NANO LETTERS
Study of Electronic Defects in CdSe Quantum Dots and Their Involvement in Quantum Dot Solar Cells
2009 Vol. 9, No. 2 856-859
Ruben Loef,*,† Arjan J. Houtepen,† Elise Talgorn,† Joop Schoonman,‡ and Albert Goossens† Opto-Electronic Materials and Materials for Energy ConVersion and Storage, Delft UniVersity of Technology, Julianalaan 136, 2628 BL Delft, The Netherlands Received December 11, 2008; Revised Manuscript Received January 15, 2009
ABSTRACT To enhance efficiencies of quantum dot CdSe/TiO2 based solar cells, understanding of the space charge at the CdSe/TiO2 interface is crucial. In this paper, the presence of a shallow acceptor in the CdSe quantum dots is found by means of a detailed impedance and Mott-Schottky (C -2-O) study. Furthermore, it is clearly shown that this acceptor density decreases strongly with increasing quantum dot size. The presence of these defect states may give rise to Auger recombination in small quantum dots and therewith decrease the efficiency of quantum-dotsensitized solar cells.
One of the main reasons for the growing interest in quantum dots is their use in cheap solar cells, which have the possibility to increase the thermodynamic conversion efficiency above the Shockley-Queisser limit.1 Of significant practical interest are solar cells based on sensitization of nanostructured TiO2, ZnO, and SnO2 with quantum dots,2-9 although other configurations are suggested as well.1,10-13 Efficiencies up to 1.7% are reported for TiO2 nanoparticles sensitized with CdSe quantum dots.14 This efficiency should clearly be enhanced, but such improvement is hampered by the lack of understanding of the space charge at the quantum dot/metal oxide interface. To date, little is known about the space charge formation at quantum dot CdSe/metal-oxide junctions. Impedance spectroscopy (IS) has been applied on quantum dot CdSe/ Au heterojunctions to determine the nature of the electrical bistability of the system.15 Furthermore, quantum dot CdS/ Au heterojunctions have been studied by capacitance-voltage (C-φ) characterization. For n-type CdS quantum dots, donor densities of (0.35-7.7) × 1014 cm-3 for 3-5 nm particles have been reported.16 Both studies suggest the presence of surface/interface states. In the present paper we study quantum dot CdSe/TiO2 heterojunctions, by means of current-voltage, IS, and C-φ measurements at different temperatures. We unambiguously find that CdSe quantum dots show p-type behavior, due to the presence of shallow electron acceptor states. Furthermore, we are able to conclude * Corresponding author,
[email protected]. † Opto-Electronic Materials. ‡ Materials for Energy Conversion and Storage. 10.1021/nl803738q CCC: $40.75 Published on Web 01/26/2009
2009 American Chemical Society
Table 1. Overview of the Samplesa absorption peak (nm)
absorption peak (eV)
DQD (nm)
NQD,max (cm-3)
497 2.49 2.4 8.9 × 1018 508 2.44 2.6 8.1 × 1018 514 2.41 2.7 7.7 × 1018 516 2.40 2.7 7.6 × 1018 534 2.32 3.1 6.4 × 1018 554 2.23 3.5 5.1 × 1018 a The quantum dot diameter (without TOPO/HDA capping), DQD, is determined from the maximum of the first peak of the absorption spectra. The maximum quantum dot density, NQD,max, is calculated for a closepacked quantum dot layer according to eq 2.
that the acceptor density decreases with increasing quantum dot diameter, which plays a role in the decreasing efficiences of quantum-dot-sensitized solar cells with decreasing particle size.2 CdSe quantum dots are prepared following a “Greener” synthesis as described by Mekis et al.17 Samples at different time intervals are taken during crystal growth to obtain quantum dots with different sizes. Quantum dot diameters, DQD, are determined from the maximum of the first peak of the absorption spectra. These peak values are compared with absorption peak maxima reported earlier for particles with known diameter.18-24 The results are summarized in Table 1. The quantum dot density, NQD, in the quantum dot film is calculated according to NQD )
number of quantum dots film volume
(1)
In close-packed quantum dot films, either face-centered cubic (fcc) and hexagonal close-packed (hcp), it is known
that the packing factor is π/181/2. Thus, the maximum quantum dot density NQD,max is given by NQD,max )
π
√18 · VQD
)
√2 1.41 ≈ 3 DQD DQD3
(2)
Here, VQD is the film volume. It has to be noted that in our samples a TOPO/HDA capping is present around the CdSe particles. Therefore, the length of the TOPO and HDA molecules must be included into eq 2 as well. The length of the TOPO and HDA molecules is estimated from bond length and angles to be around 1.2 and 2 nm, respectively. In our calculations we assume some intertwining of the capping molecules of different quantum dots and, therefore, add 3 nm (≈1.5 × length of the longer HDA molecule) to DQD when calculating NQD,max. Quantum dot diameters given in this paper refer to diameters without TOPO/HDA capping, unless stated otherwise. In drop-casted films it is likely that the quantum dots are not exactly close packed, i.e., the packing factor 0.5 V and a flatter one, φ < 0 V. The intermediate region shows a valley superimposed on the two slopes. This intermediate region is probably related to interface traps at the quantum dot CdSe/ TiO2 interface. The nature of these traps is still under investigation and will not be further discussed in the present paper. The small hysteresis that is observed in the C -2-φ curves may also be related to interfacial charging. We will focus now on the two slopes. The slopes are related to the situations where the TiO2- and quantum dot layers are dominant. To distinguish between the different slopes, the quantum dots are removed from the samples after the measurements and the remaining TiO2/TCO samples are analyzed with impedance and C -2-φ measurements in a 0.1 857
Figure 3. C -2-φ curves for 3.1 nm quantum dot CdSe/TiO2 devices at 310 K (squares), 330 K (circles), and 350 K (triangles). The direction of the potential scan is identical for all measurements and indicated by the arrow. The potential regions A, B, and C are related to the band diagrams in Figure 4.
Figure 4. Band diagrams for the CdSe quantum dot/TiO2 heterojunction for the case where the CdSe quantum dot layer is fully depleted (A), both layers are not fully depleted (B), and the flatband situation (C). Figure 3 shows the corresponding potential ranges for the three cases. Vbi is the built-in potential.
M KOH electrolyte with a Ag/AgCl reference electrode and a Pt counter electrode. Following the same argumentation as above, the C -2-φ analysis of TiO2/TCO samples with a liquid junction contact is performed at 10 kHz. From the C -2-φ analysis, donor densities, ND, are derived from the slopes of the plot according to 2 dC -2 )dφ qε0εr ND A2
(3)
Here, q is the elementary charge (1.60 × 10-19 C), ε0 is the vacuum permittivity (8.85 × 10-12 F m-1), εr the relative dielectric constant of the semiconductor layer, and A the surface area of the contact. A donor density around 1020 cm-3 is found for TCO and 1018 cm-3 for TiO2. For the calculations, dielectric constants of 3.7 and 55 are used, respectively. The latter values correspond closely to those found at φ < 0 V for the samples with quantum dots. From the above it is concluded that the slope at φ < 0 V is related to the situation where the CdSe quantum dot layer is fully depleted and the capacitance change is caused by the change of the depletion layer width in TiO2. The slope at φ > 0.5 V is related to the situation where the quantum dot film is dominant. Still two cases are possible, i.e., the CdSe quantum dot layer can have either n-type or p-type semiconductor properties. In the case of an n-type layer, the sample should behave like an n-n heterojunction, with a 858
Figure 5. Arrhenius plots of acceptor densities of the quantum dot layer for 2.4 nm (squares), 2.6 nm (circles), 2.7 nm (up triangles), 2.7 nm (down triangles), 3.1 nm (diamond), and 3.5 nm (left triangle) diameter quantum dots.
Schottky barrier at the gold/quantum dot interface. In this case the forward bias situation is expected at negative potentials. Because this is not the case, we conclude that a p-n heterojunction at the quantum dot CdSe/TiO2 interface is formed, as illustrated in Figure 4. This is quite surprising, because it has been suggested that electron transport is dominant in TOPO-capped CdSe quantum dots.30 However, a different preparation method is used for the CdSe quantum dots under investigation here. For φ > 0.5 V both the CdSe quantum dot and TiO2 layer are not fully depleted, and both layers contribute to the slope in the C -2-φ plot, according to 2(εQDNA,QD + εTiO2ND,TiO2 dC -2 )dφ qε0εQDNA,QDεTiO2ND,TiO2
(4)
Following this equation, the acceptor densities of the CdSe quantum dot layers have been determined using the slopes for descending voltages. All values are shown in the Arrhenius plot in Figure 5. No clear temperature dependence is observed for all quantum dot sizes, which indicates that the electron acceptors are completely ionized around room temperature. These shallow electron acceptors are therefore located within kT (∼25 meV) from the valence band. Figure 5 shows a clear decrease of the acceptor density with increasing particle size. To investigate if there is a relationship between the acceptor density and the quantum dot density, NA,QD is plotted as NA,QD/NQD,max in Figure 6. It can be seen that effectively only a few percent of the quantum dots has an acceptor trap. For all temperatures the charged fraction decreases with increasing quantum dot size. On first sight it seems that there is an exponential relation between NA,QD/NQD,max and the quantum dot size. However, we cannot think of a physical effect that describes such a relation. Instead, we believe that the decreasing NA,QD/NQD,max with DQD is related to some effects during the synthesis of the quantum dots. For example, larger quantum dots are grown for a longer time at high temperatures. During growth the particles are simultaneously annealed, by which part of originally present defects are removed from the quantum dots. The above results are compared with the findings of Kongkanand et al.,2 who found an optimal efficiency in quantum-dot-sensitized solar cells with CdSe quantum dots Nano Lett., Vol. 9, No. 2, 2009
References
Figure 6. NA,QD/NQD,max as function of quantum dot size at room temperature (297 K, squares), 310 K (circles), 320 K (up triangles), 330 K (down triangles), 340 K (diamond), and 350 K (left triangle). It is clearly observed that for larger quantum dot sizes, the fraction of charged quantum dots decreases strongly.
with a diameter of 3 nm. This behavior is assigned to two opposing effects related to the increasing bandgap energy with decreasing quantum dot size. The visible light absorption will be lower in smaller particles, while the higher conduction band energy will increase the driving force for electron injection in TiO2 injection. Here, we add a third aspect that determines the optimal quantum dot size in quantum-dot-sensitized solar cells. The higher concentration of electron acceptors and associated free holes in smaller particles give rise to Auger recombination, which is much faster than radiative recombination. Accordingly, recombination plays a more important role in smaller particles and their use will have a negative effect on the solar cell efficiency. In conclusion, it is shown with a detailed C -2-φ study that a Schottky barrier is present at the quantum dot CdSe/ TiO2 interface. The quantum dot layer shows p-type behavior. A relation between the quantum dot size and acceptor density is observed. Layers with smaller quantum dots, and concomitant higher quantum dot densities, show higher shallow acceptor densities. The percentage of charged quantum dots varies between ∼10% and ∼0.01% for CdSe quantum dots of 2.4 and 3.5 nm, respectively. The higher acceptor densities for smaller quantum dots will have a negative effect on the efficiency when used in quantum-dot-sensitized solar cells, due to Auger recombination. Acknowledgment. Everest Coatings (Delft, The Netherlands) is acknowledged for supplying the TiO2 samples. Supporting Information Available: Detailed experimental information. This material is available free of charge via the Internet at http://pubs.acs.org.
Nano Lett., Vol. 9, No. 2, 2009
(1) Nozik, A. J. Phys. E 2002, 14 (1-2), 115–120. (2) Kongkanand, A.; Tvrdy, K.; Takechi, K.; Kuno, M.; Kamat, P. V. J. Am. Chem. Soc. 2008, 130 (12), 4007–4015. (3) Robel, I.; Subramanian, V.; Kuno, M.; Kamat, P. V. J. Am. Chem. Soc. 2006, 128 (7), 2385–2393. (4) Leschkies, K. S.; Divakar, R.; Basu, J.; Enache-Pommer, E.; Boercker, J. E.; Carter, C. B.; Kortshagen, U. R.; Norris, D. J.; Aydil, E. S. Nano Lett. 2007, 7 (6), 1793–1798. (5) Diguna, L. J.; Shen, Q.; Kobayashi, J.; Toyoda, T. Appl. Phys. Lett. 2007, 91 (2), 023116. (6) Yu, P. R.; Zhu, K.; Norman, A. G.; Ferrere, S.; Frank, A. J.; Nozik, A. J. J. Phys. Chem. B 2006, 110 (50), 25451–25454. (7) Shen, Q.; Yanai, M.; Katayama, K.; Sawada, T.; Toyoda, T. Chem. Phys. Lett. 2007, 442 (1-3), 89–96. (8) Tachibana, Y.; Akiyama, H. Y.; Ohtsuka, Y.; Torimoto, T.; Kuwabata, S. Chem. Lett. 2007, 36 (1), 88–89. (9) Lopez-Luke, T.; Wolcott, A.; Xu, L.; Chen, S.; Wen, Z.; Li, J.; DeLaRosa, E.; Zhang, J. Z. J. Phys. Chem. C 2008, 112 (4), 1282– 1292. (10) Luque, A.; Martı´, A. Phys. ReV. Lett. 1997, 78 (26), 5014. (11) Barnham, K. W. J.; Duggan, G. J. Appl. Phys. 1990, 67 (7), 3490– 3493. (12) Nozik, A. J.; Parsons, C. A.; Dunlavy, D. J.; Keyes, B. M.; Ahrenkiel, R. K. Solid State Commun. 1990, 75 (4), 297–301. (13) Luther, J. M.; Law, M.; Beard, M. C.; Song, Q.; Reese, M. O.; Ellingson, R. J.; Nozik, A. J. Nano Lett. 2008, 8 (10), 3488–3492. (14) Lee, H. J.; Yum, J. H.; Leventis, H. C.; Zakeeruddin, S. M.; Haque, S. A.; Chen, P.; Seok, S. I.; Gratzel, M.; Nazeeruddin, M. J. Phys. Chem. C 2008, 112 (30), 11600–11608. (15) Mohanta, K.; Majee, S. K.; Batabyal, S. K.; Pal, A. J. J. Phys. Chem. B 2006, 110 (37), 18231–18235. (16) Mandal, S. K.; Maity, A. B.; Dutta, J.; Pal, R.; Chaudhuri, S.; Pal, A. K. Phys. Status Solidi A 1997, 163 (2), 433–443. (17) Mekis, I.; Talapin, D. V.; Kornowski, A.; Haase, M.; Weller, H. J. Phys. Chem. B 2003, 107 (30), 7454–7462. (18) Katari, J. E. B.; Colvin, V. L.; Alivisatos, A. P. J. Phys. Chem. 1994, 98 (15), 4109–4117. (19) Yu, W. W.; Qu, L. H.; Guo, W. Z.; Peng, X. G. Chem. Mater. 2003, 15 (14), 2854–2860. (20) Peng, X.; Wickham, J.; Alivisatos, A. P. J. Am. Chem. Soc. 1998, 120 (21), 5343–5344. (21) Murray, C. B.; Norris, D. J.; Bawendi, M. G. J. Am. Chem. Soc. 1993, 115 (19), 8706–8715. (22) Soloviev, V. N.; Eichhofer, A.; Fenske, D.; Banin, U. J. Am. Chem. Soc. 2000, 122 (11), 2673–2674. (23) Jacobsohn, M.; Banin, U. J. Phys. Chem. B 2000, 104 (1), 1–5. (24) Leatherdale, C. A.; Woo, W. K.; Mikulec, F. V.; Bawendi, M. G. J. Phys. Chem. B 2002, 106 (31), 7619–7622. (25) Efros, A. L.; Shklovskii, B. I. J. Phys. C: Solid State Phys. 1975, 8 (4), L49-L51. (26) Mott, N. F. Philos. Mag. 1969, 19 (160), 835–852. (27) Miller, A.; Abrahams, E. Phys. ReV. 1960, 120 (3), 745–755. (28) Yu, D.; Wang, C. J.; Wehrenberg, B. L.; Guyot-Sionnest, P. Phys. ReV. Lett. 2004, 92 (21), 216802. (29) Houtepen, A. J.; Kockmann, D.; Vanmaekelbergh, D. Nano Lett. 2008, 8 (10), 3516–3520. (30) Ginger, D. S.; Greenham, N. C. J. Appl. Phys. 2000, 87 (3), 1361–1368.
NL803738Q
859
