Lysinol: A renewably resourced alternative to petrochemical polyamines and aminoalcohols

Article abstract of DOI:10.1039/c4gc01167h

This paper reports the preparation of lysinol (2,6-diamino-1-hexanol) by the hydrogenation of lysine and an example of its use as a replacement for petrochemical-derived amines. Lysine is presently manufactured by fermentation of sugars and other carbon sources at scale exceeding 10⁹ kg per year. Therefore, lysinol is potentially a renewable, platform aminoalcohol of previously unrecognized potential. Lysine hydrogenation proceeds under relatively modest conditions with Ru/C catalyst in water (100-150 °C, 48-70 bar, pH 1.5-2) to give lysinol in good yield (100% conversion, >90% selectivity; 50-70% isolated yield after purification by distillation). The impact of the various reaction parameters on conversion and selectivity are presented and discussed. Lysine hydrogenation at higher temperatures provides a pathway to piperidines and other products via further reduction and elimination of lysinol. The feasibility of lysinol synthesis from commodity, animal feed-grade lysine sources is presented as well. An example of the potential utility of lysinol is demonstrated by its use as a diamine curing agent with a standard epoxy resin. The properties of the resulting thermoset are contrasted with that obtained with a typical petrochemical amine used in this application (diethylenetriamine, DETA). This journal is

Full text of DOI:10.1039/c4gc01167h

Green Chemistry
View Article Online
View Journal | View Issue
PAPER
Lysinol: a renewably resourced alternative to
petrochemical polyamines and aminoalcohols†
Cite this: Green Chem., 2014, 16,
4575
Pranit S. Metkar, Mark A. Scialdone‡ and Kenneth G. Moloy*
This paper reports the preparation of lysinol (2,6-diamino-1-hexanol) by the hydrogenation of lysine and
an example of its use as a replacement for petrochemical-derived amines. Lysine is presently manufac-
tured by fermentation of sugars and other carbon sources at scale exceeding 109 kg per year. Therefore,
lysinol is potentially a renewable, platform aminoalcohol of previously unrecognized potential. Lysine
hydrogenation proceeds under relatively modest conditions with Ru/C catalyst in water (100–150 °C,
48–70 bar, pH 1.5–2) to give lysinol in good yield (100% conversion, >90% selectivity; 50–70% isolated
yield after purification by distillation). The impact of the various reaction parameters on conversion and
selectivity are presented and discussed. Lysine hydrogenation at higher temperatures provides a pathway
to piperidines and other products via further reduction and elimination of lysinol. The feasibility of lysinol
synthesis from commodity, animal feed-grade lysine sources is presented as well. An example of the
potential utility of lysinol is demonstrated by its use as a diamine curing agent with a standard epoxy resin.
The properties of the resulting thermoset are contrasted with that obtained with a typical petrochemical
amine used in this application (diethylenetriamine, DETA).
Received 22nd June 2014,
Accepted 28th July 2014
DOI: 10.1039/c4gc01167h
www.rsc.org/greenchem
modity-scale volumes. Nearly 3 × 109 kg of glutamic acid and
1.2 × 109 kg of lysine are produced annually, much larger
Introduction
The bulk of current research on renewable organic chemicals volumes than any other amino acids. It should be further
is directed at hydrocarbons and oxygenates. This is under- noted that lysine capacity is increasing at a significant rate, with
standable as these classes of organic molecules dominate the a current annual volume growth of about 4%.3 Significant
organic petrochemical industry. However, there are other current research efforts by lysine manufacturers continue to opti-
useful functional organics that also should be considered, mize the process to reduce costs and environmental impact. For
amines being one particular example.
example, Ajinomoto Co., the world’s largest lysine manufacturer,
With respect to amines that can be considered has announced a multi-generational plan to this end and ulti-
renewable, the few available examples are limited to commer- mately plan to convert their amino acid manufacturing to the fer-
cially-manufactured amino acids such as glutamic acid mentation of non-edible cellulose feedstock such as corn stover.4
and lysine. Both of these amino acids are manufactured by
Both glutamic acid and lysine have received attention as
fermentation of sugars (molasses, sucrose, corn steep liquor, potential platform renewable chemicals.5–7 The conversion of
etc.) although other carbon feeds (acetate or acetic acid, alco- glutamic acid to potentially useful chemicals, for example by
hols, paraffins) can also be used.1–3 Glutamic acid is primarily hydrogenation, decarboxylation, and cyclization chemistry,
used as a well-known flavor enhancer (monosodium gluta- has been investigated.8,9 Similarly, the conversion of lysine to
mate) while lysine, being an essential amino acid, is used 1,5-pentanediamine, caprolactam, and 5-aminovaleric acid has
nearly exclusively as a dietary supplement in animal feed. received recent attention.10–13
These two renewable chemicals are produced at large, com-
With respect to their reactivity, the zwitterionic nature of
α-amino acids typically limits the amine-like reactivity of the
α-amino group. Thus, lysine itself does not react like a typical
organic diamine because under most conditions the α-amino
group exists in the ammonium form. We reasoned that elimin-
ation of the zwitterion by reduction of the carboxylic acid func-
tionality to a non-acidic hydroxymethyl group (Scheme 1)
would provide a simple and scalable transformation to convert
readily available lysine into the reactive, renewable diamine
lysinol (Scheme 1).
E.I. du Pont de Nemours and Company, Central Research and Development,
Experimental Station, 200 Powder Mill Road, Wilmington, DE 19803, USA.
E-mail: kenneth.g.moloy@dupont.com
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
c4gc01167h
‡Present address: BetterChem Consulting LLC, 77 Allsmeer Drive, West Grove,
PA 19390, USA, E-mail: scialdon@comcast.net, www.betterchemconsulting.com,
Tel: 1-610-299-6750.
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 4575–4586 | 4575

View Article Online
Paper
Green Chemistry
than ammonia, and dialkyl amines are more reactive than
monoalkyl amines. As a result, nucleophilic reactions of
ammonia with EO produce a mixture of mono-, di-, and
triethanolamine. Reaction of ammonia with DCE produces a
more complex mixture of isomeric, cyclic, and oligomeric poly-
alkylated amines. This reactivity pattern also impacts the
metal-catalysed amination of ethanolamines to produce ethylene-
Scheme 1 Proposed conversion of lysine to lysinol.
Lysinol, an aliphatic diamino alcohol, possesses clear amines, as these reactions proceed via attack of nitrogen
structural and functional similarity to two major classes of nucleophiles on carbonyl intermediates generated by reversi-
petrochemical-derived amines, the ethyleneamines and the ble dehydrogenation.19 In commercial practice, the product
ethanolamines. World production of these two classes of poly- distributions can be and are shifted to some extent by the
functional amines exceeds 1.5 × 109 kg per year (ethylene- process parameters (for example, reactant ratios) but ulti-
amines: 3.9 × 108 kg per year in 2001;14,15 ethanolamines: 1.1 × mately rely on fractional distillation to separate the resulting
109 kg per year in 199916,17). The processes used to manufac- product mixtures. In the case of the ethyleneamines, the di-
ture these amines are shown in Scheme 2.18 The ethanol- and triamines are isolated as pure, single components.
amines are manufactured exclusively from ethylene oxide (EO) However, the heavier tetra- and pentaamines are generally not
and ammonia. The ethyleneamines are manufactured by two separated as pure components and are sold and used as tech-
routes, reaction of ammonia with either 1,2-dichloroethane nical-grade mixtures owing to the similar boiling points of the
(DCE) or with the ethanolamines, primarily monoethanol- individual components. In addition to the non-selective nature
amine. An inherent selectivity problem plaguing many amine of the chemistry to manufacture these amines, these processes
syntheses is readily exemplified by the chemistry depicted in utilize hazardous starting materials (EO, DCE) and, where
Scheme 2. These reactions rely on nucleophilic attack of DCE is used, generate large quantities of by-product
nitrogen on an alkylating agent (EO, DCE) and nitrogen salts. Clearly, a renewable alternative replacement to these
becomes more nucleophilic with increasing alkyl substitution. petrochemically-derived amines and aminoalcohols, manufac-
Therefore, the alkylated amine intermediates are more reactive tured by the efficient conversion of a renewable feedstock, is of
Scheme 2 Ethanolamine and ethyleneamine manufacturing processes.
4576 | Green Chem., 2014, 16, 4575–4586
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
of lysinol from lysine has been reported but involves nitrogen
protection, hydride-reagent reduction, and nitrogen deprotec-
tion.32 Multi-step, stoichiometric chemistry such as this is
obviously impractical to be considered for large scale appli-
cation and also entirely negates the green advantages of using
a renewable feedstock. It is possibly for these reasons that
there are surprisingly few literature citations to lysinol or its
derivatives. Lysinol derivatives have been employed primarily
in pharmaceutical applications in addition to a few studies of
chiral dendrimers and polymers.32–38
Scheme 3 Hydrogenation of α-aminoacids.
In this contribution, we describe our efforts to develop a
practical preparation of lysinol by the catalytic, aqueous phase
hydrogenation of lysine salts (sulphate, phosphate, chloride).
We then show a comparison of lysinol with a petrochemically-
derived ethyleneamine in a prototypical application, as a
hardener/curing agent in epoxy thermosets.
significant interest and potential utility in a wide range of
applications.§
The catalytic hydrogenation of α-amino acids to the corres-
ponding 1,2-aminoalcohols is known.20–30 In general, car-
boxylic acids can be hydrogenated to alcohols but require
fairly extreme conditions (≥200 °C, 50–140 bar) even with the
most active catalysts.31 However, the carboxylic acid group in
α-amino acids can be hydrogenated under significantly milder
conditions (100–150 °C, 17–70 bar) when performed at acidic
pH. Below approximately pH 3, α-aminoacids are largely
present as the cation and not the familiar zwitterion
(Scheme 3). It is speculated that the electron-withdrawing
effect of the adjacent ammonium cation renders the carboxylic
acid functionality more reactive towards reduction.
Experimental
Materials
Lysine was obtained from multiple sources. L-Lysine (free
base), L-lysine monohydrochloride, D-lysine monohydrochlo-
ride, and DL-lysine monohydrochloride were purchased from
Sigma Aldrich®. Commercial animal-feed grade lysine
samples were provided by Archer Daniels Midland Co.,
Decatur, Illinois, USA (Liquid lysine 50% feed grade, a dark
brown 50% aqueous solution of L-lysine free base; L-lysine HCl,
98.5% feed grade monohydrochloride as off-white granules)
and Evonik Industries AG, Essen, Germany (BioLys®, feed
grade lysine sulfate as brown granules containing the equi-
While the hydrogenation of several amino acids to alcohols
has been reported, there are, surprisingly, no reports describ-
ing the direct hydrogenation of lysine to lysinol.¶ The synthesis
§Relevant to this discussion are the relative prices of the ethylene- and ethanol-
amines compared to lysine and lysinol. While current pricing information is not valent of 50.7% by weight L-lysine free base).
readily or easily obtainable the available data suggest that they may be comparable.
Several catalysts were screened for the hydrogenation of
lysine to lysinol. For most of the reactions discussed in this
work, 5% ruthenium on carbon (Product no. 206180) powder
obtained from Sigma-Aldrich® was used. Catalysts were also
obtained from Johnson Matthey®. All the catalysts were
obtained in powder form.
Various reference standards including 2-piperidinemetha-
nol, 2-(aminomethyl)piperidine, 2-amino-1-hexanol, and
6-amino-1-hexanol were used for GC, LC, and NMR analysis
For example, 1999 ethyleneamines market prices were US$ 3–4 kg−1 15
In the early
.
to mid-1990s the manufacturing cost for lysine monohydrochloride was estimated
at US$ 1.5–1.8 kg−1, equivalent to US$ 1.9–2.3 kg−1 on a lysine content basis
(J. M. Connor, Rev. Ind. Org., 2001, 18, 5–21). L-Lysine prices during the infamous
1990s price-fixing case reached US$ 2.2 kg−1 (equal to US$ 2.7 kg−1 free base).
¶During the course of our work a patent published (P. Ouyang, K. Guo, H. Zhong
and Z. Fang, CN Pat, 102 617 364, 2012) that claims the hydrogenation of a variety
of α-amino acids and esters bearing pendant amino groups to produce diaminoal-
cohols. The patent further describes the preparation of various polymers, including
epoxy thermosets, from these diaminoalcohols. The preparation and use of lysinol
is included. Careful translation and consideration of this work indicate that it is in and were obtained from Sigma-Aldrich®.
error. For example, the patent describes lysinol only as a colorless solid that is puri-
fied by washing with a mixture of water and ethanol. No analytical or spectroscopic
information is provided. However, as demonstrated by others32 and in this present
Product analysis
The reaction/product mixtures were analyzed using GC, GCMS,
work, lysinol is a liquid and completely miscible with water and ethanol. Further-
more, the patent claims the hydrogenation of lysine methyl ester with RANEY® LCMS, and NMR. The high-pressure liquid chromatography
nickel (6 bar H2, 60 °C, 6 h, H2O). These exceptionally mild hydrogenation con-
(HPLC) instrument (Hewlett-Packard 1100) was equipped with
a mass spectrum detector. The separation was carried out on a
ditions are suspect. RANEY® nickel is a poor catalyst for the hydrogenation of esters
to alcohols, requiring much higher temperature and pressure, and is strongly basic.
Zorbax Eclipse AAA column (4.6 × 150 mm) operating at a flow
As discussed in this report, acidic conditions, incompatible with RANEY® nickel,
rate of 1 mL min−1 at 60 °C. The mobile phase consisted of a
are required for amino acid hydrogenation under mild conditions. These suspicions
were confirmed by exactly repeating the patent example using lysine methyl ester. A mixture of water and 0.1% formic acid (Phase A) and aceto-
colorless solid was indeed isolated as described. Analysis (NMR, LCMS) clearly
nitrile and 0.1% formic acid (Phase B) at a gradient of 95%
A/5%B to 0% A/100%B over 10 min, followed by a 4 min hold.
Gas chromatography was conducted with a Hewlett-Packard
proved the product is lysine, resulting from saponification of the methyl ester start-
ing material, presumably from the base introduced with the RANEY® nickel. No
hydrogenation product or gas uptake was observed. Polymers, including epoxy ther-
5890 (FID detection) or a Hewlett-Packard 6890 (MS detection)
mosets, prepared with this solid are, unsurprisingly, dramatically different than
those prepared from authentic lysinol as described in this paper.
and a 30 m × 0.32 mm × 0.25 μm ZB-5 column. Sample injec-
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 4575–4586 | 4577

View Article Online
Paper
Green Chemistry
tions were made at a 40 °C initial column temperature and Lysinol synthesis and isolation, representative procedure
after a 4 min hold were ramped to 300 °C at 20 °C min−1
.
L-Lysine (420.0 g, 2.79 mol) and 2.2 L of deionized water were
charged to a container. Concentrated sulfuric acid (160 mL,
2.94 mol) was added gradually with stirring, and using ice
cooling as required to maintain temperature <30 °C. The
mixture was stirred until all of the lysine had dissolved and the
temperature had returned to ambient. The resulting solution
and 100.0 g of powdered 5% ruthenium on carbon were
charged to a 4 L Hastelloy autoclave. The vessel was purged
with hydrogen and then brought to 120 °C and 69 bar hydro-
gen pressure with vigorous stirring, recharging with hydrogen
as needed until the hydrogen uptake ceased. After 22.5 hours
the reactor was cooled and depressurized. The catalyst was
removed by filtration, rinsed twice with deionized water, dried
under nitrogen, and saved for later reuse. LCMS (API-ES+) of
the combined filtrates showed lysinol (M + H = 133) and no
unreacted lysine (M + H = 147), as well as small amounts of
byproducts. The filtrate pH was increased to 12.2 by dropwise
addition of 50% NaOH (ca. 300 mL, 5.7 mole). Water was
removed by vacuum distillation to give a mixture of lysinol
and sodium sulfate as a colorless semi-solid. This material was
extracted with ethanol (2 × 500 mL) and then warm MeOH
(500 mL). The extracts were combined and concentrated to
give a light amber liquid. The product was fractionally distilled
to give 62 g of an impure forecut boiling at ca. 40–100 °C and
0.1 mm. Pure lysinol was then distilled at 105–118 °C and
0.1 mm. Yield of distilled lysinol: 259.8 g, 1.97 mol, 71%.
1H NMR analysis was performed on an Agilent DD2 Spectro-
meter operating at 500 MHz equipped with a room tempera-
ture 4-nucleus probe. The data were collected with 64
transients, 45° pulse width, a 12.9 kHz spectral width, 3.2 s
acquisition time, and a 45 s recycle delay. Samples for product
composition analysis were prepared with ca. 20 mg sample in
0.8 g D2O stock solution containing a known concentration of
3-(trimethylsilyl)-1-propanesulfonic acid sodium salt (DSS) as
internal reference.
Reactors
Small scale reactions were conducted in a custom high
pressure reactor of Hastelloy C construction (50 mL liquid
capacity). Mixing was achieved with a Teflon® coated stir bar
and magnetic stirring; the stir bar length was matched to the
reactor inner diameter to ensure catalyst fluidization. Reaction
progress was monitored by pressure drop using a Setra
Systems Model 206 Pressure Transducer (0–138 bar range) and
digital readout. Larger scale reactions were conducted in a
600 mL Parr® (Model no. 4563) Hastelloy C276 batch reactor
equipped with a magnetically-driven impeller, heating mantle
(Temperature Controller Model 4848), and a liquid sampling
system. Preparative scale reactions were performed in 4 or 19 L
autoclaves.
Hydrogenation screening
NMR (ppm, D2O/DSS): 13C: 68.5 (C1), 53.8 (C2), 42.8
The evaluation of reaction parameters such as catalyst and pH
were performed as described by the following procedure. The
50 mL pressure reactor was charged with catalyst (1–3 g), 30 g
reaction solution, and a stir bar. A typical reaction solution
contained 10 wt% lysine (free base), with the remaining
balance the requisite amounts of water and H2SO4. Catalyst
screening was conducted at an initial pH of about 2.1. The
reactor was then sealed and pressure tested. The vessel was
then pressurized with 34.5 bar H2 followed by heating to
120 °C. Upon reaching this temperature the reactor pressure
was adjusted to 55.2 bar by introducing additional H2. The
pressure drop was monitored regularly and the reactor was re-
pressurized once the pressure dropped below 48.3 bar. The
typical total reaction times were 24 h. The reactor was then
cooled and vented, and the catalyst was separated from the
reaction mixture by filtration with a 0.45 μm filter. The filtrate
was analyzed using LCMS and NMR techniques as described
earlier.
2
(C6), 34.7, 34.2 (C3, C5), 24.8 (C4). 1H: 3.52 (1H, dd, JHH
=
=
3
2
3
11.0 Hz, JHH = 4.5 Hz), 3.35 (dd, JHH = 11.0 Hz, JHH
7.0 Hz), 2.78 (1H, m), 2.58 (2H, t, J = 6.8 Hz), 1.45–1.3 (4H, m),
1.3–1.2 (2H, m).
Preparation of Mosher amides and determination of lysinol
enantiopurity – representative example
A vial was charged with a micro stir bar, S-lysinol (4 mg,
0.03 mmol, prepared by hydrogenation of S-lysine), diisopropyl-
ethylamine (10 μL, 8 mg, 6 mmol), and 0.5 mL of CD2Cl2.
(S)-(+)-α-Methoxy-α-trifluoromethylphenylacetyl chloride
[(S)-(+)-MTPA-Cl, S-Mosher’s acid chloride; 11 μL, 15 mg,
0.06 mmol] was added by syringe. After 2 h at room tempera-
ture another 0.3 mL CD2Cl2 and 1 drop of hexafluorobenzene
internal standard (δ −164.90) were added. The solution was
filtered through a 0.2 μ filter into a 5 mm NMR tube and
analyzed by 1H and 19F NMR. Enantiopurity was determined
by both integration and peak deconvolution of the CF3 reson-
ances in the 19F NMR. Peak assignments were made by prepar-
ing solutions of all combinations of R-, S-, and racemic lysinol
(in turn prepared by hydrogenation of R-, S-, and racemic
lysine) and R- and S-Mosher acid chloride.
Mass transfer
The rate of lysine hydrogenation is subject to mass transfer
limitations in the case of inefficient stirring. At 600 mL scale
the hydrogenation rate was found to increase with increasing
stirrer speed over the range 300–900 rpm. Above approximately
900 rpm the hydrogenation rate was observed to plateau.
Epoxy thermoset preparation and evaluation – instrumentation
Therefore, experiments were conducted at this higher stirrer Epoxy thermosets were prepared from lysinol or DETA and
speed to ensure that experiments were not operated under bisphenol-A-diglycidyl ether (BADGE) using stoichiometries of
H2-starved conditions.
1 equiv. N–H per 1 equiv. epoxide. Nanoindentation evalu-
4578 | Green Chem., 2014, 16, 4575–4586
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
ations were performed with a CSM Instruments Ultra Nano- Ru/C catalyst performed much better than any of the others
Hardness Tester equipped with a Berkovich indenter. Com- examined. The analogous 5% Rh/C catalyst showed marginal
pression and lap shear tests were conducted on a screw-driven activity (ca. 60% conversion), while carbon-supported Re, Ir,
universal mechanical tester under standard conditions (23 °C Ni, Pd, and Pt catalysts showed little to no activity. Bimetallic
and 50% RH). Evaluations were conducted in accordance with catalysts were not examined but have been reported in the
established ASTM methods (see ESI†).
patent literature for the hydrogenation of other amino
acids.29,30
Alumina and silica were investigated as alternative supports
to carbon, limiting the supported metal to ruthenium and to
5 wt% metal loading. Both 5% Ru/alumina and 5% Ru/silica
catalysts hydrogenate lysine but with lower activity than the
Results and discussion
Lysine hydrogenation
As discussed earlier, the catalytic hydrogenation of α-amino carbon support. In addition, lysinol selectivities are much
acids to the corresponding 1,2-aminoalcohols has been poorer with these oxides than that obtained with the carbon
reported.22,24,25,28–30 In particular, Jere et al. reported
a
support. While lysinol remains the primary product, the oxides
detailed kinetic study of the hydrogenation of alanine to alani- produce larger amounts of piperidines (2-(hydroxymethyl)-
nol.26 That report addresses the impact of a number of impor- piperidine, 2-(aminomethyl)piperidine, 2-methylpiperidine) as
tant reaction parameters on amino acid hydrogenation such as well as 1-hexylamine.
temperature, pressure, and acid concentration. The hydrogen-
We have not performed a detailed kinetic analysis but the
ation of lysine to lysinol has not been reported and therefore following observations are notable. Representative gas uptake
the impact of the additional pendant amino group is data are provided in Fig. 1. These hydrogenations exhibit
unknown. We undertook an investigation of this hydrogen- linear hydrogen uptake with time for the majority of the reac-
ation, including a cursory study of the hydrogenation con- tion as is seen by the linear least-squares fits to the initial
ditions to develop a semi-quantitative understanding of the uptake rates of these data sets. The uptake rate abruptly
key parameters to allow for the synthesis of lysinol on a pre- decreases and then stops. Analysis of the reaction solution
parative (ca. 0.1–1.0 kg) laboratory scale.
shows quantitative lysine conversion at the cessation of hydro-
Lysine hydrogenation was first investigated using the reac- gen uptake.
tion conditions reported in the literature as guidance: 5% Ru/
A few conclusions can be drawn from these data. Under
C, 120–160 °C, 48–69 bar, aqueous H3PO4.26,27 Two molar these conditions the hydrogenation can be considered to be
equivalents of H3PO4 were used to ensure protonation of both approximately pseudo-first order in hydrogen pressure
of amino groups (vide infra). Under these conditions lysine is through a substantial portion of the reaction. Thus, at initial
indeed hydrogenated to give lysinol as the major product. pressures Pi of about 69 bar (Fig. 1, curves b and c), the total
Lysine conversions of 100% are easily achieved at unoptimized
productivities on the order of ca. 0.5 M h−1. Lysinol is the
major product, exceeding 90% selectivity under some reaction
conditions. The primary by-products are the result of cycliza-
tion to produce piperidines. For example, 2-hydroxymethyl-
piperidine, 2-aminomethylpiperidine, 2-methylpiperidine, and
piperidine are observed.
After establishing that lysinol could be prepared in excellent
conversion and selectivity, modifications to the original con-
ditions were explored. First, the lysine concentration was
increased to 10–20 wt% from the 2–4 wt% amino acid and
phosphoric acid was replaced with sulfuric acid. In addition,
in our hands the catalyst preactivation step described (13.8 bar
H2, 200 °C, 15–18 h) prior to hydrogenation is not necessary;
we observe no difference in activity or selectivity between
experiments using preactivated catalyst and those with no pre-
activation and therefore this step was omitted. Apparently, any
requisite activation of the 5% Ru/C catalyst is rapid and com-
plete when the reaction solutions reach the target temperature
and pressure.
Further examination of the reaction parameters was con-
ducted. First, commercially-available hydrogenation catalysts
were screened for lysine hydrogenation under a standard set of
Fig. 1 Representative hydrogen uptake rate for the lysine hydrogen-
ation reaction. All experiments were carried out with 2.1 g (14.4 mmol)
L-lysine, 11.0 g H2O, 1.0–1.3 equiv. H2SO4, 120 °C. (a) Diamonds: 0.253 g
conditions: 30 g 10 wt% aqueous L-lysine (free base), 0.74
5% Ru/C, 52.0 bar Pi H2; (b) circles: 0.25 g 5% Ru/C, 69.7 bar Pi H2; (c)
equiv. H2SO4, 3 g catalyst, 120 °C. 55.2 bar H2, 24 h. The 5% triangles: 0.50 g 5% Ru/C, 68.8 bar Pi H2.
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 4575–4586 | 4579

View Article Online
Paper
Green Chemistry
pressure drop to complete lysine consumption is 16 bar, or
23% of the initial pressure. However, the uptake rate is linear
for over 90% of the total uptake and thus lysine conversion.
That may be contrasted with a hydrogenation at lower initial
pressure (Pi 52.0 bar, Fig. 1 curve a) where the initial uptake
rate begins to deviate from linearity at 75% conversion. The
change in initial gas uptake rates in experiments a and b of
Fig. 1, varying only in Pi, is consistent with first-order behavior
in hydrogen pressure. Thus the initial rate at Pi 52.0 bar (0.014
bar min−1) is decreased from that obtained at Pi 69.7 bar
(0.019 bar min−1) by exactly the same amount expected for
first order dependence.
Unlike hydrogen, lysine is completely consumed in these
reactions. Therefore, the linearity of the gas uptake through
3–4 half-lives (>90% lysine conversion) is consistent with zero
order dependence on lysine concentration. Finally, compari-
son of experiments b and c in Fig. 1 indicate a first order
dependence on catalyst.
These observations are consistent with the literature which,
for other amino acids, have shown dependencies on hydrogen
that are approximately first order below 69 bar but change to
zero order at higher pressures.26 The previously reported rate
dependencies on amino acid concentration are complex and
structure dependent, but zero order behavior has been
observed with some amino acids. The lysine hydrogenation
rates reported compare favorably with those reported for other
amino acids, for example alanine.k
Fig. 2 Dependence of lysine conversion on initial pH. Reaction con-
ditions – feed: 10 wt% lysine (aqueous free base or monohydrochloride),
3 wt% Ru/C catalyst, H2SO4 as required to achieve the indicated initial
pH, Temperature – 120 °C, pressure – 55.2 bar. Reaction time – 24 h.
Table 1 pKa data
Compound
pKa1
pKa2
pKa3
10.5
12.4
Lysine
Lysinola
H3PO4
H2SO4
2.2
10.5
2.2
8.8
12.8
7.2
Lysine hydrogenation is pH dependent. From the foregoing
discussion it would be expected that a minimum pH must be
achieved to establish a substantial concentration of the requi-
site lysine dication substrate. As shown in Fig. 2 this is indeed
observed, and the initial pH must be below 1.8–2.0 to achieve
high lysine conversion. At higher pH the conversion drops, but
not completely at intermediate pH, and has been attributed to
the action of acid sites present on the carbon support.
−3
2.0
a Estimated pKa obtained from: Scifinder, Chemical Abstracts Service:
Columbus, OH, 2014; RN 25441-01-4 (accessed 6 May 2014); calculated
using ACD/Labs Software version 11.02; ACD/Labs 1994–2011.
Table 2 pH shift during lysine hydrogenation
examplesa
–
representative
Because lysinol is a stronger base than lysine (Table 1), it is
expected that the solution pH should increase as the hydrogen-
ation progresses. As shown in Table 2, this is in fact observed,
where the pH increases by 2–4 units upon reaction com-
pletion. This result has significant implications for batch
Initial pH
Final pH
Lysine conv. %
2.2
1.7
1.3
6.9
5.7
3.5
75
Not determined
>99
hydrogenations conducted without in situ pH control such as a Conditions: 20 wt% L-lysine, H2SO4 as required, 8 wt% Ru/C, 120 °C,
48.3 bar, 7–24 h.
those reported here. If insufficient acid is present at the begin-
ning of the hydrogenation the pH rise will eventually cause the
lysine dication concentration to decrease to the point that
improved process would maintain constant and higher pH
range by continuously feeding acid during the hydrogenation.
This concept requires further study and demonstration.
hydrogenation ceases and incomplete conversion occurs; this
phenomenon has been examined in more detail in the case of
alanine. Unlike the case of alanine, while we are able to
achieve quantitative lysine conversion at an initial pH of
1.6–1.8 it is also the case that lysinol selectivity is pH depen-
dent. Thus, at very low initial pH (<1.3–1.5) the selectivity
decreases due to increased cyclization to form piperidines. An
The data in Table 1 also provide guidance with respect to
optimum acids to use for pH control. Our initial investigation
of lysine hydrogenation employed phosphoric acid, using gui-
dance from the literature. Examination of the pKa data indi-
cates that two equivalents of H3PO4 are required to convert
lysine to substantial concentrations of the requisite dication
kFrom Fig. 1 of ref. 24 we estimate alanine hydrogenation rates of ca. 4–8 mmol
(g-cat hour)−1 at 100–125 °C (69 bar, 100% conversion). The rate of lysine hydro-
genation at 120 °C in the present work is ca. 3–4 mmol (g-cat hour)−1 (69 bar,
100% conversion).
[lysineH2]2+. On the other hand, the acidity of HSO4 is
−
sufficient to generate lysine dication and therefore only a
single equivalent of H2SO4 is needed to achieve complete
4580 | Green Chem., 2014, 16, 4575–4586
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
lysine conversion. The choice of acid also has significant
implications in product isolation, which involves the addition
of base to convert the aqueous lysinol salt to the free base.
Phosphoric acid is triprotic and requires at least two molar
equivalents to effect lysine hydrogenation, thus requiring at
least 6 molar equivalents of base to liberate free base lysinol.
Sulfuric acid, diprotic and requiring a single molar equivalent,
requires only 2 equivalents of base for neutralization and
therefore generates far less salt byproduct. As discussed in
further detail below, it is notable that lysine manufacturers
produce the chloride and sulfate salts directly from their
fermentation processes.
Lysinol selectivity decreases with increasing temperature as
shown in Fig. 3. At 100 °C the lysinol selectivity exceeds 95%,
dropping slightly to 93–95% at 120 °C. The primary by-product
observed by 1H NMR is 1,5-diaminohexane, characterized by
the doublet at δ 1.29 (J = 6.7 Hz) and supported by LCMS
([MH+] = 117 amu). Trace amounts of other products such as
2-hydroxymethylpiperidine ([MH+] = 116 amu), 2-aminomethyl-
piperidine ([MH+] = 114 amu) 1-hexylamine ([MH+] = 102 amu)
are also observed by LCMS and confirmed by comparison to
authentic samples. At 150 °C the selectivity drops more
severely to 65–70%, with increasing amounts of these bypro-
ducts in addition to 2-methylpiperidine (δ 1.28 d, J = 6.6 Hz;
[MH+] = 100 amu)) and piperidine (δ 3.15 d, 1.77 quin, 5.8 Hz;
[MH+] = 85 amu). At 180 °C no lysinol is observed and the
primary products are 2-methylpiperidine and piperidine.
Most of the experiments shown in Fig. 3 were conducted for
24 h, sufficient time to ensure complete lysine conversion
although gas uptake ceased well before this time (120 °C, 10 h;
150 °C, 6 h; 180 °C, 3 h); the 100 °C experiment was held at
Fig. 4 Proposed reaction scheme for the by-product pathways during
lysine hydrogenation reaction.
that temperature for 45 h prior to analysis. Additional details
(NMR, LCMS) may be found in the ESI.†
A reaction scheme consistent with these observations is
shown in Fig. 4 and is supported by the following obser-
vations. First, treating aqueous lysinol at 180 °C, 70 bar H2,
and pH 3.1 in the presence of 5% Ru/C resulted in gas uptake
for approximately 7 h before halting. Analysis at this time
showed complete lysinol consumption and formation of
2-methylpiperidine (80%) and piperidine (10%) as the major
products, in addition to smaller amounts of 1-hexylamine and
2-amino-1-hexanol. A similar reaction, again starting with
lysinol but at 150 °C, resulted in ca. 15% lysinol conversion
with 2-hydroxymethylpiperidine obtained as the major product
(selectivity – 70%) and piperidine (20%) and 2-methylpiperi-
dine (10%) as side products. Finally, treating 2-hydroxymethyl-
piperidine with H2 (69 bar) at 180 °C produced largely
piperidine (84%), 2-methylpiperidine (5%), and the usual slate
of other trace products as observed with lysinol. These exper-
iments are consistent with and support the reaction scheme
proposed in Fig. 4.
Catalyst lifetime is critical in any process and was con-
sidered in the present study. A series of batch lysine hydrogen-
ations was performed wherein the 5% Ru/C was separated
from the aqueous reaction solution by filtration, rinsed with
water, dried, and then added back to the reactor for another
hydrogenation. No attempt was made to replenish catalyst
losses due to physical handling during the filtration or other
manipulations involved. Lysine conversions and selectivities
were unchanged over 6 batch hydrogenations conducted in
this manner (experimental details are provided in the ESI†).
Analyses of the filtrates from these experiments showed no
detectable ruthenium (<1 ppm). Catalyst lifetimes for a com-
Fig. 3 Lysine hydrogenation product distribution (selectivity in %) at
different temperatures. Experimental conditions: 14 wt% aqueous
L-lysine, 1.0 equiv. H2SO4, 3.4 wt% Ru/C catalyst, 68.9 bar H2, tempera-
ture as indicated. Lysinol, black; 1,5-diaminohexane, white; 2-substituted
pyridines, gray; piperidine, diagonals; other, dash fill. Reaction time –
24 h for runs at 120–180 °C; 45 h at 100 °C run.
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 4575–4586 | 4581

View Article Online
Paper
Green Chemistry
mercially viable process will likely be on the order of weeks to pure lysinol in 50–70% isolated yields. Given the high lysine
months and therefore these encouraging results should be conversion and lysinol selectivity observed by analyses (GC,
considered as preliminary.
LC, NMR) of the raw hydrogenation solutions, higher isolated
yields should be possible. The efficiency with which free base
lysinol is extracted from the sulfate salt mass has been identi-
Lysinol from commercial lysine
The experiments described to this point provide an insight fied as one critical parameter impacting the final yield.
regarding experimental conditions (e.g., catalyst, acidity/pH, Another yield loss occurs during the vacuum fractional distilla-
temperature, etc.) required for the effective conversion of tion of the crude lysinol where an as yet unidentified
lysine to lysinol. All of these experiments were carried out decomposition reaction occurs in the distillation pot. Pure
using high purity L-lysine free base obtained from Sigma- lysinol can be redistilled with very high recovery, suggesting
Aldrich®. As discussed earlier, more than 1.2 × 109 kg of lysine that an impurity or impurities present in the crude product
are produced annually throughout the world. This commer- result in decomposition when exposed to the high tempera-
cially manufactured material is almost entirely in the form of tures in the distillation pot. Process optimization and yield
salts, usually the monohydrochloride but more recently the improvement are the subject of future study.
sulfate has also become available. Much smaller amounts of
free base lysine are available as a 50% aqueous solution. In all
cases this lysine is of adequate purity for animal feed but is
not laboratory reagent or pharmaceutical grade. For example,
Determination of lysinol enantiopurity
Jere et al.27 report that the hydrogenation of alanine to alani-
nol under similar conditions to those reported here proceeds
with a high degree of stereoretention. For example, hydrogen-
ation of L-alanine at 100 °C gives L-alaninol with the enantio-
meric excess ranging from 83% to above 99% depending on
the specific reaction conditions. It was therefore of interest to
determine the enantiopurity of lysinol prepared by lysine
hydrogenation.
As expected, lysinol is rapidly and cleanly converted to the
bis-Mosher amides upon treatment with the Mosher acid
chloride (Scheme 4), presenting the use of NMR as a method
to determine enantiopurity.39,40 The reaction is selective for
the amine groups; the alcohol group reacts much more slowly
to form the ester, even in the presence of excess acid chloride
reagent. This allowed for the rapid and simple in situ gene-
ration of the bis-amide and subsequent evaluation by NMR, in
particular by 19F NMR. Fig. 5 shows representative 19F NMR
spectra of the bis-Mosher amides generated from reaction of
S-Mosher acid chloride with lysinol prepared from S-lysine (top),
racemic lysine (middle), and R-lysine (bottom). These spectra
clearly indicate a significant degree of enantiopurity in these
hydrogenation products. The top and bottom spectra are the
non-mirror image diastereomers RSR (top) and RRR (bottom).
The product prepared from racemic lysine is, as expected, an
the color of these feed grade products varies from off-white to
dark brown. To assess the feasibility of lysinol production, it is
important to determine if commodity lysine is a suitable raw
material and that impurities and the chloride counterion
found in most commercial material do not interfere with the
hydrogenation.
Three samples of commercial, feed-grade L-lysine were eval-
uated: L-lysine-HCl (off-white granules) and lysine (50 wt%)
free base (dark coffee-colored liquid) were obtained from
ADM. BioLys®, L-lysine sulfate, was obtained from Evonik and
received in the form of brown granules. These three samples
were each tested under typical hydrogenation conditions after
adjusting the solution to pH ∼ 1.5 as needed with H2SO4 (30 g
of 10 wt% aqueous lysine, 3 g 5% Ru/C, 120 °C, 55.2 bar,
24 h). All three commercial samples were cleanly hydrogenated
to lysinol with quantitative lysine conversion and selectivities
of approximately 95% and thus showed no difference in
activity or selectivity compared to high purity lysine.
Lysinol synthesis and isolation
Once the hydrogenation conditions became well-understood
the reaction was scaled up to produce sufficient quantities for
evaluation in various applications. As described in the Experi-
mental section, this hydrogenation has been conducted at
multi-liter scale (typically 15–20 wt% aq. lysine, 1.05 equiv.
H2SO4, 120 °C, 48.3 bar, 5% Ru/C, 4–20 L autoclave). Although
the catalyst loadings relative to lysine substrate are somewhat
large (20–25 wt/wt%), the catalyst from large scale preparations
is routinely isolated and reused with little to no loss in activity.
Lysinol isolation involves catalyst filtration and adjustment
to high pH (ca. 12) by the addition of NaOH. Calcium hydro-
xide and strong base ion exchange resins may also be used to
generate the free base solution. Water is then removed by dis-
tillation and the resulting semi-solid, a mixture of Na2SO4 and
lysinol free base, is extracted with an alcohol (EtOH, MeOH) to
separate the lysinol from the insoluble salt. After solvent
removal the resulting liquid is fractionally distilled to provide,
after a forecut containing the aforementioned piperidines, Scheme 4 Formation of RSR lysinol Mosher amide.
4582 | Green Chem., 2014, 16, 4575–4586
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
in diastereomer 2 (SSS) relative to diastereomer 1 (RSR).
Indeed, the diastereotopic methylene protons shift upfield by
0.0131 and 0.0604 ppm in 2 relative to 1. The 19F NMR chemi-
cal shifts are also diagnostic. Thus, the CF3 in diastereomer 2
is shifted 0.090 ppm downfield relative to that in 1, again
exactly as predicted by the Mosher analysis. These results
demonstrate retention of configuration in the hydrogenation.
Epoxy thermosets
As previously discussed, the ethyleneamines and ethanol-
amines are used in a number of applications. One of the more
familiar applications is in epoxy thermosets commonly used
as adhesives and coatings. In this application, an epoxy resin
is combined with a “hardener” to produce the thermoset.45
Polyamines such as diethylenetriamine (DETA) and triethyl-
enetetramine (TETA) are common hardeners and, for example
are employed in consumer two-part epoxy adhesives. Besides
being petrochemical-based, non-renewable materials, DETA
and TETA are potent skin irritants and sensitizers,15,46 provid-
ing additional incentive for alternatives. For these reasons, as
well as the obvious structural and functional group similarity,
the use of lysinol as an epoxy hardener was investigated and
its performance was compared to DETA.
Fig. 5 19F NMR (377 MHz) spectra of lysinol bis-Mosher amide dia-
stereomers. Legend: A: Product of reaction of S-Mosher acid chloride
and lysinol prepared by S-lysine hydrogenation and assigned to RSR dia-
stereomer (refer to Scheme 4 and text for diastereomer designations);
B, racemic mixture of RSR and RRR obtained from S-Mosher acid
chloride and lysinol prepared by S,R-lysine hydrogenation; C, RRR dia-
stereomer obtained from S-Mosher acid chloride and lysinol prepared
by R-lysine hydrogenation.
Bisphenol A diglycidyl ether (BADGE, Fig. 6) was chosen as
the epoxy resin due to its ubiquitous presence in amine-cured
epoxy thermosets. Indeed, over 75% of epoxy thermosets
employ BADGE as the epoxy resin.45 During thermoset for-
mation each N–H bond in the amine hardener ring opens an
epoxide to give a 1,2-aminoalcohol group. It is for this reason
that the stoichiometry of these epoxy thermosets is typically
set at two epoxy groups per primary amine (RNH2). The mul-
tiple number of amine and epoxy groups in the hardener and
resin, respectively, results in a highly cross-linked three dimen-
sional network as depicted in Scheme 5.
In commercial applications epoxy thermosets are often
complex formulations tailored to their intended application.
These formulations may include, besides the epoxy resin and
hardener, curing catalysts, fillers, strengthening agents such
as silica and fiberglass, and diluents for rheology control. For
equal mixture of this diastereomeric pair. The results
were further confirmed by 19F NMR evaluation of the products
of reaction of enantioenriched and racemic lysinol with
R-Mosher acid chloride.
The enantiopurity was determined by peak deconvolution
(details provided in ESI†) or integration of the 19F NMR
spectra and indicate 88–96% ee for lysinol prepared under
these conditions. These results are entirely consistent with
those reported previously, where alanine hydrogenation at
125 °C provided alaninol with ee’s spanning 83–98% depend-
ing on the other reaction parameters.27
The absolute configurations of these derivatives can be
1
deduced from the H and 19F NMR chemical shift data follow-
ing the analysis developed by Sullivan et al.41–44 Consider, for
example, stereoisomers 1 and 2, differing in the configuration
of the α-methoxy-α-trifluoromethylphenyl groups only. If the
configuration at C2 is S as indicated, and assuming that the
hydroxymethyl group at C2 is sterically smaller than
the (CH2)4NH(COR) group, the Mosher analysis predicts that
the hydroxymethyl methylene protons will be shifted upfield
Fig. 6 Structure of BADGE.
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 4575–4586 | 4583

View Article Online
Paper
Green Chemistry
Table 3 Physical property comparison of epoxy thermosets derived
from renewable lysinol and petrochemical DETAa
Lysinol
4.6(1.1)
DETA
Lap shear
Tensile strength (MPa)
4.7(0.5)
Compression
Compression modulus (MPa)
Strain at yield (%)
Stress at yield (MPa)
1678(133)
11.6(0.5)
103(1)
1489(69)
16.3(0.4)
104(1)
Nanoindentation
Instrumented hardness (MPa)
228(48)
199(35)
Scheme 5 Depiction of epoxy thermoset formation between lysinol
and a generic bis-epoxide.
Reduced modulus (MPa)
3506(412)
3287(279)
Chemical resistance
% Mass gain
50% aq. NaOH
30% aq. H2SO4
H2O
3.5% aq. NaCl
Acetone
the purposes of this investigation, which is simply to compare
lysinol with DETA, thermosets were prepared from these
amines and BADGE only and no optimization or formulation
development was performed. The thermosets were prepared
using stoichiometrically precise amounts of amine and
BADGE, thus 1 mole lysinol (4 moles N–H) per 2 moles of
resin (4 moles epoxide; see Scheme 5), and 1 mole DETA
(5 moles N–H) per 2.5 moles resin (5 moles epoxide). After
mixing the two components thoroughly at room temperature
the resulting viscous liquid were poured into a mold or
applied to coupons as required. The samples were allowed to
set at ambient temperature for up to 24 h. The resulting tack-
free solids were then cured at 60–100 °C. This procedure gave
clear, colorless, thermosets in the case of both lysinol and
DETA. Samples with stoichiometrically unbalanced amounts
of hardener and epoxy ( 15%) were also prepared and evalu-
ated to determine the sensitivity of the properties investigated
to this parameter. Although some variation was observed,
within this range of stoichiometry the effects were small and
of no impact or importance to the purposes of this evaluation
and comparison.
5.5(2.5)
3.8(1.0)
2.4(0.2)
2.6(0.3)
6.9(2.3)
0.4(0.1)
5.0(1.1)
4.4(0.5)
2.3(0.4)
1.9(1.5)
5.4(1.9)
0.2(0.4)
Toluene
a Value in parentheses is one standard deviation.
The resulting thermosets were subjected to a number of
tests to evaluate their mechanical and adhesive properties, and
chemical resistance, as described in the Experimental section.
The results are summarized in Table 3. Within the experi-
mental errors, the thermosets prepared from lysinol and DETA
are indistinguishable.
Fig. 7 Thermogravimetric analyses (TGA) of lysinol and DETA epoxy
thermosets. Experimental conditions: initial temperature: 30 °C; temp-
erature ramp: 10 °C min−1. Solid line: lysinol thermoset; dashed line:
DETA thermoset.
The thermal stabilities of the two thermosets were also eval-
uated using TGA, and the results are provided in Fig. 7.
Although these data indicate a slightly lower thermal stability
for the lysinol thermoset it shows excellent stability to
300–310 °C, as compared to 350–360 °C observed for DETA.
hydrogenation reaction, Ru/C gives the highest lysinol yields.
Lysine hydrogenation at higher temperature results in the for-
mation of piperidines. The feasibility of lysinol synthesis from
commercially available animal feed-grade lysine sources is
shown.
Structurally and chemically lysinol resembles di- and poly-
functional petrochemical amines, particularly ethyleneamines
and ethanolamines. These amines, besides being derived from
Conclusions
A comprehensive experimental study of lysinol synthesis from
lysine, a renewably sourced large scale commodity amino acid,
has been conducted. Lysine hydrogenation occurs under rela- non-renewable resources, are manufactured from toxic and
hazardous raw materials via processes that produce complex
tively modest conditions (100–150 °C, 48–70 bar, pH 1.5–2) in
mixtures of products requiring energy-intensive fractional dis-
benign water solvent and yields water as the only direct reac-
tion byproduct. Amongst the various catalysts screened for this tillation. These facts alone suggest the potential for lysinol as a
4584 | Green Chem., 2014, 16, 4575–4586
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
replacement for ethylene- and ethanolamines. This point is Finally, the authors thank E. I. DuPont de Nemours and Co.
further underscored by the demonstration that an epoxy for support and permission to publish this work.
thermoset prepared from lysinol possesses properties indistin-
guishable from one derived from a commercial ethyleneamine.
Bio-based epoxy thermosets have received considerable recent
attention but in all cases the focus has been on renewably
References
sourced epoxy resins and not the hardeners.47–53 Lysinol pro-
vides a bio-based amine hardener to complement bio-based
epoxy resins and thus advances the area closer to a fully bio-
based, renewable thermoset.
1 W. Pfefferle, B. Möckel, B. Bathe and A. Marx, in Microbial
Production of l-Amino Acids, ed. R. Faurie, J. Thommel,
B. Bathe, V. G. Debabov, S. Huebner, M. Ikeda, E. Kimura,
A. Marx, B. Möckel, U. Mueller and W. Pfefferle, Springer,
Berlin, Heidelberg, 2003, ch. 3, vol. 79, pp. 59–112.
2 K. Drauz, I. Grayson, A. Kleemann, H.-P. Krimmer,
W. Leuchtenberger and C. Weckbecker, in Ullmann’s Ency-
clopedia of Industrial Chemistry, Wiley-VCH Verlag, 2000.
3 K. Araki and T. Ozeki, in Kirk-Othmer Encyclopedia of Chemi-
cal Technology, John Wiley & Sons, 2000.
4 Ajinomoto Co. Inc., Contributing to a Low-Carbon Society,
http://www.ajinomoto.com/en/activity/csr/earth/climate.html,
accessed June 12, 2014.
5 S. Schaffer and T. Haas, Org. Process Res. Dev., 2014, 18,
752–766.
6 R. A. Sheldon, Green Chem., 2014, 16, 950–963.
7 T. Werpy, G. Petersen, A. Aden, J. Bozell, J. Holladay,
J. White, A. Manheim, D. Eliot, L. Lasure and S. Jones, Top
value added chemicals from biomass. Volume 1-Results of
screening for potential candidates from sugars and synthesis
gas, DTIC Document, 2004.
8 T. M. Lammens, J. Potting, J. P. M. Sanders and I. J. M. De
Boer, Environ. Sci. Technol., 2011, 45, 8521–8528.
9 T. M. Lammens, D. De Biase, M. C. R. Franssen, E. L. Scott
and J. P. M. Sanders, Green Chem., 2009, 11, 1562–1567.
The concept and results presented here address most of the
12 Principles of Green Chemistry.54 Conversion and selectivity
to lysinol from lysine is high (Principle 1), a renewable feed-
stock is used as raw material (Principle 7), and the use of
highly toxic and explosive starting materials (EO, EDC) is
avoided (Principles 3, 4, and 12). The catalytic hydrogenation
is conducted using water as solvent under relatively mild con-
ditions (Principles 5, 6 and 9), although reduced pressure and
catalyst loadings represent significant potential improvements.
The E-factor (Principle 2) for lysine hydrogenation is 1.0–1.2,
in the range of bulk chemical processes. Besides co-product
water this value takes into account the salt formation (Na2SO4
or 2 NaCl) resulting from the counterion introduced with the
lysine starting material and required to achieve the low pH
conditions for the hydrogenation and the resulting neutraliz-
ation. Because the fermentation process produces lysine as a
salt almost all downstream uses of lysine to manufacture other
chemicals will generate salt co-products. Nonetheless, this
value compares favorably with the current processes to manu-
facture ethyleneamines with E-factors ranging from theoretical
maxima of 0.3 (2EO + 3NH3 = DETA + H2O) to 2.3 (2EDC +
3NH3 + 4NaOH = DETA + 4NaCl). Note that the latter calcu- 10 S. Kind and C. Wittmann, Appl. Microbiol. Biotechnol., 2011,
lations do not attempt to include correction for the complex 91, 1287–1296.
product mixtures generated by these processes in practice, but 11 J. W. Frost, WO Pat, 2005/123669A1, 2005.
instead assume direct conversion to a single product. Recycle 12 D. A. Wicks, WO Pat, 2010/011967A1, 2010.
or treatment of the water used and generated in this process 13 A. V. Pukin, C. G. Boeriu, E. L. Scott, J. P. M. Sanders and
also must eventually be considered. We have conducted lysine
M. C. R. Franssen, J. Mol. Catal. B: Enzym., 2010, 65, 58–62.
hydrogenation at up to 35 wt% lysine concentration without 14 S. Sridhar and R. G. Carter, in Kirk-Othmer Encyclopedia of
difficulty, thus demonstrating minimum water use. Finally, the
toxicity, biodegradation, and environmental fate character-
istics of lysinol require investigation.
To the best of our knowledge, this is the first study report-
ing a potentially scalable and economical lysinol synthesis and
demonstration of the potential of both lysine and lysinol as a
Chemical Technology, John Wiley & Sons, Inc., 2000, DOI:
10.1002/0471238961.0409011303011820.a01.pub2.
15 K. Eller, E. Henkes, R. Rossbacher and H. Höke, in Ull-
mann’s Encyclopedia of Industrial Chemistry, Wiley-VCH
Verlag GmbH & Co. KGaA, 2000, DOI: 10.1002/14356007.
a02_001.
replacement for petrochemical amines. The use of lysinol in 16 M. Frauenkron, J.-P. Melder, G. Ruider, R. Rossbacher and
the dozens of other applications currently addressed by petro-
chemical amines awaits further exploration and exploitation.
H. Höke, in Ullmann’s Encyclopedia of Industrial Chemistry,
Wiley-VCH Verlag GmbH & Co. KGaA, 2000, DOI: 10.1002/
14356007.a10_001.
17 C. Jones, M. R. Edens and J. F. Lochary, in Kirk-Othmer
Encyclopedia of Chemical Technology, John Wiley & Sons,
Inc., 2000, DOI: 10.1002/0471238961.0112110105040514.
a01.pub2.
Acknowledgements
The authors would like to thank Janice Hytrek, Ellen Baldas-
sare, Charles Bellini, Scott Hutchison, and Joe Colombo for 18 K. A. Weissermel and H.-J. Arpe, Industrial Organic Chem-
their help and dedication in conducting experiments reported istry, VCH, New York, NY, 2nd edn, 1993.
in this paper. Also, thanks are expressed to Laurie Howe for 19 A. Peeters, L. Claes, I. Geukens, I. Stassen and D. De Vos,
assistance with NMR product analysis and quantification.
Appl. Catal., A, 2014, 469, 191–197.
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 4575–4586 | 4585

View Article Online
Paper
Green Chemistry
20 M. Tamura, R. Tamura, Y. Takeda, Y. Nakagawa and 38 A. K. Saund, B. Prashad, A. K. Koul, J. M. Bachhawat and
K. Tomishige, Chem. Commun., 2014, 50, 6656–6659. N. K. Mathur, Int. J. Pept. Protein Res., 1973, 5, 7–10.
21 H. Urtel, M. Roesch, A. Hauner and M. Schubert, US Pat, 39 S. Lee, Chem. Educ., 2004, 9, 359–363.
7 507 866, 2009.
40 J. A. Dale, D. L. Dull and H. S. Mosher, J. Org. Chem., 1969,
22 K. P. Pimparkar, D. J. Miller and J. E. Jackson, Ind. Eng.
Chem. Res., 2008, 47, 7648–7653.
23 W. Mägerlein, C. Dreisbach, H. Hugl, M. K. Tse,
34, 2543–2549.
41 T. R. Hoye, C. S. Jeffrey and F. Shao, Nat. Protoc., 2007, 2,
2451–2458.
M. Klawonn, S. Bhor and M. Beller, Catal. Today, 2007, 121, 42 B. S. Joshi and S. W. Pelletier, Heterocycles, 1999, 51, 183–
140–150. 184.
24 H. Urtel, M. Roesch, A. Hauner and M. Schubert, WO Pat, 43 M. J. Rieser, Y. H. Hui, J. K. Rupprecht, J. F. Kozlowski,
2005/077871A1, 2005.
25 H. Urtel, M. Roesch and A. Hauner, WO Pat, 2005/
077870A1, 2005.
26 F. T. Jere, J. E. Jackson and D. J. Miller, Ind. Eng. Chem.
Res., 2004, 43, 3297–3303.
27 F. T. Jere, D. J. Miller and J. E. Jackson, Org. Lett., 2003, 5,
527–530.
K. V. Wood, J. L. McLaughlin, P. R. Hanson, Z. Zhuang and
T. R. Hoye, J. Am. Chem. Soc., 1992, 114, 10203–10213.
44 G. R. Sullivan, J. A. Dale and H. S. Mosher, J. Org. Chem.,
1973, 38, 2143–2147.
45 H. Q. Pham and M. J. Marks, Epoxy Resins, Kirk-Othmer
Encyclopedia of Chemical Technology, John Wiley & Sons,
Inc., New York, 2000.
28 S. Antons, A. Tilling and E. Wolters, US Pat, US 6 310 254 46 Hazardous Substances Data Bank [Internet]. Bethesda
B1, 2001.
(MD): National Library of Medicine (US), Division of
Specialized Information Services. 1986 [cited 2014
June 18]. Available from: http://toxnet.nlm.nih.gov/cgi-bin/
sis/search2/r?dbs+hsdb:@term+@rn+111-40-0.
29 S. Antons, A. S. Tilling and E. Wolters, WO Pat, 9938838A1,
1999.
30 S. Antons and B. Beitzke, EP Pat, 696575A1, 1996.
–
31 S. Nishimura, in Handbook of Heterogeneous Catalytic 47 S. Benyahya, C. Aouf, S. Caillol, B. Boutevin, J. P. Pascault
Hydrogenation for Organic Synthesis, John Wiley & Sons,
2001, ch. 10.
32 N. Kihara, Y. Kushida and T. Endo, J. Polym. Sci., Part A:
Polym. Chem., 1996, 34, 2173–2179.
33 K. Liu, X. Zhang, X. Tao, J. Yan, G. Kuang, W. Li and
A. Zhang, Polym. Chem., 2012, 3, 2708–2711.
34 E. K. Efthimiadou, C. Tapeinos, P. Bilalis and G. Kordas,
J. Nanopart. Res., 2011, 13, 6725–6736.
and H. Fulcrand, Ind. Crops Prod., 2014, 53, 296–307.
48 S. Ma, X. Liu, L. Fan, Y. Jiang, L. Cao, Z. Tang and J. Zhu,
ChemSusChem, 2014, 7, 555–562.
49 K. Huang, Z. Liu, J. Zhang, S. Li, M. Li, J. Xia and Y. Zhou,
Biomacromolecules, 2014, 15, 837–843.
50 P.-Y. Kuo, M. Sain and N. Yan, Green Chem., 2014, 16,
3483–3493.
51 L. Cao, X. Liu, H. Na, Y. Wu, W. Zheng and J. Zhu, J. Mater.
Chem. A, 2013, 1, 5081–5088.
35 A. Zhang, Macromol. Rapid Commun., 2008, 29, 839–845.
36 S. Miyanaga, T. Obata, H. Onaka, T. Fujita, N. Saito, 52 C. Aouf, J. Lecomte, P. Villeneuve, E. Dubreucq and
H. Sakurai, I. Saiki, T. Furumai and Y. Igarashi, J. Antibiot.,
2006, 59, 698–703.
37 X. Huang, B. H. Rickman, B. Borhan, N. Berova and
K. Nakanishi, J. Am. Chem. Soc., 1998, 120, 6185–6186.
H. Fulcrand, Green Chem., 2012, 14, 2328–2336.
53 M. Chrysanthos, J. Galy and J.-P. Pascault, Polymer, 2011,
52, 3611–3620.
54 R. A. Sheldon, Chem. Soc. Rev., 2012, 41, 1437–1451.
4586 | Green Chem., 2014, 16, 4575–4586
This journal is © The Royal Society of Chemistry 2014