Article Cite This: Chem. Mater. 2019, 31, 1990−2001
pubs.acs.org/cm
Controlling Ion-Exchange Balance and Morphology in Cation Exchange from Cu3−xP Nanoplatelets into InP Crystals Sungjun Koh,† Whi Dong Kim,† Wan Ki Bae,‡ Young Kuk Lee,*,§ and Doh C. Lee*,† †
Downloaded via GUILFORD COLG on July 18, 2019 at 15:45:03 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
Department of Chemical and Biomolecular Engineering, KAIST Institute for the Nanocentury, Korea Advanced Institute of Science and Technology (KAIST), Daejeon 34141, Republic of Korea ‡ SKKU Advanced Institute of Nano Technology (SAINT), Sungkyunkwan University, Suwon-si, Gyeonggi-do 16419, Republic of Korea § Advanced Materials Division, Korea Research Institute of Chemical Technology (KRICT), Daejeon 34114, Republic of Korea S Supporting Information *
ABSTRACT: Synthesis of colloidal nanocrystals (NCs), which are not readily available via the wet-chemical approach based on arrested precipitation, has often relied on templated growth. Cation exchange, in which guest cations in bulk solution replace host cations in template NCs, has evolved as one of the most powerful examples. Despite its versatility and facileness, there are caveats because most of the cationexchange processes presuppose the formation of crystalline defects, which are more or less uncontrolled in terms of population and locations. The defect formation is a consequence of the imbalance between extraction and incorporation of host and guest cations. Here, we demonstrate the controlling of ion-exchange balance in the Cu+-to-In3+ cation-exchange reaction of Cu3−xP nanoplatelets (NPLs), which triggers the nanoscale Kirkendall effect, a representative phenomenon of crystal defect generation, clearly shown by morphology change of NPLs. Cation-exchanged NPLs exhibit various morphologies depending on the ligand composition introduced in the cation-exchange reaction. For example, Kirkendall voids appear inside NPLs when Cu+ is expected to undergo solvation more than In3+ is dissolved. High-resolution transmission electron microscopy images of partially cation-exchanged NPLs show separate steps of Kirkendall void nucleation and growth at the Cu3−xP/InP interface. Elemental analysis combined with a mathematical description of stoichiometry of the Cu3−xP segment enables the quantification of Cu+ vacancy concentration, which has to do with Kirkendall void nucleation. Estimation of ion solvation energetics reveals that the composition of solvating ligands is responsible for imbalance between cation solvation and desolvation, resulting in an increase in Cu+ vacancy. In addition to defect generation inside a crystal, defect removal is visualized via the transformation of cracked InP NPLs into hollow NPLs after annealing at high temperature. Our indepth study on ion-exchange balance and morphology change in cation-exchange reaction not only provides mechanistic insights into internal defect formation and removal process accompanied with cation exchange but also opens up the possibility of diverse morphology control of NCs.
■
INTRODUCTION Chemical synthesis of colloidal nanocrystals (NCs) has shown remarkable progress in the past three decades.1−4 With growing demands for application of NCs in a wide range of fields, advanced techniques such as templated growth have evolved and enabled tailoring of size, shape, composition, structure, and morphology of NCs.5−10 One of the powerful techniques is utilization of cation exchange, a phenomenon describing replacement of the cations in an ionic compound by externally introduced cations.11 By virtue of high surface reactivity and low solid-state ionic diffusion barrier of NCs compared to bulk, it is facile to induce cation-exchange reaction as postsynthetic modification of NCs.12,13 In most cases, template NCs usually retain the overall shape after conversion.14,15 In this regard, cation exchange could allow for © 2019 American Chemical Society
the growth of NCs with composition or shape that is not likely to be achieved by a typical bottom-up strategy.16−20 Although the cation exchange of colloidal semiconductor NCs has been in the spotlight as a versatile synthetic method, lack of understanding in dynamics and crystallographic effect impedes the realization of technological application. It is noteworthy that cation-exchange reaction often results in crystalline defects in product NCs whose population and locations are not controlled.14 Indeed, cation-exchanged NCs contain high density of defects compared to bottom-up synthesized NCs. Low photoluminescence and optical stability Received: November 21, 2018 Revised: February 28, 2019 Published: March 1, 2019 1990
DOI: 10.1021/acs.chemmater.8b04859 Chem. Mater. 2019, 31, 1990−2001
Article
Chemistry of Materials
Scheme 1. Schematic Description of Morphology and Structure Changes of NPLs Resulting from (a) Ion-Exchange Balance in the Cu+-to-In3+ Cation-Exchange Reaction of Cu3−xP NPLs and (b) Annealing at High Temperature
complex.35 In this regard, the ion solvation and desolvation balance would depend on the selection or composition of the ligand mixture utilized in the cation exchange. Previous studies focusing on cation-exchange kinetics and dynamics with respect to organic ligands imply that controlling the type or composition of ligands would be key to manipulate the ionexchange balance.18,36 Here, we report controlling of the ion-exchange balance and the following morphology change during the Cu+-to-In3+ cation-exchange reaction of Cu3−xP nanoplatelets (NPLs). Controlling the compositions of organic ligands, trioctylphosphine (TOP) and oleylamine (OAm), introduced in cation exchange enables manipulating the extraction and incorporation balance between Cu+ and In3+, that is, solvation of Cu+ and desolvation of In3+. Depending on the ligand composition, cation-exchanged NPLs show clearly different morphologies as depicted in Scheme 1a. Especially, the nanoscale Kirkendall effect takes place when the solvation of Cu+ outweighs the stoichiometric quantity of In3+ desolvation, resulting in cracked morphology. High-resolution transmission electron microscopy (HR-TEM) analysis visualizes the nucleation and growth of Kirkendall voids at the Cu3−xP/InP heterointerface during cation exchange. In addition, energy-dispersive X-ray (EDX) spectroscopy analysis on partially cation-exchanged Cu3−xP/ InP NPLs shows that the elemental composition of the heteroNPLs also depends on the ligand composition. By relating the composition of the Cu3−xP segment to the reaction coordinate, we quantify the concentration of Cu+ vacancy, which is the source of Kirkendall void nucleation. The prediction of ionexchange balance during the cation exchange has heretofore relied on the HSAB theory, a concept that describes stability of a metal−ligand complex in a qualitative fashion.35 We provide quantitative insights into solvation energetics that governs ionexchange balance. Moreover, in addition to the defect generation by diffusion imbalance during cation exchange, we present the defect removal pathway by proposing a hollow NC formation mechanism based on the observation of void coalescence inside cracked NPLs (Scheme 1b).
of cation-exchanged Cd- or Zn-chalcogenide core−shell NCs are attributed to the poor crystallographic quality of cationexchanged products.16,21 Occasionally, such defects lead to collapse of the anion framework of product NCs. Son et al. observed complete disruption of the anion sublattice of CdSe NRs after cation exchange to Ag2Se.12 In our previous work on direct Cd-to-Pb cation exchange of CdSe NRs, we also observed the transformation of CdSe NRs to spherical PbSe NCs in a certain reaction condition.22 We have raised the possibility of defect formation within NRs during cation exchange due to the diffusion difference between Cd2+ and Pb2+. However, the mechanism of defect formation at the atomic scale remains unclear. For better design of the cationexchange reaction and practical use of cation-exchanged NCs, it is necessary to understand the defect formation pathway in the context of the imbalance between extraction and insertion of host and guest cations. The nanoscale Kirkendall effect entails a process in which the imbalance in the ion diffusion rate causes formation of defects or voids within colloidal NCs.23,24 In bulk, the Kirkendall effect normally takes place in intermetallic diffusion processes.25 Kirkendall voids evolve at the interface such as solder joints in microelectronic devices.26,27 In the nanoscale, the Kirkendall effect takes place in processes involving solidstate interdiffusion such as cation exchange,28 galvanic replacement,29,30 and oxidation, sulfidation, or phosphidization of metal NCs.23,31,32 Similar to the case of bulk, the nanoscale Kirkendall effect results in defective sites inside NCs. While defect formation is considered detrimental for optoelectronic applications of product NCs, it could often be beneficial for catalytic or electrochemical properties of NCs.33,34 A bigger issue is that no chemical protocol for cation exchange of NCs has come to grips with controlled ion-exchange balance. In this sense, it is difficult to predict the condition that prevents or induces the nanoscale Kirkendall effect. Considering the nature of wet-chemical reaction, organic ligands are expected to have a profound bearing on the ion-exchange balance. Utilized in cation exchange to form complexes with guest and host cations in bulk solution, organic ligands with different functional groups have different binding affinities for various cations based on the hard and soft acids and bases (HSAB) theory, a qualitative concept that covers stability of the metal−ligand
■
EXPERIMENTAL SECTION
Chemicals. Copper(II) chloride dihydrate (≥99%), copper(I) iodide (≥99.999%), n-trioctylphosphine (≥97%), n-trioctylphosphine 1991
DOI: 10.1021/acs.chemmater.8b04859 Chem. Mater. 2019, 31, 1990−2001
Article
Chemistry of Materials
Figure 1. TEM images of (a) Cu3−xP NPLs and (c) InP NPLs. HR-TEM images and the corresponding FFT patterns of (b) Cu3−xP NPL and (d) InP NPL. (e) Description of the phosphorus framework and unit cells of hexagonal Cu3P and InP. (f) Absorption spectra of Cu3−xP NPLs and InP NPLs. (g) XRD patterns of Cu3P NPLs and InP NPLs. oxide (≥99%), oleylamine (≥70%), and 1-octadecene (≥90%) were purchased from Sigma-Aldrich Co. Indium(III) iodide (≥99.999%) was purchased from Alfa Co. Tris(trimethylsilyl)phosphine (≥95%) was purchased from JSI Silicone Co. All the purchased chemicals were used without further purifications. Synthesis of Cu3−xP NPLs. First, 68.2 mg of copper(II) chloride dihydrate (0.4 mmol), 4 g of n-trioctylphosphine oxide, and 20 mL of oleylamine were added into a round-bottom three-neck flask (100 mL, 14/20) equipped with a reflux condenser and a thermocouple. The flask was degassed under vacuum (
