Systematically Tuning the Electronic Structure of Gold Nanoclusters through Ligand Derivatization

Article abstract of DOI:10.1002/anie.201907586

While the ability to crystallize metal nanoclusters has revealed their geometric structure, the lack of a similarly precise measure of their electronic structure has hampered the development of synthetic design rules to precisely engineer their electronic properties. We track the evolution of highly-resolved electronic absorption spectra of gold nanoclusters with precisely mass-selected chemical composition in a controlled environment. Simple derivatization of the ligands yields larger spectral changes than varying the overall atomic composition of the cluster for two clusters with similar symmetry and size. The nominally metal-localized HOMO–LUMO transition of these nanoclusters lowers in energy linearly with increasing electron donation from the exterior of the ligand shell for both cluster sizes. Very weak surface interactions, such as binding of He or N₂, yield significant state-dependent shifts, identifying states with significant interfacial character. These observations demonstrate a pathway for deliberate tuning of interfacial chemistry for chemical and technological applications.

Full text of DOI:10.1002/anie.201907586

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: Systematically tuning the electronic structure of gold nanoclusters
through ligand derivatization
Authors: Anthony Cirri, Hanna Morales Hernandez, Christina Kmiotek,
and Christopher J. Johnson
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.201907586
Angew. Chem. 10.1002/ange.201907586
Link to VoR: http://dx.doi.org/10.1002/anie.201907586
http://dx.doi.org/10.1002/ange.201907586

10.1002/anie.201907586
Angewandte Chemie International Edition
COMMUNICATION
Systematically tuning the electronic structure of gold
nanoclusters through ligand derivatization
Anthony Cirri, Hanna Morales Hernández, Christina Kmiotek, and Christopher J. Johnson*
Abstract: While the ability to crystallize metal nanoclusters has
revealed their geometric structure, the lack of a similarly precise
measure of their electronic structure has hampered the development
of synthetic design rules to precisely engineer their electronic
properties. We track the evolution of highly-resolved electronic
absorption spectra of gold nanoclusters with precisely mass-selected
ease of preparation, small cluster size (≤1 nm), and ease of ligand
derivatization.[21,24] We prepared a set of clusters of varying size and
ligand composition using standard wet synthetic protocols.
Electrospray ionization mass-spectrometry shows that clusters of the
form Au8(PPh3)2+7 and Au9(PPh3)3+8 are both readily obtained from
the same reaction mixture (Figure S2A–F). Spectra of mass-
selected clusters are acquired by physisorbing weakly-bound He
atoms (∼1 meV/atom adsorption energy) to the cluster at cryogenic
temperatures and observing the desorption of tags as a function of
wavelength upon the absorption of one photon (see SI for details).
chemical composition in
a
controlled environment. Simple
derivatization of the ligands yields larger spectral changes than
varying the overall atomic composition of the cluster for two clusters
with similar symmetry and size. The nominally metal-localized
HOMO-LUMO transition of these nanoclusters lowers in energy
linearly with increasing electron donation from the exterior of the
ligand shell for both cluster sizes. Very weak surface interactions,
such as binding of He or N2, yield significant state-dependent shifts,
identifying states with significant interfacial character. These
observations demonstrate
a pathway for deliberate tuning of
interfacial chemistry for chemical and technological applications.
Atomically-precise gold nanoparticles are valued for their
potential as photocatalysts, photovoltaics, chemical sensors,
non-linear optical media, etc., and our fundamental
understanding of the chemical/physical properties relevant to
such applications has undergone a renaissance over the past
decade due to advances in synthetic protocols selective for their
isolation and crystallization.[1-16] Crystallographic characterization
of ligand-protected gold nanoparticles has revealed high-
symmetry core structures protected by the so-called ‘staple
motif’ for thiols or, in the case of phosphines, halides, or
hydrocarbons, direct metal core-ligand binding.[14] Such detailed
geometric characterization has been pursued due to its potential
to establish structure/function relationships in the synthesis of
‘designer’ nanomaterials.[6]
Electronic structure plays a similarly critical role in determining
the properties and reactivity of nanoparticles. To date, however, a
well-defined parallel between nanocluster geometric and electronic
structure has yet to emerge, particularly at the particle-environment
interface.[1] The development of a framework for predicting electronic
properties of nanoparticles, analogous to that established for organic
molecules or coordination complexes, could be realized via a similar
approach of careful chemical derivatization and precise
measurement.[17] However, the need for atomically precise syntheses
or separation methods limits the surface chemistries that can be
explored, and experimental methods sensitive to electronic structure
typically do not provide the necessary detail to identify trends or
benchmark quantum chemical studies.[14,18,19] The combination of
anisotropic solvent environments, thermal averaging, impurities, and
ligand dissociation or core isomerization obfuscate quantitative
comparisons between species or with quantum chemical predictions,
hampering efforts to develop a cohesive framework to rationalize
electronic structure and necessitating a new approach.[8, 14, 20-22]
In order to overcome such challenges, we turned to cryogenic
ion trap photofragmentation spectroscopy, a measurement which
combines the selectivity of mass-spectrometry with the spectral
detail of low-temperature UV-Vis spectroscopy to provide a probe of
electronic structure approaching the physical limit of spectroscopic
resolving power.[23] Triphenylphosphine-protected gold nanoclusters
serve as a model system given their narrow synthetic size dispersity,
Figure 1. (A) Crystallographically-resolved structures of Au9(PPh3)3+8 (left) and
Au8(PPh3)2+ clusters (right). Yellow and orange spheres depict Au and P,
7
respectively. The AuPPh3 unit highlighted in red is essentially cleaved from
Au9(PPh3)3+ to give Au8(PPh3)2+
. (B) Mass-selective UV-Vis absorption
8
7
spectra for the compositionally-related Au9(PPh3)3+8 and Au8(PPh3)2+7 clusters
(blue and red, respectively). For comparison, the black trace depicts the
solution-phase UV-Vis spectrum for a chromatographically-purified sample of
Au-phosphine clusters.
The geometric structures of both Au9(PPh3)3+ and
8
Au8(PPh3)2+7 have been crystallographically resolved, revealing that
the metal core of Au9(PPh3)3+8 takes on nearly D2h-symmetry while
that of Au8(PPh3)2+7 is similar but lowered to Cs, as shown in Figure
1A.[25,26] These structures reveal that the two clusters are related by
cleavage of a single Au(I)PPh3 unit, highlighted in red. Both of these
clusters are isovalent, within the superatomic model, such that the
number of possible electronic transitions making up their electronic
absorption spectra is expected to be conserved.[27] Figure 1B
[*]
Dr. A. Cirri, H. Morales Hernández, C. Kmiotek, Prof. C.J. Johnson
Department of Chemistry
Stony Brook University
100 Nicolls Rd., Stony Brook, NY 11794
E-mail: chris.johnson@stonybrook.edu
Supporting information for this article is given via a link at the end of
the document.
compares
the
solution-phase
UV-Vis
spectrum
of
a
This article is protected by copyright. All rights reserved.

10.1002/anie.201907586
Angewandte Chemie International Edition
COMMUNICATION
chromatographically purified sample from the reaction mixture to the
photofragmentation spectra of clusters mass-selected from the same
reaction mixture. The solution-phase spectrum reveals five partially
resolvable transitions, centered at 291, 316, 344, 416, and 479 nm.
Fourteen bands emerge for each cluster in the substantially higher
resolution photofragmentation spectra, with key transitions in the
lowest energy region including bands centered at 505 nm and 601
verified this assignment by measuring the UV spectrum for
Au(PPh3)+ , where the onset of ligand-centered π→π* occurs at
2
∼4.50 eV (see SI).
Chemical derivatization is routinely employed to fine tune the
properties of molecules and coordination complexes, though its use
in nanoscience is still emerging.[30] The substituent in the para-
position of the ligand phenyl group can be systematically tuned to
control the electron-donating or -withdrawing strength of the ligand.
The Hammett parameter (σP) is a quantitative parameterization of
this chemical property, which has been shown to correlate with
physical phenomena such as conductivity, acid/base chemistry,
electric field effects, and crystal-field splitting of monometallic
transition metal complexes.[31] Figures 2A–C compares the
triphenylphosphine, tris(4-methylphenyl)phosphine, and tris(4-
methoxyphenyl)phosphine ligated derivatives of the Au8(PPh3)2+7 and
nm for Au9(PPh3)3+
. The putative HOMO-LUMO transition is
8
identically centered at 601 nm for each spectrum, albeit with an
approximate order of magnitude difference in intensity likely due to
symmetry-based selection rules. At shorter wavelengths, a close
correspondence exists between the major features in these two
spectra, suggesting that they share striking similarities in electronic
structure despite their different compositions.
While comparison to computational results for Au9(PPh3)3+8 is
not quantitative, they allow us to generalize the origin of our
measured transitions.[28,29] Figure 2A compares the spectra given in
Figure 1 in units of eV. The broken lines of Figure 2A identify
electronic transitions between Au9(PPh3)3+8 and Au8(PPh3)2+7 that we
find to be conserved. The four lowest energy transitions between 2.0
and 2.8 eV are predicted to arise from transitions between
delocalized gold-dominant states. Transitions between 2.8 eV and
4.4 eV are nominally metal-to-ligand charge transfer (MLCT) bands
involving promotion of an electron from a metal-core orbital to an
orbital localized on the ligand. These are best classified as
originating either from the delocalized Au 6s- or the localized Au 5d-
states to the phenyl π* unoccupied orbitals. Above 4.4 eV, we
observe a broad absorption feature in each spectrum that we assign
to intra- and interligand π →π* transitions. We have independently
Au9(PPh3)3+ nanoclusters. These spectra are also presented
8
overlaid in the Supplementary Information for ease of comparison.
The electron-donating strength of the substituent goes as –H <–Me
<–OMe.
We observe substantial changes in the electronic absorption
spectrum as the substituent is changed for both the Au9(P(p-X-
Ph)3)3+
-
8
and Au8(P(p-X-Ph)3)2+7-series of clusters. However,
comparison between the two cluster sizes for each substituent
shows that these changes are largely conserved, suggesting that the
two clusters react similarly to derivatization. The change in electron
donating strength of the substituent introduces quantitative,
systematic changes to the electronic absorption spectrum. Focusing
on the HOMO-LUMO transition (Figure 2D, marked by arrows), we
find that its energy undergoes a monotonic red-shift as the electron-
Figure 2. Effect of ligand chemistry over the electronic structure of the Au9(P(p-X-Ph)3)3+ and Au8(P(p-X-Ph)3)2+ clusters (blue and red,
8
7
respectively), where X is (A) -H, (B) -CH3, or (C) -OCH3. The Lewis structure of the ligand is given for reference. (D) Enlarged view of the lowest
energy electronic transitions measured for the six clusters. Arrows denote the central position of the lowest energy transition, corresponding to
the cluster ‘bandgap’ or HOMO-LUMO transition energy. (E) Linearly fit correlation between the HOMO-LUMO transition energy and the para-
substituent Hammett parameter (σP) for Au8(P(p-X-Ph)3)2+7 and Au9(P(p-X-Ph)3)3+8. (F) Crystallographically-resolved metal core structures for the
Au9 series of phosphine-protected clusters.[25,34,35]
This article is protected by copyright. All rights reserved.

10.1002/anie.201907586
Angewandte Chemie International Edition
COMMUNICATION
donating strength of the ligand increases. A strong linear correlation
between the center of this peak and σP, plotted in Figure 2E, occurs
for both cluster sizes. The origin of this shift can be linked to a
ligand-induced change in charge distribution of the cluster. Quantum
chemical simulations of select nanoclusters have suggested that
increased electron density on the metal core should lead to a red
shift of low-energy electronic transitions, while electron spin
resonance studies have demonstrated that electron-donating ligands
increase the charge density on the metal core of nanoparticles.[18,32]
Though both clusters show the same trend, the slope of the fit for the
Au9(PPh3)3+ cluster (0.84) is greater than that of the Au8(PPh3)2+
cluster (0.36). Given the correlation with electron donating strength
of the ligand, we speculate that the increased charge density of the
3+ core enhances its ability to accept charge from the ligand shell,
increasing the magnitude of the shift of the HOMO-LUMO transition.
Taken together, these results demonstrate that the electronic
structure of the cluster core can be systematically and quantitatively
controlled through the electronic properties of ligand, and that this
approach can provide a similar level of control over electronic
structure as does changing the core composition.
electronic structure, at least for the HOMO-LUMO gap. While we
particularly focus on the HOMO-LUMO transition, as these are often
the most chemically- or technologically-relevant features of the
electronic structure, not all bands appear to exhibit such strong
linear correlations with σP. This likely reflects the electronic structure
of the ligand itself, particularly for transitions between states that
feature substantial orbital contribution from the substituent. Thus, the
ligands provide two modes of control of electronic structure: (1)
directly via manipulation of the energies of the ligand orbitals
themselves, or (2) indirectly via redistribution of charge density in the
cluster. Our results indicate that the former dominates the MLCT
transitions, while the latter controls transitions local to the metal core.
We hypothesized that species with even weaker binding could
serve as probes for the interfacial electronic structure. Transitions
involving similar orbitals should show similar spectral shifts, allowing
us to positively identify related transitions between the two clusters,
and the magnitude of the shift is expected to serve as an indirect
measure of the interfacial nature of the involved electronic states.
We thus collected spectra substituting N2 for He as the mass-action
tag for the triphenylphosphine-protected clusters, as shown in Figure
3. Both spectra respond similarly to the identity of the probe,
undergoing up to 50 meV shifts in the maxima of peaks, confirming
the similarity between their electronic structure. The preponderance
of low energy peaks tend to blue-shift with the introduction of N2,
indicating that adsorption by N2 notably reduces charge density on
the metal core, likely through bonding between occupied gold and
unoccupied N2 orbitals. Perhaps coincidentally, the magnitude of
solvatochromism introduced by exchanging He for N2 is comparable
to that measured for the plasmon resonance of gold nanoparticles
dispersed in organic solvents (e.g., cyclohexane to toluene yields a
∼50 meV peak shift).[36] Particularly large shifts are seen in the
bands labeled ‘C,’ occurring in the range of 2.50 – 2.80 eV for
Au9(PPh3)3+ and Au8(PPh3)2+ (highlighted green in Figure 3),
8
7
8
7
indicating that at least one orbital involved in both transitions
displays significant interfacial character. Other smaller, but still
notable, changes are highlighted as well.
These results emphasize that the ligand is more than just a
means to stabilize the metal core; it exerts a cluster-spanning
influence over electronic structure, even for orbitals with little direct
probability density on the ligand. Ligand chemistry can be at least as
influential towards electronic structure as atomic composition for
nanoclusters with similar size and symmetry. The primary difference
between these two synthetic controls is that ligand properties can be
nearly continuously tuned using established chemical reasoning,
whereas the effects of composition are discrete and, particularly for
atomically-precise magic clusters, difficult to reliably vary.
Furthermore, even extremely weak surface interactions induce
perturbations that range up to one-sixth of the total impact of ligands
for certain states, suggesting that solvent and interfacial effects can
be used to fine tune electronic structure state-specifically and
selectively for states directly involved in interfacial processes. This
hierarchy forms a preliminary framework to aid rational nanocluster
design and, given the generality of the technique used in these
studies, can be extended to include steric effects, ligand binding
group effects, and alloy composition at a similarly detailed level
.
Figure 3. Effect of solvent environment over electronic absorption spectrum of
Au9(PPh3)3+8 and Au8(PPh3)2+7 using He- or N2-physisorption. The alphabetical
assignments reflect correlated spectral features.
The high coefficients of determination (>0.99) of these trends
indicate that the changes in Hammett parameter are sufficient to
explain the entire red shift of the HOMO-LUMO transition, and
alternative explanations cannot explain this trend. The nearly
identical cone angle of the three para-substituted triphenylphosphine
ligands, as well as the similar crystal structures of the protic- and
Conflicts of Interest
The authors declare no conflicts of interest.
methyl-Au9(P(p-X-Ph)3)3+ clusters, suggest that electronic rather
8
Acknowledgements
than geometric perturbations are responsible for the observed
differences.[25,33,34] Furthermore, the crystal structures (Figure 2F)
show a non-monotonic change in core structure from –H to –OMe
substituents, with the core structure of –H-substituted clusters
intermediate between –Me and –OMe, which does not comport with
the energetic ordering of HOMO-LUMO transitions observed (the
respective Au8(PPh3)2+7 derivatives have not been reported).[25, 34, 35]
The high degree of linearity of this trend, despite the geometric
changes induced by ligand derivatization, thus suggests that ligand
chemistry also dominates metal core geometry in determining
We thank C. M. Aikens, B. J. Lear, M. G. White, and M. A. Johnson
for helpful discussions, and Y. Yang and J. J. Kreinbihl for
assistance in developing data acquisition and analysis software. This
work was supported by the Air Force Office of Scientific Research
grant FA9550-17-1-0373 and Stony Brook University.
Keywords: Electronic structure • Gold nanocluster • Ligand
effects • Linear free energy relationships • UV/Vis spectroscopy
This article is protected by copyright. All rights reserved.

10.1002/anie.201907586
Angewandte Chemie International Edition
COMMUNICATION
[19] A. Dass, S. Theivendran, P.R. Nimmala, C. Kumara, V.R. Jupally, A.
Fortunelli, L. Sementa, G. Barcaro, X. Zuo, B.C. Noll, J. Am. Chem.
Soc. 2015, 137, 4610–4613.
[1] a) M.A. Abbas, P.V. Kamat, J.H. Bang, ACS Energy Lett. 2018, 3, 840–
854; b) M. Girod, M. Kristić, R. Antoine, L. MacAleese, J. Lemoine, A.
Zavras, G.N. Khairallah, V. Bonačić-Koutecký, P. Dugourd, R.A.J.
O’Hair, Chem. Eur. J. 2014, 20, 16626—16633.
[2] K.G. Stamplecoskie, G. Yousefalizadeh, L. Gozdzialski, H. Ramsay, J.
Phys. Chem. C 2018, 122, 13738–13744.
[3] S. Knoppe, H. Häkkinen, T. Verbiest, K. Clays, J. Phys. Chem. C 2018, 122,
4019–4028.
[4] N.A. Sakthivel, M. Stener, L. Sementa, A. Fortunelli, G. Ramakrishna, A.
Dass, J. Phys. Chem. Lett. 2018, 9, 1295–1300.
[5] T. Higaki, M. Zhou, K.J. Lambright, K. Kirschbaum, M.Y. Sfeir, R. Jin, J. Am.
Chem. Soc. 2018, 140, 5691–5695.
[6] L. Sumner, N.A. Sakthivel, H. Schrock, K. Artyushkova, A. Dass, S.
Chakraborty, J. Phys. Chem. C 2018, 122, 24809–24817.
[7] C. Yi, H. Zheng, L.M. Tvedte, C.J. Ackerson, K.L. Knappenberger Jr, J.
Phys. Chem. C 2015, 119, 6307–6313.
[8] K. Konishi, M. Iwasaki, Y. Shichibu, Acc. Chem. Res. 2018, 51, 3125–3133.
[9] X.-K. Wan, J.-Q. Wang, Z.-A. Nan, Q.-M. Wang, Sci. Adv. 2017, 3,
e1701823.
[10] Q. Li, T.-Y. Luo, M.G. Taylor, S. Wang, X. Zhu, Y. Song, G. Mpourmpakis,
N.L. Rosi, R. Jin, Sci. Adv. 2017, 3, e1603193.
[20] W. R. Mason, Inorg. Chem. 2000, 39, 370–374.
[21] D.M.P. Mingos, Philos. Trans. Royal Soc. A 1982, 308, 75–83.
[22] T. Huang, L. Huang, Y. Jiang, F. Hu, Z. Sun, G. Pan, S. Wei, Dalton Trans.
2017, 46, 12239–12244.
[23] E.K. Campbell, M. Holz, D. Gerlich, J.P. Maier, Nature 2015, 523, 322–
323.
[24] a) G.E. Johnson, A. Olivares, J. Laskin, Phys. Chem. Chem. Phys. 2015,
17, 14636–14646; b) G.E. Johnson, J. Laskin, Analyst 2016, 141,
3573—3589; c) K.A. Parish, M. King, M.R. Ligare, G.E. Johnson, H.
Hernández, Phys. Chem. Chem. Phys. 2019, 21, 1689—1699.
[25] F. Wen, U. Englert, B. Gutrath, U. Simon, Eur. J. Inorg. Chem. 2008, 106–
111.
[26] J. W. van der Velden, J.J. Bour, W.P. Bosman, J.H. Noordik, J. Chem.
Soc., Chem. Commun. 1981, 1218–1219.
[27] M. Walter, J. Akola, O. Lopez-Acevedo, P.D. Jadzinsky, G. Calero, C.J.
Ackerson, R.L. Whetten, H. Grönbeck, H. Häkkinen, Proc. Natl. Acad.
Sci. U.S.A. 2008, 105, 9157–9162.
[28] N.V. Karimova, C.M. Aikens, J. Phys. Chem. C 2017, 121, 19478–19489.
[29] N.V. Karimova, C.M. Aikens, J. Phys. Chem. A 2016, 120, 9625–9635.
[30] Z. Cao, D. Kim, D. Hong, Y. Yu, J. Xu, S. Lin, X. Wen, E.M. Nichols, K.
Jeong, J.A. Reimer, P. Yang, C.J. Chang, J. Am. Chem. Soc. 2016,
138, 8120–8125.
[31] C. Hansch, A. Leo, R.W. Taft, Chem. Rev. 1991, 91, 313–348.
[32] A. Cirri, A. Silakov, L. Jensen, B.J. Lear, Phys. Chem. Chem. Phys. 2016,
18, 25443–25451.
[11] P.D. Jadzinsky, G. Calero, C.J. Ackerson, D.A. Bushnell, R.D. Kornberg,
Science 2007, 318, 430–433.
[12] Q. Yao, T. Chen, X. Yuan, J. Xie, Acc. Chem. Res. 2018, 51, 1338–1348.
[13] I. Chakraborty, T. Pradeep, Chem. Rev. 2017, 117, 8208–8271.
[14] R. Jin, C. Zeng, M. Zhou, Y. Chen, Chem. Rev. 2016, 116, 10346– 10413.
[15] N. Yan, N. Xia, L. Liao, M. Zhu, F. Jin, R. Jin, Z. Wu, Sci. Adv. 2018, 4,
eaat7259.
[33] C. Tolman, Chem. Rev. 1977, 77, 165–195.
[34] P. Bellon, F. Cariati, M. Manassero, L. Naldini, M. Sansoni, J. Chem. Soc.
D 1971, 1423–1424.
[35] K.P. Hall, B.R.C. Thobald, D.I. Gilmour, D.M.P. Mingos, A.J. Welch, J.
Chem. Soc., Chem. Commun. 1982, 528–530.
[36] S.K. Ghosh, S. Nath, S. Kundu, K. Esumi, T. Pal, J. Phys. Chem. B 2004,
108, 13963–13971.
[16] C. Zeng, Y. Chen, K. Kirschbaum, K.J. Lambright, R. Jin, Science 2016,
354, 1580– 1584.
[17] R.H. Crabtree, The Organometallic Chemistry of the Transition Metals, 6th
Ed., Wiley, Hoboken, NJ, 2014.
[18] G. Lugo, V. Schwanen, B. Fresch, F. Remacle, J. Phys. Chem. C 2015,
119, 10969–10980.
This article is protected by copyright. All rights reserved.

10.1002/anie.201907586
Angewandte Chemie International Edition
COMMUNICATION
COMMUNICATION
Atomically-precise gold nanoclusters
are known for their diverse geometric
structures, yet comparatively little has
been resolved regarding their
Anthony
Hernández,
Christopher J. Johnson
Cirri,
Hanna
Morales
Kmiotek,
Christina
Page No. – Page No.
electronic structure. Using mass-
selective UV/Vis spectroscopy and
ligand derivatization, we show that the
cluster electronic structure may be
systematically tuned via precise
Systematically
nanocluster
tuning
electronic
gold
structure
through ligand derivatization
chemical control over the interfacial
chemi- and physisorptive interactions.
This article is protected by copyright. All rights reserved.