Tantalum Oxide

Subscriber access provided by GAZI UNIV

Article

Low Dimensional Polyoxometalate Molecules/Tantalum Oxide Hybrids for Non-Volatile Capacitive Memories Angelika Balliou, Giorgos Papadimitropoulos, George Skoulatakis, Stella Kennou, Dimitrios Davazoglou, Spiros Gardelis, and Nikos Glezos ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.5b11204 • Publication Date (Web): 29 Feb 2016 Downloaded from http://pubs.acs.org on March 2, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Applied Materials & Interfaces is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Low Dimensional Polyoxometalate Molecules/Tantalum Oxide Hybrids for Non-Volatile Capacitive Memories Angelika Balliou,

Kennou,

†Institute

¶

∗,†,‡

Giorgos Papadimitropoulos,

Dimitrios Davazoglou,

†

†

George Skoulatakis,

Spiros Gardelis,

§

¶

and Nikos Glezos

Stella

†

of Nanoscience and Nanotechnology, NCSR Demokritos, Aghia Paraskevi, Athens 15310, Greece

‡National

Technical University of Athens, Iroon Polytexneiou 9, 157 80 Athens, Greece

¶Department

of Chemical Engineering, University of Patras, University Campus, Patras 26504, Greece

§Department

of Solid State Physics, National and Kapodistrian University of Athens, University Campus, Zografos, 157 84, Athens, Greece

E-mail: [email protected]

Phone: +30 210 6503202. Fax: +30 210 6511723

Abstract Transition metal oxide hybrids comprising of high surface-to-volume ratio Ta2 O5 matrices and a molecular analogue of transition metal oxides, tungsten polyoxometalates ([PW12 O40 ]3 ), are introduced herein as charge storage medium in molecular nonvolatile capacitive memory cells. The polyoxometalate molecules are electrostatically self-assembled on a low dimensional Ta2 O5 matrix, functionalized with an aminosilane molecule with primary amines as the anchoring moiety. The charge trapping sites are 1

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

located onto the metal framework of the electron-accepting molecular entities as well as on the molecule/oxide interfaces which can immobilize negatively charged mobile oxygen vacancies. The memory characteristics of this novel nano-composite were tested using no blocking oxide for extraction of structure-specic characteristics. The lm was formed on top of 3.1 nm-thick SiO2 /n-Si(001) substrates and has been found to serve as both SiO2 /Si interface states reducer (i.e. quality enhancer) and electron storage medium. The device with the polyoxometalates sandwiched between two Ta2 O5 lms results in enhanced internal scattering of carriers. Thanks to this, it exhibits signicantly larger memory window than the one containing the plain hybrid and comparable retention time, resulting in a memory window of 4.0 V for the write state and a retention time around 10 4 sec without blocking medium. Dierential distance of molecular trapping centers from the cell's gate and electronic coupling to the space charge region of the underlying Si substrate were identied as critical parameters for enhanced electron trapping for the rst time in such devices. Implementing a numerical electrostatic model incorporating structural and electronic characteristics of the molecular nodes derived from scanning probe and spectroscopic characterization we are able to interpret the hybrid's electrical response and gain some insight in the electrostatics of the trapping medium.

Keywords Molecular capacitive memories, transition metal oxide/molecular hybrids, self-assembled molecular layers, interface engineering, electron trapping, CMOS-compatible technology

1 Introduction In the quest for solid state memories' downscaling approaches other than photolithographic size reduction, like the introduction of molecular components, hybrid materials and three dimensional nanostructuring are gaining ground. 1 The remarkable progress made lately in 2


Page 2 of 31

Page 3 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


the elds of organic and molecular electronics has enabled successful fabrication of moleculebased hybrid devices, among which promising alternatives in the eld of non-volatile memories. 25 Such devices aim at exploiting the bi-stability of states either in conductance, 69 magnetic 10 and optical properties, 11 polarization, 12 or structure 4 upon an external stimulus to obtain binary switching functionality. The appeal of molecular devices lies on properties that are inherent in molecules like discrete quantum states and, thus, charge localization, oering tolerance on oxide defects pretty much as with oating dots 13 and, additionally, signicantly lower dimensions and multiple charging state addressing. To this end molecules with reversible redox properties seem to present an advantage. 1417 On the other hand, beyond charging eciency, one of the most fundamental limits in speeding up memory devices is the increase of the tunneling current which follows thickness downscale of the dielectric layer and renders the device unreliable. 18 High-k oxides are now being used to reduce the leakage current while maintaining the same capacitance. 19,20 Promising as it may seem, the combination of these two approaches is a challenging task due to constrains like the demand for low temperature growth and the precision control of the oxide/molecule interface. 21 Moreover, molecular components suer from variability and high extraction losses at interfaces, 22 facts impeding their technological competitiveness. Therefore, the need for molecule-based devices with controlled coupling of the functional molecules to the support layer and/or the addressing leads and low contact resistance remains an open issue. This has triggered the research on defect-tolerant 23 CMOS-compatible processes 24 and minimization of charge injection barriers. 25 We propose here that transition metal oxides, in which bonding is intrinsically neither purely covalent nor ionic or metallic, can be a promising candidate for the formation of interfaces of low energy cost with molecular assemblies. 26 Salient features of these oxides are their potential for molecular-friendly (i.e. low temperature) growth and their typically high dielectric constant, corroborating their appeal for molecular memory applications. Hybrid molecular/Ta 2 O5 memory cells based on self-assembled tungsten polyoxometa3



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

late molecules (POMs, [PW 12 O40 ]3 ) were fabricated and characterized. POMs are able to attach and delocalize electrons on their metal framework without undergoing structural damage. 14,27,28 The interfaces involved in this cell bear localized dipoles of molecular origin, able to inuence the band alignment of the MOS stack. 29,30 This phenomenon along with the attainment of electronic coupling of the trapped electrons to the space charge region (scr) of the Si body can be exploited to build a hybrid memory cell 31 whilst improving the properties of the high-k dielectric by mitigating xed and/or interface charges. 32 These capacitive memory structures extend current achievements in the eld of CMOS-compatible processes for the fabrication of molecular devices, 3335 provide, similarly to, 36 a way for direct access of the POM switching abilities in a MOS ash memory and introduce a new molecular hybrid of signicantly improved memory window and commendable charge retention characteristics. The presented results demonstrate the suitability of this hybrid to be incorporated as high density storage node in advanced, solution-processed molecular nonvolatile ash memory arrays.

2 Experimental Details 2.1

Materials

12-Tungstophosphoric acid hydrate (POM, HPW 12 O40 · xH2 O, Aldrich), 3-aminopropyl triethoxysilane (APTES, Aldrich) and isopentylamine (IPA, Aldrich) were all of analytical grade and used without further purication. Hydrochloric acid, sulfuric acid and hydrogen peroxide (all obtained from Aldrich, analytical grade), were used for the pH adjustment of the deposition solutions and the purication and hydrophilization of the SiO 2 surface respectively. Deionized water with a resistivity of 15 M Ωcm−1 prepared from the Milli-RO plus 90 apparatus (Millipore) was used in all experiments. The substrates used in the molecular nanostructures fabrication varied according to the characterization method applied; quartz slides were used for the UV study and silicon wafers with 3.1 nm dry oxide were used for 4


Page 4 of 31

Page 5 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


the Fourier transform infrared spectroscopy (FTIR), scanning electron microscopy (SEM), and X-ray photoelectron spectroscopy (XPS) studies. All substrates were used after cleaning with piranha solution (H 2 O2 :H2 SO4 , 1:1 v/v).

2.2

Fabrication of capacitive memory cells

We fabricate memory capacitor testbeds which incorporate the material under test in the active region. The process consists of standard photo-lithography and etching on wet-oxidized

Figure 1: Schematics of the fabricated stacks along with three SEM snapshots depicting the growing Ta2 O5 lm. In gure a) Ta 2 O5 is grown on POM molecules self-assembled on the active region in b) Ta 2 O5 is directly grown on SiO 2 and the molecules are post assembled on top and in c) a passivated POM layer is being sandwiched between two Ta 2 O5 lms. (500 nm SiO2 ) n-type Si substrates 1-2 Ωcm dening square patterns with side dimensions of 100 µm. These constitute the active areas of the devices. A 3.1 nm-thick tunneling oxide is grown on top via dry oxidation, in order to minimize the trap density at the Si/oxide inter5



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

face. 37,38 For the assembly of the molecular traps the underlying oxide is chemically modied with APTES, a type of aminosilane molecule used for the electrostatic self-organization of POM anions via the Layer-by-layer (LBL) process. 34,39,40 Passivation of the immobilized POMs is attained through an IPA capping monolayer and is crucial in providing electronic decoupling from the non-injecting side of the device. 33,41 The device is sealed with a 500nm-thick Al gate electrode. Three types of devices are fabricated as depicted in Figure 1: a) P-Ox where Ta2 O5 epitaxy follows the assembly of molecular layers on SiO 2 , b) Ox-P where Ta2 O5 is grown directly on top of SiO 2 and the molecules follow, and c) Ox-P-Ox where the passivated POM layer is sandwiched between two Ta 2 O5 lms. For each device type three dierent Ta2 O5 thicknesses were tested, namely 25 nm, 80 nm and 145 nm. Ta 2 O5 is grown via evaporative deposition of oxidized tantalum wire under N 2 environment and ambient substrate temperature.

2.3

Characterization

All electric measurements were performed in shielded probe station using a HP4041B picoammeter, a HP4284A precision LCR meter and an HP 8110A 150 MHz pulse voltage generator. For the XPS measurements unmonochromatized Mg Ka line at 1253.6 eV (12 kV with 20 mA anode current) and an analyzer (Leybold EA-11) pass energy of 100 eV were used at 0 degree take-o angle. For the UV photoemission spectra (UPS) the He I (21.2 eV) excitation line was used and a voltage of 12.23 eV was applied to the specimen in order to separate the high binding energy cut-o from the analyzer. UV-vis absorption spectra were obtained on a Perkin-Elmer UV-vis Lambda 40 spectrophotometer. FTIR transmittance spectra were recorded on a Bruker, Tensor 27 spectrometer using 1000 scans at 4 cm−1 spectral resolution. For the Photoluminescence (PL) measurements the 458 nm line of an Ar+ laser was used. The signal was analyzed with a Jobin Yvon SPEX HR320 monochromator and detected by a photomultiplier tube in reection geometry. Atomic force microscopy (AFM) measurements were carried out using a TS-150 NT-MDT Solver Scanning Probe Microscope and an ultra 6


Page 6 of 31

Page 7 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


sharp diamond-like carbon tip (curv. rad. 1-3 nm) with Au reective side in tapping mode.

3 Results and discussion 3.1

Investigation of the hybrid's morphology and electronic structure

The Ta2 O5 matrix morphology is investigated with FE-SEM and AFM. In early deposition stages (i.e. 25 nm thickness) uniform tantalum oxide islands are being developed (Figure 1 right). As growth evolves we meet the cases of 80 and 145 nm, ending up with clear formation of dendrite nanocrystalline structures. For the 25 nm case, the average rms roughness (Sq) obtained via AFM decreases with the addition of POM molecules from 1.7 nm to 0.3 nm. Although this indicates a surface modication, fractal analysis reveals a hidden correlation between the two structures. When analyzed with a box counting algorithm 42,43 both surfaces are found to exhibit scale invariance with a fractal dimension (df) of 1.8 for the molecular layer and 1.7 for the Ta 2 O5 substrate. The AFM-derived binary images used for fractal analysis can be seen in Figure 2 against the original scans to enable comparison. In the case of the molecule-modied structures this df value is consistent with a Diusion Limited Aggregation formation process. 44 The lower fractal dimension of Ta 2 O5 conveys its inuence on the topology of the molecular layer, 42 i.e. the electronic and structural coupling of POM molecules to the tantalum oxide matrix. XPS of the 25 nm as-grown lm indicates stoichiometric Ta 2 O5 , containing a certain amount of oxygen vacancies. As seen in the right part of Figure 3 the O1s peak can be analyzed in two peaks, with the rst located at a binding energy of 530.7 ± 0.1eV corresponding to oxygen bonded to tantalum in Ta 2 O5 , in agreement to the literature 45 and the other corresponding to oxygen related to surface contamination ( 532.0 ± 0.1eV ). 46 The main peak of Ta4f7/2 is located at a binding energy of 26.4 ± 0.1eV (Figure 3 left) corresponding to tantalum bonded with oxygen in stoichiometric Ta 2 O5 . 45,47 The higher binding energy 7



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 2: AFM images (top) and created binary images used for fractal analysis (bottom) of the 25-nm-thick Ta 2 O5 template before (a) and after (b) the self-assembly of the molecular clusters. Inset (c) corresponds to the morphology of the POM molecular layer when developed on a "at" (sub-nm-roughness) SiO 2 surface. (d) and (e) are the binary images produced from AFM scans and used for fractal analysis after convergence of the box-counting algorithm.

8


Page 8 of 31

Page 9 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


peak of the 4f doublet appears at 28.4 eV and corresponds to Ta4f5/2 in Ta 2 O5 , according to literature. 47,48 In Figure 4 UV-Vis absorption measurements indicate a shift of the

Figure 3: Ta4f (left) and O1s (right) XP spectra of the 25-nm-thick as-grown Ta 2 O5 lm. Ta2 O5 absorption edge towards higher wavelengths upon POM deposition, indicating an effective narrowing of the gap. The characteristic POM peaks remain intact in the absorption spectrum of the composite material, indicating the integrity of the molecular structures. Moreover, the intensity of the POM W-O transfer peak at 270 nm is tripled in the composite material, owing to its large eective surface. Along the same lines, the four FTIR bands of the Keggin structure POM 49,50 remain unaltered upon deposition of Ta 2 O5 indicating no bonding disruption upon the Ta 2 O5 -POMs interaction. The UPS valence band spectra of sample Ox-P reveal the presence of a shallow intra-gap state centered at 1.3 eV within the Ta2 O5 gap facilitating charge injection onto the states of the hybrid. The appearing gap states are attributed to the W5d band of W in the POM framework, and result from the tailing of electron-rich orbitals of reduced POM structures within the gap. 33 When POM is sandwiched between tantalum oxide lms there is a sloping up tail before the high binding energy edge. This is indicative of increased inelastic scattering of the emitted valence band electrons on their way to the sample's surface. Oxygen vacancies and oxygen interstitials have been known to emit in the visible part of the spectrum when excited, owning to the provision of alternate paths for the photo-induced 9



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 4: Top: UV-Vis absorbance of pure POM and Ta 2 O5 lms along with the spectrum of the composite Ox-P type material and its spectral deconvolution next to FT-IR spectra. Down: Valence band spectra of the materials under test. On the right the region near the Fermi level is presented in magnication and on the left the high binding energy cut-o.

10


Page 10 of 31

Page 11 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


carrier generation and recombination leaving their ngerprint on both PL and photoabsorption characteristics. 5153 Figure 5 shows the PL spectra of three sample congurations, Ox, P-Ox and Ox-P excited with the 458 nm line of an Ar+ laser. These spectra all exhibit an asymmetric broad PL band mainly in the cyan-green region. This band can be resolved by

Figure 5: PL spectra of as-deposited Ta 2 O5 (Ox) (a), Ta2 O5 grown on POM-modied substrate (P-Ox) (b) and POM post-functionalized Ta 2 O5 (Ox-P) (c). The main band is resolved by means of Gaussians in three discrete sub-bands presented in (d) and attributed to surface (lower wavelength peak) and volume emission centers, as well as to contribution from POM luminescence (red peak). In Figure (d) the lines are displayed as a guide for the eye. means of Gaussians in three discrete sub-bands attributed to specic emission centers and identied in all three cases ( Ox, P-Ox and Ox-P ). The rst band, centered in green (around 500 nm for Ox-P and P-Ox, and at 517 nm for Ox), is almost identical in width and intensity for samples Ox and P-Ox, while it is slightly suppressed in sample Ox-P, where the Ta 2 O5 surface is post-modied, highlighting its surface origin. The intensity of the second band (located in green as well) is related to the oxygen content of the sample (the oxygen percent11



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

age is estimated from XPS stoichiometric analysis) pointing out its correlation to volume centers. Interestingly this band appears centered towards yellow (576 nm) and broadened in the case the POM molecules are not present. Yellow emissions are being attributed to the presence of oxygen interstitials 54 which, in our case co-exist with oxygen vacancies (thus the broad band) in the volume of the Ta 2 O5 , constituting its volume oxygen defects. This band shifts further to green when POM-originating oxygen vacancies 33 are entering the structure, as these vacancies tend to passivate part of the interstitials. As a result, the band is also narrowed in accordance to the suppression of the energy distribution of emissive centers. The third contribution on the PL emission is directly related to the presence of POM molecules and is, therefore, not present in the case of Ox sample. This band centered at around 630 nm is attributed to the O2p → W 5d intra-band transition in reduced POM molecules. 33 Since both the intensity and the wavelength of the green emissions are correlated to oxygen related defects in the lm, these bands can be attributed to oxygen point defects in agreement with previously reported results. 55

3.2

Electrical characterization

Typical bidirectional capacitance-voltage (C-V) characteristics are shown in Figure 6. The voltage sweep direction was from inversion to accumulation and back to inversion. No hysteresis has been detected for the reference SiO 2 sample. In contrast, all Ta 2 O5 samples exhibit signicant hysteresis that tends to increase with increasing thickness. The hysteresis width is quite large in the case of 145-nm-thick oxide (1.83 V) and reveals a favorable hole injection mechanism, a fact attributed to the presence of negatively charged oxygen vacancies in its structure. The reference sample is suering from the presence of xed positive charges, with a surface density of 3.8 × 10−7 Cb/cm2 (Figure 6). The fact that these charges are compensated upon the deposition of thick layers of tantalum oxide is compatible with the PL results indicating the presence of negative vacancies in Ta 2 O5 . In all cases addition of POM molecules on Ta 2 O5 shifts the at band voltage to more positive values, further 12


Page 12 of 31

Page 13 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 6: 1 MHz C-Vs under bi-directional voltage sweep for the two extreme Ta 2 O5 thicknesses of a) 145 and b) 25 nm. Electron trapping (positive Vfb shift) is signicant only for cases Ox-P. compensating the xed charge. In the case of 25nm thickness (Figure 6b) the hysteresis loop is mainly dependent on the presence of the POM molecules and involves pure electron storage. This is the case of interest for a memory cell and it will be further investigated. Figure 7a, b show the room temperature conductance characteristics of the control (only SiO2 ) and Ox samples respectively for several gate voltages. Both samples show qualitatively almost identical behavior, although the losses in sample Ox are almost two times lower than in the control SiO2 sample. This provides further evidence on interface trap reduction upon Ta2 O5 deposition. In weak inversion a quite narrow conductance peak dominates the normalized conductance over frequency, Gp /ω , spectrum with a maximum at quite high frequencies, the same for both samples. From this value the interfacial trap density (Dit) is estimated at 7.3 × 1011 cm−2 eV −1 (time constant distributed around 1.12 µsec). The rapidly diminishing maximum as gate bias shifts the Fermi level some kT/q quanta away from the interface trap energy level, indicates a single-level interface trap. 56 Upon Ta2 O5 deposition 13



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 7: Normalized conductance vs frequency for several gate voltages for samples a) SiO 2 and b) Ox.

14


Page 14 of 31

Page 15 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(Figure 7b) the existence of a signicant amount of slow traps induces signicant losses at low frequencies as well. As a result, there is exhibition of a second trap level with time constants distributed around 8 msec. Despite the fact that these traps are spaced away from the SiO2 -Si interface they are detectable in the admittance measurements, implying electronic transitions between electrons localized on Ta 2 O5 trapping sites and the scr. The very weak bias dependence of the amplitude of maximum loss of the latter peak suggests the presence of a whole distribution of traps with energies throughout the gap subjected to band bending variations. 56

3.3

Memory cell performance

In order to investigate the electron injection/trapping process under programming operation, the incremental-step-pulse programming (ISPP) method was employed, 57 i.e., programming (positive) voltage pulses with constant increasing step height (1 V) were applied on the gate and the at band voltage (Vfb) was measured between two successive pulses. Pulse widths of 100 µsec and 100 msec were tested. The results are presented in Figure 8 for the 25 nm-thick Ta2 O5 /POM hybrid which exhibited signicant electron storage capacity. The

Figure 8: Programming windows for cells P-Ox, Ox-P and Ox-P-Ox obtained by the ISPP method (see text) utilizing two dierent pulse durations in order to investigate the programming speed. memory window is signicantly aected from the duration of the applied pulse. The slopes 15



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

dVf b /dVgate of the characteristics of devices Ox-P and P-Ox (Figure 8) are not aected by the width of the pulse. Charge trapping, though, is more ecient (i.e. requires signicantly lower voltages) at shorter pulses. On the contrary, the Ox-P-Ox sandwich cell is responsive at slow (100 msec) pulses. This is a direct consequence of the inuence of the Ta 2 O5 slow traps discussed in the context of Figure 7b. Enhanced scattering within the Ta 2 O5 layer (UPS discussion) prohibits short time back-tunneling, boosting the trapping ability. The process, though, is limited by the vacancy response frequency, which is of the order of msec (Figure 7b). Thus, the 100 msec pulse results in enhanced electron trapping onto the available POM orbitals, extending the time over which charge capture can occur. Annealing of the Ox-P-Ox device at 320 o C (this temperature does not disrupt POM structure) under inert environment (N2 ) results in a Gaussian type ISPP response, typical for single level traps, suggesting that the distribution of trapping sites detected in the Gp /ω versus f characteristics might play a crucial role in the enhanced trapping performance of Ox-P-Ox cells. To examine charge retention cells Ox-P, P-Ox and Ox-P-Ox were programmed via single pulses of +14 V, +10 V and +20 V (charging favorable pulses) respectively. The transient memory window (Vfb shift vs time) is shown in Figure 9. The device containing the POM

Figure 9: Charge retention characteristics of single-pulse programmed cells. layer buried under the tantalum oxide (P-Ox) maintains almost the half charge than the device with POM onto the porous tantalum oxide (Ox-P). This manifests the impact of 16


Page 16 of 31

Page 17 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


the large Ta2 O5 surface. In the Ox-P-Ox structures multiple inelastic scattering across the structure (see UPS) is proposed as the reason for their enhanced trapping eciency, owing to reduced back-tunneling. For this structure we get a large amount of trapped charge constantly increasing up until 20 V. All congurations exhibited comparable retention time and did not show closing until after 104 sec. In short, Ox-P cells exhibit enhanced speed, but limited charging ability due to pronounced back-tunneling, while Ox-P-Ox cells, although performing at lower programming speed, oer large memory window.

3.4

Electrostatics of the trapping medium

In order to quantify the distribution of charge across the Ta 2 O5 /POM trapping medium and investigate the details of the charging mechanism we consider the simple case of a Si/SiO2 /Ta2 O5 /POM/Al stack. The alignment of all relevant electronic levels and states before and after voltage application is depicted in Figure 10 along with the UPS valence band spectra of the Ox-P structure. The distribution of POM molecules across the stack is approximated with successive discrete monolayers which are embedded in the Ta 2 O5 lm with the molecules closely packed. The hybrid exhibits sub-nm surface roughness (AFM, subsection 3.1) meaning that, most probably, the amount of POM molecules protruding from the Ta2 O5 matrix and contributing mainly to the POM/Al interface are of the order of a monolayer. The LUMO of these interfacial molecules lies slightly below the Al fermi level, 58 forming locally an ohmic contact (Figure 10). All molecular HOMOs and LUMOs are treated as continuous densities of states, rather than discrete levels, 41 with characteristics derived from UV-Vis spectroscopy and UPS. In Figure 10 the POM HOMOs and LUMOs, depicted as bands for simplicity, are subjected to dierential energetic shift according to their distance from the Al gate and the Si scr. The LUMO shift of the top POM layer towards the Al fermi level has also been taken into account according to 58 The POM lm is discretized into layers of thickness 1nm, the diameter of POM molecules, 17



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 10: (Top): Scheme of molecular Density of States (DOS) shift due to dierential distance from the scr a) Upon contact and before voltage application, b) After voltage application. (Bottom): UPS valence band spectra and deconvolution of signal by means of Gaussians for cell Ox-P. The lateral graphs depict the extracted high binding energy cut-o position and valence band onset.

18


Page 18 of 31

Page 19 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


along the growth axis of the Ta 2 O5 /POM nanocomposite. An iterative 1-D electrostatic model is applied and the equations that involve the voltage-dependent occupancy of trapping sites in conjunction with the Poisson equation are solved across the discretized structure in a self-consistent manner. The thickness of the hybrid, zmax , is determined from cross sectional FE-SEM images. The Ta 2 O5 conduction band minimum ( Ec ) and valence band maximum (Ev ) values are derived from UV-Vis absorbance and UPS measurements while the corresponding values for Si, SiO 2 and the Al Fermi level are obtained bibliographically. 59,60 Following 41 we model the HOMO as a Gaussian with its maximum and standard deviation (σ ) derived from UV-Vis spectroscopy and UPS respectively. More specically, σ is approximated by the width of the HOMO-related feature of UPS valence band spectra (i.e. σ =0.5 eV). The bandgap value, Eg , is obtained via calculation of the absorption coecient from UV-Vis spectra, linear t of the experimental data in the vicinity of bandgap energy and calculation of the intersection point of linear t with the abscissa (i.e. Eg =3.8 eV). From UPS the ionization energy, IE, is obtained (IE=6.46 for POM, 6.5 eV for Ta 2 O5 ) as the energy dierence between the vacuum level and the low-binding energy onset of the UPS spectrum. The position of the HOMO maximum (peak of the Gaussian distribution) is located 2 σ away from the IE onsets (i.e. EHOM O =7.5 eV). The LUMO maximum is obtained by shifting the onset of this distribution by Eg + 2σ along the energy axis according to. 41 The molecular HOMO and LUMO of each monolayer are being shifted dierentially, by the local electron potential energy q × V (zi ), according to their distance from the scr (Figure 10). This means that with increasing distance z from the accumulation region of the semiconductor, the molecular DOS shifts to progressively higher binding energy by the local electron potential energy, while the electron injection barrier becomes progressively lower. 41 The charge density σ(zi ) on each of the POM layers is then calculated via Eq. 1. 41

19



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Z∞ σ(zi ) = q × nP OM × α ×

1

dE ×

1 gHOM O

−∞

Z∞ −

dE ×

−∞

e−β(E−EF )+1

Page 20 of 31

× GHOM O [E + qV (zi )] (1)

1 1 e+β(E−EF )+1 gLU M O

× GLU M O [E + qV (zi )]

,where q is the absolute value of elementary charge, nP OM the number of POM molecules per unit area and per discretization interval ∆zi calculated as 7.89 × 1013 /(cm2 ∆z) considering adaptation of a bcc lattice for α-Keggin POMs 61 and a 78.9% lling factor ( α) of the hybrid with molecules (see inclusions estimation below), Dirac distribution for HOMO and

1 1 e+β(E−EF )+1 gLU M O

1 1 e−β(E−EF )+1 gHOM O

is the Fermi-

for LUMO respectively. gHOM O , gLU M O

are the spin degeneracies of levels both considered to be 2, β = 1/kB T with kB the Boltzmann constant and T the temperature, set to 300 K. The energy distributions of HOMO and LUMO levels are GHOM O [E + qV (zi )] and GLU M O [E + qV (zi )]. The total thickness of the hybrid, i.e. zmax (25 nm), is determined from cross sectional FE-SEM images. The electrostatic potential across the lm is initially assumed that of a parallel plate capacitor under bias (V). The corresponding electric potential Ecap z is fed in Eq. 1 to obtain an initial guess for the charge density of the rst layer, σ(z0 ). The rst Al/POM interfacial layer is considered fully occupied, as indicated from UPS spectra, 58 and is incorporated as a xed charged sheet at z=0. This σ(z0 ) is plugged in as a starting value into the one dimensional i) to obtain the value of next layer's local potential V (zi ), Poisson equation: 52 V (zi ) = − σ(z 0

which is calculated via application of superposition principle as the vector sum of the capacitive electrostatic eld and the elds originating from the trapped charge of all preceding layer(s). 0 is the vacuum permittivity and an eective dielectric constant for the hybrid. The volume fraction of inclusions in as-grown Ta 2 O5 has been estimated at 78.9% via highfrequency (1MHZ) C-V and application of the Maxwell Garnett equation. 62 We assume a static dielectric constant of 25, 63,64 typical of amorphous Ta 2 O5 , for the nanocrystalline

20


Page 21 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Ta2 O5 matrix (the diameter of the nanocrystals has been determined via transmission electron microscopy (TEM) imaging and is of the order of 1 nm with no preferred orientation. The material is, thus, considered to be amorphous in terms of dielectric constant). Since POM molecular structures are expected to be highly hydrated due to crystallization water 65 a water dielectric constant is used for the POM-lled inclusions in the hybrid. This gives an eective dielectric constant of 62.7 for the composite. The DOS of each discretization interval is then shifted in energy by the respective q × V (zi ), as shown in Figure 10, and the next σ(zi ) is again calculated by occupying all states in the molecular hybrid up to the Fermi level, EF . The process repeats itself iteratively by re-calculating the discrete V (zi ) of the next layer every time one σ(zi ) is obtained. We use a Dirichlet [V(0)=0] boundary condition and apply Gauss law. The result of the calculation is presented in Figure 11.

Figure 11: Charge density across the trapping (Ta 2 O5 /POM) layer. The plot indicates a distribution of charge density across the hybrid, owing to both dierential distance of trapping levels from the accumulation layer and electronic coupling of the molecular HOMOs-LUMOs to the Ta 2 O5 matrix (observed energetically as LU M OP OM −

EvTa2 O5 alignment). The existence of, essentially, a continuum of trapping states instead of a single level and the lowering of electron injection barrier for layers spaced away from the scr, are the origin of enhanced trapping eciency in the POM/Ta 2 O5 hybrid. When an additional Ta 2 O5 lm is deposited on top of the hybrid (sandwich structure) and positive bias is applied on gate, a Schottky barrier is being formed onto the metal-oxide 21



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

interface. Thus, the mobile O 2 ions present in the Ta 2 O5 structure cannot migrate with the same eciency towards the gate, remaining immobilized onto the interfaces. Driving force of this ionic charge immobilization is the lattice mismatch on the interfaces of POMmodied and non-modied transition metal oxides. Such mismatches cause modulation of the structural and electronic properties of the lm, changing the local work function and resulting in adsorption of charged species 66 at the Ta2 O5 /POM interfaces. This results in the monitored expansion of the memory window. The process is limited by the vacancy response frequency, which corresponds to a lifetime of the order of msec, giving rise to the response enhancement of the Ox-P-Ox cell at slow pulse duration. At the same time, enhanced scattering within the Ta 2 O5 layer prohibits short time back-tunneling enhancing the trapping ability.

4 Conclusions Hybrids comprising of high surface-to-volume ratio transition metal oxides (i.e. Ta 2 O5 ) and a transition metal oxide molecular analogue, Keggin type tungsten polyoxometalates, consist a novel composite material that can be fabricated using CMOS-compatible technology and serve as a memory cell based on electron trapping. Non-volatile memory devices with commendable memory characteristics such as large memory window (4.0 Volt for the write state) and signicant retention (10 4 sec) even without the use of blocking oxide have been fabricated utilizing a MIS capacitive cell architecture. We demonstrate for the rst time that key issues for boosting the width of memory window of such devices are dierential distance of trapping sites from the injection layer and eective electronic coupling to the scr of the underlying Si substrate. The promising memory and retention characteristics of this hybrid highlight its potential as storage node for molecular nonvolatile ash memories. The fabricated memory cells are able to bring together signicant gures of merit like high density of trapping sites via the combination of stable molecular components and 3-dimensional

22


Page 22 of 31

Page 23 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


nanostructuring, electronic coupling to Si and direct accessibility of electronic levels in solid state, inherent retention of information under room temperature, bottom up fabrication logic and solution processing. They are, thus, of great potential for advanced molecule-based storage and memory technologies.

References (1) International Technology Roadmap for Semiconductors: 2014 ITRS White Papers. Semiconductor Industry Association, San Francisco, CA

2014, ITRS 2.0.

(2) Tseng, C.-W.; Huang, D.-C.; Tao, Y.-T. Organic Transistor Memory with a Charge Storage Molecular Double-Floating-Gate Monolayer. ACS Appl. Mater. Interfaces

2015, 7, 97679775. (3) Chi, H.-Y.; Hsu, H.-W.; Tung, S.-H.; Liu, C.-L. Nonvolatile Organic Field-Eect Transistors Memory Devices Using Supramolecular Block Copolymer/Functional Small Molecule Nanocomposite Electret. ACS Appl. Mater. Interfaces 2015, 7, 56635673. (4) Radha, B.; Sagade, A. A.; Kulkarni, G. U. Metal-Organic Molecular Device for NonVolatile Memory Storage. Appl. Phys. Lett. 2014, 105 . (5) Shaw, J.; Zhong, Y.-W.; Hughes, K. J.; Hou, T.-H.; Raza, H.; Rajwade, S.; Bellfy, J.; Engstrom, J. R.; Abruña, H. D.; Kan, E. C. Integration of Self-Assembled Redox Molecules in Flash Memory Devices. IEEE Trans. Electron Devices 2011, 58, 826 834. (6) Tang, W.; Shi, H.; Xu, G.; Ong, B. S.; Popovic, Z. D.; Deng, J.; Zhao, J.; Rao, G. Memory Eect and Negative Dierential Resistance by Electrode-Induced Two-Dimensional Single-Electron Tunneling in Molecular and Organic Electronic Devices. Adv. Mater.

2005, 17, 23072311. 23



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(7) Green, J. E.; Choi, J. W.; Boukai, A.; Bunimovich, Y.; Johnston-Halperin, E.; DeIonno, E.; Luo, Y.; Sheri, B. A.; Xu, K.; Shin, Y.; Shik, A 160-Kilobit Molecular Electronic Memory Patterned at 1011 Bits per Square Centimetre. Nature 2007, 445,

414417.

(8) Wang, C.; Gu, P.; Hu, B.; Zhang, Q. Recent Progress in Organic Resistance Memory with Small Molecules and InorganicOrganic Hybrid Polymers as Active Elements. J. Mater. Chem. C

2015, 3, 1005510065.

(9) Gu, P.-Y.; Zhou, F.; Gao, J.; Li, G.; Wang, C.; Xu, Q.-F.; Zhang, Q.; Lu, J.-M. Synthesis, Characterization, and Nonvolatile Ternary Memory Behavior of a Larger Heteroacene with Nine Linearly Fused Rings and Two Dierent Heteroatoms. J. Am. Chem. Soc.

2013, 135, 1408614089.

(10) Khan, M.; Bhansali, U. S.; Alshareef, H. High-Performance Non-Volatile Organic Ferroelectric Memory on Banknotes. Adv. Mater. 2012, 24, 21652170. (11) Simão, C.; Mas-Torrent, M.; Crivillers, N.; Lloveras, V.; Artés, J. M.; Gorostiza, P.; Veciana, J.; Rovira, C. A Robust Molecular Platform for Non-Volatile Memory Devices with Optical and Magnetic Responses. Nat. Chem. 2011, 3, 359364. (12) Ryndyk, D. A.; DâAmico, P.; Richter, K. Single-spin Polaron Memory Eect in Quantum Dots and Single Molecules. Phys. Rev. B 2010, 81, 115333. (13) Lai, S. Flash Memories: Successes and Challenges. IBM J. Res. Dev. 2008, 52, 529535. (14) Volatron, F.; Noël, J.-M.; Rinfray, C.; Decorse, P.; Combellas, C.; Kanou, F.; Proust, A. Electron Transfer Properties of a Monolayer of Hybrid Polyoxometalates on Silicon. J. Mater. Chem. C 2015, 62666275. (15) Flood, A. H.; Wong, E. W.; Stoddart, J. F. Models of Charge Transport and Transfer

24


Page 24 of 31

Page 25 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


in Molecular Switch Tunnel Junctions of Bistable Catenanes and Rotaxanes. Chem. Phys.

2006, 324, 280290.

(16) Zhu, H.; Hacker, C. A.; Pookpanratana, S. J.; Richter, C. A.; Yuan, H.; Li, H.; Kirillov, O.; Ioannou, D. E.; Li, Q. Non-Volatile Memory with Self-Assembled Ferrocene Charge Trapping Layer. Appl. Phys. Lett. 2013, 103 . (17) Hu, B.; Wang, C.; Wang, J.; Gao, J.; Wang, K.; Wu, J.; Zhang, G.; Cheng, W.; Venkateswarlu, B.; Wang, M.; Lee, P. S.; Zhang, Q. InorganicOrganic Hybrid Polymer With Multiple Redox for High-Density Data Storage. Chem. Sci. 2014, 5, 34043408. (18) Buchanan, D. A. Scaling the Gate Dielectric: Materials, Integration, and Reliability. IBM J. Res. Dev.

1999, 43, 245264.

(19) Groner, M.; George, S. High-k Dielectrics Grown by Atomic Layer Deposition: Capacitor and Gate Applications. Interlayer Dielectrics for Semiconductor Technologies

2003, 1, 327. (20) Bohr, M.; Chau, R.; Ghani, T.; Mistry, K. The High-k Solution. IEEE Spectrum 2007, 44,

2329.

(21) Flood, A. H.; Stoddart, J. F.; Steuerman, D. W.; Heath, J. R. Whence Molecular Electronics? Science 2004, 306, 20552056. (22) Lörtscher, E. Wiring Molecules into Circuits. Nat. Nanotechnol. 2013, 8, 381384. (23) Heath, J. R.; Kuekes, P. J.; Snider, G. S.; Williams, R. S. A Defect-Tolerant Computer Architecture: Opportunities for Nanotechnology. Science 1998, 280, 17161721. (24) Rose, G. S.; Cabe, A. C.; Gergel-Hackett, N.; Majumdar, N.; Stan, M. R.; Bean, J. C.; Harriott, L. R.; Yao, Y.; Tour, J. M. Design Approaches for Hybrid CMOS/Molecular Memory Based on Experimental Device Data. Proceedings of the 16th ACM Great Lakes symposium on VLSI. 2006; pp 27. 25



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(25) Niederhausen, J.; Amsalem, P.; Frisch, J.; Wilke, A.; Vollmer, A.; Rieger, R.; Müllen, K.; Rabe, J.; Koch, N. Tuning Hole-Injection Barriers at Organic/Metal Interfaces Exploiting the Orientation of a Molecular Acceptor Interlayer. Phys. Rev. B

2011, 84, 165302. (26) Demkov, A. A.; Posadas, A. B. Integration of Functional Oxides with Semiconductors ; Springer, 2014; pp 1718. (27) Pope, M. T.; MÃ¼ller, A. Polyoxometalate Chemistry: An Old Field with New Dimensions in Several Disciplines. Angew. Chem., Int. Ed. Engl. 1991, 30, 3448. (28) Ammam, M. Polyoxometalates: Formation, Structures, Principal Properties, Main Deposition Methods and Application in Sensing. J. Mater. Chem. A 2013, 1, 62916312. (29) Huang, A.; Zheng, X.; Xiao, Z.; Wang, M.; Di, Z.; Chu, P. Interface Dipole Engineering in Metal Gate/High-k Stacks. Chin. Sci. Bull. 2012, 57, 28722878. (30) Yae, O.; Pujari, S.; Sinai, O.; Vilan, A.; Zuilhof, H.; Kahn, A.; Kronik, L.; Cohen, H.; Cahen, D. Eect of Doping Density on the Charge Rearrangement and Interface Dipole at the Molecule-Silicon Interface. J. Phys. Chem. C 2013, 117, 2242222427. (31) Lim, A.-J.; Hou, J.; Kwong, D.-L.; Yeo, Y.-C. Manipulating Interface Dipoles of Opposing Polarity for Work Function Engineering within a Single Metal Gate Stack. Electron Devices Meeting, 2008. IEDM 2008. IEEE International. 2008; pp 14. (32) Miyata, N. Study of Direct-Contact HfO 2 /Si Interfaces. Materials 2012, 5, 512527. (33) Balliou, A.; Douvas, A.; Normand, P.; Tsikritzis, D.; Kennou, S.; Argitis, P.; Glezos, N. Tungsten Polyoxometalate Molecules as Active Nodes for Dynamic Carrier Exchange in Hybrid Molecular/Semiconductor Capacitors. J. Appl. Phys. 2014, 116, 143703. (34) Douvas, A. M.; Makarona, E.; Glezos, N.; Argitis, P.; Mielczarski, J. A.; Mielczarski, E.

26


Page 26 of 31

Page 27 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Polyoxometalate-Based Layered Structures for Charge Transport Control in Molecular Devices. ACS Nano 2008, 2, 733742. (35) He, T.; He, J.; Lu, M.; Chen, B.; Pang, H.; Reus, W. F.; Nolte, W. M.; Nackashi, D. P.; Franzon, P. D.; Tour, J. M. Controlled Modulation of Conductance in Silicon Devices by Molecular Monolayers. J. Am. Chem. Soc. 2006, 128, 1453714541. (36) Busche, C.; Vilà-Nadal, L.; Yan, J.; Miras, H. N.; Long, D.-L.; Georgiev, V. P.; Asenov, A.; Pedersen, R. H.; Gadegaard, N.; Mirza, M. M.; Paul, D. J.; Poblet, J. M.; Cronin, L. Design and Fabrication of Memory Devices Based on Nanoscale Polyoxometalate Clusters. Nature 2014, 515, 545549. (37) Mahata, C.; Oh, I.-K.; Yoon, C. M.; Lee, C. W.; Seo, J.; Algadi, H.; Sheen, M.-H.; Kim, Y.-W.; Kim, H.; Lee, T. The Impact of Atomic Layer Deposited SiO 2 Passivation for High-k Ta1-x Zrx O on the InP Substrate. J. Mater. Chem. C 2015, 3, 1029310301. (38) Zhang, S. K.; Fu, Z. W.; Ke, L.; Lu, F.; Qin, Q. Z.; Wang, X. Properties of Interface States at Ta2 O5 /n-Si Interfaces. J. Appl. Phys. 1998, 84 . (39) Broekmann, D. K. H. S. C. A., Peter, Ed. Templates in Chemistry III ; Springer Berlin Heidelberg, 2009; Vol. 287. (40) Pope, M.; MÃ¼ller, A. In Polyoxometalate Chemistry From Topology via Self-Assembly to Applications ;

Pope, M., Mueller, A., Eds.; Springer Netherlands, 2001; pp 16.

(41) Oehzelt Martin, H. G., Koch Norbert Organic Semiconductor Density of States Controls the Energy Level Alignment at Electrode Interfaces. Nat. Commun. 2014, 5 . (42) Constantoudis, V.; Patsis, G.; Gogolides, E. Fractals and Device Performance Variability: The Key Role of Roughness in Micro and Nanofabrication. Microelectron. Eng.

2012, 90, 121 125. (43) Barabási, A.-L. Fractal Concepts in Surface Growth ; Cambridge university press, 1995. 27



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(44) Witten, T. A.; Sander, L. M. Diusion-Limited Aggregation. Phys. Rev. B 1983, 27, 56865697. (45) E. Atanassova, D. S. X-ray Photoelectron Spectroscopy of Thermal Thin Ta 2 O5 Films on Si. Appl. Surf. Sci. 1998, 135, 7182. (46) Kim, Y. S.; Shimogaki, Y. X-ray Photoelectron Spectroscopic Characterization of the Adhesion Behavior of Chemical Vapor Deposited Copper Films. J. Vac. Sci. Technol., A

2001, 19, 26422651.

(47) Song, W.; Ying, J.; He, W.; Zhuo, V.-Q.; Ji, R.; Xie, H.; Ng, S.; Ng, S. L.; Jiang, Y. Nano Suboxide Layer Generated in Ta 2 O5 by Ar+ Ion Irradiation. Appl. Phys. Lett.

2015, 106, 031602. (48) Dhanaraj, G.; Byrappa, K.; Prasad, V.; Dudley, M. Springer Handbook of Crystal Growth ;

Springer Science & Business Media, 2010; pp 955956.

(49) Rocchiccioli-Deltche, C.; Fournier, M.; Franck, R.; Thouvenot, R. Vibrational Investigations of Polyoxometalates. 2. Evidence for Anion-Anion Interactions in Molybdenum (VI) and Tungsten (VI) Compounds Related to the Keggin Structure. Inorg. Chem.

1983, 22, 207216. (50) Keggin, J. F. Structure of the Molecule of 12-Phosphotungstic Acid. Nature, Letters to Editor

1933, 132 .

(51) Cai, J.-w.; Xu, J.-p.; Zhang, X.-s.; Niu, X.-p.; Xing, T.-y.; Ji, T.; Li, L. Defect-Related Visible Luminescence of ZnO Nanorods Annealed in Oxygen Ambient. Optoelectronics Letters

2012, 8, 48.

(52) Miriyala, N.; Prashanthi, K.; Thundat, T. Oxygen Vacancy Dominant Strong Visible Photoluminescence from BiFeO 3 Nanotubes. Phys. Status Solidi RRL 2013, 7, 668671.

28


Page 28 of 31

Page 29 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(53) Zhu, M.; Zhang, Z.; Miao, W. Intense Photoluminescence from Amorphous Tantalum Oxide Films. Appl. Phys. Lett. 2006, 89 . (54) Hai-Bo, F.; Shao-Yan, Y.; Pan-Feng, Z.; Hong-Yuan, W.; Xiang-Lin, L.; Chun-Mei, J.; Qin-Sheng, Z.; Yong-Hai, C.; Zhan-Guo, W. Investigation of Oxygen Vacancy and Interstitial Oxygen Defects in ZnO Films by Photoluminescence and X-ray Photoelectron Spectroscopy. Chin. Phys. Lett. 2007, 24, 2108. (55) Zhu, M.; Zhang, Z.; Miao, W. Intense Photoluminescence from Amorphous Tantalum Oxide Films. Appl. Phys. Lett. 2006, 89 . (56) E.H.Nicolian,; J.R.Brews, MOS (Metal Oxide Semiconductor) Physics and Technology ; Wiley, New York, 1982. (57) Dimitrakis, P.; Normand, P.; Bonafos, C.; Papadomanolaki, E.; Iliopoulos, E. GaN Quantum-Dots Integrated in the Gate Dielectric of Metal-Oxide-Semiconductor Structures for Charge-Storage Applications. Appl. Phys. Lett. 2013, 102 . (58) Vasilopoulou, M.; Douvas, A. M.; Palilis, L.; Kennou, S.; Argitis, P. Old Metal Oxide Clusters in New Applications: Spontaneous Reduction of Keggin and Dawson Polyoxometalate Layers by a Metallic Electrode for Improving Eciency in Organic Optoelectronics. J. Am. Chem. Soc. 2015, 68446856. (59) Palik, E. D. Handbook of Optical Constants of Solids ; Academic press, 1998; Vol. 3. (60) Nalwa, H. S. Handbook of Surfaces and Interfaces of Materials: Surface and Interface Phenomena ;

Academic Press, 2001; Vol. 1.

(61) Keggin, J. The Structure and Formula of 12-Phosphotungstic Acid. Proceedings of the Royal Society of London A: Mathematical, Physical and Engineering Sciences. 1934; pp 75100.

29



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(62) Levy, O.; Stroud, D. Maxwell Garnett Theory for Mixtures of Anisotropic Inclusions: Application to Conducting Polymers. Phys. Rev. B 1997, 56, 8035. (63) Chaneliere, C.; Four, S.; Autran, J.; Devine, R. Dielectric Permittivity of Amorphous and Hexagonal Electron Cyclotron Resonance Plasma Deposited Ta 2 O5 Thin Films. Electrochem. Solid-State Lett.

1999, 2, 291293.

(64) Chaneliere, C.; Four, S.; Autran, J.; Devine, R.; Sandler, N. Properties of Amorphous and Crystalline Ta 2 O5 Thin Films Deposited on Si from a Ta(OC 2 H5 )5 Precursor. J. Appl. Phys.

1998, 83, 48234829.

(65) Pope, M.; Müller, A. Polyoxometalate Chemistry from Topology via Self-Assembly to Applications ;

Springer Science & Business Media, 2001.

(66) Netzer, F. P.; Allegretti, F.; Surnev, S. Low-Dimensional Oxide Nanostructures on Metals: Hybrid Systems with Novel Properties. J. Vac. Sci. Technol., B 2010, 28, 116.

30


Page 30 of 31

Page 31 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Graphical TOC Entry

31


Tantalum Oxide

Recommend Documents