Melting Points of Potential Liquid Organic Hydrogen Carrier Systems Consisting of N-Alkylcarbazoles

Article abstract of DOI:10.1021/acs.jced.5b00679

Liquid organic hydrogen carriers (LOHCs) represent an attractive concept for storing hydrogen by the hydrogenation of usually aromatic compounds. One of the best investigated LOHCs is N-ethylcarbazole because of its favorable thermodynamic properties. However, its high melting point of 343.1 K could be a major drawback particularly in mobile applications. Therefore, it is desired to decrease the melting point of N-ethylcarbazole without significantly changing favorable properties such as the storage density or the reaction behavior of the carrier compound. To investigate the solid-liquid behavior during hydrogenation, the melting points of pure N-ethylcarbazole derivatives with increasing degree of hydrogenation as well as the liquidus line of the binary mixture of N-ethylcarbazole and N-ethyl-dodecahydro-carbazole were measured. Because of their structural and chemical resemblance binary mixtures consisting of different alkylcarbazole combinations were analyzed regarding their potential for a melting point depression. By the appropriate combination of N-alkylcarbazoles, it is possible to achieve a considerable melting point decrease to 297.1 K.

Full text of DOI:10.1021/acs.jced.5b00679

Article
pubs.acs.org/jced
Melting Points of Potential Liquid Organic Hydrogen Carrier Systems
Consisting of N‑Alkylcarbazoles
Katharina Stark,† Philipp Keil,† Sebastian Schug,† Karsten Muller,*,† Peter Wasserscheid,‡
̈
and Wolfgang Arlt†
‡
†Institute of Separation Science and Technology, and Institute of Chemical Reaction Engineering, Friedrich-Alexander-Universitat
̈
Erlangen-Nurnberg, Egerlandstraße 3, 91058 Erlangen, Germany
̈
S
* Supporting Information
ABSTRACT: Liquid organic hydrogen carriers (LOHCs) represent an
attractive concept for storing hydrogen by the hydrogenation of usually
aromatic compounds. One of the best investigated LOHCs is N-ethyl-
carbazole because of its favorable thermodynamic properties. However, its
high melting point of 343.1 K could be a major drawback particularly in
mobile applications. Therefore, it is desired to decrease the melting point of
N-ethylcarbazole without significantly changing favorable properties such as
the storage density or the reaction behavior of the carrier compound. To
investigate the solid−liquid behavior during hydrogenation, the melting points
of pure N-ethylcarbazole derivatives with increasing degree of hydrogenation
as well as the liquidus line of the binary mixture of N-ethylcarbazole and N-
ethyl-dodecahydro-carbazole were measured. Because of their structural and
chemical resemblance binary mixtures consisting of different alkylcarbazole
combinations were analyzed regarding their potential for a melting point
depression. By the appropriate combination of N-alkylcarbazoles, it is possible to achieve a considerable melting point decrease to
297.1 K.
Therefore, it was the objective of this work to examine
different options for decreasing the melting point of N-
ethylcarbazole to melting points at least below ambient
temperature.
Pure N-ethylcarbazole is considered to be the highest melting
compound in the LOHC system. During the hydrogenation of
N-ethylcarbazole a multinary mixture containing different partly
hydrogenated derivatives is formed (refer to Figure 1 for the
1. INTRODUCTION
Liquid organic hydrogen carriers (LOHCs) are novel media for
the storage and transport of energy in the form of hydrogen.
The concept is based on a reversible catalytic hydrogenation
reaction of organic compounds that offer a high hydrogen
storage capacity. Among others, N-ethylcarbazole (0H-NEC) is
a very promising candidate which was first proposed as a
potential LOHC by Pez et al. in 2004.1 Since then N-
ethylcarbazole has become one of the most intensively
investigated LOHCs.2−11 N-ethylcarbazole stores 12 atoms of
hydrogen in the form of its fully hydrogenated derivative N-
ethyl-dodecahydro-carbazole.12 This results in a high storage
density of 5.8 wt %. Moreover also several other properties of
N-ethylcarbazole are beneficial for its use as LOHC, such as
nontoxicity, low vapor pressure, low dehydrogenation temper-
atures due to favorable thermodynamics, and a comparatively
high rate of reaction for the dehydrogenation of the
hydrogenated form.
Many authors consider the high melting point of N-
ethylcarbazole of 343.1 K13 as the major drawback for its use
as LOHC in stationary and particularly mobile applica-
tions.14−16 To ensure that the compound does not solidify
during storage or transport, it would be necessary to heat all
process equipment which is in contact with LOHC constantly
to a temperature above 343.1 K. This results in an additional
energy demand as well as higher capital expenditure.9
Figure 1. Partly and fully hydrogenated derivatives of N-ethyl-
carbazole.4−6
Received: August 9, 2015
Accepted: March 3, 2016
Published: March 17, 2016
© 2016 American Chemical Society
1441
DOI: 10.1021/acs.jced.5b00679
J. Chem. Eng. Data 2016, 61, 1441−1448

Journal of Chemical & Engineering Data
Article
Table 1. Investigated N-Alkylcarbazoles Purified by Means of Crystallization and Analyzed by GC−MS
chemical name
carbazole
abbreviation
source
Merck KGaA
CAS
final mole fraction purity
C
86-74-8
0.995
0.995
0.995
0.999
0.995
0.999
0.995
tetrahydrocarbazole
N-methylcarbazole
N-ethylcarbazole
4H-C
NMC
(0H)-NEC
NPC
Alfa Aesar
942-01-8
1484-12-4
86-28-2
Sigma-Aldrich
Sigma-Aldrich
N-propylcarbazole
N-isopropylcarbazole
N-butylcarbazole
Akos Cons. & Solutions GmbH
Apollo Scientific
1484-10-2
1484-09-9
1484-08-8
NiPC
NBC
Akos Cons. & Solutions GmbH
Table 2. Investigated Partly and Fully Hydrogenated Derivatives of N-Ethylcarbazole and Carbazole Purified by Means of
Fractional Batch Distillation and Analyzed by GC−MS. Source of Components: Institute of Reaction Engineering, University of
Erlangen-Nuremberg
most important intermediates). In the following those
derivatives are abbreviated as xH-NEC, where x denotes the
number of hydrogen atoms reversibly bound to the N-
ethylcarbazole molecule (or just xH-C, if the dehydrogenated
form is carbazole). It was reported that the resulting multinary
mixture of different partly hydrogenated derivatives leads to a
melting point depression.15
Thus, incomplete dehydrogenation might be an option for a
melting point decrease. However, to the best of our knowledge
no quantitative data on the melting points of the partly
hydrogenated N-ethylcarbazoles N-ethyl-tetrahydro-carbazole
(4H-NEC), N-ethyl-hexahydro-carbazole (6H-NEC), and N-
ethyl-octahydro-carbazole (8H-NEC) are available in the
literature. To close this gap the melting points of pure
intermediates were measured in this work.
Another approach for decreasing the melting point is based
on the substitution of N-ethylcarbazole by another alkylcarba-
zole derivative with a lower melting point. In that way
comparable hydrogenation reaction properties can be
achieved.17 The melting point of the homologous series of N-
alkylcarbazoles strongly depends on the length of the alkyl
chain.6 It is known from literature that the melting point
decreases from the nonalkylated carbazole (Tm = 518.3 K18) to
N-propylcarbazole (Tm = 322.2 K19). N-Isopropycarbazole with
a branched alkyl chain shows a considerably higher melting
point (Tm = 395.2 K20) than its linear isomer. By a further
increase of the chain length the melting point starts to increase
again which is indicated by the melting point of N-
butylcarbazole (Tm = 331.0 K19). Because the availability of
melting point data especially for N-propylcarbazole and N-
butylcarbazole in the literature is rather poor, additional
property data are reported in this work.
The solid−liquid behavior of the multinary hydrogenation
mixtures is very complex due to the number of involved
components, and the resulting complex crystallization behavior.
For simplification reasons, the liquidus curve of the binary
system N-ethylcarbazole and its perhydrogenated derivative N-
ethyl-dodecahydro-carbazole was measured in the first place.
Combining different N-alkylcarbazoles might be also an
attractive option for achieving a melting point depression in the
case of eutectic solid−liquid phase behavior. Owing to the lack
of literature data, melting enthalpies of pure N-alkylcarbazoles
1442
DOI: 10.1021/acs.jced.5b00679
J. Chem. Eng. Data 2016, 61, 1441−1448

Journal of Chemical & Engineering Data
Article
as well as their binary solid−liquid phase behavior have been
determined in this work.
solution was obtained and filled as a liquid into aluminum
crucibles.
Reaction Intermediates. Crucibles containing N-ethyl-
octahydro-carbazole or tetrahydro-carbazole which were sealed
under air atmosphere did not show reproducible experimental
results. The enthalpy of fusion decreased with repeated
measurements. Additionally an exothermic effect was observed
directly after the solid−liquid transition. Therefore, it was
assumed that the samples showed degradation during DSC
analysis because of a reaction of N-ethyl-octahydro-carbazole or
tetrahydro-carbazole, respectively, with the oxygen present in
air in the crucibles.
Crucibles containing N-ethyl-octahydro-carbazole or tetrahy-
dro-carbazole, respectively, which were sealed in a glovebox
(oxygen, ≤20 ppm; water, ≤0.5 ppm) to ensure the absence of
oxygen, did not show this exothermic effect but gave
reproducible results.
Samples containing N-ethyl-tetrahydro-carbazole had to be
stored in a refrigerator for several days until they were fully
crystallized because of the very slow crystallization kinetics of
the compound.
Experimental Procedure. The DSC measurements were
performed using a Maia 200 differential scanning calorimeter
(DSC) from Netzsch. The temperature range and program of
the measurements was adapted depending on the melting
points of the components.
The DSC was calibrated with respect to enthalpy of fusion
and melting temperature using a Netzsch certified calibration
set containing the six compounds indium, adamantane, tin,
bismuth, zinc, and cesium chloride at very high purity 99.9% to
99.999% with melting point temperatures in the range of 218.7
K to 749.2 K and enthalpies of fusions in the range of 17.2 J/g
to 107.5 J/g. Additionally the certified calibration compound
biphenyl with a purity of 99.5%, a melting point of 342.1 K, and
enthalpy of fusion of −120.4 J/g, was used for calibration to
incorporate an organic compound with structural similarity to
the N-alkylcarbazoles. After calibration, the melting point of
indium was determined with a standard uncertainty of 0.1 K,
and the standard uncertainty of enthalpy of fusion was
determined to be 0.1 J/g.
The measurements have been evaluated using the software
Proteus 5.2 (Netzsch). For pure substances the onset point of
the peak represents the start of the solid−liquid phase change
which indicates the melting point. If the binary mixture
demonstrates eutectic behavior, two peaks are formed. The
onset of the first peak indicates the eutectic temperature while
the peak point of the second peak represents the melting
temperature owing to the dissolution of the last solid. In case of
eutectic composition of the sample, only one peak is formed.22
2.3. Static Measurements. Static measurements have been
performed for the determination of the liquidus curve for the
binary mixture of N-ethylcarbazole and N-ethyl-dodecahydro-
carbazole. DSC measurements were found to be not suitable for
mixtures of these two substances because of the slow
crystallization kinetics. Therefore, both substances were put
into a 100 mL stirred glass vessel in defined shares and heated
to the desired temperature so that both phases, solid and liquid,
coexisted. The total mass during the experiments was
approximately 50−100 g of substance depending on the
desired concentration. Then both substances were stirred
intensively at defined, constant temperatures controlled by a
LAUDA PROLINE RP845 thermostat until equilibrium was
reached. After that the stirrer was switched off and a liquid
2. EXPERIMENTAL SECTION
2.1. Materials. Table 1 shows the N-alkylcarbazoles
investigated in this contribution. The purity was confirmed
by gas chromatography−mass spectroscopy (GC−MS), and if
required, the purchased substances were further purified by
recrystallization from cyclohexane or acetone to obtain a final
purity of ≥99.5% (GC−MS).
Table 2 shows the partly and fully hydrogenated derivatives
of N-etylcarbazole and carbazole investigated in this paper. A
reaction mixture containing different hydrogenation products
was prepared by hydrogenation of N-ethylcarbazole (0H-NEC)
at 150 °C and 40 bar hydrogen pressure in a stainless steel
autoclave equipped with a gas entry stirrer. Ru (5 wt %) on
alumina support was used as catalyst. The reaction product was
filtered through a frit at the bottom of the autoclave.13
The different intermediates as well as the fully hydrogenated
12H-NEC were separated from the reaction mixture by
fractional batch distillation and analyzed by means of NMR
and GC−MS. 0H-NEC, 4H-NEC, 6H-NEC, 8H-NEC and at
least three diastereomers of 12H-NEC were identified.21 12H-
NEC was separated from the mixture by distillation with a high
purity of >99.9%. On the basis of the NMR measurements it is
assumed that the asymmetric dodecahydro-N-ethylcarbazole
(see Table 2) was obtained, which has already been described
by Eblagon4−6 to be the most stable one. The purity of the
partly hydrogenated substances 4H-NEC and 8H-NEC which
were also separated from the reaction mixture by distillation
was determined to be >99.8% (GC−MS). Because of 6H-NEC
being formed as a stable intermediate during (de)-
hydrogenation to a far lesser extent than 4H-NEC and 8H-
NEC, it could not be separated in a sufficient amount and
purity. A diastereomeric mixture of 12H-C was prepared by
catalytic hydrogenation of 4H-C analogous to the procedure
described above. 12H-C was recrystallized from a mixture of
acetone and toluene (50/50 wt %) to obtain a final purity
>99.0%. The diasteromeric mixture was not further separated.
The partly and fully hydrogenated products were stored at
temperatures below 253 K under argon atmosphere to prevent
degradation.
2.2. DSC Measurements. Sample Preparation. N-
Alkylcarbazoles. For the pure component measurements, 3−
8 mg of sample in solid or liquid state was sealed in an
aluminum crucible. At least three crucibles were prepared for
each substance.
The different binary mixtures of two N-alkylcarbazoles were
prepared by weighing at least 0.1 g of each solid component
into a glass vial depending on the desired molar ratio.
Subsequently the substances were melted by heating them
and then mixed by shaking the vial until a homogeneous
solution was obtained. After that the samples were stored in a
refrigerator at 253 K, ensuring a quick solidification of the
mixture. Briefly before measurement the solid samples were
ground at room temperature to homogenize the mixture.
Samples containing carbazole (nonalkylated) were not melted
before grinding up to avoid degradation because of the high
melting point of carbazole (Tm = 518 K).18 Consequently they
were not stored in a refrigerator. For each DSC measurement
3−8 mg of sample was sealed in an aluminum crucible. Liquid
samples were mixed by shaking the vial until a homogeneous
1443
DOI: 10.1021/acs.jced.5b00679
J. Chem. Eng. Data 2016, 61, 1441−1448

Journal of Chemical & Engineering Data
Article
sample was withdrawn using a syringe. The temperature of the
vessel was measured with a calibrated Pt-100 placed in the
middle of the liquid phase during sample taking. To minimize
the amount of solid particles in the withdrawn sample, a PTFE
syringe filter with a pore size of 5.0 μm (Rotilabo) was used.
The concentration of the samples was determined quantita-
tively by GC (FID)-analysis. Sampling for each temperature
was performed at least two times.
decreases to 349.5 K. The higher experimental error of 12H-C
in comparison to the other substances can be explained by the
fact that 12H-C is not a pure compound but a diastereomeric
mixture.
3.1.2. Variation of the Alkyl Chain. The melting points of
the N-alkylcarbazoles with linear alkyl chain decrease with
increasing chain length up to N-propylcarbazole (see Table 3).
The higher melting point of N-butylcarbazole as compared to
N-propylcarbazole indicates that the melting point increases
with a further increase of the chain length. N-Iso-
propylcarbazole with a branched alkyl chain shows a higher
melting point compared to the linear N-propylated derivative,
which might be because of its rather symmetric structure. The
measured melting temperatures are in good agreement with
literature data.
The enthalpy of fusion of carbazole (27.70 kJ/mol) is higher
than for N-methylcarbazole (16.43 kJ/mol) or N-ethylcarbazole
(16.55 kJ/mol). This effect propably occurs because of the
larger influence of the N atom inducing a slight polarity in
carbazole, which is weaker in N-methylcarbazole and N-
ethylcarbazole. With a further increase of the length of the side
chain, the enthalpy of fusion increases again but still does not
reach the value of carbazole. The measured enthalpies of fusion
are in good agreement with literature data.
3. RESULTS AND DISCUSSION
3.1. Pure Component Data. 3.1.1. Melting Point
Behavior during (De)hydrogenation. Table 3 shows the
Table 3. Melting Points Tm,0i and Enthalpies of Fusion
Δfush0i (at Tm,0i) of Different N-Alkylcarbazoles and Their
a
(Partly) Hydrogenated Derivatives at Pressure p = 0.1 MPa
substance
carbazole (C)
Tm,0i/K
Δ
fush0i/kJ mol−1
source
517.1 0.2
518.33 0.01
521.2
27.70 0.31
27.08 0.48
27.2
this work
18
23
518.7
26.9
24
N-methylcarbazole (NMC)
N-ethylcarbazole (NEC)
N-propylcarbazole (NPC)
361.0 0.3
362.5
16.43 0.22
17.15
this work
25
342.4 0.1
343.1
16.55 0.17
15.10 0.40
19.57 0.13
this work
13
3.2. Binary Mixtures. 3.2.1. Modeling of the Phase
Behavior. Binary solid−liquid equilibria (SLE) were modeled
using eq 1.
320.4 0.2
322.2 4.0
393.9 0.3
this work
19
fi0,L
fi0,S
L
L
⎛
⎝
⎞
⎠
⎡
⎢
⎤
⎥
xi ·γ
N-iso-propylcarbazole
(NiPC)
18.26 0.33
this work
Δfush0i
RT
T
Tm,0i
i
⎜
⎜
⎟
⎟
ln
= ln
= −
1 −
xS·γS
⎢
⎣
⎥
⎦
i
i
395.2
17.73
20
N-butylcarbazole (NBC)
330.6 0.1
331.0 4.0
282.4 0.1
22.80 0.16
this work
19
Δcp
R
⎡
⎤
Tm,0i
T
Tm,0i
−
1 −
− ln
⎢
⎣
⎥
⎦
T
(1)
N-ethyl-tetrahydro-
carbazole
12.41 0.04
20.58 0.19
this work
Equation 1 describes the ratio of standard fugacity f0i of the
liquid phase L and solid phase S with the measurable
thermodynamic variables of the pure component enthalpy of
fusion Δfush0i, the melting temperature Tm*,0i and the difference
in heat capacities Δcp of solid and liquid state. Further xi is the
molar fraction, R is the ideal gas constant, and γφi is the activity
coefficient in the respective phase φ.
N-ethyl-octahydro-carbazole 316.0 0.1
this work
this work
N-ethyl-dodecahydro-
carbazole
188.6 (TG)
tetrahydro-carbazole
391.3 0.1
349.5 0.4
17.85 0.06
22.91 0.72
this work
this work
dodecahydro-carbazole
a
The uncertainties of the melting point and enthalpy of fusion
correspond to expanded uncertainties of the mean (0.95 confidence
level). Standard uncertainties u are u(T) = 0.1 K, u(p) = 5 kPa.
Equation 1 can be simplified using the following
assumptions: (i) ideal eutectic behavior with crystallization of
pure solid phase i: xiS = 1; γSi = 1; (ii) difference in heat
capacities between solid and liquid phase is small: Δcp ≈ 0.
With these assumptions eq 2 is obtained:
pure component data for the N-ethylcarbazole and carbazole
derivatives with different degrees of hydrogenation which were
obtained by DSC measurements and a comparison with
literature values.
The melting points of the N-ethylcarbazole derivatives do
not show a clear trend with increasing degree of hydrogenation,
as the melting point of 8H-NEC is higher than that of 4H-
NEC. However, both components show a lower melting point
than 0H-NEC.
For 12H-NEC only a glass transition was observed during
DSC measurements. The substance did not crystallize even
after storing it in a glass vial at temperatures less than 245 K for
more than a year. Subsequently also no enthalpy of fusion was
obtained.
The change in melting point in dependence of the degree of
hydrogenation was compared to carbazole derivatives which
constitute the main impurity in N-ethylcarbazole reaction
mixtures. By the addition of four hydrogen atoms the melting
point of carbazole with 517.1 K drops to 391.3 K for 4H-C. By
the addition of 12 hydrogen atoms the melting point further
⎡
⎢
⎤
⎥
Δfush0i
RT
T
Tm,0i
ln(xiL·γL) = −
1 −
i
⎢
⎣
⎥
⎦
(2)
In the case that the liquid phase behaves as an ideal mixture, γiL
= 1 holds and eq 3 is obtained which represents the most
simplified expression for describing the solubility of a
component i.
⎡
⎢
⎤
⎥
Δfushi
RT
T
Tm,i
ln(xiL) = −
1 −
⎢
⎣
⎥
⎦
(3)
Only the pure component properties, enthalpy of fusion and
melting point of the involved components, are required for
modeling the liquidus line of an ideal eutectic mixture.25,26
3.2.2. Binary Mixtures of N-Ethylcarbazole Derivatives
with Different Degree of Hydrogenation. During hydro-
1444
DOI: 10.1021/acs.jced.5b00679
J. Chem. Eng. Data 2016, 61, 1441−1448

Journal of Chemical & Engineering Data
Article
genation the melting point of the reaction mixture decreases.15
The solid−liquid behavior of these multinary hydrogenation
mixtures is very complex. To get a basic understanding of the
phase behavior, the liquidus lines for the binary mixture N-
ethylcarbazole (NEC) and N-ethyl-dodecahydro-carbazole
(12H-NEC) was measured (see Figure 2).
known NRTL-model26 (eqs 6−11). The parameters are
summarized in Table 4.
⎛
⎝
⎞
⎠
∑j xjτjiGji
∑k xkGki
xjGij
∑m xmτmjGmj
∑k xkGkj
ln γL =
+
τ −
⎜
⎜
⎟
⎟
∑
ij
i
∑ xkGkj
k
k
(6)
(7)
Gij = exp(−αijτij)
b
ij
τ = aij +
ij
(8)
(9)
T
αij = cij + dij(T − 273.15 K)
τ = 0
(10)
(11)
ii
Gii = 1
Table 4. NRTL Regression Parameters for the Binary
Mixture 0H-NEC−12H-NEC at Pressure p = 0.1 MPa
NRTL T/K
i = 0H-NEC, j = 12H-NEC
Figure 2. Solid−liquid phase diagram of the system N-ethylcarbazole
(NEC) and N-ethyl-dodecahydro-carbazole (12H-NEC). Experimen-
tal data (symbols), calculated liquidus lines assuming ideal phase
behavior (ideal), and liquidus lines calculated for real phase behavior
with activity coefficients obtained by fitting the experimental data to
the NRTL-model (NRTL) and with activity coefficients predicted by
COSMO-RS (COSMO-RS).
aij
aji
bij
bji
cij
0
0
−3274.28415
4100.5343
9.87 × 10−03
0
dij
For an a priori prediction of liquid phase activity coefficients γiL
in order to determine the solid−liquid phase behavior,
COSMO-RS (COSMOthermX, C21_0111) was used.27 The
COSMO-RS calculations predict activity coefficients greater
than 1. The trend of the liquidus line was predicted correctly,
resulting in lower values for σ (0.05) and δ (22.9%). However,
the absolute values of the activity coefficients are still
underestimated.
As described previously, the solid−liquid equilibrium of the
system was measured by using a static method. With increasing
molar fraction of N-ethyl-dodecahydro-carbazole the melting
point decreases. At 298.6 K, a solid−liquid phase separation
experiment was performed to determine the composition of the
formed solid phase (indicated by arrows in Figure 2). An
aliquot of the well-mixed suspension was extracted with a
syringe. The two phases were separated in a filter centrifuge
and the crystals were washed twice with cyclohexane to remove
adhering liquid phase. The concentration of N-ethyl-dodeca-
hydro-carbazole in the solid phase after washing was
determined to be 0.6% while the concentration of the
coexisting liquid phase lies on the liquidus line. Therefore, it
can be stated that rather pure N-ethylcarbazole is crystallized as
solid phase from the binary mixture.
3.2.3. Binary Mixtures of Different N-Alkylcarbazoles.
Figure 3 shows the binary SLE of six different N-alkylcarbazole
combinations.
The eutectic temperatures for the six binary solid−liquid
equilibria are summarized in Table 5. The binary mixtures
NEC−C, NEC−NMC, and NEC−NPC show only a slight
positive deviation of the ideal calculated liquidus curve. The
minor differences between ideal calculated and experimental
obtained values are within the uncertainty of measurement.
The binary SLE of NPC−NBC also demonstrates eutectic
behavior with a further solid−solid transition at T = 284 K for
compositions with xNEC > 0.5, that is, on the N-ethylcarbazole-
rich side of the phase diagram. A similar phase behavior has
been observed in a recent study on the solid−liquid equilibria
of triolein with fatty acids.28
The deviation between the experimental data xi,exp and the
calculated points on the ideal liquidus line xi,calc was evaluated
quantitatively with the standard deviation σ and the relative
error δ (see eqs 4 and 5).
n
2
∑
i=1 (xi,exp − xi,calc
)
σ =
(n − 1)
(4)
The system NEC and NiPC shows a strong deviation from
the ideal calculated liquidus curve. Up to a concentration of
NEC = 0.5, a phase transition at a constant temperature of T =
xi,exp − xi,calc
n
i=1
∑
xi,exp
x
δ =
(5)
n
367.7 K is found, while for higher concentrations, a phase
transition at a constant temperature of T = 336.9 K is found.
The data of the liquidus line indicate a discontinuity in the
region around xNEC = 0.55 which can be interpreted to be a
peritectic point. On the basis of the results obtained in this
work, a peritectic phase behavior with incongruent melting
behavior is proposed. In region (I) the melt and solid N-
isopropyl-carbazole are in equilibrium. In region (II), the melt
With increasing amount of 12H-NEC the measured values
show an increasing positive deviation from the ideal predicted
liquidus line which is the reason for the resulting values of σ =
0.2 and δ = 98.6% being quite high.
Owing to the differences between experimental and ideal
calculated values, activity coefficients greater than 1 result. The
thus obtained activity coefficients were fitted using the well-
1445
DOI: 10.1021/acs.jced.5b00679
J. Chem. Eng. Data 2016, 61, 1441−1448

Journal of Chemical & Engineering Data
Article
□
●
Figure 3. Binary phase diagrams of six combinations of different N-alkylcarbazoles: , Teu = eutectic temperature; , Tfus = temperature of fusion;
▼
, Ttr = temperature of transition; ---, ideal calculated liquidus lines (eq 3).
Table 5. Ideal Calculated Eutectic Temperature Teu,id and Eutectic Temperature Teu,exp of the Binary SLEs at the Eutectic
Concentration xNEC,eu (for Abbreviations of the Compounds See Table 3)
mixture
xNEC,eu
Teu,id/K
Teu,exp/K
σ/K
δ/%
hydrogen storage capacity at xNEC,eu/wt %
NEC−C
0.96
0.57
0.42
0.72
0.49
0.59
340.3
312.5
298.2
323.1
305.5
298.7
340.2
314.9
297.1
336.9
303.8
297.7
3.9
1.0
0.76
0.25
0.44
7.10
0.16
0.43
5.8
6.0
5.6
5.7
5.4
5.3
NEC−NMC
NEC−NPC
NEC−NiPC
NEC−NBC
NPC−NBC
1.7
30.7
0.71
1.8
a
Standard uncertainties u are u(Tm) = 1 K, u(Teu) = 2 K, u(x) = 0.001, u(p) = 5 kPa.
and a solid 1:1-mixture of the components coexist, and in
region (III) the melt and solid N-ethylcarbazole coexist. The
eutectic point is around xNEC = 0.95; however, this work
focused on the investigation of the liquidus curve and hence the
solid phenomena were not further investigated.
By enlarging the chain length of the alkyl group the
gravimetric hydrogen storage capacity decreases. However,
using a NEC−NPC mixture at eutectic composition, the
hydrogen capacity is still 5.6 wt % while the melting point is
reduced to 298 K compared to 5.8 wt % and 342 K for pure
NEC.
Multinary mixtures, derived by mixing more than two N-
alkylcarbazoles, might offer an even lower melting point.
Because of the low eutectic temperatures measured for the
binary systems NEC−NPC, NEC−NBC, and NPC−NBC, it
seems reasonable to investigate the solid−liquid phase behavior
of a ternary mixture consisting of these three substances.
Therefore, the ternary eutectic temperature was calculated
using the simplified Schroeder-van-Laar equation for ideal
eutectic systems assuming that like the binary subsystems, also
the ternary system shows ideal liquid phase behavior (see
Figure 4).
The deviation between the experimental data and the
calculated points on the ideal calculated liquidus line Ti,exp
−
Ti,calc was evaluated quantitatively with the standard deviation σ
and the relative error δ.
n
i=1
2
∑
(Ti,exp − T
)
i,calc
σ =
(n − 1)
(12)
(13)
T
i,exp − T
n
i=1
i,calc
∑
T
i,exp
δ =
n
Comparing the eutectic temperatures of the analyzed
mixtures it can be concluded that the lowest melting
temperatures in binary systems are obtained using the mixtures
NEC−NPC, NEC−NBC, and NPC−NBC which have eutectic
points around room temperature. Therefore, a significant
melting point decrease can be achieved compared to pure N-
ethylcarbazole.
A ternary eutectic temperature of T = 285.7 K was calculated
for the mixture NEC−NPC−NBC while still offering a storage
capacity of 5.5 wt % which is 95% of the capacity of pure NEC.
The ideal calculations show that multinary mixtures containing
three or more different alkylcarbazoles offer a simple solution
with a very attractive potential for further decreasing the
1446
DOI: 10.1021/acs.jced.5b00679
J. Chem. Eng. Data 2016, 61, 1441−1448

Journal of Chemical & Engineering Data
Article
Moreover, none of the investigated derivatives fulfill the
objective of a melting point below ambient temperature.29
Therefore, binary mixtures of N-alkylcarbazoles were
investigated regarding their melting point behavior at different
concentrations, preferably in the case of eutectic phase
behavior. Ideal eutectic behavior was observed for the binary
mixtures NEC−C, NEC−NMC, NEC−NPC, and NEC−NBC.
For the binary mixture NEC−NiPC a nonideal phase behavior
was observed. On the basis of the experimental results, the
system is proposed to be peritectic. This work shows that
binary mixtures of N-alkylcarbazoles offer an attractive option
to achieve a significant melting point decrease while keeping a
high storage capacity. The mixture NEC−NPC at eutectic
composition offers a hydrogen storage capacity of 5.6 wt % and
is liquid down to Teu = 297.1 K.
Multinary mixtures might be an option for a further melting
point decrease. The prediction of the ternary eutectic point for
the mixture NEC−NPC−NBC indicates a melting point of
285.7 K while still reaching a storage capacity of 5.5 wt %.
Mixtures of different N-alkylcarbazoles hence are a very
attractive option to provide a LOHC system which combines a
high effective hydrogen storage density and melting points
which are over the whole degree of hydrogenation significantly
below ambient temperature. Consequently the focus of further
investigations should lie on the reaction and property behavior
of N-alkylcarbazole mixtures during hydrogenation and
dehydrogenation reaction in order to model the time-
dependent occurring equilibria in the system as well as the
whole LOHC-process and finally ensure a reliable process
control. Besides the melting point future works should also
investigate other important substance properties like the
viscosity of these multicomponent mixtures.
Figure 4. Ternary SLE diagram of the system NEC−NPC−NBC
calculated assuming ideal liquid phase behavior.
melting point of a LOHC system to temperatures below
ambient.
Because of the very slow crystallization behavior of this
ternary mixture, DSC measurements were not performed in
this work, but should be the subject of further investigations.
The slow crystallization kinetics of mixtures consisting of
carbazole derivatives may mislead experimentalists to believe
that the mixtures have melting points far below room
temperature. Therefore, one should keep in mind that liquid
carbazole mixtures are able to crystallize even after more than a
month storage time.
4. SUMMARY
ASSOCIATED CONTENT
The high melting point of N-ethylcarbazole is often mentioned
as the main drawback for its use as a liquid organic hydrogen
carrier;14−16 therefore, different approaches for achieving a
melting point depression preferably to temperatures below
ambient were investigated.
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.jced.5b00679.
Conditions of the DSC measurements (PDF)
Incomplete dehydrogenation might be an efficient measure
for a sufficient melting point depression. The melting points of
N-ethylcarbazole derivatives with different degrees of hydro-
genation decrease in the following order: 0H-NEC, 8H-NEC,
4H-NEC, 12H-NEC. Because of a partly hydrogenated LOHC
system being a multinary mixture, the solid−liquid behavior is
very complex. The melting point decreases with increasing
amount of 12H-NEC. However, 80% of 12H-NEC are
necessary to keep a binary mixture of NEC−12H-NEC liquid
at room temperature. This would drastically reduce the effective
hydrogen storage density. Hence, on the basis of the current
knowledge an incomplete dehydrogenation does not seem to
be a reasonable solution particularly considering the loss in
hydrogen storage capacity.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: karsten.mueller@fau.de.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors thank Prof. Dr. Axel Konig and Dr. Matthias Kick
for helpful discussions and support of this work. We want to
thank Roswitha Eckstein for assistance during the experimental
̈
work and Katharina Obesser, Michael Muller, and Marcus Koch
from the group of Prof. Dr. Peter Wasserscheid for the
synthesis of chemical species.
An alternative approach is to change the length of the alkyl
chain in order to find an alkylcarbazole having a significantly
lower melting point than N-ethylcarbazole. The melting point
decreases with increasing chain length from carbazole to N-
propylcarbazole and tends to increase again from N-
butylcarbazole on. The branched N-isopropyl-carbazole has a
comparatively higher melting point in comparison to its linear
isomer N-propylcarbazole. Substituting N-ethylcarbazole by N-
propylcarbazole, the melting point of the fully dehydrogenated
LOHC can be reduced from 342.4 to 320.4 K. However, this is
accompanied by a slight reduction in hydrogen storage density.
̈
REFERENCES
■
(1) Pez, G. P.; Scott, A. R.; Cooper, A. C.; Cheng, H. Hydrogen
storage by reversible hydrogenation of pi-conjugated substrates. EP
Patent EP1475349A2, 2004.
(2) Pez, G. P.; Scott, A. R.; Cooper, A. C.; Cheng, H.; Wilhelm, F. C.;
Abdourazak, A. H. Hydrogen storage by reversible hydrogenation of
pi-conjugated substrates. U.S. Patent US7351395B1, 2008.
1447
DOI: 10.1021/acs.jced.5b00679
J. Chem. Eng. Data 2016, 61, 1441−1448

Journal of Chemical & Engineering Data
Article
(3) Pez, G. P.; Scott, A. R.; Cooper, A. C.; Cheng, H. Hydrogen
storage by reversible hydrogenation of pi-conjugated substrates. U.S.
Patent US20050002857A1, 2005.
(4) Eblagon, K. M.; Tam, K.; Yu, K. M. K.; Tsang, S. C. E.
Comparative Study of Catalytic Hydrogenation of 9-Ethylcarbazole for
Hydrogen Storage over Noble Metal Surfaces. J. Phys. Chem. C 2012,
116, 7421−7429.
(5) Eblagon, K. M.; Tam, K.; Tsang, S. C. E. Comparison of catalytic
performance of supported ruthenium and rhodium for hydrogenation
of 9-ethylcarbazole for hydrogen storage applications. Energy Environ.
Sci. 2012, 5, 8621−8630.
(6) Eblagon, K. M.; Rentsch, D.; Friedrichs, O.; Remhof, A.; Zuettel,
A.; Ramirez-Cuesta, A. J.; Tsang, S. C. Hydrogenation of 9-
ethylcarbazole as a prototype of a liquid hydrogen carrier. Int. J.
Hydrogen Energy 2010, 35, 11609−11621.
(7) Sotoodeh, F.; Smith, K. J. Structure sensitivity of dodecahydro-N-
ethylcarbazole dehydrogenation over Pd catalysts. J. Catal. 2011, 279,
36−47.
(8) Sotoodeh, F.; Huber, B. J. M.; Smith, K. J. Dehydrogenation
kinetics and catalysis of organic heteroaromatics for hydrogen storage.
Int. J. Hydrogen Energy 2012, 37, 2715−2722.
components and charge-transfer complexes. Thermochim. Acta 1980,
40, 405−414.
́
(24) Lisicki, Z.; Jamroz, M. E. (Solid + liquid) equilibria in
(polynuclear aromatic + tertiary amide) systems. J. Chem. Thermodyn.
2000, 32, 1335−1353.
(25) Domalski, E. S.; Hearing, E. D. Heat Capacities and Entropies of
Organic Compounds in the Condensed Phase. Volume III. J. Phys.
Chem. Ref. Data 1996, 25, 1−525.
(26) Renon, H.; Prausnitz, J. M. Local compositions in
thermodynamic excess functions for liquid mixtures. AIChE J. 1968,
14, 135−144.
(27) Klamt, A.; Schuurmann, G. COSMO: a new approach to
dielectric screening in solvents with explicit expressions for the
screening energy and its gradient. J. Chem. Soc., Perkin Trans. 2 1993,
799−805.
(28) Maximo, G. J.; Costa, M. C.; Meirelles, A. J. A. Solid-liquid
equilibrium of triolein with fatty alcohols. Braz. J. Chem. Eng. 2013, 30,
33−43.
(29) DOE. Targets for onboard hydrogen storage for light duty fuel
cell vehicles. http://energy.gov/sites/prod/files/2014/03/f12/
targets_onboard_hydro_storage.pdf (accessed 05.12.2014).
(9) Teichmann, D.; Stark, K.; Muller, K.; Zottl, G.; Wasserscheid, P.;
Arlt, W. Energy storage in residential and commercial buildings via
Liquid Organic Hydrogen Carriers (LOHC). Energy Environ. Sci. 2012,
5, 9044−9054.
(10) Teichmann, D.; Arlt, W.; Wasserscheid, P.; Freymann, R. A
future energy supply based on Liquid Organic Hydrogen Carriers
(LOHC). Energy Environ. Sci. 2011, 4, 2767−2773.
(11) Teichmann, D.; Arlt, W.; Wasserscheid, P. Liquid Organic
Hydrogen Carriers as an efficient vector for the transport and storage
of renewable energy. Int. J. Hydrogen Energy 2012, 37, 18118−18132.
(12) Mehranfar, A.; Izadyar, M.; Esmaeili, A. A. Erratum to
“Hydrogen storage by N-ethylcarbazol as a new liquid organic
hydrogen carrier: A DFT study on the mechanism. Int. J. Hydrogen
Energy 2015, 40, 7420−7421.
(13) Verevkin, S. P.; Emel’yanenko, V. N.; Pimerzin, A. A.;
Vishnevskaya, E. E. Thermodynamic Analysis of Strain in the Five-
Membered Oxygen and Nitrogen Heterocyclic Compounds. J. Phys.
Chem. A 2011, 115, 1992−2004.
(14) Brueckner, N.; Obesser, K.; Boesmann, A.; Teichmann, D.; Arlt,
W.; Dungs, J.; Wasserscheid, P. Evaluation of industrially applied heat-
transfer fluids as liquid organic hydrogen carrier systems. Chem-
SusChem 2014, 7, 229−235.
(15) Sotoodeh, F.; Smith, K. J. An overview of the kinetics and
catalysis of hydrogen storage on organic liquids. Can. J. Chem. Eng.
2013, 91, 1477−1490.
(16) Klebanoff, L. Hydrogen Storage Technology: Materials and
Applications; CRC Press: Boca Raton, 2013.
(17) Yang, M.; Dong, Y.; Fei, S.; Pan, Q.; Ni, G.; Han, C.; Ke, H.;
Fang, Q.; Cheng, H. Hydrogenation of N-propylcarbazole over
supported ruthenium as a new prototype of liquid organic hydrogen
carriers (LOHC). RSC Adv. 2013, 3, 24877−24881.
(18) Sabbah, R.; Watik, L. E. New reference materials for the
calibration (temperature and energy) of differential thermal analysers
and scanning calorimeters. J. Therm. Anal. 1992, 38, 855−863.
(19) Lopatinskii, V. P.; Sirotkina, E. E.; Sukhoroslova, M. M. N-
Alkylcarbazoles. Metody Poluch. Khim. Reakt. Prep. 1964, 11, 69−72.
(20) Krajewska-Cizio, A. Phase diagrams in the binary systems of N-
isopropylcarbazole with 2,4,7-trinitrofluoren-9-one and carbazole.
Thermochim. Acta 1990, 158, 317−325.
(21) Stark, K.; Westerholt, A.; Wasserscheid, P.; Arlt, W. Trennung
komplexer Mischungen von partiell hydrierten organischen Verbin-
dungen und ihrer Isomere durch Destillation. Chem. Ing. Tech. 2012,
84, 1269−1269.
(22) Hohne, G.; Hemminger, W.; Flammersheim, H. J. Differential
̈
Scanning Calorimetry; Springer-Verlag: Berlin Heidelberg, 2003.
(23) Radomska, M.; Radomski, R. Calorimetric studies of binary
systems of 1,3,5-trinitrobenzene with naphthalene, anthracene and
carbazole. I. Phase transitions and heat capacities of the pure
1448
DOI: 10.1021/acs.jced.5b00679
J. Chem. Eng. Data 2016, 61, 1441−1448