ARTICLES
experimentally. Moreover, this integrated Co atom can further 15 min to remove unexfoliated precursor and only the supernatant was collected by
pipette. Eventually, the exfoliated MoS
aqueous solution. The pH value of the dispersion became alkaline due to the
presence of trace lithium hydroxide. To collect the MoS
added dropwise into the above colloid until the pH value reached around 7. The
2
layers remained totally suspended in
promote new S vacancies in proximal sites in an energetically favour-
able manner during our hydrogen activation (Fig. 4e) due to the
electronic promotion effect of Co–Mo (presumably with the mixing
S
2
powder HCl solution was
of d-orbitals of Co with Mo) as discussed in ‘Assessment of electronic precipitate was then washed and collected by centrifugation several times and the
S
final product was dried under vacuum for 12 h.
influence of single Co atoms on MoS ’ in the Supplementary
2
Information (also see Supplementary Figs 17–20). Indeed, HAADF-
S
S
Preparation of Co– MoS
2
catalyst. The isolated Co atom promoted MoS
2
STEM, EELS and image simulations (Figs 2 and 3) confirmed the pres-
S
S
(Co– MoS
2
) catalyst was prepared using a hydrothermal method. 50.6 mg of cobalt
acetate tetrahydrate and 36 mg of thiourea were added to 1 ml of water and left
ence of all calculated Co sites in the post-reaction Co– MoS sample
2
2+
(
Coon edge; Co atop to Mo;Coon V ; Co on hollow site, etc). The low overnight to form the complex Co ion Co(thiourea)4 (ref. 28). Subsequently this Co
S
−1 S
V formation energy at the interface of Co–[S]–Mo on the edge site thiourea complex solution was then added into 50 ml of 1 mg ml stabilized MoS
has been discussed previously
2
S
3,32
colloid (30 v/v% isopropanol/water with 50 mg of polyvinylpyrrolidone). A mixed
homogeneous solution was then transferred to an autoclave, followed by
hydrothermal treatment at 160 °C for 24 h. In order to anchor Co atoms onto the
.
The availability of coordinatively unsaturated sulfur vacancies
and hollow basal sites to accommodate additional Co atoms in
S
defects of MoS
2
rather than inducing a fast self-nucleation, we used a very low
this unconventionally prepared single molecular layer MoS , akin concentration of the Co complex. During the hydrothermal process, the S
2
2
to the edge site are clearly evident. Indeed, we have showed Co thiourea–Co complex refilled some of the surface sulfur vacancy sites of MoS .
Previous studies confirm that defective vacancy sites possess high molecular affinity,
especially for thiol-based molecules
washed three times using deionized water and then dried under vacuum for 12 h and
2
finally treated with H at 300 °C for 1 h prior to storage.
atoms can take residence on both the edge sites and basal sites of
1
7–19
S
. After the reaction, the precipitate was
Co– MoS . Studies have pointed to edges as being the active sites,
2
however, basal S vacancies have also been proposed as possible insti-
33–35
gators of chemical reactivity of the MoS basal plane
. Although
2
F
our present data reveal that the general increase in basal S vacancies Preparation of MoS . 6 g of bulk MoS powder was dispersed in 400 ml of
can contribute toward catalytic activity, it is not conclusive exactly isopropanol/water at 30% v/v, and 4 ml of hydrazine monohydrate . The system
which sites are active and to what degree they contribute to the
overall activity, thus we cannot exclude the possibility that catalysis
is dominated by the edges. However, the much larger number of
2
2
2
6
was sonicated for 12 h, and then centrifuged at 2,000 r.p.m. for 60 min to remove
non-exfoliated sediments. In order to collect the exfoliated MoS
suspension, we changed the pH to around 7 by adding HCl solution. The charge
balance in the dispersion would be disrupted, which would cause a fast flocculation.
F
2
powder from
basal versus edge sites and the fluxional (dynamic) behaviour of Co The final products were centrifuged, washed with deionized water three times and
at elevated temperatures, that is, their ability to diffuse between dried in the vacuum container for 12 h.
edges and basal planes, means that a sizeable population of Co
atoms reside in these planes facilitating the formation of sulfur
Hydrodeoxygenation of 4-methylphenol. Hydroxygenation (HDO) of
4-methylphenol was carried out in a 50 ml sealed autoclave reactor (Parr 4792,
vacancies. Thus, our assertion is that the dramatic increase in the
pressure vessel) with an inner glass liner. In a typical procedure, 280 mg of
4-methylphenol, 20 mg of catalyst and 280 mg of octane (internal standard) were
S
activity of MoS towards HDO of 4-methylphenol upon Co pro-
2
motion must be related to the consequent wider availability of dispersed in 10 ml decalin and then transferred to the reactor. The autoclave was
2
purged with H several times, then pressurized to 3 MPa and heated to the chosen
these basal vacant sites, which act as catalytically active centres for
temperature. After reaction, the autoclave was cooled down naturally to room
temperature. A small volume of liquid sample was collected and then analysed by gas
chromatography (Agilent 6890N). The products were confirmed against pure
this reaction. Finally, we assessed the tendency of sulfur loss of the
S
Co– MoS catalyst (see ‘DFT calculations’ in the Supplementary
2
S
Information). In both the pristine and Co-promoted MoS , S-loss samples by mass spectrometry (MS). For all the experimental data, the carbon
2
is uphill in energy, whereas HDO is exothermic when oxygenated balance was >95% and experiments were repeated at least three times to ensure the
repeatability of the data. The rate of the HDO of 4-methylphenol was calculated
assuming a pseudo-first-order reaction as below,
compound is present. This calculation supports the lack of sulfur
loss and the stability of the catalyst observed during the HDO
of 4-methylphenol.
ln(1 − x) = −kCcat
t
(1)
−
1
−1
Conclusions
where k is the pseudo-first-order rate constant (ml s mol ), x is the conversion of
-methylphenol, Ccat is the concentration of catalyst under reaction system and t is
the reaction time (s).
4
In summary, a strong covalent attachment of Co atoms to mono-
layer MoS through defect engineering significantly enhances the
2
Sample characterizations as well as the theoretical methods and computational
details, are described in the Supplementary Information.
number of Co–S–Mo interfacial sites on the basal plane (Co on
atop site to Mo; Co in hollow site; and Co substituting S site, etc),
as shown by STEM and EELS. In this work, we employed the strat- Data availability. All relevant data are available from S.C.E.T. and detailed
Supplementary Information is available online containing descriptions of methods;
Supplementary Figs 1–20 with results and discussions; assessment of electronic
influence; DFT calculations; anticipated advantages of the new catalyst; background
egy of creating a large number of basal sulfur vacancies within Co-
doped single-layer MoS through high-temperature treatment in H2
2
(
300 °C), which provided a sufficient number of Co–S–Mo active
sites for the HDO reaction to occur at a low operating temperature
180 °C). The catalyst described is extremely active, selective and Received 9 September 2016; accepted 23 January 2017;
of the desulfurization reaction and references.
(
stable for the conversion of 4-methylphenol to toluene in hydrogen published online 6 March 2017
at low operation temperature and opens the possibility of using this
novel type of catalyst for HDO for biomass conversion without References
1.
Prins, R., De Beer, V. H. J. & Somorjai, G. A. Structure and function of the
catalyst and the promoter in Co–Mo hydrodesulfurization catalysts. Catal. Rev.
sulfur loss from the catalyst.
31, 1–41 (1989).
Methods
2. Grange, P. Catalytic hydrodesulfurization. Catal. Rev. 21, 135–181 (1980).
3. Byskov, L. S., Nørskov, J. K., Clausen, B. S. & Topsøe, H. DFT calculations of
S
Preparation of single-layered MoS
2
( MoS
2
). In a typical experiment, 0.5 g of bulk
MoS
under nitrogen atmosphere. After the intercalation of MoS
produced Li MoS was washed using vacuum filtration with 50 ml of hexane to
remove excess butyllithium and organics and then dried under N . Then the powder
2
crystalline powder was soaked in 4 ml of 1.6 M n-butyllithium/hexane for 48 h
2
unpromoted and promoted MoS -based hydrodesulfurization catalysts. J. Catal.
187, 109–122 (1999).
2
by lithium, the
x
2
4. Karunadasa, H. I. et al. A molecular MoS
2
edge site mimic for catalytic hydrogen
2
generation. Science 335, 698–702 (2012).
was immersed in 250 ml of water and the resulting suspension was sonicated to assist
the completion of the exfoliation process, as any intercalated lithium would react
5. Elliott, D. C. Historical developments in hydroprocessing bio-oils. Energy Fuels
21, 1792–1815 (2007).
with water to form H
the separation of MoS
2
gas between the layers. The evolution of H
layers. The dispersion was centrifuged at 5,000 r.p.m. for
2
tended to assist
6. Saidi, M. et al. Upgrading of lignin-derived bio-oils by catalytic hydrodeoxygenation.
Energy Environ. Sci. 7, 103–129 (2014).
2
6
©
2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.