Quantum dot (QD), or semiconductor nanocrystal, thin films are being explored for making solution-processable devices due to their size- and shape-tunable bandgap and discrete higher energy electronic states. While DNA has been extensively used for the self-assembly of nanocrystals, it has not been investigated for the simultaneous conduction of multiple energy charges or excitons via exciton shelves (ES) formed in QD-DNA nano-bioelectronic thin films. Here, we present studies on charge conduction through exciton shelves, which are formed via chemically coupled QDs and DNA, between electronic states of the QDs and the HOMO-LUMO levels in the complementary DNA nucleobases. While several challenges need to be addressed in optimizing the formation of devices using QD-DNA thin films, a higher charge collection efficiency for hot-carriers and our detailed investigations of charge transport mechanism in these thin films highlight their potential for applications in nano-bioelectronic devices and biological transducers.

Quantum dots (QDs) are semiconductor nanocrystallites that allow for solution processable device fabrication, which utilize the size-dependent multiple quantum-confined electronic states of the materials for capturing broadband radiation.1–6 Especially, the size- and shape-tunable molecule-like energy levels of QDs are being explored for utilizing bandedge and hot-carriers in photovoltaic,4 photodetection,5,6 and photocatalytic devices.7 While short cooling times and charge recombination at QD surface traps limit improvement in device efficiency,8 several strategies are being examined to improve the conduction of photogenerated charges, including exploration of band-like conduction in self-assembled superlattice structures instead of glassy QD films.9,10 DNA-mediated self-assembly provides a route for arranging these semiconductor nanostructures in desired three-dimensional architectures, using bottom-up fabrication.10–13 Furthermore, the highest occupied and lowest unoccupied molecular orbitals (HOMO-LUMO) states of complementary nucleobases (e.g., adenine (A) and thymine (T)) also provide a pathway for the simultaneous conduction of different energy charge carriers (bandedge and hot-carriers without cooling) from the QD states energetically aligned with the DNA levels, termed exciton shelves (ES).6 Therefore, we studied the transport of dark and photogenerated charges through exciton shelves formed in QD-DNA nano-bioelectronic thin films.

The energy level alignment of different cadmium chalcogenide QDs with DNA was tested using scanning tunneling spectroscopy (STS) measurements. Hybridized dsDNA nucleotides (double-stranded DNA) show the energetic separation of the HOMO-LUMO levels of the constituent nucleotides (A and T here),6,14 arising from their different degrees of conjugation (Figure 1(a)). Chemically coupling dsDNA to CdSe QDs (using a thiolated end group on dsDNA, CdSe-dsDNA constructs) reveals a mismatch between the DNA HOMO states and the valence band (VB) of the QDs, leading to the formation of intra-bandgap states (Figure 1(a)), likely causing charge trapping and recombination. A similar mismatch and formation of undesirable intragap states are also seen with CdS-dsDNA constructs (Figure S314). In contrast, the well-matched energy levels of the CdTe conduction band (CB)-VB and respective DNA LUMO-HOMO states (Figure 1(a)) form a two-level system for charge transfer with a lower energy exciton shelf for conducting band-edge electrons, and a high energy exciton shelf for hot electrons (illustrated in Figure 1(b)). Changing the size of CdTe QDs primarily shifts the CB states (due to the heavy-hole6) while leaving the VB states virtually unchanged, leading to the formation of exciton-shelves with different electron injection efficiencies (Figure S3 shows the change activation energy for electron transport, and Table S1 lists device efficiencies with different-sized CdTe QDs14). These energetically distinct pathways can allow the simultaneous conduction of different energy charges without cooling, since the energy states of the nucleobases do not mix (no convolution or mixing of states observed experimentally or computationally).

FIG. 1.

(a) STS measurements of CdSe and CdTe QD-dsDNA nanobio-hybrids, with the probability density curves describing the DNA DOS overlaid in green.14 The inset shows an STM micrograph showing quantum dots on a ITO surface (scale bar is 20 nm). (b) Schematic showing the energy level alignment between the conduction/valence bands of CdSe and CdTe QDs and the HOMO/LUMO levels of poly(adenine-thymine) oligonucleotides. The exciton shelves are represented as energetic gradients to represent the distribution in conformational entropy of the DNA nucleotides. The red X on CdSe indicates a hole trap (non-radiative relaxation). Energetic separations between the CB/VB of CdTe (black), 1st exciton shelf (red), and 2nd exciton shelf (blue) are marked. (c) Schematic showing the structure of the thin film device used to characterize the photoresponse of QD-DNA materials. The expanded view of the QD-DNA layer shows the charge conduction pathways for different energy electrons and holes.

FIG. 1.

(a) STS measurements of CdSe and CdTe QD-dsDNA nanobio-hybrids, with the probability density curves describing the DNA DOS overlaid in green.14 The inset shows an STM micrograph showing quantum dots on a ITO surface (scale bar is 20 nm). (b) Schematic showing the energy level alignment between the conduction/valence bands of CdSe and CdTe QDs and the HOMO/LUMO levels of poly(adenine-thymine) oligonucleotides. The exciton shelves are represented as energetic gradients to represent the distribution in conformational entropy of the DNA nucleotides. The red X on CdSe indicates a hole trap (non-radiative relaxation). Energetic separations between the CB/VB of CdTe (black), 1st exciton shelf (red), and 2nd exciton shelf (blue) are marked. (c) Schematic showing the structure of the thin film device used to characterize the photoresponse of QD-DNA materials. The expanded view of the QD-DNA layer shows the charge conduction pathways for different energy electrons and holes.

Close modal

To study charge transport through ES in QD-DNA thin films, we constructed a simple thin film device architecture which allows optoelectronic characterization of these nano-bio hybrid materials (Figure 1(c)). Spherical dodecylphosphonic acid/trioctylphoshine oxide capped CdTe QDs with a diameter of 7 ± 1 nm were synthesized according to published methods, then ligand exchanged with thioglycerol and thiol-terminated DNA in aqueous media.14,15 The thin film devices were fabricated by spin-coating a layer of TiO2 nanoparticles on indium-tin-oxide coated glass substrate, followed by annealing at 450 °C for 10 min and treatment with 50 mM TiCl4 at 70 °C for 30 min, preceding a second annealing at the same conditions. To deposit the QD-DNA film, 5′ and 3′ thiolated DNA coated QDs were mixed in 10 mM MgCl2 with the linking DNA strand, and drop cast onto an area of the substrate defined by O3 treatment, and allowed to adsorb for 1 h under humid conditions, followed by a vacuum treatment to remove excess solvent. Excess salts were removed with a brief wash of 90% ethanol. The film was annealed at 60 °C for 1 h under humid conditions, with a slow cooling rate of 30 K/h. Any remaining water was removed by vacuum drying at 80 °C for 40 min. The gold contact was deposited on the QD-DNA film via vacuum evaporation at rates between 0.4 and 1.5 Å/s and was annealed at 80 °C for 40 min.14,15

The charge transport properties of electrons and holes in dark (electrically injected) through the QD-DNA thin films were probed by monitoring the current-voltage (I-V) characteristics at different temperatures. I-V characteristics of the thin films were measured in an Advanced Research Systems vacuum cryostat chamber, using a Keithley 2612 A Source Meter. The barriers (ϕ) for electron and hole transport were calculated from the change in conductivity with temperature

I=CVeqϕkT
(1)

where C is the constant, q is the elementary charge, and k is the Boltzmann constant (Figure 2(a)), and Richardson's plot

I=AGT2eqϕkT
(2)

where AG is the material dependent constant (Figure 2(b)).16 Using both methods, the barrier for electron transport was measured as 0.35 ± 0.03 eV, which matched well with our STS measurements (Figure 1(b)). The energetic spacing between the CB of CdTe QD and the LUMO level of adenine nucleobases was measured as ∼0.3–0.4 eV using our single QD-DNA STS data. The hole transport barrier was much lower (∼0.01–0.1 eV) and is also consistent with the STS measurements showing a close overlap between the CdTe QD VB and the HOMO levels of the dsDNA.

FIG. 2.

(a) Determination of the barrier heights for charge conduction for the CdTe QD-DNA thin film using a dark conductance as a function of temperature. (b) Richardson plot.

FIG. 2.

(a) Determination of the barrier heights for charge conduction for the CdTe QD-DNA thin film using a dark conductance as a function of temperature. (b) Richardson plot.

Close modal

To further understand the charge conduction mechanism for electrically injected charge carriers in QD-DNA thin films, we used a Fowler-Nordheim analysis to identify the different regimes of charge transport under application of external bias (Figure 3(a)).17 From this analysis, two different regions can be easily identified: a high bias region where I·V−2 is independent of the applied bias, and a linear region with low applied bias where ln[I/V2] is proportional to ln[1/V]. The current dependence on external bias in high bias region (IV2) is characteristic of space-charge limited current (SCLC)

I=8εμV29d3+VR
(3)

where μ is the mobility, d is the inter-particle spacing, and ε is the dielectric constant.18–21 The linear dependence of current on voltage (IV) at low bias is indicative of Ohmic transport, which is also common to SCLC conduction. While these charge transport mechanisms are consistent with our observations and the proposed model for charge transport through QD-DNA exciton shelves, we also investigated other possibilities for conduction through the QD-DNA thin films.

FIG. 3.

Analysis of the charge transport mechanism in the CdTe-dsDNA thin film. (a) Fowler-Nordheim plot of dark I-V data with characteristic regions labeled. (b) Various charge transport mechanisms fitted to a dark I-V characteristic. Poor model fits are shown as dashed lines while space-charge limited current, as the best fit, is shown as a solid line.

FIG. 3.

Analysis of the charge transport mechanism in the CdTe-dsDNA thin film. (a) Fowler-Nordheim plot of dark I-V data with characteristic regions labeled. (b) Various charge transport mechanisms fitted to a dark I-V characteristic. Poor model fits are shown as dashed lines while space-charge limited current, as the best fit, is shown as a solid line.

Close modal

If the electronic states of DNA played no role in charge conduction and DNA only acts as a trap-free insulator, charge tunneling between QDs would be the likely transport mechanism

IV2exp[ 4m*(qϕ)3/23qV ]
(4)

where m* is the charge carrier effective mass. However, the clear temperature dependence of the I-V characteristics (Figure 2) largely eliminates this possibility. Another possible charge transport mechanism where the DNA plays a limited role could be ionic conduction

IVTexp[ ΔEakT ]
(5)

where ΔEa is the activation energy. In this mechanism, the DNA could be inert and the metal ions (Na+, Mg2+, etc.) which counter the negatively charged phosphate backbone, can act as mobile charge carriers through the film. A pronounced hysteresis is usually observed under these conditions, which was absent from our measurements. Furthermore, this functional dependence of electronic current on bias and temperature does not match our observed charge transport measurements (Figure 3(b)). If the dsDNA was the charge conduction medium, but the HOMO-LUMO electronic states (measured by our STS) do not pervade the entire molecule such that each nucleotide is energetically separate, a mechanism resembling Frenkel-Poole emission

IVexp[ qkT(ϕV1/2qπε) ]
(6)

may be expected. If this is the case, the HOMO-LUMO levels which are capable of accepting charge would be analogous to trap states in an insulator and require the electrons or holes to hop across the sequence instead of being smoothly conducted. This is inconsistent with all our STS measurements, and charge transport studies through electrically injected electrons and holes in dark (Figure 3(b)). Another alternative is if the QD-DNA thin film acted as a composite semiconductor, charge separation following Schottky emission-like behavior would be observed:

IT2exp[ qkT(ϕV1/2q4πε) ]
(7)

However, these charge transport studies, when compared against different potential charge transport mechanisms (Figure 3(b), Eqs. (3)–(7)), consistently indicate SCLC as the transport mechanism which matches the observed I-V-T characteristics and our STS measurements. In these QD-DNA thin films, QDs likely act as charge centers with a conduction gradient extending radially outward. Because of the well-aligned and coupled energy levels between QDs and DNA, the charge carriers seamlessly extend into the DNA molecules, which mediate the transport through these thin films. The conductivity gradient likely arises due to the relatively higher charge carrier mobility and conductivity in the semiconductor QDs compared to the DNA.

While measurements of electrically injected charges (in dark) corroborate the band alignment observed by STS measurements and provide insights into the role of DNA, we investigated the transport of photogenerated charge carriers through different energy exciton shelves in QD-DNA thin films. We illuminated the transparent side (ITO end) of our exploratory thin film device with monochromatic light (Figure 1(a)) and quantified the charge transport (or photoresponse) of different energy photogenerated charge carriers through the QD-DNA thin film. While hot electrons would ideally yield an increased voltage as their extra energy is maintained through conduction, the output voltage is governed by the work function of the metal electrodes. Thus, while we can probe the effect of the exciton shelves as conduction pathways in these exploratory photodetection devices, a more optimized device design and advanced electrodes (with two separate work functions) would be required to fully use of both bandedge and hot carriers.

The I-V curves under illumination yielded a clear photovoltaic response above the bandgap of the CdTe QDs, and a significant increase in current with light illumination (due to photogenerated charges, Figure 4(a)). Comparing the photocurrent normalized by light intensity (P) across the UV-VIS spectrum, we observed a several orders of magnitude increase in photocurrent for ultraviolet photons (Figure 4(b)). Nominally in thin film QD devices, the photocurrent is limited by light absorption, and hence mimics the QD absorbance (inset, Figure 4(b)) as there is only one pathway for charge conduction.1–6 While the absorbance of our QDs does increase with the photon energy (Figure 4(b), inset), it does not match the orders of magnitude increase we observed.

FIG. 4.

Photoresponse measurements of the CdTe-dsDNA thin film. (a) I-V characteristics in dark and light exhibiting increased photocurrent upon illumination and a photovoltaic response. (b) Photocurrent normalized to light intensity as a function of monochromatic photon energy. Inset shows the absorbance spectrum of the CdTe quantum dots. (c) Analysis of the photocurrent using the Fowler-Nordheim equation. This figure-of-merit at a fixed bias represents the barrier for charge conduction at each photon energy, with a lower value corresponding to a higher barrier for charge (electron here) transport.

FIG. 4.

Photoresponse measurements of the CdTe-dsDNA thin film. (a) I-V characteristics in dark and light exhibiting increased photocurrent upon illumination and a photovoltaic response. (b) Photocurrent normalized to light intensity as a function of monochromatic photon energy. Inset shows the absorbance spectrum of the CdTe quantum dots. (c) Analysis of the photocurrent using the Fowler-Nordheim equation. This figure-of-merit at a fixed bias represents the barrier for charge conduction at each photon energy, with a lower value corresponding to a higher barrier for charge (electron here) transport.

Close modal

To gain a better understanding of the charge conduction for photogenerated charges, we developed a figure-of-merit for analyzing energetically diverse charge carriers (Figure 4(c))

F(λ)=Vln[ I(λ)IdarkV2×P(λ)×A(λ) ]
(8)

Using this figure-of-merit, the photoresponse (I(λ)Idark) is normalized by QD absorbance (A) and is analyzed using the Fowler-Nordheim plot

IV2exp[ C(ϕ)V ]
(9)

to separate the effect of different charge conduction regimes.17 This figure-of-merit evaluated for a given applied bias, plotted for different incident photon energies for photogenerated charges, represents a measure of the relative barrier for charge transport as a function of photon energy (Figure 4(c)). In traditional QD thin films, this functional form would yield a flat curve due to the presence of a single pathway for charge conduction through the QD bandedge states. Based on the shape of our curve, there are two distinct regions of charge conduction in our system for the same applied bias. The first region, at low photon energies, corresponds to a relatively high barrier for charge transport, where the lower energy photogenerated charge carriers in CdTe are being injected into the lower energy adenine LUMO level (Figure 1(c)). After a transition (∼2.8 eV), there is a second region corresponding to lower barrier for injection of photogenerated charges into the complementary thymine LUMO level. The inflection point of the transition region matches the separation of the QD CB and the thymine LUMO level measured by STS (Figure 1(b)). Taken together, these charge conduction pathways match our STS measurements and are an evidence of different energy conduction pathways using exciton shelves in QD-DNA thin films.

In conclusion, we have demonstrated a QD-DNA nanobio-hybrid architecture using the alignment of the energetic states of these materials. Using temperature- and external bias-dependence of transport of electrically injected charges in dark, we showed that the charge conduction likely proceeds through a space-charge limited mechanism. We have also shown that it is possible to transport different energy photogenerated charge carriers in QDs through the exciton shelves in these QD-DNA thin films. This provides a direct evidence for the existence of exciton shelves and may be a useful consideration when developing new device architectures. While providing a proof of concept for such architectures, the potential charge trapping due to the negatively charged phosphate-sugar backbone, improvements required in device architecture, low-conductivity of the DNA, low temperature tolerance, and a high-cost of DNA also limits the potential application of these devices. However, the insights gained from measurements of DNA electronic levels and its integration with optoelectronic nanomaterials opens up opportunities to integrate them as biological transducers and provide a pathway to intervene with electronic or optical stimuli, while building a foundation for investigating other conjugated polymers as molecular wires for the formation of exciton shelves with QDs.

This work was funded by National Science Foundation CAREER Award CBET 1351281 (S.M.G. and P.N.) and W. M. Keck Foundation Research Award (V.S. and P.N.), U.S. Department of Energy (DOE), Office of Science, Basic Energy Sciences (BES) under Award No. DE-SC0006398 (synthesis and characterization, H.N. and J.N.C. were supported by DOE DE-SC0006398), and NSF Soft Materials MRSEC at the University of Colorado through NSF Award DMR 1420736 (J.N.C. and P.N.).

1.
S.
Coe
,
W. K.
Woo
,
M.
Bawendi
, and
V.
Bulović
,
Nature
420
,
800
(
2002
).
2.
I.
Gur
,
N. A.
Fromer
,
M. L.
Geier
, and
A. P.
Alivisatos
,
Science
310
,
462
(
2005
).
3.
D. V.
Talapin
and
C. B.
Murray
,
Science
310
,
86
(
2005
).
4.
O. E.
Semonin
,
J. M.
Luther
,
S.
Choi
,
H. Y.
Chen
,
J.
Gao
,
A. J.
Nozik
, and
M. C.
Beard
,
Science
334
,
1530
(
2011
).
5.
V.
Sukhovatkin
,
S.
Hinds
,
L.
Brzozowski
, and
E. H.
Sargent
,
Science
324
,
1542
(
2009
).
6.
S. M.
Goodman
,
V.
Singh
,
J. C.
Ribot
,
A.
Chatterjee
, and
P.
Nagpal
,
J. Phys. Chem. Lett.
5
,
3909
(
2014
).
7.
V.
Singh
,
I. J. C.
Beltran
,
J. C.
Ribot
, and
P.
Nagpal
,
Nano Lett.
14
,
597
(
2014
).
8.
P.
Nagpal
and
V.
Klimov
,
Nat. Commun.
2
,
486
(
2011
).
9.
J. J.
Urban
,
D. V.
Talapin
,
E. V.
Shevchenko
, and
C. B.
Murray
,
J. Am. Chem. Soc.
128
,
3248
(
2006
).
10.
C. A.
Mirkin
,
R. L.
Letsinger
,
R. C.
Mucic
, and
J. J.
Storhoff
,
Nature
382
,
607
(
1996
).
11.
D.
Han
,
S.
Pal
,
J.
Nangreave
,
Z.
Deng
,
Y.
Liu
, and
H.
Yan
,
Science
332
,
342
(
2011
).
12.
H.
Noh
,
A. M.
Hung
, and
J. N.
Cha
,
Small
7
,
3021
(
2011
).
13.
S. Y.
Park
,
A. K.
Lytton-Jean
,
B.
Lee
,
S.
Weigand
,
G. C.
Schatz
, and
C. A.
Mirkin
,
Nature
451
,
553
(
2008
).
14.
See supplementary material at http://dx.doi.org/10.1063/1.4913563 for chemical synthesis, device fabrication, STS, and electronic characterization data.
15.
H.
Noh
,
S. M.
Goodman
,
P.
Mohan
,
A. P.
Goodwin
,
P.
Nagpal
, and
J. N.
Cha
,
RSC Adv.
4
,
8064
(
2014
).
16.
S. M.
Sze
,
Physics of Semiconductor Devices
, 2nd ed. (
Wiley-Interscience
,
New York, NY, USA
,
1981
).
17.
R. H.
Fowler
and
L.
Nordheim
,
Proc. R. Soc. London
119
,
173
(
1928
).
18.
S. H.
Choi
,
B. S.
Kim
, and
C. D.
Frisbie
,
Science
320
,
1482
(
2008
).
20.
J. M.
Beebe
,
B.
Kim
,
J. W.
Gadzuk
,
C. D.
Frisbie
, and
J. G.
Kushmerick
,
Phys. Rev. Lett.
97
,
026801
(
2006
).
21.
M. A.
Lampert
,
Phys. Rev.
103
,
1648
(
1956
).

Supplementary Material