by dr. rudy schlaf semiconductor nanowires:...
TRANSCRIPT
Intro to Nanotechnology NanowiresBy Dr. Rudy Schlaf
Semiconductor Nanowires: Motivation
• Patterning into sub 50 nm range is difficult withoptical lithography. Self-organized growth ofnanowires enables 2D confinement of carrierswith large splitting of discrete energy states.
• Introduction of heterojunctions betweensemiconductors with different band gaps allows3D confinement in sub 50 nm range.
• Room temperature single electron transistors arepossible.
• Single photon devices are possible.
Intro to Nanotechnology NanowiresBy Dr. Rudy Schlaf
“Vapor-Liquid-Solid (VLS)” Growth of Si-Wire• Si substrate with Au
cluster is heated untilliquid Au-Si alloyforms
• Si-vapor/Si-precursorimpinges on surfaceand Si gets dissolvedin Au-Si liquidsupersaturating it
• Si precipitates atliquid/wafer interface
• Wire starts growingwhile lifting alloydrop up
• Wire assumes samecrystal structure assubstrate (epitaxialgrowth)
Questions to answer:(1) Why does the Si-precursor not deposit on areasurrounding the liquid alloy?(2) What is going on in the alloy droplet?(3) Why does the Si precipitate at all?
From
: A.P
.Lev
itt: W
hisk
er T
echn
olog
y, W
iley,
1970
Intro to Nanotechnology NanowiresBy Dr. Rudy Schlaf
(1) Growth Rate Vs. Supersaturation/Sticking Coefficient• Plot shows growth rate J [layers per
sec] versus the supersaturation σ.• J~ σ∗sticking coefficient∗T-1/2
• σ = (p-po)/po; p > p0 needed forgrowth
• p=vapor pressure of precursor abovesurface
• po=vapor pressure of sample surface• Liquid has much higher growth rate
below σ1, i.e. careful adjustment ofprecursor pressure allows toexclusively absorb precursor atomsin the liquid.
•Clean liquid has highest stickingcoefficient because it has an “infinitenumber of edges/defects”•Imperfect crystal has medium stickingcoefficient•Perfect crystal has lowest stickingcoefficient due to absence of edges, holesetc…where impinging atoms can bind
p
poFrom: A.P.Levitt: Whisker Technology, Wiley,1970
Intro to Nanotechnology NanowiresBy Dr. Rudy Schlaf
TL~850C
possible concentrations of Au-Si solution at TL
C1 C2
liquid and solid Si coexist
everything is eutectic solid
liquid and solid Au coexist
everything is liquid
•Au and Si do not exist in homogenous solution,only as eutectic solid (crystallites of pure Auand Si form a solid body)
(2,3) Au-Si Phase Diagram: Growth by Supersaturation• Au and Si form eutectic
mixture (i.e. the mixturecan have a lower meltingpoint than the purematerials.
• As the sample is heated,the Au particle on Si-wafer melts and due to“infinite” Si supply asolution of concentrationC2 develops
• As more Si is suppliedfrom gas phase, thesolution becomessupersaturated and excessSi precipitates out as solidSi particles whicheventually bind to the Sisurface under the Audroplet
Intro to Nanotechnology NanowiresBy Dr. Rudy Schlaf
Si-Whisker Growth
a) Piece of Au foil lying on Sisubstrate.
b) Au-Si alloy droplet at1050C.
c) After 10 min of growth fromSiCl4 precursor. Contactangle considerably increaseswith the consequence of areduction of the dropletdiameter.
d) After 30 min of growth:Contact angle has stabilizedand whisker has grown.
From: A.P.Levitt: Whisker Technology, Wiley,1970
Intro to Nanotechnology NanowiresBy Dr. Rudy Schlaf
Life is Not Easy: VLS Growth Artifacts
• Top: Temperature gradient onwafer: Alloy droplet originallyat A migrates towards highertemp. B.
• Left: Too much precursor vaporresults in breakup of originaldroplet, resulting in smallerwhiskers.
• Right: Sudden drop intemperature from 1000C to600C for 30 sec and subsequentre-heating to 950C resulted inrapid formation of small Sicrystals with randomorientation serving as seeds forsmall whiskers.
Intro to Nanotechnology NanowiresBy Dr. Rudy Schlaf
Size, Shape and Position Controlled GaAs Nanowires
From: B.J. Ohlson et al., Appl.Phys.Lett. 79(20),pp.3335,2001
GaAs(111)A-surface
(111)B-surface
• Size selected Au particles in N2stream are deposited onGaAs(111)B surface inside aglove box.
• AFM inside glove box is used topush the Au particles intoposition.
• Growth occurs at Au particlelocation using VLS mechanism.
• Precursors: triethylgallium (TEG)and tertiarybutylarsine (TBA).TBA is thermally precracked torelease As2 molecules.
Intro to Nanotechnology NanowiresBy Dr. Rudy Schlaf
Small Particle Mobility Analyzer• Selects particles of
certain mobility(mass) fromdistribution
• High voltage isapplied betweencenter rod and outerwall
• Particles (aerosol,red) enter main gas(air or N2) stream(blue)and getelectrostaticallyattracted to centerrod.
• Only particles ofcertain mass enterexit hole to“sampling outlet”
• A stream of mono-disperse particles isgenerated
From
: E.O
.Knu
dson
et a
l., J.
Aer
osol
Sci
.,6,p
p.44
3-45
1,19
75
Intro to Nanotechnology NanowiresBy Dr. Rudy Schlaf
GaAs Nanowires Growth Results• (top): Au particles moved into growth
positions(a), with grown GaAs wires, topview(b), side view(c).
• (bottom): Randomly distributed wires (b),TEM image of wire end (c); Differentcrystal structures and defects are visible(W=wurzite, Z=zinc blende, SF=stackingfault,T=twin plane). Cap at the end: Auparticle.
From: B.J. Ohlson et al., Appl.Phys.Lett. 79(20),pp.3335,2001
Intro to Nanotechnology NanowiresBy Dr. Rudy Schlaf
InAs/InP Heterostructure Nanowires
• Same VLS-procedure as for purewires using trimethyl In,tertiarybutylarsine, andtertiarybutylphosphine asprecursors.
• 40 nm width.• Alternating between arsine and
phosphine precursors producesalternating InP and InAs growthregions.
• TEM shows atomically abruptinterfaces.
From
: M.T
.Bjö
rk e
t al.
App
l.Phy
s.Let
t. 80
(6),1
058,
2002
Intro to Nanotechnology NanowiresBy Dr. Rudy Schlaf
InAs/InP Devices and Their Characterization
From: M.T.Björk et al. Appl.Phys.Lett. 80(6),1058,2002
• E-beam litho was used to makecontacts for electricalcharacterization.
• Wires were dispersed on oxidizedSi wafer (100 nm). Then 20 nmNi/70 nm Au electrodes werepatterned across the ends of thewire.
• Two kinds of wires werefabricated: (1) Pure InAsreferences and (2) InAs with one80 nm long InP section in themiddle.
• Graph shows current through pureInAs wire (at 4.2K) depending ona bias applied to the back side ofthe Si wafer(=gate). Linear curvesdemonstrate Ohmic behavior.
• Conductivity changes strongly dueto changing carrier density ofwire.
Intro to Nanotechnology NanowiresBy Dr. Rudy Schlaf
I-V Characteristic of InAs Wire with InP Barriera) I-V curves through pure InAs and
InAs/InP wire in comparison.Strong current increases are aresult from the barrier heightreduction due to the applied bias.
b) Current depends on temperature:I=const*T2*exp[-q*ΦB/kT]. Plotshows that ΦB varies dependingon the applied bias (voltages seec). Note In(I/T2) vs. 1/T plot; ΦBis extracted from slope.
c) Barrier height vs.bias voltage.Extrapolation to zero bias yields0.6eV barrier height, which isclose to the bulk heterojunctionvalue.
d) Band line-up of InAs/InPheterojunction; ~0.6 eVconduction band barriers.
pure
with InP
From: M.T.Björk et al. Appl.Phys.Lett. 80(6),1058,2002
Intro to Nanotechnology NanowiresBy Dr. Rudy Schlaf
Resonant Tunneling InAs/InP Device
From: M.T. Björk et al. Appli.Phys.Lett. 81(23), pp.4458-61,2002
a) 5 nm wide InP barriers inInAs. 15 nm wide InAsquantum dot in between.
b) Band diagram withelectronic states. InAs is n-type; Fermi energy liesabove E1, i.e. E1 is the onlyoccupied state in theemitter. On the dot thestates are split due to thedifferent z-direction andtransversal quantization.The smallest energy barrieris E1
z.c) I-V curve through wire.
Current is zero below 70mV applied bias. Sharppeak corresponds totunneling through E1
z.Insert shows repeatmeasurements.
Intro to Nanotechnology NanowiresBy Dr. Rudy Schlaf
Quantum Dot Devices Schematica) (left) Confinement of
electrons in potential wellresults in discrete states.(right) Electron can tunnelthrough potentiaal barriers.Their energy is unaffected,only the amplitude of thewave function changes.
b) Heterojunctions betweenmaterials having differentband gaps result inpotential barriers. This canbe used to prepare quantumdot devices.
c) When a SD bias is appliedelectrons tunnel throughsingle states on the dot asthey become available.Current spikes result.
From: L.Samuelson, Materials Today 6(10), pp.22-31,2003
Intro to Nanotechnology NanowiresBy Dr. Rudy Schlaf
Effects of 80 nm Long InAs Quantum Dot• Longer InAs section in wire
results in more closely spacedenergy states in the dot, i.e.many more peaks are beingseen in I/VG curve.
• (a) shows contacted wirebetween EBL contacts.
• (b) I/VSD curves across thewire. Depending on biasapplied to the wafer backside(gate) voltage VG different“wiggles” (representing energystates that are being tunneledthrough) appear.
• (c) I/VG curves at VSD=0.5mV. Note the pA scale.
From: C.Thelander et al. Appl.Phys.Lett. 83(10),pp.2052-55,2003
Intro to Nanotechnology NanowiresBy Dr. Rudy Schlaf
Effects of 80 nm Long InAs Quantum Dot (2)
• “Stability Plot” showsI/VSD curves for whole VGrange. Gray scalecorresponds to absolutecurrent (black is zero,white ±1 nA).
• The dotted linescorrespond to the twocurves in the graph.
• Top plot: data• Bottom plot: simulation
VG(V)
From
: C.T
hela
nder
et a
l. A
ppl.P
hys.L
ett.
83(1
0),p
p.20
52-5
5,20
03