Solvent effects on the hydrogenolysis of diphenyl ether with raney nickel and their implications for the conversion of lignin

Article abstract of DOI:10.1002/cssc.201200040

The conversion of lignin, the most recalcitrant of the biopolymers, is necessary for a carbon-efficient utilization of lignocellulosic materials. In this context, hydrogenolysis of lignin is a process receiving increasing attention. In this report, the solvent effects on the hydrogenolysis of diphenyl ether and lignin with Raney Ni are addressed. The Lewis basicity of the solvent very much affects the catalytic activity, so Raney Ni in nonbasic solvents is an extremely active catalyst for hydrogenolysis and hydrogenation. In basic solvents, however, Raney Ni is a less active, but much more selective catalyst for hydrogenolysis while preserving the aromatic products. With regard to the reactions with lignin, assessing the complexity of the product mixtures by two-dimensional GC×GC-MS revealed solvent effects on the product distribution. Reaction in methylcyclohexane resulted in cyclic alcohols and cyclic alkanes, whereas reaction in 2-propanol led to cyclic alcohols, cyclic ketones, and unsaturated products. The hydrogenolysis of lignin in methanol, however, produced mostly phenols. Overall, these results demonstrate that the solvent plays a key role in directing the selectivity and, thus, it must be taken into consideration in the design of catalytic systems for conversion of lignin by hydrogenolysis of C-O ether bonds. Copyright

Full text of DOI:10.1002/cssc.201200040

DOI: 10.1002/cssc.201200040
Solvent Effects on the Hydrogenolysis of Diphenyl Ether
with Raney Nickel and their Implications for the
Conversion of Lignin
Xingyu Wang and Roberto Rinaldi*[a]
The conversion of lignin, the most recalcitrant of the biopoly-
mers, is necessary for a carbon-efficient utilization of lignocel-
lulosic materials. In this context, hydrogenolysis of lignin is
a process receiving increasing attention. In this report, the sol-
vent effects on the hydrogenolysis of diphenyl ether and lignin
with Raney Ni are addressed. The Lewis basicity of the solvent
very much affects the catalytic activity, so Raney Ni in nonbasic
solvents is an extremely active catalyst for hydrogenolysis and
hydrogenation. In basic solvents, however, Raney Ni is a less
active, but much more selective catalyst for hydrogenolysis
while preserving the aromatic products. With regard to the re-
actions with lignin, assessing the complexity of the product
mixtures by two-dimensional GCꢀGC–MS revealed solvent ef-
fects on the product distribution. Reaction in methylcyclohex-
ane resulted in cyclic alcohols and cyclic alkanes, whereas reac-
tion in 2-propanol led to cyclic alcohols, cyclic ketones, and
unsaturated products. The hydrogenolysis of lignin in metha-
nol, however, produced mostly phenols. Overall, these results
demonstrate that the solvent plays a key role in directing the
selectivity and, thus, it must be taken into consideration in the
design of catalytic systems for conversion of lignin by hydroge-
nolysis of CꢀO ether bonds.
Introduction
Lignin occurs alongside hemicellulose and cellulose
in plant cell walls,[1] and protects the plant tissue
from mechanical stress and microbial attack.[2] Ac-
cordingly, lignin is the most recalcitrant of the biopo-
lymers, and perhaps the most difficult hurdle facing
the efficient biorefining of lignocellulosic materials,
which comprise as much as 30% lignin.[3] Catalysis
has an enormous potential for the development of
efficient processes for the valorization of this raw ma-
terial, available in vast quantities.[1,3] The structural
model of native lignin (Scheme 1)[2] suggests that this
feedstock could be converted into aromatics, phe-
nols, and cyclic saturates.[3] However, serious knowl-
edge gaps in catalysis for conversion of lignin pres-
ent a major barrier that must be overcome to fully
exploit this feedstock for the chemical industry.[3a]
Unlike cellulose, which is a linear homopolymer of
glucose,[5] lignin is a network polymer[1a,4] comprising
phenylpropenyl units (coumaryl, syringyl, and conifer-
yl alcohols) randomly connected through carbonꢀ
carbon and carbonꢀoxygen bonds (Scheme 1).[3] As
much as 80% of the primary inter-unit linkages of
lignin are ether bonds.[5] The distribution of the link-
ages in the biopolymer structure depends not only
Scheme 1. Example of lignin structure depicting some of the primary inter-unit linkages
and their estimated BDE values (adapted from Ref. [2]). The BDE values were obtained
from Ref. [6a].
upon the type of wood but also upon the part of the plant.
The most predominant ether linkage is the b-O-4 (softwood:
39 to 48%, hardwood: 32 to 37%), followed by the a-O-4 (soft-
wood: 11 to 16%, hardwood: 28 to 32%).[5]
to about 7% of the inter-unit linkages in softwood and to ca.
2% in hardwood. In turn, the carbonꢀcarbon bonds, a-1 and
[a] X. Wang, Dr. R. Rinaldi
Max-Planck-Institut fꢀr Kohlenforschung
Kaiser-Wilhelm-Platz 1, 45470 Mꢀlheim (Ruhr) (Germany)
Fax: (+49)208-306-2368
The bond dissociation enthalpies (BDE) of the ether linkages
reveal the order of bond strengths: a-O-4<b-O-4<4-O-5
(Scheme 1).[6] The strongest ether linkage, 4-O-5, corresponds
E-mail: rinaldi@mpi-muelheim.mpg.de
ChemSusChem 2012, 5, 1 – 13
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
ÞÞ
These are not the final page numbers!

X. Wang, R. Rinaldi
5’-5, have BDEs of ca. 100 and 200 kJmolꢀ1, respectively, higher
than those of the b-O-4 linkages (Scheme 1).[6] Therefore, the
hydrogenolysis of the weak ether linkages, a-O-4 and b-O-4, is
the most enthalpy-favorable strategy for the hydrogenolysis of
lignin towards simpler molecules.
paid to the importance of the solvent as part of the catalytic
system. Unlike the organic synthesis literature, in which solvent
effects are well-described and understood, the influence of the
solvent in the hydrogenation and hydrogenolysis catalyzed by
solids are often neglected and/or poorly comprehended.[14] Al-
though the hydrogenation and hydrogenolysis of ethers, cata-
lyzed by Pt, Pd and Raney Ni, are reactions already known for
a century,[27,28] there are only few reports addressing solvent ef-
fects on the hydrogenation of aromatic alcohols and ketones
with Pt/Al2O3,[13a] Ru/Al2O3,[13a] Pd on several supports,[13b] and
Ni/SiO2 catalysts,[13c] and on the hydrogenolysis of lignin.[21b]
These limited, but valuable reports indicate that the solvent is
a parameter affecting the reaction performance. However, ad-
vances towards understanding solvent effects on the hydroge-
nolysis of lignin are still needed to lay the foundations for the
rational design of catalytic systems for valorization of lignin.
Herein, solvent effects on the hydrogenolysis of diphenyl
ether and lignin with Raney Ni are addressed in detail. In the
first part, the critical factors responsible for the solvent effects
on the hydrogenolysis of diphenyl ether are discussed. In the
second part, we draw out the major implications of the solvent
effects on the conversion of organosolv lignin in several sol-
vents. One of the more significant findings to emerge from
this study is that the Lewis basicity of the solvent holds the
key to control the activity as well as the selectivity of Raney Ni
in the conversion of lignin into saturates with reduced oxygen
content or into phenols.
Harris and Adkins gave the first report on hydrogenolysis of
lignin in 1938.[7] Lignins isolated by different methods were
subjected to a temperature of 2508C, using a copper-chromite
catalyst in 1,4-dioxane under 170 to 400 bar of H2 for 18 h.[7]
The main isolated products were methanol, 4-propylcyclohexa-
nol, 4-propylcyclohexane-1,2-diol and 4-(3-hydroxypropyl)cy-
clohexanol. Conversions of lignin between 34 and 78% were
achieved.
In the 1960s Pepper and Steck[8] reported that lignin suffers
hydrogenolysis yielding mostly 4-propylsyringol, dihydroxysi-
napyl alcohol, and dihydroconiferyl alcohol. The reactions ach-
ieved up to 52% conversion using Raney Ni at 150–2208C,
34 bar of H2 in dioxane-water (1:1, v/v). In a later report,[8b] they
showed that Ru/C and Ru/Al2O3 are very active for the hydro-
genolysis of lignin leading to phenols and cyclohexanols.
Recently, Ragauskas et al.[9] screened several solid catalysts
(Co/Mo, Raney Ni, Pd/C, and Pt/C), homogeneous catalysts
[NaBH4/I2, RhCl(PPh3)3, Ru(Cl)2(PPh3)3, and Ru(H)(Cl)(PPh3)3], and
Ru-(PVP) nanoparticles in the hydrogenolysis of ethanol orga-
nosolv lignin. The reactions were performed in ethanol. Al-
though the products were not identified, the potential of the
systems for hydrogenation and hydrogenolysis of lignin was
1
demonstrated using solubility tests and H and 13C NMR spec-
troscopy. The reaction catalyzed by Ru(Cl)2(PPh3)3 resulted in
96.4% of ethanol-soluble products.
Results and Discussion
Yan et al.[10] proposed a two-step process for hydrogenolysis
of lignin. The reaction was first performed in water or water/di-
oxane (1:1) in the presence of H3PO4 and Pt/C at 2008C, 40 bar
of H2 for 4 h. In sequence, the product mixture was hydrogen-
ated over a Pd/C catalyst at 2508C, 40 bar H2, for 0.5 to 2 h, af-
fording C8–C9 (42%) and C14–C18 (10%) hydrocarbons.
Ford et al.[11] introduced recently the hydrogenolysis of lignin
using supercritical methanol as hydrogen source in the pres-
ence of mesoporous copper-based catalysts at 3008C. The re-
action produced mainly substituted cyclohexyl derivatives with
reduced oxygen-content and a minor amount of aromatics. Be-
ginning the process with wooden biomass led mostly to alco-
hols.
Solvent effects on the hydrogenolysis of diphenyl ether
Aiming to ultimately assess the solvent effects on the hydroge-
nolysis of lignin, we first explored the influence of several sol-
vents in the reaction with diphenyl ether. The solvents were
classified into four groups according to their properties
(Table 1):
1. Protic solvents displaying Lewis basicity: methanol, ethanol,
2-propanol, 1-butanol, 2-butanol, and tert-butanol. They show
high polarity, as indicated by the largest values of ETN. These
solvents are both good H-bond donors (a¼0) and good H-
bond acceptors (b¼0). Moreover, they are the most Lewis-
basic solvents, as indicated by the highest values of donor
number (DN>100 kJmolꢀ1).
Selective hydrogenolysis of lignin model compounds to-
wards phenols and aromatics with homogeneous catalysts
were recently reported.[12] Sergeev and Hartwig[12a] showed
that aryl ethers undergo selective hydrogenolysis in the pres-
ence of a Ni-carbene complex under 1 bar of H2 at 80 to 1208C
in m-xylene. Bergman and Ellman[12b] et al. demonstrated the
redox neutral, ruthenium-catalyzed cleavage of 2-aryloxy-1-ary-
lethanols in toluene. The conversion of a polyether lignin
model, performed in 1,4-dioxane due to the low solubility of
the polymer in toluene, achieved 99% yield of monomer after
reaction at 1758C for 3 h.
2. Protic solvent displaying no Lewis basicity: Hex-F-2-PrOH,
or 1,1,1,3,3,3-hexafluoro-2-isopropanol, has the highest a value
(1.96) and is, thus, the best H-bond donor among the solvents
listed in Table 1. This solvent is not an H-bond acceptor (b=0)
or a Lewis base (DN=0).
3. Aprotic polar solvents: ethyl acetate, tetrahydrofuran
(THF), 2-methyltetrahydrofuran (2-Me-THF), and 1,4-dioxane are
not H-bond donors (a=0), but they are H-bond acceptors (b¼
0). Furthermore, they are less basic (50<DN<84 kJmolꢀ1) than
the solvents from group 1.
These early and the recent advances towards hydrogenolysis
of lignin have been focused exclusively on the development
and screening of catalysts.[1a] Far too little attention has been
4. Aprotic nonpolar solvents: methylcyclohexane (MCH),
decaline, and n-heptane. They are not H-bond donors or ac-
ceptors (a, b=0). They show no Lewis basicity (DN=0).
&2
&
www.chemsuschem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 13
ÝÝ
These are not the final page numbers!

Solvent Effects on the Hydrogenolysis of Diphenyl Ether with Raney Ni
Table 2 summarizes the results from the hydrogenolysis of
diphenyl ether with Raney Ni conducted in several solvents
under 50 bar H2 (RT) at 908C for 2.5 h. The results reveal that
the reaction depends heavily upon the solvent. Among the sol-
vents from group 1, the conversion decreases in the following
order: 2-propanol (72.7%), 2-butanol (55.1%), ethanol (33.0%),
1-butanol (32.0%), tert-butanol (21.7%), and methanol (12.4%).
Markedly, the reaction in Hex-F-2-PrOH led to full conversion
of the substrate. Among group 3, the reaction showed the
best conversions in 2-Me-THF (61.1%) and ethyl acetate
(47.0%). Poor conversions were obtained in THF (29.6%) and
1,4-dioxane (17.8%). Conducting the reaction in alkanes, how-
ever, resulted in very high to full conversion: n-heptane
(99.0%), decaline (99.6%), and MCH (100%).
Table 1. The property parameters of some organic solvents.[15]
Solvent
ETN
a
b
p*
AN
[kJmolꢀ1
DN
]
[kJmolꢀ1
]
methanol
ethanol
0.762 0.98 0.66 0.60
0.654 0.86 0.75 0.54
0.546 0.76 0.84 0.48
1.068 1.96 0.00 0.65
0.602 0.84 0.84 0.47
0.506 0.69 0.80 0.40
0.389 0.42 0.93 0.41
0.228 0.00 0.45 0.55
0.207 0.00 0.55 0.58
172.6
155.1
140.0
278.8
153.8
109.0[a]
113.3
38.9
33.4
50.6[a]
43.1
0
125.5
133.9
150.6
0
121.3
136.4[a]
159.0
71.5
83.7
50.2
59.8
0
2-propanol
Hex-F-2-PrOH
1-butanol
2-butanol
t-butanol
ethyl acetate
tetrahydrofuran
2-Me-THF
1,4-dioxane
n-heptane
decaline
0.179 0.00 0.45
–
0.164 0.00 0.37 0.55
0.012 0.00 0.00 ꢀ0.08
0.015 0.00 0.08 0.11
0.006 0.00 0.00 0.00
0
0
0
0
Raney Ni catalyzes both hydrogenolysis of ether bonds and
hydrogenation of arenes, as seen in Table 2. Analyzing the
ratios phenol/cyclohexanol (4:6) and benzene/cyclohexane
(5:7) reveals that solvents can be part of a strategy to improve
the selectivity for aromatics. In fact, the reactions performed in
methanol and in 1,4-dioxane produced a large fraction of
phenol and benzene, whereas the reaction in Hex-F-2-PrOH
and in alkanes resulted in a very low yield to no aromatics.
Moreover, the values of the ratio S(4,6)/S(5,7) are very close to
unity. This indicates that the formation of phenol, benzene, cy-
clohexanol, and cyclohexane happens exclusively through hy-
drogenolysis of the ether bond. Although Raney Ni catalyzes
the hydrogenation of phenol to cyclohexanol, this catalyst
itself cannot cause the dehydration of alcohols to alkanes.[28]
The efficient conversion of phenols to alkanes requires both
a hydrogenation and an acid catalyst.[16]
MCH[b]
[a] DN and AN values were estimated by binary correlations of parame-
ters DN [kcalmolꢀ1]=0.25+40.4b and AN [kcalmolꢀ1]=12.1+33.6b, as
described in Ref. [15]. [b] Because the solvent parameters of MCH are not
available in the literature, the values for cyclohexane were used instead.
To unveil the causes of solvent effects on heterogeneously-
catalyzed reactions, it is essential to conduct the reactions in
the kinetic regime, that is, in absence of rate-determining mass
transfer of H2 at the gas/liquid interface as well as without
liquid/solid and pore diffusion limitations.[14] To check if we
were working under these conditions, we performed the reac-
tions in methanol, ethanol, or cyclohexane at different propel-
ler speeds (300 and 400 rpm). As no changes in the values of
conversion of diphenyl ether and in the product distribution
were detected under both conditions, the transport of H2 from
the gas phase through the liquid phase onto the active sites is
not a rate-determining step.
Even in the absence of mass transport problems, the solubil-
ity of H2 in the reaction medium can affect the catalyst perfor-
mance.[14] Nonetheless, comparing the conversion values ach-
ieved in n-heptane (99.0%), ethyl acetate (47.0%), and in
Table 2. Products of the hydrogenolysis of diphenyl ether conducted in different solvents.[a]
Entry
Solvent
Conversion [%]
Selectivity [%]
2
3
4
5
6
7
4:6
5:7
S(4,6)/S(5,7)
1
2
3
4
5
6
7
8
methanol
ethanol
12.4
33.0
72.7
100
13.7
11.8
15.0
0
0
25.7
10.7
1.6
0
9.2
10.3
18.4
8.1
14.2
2.9
24.1
0
32.3
26.7
15.4
0
22.5
18.0
17.2
15.7
25.3
12.8
22.0
1.5
16.8
33.6
40.5
16.6
36.8
32.1
21.4
30.1
27.6
34.4
13.3
27.5
26.6
22.8
11.5
16.9
25.6
16.1
21.5
23.4
21.2
23.3
17.4
24.1
16.1
24.9
25.7
21.8
1.53
0.32
0.04
0
0.25
0.32
0.86
0.27
0.51
0.08
1.81
0
2.81
1.58
0.60
0
1.05
0.77
0.81
0.67
1.45
0.53
1.37
0.06
0
0.97
1.02
1.03
1.03
1.05
1.02
1.04
0.98
0.98
1.01
0.98
1.04
1.04
1.05
0.3
1.9
67.3
0.7
0.8
1.4
1.8
0
1.7
0
36.5
41.4
55.4
2-propanol
Hex-F-2-PrOH
1-Butanol
2-Butanol
t-butanol
ethyl acetate
tetrahydrofuran
2-Me-THF
1,4-dioxane
n-heptane
decaline
32.0
55.1
21.7
47.0
29.6
61.1
17.8
99.0
99.6
100
9.3
15.4
20.4
21.0
15.5
24.1
24.5
9.6
9
10
11
12
13
14
6.3
0
0
0
0
0
0
0
MCH
0
Reaction conditions: diphenyl ether (2.9 mmol), dibutyl ether (internal standard, 0.78 mmol), Raney Ni (100 mg), and solvent (15 mL), 50 bar H2 (RT), 908C
for 2.5 h.
ChemSusChem 0000, 00, 1 – 13
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
&
3
&
ÞÞ
These are not the final page numbers!

X. Wang, R. Rinaldi
methanol (12.4%), with the solubility of H2 in the respective
solvents (n-heptane, 4.70 molmꢀ3; ethyl acetate, 3.54 molmꢀ3
sion of diphenyl ether (adjusted R2 =0.71).
;
and methanol, 3.54 molmꢀ3 at 1 bar and 258C)[17] shows that
the solubility of H2 is not responsible for the different results
displayed in Table 2. For example, the solubility of H2 in ethyl
acetate is similar to that in methanol; however, the reaction in
ethyl acetate achieved a conversion three times higher than
that in methanol. Overall, this indicates that solvent interac-
tions, established with the substrate–surface complex and with
the catalyst surface itself, are of first importance.
*
Conversion ð%Þ ¼ 102:6þ24:7aꢀ47:0bꢀ90:1p
ð1Þ
Table 3 displays the experimental conversions as well as
those predicted by Equation (1) and the deviation values. For 9
of the 14 reactions, the predicted conversions showed a devia-
tion of 11% or less. However, the predicted conversion for the
Plotting the parameters of solvent properties (Table 1)
against the substrate conversion is an approach often adopted
in order to shed light onto the interactions accounting for the
solvent effects on catalytic reactions.[13,14] At first sight, Table 2
suggests that the hydrogenolysis of diphenyl ether be favored
in nonpolar solvents (e.g., alkanes). However, Figure 1 does not
show an apparent correlation between conversion of diphenyl
ether and the empirical parameter of solvent polarity (ETN).
Indeed, the reactions performed in the solvents of group 3 led
to totally different results, although these solvents display
fairly similar values of ETN (Table 2, entries 8–11).
Table 3. Experimental and predicted conversions of diphenyl ether in the
presence of Raney Ni conducted in different solvents.[a]
Solvent
Experimental
conversion [%]
Predicted
conversion [%]
Deviation[b]
[%]
methanol
ethanol
12.4
33.0
72.7
100
32.0
55.1
21.7
47.0
29.6
17.8
99.0
99.6
100
41.7
39.9
38.6
92.4
41.5
46.0
32.3
31.9
24.5
35.7
109.8
88.9
102.6
29.3
6.9
ꢀ34.1
ꢀ7.6
9.5
ꢀ9.1
10.6
ꢀ15.1
ꢀ5.1
17.9
2-propanol
Hex-F-2-PrOH
1-butanol
2-butanol
t-butanol
ethyl acetate
tetrahydrofuran
1,4-dioxane
n-heptane
decaline
10.8
ꢀ10.7
2.6
MCH[b]
[a] As the p* parameter of 2-Me-THF is not available in the literature, we
omitted this solvent in the multilinear model and in the prediction.
[b] Deviation (%)=(Predicted Conversion)ꢀ(Experimental Conversion).
reactions in 1,4-dioxane and in ethyl acetate showed high de-
viations, ꢀ15.1 and 17.9%, respectively. In addition, the model
could not predict properly the conversion of diphenyl ether in
methanol and in 2-propanol. In methanol the reaction ach-
ieved a lower conversion (12.4%) than the value predicted by
the multilinear model (41.7%). In 2-propanol, the experimental
conversion (72.7%) was much higher than that predicted by
the model (38.6%). Despite the outliers, the multilinear model
suggests which factors should be responsible for the perfor-
mance of Raney Ni in the liquid-phase reaction. A mathemati-
cal analysis of Equation (1) reveals that parameter a has a posi-
tive effect on the conversion of diphenyl ether in the presence
of Raney Ni, whereas parameters b and p* exert a negative
effect on the catalyst performance. In addition, the angular co-
efficient of the explanatory variable p* is twice smaller than
that of b. This suggests that Raney Ni is more sensitive to sol-
vents with high values of p* than to those with high values of
b.
Figure 1. Conversion of diphenyl ether versus solvent polarity (ETN).
The solvent polarity, as described by ETN, corresponds to an
average of specific and nonspecific interactions that a solvent
establishes with a solute.[15] To measure separately the specific
interactions contributing to the overall solvent polarity,
a better approach is the use of the parameters obtained from
Kamlet–Taft solvatochromic relationships, namely H-bond
donor (a), H-bond acceptor (b), and polarity/polarizability (p*)
properties.[15] In a different analysis than that shown in
Figure 1, we considered the Kamlet–Taft parameters as explan-
atory variables of a multilinear regression, in which the conver-
sion of diphenyl ether was the response variable. The multilin-
ear regression resulted in Equation (1) as a mathematical
model correlating the Kamlet–Taft parameters with the conver-
One question that needs to be asked, however, is how the
parameters a, b, and p* are capable of governing the activity
of Raney Ni. In fact, there is no reason to assume that the
weak specific (H-bonding) and/or nonspecific (van der Waals)
interactions involving solvent and substrate would influence
the hydrogenolysis of diphenyl ether, which takes place on the
catalyst surface due to strong interactions established between
substrate and active sites.
&4
&
www.chemsuschem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 13
ÝÝ
These are not the final page numbers!

Solvent Effects on the Hydrogenolysis of Diphenyl Ether with Raney Ni
In a seminal work, Marcus[15] demonstrated through several
binary correlations of solvent parameters that the parameter
a is closely associated with the acceptor number (AN), a quanti-
tative measure of Lewis acidity. Likewise, he also showed that
the parameter b strongly correlates with the donor number
(DN), a quantitative measure of Lewis basicity. In turn, the pa-
rameter p*, which describes a combination of polarity and po-
larizability of a solvent, does not correlate with acidity or basic-
ity.[15] Taken together, these pieces of information suggest that
the multilinear model be ultimately correlating the Lewis acidi-
ty and Lewis basicity of the solvents with the performance of
Raney Ni in the hydrogenolysis of diphenyl ether.
To explore how the Lewis acid-base interactions, established
between solvent and catalyst surface, affect the hydrogenolysis
of diphenyl ether, the substrate conversion was plotted against
AN and DN separately. On one hand, Figure 2 shows that the
Figure 3. Conversion of diphenyl ether versus the Lewis basicity of the sol-
vents as described by the donor number (DN).
Table 4. Experimental and predicted conversions of diphenyl ether in the
presence of Raney Ni conducted in different solvents.
Solvent
Experimental
conversion [%]
Predicted
conversion [%][a]
Deviation[b]
[%]
methanol
ethanol
12.4
33
72.7
100
32
55.1
21.7
47
29.6
61.1
17.8
99
29.3
27.7
25.2
99.8
30.1
27.3
24.2
45.2
40.5
56.0
50.7
99.8
99.8
99.8
ꢀ16.9
5.3
47.5
0.2
2-propanol
Hex-F-2-PrOH
1-butanol
2-butanol
t-butanol
1.9
27.8
ꢀ2.5
1.8
ꢀ10.9
5.1
ꢀ32.9
ꢀ0.8
ꢀ0.2
0.2
ethyl acetate
tetrahydrofuran
2-Me-THF
1,4-dioxane
n-heptane
decaline
Figure 2. Conversion of diphenyl ether versus the Lewis acidity of the sol-
vents as described by the acceptor number (AN).
99.6
100
MCH[b]
[a] Values predicted from Equation (2). [b] Deviation (%)=(Predicted con-
version)ꢀ(Experimental conversion)
Lewis acidity of the solvents does not affect the conversion of
diphenyl ether. This result contrasts with that obtained from
the multilinear model, which shows that parameter a exerted
a positive effect on the conversion. On the other hand, in
Figure 3 it is possible to identify a correlation between the
Lewis basicity of the solvents and the substrate conversion. As
indicated by the dashed line in Figure 3, an exponential decay
model fits the conversions—achieved in MCH, decaline, n-hep-
tane, Hex-F-2-PrOH, 2-Me-THF, ethyl acetate, tetrahydrofuran,
1-butanol, ethanol, and tert-butanol—very well with the re-
spective Lewis basicity of the solvent (adjusted R2 0.97), as
given by Equation (2):
values. For 10 of the 14 reactions, the predicted conversions
showed a deviation of ꢁ5.3% or less. However, the model fails
to predict the conversions of the reactions performed in meth-
anol, 2-propanol, 2-butanol, or 1,4-dioxane. Large positive devi-
ations were obtained from Equation (2) for 2-propanol (47.5%)
and 2-butanol (27.8%), whereas large negative deviations were
found for methanol (ꢀ16.9%) and 1,4-dioxane (ꢀ32.9%). In
spite of this, it is clear from Figure 3 that solvents possessing
a high Lewis basicity, such as those from group 1, hinder the
performance of Raney Ni. This is confirmed when the reaction
is conducted in Hex-F-2-PrOH, a solvent showing no Lewis ba-
sicity (DN=0). In the fluorinated solvent, full conversion of the
substrate was achieved (Table 2, entry 4). Such a result was
only achieved in reactions carried out in alkanes (Table 2, en-
tries 12–14), which also display no Lewis basicity.
DN þ 2:61
ð2Þ
Conversion ð%Þ ¼ 16:17 þ 86:94exp
67:67
Table 4 summarizes the experimental and the predicted con-
versions, calculated from Equation (2), and the deviation
ChemSusChem 0000, 00, 1 – 13
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
&
5
&
ÞÞ
These are not the final page numbers!

X. Wang, R. Rinaldi
No direct assessment of the effect of water, as solvent, on
the catalytic activity of Raney Ni could be undertaken since di-
phenyl ether is insoluble in water. In a biphasic system, our
analysis could be biased by mass transport phenomena and
other factors. The current results suggest, however, that water,
the interactions of secondary alcohols with Ni surface should
also be taken into account. In a surface study, Raval et al.[19] re-
ported that the interaction of 2-propanol with a Ni(111) surface
cleaves both the OꢀH and the a-CꢀH bonds. As a result, ace-
tone is released into the gas phase and two H atoms remain
adsorbed on the Ni surface. The H atoms can be used in hy-
drogenation or hydrogenolysis.[20]
having a DN of 102 kJmolꢀ1 [30]
should have an impact on the
,
activity of Raney Ni comparing to that of ethanol or 1-butanol
(Figure 3).
To verify whether hydrogenolysis of diphenyl ether takes
place at 908C, in the absence of H2, reactions in methanol, 2-
propanol, 2-butanol, Hex-F-2-PrOH, decaline, or MCH were car-
ried out with Raney Ni for 2.5 h. Table 5 shows that the reac-
tion in 2-propanol or 2-butanol also happened in the absence
of H2. In methanol, the reaction took place only to a small
extent, whereas no reaction happened in Hex-F-2-PrOH, deca-
line, or MCH at 908C.
The reactions performed in secondary alcohols demonstrate
that an H transfer reaction is cooperating with the reaction in-
volving H2. In addition, considering the predicted conversion in
2-propanol (Table 4), one can conclude that if the reaction
through H transfer took place in parallel to the “regular” hydro-
genolysis using H2, a conversion of ca. 43% would be achieved
in 2-propanol, instead of 72.7%. Likewise, in 2-butanol, a con-
version of ca. 45% would be achieved instead of 55.1%. These
results reveal a synergism between H transfer and “regular” ac-
tivation of H2 promoting the conversion of diphenyl ether.
While there is strong evidence that one of the causes of the
solvent effect is the Lewis basicity of the solvent, which sup-
posedly favors the adsorption of the solvent blocking partially
the active Ni sites for hydrogenolysis and hydrogenation, the
unexpected low activity of Raney Ni in methanol and in 1,4-di-
oxane is thought-provoking. This suggests that methanol and
1,4-dioxane establish much stronger interactions with Raney Ni
than other solvents do. With regard to 1,4-dioxane, this mole-
cule can indeed use both O sites to simultaneously interact
with the Ni surface, establishing a strong interaction with the
catalyst surface. Methanol, however, establishes another type
of interaction.
Studies on the adsorption of methanol on Ni(111) and
Ni(100) surfaces showed that the process involves the cleavage
of the OꢀH bond, generating a surface methoxide group and
a H adatom.[18] Likewise, ethanol adsorbs on Ni surfaces lead-
ing to a surface ethoxide group and a H adatom.[18c] As paral-
lels do exist between the adsorption of methanol and etha-
nol,[18b,c] these surface reactions should not be the only cause
of the unpredicted low activity of Raney Ni in methanol, but
the fact that methanol covers the catalyst surface to a larger
extent than ethanol, does. Indeed, Saleh et al.[18b,c] showed that
Ni films take up 1.9 times more methanol than ethanol in the
adsorption process. Hence, it is reasonable to assume that the
surface coverage by alkoxy groups decreases with the size of
the alcohol chain. As a result, a long-chain linear alcohol
should leave more active sites available than methanol does.
Apparently, it is because of this that Raney Ni displays higher
activity in 1-butanol than in methanol although both solvents
have similar Lewis basicities.
Implications of solvent effects for the conversion of lignin
The solvent effects on the hydrogenolysis of diphenyl ether
demonstrate that the reaction medium is not a bystander in
the catalytic conversion. Usually, hydrogenolysis of lignin is car-
ried out in methanol, ethanol, or 1,4-dioxane.[1a,8–12,21b] Al-
though these solvents and/or their aqueous solutions can dis-
solve some types of technical lignins, they can decrease the
catalytic activity of Raney Ni. On the contrary, using solvents
that enable the maximum activity of Raney Ni (e.g., alkanes)
raises difficulties with regard to the solubility of lignin. These
facts pose, thus, a key question: Is a compromise between cata-
lytic activity and substrate solubility required for the efficient con-
version of lignin?
Remarkably, better conversions than those predicted from
the Lewis basicity are achieved in 2-propanol and in 2-butanol
(Table 4). To understand this positive deviation, the nature of
Table 5. Products of the hydrogenolysis of diphenyl ether, conducted in the absence of H2, with Raney Ni.[a]
Entry
Solvent
Conversion [%]
Selectivity [%]
2
3
4
5
6
7
4:6
5:7
S(4,6)/S(5,7)
1
2
3
4
5
6
methanol
2-propanol
2-butanol
Hex-F-2-PrOH
decaline
3.0
16.6
18.3
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
50.0
29.8
21.5
0
0
0
50.0
49.8
49.3
0
0
0
0
0
0
0
0
0
0
–
–
–
-
–
–
–
1.00
1.01
1.03
–
–
–
20.4
29.2
0
0
0
1.46
0.73
–
–
–
MCH
[a] Reaction conditions: diphenyl ether (2.9 mmol), dibutyl ether (internal standard, 0.78 mmol), Raney Ni 4200 (100 mg), and solvent (15 mL) at 908C for
2.5 h.
&6
&
www.chemsuschem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 13
ÝÝ
These are not the final page numbers!

Solvent Effects on the Hydrogenolysis of Diphenyl Ether with Raney Ni
In the hydrogenolysis of lignin, we chose organosolv lignin
as substrate for several reasons. First, this substrate is sulfur-
free, obviating the need for sulfur-resistant catalysts. Second,
the organosolv process is relatively easy to implement in the
laboratory.[26] In this study, lignin was extracted by subjecting
a suspension of Poplar pellets in ethanol-water (1:1, v/v) to
1788C for 3.3 h. This led to a red–brownish liquor comprising
soluble lignin fragments with low apparent molecular weight
(Mw =1320 Da, Mn =750 Da). In these fragments, much of the
native structure of lignin is retained in the form of b-O-4 inter-
unit linkages, as reported elsewhere.[3a,31] A red–brownish
powder was obtained after removal of the solvent mixture. Al-
ternatively, lignin can be selectively precipitated by increasing
the water content in the mixture. This is a valuable method to
separate lignin from eventual water-soluble hemicellulose frag-
ments.[29] Third, organosolv lignin is soluble in 50% (v/v) aque-
ous solutions of lower alcohols (e.g., methanol, ethanol, 2-
propanol) and in cyclic ethers (e.g., THF, 1,4-dioxane).[3a] This is
often regarded as an advantageous feature while considering
the use of a solid catalyst in the reaction.[1,5]
With regard to the solubility of the organosolv lignin in 2-
propanol and in MCH, this substrate was slightly soluble in 2-
propanol (4 gLꢀ1) and practically insoluble in MCH (0.56 gLꢀ1)
at 258C. We detected, however, that subjecting dispersions of
the lignin to 3008C for 8 h, in the absence of catalyst, led to
80% solubilization of lignin in 2-propanol and 42%, in MCH.
At this point, a distinction between diphenyl ether and orga-
nosolv lignin should be made. While the model compound
shows high thermal stability due to the strong CꢀO aryl ether
bond,[6c] organosolv lignin can easily suffer thermolysis above
2508C due to the low thermal stability of the b-O-4 linkages.[21]
To assess, in a bit more detail, the solvent effects on the
liquid-phase thermolysis of lignin in absence of Raney Ni, sus-
pensions of organosolv lignin in MCH, 2-propanol, or methanol
were heated to 3008C within 0.5 h and immediately quenched
upon reaching this temperature. The soluble products were
analyzed by gel permeation chromatography (GPC) in THF.
Figure 4 displays the diode-array UV/Vis detector response for
the eluting species.
Figure 4. GPC analysis of the thermally treated lignin in a) methanol, b) 2-
propanol, c) MCH, and of d) organosolv lignin. The apparent molecular
weight (MW) is given relative to polystyrene standards (indicated by ‘+’ on
the bottom chromatogram).
with Raney Ni in 2-propanol and in MCH at 200, 250, or 3008C,
under 70 bar H2 (RT) for 8 h. Table 6 summarizes the results
achieved in the hydrogenolysis of lignin at different tempera-
tures. The conversion of lignin is expressed as the solubiliza-
tion of the substrate in the reaction medium.
Table 6. Hydrogenolysis of lignin, conducted in 2-propanol or MCH, in
the presence of Raney Ni at different temperatures under 70 bar H2 (RT)
for 8 h.
Figure 4 displays that the thermolysis of lignin is also affect-
ed by the solvent. Overall, the UV-absorption spots at around
275 nm indicate the presence of several phenolic species in
the solution, produced by heating the suspension of lignin in
MCH, 2-propanol or methanol up to 3008C. However, the
higher retention time of the products, obtained by the ther-
molysis in MCH, shows that smaller phenolic fragments are
formed in MCH than in 2-propanol or methanol. In fact, most
of the soluble products obtained in MCH possess an apparent
molecular weight (MW) lower than 300 Da, while those formed
in 2-propanol or methanol are mostly in the region between
300 and 1000 Da.
Solvent
T
[8C]
Solubilization of
lignin [%]
2-propanol
200
250
300
200
250
300
67
83
91
27
61
81
MCH
Table 6 clearly shows that high conversion of lignin is only
achieved at 3008C. There are several possible explanations for
this result. First, the thermolysis of lignin does not happen to
a large extent below 2508C.[21] This process apparently regu-
lates the amount of soluble fragments that Raney Ni is capable
of converting. Second, both 2-propanol and MCH are supercrit-
ical fluids under the reaction conditions at 3008C. Accordingly,
This clearly shows that soluble phenolic fragments can be
released into the reaction medium depending on the reaction
temperature.[23c] Therefore, this parameter is of central impor-
tance in reaching the compromise between catalytic activity
and substrate solubility. Accordingly, to find the optimal tem-
perature for the conversion of lignin, we carried out reactions
ChemSusChem 0000, 00, 1 – 13
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
&
7
&
ÞÞ
These are not the final page numbers!

X. Wang, R. Rinaldi
the mass transport between the bulk fluid and the catalyst sur-
face is highly improved.[22] In addition, H2 becomes extremely
soluble in the reaction medium under supercritical conditions,
contributing as well to the reaction achieving high conver-
sions.[22] Third, the presence of alcohol groups in the lignin
fragments poses a problem for the conversion of the frag-
ments with Raney Ni. Table 7 shows that the aromatic alcohols
possess higher DN’s than phenols and aromatic ethers. Conse-
quently, the highly basic functional groups of the lignin frag-
ments may also compete with ether linkages for the adsorp-
tion, reducing the activity of Raney Ni for hydrogenolysis and
hydrogenation.
from the experiments in the absence of catalyst. With Raney
Ni, the soluble fragments promptly undergoes hydrogenolysis
and hydrogenation.[21b] This suppresses the recombination of
the fragments.[23] The visual appearance of the product solu-
tions indicates, however, that the conversion of the fragments
in MCH happened quicker than in 2-propanol. In fact, the reac-
tion conducted in 2-propanol led to orange to light-yellow
product solutions throughout the reaction course, while the re-
action in MCH resulted in colorless solutions.
To gain more insight into the solvent effects on the reaction
with lignin, the samples were analyzed by comprehensive two-
dimensional GCꢀGC–MS. This technique has been applied to
the analysis of many complex mixtures (e.g., petrochemi-
cals).[24–25] However, it has never been used in the analysis of
lignin products from liquid-phase reactions. Figure 6 compares
the GCꢀGC–MS images of the reactions in a) 2-propanol,
b) MCH, and c) methanol, with Raney Ni at 3008C under 70 bar
H2 (RT) for 8 h.
Table 7. Donor numbers of some lignin model compounds.[15]
Compound
DN [kJmolꢀ1
]
benzyl alcohol
3-phenylpropanol[a]
phenol
96.6
94.3
46.2
58.7
79.8
23.1
The reaction carried out in 2-propanol achieved 91% conver-
sion, resulting in a complex product mixture. In this mixture, of
the more than 400 peaks detected in the image (Figure 6a),
we promptly identified 89 compounds by comparing the EI-MS
spectra to the NIST library. They are alkanes (particularly, cyclic
alkanes), monocyclic alcohols, monocyclic ketones—grouped
as saturates; and alkenes—grouped as unsaturated com-
pounds. Hydrogenolysis of lignin in MCH achieved 81% con-
version and had the selectivity markedly shifted towards satu-
rated compounds, such as monocyclic and polycyclic alkanes
and cyclic alcohols (Figure 6b). The reaction in methanol was
performed due to the importance of this solvent in lignin
chemistry. Surprisingly, the reaction achieved 86% conversion
and led mostly to phenols (Figure 6c). In addition, a product
containing a diaryl ether linkage was detected. These results
indicate that methanol inhibits the hydrogenating activity of
Raney Ni, turning the material into a highly selective catalyst
for the hydrogenolysis of the aryl alkyl ether linkages convert-
ing lignin into phenols.
m-cresol[a]
dibenzyl ether
diphenyl ether[a]
[a] DN values estimated by binary correlation of parameters: DN [kcal
molꢀ1]=0.25+40.4b, as described in Ref. [15].
To evaluate the solvent effects on the hydrogenolysis of
lignin with Raney Ni at 3008C, reactions in 2-propanol and in
MCH were conducted at different reaction times. Figure 5 dis-
plays the temporal evolution of the solubilization of lignin in
2-propanol as well as in MCH, and the visual appearance of
the reaction media after filtering off the catalyst and residual
lignin. Upon reaching 3008C (t=0 h), 76 and 52% of lignin is
solubilized in 2-propanol and in MCH, respectively. In addition,
the product solutions were much lighter in color than those
To estimate the composition of the product solutions, in
terms of saturated and unsaturated compounds, GCꢀGC–FID
images were used. To make a semi-quantitative assessment of
the content of each product class, estimated response factors
were used. Table 8 summarizes the results of hydrogenolysis of
lignin in 2-propanol, MCH, and methanol. As already noticed in
Figure 6, the solvent drastically affects the distribution of prod-
ucts obtained from lignin. Remarkably, just by changing the
solvent from methanol to MCH, the selectivity was shifted
from mostly phenols (unsaturates) to saturated cyclic alkanes
and alcohols (saturates).
Table 8 shows carbon balances ranging from 43 to 55%. This
is not a bad result considering the complexity of the mixtures,
the conservative assumptions taken in the product quantifica-
tion, and the product volatility. With regard to the quantifica-
tion of the product contents in the mixture, we did not consid-
er in the analysis the products that could be also produced by
solvent decomposition. Accordingly, cyclohexane and MCH,
which are also two major products detected in the reaction
conducted in 2-propanol, were not taken into consideration in
Figure 5. Temporal solubilization of lignin in a) 2-propanol, and b) MCH. The
inset shows the visual appearance of the product solutions at the indicated
times. Reaction conditions: lignin (0.5 g), solvent (15 mL), Raney Ni (150 mg),
7 MPa H2 (RT).
&8
&
www.chemsuschem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 13
ÝÝ
These are not the final page numbers!

Solvent Effects on the Hydrogenolysis of Diphenyl Ether with Raney Ni
the quantification of the prod-
ucts obtained from the reaction
in MCH. Second, gaseous prod-
ucts (e.g., methane) were not
considered in the carbon bal-
ance because they can also be
a result of solvent decomposi-
tion at 3008C.[11] A simple sub-
traction of solvent decomposi-
tion products from the lignin
products is not a reliable proce-
dure because Raney Ni displays
lower activity in the presence of
lignin. Finally, and most impor-
tantly, comprehensive two-di-
mensional GCꢀGC technique an-
alyzes only volatile products. In
the current method, compounds
higher than C25 (Mꢂ350 Da)
could not be analyzed.
Conclusions
Solvents are not bystanders in
the hydrogenolysis of diphenyl
ether and lignin with Raney Ni.
They affect the activity as well as
control the selectivity of the cat-
alyst. Accordingly, the following
conclusions can be drawn from
the present study:
1. The Lewis basicity of the
solvent, as expressed by the
donor number, is an important
parameter in the reaction perfor-
mance with Raney Ni. Basic sol-
vents (e.g., methanol, 1,4-diox-
ane, and THF) markedly reduce
the catalytic activity towards hy-
drogenation of aromatic prod-
ucts. Nonetheless, the activity to-
wards hydrogenolysis of the
ether bonds, is not affected in
the same proportion. As a result,
performing the reaction in meth-
anol was demonstrated to effec-
tively suppress the hydrogena-
tion of aromatic products, im-
proving the selectivity for phe-
nols from lignin.
2. If saturates are the desired
products from lignin, solvents
possessing no Lewis basicity are
the best choices.
Figure 6. GCꢀGC-MS images showing some products obtained from the hydrogenolysis carried out in a) 2-propa-
nol, b) MCH, and c) methanol, in the presence of Raney Ni at 3008C under 70 bar H2 (RT) for 8 h. The chromato-
graphic internal standard (dibutyl ether) was added into the mixture of products after reaction. Images (a) and (b)
were obtained from mixtures without addition of internal standard (ISTD).
3. Thermolysis of lignin in the
liquid phase is a key step that
brings aromatic fragments into
ChemSusChem 0000, 00, 1 – 13
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
&
9
&
ÞÞ
These are not the final page numbers!

X. Wang, R. Rinaldi
lene (cis/trans mixture, Aldrich, 99%) were used as-received. Raney
Ni 4200 (Aldrich) was washed with degassed water, dried under
argon, and stored in a glove box.
Table 8. Results of hydrogenolysis of lignin in 2-propanol, MCH, or meth-
anol, in the presence of Raney Ni, at 3008C under 70 bar H2 (RT) for 8 h.
Solvent
Conversion [%] Selectivity [%]
Saturates Unsaturates
Carbon balance [%]
2-propanol 91
60
40
55
45
43
Organosolv lignin
MCH
81
86
71 29
24 76
Poplar wood (500 g, 2 mm pellets, Fa. J. Rettenmaier & Sçhne) was
suspended in 3 L of ethanol/water (1:1, v/v) in a 5 L autoclave
equipped with an overhead mechanical stirrer. The suspension was
heated from 258C to 1788C at 1.58Cminꢀ1 under mechanical stir-
ring. The autogenous pressure at 1788C was 14.9 bar. The suspen-
sion was further processed at 1788C for 3.3 h. Next, the mixture
was left to cool to 258C. A red–brownish solution was obtained
after filtering off the wood fibers. The solvent was removed at
408C using a rotavap. The red–brownish solid residue was dried
under high vacuum at 258C for 2 days. From 500 g poplar wood,
116 g of organosolv lignin were obtained. Elemental analysis: C,
59.0%; H, 6.2%; N, 0.1%, O, 34.7%. These values are in agreement
with values found for technical hardwood lignins.[2] GPC characteri-
zation: Mw 1320 Da; Mn 750 Da (relative to polystyrene standards in
THF).
methanol
solution and, consequently, enables the action of solid cata-
lysts. This process usually starts above 2508C.[21] At this tem-
perature, conversions of lignin higher than 50% were achieved
either in MCH or 2-propanol. As lignin thermally decomposes,
releasing soluble fragments into the reaction medium, the hy-
drogenolysis can be performed even in solvents in which
lignin is insoluble. 4. Secondary alcohols can also serve as
source of hydrogen for the reaction. The hydrogenolysis of di-
phenyl ether in 2-propanol and in 2-butanol was positively in-
fluenced by an H transfer reaction from the solvent to the sub-
strate. The reactions with lignin shown that 2-propanol induces
the formation of a mixture of products comprising cyclic alka-
nes, monocyclic alcohols, monocyclic ketones, and some al-
kenes.
Hydrogenolysis of diphenyl ether
Diphenyl ether (2.9 mmol), dibutyl ether (internal standard - ISTD,
0.78 mmol), Raney Ni 4200 (100 mg), and solvent (15 mL) were
placed in a batch reactor (30 mL) under argon (glove box). After
purging the reactor with H2, the reaction vessel was loaded with
50 bar H2 (258C). The experiments were performed at 908C for
2.5 h (RT to 908C: 10 min) using overhead mechanical stirring
(300 rpm). Under these conditions dibutyl ether is stable in the re-
action media, as evaluated by us and by others.[27] The reactions
were conducted in methanol, ethanol, 2-propanol, 1,1,1,3,3,3-hexa-
fluoro-2-propanol, 2-butanol, tert-butanol, dioxane, tetrahydrofur-
an, methylcyclohexane, 2-methyltetrahydrofuran, n-heptane, or
decaline. The products were analyzed by GC-FID and GC-MS.
5. Lignin is a highly functionalized polymer. Its own organic
functionalities (e.g. alcohol groups) occupy the active sites re-
quired for hydrogenolysis and hydrogenation. As a result, part
of the catalytic activity is inhibited by the substrate. Studies
are in progress to clarify the effects of lignin itself on the per-
formance of Raney Ni.
6. A comprehensive two-dimensional GCꢀGC technique has
been successfully applied to the characterization of lignin
products. This technique is capable of (i) resolving hundreds of
peaks, and (ii) improving the sensitivity. Both features are of
great importance for lignin chemistry as a large number of
products in low concentrations are produced by hydrogenoly-
sis due to the no regularity of lignin structure.
Hydrogenolysis of lignin
In summary, the rational choice of the reaction medium is
extremely important in the design of catalytic systems for hy-
drogenolysis of lignin. Surely, the solvent effects, discussed in
this report, are closely connected with the metal used in the
catalysis, that is, Raney Ni. Nonetheless, the findings from this
study make several contributions to the current literature. Fur-
ther systematic studies are needed for rationalizing the solvent
effects on a comprehensive list of solid catalysts for hydroge-
nolysis and hydrogenation of lignin.
Organosolv lignin (500 mg), Raney Ni 4200 (150 mg, Aldrich), and
15 mL solvent (methylcyclohexane or 2-propanol) were placed in
a batch reactor under argon (glove-box). After purging the reaction
vessel with H2, the reactor was loaded with 70 bar H2 (258C). The
reaction was performed at 3008C for 8 h (the reaction time was re-
corded upon reaching 3008C) under overhead mechanical stirring.
After the reaction, dibutyl ether was added into the reaction mix-
ture, as internal standard, for the GCꢀGC-FID analyses. This proce-
dure is required because dibutyl ether may eventually undergo hy-
drogenolysis in the presence of Raney Ni under H2 pressure at
3008C.[27] The reaction mixture was filtered. The solid residue (cata-
lyst and processed lignin) was washed with reaction solvent, dried
at 608C under vacuum, and weighed. The conversion (or solubiliza-
tion) of lignin was determined as described by Equation (3):
Experimental Section
Chemicals
ꢀ
ꢁ
Diphenyl ether (Aldrich, 99%), dibutyl ether (Aldrich, 99%), metha-
nol (Aldrich, 99.5%), ethanol (Aldrich, 99.8%), 2-propanol (Aldrich,
99.9%), 1,1,1,3,3,3-hexafluoro-2-propanol (Aldrich, 99%), 1-butanol
(Aldrich, 99.8%), 2-butanol (Aldrich, 99.7%), tert-butanol (Aldrich,
99.3%), tetrahydrofuran (Aldrich, 99.9%), 2-methyltetrahydrofuran
(Aldrich, 99.9%), dioxane (Riedel-de Haꢂn, 99.5%), methylcyclohex-
ane (Aldrich, 99%), n-heptane (Aldrich, 99%), decahydronaphtha-
mresidue ꢀ mcatalyst
Clignin ð%Þ ¼ 1 ꢀ
ꢃ 100
ð3Þ
mligning
where Clignin is the conversion (or solubilization) of lignin, mresidue is
the weight of solid residue, mcatalyst is the weight of catalyst, and
mlignin is the weight of lignin.
&10
&
www.chemsuschem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 13
ÝÝ
These are not the final page numbers!

Solvent Effects on the Hydrogenolysis of Diphenyl Ether with Raney Ni
[3] a) J. E. Holladay, J. F. White, J. J. Bozell, D. Johnson, Pacific Northwest Na-
tional Laboratory, Richland, WA., 2007. Available from http://
www.pnl.gov/main/publications/external/technical_reports/PNNL-
16983.pdf (accessed April 2012); b) G. W. Huber, S. Iborra, A. Corma,
Chem. Rev. 2006, 106, 4044–4098.
[4] J. O. Metzger, C. Bicke, O. Faix, W. Tuszynski, R. Angermann, M. Karas, K.
Strupat, Angew. Chem. 1992, 104, 777–779; Angew. Chem. Int. Ed. Engl.
1992, 31, 762–764.
GC analysis of the diphenyl ether products
The samples were analyzed by GC using a Shimadzu QP2010 Plus
gas chromatograph, equipped with an Rxi-1ms column (30 m). The
temperature program started with an isothermal step at 408C for
6 min. Next, the temperature was increased, at 308C minꢀ1, to an-
other isothermal step at 1608C for 1 min. Again, the temperature
was increased to 2008C at 308Cminꢀ1. From this point on, the tem-
perature was increased, at 108Cminꢀ1, until reaching the last iso-
thermal step at 2408C for 1 min. The retention times (min) were:
dibutyl ether (IS) 10.9, benzene 6.2, cyclohexane 6.4, cyclohexanol
10.8, phenol 12.1, dicyclohexyl ether 15.6, diphenyl ether 15.9 and
cyclohexyl phenyl ether 16.1. The components were quantified
using calibration curves.
[5] a) R. Rinaldi, F. Schꢃth, Energy Environ. Sci. 2009, 2, 610–626; b) R. Rinal-
di, F. Schꢃth, ChemSusChem 2009, 2, 1096–1107.
[6] a) R. Parthasarathi, R. A. Romero, A. Redondo, S. Gnanakaran, J. Phys.
Chem. Lett. 2011, 2, 2660–2666; b) J. M. Younker, A. Beste, A. C. Buchan-
an, III, ChemPhysChem 2011, 12, 3556–3565; c) E. Dorrestijn, L. J. J.
Laarhoven, I. W. C. E. Arends, P. Mulder, J. Anal. Appl. Pyrolysis 2000, 54,
153–192.
[7] a) E. E. Harris, H. Adkins, Pap. Trade J. 1938, 236–238; b) E. E. Harris, J.
D’Ianni, H. Adkins, J. Am. Chem. Soc. 1938, 60, 1467–1470.
[8] a) J. M. Pepper, W. Steck, Can. J. Chem. 1963, 41, 2867–2875; b) J. M.
Pepper, Y. W. Lee, Can. J. Chem. 1970, 48, 477–479.
[9] M. Nagy, K. David, G. J. P. Britovsek, A. J. Ragauskas, Holzforschung 2009,
63, 513–520.
[10] a) N. Yan, C. Zhao, P. J. Dyson, C. Wang, L. Liu, Y. Kou, ChemSusChem
2008, 1, 626–629; b) N. Yan, Y. A. Yuan, R. Dykeman, Y. A. Kou, P. J.
Dyson, Angew. Chem. 2010, 122, 5681–5685; Angew. Chem. Int. Ed.
2010, 49, 5549–5553.
[11] a) T. D. Matson, K. Barta, A. V. Iretskii, P. C. Ford, J. Am. Chem. Soc. 2011,
133, 14090–14097; b) K. Barta, T. D. Matson, M. L. Fettig, S. L. Scott, A. V.
Iretskii, P. C. Ford, Green Chem. 2010, 12, 1640–1647; c) G. S. Macala,
T. D. Matson, C. L. Johnson, R. S. Lewis, A. V. Iretskii, P. C. Ford, ChemSu-
sChem 2009, 2, 215–217.
GCꢀGC analysis of lignin products
The reaction mixtures were analyzed using 2D GCꢀGC-MS (1st
column: Rxi-1 ms 30 m, 0.25 mm ID, df 0.25 mm; 2nd column:
BPX50, 1 m, 0.15 mm ID, df 0.15 mm) in a GC-MS 2010 Plus (Shi-
madzu) equipped with a ZX1 thermal modulation system (Zoex).
The temperature program started with an isothermal step at 308C
for 5 min. Next, the temperature was increased from 308C to
3008C at 5.28C minꢀ1. The program finished with an isothermal
step at 3008C for 5 min. The modulation applied for the compre-
hensive GCꢀGC analysis was a hot jet pulse (400 ms) every
9000 ms. The 2D chromatograms were processed with GC Image
software (Zoex). The products were identified by a search of the
MS spectrum with the MS library NIST 08 and NIST 08 s with a re-
tention index allowance of ꢁ100. The semi-quantification of the
products was performed using the GCꢀGC-FID images.
[12] a) A. G. Sergeev, J. F. Hartwig, Science 2011, 332, 439–443; b) J. M. Nich-
ols, L. M. Bishop, R. G. Bergman, J. A. Ellman, J. Am. Chem. Soc. 2010,
132, 12554–12555.
[13] a) H. Takagi, T. Isoda, K. Kusakabe, S. Morooka, Energy Fuels 1999, 13,
ˇ
ˇ
´
´
1191–1196; b) M. Bejblovꢄ, P. Zꢄmostny, L. Cerveny, J. Cejka, Appl. Catal.
A 2005, 296, 169–175; c) N. M. Bertero, A. F. Trasarti, C. R. Apesteguꢅa,
A. J. Marchi, Appl. Catal. A 2011, 394, 228–238.
GPC analysis
[14] U. K. Singh, M. A. Vannice, Appl. Catal. A 2001, 213, 1–24.
[15] Y. Marcus, Chem. Soc. Rev. 1993, 22, 409–416.
The GPC analyses were performed at 608C on a Perkin–Elmer HPLC
200 equipped with 4 columns (2ꢀTSKgel Super HZ1000, TSKgel
Super HZ2000 and TSKgel Super HZ3000, 4.6 mmꢀID 15.0 cm),
and using inhibitor-free THF as eluent (0.4 mLminꢀ1, Aldrich). Solu-
tions containing 2 mgmLꢀ1 of the analyte were injected. For detec-
tion, a diode-array detector was used. The system was calibrated
with polystyrene standards (200 to 60000 Da, Aldrich).
[16] a) C. Zhao, J. Y. He, A. A. Lemonidou, X. B. Li, J. A. Lercher, J. Catal. 2011,
280, 8–16; b) C. Zhao, Y. Kou, A. A. Lemonidou, X. B. Li, J. A. Lercher,
Chem. Commun. 2010, 46, 412–414; c) C. Zhao, Y. Kou, A. A. Lemoni-
dou, X. B. Li, J. A. Lercher, Angew. Chem. 2009, 121, 4047–4050; Angew.
Chem. Int. Ed. 2009, 48, 3987–3990.
[17] M. Boudart, D. J. Sajkowski, Faraday Discuss. 1991, 92, 57–67.
[18] a) Y.-H. Zhou, P.-H. Lv, G.-C. Wang, J. Mol. Catal. A 2006, 258, 203–215;
b) D. Al-Mawlawi, J. M. Saleh, J. Chem. Soc. Faraday Trans. 1 1981, 77,
2965–2976; c) D. Al-Mawlawi, J. M. Saleh, J. Chem. Soc. Faraday Trans.
1 1981, 77, 2977–2988.
Acknowledgements
[19] L. J. Shorthouse, A. J. Roberts, R. Raval, Surf. Sci. 2001, 480, 37–46.
[20] F. Alonso, P. Riente, M. Yus, Acc. Chem. Res. 2011, 44, 379–391.
[21] a) T. Watanabe, H. Kawamoto, S. Saka, Holzforschung 2009, 63, 424–
430; b) J. D. Benigni, I. S. Goldstein, J. Polym. Sci. Part C 1971, 36, 477–
488.
[22] E. P. Martins, D. A. G. Aranda, F. L. P. Pessoa, J. L. Zotin, Braz. J. Chem.
Eng. 2000, 17, 361–366.
[23] a) V. M. Roberts, R. T. Knapp, X. B. Li, J. A. Lercher, ChemCatChem 2010,
2, 1407–1410; b) V. M. Roberts, V. Stein, T. Reiner, A. Lemonidou, X. B. Li,
J. A. Lercher, Chem. Eur. J. 2011, 17, 5939–5948; c) J. Zakzeski, B. M.
Weckhuysen, ChemSusChem 2011, 4, 369–378.
The funds for this research have been provided by the Alexander
von Humboldt Foundation in the framework of the Sofja Kova-
levskaja Award 2010 endowed by the German Federal Ministry of
Education and Research. This work was performed as part of the
Cluster of Excellence “Tailor-Made Fuels from Biomass”, which is
funded by the Excellence Initiative of the German federal and
state governments to promote science and research at German
universities.
[24] L. Mondello, P. Q. Tranchida, P. Dugo, G. Dugo, Mass Spectrom. Rev.
2008, 27, 101–124.
[25] J. Harynuk (2009, June 18). Comprehensive Two-dimensional Gas Chro-
matography (GCxGC). SciTopics. Available from http://www.scitopics.-
com/Comprehensive_Two_dimensional_Gas_Chromatography_
GCxGC.html (accessed April 2012).
[26] A. Johansson, O. Aaltonen, P. Ylinen, Biomass 1987, 13, 45–65.
[27] E. M. van Duzee, H. Adkins, J. Am. Chem. Soc. 1935, 57, 147–151.
[28] a) G. S. Chandler, W. H. F. Sasse, Aust. J. Chem. 1963, 16, 20–30;
b) G. D. F. Jackson, W. H. F. Sasse, J. Chem. Soc. 1962, 3746–3751.
Keywords: biomass · catalysis · nickel · renewable resources ·
solvent effects
[1] a) J. Zakzeski, P. C. A. Bruijnincx, A. L. Jongerius, B. M. Weckhuysen,
Chem. Rev. 2010, 110, 3552–3599; b) F. G. Calvo-Flores, J. A. Dobado,
ChemSusChem 2010, 3, 1227–1235.
[2] D. Fengel, G. Wegener, Wood: Chemistry Ultrastructure Reactions, Walter
De Gruyter, Berlin 1984.
ChemSusChem 0000, 00, 1 – 13
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
&
11
&
ÞÞ
These are not the final page numbers!

X. Wang, R. Rinaldi
[29] P. Kçll, J. Metzger, B. Brçnstrup, Makromol. Chem. Rapid Commun. 1982,
3, 365–369.
[31] F. Abdelkafi, H. Ammar, B. Rousseau, M. Tessier, R. El Gharbi, A. Fradet,
Biomacromolecules. 2011, 12, 3895–3902.
[30] C. Laurence, J.-F. Gal, Lewis Basicity and Affinity Scales, Data and Mea-
surement, Wiley, United Kingdom, 2010.
Received: January 16, 2012
Revised: February 20, 2012
Published online on && &&, 0000
&12
&
www.chemsuschem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 13
ÝÝ
These are not the final page numbers!

FULL PAPERS
X. Wang, R. Rinaldi*
&& – &&
Solvent Effects on the Hydrogenolysis
of Diphenyl Ether with Raney Nickel
and their Implications for the
Conversion of Lignin
Solvents are not bystanders: The Lewis
basicity of the solvent very much affects
the catalytic activity of Raney nickel, so
in nonbasic solvents the catalyst is an
extremely active catalyst for hydroge-
nolysis and hydrogenation. In basic sol-
vents, however, Raney nickel is a less
active, but much more selective catalyst
for hydrogenolysis while preserving the
aromatic products.
ChemSusChem 0000, 00, 1 – 13
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
&13
&
ÞÞ
These are not the final page numbers!