M. L. W. Thewalt , A. Yang, M. Steger, T. Sekiguchi, K. Saeedi, Dept. of Physics, Simon Fraser University, Burnaby BC, Canada V5A 1S6

T. D. Ladd, E. L. Ginzton Laboratory, Stanford University, Stanford CA, USA (now HRL)

K. M. Itoh, Keio University, Yokohama, Japan

H. Riemann, N. V. Abrosimov, IKZ, Berlin, Germany

A. V. Gusev, A. D. Bulanov, ICHPS, Nizhny Novgorod, Russia

A. K. Kaliteevskii, O. N. Godisov, “Centrotech-ECP”, St Petersburg, Russia

P. Becker, PTB, Braunschweig, Germany

H. J. Pohl, VITCON Projectconsult GmbH, Jena, Germany

J. J. L. Morton, Oxford, UK

S. A. Lyon, Princeton, USA

Avogadro

Project

Highly Enriched 28Si – new frontiers in semiconductor spectroscopy

SILICON – 2010 Nizhny Novgorod

E

T0 R.T.

bare EG

actual EG

EG depends on M due to zero point

motion

ΔE

The renormalization goes as M -1/2

recent review:Cardona and ThewaltReviews of Modern Physics 77, 1173 (2005)

Natural Si: 92.2% 28Si + 4.7% 29Si + 3.1% 30Si I = 0 I = ½ I = 0

Why does the isotopic composition affect electronic and optical properties?

the electron-phonon interaction

Eg(0) increases by ~2 meV from 28Si to 30Si

~62 meV for natSi

9275.0 9280 9281

T=1.3 KB1NPa1

NP

28Si

natural Si

Pho

tolu

min

esce

nce

Inte

nsity

Photon Energy for natural Si (cm-1)

D. Karaiskaj et al., Phys. Rev. Lett. 86, 6010 (2001) [ FWHM < 0.005 cm-1 ]

P BE Boron BE

The spectroscopic challenge of highly enriched 28Si

( 8 cm-1 ~ 1 meV )

Boron BE

200 neV !

a

Improved shallow bound exciton linewidths – observation of the 31P ground state hyperfine splitting in the donor bound exciton transition.Phys. Rev. Lett. 97, 227401 (2006). Nuclear and electron spin readout…

31P prime Si qubit

candidate

can we use this to

polarize the spins?

Initialization problem

Almost zero equilibrium polarization

Hyperpolarizationusing optical pumping

Phys. Rev. Lett.102, 257401 (2009)

Resolved bound exciton hyperfine spectroscopy gives the populations of all four donor hyperfine levels in a single measurement

Nuclear polarization of 76% and electronic polarization of 90% are achieved simultaneously – in less than 1 second!

Higher polarization should be possible with reduced linewidths

New experiments: NMR on dilute 31P in 28Si using optical polarization and optical nuclear spin readout

CW NMR

55,846,700 55,846,750

FWHM = 24 Hz

RF power: -75dBm (40 µV)B = 844.77 G

PL

inte

nsity

RF frequency (Hz)

pump 8, probe 10

Why 845 Gauss?Schematic energy diagram

Magnetic field (nonlinear)

Max. 55.85 MHz

Min. 61.68 MHz

844.9G 34,043G

A = 117.53 MHz

Ener

gy (n

onlin

ear)

A

ge = 1.99850g31P = 2.26320e, 31P

Magnetic Field Dependence

55,846.65

55,846.70

55,846.75

61,677.15

61,677.20

61,677.25

Magnetic field (Gauss)

e: probe 4, pump 6RF: -75 dBm (40 µV)

f0=61,677.172 kHzB0=844.85 G

RF

frequ

ency

(kH

z)

e: probe 10, pump 8 RF: -75 dBm (40 µV)

840 845 850

f0=55,846.712 kHzB

0=844.89 G

transient NMR - Rabi Oscillations

0.0 0.5 1.0 1.5 2.0

PL

Inte

nsity

Time (ms)

pump 8, probe 10lasers always on

RF: CW 15 dBm (1.26V) at 55,846,714 Hz

First pulsed NMR -Ramsey Fringe Experiment

p/2

t

p/2

Lasers

on (polarizing)

measurement

off

RF pulseson

off

RF freq. = Resonance freq. +/– 1 kHz fringes of 1 kHz

signal

Ramsey FringesTemperature (Pressure) dependence

fRamsey

= 2,453.6 Hz, T2* = 21.0 ms

T=4.2K (P=1atm)

0 10 20 30

PL

inte

nsity

t (ms)

T=1.3K (P=1.0Torr)fRamsey

= 975.4 Hz, T2* = 16.4 ms

RF=55,847,712 HzPump 8(), Probe 10(), ; RF = 55,847,712 Hz;

probe

pump

RF

31P Donor Hyperfine Constant

me RF freq. (Hz)

Fringe freq. (Hz)

NMR freq. (Hz)

55,845,712.0 1,024.9

55,846,736.8(1)55,847,712.0 975.4

Σ = 2,000.3

61,676,172.0 1,027.2

61,677,199.1(1)61,678,172.0 972.9

Σ = 2,000.1

Hyperfine constant A = 117.523 935 9(2) MHz at T=1.3K

[existing value 117.53(2) MHz]

At 4.2 K and 1 atm, A is 3.116 kHz (26.5 ppm) lower! Why?

At T=1.3K (P=1.0Torr)

Pulsed NMR—Hahn echo method—

p/2

t

p/2

Lasers(Pump&Probe)

on

measurement

off

RF pulseson

off

fRF = fNMR

Signal decay with 2t measurement of T2

PL signal

rotation for optical readout

refocusing

(initialization)

t

p

0 500 1000

10

100

1000

ampl

itude

2 tau (ms)

Hahn echo result: 61,677,199 Hz, pump 6, probe 4, B = 845.3 G, T = 1.3 K (P = 1.0 Torr)

T2 = 250 ms

0 500 1000

10

100

1000

ampl

itude

2 tau (ms)

Hahn echo result: 61,677,199 Hz, pump 6, probe 4, B = 845.3 G, T = 1.3 K (P = 1.0 Torr)

T2 = 250 ms

T2 from Hahn Echo Method

probe

pump

RF

T2 = 250 ms

2t (ms)

PL

inte

nsity

Pump 6(), Probe 4(); RF = 61,677,199 Hz; B = 845.3 G, T = 1.3 K (P = 1.0 Torr)

Highly Enriched 28Si – new frontiers in semiconductor spectroscopy

Resolved donor bound exciton hyperfine structure:

- optical readout of 31P electron and nuclear spin - fast optical hyperpolarization of electron and nuclear spin - possibility of single donor readout - promising new approaches to silicon quantum computing - have begun NMR of 31P in 28Si combining optical readout and optical hyperpolarization

Studies which were thought to be limited to isolated single atoms and ions in vacuum are now becoming possible in semiconductors.