In the Laboratory

Quantum Dots: An Experiment W

for Physical or Materials Chemistry
L. D. Winkler, J. F. Arceo, W. C. Hughes,* B. A. DeGraff,** and B. H. Augustine***
Departments of Chemistry and Physics, James Madison University, Harrisonburg, VA 22807; *[email protected],
**[email protected], ***[email protected]

Semiconductor quantum dots are nanometer-scale ag- tions. The average CdS particle size and distribution of sizes
gregates and can exhibit luminescence. They have several in- depends on the characteristics of the polymer used. The ex-
teresting properties that have led to diverse applications periment can be carried out in air at room temperature, al-
ranging from labeling biomolecules (1, 2) to harvesting solar though slightly better results are obtained at 10 ⬚C. The
energy (3). Their unique photophysical properties derive from preparations can be completed in under an hour. Students
quantum confinement of the charge carriers and serve as an can see that a luminescent material is present by simply shin-
excellent solid-state illustration of the classical particle-in-a- ing a standard 366-nm black light on the reaction flask at
box concept that is basic to both physical and materials chem- the end of the preparation. Use of a high-MW, branched poly-
istry. The synthesis and characterization of CdS quantum dots mer produces polydisperse nanoparticles of CdS with typi-
can introduce and dramatically illustrate a number of con- cal radii in the 3–4 nm range. Using the same polymer, but
cepts and techniques. This laboratory is suitable for students a different addition sequence, largely-colloidal CdS can be
in either physical or materials chemistry courses and because produced for comparison. Black light illumination shows a
it involves brightly colored luminescence often generates con- clear emission color difference between the two preparations.
siderable interest and enthusiasm.
Quantum dots can be obtained in a variety of colors, Characterization
coatings, and functional-group “handles”; a selection of dots The preparations are then characterized by their UV–
with various features is available commercially (4). The basic vis absorption and emission spectra. The absorption spectra
concepts behind the technology have been reviewed at a num- exhibit two strong bands in the near- and deep-UV regions,
ber of levels (5–7). A Google search of “quantum dots” will respectively. The near-UV band is between 320–370 nm and
generate over 250,000 hits. Despite the mature nature of can show some structure. The near-UV band’s position and
some aspects of quantum dots, it is still an active field of re- shape depends on the preparation, with nanoparticulate CdS
search where new materials and more detailed understand- showing a blue shift compared to the more colloidal CdS (Fig-
ing are vigorously pursued (8). While a wide variety of dots ure 1). The emission spectra resemble skewed Gaussian func-
have been reported, we will focus on semiconductor dots with tions with λ max dependent on the preparation. The
binary composition suspended in solution. nanoparticulate CdS has a λmax near 460–470 nm, while the
Broadly viewed, the basic synthetic approach is to com- colloidal material’s λmax is about 545–555 nm. The prepara-
bine the two dot components in a solvent in which the prod- tions are allowed to stand overnight. The emission spectra
uct material has limited solubility. The incipient product is are retaken and overlaid on the spectra taken immediately
allowed limited aggregation before being sequestered with a after synthesis. Several interesting changes have occurred in
suitable passivating coating, usually some type of polymer. both the materials. The emission λmax shows a modest blue
Constraining the aggregation number and ensuring a rela- shift on standing (∼30 nm for both preparations), the over-
tively narrow distribution of particle sizes have proved to be all emission intensity increases (∼30%), and the distribution
major synthetic challenges. Unfortunately, the more elegant
syntheses that give tight control over both particle size and
distribution require conditions that are less suitable for an
undergraduate audience (9). This experiment is adapted from
a reported synthesis that uses simple, inexpensive, and user- 1.2

friendly materials and techniques (10). Indeed, this synthe-

Absorption / Emission

1.0
sis, which can be completed in a half an hour using simple
equipment and inexpensive chemicals, is very robust.
Normalized

0.8

The Experiment 0.6

Synthesis 0.4

The core experiment consists of synthesizing nanopartic- 0.2

ulate CdS (quantum dots) and largely-colloidal (bulk) CdS,
both in fluid solution using a similar set of conditions. 0
Aliquots of Cd(NO3)2 and Na2S in water兾methanol are added 300 350 400 450 500 550 600 650 700

to a rapidly stirred methanol solution of an ethylenimine Wavelength / nm

polymer that acts to terminate aggregation of the incipient
CdS. These polymers are available in several molecular Figure 1. Normalized absorption and emission spectra for nano-
weights (MW) and in both branched and linear configura- particulate CdS (-----) and colloidal CdS(- - -).

1700 Journal of Chemical Education • Vol. 82 No. 11 November 2005 • www.JCE.DivCHED.org

 In the Laboratory

narrows for the nanoparticulate CdS as shown by loss of in- part in discrete molecules). In short, effective masses are a
tensity in the longer wavelength region. The effects of aging function of the curvature of the energy bands in the solid
can be a point of departure for more detailed discussions. state. A good discussion of the physical and mathematical
This phenomenon is rather common (10) and is attributed origins of effective masses can be found in Hummel (13).
to the removal or termination of dangling bonds at the sur- Here we have simplified eq 2 by ignoring the electron–hole
face of the particle. These dangling bonds result in traps that correlation effects. A discussion of the approximations used
allow nonradiative decay of the excited state and their termi- to obtain eq 2 and their limitations is available (14). With
nation enhances the emissive decay path. The mechanism of λmax = 465 nm, ∆Eqc ∼ 4.25 × 10᎑20 J, eq 2 will yield R ∼
termination is not completely understood. 3.04 nm. This is an approximate value since the spread of
emission wavelengths is large. However, even cursory com-
Hazards parison with the largely-colloidal CdS with λmax ∼ 530 nm
clearly shows a shift to higher transition energy with smaller
Methanol, Cd(NO3)2⭈4H2O, and Na2S⭈9H2O are all particle size. It is quite instructive for the student to con-
toxic so good practice requires use of suitable gloves for all struct a plot of λmax versus Rparticle, a task easily done with a
manipulations. The ethylenimine polymer is an irritant. mathematical-modeling program such as Mathcad. From this
Methanol solutions, even at 10 ⬚C, should be handled in a plot it is easily seen that when the average particle radius ex-
hood. Proper disposal of reaction products and excess reagents ceeds 10 nm, the effects of quantum confinement are mini-
is essential. mal. Students should note the similarity of eq 2 to the
Schrödinger expression for a particle-in-a-box modified for
Discussion a spherical container.
The mechanism for the production of emission in both
The emission energy taken at λmax for the nanoparticu- the nanoparticulate and the bulk CdS material is electron–
late CdS can be thought of as composed of the band-gap en- hole recombination. In the bulk solid state, this electron–
ergy, Ebg, and the quantum-confinement energy, ∆Eqc, which hole pair is called an exciton that can be modeled
is the energy increment due to confinement of the charge approximately as a hydrogen atom where the hole serves as
carriers in a smaller “box”. When the CdS particle absorbs a the nucleus. As with eq 2, effective masses must be used, but
photon, an electron–hole pair is created. The electron rap- one can calculate a radius of separation, aB (Bohr radius),
idly loses any excess energy to the lattice and resides at the using
lowest energy state of the conduction band. If the particle is 2
large (i.e., greater than 10 nm) the electron–hole recombi- h ε 1 1
aB = + (3)
nation can result in emission of a photon with energy corre- π e 2 me* m h*
sponding to the band gap of CdS. In practice the particles
are not perfect so that defects and impurities alter the emis- where ε = εCdSε0 is the dielectric constant and e is the charge
sion energy and a spread of photon energies is observed, usu- on the electron. me* and mh* have the same meaning as eq
ally centered around the band-gap energy. When the particles 2. Using a value of 5.7 for εCdS at optical frequencies (12a),
are small, less than 10 nm, quantum mechanics predicts a we predict ∼2.0-nm radius of separation for the exciton pair.
shift in the energy levels owing to the confinement of the This compares favorably with our particle radius of 3.04 nm.
electron–hole pair. This energy shift, ∆Eqc, increases with As the particle size approaches the exciton separation, the
decreasing particle size. Because this preparation results in a charges are confined to the particle’s surface and quantum
variety of particle sizes, ∆Eqc has a range of values. For con- confinement becomes a major effect as the charge carriers
venience we select the most probable value of ∆Eqc , which can not be swept away but can “classically” orbit.
corresponds to λmax in the emission spectrum. Thus It is also useful for students to estimate how many CdS
molecules are in the average aggregate. Using ionic radii found
hc in widely-available sources (e.g., Handbook of Chemistry and
E photon = = E bg + ∆Eqc (1) Physics) and a simple packing model, we estimate that for a
λ max
∼3-nm radius particle there are about 1200 CdS molecules
The Ebg for bulk CdS at room temperature is ∼2.42 eV = per particle, depending on the packing model.
3.85 × 10᎑19 J (11). For a typical λmax = 465 nm, ∆Eqc ∼ The polydisperse nature of this preparation can be viv-
4.25 × 10᎑20 J, which is the additional energy resulting from idly illustrated by comparison to the commercial products
quantum confinement and can be related to the most prob- such as CdSe dots. The emission spectrum for these com-
able particle radius through the expression (12) mercial quantum dots shows a full-width at half maximum
of about 30 nm while the same measure for this preparation
h2 1 1 is ∼125 nm. Despite the smaller intrinsic band gap of CdSe,
∆Eqc = 2
+ (2) these dots are more widely used and more readily available
8R me* m h*
at reasonable cost (4). CdSe dots are available in colors rang-
where R is the particle radius, me* is the effective mass of the ing from deep red to blue–green with narrow dispersion due
electron, and mh* is the effective mass of the hole. These have to tight aggregation control using a more sophisticated syn-
values of 0.19 me and 0.80 me respectively, where me is the thetic technique. The polymer-passivation technique used in
rest mass of the electron. The use of effective masses is nec- this experiment is much less efficient than that used in ref 9,
essary owing to the different behavior of electrons in the solid but this experiment uses much less hazardous conditions and
state versus discrete molecules (holes have no real counter- reagents.

www.JCE.DivCHED.org • Vol. 82 No. 11 November 2005 • Journal of Chemical Education 1701

 In the Laboratory

Additional Experiments W
Supplemental Material
There are a number of modifications and extensions that Instructions for the student laboratory writeup, notes for
can be considered. Water can be used as the only solvent, the instructor, suggestions for supplemental work, and a
but the emission intensity drops and the average particle size Mathcad template are available in this issue of JCE Online.
increases. A better substitution is 95% ethanol, which gives
particles only slightly less emissive and larger than metha- Literature Cited
nol. There are four different polyethylenimines available from
Aldrich. Each gives a slightly different dominant particle size 1. Henry, C. Chem. Eng. News 2003, June 9, 10.
and distribution. The particle suspensions are stable in air at 2. Bruchez, M., Jr.; Moronne, M.; Gin, P.; Weiss, S.; Alivisatos,
room temperature for several days. Our attempts to narrow A. P. Science 1998, 281, 2013–2019.
the distribution of particle sizes by ultra-centrifugation were 3. Hanson, T. Squeezing More Juice Out of Solar Pannels. http:/
inconclusive and this could be further pursued. As noted by /www.lanl.gov/worldview/news/releases/archive/04-040.shtml (ac-
others (15), the suspensions of nanoparticles scatter visible cessed Jul 2005).
light. This can easily be seen by eye with a 10-mW DPSS 4. Nano Materials. http://www.evidenttech.com/products/quantum-
Nd兾YAG laser at 532 nm (while any laser will do, the eye is dot-nanomaterials.php (accessed Jul 2005).
most sensitive to green light). Unfortunately, the higher-MW 5. Murphy, C. J.; Coeffer, J. L. Appl. Spectroscopy 2002, 56, 16A–
passivating polymers also scatter light, so the interpretation 26A.
of this experiment is not crisp. Finally, it is possible to make 6. Gammon, D.; Steel, D. G. Physics Today 2002, 36–41.
cast luminescent thin films of the nanoparticles in suitable 7. Bimburg, D.; Grundmann, M.; Ledentsov, N. N. Quantum
polymer hosts such as polyvinylalcohol or a sol-gel glass. Dot Heterostructures; Wiley: Chichester, UK, 1999.
These films are stable in air for at least a week. 8. van Stark, W. G. J. H. M; Frederix, P. L. T. M.; van den Heuvel,
D. J.; Bol, A. A.; van Lingen, J. N.; de Mello Donega, C.;
Conclusion Gerritsen, H. C.; Meijerink, A. J. Fluorescence 2002, 12, 69–76.
9. Kippeny, T.; Swafford, L. A.; Rosenthal, S. J. J. Chem. Educ.
This experiment has been used successfully for three years 2002, 79, 1094–1100.
in a senior physical chemistry laboratory. It follows the tra- 10. Huang, J.; Sookal, K.; Murphy, C. J.; Ploehn, H. J. Chem.
ditional particle-in-a-box cyanine dye experiment and serves Mater. 1999, 11, 3595–3602.
to both reinforce the basic concept of quantum confinement 11. CRC Handbook of Chemisty and Physics, 71st ed.; Lide, D. R.,
as well as providing a useful bridge between the molecular Ed.; CRC Press: Boca Raton, FL, 1991.
and solid-state world. 12. (a) Brus, L. E. J. Chem. Phys. 1983, 79, 5566–5571. (b) Brus,
L. E. J. Chem. Phys. 1984, 80, 4403–4409.
Acknowledgments 13. Hummel, R. E. Electronic Properties of Materials, 3rd ed.;
Springer Verlag: New York, 2001; pp 70–73.
The authors gratefully acknowledge the support of the 14. Steigerwald, M. L.; Brus, L. E. Acctn. Chem. Res. 1990, 23, 183–
National Science Foundation (Grants DMR-0071717, 188. Wang, Y.; Herron, N. J. Phys. Chem. 1991, 95, 525–532.
DMR-0097449, CHE-97-26999 ). Additionally, the good 15. Sooklal, K.; Hanus, L. H.; Ploehn, H. J.; Murphy, C. J. Adv.
humor of three classes of Chem. 438L is appreciated. Mater. 1998, 10, 183–188.

1702 Journal of Chemical Education • Vol. 82 No. 11 November 2005 • www.JCE.DivCHED.org