On-surface azide-alkyne cycloaddition on Cu(111): Does it "click" in ultrahigh vacuum?

Article abstract of DOI:10.1021/ja312303a

Using scanning tunneling microscopy, we demonstrate that the 1,3-dipolar cycloaddition between a terminal alkyne and an azide can be performed under solvent-free ultrahigh vacuum conditions with reactants adsorbed on a Cu(111) surface. XPS shows significa

Full text of DOI:10.1021/ja312303a

Communication
pubs.acs.org/JACS
On-Surface Azide−Alkyne Cycloaddition on Cu(111): Does It “Click”
in Ultrahigh Vacuum?
Fabian Bebensee,† Christian Bombis,† Sundar-Raja Vadapoo,† Jacob R. Cramer,‡ Flemming Besenbacher,†
Kurt V. Gothelf,*,‡ and Trolle R. Linderoth*,†
†Sino-Danish Center for Molecular Nanostructures on Surfaces, Interdisciplinary Nanoscience Center (iNANO) and Department of
Physics and Astronomy, Aarhus University, 8000 Aarhus C, Denmark
‡Sino-Danish Center for Molecular Nanostructures on Surfaces, Danish National Research Foundation: Center for DNA
Nanotechnology, Interdisciplinary Nanoscience Center (iNANO) and Department of Chemistry, Aarhus University, 8000 Aarhus C,
Denmark
S
* Supporting Information
Scheme 1. General Huisgen 1,3-Dipolar Cycloaddition of
Azides and Alkynes
a
ABSTRACT: Using scanning tunneling microscopy, we
demonstrate that the 1,3-dipolar cycloaddition between a
terminal alkyne and an azide can be performed under
solvent-free ultrahigh vacuum conditions with reactants
adsorbed on a Cu(111) surface. XPS shows significant
degradation of the azide upon adsorption, which is found
to be the limiting factor for the reaction.
he field of on-surface synthesis aims at forming covalently
T
interlinked molecular nanostructures adsorbed on solid
surfaces1−3 to complement less robust counterparts formed by
supramolecular (noncovalent) surface chemistry.4,5 This target
has spurred much activity investigating to what extent a range of
organic reactions can be performed under ultrahigh vacuum
(UHV) conditions with reactants adsorbed on surfaces of metal
single-crystals, e.g., imine condensation reactions,6,7 Ullmann
coupling type reactions,8−10 boronic acid based chemistry,9,11
cyclodehydrogenation,12 and formal cycloaddition reactions.13
A highly interesting class of reactions with potential for surface
chemistry are “click” reactions.14 Click transformations are
general, give rise to their expected products in high yields, and
have been used in numerous applications including bioconjuga-
tion,15,16 materials science,17 and drug discovery.18 The
prototypical click reaction is the copper-catalyzed Huisgen 1,3-
dipolar cycloaddition19,20 between a terminal alkyne and an azide
to produce a 1,4-disubstituted 1,2,3-triazole (CuAAC reaction,
Scheme 1). In contrast to the uncatalyzed Huisgen cycloaddition
reaction, which often produces a mixture of 1,5- and 1,4-triazole
regioisomers, the CuAAC reaction selectively produces the 1,4-
triazole. The CuAAC reaction is especially interesting for on-
surface synthesis, since it is byproduct free and proceeds with
little thermal activation owing to the “spring-loaded” nature of
the azide reactant. In particular, the calculated (gas-phase)
activation energy barrier for the CuAAC reaction of ∼0.65 eV21 is
comparable to the binding energy in supramolecular surface
assemblies stabilized by multiple hydrogen bonds,5 rendering the
CuAAC reaction interesting for post self-assembly covalent
capture.22 The CuAAC reaction has been used in solution for
grafting/functionalizing self-assembled monolayers (SAMs)
onto surfaces.23−25 However, it remains to be shown if the
CuAAC reaction can be performed in an on-surface synthesis
a
(A) The Cu catalyzed CuAAC reaction regioselectively produces the
1,4-triazole, while the uncatalyzed reaction often produces a mixture of
1,4- and 1,5-triazole regioisomers. (B) The on-surface reaction
investigated here between 9-ethynylphenanthrene (alkyne) and 4-
azidobiphenyl (azide) produces a 1,4-triazole.
scheme with the reacting groups adsorbed directly on a surface
and in the absence of a solvent and the solvated Cu(I) ions
normally catalyzing the reaction.
Here, we use a combination of STM and XPS to explore the
feasibility of the CuAAC reaction under UHV conditions. We
have designed the reactants 9-ethynylphenanthrene (alkyne) and
4-azidobiphenyl (azide) shown in Scheme 1. The aromatic
residues on these reactant molecules have sizes that allow for
thermal sublimation onto the surface and are sufficiently different
to make them distinguishable by UHV-STM. Inspired by the use
of Cu as catalyst for the solution-phase reaction, the experiments
were performed on a Cu(111) surface. Our study demonstrates
that it is indeed possible to react these two species on the
Cu(111) surface to form the 1,4-triazole product. Implications of
this finding regarding a catalyzed/uncatalyzed on-surface
reaction path are discussed. Importantly, we furthermore find
Received: December 21, 2012
Published: January 29, 2013
© 2013 American Chemical Society
2136
dx.doi.org/10.1021/ja312303a | J. Am. Chem. Soc. 2013, 135, 2136−2139

Journal of the American Chemical Society
Communication
held at rt. This procedure resulted in a surface covered by a large
variety of different structures which were carefully surveyed by
STM (e.g., Figure 1C). By comparing these images to the data set
obtained for the individual molecules, we can assign a large
portion of the observed features to either the clusters formed
after deposition of pure azides (e.g., the cross like structures), the
alkyne molecules (individual or dimers), or a combination of
azide clusters/alkynes in close proximity on the surface.
Interestingly, however, the survey also revealed a distinct and
new structure (marked in green in Figure 1C). This new
structure is considerably larger than any one of the reactant
molecules alone, yet it behaves as one entity, e.g., it is observed to
rotate as a whole. Overlaying this STM feature with a scaled
molecular structure of the expected 1,4-triazole reaction product
shows an excellent structural correspondence regarding the bent
molecular shape and the positions/appearance of the assumed
diphenyl and phenanthrenyl groups. In total, 35 observations of
this new feature were made in 8 independent experiments
(sampling ∼3200 deposited molecules), and both anticipated
surface enantiomers of the triazole product were observed.
To substantiate that this new species is the expected reaction
product, we prepared the 1,4-triazole by conventional ex situ
synthesis and deposited it on the Cu(111) surface. Figure 2
shows a side-by-side comparison of the STM signature of
individually adsorbed, ex situ prepared 1,4-triazole molecules and
the in situ product (cf. Figure S2 for more ex situ product STM
images). The general appearance of the two species matches very
well. The length (long molecular axis) and apparent height of the
ex situ (in situ) synthesized 1,4-triazole are 23.8 1.6 (22.2
Figure 1. STM images showing reactants and products of the in situ
Huisgen azide−alkyne cycloaddition on Cu(111). (A) 9-Ethynylphen-
anthrene (alkyne) adsorbed individually. (B) Clusters formed after
deposition of 4-azidobiphenyl (azide). (C) Surface after codeposition of
alkyne and azide. Two triazole reaction products are marked in green
and shown to the right at high magnification and with a superimposed
molecular model. All images acquired at 100−125 K with It ≈ 300 pA, Vt
= −1.4 V. Full color range (black to yellow) corresponds to ∼1.5 Å .
1.8) and 1.6
0.1 (1.4
0.4) Å , respectively, in good
quantitative agreement. The ex situ product shows a strong
tendency toward dimerization (cf. Figure S2). Dimerization is
less likely to occur for the in situ product due to the low coverage,
but a few instances of dimers from triazoles formed in situ were
observed, further corroborating that it is indeed the same species.
that interaction with the surface causes the azide group to
decompose in the majority of deposited azide reactants. This
decomposition constitutes the limiting factor for the on-surface
azide−alkyne cycloaddition reaction.
Figure 1A shows an STM image acquired at ∼120 K after
depositing the alkyne alone on a Cu(111) surface held at room
temperature (rt, 300 K). The molecules adsorb as isolated
entities, and the characteristic bent shape of the phenanthrenyl
group is clearly resolved. The two outermost lobes of the
phenanthrenyl group are imaged with different STM contrast,
suggesting they have different interaction to the substrate. This
asymmetry allows one to distinguish the two anticipated surface
enantiomers for the prochiral alkyne molecule.26 Statistical
analysis of the rotation angles for the phenanthrenyl units in
combination with overlay of molecular models (see Figure S1)
shows a preferential alignment of the ethynyl group along the
close-packed [1−10] directions. Such an alignment is consistent
with di-σ bonding of the ethynyl group to the Cu(111) substrate,
as proposed for phenyl-acetylene.27,28 At higher coverage, some
dimers of the alkynes are observed, but condensed islands do not
form. Deposition of the azide molecules alone results in a range
of small clusters (Figure 1B). The clusters comprise two to five
identical elongated entities, attributed to the biphenyl units, and
exhibit a number of characteristic motifs in which the biphenyl
units all point toward a central interaction node. The larger
clusters comprising four or five biphenyl units are most common,
especially at higher coverage. To explore the possibility of an on-
surface CuAAC click reaction, the alkynes and azides were
codeposited from separate crucibles onto the Cu(111) surface
Figure 2. Side by side comparison of STM images of the ex situ prepared
1,4-triazole (left) and the in situ reaction product (right). Note that due
to the pro-chirality of the alkyne, two different surface enantiomers are
expected and observed for the triazole. Full color range corresponds to
1.6 Å.
2137
dx.doi.org/10.1021/ja312303a | J. Am. Chem. Soc. 2013, 135, 2136−2139

Journal of the American Chemical Society
Communication
We therefore conclude that the click reaction proceeds in the
solvent-free UHV environment with reactants adsorbed on the
Cu(111) surface at rt.
intact on the surface. (Direct comparison of the binding energies
on the surface and in the SAM is not possible due to the different
bonding environments.) To further test for degradation of the
azide group, we evaluated C/N ratios as shown in Table 1. The
The observation of quite a number of unreacted molecules on
the surface after codeposition indicates an incomplete reaction,
which is somewhat surprising given the normally high yield of
click reactions. Attempts to increase the number of successfully
coupled molecules by varying the deposition order of the
reactants, raising the substrate temperature during deposition to
∼400 K, or post annealing at 400 K were not successful requiring
us to investigate possible explanations.
One requirement for an on-surface reaction to occur is that the
reactants can meet by surface diffusion. The observation of
clusters formed from the azide (Figure 1B) and step edge
decoration with alkynes after deposition at rt (cf. Figure S3) as
well as occasional STM observations of lateral displacements of
individual alkynes at 100−120 K, allows us to rule out kinetic
limitations in mass transport as a limiting factor.
Table 1. Quantitative Analysis of XPS in Figure 3
a
b
c
d
compound
1,4-triazole
C/N ratio
9.33
integral area
C/N by area
6.2
expt C/N
C 1s: 3.2
N 1s: 0.52
C 1s: 1.3
N 1s: 0.07
−
azide
4
18.6
28
a
b
From stoichiometry of intact molecules. In arbitrary units from the
integrated areas under the curves after background subtraction.
c
Calculated from int. areas stated; note that this ratio may deviate
from the stoichiometry due to different sensitivities for the two
different elements under the given experimental conditions.
d
Calculated stoichiometry with the 1,4-triazole taken as the reference.
The scaling factor translating the measured C/N peak area ratio for
the triazole to the stoichiometric ratio was applied to obtain the
experimental C/N ratio for the azide.
triazole was taken as reference compound since its XPS had the
anticipated spectral features and the STM signature clearly
corresponds to molecules adsorbed without decomposition. The
analysis presented in Table 1 reveals an experimentally
determined atomic C/N ratio of 28 for the deposited azides,
much higher than the stoichiometric C/N ratio of 4 for the intact
azide molecule. We therefore conclude from the XPS experi-
ments that the predominant portion of the azides are not
adsorbed intact on the surface but decompose and lose nitrogen
(the upper limit for intact azides from the XPS data is 14%
assuming the unrealistic scenario that all N atoms are in intact
azides). A nitrene species formed via abstraction of dinitrogen
from the azide group poses a potential intermediate degradation
product on the Cu surface. The C/N ratio for biphenylnitrene is
12, so the experimental C/N ratio of 28 (cf. Table 1) strongly
suggests degradation beyond this stage for some azides. The
clusters in Figure 1B are thus ascribed to fragmented azides.
A number of possible causes for the azide degradation were
tested. Storage prior to the experiments and heating under
ambient conditions to 383 K (well above the UHV sublimation
temperature) did not lead to degradation. In a control
experiment, the azide was deposited onto a glass plate using
our standard experimental parameters but for a prolonged dosing
time (days). Subsequent analysis of the sublimate by FTIR and
TLC both showed intact azides, demonstrating that the
sublimation process is not the cause of degradation either. The
degradation must therefore occur after adsorption on the
Cu(111) surface. To test if the degradation is particular to
interaction with Cu, we dosed the azide onto an inert Au(111)
surface. Also here, XPS showed severe degradation, interestingly
suggesting that the degradation is caused by adsorption on metal
surfaces in general and not Cu(111) in particular. Possibly,
charge transfer from the metal into the aromatic system of the
adsorbed molecules plays a role in the observed degradation of
the azide.
Figure 3. XPS of the C 1s and N 1s regions of ex situ prepared 1,4-
triazole and 4-azidobiphenyl (azide) on Cu(111). To have appreciable
intensity in the N 1s region, the spectra were acquired at higher coverage
than shown in the STM images but still in the monolayer coverage
regime.
To obtain further chemical information, we performed XPS
measurements. Figure 3 shows XPS of the C 1s (left) and the N
1s regions (right) for the ex situ prepared 1,4-triazole (top) and
the azide compound (bottom). For both species, the C 1s
spectrum shows a single peak at a binding energy of 285.2
(triazole) and 284.6 eV (azide). A slight asymmetry resulting
from the small proportion of carbon atoms in direct contact with
nitrogen atoms is marginally visible. In the triazole N 1s
spectrum, there are two main peaks at binding energies of 400.2
and 402.4 eV with an ∼2:1 intensity ratio. These peaks
correspond to N atoms at 1/3 and 2 positions of the triazole
ring, respectively, and are in agreement with published spectra for
the triazole moieties in alkane-thiol and Si−C grafted
SAMs.29−31 (Further analysis of the spectrum is hampered by a
highly nonlinear background from the clean Cu substrate, see
Figure 3B). The azide group is expected to give rise to two N 1s
XPS peaks, and XPS for azide moieties in SAMs has shown30
these peaks to be separated by ∼5 eV. The present N 1s spectrum
for the azide on Cu(111), in contrast, shows only a single peak at
398.2 eV with a shoulder at the low-binding energy side (∼396.7
eV). This XPS signature suggests that the azide group is not
The apparent yield calculated from the number of reactant and
product molecules observed on terraces of the Cu(111) surface is
∼2%, conservatively assuming a 1:1 ratio of reactants and that all
molecules are intact. The observed click reaction products must,
however, stem from the small fraction of intact azide molecules
present on the surface. The XPS data place an upper limit for this
fraction at ∼14%. Consequently, the reaction yield based solely
2138
dx.doi.org/10.1021/ja312303a | J. Am. Chem. Soc. 2013, 135, 2136−2139

Journal of the American Chemical Society
Communication
on intact azides is at least 15%, suggesting that the cycloaddition
reaction may proceed quite readily on the surface.
ACKNOWLEDGMENTS
■
We acknowledge financial support from the Danish National
Research Foundation, The Danish Council for Independent
Research, Natural Sciences, the Marie-Curie ITN SMALL, and
the Alexander von Humboldt-Foundation.
The reaction mechanism for the liquid-phase CuAAC reaction
involves a Cu(I) ion which forms a Cu-acetylide with the
terminal alkyne group under replacement of the alkyne proton.21
The regioselectivity toward the 1,4-regiosiomer results from
preferential binding of the azide nitrogen closest to the residue to
this Cu atom, aligning the alkyne and azide groups with their
residues pointing in opposite directions. In a later reaction step,
the Cu atom in the intermediate is replaced by a proton from the
solvent. The activation energy barrier for this CuAAC reaction
mechanism has been calculated to 0.65 eV.21 An equivalent on-
surface scenario could be envisioned involving free Cu adatoms
released thermally (or by interaction with alkynes) from Cu step-
edges. Here, the Cu surface would have to replace the solvent as
reservoir for released hydrogen, although it is noted that
recombinant hydrogen desorption from Cu(111) occurs around
rt,32 limiting the lifetime of H species on the surface. If, for some
reason, the CuAAC catalyzed channel is blocked on the surface, a
direct reaction should be considered. The energy barrier for the
(gas-phase) direct reaction mechanism has been calculated to
∼1.1 eV,21 i.e., within a few tenths of an eV from rendering it
relevant at rt. The observed likely di-σ bonding of the alkyne to
the Cu(111) substrate is somewhat analogous to the “concerted
mechanism” involving formation of a Cu(I) π complex with the
alkyne triple bond21 and could possibly help to lower the barrier
for a direct (surface-catalyzed) reaction path. While the liquid-
phase uncatalyzed reaction often produces a mixture of 1,5- and
1,4-regioisomers, we only find evidence for the 1,4-regioisomer
in the STM data, apparently pointing toward the CuAAC
mechanism. However, when we perform liquid-phase synthesis
with the protocol of the uncatalyzed path, we also obtain only the
1,4-product. This regioselectivity is ascribed to steric hindrance
making a prereaction arrangement with both bulky residues to
the same side of the reacting azide and alkyne groups
inaccessible. Steric hindrance will be even more important in a
surface-confined situation. To fully establish the on-surface
reaction path, further modeling will be required.
REFERENCES
■
(1) Gourdon, A. Angew. Chem. Int. Ed. 2008, 47, 6950.
(2) Mendez, J.; Lopez, M. F.; Martin-Gago, J. A. Chem. Soc. Rev. 2011,
40, 4578.
(3) Franc, G.; Gourdon, A. Phys. Chem. Chem. Phys. 2011, 13, 14283.
(4) Elemans, J. A. A. W.; Lei, S. B.; De Feyter, S. Angew. Chem. Int. Ed.
2009, 48, 7298.
(5) Barth, J. V. Annu. Rev. Phys. Chem. 2007, 58, 375.
(6) Weigelt, S.; Busse, C.; Bombis, C.; Knudsen, M. M.; Gothelf, K. V.;
Laegsgaard, E.; Besenbacher, F.; Linderoth, T. R. Angew. Chem. Int. Ed.
2008, 47, 4406.
(7) Weigelt, S.; Busse, C.; Bombis, C.; Knudsen, M. M.; Gothelf, K. V.;
Strunskus, T.; Woll, C.; Dahlbom, M.; Hammer, B.; Laegsgaard, E.;
̈
Besenbacher, F.; Linderoth, T. R. Angew. Chem. Int. Ed. 2007, 46, 9227.
(8) Gutzler, R.; Walch, H.; Eder, G.; Kloft, S.; Heckl, W. M.; Lackinger,
M. Chem. Commun. 2009, 4456.
(9) Faury, T.; Clair, S.; Abel, M.; Dumur, F.; Gigmes, D.; Porte, L. J.
Phys. Chem. C 2012, 116, 4819.
(10) Grill, L.; Dyer, M.; Lafferentz, L.; Persson, M.; Peters, M. V.;
Hecht, S. Nat. Nanotechnol. 2007, 2, 687.
(11) Zwaneveld, N. A. A.; Pawlak, R.; Abel, M.; Catalin, D.; Gigmes,
D.; Bertin, D.; Porte, L. J. Am. Chem. Soc. 2008, 130, 6678.
(12) Cai, J. M.; Ruffieux, P.; Jaafar, R.; Bieri, M.; Braun, T.;
Blankenburg, S.; Muoth, M.; Seitsonen, A. P.; Saleh, M.; Feng, X. L.;
Mullen, K.; Fasel, R. Nature 2010, 466, 470.
(13) Fesser, P.; Iacovita, C.; Wackerlin, C.; Vijayaraghavan, S.; Ballav,
N.; Howes, K.; Gisselbrecht, J. P.; Crobu, M.; Boudon, C.; Stohr, M.;
̈
Jung, T. A.; Diederich, F. Chem. Eur. J. 2011, 17, 5246.
(14) Kolb, H. C.; Finn, M. G.; Sharpless, K. B. Angew. Chem. Int. Ed.
2001, 40, 2004.
(15) Moses, J. E.; Moorhouse, A. D. Chem. Soc. Rev. 2007, 36, 1249.
(16) Voigt, N. V.; Torring, T.; Rotaru, A.; Jacobsen, M. F.; Ravnsbaek,
J. B.; Subramani, R.; Mamdouh, W.; Kjems, J.; Mokhir, A.; Besenbacher,
F.; Gothelf, K. V. Nat. Nanotechnol. 2010, 5, 200.
(17) Binder, W. H.; Sachsenhofer, R. Macromol. Rapid Commun. 2007,
28, 15.
(18) Kolb, H. C.; Sharpless, K. B. Drug Discovery Today 2003, 8, 1128.
(19) Rostovtsev, V. V.; Green, L. G.; Fokin, V. V.; Sharpless, K. B.
Angew. Chem. Int. Ed. 2002, 41, 2596.
In conclusion, we have demonstrated that a cycloaddition can
be performed between alkyne and azide reactants adsorbed on a
Cu(111) surface under UHV conditions at rt to produce the 1,4-
triazole click reaction product. The reaction has a low apparent
yield, because a sizable fraction of the deposited azide molecules
degrade on the surface. The reaction itself, however, proceeds
readily provided that intact reactants are available on the surface.
For the azide−alkyne cycloaddition reaction to be useful for on-
surface synthesis of organized covalently bonded surface
structures, conditions must be identified, e.g., molecular
structure, substituent groups, substrates, adsorption temper-
atures, where the azide is stable upon adsorption, yet maintains
sufficient reactivity.
(20) Tornoe, C. W.; Christensen, C.; Meldal, M. J. Org. Chem. 2002,
67, 3057.
(21) Himo, F.; Lovell, T.; Hilgraf, R.; Rostovtsev, V. V.; Noodleman,
L.; Sharpless, K. B.; Fokin, V. V. J. Am. Chem. Soc. 2005, 127, 210.
(22) Prins, L. J.; Scrimin, P. Angew. Chem. Int. Ed. 2009, 48, 2288.
(23) Das, M. R.; Wang, M.; Szunerits, S.; Gengembre, L.;
Boukherroub, R. Chem. Commun. 2009, 2753.
(24) Watson, M. A.; Lyskawa, J.; Zobrist, C.; Fournier, D.; Jimenez, M.;
Traisnel, M.; Gengembre, L.; Woisel, P. Langmuir 2010, 26, 15920.
(25) Stoddart, J. F.; Paxton, W. F.; Spruell, J. M. J. Am. Chem. Soc. 2009,
131, 6692.
(26) Bohringer, M.; Morgenstern, K.; Schneider, W. D.; Berndt, R.;
̈
ASSOCIATED CONTENT
■
Mauri, F.; De Vita, A.; Car, R. Phys. Rev. Lett. 1999, 83, 324.
(27) Iucci, G.; Carravetta, V.; Altamura, P.; Russo, M. V.; Paolucci, G.;
Goldoni, A.; Polzonetti, G. Chem. Phys. 2004, 302, 43.
(28) Sohn, Y.; Wei, W.; White, J. M. J. Phys. Chem. C 2008, 112, 18531.
(29) Ciampi, S.; Bocking, T.; Kilian, K. A.; James, M.; Harper, J. B.;
Gooding, J. J. Langmuir 2007, 23, 9320.
(30) Collman, J. P.; Devaraj, N. K.; Eberspacher, T. P. A.; Chidsey, C.
E. D. Langmuir 2006, 22, 2457.
(31) Guan, B.; Ciampi, S.; Le Saux, G.; Gaus, K.; Reece, P. J.; Gooding,
S
* Supporting Information
Experimental details and characterization data. This material is
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
trolle@inano.au.dk; kvg@chem.au.dk
J. J. Langmuir 2011, 27, 328.
Notes
(32) Kammler, T.; Kuppers, J. J. Chem. Phys. 1999, 111, 8115.
̈
The authors declare no competing financial interest.
2139
dx.doi.org/10.1021/ja312303a | J. Am. Chem. Soc. 2013, 135, 2136−2139