J. Am. Chem. Soc. 1997, 119, 7019-7029
`
`7019
`
`Epitaxial Growth of Highly Luminescent CdSe/CdS Core/Shell
`Nanocrystals with Photostability and Electronic Accessibility
`
`Xiaogang Peng, Michael C. Schlamp, Andreas V. Kadavanich, and
`A. P. Alivisatos*
`Contribution from the Department of Chemistry, UniVersity of California, Berkeley, and
`Molecular Design Institute, Lawrence Berkeley Laboratory, Berkeley, California 94720
`ReceiVed March 10, 1997X
`
`Abstract: The synthesis of epitaxially grown, wurtzite CdSe/CdS core/shell nanocrystals is reported. Shells of up
`to three monolayers in thickness were grown on cores ranging in diameter from 23 to 39 Å. Shell growth was
`controllable to within a tenth of a monolayer and was consistently accompanied by a red shift of the absorption
`spectrum, an increase of the room temperature photoluminescence quantum yield (up to at least 50%), and an increase
`in the photostability. Shell growth was shown to be uniform and epitaxial by the use of X-ray photoelectron
`spectroscopy (XPS), X-ray diffraction (XRD), high resolution transmission electron microscopy (HRTEM), and optical
`spectroscopy. The experimental results indicate that in the excited state the hole is confined to the core and the
`electron is delocalized throughout the entire structure. The photostability can be explained by the confinement of
`the hole, while the delocalization of the electron results in a degree of electronic accessibility that makes these
`nanocrystals attractive for use in optoelectronic devices.
`
`Introduction
`During the past decade, tremendous advances in colloid
`chemistry have led to the preparation of high quality nanometer
`sized semiconductor crystals.1-9 A current challenge is to apply
`the full range of techniques employed in two-dimensional (2D)
`artificial structure growth, such as epitaxial growth and band
`gap patterning,10,11 to nanocrystals fabricated by wet chemical
`routes. This paper details the synthesis of an epitaxial core/
`shell system. In 2D systems, it is possible to prepare either a
`type I structure, in which both electrons and holes are in their
`lowest energy states in the same material, or a type II structure,
`in which their lowest energy states are spatially separated in
`different materials. In certain 2D structures12,13 this CdSe/CdS
`system has yielded a type II quantum well. However, because
`the kinetic energy of electrons and holes are relatively large in
`0D structures,14 the potential energy steps created by alternating
`materials in 2D may not always be sufficient to confine carriers
`in 0D. Such an example is presented here, where the hole is
`largely confined to the core and the electron is delocalized
`throughout the core/shell structure.
`Because the surface of a nanocrystal is made up of atoms
`that are not fully coordinated, it is highly active and invites the
`possibility of epitaxial overgrowth of another semiconductor
`X Abstract published in AdVance ACS Abstracts, July 1, 1997.
`(1) Henglein, A. Top. Curr. Chem. 1988, 143, 113.
`(2) Bawendi, M. G.; Steigerwald, M. L.; Brus, L. E. Annu. ReV. Phys.
`Chem. 1990, 41, 477.
`(3) Brus, L. E. Appl. Phys. A 1991, 53, 465.
`(4) Wang, Y.; Herron, N. J. Phys. Chem. 1991, 95, 525.
`(5) Heath, J. R. Science 1992, 258, 1131-1133.
`(6) Weller, H. AdV. Mater. 1993, 5, 88.
`(7) Micic, O. I.; Curtis, C. J.; Jones, K. M.; Sprague, J. R.; Nozik, A. J.
`J. Phys. Chem. 1994, 98, 4966-4969.
`(8) Alivisatos, A. P. J. Phys. Chem. 1996, 100, 13226.
`(9) Alivisatos, A. P. Science 1996, 271, 933.
`(10) Tsu, R.; Esaki, L. App. Phys. Lett. 1973, 22, 562-564.
`(11) Sze, S. M. Physics of Semiconductor DeVices; Wiley-Interscience:
`New York, 1986.
`(12) Langbein, W.; Hetterich, M.; Grun, M.; Klingshirn, C.; Kalt, H.
`Appl. Phys. Lett. 1994, 65, 2466.
`(13) Halsall, M. P.; Nicholls, J. E.; Davies, J. J.; Cockayne, B.; Wright,
`P. J.; Cullis, A. G. Semicond. Sci. Technol. 1988, 3, 1126.
`(14) Rosetti, R.; Nakahara, S.; Brus, L. E. J. Chem. Phys. 1983, 79, 1086.
`
`Nanoco Technologies, Ltd
`S0002-7863(97)00754-3 CCC: $14.00 © 1997 American Chemical Society
`1 of 11
`EXHIBIT 1010
`
`or other inorganic material. In addition, surface atoms act like
`defects unless passivated.15 To remove these defects, high
`quality homogeneous II-VI and III-V nanocrystals have been
`passivated with long chain organic surfactants. These “capped”
`nanocrystals have room temperature photoluminescence quan-
`tum yields as high as 10% with a very long fluorescence lifetime
`and often have some non-band-edge luminescence.16,17 How-
`ever, it is generally very difficult to simultaneously passivate
`both anionic and cationic surface sites by organic ligands; there
`are always some dangling bonds on the surface. For nano-
`crystals, inorganic epitaxial growth can not only eliminate both
`the anionic and cationic surface dangling bonds but also generate
`a new nanocrystal system with novel properties. Most of the
`properties of these “core/shell” systems are dependent on both
`core and shell materials.
`Inorganic passivation has been explored in such core/shell
`systems as Si/SiO2,18 CdS/Cd(OH)2,15 CdSe/ZnSe,19,20 CdSe/
`ZnS,21,22 CdS/HgS/CdS,23,24 and CdSe/CdS25 and has been
`shown to improve luminescence quantum yields,15,19-23 decrease
`fluorescence lifetimes,21 and have other benefits26 related to the
`tailoring of relative bandgap positions between the two materials.
`
`(15) Spanhel, L.; Haase, M.; Weller, H.; Henglein, A. J. Am. Chem. Soc.
`1987, 109, 5649.
`(16) Norris, D. J.; Sacra, A.; Murray, C. B.; Bawendi, M. G. Phys. ReV.
`Lett. 1994, 72, 2612.
`(17) Hoheisel, W.; Colvin, V. L.; Johnson, C. S.; Alivisatos, A. P. J.
`Chem. Phys. 1994, 101, 8455.
`(18) Wilson, W. L.; Szajowski, P. F.; Brus, L. E. Science 1993, 262,
`1242-1244.
`(19) Danek, M.; Jensen, K. F.; Murray, C. B.; Bawendi, M. G. Chem.
`Mater. 1996, 8, 173.
`(20) Hoener, C. F.; Allan, K. A.; Bard, A. J.; Campion, A.; Fox, M. A.;
`Mallouk, T. E.; Webber, S. E.; White, J. M. J. Phys. Chem. 1992, 96, 3812.
`(21) Hines, M. A.; Guyot-Sionnest, P. J. Phys. Chem. 1996, 100, 468.
`(22) Kortan, A. R.; Hull, R.; Opila, R. L.; Bawendi, M. G.; Steigerwald,
`M. L.; Carroll, P. J.; Brus, L. E. J. Am. Chem. Soc. 1990, 112, 1327.
`(23) Mews, A.; Eychmuller, A.; Giersig, M.; Schooss, D.; Weller, H. J.
`Phys. Chem. 1994, 98, 934.
`(24) Mews, A.; Kadavanich, A. V.; Banin, U.; Alivisatos, A. P. Phys.
`ReV. B. 1996, 53, 13242.
`(25) Tian, Y.; Newton, T.; Kotov, N. A.; Guldi, D. M.; Fendler, J. H. J.
`Phys. Chem. 1996, 100, 8927.
`(26) Danek, M.; Jensen, K. F.; Murray, C. B.; Bawendi, M. G. Appl.
`Phys. Lett. 1994, 65, 2795.
`
`

`

`7020 J. Am. Chem. Soc., Vol. 119, No. 30, 1997
`
`Peng et al.
`
`Figure 1. Schematic synthesis of CdSe/CdS core/shell nanocrystals.
`
`Often these efforts have been hindered by a lack of size control
`or amorphous or nonepitaxial growth.
`Reported here is a synthesis of wurtzite CdSe/CdS core/shell
`nanocrystals. A shell having uniform thickness is grown
`epitaxially on the core, and the resulting nanocrystals have high
`room temperature quantum yields (>50%). The wide range of
`core and shell sizes studied provide tunability of band edge
`luminescence at these high quantum yields. In this system, the
`holes are confined to the core, while the electrons are delocalized
`as a result of the similar electron affinities of the core and shell.
`As a consequence, the nanocrystals are extremely stable with
`respect to photooxidation, which requires a hole trapped at the
`surface. The increase in photochemical stability does not come
`at the expense of electrical access; electrons are fully delocalized
`and may be accessed for applications where the ability of a dense
`matrix of these nanocrystals to both conduct an electron current
`and luminesce is paramount.
`The wurtzite CdSe/CdS system is ideal in many respects. The
`lattice mismatch of 3.9% is small enough to allow epitaxial
`growth while still preventing alloying, and the difference in band
`gaps is large enough for shell growth to increase the quantum
`yield and stability of the cores. Some of these characteristics
`of CdSe/CdS epitaxial systems have been observed in 2D
`superlattices.13,27-29 During the past several years, the synthesis
`of CdSe nanocrystals has become well established.30,31 Mono-
`disperse, faceted, and highly crystalline CdSe nanocrystals can
`now be reproducibly and controllably synthesized on a gram
`scale. A high quality core is very helpful for the growth and
`characterization of the core/shell structures as shown below.
`The synthetic conditions presented here allow shells to be
`grown of varying thicknesses on varying core sizes. The
`conditions also prevent core dissolution and CdS-only nano-
`crystal formation during shell growth. The chemistry of shell
`growth is cleansthe added reagents react quantitatively. The
`resulting reaction provides an additional means for tuning the
`band gap of the nanocrystals while yielding a stable product
`which can be treated as a chemical reagent for building up more
`complex structures.
`
`Experimental Section
`
`(A) Synthesis of Nearly Monodisperse CdSe and CdS Nano-
`crystals. Chemicals. The dimethylcadmium (Cd(CH3)2) and tribut-
`ylphosphine (TBP) were purchased from Strem. Cd(CH3)2 was stored
`in a refrigerator in the drybox after being vacuum transferred. Selenium
`
`(27) Halsall, M. P.; Nicholls, J. E.; Davies, J. J.; Wright, P. J.; Cockayne,
`B. Surf. Sci. 1990, 228, 41.
`(28) Grun, M.; Hetterich, M.; Klingshirn, C.; Rosenauer, A.; Zweck, J.;
`Gebhardt, W. J. Crystal Growth 1994, 138, 150-154.
`(29) Grun, M.; Hetterich, M.; Becker, U.; Giessen, H.; Klingshirn, C. J.
`Crystal Growth 1994, 141, 68.
`(30) Murray, C. B.; Norris, D. J.; Bawendi, M. G. J. Am. Chem. Soc.
`1993, 115, 8706.
`(31) Bowen Katari, J. E.; Colvin, V. L.; Alivisatos, A. P. J. Phys. Chem.
`1994, 98, 4109.
`
`Figure 2. Absorption spectra of a core/shell and CdS-only nanocrystal
`mixture as synthesized (dashed) and after successive separations (solid
`and long dashed). The peak at 3.5 eV represents the CdS-only
`nanocrystals (see text), while the lower energy features are characteristic
`of the core/shell nanocrystals. During successive separations, the CdS-
`only peak intensity decreases relative to that of the core/shells, indicating
`successful separation.
`
`(Se), tri-n-octylphosphine oxide (TOPO), anhydrous methanol, anhy-
`drous pyridine, and anhydrous toluene were purchased from Aldrich.
`Bis-trimethylsilane sulfide ((TMS)2S) was purchased from Fluka.
`Stock Solution for CdSe Nanocrystals Synthesis. The desired
`amount of Cd(CH3)2 was added to a solution of Se powder dissolved
`in TBP. Cadmium to selenium molar ratios were 1.0/0.7 or 1.0/0.9.
`The Se concentration was maintained near 0.20 mol/kg. Stock solutions
`were prepared in the drybox and stored under N2 in the refrigerator.
`Stock Solution for CdS Nanocrystal Synthesis. A solution of Cd
`(CH3)2 was dissolved in 3/1 TBP/toluene in the drybox. To this was
`added (TMS)2S. The cadmium to sulfur molar ratio was 1.0/0.4. The
`sulfur concentration was maintained near 0.05 mol/kg.
`Synthesis Procedure of CdSe and CdS Nanocrystals. For a typical
`synthesis, 12 g of TOPO was transferred to a three-necked flask and
`briefly heated to 360 (cid:176)C under Ar flow on a Schlenk line. The stock
`solution (6 mL) was quickly (<1 s) injected into the hot TOPO using
`a large bore needle (17-12 gauge). The reaction was either stopped
`immediately by quickly removing the heating or allowed to continue
`after lowering the temperature to 300 (cid:176)C for CdSe nanocrystals or 250
`(cid:176)C for CdS nanocrystals. Nanocrystal size was monitored by UV-vis
`spectra of aliquots taken from the reaction solution. The reaction was
`stopped by removing the heating after the desired size was achieved.
`This heating process can last from minutes to hours. Nanocrystals were
`precipitated by the addition of methanol to the cooled, room temperature
`reaction mixture. After centrifugation and drying under N2 flow, a
`powder was obtained. Typically, hundreds of milligrams to a few grams
`of TOPO capped nanocrystals are synthesized by each reaction. No
`further size selective purification was done for the samples mentioned
`here. CdS nanocrystals of 35 Å diameter synthesized in this way were
`used in all experiments presented here, except those of Figure 2.
`(B) Synthesis of CdSe/CdS Core/Shell Structure. Stock Solution
`(TMS)2S (100 (cid:237)L) was added to a solution
`for CdS Shell Growth.
`of 0.033 g of Cd(CH3)2 dissolved in 3.81 g of TBP and stored under
`N2 in refrigerator. The cadmium to sulfur ratio in this stock solution
`is 1.0/2.1. All the core/shell nanocrystals were synthesized using this
`stock solution except where noted.
`
`2 of 11
`
`

`

`Epitaxial Growth of CdSe/CdS Core/Shell Nanocrystals
`
`Procedure for Shell Growth onto CdSe Nanocrystals. TOPO (2-
`13 mg) capped CdSe nanocrystals were transferred to a three-necked
`flask, degassed, and purged with Ar on the Schlenk line. To this was
`added 15 mL of anhydrous pyridine. The nanocrystals dissolved
`readily, and the solution was allowed to reflux overnight under Ar flow.
`From 0.5 to 2 mL of diluted CdS stock (1:3 volume ratio of stock:
`TBP) was then added dropwise (1 drop per s) to the reaction solution
`at 100 (cid:176)C. Absorption spectra of removed aliquots were recorded after
`each CdS addition.
`Stopping the CdS addition and removing the heating completed the
`growth. Dodecylamine was added to the reaction solution at room
`temperature until the nanocrystals precipitated. Nanocrystals were
`isolated from solvents and unreacted reagents by centrifugation. These
`isolated, capped nanocrystals were readily soluble in CHCl3 or CH2-
`Cl2 but not pyridine.
`(C) Characterization. UV-Vis Absorption Spectroscopy. Ab-
`sorption spectra were obtained using an HP Model 8452 ultraviolet-
`visible absorption (UV-vis) diode array spectrometer with a collimated
`deuterium lamp source having a resolution of 2 nm. Samples were
`dispersed in toluene, chloroform, or pyridine for measurement.
`Photoluminescence (PL) and Photoluminescence Excitation (PLE)
`Spectroscopy. Photoluminescence experiments were conducted on a
`Hitachi F-4500 fluorescence spectrometer with collection at 90(cid:176) and
`resolution of 2.5 nm. Quantum yields of nanocrystal solutions were
`calculated by comparing the integrated emission to that of Rhodamine
`6G or Rhodamine 640 in methanol. Optical densities of all solutions
`were below 0.3 at the excitation wavelengths used.
`X-ray Diffraction (XRD). X-ray powder diffraction patterns were
`recorded using a Siemens Model D5000 X-ray diffraction with Cu KR
`radiation. Nanocrystals were placed on quartz plates for measurement.
`Transmission Electron Microscopy (TEM). Samples for TEM
`were deposited onto copper TEM grids coated with thin (5-50 nm
`thickness) carbon films. To improve adhesion of the nanocrystals to
`the carbon film, the grids were pretreated in an argon ion glow discharge
`(30 s in 50 mTorr argon). Immediately after the glow discharge a drop
`of nanocrystal dissolved in dimethyl chloride or chloroform was placed
`on a grid. The excess liquid was then wicked away with tissue, and
`the grid was allowed to dry in air. The grids were examined in a
`TopCon EM002B microscope with an Ultra-High Resolution (UHR)
`polepiece using a LaB6 filament operated at 200 kV.
`Nanocrystal sizes were measured by counting HRTEM fringe
`spacings directly off the micrograph negatives. The spacings used for
`size measurements were the a spacing (100 planes, 3.72 Å) and the c
`spacing (002 planes, 3.51 Å). These large spacings are easy to measure
`and not likely to be confused with other materials, such as oxides. When
`cross fringes are present in a nanocrystal, these two spacings can be
`uniquely identified based solely on the symmetry of the fringes
`observed. Even without full cross fringes, such spacings can frequently
`be identified by careful study of the detailed fringe contrast near the
`edges of the crystal, the location and orientation of stacking faults
`(which are parallel to 002 planes), and other considerations. For
`instance, as has been previously observed (see ref 30) and verified in
`this study, these nanocrystals are usually elongated along the c-axis.
`Hence simple fringes running perpendicular to the long axis are usually
`assigned to be (002) planes. Any misidentification of such planes will
`result in a systematic error in the measured nanocrystal diameter of at
`most (j3.72-3.51j)/3.51 ) 6%. This is considered to be an acceptable
`error as it allows the counting of many more nanocrystals to obtain
`good sizing statistics. In fact, this error is smaller than the statistical
`uncertainty in the measured diameters and is comparable to the
`discrepancy in describing prolate particles by a single average diameter.
`On the order of 150 nanocrystals were measured for each size
`distribution cited in this study.
`Measurements of aspect ratio were undertaken only for nanocrystals
`exhibiting unambiguous cross-fringe contrast.
`X-ray Photoelectron Spectroscopy (XPS). XPS was performed
`using a Perkin-Elmer PHI 5300 ESCA System. Thin cast films of
`nanocrystals were placed on evaporated gold films for measurement.
`
`Results
`
`Synthesis of Core/Shell Nanocrystals. Dropwise injection
`of CdS stock into a solution of CdSe nanocrystals as synthesized
`
`J. Am. Chem. Soc., Vol. 119, No. 30, 1997 7021
`in TOPO and heated to 200 (cid:176)C yields a product for which there
`was no evidence for the formation of a core/shell structure. It
`was then decided to attempt shell growth on a sample of CdSe
`nanocrystals precipitated and redissolved in a solvent more
`amenable to shell growth than TOPO. Pyridine was chosen as
`this solvent based on previous IR and XPS31 experiments which
`revealed that, after refluxing in pyridine overnight, TOPO could
`almost completely be removed from CdSe nanocrystals without
`affecting the nanocrystal structure. Pyridine displaces TOPO
`and forms a weak bond to surface Cd atoms of CdSe nano-
`crystals. Furthermore, NMR studies have revealed that pyridine
`forms a labile bond to the surface Cd atoms when CdS
`nanocrystals are dissolved in pyridine.32 Nanocrystals refluxed
`in pyridine are dynamically capped, providing simultaneous
`chemical stability and access to the surface. In addition to the
`use of pyridine, the reaction temperature was lowered to 100
`(cid:176)C, and the starting core (CdSe nanocrystal) concentration was
`lowered. This reaction is schematically shown in Figure 1.
`Using these new conditions, and a Cd/S molar ratio in the
`stock of 1.0/0.7, the growth of a CdS shell was observed by
`UV-vis, PL, and TEM. However, the formation of CdS-only
`nanocrystals (having a first exciton peak in their absorption
`spectra near 3.4 eV (360 nm)) was also observed. CdS-only
`formation was also observed about 10 min after the injection
`of the same shell stock solution into pyridine (without cores)
`at 100 (cid:176)C. These nanocrystals, suspected to be similar to
`Cd32S50 molecular species,33 were found to be insoluble in
`pyridine, and for this reason could easily be separated from the
`core/shell nanocrystals. Figure 2 documents this separation for
`a somewhat anomalous sample in which a large amount of CdS-
`only nanocrystals were formed. As can be seen, each successive
`precipitation with methanol, centrifugation, and redissolution
`in methylene chloride removed a large percentage of the CdS
`while retaining the core/shell nanocrystals. These CdS nano-
`crystals were too small to have their crystal structure unambigu-
`ously determined using XRD.
`By changing the Cd/S ratio in the stock solution to 1.0/2.1,
`the formation of CdS-only nanocrystals was prevented; only
`shell growth was observed. By checking the absorption
`spectrum of aliquots,
`it was found that shell growth was
`completed immediately after the addition of the stock solution.
`The final achievable shell thickness was limited by the
`solubility of the core/shell nanocrystals; the reaction had to be
`stopped before the solution became turbid. This turbidity is
`thought to be the result of increased van der Waals interactions
`between nanocrystals brought about by their increased size, their
`dynamically capped surface, and the slightly polar nature of
`the solvent. The maximum achievable shell thickness depended
`on nanocrystal concentration, stock injection volume, and
`nanocrystal size.
`Judicious control of the conditions for shell growth lead to
`quantitative use of the injected CdS stock solution. This is
`indicated both by the absence of a CdS-only peak in the
`absorption spectra and also by a quantitative correlation between
`the amount of injected stock and the average increase in
`nanocrystal size. This correlation was used to determine
`nanocrystal sizes in samples having very thin shells.
`Absorption, Photoluminescence, and Quantum Yields.
`Plots of the photoluminescence (PL) and absorption spectra for
`two series of nanocrystals having two different core sizes and
`varying shell thicknesses are given in Figure 3. The core/shell
`nanocrystals were capped with dodecylamine and dissolved in
`methylene chloride. Unless otherwise stated, this capping group
`
`(32) Sachleben, J. R.; Wooten, E. W.; Emsley, L.; Pines, A.; Colvin, V.
`L.; Alivisatos, A. P. Chem. Phys. Lett. 1992, 198, 431.
`(33) Herron, H.; Calabrese, J. C.; Farneth, W. E.; Wang, Y. Science 1993,
`259, 1426.
`
`3 of 11
`
`

`

`7022 J. Am. Chem. Soc., Vol. 119, No. 30, 1997
`
`Peng et al.
`
`Figure 4. Photoluminescence quantum yield variation with shell
`thickness for three different core/shell syntheses starting with core
`diameters of 23 Å (circles), 34 Å (X’s), and 39 Å (squares).
`
`nanocrystals were capped with amine and used as a standard
`sample for comparison to core/shell samples in all experiments.
`During shell growth, the absorption spectrum roughly main-
`tained its overall shape, with a slight broadening of features,
`while shifting to lower energy. For the smaller cores a shift of
`approximately 0.05 eV accompanied each CdS stock injection,
`while smaller shifts of 0.02 eV per injection were observed for
`the larger cores. Smaller cores also showed a total absorption
`shift of near 0.17 eV, while larger cores showed a total shift of
`about 0.06 eV before the reaction mixture became turbid. These
`changes in the absorption spectra are caused by the closing
`together of electronic levels that accompanies a size increase14
`and by a slight worsening of the size distribution.
`The PL spectra of Figure 3 result from excitation at or slightly
`blue of the first exciton peak in the absorption spectrum. The
`PL peaks reveal band edge luminescence for all shell thick-
`nesses; no deep trap luminescence was detected. The shift of
`the peak of the PL from the first peak in the absorption spectrum
`is the result of a convolution of the size distribution and the
`emitting state.34 This shift is also highly dependent on excitation
`energy for excitation near the first peak.17
`Figure 4 shows the increase in quantum yield with shell
`thickness for three series of core/shell nanocrystals with cores
`of 23, 34, and 39 Å diameter. This increasing trend is observed
`up to a shell thickness of about three monolayers.
`(A shell
`thickness of (cid:24)3.5 Å corresponds to one full monolayer of
`growth.) The photoluminescence quantum yields (QY) increased
`from a few percent for core samples to at least 50% for samples
`with the thickest shells.
`It should be stressed that this 50%
`value is not an absolute limit. Samples have been synthesized
`with quantum yields near 100%. The core/shell nanocrystals
`of highest QY are seen to luminesce strongly under room lights.
`PL, absorption spectra, and quantum yields were also
`collected for both core and core/shell nanocrystals lacking an
`organic capping group. For these, aliquots of nanocrystals in
`pyridine were removed from the reaction flask and immediately
`placed in a dark, air-free environment to eliminate oxidation of
`their bare surfaces. No dodecylamine was added. These air-
`free, uncapped nanocrystals showed absorption spectra un-
`changed from those having organic caps. The shape of their
`PL spectra, including their lack of deep trap luminescence and
`their trend of increasing intensity with shell thickness, was also
`similar to the capped nanocrystals. However, the quantum
`yields of these uncapped nanocrystals were between 15 and 20
`times lower than their capped counterparts having identical core
`(34) Nirmal, M.; Norris, D. J.; Kuno, M.; Bawendi, M. G. Phys. ReV.
`Lett. 1995, 75, 3728.
`
`Figure 3. Absorption (dashed) and photoluminescence (solid) spectra
`of two series of core/shell nanocrystals. Spectra were taken after
`successive injections of CdS stock solution. The increase in quantum
`yield and of coverage of CdS with each injection is also shown. Q.Y.:
`quantum yield of photoluminescence. ı: number of monolayers of shell
`growth. All spectra were taken at a concentration corresponding to an
`optical density (OD) of roughly 0.2 at the peak of the lowest energy
`feature in the absorption spectrum. Figure 3A: 30 Å CdSe core diameter
`series; Figure 3B: 23 Å CdSe core diameter series.
`
`was present on all samples. Throughout this paper, average
`sizes of core and core/shell nanocrystals (having thicker shells)
`were determined using HRTEM (see TEM section). Sizes of
`core/shell nanocrystals having shells thinner than approximately
`3 Å were difficult to determine accurately using TEM. These
`sizes were determined based on the amount of injected CdS
`stock used in their synthesis. This sizing method was normal-
`ized to those sizes measured by TEM for thicker shell samples.
`The core absorption spectrum is characteristic of nearly
`monodisperse nanocrystals with an absorption onset significantly
`shifted from the bulk value of wurtzite CdSe (1.74 eV). After
`refluxing the TOPO capped cores overnight in pyridine, no
`change was detected in their absorption spectra. These core
`
`4 of 11
`
`

`

`Epitaxial Growth of CdSe/CdS Core/Shell Nanocrystals
`
`J. Am. Chem. Soc., Vol. 119, No. 30, 1997 7023
`
`Figure 5. Photoluminescence excitation (solid) and absorption (dashed)
`spectra of a core/shell sample of 39 Å core diameter and 11 Å shell
`thickness. The PL was detected at 1.9 eV. The PLE spectrum closely
`tracks the absorption spectrum at all energies.
`
`Figure 6. Photostability comparison of core and core/shell nanocrystals.
`Absorption spectra of core (top) and core/shell (bottom) samples before
`(solid) and after (dashed) continuous wave irradiation at 514 nm with
`an average power of 50 mW. The core is of 39 Å diameter, and the
`core/shell is made from this core with a 7 Å shell thickness. The
`solutions were saturated with oxygen and had identical optical densities
`at the excitation wavelength at the start of the experiment.
`
`diameters and shell thicknesses. The uncapped core nanocrys-
`tals gave no detectable luminescence.
`The PLE spectra of a very dilute (first exciton peak OD of
`0.003, nanocrystal concentration of 1.4 (cid:2) 10-8 M) solution of
`core/shell nanocrystals with a 39 Å diameter core and an 11 Å
`shell is given in Figure 5. The emission was detected at the
`peak of the sample photoluminescence (1.9 eV). At these low
`concentrations, the PLE spectrum closely tracks the absorption
`spectrum at all wavelengths of excitation. At higher concentra-
`tions, the PLE spectra diverged markedly from the absorption
`spectra at higher energies.17
`Stability. Stabilities of core and core/shell nanocrystals were
`compared by a continuous wave laser photooxidation experi-
`ment. Figure 6 shows the results of an experiment on both core
`and core/shell sample of the same series. The core is of 39 Å
`diameter and the shell thickness is 7 Å. Both samples were
`dissolved in pyridine and lacked an organic capping molecule.
`The samples were saturated with oxygen throughout
`the
`experiment. The ODs of the two samples at the exciting
`wavelength (514 nm) were identical at the start. The laser power
`incident on each sample was 50 mW for approximately 2 h. A
`total of approximately 3 (cid:2) 106 photons/nanocrystal were
`
`Figure 7. Quantum yield stability of core/shell nanocrystals. Photo-
`luminescence (solid) and absorption (dashed) spectra of a core/shell
`sample of 27 Å diameter and 4 Å shell thickness before (a) and after
`(b) being left in air and room lights for 4 months.
`
`absorbed by the core/shell sample and about 50% less by the
`core sample because of its decreasing OD during the experiment.
`The before and after absorption spectra of the two samples
`are in stark contrast. Relative to its absorption spectrum before
`irradiation, the core sample after irradiation showed a 0.16 eV
`(40 nm) shift of the first exciton peak to the blue, a general
`washing out of absorption features, and an overall decrease in
`OD throughout the energy range corresponding to nanocrystal
`absorption. This is a result of photooxidation of the nanocrys-
`tal.15 In contrast, the core/shell absorption spectra showed little
`change after the photooxidation treatment. The first exciton
`peak showed no shift, and the overall OD and spectral shape
`were effectively unchanged.
`A related photooxidation experiment was done on CdSe, CdS,
`and core/shell nanocrystals dissolved in chloroform saturated
`with oxygen. The core/shell sample had dodecylamine capping
`group, while the other two were capped with a mixture of
`dodecylamine and TOPO. The optical densities of the three
`samples were within 10% of each other below 375 nm. The
`CdS sample had an absorption peak at 418 nm, the CdSe at
`582 nm, and the core/shell sample at 588 nm.
`A 200 W xenon lamp illuminated the three samples simul-
`taneously for 1 h at alarge enough distance from the lamp to
`prevent significant heating. As above, the CdS and CdSe
`nanocrystals showed significant blue shifting (from 30 to 40
`nm) of their absorption maximums and a general loss of optical
`density at all wavelengths of absorption. The core/shell
`spectrum showed neither, though this sample absorbed more
`photons than the others due to its higher optical densities at
`longer wavelengths. These qualitative results demonstrate the
`enhanced photostability of the core/shell nanocrystals as com-
`pared to the cores. Quantitative kinetic studies are of significant
`interest for the future.
`The stability of core/shell quantum yields to air and room
`lights is shown in Figure 7. Absorption spectra and quantum
`yields were measured on a sample of core/shell nanocrystals
`capped with dodecyl amine before and after being left in air
`and room lights for 4 months. Little change in absorption
`spectra and an approximate 15% decrease in quantum yield were
`detected.
`XRD. Figure 8 shows a series of core, core/shell, and CdS-
`only XRD patterns. The cores are 39 Å diameter, and the core/
`shell nanocrystals are grown from these. The CdSe core and
`
`5 of 11
`
`

`

`7024 J. Am. Chem. Soc., Vol. 119, No. 30, 1997
`
`Peng et al.
`
`Figure 8. X-ray powder diffraction patterns of a pure CdS nanocrystal
`sample of 35 Å diameter (dotted), a CdSe core sample of 39 Å diameter
`(dashed), and core/shell samples having the same 39 Å core and shell
`thicknesses of (a) 2 Å, (b) 7 Å, and (c) 11 Å. The dashed vertical lines
`represent peak positions for pure CdSe; the solid lines represent pure
`CdS. A shift in core/shell peak positions from pure CdSe to pure CdS
`is observed upon shell growth.
`
`CdS-only (high temperature TOPO synthesis; about 35 Å
`diameter) diffraction patterns exhibited peak positions corre-
`sponding to those of their bulk wurtzite crystal structures. The
`core/shell nanocrystal diffraction patterns were roughly the same
`shape (i.e., peak widths and relative peak intensities) as both
`the CdS-only and the CdSe core patterns. Upon shell growth,
`the entire spectrum systematically shifted in reflection angles
`from those characteristic of pure CdSe to those of pure CdS.
`For all samples, peak broadening due to finite size was observed
`at all reflections.
`XPS. Figure 9 shows the results of XPS collected on three
`samples: 35 Å diameter CdS, 34 Å diameter CdSe, and core/
`shell nanocrystals having a 34 Å diameter core and a shell
`thickness of 18 Å. All samples were prepared air-free in
`identical manners, and for simplification in analysis, signal was
`collected over the same energy range for all samples.
`The core/shell signal was fit (using commercial peak fitting
`software) with Gaussians representing the S2p1/2, S2p3/2, Se3p1/2,
`and Se3p3/2 peaks. For the fit, experimental values of the peak
`widths, intensity ratios between spin-orbit split states, and peak
`positions (as determined by the homogeneous nanocrystal
`signals) were used. Intensities were allowed to vary to obtain
`the best fit to the experiment. The atomic ratio of S/Se in the
`core/shell sample was determined to be 4.7/1 by dividing the
`integrated peak areas by the electronic core level sensitivity
`factors for each element. Note that the effect of the finite mean
`free path on the escape of