�����������

� �� ���� � ���� ����� ��������������

����

Single photon sources based upon single quantum dots

in semiconductor microcav ity pillars

J. A. TIMPSON*y, D . SANV ITTOy, A. D AR AE Iy, P. S. S. G U IMAR AE Sy,

H . V INC K y, S. L AMy, D . M. W H ITTAK E R y, M. S. SK OL NIC K y,

A. M. F OX y, C . Y . H U z, Y .- L . D . H Oz, R . G IB SONz, J. G . R AR ITY z,

S. PE L L E G R INI§ , K . J. G OR D ON}, R . E . W AR B U R TON}, G . S. B U L L E R },

A. TAH R AOU I�, P. W . F R Y � a n d M. H OPK INSON�

y D e p a r t m e n t o f Ph y s i c s a n d As t r o n o m y , U n i v e r s i t y o f Sh e ffi e l d ,

Sh e ffi e l d S3 7 R H , U K

z D e p a r t m e n t o f E l e c t r i c a l & E l e c t r o n i c E n g i n e e r i n g , U n i v e r s i t y o f B r i s t o l ,

B r i s t o l B S8 1 U B , U K

xSc h o o l o f E n g i n e e r i n g a n d Ph y s i c a l Sc i e n c e s , H e r i o t W a t t U n i v e r s i t y ,

E d i n b u r g h , E H 1 4 4 AS, U K

�D e p a r t m e n t o f E l e c t r o n i c a n d E l e c t r i c a l E n g i n e e r i n g , U n i v e r s i t y o f Sh e ffi e l d ,

Sh e ffi e l d S1 3 JD , U K

(Received 10 February 2006; in final form 3 May 2006)

Se m i c o n d u c t o r m i c r o c a v i t y p i l l a r s w i t h b o t h c i r c u l a r a n d e l l i p t i c a l c r o s s - s e c t i o nc o n t a i n i n g s e m i c o n d u c t o r q u a n t u m d o t s a r e s h o w n t o b e g o o d c a n d i d a t e s f o re ffi c i e n t s i n g l e p h o t o n s o u r c e s . Pi l l a r s w i t h s m a l l d i a m e t e r s a r e s h o w n t o h a v ee x c e p t i o n a l l y h i g h q u a l i t y f a c t o r s a n d t h e r e d u c t i o n i n t h e m e a s u r e d q u a l i t yf a c t o r a s t h e p i l l a r d i a m e t e r i s r e d u c e d i s s h o w n t o a g r e e w e l l w i t h fi n i t ed i ff e r e n c e t i m e d o m a i n s i m u l a t i o n . Th e s e p i l l a r s e x h i b i t a Pu r c e l l e n h a n c e m e n to f t h e q u a n t u m d o t e m i s s i o n w h e n t h e d o t s a r e o n - r e s o n a n c e w i t h t h e c a v i t ym o d e a n d s t r o n g p h o t o n a n t i b u n c h i n g . Th e u s e o f t h e p o l a r i z e d m o d e s o fa n e l l i p t i c a l m i c r o p i l l a r a l l o w s t h e p o l a r i z a t i o n o f t h e e m i t t e d s i n g l e p h o t o n s t o b es e l e c t e d .

1 . I ntroduction

An e ffi c i e n t s o u r c e o f s i n g l e p h o t o n s w o u l d a l l o w s i g n i fi c a n t a d v a n c e s i n t h e

fi e l d o f q u a n t u m i n f o r m a t i o n p r o c e s s i n g [ 1 ] , m o s t i m m e d i a t e l y i n q u a n t u m k e y

d i s t r i b u t i o n [ 2 ] . Si n g l e p h o t o n s o u r c e s h a v e b e e n d e m o n s t r a t e d , f o r e x a m p l e ,

u s i n g s i n g l e m o l e c u l e s [ 3 ] , c o l o u r c e n t r e s i n d i a m o n d [ 4 ] a n d s e m i c o n d u c t o r q u a n t u m

d o t s [ 5 , 6 ] . On e a d v a n t a g e o f s e m i c o n d u c t o r q u a n t u m d o t s i s t h e i r c o m p a t i b i l i t y w i t h

t h e a d v a n c e d p r o c e s s i n g t e c h n i q u e s d e v e l o p e d b y t h e s e m i c o n d u c t o r i n d u s t r y ,

a l l o w i n g t h e m t o b e i n c o r p o r a t e d e a s i l y w i t h i n m i c r o c a v i t i e s , i m p r o v i n g b o t h t h e

*C o r r e s p o n d i n g a u t h o r . E m a i l : J.A.Ti m p s o n @ s h e f f i e l d .a c .u k

J ournal of Modern O p t ics

V o l . 5 4 , No s . 2 – 3 , 2 0 Ja n u a r y – 1 5 F e b r u a r y 2 0 0 7 , 4 5 3 – 4 6 5

J ournal of Modern O p t icsISSN 0 9 5 0 – 0 3 4 0 p r i n t / ISSN 1 3 6 2 – 3 0 4 4 o n l i n e # 2 0 0 7 Ta y l o r & F r a n c i s

h t t p : / / w w w .t a n d f .c o .u k / j o u r n a l sD OI: 1 0 .1 0 8 0 / 0 9 5 0 0 3 4 0 6 0 0 7 8 5 0 5 5

� !"# $%&%'( )

* +",-&./,0( 1'.,/0 #2 30)45)6789&:.;$.

( 5<<=

quantum dot emission rate via the Purcell effect [7, 8] and the photon collection

efficiency [7, 8].

In this paper we report small diameter microcavity pillars with exceptionally high

quality factors. The Purcell enhancement factor (PF) of the spontaneous emission of

a quantum dot embedded within the microcavity is calculated to be very high,

with PF > 70 for a 1.5mm diameter pillar. When the quantum dot emission

energy is on-resonance with the cavity mode, the spontaneous emission rate is

shown to be enhanced and, as required for a single photon source, the multi-photon

emission probability is reduced to a very low level (in the case of a 0.6 mm diameter

pillar the multi-photon emission is suppressed by at least a factor of 20 compared

to a Poissonian source with the same intensity). Using the polarized modes of

a micropillar with elliptical cross-section [9–11] the emission from a quantum

dot can be tuned so that the emission has a degree of linear polarization in excess

of 0.9. (Note that similar results have also been observed in certain designs of

photonic crystal [12].)

Section 2 will describe the sample details and section 3 will describe the two

experimental arrangements used in this work. Section 4 contains the experimental

results, starting with the measured and simulated quality (Q) factors. The control of

the polarization of the single photon sources in elliptical micropillars is then

described followed by a discussion of quantum dot lifetimes. Finally, results

demonstrating single photon generation in both circular and elliptical micropillars

will be presented.

2. Sample details

The microcavity structure consists of a one wavelength thick GaAs cavity

surrounded by 27 pairs of alternating Al0.9Ga0.1As/GaAs layers in the bottom

distributed Bragg refl ector (DBR) and 20 pairs in the top DBR. The cavity contains

a single layer of low density InAs quantum dots positioned at the cavity antinode.

The samples were grown by molecular beam epitaxy using in situ refl ectivity

monitoring.

The samples were processed using a combination of electron beam lithography

(EBL) and inductively coupled plasma (ICP) etching. First, a 600 nm layer of SiO2

was deposited, followed by an �300 nm thick poly(methylmethacrylate) layer

which was patterned using EBL. A 100 nm thick layer of aluminium was evaporated

on the sample and the desired pattern transferred to this using lift-off to leave

metal discs; these discs served as a mask allowing the SiO2 to be etched using a

CHF3 based reactive ion etching. The metal discs were then chemically

removed, leaving a SiO2 mask. Pillars were then formed using a SiCl4/Cl2-based

ICP. The SiO2 layer facilitates the etching of high aspect ratio pillars and inhibits

degradation of the side walls which would occur during the removal of the

aluminium mask. A micrograph of a 1.5 mm diameter pillar is shown in the inset

to figure 1. The smooth, nearly vertical side walls clearly show the high quality of

the etching.

454 J. A . T impson et al.

>?@AB?CDEDFG H

I JAKLEMNKOG ?PFMKNO?BQ ROHSTHUVWXEYMZCM

G T[[\

Each sample contained a series of micropillars with both circular and elliptical

cross-sections, the circles ranging in diameter from 50 mm to 600 nm and the ellipses

having major and minor axes ranging from 5 to 1.5 mm and 4 to 0.5 mm, respectively.

3. Experimental details

In this paper two different types of measurements are reported: continuous wave (cw)

and time-resolved measurements. The experimental layout is shown in figure 2.

To measure micropillar Q factors and the polarization of dot emission both on- and

off-resonance with the cavity mode, the following arrangement was used. The sample

was cooled to temperatures between 4.2 and 50K in a liquid helium continuous flow

cryostat. In the cryostat the sample was mounted close to a thin window allowing

high resolution micro-photoluminescence (mPL) spectra to be measured. Excitation

was provided by the 633 nm line from a He:Ne laser. A microscope objective with

a numerical aperture of 0.5 focused the laser to a spot of diameter �2 mm and

collected the emitted PL emission which was then dispersed in a 0.75m double

spectrometer. Using high excitation power (�200 mW in a 2 mm diameter spot) the

measured spectra show broadband emission from, for example, the tail of the wetting

layer, modulated by the cavity mode. Under these conditions single dot emission

was not resolved. Single dot emission was measured using low power excitation

(�5 mW in a 2 mm diameter spot).

A similar apparatus was used for the time-resolved micro-photoluminescence

(TRPL) experiments, except that the laser used was a picosecond mode-locked

Ti:Sapphire laser with a pulse length of 3 ps, a repetition rate of 82MHz and

1 2 30

2

4

6

8

10

12

14

16

Q (

×1000)

P illa r d ia m e te r (µm )

Figure 1. Q uality factors for microcavity pillars with circular cross-sections. Hollow pointsshow the measured Q factors, while solid black circular (solid grey square) points showQ factors predicted by FDTD simulations for micropillars with circular (square) cross-section.The inset shows an electron microscope image of a micropillar with circular cross-sectionand a diameter of 1.5mm, measured at the cavity centre.

S ing le ph oton sources based upon sing le q uantum dots 455

]^_`a^bcdcef g

h i`jkdlmjnf ^oeljmn^ap qngrsgtuvwdxlybl

f szz{

a maximum pulse energy of �1:2� 10�14 J per pulse. The laser wavelength was tuned

to 808 nm and the laser beam was focused by a microscope objective lens (NA ¼ 0.7)

onto the sample with an effective spot size of �3 mm. The luminescence was collected

by the same microscope objective and was sent to a 0.55m monochromator with two

outputs allowing connection to a liquid-nitrogen cooled CCD for mPL measurements

or to a Hanbury Brown and Twiss (HBT) photon correlation apparatus, which

measures the second-order intensity autocorrelation function (gð2Þð�Þ). The HBT

set-up consists of a 50/50 non-polarizing beamsplitter, two fibre-coupled silicon

TRPL setup for

m ea surin g lifetim es

H e-N e ( c w ) / Ti:S a ppire (ps)

S pec trom eter

TIA

sta rt

stop

4 .2 K – 3 0 0 K

C ry osta t

O b jec tiv e

len s

I llum in a tor

B S 1 B S 2

N D

M 2 M 3

TC S PC

H B T setup for g( 2 ) m ea surem en ts

M 4

M 5

B S 3

S PC M

S PC M

S PC M

S y n c

S y n c

V id eo

c a m era

M 1

Figure 2. Optical arrangement for micro-PL, HBT (gð2Þ) and TRPL measurements.A He:Ne laser or a picosecond laser (Ti:Sapphire Tsunami) delivers the laser light to thesample via mirror M1 and the objective lens. BS1 is a dichroic beamsplitter that reflectsthe pump light and transmits the photoluminescence. The photoluminescence is thus collectedand returns along the same route to BS2 where it is directed through M2, M4 and focusedonto the spectrometer slit. A 0.75m double spectrometer (SPEX 1402) or a 0.55m (TRIAX550, Jobin Yvon) spectrometer is used to study spectra (with a charged coupled detector(CCD)) or to filter the light around the quantum dot emission. For gð2Þ measurements mirrorM5 is flipped away and the light transmitted through the spectrometer is recollimated andpasses through a 50:50 beamsplitter BS3, to two fibre-coupled single photon counting modules(SPCM). The photo-pulses are used as start and stop inputs to a time interval analyser (TIA).To measure dot lifetimes, the light transmitted through the spectrometer is again recollimatedthen reflected by mirror M5 and coupled into a fibre connected to the single photon countingmodule (SPCM) and photodetections are recorded using the time-correlated single photoncounting (TCSPC) card. The sample is housed in an optical access cryostat under a 0.5mmthick window. White light illumination can be used with a video camera to allow samplealignment when M2 is flipped away.

456 J. A. Timpson et al.

|}~��}������ �

� ���������� }������}�� ����������������

� ����

single photon counting modules (SPCMs) and a modulation domain analyser

(HP53310A). To save measurement time, a passive time-out circuit was added

between the start and stop inputs. Lifetime measurements were made using a

modified Perkin Elmer SPCM [13, 14] with a timing resolution of �400 ps

(full width at half-maximum) in conjunction with a Becker and Hickl SPC600

time-correlated single photon counting (TCSPC) card with a timing resolution of

11 ps (full width at half-maximum).

4. Results

4.1 Quality factors

The efficiency of a single photon source based upon a single InAs quantum dot

embedded w ithin a planar G aAs layer is limited by the poor ex traction efficiency

of photons from the semiconductor. This efficiency can be improv ed dramatically

by placing the dot w ithin a microcav ity [ 7 , 8 ] . A suitable microcav ity w ill modify

the spontaneous emission of a dot on- resonance w ith the cav ity mode, both

increasing the spontaneous emission rate and improv ing the collection efficiency,

thereby increasing the efficiency of the source. The decrease in lifetime of a quantum

dot on- resonance w ith a cav ity mode is characteriz ed by the P urcell factor ( PF) [ 1 5 ] .

To obtain a high PF , microcav ities w ith high Q factors and low modal v olumes ( V)

are required ( PF ¼ ð3Q=4p2VÞ �c=nð Þ3 , w here �c is the air w av elength of the cav ity

mode and n is the refractiv e index of the medium). The measured Q factors for

small diameter microcav ity pillars are plotted in fi gure 1 . The measured Q factors of

1 2 0 0 0 for a 1 . 5 mm diameter pillar and 30 0 0 for a diameter of 0 . 9 mm are amongst the

highest reported for small diameter pillars ( diameter <2 mm) [ 1 1 ] . From these

measured Q factors, a max imum v alue of PF in ex cess of 7 0 is calculated for

a 1 . 5 mm diameter pillar.

For larger pillars ( w ith diameter �2 mm) or planar microcav ities, the w av elength

of the fundamental cav ity resonance is principally determined by the optical

thick ness of the cav ity spacer. As the diameter is reduced until it approaches 1 mm

( i. e. w av elength scale) the fundamental resonance is blue- shifted due to the increasing

transv erse component in the H E 1 1 w av eguide mode. This can be quantitativ ely

understood by using approx imate simulations, such as the transfer- matrix method

coupled to a calculation of the w av eguide modes using an av erage pillar refractiv e

index . H ow ev er for detailed estimates of the cav ity Q, w e used the 3D Finite

D iff erence Time D omain ( FD TD ) simulation method. For the FD TD method

[ 1 6 ] , w e place a broad band dipole source in the centre of the microcav ity and

apply a short ex citation pulse to probe the mode spectrum. The cav ity then rings at

its resonant frequency and w e monitor the cav ity ringdow n using a probe abov e the

pillar. Tak ing the Fourier transform of the ringdow n signal ( in the time domain)

allow s us to determine the resonant frequencies of the w av eguide cav ity and also

giv es us an estimate of the Q factor ( Q ¼ �=��). W e hav e calculated Q factors for

v arious pillar diameters w ith the same cav ity design as the pillars used in these

S i ng l e p h o t o n s o u r c e s b a s e d u p o n s i ng l e q u a nt u m d o t s 45 7

������ ¡¢¡£¤ ¥

¦ §�¨©¢ª«¨¬¤ �­£ª¨«¬��® ¯¬¥°±¥²³´µ¢¶ª· ª

¤ ±¸¸¹

experiments (20 top mirror pairs and 27 bottom pairs). The simulation results and

the experimental data are shown in figure 1.

The Q factors obtained from both experiment and FDTD simulations show the

expected behaviour. The high Q factors found in large diameter micropillars are

maintained down to a pillar diameter of 1– 2 mm at which point the field intensity

at the pillar edge becomes significant leading to increasing loss for smaller pillars and

a corresponding reduction in Q factor [17]. Although the Q factors obtained from

both measurement and simulations follow the same trend, the measured values are

lower than the simulated ones, particularly for small pillars with diameter less

than �1.5mm where the Q factors are being degraded due to surface imperfections

not included in the FDTD simulations. This indicates that although the Q factors

reported here are amongst the best reported for pillars of such small dimensions,

the theoretical limit achievable in perfect pillars is yet to be reached.

4.2 Polarized single dot emission

A single photon source which emits photons with a defined polarization is

important for many applications in quantum optics and quantum computation

(where quantum indistinguishability is important) and in some implementations of

quantum cryptography [18]. S uch a polarized single photon source will be twice as

efficient as using an unpolarized source in conj unction with a polarizer since using

a polarizer involves rej ection of half the emitted photons.

The high power mPL spectrum for an elliptical micropillar with maj or and minor

axis 2.0 and 0.8mm respectively, measured using the 633 nm, cw, HeN e laser, is

shown in figure 3(a). Two cavity modes are shown with opposite linear polarizations.

The lower energy mode is polarized with the E field parallel to the maj or axis of the

ellipse (x-polarized) and the higher energy mode is polarized with the E field parallel

to the minor axis of the ellipse (y-polarized). The emission of a quantum dot emitting

off-resonance with a cavity mode is, in general, unpolarized. As the emission

energy of the quantum dot is tuned by varying the temperature, the component of

emission from the dot having the same polarization as the resonant mode is

selectively enhanced. When the dot is on-resonance with the polarized cavity

mode, the emission from the quantum dot can have a degree of linear polarization

in excess of 0.9y ; the degree of enhancement depending upon how well the quantum

dot couples to the cavity.

Figure 3(b) shows the low power mPL spectrum from the pillar described above.

The x-(y-)polarized emission spectrum is shown in black (grey). The spectra were

measured at a temperature of 35 K where one quantum dot (labelled D) is

on-resonance with the x-polarized cavity mode. It is clear that this dot is predomi-

nantly x-polarized in contrast to the other dots shown, which are not on-resonance

with either cavity mode and are unpolarized. The high degree of polarization of this

dot is due to the selective enhancement of the x-polarized PL emission when the dot

yDegree of polarization is defined as ½IðPxÞ � IðPyÞ� =½IðPxÞ þ IðPyÞ� , where IðPxÞ and IðPyÞ arethe x- and y-polarized mPL intensities respectively.

458 J . A . T impson et al.

º»¼½¾»¿ÀÁÀÂà Ä

ŠƽÇÈÁÉÊÇËà »ÌÂÉÇÊË»¾Í ÎËÄÏÐÄÑÒÓÔÁÕÉÖ¿É

à Ð××Ø

is on-resonance with the x-polarized cavity mode and not due to any intrinsic

polarization of the dot. This is clearly shown in figure 3(c) which shows the

x-polarized mPL spectra for this pillar at temperatures between 5 and 43K. B y

varying the temperature, the dot (D) emission is tuned into and out of resonance

with the cavity mode (C ). A clear enhancement of the PL emission is observed when

the dot is on-resonance with the cavity mode. This is emphasized in figure 3(d) which

shows the intensity of both polarization components of the dot emission as

a function of detuning from the cavity mode. The x-polarized component shows a

clear peak on-resonance while the y-polarized emission shows only the small decrease

in intensity associated with raising the sample temperature.

4.3 Lifetimes and Purcell factors

Figures 4(a) and (b) show the low power mPL spectra from a micropillar

with elliptical cross-section where the major and minor axes are 1.5 and

939 940 941 942 943

941 942 943 944

0.0

0.3

0.6

0.9

35 K(b)

De tu n in g (n m )

D

W a v e le n gth (n m )

W a v e le n gth (n m )

0

5

10

15

20(a)

Cy

Cx

PL

in

ten

sity (

a.u

.)P

L inte

nsity (

a.u

.)P

L in

ten

sity

(a.u

.)P

L in

ten

sity

(a.u

.)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

(d)

(c)

T = 43 K

T = 5 K

C D

DC

−1.2 −0.9 −0.6 −0.3 0.0 0.3

0.0

0.5

1.0

x-p o la ris e d

y-p o la ris e d

Figure 3. (a) High power mPL spectra from an elliptical micropillar with major and minoraxes of 2.0 and 0.8mm respectively. Emission polarized parallel to the major (x-polarized)and minor (y-polarized) axis of the ellipse is shown in black and grey respectively. (b) Lowpower mPL spectra from the same pillar measured at 35K: x-(y-)polarized emission isshown in black (grey). A dot, D, is on-resonance with the x-polarized cavity mode andis highly polarized. (c) Low power x-polarized mPL spectra for the sample at temperaturesbetween 5 and 43K. As the temperature is varied the dot, D, moves into and out of resonancewith the cavity mode C . (d ) PL intensity of the x- and y-polarized components for dot Das a function of detuning from the cavity mode.

Single photon sources based upon single quantum dots 459

ÙÚÛÜÝÚÞßàßáâ ã

ä åÜæçàèéæêâ ÚëáèæéêÚÝì íêãîïãðñòóàôèõÞè

â ïöö÷ 0.5mm respectively. A pillar was selected that had a quantum dot emitting on

resonance with the fundamental cavity mode at 4.2K as shown in figure 4(a). This

dot was moved off-resonance by raising the temperature to 35K (figure 4(b)).

Figure 4(c), shows the normalized TR PL [19] traces taken when the dot is both

on- and off-resonance at temperatures of 4.2 and 35K respectively; an excitation

wavelength of 811 nm and a repetition rate of 82 M Hz was used to excite the pillar.

The instrumental response of the system is shown for comparison. An exponential

fit to the decays over the first few nanoseconds revealed that the quantum dot

lifetime (�) shortened from 1350 ps when the dot was off resonance with the cavity

mode, to 550 ps when the dot was on-resonance, giving a measured Purcell factor

of �2.5. As noted by other groups [7, 20, 21], this measured Purcell enhancement

is smaller than that calculated using the measured Q factor for this pillar, indicating

non-optimal coupling between the cavity and the dot.

O ne issue arising from the use of the Ti:Sapphire laser was that its repetition

frequency of 82 M Hz was found to be too high and did not permit the dot emission to

decay fully before the incidence of the next pulse. The effect of the laser repetition rate

on the decay time was investigated by repeating the experiment using a semiconductor

laser with a wavelength 784 nm, pulse width of 80 ps and a repetition frequency of

5 M Hz. A comparison of the results is shown in figure 5. It is clear that the TR PL has

5

10

15

20

25

30

35

Dot off resonanc e

Dot on resonanc e

(b)

(a)

C

35 K

4.2 K

Inte

nsity (

a.u

.)In

tensity (

a.u

.)

910 912 914 916 918

5

10

15

Wavelength (nm)

0 2 4 6

0.01

0.1

1

Norm

alis

ed c

ounts

Dot on resonanc e at 4.2KDot off resonanc e at 35K

(c)

T ime (ns)

Instrumental

Figure 4. (a) mPL spectrum of an elliptical micropillar (major and minor axis 1.5 and 0.5 mmrespectively) at a temperature of 4.2K where a quantum dot is on-resonance with the cavitymode at a wavelength of 912.9 nm. (b) PL spectrum of the same pillar at a temperatureof 35.0K where the dot is off-resonance at 914.6 nm and the cavity mode C is at 913.3 nm.(c) TR PL measurements of the dot on-resonance at 4.2K in black and off-resonance at 35Kin grey. The instrumental response of the system is also shown.

460 J. A. Timpson et al.

øùúûüùýþÿþ�� �

� �û��ÿ���� ù����ùü� �� ������ÿ���ý�

� ����

not fully decayed to the background count level in the 82MHz repetition rate case,

whilst the signal has fully decayed in the lower repetition frequency excitation

measurement. This indicates the presence of a long-lived component in the lumines-

cence decay with a lifetime � � 10–40 ns depending upon the excitation conditions.

Exponential fits to the TRPL data obtained using the 784 nm, 5MHz excitation

laser did not show as significant a reduction in lifetime as the dot was moved

on-resonance. Further studies are in progress to determine the optimum excitation

conditions for this type of measurement. However it is clear that a combination of low

excitation frequency and close-to-resonant photon energy will be required.

4.4 Single photon generation in circular and elliptical pillars

Single Q D emission can be isolated in small diameter micropillars (<4 mm) under low

pump power (inset to figure 6(b)), and can be tuned on- and off-resonance with the

cavity mode by varying temperature. At higher excitation powers, the cavity mode

dominates with single Q D emissions washed out. In this subsection we report single

photon generation in circular pillars and polarized single photon generation in

elliptical pillars.

Single photon generation was examined by measuring the second-order intensity

autocorrelation function (gð2Þð�Þ) using the HBT arrangement shown in figure 2.

gð2Þð�Þ is defined as

gð2Þð�Þ ¼hnðtÞnðtþ �Þi

hni2¼

pðt : tþ �Þ

pðtÞ,

0 4 6 82

1

10

100

1000(a)

λex = 811 nm

82 M H z

Counts

(a.u

.)

Time (ns)

0 25 50 7 5 100

(b)λex = 7 84 nm

5 M H z

Time (ns)

Figure 5. Comparison of TRPL traces of a quantum dot on resonance with the cavity modeat a temperature of 5K obtained using different repetition rate lasers. The backgroundsignal measured with no laser excitation is shown for comparison in grey. (a) TRPL traceobtained using an excitation wavelength of 811 nm, repetition frequency of 82MHz andperiod of 12.2 ns. (b) TRPL trace obtained using an excitation wavelength or 784 nm,repetition frequency of 5MHz and period of 200 ns. The time t¼ 0 is set by the timingelectronics. Note that the decay at 82MHz does not reach the background level, whileit does reach the background level when a frequency of 5MHz is used.

Single photon sources based upon single quantum dots 461

����������� !

" #�$%�&'$( �)�&$'(��* +(!,-!./01�2&3�&

-445

where n(t) is the photocount probability at time t and the angular brackets denote

ensemble averages. This can also be cast in terms of the conditional probability of

recording photodetections at times t and tþ �, pðt : tþ �Þ normalized by the

probability for recording a single detection p(t). At low count rates the time interval

analyser (TIA) measures the probability that a stop pulse at delay time � follows the

start pulse; this probability is proportional to pðt : tþ �Þ, allowing us to estimate

gð2Þð� ¼ 0Þ after normalization. Of course, in this set-up there were two detectors

viewing the same source through the beamsplitter as a single detector cannot

measure a second photo-pulse arriving during the detector dead time after the first

pulse. However this still provides a good estimate of gð2Þð�Þ. Also in the case of pulsed

900 910

1 µW

11.4 µW

PL inte

nsity (

a.u

.)

λ (nm)

T= 35 K

1 2 3 4 5 6 70.0

0.2

0.4

0.6

0.8

1.0

Pump pow er (µW)

g(2

) (0)

(b)

−60 −40 −20 0 4020 60 80

0.0

0.5

1.0

1.5

Coin

cid

ence p

robabili

ty (

a.u

.)

Time (ns)

0.05

0.881.07

0.990.93 0.99

1.030.89

1.05

0.88

1.03

(a)

Figure 6. (a) Measured second-order correlation function for emission from the quantumdot on resonance with the cavity as shown in the inset to figure 6(b). The multi-photonprobability is very low. (b) The power dependence of g2ð0Þ. Inset, PL spectrum of a quantumdot on resonance with the cavity mode of a circular micropillar of diameter 600 nm excitedat � ¼ 810 nm and power P¼ 1 mW (black line) and 11.4 mW (dotted line).

462 J. A. Timpson et al.

6789:7;<=<>? @

A B9CD=EFCG? 7H>ECFG7:I JG@KL@MNOP=QER;E

? LSST

pumping we will record a set of peaks in the time interval plot due to the fact that the

probability of stops falling outside the pump pulses is low. For ideal single photon

sources, the peak at zero time delay should disappear completely, representing

a probability of zero for a multi-photon event [3, 6, 22].

In a circular pillar with diameter of 0.6 mm, we observed a rather weak central

coincidence peak (figure 6(a)). The normalized area of the central peak was

calculated using a Lorentzian fitting and gave gð2Þð� ¼ 0Þ ¼ 0:05, indicating that

multi-photon emission is suppressed by a factor of at least 20. With increasing pump

power, the single QD emission saturates and the cavity mode intensity develops

rapidly. As a result, gð2Þð� ¼ 0Þ increases rapidly with the pump power (as shown in

figure 6(b)).

910 912 914 916 918

polarization along minor axispolarization along major axis

PL

in

ten

sity (

a.u

.)

Wavelength (nm)

T= 4.3 K

Pex = 1 µW D

−60 −40 −20 200 40 600.0

0.2

0.4

0.6

0.8

1.0

1.2

Time (ns)

Co

incid

en

ce

pro

ba

bili

ty (

a.u

.)

T=4.3 KPex = 0.5 µW

(b)

(a)

Figure 7. (a) Spectrum showing the emission from an elliptical micropillar with majorand minor axes 1.5 and 0.5 mm respectively. The emission polarized parallel to themajor (dashed line) and minor axis (solid line) is shown. A quantum dot (D) is onresonance with the cavity mode at 912.9 nm. (b) gð2Þð�Þ of the emission from the quantumdot labelled (D) in (a).

Single photon sources based upon single quantum dots 463

UVWXYVZ[\[]^ _

` aXbc\debf^ Vg]dbefVYh if_jk_lmno\pdqZd

^ krrs

Progressing from circular pillars to elliptical pillars, the two-fold degenerate

cavity is split into two modes: one polarized along the major axis (x-polarized) and

another polarized along the minor axis (y-polarized) as discussed in section 4.2. In an

elliptical pillar with a major axis of 1.5 mm and a minor axis of 0.5 mm, we observe

a QD on resonance with the cavity mode at 912.9 nm, polarized along the minor axis

(figure 7(a)). The polarized single photon generation is verified by observation

of a weaker central coincidence peak with g2 � 0:3 (figure 7(b)) [10, 22].

5. Conclusions

Semiconductor microcavity pillars are shown to provide the basis for efficient

sources of single photons, both polarized and unpolarized. A Purcell reduction of

the lifetime of a quantum dot when on-resonance with a cavity mode of �2.5 has

been demonstrated. The multiphoton emission probability from a circular micro-

pillar is shown to be reduced by a factor of 20, while that of a quantum dot within

an elliptical microcavity is also shown to be reduced, but by a smaller factor. It has

been demonstrated that the use of elliptical micropillars allows the polarization of

the emitted photon to be controlled. All of these features demonstrate the excellent

potential of quantum dot micropillar sources in quantum information processing

applications.

A ck now le d g e m e nt s

The authors would like to acknowledge the support of EPSRC and MOD/ DSTL

through the J oint Research Grants Scheme ref. no. Dstl/ GR/ T09408, support from

QIP IRC (GR/ S822176) and EU IP:015848 QAP and use of Edinburgh Instruments

Ltd, Analytical T900 Software V 6.0.

R e f e r e nce s

[1] D.J . Mowbray and M.S. Skolnick, J . Phys. D Appl. Phys. 38 2059 (2005).[2] S. Benjamin, Science 29 0 2273 (2000).[3] B. Lounis and W.E. Moerner, Nature 40 7 491 (2000).[4] C. Kurtsiefer, S. Mayer, P. Z arda, et al., Phys. Rev. Lett. 8 5 290 (2000).[5] P. Michler, A. Imamoglu, M.D. Mason, et al., Nature 40 6 968 (2000).[6] C. Santori, M. Pelton, G. Solomon, et al., Phys. Rev. Lett. 8 6 1502 (2001).[7] J .M. Gerard, B. Sermage, B. Gayral, et al., Phys. Rev. Lett. 8 1 1110 (1998).[8] G.S. Solomon, M. Pelton and Y . Y amomoto, Phys. Rev. Lett. 8 6 3903 (2001).[9] B. Gayral, J .M. Gerard, B. Legrand, et al., Appl. Phys. Lett. 7 2 1421 (1998).[10] D.C. U nitt, A.J . Bennett, P. Atkinson, et al., Phys. Rev. B 7 2 033318 (2005).[11] A. Daraei, A. Tahraoui, D. Sanvitto, et al., Appl. Phys. Lett. 8 8 051113 (2006).[12] D. Englund, D. Fattal, E. Waks, et al., Phys. Rev. Lett. 9 5 013904 (2005).[13] I. Rech, I. Labanca, M. Ghioni, et al., Rev. Sci. Instrum. 7 7 033104 (2006).[14] K.J . Gordon, V . Fernandez, G.S. Buller, et al., Opt. Express 1 3 3015 (2005).[15] E.M. Purcell, Phys. Rev. 6 9 681 (1946).

464 J. A. Timpson et al.

tuvwxuyz{z|} ~

� �w��{����} u�|����ux� ��~��~����{���y�

} ����

[16] Y.-L.D. Ho, T. Cao, P.S. Ivanov, et al., Int. J. Quantum Informat. 3 229 (2005).[17] D. Sanvitto, A. Daraei, A. Tahraoui, et al., Appl. Phys. Lett. 86 19, 1109 (2005).[18] N. Gisin, G. Ribordy, W. Tittel, et al., Rev. Mod. Phys. 74 145 (2002).[19] J.M. Smith, P.A. Hiskett, I. Gontijo, et al., Rev. Sci. Instrum. 72 2325 (2001).[20] C. Santori, D. Fattal, J. Vuckovic, et al., New J. Phys. 6 89 (2004).[21] A.J. Bennett, D.C. Unitt, P. Atkinson, et al., Opt. Express 50 5759 (2004).[22] E. Moreau, I. Robert, J.M. Gerard, et al., Appl. Phys. Lett. 79 2865 (2001).

Single photon sources based upon single quantum dots 465