Overview

Electrical transport and charge detection in nanoscale phosphorus-in-silicon islands

Fay Hudson, Andrew Ferguson, Victor Chan, Changyi Yang, David Jamieson, Andrew Dzurak, Bob Clark

24th January 2006

Overview

• Introduction to previous work by PhD student, Victor Chan

• Large buried P islands with buried leads

• Motivation for this work

• Smaller islands, approaching few hundred atoms per island

• Fabrication – e-beam lithography, phosphorus implantation through

PMMA mask, activation of phosphorus, surface gates and SETs

• Single island measurements – direct transport

• Double island measurements – with SET charge detection

• Future work and conclusions

Previous work on buried islands with leads

• Victors device comprises:

- Diffused ohmic contacts

- Implanted phosphorus leads

- Implanted phosphorus island

- Barrier control gates (B1 and B2)

- Control gates (C1 and C2)

• Phosphorus implanted at 14keV to a depth of ~ 20 nm below the Si surface

• Island contains a few tens of 1,000’s of phosphorus atoms, 500 nm x 100 nm

• Gap between leads and island ~ 80-100 nm

V.C. Chan et. al., condmat/0510373

Experiments by Victor Chan

• To learn about the phosphorus-in-silicon system

• Also, to further confirm that fabrication works – P ion implantation through a mask, P ion activation

Previous work on buried islands with leads

Experiments by Victor Chan

• Coulomb blockade was observed (data left)

• Tunnel barriers could be independently controlled via surface gates, B1 and B2

V.C. Chan et. al., condmat/0510373

Gsd

(10-2

e2/h

)

Gsd

(10

-2 e2/h

)

Vsd = 0 V

Vsd = 350 V

Scaling-down: nanoscale island with leads (no SET)

Smaller devices: ~ few hundred P atoms (c.f. Victor’s ~ 104)

Island implant aperture ~ 30 nm diameter

Areal implant doses ~ 8.5 × 1013 cm-2 (~ 10 x bulk MIT for P in Si) 600 P atoms per island

~ 4.2 × 1013 cm-2 (~ 5 x bulk MIT for P in Si) 300 P atoms per island

• Diffused ohmic contacts • Buried leads• Buried island

• Barrier gate (Vb)

(too small for separate barrier gates)

• Control gate (Vg)

d = 20, 40, 60, 80 nm (Victor’s devices ~ 80nm)

30 nm island65 nm gaps

Test implant mask

50 nm island35-45 nm gaps

P implanted through mask

Fabrication (no SETs)

1 m20 m

Bond pads(Optical lithography)

Diffused ohmicContact(Optical)

Thin oxideregion

(Optical)Ion implanted

source-drain leadsand island

(EBL)Vb – barrier

gate

Vg – controlgate

(EBL)

Ti/Pt Alignment marks

(EBL)

Ion implanted source-drain and island (pre-anneal)

50 nm island35-45 nm gaps

50 nm island50-60 nm gaps

50 nm island65-75 nm gaps

no island120 nm gap

• SEM images of device: post-implant, pre-anneal

• Dark areas show location of implanted 31P+ and damage in Si

• Can check fabrication and record individual dimensions

50 nm island17-20 nm gaps

Looking for gate dependent eventsCoulomb blockade seenin bias spectroscopy

DC measurements (50 nm island, 35 nm gaps)

1 m

Vb

Vg

S D

Device turns on ~ 1.5V

(looks like a MOSFET)

DC measurements (50 nm island, 35 nm gaps)

3 mV

3 mV

Vsd (mV)

• Coulomb blockade peaks have period ~ 3mV

• Also found at a higher barrier voltage range

• Indicates a structure with constant capacitance – charging to a regular potential

3 mV

Vg = 175 mV

F.E. Hudson et. al., to appear in Microelectronic Engineering 2006

Devices with different gaps

• 20 nm device – shows weaker gate dependent events with similar period (6 mV)

• All devices with ~20 nm gaps are conducting at zero barrier-gate voltages

• All devices with >40 nm gaps are not conducting at zero barrier-gate voltage

• Small change in gap size large changes in device characteristics

• Barrier control is important to adjust for fabrication variations

6 mV

Previous work on double island devices with SETs

Previous work by Victor Chan: Large (500nm) double-dot structure with two rf-SETs:

Double-dot hexagon cell structure was observed

V.C. Chan et. al., paper in preparation

van der Wiel et. al.,Rev. Mod. Phys. 75 1 (2003)

Previous work on double island devices

Demonstrated control over interdot coupling

• Voltage on middle gate, VM, is increased

• Interdot coupling is increased – seen is a separation of the triple points

• Eventually, the two dots merge and exhibit single island characteristics (parallel lines)

Previous work by Victor Chan:

V.C. Chan et. al., paper in preparation

VM = 0.81 V

VM = 1.0 V

Double island devices – fabrication

Implant test mask:50 nm islands 60 nm gaps

Implant test mask:50 nm islands80 nm gaps

60 nm

50 nm

80 nm

50 nm

•Make smaller islands, with few hundred electrons in each

•Two surface control gates and use just one SET (no room for two)

Si:P double island – device 1 (50 nm, 80 nm gaps) with SET

500 nmVL VR

SET

SETgate

n, m n+1, m

n+1, m+1

n, m+1

SET aligned centrally between dots

Device 1 – 600 ions per island

n m

Si:P double island – device 2 (50 nm, 60 nm gaps)

SET aligned slightly closer to right dot (m)

n, m

n+1, m

n+1, m+1

n, m+1

Device 2 – 300 ions per island

• Coupling between dots is much smaller than coupling of dots to gates (by 50-100x)

• Need control over tunnel coupling between dots

• Difficult to fit tunnel barrier gates, plus SET, plus control gates…

• Would be interesting to see the effect of tunnel coupling control - could float an rf-SET relative to buried structure to control tunnel barriers

Si:P double island – device 1 (50 nm, 80 nm gaps) with SET

600 ions per island

Future work

rf in

dc offset

0.4 pF 100 pFSET

SET control gate

• Add capacitor to one side of the SET – rf sees ground but dc floating

• rf-SET is still operational and also a DC bias can be applied which acts as the tunnel barrier control gate

• Increase the SET antenna size to cover barriers• Lose ability to bias the SET, but can operate normal (B ~ 0.5T) - no need for bias

2) Is it possible to make these devices smaller – towards observing quantum states? – with current fabrication, N ~ 50

1) Combined SET and interdot coupling control

100 pF