Synthesis and use of a hyper-connecting cross-linking agent in the hole-transporting layer of perovskite solar cells

Article abstract of DOI:10.1039/c7ta05004f

Solution-processed organic semiconducting materials feature prominently in modern optoelectronic devices, especially where low-cost and flexibility are specific goals, such as perovskite solar cells. Their intrinsic solubility, poor cohesion and lack of a

Full text of DOI:10.1039/c7ta05004f

View Article Online
View Journal
Journal of
M at erials Chemistry A
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: B. L. Watson, N. J.
Rolston, K. A. Bush, L. Taleghani and R. H. Dauskardt, J. Mater. Chem. A, 2017, DOI:
1
0.1039/C7TA05004F.
Volume
4
Number
1
7
January 2016 Pages 1–330
This is an Accepted Manuscript, which has been through the
Royal Society of Chemistry peer review process and has been
accepted for publication.
Journal of
Materials Chemistry A
Accepted Manuscripts are published online shortly after
Materials for energy and sustainability
www.rsc.org/MaterialsA
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
author guidelines.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the ethical guidelines, outlined
in our author and reviewer resource centre, still apply. In no
event shall the Royal Society of Chemistry be held responsible
for any errors or omissions in this Accepted Manuscript or any
consequences arising from the use of any information it contains.
ISSN 2050-7488
PAPER
Kun Chang, Zhaorong Chang et al.
Bubble-template-assisted synthesis of hollow fullerene-like
MoS nanocages as a lithium ion battery anode material
2
rsc.li/materials-a

Page 1 of 28
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA05004F
Synthesis and Use of a Hyper-Connecting Cross-linking Agent in the
Hole-transporting Layer of Perovskite Solar Cells
1
, *
2, *
1
1
Brian L. Watson , Nicholas Rolston Kevin A. Bush , Leila Taleghani and Reinhold H.
1
,†
Dauskardt
1
2
Department of Materials Science and Engineering, Stanford University, Stanford, CA, USA
Department of Applied Physics, Stanford University, Stanford, CA, USA
*
These authors contributed equally to this work.
Corresponding author:
†
Prof. Reinhold H. Dauskardt
Department of Materials Science and Engineering
Stanford University, Stanford, CA 94305ꢀ2205
Phone: +1 650 725 0679
Fax: +1 650 725 4034
eꢀmail: dauskardt@stanford.edu
Keywords: solventꢀresistance; crossꢀlinking agent; organic semiconductors; perovskite solar
cells; holeꢀtransport material.

Journal of Materials Chemistry A
Page 2 of 28
View Article Online
DOI: 10.1039/C7TA05004F
Synthesis and Use of a Hyper-Connecting Cross-linking Agent in the
Hole-transporting Layer of Perovskite Solar Cells
1
, *
2, *
Brian L. Watson , Nicholas Rolston , Kevin A. Bush, Leila Taleghani and Reinhold H.
1
,†
Dauskardt
Abstract
Solutionꢀprocessed organic semiconducting materials feature prominently in modern
optoelectronic devices, especially where lowꢀcost and flexibility are specific goals, such as
perovskite solar cells. Their intrinsic solubility, poor cohesion and lack of adhesion to underlying
substrates, however, curtail their scope of application and durability. To overcome this, a
mechanically stiff, lightꢀactivated, tetraꢀazide crossꢀlinking agent, 1,3,5,7ꢀtetrakisꢀ(pꢀ
benzylazide)ꢀadamantane (TPBA), has been developed to transform solution processed organic
polymers into solventꢀresistant and mechanically tough films. The use of 3ꢀ
azidopropyltrimethoxysilane (AzPTMS) has been developed as a lightꢀactivated adhesion
promotor,
enabling
testing
of
toughened,
(LiTFSI)
crossꢀlinked
doped
polymers.
Lithium
bis(trifluoromethane)sulfonimide
poly[bis(4ꢀphenyl)(2,4,6ꢀ
trimethylphenyl)amine, poly(triaryl amine) (PTAA), a holeꢀtransporting material used in
perovskite solar cells, has been selected as a candidate system for demonstrating the utility of
TPBA to transform a fragile, and highlyꢀsoluble holeꢀtransporting organic semiconductor into a
mechanically tough and solventꢀresistant semiconducting composite. TPBA enables the solvent
resistance and mechanical toughness of PTAA to be tuned without compromising the electronic
functionality of the semiconducting material. While increasing the fracture toughness of PTAA
by over 300%, TPBA crossꢀlinking also enables fabrication of perovskite solar cells with
increased photovoltaic efficiencies in nꢀiꢀp and pꢀiꢀn geometries, and promotes adhesion of the
doped polymer to the perovskite layer, mitigating interfacial device failure.

Page 3 of 28
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA05004F
Introduction
Solution processed semiconducting organic polymers are used in several promising electronic
1
ꢀ4
and optoelectronic technologies as chargeꢀtransporting layers , lightꢀabsorbing photovoltaic
5
,6
7ꢀ10
materials
techniques compatible with rollꢀtoꢀroll printing,
and as electroluminescent materials
.
Moreover, solutionꢀstate deposition
1
1,12
13,14
such as slotꢀdie coating, shear coating
,
1
5,16
and inkꢀjet printing
flexible electronic devices .
A challenge limiting the scope of solution processed flexible electronic technologies is
their inherent solubility once deposited, rendering them susceptible to etching or complete
present the opportunity for lowꢀcost, highꢀthroughput fabrication of
1
7
18ꢀ20
dissolution by the solvents of subsequent deposition steps
. Consequently, the fabrication of
multilayer devices by solution processing is severely limited. The fabrication of solutionꢀ
processed, pꢀiꢀn perovskite solar cells is a particular example of this limitation. In these devices,
the ability of solution processed organic hole transporting layers to resist etching or complete
dissolution during the deposition of the perovskite layer (which frequently require the use of
aromatic antiꢀsolvents) heavily influences the performance of the resulting devices.
Along with solvent resistance, the mechanical resilience of these layers is important for
ensuring a high yield in fabrication, a strict requirement for lowꢀcost devices, and reliability,
ensuring performance lifetimes, which in the case of solar cells, may be 2ꢀ3 decades. The
cohesion energy, G , of solution processed semiconducting polymer layers are often low,
c
2
21,22
typically below ~ 5 J/m
as was the case of organic polymer photovoltaics, the predecessor
to perovskite solar cells as a solutionꢀprocessed, lowꢀcost solar technology, which encountered
significant reliability challenges and were generally regarded as being thermomechanically
2
3
fragile .
The fragility of solution processed semiconducting polymers stems from the low density
of covalent bonding within these materials. When meaningful interactions between constituent
polymer strands in these types of materials do not exist, such as πꢀorbital overlaps or hydrogen
bonding for example, the degree to which energyꢀdissipating mechanism like plastic deformation
ahead of an advancing crack tip may be engaged is restricted, lowering the barrier to critical
fracture propagation. Increasing the MW of polymers, allowing for increased interactions and
2
4
entanglements per strand, can effectively increase the cohesion of a layer . This strategy for
increasing polymer cohesion, however, requires that polymer solubility can be maintained during

Journal of Materials Chemistry A
Page 4 of 28
View Article Online
DOI: 10.1039/C7TA05004F
synthesis to achieve these higher molecular weights. Similarly, the absence of covalent chemical
bonding and interactions at the interfaces of these organic semiconducting layers has resulted in
2
5
some of the lowest values for adhesive fracture reported for electronic devices .
A survey of the critical fracture energy, G , an essential indicator of thermomechanical
c
2
6
reliability , of perovskite solar cells highlights how their development within the academic
research environment to increasingly higher efficiencies has occurred at the expense of the
2
2
mechanical integrity of the devices . In this work, we demonstrate that this need not be the case,
and that increases in efficiency can be achieved in the pursuit of thermomechanical stability.
Imbuing solutionꢀprocessed semiconducting polymers with increased resistance to
fracture and dissolution following deposition, will enable fabrication of a new variety of robust,
solutionꢀprocessed polymer devices, increasing both their manufacturability and reliability.
Generating covalent crossꢀlinks within polymers following solutionꢀstate deposition is a simple
strategy to achieve this. In the case of semiconducting polymers, crossꢀlinking can be achieved
8
,27
by redesigning the polymer to feature crossꢀlinking chemical functionalities , or more simply,
2
8ꢀ30
by addition of a molecular crossꢀlinking agent
. The latter approach has the advantage of
using existing semiconducting polymers, avoiding the development of complex new polymerꢀ
specific syntheses and purification protocols.
To date, aromatic azide crossꢀlinking agents, like the sterically hindered bis(fluorphenyl
azide) (sFPA), bis(4ꢀazidoꢀ2,3,ꢀtrifluorꢀ6ꢀisopropylbenzoate) shown in Fig. 1a, have been
successfully used in organic polymer solar cells for creating semiconductingꢀpolymer
2
8
31
heterostructures , nanotemplated polymer donor network films, and for controlling the
3
2
morphology of P3HT domains in bulkꢀheterojunctions . The covalent crossꢀlinks of azide
crossꢀlinking agents are formed via thermal or deep UV activation of the azide group, forming
2
8,30ꢀ32
singlet nitrene radicals which proceed to insert into CꢀH bonds present on the polymer
Singlet nitrene radicals formed on activation of aromatic and alkyl azides have been
.
3
3ꢀ35
studied extensively in the solution state via photolysis experiments
. These complex
experiments often require indirect detection of reaction products due to overlapping signals in
3
3
the transient spectra , and highlight the wide variety of reaction pathways along which singlet
nitrene radicals may proceed in the solution state, most of which do not result in CꢀH insertion
reactions required for effective crossꢀlinking.

Page 5 of 28
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA05004F
However, in reporting the use of sFPAs in heterojunction polymer solar cells, Png et al.
observed that CꢀH insertion of nitrene radicals polymers is the dominant outcome of azide
activation in the solid state and that the quantum yield of known triplet and ketenimine side
reactions were suppressed to < 0.02, well below the solution state values that typically range
2
8
from 0.5ꢀ0.8 in the solution state . This work indicates that direct comparisons of solution state
photolysis and photochemical crossꢀlinking in the solid state may not be made with confidence
and that further work is required specifically studying the solidꢀstate reactivity of activated
azides.
Derue et al. expanded on the use of azide crossꢀlinking agents presenting a simpler
design strategy, departing from the sFPA design using a bisꢀbenzylazide crossꢀlinking agent,
4,4’ꢀbis(azidomethyl)ꢀ1ꢀ1’ꢀbiphenyl (BABP) (Fig. 1a) and demonstrating that in addition to CꢀH
insertion reactions, azides crossꢀlinking agents can be used to stabilize the morphology of
polymer:fullerene blends through selective, thermallyꢀactivated addition reactions with
fullerenes forming azafulleroids and closed [6.6]aziridinofullerenes. In this work, neither
fluorination of the aromatic ring nor the presence of sterically inhibiting substituents defining the
design of sFPAs were required for effective crossꢀlinking. In addition to the simplified chemical
makeꢀup and corresponding ease of synthesis of BABP, Derue et al. reported that in contrast to
sFPAs, which are liquid at RT, the solid form of BABP resulted in increased stability enabling
3
0
storage at RT .
While the solidꢀstate reactivity of azides in the solid state do exhibit very different
quantum yields in comparison to those characterized in solution, which in the case of benzyl
3
6ꢀ38
azide in particular, has proven to be a controversial area of research
, even in the solid state
limited side reactions can be expected that would reduce the crossꢀlinking efficiency.
Consequently, a crossꢀlinking agent featuring only two azide groups is rendered ineffective if
just one of the nitrene radicals formed proceeds along a reaction pathway that does not result in
CꢀH insertion. The crossꢀlinking efficiency of azide crossꢀlinking agents can therefore be
increased simply by increasing the number of azide groups present on an agent. When four azide
groups are present on a crossꢀlinking agent, failure to achieve crossꢀlinking requires three of the
azides to proceed along reaction pathways that do not result in nitrene insertion, which is far less
probable than a twoꢀazide agent that only requires one to render the agent ineffective. The
presence of more azides significantly increases the opportunity for creating points of hyperꢀ

Journal of Materials Chemistry A
Page 6 of 28
View Article Online
DOI: 10.1039/C7TA05004F
connectivity within polymeric materials where more than two independent polymer strands may
be crossꢀlinked (Fig. 1b).
In this work, focused on addressing the poor mechanical stability of semiconducting
polymers and their intrinsic solubility through development of a crossꢀlinking agent capable of
creating these nodes of hyperꢀconnectivity, we have selected to use the chemical simplicity of
3
0
benzyl azide crossꢀlinking agents like BABP to create 1,3,5,7ꢀtetrakisꢀ(pꢀformylphenyl)ꢀ
adamantane (TPBA) shown in Fig. 1a. TPBA features four reactive benzylazide groups
positioned on a tetrahedral, sigmaꢀbonded adamantane core. The selection of the rigid
adamantane core is threeꢀfold. Firstly, the rigid core prevents benzylazide groups from bending
back upon one another, preventing selfꢀreaction of the agent. Secondly, regardless of the
orientation of TPBA in a polymer blend, at least one crossꢀlinking arm is always subtended with
a in each of the three dimensions of space, preventing isotropic toughening of polymers. Finally,
the sigmaꢀbonded adamantane core prevents electronic conjugation across TPBA, limiting the
degree to which the crossꢀlinking agent may interfere with electronic orbitals of semiconducting
polymers.
In this work, we present the synthesis of TPBA and demonstrate its utility as a
crosslinking agent for systematically transforming the holeꢀtransporting polymer used in some of
3
9,40
the most efficient perovskite solar cells fabricated to date
, poly[bis(4ꢀphenyl)(2,4,6ꢀ
trimethylphenyl)amine, poly(triaryl amine) (PTAA), into mechanically tough films, increasing
their fracture resistance by 345% and significantly increasing their resistance to solvent etching.
Through our work investigating the fracture behavior of solution processed perovskite
solar cells, cells featuring the PTAA HTM doped with LiTFSI (HTM) fractured adhesively at the
2
22
dꢀPTAA/perovskite interface at G values well below 0.5 J/m . This doped PTAA employs
c
3
9,40
roughly one LiTFSI molecule to every eight PTAA mers
, a significant quantity, prompting
an investigation of the influence of LiTFSI on PTAA adhesion and cohesion. We report that
LiTFSI significantly compromises the adhesion of PTAA to substrates, that it reduces the
cohesion of this PTAA as a whole, and that TPBA crossꢀlinking can be used to overcome this
decrease in fracture energy.
In addition, to address the poor adhesive properties of semiconducting polymers and to
enable the fracture testing featured in this work, we have further leveraged azide chemistry and
report the use of a light activated adhesion promoter, 3ꢀ(azidopropyl)triemthoxysilane

Page 7 of 28
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA05004F
(AzPTMS) for covalently bonding organic materials to oxide substrates that react readily with
trimethoxy organosilanes, like glass and silicon (Fig. 2b), a significant advance for promoting
adhesion of solution processed organic materials to substrates.
Results
The synthesis of TPBA was achieved from 1,3,5,7ꢀtetrakisꢀ(pꢀformylphenyl)ꢀadamantane (1) in
three, simple, highꢀyielding steps as shown in Fig. 2a. All steps in the synthesis can be
accomplished quickly providing multiꢀgram quantities of pure TPBA. This synthesis involves
reduction of 1 to produce 1,3,5,7ꢀtetrakisꢀ(pꢀbenzylalcohol)ꢀadamantane (2) in 96% yield using
NaBH . Introduction of the bromine leavingꢀgroups of 1,3,5,7ꢀtetrakisꢀ(benzylbromide)ꢀ
4
adamantane (3) is achieved by reaction of 2 with PBr . Finally, the azide functionality of TPBA
3
is introduced by substitution of the bromine atoms using NaN , affording TPBA in 86 % yield.
3
In this work, we employed a known, twoꢀstep synthesis to obtain large quantities of the
precursor 1,3,5,7ꢀtetrakisꢀ(pꢀformylphenyl)ꢀadamantane (1) for the synthesis of TPBA.
Specifically, 1,3,5,7ꢀtetraphenyladamantane (TPA) was obtained using the literature method
41
reported by Reichert et al utilizing an AlCl promoted FriedelꢀCrafts arylation of 1ꢀ
3
41
bromoadamantane with benzene in the presence of tertꢀbutyl bromide. This method affords
large (> 20g) quantities of the benzeneꢀfunctionalized adamantane core, 1,3,5,7ꢀtetrakisꢀ
(
phenyl)adamantane (TPA). Synthesis of 1, was achieved by a TiCl promoted Rieche
4
formylation of TPA with α,αꢀdichlomethylꢀmethylether utilizing the modification to a patent
4
2
procedure reported by Duncan et al . TPBA dissolves readily in a wide variety of chlorinated
and aromatic solvents and does not exhibit absorbance in the visible region of the
electromagnetic spectrum or result in changes in the absorption spectrum of PTAA films crossꢀ
linked with TPBA (Fig. S12). The high ratio of carbon to azide in the molecule, 9:1, satisfies the
4
3
“
rule of six” reducing concerns related to potential explosive hazards related to the azide group.
An upper limit of 0.2 w.eq. was selected for the quantity of TPBA added to PTAA films
in this work. 0.2 w.eg represents a larger that 10 % concentration. We regarded this a
significantly large addition in a study focused on probing the efficacy of small additions of a
potentially insulating agent to the semiconducting films. Films of PTAA containing increasing

Journal of Materials Chemistry A
Page 8 of 28
View Article Online
DOI: 10.1039/C7TA05004F
weight equivalents (0.01, 0.05, 0.1 and 0.2) of TPBA were deposited on silicon substrates by
spinꢀcoating and crossꢀlinked by irradiation with UVꢀC (254 nm).ATR IR spectra revealed
complete reaction of the azide groups within 10 min of irradiation (Fig. S13) with a handꢀheld
UV lamp. To probe the influence of TPBA crossꢀlinking on PTAA solvent resistance, films were
dipped in toluene for 30 s. Films containing less than 0.1 weight equivalents of TPBA dissolved
readily, while films containing 0.2 weight equivalents (PTAAꢀX ) were significantly more
0
.2
resistant to dissolution by toluene (Fig. S14) with the presence of the film persisting on
substrates tested.
Influence of LiTFSI doping on PTAA fracture
On increasing the concentration of LiTFSI in PTAA polymer films deposited on glass
substrates from 0 to 0.15 molar equivalents (equivalent to PTAA monomeric units), we observed
2
that the adhesion of PTAA to glass decreased by more than 90%, from 2.04 J/m to only ~ 0.14
2
J/m (Fig. 3a). We propose that with the increase of LiTFSI in the polymer film, interactions
between the polymer and the substrate are increasingly inhibited by the presence of the TFSI
anion and that the resulting decrease in contact of the polymer with the glass substrate reduces
the adhesion at this interface.
Test structures of undoped PTAA deposited on perovskite exhibited failure within the
2
perovskite layer at a G of ~ 0.5 J/m , as shown in Fig. 3b. Again, addition of LiTFSI resulted in
c
failure to occur at the interface of dꢀPTAA with a substrate, in this case, the perovskite. Here the
2
two materials delaminated with a G of ~ 0.47 J/m . By addition of 0.1 w.eq. and 0.2 w.eq. of
c
TPBA to dꢀPTAA, this adhesive failure was inhibited, indicating that TPBA promotes adhesion
4
4
of dꢀPTAA to the perovskite surface. A similar achievement has been reported for perovskite
cells using spiroꢀOMeTAD as the HTM through the use of polyethylenimine (PEI), but to date,
has not been achieved for mitigating the delamination of dꢀPTAA from the perovskite surface.
Advances required to mitigate the cohesive failure of the brittle perovskite layer further leverage
TPBA and are reported elsewhere and which could not be achieved through the use of PEI.
If LiTFSI is found to decrease the cohesion of PTAA too, it would indicate that PTAA
does not exhibit strong interactions with LiTFSI and so, enables a conclusion that the presence of

Page 9 of 28
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA05004F
LiTFSI in PTAA and similar polymers would compromise their adhesion to other substrates
regardless of the interaction between LiTFSI or the TFSI anion and a given substrate.
AzPTMS adhesion promotion
The poor adhesion of PTAA and dꢀPTAA to glass required a strategy for covalently
bonding these films to the glass beams used in the double cantilever beam fracture test. This
development would enable the cohesion of PTAA and the influences of LiTFSI doping and
TPBA crossꢀlinking on this cohesion to be measured. To date AzPTMS (Fig. 2b) has been used
to functionalize a variety of substrates with azide groups to enable surface functionalization via
4
5ꢀ47
“click chemistry” reactions
. To date, the use of AzPTMS functionalized substrates for
promoting adhesion through solid state thermal or deep UV activation of the azide groups has
not been investigated.
AzPTMS functionalized glass beams were prepared according to the scheme shown in
Fig. 2b. Covalent bonding of films deposited on AzPTMSꢀfunctionalized substrates were
activated by UVꢀC irradiation to generate reactive nitrene radicals, which readily insert into CꢀH
bonds of material deposited upon the functionalized surface, in our case, PTAA, covalently
bonding our PTAA films to the glass substrates. This novel use of AzPMTS enabled the cohesive
fracture energy of PTAA and the influence of LiTFSI doping on the cohesion of PTAA to be
2
determined (Fig. 3a). The cohesion of PTAA was measured to be 9.10 J/m . Additions of LiTFSI
to PTAA, as expected, decreased the cohesion of PTAA. Addition of 0.127 eq. of LiTFSI, the
2
concentration used in perovskite solar cells, resulted in a 20 % decrease in the G to 7.3 J/m .
c
This result indicates that the TFSI anion does not exhibit favorable interactions with PTAA and
that its presence can be expected to compromise the adhesion of PTAA to a variety of substrates
as previously discussed, increasing the risk of mechanical failure in devices.
Influence of TPBA cross-linking on PTAA cohesion
Using AzPTMSꢀfunctionalized glass beams, the cohesion of undoped PTAA films with
increasing weight equivalents of TPBA (0.01, 0.05, 0.1 and 0.2) could be measured, exceeding
2
energies of 40 J/m without films delaminating. In this work, an upper limit of the adhesion
provided by AzPMTS surface functionalization and bonding was not encountered. The crossꢀ
linking activity of the TPBA crossꢀlinking agent resulted in a pronounced increase in the fracture

Journal of Materials Chemistry A
Page 10 of 28
View Article Online
DOI: 10.1039/C7TA05004F
energy of PTAA crossꢀlinked by TPBA as shown in Fig. 3c. With addition of 0.01 w.eq. of
2
2
TPBA, the G increased by 118 % from 9.1 J/m to 19.9 J/m . Addition of 0.05 w.eq. TPBA
c
resulted in a 151 % increase, 0.1 w.eq., a 264 % increase and finally, 0.2 w.eq. a 345 % increase
2
to 40.5 J/m indicating that TPBA is a highly efficient crossꢀlinking agent and that the
hyperconnectivity introduced into the PTAA polymer results in sharp increases in PTAA fracture
toughness.
The presence of LiTFSI, was found to reduce the ability of TPBA to increase the fracture
toughness of PTAA. Nonetheless, TPBA crossꢀlinking was still found to increase the cohesive
fracture energy of dꢀPTAA, as shown in Fig. 3c. Additions of TPBA to dꢀPTAA resulted in a
2
2
linear increase in the fracture energy from 7.3 J/m without TPBA, to 23.5 J/m with 0.2 w.eq
TPBA, a 221 % increase. TPBA is thus highly effective at increasing the fracture energy of a
fragile semiconducting polymer, even one containing molecular dopants to values well beyond
those observed for borosilicate glass, CIGS and cꢀSi solar cells, which are on the order of ~10
2
J/m .
DFT Energy Level Analysis
An important design consideration of crossꢀlinking agents for semiconducting polymers
is to minimize the electronic hybridization of the crossꢀlinker with the polymer to limit altering
the energy and spatial distribution of the polymer’s electronic orbitals. The crossꢀlinking agent
should feature sigmaꢀbond connectivity between respective azide groups to decouple these
2
8
groups from one another. This is a feature of sFPAs . We use density functional theory (DFT)
calculations performed using the Gaussian09 suite at the B3LYP/6ꢀ31G(d,p) level of theory to
investigate the degree to which the sigmaꢀbonded adamantane core of TPBA limits orbital
hybridization across the crossꢀlinking agent, resulting in reactions of TPBA with polymers being
likened electronically to reactions of polymers with benzylazide. The frontier orbital energies of
TPBA (Fig. 4a), the model compound benzyl azide (BA) (Fig. 4b), and corresponding reaction
products of these two azides with a series of alkane and aromatic small molecules, (R groups)
ranging from short chain alkanes, through benzene, to short acenes; naphthalene, anthracene and
finally coronene were calculated (Fig. 4). This series of small molecules was selected for
computational simplicity, distinct from computationally expensive polymers and provides insight
to the trends in the hybridization of TPBA with molecules of decreasing HOMOꢀLUMO gaps.

Page 11 of 28
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA05004F
In addition, the energies of 1,3,5,7ꢀtetrakisꢀ(pꢀbenzylamine)ꢀadamantane (TPBAꢀ1) and
benzylamine (BAꢀ1) were calculated and are shown in Fig. 4a and b. While these two structures
are not the products of nitrene insertion into CꢀH bonds, but rather, into diatomic hydrogen, the
energies of the frontier orbitals of these two structures better resemble the energies of the states
that will influence the orbital energies of the Rꢀgroups. Comparison of the results of Fig. 4a and
Fig. 4b reveals a similarity in the influence of addition of TPBA and BA to the R groups, and the
resulting TPBAꢀR and BAꢀR constructs.
The differences between calculated frontier orbital energies of BAꢀR and TPBAꢀR,
ꢀEapprox = EBAꢀR ꢀ ETPBAꢀR, were calculated. The distribution of along ꢀEapprox = 0 eV indicates the
value of the benzylamine approximation, the degree to which crossꢀlinking with TPBA is
electronically analogous to benzylamine addition (Fig. 4c). In this light, addition of TPBA to an
alkane group, results in frontier orbital energies characteristic of benzyl amine influenced by the
alkane, whereas, addition of a TPBA to an increasingly π conjugated system with a smaller
HOMOꢀLUMO gaps will exhibit orbital energies dominated by those of the π system, influenced
by addition of benzylamine groups. In the context of mitigating the impact on semiconducting
polymers, which typically exhibit HOMOꢀLUMO gaps smaller than those of TPBAꢀ1, the latter
result can be expected.
The difference in frontier orbital energies between R groups and the TPBAꢀR constructs,
revealing how ꢀE (here ꢀE = E ꢀ ETPBAꢀR) tends to zero with increasing conjugation of the R
R
group (Fig. 4d). These results suggest that when semiconducting polymers are crossꢀlinked by
reaction of TPBA with the aromatic backꢀbones of the polymers (reacting with alkyl sideꢀchains
are not expected to significantly influence the electronic properties of the polymers) the addition
of TPBA will act to destabilize the orbitals slightly, again, being likened to addition of benzyl
amine to the polymer.
The similarity of the first four HOMOs of TPBA further reveals the effective decoupling
of the benzylazide groups by the adamantane core (Fig. 5a). This is even more pronounced in the
reaction product of TPBA with benznene (TPBAꢀ4), in which these first four HOMOs are
largely degenerate and lie predominantly on the added benzene, as can be seen in Fig. 5b. The
LUMO orbitals are less degenerate, yet do exhibit decoupling of the peripheral benzene moieties
from one another. With addition of molecules and polymers with narrower HOMOꢀLUMO gaps,
this further localization of the HOMO away from the TPBA core can be expected.

Journal of Materials Chemistry A
Page 12 of 28
View Article Online
DOI: 10.1039/C7TA05004F
PESA Ionization Potential Analysis
Recently, Belisle et al. reported that the VOC and associated performance of perovskite
4
8
solar cells are quite forgiving to variations in the ionization potential (IP) of the HTM . As such,
for perovskite solar cells, changes in the IP that result from TPBA crossꢀlinking may not have
significant consequence on device performance if the changes are not too large, ca. 300 meV,
while larger changes may influence the performance of these devices, not to mention other
devices using TPBA crossꢀlinked PTAA.
We investigated the influence and synergy of additions of LiTFSI, 4ꢀtertꢀbutyl pyridine
(tBP) and TPBA along with UVꢀC irradiation to activate TPBA on the IP of PTAA using photo
electron spectroscopy in air (PESA) (Fig. 5c). While PESA measurements produce values to the
second decimal, which in some cases are informative, in general it is best practice to place
confidence in only the first decimal, probing 100 meV or larger variances of materials under
investigation. While it is known that LiTFSI oxidizes PTAA, our TPBA crossꢀlinking agent
results in addition of new secondary amines to the PTAA polymer which may either decrease the
IP by adding electron rich groups to the periphery of the polymer, or increase it by facilitating
more efficient oxidation of the resulting crossꢀlinked polymer by LiTSFI. In addition, we
suspected that the presence of tBP, which has been determined to enhance perovskite solar cell
4
9
efficiency through direct interaction with the perovskite material , may also alter the IP of
LiTFSI doped PTAA through direct interaction with LiTFSI or though donation of electron
density to the oxidized polymer, reducing the effective IP of LiTFSIꢀdoped PTAA.
Addition of 0.2 w.eq. of TPBA to PTAA without UVꢀC irradiation results in an increase
of the IP from 5.16 to 5.19 eV. Here, the observed IP is a function of PTAA and the presence of
the deep lying HOMO of TPBA. UVꢀC irradiation of the PTAAꢀTPBA composite film to induce
crossꢀlinking lowers the ionization potential to 5.17 eV due to addition of the benzylꢀazide
groups of TPBA to the polymer, leaving the IP largely unaltered in comparison to PTAA. This
result is consistent with the trends observed in the DFT calculations in which TPBA addition to
increasingly conjugated molecules results in only a slight destabilization of the HOMO.
Addition of LiTFSI to PTAA raises the ionization potential to 5.44 eV, typical of hole
doping. Inclusion of tBP along with LiTFSI does alter the IP of LiTFSI doped PTAA, though far
more than expected, largely nullifying the influence of LiTFSI on the IP, resulting in a film with

Page 13 of 28
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA05004F
an IP of 5.17 eV. This indicates that the activity of tBP in the PTAA HTM may extend beyond
modification of the HTMꢀperovskite interface. Addition of TPBA to a PTAA film containing
both LiTFSI and tBP results in a 0.09 eV increase in the IP to 5.26 eV again with the observed IP
being a function of the composite structure of dꢀPTAA and TPBA. Finally, UVꢀC activation of
TPBA in this mixture, again, results in a marginal reduction of the IP to 5.25 eV indicating that
TPBA crossꢀlinking of PTAA may reꢀenable LiTFSI doping in the polymer typically inhibited
by tBP. Altogether, the IP of dꢀPTAA and dꢀPTAAꢀX differ by only 0.08 eV which is certainly
within the variation in IP tolerated by perovskites (valence band edge ꢀ5.39 eV) in perovskite
solar cells.
Solar Cells using PTAA-X
Increases in fracture resistance and solvent resistance of dꢀPTAA through TPBA crossꢀ
linking do not come at the expense of solar cell performance. In fact, the use of TPBA crossꢀ
linked dꢀPTAAꢀX0.2 results in appreciable increases in the PCE of perovskite solar cells using
PTAAꢀX in both nꢀiꢀp and pꢀiꢀn architectures as shown in Fig. 6 and Table 1, and verified by
maximum power point tracking (Fig. S19 and S20) The nꢀiꢀp cells utilized MAPbI perovskite
3
50
deposited by the acetate method , and the pꢀiꢀn cells the Cs0.05(MA0.17FA
)
Pb(Br I )
0.83 0.95 .17 .83 3
5
1
triple cation perovskite . The perovskite used in each case was determined by the underlying
layer on which perovskite deposition would occur. In the nꢀiꢀp architecture, the chlorobenzene
antiꢀsolvent used in the deposition of the triple cation perovskite etched the underlying fullerene
layer, resulting in the formation of deviceꢀkilling shunts. Similarly, in the case of the pꢀiꢀn
architecture, the precursor solution of the acetate method did not suitably wet the underlying
PTAA film preventing device fabrication. However, the added DMSO of the tripleꢀcation
enabled suitable wetting.
TPBA crossꢀlinking resulted in systematic increases in the efficiencies of nꢀiꢀp cells
(Fig. 6a) from 16.7 % for the dꢀPTAA control, to 16.8 % with addition of 0.1 w.eq of TPBA, and
finally, to 17.2% for 0.2 w.eq. of TPBA. This increase is a product of improvements in the cells’
VOC and FF. Addition of 0.2 w.eq. of TPBA increases the FF from 0.78 to 0.81, and the VOC from
ꢀ2
1
.00 V to 1.04 V. The JSC of cells using dꢀPTAAꢀX0.2 is lowered from 21.5 mA.cm to 20.0
ꢀ2
mA.cm , but without reducing the overall efficiency. Maximum power point tracking of dꢀ
PTAAꢀX cells showed sustained PCEs above that of control dꢀPTAA cells (Fig. S19).
0
.2

Journal of Materials Chemistry A
Page 14 of 28
View Article Online
DOI: 10.1039/C7TA05004F
In the pꢀiꢀn cells studied cells (Fig. 6b), the use of TPBA to form a solvent resistant,
mechanically robust film proved to be tremendously advantageous. Cells utilizing solvent
resistant dꢀPTAAꢀX0.2 dramatically outperformed cells using soluble dꢀPTAA as the HTM. The
solvent resistance of dꢀPTAAꢀX0.2 enabled the fabrication of a pꢀiꢀn solar cell exhibiting a PCE
of 17.7 %, a marked 4.2 percentageꢀpoint efficiency increase in comparison to the dꢀPTAA
control. The contrast in fillꢀfactors of these two cells, 0.53 for dꢀPTAA and 0.78 for dꢀPTAAꢀ
X0.2 highlights the benefits of TPBA induced solvent resistance which prevents etching by the
chlorobenzene antisolvent used in the deposition of the tripleꢀcation perovskite and the formation
of FFꢀkilling shunts in the control cell.
TPBA crossꢀlinking in pꢀiꢀn cells also resulted in increases in the J , from 21.1 to 22.4
SC
ꢀ2
mA.cm and produced cells that exhibited sustained PCEs during maximum power point
tracking. (Fig. S20). These results show that increases in photovoltaic efficiency can be achieved
by pursuing metrics related to reliability, which not only stands to enhance the yield of
manufacture and reliability, but enables new device architectures to be created using solution
state methods. This stands to be particularly advantageous for the field of tandem siliconꢀ
5
2
perovskite and perovskiteꢀperovskite solar cells that will benefit from the use of a roomꢀ
temperature, crossꢀlinkable HTM.
Methods
General Experimental
All anhydrous solvents were purchased from Acros Organics. NaBH , PBr , PbBr , Pb(AcO) , 4ꢀ
4
3
2
2
tert butylypyridine (tBP), formamadinium Iodide (FAI), and hypophosphorous acid (HPA) were
purchased from Sigma Aldrich, PbI , from Tokyo Chemical Industry Co., Methyl Ammonium
3
Iodide (MAI) from Dyesol and NaN from Acros Organics. PTAA was purchased from Solaris
3
Chem. 10 ꢁ/□ ITO coated 2 cm x 2 cm glass pieces were purchased from Xin Yan Technology.
Deuterated NMR solvents were purchased from Acros Organics, Sigma Aldrich and Cambridge
Isotope Laboratories.
o
NMR spectra were acquired at 25 C using a Varian VNMRS spectrometer fitted with a
1
4
ꢀnucleus (CP/HF) ASW probe, operating at a Larmor frequency of 399.7 MHZ for H. MALDIꢀ
TOF mass spectra were acquired using a Bruker Daltronics spectrometer in positive or negative
ion mode using 1,4ꢀdiphenylbutaꢀ1,3ꢀdiene as the matrix. Waters GCT Premier was used in

Page 15 of 28
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA05004F
electron ionization (EI) mode via direct insertion probe. Solid sample was applied to a quartz rod
°
probe tip and the probe inserted into the source. Within the source, the probe was taken to 500 C
°
to aid in volatilization of the samples. The EI source was set at 250 C and the electron energy
was 70 V. PFTBA was used as internal calibration standard for AMM measurement. UV−vis
spectra of MPMIC60 were acquired using an Agilent Carey 6000i UV−vis−NIR spectrometer.
ATR IR spectra were acquired using a Nicolet iS50 FT/IR Spectrometer.
General cleaning procedure for glass and 10 ꢁ/□ ITO coated 2 cm x 2 cm glass (Xin Yan
Technology), used in this work was as follows: slides were cleaned by scrubbing with a
toothbrush using a 1:10 Extran:DI water solution followed by sequential sonication and rinsing
with DI water, acetone and IPA for 15 minute periods. Samples were exposed to UVꢀozone for
15 minutes immediately before deposition of subsequent layers. Tripleꢀcation perovskite films
4
were deposited using the method reported by Saliba et al. MAPbI films were deposited using
3
3
the acetate method described by Zhang et al. C , BCP and Al/Ag layers were deposited by
60
thermal evaporation using an evaporator (Angstrom Åmod) contained within a N glovebox.
2
Synthesis
The synthesis of 1,3,5,7ꢀtetraphenyladamantane was achieved according the procedure reported
4
1
by Reichert et al . and 1,3,5,7ꢀtetrakisꢀ(pꢀformylphenyl)adamantane according to the procedure
4
2
reported by Duncan et al . AzPTMS was synthesized in accordance with the procedure reported
5
3
by Paoprasert et al .
,3,5,7-tetrakis-(p-phenylmethanol)-adamantane (2) A dry, 1L 2ꢀneck round bottom flask
1
containing a magnetic stir bar was charged with 3.08 g (5.57 mmol) of 1,3,5,7ꢀtetrakisꢀ(4ꢀ
formylphenyl)adamantane, 400 ml of dichloromethane and 150 ml of methanol. While stirring,
1
.26 g (33.43 mmol) of NaBH was added and allowed to stir at RT for 16 h. The solution was
4
washed with 2M HCl and the organic layer filtered. The filtrate was added to an equivalent
volume of hexanes, which caused precipitation of more of the product and further isolated by
filtration. The isolated solids were combined and dried under air to yield 3.00 g (5.35 mmol, 96
ꢀ1
%
yield). ). ATR ν/cm : 3261 (m), 2922 (m), 2848 (m) 1409 (m) 1047 (m), 1013 (m). MS (EI):
1
m/z calcd for C H O , 560.2927; found, 560.2905. H NMR (400 MHz, DMSOꢀd6, δ): 7.50 (d,
3
8
40
4
J = 8.4 Hz, 8H), 7.27 (d, J = 8.4 Hz, 8H), 5.08 (t, 5.6 Hz, 4H), 4.46 (d, 5.6 Hz, 8H), 2.06 (s,
1
3
1
2H). C NMR (100 MHz, DMSOꢀd6, δ): 148.15, 139.94, 126.39, 124.80, 62.71, 46.80, 38.70.

Journal of Materials Chemistry A
Page 16 of 28
View Article Online
DOI: 10.1039/C7TA05004F
1,3,5,7-tetrakis-(p-benzylbromide)-adamantane (3) 3 g (5.35 mmol) of 1 and 100 ml of
anhydrous THF were added to an oven dried 500 ml roundꢀbottom flask containing a magnetic
o
stir bar and cooled to 0 C under N . While stirring, 5.07 g (18.73 mmol) of PBr was added
2
3
dropwise. The vessel was allowed to warm to RT slowly and was left to stir for 16 h. 300 ml of
saturated NaHCO was added and allowed to stir for 5 minutes, causing precipitation of the
3
white product. The precipitate was isolated by filtration, washed with 2 x 200 ml water and 2 x
o
2
9
6
00ml methanol and then dried in a vacuum tube at 50 C for 1 day, yielding 4.03 g (4.96 mmol,
ꢀ1
3 % yield). ATR ν/cm : 3544 (sh, m), 3463 (sh, m), 3412 (s), 2930 (m), 2849 (m), 1206 (m),
06 (m). MS (EI): m/z calcd for C H Br , 807.9550; found, 807.9557. The product was largely
3
8
36
4
insoluble in all NMR solvents tested, which included CDCl , Tolueneꢀd , Benzeneꢀd , 1,1,2,2ꢀ
3
8
6
Tetrachloroethaneꢀd , DMSOꢀd and trifluoroacetic acidꢀd . A proton spectrum, however, could
2
6
1
o
be acquired by heating a suspension in 1,1,2,2ꢀtetrachloroethaneꢀd to 145 C. Residual
2
o
aggregate/impurity peaks were present at δ 1.86 ppm which was the chief signal at 25 C. With
1
heating, this peak diminished in size as peaks for 3 grew in. H NMR (400 MHz, 1,1,2,2ꢀ
tetrachloroethaneꢀd , δ): 7.51 (d, J = 8.4 Hz, 8H), 7.57 (d, J = 8.4 Hz, 8H), 4.57 (s, 8H), 2.23 (s,
2
12H), 1.86 (s, signal of residual aggregate/impurity).
1,3,5,7-tetrakis-(p-benzylazide)-adamantane (TPBA) 60 ml of anhydrous DMF was added to
a rubber septumꢀsealed, ovenꢀdried 200 ml roundꢀbottom flask and wrapped in aluminum foil
containing 3.5 g (4.31 mmol) of 2. The vessel was purged with nitrogen gas for 10 minutes,
o
followed by the addition of 1.4 g (21.54 mmol) of NaN . The flask was heated to 65 C and
3
allowed to react with stirring for 16 h, following which 60 ml of water was added. The white
precipitate was isolated by filtration in the dark and washed with 2 x 50 ml water, 2 x 50 ml
o
ethanol and dried in a vacuum tube overnight at 65 C. The product was then recrystallized from
ꢀ1
CHCl :MeOH to yield 2.45 g (3.71 mmol, 86 % yield). ATR ν/cm : 2924 (m), 2851 (m), 2082
3
(
s, N=N=N stretch), 1503 (m), 1241 (s), 769 (m). MS (EI): m/z calcd for C H N , 660.3186;
38 36 12
1
found, 660.3172. H NMR (400 MHz, CDCl , δ): 7.51 (d, J = 8.4 Hz, 8H), 7.32 (d, 8.4 Hz, 8H),
3
1
3
4
.34 (s, 8H), 2.18 (s, 12H). C NMR (100 MHz, CDCl , δ): 149.41, 133.51, 128.50, 125.67,
3
54.58, 47.30, 39.36.
Film Preparation and Cross-linking.

Page 17 of 28
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA05004F
All PTAA films featured in this work for mechanical testing or PESA analysis were deposited
from 15 mg/ml solutions of PTAA (Solaris Chem) in anhydrous toluene (Acros Organics). The
3
9
ratio of LiTFSI (Acros Ogranics) and tBP added were taken from the reports by Yang et al . and
54
Jeon et al. . Accordingly, 11.25 ꢁl of a 170mg/ml solution of LiTFSI in acetonitrile and 6ul of
distilled tBP were added to 1ml of 15mg/ml PTAA to create doped PTAA. To probe the
influence of LiTFSI on PTAA cohesion, 6 ul of tBP was added to 15mg/ml PTAA solution along
with varying quantities of the 170 mg/ml LiTFSI solution as required. Similarly, relevant weight
o
equivalents (0.01, 0.05, 0.1 and 0.2) TPBA were added to films and heated to 80 C prior to
addition of dopants when relevant to ensure complete dissolution of TPBA. Films were deposited
on glass beams, AzPTMSꢀfunctionalized glass or silicon beams (for testing solvent resistance)
by spin coating (spun at 4000 rpm with a 2000 rpm/s ramp rate for 45 seconds) then heated on a
o
hotꢀplate at 95 C for 5 min in a dry air glovebox. Crossꢀlinking was conducted in an N2
2
glovebox by irradiation with 254 nm UV light for 5 minutes (~ 0.01 W/cm ). Control films
containing no TPBA crossꢀlinker were also irradiated. Solvent resistance was tested by dipping
films deposited on silicon for 30 seconds in toluene, removing them and allowing to dry
vertically in air.
AzPTMS Surface Functionalization
Glass beams for AzPTMS surface functionalization were cleaned with freshlyꢀprepared pirhana,
rinsed with DI water, sonicated in methanol, dried under a stream of N and dried in an oven at
2
o
1
00 C for 10 min. AzPTMS surfaceꢀfunctionalization of glass substrates was accomplished in
glass jars with screwꢀtop Teflon lids containing 40 mM solutions of AzPTMS in anhydrous
o
toluene. Jars were sealed and heated to 100 C for 12 hours in the dark. Substrates were then
sequentially rinsed and sonicated (5 min) in toluene (x 2) and isoꢀpropyl alcohol and blown dry
with N gas under low light conditions.
2
Fracture Testing
Doubleꢀcantilever mechanical test structures were fabricated by capping layers of interest
deposited on glass or AzPTMSꢀtreated glass beams with a 5 nm Ti, 200 nm Al metal adhesion
and blocking layer deposited by eꢀbeam evaporation in a Kurt J. Lesker evaporator after film
deposition. Loctite Eꢀ20NS epoxy was then used to bond a second glass beam to this metal

Journal of Materials Chemistry A
Page 18 of 28
View Article Online
DOI: 10.1039/C7TA05004F
surface, completing the sandwiched DCB test structure. The epoxy resin was cured at room
temperature in a drybox with <1 ppm O for 24 hours under pressure to produce a thin, brittle
2
layer before mechanical testing. Edges of the test structures were scraped clean of epoxy with a
razorꢀblade prior to testing.
Fracture testing of double cantilever beams (DCBs) was performed in laboratory air (~25
°
C and ~40 % R.H.) under displacement control in a thinꢀfilm cohesion testing system
(
Delaminator DTS, Menlo Park, CA) which measured the load, P, versus displacement, . The
ꢀ
–
2
critical fracture energy, G (J. m ), can then be calculated from equation 1:
c
ꢄ
ꢄ
ꢄ
1
2ꢃ ꢅ
ℎ
ꢂ
ꢀ
ꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢁ =
ꢊ1 + 0.64 ꢋ ꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀ(1)
ꢉ
ꢂ
ꢄ
ꢈ
ꢆ ꢇ ℎ
ꢅ
where P is the critical load at which crackꢀgrowth occurs, a is the crack length, E′ is the planeꢀ
c
strain elastic modulus, and b and h are the width and halfꢀthickness of the substrates,
respectively. The crack length, a, was estimated from the elastic compliance, dꢀ/dP, using the
relationship in equation 2:
ꢏ/ꢉ
ꢍ∆ ꢆꢇꢈℎꢉ
ꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢅ = ꢌ
×
ꢎ
− 0.64ℎꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀ(2)
ꢍꢃ
8
After mechanical testing, a survey xꢀray photoelectron spectroscopy (XPS, PHI 5000
Versaprobe) scan (0 ꢀ 1000 eV) was made of each of the fractured specimens using
monochromatic Al K xꢀray radiation at 1487 eV to locate the fracture path and thus the weakest
α
layer or interface within the test structure.
PESA Analysis
Ionization potentials of thin films were acquired with a Riken Instruments ACꢀ2 PESA
spectrometer with a light intensity of 5ꢀ10 nW, depending on the samples.
Computational Analysis
Ground state optimizations of the lowest energy structures were calculated using Gaussian 09
software using density functional theory (DFT) employing the Becke, threeꢀparameter, Leeꢀ
YangꢀParr (B3LYP) exchangeꢀcorrelation functional with the 6ꢀ31G (d, p) basis set, for which

Page 19 of 28
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA05004F
there is precedent in evaluating semiconducting materials for use in perovskite solar cells and
5
5ꢀ58
solidꢀstate dyeꢀsensitized solar cells
.
Solar Cells
Conventional (nꢀiꢀp) devices were prepared on cleaned ITO slides. Compact titania layers were
formed by spinꢀcoating an acidic solution of titanium isoꢀpropoxide (Sigma Aldrich) in ethanol
onto cleaned ITOꢀglass substrates at 4000 rpm for 30 seconds before drying at 150 °C for 60
minutes. A 15 nm layer of C was deposited by thermal evaporation. The acetate perovskite
6
0
3
layers featured were deposited using a method based on that reported by Zhang et al.
Accordingly, a 40 wt.% solution of 3:1 molar ratio methylammonium iodide to lead(ii) acetate in
anhydrous N,Nꢀdimethylformamide with 0.75% HPA/Pb(AcO) hypophosphorous acid was
2
passed through a 0.2 ꢁm PTFE filter. The solution of MAI and Pb(AcO) was spun at 2000 rpm,
2
2000 rpm/s ramp rate for 60 seconds. A constant stream of dry air was directed at the sample to
aid solvent evaporation. Perovskite films were allowed to dry further at room temperature and
then annealed at 100 °C for 5 minutes to complete the formation of the perovskite layer. dꢀ
PTAAꢀX0.2 was prepared as described above with additives of 10 ul LiTFSI at 170 mg/mL in
acetonitrile and 6 ul tBP and deposited on the perovskite by spin coating at 4000 rpm, 2000
rpm/s ramp rate for 45 seconds and dried at 75°C for 5 minutes. dꢀPTAAꢀX0.2 were transferred to
a dry N glovebox and irradiated with UVꢀC light for 8 minutes. A top electrode of 100 nm Ag
2
was then deposited by thermal evaporation. Inverted (pꢀiꢀn) cells were fabricated as follows: 5
mg/mL dꢀPTAAꢀX were deposited on freshly cleaned ITO coated glass by spin coating at 4000
0
.2
o
rpm, 2000 rpm/s ramp rate for 45 seconds and subsequently dried on a hotꢀplate at 95 C for 15
minutes. dꢀPTAAꢀX0.2 films requiring UVꢀC activation were transferred into a dry N glovebox
2
and irradiated with UVꢀC light for 8 minutes. Tripleꢀcation perovskite layers were then deposited
4
using the method reported by Saliba et al. Finally, a 45 nm layer of C , a 7 nm layer of BCP
6
0
and a top electrode of 150 nm Al were deposited sequentially by thermal evaporation (Angstrom
Engineering Aꢀmod). Currentꢀvoltage measurements were performed using a Keithley model
2400 digital source meter and 300W xenon lamp (Oriel) solar simulator was used for irradiation.
The lamp was calibrated with an NRELꢀcalibrated KG5 filtered Si reference cell. JꢀV curves
were taken from forward to reverse bias sampled at 0.05V intervals with a 0.5 second delay time
2
at each voltage step before taking data. Cells were masked to expose a 0.12 cm window.

Journal of Materials Chemistry A
Page 20 of 28
View Article Online
DOI: 10.1039/C7TA05004F
Conclusion
In summary, the synthesis of a novel, tetrahedral, azideꢀfunctionalized crossꢀlinking agent
featuring four reactive azide groups, TPBA, has been achieved. The use of 0.2 w.eq. of TPBA in
PTAA and dꢀPTAA hole transporting materials enables their fracture toughness to be enhanced
significantly by 345 % and 220 % respectively, significantly increasing the resistance to solvent
etching and promoting adhesion of dꢀPTAA to organometal trihalide perovskite surfaces. LiTFSI
was found to compromise the cohesion of PTAA and its adhesion to glass, which can be
overcome through the use of AzPTMS lightꢀactivated adhesion promoter. The use of TPBA
crossꢀlinked PTAA in perovskite solar cells results in nꢀiꢀp and pꢀiꢀn perovskite solar cells with
notable increases in efficiency. TBPA crossꢀlinking will enable the fabricating entirely new
device architectures through its ability to transform semiconducting organic films into robust,
solventꢀresistant layers without negatively impacting the performance of devices using them.
Supporting Information
Supporting information file contains ATR, NMR and Mass Spectral data of compounds
synthesized, solvent resistance data, ATR spectra of PTAAꢀX films, UVꢀVis spectra of PTAAꢀX
films and TPBA in solution, XPS analysis of mating fracture surfaces, and maximum power
point tracking of solar cells.
Acknowledgements
This material is based upon work supported by the Department of Energy through the
Bay Area Photovoltaic Consortium under Award Number DEꢀEE0004946
.
References
1
2
3
4
J. H. Heo, S. H. Im, J. H. Noh, T. N. Mandal and C. S. Lim, Nature, 2013, 7, 486–491.
I. McCulloch, M. Heeney, C. Bailey and K. Genevicius, Nat. Mater., 2006, 5, 328–333.
A. Facchetti, Chem. Mater., 2010, 23, 733–758.
P. Heremans, G. H. Gelinck, R. Muller, K.ꢀJ. Baeg, D.ꢀY. Kim and Y.ꢀY. Noh, Chem.
Mater., 2011, 23, 341–358.
5
G. Li, V. Shrotriya, J. Huang, Y. Yao, T. Moriarty, K. Emery and Y. Yang, Nat. Mater.,

Page 21 of 28
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA05004F
2
005, 4, 864–868.
6
7
G. Li, R. Zhu and Y. Yang, Nature Photon, 2012.
M. Berggren, O. Inganäs, G. Gustafsson, J. Rasmusson, M. R. Adresson, T. Hjertberg and
O. Wennerstrom, Nature, 1994, 372, 444–446.
8
9
1
1
M. C. Gather, A. Köhnen, A. Falcou, H. Becker and K. Meerholz, Adv. Funct. Mater., 2007,
1
7, 191–200.
C. Y. Chuang, P. I. Shih, C. H. Chien, F. I. Wu and C. F. Shu, Macromol, 2007, 40, 247–
52.
0 F. Huang, Y. Zhang, M. S. Liu, Y. J. Cheng and A. K. Y. Jen, Adv. Funct. Mater., 2007, 17,
808–3815.
2
3
1 J.ꢀM. Verilhac, M. Benwadih, A.ꢀL. Seiler, S. Jacob, C. Bory, J. Bablet, M. Heitzman, J.
Tallal, L. Barbut, P. Frère, G. Sicard, R. Gwoziecki, I. Chartier, R. Coppard and C.
Serbutoviez, Org. Electron., 2010, 11, 456–462.
12 F. Pastorelli, T. M. Schmidt, M. Hösel, R. R. Søndergaard, M. Jørgensen and F. C. Krebs,
Adv. Eng. Mater., 2016, 18, 51–55.
1
1
3 M. Chang, D. Choi and E. Egap, ACS Appl. Mater. Interfaces, 2016, 8, 13484–13491.
4 J. Shin, T. R. Hong, T. W. Lee, A. Kim, Y. H. Kim, M. J. Cho and D. H. Choi, Adv. Mater.,
2
014, 26, 6031–6035.
15 K.ꢀJ. Baeg, D. Khim, D.ꢀY. Kim, S.ꢀW. Jung, J. B. Koo, I.ꢀK. You, H. Yan, A. Facchetti
and Y.ꢀY. Noh, J. Polym. Sci., Part B: Polym. Phys., 2011, 49, 62–67.
1
1
1
6 A. Teichler, J. Perelaer and U. S. Schubert, J. Mater. Chem. C, 2013, 1, 1910–1925.
7 F. C. Krebs, Sol. Energy Mater. Sol. Cells, 2009, 93, 394–412.
8 A. A. Zakhidov, J. K. Lee, H. H. Fong, J. A. DeFranco, M. Chatzichristidi, P. G. Taylor, C.
K. Ober and G. G. Malliaras, Adv. Mater., 2008, 20, 3481–3484.
1
9 J. K. Lee, P. G. Taylor, A. A. Zakhidov, H. H. Fong, H. S. Hwang, M. Chatzichristidi, G. G.
Malliaras and C. K. Ober, J. Photopolym. Sci. Technol, 2009, 22, 565–569.
0 K. Meerholz, Nature, 2005, 437, 327–328.
2
2
1 C. Bruner, F. Novoa, S. Dupont and R. Dauskardt, ACS Appl. Mater. Interfaces, 2014, 6,
2
1474–21483.
2
2 N. Rolston, B. L. Watson, C. D. Bailie, M. D. McGehee, J. P. Bastos, R. Gehlhaar, J.ꢀE.
Kim, D. Vak, A. T. Mallajosyula, G. Gupta, A. D. Mohite and R. H. Dauskardt, Extreme
Mech., 2016, 9, 353–358.
23 T. F. O'Connor, A. V. Zaretski, B. A. Shiravi, S. Savagatrup, A. D. Printz, M. I. Diaz and D.
J. Lipomi, Energy Environ. Sci., 2013, 7, 370.
2
2
4 C. Bruner and R. Dauskardt, Macromol, 2014, 47, 1117–1121.
5 S. R. Dupont, E. Voroshazi, P. Heremans and R. H. Dauskardt, Org. Electron., 2013, 14,
1
262–1270.
6 J.ꢀH. Kim, I. Lee, T.ꢀS. Kim, N. Rolston, B. L. Watson and R. H. Dauskardt, MRS Bulletin,
017, 42, 115–123.
7 B. Ma, F. Lauterwasser, L. Deng, C. S. Zonte, B. J. Kim, J. M. J. Fréchet, C. Borek and M.
E. Thompson, Chem. Mater., 2007, 19, 4827–4832.
2
2
2
2
8 R.ꢀQ. Png, P.ꢀJ. Chia, J.ꢀC. Tang, B. Liu, S. Sivaramakrishnan, M. Zhou, S.ꢀH. Khong, H.
S. O. Chan, J. H. Burroughes, L.ꢀL. Chua, R. H. Friend and P. K. H. Ho, Nat. Mater., 2010,
9
, 152–158.
9 L. Derue, C. Lecourtier, T. Gorisse, L. Hirsch, O. Dautel and G. Wantz, RSC Adv., 2014, 5,
840–3843.
2
3

Journal of Materials Chemistry A
Page 22 of 28
View Article Online
DOI: 10.1039/C7TA05004F
3
0 L. Derue, O. Dautel, A. Tournebize, M. Drees, H. Pan, S. Berthumeyrie, B. Pavageau, E.
Cloutet, S. Chambon, L. Hirsch, A. Rivaton, P. Hudhomme, A. Facchetti and G. Wantz,
Adv. Mater., 2014, 26, 5831–5838.
3
3
3
3
3
1 B. Liu, R.ꢀQ. Png, L.ꢀH. Zhao, L.ꢀL. Chua, R. H. Friend and P. K. H. Ho, Nat. Commun,
2
012, 3, 1321.
2 C. Tao, M. Aljada, P. E. Shaw, K. H. Lee, H. Cavaye, M. N. Balfour, R. J. Borthwick, M.
James, P. L. Burn, I. R. Gentle and P. Meredith, Adv. Energy Mater., 2013, 3, 105–112.
3 A. Marcinek, M. S. Platz, S. Y. Chan, R. Floresca, K. Rajagopalan, M. Golinski and D.
Watt, The Journal of Physical Chemistry, 1994, 98, 412–419.
4 R. Poe, K. Schnapp, M. J. T. Young, J. Grayzar and M. S. Platz, J. Am. Chem. Soc., 1992,
1
14, 5054–5067.
5 S. Bräse, C. Gil, K. Knepper and V. Zimmermann, Angew. Chem. Int. Ed., 2005, 44, 5188–
240.
5
3
3
3
3
6 R. M. Pinto, M. Guerra and G. Copeland, J. Phys. Chem. A, 2015, 119, 4118–4126.
7 C. Wentrup, J. Phys. Chem. A, 2015, 119, 8256–8257.
8 R. M. Pinto, M. Guerra and M. L. Costa, J. Phys. Chem. A, 2015, 119, 8258–8259.
9 W. S. Yang, J. H. Noh, N. J. Jeon, Y. C. Kim, S. Ryu, J. Seo and S. Il Seok, Science, 2015,
3
48, 1234–1237.
0 N. J. Jeon, J. H. Noh, Y. C. Kim, W. S. Yang, S. Ryu and S. I. Seok, Nat. Mater., 2014, 13,
97–903.
4
8
41 V. R. Reichert and L. J. Mathias, Macromol, 1994, 27, 7015–7023.
42 N. C. Duncan, B. P. Hay, E. W. Hagaman and R. Custelcean, Tetrahedron, 2012, 68, 53–
6
4.
43 H. C. Kolb, M. G. Finn and K. B. Sharpless, Angew. Chem. Int. Ed., 2001, 40, 2004–2021.
44 J. H. Yun, I. Lee, T. S. Kim, M. J. Ko and J. Y. Kim, J. Mater. Chem. A, 2015.
45 B. Uszczyńska, T. Ratajczak, E. Frydrych, H. Maciejewski, M. Figlerowicz, W. T.
Markiewicz and M. K. Chmielewski, Lab Chip, 2012, 12, 1151–1156.
46 P. Sharma and A. P. Singh, RSC Adv., 2014, 4, 43070–43079.
47 J. Gao, X. Zhang, S. Xu, F. Tan, X. Li, Y. Zhang, Z. Qu, X. Quan and J. Liu, Chem. Eur. J.,
2
014, 20, 1957–1963.
8 R. A. Belisle, P. Jain, R. Prasanna, T. Leijtens and M. D. McGehee, ACS Energy Letters,
016, 1, 556–560.
9 S. N. Habisreutinger, N. K. Noel, H. J. Snaith and R. J. Nicholas, Adv. Energy Mater., 2017,
4
4
5
2
7
.
0 W. Zhang, M. Saliba, D. T. Moore, S. K. Pathak, M. T. Hörantner, T. Stergiopoulos, S. D.
Stranks, G. E. Eperon, J. A. AlexanderꢀWebber, A. Abate, A. Sadhanala, S. Yao, Y. Chen,
R. H. Friend, L. A. Estroff, U. Wiesner and H. J. Snaith, Nat. Commun, 2015, 6, 6142.
1 M. Saliba, T. Matsui, J.ꢀY. Seo, K. Domanski, J.ꢀP. CorreaꢀBaena, M. K. Nazeeruddin, S.
M. Zakeeruddin, W. Tress, A. Abate, A. Hagfeldt and M. Grätzel, Energy Environ. Sci.,
5
2
016.
5
2 G. E. Eperon, T. Leijtens, K. A. Bush, R. Prasanna, T. Green, J. T.ꢀW. Wang, D. P.
McMeekin, G. Volonakis, R. L. Milot, R. May, A. Palmstrom, D. J. Slotcavage, R. A.
Belisle, J. B. Patel, E. S. Parrott, R. J. Sutton, W. Ma, F. Moghadam, B. Conings, A.
Babayigit, H. G. Boyen, S. Bent, F. Giustino, L. M. Herz, M. B. Johnston, M. D. McGehee
and H. J. Snaith, Science, 2016, 354, 861–865.
53 P. Paoprasert, J. W. Spalenka, D. L. Peterson, R. E. Ruther, R. J. Hamers, P. G. Evans and

Page 23 of 28
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA05004F
P. Gopalan, J. Mater. Chem., 2010, 20, 2651–2658.
4 N. J. Jeon, J. H. Noh, W. S. Yang, Y. C. Kim, S. Ryu and J. Seo, Nature, 2015, 517, 476–
80.
5
4
5
5
5 H. Chen, D. Bryant, J. Troughton and M. Kirkus, Chem. Mater., 2016, 28, 2515–2518.
6 S. Park, J. H. Heo, C.ꢀH. Cheon, H. Kim, S. H. Im and H. J. Son, J. Mater. Chem. A, 2015,
3
, 24215–24220.
5
5
7 W. J. Chi, Q. S. Li and Z. S. Li, J. Phys. Chem. C, 2015, 119, 8584–8590.
8 D. Alberga, G. F. Mangiatordi and F. Labat, J. Phys. Chem. C, 2015, 119, 23890–23898.
Table 1. Solar cell performance data of pꢀiꢀn and nꢀiꢀp solar cells using dꢀPTAA and TPBA
crossꢀlinked dꢀPTAAꢀX holeꢀtransport materials.
Architecture
HTM
VOC
V)
JSC
FF
PCE
(%)
ꢀ2
(
(mA.cm )
21.5
n-i-p
dꢀPTAA
1.00
1.02
1.04
1.10
1.08
0.78
0.79
0.81
0.58
0.73
16.7
16.8
17.2
13.5
17.7
dꢀPTAAꢀX0.01
dꢀPTAAꢀX0.2
dꢀPTAA
20.0
20
p-i-n
21.1
dꢀPTAAꢀX0.2
22.4

Journal of Materials Chemistry A
Page 24 of 28
View Article Online
DOI: 10.1039/C7TA05004F
Fig. 1. a. Azide crossꢀlinking agents reported previously featuring reactive azide groups and the new tetrahedral
azide crossꢀlinking node of this work, TPBA featuring four reactive azide crossꢀlinking groups. b. Schematic of
crossꢀlinking of polymer strands by the tetrahedral, fourꢀsite crosslinking agent, TPBA via CꢀH insertion of nitrene
radicals generated by UVꢀC (254 nm) irradiation.

Page 25 of 28
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA05004F
Fig. 2. a. Schematic of the synthesis of TBPA. b. Schematic of surface functionalization of an oxide substrate with
AzPTMS and subsequent film deposition, UVꢀC activation of the azide groups in the solidꢀstate resulting in covalent
bonding of the film to the substrate through insertion of nitrene radicals into CꢀH bonds of the polymer film (dark
blue).

Journal of Materials Chemistry A
Page 26 of 28
View Article Online
DOI: 10.1039/C7TA05004F
Fig 3. a. Influence of LiTFSI concentration on fracture energy and pathway of PTAA deposited on glass beams with
AzPTMS adhesion promotion, (red squares ꢀ cohesive fracture) and without AzPTMS adhesion promotion (black squares
–
adhesive fracture). The dashed blue line indicates the increase in fracture energy sustained by AzPTMS surface
treatment. b. Influence of LiTFSI and TPBA concentration on the fracture behavior of dꢀPTAA (red squares) deposited
on CH NH PbI perovskite. When undoped PTAA (blue data point) is used, adhesive failure is not observed and fracture
3
3
3
occurs within the perovskite layer. With addition of LiTFSI (red squares), adhesion at this interface is compromised
resulting in adhesive failure. This adhesive fracture is mitigated with addition of TPBA, resulting in cohesive failure
within the perovskite film. c. Influence of TPBA concentration on PTAA (black squares) and LiTFSIꢀdoped PTAA (dꢀ
PTAA) (red squares) cohesive fracture. TPBA crossꢀlinking significantly increases the fracture energy of both doped and
undoped PTAA.

Page 27 of 28
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA05004F
Fig 4. DFT B3LYP/6ꢀ31G(d,p)computational results of a, TPBA, R groups and products of nitrene
insertion of TPBA with the R groups at the positions indicated by dashed bonds; b, TPBA, R groups and
products of nitrene insertion of BA with the R groups at the positions indicated by dashed bonds. c. The
difference in the HOMO and LUMO energy of the structures BAꢀR and TPBAꢀR showing the strength
with which the influence of TPBA reactivity with various R groups may be likened simply to reactivity
with BA. d. The difference in energy of the frontier orbitals of TPBAꢀR constructs and individual R
groups, indicating the change in energy of these orbitals on reaction of differing R groups with TPBA.
These computationally accessible calculations give insight to the influence of TPBA reactivity on the
band energies of semiconducting polymers.

Journal of Materials Chemistry A
Page 28 of 28
View Article Online
DOI: 10.1039/C7TA05004F
Fig 5. a. DFTꢀcalculated orbital energies of TPBA and TPBAꢀ4 showing how the four highest occupied
molecular orbitals of TBAꢀ4 are pinned by contributions from the benzene to all lie very close to ꢀ5.19
eV. b. DFTꢀcalculated orbital energies and HOMO and LUMO diagrams TPBAꢀ4 showing the absence of
orbital density on the central adamantane core of TPBAꢀ4 and that these four highest occupied molecular
lie predominantly on the periphery of the molecule. The four lowest unoccupied molecular orbitals of
TPBAꢀ4 lie predominantly on the crossꢀlinking agent while still decoupled by the adamantane core of the
agent. c, Influence and synergy of additions of LiTFSI, tBP, TPBA and UVꢀC crossꢀlinking on the
ionization potential of PTAA.
Fig 6. a, JVꢀcurves of conventional, nꢀiꢀp perovskite solar cells using dꢀPTAA and PTAAꢀX HTM. b, JVꢀcurve
of inverted, pꢀiꢀn perovskite solar cell using dꢀPTAA and dꢀPTAAꢀX as the holeꢀtransporting materials. Use of
TPBA to crossꢀlink dꢀPTAA results in increases in the PCE of solar cells in both architectures.