Layered insulator hexagonal boron nitride for surface

Layered insulator hexagonal boron nitride for surface passivation in quantum dot solar cell Mariyappan Shanmugam, Nikhil Jain, Robin Jacobs-Gedrim, Yang Xu, and Bin Yu Citation: Applied Physics Letters 103, 243904 (2013); doi: 10.1063/1.4848235 View online: http://dx.doi.org/10.1063/1.4848235 View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/103/24?ver=pdfcov Published by the AIP Publishing Articles you may be interested in High open circuit voltages of solar cells based on quantum dot and dye hybrid-sensitization Appl. Phys. Lett. 104, 013901 (2014); 10.1063/1.4861163 CdS quantum dots grown by in situ chemical bath deposition for quantum dot-sensitized solar cells J. Appl. Phys. 110, 044313 (2011); 10.1063/1.3624944 Quantum-dot-sensitized solar cells: Assembly of CdS-quantum-dots coupling techniques of self-assembled monolayer and chemical bath deposition Appl. Phys. Lett. 90, 143517 (2007); 10.1063/1.2721373 Ultrahigh vacuum preparation and characterization of TiO 2 / CdTe interfaces: Electrical properties and implications for solar cells J. Appl. Phys. 91, 1984 (2002); 10.1063/1.1435413 Novel method for growing CdS on CdTe surfaces for passivation of surface states and heterojunction formation J. Vac. Sci. Technol. A 15, 1119 (1997); 10.1116/1.580440

Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. Download to IP: 210.32.190.6 On: Sun, 30 Oct 2016 09:42:37

APPLIED PHYSICS LETTERS 103, 243904 (2013)

Layered insulator hexagonal boron nitride for surface passivation in quantum dot solar cell Mariyappan Shanmugam,1 Nikhil Jain,1 Robin Jacobs-Gedrim,1 Yang Xu,2 and Bin Yu1,a) 1

College of Nanoscale Science and Engineering, State University of New York, Albany, New York 12203, USA Institute of Microelectronics and Optoelectronics, Zhejiang University, Hangzhou 310027, People’s Republic of China 2

(Received 22 October 2013; accepted 29 November 2013; published online 11 December 2013) Single crystalline, two dimensional (2D) layered insulator hexagonal boron nitride (h-BN), is demonstrated as an emerging material candidate for surface passivation on mesoporous TiO2. Cadmium selenide (CdSe) quantum dot based bulk heterojunction (BHJ) solar cell employed h-BN passivated TiO2 as an electron acceptor exhibits photoconversion efficiency 46% more than BHJ employed unpassivated TiO2. Dominant interfacial recombination pathways such as electron capture by TiO2 surface states and recombination with hole at valence band of CdSe are efficiently controlled by h-BN enabled surface passivation, leading to improved photovoltaic performance. Highly crystalline, confirmed by transmission electron microscopy, dangling bond-free 2D layered h-BN with self-terminated atomic planes, achieved by chemical exfoliation, enables efficient passivation on TiO2, allowing electronic transport at TiO2/h-BN/CdSe interface with much lower recombination C 2013 AIP Publishing LLC. rate compared to an unpassivated TiO2/CdSe interface. V [http://dx.doi.org/10.1063/1.4848235] Bulk heterojunction (BHJ) solar cells, employing titanium dioxide (TiO2) as an electron acceptor and poly(3-hexylthiophene) (P3HT) as a hole conductor, have widely been studied as an emerging, cost effective hybrid photovoltaic technology.1–4 Cadmium selenide (CdSe) quantum dots have become the most important semiconductor material choice for BHJ solar cells due to efficient photo-absorption in the visible spectrum.5–8 While mesoporous TiO2 is widely used as an electron acceptor to adsorb CdSe quantum dots, various kinds of functional nanostructures as electron acceptors have been explored for BHJ solar cell applications.9,10 Photovoltaic performance of the BHJ solar cells depends on many factors including photon harvesting capability of the quantum-dot sensitizer employed, electron injection into the electron acceptor, carrier transport, and interfacial recombination mechanism. Mesoporous TiO2 is a n-type wide-bandgap (EG 3.1 eV) semiconductor material consists of nanoparticles with size on the order of 15 nm–20 nm, interconnected by van der Waals forces through which photo-injected electrons are transported by diffusion process. The porosity and nanoparticle size are highly controllable by the synthesis process to make the material suitable and efficient for photovoltaic applications.11,12 High porosity in the bulk of TiO2 is preferred to coat the semiconductor quantum dots on the surface. Porosity of the TiO2 facilitates semiconductor quantum dots to diffuse into the bulk. While the porosity and nanoparticle size in TiO2 are engineered to favor the sensitizer loading onto the bulk, the associated surface area is significantly increased. However, the density of defects associated with the surface of TiO2 nanoparticles is also increased with surface area. The defects, present throughout the bulk of TiO2 surface, are surface states which play a vital role in electronic transport and a)

Author to whom correspondence should be addressed. Electronic mail: [email protected].

0003-6951/2013/103(24)/243904/5/$30.00

recombination in hybrid solar cells.13–15 The TiO2 surface states, close to the conduction band of TiO2, can capture photo-injected electrons and trap them in the bulk. Similarly, TiO2 surface states can capture free holes from the valance band of the CdSe quantum dots. The captured electrons and holes are recombined in the bulk of TiO2 through surface states, a major performance-limiting interfacial recombination process at TiO2/CdSe interface in BHJ solar cells. Surface passivation is a special measure to modify the defective semiconductor surface and hence defect-free interfaces can be formed. This helps to facilitate electronic transport with the lowest possible probability of recombination in solar cells. Various high-k dielectrics such as Al2O3, ZnO, ZrO2, and MgO were demonstrated as the candidates for surface passivation on TiO2 to enhance solar cell performance.16,17 The surface passivation materials are expected to be in high quality and ultra-thin to saturate the dangling bonds present on the semiconductor surface. Increasing thickness and defect density of the surface passivation layer may impede electronic transport at the respective interface in solar cells. In general, high-k dielectrics prepared by atomic layer deposition provide highly conformal coverage on semiconductor surface.18 Two dimensional (2D) hexagonal boron nitride (h-BN) is a layered insulator exhibiting wide bandgap (EG 5.5 eV) and a dielectric constant of 4. While weak van der Waals forces coordinate individual atomic planes together to form bulk h-BN crystal, strong in-plane covalent bonds make hexagonal lattice structure analogous to the well-known graphene. Recently, it has been demonstrated that micro-mechanically exfoliated or chemical-vapor deposited h-BN serves as a superb substrate material for graphene nanoelectronics.19–22 In this Letter, we present single-crystalline 2D layered insulator h-BN as a dielectric material to passivate the surface of defective mesoporous semiconducting TiO2. The potential application of h-BN as a viable surface passivation material is demonstrated in BHJ solar cell employing CdSe quantum dots as a

103, 243904-1

C 2013 AIP Publishing LLC V


243904-2

Shanmugam et al.

FIG. 1. (a) Raman spectra of h-BN nanoflake-coated TiO2 electron acceptor showing active modes of TiO2 at 145.7 cm 1, 397.1 cm 1, 513.5 cm 1, and 636.1 cm 1, respectively, and h-BN at 1367.2 cm 1. (b) A closer view of the major TiO2 peak showing no shift in wavenumber. (c) A closer view of h-BN peaks shows small shift, confirming the presence of the mixed monolayer/few layer on TiO2 surface.

photo-sensitizer to harvest solar energy in the visible spectral window. The effect of h-BN surface passivation on TiO2, interfacial recombination processes in BHJ solar cell, and photovoltaic performance are explored to elucidate 2D layered h-BN as an emerging material for surface passivation in hybrid photovoltaic cells. Two kinds of BHJ solar cells were fabricated and characterized: (1) with a stack of ITO/TiO2/CdSe/P3HT/Au and (2) ITO/TiO2/h-BN/CdSe/P3HT/Au. For the fabrication of BHJ with no h-BN surface passivation, colloidal TiO2 nanoparticle film was deposited onto an indium tin oxide (ITO)-coated glass substrate annealed at 450 C for 30 min, resulted in 15 lm porous TiO2 layer. Commercially available CdSe quantum dots were used to cover the porous TiO2 film, followed by an annealing process at 100 C for 10 min to evaporate excessive solvent present in the bulk of the porous TiO2/CdSe film. P3HT solution, prepared by dissolving

Appl. Phys. Lett. 103, 243904 (2013)

1 wt. % of P3HT in chlorobenzene, was spin-coated onto the TiO2/CdSe film. The spin-coated P3HT layer on TiO2/CdSe film diffuse into the CdSe-coated porous TiO2 nanoparticle network and forms a large surface BHJ structure to facilitate effective charge separation. A 50 nm gold thin film was deposited by electron-beam evaporation as the back contact of the BHJ solar cell. For the BHJ with h-BN surface passivation, the h-BN powders (Momentive Specialty Chemicals, Inc.,), with an average particle size of 10 lm, were manually grinded to obtain particles with smaller sizes. The grinded fine powder of h-BN (5 mg) was dissolved in 15 ml of 2-propanol solution and ultra-sonicated for 20 min. The ultra-sonication process breaks the weak van der Waals bonds which hold individual atomic planes in the h-BN. This process yields nanoflakes of h-BN in 2-propanol solution. The solution was spin-coated onto the TiO2 surface and subsequently annealed at 100 C to evaporate excessive solvent present in the film. The rest of the solar cell fabrication process is exactly the same as that for the BHJ without h-BN surface passivation layer. The h-BN nanoflake-coated TiO2 electron acceptor layer was characterized by Horiba LabRam HR 800 confocal Raman microscope to confirm the h-BN coating on TiO2 surface. The chemically exfoliated nanoflakes of h-BN were transferred onto a Transmission Electron Microscope (TEM) grid for micro-structural analysis. The JEOL 2010 Hi-Resolution Transmission Electron Microscope (HRTEM) was used to characterize the h-BN nanoflakes transferred onto a grid. The current density vs. voltage (J-V) characteristics of the fabricated BHJ solar cells were measured under both dark and standard AM 1.5 illumination conditions (calibrated using a standard silicon solar cell) by Agilent B1500A semiconductor device analyzer. A Xe arc lamp was used to simulate the AM 1.5 photon flux for the J-V characteristic measurement. The external quantum efficiency (EQE) of the BHJ solar cells was measured in the wavelength range of 350 nm–750 nm using a Newport monochromator equipped with the same Xe arc lamp used to measure the J-V characteristics of the BHJ solar cells. Figure 1 shows the Raman spectra obtained on the h-BN nanoflake-coated TiO2 nanoparticle electrode. The measurements were carried out on four different positions on the same sample to confirm the complete surface coverage on TiO2 by the h-BN. The major four active vibrational modes of TiO2 nanoparticle film were identified at 145.7 cm 1 and 636.1 cm 1 are corresponding to two Eg modes, while 397.1 cm 1 and 513.5 cm 1 represent A1g and B1g modes, respectively. The major Raman peak identified at 1367.2 cm 1 confirms the h-BN nanoflakes coated on TiO2.

FIG. 2. (a) TEM image of the h-BN nanoflake transferred onto a grid, showing atomically smooth surface. (b) HRTEM image of the atomic planes of h-BN, showing the highly crystalline lattice structure. The inset showing the magnified image of the lattice structure with no apparent defects present. (c) SAED pattern obtained from the h-BN nanoflake confirming the hexagonal crystal structure.


243904-3

Shanmugam et al.

The four measurements carried out on different positions of the sample show exactly the same Raman signature peaks for TiO2, suggesting that size of the nanoparticles is homogeneous in the film. However, we observed that there is a small variation in the Raman peak of h-BN, while the measurement position was changed. The h-BN nanoflakes in 2-propanol solution consist of a combination of monolayer/ few layers distributed randomly. The spin-coating process deposits this solution everywhere on the surface of TiO2, resulting in the presence of monolayer/few layer h-BN nanoflakes on TiO2 electron acceptor which caused the small change in the Raman peak position. Gorbachev et al.23 reported that while major Raman peak for h-BN monolayer is expected to be at 1366 cm 1, the hardening of E2g phonon mode and the strain on h-BN can cause blue and red shifts by up to 4 cm 1 and 2 cm 1, respectively. While the blue shift caused by slightly shorter B-N bonds in case of monolayer h-BN, the red shift may be due to the strain. We expect that the strain on h-BN was induced during the exfoliation of individual atomic planes by ultra-sonication process. We observe both blue and red shifts in the Raman spectroscopic characteristics of h-BN on TiO2 confirming the presence of monolayer/few layers. Figure 2(a) shows the TEM image of the h-BN nanoflake along with the high-resolution lattice image (Figure 2(b)) and the corresponding selective area energy diffraction (SAED) pattern (Figure 2(c)). The surface of h-BN nanoflake is observed to be ultra-smooth and wrinkles-free as shown in Figure 2(a). Due to the weak van der Waals interaction, the self-terminated individual atomic layers of h-BN results in ultra-smooth surface after the exfoliation process. It is an essential feature for a surface passivation material to possess surface with self-saturated atomic bonds. The lattice pattern of h-BN is illustrated in Figure 2(b) confirms the singlecrystalline phase along with a high-resolution lattice image, as shown in the inset. The measured lattice constant of the h-BN nanoflake along the direction of [1100] is 0.25 nm. We observe a highly periodic lattice structure without any significant defects on the sites. However, few regions can be seen as mixed single-crystalline phase along with small disorders, possibly due to the manual transfer process on the TEM grid. Further, the single-crystalline phase of h-BN nanoflake is confirmed by the SAED pattern shown in Figure 1(c). The incident electron beam is perpendicular to the [0001] plane of the h-BN nanoflake. We observe the atomic spacing of 0.135 nm along the direction of [1010] and 0.235 nm along that of [1120]. The angle between these two directions confirms the hexagonal crystal structure similar to graphene. Figures 3(a) and 3(b) show the J-V characteristics of BHJ solar cells in a forward applied bias region of 0 V–1 V. The BHJ solar cell employed non-passivated TiO2 nanoparticle electrode as an electron acceptor exhibits short-circuit current density (JSC), open-circuit voltage (VOC), maximum power output (PMAX), and photoelectric conversion efficiency (g) of 10.9 mA/cm2, 663 mV, 3.4 10 4 W, and 4.8%, respectively, as shown in Figure 2(a). On the other hand, the BHJ solar cell employed h-BN nanoflake-enabled surface passivation on TiO2 yields JSC, VOC, PMAX, and g of 15.3 mA/cm2, 719 mV, 4.9 10 4 W, and 7.0%, respectively (Figure 2(b)). To understand the significant

Appl. Phys. Lett. 103, 243904 (2013)

FIG. 3. Dark and illuminated J-V characteristics of the BHJ solar cells (a) with no surface passivation on TiO2 and (b) with h-BN surface passivation on TiO2. The output power characteristics are shown in the respective insets. (c) Transport and recombination processes involve at the TiO2/CdSe interface and TiO2/h-BN/CdSe interfaces. Process 1 shows optical excitation generating electron-hole pair at the valence and conduction band of the CdSe quantum dots, injecting electron to the TiO2 nanoparticles. Process 2 shows electron capture by TiO2 surface states from the conduction band. Process 3 represents the hole capture from the valence band of the CdSe leading to recombination. h-BN helps to block Process 3, leading to electron emission from the TiO2 surface state (Process 4).

enhancement in photovoltaic performance due to h-BN surface passivation, we explored the possible transport and recombination mechanisms at TiO2/CdSe/P3HT and TiO2/h-BN/CdSe/P3HT interfaces. Figure 2(c) shows the


243904-4

Shanmugam et al.

key physical processes occurring during the solar cell operation under the illumination condition. Photons generate excitons in CdSe quantum dots (Process 1), injecting electrons into the conduction band of the TiO2 nanoparticles. The photo-injected electrons are transported towards the ITO electrode to be collected. The electron capture by TiO2 surface states (Process 2) impedes the electron transport in the TiO2 due to the presence of surface states. While Process 1 leaves holes in the valence band of CdSe quantum dots, these holes can be captured by TiO2 surface states as illustrated in Process 3. The electron and hole capture processes lead to interfacial recombination mediated by TiO2 surface states. We use nanoflakes of crystalline, layered h-BN as the material to passivate the TiO2 surface states. The potential barrier provided by the h-BN on the surface of TiO2 does not allow the trapped electrons at the TiO2 surface states to interact with holes present at the valence band of CdSe quantum dots. Thus, a major recombination pathway is blocked by h-BN surface passivation, resulting in improved photovoltaic performance. Since the recombination probability is significantly controlled by the h-BN passivation, the trapped electrons at the TiO2 surface states migrate in the bulk of TiO2 by diffusion process called electron emission (Process 4). This process helps to collect the photo-injected electrons, resulting in effective electronic transport. We observe 40% enhancement in JSC of the BHJ employing h-BN surface passivation due to efficient charge collection at the ITO electrode. The h-BN surface passivation layer reduced the probability of electron capture by TiO2 surface states leading to longer carrier lifetime at the TiO2 conduction band facilitating efficient collection. The enhancement in VOC (56 mV) suggests that the TiO2/CdSe interface quality is significantly improved by h-BN passivation. We believe that TiO2/CdSe interface is critical in the presented solar cell structure, as the surface states are directly exposed to CdSe quantum dots. The dominant recombination process (between trapped electrons at TiO2 surface states and holes from CdSe) occurs at the TiO2/CdSe interface where h-BN efficiently passivates the surface states and helps in reducing the probability of recombination. Figure 4 shows the EQE spectra of the two BHJ solar cells in the spectral window of 350 nm–750 nm. The major photoactive material used in both solar cells is CdSe quantum dots which exhibit effective optical absorption in the same spectral window. We observe that the EQE of the BHJ solar cell employed h-BN surface passivation improved significantly in the visible spectral window. The maximum EQE of the BHJ with h-BN surface passivation is 79% at 510 nm, while that with no surface passivation exhibited 64%. The enhancement in EQE is attributed to the effective electron collection in the BHJ solar cell with h-BN surface passivation, different from the device with no surface passivation. The photo-injected electrons in the TiO2 nanoparticle electrode gains longer lifetime, as the electron and hole capturing processes are suppressed by h-BN, as compared with non-passivated TiO2. The increment in electron lifetime in the bulk of TiO2 electrode facilitates effective electron collection at the transparent-conductor-oxide electrode, resulting in enhanced EQE. In other words, while

Appl. Phys. Lett. 103, 243904 (2013)

FIG. 4. EQE spectra of the two BHJ solar cells in a spectral window of 350 nm–750 nm showing the enhancement achieved in solar cell with h-BN surface passivation on TiO2, as compared with the device with no surface passivation.

surface passivation provides interfaces with reduced defects, electron re-emission helps to achieve effective electron transport in bulk TiO2, which helps to reduce the chance of losing the electrons by recombination process. In summary, we demonstrated h-BN enabled surface passivation to enhance solar cell performance significantly by suppressing the dominant interfacial recombination process. Single-crystalline, layered insulator h-BN can be used in photovoltaic applications, benefiting from its selfterminated, atomically smooth planes that effectively passivate the surface of defective semiconductor. As compared with a number of amorphous and polycrystalline dielectrics previously explored for surface passivation in solar cells, hBN is expected to be highly advantageous due to its very unique, graphene-like 2D lattice structure. 1

Y. Y. Yu, W. C. Chien, Y. H. Ko, Y. C. Chan, and S. C. Liao, Curr. Appl. Phys. 12, S7–S13 (2012). 2 S.-Jin Moon, E. Baranoff, S. M. Zakeeruddin, C.-Yu Yeh, E. W.-Guang Diau, M. Gr€atzel, and K. Sivula, Chem. Commun. 47, 8244–8246 (2011). 3 L. Shen, G. Zhu, W. Guo, C. Tao, X. Zhang, C. Liu, W. Chen, S. Ruan, and Z. Zhong, Appl. Phys. Lett. 92, 073307 (2008). 4 Z. J. Wang, S. C. Qu, X. B. Zeng, J. P. Liu, C. S. Zhang, F. R. Tan, L. Jin, and Z. G. Wang, Appl. Surf. Sci. 255, 1916–1920 (2008). 5 Y. L. Lee, B. M. Huang, and H. T. Chien, Chem. Mater. 20, 6903–6905 (2008). 6 S. Q. Fana, D. Kima, J. J. Kima, D. W. Jungb, S. O. Kanga, and J. Ko, Electrochem. Commun. 11, 1337–1339 (2009). 7 K. Prabakar, S. Minkyu, S. Inyoung, and K. Heeje, J. Phys. D: Appl. Phys. 43, 012002 (2010). 8 S. Sun, L. Gao, Y. Liu, and J. Sun, Appl. Phys. Lett. 98, 093112 (2011). 9 Y. Hao, Y. Cao, B. Sun, Y. Li, Y. Zhang, and D. Xu, Sol. Energy Mater. Sol. Cells 101, 107–113 (2012). 10 Y. Li, C. W. Wang, Y. Zhao, J. Wang, and F. Zhou, J. Solid State Chem. 196, 349–355 (2012). 11 J. B. Joo, M. Dahl, N. Li, F. Zaera, and Y. Yin, Energy Environ. Sci. 6, 2082–2092 (2013). 12 C. H. Huang, Y. T. Yang, and R. A. Doong, Microporous Mesoporous Mater. 142, 473–480 (2011). 13 S. H. Kang, J. Y. Kim, and Y. E. Sung, Electrochim. Acta 52, 5242–5250 (2007). 14 F. Nunzi, E. Mosconi, L. Storchi, E. Ronca, A. Selloni, M. Gr€atzel, and F. D. Angelis, Energy Environ. Sci. 6, 1221–1229 (2013). 15 J. Bisquert, J. Electroanal. Chem. 646, 43–51 (2010). 16 T. C. Li, M. S. G oes, F. Fabregat-Santiago, J. Bisquert, P. R. Bueno, C. Prasittichai, J. T. Hupp, and T. J. Marks, J. Phys. Chem. C 113, 18385–18390 (2009).


243904-5 17

Shanmugam et al.

A. Kay and M. Gr€atzel, Chem. Mater. 14, 2930–2935 (2002). G. Dingemans and W. M. M. Kessels, J. Vac. Sci. Technol. A 30, 40802 (2012). 19 N. Jain, T. Bansal, C. A. Durcan, Y. Xu, and B. Yu, Carbon 54, 396–402 (2013). 20 G. H. Lee, Y. J. Yu, X. Cui, N. Petrone, C. H. Lee, M. S. Choi, D. Y. Lee, C. Lee, W. J. Yoo, K. Watanabe, T. Taniguchi, C. Nuckolls, P. Kim, and J. Hone, ACS Nano 7, 7931–7936 (2013). 18

Appl. Phys. Lett. 103, 243904 (2013) 21

Z. Liu, L. Ma, G. Shi, W. Zhou, Y. Gong, S. Lei, X. Yang, J. Zhang, J. Yu, K. P. Hackenberg, A. Babakhani, J. C. Idrobo, R. Vajtai, J. Lou, and P. M. Ajayan, Nat. Nanotechnol. 8, 119–124 (2013). 22 N. Jain, C. A. Durcan, R. Jacobs-Gedrim, Y. Xu, and B. Yu, Nanotechnology 24, 355202 (2013). 23 R. V. Gorbachev, I. Riaz, R. R. Nair, R. Jalil, L. Britnell, B. D. Belle, E. W. Hill, K. S. Novoselov, K. Watanabe, T. Taniguchi, A. K. Geim, and P. Blake, Small 7, 465–468 (2011).
