**4. A new approach: The fabrication of high-performance hybrid polymernanocrystal heterostructures for near-infrared optoelectronics**

As we have seen, PbS nanocrystals are typically synthesized using slightly modified versions of the widely-popular hot-colloidal method (Hines & Scholes, 2003). As such, the oleate capping group keeping colloidal nanocrystals stable in solution also severely limits charge transport between nanocrystals (S. Zhang et al., 2005). For this reason, previouslyproposed near-infrared hybrid LED structures rely on colloidal quantum dots embedded within a polymer host matrix (Choudhury et al., 2010; Konstantatos et al., 2005), or use a monolayer of nanocrystals located directly at the junction of an organic heterostructure (Steckel et al., 2003). In both cases, we have seen that their performances greatly suffer from poor injection efficiencies and from significant carrier losses into the organic layers, thus providing output powers of a few micro-Watts (µW) at best (Konstantatos et al., 2005; X. Ma et al., 2010).

More recently, dithiol-based ligand exchange has been explored as a way of producing highquality self-assembled PbS nanocrystalline film structures such as shown in Figure 12, with application in low-cost photovoltaic and photo-detector platforms (Klem et al., 2008). In this process, short dithiol linker molecules with strong thiolated bonds on both ends are used to exchange the long capping groups around colloidal nanocrystals, resulting in highlyconductive films of strongly-coupled cross-linked nanocrystals. As we will see, this breakthrough also provides new and exciting possibilities for novel near-infrared lightemitting device structures.

## **4.1 Conductivity and mobility of dithiol-treated nanocrystalline PbS film structures**

For the first time, we investigated the optoelectronic properties of PbS nanocrystalline film structures cross-linked using carefully-controlled ethanedithiol (EDT) and benzenedithiol (BDT) ligand-exchange processes. To characterize the electronic properties of those selfassembled nanocrystalline film structures, we use two established methods. First, the charge extraction in linearly increasing voltage (CELIV) method described in Figure 13 can be used to measure both the conductivity and the majority carrier mobility. In view of the

greatly suffer from poor injection efficiencies and from significant carrier losses into the organic layers (Konstantatos et al., 2005), most especially in the electron-transporting layer.

Fig. 11. Incorporation of PbS nanocrystals in polyfluorene-based bilayered type-II

**4. A new approach: The fabrication of high-performance hybrid polymer-**

As we have seen, PbS nanocrystals are typically synthesized using slightly modified versions of the widely-popular hot-colloidal method (Hines & Scholes, 2003). As such, the oleate capping group keeping colloidal nanocrystals stable in solution also severely limits charge transport between nanocrystals (S. Zhang et al., 2005). For this reason, previouslyproposed near-infrared hybrid LED structures rely on colloidal quantum dots embedded within a polymer host matrix (Choudhury et al., 2010; Konstantatos et al., 2005), or use a monolayer of nanocrystals located directly at the junction of an organic heterostructure (Steckel et al., 2003). In both cases, we have seen that their performances greatly suffer from poor injection efficiencies and from significant carrier losses into the organic layers, thus providing output powers of a few micro-Watts (µW) at best (Konstantatos et al., 2005; X. Ma

More recently, dithiol-based ligand exchange has been explored as a way of producing highquality self-assembled PbS nanocrystalline film structures such as shown in Figure 12, with application in low-cost photovoltaic and photo-detector platforms (Klem et al., 2008). In this process, short dithiol linker molecules with strong thiolated bonds on both ends are used to exchange the long capping groups around colloidal nanocrystals, resulting in highlyconductive films of strongly-coupled cross-linked nanocrystals. As we will see, this breakthrough also provides new and exciting possibilities for novel near-infrared light-

**4.1 Conductivity and mobility of dithiol-treated nanocrystalline PbS film structures**  For the first time, we investigated the optoelectronic properties of PbS nanocrystalline film structures cross-linked using carefully-controlled ethanedithiol (EDT) and benzenedithiol (BDT) ligand-exchange processes. To characterize the electronic properties of those selfassembled nanocrystalline film structures, we use two established methods. First, the charge extraction in linearly increasing voltage (CELIV) method described in Figure 13 can be used to measure both the conductivity and the majority carrier mobility. In view of the

**nanocrystal heterostructures for near-infrared optoelectronics**

thin quantum-dot layer at the interface.

et al., 2010).

emitting device structures.

heterostructures. (a) Device schematics. (b) Self-assembled monolayer of PbS quantum dots. (c) Cross-sectional TEM of the structure after lift-off and deposition of the polymer layer atop the structure. (d) Cross-sectional confocal fluorescence intensity mapping showing the

Fig. 12. Self-assembled PbS nanocrystalline film structure. (a) Cross-sectional SEM micrograph of a self-assembled film of PbS colloidal quantum dots formed by dithiol ligandexchange chemistry. The inset shows a typical AFM mapping of the nanocrystalline PbS film surface revealing a 2 nm roughness parameter. (b) Cross-sectional map of the intensity of the 451 cm-1 Raman line associated with the 2LO-phonon vibration of the PbS film structure measured by confocal micro-Raman spectroscopy. (c) Conductive AFM (TUNA) mapping of the nanocrystalline film showing a good conductivity uniform across the surface.

p-type doping of the PbS nanocrystalline films, the CELIV measurement provides us with the hole-mobility for the dithiol-treated nanocrystalline films.

For CELIV measurements, the samples consist of a 250~350 nm-thick nanocrystalline films fabricated using the layer-by-layer spin coating method and sandwiched between ITO and Al contacts such as shown in Figure 13(a). As shown in Figure 13(b,c), a linearly-increasing bias is then applied across the sample while the transient currents are measured through the voltage drop across a 200 Ω load. The conductivity σ can then be obtained using (Juška et al., 2001):

$$
\sigma = \frac{3}{2} \frac{d\Delta \mathbf{j}}{t\_{\text{max}} A} \tag{1}
$$

while the hole mobility µh is calculated using (Juška et al., 2001):

$$\mu\_h = \frac{2d^2}{3At\_{\text{max}}^2 \left(1 + 0.36 \frac{\Delta j}{j\_0}\right)}\tag{2}$$

The parameters tmax, j0 and Δj in those equations can be obtained directly from the CELIV measurement, such as shown in Figure 13(b,c).

In contrast, the time-of-flight (TOF) method can be used to measure the minority carrier mobility. For TOF measurements, the samples have the same structure as for the CELIV measurements but require a much thicker nanocrystalline layer (1.0 ~ 1.5 µm) formed by depositing multiple layers of quantum dots. The 5 ns pulses of a Q-switched Nd:YAG laser are then used to excite the sample from the ITO side, while a reverse-bias is applied on the sample to extract the photo-generated minority carriers. In this case, the transient currents can also be recorded using the voltage drop across a 200 Ω load as shown in Figure 13(d). The electron mobility µe is then directly obtained as (Tiwari & Greenham, 2009):

$$
\mu\_e = \frac{d^2}{V\pi} \tag{3}
$$

Hybrid Polyfluorene-Based Optoelectronic Devices 189

With BDT, the conjugation also provides substantial weight on the thiolated bonds that are directly coupled with the nanocrystals carriers wave-functions (Dadosh et al., 2005), thus directly increasing carrier transport for electrons. Indeed, previous experimental and theoretical models imply that the carrier transport through conjugated aryl dithiol molecules (such as BDT) occurs through the LUMO (electrons) level and not the HOMO (holes) (Nitzan & Ratner, 2003). This is consistent with both the increase of electron mobility and the large offset in electron and hole mobilities we observe for the BDT-treated

**4.2 Hybrid polymer-nanocrystal heterostructures for near-infrared optoelectronics** 

overall performances of polymer-based optoelectronic devices (Moons, 2002).

transport, thus leading to highly-improved LED devices.

As we have seen with both the hybrid polymer-nanocrystal blends discussed in section 3.3 and the hybrid bilayered heterostructures discussed in section 3.4, their performances greatly suffer from poor injection efficiencies and from significant carrier losses into the organic layers (Choudhury et al., 2010; Konstantatos et al., 2005; Steckel et al., 2003). As we mentioned before, constraints on the electron-transporting material especially limit the

As such, a new approach summarized in Figure 15 consists in substituting entirely the electron-transporting organic material, with a self-assembled film of cross-linked PbS nanocrystals deposited atop a hole-transporting polymer film structure. As shown in Figure 15(c), this new hybrid polymer-nanocrystal architecture allows direct electron injection without any additional injection layer and provides significantly better electron

Fig. 15. Novel hybrid polymer-nanocrystal heterostructures for near-infrared LEDs. (a) Device schematics. (b) Energy diagram of the previously studied near-IR LED structures using nanocrystals embedded within a polymer host matrix, or a monolayer of quantum dots introduced at the junction of the organic heterostructure in-between the hole- and electron-transporting organic materials. (c) Energy diagram of the proposed near-IR LED architecture replacing the electron-transporting organic layer, with a self-assembled film of cross-linked PbS nanocrystals deposited atop a hole-transporting polymer film structure.

**4.3 High-performance near-infrared LEDs using hybrid polymer-nanocrystal** 

cost using a very simple all solution-based processing approach.

As shown in Figure 16, high-performance near-IR LEDs can be fabricated at extremely low

nanocrystalline films.

**heterostructures** 

Fig. 13. CELIV and TOF measurements of self-assembled PbS nanocrystalline film structure. (a) Device schematics. (b) For CELIV measurements, a linearly-increasing bias is applied across the sample while the transient currents are measured through the voltage drop across a 200 Ω load. (c) Typical CELIV transient current measured for an EDT-treated PbS film of thickness d= 250 nm using a linearly-increasing voltage with slope A = 400 KV/s. (d) TOF measured for EDT-treated nanocrystalline film structure with d = 1.2 µm under V = 15 V bias (black) and for BDT-treated nanocrystalline film structure with d = 1.1 µm under V = 27 V bias (red).

where d is the sample thickness, V the applied bias across the sample and τ is the transit time obtained from the TOF measurement such as shown in Figure 13(d).

The results of the conductivity, hole- and electron-mobility measurements for both EDTand BDT-treated nanocrystalline film structures are summarized in Figure 14.

Fig. 14. Conductivity, hole- and electron-mobility measurements for EDT- and BDT-treated nanocrystalline film structures. (a) Conductivity measured using CELIV. (b) Electronmobility measured using TOF (black) and hole-mobility measured using CELIV (red).

The smaller conductivity in the BDT-treated films simply suggests a significantly lower p-type doping concentration, most likely resulting from a better passivation of surface states compared with EDT. Moreover, the electron- and hole-mobilities are comparable for EDTtreated films. While we observe a modest increase in the electron mobility for the BDT-treated film despite using longer BDT molecules, it also leads to a significant drop in hole mobility. This is mostly because conjugated-dithiol molecular conductors (such as BDT) were previously shown not only to provide a physical linking between nanocrystals, but also a *conductive path* for electron transfer between nanocrystals through delocalization of the molecular electronic orbitals (Dadosh et al., 2005; Nitzan & Ratner, 2003). As such, the conjugated linker's LUMO and HOMO now provide additional energy barriers between nanocrystals for electrons and holes respectively. Since this energy barrier between nanocrystals is now significantly higher for holes compared with electrons, the carrier transport gets affected accordingly.

Fig. 13. CELIV and TOF measurements of self-assembled PbS nanocrystalline film structure. (a) Device schematics. (b) For CELIV measurements, a linearly-increasing bias is applied across the sample while the transient currents are measured through the voltage drop across a 200 Ω load. (c) Typical CELIV transient current measured for an EDT-treated PbS film of thickness d= 250 nm using a linearly-increasing voltage with slope A = 400 KV/s. (d) TOF measured for EDT-treated nanocrystalline film structure with d = 1.2 µm under V = 15 V bias (black) and for BDT-treated nanocrystalline film structure with d = 1.1 µm under V = 27

where d is the sample thickness, V the applied bias across the sample and τ is the transit

The results of the conductivity, hole- and electron-mobility measurements for both EDT-

Fig. 14. Conductivity, hole- and electron-mobility measurements for EDT- and BDT-treated nanocrystalline film structures. (a) Conductivity measured using CELIV. (b) Electronmobility measured using TOF (black) and hole-mobility measured using CELIV (red).

The smaller conductivity in the BDT-treated films simply suggests a significantly lower p-type doping concentration, most likely resulting from a better passivation of surface states compared with EDT. Moreover, the electron- and hole-mobilities are comparable for EDTtreated films. While we observe a modest increase in the electron mobility for the BDT-treated film despite using longer BDT molecules, it also leads to a significant drop in hole mobility. This is mostly because conjugated-dithiol molecular conductors (such as BDT) were previously shown not only to provide a physical linking between nanocrystals, but also a *conductive path* for electron transfer between nanocrystals through delocalization of the molecular electronic orbitals (Dadosh et al., 2005; Nitzan & Ratner, 2003). As such, the conjugated linker's LUMO and HOMO now provide additional energy barriers between nanocrystals for electrons and holes respectively. Since this energy barrier between nanocrystals is now significantly higher

for holes compared with electrons, the carrier transport gets affected accordingly.

time obtained from the TOF measurement such as shown in Figure 13(d).

and BDT-treated nanocrystalline film structures are summarized in Figure 14.

V bias (red).

With BDT, the conjugation also provides substantial weight on the thiolated bonds that are directly coupled with the nanocrystals carriers wave-functions (Dadosh et al., 2005), thus directly increasing carrier transport for electrons. Indeed, previous experimental and theoretical models imply that the carrier transport through conjugated aryl dithiol molecules (such as BDT) occurs through the LUMO (electrons) level and not the HOMO (holes) (Nitzan & Ratner, 2003). This is consistent with both the increase of electron mobility and the large offset in electron and hole mobilities we observe for the BDT-treated nanocrystalline films.

## **4.2 Hybrid polymer-nanocrystal heterostructures for near-infrared optoelectronics**

As we have seen with both the hybrid polymer-nanocrystal blends discussed in section 3.3 and the hybrid bilayered heterostructures discussed in section 3.4, their performances greatly suffer from poor injection efficiencies and from significant carrier losses into the organic layers (Choudhury et al., 2010; Konstantatos et al., 2005; Steckel et al., 2003). As we mentioned before, constraints on the electron-transporting material especially limit the overall performances of polymer-based optoelectronic devices (Moons, 2002).

As such, a new approach summarized in Figure 15 consists in substituting entirely the electron-transporting organic material, with a self-assembled film of cross-linked PbS nanocrystals deposited atop a hole-transporting polymer film structure. As shown in Figure 15(c), this new hybrid polymer-nanocrystal architecture allows direct electron injection without any additional injection layer and provides significantly better electron transport, thus leading to highly-improved LED devices.

Fig. 15. Novel hybrid polymer-nanocrystal heterostructures for near-infrared LEDs. (a) Device schematics. (b) Energy diagram of the previously studied near-IR LED structures using nanocrystals embedded within a polymer host matrix, or a monolayer of quantum dots introduced at the junction of the organic heterostructure in-between the hole- and electron-transporting organic materials. (c) Energy diagram of the proposed near-IR LED architecture replacing the electron-transporting organic layer, with a self-assembled film of cross-linked PbS nanocrystals deposited atop a hole-transporting polymer film structure.

#### **4.3 High-performance near-infrared LEDs using hybrid polymer-nanocrystal heterostructures**

As shown in Figure 16, high-performance near-IR LEDs can be fabricated at extremely low cost using a very simple all solution-based processing approach.

Hybrid Polyfluorene-Based Optoelectronic Devices 191

magnitude larger than for BDT-treated films. Moreover, the large hole-current would for EDT-treated films would ideally require a hole-barrier at the metal-nanocrystal interface. While we tried to use a TiO2 barrier to reduce the hole-current for such films, we observed

With the BDT-treatment, the hole-mobility drops significantly. As such, there is no need to have a hole-blocking barrier at the metal-PbS interface since the holes don't make it to this interface anyway. Even better, this dramatic reduction in hole mobility is associated with a modest increase in electron mobility. As we know, everything happens at the junction of this hybrid polymer-nanocrystal heterostructure. Using the BDT-treated nanocrystalline films, electrons can be very efficiently injected from one side and holes from the other. Moreover, the hole-transporting polymer bilayer provides an efficient electron barrier while the BDT-treatment provides a good mobility-barrier for holes in the nanocrystalline film. As such, the carriers are efficiently delivered and confined close to the junction (active region). Due to the low hole-mobility, the excitons then bind and stay close to the junction, having plenty of time to recombine radiatively while avoiding metal quenching from the metal-PbS

While π-conjugated polymer-based light-emitting diodes are perfectly suited for the visible, their potential for near-infrared operation remains limited. However, the hybrid integration of semiconductor nanocrystals and conjugated polymer material systems can provide an easy pathway for (1) improving the conjugated polymer-based devices optoelectronic properties and/or (2) providing added functionality to the conjugated polymer-based device structures. Because the oleate capping groups keeping colloidal lead-chalcogenide nanocrystals stable also inhibit carrier transport, previously-proposed hybrid near-infrared LED structures usually rely on nanocrystals embedded within a polymer host matrix, or use a self-assembled monolayer of colloidal quantum dots located at the junction of an organic heterostructure directly between hole- and electron-transporting organics. We have demonstrated why both these hybrid polymer-nanocrystal blends and the hybrid bilayered heterostructures greatly suffer from poor injection efficiencies and from significant carrier losses into the organic layers, while limited electron-transporting materials especially limit

Here, we report an all solution-based method producing efficient hybrid polymernanocrystal multilayered heterostructures for light-emission in the near-infrared (1050-1600 nm). After optimization device structure, we obtain low-cost near-infrared light-emitting diodes with external quantum efficiency (EQE) as high as 0.7% and up to 80 µW output from devices entirely processed in ambient air and with no encapsulation. This approach relies on a carefully-controlled layer-by-layer benzenedithiol (BDT) ligand-exchange to achieve direct charge injection and better transport. In comparison with this BDT treatment, the conventional ethanedithiol (EDT)-based treatment provides poor LED structures. As we show, the high performances of our devices can be explained by the different doping levels

In the future, this easy, robust, low-temperature and substrate-independent approach has the potential to become extremely useful for flexible and/or reconfigurable integrated opto-

the overall performances of those polymer-based optoelectronic devices.

and electron & hole mobilities resulting from the BDT versus EDT treatments.

that this barrier significantly impedes the electron injection.

interface.

**5. Conclusion** 

To fabricate those structures, the ITO substrate is first patterned using standard photolithography and wet-etching processes before spin-coating of a very thin 30 nm-thick layer of PEDOT:PSS. As we mentioned before, this thin layer of PEDOT:PSS plays the dual role of facilitating the hole-injection while alleviating the detrimental effects of the surface roughness of the ITO film on the structural properties of the hole-transporting polymer. A toluene-based TFB solution is then spin-coated to provide a 120 nm-thick TFB film atop the PEDOT:PSS. Since the PEDOT:PSS is immune to toluene, there is no solvent-compatibility issues. Then, a solution of PbS quantum dots suspended in hexane is spin-coated atop the TFB, prior to performing the ligand-exchange process using the dithiol molecule diluted in acetonitrile. While the PEDOT:PSS would be affected by both the hexane and acetonitrile solutions, the TFB provides an efficient protection barrier for both solvents, thus preserving the structural integrity of the whole structure.

Fig. 16. High-Performance near-infrared LEDs using all solution-based processing. (a) Cross-sectional SEM micrograph showing the optimized structure of the device. (b) Device current-voltage characteristics for hybrid LED structure (•) and a polymer-only control device (◦). The inset shows the actual device atop the entry port of an integrating sphere while the near-IR electroluminescence is collected through the transparent substrate. (c) Emission spectra of a typical 1050 nm LED. The inset image shows the actual near-IR emission of a 1 mm2 device measured using a near-IR camera coupled to a 2X objective.

As shown in Figure 16(b), the polymer-only control device reaches a clear single-carrier (hole) trap-limited regime around 1 Volt, before reaching a space-charge limited operation regime around 2 Volts. This is consistent with the large energy barrier at the TFB-aluminum interface. For the LED device with the BDT-treated nanocrytalline film structure, measurements indicate a much higher current density at low voltages originating from the efficient electron-injection at the metal-PbS interface. Here, the lower slope in the traplimited region simply suggests different transport and trapping mechanisms in the nanocrystalline film compared with the TFB. These highly-efficient LED structures can operate anywhere between 1000 and 1600 nm depending on the nanocrystals used while providing external quantum efficiencies as high as 0.7% and ouput powers close to 80 µW.

While the conventional ethanedithiol (EDT)-based ligand-exchange treatment is known to work well for photovoltaic structures (Luther et al., 2008), it yields only relatively poor LED structures compared to the phenomenal results achieved using benzendithiol (BDT)-based treatment. This disparity can be readily explained now based on the conductivity and mobility results presented in Figure 14. Indeed, the EDT treatement provides higher conductivities due to higher p-type doping and comparable electron- and hole- mobilities. As such, both the hole current and nonradiative Auger recombination will be orders of magnitude larger than for BDT-treated films. Moreover, the large hole-current would for EDT-treated films would ideally require a hole-barrier at the metal-nanocrystal interface. While we tried to use a TiO2 barrier to reduce the hole-current for such films, we observed that this barrier significantly impedes the electron injection.

With the BDT-treatment, the hole-mobility drops significantly. As such, there is no need to have a hole-blocking barrier at the metal-PbS interface since the holes don't make it to this interface anyway. Even better, this dramatic reduction in hole mobility is associated with a modest increase in electron mobility. As we know, everything happens at the junction of this hybrid polymer-nanocrystal heterostructure. Using the BDT-treated nanocrystalline films, electrons can be very efficiently injected from one side and holes from the other. Moreover, the hole-transporting polymer bilayer provides an efficient electron barrier while the BDT-treatment provides a good mobility-barrier for holes in the nanocrystalline film. As such, the carriers are efficiently delivered and confined close to the junction (active region). Due to the low hole-mobility, the excitons then bind and stay close to the junction, having plenty of time to recombine radiatively while avoiding metal quenching from the metal-PbS interface.
