, 15.24.
Results and Discussion
The aqueous preparation of K2Mn[Mn(CN)6] from KCN
and MnCl2 4H2O (1a), or with Mn(OAc)2 4H2O in a ∼3:1
3
3
ratio (1b), exhibits a sharp υOH absorption at 3628 cm-1 in
the IR spectra (Figure 1) that is characteristic of free
(nonhydrogen bonded) water or hydroxide.15,26 The two
previous studies of K2Mn[Mn(CN)6] did not report this
(26) Nakamoto, K. Infrared and Raman Spectra of Inorganic and
Coordination Compounds,5th ed.; John Wiley and Sons: New York, 1997; Part
A, p 161, also Part B, pp 53-57. Hornick, C.; Rabu, P.; Drillon, M. Polyhedron
2000, 19, 259.

Article
Inorganic Chemistry, Vol. 49, No. 4, 2010 1527
Table 1. Summary of Crystallographic Parameters for K2MnII[MnII(CN)6], 1a,
Rb2MnII[MnII(CN)6], 2, and Cs2MnII[MnII(CN)6], 3
K2MnII
Rb2MnII
Cs2MnII
[MnII(CN)6], 1a [MnII(CN)6], 2 [MnII(CN)6], 3
MW, g/mol
˚
˚
b, A
˚
c, A
344.2
436.92
10.4102(1)
7.4492(1)
7.2132(1)
90.072(1)
559.36(1)
2
531.8
a, A
10.1786(1)
7.4124(1)
6.9758(1)
90.206(1)
526.30(1)
2
10.6061(1)
10.6061(1)
10.6061(1)
90
1193.06(4)
4
Fm3m
2.1719(1)
0.079
0.071
293
1.117
β, deg
3
˚
V, A
Z
space group
P21/n
P21/n
F
calc, g/cm3
2.9612(1)
0.060
0.046
293
1.309
2.5941(1)
0.054
0.048
293
1.145
a,b
Rwp
Rexp
b,c
T, K
GOF (Rwp/Rexp
)
P
P
a Rwp = {[ iwi(yicalc - yiobs)2]/[ iwi(yiobs)2]}1/2
.
b yicalc and yiobs are
the calculated and observed intensities at the ith point in the profile,
normalized to monitor intensity. The weight wi is 1/σ2 from the counting
statistics, with the same normalization factor. N is the number of
=
c
points in the measured profile minus number of parameters. Rexp
P
{N/[ iwi(yiobs)2]}1/2
.
symmetry-breaking Jahn-Teller distortion for the low spin
MnII ion consistent with the observed low-symmetry struc-
ture. Furthermore, absorptions at 2138, 2121, or 2097 cm-1
characteristic of MnIII[MnIII(CN)6],7 K3[MnIII(CN)6],29a or
[NBu4]3[MnIII(CN)6]29b are not observed. Evidence for the
oxidative hydrolysis to Mn3O430 was also not observed.
Cs2MnII[MnII(CN)6] (3) has been prepared for the first
time and has a υCN absorption at 2073 cm-1, but does not
have any absorptions assignable to υOH. The bridged υCN
absorption in 3 is ∼15 cm-1 higher in frequency than those
observed for 1. This indicates that Mn-C bonding in 3 is
weaker than in 1. The presence of one υCN absorption
indicates Oh symmetry, and is expected for a Prussian-blue
structured material (vide infra).
Thermal Properties. The TGA traces for both 1a and 1b
(Figure 2) show the weight loss to be <0.25% upon
heating up to 150 °C, and 6 ( 2% weight loss upon
heating up to 500 °C. The former value sets the upper limit
of 0.05H2O per formula. This small amount of the water
implies the absence of water in the lattice. This is in accord
with the observation that the intensity of the υOH peaks in
1a and 1b is attenuated upon heating to 70 °C for 14 h, but
without a change in cell parameters, as evidenced from
the XRPD pattern. The 6% weight loss below 500 °C
indicates that K2Mn[Mn(CN)6] is relatively thermally
stable lattice because of the interaction of cyanide with
the Kþ. In contrast, Fe[Mn(CN)6] loses all of its cyanides
below 500 °C.13 Furthermore, M[Mn(CN)6] (M = V, Cr,
Mn, Co, Ni) decomposes below 400 °C.7 This suggests
that the Kþ in the K2MnII[MnII(CN)6] plays an impor-
tant role of its thermal stability. Compounds 1c and 1d,
are similar to 1a.
Figure 3. High-resolution synchrotron powder diffraction data (dots)
and Rietveld fit (line) of the data for K2MnII[MnII(CN)6], 1a (top),
Rb2MnII[MnII(CN)6], 2 (middle), and Cs2MnII[MnII(CN)6], 3 (bottom).
The lower traces for each plot are the differences, measured - calculated,
plotted to the same vertical scale.
of trivalent {e.g., MnIII[MnIII(CN)6] (2138 cm-1)} and tetra-
valent hexacyanomanganate {e.g., MII[MnIV(CN)6] 1.4H2O
3
(2183 cm-1)} Prussian blue type materials,7 because of the
lower oxidation state of manganese ion. In contrast, hex-
acyanometalates with only M coordinated to the cyanide
carbon atom have absorptions shifted to lower frequency by
20 to ∼75 cm-1 with respect to when both the C and N bond
to metal ions, as occurs for Prussian blue-type hexacyano-
Structure. The structures of 1 and 2 were solved by the
simulated annealing method, and refined from the syn-
chrotron powder diffraction data (Figure 3), and the
structure of 3 was refined as the expected Prussian blue
metalates6,27 {i.e., 2037 cm-1 in K4[Mn(CN)6] 3H2O28 vs
3
2058 cm-1 in K2Mn[Mn(CN)6]}. The two peaks are attrib-
uted to the asymmetric υCN stretching modes arising from a
(27) (a) Shriver, D. F.; Shriver, S. A.; Anderson, S. E. Inorg. Chem. 1965,
4, 725. (b) Dunbar, K. R.; Heintz, R. A. Prog. Inorg. Chem. 1997, 45, 288.
(c) Jones, L. H.; Swanson, B. I. Acc. Chem. Res. 1976, 9, 128.
(28) Sharpe, A. G. The Chemistry of Cyano Complexes of the Transition
Metals; Academic Press: New York, 1976.
(29) (a) Gupta, M. P.; Milledge, H. J. Acta Crystallogr. 1974, B30, 656.
(b) Buschmann, W. E.; Liable-Sands, L.; Rheingold, A. L.; Miller, J. S. Inorg.
Chim. Acta 1999, 284, 175.
ꢀ
ꢀ
(30) Buckelew, A.; Galan-Mascaros, J. R.; Dunbar, K. R. Adv. Mater.
2002, 14, 1646.

1528 Inorganic Chemistry, Vol. 49, No. 4, 2010
Her et al.
Figure 4. MnC6 (a), and MnN6 (b) octahedral coordination spheres for K2MnII[MnII(CN)6], 1a, showing the alkaliþ N interactions (dashed lines).
3 3 3
Rb2MnII[MnII(CN)6], 2, is isostructural (see TOC).
structure. At the final refinement step, all atoms were
freely refined with isotropic displacement parameters
(IDP). The C and N atoms of each cyano group were
set to have the same IDP value. The present powder
diffraction measurements do not clearly distinguish the
orientation of the cyano groups; they were assigned from
their bond lengths. In the Cs2Mn[Mn(CN)6] XRPD
pattern refinement, some impurity lines were observed
at the higher angle tail of the major Bragg reflections, and
were described as individual peaks whose centroid posi-
tion, integrated intensity, and full width at half maximum
were freely varied in the Rietveld process. Therefore, the
structural model of Cs2Mn[Mn(CN)6] was affected only
by the sharp Bragg reflections. The crystallographic data
for 1a, 2, and 3 are summarized in Table 1. The mono-
(average 177°) for 2 approach linearity, as expected. However,
the Mn-N-C angles of 152.7(7), 148.2(7), and 145.6(7)o
(average = 148.8°) for 1, and 150(1), 153(1), and 157(1)o
(average = 153.3°) for 2 deviate substantially from linearity,
an unexpected result for a Prussian blue structured material.
Each cyanide bridges a low- and a high-spin Mn(II) ion to
form a three-dimensional (3-D) framework structure with
void spaces occupied by Kþ (1a, Figures 4 and 5) or Rbþ
II
ions (2). The MnII Mn separations are 5.09 and 5.19 A
˚
3 3 3
for 1a and 2, respectively, which are slightly shorter than the
12
Mn Cr separations in NaMn[Cr(CN)6] (5.41 A) [and
˚
3 3 3
˚
5.30 A for Cs2Mn[Mn(CN)6] (vide infra)] that also pos-
sesses nonlinear M-N-C bonding. As a consequence,
MnC6, and MnN6 occur in alternating rows and have
their octahedra tilting in opposing directions (Figure 5). The
Kþ and Rbþ ions interact with seven cyanide nitrogens with
Kþ N and Rbþ N distances that range from 2.98 to
˚
clinic unit cell of 1 differs from the cubic a = 10.15 A
previously reported for 1;8 hence, 1 is not a face-centered
cubic that is typical of Prussian blue structured materials.9,12
K2MnII[MnII(CN)6] (1a) and Rb2MnII[MnII(CN)6] (2).
Both K2MnII[MnII(CN)6] (1a) and Rb2MnII[MnII(CN)6]
(2) are isostructural. The structure of 1a is composed of
six C-bound cyanides to a Mn(II) ion, and has two shorter
3 3 3
3 3 3
˚
˚
3.50 A, and 3.15 and 3.59 A for 1 and 2, respectively. The
alkali ion interactions with the cyanide ligands occurs in all
directions, and leads to a 3-D zigzag network of the
Mn-CN-Mn bonds. Due to the observation of the νOH
absorption for 1a, crystallographic evidence was sought, but
could not be identified, in the Fourier difference map, and is
in accord with the presence of surface water.
˚
Mn-C distances of 1.90(1) and 1.92(1) A, and a longer
distance of 2.01(1) A, Figure 4a. This is attributed to a
˚
slight tetragonally elongated, Jahn-Teller distorted low-
In contrast to a cubic Prussian blue structural motif
where the void spaces are defined by the cubic framework
arrangement of atoms arising from covalent M-CN and
M0-NC bonding leading to linear M-CN-M0 angles,
the less covalent MnII (more ionic)32 can accommodate a
reduced M-CN-M0 angle induced from the electro-
static interactions increase with decreasing cation size
(Kþ > Rbþ > Csþ) minimizing void space, as occurs.
Similar, bent cyanide bridged ligands with nonlinear
Mn-N-C angles of 152.7 and 153.4° have been observed
for A[(Me3tacn)6MnCr6(CN)18](ClO4)3 (A = Na, K;
Me3tacn = N,N0,N00-trimethyl-1,4,7-triazacyclononane)
II
˚
spin Mn by ∼0.1 A. The other Mn(II) ion is high spin,
and it bonds to 6 Ns with similar Mn-N bond lengths
[2.18(1), 2.23(1), and 2.25(1) A], Figure 4b. For compar-
˚
ison, the MnII-C bond distances vary between 1.91 and
31a,b
˚
1.98 A for Na4[Mn(CN)6] 10H2O.
The Mn-C dis-
tances are 1.90(2), and 1.92(2) and slightly tetragonally
3
˚
elongated 1.97(2) A, and the Mn-N distances are 2.22(1),
II
˚
2.23(1), and 2.24(1) A for 2. The Mn -N bond for
NaMn[Cr(CN)6] is 2.207(9) A, and that for Mn3[Co-
13
˚
(CN)6]2 12H2O is 2.184(9) A.31c Hence, this is a nondis-
˚
3
II
˚
torted high-spin Mn ion. The CN bond distance (∼1.14 A)
is not sensitive to the chemical environment.10,31d,e
The Mn-C-N angles of 168.7(9), 172.9(7), and 178(1)o
(average = 173°) for 1a, and 176(2), 177(1), and 179(1)
(32) Covalency increases with crystal field stabilization, Δo; hence, high
spin MnII-N6 is more ionic than low spin MnII-C6.Mingos, D. M. P.
Essential Trends in Inorganic Chemistry; Oxford University Press: New York,
1998; p 331.
(31) (a) Tullberg, A.; Vannerberg, N. Acta Chem. Scand. A 1974, 28, 551.
(b) Tullberg, A.; Vannerberg, N.-G. Acta Chem. Scand. 1971, 25, 343. (c) Beall,
G. W.; Milligan, W. O.; Korp, J.; Bernal, I. Inorg. Chem. 1977, 16, 2715.
(d) Brown, D. B.; Shriver, D. F. Inorg. Chem. 1969, 8, 37. (e) Ludi, A.; G€udel,
H. U. Struct. Bonding (Berlin) 1973, 14, 1.
(33) (a) Heinrich, J. L.; Sokol, J. J.; Hee, A. G.; Long, J. R. J. Sol. State
ꢀ
Chem. 2001, 159, 293. (b) Rodríquez-Hernandez, J.; Reguera, E.; Lima, E.;
Balmaseda, J.; Martínez-Garcia, R.; Yee-Maderia, H. J. Phys. Chem. Sol. 2007,
ꢀ
68, 1630. (c) Escorihuela, I.; Falvello, L. R.; Tomas, M. Inorg. Chem. 2001, 40,
636.

Article
Inorganic Chemistry, Vol. 49, No. 4, 2010 1529
Figure 5. View normal to the bc plane of the 3-D network lattice for
K2MnII[MnII(CN)6], 1a, showing MnC6 surrounded by four MnN6
octahedra pointing in different directions (Kþ are not shown for clarity).
Rb2MnII[MnII(CN)6], 2, is isostructural.
Figure 6. Structure of Cs2MnII[MnII(CN)6], 3, showing the usual fcc
Prussian blue structure.
also possessing high spin MnII.33a Likewise, Zn-N-C
angles ranging from 151.4 and 165.1° have been reported
for Zn3[MIII(CN)6]2 (M = Fe, Co), Zn3A2[FeII(CN)6]2
(A = Rbþ, NH4þ) possessing ionic ZnII,33b and an even
further reduced CuII-N-C angle of 120.1° has been
documented.33c
Magnetic Properties. The magnetic susceptibilities, χ(T),
of 1-3 as well as K4[Mn(CN)6] 3H2O were measured
3
between 5 to 300 K, and are plotted as χT(T), Figures 9
and 10, and the results are summarized in Table 2. [See
Supporting Information, Figures S1 and S2 for the χ(T)
data.] K4[Mn(CN)6] has a room temperature χT value of
0.50 emu K/mol that exceeds the low-spin spin-only value
0.375 emu K/mol for MnII indicative of an orbital contribu-
tion and g = 2.31. The χT value decreases slightly with
decreasing temperature to 0.45 emu K/mol at 60 K, fol-
lowed by a big decrease to 0.275 emu K/mol at 2 K,
indicating either an intermolecular interaction or a
spin-orbit interaction (λ). For K4[Mn(CN)6], it is known
that the magnetic moment has a range of 0.48 to 0.59 emu
K/mol,17 and attempts to obtain the spin-orbit interaction
(λ) for K4[Mn(CN)6]17 were unsuccessful because of the
large deviation below ∼100 K.
Cs2MnII[MnII(CN)6] (3). Cs2MnII[MnII(CN)6] (3) ex-
hibits the typical cubic Prussian Blue structure34 [a =
˚
10.6061(1) A] with one hexacoordinate Mn(II) sur-
rounded by six 6 C-bound cyanides, and the other Mn(II)
surrounded by six 6 N-bound cyanides, with linear
MnCNMn linkages. The Mn-C, Mn-N, and C-N
˚
distances are 1.93(1), 2.19(1), and 1.18(1) A, respectively,
Figure 6. Hence, structural evidence for Jahn-Teller
distortion about the low spin Mn(II) site is not present.
˚
The Mn(II) Mn(II) separation is a/2 (5.303 A). The
cubic unit cell is very similar to that of NaMn[Cr(CN)6]
3 3 3
K2MnII[MnII(CN)6] (1). 1a and 1e have room-tempera-
ture χTvalues of 5.56 and 5.10 emu K/mol, respectively, and
like the previously reported values of 4.998 and 5.1311 emu
K/mol for K2Mn[Mn(CN)6] exceed the 4.75 emu K/mol
expected for the spin-only value for one low- and one high-
spin Mn(II) site. This is indicative of an anisotropic orbital
contribution from the low spin Mn(II), and the differing
average orientations of the polycrystalline samples. These
values decrease upon cooling and reach shallow minima of
4.99 and 5.11 emu K/mol at 90 and 105 K, respectively, and
then increase and reach maxima of 100 ( 2 emu K/mol at
32 ( 1 K for 1a and 1e, Figure 10. Above 120 K, the χ-1(T)
12
[a = 10.8159(12) A]. Hence, 3 has a structure that is
˚
typical of Prussian blue.
Comparison of the A2MnII[MnII(CN)6] (A = K, Rb, Cs)
Structures. The A2MnII[MnII(CN)6] (A = K, Rb) struc-
tural motif is unusual and differs from the well-known fcc
Prussian blue motif31e as a consequence of its monoclinic
unit cell. Perspective views of the Prussian blue and
Rb2MnII[MnII(CN)6] structures are shown in Figure 7.
The new motif is related to the Prussian blue lattice
√
√
parameters that are approximately a, a/ 2, a/ 2, β ∼
90°. The distortion from an orthorhombic lattice is small,
but it is clearly required by the raw data, as shown by
profile fits in the Supporting Information, Figures S3 and
S4. The nonlinear Mn-C-N-Mn linkages for the latter
lead to canting of the octahedra, as described above
(Figure 5), and buckling of the 3-D lattice, Figure 7. This
is also observed looking down the a-axis, Figure 8. For
both structure types 3-D channels are present that are
filled by the alkali cations, Figure 8.
can be fit to the Curie-Weiss expression, χ µ (T - θ)-1
,
with θ = -63 ( 3 K, indicating significant antiferromag-
netic couplings. The significant increase in χT(T) at low
temperature suggests the onset of magnetic ordering. Like-
wise, previously K2Mn[Mn(CN)6] was reported to reach a
minimum at 105 K, and then increase and reach maxima of
1325 emu K/mol at 24 K as well as θ = -19 K from data
taken between 100 and 290 K.8a
A fit of the χ-1(T) data to the Neel hyperbolic equation,
ꢀ
(34) The observed XRPD pattern of 3 can also be equally well-explained
with the Fm3c space group symmetry whereby all C and N are positionally
disordered, but this was discarded, as it is inconsistent with the structure of
the other PBAs, and the magnetic data.
eq 1 {where C and θ are the Curie and Weiss constants,
respectively, and θ0 and ζ are proportional to ηAηBC(ηA
- ηB) and ηAηBC (where ηi is the fractional occupancy of

1530 Inorganic Chemistry, Vol. 49, No. 4, 2010
Her et al.
Figure 7. Perspective views of the Prussian blue structured Cs2MnII[MnII(CN)6], 3, (a), and non-Prussian blue structured Rb2MnII[MnII(CN)6] (b).
Figure 8. Views along the a-axis for Cs2MnII[MnII(CN)6] (3) (a), and non-Prussian blue structured Rb2MnII[MnII(CN)6] (2) (b) showing the location of
the respective cations, and buckling of the lattice for (2).
Figure 9. χT(T) for K4[Mn(CN)6] (solid blue circles), and Cs2MnII-
[MnII(CN)6], 3 (open red circles).
each sublattice site), respectively}35 (Figure 11) gives a Tc
of 41 K for both 1a and 1e.
Figure 10. χT(T) for 1a (ꢀ), 1e (þ), and 2 (solid blue circles). The lines
are guides for the eye.
magnetic field (Figure 12). The bifurcation temperature,
Tb, of the FC/ZFC magnetizations of 47 and 46 K for 1a
and 1e, respectively.
T -θ
C
ζ
χ-1
¼
-
ð1Þ
T -θ0
The field-cooled (FC) and zero-field-cooled (ZFC)
magnetizations for 1a and 1e were measured in a 5 Oe
The 5 K field dependence of the magnetization, M(H),
shows a rapid rise with increasing field, H, with respect to

Article
Inorganic Chemistry, Vol. 49, No. 4, 2010 1531
Table 2. Summary of Magnetic Properties for 1-3
1a
1e
Kþ
5.10
4.75
-65
-20
41
K2Mn[Mn(CN)6]8
2
3
cation
χobsT, emu K/mol (295 K)
Kþ
Kþ
Rbþ
4.83
4.75
-30
-19
34.6
38
Csþ
4.68
4.75
-15
-15
21
5.56
4.75
-60
-33
41
4.99,8 5.1311
χcalcT, emu K/mol (spin-only)
4.75
-19
b
41
b
θ, K (Curie-Weiss)
ꢀ
θ, K (Neel), eq 1
ꢀ
TN, K (Neel), eq 1
Tb, K (ZFC/FC)
Tf, K (ac)
Hcr, Oe
M, emu Oe/mol (5 K, 50 kOe)
Mr, emu Oe/mol
47
38
46
38
24
22
b
34
3900
16,100
8600
9150
15,800
9350
370
24,000 a
430
8850
16,300 c
9300
5000
9360 c
640
a 4.5 K, 70 kOe.8 b Not reported. c Not saturated. d 19,150 emu Oe/mol at 90 kOe.
Figure 11. χ-1(T) for 1a (O, red), 1e (ꢀ, blue), 2 (þ, green), and 3
ꢀ
(b, black). The solid lines represent the best-fit curves to the Neel
hyperbolic equation, eq 1.
the expectation from the Brillouin expression, is sugges-
tive of magnetic ordering for 1a and 1e (Figure 13). The
50 kOe magnetizations at 5 K are 16, 100, and 15,800 emu
Oe/mol for 1a and 1e, respectively, but are still gradually
rising with increasing field. Hence, they are in accord with
the expectation of 22,340 emu Oe/mol for antiferromag-
netic coupling leading to ferrimagnetic ordering, and are
consistent with the reported 70 kOe magnetization at 4.5 K
that is 24,000 emu Oe/mol for K2Mn[Mn(CN)6].8 At 5 K
the remanent magnetizations, Mr, are 8600 and 9350 emu
Oe/mol, and the coercive fields, Hcr, are 3900 and 9150 Oe
for 1a and 1e, respectively (Figure 13). Hence, the biggest
difference between the sample that has a υOH absorption
(i.e., 1a), and the one that does not (i.e., 1e), is the more
than doubling of the Hcr, but in both cases they exceed the
previously reported value of 370 Oe.8
Figure 12. Field-cooled (FC, filled circles) and zero-field-cooled (ZFC,
open circles) M(T) for 1a (blue), and 1e (red) showing the bifurcation
temperatures.
In addition, peaks in both the frequency-independent in-
phase, χ0(T), and out-of-phase, χ00(T), components of the
AC susceptibility for 1a and 1e occur, and are characteristic
of ferri- or ferromagnetic ordering, Figure 14. Using the
maximum in the 10 Hz χ0(T) data, Tc = 38 K, which is a bit
reduced from the bifurcation temperature, Tb.
The 41 K Tc for K2Mn[Mn(CN)6 is in contrast to that
of the isoelectronic FeIII[FeIII(CN)6] 4H2O with low spin
3
2
FeIII (t2g5, S = 1/2) and high spin FeIII (t2g3eg , S = 5/2)
ꢀ
(35) (a) Neel, L. Ann. Phys. 1948, 3, 137. (b) Smart, J. S. Am. J. Phys. 1955,
23, 356.
Figure 13. M(H) for 1a (black circles), 1e (red circles), and 2 (blue
circles).

1532 Inorganic Chemistry, Vol. 49, No. 4, 2010
Her et al.
Figure 16. χ0(T) (b) and χ00(T) (O) for 2 at 10, 100, and 1000 Hz.
with respect to the Fe-N-C angle, stabilizing the observed
solvent-free non-Prussian blue structure
Rb2MnII[MnII(CN)6] (2). The room temperature χT
value for 2 is 4.83 ( 0.07 emu K/mol, and is also in
accord with 4.75 emu K/mol expected for the spin-only
value for one low- and one high-spin Mn(II) site
(Figure 8).37 The χT(T) decreases upon cooling and
reaches a minimum of 4.40 emu K/mol at 100 K, before
again increasing to 117 emu K/mol at 28 K. Above 170 K,
the χ-1(T) data can be fit to the Curie-Weiss expression
with θ = -30 K, indicating antiferromagnetic coupling,
albeit weaker, than observed for 1a and 1e. The large
increase in χT(T) at low temperature suggests the onset of
Figure 14. χ0(T) for 1a (blue circles) and 1e (red circles) showing the
38 K ordering temperature.
magnetic ordering. The Tc of 34.6 K was obtained from
0
ꢀ
fitting the data to Neel s hyperbolic equation, eq 1
(Figure 11).8,10 This is reduced from 41 K for K2Mn-
[Mn(CN)6].
The 1 Oe FC/ZFC magnetizations for 2 exhibit a Tb of
38 K (Figure 15, H e 10 Oe). Unlike for 1a and 1e, the
ZFC magnetizations exhibit an anomaly (bump) at ∼34 K,
whose intensity increases with applied field (Figure 15).
This suggests a transition to a second magnetic phase
upon cooling; however, the nature of this is unknown, but
might arise from canting. To rule out an extrinsic im-
purity, an independent sample was studied, and the
magnetic data were in quantitative agreement.
The 5 K M(H) shows a rapid rise with increasing field,
H, which with respect to the expectation from the Bril-
louin expression is suggestive of magnetic ordering
(Figure 13). At 90 kOe the magnetization is 19,250 emu
Oe/mol and still rising; hence, it is in accord with the
expectation of 22,340 emu Oe/mol for antiferromagnetic
coupling leading to ferrimagnetic ordering. As shown in
Figure 13, 2 has a significant coercive field of 8850 Oe,
remanent magnetization, Mr, of 9300 emu Oe/mol at 5 K,
and the hysteresis is almost identical to that for 1e
(Figure 13).
In addition, peaks in both the frequency-independent
χ0(T) and χ00(T) for 2 occur and are characteristic of ferri-
or ferromagnetic ordering, Figure 16. Using the max-
imum in the 10 Hz χ0(T) data, Tc = 34 K, which is a bit
reduced from the 38 K bifurcation temperature, Tb. χ0(T)
has a shoulder at ∼29 K and χ00(T) exhibits a second
Figure 15. Field-cooled (FC) (open circles) and zero-field-cooled (ZFC)
(filled circles) M(T) for 2 at 0.1 (red), 1 (blue), 10 (black), and 100 Oe
(green) applied fields.
that exhibits a lower Tc of 17.4 K and Hcr of 465 Oe.36
FeIII[FeIII(CN)6] 4H2O possesses a fcc structure with
3
water molecules in the interstitial sites. Thus, the bent
Mn-CN-Mn bond contributes to the higher magnetic
ordering temperature and coercive field of K2Mn[Mn-
(CN)6] with respect to FeIII[FeIII(CN)6] 4H2O. The
3
MnII MnII separations of 5.09 A are identical to the
˚
3 3 3
FeIII FeIII distance of 5.1090(5) A; thus, the bent
˚
cyanide bridges, rather than the short distance between
3 3 3
the spin centers, play a crucial role in the superexchange
mechanism. The more ionic character (less directional
covalency) associated with MnII with respect to FeII 32
accommodates the significantly reduced Mn-N-C angle
(37) It is also substantially greater than the 1.0 emu Oe/mol expected for
two low-spin, S = 1/2 MnII (g ∼ 2.3) sites, and is substantially lower than the
8.75 emu Oe/mol expected for two high-spin, S = 5/2 MnII (g ∼ 2.0) sites.
(36) Kumar, A.; Yusuf, S. M.; Keller, L. Phys. Rev. B 2005, 71, 054414.

Article
Inorganic Chemistry, Vol. 49, No. 4, 2010 1533
Figure 17. Field-cooled (FC) (red circles) and zero-field-cooled (ZFC)
(blue circles) M(T) for 3 showing the bifurcation temperature, Tb.
Figure 19. χ0(T) (O) and χ00(T) (ꢀ) for 3 showing the their 22 K Tc.
Figure 20. Correlation of increasing Tc with deviation from linearity of
the Mn-N-C angle for A2Mn[Mn(CN)6] (A = K, Rb, Cs).
Figure 18. M(H) for 3.
The magnetization at 50 kOe is 9360 emu Oe/mol and is
still rising without evidence of saturating. As shown in
Figure 18, 3 has a significant coercive field of 5000 Oe
and remanent magnetization of 640 emu Oe/mol at 5 K,
and the shape of the hysteresis curve is quite unusual.
The source of the unusual shape of the M(H) curve, the
high coercivity, and lack of approaching saturation is
unknown.
Peaks in both χ0(T) and χ00(T) characteristic of ferri- or
ferromagnetic ordering are also observed, Figure 19.
Using the maximum in the 10 Hz χ0(T) data, a Tc of
22 K, that is reduced from the bifurcation temperature,
Tb, is observed. The χ0(T) data of 3 has a small frequency
dependence with the temperatures of each maximum
absorption at ∼ 30 K that are suggestive of an unknown
second magnetic phase, and is also indicated from the FC/
ZFC magnetization data, Figure 15.
Cs2MnII[MnII(CN)6] (3). The room temperature χT
value for 3 is 4.68 emu K/mol, and like 1a to 1e, is in
accord with 4.75 emu K/mol expected for the spin-only
value for one low- and one high-spin Mn(II) site
(Figure 9).38 χT(T) decreases upon cooling and reaches
a minimum of 3.46 emu K/mol at 26 K, before again
increasing to 6.3 emu K/mol at 22 K. Above 25 K, the
χ-1(T) data can be fit to the Curie-Weiss expression with
θ = -15 K, indicating antiferromagnetic coupling, albeit
weaker, than observed for 1a, 1e, 2, and 3. The increase in
χT(T) at low temperature suggests the onset of magnetic
ordering.
being almost constant {φ ∼ 0; φ = ΔTf[TfΔ(log f)]-1
;
where Tf is the temperature of the peak in the lowest
frequency (10 Hz) data; and f is the frequency in hertz}
indicating no spin glass behavior. The out-of-phase,
imaginary, χ00(T) data displays a peak below 24 K,
indicating magnetic ordering.
The 5 Oe FC/ZFC magnetizations for 3(Figure 17) exhibit
a Tb of 24 K. The Tc of 21 K was obtained from fitting the
data to Neel s hyperbolic equation, eq 1, (Figure 11).8,10 This
0
ꢀ
is reduced from that reported for 1a and 1e.
Conclusion
(38) It is also substantially greater than the 1.0 emu Oe/mol expected for
two low-spin, S = 1/2 MnII (g ∼ 2.3) sites, and is substantially lower than the
8.75 emu Oe/mol expected for two high-spin, S = 5/2 MnII (g ∼ 2.0) sites.
K2Mn[Mn(CN)6] was made in aqueous (1a) or non-aqu-
eous (1e) media. 1a has a sharp υOH absorption in the IR

1534 Inorganic Chemistry, Vol. 49, No. 4, 2010
Her et al.
spectrum, while 1e does not. Use of D2O shifts this peak as
expected for deuteration. While this supports the presence of
H2O, that is not evident from the structure, TGA, XRPD, or
elemental analysis; thus, it is attributed to surface water, and
it may arise from surface hydrolysis. K2Mn[Mn(CN)6] has a
bent (148.8°) Mn-CN-Mn framework that is different with
respect to the face-centered structures typical of PBAs. This
nonlinearity is attributed to the ionic nature if high spin
Mn(II) strong interactions between Kþ and cyanide ions. 1a
and 1e exhibit essentially identical structures and similar
magnetic properties. Rb2Mn[Mn(CN)6] (2) is isostructural
(Csþ), Figure 20, and decreasing Mn(II) Mn(II) separa-
tions of 5.09 (K ) < 5.19 (Rb ) < 5.30 A (Cs ). This
suggests that the bent cyanide bridgesplay a crucial role inthe
superexchange mechanism by increasing the coupling via
3 3 3
þ
þ
þ
˚
shorter Mn(II) Mn(II) separations, and perhaps enhanced
3 3 3
overlap. MO calculations are needed to provide insight into
this. Furthermore, use of the smaller Naþ cation should lead
tostronger interactionsincreasing Tc, andincreasing pressure
should lead to greater deviation of the MnNC angle from
linearity and increase Tc.
to K2Mn[Mn(CN)6]; however, the longer Rbþ N dis-
Acknowledgment. We appreciate the sample of RbCN
provided by Mathew C. DeLong (Department of Physics,
University of Utah) and the helpful discussions with
Profs. Larry R. Falvello (University of Zaragoza), Jeffrey
Long (University of California, Berkeley), and assistance
from Saul Lapidius (Department of Physics and Astron-
omy, Stony Brook University) and Kendric J. Nelson
(Utah), and the continued partial support by the Depart-
ment of Energy Division of Material Science (Grants DE-
FG03-93ER45504 and DE-FG02-01ER4593). Use of the
National Synchrotron Light Source, Brookhaven Na-
tional Laboratory, was supported by the U.S. Depart-
ment of Energy, Office of Basic Energy Sciences, under
Contract No. DE-AC02-98CH10886.
3 3 3
tances with respect to Kþ N provide weaker electrostatic
3 3 3
interactions decreasing the nonlinearity of the Mn-N-C
linkages to 153.3°. This distortion from the traditional fcc
Prussian blue structure is not observed for the isoelectronic
FeIII[FeIII(CN)6] 4H2O because of the greater charges and
3
greater covalency (reduced ionic character).
Cs2Mn[Mn(CN)6] has the archetypal fcc Prussian blue
structure and is magnetically characterized to order as a
ferrimagnet at 21 K. While saturation does not occur at
50,000 Oe, it has a remanent magnetization of 640 emu Oe/
mol and a large coercive field of 5 kOe at 5 K. The latter is
attributed to the positional disorder of the cyanides.
The 41 K Tc for K2Mn[Mn(CN)6] is in contrast to isoelec-
tronic FeIII[FeIII(CN)6] 4H2O with low spin FeIII (t2g5, S =
3
2
Supporting Information Available: The powder X-ray crystal-
lographic information files (CIF) for K2MnII[MnII(CN)6],
Rb2MnII[MnII(CN)6], and Cs2MnII[MnII(CN)6] (CCDC
709700, 723397, and 709701), χ(T) for K4[Mn(CN)6], 1a, 1e, 2,
and 3, and fits of the observed PXRD data for 1a and 2 assuming
orthorhombic and monoclinic space groups. This material is
available free of charge via the Internet at http://pubs.acs.org.
1/2) and high spin FeIII (t2g3eg , S = 5/2) that exhibits a lower
Tc of 17.4 K and Hcr of 465 Oe.36 The Tc is comparable to that
for Prussian blue structured 3. Furthermore, Tc increases as
Kþ > Rbþ > Csþ for A2Mn[Mn(CN)6] (A = K, Rb, Cs)
and this correlates with the increasing deviation of linearity of
the Mn-N-C linkages of 31.2 (Kþ) > 26.7 (Rbþ) > 0°