Multiexciton Generation in Seeded Nanorods - The Journal of Physical

Jul 5, 2014 - *E-mail: [email protected] (H.E.), *E-mail: [email protected] (R.B.), ... The formalism is applied to a series of CdSe/CdS seeded ...
0 downloads 0 Views 639KB Size
Letter

Subscriber access provided by Otterbein University

Multixciton Generation in Seeded Nanorods Hagai Eshet, Roi Baer, Daniel Neuhauser, and Eran Rabani J. Phys. Chem. Lett., Just Accepted Manuscript • DOI: 10.1021/jz5010279 • Publication Date (Web): 05 Jul 2014 Downloaded from http://pubs.acs.org on July 15, 2014

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

The Journal of Physical Chemistry Letters is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry Letters

Multiexciton Generation in Seeded Nanorods Hagai Eshet,



Roi Baer,



Daniel Neuhauser,



and Eran Rabani

School of Chemistry, The Sackler Faculty of Exact Sciences, Tel Aviv University,Tel Aviv 69978, Israel, Fritz Haber Center for Molecular Dynamics, Institute of Chemistry, The Hebrew University of Jerusalem, Jerusalem 91904, Israel, and Department of Chemistry, University of California at Los Angeles, CA-90095 USA

E-mail:

∗ †

To whom correspondence should be addressed School of Chemistry, The Sackler Faculty of Exact

Sciences, Tel Aviv University,Tel Aviv 69978, Israel



Fritz Haber Center for Molecular Dynamics, Insti-

tute of Chemistry, The Hebrew University of Jerusalem, Jerusalem 91904, Israel



Department of Chemistry, University of California

at Los Angeles, CA-90095 USA

ACS Paragon Plus Environment 1

The Journal of Physical Chemistry Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 10

Abstract The stochastic formulation of multiexciton generation (MEG) rates is extended to provide access to MEG eciencies in nanostructures containing thousands of atoms. The formalism is applied to a series of CdSe/CdS seeded nanorod hetero-structures with dierent core and shell dimensions. At energies above 3Eg (where Eg is the band gap) the MEG yield increases with decreasing core size, as expected for spherical nanocrystals. Surprisingly, this behavior is reversed for energies below this value. This anomalous behavior is explained by the behavior of the density of states near the valence band edge, which increases with the core diameter. Our predictions indicate that the onset of MEG can be shifted to lower energies by manipulating the density of states in complex nanostructure geometries.

Graphical TOC Entry

Multiexciton generation yield in CdSe/CdS seeded nanorods.

Keywords: Semiconductors, heterostructures, electronic structure, solar cells, nanostructures. Multiexciton generation by which more than

of impact ionization.

18,19,21,29

Using this frame-

a single electron-hole pair is generated from

work and a unied Green function approach to

a single photon upon optical excitation is a

MEG it was shown that size scaling of the band

promising direction to collect hot carriers, po-

gap, Coulomb couplings and density of states

tentially increasing photovoltaic power conver-

(DOS) lead to an overall increase in MEG rates

sion eciencies in solar cells to above

and eciencies with decreasing NC size.

45%. 1,2

It is common to dene an onset photon energy

Eon

Several dierent approaches have been pro-

beyond which MEG eciencies are

posed recently to lower the onset energy ra-

noticeable. One of the necessary conditions for MEG to be of technological use is that

Eon

32

tio

be

Eon /Eg .

to decrease

Gabor et al.

Eon /Eg

33

used internal elds

in carbon nanotubes pho-

close to the lower bound of twice the quasi-

todiodes,

particle band gap (2 Eg ).

eld-induced acceleration of the charge carri-

tors

Eon /Eg > 5

1

In bulk semiconduc-

and thus MEG is not use-

ers.

34

a result that was rationalized by

Sandberg et al.

35

studied the role of

ful for photovoltaic applications. This large ra-

shape on MEG eciency and argued for a

tio is a consequence of energy and momentum

60%

conservation constraints which severely limit

tor nanorods (NRs) compared to spherical NCs.

the number of multiexciton states that can be

The eect of NR diameter and length on MEG

formed coupled with the existence of very rapid

rates was analyzed by considering the scaling

non-radiative relaxation pathways.

of the density of trion states (DOTS) and the

3,4

Conned

increase in MEG yields in semiconduc-

36

semiconductor nanocrystals (NCs) on the other

Coulomb couplings.

hand, oer improved MEG eciencies,

with

ing from spherical NCs leading to a surprising

This re-

result where the MEG rates, which are given by

sult was rationalized by several dierent theo-

the product of the two, were roughly indepen-

retical treatments.

dent of the NR length.

an onset energy below

1831

Eon < 3Eg . 717

5,6

Perhaps the most intu-

itively appealing picture of MEG in NCs is that

Both show distinct scal-

36

This prediction was

recently veried by experiments.

ACS Paragon Plus Environment 2

37,38

Page 3 of 10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry Letters

Two additional handles can be used to control the onset energy ratio

Eon /Eg .

generating an exciton

The rst

Klimov and cowork-

In the above equations,

energy in Type-II core-shell NCs can be re-

dressing the role of phonon emission in nanos-

40,41

a, b, c... for unocr, s, t, u... are general indices, and σ =↑, ↓ is the spin index. µ0S is the transition dipole from |0i to |Si, γ is the phonon emission rate, and ΓS is the MEG rate from state|Si.

This allows for reduction of In order to achieve this, we

study MEG eciency in seeded nanorods heterostructures

42,43

It is interesting to note two limits of Eq. (1):

using a stochastic formulation

of the MEG formation rates.

30

When the MEG rate is much larger than the

We then extend

the formalism to calculate the MEG eciency in a series of seeded nanorods with dierent seed

MEG becomes

highly ecient ( nex

and when the

(nex

the MEG eciency increases with the

of

Eg .

γ

32

approach, which assumes that each exciton gen-

is computation-

).

ΓS

is written as as sum of rates

+ ΓS ≡ Γia = Γ− a + Γi ,

erated upon absorption can either decay to the

γ

(3)

and a stochastic formula for the negative and positive trion formation rates is adopted:

or produce another electron-hole pair with a

30

The master equation approach was

previously applied to study MEG in spherical nanocrystals and predicts MEG eciencies similar to those calculated by a direct absorption Green function formalism within the wide band limit.

Γs

of creating positive and negative trions:

seeded nanorods we adopt the master equation

32

and

from experiments and assume it is energy

independent

In order to calculate the MEG eciencies in

ΓS .

γ

ally expensive for large NCs. We take the value

eciency even when the photon energy is scaled

rate

MEG is suppressed

(ω) → 1).

The calculation of

seed radius, i.e., larger seeds show higher MEG

band edge by phonon emission with a rate

 ΓS )

to the MEG rate ( γ

the case of spherical NCs, we nd that below

by

 γ), (ω) → 2)

phonon emission rate (ΓS

phonon emission rate is much larger compared

size and dierent NR diameter. In contrast to

Eon /Eg

i, j, k...

cupied (electron) states.

and at the same time increase of the

MEG eciency.

We use indices

for occupied (hole) states and

is based on controlling the density of states near

Eon /Eg

|0i.

Slater determinant,

The other

handle, which forms the crux of our approach, the band edge.

aσ i = a†aσ aiσ |0i |Si ≡ |Siσ

obtained by applying the annihilation ( aiσ ) and † creation (aaσ ) operators on the ground state

Theoretically, ad-

tructures is rather challenging.

(2)

is a shorthand notation for a single exciton state

duced markedly by partially blocking the non-

39

ES :

PS (ω) = P

ers have recently shown that the MEG onset

radiative decay channel.

with energy

|µ0S |2 δ (ES − ~ω) . 2 S 0 |µ0S 0 | δ (ES 0 − ~ω)

is based on suppressing the competing channel of phonon emission.

|Si

2π ~ 2π = ~

= Γ− a



Wa2 ρ− T (εa )

Γ+ i



Wi2 ρ+ T (εi ) ,

Using the master equation approach it

(4)

is straightforward to show that the MEG e-

where the negative and positive trion density of

ciency is related to the number of excitons gen-

state (DOTS) is given by

erated at steady state, which is given by

nex (ω) =

X S

2ΓS + γ , PS (ω) ΓS + γ

21

ρ− T (ε) =

X

ρ+ T

X

δ (ε − (εb + εc − εj ))

bcj

(1)

(ε) =

δ (ε − (εj + εk − εb )) ,

(5)

jkb

nex (ω) represents the number of excitons generated at a photon energy ~ω (the MEG eciency is simply given by nex (ω)−1) and PS (ω) where

and the average square Coulomb coupling for electrons and holes is given by

is the probability for absorbing the photon and

ACS Paragon Plus Environment 3

The Journal of Physical Chemistry Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 10

tic procedure.

X

2 2 Wa = Wa;cbj δ (εa − (εb + εc − εj )) /ρ− T (εa )

3. Filter target electron and hole states with energies sampled from

bcj

X

2 2 Wi = Wi;jkb δ (ε − (εj + εk − εb )) /ρ+ ). T (εi(6) jkb

Γ− a

and

ES .

Γ+ i

an average over the sampled electronP 2ΓS +γ 1 hole pairs: nex (ω) = S PS (ω) ΓS +γ NS where PS (ω) = ρ (εa ) ρ (εi ) PS (ω). The

where

product

is the screened Coulomb interactions and

at energy

summation expression (cf., Eq. (1)) with

2 Wi;jkb = 2 |Vjibk − Vkijb |2 + |Vkijb |2 + |Vjibk(7) |2 ,

ψr (r)ψs (r)ψu (r0 )ψt (r0 )  |r − r0 |

µ0S

to generate the MEG rate − for each electron-hole pair: ΓS≡ia = Γa + + Γi . Obtain nex (ω) by replacing the full

2 = 2 |Vacjb − Vjcab |2 + |Vjcab |2 + |Vacjb |2 Wa;cbj

d3 rd3 r0

(above and

for each electron-hole pair, calculate the

4. Use

couplings are given by

¨

ρ (ε)

below the Fermi energy, respectively) and transition dipole

In the above equations, the squared Coulomb

Vrsut =

30

ρ (εa ) ρ (εi )

is used to correct for

the density of electron and hole pairs at energy

(8)

ES . NS

is the number of excited

ES = ~ω , NS = 200.

states sampled at energy cally taken to be

ψr (r)

typi-

are the single-particle orbitals (see more details

The rst two steps have been developed previ-

below). We assume a static dielectric screening,

ously by Baer and Rabani and applied to both

,

NCs and NRs.

for the Coulomb matrix elements.

30

The last two provide an exten-

To calculate the MEG eciency given by

sion to calculate the MEG eciencies by sam-

Eq. (1), one requires as input the MEG rates

pling the singly excited states rather than cal-

and the transition probabilities for all possiaσ ble single excitons |Si ≡ |Siσ i. For spherical nanocrystals these are readily available for

culating them all.

moderate excitation energies ( ≤

in Ref.

3Eg ). 21

In Fig. 1 we plot the negative and positive trion formations rates (upper panel, shown also

How-

30

) and the MEG eciency (lower panel,

3nm

ever, for nanorods (and also seeded nanorods),

shown here for the rst time) for a

it is practically impossible to generate all possi-

ameter CdSe nanocrystal.

ble singly excited states, even at relatively low

sults are compared with the full summation ap-

excitation energies, due to the large density of

proach, labeled deterministic. The agreement

states.

Therefore, in order to obtain the MEG

between the stochastic and deterministic calcu-

eciency, we propose a stochastic approach,

lations is excellent, albeit the fact that stochas-

where we sample the singly excited states and

tic rates are slight underestimated compared

the corresponding MEG rates rather than cal-

with the average deterministic rates.

culating them all, to perform the full summa-

ing the results near the valence and conduction

36

tion formula for

nex (ω).

The approach can be

Compar-

approach of the energy falls short in describing the trion formation rates for

1. Calculate the single particle density of

the phonon emission rate,

Use convolutions to generate the

density of single excited states ( ρX

ε/Eg → ±1.

How-

ever, since these rates are very low compared to

states (ρ (ε)) using a stochastic trace for-

30

30

band edge, it is clear that the uniform sampling

summarized as follows:

mula.

di-

The stochastic re-

γ , assuming a vanish-

ing value does not aect the MEG eciencies.

(E))

and the negative/positive trion density of − + 30 states (ρT (ε)/ρT (ε)). 2. Generate the negative/positive trion formation rates (Eq. (4)) using the stochas-

ACS Paragon Plus Environment 4

Page 5 of 10

only one trajectory was used for each structure.

2

-1

Γ (ps )

10

The single particle orbitals ( ψr (r)) required for

0

10

the calculation of the MEG rates were then

-2

generated for the relaxed congurations using

10

a semi-empirical pseudopotential models on a

-4

10 -2

-1

0

1

2

3

4

real space grid,

ε / Eg

2

function.

Stochastic Deterministic

1.8

n(ω)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry Letters

46,47

48

with a Gaussian ltering

1.6

In Fig.2 we show the MEG eciencies (left

1.4

panels) and the trion formation rates (right

1.2

panels) for the two sets of rod diameters. The MEG eciency increases with photon energy

1 2

3

2.5

4

3.5

with an onset close to 3 Eg and

ω / Eg Figure 1:

Drod = 4

The trion formation rate (upper

and

6nm,

respectively.

3.5Eg

for the

This is con-

sistent with experiments for a wide range of

panel) and the MEG eciency (lower panel) for

conned nanostructures.

5

At energies above the

3nm diameter CdSe nanocrystal (Cd 232 S e251 )

onset energy, we nd that the MEG eciency

as a function of scaled energy ( ω is the pho-

at a scaled energy increases as the size of the

a

ton energy while

ε

is the energy of the electron

nanocrystal seed decreases, consistent with cal-

21

(hole), for negative (positive) trions, measured

culations of spherical nanocrystals.

from the bottom of the conduction (top of the

ton energies below the onset (between

valence) band). The full summation approach

3Eg ) the reverse is true and the MEG eciency

labeled deterministic is shown by red circles

is higher for larger seeds.

and the stochastic approach spanning a larger

the inset of Fig. 2 and is qualitatively dierent

energy range is shown by black circles.

from spherical nanocrystals.

For pho-

2Eg

and

This is depicted in

Despite the fact that the MEG eciencies for

In the lower panel of Fig. 1 we compare

2Eg

energies between

ministic approach.

in seeded nanorods, understanding the origin of

The stochastic approach

and

3Eg

the stochastic MEG eciencies to the deter-

are rather small

resulting from

the behavior below the onset is still signicant

the small systematic deviation of the trion for-

as it may prove useful in designing materials

mation rates. However, the overall agreement is

with higher MEG in this important range of

excellent and the stochastic approach is reliable

energies. In the right panels of Fig. 2 we plot − the negative trion formation rate ( Γe (ε)) for all corresponding cases. Since the MEG pro-

slightly underestimates

nex (ω),

at energies much higher than those accessible by the deterministic calculations. Moving

on

nanorods of

to

the

20nm

CdSe/CdS

cess is dominated by the formation of negative

core/shell

trions for CdSe NCs,

length, we calculated the

Drod

core diameters

in the calculations of the MEG eciencies), the − positive channel. For all cases studied, Γe (ε) increases with the electron energy, similar to

4nm and 6nm and for several Dcore in the range 2 − 4.5nm in of

each case. The rods were faceted and the center of the spherical CdSe core was placed at of the length of the rod.

the behavior of spherical

1/3

nanostructures.

The congurations

100ps

recently developed force-eld

at

44

300K

with a

accurately de-

45

29,32

and rod shaped

36

However, there seems to be

an isosbestic point about ε/Eg = 1.75 where Γ− e (ε) changes behavior (this is more clearly observed in the DOTS discussed below). This

of the seeded rods were relaxed using molecular dynamics runs for

we focus on this channel

and ignore, for the sake of the discussion (not

MEG formation rates and eciencies for rod diameters

21

The

isosbestic point is correlated with the crossover

nal congurations were quenched to remove

of the MEG eciency with the nanocrystal seed

structural eects of thermal uctuations, thus,

size.

scribing

CdSe/CdS

hetero-structures.

To delineate the factors governing this be-

ACS Paragon Plus Environment 5

The Journal of Physical Chemistry Letters 2

4

10

2

10

-1

Γe (ps )

nex(ω)

1.75 1.5 1.25

2

2.5

3

D=4nm

0

10

-2

10

D=4nm

-4

10

1

2

10 -1

Γe (ps )

1.75

nex(ω)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 10

1.5 1.25

D=6nm

1 2

2.5

3

4

3.5

4.5

0

10

-2

10

D=6nm

-4

10

5

0

1

ω / Eg Figure 2: MEG eciency for

2

3

4

ε / Eg

Drod = 4

6

and

nm diameter nanorods (upper and lower left panels,

respectively) and the corresponding trion formation rates (right panels) for scaled energies. For the

Drod = 4nm rods the blue, black, red and green curves represent seed sizes of Dcore = 0, 2, 2.5 and 3nm, respectively. For the Drod = 6nm rod the black, red and green curves represent seed sizes of Dcore = 2.5, 3.5, and 4.5nm, respectively. Inset: zoom in on the low energy range. havior, we rst note that the trion formation

crease of the band gap was mainly attributed

rates are given in terms of a product of the av-

to the change of the highest hole energy, which

erage square Coulomb coupling and the trion

is localized on the seed.

density of states (cf., Eq. (4)).

47

The former

In the right panels of Fig. 3 we plot the DOTS

shows a monotonic core-size dependence and

for two additional cases. The bottom panel as-

a weak energy dependence (not shown).

sumes that the DOS of the holes for all seed

30,36

Therefore, the isosbestic point must reect a

sizes is identical so the observed change in the

non-monotonic behavior of the DOTS with

DOTS reects the dierences in the DOS of the

Dcore ,

which is indeed the case, as shown by

electrons (i.e., the portion of the DOS above

the left panels of Fig. 3. We nd that both + ρ− T (ε) and ρT (ε) exhibit an isosbestic point, while that for the negative DOTS (circled in

the Fermi energy). Clearly, all three sizes give

black) is correlated with the isosbestic point of Γ− e (ε).

tive DOTS originates from the hole DOS (i.e.,

Since the DOTS is given in terms of a triple

ergy). The upper right panel of Fig. 3 shows the

convolution of the DOS, we also plot the DOS

DOTS for the case where the only dierence in

as insets to each panel in Fig. 3.

the DOS for the three values of

the same DOTS. Therefore, one can conclude that the isosbestic point observed in the negathe portion of the DOS below the Fermi en-

We focus

Dcore is assumed

on the energy regime about the Fermi energy,

to be the aforementioned shoulder near the va-

which is assumed at the top of the valence band

lence band edge.

and is set as the origin of the energy axis. Per-

account for the isosbestic point. Thus, we con-

haps the most notable eect of the seed size

clude that the crossover behavior observed in

on the DOS is the appearance of a shoulder

the MEG eciency can be attributed to the in-

near the valence band edge as

Dcore

Clearly, this is sucient to

increases.

crease in the DOS near the valence band edge

This shoulder reects the increasing DOS on

and at the same time the rather insignicant

the seed due to volume eects. In addition, the

decrease in the band gap energy.

DOS shows a systematic shift in the position

In conclusion, we have extended our stochas-

of the valence and conduction peaks resulting

tic approach to calculate the MEG eciency in

from small changes in the band gap as

Dcore

in-

systems containing thousands of atoms within

creases (note that we scale the energy by the

a master equation formalism. Rather than cal-

band gap

Eg

for each system).

Here, the de-

culating all possible excitations and the cor-

ACS Paragon Plus Environment 6

Page 7 of 10

The Journal of Physical Chemistry Letters 12

D=4nm

D=4nm

10

ρDOTS(ε)

2

10

ρ(ε)

ρDOTS(ε)

10

1

8

10

0 -3 -2 -1 0

6

10

1 2 ε / Eg

3

4

10

10

8

10

5 6

10

12

10

D=6nm

10

ρDOTS(ε)

4

10

ρ(ε)

ρDOTS(ε)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

2

8

10

0 -3 -2 -1 0

6

10

1 2 ε / Eg

3

4

D=4nm

10

10

8

10

5 6

-3

-2

-1

0

1

2

3

10 -3

-2

-1

0

ε / Eg

1

2

3

ε / Eg

Drod = 4nm and 6nm nanorods (upper and lower panels, respectively) and the articial DOTS for a Drod = 4nm diameter (right panels) as described in the text. Insets show the corresponding DOS for each Drod . For the Drod = 4nm rod the blue, black, red and green curves represent seed sizes of Dcore = 0, 2, 2.5 and 3nm, respectively. For the 6nm rod the black, red and green curves represent seed sizes of Dcore = 2.5, 3.5, and 4.5nm, respectively. Figure 3: The DOTS for

responding MEG rates, we proposed to sam-

Grants No.

ple the electron-hole pairs and for each pair,

spectively.

obtain the MEG rates (sum of the positive

support of the US-Israel Bi-National Science

and negative trion formation rates) using our

Foundation.

recently developed expeditious stochastic ap-

Center for Re-Dening Photovoltaic Eciency

proach.

Application of the approach to quasi-

Through Molecule Scale Control, an Energy

type II CdSe/CdS seeded nanorods uncovered

Frontier Research Center funded by the U.S.

an interesting and unexpected behavior. Above

Department of Energy, Oce of Science, Oce

30

the onset of MEG (around

Dcore

increases as

3Eg )

the eciency

1020/10 and No.

611/11, re-

R. B. and D. N. acknowledge the E. R. would like to thank the

of Basic Energy Sciences under Award Number DE-SC0001085.

decreases, as expected. How-

ever, below this onset, the opposite is true and

References

larger seeds show higher MEG eciencies, contrary to the situation in spherical nanocrystals.

(1) Beard,

This anomalous behavior was caused by the

M.

C.;

A.

G.;

Luther,

J.

M.;

marked increase in DOS near the valence band

Hanna,

Dcore drops and the relative insensitivity of Eg to Dcore (due to the quasi-type II band

Hughes,

alignment) .

Quantum Dots To Impact Ionization in

edge as

paring Bulk

Manipulating the density of states without af-

M.

Midgett,

B.

C.; K.;

Multiple

Nozik, Exciton

Semiconductors:

A.

J.

Com-

Generation

Implications

in for

fecting the band gap can be used to control

Enhancement of Solar Energy Conversion.

MEG eciencies in other semiconducting het-

Nano Lett.

erostructures. Of particular relevance are type II nanorods

49

(2) Klimov,

which are expected to yield larger

of

These, and

vestigation.

Detailed-Balance

Limits

of

Power

Nanocrystal-

Multiplication.

2006, 89, 123118.

Appl.

Phys.

(3) Christensen, O. Quantum Eciency of In-

Acknowledgement thank

Carrier

Lett.

other possible scenarios are currently under in-

fully

I.

Quantum-Dot Solar Cells in the Presence

internal elds across the interface between the

33,34

V.

Conversion

MEG eciencies, perhaps also due to the large two semiconducting materials.

2010, 10, 30193027.

the

ternal Photoelectric Eect in Silicon and

R. B. and E. R. grate-

Israel

Science

Germanium.

Foundation,

689695.

ACS Paragon Plus Environment 7

J.

Appl.

Phys.

1976,

47,

The Journal of Physical Chemistry Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(4) Kolodinski,

H.;

Ellingson, R. J.; Nozik, A. J. PbTe Col-

Quan-

loidal Nanocrystals: Synthesis, Character-

tum Eciencies Exceeding Unity Due to

ization, and Multiple Exciton Generation.

Impact Ionization in Silicon Solar-Cells.

J. Am. Chem. Soc.

Wittchen,

S.; T.;

Werner,

Queisser,

H.

J.

1993, 63, 24052407.

Appl. Phys. Lett.

(5) Padilha, L. A.; berg, R. L.;

J.

Page 8 of 10

Stewart, J. T.;

Bae, W. K.;

(12) Schaller, R. D.;

Sand-

2006, 128, 32413247.

Sykora, M.;

Jeong, S.;

Klimov, V. I. High-Eciency Carrier Mul-

Koh, W.-K.;

tiplication

and

Pietryga, J. M.; Klimov, V. I. Carrier Mul-

aration

Semiconductor

tiplication

Studied via Time-Resolved Photolumines-

in

Semiconductor

Nanocrys-

in

tals: Inuence of Size, Shape, and Com-

cence.

position. Acc. Chem. Res.

2533225338.

2013, 46, 1261

1269.

J.

Phys.

(13) Schaller,

(6) Shabaev, A.; Hellberg, C. S.; Efros, A. L.

Ultrafast

Chem.

R.

Pietryga,

J.

Sep-

Nanocrystals

2006,

B

D.;

M.;

Charge

110,

Sykora,

Klimov,

V.

M.;

I.

Seven

Eciency of Multiexciton Generation in

Excitons at a Cost of One:

Colloidal

the Limits for Conversion Eciency of

Res.

Nanostructures.

2013, 46, 12421251.

Acc.

Chem.

Redening

Photons into Charge Carriers. Nano Lett.

2006, 6, 424429.

(7) Schaller, R. D.; Klimov, V. I. High Efciency Carrier Multiplication in PbSe

(14) Beard, M. C.; Knutsen, K. P.; Yu, P. R.;

Nanocrystals: Implications for Solar En-

Luther, J. M.; Song, Q.; Metzger, W. K.;

ergy Conversion. Phys. Rev. Lett.

Ellingson, R. J.;

2004,

92, 186601.

(8) Schaller,

R.

D.;

Agranovich,

V.

Nanocrystals. Nano Lett.

M.;

tiplication Through Direct Photogenera-

(15) Schaller,

tion of Multi-Excitons via Virtual Single-

Klimov,

Exciton States. Nat. Phys.

in

2005,

1, 189

194.

J.

C.;

Beard, M. C.;

R. V.

InAs

D.; I.

Pietryga,

Carrier

Nanocrystal

J.

M.;

Multiplication

Quantum

Dots

Yu,

P.

R.;

Micic,

O.

(16) Trinh, M. T.; Polak, L.; Schins, J. M.; Houtepen,

tion in Colloidal PbSe and PbS Quantum

Maikov,

Dots. Nano Lett.

Midgett,

R.

2005, 5, 865871.

D.;

Petruska,

M.

Beard,

A.;

on

ture on Carrier Multiplication Eciency:

J.

G.

G.; C.;

E.;

Beard,

Ahrenkiel, Yu,

P.

R.;

M. S.

C.; P.;

Micic,

et

Grinbom,

G.;

Luther, al.

Group

J.

Qazilbash,

Nor-

Midgett,

JohnO.

G.;

C.

Dierent

(17) Stewart, J.

R.;

I.;

A.

M.

Vaxenburg, J.

Anomalous

M.; Inde-

IV-VI

Quantum

2011,

11,

16231629.

2005, 87,

253102.

A.

J.;

Dot Architectures. Nano Lett.

Comparative Study of PbSe and CdSe

(11) Murphy,

A.

pendence of Multiple Exciton Generation

Klimov, V. I. Eect of Electronic Struc-

Nanocrystals. Appl. Phys. Lett.

7,

34693476.

I.;

Highly Ecient Multiple Exciton Genera-

(10) Schaller,

2007,

Conservation Limit. Nano Lett.

John-

Nozik, A. J.; Shabaev, A.; Efros, A. L.

son,

7, 2506

with an Onset Dened by the Energy

(9) Ellingson, R. J.;

man,

2007,

2512.

Klimov, V. I. High-Eciency Carrier Mul-

son,

Nozik, A. J. Multiple

Exciton Generation in Colloidal Silicon

T.;

M. A.

M.; G.;

Padilha, Pietryga, Luther,

L.

A.;

J.

M.;

J.

M.;

Beard, M. C.; Nozik, A. J.; Klimov, V. I.

I.;

Comparison

of

Carrier

Multiplication

Yields in PbS and PbSe Nanocrystals:

ACS Paragon Plus Environment 8

Page 9 of 10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry Letters The

Role

of

Competing

Processes. Nano Lett.

Energy-Loss

(26) Fischer,

2012, 12, 622628.

S.

A.;

born, C. M.;

2006, 6, 21912195.

citon Scattering Model for Carrier Multiplication

ization in Multiple Exciton Generation in PbSe Nanocrystals. Phys. Rev. B

(28) Allan, G.;

Dependence of the Multiple Exciton Generation Rate in CdSe Quantum Dots. ACS

8,

Nano

44884492.

tic Calculation of Multiexciton Genera-

Prezhdo, O. V. Generation of Multiple

tion Rates in Semiconductor Nanocrys-

Quantum

tals. Nano Lett.

First-

Principles Calculations on Small PbSe and CdSe Clusters. J. Phys. Chem. C

2008,

in

Semiconductor

Principles

Mechanisms in Nanocrystalline and Bulk

Exciton

Forms of PbSe and PbS. Phys. Rev. B

2012, 86, 165319.

Generation

Nanocrystals:

Calculations

on

First-

Small

(32) Rabani,

PbSe

2009,

Clusterse. J. Phys. Chem. C

Baer,

R.

Generation

Theory in

of

Mul-

Semiconductor

Nanocrystals. Chem. Phys. Lett.

2010,

496, 227235.

J.

(33) Gabor, N. M.; Zhong, Z. H.; Bosnick, K.;

Bonn, M. Carrier Multiplication

Park, J.; McEuen, P. L. Extremely E-

in Bulk and Nanocrystalline Semiconduc-

cient Multiple Electron-Hole Pair Gener-

tors: Mechanism, Eciency, and Interest

ation in Carbon Nanotube Photodiodes.

for Solar Cells. Phys. Rev. B

Science

J. H.;

C.;

E.;

tiexciton

113,

1261712621. (24) Delerue,

Piryatinski, A. Nu-

merical Analysis of Carrier Multiplication

(23) Isborn, C. M.; Prezhdo, O. V. Charging Multiple

2012, 12, 21232128.

(31) Velizhanin, K. A.;

112, 1829118294.

Quenches

2011, 5, 25032511.

(30) Baer, R.; Rabani, E. Expeditious Stochas-

(22) Isborn, C. M.; Kilina, S. V.; Li, X. S.;

Dots by Direct Photoexcitation:

2011, 5, 73187323.

(29) Lin, Z.; Franceschetti, A.; Lusk, M. T. Size

tiexciton Generation Rates in CdSe and

CdSe

133,

The Case of Sn Quantum

Dots. Acs Nano

(21) Rabani, E.; Baer, R. Distribution of Mul-

and

2010,

Delerue, C. Optimization of

Solar Cells:

6,

28562863.

2008,

Nanocrys-

Carrier Multiplication for More Ecient

Multiexciton Generation by a Single Pho-

2006,

Semiconductor

084508.

(20) Shabaev, A.; Efros, A. L.; Nozik, A. J. ton in Nanocrystals. Nano Lett.

in

tals: Theory. J. Chem. Phys.

2006,

73, 205423.

PbSe

Prezhdo, O. V. Multiple

(27) Piryatinski, A.; Velizhanin, K. A. An Ex-

(19) Allan, G.; Delerue, C. Role of Impact Ion-

in

Is-

2010, 1, 232237.

Chem. Lett.

Multiplication in PbSe Quantum Dots.

Excitons

B.;

A High-Level, Ab Initio Study. J. Phys.

Impact Ionization Can Explain Carrier

InAs Nanocrystals. Nano Lett.

A.

Exciton Generation in Small Si Clusters:

(18) Franceschetti, A.; An, J. M.; Zunger, A.

Nano Lett.

Madrid,

Allan,

G.;

Pijpers,

2010,

81,

125306. (25) Witzel,

2009, 325, 13671371.

(34) Baer, R.; Rabani, E. Can Impact ExciW.

M.;

Shabaev,

A.;

Hell-

tation Explain Ecient Carrier Multipli-

berg, C. S.; Jacobs, V. L.; Efros, A. L.

cation in Carbon Nanotube Photodiodes?

Quantum Simulation of Multiple-Exciton

Nano Lett.

Generation in a Nanocrystal by a Single Photon. Phys. Rev. Lett.

2010,

2010, 10, 32773282.

(35) Sandberg, R. L.; Padilha, L. A.; Qazil-

105,

bash, M. M.; Bae, W. K.; Schaller, R. D.;

137401.

Pietryga, J. M.; Stevens, M. J.; Baek, B.;

ACS Paragon Plus Environment 9

The Journal of Physical Chemistry Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nam, S. W.;

Klimov, V. I. Multiexci-

CdSe/CdS

ton Dynamics in Infrared-Emitting Col-

Seeded

conducting Nanowire Single-Photon De-

(36) Baer,

R.;

2012, 6, 95329540.

Rabani,

E.

Generation

Rates

Nanorods

Are

Independent.

Chem. Phys.

Length

Bae, W. K.;

tals:

State of the Art and Perspectives.

Acc. Chem. Res.

Sand-

Pair-Potentials for CdS and ZnS Crystals. J. Chem. Phys.

(45) Grünwald,

Ratio Dependence of Auger Recombina13,

2012, 136, 234111.

M.;

Alivisatos,

tion and Carrier Multiplication in PbSe

A.

Lutker, P.;

K.;

Rabani,

E.;

Geissler, P. L. Metastability in Pressure-

1092

Induced

1099.

Structural

Transformations

of

CdSe/ZnS Core/Shell Nanocrystals. Nano

(38) Karki, K. J.; Ma, F.; Zheng, K.; Zidek, K.; Mousa,

A.;

Abdellah,

M.

A.;

Lett.

Mess-

Pullerits,

T.

Multiple

Exciton

Genera-

of a Dielectric Medium. J. Chem. Phys.

1999, 110, 5355.

2013, 3 .

(47) Eshet,

(39) Klimov, V. I. Multicarrier Interactions in

H.;

Grunwald,

Quasi-Type-II?

tion and Carrier Multiplication. Ann. Rev.

2014, 5, 285316.

(48) Neuhauser, Resolution.

ductor Quantum Dots: An ab initio Per-

2008, 460, 1

D. J.

(49) Luther, J. M.; H.;

Tan,

T.

Alivisatos,

Z.;

Nanorods

Prezhdo, O. V.; Ruan, X. L. Shape and Dynamics

in

Spherical

Chem. Soc.

2011, 115, 1140011406.

(42) Carbone, L.; Nobile, C.; De Giorgi, M.; F.

Hytch,

D.; M.;

Franchini,

I.

Morello,

G.;

Pompa,

P.;

Snoeck,

E.;

Fiore,

A.;

Synthesis

and

R.

et

al.

13,

Bound-State

Eigenfunc-

Chem.

1990,

Phys.

93,

A. and

Zheng, H.; P.

Sadtler, B.;

Synthesis

Other

Ionic

of

PbS

Nanocrys-

tial Cation Exchange Reactions. J. Am.

and

Elongated CdSe Quantum Dots. J. Phys. Chem. C

Type-I or

2013,

tals of Complex Morphology by Sequen-

Temperature Dependence of Hot Carrier Relaxation

Lett.

26112616.

9. Bao,

Nano

tions from Wave-Packets - Time-Energy

Electron-Phonon Bottleneck in Semicon-

L.;

E.

58805885.

(40) Prezhdo, O. V. Multiple Excitons and the

L.

Rabani,

Core/Shell Seeded Nanorods:

to the Phenomena of Auger Recombina-

(41) Chen,

M.;

The Electronic Structure of CdSe/CdS

Semiconductor Nanocrystals in Relation

spective. Chem. Phys. Lett.

Berne, B. J.;

Nanocrystals in the Absence and Presence

Calculation of the Yield Based on Pump-

Conden. Mat. Phys.

Hetenyi, B.;

Brus, L. E. Electronic Properties of CdSe

tion in Nanocrystals Revisited: Consistent Probe Spectroscopy. Sci. Rep.

2013, 13, 13671372.

(46) Rabani, E.;

ing, M. E.; Wallenberg, L. R.; Yartsev, A.;

Sala,

2013, 46, 13871396.

(44) Grünwald, M.; Zayak, A.; Neaton, J. B.;

Koh, W.-K.;

2013,

Lett.

J.

Klimov, V. I. Aspect

Nanorods. Nano Lett.

Nano

a

Geissler, P. L.; Rabani, E. Transferable

Stewart, J. T.;

Pietryga, J. M.;

Approach.

by

loidal Branched Semiconductor Nanocrys-

CdSe

2013, 138, 051102.

(37) Padilha, L. A.; berg, R. L.;

in

Growth

Prepared

(43) Li, H.; Kanaras, A. G.; Manna, L. Col-

Communication:

Biexciton

Nanorods

2007, 7, 29422950,.

loidal Nanostructures Probed by a Supertector. ACS NANO

Page 10 of 10

Micrometer-Scale Assembly of Colloidal

ACS Paragon Plus Environment 10

2009, 131, 1685116857.