Photocatalytic carbon disulfide production via charge transfer quenching of quantum dots

Article abstract of DOI:10.1021/ja4083599

Carbon disulfide, a potentially therapeutic small molecule, is generated via oxidative cleavage of 1,1-dithiooxalate (DTO) photosensitized by CdSe quantum dots (QDs). Irradiation of DTO-QD conjugates leads to λ_irr independent photooxidation with a quantum yield of ～4% in aerated pH 9 buffer solution that drops sharply in deaerated solution. Excess DTO is similarly decomposed, indicating labile exchange at the QD surfaces and a photocatalytic cycle. Analogous photoreaction occurs with the O-tert-butyl ester ^tBuDTO in nonaqueous media. We propose that oxidation is initiated by hole transfer from photoexcited QD to surface DTO and that these substrates are a promising class of photocleavable ligands for modifying QD surface coordination.

Full text of DOI:10.1021/ja4083599

Communication
pubs.acs.org/JACS
Photocatalytic Carbon Disulfide Production via Charge Transfer
Quenching of Quantum Dots
Christopher M. Bernt,† Peter T. Burks,† Anthony W. DeMartino, Agustin E. Pierri, Elizabeth S. Levy,
David F. Zigler, and Peter C. Ford*
Department of Chemistry and Biochemistry, University of California, Santa Barbara, California 93106, United States
S
* Supporting Information
dictory.12 Thus, the benefit of balancing the therapeutic and
toxic effects of CS2 make the prospect of controlled, targeted
delivery an attractive goal.
ABSTRACT: Carbon disulfide, a potentially therapeutic
small molecule, is generated via oxidative cleavage of 1,1-
dithiooxalate (DTO) photosensitized by CdSe quantum
dots (QDs). Irradiation of DTO−QD conjugates leads to
λirr independent photooxidation with a quantum yield of
∼4% in aerated pH 9 buffer solution that drops sharply in
deaerated solution. Excess DTO is similarly decomposed,
indicating labile exchange at the QD surfaces and a
photocatalytic cycle. Analogous photoreaction occurs with
In these contexts, we are probing DTO−QD conjugates as
photochemical CS2 precursors. These were readily prepared by
exchanging DTO dianions for the myristate ligands (n-
−
C13H27CO2 ) originally terminating the QD surface (see
Supporting Information (SI) for procedures). Ligand exchange
is evidenced by the solubility shift from organic to aqueous
media as dianionic DTO replaces the hydrophobic myristate.
Purified conjugates show a very strong UV absorption band
t
the O-tert-butyl ester BuDTO in nonaqueous media. We
propose that oxidation is initiated by hole transfer from
photoexcited QD to surface DTO and that these
substrates are a promising class of photocleavable ligands
for modifying QD surface coordination.
(∼335 nm) nearly the same as free DTO (λmax ∼335 nm, εmax
=
1.5 × 104 M−1 cm−1 in aq. solution) and a QD exciton band
red-shifted by as much as 40 nm (Figure 1, SI Table S-1).
emiconductor quantum dots (QDs) exhibit size and
S
composition dependent optical properties, strong absorp-
tion cross sections, high photoluminescence (PL) quantum
yields, and customizable solubility through surface ligand
exchange.1 These properties position QDs as attractive
sensitizers for photodynamic therapy,2 photoactivated drug
delivery,3 solar energy conversion,4 and photocatalysis.5 Here
we describe the photocatalytic cleavage of 1,1-dithiooxalate
(DTO) to CS2 and CO2 mediated by CdSe QDs. This process
serves to “uncage” carbon disulfide (CS2), a potentially
therapeutic agent. Similar reactions are observed with the
Figure 1. Absorption spectrum of purified DTO−QD550 conjugates in
pH 9 buffer (50 mM sodium borate) solution (solid line) and that of
QD550 nanoparticles in toluene prior to ligand exchange (dashed line).
Inset: Expanded view of the exciton peak shift. (QD subscript refers to
exciton λmax before ligand exchange.)14
t
DTO ester BuDTO, and light activated cleavage of such
ligands suggests strategies for the controlled modification of
quantum dot surfaces.
Analogous red shifts in exciton absorptions have been seen
when dithiocarbamates were exchanged onto semiconductor
QD surfaces,13 and this effect was attributed to relaxation of
exciton confinement in the QD−ligand conjugate owing to
alignment of interfacial orbital energies. Lastly, the strong QD
PL was completely quenched. Aqueous solutions of the DTO−
QD conjugates are stable in the dark for at least 6 h at 37 °C
(Figure S-2) and indefinitely stable when stored in a
refrigerator.
Fragmentary evidence points toward possible therapeutic
roles for carbon disulfide.6 For example, CS2 may react with
biological amines to form dithiocarbamates, known inhibitors7
of nuclear factor κB (NF-κB),8 a crucial mediator in
inflammation induced tumor growth and progression.9 Indeed,
dithiocarbamates have been proposed as anticancer agents.10
Physiological responses to CS2 or dithiocarbamates include cell
growth, apoptosis, and neurotransmission, which are also
functions of the small molecule bioregulators NO, CO, and
H2S.11 Like each of these, CS2 is considered toxic at higher
concentrations, although epidemiological data are contra-
PL quenching in thiolate terminated QDs has been attributed
to hole transfer from the valence band to the surface bound
Received: August 12, 2013
Published: October 23, 2013
© 2013 American Chemical Society
2192
dx.doi.org/10.1021/ja4083599 | J. Am. Chem. Soc. 2014, 136, 2192−2195

Journal of the American Chemical Society
Communication
ligands.15 Such a mechanism suggested that CdSe QDs could
mediate DTO cleavage (eq 1).
hv, QD
2−
S2C−CO2 ⎯⎯⎯⎯⎯⎯→ CS2 + CO2
(1)
oxidant
In order to test this hypothesis, aerated pH 9 aqueous
solutions of the purified DTO−QD550 conjugate14 were
irradiated with 365 nm light. This led to a systematic decrease
in the 335 nm absorption band indicating that the surface-
coordinated DTO was undergoing photodecomposition. The
quantum yield for the disappearance of this band (Φdis) was
0.029 0.008 (see SI, Sec. I and J for details). Notably, this
absorbance decrease was accompanied by recovery of the QD
PL (Figure 2). At higher photochemical conversion, the QD
Figure 3. (Left) Φdis values for the photolysis of DTO−QD550
conjugates at different λirr (UV = 365, blue = 479, cyan = 498, and
green = 530 nm). (Center) Φdis values for λirr 498 nm of DTO−
QD510, DTO−QD550, and DTO−QD580. (Right) Φdis values for 498
nm photolysis of DTO−QD550 under Ar, air, and O2. All in pH 9
buffer solution.
DTO, this does not appear to be rate-limiting under normal
conditions.
If these spectral changes do indeed reflect DTO photo-
decomposition, then other ligands in solution should
coordinate to the resulting open surface sites. Thus, the QDs
should be photocatalysts for oxidative decomposition of free
DTO in solution.This reactivity was demonstrated using an
aerated pH 9 solution of DTO−QD550 to which a 2-fold excess
of the K+ salt of DTO (115 μM) had been added. Photolysis at
498 nm led to a rapid decrease in the characteristic DTO
absorbance at 335 nm (Figure S-7), and a Φdis = 0.052 0.006
was measured. Furthermore, the QDs remained in solution and
showed little PL until the DTO was nearly consumed according
to λmax 335 nm. Thus, the system is indeed photocatalytic.
Additionally, this confirms that the decomposition of DTO
frees the QD surface from dithiolcarboxylate coordination and
suggests a strategy for syntheses of new quantum dot
conjugates via incorporation of other ligands.
Figure 2. Spectral changes for pH 9 solution of DTO−QD550 upon
365 nm excitation over the course of 120 s at 20 s intervals. Inset:
Photograph showing PL return as photolysis proceeds. (See video in
the SI.)
cores precipitated, owing to loss of the water-solvating surface
ligands. Photolysis (365 nm) of a comparable solution of DTO
alone (∼70 μM) showed minimal photoinduced bleaching
(Figure S-4) and a much smaller Φdis (0.004 0.001).
The difference between the DTO−QD conjugates and DTO
itself is even more dramatic at longer irradiation wavelengths
(λirr). When buffered solutions of DTO−QD550 were excited
with light emitting diodes centered at 479, 498, or 530 nm
(Figure S-5), bleaching of the 335 nm band occurred with
comparable efficiency (Φdis = 0.045 0.005, 0.032 0.003 and
Although it has been speculated16 that photooxidative
decomposition of DTO would occur according to eq 1, CS2
has not previously been detected as a photoproduct. To address
this issue, we exhaustively photolyzed (λirr ≥ 479 nm) aerated
pH 9 solutions of DTO−QD550 in sealed, septa-capped vials.
GC-MS analysis of headspace gases by Weck Laboratories (City
of Industry, CA) showed that photolyzed samples gave
significantly larger amounts of CS2 than “dark” controls,17
thereby confirming, qualitatively, formation of CS2 as a
photoproduct.
In order to evaluate the CS2 production quantitatively, we
developed a colorimetric assay based on the rapid reaction of
CS2 with amines to form dithiocarbamates.18 The latter have
strong UV absorptions. This procedure clearly confirmed
photochemical CS2 formation, although the amount detected
was systematically about 50% that predicted by assuming eq 1
to be the only reaction leading to photobleaching of the DTO
peak at 335 nm (SI, Sec. K for details). No CS2 formation was
apparent without photolysis (see Figure S-9).
The other product predicted is CO2. This was determined by
GC analysis of the headspace before and after acidifying the
photoproduct solution to ∼pH 1, since some of the CO2 would
be trapped as bicarbonate in the pH 9 buffered solutions.
Consistent with the colorimetric analysis for CS2, the CO2
generated by photolysis of DTO−QD550 was about half that
predicted by eq 1 for the analysis before acidification and about
70% after acidification (SI, Sec. M).
0.039 0.005 for these respective λirr) to that seen for λirr
=
365 nm. Again, the PL of the QD cores was restored (Figure S-
6). In contrast, solutions of free DTO were unaffected owing to
their optical transparency at these longer λirr, so there is little
question that the QDs sensitize the photochemical decom-
position of the surface bound DTO.
Similar sensitization of photodecomposition was observed
with other DTO−QD conjugates, although there were some
differences in the quantum efficiencies. Respective Φdis values
of 0.016
0.003, 0.033
0.003, and 0.035
0.001 were
measured for 498 nm irradiation of DTO−QD510, DTO−
QD550, and DTO−QD580 (Figure 3); however, a larger data set
is needed before considering a possible systematic trend.
When the photoreactions were carried out under deaerated
conditions, Φdis values dropped precipitously. For example, 498
nm irradiation of DTO−QD550 solutions deaerated by
extensive bubbling with argon gave Φdis = 0.002
However, an atmosphere of pure O2 did not increase
0.001) significantly above that
0.001.
photoactivity (Φdis = 0.033
seen in aerobic media. Thus, while the oxidant O2 appears
necessary for efficient net photodecomposition of coordinated
2193
dx.doi.org/10.1021/ja4083599 | J. Am. Chem. Soc. 2014, 136, 2192−2195

Journal of the American Chemical Society
Communication
These products and the requirement for O2 as a coreactant
suggest that photolysis of surface coordinated DTO reversibly
leads to transient species that are trapped by the external
oxidant (Scheme 1). Alternatively, it might be argued that
with modest quantum yields. By using a colorimetric assay, it
was shown in both cases that CS2 is generated. This represents
the first demonstrated photochemical release of this potentially
therapeutic small molecule.
We propose that the photooxidation mechanism involves
QD excitation-induced hole transfer to surface bound DTO.
The efficiency is highly dependent upon the presence of
another oxidant, O2, which serves as the ultimate electron
acceptor. The reaction not only decomposes the multiple
DTOs bound to the QD surface, but the facile ligand exchange
leads to a catalytic cycle for photooxidation of excess free DTO
in solution. In addition, this offers a potential pathway for the
systematic removal or replacement of QD surface ligands and
for the photochemical syntheses of new QD−ligand conjugates.
O-Ester-DTO−QD conjugates undergo analogous photo-
decomposition. We propose that such a photocleavable ligand
may allow the controlled release of organic radicals from QD
surfaces. Ongoing studies are directed toward a better
quantification of the products as well as expanding the
photochemistry to a library of other esters in order to probe
the utility of these as photocleavable surface anchors.
Scheme 1
photoinduced two-electron transfer from DTO to the QD
occurs either simultaneously or sequentially and that “charged”
QDs are no longer photoactive until “discharged” by reacting
with O2 (SI, Scheme S-1). The formation of the DTO radicals
suggested by Scheme 1 offers a possible explanation for the
nonstoichiometric formation of CS2 and CO2, given that such
species might dimerize to give a disulfide linkage.19 Regardless
of the actual mechanism, the catalytic behavior shows that the
QD sensitized photooxidation and cleavage of DTO facilitates
removal and replacement of the surface ligands. Ongoing
studies will attempt to differentiate the mechanistic possibilities.
While DTO functionalization provides aqueous solubility,
ASSOCIATED CONTENT
■
S
* Supporting Information
Experimental details regarding the synthesis and character-
t
ization of DTO−QD and BuDTO−QD conjugates, photo-
chemical procedures and photoproduct analyses, a video
illustrating the photoreaction, and additional data. This material
is available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
ford@chem.ucsb.edu
■
t
QD conjugates of BuDTO remain soluble in organic media.
Given that CO2 formation must contribute to the driving force
for eq 1, analogous oxidative activation of an O-ester such as
tBuDTO may release a radical, e.g., eq 2.
Author Contributions
†These authors contributed equally.
hv, QD
S2C−CO2R− ⎯⎯⎯⎯⎯⎯→ CS2 + CO2 + R•
Notes
(2)
oxidant
The authors declare no competing financial interest.
Organic soluble conjugates were prepared as described in the
t
ACKNOWLEDGMENTS
SI by exchanging BuDTO for the myristate surface ligands of
■
t
QD510. The spectrum of BuDTO−QD510 displayed the strong
This work was funded by the National Science Foundation
(NSF-CHE-1058794). P.T.B. thanks ConvEne-IGERT (NSF-
DGE 0801627), C.M.B. thanks the UCSB Graduate Division,
A.E.P. and A.W.D. thank PIRE-ECCI (NSF-OISE-0968399),
and E.S.L. thanks IRES-ECCI (NSF-OISE-1065581) for
fellowships. We thank Weck Laboratories for GC-MS measure-
ments .
absorption band at 350 nm characteristic of the DTO
chromophore and an exciton peak at 523 nm red-shifted
from that of the core QD exciton peak (Figure S-8). In addition
t
the BuDTO conjugate was not luminescent, as seen for the
QD conjugates with DTO. Photolysis of tBuDTO−QD510 with
λirr 498 nm in aerated solution led to bleaching of the 350 nm
DTO band, but Φdis values were small, 0.007 in toluene and
0.0024 in chloroform. Exhaustive photolysis led to PL recovery
and a shift of the exciton peak to its original maximum (Figure
S-8), while the QDs remained soluble. The colorimetric
analytical method described above demonstrated that CS2 is
clearly formed during this photoreaction, but the yields are
even lower (∼12%) than with analogous DTO conjugates (SI,
Sec. M).
REFERENCES
■
(1) (a) Choi, C. L.; Alivisatos, A. P. Annu. Rev. Phys. Chem. 2010, 61,
369. (b) Medintz, I. L.; Uyeda, H. T.; Goldman, E. R.; Mattoussi, H.
Nat. Mater. 2005, 4, 435.
(2) Samia, A. C. S.; Chen, X.; Burda, C. J. Am. Chem. Soc. 2003, 125,
15736.
(3) (a) Burks, P. T.; Ford, P. C. Dalton Trans. 2012, 41, 13030.
(b) Sortino, S. J. Mater. Chem. 2012, 22, 301.
(4) Robel, I. n.; Subramanian, V.; Kuno, M.; Kamat, P. V. J. Am.
Chem. Soc. 2006, 128, 2385.
(5) (a) Ruberu, T. P. A.; Nelson, N. C.; Slowing, I. I.; Vela, J. J. Phys.
Chem. Lett. 2012, 3, 2798. (b) Zhao, J.; Holmes, M. A.; Osterloh, F. E.
ACS Nano 2013, 7, 4316.
(6) (a) Huang, X.; Zhou, Y.; Ma, J.; Wang, N.; Zhang, Z.; Ji, J.; Ding,
Q.; Chen, G. Environ. Toxicol. Pharm. 2012, 34, 679. (b) Munger, S.
D.; Leinders-Zufall, T.; McDougall, L. M.; Cockerham, R. E.; Schmid,
In summary, we have described the preparation of photo-
sensitive conjugates by displacing the native surface ligands on
CdSe QDs with 1,1-dithiooxalate and with O-tert-butyl-
dithiooxalate. The resulting DTO−QD conjugates are water-
soluble while the R-DTO conjugates are organic-soluble. For
both, the QD exciton bands are shifted to the red and PL is
quenched. Photolysis in aerated solutions leads to the
photocatalytic oxidative decomposition of these surface ligands
2194
dx.doi.org/10.1021/ja4083599 | J. Am. Chem. Soc. 2014, 136, 2192−2195

Journal of the American Chemical Society
Communication
A.; Wandernoth, P.; Wennemuth, G.; Biel, M.; Zufall, F.; Kelliher, K.
R. Curr. Biol. 2010, 20, 1438. (c) Tan, X.; Peng, X.; Wang, F.; Joyeux,
M.; Hartemann, P. Int. J. Hyg. Environ. Health 2002, 205, 473.
(d) Shreck, R.; Meier, B.; Mannel, D. N.; Droge, W.; Baeuerle, P. A. J.
Exp. Med. 1992, 175, 1181. (e) Valentine, W. M.; Amarnath, V.;
Graham, D. G.; Anthony, D. Chem. Res. Toxicol. 1992, 5, 254.
(7) Cvek, B.; Dvorak, Z. Curr. Pharma. Des. 2007, 13, 3155.
(8) Gilmore, T. D. Oncogene 2006, 25, 6680.
(9) Karin, M.; Greten, F. R. Nat. Rev. Immunol. 2005, 5, 749.
(10) (a) Ali, I.; Wani, W. A.; Saleem, K.; Hseih, M.-F. Polyhedron
2013, 56, 134. (b) Zhang, X.; Frezza, M.; Milacic, V.; Ronconi, L.; Fan,
Y.; Bi, C.; Fregona, D.; Dou, Q. P. J. Cellular Biochem. 2010, 109, 162.
(11) Fukuto, J. M.; Carrington, S.; Tantillo, D. J.; Harrison, J.;
Ignarro, L.; Freeman, B. A.; Chen, A.; Wink, D. A. Chem. Res. Toxicol.
2012, 25, 769.
(12) Silva, M. Birth Defects Res. B 2013, 98, 119.
(13) (a) Frederick, M. T.; Weiss, E. A. ACS Nano 2010, 4, 3195.
(b) Frederick, M. T.; Amin, V. A.; Weiss, E. A. J. Phys. Chem. Lett.
2013, 4, 634.
(14) QD samples were indexed according to the exciton λmax prior to
DTO exchange rounded to the nearest 10 nm. Estimated CdSe QD
diameters are discussed in the SI, Table S-1.
(15) Wuister, S. F.; de Mello Donego, C.; Meijerink, A. J. Phys. Chem.
B 2004, 108, 17393.
(16) (a) Zigler, D. F.; Tordin, E.; Wu, G.; Iretskii, A. V.; Cariati, E.;
Ford, P. C. Inorg. Chem. Acta 2011, 374, 261. (b) Strauch, P.; Dietzsch,
W.; Golic, L. Z. Anorg. Allg. Chem. 1997, 623, 129.
(17) In the qualitative GCMS measurements, the dark samples
showed some CS2 release, which we attribute to lengthy sample
handling under ambient light conditions during analysis by the Weck
Laboratories. However, samples deliberately photolyzed prior to
transferring to that laboratory for analysis gave GCMS signals about
twice those for the dark samples.
(18) Critchfield, F. E.; Johonson, J. Anal. Chem. 1956, 28, 430.
(19) (a) Aldana, J.; Wang, Y. A.; Peng, X. J. Am. Chem. Soc. 2001,
123, 8844. (b) Lieder, M. Electrochim. Acta 2004, 49, 1813.
2195
dx.doi.org/10.1021/ja4083599 | J. Am. Chem. Soc. 2014, 136, 2192−2195