pubs.acs.

org/journal/ascecg Research Article

Pretreatment with Sodium Methyl Mercaptide Increases

Carbohydrate Yield during Kraft Pulping
Connor J. Cooper,⊥ Ravikant Patil,⊥ Lintao Bu,⊥ Zhongyu Mou, David B. Turpin,
Adriaan van Heiningen,* Jerry M. Parks,* and Brandon C. Knott*
Cite This: ACS Sustainable Chem. Eng. 2021, 9, 11571−11580 Read Online

ACCESS Metrics & More Article Recommendations *

sı Supporting Information
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
Downloaded via NATL INST OF TECH WARANGAL on June 14, 2023 at 12:06:24 (UTC).

ABSTRACT: Kraft pulping is the predominant technology in the pulp

and paper industry for removing lignin from wood carbohydrates to
produce paper, board, packaging, tissue, and specialty cellulose. However,
the kraft process is energy intensive and expensive, and its yield is limited
by the degradation of carbohydrates. Pretreatment can increase
carbohydrate yield by limiting degradation via primary peeling of
reducing end groups. However, protection of galactoglucomannan
(GGM), the primary hemicellulose component of softwood, is minimal
when conventional pretreatments are used. Here, we investigated the
effectiveness of sodium methyl mercaptide pretreatment on southern pine
wood chips under a range of experimental conditions. We found that
pretreatment of biomass with 4.38% sodium methyl mercaptide at pH 12
and 105 °C for 60 min provided small but significant increases in xylan
and cellulose yields relative to control conditions, but preservation of GGM was minimal. To provide insight into molecular-scale
details of primary peeling, pretreatment, and alkaline hydrolysis, we performed classical molecular dynamics (MD) simulations under
selected process conditions and quantum mechanical (QM) calculations of selected reactions. MD simulations showed that C1 of
the GGM reducing end is more readily accessible by HO− and CH3S− ions than in cellulose. The free energy barrier for peeling
calculated with QM is lower for GGM than for cellulose, indicating increased susceptibility to peeling. In addition, we found that
GGM may be more susceptible to internal chain cleavage than cellulose. Thus, even though reducing end groups may be protected
initially through pretreatment, new unprotected reducing end groups may be generated through alkaline hydrolysis. Taken together,
these findings show the promise of methyl mercaptide as a pretreatment technology for cellulose retention and also provide
molecular insight for improving its effectiveness toward GGM.
KEYWORDS: Kraft pulping, Pretreatment, Biomass, Degradation, Peeling, Alkaline hydrolysis, MD, DFT

■ INTRODUCTION
The forest products industry creates consumer goods necessary
known for being odorous. The odor comes from a combination
of methyl mercaptan, dimethyl sulfide, dimethyl disulfide, and
for everyday life from renewable resources, accounting for H2S, although the first two are dominant. Methyl mercaptide
(CH3S−) is formed by a nucleophilic attack of HS− on the
approximately 4% of the total U.S. manufacturing gross
methoxy groups of lignin.
domestic product, manufacturing nearly $300 billion in
Wood is composed of 40%−45% cellulose, 25%−30%
products annually and employing approximately 950,000
hemicellulose, and 25−30% lignin, whereas the final kraft
men and women. 1 The industry meets a payroll of
pulp consists of 35%−40% cellulose, 8%−13% hemicellulose,
approximately $55 billion annually and is among the top 10
and 1%−2% lignin.4 Cellulose is a homopolymer consisting of
manufacturing sector employers in 45 states. In the
repeating D-anhydroglucopyranose units linked by β-1,4
manufacture of pulp, the removal of lignin from wood
glycosidic bonds (Figure 1). Galactoglucomannan (GGM),
carbohydrates is predominantly accomplished via the kraft
the primary hemicellulose in softwood, is a heteropolymer
process, resulting in global pulp production of about 200
million metric tonnes annually.2 The kraft process is based on
dissolution of lignin from wood by reaction at high Received: June 25, 2021
temperature (150−170 °C) with an aqueous solution of Revised: August 4, 2021
NaOH and Na2S.3 This process is energy and capital intensive Published: August 20, 2021
and provides a less efficient use of wood resources than desired
due to substantial degradation of wood polysaccharides,
hemicelluloses and cellulose.3 In addition, the kraft process is

© 2021 American Chemical Society https://doi.org/10.1021/acssuschemeng.1c04332

11571 ACS Sustainable Chem. Eng. 2021, 9, 11571−11580
ACS Sustainable Chemistry & Engineering
Figure 1. Chemical structures of cellulose and galactoglucomannan (GGM).
Figure 2. Selected reactions for (A) primary peeling, (B) stopping, (C) pretreatment, and (D) alkaline hydrolysis.
consisting of β-1,4-linked glucose and mannose (1:3−4 ratio)
as a backbone with some mannose units substituted by α-1,6-
linked galactose. Maximizing retention of these polysaccharides
through the pulping process is essential to plant economics.
At high pH, the reducing end groups (REGs) of cellulose
and GGM polymers can undergo a stepwise depolymerization
reaction called primary peeling (Figure 2A), resulting in
carbohydrate loss to the black liquor. Primary peeling of
cellulose becomes significant above 130 °C,5,6 although the
onset of primary peeling for GGM occurs at much lower
temperature (around 80 °C).7 Chain ends can also undergo
stopping reactions, in which the reducing end groups are
converted into alkali-stable carboxylate groups (Figure 2B). In
addition, at temperatures above 130 °C, the internal glycosidic
bonds in a polysaccharide chain are susceptible to alkaline
hydrolysis (Figure 2D). When hydrolysis occurs, newly formed
chain ends constitute REGs and thus possible sites for further
stepwise degradation (i.e., secondary peeling). Temperature,
alkali, and other pulping conditions all influence the relative
rates for peeling, stopping, and hydrolysis.
The dissolution of xylan is governed mostly by alkaline
solubility of the residual polymers because the uronic acid side
group protects against extensive peeling. As a result, more
11572
pubs.acs.org/journal/ascecg Research Article
xylan is retained when the alkalinity is kept low throughout the
pulping process. In contrast, some experiments have shown
primary peeling to degrade more than half of the original GGM
during impregnation with NaOH and Na2S,8 with further
losses due to secondary peeling. As a result, the final kraft pulp
contains only ∼25% of the native glucomannan.4
In general, pulp yield is less than 50% compared to a
potential maximum yield of 70%−75% when both cellulose
and hemicelluloses are fully retained. Estimated pulp yield
increases from completely eliminating primary peeling is ∼12%
for softwoods (4% cellulose, 8% GGM) and ∼6% for
hardwoods (4% cellulose, 2% glucomannan).4 Yield increases
of this magnitude could have a profound impact on plant
economics. Various approaches have been developed to
minimize primary peeling of carbohydrates by converting the
reducing end groups into alkali-stable groups through both
oxidative and reductive reactions (Figure 2C). Oxidative
pretreatments include the use of anthraquinone or polysulfides,
and examples of reductive pretreatments include hydrogen
sulfide, ammonium sulfide, and sodium borohydride.9−12
Suitable pretreatments should meet the following criteria: (i)
They should have low cost (or efficient recoverability). (ii)
They should avoid nonprocess elements (i.e., include only H,
https://doi.org/10.1021/acssuschemeng.1c04332
ACS Sustainable Chem. Eng. 2021, 9, 11571−11580
ACS Sustainable Chemistry & Engineering pubs.acs.org/journal/ascecg Research Article

C, O, S, and Na).(iii) They should be reducing, avoiding the
formation of unsaturated bonds that can undergo further
peeling.13 Although sodium borohydride is commonly used, it
is expensive and adds the nonprocess element boron to the
reaction conditions.12 Recently, sodium dithionite was
investigated but was unsuccessful in stabilizing mannose
under kraft pulping conditions.13
Sodium methyl mercaptide can be prepared by dissolving
methyl mercaptan gas in an aqueous solution of sodium
hydroxide (Scheme 1).
Scheme 1. Formation of Sodium Methyl Mercaptide
Methyl mercaptide is a stronger nucleophile than HS−, and a
similar attack on the methoxy groups of lignin by CH3S−
produces dimethyl sulfide (CH3SCH3). Dimethyl disulfide
(CH3S2CH3) is then formed by oxidation of CH3SCH3. The
beneficial effect of methyl mercaptide on carbohydrate
retention during kraft pulping has been reported in two
patents. Tirado et al. demonstrated a 3% increase in pulp
Assuming a yield loss on odw of 0.15%/kappa,4 these findings
imply that methyl mercaptide addition would lead to a pulp
yield increase of about 2% at the same kappa value. In addition,
methyl mercaptide can be generated from spent kraft pulping
liquor, greatly enhancing the commercializability of the
technology compared to prior proposed pretreatment tech-
nologies.
Here, we use aqueous SMM as a pretreatment for wood
chips to achieve a a kraft pulp yield increase of about 2%−3%
at optimal process conditions. The experiments to identify the
optimum conditions will be reported in a subsequent paper.
We perform quantitative pulp product composition analysis
and show that SMM is an effective pretreatment for improving
carbohydrate yield, but that it is substantially less effective for
GGM than for xylan and cellulose. We then deploy molecular
modeling studies to investigate the molecular roots for these
varying degress of effectiveness. Quantum mechanical
investigation of the peeling and hydrolysis reactions reveal
slightly lower free energy barriers for GGM compared to
cellulose for both types of degradation reactions. Molecular
dynamics simulations suggest that increased accessibility to
reaction sites by hydroxide ions may render GGM more
susceptible to degradative reactions under cooking conditions.

production (or about 1.5% based on the original oven dry
wood (odw) weight, assuming a pulp yield of about 50%)
when digester relief gases in a kraft pulp mill containing methyl
mercaptide, methyl sulfide, and dimethyl sulfide were
introduced into the fresh kraft cooking liquor, while
maintaining the same degree of delignification.14 More
recently, a Finnish patent reported an increase in the pulp
yield of about 1% when sodium methyl mercaptide (SMM,
Na+CH3S−) was used as an additive in polysulfide (PS) kraft
pulping of pine, while the kappa number (a measure of the
lignin content in pulp15) decreased from 27.9 to 21.7.16
■
MATERIALS AND METHODS
Pulping Experiments. Pulping experiments were performed
using southern pine chips obtained from U.S. mills. All experiments
were performed in quadruplet using four 235 mL cylindrical rocking
digesters. Each digester contained approximately 30 g of wood chips
(on oven dry basis, o.d.). To avoid the handling of methyl mercaptan
gas, an aqueous solution of SMM (21% w/w) containing a small
amount of NaOH (0.4% w/w) at pH 12.9 was obtained from Arkema
(King of Prussia, PA 19406). The SMM charge was 4.38% on wood
(o.d. basis). The liquor-to-wood ratio for SMM pretreatment was 3

Figure 3. Atomistic models for MD simulations. (A) Model of galactoglucomannan (GGM) composed of three repeating units (one repeating unit
is shaded in red), each consisting of three mannose, one glucose, and one galactose residue. (B) Edge-on view of an example starting configuration
for a 36-chain cellulose microfibril (center, shown with green carbon atoms) surrounded by 18 GGM molecules (various colors). (C) Structures of
wild type glucose dimer (Glu-Glu, top) and mannose dimer (Man-Man, bottom). Note the difference in orientation of the hydroxy group at the O2
position between glucose and mannose. (D) Structures of pretreated Glu-Glu (top) and Man-Man (bottom) dimers, in which the C1 hydroxy has
been replaced by hydrogen to create an alkali-stable 1-deoxypyranose species.

11573 https://doi.org/10.1021/acssuschemeng.1c04332
ACS Sustainable Chem. Eng. 2021, 9, 11571−11580
ACS Sustainable Chemistry & Engineering
L/kg, and the pretreatment temperature was between 105 to 130 °C.
The initial pH of the pretreatment liquor was approximately 12.
Following pretreatment, the digesters were cooled to room
temperature using a water bath, while makeup white liquor was
being prepared. The final pH of the pretreatment liquor was between
9.5 to 10.5. The total liquor-to-wood ratio after adding makeup white
liquor was 4.5 L/kg. The white liquor charge on wood was one of the
process variables, expressed as effective alkali (EA) charge. EA charge
is calculated as the weight of NaOH (as Na2O) plus half the weight of
Na2S (as Na2O), reported as a percentage of the dry wood chip
weight. After adding the makeup white liquor (sulfidity of 30% and
causticization efficiency of 80%), the digesters were placed in an oil
bath for 1 h at 115 °C to achieve uniform impregnation of white
liquor into the wood chips. At the end of impregnation, the digesters
were removed from the oil bath and cooled to room temperature,
while the oil bath was heated to the cooking temperature of 170 °C.
The digesters were then immersed in the bath and cooked to a target
H-factor of ∼1960 h. The H-factor is a measure of the pulping
intensity, which takes into account the cooking time and temperature
compared to pulping at 100 °C (eq 1).4
i 16115 yzz
expjjjj43.2 −
zdt
T (K ) z{
t assisted activation free energy (Ea) for the ring opening of cyclic α-D-
k
H= ∫t 0 (1)
Screened pulp was used to determine the kappa number (a
measure of the lignin content of pulp, calculated according to TAPPI
Standard T-236) and chemical composition. The carbohydrate
composition of wood and pulp was determined using high-
performance anion exchange chromatography (HPAEC) after acid
hydrolysis of the pulp sample. TAPPI standard methods (T-249, T-
222, and T-204) were used to estimate the composition of lignin and
extractives in the wood/pulp samples. The uronic anhydride content
was determined using the chromophoric group analysis method.17
Residual alkali was determined using a modified SCAN-N 2:88
method.
Molecular Dynamics (MD) Model Construction. GGM is a
heteropolymer consisting of uneven β-(1,4)-linked glucose and
mannose as a backbone with some mannose units substituted by α-
(1,6)-linked galactose. The molar ratio galactose:glucose:mannose in
GGM is approximately 1:1:3, and the atomic model used here for
GGM is shown in Figure 3A. To construct an atomistic model of
GGM bound to cellulose, a 36-chain cellulose Iβ microfibril with a
degree of polymerization (DP) of 20 was first generated with a
diamond cross-sectional geometry (Figure 3B). Subsequently, 18
GGM molecules with DP of 15 were then pseudorandomly arranged
around the cellulose microfibril, resulting in a cellulose/GGM ratio of
2.7, which is consistent with the experimentally observed cellulose/
GGM ratio of ∼2.6.18
In addition, because cellulose is a homopolymer consisting of
repeating D-anhydroglucopyranose units linked by β-(1,4)-glycosidic
bonds, we used a β-(1,4)-glycosidic bonded glucose dimer (Glu-Glu)
to represent the cellulose for additional MD simulations (Figure 3C).
Because of the 1:3 molar ratio between glucose and mannose, we used
a β-(1,4)-linked mannose dimer (Man-Man) instead of a glucose-
mannose dimer to represent GGM in the classical MD simulations
(Figure 3D). The dimer models were constructed, along with
pretreated forms of Glu-Glu and Man-Man dimers (i.e., with a 1-
deoxypyranose terminus), for computational efficiency as well as to
remove supramolecular effects given that cellulose forms microfibrillar
bundles whereas GGM does not. All generated models were solvated
with explicit TIP3P water molecules.19 When necessary in all
simulations, charge neutrality was achieved by adding sodium cations.
MD Simulation Details. All MD simulations were conducted
using the program CHARMM20 with the CHARMM carbohydrate
force field, all36_carb.21 The DOMDEC fast parallel CHARMM
method22 was used to accelerate performance. All simulations used a
2 fs time step and the SHAKE algorithm23 to keep the length of
covalent bonds to hydrogen fixed. Periodic solvated systems used a
nonbonded cutoff of 11 Å, with a 2 Å list buffer and heuristic list
update using a 1 Å switching function for dispersion and the particle-
11574
pubs.acs.org/journal/ascecg Research Article
mesh Ewald method24 to calculate electrostatics accurately to long-
range. All systems were initially minimized and then density
equilibrated via 200 ps constant pressure simulations (NPT) at 300
K. Subsequently, production runs of 300 ns for the GGM/cellulose
microfibril complex and 100 ns for the dimeric systems were
performed in the NVT ensemble.
Quantum Mechanical (QM) Model Construction. Similar to
the simplified models used in the classical MD simulations, we used a
β-(1,4) glucose dimer (Glu-Glu) to represent cellulose in the QM
calculations. Differing from the MD model, β-(1,4)-linked Glu-Man
was chosen as a model for GGM for the QM calculations of the
alkaline hydrolysis reaction pathway because Man-Man is unable to
undergo epoxide formation due to the stereochemistry at C2.
QM Method Validation. The dispersion-corrected hybrid meta-
GGA density functional approximation M062X-D325 and the 6-31G
(d,p) basis set were used throughout the study. M062X-D3 was
chosen because it is among the most accurate density functionals for
calculating conformational energies of carbohydrates,26 reaction
energies for small systems, and noncovalent interactions in the
GMTKN55 benchmarks.27 As an additional evaluation of this model
chemistry, we calculated the reaction free energy (ΔGr) and water-
glucose to form acyclic D-glucose. We obtained ΔGr = 9.5 kcal/mol at
the M062X-D3/6-31g(d,p) level of theory, which is in good
agreement with the available reference value of 10.3 kcal/mol
obtained at the CCSD(T)/G4 level of theory.28 Thus, we expect that
the M062X-D3/6-31g(d,p) level of theory should provide sufficiently
accurate energetics to investigate the model systems in the present
work.
To account for solution-phase effects, we used the SMD polarizable
continuum model29 with water as the solvent to calculate aqueous free
energies. However, using the continuum solvent representation alone
to calculate hydration free energies of charged solutes can be
inaccurate because it neglects specific solute−solvent interactions.
Thus, we included one or two explicit water molecules that interact
directly with the solute to account for explicit hydrogen bonding.
Vibrational frequencies were calculated for all optimized geometries
to confirm the presence of zero or one imaginary frequencies for
minima and transition states, respectively. Free energies were
calculated at 80, 105, 115, and 130 °C. All QM calculations were
performed with Gaussian 16, revision A.03.30
■
RESULTS
Experimental Kinetic Data for Real Wood Chips. We
investigated the effectiveness of 4.38% (w/w% on original
wood) SMM pretreatment in preserving carbohydrate yield
from southern pine wood chips. The operating conditions were
adjusted to obtain a kappa number in the target range of 25−
30. The kappa number, a measure of the pulp lignin content, in
this range is desirable because it is suitable for subsequent
bleaching.31 It is also desirable to stop the pulping process in
this kappa range to avoid excessive degradation and dissolution
of the wood carbohydrates.
On the basis of literature data,32,33 about 0.18% methyl
mercaptan (on-wood basis) is generated during kraft pulping
of softwood to produce a standard unbleached pulp. The
amount of methyl mercaptan produced in a typical softwood
kraft cooking process, regardless of cooking temperature, is
about 0.235 mol/L of the spent black liquor.33 Using
McKean’s data,32 which correlates the concentration of methyl
mercaptide in the spent black liquor with rate of production of
methyl mercaptan in the kraft process, it can be determined
that ∼0.18% methyl mercaptan (on-wood basis) is produced in
the softwood kraft pulping process. Thus, 1.8 kg of methyl
mercaptan per metric ton of oven-dry wood is equivalent to
0.18% (on-oven dry wood basis). This amount is significantly
less than the 3% (on-wood basis) of methyl mercaptan
https://doi.org/10.1021/acssuschemeng.1c04332
ACS Sustainable Chem. Eng. 2021, 9, 11571−11580
ACS Sustainable Chemistry & Engineering pubs.acs.org/journal/ascecg Research Article

Table 1. Composition of Pulp Samples, Rejects, and Residual Alkali

Δ pulp
composition (%)b yield (%)
experimental conditionsa kappa residual EA
(pretreatment; impregnation; pulping) xylan GGM cellulose lignin calcc measd number % rejects (g/L Na2O)
0 southern pine chips 11.76 15.61 41.73 28.88 n/a 100 − − −
1 1 mM NaOH, 115 °C (ctrl); 115 °C; 15% EA 4.61 3.64 35.09 1.86 n/a n/a 29.4 ± 0.7 0.04 ± 0.02 7.7 ± 0.15
composition (%) rel to ctrl rel pulp
yield (%)
2 none; 115 °C; 16% EA 0.21 −0.04 −0.91 −0.36 −1.10 −1.10 25.1 ± 0.2 0.05 ± 0.00 9.3 ± 0.20
3 none; 115 °C; 17% EA −0.10 −0.11 −0.74 −0.55 −1.50 −1.50 22.7 ± 0.5 0.06 ± 0.05 10.6 ± 0.20
4 4.38% SMM 105 °C; 115 °C; 15% EA 0.20 0.29 −0.50 −0.39 −0.40 −0.40 19.1 ± 0.5 0.03 ± 0.02 9.5 ± 0.68
5 4.38% SMM 105 °C; 115 °C; 12% EA 0.96 0.23 1.07 0.24 2.49 2.50 29.0 ± 0.8 0.01 ± 0.01 6.2 ± 0.04
6 4.38% SMM 115 °C; 115 °C; 12% EA 0.87 0.12 1.85 0.06 2.90 2.90 30.0 ± 0.5 0.04 ± 0.00 5.6 ± 0.06
7 4.38% SMM 130 °C; 115 °C; 12% EA 1.31 0.11 0.31 0.38 2.10 2.10 29.7 ± 0.9 0.01 ± 0.01 5.8 ± 0.08
a
Pretreatment at condition 1 was carried out at 1 mM NaOH at pH 12 for 60 min. No pretreatment in conditions 2 and 3. Pretreatments at other
conditions were at pH 12 for 60 min unless otherwise noted. bBased on dry wood mass. cCalculated from pulp analysis data. dMeasured from pulp
yield data.

Figure 4. Effect of kappa number on total yield for the experiments included in Table 1. Seven experimental conditions were examined, each in
triplicate.

(equivalent to 4.38% as SMM) that was added in most of the
present experiments. Thus, additional mercaptan was needed
in the present experiments to achieve 3% methyl mercaptan.
Prior to pulping, we determined the composition of
southern pine wood chips to be 11.76% xylan, 15.61%
GGM, 41.73% cellulose, and 28.88% lignin (Table 1, condition
0). Under pulping conditions of 1 mM NaOH, 115 °C, and pH
12 for 60 min, we obtained a kappa number of 29.4 ± 0.7
when the effective alkalinity (EA) was 15% (Table 1, condition
1). Under these conditions, the relative compositions of xylan,
GGM, and cellulose were reduced to 4.61, 3.64, and 35.09,
respectively, reflecting a loss of 60.79% of xylan, 76.68% of
GGM, and 15.91% cellulose at 93.56% lignin removal. The
largest carbohydrate loss occurred for GGM. A significant
increase in GGM retention would have a major positive effect
on the pulp yield.
We then analyzed the effect of changing the EA on kappa
number to identify pulping conditions that led to kappa
numbers in the desired range (Table 1 and Figure 4). As
expected, there is a linear relationship between total pulp yield
11575
and kappa number for the control cooks (Figure 4). Increasing
EA to 16% (condition 2) and 17% (condition 3) resulted in a
reduction of the kappa number to 25.1 ± 0.2 and 22.7 ± 0.5,
respectively. Thus, we selected condition 1 as the control.
Next, we considered the effects of SMM pretreatment and
found that at 4.38% SMM charge it was possible to reduce the
EA charge during pulping from 15% to 12% (conditions 4 and
5, respectively) to obtain a kappa number similar to that of the
control (condition 1). This approach led to a measured pulp
yield increase of 2.5% (on wood) for experiment 5 compared
to the control (Table 1 and Figure 4). For most of the
experiments, there is good agreement between the calculated
and experimentally measured pulp yield increase (Table 1).
The calculated pulp yield increase is the sum of the changes in
the percentages (on-wood basis) of xylan, GGM, cellulose, and
lignin, all relative to the control. The percentage of rejects was
less than 0.1%, and the residual alkali was generally within the
acceptable range of 4−10 g/L.
The initial SMM pretreatment experiments were performed
at a charge of 4.38% SMM on o.d. wood for 60 min at 105 °C
https://doi.org/10.1021/acssuschemeng.1c04332
ACS Sustainable Chem. Eng. 2021, 9, 11571−11580
ACS Sustainable Chemistry & Engineering
Figure 5. Snapshot of the GGM/cellulose complex at the end of a 300 ns MD simulation. (A) End-on view and (B) side view rotated 90° with
cellulose microfibril shown in green surface representation, and bound GGM residues (determined as those within 3 Å of the cellulose surface)
shown in solid surface model of various colors. Unbound GGM residues are shown in stick representation. Hydroxide ions are shown as spheres
(red, oxygen; white, hydrogen). (C) Radial distribution function of OH− ions with respect to O1 and O2 at the reducing end of cellulose or GGM.
Results shown are with NaOH concentration of 0.9 M; 1.5 M NaOH loading showed qualitatively similar behavior.
(condition 4). The temperature of 105 °C was chosen
assuming that, at this temperature and pH ∼ 12 of the
pretreatment liquor, the kinetics of the primary peeling
reactions of GGM and cellulose would be slower than the
kinetics of the reducing end group stabilization reactions by
methyl mercaptide ions. Under these pretreatment conditions
with 15% EA, the kappa number dropped to 19.1. Therefore,
to obtain a kappa number that is closer to the target value we
performed a 12% EA cook after pretreatment with 4.38% SMM
(condition 5), which resulted in a kappa number of 29. The
results show that the addition of SMM increased the retention
of wood carbohydrates with an increase of ∼2.5% in the pulp
yield and the same kappa as the control. It is also important to
note that the 4.38% SMM pretreatment allowed a decrease of
the EA charge in kraft cooking from 15% to 12%, while still
achieving the same kappa number as the control kraft cook at
15% EA. Under these conditions, we observed increases in the
yields of xylan (0.96%) and cellulose (1.07%) but a much less
significant increase in the GGM yield (0.23%). Details on the
calculations of the pulp composition will be reported
separately.
Next, we increased the temperature to 115 °C (condition 6).
We observed a 1.85% increase in cellulose yield relative to the
control and a 0.12% increase in GGM yield. Further increasing
the temperature to 130 °C (condition 7) resulted in a slight
reduction in the cellulose yield, likely due to increased peeling
reactions. We hypothesized that the mercaptide ion stabiliza-
tion reaction of cellulose occurs during the 60 min
impregnation phase at 115 °C, but apparently the stabilization
reaction of GGM by mercaptide is too slow compared to the
GGM primary peeling reaction to affect the retention of GGM.
Because primary peeling of GGM starts at about 80 °C7 but
primary peeling of cellulose is only significant at or above 130
°C,5,6 it is possible that mercaptide ion stabilizes some
cellulose before peeling but that GGM stabilization by
mercaptide is slow compared to primary peeling. The degree
of xylan retention was mostly related to alkali charge because
more xylan is dissolved at higher alkaline concentrations. The
experiments with 4.38% SMM pretreatment at 130 °C
(condition 7) resulted in a higher retention of xylan,
presumably because more alkali was consumed by cellulose
peeling, which in turn lowered the solubility of xylan in the
cooking liquor.
11576
pubs.acs.org/journal/ascecg Research Article
Molecular Dynamics (MD) Simulations. Molecular
dynamics (MD) simulations were performed to determine
the spatial distributions of ionic additives of the pretreatment
chemistry around each biomass polysaccharide. The inter-
actions of these ions were probed with the unmodified forms
of cellulose and GGM as well as with the corresponding
polymers with terminal 1-deoxypyranoses expected to be
formed during pretreatment.
Cellulose/GGM Simulations. To investigate the distribu-
tion of OH− ions under cooking conditions, i.e., in the
presence of NaOH, we performed a 300 ns simulation of a
GGM/cellulose microfibril complex (Figure 5A, B). Simu-
lations were performed at both 0.9 and 1.5 M NaOH.
Although much more concentrated than the experiments
described above, these conditions are consistent with past
investigations of GGM stabilization during kraft cooking.5 In
addition, whereas the experimental conditions were chosen to
maintain a balance between lignin removal and polysaccharide
retention, the higher ion concentrations used in the
simulations resulted in a much larger data set from which to
gather statistics. Simulations at both NaOH concentrations
indicate that OH− ions have a much higher propensity to bind
to the reducing end groups of GGM than to cellulose, as
revealed by comparing the radial distribution functions of each
polymer (Figure 5C).
Dimer Simulations under Cooking Conditions. A
higher affinity of hydroxide ions for GGM than for cellulose
(as deduced from the radial distribution functions) was also
observed in 100 ns simulations of Glu-Glu and Man-Man
dimers (Figure S1). Glucose and mannose differ only in the
stereochemistry at the C2 carbon, and our results demonstrate
that OH− ions exhibit a stronger propensity to localize at the
bridging position between the O1−O2 atoms in mannose.
Analogous simulations with pretreated dimers indicate that
OH− ions retain a stronger affinity to the hydroxy group in
mannose but in this case around the O2−O3 atoms.
Dimer Simulations under Pretreatment Conditions.
CH3S− ions form upon dissociation of sodium methyl
mercaptan and are thought to facilitate the conversion of
reducing end groups to alkali-stable 1-deoxypyranoses and
possibly other products during the pretreatment reaction.
Dimer simulations indicate that, similar to the behavior of
HO−, CH3S− ions exhibit affinity for C1 in both Glu-Glu and
Man-Man dimers (Figure 6). There is a slightly higher
https://doi.org/10.1021/acssuschemeng.1c04332
ACS Sustainable Chem. Eng. 2021, 9, 11571−11580
ACS Sustainable Chemistry & Engineering
Figure 6. Simulations of pretreated Glu-Glu and Man-Man dimers in SMM solution. Snapshots of CH3S− ions around C1 at the reducing end of
(A) a Glu-Glu dimer and (B) Man-Man dimer. The sulfur atom of each CH3S− ion within the first well in the RDF (approximately 4.25 Å of C1) is
shown every 100 ps of the 100 ns trajectory. (C) Radial distribution function of CH3S− ions with respect to C1 at the reducing end of a Glu-Glu
(green) or Man-Man (blue) dimer (CH3S− concentration = 1.5 M).
probability of CH3S− ions being located in close proximity to
C1 in Glu-Glu than in Man-Man.
MD simulations may also provide insight into the
susceptibility of cellulose and GGM polymers to alkaline
hydrolysis, which can lead to secondary peeling. Thus, we
measured the radial distribution of OH− ions with respect to
the glycosidic oxygen (O4) of cellulose and GGM. The
simulations indicate a much higher local concentration of OH−
ions around O4 in GGM than in cellulose (Figure 7).
Figure 7. Radial distribution function of OH− ions at the glycosidic
bond (O4 atom) in cellulose and GGM polymer MD simulations.
Quantum Mechanical (QM) Calculations of Peeling
and Hydrolysis Reactions. We performed quantum
mechanical calculations to estimate the thermodynamic and
kinetic feasibility of peeling and alkaline hydrolysis reactions.
These two reactions were chosen to determine whether there
are differences in free energy barriers between the dimeric
models representing cellulose and GGM. These calculations
provide insight into detailed mechanisms that have not been
tested or are unable to be tested experimentally.
Alkali-catalyzed endwise degradation, or primary peeling, is a
major cause of the loss of carbohydrates during kraft pulping.
The generally accepted mechanism of primary peeling involves
ring opening, enolization, β-elimination, and benzylic acid
rearrangement.4 Using this mechanism, we calculated free
energies for the primary peeling reaction at the experimentally
relevant temperatures of 80, 105, 115, and 130 °C (Figure 8
and Figure S2). As was done for the classical MD simulations
of dimeric models, we used a β-(1,4)-linked glucose dimer
11577
pubs.acs.org/journal/ascecg Research Article
Figure 8. Selected free energies for the primary peeling reaction of
Glu-Glu (representing cellulose) and Man-Man (representing GGM).
Free energies were calculated at 105 °C, and structures for the Glu-
Glu reactions are shown.
(Glu-Glu) and mannose dimer (Man-Man) to represent
cellulose and GGM, respectively.
Comparing the GGM and cellulose pathways, the corre-
sponding free energies for intermediates 1−4 are generally
within 1−2 kcal/mol of each other, indicating similar
thermodynamic favorability in the two systems. The overall
free energy barriers for the reactions were calculated as the
difference in free energy between the lowest-energy state, 1,
and that of the transition state (TS-1). The calculated barrier is
slightly lower for Man-Man (22.9 kcal/mol) than for Glu-Glu
(26.5 kcal/mol).
We then calculated reaction free energy diagrams for the
generally accepted mechanism of alkaline hydrolysis of
cellulose and GGM, with a particular emphasis on the
presumed rate-limiting step of converting the O2-deprotonated
species to the epoxide intermediate, which then results in the
formation of a terminal α-D-glucose unit (Figure 9). We used
β-(1,4)-linked Glu-Man as a model of GGM for the QM
calculations of the alkaline hydrolysis reaction pathway because
Man-Man is unable to undergo epoxide formation due to the
stereochemistry at C2. The only difference between the two
models is in the stereochemistry of C2 in one sugar, so similar
results for the two systems were anticipated. We obtained free
energy barriers of 40.8 kcal/mol for Glu-Glu and 38.7 kcal/mol
for Glu-Man (Figure 9). These values should be considered
essentially the same because they are within the expected error
for the QM method used. However, it is possible that GGM
may be slightly more readily hydrolyzed than cellulose.
https://doi.org/10.1021/acssuschemeng.1c04332
ACS Sustainable Chem. Eng. 2021, 9, 11571−11580
ACS Sustainable Chemistry & Engineering
Figure 9. Selected free energies for the alkaline hydrolysis reaction of
Glu-Glu (representing cellulose) and Glu-Man (representing GGM).
Free energies were calculated at 105 °C, and structures for the Glu-
Glu reactions are shown.
■ DISCUSSION AND CONCLUSIONS
Wood remains the largest cost in the manufacture of wood
pulp for paper, making up over half of the noncapital operating
cost of producing pulp.34 Therefore, increasing pulp yield
would have a major financial impact on pulp mill economics.
Unfortunately, and despite intensive R&D efforts, attempts to
increase kraft pulp yield by an industry-relevant process have
not yet been successful. Here, we have demonstrated a 3%
increase in pulp yield from softwood via pretreatment with
SMM. The primary means by which pulp yield is increased is
in the stabilization of cellulose against degradation by alkaline
peeling (Table 1). Unfortunately, stabilization of the primary
softwood hemicellulose galactoglucomannan (GGM) by this
pretreatment is minimal, although the cause for the
discrepancy between the two polysaccharides is not immedi-
ately clear.
To obtain a better fundamental understanding of the
molecular basis for the experimentally observed differences in
stabilization of cellulose and GGM by SMM pretreatment, we
carried out a computational analysis of relevant model systems.
Numerous kinetic models have been developed to predict
carbohydrate degradation during pulping.5,6,35,36 While useful
for predicting the effect of various kinetic parameters on
product distributions, such models do not provide the
molecular level insight critical to developing actionable
hypotheses for pulp yield enhancement. Molecular dynamics
studies have investigated various aspects of carbohydrate
conformations and dynamics in aqueous solution at the
molecular scale,23,37−39 but detailed molecular-scale insight
under kraft pulping conditions has not been demonstrated.
Quantum mechanical results presented here indicate a
slightly lower peeling barrier for Man-Man compared to Glu-
Glu. The DFT-derived energy barriers can be compared with
experimentally derived values for cellulose and GGM peeling,
calculated from experiments as 23.4 and 6.2 kcal/mol,
respectively.6 The latter value was noted in the original
publication to be surprisingly lower than values from previous
work.5 The value computed for GGM peeling (27.7 kcal/mol)
by Montagna et al.7 is much more in line with the results
presented here. The lower peeling barrier for the mannose
dimer suggests that more successive peeling events will occur
for GGM than for cellulose. However, we note that the peeling
rates will also be affected by the whole polymer context, and
11578
pubs.acs.org/journal/ascecg Research Article
larger scale structural features such as crystallinity may play an
important role in governing accessibility of reactive sites and
therefore peeling susceptibility. In fact, our classical MD
simulations of cellulose and GGM indicate that OH− ions have
a greater propensity to localize near the reducing end groups of
GGM than in cellulose. This difference in affinity could be due
to the chemical differences between cellulose and GGM or to
the crystalline nature of cellulose possibly rendering physical
protection against peeling. We posit that both effects are likely
at work, on the basis of these polymer-scale results as well as
our findings from dimer models. MD simulations of a mannose
dimer reveal an enhanced propensity for OH− ions to locate at
the bridging position between the O1 and O2 atoms at the
reducing end. This finding indicates that, even without a
polymer chain and in the absence of supramolecular structure
(i.e., a cellulose microfibril), peeling of mannose from GGM
may be promoted more so than glucose from cellulose. This
finding holds even in analogous simulations wherein the
reducing end group is converted to an alkali-stable 1-
deoxypyranose functionality. The enhanced affinity is possibly
due to cis-OH groups at these two carbon ring atoms and may
promote both the peeling reaction in the WT Man-Man dimer
and the dehydration reaction in the pretreated Man-Man
dimer.
In addition to primary peeling, internal chain scission by
alkaline hydrolysis leads to the creation of new chain ends that
can subsequently undergo secondary peeling. Thus, even
though SMM pretreatment can protect REGs against primary
peeling, new unprotected REGs may be generated during
pulping. Our DFT calculations on dimeric models indicate that
GGM may be slightly more susceptible to internal alkaline
hydrolysis of a carbohydrate linkage than cellulose, although
the calculated free energy barriers are within ∼2 kcal/mol of
each other. The DFT-derived energy barriers can be compared
qualitatively with previously reported experimental values. Our
GGM model for alkaline hydrolysis yielded a substantially
higher value (38.7 kcal/mol) than the results reported by
Montagna et al. (23.7 kcal/mol)7 and by Nieminen et al. (14.9
kcal/mol).6 The origin of the large discrepancies among these
values is unclear. In contrast, our calculated value of 40.8 kcal/
mol for alkaline hydrolysis with the cellulose model is quite
close to the corresponding value from kinetic modeling of 41.8
kcal/mol.6 Overall, our calculations follow the same trend in
reactivity; i.e., the barrier for cellulose hydrolysis is higher than
that for GGM.
Finally, we reiterate that the 2.5% pulp yield increase
obtained with the addition of 3% methyl mercaptan (on o.d.
wood chips, equivalent to 4.38% as SMM) is rather modest.
However, it has been proposed that a yield increase of 12% for
softwood (4% cellulose and 8% galactoglucomannan) could be
achieved if primary peeling could be eliminated by the addition
of methyl mercaptan. In addition, during regular softwood kraft
pulping, about 0.2% methyl mercaptan (on wood) is
formed.32,33 Thus, methyl mercaptan is already part of normal
industrial operations, albeit at lower levels than those used in
the present experiments. As a result, implementation of the
concept would require emission control techniques similar to
what is presently done in modern “odorless” kraft mills. Of
course, the methyl mercaptan would also have to be
regenerated from the spent pulping so that addition of makeup
methyl mercaptan would be kept to a minimum in a closed-
loop process; recycling of methyl mercaptan is a topic that will
be described in a subsequent publication. The primary
https://doi.org/10.1021/acssuschemeng.1c04332
ACS Sustainable Chem. Eng. 2021, 9, 11571−11580
ACS Sustainable Chemistry & Engineering
motivation for the present work, and especially the MD and
QM calculations, is to provide a better understanding of why
methyl mercaptan addition produced only a pulp yield increase
of 2.5% and that this fundamental insight might allow
identification of conditions that would minimize or even
suppress primary peeling of galactoglucomannan.
In conclusion, we have demonstrated that pretreatment of
softwood with sodium methyl mercaptan can increase kraft
pulp yield by 3%, while maintaining a kappa value in the
desirable range. This result is thought to be achieved by the
capping cellulose REGs with alkali-stable functionality that
protects against peeling. Hemicellulose GGM is left largely
unprotected, representing a further target for increasing pulp
yield. The results from atomistic MD simulations and DFT
calculations indicate that the molecular roots for these
observations include both the chemical differences between
mannose and glucose as well as differences in the supra-
molecular structures, namely, that cellulose forms microfibrils
whereas GGM does not. This molecular insight may be useful
for guiding future strategies for increased preservation of wood
carbohydrates during kraft pulping.
■
*
ASSOCIATED CONTENT
sı Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acssuschemeng.1c04332.
Additional figures (PDF)
Calculated energies (XLSX)
Cartesian coordinates for all optimized geometries (ZIP)
■ AUTHOR INFORMATION
Corresponding Authors
Adriaan van Heiningen − Department of Chemical and
Biomedical Engineering, University of Maine, Orono, Maine
04469-5737, United States; Email: adriaan.vanheiningen@
maine.edu
Jerry M. Parks − Biosciences Division, Oak Ridge National
Laboratory, Oak Ridge, Tennessee 37831-6309, United
States; orcid.org/0000-0002-3103-9333;
Email: parksjm@ornl.gov
Brandon C. Knott − Renewable Resources and Enabling
Sciences Center, National Renewable Energy Laboratory,
Golden, Colorado 80401, United States; orcid.org/0000-
0003-3414-3897; Email: brandon.knott@nrel.gov
Authors
Connor J. Cooper − Biosciences Division, Oak Ridge National
Laboratory, Oak Ridge, Tennessee 37831-6309, United
States; orcid.org/0000-0002-5527-9948
Ravikant Patil − Department of Chemical and Biomedical
Engineering, University of Maine, Orono, Maine 04469-5737,
United States
Lintao Bu − Renewable Resources and Enabling Sciences
Center, National Renewable Energy Laboratory, Golden,
Colorado 80401, United States
Zhongyu Mou − Biosciences Division, Oak Ridge National
Laboratory, Oak Ridge, Tennessee 37831-6309, United
States; orcid.org/0000-0003-2240-3129
David B. Turpin − Alliance for Pulp and Paper Technology
Innovation (APPTI), Washington, DC 20005, United States
Complete contact information is available at:
https://pubs.acs.org/10.1021/acssuschemeng.1c04332
11579
pubs.acs.org/journal/ascecg Research Article
Author Contributions
⊥
C. J. Cooper, R. Patil, and L. Bu contributed equally.
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS
This research was supported by the High-Performance
Computing for Manufacturing Project Program (HPC4Mfg),
which is managed by the U.S. Department of Energy (DOE),
Advanced Manufacturing Office, within the Energy Efficiency
and Renewable Energy Office. This work was partially
authored by the Alliance for Sustainable Energy, LLC, the
manager and operator of NREL for the U.S. Department of
Energy (DOE) under Contract No. DE-AC36-08GO28308.
The views expressed in the article do not necessarily represent
the views of the DOE or the U.S. Government. The U.S.
Government retains and the publisher, by accepting the article
for publication, acknowledges that the U.S. Government
retains a nonexclusive, paid-up, irrevocable, worldwide license
to publish or reproduce the published form of this work, or
allow others to do so, for U.S. Government purposes.
Computer time was provided by the National Renewable
Energy Laboratory Computational Sciences Center supported
by the DOE Office of EERE under Contract No. DE-AC36-
08GO28308. This research also used resources of the
Compute and Data Environment for Science (CADES) at
ORNL, which is managed by UT-Battelle, LLC, for the U.S.
DOE under Contract No. DE-AC05-00OR22725.
■
REFERENCES
(1) State Industry Economic Impact Report; American Forest & Paper
Association: January 2021.
(2) Pulp and Paper Capacities − Survey; FAO: Rome, 2020.
(3) Smook, G. A. Handbook for Pulp & Paper Technologists, 3rd ed.;
Angus Wilde Publications, 2002; Vol. 75.
(4) Sixta, H. Handbook of Pulp; Wiley-VCH Verlag GmbH & Co.
KGaA: Weinheim, Germany, 2006; pp 3−20.
(5) Paananen, M.; Tamminen, T.; Nieminen, K.; Sixta, H.
Galactoglucomannan stabilization during the initial kraft cooking of
Scots pine. Holzforschung 2010, 64 (6), 683−692.
(6) Nieminen, K.; Paananen, M.; Sixta, H. Kinetic Model for
Carbohydrate Degradation and Dissolution during Kraft Pulping. Ind.
Eng. Chem. Res. 2014, 53 (28), 11292−11302.
(7) Montagna, P. N.; Inalbon, M. C.; Paananen, M.; Sixta, H.;
Zanuttini, M. A. Diffusion Dynamics in Pinus sylvestris Kraft
Impregnation: Effect of Deacetylation and Galactoglucomannan
Degradation. Ind. Eng. Chem. Res. 2013, 52 (10), 3658−3662.
(8) Matthews, C. H. Carbohydrate losses at high temperature in
kraft pulping. Svensk Papperstidn 1974, 77 (17), 629−635.
(9) Hartler, N.; Olson, L. A. Hydrogen Sulfide Cooking Part I. First
Stage Variables. Svensk Papperstidn 1972, 75 (13), 559−565.
(10) Pekkala, O.; Palenius, I. Hydrogen sulfide pretreatment in
alkaline pulping. Paperi ja Puu 1973, 55 (9), 659−668.
(11) Chiang, V. L.; Sarkanen, K. V. Ammonium sulfide organosolv
pulping. Wood Sci. Technol. 1983, 17, 217−226.
(12) Wang, Y.; Azhar, S.; Lindström, M. E.; Henriksson, G.
Stabilization of polysaccharides during alkaline pre-treatment of wood
combined with enzyme-supported extractions in a biorefinery. J. Wood
Chem. Technol. 2015, 35 (2), 91−101.
(13) Lindén, P. A.; Lindström, M. E.; Lawoko, M.; Henriksson, G.
Stabilising mannose using sodium dithionite at alkaline conditions.
Holzforschung 2020, 74 (2), 131−140.
(14) Tirado, A.; Guevara, M.; Gonzalez, V. Method of digesting
wood by the kraft process utilizing organic sulphides. U.S. Patent
US3,451,889, 1969.
https://doi.org/10.1021/acssuschemeng.1c04332
ACS Sustainable Chem. Eng. 2021, 9, 11571−11580
ACS Sustainable Chemistry & Engineering pubs.acs.org/journal/ascecg Research Article

(15) TAPPI Standard T236 cm-85: Kappa Number of Pulp; TAPPI: Rendell, A. P.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.;
Atlanta, GA, 1993. Millam, J. M.; Klene, M; Adamo, C.; Cammi, R.; Ochterski, J. W.;
(16) Vaelttilae, O.; Pekkala, O. Method for making chemical Martin, R. L.; Morokuma, K.; Farkas, O.; Foresman, J. B.; Fox, D. J.
cellulose pulp by a polysulfide cooking process with the addition of Gaussian 16, Revision A.03; Gaussian, Inc.: Wallingford, CT, 2016.
organic mercaptans. Finland Patent FI118347B20071015, 2007. (31) Annergren, G.; Bowman, G. M.; Sandstrom, P. Principles of
(17) Scott, R. W. Colorimetric determination of hexuronic acids in Multi-stage Bleaching of Softwood Kraft Pulp. In Proceedings of the
plant materials. Anal. Chem. 1979, 51 (7), 936−941. International Pulp Bleaching Conference, Helsinki, Finland, 1998.
(18) Negahdar, L.; Delidovich, I.; Palkovits, R. Aqueous-phase (32) McKean, W. T.; H, B. F.; Sarkanen, K. V. Kinetic analysis of
hydrolysis of cellulose and hemicelluloses over molecular acidic odor formation in the kraft pulping process. Tappi J. 1965, 48 (12),
catalysts: Insights into the kinetics and reaction mechanism. Appl. 699−704.
Catal., B 2016, 184, 285−298. (33) McKean, W. T.; H, B. F.; Sarkanen, K. V.; Price, L.; Douglass, I.
(19) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. B. Effect of kraft pulping conditions on the formatino of methyl
W.; Klein, M. L. Comparison of simple potential functions for mercaptan and dimethyl sulfide. Tappi J. 1967, 50 (8), 400−405.
simulating liquid water. J. Chem. Phys. 1983, 79 (2), 926−935. (34) Kangas, P.; Kaijaluoto, S.; Määttänen, M. Evaluation of future
(20) Brooks, B. R.; Brooks, C. L., 3rd; Mackerell, A. D., Jr.; Nilsson, pulp mill concepts − Reference model of a modern Nordic kraft pulp
L.; Petrella, R. J.; Roux, B.; Won, Y.; Archontis, G.; Bartels, C.; mill. Nord. Pulp Pap. Res. J. 2014, 29 (4), 620−634.
Boresch, S.; Caflisch, A.; Caves, L.; Cui, Q.; Dinner, A. R.; Feig, M.; (35) Fearon, O.; Nykänen, V.; Kuitunen, S.; Ruuttunen, K.; Alén, R.;
Fischer, S.; Gao, J.; Hodoscek, M.; Im, W.; Kuczera, K.; Lazaridis, T.; Alopaeus, V.; Vuorinen, T. Detailed modeling of the kraft pulping
Ma, J.; Ovchinnikov, V.; Paci, E.; Pastor, R. W.; Post, C. B.; Pu, J. Z.; chemistry: carbohydrate reactions. AIChE J. 2020, 66 (8), e16252.
Schaefer, M.; Tidor, B.; Venable, R. M.; Woodcock, H. L.; Wu, X.; (36) da Silva Perez, D.; van Heiningen, A. Prediction of alkaline
Yang, W.; York, D. M.; Karplus, M. CHARMM: the biomolecular pulping yield: equation derivation and validation. Cellulose 2015, 22
simulation program. J. Comput. Chem. 2009, 30 (10), 1545−614. (6), 3967−3979.
(21) Guvench, O.; Greene, S. N.; Kamath, G.; Brady, J. W.; Venable, (37) Peric-Hassler, L.; Hansen, H. S.; Baron, R.; Hunenberger, P. H.
R. M.; Pastor, R. W.; Mackerell, A. D., Jr. Additive empirical force Conformational properties of glucose-based disaccharides investigated
field for hexopyranose monosaccharides. J. Comput. Chem. 2008, 29 using molecular dynamics simulations with local elevation umbrella
(15), 2543−2564. sampling. Carbohydr. Res. 2010, 345 (12), 1781−801.
(22) Hynninen, A. P.; Crowley, M. F. New faster CHARMM (38) Berglund, J.; Angles d’Ortoli, T.; Vilaplana, F.; Widmalm, G.;
molecular dynamics engine. J. Comput. Chem. 2014, 35 (5), 406−413. Bergenstrahle-Wohlert, M.; Lawoko, M.; Henriksson, G.; Lindstrom,
(23) Krautler, V.; Muller, M.; Hunenberger, P. H. Conformation, M.; Wohlert, J. A molecular dynamics study of the effect of glycosidic
dynamics, solvation and relative stabilities of selected beta- linkage type in the hemicellulose backbone on the molecular chain
hexopyranoses in water: a molecular dynamics study with the flexibility. Plant J. 2016, 88 (1), 56−70.
GROMOS 45A4 force field. Carbohydr. Res. 2007, 342 (14), 2097− (39) Berglund, J.; Azhar, S.; Lawoko, M.; Lindström, M.; Vilaplana,
2124. F.; Wohlert, J.; Henriksson, G. The structure of galactoglucomannan
(24) Darden, T.; York, D.; Pedersen, L. Particle mesh Ewald: An N· impacts the degradation under alkaline conditions. Cellulose 2019, 26
log(N) method for Ewald sums in large systems. J. Chem. Phys. 1993, (3), 2155−2175.
98 (12), 10089−10092.
(25) Zhao, Y.; Truhlar, D. G. The M06 suite of density functionals
for main group thermochemistry, thermochemical kinetics, non-
covalent interactions, excited states, and transition elements: two new
functionals and systematic testing of four M06-class functionals and
12 other functionals. Theor. Chem. Acc. 2008, 120 (1−3), 215−241.
(26) Marianski, M.; Supady, A.; Ingram, T.; Schneider, M.; Baldauf,
C. Assessing the accuracy of across-the-scale methods for pedicting
carbohydrate conformational energies for the examples of glucose and
alpha-maltose. J. Chem. Theory Comput. 2016, 12 (12), 6157−6168.
(27) Goerigk, L.; Hansen, A.; Bauer, C.; Ehrlich, S.; Najibi, A.;
Grimme, S. A look at the density functional theory zoo with the
advanced GMTKN55 database for general main group thermochem-
istry, kinetics and noncovalent interactions. Phys. Chem. Chem. Phys.
2017, 19 (48), 32184−32215.
(28) Assary, R. S.; Curtiss, L. A. Comparison of sugar molecule
decomposition through glucose and fructose: a high-level quantum
chemical study. Energy Fuels 2012, 26 (2), 1344−1352.
(29) Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. Universal
solvation model based on solute electron density and on a continuum
model of the solvent defined by the bulk dielectric constant and
atomic surface tensions. J. Phys. Chem. B 2009, 113 (18), 6378−6396.
(30) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Petersson,
G. A.; Nakatsuji, H.; Li, X.; Caricato, M.; Marenich, A. V.; Bloino, J.;
Janesko, B. G.; Gomperts, R.; Mennucci, B.; Hratchian, H. P.; Ortiz, J.
V.; Izmaylov, A. F.; Sonnenber, J. L.; Williams-Young, D.; Ding, F.;
Lipparini, F.; Egidi, F.; Goings, J.; Peng, B.; Petrone, A.; Henderson,
T.; Ranasinghe, D.; Zakrzewsk, V. G.; Gao, J.; Rega, N.; Zheng, G.;
Liang, W.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.;
Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H; Vreven, T.;
Throssell, K.; Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.;
Bearpark, M. J.; Heyd, J. J.; Brothers, E. N..; Kudin, K. N.; Staroverov,
V. N.; Keith, T. A.; Kobayashi, R.; Normand, J.; Raghavachari, K.;

11580 https://doi.org/10.1021/acssuschemeng.1c04332
ACS Sustainable Chem. Eng. 2021, 9, 11571−11580