Local Charge Injection and Extraction on Surface ... - ACS Publications

Aug 17, 2016 - Nano Letters .... Fritjof Nilsson , Mattias Karlsson , Love Pallon , Marco Giacinti , Richard T. Olsson , Davide Venturi , Ulf W. Gedde...
0 downloads 0 Views 2MB Size
Letter pubs.acs.org/NanoLett

Local Charge Injection and Extraction on Surface-Modified Al2O3 Nanoparticles in LDPE Riccardo Borgani,*,† Love K. H. Pallon,*,‡ Mikael S. Hedenqvist,‡ Ulf W. Gedde,‡ and David B. Haviland† †

Nanostructure Physics, KTH Royal Institute of Technology, 10691 Stockholm, Sweden Fibre and Polymer Technology, KTH Royal Institute of Technology, 10044 Stockholm, Sweden



S Supporting Information *

ABSTRACT: We use a recently developed scanning probe technique to image with high spatial resolution the injection and extraction of charge around individual surface-modified aluminum oxide nanoparticles embedded in a lowdensity polyethylene (LDPE) matrix. We find that the experimental results are consistent with a simple band structure model where localized electronic states are available in the band gap (trap states) in the vicinity of the nanoparticles. This work offers experimental support to a previously proposed mechanism for enhanced insulating properties of nanocomposite LDPE and provides a powerful experimental tool to further investigate such properties. KEYWORDS: polyethylene nanocomposites, nanodielectrics, HVDC, surface potential, intermodulation, KPFM

H

systems have been characterized at the macro scale, detailed studies of electrical charging and discharging at the single nanoparticle level in polyethylene have not been reported. In this study, we use a recently developed mode of electrostatic force microscopy to image the charging and discharging of individual surface-modified Al2O3 nanoparticles embedded in low-density polyethylene (LDPE). Intermodulation electrostatic force microscopy14 (ImEFM) allows for nanometer scale mapping of the surface potential in the vicinity of a nanoparticle under different DC bias conditions. Comparing surface potential maps at both positive and negative bias allows us to conclude that additional electron states, behaving as shallow traps, are localized in the vicinity of the nanoparticles. The industry standard for measuring the local surface potential (VSP) of a sample relative to a conductive AFM tip is Kelvin probe force microscopy (KPFM). With KPFM, the measurement is performed by means of a voltage feedback, that is, a potential bias is adjusted on the tip to nullify VSP. Therefore, it is impossible to perform an experiment where the surface potential is investigated under different tip bias conditions. ImEFM performs the measurement in an open-loop fashion, that is, without voltage feedback. The absence of the voltage feedback allows the DC bias of the tip to be used as a free measurement parameter. By changing this DC bias, we can locally gate the sample, or change the conditions for injection of electrons or holes when the tip contacts the surface, something that is not possible with traditional KPFM. In

igh voltage direct current (HVDC) cables for long distance (>2000 km) transmission of electrical power (several GW) are key elements for the distribution of electrical energy.1 These electrical superhighways enable interconnection of transcontinental power systems and thereby the harnessing of distant renewable energy resources. To reach the desired transmission distances and capacity, the electrical insulation of cables needs to be improved to tolerate transmission voltages as high as 1 MV.2 Present day 525 kV extruded cable transmission systems employ polyethylene as an insulator.3 In recent years metal oxide (MgO, Al2O3, SiO2) nanoparticles embedded in polyethylene have emerged as a promising route to reaching the insulation properties required for ultrahigh voltage systems. These nanocomposite insulation materials have shown a 10fold reduction of conductivity, suppressed space charge accumulation and a 50% increase in breakdown strength.4−8 It has been reported that the improved properties are related to the generation of a greater number of electron traps coupled to the nanoparticles, which reduce the charge mobility.5 It has also been speculated that electron traps localized to the nanoparticles set up a counteracting electric field, resulting in higher charge injection barrier at the electrode−insulator interface.5,9 The traps in turn have been attributed to induced dipoles,10 modified electron states in the polymer interphase,11 surface states of the nanoparticles attributed to oxygen defects7 and charge difference between the nanoparticle and the polymer matrix.12,13 Whatever the mechanism involved, it is clear that the addition of nanoparticles has a strong influence on the electrical properties of the bulk material, disproportionate to the volume they occupy.4 Though a large number of nanocomposite © 2016 American Chemical Society

Received: July 14, 2016 Revised: August 12, 2016 Published: August 17, 2016 5934

DOI: 10.1021/acs.nanolett.6b02920 Nano Lett. 2016, 16, 5934−5937

Letter

Nano Letters

under the AFM tip. The average tip−sample distance is smaller than the tip radius, causing the local field to be dominantly in the direction perpendicular to the sample surface. Figure 2a−c shows the surface potential measured above an Al2O3 nanoparticle pair embedded in the PE matrix. Repetitive ImEFM scans over the same area of the sample are preformed with different applied DC bias. Figure 2 shows the three images with the same colorbar spanning the same potential range. However, to highlight change in the contrast between the nanoparticles and the LDPE matrix, the zero potential point of each color scale is offset by an amount V0, placing it at the average potential measured over the PE matrix. Figure 2e shows cross sections of the raw VSP across the nanoparticles. The scan with no applied DC bias (Figure 2a) shows the equilibrium condition between the nanoparticles and the matrix. Two nanoparticles are clearly visible at the center of the scan area, charged positive with respect to the matrix, as one would expect from the LDPE being more electronegative than the Al2O3.15,16 A third nanoparticle, buried in the matrix, is faintly visible in the lower-right corner. Figure 2d is the corresponding height image (image of the scanning feedback signal) showing the surface topography. While the presence of the three nanoparticles is clear from the potential image, it is difficult to distinguish the particles from the crystalline phase of the LDPE that protrude to create a rough topography (see also SEM images in Supporting Information, Figure S1). By comparing height and potential images, we also notice how the potential imbalance due to excess charge near the nanoparticles, spreads over a larger area than the nanoparticles themselves. We attribute this to a modification of the available energy states in the matrix due to the presence of the charged nanoparticles. When scanning with a positive DC bias applied to the AFM tip (Figure 2b), there is a marked change in image contrast, or

addition to being open-loop, ImEFM allows for the mapping of VSP in a single-pass scan, from a measurement at each image pixel of multiple frequency components of the force near the cantilever mechanical resonance, where sensitivity is greatest. The enhanced sensitivity results in higher signal-to-noise ratio (SNR) imaging, or a faster scan rate for a given measurement bandwidth. As with all AFM methods, probe height feedback is used to track the changing topography of the sample surface. A schematic representation of the experimental setup is depicted in Figure 1. A thin insulating film (100 nm) of

Figure 1. Schematic representation of the experimental setup (not to scale). A DC and an AC voltage are applied to the oscillating AFM tip. The sample’s gold substrate is connected to ground. Three nanoparticles (NP) are shown, one of which is completely embedded in the low-density polyethylene (LDPE) matrix.

nanoparticles in a LDPE matrix is spin-coated on a gold conducting substrate. ImEFM uses a voltage applied to a conducting tip. The AC component of the voltage generates an oscillating electrostatic force that is detected and analyzed to give the surface potential, VSP. Scanning over the surface we create an image of VSP (see Supporting Information). The image of VSP is strongly affected by the DC component of the applied voltage, because the latter creates a local electric field

Figure 2. ImEFM study of the Al2O3 nanoparticles embedded in LDPE. (a), (b), (c) Surface potential VSP under condition of none, positive, and negative DC bias, respectively. An offset is applied to all color scales so that they are centered on the average value of the surface potential over the LDPE matrix. (d) Height image. The black scale bar in (a)−(d) is 100 nm. (e) Cross sections of VSP and height along the lines shown in (a)−(d). In the cross sections the measured surface potential is shown without offset correction. 5935

DOI: 10.1021/acs.nanolett.6b02920 Nano Lett. 2016, 16, 5934−5937

Letter

Nano Letters

coated AFM tip under different DC bias conditions. To understand the features observed in Figure 2, we first consider the band alignment in the absence of nanoparticles. The Fermi energy EF is pinned between the substrate and the tip, and a surface potential arises given by the difference in their work functions: VSP ≔ ϕPt − ϕAu. No charge transfer between the insulating LDPE and the conducting substrate is possible and the vacuum level EV is constant between the two materials (Figure 3a). Therefore, VSP measured over the LDPE, which we call V0, depends only on the difference in ϕ between the conducting substrate and the conducting tip, and no contrast in the VSP image would be observed. The alignment of the conduction and valence band of the LDPE with respect to EF is given by the electron affinity χ and the energy gap EG. In the presence of the nanoparticle, additional localized energy states become available within the energy gap (Figure 3b) close to the conduction band (shallow traps). Therefore, it is energetically favorable for holes in the tip to move onto the surface, causing the nanoparticle to charge positively. Thus, VSP measured over a nanoparticle is higher than V0, in agreement with the contrast shown in Figure 2a. With a positive DC bias between the tip and the substrate, the Fermi energy on the tip shifts and hole injection from the tip to the surface becomes even more energetically favorable (Figure 3c). More traps become filled and the measured surface potential increases further, as seen in Figure 2b. Finally, a negative DC bias causes the filling of traps to become energetically unfavorable and holes are extracted by the tip (Figure 3d). This process can be slower than the injection at positive bias because the difference in energy between traps and EF on the tip is very small. Eventually, all the traps are empty and the vacuum level of the LDPE matrix aligns with that of the substrate, restoring the surface potential to the value of V0 so that no contrast is seen between the nanoparticle and the matrix, as observed in Figure 2c. Using the unique properties of ImEFM, we were able to observe for the first time charge injection and extraction in the vicinity of Al2O3 nanoparticles embedded in LDPE. By means of a simple band alignment model, we can relate our experimental observations to the presence of localized energy states close to the conduction band (shallow traps) in the vicinity of the nanoparticles. The presence of the traps explains the contrast with no applied DC bias, as well as its increase and its disappearance with positive and negative applied bias, respectively. The technique opens up the possibility of studying how different kinds of nanoparticle and surface modification influence the charge distribution in nanocomposites, allowing optimization of the electrical properties required to insulate 1 MV HVDC cables.

change in the difference between the surface potential near the three nanoparticles, and that of the surrounding matrix. We attribute this change to hole injection from the positively charged tip when it repeatedly taps the surface while scanning, thereby filling shallow traps near the nanoparticles. Similarly, while scanning with negative DC bias (Figure 2c), we are able to extract holes from the nanoparticles such that the contrast in the image is completely lost. It is worth mentioning that while the injection process appears to be relatively fast, requiring less than one scan (4.5 min) to reach local equilibrium, several scans at negative DC bias are needed before equilibrium is reached (see Supporting Information, Figure S2). The observed hole injection and extraction phenomena can be explained with the Al2O3 nanoparticles being host to localized electronic energy states near the LDPE conduction band, or LUMO level. Such trap states have been proposed in the literature.7,17 Figure 3 shows the alignment of the relevant electron energy bands between the Au substrate, the LDPE matrix and the Pt-



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.nanolett.6b02920. Detailed experimental description, scanning electron microscopy (SEM) images of the sample, time sequence of ImEFM scans at different DC bias, and ImEFM scans showing slow diffusion of charge. (PDF)

Figure 3. Band alignment diagram of the Au substrate, composite insulating film, and Pt-coated AFM tip under different bias conditions. (a) Over the LDPE matrix, with no trap states and no applied DC bias, VSP ≡ V0. (b) Over the nanoparticles, holes are injected from the tip on to shallow traps: VSP > V0. (c) With positive DC bias, hole injection is more favorable and more traps are occupied: VSP ≫V0. (d) With negative DC bias, hole injection is inhibited and the traps are emptied: VSP ≈V0. 5936

DOI: 10.1021/acs.nanolett.6b02920 Nano Lett. 2016, 16, 5934−5937

Letter

Nano Letters



AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. Author Contributions

(R.B. and L.K.H.P.) Contributed equally to this work. Notes

The authors declare the following competing financial interest(s): DBH is part owner of the company Intermodulation Products AB, which manufactures and sells the multifrequency lockin amplifier used in this work.



ACKNOWLEDGMENTS The authors acknowledge financial support from the Swedish Research Council (V.R.), the Knut and Alice Wallenberg Foundation, and the Swedish Foundation for Strategic Research (EM11-0022).



REFERENCES

(1) Hjertberg, T.; Englund, V.; Hagstrand, P.-O.; Loyens, W.; Nilsson, U.; Smedberg, A. Rev. Electr. Electron. 2014, 4. (2) Sellerholm, H. Elforsk AB, Stockholm 2014, Rapport 14:21. (3) Gustafsson, A.; Jeroense, M.; Ghorbani, H.; Quist, T.; Saltzer, M.; Farkas, A. Jicable 2015, A7.1.A7.1 (4) Pallon, L. K. H.; Hoang, A. T.; Pourrahimi, A. M.; Hedenqvist, M. S.; Nilsson, F.; Gubanski, S.; Gedde, U. W.; Olsson, R. T. J. Mater. Chem. A 2016, 4, 8590−8601. (5) Smith, R.; Liang, C.; Landry, M.; Nelson, J.; Schadler, L. IEEE Trans. Dielectr. Electr. Insul. 2008, 15, 187−196. (6) Murakami, Y.; Nemoto, M.; Okuzumi, S.; Masuda, S.; Nagao, M.; Hozumi, N.; Sekiguchi, Y.; Murata, Y. IEEE Trans. Dielectr. Electr. Insul. 2008, 15, 33−39. (7) Roy, M.; Nelson, J. K.; MacCrone, R. K.; Schadler, L. S. J. Mater. Sci. 2007, 42, 3789−3799. (8) Hoang, A.; Pallon, L.; Liu, D.; Serdyuk, Y.; Gubanski, S.; Gedde, U. Polymers 2016, 8, 87. (9) Milliere, L.; Makasheva, K.; Laurent, C.; Despax, B.; Teyssedre, G. Appl. Phys. Lett. 2014, 105, 122908. (10) Takada, T.; Hayase, Y.; Tanaka, Y.; Okamoto, T. Space charge trapping in electrical potential well caused by permanent and induced dipoles. Annual Report: Conference on Electrical Insulation and Dielectric Phenomena. 2007; pp 417−420.10.1109/CEIDP.2007.4451547 (11) Takada, T.; Kikuchi, H.; Miyake, H.; Tanaka, Y.; Yoshida, M.; Hayase, Y. IEEE Trans. Dielectr. Electr. Insul. 2015, 22, 1240−1249. (12) Lewis, T. IEEE Trans. Dielectr. Electr. Insul. 2004, 11, 739−753. (13) Kubyshkina, E.; Unge, M.; Jonsson, B. L. G. 2016, arXiv:1604.05984. arXiv.org e-Print archive. http://arxiv.org/abs/ 1604.05984 (accessed May 2016). (14) Borgani, R.; Forchheimer, D.; Bergqvist, J.; Thorén, P.-A.; Inganäs, O.; Haviland, D. B. Appl. Phys. Lett. 2014, 105, 143113. (15) Less, K. J.; Wilson, E. G. J. Phys. C: Solid State Phys. 1973, 6, 3110−3120. (16) Yeo, Y.-C.; King, T.-J.; Hu, C. J. Appl. Phys. 2002, 92, 7266− 7271. (17) Ishimoto, K.; Tanaka, T.; Ohki, Y.; Sekiguchi, Y.; Murata, Y. Elect. Eng. Jpn. 2011, 176, 1−7.

5937

DOI: 10.1021/acs.nanolett.6b02920 Nano Lett. 2016, 16, 5934−5937