Comparison of performances of three active layers cascade OPVCs with those obtained from corresponding bi-layers

(1)

Contents lists available atScienceDirect

Solar Energy

journal homepage:www.elsevier.com/locate/solener

Comparison of performances of three active layers cascade OPVCs with those obtained from corresponding bi-layers

L. Cattin

^a

, Z. El Jouad

^b,f

, L. Arzel

^a

, G. Neculqueo

^c

, M. Morsli

^d

, F. Martinez

^e

, M. Addou

^f

, J.C. Bernède

^b,⁎

aInstitut des Matériaux Jean Rouxel (IMN), Université de Nantes, CNRS, 2 rue de la Houssinière, BP 32229, 44322 Nantes Cedex 3, France

bMOLTECH-Anjou, CNRS, UMR 6200, Université de Nantes, 2 rue de la Houssinière, BP 92208, Nantes F-44000, France

cDepartamento de Materiales Avanzados, Comisión Chilena de Energía Nuclear, Amunátegui 95, Santiago de Chile 8340701, Chile

dFaculté des Sciences et des Techniques, 2 rue de la Houssinière, BP 32229, 44322 Nantes Cedex 3, France

eDepartamento de Ciencia de los Materiales, Facultad de Ciencias Físicas y Matemáticas, Universidad de Chile, Casilla 2777, Santiago, Chile

fLMVR, FST, Université Abdelmalek Essaidi, Tanger, Ancienne Route de l’Aéroport, Km 10, Ziaten, BP: 416, Morocco

A R T I C L E I N F O

Keywords:

Ternary organic solar cells Thiophene derivative Phthalocyanine dye Bande structure

A B S T R A C T

In this study, organic photovoltaic cells based on planar heterojunctions using small-molecules were fabricated.

Two cell configurations were used, planar heterojunctions based on a classical electron donor/electron acceptor couple and planar heterojunctions using three active layers, one donor, one acceptor and a central ambipolar layer. The donor is a thiophene derivative called BSTV, the acceptor is the fullerene while the ambipolar material is a phthalocyanine dye, the SubPc. After studying the performances of cells based on the couples SubPc/C60, BSTV/C60and BSTV/SubPc, we explore the dependence of the short circuit current and the external quantum efficiency on the SubPc interlayer thickness in cells with three active layers. By comparison with the optimum couple SubPc/C60the ternary structure BSTV/SubPc/C60exhibits higher short circuit current but smallerfill factor. Nevertheless, for an optimum interlayer thickness of 12.5 nm, the ternary structure allows an improvement of the cell efficiency of 11% by comparison with the highest result obtained with the best couple. As shown by optical measurements, long range Föster energy transfer from the larger band gap donor, BSTV, to the smaller band gap, SubPc, is possible. Therefore, the increase of Jsc is attributed to two-step exciton dissociation process.

Excitons produced in the outer BSTV layer are transferred to the SubPc central layer and then dissociated at the interface.

1. Introduction

Organic photovoltaic cells (OPVCs) constitute a promising source of renewable energy due to some interesting advantages such as ﬂex- ibility, lightweight, possible semi-transparency… They made rapid progress during the last decades (Søndergaard et al., 2013; Oseni and Mola, 2017). To continue this progression it is now necessary to explore new materials and conﬁgurations making it possible to improve various parameters of OPVCs. For instance, synthesis of new organic molecules with optimized HOMO (Highest Occupied Molecular Orbital) value in order to increase the open circuit voltage, Voc (Badgujar et al., 2016;

Baran et al., 2016; Benduhn et al., 2017), use of tandem OPVCs. Ac- tually, it is possible to broaden the absorption range through the tandem OPVCs approach which has granted high eﬃciency. However, the short circuit current density (Jsc) of such device corresponds to that

of the diode giving the smaller Jsc (Lassiter et al., 2013). Therefore, it is required to control carefully the fabrication receipt, which makes it a difficult structure to implement in practical OPVCs. To overcome this difficulty, it is possible to use other device architectures employing three molecular species in a cascade energy alignment. In such structures, the cascade energy alignment allows exciton dissociation and carriers collection without needing recombination layers (Yuen et al., 2011). Therefore the ternary structures used in the present work were proposed to avoid the technical difficulty in growing efficient tandem cells.

We used a ternary organic active layer composed of an electron donor, an acceptor and a central ambipolar layer with complementary absorption in order to increase light harvesting and therefore the short circuit current, Jsc (Hong et al., 2009). In such structures, the interest of using small molecules is their well deﬁned molecular structures, high

https://doi.org/10.1016/j.solener.2018.07.008

Received 20 March 2018; Received in revised form 11 June 2018; Accepted 4 July 2018

⁎Corresponding author.

E-mail address:[email protected](J.C. Bernède).

T

(2)

reliability, easy purification which allows obtaining good reproduci- bility of the OPVCs performances. Moreover small molecules can be easily deposited by sublimation under vacuum, which induces purity improvement and permits easy fabrication of multilayered architectures such as multilayer planar heterojunction (PHJ) using three active organic layers (Chang et al., 2017). The bulk-heterojunction (BHJ) configuration being the most often used, ternary BHJ-OPVCs were already probed with some promising results (Sharapov et al., 2018, Zhu et al., 2017; Lu et al., 2015; Liu et al., 2015). Recently, efficiency higher than 13.7% using ternary nonfullerene polymer solar cells were reported (Ma et al., 2018). However, if not so many ternary PHJ-OPVCs were studied, works concerning these structures are regularly published (Sista et al., 2011; Heidel et al., 2011; Schlenker et al., 2011; Cnops et al., 2014; Stevens and Arango, 2016; Lin et al., 2017). Among these publications, very promising results, 8.4% efficiency, were obtained (Cnops et al., 2014), while new exploratory pathways using these structures are being studied (Lin et al. 2017). More precisely, in OPVCs, the values of the HOMO and the LUMO (lowest unoccupied molecular orbital) of the organic materials are primordial. The values of the LUMO and HOMO of electron donor (ED) relative to those of the electron acceptor (EA) are very important. Actually, in order to obtain a high efficiency of the separation of the exciton charges, the offset of the LUMOs (HOMOs) values must be higher than the energyΔEexcof the exciton. In ternary PHJ the LUMO and HOMO levels of the three active layers must be progressively offset to create cascading heterojunctions.

The central layer enables exciton dissociation on the donor and acceptor side, which allows improving the short circuit current Jsc (Barito et al., 2014). It must be noted that, to be eﬃcient, the central layer of the ternary structure must be ambipolar. For instance, if boron subphthalocyanine chloride (SubPc) is a well known eﬃcient ED (El Jouad et al, 2015) it was also used as electron acceptor (Beaumont et al.

2012), due to it quite high LUMO and HOMO values and to the fact that it is ambipolar with an electron mobility is 8 × 10⁻³cm²V⁻¹s⁻¹ (Ebenhoch et al., 2015). Recently, researches on new small molecules based on thiophene derivatives have shown that an original branched sexithienylene vinylene oligomer ((E)-Bis-1,2-(5,5″-Dimethyl- (2,2′:3′,2″-terthiophene)vinylene (BSTV)) (Fig. 1a) can be used as eﬃ- cient ED in OPVCs (Martinez et al., 2015). In the present work, we introduce this new molecule as ED layer in classical PHJ based on the couple phthalocyanine dye /fullerene acceptor (El Jouad et al, 2015) and we show that the maximum power conversion eﬃciency (η) of the ternary system is 5.18%, which is higher than that (4.66%) of the binary system based on SubPc/C60(boron subphthalocyanine chloride/

fullerene).

2. Experimental section 2.1. Device fabrication

Devices with traditional PHJ configuration ITO/MoO3/CuI/organic active layers/Alq3/Al were fabricated in one run under vacuum (Martinez et al., 2015; El Jouad et al., 2015; Cattin et al., 2009). Here, MoO3/CuI is the hole transporting layer (HTL), while Alq3is the electron transporting layer (ETL). In PHJ, the electron transport layer is usually called “Exciton blocking layer” (Peumans et al., 2000), this layer is often a thin layer of bathocuproine (BCP), it allows to block exciton and to protect the electron acceptor layer from Al diffusion. If BCP is very efficient it is not very stable. Therefore different other layers, more stable are also often used. Among them, it was shown that Alq3is an efficient stable EBL (Berredjem et al., 2007). We use sys- tematically this EBL in all our structures.

ITO coated glass substrates were provided by SOLEMS, while che- mical products were provided by Aldrich and CODEX (France). It is well known that it is necessary to improve the band matching between the electrodes and the organic materials. Here, MoO3/CuI is the hole transporting layer (HTL), while Alq3is the exciton blocking layer (El

Jouad et al., 2015). The deposition rates and thicknesses of the layers were measured with the help of a quartz monitor. These values, except for the BSTV/SubPc binary and the BSTV/SubPc/C60ternary structures, were already optimised and they are: 0.02 nm/s, 3 nm for MoO3(Cattin et al., 2009), 0.005 nm/s, 1.5 nm for CuI, 0.05 nm/s, 20 nm for SubPc (El Jouad et al., 2015), 0.05 nm/s, 18 nm for BSTV (Martinez et al., 2015), 0.05 nm/s and 40 nm for C60, 0.05 nm/s and 9 nm for Alq3and 1 nm/s, 100 nm for Al. It must be noted that we failed in our attempts in improving the OPVC eﬃciency through the modiﬁcation of the HTL or EBL, therefore all the given results were obtained using MoO3(3 nm)/

CuI (1.5 nm) and Alq3(9 nm) as HTL and ABL respectively. During the deposition process the vacuum was 10⁻⁴Pa. The effective area of the devices was 0.10 cm². In order to compare the results obtained with the ternary structure, BSTV/SubPc/C60, to those obtained with the different binary configurations we have fabricated OPVCs with the three possible binary structures: BSTV/C60, SubPc/C60 and BSTV/SubPc. Therefore two devices configurations have been used (Fig. 1b, c), corresponding two four series of OPVCs:

–ITO/MoO3/CuI/BSTV/SubPc/C60/Alq3/Al;

–ITO/ MoO3/CuI/BSTV/C60/Alq3/Al;

–ITO/ MoO3/CuI/SubPc/C60/Alq3/Al;

–ITO/ MoO3/CuI/BSTV/SubPc/Alq3/Al.

2.2. Device characterizations

Electrical characterizations were performed with an automated I-V tester under sun global AM 1.5 simulated solar illumination and in the dark. Performances of photovoltaic cells were measured using a calibrated solar simulator (Oriel 300 W) at 100 mW/cm²light intensity adjusted with a PV reference cell (0.5 cm²CIGS solar cell, calibrated at NREL, USA). Measurements were performed at an ambient atmosphere.

All devices were illuminated through TCO electrodes.

External Quantum Eﬃciency (EQE) was also measured on apparatus

SubPc C₆₀

(a)

(b)

(c)

Fig. 1.(a) Molecules used in this study, (b) binary device structure, (c) ternary device structure.

(3)

built up in our laboratory. The measurements of incident photon-to- current conversion eﬃciency are done with monochromatic continuous light without modulation. The measurement duration for a given wavelength was long enough to reach a steady-state value and to minimize noise.

We proceeded also to some photoluminescent measurements of BSTV. The photoluminescence (PL) spectra were recorded using the Fluorolog Jobin-Yvon spectrophotometer. The sample was excited with a 400 nm line of a ﬁltered Xenon lamp (150 W). Experiments were performed at room temperature.

2.3. Additional characterizations

Optical transmission spectra were recorded using a Carry spectrophotometer. The optical absorption has been measured at wavelengths of 1–0.30μm.

Also we proceeded to some carrier mobility measurement following the same process as that described in (El Jouad et al.; 2015, Blakesley, 2014),i.e.the space charge limited current (SCLC) method was used to estimate hole mobility of SubPc and BSTV. Hole only devices were as follow: ITO/MoO3 (3 nm)/CuI (1.5 nm)/SubPc (or BSTV) (120 nm)/

MoO3(7 nm)/Al. The hole only devices were fabricated by covering the SubPc or BSTV layer deposited onto the HTL, by a high work function ﬁlm. The work function of Au being not more than 5.1 eV, it is expected to lie within the band gap of SubPc, thereby inducing barrier to injec- tion (or collection) of holes. In order to avoid such barrier, a thin MoO3

buffer layer was introduced between the SubPc or BSTVfilm and the top electrode. It was shown that MoO3acts as efficient hole injecting- collecting layer into most conjugated organic material (Park et al., 2010; Blakesley et al., 2014). Hole only devices have been made using the same ITO covered glass substrate than those used to grow OPVCs.

After deposition of the HTL, an organicfilm thick of 120 nm was deposited. The organicfilm was covered with a MoO3film thick of 7 nm.

Finally aluminium was used as the top electrode. The organic ﬁlm probed was thick of 120 nm because it must be thick enough to prevent that interface phenomena dominate those of bulk (Blakesley, 2014).

3. Results

In this work, we investigated a ternary OPVC architecture with three active layers, two ED, BSTV and SubPc and one EA, C60. Firstly, in order to check that the multiple interfaces to be used within the ternary structure are all active, we fabricated series of planar heterojunction OPVCs based upon the binary structures, BSTV/C60, SubPc/C60 and BSTV/SubPc, as control devices. Then, to construct the ternary OPVC, we used the optimal BSTV/C60binary structure, but knowing that we added a third organic layer, we reduced the thickness of BSTV by 2 nm and have inserted a thin SubPc interlayer to achieve the ternary OPVC BSTV/SubPc/C60. The thickness of this SubPc layer was used as parameter.

The J-V characteristics obtained with the binary structures after optimization of the organic layers thicknesses are presented inFigs. 1–3 while their parameters, Voc, Jsc, FF (Fill Factor) andη, are summarized inTable 1.

The reference SubPc (20 nm)/C60(40 nm) OPVC exhibits optimum result with Voc = 1.03 V, Jsc = 8.31 mA/cm², FF = 54.5% leading to η= 4.66% (El Jouad et al., 2015). Replacing SubPc by BSTV (18 nm) leads to interesting, but smaller device performance with Voc = 0.84 V, Jsc = 5.66 mA/cm², FF = 45% leading toη= 2.2 8% (Martinez et al., 2015). The third, and more original OPVC family based on the BSTV/

SubPc bilayer heart, after optimization, gives the lowest performances with Voc = 1.35 V, Jsc = 3.45 mA/cm², FF = 32% andη= 1.35%. The optimum result is obtained with a BSTV layer thick of 18 nm and that of SubPc thick of 15 nm. The very large value of Voc must be related to the large value of LUMOA-HOMOD(♯2 eV) (Fig. 4b), while the poor device performance can be related the absence of HOMO diﬀerence (Fig. 4b).

Then, starting from the structure BSTV (16 nm)/C60 (40 nm), we added a SubPc interlayer to achieve the ternary OPVC, BSTV/SubPc/

C60, the SubPc layer thickness being used as parameter. The corresponding photovoltaic performance parameters are shown inTable 2.

Except in the case of too thick (17.5 nm) or too thin (7.5 nm) layer, by comparison with optimum SubPc/C60binary structure, the addition of a SubPc interlayer in BSTV/C60increases Jsc and Voc, unfortunately, it decreases FF. All this makes that, for the optimal SubPc layer thickness (12.5 nm), the OPVC efficiency is increased by 11% over the optimal SubPc/C60 binary structure (Tables 1 and 2). By introducing 12.5 nm of SubPc in BSTV/C60 structure, η was 5.18% with Voc = 1.06 V, Jsc = 9.98 and FF = 49%. We would note that the main contribution to the improvement of power conversion efficiency is due to the increase of Jsc. The optimal results obtained with the different structures are visualized inFig. 5.

To check the origin of the increase of Jsc in the ternary structures, we have proceeded to the measure of the External Quantum Eﬃciency (EQE) of binary and ternary structures. We also investigated the evo- lution of EQE with SubPc thickness in ternary structures. On the other hand, in order to try to discriminate between the contributions of the three organic layers, we measure the absorption of the monolayer, binary and ternary structures. We proceeded to the measure of the hole mobility in BSTV layers, we also checked the hole mobility in SubPc layers. The hole mobility of SubPc and BSTV was estimated by SCLC method.

First of all, from EQE measurements, integration of the photocurrent over the solar spectrum gives JEQEvalues in good agreement with Jsc deduced from J-V characteristics. InFig. 6the absorption spectra of the different binary and ternary structures are presented, while the normalized EQE curves are shown inFig. 7. It can be seen that, for each structure, the shape of the EQE spectrum reproduces quite well that of the absorption. In the case of the BSTV/C60couple, while the absorption maximum is situated atλ= 385 nm (Fig. 6a), the EQE spectrum of the BSTV/C60 device exhibits a prominent peak visible around λ= 430 nm (Fig. 7). The slight red-shift effect between the two curves is due to some contribution of the C60layer which absorption curve exhibits a shoulder around 430 nm. This effect indicates that carriers issued from excitons created in the C60layer participate to the photocurrent.

In the case of the SubPc/C60binary structure the shape of the EQE spectrum (Fig. 7) follows clearly that of the absorption spectrum of SubPc (Fig. 6b) with a small contribution of C60for wavelengths below 450 nm. In the case of the BSTV/SubPc binary structure, the absorption of BSTV being far higher than that of C60, the contributions of BSTV and SubPc are clearly visible in both spectra (Figs. 6c, 7). A ﬁrst EQE maximum appears atλ= 430 nm, which can be attributed to excitons issued from BSTV and the prominent peak, situated atλ= 590 nm must be attributed to SubPc as shown by the absorption curve ofFig. 6c.

Therefore the EQE curve of the BSTV/SubPc structures demonstrates that both layers are clearly active, though they have the same HOMO value. Some years ago an energy offset of around 0.3 eV was regarded necessary for charge separation, however more recent studies have shown that efficient charge separation can be achieved with smaller energy offset (Chen et al., 2017). Nowadays different studies have shown that efficient junctions with HOMO energy offset smaller than the empirical threshold of 0.3 eV for effective exciton dissociation to overcome the binding energy of the excitons can be very efficient (Bin et al., 2016). Similar results were obtained in the case in the case of nearly equal LUMOs (Baran et al. 2016).

The absorption and EQE spectra of the BSTV/SubPc/C60structures are presented inFigs. 6d and7. It can be seen that the shape of the curve depends on the thickness ratio tBSTV/tSubPc. The contribution of BSTV is clearly visible between 380 nm and 480 nm, while that of SubPc is visible between 480 nm and 750 nm. The SubPc contribution increases when its thickness increases, while that of BSTV decreases progressively. The optimum SubPc thickness, which allows achieving

(4)

the highest eﬃciency, is found to be 12.5 nm.

Finally, the hole mobility of BSTV was estimated by SCLC method, the values obtained were μh= 1.69 × 10⁻⁵cm²/(V s) and μh= 9.5 × 10⁻⁶cm²/(V s) for SubPc and BSTV respectively.

About the value obtained for SubPc, it must be noted that it is in good agreement with values encountered in the literature: from 5.72 × 10⁻⁵ to 1.09 × 10⁻⁶cm²/(V s) following the purity of the SubPcﬁlms (Liu et al., 2012).

4. Discussion

Among the parameters of the J-V characteristics, it is established that, if good ohmic contacts between organic layers and electrodes are achieved, Voc increases with the diﬀerence between LUMOA and HOMODaccording to the expression (Fig. 8):

=

qVoc LUMO - HOMO - Δ_A _D (1)

TheΔvalue depends of the recombination losses which depend on the ED/EA interface quality (Benduhn et al., 2017). Moreover, Voc, and FF, depend also of the possible presence of leakage current which can be estimated though the measure of Js the reverse saturation dark current.

High Js value induces a decrease of Voc and FF (Li et al, 2012, Kulshreshtha et al, 2011). Nevertheless, we have shown earlier that the leakage currents of the OPVCs based on the couples BSTV/C60

(Martinez et al., 2016) or SubPc/C60(El Jouad et al., 2015) are low and of the same order of magnitude (10⁻⁴mA/cm²at−200 mV), which means that the diﬀerence in the Voc value of the OPVC using these

couple is not due to any diﬀerence in the leakage current. Therefore, since the LUMOA-HOMOD value is nearly the same for BSTV/C60

(1.15 eV) and SubPc/C60 (1.10 eV) and the Voc values are quite different, 0.84 eV and 1.03 eV in the former and latter case respectively it can be think that the SubPc/C60 interface quality is better than that of BSTV/C60.

For instance, while the LUMOA-HOMODvalue is nearly the same for BSTV/C60(1.15 eV) and SubPc/C60(1.10 eV) the Voc values are quite diﬀerent, 0.84 eV and 1.03 eV in the former and latter case respectively, which means that the SubPc/C60interface quality is better than that of BSTV/C60. Nevertheless, the tendency given by Eq.(1)is corroborated by the result obtained with the BSTV/SubPc binary structure, even if Voc, 1.35 eV, is far smaller than the maximum possible value, LUMOA- HOMOD= 2.05 V. Nevertheless, this large Voc value obtained, 1.35 eV, is one of the largest ever measured. For instance Voc values of 1.12 eV (Baran et al., 2016), 1.14 eV (Li et al., 2016) were measured in BHJ

0 200 400 600 800

-6 -4 -2 0 2

4

Voc = 0.84 V

Jsc = 5.66 mA/cm

²

FF = 48 %

= 2.28 %

J (mA/cm2 )

V (mV)

(a)

C

60

4.5 eV

6.3 eV 3.18 eV

5.65 eV 1.15 eV (b)

Fig. 2.(a) J-V characteristics of ITO/MoO3/ BSTV (18 nm)/C60(40 nm)/Alq3/Al; (b) Energy levels of SubPc/C60interface.

0 200 400 600 800 1000 1200

-10 -8 -6 -4 -2 0 2 4

J (mA/cm

2

)

V (mV) (a)

Voc = 1.03 V Jsc = 8.31 mA/cm

²

FF = 54.5%

= 4.66%

C

60

4.5 eV

6.3 eV 1.10 eV

3.60 eV

5.60 eV (b)

Fig. 3.(a) J-V characteristics of ITO/MoO3/ SubPc (20 nm)/C60(40 nm)/Alq3/Al; (b) Energy levels of BSTV/C60interface.

Table 1

Parameters for the series of OPVCs with optimal active binary and ternary (ternary) structures.

Structure ED/EA Voc (V) Jsc (mA/cm²) FF (%) η(%)

Binary BSTV/C60 0.84 5.66 48 2.28

SubPc/C60 1.03 8.31 54.5 4.66

BSTV/SubPc 1.35 3.45 32 1.50

Ternary BSTV/SubPc/C60 1.06 9.98 49 5.18

(5)

OPVCs. Actually these high values correspond nearly to the maximum Voc value in BHJ OPVC using optimal optical gap of 1.45–1.65 eV, since it was estimated that for such optimal optical gap, the corresponding maximum Voc values should be 0.93–1.17 eV (Benduhn et al. 2017).

In the case of the ternary structure, Voc is only slightly higher than that of the junction SubPc/C60, which means that as expected, the Voc value is limited by LUMOC60-HOMOBSTV and not by LUMOSubPc-HOMOBSTV(Fig. 5b).

If, it is desirable to maximize LUMOA-HOMODto achieve high Voc value, we must keep in mind that the LUMOD- LUMOAand HOMOD– HOMOAlevel offsets must be sufficiently large to drive charge separation across the ED/EA junction (Fig. 8). These requirements are clearly fulfilled by the BSTV/C60(Fig. 2b) and SubPc/C60(Fig. 3b) structures, which justifies their efficiency as OPVCs. In the case of the of the BSTV/

SubPc interface these requirements are not fulﬁlled, sinceΔHOMO♯0 (Fig. 4b), however this structure exhibits some eﬃciency as OPVC (Fig. 4a), with the contribution of excitons issued from both layers

(Fig. 7). Similar results, but withΔLUMO♯0, were already obtained by Baran et al. (2016). In the present case, even ifΔHOMO♯0, the dif- ference between LUMOs of BSTV and SubPc is about 0.4 eV, providing an enough driving force for exciton dissociation. Nevertheless the poor performance of these OPVCs (Fig. 4a), low FF an Jsc, may be related to the band scheme conﬁguration of this interface.

In ternary OPVCs the results show enhanced short circuit current density (Jsc) due to broader spectral responses. The value of Jsc depends on the charge generation eﬃciency. Generally, the charge generation at an ED/EA interface corresponds to a four steps process (Peumans and Forrest, 2004): (1) Absorption of light and excitons generation, (2) exciton diﬀusion to an interface ED/EA, (3) exciton dissociation and charge carrier generation at the interface and (4) charge carrier collection at the electrodes.

This four steps process is the most commonly used in the case of binary structures. Nevertheless, an alternative migration route is possible. An efficient two step exciton generation at the ED/EA interface was also put in evidence (Coffey et al., 2010). It consists in two-steps exciton dissociation,firstly diffusion from the donor to the acceptor over long length scales, then dissociation. This diffusion over long length is possible through energy transfer by «Föster resonance energy transfer-FRET». The occurrence Föster resonance energy transfer-FRET is well documented in review papers such as those ofStoltzfus et al.

(2016) and An et al. (2016).

This Föster resonance energy transfer from a layer to an adjacent one is eﬃcient if their constituents have complementary optical properties,i.e.if there is a good spectral overlap between the emission of the donor and the absorption of the acceptor. This possibility opens some opportunity for ternary devices. Excitons created in the outer ED layer 0 200 400 600 800 1000 1200 1400 1600 1800

-3 -2 -1 0 1 2

Y Axis Title

V (mV)

Voc = 1.35 V Jsc = 3.45 mA/cm

²

FF = 32%

= 1.50%

(a) 3.18

5.65

3.60 5.60 2.05

SubPc (b)

Fig. 4.(a) J-V characteristics of ITO/MoO₃/BSTV (18 nm)/SubPc (15 nm)/Alq₃/Al; (b) Energy levels of BSTV/SubPc interface.

Table 2

Variation of the ternary OPVCs parameters with SubPc layer thickness, the BSTV thickness being 16 nm and that of C6040 nm.

SubPc (nm) Voc (V) Jsc (mA/cm²) FF (%) η(%)

17.5 1.10 7.96 45 3.94

15.0 1.10 8.65 46 4.81

12.5 1.06 9.98 49 5.18

10.0 1.04 8.55 50 4.49

7.5 0.98 7.49 50 3.67

0 1.03 8.31 54.5 4.66

0 200 400 600 800 1000 1200 1400 -10

-8 -6 -4 -2 0

J (mA/cm2 )

V (mV)

(a)

BSTV(18nm)/C₆₀ SubPc (20nm)/C₆₀

BSTV(16nm)/SubPc(12.5nm)/C₆₀ BSTV(18nm)/SubPc(15 nm)

C

60

4.5 eV

6.3 eV 1.10 eV

3.60 eV

5.60 eV 3.18 eV

5.65 eV 2.05 eV (b)

Fig. 5.(a) Optimal J-V curves obtained with the diﬀerent structures (C60thickness = 40 nm); (b) Energy levels of BSTV/SubPc/C60interfaces.

(6)

of the ternary structure can diffuse into the intermediate donor layer and be dissociated at the interface ED/EA. However, to be efficient, the interlayer must not be too thick, the FRET diffusion length being limited. These charge generation processes, can be applied to our ternary structures. The absorption range is optimized through absorption of light by BSTV and SubPc. The classical process, exciton diffusion to the interface ED/EA and then exciton dissociation and charge carrier generation at this interface, takes place in the PHJ SubPc/C60. The second process, the FRET transfer, takes place for the transfer of the excitons

from BSTV to SubPc, where they can be dissociated at the interface SubPc/C60. The EQE signal between 380 nm and 480 nm proves the participation of excitons created in the BSTV. Actually, as shown in Fig. 9, efficient exciton energy transfer (FRET) from BSTV to SubPc is possible due to their complementary optical properties: there is a good spectral overlap between the emission of BSTV and the absorption of SubPc. BSTV showed a broad emission peak from 490 nm to more than 650 nm, which overlapped well with the absorption spectrum of SubPc, making energy transfer between BSTV and SubPc favourable. Actually, when the BSTV layer is covered with a thin SubPc (10 nm) layer, the photoluminescence signal decreases significantly (Fig. 9). The high efficiency of the SubPc quenching layer indicates that FRET from BSTV to SubPc is very efficient. This Föster resonance energy transfer-FRET is already put in evidence in structure similar to those used in the present worki.e.ternary PHJ-OPVCs such as a-6T/SubNc/SubPc described by

400 500 600 700 800

0,00 0,05 0,10 0,15 0,20 0,25

O. D. (a. u.)

(nm)

(a)

C

₆₀

BSTV

BSTV/C

₆₀

400 600 800

0,00 0,05 0,10 0,15 0,20 0,25

O. D. (a. u.)

(nm)

(

b

⁾

C₆₀ SubPc SubPc/C₆₀

400 500 600 700 800

0,0 0,1 0,2 0,3 0,4 0,5

O. D. (a. u.)

(nm)

(C)

SubPc BSTV

BSTV/SubPc

400 500 600 700 800

0,0 0,1 0,2 0,3

O. D. (a. u.)

(nm)

(d)

BSTV

SubPc C

₆₀

BSTV/SubPc/C

₆₀

Fig. 6.Absorption spectra of (a) BSTV/C60, (b) SubPc/C60, (c) BSTV/SubPc binary structures and (d) BSTV/SubPc/C60ternary structure.

400 500 600 700 800

0,0 0,2 0,4 0,6 0,8 1,0

EQE (a. u.)

(nm)

BSTV (16 nm)/

SubPc (12.5nm)/C₆₀ BSTV (18nm)/C₆₀ BSTV (18 nm)/

SubPc (15 nm) SubPc (20nm)/C₆₀ BSTV(16nm)/

SubPc (17.5nm)/C₆₀ BSTV (16 nm)/

SubPc (7.5nm)/C₆₀

Fig. 7.Normalized External Quantum Eﬃciency spectra of ( ) BSTV (18 nm)/

C₆₀(40 nm), ( ) SubPc (20 nm)/C₆₀(40 nm), ( ) BSTV (18 nm)/SubPc (15 nm) binary structures and (■) BSTV (16 nm)/SubPc (12.5 nm)/C60 (40 nm), ( ) BSTV (16 nm)/SubPc (17.5 nm)/C60 (40 nm) and ( ) BSTV (16 nm)/SubPc (7.5 nm)/C₆₀(40 nm) ternary structures.

(LUMO

_A

- HOMO

_D

) LUMO

ED HOMO LUMO

Eg

HOMO

(LUMO

_D

-LUMO

_A

)

(HOMO

_D

-HOMO

_A

)

Fig. 8.Schematic band structure of an interface electron donor/electron acceptor.

(7)

Cnops et al., 2014. This example is not isolated, for instance it was shown that a stack (thiophene/phenylene)co-oligomers/metal-phthalocyanine/fullerene allows improving the short circuit current through eﬃcient FRET (Ichikawa et al. 2010). In the same way, it was shown that using ternary PHJ pentacene/ZnPc/C60allows improving the EQE and eﬃciency by comparison with pentacene/C60and ZnPc/C60binary PHJ (Hong et al., 2009).

It must be noted that SubPc can act as an active ambipolar interlayer between BSTV donor and C60 acceptor, its electron mobility being μe= 8 × 10⁻³cm²V⁻¹s⁻¹ (Ebenhoch et al., 2015), while its hole mobility isμh= 1.69 × 10⁻⁵cm²V⁻¹s⁻¹. This hole mobility is of the same order of magnitude than that of BSTV, μh= 9.5 × 10⁻⁶cm²/ (V s), which avoid any space charge formation in one of the donor layers and facilitates the hole collection. Moreover, as shown by the results obtained with the BSTV/SubPc binary structure, the BSTV/

SubPc junction can works as dissociation interface for excitons.

Therefore these two mechanisms, FRET transfer and eﬃcient BSTV/

SubPc junction, can explain the contribution of BSTV to the photocurrent. However, the short circuit current density Jsc was shown to increase initially with SubPc layer thickness, then decrease at thickness larger than 12.5 nm (Table 2), indicating that the charge collection and/or optical absorption efficiency was maximized at a 12.5 nm. BSTV has a band gap of 2.45 eV, wider than that of SubPc (2.0 eV), and both EDs have nearly same HOMO level value. Even if the absorption of BSTV is weaker than that of SubPc, it can be used as active layers because of the fact that FRET diffusion length is larger than classical exciton diffusion length, therefore resulting in added absorption. Due to this advantage, improved conversion efficiency by inserting a thin SubPcfilm between BSTV and C60is obtained. The optimal thickness of the SubPc interlayer is 12.5 nm. The contribution to the photocurrent of BSTV and SubPc is asserted from the EQE versus wavelength curves (Fig. 7). Increasing the SubPc thickness above 12.5 nm results in sig- nificant loss in OPVC efficiency, suggesting that FRET diffusion length is below this value (Hong et al., 2009).

FF was the largest (50%) for the OPVC with d = 10 nm or less.

While it was slightly reduced to 49% for the OPVC with d = 12.5 nm, it rapidly decreased to 46% and 44% for thicker films (d = 15 nm and 17.5 nm respectively). As a matter of fact, the series resistance increases with the SubPc thickness from 10Ωto 60Ω(d = 7.5 nm and 17.5 nm), which is likely to reduce the charge transport efficiency and conse- quently lower FF. This negative effect of the insertion of the SubPc interlayer can be attributed to poor charge collection resulting from low charge carrier motilities of SubPc (Cnops et al., 2012). This rapidly decreasing FF therefore shifts the optimal SubPc thickness to lower value (12.5 nm) than in the case of binary structures (20 nm): the ternary cascade devices produce a maximal efficiencyηof 5.18% for a

12.5 nm thick SubPc interlayer.

5. Conclusion

The idea was to create two charge-generating heterojunctions op- erating in series. Firstly we have checked the ability of BSTV to be an eﬃcient ED. Its hole mobility,μh= 9.5x10^-6cm²/(V s), is of the same order than that of SubPc and it gives eﬃcient binary OPVC structures, BSTV/C60and BSTV/SubPc. More precisely, while, in the case of BSTV/

SubPc interface,ΔHOMO♯0, both layers of the couple participate ef- ﬁciently to carrier creation. After optimization of the layers of the ternary structures, the ternary structures give optimum eﬃciency of 5.18%,i.e.11% more than in the case of SubPc/C60the best binary structure, through Jsc increase. This increase corresponds to BSTV contribution through too possible mechanisms: possible exciton energy transfer (FRET) from BSTV to SubPc and active BSTV/SubPc junction.

Work is currently underway on the study of structures using a co- evaporated active layer to improve the exciton dissociation and thus to increase the current Jsc andη.

Acknowledgments

The authors acknowledge funding received from CNRST (PPR/

2015/9) Ministère, Morocco).

References

An, Q., Zhang, F., Zhang, J., Tang, W., Deng, Z., Hu, B., 2016. Versatile ternary organic solar cells: a critical review. Energy Environ. Sci. 9, 281–322.

Badgujar, S., Lee, G.-Y., Park, T., Song, C.E., Park, S., Oh, S., Shin, W.S., Moon, S.-J., Lee, J.-C., Lee, S.K., 2016. High-performance small molecule via tailoring intermolecular interactions and its application in large-area organic photovoltaic modules. Adv.

Energy Mater 1600228.

Baran, D., Kirchartz, T., Wheeler, S., Dimitrov, S., Abdelsamie, M., Gorman, J., Ashraf, R.S., Holliday, S., Wadsworth, A., Gasparini, N., Kaienburg, P., Yan, H., Amassian, A., Brabec, C.J., Durrant, J.R., McCulloch, I., 2016. Reduced voltage losses yield 10%

eﬃcient fullerene free organic solar cells with > 1 V open circuit voltages. Energy Environ. Sci. 9, 3783–3793.

Barito, A., Sykes, M.E., Huang, B., Bilby, D., Frieberg, B., Kim, J., Green, P.F., Shtein, M., 2014. Universal design principles for cascade heterojunction solar cells with highﬁll factors and internal quantum eﬃciencies approaching 100%. Adv. Energy Mater. 4, 1400216.

Beaumont, N., Cho, S.W., Sullivan, P., Newby, D., Smith, K.E., Jones, T.S., 2012. Boron subphthalocyanine chloride as an electron acceptor for high-voltage fullerene-free organic photovoltaics. Adv. Funct. Mater. 22, 561.

Blakesley J.C., 2014a.http://www.npl.co.uk/sciencetechnology/electrochemistry/

research/organic-elecronics/mobility-protocol.

Blakesley, J.C., Castro, F.A., Kylberg, W., Dibb, G.F.A., Arentes, C., Valaski, R., Cremona, M., Kim, J.S., Kim, J.-S., 2014. Towards reliable charge-mobility benchmark measurements for organic semiconductors. Org. Electron. 15, 1263.

Benduhn, J., Tvingstedt, K., Piersimoni, F., Ullbrich, S., Fan, Y., Tropiano, M., McGarry, K.A., Zeika, O., Riede, M.K., Douglas, C.J., Barlow, S., Marder, S.R., 2017. Intrinsic non-radiative voltage losses in fullerene-based organic solar cells. Nat. Energy. 2, 17053.

Berredjem, Y., Karst, N., Boulmokh, A., Gheid, A.H., Drici, A., Bernède, J.C., 2007.

Optimisation of the interface“organic material/aluminium”of CuPc/C60based photovoltaic cells. Eur. Phys. J. Appl. Phys 40, 163–167.

Bin, H., Zhang, Z.-G., Gao, L., Chen, S., Zhong, L., Xue, L., Yang, C., Li, Y., 2016. JACS 138, 4657–4664.

Cattin, L., Dahou, F., Lare, Y., Morsli, M., Tricot, R., Houari, S., Mokrani, A., Jondo, K., Khelil, A., Napo, K., Bernède, J.C., 2009. MoO3surface passivation of the transparent anode in organic solar cells using ultra-thinﬁlms. J. Appl. Phys. 105, 034507.

Chang, Y.J., Hsu, J.-L., Li, Y.-H.S., Biring, Yeh T-H., Guo, J.-Y., Liu, S.-W., 2017.

Carbazole-based small molecules for vacuum-deposited organic photovoltaic devices with open-circuit voltage exceeding 1 V. Org. Electron. 47, 162–173.

Chen, S., Liu, Y., Zhang, L., Chow, P.C.Y., Wang, Z., Zhang, G., Ma, W., Yan, H., 2017. J.

Am. Chem. Soc. 139, 6298–6301.

Cnops, K., Rand, B.P., Cheyns, D., Heremans, P., 2012. Enhanced photocurrent and open- circuit voltage in a 3-layer cascade organic solar cell. Appl. Phys. Lett. 101, 143301.

Cnops, K., Rand, B.P., Cheyns, D., Verreet, B., Empl, M.A., Heremans, P., 2014. 8.4%

eﬃcient fullerene-free organic solar cells exploiting long-range exciton energy transfer. Nat. Commun. 5, 3406.https://doi.org/10.1038/ncomms4406.

Coﬀey, D.C., Ferguson, A.J., Kopidakis, N., Rumbles, G., 2010. Photovoltaic charge generation in organic semiconductors based on long-range energy transfer. ACS Nano 4, 5437–5445.

Ebenhoch, B., Prasetya, N.B.A., Rotello, V.M., Cooke, G., 2015. Solution-processed boron subphthalocyanine derivatives as acceptors for organic bulk-heterojunction solar

500 550 600 650 700

0 2000 4000 6000 8000

Photoluminescence (a. u.)

(nm)

).u.a(.D.O SubPc

BSTV

BSTV/SubPc

Fig. 9.Photoluminescence spectrum of a BSTV layer and optical density spectrum of SubPc.

(8)

cells. J. Mater. Chem. A 3, 7345–7352.

El Jouad, Z., Morsli, M., Louarn, G., Cattin, L., Addou, M., Bernède, J.C., 2015. Improving the eﬃciency of subphthalocyanine based planar organic solar cells through the use of MoO3/CuI double anode buﬀer layer. Solar Energy Mater. Solar Cells 141, 429–435.

Heidel, T.D., Hochbaum, D., Sussman, J.M., Songh, V., Bahlke, M.E., Hiromi, I., Lee, J., Baldo, M.A., 2011. Reducing recombination losses in planar organic photovoltaic cells using multiple step charge separation. J. App. Phys. 109, 104502.

Hong, Z.R., Lessmann, R., Maennig, B., Huang, Q., Harada, K., Riede, M., Leo, K., 2009.

Antenna eﬀects and improved eﬃciency in multiple heterojunction photovoltaic cells based on pentacene, zinc phthalocyanine, and C60. J. Appl. Phys. 106, 064511.

Ichikawa, M., Suto, E., Jeon, H.-G., Taniguchi, Y., 2010. Sensitization of organic photo- volataic cells based on interlayer excitation energy transfert. Org. Electron. 11, 700–704.

Kulshreshtha, C., Choi, J.W., Kim, J.-K., Jeon, W.S., Suh, M.C., Park, Y., Kwon, J.H., 2011.

Open-circuit voltage dependency on hole-extraction layers in planar heterojunction organic solar cells. Appl. Phys. Lett. 99, 023308.

Lassiter, B.E., Renshaw, C.K., Forrest, S.R.J., 2013. Understanding tandem organic photovoltaic cell performance. Appl. Phys. 113, 214505.

Lin, Y.H.L., Koch, M., Brigeman, A.N., Freeman, D.M.E., Zhao, L., Bronstein, H., Giebink, N.C., Scholes, G.D., Rand, B.P., 2017. Enhanced sub-bandgap eﬃciency of a solid- state organic intermediate band solr cell using triplet-triplet annihilation. Energy Environ. Sci. 10, 1465–1475.

Li, N., Stubhan, T., Luechinger, N.A., Halim, S.C., Matt, G.J., Ameri, T., Brabec, C.J., 2012. Inverted structure organic photovoltaic devices employing a low temperature solution processed WO3anode buﬀer layer. Org. Electron. 13, 2479–2484.

Li, S., Liu, W., Shi, M., Mai, J., Lau, La.u.T.-K., Wan, J., Lu, X., Li, C.-Z., Chen, H., 2016. A spirpbiﬂuorene and diketopyrrolopyrrole moities based non-fullerene acceptor for eﬃcient and thermally stable polymer solar cells with high open-circuit voltage.

Energy Environ. Sci. 9, 604–610.

Liu, W., Shi, H., Fu, W., Zuo, L., Wang, L., Chen, H., 2015. Eﬃcient ternary blend polymer solar cells with a bipolar diketopyrrolopyrrole small molecule as cascade material.

Org. Electron. 25, 219–224.

Liu, W., Su, W.-C., Lee, C.-C., Chou, C.-C., Cheng, C.-W., 2012. Eﬃcient organic photovoltaic device using a sublimated subphthalocyanine as an electron donor. ECS Solid State Lett. 1, 70.

Lu, L., Chen, W., Xu, T., Yu, L., 2015. High-performance ternary blend polymer solar cells involving both energy transfer ad hole relay processes. Nat. Commun.https://doi.

org/10.1038/ncomms8327.

Ma, X., Gao, W., Yu, J., An, Q., Zhang, M., Hu, Z., Wang, J., Tang, W., Yang, C., Zhang, F., 2018. Ternary nonfullerene polymer solar cells with eﬃciency > 13.7% by

integrating the advantages of the materials and two binary cells. Energy Environ. Sci.

https://doi.org/10.1039/c8ee01107a.

Martinez, F., Neculqueo, G., Bernède, J.C., Cattin, L., Makha, M., 2015. Inﬂuence of the presence of Ca in the cathode buﬀer layer on the performance and stability of organic photovoltaic cells using a branched sexithienylenevinylene oligomer as electron donor. Phys. Stat. Sol. A 212, 1767–1773.

Martinez, F., El Jouad, Z., Neculqueo, G., Cattin, L., Dabos-Seignon, S., Pacheco, L., Lepleux, E., Predeep, P., Manuvel, J., Thapppily, P., Addou, M., Bernède, J.C., 2016.

Classical or inverted photovoltaic cells: on the importance of the morphology of the organic layers on their power conversion eﬃciency. Dyes Pigm. 132, 185–193.

Oseni, S.O., Mola, G.T., 2017. Properties of functional layers in inverted thinﬁlm organic solar cells. Sol. Energy Mater. Solar Cells 160, 241–256.

Park, H.J., Kang, M.-G., Ahn, S.H., Guo, L.J., 2010. A facile route to polymer solar cells with optimum morphology readily applicable to a Roll-to-Roll process without sa- criﬁcing high device performance. Adv. Mater. 22, E247–E253.

Peumans, P., Bulovic, V., Forrest, S.R., 2000. Eﬃcient photon harvesting at high optical intensities in ultrathin organic double-heterostructure photovoltaic diodes. Appl.

Phys. Lett. 79, 2650.

Peumans, P., Forrest, S.R., 2004. Separation of geminate charge-pairs at donor-acceptor interfaces in disordered solids. Chem. Phys. Lett. 398, 27–31.

Schlenker, C.W., Barlier, V.S., Chin, S.W., Whited, M.T., McAnally, R.E., Forrest, S.R., Thompson, M.E., 2011. Cascade organic solar cells. Chem. Mater. 23, 4132–4140.

Sharapov, V., Wu, Q., Neshchadin, A., Zhao, D., Cai, Z., Chen, W., Yu, L., 2018. High performance ternary organic solar cells due to favored interfacial connection by nonfullerene electron acceptor with cross like molecular geometry. J. Phys. Chem. C 122, 11305–11311.

Sista, S., Yao, Y., Yang, Y., Tang, M.L., Bao, Z., 2007. Enhancement in open circuit voltage through a cascade-type energy band structure. Appl. Phys. Lett. 91, 223508.

Søndergaard, R.R., Hösel, M., Krebs, F.C., 2013. Roll-to-roll fabrication of large area functional organic materials. J. Polym. Sci. Part B: Polym. Phys. 51, 16–34.

Stevens, M.A., Arango, A.C., 2016. Open-Cirduit voltage exceeding the outmost HOMO- LUMO oﬀset in cascade organic solar cells. Org. Electron. 37, 80–84.

Stoltzfus, D.M., Donaghey, J.E., Armin, A., Shaw, P.E., Burn, P.L., Meredith, P., 2016.

Charge generation pethways in organic solar cells: assessing the contribution from the electron acceptor. Chem. Rev. 116, 12920–12955.

Yuen, A.P., Hor, A.-M., Preston, J.S., Kienkler, R., Bamsey, N.M., Loutfy, R.O., 2011. A simple parallel tandem organic solar cell based on metallophthalocyanines. Appl.

Phys. Lett. 98, 17301.

Zhu, K., Tang, D., Zhang, K., Wang, Z., Ding, L., Liu, Y., Yuan, L., Fan, J., Song, B., Zhou, Y., Li, Y., 2017. A two-dimension-conjugated small molecule for eﬃcient ternary organic solar cells. Org. Electron. 48, 179–187.