—

11.1.}; .1. .2.
.§:,...,,.Efi+mfi .
.. . y
hr$ .1.
$3.»

I:

a v3

.
.. .
.n. in:
‘ 4%“... as.

.n!’

. 5‘
.m .9
5 5 L..." as n

5.7”... 3.31

:12:
p.

i 52......
9~’{I'?'plln£ I.
1).. . .93.!
l. a . :‘mew.
ha. .P..3.I‘o:
.lt! . 5‘ .
v» .. t...

n r!)

“fir... .:

.. i I ;K
£1155fi

st!
.. I t. 3
:1

.1. a...
(1.72....

N. (\rlr

 

 

 

 

 

 

ii...
at .14.. hark unfit!
.5. 511““..ch ;
.-

Dal}. .
an. 1.170..
. 9.6 . .
txft,.:..so"¥a
9&1.

I {nfi
hurd- .. i5-..a:.§.«wm$-UM¥1
.l...§.!..st:)snfl‘d.lz.ls# “MS
all..rur..r..15. . .531...
arkg3uxtfi.§hu
_ $.4..9.;finunh£§§xf
1.1.. 1...»?!
, {zill1v1fl
flint}: v
. to. f, I
80““... 2 .
2. .93....xmtkunfimfful .31;
.1. .MH. 1.311.! ..E1axf.£¥iN-|
:9 iRWfi-‘flrhv t u. input; .0!!! $0.1
i. .e .2...» 21...! has»)! 44.21....)
I: [Ab-ior\ -
.E.
1.1.5.3..
, .2 95.21
. 14‘s.;
23).??on la . V
€2.53:

1...: $1.7. 1.11
l.‘;illl

‘ .ullttly: K x

7‘ I‘llt‘ltl
.nuv.“ Iv..- 1.x.
Ii 5;")?!
90. ..¢1\lftst.

THE§S

4'

,0: O

 

LIBRARY

Michlgan State
University

 

 

 

This is to certify that the

dissertation entitled

TOPOCHEMICAL ASSEMBLY OF COVALENT MATERIALS
USING DIHYDROGEN BONDING

presented by

Radu Custelcean

has been accepted towards fulfillment
of the requirements for

Ph . D. degree in Chemistry

 

 

f 0/».

/ Major profesch

Date 7 Aujgét LOCO

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

JUL 0 2 2002

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11m elem-gasp.“

 

TOPOCHEMICAL ASSEMBLY OF COVALENT MATERIALS
USING DIHYDROGEN BONDING
By

Radu Custelcean

A DISSERTATION
Subnfinedto
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHYLOSOPHY
Department of Chemistry

2000

ABSTRACT
TOPOCHEMICAL ASSEMBLY OF COVALENT MATERIALS
USING DIHYDROGEN BONDING
By

Radu Custelcean

This study explores the potential of the unconventional hydrogen bonds
(dihydrogen bonds) between hydridic hydrogens in X-BH3' (X = H, CN) and traditional
AH (A = O, N) proton donors to serve as preorganizing interactions for the topochemical
assembly of covalent materials. Such weak A-H---H-B interactions can indeed be used to
organize and hold a structure’s form while it is more firmly fastened together by A-B
bond formation, transferring thus the initial order from the starting crystal to the newly
formed covalent frame. This strategy makes dihydrogen bonding a potentially powerful
tool for the rational assembly of crystalline covalent solids with controlled structures and
properties, a class of compounds generally accessible only by empirical methods or
serendipitous discoveries.

Evidence of such crystal lattice control in the transformation O-H---H-B —) O-B +
H2 was first obtained in studies of the solid-state structures and reactivities of
triethanolamine (TEA) complexes with various metal borohydrides (MBH3X-TEA: M =
Na, X = H; M = Na, X = CN; M = Li, X = H). Thus, solid-State decomposition of
N aBH4-TEA, whose crystal structure shows multiple dihydrogen bonds between the BH.{
ions and the OH groups from TEA, is topochemical, leading to a polymeric

trialkoxyborohydride, in sharp contrast to the N aBH4 and hydride-free polymeric borate

disproportionation products obtained from decomposition in solution or melt. With its
less basic hydridic sites, the dihydrogen-bonded N aNCBH3-TEA complex could not be
decomposed in the solid state, losing H2 only at ca. 100 °C above its melting point. Thus,
in addition to close H-H contacts, the relative acidity/basicity of the proton/hydride
partners is also critical for the reactivity of the dihydrogen-bonded systems. The Hn-H
interactions were conveniently tuned by substituting Li+ for Na” in the NaBH4-TEA
complex, which led to shorter H-H contacts and enhanced solid-state reactivity, as a
result of stronger complexation by Li+ and consequently higher acidity of the OH sites.
Mechanistic investigations using in Situ solid-State llB NMR, optical microscopy, FT-IR,
and XRD, found that, like most solid-state reactions, the topochemical decomposition of
LiBH4-TEA is heterogeneous, with nucleation of the product phase from the parent
crystal. Kinetic analysis and HID exchange experiments established that proton transfer
along the H---H bond, at the reactant/product interface, is the rate limiting step, with
activation parameters AH" and AS" of 20.1 :i: 2.4 kcal/mol and -16.8 i 6.2 cu,
comparable with the analogous values found for the aqueous hydrolysis of BH4‘ in
neutral water.

Finally, the concept of crystallinity transfer from dihydrogen- to covalent-bonded
crystals was demonstrated by the crystal-to-crystal decomposition of NaBH4-THEC
(THEC = N,N’,N’ ’,N’ "-tetrakis-(2-hydroxyethyl)cyclen), in which self-assembly into
dihydrogen-bonded closed-loop large globular dimers minimized the unit cell Shrinkage
accompanying the O-H---HB to 0-3 conversion. Further elaboration of these dihydrogen-
bonded building blocks offers real prospects for the rational assembly of extended H-c-H

bonded, and ultimately covalent solids with controlled architectures.

 

 

 

TABLE OF CONTENTS

LIST OF TABLES .............................................................................. v
LIST OF FIGURES ............................................................................. vi
1. Dihydrogen Bonding: A Literature Review ........................................... l
1 . 1. Introduction ................................................................................. 1
1.2. Structural and Energetic Characterization .............................................. 2

A. Dihydrogen Bonding in the Main Group Hydrides ...................... . .......... 2

B. Dihydrogen Bonding in Transition Metal Hydrides ................................ 14
1.3. Self-Assembly of Extended Dihydrogen-Bonded Systems ........................... 25
1.4. Dynamics of Dihydrogen Bonds ......................................................... 29
2. Organic Solid-State Reactions: Mechanisms and Concepts ........................ 46
2.1. Topochemical Polymerizations and Polycondensations ............................... 51
2.2. Hydrogen Bonding Assisted Topochemical Reactions ................................. 55
2.3. Kinetics of Solid—State Decompositions ................................................. 60

3. Topochemical Assembly of Covalent Materials Using Dihydrogen Bonding 64

3.1. Results and Discussion ..................................................................... 67
3.2. Conclusions .................................................................................. 1 19
3.3. Experimental Section ....................................................................... 122
BIBLYOGRAPHY ............................................................................. 145

iv

LIST OF TABLES

Table 3.1. NMR data for BHx(OCH3)4-x', 47, and 48 ...................................................... 85
Table 3.2. IR data for LiBH4o TEA (47), LiBD4-TEA (47a), LiBH4(50%D)-TEA (47b),
and their corresponding solid state decomposition products, 48, 48a, and 48b. LiBH4 is
included for comparison .................................................................................................. 87
Table 3.3. Rate constants k for the solid state decomposition of 47 , calculated from the
Avrami-Erofeev law for nucleation and two-dimensional growth in the specified

conversion ranges, together with the correlation coefficients of the linear regression

analysis R2 ....................................................................................................................... 94
Table 3.4. Crystal data and structure refinement for N APA H3BCN (44) ..................... 125
Table 3.5. Crystal data and structure refinement for N aBH4-TEA (45) ......................... 126
Table 3.6. Crystal data and structure refinement for N aN CBH3'TEA (46) ................... 127
Table 3.7. Crystal data and structure refinement for LiBH4-TEA (47) .......................... 128
Table 3.8. Crystal data and structure refinement for NaBI-I4-THEC (49) ...................... 129

Table 3.9. Crystal data and structure refinement for the decomposition product of

NaBH4-THEC in DMSO ................................................................................................ 130
Table 3.10. Crystal data and structure refinement for N aNCBH3'THEC (50) .............. 131
Table 3.11. Crystal data and structure refinement for NaBH4-THPC (51) .................... 132

Table 3.12. Crystal data and structure refinement for 9-(1’,4’,7’-triS-hydroxyethyl-
cyclen)methylanthracene (52) ........................................................................................ 133

Table 3.13. Crystal data and structure refinement for N aBI-I4-[9-( l ’,4’,7’-tris-

hydroxyethylcyclen)methylanthracene] -EtzO (53) ....................................................... 1 34

LIST OF FIGURES

Figure 1.1. O-H---H-B dihydrogen bonds found in N3BH4'2H20 in the solid-state ........ 3
Figure 1.2. C2 isomer of the N H3BH3 dimer .................................................................... 4
Figure 1.3. C2}. isomer of the NH3BH3 dimer .................................................................. 5
Figure 1.4. The shortest N-Hu-H-B contacts found in NH3BH3 in the solid-state .......... 6

Figure 1.5. Dihydrogen-bonded dimers found in (CH3)2NH-BH2-N(CH3)2-BH3 in the

solid-State ......................................................................................................................... 6
Figure 1.6. C-H---H-B close contacts in 7-10 ................................................................. 9
Figure 1.7. Calculated structure of the NH3A1H3 dimer ................................................ 10
Figure 1.8. Calculated structure of (NH2A1H2)3 ............................................................ 10
Figure 1.9. Calculated structure of the (NH2A1H2)3 dimer ............................................ 11
Figure 1.10. Self-assembly of cyclotrigallazane in the solid-state ................................ 12
Figure 1.11. Calculated structure of the (NHzGaH2)3 dimer ......................................... 13

Figure 1.12. Calculated structure for dihydrogen-bonded complex 26a; < FH---H-Mo =
127°, < F—H---HMo = 174° ............................................................................................. 23
Figure 1.13. Generic representation of C-H---H-M dihydrogen bonds in complexes of

transition metal hydrides with R3-x(Ph)xP ligands ......................................................... 24
Figure 1.14 Self-assembly of dihydrogen-bonded complex 28 in one-dimensional
chains ............................................................................................................................. 26
Figure 1.15. Self-assembly of dihydrogen-bonded complexes 29-[K(1,10-diaza-18-

crown-6)] in one-dimensional chains ............................................................................ 27

vi

 

 

Figure 1.16. Self-assembly of dihydrogen-bonded complexes 29a,b-[K(THF)(1-aza-18-
crown-6)] in one-dimensional chains ............................................................................... 27
Figure 1.17. Self-assembly of dihydrogen-bonded complexes 29b-[K(THF)(18-crown-
6)] in one-dimensional chains ......................................................................................... 28
Figure 1.18. Self-assembly of guanidinium borohydride in one-dimensional extended
tapes ................................................................................................................................ 28
Figure 3.1. X-Ray Crystal Structure of NAPA H3BCN Showing the H---H contacts in

A ...................................................................................................................................... 69
Figure 3.2. X-Ray crystal structure of NaBH4-TEA: (a) coordination of Na+;

(b) dihydrogen bonds connecting the chains, with H-H contact distances in A .............. 72
Figure 3.3. Progress of N aBH4-TEA decomposition monitored by X-ray powder
diffraction: (a) initial complex; (b) high-temperature polymorph; (c) final decomposed
material ............................................................................................................................ 75
Figure 3.4. X-Ray crystal structure of N aNCBH3-TEA: (a) coordination of N a+;

(b) dihydrogen bonds connecting the chains, with H-H contact distances in A .............. 79
Figure 3.5. X-Ray crystal structure of LiBI-h-TEA showing the H~--H close contacts in

A ...................................................................................................................................... 81
Figure 3.6. X-Ray powder diffraction pattern of LiBH4-TEA (a) and its solid-state
decomposition product (b) ............................................................................................... 83
Figure 3.7. Calculated structure of LiBH(OCH3)3 with the interatomic distances in A.. 89
Figure 3.8. Typical microscopic view (transmitted light) of LiBH4-TEA solid-state
decomposition: (a) initial crystal, (b) 10 min at 110 °C, (c) final decomposed crystal;

Scale bar: 100 um ............................................................................................................ 91

vii

Figure 3.9. Conversion (0t) vs. time curves for the solid-state decomposition of 47: O =

105°C;CI=110°C;A=115°C;X=120°C ................................................................. 92

“2

Figure 3.10. Plots of the Avrami-Erofeev law, [-ln(l-a)] against time, for the solid

state decomposition of 47: O = 105 °C; C] = 110 °C; A = 115 °C; x = 120 °C .............. 93
Figure 3.11. Arrhenius plot for the solid-state decomposition of 47 at 105-120 °C ....... 95
Figure 3.12. Variation of activation energy with conversion for the solid-state
decomposition of 47 , obtained using the isoconversional method .................................. 95
Figure 3.13. Proposed mechanism for the first O-H---H-B to 0-8 topochemical
conversion in 47 ............................................................................................................... 97
Figure 3.14. X-Ray crystal structure of NaBH4-THEC illustrating the self-assembly in
dihydrogen-bonded dimers .............................................................................................. 99
Figure 3.15. Crystal packing in NaBH4oTHEC .............................................................. 100
Figure 3.16. Polarizing microscopic view of NaBH4~THEC crystal-to-crystal
decomposition: a) initial crystal; b) after 20 h at 160 °C. Scale bar = 100 um ............... 101
Figure 3.17. Powder XRD patterns for: a) dihydrogen-bonded complex 49; b) solid-state
decomposition product; 0) product of decomposition in the melt .................................. 102
Figure 3.18. Evolution of unit cell axes (a) and volume (b) during the solid-state
decomposition of NaBH4-THEC at 120 °C, monitored by powder XRD ....................... 103
Figure 3.19. Molecular model of the product from the solid-state decomposition of 49;
selected bond lengths [A] and angles [°]: B-O(1) 1.458, B-O(2) 1.482, B-O(3) 1.497, B-
O(4) 1.474, O(1)-B-O(2) 112.5, O(l)-B-O(3) 109.1, O(l)-B-O(4) 114.6, B-O(1)—C 123.7,
B-O(2)-C 122.3, B-O(3)-C 125.9, B-O(4)-C 125.6, N(l)-Na 3.003, N(2)-Na 2.714, N(3)-

Na 2.435, N(4)-Na 2.565, O(2)-Na 2.670, O(3)-Na 2.276, O(4)-Na 2.228 .................... 105

viii

Figure 3.20. X-Ray crystal structure of the major conformation found in the covalent
dimer obtained by decomposing 49 in DMSO; selected bond lengths [A] and angles [°]:
B-O(l) 1.502(2), B-O(2) 1.511(2), B-O(3) 1.450(2), B-O(5) 1.457(2), O(1)-B-O(5)
111.6(1), O(2)-B-O(5) 109.8(1), O(3)-B-O(5) 111.6(1), B-O(1)-C 121.7(1), B-O(2)-C
113.2(1), B-O(5)-C 122.8(1), N(1)-Na 2.519(1), N(2)-Na 2.540(1), N(3)-Na 2.565(1),
N(4)-Na 2.682(1), O(1)-Na 2.371(1), O(2)-Na 2.796(1), O(3)-Na 2.391(1), O(4)-H
082(2), O(4)H---O(2) 1.92(3), O(4)---O(2) 2.735(4), O(4)-H-O(2) 175.1(3) ................. 106
Figure 3.21. Conversion (0t) vs time curve for the solid-state decomposition of
NaBI-h-THEC at 160 °C, obtained using in situ "B MAS NMR ................................... 108

Figure 3.22. Lattice conversion (VO'VNO‘Vf) vs time curve for the solid-state

decomposition of N aBH4-THEC at 120 °C, obtained using powder XRD ..................... 108
Figure 3.23. X-Ray crystal structure of NaNCBH3-THEC ............................................ 110
Figure 3.24. or-Helix formed by NaNCBHyTHEC in the solid-state ............................ 111

Figure 3.25. X-Ray crystal structure of NaBH4-THPC; dihydrogen bonding parameters
(A, °): O(4)H---HB(1) (lower) 2.106, O(4)H---HB(1) (upper) 2.266, O(3)H---HB(1) 2.005,
O(7)H---HB(2) 1.927, O(8)H---HB(2) 1.676, O(2)H---HB(2) 1.904, O(4)H-H-B(1) (lower)
96.1, O(4)H-H-B(l) (upper) 88.7, O(3)H-H-B(1) 113.8, O(7)H-H-B(2) 110.7, O(8)H-H-
B(2) 120.2, O(2)H-H-B(2) 133.3, O(4)-H-HB(1) (lower) 162.5, O(4)-H-HB(1) (upper)
143.9, O(3)-H-HB(1) 153.4, O(7)-H-HB(2) 155.4, O(8)-H-HB(2) 153.0, O(2)-H—HB(2)
169.3 ................................................................................................................................ 113
Figure 3.26. X-Ray crystal structure of 9-(1’,4’,7’-triS-hydroxyethyl-

cyclen)methylanthracene ................................................................................................ l 15

ix

Figure 3.27. X-Ray crystal structure of 53; normalized dihydrogen bonding parameters
(A, °): O(1)H~-~HB 1.692, O(2)H---HB 1.661, O(3)H---HB 1.730, O(1)H-~H-B 114.5,
O(2)H---H-B 117.8, O(3)H---H-B 116.1, O(1)-H---HB 166.8, O(2)-H-~-HB 155.1, O(3)-
H---HB 154.6 ................................................................................................................... 116

Figure 3.28. Prospective multidentate ligands for the construction of extended

dihydrogen-bonded and covalent networks .................................................................... 118

 

 

 

l. Dihydrogen Bonding: A Literature Review

1.1. Introduction

Hydrogen bondingl occupies a prominent position in modern chemistry. It is
fundamental for molecular recognition and supramolecular synthesis, and holds a central
role in biology.2 Typically, this noncovalent interaction occurs between the positively
charged hydrogen of an A—H (A = O, N, halogen, C) proton donor, and the lone pair of an
electronegative element, or the It electrons of a multiple bond or aromatic ring,
representing the proton acceptor. Recently, an unusual type of hydrogen bonding, in
which a o M-H bond (where M is less electronegative than H) acts as the electron donor,

has attracted considerable attention.3

With strength and directionality comparable with those found in conventional hydrogen
bonding, this proton-hydride interaction, or dihydrogen bonding, can influence the
structure, reactivity, and selectivity in solution and solid state, thus finding potential
utilities in catalysis, crystal engineering, and materials chemistry. This review
summarizes the emergence and development of this topic, starting with the early
observations of H---H bonding, continuing with its comprehensive structural and
energetic description, and concluding with the implications of this interaction in
supramolecular synthesis, as well as its influence on reactivity and selectivity in both

fluid and condensed phases.

1.2. Structural and Energetic Characterization

A. Dihydrogen Bonding in the Main Group Hydrides. Chemists had intuitively
thought about proton-hydride interactions long before these associations were formally
categorized as hydrogen bonds. In 1934, Zachariasen and Mooney reported the crystal
structure of ammonium hypophosphite (NI-14+H2P02’),4 and noted that “the hydrogen
atoms of the hypophosphite group behave toward ammonium as if they were H' ions".5
Thirty years later, Burg suggested the presence of N-H---H3B interactions comparable to
hydrogen bonds in liquid (CH3)2NH-BH3, based on the perturbation of the N-H and B-H
bands in the IR spectrum.6 Titov et a1. explained the enhanced chemical reactivity of
aminoboranes toward H2 loss by the “close spatial arrangement of the oppositely-charged
hydrogen atoms”.7 However, the first to recognize this interaction as a true hydrogen
bond were Brown and coworkers in the late 1960’s.8 Based on a thorough analysis of the
IR Spectra of the boron coordination compounds of the type L-BH3 (L = Me3N, Et3N, Py,
Et3P) and Me3N-BH2X (X = Cl, Br, I) in the presence of proton donors such as MeOH,
PhOH and p-F-C6H4-OH in CC14, they proposed the formation of a novel type of
hydrogen bond in which the BH3 and BHz groups acted as proton acceptors, despite their
lack of lone pairs or 1: electrons. They measured the strengths of these interactions by
variable temperature IR spectroscopy, finding association energies in the range of 1.7-3.5
kcal/mol, comparable with moderately strong conventional hydrogen bonds.9 Similarly,
they inferred the occurrence of intermolecular N H---HB hydrogen bonding in
MCzNH°BH3 and (RNH-BH2)3 (R = Pr, Bu), from the temperature and concentration

dependence of their N-H Stretching absorptions in CC14.8

In a recent study, Epstein and coworkers confirmed the ability of boron hydrides
to act as proton acceptors in hydrogen bonds.10 They Studied the interaction of neutral
NEthH3 and P(OEt)3BH3 as well as ionic Bu4N+BH4' with different alcohols as proton
donors, by IR and NMR spectroscopy in CHzClz, C6H14 and C6D12, and concluded that
the properties of these unconventional OH---HB interactions are similar to those found in
classical hydrogen bonds. Their association energies were found to increase
proportionally with the proton donors’ acidities, being situated in the range 1.1-3.6 and
2.5-6.5 kcal/mol, for the neutral and ionic boron hydrides, respectively. Theoretical
calculations (RHF/6-3 16) confirmed the attractive nature of these proton-hydride
interactions.

The solution studies, however, cannot unambiguously establish whether these
unusual interactions involve the boron atom, the hydridic hydrogen, or the BH group as a

whole. The X-ray and neutron crystal structures of NaBH4-2HZO were recently

determined by Jackson et al., in order to probe the existence of O-H---H-B dihydrogen

bonding in the solid state, and provide a detailed structural description of it (Figure 1.1).ll

H_.
H ’0
H
l O
H B\ .uHr ‘H
H/ H
+3
O<H

Figure 1.1. O-H--~H-B dihydrogen bonds found in N aBH4-2H20 in the solid-state.

Three close H---H contacts of 1.79, 1.86, and 1.94 A were found, substantially
Shorter than the 2.4 A distance corresponding to twice the Van der Waals radius of a H
atom. The DH vectors clearly point toward the middle of the B-H bonds, suggesting
association with the o—bond electrons, rather than B or H atoms.

The same conclusion can be drawn from the systematic Cambridge Structural
Database (CSD) search done by Crabtree et al. for boron-nitrogen compounds.l2 The
twenty-six intermolecular N -H---H-B short contacts found in the range 1.7-2.2 A, for
which Crabtree suggested the term “dihydrogen bonds”, showed a strong preference for a
bent geometry, with NH---H-B angles typically situated between 95 and 120°, and N -
HmHB angles tending to be larger (150-170° in most of the cases). These side-on
structures were rationalized by the presence of negative charges on both B and H atoms,
which avoid the unfavorable alignment of the dipoles that would result from a linear N -
H---H-B arrangement. They also investigated theoretically the NH3BH3 dimer, whose C2
symmetrical geometry (Figure 1.2) optimized at the PCI-80/BBLYP level of theory
showed two identical HmH interactions, with contact distances of 1.82 A, and NH-nH-B
and N-H---HB angles of 98.8 and 158.7°, respectively, falling in the range found by the

CSD search.

Figure 1.2. C2 isomer of the NH3BH3 dimer.

The calculated dimerization energy of -12.1 kcal/mol corresponds to 6.1 kcal/mol per N-
H---H-B interaction, which, as suggested by Crabtree, could account for the strikingly
higher melting point of aminoborane (+104 °C) relative to the isoelectronic ethane (mp
-181 °C).12

A similar head-to-tail arrangement was also found by Cramer and Gladfelter in
their theoretical study of the (NH3BH3)2 dimer.l3 However, using HF, DPT, or MP2
methods, they found a C2,, symmetrical structure as the global minimum (Figure 1.3),
which lies only 0.2 kcal/mol lower in energy than the C2 isomer reported by Crabtree et
a]. This geometry allows the formation of bifurcated dihydrogen bonds with H-H
distances of 1.990 A, and NH---H-B and N-H---HB angles of 88.6 and 144.8°,
respectively, as calculated at the MP2/cc-pVDZ level. The association energy obtained at

the same level of theory is —15 .1 kcal/mol.

HEN/H: 2 1 ‘, HEB/H
l l
H/Bc'lf“ f I :H’N§:

Figure 1.3. C 2,. isomer of the NH3BH3 dimer.

For comparison, the crystal structure of NH3BH3 recently determined by Crabtree and
coworkers by neutron diffraction shows a packing that results in three short
intermolecular N-H---H-B interactions, with the shortest one exhibiting a H-H distance of
2.02 A and values for the NH---H-B and N-H---HB angles of 106(1) and 156(3)°,

respectively (Figure 1.4).14 Again, the N-H vectors point toward the middle of the B-H

bonds, suggesting that the o-bond as a whole represents in fact the proton acceptor

partner in these interactions.

H
H H
H H
H 5“”,
’/H H
H H H

Figure 1.4. The shortest N-H---H-B contacts found in NH3BH3 in the solid-state.

Further insight into the nature of the N-Hu-H-B interaction was provided by Popelier,
who applied the “atoms in molecules” theory on the same (NH3BH3)2 dimer, and
concluded that this interaction can indeed be classified as a hydrogen bond.”
Intermolecular N-H---H-B interactions have also been described by Noth et al., who
recently reported the crystal structure of (CH3)2NH-BH2-N(CH3)2-BH3, which self-

assembles into dihydrogen-bonded dimers, as illustrated in Figure 1.5.16

Me_ ll-l =1 /Me
-. /B'”HH ' 'iiiiH’Nx
MerN ‘H‘ 8H2
I I .IMe
HZBTN’H” '3 -------- Abs/mm
Me". \ H

Figure 1.5. Dihydrogen-bonded dimers found in (CH3)2NH-BH2-N(CH3)2-BH3 in the

solid-state.

The N-H---H-B dihydrogen bonding can also form intramolecularly, as found in
the crystal structure of the 2’-deoxycytidine-N(3)-cyanoborane ( 1), which shows a H---H

close contact of 2.05 A.17

”H H
9 w
NZC’P\N/
H I
0 CHZOH
H'O'H
H
H OH
1

Intramolecular C-H---H-B close contacts are present in the aminoboron hydrides 2-6,
which could be responsible for the stabilization against disproportionation in these

complexes. 18

HH H ----- H H ----- H
2 3 4
o .
ti / H Fa
N..'B/N / /N\ N\CH
\ N—B Y 3
,H H ....... H
H
5 6

Their X-ray crystal structures show multiple H-H distances below 2.65 A, which was
considered the threshold intermolecular distance for H-H interactions in this study. The
heterocyclic rings adopt almost coplanar orientations relative to the B-H bonds, thus
maximizing the intramolecular HnH associations. The relatively small H-C-N exocyclic
angles next to the B-H bonds in some of these complexes, compared to the free
heterocycles, also suggest attractive interactions between the protonic hydrogens on the
(it-carbons and the hydridic BH hydrogens. In solution, the formation of similar
dihydrogen bonds was explored by NOE experiments. For instance, 2 adopts a
conformation comparable to the one found in the solid state, allowing again short C-
HmH-B contacts.

Intramolecular C-H---H-B dihydrogen bonds were also demonstrated to play a

decisive role in controlling the conformation of the heterocyclohexane-borane adducts 7-

10.”
(9H3 (EH3
SvN\BH S N\
Ls\/ 3 pm BHZc'
7 8
(EH3 $H3 (EH3
H B..-°N\/N\BH H c,.-'N\/N\BH
3 ps\/ 3 3 4w 3
H30
9 10

The BH3 groups are always found in the equatorial position in these complexes, which
appears to be the result of favorable attractive interactions between the hydridic B-H

hydrogens, and the positively charged H atoms of the a—CHZ groups (Figure 1.6).

 

Associations with the OH hydrogens of the N-CH; group also seem to stabilize these
structures, as indicated by the short H-H distances and the decrease of the H3C-N-BH3

angles relative to the H3C-N-CH3 angle in the (CH3)2N+ derivative.

Figure 1.6. C-H---H-B close contacts in 7-10.

The hydrides of the heavier group 3 elements are also capable of forming
dihydrogen bonds. Thus, in 1994, Raston and coworkers provided X-ray crystallographic
evidence for an intramolecular N -H~--H-Al interaction in the alane-piperidine adduct 11.20
The H-Al—N-H unit has an eclipsed conformation in the solid state, allowing the two
oppositely charged hydrogen atoms to approach to 2.31 A, in direct contrast to the
previously reported structures of aminoalanes, which are known to exhibit a staggered

conformation about the Al-N bond. This arrangement, Raston noted, represents an

intermediate prior to H2 evolution, to form an amidometal species.

Computational studies at the MP2/cc-pVDZ level, by Cramer and Gladfelter,l3
revealed a staggered, C3v symmetrical geometry for NH3A1H3, which upon dimerization
forms a C2 symmetrical structure (Figure 1.7) that contains two short intermolecular N-
H-o-H-Al hydrogen bonds, with an H-H separation of 1.781 A, and N H-nH-Al and N-
H---HA1 angles of 119.4 and 172.0°, respectively. The dimerization energy calculated at
the same level of theory is —11.8 kcal/mol, which corresponds to about 6 kcal/mol per N-

Hu-H-Al dihydrogen bond.

- .--H

- H 1' __
H~T/ H‘Al/H
/Al 3“ [Iv -~H
H ~H ...... H/ ‘H

Figure 1.7. Calculated structure of the NH3A1H3 dimer.

Cyclotrialumazane ((NH2A1H2)3) was found theoretically21 (MP2/VDZ) to prefer the
twist-boat over the chair conformation in the gas phase, by as much as 2.8 kcal/mol, due

to favorable electrostatic N-HsflflH-Al flagpole interactions (Figure 1.8).

Figure 1.8. Calculated structure of (NH2A1H2)3.

10

In the solid state, however, it is likely that it will adopt the chair arrangement for a more
efficient packing, as also observed in the analogous boron and gallium systems. Based on
this consideration, the chair conformer was used for the calculation (RHF/cc-pVDZ) of
the preferred geometry in the [(NH2A1H2)3]2 dimer. As shown in Figure 1.9, the resulting

C 3,, symmetrical structure exhibits six short H-H contacts, and the enthalpy of

dimerization is predicted to be —9.9 kcal/mol at the MP2/cc-pVDZ//RHF/cc-pVDZ level.

\‘5 x \H 1“

'- A'l‘.‘ ’H “‘\ AIH

H 4‘ ‘Al F ‘H
1“ , \HN
H H \H

Figure 1.9. Calculated structure of the (NH2A1H2)3 dimer.

The next element in group 3, gallium, can also be involved in dihydrogen

bonding, as Gladfelter’s neutron diffraction crystal structure of cyclotrigallazane (12)

demonstrates.”

ll

In the solid state, 12 forms an (at-network, by participating in four N-H---H-Ga
intermolecular interactions, with H-H distances of 1.97 A (Figure 1.10). The observed

NH---H-Ga and N-H—--HGa angles are rather close, with values of 131 and 145°,

respectively.

H H \ /H
H\ /H H Ga/ H\ Ga H
H\/Ga H \/\/H N/\

N H

/N\ L/H ------ H/ X NXH"'1"|"H<H\/‘c§a’/ /H
H / ,\/ /H ------ H Ga Ga/ / Ga
H Ga Ga /\/\ \/\
H/\/\ /N\H /N\n

”\n H H H

Figure 1.10. Self-assembly of cyclotrigallazane in the solid-state.

The strength of these dihydrogen bonds was estimated by theoretical calculations on the
[(NHzGaH2)3]2 dimer. As in the aluminum analog, the monomer prefers the twist-boat
conformation by 2.6 kcal/mol, thus favoring intramolecular HmH interactions involving
the oppositely charged H atoms from the flagpole positions. However, for direct
comparison with the solid-state structure, the chair conformation was considered for the
geometry optimization of the dimer. The highest dimerization energy was found for the
CS symmetrical structure illustrated in Figure 1.11, from which an interaction energy of

about 3 kcal/mol could be estimated for each N-H~--H—Ga dihydrogen bond.

12

HGa—/—N
\ / \
[Ga—H H
H /N
/ \
Ga—NH\
,/ H ': h'
\‘\“ T i
\‘ . H
H GaH H
\Z—GbN‘éGg

Figure 1.11. Calculated structure of the (NHzGaH2)3 dimer.

In the case of the (NHgGaH3)2 dimer, theoretical calculations by Cramer and Gladfelterl3
predicted a C2 symmetrical geometry similar to the one found in the A1 analog, with N-
H---H-Ga dihydrogen bonds of approximately 5 kcal/mol in strength.

Formation of dihydrogen-bonded complexes by other main group hydrides such
as Lil-I, BeHz, or the recently discovered XeHz, has been investigated theoretically by a
number of researchers.22 A theoretical study (MP2 and B3LYP) of the dihydrogen-
bonded complexes between the hydrides LiH, NaH, BeHz, MgHz, CH4, SiH4, Gel-I4,
SnH4, and hydrofluoric acid, reported by Grabowski, demonstrated the existence of direct

correlations between the H--H distances and the H-bonding energies.23 Also, the HmH

l3

separations have been found to be inversely proportional to the F-H bond lengths, as is

seen in conventional O-H---O or N-H---O hydrogen bonds.

B. Dihydrogen Bonding in Transition Metal Hydrides. In 1986, Milstein and
coworkers reported the X-ray crystal structure of the iridium hydride complex 13, in
which they noted an unusually small Ir-O—H angle of 91° and an eclipsed H-Ir-O-H
conformation, indicating an attractive H---H interaction.24 However, the H-H distance of
2.441 A was too long for a hydrogen bond. The later neutron diffraction study of the
same compound revealed a shorter H-H distance of 2.40 A, but a wider Ir-O-H angle of

104.40.25

/H
(PL;

'17? ‘H L = PMea
L

13

In 1994, Berke and collaborators suggested that the formation of M-H&---H5+-X
interactions might precede the protonation of transition metal hydrides to yield

dihydrogen complexes (Scheme 1.1).26

LnM—H + H—OR ——-* LnM—H---H—OR ——>

+ H

LnM:::OR ——-> LnM<—-|_I_' 'OR

Scheme 1.1

14

The first unequivocal evidences of dihydrogen bonding involving a transition
metal hydride came independently from the groups of Crabtree and Morris in 1994. The
X-ray crystal structure of 14 determined by Crabtree et al. showed an unexpected
tautomerization of the amide group from the ligand into the iminol form, facilitating thus
the interaction between the hydridic Ir-H and the OH proton.27 While the exact H-H
distance could not be determined in the solid state due to the failure to locate the Ir-H
hydrogen, 1H NMR T1 relaxation time measurements in solution gave a HmH contact
distance of about 1.8 A. The interaction seems to have some covalent character, as

demonstrated by the observed coupling of 3 Hz between Ir-H and O-H hydrogens.
+ _
—] SbF6

(x = PPha)

 

An even higher coupling of 5.5-5.6 Hz was found in 15 (Y = H, Cl, Br, I), in which the
HmH contact distance was estimated around 1.7 A (Y = Cl), based on T1 relaxation time
measurements.28

a
mail? .0
Y—lr-H’

I'l'l\PPh3

15

15

These H---H interactions are fairly strong, competing favorably with the conventional
Y---H hydrogen bonds, as demonstrated by the predominance of the O-H---H-Ir
dihydrogen-bonded rotamers over the opposite O-HmY-Ir hydrogen-bonded ones. A
quantitative estimation of the strength of these unconventional hydrogen bonds involving
iridium hydride complexes was elegantly made by Crabtree’s group in the analogous 2-
aminopyridine complexes 16 (Y = H, F, Cl, Br, 1, CN, CO). They measured the rotation
barrier for the NH; group, which represents the sum of the HuH interaction energy and
the intrinsic rotation barrier around the C-N bond in the ligand. The strongest N-H---H-Ir
hydrogen bond (5.0 kcal/mol) was found for the case where Y = H, while the trans Y
ligand tends to weaken the interaction in the order: F > C1 > Br > I > CN > CO > H,

which can be rationalized by their decreasing electronegativity in the same order.

Morris and coworkers demonstrated the presence of Hull interactions in the
iridium hydride complex 17 by X-ray diffraction in the solid-state, and NMR
spectroscopy in solution.29 The crystal structure of 17 indicated a close contact between
the pyridinium protons and the Ir—H hydridic hydrogens, which unfortunately could not
be located precisely from the electron density difference maps. Nevertheless, their lH

NMR data provided clear evidence for N-H---H-Ir hydrogen bonding in CD2C12, with a

16

H-H contact distance of about 1.75 A, calculated from the observed T1 relaxation times of

the protonic and hydridic hydrogens.

S .i’ L H
_ L/IIr\H”
\ /N-H--~H

l7 (L=PCY3)

18 (L=PH3)

Theoretical calculations by Hoffmann et al.223 on the model complex 18 confirmed the
attractive H~--H interaction, and concluded that its nature is mostly electrostatic.
Interestingly, when THF was used as a solvent, the dihydrogen bonds were switched off
in 17, presumably by the formation of conventional N—HmO hydrogen bonds with the
solvent.

Bifurcated N -H---H(Ir)---H—N dihydrogen bonds were also detected by Morris et
al. in 19 and 20 by X-ray crystallography and NMR spectroscopy, and their H-H contact

distances were estimated around 1.80 and 1.86 A, respectively.30

 

l7

They also reported an interesting bifurcated Ir-H---H(N)---F-B interaction in complex 21,

in which the N-H proton is shared by a hydridic Ir-H hydrogen and a conventional B-F

electron donor from the BF4' counterion.31

PPhg
" Ir " N—
PPha/H\H~H
' / F.
/ BF3
21

The occurrence of intermolecular X-H---H-M dihydrogen bonds was first
documented by Crabtree and his collaborators, with the neutron diffraction crystal
structure of the rhenium polyhydride complex 22, which exhibits a three-center
interaction between two Re-H hydridic hydrogens and the N -H proton from an indole
molecule of crystallization.32 The geometrical parameters for the two HnH contacts, i.e.
H-H distances of 1.73 and 2.21 A, and strongly bent NH---H-Re angles of 119 and 97°,
respectively, fall in the range expected for dihydrogen bonding. The strength of the
interaction was estimated around 4.3 kcal/mol, from the shift of the N-H stretching band

in the solid-state IR spectrum of 22, relative to free indole.

/
LLH\ //L¥ H

H HFI

22

18

Theoretical calculations using the DFI‘ method on a [ReH5(PH3)3]-NH3 model confirmed
the attractive nature of the interaction, and predicted a similar three-center H-bond with
H--H distances of 1.92 and 2.48 A, and an interaction energy of 8 kcal/mol in the gas
phase. The intermolecular N-H---H-Re hydrogen bonding appears to be general, as
demonstrated by the X-ray crystal structure of the analogous complex 23, in which the
role of the proton donor is played by an imidazole molecule.32b’ 33 The H-H distances
could not be satisfactorily determined due to the failure to locate the N-H hydrogen from
the disordered imidazole. However, the strength of the interaction (5.3 kcal/mol,
estimated by IR spectroscopy) is greater than in the similar complex 22, as expected

considering the higher acidity of imidazole relative to indole.

/ N
H H\ /N__.//
H H
NJ
H
23

Other weak acids such as 2,4,6-Me3C6HzOH, 2-tBu-6-MeC6H30H, pyrrole, PhNHPh,
PhNHBn, PhNHMe, also associate with ReH5(PPh3)3, as shown by the IR of the thin
films obtained by evaporation of CH2C12 solutions containing a 1:1 mixture of the
polyhydride and the proton donor.34 The ~AH° of the interactions, evaluated from the
observed shifts of the NH or OH bands, were found to generally correlate with the
acidities of the proton donors, and vary between 3.0 and 5.6 kcal/mol. Replacing the

hydridic partner with the WH4(PMePh2)4 complex resulted in slightly weaker dihydrogen

l9

bonds of 1.1-5.2 kcal/mol. That these interactions involve primarily the hydridic
hydrogen and not the d2 non-bonding electron pairs of the Re or W metals was
demonstrated by the analogous d0 complex ReH7(dppe) (dppe = thPCHZCHzPth), that
also associated with the same proton donors, with -AHO in the range 1.3-4.7 kcal/mol.

In solution, the N -H---H-Re interactions were studied by Crabtree et al. using UV-
vis spectroscopy.35 Thus, the ReH5(PPh3)2L (L = pyridine, 4-picoline, 4-
dimethylaminopyridine, and 4-carbomethoxypyridine) hydrogen bond acceptors interact
with indole as the hydrogen bond donor, showing free energies of association, -AG, of
3.8-5.0 kcal/mol.

In a thorough analysis, Epstein, Berke, and coworkers surveyed dihydrogen
bonding in solution in the tungsten hydride-alcohol complexes 24.36 Using IR and NMR
spectroscopies, they ruled out hydrogen bonding to the CO or NO groups, and proved the
exclusive formation of the unconventional O-H---H-W interactions. As expected, the
strengths of these dihydrogen bonds increase with the donator abilities of the ligand L
(PMe3 > PEt3 > P(OiPr)3 > PPh3) and are directly proportional to the acidities of the
proton donors phenol (PhOH), hexafluoro-2-propanol (HFIP), and perfluoro-2-methyl-2-
propanol (PFI‘B) (PFI‘B > HFIP > PhOH). Their magnitude was estimated in the range
4.1-6.9 kcal/mol, from the observed shifts in the corresponding Von bands, as well as
from the variation of the association constants K with temperature. In addition, NMR
experiments (6 shifts, NOE, and T1 relaxation times), all supported the formation of O-

H---H-W dihydrogen bonds, with H-H contact distances as short as 1.77 A in the case of

HFIP. A linear OH---H-W orientation was arguably suggested for these interactions, in

20

sharp contrast to the previously established propensity of dihydrogen bonds for a strongly

bent geometry.

CO

\

0N7w'L-H-«H—on

0‘3 L L = PMe3, PEtg, P(iPr0)3. PPh3

R = Ph, CH(CF3)2, C(CF3)3
24

In an analogous series of rhenium hydride complexes (25), the same two research groups
proved the occurrence of intermolecular O—H---H-Re dihydrogen bonding in solution.” 38
When PFI'B was used as proton donor, interaction energies between 4.5 and 6.1 kcal/mol
were calculated from the observed v0" shifts in the IR spectra in hexane. In toluene,
however, the AH values, derived from variable temperature N MR spectroscopy, are
smaller by about 3 kcal/mol, apparently due to competitive O-H-o-n interactions with the
solvent.38 The H---H contact distances calculated from T1 relaxation times range between
1.78 and 1.94 A. The phosphine ligand appears to have an important influence over the
regioselectivity of the H-bonding formation. Thus, while in 25a interaction with one of
the hydridic hydrogens is preferred, the NO group competes more and more effectively
for the proton donor as the bulk of L increases, to the point where only O-H---ON
hydrogen bonds are observed for 25¢. However, DFI‘ calculations on a
ReH2(CO)(NO)(PH3)2.H20 model indicated that the H---H interaction is energetically

preferred by about 3.0-3.5 kcal/mol.37 Also, a stronger interaction was predicted with the

Re-H hydride trans to the NO group (dim = 1.49 A) compared to the Re-H trans to CO

21

(dim = 1.79 A), and confirmed experimentally by the high regioselectivity displayed by

both PFI'B and HFIP alcohols toward the former, as shown by NMR spectroscopy.38

L
l v”
ON—/—Rle——H
CO L
25

L = PMe3 (a), PEt3 (b), PiPr3 (c)

The experimental results obtained by Epstein and Berke on the intermolecular
dihydrogen bonding in solution were complemented by the theoretical work of Scheiner
et al. on the Mo and w hydride complexes 26.39 Their HF/3-21G and DFT (BBLYP,
BLYP, B3PW91) calculations confirmed that the HmH interactions are favored over

conventional hydrogen bonding involving the N 0 group.

L M = M0, W
, l .~‘CO L= PH3, NH3
'- 7'\|"—H\.\ L’ = NO, Cl, H
CO L “R R = F, OH, H20+
26

The dihydrogen bonds in 26 become stronger and shorter with the increases in the
donating ability of the cis-ligand or the acidity of the proton donor, consistent with
experiment. However, the strongly acidic H3O+ induces complete proton transfer,

resulting in the formation of an nz-Hz dihydrogen complex. The H-H-M angles are

22

strongly bent in all the optimized structures, as illustrated in Figure 1.12 for the
representative complex 26a. For this dihydrogen-bonded system, a 2.6 kcal/mol

destabilization energy was calculated for a linear F-HmH-Mo orientation.

PH3
(co
0N7Mo‘é—H
00 l
PH3 H
F
263

Figure 1.12. Calculated structure for dihydrogen-bonded complex 26a; < FHmH-Mo =
127°, < F—HmHMo = 174°.

As with the boron hydrides, C-H sites may participate as protonic partners with
transition metal hydrides. Recent X-ray structural studies and CSD surveys confirmed the
existence of intra-40 as well as intermolecular“ C-HmH-M close contacts in transition
metal hydrides. A large number of these examples were observed in complexes
containing R3-X(Ph)xP (x = 1-3) ligands, in which one or more ortho C-H bonds point
toward the M-H hydridic hydrogens (Figure l.13).4°°°’d While the HmH distances and M-
HmH angles in these complexes were found to fall essentially in the same range as
observed for the more “conventional” dihydrogen bonds involving N—H or O-H proton
donors, the C-HmH angles usually tend to be smaller, due to the inherent constraints

imposed by the chelation.4oa However, caution is advisable in interpreting some of these

23

C-HmH-M short contacts, as steric compression by bulky ligands or packing forces may

make a significant contribution to the observed HmH close proximities.40b

Figure 1.13. Generic representation of C-HmH-M dihydrogen bonds in complexes of

transition metal hydrides with R3-X(Ph)xP ligands.

The comprehensive analysis of the manganese hydride complex 27 carried out by
Brammer and collaborators provided convincing evidence for an intramolecular C-HmH-
Mn dihydrogen bond.42 Their combined low-temperature neutron and X-ray diffraction
study revealed a short intramolecular C-HmH-Mn contact of 2.10 A, with HmH-Mn and
C-HmH angles of 126.5 and 129.0°, respectively, and an essentially coplanar relative
orientation of the Mn—H and OH bonds ((1 Mn—HmH-C = 07°). The experimental atomic
charges found for the Mn-H hydridic and ortho CH protonic hydrogens, of —0.40 and +
0.32, clearly indicate an attractive electrostatic interaction, whose magnitude was
calculated to be 5.7 kcal/mol. Moreover, topological analysis of the charge density using
the “atoms in molecules” theory unequivocally supported the existence of a moderately

strong intramolecular C-HmH-Mn hydrogen bond.

24

CO

I __00
00—Mrlm\'—H\
‘H
.. Pco
0
Ph
27

1.3. Self-Assembly of Extended Dihydrogen-Bonded Systems.

In the previous section it was demonstrated that dihydrogen bonding is a
significant interaction, with energetic, electronic, and spectral characteristics, as well as
directionality, comparable with those found in conventional hydrogen bonding. A direct
consequence of this similarity is that dihydrogen bonds, like traditional H-bonds, could
find potential utility in crystal engineering and supramolecular synthesis. That these
HmH noncovalent interactions are indeed capable of controlling crystal packing was
already suggested by the crystal structure of cyclotrigallazane (12), which self-assembles
into an extended dihydrogen-bonded tut-network in the solid state (Figure 1.10), as
demonstrated by Gladfelter’s group.2|

The osmium polyhydride complex [K( 1-aza-18-crown-6)][mer-
OsH3(CO)(iPr3P)2] (28) synthesized by Morris et al. also forms polymeric one-

dimensional chains in the solid-state (Figure 1.14).43 The N-HmH-Os dihydrogen bonds

are however complemented by COmK+ interactions in the assembly of the chains.

25

Mm

. . o
( x fifPiPr3
‘ ‘~K’+ _____ _N/H---H—osI—co-~

0" _____

K )iPrgP/H

D0 n
28

Figure 1.14 Self-assembly of dihydrogen-bonded complex 28 in one-dimensional chains.

When 1,10-diaza-18-crown-6 was used to complex K+, similar chains held together
exclusively by N-HmH-M dihydrogen bonds (Figure 1.15) could be assembled from the
osmium, ruthenium, and iridium anionic polyhydrides [MHX+3(iPr3P)2]' (M = Os (3), Ru
(b), x = 2; M = Ir (c), x = 1) (29).“ 45 The intermolecular HmH distances in these
complexes were estimated from their X-ray crystal structures to be in the range 1.8-1.9 A.
The NH bands in their solid-state IR spectra are broadened relative to the free diaza-
crown ether, and shifted to lower numbers by 96, 107, and 132 cm'l, for the Os, Ru and Ir
complexes, respectively, in accord with the increasing basicity (and hydridicity) of these
polyhydrides in the same order. The observed shifts correspond to N-HmH-M interaction
energies of about 3 kcal/mol.4S Weaker C-HmH-M hydrogen bonds can also be involved
in the supramolecular association of the polyhydrides 29a,b with K(THF)(1-aza-18-

crown-6), in which alternating NHmH-M-HmHN and CHmH-M-HmHC units lead to the

formation of zigzag chains, as illustrated in Figure 1.16.45 The NHmHOs and NHmHRu

26

separations were estimated at 1.7 A, while the weaker CHmHOs and CHmHRu

interactions exhibit longer H-H contact distances of 2.2 and 2.1 A, respectively.

/—\ /—\
0. .0 Q. .0
{ ‘ H PiPr3 ( V
— H—N— ,Kf -N—H -H—M‘—H---H—N---—--‘K;+--- - —H -
x /' x
O \O O
\__/ \_/
29 M=OsH,RuH,lr

Figure 1.15. Self-assembly of dihydrogen-bonded complexes 29-[K( l ,10-diaza-18-

crown-6)] in one-dimensional chains.

0. ‘
.PIPr ( 03
H s ..... .K? ------ o M = Os, Ru

ingP H H

Figure 1.16. Self-assembly of dihydrogen-bonded complexes 29a,b-[K(THF)( l-aza-18-

crown-6)] in one-dimensional chains.

27

Finally, a similar polymeric chain can be assembled exclusively by weak C-HmH-M
interactions (Figure 1.17), as illustrated by the crystal structure of 29b-[K(THF)(18-

crown-6)].45 In this case, the observed C-H---H-Ru contact distance is 2.2 A.

o‘ \
\ ‘. (3 +1
i4 [ll ..uPiPl'3 ( V
\Ru'—H\ 4 H
iPrapfi/ \H “H O- ---- (K; ------ O x
Fad” )
\__/O

Figure 1.17. Self-assembly of dihydrogen-bonded complexes 29b-[K(THF)(18-crown-

6)] in one-dimensional chains.

As part of early endeavors in the Jackson group directed toward the structural
characterization of dihydrogen-bonded systems involving anionic borohydrides, the
crystal structure of guanidinium borohydride was explored.46 In the solid-state, this salt is
organized into extended tapes, in which the alternating BH4' and C(NH2)3+ ions are

connected by multipoint dihydrogen bonds, as illustrated in Figure 1.18.

a- .-—-P1 fl--- -
H\N’H \3/ “Hm/H” Ike/w“
-’ \ .5 \
\ |+ ,H‘,'H‘ I H ”H
~H\ C\ H ““‘s C”
l I
H H H H

Figure 1.18. Self-assembly of guanidinium borohydride in one-dimensional extended

tapes.

28

1.4. Dynamics of Dihydrogen Bonds

Numerous studies have now established that dihydrogen bonding is an important
and general interaction involving element-hydride 0' bonds, and its geometrical and
energetic features have been described in great detail. The significance of these unusual
hydrogen bonds extends, however, beyond their fundamental aspects. With their
substantial strength and directionality, they can be used to control reactivity and
selectivity of chemical reactions, in the same time finding a place alongside conventional
hydrogen bonding in the supramolecular chemists’ arsenal of noncovalent interactions.
However, what makes dihydrogen bonding particularly interesting is the special
reactivity conferred by its peculiar nature. It has been recently demonstrated that HmH

bonds have a role in the formation of dihydrogen Ill-H2 complexes and the reverse

heterolytic splitting of H2, as well as o-bond metathesis (Scheme 1.2).

M—H + H—A —-+ M—H---H—A :

 

HT H
M~-I A' ‘——2 M—A
+H2

Scheme 1.2

The Ir-Hb bond in 30 has been found to be activated by dihydrogen bonding for a
number of reactions.47 Thus, the hydridic and protonic hydrogens l-la and Hb involved in

the HmH interaction can interchange relatively easily whereas the non-interacting Hc is

29

exchanged much more slowly with H8 and Hb. The AH‘ for the Ha/Hb exchange has been
estimated by variable temperature NMR spectroscopy at around 14-16 kcal/mol and
found to go down as the R group becomes more electron-withdrawing, consistent with a
mechanism involving proton transfer from the OH group to the Ir-Ha bond, to give an 772-
H2 intermediate complex (Scheme 1.3). Rotation of the H2 ligand in this complex and
transfer of the proton back to the oxygen completes the exchange. When the reaction was
performed in the presence of benzonitrile, the H2 ligand could be displaced by PhCN, in a

rate-limiting step.

 

30

L = PPha
R = Me, n-Bu, p-tolyl, Ph, p-FCsH4. 3.4-CeH3F2

Scheme 1.3

Alternatively, H2 elimination from 30 by heating in a sealed tube at 80 °C yielded the
chelate complex 30a, in a o-bond metathesis reaction. The initial complex 30 could be

recovered by exposure to H2 in CH2Cl2 at room temperature, via the isomer 30b (Scheme

1.4).

30

   

Scheme 1.4

Similarly, hydrogen scrambling between Ha and H, is facilitated by dihydrogen bonding
in complexes 31, apparently via an nz-H2 intermediate.47 H2 loss at room temperature was
also observed, with the formation of 31a in a first order reaction, with measured
activation parameters AH‘ and AS" of 14 i 2 kcal/mol and —32 :t 6 eu, respectively
(Scheme 1.5). The highly negative activation entropy suggests an associative process
with a highly ordered transition state.

The first direct observation of a dynamic equilibrium between a HmH bonded system and
an If—H2 complex resulting from proton transfer along a dihydrogen bond was made by
Chaudret and coworkers, using NMR spectroscopy.48 Thus, in the presence of phenol, the

ruthenium hydride complex RuI-I2(dppm)2 (32) exists as a mixture of dihydrogen-bonded

31

cis and trans isomers in benzene or toluene solutions. The trans isomer is also involved in
a dynamic equilibrium with the dihydrogen complex 32a, which lies 17 kcal/mol lower in
enthalpy than 32-PhOH (Scheme 1.6). It was proposed that the reversibility of the process
originates in the strong dihydrogen bonding between 32 and phenol. In the presence of
the more acidic hexafluoroisopropanol, the corresponding dihydrogen complex 32a

further reacts by H2 loss, to ultimately give 32c via 32b.

 

/ I
\ R -H2
N N’ —*
PEP-.2 | ,Ha Ph3P
Hd—lr-Hfi _
l\ Hd l
Hc PPh3
31 31a
R = H, Ph
Scheme 1.5

DFT calculations by Scheiner et al. on a HOH--.H2Ru(PH2CH2PH2)2 model yielded AE
and Ali“ values of -10.7 and 10.0 kcal/mol, respectively, in qualitative agreement with
experiment.49 However, when the stronger proton donor HF was used in the calculations,
no F—H---H-Ru adduct could be identified, and the system evolved directly toward the
dihydrogen complex, which in this case lies 23.8 kcal/mol lower in energy than the

separated HF and ruthenium hydride complex.

32

H-OR /
I H I P-
EPH" | ..IP/ >ph I ”.H "H—OR
"'Ru "" _ a "" Fl
/P/| \PZ \p/ |u\H
I H | /\/P—-
trans-32 \

,.

\l HTH I 7 I 0" I

(P """" Ru """ P; QR —>'H2 2PM“; ""P/
. u..

/ /| \P\ /P/| \PZ
H l I H |
32a 32b

1+ROH
/

p...
>p... I NOR
“Ru"
>P/I \OR
P...—

\
32c

Scheme 1.6

33

Using in situ IR and NMR spectroscopy, Epstein et al. studied the proton transfer
in the dihydrogen-bonded complexes between (triphos)Re(CO)2H (33) and phenol,
tetrafluoroboric acid (HBF4-0Me2), chloroacetic acid (ClCH2CO2H), hexafluoro-2-
propanol (HFIP) or perfluoro-2-methyl—2-propanol (PFTB), as proton donors, at 200-260
K (Scheme 1.7).50 The 172-H2 complexes 33a were found again to be thermodynamically
more stable than their HmH bonded precursors. Higher temperatures induced H2 loss with

the formation of the covalent products 33b.

 

Scheme 1.7

In the case of H2Re(CO)(NO)(PMe3)2 + CF3COOH, the dihydrogen-bonded complex
coexists in equilibrium with the corresponding nz-H2 complex, which loses H2 upon

heating.5 l Similarly, the dihydrogen-bonded adduct

34

(CF3)2CHOH---HW(CO)2(NO)(PMe3)2 was found to undergo proton transfer in the rate-
limiting step with the formation of an unstable nz-H2 complex, which subsequently
eliminates H2 to form a W-OR covalent product.
The proton transfer in the dihydrogen-bonded complex
(CF3)3COH---HRu(Cp)(CO)(PCy3) was also studied by Epstein et al. using in situ IR
spectroscopy, and a relatively high barrier of 15 kcal/mol was found for this process.5 I To
model this reaction, Scheiner et al. used the HRu(Cp)(CO)(PH3) ruthenium hydride
model, which was allowed to interact with H3O+, CF3OH or H2O, representing strong,
moderate, and weak proton donors, respectively.52 While in the first case spontaneous
transfer of proton with the formation of a corresponding hydrated 02-H2 complex was
observed, the other two weaker acids did not transfer the proton at all, suggesting that the
activation barrier for this process is largely determined by the proton donor ability of the
acidic partner. The critical role of the proton donor acidity has also been recognized
recently by Lau et al., who concluded that strongly acidic conditions give nz-H2
complexes, while weakly acidic conditions favor dihydrogen-bonded species.53
In the ruthenium polyhydride complex 34, Chaudret et al. noted a substantial
increase of the H-H coupling Jab upon formation of dihydrogen bonding with various
proton donors in toluene, which was tentatively explained by the decrease of the electron
density on Ru, caused by the partial charge transfer from the metal hydride to the
hydrogen bond donor.54 When the CDC12F/CDF3 (2:1) solvent system was used instead,
proton transfer with the formation of the (Cp*)(PCy3)RuH4 complex was observed at low
temperatures, as a result of the unusual property of the Freon mixture to strongly increase

its dielectric constant upon cooling, assisting thus the protonation of 37 .5 5 Interestingly,

35

the reactants can be mixed at room temperature, but the proton transfer only occurs when
the temperature is sufficiently lowered to induce an adequate increase in the dielectric
constant of the solvent. The same behavior was observed in CD2Cl2, a solvent which is
also known to have a strong temperature-dependent dielectric constant.
CP\ ..Hb
Hui—Ha

34

According to Lau et al., intramolecular N-H~--H-Ru dihydrogen bonds also appear
to mediate proton transfer and subsequent formation of N-Ru bonds in 35, as illustrated
in Scheme 1.8.56 HID exchange of both protonic and hydridic hydrogen atoms with D20
strongly suggest the existence of nz-H2 intermediate species in equilibrium with the HmH
bonded complexes. H2 loss with the formation of a Ru-N bonded chelate structure is
facile in 35b, and the reverse Ru-N bond hydrogenolysis can be done at 60 °C under 60
atm (Scheme 1.8). This system was found to catalyze the reduction of CO2 to formic acid,
although with low yields.56 The heterolytic splitting of the H2 ligand is believed to be a
crucial step in the proposed mechanism, which is depicted in Scheme 1.9. The only NMR
detectable metal-containing species throughout the reaction is 35, suggesting that the
insertion of CO2 into the Ru-H bond is the rate-determining step, as also supported by

recent theoretical calculations on a similar system.57

36

 

 

\
\\'R till
Phi l‘k ,N\ Me thP“ ;u\ H N ...Me
Me
PPh2 LPth Me
35 n = 2 (a)
n = 3 (b) + H2 H
' 2
14-
@4042»:
th P“ L/U\ ”,Me
LPth Me
Scheme 1.8 n = 3

©(CH2M 1+

 

 

 

 

©(CH230 1

+ H2
“I Ru ,.. ,.
Phi / \N N'IMe Phi ;UI\+’H N\"M°
Pth Me PPh2 M9
-HCOOH "

? (CH2)n ©(CH2)n
thp‘v/Ru\o H’N\'"Me + C02 thP‘“ /U\H ---‘H NimMe
Agh )‘H Me Lppth \Me

2
36 35 (n = 2 (a); 3 (b))
Scheme 1.9

37

While the proposed metal formate intermediate 36 could not be detected, the analogous
dithioformate complex 37 was easily prepared from 35b and excess CS2 (Scheme 1.10),

and its identity was unambiguously established by IR and NMR spectroscopies.56

9) (CH2\)3 032 ©(CH223

 

thP‘“';u\H____ ,Nf‘ 'Me thP“"/Ru \s H,N'+{"Me
th
35b 37

Scheme 1.10

Hydrogen exchange in the structurally related complex 38 was studied by
Chaudret et a1. using 1H N MR spectroscopy, and the activation energy for this process
was determined to be around 11 kcal/mol.58 However, extensive DFT calculations
suggested that the mechanism for the exchange does not involve any proton transfer

within the N-H---H-Ru dihydrogen bond.

©(CH222

w°RU N3”
Ph P‘ , 'Me
3 / \H-u-H \

Ph3P Me
38

A very interesting dihydrogen-bonded system with its hydrogen exchange
dynamics has been recently described by Jalon et al.59 They reported a three-center

py2H---H-Ru intramolecular interaction in 39, in which fast scrambling between the

38

hydridic and protonic hydrogen atoms occurs most probably via an "2_H2 complex
intermediate (Scheme 1.11). An activation energy of about 13.6 kcal/mol was determined
for this process, using variable temperature 1H NMR spectroscopy. Moreover, this system
proved to be a very active catalyst for DVHZ exchange. Thus, when a solution of 39 in
CD3OD was exposed to a dihydrogen atmosphere (1 atm) at room temperature, more than

90% of Hg was exchanged for D2 in about half an hour.

 

Scheme 1.11

A similar exchange was reported by Morris et al. in the ruthenium polyhydride

complex 29b-[K(l,lO-diaza-l8-crown-6)] (Scheme 1.12).” Upon exposure to D2 gas at l

39

atm and room temperature for 5 min, the intensities of the NH and RuH 1H NMR (THF-
(13) signals were depleted by 100% and 90%, respectively. For comparison, only 13%
decrease in the hydride resonance was observed after 10 days when the much less acidic
l8-crown-6 ether was used for complexation, implying efficient activation of the M-H
bonds by N-H---H-Ru dihydrogen bonding. The exchange is also significantly slower if
the ruthenium hydride is replaced by the less basic analogous osmium hydride (29a).
While the conjugate acid of 29b, the known RuH2(H2)2(iPr3)2 dihydrogen complex, could
reasonably play the role of the intermediate in the exchange process depicted in Scheme

1.12, its involvement in this transformation was ruled out by control experiments.

PiPr3 ( PiPra (
H ’ .H 1(wa DZ,THF,5min> D" I “D _ VN D
H’| \H \LOD—H _H2 0.. \D k )
PiPr3 PiPr3 o

 

Scheme 1.12

Intramolecular N-H---H-Re interactions can affect hydride fluxionality in
ReH5(PPh3)2L (L = N-acetyl-2-aminopyridine) (40).60 Thus, the free energy of activation
for the tumstile rotation involving the H1, H4, and H5 atoms in this complex is 0.7
kcal/mol smaller than for the analogous complex with the NHAc group in the para
position of the pyridine ring. Stabilization of the transition state by strong N-HmH-Re

dihydrogen bonding, which is only possible in the ortho-NHAC isomer, seems to be

40

responsible for the observed difference. However, this effect is partly offset by non-
bonding repulsive interactions between the two HmH hydrogen atoms, which are forced

to approach to 1.49 A in the transition state, according to theoretical calculations.61

40

Dihydrogen bonding can have important consequences on the selectivity and
stereochemistry of reactions in solution. Thus, the formation of directional and strong
HmH interactions can differentially stabilize one particular transition state among two or
more possibilities, ultimately controlling the product distribution or stereochemical
outcome. An illustration of this concept is the selective imination of the nl-aldehyde
complex 41 with ortho- vs. para—aminophenol, carried out by Crabtree et al.62 In a
competitive experiment using an equimolecular mixture of 2-aminophenol,
4-aminophenol and 41, a 4.2:1 ratio of the resulting products 41a and 41b was obtained
(Scheme 1.13), which was calculated to correspond to a k5/k6 value of 6. This outcome
appears to be the result of O-H---H-Ir dihydrogen bonding stabilization of 413 (and
presumably of the TS leading to it). The two isomers do not interconvert, indicating that
the observed product distribution is dictated by kinetic not thermodynamic control. While

the observed ratio between the two rate constants is translated into a AAGat value of 1.1

41

kcal/mol, the 4.5 kcal/mol estimated HmH interaction energy in 41a is substantially
larger, the difference being apparently offset by the unfavorable chelate ring

conformation required for efficient H-bonding in the transition state.

 

 

 

 

 

Scheme 1.13

Dihydrogen bonding can also direct the regiochemistry of ligand attachment to transition
metal clusters, as demonstrated by Aime and coworkers.‘53 Thus, reaction of the
electronically unsaturated osmium cluster 42 with EtNHz or EtzNH yields exclusively the
syn product 423, stabilized by an intramolecular N-H---H-Os interaction, which would not
be possible in the anti isomer (Scheme 1.14). Notably, with Et3N, which lacks the acidic

hydrogens required for dihydrogen bond formation, no reaction occurs.

42

-

-

- -

"' .

0' .

'4
o
o
-
‘9

N-—-R

H
(00) (gszosmok L co (I) 40s(00)4\
3 \H/OS(CO)3 —> ( )3 S\H/05(CO)3

 

 

42 423
L = EtNH 2, Et2NH
Scheme 1.14

The syn isomer is also preponderantly formed when 42 reacts with NH3. Subsequent
treatment with acetaldehyde or acetone in chloroform leads to the exclusive formation of
the dihydrogen-bonded imino-derivatives 42b with a syn configuration

(Scheme 1.15).“ 65 The intramolecular N-H-uH-Os interaction is disrupted in more polar,
hydrogen bonding solvents such as methanol or acetone. In this case, the interconversion
between the syn and anti isomers was demonstrated by variable temperature NMR
spectrosc0py, which suggests that the observed regioselectivity is a result of

thermodynamic, rather than kinetic control.

 

 

 

33““
...... H\ /C\R
WOO) 1. NH3 T” N
4
2. CH C(R)=O Os(CO)4 |
(CO)3054 > 3 _ 4
\H/Os(CO)3 CHCI3 r (CO)aOs\H/—\-05(co)3
42 42b
H = H, CH 3
Scheme 1.15

43

The ruthenium hydride 43, whose crystal structure shows an intramolecular N-
H---H-Ru short contact, was found to efficiently catalyze the asymmetric hydrogenation
of ketones to chiral alcohols.66 However, the specific contribution of the H--H interaction

to the high enantioselectivity observed in these reductions was not analyzed.

  

'Irs
Ph,, N
- \
L R
NI
Ph :_. \ .
H H

Work in our group by Gatling established the ability of O—H---H-B dihydrogen
bonds to direct the borohydride reduction of ketones to alcohols.67 Thus, reductions of 2-
hydroxycyclobutanone or 2-hydroxycyclopentanone with tetrabutylammonium
borohydride in the non hydrogen bonding solvents CH2C12, ClCHzCH2Cl, or o-
dichlorobenzene are accelerated about 150 times relative to the reductions of the
corresponding unsubstituted cycloalkanones, and yield almost exclusively trans diols
after workup (Scheme 1.16). These effects are greatly reduced in the presence of
competing hydrogen bonding alcohols or anions like F, Cl', or Br'. Capping of the OH
with a trimethylsilyl group also shuts off both the stereodirection and the rate
acceleration. AMl semiempirical calculations predicted a 3.5 kcal/mol preference for the
hydride delivery from the OH substituted face of 2-hydroxycyclobutanone. This value is
in reasonable agreement with the experimental findings, despite the crude level of theory

and the absence of counterions or solvent in the model.

1H, /B\H
O H HO H
H
(CH2)n§—O C 202 (CH2)n2><
\I OH
n = 1 , 2

Scheme 1.16

In the solid-state, dihydrogen bonding can control the transformation of
cyclotrigallazane (12) into nanocrystalline gallium nitride, according to Gladfelter and
coworkers.68 Initial loss of H2 at 150 °C resulted in an amorphous GaN phase, which
upon annealing at 600 °C led to the metastable crystalline cubic gallium nitride, as a 1:1
mixture with the thermodynamically favored hexagonal GaN. The crystallization in the
cubic system appears to be dictated by the initial crystal packing in 12, consisting of N-
H---H-Ga dihydrogen-bonded chains (Figure 1.10), which can be considered essentially
“hydrogenated” cubic GaN. For comparison, decomposition of cyclotrigallazane in thin
films obtained by vapor deposition, a process that presumably disrupts the dihydrogen-
bonded network, yields exclusively hexagonal GaN. It is remarkable that despite the huge
contraction of the unit cell accompanying the conversion of 12 into cubic GaN, the

reaction still maintains partial topochemical character.

45

2. Organic Solid-State Reactions: Mechanisms and Concepts

Organic solid-state reactions have been reported to occur since the beginning of
organic chemistry. The 1828 Wohler’s famous transformation of ammonium cyanate into
urea was proven to take place both in solution69 and solid-state.7O However, it was not
until relatively recently that this discipline has developed as a mature field, with the
invaluable assistance of physical methods like X-ray crystallography, optical and
electronic microscopies, thermal analysis, and solid-state NMR spectroscopy. A
fundamental difference between the reactivity in fluid and solid phases is that while the
former is dominated by the electronic properties of molecules, reactivity in solids is a
balance between packing and electronic effects. In the solid-state, the reactants are
usually locked in a fixed orientation with relatively low mobility, and consequently the
intrinsic reactivity of the molecules is often less important than their spatial arrangement
relative to the neighboring molecules in the crystal. Reactions under such geometrical
control can therefore be highly selective, leading sometimes to products and materials
otherwise inaccessible.71 The first to recognize this powerful concept was G. M. J.
Schmidt, who in the 1960’s articulated the “topochemical” principle, of lattice control
over the course of solid-state reactions and stereochemistry of the products.72 Thus, in
their seminal study of photochemical dimerizations of trans-cinnamic acid derivatives,
Schmidt and coworkers found correlations between their orientations in the starting
crystals, and their solid-state reactivity, as well as stereochemistry of the resulting dimers

(Scheme 2.1).

46

X—Ph

"P h—— X

X—Ph

X-Ph [I"3.8-4.2A —_” X—Ph .i-‘COR
"0

OR

Y'IYPC

X = H, OH, CI, N02, CH3, OCH3, OC2H5, OP!“

R = OH, oer-:3, OCsHs, NH2, CSHS

Scheme 2.1

47

The monomers may have three major types of crystal packings. In the or arrangement, the
molecules are anti-parallel, with 36-42 A between the double bonds, and yield almost
exclusively centrosimmetric dimers. In contrast, dimers with a plane of symmetry are

obtained by the irradiation of the B-type crystals, in which the double bonds are oriented

in a parallel fashion, approximately 38-42 A apart. Finally, the y—type polymorphs, with
distances between the closest double bonds greater than 4.8 A, are photochemically inert
in the solid-state. This study made Schmidt soon realize the imperative of understanding
the rules governing the packing of molecules in crystals, which led to the emergence of
crystal engineering.72

According to Schmidt, for a reaction to be topochemical, it has to proceed with a
minimum of atomic and molecular movement; otherwise, disorganization of the
crystalline medium due to excessive molecular motions may lead to loss of lattice control
over the reaction. Later, Cohen introduced a new concept, that of the “reaction cavity”,
defined as the space occupied by the atoms directly involved in the solid-state reaction.73
The topochemical principle can be thus reformulated to state that those reactions which
proceed under lattice control occur with minimal distortion of the reaction cavity, and
that formation or removal of any empty space within the cavity should be energetically
unfavorable, since they imply substantial changes in attractive and repulsive forces.
Gavezzotti quantified this theory by proposing precise methods for molecular volume
calculation, and concluded that a prerequisite for crystal reactivity is the availability of
free space around the reaction site.74

Topochemical control was also noticed in the solid state decomposition of diacyl

peroxides which photolyze to a pair of C02 molecules and a pair of free radicals (Scheme

48

2.2).75 While in solution or melt the fate of the radicals is complex and variable and
mixtures of products are obtained, photolysis of single crystals give radical-radical

coupling products in high yields.

O—O C02
__—_—’ .
Rflo chi? R—/ (:02 \—R

Scheme 2.2

At the first inspection one might conclude that the topochemical least-motion principle is
obeyed since radical coupling predominates in the solid state over disproportionation,
which requires that one radical abstract 3 hydrogen atom from the more distant second
carbon in the other radical. However, low temperature EPR studies of the generated
radical pair suggest that the reaction does not follow a least-motion path, as the reactive
carbons initially move apart before they finally combine. This behavior appears to be
caused by the two C02 molecules resulting from photolysis, which cannot be
accommodated in the reaction cavity. The pressure created by the forming C02 was
estimated around 20 kbar from its asymmetric stretching mode, which is sensitive to
pressure. The conclusion is that in such gas evolving systems, where there is no

possibility that the product molecules would fit into the reaction cavity, the developed

49

stress could control the course of the reaction. Recently, a unified theoretical formulation
of solid-state reactivity, that can be applied to thermal-, light-, and shock-induced
reactions as well has been proposed, in which the reaction cavity and steric compression
concepts have been quantified.76

There are two main mechanisms that may operate in organic solid-state
transformations.77 The first one, the homogeneous mechanism, involves conversion of
"A" molecules of the starting material in the initial crystal lattice, to "B" molecules of the
product, with the B molecules remaining in the same region of the lattice initially
occupied by the A molecules. A single crystal of the reactant will thus pass through a
continuous series of solid solutions, first of B in A and later of A in B, before becoming a
single crystal of the product. Minimal structural change is required on going from A to B
so that the crystal lattices of reactant and product, which should have very close
geometrical parameters, can accommodate each other in the process of conversion. These
severe requirements make candidates for the homogeneous mechanism extremely rare. In
most of the solid-state reactions the product is structurally incompatible with the reactant
lattice, and therefore a heterogeneous mechanism operates, with initiation of reaction at
nucleation sites frequently located at crystal imperfections or surfaces, and subsequent
growth of the product phase, which usually does not retain the three-dimensional
periodicity of the initial crystal, and appears amorphous or polycrystalline. Even under
these conditions, the process can be topochemical; a great number of solid-state reactions
have shown that there is no strong correlation between the chemical specificity and the

long-range crystallographic order of the product phase.78

50

Another term frequently used by solid-state chemists is “topotaxy”, which is
concerned with the relationship between the three-dimensional crystallographic
orientation of reactants and products.79 In a “topotactic” transformation, the lattice
parameters, crystal system, and space group of the product are closely related to those of
the reactant. The solid-state process does not necessarily have to be homogeneous, as
nucleation of the crystalline product at the surface of the initial crystal can lead to
oriented growth under the influence 6f the surface tension, resulting in coincidence
between certain crystallographic axes of reactant and product phases. It thus follows that
while topochemical control often leads to topotaxy, the later phenomenon may be
completely independent from the occurrence of a topochemical process.

A potential problem in solid-state organic chemistry is that the intermblecular
forces in organic crystals are predominantly of van der Waals or hydrogen bonding type,
which are significantly weaker than typical covalent bonds. As a consequence, melting or
sublimation may occur very often in thermally activated reactions, before any detectable
chemical transformation. Another possible complication is the formation of liquid
products. It is therefore imperative that these possibilities be ruled out any time the

exertion of topochemical control is invoked.

2.1. Topochemical Polymerizations and Polycondensations

In ordinary solid-state polymerizations there is essentially no lattice control over

the polymer chain growth, apparently due to the considerable movement that the

monomers must undergo to add to the end of the growing chain.80 This is the result of the

51

significant shrinkage that occurs when going from van der Waals distances between
molecules to much shorter covalent bonds, leading to a growing gap between the chain
end and the next reacting monomer. This means that ultimately the monomer can add to
the polymer only by breaking away from the lattice, resulting in the loss of
stereoregularity and crystallinity. Hirshfeld and Schmidt argued that this effect could be
avoided in crystals of carefully designed bifunctional monomers.81 It was predicted that
in suitable structures, the monomer would attach itself to its neighbor via a rotation about
its center of mass, so that the polymerization would not require diffusive motion, and
lattice control would be maintained throughout the transformation. Realization of this
model has indeed been achieved in divinyl arenes82 (Scheme 2.3) and diacetylenes79
(Scheme 2.4), which can photopolymerize topochemically, leading to highly

stereoregular, sometimes crystalline polymers.

 

Scheme 2.3

52

C$C \C
\C\\ \\\
\ C\ x
R / C— R
/
FL R— C
CS ‘0
C\c\ *’
\\C\ \
R / C— R
/
R\C R—C\
§\C\ C \
C\\ “
\C\R ‘c R
//
Scheme 2.4

These topochemical polymerizations may proceed either homogeneously or
heterogeneously.79 In homogeneous cases the polymer is formed as a solid solution,
which grows from points randomly distributed throughout the parent crystal. The product
is thus isomorphous with the monomer crystal. This mechanism usually operates in the
polymerization of diacetylenes, and in some cases, singleocrystal to single-crystal
transformations can be achieved. In the case of [2+2] photopolymerizations,
heterogeneous mechanisms dominate; the reaction starts preferentially at defect sites of
the monomer crystal, with nucleation of the polymer phase. The parent crystal typically
shatters into a polycrystalline aggregate, or, sometimes, an amorphous phase. Once the
crystal is broken, polymerization continues into the crystallites, usually starting from the
edges. Comparison of the monomer and polymer crystals showed that in many cases the

space group does not change and the modification of the unit cell volume during

53

polymerization is very small, implying topotactic control. Moreover, when the initial
diolefin crystallized in a chiral space group, an absolute asymmetric synthesis with the
formation of an optically active polymer could be achieved.83

An interesting solid-state thermal polycondensation occurs in salts of
halogenoacetic acids, leading to polyglycolide, the simplest possible polyester (Scheme

2.5).84
x—CH2-000'*M ———>1/nE——CHz-eoo-—]In + MX

x = Cl, Br, I;
M = Na, K, Rb, NH4, Ag

Scheme 2.5

Interestingly, the metal halide byproducts are deposited as small cubic crystallites in the
polyglycolide matrix, implying long-range diffusion of the formed M+X' ion pairs in the
crystals. The salts can be subsequently washed out with water, to leave a highly
microporous polymer. Crystal structure analysis in the case of silver chloroacetate
indicated that this polycondensation is topochemical, and is driven by the very short Ag-
Cl interatomic distance of 2.903 A, which is only 0.128 A longer than in silver chloride.85

The observed interatomic distance between the reacting carbon and oxygen atoms of 3.25

A is also within the range expected for high solid-state reactivity.

54

2.2. Hydrogen Bonding Assisted Topochemical Reactions

The topochemical principle is a very powerful concept that can be applied to the
assembly of both small molecules and extended covalent networks with controlled
topologies. However, the major disadvantage of this approach is the lack of solid-state
reactivity in many systems due to misalignment of the reactive functionalities in their
crystals. This obstacle can, however, be eliminated by successful application of crystal
engineering in designing solid-state reactive molecular solids. Schmidt and Leiserowitz
noted as early as 1969 that the probability of mB-unsaturated amides to crystallize in
photodimerizable structures was very high, due to favorable hydrogen-bonded,
predictable packings.86 Alternatively, they could control crystal packing by introduction
of stackable dichlorophenyl units, or by co—crystallization with mercuric chloride, which
were shown to be capable of imposing 4 A periodic arrangements.72

More recently, Feldman and coworkers used the predictable hydrogen bonding
dimer motif of the carboxylic acid functionality to properly align olefins for
photodimerization (Scheme 2.6).87 This example is a nice illustration of using hydrogen
bonds as preorganizing interactions, which can be subsequently reinforced by robust
covalent linkages. While in this case a discrete dimer was assembled, extension of this
strategy to higher-dimensional systems by appropriate design can be easily imagined.88
A similar approach was used by Matsumoto et al. for the topochemical
photopolymerization of benzylammonium salts of butadiene-1,4-dicarboxylic acid

(Scheme 2.7).89

55

 

x” 'o c 1 002-)?
2 —-\ -
—/ 002- X+ — _ / n
cogx”
NH3 NH3
Cl
C C

Scheme 2.7

56

Hydrogen bonding between the N H3+ and COO' groups enforce the required orientation

of the butadiene units for solid-state polymerization in this case (Scheme 2.8).

 

NH3+
o - —
O\c/—\:/C\0' / \
+ H *HaN \
NH3 F?
O -O\/(/:/:\.:/C\O' 0
+H N
NH3+ 9 3
O -O\/(’:/_\\:/C‘O' // \
O +H3N \

Scheme 2.8

The reliable self-assembly of urea derivatives into hydrogen-bonded a—networks was
exploited by Lauher and Fowler to preorganize diacetylenes for solid-state topochemical
photopolymerization.90 Thus, by co-crystallization of a pyridyl-substituted diacetylene
with dicarboxylurea, an anticipated B-network could be assembled, in which the
diacetylene units were forcibly aligned into the desired solid-state reactive arrangement
(Scheme 2.9). Using a similar strategy, they were recently able to align a triacetylene for
solid-state polymerization, which led to an unprecedented 1,6-polytriacetylene via a

single-crystal to single-crystal reaction (Scheme 2.10).91

57

O H H O \ \ é , N O H H 0
o O H I'I IRA H
o o ' O \0 \I/ 0’
, . o l
H N N H / O 0
I I , \ .
o H H o N / \ " ‘
t‘ I} \ \ O/ , N O H 1‘". O
H’0 N/U\N O‘H O I
| I ~‘N ’ // O \ \ O

H’ '7‘ l‘ I N // VN. , 'fi fi' 0
o H H o ’
,0 NA 0. /N
H \H/\‘I 1:1/Y H~.\N/ // H ? E '7' O H
o H H 0 SY0 ‘0 Y o/
o 0

Scheme 2.9

58

com--- N

- / H’xt
QN \ \\ \ ¢N
H/\©N H020 // ©
(H)
C: 7

H020
l hv

/~COQH- H
OVI
// m
o // I")

ré

ICOZH

AJVN/
Ho2 c] K H
0'

H020

Scheme 2.10

59

Besides hydrogen bonding, other noncovalent associations such as phenyl-
perfluorophenyl stacking interactions have also been explored as preorganizing elements

in the solid-state topochemical polymerization of diacetylenes92 and diolefins.93

2.3. Kinetics of Solid-State Decompositions

In this section the focus will be mainly on organic solids which decompose to
yield solid and gaseous products?4 as also expected from dihydrogen-bonded systems

upon heating in the solid-state. This can be represented as:
Asolid '9 Bsolid + Cgas

Decomposition may be initiated by heat, light, or ionizing radiations. The stability toward
thermal decomposition may be affected by the previous history of the sample, presence of
impurities, aging, sample size, or the nature and pressure of gas atmosphere above the
solid.95 The rate of the process can be most conveniently followed by measuring the
pressure developed in a constant volume system or by using a thermobalance to
determine the weight loss under isothermal conditions, although XRD or spectroscopic
methods can also be used.94b The fractional decomposition 0t vs. time curves are typically
sigmoid in shape, indicating an autocatalytic reaction. This pattern is almost universally
associated with the initiation of reaction at specific sites followed by the growth of
nuclei, with the reaction mostly confined to the product/reactant interface. The kinetics is
therefore controlled by the number of nuclei present and the total area of the expanding
interface. After the inflection point, the growing nuclei start to coalesce, slowing the

decomposition rate as the interfacial area decreases, until the reaction eventually st0ps.

6O

Different kinetic models with their corresponding equations have been elaborated to
account for various solid-state decomposition mechanisms.94 Among them, the most
common are the Avrami-Erofeev equation (1) and the phase boundary model (2):
[-ln(1-0t)]""=kt; it: 1-4 (1)
1-(1-ot)“"=kt; n= 13 (2)
The first one corresponds to a nucleation and growth mechanism, while the second is
associated with an inward advancement of the reaction interface from the crystal’s edges.
The exponent n in Eq. 1 is given by B + X, where B is the number of steps involved in

nucleus formation (typically B = O or 1, the former corresponding to spontaneous

nucleation), and 2. is the dimensionality of the nuclei growth.94b

For some decompositions the acceleratory region can be described by an
exponential expression (3).94a
(X = Ce,“ (3)

Due to molecular volume change, the formation of product molecules induces strain in
the crystal, which produces cracks forming fresh surfaces on which decomposition can
occur. The reaction will therefore spread down these crevices into the crystal.
Decomposition on these surfaces produces additional strains and hence further cracking
and a type of chain branching process thus develops.94c This mechanism can be
mathematically described using the Prout-Tompkins equation (4):

ln [oz/(1-oc)] = kt (4)
It should be pointed out, however, that very often there is no single model that can
acceptably describe the whole conversion range, as different mechanisms may operate for

different stages of decomposition.

61

A substantial autocatalytic effect has been also observed for some single-phase
solid-state polymerizations, which could be explained quantitatively by a strain-
dependent rate of chain initiation and propagation model.96 Thus, it was demonstrated
that the conversion—time curves display a more pronounced sigmoid character as the
differences in unit cell parameters between the monomers and polymers increase.

The rate constants obtained from solid-state reactions usually obey the Arrhenius
equation. However, special care must be exercised when interpreting the obtained
activation parameters, since decompositions of solids are complex processes involving
not only chemical steps such as breaking or formation of bonds, but also physical
transformations such as phase transitions, diffusion and desorption of gaseous products,
and heat transfer. Therefore, chemically specific techniques may be necessary, to obtain
mechanistic details at the molecular level. Another important issue is whether the
Arrhenius equation has any physical meaning in the solid-state, as its theoretical
foundation has been established in the context of collision theory in the gas phase.
However, even if the Maxwell-Boltzmann energy distribution is not applicable to the
immobilized constituents of a solid, it was demonstrated that energy-distribution
functions of similar form (Fermi-Dirac or Bose-Einstein) also give rise to an Arrhenius—
type equation.97 Thus, its application to solid-state transformations is justifiable not only
in terms of a useful empirical parametrization, but is also supported by a rigorous
theoretical foundation.

An alternative approach for the kinetic analysis of solid-state reactions is the
isoconversional method, applied under isothermal or nonisothermal conditions.98 This

strategy allows the estimation of the activation energy without assuming a particular

62

reaction model, and is particularly convenient for analysis of nonisothermal data such as
that obtained from DSC and TGA experiments. Under isothermal regime, the activation

energy at a particular conversion, Ea, can be evaluated using Equation 5:98b

- In to“ = ln[A/g(0t)] - Eel/RT, (5)
While in some cases it can unmask the complexity of the solid-state process studied, the
major disadvantage of this method, however, is that the obtained activation energy varies
significantly with the extent of reaction, which complicates the interpretation of the

kinetic data.

63

3. Topochemical Assembly of Covalent Materials Using

Dihydrogen Bonding

As illustrated in Chapter 1, many studies regarding the dynamics of dihydrogen-
bonded systems in solution have demonstrated that proton transfer from the acidic AH
partners to the transition metal hydrides MH, along the H---H bonds, generally leads to
172-H2 non-classical complexes, which subsequently eliminate hydrogen upon heating,
with the formation of covalent M-A bonds (Scheme 1.2). An analogous process appears
to occur in the case of the borohydride anion. In aqueous solutions, BH4' is very likely
dihydrogen-bonded to H20, as suggested by the crystal structure of NaBHn-ZHZO,ll as
well as theoretical and experimental studies by Epstein et al.10 Under neutral or acidic
conditions, borohydrides undergo hydrolysis to boric acid (B(OH)3), for which the
established mechanism involves slow proton transfer resulting in a BH5 intermediate,
followed by fast H2 loss and B-0 bond formation (Scheme 31).” Activation parameters
AH“ and AS“ of 20.6 i 1 kcal/mol and —22.3 i 3 en were measured for the neutral
hydrolysis, while under acidic conditions the corresponding obtained values were 8.0 i 1
kcal/mol and —3 _+.. 3 eu, respectively.100 The structure of BH5, as deduced by theoretical

calculations, is best described as an almost planar BH3 molecule, loosely coordinated by

10]
H2.

H - -
BH4i-—Hzo s'°‘” Has—H OH fi+ HaB-OH + H2

 

 

Scheme 3.1

Theoretical work by Elguero et al. indicates that H2 generation from dihydrogen-bonded
borohydrides can also be induced by the internal forces within a crystal.‘02 All these
premises, together with the established ability of borohydrides to self-assemble into
extended dihydrogen-bonded networks, suggest that A-Hn-H-B dihydrogen bonds could
be employed in topochemical assembly of covalent materials. Such weak HmH
interactions, in principle, may be used to organize and hold a structure’s form while it is
more firmly fastened together by A-B bond formation, transferring thus the initial order
from the starting crystal to the newly formed covalent frame. This strategy makes
dihydrogen bonding a potentially powerful tool for rational assembly of new covalent
crystalline materials with controlled structures and properties.

In the topochemical transformation outlined above, the initial H-bond
arrangement should detemiine the final covalent structure. However, when going from
the A-H---H-B interaction to a covalent A-B bond there is approximately a 1.5-2 A
change in distance between A and B. This is a considerable shrinkage, and if it were
cumulative in the reaction direction, it would lead to an increasing gap between the
growing covalent network and the unconverted dihydrogen-bonded molecular units as the
decomposition advanced, with eventual loss of lattice control over the reaction. Even
under these circumstances, however, the transformation may still be topochemical, as
demonstrated by the solid-state conversion of cyclotrigallazane into nanocrystalline
gallium nitride reported by Gladfelter et al.“, which is accompanied by 62.9% shrinkage
of the unit cell. The price to pay was the initial loss of crystallinity and the consequent

requirement for high annealing temperatures to restore it. Although this thermal treatment

65

had no detrimental effect upon the robust GaN product, more delicate structures would
not tolerate such high temperatures, limiting the general applicability of this approach.

A low temperature procedure for topochemical dihydrogen to covalent bonding
transformations would allow the extension of this strategy into the structurally more
diverse domain of organic materials. Like many solid-state processes, this reaction
includes two threats to the crystalline order: (a) geometry change upon bond
reorganization, and (b) gas generation and diffusion within the lattice. Clearly, careful
design of the starting dihydrogen-bonded networks is necessary in order to meet these
challenges. Success, however, would mean that the well-developed tools of molecular
synthesis could now be applied to the rational construction of crystalline covalent solids
with desired structures and functions.

This work involves two strategies to address this problem: (a) design of cations to
form closed loops in coordination with hydride-bearing anions, in which case the lattice
distortion would not be cumulative, and (b) selection of globular cations large enough
that their close packing determines the lattice parameters, with the hydridic anions fitting
into the interstitial holes, in which bond formation via flexible arms would induce

103 A dividend of this latter strategy is that

minimal change in the unit cell (Scheme 3.2).
the anticipated looseness of the lattice should allow the released H2 to diffuse readily

through and out of the crystal.

66

(a) A
y.
I

I
M
—m

Scheme 3.2

3.1. Results and Discussion

N-[2-(6-Aminopyridyl)]-Acetamidine (NAPA) Cyanoborohydride: A Closed
Loop Dihydrogen-Bonded Structure by Design.'03 NAPA H3BCN (44) was
synthesized (Scheme 3.3) to explore the prospect of convergent coordination in closed
loops, which appeared as a likely possibility due to the bent geometry of the
aminopyridyl-acetamidine unit. The cyanoborohydride anion was chosen as the hydridic
partner because of its enhanced stability toward acidic substrates relative to BHa‘,IOIC
which was decomposed instantaneously in solution by the NAPA cation. The white

crystalline compound melts at 119-120 °C with decomposition and gas evolution. The

1H, '3 C, and 11B NMR, as well as IR spectra confirmed its structure.

67

.,
/ OCH /
\ I + H3C_< + 3 - —_’ £1 —]
H2 NH2 NH2 CI H2 N NH
ct Wk

CH3

NaNCBHa
-NaC|

CH30H
NCBH3 H2N’J\ H3

Scheme 3.3

Figure 3.1 shows the X-ray crystal structure for 44. There are two independent
centrosymmetric (NAPA H3BCN)2 dimers in the unit cell, each of them exhibiting close
H-H contacts: 1.98, 2.12, 2.26 A, and 2.04, 2.09, 2.31 A, respectively. Applying the
corrections for the N -H and B-H bonds that appear too short from X-ray compared to
typical literature values, these H-H contacts become 01-015 A shorter. This is well
below the 2.4 A van der Waals contact radius, implying strong and specific interactions.
The NH---H-B angles are strongly bent, ranging between 91.6° and 126.3°, with an
average of 108.2°, while the N-H-~HB angles are larger (range 148.1 - l75.6°; average
158.5°), as expected for dihydrogen bonds. It thus appears that despite the less negative
charge on the hydridic hydrogens (-0. 18 vs. -0.27 by Mulliken population analysis for
MP2/6-311++G** wavefunctions), the NCBH3' ion is capable of strong dihydrogen
bonding, comparable to that of the BH4’ ion in this case, owing to the increased acidity of
the proton donor partner. More important, the structure of this very first attempt showed
two independent occurrences of the intended closed loop packing, with its potential for

topochemical control.

68

    
    

   

N(5A)€:’:‘ i A A

C(BA “sf- f
.L, B(A)\3

 

Figure 3.1. X-Ray Crystal Structure of NAPA H3BCN showing the HmH contacts in A.

69

Decomposition in both solid and solution unfortunately led to complex mixtures (Scheme
3.4) via the unwanted reduction of the amidine group by the cyanoborohydride, which
accounts for the majority of the decomposition products identified by MS, lH, '3 C and
1'B NMR. Although the decomposition product ratios were slightly different in solid state
and solution, we could not attribute that result to topochemical control, since the solid
state reaction occurred with partial liquefaction due to the low melting point of some of

the products, relative to the temperature required for decomposition (~65 °C).

/

/
H \N I NH solid-state, 65 deg N11 +
2 or THF refl. H2 \N NH2

NCBngb * CH3

 

44
/ / /
\ I . \ I \ I
H2N NH—CHZCH3 H2 NH CN N N NH2
1
AN+IB\NJK
l I
H H
NCBHg
443

Scheme 3.4

70

Despite the lack of observable topochemical control, this system demonstrated the ability
of NCBH3' to form dihydrogen bonds and showed that the closed loop coordination is a
viable packing arrangement that can be readily designed into dihydrogen-bonded
systems. Additionally, another critical design feature for the next dihydrogen-bonded
systems became apparent: their melting point, which should be sufficiently high to allow
the decomposition to be carried out in the solid.

Topochemical Control by Dihydrogen Bonding. Structure and Reactivity of

NaBH4tTEA (45).103 The complex of NaBH4 with triethanolamine (TEA) was

synthesized as a candidate for the globular cation strategy.

I‘\

The complex precipitated from a stirred mixture of NaBH4 and TEA in THF as a white
crystalline compound, which melts at 107-108 °C with decomposition and gas evolution.
The X-ray powder diffraction pattern of the complex is unique and confirms the absence
of NaBH4. The 11B and 23 Na solid state MAS NMR chemical shifts are -47.9 and -9.9
ppm respectively, 2.5 and 6.1 ppm downfield relative to sodium borohydride, indicating
different environments for the BH.{ and Na+ ions. However, the B-H stretching vibration
frequency for 45 (2288 cm") is almost identical with the corresponding value in N aBH4

(229lcm"). The X-ray crystal structure is shown in Figure 3.2.

71

(a)

     

I' ‘t
I. '
7 1
I /
l
1‘ / I»).
A, _ five/'2'

C(23)“ - Has)?"
U ."I

(b)

Figure 3.2. X-Ray crystal structure of NaBHa-TEA: (a) coordination of Na+;

(b) dihydrogen bonds connecting the chains, with H-H contact distances in A.

72

Each Na+ is hexacoordinated by the N and two 0 atoms from a TEA molecule and by
three 0 atoms from the two neighboring TEA molecules, forming TEA-Na+ linear chains
(Figure 3.2a). Two 0 atoms from each TEA molecule are shared by two adjacent Na+
cations, but the third oxygen only coordinates the next Na+ in the chain. Thus, 45 is
unlike the previously reported structures of TEA complexes where each cation was
complexed by all four heteroatoms from a given TEA molecule.104 Multiple dihydrogen
bonds connect the chains via the BH4’ anions, giving rise to extended two-dimensional
layers (Figure 3.2b). Each BHa' H-bonds with two OH groups in one chain and one OH
group from the next chain in a total of 5 dihydrogen bonds. The H—H contacts from the
three different OH sites are 1.94, 1.93 and 2.12 A, and 2.16 and 2.13 A in distance
respectively, the latter two being bifurcated H-bonds. Again, typical corrections for O-H
and B-H bond lengths lead to H-H distances which are 0.15-0.2 A shorter. The OHmH-B
angles vary between 91.8 and 111.70 (average 98.0°) while the O-H-HHB angles range
between 143.9 and 166.7° (average 157.0°). There is no conventional H-bonding; instead
all the hydroxylic protons point to the interchain space to form dihydrogen bonds with the
borohydrides, demonstrating the importance of these interactions in defining the crystal
packing. However, an important driving force for the formation of 45 must be the
complexation of Na” by TEA, since no complex was formed with (CH3)4NBH4 or KBHa.
Solid-state decomposition of NaBH4-TEA at 82 °C under Ar or open atmosphere
for 30-58 days resulted in a loss of 3 moles of H2 for each mole of 45, as indicated by the
H“ content of the initial complex and final decomposed material. Attempts to use higher
decomposition temperatures induced melting. The rate of decomposition seems to be

affected by different factors such as sample size, humidity, or the nature or pressure of

73

gas atmosphere above the solid, which is characteristic for many solid-state reactions.95
Thus, decomposition is slower under vacuum than under Ar, and it is considerably faster
in open atmosphere, probably due to the presence of humidity. The resulting white solid
is insoluble in common organic solvents and it does not melt up to 300 oC, suggesting a
polymeric structure. Its llB solid state MAS NMR shows a single peak at 8 = -7.1 ppm.
X-Ray powder diffraction of the same material exhibits only a semi-amorphous phase
with a broad peak at 20 = 94°, but no peaks corresponding to NaBH4 or to the initial
complex (Figure 3.3c). Furthermore, the 1‘B NMR taken after hydrolysis of this material
in neutral D20 showed only traces of BH4’, indicating the virtual absence of this boron
species in the solid-state decomposition product. These data, correlated with the hydride
content of the decomposed material (one H' left), suggest a trialkoxyborohydride
structure for the decomposed material. The B-H stretching frequency is shifted 5 cm'1
relative to the initial complex (from 2288 to 2293 cm"'). This is not a dramatic change,
but it is not surprising since the variation of VB“ in the BHX(OR)4-,[ series is generally

1’05 The loss of 3 moles of H2 is in agreement with the crystal structure of 45, where

smal
each BH4‘ is H-bonded to three -OH groups, two from one chain and one from the next
chain. This fact suggests a topochemical relationship between the starting and final
materials. Additionally, the two OH from the same chain in 45 belong to different
adjacent TEA molecules. A topochemical reaction should therefore lead to a two-
dimensional, extended covalent structure, and indeed, the peak at 29 = 9.4 in the powder

XRD (Figure 3.3c) corresponds to a d spacing of 9.41 A in a layered structure. This value

is very close to the 9.24 A inter-layer distance in the initial NaBHa-TEA.

74

 

l_
I I V ‘

 

 

 

;
I.
I
I.
I
-I-
I b
p
:
‘EJ‘j
I. "“
C

   

hllLllllllnlllJlell

 

 

 

O 1 O 20 30 40 50 60

20/°

Figure 3.3. Progress of NaBHa-TEA decomposition monitored by X-ray powder
diffraction: (a) initial complex; (b) high-temperature polymorph; (c) final decomposed

material.

75

Monitoring the solid-state decomposition by X-ray powder diffraction (Figure
3.3) and H' analysis revealed a slow (~ 2 weeks), reversible phase transition of the initial
complex, prior to any H2 loss. The IR spectrum for this high temperature polymorph
shows the B-H stretching vibration at 2291 cm‘1 and the “13 solid state MAS NMR
spectrum shows a chemical shift at -51.4 ppm. Structure elucidation of this high
temperature crystalline intermediate, which somewhat complicates the topochemical
relationship between the initial and final structures, would provide useful insight into the
intimate mechanism of conversion from the dihydrogen bonding to the covalent
networks. Unfortunately, only the low temperature polymorph of 45 has been obtained in
recrystallization attempts to date.

An indication for topochemical control is that different reaction products or
stereoselectivity are observed in the solid state than in solution or melt. Therefore, we
also studied decomposition of 45 in DMSO solution and in the melt. A solution of 45 in
DMSO was stirred and heated under Ar at 110 °C for 55 hours. Like the solid-state
decomposition product, the resulting precipitate is insoluble in common organic solvents
and does not show any melting below 300 °C. However, H’ analysis and IR showed that
this compound has virtually no hydridic hydrogen left. Its llB solid state MAS NMR
chemical shift is slightly different from the solid-state decomposition material (6 = -5.0
ppm). No other product of decomposition was found by 11B NMR of the mixture resulting
from a parallel reaction run in DMSO-d6, BH4' being the only boron species left in
solution. On the other hand, when the solid-state decomposition product was stirred in
DMSO under the same conditions, no change was observed in the hydridic content or IR

spectrum, implying again different products for the solid and solution decompositions.

76

This experiment demonstrated that unlike in the solid state, where the reaction stops at
trialkoxyborohydride, decomposition of 45 in solution resulted in complete alcoholysis of
BHa' to B(OR)4' as expected for a reaction in which the BHX(OR)4-,{ intermediates, which
are more reactive than the starting borohydride, are mobile and therefore susceptible to
disproportionation. '06
Similarly, the reaction in the melt at 130 0C under Ar for 2 h resulted in complete
alcoholysis of half of the BH4' from the initial complex, NaBH4 being the byproduct of
decomposition as indicated by X-ray powder diffraction and HB NMR. It is remarkable
that despite the considerably higher temperature of decomposition in the melt, only 2
moles of H2 were lost compared to 3 in solid state. The decomposition process in the
solid state thus differs significantly from those in solution or melt. These results
demonstrate that the particular packing of the molecules in the original crystal, dominated
by H---H interactions, induced the reaction to take place under topochemical control,
leading to an otherwise inaccessible poly-trialkoxyborohydride structure. However, this

product showed poor crystallinity, and the reaction times were exasperatingly long due to

the relatively low temperature required for decomposition in order to avoid melting.

Influence of the Relative Acidity/Basicity of the Proton/Hydride Partners
Upon Solid-State Reactivity of Dihydrogen-Bonded Systems: Structure and
Reactivity of NaNCBH3~TEA (4t5).'03 NaNCBH3-TEA complex was synthesized in
order to explore the influence of varying the hydridic donor partner upon the structure
and solid—state reactivity of dihydrogen-bonded systems. Due to the smaller negative

charge on the hydridic hydrogens, the dihydrogen bonds were expected to make smaller

77

contributions in defining the solid-state structure and reactivity, compared to the
borohydride analog. The complex was crystallized by slow evaporation of a 1:1
triethanolamine-NaNCBH3 mixture in isopropanol. The X-ray structure is shown in
Figure 3.4. The Na+ cations are heptacoordinated by the four heteroatoms from one TEA
molecule, the nitrogen from the cyanoborohydride and two oxygens from neighboring
TEAS which bridge adjacent N a” cations to form extended chains (Figure 3.4a),
crosslinked by dihydrogen bonds. One H from each CNBH3' hydrogen bonds to a
hydroxylic proton from another chain (Figure 3.4b) with an uncorrected H-H contact
distance of 2.16 A (OH---H-B angle = lOO.6°; o-H---HB angle = 159.6°), forming a three-
dimensional network. There are also O(1)-H---O(2) intra-chain conventional H-bonds
connecting neighboring TEA molecules. It appears that dihydrogen bonds are less
important in the NaNCBH3-TEA crystal packing than in the borohydride analog, with its
shorter H---H contacts and complete absence of conventional H-bonding. This difference
is also reflected in the significantly different solid-state reactivity of the two complexes.
While the borohydride complex melted at 107-108 °C with decomposition, the
cyanoborohydride analog melts at 85-86 °C, and it takes approximately 100 more degrees
to finally start decomposing. Thus, close HmH contacts do not automatically confer
solid-state reactivity; the relative acidity and basicity of the protonic and hydridic
partners are also critical. Precise estimates of the basicities of BH4‘ and BH3CN' are
unavailable, but the bimolecular rate constants for their H3O+-catalyzed hydrolysis are
~106 and 10'2 l/mol x sec respectively which may be translated into a pKB difference of 2

8 units.'07

78

(a)

 

 

Figure 3.4. X-Ray crystal structure of NaNCBH3-TEA: (a) coordination of N a+;

(b) dihydrogen bonds connecting the chains, with H-H contact distances in A.

79

Tuning Dihydrogen Bonds: Enhanced Solid-State Reactivity in LiBHa-TEA
(47).108 The simplest way to tune the solid-state reactivity in 45 appeared to be the
substitution of the Li+ cation for Na“, which should result in stronger complexation by
TEA, thus making the OH sites more acidic, and consequently more reactive. The LiBHa-
TEA complex (47) was obtained by slow evaporation of a 1:1 TEA-LiBH4 mixture in 2-
propanol, as a white crystalline compound, which does not melt up to 300 °C. The X-ray
crystal structure of 47 is presented in Figure 3.5. Each Li+ is pentacoordinated by the N
and three 0 atoms from a TEA molecule and by an 0 atom from a different TEA
molecule, forming (Li+TEA)2 dimers, in contrast to the Na analogue, where (Na‘TEA)n
chains were present. Multiple dihydrogen bonds connect the dimers via BH4' pairs, giving
rise to extended ribbons. The dihydrogen bonding network is asymmetrical: one BHa' H-
bonds with three OH groups (0(2’), 0(1) and 0(3A)) in a total of 6 dihydrogen bonds,
while the second BH4' only H-bonds with two of the remaining OH groups (0(2A’) and
0(3)) in a total of 4 dihydrogen bonds. The 0(1A)H groups are not involved in any
dihydrogen bonding; instead they form conventional H-bonds with 0(l)s from the
neighboring ribbon, creating thus overall two-dimensional layers. The H-H contact
distances range between 1.69 and 2.32 A. The typical normalization of 0-H and B-H
bonds to 0.96 and 1.21 A respectively leads to the more realistic H-H distances of 1.62 -
2.28 A. All hydrogen atoms were unambiguously located from the difference Fourier
map, making these distances sufficiently reliable. The 0H---H-B angles range between
75.8 and 106.1° (average 95.0°), while the 0-H---HB angles are larger, ranging between

130.0 and 170.7° (average 150.8°).

80

2.30" IONA)

\_.

 

Figure 3.5. X-Ray crystal structure of LiBHa-TEA showing the H---H close contacts in A.

81

The shortest dihydrogen bonds are 1.69 and 1.76 A in distance, which after the
normalization of the 0-H and B-H bonds become 1.62 and 1.67 A respectively. These
are the shortest H-H distances reported so far for dihydrogen bonds. Notably, both OH
groups involved in these two dihydrogen bonds contain bridging 0 atoms coordinating
two Li+, which undoubtedly results in increased acidity of their corresponding protons,
and thus enhanced hydrogen bonding ability. All the other OH groups complex only one
Li+, leading thus to H-H separations that are significantly longer.

The very short H-H distances and the expected acidity increase of the OH groups
due to their complexation of Li“ predicted enhanced solid-state reactivity for this system.
Indeed, when heated approximately 1 h at 120 °C under Ar, the complex completely
decomposed, as indicated by X-ray powder diffraction (Figure 3.6), HB solid-state MAS
NMR and IR spectra of the resulted material. Only about 24 h were necessary for
decomposition at 82 °C, the temperature used for the solid-state decomposition of NaBHa
-TEA, which took about 6 weeks. Three moles of H2 per mole of 47 were lost, as
indicated by the H' content of the initial complex and final decomposed material. The
resulting solid is insoluble in common organic solvents and does not melt up to 300 °C,
indicating a polymeric nature. Its llB solid state MAS NMR exhibits a single peak at 8
-4.6 ppm, which, together with the H' content, suggests a trialkoxyborohydride structure
for the decomposed material. The v3-“ of 2247 cm'1 for the same material is also
significantly shifted from the 2290 cm‘] in the initial complex. No phase transition was
observed in this case when the decomposition was monitored by X-ray powder

diffraction.

82

 

 

lljl4lllllllllll_lJ_lllllll

0 1 0 20 30 4O 50 60

29l°

Figure 3.6. X-Ray powder diffraction pattern of LiBH4-TEA (a) and its solid-state

decomposition product (b).

The loss of three moles of H2 points to a topochemical relationship between the
final decomposed material and the initial complex, where two BH.{ and six OH groups
are in close proximity due to the crystal packing imposed by the dihydrogen bonding
network. Unlike in the NaBH4-TEA case where two-dimensional covalent layers were
likely to have formed, a topochemical reaction in the present example is expected to yield
a one-dimensional covalent structure (48), according to the crystal structure of 47
(Scheme 3.5). However, due to the modest crystallinity of the final material (Figure 3.6b),
a definitive knowledge of its structure remains elusive. Using lower temperatures for

decomposition (down to 65 °C) did not improve the crystallinity of this material.

83

 

E LI-. / . I 2)
x-,\' ‘~ s o ‘
Q 9 2H0 t E W
”’4 N~~-L- ,IJ- —~.,,
It Ir H2 Q a
H\ -/H HM. B \\H O\ X /
l 1 BH BH
H ,,,,,,, H r (I)
,,,,, “0‘0 0 ‘
um. :'
<5 9’» N“-l'.i a“ ' "I
‘1 E. ’ E/fll \\ 3(2' ‘
€;Zl.-i‘~~:j|'z'{ N. " d:o\ /OJ
0.’ ‘9" 3,? BH‘BH
H131 - I I
48
H g : ’Hlu. .MH
I H
47
Scheme 3.5

In direct contrast, decomposition of 47 in DMSO at 120 °C under Ar for 4 days
yielded a material with no hydridic hydrogen left as shown by its H' content and IR
spectrum. Its llB solid state MAS NMR chemical shift (-5.9 ppm) is slightly different
from the corresponding value in the solid-state decomposition product. In a parallel
reaction run in DMSO-d6, unreacted BHa' and completely alcoholyzed B(0R)4‘ were the
only boron species found by HB N MR in solution throughout the reaction. Such behavior
is expected for a reaction in which the BHX(0R)4-,( intermediates are more reactive than

the starting BH4' and susceptible to disproportionation.106

84

It is important to rule out such disproportionation in the solid-state. As depicted in
Scheme 3.5, pairs of trialkoxyborohydrides are presumably present in 48.
Disproportionation into dialkoxyborohydn'de and borate is therefore conceivable in
principle. For instance, in LiBH(0CH3)3, slow disproportionation into LiBH2(0CH3)2
and LiB(0CH3)4 was observed in THF.105 a’ 109 Table 3.1 presents the experimental llB
NMR chemical shifts for the BHX(0CH3)4.,{ (x = 0-4) anions, for comparison with the
experimental data found for 48. Unfortunately, no N MR data in the literature referring to
BH2(0CH3)2' were found. Therefore the 11B NMR chemical shifts for the same series
were calculated at the RHF/6-3 10* level (Table 3.1), finding good agreement between

the experimental103 ’ 108’ ”0 and calculated values for the known BHX(0CH3)4-,( anions.

Table 3.1. NMR data for BHX(0CH3)4-X', 47, and 48.

 

 

llB NMR 8“ (ppm) 7Li N MR 8" (ppm)
calculated experimental
BH4' -507 49.7 (solid)c 1.5 (solid Linn.)d
49.5 (solid)d
BH3(0CH3)' -26.4 -25.1 (THF)f -
BH2(0CH3)2' -12.8 - -
BH(0CH3)3' -10.8 -8.6 (THF)f -
B(OCH3)4' -l4.9 -152 (CH30H)° -
-13.0 (THF)f
47 - 48.0 (solid)d 2.8 (solid)d
48 initial - -39 (solid) 3.1 (solid)
annealed - -2.6 (solid) 1.0 (solid)

 

a relative to B(0CH3)3. b relative to LiCl. ° reference 103. d reference 108. ° reference
1103. freference llOb.

Compared to these values, however, our decomposed borohydrides exhibit 8 values that

are consistently shifted downfield. For example, decomposition of 47 or of its Na

85

analogue in DMSO yielded polymeric borates which showed llB solid state MAS NMR
peaks at -5.2 and -4.3 ppm," respectively, significantly different from the corresponding
values in B(0CH3)4'. However, for the solid state decomposition product 48 and its Na
analogue, 8 values of -3.9 and -6.4 ppm,‘ respectively, were observed, in better
agreement with the corresponding experimental value of -8.6 ppm in BH(0CH3)3'. The
only possibilities that can be ruled out unambiguously based on the available data are the
presence of BH3(0R)' or 8114' in 48, which would display distinctive chemical shifts
around -26 or -50 ppm, respectively. The chemical shifts for the other alkoxyborohydride
species are too close to each other for a safe conclusion to be drawn. It would be very
difficult to distinguish, especially in the solid state, between a BH(0R)3' structure and the
1:1 mixture of BH2(0R)2' and B(0R)4' that would result from disproportionation.

An experiment that would possibly differentiate between the two possibilities is to
decompose a sample of 47 that contains 50% deuterated borohydride. If any
disproportionation occurred during decomposition, mixed BHD(0R)2' species would
result, which should display IR bands significantly different from those of BH2(0R)2' or
BD2(0R)2'. Ab initio calculations at the RHF/6-3IG* level predict the BH and BD
stretching frequencies in BHD(0R)2' to be shifted by -10 and -53 cm], respectively,
relative to the corresponding values in dialkoxyborohydride or -borodeuteride,
respectively. The reference LiBDaoTEA (47a) was first synthesized. The solid-state
decomposition product resulting from this material (48a) exhibits the V31) at 48 cm'|
lower than the starting compound 47a (Table 3.2). This change is comparable with the

corresponding shift of -43 cm’1 accompanying the transformation of 47 to 48. The 50%

 

' Corrected by +0.7 ppm for adjustment to the B(0CH3)3 reference.1 '0“

86

deuterated complex (47b) was then obtained, starting from a 1:1 mixture of LiBH4 and
LiBD4 in THF, by precipitation with TEA. It exhibits VB" and Van values that are very
similar to the corresponding values in 47 and 47a (Table 3.2). The IR spectrum of its
solid-state decomposition product (48b) is virtually identical with the spectrum obtained
by addition of the spectra corresponding to 48 and 483 (Table 3.2). This result suggests
that no disproportionation occurs during the solid-state decomposition of 47, as a result of

the isolation and reduced mobility of the borohydride units in 47 and 48.

Table 3.2. IR data for LiBH4- TBA (47), LiBDaoTEA (47a), LiBH.t(50%D)-TEA (47b),
and their corresponding solid state decomposition products, 48, 48a, and 48b. LiBl-I4 is

included for comparison.

 

-1
IR VBH (BD) (cm )

 

47 2231, 2290, 2369
48 initial 2171, 2247, 2301
annealed 2340, 2368
473 1652, 1704
488 initial 1656
annealed 1615a
47b 1649, 1703, 1752
2226, 2292, 2384
48b initial 1656
2169, 2247, 2291
annealed 1613a
2340, 2369
LiBH4 2219, 2301, 2378

 

a uncertain value due to the low intensity of the absorption and overlap with other
vibration modes.

87

Thus, the decomposition process in the solid-state differs significantly from that in
solution, demonstrating again the ability of dihydrogen bonding to exert topochemical
control in the 0H---HB to O-B conversion.

Attempts to enhance the crystallinity of 48 by annealing at 120 °C under Ar
resulted in a decrease of the XRD peaks’ intensities, until they completely disappeared
after approximately 2 weeks, or 4 days if heated as 3 DMSO suspension. While the
hydridic content remained unchanged (one H'), the 11B and 7Li MAS NMR chemical
shifts moved 1.3 and 2.1 ppm downfield and upfield, respectively (Table 3.1). These
changes point up the metastable nature of this topochemically controlled product.
Moreover, the appearance of the BH stretching region in the IR spectrum changed
considerably (Table 3.2) and apparently irreversibly, as the BH stretching absorptions did
not revert to the original values even after 8 months at room temperature. A Van shift to
shorter wavelengths in metal borohydrides generally indicates a stronger BH‘---W
interaction,l '1 and therefore a similar situation could be present in 48. The observed shifts
in the solid-state HB and 7Li MAS NMR spectra are in agreement with this supposition.
0n the other hand, no disproportionation seems to have occurred, since again the IR
spectrum of annealed 48b is basically a summation of the corresponding spectra for 48
and 483 (Table 3.2). The result of a topochemical reaction, the structure of 48 was very
likely controlled by kinetic factors, and its subsequent annealing presumably allowed
structural relaxation into a thermodynamically more stable arrangement in which the
oppositely charged BH(0R)3’ and Li+ units are closer in space. The profound
modifications in the IR spectrum of 48 and the relatively small downfield shift in its HB

NMR spectrum are in good qualitative agreement with a BH(0CH3)3'-~Li+ model, whose

88

Optimized Cs symmetrical structure at the RHF/6-3IG* level is shown in Figure 3.7. Its
VBH frequency (scaled by 0.97 based on the ratio between the theoretical and
experimental values of the strongest absorption in BI-Lt’) is estimated at 2418 cm’l,
comparable with the 2368 cm'1 value found in 48 after annealing. The 11B NMR 8 value
of -10.7 ppm calculated for this model structure is virtually identical with the
corresponding value in BH(0CH3)3', predicting that the above mentioned interaction with

Li+ should have minimal influence over the “B NMR chemical shift.

H301 H CH3
Olfi;O-" do-Li = 1.785
Hap—o“ ,: dB-Li = 2.346

Figure 3.7. Calculated structure of LiBH(0CH3)3 with the interatomic distances in A.

The fresh or annealed 48 is inert to ethanol, phenol, or acetone. However, it reacts
with Pd(Ac0)2 in THF at room temperature to yield an amorphous hygroscopic green
powder, soluble in DMSO, and presumably containing Pdo, as indicated by deposition of
Pd black upon H20 hydrolysis. IR, 1H and 11B NMR spectroscopies showed the
consumption of the BH, and attachment of the acetate group to the B (See Experimental).
This material might be a very effective catalyst for various transformations, considering
the presumable high dispersion of the Pd centers on the polymeric borate support.

Mechanistic Study of the Topochemical Dihydrogen to Covalent Bonding

Transformation in LiBH4~TEA.112 There are a few fundamental issues to be considered

regarding the mechanism of the topochemical decomposition of 47 in particular, and of

dihydrogen-bonded complexes in general. As for many other solid-state reactions,

89

conventional concepts like concentration, reaction order, or molecularity have little
applicability here. More relevant in this context are the appearance and morphological
evolution of the product phase, as well as its compatibility and "solubility" in the reactant
phase (See Chapter 2). As thermally initiated solid state reactions that yield both solid
and gaseous products, decompositions of 47 (and other dihydrogen-bonded systems), are
complex processes involving not only chemical steps such as breaking and formation of
bonds, but also physical transformations like destruction of the initial lattice,
reactant/product solid solution formation (with possible separation of the product phase),
diffusion and desorption of H2, and heat transfer. The variations in the crystals’
morphology are of particular interest as they can provide critical information about the
mechanism of decomposition.1 '3 The mechanism of this solid state decomposition at the
molecular level is particularly relevant for the successful design of the next generation of
dihydrogen-bonded systems, and also for a better understanding of such fundamental
processes as proton transfer or covalent bond formation, extensively studied in the gas
phase and solution, but considerably less in the solid state.

Figure 3.8 presents typical optical micrographs showing crystals of 47 at different
stages of decomposition at 110 °C. The initial transparent crystals gradually become
opaque as the reaction progresses toward completion, suggesting the separation of a new
phase. There is no visible reaction front advancing through the crystal; instead, the
process appears to start randomly and proceed uniformly in the crystal bulk. The size and
morphology of the crystals remained virtually unchanged during decomposition,

practically eliminating the possibility of melting.

90

 

Figure 3.8. Typical microscopic view (transmitted light) of LiBH4-TEA solid-state
decomposition: (3) initial crystal, (b) 10 min at 110 °C, (c) final decomposed crystal;

Scale bar: 100 um.

91

When decomposition of 47 was monitored in situ by HB solid-state MAS NMR
spectroscopy, no other boron species than the initial BH.{ and the final
trialkoxyborohydride could be detected, probably because of the increased reactivity and
the short lifetime of the intermediate mono- and dialkoxyborohydrides. Integration of the
two well-separated peaks proved to be a suitable way to measure the extent of the solid-
state decomposition, and thus to study the kinetics of this topochemical reaction
independent of other physical transformations that occur during the process. The reaction
was studied at temperatures between 105 and 120 °C, using LiBH4-TEA samples from
the same batch for each experiment to eliminate any possible error introduced by sample
variation, as significant differences in decomposition rates was noticed among various
preparations, a feature characteristic of many solid-state reactions.95 The resulting
conversion-time curves (Figure 3.9) have the typical sigmoid shape characteristic for

most solid-state decompositions of the type Asond —) Bsohd + Cgas (See Section 2.3).94

 

 

 

 

0.9
0.8 +- x A ‘0300‘ Q
T xx“x°‘. DOD
0.7+ g? ‘
X A
0.6» and?
1‘ “
0.5+ xta?
“f"
a 0.4-“ x Q
nu.
0.3‘”x{‘~
02» a?
. x“
0.1}?
04: i i a
0 200 400 600 800
t(min)

Figure 3.9. Conversion (0t) vs. time curves for the solid-state decomposition of 47 : O =

105°C;D=110°C;A=115°C;X= 120°C.

92

The agreement of our data to various solid-state decomposition models (Section 2.3)94
were compared, and the best match was found for a nucleation and two—dimensional

”2 against time for different

growth mechanism. The corresponding plots of [- ln(l - (1)]
temperatures studied are presented (Figure 3.10), together with the obtained rate

constants k, the correlation coefficients of the linear regression analysis R2, and the

conversion ranges for which the Avrami-Erofeev law is obeyed (Table 3.3).

 

 

 

 

1.2
3 0.8 ‘r ,(‘x “A“ n D
C 06 _0_ Xx A on D
5 I “‘ an O. .
—'-: 0.4 -- x a . °°
ll xz?‘ u n .0 °
8 0.2 ~~ , 0°
70 0 h 1r .

0 50 100 150 200 250

t (min)

112

Figure 3.10. Plots of the Avrami-Erofeev law, [-ln( l-a)] against time, for the solid

state decomposition of 47: O = 105 °C; CI = 110 °C; A = 115 °C; x = 120 °C.

It could be speculated that the two-dimensional expansion of nuclei might originate in the
crystal structure of 47, consisting of one-dimensional dihydrogen-bonded ribbons linked
by conventional H-bonds in overall extended layers. Once decomposition has started, it is
more likely it will propagate within the same layer, where the H-bonding network is
disrupted, weakening thus the compactness of the crystalline environment. By
comparison, a similar mechanism, but with a three-dimensional growth of nuclei (n = 3),

gave a significantly worse match (av. R2 = 0.9852 vs. 0.9984 for n=2).

93

Table 3.3. Rate constants k for the solid state decomposition of 47, calculated from the
Avrami-Erofeev law for nucleation and two-dimensional growth in the specified

conversion ranges, together with the correlation coefficients of the linear regression

 

 

analysis R2.
Temp (°C) k (min’1)x104 Conversion Range R2
105 37 i 1.4 0.284-0.681 0.9980
110 56 i 2.6 0.090-0.480 0.9988
115 66 :1: 2.8 0.224-0.606 0.9979
120 111 i 0.9 0064-0531 0.9990

 

The measured rate constants for the solid-state decomposition of 47 at
temperatures between 105 and 120 °C obey the Arrhenius equation (R2 = 0.9755), as
illustrated by the linear dependence of log k against l/T (Figure 3.11). At temperatures
above 120 °C, the solid-state decomposition of 47 becomes faster than predicted by the
Arrhenius equation. A possible explanation is that the dissipation of the resulting heat
becomes too slow compared to decomposition rate, eventually leading to the
autoacceleration of reaction. An activation energy of 21.0 :1: 2.4 kcal/mol for the solid-
state decomposition of 47 can be estimated from our data. Similarly, using the Eyring
equation, the activation parameters for decomposition, AH“2 and AS“, were calculated to

be 20.1 :1: 2.4 kcal/mol and -16.8 d: 6.2 e.u., respectively.

94

 

5.5 —~

541. O

-lnk

4.5 «~ .

 

 

4 i #
0.0025 0.00255 0.0026 0.00265

in (K")

 

Figure 3.11. Arrhenius plot for the solid-state decomposition of 47 at 105-120 °C.

When the isoconversional method98 (Eq. 5, Section 2.3) was used, comparable activation

energies within experimental errors were obtained (Figure 3.12).

 

 

 

.a: Efiffilifigm
:T I I L I l L
0 05 0'2 ole 0!4 0!5 0!6 0.7

 

(I
Figure 3.12. Variation of activation energy with conversion for the solid-state

decomposition of 47, obtained using the isoconversional method.

However, analyses starting with t0,,- initial times set at various extents of reaction 0t,

yielded different Ea values. For instance, when ta,- was set for CL = 0.1, the resulting

95

activation energies varied between 14.9 and 24.1 kcal/mol. In comparison, the Avrami-
Erofeev (n = 2) model analysis did not show much variation.

Since the kinetic measurements were done by in situ 11B NMR spectroscopy
which allowed direct monitoring of the appearance of the final trialkoxyborohydride
product, independent of other physical processes such as nucleation, phase separation, or
diffusion and desorption of H2, and considering the fact that no phase transition occurred
prior to decomposition, as shown by powder X-ray diffraction, it is reasonable to assume
that the kinetic parameters found in this study are directly related to the chemical
transformations responsible for decomposition.114 Moreover, the present activation
parameters are comparable with the activation enthalpy of 20.6 i 1 kcal/mol and
activation entropy of -22.3 i 3 en. found for the hydrolysis of ER; in neutral water, and
associated with the rate limiting proton transfer step.loo However, the mechanism could
be different in the solid-state, with the H2 evolution slowed down by the crystal
constraints, possibly becoming the rate-determining step. To address this question, we
studied the solid state decomposition of LiBH4-N(CH2CH2OD)3, the analogue of 47, with
the OH groups deuterated. H/D exchange between Eli; and OD groups of the TEA is
expected for a scheme involving fast, reversible proton transfer followed by slow H2 loss.
However, no HID exchange was observed at any stage of decomposition, under various
conditions, as indicated by the IR or 1H NMR of the partly or fully decomposed
deuterated complex. This experiment suggests that as in solution, the proton transfer is
slow compared to H2 loss and B-0 covalent bond formation (Figure 3.13). It should be
stressed here that these results are interpreted using a stepwise mechanism for the solid

state decomposition based on analogy with the aqueous hydrolysis of BH4' in neutral

96

water, and theoretical calculations supporting the existence of the BH5 intermediate'o' In
solution, a one step reaction with a four center transition state could be unambiguously
eliminated based on kinetic isotope effects and hydrolysis experiments in the presence of
trimethylammonium ion.l '5 In our solid-state system, on the other hand, the possibility of
a concerted mechanism cannot be ruled out. However, considering the similarity of the
reacting partners and the activation parameters found in this study and in solution, it

appears the same mechanism seems likely to apply for both.

 

 

H\
BH
3 BH
,H 5 H
N f? / WW
N~--.:l‘|_'. slow : Elu“2i,-I+ fast N~-~-:l'_l+
/,;’| ‘/.;’\| ‘/,;’\l
O O O O O O
\ I \ I \ I
HH HH HH

Figure 3.13. Proposed mechanism for the first O-H---H-B to O-B topochemical

conversion in 47.

To all appearances, the measured activation parameters in this study correspond to
rate-limiting proton transfer at the interface between the crystalline dihydrogen-bonded
system 47 and the newly formed covalent product 48. The reacting partners in this region
are presumably more flexible than in the perfectly ordered environment of the
undisturbed crystal. The somewhat lower activation entropy found for the solid-state
process, compared to the similar reaction in water, could be attributed to the

preorganization imposed by the dihydrogen bonding network in the crystal. Despite the

97

heterogeneous nature of decomposition, the topochemical information is nonetheless
transferred through the interface to the newly formed covalent product.

Toward Crystalline Covalent Solids: Crystal-to-Crystal Dihydrogen to
Covalent Bonding Transformation in NaBI-L-THEC (49).l '6 The N aBH4-THEC
(THEC = NNJV”,N”’-tetrakis—(2-hydroxyethyl)cyclen) complex exemplifies a system that
combines both the closed loop and the globular cation strategies, leading as a result for
the first time to a crystalline covalent product, in a crystal-to-crystal thermal

decomposition.

The complex precipitates from a 1:1 mixture of N aBH4 and THEC in 2-propanol, and its
crystal structure is shown in Figure 3.14. The THEC ligand adopts a conformation with
all hydroxyethyl arms oriented toward the same face of the azacrown ring, as observed in
previously reported metal complexes containing this ligand.117 The nitrogen and oxygen
atoms are thus organized in a pseudo-cubic geometry, encapsulating the octacoordinated
Na cation. The N a+THEC units have thus the appropriate geometry for the formation of

the desired closed-looped dihydrogen-bonded assembly.

98

,3;

  
 
 
 

012AA)
‘92

,W0(1AA)" .."- .', j. m
i i ‘3 011A) ' ‘j ""

 

Figure 3.14. X-Ray crystal structure of N aBIheTHEC illustrating the self-assembly in

dihydrogen-bonded dimers.

Indeed, these cations self assemble in D2 symmetrical dimers, held together by four

conventional O(2)-H---O(1) hydrogen bonds, complemented by four orthogonal O(1)-

HmH-B proton-hydride interactions involving the borohydride anions (Figure 3. 14). The

observed H---H distance is 2.00 Along, and the OH---H-B and O-H---HB angles display

values of 115 and 151°, respectively. There is no hydrogen bonding interconnecting the

dimers, which are packed in two-dimensional layers (Figure 3.15). The layers are stacked

in a parallel fashion, creating thus one—dimensional channels along the (1 axis, in which

the BH4' anions reside.

99

 

 

 

 

 

 

Figure 3.15. Crystal packing in NaBH4-THEC.

The association in closed loops, as well as the packing mode of the large globular
{Na2(TI-IEC)2}+2 cations, appear to fulfill the two geometrical requirements that were
predicted to favor the transfer of crystallinity from dihydrogen- to covalent—bonded
networks. Solid—state decomposition of 49, induced by heating under an inert atmosphere
at 160 °C for 20 h, resulted in complete elimination of H2, as indicated by the FT-IR and
HB solid state MAS-NMR spectra of the resulting material. Thermogravimetric analysis
showed a remarkably sharp 1.88 % weight loss between 168.0 and 168.8 °C, which

corresponds to 3.63 mol H2. The solid-state decomposition product exhibits high

crystallinity, as indicated by microscopic examination, which showed good transparency
to polarized light for the decomposed crystals (Figure 3.16). The reaction appears to be
crystallographically homogeneous, in direct contrast to the LiBH4-TEA dihydrogen-

bonded system, where the covalent product nucleated as a separate phase.

 

Figure 3.16. Polarizing microscopic view of NaBH4-THEC crystal-to—crystal

decomposition: a) initial crystal; b) after 20 h at 160 °C. Scale bar = 100 um.

That the solid state decomposition of 49 is a single-phase transformation is also
supported by powder XRD analysis, which shows a gradual shift of the diffraction pattern
as the reaction progresses. As depicted in Figure 3.17, the final covalent product exhibits
high crystallinity, in agreement with the microscopic observations. For comparison, an
amorphous solid results when 49 is decomposed in the melt at 200 °C for 10 min (Figure

3.17c).

101

3)

 

 

 

b)
C)

”W" 2 l l TMWTWH 7
0 1 30 40 50

20
20/0 —————>

Figure 3.17. Powder XRD patterns for: a) dihydrogen-bonded complex 49; b) solid-state

decomposition product; c) product of decomposition in the melt.

The unit cell parameters for the decomposition product, calculated using the powder
XRD data, are: a = 8.600(34) A, b = 13.86404) A, c = 16.895(27) A, and v: 2014(11)
A3. They correspond to 3.3, 2.5, 3.5, and 9.0 % shrinkage, respectively, relative to 49.
For this comparison, the following values derived from powder XRD at room
temperature were used for 49: a = 8.894(7), b = l4.214(13), c = 17.509(17), V= 2214(3).
The evolution of unit cell parameters during the solid-state decomposition of 49 at 120 °C

could be monitored by powder XRD (Figure 3.18).

102

V/A‘

 

 

_4p_

50

4.

50

t/h

t/h

Figure 3.18. Evolution of unit cell axes (a) and volume (b) during the solid-state

decomposition of NaBH4-THEC at 120 °C, monitored by powder XRD.

103

As shown in Figure 3.18, the initial crystal lattice smoothly transforms into the
corresponding lattice of the product, indicating a crystal-to-crystal process.

Despite sustained efforts, the detailed structure of the decomposed crystal has
remained elusive to date. The relatively big change in the unit cell accompanying the
solid-state decomposition ultimately results in deterioration of the single crystals’ quality,
precluding their X-ray structural investigation. Also, the unexpected insolubility of the
product in common organic solvents does not allow its recrystallization. Considering the
crystal packing of 49, with no interdimer hydrogen bonding present, formation of discrete
molecular cages is expected from the solid-state decomposition. Crosslinking between
dimers would require substantial intermolecular rearrangements, which is highly unlikely
in a crystallographically homogeneous, lattice-controlled transformation. Figure 3.19
presents a possible structure for the putative molecular cages, optimized with the DFT
method at the B3LYP/6-3 10* level.1 '8 No unusual values for the bond lengths and angles
were observed in the resulting C2 symmetrical structure, suggesting minimal strain in
these covalent dimers.

When 49 was decomposed in DMSO by heating at 120 °C for 48 h, a crystalline
product precipitated in high yield. X-ray structural analysis of the resulting crystals
revealed a centrosymmetric molecular cage very similar to that expected for the solid-
state reaction (Figure 3.20). However, one of the hydroxyethyl arms in each of the two
THEC ligands in the dimer does not bind the boron atom, nor does it complex the Na
cation, being replaced by an OH group, presumably from water present in the solvent.
Except for the Na, the four N atoms, and C(4) and C(6), all the atoms in the structure are

disordered over two sites, with 90.1 and 9.9 % occupancies, respectively.

104

 

Figure 3.19. Molecular model of the product from the solid-state decomposition of 49;
selected bond lengths [A] and angles [°]: B-O(1) 1.458, B—O(2) 1.482, B-O(3) 1.497, B-
O(4) 1.474, O(1)-B-O(2) 112.5, O(1)-B-O(3) 109.1, O(1)-B-O(4) 114.6, B-O(1)-C 123.7,
B-O(2)-C 122.3, B-O(3)-C 125.9, B-O(4)-C 125.6, N(l)-Na 3.003, N(2)-Na 2.714, N(3)-

Na 2.435, N(4)-Na 2.565, O(2)-Na 2.670, O(3)-N a 2.276, C(4)-Na 2.228.

105

   

‘nnt,
.. AD N12A1

’30»
'v

  

Figure 3.20. X-Ray crystal structure of the major conformation found in the covalent
dimer obtained by decomposing 49 in DMSO; selected bond lengths [A] and angles [°]:
B-O(1) 1.502(2), B-O(2) 1.511(2), B—O(3) 1.450(2), B-O(5) 1.457(2), O(1)-B-O(5)
111.6(1), O(2)-B-O(5) 109.8(1), O(3)-B-O(5) 111.6(1), B-O(1)-C 121.7(1), B-O(2)-C
113.2(1), B-O(5)-C 122.8(1), N(1)-Na 2.519(1), N(2)-Na 2.540(1), N(3)-Na 2.565(1),
N(4)-Na 2.682(1), O(1)-Na 2.371(1), O(2)-Na 2.796(1), O(3)-Na 2.391(1), C(4)-H

082(2), O(4)H---O(2) 1.92(3), O(4)---O(2) 2.735(4), O(4)-H-O(2) 175.1(3).

106

Like the solid-state decomposition product, the compound obtained from DMSO is
surprisingly insoluble in common organic solvents, despite its relatively small size,
overall neutral charge, and lack of intermolecular hydrogen bonding.

In situ llB MAS NMR spectroscopy proved again to be a convenient method for
monitoring the solid-state dihydrogen to covalent bonding transformation and measuring
the kinetics of decomposition. As in the case of the heterogeneous transformation in

LiBH4-TEA, no boron species other than the initial borohydride and the final borate could

be observed. The extent of reaction at 160 °C could be easily measured by integration of
the two peaks, and the conversion-time curve is depicted in Figure 3.21. After an
initiation period, the reaction is accelerated up to approximately 75% conversion, after
which time it is slowed down significantly. No common kinetic model for solid-state
decompositions was found to fit the existing data. For comparison, the extent of lattice
conversion vs time was plotted, using the data from the solid-state decomposition of 49 at
120 °C, monitored by powder XRD. As illustrated in Figure 3.22, the obtained curve is
qualitatively comparable with the one obtained from the llB NMR measurements. The
enhanced reaction rate at the intermediate stages of decomposition might originate in the
increased strain caused by the enforced coexistence of the initial and final crystal lattices.
On the other hand, powder XRD examinations at different extents of decomposition
showed a gradual deterioration of crystallinity as the reaction progressed, followed by
significant recovery of the diffraction pattern at the late stages of reaction. This increased
lattice looseness at intermediate conversions appears to facilitate the chemical reaction
relative to the initial and final stages, when the reacting molecules are constrained in a

more rigid, unperturbed crystalline lattice.

107

 

 

 

 

1
00”
0.8 _. o ’
.0
0.6 ~~ v’
5 04 ¢
. 1r
0”
0 2 —— O ,o o
0 e i
0 100 200 300
t (min)

Figure 3.21. Conversion (or) vs time curve for the solid-state decomposition of

NaBH4-THEC at 160 °C, obtained using in situ “B MAS NMR.

 

 

 

 

1
I

0.8 . .
$- - - '
g: 0.6 7
Z
>.c 0.4 « ' '
>

0.2. _

o a . .
0 50 100 150
Nb)

Figure 3.22. Lattice conversion (VO‘V/Vo-Vf) vs time curve for the solid-state

decomposition of NaBHa-THEC at 120 °C, obtained using powder XRD.

108

In summary, the first crystal-to-crystal dihydrogen to covalent bonding
transformation was carried out. This outcome appears to be the result of the judicious
engineering of the starting dihydrogen-bonded system, which was designed to satisfy
both the closed loop and globular cation geometrical criteria.

Structural Variations from the NaBI-h-THEC System. The N aBHa-THEC

complex demonstrated that careful design of dihydrogen-bonded crystals permits transfer
of crystallinity to the covalent products resulting from their solid-state decomposition.
The next step is to extend this concept to the construction of higher-dimensional
crystalline covalent systems, a class of compounds to which few purposeful synthetic
paths exist. Further elaboration of the dihydrogen-bonded building blocks present in 49,
which proved to possess the required geometrical and solid-state reactivity prerequisites,
appears to be the most potentially successful approach toward our objective. However, an
important and necessary condition is that the dihydrogen-bonded dimeric motif observed
in 49 be robust enough to survive substantial modifications in molecular structure. This
section presents some preliminary results from the systematic investigation of the
relationship between molecular structure and supramolecular organization in metal
borohydride complexes with THEC-type ligands. We seek an understanding of the
factors that govern the crystal packing in these dihydrogen-bonded systems, and possibly
to control them for the crystal engineering of the desired extended dihydrogen-, and
ultimately covalent-bonded networks.

We first explored the role of the metal borohydride in determining the self-
assembly into dihydrogen-bonded dimers. The LiBH4 and KBH4 complexes with THEC

were therefore synthesized to investigate the effect of changing the metal cation upon

109

crystal structure. Unfortunately, the modest crystallinity of the former, and the propensity
of the latter to form polycrystalline aggregates instead of single-crystals, precluded their
detailed structural elucidation. The NaCNBH3~TI~IEC (50) complex was then synthesized
to explore the contribution of the hydridic partner in the formation of the closed-looped
dimers. The white crystalline compound melts at 230-234 °C with decomposition and gas
evolution. However, its crystal structure, which is depicted in Figure 3.23, reveals the
absence of any O-HmH-B close contacts, the cyanoborohydride anion participating with

the N atom via conventional H-bonding instead.

 

Figure 3.23. X-Ray crystal structure of NaNCBHyTHEC.

110

 

Figure 3.24. a-Helix formed by N aNCBH3-THEC in the solid-state.

lll

Notably, instead of forming dimers, in this case the Na-THEC units self-assemble into
extended a-helices, via bifurcated O-H---N---H-O hydrogen bonds involving the CN
groups (Figure 3.24). The helices are further associated in two-dimensional layers by O-
H---O interactions. It thus appears that the O-H---H-B dihydrogen bonds in 49 are
essential for the formation of the {N a2(THEC)2}+2(BH4')2 dimers, as their absence in 50
resulted in a completely different packing.

NaBHa-THPC (51) (THPC = 1,4,7,10-tetrakis((S)-2-hydroxypropyl)-1,4,7,10-
tetraazacyclododecane) was synthesized to explore the effect of subtle molecular
modifications in the ligand upon crystal packing. It was also hoped that the introduction
of the Me groups would improve the solubility of the covalent dimers (if they are formed)

in organic solvents.

H,’ Me
Me e\<\NH 473$
H OH Na H0 3H4-

$7? 3%.

51

The complex precipitates as a white solid upon mixing NaBHa and THPC in 2-propanol.
Its crystal structure is presented in Figure 3.25. The most notable feature is the formation
of hydrogen-bonded dimers, which, however, display an S-shaped topology, as opposed

to the O-shaped one in 49. The most probable cause of this difference is the steric

repulsion that would result between the Me groups, in an O-shaped configuration in 51.

112

 

Figure 3.25. X-Ray crystal structure of NaBHa-THPC; dihydrogen bonding parameters
(A, °): O(4)H-~-HB(1) (lower) 2.106, O(4)H---HB(1) (upper) 2.266, O(3)H---HB(1) 2.005,
O(7)H---HB(2) 1.927, O(8)H---HB(2) 1.676, O(2)H~-HB(2) 1.904, O(4)H-H-B(1) (lower)
96.1, O(4)H-H-B(1) (upper) 88.7, O(3)H-H-B(1) 113.8, O(7)H-H-B(2) 110.7, O(8)H-H-
B(2) 120.2, O(2)H-H-B(2) 133.3, O(4)-H-HB(1) (lower) 162.5, O(4)-H-HB(1) (upper)
143.9, O(3)-H-HB(1) 153.4, O(7)-H-HB(2) 155.4, O(8)-H-HB(2) 153.0, 0(2)-H-HB(2)

169.3.

113

The two THPC units in the dimer are held together by O(1)H---O(6) and O(5)H---O(2)
conventional hydrogen bonds, as well as by O(7)H-~HB(2)H~~-HO(2) and
O(8)H---HB(2)H---HO(2) dihydrogen bonds. The second borohydride anion interacts with
the remaining O(3)H and O(4)H from one THPC ligand. As in 49, no hydrogen bonding
interconnects the dimers.

The solid-state decomposition of 51 was unfortunately precluded by partial
liquefaction during heating, apparently due to the low melting point of the product(s).

More profound structural variations in the THEC supramolecular building block
were brought about by replacing one of the hydroxyethyl arms with a 9-methylanthracene
group. The intention was to test the ability of the resulting tripodal ligand to maintain the
association in dimers. Recurrence of this supramolecular motif could ultimately open the
way to extended networks, by simple attachment of the tris-hydroxycyclen units to
multidentate spacers, using the remaining N atoms. Furthermore, monitoring the
orientation of the anthracene chromophores via fluorescence spectroscopy might offer an
additional handle on the evolution of order during the solid-state decomposition.

The 9-(1’,4’,7’-tris-hydroxyethylcyclen)methylanthracene (52) ligand was
synthesized from 9-(1’,4’,7’,10’-tetraazacyclododecyl)methylanthracenel '9 and ethylene
oxide (Scheme 3.6), and its structure was confirmed by X-ray crystallography (Figure
3.26). Upon mixing with NaBHa in 2-propanol, the NaBH4-52 complex (53) was formed,
which precipitated as a white crystalline solid. Elemental analysis and 1H NMR indicated
the crystallization with 1 mol of solvent. The inclusion of 2-propanol was also supported
by TGA, which showed a 10.35% weight loss, in good agreement with the 10.14%

theoretical value. Single crystals, however, were grown from CH3CN/Et2O.

114

A O ”OH O

 

0 ° ‘ ‘ O
N N
E * E
N 1:] EtOH, H20 N :1

\__/ \_\

HO OH
52
Scheme 3.6

 

Figure 3.26. X—Ray crystal structure of 9-(1’,4’,7’-tris-hydroxyethyl-

cyclen)methylanthracene.

115

 

Figure 3.27. X-Ray crystal structure of 53; normalized dihydrogen bonding parameters
(A, °): O(1)H---HB 1.692, O(2)H---HB 1.661, O(3)H~--HB 1.730, O(1)H---H-B 114.5,
O(2)H--~H-B 117.8, O(3)H-~H-B 116.1, O(1)-H--~I-IB 166.8, O(2)-HmHB 155.1, O(3)-

H---HB 154.6.

116

X-Ray structural examination of crystals of 53 obtained from acetonitrile-ether revealed
the inclusion of 1 mol of disordered 320. The crystal structure (EtzO was omitted) is
depicted in Figure 3.27, and shows the self—assembly in dihydrogen-bonded
centrosymmetric dimers. However, compared with the previous structures, no O-H---O
interaction is present in this case; instead, the BH4' anions are intercalated between the
two ligands, forming three dihydrogen bonds with the OH groups. No conclusive results
from the solid-state decomposition of 53 have been obtained to date, due to the
insufficient quantity of its ether clathrate available.

In summary, NaBH4 and various hydroxyethylcyclen ligands persistently self-
assemble in dihydrogen-bonded dimers in the solid-state, as demonstrated by the crystal
structures of 49, 51, and 53. However, the exact topology of the dimers varies from one
case to another, in response to structural modifications in the ligand. Nevertheless, this
consistent recurrence of the general dimeric motif makes these dihydrogen-bonded units
sufficiently reliable building blocks for supramolecular synthesis. Their incorporation
into more elaborate structures, as suggested in Figure 3.28 would offer real prospects for
the construction of extended dihydrogen-bonded, and ultimately covalent solids, with

controlled architectures.

117

 

Figure 3.28. Prospective multidentate ligands for the construction of extended

dihydrogen-bonded and covalent networks.

118

3.2. Conclusions

0 Dihydrogen bonds can play an important role in defining solid-state
supramolecular structures, comparable to conventional hydrogen bonds, as demonstrated
by the crystal structures of all H---H bonded compounds studied, where this interaction
significantly influenced the arrangement of molecules in crystals. Thus, with its capacity
to control crystal packing, dihydrogen bonding, like traditional H—bonding, finds its place
among the tools that chemists can use for supramolecular synthesis. The additional
feature that makes this unconventional H-bonding particularly interesting is the ability to
react, trading the weak HmH interactions for strong covalent bonds. As a direct
consequence of this character, dihydrogen bonding has the potential to induce
topochemical control, providing access to new materials otherwise not achievable.

0 The NaBH4-TEA complex has demonstrated the topochemical control concept.
Its crystal structure shows multiple dihydrogen bonds between 3114' and the three OH
groups from TEA. Loss of 3 moles of H2 in the solid-state leads to a polymeric
trialkoxyborohydride structure, whereas solution or melt decompositions yield NaBH4
and a polymeric borate with no hydridic hydrogen left. Remarkably, more H2 is lost in
the solid-state decomposition than in the melt, despite the much higher reaction
temperature in the latter (130 °C vs 82 °C). However, the product of the solid-state
decomposition showed poor crystallinity and the reaction times were exasperatingly long
due to the relatively low temperature required for decomposition in order to avoid

melting.

119

o In addition to close H---H contacts, the relative acidity/basicity of the proton-
hydride pairs significantly affect the solid-state reactivity of the dihydrogen-bonded
systems. This was demonstrated by changing the H’ donor from BI-L{ to CNBH3‘ in the
TEA complexes, which resulted in substantial decreases in dihydrogen bonding and
solid-state reactivity.

o Fine-tuning of the dihydrogen bonds was possible in the TEA complexes by
replacing the Li” for Na+, which resulted in exceptionally short HmH contacts and
enhanced solid-state reactivity. Decomposition of LiBI-IarTEA is also topochemical,
apparently leading to a metastable one-dimensional polymeric trialkoxyborohydride
structure, in direct contrast to decomposition in DMSO solution, which yielded a
polymeric hydride free borate and unconverted LiBI-I4, the disproportionation products.
For comparison, no such disproportionation occurs during the solid-state decomposition,
as demonstrated by H/D isotopic labeling experiments.

0 The high solid-state reactivity of LiBH4-TEA allowed the detailed study of the
mechanism of H2 loss and covalent bond formation in this dihydrogen-bonded complex,
both at macroscopic and molecular level. Using in situ solid-state llB NMR and optical
microscopy, it was found that, like most solid-state reactions, this decomposition is
heterogeneous, with separation of the product phase from the parent crystal. Kinetic
analysis and H/D exchange experiments established that proton transfer between the OH
groups of the TEA and the BHa' anions, at the reactant/product interface, is the rate-
limiting step, with activation parameters AHat and AS"e of 20.1 i 2.4 kcal/mol and —l6.8 i

6.2 eu. These values are comparable with the analogous values found for the aqueous

120

hydrolysis of BHa' in neutral water, suggesting similar mechanisms for the solid and
solution decompositions.

o Preservation of crystallinity was demonstrated to be possible in the
topochemical dihydrogen to covalent bonding transformation of N aBH4-THEC, which
was designed to satisfy both the closed loop and large globular cation geometrical
criteria. Remarkably, its solid-state decomposition is a crystal-to-crystal process, as
indicated by polarizing microscopy and powder XRD. However, the observed 9%
shrinkage of the unit cell led to significant deterioration of the single crystals’ quality,
which in conjunction with the insolubility of the covalent product in organic solvents,
precluded its detailed structural characterization. Nevertheless, this study is the first
example to demonstrate that judicious engineering of dihydrogen-bonded crystals permits
transfer of crystallinity to the covalent products resulting from their solid-state
decomposition.

0 Preliminary results suggest that NaBH4 and various hydroxyethylcyclen ligands
consistently self-assemble in dihydrogen-bonded dimers, as demonstrated by the crystal
structure analysis of NaBI-14~TI~{EC, NaBH4-THPC, and N aBHa-[9-(l’,4’,7’-tris-
hydroxyethyl-cyclen)methylanthracene]~Et2O. Consequently, incorporation of these
dimeric building blocks in more elaborated systems may allow the rational construction
of extended dihydrogen-bonded networks, and ultimately covalent solids, with controlled

structures and functions.

121

3.3. Experimental Section

Materials. Common chemicals were purchased from commercial sources and
used without further purification. N—[2-(6-Aminopyridyl)]-acetamidine hydrochloride
was prepared by a literature procedure. '20 TEA-d3 was obtained by repeated dissolution
of TEA in D2O (99.9% D) followed by removal of the water at 85 °C under vacuum, until
the 1H N MR (DMSO-d6) signal for the OH protons completely disappeared. LiBD4 was
synthesized by metathesis,121 from NaBD4 (98%D). THEC was prepared by a literature
procedure. 122 1 ,4,7, 10-Tetrakis((S)-2-hydroxypropyl)- 1 ,4,7, 10-tetraazacyclododecane)
(THPC) was synthesized from cyclen and S(-) propylene oxide.123 9-(1’,4’,7’,10’-
Tetraazacyclododecyl)methylanthracene was obtained by a literature procedure.“9

Physical and Chemical Characterizations. Melting points were taken on a
Thomas Hoover instrument and are uncorrected. FT-IR spectra were measured in KBr
pellets on a Perkin Elmer Spectrum 2000 instrument. 1H N MR (300.1 MHz) spectra were
recorded on a Gemini-300 instrument. l3C NMR (75.4 MHz) and “B (96.23 MHz) NMR
spectra were recorded on a Varian VXR-300 instrument. “8 (128.33 MHz), 23Na (105.81
MHz), and 7L1 (155.44 MHz) solid-state MAS NMR spectra were measured on a Varian
VXR-400 instrument. llB solution NMR chemical shifts were referenced to B(OCH3)3 in
CDC13. llB solid state MAS NMR chemical shifts were referenced to solid boric acid;

23 N a and 7Li chemical shifts were referenced to solid N 8C] and LiCl, respectively. A
negative sign indicates chemical shifts upfield from references. X-ray powder diffraction

measurements were conducted on a Rigaku-Denki RW400F2 diffractometer with

monochromatic Cu Kat radiation, operated at 45 kV and 100 mA. Optical micrographs

122

were obtained with a Nikon AFX-DX microscope equipped with a Mettler FP82HT hot
stage. TGA was recorded on a CAI-IN TGSystem 121 instrument at a heating rate of 5
°/min.

Hydridic Content Analyses. A 30-120 mg sample of the hydride-containing
material was placed in a 50 mL side-armed round bottom flask, connected by a rubber
tube to a 50 mL burette, half-filled with distilled water. About 10 mL of diluted HCl (1-
2%) was rapidly added from a 25 mL addition funnel. The amount of the evolved H2 was
obtained by measuring the volume of the displaced water in the burette, after adjusting
the gas pressure in the flask (by rising it to the corresponding height) to the atmospheric
pressure. Typical uncertainty: :1: 0.3 H' equivalents.

In Situ llB MAS Solid State NMR. All spectra were recorded at a MAS
frequency of 3.8 kHz, using 2.0 [.18 N2 pulses, with a recycle delay of 5 s, to allow full
relaxation of all boron species.124 For each experiment, 48 scans were acquired. The
observed spectra were deconvoluted into Lorentzian lines, and the ratios of the resulting
integrals, corresponding to product and starting material, respectively, were used to
estimate the extent of decomposition.

Theoretical Calculations. All calculations were carried out with the GAUSSIAN
94 and 98 packages.'25 All structures were fully optimized and confirmed (except for the
model of the decomposition product of 49) by vibrational analysis to be minima on the
potential energy surface.

Single Crystal X-Ray Structure Determination. X-Ray crystallographic
measurements were carried out on a Siemens SMART CCD diffractometer with graphite-

monochromated Mo Kat radiation (A = 0.71073 A), operated at 50 kV and 40 mA. The

123

structures were solved by direct methods and refined on F2 using the SHELXTL software

'26 Absorption corrections were applied using SADABS, part of the SHELXTL

package.
software package. All non-hydrogen atoms were refined anisotropically. Hydrogen atoms
were located from difference Fourier maps and refined isotropically for all the structures.
For the borohydride ions in 49, however, only the hydridic hydrogens involved in
dihydrogen bonding could be found, and the remaining H atoms from 8114' were
calculated and placed in idealized positions. Similarly, for B(2)H4' in 51, the two
noninteracting hydridic hydrogens were calculated, and, in this case, the 4 H were

subsequently constrained in an idealized tetrahedral geometry, with B-H distances of 1.21

A.

124

Table 3.4. Crystal data and structure refinement for NAPA H3BCN (44).

 

Empirical formula
Formula weight
Temperature
Wavelength

Crystal system
Space group

Unit cell dimensions

Volume

Z

Density (calculated)
Absorption coefficient
F(000)

Crystal size

0 range for data collection
Index ranges

Reflections collected / unique
Refinement method

Data / restraints / parameters
Goodness-of-fit on F2

Final R indices [I>20(I)]

R indices (all data)
Extinction coefficient
Largest diff. peak and hole

125

C8H14BN5

191.05

143(2) K
0.71073 A

Triclinic

P]

a = 7.104(2) A

b = 9.390(4) A

c = 17.736(10) A

a = 99.38(4)°

13 = 95.99(4)°

y = 111.53(3)°

1068.3(9) A3

4

1.188 g/cm3

0.077 mm"

408

0.6 x 0.3 x 0.15 mm

2.37 to 28.47°
-8$h$8,-11$k$12,-10$1$23
3118 / 3095 [R(int) = 0.0469]
Full-matrix least-squares on F
3095 IO / 366

1.095

R, = 0.0455, sz = 0.1130
R1 = 0.0577, sz = 0.1238
0.022(3)

0.262 and -0.163 eA'3

Table 3.5. Crystal data and structure refinement for NaBHa-TEA (45).

 

Empirical formula
Formula weight
Temperature
Wavelength

Crystal system
Space group

Unit cell dimensions

Volume

Z

Density (calculated)
Absorption coefficient
F(000)

Crystal size

0 range for data collection
Index ranges

Reflections collected / unique
Refinement method

Data / restraints / parameters
Goodness-of-fit on F2

Final R indices [I>20'(I)]

R indices (all data)
Extinction coefficient
Largest diff. peak and hole

126

C6H|9BNNaO3
187.02
173(2) K
0.71073 A
Monoclinic
P21/Il
a = 9.23510(10) A
b = 7.35120(10)A
= 15.8083(2) A
or = 90°
8 = 103.6880(10)°
Y = 90°
1042.73(2) A3
4
1.191 g/cm3
0.123 mm"
408
0.52 x 0.38 x 0.31 mm
2.34 to 28.15°
-11$h$12,-9Sks9,-20.<_1S.20
9787 / 2482 [R(int) = 0.0211]
Full-matrix least-squares on F2
2482 / 0 / 186
1.060
R1 = 0.0250, sz = 0.0726
R. = 0.0292, sz = 0.0743
0.0102(31)
0.357 and -O.165 eA'3

Table 3.6. Crystal data and structure refinement for NaNCBH3-TEA (46).

 

Empirical formula
Formula weight
Temperature
Wavelength

Crystal system
Space group

Unit cell dimensions

Volume

Z

Density (calculated)
Absorption coefficient
F(000)

Crystal size

0 range for data collection
Index ranges

Reflections collected / unique
Refinement method

Data / restraints / parameters
Goodness-of-fit on F2

Final R indices [I>20'(I)]

R indices (all data)

Largest diff. peak and hole

127

C7H13BN2NaO3

212.03

173(2) K

0.71073 A

Monoclinic

sz/C

a = 7.55160(10)A

b = 15.1232(2) A

c = 10.57070(10) A

or = 90°

[3 = 105.5380(10)°

Y = 90°

1163.10(2) A3

4

1.211 g/cm3

0.121 mm'1
456
0.62 x 0.18 x 0.10 mm

2.41 to 28.17 deg.
-9.<.h510,-195k$20,-13515 14
9974 / 2753 [R(int) = 0.0549]
Full-matrix least-squares on F2
2753 / 0/ 199

1.053

R. = 0.0352, sz = 0.0891
R. = 0.0442, sz = 0.0945
0.232 and -0.219 eA'3

Table 3.7. Crystal data and structure refinement for LiBHa-TEA (47).

 

Empirical formula
Formula weight
Temperature
Wavelength

Crystal system
Space group

Unit cell dimensions

Volume

Z

Density (calculated)
Absorption coefficient
F(000)

Crystal size

0 range for data collection
Index ranges

Reflections collected / unique
Refinement method

Data / restraints / parameters
Goodness-of-fit on F2

Final R indices [I>2o(I)]

R indices (all data)

Absolute structure parameter
Extinction coefficient
Largest diff. peak and hole

128

C6HloBLiN03
170.97

173(2) K

0.71073 A
Monoclinic

P21

a = 7.7156(5) A

b = 17.0039(11) A
c = 7.8656(5) A

01 = 90°

13 = 98.1330(10)°
7: 90°
1021.55(11) A3

4

1.112 g/cm3

0.081 mm"

376
0.44 x 0.23 x 0.21 mm
2.40 to 25.00°

-9Sh$6,-l7$kS20,-8SIS9

5330/ 3229 [R(int) = 0.0339]
Full-matrix least—squares on F2
3229/ l / 370

0.979

R. = 0.0417, wR2 = 0.0768
R1 = 0.0619, wR2 = 0.0839
1.7(10)

0.015(3)

0.145 and -0.123 eA’3

Table 3.8. Crystal data and structure refinement for NaBHatTHEC (49).

 

Empirical formula
Formula weight
Temperature
Wavelength

Crystal system
Space group

Unit cell dimensions

Volume

Z

Density (calculated)
Absorption coefficient
F(000)

Crystal size

0 range for data collection
Index ranges

Reflections collected / unique
Refinement method

Data / restraints / parameters
Goodness—of-fit on F2

Final R indices [I>20(I)]

R indices (all data)

Absolute structure parameter
Extinction coefficient
Largest diff. peak and hole

129

CtoH4oBN4NaO4
386.32

173(2) K

0.71073 A

Orthorombic

1222

a = 8.8240(3) A

b = 14.1344(6) A

c = 17.35330) A

0: = 90°

B = 90°
Y = 90°
2164.34(15) A3

4

1.186 g/cm3
0.100 mm'1

848

0.23 x 0.18 x 0.13 mm
1.8610 28.20°
-11ShSll,-185k518,-225.1.<_22
12854 / 2608 [R(int) = 0.0979]
Full-matrix least-squares on F
2608/ 11 / 200

1.029

R1 = 0.0487, wR2 = 0.1221

R. = 0.0599, wR2 = 0.1399
0.7(5)

0.0027(9)

0.488 and 0250 eA'3

Table 3.9. Crystal data and structure refinement for the decomposition

product of NaBHa-THEC in DMSO.

 

Empirical formula
Formula weight
Temperature
Wavelength

Crystal system
Space group

Unit cell dimensions

Volume

Z

Density (calculated)
Absorption coefficient
F(000)

Crystal size

0 range for data collection
Index ranges

Reflections collected / unique
Absorption correction
Refinement method

Data / restraints / parameters
Goodness-of-fit on F2

Final R indices [I>20’(I)]

R indices (all data)
Extinction coefficient
Largest diff. peak and hole

130

C16H34BN4N305
396.27
173(2) K
0.71073 A
Monoclinic
P21/n
a = 8.7696000) A
b = 130571000) A
c = 17.7724(2) A
01 = 90°
13 = 96.36°
Y = 90°
2022.500) A3
4
1.301 g/cm3
0.112 mm'1
856
0.47 x 0.39 x 0.29 mm
1.94 to 28.26°
-11 shs 11,-16Sks 17,-23s1s22
20248 /4840 [R(int) = 0.0261]
None
Full-matrix least-squares on F2
4840 / 0 / 462
1.043
R. = 0.0352, wR2 = 0.0921
R1 = 0.0439, wR2 = 0.0966
0.0000(9)
0.261 and 0250 eA'3

Table 3.10. Crystal data and structure refinement for N aNCBH3-THEC (50).

 

Empirical formula
Formula weight
Temperature
Wavelength

Crystal system
Space group

Unit cell dimensions

Volume

Z

Density (calculated)
Absorption coefficient
F(000)

Crystal size

0 range for data collection
Index ranges

Reflections collected / unique
Refinement method

Data / restraints / parameters
Goodness-of-fit on 1:2

Final R indices [I>20'(I)]

R indices (all data)

Absolute structure parameter
Extinction coefficient
Largest diff. peak and hole

131

C 17H3gBN5NaO4
411.33

173(2) K

0.71073 A

Orthorombic

P112121

a = 25.2624(3) A

b = 9.4053(2) A

c = 9.4430(2) A

01 = 90°

B = 90°

7 = 90°

2243.660) A

4

1.218 g/cm3

0.102 mm"1

896
0.65 x 0.29 x 0.26 mm

1.61 to 28.20°
-32$h$32,-12$k$12,-12$1S 12
22515 / 5336 [R(int) = 0.0187]
Full-matrix least-squares on F2
5336/ 1 /410

1.022

R1 = 0.0251, sz = 0.0637

R1 = 0.0276, WRz = 0.0646
007(17)

0.0058(9)

0.205 and -0125 eA'3

Table 3.11. Crystal data and structure refinement for N aBHa-THPC (51).

 

Empirical formula
Formula weight
Temperature
Wavelength

Crystal system
Space group

Unit cell dimensions

Volume

Z

Density (calculated)
Absorption coefficient
F(000)

Crystal size

0 range for data collection
Index ranges

Reflections collected / unique
Refinement method

Data / restraints / parameters
Goodness-of-fit on F2

Final R indices [I>20’(I)]

R indices (all data)

Absolute structure parameter
Largest diff. peak and hole

C20H483N4N304

442.42

173(2) K

0.71073 A

Monoclinic

P21

a = 9.7915(4) A

b = 18.14490) A

c = 14.6111(6) A

01 = 90°

13 = 96.867(2)°

Y = 90°

25772708) A3

4

1.140 g/cm3

0.092 mrn'I

976

0.49 x 0.41 x 0.18 mm

1.80 to 28.24°
-125hS12,-24S.k_<.24,-19Sl_<. 16
23846 / 11681 [R(int) = 0.0198]
Full-matrix least-squares on F2
11681 I20 / 923

0.789

R1 = 0.0358, WRZ = 0.0968

R. = 0.0428, sz = 0.1023
0.09( 18)
0.517 and -0203 eA'3

132

 

Table 3.12. Crystal data and structure refinement for 9-(1’,4’,7’-tris-hydroxyethyl-

cyclen)methylanthracene (52).

 

Empirical formula
Formula weight
Temperature
Wavelength

Crystal system
Space group

Unit cell dimensions

Volume

Z

Density (calculated)
Absorption coefficient
F(000)

Crystal size

6 range for data collection
Index ranges

Reflections collected / unique
Refinement method

Data / restraints / parameters
Goodness-of-fit on F2

Final R indices [I>20'(I)]

R indices (all data)

Largest diff. peak and hole

133

C29H42N403
494.67
173(2) K
0.71073 A
Triclinic
P1
a = 8.7352(3) A
b = 1229710) A
c = 13.7789(3) A
at = 69.152(2)°
13 = 79.456(2)°
y: 78.736(2)°
1346.11(7) A3
2
1.220 g/cm3
0.080 mm'l
536
0.39 x 0.18 x 0.08 mm
1.59 to 28.27°
-11 shs 11,-16Sks 15,-18s15 17
12550 / 6195 [R(int) = 0.0383]
Full-matrix least-squares on F2
6195 / 0 I489
1.268
R. = 0.0825, wR2 = 0.2187
R. = 0.1467, wR2 = 0.2369
1.571 and -0.708 eA'3

Table 3.13. Crystal data and structure refinement for NaBHa-[9-( l ’,4’,7’-tris-

hydroxyethyl-cyclen)methylanthracene]rEt2O (53).

 

Empirical formula C33H56BN4Na04
Formula weight 590.62
Temperature 173(2) K
Wavelength 0.71073 A
Crystal system Monoclinic
Space group P21/n

Unit cell dimensions

a = 12.7329(6) A
b = 19.0589(2) A
c = 13.7260(6) A

01 = 90°
13 = 95.459(2)° “
y = 90°
Volume 3315.8(2) A3
Z 4
Density (calculated) 1.183 g/cm3
Absorption coefficient 0.086 mm'l
F(000) 1288
Crystal size 0.49 x 0.18 x 0.13 mm
0 range for data collection 1.83 to 28.25°
Index ranges -16ShS16,-2SSkS25,-1851518
Reflections collected / unique 30419 / 7907 [R(int) = 0.0676]
Refinement method Full-matrix least-squares on F
Data / restraints / parameters 7907 / 0 / 541
Goodness-of-fit on F"2 1.027
Final R indices [I>20'(I)] R. = 0.0672, wR2 = 0.1646
R indices (all data) R. = 0.1547, wR2 = 0.1969

Largest diff. peak and hole 0.767 and -0.478 eA'3

134

Syntheses:

N-[2-(6-Aminopyridyl)]-acetamidine cyanoborohydride (44). 2.61 g (0.014
mol) of N—[2-(6-Aminopyridyl)]-acetamidine hydrochloride and 0.88 g (0.014 mol) of
NaNCBH3 were dissolved in 65 mL methanol and the resulting clear solution was left
standing at rt for 12 h. After the removal of methanol in vacuo, 20 mL of THF were
added and the solution was filtered. The filtrate was concentrated in vacuo and 100 mL of
CH2C12 were added. The resultant precipitate was collected to yield 2 g (75%) of product
as a white solid. X-Ray quality crystals were grown by diffusion of cyclohexane vapors
into an isopropanol solution of 44. Mp = 119-120 °C; IR (KBr): v3” = 2338, 2219 cm", )
VCN = 2168 cm"; ‘H NMR (DMSO-d6): 5 7.47 (t, J = 8 Hz, 1H, py), 6.52 (s, 2H, py), 6.26
(t, J = 8 Hz, 2H. py), 2.31 (s, 3H, CH3), 0.21 (q, 1.... = 88.5 Hz, 3H, NCBH3); ”B NMR
(DMSO-d6): 5 - 55.95 (q, JBH = 88.8 Hz, NCBH3); l3C NMR (DMSO-d6): 8 164.0, 158.1,
150.6, 140.1, 104.4, 100.1, 20.0. Anal. Calcd for C3H14N5B: C, 50.26; H, 7.33; N, 36.65.
Found: C, 49.64; H, 7.57; N, 36.11.

NaBH4-TEA (45). A mixture of 0.38 g (0.01 mol) NaBH4 and 2.8 mL (0.02 mol)
triethanolamine in 25 mL THF was stirred at rt for 2 h. Filtration of the precipitate
obtained and washing with THF yielded 0.95 g (51%) of 45. Crystals suitable for X-Ray
crystallography were obtained by slow cooling of a solution of 45 in isobutanol from rt to
-25 °C. Mp = 107-108 °C (dec.); IR (KBr): VBH = 2288 cm"; 1H NMR (pyridine-d5): 5
6.68 (br, 3H, OH), 3.87 (t, J = 5.1 Hz, 6H, OCH2), 2.67 (t, J = 5.1 Hz, 6H, NCH2), 1.30
(q, 1511 = 81.3 Hz, 4H, BH4); llB NMR (DMSO-d6): 5 —49.28 (quintet, 1311 = 81.4 Hz,

BH4); ”B MAS NMR (6000 Hz): 5 - 47.9 (Aw/2 = 1826 Hz); ”C NMR (DMSO-d6): 5

135

58.9 (OCH2), 56.9 (NCH2); 23Na MAS NMR (6029 Hz): 5 -9.9 (40.,2 = 1649 Hz). Anal.
Calcd for C6H19NO3BNa: C, 38.50; H, 10.16; N, 7.49. Found: C, 38.09; H, 10.47; N,
7.39; H' content: 4.11.

NaNCBH3-TEA (46). A mixture of 0.63 g (0.01 mol) of NaCNBH; and 1.4 mL
(0.0] mol) of triethanolamine in 30 mL isopropanol was stirred at rt for 2 h. The
undissolved material was filtered out and the isopropanol was slowly evaporated from the
filtrate using a gentle stream of nitrogen. When it was almost dry, 50 mL of CH2Cl2 were
added and the white solid was collected. Yield 1.1 g (52%). X-Ray quality crystals were
grown by slow evaporation of a solution of 46 in isopropanol. Mp = 85-86 °C; IR (KBr):
w... = 2341 cm“, vCN = 2180 cm"; 'H NMR (CH3CN-d3): 5 3.55 (t, J = 5.4 Hz, 6H,
OCH2), 3.46 (s, 3H, OH), 2.52 (t, J = 5.4 Hz, 6H, NCH2), 0.26 (q, 13.. = 88.6, 3H,
NCBH3); “B NMR (DMSO-d6): 5 -559 (q, 1.... = 88.6, NCBH3); 13C NMR (DMSO-d6):
5 59.0 (OCH2), 56.9 (NCH2). Anal. Calcd for C7H18N2O3BNa: C, 39.62; H, 8.49; N,
13.21. Found: C, 38.73; H, 9.23; N, 13.21.

LiBH4-TEA (47). 0.22 g (10 mmol) LiBH4 and 1.4 mL (10 mmol) TEA were
dissolved in 20 mL 2-propanol. Slow evaporation of most of the solvent under Ar
followed by addition of 20 mL CH2C12 afforded 0.95 g (56 %) complex. X-Ray quality
crystals were grown by diffusion of a diethyl ether layer into a solution of 47 in 2-
propanol. 1H NMR (CH3CN-d3): 5 = 4.21 (s, 3H; OH), 3.60 (t, J = 5.1 Hz, 6H; OCH2),
2.58 (t, J = 5.1 Hz, 6H; NCH2), -O.23 (q, JBH = 81.0 Hz, 4H; BH4); 13C NMR (CH3CN-
d3): 5 53.5 (NCH2), 58.4 (OCH2); ”B NMR (CH3CN-d3): 5 -55.98 (quintet, 1.... = 80.9

Hz; B114); "8 MAS NMR (5997 Hz): 5 -48.7 (Aw/2 = 1586 Hz); 7L1 MAS NMR (6000

136

Hz): 5 2.8 (411.,2 = 1735 Hz); IR (KBr): w... = 2290 cm"; Anal. Calcd for C6H1903NBLi:
C, 42.15; H, 11.11; N, 8.19. Found: C, 42.39; H, 11.56; N, 8.17.

LiBH4-TEAod3. A mixture of 0.22 g (10 mmol) of LiBHa and 1.4 mL ( 10.3
mmol),of TEA-d3 in 30 mL THF was stirred under Ar at rt for 1 h. The resulting
precipitate was filtered and washed with THF. Yield quantitative. 1H NMR (CH3CN-d3):
no OH signal; IR (KBr): V01) = 2516 cm"; Anal. hydridic content: 4.06.

LiBDa-T EA (47a). A mixture of 0.065 g (2.5 mmol) of LiBDa and 0.4 g (2.7
mmol) TEA in 8 mL THF was stirred under Ar at rt for 2h. The resulting precipitate was
filtered and washed with THF. IR (KBr): V31) = 1704, 1652 cm"; Anal. hydridic content:
3.90, 90% D based on 'H NMR (DMSO-d6).

LiBI-Ia (50% D)-TEA (47b). A mixture of 0.044 g (2 mmol) of LiBHa and 0.052
g (2 mmol) of LiBDa was dissolved in 12 mL THF. An amount of 0.64 g (4.3 mmol) of
TEA was subsequently added and the mixture was stirred under Ar at rt for 2h. The
resulting precipitate was collected and washed with THF. IR (KBr): v31) = 1752, 1703,
1649 cm", w... = 2384, 2292, 2226 cm"; Anal. hydridic content: 3.80, 54% D based on
‘H NMR (DMSO-d6).

NaBHa-THEC (49). 0.038 g (1 mmol) of NaBHa were dissolved in 10 mL 2-
propanol and an amount of 0.366 g (1 mmol) of THEC-H2O was subsequently added.
After approximately 1 min a white precipitate was formed. The mixture was stirred for 1
h at rt under an Ar atmosphere. Filtration of the mixture yielded 0.3 g (78%) of 49. Single
crystals suitable for X-ray crystallography were grown by diffusion of Et2O in a solution
of 49 in CH3CN. Mp = 180-181 °C (dec.); IR (KBr): var. = 2298, 2225 cm"; 1H NMR

(pyridine-d5): 5 6.03 (br s, 4H, OH), 3.85 (t, J = 4.5 Hz, 8H, OCH2), 2.49-2.41 (br, 24H,

137

NCH2), 1.51 (q, 13.. = 81.3 Hz, 4H, B114); 13C NMR (CH3CN-d3): 5 51.17 (NCH2), 56.45
(NCH2), 58.96 (OCH2); ”B NMR (CH3CN-d3): 5 -54.21 (quintet, 1.3.. = 81.4 Hz, B114);
“13 MAS NMR (4013 Hz): 5 - 45.4 (Aw/2: 2198 Hz); 23Na MAS NMR (6028 Hz):
5 -12.3 (Av1/2 = 1440 Hz). Anal. Calcd for C16H40N404BNa: C, 49.74; H, 10.36; N, 14.51.
Found: C, 49.37; H, 10.87; N, 14.80.

NaNCBH3-THEC (50). A solution of 0.032 g (0.5 mmol) of NaNCBH3 in 3 mL
THF was added to 0.183 g (0.5 mmol) THEC-H2O in THF. A white precipitate was
formed immediately. The mixture was stired 1 h at rt under an Ar atmosphere. The solid
was subsequently filtered and washed with THF to yield 0.17 g (83%) of 50. Single
crystals suitable for X-ray crystallography were grown by diffusion of Et2O in a solution
of 50in CH3CN. Mp = 230-234 °C (dec.); IR (KBr): v3" = 2348 cm'l, VCN = 2162 cm";
lH NMR (CH3CN-d3): 5 3.61 (t, J = 4.8 Hz, 8H, OCH2), 3.25 (br s, 4H, OH), 2.47 (br,
24H, NCH2), 0.25 (q, 3H, NCBH3); ”C NMR (CH3CN-d3): 5 51.17 (NCH2), 56.50
(NCH2), 59.06 (OCH2); llB NMR (CH3CN-d3): 5 -56.60 (q, 13.. = 88.3 Hz, NCBH3).

NaBHa-THPC (51). 0.285 g (0.7 mmol) of THPC were added to a solution of
0.027 g (0.7 mmol) in 8 ml 2-propanol. In approximately 2 min a white precipitate was
formed. The mixture was left standing for l h at rt, then the solid was filtered. Yield
0.085 g (27%). No melting occurs when the solid is heated up to 200 °C, at which
temperature it turns brown, and eventually starts to melt at approximately 210 °C. Single
crystals suitable for X-ray crystallography were grown by slow evaporation from 2-
propanol. IR (KBr): VB" = 2290 cm'l; 1H NMR (H20-d2): 5 3.89 (br m, 4H, OCH), 2.90-
2.87 (br (1, 8H, NCH2), 2.42-2.34 (br t, 4H, NCH2), 2.09 (br s, 12 H, NCH2) -0.27 (q, 1311

= 80.72 Hz, 4H, BH4); 13C NMR (DMSO-d5): 5 21. 79 (CH3), 49.17 (br, NCH2), 51.03

138

 

(br, NCH2), 61.58 (NCH2), 62.31 (OCH2); “8 NMR (DMSO-d6): 5 -52.55 (quintet, 1.3..
= 82.8 Hz, BH4); Anal. Calcd for C20H43N4OaBNa: C, 54.30; H, 10.86; N, 12.67. Found:
C, 54.66; H, 11.39; N, 12.41.

9-(1’,4’,7’-tris-hydroxyethyl-cyclen)methylanthracene (52). Warning: Ethylene
oxide is a volatile (Bp = 11 °C), extremely toxic, carcinogenic substance. 0.2 g (0.57
mmol) of 9-(1’,4’,7’,10’~tetraazacyclododecyl)methylanthracene were dissolved in 3 mL
absolute ethanol. 2 mL of water were added, and the solution was cooled to 0 °C. To this
solution, 0.158 g (3.6 mmol) of ethylene oxide (cooled to 0 °C) were added, and the
mixture was stirred at 0 °C for 2 h, and then another hour at rt. Solvent evaporation under
vacuum at 35 °C yielded a yellow oil. A few mL of water were subsequently added, and
the resulting aqueous solution was extracted with chloroform. Evaporation of CHC13
yielded a yellow oil, which was dissolved in minimal amount of 2-propanol, and a large
excess of ether was added. Upon standing in the freezer over night, a yellow crystalline
solid precipitated, which was filtered and washed with ether, to yield 0.12 g (42%) of 52.
Single crystals suitable for X-ray crystallography were grown by diffusion of Et2O in a
solution of 52 in 2-propanol. Mp = 131-133 °C; 1H NMR (CH3Cl—d3): 5 8.36 (d, 2H,
ArH), 8.29 (s, 1H, ArH), 7.87 (d, 2H, ArH), 7.38 (m, 4H, ArH), 4.8 (br, OH), 4.39 (s, 2H,
CH2-Ar), 3.47 (t, J = 4.5 Hz, 2H, OCH2), 3.40 (t, J = 4.5 Hz, 4H, OCH2), 2.73 (t, J = 4.5
Hz, 6H, NCH2), 2.43-2.32 (br, 16H, NCH2); 13C NMR (CH3C1-d3): 5 130.83, 128.51,
127.11, 125.32, 124.50, 58.35, 58.01, 57.73, 55.65, 52.37, 52.11, 51.90, 50.88, 50.56;
Anal. Calcd for C29H42N403: C, 70.45; H, 8.50; N, 11.34. Found: C, 70.23; H, 8.83; N,

11.04.

139

NaBHa-[9-(1’,4’,7’-tris-hydroxyethyl-cyc1en)methylanthracene]~iPrOH (53).
0.1 g (0.2 mmol) of 9-(1’,4’,7’-tris-hydroxyethyl-cyc1en)methylanthracene were added to
a solution of 0.008 g (0.21 mmol) of N aBH4 in 4 mL 2-propanol. In approximately 4 min
a white precipitate was formed. The mixture was stirred 2 h at rt under an Ar atmosphere,
and subsequently filtered to yield 0.09 g (53%) of 53. The solid turns red when heated,
and partly melts at 190 °C. Single crystals suitable for X-ray crystallography were grown
by diffusion of Et2O in a solution of 53 in CH3CN. IR (KBr): V311 = 2384, 2293, 2228
cm"; 'H NMR (CH3CN-d3): 5 8.54 (s,1H, ArH), 8.43 (d, 2H, ArH), 8.06 (d, 2H, ArH),
7.60 (m, 2H, ArH), 7.49 (m, 2H, ArH), 4.61 (s, 2H, CH2-Ar), 3.72 (septet, 1H, iPrOH),
3.6-1.9 (br), 1.08 (d, 6H, iPrOH); 13C NMR (CH3CN-d3): 5 132.47, 130.23, 128.98,
127.77, 126.22, 125.46, 58.54, 57.98, 56.54, 56.15, 54.51, 52.65, 51.59, 51.26, 50.74; HB
NMR (CH3CN-d3): 5 -53.75 (quintet, JBH = 80.4 Hz, BH4); Anal. Calcd for

C32H54N4O4NaB: C, 64.87; H, 9.12; N, 9.46. Found: C, 64.20; H, 9.06; N, 9.33.

Decomposition Studies:

Solid-State Decomposition of NaBH4-TEA. 05-07 g of N aB114~TEA were
heated at 82 °C under Ar or open atmosphere in a 7 x 27 mm glass tube immersed in
refluxing acetonitrile, for 30-58 days. The resulting white solid is insoluble in common
organic solvents and it does not melt up to 300 °C. The powder XRD shows a single
phase with a broad peak at 20 = 94° but no NaBHa or initial complex. IR (KBr): VBH =

2387, 2293, 2226 cm"; ”8 MAS NMR (6058 Hz): 5 -7.1 (Aw/2 = 1390 Hz); 23Na MAS

140

 

 

NMR (6004 Hz): 5 -14.6 (Aw/2 = 2769 Hz). Anal. NIB/Na = 1/ 1/0.96; H' content: 1.05-
1.38.

Decomposition of NaBHa-TEA in DMSO. A solution of 0.5 g of NaBH4-TEA in
10 mL DMSO was heated at 110 °C under Ar for 55 h. The resulting precipitate was
collected and washed with CH2Cl2 to yield 0.17 g of white solid insoluble in common
organic solvents. Mp > 300 °C; IR (KBr): no B-H stretching vibration; llB MAS NMR
(6000 Hz): 5 -5.0 (Av.,.= 1559 Hz); ”Na MAS NMR (6011 Hz): 5 .135 (Av... = 2665
Hz). Anal. H' content: 0.22.

Decomposition of NaBH4-TEA in the Melt. 0.5 g of NaBH4-TEA were placed in
a 21 x 50 mm vial and heated under Ar at 130 °C in an oil bath for two hours. In about 1
min the solid melted with effervescence. After approximately another 10 min the melt
solidified. The powder XRD confirmed the presence of NaBH4. llB MAS NMR (5998
Hz): 5 -2.3 (Aw/2: 1394 Hz, 1 B), 5 -50.4 (Aw/2: 1017 Hz 1.2 B). Anal. H' content:
2.03.

Solid-State Decomposition of LiBIL-TEA. Aliquots of 0.1-0.5 g of LiBHa-TEA
were placed in a 21 x 50 mm vial and were preheated under Ar at 80-100 °C for 10-15
min. The temperature was subsequently raised to 120-130 °C, and heating was continued
for about 2 h. The resulting white solid is insoluble in common organic solvents and it
does not melt up to 300 °C. IR (KBr): v8.. = 2301, 2247, 2171 cm"; “13 MAS NMR
(6000 Hz): 5 -4.6 (Av.,. = 1693 Hz); 7L1 MAS NMR (6000 Hz): 5 3.1 (Av.,. = 1611 Hz).

Anal. H’ content: 0.97-1.08.

141

Annealing of the LiBH4-TEA Solid-State Decomposition Product. A
suspension of 0.07 g of the solid-state decomposition product 48 (H' = 1.05) in 5 ml. of
freshly distilled DMSO was heated at 120 °C under an Ar atmosphere for 4 days, then
filtered and the resulting solid was washed with CH2Cl2. IR (KBr): v3" = 2368, 2340
cm"; ”13 MAS NMR (6000 Hz): 5 -33 (Av.,. = 1677 Hz); 71.1 MAS NMR (3540 Hz): 5
1.0 (Aw/2 = 2496 Hz). Anal. H' content: 1.05.

Decomposition of LiBHa-TEA in DMSO. A solution of 0.5 g of LiBH4-TEA in
10 mL DMSO was heated at 120 °C under Ar for 4 days. The resulting precipitate was
collected and washed with CH2Cl2. Yield 0.2 g. Mp > 300 °C; IR (KBr): no EH; "8
MAS NMR (6000 Hz): 5 -5.9 (Aw/2 = 1726 Hz). Anal. H' content: 0.26.

HID Isotope Exchange Experiments. Aliquots of 0.1-0.3 g of LiBHa-TEA-d3
were heated under Ar at 85-110 °C, monitoring the decompositions by H' analysis. The
partly or fully decomposed samples were analyzed by IR (KBr) and 1H NMR (CH3CN-
d3). No signals for the V131) in the IR or for OH in the 1H N MR spectra were observed
throughout the reaction.

Reaction of the LiBIL-TEA Solid-State Decomposition Product (48) with
Pd(AcO)2. A sample of 0.1 g (z 0.61 miliequivalents of H’) of the solid-state
decomposition product 48 was suspended in 10 mL freshly distilled THF containing
0.136 g (0.61 mmol) Pd(AcO)2. The mixture was stirred at rt under an Ar atmosphere for
24 h. After filtering and washing with THF, a green hygroscopic powder was obtained,
which turned into a black paste if left in open atmosphere. Powder XRD analysis Showed
that this compound was completely amorphous, and no Pd black, or unconverted

Pd(AcO)2 was present. IR (KBr): vAco = 1580 cm"; no B-H; ‘H NMR (DMSO-d6): very

142

complex, and indicates the presence of at least three different AcO groups: 5 1.72, 1.69,
1.63. '3C NMR (DMSO-d6): 5 22.94, 24.02, 25.17, 59.20, 61.64, 64.43, 66.26, 67.08; “B
NMR (DMSO-d6): 5 -14.61.

Solid-State Decomposition of NaBHa-THEC. A sample of 0.07 g of
NaBI-I4-THEC was heated at 160 °C under an Ar atmosphere for 20 h. The resulting
white solid is insoluble in common organic solvents and it does not melt up to 300 °C. IR
(KBr): no B-H; “13 MAS NMR (4038 Hz): 5 80 (Av.,. = 2462 Hz); ”Na MAS NMR
(6001 Hz): 5 -15.6 (Aw/2: 3421 Hz).

Decomposition of NaBHa-THEC in the Melt. A sample of 0.04 g of
NaBHa-THEC was rapidly heated to 200 °C under Ar. Melting occurred under these
conditions, followed by immediate solidification. Heating was continued for 10 min. The
IR spectrum of the resulting solid is virtually identical with the corresponding spectrum
of the solid-state decomposition product. Powder XRD indicates an amorphous phase,
and annealing at 160 °C for 24 h did not improve its crystallinity. llB MAS NMR (6000
Hz): 5 -6.9 (Aw/2 = 1745 Hz); ”Na MAS NMR (6000 Hz): 5 -8.9 (Av.,. = 2704 Hz).

Decomposition of NaBHa-THEC in DMSO. A solution of 0.15 g of
N aBHa-THEC in 5 mL DMSO was heated at 120 °C under Ar for 48 h. A crystalline
solid precipitated, which was filtered and washed with CH2Cl2. Yield 0.13 g. “B MAs
NMR (6060 Hz): 5 -7.5 (Aw/2 = 1880 Hz); 23Na MAS NMR (6058 Hz): 5 -9.0 (Aw/2 =

2476 Hz).

143

BIBLIOGRAPHY

144

BIBLYOGRAPHY

 

(1) Jeffrey, G. A. An Introduction to Hydrogen Bonding, Oxford University Press,
Oxford, 1997.

(2) Jeffrey, G. A.; Saenger, W. Hydrogen Bonding in Biological Structures, Springer-
Verlag, Berlin, 1991.

(3) (a) Crabtree, R. H.; Siegbahn, P. E. M.; Eisenstein, O.; Rheingold, A. L.; Koetzle, T.
F. Acc. Chem. Res. 1996, 29, 348. (b) Crabtree, R. H. J. Organomet. Chem. 1998, 577,
111. (c) Shubina, E. S.; Belkova, N. V.; Epstein, L. M. J. Organomet. Chem. 1997, 536-
537, 17. (d) Alkorta, 1.; Rozas, I.; Elguero, J. Chem. Soc. Rev. 1998, 27, 163. (e)
Crabtree, R. H. Science 1998, 282, 2000. (f) Crabtree, R. H.; Eisenstein, O.; Sini, G.;
Peris, E. J. Organomet. Chem. 1998, 567, 7. (g) Kelly, R; Loza, M. Chemistry in Britain
1999, 35, 26. (h) Calhorda, M. J. Chem. Commun. 2000, 801

(4) Zachariasen, W. H.; Mooney, R. C. L. J. Chem. Phys. 1934, 2, 34.

(5) We have recently redetermined the crystal structure of ammonium hypophosphite at
higher resolution, which allowed us to locate the H atoms and thus probe the existence of
dihydrogen bonding. However, we found no close H-r-H contacts in this salt. Stein, R.;
Custelcean, R.; Jackson, J. E. Manuscript in preparation.

(6) Burg, A. B. Inorg. Chem. 1964, 3, 1325.

(7) Titov, L. V.; Makarova, M. D.; Rosolovskii, V. Y. Doklady Akademii Nauk 1968,

180, 381.

145

 

 

(8) (a) Brown, M. P.; Heseltine, R. W. Chem. Commun. 1968, 1551. (b) Brown, M. P.;
Heseltine, R. W.; Smith, P. A.; Walker, P. J. J. Chem. Soc. A 1970, 410.

(9) Brown, M. P.; Walker, P. J. Spectrochim. Acta 1974, 30A, 1125.

(10) (a) Shubina E. S.; Bachmutova, E. V.; Saitkulova, L. N.; Epstein, L. M. Mendeleev
Commun. 1997, 83. (b) Epstein, L. M.; Shubina, E. S.; Bakhmutova, E. V.; Saitkulova, L.
N.; Bakhmutov, V. 1.; Chistyakov, A. L.; Stankevich, I. V. Inorg. Chem. 1998, 37, 3013.
(c) Shubina, E. S.; Belkova, N. V.; Bakhmutova, E. V.; Saitkulova, L. N.; Ionidis, A. V.;
Epstein, L. M. Russ. Chem. Bull. 1998, 4 7, 817.

(11) Jackson, J. E.; Huang, R.; Eckert, J.; Sharma, M.; Argyriou, D.; Sheldon, R.
Unpublished results.

(12) Peris, E.; Lee, J. C.; Rambo, J. R.; Eisenstein, O.; Crabtree, R. H. J. Am. Chem. Soc.
1995, 117, 3485.

(13) Cramer, C. J.; Gladfelter, W. L. Inorg. Chem. 1997, 36, 5358.

(14) Klooster, W. T.; Koetzle, T. F.; Siegbahn, Per E. M.; Richardson, T. B.; Crabtree, R.
H. J. Am. Chem. Soc. 1999, 121, 6337.

(15) Popelier, P. L. A. J. Phys. Chem. A 1998, 102, 1873.

(16) N5th, H.; Thomas, S. Eur. J. Inorg. Chem. 1999, 1373.

(17) (a) Zottola, M. A.; Pedersen, L. G.; Singh, P.; Shaw, B. R. Modeling the Hydrogen
Bond; ACS Symposium Series, Smith, D. A., Ed.; American Chemical Society,
Washington, DC, 1994, Vol. 569, pp. 277-285 (b) Singh, P.; Zottola, M.; Huang, 8.;
Shaw, B. R.; Pedersen L. G. Acta Cryst. 1996, C52, 693.

(18) Padilla-Martinez, I. R.; Rosalez-Hoz, M. de J.; Tlahuext, H.; Camacho-Camacho, C.;

Ariza-Castolo, A.; Contreras, R. Chem. Ber. 1996, 129, 441.

146

 

 

(19) Flores-Pana, A.; Sénchez—Ruiz, S. A.; Guadarrarna, C.; N5th, H.; Contreras, R. Eur.
J. Inorg. Chem. 1999, 2069.

(20) Atwood, J. L.; Koutsantonis, G. A.; Lee, F.-C.; Raston, C. L. Chem. Commun. 1994,
91.

(21) Campbell, J. P.; Hwang, J. W.; Young, V. G.; Von Dreele, R. B.; Cramer, C. J.;
Gladfelter, W. L. J. Am. Chem. Soc. 1998, 120, 521.

(22) (a) Liu, Q.; Hoffmann, R. J. Am. Chem. Soc. 1995, 117, 10108. (b) Alkorta, 1.;
Elguero, J.; Foces-Foces, C. Chem. Commun. 1996, 1633. (c) Remko, M. Molecular
Physics 1998, 94, 839. (d) Kulkami, S. A. J. Phys. Chem. A 1998, 102, 7704. (e)
Kulkami, S. A.; Srivastava, A. K. J. Phys. Chem. A 1999, 103, 2836. (f) Petterson, M.;
Lundell, J. Phys. Chem. Chem. Phys. 1999, I, 1601.

(23) Grabowski, S. J. Chem. Phys. Lett. 1999, 312 (5-6), 542.

(24) Milstein, D.; Calabrese, J. C.; Williams, I. D. J. Am. Chem. Soc. 1986, 108, 6387.
(25) Stevens, R. C.; Bau, R.; Milstein, D.; Blum, O.; Koetzle, T. F. J. Chem. Soc. Dalton
77ans.1990,l429.

(26) Feracin, S.; Biirgi, T.; Bakhmutov, V. I.; Eremenko, 1.; Vorontsov, E. V.; Vimenits,
A. B.; Berke, H. Organometallics 1994, I3, 4194.

(27) Lee, J. C.; Rheingold, A. L.; Muller, B.; Pregosin, P. S.; Crabtree, R. H. Chem.
Cbnunun.l994,1021.

(28) Peris, E.; Lee, J. C.; Rambo, J. R., Eisenstein, O.; Crabtree, R. H. J. Am. Chem. Soc.
1995,117,3485.

(29) Lough, A. J .; Park, S.; Ramachandran, R.; Morris, R. H. J. Am. Chem. Soc. 1994,

116, 8356.

147

 

 

(30) Park, S.; Ramachandran, R.; Lough, A. J.; Morris, R. H. Chem. Commun. 1994,
2201.

(31) Park, S.; Lough, A. J.; Morris, R. H. Inorg. Chem. 1996, 35, 3001.

(32) (a) Wessel, J.; Lee, J. C.; Peris, E.; Yap, G. P. A.; Fortin, J. B.; Ricci, J. S.; Sini, G.;
Albinati, A.; Koetzle, T. F.; Eisenstein, O.; Rheingold, A. L.; Crabtree, R. H. Angew.
Chem. Int. Ed. 1995, 34, 2507. (b) Patel, B. P.; Wessel, J .; Yao, W.; Lee Jr., J. C.; Peris,
E.; Koetzle, T. F.; Yap, G. P. A.; Fortin, J. B.; Ricci, J. S.; Sini, G.; Albinati, A.;
Eisenstein, O.; Rheingold, A. L.; Crabtree, R. H. New J. Chem. 1997, 21 , 413. (c) Sini,
G.; Eisenstein, O.; Yao, W.; Crabtree, R. H. Inorg. Chim. Acta 1998, 280, 26.

(33) Patel, B. P.; Yao, W.; Yap, G. P.A.; Rheingold, A. L.; Crabtree, R. H. Chem.
Commun. 1996, 991.

(34) Peris, E.; Wessel, J.; Patel, B. P.; Crabtree, R. H. Chem. Commun. 1995, 2175.
(35) Desmurs, P.; Kavallieratos, K.; Yao, W.; Crabtree, R. H. New J. Chem. 1999, 23,
1111.

(36) Shubina, E. S.; Belkova, N. V.; Krylov A. N .; Vorontsov, E. V.; Epstein, L. M.;
Gusev, D.G.; Niederrnann, M.; Berke, H. J. Am. Chem. Soc. 1996, 118, 1105.

(37) Belkova, N. V.; Shubina, E. S.; Ionidis, A. V.; Epstein, L. M.; Jacobsen, H.;
Messmer, A.; Berke, H. Inorg. Chem. 1997, 36, 1522.

(38) Messmer, A.; Jacobsen, H.; Berke, H. Chem. Eur. J. 1999, 5, 3341.

(39) Scheiner, S.; Orlova, G. J. Phys. Chem. A 1998, 102, 260.

(40) (a) Richardson, T. B.; Koetzle, T. F.; Crabtree, R. H. Inorg. Chim. Acta 1996, 250,

69. (b) Xu, W.; Lough, A. J.; Morris, R. H. Can. J. Chem. 1997, 75, 475. (c) Huang, L.-

148

 

Y.; Aulwurrn, U. R.; Heinemann, F. W.; Knoch, F.; Kisch, H. Chem. Eur. J. 1998, 4,
1641. (d) Junk, P. C.; Steed, J. W. J. Organomet. Chem. 1999, 587, 191.

(41) Braga, D.; De Leonardis, P.; Grepioni, F.; Tedesco, E.; Calhorda, M. J. Inorg. Chem.
1998, 37, 3337.

(42) Abramov, Y. A.; Brammer, L.; Klooster, W. T.; Bullock, R. M. Inorg. Chem. 1998,
37, 6317.

(43) Gusev, D. G.; Lough, A. J .; Mom's, R. H. J. Am. Chem. Soc. 1998, 120, 13138.
(44) Abdur-Rashid, K.; Gusev, D. G.; Landau, S. E.; Lough, A. J.; Morris, R. H. J. Am.
Chem. Soc. 1998, 120, 11826.

(45) Abdur-Rashid, K.; Gusev, D. G.; Lough, A. J .; Morris, R. H. Organometallics 2000,
19, 834.

(46) Jackson, J. E.; Bass, S.; Huang, R. Unpublished results.

(47) Lee, J. C.; Peris, E.; Rheingold, A. L.; Crabtree, R. H. J. Am. Chem. Soc. 1994, 116,
11014.

(48) Ayllon, J. A.; Gervaux, C.; Sabo-Etienne, S.; Chaudret, B. Organometallics 1997,
16, 2000.

(49) Orlova, G.; Scheiner, S.; Kar, T. J. Phys. Chem. A 1999, 103, 514.

(50) Shubina, E. S.; Belkova, N. V.; Bakhmutova, E. V.; Vorontsov, E. V.; Bakhmutov,
V. 1.; Ionidis, A. V.; Bianchini, C.; Marvelli, L.; Peruzzini, M.; Epstein, L. M. Inorg.
Chim. Acta 1998, 280, 302.

(51) Epstein, L. M.; Shubina, E. S. Ber. Bunsenges. Phys. Chem. 1998, 102, 359.

(52) Orlova, G.; Scheiner, S. J. Phys. Chem. A 1998, 102, 4813.

(53) Chu, H. S.; Xu, 2.; Ng, S. M.; Lau, C. P.; Lin, Z. Eur. J. Inorg. Chem. 2000, 993.

149

 

 

(54) Ayllon, J. A.; Sabo-Etienne, S.; Chaudret, B.; Ulrich, S.; Limbach, H.-H. Inorg.
Chim. Acta 1997, 259, 1.

(55) Grfindemann, S.; Ulrich, S.; Limbach, H.-H.; Golubev, N. S.; Denisov, G. S.;
Epstein, L. M.; Sabo-Etienne, S.; Chaudret, B. Inorg. Chem. 1999, 38, 2550.

(56) Chu, H. S.; Lau, C. P.; Wong, K. Y.; Wong, W. T. Organometallics 1998, 17, 2768.
(57) Musashi, Y.; Sakaki, S. J. Am. Chem. Soc. 2000, 122, 3867.

(58) Ayllon, J. A.; Sayers, S. F.; Sabo-Etienne, S.; Donnadieu, B.; Chaudret, B.
Organometallics 1999, 18, 3981.

(59) Caballero, A.; Jalén, F. A.; Manzano, B. R. Chem. Commun. 1998, 1879.

(60) Patel, B. P.; Kavallieratos, K.; Crabtree, R. H. J. Organomet. Chem. 1997, 528, 205.
(61) Bosque, R.; Maseras, F.; Eisenstein, O.; Patel, B. P.; Yao, W.; Crabtree, R. H. Inorg.
Chem. 1997, 36, 5505.

(62) Yao, W.; Crabtree, R. H. Inorg. Chem. 1996, 35, 3007.

(63) Aime, S.; Gobetto, R.; Valls, E. Organometallics 1997, 16, 5140.

(64) Aime, S.; Fém'z, M.; Gobetto, R.; Valls, E. Organometallics 1999, 18, 2030.

(65) Aime, S.; Férriz, M.; Gobetto, R.; Valls, E. Organometallics 2000, 19, 707.

(66) Haack, K.-J.; Hashiguchi, S.; Fujii, A.; Ikariya, T.; Noyori, R. Angew. Chem. Int. Ed.
1997, 36, 285.

(67) Gatling, S. C.; Jackson, J. E. J. Am. Chem. Soc. 1999, 121, 8655.

(68) Hwang, J .-W.; Campbell, J. P.; Kozubowski, J.; Hanson, S. A.; Evans, J. F.;
Gladfelter, W. L. Chem. Mater. 1995, 7, 517.

(69) W5hler, F. Pogg. Ann. 1828, 12, 253.

150

 

 

(70) (a) v Liebig, J .; W5hler, F. Ann. Phys. Leipzig, Ser. 2 1830, 20, 369; (b) Dunitz, J.
D.; Harris, K. D. M.; Johnston, R. L.; Kariuki, B. M.; MacLean, E. J .; Psallidas, K.;
Schweizer, W. B.; Tykwinski, R. R J. Am. Chem. Soc. 1998, 120, 13274.

(71) (a) Ohashi, Y. Reactivity in Molecular Crystals, Kodansha, Tokyo, 1993. (b) Singh,
N. B.; Singh, R. J .; Singh, N. P. Tetrahedron 1994, 50, 6441.

(72) Schmidt, G. M. J. Pure Appl. Chem. 1971, 27, 647.

(73) Cohen, M. D. Angew. Chem. Int. Ed. 1975, 14, 386.

(74) Gavezzotti, A. J. Am. Chem. Soc. 1983, 105, 5220.

(75) McBride, J. M.; Segmuller, B. E.; Hollingsworth, M. D.; Mills, D. E.; Weber, B. A.
Science 1986, 234, 830.

(76) Luty, T.; Eckhartd, C. J. J. Am. Chem. Soc. 1995, 117, 2441.

(77) Curtin, D. Y.; Paul, C. I.; Duesler, E. N.; Lewis, T. W.; Mann, B. J .; Shiau, W. 1.
Mol. Cryst. Liq. Cryst. 1979, 50, 25.

(78) Gougoutas, J. Z. Pure Appl. Chem. 1971, 27, 305.

(79) Wegner, G. Pure Appl. Chem. 1977, 49, 443.

(80) Adler, G. Mol. Cryst. Liq. Cryst. 1988, 156, l.

(81) Hirshfeld, F. L.; Schmidt, G. M. J. J. Polymer Sci. A 1964 , 2, 2181.

(82) Hasegawa, M. Chem. Rev. 1983, 83, 508.

(83) Addadi, L.; Lahav, M. J. Am. Chem. Soc. 1979, 101, 2152.

(84) Epple, M.; Kirschnick, H. Chem. Ber. 1996, 129, 1123.

(85) Epple, M.; Kirschnick, H. Chem. Ber. 1997, I30, 291.

(86) Leiserowitz, L.; Schmidt, G. M. J. J. Chem. Soc. A, 1969, 2372.

(87) Feldman, K. S.; Campbell, R. F. J. Org. Chem. 1995, 60, 1924.

151

 

(88) Feldman, K. S.; Campbell, R. F.; Saunders, J. C.; Ahn, C.; Masters, K. M. J. Org.
Chem. 1997, 62, 8814.

(89) Matsumoto, A.; Odani, T.; Chikada, M.; Sada, K.; Miyata, M. J. Am. Chem. Soc.
1999. 121, 11122.

(90) Kane, J. J .; Liao, R.-F.; Lauher, J. W.; Fowler, F. W. J. Am. Chem. Soc. 1995, 117,
12003.

(91) Xiao, J .; Yang, M.; Lauher, J. W.; Fowler, F. W. Angew. Chem. Int. Ed. 2000, 39,
2132.

(92) Coates, G. W.; Dunn, A. R.; Henling, L. M.; Dougherty, D. A.; Grubbs, R. H.
Angew. Chem. Int. Ed. 1997, 36, 248.

(93) Coates, G. W.; Dunn, A. R.; Henling, L. M.; Ziller, J. W.; Lobkovsky, E. B.;
Grubbs, R. H. J. Am. Chem. Soc. 1998, 120, 3641.

(94) For excellent reviews about kinetics in the solid-state see: (a) Hannay, N. B. Treatise
0n Solid State Chemistry 1976, Reactivity of Solids, Vol 4, Plenum Press, New York. (b)
Brown, M. E.; Dollimore, D.; Galwey, A. K. Bamford, C. H. (Editor) Comprehensive
Chemical Kinetics, Vol. 22: Reactions in the Solid State 1980, Elsevier Scientific
Publishing Company, Amsterdam-Oxford-New York. (c) Garner, W. E. (Editor)
Chemistry of the Solid State 1955, Academic Press Inc., New York.

(95) (a) Sekiguchi, K.; Shirotani, K—I.; Sakata, 0.; Suzuki, E. Chem. Pharm. Bull. 1984,
32, 1558. (b) Brill, T. B.; Gongwer, P. E.; Williams, G. K. J. Phys. Chem. 1994, 98,
12242.

(96) Baughman, R. H. J. Chem. Phys. 1978, 68, 3110.

(97) Galwey, A. K.; Brown, M. E. Proc. R. Soc. London A 1995, 450, 501.

152

 

 

(98) (a) Vyazovkin, S.; Wight, C. A. Annu. Rev. Phys. Chem. 1997, 48, 125. (b)
Vyazovkin, S.; Wight, C. A. J. Phys. Chem. A 1997, 101, 8279.

(99) (a) Davis, R. E. J. Am. Chem. Soc. 1962, 84, 892. (b) Pepperberg, I. M.; Halgren, T.
A.; Lipscomb, W. N. J. Am. Chem. Soc. 1976, 98, 3442.

(100) Mesmer, R. E.; Jolly, W. L. Inorg. Chem. 1962, 1, 608.

(101) (a) Schreiner, P. R.; Schaefer, H. F.; Schleyer, P. v. R. J. Chem. Phys. 1994, 101,

7625. (b) Watts, J. D.; Bartlett, R. J. J. Am. Chem. Soc. 1995, 117, 825. (c) Custelcean, R.

ow‘a

J. Mol. Struct. Theochem 2000, 505, 95.

(102) Rozas, I.; Alkorta, 1.; Elguero, J. Chem. Phys. Lett. 1997, 275, 423.

(103) Custelcean, R.; Jackson, J. E. J. Am. Chem. Soc. 1998, 120, 12935.

(104) Naiini, A. A.; Pinkas, J .; Plass, W.; Young, V. G.; Verkade, J. G. Inorg. Chem.
1994, 33, 2137.

(105) a) Ashby, E. C.; Dobbs, F. R.; Hopkins, H. P.; J. Am. Chem. Soc. 1975, 97, 3158.
b) Kadlec, V.; Hanzlik, J. Collect. Czech. Chem. Commun. 1974, 39, 3200.

(106) Davis, R. E.; Gottbrath, J. A.; J. Am. Chem Soc. 1962, 84, 895.

(107) (8) Davis, R. E.; Bromels, E.; Kibby, C. L. J. Am. Chem. Soc. 1962, 84, 885; (b)
Kreevoy, M. M.; Hutchins, J. E. C. J. Am. Chem. Soc., 1969, 91, 4329.

(108) Custelcean, R.; Jackson, J. E. Angew. Chem. Int. Ed. 1999, 38, 1661.

(109) Brown, H. C.; Mead, E. J .; Scoaf, C. J. J. Am. Chem. Soc. 1956, 78, 3616.

(110) (a) Onak, T. P.; Landesman, H.; Williams, R. E.; Shapiro, I. J. Phys. Chem. 1959,
63, 1533. (b) Golden, J. H.; Schreier, C.; Singaram, B.; Williamson, 8. M. Inorg. Chem.

1992, 31, 1533.

153

 

(111) (a) Price, W. C. J. Chem. Phys. 1949, 17, 1044. (b) Volkov, V. V.; Sobolev, E. V.;
Grankina, Z. A.; Kalinina, I. S. Russ. J. Inorg. Chem. 1968, 13, 343. (c) Marks, T. J.;
Kennelly, W. J .; Kolb, J. R.; Shimp, L. A. Inorg. Chem. 1972, 11, 2540.

(112) Custelcean, R.; Jackson, J. E. J. Am. Chem. Soc. 2000, 122, 5251.

(1 13) (a) Galwey, A. K. Pure Appl. Chem. 1995, 67, 1809. (b) Tanaka, H.; Koga, N. J.
Phys. Chem. 1988, 92, 7023. (c) Koga, N.; Tanaka, H. J. Phys. Chem. 1994, 98, 10521.
(114) For examples of other solid-state reactions where the kinetics are directly coupled
to chemical transformations see: (a) Galwey, A. K.; Mohamed, M. A. J. Chem. Soc.
Faraday Trans. 1985, 81, 2503. (b) Son, S. F.; Asay, B. W.; Henson, B. F.; Sander, R.
K.; Ali, A. N .; Zielinski, P. M.; Phillips, D. S.; Schwarz, R. B.; Skidmore, C. B. J. Phys.
Chem. B 1999, 103, 5434.

(115) Davis, R. E. J. Am. Chem. Soc. 1962, 84, 892.

(116) Custelcean, R. Vlassa, M.; Jackson, J. E. Angew. Chem. in press.

(117) (a) Buoen, S.; Dale, J.; Krane, J. Chem. Commun. 1982, 1172. b) Whitbread, S. L.;
Politis, S.; Stephens, A. K. W.; Lucas, J. B.; Dhilon, R. S.; Lincoln, S. F.; Wainwright, K.
P. J. Chem. Soc. Dalton Trans. 1996, 1379.

(118) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.

(119) Akkaya, E. U.; Huston, M. E.; Czamik, A. W. J. Am. Chem. Soc. 1990, 112, 3590.
(120) Bernstein, J .; Stearns, B.; Shaw, E.; Lott, W. A. J. Am. Chem. Soc. 1947, 69, 1151.
(121) Schlesinger, H. I.; Brown, H. C.; Hyde, E. K. J. Am. Chem. Soc. 1953, 75, 209.

(122) Buoen, S.; Dale, J .; Krane, J. Acta. Chem. Scand. 1984, B 38, 773.

154

 

(123) (a) Hancock, R. D.; Shaikjee, M. S.; Dobson, S. M.; Boeyens, J. C. A. Inorg. Chim.
Acta 1988, 154, 229. (b) Dhillon, R. S.; Madbak, S. E.; Ciccone, F. G.; Buntine, M. A.;
Lincoln, S. F.; Wainwright, K. P. J. Am. Chem. Soc. 1997, 119, 6126.

(124) Eckert, H. Prog. Nucl. Mag. Res. Sp. 1992, 24, 159.

(125) Gaussian 98: Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A.; Stratmann, R. E.;
Burant, J. C.; Dapprich, S.; Millam, A. D.; Daniels, A. D.; Kudin, K. N .; Strain, M. C.;
Farkas, O.; Thomasi, J .; Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.;
Adamo, C.; Clifford, S.; Ochterski, J .; Petersson, G. A.; Ayala, P. Y.; Cui, Q.;
Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.;
Cioslowski, J .; Ortiz, J. V.; Stefanov, B. B.; Lin, G.; Liashenko, A.; Piskorz, P.;
Komarorni, 1.; Gomperts, R.; Martin, R. L.; Fox, D. J .; Keith, T.; Al-Laham, M. A.; Peng,
C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; Johnson, B. G.;
Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon, M.; Replogle, E. S.; Pople, J. A.;
Gaussian, Inc., Pittsburgh PA, 1998.

(126) SHELXTL: Structure Analysis Program 5.1; Bruker AXS, Inc., Madison, WI, 1997.

155

MICHIGAN STATE UNIVERSITY LIBRARI-

lwmmww