Figure 1 presents the energy diagram of a bulk heterojunc-tion (BHJ) solar cell. The photon-to-electron conversion mechanism involves four fundamental steps. In the fi rst step, the absorption of light causes photoinduced excitation of an electron in the donor from the HOMO to the LUMO, forming a Frenkel exciton [i.e., a coulombically bound a built-in internal fi eld at the interface, thereby dissociating any excitons into their free carriers. Nanocrystals (n-type materials) can, for example, be used as acceptor materials to replace the fullerene units while polymers remain as the donor materials in polymer/fullerene–based solar cells [ 24,25 ] to form donor–acceptor (D–A) interfaces. In the second step, the excitons diffuse to the D–A interfaces within the diffu-sion length to prevent recombination for releasing photon energy and returning to the ground state. In the third step, an exciton at a D–A interface undergoes ultra-rapid CT to form a CT exciton with the hole and electron remaining in the D and A phases, respectively, held together through cou-lombic attraction. In the fourth step, the built-in electric fi eld causes the CT excitons to dissociate into free holes and elec-trons, which then travel through the D and A phases to their respective electrodes. These four steps in the photon-to-elec-tron conversion process are characterized by four conver-sion effi ciencies: the absorption effi ciency ( η A ), the exciton
diffusion effi ciency ( η ED), the charge separation effi ciency ( η CS ), and the charge collection effi ciency ( η CC ), respectively.
The EQE is defi ned, in terms of energy, as the ratio between the number of collected photogenerated charges at the elec-trodes and the number of incident of photons at a particular wavelength ( λ ); it can be expressed as
EQE( ) =λ η λ ηA( )× ED( )λ η× CS( )λ η× CC( )λ (1) The PCE (%), of a photovoltaic device is defi ned as PCE = (Jsc×Voc×FF) / (Pin) (2) where J sc is the short-circuit current density, V oc is the open-circuit voltage, FF is the fi ll factor, and P in is the power of the incident light. Equation ( 3) can be used to provide a theo-retical value of J sc for comparison with the value obtained experimentally from the current density–voltage ( J – V ) measurements:
Yung-Jung Hsu is an Associate Professor in the Department of Materials Science and
Figure 1. Fundamental mechanism of the photon-to-electron conversion process in BHJ solar cells. Reproduced with permission. [ 24 ] Copyright 2012, Elsevier Ltd.
( )EQE( )
sev-eral benefi ts: [ 14 ] i) the light absorption contribution of chal-cogenide QDs is greater than that of fullerene derivatives, leading to enhanced photogeneration of charge carriers; ii) the light absorption of nanocrystals can be tuned to cover a broad spectral range through modifi cations of the sizes and shapes; iii) nanocrystals provide ultrafast and effi cient pho-toinduced charge carrier transfer between D and A materials;
iv) nanocrystals have relatively high electron mobilities and good photostability and chemical stability. Despite all these benefi ts, the capping ligands on the QDs play an important role in their interface properties that determine the extent of mixing with the conjugated polymers as well as the carrier transport in these nanocomposites.
2.1. Preparation and Morphology of Polymer/Nanocrystal QDs Colloidal CdSe QDs are typically synthesized through thermal decomposition (280–300 °C) of an organometallic Cd precursor with a Se precursor in a hot solution of hexylphos-phonic acid, trioctylphosphine oxide (TOPO), tributyl- or trioctylphosphine (TOP), or hexadecylamine (HDA). When the reaction is complete, the solution is suddenly cooled to room temperature to quench any further nucleation; it is then added into methanol to precipitate the nanocrystal CdSe QDs. After washing to remove any excess capping ligands, monodisperse nanocrystal QDs are obtained; their size and
shape can be controlled by the reaction time and the ratio of surfactants, injection volume and the monomer concen-tration, respectively. [ 26,27 ] Figure 2 displays the photolumi-nescence (PL) spectra of CdSe nanocrystals synthesized in TOPO and in a TOPO/HDA mixture. Upon increasing the reaction time, pure TOPO as solvent resulted in a nearly linear increase in the emission line widths, whereas the mixed solvent led to constant emission line widths. [ 28 ]
Figure 3 a displays transmission electron microscopy (TEM) images of CdSe QDs stabilized with TOP and oleic acid (OA) (both pristine and washed), pyridine, and dithiol molecules. The pristine TOP/OA–capped QDs exhibited a regular hexagonal arrangement with the largest averaged interparticle distance of 7.7 nm between the QD centers. An additional ligand exchange process with 1,4-benzenedithiol (dithiol) decreased the distance to 5.5 nm. Figure 3 b pre-sents PL spectra for the CdSe QDs coated with TOP and OA (both pristine and washed), pyridine, and dithiol molecules.
The relative intensity of the fi rst excitonic peak at 640 nm followed the sequence pristine > washed > pyridine-coated QDs; this peak was not observable for the dithiol-coated QDs. Pyridine ligand exchange provided an even higher number of trap states, enabling a decrease in radiative recom-bination as a result of the stabilized electrons in the π rings of the pyridine units. Exchange with dithiol decreased the domain and a connected network in the plane of the fi lm.
In contrast, the TOPO-coated CdSe nanocrystals underwent almost no aggregation at low concentration, because TOPO and MEH-PPV both have nonpolar faces. This morphology provides a rationale for why, in subsequent studies, solar cells required a suffi ciently high weight percentage of nanocrystals (>80%) to ensure effi cient quenching of the PL. [ 30 ]
Figure 3. a) TEM images and b) PL measurement of CdSe QDs presenting various surfactants.
Reproduced with permission. [ 29 ] Copyright 2012, American Chemical Society.
Figure 2. Evolution of the PL spectra during the synthesis of CdSe nanocrystals in a) TOPO and b) TOPO/HDA (molar ratio: 1:4). Reproduced with permission. [ 28 ] Copyright 2002, American Chemical Society.
In 1992, fullerene (C 60 or C 70 )-doped polymers were reported to exhibit pho-toconductivity; [ 31 ] subsequent studies revealed that photocurrents could be achieved when using semiconductor nanocrystals to photosensitize needlelike C 60 crystals. [ 32 ] A cluster shell comprising CdSe QDs surrounded by functionalized C 60 molecules exhibited a superior pho-tocurrent response, with ultrafast electron transfer between the CdSe QD and the functionalized fullerene. [ 33,34 ] An alterna-tive method was developed for the in situ generation of CdSe QDs in a polymer nanocomposite fi lm through thermal decomposition of a complex cadmium pre-cursor blended with the polymer. [ 35 ]
2.2. Cadmium-Based Nanocrystals
Figure 5 presents the chemical structures of some representative polymers that have been applied in the solar cells based on polymer/nanocrystal nanocomposites that are discussed in this Review. Poly-(3-hex-ylthiophene) (P3HT) as an electron-donor material has been attractive for forming P3HT/CdX (X = S, Se) BHJ solar cells
because it allows effective hole transport in its regioregular form. Table 1 lists all reported device characterization data for nanocomposite solar cells incorporating polymer/Cd-based nanocrystal blends. The size and concentration of the CdSe nanorods in the P3HT/CdSe blend solution can be tuned to control the degree of nanocrystal aggregation and to obtain high EQEs (>54%) and PCEs (ca. 2%). [ 36,37 ] In addi-tion, the incorporation of three-dimensional hyperbranched CdSe nanocrystals into solution blends produces an active layer having large and distributed surface area for charge separation. [ 38 ] Other parameters, including the nature of the solvent, [ 39 ] the annealing temperature, the annealing time, [ 40 ] the nanocrystal size, [ 41 ] and the nature of the ligand on CdSe, [ 42,43 ] have also been studied intensively to optimize the device fabrication process. A modifi ed device architecture featuring ZnO nanoparticles as a buffer layer between the active layer (P3HT/CdSe) and the metal electrode (Al) has been demonstrated that the stability of the device enhanced to 60–70% of the original PCE after exposed to ambient condition for 70 days, as compared to a few hours life time without the ZnO layer. [ 44 ]
Recently, a chemical grafting approach was developed to bind CdS QDs onto P3HT nanowires, [ 45 ] allowing control over organic/inorganic phase separation and signifi cantly improving the PCE to 4.1% from 0.6% for the approach without chem-ical grafting. A post-treatment process—immersing the sub-strate with the polymer/CdSe blend fi lm into an acetonitrile solution containing 1% ethanedithiol (EDT)—was used to remove the charged phosphonic acid ligands from the sur-faces of the CdSe nanorods and form monolayer-passivated Figure 4. TEM images of blends of MEH-PPV with pyridine-treated CdSe
and TOPO-coated CdSe at nanocrystal concentrations of 5, 20, and 65% (w/w). Reproduce with permission. [ 30 ] Copyright 1996, American Physical Society.
Figure 5. Chemical structures of representative polymers that have been incorporated in the solar cells based on polymer/nanocrystal nanocomposites that are discussed in this Review.
Cd-thiolate nanorods. Such EDT treatment decreased exciton and charge carrier recombination and increased charge trans-port in the polymer/CdSe hybrid fi lm, resulting in an increase in PCE to 2.9% from 2.2%. [ 46 ] A comparative study indicated that a major fraction of the successfully separated charge car-riers in the P3HT/CdSe blends were trapped in deep localized states, whereas the geminate recombination resulted from only shallow traps in P3HT/fullerene blends. [ 47 ]
Low band gap conjugated polymers have also been studied for organic/inorganic hybrid photovoltaic applications.
Before P3HT was used as an electron-donor material, MEH-PPV [ 30 ] had been blended with CdSe and CdS nanocrystals for use in solar cell devices. The absorption intensity in the vis-ible light range (600–650 nm) increases with the weight per-centage of CdSe in MEH-PPV/CdSe blends; the PCE reached to 0.2% under AM 1.5 conditions at an incident energy of 0.5 mW cm −2 when 90 wt% CdSe was incorporated. Figure 6 presents three possible routes for exciton separation and CT in MEH-PPV/CdSe blends. Excitons created in MEH-PPV can experience charge separation through electrons transferring from the polymer to the nanocrystal (route a). Alternatively, if the exciton is created on the nanocrystal (route c), or trans-fers onto the nanocrystal (route b), the hole can subsequently transfer to the polymer, producing a separated state of an electron on the nanocrystal. If CdSe nanocrystals are replaced by CdS nanocrystals, the excitons generated in the polymer are unable to transfer to the nanocrystals, because the band gap of CdS is larger than that of the polymer. [ 30 ] Photoinduced absorption spectroscopy has been used to observe long-lived positive polarons in MEH-PPV/CdSe blends following elec-tron transfer to the nanocrystals. [ 48 ] Many subsequent studies have focused on using polymer/CdSe nanocrystal systems to
keep improving device performance. Polymer derivatives fea-turing modifi ed side chains, including poly(2-methoxy-5-(3 ′,7′-dimethyloctyloxy)- p -phenylenevinylene) (MDMO-PPV) [ 49,50 ] and poly(triphenylamine acrylate) (PTPAA) [ 51 ] can improve
Figure 6. Routes for exciton transfer and CT in MEH-PPV/CdSe blends.
a) Exciton created in the polymer after absorption, followed by electron transfer onto the nanocrystal. b) Exciton created in the polymer after absorption, followed by exciton transfer onto the nanocrystal, followed by hole transfer onto the polymer. c) Exciton created in the polymer after absorption in the nanocrystal, followed by hole transfer onto the polymer. Reproduced with permission. [ 30 ] Copyright 1996, American Physical Society.
Table 1. Device characterization data for nanocomposite solar cells incorporating polymer/Cd–based QD blends.
Ref. Polymer QD V oc
[V]
J SC [mA cm −2 ]
FF PCE
[%]
Huynh[ 37 ] P3HT CdSe 0.70 5.7 0.40 1.7
Sun a)[ 39 ] P3HT CdSe 0.62 8.79 0.50 2.6
Olson[ 40 ] P3HT CdSe 0.55 6.87 0.47 1.8
Zhou[ 42 ] P3HT CdSe 0.62 5.80 0.56 2.0
Yang[ 41 ] P3HT CdSe 0.70 6.50 0.40 1.9
Qian[ 44 ] P3HT CdSe 0.64 6.30 0.54 2.2
Ren[ 45 ] P3HT CdS 1.10 10.9 0.35 4.1
Albero[ 43 ] P3HT CdS 0.55 5.61 0.66 2.0
Zhou[ 46 ] P3HT CdSe 0.73 7.40 0.54 2.9
Greeham b)[ 30 ] MEH-PPV CdSe 0.50 0.014 0.26 0.2
Sun c)[ 50 ] MDMO-PPV CdSe 0.76 6.42 0.44 2.4
Böhm[ 52 ] MDMO-PPV CdSe 0.72 5.7 N/A 1.7
Snaith d)[ 51 ] PTPAA CdSe 0.85 N/A N/A N/A
Wang[ 54 ] APFO-3 CdSe 0.95 7.23 0.35 2.4
Dayal[ 55 ] PCPDTBT CdSe 0.67 9.02 0.51 3.1
Albero[ 57 ] PCPDTBT CdSe 0.61 6.89 0.28 1.2
Zhou[ 56 ] PCPDTBT CdSe 0.78 9.20 0.49 3.7
Kuo[ 59 ] PDTTTPD CdSe 0.88 7.26 0.46 2.9
a–c) Incident power: a) 92, b) 0.5, and c) 90 mW cm −2 ; d) Illumination conditions: monochromic light; wavelength: 400 nm.
the PCE to over 2%. Treatment of n -butylamine–capped CdSe with hexane could expel aggregated and poorly passivated CdSe nanocrystals, thereby resulting in a fi ner morphology and better long-range connectivity of the nanocrystal network. [ 52 ] Compared with oleic acid as the capping ligand, n -butylamine, with its shorter chain length, provided stronger evidence for hole transfer from the CdSe nanocrystals to MDMO-PPV. [ 53 ] A polyfl uorene copolymer, poly[2,7-(9,9-dioctyl-fl uorene)- alt -5,5-(4′,7′-di-2-thienyl-2′,1′,3′-benzothiadiazole)] (APFO-3), [ 54 ] was developed by tuning the chemical structures in the conjugated a ZnO buffer layer was present. [ 56 ] Using time-resolved spec-troscopic techniques, it was found that the carrier recombina-tion dynamics depended strongly on the charge density and not on the electric fi eld, a phenomenon that was related to the capping agents or the defects at the surface between the semi-conductor nanocrystals and the polymer. [ 57 ] Another electron-withdrawing moiety, the thieno[3,4- c ]pyrrole-4,6-dione (TPD) unit, having a symmetrical, rigidly fused, and coplanar struc-ture, is also benefi cial for solar cells when it is incorporated in conjugated polymers. [ 58 ] The conjugated polymer PDTTTPD, comprising 2,5-di(thiophen-2-yl)thieno[3,2- b ]thiophene and absorp-tion, equal confi nement of electrons and holes, and long-lived excited states. A device prepared with MEH-PPV/PbS as a alt -4,7-(2,1,3-benzothiadiazole)] (PDTPBT), [ 63 ] have been blended with PbS QDs and incorporated with a TiO 2 buffer
PbS) was soaked several times in an acetonitrile solution of 1,4-benzenedithiol to replace the oleic acid ligands. [ 66 ]
Schottky junction solar cells incorporating ternary PbS x Se 1– x QDs have displayed a relatively high V oc of 0.4V from their PbS QDs ( x = 1) and high J sc of 21 mA cm −2 [ 67 ] from their PbSe QDs ( x = 0), achieving PCEs of up to 3.3%. [ 68 ] The BHJ device can be further improved by using a mixture nanostructures of QDs and nanorods with PbS x Se 1– x [ 69 ] or PbS QDs blended with PSBTBT. [ 70 ] A low band gap ( E g = 1.7eV) polymer, PTB1, introduced into the PbS QDs phase exhibited the highest PCE of 2.8% when the PbS QDs were reduced to a size having a corresponding band gap energy of 1.32 eV. [ 71 ] A series of alternative planar bilayer structure of poly(3-octylthiophene)/PbS solar cells has been developed since 2005, [ 72 ] achieving a PCE of 4.2% after controlling the thickness of the polymer and QDs to 15 and 90 nm, respectively. [ 73 ]
When the S atoms are substituted by Se atoms, charge car-rier kinetics probed with femtosecond transient absorption demonstrates that PbSe QDs has potential to exceed the Table 2. Device characterization data for nanocomposite solar cells incorporating polymer/Pb–based QD blends.
Refs. Polymer QD V oc
theoretical maximum thermodynamic conversion effi -ciency. [ 5,74 ] PbSe QDs has also been blended with the poly-mers P3HT and MEH-PPV for use as the active layer for the nanocomposite solar cells. [ 75,76 ] These polymer/PbSe devices exhibited lower PCEs because of the absent of photoin-duced CT in devices based on MDMO-PPV/PbSe and P3HT/
PbSe. [ 77 ] Therefore, most PbSe QDs have been based on Schottky junction structures, which rely on the built-in fi eld driving extraction of electrons between the indium tin oxide and the back contact (Mg or Al). [ 78–83 ] Recently, depleted-heterojunction–type PbSe QDs solar cells incorporating transparent TiO 2 contacts have led to signifi cantly break-throughs in PCEs (>7%). [ 84–87 ]