In literature, a wide range of catalysts have been described and tested to conduct the

hydrogenolysis of glycerol to propylene glycol. Different metals, i.e. Ru, Rh, Re, Ir, Pd, Pt, Au,
Cu, Ni, Co, Mn, Mo have been used as mono- or bimetallic catalysts supported on a range of
different materials, i.e. C, Al2O3, SiO2, ZnO, Amberlyst, zeolites, in the Raney form (Ni and
Cu), and as mixed oxides or salts (Mg, Al, Zn, Ni, Co and Cu oxides in different combinations).
The use of acidic or basic co-catalysts together with these metallic catalysts has also been
reported. The hydrogenolysis reaction is known as a bifunctional reaction which requires
catalysts for both dehydration and hydrogenation functionality.
These catalysts used can be divided into several, sometimes overlapping categories. These
catalysts have, in different reviews, been categorized in different ways. Nakagawa et al. [25]
divided the catalysts in noble metal catalysts, which could be monometallic or combined with an
acid, base, or oxide; and first row transition metal catalysts. Palkovits et al. [23] divided the
catalysts based on the kind of metal they contained. Ten Dam et al. [26] made a distinction
between alkaline and acidic catalysts. In this section, various hydrogenolysis catalysts will be
discussed, which will be divided into two main types: noble metal catalysts and non-noble metal
catalysts. The main challenge that existsin the hydrogenolysis of glycerol is the selective
cleavage of C-O bonds over C-C bonds. Although noble metal catalysts are usually more active
than Cu- or Ni-based ones, their selectivity towards propylene glycols is lower since they are
able to cleave both C-C and C-O bonds.
2.3.3.1 Noble metal catalysts
Hydrogenolysis uses hydrogen as a reactant. Therefore, one of the most important abilities a
hydrogenolysis catalyst must have, is the ability to activate hydrogen molecules. Noble metals
are well known to be able to activate hydrogen molecules and are widely used as hydrogenation
catalysts [25]. Commonly used noble metals include Ru, Rh, Pd and Pt. Miyazawa et al. [21]
examined various carbon-supported noble metal catalysts (Ru, Rh, Pd and Pt).
The results of an activity test they performed are given in Table 2-3. Ru resulted in the highest
conversion, but the use of Rh/C resulted in a higher selectivity towards 1,2-PDO. In a later
report, they showed that a Rh/SiO2 catalyst was more active than the previously examined Ru/C
catalyst.
Recently, supported mono- and dimetallic Ru catalysts have been used for the Glycerol
hydrogenolysis reaction. Feng et al. [27] studied the effect of the support and catalyst reduction
temperature on the catalytic performance of ruthenium catalysts in the hydrogenolysis of
glycerol. A range of different ruthenium catalysts were tested which included Ru/SiO2, RuNaY,
Ru/-Al2O3, Ru/C a,d Ru/TiO2. The results are shown in Figure 2-14. Among the tested
catalysts, the TiO2 supported catalyst exhibited high activity giving the highest conversion of
glycerol. It did however favor the
production of EG over 1,2-PDO under the tested reaction conditions, i.e. 180 C, 5 MPa. The
Ru/SiO2 catalyst showed the lowest activity, but resulted in a much higher selectivity to 1,2-
PDO than that of EG. By varying the reaction temperature and hydrogen pressure, the authors
were able to increase the selectivity of the Ru/TiO2 catalyst, but at a lower glycerol conversion.
According to the authors, the reason for this difference in activity and selectivity can be
explained by the fact that the different supports determine the mean Ru particle size of the
catalysts before reaction. By using XRD and TEM techniques, it was determined that the mean
Ru particle size before reaction decreases in the order Ru/SiO2 > Ru/NaY > Ru/-Al2O3, Ru/C,
Ru/TiO2 which is consistent with their increasing activity. This result indicates that the
hydrogenolysis of glycerol is more active on small metal particles. The difference in selectivities
towards different products was explained based on the intrinsic properties of the support. The
support material can influence the reaction routes in the presence of metal catalyst. They
assumed the glycerol hydrogenolysis occurs through the dehydrogenation mechanism originally
proposed by Montassier et al. As mentioned above, glyceraldehyde can either be dehydrated to
form 2-hydroxyacrolein which can then be hydrogenated to propylene glycol, or it can undergo a
retro-aldol reaction towards glycolaldehyde which is then hydrogenated to ethylene glycol.
Under the same reaction conditions, SiO2 or -Al2O3 favors the dehydration of glyceraldehyde
over the retro-aldolization route, leading to higher selectivity to 1,PDO. In contrast, TiO2 favors
the retro-aldolization route which results in higher selectivity to EG.
The use of active carbon as support results in similar selectivities towards EG and PG because
Ru/C catalyst have no preference for either reaction route. Feng et al. also investigated the
influence of catalyst reduction temperature. The results are given in Table 2-4. The lower activity
of high-temperature reduced catalysts can be attributed to the growth of Ru particles caused by
the heat treatment and by the strong metal-support interaction, i.e., SMSI effect, which causes
partial covering of the surface with Ti2O3 species. The change in selectivities is explained by the
SMSI effect. The covering of particles by Ti2O3 may favor the retro-aldolization reaction route
over the dehydration route which causes the selectivity towards EG to increase and the
selectivity to PG to decrease.
Since noble metal catalysts without additives generally produce more than one product [25],
different ways have been investigated to improve product selectivities. Miyazawa et al. studied
the effect of the addition of different acid catalysts, i.e. acid type cation-exchange resin
Amberlyst 15, H2SO4 and HCl, on different noble metal catalysts, i.e. carbon-supported Ru, Rh,
Pt and Pd. The effects were most clear for the carbon-supported Ru catalyst. On promotion, the
hydrogenation activity for the different metals decreased in the order Ru/C >>>Rh/C > Pt/C >
Pd/C. The promoting effect of Amberlyst 15 was much higher than that of H2SO4 and the
addition of HCl even suppressed
the glycerol reaction on Ru/C. This is due to the poisoning of the surface with Cl- ions. The
influence of the addition of Amberlyst 15 on the activity of Rh/C, Pd/C and Pt/C was also
examined. It also enhanced catalytic activity of Rh/C, but the effects were smaller than for Ru/C.
In the case of Pt/C and Pd/C, the addition of Amberlyst 15 was less effective. One problem with
the addition of Amberlyst 15 is the limitation of the reaction temperature by the thermal stability
of the ion- exchange resin. This can be solved by using a heat-resistant ion-exchange resin such
as Amberlyst 70. The addition of inorganic acids, which are thermally more stable than ion-
exchange resins and can thus be used at much higher temperatures, was investigated by Balaraju
et al. [9]. Several different acid catalysts were used including Nb2O5, TPA/ZrO2,CsTPA,
CsTPA/ZrO2. Balaraju et al. concluded that the total acidity of the catalyst plays a major role in
the overall glycerol conversion. Glycerol conversion could be linearly correlated to the number
of acid sites on the solid acid. This relation between the total acidity of the solid acid catalyst and
the glycerol conversion is shown in Figure 2-15. The increased glycerol conversion with
increased acidity indicates that dehydration of Glycerol takes place on acid sites to yield acetol
and its further hydrogenation on Ru.

Instead of the addition of an external acid, noble-metal catalysts supported on the solid acid have
also already been tested. Vasiliadou et al. [28] prepared alumina-supported ruthenium catalysts
with moderate acid sites and tested them in the hydrogenolysis of neat glycerol at 513 K. They
found that both the conversion of glycerol and the selectivity to over hydrogenolysis products
increased with the increase of the total acidity of the catalytic materials. Gandarias et al.
investigated the hydrogenolysis of glycerol over a platinum catalyst supported on acidic
amorphous silica-alumina (Pt/ASA). The ASA acid sites were responsible for glycerol
dehydration while Pt metal sites catalyze acetol hydrogenation.
Besides the addition of an acid, it is also possible to use a base as co-catalyst in the
hydrogenolysis of glycerol. Feng et al. reported the hydrogenolysis of glycerol catalyzed by
TiO2 in the presence of various bases at 443 K. The addition of lithium or sodium base increased
the conversion of Glycerol and the selectivity to 1,2-PDO quite dramatically. It also decreased
the selectivity to EG. When increasing the amount of base, the conversion of glycerol and the
selectivity to 1,2-PDO first increased and then reached a maximum after which the conversion
and selectivity began to decrease again. This effect was explained based on the glyceraldehyde
based mechanism originally proposed by Montassier et al. The addition of a base aids the initial
dehydrogenation of glycerol to glyceraldehyde. The decrease in selectivity towards EG is
explained by the fact that the retro-aldol reaction, which leads to EG, is very difficult in basic
conditions [25]. Maris and Davis. investigated the hydrogenolysis of glycerol over carbon-
supported Ru and Pt catalysts. Ru/C was more active than Pt/C but was much less selective. The
addition of a base, i.e., both NaOH and CaO, enhanced the rate of reaction over both catalysts.
The activity of Pt/C increased more than that of Ru/C. This is consistent with the fact that the
rate of alcohol dehydrogenation over Pt is enhanced by a base to greater extent than over Ru.
Since the dehydrogenation of glycerol to glyceraldehyde is proposed to be the first step in the
glycerol hydrogenolysis mechanism, the activity of Pt/C is enhanced more than Ru/C with the
addition of a base [16].
Another commonly used strategy to improve catalyst activity is the co-immobilization of a
second metal with the hydrogenating metal [26]. Examples include the addition of chromium to
promote the activity of copper, or the combination of Ru and Cu on a bentonite carrier.
Supported dimtallic catalysts with Ru and Re or Cu have recently been applied for the
hydrogenolysis of glycerol. He et al. [29] investigated the influence of rhenium on the activity
and selectivity of several heterogeneous ruthenium catalysts. They compared various Ru-Re
bimetallic catalysts( Ru-Re/SiO2, Ru-Re/ZrO2, Ru- Re/TiO2, Ru-Re/H-, Ru-Re/H-ZSM5) and
Ru monometallic catalysts (Ru/SiO2, Ru/ZrO2, Ru/TiO2, Ru/H-, Ru/H-ZSM5). Compared with
monometallic catalysts, the Ru-Re bimetallic catalysts showed much higher activity in the
hydrogenolysis of glycerol, and Re exhibited obvious promoting effect on the performance of
catalysts. The addition of heterogeneous Re2(CO)10 was most effective in combination with
Ru/Al2O3. It not only increased glycerol conversion, but also improved 1,2-PD and 1,3-PD
selectivity, by preventing the degradation of glycerol into EG. In combination with Ru/C, the
rhenium improved both conversion and selectivity but to a lesser extent. When adding
Re2(CO)10 to Ru/ZrO2, the selectivity towards propanediols was increased, but the glycerol
conversion stayed about the same.
According to the authors, the presence of Re improves the dispersion of rhodium on the surface.
The surface acidity of ReOx can also promote the dehydration step in the hydrogenolysis of
glycerol. An interesting report concerning a RhRe catalyst was written by Shinmi et al. [24] in
which they proposed an alternative direct hydrogenolysis mechanism on the catalyst. This
mechanism was described above.
2.3.3.2 Non-noble metal catalysts
Besides noble metals, Cu, Ni and Co are known to be able to activate hydrogen molecules and
are often used as hydrogenation catalysts. Although the hydrogenation activity is generally lower
than that of noble metals, the much lower prices and the higher resistance to poisoning by trace
impurities make non-noble-metal catalysts practically important [25]. Copper catalysts are of
high interest for the hydrogenolysis of glycerol. Cu is not as active as noble metals but it does
not catalyze C-C bond cleavage and thus increases the selectivity for PG. Copperchromite is a
conventional hydrogenation catalyst which is industrially used because of its high selectivity
towards PG. Nearly 90% selectivity at 65 % conversion is obtained at 473 K. The use of Cu-ZnO
catalysts, which are industrially used in methanol synthesis and methanol steam reforming, has
also already been well studied. Selectivities of over 80% to PG are reported at relatively low
conversions. The pre-reduction of these catalysts is necessary to avoid the aggregation of Cu
particles in the reaction solutions and to obtain high selectivity towards PG. Other investigated
catalysts include Cu/Zn/Al mixed oxide catalysts, supported Cu catalysts on silica or alumina, a
co-precipitated Cu-Al catalyst, a Cu-Zn catalyst supported on alumina, a Cu/MgO catalyst
prepared through coprecipitation. The activity and selectivity of copper immobilized on various
aluminum-containing acidic supports was investigated by Zhang et al. [30]. They believe that
Al2O3 could possess an appropriate acidity to catalyze the dehydration of glycerol to acetol.
They concluded that the support plays an important role in the catalysis besides immobilizing
and dispersing the active Cu species. When looking at the results in Table 2-5, it can be seen that
pure alumina was more active and selective than the more acidic zeolitic supports (H, HY,
HZSM-5, and 13X), which show minimal or no activity at all. This could be because of the
preferred selectivity of the zeolites towards acrolein. The strong metalsupport interaction due to
the acidity and high surface area of these zeolites could also make the reduction of CuO to active
metallic copper difficult under the conditions used (300 C).
Most Cu-based catalysts need a pre-reduction step to have high activity. Reduction conditions
thus have a great influence on the catalytic performance of the catalyst. Guo et al. [30]
investigated the influence of the reduction temperature on the catalytic performance of a
Cu/Al2O3 catalyst, the results are given in Table 2-6. With an increase of reduction temperature,
glycerol conversiondisplayed a volcano shape. Maximal conversion was obtained at a reduction
temperature of 300C.
The selectivity values are relatively insensitive to the reduction temperature. Guo et al.
investigated TPR patterns of the Cu/Al2O3 catalyst and found that the reduction of CuO on the
sample started at a temperature of about 200C and was completed around 280C. Next, they
investigated XRD patterns of Cu/Al2O3 reduced at 200, 300 and 400C in H2. The results they
obtained are shown in Figure 2-16. At 200C, the signal of the crystallite phases of metallic Cu
was very weak, which indicated that part of the Cu species existed as CuO or Cu2O. XRD
patterns of the catalyst reduced at 300C showed crystallite phases of Cu metal, consistent with
the TPR results that at this temperature, the Cu was completely reduced. At 400C, the signal
that corresponds with the metallic Cu phase became much stronger, indicating growth of the Cu
particles. This suggests that a high-temperature reduction could cause the sintering of well
dispersed active Cu species on the support, which could be responsible for the decreased
catalytic activity of the Cu/Al2O3 reduced at 400C.
Dasari et al. [20] also performed studies in order to determine the activity of a copper-chromite
catalyst as a function of catalyst reduction temperature. They obtained similar results, i.e., yield
and selectivity of propylene glycol increase with catalyst reduction temperatures up to 300C and
decrease for temperatures greater than 300C. The effect of the Cu loading on the conversion and
selectivity of a Cu/Al2O3 catalyst was investigated by Guo et al. The results of several reactions
on the catalysts with different amounts of Cu are given in Table 2-7.Gcon/Cu represents the
molar ratio of the amount of glycerol converted per hour during 10 h reaction and the Cu
quantity used in the reaction. When the loading passed 2.7 mmol Cu g-1, this quantity started to
drop quickly. Glycerol conversion rises almost proportionally with the amount of Cu and reaches
a maximum at a loading of 2.7 mmol g-1, after which the conversion decreased with increased
copper loading. When investigating XRD patterns of the catalysts with varying amounts of Cu
loading, the strength of the signals increased with the amount of Cu loading. From these results,
Guo et al. concluded that when the loading is less or equal to 2.7 mmol g-1, the active Cu species
are well dispersed and can efficiently be used, while at higher loading, larger Cu crystallites are
formed.
These deteriorate the dispersion of the active metal, but can also block the pores of the support.
Therefore the total amount of accessible active metal sites decreases, although the total metal
loading increases. This explains why both the conversion of glycerol and the GCon/Cu decreases
after the loading exceeds 2.7 mmol g-1.
Nickel-based catalysts include Raney Nickel and other supported catalysts are also tested. The
main product is also propylene glycol. Ni-catalysts however also catalyze the C-C
hydrogenolysis to EG or ethanol and the selectivity towards PG is lower than that of copper-
based catalysts. The addition of a secondary metal has also been tested for Ni catalysts. The
addition of cerium to Ni supported on active carbon decreased the reduction temperature of Ni,
resulted in smaller particle sizes and enhanced catalytic activity. When adding Pt to a Ni/SiO2
catalyst, the C-C hydrogenolysis was enhanced, resulting in more EG.
The use of cobalt-based catalyst has been less studied. Liu et al. however prepared various Co
nanostructures and used them as hydrogenolysis catalysts of glycerol. This resulted in selectivity
towards PG between 57 and 72% at 493K [25].