Naphthalene diimide based materials with adjustable redox potentials: Evaluation for organic lithium-ion batteries

Article abstract of DOI:10.1021/cm503800r

The promising crystallinity and tunable redox capabilities of naphthalene diimides make them attractive candidates as electroactive materials for organic-based lithium-ion batteries. In this study, a family of naphthalene diimide derivates was synthesized and their redox properties explored with the intent of unveiling structures with reduction potentials that are higher than those encountered in previous organic redox processes. Changes in the electronic characteristics of the aryl substituents resulted in materials with discharge potentials that vary from 2.3 to 2.9 V vs Li/Li+, with discharge capacities as high as 121 mAh/g.

Full text of DOI:10.1021/cm503800r

Subscriber access provided by Penn State | University Libraries
Article
Naphthalene Diimide Based Materials with Adjustable Redox
Potentials: Evaluation for Organic Lithium-Ion Batteries
Geeta S. Vadehra, Ryan P Maloney, Miguel A. Garcia-Garibay, and Bruce Dunn
Chem. Mater., Just Accepted Manuscript • DOI: 10.1021/cm503800r • Publication Date (Web): 01 Dec 2014
Downloaded from http://pubs.acs.org on December 5, 2014
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Chemistry of Materials is published by the American Chemical Society. 1155 Sixteenth
Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 8
Chemistry of Materials
1
2
3
4
5
6
7
8
9
Naphthalene Diimide Based Materials with Adjustable Redox Poten-
tials: Evaluation for Organic Lithium-Ion Batteries
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Geeta S. Vadehra, Ryan P. Maloney, Miguel A. Garcia-Garibay*, Bruce Dunn*
Department of Chemistry and Biochemistry, University of California, Los Angeles, California 90095, Department of
Materials Science and Engineering, University of California, Los Angeles, California 90095.
ABSTRACT: The promising crystallinity and tunable redox capabilities of naphthalene diimides make them attractive
candidates as electroactive materials for organic-based lithium-ion batteries. In this study, a family of naphthalene
diimide derivates was synthesized and their redox properties explored with the intent of unveiling structures with reduc-
tion potentials that are higher than those en-
countered in previous organic redox processes.
Changes in the electronic characteristics of the
aryl substituents resulted in materials with
discharge potentials that vary from 2.3 to 2.9 V
vs. Li/Li+, with discharge capacities as high as
121 mAh/g.
has focused on increasing cycling capabilities by decreas-
ing their solubility6-10 with the use of rigid aromatics9-10,11
or insoluble salts6,9. However, their lithium-insertion po-
tentials are still too low to compete with the nearly 4 V vs
Li/Li+ observed with other transition metal oxides. As
illustrated in Table 1, research in organic electrodes thus
far has incorporated three main categories of organic re-
dox-active functionalities:7,12 nitroxyl radicals, disulfides,
and conjugated dicarbonyls.
INTRODUCTION
Lithium ion batteries have achieved enormous success in
the area of energy storage in small electronics due to their
impressive energy densities and cyclability.1,2 These bat-
teries are based on the use of Li intercalation compounds.
Li ions are transported between two hos+t structures, a
graphite negative electrode that is electrochemically con-
nected by a non-aqueous electrolyte to the positive elec-
trode, commonly a transition metal oxide compound.
With an output voltage of ~4 V this process works rela-
tively well, however, the required metal oxides (LiCoO2,
LiNiO2, LiMn2O4) have some drawbacks: they are relative-
ly scarce, heavy, toxic and their production requires ex-
pensive high temperature processing.3 In recent years,
there has been growing interest in creating electroactive
organic electrode materials that use natural organic
sources as precursors and generally lead to environmen-
tally benign processing.4
Nitroxyl radicals13 can be oxidized with potentials of ca.
3.5-3.6 V vs. Li/Li+13. The cationic charged state of this
mechanism designates it as a “p-type” redox mechanism.
This narrow range is determined by its relatively isolated
two-atom (N-O.) redox system. Furthermore, having a
one-electron redox mechanism results in a more limited
energy storage capacity in comparison to the two-electron
processes displayed by disulfides14,15 and conjugated di-
carbonyls.6,8,9,16 In addition, considering that nitroxyls are
oxidized in the charging process, their function depends
on the insertion of charge-compensating stable anions.
Many of these concerns can be addressed by taking ad-
vantage of organic electrodes based on bulk crystals of
suitable redox active compounds. First explored in the
early 1970s,5 organic redox-active compounds tend to be
less expensive, less toxic and potentially much lighter
than their inorganic counterparts. Their performance,
however, has been limited by their inability to reach high
redox potentials, and by their tendency to dissolve in
non-aqueous electrolytes, which limits their use to a few
charge-discharge cycles. Nonetheless, as the limits of
inorganic insertion metal oxides are being reached, or-
ganic redox materials are being revisited.6 Recent work
-
The large anions such as PF6 that are common in Li-ion
electrolytes cannot be inserted readily into the corre-
sponding material.
The two-electron reduction of disulfides into the corre-
sponding lithium thiolate species, illustrated with deriva-
tive 2, is also not ideal. Disulfides display a relatively low
potential, typically ranging from 1.8-2.3 V vs. Li/Li,+7 and
have a redox mechanism that seems not to lend itself to
large electronic variations due to the isolated nature of
ACS Paragon Plus Environment

Chemistry of Materials
Page 2 of 8
1
2
3
Table 1. Summary of processes typical to each category: disulfide bond, nitroxide radicals and conjugated car-
bonyls.
4
Redox Structure
Category
Discharge Redox Theoretical Reference
Potential
Type
Capacity
5
(V vs
(mAh/g)
6
7
Li/Li++)
8
9
Nitroxyl
Radical
3.5
p-type
111
13
OR
OR
N
-1 e- + A-
+1 e- - A-
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1+
1
N
O
O
.
Disulfide
1.8
1.7
n-type
n-type
282
235
15
9
S
S
SLi SLi
+ 2 e- + 2 Li+
- 2 e- - 2 Li+
2
22-
Dicarbonyl
O
OLi
O
O
O
+ 2 e- + 2 Li+
LiN
LiO
NLi
LiN
NLi
3
32-
- 2 e- - 2 Li+
O
O
-
+
Dicarbonyl
Dicarbonyl
0.8
2.3
n-type
n-type
301
257
18
17
O
OLi
O
+ 2 e + 2 Li
LiO
LiO
OLi
OLi
4
42-
- 2 e- - 2 Li+
LiO
O
O
OLi
OLi
+ 2 e- + 2 Li+
- 2 e- - 2 Li+
52-
5
this particular functional group. In contrast, the reduc-
tion of conjugated carbonyls into the corresponding
dialkoxides seem to have a larger range of molecular
well compounds with electron donating X=NMe2 (7 and
11), and electron withdrawing X=F (8 and 12) and X=CN
(9 and 13) substituents (Figure 1).19 The decreased elec-
tron density of the NDI core was expected to increase
the potential of 9 as compared to that of 6 without dras-
tically compromising its capacity.
design and electronic tunability.8,16-18
For example,
compounds 3, 4 and 5 in Table 1 display two-electron
redox processes with potentials that cover a range from
0.8 V for dicarboxylate 4 and up to 2.3 V for anthraqui-
none 5 with a mid-point value of 1.7 V for pyromellitc
diimide 3. These variations demonstrate the relatively
large range available for this redox process and the po-
tential for fine-tuning and optimizing the observed po-
tential.
X
X
LiO
O
O
N
O
O
N
O
+ 2 e- + 2 Li+
- 2 e- - 2 Li+
R
N
O
N
R
R
R
OLi
X
X
In this paper, we use a model system, naphthalene
diimide (NDI; see Figure 1), to explore the possibility of
extending the range of physical properties and redox
parameters. We expect that an extended aromatic core
as compared to that of pyromellitic diimide 3 should
serve the dual purpose of decreasing the solubility of the
crystalline material while providing opportunities to
modify the LUMO levels of the parent hydrocarbon
structure 6 (Figure 1, X=H) with the help of synthetically
accessible substituents. Knowing that a highly desirable
low solubility translates into very challenging sample
tractability during synthesis and analysis, we decided to
analyze two sets of structures with (R=C6H13) and with-
out (R=H) an N-hexyl substituent. In this article we
report the synthesis and electrochemical evaluation of
compounds with a hydrocarbon aromatic core (X=H), as
X
R
Compound
H
6
Hexyl
Hexyl
Hexyl
Hexyl
H
NMe2
7
F
CN
H
8
9
10
11
12
13
NMe2
H
F
H
CN
H
Figure 1. Structure and redox process of naphththalene
diimides (NDIs) with different aromatic (X) and imide (R)
substituents. Although there is a potential four electron
mechanism resulting in the lithiation of the remaining car-
bonyls, attempts to reach sufficiently low potentials to ac-
cess this state resulted in irreversible decomposition.19
ACS Paragon Plus Environment
2

Page 3 of 8
Chemistry of Materials
1
2
Scheme 1
O
N
O
O
3
4
5
6
7
8
9
O
O
O
O
O
N(Me)2
O
a
O
N
Hex
N
O
Hex
d
e
f
Hex
N
O
Hex
O
14
6
O
N(Me)2
7
Br
F
Br
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
N
O
O
N
O
O
O
N
O
O
O
O
c
b
Hex
Hex
Hex
N
O
Hex
14
O
O
O
F
8
Br
Br
15
16
NC
O
O
Hex
N
O
N
O
Hex
CN
9
(a) Hexylamine, pyridine, 120⁰C, 83%; (b) Dibromoisocyanuric acid, sulfuric acid, 130⁰C; (c) Hexylamine, glacial acetic acid,
120⁰C, 40% over two steps; (d) Dimethylamine, THF, 80⁰C, 93%; (e) KF, 18-crown-6, 3 Å molecular sieves, sulfolane, 165⁰C, 53%;
(f) CuCN, NMP, 100⁰C, 87%.
tor. The cells were assembled in an Argon filled glove
box containing less than 1 ppm water and oxygen. Elec-
trochemical characterization was performed using a
Biologic VMP-3 system. Each cell was characterized by
cyclic voltammetry (CV) at multiple sweep rates and
galvanostatic cycling at a C/5 rate based on the theoreti-
cal two-electron capacity of the electroactive material.
EXPERIMENTAL
Materials Synthesis. All the naphthalene diimides stud-
ied in this work were synthesized by the procedures
indicated in Schemes 1 and 2. Their characterization
and purity were determined by standard spectroscopic
techniques (please see SI).
Electrode preparation. Electrodes were prepared by drop
casting a slurry of the electroactive material, acetylene
black (Alfa Aesar #39724) conductive additive, and poly-
vinylidene fluoride (PVDF) binder (Kynar PVDF) onto
pre-cut current collectors, which were allowed to dry
under ambient conditions. The slurries had a nominal
RESULTS AND DISCUSSION
LUMO Energies. As a starting point for this study, and
to guide our electrochemical measurements, we looked
for qualitative insight into the effects of substituents on
the aromatic core of the NDI units. Calculations were
carried out using gas phase electronic structure calcula-
tions with the B3LYP/6-311+G(d,p) level of theory for
both the N-hexyl and the N-H compounds in order to
establish the LUMO values versus their neutral states,
which are known to correlate directly to redox poten-
tials.19 As shown in Table 2, the LUMO levels of N-hexyl
compounds 6-9 span an energy range of 1.5 eV from the
more electron-rich dimethyl-amino substituted NDI 7,
with a value of -3.0 eV, to the electron-poor cyano sub-
stituted NDI 9, which has a value of -4.5 eV. The unsub-
composition
of
45:50:5
electroactive
materi-
al:carbon:PVDF, with the PVDF dissolved in N-methyl
pyrrolidone (NMP) prior to addition to the slurry.
Enough NMP (Sigma Aldrich #443778) was added to
form a paint-like consistency; nominally 800 μL NMP
per 100mg of dry material (inclusive of the NMP used to
dissolve the PVDF). A volume of 30μL of slurry was pi-
petted onto each 3/8” diameter carbon-coated alumi-
num current collector (Exopac Z-Flo 2650), resulting in
an average areal loading of 6 mg cm-2.
Electrochemical Characterization in Solution: The com-
pounds were cycled in 0.1 M tetrabutylammonium hex-
afluorophosphate in CH2Cl2. A platinum wire and plati-
num disk were used for working and counter electrodes,
respectively, and the working potential was measured
versus a Ag/AgCl reference electrode.20
stituted core of compound
6 and the fluorine-
substituted core of NDI 8 have similar values of -3.7 eV
and -3.9 eV, respectively. The larger variation observed
with respect to the parent compound in the case of 7
and 9 as compared to 8 confirm a greater effect for res-
onance effects relative to the inductive effects of the
electronegative fluorine atom. Calculations also showed
that removal of the N-hexyl groups in compounds 10-13
systematically shifts the corresponding energies by ca. -
0.2 to -0.3 eV, with an increase in the range of values up
to 1.6 eV between 11 (X=NMe2) and 13 (X=CN).
Electrochemical Characterization in the Solid State: Elec-
trodes were tested in a 2016 coin cell with 1M LiClO4 in
1:1 vol/vol ethylene carbonate : dimethyl carbonate
(EC:DMC) electrolyte, using lithium metal as counter
electrode and Celgard ##3401 separator. Minimum elec-
trolyte volume was used in order to moisten the separa-
ACS Paragon Plus Environment
3

Chemistry of Materials
Page 4 of 8
1
2
Table 2. Solid State Electrochemical Properties of Naphthalene Diimides 6-13.
3
4
5
6
7
8
9
NDI
X
R
MW
Theoretical
Capacity
Experimental Experimental
LUMO
[eV]b
1st Reducton
Potential
Capacity at
Capacity at
(2e) [mA h g- C/5 rate [mA h C/5 rate [# Li]
[V vs Li/Li+]c
1
]
g-1]a
92
6
7
H
NMe2
F
Hexyl
Hexyl
Hexyl
Hexyl
H
434.5
520.7
470.5
484.5
266.2
352.3
316.2
123
103
114
111
1.5
0.8
0.8
1.8
-3.7
-3.0
-3.9
-4.5
-4.0
-3.2
-4.8
2.5
2.3
41
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
8
45
2.6
2.8
2.55
2.4
2.9
9
CN
100
121
30
10
11
13
H
201
152
170
1.2
NMe2
H
0.4
CN
H
34
0.4
a A C/5 rate is the value needed for the current to reach maximum theoretical capacity to be delivered in 5 h. Gas phase values
obtained with the B3LYP/6-311+G(d,p) level of theory. cValues obtained using solid state NDI electrodes during the galvanostatic
cycling experiments.
b
Aryl-Substituted N-Hexyl Napthalene Diimides. An ex-
perimental evaluation of the effect of aryl substituents
on the reduction potential of the more soluble N-hexyl
substituted NDIs 6-9 was carried out both in solution
and in a coin cell format. The synthesis of these com-
pounds was accomplished as illustrated in Scheme 1
starting from naphththalene tetracarboxylic dianhydride
14. Compound 6 was obtained in a single step in 83%
yield by reaction of 14 with hexylamine. Compounds 7,
8 and 9 were obtained from the N-hexyl-dibromo-
naphthalenetetracarboxylic diimide 16, which was ob-
tained from readily available dibromo naphthalenetetra-
carboxylic dianhydride 15.
most electron withdrawing cyano-aryl substituted NDI
9, with least negative potential at the
E(1/2) V vs Ag/AgCl (vs Li/Li+) in Dichloromethane
X/X-
X-/X-2
LUMO (eV)
6
7
8
9
-0.64 (2.6)
-1.03 (2.21
-3.7
-3.0
-3.9
-4.5
-0.97 (2.27) -1.23 (2.01)
-0.55 (2.69) -0.98 (2.26)
-0.10 (3.14) -0.64 (2.6)
Compound 7 was isolated as a dark blue solid in 93%
yield via SNAr reaction of NDI 16 with dimethyl amine.
The fluorinated NDI 8 could be obtained via the halex
process21 in 53% yields, and the cyano derivative 9 via
copper-mediated coupling.22 All compounds were ob-
tained as solids and the purity was established by thin-
layer chromatography and NMR spectroscopy. Cyclic
voltammetric analysis of N-Hexyl NDIs 6-9 was carried
in out at ambient temperature at a 0.01 mM concentra-
tion of active material in 0.1 M tetrabutylammonium
hexafluorophosphate in CH2Cl2 (Figure 2). While the
result obtained in dilute solution are not representative
of the electrochemical behavior of the solid state mate-
rial in a typical lithium-ion cell, they provide valuable
information on the number of redox events and their
reversibility or lack thereof. As shown in Figure 2 and in
agreement with literature reports with other NDIs,21-23
each compound exhibits two redox waves characteristic
of the two-electron redox process indicated in Figure 1.
The first redox wave results in a charge delocalized radi-
cal anion, and the second then results in the previously
discussed dianion. It is important to point out that all
the samples could be reversibly cycled. The effect of the
substituent can be seen by the shift of the redox poten-
tials in the figure from the electron donating N,N-
dimethyl-aryl substituted NDI 7, shown at the bottom
with the most negative potential vs Ag/AgCl, to the
Figure 2. (a) Cyclic voltammograms of NDIs 6-9 measured
vs Ag/AgCl in 0.1 M Bu4NPF6 in CH2Cl2 (Li/Li+ is -3.237 vs
Ag/AgCl).
top. It is clear from the figure and Table 2 that the ef-
fects of substituents are greatest on the first reduction
potential as the difference between the first and second
reduction wave decreases as the aromatic substituents
change from electron withdrawing –CN (9) and –F (8),
to the reference compound -H (6) to the electron rich –
NMe2 (7). In the case of NDI 9, the first reduction of the
electron poor aromatic core occurs at a potential of -0.10
V vs Ag/AgCl and the second reduction wave occurs at a
ACS Paragon Plus Environment
4

Page 5 of 8
Chemistry of Materials
more negative potential of -0.64 V. By contrast, the
3h) with a high potential of 2.8 V vs. Li/Li+, a single flat
plateau, and an ability to reach close to 100% of its theo-
retical capacity (100 mAh g-1).
1
2
3
4
5
6
comparatively electron-rich aromatic core of NDI 7 un-
dergoes the first and second reduction waves much
closer together, at values of -0.97 V and -1.23 V, with a
difference between them that is about half as large as
that of the two reduction peaks in 9.
Having explored the effect of aryl substitution on the
redox potential of NDIs 7-9 in solution, we next ex-
plored their effect on the potential and capacity in a
coin cell format. The four N-Hexyl compounds were
characterised both by cyclic voltammetry and galvanos-
tatic cycling. As shown on the left column in Figure 3,
the CV curves obtained with the solid electrodes when
cycled between 1 and 4 V vs Li/Li+ revealed pronounced
differences in shape as compared to those obtained in
solution (Figure 2). Only NDI 6 exhibits two clear redox
peaks (Figure 3a) while the others have a relatively large
shift in the position of the peak and a significant broad-
ening, suggesting kinetic limitations associated with
electron and Li+ ion transport in the bulk solid material.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The galvanostatic cycling experiments measure the cell
potential during the time that it takes to cycle between
two potentials as
a function of a constant pre-
determined applied current. The current used in these
experiments (generally known as the C-rate) was based
on the theoretical capacity for a given sample being
reached in 5 h; a value commonly referred to as a C/5
cycling rate (Table 1). The galvanostatic data plotted in
the right column in Figure 3 includes representative
traces of up to ten cycles for each NDI. The current
flowing through the cell was used to calculate the num-
ber of Lithium ions per molecule inserted into the mate-
rial during the charging process (dotted lines), or the
number ions released during the discharge (solid lines).
On charging, the potential increases rapidly from 1 V to
a value that is characteristic of each solid NDI. The po-
tential remains at this value until either there is a new
redox event, or until the material reaches the limit of its
capacity at a set value of 4V. The discharge cycles (solid
lines) start at 4V, rapidly reach the corresponding
pleateau(s), and eventually decrease to 1 V. One can
appreciate from the galvanostatic cycling measurements
that the relationship between aryl substitution and re-
dox potential can be determined with more certainty
from the plateau value. First reduction potential values
of 2.3, 2.5, 2.6 and 2.8 V vs. Li/Li+ can be determined for
compounds 7, 6, 8 and 9, respectively. These values are
also included in Table 1. In agreement with the cyclic
voltammetry, only compound 6 (Figure 3e) displays two
clear plateaus and all others display only one, despite
the two-electron redox mechanism. The initial capaci-
ties achieved by compounds 6 (Figures 3e) and 9 (Fig-
ures 3h) are determined by the extent of the plateau,
and approach the theoretical maximum, with the inser-
tion of two Li+ per NDI. By comparison, compounds 7
(Figures 3f) and 8 (Figures 3g) saturate at about 0.75-1.0
Li+ per NDI. Another significant feature of the galvanos-
tatic cycling data is that the capacity for all N-hexyl
compounds decreases with increasing number of cycles,
likely due to the partial solubility in the electrolyte.
Overall, nitrile 9 has the best initial performance (Figure
Figure 3. Cyclic voltammetry and cyclability of a) com-
pound 6, b) compound 7, c) compound 8 and d) compound
9. The galvanostatic cycling with electrodes prepared with
e) compound 6, f) 8 and g) 9 was carried out between 1 and
4 V vs Li/Li+. The N,N-Dimethyl NDI 7 in f) was cycled
between 1 and 3.5 V in order to avoid an irreversible oxida-
tion above that value.
Aryl-Substituted N-H Naphthalene Diimides. In a sepa-
rate series of experiments, we examined the question of
whether the loss of capacity as a function of cycle num-
ber is determined by the solubility of the N-hexyl com-
pounds. For this study, we
ACS Paragon Plus Environment
5

Chemistry of Materials
Page 6 of 8
1
2
Scheme 2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
o
(a) Dialkoxybenzyl amine (DAB), acetic acid 120 C; (b)
o
o
CuCN, NMP 100 C, 70%; (c) Dimethylamine, THF, 80 C,
13% over 3 steps; (d) i. MeSO3H, PhMe, 110 C; ii. SOCl2,
o
DMF, 81%; (e) MeSO3H, PhMe, 110 oC, 79%.
decided to evaluate the properties of unsubstituted sec-
ondary (N-H) NDIs 10, 11 and 13, which are highly insol-
uble and nearly intractable, but can be obtained by in-
stalling a removable protecting group to assist with sol-
ubility during the synthesis and purification of all in-
termediates (Scheme 2). It should be noted that while
one may be concerned about the acidity of this proton,
previous reports show that for diimides, this proton re-
mains uninvolved.9,10 Compound 10 was prepared as
reported in the literature.24 The synthesis of NDIs 11
and 13 started with previously prepared dibromo substi-
Figure 4. Cyclic voltammetry and galvanostatic cyclability
of a) and d) compound 10, b) and e) compound 13, and c)
and f) compound 11.
Solid samples of compounds 10, 11 and 13 were charac-
terized by both cyclic voltammetry and galvanostatic
cycling as shown in Figure 4. As expected, removal of
the bulky N-Hexyl group reduces the solubility of the
corresponding N-H compound making their handling
difficult. The cyclic voltammograms are closely related
to those of their N-hexyl counterparts, although com-
pound 11 (Figure 4c) showed only minimal redox activi-
ty. A single very broad redox wave was observed for
compounds 10 (Figure 4a) and 13 (Figure 4b). Although
the theoretical capacity of compounds 10, 11 and 13 in-
creases in comparison to the N-hexyl series due to their
lower molecular weight, the values obtained from the
galvanostatic cycling experiments revealed significantly
lower Li+ insertion, with first cycle values of ~ 1.0 Li+ for
10 (Figure 4a), 0.75 Li+ for 13 (Figure e) and <0.5 for 11
(figure 4f). Despite the significantly lower solubility of
all three NDIs, only the unsubstituted derivative 10
showed better capacity retention during cycling. In
general, the N-H compounds have a lower utilization
than the N-Hexyl analogs, suggesting that the N-H
compounds may have tighter packing arrangements that
prevent the intercalation of the Li+ ions. Another pos-
sible explanation is that the N-Hexyl compounds may be
cycling in solution rather than in the solid state, but this
tuted
dianhydride
15
to
prepare
the
2,4-
dimethoxybenzyl protected NDI 16a by reaction with
2,4-dimethoxybenzyl amine, or the more soluble 16b, by
reaction with 2,4-di-n-butoxybenzyl amine (Scheme 2).
The dialkoxybenzyl substituted dibromo intermediates
were used to prepare the NDI 11 and 13 by SNAr reaction
with dimethylamine and Rosenmund von Braun reac-
tion with copper cyanide followed by acid catalyzed re-
moval of the 2,4-dialkoxybenzyl group. The use of a di-
n-butoxybenzyl protected nitrile 17 for the synthesis of
13 was dictated by the low solubility of the dimethoxy
analog, which had been suitable in the case of 18. The
required di-n-butoxy benzylamine was synthesized as
described in the SI section (Scheme SI 1). The desired
NDI 11 was obtained as a dark blue solid. It should be
noted that acid catalyzed deprotection of 17 was accom-
panied with hydrolysis. However, the resulting solid
could be easily dehydrated with thionyl chloride in di-
methylformamide to yield desired NDI 13 as a grey solid.
ACS Paragon Plus Environment
6

Page 7 of 8
Chemistry of Materials
triangles to heterocyclic anthraquinone analogs reported in
ref. 23.
is confounded by the poor capacity fade of the N-Hexyl
1
2
3
4
5
6
7
8
compounds and the superior cyclability of the highly
insoluble NDI 10. Figure 5 highlights the cyclability of
the three best performing NDIs. The two relatively sol-
uble N-hexyl NDIs 9 (CN) and 6 (H) have the highest
initial utilization, but their capacity decreases by ~ 30%
over 20 cycles. In contrast, the unsubstituted NDI 10
begins with a lower utilization value but is able to main-
tain ~ 90% after 20 cycles. While it is possible that con-
formational degrees of freedom provided by the hexyl
chains may help their crystal tolerate the structural
changes required for the NDI reduction and initial in-
sertion of the lithium ions, one could also conclude that
the higher solubility compromises the integrity of the
solid electrodes. Moreover, the much stronger lattice
energies of the un-substituted NDI 10 appear to reduce
the loss of material but also limit the amount of lithium
that can be inserted in the crystal lattice.
It is worth noting that the experimental reduction po-
tentials for the organic electrodes prepared with NDIs
bearing various aromatic substituents have a relatively
good correlation with the gas phase electronic structure
calculations computed at the B3LYP/6-311+G(d,p) level
of theory. This can be readily appreciated in the calcu-
lated LUMO vs. the solid state reduction potential
shown in Figure 6, which shows a linear trend. A simi-
lar correlation was recently reported by Liang et al.,16 in
a study of heterocyclic anthraquinone analogues, which
also displayed a linear correlation but with a somewhat
steeper slope. While two data sets are clearly too lim-
ited to draw conclusions, these results suggest that
structural factors within a given compound family may
dictate the relationship between the reduction potential
in the gas and that in the solid state.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
CONCLUSIONS
With a starting potential of 2.55 V vs Li/Li+ and 60%
theoretical capacity at C/5 rate, parent naphthalene
diimide 10 is both an intriguing material for lithium-ion
energy storage and an excellent model system for study-
ing the tailorability of organic electrode materials. The
use of N- and aryl-substitution detailed in this article
illustrates ways to control cycling performance and re-
dox potential by influencing the solubility and electron-
ic properties of the corresponding materials. The N-
hexyl- nitrile-substituted NDI 9 has the highest utiliza-
tion (90%) and second highest potential (2.8 V vs.
Li/Li+) of the compounds studied. This material main-
tains an appreciable capacity (100 mAh/g) despite a rela-
tively high molecular weight. Furthermore, aryl substi-
tution with CN-, F-, and Me2N- groups allows the redox
potential of the NDIs to be tailored between 2.9 and 2.3
V vs Li/Li+ without significantly affecting the solubility
or molecular weight. This could be a distinct advantage
over standard transition-metal based lithium-ion elec-
trode materials, whose potentials are largely confined to
the energy levels of the valence states of the metal. We
expect the interest in organic electrode materials such
as NDI to increase as their tailorability gains greater
recognition.
Figure 5. Discharge capacity vs. cycle number for three of
the compounds tested. Cycling data for all compounds
studied can be found in the supplementary information.
AUTHOR INFORMATION
Corresponding Author.
*Email: Bdunn@ucla.edu, mgg@chem.ucla.edu
ASSOCIATED CONTENT
Supporting Information. Synthetic details as well as
characterization information (1H, 13C and 19F NMR, IR,
and Mass Spectroscopy) can be found in associated
Supporting Information. This material is available free
of charge via the Internet at http://pubs.acs.org.
Figure 6. Solid state reduction potential vs. calculated gas-
phase B3LYP/6-311+G(d,p) LUMO energy levels. Numbered
solid squares correspond to the NDI studies here and open
ACS Paragon Plus Environment
7

Chemistry of Materials
Page 8 of 8
ACKNOWLEDGMENT
1
We thank Y. Cao and H. Pham for computational assis-
tance as well as E. Pierre and J. Fang for early electro-
chemical measurements. This work was supported by
NSF IGERT: Materials Creation Training Program
(MCTP) − DEG-0654431 and by grant NSF CHE-1266405
and by the Office of Naval Research (R.M. and B.D.).
This material is based upon work supported by the Na-
tional Science Foundation under equipment grant no.
CHE-1048804.
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
REFERENCES
1 Manthiram, A. J. Phys. Chem. Lett., 2011, 2, 176.
2 Goodenough, J.B.; Kim, Y. J. Power Sources, 2011, 196,
6688.
3 Poizot, P.; Dolhem, F. Energy Environ. Sci., 2011, 4, 2003.
4 B. Dunn , H. Kamath , J. M. Tarascon , Science 2011 , 334 ,
928.
5 Alt, H.; Binder, H.; Köhling, A.; Sandsteede, G. Electro-
chim. Acta., 1972, 17, 873.
6 Chen, H.; Armand, M.; Demailly, G.; Dolhem, F.; Poizot,
P.; Tarascon, J. M. ChemSusChem, 2008, 1, 348.
7 Liang, Y.; Tao, Z.; Chen, J. Adv. Energy Mater., 2012, 2, 742.
8 Walker, W.; Grudgeon, S.; Vezin, H.; Laruelle, S.; Armand,
M.; Wudl, F.; Tarascon, J. M. J. Mater.Chem., 2011, 21, 1615.
9 Renault, S.; Geng, J.; Dolhem, F.; Poizot, P. Chem. Com-
mun., 2011, 47, 2414.
10 Kim, D. J.; Je, S. H.; Sampath, S.; Choi, J. W.; Coskun, A.
RSC Adv., 2012, 2, 7968.
11 Walker, W.; Grugeon, S.; Mentre, O.; Laruelle, S.;
Tarascon, J. M.; Wudl, F. J. Am. Chem. Soc., 2010, 132, 6517.
12 Song, Z.; Zhou, H. Energy, Environ. Sci., 2013, 6, 2280.
13 Nakahara, K.; Iwasa, S.; Satoh, M.; Morioka, Y.; Iriyama, J.;
Suguro, M.; Hasegawa, E. Chem. Phys. Lett., 2002, 359, 351.
14 Sarukawa, T.; Yamaguchi, S.; Oyama, N. J. Electro.Soc.
2010, 12 196.
15 Inamasu, T.; Yoshitoku, D.; Sumi-otorii, Y.; Tani, H.; Ono,
N. J. Electrochem. Soc., 2003, 150, A128.
16 Liang, Y.; Zhang, P.; Yang, S.; Tao, Z.; Chen, J. Adv. Energy,
Mater., 2013, 3, 600.
17 Song, Z.; Zhan, H.; Zhou, Y. Chem. Comm., 2009, 448.
18 Armand, M.; Grudgeon, S.; Vezin, H.; Laruelle, S. Ribiére,
P.; Poizot, P.; Tarascon, J.M. Nature Mat., 2009, 8, 120.
19 Song, Z.; Zhan, H.; Zhou, Y. Angew. Chem. Int. Ed., 2010,
49, 8444.
20 Only the N-Hexyl compounds were tested in solution,
due to the lack of solubility of the X=H compounds in di-
chloromethane.
21 Würthner, F.; Osswald, P.; Schmidt, R.; Kaiser, T.;
Mansikkamaki, H.; Konemann, M. Org. Lett., 2006, 8, 3765.
22 Dawson, R.; Hennig, A.; Weimann, D.; Emery, D.; Ra-
vikumar, V.; Montenegro, J.; Takeuchi, T.; Gabutti, S.; Mayor,
M.; Mareda, J.; Schalley, C.; Matile, S. Nature Chem., 2010, 2,
533.
23 Röger, C. and Wurthner, F. J. Org. Chem., 2007, 72, 8070.
24 Sotiriou-Leventis, C.; Mao, Z. J. Heterocyclic Chem., 2000,
37, 1665.
ACS Paragon Plus Environment
8