New version page

Overview of Phase-Change Chalcogenide Nonvolatile Memory Technology

This preview shows page 1 out of 4 pages.

View Full Document
View Full Document

End of preview. Want to read all 4 pages?

Upload your study docs or become a GradeBuddy member to access this document.

View Full Document
Unformatted text preview:

IntroductionThe use of phase-change chalcogenidealloy films to store data electrically andoptically was first reported in 19681and in1972,2respectively. Early phase-changememory devices used tellurium-rich, multi-component chalcogenide alloys with a typi-cal composition of Te81Ge15Sb2S2. Both theoptical and electrical memory devices wereprogrammed by application of an energypulse of appropriate magnitude and dura-tion. A short pulse of energy was used tomelt the material, which was then allowedto cool quickly enough to “freeze in” theglassy, structurally disordered state. To reverse the process, a somewhat lower-amplitude, longer-duration pulse wasused to heat a previously vitrified region ofthe alloy to a temperature below the meltingpoint, at which crystallization could occurrapidly, as shown in Figure 1. Differences inelectrical resistivity and the optical constantsbetween the amorphous and polycrys-talline phases were used to store data.During the 1970s and 1980s, significantresearch efforts by many industrial and aca-demic groups were focused on under-standing the fundamental properties ofchalcogenide alloy amorphous semicon-ductors.3,4Prototype optical memory disksand electronic memory device arrays alsowere announced,5,6beginning in the early1970s. Rapidly crystallizing chalcogenidealloys were later reported7–10 by severaloptical memory research groups. These newmaterial compositions, derived from theGe-Te-Sb ternary system, did not phase-segregate upon crystallization like the ear-lier Te-rich alloys, but instead exhibitedcongruent crystallization11 with no large-scale atomic motion.In the 1990s, researchers at Energy Con-version Devices Inc. and Ovonyx Inc. developed new, thermally optimized phase-change memory device structures that ex-ploited rapidly crystallizing chalcogenidealloy materials to achieve increased pro-gramming speed and reduced program-ming current.12,13These devices14could beprogrammed in 20 ns—about six orders ofmagnitude faster than the early phase-change memory cells, and their much lowerprogramming current requirements per-mitted the design of memory arrays usingmemory bit access devices (transistors ordiodes) fabricated at minimum litho-graphic dimensions. Ovonyx is now com-mercializing its phase-change memorytechnology, called Ovonic Unified Memory(OUM),15through a number of license agree-ments and joint development programswith semiconductor device manufacturers.Storage MechanismThe Amorphous–CrystallineStructural Phase ChangeGlassy materials are produced by rapidlysupercooling a liquid below its meltingpoint to a temperature at which the atomicmotion necessary for crystallization cannotreadily occur. Chalcogenide alloys—materials containing one or more elementsfrom Group VI of the periodic table—aretypically good glass-formers, in large partbecause the Group VI elements form pre-dominantly twofold-coordinated covalentchemical bonds that can produce linear,tangled, polymerlike clusters in the melt.This increases the viscosity of the liquid,inhibiting the atomic motion necessary forcrystallization. Many amorphous chalco-genide alloys have been reported in the literature.3 The Ge2Sb2Te5(GST 225) chalco-genide alloy currently used in OUM mem-ory devices melts at approximately 610Cand has a glass-transition temperature of350C. In order to crystallize an amor-phous region of GST 225, the material mustbe heated to a temperature somewhatbelow the melting point and held at thisMRS BULLETIN/NOVEMBER 2004 829Overview of Phase-ChangeChalcogenideNonvolatile MemoryTechnologyS. Hudgens and B. JohnsonAbstractPhase-change nonvolatile semiconductor memory technology is based on anelectrically initiated, reversible rapid amorphous-to-crystalline phase-change process inmulticomponent chalcogenide alloy materials similar to those used in rewriteable opticaldisks. Long cycle life, low programming energy, and excellent scaling characteristics areadvantages that make phase-change semiconductor memory a promising candidate toreplace flash memory in future applications. Phase-change technology is beingcommercialized by a number of semiconductor manufacturers. Fundamental processesin phase-change semiconductor memory devices, device performance characteristics,and progress toward commercialization of the technology are reviewed.Keywords: chalcogenides, nonvolatile memory, phase change.Figure 1. Schematic temperature–timerelationship during programming in aphase-change rewriteable memorydevice. Taand Txare the amorphizationand crystallization temperatures,respectively. The SET and RESETstates of the memory correspond to astored binary 1 or binary 0.www.mrs.org/publications/bulletintemperature for a time sufficient to allowthe crystallization to occur. The composi-tional dependence of crystallization kinet-ics in the GeSbTe ternary system has beenextensively studied and reported in the lit-erature.16,17OUM cells based on GST 225that can be programmed (crystallized) to the“SET” state in 20 ns have been reported.18Electronic Properties of Crystallineand Amorphous GST AlloysTwo special electronic properties ofchalcogenide amorphous semiconductor al-loys are required for the operation of OUMmemory—the strong dependence of elec-trical resistivity on the structural state ofthe material and the high-field thresholdswitching phenomenon. PolycrystallineGST 225 alloy has a resistivity of 25 m cm,while resistivity in the vitreous state is threeorders of magnitude higher—sufficient toenable good memory read capability.Both structural states of the alloy aresemiconductors with comparable energybandgaps. The bandgap Egis 0.7 eV in theamorphous state and 0.5 eV in the poly-crystalline state. The conductivity activationenergy Eais 0.3 eV for the amorphousstate and 0.02 eV for the polycrystallinestate. In addition, the amorphous phaseexhibits a very low, trap-limited hole mo-bility of 2  105cm2/V s, while the poly-crystalline phase shows band-type mobilityof 10 cm2/V s. These large differencescome about because of disorder-inducedlocalized electronic states as originally de-scribed by Mott19and by Cohen, Fritzsche,and Ovshinsky20(CFO) and later by Kastner,Adler, and Fritzsche.21When chalcogenidealloy semiconductors are amorphized, elec-tronic energy levels originating in the va-lence and conduction bands are pulled intothe what was originally the empty energybandgap of the crystalline material. As de-scribed by Mott–CFO, these


Loading Unlocking...
Login

Join to view Overview of Phase-Change Chalcogenide Nonvolatile Memory Technology and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Overview of Phase-Change Chalcogenide Nonvolatile Memory Technology and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?