Essays & Theses

Bioorganic nanodots for non-volatile memory devices

Bioorganic nanodots for non-volatile memory devices
of 7
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Bioorganic nanodots for non-volatile memory devices Nadav Amdursky, Gil Shalev, Amir Handelman, Simon Litsyn, Amir Natan, Yakov Roizin, Yossi Rosenwaks, Daniel Szwarcman, and Gil Rosenman   Citation: APL Materials 1 , 062104 (2013); doi: 10.1063/1.4838815   View online:   View Table of Contents:   Published by the AIP Publishing   This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: Downloaded toIP: On: Wed, 25 Dec 2013 16:56:26  APL MATERIALS  1 , 062104 (2013) Bioorganic nanodots for non-volatile memory devices Nadav Amdursky, 1 Gil Shalev, 1 Amir Handelman, 1 Simon Litsyn, 1,2 Amir Natan, 1 Yakov Roizin, 1,3 Yossi Rosenwaks, 1 Daniel Szwarcman, 1,2 and Gil Rosenman 1,2,a 1 School of Electrical Engineering, Iby and Aladar Fleischman Faculty of Engineering,Tel Aviv University, Tel Aviv 69978, Israel 2 StoreDot LTD, 16 Menahem Begin St., Ramat Gan, Israel 3 TowerJazz, P.O. Box 619, Migdal HaEmek 23105, Israel (Received 2 August 2013; accepted 20 November 2013; published online 10 December 2013) Inrecentyearswearewitnessinganintensiveintegrationofbio-organicnanomaterialsin electronic devices. Here we show that the diphenylalanine bio-molecule can self-assemble into tiny peptide nanodots (PNDs) of   ∼ 2 nm size, and can be embeddedinto metal-oxide-semiconductor devices as charge storage nanounits in non-volatilememory. For that purpose, we first directly observe the crystallinity of a single PNDbyelectronmicroscopy.WeusethesenanocrystallinePNDsunitsfortheformationof adense monolayer onSiO 2  surface,and studytheelectron/hole trapping mechanismsand charge retention ability of the monolayer, followed by fabrication of PND-basedmemory cell device.  © 2013 Author(s). All article content, except where otherwisenoted, is licensed under a Creative Commons Attribution 3.0 Unported License. []Nature’s diverse building blocks that self-assemble from DNA, proteins, and peptides becamean object of intensive nanotechnological research, due to their integration feasibility in electronicdevices. In this letter we focus on the integration of nm-scaled peptide structures, which were self-assembled from the short diphenylalanine (FF) peptide, into bio-electronic devices. FF is a shortaromatic peptide, which was srcinally suggested as the core recognition motif of the alzheimer-related amyloid- β  protein. 1,2 FF can self-assemble into nanotubes similar to amyloid fibrils, wherethe supramolecular architecture is being stabilized by noncovalent interactions. 1 In the last decade,severaluniquephysicalpropertieshavebeenassignedtotheFFtubularstructure,andseveralpossibleapplications have been purposed for them. A partial list includes: using the FF tubes as templatesfor metal nanowires or coaxial wires, 1,3 as templates for the incorporation of photo-sensitizers, 4 aschannels for nano-fluidic devices, 5 as efficient carbon electrode coating of supercapacitors, 6,7 theyexhibit one of the strongest measured piezoelectric signal among biological structures, 8 demonstratepronounced non-linear optical response, 9 and possess unique photoluminescence properties. 10,11 Though the tubular structure of the FF tubes has an advantage for the described above applications,its main disadvantage is in its relatively large size, reaching the µ m scale. This disadvantage limitsthe integration of the FF tubes in nm-scaled bio-electronics devices (as in here). Recently wehave found that in anhydrous conditions the tubes can disassemble into stable building blocks of peptide nanodots (PNDs). 12 Thus, the nm-scaled ( ∼ 2 nm in diameter) PND architecture opens newopportunities for the integration of FF-based structures in bio-electronics devices.We report in this letter on the successful integration of the PNDs as the charge-storage elementsin non-volatile memory (NVM) devices. Currently, charge trapping NVM is typically composed of ONO(oxide-nitride-oxide)dielectric, 13 wherethechargesarestoredindeeptrapsinSi 3 N 4  depositedon a bottom oxide and capped with a top oxide. In order to reduce the program-erase voltages andto enhance the charge retention, there is a need to replace the commonly used ONO layer with a a Author to whom correspondence should be addressed. Electronic mail: 2166-532X/2013/1(6)/062104/6 ©Author(s) 2013 1 , 062104-1  This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: Downloaded toIP: On: Wed, 25 Dec 2013 16:56:26  062104-2 Amdursky  et al.  APL Mater.  1 , 062104 (2013) FIG. 1. Electron microscopy analysis of the FF PND. (a) TEM images of single FF PNDs. The inset of (a) shows the Fouriertransform of a single dot. (b) ED ring pattern of the FF PND. (c) Size-distribution histogram of several TEM acquisitions. memorymediathatcontainsdiscretechargestoragecrystallinenano-dots(NDs). 14–21 TheNDslayershould have high NDs surface density, nm-scaled structure, size and shape uniformity, and lateralisolation (i.e., electrically isolated NDs). As of now, Au or Pt nanocrystals are the most popular NDsfor new-generation NVM. Recently, Paydavosi  et al. 22 have shown that also organic dye molecules(perylene derivatives) can serve as the discrete ND element in NVM, however the charge retentionperformance was rather poor. The possible integration of biological molecules in this type of NVMwasdiscussedinthecontextoftheiruseas merely atemplatefortheformationofinorganicNDs. 23–25 The nano-crystallinity of the FF PNDs, together with their uniform nm-scaled size distribution andlow temperature deposition, makes them promising candidates for the integration in NVM. Thistype of NVM is fundamentally different from the latter examples of using biological molecules as atemplate for inorganic ND, since we show here for the first time that the nano-crystalline unit is thebio-organic material itself.In order to utilize the FF PND as a new nanotechnological material (in general or specificallyfor NVM as we show here), one should first be able to both isolate and explore the propertiesof a single PND, as well as to form a dense monolayer of the FF PNDs. Thus, we first explorethe properties of individual PNDs by electron microscopy, while confirming their nano-crystallinitynaturebyfollowingtheirelectrondiffraction(ED)pattern.Wefurthershowthesuccessfulformationof a PNDs dense monolayer, and the ability of the monolayer to retain charge by using Kelvin ProbeForce Microscopy (KPFM). In the last stage, we demonstrate how the FF PNDs can be successfullyembedded in a NVM device as the charge storage elements.Transmittance- and scanning-electron microscopy (TEM and SEM, respectively) of isolatedPNDs, which were self-assembled in anhydrous methanol solution are shown in Figure 1(a) and inFigure S1 of the supplementary material. 40 The images of the FF PNDs were obtained without anycoating by metallic layer or using a contrast agent, although a fast acquisition time and reducedbeam intensity were used in order not to damage the PND (see further in the materials and methodssection in the supplementary material 40 ). The energy-dispersive X-ray spectroscopy (EDX) analysis(shown in Fig. S2 of the supplementary material 40 ) confirms that the observed structures are theFF PNDs, and not metallic artifacts. The Fourier transform image (inset of Fig. 1(a)), displayingdiscrete diffraction spots, indicates the single crystalline structure of the PND. Further, ED patternsobtained from numerous PNDs, such as those depicted in Fig. 1(a), validate the crystalline structureof the PND. The ED image (Fig. 1(b)) unveils diffraction ring patterns from (130), (101), (500),(221), (251), and (212)  hkl  planes, which correspond to 5.8 Å, 4.2 Å, 4.0 Å, 2.9 Å, and 2.6 Åd-spacings, respectively. These d-spacings are consistent with the crystalline structure of the FF tubeformation 26 and coincide with XRD of the PND (this latter XRD were obtained by powder-XRD of PNDs aggregates and not from single PND as in here). 12 The lateral size distribution of the FF PNDthat was estimated from the TEM images is 2.4 ± 0.6 nm (Fig. 1(c)). The size distribution is similarto what we previously found for the FF PNDs (of 2.1 nm). 12 Though earlier mass- and secondary ion mass-spectrometries suggested that the PND is com-posed from a dimer of FF, 12 there is no purposed structure for the FF PND yet. An upper bound  This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: Downloaded toIP: On: Wed, 25 Dec 2013 16:56:26  062104-3 Amdursky  et al.  APL Mater.  1 , 062104 (2013) FIG. 2. AFM morphology of the FF PND. (a) Monolayer of FF PND on Si/SiO 2  /APTMS surface. The bar corresponds to500 nm, and the z-scale is 4 nm. (b) Size-distribution histogram of (c). (c) Raman spectrum of (a) using a Raman NSOMapparatus. estimate for the number of molecules in a PND can be obtained by taking the crystalline moleculardensity as a limit. In the hexagonal cell, proposed by G¨orbitz, 26 there are 6 molecules in a volumeof  ∼ 2730 Å 3 , suggesting that at most 9 FF molecules can be packed in a sphere of 2 nm diameter.The crystalline structure attains an optimal arrangement, in which the non-polar rings interact withone another via aromatic stacking and van der Waals (VdW) forces, while the polar backbone sidesinteract with the solvent molecules and with adjacent FF molecules via hydrogen bonds. Althoughthe same forces should affect the PND structure, we expect that inside a PND the packing is lessoptimal and hence the number of FF molecules can be significantly lower. The influence of solventontheforcesbetweentheFFmoleculesintheprocessofself-assemblywasrecentlydemonstratedbyRissanou  et al. 27 Since both, simulations and experiments, indicate susceptibility for self-assembly,one can estimate the number of FF molecules inside a PND to be between 2 (a dimer) and 10.However, molecular dynamic simulations performed by Shell and co-workers 28 showed that FFmolecules in a polar solvent can form dimers to hexamers, where the dimer configuration is the mostfrequent one having ∼ 70% probability, which is consistent with experimental data. 12 A mandatory step in the integration of PNDs (or any other crystalline NDs for that matter)in a NVM is to form a uniform monolayer of the NDs on an oxide surface (the bottom oxide of the memory stack). Since all the deposition techniques, which involve merely the evaporation of the PNDs containing organic solvent (such as drop-casting, spin coating, or langmuir-blodgett),induce the formation of aggregates or isolated PNDs (depends on the initial concentration of thePNDs in solution) and not the formation of a monolayer, we had to use a linker to form the PNDsdense monolayer. The chosen linker molecule was the short amine-terminated propyl-silane linker(3-aminopropyl trimethoxysilane, APTMS) in order to electrostatically bind the FF PND to the SiO 2 surface. In this way, a monolayer has been formed on the oxide surface (Fig. 2(a)) with an estimatedparticle density of 6 × 10 11 PND/cm 2 . The narrow size (height) distribution of the structures withinthe monolayer (2.2 ± 0.4 nm, Fig. 2(b)) is highly similar to the histogram that was obtained from theTEM measurements (2.4 ± 0.6 nm, Fig. 1(c)), and also to the one that we previously observed (2.1 ± 0.2 nm), 12 thus suggesting that the structures of the monolayer are indeed the FF PNDs. Ramannear-field scanning optical microscopy (NSOM) confirmed that the deposited elements are the FFPNDs by showing that the FF-PND/APTMS surface exhibits the typical Raman peaks of the FFaromatic moieties (indicated by arrows in Fig. 2(c)): 29,30 1003, 1032, 1587, and 1605 cm − 1 . Boththe later indexed peaks and the un-indexed ones ( ∼ 800 and 1200 cm − 1 ) correlate with the knownRaman spectrum of the FF tubes. 31 An additional obligatory step in the verification of the FF PND suitability for NVM was toassure the presence of deep trapping energy levels. 32 We used KPFM for this purpose, which is anAFM-based technology enabling spatially resolved work function measurement with a nanometerresolution. KPFM measurements utilize a conductive tip to measure the contact potential difference(CPD) defined as:  CPD  = − (  tip  −   sample )/  q , where   tip  and   sample  are the work functions of   This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: Downloaded toIP: On: Wed, 25 Dec 2013 16:56:26  062104-4 Amdursky  et al.  APL Mater.  1 , 062104 (2013) FIG. 3. Charge retention of FF PND. (a) AFM topography of the patterned FF PND surface, exhibiting an array of PND andSiO 2 . CPD images of (b) the array immediately after charge injection and (c) after 3 h. (d) and (e) CPD cross sections alongthe lines in (b) and (c). the tip and the sample, respectively, and  q  is the elementary charge. 33 One instrumental featureof KPFM measurements is the CPD nullification, allowing the exclusion of any tip-induced bandbending. 34 The FF PND monolayer on an APTMS-processed silicon dioxide surface was patternedusing focused ion beam. The 300 nm pitch included a 200 nm line of PNDs and 100 nm (SiO 2 surface) space (Fig. 3(a)). Figures 3(b) and 3(c) present the CPD images of the trapped charge immediately after the charge injection and after 3 h of retention, respectively. During the injectionthe tip was grounded and the substrate was biased with –1 V. Figures 3(d) and 3(e) are line plots along the lines indicated in Figures 3(b) and 3(c), that reflect the presence of charges at the interface space (A and A ′ ) and the PNDs (B and B ′ ). CPD decrease ( ∼ 120 mV) after injection confirmshole-injection from the AFM tip and trapping in the PNDs. The decrease of the CPD in the spacesfollowing ∼ 3hofstorageisprobablyduetoelectrostaticscreeningofthenegativechargeadjacenttothe positively charged PNDs. This is consistent with migration of water ions (both H 3 O + and OH − )along the interface with the PNDs. 35–38 The presence of deep traps in the PNDs can be indicated bythe results, otherwise, no charge trapping effects and fast charge decay would have been observed.Following the verification of charge retention within FF PNDs, we used the PNDs as the chargestorage elements in a NVM device. Memory MOS capacitors (see the scheme in Fig. 4) were formedonp-typeSisubstrateswith50Åthermalbottomoxide.TheformationoftheAPTMS/PNDslayeronthebottomoxidewasfollowedbydepositionof30nmaluminaat200 ◦ Cbyatomiclayerdeposition.No special post-deposition anneals were employed. Sputtered Al or InGa electrodes were usedas metal electrodes of MOS capacitors. C-V (Capacitance-Voltage) characteristics of the formedMOS capacitors were measured by applying voltage sweeps ( +  /  − 6 V). The C-V characteristics(Fig. 4) show that stable trapping of both positive and negative charges takes place. It is necessaryto emphasize once again that both the PNDs and the top alumina dielectric layers were depositedat low temperatures. We cannot exclude some residual trapping in the alumina and Si-alumina  This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: Downloaded toIP: On: Wed, 25 Dec 2013 16:56:26
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks