10
M.A. Jackson et al. / Carbohydrate Research 432 (2016) 9e16
where tetrahydrofurfurylalcohol is the main product. One of the
byproducts, described as high-boiling, was isolated as a colorless
crystal melting at 103 ꢀC. Following discussions between Swadesh
and Hurd, it was concluded that each group was studying the same
compound, namely 1. After a fifty year absence from the literature,
the structure of 1 was determined by Gaede et al. [26] This group
used a Pd/C catalyst for the hydrogenation of furfuryl alcohol and
recovered 1 at a yield of less than 5%. They found the product
deposited in the condenser while attempting a distillation. Each of
these groups also noted the monomeric, ring-opened form of the
compound,1,5-dihydroxy-2-pentanone. More recently, Zhang, et al.
[27] isolated 1 and several derivatives including the mono- and
diacetylated derivatives of 1,5,-dihydroxy-2-pentanone from the
fruiting spores of the basidiomycete Catathalesma imperial. The
compounds were isolated by ethyl acetate extraction and purified
by column chromatography in quantities ranging from 2 to 110 mg/
kg of fungal fruiting tissue. Three of these isolated compounds in
open chain form, though not the cyclic species, were found to be
pore diameter and volume. Calcination at 973 K causes a greater
loss in surface area compared to the lower temperatures. PdO
crystallite size is also affected by calcination temperature with the
higher temperatures giving larger crystallites. Metal oxide supports
are known to be susceptible to degradation in hot water and this
can be seen in both the table and in the shape of the N2 isotherms.
The loss in surface area and expansion of the pores reveals the
change in the support structure. Fig. 1 shows the hysteresis loop of
the fresh catalyst is a Type H2 typical of metal oxide gels and
indicative of complex, interconnected pore structure. After use, the
pore structure appears to collapse into fewer but larger pores
resulting in a hysteresis loop that takes on the appearance of Type
H1. This reorganization also affects the PdO crystallite size as the X-
ray diffraction results from the fresh and used catalysts indicate
that the crystallites are smaller after the catalysts have been used.
The energy dispersive X-ray spectroscopic analysis of Pd and Al also
indicate that the Pd migrates to the surface as a result of both
calcination temperature and use. Overall, this means that the
calcination temperature becomes irrelevant after the catalysts are
exposed to hot water. (See the SI for the XRD and EDS spectra in
Figs. S1 and S3).
inhibitors of both human and mouse 11b-hydroxysteroid de-
hydrogenases. These enzymes are seen as targets for treating hy-
pertension, type II diabetes, and metabolic syndrome [28]. This
report offers some expectation for the utility of 1, either as the
tricyclic or its open chain form.
Fig. S2 displays the temperature programmed hydrogenation of
the catalysts after being calcined at temperatures of 673 K, 773 K,
and 973 K. The negative peak, an area where hydrogen is being
We rediscovered the title compound while screening catalysts
for the hydrodeoxygenation of biomass hydrolysates. The hydro-
lysates were prepared by passing hot, compressed water over ball-
milled switchgrass. These screening reactions were performed on
these murky suspensions at 433 K and 7 MPa hydrogen pressure
and resulted in solutions containing products from both lignin and
sugar components of biomass. Analysis of these solutions by gas
chromatography/mass spectrometry and HSQC NMR led to the
tentative identification of 1. The structure of 1 suggested that it
originated with xylose and/or arabinose that were hydrolyzed from
hemicellulose. Therefore, these pentoses were hydrogenated under
the conditions of our screening reactions to see if 1 could be syn-
thesized in high yields and to see if it could serve as biobased
platform chemical. The results of these studies are presented here.
evolved from the catalyst, is indicative of the b-hydride capacity of
the reduced Pd0 on the surface of the catalyst. The TPR trace of the
catalyst calcined at 673 K shows a smaller negative peak at 358 K
and a large hydrogen absorption peak near 573 K. The peak around
573 K results from better dispersed Pd on the surface. At higher
calcination temperatures, the b-hydride peak increases in size and
the dispersed Pd is reduced in relative amounts.
2.2. Hydrodeoxygenation of D-(þ)-xylose
The title compound is a colorless solid with a literature melting
point of 376 K; at high purity, it sublimes; it has a fragrance vari-
ously described as like that of hazelnut or bourbon. The hydro-
deoxygenation (HDO) of xylose to 1 and the successive conversion
of this to the ring-opened species 1,5-dihydroxy-2-pentanone (3)
and on to the acetylated derivative (4) are shown in Fig. 2. The
yield-limiting step in the reaction series is the reaction of xylose
with a yield of 1 typically at 35e40%. Fig. 3 displays the product
selectivity at high conversion of xylose. The identified side products
of this reaction result from as many as four competing pathways
that need to be inhibited to increase the yield of 1. These include
hydrogenolysis in which carbon-carbon bonds are broken giving
hydroxypropanone and 1-hydroxy-2-butanone. The second
pathway is the HDO of xylose in which carbon-oxygen bonds are
broken resulting in the formation of water and the C5 products 1-
hydroxy-2-pentanone, cyclopentanone, and THFA. The third
pathway is the hydrogenation of the carbonyl group to give xylitol.
A fourth pathway is the dehydration of xylose to furfural, which is a
well known route to platform chemicals from biomass [4e7,9,29].
But furfural is only detected at very low levels during the HDO
reaction. Although furfural could be seen as the logical precursor to
1 if it was hydrogenated to furfuryl alcohol, the starting material in
the previous reports for the synthesis of 1. In fact, the hydrogena-
tion of furfuryl alcohol over Pd/SA does not give 1. Furfural could
also be proposed as the precursor to both THFA [13] and the
cyclopentanone [30,31]. The hydrogenation of furfural proceeds
only slightly under these reaction conditions over Pd/SA so the
THFA likely forms from the HDO pathway. Furfural is not formed if
the catalyst is more active toward the reactions involving hydrogen
than toward dehydration. That is, palladium catalysts activate
hydrogen and produce the reduced organic species. Other
2. Results and discussion
2.1. Catalyst selection and characterization
From our screening reactions it was apparent that palladium was
the active metal with the greatest tendency to produce 1 from
xylose and it performed best when supported on silica-alumina
(SA). Other supports that were examined included commercial 5%
Pd/CaCO3, 5%Pd/C, and 5%Pd/Al2O3. Of these, only the 5%Pd/Al2O3
gave 1 in yields near that obtained with Pd/SA. 5%Pd/C gave 1 at <5%
and xylitol and tetrahydrofurfuryl alcohol (THFA) at selectivities of
35% and 41%, respectively. 5%Pd/CaCO3 produced mostly the
hydrogenolysis products acetol, 3-hydroxy-2-butanone, and acetic
acid and 1 at only 6%. Palladium was also impregnated onto other
supports. 5%Pd/AlPO4 was prepared to examine the role support
acidity might play. This catalyst did produce 1 at 12% selectivity
(and THFA at 38%) but it was visibly degraded after one use. Palla-
dium was loaded onto H-Beta zeolite, another acidic support, which
gave 1 at 20% yield along with THFA and acetol. Following these
results, we focused on the Pd/SA catalysts for the synthesis of 1.
A series of Pd/SA catalysts with Pd loadings ranging from 3% to
10% were prepared by wet impregnation using PdCl2 dissolved in
aqueous ammonium hydroxide. The surface properties of fresh and
used 5Pd/SA catalysts calcined at 673e973 K are shown in Table 1
and the N2 isotherms of the catalyst calcined at 973 K are shown
in Fig. 1. It can be seen in the table that addition of the palladium to
the support causes a decrease in surface area and small changes in
Jackson, Michael A.
