Fast carbon dioxide fixation by 2,6-pyridinedicarboxamidato-nickel(II)- hydroxide complexes: Influence of changes in reactive site environment on reaction rates

Article abstract of DOI:10.1021/ic200942u

The planar complexes [Ni^II(pyN₂^R2)(OH)] ^-, containing a terminal hydroxo group, are readily prepared from N,N′-(2,6-C₆H₃R₂)-2,6- pyridinedicarboxamidate(2-) tridentate pincer ligands (R₄N)(OH), and Ni(OTf)₂. These complexes react cleanly and completely with carbon dioxide in DMF solution in a process of CO₂ fixation with formation of the bicarbonate product complexes [Ni^II(pyN₂ ^R2)(HCO₃)]^- having η¹-OCO ₂H ligation. Fixation reactions follow second-order kinetics (rate = k₂′[Ni^II-OH][CO₂]) with negative activation entropies (-17 to -28 eu). Reactions were monitored by growth and decay of metal-to-ligand charge-transfer (MLCT) bands at 350-450 nm. The rate order R = Me > macro > Et > Prⁱ > Buⁱ > Ph at 298 K (macro = macrocylic pincer ligand) reflects increasing steric hindrance at the reactive site. The inherent highly reactive nature of these complexes follows from k₂′ ≈ 10⁶ M^-1 s ^-1 for the R = Me system that is attenuated by only 100-fold in the R = Ph complex. A reaction mechanism is proposed based on computation of the enthalpic reaction profile for the R = Prⁱ system by DFT methods. The R = Et, Prⁱ, and Buⁱ systems display biphasic kinetics in which the initial fast process is followed by a slower first order process currently of uncertain origin.

Full text of DOI:10.1021/ic200942u

ARTICLE
pubs.acs.org/IC
Fast Carbon Dioxide Fixation by 2,6-Pyridinedicarboxamidato-
nickel(II)-hydroxide Complexes: Influence of Changes in Reactive Site
Environment on Reaction Rates
Deguang Huang,† Olga V. Makhlynets,‡ Lay Ling Tan,§ Sonny C. Lee,*,§ Elena V. Rybak-Akimova,*,‡ and
R. H. Holm*,†
†Department of Chemistry and Chemical Biology, Harvard University, Cambridge, Massachusetts 02138, United States
‡Department of Chemistry, Tufts University, Medford, Massachusetts 02155, United States
§Department of Chemistry, University of Waterloo, Waterloo, Ontario N2L3G1, Canada
S Supporting Information
b
ABSTRACT: The planar complexes [NiII(pyN2R2)(OH)]ꢀ, contain-
ing a terminal hydroxo group, are readily prepared from N,N0-(2,6-
C6H3R2)-2,6-pyridinedicarboxamidate(2-) tridentate pincer ligands
(R4N)(OH), and Ni(OTf)2. These complexes react cleanly and
completely with carbon dioxide in DMF solution in a process of
CO2 fixation with formation of the bicarbonate product complexes
[NiII(pyN2R2)(HCO3)]ꢀ having η1-OCO2H ligation. Fixation reac-
II
0
tions follow second-order kinetics (rate = k2 [Ni ꢀOH][CO2]) with
negative activation entropies (ꢀ17 to ꢀ28 eu). Reactions were
monitored by growth and decay of metal-to-ligand charge-transfer (MLCT) bands at 350ꢀ450 nm. The rate order R = Me >
macro > Et > Pri > Bui > Ph at 298 K (macro = macrocylic pincer ligand) reflects increasing steric hindrance at the reactive site. The
6
inherent highly reactive nature of these complexes follows from k2 ≈ 10 Mꢀ1 sꢀ1 for the R = Me system that is attenuated by only
0
100-fold in the R = Ph complex. A reaction mechanism is proposed based on computation of the enthalpic reaction profile for the R
= Pri system by DFT methods. The R = Et, Pri, and Bui systems display biphasic kinetics in which the initial fast process is followed by
a slower first order process currently of uncertain origin.
’ INTRODUCTION
highest reported value for nonenzymatic carbon dioxide fixation
and overlaps with the lower end of estimated rate constants for
human carbonic anhydrase in aqueous solution (k2 = kcat/
KM ∼106ꢀ108 Mꢀ1 sꢀ1).10 The enthalpic profile of this reaction
computed by density functional theory (DFT) methods provides
useful insight into the stationary point structures along the reac-
tion pathway. In this investigation, we have sought additional
characterization of this mode of carbon dioxide fixation. The
influence of changes in environment at the NiIIꢀOH reactive
group has been examined by determining the kinetics of five addi-
tional systems based on the pyN2R2 ligand platform. In addition,
we have found in several fixation reaction systems a second mea-
surable kinetics process.
In the present context, carbon dioxide fixation is a recognized
reaction type by which CO2 is converted to bicarbonate or
carbonate by reaction with metal-bound hydroxo groups. With
trivalent metals, the reactions frequently employ terminal MIIIꢀ
OH groups of chromium and cobalt and afford η1-bicarbonate
species as the initial product.1ꢀ4 With divalent metals, the
reactions proceed with terminal or bridging hydroxo ligation
and yield products that often contain carbonate bound in
binuclear or trinuclear bridging modes. Numerous examples of
these reactions are found with complexes of CoII, NiII, CuII, and
ZnII in particular.5ꢀ14 In the case of NiII, nearly all reactions
utilize bridged Ni-(OH)1,2-Ni units and lead to products in
which carbonate is stabilized by η2- or η3-coordination to two
NiII atoms.5,15ꢀ18
In contrast to foregoing reactions, we have discovered reaction
1 in Figure 1, which has several notable features.19 The reac-
tant contains an infrequent example of the terminal NiIIꢀOH
group,20ꢀ24 apparently stabilized by the steric protection pro-
vided by the 2,6-dimethylphenyl substituents of the tridentate
2,6-pyridinedicarboxamidate pincer ligand. Although compari-
sons are inexact owing to differences in solvent, the second-order
rate constant k2 = 9.5 ꢁ 105 Mꢀ1 sꢀ1 in DMF at 298 K19 is the
’ EXPERIMENTAL SECTION
Preparation of Compounds. All reactions and manipulations in
the preparation of NiII compounds were performed under a dinitrogen
atmosphere using either Schlenk techniques or an inert atmosphere
box. Volume reduction and drying steps were performed in vacuo.
Received: May 5, 2011
Published: September 09, 2011
r
2011 American Chemical Society
10070
dx.doi.org/10.1021/ic200942u Inorg. Chem. 2011, 50, 10070–10081
|

Inorganic Chemistry
ARTICLE
(t, 1), 8.45 (d, 2), 11.44 (s, 2). N,N0-bis(1-adamantyl)-2,6-pyridinedi-
carboximidate: 10 mmol scale, 1.62 g, 75%. 1H NMR (CDCl3): δ 1.74
(s, 18), 2.15 (s, 12), 7.48 (s, 2), 7.80 (t, 1), 8.27 (d, 2).
Ph2
H2pyN2 . An adaptation of a previous procedure27 was used for
the synthesis of precursor 2,6-diphenylaniline. Under anaerobic condi-
tions a solution of PhB(OH)2 (2.19 g, 18.0 mmol) in ethanol (12 mL)
was added to a solution of 2,6-dibromoaniline (1.51 g, 6.00 mmol) in
toluene(60mL). AqueousNa2CO3 solution(2M, 25mL) andPd(PPh3)4
(0.83 g, 0.72 mmol) were added, and the mixture was refluxed for 20 h at
85 °C. The organic layer was separated, and the aqueous phase extracted
with ether (3 ꢁ 50 mL). The combined organic phases were dried over
MgSO4 in air, and solvent was removed. The black residue was purified
on a silica column eluted with ethyl acetate/petroleum ether (1/9 v/v).
Solvent was removed from the eluate, and the residue was crystallized
from hot hexanes to afford pure 2,6-diphenylaniline as a white solid
(1.25 g, 85%). 1H NMR (CDCl3): δ 3.87 (br, 2), 6.93 (t, 1), 7.17 (d, 2),
7.39 (t, 2), 7.49 (t, 4), 7.55 (d, 4). A solution of pyridine-2,6-dicarbo-
nylchloride (0.61 g, 3.00 mmol) was added slowly to a mixture of 2,
6-diphenylaniline (1.25 g, 5.10 mmol) and triethylamine (0.83 mL,
6.00 mmol) in THF (120 mL) at 0 °C. The mixture was stirred for 40 h
and filtered. Solvent was removed, and the residue was purified on a silica
column eluted with dichloromethane and dichloromethane/methanol/
NH4OH (150:10:1 v/v/v) in succession to give the pure product as a
white solid (1.27 g, 80%). 1H NMR (CDCl3): δ 7.26 (t, 4), 7.34 (t, 8),
7.45 (d, 8), 7.51 (m, 6), 7.78 (t, 1), 7.97 (d, 2), 8.58 (s, 2).
Et2
Et2
(Et4N)[Ni(pyN2 )(OH)]. H2pyN2 (86 mg, 0.20 mmol) and
Ni(OTf)2 (71 mg, 0.20 mmol) were stirred in DMF (3 mL) for 1 h. The
light yellow suspension was treated with Et4NOH (25% in methanol,
176 mg, 0.30 mmol) and stirred for 1 h to give a red suspension. A
second equal portion of Et4NOH was added, and the mixture was stirred
for 10 h to form a deep red solution. The mixture was filtered through
Celite and ether (20 mL) was added to the filtrate to deposit a red oil
which was taken up in THF (3 mL) and treated with ether (10 mL) to
afford a red solid. This material was washed with ether, THF, and ether
and dried to give the product as a red crystalline solid (78 mg, 62%).
Absorption spectrum (DMF): λmax (εM) 413 (5200), 486 (sh, 2300) nm;
(CH2Cl2):λmax (εM) 407 (5180), 477 (sh, 2200) nm. 1HNMR(DMF-d7):
δ ꢀ5.08 (s, 1), 1.34 (t, 12), 2.98 (q, 8), 6.88 (s, br, 6), 7.52 (d, 2), 8.07
(t, 1). Anal. Calcd. for C35H50N4NiO3: C, 66.36; H, 7.96; N, 8.84.
Found: C, 66.44; H, 8.00; N, 8.86.
Figure 1. Depiction of CO2 fixation reactions of N,N0-disubstituted 2,6-
pyridinedicarboxamidato-NiIIꢀOH complexes and a macrocyclic com-
plex containing the same pincer chelate group. Numerical designations
of initial hydroxo and product bicarbonato complexes are indicated.
Commercial grade chemicals were used without further purification.
Tetrahydrofuran (THF) and diethyl ether were purified by an Innova-
tive Technology or MBraun purification system. Dimethylformamide
(DMF) was freshly distilled from CaH2 in vacuo at 55 °C and purged
with dinitrogen gas. All solvents were degassed before use. Compounds
were identified by a combination of spectroscopic and crystallographic
methods; representative compounds were analyzed (Midwest Microlab).
(Et4N)[Ni(pyN2iPr2)(OH)]. The preceding method was employed
on the same scale with use of H2pyN2iPr2. The red oil was treated with
ether (5 mL) to form an orange-red solid, which was dissolved in a
minimal volume of THF. The solution was stored at ꢀ20 °C overnight,
filtered, and ether added to the filtrate. The red material deposited was
washed with ethyl acetate (4 ꢁ 5 mL) and ether, and dried to afford the
product as a red solid (69 mg, 50%). Absorption spectrum (DMF): λmax
(εM) 413 (4580), 486 (sh, 2070) nm. 1H NMR (THF-d8): δ ꢀ5.25 (s, 1),
1.15 (d, 12), 1.38 (d, 12), 4.01 (m, 4), 6.89 (s, 6), 7.44 (d, 2), 7.86 (t, 1).
Anal. Calcd. for C39H58N4NiO3: C, 67.93; H, 8.48; N, 8.12. Found: C,
67.85; H, 8.55; N, 8.03.
1
Cation resonances are omitted from H NMR data. The compounds
H2pyN2Me2, (Et4N)[Ni(pyN2Me2)(OH)], (Et4N)[Ni(pyN2Me2)(HCO3)],
(Et4N)[Ni(⊂-pyN2dienMe3)(OH)], and (Et4N)[Ni(⊂-pyN2dienMe3)-
(HCO3)] were prepared as described.25,26
Pincer Ligands: N,N0-disubstituted 2,6-Pyridinedicarbox-
amidates, H2pyN2R2. The procedure for the R = Et ligand is typical.
To a mixture of 2,6-diethylaniline (1.49 g, 10.0 mmol) and triethylamine
(1.67 mL, 12.0 mmol) in THF (150 mL) at 0 °C was added slowly a
solution of pyridine-2,6-dicarbonylchloride (1.22 g, 6.00 mmol). The
reaction mixture was warmed to room temperature, stirred for 15 h, and
filtered. Solvent was removed and the light yellow mixture was purified
on a silica column eluted with dichloromethane/methanol/NH4OH
(150/5/1 v/v/v). The colorless oily residue after solvent removal was
treated with hexane (10 mL) and dried to give the pure product as a
white powder (1.97 g, 92%). 1H NMR (CDCl3): δ 1.21 (t, 12), 2.66 (q, 8),
7.19 (d, 4), 7.28 (m, 2), 8.17 (t, 1), 8.55 (d, 2), 9.08 (s, 2). R = Pri:
10 mmol scale, 2.27 g, 94%. 1H NMR (CDCl3): δ 1.23 (d, 24), 3.15 (m, 4),
7.24 (d, 4), 7.36 (t, 2), 8.19 (t, 1), 8.56 (d, 2), 9.01 (s, 2). R = Bui:
2.4 mmol scale, 0.57 g, 88%. 1H NMR(CDCl3): δ0.84 (d, 24), 1.84 (m, 4),
2.49 (d, 8), 7.13 (d, 4), 7.23 (t, 2), 8.16 (t, 1), 8.57 (d, 2), 9.08 (s, 2).
N,N0-bis(1-naphthyl)-2,6-pyridinedicarboximidate: 5.0 mmol scale,
(Me4N)[Ni(pyN2iBu2)(OH)]. The preceding method was used on
iBu2
a 0.10 mmol scale with use of H2pyN2
and Me4NOH (25% in
methanol). After the addition of the second portion of this reagent, the
mixture was stirred for 15 h, filtered through Celite, and ether (20 mL)
was added. The mixture was filtered and addition of a second portion
of ether (30 mL) resulted in precipitation of a red oily material that was
washed with THF (4 mL) and dissolved in THF/DMF (1 mL, 7:3 v/v).
Ethyl acetate was added slowly to precipitate a colorless solid. The
mixture was filtered, and ether was added to the filtrate to form a red
solid, which was washed with THF (4 mL) and dissolved in toluene/
DMF (2 mL, 10:1 v/v). Ether was diffused into this solution to give
the product as a red crystalline solid (26 mg, 38%). Absorption spec-
1
solvent acetonitrile (no chromatography step), 1.63 g, 78%. H NMR
1
((CD3)2SO): δ 7.70 (d, 2), 7.91 (d, 2), 7.99 (d, 2), 8.10 (d, 2), 8.35
trum (DMF): λmax (εM) 414 (4700), 486 (sh, 2100) nm. H NMR
10071
dx.doi.org/10.1021/ic200942u |Inorg. Chem. 2011, 50, 10070–10081

Inorganic Chemistry
ARTICLE
(DMF-d7): δ ꢀ5.15 (s, 1), 0.94 (m, 24), 2.55 (m, 4), 2.65 (m, 4), 3.15
379 (5600), 477 (sh, 990) nm. 1H NMR (DMF-d7): δ 1.28 (m, 12),
3.13 (m, br, 8), 6.84 (s, 6), 7.55 (s, 2), 8.16 (t, 1).
In the sections that follow, complexes are numerically designated as in
the Chart 1.
X-ray Structure Determinations. The structures of the seven
compounds in Tables 1 and 2 were determined; for brevity, compounds
are referred to by their anion number. Diffraction-quality crystals were
obtained from the following solvents: DMF/ether (2, 8, 9, 12), DMF/
ethyl acetate (3), DMF/toluene/ether (4, 5). Diffraction data were
collected on a Bruker CCD area detector diffractometer equipped
with an Oxford 700 low-temperature apparatus. Structures were solved
by direct methods using the SHELX program package.28 The ether
solvate molecule in 4 was treated as disordered with two equally occupied
(m, 4), 6.82 (s, 6), 7.49 (d, 2), 8.03 (t, 1).
Ph2
(Me4N)[Ni(pyN2 )(OH)]. H2pyN2Ph2 (62 mg, 0.10 mmol) and
Ni(OTf)2 (36 mg, 0.10 mmol) were stirred in DMF for 1 h. The light
yellow mixture was treated with Me4NOH (25% in methanol, 55 mg,
0.15 mmol) and stirred for 1 h to give a red suspension. A second equal
portion of Me4NOH was added, and the mixture was stirred for 5 h to
afford a dark red solution, which was filtered through Celite. Ether
(15 mL) was added to the filtrate to deposit an orange-red solid. The
solid was stirred in THF/propionitrile (8 mL, 3:1 v/v) for 45 min and
filtered through Celite to remove a sticky precipitate. Ether was added to
the filtrate to deposit a red solid that was taken up in THF/DMF
(3.5 mL, 6:1 v/v) to form a saturated solution. Toluene (3 mL) was
slowly added; after 1 h the solution was filtered to remove a colorless
solid. The red filtrate was diffused with ether to deposit the product as
block-like red crystals (23 mg, 30%) together with a small amount of
colorless crystals which were separated manually. Absorption spectrum
(DMF): λmax (εM) 414 (4050), 486 (sh, 1800) nm. 1H NMR (DMF-d7):
δ ꢀ4.86 (s, 1), 7.02 (d, 2), 7.10 (d, 4), 7.l6 (t, 2), 7.31 (m, 12), 7.64 (t, 1),
8.05 (m, 8).
Table 2. Crystallographic Data for Compounds Containing
[Ni(pyN2R2)(HCO3)]ꢀa
compounds
(Et4N)[8]
(Et4N)[9]
(Et4N)[12]
formula
C
36H50N4NiO5 C40H58N4NiO5 C37H53N7NiO5
M
677.51
monoclinic
P21/n
733.61
orthorhombic
Pnma
734.57
Et2
(Et4N)[Ni(pyN2 )(HCO3)]. Asolutionof(Et4N)[Ni(pyN2Et2)(OH)]
crystal system
space group
a, Å
triclinic
P1
(25 mg, 0.040 mmol) in DMF (2 mL) was bubbled with carbon dioxide
for 20 min. Diffusion of ether saturated with carbon dioxide into the red
solution resulted in separation of the product as block-like red crystals
(22 mg, 81%). Absorption spectrum (DMF): λmax (εM) 306 (7250),
378 (5600), 477 (sh, 940) nm; (CH2Cl2): λmax (εM) 304 (7070),
15.172(2)
12.394(1)
19.098(2)
90
22.679(2)
11.6237(8)
14.917(1)
90
10.521(5)
13.400(6)
15.030(7)
95.953(9)
104.788(9)
112.368(8)
1846.1(15)
2
b, Å
c, Å
α, deg
Chart 1. Abbreviations and Designation of Complexes
β, deg
102.029(2)
90
90
γ, deg
V, Å3
90
3512.5(6)
4
3932.3(5)
4
Z
μ, mmꢀ1
0.599
0.540
0.577
independent data
6427
4217
6626
refined parameters
768
338
605
R1 , wR2c (I >2σ(I)) 0.0807, 0.2312 0.0798, 0.2290 0.0615, 0.1629
b
R1, wR2 (all data)
0.1039, 0.2666 0.0951, 0.2498 0.0955, 0.1884
a T = 100(2) K, Mo Kα radiation (λ = 0.71073 Å). b R1 = ∑||Fo| ꢀ |Fc||/
∑|Fo|. c wR2 = {∑[w(Fo ꢀ Fc2)2/(Fo ) ]}1/2
.
2
2 2
Table 1. Crystallographic Data for Compounds Containing [Ni(pyN2R2)(OH)]ꢀa
(Et4N)[2]
(Et4N)[3] MeCO2Et
(Me4N)[4] Et2O
{(Me4N)[5]}5 3DMF
3
3
3
formula
C
35H50N4NiO3
C43H66N4NiO5
777.71
C43H68N4NiO4
763.72
C244H231N23Ni5O18
2967.3
M
633.50
crystal system
space group
a, Å
orthorhombic
Pca21
monoclinic
P21/c
orthorhombic
P212121
15.7551(6)
16.2340(6)
17.6627(7)
90
monoclinic
P21/n
16.1746(6)
25.954(1)
15.7902(6)
90
14.817(2)
14.424(2)
19.689(2)
90
18.031(1)
50.315(3)
23.340(2)
90
b, Å
c, Å
α, deg
β, deg
90
93.352(2)
90
90
98.914(1)
90
γ, deg
90
90
V, Å3
6628.8(4)
8
4200.6(7)
4
4517.6(3)
4
20919(2)
4
Z
μ, mmꢀ1
independent data
refined parameters
0.625
0.509
0.471
0.513
12593
10422
8565
39803
776
497
493
2616
R1 , wR2c (I >2σ(I))
0.0360, 0.0847
0.0428, 0.1077
0.0688, 0.1922
0.0532, 0.1109
0.0752, 0.1201
b
R1, wR2 (all data)
0.0400, 0.0877
0.0590, 0.1185
0.0837, 0.2083
2
a T = 100(2) K, Mo Kα radiation (λ = 0.71073 Å). b R1 = ∑||Fo| ꢀ |Fc||/∑|Fo|. c wR2 = {∑[w(Fo ꢀ Fc2)2/(Fo ) ]}1/2
.
2 2
10072
dx.doi.org/10.1021/ic200942u |Inorg. Chem. 2011, 50, 10070–10081

Inorganic Chemistry
ARTICLE
positions and atoms were described isotropically. The unit cell of 5
presents a very long b-axis owing to an unusual independent unit in the
structure; the cell contains four crystallographically independent units,
each of which has five anions, five cations, and three DMF solvate
molecules. All atoms in 8 were disordered in two positions with 3:2
occupancies in the cation and 1:1 in the cation, except for two carbon
atoms of one ethyl group which was treated as disordered over three
positions with occupancies of 4:3:3. Structure 9 has a crystallographic
mirror plane containing NiN3. The two independent isopropyl groups
were disordered over two positions with 85/15 and 60/40 occupancies;
carbon atoms of two ethyl groups of the cation also showed a 60/40
disorder. The two phenyl rings and the dien chain of 12 are disordered in
two positions with 55/45 occupancies. Hydrogen atoms were not added
to cations or solvate molecules in 4 and 9 owing to disordered carbon
atoms. The hydrogen atoms of the ꢀOH groups of 2ꢀ5 were located
geometrically. However, the hydrogen atom of the -HCO3 group of 12
was located from difference Fourier maps and refined isotropically. The
hydrogen atom was not added to the -HCO3 group of 8 because of its
highly disordered oxygen atoms. Crystal data and refinement details are
given in Tables 1 and 2.29 The compound tetraethylammonium [N,N0-
bis(1-naphthyl)-2,6-pyridinedicarboximidato]NiII crystallizes in monocli-
nic space group P21/c with a = 30.170(7) Å, b = 24.581(6) Å, c =
19.054(4) Å, β= 104.251(4)°, V =13.696(5) Å3; thestructurewasrefined
to R1 (wR2) = 0.0959 (0.1491) (all data).29 Other refinement details and
explanations (wherever necessary) are included in individual CIF files.
Kinetics Measurements. Kinetics of the reactions between NiII
complexes 2ꢀ6 and carbon dioxide were determined in DMF solutions.
Commercial DMF was dried over 3 Å molecular sieves for 3 d and then
distilled from CaH2 under vacuum (55 °C) and degassed by purging
with dry dinitrogen before use in an MBraun inert atmosphere box filled
with argon. Solutions of each complex were prepared immediately
before a kinetics run and were placed in Hamilton gastight syringes
equipped with three-way valves. Saturated solutions of CO2 in DMF
(0.20 M30,31) were prepared by bubbling CO2 gas through argon-
saturated DMF in a syringe at 298 K for 20 min. Solutions of lower
concentrations of CO2 were prepared by serial dilution of the 0.20 M
solution with argon-saturated DMF using graduated gastight syringes;
liquid phase occupied the entire available volume in the closed system
and was never exposed to the atmosphere and/or the gas/vapor phase.
Kinetics measurements were performed on a TgK Scientific (formerly
HiTech Scientific, Salisbury, Wiltshire, U.K.) SF-61DX2 cryogenic
stopped-flow system (mixing time 2ꢀ4 ms) equipped with tungsten
and xenon arc lamps. Solutions of the NiII complexes and CO2 in DMF
were separately thermostatted at low temperature (218ꢀ283 K) and
mixed in a 1:1 volume ratio. The mixing cell temperature was maintained
at (0.1 K. Concentrations of all reagents are reported for the onset of
the reaction (after mixing). CO2 (0.4ꢀ10 mM) was always in excess
to the NiII complex (usually 0.1 mM). Because the reactions are fast,
data were collected in a single-wavelength mode; repeated scans over a
range of wavelengths allowed reconstruction of time-resolved spectra.
For each complex, a series of kinetics traces in the range 370ꢀ450 nm
were collected in steps of 3ꢀ5 nm. In this way, the two processes were
separated and rate constants determined for each.
where A∞ is the absorbance of the reaction mixture after the reaction is
complete and ΔA = A0 ꢀ A∞ where A0 is the initial absorbance.
Similarly, double-exponential kinetics were analyzed using A = A∞
+
ΔA1 exp(ꢀk1t) + ΔA2 exp(ꢀk2t) in which ΔA1,2 and k1,2 were used as
fitting parameters. Each reported rate constant is an average of at least
7 runs with a standard deviation within 2%. To evaluate activation
parameters for each complex, variable temperature measurements were
performed at a wavelength that showed most clearly the growth of the
0
intermediate followed by its decay; linear Eyring plots (k2 = [(kBT/h)
exp(ꢀΔGq/RT) with ΔGq = ΔHq ꢀ TΔSq) were obtained for the
fast (direct reaction with CO2) and slow processes of several complexes
(2, 3, 4) and for the fast process of the remaining complexes (5, 6) which
did not exhibit sufficient absorbance changes for evaluation of the slow
process.
Computational Methods. Density functional calculations were
performed using the Gaussian 09 software package.32 Stationary points
and associated transition energies were characterized using the BP86
functional33,34 and 6-31G(d) basis set35,36 following our previously re-
ported protocol.19 For time-dependent DFT (TD-DFT) calculations,37
the 200 lowest energy singlet transitions, fully spanning the 250ꢀ
650 nm range, were determined using the B3LYP functional33,38,39
and 6-31G(d) basis set;35,36 additional test calculations investigated the
effects of other functional (BP86) and basis set (6-31+G(d),40 LanL2DZ41)
combinations and solvation (polarizable continuum model).42 Electro-
nic spectra were simulated with transitions represented by Gaussian fun-
ctions with full-width at half-maximum values of 3000 cmꢀ1 (GaussSum
2.2).43 Different functionals (BP86, B3LYP) and basis sets (LanL2DZ,
6-31G(d), 6-31+G(d)), in the gas phase and with a continuum solvent
model (PCM, DMF solvent), were investigated, and the simulated
spectra compared against the experimental spectra of selected hydroxo
and bicarbonate complexes. Empirically, B3LYP calculations were signi-
ficantly better at reproducing the general features in the experimental
spectra. Gas phase B3LYP/6-31G(d) calculations yielded the best quan-
titative results in terms of band maxima and intensities at reasonable
computational cost. Qualitative assignments for the electronic transi-
tions were obtained by visual inspection of orbital surfaces (GaussView)44
and Mulliken population analysis45 of relevant fragment orbital con-
tributions (QMForge).46
’ RESULTS AND DISCUSSION
A DFT analysis of the enthalpic profile of reaction 1 in which
hydroxo complex 1 is converted to bicarbonate product 7
indicates a low activation enthalpy and attendant unusually high
reaction rate. We interpret this property to originate in major part
from the favorable steric accessibility of the small substrate to the
NiIIꢀOH functional group. In addition, there are minimal struc-
tural changes of the complex in the fixation process.19 To better
understand this reaction type, it is significant to identify factors
that increase or decrease the reaction rate in a constant medium
while maintaining the essential NNN pincer ligand platform.
Reactant Complexes Containing the NiIIꢀOH Group. The
most accessible molecular structural variables are the N-substitu-
ents, the decreased size of which compared to 2,6-dimethylphenyl
in 1 might be expected to increase rates because of enhanced access
to the reactive site. However, phenyl47 or other groups of about the
same size or smaller48ꢀ50 stabilize octahedral bis-ligand complexes
instead of the desired planar stereochemistry. Also, a 1:1 reaction of
the N,N0-bis(1-naphthyl) ligandand Ni(OTf)2 gave octahedral [N,
N0-bis(1-naphthyl)-2,6-pyridinedicarboximidato]NiII, identified
crystallographically.29 In an attempt to obtain a 1:1 complex
with an aliphatic substituent, the N,N0-bis(1-adamantyl) ligand
was prepared, but it did not form a NiII complex. Ligands with
All complexes showed the formation and decay of an intermediate at
380ꢀ390 nm. The spectroscopic contributions of the second process
were negligible for 6 where essentially single-exponent behavior over the
entire spectral range was observed. Similarly, for 1 absorbance changes
due to the second process were very small at 370ꢀ390 nm, and the
complex displayed single-exponent kinetics when the reaction was
monitored at 400ꢀ500 nm.19 Rate constants were determined with
Kinetic Studio (TgK Scientific) and IgorPro programs using single-
exponential (AfB) or double-exponential (AfBfC) models. Kinetics
parameters for single-step processes were determined from single-
exponential fits to the pseudo-first-order rate law A = A∞+ ΔA exp(ꢀkt)
10073
dx.doi.org/10.1021/ic200942u |Inorg. Chem. 2011, 50, 10070–10081

Inorganic Chemistry
ARTICLE
Figure 2. Structures of the complexes [NiII(pyN2R2)(OH)]ꢀ with R = Et (2), Pri (3), Bui (4), Ph (5), and [NiII(pyN2R2)(HCO3)]ꢀ with R = Et (8), Pri
(9), and macro (12) showing 50% probability ellipsoids. Bond distances (Å) and angles (deg) in the hydroxo series are nearly invariant and are
represented by 2: NiꢀN1 1.924(2), NiꢀN2 1.825(2), NiꢀN3 1.917(2), NiꢀO3 1.828(2), N1ꢀNiꢀN2 82.7(1), N2ꢀNiꢀN3 82.4(1), N1ꢀNiꢀO3
96.0(1), N3ꢀNiꢀO3 98.9(1), N2ꢀNiꢀO3 177.6(1), N1ꢀNiꢀN3 165.1(1). For other complexes, NiꢀO3 = 1.820(2) (3), 1.822(3) (4), 1.835(9)
(5). The structures of 8, 9, and 12 in the NiꢀN3 region are very similar to those of 2. For 8: NiꢀO3 1.87(1), NiꢀO3ꢀC28 = O4ꢀC28ꢀO5 117(2),
O3ꢀC28ꢀO4 128(2), O4ꢀC28ꢀO5 117(2).
2,6-disubstituted phenyl groups are sufficiently hindered to form
only 1:1 complexes, as was originally shown by the isolation of
[Ni(pyN2iPr2)L] species with L = H2O and PMe351 and thereafter
complex 1.19 Consequently, new members 2ꢀ5 of the hydroxo
series 1ꢀ6 were prepared for kinetics study with CO2.
generated at 425 nm. Hydroxo ligand rotation splits the MLCT
transition between different 3d orbitals, resulting in a blue-shifted
main absorption and a lower energy shoulder. Although the in-
plane hydroxo conformation is calculated in the gas phase and
observed in the solid state, solvent interactions may reorient the
ligand in solution.
Carbon Dioxide Fixation. To compare the new fixation re-
actions reported here, it is useful to summarize previous results
for complexes 1 and 7 and 6 and 12.19,25 (i) Absorption spectra of
isolated (Et4N)[7] and the complex generated in solution by the
minimal reaction NiIIꢀOH + CO2 f NiIIꢀHCO3 are identical.
(ii) The time dependence of UV/visible absorbance changes of
the reaction system 1/CO2 in DMF at 400ꢀ450 nm fits
accurately to a single exponential function indicating one resol-
va0ble process. (iii) The system 1/CO2 is second order with rate =
k2 [1][CO2] and is described by the kinetics parameters in
Table 4. (iv) Complex 6 with CO2 forms 12, which was isolated
but reaction kinetics and the structure of 12 were not determined
in previous work.
Structures of hydroxo complexes 2ꢀ5, set out in Figure 2, are
planar; dimensions of 2 are typical.29 Over the set 1ꢀ6, the
complexes are characterized by NiꢀOH bond lengths in the
interval 1.802(4)ꢀ1.835(9) Å and upfield hydroxyl proton shifts
at δ ꢀ4.40 to ꢀ5.25. These signals are abolished by addition of
small amount of D2O. As seen in Figure 3, absorption spectra in
DMF solutions show maxima near 410 nm and prominent
shoulders at or near 450 nm. Given the utility of the main visible
feature in monitoring reaction kinetics (see below), the origin of
the UV/visible absorption bands was investigated by time-
dependent DFT (TD-DFT). Computed spectra of 3 are shown
in Figure 4, and additional details are found in Table 3. When the
OꢀH group is constrained to the coordination plane, the prin-
cipal band, a composite of two metal-to-ligand charge-transfer
(MLCT) absorptions involving the pyridyl π* acceptor orbitals,
occurs at 403 nm together with a second feature which is very
weak compared to the experimental spectrum. When the OꢀH
group is rotated out the plane by, for example, 60°, the principal
band shifts to 370 nm and a prominent low-energy shoulder is
(a). Reaction Systems. Kinetics of the reactions of hydroxo
complexes 2ꢀ6 with carbon dioxide in DMF (Figure 1) were in-
vestigated by stopped-flow techniques appropriate to processes
that are complete in several seconds or less. Spectral changes are
depicted in Figure 3 for the reactions of 2ꢀ5 and are similar to
10074
dx.doi.org/10.1021/ic200942u |Inorg. Chem. 2011, 50, 10070–10081

Inorganic Chemistry
ARTICLE
Figure 3. Absorption spectra of [NiII(pyN2R2)(OH)]ꢀ and their reaction products after passing CO2 through DMF solutions for 5ꢀ10 min. (a) R = Et,
(b) R = Pri, (c) R = Bui, (d) R = Ph. Band maxima are indicated.
each other and to the reactions of 1 and 6.25 Spectra of typical
complexes 2 and 8 are nearly identical in DMF and dichlor-
omethane (Experimental Section), indicating that DMF is not
acting as a ligand to hydroxo or bicarbonate complexes. Hereafter
we designate the system with initial reactants 1 + CO2 as the
DMP system and that with 3 + CO2 as the DIPP system. In all
reaction systems, band intensities of the starting complexes near
415 nm and at 360ꢀ380 nm decrease and increase, respectively,
as the bicarbonate reaction products 8ꢀ12 are formed. Products
8 and 9 (and 7 previously25) were isolated and identified by
structure determinations. A preparative method is given for
(Et4N)[8] (81% yield) but not for any compounds of 9ꢀ11
whose bicarbonate complexes proved too labile to loss of CO2 to
be isolated. Fortunately, a minute amount of crystalline (Et4N)[9]
was isolated from DMF/ether after several weeks, allowing a
structure determination. The structures of 8, 9, and 12 are
included in Figure 2 and are shown from a side-on perspective
in Figure 5 together with 7. These views emphasize the η1
binding mode of bicarbonate, the small deviations of the bound
oxygen atoms from the NiN3 plane, and the near or exact perpen-
dicular orientations of this plane and the CO3 plane. Among
bicarbonate complexes, 7, with the smallest ring substituents, is
the most stable in DMF solution.
are shown in Figure 6. At the beginning, values at 385 and 421 nm
increase and decrease, respectively, as formation of bicarbonate
species begins (Figure 3). The data at 385 nm are the more
revealing, showing a rapid growth in absorbance and subsequent
slow decay. Similar observations were made for reactions of 2 and
4. The fast process results in the formation of an intermediate
from [Ni(pyN2R2)(OH)]ꢀ and CO2 and the slower process,
which is independent of CO2 concentration, affords the product
bicarbonate complexes 8ꢀ10. The time course of the reaction
from reconstructed spectra of the DIPP system at 375ꢀ380 nm is
provided in Figure 7. Initial complex 3 (λmax 415, 470ꢀ480
(sh) nm, spectrum a) forms an intermediate (λmax 375ꢀ385 nm,
spectrum b) which converts to product 9 (λmax ∼380, 450ꢀ460
(sh) nm, spectrum c). As shown in Figure 8, the fast process is
dependent on CO2 concentration whereas the slowprocessis not.
Noting that the ligand R-substituents are alkyl in the reactions
of 2, 3, and 4, we examined the similarly substituted DMP system
at 380ꢀ390 nm in addition to 450 nm as previously. Rapid
growth of absorbance was followed by slow and very small decay,
suggesting that the intermediate and product spectra are very
similar. Fits of the absorbance at 380 nm to a two-exponential
rate law gave a pseudo-first-order rate constant k2 identical to that
evaluated from a one-exponential fit of the data at 450 nm.19
For phenyl-substituted complex 5 and macrocylic 6, absor-
bance data at 370ꢀ500 nm and 223ꢀ243 K were best treated
with single-exponential fits. For the reaction of 6, data at three
wavelengths show excellent fits to a single-exponential equation
at 223 K in DMF, as seen in Figure 9. Note the absence of
(b). Kinetics. Two types of kinetics behavior emerge. Absor-
bance data for the reactions of complexes 2, 3, and 4 displayed
biphasic behavior indicative of two consecutive processes, and
were fitted to a two-exponential equation. Taking the DIPP
system as an example, absorbance changes at three wavelengths
10075
dx.doi.org/10.1021/ic200942u |Inorg. Chem. 2011, 50, 10070–10081

Inorganic Chemistry
ARTICLE
absorbance decay at 370 and 382 nm as the reaction proceeds.
At 263 K and 370ꢀ390 nm, a rise and slight decay of absorbance
was observed for 6, suggesting an intermediate. However,
spectral differences between intermediate and product in systems
initially containing 5 and 6 are much less pronounced at this
temperature than in reactions of 2ꢀ4, such that kinetics param-
eters of a second step could not be reliably determined. The data
for 6 in the lower temperature range demonstrate one fast, clean
process of CO2 binding uncomplicated by a subsequent slow
process.
The kinetics results demonstrate that all hydroxo complexes
1ꢀ6 react rapidly with CO2 to form an intermediate which
undergoes a much slower transformation to the product bicar-
bonate complexes 7ꢀ12, completing the fixation reaction.
Biphasic behavior is most clearly developed with 2, 3, and 4, while
6 at low temperature is the most evident example of the fast process
kinetically isolated from the slower process. If the slow process
does occur, negligible spectral differences between the interme-
diate and bicarbonate product do not allow its spectrophoto-
metric observation. Reactions were studied under pseudo-first-
order conditions (large excess of CO2). Exponential shapes of
the kinetics traces of the fast reaction suggest that the reaction is
first-order in NiII, and values of the rate constants do not depend
on the initial NiII concentration. Values of the observed rate
constants increase with CO2 concentration (Figure 8), establish-
ing that the reaction is first order in CO2. Consequently, for the
II
0
0
fast process rate = k2 [Ni ꢀOH][CO2] where k2 is the second-
order rate constant. The slow process is independent of CO2
concentration and is described by rate = k1[NiII]. An extensive
Table 4. Kinetics Data for Carbon Dioxide Fixation by
[Ni(pyN2R2(OH)]ꢀ Complexes and a Slow Subsequent
Reaction in DMF Solutions
233
298
ΔHq
(kcal/mol) (cal/(mol K))a
ΔSq
k2
(Mꢀ1 sꢀ1
k2
(Mꢀ1 s ꢀ1 b
0
0
R
)
)
Fast Process: CO2 Fixation
Mec
Et
3.2 (5)
2.0 (2)
ꢀ20
ꢀ28
ꢀ17
ꢀ27
ꢀ25
ꢀ27
1.75 (1) ꢁ 105 9.5 ꢁ 105
0.46 (2) ꢁ 105 1.5 ꢁ 105
1.90 (4) ꢁ 104 2.7 ꢁ 105
1.03 (2) ꢁ 104 5.1 ꢁ 104
3.40 (1) ꢁ 103 2.8 ꢁ 104
1.17 (2) ꢁ 103 9.5 ꢁ 103
macro 5.0 (1)
Pri
Bui
Ph
3.0 (1)
4.0 (1)
4.0 (2)
k1233 (sꢀ1
)
d
R
ΔHq (kcal/mol)
ΔSq (cal/(mol K))a
Slow Process
Et
7.0 (1)
ꢀ27
ꢀ18
ꢀ19
1.4 (1)
Figure 4. Absorption spectra (gas phase) computed by TD-DFT
methodology. Upper panel: hydroxo complex 3 with NiꢀOꢀH group
in the coordination plane (red) or rotated by 60° out of the plane around
the NiꢀO bond (blue). Lower panel: bicarbonate complex 9 showing
intermediate P1 (red) and conformers P2 (blue) and P4 (green).
Pri
Bui
10.0 (5)
0.25 (1)
10.0 (1)
0.118 (5)
a Estimated standard deviation (3 cal/(mol K). b Extrapolated value.
c Ref 19. d Values independent of [CO2].
Table 3. Computed Electronic Absorption Bands by TD-DFT in the DIPP Reaction Systema
species
λmax, nm (εM)
assignments
Ni (z2) f pyridyl (π*)
3 (in-plane)b
403 (9400)
Ni (xz) + carboxamide (π) f pyridyl (π*)
Ph (π) + carboxamide (π) f pyridyl (π*)
Ni (yz) + OH (p)d f pyridyl (π*)
Ni (mixed d) + OH (p) f pyridyl (π*)
Ni (xz) + carboxamide (π) f pyridyl (π*)
Ph (π) f Ni(x2-y2)
3 (60° out-of-plane)b
370 (7100)
425 (sh, 2700)
290 (12,800)
359 (8700)
P1c
carboxamide (O) f pyridyl (π*)
Ni (yz) f pyridyl (π*)
Ph (π) + carboxamide (π) f pyridyl (π*)
419 (sh, 1500)
Ni (xz) + carboxamide (π) f pyridyl (π*)
a B3LYP/6-31G(d), gas phase; coordinate system: x-axis along NꢀNiꢀN, y-axis along NpyꢀNiꢀO, z-axis perpendicular to coordination plane. b OꢀH
group in coordination plane or rotated by 60° around the NiꢀO bond out of the plane. c P2 and P4 (Figure 4) show nearly identical features to P1. P2:
287 (13,600), 356 (8200), 417 (sh, 1600) nm. P4: 285 (13,900), 355 (8300), 418 (sh, 1500) nm. d O p-orbital perpendicular to NiꢀOꢀH plane.
10076
dx.doi.org/10.1021/ic200942u |Inorg. Chem. 2011, 50, 10070–10081

Inorganic Chemistry
ARTICLE
Figure 7. Spectra of 3 (a), intermediate in the reaction with CO2 (b),
and bicarbonate product 9 (c) reconstructed from single wavelength
kinetics traces acquired at 253 K for the DIPP system in DMF: [3]0 =
0.1 mM, [CO2]0 = 1.0 mM. Inset: kinetics trace at 223 K and a fit of the
data to a double-exponential rate law.
Figure 5. Structures of bicarbonate complexes 7, 8, 9, and 12 viewed
perpendicular to the chelate ring plane and showing the dihedral angles
between planes: α = NiN3/CO3, β and β0 = NiN3/phenyl (two values).
Also shown is the dihedral angle NiN3/NiN2O, the displacement (Å) of
the coordinated oxygen atom from the NiN3 plane, and the NiꢀO and
CꢀO distances (Å). The CꢀO(5) bond distance of 9 is not given
because of disorder.
Figure 8. Plot showing dependence of rate constant on [CO2] for the
fast process (k2) and lack of dependence for the slow process (k1). Data
were acquired at 390 nm and 253 K in DMF with [3]0 = 0.1 mM and
variable [CO2].
of 3200 equiv of H2O or D2O followed by kinetics measure-
ments of the reaction between 3 and CO2 in DMF at 223 K
0
showed that k2 was reduced by about 50% compared to the
system with no water added, consistent with a previous finding
for 1 and CO2.19 Further, the ratio k2 (D2O)/k2 (H2O) =
1.27, a result showing that breaking the OꢀH(D) bond is not
involved in the rate-limiting step of CO2 fixation with 3 and by
implication with other members of the set 1ꢀ6. Also, the
presence or absence of water (H2O and D2O) has no effect on
the rate constant of the second process (k1 = 0.09ꢀ0.11 sꢀ1).
Variation of the concentration of 3 over the range 0.05ꢀ
0.20 mM with excess CO2 at 223 K yielded absorbance traces at
390 nm that were fit accurately with a double exponential rate
0
0
law first-order in NiII for both the fast and the slow processes.
3
For the fast reaction, k2 = (9.5ꢀ8.9) ꢁ 10 Mꢀ1 sꢀ1, while for
0
the slow reaction k1 = 0.24ꢀ0.46 sꢀ1. Attempts to fit the latter
reaction to a second-order rate law were unsatisfactory. The
linear increase in k1 with increasing NiII concentration suggests
a possible second-order NiII contribution. However, a plot of
the data reveals a large nonzero intercept corresponding to,
primarily, a reaction first-order in NiII. These data do not
clearly point to a process such as μ2-carbonate bridge forma-
tion as the origin of the slow reaction.
Figure 6. Kinetics traces at 370, 385, and 421 nm acquired for the DIPP
system at 253 K in DMF: [3]0 = 0.1 mM, [CO2]0 = 1.0 mM.
body of kinetics data on which the foregoing results are based is
available.29
Reactions have been conducted under two other variable
conditions that could bear on reaction mechanism.29 Addition
10077
dx.doi.org/10.1021/ic200942u |Inorg. Chem. 2011, 50, 10070–10081

Inorganic Chemistry
ARTICLE
shows the lowest activation entropy, such that the contribu-
tions to ΔGq from the activation entropy and enthalpy are
practically equal.
(c). Reaction Mechanism. We have previously carried out a
computational analysis of the energetics of the DMP system by
DFT methods.19 The DIPP system has been similarly analyzed in
this work; the computed enthalpic reaction profile is given in
Figure 11, and the energies and entropies of the equivalent
stationary points in the two systems are compared in Table 5.
The lowest energy pathway is summarized by the following
sequence: (i) formation of a precursor complex (PC) by ap-
proach (ca. 2.5ꢀ3.0 Å) of the carbon atom of CO2 to the
hydroxo oxygen atom; (ii) insertion of CO2 into the NiꢀO bond
to form a four-centered transition state (TSins); (iii) detachment
of the NiꢀOH bond to generate the Ni-η1-OCO2H insertion
product (P1); (iv) NiꢀO bond rotations to the most stable
bicarbonate conformers (P2 or P4 (not shown), in which the
carbonyl oxygen atom and the hydroxyl proton are mutually syn
or anti). The profiles for the two systems are quite similar; a
detailed description of the DMP system is given elsewhere.19
For the DMP system, gas phase BP86/6-31G(d) calculations
yield results that agree well with experimental findings in DMF
solution. Overall, the computed energetics for the DIPP system
differ little from those of the DMP system, with the calculated
activation parameters for CO2 insertion lower for the more
hindered system by about 1.3 kcal/mol enthalpically and 3 cal/
(mol K) entropically. Assuming that this step corresponds to the
initial fast process observed experimentally, the changes in
activation parameters agree with the kinetics measurements
within acceptable errors for calculations of this type. On an
absolute scale, the calculated activation enthalpy for the DIPP
system is essentially identical with experiment, while the activa-
tion entropy deviates substantially in the negative direction. The
same entropy error was observed in the DMP system and prob-
ably originates from the inaccurate description of low frequency
vibrational modes.
Figure 9. Kinetics traces at 370, 382, and 418 nm and their fits to a
one-exponential rate law for the reaction between 6 and CO2 in DMF at
223 K: [6]0 = 0.1 mM, [CO2] = 1.0 mM. Rate constants are the averages
of at least seven runs.
Figure 10. Eyring plots for the CO2 fixation reactions of
[NiII(pyN2R2)(OH)]ꢀ and [Ni(⊂-pyN2dienMe3)(OH)]ꢀ (macro) in
0
DMF (Figure 1) as dependent on substituents R. k2 = kobs/[CO2];
Following the initial fast process of CO2 insertion, what is the
nature of the subsequent slow process found with the apparently
more hindered ligand systems (2ꢀ4)? Given that the initial
insertion product P1 is not predicted to be the most stable
conformer, and P2 (or possibly P4 for the DIPP system) is the
actual conformer observed crystallographically, it is tempting to
ascribe the slow process to the rearrangement of the kinetic
product P1 to the more stable bicarbonate orientations found in
P2 and P4. The calculated barriers for these rearrangements
relative to P1, however, differ little between the DIPP and DMP
systems, and there is no slow process cleanly resolvable for the
latter system. The CO2 insertion is the rate-limiting step in our
computational models for both systems. We have found no
computational evidence for a subsequent high-barrier process in
any of the reaction paths investigated. We also find a non-
insertion pathway in which CO2 forms an adduct NiꢀO(H)CO2
following by proton transfer to form a bicarbonate product to be
substantially higher in energy (>10 kcal/mol) than the pathway
of Figure 11.
[CO2] = 1.0 mM after mixing in the stopped flow instrument.
Temperature dependencies of rate constants for the fast
processes of all reaction systems and slow processes in selected
systems give excellent fits to the Eyring equation. Plots for the
fast process are set out in Figure 10, and activation parameters are
collected in Table 4 for the fast and slow processes. The negative
activation entropies are consistent with a second0order reaction
and an associative transition state. At 233 K, k2 values follow
order (3) written in terms of variable ligand substituents R.
Numerical values are ratios of rate constants; for example, 5
reacts 123 times slower than 1.
Með1Þ > Etð3:2Þ > macroð7:6Þ > Prið14Þ > Buið43Þ > Phð123Þ ð3Þ
Eyring plots for systems with acyclic substituents are essentially
parallel such that the rate order is unchanged over the tempera-
ture range. There is a qualitative trend toward slower rates as the
steric size of the R groups increases. However, because of
experimental uncertainties, this order does not correlate clearly
with either ΔHq or ΔSq values. A rather typical trend in binding
small molecules to metals is that increased activation entropies
relate to less favorable binding because of limited access to the
reactive site.52 The macrocyclic reaction system differs from the
others in that the Eyring plot is not parallel, and the position of
macro and Et in series 3 must be inverted at 298 K. This system
Notwithstanding the predicted reaction energetics, the UV/
visible spectra of the bicarbonate conformers in the DIPP system
were calculated by TD-DFT to see if the intermediate P1 and
products differ to any significant extent. As for hydroxo complex
3, all primary bands of 9 are simulated by this approach using
optimized geometries (Figure 4, Table 3). The theoretical
maxima at 287, 356, and 417 (sh) nm of P2, for example, have
10078
dx.doi.org/10.1021/ic200942u |Inorg. Chem. 2011, 50, 10070–10081

Inorganic Chemistry
ARTICLE
Figure 11. BP86/6-31G(d) reaction profile and stationary point structures for CO2 fixation in the DIPP reaction system. Energy values (kcal/mol) are
standard enthalpies of reaction or activation.
Table 5. Computed Standard Reaction Enthalpies and
Entropies (BP86/6-31G(d)) for the DMP and DIPP CO2
Fixation Systemsa
While a close agreement cannot be expected between an experi-
mental spectrum, particularly a reconstituted spectrum of non-
optimal resolution, and a theoretical spectrum, we do associate
spectrum b of the reaction intermediate in the DIPP system
(λmax 375ꢀ380 nm, with a broad low-energy shoulder, Figure 7)
with that of P1 (λmax 359, 419 (sh) nm). Spectrum c would then
be related to the P2 and P4 spectra. The differences between the
three computed spectra are so small that no structural differences
can be inferred between the kinetic product of the fast process
(P1) and the product of the slow process (P2 and/or P4). The
nature of the latter remains to be determined.
speciesb
DMP
DIPP
PC
ꢀ3.1c (ꢀ26)d
4.5 (ꢀ36)
18.4 (ꢀ34)
ꢀ7.9 (ꢀ34)
ꢀ2.2 (ꢀ38)
1.3 (ꢀ40)
ꢀ9.2 (ꢀ35)
ꢀ0.3 (ꢀ35)
ꢀ6.0 (ꢀ34)
ꢀ1.4 (ꢀ34)
ꢀ9.6 (ꢀ35)
ꢀ1.1 (ꢀ38)
ꢀ1.6 (ꢀ41)
ꢀ2.9 (ꢀ26)
3.2 (ꢀ39)
18.5 (ꢀ36)
ꢀ8.0 (ꢀ35)
ꢀ3.3 (ꢀ40)
1.3 (ꢀ41)
ꢀ9.8 (ꢀ37)
ꢀ1.1 (ꢀ38)
ꢀ7.0 (ꢀ35)
ꢀ2.2 (ꢀ37)
ꢀ9.8 (ꢀ38)
ꢀ2.0 (ꢀ37)
ꢀ2.2 (ꢀ42)
TSins
TSpt
P1
TSpiv12
TSrot12
P2
Summary. The following are the principal results and con-
clusions of this investigation, including certain results from prior
work.19,25
TSrot13
P3
(1) Members of the family of planar complexes [NiII(pyN2R2)-
(OH)]ꢀ containing tridentate 2,6-pyridinedicarboxamidate
pincer ligands with variable 2,6-C6H3R2 N-substituents
and terminal hydroxo ligands are readily prepared.
TSrot24
P4
TSpiv34
TSrot34
(2) The complexes in (1) react cleanly and rapidly with CO2
in DMF solution to afford the products [NiII(pyN2R2)-
(HCO3)]ꢀ, some of which have been isolated. Bicarbo-
nate complexes contain the η1-OCO2H ligation mode
with the CO3 group nearly or exactly perpendicular to the
coordination plane. These reactions are examples of CO2
fixation.
(3) Fixation reactions are very rapid, low-barrier second-
order processes with negative activation entropies indi-
cative of associative transition states. At 298 K, rate
constants k20 range from 9.5 ꢁ 105 (R = Me) to 9.5 ꢁ
103 Mꢀ1 sꢀ1 (R = Ph), with the rate order overall
correlating with the size of the R-substituent. Kinetics
a Energies are referenced against the separated reactants using the
conformer of 1 or 3 in which the phenyl rings of are canted toward
each other on one side of the NiN3O coordination plane; this is the
conformer observed crystallographically. An alternate conformation in
which the rings are oriented roughly parallell to each other19 is slightly
lower in energy (ꢀ0.3 kcal/mol). b Bolded entries are defined in
Figure 11; other entries refer to stationary points on the reaction
c
pathway and are specified in ref 19 (Scheme 1). Enthalpy, kcal/mol.
d Entropy, cal/(mol K).
apparent experimental counterparts at 306, 369, and 475 nm.
The two lower energy features contain MLCT components.
10079
dx.doi.org/10.1021/ic200942u |Inorg. Chem. 2011, 50, 10070–10081

Inorganic Chemistry
comparisons with other metal-based CO2 fixation sys-
tems are not highly accurate because these systems have
been conducted in aqueous media.
(4) The majority of reaction systems are biphasic, showing
a fast fixation reaction followed by a slow first order
process (k1 ∼ 0.1ꢀ1.4 sꢀ1, 233 K) of currently unknown
origin.
ARTICLE
(6) Kitajima, N.; Fujisawa, K.; Koda, T.; Hikuchi, S.; Moro-oka, Y.
J. Chem. Soc., Chem. Commun. 1990, 1357–1358.
(7) Murthy, N. N.; Karlin, K. D. J. Chem. Soc., Chem. Commun. 1993,
1236–1238.
(8) Bazzicalupi, C.; Bencini, A.; Bencini, A.; Bianchi, A.; Corana, F.;
Fusi, V.; Giorgi, C.; Paoli, P.; Paoletti, P.; Valtancoli, B.; Zanchini, C.
Inorg. Chem. 1996, 35, 5540–5548.
(9) Looney, A.; Han, R.; McNeill, K.; Parkin, G. J. Am. Chem. Soc.
1993, 115, 4690–4697.
(10) (a) Zhang, X.; van Eldik, R. Inorg. Chem. 1995, 34, 5606–5614.
(b) Nakata, K.; Shimomura, N.; Shiina, N.; Izumi, N.; Ichikawa, K.;
Shiro, M. J. Inorg. Biochem. 2002, 255–266. (c) Notni, J.; Schenk, S.;
G€orls, H.; Breitzke, H.; Anders, E. Inorg. Chem. 2008, 47, 1382–1390.
(11) Bergquist, C.; Fillebeen, T.; Morlok, M. M.; Parkin, G. J. Am.
Chem. Soc. 2003, 125, 6189–6199.
(5) Enthalpic reaction profiles for the R = Me and Pri systems
computed by DFT methods present a five-stage mechan-
ism implicating a four-center (Ni-η2-OCO2H) transition
state, an insertion product intermediate (Ni-η1-OCO2H),
and bicarbonate reorientation to the stable product
conformer. The insertion of CO2 into a NiꢀOH bond
is a previously recognized event in the fixation process. In
comparison, CoIII complexes react without bond-breaking,1
as might be expected for six-coordinate low-spin reactants
with lesser kinetic lability.
(12) Peter, A.; Vahrenkamp, H. Z. Anorg. Allg. Chem. 2005, 631,
2347–2351.
(13) Zhang, X.; van Eldik, R.; Koike, T.; Kimura, E. Inorg. Chem.
1993, 32, 5749–5755.
(6) Reaction rates are among the fastest reported for CO2
fixation although comparisons are necessarily imprecise
owing to solvent differences, with the large majority of
existing data having been obtained in aqueous systems.
However, the intrinsic facility of the fixation reaction is
emphasized by the decrease in rate constant of only 100
at 298 K upon replacement of a small steric impediment
(R = Me) near the reactive site with much larger groups
(R = Bui, Ph).
(14) Company, A.; Jee, J.-E.; Ribas, X.; Lopez-Valbuena, J. M.;
Gꢀomez, L.; Corbella, M.; Llobet, A.; Mahia, J.; Benet-Buchholz, J.;
Costas, M.; van Eldik, R. Inorg. Chem. 2007, 46, 9098–9110.
(15) Carmona, E.; Marín, J. M.; Palma, P.; Paneque, M.; Poveda,
M. L. Inorg. Chem. 1989, 28, 1895–1900.
(16) Ito, M.; Takita, Y. Chem. Lett. 1996, 929–930.
(17) Lozano, A. A.; Sꢀaez, M.; Pꢀerez, J.; García, L.; Lezama, L.; Rojo,
T.; Lꢀopez, G.; García, G.; Santana, M. D. Dalton Trans. 2006, 3906–3911.
(18) Wikstron, J. P.; Filatov, A. S.; Mikhalyova, E. A.; Shatruk, M.;
Foxman, B. M.; Rybak-Akimova, E. V. Dalton Trans. 2010, 39, 2504–2514.
(19) Huang, D.; Makhlynets, O. V.; Tan, L. L.; Lee, S. C.; Rybak-
Akimova, E. V.; Holm, R. H. Proc. Natl. Acad. Sci. U.S.A. 2011, 108,
1222–1227.
Current research is directed toward an elucidation of the slow
first-order process and of metal variation on the reaction rate.
(20) Kieber-Emmons, M. T.; Schenker, R.; Yap, G. P. A.; Brunold,
’ ASSOCIATED CONTENT
T. C.; Riordan, C. G. Angew. Chem., Int. Ed. 2004, 43, 6716–6718.
ꢀ
(21) Cꢀampora, J.; Palma, P.; del Rio, D.; Alvarez, E. Organometallics
S
Supporting Information. Crystallographic data for the
b
2004, 23, 1652–1655.
(22) Cꢀampora, J.; Matas, I.; Palma, P.; Graiff, C.; Tiripicchio, A.
Organometallics 2005, 24, 2827–2830.
(23) Adhikari, D.; Mossin, S.; Basuli, F.; Dible, B. R.; Chipara, M.;
Fan, H.; Huffman, J. C.; Meyer, K.; Mindiola, D. J. Inorg. Chem. 2008,
47, 10479–10490.
compounds in Tables 1, 2 and the Et4N+ salt of [N,N'-bis(1-
naphthyl)-2,6-pyridinedicarboximidato]NiII, kinetics data perti-
nent to the reactions of [NiII(pyN2R2)(OH)]ꢀ with carbon
dioxide. This material is available free of charge via the Internet
at http://pubs.acs.org.
(24) Powell-Jia, D.; Ziller, J. W.; DiPasquale, A. G.; Rheingold, A. L.;
Borovik, A. S. Dalton Trans. 2009, 2986–2992.
(25) Huang, D.; Holm, R. H. J. Am. Chem. Soc. 2010, 132, 4693–
4701.
’ AUTHOR INFORMATION
Corresponding Authors
*E-mail: sclee@uwaterloo.ca (S.C.L.), elena.rybak-akimova@
tufts.edu (E.V.R.-A.), holm@chemistry.harvard.edu (R.H.H.).
(26) Abbreviations are given in the Chart 1.
(27) Miura, Y.; Oka, H.; Momoki, M. Synthesis 1995, 1419–1422.
(28) Sheldrick, G. M. Acta Crystallogr. 2008, A64, 112–122.
(29) See paragraph at the end of this article for Supporting Informa-
tion available.
(30) Gennaro, A.; Isse, A. A.; Vianello, E. J. Electroanal. Chem. 1990,
289, 203–215.
(31) Tezuka, M.; Iwasaki, M. Chem. Lett. 1993, 427–430.
(32) Frisch, M. J. et al. Gaussian 09; Gaussian, Inc.: Wallingford,
CT, 2009.
(33) Becke, A. D. Phys. Rev. A 1988, 38, 3098–3100.
(34) Perdew, J. P. Phys. Rev. B 1988, 33, 8822–8824.
(35) Rassolov, V. A.; Pople, J. A.; Ratner, M. A.; Windus, T. L.
J. Chem. Phys. 1998, 109, 1223–1229.
’ ACKNOWLEDGMENT
This research was supported by grants from the National
Science Foundation at Harvard University (CHE 084639) and
at Tufts University (CHE 0750140 and CRIF CHE 0639138),
and at the University of Waterloo by NSERC, CFI, ORF, and
SHARCNET (computational facilities).
’ REFERENCES
(36) Rassolov, V. A.; Ratner, M. A.; Pople, J. A.; Redfern, J. A.;
Curtiss, L. A. J. Comput. Chem. 2001, 22, 976–984.
(37) Stratmann, R. E.; Scuseria, G. E.; Frisch, M. J. J. Chem. Phys.
1998, 109, 8218–8224.
(38) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785–789.
(39) Miehlich, B.; Savin, A.; Stoll, H.; Preuss, H. Chem. Phys. Lett.
1989, 157, 200–206.
(1) Palmer, D. A.; van Eldik, R. Chem. Rev. 1983, 83, 651–731.
(2) Acharya, A. N.; Dash, A. C. Proc. Indian Acad. Sci. (Chem. Sci.)
1993, 105, 225–233.
(3) Jacewicz, D.; Dabrowska, A.; Kolisz, I.; Chmurzynski, L.; Banecki,
B.; Wozniak, M. Transition Met. Chem. 2005, 30, 209–216.
(4) Jacewicz, D.; Chylewska, A.; Daakbrowska, A.; Chmurzynski, L.
Z. Anorg. Allg. Chem. 2007, 633, 1493–1499.
(5) Kitajima, N.; Hikuchi, S.; Tanaka, M.; Moro-oka, Y. J. Am. Chem.
Soc. 1993, 115, 5496–5508.
(40) Clark, T.; Chandrasekhar, J.; Spitznagel, G. W.; Schleyer,
P. v. R. J. Comput. Chem. 1983, 4, 294–301.
10080
dx.doi.org/10.1021/ic200942u |Inorg. Chem. 2011, 50, 10070–10081

Inorganic Chemistry
ARTICLE
(41) Dunning, T. H., Jr.; Hay, P. J. In Modern Theoretical Chemistry;
Plenum: New York, 1976; pp 1ꢀ26.
(42) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. Rev. 2005, 105,
2999–3093.
(43) O’Boyle, N. M.; Tenderholt, A. L.; Langner, K. M. J. Comput.
Chem. 2008, 29, 839–845.
(44) Dennington, R.; Keith, T.; Millam, J. GaussView, Version 5;
Semichem, Inc.: Shawnee Mission, KS, 2009.
(45) Mulliken, R. J. Chem. Phys. 1955, 23, 1833–1840.
(46) Tenderholt, A. L. QMForge, v. 2.1; Stanford University:
Stanford, CA, 2007; http://qmforge.sourceforge.net.
(47) Patra, A. K.; Mukherjee, R. Inorg. Chem. 1999, 38, 1388–1393.
ꢁ
(48) Swiatek-Koslowska, J.; Gumienna--Kontecka, E.; Dobosz, A.;
Golenya, I. A.; Fritsky, I. O. Dalton Trans. 2002, 4639–4643.
(49) Alcock, N. W.; Clarkson, G.; Glover, P. B.; Lawrance, G. A.;
Moore, P.; Napitupulu, M. Dalton Trans. 2005, 518–527.
(50) Liu, S.-G.; Ni, C.-L. Acta Crystallogr. 2007, E63, m1373–m1374.
(51) Wasilke, J.-C.; Wu, G.; Bu, X.; Kehr, G.; Erker, G. Organo-
metallics 2005, 24, 4289–4297.
(52) Rybak-Akimova, E. V. In Physical Inorganic Chemistry; Bakac, A.,
Ed.; Wiley: New York, 2010; Vol. 2, pp 109ꢀ188.
10081
dx.doi.org/10.1021/ic200942u |Inorg. Chem. 2011, 50, 10070–10081