Mechanistic study of the synthesis of CdSe nanocrystals: Release of selenium

Article abstract of DOI:10.1021/ja209246z

We outline a reaction pathway for the cleavage of the P=Se bond in trialkylphosphine selenide during the synthesis of CdSe nanocrystals. The reaction between cadmium carboxylate and trimethylphosphine selenide in the presence of an alcohol produces alkoxytrimethylphosphonium (2). Control experiments and density functional theory calculations suggested that the cleavage of the P=Se bond is initiated by nucleophilic attack of carboxylate on a Cd²⁺-activated phosphine selenide to produce an acyloxytrialkylphosphonium intermediate (1), which is converted to 2 in the presence of an alcohol.

Full text of DOI:10.1021/ja209246z

Communication
pubs.acs.org/JACS
Mechanistic Study of the Synthesis of CdSe Nanocrystals: Release of
Selenium
Raul García-Rodríguez and Haitao Liu*
́
Department of Chemistry, University of Pittsburgh, Pittsburgh, Pennsylvania 15260, United States
S
* Supporting Information
The previously proposed nucleophilic attack of carboxylate
on Cd2+-activated phosphine selenide should produce 1 upon
cleavage of the P−Se linkage (Scheme 1). Compound 1 has
ABSTRACT: We outline a reaction pathway for the
cleavage of the PSe bond in trialkylphosphine selenide
during the synthesis of CdSe nanocrystals. The reaction
between cadmium carboxylate and trimethylphosphine
selenide in the presence of an alcohol produces
alkoxytrimethylphosphonium (2). Control experiments
and density functional theory calculations suggested that
the cleavage of the PSe bond is initiated by nucleophilic
attack of carboxylate on a Cd2+-activated phosphine
selenide to produce an acyloxytrialkylphosphonium
intermediate (1), which is converted to 2 in the presence
of an alcohol.
Scheme 1. Cleavage of the PSe bond
emiconductor nanocrystals have made a great impact in
S
numerous applications, including lasers,1−3 photovoltaic
cells,4,5 biomarkers,6,7 and light-emitting diodes.8−10 The
synthesis of group II−VI nanocrystals from trialkylphosphine
chalcogenides (R3PE; E = S, Se, Te) and metal salts has been
used for almost 20 years.11 Over the years, a large number of
recipes have been developed, and a good control over size and
shape has been achieved.12−14 More recently, a number of
studies have aimed at understanding the chemical mechanism
of nanocrystal formation15−21 and in particular the mechanism
by which the monomers are generated.22−25 These studies have
been motivated by the need for better control of the
nanocrystals synthesis and potentially will allow new synthesis
recipes developed by rational design.
been suggested as the key reactive intermediate in many other
synthetically important reactions, including the Mitsunobu
reaction,26−29 the synthesis of amides30 and thioesters,31 and a
number of peptide coupling reactions.32,33 Its structural
characterization, however, has not been well described in the
literature because of its instability, which has made its isolation
impossible.28,29,34
We prepared compound 1a by reacting a substoichiometric
amount of oleic acid or cadmium oleate with hexamethylox-
odiphosphonium triflate in CDCl3.35 While compound 1a was
too reactive to be isolated in pure form, its structure could be
unambiguously identified through solution-phase NMR spec-
troscopy and high-resolution mass spectrometry. The 31P{1H}
NMR spectrum of 1a gave a singlet at 98.8 ppm;28 its 13C{1H}
NMR spectrum showed a doublet at 169.5 ppm for the
carbonyl carbon with the expected P−C coupling constant
Among the two precursors, the trialkylphosphine chalcoge-
nide (R3PE) is special in that the chalcogen atom is
covalently bound to the phosphorus atom. The PE bond
must be cleaved to “release” the chalcogen atom for nanocrystal
growth. Previous studies have suggested that this bond cleavage
is initiated by nucleophilic attack of carboxylate on the
phosphine chalcogenide.15,23,24 However, the details of this
important PE bond cleavage are largely unknown.
1
(2JP−C = 10 Hz). The H NMR spectrum showed, among the
signals expected for the oleyl fragment and the P(CH3)3
fragment, a resonance for the α-CH2 group adjacent to the
carbonyl that was ∼0.3 ppm downfield relative to that of oleic
acid; this CH2 peak also showed a cross-peak with the
Herein we report a mechanistic study of the synthesis of
CdSe nanocrystals from cadmium carboxylate and trimethyl-
phosphine selenide. We identified acyloxytrialkylphosphonium
(1) as a key intermediate following cleavage of the PSe bond.
A kinetics study suggested that the rate-limiting step of the
CdSe nanocrystal synthesis is or precedes the PSe bond
cleavage. Density functional theory (DFT) calculations
indicated that the PSe bond is cleaved in two steps:
nucleophilic attack by carboxylate on a Cd2+-activated
phosphine selenide followed by proton transfer from carboxylic
acid to the selenium atom.
1
phosphorus signal (98.8 ppm) in the H−31P heteronuclear
multiple bond correlation (HMBC) spectrum. High-resolution
spectrometry of the crude reaction mixture showed an [M +
H]+ peak at m/z 507.2474 (calcd 507.2521). To the best of our
knowledge, this is the most complete characterization of an
acyloxyphosphonium species reported to date [see the
Supporting Information (SI)].
Received: September 30, 2011
Published: January 5, 2012
© 2012 American Chemical Society
1400
dx.doi.org/10.1021/ja209246z | J. Am. Chem.Soc. 2012, 134, 1400−1403

Journal of the American Chemical Society
Communication
Upon its addition to a mixture of Cd(OA)2 (OA = oleate)
and trimethylphosphine selenide, compound 1a was immedi-
ately consumed, and trimethylphosphine oxide and oleic acid
anhydride were produced (see the SI).23 The rapid conversion
of 1a in this reaction mixture suggests that if this compound is
formed as an intermediate during the synthesis of CdSe
nanocrystals, its direct observation would be difficult. Indeed,
the reaction mixture of Cd(OA)2 and trimethylphosphine
selenide only showed 31P NMR signals from trimethylphos-
phine selenide and trimethylphosphine oxide; no other 31P
peak was observed. Similar observations were also made when
triphenylphosphine selenide or methyldiphenylphosphine
selenide was used as the selenium precursor for nanocrystal
synthesis (see the SI).
Failing to observe this key intermediate directly, we then
sought indirect ways to detect its formation during the
synthesis of CdSe nanocrystals. The strong electrophilicity of
1a suggests that it may be trapped by reaction with a suitable
nucleophile. We found that the reaction between 1a and an
alkyl alcohol resulted in the immediate and quantitative
transformation of 1a to alkoxytrimethylphosphonium (2),
which is considerably more stable than 1a (see the SI). This
result suggests that trapping of 1 in the synthesis of CdSe
nanocrystals may be achieved by addition of alcohol to the
reaction mixture.36
peaks in the 31P{1H} spectrum; no additional peak was
observed (see the SI).
The conversion of 1 to 2 was quantitative within several
minutes of addition of alcohol. However, the formation of 2
during the CdSe nanocrystal synthesis was apparently limited
by the much slower decay of phosphine selenide, which is
known to occur at the same rate as the formation of CdSe
nanocrystals. In addition, the addition of alcohol did not
significantly change the decay rate of phosphine selenide (see
the SI). These facts indicate that the rate-limiting step of the
CdSe nanocrystal synthesis must be or precede the complete
cleavage of the P−Se linkage.
We found that compound 2 was also formed when an
alcohol was added to the synthesis of CdS or CdTe
nanocrystals using trimethylphosphine sulfide or telluride as
the precursor, respectively (see the SI). These results suggest
that a common mechanism is responsible for the release of the
chalcogen atom from trialkylphosphine chalcogenide precur-
sors.
The reaction between Cd(OA)2 and a secondary phosphine
selenide showed a very different behavior. The crude reaction
mixture of diphenylphosphine selenide and Cd(OA)2 gave a
31P NMR peak at 98.6 ppm. We assigned this peak to
oleyloxydiphenylphosphine on the basis of 1H−13C and
1H−31P HMBC experiments (see the SI). In this case, however,
we also observed 31P peaks at 77.2 and −15.0 ppm that were
tentatively assigned to oleyloxydiphenylphosphine selenide and
tetraphenyldiphosphine, respectively.22,37 Diphenylphosphine
selenide was consumed within 30 min at room temperature
with very little formation of diphenylphosphine oxide. Similar
observations were recently reported by Krauss et al.22 and
Owen et al.37 for nanocrystal syntheses using the same
selenium precursor. We believe that a different mechanism is
responsible for the cleavage of the PSe bond in secondary
phosphine selenides.
We investigated the effect of alcohol on the synthesis of
CdSe nanocrystals. In the presence of ROH [R = −CH3,
−CH(CH3)2], we observed a new 31P NMR signal at 99.9 ppm
(R = −CH3) or 93.0 ppm [R = −CH(CH3)2] from the crude
reaction mixture (Figure 1). Detailed analysis showed that this
We modeled the PSe bond-cleavage process during the
reaction between Cd(OAc)2 and Me3PSe using DFT/B3LYP
calculations.38−41 We used a triple-ζ-quality basis set for Cd, Se,
and P and the 6-31+G* basis set for O, C, and H.42−49 The
electrostatic solvation energy was taken into account using a
polarizable continuum model.50 Our calculations showed that
the formation of 1 is consistent with nucleophilic attack of
carboxylate on phosphine selenide.
Scheme 2 shows the calculated reaction pathway.51 It was
found that Me3PSe binds to Cd(OAc)2 to give complex 3
Figure 1. 31P{1H} spectrum of the reaction between Cd(OA)2 and
Me3PSe (A) without and (B) with added 2-propanol.
Scheme 2. DFT-Calculated Reaction Pathway for PSe
Cleavage
new peak originated from the expected alkoxytrimethylphos-
phonium 2 (see the SI). For example, the P−OCH3 fragment
1
of 2a gave a doublet at 3.89 ppm (3JP−H = 12 Hz) in the H
NMR spectrum and a doublet at 55.8 ppm (2JP−C = 8 Hz) in
the 13C{1H} NMR spectrum; signals from the −P(CH3)3
2
portion were located at 2.17 ppm (doublet, JP−H =14 Hz) in
1
1
the H spectrum and 11.39 ppm (doublet, JP−C = 66 Hz) in
the 13C{1H} spectrum. These 1D NMR assignments were
confirmed by 1H−13C heteronuclear multiple quantum
coherence (HMQC) and H−31P HMBC experiments. The
1
final confirmation of our structural assignment came from the
addition of authentic samples of 2a to the crude reaction
mixture. Such additions resulted in an increase of the expected
1401
dx.doi.org/10.1021/ja209246z | J. Am. Chem.Soc. 2012, 134, 1400−1403

Journal of the American Chemical Society
Communication
[ΔH = −14.9 kcal mol−1, ΔS = −37.0 cal mol−1 K−1; all
calculated thermodynamic quantities are referenced to Cd-
(OAc)2, Me3PSe, and HOAc]. The Me3PSe moiety in
complex 3 can be attacked by an acetate of the same complex,
going through a transition state featuring a six-membered ring
(TS1, ΔH⧧ = 24.0 kcal mol−1, ΔS⧧ = −49.1 cal mol−1 K−1) to
give intermediate 4 (ΔH = 15.7 kcal mol−1, ΔS = −38.6 cal
mol−1 K−1). The nucleophilic attack on the phosphorus atom
converts the tetravalent phosphorus atom in complex 3 into a
pentavalent one in 4. Essentially, this reaction partially breaks
the PSe double bond into a P−Se single bond and forms the
initial Cd−Se bond. Intermediate 4 binds an acetic acid to give
5 (ΔH = 6.7 kcal mol−1, ΔS = −75.3 cal mol−1 K−1). The Se
atom of 5 accepts the acid proton from the incoming acetic
acid, going through a proton-transfer transition state (TS2, ΔH⧧
= 11.3 kcal mol−1, ΔS⧧ = −79.9 cal mol−1 K−1) to break the P−
Se linkage completely, giving 1b and 6. We suggest that
complex 6 may be considered as the precursor for the
subsequent growth of CdSe clusters and nanocrystals.
ASSOCIATED CONTENT
* Supporting Information
Details of experimental procedures and DFT calculations and
additional figures and tables. This material is available free of
charge via the Internet at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
hliu@pitt.edu
■
ACKNOWLEDGMENTS
■
This research was supported in part by Computational
Resources on PittGrid (www.pittgrid.pitt.edu) and by the
National Science Foundation through TeraGrid resources
provided by the Texas Advanced Computing Center under
Grant TG-CHE100132; we specifically acknowledge the
assistance of Hang Liu. We thank Dr. Jeffery C. Grossman
for providing computational resources at the early stage of this
́ ́
work. R.G.-R. acknowledges Fundacion Ramon Areces for a
postdoctoral fellowship.
Two transition states were identified before the complete
PSe bond cleavage: one for the nucleophilic attack of acetate
on the phosphorus atom (TS1, ΔH⧧ = 24.0 kcal mol−1, ΔS⧧ =
−49.1 cal mol−1 K−1) and the other for the proton transfer from
carboxylic acid to the Se atom (TS2, ΔH⧧ = 11.3 kcal mol−1,
ΔS⧧ = −79.9 cal mol−1 K−1). The calculated activation
parameters of both transition states are in reasonable agreement
REFERENCES
■
(1) Klimov, V. I.; Bawendi, M. G. MRS Bull. 2001, 26, 998.
(2) Klimov, V. I.; Mikhailovsky, A. A.; Xu, S.; Malko, A.;
Hollingsworth, J. A.; Leatherdale, C. A.; Eisler, H.; Bawendi, M. G.
Science 2000, 290, 314.
(3) Eisler, H.-J.; Sundar, V. C.; Bawendi, M. G.; Walsh, M.; Smith, H.
I.; Klimov, V. Appl. Phys. Lett. 2002, 80, 4614.
(4) Huynh, W. U.; Dittmer, J. J.; Alivisatos, A. P. Science 2002, 295,
2425.
(5) Gur, I.; Fromer, N. A.; Geier, M. L.; Alivisatos, A. P. Science 2005,
310, 462.
(6) Alivisatos, P. Nat. Biotechnol. 2004, 22, 47.
(7) Han, G.; Mokari, T.; Ajo-Franklin, C.; Cohen, B. E. J. Am. Chem.
Soc. 2008, 130, 15811.
(8) Achermann, M.; Petruska, M. A.; Koleske, D. D.; Crawford, M.
H.; Klimov, V. I. Nano Lett. 2006, 6, 1396.
(9) Colvin, V. L.; Schlamp, M. C.; Alivisatos, A. P. Nature 1994, 370,
354.
(10) Coe, S.; Woo, W.-K.; Bawendi, M.; Bulovic, V. Nature 2002,
420, 800.
(11) Murray, C. B.; Norris, D. J.; Bawendi, M. G. J. Am. Chem. Soc.
with the experimental value (ΔH⧧
= 14.8 0.7 kcal mol−1,
exptl
ΔS⧧exptl = −34.7 1.9 cal mol−1 K−1).23 More interestingly, the
relative magnitudes of the two activation free energies are
dependent on the temperature. Below ∼412 K, the activation
free energy of the nucleophilic-attack transition state (TS1) is
higher; above ∼412 K, that of the proton-transfer transition
state (TS2) is higher. We note that this analysis should be
considered qualitative and not quantitative: the DFT/B3LYP
method we used here has an average deviation of 4−5 kcal
mol−1 (up to 19 kcal mol−1) in enthalpy calculations52 and
tends to grossly overestimate the entropy penalty for solution-
phase reactions.53 Nevertheless, the calculation predicts that the
activation parameters of the CdSe nanocrystal synthesis are
temperature-dependent. In addition, one also expects a primary
hydrogen kinetic isotope effect under conditions where the
proton-transfer step is rate-limiting. Ongoing research in our
lab is focused on validating these predictions.
1993, 115, 8706.
(12) Peng, Z. A.; Peng, X. J. Am. Chem. Soc. 2002, 124, 3343.
(13) Li, J. J.; Wang, Y. A.; Guo, W.; Keay, J. C.; Mishima, T. D.;
Johnson, M. B.; Peng, X. J. Am. Chem. Soc. 2003, 125, 12567.
(14) Dayal, S.; Kopidakis, N.; Olson, D. C.; Ginley, D. S.; Rumbles,
G. J. Am. Chem. Soc. 2009, 131, 17726.
(15) Owen, J. S.; Chan, E. M.; Liu, H.; Alivisatos, A. P. J. Am. Chem.
Soc. 2010, 132, 18206.
(16) Viswanatha, R.; Amenitsch, H.; Sarma, D. D. J. Am. Chem. Soc.
2007, 129, 4470.
In conclusion, we have proposed a detailed mechanism for
the cleavage of the PSe bond in trialkylphosphine selenide
during the synthesis of CdSe nanocrystals. Our findings suggest
that this bond cleavage is initiated by nucleophilic attack of
carboxylate on a Cd2+-activated phosphine selenide. This attack
is followed by proton transfer from carboxylic acid to the Se
atom to break the P−Se linkage completely, forming the initial
Cd−Se bond. A signature of the PSe bond cleavage is the
formation of an acyloxytrialkylphosphonium ion, which could
be trapped using an alkyl alcohol. Kinetic experiments showed
that the rate-limiting step of the CdSe nanocrystal synthesis is
or precedes the formation of this acyloxytrialkylphosphonium
intermediate. We hope that these results will lead to a better
understanding of the synthesis of group II−VI nanocrystals and
the rational design of new synthetic methodologies.
(17) Bullen, C. R.; Mulvaney, P. Nano Lett. 2004, 4, 2303.
(18) Rempel, J. Y.; Bawendi, M. G.; Jensen, K. F. J. Am. Chem. Soc.
2009, 131, 4479.
(19) Kloper, V.; Osovsky, R.; Kolny-Olesiak, J.; Sashchiuk, A.;
Lifshitz, E. J. Phys. Chem. C 2007, 111, 10336.
(20) Joo, J.; Pietryga, J. M.; McGuire, J. A.; Jeon, S.-H.; Williams, D.
J.; Wang, H.-L.; Klimov, V. I. J. Am. Chem. Soc. 2009, 131, 10620.
(21) Jiang, Z.-J.; Kelley, D. F. ACS Nano 2010, 4, 1561.
(22) Evans, C. M.; Evans, M. E.; Krauss, T. D. J. Am. Chem. Soc.
2010, 132, 10973.
(23) Liu, H.; Owen, J. S.; Alivisatos, A. P. J. Am. Chem. Soc. 2007,
129, 305.
(24) Steckel, J. S.; Yen, B. K.; Oertel, D. C.; Bawendi, M. G. J. Am.
Chem. Soc. 2006, 128, 13032.
1402
dx.doi.org/10.1021/ja209246z | J. Am. Chem.Soc. 2012, 134, 1400−1403

Journal of the American Chemical Society
Communication
(25) Allen, P. M.; Walker, B. J.; Bawendi, M. G. Angew. Chem., Int. Ed.
2010, 49, 760.
(26) McNulty, J.; Capretta, A.; Laritchev, V.; Dyck, J.; Robertson, A.
J. Angew. Chem., Int. Ed. 2003, 42, 4051.
(27) McNulty, J.; Capretta, A.; Laritchev, V.; Dyck, J.; Robertson, A.
J. J. Org. Chem. 2003, 68, 1597.
(28) Ahn, C.; Correia, R.; DeShong, P. J. Org. Chem. 2002, 67, 1751.
(29) Schenk, S.; Weston, J.; Anders, E. J. Am. Chem. Soc. 2005, 127,
12566.
(30) Hans, J. J.; Driver, R. W.; Burke, S. D. J. Org. Chem. 2000, 65,
2114.
(31) Gopinath, P.; Ravindran, S. V.; Chandrasekaran, S. J. Org. Chem.
2009, 74, 6291.
(32) Klausner, Y. S.; Bodansky, M. Synthesis 1972, 453.
(33) Gawne, G.; Kenner, G. W.; Sheppard, R. C. J. Am. Chem. Soc.
1969, 91, 5669.
(34) Ohmori, H.; Maeda, H.; Kikuoka, M.; Maki, T.; Masui, M.
Tetrahedron 1991, 47, 767.
(35) Hendrickson, J. B.; Hussoin, M. S. J. Org. Chem. 1989, 54, 1144.
(36) It is important to note that the alkyl alcohol, being a weaker
nucleophile than carboxylate, does not attack Me3PSe directly to
produce the alkyloxyphosphonium ion. See the SI for details.
(37) Cossairt, B. M.; Owen, J. S. Chem. Mater. 2011, 23, 3114.
(38) Granovsky A. A. Firefly, version 7.1.G; http://classic.chem.msu.
su/gran/firefly/index.html (accessed Sept 30, 2011).
(39) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J. Phys. 1980, 58, 1200.
(40) Lee, C. T.; Yang, W. T.; Parr, R. G. Phys. Rev. B 1988, 37, 785.
(41) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
(42) Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J. J.
Phys. Chem. 1994, 98, 11623.
(43) Hehre, W. J.; Ditchfield, R.; Pople, J. A. J. Chem. Phys. 1972, 56,
2257.
(44) Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta 1973, 28, 213.
(45) Clark, T.; Chandrasekhar, J.; Spitznagel, G. W.; Schleyer, P. v. R.
J. Comput. Chem. 1983, 4, 294.
(46) Woon, D. E.; Dunning, T. H. J. Chem. Phys. 1993, 98, 1358.
(47) Feller, D. J. Comput. Chem. 1996, 17, 1571.
(48) Martin, J. M. L.; Sundermann, A. J. Chem. Phys. 2001, 114, 3408.
(49) Schuchardt, K. L.; Didier, B. T.; Elsethagen, T.; Sun, L. S.;
Gurumoorthi, V.; Chase, J.; Li, J.; Windus, T. L. J. Chem. Inf. Model.
2007, 47, 1045.
(50) Tomasi, J.; Persico, M. Chem. Rev. 1994, 94, 2027.
(51) We found that the binding of HOAc to Cd(OAc)2 is not
favored at the temperature commonly used for CdSe nanocrystal
synthesis. For the reaction Cd(OAc)2 + HOAc → Cd(HOAc)(OAc)2,
our DFT calculations gave ΔH = −13.1 kcal mol−1 and ΔS = −36.4 cal
mol−1 K−1, indicating that ΔG > 0 at T > 360 K. For this reason, we
chose Cd(OAc)2 as the model cadmium precursor.
(52) Zhang, Y.; Xu, X.; Goddard, W. A. III. Proc. Natl. Acad. Sci.
U.S.A. 2009, 106, 4963.
(53) Fang, H.; Lu, G.; Zhao, L.; Li, H.; Wang, Z.-X. J. Am. Chem. Soc.
2010, 132, 12388.
1403
dx.doi.org/10.1021/ja209246z | J. Am. Chem.Soc. 2012, 134, 1400−1403