Effect of ancillary (aminomethyl)phenolate ligand on efficacy of aluminum-catalyzed glucose dehydration to 5-hydroxymethylfurfural

Article abstract of DOI:10.1016/j.poly.2018.03.035

Air-stable dimethylaluminum complexes L^RAlMe₂ that contain (aminomethyl)phenolate (L^R) ligands were prepared in high yield. NMR data and X-ray crystallographic characterization of the molecular structures of several of the complexes confirmed bidentate coordination of the (aminomethyl)phenolate ligand to aluminum. Efficient aluminum catalysts for glucose dehydration to HMF were generated via modification of the (aminomethyl)phenolate ligand. L^RAlMe₂ complexes containing bidentate (aminomethyl)phenolate ligands with an aryl substituent on the amino moiety are efficient catalysts for glucose dehydration to HMF in ionic liquid solvents. In [EMIM]Br and [BMIM]Br, the reaction proceeds at 120 °C to very high conversion in 2 h to produce HMF with 60–63% selectivity and in 58–60% yield. Evidently, L^RAlMe₂ complexes catalyze glucose isomerization to fructose at ≥120 °C while the HMF yield depends on the degree of competing HMF loss to humins formation. These results indicate that additional studies of ancillary ligand effects on aluminum-catalyzed glucose dehydration are needed to improve knowledge of structure–function relationships that are key to increasing the efficiency of aluminum catalysts for dehydration of glucose (and ultimately cellulose) to HMF.

Full text of DOI:10.1016/j.poly.2018.03.035

Accepted Manuscript
Effect of ancillary (aminomethyl)phenolate ligand on efficacy of aluminum-
catalyzed glucose dehydration to 5-hydroxymethylfurfural
Daudi Saang'onyo, Sean Parkin, Folami T. Ladipo
PII:
S0277-5387(18)30167-0
https://doi.org/10.1016/j.poly.2018.03.035
POLY 13101
DOI:
Reference:
Polyhedron
To appear in:
Received Date:
Accepted Date:
29 December 2017
30 March 2018
Please cite this article as: D. Saang'onyo, S. Parkin, F.T. Ladipo, Effect of ancillary (aminomethyl)phenolate ligand
on efficacy of aluminum-catalyzed glucose dehydration to 5-hydroxymethylfurfural, Polyhedron (2018), doi:
https://doi.org/10.1016/j.poly.2018.03.035
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

REVISED
Polyhedron
Effect of ancillary (aminomethyl)phenolate ligand on efficacy of aluminum-catalyzed
glucose dehydration to 5-hydroxymethylfurfural
Daudi Saang’onyo, Sean Parkin, Folami T. Ladipo*
Chemistry Department, University of Kentucky, Lexington, Kentucky 40506-0055, USA
ARTICLE INFO
ABSTRACT
Article history:
Received
Received in revised form
Accepted
Air-stable dimethylaluminum complexes LRAlMe2 that contain (aminomethyl)phenolate (LR)
ligands were prepared in high yield. NMR data and X-ray crystallographic characterization of
the molecular structures of several of the complexes confirmed bidentate coordination of the
(aminomethyl)phenolate ligand to aluminum. Efficient aluminum catalysts for glucose
dehydration to HMF were generated via modification of the (aminomethyl)phenolate ligand.
LRAlMe2 complexes containing bidentate (aminomethyl)phenolate ligands with an aryl
substituent on the amino moiety are efficient catalysts for glucose dehydration to HMF in ionic
liquid solvents. In [EMIM]Br and [BMIM]Br, the reaction proceeds at 120 ˚C to very high
conversion in 2 hours to produce HMF with 60-63% selectivity and in 58-60% yield. Evidently,
LRAlMe2 complexes catalyze glucose isomerization to fructose at ≥120 ˚C while the HMF yield
depends on the degree of competing HMF loss to humins formation. These results indicate that
additional studies of ancillary ligand effects on aluminum-catalyzed glucose dehydration are
needed to improve knowledge of structure-function relationships that are key to increasing the
efficiency of aluminum catalysts for dehydration of glucose (and ultimately cellulose) to HMF.
Available online
Keywords:
Glucose
5-Hydroxymethylfurfural
Aluminum catalysts
Dehydration
Ionic liquids
2009 Elsevier Ltd. All rights reserved.
1. Introduction
liquids and high boiling organic solvents with various Lewis acid
metal salts, such as CrCl2 [11-13], SnCl4 [14, 15], and AlCl3 [10,
Glucose is the most abundant monosaccharide in cellulosic
biomass hence efficient catalytic processes for its conversion into
chemicals and biofuels are highly desirable [1-3]. Glucose
dehydration is a promising method for synthesis of 5-
hydroxymethylfurfural (HMF), an emerging bio-derived platform
chemical that potentially could be used to produce a wide variety
of high-value chemicals [4, 5]. For example, HMF can be
converted by selective oxidation into 2,5-furandicarboxylic acid
(FDCA) which is attractive as a substitute for terephthalic acid in
plastics production [6, 7]. HMF can also undergo rehydration to
produce levulinic acid (LA) which itself is a promising platform
chemical that can be used as a feedstock for production of liquid
hydrocarbon fuels [8, 9].
16-18] as catalysts.
Given the much lower toxicity and cheaper cost of Al in
comparison to Cr and Sn, the development of efficient Al
catalysts for glucose conversion to HMF is receiving increased
attention [10, 16-20]. For example, Abu-Omar and coworkers
have reported that AlCl3 exhibits high glucose conversion
activity in water/THF biphasic medium to give HMF in 61%
yield [16]. Dumensic and coworkers have found that catalytic
conversion of glucose with the combination of AlCl3 and a
Brønsted acid (such as HCl) in a biphasic water/alkylphenol
solvent system gave 62% yield of HMF [10]. Rasrendra et al.
used both AlCl3 and Al(OTf)3 in DMSO for glucose conversion
to produce HMF in 50% and 60% yield, respectively [21]. Liu
and Chen showed that aluminum trialkyls (such as pyrophoric
AlMe3 and AlEt3) and trialkoxides (such as Al(OPri)3 and
Al(OBut)3) can give up to 50% HMF yield from glucose
conversion in [EMIM]Cl [20]. These studies indicate that
aluminum species hold strong promise as Lewis acid catalysts for
glucose conversion to HMF. However, the majority of studies
used (10-30%) AlCl3 in different solvents, and current knowledge
of ancillary ligand effects on the efficiency of glucose conversion
to HMF with aluminum Lewis acid catalysts is lacking. Herein,
we report a systematic study of the efficacy of easily prepared,
Investigations of glucose dehydration [5] using different
catalysts (such as organic and inorganic acids, Lewis acids, salts,
and zeolites) and solvents (including aqueous, organic, mixed
aqueous/organic, and ionic liquids) have established that glucose
conversion to HMF with Brønsted acids (such as HCl and H2SO4)
typically proceeds via direct dehydration of glucose to HMF
while with Lewis acid catalysts, the reaction typically proceeds
via formation of fructose [6, 10]. However, while mineral acids
usually give HMF in low yield and produce other byproducts,
moderate-to-high yields of HMF have been reported in ionic

air –stable dimethylaluminum complexes containing bidentate
(aminomethyl)phenolate ligands as catalysts for the conversion
of glucose to HMF in ionic liquids. We demonstrate that
effective catalysts for glucose dehydration to HMF can be
produced via modification of the (aminomethyl)phenolate ligand.
2-tert-Butyl-4-methylphenol (3.28 g, 19.98 mmol), N-
isopropylbenzylamine (2.98 g, 20.04 mmol) and
paraformaldehyde (0.90 g, 29.97 mmol) were charged into a
heavy-walled reaction vessel, equipped with a magnetic stir bar.
The vessel was capped tightly, then placed in an oil bath
maintained at 105 °C, and heated with stirring for 1 h. After
cooling to room temperature, the light-yellow reaction mixture
was dissolved in chloroform (50 mL). The solution was washed
with distilled water (5 x 15 mL) and dried over anhydrous
Na2SO4 for 12 h. After filtering off Na2SO4, the filtrate was
evaporated under reduced pressure to give a pale-yellow oil,
which was purified by silica gel column chromatography using
20:1 petroleum ether:ethyl acetate as eluent. The solution was
evaporated under reduced pressure to give 1c as a colorless oil.
The material was collected and dried under reduced pressure.
Yield: 5.40 g, 82.9 %. 1H NMR (400 MHz, CDCl3): δ 11.15 (br
2. Experimental
2.1. General comments
All manipulations of air- and/or moisture-sensitive
compounds were carried out under dry nitrogen atmosphere using
standard Schlenk or glovebox techniques. All solvents were dried
and distilled by standard methods [22] prior to use and stored in a
glovebox over 4A molecular sieves that had been dried in a
vacuum oven at 150 C for at least 48 h. All other chemicals
were used as received, unless otherwise stated. Toluene, THF,
ethanol, petroleum ether, n-hexane, chloroform, methylene
chloride, and methanol (all ACS grade) were purchased from
Pharmco-Aaper. Ethyl acetate and acetonitrile were purchased
from Fisher Scientific. 5-Hydroxymethylfurfural, (~99%), D-(–)-
fructose, D-(+)-glucose (≥ 99.5%), AlMe3 (2.0 M in hexane), 2-
tert-butyl-4-methylphenol (99%), N-methylbenzylamino (97%),
N-ethylbenzylamino (97%), N–isopropylbenzylamine (97%), N-
phenylaniline (99%), 4-methylaniline (99.6%), 4-chloroaniline
(98%), benzaldehyde (≥ 98%) and poly(methylhydrosiloxane)
were purchased from Sigma-Aldrich. Paraformaldehyde (96%),
1-ethyl-3-methylimidazolium chloride ([EMIM]Cl, 97%), and 1-
ethyl-3-methylimidazolium bromide ([EMIM]Br, 97%) were
purchased from Acros Organics. [EMIM]Cl and [EMIM]Br were
purified before use, via recrystallization according to the
literature method [23]. [BMIM]Br was synthesized and purified
by following literature methods [23, 24]. N-(tert-
butyl)benzylamino (99%) was purchased from Alfa Aesar. N-
benzyl-4-chloroaniline [25], N-benzyl-p-toluidine [25], 2-[(N-
benzyl-N-methyl)aminomethyl]-6-tert-butyl-4-methylphenol (1a)
[26], 2-(N-benzyl-N-ethyl-aminomethyl)-4-methyl-6-tert-butyl-
phenol (1b) [27], and LEtAlMe2 (2b, LEt = 2-(N-benzyl-N-ethyl-
4
s, 1H, OH), 7.28–7.15 (m, 5H, ArH), 6.92 (d, J = 2.0 Hz, 1H,
4
ArH), 6.64 (d, J = 2.0 Hz, 1H, ArH), 3.67 (s, 2H, ArCH2), 3.51
3
(s, 2H, PhCH2), 2.99 (sept., 1H, J = 7.2 Hz, CH(CH3)2), 2.18 (s,
3
3H, ArCH3), 1.37 (s, 9H, ArC(CH3)3), 1.06 (d, 6H, J = 7.2 Hz,
NCH(CH3)2). 13C{H} NMR (100 MHz, CDCl3): δ 154.7, 138.5,
136.4, 129.6, 128.7, 127.8, 127.5, 127.1, 126.6, 122.4 (all Ar–C),
54.0 (ArCH2), 52.7 (PhCH2), 48.3 (CH(CH3)2), 34.8 (C(CH3)3),
29.7 (ArC(CH3)3), 21.0 (ArCH3), 17.1 (NCH(CH3)2).
2.2.2. Synthesis of 2-[(N-benzyl-N-tert-butyl)aminomethyl]-6-
tert-butyl-4-methylphenol (1d).
This compound was prepared following a similar procedure to
that described for 1c, from 2-tert-Butyl-4-methylphenol (3.28 g,
19.98 mmol), N-tert-butylbenzylamine (3.27 g, 20.04 mmol) and
paraformaldehyde (0.60 g, 19.98 mmol). After purification of the
reaction product, a light-yellow oil, by silica gel column
chromatography (using 20:1 petroleum ether:ethyl acetate as
eluent), removal of the organic volatiles under reduced pressure
furnished a light-yellow oil that was recrystallized from hexane at
-20 ˚C, giving 1d as white crystals. The material was collected
and dried under reduced pressure. Yield: 4.46 g, 65.7 %. 1H
NMR (400 MHz, CDCl3): δ 10.98 (s, 1H, OH), 7.23–7.06 (m,
aminomethyl)-4-methyl-6-tert-butyl-phenolate)
[27]
were
4
4
prepared by the literature methods or modification thereof.
5H, ArH), 6.88 (d, J = 1.6 Hz, 1H, ArH), 6.61 (d, J = 1.6 Hz,
1H, ArH), 3.84 (s, 2H, ArCH2), 3.70 (s, 2H, PhCH2), 2.19 (s, 3H,
ArCH3), 1.37 (s, 9H, NC(CH3)3), 1.21 (s, 9H, ArC(CH3)3).
13C{H} NMR (100 MHz, CDCl3): δ 154.5, 140.9, 136.4, 128.9,
128.3, 127.2, 127.1, 126.8, 126.3, 124.0 (all Ar–C), 57.1
(NC(CH3)3), 54.6 (ArCH2), 54.4 (PhCH2) 34.7 (C(CH3)3), 29.7
(NC(CH3)3), 27.2 (ArC(CH3)3), 21.0 (ArCH3).
1H and 13C{1H} NMR spectra were recorded on a Varian
VXR-400 spectrometer at room temperature. All chemical shifts
are reported in units of δ (downfield from tetramethylsilane) and
were referenced to residual solvent peaks. FTIR spectra were
collected on a Thermo Scientific Nicolet 6700 ATR-FTIR
spectrometer fitted with a ZnSe crystal with a Smart iTR
accessory. The resolution of the instrument was set to 4 cm-1. The
background of the IR spectrum of air was first collected, and then
powdered samples were placed on the ZnSe crystal, pressed
against the crystal using the inbuilt high-pressure clamp and their
absorbance was measured. A total of 40 s scans were used for
both background and the samples. Raman spectra were collected
on a DXR Raman microscope (Thermo Fisher) spectrometer. The
source of radiation was a laser operated at 532 nm. The excitation
laser beam was focused on the sample using a microscope
equipped with a 10X lens. The laser power at the sample surface
was about 2 mW and the acquisition time for each spectrum was
20 s and recorded in the range of 50 – 3500 cm-1. X-ray
diffraction data were collected at 90.0(2) K on either a Nonius
kappaCCD, Bruker-Nonius X8 Proteum, or a D8 Venture
diffractometer. Elemental analysis for C, H, and N was
performed by Robertson Microlit Laboratories, Ledgewood, NJ.
2.2.3. Synthesis of 2-[(N-benzyl-N-phenyl)aminomethyl]-6-tert-
butyl-4-methylphenol (1e).
This compound was prepared following a similar procedure to
that described for 1c, from 2-tert-butyl-4-methylphenol (3.28 g,
19.98 mmol), paraformaldehyde (0.60 g, 19.98 mmol), and N-
phenylbenzylamine (3.67 g, 20.04 mmol). After purification of
the reaction product, a light-yellow oil, by silica gel column
chromatography (using 5:1 hexane:ethyl acetate as eluent),
removal of the organic volatiles under reduced pressure gave 1e
as white crystals. Yield: 4.53 g, 63.1 %. 1H NMR (400 MHz,
CDCl3): δ 9.61 (br s, 1H, OH), 7.28 –7.16 (m, 5H, ArH), 7.10–
4
6.96 (m, 6H, ArH), 6.71 (d, J = 1.6 Hz, 1H, ArH), 4.26 (s, 2H,
ArCH2), 4.22 (s, 2H, PhCH2), 2.24 (s, 3H, ArCH3), 1.41 (s, 9H,
C(CH3)3). 13C{H} NMR (100 MHz, CDCl3): δ 154.2, 149.4,
136.8, 136.3, 129.4, 129.3, 128.4, 128.0, 127.9, 127.6, 127.2,
123.7, 122.1, 122.0 (all Ar–C), 57.6 (ArCH2), 57.0 (PhCH2), 34.8
(C(CH3)3), 29.8 (C(CH3)3), 21.0 (ArCH3).
2.2. Synthesis of the proligands
2.2.1. Synthesis of 2-[(N-benzyl-N-isopropyl)aminomethyl]-6-
tert-butyl-4-methylphenol (1c).

2.2.4. Synthesis of 2-[(N-benzyl-N-p-toluidine)aminomethyl]-6-
tert-butyl-4-methylphenol (1f).
Complex 2c was obtained as a white powder, by following a
similar procedure to that described for 2a, from reaction between
AlMe3 (1.40 mL, 2.87 mmol, 2.0 M in hexane) and 1c (0.94 g,
This compound was prepared following a similar procedure to
that described for 1c, from 2-tert-butyl-4-methylphenol (1.42 g,
8.65 mmol), paraformaldehyde (0.26 g, 8.65 mmol), and N-
benzyl-p-toluidine [25] (1.71 g, 8.65 mmol), except the reaction
mixture was heated for 6 h. After purification by silica gel
column chromatography using 5:1 hexane:ethyl acetate as eluent,
1f was obtained as a white powder. Yield: 1.63 g, 50.4 %. 1H
NMR (400 MHz, CDCl3): δ 9.99 (s, 1H, OH), 7.22–7.18 (m, 3H,
1
2.87 mmol). Yield: 0.92 g, 83.9%. H NMR (400 MHz, CDCl3):
4
δ 7.37–7.21 (m, 5H, ArH), 7.03 (d, J = 2.0 Hz, 1H, ArH), 6.68
4
2
(d, J = 2.0 Hz, 1H, ArH), 4.23 (d, J = 14.0, 1H, ArCH2), 4.16
(d, 2J = 14.0 Hz, 1H, ArCH2), 3.95 (d, 2J = 14.0 Hz, 1H, ArCH2),
2
3
3.59 (d, J = 14.0 Hz, 1H, ArCH2), 3.17 (sept., 1H, J = 6.8 Hz,
CH(CH3)2), 2.27 (s, 3H, ArCH3), 1.38 (s, 9H, C(CH3)3), 1.36 (d,
3H, 3J = 7.2 Hz, CH(CH3)2), 1.25 (d, 3H, 3J = 6.8 Hz, CH(CH3)2),
-0.64 (AlCH3), -0.68 (AlCH3). 13C{H} NMR (100 MHz, CDCl3):
δ 156.6, 139.0, 132.2, 132.0, 129.0, 128.7, 128.6, 128.1, 125.3,
120.8 (all Ar–C), 55.4 (ArCH2), 53.8 (ArCH2), 52.2 (CH(CH3)2),
35.0 (C(CH3)3), 29.7 (C(CH3)3), 21.1 (ArCH3), 19.4 (CH(CH3)2),
19.1 (CH(CH3)2), -7.1 (AlCH3), -8.0 (AlCH3). Anal. Calcd. for
C24H36AlNO: C, 75.55; H, 9.51; N, 3.67. Found: C, 74.50; H,
9.48; N, 3.64%.
4
ArH), 7.06–6.95 (m, 7H, ArH), 6.72 (d, J = 1.6 Hz, 1H, ArH),
4.23 (s, 2H, ArCH2), 4.17 (s, 2H, ArCH2), 2.27 (s, 3H, NArCH3),
2.25 (s, 3H, ArCH3), 1.42 (s, 9H, C(CH3)3). 13C{H} NMR (100
MHz, CDCl3): δ 154.4, 146.8, 136.7, 136.5, 133.6, 129.9, 129.6,
128.4, 127.9, 127.7, 127.5, 127.1, 122.5, 122.2 (all Ar–C), 58.1
(ArCH2), 57.7 (PhCH2), 34.8 (C(CH3)3), 29.8 (C(CH3)3), 21.1
(NArCH3), 21.0 (ArCH3).
2.3.3. Synthesis of Lt-BuAlMe2 complex (2d)
Complex 2d was obtained as a light-yellow powder, by
following a similar procedure to that described for 2a, from
reaction between AlMe3 (1.0 mL, 2.00 mmol, 2.0 M in hexane)
and 1d (0.79 g, 2.00 mmol). Yield: 0.637 g, 80.5%. 1H NMR
(400 MHz, CDCl3): δ 7.37–7.31 (m, 2H, ArH), 7.30–7.25 (m,
2.2.5.
Synthesis
of
2-[(N-benzyl-N-(p-
chlorophenyl))aminomethyl]-6-tert-butyl-4-methylphenol (1g).
This compound was prepared following a similar procedure to
that described for 1c, from 2-tert-butyl-4-methylphenol (0.37 g,
2.25 mmol), paraformaldehyde (0.10 g, 3.38 mmol), and N-
benzyl-4-chloroaniline [25] (0.49 g, 2.25 mmol), except the
reaction mixture was heated for 50 h. After purification by silica
gel column chromatography using 20:1 petroleum ether:ethyl
acetate as eluent, 1g was obtained as a white powder. Yield: 0.66
4
4
3H, ArH), 7.00 (d, J = 2.0 Hz, 1H, ArH), 6.72 (d, J = 2.0 Hz,
2
2
1H, ArH), 4.47 (d, J = 14.8, 1H, ArCH2), 4.36 (d, J = 15.2 Hz,
2
2
1H, ArCH2), 4.26 (d, J = 15.2 Hz, 1H, ArCH2), 4.11 (d, J =
14.8 Hz, 1H, ArCH2), 2.26 (s, 3H, ArCH3), 1.37 (s, 9H,
NC(CH3)3), 1.26 (s, 9H, ArC(CH3)3), -0.56 (AlCH3), -0.63
(AlCH3). 13C{H} NMR (100 MHz, CDCl3): δ 156.3, 139.4,
135.1, 131.8, 128.8, 128.5, 128.1, 127.6, 125.4, 121.2 (all Ar–C),
62.4 (NC(CH3)3), 52.6 (ArCH2), 52.3 (PhCH2) 35.0 (C(CH3)3),
29.6 (NC(CH3)3), 28.2 (ArC(CH3)3), 21.1 (ArCH3), -4.7 (AlCH3),
–7.3 (AlCH3). Anal. Calcd. for C25H38AlNO (%): C, 75.91; H,
9.68; N, 3.54. Found: C, 75.31; H, 9.97; N, 3.51.
1
g, 73.9 %. H NMR (400 MHz, CDCl3): δ 9.26 (s, 1H, OH),
7.23–7.17 (m, 5H, ArH), 7.03–6.96 (m, 5H, ArH), 6.70 (d, 1H, 4J
= 1.6 Hz, ArH), 4.23 (s, 2H, ArCH2), 4.22 (s, 2H, ArCH2), 2.24
(s, 3H, ArCH3), 1.40 (s, 9H, C(CH3)3). 13C{H} NMR (100 MHz,
CDCl3): δ 154.0, 147.9, 136.9, 136.0, 129.4, 129.3, 128.6, 128.1,
127.9, 127.8, 127.4, 123.2, 121.8 (all Ar–C), 57.8 (ArCH2), 56.8
(ArCH2), 34.8 (C(CH3)3), 29.8 (C(CH3)3), 21.1 (ArCH3).
2.3.4. Synthesis of LPhAlMe2 complex (2e)
2.3. Synthesis of Aluminum Complexes
Complex 2e was obtained as a white powder, by following a
similar procedure to that described for 2a, from reaction between
AlMe3 (1.74 mL, 3.48 mmol, 2.0 M in hexane) and 1e (1.25 g,
2.3.1. Synthesis of LMeAlMe2 complex (2a)
1
AlMe3 (4.50 mL, 8.97 mmol, 2.0 M in hexane) was added
dropwise to a toluene (20 mL) solution of 2-[(N-benzyl-N-
methyl)aminomethyl]-6-t-butyl-4-methylphenol [26] (1a, 2.67 g,
8.97 mmol) at room temperature. Evolution of methane was
immediately observed. The reaction mixture was stirred for 24 h
at room temperature. All of the volatiles were removed under
reduced pressure to give a foam-like white solid, which was
dissolved in n-hexane and filtered to remove trace impurities.
The filtrate was concentrated and kept at -20˚ C overnight.
Subsequently, 2a was collected as a white precipitate and dried
3.48 mmol). Yield: 1.16 g, 79.9%. H NMR (400 MHz, CDCl3):
δ 7.48–7.36 (m, 4H, ArH), 7.35–7.29 (m, 1H, ArH), 7.25–7.19
(m, 1H, ArH), 7.13–7.04 (m, 3H, ArH), 6.59–6.52 (m, 3H, ArH),
2
2
4.57 (d, J = 14.0 Hz, 1H, ArCH2), 4.43 (d, J = 12.8 Hz, 1H,
ArCH2), 4.20 (d, 2J = 14.0 Hz, 1H, ArCH2), 3.81 (d, 2J = 12.8 Hz,
1H, ArCH2), 2.28 (s, 3H, ArCH3), 1.44 (s, 9H, C(CH3)3), -0.43
(AlCH3), -1.26 (AlCH3). 13C{H} NMR (100 MHz, CDCl3): δ
157.0, 145.8, 138.9, 131.5, 130.8, 129.9, 129.8, 129.0, 128.5,
128.0, 127.0, 125.0, 122.4, 119.5 (all Ar–C), 58.5 (ArCH2), 53.3
(PhCH2), 35.0 (C(CH3)3), 29.7 (C(CH3)3), 21.0 (ArCH3), -9.5
(AlCH3), -10.4 (AlCH3). Anal. Calcd. for C27H34AlNO (%): C,
78.04; H, 8.23; N, 3.37. Found: C, 78.50; H, 8.71; N, 3.38.
1
under reduced pressure. Yield: 2.32 g, 73.3%. H NMR (400
MHz, CDCl3): δ 7.44–7.37 (m, 3H, ArH), 7.33–7.27 (m, 2H,
4
4
ArH), 7.04 (d, 1H, J = 2.0 Hz, ArH), 6.56 (d, J = 2.0 Hz, 1H,
2.3.5. Synthesis of Lp-TolAlMe2 complex (2f)
2
2
ArH), 3.97 (d, J = 13.2 Hz, 1H, ArCH2), 3.92 (d, J = 13.2 Hz,
2
2
1H, ArCH2), 3.89 (d, J = 13.2 Hz, 1H, ArCH2), 3.55 (d, J =
13.2 Hz, 1H, ArCH2), 2.24 (s, 3H, NCH3), 2.22 (s, 3H, ArCH3),
1.39 (s, 9H, C(CH3)3), -0.62 (s, 3H, AlCH3), -0.86 (s, 3H,
AlCH3). 13C{H} NMR (100 MHz, CDCl3): δ 156.7, 138.9, 132.4,
129.8, 129.4, 128.8, 128.5, 128.2, 125.1, 120.4 (all Ar–C), 59.4
(ArCH2), 59.0 (PhCH2), 40.2 (NCH3), 35.0 (C(CH3)3), 29.7
(C(CH3)3), 20.9 (ArCH3), –10.3 (AlCH3), –10.9 (AlCH3). Anal.
Calcd. for C22H32AlNO (%): C, 74.75; H, 9.12; N, 3.96. Found:
C, 74.80; H, 9.54; N, 4.00.
Complex 2f was obtained as a white powder, by following a
similar procedure to that described for 2a, from reaction between
AlMe3 (0.34 mL, 0.67 mmol, 2.0 M in hexane) and 1f (0.24 g,
1
0.67 mmol). Yield: 0.21 g, 73.0%. H NMR (400 MHz, CDCl3):
δ 7.30–7.18 (m, 5H, ArH), 7.12–7.04 (m, 3H, ArH), 6.58–6.52
2
2
(m, 3H, ArH), 4.52 (d, J = 14.4 Hz, 1H, ArCH2), 4.39 (d, J =
13.2 Hz, 1H, ArCH2), 4.16 (d, 2J = 14.4 Hz, 1H, ArCH2), 3.76 (d,
2J = 13.2 Hz, 1H, ArCH2), 2.38 (s, 3H, NArCH3), 2.27 (s, 3H,
ArCH3) 1.43 (s, 9H, C(CH3)3), -0.46 (AlCH3), -1.26 (AlCH3).
13C{H} NMR (100 MHz, CDCl3): δ 157.0, 143.1, 138.8, 136.7,
131.6, 130.8, 130.4, 129.9, 128.9, 128.4, 127.9, 125.0, 122.2,
119.5, (all Ar–C), 58.4 (ArCH2), 53.2 (PhCH2), 35.0 (C(CH3)3),
2.3.2. Synthesis of Li-prAlMe2 complex (2c)

29.7 (C(CH3)3), 21.1 (NArCH3), 21.0 (ArCH3), -9.5 (AlCH3), -
10.3 (AlCH3). Anal. Calcd. for C28H36AlNO (%): C, 78.29; H,
8.45; N, 3.26. Found: C, 77.82; H, 8.94; N, 3.17.
2.3.6. Synthesis of L4-ClArAlMe2 complex (2g)
Complex 2g was obtained as a white powder, by following a
similar procedure to that described for 2a, from reaction between
AlMe3 (0.82 mL, 1.63 mmol, 2.0 M in hexane) and 1g (0.65 g,
1
3. Results and Discussion
1.63 mmol). Yield: 0.63 g, 86.2%. H NMR (CDCl3, 400 MHz):
δ 7.46–7.38 (m, 1H, ArH), 7.36–7.29 (m, 2H, ArH), 7.27–7.18
3.1. Synthesis and characterization of proligands and complexes
(m, 2H, ArH), 7.16–7.07 (m, 3H, ArH), 6.61–6.53 (m, 3H, ArH),
2
2
4.51 (d, J = 14.4 Hz, 1H, ArCH2), 4.39 (d, J = 12.8 Hz, 1H,
ArCH2), 4.21 (d, 2J = 14.4 Hz, 1H, ArCH2), 3.79 (d, 2J = 12.8 Hz,
1H, ArCH2), 2.28 (s, 3H, ArCH3), 1.43 (s, 9H, C(CH3)3), -0.44
(AlCH3), -1.23 (AlCH3). 13C{H} NMR (CDCl3, 100 MHz): δ
156.9, 144.5, 139.1, 132.8, 131.5, 130.5, 129.9, 129.8, 129.2,
128.7, 128.2, 125.3, 123.9, 119.2 (all Ar–C), 58.6 (ArCH2), 53.7
(ArCH2), 35.0 (C(CH3)3), 29.8 (C(CH3)3), 21.0 (ArCH3), -9.5
(AlCH3), -10.1 (AlCH3). Anal. Calcd. for C27H33AlClNO (%): C,
72.07; H, 7.39; N, 3.11. Found: C, 71.51; H, 7.25; N, 3.08
The new (aminomethyl)phenol derivatives 1c-g (Scheme 1)
were synthesized in good yield by modification of the method
reported by Kim and Ishida [28], via neat reaction of 2-tert-butyl-
4-methylphenol with paraformaldehyde and appropriate amine at
105 °C. In addition, compounds 1a [26] and 1b [27] were
prepared by literature methods. LRAlMe2 complexes 2a-g
(Scheme 1) were obtained in good yield via modification of the
method reported by Wang and Ma [27] for preparation of
LEtAlMe2 (2b), by treatment of proligands 1a-g with one
equivalent of AlMe3 in toluene at room temperature for 24 h. The
reaction proceeded cleanly with evolution of methane to produce
2a-g which were isolated as moisture-sensitive light-yellow or
white powders. The compounds are readily soluble in nonpolar
and polar aprotic hydrocarbon solvents, such as chloroform,
methylene chloride, diethyl ether, and THF, as well as aromatic
hydrocarbon solvents such as benzene and toluene. However, the
compounds are only moderately soluble in aliphatic hydrocarbon
solvents, and thus could be recrystallized from hexane at low
temperatures.
2.4. Crystallographic Studies
Single crystals of LRAlMe2 complexes 2a-2c (R = Me, Et, i-
Pr), 2e (R = Ph), and 2f (R = p-tolyl) suitable for X-ray
crystallographic analysis were obtained by slow recrystallization
from a 1:1 n-hexane:toluene solution of the complex in the
glovebox at room temperature. Colorless single crystals of each
complex were placed in dry and degassed paratone oil on a glass
plate and used for X-ray diffraction analysis. Crystallograhic data
for the complexes are collected in Table 1. Further details of the
crystallographic study are given in the supplementary material.
2.5. General procedure for catalytic dehydration of glucose
All the reactions were performed in a 5-mL reaction vial
sealed with a solid cap with PTFE faced silicone septum. In a
typical experiment, D–(+)–glucose (50 mg, 0.28 mmol),
[EMIM]Cl (500 mg, 3.41 mmol), and a specified amount of
aluminum precatalyst were charged into the reaction vial along
with a magnetic stir bar under nitrogen atmosphere. The mixture
was placed in a preheated oil bath at the desired temperature and
let stir for a specified period of time. The reaction was quenched
by immediately placing the vial in an ice bath. The mixture was
diluted with 3 mL of deionized water, let stir for 5 minutes and
centrifuged for 1 h. Subsequently, the supernatant was collected
(via decantation to exclude insoluble solids) and analyzed by
HPLC.
Scheme 1. Synthesis of proligands 1c-g and LRAlMe2 complexes 2a-g.
The formulation and molecular structure of LRAlMe2
complexes 2a and 2c-g were established by microanalysis and
1
solution NMR data. Their H NMR spectra did not show the
downfield resonance characteristic of the phenolic OH group of
the proligands, supporting coordination of phenolate oxygen with
aluminum. Consistent with bidentate coordination of the
(aminomethyl)phenolate ligand, with tight binding of the amino
nitrogen to aluminum resulting in hindered rotation of N-benzyl
group on the NMR timescale at room temperature, the 1H NMR
spectra of LRAlMe2 complexes 2a and 2c-g contained four
doublet resonances in the 4.57-3.55 ppm range for the four
benzylic protons. Similarly, two chemically inequivalent methyl
resonances were observed for the N-isopropyl group of 2c,
consistent with coordination of amino nitrogen to aluminum and
hindered rotation about the N–CH(CH3)2 bond. In contrast, the
1H NMR spectrum for the proligand 1c contained a single
resonance for chemically equivalent methyl groups of the N-
isopropyl unit. Furthermore, consistent with the C1 symmetry
expected for tetrahedral LRAlMe2 complexes 2a and 2c-g, two
2.6. Product Analysis
Quantitative analysis of the products was performed by HPLC
using a Thermo Scientific Dionex Ultimate 3000 HPLC system
equipped with a Dionex quaternary pump, a Shodex RI-101
refractive index detector, and a Biorad Aminex HPX-87H
column (300 7.8 mm). 5.0 mM H2SO4 was used as the mobile
phase at a flow rate of 0.6 mL/min, and the column temperature
was maintained at 50 °C. The injection volume was 20 µL. All
concentrations of glucose, fructose, and HMF in the aqueous
phase were determined by comparison to standard calibration
curves.
Glucose conversion and products selectivity are defined as
follows:
1
different Al–CH3 resonances were observed in their H NMR
spectra in the upfield region of -0.56 to -0.86 ppm for complexes
2a-2d (with N-alkyl substituent), and -0.43 to -1.26 ppm for 2e-g
(with N-aryl substituent). 13C{1H}

Table 1. Crystallographic Data for LRAlMe2 complexes 2a-c, 2e, and 2f

complex
2a
2b
2c
2e
2f
formula
C22H32AlNO
353.46
C23H34AlNO
367.49
C24H36AlNO
381.52
C27H34AlNO
415.53
C28H36AlNO
429.56
formula wt.
T (K)
90.0(2)
90.0(2)
90.0(2)
90.0(2)
90.0(2)
crystal system
Monoclinic
Monoclinic
Monoclinic
Monoclinic
Monoclinic
space group
a (Å)
P2(1)/n
P2(1)/n
P2(1)/n
P2(1)/c
P2(1)/c
9.6557(2)
12.6012(2)
17.5369(3)
90
11.1393(4)
9.6143(3)
20.1941(8)
90
11.5303(4)
18.1337(7)
12.0981(2)
17.9578(4)
12.0051(2)
90
9.2418(3)
21.7215(6)
12.4113(3)
90
b (Å)
12.0678(5)
c (Å)
α (deg)
β (deg)
γ (deg)
V (Å3)
90
99.0269(8)
90
91.854(3)
90
117.291(2)
116.0401(12)
90
98.493(1)
90
90
2242.36(15)
2107.35(7)
2161.59(13)
2343.41(8)
2464.19(12)
Z
4
4
4
4
4
Dcalcd (g/cm3)
Final R indices [I > 2σ(I)]
R1, wR2
1.114
1.129
1.130
1.178
1.158
0.0445, 0.1128
0.0548, 0.1458
0.0336, 0.0857
0.0430, 0.1053
0.0344, 0.0911
R indices (all data)
R1, wR2
0.0595, 0.1251
1489633
0.0700, 0.1552
1489630
0.0365, 0.0883
1812487
0.0654, 0.1162
1489632
0.0350, 0.0916
1489631
CCDC no.
Table 2. Selected Bond Lengths (Å) and Bond Angles (˚) for LRAlMe2 complexes 2a-2c, 2e, and 2f
2b 2c 2e
2a
2f
Al(1)-O(1)
Al(1)-C(22)
Al(1)-C(21)
Al(1)-N(1)
1.7559(10)
1.9470(16)
1.9542(16)
2.0284(12)
111.53(7)
111.66(6)
117.33(8)
96.53(5)
Al(1)-O(1)
Al(1)-C(22)
Al(1)-C(23)
Al(1)-N(1)
1.7474(16)
Al(1)-O(1)
Al(1)-C(23)
Al(1)-C(24)
Al(1)-N(1)
1.7589(8)
1.9619(11)
1.9647(11)
2.0727(9)
110.95(4)
108.86(4)
113.74(5)
97.36(3)
Al(1)-O(1)
Al(1)-C(27)
Al(1)-C(26)
Al(1)-N(1)
1.7571(11)
1.9560(16)
1.9589(16)
2.0802(13)
111.47(6)
110.67(6)
116.50(7)
97.09(5)
Al(1)-O(1)
Al(1)-C(28)
Al(1)-C(27)
Al(1)-N(1)
1.7616(9)
1.9570(14)
1.9638(13)
2.0532(10)
110.42(5)
113.27(5)
118.92(6)
94.95(4)
1.965(3)
1.966(3)
2.027(2)
O(1)-Al(1)-
C(22)
O(1)-Al(1)-
C(22)
109.99(10)
110.45(10)
119.21(12)
97.17(8)
O(1)-Al(1)-
C(23)
O(1)-Al(1)-
C(27)
O(1)-Al(1)-
C(28)
O(1)-Al(1)-
C(21)
O(1)-Al(1)-
C(23)
O(1)-Al(1)-
C(24)
O(1)-Al(1)-
C(26)
O(1)-Al(1)-
C(27)
C(22)-Al(1)-
C(21)
C(22)-Al(1)-
C(23)
C(23)-Al(1)-
C(24)
C(27)-Al(1)-
C(26)
C(28)-Al(1)-
C(27)
O(1)-Al(1)-
N(1)
O(1)-Al(1)-
N(1)
O(1)-Al(1)-
N(1)
O(1)-Al(1)-
N(1)
O(1)-Al(1)-
N(1)
C(22)-Al(1)-
N(1)
108.36(6)
109.35(6)
C(22)-Al(1)-
N(1)
110.44(10)
107.31(10)
C(23)-Al(1)-
N(1)
109.30(4)
115.45(4)
C(27)-Al(1)-
N(1)
110.12(6)
109.26(6)
C(28)-Al(1)-
N(1)
110.52(5)
105.97(5)
C(21)-Al(1)-
N(1)
C(23)-Al(1)-
N(1)
C(24)-Al(1)-
N(1)
C(26)-Al(1)-
N(1)
C(27)-Al(1)-
N(1)

C(20)-N(1)-
Al(1)
109.58(8)
C(20)-N(1)-
Al(1)
109.88(15)
C(20)-N(1)-
Al(1)
111.59(7)
C(20)-N(1)-
Al(1)
106.78(8)
C(20)-N(1)-
Al(1)
109.24(7)
(b)
(a)
)
(c)
)
Figure 1. ORTEP diagrams of (a) 2a, (b) 2b and (c) 2c. Thermal ellipsoids are drawn at 50% probability level. Hydrogens are omitted for clarity.

Figure 2. ORTEP diagrams of 2e, (left) and 2b (right). Thermal ellipsoids are drawn at 50% probability level. Hydrogens are omitted for clarity.
NMR spectra of the complexes are also consistent with their C1 symmetry; together with two Al–CH3 and two benzylic
carbon resonances, 2a, 2c, and 2d each displayed ten aromatic carbon resonances while 2e-g each displayed fourteen
aromatic carbon resonances.
X-ray diffraction analysis on single-crystals of 2a-2c, 2e, and 2f confirmed the structure assigned by spectroscopy.
Structures of the complexes are depicted in Figures 1 and 2, and crystallographic data and selected metrical parameters for
the complexes are collected in Tables 1 and 2. The compounds adopt a distorted tetrahedral structure with the
(aminomethyl)phenolate ligand coordinated to aluminum in bidentate fashion, via phenolate oxygen and amino nitrogen
atoms. The aluminum center is also coordinated by two carbon atoms from two methyl groups. The distortion from
idealized tetrahedral geometry arises from the acute bite angle of the chelating (aminomethyl)phenolate ligand [O(1)-
Al(1)-N(1) bond angles range from ca. 95 to 97°], which is compensated for by opening of the C–Al–C, C–Al–O, and C–
Al–N bond angles (Table 2). All of the Al–O, Al–N and Al–C bond distances are within the range reported for related
complexes [29-31]. However, 2a and 2b (with NMe(CH2Ph) or NEt(CH2Ph) moiety, respectively) possessed shorter Al–N
bond distances (<2.03 Å) than were observed (>2.05 Å, Table 2) for
2c (with N(i-Pr)(CH2Ph) moiety), 2e (with NPh(CH2Ph) moiety) or
2f (with N(p-MeC6H4)(CH2Ph) moiety). Presumably, this is because
electron-releasing methyl and ethyl substituents increase electron
donation by amino nitrogen atom to aluminum, relative to the bulkier

isopropyl substituent or less electron donating aryl substituents. The molecular structures (Figures 1 and 2) confirmed that
1
in 2e and 2f the two Al–CH3 groups reside in a more dissimilar chemical environment than in 2a-c, consistent with H
NMR data (vide supra). In 2e and 2f, one Al–Me group lies in close proximity to the N-aryl ring; for 2e, the torsion angle
between Al-Me and N-phenyl ring (C27-Al(1)-N1-C20) is 27.52˚ and the C20–C27 distance is 3.297Å.
3.2. Glucose Dehydration Studies
3.2.1. Effect of glucose loading
Figure 3 shows the effect of the weight percent (wt%) of glucose in 1-ethyl-3-methylimidazolium chloride ([EMIM]Cl) on
the conversion and the product distribution of glucose dehydration at 120 ˚C for 2 hours using 5 mol% (relative to moles
of glucose) LPhAlMe2 (2e) as catalyst. The glucose conversion ranged between
Figure 3. The effect of glucose weight percent in [EMIM]Cl on
the products distribution. Reaction conditions: 50 mg glucose
using 5 mol% [LPhAlMe2] (2e) at 120˚ C for 2 h.
73 to 80% for glucose concentrations in [EMIM]Cl
ranging between 9.1 to 28.6 wt%. However, the
highest HMF selectivity and yield (58% and 42%,
respectively) were both obtained when 9.1 wt%
glucose was employed. It is known that Lewis acid-
catalyzed glucose dehydration generally proceeds via
glucose isomerization to fructose, followed by
fructose dehydration to HMF (Scheme 2) [6].
Predictably, all of the product mixtures also contained
a small amount of fructose (2-3%) except for when
9.1 wt% glucose in [EMIM]Cl was employed,
whereupon fructose was present only in trace amount.
No other soluble products were detected by HPLC
analysis in the supernatants obtained after aqueous
extraction of any of the dark brown reaction mixtures; these results and all other results reported herein were reproduced at
least 3 times. Since glucose concentrations ≥9.1 wt% resulted in comparable conversions while both the HMF selectivity
and HMF yield decreased when >9.1 wt% glucose in [EMIM]Cl was employed, all other experiments reported herein were
conducted using 9.1 wt% sugar in ionic liquid solvent, unless otherwise indicated.
3.2.2. Effect of temperature and time
Table 3 shows the effects of temperature and time on glucose dehydration in [EMIM]Cl using [LPhAlMe2](2e, 5 mol %) as
catalyst. The reaction was investigated in the absence and presence of catalyst over the 100 – 140 ˚C temperature range. At
all temperatures in the absence of a catalyst, both the glucose conversion (3-23%) and the HMF yield (<1%) were quite
poor, consistent with previous literature reports.[11, 32] For example, Zhao et al. reported 40% glucose conversion and
<4% HMF yield when 9.1 wt% glucose in [EMIM]Cl was heated at 180 ˚C for 3 hours in the absence of a catalyst.[11] In
the presence of [LPhAlMe2] (2e), the conversion of glucose at 100 ˚C increased gradually with time, reaching a maximum
of 59% after 6 hours (Table 3, entries 3-6). The HMF selectivity increased up to 56% over four hours of reaction, and
remained at 56% after 6 hours, resulting in 33% HMF
yield. As expected, glucose conversion increased with
an increase in temperature. Consequently, 95%
Scheme 2. Possible pathways for glucose conversion to HMF and
other products.
glucose conversion was achieved at 120 ˚C after 6
hours. However, as the data in Table 3 (entries 11-14) show, glucose conversion slowed dramatically as the reaction
progressed, with only a small increase in glucose conversion observed after 4 hours. While this likely reflects the
reduction in reaction rate as the concentration of glucose decreases, it is noteworthy that the HMF selectivity remained
more or less constant (53-54%) over 4 hours, and decreased only slightly (to 49%) after 6 hours. This result argues against
significant catalyst deactivation occurring during the reaction since the HMF selectivity remained essentially constant as
the glucose conversion increased. Accordingly, the HMF yield increased up to 46% after 4 hours and was essentially
unchanged after 6 hours.

Table 3. Temperature and time effects on glucose dehydration in [EMIM]Cl in absence and presence of a catalyst.a
a Reaction conditions: 9.1 wt% glucose in [EMIM]Cl. b 5 mol% [LPhAlMe2] (2e)
Entry
Temp
Time
(min)
Cat.b
Glucose
Conv.
(%)
3
HMF
HMF
Yield
(%)
0
used as catalyst. c 20 min. d 40 min.
(˚C)
Selectivity
(%)
0
Raising the reaction temperature to 140 ˚C resulted in 92%
glucose conversion after 1 hour, along with 48% HMF selectivity
and 44% HMF yield. Consequently, we investigated the effect of
shorter reaction time for [LPhAlMe2] (2e)-catalyzed glucose
conversion at 140 ˚C (Table 3, entries 16 and 17). 70% glucose
conversion was observed after 20 minutes but the reaction progress
slowed dramatically once again, with only 15% additional glucose
conversion observed after another 20 minutes of reaction. However,
while glucose conversion increased on raising the reaction
temperature from 120 to 140 ˚C, the HMF selectivity and hence the
HMF yield decreased slightly although shorter time was required to
reach high conversion (Table 3).
1
100
100
100
100
100
100
120
120
120
120
120
120
120
120
140
140
140
140
1
-
2
2
-
4
0
0
3
1
2e
2e
2e
2e
-
17
31
46
59
4
38
43
56
56
0
6
4
2
13
26
33
0
5
4
6
6
7
1
8
2
-
6
0
0
9
4
-
8
5
< 1
<1
28
38
46
47
<1
33
39
44
10
11
12
13
14
15
16
17
18
6
-
15
52
69
88
95
23
70
85
92
5
3.2.3. Effect of catalyst loading
1
2e
2e
2e
2e
-
53
54
53
49
<1
47
45
48
Table 4 shows results of our study of the effect of catalyst
loading on glucose conversion and the product distribution for
[LPhAlMe2] (2e)-catalyzed dehydration of glucose in [EMIM]Cl at
both 100 and 120 ˚C for 4 hours; the catalyst loading was varied in
5% increments from 5 to 20 mol%. At both temperatures, a modest
increase in glucose conversion accompanied an increase in the
catalyst loading from 5 to 10 mol% while further increase in the
catalyst loading had little effect on the extent of reaction.
Conversely, both the HMF selectivity and yield decreased
significantly upon increasing the catalyst loading from 5 to 10 mol%
2
4
6
1
0.33c
0.66d
1
2e
2e
2e
while further increase in the catalyst loading resulted in unchanged or slightly decreased HMF selectivity and yield. Thus,
it appears that catalyst loadings higher than 5 mol% enhance side reactions that lead to formation of humins (vide infra).
Table 4. The effect of catalyst ([LPhAlMe2], 2e) loading on glucose dehydration in [EMIM]Cl.a
Entry
Temp
Catalyst
mol%
Glucose
Conv.
(%)
46
HMF
HMF
Yield
(%)
26
(˚C)
Selectivity
(%)
56
42
43
39
53
45
45
42
1
2
3
4
5
6
7
8
100
100
100
100
120
120
120
120
5
10
15
20
5
65
27
69
30
71
28
88
46
10
15
20
94
42
95
43
97
40
a Reaction conditions: 9.1 wt% glucose at the indicated temperature for 4 h.
3.2.4. Ligand effects
The potential of LRAlMe2 complexes 2a-g as catalysts for glucose dehydration to HMF was investigated by conducting
the reaction in [EMIM]Cl at 120 ˚C for 4 hours using 5 mol% of 2a-g as catalyst. As shown by the glucose conversion and
product distribution data in Figure 4, LRAlMe2 complexes 2a-d for which the R substituent was an alkyl group (Scheme 1)
were ineffective

Figure 4. Glucose conversion in [EMIM]Cl with [LRAlMe2]
catalysts 2a-2g. Reaction conditions: 9.1 wt% glucose using 5
mol% [LRAlMe2] as catalyst at 120˚ C for 4 h.
catalysts for selective formation of HMF. The glucose
conversion was modest (~50%), even if significantly
higher than in absence of a catalyst (Table 3, entry 9).
But more importantly, both the HMF selectivity and
yield were extremely poor. The HMF yield in fact
decreased as size of the amino moiety’s alkyl
substituent increased, with only a trace amount of
HMF produced when 2d (LRAlMe2, R = But) was the
catalyst.
The difference in catalytic efficiency of LRAlMe2
complexes containing alkyl-substituted amino group
(2a-d) versus aryl-substituted amino group (2e-g) is
remarkable. All of the aryl-substituted aluminum
(aminomethyl)phenolate complexes 2e-g afforded much higher glucose conversion (>87%) and much better HMF
selectivity (49-54%) and yield (42-49%) than alkyl-substituted aluminum (aminomethyl)phenolate complexes 2a-d
(Figure 4). As the data in Table 2 show, bond angles about the Al and N atoms are similar for all of the complexes.
However, Al–N bond distances for 2a and 2b are significantly shorter than those for 2e and 2f, due presumably to stronger
sigma electron donation to aluminum by alkyl-substituted nitrogen relative to aryl-substituted nitrogen. On the other hand,
the significantly longer Al–N bond distance for 2c (compared to 2a and 2b) is most probably due to its sterically more
crowded coordination sphere. Thus, we presume that the markedly decreased efficiency of 2a-d as glucose dehydration
catalysts (versus 2e-g) is due to the reduced Lewis acidity of 2a-d, and/or greater steric hindrance at the aluminum center
in complexes 2c and 2d. In this regard, a slight increase in both the HMF selectivity and yield was observed as electron
donation from aryl-substituted amino group was decreased by decreasing the electron releasing ability of the para
substituent of the aryl group (Figure 4), that is, from R = p-MeC6H4 (2e) to R = C6H5 (2d) to R = p-ClC6H4 (2f) [33].
Clearly, the different (aminomethyl)phenolate ligands impose different chemical (coordination) environments about the Al
center, consistent with the different chemical shifts observed for the Al–Me groups of 2a-d versus 2e-g (see Section 3.1).
3.2.5. Effect of ionic liquid
Table 5. Ionic liquid effects on glucose dehydration with [LPhAlMe2] (2e) as catalyst.
Entry
Ionic
Time (h)
Glucose
Conv.
(%)
56
HMF
HMF
Yield
(%)
32
liquid (IL)
Selectivity
(%)
56
59
64
55
56
60
55
63
1a
2a
3a
4a
5a
6a
7a
8b
[EMIM]Br
[EMIM]Br
[EMIM]Br
[EMIM]Br
[BMIM]Br
[BMIM]Br
[BMIM]Br
[BMIM]Br
1
2
4
6
1
2
4
2
74
44
84
54
91
50
92
52
97
58
100
95
55
60

a Reaction conditions: 9.1 wt% glucose with 5 mol% [LPhAlMe2] (2e) at 120˚ C. b 4.8 wt% glucose and 5 mol% [LPhAlMe2] (2e) at 120˚ C.
Table 5 shows the effect of ionic liquid on efficiency of glucose dehydration at 120 ˚C using 2e (5 mol%) as catalyst.
While the reaction progressed similarly in [EMIM]Br (Table 5,
entries 1-4) as in [EMIM]Cl (Table 3, entries 11-14), the HMF
selectivity and yield were significantly higher in [EMIM]Br, peaking
after 4 hours at 64% and 54%, respectively. Higher HMF selectivity
and yield have previously being observed in the presence of bromide
ion relative to chloride ion, and have been attributed to acceleration
of fructose dehydration as a result of better nucleophilicity and
leaving group properties of bromide ion [34, 35]. As mentioned
earlier, fructose is the only other soluble product observed in our
reactions, and Lewis acid-catalyzed glucose dehydration generally
occurs via glucose isomerization to fructose [6].
Glucose conversion progressed significantly faster in [BMIM]Br
(1-butyl-3-methylimidazolium bromide) than in [EMIM]Br, reaching
92% in 1 hour, and giving HMF selectivity and yield of 56% and
52%, respectively. Increasing the reaction time to 2 hours resulted in
slightly higher glucose conversion and an increase in the HMF
selectivity and yield to 60% and 58%, respectively. However, further
increase in the reaction time resulted in a decrease in the HMF selectivity and yield (Table 5, entry 7). Since glucose
conversion was much faster in [BMIM]Br, we investigated the effect of lowering the concentration of glucose in
[BMIM]Br from 9.1 wt% to 4.8 wt% on the HMF selectivity and yield: 95% glucose conversion was achieved after 2
hours, along with slight increases in the HMF selectivity and yield, up to 63% and 60%, respectively (Table 5, entry 8).
3.3. Humins Analysis
Sugar dehydration is routinely accompanied by formation of humins, which studies have indicated may be formed by
condensation reactions between sugars, HMF, and intermediates formed during dehydration of sugars [9, 36, 37]. In this
regard, Lund et al.[38, 39] have proposed a mechanism for humin formation in which 2,5-dioxo-6-hydroxyhexanal
(DHH), formed by HMF rehydration, is a key intermediate (Scheme 2). Humins were proposed to be formed via
subsequent aldol condensations of DHH with the carbonyl group of HMF, with the extent of HMF incorporation in the
humin structure being dependent on the accumulation of HMF during the reaction. Furthermore, it was suggested that
humins could not be directly formed from sugars. Zandvoort et al.[40] have similarly suggested that humins are mainly
derived from HMF based on their finding that addition of HMF to the glucose feed barely changed the elemental
composition of the humin obtained from acid-catalyzed dehydration of glucose. HPLC analysis of the product mixtures
from glucose and fructose dehydration catalyzed with aluminum (aminomethyl)phenolate complexes 2a-g detected HMF
as well as glucose and/or fructose as the only products. Thus, formation of humins rather than HMF rehydration to form
levulinic acid (LA) and formic acid (FA) appears to be the main route for HMF loss in these reactions (Scheme 2).
The nature of the insoluble brown solids produced during LPhAlMe2 (2e)-catalyzed dehydration of glucose (for 4 hours)
in [EMIM]Cl at 120 ˚C was investigated by Raman and ATR-FTIR spectroscopy. The Raman data are suggestive of the
presence of aromatic groups with oxygen-rich functionalities. The signals at 1385 and 1585 cm-1 are characteristic of the D
and G bands of disordered graphite-like carbon [38, 40, 41]. Figure 5 compares ATR-FTIR spectra of the humins with the
IR spectrum of HMF. The humins show broad absorbance peaks in ca. 1100-1400 cm-1 range, along with peaks that arise
from the furan ring of HMF [38, 42]. Specifically, the two peaks in the 750- 850 cm-1 range, the
Figure 5. ATR-IR spectra of (a) HMF and (b) humins formed during LPhAlMe2 (2e)-catalyzed glucose dehydration in [EMIM]Cl at 120 ˚C for 4
hours.
peak at 1020 cm-1, and the peak at 1512 cm-1 have been attributed the furan ring of HMF. These data strongly support
significant incorporation of HMF into the humin structure.

4. Conclusions
LRAlMe2 complexes 2e-g, which contain a bidentate (aminomethyl)phenolate ligand with an aryl substituent on the
amino group, are efficient catalysts for glucose dehydration in ionic liquid solvents to give HMF. In [EMIM]Br and
[BMIM]Br, the reaction proceeds at 120 ˚C with very high conversion in 2 hours to produce HMF with 60-63% selectivity
and in 58-60% yield. Both the HMF selectivity and yield were lower in [EMIM]Cl, up to 54% and 49%, respectively. The
HMF selectivity of glucose dehydration decreased as the concentration of the aluminum catalyst was increased from 5 to
20 mol%. Giving that no other soluble products (besides glucose, fructose, and/or HMF) were detected by HPLC analysis
of the supernatants obtained after aqueous workup of the reaction mixtures, and that Raman and ATR-FTIR studies of the
humins produced during glucose dehydration established significant incorporation of HMF in their structure, the HMF
selectivity of LRAlMe2-catalyzed glucose dehydration appears to be limited by competing loss of HMF to humins
formation.
The reasonably high yield of HMF (60%) obtained herein from LPhAlMe2-catalyzed glucose dehydration in ionic
liquids is encouraging, as is our finding that the catalytic efficiency of LRAlMe2 complexes can be tuned via modification
of the (aminomethyl)phenolate ligand. To date, the vast majority of studies of aluminum-catalyzed glucose conversion to
HMF have focused on AlCl3. The findings from this study are useful toward developing better understanding of the
relationship between the structure and function of aluminum catalysts for glucose (and ultimately cellulose) conversion
into HMF. Towards this end, we have recently initiated a study of the reactions of LMeAlMe2 (2a) and LPhAlMe2 (2e) with
glucose and cycloalkane diols in ionic liquid solvents.
Acknowledgements
We are grateful to the Kentucky Science Foundation (grant number KSEF-2428-RDE-014) for partial financial support
of this work. We are also grateful to US National Science Foundation (grant number CBET 1604491) for partial financial
support of this research. NMR instruments utilized in this research were funded in part by the CRIF program of the US
National Science Foundation (grant number CHE-9974810). The X8 Proteum and D8 Venture diffractometers were
funded by the NSF (MRI CHE0319176 and CHE1625732 respectively). The authors are also grateful to Professor Darrin
L. Smith (Chemistry Department, Eastern Kentucky University) for generously providing access and helping with HPLC
analysis. DS is grateful to Professor Anne-Frances Miller and Mr. John Layton for helpful advice with regard to NMR.
Appendix A. Supplementary Data
CCDC 1489633, 1489630, 1812487, 1489632, and 1489631 contains the supplementary crystallographic data for
complexes 2a-c, 2e and 2f, respectively. These data are available free of charge
via
http://www.ccdc.cam.ac.uk/conts/retrieving.html, or from the Cambridge Crystallographic Data Centre, 12 Union Road,
Cambridge CB2 1EZ, UK; fax: (+44) 1223-336-033; or email: deposit@ccdc.cam.ac.uk. Additional supplementary data
associated with this article can be found online.
5. References
[1] A. Chinnappan, C. Baskar, H. Kim, RSC Adv., 6 (2016) 63991.
[2] C.B. Rasrendra, I.G.B.N. Makertihartha, S. Adisasmito, H.J. Heeres, Top. Catal., 53 (2010) 1241.
[3] B.R. Caes, R.E. Teixeira, K.G. Knapp, R.T. Raines, ACS Sustainable Chem. Eng., 3 (2015) 2591.
[4] H. Li, F. Chang, Y. Zhang, D. Hu, L. Jin, B. Song, S. Yang, Curr. Catal., 1 (2012) 221.
[5] R.-J. van Putten, J.C. van der Waal, E. de Jong, C.B. Rasrendra, H.J. Heeres, J.G. de Vries, Chem. Rev. (Washington, DC, U.
S.), 113 (2013) 1499.
[6] A. Corma, S. Iborra, A. Velty, Chem. Rev. (Washington, DC, U. S.), 107 (2007) 2411.
[7] M. Bicker, J. Hirth, H. Vogel, Green Chem., 5 (2003) 280-284.
[8] G.W. Huber, S. Iborra, A. Corma, Chem. Rev. (Washington, DC, U. S.), 106 (2006) 4044.
[9] J.-P. Lange, R. Price, P.M. Ayoub, J. Louis, L. Petrus, L. Clarke, H. Gosselink, Angew. Chem., Int. Ed., 49 (2010) 4479.
[10] Y.J. Pagan-Torres, T. Wang, J.M.R. Gallo, B.H. Shanks, J.A. Dumesic, ACS Catal., 2 (2012) 930.
[11] H. Zhao, J.E. Holladay, H. Brown, Z.C. Zhang, Science (Washington, DC, U. S.), 316 (2007) 1597.
[12] Y. Zhang, E.A. Pidko, E.J.M. Hensen, Chem. - Eur. J., 17 (2011) 5281.
[13] J.B. Binder, R.T. Raines, J. Am. Chem. Soc., 131 (2009) 1979.
[14] S. Hu, Z. Zhang, J. Song, Y. Zhou, B. Han, Green Chem., 11 (2009) 1746.
[15] K.R. Enslow, A.T. Bell, Catal. Sci. Technol., 5 (2015) 2839.
[16] Y. Yang, C. Hu, M.M. Abu-Omar, Bioresour. Technol., 116 (2012) 190.
[17] Y. Yang, C. Hu, M.M. Abu-Omar, J. Mol. Catal. A Chem., 376 (2013) 98.

[18] X. Zhang, P. Murria, Y. Jiang, W. Xiao, H.I. Kenttamaa, M.M. Abu-Omar, N.S. Mosier, Green Chem., 18 (2016) 5219.
[19] C. Rasrendra, J. Soetedjo, I. Makertihartha, S. Adisasmito, H. Heeres, Topics in Catalysis, 55 (2012) 543.
[20] D. Liu, E.Y.X. Chen, Appl. Catal., A, 435-436 (2012) 78.
[21] C.B. Rasrendra, J.N.M. Soetedjo, I.G.B.N. Makertihartha, S. Adisasmito, H.J. Heeres, Top. Catal., 55 (2012) 543.
[22] W.L.F. Armarego, D.D. Perrin, Purification of Laboratory Chemicals, Fourth Edition, 1997.
[23] P. Nockemann, K. Binnemans, K. Driesen, Chem. Phys. Lett., 415 (2005) 131.
[24] P.N. Tshibangu, S.N. Ndwandwe, E.D. Dikio, Int. J. Electrochem. Sci., 6 (2011) 2201.
[25] S. Enthaler, Catal. Lett., 141 (2011) 55.
[26] J.-B. Cazaux, P. Braunstein, L. Magna, L. Saussine, H. Olivier-Bourbigou, Eur. J. Inorg. Chem., (2009) 2942.
[27] Y. Wang, H. Ma, J. Organomet. Chem., 731 (2013) 23.
[28] H.-D. Kim, H. Ishida, J. Phys. Chem. A, 106 (2002) 3271.
[29] L. Chen, W. Li, D. Yuan, Y. Zhang, Q. Shen, Y. Yao, Inorg. Chem., 54 (2015) 4699.
[30] A. Maisse-Francois, L. Azor, A.-L. Schmitt, A. Coquel, L. Brelot, R. Welter, S. Bellemin-Laponnaz, S. Dagorne, J.
Organomet. Chem., 696 (2012) 4248.
[31] S. Dagorne, J. Organomet. Chem., 691 (2006) 4797.
[32] W. Liu, J. Holladay, Catal. Today, 200 (2013) 106.
[33] M.B. Smith, J. March, March's Advanced Organic Chemistry: Reactions, Mechanisms, and Structure, 5th Edition, 2000.
[34] P. Wrigstedt, J. Keskivaeli, M. Leskelae, T. Repo, ChemCatChem, 7 (2015) 501.
[35] Y. Yang, W. Liu, N. Wang, H. Wang, W. Li, Z. Song, Chin. J. Chem., 33 (2015) 583.
[36] B. Girisuta, L.P.B.M. Janssen, H.J. Heeres, Ind. Eng. Chem. Res., 46 (2007) 1696.
[37] J. Horvat, B. Klaic, B. Metelko, V. Sunjic, Tetrahedron Lett., 26 (1985) 2111.
[38] S.K.R. Patil, J. Heltzel, C.R.F. Lund, Energy Fuels, 26 (2012) 5281.
[39] S.K.R. Patil, C.R.F. Lund, Energy Fuels, 25 (2011) 4745.
[40] I. van Zandvoort, Y. Wang, C.B. Rasrendra, E.R.H. van Eck, P.C.A. Bruijnincx, H.J. Heeres, B.M. Weckhuysen,
ChemSusChem, 6 (2013) 1745.
[41] C. Yao, Y. Shin, L.-Q. Wang, C.F. Windisch, Jr., W.D. Samuels, B.W. Arey, C. Wang, W.M. Risen, Jr., G.J. Exarhos, J.
Phys. Chem. C, 111 (2007) 15141.
[42] Y.-S. Lim, H.-S. Kim, M.-S. Kim, N.-H. Cho, S. Nahm, Macromol. Res., 11 (2003) 122.

Effect of ancillary
(aminomethyl)phenolate ligand on
aluminum-catalyzed glucose
dehydration to 5-
hydroxymethylfurfural
Daudi Saang’onyo, Sean Parkin, Folami T. Ladipo*
Chemistry Department, University of Kentucky, Lexington, Kentucky 40506-0055, USA
Modification of (aminomethyl)phenolate ligand (LR) furnished LRAlMe2 complexes that
efficiently catalyze glucose dehydration to 5-hydroxymethylfurfural in ionic liquids. In
[BMIM]Br at 120 ˚C, LphAlMe2 (2e) produced HMF with 63%