How Inter- and Intramolecular Reactions Dominate the Formation of Products in Lignin Pyrolysis

Article abstract of DOI:10.1002/chem.201700639

One of the key challenges in renewable chemical production is the conversion of lignin, especially by fast pyrolysis. The complexity of the lignin pyrolysis process has hindered the elucidation of the mechanism, inhibiting further industrial implementation. By combining pyrolysis of model compounds (4-phenoxyphenol and 2-methoxy-phenoxybenzene) with lignin bond characteristics both under vacuum and under realistic pressure conditions, the roles of inter- and intramolecular reactions were established. On the one hand, the stable 4-O-5 ether bond enables, without breaking, C?C bond formation and even directly forms naphthalene depending on the position and type of the substituent. p-Benzoquinone intermediates, on the other hand, are highly unstable at ambient pressure and directly decompose into coke and carbon monoxide. The system pressure (radical concentration) plays a crucial role in the dominant reaction mechanism by initiating intramolecular reactions, interfering with intramolecular reactions. H-transfer and recombination reactions suppress the decarbonylation of phenoxy radicals, thus yielding a very different product distribution.

Full text of DOI:10.1002/chem.201700639

A Journal of
Accepted Article
Title: How inter- and intramolecular reactions dominate formation of
products in lignin pyrolysis
Authors: Victoria B.F. Custodis, Patrick Hemberger, and Jeroen A. van
Bokhoven
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201700639
Link to VoR: http://dx.doi.org/10.1002/chem.201700639
Supported by

10.1002/chem.201700639
Chemistry - A European Journal
FULL PAPER
How inter- and intramolecular reactions dominate formation of
products in lignin pyrolysis
Victoria B. F. Custodis[a,c], Patrick Hemberger[b] and Jeroen A. van Bokhoven[a,c,*]
Abstract: One of the key challenges in renewable chemical
production is the conversion of lignin, especially by fast pyrolysis.
The complexity of the lignin pyrolysis process has prevented the
identification of the mechanism, inhibiting further industrial
implementation. By combining pyrolysis of model compounds (4-
phenoxyphenol and 2-methoxy-phenoxybenzene) with lignin bonds
characteristic under vacuum and realistic pressure conditions, the
roles of inter- and intramolecular reactions were established. On the
one hand the stable 4-O-5 ether bond enables, without breaking, C-
C bond formation and even directly forms naphthalene depending on
the position and type of the substituent. P-benzoquinone
intermediates, on the other hand are highly unstable at ambient
pressure and directly decompose into coke and carbon monoxide.
The system pressure (radical concentration) plays a crucial role in
the dominant reaction mechanism by initiating intramolecular
reactions, interfering with intramolecular reactions. H-transfer and
recombination reactions suppress the decarbonylation of phenoxy
radicals, thus yielding a very different product distribution.
depolymerization and enable the study of a single bond or
building block. Due to the complexity and diversity of the lignin
many chemical reactions and processes occur simultaneously
during lignin treatment, so that specific behavior is generally
hard to correlate to one chemical/physical property of the
lignin.[5,6] Model compounds, such as guaiacol and diphenylether,
resemble a common intermediate (and building block) and the 4-
O-5 bond.
Fast pyrolysis, in particular fast pyrolysis, is a promising method
to depolymerize lignin with all its different bonds to obtain
valuable chemicals.[7] Here, the lignin polymer is thermally
broken down and yields a mix of oxygen-containing molecules,
mostly
phenols.
Thermal-induced
decomposition
of
macromolecules results from the formation of radicals, which
can recombine easily and, in times form undesired char. The
liquid product of fast pyrolysis is bio-oil, which is unstable and
corrosive. A great deal of research has been devoted to the
improvement of the quality of the products, as well as to a
decrease in char by catalytic fast pyrolysis (CFP).[7–12]
Both pyrolysis and catalytic pyrolysis depend on the thermal
depolymerization of the lignin polymer. The mechanism of
depolymerization and monomer formation is crucial to predict
and ultimately to influence product selectivity. The quantitative
Introduction
Lignin is one of three main components lignocellulosic biomass
and is regarded to be the largest renewable biofuel in Europe
(2007).[1] Pulp and paper mills produce a surplus of energy by
simple incineration. The specific structure of lignin makes it a
possible source of chemicals. It has high potential as a
feedstock for bulk and fine chemicals, especially aromatic and
phenolic compounds.[2],[3] Lignin is polymerized from three
phenolic monomers, resulting in a complex and irregular
polymer that is very hard to characterize.[4] Scheme 1 is a
sample sketch of the lignin polymer together with the three
building blocks: coniferyl, sinapyl and coumaryl units. Prominent
bonds are β-O-4 (aryl ether), α-O-4 (aryl ether), 4-O-5 (aromatic
ether), 5-5 (C-C aromatic) or spyro-bonds (C-C). The lignin
structure varies depending on the bioresource and separation
method as well as storage time. Model compounds are
commonly studied to understand chemical upgrading and
and qualitative characterization of the radical depolymerization
[13]
of the lignin polymer is very difficult
are investigated instead.
Thus, typical monomers
Many
model
compounds
have
been
studied:
phenol,[16,17]
phenylphenethylether,[14],[15]
diphenylether,
anisole,[18–20] and guaiacol.[21,22] These model compounds
represent a typical feature of lignin, such as frequently occurring
bonds, building blocks or a component of the pyrolysis oil.[15,23]
Brebu et al.[24] give a summary of many model compound
studies. Phenylphenethylether (PPE) is the model compound
mimicking the β-O-4 bond and its decomposition mechanism
depends strongly on the temperature. Extensive studies have
shown that there are two dominating mechanisms, the 6-centred
retro-ene and the Maccoll elimination,[25,26] both of which lead to
phenol and styrene as end-products.[27,28] The β-O-4 bond is
only present in small quantities in most industrial lignins, since it
can be easily hydrolysed during the separation process.[29]
Phenol, a very simple model compound, forms cyclopentadienyl
radicals by decarbonylation.12,17 Guaiacol also decarbonylates
after phenoxy formation under collision reduced-conditions.
Collision-reduced conditions exist when the sample
concentration is very low and the vacuum high, so that direct
radicals are detectable, because they are not quenched by
bimolecular reactions.[31,32]
[a]
Victoria. B. F. Custodis and Prof. Dr. Jeroen. A. van Bokhoven,
Institute for Chemical and Bioengineering, Department of Chemistry
and Applied Biosciences, ETH Zurich, HCI E 127, Vladimir-Prelog-
Weg 1, 8093 Zurich, Switzerland.
jeroen.vanbokhoven@chem.ethz.ch
[b]
[c]
Dr. Patrick. Hemberger, Laboratory for Femtochemistry and
Synchrotron Radiation, Paul Scherrer Institute, CH-5232 Villigen-
PSI, Switzerland
Prof. Dr. J. A. van Bokhoven, Laboratory for Catalysis and
Sustainable Chemistry, Paul Scherrer Institute, WLGA 135, 5232
Villigen, Switzerland.
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/chem.201700639
Chemistry - A European Journal
FULL PAPER
This study will focus on the most stable ether bond, the 4-O-5
bond, which is even more stable than some C-C bonds.[33]
Which primary radicals form and in which further reactions do
they participate? Aim is to control those intermediates and to
steer the reactions towards producing more valuable products.
on the kinetics of the decomposition mechanism, but also relies
strongly on the environment, especially in radical reactions.[32,35]
Ambient pressure enhances bimolecular reactions and in
particular H-transfer and -loss become important radical transfer
and formation steps, while the monomolecular decomposition at
high vacuum depends mainly on homolytic fission of the
weakest bond.
a)
OH
OH
OH
O
O
O
Results and Discussion
p-coumaryl building block
H-Unit
Coniferyl building block
G-Unit
Synapyl building block
S-Unit
Results
4-phenoxyphenol (PhOPhOH)
b)
HO
HO
coniferyl
O
Figure 1 shows the most common products detected upon fast
pyrolysis of 4-phenoxyphenol in the py-GC/MS setup. First
decomposition was observed at temperatures higher than
600 °C during pyrolysis. The first products were phenol,
benzene and dibenzofuran. The phenol selectivity decreases
quickly with temperature, while benzene selectivity increases
steadily. Other minor products above 800 °C are toluene,
O
HO
4-O-5
-O-4
O
O
O
sinapyl
O
O
O
OH
O
5-5
O
HO
O
O
OH
HO
HO
ß-O-4
O
O
O
HO
O
OH
OH
O
OH
OH
OH
O
O
styrene,
diphenylether,
indene,
hydroxyl-biphenyl
and
O
anthracene. Above 800 °C, a noticeable amount of biphenyl
forms, and at 850 °C naphthalene and more benzene forms. The
pyrolysis temperatures are considerably high, which is
necessary, because this compound evaporates very quickly.
Scheme 1. a) building blocks and b) lignin sample structure.
There are two possible approaches to studying the pyrolysis
mechanism of lignin. Top down: looking at all the radicals of the
lignin during pyrolysis,[13,34,35] or bottom up: focusing on one
model compound at a time.[31,32,36] Both approaches have
advantages and disadvantages. On the one hand, information is
incomplete, when only the integrated signal of all the different
components are determined, while on the other hand, many
side-reactions and other components are ignored in hope of
representing the major component. This conflict is particularly
evident in research on lignin due to its heterogeneity. We focus
on both approaches and their differences in order to identify
possible dependencies and connections.
OH
O
We employed two types of pyrolysis setups, which detect
isolated radicals as well as the stable end-products: py-GC/MS
at ambient pressure and iPEPICO at the VUV (x04db) beamline
of the Swiss Light Source under reactive collision-reduced
conditions. Py-GC/MS detects products of recombination,
rearrangement and stabilization. In contrast, in the iPEPICO
setup, the detection of molecules is so immediate and the
residence time in the reactor is so short (hundreds of
microseconds) that even radicals are detectable. The iPEPICO
Figure 1. Product selectivity of the main products of 4-phenoxyphenol
pyrolysis at different temperatures and at ambient pressure in py-GC/MS.
Figure 2 shows the relative signals (normalized to total peak
area) at each pyrolysis temperature under high vacuum at a
constant sample temperature and 10.5 eV ingoing photon
energy. The relative signal strength mirrors the mechanism in
terms of primary, secondary and tertiary products. At the lowest
temperature, at which the parent molecule decomposes, various
products emerge: m/z=108/109, m/z=93, m/z=80, m/z=77,
m/z=65 and m/z=52. According to their ionization energy and
TPE spectra at 760 °C the products are the hydroxyl-phenoxy
radical (109), p-benzoquinone (108), the phenoxy radical (93),
cyclopentatienone (80), the phenyl radical (77), cyclopentadienyl
radical (65) and butane-3yne (52). Some of these species are
known from the decomposition of guaiacol and could be
identified based on previous studies and recorded spectra.[31,32]
system determines each radical and intermediate by threshold
[37]
photoelectron spectra (TPES) isomer-specifically.
iPEPICO
yields the reaction mechanisms of the initial reactions and py-
GC/MS the stabilized products. This combination bridges the
pressure gap and enables identification of the role of reaction
conditions on mechanisms and possible products.
We determine the importance of hydroxyl and methoxy
substituents on the model dimers 4-phenoxyphenol and 1-
methoxy-2-phenoxybenzene and how they affect the
decomposition pathway and the breakup of the stable aromatic
ether bond. The radical environment, in our case the strong
dilution of the model compound vs. ambient pressure, is crucial
to take into account during the thermal degradation of lignin.
Previous studies have shown that selectivity not only depends
This article is protected by copyright. All rights reserved.

10.1002/chem.201700639
Chemistry - A European Journal
FULL PAPER
Above 750 °C, m/z=39, propargyl radicals, are also detected as
well as m/z=26, which is acetylene and is not visible in Figure 2,
since it ionizes only above 11.4 eV. At 820 °C, the signal at
m/z=186 (PhOPhOH) almost depleted, indicating complete
decomposition of the parent molecule. Additional measurements
at 13.9 eV and 14.1eV photon energy, at 760 °C, revealed also
the formation of carbon monoxide, m/z=28.
Figure 2c shows the fractional abundances (relative selectivites
by peak-areas) of the ionized products and intermediates. Due
to the different ionization cross-sections of each molecule the
signal intensity alone provides only limited quantitative
information. However, if the ionization energy and the inlet
stream are kept constant, the relative increase and decrease in
signal intensity identifies reaction sequences. As the parent
signal decreases, all products form at the same pyrolysis
temperature (640 °C). The relative increase of the
cyclopentadienyl radical and the butane-3-yne is the steepest,
followed by phenoxy and phenyl radicals. Above 750 °C phenyl
radicals and the benzene signals decrease, and benzyne
(m/z=76) is detected. In general, the phenyl and phenoxy
radicals as well as benzoquinone and benzene show similar
trends, while cyclopentadienon (m/z=80) and benzyne (m/z=76)
peak at higher temperatures.
Comparing the products in the iPEPICO experiment (Fig 2b)
there is a lack of benzoquinone (BQ) and benzenediol products
in py-GC/MS. Further investigations with pyrolysis of a mixture of
PhOPhOH and p-benzoquinone at ambient pressure showed
high conversion of the quinone (over 60% at 700°C compared to
8% BQ alone), while most of it converted to coke and trace
amounts (< 0.05 peak area %) of products. Copyrolysis led to an
increase in carbon monoxide up to 2 mol % compared to 0.04
mol % (PhOPhOH) and 0.16 mol % (BQ).
1-methoxy-2-phenoxybenzene (PhOPhOMe)
PhOPhOMe at ambient pressure (py-GC/MS) decomposes in a
different way to 2-phenoxy-phenol and dibenzofuran-2-ol. Figure
3 shows that the primary product is the corresponding hydroxide
2-phenoxyphenol. At very low conversion, this is the only
product besides recombination products and 2-hydroxyphenyl-
phenylmethanone. At higher temperatures selectivity of
dibenzofuranol and dibenzofuran increases, while that of 2-
phenoxyphenol decreases. Benzene, dibenzodioxin and phenols
are also detected, the latter being mainly phenol and trace
amounts of methyl-phenols. Even though 2-phenoxy-phenol is a
major intermediate in py-GC/MS, the other products do not
decompose in the same way as 4-phenoxyphenol.
OH
O
OH
O
O
OH
O
O
O
Figure 3. Product selectivity of 1-hydroxy-2-phenoxybenzene (PhOPhOMe
Figure 2. a) Temperature-dependent mass spectra in the iPEPICO-TOF setup
at a sample temperature of 90°C and an ionization beam energy of 10.5eV
and b) a detailed mass spectrum at 760°C. c) Thermal breakdown diagram
with primary and secondary products using the iPEPICO setup. Data is on the
integration of the mass spectra from Tsample = 90°C and 10.5 eV
pyrolysis in py-GC/MS at ambient pressure) as dependent on the temperature.
This article is protected by copyright. All rights reserved.

10.1002/chem.201700639
Chemistry - A European Journal
FULL PAPER
Figure 4. c) Thermal breakdown of PhOPhOMe in the iPEPICO
setup at constant 10.5 eV ionization energy.
To sum up, the pyrolysis under collision-reduced conditions,
results in the molecule with m/z=184 as the primary product. The
secondary product is cyclopentabenzofuran (m/z=156), which
decomposes to naphthalene and diethynylbenzene. Furthermore
ethynylbenzene, m/z=102 was found. Due to the lack of
references for the ionization energy of m/z=184, further study
must be conducted to determine the exact structure. Since the
methyl radical forms simultaneously, this molecule results from
abstraction of
a hydrogen atom from the corresponding
phenoxy-type radical (m/z=185). Possible isomers are
dibenzodioxin, dibenzofuran-4-ol, dibenzofuran-4-one, and
biphenyl-2,3-dione. Therefore, the ionization energies of the
listed isomers were calculated using Gaussian09 on the CBS-
QB3 level of theory. Table S1 gives the calculated ionization
energies (IE) and the experimental value of m/z=184 (8.03eV).
Dibenzodioxin shows an IE of 7.48 eV (7.5 eV in literature).[38]
Dibenzofuranol has a calculated IE of 8.06 eV, dibenzofuran-4-
one 7.74 eV, and biphenyl-2,3-one 8.45 eV. For the intermediate
with m/z=156 the structure of cyclopenta-benzofuran was
reconstructed, which shows an experimental ionization energy of
7.5 eV, while the calculated energy was 7.49 eV(B3LYP) and
7.42 eV(CBS-QB3).
Based on the calculations, the isomer with the best fit,
dibenzofuran-4-ol, was synthesized and its TPES measured and
compared to the experimental TPES of the intermediate product
during pyrolysis of PhOPhOMe at 750 °C. The spectra in Figure
S2 show good agreement. Dibenzofuran-4-ol pyrolysis, also
yields the following familiar fragments: first m/z=156, then
m/z=128 (naphthalene) and m/z=126 (diethynylbenzene)
followed by m/z=102 (ethynylbenzene). Figure S3 shows the
corresponding mass spectra. Those detected products are the
same as in PhOPhOMe pyrolysis, which again proves that
dibenzofuran-4-ol is the main intermediate with m/z=184. We
have performed quantum chemical calculations to evaluate
reaction pathways yielding dibenzodioxin and dibenzofuran-4-ol,
which clearly favors the latter one as presented in Figure 5.
Figure 4. a) Temperature-dependent mass spectra (normalized to the highest
peak) of the pyrolysis of PhOPhOMe in the iPEPICO-TOF setup at a sample
temperature of 90 °C and an ionization beam energy of 10.5 eV and b) a
detailed mass spectrum at 750 °C.
Figure 4a shows the primary products of 1-methoxy-2-
phenoxybenzene (PhOPhOMe), measured in iPEPICO under
high vacuum. Figure 4b shows the detailed mass spectrum of
the products at 750 °C. At this temperature, the parent molecule
decomposes almost completely, and the signal at m/z=200 is
quite low. The detected product molecules and radicals are
m/z=184/185, m/z=168 (dibenzofuran), m/z=156, m/z=128
(naphthalene),
m/z=126
(diethynylbenzene),
m/z=102
(ethynylbenzene), and m/z=15 (methyl radical). Small amounts
of m/z=78/77 and m/z=80 are also observed vide infra. As soon
as PhOPhOMe starts to decompose the molecule m/z=184
forms. M/z=185 is detected and its mass spectrum shown in
Figure S1 (Supplemental Information). Furthermore, the methyl
radical, m/z=15, is visible from the start (T = 500 °C). As
temperature increases the molecule with m/z=156 forms and
decomposes again above 750 °C. At this temperature
naphthalene and diethenyl benzene are the main products.
While the formation of
a new ether bond (R4) is less
energetically demanding than the H transfer, the hydrogen
abstraction in the second step requires much more energy than
the hydrogen abstraction of R3 to form dibenzofuranol.
This article is protected by copyright. All rights reserved.

10.1002/chem.201700639
Chemistry - A European Journal
FULL PAPER
TS4
43.7
OH
O
TS23
38.6
R2
TS12
31.5
OH
TS3
31.8
H
The fragments butene-3-yne and the cyclopentadienyl (cp)
radical are unique in pathway a) and b) respectively. Pathway b)
leads to the immediate formation of the cp radical, which is
known to dimerize to form naphthalene and coke.[20] At ambient
pressure (Figure 1) phenol forms and accordingly phenoxy
radicals are stabilized by H-addition first - and only above a
certain temperature the phenoxy-radicals react further to
naphthalene via decarbonylation to cp and subsequent
dimerization. As usual the five-ring species evade detection in
O
O
28.0
15.1
O
TS14
23.1
O
R3
R4
R4
R1
H
H
13.4
m/z=185
O
H
Dibenzodioxin
O
-H
-H
Dibenzofuranol
O
OH
O
-65.0
m/z=184
-79.1
Figure 5. Reaction pathways to dibenzodioxin and dibenzofuranol initiated by
methyl abstraction from 1,2-methoxy-phenoxybenzene. The latter pathway is
favored by around 5 kcal/mol. All energies are given in kcal/mol.
the py-GC/MS-setup.[31,32]
In contrast, in the PEPICO
experiment, the phenoxy radical thermalizes and immediately
decomposes to cp and even smaller radical fragments (m/z=39
and m/z=26) at higher temperatures.[40,41] Similarly to the
decomposition of guaiacol[32] we see immediately that
intermolecular radical reactions play an important role in
pyrolysis under actual conditions and prevent fragmentation into
molecules smaller than aromatics.
When stabilization by H-addition (radical transfer reactions) is
apparently common in a radical-rich environment, the question
arises, why the hydroxy-phenoxy radical is not stabilized by the
same quenching pathway forming 1,4-benzenediol. The latter
radical was also stabilized by H-abstraction to form
benzoquinone, which evades detection in py-GC/MS similar to
the five-ring species. The mere presence of benzene and
biphenyl products in py-GC/MS proves the existence of the
second decomposition pathway a). In pyrolysis tests at ambient
pressure the benzoquinone cannot be detected anymore in the
Discussion
4-phenoxyphenol
Due to the introduction of
a
hydroxyl group to the
diphenylether[32] and similar bond energies of the ether bonds
two reaction pathways can open up, which are depicted in
Figure 6. We can now determine how this dimer decomposes at
elevated temperatures and introduce a (temperature dependent)
decomposition mechanism and show how the primary radicals
convert into stable products at ambient pressure. Phenyl radical
loss (Figure 6, route a on the right) affords p-hydroxyphenoxy
radicals, which yields benzoquinone after further hydrogen
abstraction as commonly observed in lignin pyroylsis.[39] The
quinone usually remains in the polymer and contributes to the
char.[35] In the following, the reactions of the benzoquinone and
phenoxy radicals are discussed explaining the induced
selectivity of the intermediately generated radicals.
presence of 4-phenoxyphenol. In
a reactive environment
quinones may quickly be depleted to yield carbon monoxide or
recombine to naphthalene-dione, styrene or naphthalene. The
increased reactivity of benzoquinone (BQ) in copyrolysis may
explain why it was not detected by GC/MS. Upon guaiacol
pyrolysis, the similar ortho hydroxyl-phenoxy radicals cannot be
detected, but the corresponding o-benzenediol, especially at
elevated temperatures (ambient pressure).[32] Robichaud et al.
detected p-benzoquinone in p-dimethoxybenzene pyrolysis; it
decomposes further to cyclopetadienone and butane-3-yne.[36] In
their study kinetic calculations showed that the formation of p-
benzoquinone is 106 times faster than its decomposition by
decarbonylation.[36,42] The absence of benzoquinone in the
GC/MS emphasizes the large difference between theoretically
favored reaction pathways and actual reactivity in the radical,
gas phase environment. While stabilization to benzoquinone is,
in theory, much more likely, the product is hardly detected.
Figure 6 shows the decomposition mechanism of PhOPhOH
under vacuum with all the detected intermediates. As mentioned
before, all fragments are detected at the same temperature.
Therefore, both reaction pathways (a and b) are energetically
equal. In pathway b the carbon-based hydroxyphenyl radical
rearranges immediately to the most stable C6H5O isomer, the
phenoxy radical (Figure 6, on the left).[40]
Also at ambient pressure both fission pathways are observed, as
benzene and phenol appear simultaneously. However, stabilized
equivalents of the phenoxy-hydroxy radical or benzoquinone
were not observed. The benzoquinone is known to decompose
quickly to carbon monoxide and butane-3-yne[36] and carbon
monoxide was detected in both experimental setups (in py-
GC/MS max 0.07 mol%), but can be found in both reaction
pathways.
O
186
OH
1,2-methoxy-phenoxybenzene
a
b
The methoxy substitution of diphenylether in ortho position is a
common moiety in lignin.[24,43] Both the methoxy-group and the
ether bond are possible breaking points in this model compound.
However O-CH3 bond cleavage is favored, due to the formation
of resonantly stabilized phenoxy-like radicals. The latter
intermediate may also yield the corresponding o-
benzoquinone.[35] In the following we will discuss the different
reaction pathways and why o-benzoquinone is not favored, but a
new carbon-carbon bond is formed accompanied by the usual
decarbonylation.[40,44] Furthermore, we will augment our
analytical observations by quantum chemical calculations,
provindg evidence why the pathway to dibenzofuranol is favored.
O
O
+ H
- H
+ H
O
+
+
109
OH
OH
108
77
78
O
O
OH
94
OH
93
93
108
O
O
- CO
- H
+ CO
76
+ H
81
+ CO
65
66
80
+
26
52
+
26
39
+ CO
52
Figure 6. Decomposition mechanism and fission pathways of 4-hydroxyphenol
(PhOPhOH), deduced from the iPEPICO experiments.
This article is protected by copyright. All rights reserved.

10.1002/chem.201700639
Chemistry - A European Journal
FULL PAPER
Altarawneh et al. also calculated the various possibilities of
catechol coupling and concluded that a phenoxy-phenyl radical
tends to undergo loss of hydrogen rather than formation of
dibenzodioxin as a result of condensation and confirmed by our
experimental observations.[45] The observation of a newly formed
C-C bond upon pyrolysis is striking, because it shows that this
molecule is able to form naphthalene, a precursor of char, even
unimolecularly. The phenyl-ether radical may form such
products easily because of low activation energy.[46] Compared
to other lignin dimers, such as phenyl-phenethyl-ether, the ether
bond is so stable especially with a more fragile MeO-substituent,
that its thermal breakup involves the formation of a new carbon-
carbon bond. This is different compared to non-substituted
diphenylether, which showed major recombination at high
pressure and lower temperature.[32]
O
O
O
O
- Me
m/z=15
m/z=200
m/z=185
-R
minor detection:
O
O
O
-RH
OH
m/z=77
m/z=108
- CO
O
O
m/z=184
m/z=80
The further decomposition of PhOPhOMe is, as in previous
studies, dominated by decarbonylation under high vacuum.
Furthermore, similar to guaiacol decomposition at ambient
pressure, 2-hydroxyphenyl-phenylmethanone is detected at
lower temperatures (Figure 3).[32] This is initiated by radical
formation and subsequent 1,2-phenyl shift, similar to the 1,2-
methyl shift in guaiacol to yield the corresponding
benzaldehyde.[21,36,47] Figure S4 shows the proposed mechanism.
Probably due to the steric hindrance of the phenyl group it is
only formed in small quantities and outcompeted by other
reactions at higher temperatures. Instead of a methyl shift the
whole phenyl ring has to shift in this reaction. As temperature
increases dibenzofuranol and dibenzofuran are generated.
Dibenzofuranol and the isomer dibenzodioxin are also detected
at ambient pressure in contrast to the iPEPICO experiment.
O
- CO
O
m/z=168
m/z=156
- CO
m/z=128
m/z=126
Figure 7. Decomposition mechanism of 1-methoxy-2-phenoxy-benzene in
Plain diphenylether decomposes at ambient pressure pyrolysis
first and forms larger recombination products (Fig 8) and at
higher temperatures dibenzofuran; benzene and phenol are only
tertiary products.[32] When comparing these results with the
collision reduced conditions deduced from the results in iPEPICO experiments.
Figure
7 summarizes the decomposition of 1-methoxy-2-
phenoxybenzene and radial intermediates as detected by
iPEPICO experiment. In contrast to 4-phenoxyphenol, the
primary intermediate radical m/z=185 is indeed very unstable,
but does not break up at the ether bond to form phenoxy-like
radicals as 4-phenoxyphenol does. There is a side reaction,
similar to pathway b) (Figure 6) in PhOPhOH, which revealed by
the presence of cyclopentadienone and phenyl radicals/benzene
in trace amounts. The o-quinone with m/z=108 was not detected,
but it probably reacted further as quickly as it is formed, as
similar observations have been made together with kinetic
calculations with methoxy-phenoxy radicals.[36] However, the
major decomposition pathway at low pressure, does not involve
break-up of the ether bond but internal stabilization by
abstraction of hydrogen. The thermal energy required to
dissociate the methoxy group is clearly sufficient for the
molecule to form a new intramolecular C-C bond, which leads
directly to the formation of naphthalene. Quantum chemical
calculations show that the formation of dibenzofuranol requires
less activation energy via several transition states and the H-
transfer to the new hydroxyl-group. The first step on the way to
dibenzodioxin requires indeed less energy (R4, Figure 5), but is
outcompeted by the lower activation barrier to yield
dibenzofuranol (TS3, Figure 5). At higher temperatures and a
denser environment with more bimolecular reactions, H-
abstraction may be favored and more dibenzodioxin can indeed
be observed compared to the low pressure of the pyrolysis
reactor in the iPEPICO experiment.
decomposition
of
hydroxy
and
methoxy
substituted
diphenylether in the same setup, the decomposition temperature
decreases from 650 to 600 °C to 500 °C. At ambient pressure
the most prominent radical reactions are not induced by the
fission of the ether bond, but rather by hydrogen transfer in all
the compounds. This always leads to selectivity towards
dibenzofuran and dibenzofuranol as secondary products.
Comparing those three model compounds, each with a different
substituent, PhOPhOH first produces benzene and phenol
followed by dibenzofuran (recombination), while diphenylether
produces
directly
recombination
products.
Methoxy-
phenoxybenzene on the other hand,
yields the hydroxyl
equivalent (2-phenoxyphenol, Fig 3), which means that major
fractions of the ether bond remains intact. In lignin pyrolysis
hydroxyl groups in the p-position are preferred, because they
have the least tendency to recombine or build carbon-carbon
bonds. As mentioned above, diphenylether is known to
recombine at first and only at higher temperatures and the
secondary oligomers are either not formed anymore or
decompose faster than they are formed.[32] However, this plain
model compound is unlikely to exist in real lignin. Most phenols
in lignin have both methoxy and hydroxyl groups. Lignin with p-
coumaryl subunits (H-units) usually occur in grasses and for
example miscanthus,[48] which however have a low lignin content.
In general, the 4-O-5 bond is present in most lignins by 5 to 10%
per 100 C9 units.[48] This may not be the major fraction of bond
This article is protected by copyright. All rights reserved.

10.1002/chem.201700639
Chemistry - A European Journal
FULL PAPER
type, but the presence of this bond in real lignin will either lead
to recombination and dibenzofuran-like compounds or not to a
break-up of the lignin under ambient conditions. The desired
monomers thus do not form. Even if a dimer forms with this
residual ether bond and even if exits in the polymeric matrix, it
will further react to form C-C bonds and will eventually lead to
char. The weaker aryl ether bonds (β-O-4 and α-O-4) break
down easily by hydrolysis during lignin separation or
pretreatment.[29]
depolymerization and repolymerization rates and thus product
selectivity to a larger extend than the mere percentage of this
ether bond in lignin.
Experimental Section
Materials
Reactions, which tend to lead to C-C bond formation rather than
the breakup of the aromatic ether bond are a major issue during
lignin pyrolysis. This aromatic ether bond directly produces
carbon-carbon bonds, even in high vacuum. The substituent in
this model compound can thus steer the decomposition
reactions. The p-hydroxyl group relative to the ether bond leads
at first to more monomers, but in a radical environment the p-
benzoquinone decomposes directly leading to carbon monoxide
and coke as well.
The model compounds are 4-phenoxyphenol (PhOPhOH) (>99 %, ABCR
Chemicals), dibenzo-p-dioxin (>99 %, TCI Europe). All quartz devices
were calcined before reaction in air at 550°C for 5h at a heating rate of
5°C/min. The following compounds were synthesized:
1-phenoxy-2-methoxybenzene
1-phenoxy-2-methoxybenzene (PhOPhOMe) was synthesized according
to Evans et al.[49] by esterification from guaiacol and phenylboronic acid.
Guaiacol and benzeneboronic acid (AlfaAesar, >98 %), together with
triethylamine and Cu(OAc)2, were stirred in dichloromethane for 18h in a
nitrogen atmosphere and with 4Å molecular sieve. After chromatographic
purification and recrystallization white crystals formed with a purity of >
O
O
O
OH
ambient
600°C
500°C
650°C
99.5
%
by GC/MS. The characterization of 1-phenoxy-2-
hvac
high vacuum
hvac
ambient
pressure
ambient
pressure
methoxybenzene was also done by 1H- and 13C-NMR.
O
O
pressure
O
OH
O
dibenzofuran-4-ol
OH
Dibenzofuran-4-ol (4-hydroxybenzofuran) was synthesized according to
Shultz et al.[50] by oxidation of the corresponding boronic acid with
hydrogen peroxyide in tetrahydrofan.
OH
OH
O
O
O
+
O
4-(dibebzofuranyl)-boronic acid (Sigma Aldrich, > 95 %) was diluted in
THF, a 2 % NaOH solution and five times molar excess of H2O2 in a
nitrogen atmosphere and stirred at room temperature for 24 h. After
removing THF by vacuum, the organic product was extracted with
diethylether and washed with water and recrystallized over night. Purity
and identity were determined by GC/MS, 1H and 13C NMR.
Figure 8. Reaction pathways of the model compounds under various
conditions.[32]
Conclusions
iPEPICO experiment and VUV-Beamline
The iPEPICO analytical setup is a powerful tool enabling the
determination of complex decomposition mechanism with
intermediates unknown to literature. P-phenoxy-phenol breaks
up to form phenoxy, phenyl and hydroxyphenoxy radicals, and
the o-methoxy-benzene-phenol forms an intramolecular C-C
bond, and eventually naphthalene instead of breaking the ether
bond of the dimer. The comparison of these results with the
products of py-GC/MS reveals the importance of the pressure
and thus the radical environment during pyrolysis. Even though
the PhOPhOMe forms ortho-PhOPhOH as an intermediate
product, the product distribution is very different due to H-
transfer reactions at higher pressures preventing further
degradation of the dimer. In the lignin polymer, where many
more substituents are present and potential H-transfer reactions
may occur, the break-up of this ether bond seems even more
unlikely. In contrast, the p-hydroxyl substituent leads to fission of
the ether bond without these recombination reactions, even at
ambient pressure. However, the resulting benzoquinone is
highly unstable in a radical environment and decomposes
immediately to carbon monoxide and char. Therefore, the
aromatic ether bond present in lignin largely influence the lignin
depolymerization due to the identified subsequent reactions in
this work, which do not only depend on pressure, radical
concentration, but also on the type and position of the
substituent in lignin. A suitable pretreatment could control
To investigate the primary thermal decomposition mechanism, pyrolysis
of 4-phenoxyphenol and 1-phenoxy-2-methoxybenzene was performed in
the iPEPICO-endstation[37,51] of the VUV-Beamline at the Swiss Light
Source.[52] The X04DB bending magnet provides the synchrotron
radiation collimated by a toroidal mirror onto a 150 l/mm grating. Another
mirror focuses the radiation on the exit slits, which together with a
mixture of neon, argon and krypton are in the gas-filter. The latter one
suppresses the high order radiation, which is also diffracted by the
grating. The pyrolysis setup consists of an resistively heated SiC-tube (1
mm iD / 2 mm oD, 10 mm heated zone).[32] An argon stream carries the
diluted (>0.1 %) starting material through the reactor resulting in average
residence times of 100-300 µs. A typical backing pressure of 500 mbar
was applied to the 100 μm nozzle generating a supersonic molecular
beam going through the reactor and into the iPEPICO source chamber at
10-4 mbar. The molecular beam enters the spectrometer chamber (1-5
x10-6 mbar) after skimming and is subsequently ionized by VUV light. The
photon-energy resolution is typically around 8 meV, which is calibrated
according to the Rydberg state of argon. The electrons are velocity map
imaged and detected by a Roentdek DLD40 position sensitive detector
and the corresponding ions, detected in a Jordan TOF (C-726) mass
spectrometer, accelerate in the opposite direction. A multiple start –
multiple stop scheme correlates the events in real time.[53] To identify
isomers and intermediates mass-selected TPES were recorded with a
resolution of 0.01 eV. The background of hot electrons is subtracted
according to the method from Sztaray and Baer.[54]
This article is protected by copyright. All rights reserved.

10.1002/chem.201700639
Chemistry - A European Journal
FULL PAPER
Signals are considered to be impurities, when the signal intensity does
not depend on the pressure - and therefore the concentration - of the
molecular inlet beam, which means that those compounds are in the
detector chamber and do not influence the reaction. Additionally,
experiments with varying sample concentrations (backing pressure of the
molecular beam) were performed in order to rule out possible changes in
selectivity due to bimolecular collisions. In the tested range limited by the
vapor pressure of the model compound no changes were observed.
[9]
Z. Ma, E. Troussard, J. A. van Bokhoven, Appl. Catal. A Gen. 2012,
423–424, 130–136.
[10]
[11]
[12]
[13]
[14]
D. J. Mihalcik, C. A. Mullen, A. A. Boateng, J. Anal. Appl. Pyrolysis
2011, 92, 224–232.
M. Zhang, F. L. P. Resende, A. Moutsoglou, D. E. Raynie, J. Anal.
Appl. Pyrolysis 2012, 98, 65–71.
D. M. Alonso, J. Q. Bond, J. A. Dumesic, Green Chem. 2010, 12,
1493–1513.
Py-GC/MS
C. Bährle, T. U. Nick, M. Bennati, G. Jeschke, F. Vogel, J. Phys.
Chem. A 2015, 119, 6475–6482.
Ambient pressure pyrolysis of the model compounds was performed in a
platinum coil pyrolyzer (5150, CDS Analytical) consisting of an open-
ended quartz reactor packed with loose quartz-wool in an helium carrier
gas stream. The model compound (1-2 μL) was pyrolyzed at a heating
rate of 20 °C/ms, and was then kept at the final temperature for 1 min.
The pyrolysis products were injected into an Agilent 7890A GC with an
Agilent 5975C MS system through the pyrolysis interface and transfer
line at 300 °C. The GC also has a thermal conductivity detector (TCD),
which is calibrated for the most predominant gases (CO, CO2, CH4 etc.).
The condensable fraction injected into the GC/MS system was
characterized according to the NIST08 mass spectrum library. The stated
selectivity was based on the percentage of the peak area and the
conversion was calculated by the integrated peak area. Variation of the
sample size at 900 °C determined the standard deviation of each product.
All reactions were at least performed in duplicate and reproduced within
95 %.
A. Demirbas, Energy Sources, Part A Recover. Util. Environ. Eff.
2009, 31, 1186–1193.
[15]
[16]
C. Zhao, J. A. Lercher, Angew. Chemie 2012, 124, 6037–6042.
A. B. Lovell, K. Brezinsky, I. Glassman, Int. J. Chem. Kinet. 1989,
21, 547–560.
[17]
[18]
A. M. Scheer, C. Mukarakate, D. J. Robichaud, M. R. Nimlos, H.-H.
Carstensen, G. B. Ellison, J. Chem. Phys. 2012, 136, 44309.
E. Barker Hemings, G. Bozzano, M. Dente, E. Ranzi, Chem. Eng.
Trans. 2011, 24, 61–66.
[19]
[20]
M. M. Suryan, J. Am. Chem. Soc. 1989, 111, 1423–1429.
A. V. Friderichsen, E. J. Shin, R. J. Evans, M. R. Nimlos, D. C.
Dayton, G. B. Ellison, Fuel 2001, 80, 1747–1755.
P. F. Britt, A. C. Buchanan, M. J. Cooney, D. R. Martineau, J. Org.
Chem. 2000, 65, 1376–89.
[21]
[22]
Gaussian calculations
P. F. Britt, A. C. Buchanan, K. B. Thomas, S.-K. Lee, J. Anal. Appl.
Pyrolysis 1995, 33, 1–19.
Gaussian 09 was utilized, applying the B3LYP functional and the 6-
311++G(d,p) basis set to calculate the equilibrium geometry and force
constant matrixes[55]a, which were used to compute Franck-Condon
factors with the ezSpectrum.OSX program.[55]a,b To the calculate the
ionization energies and the relative energy differences of unknown
[23]
[24]
[25]
M. Koyama, Bioresour. Technol. 1993, 44, 209–215.
M. Brebu, C. Vasile, Cellul. Chem. Technol. 2010, 44, 353–363.
A. Maccoll, S. W. Ramsay, R. F. Laboratories, W. C. I. London,
Chem. Rev. 1969, 69, 33.
intermediates the CBS-QB3 composite method was selected.[56,57]
.
[26]
H. Kwart, S. F. Sarner, J. Slutsky, R. December, J. Am. Chem. Soc.
1973, 95, 5234–5242.
[27]
[28]
M. T. Klein, P. S. Virk, Ind. Eng. Chem. Fundam. 1983, 22, 35–45.
M. W. Jarvis, J. W. Daily, H.-H. Carstensen, A. M. Dean, S. Sharma,
D. C. Dayton, D. J. Robichaud, M. R. Nimlos, J. Phys. Chem. A
2011, 115, 428–438.
Acknowledgements
The authors thank the Swiss National Science Foundation for
finical support (NRP66, no. 406640-136892). The experiments
were performed at the VUV beamline of the Swiss Light Source
at the Paul Scherrer Institute (PSI). The work was supported
financially by the Swiss Federal Office for Energy (BFE Contract
Number SI/501269-01). Calculations were performed at the HPC
cluster Brutus at ETH.
[29]
[30]
[31]
B. Saake, R. Lehen, Ulmann’s Encycl. 2012, 21–36.
A. Beste, A. C. Buchanan, J. Phys. Chem. A 2012, 116, 12242–8.
A. M. Scheer, C. Mukarakate, D. J. Robichaud, M. R. Nimlos, G. B.
Ellison, J. Phys. Chem. A 2011, 115, 13381–9.
V. B. F. Custodis, P. Hemberger, Z. Ma, J. A. van Bokhoven, J.
Phys. Chem. B 2014, 118, 8524–8531.
[32]
[33]
[34]
[35]
[36]
R. Parthasarathi, R. A. Romero, A. Redondo, S. Gnanakaran, J.
Phys. Chem. Lett. 2011, 2, 2660–2666.
Keywords: lignin • fast pyrolysis • model compounds • radical
mechanism • TPES iPEPICO
C. Bährle, V. Custodis, G. Jeschke, J. A. van Bokhoven, F. Vogel,
ChemSusChem 2016, 9, 2397–2403.
[1]
J. Jönsson, K. Pettersson, T. Berntsson, S. Harvey, Int. J. Energy
Res. 2013, 37, 1017–1035.
C. Bährle, V. Custodis, G. Jeschke, J. A. van Bokhoven, F. Vogel,
ChemSusChem 2014, 7, 2022–9.
[2]
[3]
[4]
[5]
[6]
A. V. Bridgwater, Biomass and Bioenergy 2012, 38, 68–94.
T. Dickerson, J. Soria, Energies 2013, 6, 514–538.
K. Freudenberg, H. Richtzenhain, Angew. Chemie 1939, 1–5.
O. Faix, Holzforschung 1991, 45, Suppl. 21-27.
V. B. F. Custodis, C. Bährle, F. Vogel, J. a. van Bokhoven, J. Anal.
Appl. Pyrolysis 2015, 115, 214–223.
D. J. Robichaud, A. M. Scheer, C. Mukarakate, T. K. Ormond, G. T.
Buckingham, G. B. Ellison, M. R. Nimlos, J. Chem. Phys. 2014, 140,
234302.
[37]
[38]
[39]
[40]
A. Bodi, P. Hemberger, T. Gerber, B. Sztáray, Rev. Sci. Instrum.
2012, 83, 83105.
F. P. Colonna, G. Distefano, V. Galasso, K. J. Irgolic, C. E. King, G.
C. Pappalardo, J. Organomet. Chem. 1978, 146, 235–244.
E. Dorrestijn, L. J. J. Laarhoven, I. W. C. E. Arends, P. Mulder, J.
Anal. Appl. Pyrolysis 2000, 54, 153–192.
[7]
[8]
T. P. Vispute, H. Zhang, A. Sanna, R. Xiao, G. W. Huber, Science
2010, 330, 1222–7.
T. R. Carlson, G. A. Tompsett, W. C. Conner, G. W. Huber, Top.
Catal. 2009, 52, 241–252.
P. Hemberger, G. da Silva, A. J. Trevitt, T. Gerber, A. Bodi, Phys.
This article is protected by copyright. All rights reserved.

10.1002/chem.201700639
Chemistry - A European Journal
FULL PAPER
Chem. Chem. Phys. 2015, 17, 30076–30083.
[41]
[42]
[43]
T. K. Ormond, P. Hemberger, T. P. Troy, M. Ahmed, J. F. Stanton,
G. B. Ellison, Mol. Phys. 2015, 113, 2350–2358.
P. Frank, J. Herzler, T. Just, C. Wahl, Symp. Combust. 1994, 25,
833–840.
R. Rinaldi, R. Jastrzebski, M. T. Clough, J. Ralph, M. Kennema, P.
C. A. Bruijnincx, B. M. Weckhuysen, Angew. Chemie Int. Ed. 2016,
55, 8164.
[44]
[45]
H.-H. Carstensen, A. M. Dean, Int. J. Chem. Kinet. 2011, 41, 498–
506.
M. Altarawneh, B. Z. Dlugogorski, Phys. Chem. Chem. Phys. 2015,
17, 1822–30.
[46]
[47]
S. Li, Q. Zhang, Comput. Theor. Chem. 2015, 1061, 80–88.
E. Dorrestijn, P. Mulder, J. Chem. Soc. Perkin Trans. 2 1999, 777–
780.
[48]
A. J. Ragauskas, G. T. Beckham, M. J. Biddy, R. Chandra, F. Chen,
M. F. Davis, B. H. Davison, R. A. Dixon, P. Gilna, M. Keller, et al.,
Science 2014, 344, 1246843.
[49]
[50]
[51]
[52]
D. A. Evans, J. L. Katz, T. R. West, Tetrahedron Lett. 1998, 39,
2937–2940.
D. A. Shultz, J. C. Sloop, G. Washington, J.Org. Chem. 2006, 71,
9104–13.
A. Bodi, M. Johnson, T. Gerber, Z. Gengeliczki, B. Sztáray, T. Baer,
Rev. Sci. Instrum. 2009, 80, 34101.
M. Johnson, A. Bodi, L. Schulz, T. Gerber, Nucl. Instruments
Methods Phys. Res. Sect. A Accel. Spectrometers, Detect. Assoc.
Equip. 2009, 610, 597–603.
[53]
A. Bodi, B. Sztáray, T. Baer, M. Johnson, T. Gerber, Rev. Sci.
Instrum. 2007, 78, 84102.
[54]
[55]
[56]
B. Sztaray, T. Baer, Rev. Sci. Instrum. 2003, 74, 3763–3768.
́
V. Mozhayskiy, A. I. Krylov, 2009.
L. A. Curtiss, P. C. Redfern, K. Raghavachari, J. Chem. Phys. 2007,
127, 124105.
[57]
J. A. Montgomery, M. J. Frisch, J. W. Ochterski, G. A. Petersson, J.
Chem. Phys. 1999, 110, 2822.
This article is protected by copyright. All rights reserved.

10.1002/chem.201700639
Chemistry - A European Journal
FULL PAPER
Entry for the Table of Contents (Please choose one layout)
Layout 1:
FULL PAPER
Text for Table of Contents
Author(s), Corresponding Author(s)*
Page No. – Page No.
Title
((Insert TOC Graphic here: max.
width: 5.5 cm; max. height: 5.0 cm))
Layout 2:
FULL PAPER
Author(s), Corresponding Author(s)*
Page No. – Page No.
LIGNIN TO CHEMICALS
O
OH
Title
O
substituent
+
R
& pressure
OH
O
Model compound pyrolysis
Text for Table of Contents
mechanism
This article is protected by copyright. All rights reserved.