REVIEW ARTICLE PUBLISHED ONLINE: 22 JANUARY 2016 | DOI: 10.1038/NMAT4526

The surface science of nanocrystals Michael A. Boles1, Daishun Ling2,3,4, Taeghwan Hyeon2,3 and Dmitri V. Talapin1,5* All nanomaterials share a common feature of large surface-to-volume ratio, making their surfaces the dominant player in many physical and chemical processes. Surface ligands — molecules that bind to the surface — are an essential component of nanomaterial synthesis, processing and application. Understanding the structure and properties of nanoscale interfaces requires an intricate mix of concepts and techniques borrowed from surface science and coordination chemistry. Our Review elaborates these connections and discusses the bonding, electronic structure and chemical transformations at nanomaterial surfaces. We specifically focus on the role of surface ligands in tuning and rationally designing properties of functional nanomaterials. Given their importance for biomedical (imaging, diagnostics and therapeutics) and optoelectronic (light-emitting devices, transistors, solar cells) applications, we end with an assessment of application-targeted surface engineering.

A

bulk solid contains only a small concentration of surface atoms; as a result, broken chemical bonds on the exterior contribute minimally to material properties. For any substance, however, the surface-to-volume ratio scales inversely with linear dimensions. With shrinking size, the role of the surface increases, eventually becoming dominant. At the nanoscale, surfaces can significantly alter some properties (for example solubility or luminescence1) and generate completely new effects (such as surface plasmon resonance2 or size-dependent catalytic activity 3). This Review shines a spotlight on the surface of nanomaterials and discusses various strategies to tame and make use of it. For the sake of consistency, we focus our discussion on nanocrystals (NCs), but most concepts can be applied to onedimensional nanowires, two-dimensional nanoplatelets and other nanoscale objects. Traditional surface science has established that the surfaces of large crystals can lower their energy by moving surface atoms away from lattice sites in the process of surface reconstruction4, dangling bonds can introduce new electronic states5, and foreign molecules (surfactants or adsorbates) can alter the energy and reactivity of a crystal surface6. All of these effects apply to NCs, although the small facet size and multiple edge- and corner sites complicate analysis and quantitative description. On the other hand, the chemical bond between a NC surface atom and surfactant molecule is similar to that between a metal ion and ligand in a coordination complex, offering a useful analogy between NCs and molecular compounds. We will use the term ‘surface ligands’ here to emphasize this surfactant–ligand duality. The set of ligands attached to a NC forms a ‘capping’ layer that saturates dangling bonds, screens the particle from its environment, and controls nucleation and growth kinetics during synthesis7. Ligands also influence the optical and electronic properties of NCs, and provide steric or electrostatic stabilization (Box 1) of the colloidal state required for NC synthesis, processing and some applications. Ligand exchange reactions extend the versatility of NC materials by allowing replacement of ligands optimized for synthesis with application-targeted species including organic or inorganic ions8,9, clusters10 and polymers11.

Structure and bonding at the nanocrystal–ligand interface

Nanocrystals consist of hundreds to thousands of atoms. Such particles are typically synthesized in a solution containing surface ligands with an anchoring headgroup tethered to the NC surface and a hydrocarbon tail directed away from it (Fig.  1a). The equilibrium shape of the inorganic core minimizes the energy of exposed surface area and facet-specific energy of broken bonds. A poly­ hedral core, displaying only high-coordination surface atoms and slightly more exposed area than a sphere, is typical. For example, Au and PbS NCs often adopt a cuboctahedral shape, terminated by (111) and (100) facets with the hexagonal and square arrangements of surface atoms shown in Fig. 1b. If surface ligands selectively bind to certain facets of a growing NC, they reduce the surface energy of these facets relative to others. The ligand layer can also block delivery of new reagents to the NC surface. These thermodynamic and kinetic factors are widely used for synthesis of NCs with anisotropic shapes such as rods12 and platelets13. The capping layer protecting each NC facet can be viewed as a miniature self-assembled monolayer (SAM). First prepared and characterized in the 1980s14, SAMs have served as the foundation for understanding organic/inorganic interfaces and provide a convenient starting point to describe NC surface ligands. For example, the binding pattern of n-alkanethiolate on extended Au (111) and (001) surfaces15 can be used to create a first approximation of the capping layer protecting a cuboctahedral Au NC. The strong interaction between gold and sulphur atoms (~2 eV)16 drives free surfactants to bind tightly to the metal surface. A weaker van der Waals interaction between hydrocarbon tails (~0.07 eV per CH2 group16) encourages dense packing in the organic overlayer. On a flat surface, given time to adsorb and relax, a crystalline arrangement of surfactant molecules is formed in registry with the underlying substrate, with sulphurs typically occupying three-fold sites and alkyl tails tilted approximately 30° from the surface normal (Fig. 1b). SAM grafting density is limited by the steric bulk of alkyl tails: the organic layer fills space completely, whereas the sulphur atoms remain separated by ~3 van der Waals diameters16 (Fig. 1c). In addition to this ‘standard model’, alternative binding motifs including the RS–Au–SR ‘staple’ structure can be important at low surface coverage17.

University of Chicago and James Franck Institute, Chicago, Illinois 60637, USA. 2Center for Nanoparticle Research, Institute for Basic Science (IBS), Seoul 151-742, Korea. 3School of Chemical and Biological Engineering, Seoul National University, Seoul 151-742, Korea. 4Institute of Pharmaceutics, College of Pharmaceutical Sciences, Zhejiang University, 866 Yuhangtang Road, Hangzhou 310058, China. 5Center for Nanoscale Materials, Argonne National Laboratory, Argonne, Illinois 60439, USA. *e-mail: [email protected] 1

NATURE MATERIALS | VOL 15 | FEBRUARY 2016 | www.nature.com/naturematerials

© 2016 Macmillan Publishers Limited. All rights reserved

141

REVIEW ARTICLE

NATURE MATERIALS DOI: 10.1038/NMAT4526

Box 1 | Surface ligands and colloidal stabilization.

In contrast to extended flat surfaces, NC surfaces are encircled by vertex and edge sites, providing grafted chains with extra volume (Fig. 1a). The relaxed competition for space between alkyl tails minimizes the role of hydrocarbon steric bulk in determining the grafting density of surfactants on NCs. As a result, higher capping-layer coverage is possible: 3-nm Au NCs can support n-decanethiolate surface densities of 6 nm–2, as compared with 3 nm–2 on extended Au(111) (ref. 18). Open space in the NC capping layer allows penetration of solvent molecules or ligand chains of neighbouring NCs. This extra space also provides room for rotational conformations not available to molecules packed in a SAM, resulting in significant structural disorder: gauche defects are concentrated in the ends of alkyl chains and propagate towards the middle with increasing temperature19. Generally, capping-layer order is maximized for longer (C12–C18) chains tethered to the surface of larger-diameter (>5 nm) NCs20. Nanometre dimensions and intrinsic heterogeneity (each NC typically exposes several facets with different patterns of surface atoms) make experimental study of NC surfaces challenging. Currently there is no technique that provides atomic-level reconstruction of the NC capping layer. Instead, a suite of methods should be applied to obtain complementary bits of information about the NC–ligand 142

a

U(r) Repulsion

c

b

Good solvent

0

r Attraction

Surface ligands enable control over colloidal stability and aggregation of nanocrystals. When exposed to a good solvent for a given type of surface ligands, NCs disperse in solution. Reducing solvent quality, for example by changing polarity, induces flocculation of NCs. Two fundamentally different mechanisms of colloidal stabilization, steric and electrostatic, provide largely complementary approaches to stabilizing NC dispersions in polar or nonpolar solvents (panel a). For sterically stabilized particles (such as NCs coated with a layer of surface-tethered hydrocarbon chains, panel b, top), a good solvent implies a negative free energy of chain-solvent mixing 30. This causes the hydrocarbon chains to repel one another, penalizing the overlap of ligand coronas and stabilizing NC dispersions. On the other hand, when immersed in a non-solvent, with positive chain–solvent mixing energy, the drive to minimize contact with the surrounding liquid induces contraction of ligand chains and aggregation of dispersed NCs. Typical good solvents for hydrocarbon-capped NCs are nonpolar liquids (for example hexane, toluene, chloroform) whereas non-solvents are polar (for example ethanol, acetone, acetonitrile). The interaction between sterically stabilized NCs for various capping ligand lengths and solvent conditions has been investigated by X-ray scattering experiments119,120 and molecular dynamics simulations121. In general, pair potentials are purely repulsive for NCs in good solvent, whereas in non-solvent or vacuum, attractive interaction dominates at intermediate separation (panel c). Colloidal NCs may also be stabilized electrostatically through adsorption of charged species to their surface. In this case, the NC surface charge is balanced by oppositely charged counterions located in a diffuse region around the particle (panel b, bottom). Good solvents for electrostatically stabilized NCs are generally those with high dielectric constant, ε, for effective screening of electrostatic attraction between oppositely charged ions (for example formamide, ε ≈ 110). In such solvents, close approach of charge-stabilized NC surfaces is prevented by the entropic penalty associated with condensation of counterions in the diffuse double layer. Addition of poorly screening solvents (for example toluene, ε ≈ 2) induces collapse of the diffuse counterion cloud and flocculation of NCs. The interaction potential between a pair of charge-stabilized NCs, including both the doublelayer repulsion and van der Waals core–core attraction, can be treated using Derjaguin, Landau, Verwey and Overbeek (DLVO) theory 30.

Non-solvent or vacuum

Stabilization of colloidal nanocrystals. a, Depending on surface ligands, NCs can form colloidal dispersions in polar or nonpolar solvents. Oleatecapped CdSe NCs disperse in nonpolar phase (toluene; upper layer in vial shown at left), whereas exchange of oleate ligands for potassium sulphide results in the phase transfer of NCs into a polar solvent (formamide; bottom layer in vial shown at right). b, Sketch of interacting pair of NCs with hydrocarbon (top) and ionic (bottom) surface ligands. c, The interaction potential U with interparticle distance r for NCs in a good solvent and non-solvent. Panel a reproduced from ref. 9, American Chemical Society.

bonding, capping-layer structure, and interactions between surface ligands and the surrounding environment 21,22. Useful techniques are summarized in Box 2. Interaction between the NC core and ligand headgroup can be rationalized using the classification of covalent bonds23, originally proposed for metal coordination complexes and adapted to NCs by Owen and co-workers (Fig.  2a)24,25. Without going into full technical details26, three classes of metal–ligand interaction may be distinguished based on the number of electrons involved, and the identity of the electron donor and acceptor groups. L-type ligands are neutral two-electron donors with a lone electron pair that datively coordinates surface metal atoms. Amines (RNH2), phosphines (R3P) and phosphine oxides (R3PO) are examples of L-type ligands. X-type ligands are species that, in neutral form, have an odd number of valence-shell electrons, requiring one electron from the NC surface site to form a two-electron covalent bond. In practice, M–X bonds often cleave heterolytically, forming ionic, closedshell fragments. As such, X-type ligands can be neutral radicals binding neutral surface sites (each with an unpaired electron) or, more commonly, monovalent ions binding oppositely charged sites at the NC surface. Examples of X-type ligands include carboxylates (RCOO–), thiolates (RS–) and phosphonates (RPO(OH)O–), NATURE MATERIALS | VOL 15 | FEBRUARY 2016 | www.nature.com/naturematerials

© 2016 Macmillan Publishers Limited. All rights reserved

REVIEW ARTICLE

NATURE MATERIALS DOI: 10.1038/NMAT4526 a

b

(111)

(100) Au

S

C

H

c

~30°

PbS

Au

Figure 1 | Capping-layer structure. a, Calculated atomic structure of 5-nm diameter PbS NC capped with oleic acid. PbS(111) surfaces are terminated by oleate and hydroxide ions, PbS(100) by oleic acid. b, Binding pattern of sulphur headgroups (red spheres) on Au (111) and (100) facets15. c, Sketch of self-assembled monolayer (SAM) of hexadecanethiol molecules adsorbed to Au(111) surface as an idealized picture of a nanocrystal capping layer. Hydrocarbon chains are fully extended, tilted with respect to the surface normal, and in all-trans configuration. Panel a reproduced from ref. 31, AAAS.

as well as inorganic ions (such as Cl–, InCl4–, AsS33–) or bound ion pairs (for example NEt4+I–) in nonpolar solvent. Nucleophilic (electron-rich) L- and X-type ligands bind to electron-deficient (electrophilic) surface sites with pronounced Lewis acidity, typically under­coordinated metal ions at the NC surface. The surface of metal chalcogenides, oxides and other compound NCs also exposes electron-rich Lewis basic sites. These sites can interact with Z-type ligands, such as Pb(OOCR)2 or CdCl2, which bind through the metal atom as two-electron acceptors24. In addition, the surface of oxide NCs can bind protons (H+), an example of positively charged, electrophilic X-type ligands27. Following this classification, the composition and surface chemistry of NCs can be expressed in a convenient way. Cadmium selenide NCs, for example, capped by a combination of L- and X-type ligands, can be described as (CdSe)m(CdnXpLq), where m relates to the size of the NC core and n, p and q describe the ligand shell composition. Depending on the nature of exposed NC facets (polar or nonpolar), L- or X-type ligation can dominate. The measurement of metal-to-chalcogen ratio provides a simple way to access this information. CdSe and PbSe NCs synthesized in the presence of X-type ligands show metal-to-selenium ratios significantly exceeding unity 28,29. In nonpolar solvents such as hexane or toluene, a large energetic penalty for charge separation requires the ratio between L- and X-type ligands to satisfy electrostatic neutrality and fit the formula (CdSe)m(CdX2)nLq. The last expression is particularly useful for describing neutral NCs capped with one kind of X-type and one kind of L-type ligand. On the other hand, in polar solvents such as water or dimethylformamide, NCs can carry charge: [(CdSe)m(CdnX2n+sLq)]s– or [(CdSe)m(CdnX2n-sLq)]s+ compensated by counterions from the diffuse ion cloud around each NC in solution. The NC can support approximately one elemental charge per square nanometre of surface, or a few tens of charges per particle9,30. Such charging plays an important role in electrostatic stabilization of NC colloids (see Box 1). Recent computational studies have revealed some counter­ intuitive aspects of NC surfaces subsequently verified by experiments. For example, density functional theory (DFT) calculations indicate that oleic acid (OAH), commonly used for NC synthesis, binds to the (100) facet of a PbS NC (the surface presenting a

‘chequerboard’ arrangement of lead and sulphur atoms; Fig. 2b,c) as a bidentate L-type ligand with energy 0.16 eV per ligand. On the other hand, (111) facets of PbS NCs present a hexagonal layer of Pb atoms (Fig. 2b,d) and develop a very different motif, with X-type oleate ions (OA–) binding to surface Pb atoms more strongly at 0.52  eV per ligand31. The density of Pb atoms on PbS (111) surface (~8 Pb atoms nm–2), however, prevents sterically demanding oleate ligands (the footprint of COO– headgroup is ~0.3 nm2) from saturating all dangling bonds. Both DFT calculations and experimental X-ray photoelectron spectroscopy (XPS) studies reported in ref. 31 suggest that compact X-type ligands (for example OH–) bind to the PbS (111) facet together with bulky OA– ligands (Fig. 2d). Surface energy minimization suggests the following composition for 5-nm PbS NCs: (PbS)m[Pb(n+p)(OH)2n(OA)2p(OAH)q], with m = 1,385, n = 149, p = 120 and q = 48, where the ratio n:p may change depending on conditions (such as the water content in the synthesis solution). Existing experimental data support the computationally derived model of a cation-rich NC surface29. It is likely that the ‘undercoat’ of small X-type ligands (such as OH– or Cl–) is a common feature of most NC surfaces, incorporated (intentionally or not) through side reactions during NC synthesis.

Ligand exchange reactions

Surface ligands with L- or X-type headgroup and a hydrocarbon tail allow impressive control over the kinetics of NC nucleation and growth. In many cases, however, these ligands must be replaced by other surface-binding species better suited to the end application. The exchange of NC surface ligands is reminiscent of substitution reactions in coordination complexes. Solvent polarity and coordinating ability can affect the kinetics and mechanism of ligand exchange reactions at the NC surface. Typically, steric crowding of molecules in the capping layer favours a dissociative pathway that requires a bound ligand to desorb from the NC before a new one may enter from solution and attach to the surface. In nonpolar solvents, all species involved in the exchange reaction should be electrically neutral. For this reason, L-type ligands (for example octylamine32 on CdSe) rapidly adsorb and desorb from the NC surface at room temperature (exchanging species highlighted in bold):

NATURE MATERIALS | VOL 15 | FEBRUARY 2016 | www.nature.com/naturematerials

© 2016 Macmillan Publishers Limited. All rights reserved

143

REVIEW ARTICLE

NATURE MATERIALS DOI: 10.1038/NMAT4526

Box 2 | Characterization of the nanocrystal surface.

A suite of spectroscopic, scattering, imaging and computational techniques has been used to characterize NC surfaces (panel a). These methods provide complementary pieces of information about the interface that can be divided into the following components: (i)  chemical bonds between the NC core and surface ligands, (ii) the composition and structure of ligand molecules and (iii) the effective properties of the whole capping layer such as effective thickness, density and dielectric constant (panel a). Vibrational spectroscopy: Fourier transform infrared spectroscopy (FTIR) is a simple and reliable way to probe the NC surface, providing information on the structure of ligand molecules. NCs capped with organic ligands such as oleic acid (OAH, CH3(CH2)7CH=CH(CH2)7COOH; see also Fig.  1a) show strong infrared absorption bands around 3,000  cm–1 and 1,500  cm–1, corresponding to the C–H stretching and bending modes of surface-bound hydrocarbon molecules18 (panel b, top). The width and position of infrared resonances can be used to investigate chain conformations and molecular order of the capping layer 19. FTIR also allows for probing the displacement of hydrocarbon ligands (for example the exchange of organic for inorganic species; panel b, bottom). Raman spectroscopy is particularly useful for characterizing inorganic ligands with heavy atoms vibrating at low frequencies10,22. Nuclear magnetic resonance (NMR) spectroscopy is used for characterizing the capping layer of NCs following NMR finger­prints of spin-active nuclei (1H, 13C, 31P and others)122. The NMR peaks of surface-tethered ligands show significant broadening caused by dipolar coupling effects which cancel out for a

fast-tumbling free ligand molecules in solution (panel c). This broadening limits the utility of traditional one-dimensional NMR for NC surface analysis. Diffusion-ordered spectroscopy (DOSY) links a diffusion coefficient, D, to each resonance (panel d) and allows the separation of signals from surface-bound and free ligand molecules. Heavy-nuclear NMR (for example 119Sn) has proven useful in studies of NCs with inorganic ligands such as SnS44– (refs 22,96). Small-angle X-ray scattering (SAXS) and dynamic light scattering (DLS) have been used to study effective thickness of the NC capping layer and its contribution to overall particle hydrodynamic radius1,8,119. Transmission electron (TEM) and scanning probe (STM, AFM) microscopy allow real-space imaging of NCs, and in some cases, their surface ligands (panel e). A key advantage of microscopy techniques is the ability to examine NCs at the single-­particle level, revealing aspects of these systems otherwise hidden in ensemble averages. Density functional theory (DFT) and molecular dynamics (MD) can, at least partially, compensate for difficulties associated with experimental probing of NC surfaces. DFT is used to model electronic interactions between different NC facets and the molecules that bind to them (panel f). Recent studies have used combinations of ab initio electronic structure and MD calculations31,123 to reconstruct an entire NC, capturing the structure of inorganic core and complete ligand shell (Fig. 1a). These and other computational techniques aid interpretation of experimental data and propose new directions for experimental work.

c

Ligand molecule

NMR FTIR TEM

NC/ligand interface

Bound ligand

CdSe–OA

9.0 Solvent (toluene) 5.8 5.4 Chemical shift (ppm)

DFT

e Transmittance (arbitrary units)

10.0

Free ligand

XPS

b

d

8

6 4 2 Chemical shift (ppm)

−log[D (m2 s−1)]

Entire DLS particle AFM

Intensity (arbitrary units)

SAXS

8.0

f

CdSe - OA CdSe–OA

3

CdSe - AsS3 3− CdSe–AsS 3

3,000 2,000 1,000 Wavenumbers (cm−1)

5 nm

Characterization of the nanocrystal surface. a, Sketch summarizing useful techniques for analysis of the NC/ligand interface, grouped by the region probed. Black and yellow spheres represent nanocrystal surface metal and chalcogen atoms, respectively, and red spheres represent ligand headgroup atoms. b, FTIR spectra showing vibrational signatures of organic oleate (OA, top) and inorganic (NH4)3AsS3 (bottom) ligands bound to the surface of CdSe NCs. c, Proton (1H) NMR spectrum of CdSe NCs with OA ligands in toluene showing vinyl resonances of bound (blue) and a mixture of bound and free (red) ligands. d, DOSY-NMR spectrum of OA-capped CdSe NCs in toluene solution containing excess ligand. e, TEM image of dodecene-capped silicon NCs imaged at the edge of a graphene support, showing the molecular structure of surface ligands. f, PbS(111) surface reconstructed using DFT and MD modelling 31. Large black and yellow spheres represent nanocrystal surface metal and chalcogen atoms, respectively, and small black, teal and red spheres represent ligand carbon, hydrogen and oxygen atoms. Figure reproduced from: c,d, ref. 33, American Chemical Society; e, ref. 124, American Chemical Society; f, ref. 31, AAAS.

144

NATURE MATERIALS | VOL 15 | FEBRUARY 2016 | www.nature.com/naturematerials

© 2016 Macmillan Publishers Limited. All rights reserved

REVIEW ARTICLE

NATURE MATERIALS DOI: 10.1038/NMAT4526 a

b

X Cd (3-coord)

X-type Terminates lattice

X X Cd (2-coord)

Bound ion pair

RCOO− RS− Cl− HO− AsS33− R4N+··············· I−

L Cd (2- or 3-coord) Z (MX2) Se (3-coord) Cd

c

Se

RNH2 R3P R3PO

L-type Neutral donor

CO H O

CO H O

Pb

Pb

S

H H C C O O O O O O

S

Pb

Pb

Pb

Pb

Pb(RCOO)2 Z-type CdCl2 Neutral acceptor

d

PbS (001)

PbS (111)

Pb S C O H

Figure 2 | Chemistry of ligand binding. a, Classification of ligand binding motifs at the surface of a CdSe nanocrystal24. Anionic X-type ligands bind to surface cations imparting electroneutrality to metal-rich NC facets. Neutral L-type electron donor ligands attach to electrophilic metal sites on stoichiometric NC facets. Neutral Z-type electron acceptor ligands bind to electron-rich undercoordinated Se atoms. Green arrows highlight examples of such atoms at the surface of a CdSe nanocrystal model. b, Modelled 5-nm cuboctahedral PbS NC (centre) with simplified illustrations of oleic acid binding as L-type ligand to (100) facets (left) and binding together with hydroxide as X-type oleate on Pb-terminated (111) facets (right)31. c,d, Sketches of DFT-relaxed surface structure on PbS(100) and PbS(111) (refs 99, 117). Panel b reproduced from ref. 117, AAAS.

(CdSe)m(CdX2)nLq + Lʹ → (CdSe)m(CdX2)nLq–1Lʹ + L

(1)

L, Lʹ = RNH2, R3P, R3PO, RCOOH or pyridine, for example. However, X-type ligands (for example oleate33 and phosphonate34 on CdSe) remain tightly bound because of the electrostatic penalty for charging induced by self-desorption of X-type ligands. Similarly, neutral Z-type ligands (for example cadmium oleate) can be displaced by other metal complexes35: (CdSe)m(CdX2)nLq + Zʹ → (CdSe)m(CdX2)n–1ZʹLq + CdX2

(2)

Zʹ = Cd(RCOO)2, Cd(RPO(OH)O)2, CdCl2 or AlCl3, for example. On the other hand, exchange of tightly bound X-type ligands in nonpolar solution probably takes place by an associative pathway 36. Replacement of these charged ligands can occur by cation transfer, whereby incoming and outgoing species exchange a proton33 or other cation (for example trimethylsilyl37 or alkylammonium): (CdSe)m(CdX2)n–1CdX2Lq + EXʹ → (CdSe)m(CdX2)n–1CdXXʹLq + EX

(3)

X, Xʹ = RCOO–, RPO(OH)O–, OH– or Cl–, for example; E = H+, (CH3)3Si+ or NR4+, for example. Direct exchange of X-type for L-type ligands disturbs charge neutrality of the NC and is therefore highly unfavourable in nonpolar environments. Observation of such reactions (for example the exchange of X-type carboxylate or phosphonate ligands with L-type pyridine38,39) can be rationalized as a ligand-promoted Z-type displacement process24, where an incoming L-type ligand aids the removal of an X-type ligand as a neutral metal–ligand complex,

followed by coordination of the L-type ligand to the metal site on the (now charge-neutral) NC surface: (CdSe)m(CdX2)nLq + 2Lʹ → (CdSe)m(CdX2)n–1LqLʹ + Lʹ-CdX2(4) X = RCOO– or RPO(OH)O–, for example; L = pyridine or RNH2, for example. Oxide NCs can bind neutral carboxylic acid ligands as pairs of negatively and positively charged X-type ligands (RCOO– and H+, respectively), thus offering a pathway for exchange between L- and X-type ligands27. Along these lines, since the L-, X-, Z-type classification of NC ligands was developed primarily from experiments with CdSe, future work beyond the model systems will be important to assess the generality of such a framework for rationalizing inorganic core composition, feasible ligand exchange reactions and other aspects of NC surface chemistry. The use of polar solvents lifts the requirement for an electrically neutral NC surface, permitting additional ligand exchange pathways. As a result, in polar solvents, charged X-type ligands can desorb and exchange via dissociative pathways40: [(CdSe)m(CdnX2n+sLq)]s– + Xʹ– → [(CdSe)m(CdnX2n+s–1XʹLq)]s– + X–

(5)

X, Xʹ = RCOO–, (NH4)S–, SCN–, In2Se42– or PbCl3–, for example. Moreover, special reagents (HBF4 or Et3OBF4) can be used to selectively attack and cleave the NC–ligand bond by protonation or alkylation of surface ligands9,41. Cleavage of X-type ligands in the presence of weakly nucleophilic anions (for example BF4– or PF6–) leaves behind uncompensated positive surface charge that allows

NATURE MATERIALS | VOL 15 | FEBRUARY 2016 | www.nature.com/naturematerials

© 2016 Macmillan Publishers Limited. All rights reserved

145

REVIEW ARTICLE a

NATURE MATERIALS DOI: 10.1038/NMAT4526 b

Conduction band

14 σ*

Passivated Cd

12

1S(e)

Electron-accepting ligand

Energy

Electron-donating ligand

10 8 6 4

CdSe–OA

2

1S3/2(h)

0

σ

HOMO

Quantum yield (%)

LUMO

σ*

Passivated Se σ

0.0 Valence band

1.0

2.0

3.0

4.0

Carboxylate coverage (nm–2)

Figure 3 | Effect of ligands on nanocrystal surface states. a, Simplified molecular orbital diagram of a CdSe quantum dot showing that the energies of Cd and Se surface electronic states (red levels) are pushed outside the bandgap upon ligand binding. HOMO, highest occupied molecular orbital; LUMO, lowest unoccupied molecular orbital. b, Experimental photoluminescence quantum yield of CdSe quantum dot samples with different cadmium oleate (OA) surface coverage. Different symbols correspond to various conditions of the nanocrystal surface treatment with the Lewis base L’ = (CH3)2NCH2CH2N(CH3)2 as described in equation 4. Error bars represent uncertainty of ligand surface density and nanocrystal solution concentration estimates. Inset: photograph of colloidal CdSe quantum dots with high coverage (right) and low coverage (left) of oleate ligands. Panel b reproduced from ref. 24, American Chemical Society.

electrostatic stabilization and further functionalization of the NC surface by weakly coordinating labile solvent molecules: (CdSe)m(CdX2)nLq + 2nEBF4 + pLʹ → [(CdSe)m(CdnLqLʹp)]2n+ + 2nBF4– +[2nEX]

(6)

E = H+, Et3O+ or NO+, for example; Lʹ = DMF. Completion of ligand exchange is influenced by the difference in ligand affinities to the NC surface and the relative abundance of incoming and outgoing ligands. Although mass action favours binding of the ligand present in excess, headgroup-specific surface affinity can prevent displacement of strongly binding species (for example phosphonate-capped CdSe in the presence of oleic acid34). The ligand affinity can be rationalized in terms of electronic, entropic and steric effects. The first case can be illustrated by application of the hard–soft acid–base (HSAB) principle42 for predicting the strength of NC–ligand binding. Classifying Lewis acids and bases into ‘hard’ and ‘soft’ categories, HSAB anticipates that strong bonds are formed by electrostatic interaction between hard Lewis acid–base pairs and by covalent interaction between soft pairs, whereas weak association is observed between members of opposite groups. The gold–sulphur bond, a classic example of robust association between soft species, is widely used to anchor ligands to the surface of Au NCs17. In contrast, hard bases (for example ligands with oxygen-containing headgroups such as carboxylates) show poor affinity to Au NCs with soft surface sites but bind strongly to NCs with more ionic lattices and harder surface sites, such as ZnSe and CdSe (ref. 9). For compound NCs, the hardness of metal surface sites depends on the hardness of the anion sublattice. For example, although the free In3+ ion is itself a hard Lewis acid, indium sites on the InAs NC surface are bound to several soft arsenic atoms, and thus rendered rather soft Lewis acids9. In molecular chemistry this is known as the ‘symbiotic effect’, where soft (hard) ligands soften (harden) the atom to which they are bound42. The effective strength of capping-layer adhesion can be significantly increased through the chelate effect, accounting for enhanced affinity of ligands containing two or more binding groups as compared with monodentate ligands. Examples of chelating ligands 146

include molecules containing carboxylate, dopamine43 and dithiol (for example dihydrolipoic acid44) anchoring groups. Steric effects also play a role in capping-layer attachment: bulky tert-butylthiolate ligands pack on CdSe NC surface at ~2  nm–2 at full coverage45 as compared with ~4 nm–2 packing of n-alkylthiolate ligands46. Although metal coordination complexes serve as a convenient foundation for understanding NC–ligand binding, the transfer of concepts explaining binding strength (such as HSAB, chelation or steric profile) from molecular to NC systems is not entirely straightforward. For example, interaction of a ligand with various crystallo­graphic facet surface patterns and edge/vertex sites opens up a manifold of potential binding modes and corresponding NC– ligand affinities. Future computational efforts must confront the intrinsic heterogeneity of such systems by considering binding of a given ligand across categories of surface atoms, taking into account potential differences in hard/soft character and steric accessibility of each unique surface site. Such work may allow identification of the factors or qualitative chemical concepts, if any, that govern the strength of NC–ligand binding.

Surface ligands and nanocrystal electronic structure

Surface ligands can directly influence the optical, electrical, magnetic and catalytic properties of NCs. Here we use semiconductor quantum dots (QDs) to demonstrate examples of such effects. In a QD, the valence and conduction bands are split into discrete, quantum-confined states47 that give rise to size-tunable luminescence colours. However, undercoordinated surface atoms with dangling bonds often contribute a set of electronic states with energies lying between the highest occupied and lowest unoccupied quantumconfined orbitals of the QD (Fig. 3a, red lines). These localized states behave as traps for electrons or holes, quenching luminescence and hampering the performance of NC-based devices. Bonding between the NC surface atom and ligand frontier orbital generates a new set of molecular orbitals with bonding (σ) and antibonding (σ*) character, with bonding orbitals stabilized and antibonding orbitals destabilized with respect to the energies of non-interacting surface atom and ligand (Fig.  3a). The formation of a strong covalent bond between the surface atom and NATURE MATERIALS | VOL 15 | FEBRUARY 2016 | www.nature.com/naturematerials

© 2016 Macmillan Publishers Limited. All rights reserved

REVIEW ARTICLE

NATURE MATERIALS DOI: 10.1038/NMAT4526 a

b

3.5

L

L

L

L

L L

L

4.0

L

Energy (eV)

L 1S(e)

4.5

L

L L

L

5.0

L

L L L

L 5.5

L L

L L

L

1S(h)

L

L 6.0 +

u) 4

B

N(





Br

+ H4

N

N SC

HS

SH O

SH

HO

HS

L

L L

L L

L

L

d HO P O



O

CdSe

H N

S



Absorbance / PL (a.u.)

c

S

LUMO

LUMO

HDA PTC

CdSe 500 600 Wavelength (nm)

1P(e) 1S(e)

σ* 1S3/2(h) 1P3/2(h)

σ

HOMO

Absorbance (a.u.)

E

OA Na4SnS4

PbS 800 1,000 1,200 Wavelength (nm)

HOMO

Figure 4 | Effect of surface ligands on electronic states within the nanocrystal inorganic core. a, Surface ligands influence the energy of highest occupied, 1S(h), and lowest unoccupied, 1S(e), states of PbS quantum dots, measured by ultraviolet photoelectron spectroscopy for representative organic and inorganic ligands: bromide, thiocyanate, benzenedithiol, 3-mercaptopropionic acid and benzenethiol. b, The shift of 1S(h) and 1S(e) states is caused by surface dipoles, shown as blue arrows, between surface atoms and ligand molecules. c, Qualitative illustration of ‘non-innocent’ ligand phenyldithiocarbamate (PTC) mixing with CdSe 1S3/2 hole state (right), giving rise to states with mixed core–ligand character (shown in red). HOMO and LUMO energies of typical hydrocarbon ligands preclude strong mixing with nanocrystal quantum-confined states (left). d, Development of electronic states with mixed nanocrystal–ligand character relaxes quantum confinement and redshifts the absorption features of CdSe quantum dots capped with PTC ligands and of PbS quantum dots capped with SnS44– ligands96. PL, photoluminescence; a.u., arbitrary units; HDA, hexadecylamine. Figure reproduced from: a, ref. 50, American Chemical Society; d, ref. 54, American Chemical Society.

ligand shifts the energies of σ- and σ*-orbitals outside the bandgap and cleans the bandgap of trap states responsible for fast non-­ radiative recombination. This molecular orbital picture agrees with DFT calculations showing the disappearance of mid-gap states in ligand-passivated NCs48. Preserving QD luminescence requires elimination of mid-gap trap states. However, the relationship between saturation of surface sites and NC electronic structure is not yet clear. For example, despite a CdSe surface atom density of ~6 nm–2, a tremendous drop in CdSe luminescence occurs at modest oleate coverage of ~3 nm–2 (ref. 24, Fig. 3b). As such, establishing the link between NC surface structure and optical properties remains a crucial open question in the field. Surface passivation upon ligand binding is common

but not universal: some ligands introduce new mid-gap electronic states and increase the rate of non-radiative relaxation: alkanethiol ligands, for example, quench luminescence of CdSe QDs by fast hole trapping 49. Ligands can also influence the absolute energy of QD electronic states. Figure 4a shows the energies of 1S(h) and 1S(e) states of PbS QDs (~3.5  nm diameter), measured by ultraviolet photoelectron spectroscopy (UPS), when capped with different surface ligands. The observed variation of band energies (~0.9  eV) across several ligand choices is large enough to be comparable to the bandgap (~1.2 eV) of these QDs. This effect has electrostatic origin: a surface-­ bound ligand molecule generates an electric dipole. If dipoles point toward the NC centre, the electric field potential shifts all energy

NATURE MATERIALS | VOL 15 | FEBRUARY 2016 | www.nature.com/naturematerials

© 2016 Macmillan Publishers Limited. All rights reserved

147

REVIEW ARTICLE

NATURE MATERIALS DOI: 10.1038/NMAT4526

Tether Carbonyl– hydrazine

Azide– alkyne Amine– carboxylic acid O

b

NHNH2

C

O

CH OH

d Norbornene– tetrazine

Function Binding specificity

N N N N

c

Targeting

R Cellular uptake

N3

Therapy

H2N

Stabilizer

O

O

O

Reversible aggregation

O

n

Poly(ethylene glycol)

Zwitterion

Photoresponse

+

O O S O



Sensing

N

P N S mono-

Hydrophobic

S S

O

N Imidazole

di-

Thiolate

a

O

Phosphine

Catechol

Anchor

Figure 5 | Modular design of surface ligands for biocompatible nanomaterials. a, Anchor links capping molecules to the crystallite through hydrophobic interaction with native ligands or through atoms that bind directly to the nanocrystal surface. b, Stabilizer region incorporates polyethylene glycol (PEG) or zwitterionic chains that strongly bind water molecules to impart hydrophilicity and prevent nonspecific adsorption. c, Capping molecules may feature reactive groups serving as a tether point for covalent conjugation to biofunctional molecules after ligand exchange. d, The biofunctional units can offer targeting, therapeutic or sensing capability.

levels down, and for the opposite case, vice versa (Fig.  4b)50. The orientation and magnitude of the surface dipole is determined by two competing contributions: the interfacial dipole formed between the surface atom and ligand headgroup, and any intrinsic dipole associated with ligand molecular structure. For Lewis-basic ligands, the interfacial dipole points from the ligand towards the metal (Lδ– → Mδ+), while the intrinsic ligand dipole depends on its chemical structure and binding mode, approaching zero for atomic ligands (halides, for example). The largest ligand-induced downward shift of electronic energy levels is observed for halide ion ligands (Fig. 4a). Because all energy levels are shifted by the same energy, this effect is not observable in absorption or luminescence spectra. However, because the energy of highest occupied and lowest unoccupied states dictates ionization potential and electron affinity, the absolute energies of electronic states are central to operation of solar cells, light-emitting diodes and other NC-based devices51,52. In the above examples, the QD absorption spectrum was set by electronic transitions within the inorganic core, and the effect of surface ligands on energy and oscillator strength of these transitions was assumed to be negligible. For typical aliphatic ligands (for example alkyl carboxylates, phosphonates, amines) this is indeed the case, as ligand frontier orbital energies are far from QD core states, maintaining strong ligand character even when bound to the QD surface. On the other hand, ligands with redox potentials comparable to the QD electron affinity and ionization potential 148

promote state mixing between ligands and the inorganic core. In coordination chemistry, ligands that create a set of electronic states with strong metal–ligand character are referred to as ‘non-innocent’ ligands53. Analogously, when the symmetry and energy of ligand and crystallite frontier orbitals align (Fig. 4c), interfacial states with mixed QD–ligand character arise, allowing core wavefunctions to extend across the inorganic/organic interface into the ligand shell. Phenyldithiocarbamate (PTC) ligands, for example, reduce the optical bandgap of CdSe NCs by up to ~0.2 eV by delocalizing the exciton hole via mixing with QD states near the valence band edge54 (Fig. 4c; 4d, top). Similarly, exchange of oleate ligands for SnS44– or AsS33– results in redshifted excitonic absorption peaks of PbS QDs (Fig. 4d, bottom). These examples show that practically every physical property of semiconductor QDs (bandgap, ionization potential, electron affinity, luminescence efficiency and others) can be tailored by surface ligands. The same holds true for other classes of NC materials, with ligands influencing the surface plasmon resonance of Au NCs55, catalytic properties of CoPt3 NCs56, magnetic properties of Fe3O4 NCs57, and so on. These are active areas of ongoing research.

Application-driven design of surface ligands

For the synthesis of colloidal NCs, ligands must be optimized to control the transformation of molecular precursors into crystallites with desired composition, size and shape. The existing body of NATURE MATERIALS | VOL 15 | FEBRUARY 2016 | www.nature.com/naturematerials

© 2016 Macmillan Publishers Limited. All rights reserved

REVIEW ARTICLE

NATURE MATERIALS DOI: 10.1038/NMAT4526 a

i.

ii.

D

ion

L

L

L

L

L

L

L L

Deposition

L L

b

L

L

L

ii.

Organic host

L = pyridine, NH4SCN, PbI2, [N(CH3)4]6Nb10O28 or (N2H5)2In2Se4, for example

i.

sit

o ep

L

L

L

L

Inorganic host

c

L

L

L

L

L

L

L

L

L

L LL

L L L LL

LL

L L L

L = 1,2-ethanedithiol or N2H4,[NBu4]I, for example

L

L

L L L LL L LL

L

L

L

L

L

L

L

L LL

L LL LL L L L

L

LL

L

L L L LL L LL

L

L

L

L

L

L

L

L L

L LL LL L L L

L

L LL

L

L L L LL LL LL

L LL L

L

L L L

L

L

L

L

L

L

Sintering

d

100 nm

50 nm

Figure 6 | Nanocrystals for device applications. a, Photograph (upper left) and sketch (lower left) of colloidal NCs capped with native hydrocarbon ligands and dispersed in a nonpolar solvent. Deposition of such NCs produces an insulating film of NCs separated with bilayers of native organic ligands (upper middle). Solid-state film treatment (i) with compact ligands produces a NC array with new, conductive ligands. Alternatively, ligand exchange of organiccapped NCs with specially designed compact ligands may be carried out in solution (ii), with subsequent deposition of such NCs into a conductive NC film. b, Integration of organic-capped luminescent NCs into a poly(methylmethacrylate) matrix produces luminescent composites. c, Dispersal of indium tin oxide NCs capped with [N(CH3)4]6Nb10O28 ligands into amorphous NbOx host yields a composite film with advanced electrochromic properties. Shown here is a scanning transmission electron micrograph of a cross-section of the composite. d, Sintering of NCs into a continuous film allows the production of functional solids with large crystalline domains. Shown here is a cross-sectional SEM image of a 12.3%-efficient solar cell with active layer made from pyridine-capped CdTe NCs sintered at 350 °C. Figure reproduced from: b, ref. 118, American Chemical Society; c, ref. 10, NPG; d, ref. 114, American Chemical Society.

knowledge indicates that organic surfactants (for example oleic acid or oleylamine) are most suitable for this job1,7. These ‘native’ capping ligands also seem to be well-suited for self-assembly of NCs into ordered superlattices1,58 and dispersal in aliphatic polymers for composite materials59. However, end applications often place demands on the NC capping layer (for example solubility in water, electrical conductivity) that require post-synthetic ligand exchange. Among the many potential uses for NC-based materials, we focus on two areas that have seen considerable development over the past decade: biomedical60 and solid-state device61 applications. Ligand design for biocompatible nanomaterials. NCs have become a compelling class of materials for biomedical research, diagnosis and treatment 62,63. Surface ligands establish the interface between the NC core and biological surroundings, playing important roles in targeting, sensing and drug delivery. In addition to providing colloidal stability, ligands have been used to tether a diverse set of molecules (dyes, polymers, peptides/proteins, DNA, saccharides and others) to the NC surface. For biomedical applications, ligand design relies on three general principles: robust anchoring of hydrophilic molecules to the NC surface, effective repulsion of proteins in blood and cytosol, and controllable attachment of biofunctional units that enable in vivo targeting, therapy or sensing.

Water-dispersible NCs have been prepared by coating over or direct replacement of native hydrophobic ligands (Fig. 5a). In the first approach, amphiphilic molecules such as phospholipids64, alkylsaccharides65 or block copolymers66 encapsulate organiccapped NCs in the hydrophobic core of a micelle. This technique minimizes introduction of dangling bonds and resulting postsynthetic changes in NC properties. But the encapsulation leaves a non-functional interior hydrocarbon shell that increases effective particle size, limiting NC access into confined regions (such as neuronal synapses) and impeding renal clearance67. The need for a compact capping layer led to development of an alternative approach using direct covalent attachment of hydrophilic ligands to the particle surface. Small molecules with thiol groups on one end and carboxylic acid on the other (for example mercaptoacetic63, mercapto­propionic68, dihydrolipoic (DHLA)69 and dimercapto­ succinic70 acids) have been used to attach hydrophilic molecules to surface metal atoms of metal or semiconductor NCs. Chelation plays a key role in the ligand anchor strength and resulting colloidal stability of NCs in a biological environment, which may present pH variations, concentrated electrolytes and competing metal-­ coordinating species. To this end, molecules based on one44,71,72 or two73,74 bidentate DHLA anchors have produced capping schemes that represent the state-of-the-art in pH, salt and thiol tolerance. Chelation is also exploited by other capping or biofunctionalization

NATURE MATERIALS | VOL 15 | FEBRUARY 2016 | www.nature.com/naturematerials

© 2016 Macmillan Publishers Limited. All rights reserved

149

REVIEW ARTICLE

NATURE MATERIALS DOI: 10.1038/NMAT4526

strategies using oligomeric phosphines75 and dopamines76, polypeptides incorporating several terminal histidine77,78 or cysteine79 residues, and imidazole-based copolymers80. Blood and cytosol are extremely crowded with macromolecules (total protein content >100 mg ml-1, >20% by volume81) that readily adsorb to surfaces. Preventing protein attachment to the NC capping layer (‘nonspecific adsorption’) prolongs in vivo circulation of particles by slowing renal clearance and preserves functions that require a pristine interface between the NC and its environment (for example targeting or sensing). For this, polyethylene glycol (PEG) has proved to be extremely useful. Covalent attachment of PEG to other molecules (‘pegylation’), originally developed to assist polypeptide drug delivery 82, is a common structural feature of most biocompatible surface ligands used so far (Fig. 5b). The utility of PEG as an inert protective coat for any surface to which it is attached, repelling and being repelled by proteins, is attributed to its unusually strong interaction with water, which produces a tightly associated hydration shell, effectively concealing PEG-capped NCs in a water sheath83. A study of protein adsorption to PEG-capped NCs suggests that PEG grafting density and chain length influence protein adsorption and cellular uptake, with long, dense capping layers minimizing both84. Recent developments of zwitterionic (ZW) surfactants that interact with water molecules even more strongly than PEG have enabled further reduction of capping-layer thickness, while preserving effective adsorbate repulsion74,85. The combination of PEG- or ZW-ligands with multidentate anchors offers highly stable and compact NCs compatible with biological systems72,86,87 and suitable for long-term bioimaging 88, living-virus tracking 89, and other purposes. Attachment of secondary molecules alongside or onto PEGor ZW-ligands enables the production of non-fouling colloidal nanomaterials with tunable biological functionality. This may be accomplished using covalent conjugation (Fig. 5c), for example in carbodiimide-mediated condensation of amine and carboxylic acid groups to form an amide linkage66,71,72,75. In addition, bioorthogonal ‘click’ reactions (for example aldehyde–hydrazine coupling 90, azide– alkyne76 or tetrazine–norbornene91 cycloadditions) circumvent the unintended crosslinking and NC precipitation that may accompany derivatization via naturally abundant amine and carboxylic acid groups. NC biofunctionality has been reviewed elsewhere60,77,92,93 and falls roughly into three categories: targeting, therapy and sensing (Fig. 5d). The anchoring, stabilizing and tethering units described above may be covalently linked, yielding a single molecule with all three, and possibility of biofunctionalization after ligand exchange. Attachment ‘in series’ yields a molecule bearing one copy of each (for example amide-linked DHLA–PEG–amine71) that binds to the NC surface end-on. Alternatively, the functional units may be grafted to a polymer scaffold ‘in parallel’ to produce a copolymer with several copies each of anchor, stabilizing and tethering modules (for example imidazole, PEG and amine units along a polyethylene backbone80). In this case, the polymer wraps around the NC, with anchor groups directed towards the grafting surface and hydrophilic/derivatizing groups extended out into the solvent. Ligand design for solid-state nanocrystal devices. Many academic and industrial labs explore colloidal NCs as solution-processed building blocks for thin-film field-effect transistors, solar cells, photo­ detectors, electrochromic windows, light-emitting diodes (LEDs) and others10,61,94. All of these applications require efficient transport of charge carriers between NCs, introducing electronic conductivity as a new requirement for NC surface ligands. Typical native ligands with long (C8–C18) hydrocarbon chains do not permit efficient electron transport. While beneficial for some applications (for example, preventing charging is important for QD luminescence), it is generally detrimental for NC-based electronic 150

devices. Exchanging insulating organic ligands with small molecules or specially designed conductive species has been key to developing competitive devices. Huge improvement in conductivity upon installation of conductive surface ligands is possible: treatment of PbSe NC film with hydrazine increases the film conductivity 1010-fold and allows electron mobility to approach ~1  cm2  V-1  s-1 (ref. 95). Similarly, replacing dodecanethiol ligands on 5-nm Au NC with chalcogenidometallate ligands results in a 1012-fold increase in the conductivity of NC film, from ~10–9 S cm–1 to 103 S cm–1 (ref. 96). Two general approaches are used for exchange of native ligands with small molecules: (i) solution exchange, where ligand-exchanged NCs retain colloidal stability, and (ii) solid-state exchange, where treatment of a NC film with molecular species displaces native ligands but does not disperse NCs in solvent (Fig.  6a). The latter approach often uses bifunctional molecules such as 1,2-ethanedithiol that can cross-link adjacent NCs. Depending on the solvent used, the elemental steps of solid-state ligand exchange reactions can follow Equations (1–4) or (5,6) above. Solid-state exchanges permit a broad range of NC surface treatments but often lead to cracks in the NC layer due to volume contraction caused by replacement of bulky initial ligands with smaller incoming molecules. This problem can be addressed using repeated deposition and ligand exchange steps in a layer-by-layer approach97. Small molecules that both aid charge transport and provide colloidal stability are more appealing for film uniformity and device manufacturability. Progress along these lines has been linked to development of inorganic ligands such as chalcogenidometallates (for example Sn2S64–, In2Se42–, CdTe22–)8,96,98, halometallates (PbCl3–, InCl4–)99,100, chalcogenide (S2–)9, halide (Cl–, I–)50,100 and pseudo­halide (SCN–, N3–)100,101 ions. These charged ligands bind to the NC surface, providing electrostatic stabilization (Box 1) in polar solvents such as H2O, N-methylformamide or dimethylsulphoxide. Carrier mobility (μ) is one of the key parameters relating material quality to device performance. It determines how quickly, for a given electric field strength E, electrons move through the film, defined by the expression vd = μE where vd is the electron drift velocity. Films of PbS and PbSe NCs treated with small organic ligands (for example 1,2-ethanedithiol or 3-mercaptopropionic acid) show electron and hole mobilities of the order of 10–4–10–2  cm2  V-1  s-1 (ref.  61). More compact and electronically transparent inorganic ligands (such as In2Se42– and SCN–) have yielded NC films with higher mobility, around 7 cm2 V-1 s-1 for PbSe NCs capped with NH4SCN (ref. 102) and 25–30 cm2 V-1 s-1 for CdSe NCs capped with NH4SCN or (N2H5)2In2Se4 ligands103,104. With mobilities among the highest measured for solution-processed semiconductors, these materials have been used for solution-processed transistors and integrated circuits105. Inorganic ligands also enable qualitatively different transport behaviour, where overlap of electronic wavefunctions of individual NCs can develop extended electronic states delocalized across multiple NCs (ref. 106). Signatures of delocalized states have been observed spectroscopically 107 and in temperature-dependent charge transport studies for CdSe and InAs NC layers103,104,108. The operation of field-effect transistors relies on shuttling one type of carrier present in excess (either electrons or holes), characteristic of unipolar devices. On the other hand, bipolar devices such as solar cells require efficient transport of both majority and minority carriers. Bipolar device design requires understanding of and control over kinetics of charge recombination and trapping. In these devices, photogenerated electron–hole pairs can either be separated at the junction and collected by electrodes, or eliminated (undesirably) through a recombination event. The fraction of carriers ultimately collected is determined by the interplay between NC absorber layer thickness and the carrier diffusion length, Ldiff = (μτreckBT/e)1/2, where τrec is the recombination-limited carrier lifetime, kB is the Boltzmann constant, T is temperature and e is the electron charge61. Accordingly, maximizing the mobility–lifetime NATURE MATERIALS | VOL 15 | FEBRUARY 2016 | www.nature.com/naturematerials

© 2016 Macmillan Publishers Limited. All rights reserved

REVIEW ARTICLE

NATURE MATERIALS DOI: 10.1038/NMAT4526 product, μτrec, is necessary for efficient harvesting of charge carriers from photovoltaic active layers109. Importantly, τrec is related to the density of trap states in the film. Incomplete surface passivation accelerates surface recombination that drains carriers from the solid and is the primary factor limiting the power conversion efficiency (PCE) of NC solar cells110. These dangling bonds create localized electronic states with energy near a band edge (‘shallow’ traps) or in the middle of the bandgap (‘deep’ traps). Shallow traps slow charge flow through the solid by repeatedly capturing and thermally reemitting the carrier (‘trap-and-release’ transport), whereas deep traps capture both types of carriers, aiding recombination. Trap concentration in QD films has been reduced using atomic ligands like halide atoms48,111, leading to part-per-million trap concentration (of ~1022 surface atoms per cm3, only ~1016 remain unpassivated110) and improved photovoltaic performance, with PbS QD solar cells now achieving over 9% PCE (refs 51,52). The remaining challenge is to design NC materials that combine high μ with the long τrec necessary for extending diffusion lengths into the micrometre range typical for CH3NH3PbI3 perovskite and bulk inorganic semiconductors. Along these lines, further elimination of dangling bonds, in particular those giving rise to mid-gap states, should allow continued improvement in NC solar cells. Combining NCs with other materials can produce various composites where surface ligands serve as an interfacial layer between the NC and its surroundings. Organic ligands seamlessly interface inorganic NC cores with organic hosts (for example polymers; Fig.  6b). For instance, luminescent CdSe/Cd1–xZnxS core–shell QDs dispersed in a polymer matrix have recently made their first appearance in a mass-produced consumer product. These composites, featured in the Sony Triluminos TV and Amazon Kindle Fire HDX 7 tablet, incorporate organic-capped NCs that absorb a fraction of blue light from the backlight to re-emit green and red photons for green and red pixels94. Inorganic surface ligands, on the other hand, are required for mixing NCs with inorganic host matrix. The broad assortment of such ligands reported in recent years has enabled integration of NCs into all-inorganic composites including chalcogenide glasses112 and amorphous oxide layers10. In the latter case, poly­oxometallate Nb10O286– ions can serve as surface ligands for indium tin oxide (ITO) NCs which, when mixed with [N(CH3)4]6Nb10O28 and annealed at 400 °C, yield an ITO-in-NbOx glass composite (Fig.  6c) with advanced electrochromic properties. By varying the applied voltage, the ITO NC plasmon absorption band can be tuned to selectively block near-infrared light while charging/discharging allows modulation of visible light absorption. These examples reveal opportunities for development of advanced materials through combining organic or inorganic nano­components and hosts with complementary properties. Thermal annealing can be used to transform films deposited from soluble NC precursor ‘inks’ into large crystallites by means of NC sintering and grain growth (Fig. 6d), an approach used for solution-processed solar cells113,114 and field-effect transistors98. Surface ligands have a great impact on NC sintering behaviour. Volatile L-type ligands (for example pyridine) aid sintering, enabling production of pure inorganic phases without organic contaminants. Alternatively, exploiting NC–ligand chemical reactions can generate a new phase. For example, pure Cu2ZnSnSe4 phase has been produced by thermal annealing mixed films of Cu2Se and ZnSe NCs, both capped with (NH4)4Sn2Se6 ligands115. NCs with inorganic ligands can also be used as a composition-matched ‘glue’ to consolidate mesoscopic micrometre-size grains or to ‘solder’ single crystals of inorganic semiconductors98,116.

Conclusion and outlook

Over the past several years, the surface of nanomaterials has been transformed from a ‘black box’ to an area of active scientific research and a powerful toolkit for materials engineers. Our understanding of

chemical bonding and electronic structure at nanoscale interfaces has improved tremendously, thanks to the development of new experimental and computational techniques capable of shedding light upon this challenging class of objects. Elucidation of ligand binding and dynamics at nanoscale surfaces, bearing analogy with the development of coordination chemistry of molecular compounds, may have a huge impact on real-world applications in the way that metal coordination complexes have done for catalysis and therapeutics. The ability to exchange ligands used in synthesis of nanostructures with custom biocompatible molecules that repel proteins and support attachment of functional units for sensing, targeting, and drug delivery has offered the first glimpse of the burgeoning field of nanomedicine. The observation of high electron mobility in printable NC films offers good prospects for cheap, solution-processed electronic materials, potentially transforming the manufacture of solar cells and electronic circuits. These signals, among others, hint that unique opportunities afforded by the surface of nanoscale materials could have considerable economic impact over the next decade in areas such as display technology and targeted drug delivery. Received 12 March 2015; accepted 27 November 2015; published online 22 January 2016; corrected after print 27 January 2016

References

1. Murray, C. B., Kagan, C. R. & Bawendi, M. G. Synthesis and characterization of monodisperse nanocrystals and close-packed nanocrystal assemblies. Annu. Rev. Mater. Sci. 30, 545–610 (2000). 2. Ozbay, E. Plasmonics: Merging photonics and electronics at nanoscale dimensions. Science 311, 189–193 (2006). 3. Bell, A. T. The impact of nanoscience on heterogeneous catalysis. Science 299, 1688–1691 (2003). 4. Giessibl, F. J. Atomic resolution of the silicon (111)-(7×7) surface by atomic force microscopy. Science 267, 68–71 (1995). 5. Chadi, D. J. Atomic and electronic structures of reconstructed Si(100) surfaces. Phys. Rev. Lett. 43, 43–47 (1979). 6. Laibinis, P. E. et al. Comparison of the structures and wetting properties of selfassembled monolayers of n-alkanethiols on the coinage metal surfaces, copper, silver, and gold. J. Am. Chem. Soc. 113, 7152–7167 (1991). 7. Yin, Y. & Alivisatos, A. P. Colloidal nanocrystal synthesis and the organic– inorganic interface. Nature 437, 664–670 (2005). 8. Kovalenko, M. V., Scheele, M. & Talapin, D. V. Colloidal nanocrystals with molecular metal chalcogenide surface ligands. Science 324, 1417–1420 (2009). 9. Nag, A. et al. Metal-free inorganic ligands for colloidal nanocrystals: S2–, HS–, Se2–, HSe–, Te2–, HTe–, TeS32–, OH–, and NH2– as surface ligands. J. Am. Chem. Soc. 133, 10612–10620 (2011). 10. Llordes, A., Garcia, G., Gazquez, J. & Milliron, D. J. Tunable near-infrared and visible-light transmittance in nanocrystal-in-glass composites. Nature 500, 323–326 (2013). 11. Pellegrino, T. et al. Hydrophobic nanocrystals coated with an amphiphilic polymer shell:  a general route to water soluble nanocrystals. Nano Lett. 4, 703–707 (2004). 12. Peng, X. et al. Shape control of CdSe nanocrystals. Nature 404, 59–61 (2000). 13. Ithurria, S. et al. Colloidal nanoplatelets with two-dimensional electronic structure. Nature Mater. 10, 936–941 (2011). 14. Nuzzo, R. G., Zegarski, B. R. & Dubois, L. H. Fundamental studies of the chemisorption of organosulfur compounds on gold(111). Implications for molecular self-assembly on gold surfaces. J. Am. Chem. Soc. 109, 733–740 (1987). 15. Dubois, L. H., Zegarski, B. R. & Nuzzo, R. G. Molecular ordering of organosulfur compounds on Au(111) and Au(100): adsorption from solution and in ultrahigh vacuum. J. Chem. Phys. 98, 678–688 (1993). 16. Dubois, L. H. & Nuzzo, R. G. Synthesis, structure, and properties of model organic surfaces. Annu. Rev. Phys. Chem. 43, 437–463 (1992). 17. Hakkinen, H. The gold–sulfur interface at the nanoscale. Nature Chem. 4, 443–455 (2012). 18. Hostetler, M. J., Stokes, J. J. & Murray, R. W. Infrared spectroscopy of threedimensional self-assembled monolayers:  n-alkanethiolate monolayers on gold cluster compounds. Langmuir 12, 3604–3612 (1996). 19. Badia, A., Cuccia, L., Demers, L., Morin, F. & Lennox, R. B. Structure and dynamics in alkanethiolate monolayers self-assembled on gold nanoparticles:  a DSC, FT-IR, and deuterium NMR study. J. Am. Chem. Soc. 119, 2682–2692 (1997).

NATURE MATERIALS | VOL 15 | FEBRUARY 2016 | www.nature.com/naturematerials

© 2016 Macmillan Publishers Limited. All rights reserved

151

REVIEW ARTICLE

NATURE MATERIALS DOI: 10.1038/NMAT4526

20. Frederick, M. T., Achtyl, J. L., Knowles, K. E., Weiss, E. A. & Geiger, F. M. Surface-amplified ligand disorder in CdSe quantum dots determined by electron and coherent vibrational spectroscopies. J. Am. Chem. Soc. 133, 7476–7481 (2011). 21. Morris-Cohen, A. J., Malicki, M., Peterson, M. D., Slavin, J. W. J. & Weiss, E. A. Chemical, structural, and quantitative analysis of the ligand shells of colloidal quantum dots. Chem. Mater. 25, 1155–1165 (2013). 22. Protesescu, L. et al. Atomistic description of thiostannate-capped CdSe nanocrystals: retention of four-coordinate SnS4 motif and preservation of Cd-rich stoichiometry. J. Am. Chem. Soc. 137, 1862–1874 (2015). 23. Green, M. L. H. A new approach to the formal classification of covalent compounds of the elements. J. Organomet. Chem. 500, 127–148 (1995). 24. Anderson, N. C., Hendricks, M. P., Choi, J. J. & Owen, J. S. Ligand exchange and the stoichiometry of metal chalcogenide nanocrystals: spectroscopic observation of facile metal–carboxylate displacement and binding. J. Am. Chem. Soc. 135, 18536–18548 (2013). 25. Owen, J. The coordination chemistry of nanocrystal surfaces. Science 347, 615–616 (2015). 26. Green, M. L. H. & Parkin, G. Application of the covalent bond classification method for the teaching of inorganic chemistry. J. Chem. Educ. 91, 807–816 (2014). 27. De Roo, J. et al. Carboxylic‑acid‑passivated metal oxide nanocrystals: ligand exchange characteristics of a new binding motif. Angew. Chem. Int. Ed. 54, 6488–6491 (2015). 28. Moreels, I. et al. Composition and size-dependent extinction coefficient of colloidal PbSe quantum dots. Chem. Mater. 19, 6101–6106 (2007). 29. Luther, J. M. & Pietryga, J. M. Stoichiometry control in quantum dots: a viable analog to impurity doping of bulk materials. ACS Nano 7, 1845–1849 (2013). 30. Israelachvili, J. N. Intermolecular and Surface Forces (Elsevier, 2011). 31. Zherebetskyy, D. et al. Hydroxylation of the surface of PbS nanocrystals passivated with oleic acid. Science 344, 1380–1384 (2014). 32. Fritzinger, B. et al. In situ observation of rapid ligand exchange in colloidal nanocrystal suspensions using transfer NOE nuclear magnetic resonance spectroscopy. J. Am. Chem. Soc. 131, 3024–3032 (2009). 33. Fritzinger, B., Capek, R. K., Lambert, K., Martins, J. C. & Hens, Z. Utilizing self-exchange to address the binding of carboxylic acid ligands to CdSe quantum dots. J. Am. Chem. Soc. 132, 10195–10201 (2010). 34. Gomes, R. et al. Binding of phosphonic acids to CdSe quantum dots: a solution NMR study. J. Phys. Chem. Lett. 2, 145–152 (2011). 35. Lingley, Z., Lu, S. & Madhukar, A. A high quantum efficiency preserving approach to ligand exchange on lead sulfide quantum dots and interdot resonant energy transfer. Nano Lett. 11, 2887–2891 (2011). 36. Hostetler, M. J., Templeton, A. C. & Murray, R. W. Dynamics of placeexchange reactions on monolayer-protected gold cluster molecules. Langmuir 15, 3782–3789 (1999). 37. Owen, J. S., Park, J., Trudeau, P.‑E. & Alivisatos, A. P. Reaction chemistry and ligand exchange at cadmium–selenide nanocrystal surfaces. J. Am. Chem. Soc. 130, 12279–12281 (2008). 38. Murray, C. B., Norris, D. J. & Bawendi, M. G. Synthesis and characterization of nearly monodisperse CdE (E = sulfur, selenium, tellurium) semiconductor nanocrystallites. J. Am. Chem. Soc. 115, 8706–8715 (1993). 39. Gur, I., Fromer, N. A., Geier, M. L. & Alivisatos, A. P. Air-stable all-inorganic nanocrystal solar cells processed from solution. Science 310, 462–465 (2005). 40. Fedin, I. & Talapin, D. V. Probing the surface of colloidal nanomaterials with potentiometry in situ. J. Am. Chem. Soc. 136, 11228–11231 (2014). 41. Rosen, E. L. et al. Exceptionally mild reactive stripping of native ligands from nanocrystal surfaces by using Meerwein’s salt. Angew. Chem. Int. Ed. 51, 684–689 (2012). 42. Pearson, R. G. Absolute electronegativity and hardness: application to inorganic chemistry. Inorg. Chem. 27, 734–740 (1988). 43. Xu, C. et al. Dopamine as a robust anchor to immobilize functional molecules on the iron oxide shell of magnetic nanoparticles. J. Am. Chem. Soc. 126, 9938–9939 (2004). 44. Uyeda, H. T., Medintz, I. L., Jaiswal, J. K., Simon, S. M. & Mattoussi, H. Synthesis of compact multidentate ligands to prepare stable hydrophilic quantum dot fluorophores. J. Am. Chem. Soc. 127, 3870–3878 (2005). 45. Webber, D. H. & Brutchey, R. L. Ligand exchange on colloidal CdSe nanocrystals using thermally labile tert-butylthiol for improved photocurrent in nanocrystal films. J. Am. Chem. Soc. 134, 1085–1092 (2011). 46. Munro, A. M., Jen-La Plante, I., Ng, M. S. & Ginger, D. S. Quantitative study of the effects of surface ligand concentration on CdSe nanocrystal photoluminescence. J. Phys. Chem. C 111, 6220–6227 (2007). 47. Brus, L. Electronic wave functions in semiconductor clusters: experiment and theory. J. Phys. Chem. 90, 2555–2560 (1986). 48. Ip, A. H. et al. Hybrid passivated colloidal quantum dot solids. Nature Nanotech. 7, 577–582 (2012). 152

49. Aldana, J., Wang, Y. A. & Peng, X. Photochemical instability of CdSe nanocrystals coated by hydrophilic thiols. J. Am. Chem. Soc. 123, 8844–8850 (2001). 50. Brown, P. R. et al. Energy level modification in lead sulfide quantum dot thin films through ligand exchange. ACS Nano 8, 5863–5872 (2014). 51. Chuang, C.‑H. M., Brown, P. R., Bulović, V. & Bawendi, M. G. Improved performance and stability in quantum dot solar cells through band alignment engineering. Nature Mater. 13, 796–801 (2014). 52. Ning, Z. et al. Air-stable n-type colloidal quantum dot solids. Nature Mater. 13, 822–828 (2014). 53. Boyer, J. L., Rochford, J., Tsai, M.-K., Muckerman, J. T. & Fujita, E. Ruthenium complexes with non-innocent ligands: electron distribution and implications for catalysis. Coord. Chem. Rev. 254, 309–330 (2010). 54. Frederick, M. T. & Weiss, E. A. Relaxation of exciton confinement in CdSe quantum dots by modification with a conjugated dithiocarbamate ligand. ACS Nano 4, 3195–3200 (2010). 55. Malinsky, M. D., Kelly, K. L., Schatz, G. C. & Van Duyne, R. P. Chain length dependence and sensing capabilities of the localized surface plasmon resonance of silver nanoparticles chemically modified with alkanethiol self-assembled monolayers. J. Am. Chem. Soc. 123, 1471–1482 (2001). 56. Kwon, S. G. et al. Capping ligands as selectivity switchers in hydrogenation reactions. Nano Lett. 12, 5382–5388 (2012). 57. Duan, H. et al. Reexamining the effects of particle size and surface chemistry on the magnetic properties of iron oxide nanocrystals: new insights into spin disorder and proton relaxivity. J. Phys. Chem. C 112, 8127–8131 (2008). 58. Shevchenko, E. V., Talapin, D. V., Kotov, N. A., O’Brien, S. & Murray, C. B. Structural diversity in binary nanoparticle superlattices. Nature 439, 55–59 (2006). 59. Lee, J., Sundar, V. C., Heine, J. R., Bawendi, M. G. & Jensen, K. F. Full color emission from II–VI semiconductor quantum dot–polymer composites. Adv. Mater. 12, 1102–1105 (2000). 60. Howes, P. D., Chandrawati, R. & Stevens, M. M. Colloidal nanoparticles as advanced biological sensors. Science 346, 1247390 (2014). 61. Talapin, D. V., Lee, J.‑S., Kovalenko, M. V. & Shevchenko, E. V. Prospects of colloidal nanocrystals for electronic and optoelectronic applications. Chem. Rev. 110, 389–458 (2010). 62. Bruchez, M., Moronne, M., Gin, P., Weiss, S. & Alivisatos, A. P. Semiconductor nanocrystals as fluorescent biological labels. Science 281, 2013–2016 (1998). 63. Chan, W. C. W. & Nie, S. Quantum dot bioconjugates for ultrasensitive nonisotopic detection. Science 281, 2016–2018 (1998). 64. Dubertret, B. et al. In vivo imaging of quantum dots encapsulated in phospholipid micelles. Science 298, 1759–1762 (2002). 65. Osaki, F., Kanamori, T., Sando, S., Sera, T. & Aoyama, Y. A quantum dot conjugated sugar ball and its cellular uptake. On the size effects of endocytosis in the subviral region. J. Am. Chem. Soc. 126, 6520–6521 (2004). 66. Gao, X., Cui, Y., Levenson, R. M., Chung, L. W. K. & Nie, S. In vivo cancer targeting and imaging with semiconductor quantum dots. Nature Biotechnol. 22, 969–976 (2004). 67. Choi, S. H. et al. Renal clearance of quantum dots. Nature Biotechnol. 25, 1165–1170 (2007). 68. Mitchell, G. P., Mirkin, C. A. & Letsinger, R. L. Programmed assembly of DNA functionalized quantum dots. J. Am. Chem. Soc. 121, 8122–8123 (1999). 69. Mattoussi, H. et al. Self-assembly of CdSe−ZnS quantum dot bioconjugates using an engineered recombinant protein. J. Am. Chem. Soc. 122, 12142–12150 (2000). 70. Lee, J.‑H. et al. Artificially engineered magnetic nanoparticles for ultra-sensitive molecular imaging. Nature Med. 13, 95–99 (2007). 71. Susumu, K. et al. Enhancing the stability and biological functionalities of quantum dots via compact multifunctional ligands. J. Am. Chem. Soc. 129, 13987–13996 (2007). 72. Liu, W. et al. Compact biocompatible quantum dots functionalized for cellular imaging. J. Am. Chem. Soc. 130, 1274–1284 (2008). 73. Stewart, M. H. et al. Multidentate poly(ethylene glycol) ligands provide colloidal stability to semiconductor and metallic nanocrystals in extreme conditions. J. Am. Chem. Soc. 132, 9804–9813 (2010). 74. Zhan, N., Palui, G., Safi, M., Ji, X. & Mattoussi, H. Multidentate zwitterionic ligands provide compact and highly biocompatible quantum dots. J. Am. Chem. Soc. 135, 13786–13795 (2013). 75. Kim, S. & Bawendi, M. G. Oligomeric ligands for luminescent and stable nanocrystal quantum dots. J. Am. Chem. Soc. 125, 14652–14653 (2003). 76. Na, H. B. et al. Multidentate catechol-based polyethylene glycol oligomers provide enhanced stability and biocompatibility to iron oxide nanoparticles. ACS Nano 6, 389–399 (2011). 77. Medintz, I. L. et al. Self-assembled nanoscale biosensors based on quantum dot FRET donors. Nature Mater. 2, 630–638 (2003). 78. Medintz, I. L., Uyeda, H. T., Goldman, E. R. & Mattoussi, H. Quantum dot bioconjugates for imaging, labelling and sensing. Nature Mater. 4, 435–446 (2005). NATURE MATERIALS | VOL 15 | FEBRUARY 2016 | www.nature.com/naturematerials

© 2016 Macmillan Publishers Limited. All rights reserved

REVIEW ARTICLE

NATURE MATERIALS DOI: 10.1038/NMAT4526 79. Pinaud, F., King, D., Moore, H.‑P. & Weiss, S. Bioactivation and cell targeting of semiconductor CdSe/ZnS nanocrystals with phytochelatin-related peptides. J. Am. Chem. Soc. 126, 6115–6123 (2004). 80. Liu, W. et al. Compact biocompatible quantum dots via RAFT-mediated synthesis of imidazole-based random copolymer ligand. J. Am. Chem. Soc. 132, 472–483 (2010). 81. Ellis, R. J. Macromolecular crowding: obvious but underappreciated. Trends Biochem. Sci. 26, 597–604 (2001). 82. Harris, J. M. & Chess, R. B. Effect of pegylation on pharmaceuticals. Nature Rev. Drug Discov. 2, 214–221 (2003). 83. Israelachvili, J. The different faces of poly(ethylene glycol). Proc. Natl Acad. Sci. USA 94, 8378–8379 (1997). 84. Walkey, C. D., Olsen, J. B., Guo, H., Emili, A. & Chan, W. C. W. Nanoparticle size and surface chemistry determine serum protein adsorption and macrophage uptake. J. Am. Chem. Soc. 134, 2139–2147 (2011). 85. Moyano, D. F. et al. Fabrication of corona-free nanoparticles with tunable hydrophobicity. ACS Nano 8, 6748–6755 (2014). 86. Yildiz, I. et al. Hydrophilic CdSe−ZnS core−shell quantum dots with reactive functional groups on their surface. Langmuir 26, 11503–11511 (2010). 87. Ling, D. et al. Multifunctional tumor pH-sensitive self-assembled nanoparticles for bimodal imaging and treatment of resistant heterogeneous tumors. J. Am. Chem. Soc. 136, 5647–5655 (2014). 88. Giovanelli, E. et al. Highly enhanced affinity of multidentate versus bidentate zwitterionic ligands for long-term quantum dot bioimaging. Langmuir 28, 15177–15184 (2012). 89. Zhang, P. et al. Click-functionalized compact quantum dots protected by multidentate-imidazole ligands: conjugation-ready nanotags for living-virus labeling and imaging. J. Am. Chem. Soc. 134, 8388–8391 (2012). 90. Blanco-Canosa, J. B., Medintz, I. L., Farrell, D., Mattoussi, H. & Dawson, P. E. Rapid covalent ligation of fluorescent peptides to water solubilized quantum dots. J. Am. Chem. Soc. 132, 10027–10033 (2010). 91. Han, H.‑S. et al. Development of a bioorthogonal and highly efficient conjugation method for quantum dots using tetrazine−norbornene cycloaddition. J. Am. Chem. Soc. 132, 7838–7839 (2010). 92. Mura, S., Nicolas, J. & Couvreur, P. Stimuli-responsive nanocarriers for drug delivery. Nature Mater. 12, 991–1003 (2013). 93. Ling, D., Hackett, M. J. & Hyeon, T. Surface ligands in synthesis, modification, assembly and biomedical applications of nanoparticles. Nano Today 9, 457–477 (2014). 94. Talapin, D. V. & Steckel, J. Quantum dot light-emitting devices. MRS Bull. 38, 685–695 (2013). 95. Talapin, D. V. & Murray, C. B. PbSe nanocrystal solids for n- and p-channel thin film field-effect transistors. Science 310, 86–89 (2005). 96. Kovalenko, M. V., Bodnarchuk, M. I., Zaumseil, J., Lee, J.‑S. & Talapin, D. V. Expanding the chemical versatility of colloidal nanocrystals capped with molecular metal chalcogenide ligands. J. Am. Chem. Soc. 132, 10085–10092 (2010). 97. Luther, J. M. et al. Structural, optical, and electrical properties of selfassembled films of PbSe nanocrystals treated with 1,2-ethanedithiol. ACS Nano 2, 271–280 (2008). 98. Dolzhnikov, D. S. et al. Composition-matched molecular ‘solders’ for semiconductors. Science 347, 425–428 (2015). 99. Dirin, D. N. et al. Lead halide perovskites and other metal halide complexes as inorganic capping ligands for colloidal nanocrystals. J. Am. Chem. Soc. 136, 6550–6553 (2014). 100. Zhang, H., Jang, J., Liu, W. & Talapin, D. V. Colloidal nanocrystals with inorganic halide, pseudohalide, and halometallate ligands. ACS Nano 8, 7359–7369 (2014). 101. Fafarman, A. T. et al. Thiocyanate-capped nanocrystal colloids: vibrational reporter of surface chemistry and solution-based route to enhanced coupling in nanocrystal solids. J. Am. Chem. Soc. 133, 15753–15761 (2011). 102. Oh, S. J. et al. Engineering charge injection and charge transport for high performance PbSe nanocrystal thin film devices and circuits. Nano Lett. 14, 6210–6216 (2014). 103. Choi, J.‑H. et al. Bandlike transport in strongly coupled and doped quantum dot solids: a route to high-performance thin-film electronics. Nano Lett. 12, 2631–2638 (2012). 104. Lee, J.‑S., Kovalenko, M. V., Huang, J., Chung, D. S. & Talapin, D. V. Band-like transport, high electron mobility and high photoconductivity in all-inorganic nanocrystal arrays. Nature Nanotech. 6, 348–352 (2011).

105. Kim, D. K., Lai, Y., Diroll, B. T., Murray, C. B. & Kagan, C. R. Flexible and lowvoltage integrated circuits constructed from high-performance nanocrystal transistors. Nature Commun. 3, 1216 (2012). 106. Shabaev, A., Efros, A. L. & Efros, A. L. Dark and photo-conductivity in ordered array of nanocrystals. Nano Lett. 13, 5454–5461 (2013). 107. Crisp, R. W., Schrauben, J. N., Beard, M. C., Luther, J. M. & Johnson, J. C. Coherent exciton delocalization in strongly coupled quantum dot arrays. Nano Lett. 13, 4862–4869 (2013). 108. Jang, J., Liu, W., Son, J. S. & Talapin, D. V. Temperature-dependent Hall and field-effect mobility in strongly coupled all-inorganic nanocrystal arrays. Nano Lett. 14, 653–662 (2014). 109. Jeong, K. S. et al. Enhanced mobility-lifetime products in PbS colloidal quantum dot photovoltaics. ACS Nano 6, 89–99 (2012). 110. Zhitomirsky, D. et al. Engineering colloidal quantum dot solids within and beyond the mobility-invariant regime. Nature Commun. 5, 1–7 (2014). 111. Tang, J. et al. Colloidal‑quantum‑dot photovoltaics using atomic-ligand passivation. Nature Mater. 10, 765–771 (2011). 112. Kovalenko, M. V., Schaller, R. D., Jarzab, D., Loi, M. A. & Talapin, D. V. Inorganically functionalized PbS–CdS colloidal nanocrystals: integration into amorphous chalcogenide glass and luminescent properties. J. Am. Chem. Soc. 134, 2457–2460 (2012). 113. Panthani, M. G. et al. Synthesis of CuInS2, CuInSe2, and Cu(InxGa1-x)Se2 (CIGS) nanocrystal inks for printable photovoltaics. J. Am. Chem. Soc. 130, 16770–16777 (2008). 114. Panthani, M. G. et al. High efficiency solution processed sintered CdTe nanocrystal solar cells: the role of interfaces. Nano Lett. 14, 670–675 (2014). 115. Jiang, C., Lee, J.‑S. & Talapin, D. V. Soluble precursors for CuInSe2, CuIn1–xGaxSe2, and Cu2ZnSn(S,Se)4 based on colloidal nanocrystals and molecular metal chalcogenide surface ligands. J. Am. Chem. Soc. 134, 5010–5013 (2012). 116. Son, J. S. et al. All-inorganic nanocrystals as a glue for BiSbTe grains: design of interfaces in mesostructured thermoelectric materials. Angew. Chem. Int. Ed. 53, 7466–7470 (2014). 117. Boles, M. A. & Talapin, D. V. Connecting the dots. Science 344, 1340–1341 (2014). 118. Protesescu, L. et al. Nanocrystals of cesium lead halide perovskites (CsPbX3, X = Cl, Br, and I): novel optoelectronic materials showing bright emission with wide color gamut. Nano Lett. 15, 3692–3696 (2015). 119. Mattoussi, H., Cumming, A. W., Murray, C. B., Bawendi, M. G. & Ober, R. Properties of CdSe nanocrystal dispersions in the dilute regime: structure and interparticle interactions. Phys. Rev. B 58, 7850–7863 (1998). 120. Saunders, A. E. & Korgel, B. A. Second virial coefficient measurements of dilute gold nanocrystal dispersions using small-angle X-ray scattering. J. Phys. Chem. B 108, 16732–16738 (2004). 121. Schapotschnikow, P., Pool, R. & Vlugt, T. J. H. Molecular simulations of interacting nanocrystals. Nano Lett. 8, 2930–2934 (2008). 122. Hens, Z. & Martins, J. C. A solution NMR toolbox for characterizing the surface chemistry of colloidal nanocrystals. Chem. Mater. 25, 1211–1221 (2013). 123. Schapotschnikow, P., Hommersom, B. & Vlugt, T. J. H. Adsorption and binding of ligands to CdSe nanocrystals. J. Phys. Chem. C 113, 12690–12698 (2009). 124. Panthani, M. G. et al. Graphene-supported high-resolution TEM and STEM imaging of silicon nanocrystals and their capping ligands. J. Phys. Chem. C 116, 22463–22468 (2012).

Acknowledgements

We thank the National Science Foundation (Award DMR-1310398) and DOD Office of Naval Research (ONR Grant N00014-13-1-0490).

Additional information

Reprints and permissions information is available online at www.nature.com/reprints. Correspondence should be addressed to D.V.T.

Competing financial interests

The authors declare no competing financial interests.

NATURE MATERIALS | VOL 15 | FEBRUARY 2016 | www.nature.com/naturematerials

© 2016 Macmillan Publishers Limited. All rights reserved

153

ARTICLES

NATURE MATERIALS DOI: 10.1038/NMAT4578

Erratum: The surface science of nanocrystals Michael A. Boles, Daishun Ling, Taeghwan Hyeon and Dmitri V. Talapin Nature Materials 15, 141–153 (2016); published online 22 January 2016; corrected after print 27 January 2016. In the version of this Review Article originally published, the righthand panel of Fig. 1b was incorrect. The correct Figure is shown below and this error has been corrected in the online versions of the Review Article. a

b

(111)

(100) Au

S

C

H

c

~30°

PbS

364

Au

NATURE MATERIALS | VOL 15 | MARCH 2016 | www.nature.com/naturematerials

© 2016 Macmillan Publishers Limited. All rights reserved

The surface science of nanocrystals.

All nanomaterials share a common feature of large surface-to-volume ratio, making their surfaces the dominant player in many physical and chemical pro...
3MB Sizes 2 Downloads 9 Views