TSpace Research Repository tspace.library.utoronto.ca

Remote Molecular Doping of Colloidal Quantum Dot Photovoltaics

Ahmad R. Kirmani, Amirreza Kiani, Marcel M. Said,

Oleksandr Voznyy, Nimer Wehbe, Grant Walters, Stephen Barlow, Edward H. Sargent, Seth R. Marder, and Aram Amassian

Version Post-Print/Accepted Manuscript

Citation (published version)

Remote Molecular Doping of Colloidal Quantum Dot Photovoltaics. Ahmad R. Kirmani, Amirreza Kiani, Marcel M. Said, Oleksandr Voznyy, Nimer Wehbe, Grant Walters, Stephen Barlow, Edward H. Sargent, Seth R. Marder, and Aram Amassian. ACS Energy Letters 2016 1 (5), 922-930. DOI: 10.1021/acsenergylett.6b00429

Publisher’s Statement This document is the Accepted Manuscript version of a Published Work that appeared in final form in ACS Energy Letters, copyright ©American Chemical Society after peer review and technical editing by the publisher. To access the final edited and published work see https://dx.doi.org/10.1021/acsenergylett.6b00429.

How to cite TSpace items

Always cite the published version, so the author(s) will receive recognition through services that track

citation counts, e.g. Scopus. If you need to cite the page number of the author manuscript from TSpace because you cannot access the published version, then cite the TSpace version in addition to the published

version using the permanent URI (handle) found on the record page.

This article was made openly accessible by U of T Faculty. Please tell us how this access benefits you. Your story matters.

1

A General Remote Molecular Doping Platform for Colloidal Quantum Dot Solids

Ahmad R. Kirmani, Amirreza Kiani, Marcel M. Said, Oleksandr Voznyy, Nimer Wehbe, Grant

Walters, Stephen Barlow, Edward H. Sargent, Seth R. Marder, Aram Amassian*

A.R. Kirmani, Dr. N. Wehbe, Prof. A. Amassian

King Abdullah University of Science and Technology (KAUST), KAUST Solar Center (KSC),

and, Physical Science and Engineering Division, Thuwal, 23955-6900, Saudi Arabia

*E-mail: aram.amassian@kaust.edu.sa

A. Kiani, Dr. O. Voznyy, Grant Walters, Prof. E.H. Sargent

Department of Electrical and Computer Engineering, University of Toronto, Toronto, Ontario

M5S 3G4, Canada.

M. M. Said, Dr. S. J. Barlow, Prof. S. R. Marder

School of Chemistry and Biochemistry and Center for Organic Photonics and Electronics,

Georgia Institute of Technology, 901 Atlantic Drive, Atlanta, GA, 30332-0400, United States

Keywords: Colloidal quantum dots, molecular doping, trap-assisted recombination,

photoemission spectroscopy, remote electron transfer

2

Abstract

In recent years the field of colloidal quantum dot (CQD) photovoltaics has develped rapidly

thanks to novel device architectures and robust surface trap passivation schemes. Achieving

controlled net doping remains an important unsolved challenge for this field. Herein, we present

a general and facile solution processed molecular doping platform for CQD solids employing a

library of metal-organic complexes. Shallow ionization energy and high electron affinity

complexes are respectively shown to n- and p-dope the CQD solids. Employing photoemission

spectroscopy, we identify two doping concentration regimes: lower concentrations lead to

efficient doping (electron removal/addition), while higher concentrations result in surface

dipoles. We demonstrate the obvious advantage of the doping recipe in solar cells by p-doping

the CQD absorber layer. Utilizing the lower concentration regime, we remove mid-gap

electrons and suppress trap-assisted recombination leading to a 30% enhancement in the power

conversion efficiency relative to undoped cells.

Colloidal quantum dot (CQD) optoelectronics and photovoltaics have seen unabated

advancement ever since the demonstration of facile and scalable routes to nanocrystal synthesis

in the early 1990s.[1-3] The possibility of achieving >100% external quantum efficiency (EQE),

coupled with bandgap tunability, make CQDs a promising prospect for modern-day

photovoltaics.[4-8] Recent demonstrations of air-stable, high power conversion efficiency (PCE)

CQD solar cells are compatible with roll-to-roll manufacturing techniques such as spray-

coating and have brought this technology a significant step closer to commercial

implementation.[9-12] CQD photovoltaics have recently shown both efficient device

architectures and fruitful deployment of robust passivation schemes.[2-3, 5, 7, 9-10]

Charge transport in CQD solids used in photovoltaics relies on hopping.[13-15] However, the

large suface-to-volume ratio of QDs leads to surface-related charge trapping that limits device

3

performance.[16-25] Surface ligand exchange through chemical routes has traditionally served

the important purpose of passivating surface traps and bringing the QDs closer for better

electronic coupling.[26] Among the ligand exchange routes, solid-state ligand exchange is the

most commonly implemented approach for passivating QDs for solar cell applications.[20, 27-28]

However, achieving controlled passivation and net doping of CQD solids without resorting to

or being limited by chemical routes remains an important and mostly unsolved challenge for

this field. The opportunity to tune the net doping without the burden of redesigning the

processing and chemistry of CQD film fabrication would greatly enhance our ability to tailor

the electronic properties of CQD solids. Remote doping is believed to be the most versatile way

to achieve this in that it does not interfere with the QD surface chemistry, interdot spacing, and

film order at the nanoscale, and that it is free from steric hindrance and other chemical binding

restrictions, meaning that it can complement existing solution and solid-state ligand-exchange

schemes. The observation of remote electron transfer by Shim and Guyot-Sionnest in CdSe

QDs opened the door to realizing successful doping schemes in QDs without resorting to

chemical modification of QD surfaces.[29] In a later study, deposition of potassium on the

surface of CdSe QD films led to significant increase in conductivity through doping.[30]

Recently, Konstantatos et al. have demonstrated remote passivation of in-gap trap states by

introducing ZnO nanocrystals into the CQD solid, leading to suppression of trap-assisted carrier

recombination.[31] We took the view that a single-step, post-deposition remote doping protocol

that can treat CQD solids could be a robust and highly desirable platform for enabling the

controlled net doping of CQD solids for a wide range of applications.

Here, we demonstrate that a soluble molecular dopant with carefully selected energetics and

sufficiently small size can infitrate and dope a lead sulfide (PbS) CQD solid. The remote doping

is achieved in a single step without altering the deposition and exchange protocols of the CQD

solid, making this a highly versatile scheme. We consider a variety of doping scenarios by

4

employing a library of organometallic complexes with different energetics (electron affinity,

EA and ionization energy, IE). Deep EA complexes are found to shift the Fermi level of the

CQD solids towards the valence band edge, as evidenced by ultraviolet photoemission

spectroscopy (UPS), suggesting p-doping. Shallow IE complexes, on the other result in n-

doping of the solids, however, the shift of the Fermi level for this case is minor. Employing

UPS, we identify two doping concentration regimes: lower concentrations lead to efficient

electron removal/addition, while higher concentrations result in significant surface-dipole

formation. The facile procedure involves soaking the CQD solid into the dopant solution. The

benefits of this solid-state remote doping scheme are demonstrated in the context of depleted

heterojunction CQD solar cells, where p-doping of the CQD absorber layer effectively removes

mid-gap electrons and suppress trap-assisted recombination rendering the absorber layer

relatively intrinsic, leading to a 30% enhancement in power conversion efficiency from 6% to

7.8%. Our demonstration is a significant leap towards facile and scalable control of carrier

concentration and in-gap trap states in CQD solids and, given the vast number of available

organic and metal-organic dopants, should provide new and facile routes to tuning the

properties of CQD solids for photovoltaics, optoelectronics and other applications.

We begin by presenting the various doping scenariosexamined in the study and the associated

changes to the CQD band structures (Fig. 1). We chose two organometallic complexes having

deep EA (p-dopants) and two complexes with shallow IE (n-dopants), with respect to the QD

energetics. Schematics represent the energetics of the molecules and the CQD solids and the

chemical structures of the various molecules involved. The doping procedure involved briefly

soaking the CQD layer in the dopant solution in acetonitrile (ACN). ACN was chosen as the

solvent based on previous results which suggest that ACN, being aprotic and a high dipole

moment solvent, is benign to the halide passivants on the QD surface.[20, 32] The standard protic

solvent, MeOH introduces in-gap trap states over prolonged exposure to QDs by leaching off

the surface chloride passivant. Nickel bis(1,2-bis(trifluoromethyl)ethane-1,2-dithiolene),

5

Ni(tfd)2, with an EA of ~5.6 eV is expected to withdraw electrons from the CQDs behaving as

a strong p-dopant (Fig 1a).[33] Molybdenum tris(1-phenyl-2-benzoyl-1,2-dithiolene), Mo(PhBz-

dt)3, on the other hand, has a shallower EA (4.8 eV) and should therefore act as a weaker p-

dopant (Fig. 1b). Ruthenium (pentamethylcyclopentadienyl)(mesitylene) dimer, (RuCp*mes)2

and pentamethylrhodocene dimer, (RhCp*Cp)2 and are known to behave as strong and weak n-

dopants (Fig. 1c, d), respectively owing to their shallow IE (2.5 eV and 3.3 eV, respectively).[34-

35]

Figure 1. General molecular doping platform for CQD solids. The doping scenarios examined

in this study are demonstrated. (a) Ni(tfd)2 is expected to behave as a strong p-dopant owing to

its deep EA, as compared to the band structure of CQDs. (b) Mo(PhBz-dt)3, with a shallower

EA, behaves as a weak p-dopant. (c) (RuCp*mes)2, having a very shallow IE, is expected to

serve as a strong n-dopant, while (d) (RhCp*Cp)2 is a comparatively weaker n-dopant. These

expectations are verified by UPS measurements which show relevant changes to the band

structures. Weak dopants show negligible shifts to the Fermi level while the strong dopants

cause major shifts to the Fermi level. Importantly, these Fermi level changes are associated with

corresponding shifts to the Pb 4f core levels, as evidenced by XPS (Fig. S1). Significant changes

to the vacuum level (dipole formation) are observed when the CQD solids are treated with high

concentrations of the dopants (1 mg mL-1).

We sought to probe band-structure changes that might give experimental evidence of electron

transfer. We used UPS, a direct probe to elucidate the changes to the electronic band structure

– the density of states below the Fermi level (EF) of a material – and widely used to study the

6

effects of changes to nanomaterial surfaces on their band structures.[19-20, 36-37] Miller et al. have

recently highlighted the challenges associated to UPS measurements on larger size QDs which

exhibit a small band gap, owing to the extremely low density of states at the valence band

maximum.[38] However, the CQDs considered in our work have a larger band gap as they are

small-size particles; hence the UPS results are expected to be reliable, commensurate with the

findings of Miller and co-workers.

Direct evidence of an electron transfer in doped MoS2 films has recently been reported by

Tarasov et al. where a shift in the EF was observed and ascribed to transfer of electrons/holes.[39]

This shift corresponded to a similar shift in the Mo core level peak measured by X-ray

photoelectron spectroscopy (XPS). QD films obtained via spin coating were treated with

various concentrations of the dopant solution to track changes to the band structure.

Corresponding to the schematics in Fig. 1 are the band strutures of the doped CQDs as measured

by UPS. Each set shows three scenarios: undoped CQD baselines, CQDs doped with low

concentration dopant solution and CQDs doped with high concentration dopant solution. The

effect of doping on the band structures can be broadly categorised into two regimes: smaller

doping concentrations lead to changes to the Fermi level, which moves towards the valence

band for p-doped films and towards the conduction band for n-doped films, while higher doping

concentrations also result in changes to the vacuum level. This is strictly true for the strong p-

and n-dopants (Fig. 1a, c), however, for the weak p- and n-dopants (Fig. 1b, d), these effects

are expectedly very mild and the band structures of the doped solids for these cases remain

almost unchanged. The observation of Pb 4f core level shifts from XPS corresponds to the

electron transfer as a result of doping (Fig. S1).

The fact that we observe effective p-doping of the CQD solids can be good news from the solar

cell perspective where the CQD solid is employed as the absorber layer. These absorber layers,

which form a heterojunction with an n-type metal oxide (for example, titania) have been

7

recently suggested to be slightly n-type.[40] We therefore took the view that a post-synthesis p-

doping step should make the absorber layer more intrinsic leading to extension of the depletion

region into the CQD solid, eventually enhancing the solar cell performance. Towards this end,

we chose the metal-organic complex demonstrated in Fig. 2a, molybdenum tris(1-

(trifluoroacetyl)-2-(trifluoromethyl)ethane-1,2-dithiolene) (Mo(tfd-COCF3)3.[35, 41] Mo(tfd-

COCF3)3 is a more soluble variant of molybdenum tris(1,2-bis(trifluoromethyl)ethane-1,2-

dithiolene), Mo(tfd)3, which has been successfully used to dope the hole transporting material

(HTM) N,N'-di-[(1-naphthyl)-N,N'-diphenyl]-1,1'-biphenyl-4,4' diamine, (α-NPD), via

coevaporation in ultra-high vacuum, giving current density enhancements of over five orders

of magnitude.[42] The high electron affinity (EA) (estimated to be >5.6 eV) makes it an

oxidizing agent and, therefore, the conductivity enhancement was attributed to compensation

of trap states and contribution of free carriers by hole-injection from the dopant. Encouraged

by the successful deployment of Mo(tfd)3 as a strong p-dopant for α-NPD, we chose the non-

symmetrical Mo(tfd-COCF3)3, having a similar EA but ca. 0.11 V more oxidizing than Mo(tfd)3

in solution[41]. The enhanced solubility of Mo(tfd-COCF3)3 makes it compatible with solution

processing and therefore better suited for scalable manufacturing of CQD solids and devices

thereof. The orbital energies for isolated Mo(tfd-COCF3)3 were calculated using density

functional theory (DFT) and are shown in Fig. S2.

8

Figure 2. (a) Chemical structure of Mo(tfd-COCF3)3 used for p-doping the CQD solid. (b)

Schematic of the energetics at the QD-dopant interface based on experimental data. The VBM

of the QD lies near the LUMO of the dopant molecule suggesting the possibility of electron

transfer. (c) DFT simulations suggest that the Mo-based molecule can indeed act as a remote

dopant for the PbS QDs. The theoretical proximity of the molecule’s LUMO (depicted by the

black arrow) with the VBM of the QDs opens the possibility of the depopulation of the mid-

gap or VB electrons, depending upon the dopant concentration, via remote doping.

We considered the scenario of Mo(tfd-COCF3)3 in close proximity to a PbS QD. The energetics

of Mo(tfd-COCF3)3 and a QD are outlined in Fig. 2b and suggest the possibility of remote

electron transfer from the QDs to the dopant. This possibility is also supported by DFT

simulation of the energetics at the interface of this system. The results are shown in Fig. 2c.

Hybrid ligand-exchanged QDs (involving a solid-state exchange with 3-mercaptopropionic acid

ligand, MPA)[43] were considered for these calculations. The density of states (DOS)

corresponding to Mo belong to the dopant. The results show a close proximity of the dopant’s

LUMO with the VBM of the QDs. This suggests the possibility that the dopant can extract mid-

gap or valence electrons from the QDs via remote electron transfer.

9

In Fig. 3a we show the electronic band structures of the QD solids for the various doping

scenarios. The band structures remain almost unchanged for lower doping concentrations, while

significant changes are observed for higher concentrations (UPS spectra are shown in Fig. S3).

Figure 3. Photoemission spectroscopy data for PbS QD films treated with solutions of varying

dopant concentrations. (a) VB edges (blue) and Fermi levels (black) were acquired using UPS

while the conduction band edges (red) were estimated by adding the optical band gap (1.3 eV)

to the VB energy. This is justified in the current case since the exciton binding energy (BE) for

PbS QDs is low (~0.09 eV). For low doping concentrations, the Fermi level is found to shift

moderately towards the VB indicating mid-gap electron transfer. Higher doping concentrations

however lead to valence electron transfer and surface dipole formation causing significant

lowering of the vacuum level and Fermi level relative to vacuum. The energies of the dipoles

formed (ED) are extracted from these results and shown in (b). The Pb 4f core level peak,

10

obtained from high resolution (XPS) is shown in (c) for the MPA baseline and the maximum

doping cases. The red component of the peak corresponds to the Pb bonded to S atoms in the

QD. The dashed blue component is the Pb bonded to the MPA ligand while the solid blue

component is the tiny amount of Pb-O. The peak shows a ca. 0.2 eV shift to lower binding

energy for the maximum doping concentration which is commensurate with the Fermi level

shift observed for that concentration due to electron transfer (EVB in (b)).

The total change in the work function of a doped QD film (compared to the undoped, MPA

baseline), denoted as , comprises the Fermi level shift with respect to the VBM owing to

remote electron transfer (EVB) and the shift in the vacuum level due to dipole formation on the

QD surface (ED). This allows us to determine the energies ED for the various doping scenarios

(Fig. 3b). We find that as the doping concentration increases, the surface dipole strengthens.

This can be expected since for 1.0 mg mL-1 doping concentration, each QD is surrounded by

ca. 12 dopant molecules, as found from XPS atomic quantification (see Table S1 and the

associated discussion). This drops to ca. 0.3 dopants per QD for 10-2 mg mL-1 dopant

concentration, which agrees with the negligible value of ED for this case (Fig. 3b). We propose

that the interface-dipole effects originate at the surface of each QD as a result of dopant

penetration through the nanoscale voids in the QD solid, rather than being localized at the

surface of the QD film; we do not observe any accumulation of the dopants on the film surface,

as evidenced from secondary ion mass spectrometry (SIMS), even for the largest concentrations

(Fig. S4). The large values for this regime also involve a significant EVB hinting towards

removal of valence electrons. For smaller concentrations, changes in are largely associated

with changes to EF from mid-gap electron transfer (Table S2); in this regime a significant shift

in EF relative to EVBM requires fewer dopants than when the Fermi level approaches the band

edge and so the resultant interface dipole is relatively small. We also observe a ca. 0.2 eV shift

of the Pb 4f core level peak (for 1.0 mg mL-1 doping concentration) towards lower binding

energy compared to an undoped QD solid, which can be ascribed to the downward shift of the

Fermi level associated with p-doping (Fig. 2c).[39] Core level shifts were observed for other

concentrations as well and agreed with the accompanying Fermi level shifts.

11

While photoelectron spectroscopies are near-surface measurements, we confirm the dopant

infiltrates the bulk of the CQD solid by secondary ion mass spectrometry (SIMS) measurements

(Fig. S4). Despite this infiltration, the dopant does not perturb the interdot spacing, as

demonstrated by grazing incidence small angle x-ray scattering (GISAXS) (Fig. S11). Also, no

observable change in the overall film thickness was observed following doping, confirming the

CQD solid is not swollen by infiltration of dopant molecules (Fig. S6).

Having experimentally established remote electron transfer from the QDs to Mo(tfd-COCF3)3

throughout the bulk of the CQD solid, we were interested in studying its effects on solar cell

performance. Based upon the experimental and computational insights provided in Figures 1

and 2, we simulate the solar cell performances for undoped, optimally doped and overdoped

CQD absorber layers. An untreated MPA film was modeled as n-type with a doping

concentration of 1×1016 cm-3.[15, 44] The optimally treated film was graded doped from 3×1016

cm-3 p-type in the last layer to 0 near the TiO2 interface, and overtreated films were considered

either as 3x1016 cm-3 throughout or graded doped from 1x1017 cm-3 to 0. The simulated J-V

curves are shown in Fig. 4b and suggest that doping leads to VOC improvement due to the

movement of the Fermi level closer to the band edge. Overdoping, however, collapses the

depletion region leading to loss of JSC and hence the overall device performance (for details,

see Fig. S7). While these simulations make certain assumptions and simplifications, they

indicate that remote molecular doping should influence the performance of CQD solar cells.

We turn our attention to incorporating the doped QD solids into solar cells. The doping strategy

adopted successfully is outlined in Scheme 1. The standard procedure of layer-by-layer (LbL)

fabrication of the CQD absorber film was followed with the difference that the final CQD film

was dipped in the dopant solution for an adjusted time before removal of the solution and

solvent washing. An alternative doping scheme wherein every layer in the LbL stack was

12

individually doped failed to show performance enhancement (Fig. S5 and Table S3), most likely

due to overdoping of the film.

Scheme 1. Procedure of molecular doping demonstrated in this study. (1) LbL deposition of

the QD absorber layer via spin coating involves sequential deposition of CQDs (capped with

oleic acid ligands) followed by solid-state ligand exchange with MPA and a subsequent

methanol (MeOH) washing step. This cycle was repeated typically ten times until the targeted

film thickness (300 nm) was achieved. (2) The film was soaked in the dopant solution

(dissolved in acetonitrile, ACN) for an optimized time duration before spinning off the excess

solution. This was followed by a rinsing step in ACN to wash off any residual dopant and dry

the sample (3) for use in subsequent solar cell fabrication steps.

In Fig. 4c we show the experimentally measured J-V curves for various doping concentrations.

The device parameters are summarized in Table 1. It is evident that the device performance

increases for low doping concentrations, whereas higher concentrations lead to performance

degradation.

13

Figure 4. (a) Schematic of the depleted heterojunction (DHJ) device architecture employed in

this study. (b) Simulated J-V curves showing an increase in overall performance enhancement

for solar cells made using optimally doped CQD films, assuming a graded doping architecture.

Overdoped films show a performance decrease. (c) J-V curves for solar cells made using CQD

films treated with various doping concentrations and treatment times. (d) Device parameters

(JSC, FF, PCE and VOC) are shown as a function of the doping concentration for a soaking time

of 3 mins. The data highlights that 10-2 mg mL-1 is the optimized doping concentration for

achieving best performing devices.

We observe device performance enhancement for solar cells soaked in low doping

concentrations (10-3 mg mL-1). Performance is found to increase with longer soaking times

(primarily, VOC and FF). However, an extended soaking (15 mins) leads to decrease in device

14

performance, an effect we ascribe to overdoping of the CQDs. Next, we test devices with higher

doping concentrations (10-2 mg mL-1) and achieve the maximum performance boost for 3 mins

soaking time. The PCE of 7.8% obtained for this doping condition is a 30% enhancement over

the undoped baseline. We observe that at this elevated doping concentration, overdoping occurs

at only 4 mins of soaking leading to performance degradation. Doping the absorber layers with

a high concentration (0.1 mg mL-1) leads to immediate overdoping of the CQD solid and hence

poorly performing solar cells. These experiments indicate that doping with an ACN solution

can be effectively carried out in a single step in contrast to the repetitive ligand exchange steps

involved in the QD solid fabrication. The soaking in dopant solution allows the dopants to

penetrate and diffuse into the entire QD film, as indicated by SIMS measurements (Fig. S4).

The optimized doping time and dopant solution concentration pair allows for the optimal dopant

concentration levels to be found for any given QD solid. This is highlighted in Fig. 4d where

the various device parameters are plotted as a function of doping concentration for a soaking

time of 3 mins.

The shunt resistance, Rshunt, for all the better-performing doped devices is higher relative to the

undoped controls. This gives evidence of suppression of trap-assisted carrier recombination.[31]

We propose that the large surface dipoles introduced in conditions of higher doping

concentration and the associated valence electron removal, as evidenced by UPS, lead together

to the formation of localized energy barriers to efficient charge transport, causing performance

degradation. In fact, we observe no change in the absorption of the QD solids (Fig. S12) for

nearly all doping conditions with the exception of 1.0 mg mL-1, which shows evidence of

exciton quenching.[45] A similar quenching is observed in the transient absorption

measurements for the highest doped solid (Fig. S13). These observations agree with the

scenario of valence electron removal at high doping concentrations, as discussed above. This is

also supported by a quenching of the photoluminescence signal at the exciton energy (see Fig.

15

S10). Hence, only in-gap electron removal through ultra low doping of the QD solid leads to

PV performance enhancements.

Table 1. Summary of the device parameters for the various doping scenarios.

Device Soaking time

(mins)

JSC

(mA/cm2)

VOC

(V)

FF

(%)

PCE

(%)

Rshunt

(Ω/cm2)

Rs

(Ω/cm2)

MPA baseline 19.0 0.600 53.3 6.0 1985 8.0

10-3

mg mL-1

3 21.5 0.600 51.4 6.6 3495 6.8

6 20.6 0.620 54.5 6.7 3407 5.5

9 20.5 0.630 60.0 7.6 4781 4.9

15 18.6 0.630 59.2 6.9 4206 5.1

10-2

mg mL-1

3 21.0 0.630 60.0 7.8 6286 5.1

4 18.0 0.610 54.5 5.9 2777 7.7

0.1 mg mL-1 3 15.2 0.550 52.7 4.5 5868 8.2

The absorber layer thickness in the solar cells reported above is ca. 300 nm which consists of a

ca. 250 nm depletion region.[15, 46] In general, doping of the depletion region would lead to a

decrease in the depletion width and hence in the solar cell performance. However, this would

not occur if the MPA-capped QD solid behaves as an n-type semiconductor. Indeed, as shown

recently, the Fermi level of the MPA-capped QD solid lies slightly above mid-gap,[40] and as

such the doping step makes the absorber layer more intrinsic via p-doping, instead of leading

to a depletion region collapse. We further tested this hypothesis by fabricating solar cells with

thin absorber layers (ca. 150 nm). Doping of these thin layers involved ~1 min soaking.

Interestingly, we found that these thin solar cells also showed performance enhancement (Fig.

S8, S9 and Table S4, S5). These observations reveal the following picture: an MPA-capped QD

layer is intrinsically n-type owing to in-gap trap states; p-doping the layer leads to removal of

16

these traps making the film more intrinsic. This helps us push the limits on the maximum PCE

that can be achieved with an MPA-based PbS CQD solar cell.

In summary, we have successfully demonstrated a facile, solution processing-enabled remote

doping strategy for PbS QD solar cells using a large EA metal-organic molecule which does

not require chemical bonding with the QD surface. The EA of the molecule is larger than the

ionization energy of the QDs, enabling remote electron transfer (as also suggested by DFT).

Employing UPS, we carried out an in-depth study of the effect of doping on the electronic band

structure of the QDs. Our study suggests the presence of two contrasting doping regimes – one

leading to in-gap electron transfer and the second resulting in valence-electron removal and

large surface dipoles. Exploiting the former – the low doping concentration regime – we

suppress trap-assisted carrier recombination, leading to ca. 30% increase in solar cell

performance. Our mechanistic study furthers the fundamental understanding on solution

processing-friendly remote molecular doping of CQD solids and presents the most versatile and

scalable approach yet for achieving controlled net doping of CQD solids.

Experimental Section

CQD synthesis: PbS CQDs were synthesized using a variation on a literature method,[47]

employing an in-synthesis halide treatment[43].

Dopant synthesis: Molybdenum tris[1-(trifluoromethylcarbonyl)-2-(trifluoromethyl)-ethane-

1,2-dithiolene] was prepared according to the literature[35].

Density Functional Theory Calculations: Simulations were performed using CP2K software

utilizing a mixed planewaves and molecular orbitals basis[48]. Goedecker–Teter–Hutter

pseudopotentials were employed with a 300 Ry grid cutoff. The dopant molecule was simulated

using the B3LYP exchange-correlation functional as it is known to better reproduce the

bandgaps and electron affinities, which, we find, allow for a trap-free p-doping of the CQDs. A

17

3 nm PbS CQD was modeled in a (40 Å)3 unit cell using a less computationally expensive PBE

functional to confirm the charge transfer from the CQD to the dopant.

Photoemission Spectroscopy: XPS measurements were carried out in an ultrahigh vacuum

(UHV) Omicron chamber equipped with a SPHERA U7 hemispherical energy analyzer,

employing X-ray photons having an incident kinetic energy of 1486.6 eV from a

monochromated Al K α X-ray source with a total energy resolution of 0.1 eV. The chamber

base pressure for these measurements was < 5 × 10-9 mbar.

For the UPS measurements, the UHV base pressure was maintained below 8 × 10-9 mbar. The

photon line width was ca. 250 eV and the minimum spot size ca. 1 mm. He I photons (21.2 eV)

were used to acquire the spectra at normal emission. The photoelectrons were collected by the

SPHERA U7 hemispherical energy analyzer with a 7 channel MCD detector, in Constant

Analyzer Energy (CAE) mode. The BE values shown with 10 meV precision should be rounded

to the nearest 100 meV value in accordance with the overall energy resolution.

Device fabrication: CQD films were deposited on a TiO2 nanoparticle-based electrode on ITO-

coated glass substrate using layer-by-layer deposition approach. Firstly, a solution of 50 mg

mL-1 quantum dots in octane was spin-coated under ambient condition at 2500 rpm and then

followed by soaking in 1% 3-mercaptopropionic acid in methanol (v/v) for three seconds. The

layer was then rinsed twice with methanol to remove the exchanged oleic acid. The process

resulted in a layer with the thickness of approximately 30 nm. The process was repeated 10-

12 times until the desired thickness was achieved. The CQD film was then doped, as outlined

in the main text, by soaking the entire film in Mo(tfd-COCF3)3-containing acetonitrile solution.

The top electrode was then deposited using thermal evaporation, which comprised of

40 nm MoO3 and 120 nm gold. The top electrode was deposited at the rate of

0.2 Å/s for MoO3 and 1 Å/s for Au, at the pressure of 1 × 10-6 mbar.

J – V characterization: J−V characterization was performed using a Keithley 2400 source-

meter at ambient temperature, with the device in a constantly purged nitrogen environment. The

18

solar spectrum at AM1.5 was simulated to within class A specifications (less than 25% spectral

mismatch) with a xenon lamp and filters (ScienceTech; measured intensity of 100 mW

cm−2). The intensity of the source was calibrated using a Melles–Griot broadband powermeter

and a Thorlabs broadband powermeter through a circular 0.049 cm2 aperture at the position of

the device. This was confirmed with a calibrated reference solar cell (Newport, Inc.).

Secondary Ion Mass Spectrometry measurements: SIMS experiments were performed on a

Dynamic SIMS instrument from Hiden analytical company (Warrington-UK) operated under

ultra-high vacuum conditions, typically 10-9 mbar. The Dynamic SIMS is equipped with a gas

source allowing for both argon or oxygen ion beams to be employed. Throughout the sputtering

process, the selected ions ascribed C, O, F, Si, S, and Pb were sequentially collected using a

MAXIM spectrometer equipped with a quadrupole analyser. Ions are collected from the sample

by a shaped extraction field and energy filtered using a parallel plate system, with the energy

resolution matched to that of the quadrupole analyser. After passing through a triple filter

system, detected ions are measured using a pulse counting detector having a 4 keV post-

acceleration potential to increase further the detection efficiency at high masses.

Prior to acquiring mass spectra and depth profiling curves, the experimental conditions

including the primary ion type, energy and current of the sputtering beam were first optimized.

The major challenge for detecting molybdenum complex doped nanoparticles is, on one hand,

to maximize the detection limit of the dopant and, on the other hand, to reduce the fragmentation

effect leading to the formation of small hydrocarbon fragments. Therefore, the inert argon beam

at energy as low as 700 eV was chosen to conduct the measurements. The raster of the sputtered

area is estimated to be 500 × 500 µm2. In order to avoid the edge effect during depth profiling

experiments, it is necessary to acquire data from a small area located in the middle of the eroded

region. Using an adequate electronic gating, the acquisition area from which the depth profiling

data were extracted was approximately 50 × 50 µm2. This condition is not required for

collecting mass spectra where both rastered areas are the same. The conversion of the sputtering

19

time to sputtering depth scale was carried out by measuring the depth of the crater generated at

the end of the depth profiling experiment using a stylus profiler from Veeco company.

Absorption measurements: Absorption measurements were carried out using a Agilent Cary

5000 UV-Vis-NIR instrument equipped with PMT (UV-Vis) and PbS (NIR) detectors. The

samples were positioned using the solid sample holder accessory. Spectra were collected

between 400 nm and 1200 nm using a spectral bandwidth of 2 nm and a scan rate of 600

nm/min. All measurements were made in double beam mode, using reduced slit height and

baseline correction.

Photoluminescence measurements: PL measurements were carried out on a confocal micro-

Raman system (Horiba Jobin Yvon Aramis), using a 785 nm laser as the excitation source. A

100× objective lens with a numerical aperture (N.A.) of 0.90 was used to focus the laser beam

and collect scattered light. The exposure time was ca. 90 s for all spectra (3 scans, ca. 30s per

scan). CQD films were spin coated on bare soda-lime glass cleaned via ultrasonication

sequentially in acetone, isopropanol and ethanol. 2 layers of QDs were deposited leading to

very smooth and shiny films of total thickness ~60 nm.

VASE measurements: An M-2000XI, J. A. Woollam Co., Inc. ellipsometer (400 – 1700 cm-1)

was used to study the variation in film thickness following doping of the QD film coated on a

thermally oxidized Si substrate. The spectra were obtained at incidence angles in the range 45-

75 with discrete increments of 5. The film properties were modelled assuming a B-Spline

dispersion relation in the absorption region using the EASE and WVASE32 software packages

from J. A. Woollam Co., Inc.

GISAXS. Measurements were carried out at the D-line of the Cornell High Energy Synchrotron

Source (CHESS) (Cornell University). A beam with a wavelength of 1.15 Å was used, obtained

from a wide bandpass (1.47%) double-bounce multilayer monochromator. The angle of

incidence was varied discretely from 0.04 to 0.25 with respect to the plane of the substrate.

20

Optoelectronic device simulations. Simulations were performed in 1D using the SCAPS 3.0.0.1

software[49], and parameters from previous work[15].

Acknowledgements

The authors thank Yadong Zhang, Georgia Institute of Technology, for the chemical synthesis

of the metal-organic complex employed in this study, Dr. Omar El Tall of the Analytical Core

Laboratory (KAUST) for his assistance with the absorption measurements and Dr. Yang Yang

of the Advanced Nanofabrication, Imaging and Characterization Core Lab (KAUST) for his

help with PL measurements. We also thank Prof. Gerasimos Konstantatos, ICFO, Spain for

fruitful discussions. The authors acknowledge the use of the D1 beam line at the Cornell High

Energy Synchrotron Source supported by the National Science Foundation (NSF DMR-

0225180) and NIH-NIGMS. The work at Georgia Institute of Technology was supported by the

Office of Naval Research (N00014-14-1-0126).

Received: ((will be filled in by the editorial staff))

Revised: ((will be filled in by the editorial staff))

Published online: ((will be filled in by the editorial staff))

[1] C. B. Murray, D. J. Norris, M. G. Bawendi, Journal of the American Chemical Society

1993, 115, 8706.

[2] Y. Yin, A. P. Alivisatos, Nature 2005, 437, 664.

[3] M. V. Kovalenko, L. Manna, A. Cabot, Z. Hens, D. V. Talapin, C. R. Kagan, V. I.

Klimov, A. L. Rogach, P. Reiss, D. J. Milliron, P. Guyot-Sionnnest, G. Konstantatos, W. J.

Parak, T. Hyeon, B. A. Korgel, C. B. Murray, W. Heiss, ACS Nano 2015, 9, 1012.

[4] O. E. Semonin, J. M. Luther, S. Choi, H.-Y. Chen, J. Gao, A. J. Nozik, M. C. Beard,

Science 2011, 334, 1530.

[5] G. H. Carey, A. L. Abdelhady, Z. Ning, S. M. Thon, O. M. Bakr, E. H. Sargent,

Chemical Reviews 2015.

[6] X. Wang, G. I. Koleilat, J. Tang, H. Liu, I. J. Kramer, R. Debnath, L. Brzozowski, D.

A. R. Barkhouse, L. Levina, S. Hoogland, E. H. Sargent, Nat Photon 2011, 5, 480.

[7] I. J. Kramer, E. H. Sargent, Chemical Reviews 2014, 114, 863.

[8] M. Graetzel, R. A. J. Janssen, D. B. Mitzi, E. H. Sargent, Nature 2012, 488, 304.

[9] C.-H. M. Chuang, P. R. Brown, V. Bulović, M. G. Bawendi, Nat Mater 2014, 13, 796.

[10] Z. Ning, O. Voznyy, J. Pan, S. Hoogland, V. Adinolfi, J. Xu, M. Li, A. R. Kirmani, J.-

P. Sun, J. Minor, K. W. Kemp, H. Dong, L. Rollny, A. Labelle, G. Carey, B. Sutherland, I.

Hill, A. Amassian, H. Liu, J. Tang, O. M. Bakr, E. H. Sargent, Nat Mater 2014, 13, 822.

[11] I. J. Kramer, J. C. Minor, G. Moreno-Bautista, L. Rollny, P. Kanjanaboos, D.

Kopilovic, S. M. Thon, G. H. Carey, K. W. Chou, D. Zhitomirsky, A. Amassian, E. H.

Sargent, Advanced Materials 2015, 27, 116.

[12] I. J. Kramer, G. Moreno-Bautista, J. C. Minor, D. Kopilovic, E. H. Sargent, Applied

Physics Letters 2014, 105, 163902.

21

[13] P. Guyot-Sionnest, The Journal of Physical Chemistry Letters 2012, 3, 1169.

[14] D. V. Talapin, J.-S. Lee, M. V. Kovalenko, E. V. Shevchenko, Chemical Reviews

2010, 110, 389.

[15] D. Zhitomirsky, O. Voznyy, L. Levina, S. Hoogland, K. W. Kemp, A. H. Ip, S. M.

Thon, E. H. Sargent, Nat Commun 2014, 5.

[16] D. Bozyigit, O. Yarema, V. Wood, Advanced Functional Materials 2013, 23, 3024.

[17] D. Bozyigit, S. Volk, O. Yarema, V. Wood, Nano Letters 2013, 13, 5284.

[18] D. Bozyigit, W. M. M. Lin, N. Yazdani, O. Yarema, V. Wood, Nat Commun 2015, 6.

[19] K. Katsiev, A. H. Ip, A. Fischer, I. Tanabe, X. Zhang, A. R. Kirmani, O. Voznyy, L.

R. Rollny, K. W. Chou, S. M. Thon, G. H. Carey, X. Cui, A. Amassian, P. Dowben, E. H.

Sargent, O. M. Bakr, Advanced Materials 2014, 26, 937.

[20] A. R. Kirmani, G. H. Carey, M. Abdelsamie, B. Yan, D. Cha, L. R. Rollny, X. Cui, E.

H. Sargent, A. Amassian, Advanced Materials 2014, 26, 4717.

[21] D. Kim, D.-H. Kim, J.-H. Lee, J. C. Grossman, Physical Review Letters 2013, 110,

196802.

[22] G. W. Hwang, D. Kim, J. M. Cordero, M. W. B. Wilson, C.-H. M. Chuang, J. C.

Grossman, M. G. Bawendi, Advanced Materials 2015, 27, 4481.

[23] Y. Zhang, D. Zherebetskyy, N. D. Bronstein, S. Barja, L. Lichtenstein, A. P.

Alivisatos, L.-W. Wang, M. Salmeron, ACS Nano 2015.

[24] S. C. Boehme, J. M. Azpiroz, Y. V. Aulin, F. C. Grozema, D. Vanmaekelbergh, L. D.

A. Siebbeles, I. Infante, A. J. Houtepen, Nano Letters 2015, 15, 3056.

[25] C.-H. M. Chuang, A. Maurano, R. E. Brandt, G. W. Hwang, J. Jean, T. Buonassisi, V.

Bulović, M. G. Bawendi, Nano Letters 2015, 15, 3286.

[26] M. M. Krause, P. Kambhampati, Physical Chemistry Chemical Physics 2015, 17,

18882.

[27] G. H. Carey, K. W. Chou, B. Yan, A. R. Kirmani, A. Amassian, E. H. Sargent, MRS

Communications 2013, 3, 83.

[28] J. M. Luther, M. Law, Q. Song, C. L. Perkins, M. C. Beard, A. J. Nozik, ACS Nano

2008, 2, 271.

[29] M. Shim, P. Guyot-Sionnest, Nature 2000, 407, 981.

[30] D. Yu, C. Wang, P. Guyot-Sionnest, Science 2003, 300, 1277.

[31] A. K. Rath, F. Pelayo Garcia de Arquer, A. Stavrinadis, T. Lasanta, M. Bernechea, S.

L. Diedenhofen, G. Konstantatos, Advanced Materials 2014, 26, 4741.

[32] A. Hassinen, I. Moreels, K. De Nolf, P. F. Smet, J. C. Martins, Z. Hens, Journal of the

American Chemical Society 2012, 134, 20705.

[33] D. J. Harrison, N. Nguyen, A. J. Lough, U. Fekl, Journal of the American Chemical

Society 2006, 128, 11026.

[34] M. L. Geier, K. Moudgil, S. Barlow, S. R. Marder, M. C. Hersam, Nano Letters 2016,

16, 4329.

[35] S. A. Paniagua, J. Baltazar, H. Sojoudi, S. K. Mohapatra, S. Zhang, C. L. Henderson,

S. Graham, S. Barlow, S. R. Marder, Materials Horizons 2014, 1, 111.

[36] A. R. Kirmani, W. Peng, R. Mahfouz, A. Amassian, Y. Losovyj, H. Idriss, K. Katsiev,

Carbon 2015, 94, 79.

[37] K. Katsiev, Y. Losovyj, Z. Zhou, E. Vescovo, L. Liu, P. A. Dowben, D. W. Goodman,

Physical Review B 2012, 85, 195405.

[38] E. M. Miller, D. M. Kroupa, J. Zhang, P. Schulz, A. R. Marshall, A. Kahn, S. Lany, J.

M. Luther, M. C. Beard, C. L. Perkins, J. van de Lagemaat, ACS Nano 2016, 10, 3302.

[39] A. Tarasov, S. Zhang, M.-Y. Tsai, P. M. Campbell, S. Graham, S. Barlow, S. R.

Marder, E. M. Vogel, Advanced Materials 2015, 27, 1175.

[40] P. R. Brown, D. Kim, R. R. Lunt, N. Zhao, M. G. Bawendi, J. C. Grossman, V.

Bulović, ACS Nano 2014, 8, 5863.

22

[41] S. K. Mohapatra, Y. Zhang, B. Sandhu, M. S. Fonari, T. V. Timofeeva, S. R. Marder,

S. Barlow, Polyhedron.

[42] Y. Qi, T. Sajoto, M. Kröger, A. M. Kandabarow, W. Park, S. Barlow, E.-G. Kim, L.

Wielunski, L. C. Feldman, R. A. Bartynski, J.-L. Brédas, S. R. Marder, A. Kahn, Chemistry of

Materials 2010, 22, 524.

[43] A. H. Ip, S. M. Thon, S. Hoogland, O. Voznyy, D. Zhitomirsky, R. Debnath, L.

Levina, L. R. Rollny, G. H. Carey, A. Fischer, K. W. Kemp, I. J. Kramer, Z. Ning, A. J.

Labelle, K. W. Chou, A. Amassian, E. H. Sargent, Nat Nano 2012, 7, 577.

[44] Z. Ning, D. Zhitomirsky, V. Adinolfi, B. Sutherland, J. Xu, O. Voznyy, P.

Maraghechi, X. Lan, S. Hoogland, Y. Ren, E. H. Sargent, Advanced Materials 2013, 25,

1719.

[45] W.-k. Koh, A. Y. Koposov, J. T. Stewart, B. N. Pal, I. Robel, J. M. Pietryga, V. I.

Klimov, Scientific Reports 2013, 3, 2004.

[46] P. Maraghechi, A. J. Labelle, A. R. Kirmani, X. Lan, M. M. Adachi, S. M. Thon, S.

Hoogland, A. Lee, Z. Ning, A. Fischer, A. Amassian, E. H. Sargent, ACS Nano 2013, 7, 6111.

[47] M. A. Hines, G. D. Scholes, Advanced Materials 2003, 15, 1844.

[48] J. VandeVondele, M. Krack, F. Mohamed, M. Parrinello, T. Chassaing, J. Hutter,

Computer Physics Communications 2005, 167, 103.

[49] M. Burgelman, K. Decock, S. Khelifi, A. Abass, Thin Solid Films 2013, 535, 296.

A highly versatile remote molecular doping platform for colloidal quantum dot (CQD)

solids, employing organomettalic complexes, is presented. Applied to a standard CQD

heterojunction solar cell, the single-step doping recipe reduces trap state recombination and

results in robust solar cells with power conversion efficiency enhancement by ca. 30%.

Keyword (see list)

Ahmad R. Kirmani, Amirreza Kiani, Marcel M. Said, Oleksandr Voznyy, Nimer Wehbe, Grant

Walters, Stephen Barlow, Edward H. Sargent, Seth R. Marder, Aram Amassian*

A General Remote Molecular Doping Platform for Colloidal Quantum Dot Solids

ToC figure ((55 mm broad, 50 mm high, or 110 mm broad, 20 mm high))

23

Supporting Information

A General Remote Molecular Doping Platform for Colloidal Quantum Dot Solids

Ahmad R. Kirmani, Amirreza Kiani, Marcel M. Said, Oleksandr Voznyy, Nimer Wehbe, Grant

Walters, Stephen Barlow, Edward H. Sargent, Seth R. Marder, Aram Amassian*

Figure S1. Pb 4f core levels for the strong and weak p-doped CQD solids, as measured by XPS.

The strong p-dopant leads to a significant shift of the core level by ~0.3 eV whereas no apparent

peak shift is observed for the weak p-dopant. These shifts correspond to the corresponding shifts

in the Fermi level observed from UPS.

Figure S2. Molecular orbital energies for the isolated dopant molecule calculated by DFT.

24

Figure S3. UV photoemission spectroscopy (UPS) data for films doped with various dopant

concentrations. (a) Secondary electron cutoffs give information about the work-function shifts,

(b) Valence band maxima.

Table S1. Relative atomic concentration of Pb, S, C and F as found from high resolution XPS.

F/Pb ratio is found to increase with the doping concentration. Accordingly, the number of

dopant molecules per QD, χ, also increases.

Sample Pb S C F F/Pb χ

MPA baseline 32.8 29.2 38.0 0.0 - -

1.0 mgmL-1 27.9 24.4 35.2 12.5 0.45 12

0.1 mgmL-1 30.4 28.8 33.6 7.1 0.23 6

0.01 mgmL-1 33.2 30.8 35.2 0.4 0.01 0.3

Discussion:

χ represents the number of dopant molecules per QD and has been calculated as follows:

Each QD has been considered to be containing ~500 atoms of Pb in accordance with published

report.[1] The dopant molecule contains 18 F atoms.

17 16 15 14

Inte

nsity (

A.U

.)

Binding Energy (eV)

SE

(a)

4 3 2 1 0 -1

VB

Inte

nsity (

A.U

.)

Binding Energy (eV)

MPA baseline

10-5 mg/ml

10-4 mg/ml

10-3 mg/ml

10-2 mg/ml

10-1 mg/ml

1.0 mg/ml

(b)

25

Experimental value for F:Pb in the doped QD solid as found from XPS = (F/Pb)

Therefore, dopant molecules per QD, χ = (F/Pb)/(18/500)

Figure S4. SIMS depth profiles for the QD films doped with (a) 0.1 mg mL-1 and (b) 1.0 mg

mL-1 dopant solutions.[2] The films were deposited on SiO2 wafers. The fluorine signal (green)

is found to be enhanced for the 1.0 mg mL-1 treated film, (b). Importantly, dopant molecules

(represented by fluorine) are found to penetrate through the films and no surface accumulation

is observed. A single layer in the layer-by-layer depostion recipe for QDs yields a ca. 30 – 40

nm layer. Films for these measurements were restricted to a single layer since depostion of more

than a layer was found to yield highly inhomogenous films incompatible with the requirement

for SIMS measurements.

0 10 20 30 40 50 60 70

102

103

104

105

(b)(a)

Inte

nsity (

cps)

Depth (nm)

S

C

PbSiF

0 10 20 30 40 50 60 70

102

103

104

105

1mg/ml doped 0.1mg/ml doped

Depth (nm)

Pb

C

S

F

Si

Mo

26

Table S2. Change in work function, , and the change in Fermi level with respect to the VBM,

EVB

are used to calculate the induced dipole, ED, for the various doping scenarios using the

equation, = EVB + ED. The changes are measured with respect to the MPA baseline.

Dopant

concentration

(mgmL-1) (eV)

(eV)

( - 4.17) E

VB (eV)

EVB

(eV)

(0.79-EVB

) E

D (eV)

10-5 4.24 0.07 0.75 0.04 0.03

10-4 4.28 0.11 0.71 0.08 0.03

10-3 4.31 0.14 0.70 0.09 0.05

10-2 4.40 0.23 0.61 0.18 0.05

0.1 5.14 0.97 0.51 0.28 0.69

1.0 5.40 1.23 0.50 0.29 0.94

Figure S5. A different doping scheme was tested involving doping of each individual QD layer

in the LbL stack. The device performances were found to decrease (Table S3 below)

(a)

-0.2 0.0 0.2 0.4 0.6-20

-15

-10

-5

0

5

MPA baseline

0.01 mg/ml doped

0.1 mg/ml doped

1.0 mg/ml doped

Cu

rre

nt D

en

sity (

mA

/cm

2)

Voltage (V)

(b)

FTO glass

TiO2

5 LbL individually doped PbS QD

MoO3

Ag

Au

27

Table S3. Device parameters for solar cells obtained by doping each individual QD layer

separately.

Device Jsc (mA/cm2) Voc (V) FF (%) PCE (%)

MPA baseline 15.01 0.500 48.0 3.60

0.01 mgmL-1 14.51 0.491 41.3 2.95

0.1 mgmL-1 3.54 0.312 26.0 0.29

1.0 mgmL-1 1.72 0.198 27.1 0.09

Figure S6. VASE was used to precisely measure the QD film thickness upon doping. A 2-layer

stack was fabricated by spin coating on SiO2 wafer. This was followed by the doping step. A

~50% reduction in film thickness was observed after the ligand exchange step,[3] however the

thickness did not show any observable change upon doping or the subsequent ACN rinse.

28

Figure S7. Simulated energy band diagrams inside the CQD absorber layer for various doping

scenarios. The Fermi levels are found to get closer to the band edges upon doping. However,

overdoping leads to a collapse of the depletion width causing a reduction in JSC and therefore

the overall device performance.

29

Figure S8. Doped solar cells employing a ca. 150 nm MPA-capped PbS QD absorber layer.

Lower doping concentrations were found to increase the PCE, as shown in Table S4 below.

30

Table S4. Device parameters for solar cells obtained by doping thin-absorber layer solar cells.

Device JSC

(mA/cm2) VOC (V) FF (%) PCE (%)

Rshunt

(Ω/cm2)

Rs

(Ω/cm2)

MPA

baseline 19.0 0.471 43.6 3.90 957 5.2

10-6

mgmL-1 20.3 0.497 47.2 4.76 1096 3.7

10-5

mgmL-1 18.6 0.505 45.7 4.31 1318 5.3

10-4

mgmL-1 17.6 0.497 46.6 4.09 1264 4.1

10-3

mgmL-1 18.7 0.466 42.3 3.69 1206 6.4

0.1 mgmL-1 9.3 0.232 29.5 0.64 337 10.5

Figure S9. The QD absorber layers were deposited on room temperature processed TiO2. No

high temperature annealing was needed. Doping was found to improve the overall performance

(see Table S5 below).

-0.2 0.0 0.2 0.4 0.6

-20

-15

-10

-5

0

5

Cu

rre

nt

De

nsity (

mA

/cm

2)

Voltage (V)

MPA baseline

1mg/ml

0.01 mg/ml

10-4 mg/ml

Room T TiO2

(b)(a)

FTO glass

Room T TiO2

5 LbL doped PbS QD

Ag

MoO3

Au

31

Table S5. Summary of the device parameters for the room temperature processed TiO2-based

solar cells. All the parameters are found to increase for lower doping concnetrations. The

significant increase in Rshunt suggests that doping leads to a reduction in trap-assisted

recombination.[4]

Device JSC

(mA/cm2)

VOC

(V)

FF (%) PCE (%) Rshunt

(Ω/cm2)

Rs

(Ω/cm2)

MPA baseline 17.5 0.434 49.5 3.75 1187 2.1

10-4 mgmL-1 19.8 0.511 52.3 5.29 1952 3.1

10-2 mgmL-1 18.7 0.492 50.8 4.68 1670 2.6

1.0 mgmL-1 6.2 0.434 35.0 0.94 1808 3.3

Figure S10. Photoluminescence spectra for QD films doped with different doping

concentrations. The higher concentrations lead to quenching of the exciton emission around

1.25 eV which might be due to the significant removal of QD valence electrons by the dopant

at those concentrations. This is evident from the relatively large shift of the Fermi level for the

high doping concentrations (EVB in Table S2).[5]

The spectra have a non-Gaussian shape with

the 1.0 mgmL-1 doped sample’s spectrum showing the maximum deviation from Gaussian

nature. We were not able to ascertain the reason behind this. We observe a reflection peak from

the glass substrate around 1.4 eV which was used to normalize the spectra.

32

Figure S11. Interdot spacing found by azimuthally integrating the GISAXS intensities. The

spacing achieved after MPA ligand exchange (ca. 3.7 nm) remains unchanged upon doping.

Figure S12. The absorbance of the films remains essentially unchanged upon doping. A slight

decrease in exciton absorption is found for 1 mgmL-1 doping concentration as highlighted in

the inset (red spectrum).

33

Figure S13. Transient absorption data for the control (undoped, MPA-exchanged films),

optimally doped (10-2 mg mL-1) and overdoped (1 mg mL-1) films is shown. The highest doped

film shows quenching at the exciton wavelength (950 nm).

[1] A. H. Ip, S. M. Thon, S. Hoogland, O. Voznyy, D. Zhitomirsky, R. Debnath, L. Levina,

L. R. Rollny, G. H. Carey, A. Fischer, K. W. Kemp, I. J. Kramer, Z. Ning, A. J. Labelle, K. W.

Chou, A. Amassian, E. H. Sargent, Nat. Nano. 2012, 7, 577.

[2] A. Dai, A. Wan, C. Magee, Y. Zhang, S. Barlow, S. R. Marder, A. Kahn, Org. Electron.

2015, 23, 151.

[3] A. R. Kirmani, G. H. Carey, M. Abdelsamie, B. Yan, D. Cha, L. R. Rollny, X. Cui, E.

H. Sargent, A. Amassian, Adv. Mater. 2014, 26, 4717.

[4] A. K. Rath, F. Pelayo Garcia de Arquer, A. Stavrinadis, T. Lasanta, M. Bernechea, S. L.

Diedenhofen, G. Konstantatos, Adv. Mater. 2014, 26, 4741.

[5] W.-k. Koh, A. Y. Koposov, J. T. Stewart, B. N. Pal, I. Robel, J. M. Pietryga, V. I. Klimov,

Sci. Rep. 2013, 3, 2004.