Novel Materials for Organic and Thin Film Novel Materials for Organic and Thin Film ElectronicsElectronics

A

B1 µµµµm

Ruud M. TrompIBM Research Division

T.J. Watson Research CenterPO Box 218

Yorktown Heights, NY 10598rtromp@us.ibm.com

The road beyond CMOS : The road beyond CMOS : NanoNano ??Nano: things smaller than 100 nm

Opportunities for revolutionary new materials,processes, and technologies

But: also many opportunities to improve, extend, andtransform present technologies, including computerhardware technologies

����

A

B1 µµµµm

������� ���

Silicon Logic Silicon Logic -- Already at the Already at the NanoscaleNanoscale

• Gate length = 6 nm• Power4 Chip

• 174 million transistors

B. Doris et al., IEDM , paper 10.6, 2002. J. Warnock et al., IBM J. R&D, p. 27, 2002

TSi=7nm Lgate=6nm

Source Drain

Gate

(That’s what got us into trouble!)

One alternative: Carbon One alternative: Carbon NanotubeNanotube FETFET

Semiconductor mobilities @ RT (cm2/V.s)InSb 77,000CdSe 650c-Si 1,500 – 100a-Si 1CNT 100,000Polymer 10-5 – 10-2

Pentacene 0.1 – 5Chalcogenides (spinc.) 1 -20

Carbon Carbon NanotubeNanotube InverterInverter

-2

-1

0

1

2

-4 -2 0 2 4

VIN (V)

VO

UT

(V)

Gain>1

Nanotube Nanotube TechnologyTechnology ??

CNTAu

100µm

How do you get from here to there?

����������������� ���������

� ����� �������������

Some of the issues• CNT synthesis and purification; control of diameter and chirality• CNT placement in integrated circuit hierarchy with nm precision• Site- and/or area-selective n- and p-type CNT doping on nm scale• Control over contacts to CNT, injection barrier to n-, p-, i-CNTs• Elimination of parasitics for high performance• Elimination of 1/f noise in CNFET devices• Theoretical modeling of CNT physics, chemistry, and devices• CNT and device physical characterization• Optoelectronic properties

1/f noise in carbon nanotubesPhilip G. Collins, M. S. Fuhrer, and A. ZettlAPL 76, 894 (2000)

3

2

1

0

Nor

mai

lized

Cou

nts

1.61.41.21.00.80.6

Diameter (nm)

Normalized diameter distribution Gaussian based curve fitting

Contacts

CVD CNT, Pd contacts, 8nm HfO2

CVD CNT, Pd contacts, L~2umCVD CNT, Pd contacts, L=50nm

Arc discharge CNT, Pd contactsArc discharge CNT, Ti contacts

0.8

0.6

0.4

0.2

0.0

Nor

malized

Cou

nts

10-4

10-3

10-2

10-1

100

101

Ion

(����A)

Pd Contacts Ti Contacts Al Contacts

Vds = -0.5V

Zhihong Chen, Joerg Appenzeller, Joachim Knoch,Yu-Ming Lin, Phaedon AvourisA26-5

CNT Optoelectronics

Vds is fixed, Vg is scanned50 µm

SEM

0.5 0.6 0.7 0.80

1

Inte

nsity

(a.u

.)

Energy [eV]

Emission spectrum

DRAIN

SOURCE

e-

h+

e-

h+h+

e- e-

0.0 0.2 0.4 0.6 0.8 1.01.0

0.8

0.6

0.4

0.2

0.0

12 13 14

0

10

20

30

40

50

gate voltage �g

spot

( �

m )

spot

( L

)

1.0

0.8

0.6

0.4

0.2

0.0

CNT optoelectronic properties are key in understanding basic transport properties, effects of defects, CNT electronic structure, excited states, etcetera.There is room for much experimental work, as well as more advanced theoretical understanding.

Exciton binding energy is CNTs (0.1-0.5 eV) much larger than in III-V’s (10’s of meV). CNT exciton bonding energy depends on environment.

Ph. Avouris, A26-4

Nanotube Nanotube TechnologyTechnology ??

CNTAu

100µm

How do you get from here to there?

����������������� ���������

� ����� �������������

Not ‘just engineering’

Lots of basic science !

A.J. Heinrich, C.P. Lutz, J.A. Gupta, D.M. EiglerScience 2002

A Molecular Computer A Molecular Computer –– Slow but it works (once)Slow but it works (once)

Molecular TransistorsMolecular TransistorsElectrostatics of Molecular TransistorsElectrostatics of Molecular Transistors

C. R. Kagan, A. Afzali, R. Martel, L. M. Gignac, P. M. Solomon, A. G. Schrott, B. Ek, Nano Letters, 3, 119 (2003).

gate

source drain

gate dielectric

L

tdielectric

Au Au

molecule Barrier lowered byEnergy level offset

Hybridizationand Charge transfer

between metal-molecule

Design molecules with L>2.5-3 nmTailor tunnel barriers through chemistry

OFF state: Limited by Tunneling1

Gate Modulation: Electrostatics2

gate

source drain

gate dielectric

tdielectric ���� L

to attain gate field not dominated by

drain fieldYet

dielectric not leaky TSi=7nm Lgate=6nm

Source Drain

Gate

Is Si really that much bigger?

Directed Assembly of Molecular DevicesDirected Assembly of Molecular Devices

Wavelength (nm)

200 400 600 800

Abs

orba

nce

0.00

0.02

0.04

0.06

0.08

0.10Average bilayer coverage = 0.9x10-10 mol/cm2

(X-ray single crystal structure: 1.0x10-10 mol/cm2)

Wavelength (nm)200 300 400 500 600

Abs

orba

nce

0.00

0.01

0.02

0.03

0.04

Number of Bilayers0 2 4 6 8 10

Incr

ease

of A

bsor

banc

e

0.00

0.01

0.02

0.03

0.04

0.05

0.06

0.07

Rh 4 loo

p at

258

nm

Dipyridine linker at 298 nm

LayerLayer--byby--Layer AssemblyLayer Assembly

n

Rh

RhO

N

N

N

N

O

OO

Rh

RhO

N

N

N

N

OOO

N N

N N

N N

S S

Au

surfacetemplate1

inorganic2

organic3

Manipulate chemistry of template, organic, and

inorganic

n+ Si

SiO2

M1 M1

Device AssemblyDevice Assembly

Manipulate device geometry and molecular assembly

V (V)-10 0 10 20 30 40 50

I (A

)

0

1e-6

2e-6

3e-6

4e-6

5e-6

Rh

RhO

N

N

N

N

O

OO

Rh

RhO

N

N

N

N

OOO

NNNN

N

N

ZnNNNN

N

N

Zn

V (V)

-60 -40 -20 0 20 40 60

I (A

)

-2e-6

-1e-6

0

1e-6

2e-6

3e-6

4e-612345

V (V)0 20 40 60 80

I (A

)

-2.0e-7

0.02.0e-74.0e-7

6.0e-78.0e-71.0e-61.2e-6

1.4e-61.6e-61.8e-6

Rh

RhO

N

N

N

N

O

OO

Rh

RhO

N

N

N

N

OOO

N N

N N

V (V)-60 -40 -20 0 20 40 60

I (A

)

-3e-7

-2e-7

-1e-7

0

1e-7

2e-7

3e-7

4e-7

5e-71234

Device CharacteristicsDevice Characteristics

Study low and high bias characteristics

C. Lin, C. R. Kagan JACS, 125, 336 (2003)

PentacenePentacene: the world: the world’’s best organic semiconductors best organic semiconductor

GateSubstrate

Insulator

����������������������������������������

DrainSourceSemiconductor

Pentacene on Au contact

Pentacene onSiO2 Insulator

����� ������������������

����������������

1 ����m

V D (V )-4-3-2-10

I D( µµ µµ

A)

-5

-4

-3

-2

-1

0

V G =-4 V

-3 V

-2 V

-1 V0 V

Mobilities up to 5 cm2/Volt.secreported (a-Si: 1 cm2/Volt.sec)

Vacuum deposition

Key Question: How do Molecules interact Key Question: How do Molecules interact with Contacts?with Contacts?

������� ���

������ ��

������× ��������� �������

��������

������� ���

����

�������

�������

Pn stands up on non-metallicsurfaces, such as oxides, semiconductors or semimetals,but lies down flat on metals.

Fractal growth and scaling is consistent with classical growth modelsdeveloped for inorganic materials

J. Sadowski, G. Thayer, F. Meyer zu Heringdorf, R. Tromp

Growth on clean Si

Field of view 65 µmRoom temperature1 image/minute

Meyer zu Heringdorf, F.-J.; Reuter, M.C.; Tromp, R.M. Nature 2001, 412, 517-520

Deposition time (minutes)

Integral coverage (ML)0 0.5 1 1.5 2 2.5

Lay

er c

over

age

(%)

0

50

100

0 30 60 90 120

Aromatic and Pi-conjugated systems:

Benzene/Si(001) also undergoes transition from sp2 to sp3 hybridizationReversibly adsorbs, desorbs

Multiple binding configurations

Benzene liquid

Benzene/Si(001)

Wavenumber (cm-1)

2800 2900 3000 3100

K.P. Weidkamp, C.A. Hacker, M.P. Schwartz, X. Cao, R.M. Tromp and R.J. Hamers,

Abs

orba

nce

310030002900Frequency (cm

-1)

22002100

(a) Bulk pentacene

÷÷÷÷200

(b) Monolayer pentacene

2x10-4

(c) Multilayer pentacene

(d) Multilayer pentaceneafter heating

Pentacene on Si(001): Infrared Spectra

• Pentacene (bulk) shows no sp3

hybridized C-H stretching vibrations

• Monolayer coverage – peaks at 2091, 2870, 2922 cm-1 indicate Si-H and sp3

hybridized C-H bonds, also thermally stable

• Multilayer coverage – small Si-H peak, peaks above 3000 cm-1 are much larger than any below 3000 cm-1

• Heating multilayer – broader Si-H peak, peaks almost identical to those of monolayer(x4)

Interfacial layer involves transition of some C atoms to sp3 hybridization and some dissocation. Layer appears to be thermally stable and irreversibly bound.

K.P. Weidkamp, C.A. Hacker, M.P. Schwartz, X. Cao, R.M. Tromp and R.J. Hamers,

Cyclo-addition reaction on Si

Organic surface termination renders Si surface inert, providing anideal substrate for subsequent pentacene growth

cyclopentene cyclohexene cyclo-octadiene

SiSiSiSiSiSi

From cyclical molecules to chains

Chainlike molecules render Si surface inert, but affectdiffusion and aggregation during pentacene growth

hexene nonene dodecene

SiSi SiSiSiSi

Cyclooctadiene Hexene

Dodecene off axisDodecene on axis

Crystals at 90 degreesincrements

Crystals at 45 degreesincrements

Selection of azimuthal angles: Role of molecular species

LEED 8/20 dodecene Si(001) on axis

LEED 6/11 COD Si(001) on axis Epitaxial growth, rectangular unit cell

Pn

Pentacene:a = 0.627 nmb = 0.777 nmγ = 84.7o

Si(001):dimer-dimer: 0.384row-row: 0.768

Mismatch ~ 1%

Epitaxial growth of pentacene on Si:diffraction analysis

tilted illumination

on axis 4 degrees off axis

PENTACENE /Si(001) EPITAXY

0.768 nm

Pentacene MD on IBM BlueGeneclassicaldiffusion

non - classical diffusion pentacene diffusion on dodecene

Soluble Soluble pentacenepentacene precursorsprecursors

hνpolymeric precursor

SNO

O

anneal

Photosensitive version for patterningSoluble pentacene for spincoated TFTs

µµµµlin = 0.4µµµµsat = 0.8on/off > 106

Highest performancefrom solution processed organic TFT

ChalcogenidesChalcogenides: a new look at some old materials: a new look at some old materials Eg µ µ µ µn µ µ µ µp

(eV) (cm2/V-sec) (cm2/V-sec)

SnS2 2.6 18

SnSe2 1.6 27

ZnSe 2.7 600

ZnTe 2.25 100

CdS 2.42 250

CdSe 1.73 650

Organic-derivatized CdSe nanocrystals -- n-type 1.0 cm2/V-s[Ridley et. al., Science 286, 746 (1999)]

Chemical Bath Technique CdSe -- n-type 15 cm2/V-s[Gan et. al., IEEE Trans. Electr. Devices 49, 15 (2002)]

Assembled Nanorod / Nanoribbon -- n-, p-type <300 cm2/V-s[Duan et. al., Nature 425, 274 (2003)]

Semiconductor mobilities @ RT (cm2/V.s)InSb 77,000CdSe 650c-Si 1,500 – 100a-Si 1CNT 100,000Polymer 10-5 – 10-2

Pentacene 0.1 – 5Chalcogenides (spinc.) 1 -20

Solution Solution ProcessableProcessable ChalcogenidesChalcogenides

��� ������������� �� ���

�� �����������

���������� �

��� ������

ChalcogenidesChalcogenides –– a new lowa new low--T spinT spin--on on semiconductor with high mobilitysemiconductor with high mobility

A B

Sn

S SeS

Sn

10 15 20 25 30

170 175 180 185 190 195 200

10 15 20 25 30

170 175 180 185 190 195 200

2 (degree)θ

Inte

nsity

(arb

. uni

ts)

Energy (keV)

2 (degree)θ

Inte

nsity

(arb

. uni

ts)

Energy (keV)

1 mµ 1 mµ

RMS Roughness: 6 Å (A) and 14 Å (B)Film thickness: only 7- 9 unit cells thick!Composition: SnS1.8 and SnS1.4Se0.5

µµµµsat = 12.0 cm2/V-sµµµµlin = 2.4 cm2/V-sIon/Ioff > 106

L = 14 µµµµm; W = 250 µµµµmHighest spin-coated channel mobility by ~ 10X

Developed new chemistriesto spin on thin chalcogenidefilms with low processingtemperatures < 300 C

Nanoscience may give rise to future revolutionarytechnologies relevant to Information Technology:Nanotube or molecular logic; Novel, dense crosspointmemories; New high performance spin-coatable semi-conductors; etcetera.

Also, Si technology itself is a rapidly changing revolutionary technology that provides a ‘ready’ platformfor nanotechnology integration:Materials; Processes; Devices; Lithography; etcetera.

Development of new technologies to replace an existing technology takes decades, not months – but we can utilize nanotechnology on shorter timescales by insertion in existing technologies, and in niche application.